MOLECULAR CHAPERONES AND FOLDING CATALYSTS
MOLECULAR CHAPERONES AND FOLDING CATALYSTS Regulation, Cellular Function and Mechanisms Edited by
Bernd Bukau Institute for Biochemistry and Molecular Biology University of Freiburg Germany
harwood academic publishers Australia • Canada • China • France • Germany • India Japan • Luxembourg • Malaysia • The Netherlands Russia • Singapore • Switzerland
This edition published in the Taylor & Francis e-Library, 2005. “To purchase your own copy copy of this or any of taylor & Francis or Routledge's collection of thousands of ebooks please go to www.eBookstore.tandf.co.uk.” Copyright © 1999 OPA (Overseas Publishers Association) N.V. Published by license under the Harwood Academic Publishers imprint, part of The Gordon and Breach Publishing Group. All rights reserved. No part of this book may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying and recording, or by any information storage or retrieval system, without permission in writing from the publisher. Printed in Singapore. Amsteldijk 166 1st Floor 1079 LH Amsterdam The Netherlands British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library. Molecular chaperones and folding catalysts: regulation, cellular functions and mechanisms 1. Molecular chaperones 2. Protein folding I. Bukau, Bernd 572.6′45 ISBN 0-203-30375-X Master e-book ISBN
ISBN 0-203-34398-0 (Adobe eReader Format) ISBN: 90-5702-370-9 (Print Edition) The cover illustration shows schematically a 100 nm window of the cytoplasm of E. coli, depicting the macromolecular components with their estimated sizes. This cellular environment, in which assisted protein folding occurs, is characterized by extraordinary macromolecular crowding. This illustration is a modified version of Figure 1a reprinted from TIBS, Vol. 16, David Goodsell: “Inside a living cell”, pages 203–206, 1991, with permission from Elsevier Science.
To Anette
CONTENTS
Preface Contributors
xi xiii
I. INTRODUCTION 1. Assisted protein folding B.Bukau , F.X.Schmid and J.Buchner
3
II. REGULATION 2. Autoregulation of the heat shock response in prokaryotes L.Connolly , T.Yura and C.A.Gross 3. Inducible transcriptional regulation of heat shock genes: The stress signal and the unfolded protein response R.I.Morimoto 4. Protein kinase cascades involved in heat shock protein expression and function O.Bensaude 5. Thermotolerance and stress response: Possible involvement of Ku autoantigen G.C.Li , L.Li , D.Kim , A.Nussenzweig , S.-H.Yang , P.Burgman , H.Ouyang and C.C.Ling
13 39
59 85
III. CELLULAR FUNCTIONS
A. Overview and physiological aspects 6. Genetic evidence for the roles of molecular chaperones in Escherichia coli metabolism W.F.Burkholder and M.E.Gottesman 7. Genetic dissection of the Hsp70 chaperone system of yeast E.Craig , W.Yan and P.James 8. Functions in development M.Morange
116
155 180
B. Assisted protein folding processes: From ribosomes to proteases 9. Early events in the synthesis and maturation of polypeptides W.J.Welch , D.K.Eggers , W.J.Hansen and H.Nagata 10. Protein transport into and folding within the endoplasmic reticulum I.G.Haas and R.Zimmermann 11. The role of molecular chaperones in transport and folding of mitochondrial proteins P.J.T.Dekker and N.Pfanner 12. Protein import into and folding within chloroplasts E.Muckel and J.Soll 13. Protein folding in the periplasm of Escherichia coli D.Missiakas , C.Dartigalongue and S.Raina 14. Role of chaperones in replication of bacteriophage lambda DNA M.Zylicz , A.Wawrzynow , J.Marszalek , K.Liberek , B.Banecki , I.Konieczny , A.Blaszczak , P.Barski , J.Jakóbkiewicz , M.Gonciarz-Swiatek , M.Duchniewicz , J.Puzewicz and J.Krzewska 15. Control of hormone receptor function by molecular chaperones and folding catalysts D.O.Toft 16. Role of chaperones in uncoating of clathrin coated vesicles E.Eisenberg and L.Greene 17. The role of Hsp104 in stress tolerance and prion maintenance S.Lindquist and E.C.Schirmer 18. Chaperones and charonins: Protein unfolding enzymes and proteolysis M.R.Maurizi , S.Wickner and S.Gottesman
196 226 260
291 310 325
346
365 384 421
IV. MECHANISMS
A. Overview 19. Spontaneous versus assisted protein folding R.Jaenicke and R.Seckler
448
B. Folding catalysts 20. Protein disulphide-isomerase: A catalyst of thiol:disulphide interchange and associated protein folding R.B.Freedman and P.Klappa 21. Peptidyl-prolyl cis/trans isomerases G.Fischer and F.X.Schmid
479
504
C. Chaperonins 22. The ATPase cycle of the GroE molecular chaperones N.A.Ranson and A.R.Clarke 23. The relationship between chaperonin structure and function S.G.Burston and H.R.Saibil 24. Composition and function of the eukaryotic cytosolic chaperonin-containing TCP-1 K.R.Willison
537 570 605
D. Chaperones 25. Structure and mechanism of Hsp70 proteins J.-H.Ha , E.R.Johnson , D.B.McKay , M.C.Sousa , S.Takeda and S.M.Wilbanks 26. The DnaK chaperone system: Mechanism and comparison with other Hsp70 systems A.Buchberger , J.Reinstein and B.Bukau 27. Mechanisms of ATP-independent vs. ATP-dependent chaperones S.Bose , M.Ehrnsperger and J.Buchner 28. Structure and function of periplasmic PapD-like chaperones involved in assembly of bacterial P pili S.J.Hultgren , D.L.Hung , C.H.Jones and S.Knight Index
625 663
693 722
747
PREFACE One of the most intriguing discoveries in molecular biology in the last decade is the existence of an evolutionary conserved and essential system, consisting of molecular chaperones and folding catalysts, which promotes the folding of proteins in the cell. This volume summarizes our current knowledge of the cellular roles, the regulation and the mechanism of action of this system. It has a broad scope, covering cell biological, genetic and biochemical aspects of protein folding in cells from bacteria to man. The first section provides an overview of the diverse families of molecular chaperones and catalysts and the general principles of their action. The second section discusses the regulation of chaperone gene expression in response to stress. The third section summarizes the roles of chaperones and catalysts in cell physiology, followed by a detailed description of their roles in the life span of proteins, from the de novo folding at translating ribosomes to the aggregation and proteolytic degradation of misfolded proteins. The fourth section presents a detailed discussion of our current knowledge on the mechanisms of action of chaperones and folding catalysts. This volume is aimed at researchers working in basic and applied aspects of molecular biology, biochemistry and molecular medicine, and should be useful as an up-to-date reference book and a textbook for specialized university courses. The editor would like to thank the authors for their contributions and their efforts to make this book as up to date as possible, and his secretary Patricia Müller for expert help in preparation of the manuscripts.
CONTRIBUTORS Bogdan Banecki Department of Molecular and Cellular Biology Faculty of Biotechnology University of Gdansk 80–822 Gdansk, Kladki 24 Poland Piotr Barski Department of Molecular and Cellular Biology Faculty of Biotechnology University of Gdansk 80–822 Gdansk, Kladki 24 Poland Olivier Bensaude Unité de Génétique Moléculaire Département de Biologie École Normale Supérieure 46 rue d’Ulm 75230 Paris Cedex 05 France Adam Blaszczak Polish Academy of Science Institute of Biochemistry and Biophysics Laboratory of Molecular Biology University of Gdansk 80–822 Gdansk, Kladki 24 Poland Suchira Bose Department of Biochemistry University of Bristol School of Medical Sciences Bristol BS8 1TD UK Alexander Buchberger Centre for Protein Engineering Medical Research Council Centre Hills Road Cambridge CB2 2QH UK
Johannes Buchner Institut für Biophysik und Physikalische Biochemie Universität Regensburg Universitätsstr. 31 D-93040 Regensburg Germany Bernd Bukau Institut für Biochemie und Molekularbiologie Universität Freiburg Hermann-Herder-Str. 7 D-79104 Freiburg Germany P.Burgman Departments of Radiation Oncology and Medical Physics Memorial Sloan-Kettering Cancer Center 1275 York Avenue New York, NY 10021 USA William F.Burkholder Department of Biochemistry and Molecular Biophysics Institute of Cancer Research College of Physicians and Surgeons Columbia University 701 W168 Street New York, NY 10032 USA Steven G.Burston Department of Genetics Boyer Center for Molecular Medicine Yale University School of Medicine 295 Congress Avenue New Haven, CT 06510 USA Anthony R.Clarke Department of Biochemistry School of Medical Sciences University of Bristol Bristol BS8 1TD UK Lynn Connolly Department of Biochemistry and Biophysics
University of California San Francisco, CA 94143 USA Elizabeth Craig Department of Biomolecular Chemistry University of Wisconsin 1300 University Avenue Madison, WI 53706 USA Peter J.T.Dekker Institut für Biochemie und Molekularbiologie Universität Freiburg Hermann-Herder-Str. 7 D-79104 Freiburg Germany Marlena Duchniewicz Department of Molecular and Cellular Biology Faculty of Biotechnology University of Gdansk 80–822 Gdansk, Kladki 24 Poland Daryl K.Eggers Departments of Medicine and Physiology Lung Biology Research Center University of California Box 0854 San Francisco, CA 94143 USA Monika Ehrnsperger Institut für Biophysik und Physikalische Biochemie Universität Regensburg Universitätsstr. 31 D-93040 Regensburg Germany Evan Eisenberg Laboratory of Cell Biology National Heart, Lung, and Blood Institute Bethesda, MD 20892 USA
Gunter Fischer Max-Planck-Gesellschaft Arbeitsgruppe “Enzymologie der Peptidbindung” Weinbergweg 16a D-06120 Halle/Saale Germany Robert B.Freedman Research School of Biosciences University of Kent Canterbury Kent CT2 7NJ UK Malgorzata Gonciarz-Swiatek Department of Molecular and Cellular Biology Faculty of Biotechnology University of Gdansk 80–822 Gdansk, Kladki 24 Poland Max E.Gottesman Department of Biochemistry and Molecular Biophysics Institute of Cancer Research College of Physicians and Surgeons Columbia University 701 W168 Street New York, NY 10032 USA Susan Gottesman Laboratory of Molecular Biology National Cancer Institute Bethesda, MD 20892 USA Lois Greene Laboratory of Cell Biology National Heart, Lung, and Blood Institute Bethesda, MD 20892 USA Carol A.Gross Departments of Stomatology, and Microbiology and Immunology
University of California Box 0512, S534 San Francisco, CA 94143 USA Jeung-Hoi Ha Department of Structural Biology Stanford University School of Medicine Stanford, CA 94305–5400 USA Ingrid G.Haas Institut für Biochemie I Universität Heidelberg Im Neuenheimer Feld 328 D-69120 Heidelberg Germany William J.Hansen Departments of Medicine and Physiology Lung Biology Research Center University of California Box 0854 San Francisco, CA 94143 USA Scott J.Hultgren Department of Molecular Microbiology Washington University School of Medicine 660 S. Euclid Avenue, Box 8230 St. Louis, MO 63110 USA Danielle L.Hung Department of Molecular Microbiology Washington University School of Medicine 660 S. Euclid Avenue, Box 8230 St. Louis, MO 63110 USA Rainer Jaenicke Institut für Biophysik und Physikalische Biochemie Universität Regensburg D-93040 Regensburg
Germany Joanna Jakóbkiewicz Department of Molecular and Cellular Biology Faculty of Biotechnology University of Gdansk 80–822 Gdansk, Kladki 24 Poland Philip James Department of Biomolecular Chemistry University of Wisconsin 1300 University Avenue Madison, WI 53706 USA Eric R.Johnson Department of Structural Biology Stanford University School of Medicine Stanford, CA 94305–5400 USA C.Hal Jones Department of Molecular Microbiology Washington University School of Medicine 660 S. Euclid Avenue, Box 8230 St. Louis, MO 63110 USA D.Kim Departments of Radiation Oncology and Medical Physics Memorial Sloan-Kettering Cancer Center 1275 York Avenue New York, NY 10021 USA Peter Klappa Research School of Biosciences University of Kent Canterbury Kent CT2 7NJ UK Stefan Knight
Swedish University of Agricultural Sciences Uppsala Biomedical Center Department of Molecular Biology P.O. Box 590 S-751 24 Uppsala Sweden Igor Konieczny Department of Molecular and Cellular Biology Faculty of Biotechnology University of Gdansk 80–822 Gdansk, Kladki 24 Poland Joanna Krzewska Department of Molecular and Cellular Biology Faculty of Biotechnology University of Gdansk 80–822 Gdansk, Kladki 24 Poland G.C.Li Departments of Radiation Oncology and Medical Physics Memorial Sloan-Kettering Cancer Center 1275 York Avenue New York, NY 10021 USA L.Li Departments of Radiation Oncology and Medical Physics Memorial Sloan-Kettering Cancer Center 1275 York Avenue New York, NY 10021 USA Krzysztof Liberek Department of Molecular and Cellular Biology Faculty of Biotechnology University of Gdansk 80–822 Gdansk, Kladki 24 Poland Susan Lindquist Howard Hughes Medical Institute Department of Molecular Genetics and Cell Biology
University of Chicago 5841 S.Maryland Avenue, MC 1028 Chicago, IL 60637 USA C.C.Ling Departments of Radiation Oncology and Medical Physics Memorial Sloan-Kettering Cancer Center 1275 York Avenue New York, NY 10021 USA Jaroslaw Marszalek Department of Molecular and Cellular Biology Faculty of Biotechnology University of Gdansk 80–822 Gdansk, Kladki 24 Poland Michael R.Maurizi Laboratory of Cell Biology National Cancer Institute Bethesda, MD 20892 USA David B.McKay Department of Structural Biology Stanford University School of Medicine Stanford, CA 94305–5400 USA Dominique Missiakas Centre National de Recherche Scientifique LIDSM-CBBM 31 Chemin Joseph-Aiguier 13402 Marseille Cedex 20 France Michel Morange Unité de Génétique Moléculaire Département de Biologie École Normale Supérieure 46 rue d’Ulm 75230 Paris Cedex 05
France Richard I.Morimoto Department of Biochemistry, Molecular Biology and Cell Biology Rice Institute for Biomedical Research Northwestern University 2153 Sheridan Road Evanston, IL 60208 USA Eva Muckel Botanisches Institut Christian-Albrechts-Universität Am Botanischen Garten 1–9 D-24118 Kiel Germany Hirsohi Nagata Departments of Medicine and Physiology Lung Biology Research Center University of California Box 0854 San Francisco, CA 94143 USA A.Nussenzweig Departments of Radiation Oncology and Medical Physics Memorial Sloan-Kettering Cancer Center 1275 York Avenue New York, NY 10021 USA H.Ouyang Departments of Radiation Oncology and Medical Physics Memorial Sloan-Kettering Cancer Center 1275 York Avenue New York, NY 10021 USA Nikolaus Pfanner Institut für Biochemie und Molekularbiologie Universität Freiburg Hermann-Herder-Str. 7 D-79104 Freiburg Germany
Joanna Puzewicz Department of Molecular and Cellular Biology Faculty of Biotechnology University of Gdansk 80–822 Gdansk, Kladki 24 Poland Satish Raina Centre Medical Universitaire Département de Biochimie Médicale 1 rue Michel-Servet CH-1211 Geneva 4 Switzerland Neil A.Ranson Department of Crystallography Birkbeck College Malet Street London WC1E 7HX UK Jochen Reinstein Abteilung Physikalische Biochemie Max-Planck-Institut für Molekulare Physiologie Rheinlanddamm 201 D-44139 Dortmund Germany Helen R.Saibil Department of Crystallography Birkbeck College Malet Street London WC1E 7HX UK Eric S.Schirmer Howard Hughes Medical Institute Department of Molecular Genetics and Cell Biology University of Chicago 5841 S. Maryland Avenue, MC 1028 Chicago, IL 60637 USA Franz X.Schmid
Laboratorium für Biochemie Universität Bayreuth D-95440 Bayreuth Germany Robert Seckler Institut für Biophysik und Physikalische Biochemie Universität Regensburg D-93040 Regensburg Germany Jürgen Soll Botanisches Institut Christian-Albrechts-Universität Am Botanischen Garten 1–9 D-24118 Kiel Germany Marcelo C.Sousa Department of Structural Biology Stanford University School of Medicine Stanford, CA 94305–5400 USA Shigeki Takeda Department of Structural Biology Stanford University School of Medicine Stanford, CA 94305–5400 USA David O.Toft Department of Biochemistry and Molecular Biology Mayo Clinic 200 1st Street SW/1601 Rochester, MN 55905 USA Alicja Wawrzynow Department of Molecular and Cellular Biology Faculty of Biotechnology University of Gdansk 80–822 Gdansk, Kladki 24
Poland William J.Welch Departments of Medicine and Physiology Lung Biology Research Center University of California Box 0854 San Francisco, CA 94143 USA Sue Wickner Laboratory of Molecular Biology National Cancer Institute Bethesda, MD 20892 USA Sigurd M.Wilbanks Department of Structural Biology Stanford University School of Medicine Stanford, CA 94305–5400 USA Keith R.Willison Chester Beatty Laboratories Institute of Cancer Research 237 Fulham Road London SW3 6JB UK Wei Yan Department of Biomolecular Chemistry University of Wisconsin 1300 University Avenue Madison, WI 53706 USA S.-H.Yang Departments of Radiation Oncology and Medical Physics Memorial Sloan-Kettering Cancer Center 1275 York Avenue New York, NY 10021 USA Takashi Yura
HSP Research Institute Kyoto Research Park Kyoto 600 Japan Richard Zimmermann Medizinische Biochemie Universität des Saarlandes D-66421 Homburg Germany Maciej Zylicz Department of Molecular and Cellular Biology Faculty of Biotechnology University of Gdansk 80–822 Gdansk, Kladki 24 Poland
I. INTRODUCTION
1. ASSISTED PROTEIN FOLDING B.BUKAU1, * , F.X.SCHMID2 and J.BUCHNER3 1 Institut
für Biochemie und Molekularbiologie, Universität Freiburg, HermannHerder-Str. 7, D-79104 Freiburg, Germany 2 Laboratorium für Biochemie, Universität Bayreuth, D-95440 Bayreuth, Germany 3 Institut für Biophysik and Physikalische Biochemie, Universität Regensburg, 93040 Regensburg, Germany
1. Protein folding in vitro and in vivo 2. Classification of folding catalysts and molecular chaperones 2.1. Folding catalysts 2.2. Molecular chaperones 3. References
1. PROTEIN FOLDING IN VITRO AND IN VIVO The classical experiments by Anfinsen and others established that the entire information required for the folding of polypeptide chains to the native three-dimensional conformation is encoded in their primary amino acid sequences (Epstein et al., 1963). This information directs the formation of multiple non-covalent and covalent interactions within polypeptide chains and between subunits of protein oligomers which drive the folding process and stabilize the folded structures (chapter Seckler and Jaenicke). It also establishes a balance between structural stability and flexibility, achieved by low conformational stability of the folded protein (Privalov, 1979), which is required for the folding process itself and the activity of the folded protein. This balance implies that correct and incorrect folding, as well as native and nonnative structures, are separated by only relatively small energy barriers. Subtle changes in the amino acid sequence or the folding milieu may therefore have dramatic consequences for the folding process and the structural integrity of proteins. Misfolding of proteins is indeed the major damaging consequence of stress situations such as heat shock. Misfolded proteins frequently expose hydrophobic surfaces that are prone to intermolecular aggregation, a largely irreversible reaction. *Corresponding author
Molecular chaperones and folding catalysts
4
Cells require a system for the proofreading of protein conformations and the control and assistance of a multitude of folding processes for several reasons (see Figure 1 for overview on major folding reactions occuring in the life span of proteins). Proteins are synthesized vectorially, which may require mechanisms to coordinate synthesis with folding and to protect nascent chains from aggregation (chapter Welch et al.). Organellar and secretory proteins have to be translocated to their subcellular destinations prior to folding which necessitates mechanisms to coordinate folding with translocation and to assist the translocation process
Figure 1 Folding processes assisted by molecular chaperones and folding catalysts in vivo. Depicted are the major categories of folding processes that occur in vivo, starting with the folding of newly synthesized proteins (co- and post-translational) and ending with the degradation by cellular proteases. Molecular chaperones and/or folding catalysts have been implicated in all reactions shown.
itself (chapters Dekker and Pfanner; Muckel and Soll; Welch et al.; Haas and Zimmermann). For survival under stress, cells require an efficient conformational proofreading and repair system for misfolded proteins (chapters Lindquist and Schirmer; Li et al.; Maurizi et al.). The importance of this latter function is indicated by disease states such as amyloidoses and prions which result from the accumulation of aggregated protein (chapter Lindquist et al.; Horwich and Weissman, 1997; Lindquist, 1997; Prusiner, 1997; Thomas et al, 1995; Wetzel, 1996), and by the death of cells occuring upon inactivation of the repair system (chapters Connolly et al; Morimoto; Li et al; Craig et al; Burkholder and Gottesman). Intense research efforts in the past decade have led to the discovery of the evolutionary conserved families of molecular chaperones and folding catalysts which constitute the
Assisted protein folding
5
cellular system for folding and repair of proteins (see Table 1 for chaperones) (Buchner, 1996; Gething and Sambrook, 1992; Hartl, 1996). They assist the folding and targeting of newly synthesized proteins, prevent the aggregation of misfolded proteins, allow the refolding of kinetically trapped folding intermediates, mediate the translocation of proteins across membranes, assist the assembly and disassembly of protein complexes, play roles in proteolysis of unstable proteins, and even control the functional states of regulatory proteins. Members of different chaperone families and folding catalysts cooperate in folding reactions which led to the suggestion that assisted protein folding in vivo is promoted by a flexible network of folding helpers (Ehrnsperger et al, 1997; Bukau et al, 1996; Johnson and Craig, 1997).
2. CLASSIFICATION OF FOLDING CATALYSTS AND MOLECULAR CHAPERONES 2.1. Folding Catalysts The formation and isomerization of disulfide bonds and the cis-trans isomerizations of prolyl peptide bonds are slow and frequendy rate-limiting events in the folding of proteins. In vivo, these folding steps can be catalyzed by two classes of enzymes, known as protein disulfide isomerases or thiol/disulfide oxidoreductases (PDI) (chapter Freedman and Klappa) and peptidyl prolyl cis-trans isomerases (PPI) (Chapter Fischer and Schmid). PDIs are active in both the oxidized and the reduced form. In the oxidized form they introduce disulfide bonds into folding protein chains by direct thiol/disulfide exchange. In the reduced form they can attack existing disulfide bonds and thus isomerize incorrectly formed crosslinks. PDIs are localized in the endoplasmic reticulum of eukaryotic cells and the periplasm of bacteria where they are essential for disulfide bond formation in secreted proteins. All PDI proteins investigated share the catalytically active motif CysX-X-Cys in structurally related catalytic domains for which thioredoxin is the prototype. Despite this structural similarity there are striking differences within the PDI family with respect to the redox properties. Some PDI homologs, such as DsbA from E. coli, act as mere catalysts of disulfide bond formation, while others, such as eukaryotic PDI and E. coli DsbC catalyze both formation and isomerization of disulfide bonds very efficiently. These enzymes are typically composed of several thioredoxin-like domains which carry the catalytic thiol/disulfide exchange site as well as additional domains that mediate good binding to the substrate proteins. Peptidyl prolyl cis-trans isomerases catalyze the intrinsically slow rotation about XaaPro peptide bonds and thus accelerate folding reactions that are rate-limited by such isomerizations. Prolyl isomerases are abundant proteins and occur in virtually all organisms and cellular compartments. It is still unknown whether the catalysis of slow steps in protein folding is their major function. Considering the diversity and wide distribution of these enzymes it is almost certain that they are involved in many different cellular functions. The bacterial trigger factors were recently discovered to belong to the prolyl isomerases. They might, in fact, be prime candidates for ribosome-associated
Molecular chaperones and folding catalysts
6
folding enzymes that act very early in the life spans of proteins. 2.2. Molecular Chaperones The term “molecular chaperone” had been coined for a group of proteins which assist polypeptide folding in the cell. Chaperones seem to play multiple, housekeeping as well as stress related, roles in cell metabolism, including the folding and
Table 1 Conserved families of molecular chaperones and their co-chaperones1
Folding Prokaryotic Eukaryotic system Members Members
Functions
Hsp100
ClpA, ClpB, ClpX, ClpY
Hsp104, Hsp78
assistance of proteolysis of Maurizi et al.; unstable proteins (bacterial Lindquist and cytosol); prevention of aggregation Schirmer of misfolded proteins; disaggregation of misfolded proteins (eukaryotic cytosol)
Hsp90
HtpG
Hsp90, Grp94, ERp99, endoplasmin, Hsp108, gp96
prevention of aggregation and Bose et al; assistance of refolding of misfolded Toft proteins; regulation of activity of kinases and steroid hormone receptors
Hsp70
DnaK, HscA (Hsc66)
Hsp70, Hsc70, Ssa1–4, Ssb1, 2, Ssc, Ssh1, Lhs1, Kar2, BiP, Grp78
prevention of aggregation and Ha et al; assistance of refolding of misfolded Buchberger et proteins; folding of newly al; Craig et al. synthesized proteins (eukaryotic cytosol); activity control of regulatory proteins; translocation of precursors across membranes
DnaJ3
DnaJ, DjlA, CbpA, HscB
Hsp40, Ydj1, Sec63, Auxilin, CSPs, Mdj1, Hdj1, Hdj2
co-chaperone of Hsp70
Buchberger et al.
GrpE
GrpE
Mge1p
co-chaperone of Hsp70 (bacteria, mitochondria and chloroplasts)
Buchberger et al.
Functions
Book Chapters 2
Folding system
Prokaryotic Members
Eukaryotic Members
Book Chapters
sHSP
IbpA, IbpB
Hsp18.1, prevention of aggregation and Bose et al. Hsp25, Hsp27, assistance of refolding of -crystallin misfolded proteins
2
Assisted protein folding
7
PapD
PapD
—
assembly of bacterial pili
SecB
SecB
—
prevention of folding and Welch et al. targeting of precursor proteins to translocase (bacteria)
Hsp47
—
Hsp47
folding and assembly of collagen
Bose et al.
Calnexin
—
Calnexin
folding of proteins in the ER
Haas and Zimmermann
Calreticulin
folding of proteins in the ER
Haas and Zimmermann
Calreticulin —
Hultgren et al.
Subfamily of Chaperonins HspGO
GroEL
Hsp60; Cpn60
prevention of aggregation and Burston and folding of newly synthesized Saibil; Ranson and misfolded proteins and Clarke (bacteria, mitochondria and chloroplasts)
Hsp10
GroES, gp31
Hsp10, Cpn10
co-chaperone of GroEL
Burston and Saibil; Ranson and Clarke
CCT
TF55
TRiC
folding of newly synthesized and misfolded proteins (eukaryotic cytosol)
Willison
1
Only selected members of each chaperone family are shown.
2 Only the chapters with the strongest focus on the particular chaperone are listed. 3 The DnaJ family consists of a large group of heterogeneous proteins with diverse metabolic
functions. DnaJ proteins share the J domain, a conserved fragment of approx. 78 residues, which is essential for interaction of DnaJ with Hsp70 proteins.
translocation of newly synthesized proteins, the refolding of conformationally damaged proteins, and the control of biological activity of specific regulatory proteins. Originally, the functional classification of chaperones was restricted to two classes of proteins, the Hsp70 and GroEL heat shock proteins, but is now used for an ever increasing number of proteins unrelated in primary sequence. Molecular chaperones are grouped into families on the basis of their evolutionary conservation. Many chaperones are designated according to their approximate molecular weight, e.g. the 70 kDa heat shock protein is a chaperone termed Hsp70. A constitutively expressed cognate is termed Hsc70, and other members of the Hsp70 chaperone family have kept the name provided to them in the context of their historical discovery (DnaK, BiP, SSA1 etc.). We cannot eliminate this confusing nomenclature but suggest to continue using the now established historic names (see Table 1). In view of the growing number of proteins designated as molecular chaperones it is rewarding to define the basic properties that a protein has to fulfill to qualify as a
Molecular chaperones and folding catalysts
8
chaperone. The most common definition for a molecular chaperone is that it assists the structure formation of proteins and prevents unproductive side reactions without becoming part of the final structure (Ellis and Hemmingsen, 1989; Ellis, 1987). Chaperones do not catalyze or accelerate folding reactions, but rather increase the number of molecules that are on a productive folding pathway. This activity relies on their ability to inhibit intermolecular aggregation reactions by reversible association with aggregation-prone folding intermediates. In addition, the subclass of ring-like chaperonins such as GroEL, is capable of unfolding protein substrates whereby they may allow kinetically trapped misfolded polypeptides to reenter the productive folding pathway. Chaperones share the ability to transiently associate with non-native conformers of proteins by recognizing exposed hydrophobic patches. There are, however, differences with respect to the molecular mechanism of substrate recognition, as illustrated for four major chaperones (Figure 2). Hsp70, in functional cooperation with DnaJ co-chaperones, is active as a monomer containing a single substrate binding site (chapters Ha et al.; Buchberger et al.). The segment of the substrate polypeptide that contacts Hsp70 is a short stretch of five consecutive residues in extended conformation that becomes enclosed by the chaperone. Tight binding appears to require that the interacting peptide segment is physically separated from the remainder of the substrate and therefore substantial, at least local unfolding. To qualify as substrate for Hsp70, a minimal requirement for a protein is to expose a single chaperone binding site. This mode of interaction explains the wide spectrum of protein conformers, which can associate with Hsp70 ranging from extended (e.g. nascent polypeptide chains) to native. Chaperonins such as the prokaryotic GroEL and the eukaryotic CCT form double rings, composed of 7 (GroEL) to 8 (CCT) subunits/ring, each ring containing a substrate binding site made up of segments from each subunit (chapters Burston and Saibil; Ranson and Clarke; Willison). The ring structure allows the simultaneous association of various segments of a polypeptide chain within one ring, and this feature is most likely a key property allowing chaperonins to unfold protein substrates before release. A broad range of conformers can associate with GroEL, but in contrast to Hsp70 there are no reports for native proteins that are natural substrates.
Figure 2 Topology of substrate binding by molecular chaperones. Shown are the major molecular chaperones and their modes of interaction with substrate polypeptides. The structural nature of the substrate binding sites of Hsp90 and sHSPs remains unclear. Black bars in substrate polypeptides represent hydrophobic segments that serve as binding
Assisted protein folding
9
motifs for chaperones.
The conformation of the polypeptide segments that directly contact GroEL remains unclear. The small heat shock proteins (sHSPs) form oligomers with an average size of 12 to 42 subunits (chapter Bose et al.). Each oligomer can bind several protein substrates, up to one molecule per subunit, and thus serves as a very efficient binding scaffold for misfolded/unfolded substrates. Hsp90 acts as a dimer capable of binding non-native polypeptides (chapters Bose et al.; Toft). While for sHSPs and Hsp90 only little information exists with respect to the molecular basis of substrate recognition, recent data indicate that sHSPs and Hsp90 chaperones share with Hsp70 and GroEL the ability to recognize a broad range of conformations. The different chaperone families are thus not specialized for defined folding states of substrates, e.g. early unfolded or late molten globule-like states. Further differences between chaperone families exist with respect to the regulation of their functional activity. Some chaperones, including the sHSPs, Hsp47 and PapD, act independently of ATP (chapters Bose et al.; Hultgren et al.). It is somewhat mysterious how substrate binding is controlled in these cases. Yet unknown co-proteins or components of ATP-dependent chaperone systems may provide the cooperating partners for this class of chaperones. In contrast, the activity of major chaperones including Hsp70, chaperonins, Hsp90 and Hsp104/ClpB, is controlled by ATP and co-proteins (chapters Lindquist and Schirmer; Maurizi et al.; Ha et al.; Buchberger et al.; Burston and Saibil; Ranson and Clarke; Bose et al.). The role for ATP has been investigated in detail only for Hsp70 and GroEL. Hsp70 uses the energy of ATP to drive conformational changes that alter its affinity for substrates. The ATPase cycle of Hsp70 can be viewed, in its simplest form, as an alternation between two states: the ATP state with low affinity and fast exchange rates for substrates (substrate binding pocket open), and the ADP state with high affinity and low exchange rates for substrates (substrate binding pocket closed). GroEL uses ATP to drive coordinated conformational changes of all subunits of one ring, and subsequently in the other ring, which allow dissociation of substrates and ligands. ATP thus provides a mechanism to tightly control the activity of both chaperone systems, by affecting the kinetics of substrate binding and release. The ATPase activities of these chaperones are prime targets for regulatory proteins which either stimulate or inhibit checkpoints of the ATPase cycle and thereby control the affinity of the corresponding chaperone partner for substrates. Examples are the Hsp70 co-proteins DnaJ (Hsp40), GrpE, Hip and Bag1, and the GroEL co-proteins GroES and gp31. ATP-dependent chaperone systems are thus sophisticated and tightly regulated machines. The possibility to regulate their binding to substrate allows them at least in the case of Hsp70 to play diverse roles in cell metabolism, ranging from general functions in protein folding to highly specific functions e.g. in control of biological activities of regulatory proteins. Members of different chaperone families have been found in association with the same substrate conformer and capable of competing for binding. This principle of kinetic partitioning of substrates between different chaperones, and possibly folding catalysts and proteases, is likely to constitute the basis for a cellular network of folding helpers that assists protein folding (Ehrnsperger et al., 1997; Bukau et al., 1996; Johnson and Craig,
Molecular chaperones and folding catalysts
10
1997). Elucidation of the molecular principles and the biological implications of this network is a central goal for future research and will require the combined input of biochemistry, genetics and cell biology.
3. REFERENCES Buchner, J. (1996). Supervising the fold: functional principles of molecular chaperones. FASEB J. , 10, 10–19. Bukau, B., Hesterkamp, H. and Luirink, J. (1996). Growing up in a dangerous environment: a network of multiple targeting and folding pathways for nascent polypetides in the cytosol. Trends Cell Biol. , 6, 480–486. Ehrnsperger, M., Gräber, S., Gaestel, M. and Buchner, J. (1997). Binding of non-native protein to Hsp25 during heat shock creates a reservoir of folding intermediates for reactivation. EMBO J. , 16, 221–229. Ellis, J. (1987). Proteins as molecular chaperones. Nature (London), 328, 378–379. Ellis, R.J., and Hemmingsen, S.M. (1989). Molecular chaperones: proteins essential for the biogenesis of some macromolecular structures. Trends Biochem. Sci. , 14, 339–42. Epstein, C.J., Goldberger, R.F. and Anfinsen, C.B. (1963). The genetic control of tertiary protein structure: studies with model systems. Cold Spring Harb. Symp. Quant. Biol. , 28, 439–449. Gething, M.-J. and Sambrook, J.F. (1992). Protein folding in the cell. Nature , 355, 33– 45. Hartl, F.U. (1996). Molecular chaperones in cellular protein folding. Nature , 381, 571– 580. Horwich, A.L. and Weissman, J.S. (1997). Deadly conformations-protein misfolding in prion disease. Cell 89, 499–510. Johnson, J.L., and Craig, E.A. (1997). Protein folding in vivo: Unraveling complex pathways. Cell , 90, 201–204. Lindquist, S. (1997). Mad cows meet Psi-chotic yeast: the expansion of the prion disease. Cell , 89, 495–498. Privalov, P.L. (1979). Stability of proteins. Adv. Protein Chem. , 33, 167–241. Prusiner, S.B. (1997). Prion diseases and the BSE crisis. Science , 278, 245–251. Thomas, P.J., Qu, B.-H., and Pedersen, P.L. (1995). Defective protein folding as a basis of human disease. Trends Biochem. Sci. , 20, 456–459. Wetzel, R. (1996). For protein misassembly, it’s the “I” decade. Cell , 86, 699–702.
II. REGULATION
2. AUTOREGULATION OF THE HEAT SHOCK RESPONSE IN PROCARYOTES LYNN CONNOLLY1, TAKASHI YURA2 and CAROL A.GROSS3, * 1 Department
of Biochemistry and Biophysics, University of California, San Francisco, San Francisco, CA 94143 2 HSP Research Institute, Kyoto Research Park, Kyoto 600, Japan 3 Departments of Stomatology and Microbiology and Immunology, University of California, San Francisco, CA 94143
1. Introduction 2. Regulation of the
heat shock response
2.1. Discovery of 2.2. How does
regulate the response to temperature shift?
2.3. Translational regulation of 2.4. Regulation of
stability
2.5. Regulation of
activity
2.6. What are the signals governing expression of the 3. Regulation of the
heat shock regulon?
(24) heat shock response
3.1. Discovery of 3.2. What is the nature of the signal inducing
activity?
3.3. Regulation of 3.4. How is the extracytoplasmic signal transduced to 3.5. The cellular role of 4. Heat shock regulation in other prokaryotic organisms 5. Summary and prospects 6. Acknowledgments 7. References *Corresponding author
?
Molecular chaperones and folding catalysts
14
1. INTRODUCTION When cells of any type are shifted to high temperature, the heat shock response (hsr) ensues and the synthesis of a small number of proteins, called the heat shock proteins (hsps), is rapidly induced. In E. coli, the hsr was discovered independently by the Neidhardt and Yura groups, who monitored the rate of synthesis of individual proteins after a temperature upshift using either 1D or 2D gels (Lemaux et al., 1978; Yamamori et al., 1978). A group of about 20 proteins exhibited a large (10 to 20-fold) but transient increase in synthetic rate upon temperature upshift and a corresponding decrease in synthetic rate upon temperature downshift (Lemaux et al., 1978; Yamamori et al., 1978; Neidhardt et al., 1987; Straus et al., 1989; Taura et al., 1989). This group of proteins comprises the E. coli hsps. Their expression is regulated at the transcriptional level (Yamamori et al., 1980; Taylor et al., 1984; Cowing et al., 1985) by the amount and/or activity of the alternative sigma factor, , which directs RNA polymerase to transcribe this set of genes (Lesley et al., 1987; Skelly et al., 1987; Straus et al., 1987). These hsps, including the chaperones DnaK-DnaJ and GroELGroES, are required for normal growth at physiological temperatures. Whereas E. coli in its natural habitat grows at temperatures between 25°C and 40°C, deletion of the gene encoding restricts growth to temperatures below 20°C (Zhou et al., 1988). Overexpression of the GroEL-GroES and DnaK-DnaJ chaperone machines restores high temperature growth, suggesting that these chaperones play a crucial role in adaptation to high temperature. E. coli also has a second heat-controlled regulon, controlled by ( ), another alternative sigma factor (Erickson et al., 1989; Wang et al., 1989) (see Missiakas and Raina, this volume). Many members of this regulon have yet to be identified. The two responses are intertwined because holoenzyme containing ( ) transcribes at extreme temperature. However, each response also has a distinct role in the cell: controlled genes respond to conditions in the cytoplasm of the cell whereas controlled genes respond to the extracytoplasmic state. The regulon plays an auxiliary role in temperature adaptation as cells lacking cannot grow at temperatures above 40° (Raina et al., 1995; Hiratsu et al., 1995; Rouvière et al., 1995). Such strains also exhibit defects in the cell envelope, emphasizing the dual role played by members of this regulon. The heat induction of several additional genes may occur by other mechanisms. controls genes involved in adaptation to stationary phase and is also somewhat induced upon shift to high temperature, suggesting that genes in the regulon exhibit temperature regulation (Hengge-Aronis, 1996). Finally, the psp operon is controlled by a dedicated activator protein that promotes psp transcription by EJ54 following shift to very high temperatures (Brissette et al., 1990). Two global approaches, one monitoring protein synthesis and the other monitoring RNA synthesis, have been used to identify most of the hsps. In the protein based approach, spots on 2D gels have been correlated with known genes (Georgopoulos et al., 1982; Neidhardt et al., 1981; Tilly et al., 1983). In the RNA based transcriptional mapping approach, radioactively labeled cDNA, made to total E. coli RNA, is hybridized
Autoregulation of the heat
15
to membrane filters containing an ordered E. coli genomic library carried in clones (the Kohara library) and clones whose transcription increases are identified (Chuang et al., 1993; Chuang et al., 1993). A compendium of the proteins whose rates of synthesis increase upon temperature upshift is presented in Table 1.
2. REGULATION OF THE
HEAT SHOCK RESPONSE
2.1. Discovery of The gene encoding was discovered in 1975 as a nonsense mutation that affected the synthesis of the GroEL hsp. The mutation was initially thought to be located in the structural gene for GroEL (Cooper et al., 1975). Subsequently, it was found
Table 1 Heat inducible proteins in Escherichia coli
Min Protein Alphanumeric designation
Molec. Function Weight
Kohara Physical Reference(s) Clones Map
21
?
?
Regulon .3
HtpY
.3
DnaK
B 066.0
69
chaperone
101, 102 11.7–15.5 [Bardwell et al., 1984]
.3
DnaJ
H 036.5
39
chaperone
101, 102 11.7–15.5 [Bardwell et al., 1986]
H 094.0
89
protease
148
464.1– 468.4
[Gayda et al., 1985]
24(22)
protease
148
464.1– 468.4
[Maurizi et al., 1990]
10.0 ClpX
46
chaperone
148
464.1– 468.4
[Gottesman et al., 1993]
10.0 HslA
65
?
70
chaperone
152
501.5– 504.2
[Bardwell et al., 1987]
80
?
212
921.6– 936.7
[Chuang et al., 1993]
35.5
dehydrogenase 330, 331 1872–
[Charpentier et
10.0 Lon
[Missiakas et al., 1993]
H 094.1 10.0 ClpP
10.8 HtpG
F 021.5
C 062.5
19.2 HslC 39.3 GapA
I 033.5
[Chuang et al., 1993]
Molecular chaperones and folding catalysts
16
1873
al., 1987]
H 034.3 39.8 HslK
49
?
334
1901.2– 1904.2
[Chuang et al., 1993]
40.3 HtpX
32
?
?
?
[Kornitzer et al., 1991]
437
2741.2– 2743.7
[Kitagawa et al., 1991; Squires et al., 1992]
56.0 ClpB
F 084.1 E 072.0
84
chaperone
56.8 GrpE
B 025.3
26
nucleotide 438, 439 2757.7– exchange factor 2763.6
[Lipinska et al., 1988]
67.0
B 082.0
70
sigma factor
509
3233.0– 3236.2
[Burton et al., 1981]
520
3331.7– 3350.3
[Herman et al., 1995; Tomoyasu et al., 1993]
3331.7– 3350.3
[Herman et al., 1995; Tomoyasu et al., 1993]
69.2 FtsJ
26
69.2 HflB
70
protease
520
75.0 HslO
33
?
620, 621 3549.6– 3552.0
75.0 HslP
30
?
81.2 HtrM (RfaD)
34
epimerase
575, 576 3815.3– 3816.4
[Raina et al., 1991]
[Chuang et al., 1993] [Chuang et al., 1993]
83.0 IbpB (HtpE, HslS)
C 014.7
16.3
chaperone
566, 567 3889.9– 3892.7
[Allen et al., 1992; Chuang et al., 1993]
83.0 IbpA (HtpN, HslT)
G 013.5
15.8
chaperone
566, 567 3889.9– 3892.7
[Allen et al., 1992; Chuang et al., 1993]
89.0 ClpY (HtpI, HslU))
D 048.5
49
chaperone
538, 539 4149.5– 4151.7
[Chuang et al., 1993, Missiakas et al., 1996]
89.0 HslV (HtpO)
G 021.0
21
protease
538, 539 4149.5– 4151.7
[Chuang et al., 1993, Missiakas et al., 1996]
21
?
60
chaperone
90.0 HtrC 94.2 GroEL
B 056.5
[Raina et al., 1990] 648, 649 4400.5–
[Hemmingsen et
Autoregulation of the heat
17 4405.7
94.2 GroES
16
chaperone
648, 649 4400.5– 4405.7
[Hemmingsen et al., 1988]
94.2 HslW
22
?
648, 649 4400.5– 4405.7
[Chuang et al., 1993]
94.8 HslX
51
?
652
4430.8– 4433.4
[Chuang et al., 1993]
94.8 HslY
45
652
4430.8– 4433.4
[Chuang et al., 1993]
94.8 HslZ
37
652
4430.8– 4433.4
[Chuang et al., 1993]
HtpK
C 015.4
al., 1988]
F 010.1
Min Protein Alphanumeric designation
[Aa]
10
Molec. Function Weight
Kohara Physical Reference(s) Clones Map
Regulon HtpT
A 039.5
40
[Aa]
Regulon: 3.9
DegP (HtrA)
50
55.5
77.5
F 033.4
74.9 tkpA
protease
117, 118 181–182
[Lipinska et al., 1988; Strauch et al., 1989]
sigma factor
435
2718
[Lonetto et al., 1994; Nashimoto 1993; Raina et al., 1995]
sigma factor
613
3614– 3625
[Landick et al., 1984; Yura et al., 1984]
PPlase
625, 626
[Danese et al., 1997]
257, 258 1374– 1378
[Lipinska et al., 1988; Yamamori et al., 1982]
Others: 29.2 PspA
E 026.0
28
29.7 HslE
60
?
260
1388.8– 1409.9
[Chuang et al., 1993]
29.7 HslF
51
?
260
1388.8– 1409.9
[Chuang et al., 1993]
Molecular chaperones and folding catalysts 29.7 HslG
18
41
?
260
1388.8– 1409.9
[Chuang et al., 1993]
36
?
265
1448.9– 1454.5
[Chuang et al., 1993]
30.6 HslJ
14
?
265
1448.9– 1454.5
[Chuang et al., 1993]
69.2 HslM
31
?
520
3331.7– 3350.3
[Chuang et al., 1993]
75.0 HslQ
24
?
620, 621 3549.6– 3552.9
[Chuang et al., 1993]
75.0 HslR
18
?
620, 621 3549.6– 3552.9
[Chuang et al., 1993]
60
LysyltRNA synthetase
646, 647 4381.8– 4383.2
[Léveque et al., 1990]
30.6 HslI (HtpH)
93.5 LysU
D 033.4
D 060.5
that mutant cells had a global defect in the hsr, suggesting instead that the gene encoded a regulator of the hsr (Neidhardt et al., 1981; Yamamori et al., 1982). The sequence of the gene revealed strong homology to (Landick et al., 1984; Yura et al., 1984) and the regulator was shown to be , the first alternative sigma factor identified in E. coli (Grossman et al., 1984). directs core RNA polymerase to promoters that differ considerably from those recognized by RNA polymerase containing , the housekeeping sigma (Cowing et al., 1985). The fact that expression of the hsps is uniquely responsive to the amount or activity of provides a means to regulate their expression separately from other cellular proteins. 2.2. How Does
Regulate the Response to Temperature Shift?
When cells experience a temperature upshift, for example after shift from 30°C to 42°C, the rate of synthesis of the hsps increases 10 to 20-fold by 5 minutes after upshift and thereafter declines to a new steady state rate of synthesis. Interestingly, at steady state, the amount of hsps at 42° is only 2-fold greater than that at 30°. The large increase in rate of hsp synthesis immediately after temperature upshift allows cells to rapidly accumulate the new steady state level of hsps (Lemaux et al., 1978; Yamamori et al., 1978; Straus et al., 1987). The response of hsps to heat induction is controlled at the transcriptional level, primarily by the amount of in the cell. At low temperature, cells contain very little , on the order of 10 to 50 molecules per cell. By 5 minutes after temperature upshift, the amount of increases about 15-fold and thereafter declines to a new steady state level (Lesley et al., 1987; Straus et al., 1987). Changes in the amount of following temperature upshift result from changes in both the stability and synthesis of (Lesley et al., 1987; Straus et al., 1987). During steady state growth, is translated at a very
Autoregulation of the heat
19
low rate. In addition, is very unstable, with a T½ for degradation of about 1 minute. As a result, little accumulates in the cell. However, for the first 5 minutes following temperature upshift the rate of translation of increases about 5-fold and is stabilized against degradation. Following this time, the rate of translation decreases and rapid degradation resumes. Together, these two regulatory changes permit the transient accumulation of . To a first approximation, changes in the rate of hsp synthesis after temperature upshift primarily mirror changes in the amount of (Lesley et al., 1987; Skelly et al., 1987; Straus et al., 1987). However, careful examination of the kinetics suggest that shutoff of hsp synthesis in the adaptation phase of the hsp response may slightly precede the decrease in the amount of . Regulation of activity (see below) may be involved in this phenomenon. When cells experience a temperature downshift, for example after shift from 42°C to 30°C, the rate of synthesis of hsps declines 10 to 20-fold within 5 minutes after downshift. This rate of hsp synthesis is considerably lower than that normally exhibited by cells growing at 30°C (Straus et al., 1989; Taura et al., 1989). By one to two doublings after downshift, the cell gradually resumes the 30°C rate of synthesis. Presumably, existing hsps are diluted out during the long shut-off period. Hsp synthesis resumes when their amounts approximate that characteristic of the cells growing continuously at low temperature. The rapid drop in transcription of heat shock genes upon temperature downshift results from a decrease in activity, rather than from a decrease in the amount of . Temperature downshift is not the only condition that promotes inactivation of . Overexpression of hsps at constant temperature also reduces activity, suggesting that cells can sense the amount of hsps and adjust the activity of accordingly (Straus et al., 1989; Craig et al., 1991). These studies indicate that the translation, stability and activity of are all regulated by the cell in response to temperature. The extent to which temperature regulation of each of these processes is understood at a mechanistic level is discussed below, and a speculative model of the regulation of activity is presented in Figure 1. 2.3. Translational Regulation of Translational regulation includes both translational induction, which occurs immediately following temperature upshift, and translational repression, which occurs
Molecular chaperones and folding catalysts
20
Figure 1 The promoters and translational regulatory regions of E. coli rpoH. (a) Regions A and B of the mRNA are involved in translational induction by modulating the secondary structure shown in (b), whereas region C of is involved in chaperone mediated translational repression and protein stability (see text). (b) A possible secondary structure of the mRNA formed under nonstress conditions. (Reproduced with permission from Yura, 1996).
subsequently during the adaptation phase of the hsr. The cis-elements and the transacting factors required for induction and repression differ, suggesting that these two processes
Autoregulation of the heat
21
are mechanistically distinct. The mechanism of translational induction has been probed by both deletion and point mutational analysis of a - -galactosidase fusion protein (Kamath-Loeb et al., 1991; Nagai et al., 1991; Yuzawa et al., 1993). These studies indicate that two regions within , termed A and B, are required for translational induction (Figure 1). Region A, located near the start of translation initiation (nucleotide 6–20), has homology to the “downstream box”, which is required for high rates of translation in several prokaryotic systems. Deletion of the downstream box leads to very low, uninducible synthesis of . Region B is a grossly defined, internal region extending from nucleotide 110–210, part of which has the capacity to base pair with a portion of Region A. Deletion of Region B, as well as some point mutations in the region, leads to high constitutive synthesis of . Initial speculation that thermal induction might simply be explained by disruption of base-pairing potential between the two regions, led to an analysis of compensating mutational changes between putative base-pairing partners. These studies indicated that recovery of base pairing is not always sufficient for regulation, leading to the suggestion that sequence, as well as structure, is important for regulation (Yuzawa et al., 1993; Yura, 1996). The current view is that an unknown transacting factor is involved in this regulatory event. The mechanism of translational repression is distinct from that of translational induction. Translational repression requires Region C of (nucleotide 364–433; amino acid 122–144) and the DnaK, DnaJ, GrpE chaperone machine (Straus et al., 1990; Nagai et al., 1994). Deletion analysis indicates that lack of Region C prevents repression, and analysis of a frameshift of Region C indicated that polypeptide rather than nucleotide sequence was involved in the response. Interestingly, a peptide scan of using a library of overlapping 13 amino acid-long peptides identified Region C as the site of two high affinity DnaK binding sites within , leading to speculation that the function of Region C may be to bind DnaK (McCarty et al., 1996). Further support for this notion comes from comparative analysis of the sigma family of polypeptides. Whereas this region of sigma is highly conserved among homologues from diverse bacteria, it is poorly conserved among sigma factors in general (Nakahigashi et al., 1995). It is certainly plausible that a nonconserved region within the sigma family of proteins has become specialized for a regulatory function specific to homologues. Cotranslational binding of DnaK to Region C may then mediate translational repression by an unknown mechanism. 2.4. Regulation of
Stability
The instability of is a key feature of the response to temperature upshift. Because is so unstable (T½=1 minutes) during steady state growth, increases in its rate of synthesis are immediately reflected in commensurate increases in the level of available to promote transcription of the heat shock genes. Great advances in understanding this process have recently been reported. Both in vivo and in vitro studies indicate that is proteolysed by HflB, an ATP dependent protease located in the inner membrane (Tomoyasu et al., 1993; Herman et al., 1995; Tomoyasu et al., 1995).
Molecular chaperones and folding catalysts
22
Depleting cells of HflB (FtsH), or inactivating mutant HflB by shift to high temperature stabilizes about 10-fold indicating that HflB is a major protease responsible for degradation. Moreover, HflB can degrade in vitro. Interestingly, HflB is a member of the regulon and the only essential protease thus far reported in E. coli. There are still important, unresolved questions concerning the physiology of degradation. Currently, the rate of degradation of in vitro (T½=18 minutes) is much slower than the in vivo T½ of 1 min. In vivo, the DnaK-DnaJ-GrpE chaperone machine is required for degradation of , and mutations in dnaK, dnaJ or grpE decrease the rate of degradation as much as 10-fold (Tilly et al., 1989; Straus et al., 1990). Region C of , described above as a possible DnaK binding site, may couple these chaperones to the process of degradation. In support of this idea, the Region C frameshift mutant inhibits degradation of in vivo (Nagai et al., 1994). However, the in vitro degradation system currently in use exhibits no requirement for these hsps (Tomoyasu et al., 1995). Moreover, the presence of core RNA polymerase inhibits the in vitro degradation of by HflB, and this inhibition is not reversed by the DnaK-DnaJ-GrpE chaperone machine. Thus, the in vitro system is not yet a faithful mimic of in vivo degradation, either because of missing components or altered conditions. 2.5. Regulation of
Activity
Inactivation of appears to be a primary mode of regulation whenever is present in excess in the cell (Straus et al., 1989; Taura et al., 1989; Straus et al., 1990). This regulatory mode features most prominently on temperature downshift, but also most likely sharpens the shut-off phase of the heat shock response. The DnaK-DnaJ-GrpE chaperone machine is involved in inactivation, as cells carrying mutations in these genes are defective in this process (Straus et al., 1989 and unpublished experiments). Inactivation is reversible as regains activity after extraction from the cell (Straus et al., 1989). These characteristics led to the proposal that the DnaK-DnaJ-GrpE chaperone machine reversibly binds to to inhibit its function (Straus et al., 1989) (Figure 2). Elegant in vitro studies from the Bukau and Georgopoulos laboratories are beginning to establish the molecular basis for inactivation of . Both DnaK and DnaJ can bind independently to (Gamer et al., 1992; Liberek et al., 1992; Liberek et al., 1993; Gamer et al., 1996). In addition, all three also form an ATP-dependent ternary complex with distinct properties from each of the binary complexes (Liberek et al., 1993; Gamer et al., 1996). It is only this ternary complex that shows decreased activity with core RNA polymerase (Liberek et al., 1993; Gamer et al., 1996). Thus, together DnaK and DnaJ function as an anti-sigma factor. When bound to , they inhibit the formation of the -core RNA polymerase complex (Gamer et al., 1996). Understanding the mechanistic details of the interactions of DnaK and DnaJ with is in its infancy. Indeed, further study of this interaction is likely to yield important insights concerning the regulatory loop governing activity, and also
Autoregulation of the heat
23
Figure 2 Speculative model for the mechanism by which DnaK, DnaJ and GrpE regulate expression of hps by controlling activity and levels. Upon temperature upshift, the increase in misfolded protein substrates leads to a decrease in the free levels of DnaK, DnaJ and GrpE resulting in increased stability. Upon temperature downshift, the increase in the free pool of these chaperones leads to inactivation of . In addition to these effects, a role for DnaK, DnaJ and GrpE in negatively regulating the increase in translation of observed upon temperature upshift has been proposed (see text). (Figure adapted from Gross, 1996).
into the nature of chaperone interaction with native substrates. The DnaKbinary complex is relatively weak (Kd=5 M), and this binding is considerably decreased by ATP (Gamer et al., 1992; Liberek et al., 1992; Liberek et al., 1993; Gamer et al., 1996). Interestingly, the low binding constant reflects a very slow on rate, as the DnaKcomplex is quite stable once formed (T½>30 minutes) (Gamer et al., 1996). In contrast, the stronger DnaJbinary complex (Kd=20nM; measured in the Biacore), actually dissociates more rapidly than the DnaKcomplex (Gamer et al., 1996). The ternary complex, which requires ATP for its formation, somehow stabilizes the -DnaK interaction and effectively competes with for binding to core RNA polymerase. It is
Molecular chaperones and folding catalysts
24
currently unknown how DnaJ promotes formation of this ternary complex. However, DnaJ binding to substrate may not be necessary for its effect. Some DnaJ mutants that do not bind still promote an ATP-resistant -DnaK interaction, and may do so catalytically (Liberek et al., 1995). It is not known, however, whether these -DnaK binary complexes inhibit mediated transcription. 2.6. What are the Signals Governing Expression of the Regulon?
Heat Shock
The challenge of the cell is to integrate diverse environmental information to program the level of hsp expression that is appropriate for the perceived cumulative stress level. Exactly how this is accomplished is still a matter of speculation. We have a great deal of information about initial inputs—expression of the regulon is triggered by heat, ethanol and other diverse insults. Likewise, we are fairly knowledgeable about the final outputs— regulation of both the activity and amount of lead to a defined rate of transcription of the heat shock genes. However, the nature of the signal-transduction pathway(s) that couple(s) the two ends of this regulatory loop remains an area of active investigation. There are at least two distinct signal-transduction pathways governing expression of the hsps. The first pathway controls translation of mRNA in a positive way: increased environmental stress leads to increased translation. This pathway is induced by exposure to heat and ethanol, but not by accumulation of unfolded proteins. To date, the only identified player in this pathway is cis-acting mRNA sequences. Neither the trans-acting factors, nor the signaling molecule (s) have been identified. Our understanding of the remainder of the regulatory events governing the amount of active is somewhat more advanced. Regulating stability, activity and translational repression have in common the involvement of the DnaK, DnaJ and GrpE chaperone machine in the signal transduction pathway. Regulation of these diverse processes may be controlled either by a single pathway, or by multiple, interconnected pathways. A homeostatic mechanism coupling the occupancy of the DnaK, DnaJ, GrpE chaperone machine to the amount and activity of has been proposed (Straus et al., 1990; Craig et al., 1991; Bukau, 1993). Cellular stress is monitored by how well can compete with all other unfolded or misfolded proteins for binding to the DnaK, DnaJ, GrpE chaperone machine. Inducing signals increase unfolded or misfolded proteins, thus titrating DnaK, DnaJ and GrpE away from and relieving their negative regulatory effects on stability and translation. As a consequence, the amount of will rise. Conversely, repressing signals will decrease unfolded or misfolded proteins, thus freeing DnaK, DnaJ and GrpE to inactivate . This response is self limiting because under or over production of DnaK, DnaJ and GrpE will restore the free pool of these chaperones to an appropriate level. Thus, the amount of free DnaK, DnaJ, and GrpE is a “cellular thermometer” that measures the “folding state” of the cell. There is some evidence in favor of this model, however, critical experiments to test the proposition that the DnaK, DnaJ and GrpE chaperones play a regulatory role have yet to be carried out.
Autoregulation of the heat
3. REGULATION OF THE
(
25
) HEAT SHOCK RESPONSE
3.1. Discovery of was originally discovered as the sigma factor responsible for maintaining transcription of rpoH at extreme temperatures. rpoH has four promoters, three of which are transcribed by E (Figure 1a). The fourth promoter, rpoHp3, is recognized by E . rpoHp3 accounts for only 2% of total rpoH transcription at 30°C, but drives over 90% at the lethal temperature of 50°C (Erickson et al., 1987). The continued production of at 50°C is critical to cellular survival, as the dependent hsps represent the majority of proteins expressed under these extreme conditions (Neidhardt et al., 1984; Pack et al., 1986). was purified based on its ability to direct transcription from rpoHp3 (Erickson et al., 1989; Wang et al., 1989), and the structural gene encoding was recently identified (Raina et al., 1995; Rouvière et al., 1995). 3.2. What is the Nature of the Signal Inducing
Activity?
In addition to being induced by the general stresses of heat and solvents, the pathway is uniquely induced in response to alterations in the expression or maturation of outer membrane proteins (OMPs) (Mecsas et al., 1993). Overexpression of OMPs induces activity, and underexpression of OMPs decreases activity. The inducing signal arises either during or after translocation because cytoplasmic accumulation of OMP precursors does not induce activity. Although activity is induced by overexpression of some periplasmic proteins with known folding defects (Missiakas et al., 1996), overexpression of most periplasmic proteins does not induce , indicating that the signal is probably not arising due to titration of the translocation machinery. Expression of a mutant OMP that is properly translocated but fails to be inserted into the outer membrane also induces activity. Taken together, these results suggest that the signal arises in the periplasmic space, after translocation but prior to insertion into the outer membrane. Outer membrane proteins undergo a complex series of folding events during their maturation into trimeric porins. Blocking this pathway at a step after the signal intermediate is generated should cause an increase in activity. Using this and related strategies, several putative periplasmic folding agents have been identified, including the peptidyl prolyl isomerases SurA and FkpA, and the Skp protein (Rouvière et al. 1996; Missiakas et al., 1996). Loss of function mutations in each of these genes induce activity. The role of SurA in maturation of the trimeric porin LamB has been investigated (Rouvière et al., 1996; Lazar et al., 1996). SurA appears to catalyze the formation of a folded monomeric species from unfolded monomer. Cells lacking SurA and cells overexpressing LamB both accumulate the unfolded monomer form at the expense of folded monomer. The observation that two different inducing conditions result in accumulation of unfolded monomer suggests that the signal for induction occurs somewhere prior to the formation of the folded monomer species (Rouvière et al., 1996).
Molecular chaperones and folding catalysts
26
3.3. Regulation of The activity of is regulated, in part, at the level of transcription. is transcribed from a -dependent promoter and transcription from this promoter reflects the level of activity in the cell under steady state conditions (Raina et al., 1995; Rouvière et al., 1995). However, both the observation that transcription of is low under steady state conditions and that activity increases rapidly in response to induction suggest additional regulatory controls. Homology arguments suggested that is under the control of negative regulators likely to be encoded in the same operon as rpoE, and this turns out to be the case. belongs to the ECF subclass of the family of proteins, most of which regulate extracytoplasmic functions (Rouvière et al., 1995; Lonetto et al., 1994). Operons encoding other ECF sigmas have previously been shown to also encode regulators of the sigma factor activity. In particular, the operon encoding the closely related algU/T sigma factor required for alginate biosynthesis in P. aeruginosa, includes two negative regulators of AlgU/T activity, MucA and MucB (Martin et al., 1993). MucA inhibits AlgU/T activity in vivo and in vitro (Schurr et al., 1996; Xie et al., 1996), and previous work had identified a partial open reading frame encoded immediately downstream of rpoE, termed mclA, that showed significant homology to mucA (Raina et al., 1995; Rouvière et al., 1995; Yu et al., 1995). Three genes, rseABC (for regulator of sigmaE), are encoded immediately downstream of rpoE, and genetic experiments reveal that rseA (formerly mclA) and rseB negatively regulate activity (De Las Peñas, et al., 1997a; Missiakas et al., 1997). Deletion of rseA leads to a 25-fold induction of activity, whereas deletion of rseB gives only 2.3-fold induction, indicating that RseA is the major negative regulator of . RseA is an inner membrane protein, whose cytoplasmic domain binds directly to and inhibits -directed transcription in vivo and in vitro. Thus, the cytoplasmic domain of RseA acts as an anti-sigma factor. The periplasmic domain of RseA interacts with RseB, which is located in the periplasm, and RseC has a slight positive effect activity. 3.4. How is the Extracytoplasmic Signal Transduced to
?
RseA is the central regulatory molecule in the signal transduction cascade to . Cells lacking RseA are unresponsive to induction because they are already maximally induced. Moreover, cells containing only RseA modulate activity in response to inducer, indicating that RseA alone or in conjunction with unknown molecules responds to the inducing signal. Several mechanisms of RseA inactivation by the inducer can be envisioned including modification, degradation, or oligomerization of the anti-sigma factor. RseB may act to fine-tune this RseA-based signal transduction pathway. Binding of RseB to the periplasmic domain of RseA might shift RseA to a conformation where it is most effective as an anti-sigma (Figure 3a). If RseB binding to RseA were competitive with binding to a signal molecule, RseB would be titrated away from RseA as the concentration of the signal increases (Figure 3b). This would leave RseA in a
Autoregulation of the heat
27
conformational state where it is a less effective anti-sigma, and lead to a small increase in activity. At still higher concentrations, the signal molecule would interact either with an intermediate factor or with RseA itself to further increase activity (Figure 3c). The direct induction signal and how it affects RseA is currently unknown. is induced by the build up of early intermediates in the maturation pathway of outer
Figure 3 Speculative model of the signal transduction cascade leading to activation of . (a) In the presence of low levels of signal, is
Molecular chaperones and folding catalysts
28
sequestered to the membrane by a protein complex consisting of RseA and RseB, leaving activity low. (b) Under conditions of low level signal, RseB is titrated off of RseA, leaving RseA in a conformation that is less active as an anti-sigma factor, resulting in a small increase in activity, (c) When the signal is high, RseA is further inactivated either by interaction with the signal molecule itself or some intermediate factor, resulting in a large induction of activity.
membrane porins, the accumulation of a few periplasmic proteins, and a deficit of any of several periplasmic folding agents (DsbA, FkpA, Skp and SurA) (Mecsas et al., 1993; Rouvière et al., 1996; Missiakas et al., 1996). The Rse proteins may detect the levels of misfolded protein directly. Alternatively, RseA and/or RseB may monitor the levels of free periplasmic folding agents, including SurA, FkpA, and the Dsb proteins. Decreases in the free levels of each of these proteins in response to the accumulation of unfolded or misfolded species in the periplasmic space may additively induce the pathway. Upon generation of a signal, is released from the complex with RseA, leading to a positive feedback loop. The newly active transcribes its own promoter to generate more and RseA. As long as the signal is present, RseA will be unable to interact with , but when the signal is removed or reduced, RseA, possibly in concert with RseB, will again repress , achieving a new steady state level. Although this model bears a superficial resemblance to the regulation of , it is unlikely that RseA targets for degradation, or that RseA interacts with the signal in the same manner as it interacts with . 3.5. The Cellular Role of is an essential sigma factor, at least at temperatures above 18°C, and cells lacking rapidly accumulate a suppressor of this lethality (De Las Peñas et al., 1997b). Cells lacking and containing this suppressor form colonies at 42°C to 43°C with greatly reduced efficiency (10-3 to 10-5), and die more rapidly than wild type cells after exposure to lethal temperatures (Hiratsu et al., 1995; Raina et al., 1995; Rouvière et al., 1995), while cells containing the suppressor alone are temperature resistant (Connolly and Gross, unpublished observations). These phenotypes confirm the importance of the regulon for resistance to thermal stress. Overexpression of sE leads to the induction of at least 10 proteins (Raina et al., 1995; Rouvière et al., 1995). However, only four members of the regulon have been identified. In addition to rpoH, EsE transcribes the periplasmic protease degP, the periplasmic peptidyl-prolyl isomerase fkpA (Danese and Silhavy, 1997), and one of the two promoters upstream of rpoE itself. Why does E. coli need two heat-inducible regulons? Part of the answer might be that the two regulons respond to stress in different cellular compartments. Some inducers, such as heat and solvents, affect all cellular compartments and thus induce both regulons. Other inducers specifically alter protein folding in either the cytoplasmic or extracytoplasmic environments, and uniquely induce or activity, respectively.
Autoregulation of the heat
29
Just as the response has a close parallel in the eukaryotic heat shock response, the pathway also has a eukaryotic counterpart. Accumulation of unfolded proteins in the endoplasmic reticulum (ER) leads to the transcriptional induction of several ER resident folding agents (Cox et al., 1993; Mori et al., 1993). Like the pathway, the ER response, known as the unfolded protein response (UPR), is controlled separately from the cytoplasmic heat shock response. Although the central regulator of the UPR shares no common features with RseA, it remains to be seen whether the two systems share common mechanisms of sensing the initial signal. E. coli has a second signal transduction pathway, the Cpx two-component system, capable of relieving extracytoplasmic stress. Although the Cpx system is not required for growth at high temperature (Connolly et al., unpublished observations), activation of the pathway suppresses the envelope-associated toxicity conferred by certain LamB mutant proteins by inducing the expression of DegP (Cosma et al., 1995; Danese et al., 1995; Snyder et al., 1995). Interestingly, activation of the Cpx pathway also restores the ability to grow at high temperature to cells lacking , in a degP-dependent manner (Connolly, et al., 1997). Overexpression of degP alone does not suppress the rpoE-temperature sensitive phenotype, indicating that other Cpx-controlled genes are required. Future work aimed at elucidating the relationship between the Cpx pathway and the -mediated response should help to clarify the roles of each system in responding to protein misfolding outside of the cytoplasm. Work on the pathway is just beginning. The next few years should provide us with exciting insights into the members of the regulon, the nature of the signal, and the regulatory network that links the cellular compartments. In addition, has already proven to be an invaluable tool in the search for periplasmic folding agents and rapid progress in the understanding of folding processes in this cellular compartment is likely to follow.
4. HEAT SHOCK REGULATION IN OTHER PROKARYOTIC ORGANISMS Study of the heat shock response in a number of different bacteria indicates that the basic E. coli regulatory paradigm is not universal. Although homologues are widespread among gram negative bacteria, additional regulatory mechanisms also affect the primary heat shock response in some of these organisms. Moreover, the gram positive organisms examined to date do not have homologues. homologues have been isolated from a number of Gram negative bacteria (Garvin et al., 1989; Benvenisti et al., 1995; Fleischmann et al., 1995; Naczynski et al., 1995; Nakahigashi et al., 1995; Yura, 1996). All of these homologues can restore growth to E. coli cells lacking functional , indicating that the transcriptional function of the protein is conserved. However, sequence analysis suggests that only some of the regulatory inputs are conserved. All homologues identified to date contain Region C, which binds DnaK with high affinity and is required for control of stability. In contrast, the regions of mRNA implicated in translational control are conserved in but not proteobacteria. If translational control of exists in a proteobacteria, it must be
Molecular chaperones and folding catalysts
30
mechanistically distinct from the E. coli model. These observations suggest that diverse mechanisms may control the amount and/ or activity of in different gram negative species. Our knowledge about the heat shock response in gram positive organisms comes from studies of Bacillus subtilis and Clostridium acetobutylicum. (Narberhaus et al., 1992; Narberhaus et al., 1992; Schmidt et al., 1992; Wetzstein et al., 1992; Zuber et al., 1994; Yura, 1996). In these organisms, the major chaperone genes are transcribed by the housekeeping sigma and are preceded by a conserved inverted repeat sequence. This inverted repeat, named CIRCE for controlling inverted repeat for chaperone expression, is the binding site for a putative represser (Yuan et al., 1995). The mechanism of thermal induction of genes regulated by the CIRCE element has not yet been elucidated. CIRCE has also been detected in some gram negative bacteria suggesting that it is rather widely involved in the heat shock response. In Bradyrhizobium japonicum, and CIRCE together control expression of heat shock genes (Babst et al., 1996), suggesting that parallel regulatory strategies may exist in some organisms. In contrast to , the degree of conservation of has not been determined. Although several sigma factors belonging to the ECF family have been described in both gram negative and positive bacteria (Lonetto et al., 1994; Rouvière et al., 1995), their possible role in the heat shock response of these organisms has not been widely studied. Only one of the ECF sigmas in addition to has been implicated in the resistance to thermal stress. Pseudomonas aeruginosa cells lacking the homologue AlgU, show increased killing at 50°C compared to AlgU+ strains (Martin et al., 1994), and the activity of AlgU is induced in response to heat shock (Schurr et al., 1995). However, AlgU carries out additional cellular functions not mediated by . For example, AlgU-cellsshowincreased sensitivity to superoxide-generating compounds (Martin et al., 1994), and AlgU plays a key role in the production of the exopolysaccharide alginate (Deretic et al., 1994). One possibility is that the -mediated response has been co-opted by other signaling systems in P. aeruginosa, and it will be interesting to determine how AlgU and utilize similar signaling molecules to respond to diverse extracellular signals.
5. SUMMARY AND PROSPECTS Although recent studies have given us insight into the mechanisms responsible for the regulation of both and , several basic questions concerning the response to thermal stress in E. coli remain unresolved. For example, the exact nature of the initial signal and sensing mechanism have not been elucidated. Further dissection of the response loops of each sigma factor should provide us with a greater understanding of not only the heat shock response but also of the process of protein folding in each cellular compartment. We have only begun to understand the in vivo role of the chaperones, and to identify periplasmic protein folding agents. The next few years should prove to be an exciting time in the dual fields of thermal stress response and protein folding.
Autoregulation of the heat
31
6. ACKNOWLEDGMENTS We thank Jonathan Tupy for help in preparing figures, and Charlotte Hedlund for excellent assistance in editing and performing the innumerable tasks required to complete this manuscript.
7. REFERENCES Allen, S.P., Polassi, J.O., Gierse, J.K. and Easton, A.M. (1992). Two novel heat shock genes encoding proteins produced in response to heterologous protein expression in Escherichia coli. J. Bacteriol. , 174 , 6938–6947. Babst, M., Hennecke, H. and Hans-Martin, F. (1996). Two different mechanisms are involved in the heat-shock regulation of chaperonin gene expression in Bradyrhizobium japonicum. Mol. Micro. , 19 , 827–839. Bardwell, J.C. and Craig, E.A. (1984). Major heat shock gene of Drosophilia and the Escherichia coli heat-inducible dnaK gene are homologous. Proc. Natl. Acad. Sci. USA , 81 , 848–852. Bardwell, J.C., Tilly, K., Craig, E., King, J., Zylicz, M. and Georgopolous, C. (1986). The nucleotide sequence of the Escherichia coli K12 dnaJ+ gene. A gene that encodes a heat shock protein. J. Biol. Chem. , 261 , 1782–1785. Bardwell, J.C. and Craig, E.A. (1987). Eukaryotic Mr 83,000 heat shock protein has a homologue in Escherichia coli. Proc. Natl. Acad. Sci. USA , 84 , 5177–5181. Burton, Z., Burgess, R.R., Lin, J., Moore, D., Holder, S. and Gross, C.A. (1981). The nucleotide sequence of the cloned rpoD gene for the RNA polymerase sigma subunit from Escherichia coli K12. Nucleic Acids Res. , 9 , 2889–2903. Benvenisti, L., Koby, S., Rutman, A., Giladi, H., Yura, T. and Oppenheim, A.B. (1995). Cloning and primary sequence of the rpoH gene from Pseudomonas aeruginosa. Gene , 155 , 73–76. Brissette, J.L., Russel, M., Weiner, L. and Model, P. (1990). Phage shock protein, a stress protein of Escherichia coli. Proc. Natl. Acad. Sci. , 87 , 862–866. Bukau, B. (1993). Regulation of the Escherichia coli heat-shock response. Mol. Micro. , 9, 671–680. Charpentier, B. and Branlaut, C. (1994). The Escherichia coli gapA gene is transcribed by the vegetative RNA polymerase holoenzyme E sigma-70 and by the heat shock RNA polymerase E sigma-32. J. Bacteriol. , 176 , 830–839. Chuang, S.E., Daniels, D.L. and Blattner, F.R. (1993). Global regulation of gene expression in Escherichia coli. J. Bacteriol. , 175 , 2026–2036. Chuang, S.E. and Blattner, F.R. (1993). Characterization of twenty-six new heat shock genes of Escherichia coli. J. Bacteriol. , 175 , 5242–5252. Chuang, S.E., Burland, V., Plunkett, G.R., Daniels, D.L. and Blattner, F.R. (1993). Sequence analysis of four new heat-shock genes constituting the hslTS/ibpAB and hsl/VU operons in Escherichia coli. Gene , 134 , 1–6. Connolly, L., De Les Peñas, A., Alba, B.M. and Gross, C.A. (1997). The response to extracytoplasmic stress in Escherichia coli is controlled by partially overlapping pathways. Genes Dev . 11 , 2012–2021. Cooper, S. and Ruettinger, T. (1975). A temperature sensitive nonsense mutation
Molecular chaperones and folding catalysts
32
affecting the synthesis of a major protein of Escherichia coli K12. Mol Gen Genet. , 139 , 167–176. Cosma, C.L., Danese, P.N., Carlson, J.H., Silhavy, T.J. and Snyder, W.B. (1995). Mutational activation of the Cpx signal transduction pathway of Escherichia coli suppresses the toxicity conferred by certain envelope-associated stresses . Mol. Micro. , 18 , 494–505. Cowing, D.W., Bardwell, J.C., Craig, E.A., Woolford, C., Hendrix, R.W. and Gross, C.A. (1985). Consensus sequence for Escherichia coli heat shock gene promoters. Proc. Natl. Acad. Sci. USA , 82 , 2679–2683. Cox, J.S., Shamu, C.E. and Walter, P. (1993). Transcriptional induction of genes encoding endoplasmic reticulum resident proteins requires a transmembrane protein kinase. Cell , 73 , 1197–1206. Craig, A.E. and Gross, C.A. (1991). Hsp70-the cellular thermometer? Trends Biochem. Sci. , 16 , 135–140. Danese, P. and Silhavy, T. (1997). The and Cpx signal transduction systems control the synthesis of periplasmic protein-folding enzymes. Genes Dev . 11 , 1183–1193. Danese, P.N., Snyder, W.B., Cosma, C.L., Davis, L.J. and Silhavy, T.J. (1995). The Cpx two-component signal transduction pathway of Escherichia coli; regulates transcription of the gene specifying the stress-inducible periplasmic protease, DegP. Genes. Dev. , 9 , 387–398. De Las Peñas, A., Connolly, L. and Gross, C.A. (1997a). The -mediated response to extracytoplasmic stress in Escherichia coli is transduced by RseA and RseB, two negative regulators of . Mol. Microbiol. 24 , 373–385. De Las Peñas, A., Connolly, L., and Gross, C.A. (1997b). is an essential sigma factor in Escherichia coli. J. Bacteriol. 179 , 6862–6864. Deretic, V., Schurr, M.J., Boucher, J.C. and Martin, D.W. (1994). Conversion of Pseudomonas aeruginosa to mucoidy in cystic fibrosis: environmental stress and regulation of bacterial virulence by alternative sigma factors. J. Bacteriol. , 176 , 2773–2780. Erickson, J.W. and Gross, C.A. (1989). Identification of the sigma E subunit of Escherichia coli RNA polymerase: a second alternate sigma factor involved in hightemperature gene expression. Genes Dev. , 3 , 1462–1471. Erickson, J.W., Vaughn, V., Walter, W.A., Neidhardt, F.C. and Gross, C.A. (1987). Regulation of the promoters and transcripts of rpoH, the Escherichia coli heat-shock regulatory gene. Genes Dev. , 1 , 419–432. Fleischmann, R.D., Adams, M.D., White, O., Clayton, R.A., Kirkness, E.F., Kerlavage, A.R., Bult, C.J., Tomb, J.F., Dougherty, B.A. and Merrick, J.M. (1995). Wholegenome random sequencing and assembly of Haemophilus influenzae Rd. Science , 269 , 496–512. Gamer, J., Bujard, H. and Bukau, B. (1992). Physical interaction between heat shock proteins DnaK, DnaJ and GrpE and the bacterial heat shock transcription factor sigma 32. Cell , 69 , 833–842. Gamer, J., Multhaup, G., Tomoyasu, T., McCarty, J.S., Rüdiger, S., Schönfeld, H.J., Schirra, C., Bujard, H. and Bukau, B. (1996). A cycle of binding and release of the DnaK, DnaJ and GrpE chaperones regulates activity of the Escherichia coli heat shock transcription factor sigma32. Embo J. , 15 , 607–617. Garvin, L.D. and Hardies, S.C. (1989). Nucleotide sequence for the htpR gene from Citrobacter freundii. Nucleic Acids Res , 17 , 4889.
Autoregulation of the heat
33
Gayda, R.C., Stephens, P.E., Hewick, R., Schoemaker, J.M., Dreyer, N.J. and Markowitz, A. (1985). Regulatory sequence of the heat shock-inducible capR (lon), gene: DNA and protein sequences. J. Bacteriol. , 162 , 271–275. Georgopoulos, C., Tilly, K., Drahos, D. and Hendrix, R. (1982). The B66.0 protein of Escherichia coli is the product of the dnaK gene. J. Bacteriol. , 149 , 1175–1177. Gottesman, S., Clark, W.P., de Crecy-Lagard, V. and Maurizi, M.R. (1993). ClpX, an alternative subunit for the ATP-dependent Clp protease of Escherichia coli. J. Biol. Chem. , 268 , 22618–22626. Gross, C.A. (1996). Function and Regulation of the Heat Shock Proteins. In Neidhardt, F.C. (chief ed.), Escherichia coli and Salmonella: Molecular and Cellular Biology , ASM Press, Washington, D.C., pp. 1382–1399. Grossman, A.D., Erickson, J.W. and Gross, C.A. (1984). The htpRgene product of E. coli is a sigma factor for heat-shock promoters. Cell , 38 , 383–390. Hemmingsen, S.M., Woolford, C., van der Vies, S.M., Tilly, K., Dennis, D.T., Georgopolous, C.P., Hendrix, R.W. and Ellis, R.J. (1988). Homologous plant and bacterial proteins chaperone oligomeric protein assembly. Nature , 333 , 330–334. Hengge-Aronis, R. (1996). Back to log phase: sigma-s as a global regulator in the osmotic control of gene expression in Escherichia coli. Mol. Micro. , 21 , 887–893. Herman, C., Thevenet, D., D’Ari, R. and Bouloc, P. (1995). Degradation of sigma 32, the heat shock regulator in Escherichia coli, is governed by HflB. Proc. Natl. Acad. Sci. USA , 92 , 3516–3520. Hiratsu, K., Amemura, M., Nashimoto, H., Shinagawa, H. and Makino, K. (1995). The rpoE gene of Escherichia coli, which encodes sigma E, is essential for bacterial growth at high temperature. J. Bacteriol. , 177 , 2918–2922. Kamath-Loeb, A.S. and Gross, C.A. (1991). Translational regulation of sigma-32 synthesis: requirement for an internal control element. J. Bacteriol. , 270 , 3904–3906. Kitagawa, M., Wada, C., Yoshioka, S. and Yura, T. (1991). Expression of ClpB, an analog of the ATP-dependent protease regulatory subunit in Escherichia coli, is controlled by a heat shock sigma factor (sigma 32). J. Bacteriol. , 173 , 4247–4253. Kornitzer, D., Teff, D., Altuvia, S. and Oppenheim, A.B. (1991). Isolation, characterization and sequence of an Escherichia coli heat shock gene, hlpX. J. Bacteriol. , 173 , 2944–2953. Landick, R., Vaughn, V., Lau, E.T., VanBogelen, R.A., Erickson, J.W. and Neidhardt, F.C. (1984). Nucleotide sequence of the heat shock regulatory gene of E. coli suggests its protein product may be a transcription factor. Cell , 38 , 175–182. Lemaux, P.G., Herendeen, S.L., Bloch, P. and Neidhardt, F.C. (1978). Transient rates of synthesis of individual polypeptides in E. coli following temperature shifts. Cell , 13 , 427–434. Lesley, S.A., Thompson, N.E. and Burgess, R.R. (1987). Studies of the role of the Escherichia coli heat shock regulatory protein sigma 32 by the use of monoclonal antibodies. J. Biol. Chem. , 262 , 5404–5407. Leveque, F., Plateau, P., Dessen, P. and Blanquet, S. (1990). Homology of lysS and lysU, the two Escherichia coli genes encoding distinct lysyl-tRNA synthetase species. Nucleic Acids Res. , 18 , 305–312. Liberek, K., Galitski, T.P., Zylicz, M. and Georgopolous, C. (1992). The DnaK chaperone modulates the heat shock response of Escherichia coli by binding to the sigma 32 transcription factor. Proc. Natl. Acad. Sci. USA , 89 , 3516–3520. Liberek, K. and Georgopoulos, C. (1993). Autoregulation of the Escherichia coli heat shock response by the DnaK and DnaJ heat shock proteins. Proc. Natl. Acad. Sci.
Molecular chaperones and folding catalysts
34
USA , 90 , 11019–11023. Liberek, K., Wall, D. and Georgopolous, C. (1995). The DnaJ chaperone catalytically activates the DnaK chaperone to preferentially bind the sigma-32 heat shock transcriptional regulator. Proc. Natl. Acad. Sci , 92 , 6224–6228. Lipinska, B., King, J., Ang, D. and Georgopolous, C. (1988). Sequence analysis and transcriptional regulation of the Escherichia coli grpE gene encoding a heat shock protein. Nuc’eic Acids Res. , 16 , 7545–7562. Lipinska, B., Sharma, S. and Georgopolous, C. (1988). Sequence analysis and regulation of the htrA gene of Escherichia coli: a sigma 32-independent mechanism of heatinducible transcription. Nucleic Acids Res. , 16 , 10053–10067. Lonetto, M.A., Brown, K.L., Rudd, K.E. and Buttner, M.J. (1994). Analysis of the Streptomyces coelicolor sigE gene reveals the existence of a subfamily of eubacterial RNA polymerase sigma factors involved in the regulation of extracytoplasmic functions . Proc. Natl. Acad. Sci. USA , 91 , 7573–7577. Martin, D.W., Schurr, M.J., Mudd, M.H., Govan, J.R., Holloway, B.W. and Deretic, V. (1993). Mechanism of conversion to mucoidy in Pseudomonas aeruginosa infecting cystic fibrosis patients. Proc. Natl. Acad. Sci. USA , 90 , 8377–8381. Martin, D.W., Schurr, M.J., Yu, H. and Deretic, V. (1994). Analysis of promoters controlled by the putative sigma factor AlgU regulating conversion to mucoidy in Pseudomonas aeruginosa: relationship to sigma E and stress response. J. Bacteriol. , 176 , 6688–6696. Maurizi, M.R., Clark, W.P., Katayama, V., Rudikoff, S., Pumphrey, J., Bowers, B. and Gottesman, S. (1990). Sequence and structure of ClpP, the proteolytic component of the ATP-dependent Clp protease of Escherichia coli. J. Biol. Chem. , 265 , 12536– 12545. McCarty, J.S., Rudiger, S., Schonfeld, H.J., Schneider-Mergener, J., Nakahigushi, K., Yura, T. and Bukau, B. (1996). Regulatory region C of the E. coli heat shock transcription factor, sigma32, constitutes a DnaK binding site and is conserved among eubacteria. J. Mol. Biol. , 256 , 829–837. Mecsas, J., Rouvière, P.E., Erickson, J.W., Donohue, T.J. and Gross, C.A. (1993). The activity of sigma E, an Escherichia coli heat-inducible sigma-factor, is modulated by expression of outer membrane proteins. Genes Dev. , 7 , 2618–2628. Missiakas, D., Betton, J.-M. and Raina, S. (1996). New components of protein folding in extracytoplasmic compartments of Escherichia coli SurA, FkpA and Skp/OmpH. Mol. Micro. , 21 , 871–884. Missiakas, D., Schwager, F., Betton, J.M., Georgopoulos, C. and Raina, S. (1996). Identification and characterization of HslV HslU (ClpQ ClpY) proteins involved in overall proteolysis of misfolded proteins in Escherichia coli. EMBO J . 15 , 6899– 6909. Missiakas, D., Mayer, M.P., Lemaire, M., Georgopoulos, C. and Raina, S. (1997). Modulation of the Escherichia coli (RpoE) heat-shock transcription-factor activity by the RseA, RseB and RseC proteins. Mol. Microbiol . 24 , 355–371. Mori, K., Ma, W., Gething, M.J. and Sambrook, J. (1993). A transmembrane protein with a cdc2+/ CDC28-related kinase activity is required for signaling from the ER to the nucleus. Cell , 74 , 743–756. Naczynski, Z.M., Mueller, C. and Kropinski, A.M. (1995). Cloning the gene for the heat shock response positive regulator (sigma 32 homolog). from Pseudomonas aeruginosa. Can. J. Microbiol. , 41 , 75–87. Nagai, H., Yuzawa, H., Kanemori, M. and Yura, T. (1994). A distinct segment of the
Autoregulation of the heat
35
sigma 32 polypeptide is involved in DnaK-mediated negative control of the heat shock response in Escherichia coli. Proc. Natl. Acad. Sci. USA , 91 , 10280–10284. Nagai, H., Yuzawa, H. and Yura, T. (1991). Interplay of two cis-acting mRNA regions in translational control of sigma 32 synthesis during the heat shock response of Escherichia coli. Proc. Natl. Acad. Sci. USA , 88 , 10515–10519. Nakahigashi, K., Yanagi, H. and Yura, T. (1995). Isolation and sequence analysis of rpoH genes encoding sigma 32 homologs from gram negative bacteria: conserved mRNA and protein segments for heat shock regulation. Nucleic Acids Res , 23 , 4383– 4390. Narberhaus, F. and Bahl, H. (1992). Cloning, sequencing and molecular analysis of the groESL operon of Clostridium acetobutylicum. J. Bacterial. , 174 , 3282–3289. Narberhaus, F., Giebeler, K. and Bahl, H. (1992). Molecular characterization of the dnaK gene region of Clostridium acetobutylicum, including grpE, dnaJ and a new heat shock gene. J. Bacteriol. , 174 , 3290–3299. Nashimoto, H. (1993). The Translational Apparatus , New York, Plenum Press. Neidhardt, F.C. and VanBogelen, R.A. (1981). Positive regulatory gene for temperaturecontrolled proteins in Escherichia coli. Biochem. Biophys. Res. Commun. , 100 , 894– 900. Neidhardt, F.C. and VanBogelen, R.A. (1987). Heat Shock Response. In F.C.Neidhardt, (ed.), Escherichia coli and Salmonella typhimuriam: Cellular and Molecular Biology , American Society for Microbiology,Washington, D.C., pp. 1334–1345. Neidhardt, F.C., VanBogelen, R.A. and Vaughn, V.A. (1984). The genetics and regulation of heat-shock proteins. Annu. Rev. Genet , 18 , 295–329. Neidhardt, F.C., Phillips, T.A., VanBogelen, R.A., Smith, M.W., Georgalis, V. and Subramanian, A.R. (1981). Identity of the B56.5 protein, the A-protein and the groE gene product of Escherichia coli. J. Bacteriol. , 145 , 513–520. Pack, K.H. and Walker, G.C. (1986). Defect in expression of heat-shock proteins at high temperature in xthA mutants. J. Bacteriol. , 165 , 763–770. Raina, S. and Georgopolous, C. (1991). The htrM gene, whose product is essential for Escherichia coli viability at elevated temperatures, is identical to the rfaD gene. Nucleic Acids Res. , 19 , 3811–3819. Raina, S., Missiakas, D. and Georgopolous, C. (1995). The rpoE gene encoding the sigma E (sigma 24). heat shock sigma factor of Escherichia coli. Embo. J. , 14 , 1043– 1055. Rouvière, P.E., De Las Peñas, A., Mecsas, J., Lu, C.Z., Rudd, K.E. and Gross, C.A. (1995). rpoE, the gene encoding the second heat-shock sigma factor, sigma E, in Escherichia coli. Embo. J. , 14 , 1032–1042. Rouvière, P.E. and Gross, C.A. (1996). SurA, a Periplasmic Protein with Peptidyl Propyl Isomerase Activity, Participates in the Assembly of Outer Membrane Porins. Genes Dev. , 10 , 3170–3182. Schmidt, A., Schiesswohl, M., Volker, U., Hecker, M. and Schumann, W. (1992). Cloning, sequencing, mapping and transcriptional analysis of the groESL operon from Bacillus subtilis . J. Bacteriol. , 174 , 3993–3999. Schurr, M.J., Yu, H., Boucher, J.C., Hibler, N.S. and Deretic, V. (1995). Multiple promoters and induction by heat shock of the gene encoding the alternative sigma factor AlgU (sigma E). which controls mucoidy in cystic fibrosis isolates of Pseudomonas aeruginosa. J. Bacteriol. , 177 , 5670–5679. Schurr, M.J., Yu, H., Martinez-Salazar, J.M., Boucher, J.C. and Deretic, V. (1996). Control of AlgU, a member of the sigma E-like family of stress sigma factors, by the
Molecular chaperones and folding catalysts
36
negative regulators MucA and MucB and Pseudomonas aeruginosa conversion to mucoidy in cystic fibrosis. J. Bacteriol. , 178 , 4997–5004. Skelly, S., Coleman, T., Fu, C.F., Brot, N. and Weissbach, H. (1987). Correlation between the 32-kDa sigma factor levels and in vitro expression of Escherichia coli heat shock genes. Proc. Natl Acad. Sci. USA , 84 , 8365–8369. Snyder, W.B., Davis, L.J., Danese, P.N., Cosma, C.L. and Silhavy, T.J. (1995). Overproduction of NlpE, a new outer membrane lipoprotein, suppresses the toxicity of periplasmic LacZ by activation of the Cpx signal transduction pathway. J. Bacteriol. , 177 , 4216–4223. Squires, C.L., Pedersen, S. and Ross, B.M. (1991). ClpB is the Escherichia coli heat shock protein F84.1. J. Bacteriol. , 173 , 4254–4262. Strauch, K.L., Johnson, K. and Beckwith, J. (1989). Characterization of degP, a gene required for proteolysis in the cell envelope and essential for growth of Escherichia coli at high temperature. J. Bacteriol. , 171 , 2689–2696. Straus, D.B., Walter, W. and Gross, C.A. (1990). DnaK, DnaJ and GrpE heat shock proteins negatively regulate heat shock gene expression by controlling the synthesis of sigma-32. Genes Dev . 4 , 2202–2209. Straus, D.B., Walter, W.A. and Gross, C.A. (1987). The heat shock response of E. coli is regulated by changes in the concentration of sigma 32. Nature , 329 , 348–351. Straus, D.B., Walter, W.A. and Gross, C.A. (1989). The activity of sigma 32 is reduced under conditions of excess heat shock protein production in Escherichia coli. Genes Dev. , 3 , 2003–2010. Taura, T., Kusukawa, N., Yura, T. and Ito, K. (1989). Transient shut off of Escherichia coli heat shock protein synthesis upon temperature shift down. Biochem. Biophys. Res. Commun. , 163 , 438–443. Taylor, W.E., Straus, D.B., Grossman, A.D., Burton, Z.F., Gross, C.A. and Burgess, R.R. (1984). Transcription from a heat-inducible promoter causes heat shock regulation of the sigma subunit of E. coli RNA polymerase. Cell , 38 , 371–381. Tilly, K., Spence, J. and Georgopolous, C. (1989). Modulation of stability of the Escherichia coli heat shock regulatory factor sigma. J. Bacteriol. , 171 , 1585–1589. Tomoyasu, T., Gamer, J., Bukau, B., Kanemori, M., Mori, H., Rutman, A.J., Oppenheim, A.B., Yura, T., Yamanaka, K., Niki, H., Hiraga, S. and Ogura, T. (1995). Escherichia coli: FtsH is a membrane-bound, ATP-dependent protease which degrades the heatshock transcription factor sigma 32. Embo J. , 14 , 2551–2560. Tomoyasu, T., Yamanaka, K., Murata, K., Suzaki, T., Bouloc, P., Kato, A., Niki, H., Hiraga, S. and Ogura, T. (1993). Topology and subcellular localization of FtsH protein in Escherichia coli. J. Bacteriol. , 175 , 1352–1357. Wang, Q.P. and Kaguni J.M. (1989). A novel sigma factor is involved in expression of the rpoH gene of Escherichia coli . J. Bacterial , 171 , 4248–4253. Wetzstein, M., Volker, U., Dedio, J., Lobau, S., Zuber, U., Schiesswohl, M., Herget, C., Hecker, M. and Schumann, W. (1992). Cloning, sequencing and molecular analysis of the dnaK locus from Bacillus subtilis. J. Bacteriol. , 174 , 3300–3310. Xie, Z.D., Hershberger, C.D., Shankar, S., Ye, R.W and Chakrabarty, A.M. (1996). Sigma factor-anti-sigma factor interaction in alginate synthesis: inhibition of AlgT by MucA. J. Bacteriol , 178 , 4990–4996. Yamamori, T., Ito, K., Nakamura, Y. and Yura, T. (1978). Transient regulation of protein synthesis in Escherichia coli upon shift-up of growth temperature. J. Bacteriol. , 134 , 1133–1140. Yamamori, T. and Yura, T. (1980). Temperature-induced synthesis of specific proteins in
Autoregulation of the heat
37
Escherichia coli: evidence for transcriptional control. J. Bacteriol. , 142 , 843–851. Yamamori, T. and Yura, T. (1982). Genetic control of heat-shock protein synthesis and its bearing on growth and thermal resistance in Escherichia coli K-12. Proc. Natl. Acad. Sci. USA , 79 , 860–864. Yu, H., Schurr, M.J. and Deretic, V. (1995). Functional equivalence of Escherichia coli sigma E and Pseudomonas aeruginosa AlgU: E. coli rpoE restores mucoidy and reduces sensitivity to reactive oxygen intermediates in algU mutants of P. aeruginosa . J. Bacteriol. , 177 , 3259–3268. Yuan, G. and Wong, S.L. (1995). Isolation and characterization of Bacillus subtilis groE regulatory mutants: evidence for or/39 in the dnaK operon as a represser gene in regulating the expression of both groE and dnaK J. Bacteriol. , 177 , 6462–6468. Yura, T. (1996). Regulation and conservation of the heat-shock transcription factor sigma-32. Genes to Cells , 1 , 277–284. Yura, T., Tobe, T., Ito, K. and Osawa, T. (1984). Heat shock regulatory gene (htpR). of Escherichia coli is required for growth at high temperature but is dispensable at low temperature. Proc. Natl. Acad. Sci. USA , 81 , 6803–6807. Yuzawa, H., Nagai, H., Mori, H. and Yura, T. (1993). Heat induction of sigma 32 synthesis mediated by mRNA secondary structure: a primary step of the heat shock response in Escherichia coli. Nucleic Acids Res. , 21 , 5449–5455. Zhou, Y.N., Kusukawa, N., Erickson, J.W., Gross, C.A. and Yura, T. (1988). Isolation and characterization of Escherichia coli mutants that lack the heat shock sigma factor sigma 32. J. Bacteriol. , 170 , 3640–3649. Zuber, U. and Schumann, W. (1994). CIRCE, a novel heat shock element involved in regulation of heat shock operon dnaK of Bacillus subtilis. J. Bacteriol. , 176 , 1359– 1363.
3. INDUCIBLE TRANSCRIPTIONAL REGULATION OF HEAT SHOCK GENES: THE STRESS SIGNAL AND THE UNFOLDED PROTEIN RESPONSE RICHARD I.MORIMOTO Department of Biochemistry, Molecular Biology and Cell Biology, Rice Institute for Biomedical Research, Northwestern University, 2153 Sheridan Road, Evanston, Il. 60208, USA
1. Introduction 1.1 Inducible Transcription by a Family of Heat Shock Factors 2. The HSFI Cycle: Proposed Role for Heat Shock Proteins in Autoregulation of the Heat Shock Response 3. Activation of HSF2 is Linked to Proteolysis and the Activity of the UbiquitinDependent Proteasome 4. HSF 3 and HSF4 5. Transcriptional Regulation of Endoplasmic Reticulum Stress Genes (GRP78): The Unfolded Protein Response 6. Conclusions 7. Acknowledgement 8. References
1. INTRODUCTION The activation of heat shock gene expression is a highly regulated response to diverse environmental and physiological conditions including heat shock, oxidative stress, heavy metals, various chemicals, bacterial and viral infection, and exposure to a number of acute and chronic disease states (Ritossa, 1962; Ashburner, 1970; Lindquist and Craig, 1988; Morimoto et al., 1990; 1994). These conditions (Figure 1) can be partitioned into three broad catergories: (1) environmental stress including heat shock, amino acid analogues, drugs, toxic chemicals, and heavy metals, (2) pathophysiologic and disease states including oxidative stress, fever, inflammation, infection, myocardial stress and ischemia, and neural degenerative diseases, and (3) non-stress conditions including the cell cycle, growth factors, serum stimulation, development, differentiation, and activation
Molecular chaperones and folding catalysts
40
by certain oncogenes. The diverse nature of these conditions has led to questions on the identity and properties of the cellular machinery that detects “stress”. This chapter will address the stress-sensing mechanisms that operate in the cytosol and lumen of eukaryotic cells and that lead to the transcriptional activation of specific genes encoding compartment specific heat shock proteins and molecular chaperones.
Figure 1 Conditions that result in the induction of heat shock gene expression in eukaryotes. Representation of three general classes of conditions known to result in the elevated expression of stress proteins including: (1) environmental and physiological stress, (2) pathophysiological states including conditions of disease and (3) nonstressful conditions such as cell growth and development. Each condition acts on the cell as diagrammed in this figure and in the case of environmental stress and certain pathophysiological states leads to the activation of heat shock gene expression and the synthesis of heat shock proteins.
1.1. Inducible Transcription by a Family of Heat Shock Factors Exposure of eukaryotic cells to elevated temperatures and other stressful conditions leads to the rapid activation of heat shock transcription factor (HSF) and the inducible transcription of genes encoding heat shock proteins and molecular chaperones (Lindquist and Craig, 1988; Morimoto, 1993; Lis and Wu, 1993; Wu, 1995). Analysis of the
Inducible transcriptional regulation of the heat shock genes
41
chromosomal structure in the vicinity of the Drosophila Hsp70 and Hsp90 genes led to the identification of binding sites for constitutive and inducible transcription factors (Wu, 1980, 1984). Heat Shock Factor (HSF) was initially identified as the inducible transcription factor which bound specifically to the heat shock element (HSE), a pentameric nucleotide sequence (5'-AGAAn-3') positioned as inverted adjacent arrays in the promoter region of eukaryotic heat shock-responsive genes (Amin et al., 1988; Xiao and Lis, 1988; Perisic et al., 1989; Williams and Morimoto, 1990; Greene and Kingston, 1990; Xiao et al., 1991). A distinction between the heat shock response of yeast and metazoans is the constitutive DNA binding properties of yeast HSF, whereas in Drosophila, plants, and vertebrates, the HSFs are constitutively expressed, negatively regulated for DNA binding, and require heat shock for acquisition for DNA binding (Sorger and Pelham, 1988; Kingston et al., 1987; Zimarino and Wu, 1987; Larson et al., 1988; Mosser et al., 1988; Nover, 1994; Zimarino et al., 1990; Westwood et al., 1991; 1993). The mechanism by which HSFs are negatively regulated involves both cis-acting negative regulatory elements and constitutive phosphorylation which function to repress DNA binding and the transactivation domain (Shi et al., 1995; Knauf et al., 1996; Kline and Morimoto, 1997). The inducibility of the latent Drosophila and mammalian HSFs can also be observed in vitro following exposure to in vitro heat shock, non-ionic detergents, low pH, or exposure to chaotropes (Larson et al., 1988; Mosser et al., 1990). The initial biochemical characterization of HSFs and the subsequent cloning of the respective genes was initially accomplished in S. cerevisiae and D. melanogaster (Sorger et al., 1987; Wiederrecht et al., 1988; Wu et al., 1987). Subsequently, the cloning of HSF genes from larger eukaryotes revealed a multi-gene family. At least three HSFs have been isolated from the human (HSFs 1, 2, 4), chicken (HSFs 1, 2, 3), and tomato genomes (Scharf et al., 1990; Rabindran et al., 1991; Sarge et al., 1991; Schuetz et al., 1991; Nakai and Morimoto, 1993; Nakai et al., 1997). The cloned HSFs vary in size from 301 amino acids (aa) for tomato HSF24, 512aa for tomato HSF8, 503aa and 529aa for mouse and human HSF1, 517aa and 536aa for mouse and human HSF2, 463aa for chicken HSF3, 467aa for human HSF4, 691aa for Drosophila HSF and 833aa for S. cerevisiae HSF (Scharf et al., 1990; Nakai and Morimoto, 1993; Rabindran et al., 1991; Sarge et al., 1991; Schuetz et al., 1991; Nakai et al., 1997). Within a species, members of the HSF family (ie., mouse HSF1-HSF2 or chicken HSF1-HSF2-HSF3) are approximately 40% related in amino acid sequences; this is primarily due to sequence identity within the DNA binding and oligomerization domains (Morimoto, 1993; Wu, 1995). Interspecies comparisons, ie., between human, mouse, and chicken HSF1 reveals a higher level of sequence identity (85–95%) revealing a common ancestral progenitor from which the contemporary HSFs evolved (Nakai and Morimoto, 1993). Functional domains of HSFs (Figure 2) were identified initially by alignment of the amino acid sequences derived from the respective cloned genes and subsequently by functional assays using collections of deletion and point mutants. Comparison of the HSFs from yeasts, Drosophila, tomato, chicken, mouse, and humans identified a conserved ≈100 amino acid DNA-binding domain of the winged helix-turn-helix motif located towards the amino-terminus (Damberger et al., 1994; Harrison et al., 1994; Vuister et al., 1994a; 1994b). The high degree of conservation in the DNA binding domain reflects the conservation of the nucleotide sequences comprising the HSE DNA
Molecular chaperones and folding catalysts
42
binding motif. Adjacent to the DNA-binding domain is a region of approximately 100 residues corresponding to an extended hydrophobic heptad repeat (HR-A/B) essential for trimerization (Sorger and Nelson, 1989; Peteranderl and Nelson, 1992; Nieto-Sotelo et al., 1990; Perisic et al., 1989;
Figure 2 Functional domains of heat shock factor 1. The organizational structure of mouse HSF1 is indicated with the DNA binding domain designated at the amino-terminus. Adjacent to the DNA binding domain are the conserved hydrophobic heptad repeats involved in oligomerization of HSF1 and the negative regulatory domain. At the carboxyl-terminus is an additional hydrophobic heptad repeat adjacent to the transcription activation domain. The positions of sites of consitutive and inducible serine phosphorylation are indicated. Depicted below is a representation of the conversion of the inert monomeric form of HSF1 to the transcriptionally active trimer.
Clos et al., 1990; Rabindran et al., 1993). The HSFs of S. pombe, Drosophila, and larger
Inducible transcriptional regulation of the heat shock genes
43
eukaryotes contain an additional array of hydrophobic heptad repeats (HR-C) positioned near the extreme carboxyl terminus which may function in the negative regulation of DNA binding (Rabindran et al., 1993). A detailed analysis of the transcriptional activation domains of the HSF from S. cerevisiae and K. lactis identified separate regions required for full transcriptional activity (Nieto-Sotelo et al., 1990; Jakobsen and Pelham, 1991; Chen et al., 1993). In contrast, the transcription activation domain of Drosophila and mammalian HSF1 were localized to the extreme carboxyl terminus of the respective proteins (Shi et al., 1995; Newton et al., 1996; Wisniewski et al., 1996; Zuo et al., 1995). The studies on mammalian HSF1 have also identified negative regulatory elements in the vicinity of the DNA binding domain which influences the activity of the transactivation domain (Shi et al., 1995; Newton et al., 1996; Zuo et al., 1995). What is the role for a family of HSFs in the transcriptional regulation of heat shock genes? One possibility is that larger organisms may require multiple HSFs to provide specialized responses to the diverse developmental and environmental cues and insults they may be exposed to during life. Consistent with this speculation, vertebrate HSF1 and HSF2, though structurally related, are distinct by a variety of regulatory and functional criteria. For example, HSF1 corresponds to the general stress-responsive transcription factor whereas HSF2 is activated in response to developmental and differentiation cues (Theodorakis et al., 1989; Sistonen et al., 1992, 1994; Sarge et al., 1993, 1994). Activation of HSF1 is a complex multi-step process which involves oligomerization from an inert monomer to active trimer, acquisition of DNA-binding ability, stress-induced phosphorylation, and nuclear localization (Sarge et al., 1993; Baler et al., 1993; Jurivich et al., 1992; Sistonen et al., 1992; Sistonen et al., 1994; Cotto et al., 1996). Acquisition of trimer formation precedes inducible phosphorylation; furthermore the inducibly phosphorylated state of HSF1 is stress dependent with heat shock, heavy metals and arachidonate treatment resulting in the fully phosphorylated state and exposure to amino acid analogues or salicylate resulting in the activation of the DNA binding competent non-inducibly phosphorylated state (Sarge et al., 1993; Jurivich et al., 1994; Cotto et al., 1996). In contrast, HSF2 is not post-translationally modified by heat shock or other stresses (Sarge et al., 1993; Sistonen et al., 1992, 1994). HSF2 DNA binding is activated, for example, following exposure of human K562 erythroleukemia cells to hemin (Singh and Yu, 1984; Theodorakis et al., 1989; Sistonen et al., 1992). The acquisition of HSF2 DNA-binding activity is accompanied by conversion from an inert dimer to a DNA binding competent trimer; the DNA binding properties of HSF1 and HSF2 is similar but not identical (Kroeger et al., 1993; Sistonen et al., 1994). Constitutive HSF2 DNA binding activity has been observed in embryonal carcinoma cells (Morange et al., 1984; Mezger et al., 1989; Mezger et al., 1994; Murphy et al., 1994). Another observation which supports a role for HSF2 during development is the elevated expression of HSF2 expression and HSF2 DNA binding activity during spermatogenesis (Sarge et al., 1991; 1994) (see Morange, this volume).
Molecular chaperones and folding catalysts
44
2. THE HSF1 CYCLE: PROPOSED ROLE FOR HEAT SHOCK PROTEINS IN AUTOREGULATION OF THE HEAT SHOCK RESPONSE One of the distinguishing features of the heat shock response is the rapid kinetics of activation, the magnitude of inducibility, and reversibility of the response. Attenuation of the transcriptional response occurs in cells continuously exposed to intermediate heat shock temperatures (42°C) or immediately upon return to control temperatures, whereas following exposure to extreme heat shock temperatures (>43°C), the heat shock response persists and is maintained at high levels (Abravaya et al., 1991a). Comparison of HSF1 DNA binding properties using in vivo genomic footprinting on the human hsp70 promoter reveals that the HSE is unoccupied in cells at controlled temperatures and fully occupied during heat shock due to inducible binding to HSF1 (Abravaya et al., 1991a, b; Sistonen et al., 1994). During the attenuation phase, HSF1 releases rapidly from the HSE, and thereafter HSF1 DNA binding activity is no longer detected. The in vivo equilibrium dissociation rate for HSF1 has a half-life of approximately 10 minutes, whereas the dissociation rate of the HSF: HSE complex formed in vitro is greater than 100 minutes. The disparity of these results have led to the suggestion that other trans-regulatory components may be involved in the release of the activated form of HSF from DNA and dissociation of the trimer to the inert monomer. These speculations are further supported by the properties of recombinant Drosophila, chicken, mouse and human HSF’s in E. coli which are purified as stable trimers exhibiting constitutive DNA binding activity (Clos et al., 1990; Nakai and Morimoto, 1993; Sarge et al., 1993; Kroeger et al., 1993). It has long been speculated from studies in Drosophila and yeast that heat shock proteins may function in an autoregulatory loop to modulate the intensity and duration of the heat shock response (Craig and Gross, 1991). Indeed, this feature appears to be evolutionarily conserved as indicated by the role of heat shock proteins in the regulation of activity (Gamer et al., 1992; see Connolly et al., this volume). Although the sensor of cell stress in eukaryotes has not been established, most models for regulation of the heat shock response have proposed that the appearance of malfolded proteins, induced during heat shock and other forms of stress, sequesters Hsp70, thus allowing HSF1 to be activated (Morimoto et al., 1990; Abravaya et al., 1992; Baler et al., 1992; Morimoto, 1993). Additional support for the autoregulatory hypothesis comes from experimental evidence that links the activation of the heat shock response to increased levels of denatured and misfolded proteins (Anathan et al., 1986; Baler et al., 1992). Exposure to inhibitors of protein synthesis blocks the activation of the heat shock response by interfering with induction of HSF1; these results suggest that the proper synthesis and folding of nascent polypeptides represents a critical target for detection of malfolded proteins (Mosser et al., 1988; Amici et al., 1992; Baler et al., 1992). Taken together, these results support the hypothesis that the appearance and accumulation of malfolded polypeptides are directly involved in the pathway of stress detection and response (DiDomenico et al., 1982). Molecular chaperones such as members of the Hsp70 family are attractive candidates in the autoregulation of the heat shock response as they have a primary role in the association with non-native proteins to prevent their aggregation and to facilitate protein folding (Craig and Gross, 1991; Gething and Sambrook, 1992).
Inducible transcriptional regulation of the heat shock genes
45
Although there is a substantial evidence to support an involvement of molecular chaperones in the regulation of HSF1 activation, it is uncertain whether this occurs through direct or indirect effects. HSF1 trimers become associated with Hsp70 during attenuation; additionally HSF1 attenuation is more rapid in cell lines expressing high levels of Hsp70 (Abravaya et al., 1992; Mosser et al., 1993; Rabindran et al., 1994). If indeed it is the balance or ratio of free Hsp70 which is either directly or indirectly involved in maintaining the non-DNA binding form of HSF1 or in the attenuation of HSF1, it might be expected that overexpression of hsp70 would negatively affect the heat shock response. In support of this is evidence that the HSF1 transactivation domain is negatively regulated by Hsp70 (Shi and Morimoto, unpublished observations). Likewise, overexpression of HSF1 obtained by transient transfection results in a constitutively active or more readily activated factor independent of exogenous stress (Sarge et al., 1993). These results reveal that heat shock is not an obligatory step in HSF1 activation and are consistent with the hypothesis that there is a critical balance between HSF1 and its negative regulatory molecules to maintain HSF1 in either the non-DNA binding or DNA binding state. Many of the current observations are consistent with a model for the regulation of HSF1 DNA binding activity which is schematically represented in Figure 3. Under nonstressful conditions, HSF1 is constitutively expressed and maintained as a non-DNA binding monomer through intramolecular interactions perhaps influenced by constitutive serine phosphorylation (Kline and Morimoto, unpublished observations). Heat shock and other stresses result in the appearance of malfolded and aggregated proteins which creates a large pool of new protein substrates which compete with HSF1 for association with Hsp70. Thus, heat shock and other stresses initiate the events that remove the negative regulatory influence on HSF1 DNA binding activity. The conversion of HSF1 monomers to trimers requires a significant change in the conformation of HSF1 which may involve other activities to facilitate or stabilize the DNA binding trimer. HSF1 also undergoes a stress-dependent inducible serine phosphorylation and acquires transcriptional activity. The activated trimer state of HSF1 is associated with HSBP1 (heat shock factor binding protein), a small hydrophobic heptad repeat containing protein, during the period when HSF1 is active as a transcription factor (Satyal et al., in press). The subsequent events are not clearly ordered and include the interaction of HSBP1 with Hsp70 and the association of HSF1 with Hsp70 which has repressive effects on HSF1 transcriptional activity. Ultimately, these events during the regulation of HSF1 transcriptional activity are linked to the conversion of the active trimeric protein to the inert control state.
3. ACTIVATION OF HSF2 IS LINKED TO PROTEOLYSIS AND THE ACTIVITY OF THE UBIQUITIN-DEPENDENT PROTEASOME In contrast to the observations that HSF1 functions as the predominant stress responsive inducible transcription factor, the function of HSF2 has been an enigma. HSF2 is expressed in a wide range of cell types in all larger eukaryotes, yet HSF2 activity has only been detected as a constitutive DNA binding activity in embryonal carcinoma cells and tissues from murine embryos, during murine spermatogenesis, and in human K562
Molecular chaperones and folding catalysts
46
erythroleukemia cells upon treatment with hemin (see Morange, this volume). In the latter cell system, HSF2 is inert and activated upon hemin
Figure 3 HSF Cycle—A model of HSF1 regulation. In the unstressed cell, HSF1 is maintained primarily in the cytoplasm in a monomeric, nonDNA binding form. Upon heat shock or other forms of stress, HSF1 translocates and relocalizes within the nucleus and assembles into a trimer; HSBP1 associates with HSF1 trimers; HSF1 trimers bind to the heat shock element located in the promoter regions of heat shock genes and undergoes inducible serine phosphorylation. Transcriptional activation of the heat shock genes leads to increased levels of Hsp70 and to formation of an HSF1—Hsp70 complex. During attenuation of the heat shock transcriptional response, HSF1 dissociates from the DNA and is eventually converted to non-DNAbinding monomers.
treatment to a DNA binding and transcriptionally active trimer that binds in vivo to the promoter of the hsp90 and hsp70 genes (Sistonen et al., 1992). Unlike the robust, yet transient activation of HSF1 and resulting transcription of heat shock genes observed following exposure to heat shock, HSF2 activity and the expression of heat shock genes following hemin treatment can be detected over an extended period of time (Theodorakis et al., 1989; Sistonen et al., 1992). A more complete understanding of the role of HSF2 has come from observations that the activation of HSF2 DNA binding and transcriptional activity appears to be due to the inhibitory effects of hemin on ubiquitin-dependent
Inducible transcriptional regulation of the heat shock genes
47
proteasome activity and the demonstration that specific proteasome inhibitors such as the peptide aldehyde, MG132, and the Streptomyces metabolite, lactacystin, also induce HSF2 activity (Mathew, Mathur, and Morimoto, unpublished observations). Exposure of yeast and mammalian cells to MG132, lactacystin or hemin (the latter in a cell specific manner) leads to the elevated expression of heat shock genes hsp90, hsp70 and hdj-1 via transcriptional induction dependent upon the HSE (Lee and Goldberg, personal communication; Mathew, Mathur and Morimoto, unpublished observations). How is HSF2 activity regulated during downregulation of proteaseome activity? The change in HSF2 DNA binding activity is neither accompanied by relocalization of HSF2 nor by changes in the levels of HSF2. Furthermore HSF2 is not polyubiquitinated, yet maintenance of the trimeric state of HSF2 requires ongoing protein synthesis. These data are best explained by the suggestion that HSF2 activation is dependent upon another labile regulatory protein, distinct from HSF2, which itself is degraded by the proteasome. In contrast, HSF1 is a stable protein whose activity is independent of protein synthesis and the activity of the ubiquitindependent proteasome. Thus, HSF2 rather than HSF1 senses the dynamics of protein degradation. The inhibition of protein degradation would be predicted to result in the accumulation of polyubiquitinated non-native proteins that maybe prone to aggregation. Therefore, the elevated expression of heat shock proteins under such circumstances affords a mechanism to prevent the formation of insoluble protein aggregates. HSF2 activation, as a response to down-regulation of proteasome activity, may enable the cell to act in a preventative manner by increasing expression of molecular chaperones whose function is to associate with these non-native proteins. The synthesis of Hsp70 rapidly attenuates upon removal of the proteasome inhibitor, consistent with a role for the accumulated Hsp70 (and perhaps other chaperones) to associate with the increasing levels of polyubiquitinated proteins which build up during the arrest of proteasome activity. Based on these results, a model (Figure 4) is proposed for HSF2 activation in which inhibition of proteasome activity leads to the elevated synthesis and accumulation of heat shock proteins and molecular chaperones which serves to associate with the accumulated proteins, presumably to maintain these non-native proteins in a soluble intermediate folded state. This feature of Hsp70 activity, the maintenance of intermediate non-native proteins in a soluble state, is supported by recent observations of in vitro protein folding experiments with molecular chaperones (Freeman et al., 1995; Freeman and Morimoto, 1996; Freeman, Toft and Morimoto, 1996). There is additional evidence to link the heat shock response and proteolysis. In E. coli, the activities of the ATP-dependent proteases Lon and Clp, which are themselves heat shock proteins (Goff and Goldberg, 1985), are influenced by the heat shock proteins DnaK, DnaJ, GrpE, GroES and GroEL (Sherman and Goldberg, 1992; Sherman and Goldberg, 1996). Likewise, in eukaryotes, ubiquitin and other components of the ubiquitin modification system are induced by heat shock and essential for survival following exposure to stress (Finley, Ozkaynak and Varshavsky, 1987; Seufert and Jentsch, 1990; Jentsch, 1992). The molecular chaperone Ydj-1 (DnaJ homologue) has been shown to associate with an artificially ubiquitinated short-lived abnormal protein and is essential for ubiquitin-dependent degradation (Lee, Sherman and Goldberg, 1996). The link between HSF2 activity and proteolysis offers an explanation for the constitutive
Molecular chaperones and folding catalysts
48
activation of heat shock genes in the ts85
Figure 4 Model for regulation of chaperone expression by proteasome activity. (1) Malfolded, mutant and short-lived proteins are polyubiquitinated (indicated by addition of Ub). (2) Ubiquitinated proteins are targeted to the proteasome for degradation which can be inhibited by the addition of MG132, lactacystin, or hemin. (3) Inhibition of proteasome activity results in an accumulation of polyubiquitinated proteins. (4) HSF2 is activated from the inert dimer to DNA-binding trimer in response to the accumulation of non-native polyubiquitinated proteins. (5) HSF2 induces expression of molecular chaperones such as Hsp70 and Hdj-1. (6) The increased transcription of molecular chaperones results in increased levels of Hsp70 and Hdj-1 proteins. (7) The molecular chaperones associate with the polyubiquitinated proteins and are suggested to maintain the polyubiquitinated proteins in an intermediate folded state primed for degradation upon resumption of proteasome activity.
mouse cell line which has a thermolabile ubiquitin-activating enzyme (Finley, Ciechanover and Varshavsky, 1984) and a proposed model, based on these studies, for activation of the heat shock response in which inhibition of the ubiquination system would promote accumulation of a non-ubiquinated active transcription factor leading to
Inducible transcriptional regulation of the heat shock genes
49
transcription of the heat shock genes (Munro and Pelham, 1985; Finley, Ciechanover and Varshavsky, 1984).
4. HSF3 AND HSF4 Chicken HSF3 was cloned during a molecular screen of a chicken cDNA library for genes homologous to murine HSF1 (Nakai and Morimoto, 1993). The three members of the chicken HSF family (HSF1–3) are closely related (40% sequence identity) and exhibit similar structural motifs for the DNA binding domain and the adjacent hydrophobic heptad repeats. HSF3 is ubiquitously expressed and the DNA binding properties are negatively regulated presumably through intramolecular interactions involving the carboxyl terminal hydrophobic heptad C, as has also been suggested for HSF1. The inert state of HSF3 is a dimer which undergoes oligomerization to an active trimer. However, in contrast with HSF1, acquisition of HSF3 DNA binding activity in response to heat shock, appears to be a cell specific event and has been detected predominantly in HD6 cells and the chicken B lymphoblast cell line, DT40, where it has the properties of a positive activator (Nakai et al., 1995). The co-activation of HSF1 and HSF3 in HD6 cells exposed to heat shock suggests the possibility of HSF regulatory redundancy, perhaps to ensure that the transcriptional activation of heat shock genes is both rapidly induced and sustained at a high level in response to extreme temperatures. In this context, it is worth noting that the optimal heat shock temperature for avian cells is 45°C as compared to 42°C for mammals. The possibility of a complementary role for these apparently redundant HSFs is supported by the rapid kinetics of HSF1 induction followed by HSF3. This observation, together with the demonstration that the in vitro activation properties of HSF3 and HSF1 differ (Nakai and Morimoto, 1993) indicates that these these two factors are similar, yet have distinct regulatory features as measured by acquisition of DNA binding activity. To establish the biological role for HSF3, however, it will be necessary to either inactivate or substantially reduce HSF1 expression which would allow us to demonstrate whether HSF3 is itself sufficient to activate the expression of the endogenous heat shock genes. HSF4 has been recently described as a new member of the human HSF gene family with the unusual feature of lacking the properties of a transcriptional activator despite having in common many of the structural and biochemical features of other previously cloned HSF genes (Nakai et al, 1997). Overexpression of HSF4 leads to the repression of target genes that are regulated through heat shock element promoter sequences. This unusual feature may be related to the fact that HSF4 lacks the carboxyl terminal hydrophobic repeat which is shared among all previously characterized members of the HSF gene family which are negatively regulated and exhibits a highly restricted expression pattern in human tissues. Therefore, it remains to be established whether HSF4 is a novel negative regulator of heat shock gene expression which exerts its activity in a highly tissue restricted manner as a represser whose activity is to balance HSE targeted transcriptional activity. Among the various HSF’s identified in diverse species, only HSF4 lacks the activity of a positive transactivator. Overexpression of HSF4 results in the repression of basal transcription of cellular heat shock genes regulated by the HSE
Molecular chaperones and folding catalysts
50
(ie., Hsp90, Hsp70, and Hsp27) presumably through direct binding of HSF4 to the HSEs and subsequent inhibitory effects on either the initiation complex or during transcriptional elongation. It is tempting to speculate that the function of HSF4 as a negative regulator of basal transcription of heat shock genes provides a novel means to modulate heat shock gene transcription.
5. TRANSCRIPTIONAL REGULATION OF ENDOPLASMIC RETICULUM STRESS GENES (GRP78): THE UNFOLDED PROTEIN RESPONSE The endoplasmic reticulum (ER) represents yet another subcellular compartment in which nascent polypeptides are imported, folded, and assembled. The accumulation of unfolded proteins (unfolded protein response) in the ER, following exposure to conditions that prevent glycosylation, leads to the elevated transcription of lumen localized chaperones (Grp94, Grp78/Kar2, protein disulfide isomerase) (Lee, 1987; Kozutsumi et al., 1988; Watowich and Morimoto, 1988). Expression of these target genes is regulated via a common unfolding protein response element (UPRE) which is necessary and sufficient to confer the response to the accumulation of malfolded ER proteins (Mori et al., 1992; Kohno et al., 1993). A second component of the UPR is a transmembrane kinase (Ire1p) which is located in the ER (Cox, Shamu and Walter, 1993; Mori et al., 1993). Ire1p is a 1115 amino acid protein of which the amino terminus lies in the lumen of the ER where it presumably recognizes unfolded proteins and the kinase domain lies either in the cytoplasm or nucleus and is likely responsible for transmitting the signal to the downstream components of the UPR. In response to unfolded proteins, Ire1p oligomerizes and is transphosphorylated by other Ire1p molecules (Shamu and Walter, 1996; Welihinda and Kaufman, 1996). The carboxyl terminal tail of Ire1p has been suggested to interact with unfolded proteins. Using the UPRE as the target for a genetic screen, the transcription factor, Hac1p, a member of the bZIP family of transcription factors, was identified which binds directly to the UPRE (Cox and Walter, 1996; Nojima et al., 1994). Disruption of the gene encoding Hac1p results in the loss of the UPR, thus revealing that Hac1p is in the cascade of events in the UPR (Cox and Walter, 1996). HAC1 RNA, however, is present under noninducing conditions and is found on polyribosomes whereas Hac1p is only detected under UPR conditions. Although Ire1p is not an essential gene, splicing of HAC1 mRNA is dependent upon Ire1p to generate a unique exon-exon junction resulting in a change in HAC1 reading frame leading to the synthesis of a stable Hac1p. The differences in the two forms of Hac1p correspond to an extended PEST region which confers degradation by the ubiquitin-dependent proteasome. Thus, Ire1p activated splicing would induce a longer-lasting form of Hac1p. Consistent with these observations, mutations in the ubiquitin conjugating enzymes (UBC) activates the UPR pathway in the absence of ER stress (Seufert and Jentsch, 1990; Seufert et al., 1990). Is Hac1p the only transcription factor that responds to ER stress? Exposure of adipogenic cell lines to low glucose, tunicamycin, and other conditions that induce grp78 expression results in the elevated expression of CHOP (C/EBP homologous protein) (Chen et al., 1992; Price and Calderwood, 1992; Carlson et al., 1993). CHOP was
Inducible transcriptional regulation of the heat shock genes
51
initially identified as a DNA damage inducible gene based on its elevated expression following exposure to UV or the alkylating agent methyl methanesulfonate and separately shown to heterodimerize with members of the C/EBP family and function as a dominant negative regulator of transcription (Fornace et al., 1989; Ron and Habner, 1992; Ubeda et al., 1996). CHOP expression is rapidly induced following ER stress relative to the induction of ER chaperones and may be a component of the ER-stress sensing apparatus in mammalian cells (Wang et al., 1996). How CHOP detects the ERstress, however has not been established.
6. CONCLUSIONS Among the fundamental, unanswered, questions that underlie the basis of the heat shock response is the nature of the intracellular signal(s) that indicates the cell is stressed. The expression of heat shock and stress responsive genes has been closely linked with the appearance, within the cell, of malfolded or mutant proteins and the differential activation of heat shock transcription factors. Heat shock factor 1 (HSF1) functions as the ubiquitous stress responsive activator which is maintained in a latent state and responds principally to the appearance of nascent stress-induced malfolded proteins, whereas HSF2 activity is linked to proteolysis and the activity of the ubiquitin-dependent proteasome. Activation of either HSF1 or HSF2 leads to transcription of heat shock element (HSE) regulated genes encoding cytosolic heat shock proteins and molecular chaperones. In contrast, the appearance of unfolded proteins in the lumen of the endoplasmic reticulum (ER) is detected by Ire1p a kinase that spans the ER membrane and senses the increase in unfolded proteins and Haclp a transcription factor which in its activated state binds to the unfolded protein response element (UPRE) and activates the transcription of genes encoding ER-localized chaperones.
7. ACKNOWLEDGEMENTS The studies from our laboratory were supported by grants from the National Institutes of General Medicine. The author extends an appreciation to members of the laboratory who contributed to these studies and provided continuous stimulating discussions.
8. REFERENCES Abravaya, K., Phillips, B. and Morimoto, R.I. (1991a). Attenuation of the heat shock response in HeLa cells is mediated by the release of bound heat shock transcription factor and is modulated by changes in growth and in heat shock temperatures. Genes & Dev. , 5 , 2117–2127. Abravaya, K., Phillips, B. and Morimoto, R.I. (1991b). Heat shock-induced interactions of heat shock transcription factor and the human hsp70 promoter examined by in vivo footprinting. Mol. Cell. Biol , 11 , 586–592. Abravaya, K., Myers, M.P., Murphy, S.P. and Morimoto, R.I. (1992). The human heat
Molecular chaperones and folding catalysts
52
shock protein hsp70 interacts with HSF, the transcription factor that regulates heat shock gene transcription. Genes &Dev. , 6 , 1153–1164. Amici, C., Sistonen, L., Santoro, M.G. and Morimoto, R.I. (1992). Antiproliferative prostaglandin’s activate heat shock transcription factor. Proc. Natl Acad. Sci. USA , 89 , 6227–6231. Amin, J., Ananthan, J. and Voellmy, R. (1988). Key features of heat shock regulatory elements. Mol. Cell. Biol. , 8 , 3761–3769. Ananthan, J., Goldberg, A.L. and Voellmy, R. (1986). Abnormal proteins serve as eukaryotic stress signals and trigger the activation of heat shock genes. Science , 232 , 522–524. Ashburner, M. (1970). Pattern of puffing activity in the salivary gland chromosomes of Drosophila. V. Response to environmental treatments. Chromosoma , 31 , 356–376. Baler, R.W., Welch, J. and Voellmy, R. (1992). Heat shock gene regulation by nascent polypeptides and denatured proteins: hsp70 as a potential autoregulatory factor. J. Cell Biol. , 117 , 1151–1159. Baler, R., Dahl, G. and Voellmy, R. (1993). Activation of human heat shock genes is accompanied by oligomerization, modification and rapid translocation of heat shock transcription factor HSF1. Mol. Cell. Biol. , 13 , 2486–2496. Carlson, S.G., Fawcett, T.W., Bartlett, J.D., Bernier, M. and Holbrook, N.J. (1993). Regulation of the C/EBP-related gene gadd153 by glucose deprivation. Mol. Cell. Biol. , 13 , 4736–4744. Chen, Q., Yu K., Holbrook N.J. and Stevens, J.L. (1992). Activation of the growth arrest and DNA damage-inducible gene gadd153 by nephrotoxic cysteine conjugates and dithiothreitol. J. Biol. Chem. , 267 , 8207–8212. Chen, Y., Barlev, N.A., Westergaard, O. and Jakobsen., B.K. (1993). Identification of the G-terminal activator domain in yeast heat shock factor: independent control of transient and sustained transcriptional activity. EMBO J. , 12 , 5007–5018. Clos, J., Westwood, J.T., Becker, P.B., Wilson, S., Lambert, K. and Wu, C. (1990). Molecular cloning and expression of a hexameric Drosophila heat shock factor subject to negative regulation. Cell , 63 , 1085–1097. Cotto, J.J., Kline, M. and Morimoto, R.I. (1996). Activation of heat shock factor 1 DNA binding precedes stress-induced serine phosphorylation. J. Biol Chem. , 271 , 3355– 3358. Cox, J.S., Shamu, C.E. and Walter, P. (1993). Transcriptional induction of genes encoding endoplasmic reticulum resident proteins requires a transmembrane protein kinase. Cell , 73 , 1197–1206. Cox, J.S. and Walter, P. (1996). A novel mechanism for regulating activity of a transcription factor that controls the unfolded protein response. Cell , 87 , 391–404. Craig. E.A. and Gross, CA. (1991). Is hsp70 a cellular thermometer? Trends Biochem. Sci . 16 , 135–140. Damberger, F.F., Pelton, J.G., Harrison, C.J., Nelson, H.C.M. and Wemmer, D.E. (1994). Solution structure of the DNA-binding domain of the heat shock transcription factor determined by multidimensional heteronuclear magnetic resonance spectroscopy. Protein Sci. , 3 , 1806–1821. DiDomenico, B.J., Bugaisky, G.E. and Lindquist, S. (1982). The heat shock response is regulated at both the transcriptional and posttranscriptional levels. Cell , 31 , 593–603. Finley, D., Ciechanover, A. and Varshavsky, A. (1984). Cell , 37 , 43–55. Finley, D., Ozkaynak, E. and Varshavsky, A. (1987). The yeast polyubiquitin gene is essential for resistance to high temperatures, starvation and other stresses. Cell , 48 ,
Inducible transcriptional regulation of the heat shock genes
53
1035–1046. Fornace, A.J., Neibert, D.W., Hollander, M.C., Luethy, J.D., Papathanasiou, M., Fragoli, J. and Holbrook, N.J. (1989). Mammalian genes coordinately regulated by growth arrest signals and DNA-damaging agents. Mol Cell. Biol , 9 , 4196–4203. Freeman, B.M.Meyers, R.Schumacher and R.I.Morimoto. (1995). Identification of a regulatory motif in Hsp70 that affects ATPase activity, substrate binding and interaction with HDJ-1. EMBO Journal , 14 , 2281–2292. Freeman, B.C. and Morimoto, R.I. (1996). The human cystolic molecular chaperones hsp90, hsp70 (hsc70). and hdj-1 have distinct roles in recognition of a non-native protein and protein refolding. EMBO J. , 15 , 2969–2979. Freeman, B.C., Toft, D. and Morimoto, R.I. (1996). Molecular chaperone machines: Chaperone activities of the cyclophilin Cyp-40 and the steroid aporeceptor-associated protein p23. Science , 274 , 718–720. Gamer, J., Bujard, H. and Bukau, B., (1992). Physical interaction between heat shock proteins DnaK, DnaJ and GrpE and the bacterial heat shock transcription factor . Cell , 69 , 833–842. Gething, M.-J. and Sambrook, J. (1992). Protein folding in the cell. Nature , 355 , 33–45. Goff, S. and Goldberg, A. (1985). Production of abnormal proteins in E. coli stimulates transcription of Ion and other heat shock genes. Cell , 41 , 587–595. Greene, J.M. and Kingston, R.E. (1990). TATA-dependent and TATA-independent function of the basal and heat shock elements of a human hsp70 promoter. Mol. Cell Biol. , 10(4), 1319–1328. Harrison, C.J., Bohm, A.A. and Nelson, H.C.M. (1994). Crystal structure of the DNA binding domain of the heat shock transcription factor. Science , 263 , 224–227. Jakobsen, B.K. and Pelham, H.R. (1991). A conserved heptapeptide restrains the activity of the yeast heat shock transcription factor. EMBO J. , 10 , 369–375. Jentsch, S. (1992). The ubiquitin-conjugation system. Ann. Rev. Genet. , 26 , 179–207. Jurivich, D.A., Sistonen, L., Kroes, R.A. and Morimoto, R.I. (1992). Effect of sodium salicylate on the human heat shock response. Science , 255 , 1243–45. Jurivich, D.A., Sistonen, L., Sarge, K.D. and Morimoto, R.I. (1994). Arachidonate is a potent modulator of human heat shock gene transcription. Proc. Natl. Acad. Sci. , 91 , 2280–2234. Kingston, R.E., Schuetz, T.J. and Larin, Z. (1987). Heat-inducible human factor that binds to a human hsp70 promoter. Mol. Cell. Biol. , 7 , 1530–1534. Kline, M.P. and Morimoto, R.I. (1997). Repression of the heat shock factor 1 transcriptional activation domain is modulated by constitutive phosphorylation. Mol. Cell Biol. , 17 , 2107–2115. Knauf, U., Newton, E.M., Kyriakis, J., Kingston, R.E. (1996). Repression of human heat shock factor 1 activity at control temperature by phosphorylation. Genes and Development , 10 , 2782–2793. Kohno, K., Normington, K., Sambrook, J., Gething, M.J. and Mori, K. (1993). The promoter region of the yeast KAR2 (BiP). gene contains a regulatory domain that responds to the presence of unfolded proteins in the endoplasmic reticulum. Mol. Cell. Biol. , 13 , 877–890. Kozutsumi, Y., Segal, M., Normington, K., Gething, M.J. and Sambrook, J. (1988). The presence of malfolded proteins in the endoplasmic reticulum signals the induction of glucose-regulated proteins. Nature , 332 , 462–464. Kroeger, P.E., Sarge, K.D and Morimoto, R.I. (1993). Mouse heat shock transcription factors 1 and 2 prefer a trimeric binding site but interact differently with the HSP70
Molecular chaperones and folding catalysts
54
heat shock element. Mol. Cell. Biol. , 13 , 3370–3383. Kroeger, P.E. and Morimoto, R.I. (1994). Selection of new HSF1 and HSF2 DNA binding sites reveals differences in trimer cooperativity. Mol. Cell. Biol. , 14 , 7592– 7603. Larson, J.S., Schuetz, T.J. and Kingston, R.E. (1988). Activation in vitro of sequencespecific DNA binding by a human regulatory factor. Nature , 335 , 372–375. Lee, A.S. (1987). Coordinated regulation of a set of genes by glucose and calcium ionophores in mammalian cells . Trends Biochem. Sci. , 12 , 20–23. Lee, D.H., Sherman, M. and Goldberg, A. (1996). Involvement of the molecular chaperone Ydj1 in the ubiquitin-dependent degradation of short-lived and abnormal proteins in saccharomyces cerevisiae. Mol. & Cell. Biol. , 16 , 4773–4781. Lindquist, S. and Craig, E.A. (1988). The heat shock proteins. Annu. Rev. Genet. , 22 , 631–677. Lis, J. and Wu, C. (1993). Protein traffic on the heat shock promoter, parking, stalling and trucking along. Cell , 74 , 1–4. Mezger, V., Bensaude, O. and Morange, M. (1989). Unusual levels of heat shock element-binding activity in embryonal carcinoma cells. Mol. Cell. Biol. , 9 , 3888– 3896. Mezger, V., Rallu, M., Morimoto, R.I., Morange, M. and Renard, J-P. (1994). Rapid communication heat shock factor 2-like activity in mouse blastocysts. Dev. Biol. , 166 , 819–822. Morange, M., Diu, A., Bensaude, O. and Babinet, C. (1984). Altered expression of heat shock proteins in embryonal carcinoma and mouse early embryonic cells. Mol. Cell. Biol. , 4 , 730–735. Mori, K., Sant, A., Kohno, K., Normington, K,, Gething, M.J. and Sambrook, J.F. (1992). A 22 bp cis-acting element is necessary and sufficient for the induction of the yeast KAR2 (BiP) gene by unfolded proteins, EMBO J. , 11 , 2583–2593. Mori, K., Ma, W., Gething, M.-J. and Sambrook, J. (1993). A transmembrane protein with a cdc2+/ CDC28-related kinase activity is required for signaling from the ER to the nucleus. Cell , 74 , 743–756. Morimoto, R.I. (1993). Cells in stress: transcriptional activation of heat shock genes. Science , 259 , 1409–1410. Morimoto, R.I., Tissieres, A. and Georgopoulos, C. (1990). The stress response, function of the proteins and perspectives. In: Stress proteins in biology and medicine . (Ed. Morimoto, R.I., Tissieres, A. and Georgopoulos, C.). Cold Spring Harbor, New York. Cold Spring Harbor Laboratory Press, pp. 1–36. Morimoto, R.I., Tissieres, A. and Georgopoulos, C. (1994). The biology of heat shock proteins and molecular chaperones. In: Stress proteins in biology and medicine . (Ed. Morimoto, R.I., Tissieres, A. and Georgopoulos, C.). Cold Spring Harbor, New York. Cold Spring Harbor Laboratory Press. Mosser, D.D., Theodorakis, N.G. and Morimoto, R.I. (1988). Coordinate changes in heat shock element-binding activity and hsp70 gene transcription rates in human cells. Mol. Cell. Biol. , 8 , 4736–4744. Mosser, D.D., Kotzbauer, P.T., Sarge, K.D. and Morimoto, R.I. (1990). In vitro activation of heat shock transcription factor DNA-binding by calcium and biochemical conditions that affect protein conformation. Proc. Natl. Acad. Sci. USA . 87 , 3748– 3752. Mosser, D.D., Duchaine, J. and Massie, B. (1993). The DNA-binding activity of the human heat shock transcription factor is regulated in vivo by hsp70. Mol. Cell. Biol. ,
Inducible transcriptional regulation of the heat shock genes
55
13 , 5427–5438. Munro, S. and Pelham, H.R.B. (1985). What turns on heat shock genes? Nature , 317 , 477–478. Murphy, S.P., Gorzowski, J.J., Sarge, K.D. and Phillips, B. (1994). Characterization of constitutive HSF2 DNA-binding activity in mouse embryonal carcinoma cells. Mol Cell. Biol. , 14 , 5309–5317. Nakai, A., Kawazoe, Y., Tanabe, M., Nagata, K. and Morimoto, R.I. (1995). The DNAbinding properties of two heat shock factors, HSF1 and HSF3, are induced in an avian erythroblast cell line HD6. Mol. Cell. Biol. , 15 , 5268–5278. Nakai, A. and Morimoto, R.I. (1993). Characterization of a novel chicken heat shock transcription factor, HSF3, suggests a new regulatory pathway. Mol. Cell. Biol. , 13 , 1983–1997. Nakai, A., M.Tanabe, Y.Kawazoe, J.Inazawa, R.I.Morimoto and K.Nagata. HSF4, A new member of the human heat shock factor gene family which lacks properties of a transcriptional activator. Mol. Cell. Biol. , (in press, 1997). Newton, E.M., Knauf, U., Green, M. and Kingston, R.E. (1996). The regulatory domain of human heat shock factor 1 is sufficient to sense heat stress. Mol. Cell Biol. , 16 , 839–846. Nieto-Sotelo, J., Wiederrecht, G., Okuda, A. and Parker, C.S. (1990). The yeast heat shock transcription factor contains a transcriptional activation domain whose activity is repressed under nonshock conditions. Cell , 62 , 807–817. Nojima, H., Sun-Hee, L., Araki, H., Sakai, A., Nakashima, N., Kanaoka, Y. and Ono, Y. (1994). Hac1: a novel yeast bZIP protein binding to the CRE motif is a multicopy suppressor for cdc 10 mutant of Schizosaccharomyces pombe. Nucl. Adds Res. , 22 , 5279–5288. Nover, L. (1994). The heat stress reponse as part of the plant stress network. In NATOASI Series on Biochemical and Cellular Mechanisms of Stress Tolerance in Plants, Cherry J. (ed.), Springer Verlag, Berlin, pp. 3–45. Perisic, O., Xiao, H. and Lis, J.T. (1989). Stable binding of Drosophila heat shock factor to head-to-head and tail-to-tail repeats of a conserved 5 bp recognition unit. Cell , 59 , 797–806. Peteranderl, R. and Nelson, H.C.M. (1992). Trimerization of the heat shock transcription factor by a triple-stranded a-helical coiled-coil. Biochemistry , 31 , 12272–12276. Price, B. and Calderwood, S., (1992). Gadd 45 and Gadd 153 messenger RNA levels are increased during hypoxia and after exposure of cells to agents which elevate the levels of glucose-regulated proteins. Cancer Res. , 52 , 3814–3817. Rabindran, S.K., Giorgi, G., Clos, J. and Wu, C. (1991). Molecular cloning and expression of a human heat shock factor, HSF1. Proc. Natl. Acad. Sci. USA . 88 , 6906–6910. Rabindran, S.K., Wisniewski, J., Li, L., Li, G.C. and Wu, C. (1994). Interaction between heat shock factor and HSP 70 is insufficient to suppress induction of DNA-binding activity in vivo. Mol. Cell. Biol. , 14 , 6552–6560. Rabindran, S.K., Haroun, R.I., Clos, J., Wisniewski, J. and Wu, C. (1993). Regulation of heat shock factor trimer formation: role of a conserved leucine zipper. Science , 259 , 230–234. Ritossa, F.M. (1962). A new puffing pattern induced by a temperature shock and DNP. Drosophilia: Experientia, 18 , 571–573. Ron, D. and Habener, J.F. (1992). CHOP, a novel developmentally regulated nuclear protein that dimerizes with transcription factors C/EBP and LAP and functions as a
Molecular chaperones and folding catalysts
56
dominant negative inhibitor of gene transcription. Genes Dev. , 6 , 439–453. Sarge, K.D., Zimarino, V., Holm, K, Wu, C. and Morimoto, R.I. (1991). Cloning and characterization of two mouse heat shock factors with distinct inducible and constitutive DNA-binding ability. Genes & Dev. , 5 , 1902–1911. Sarge, K.D., Murphy, S.P. and Morimoto, R.I. (1993). Activation of heat shock gene transcription by HSF1 involves oligomerization, acquisition of DNA binding activity and nuclear localization and can occur in the absence of stress. Mol. Cell. Biol , 13 , 1392–1407. Sarge, K.D., Park-Sarge, O.-K., Kirby, J.D., Mayo, K.E. and Morimoto, R.I. (1994). Regulated expression of heat shock factor 2 in mouse testis: potential role as a regulator of hsp gene expression during spermatogenesis. Biol. Reprod. , 50 , 334–343. Satyal, S., Chen, D., Fox, S. and Morimoto, R.I., Negative regulation of the heat shock transcriptional response by HSBP1. Genes and Development , in press. Scharf, K.-D., Rose, S., Zott, W., Schoff, F. and Nover, L. (1990). Three tomato genes code for heat stress transcription factors with a remarkable degree of homology to the DNA-binding domain of the yeast HSF. EMBO J. , 9 , 4495–4501. Schuetz, T.J., Gallo, , G.J., Sheldon, L., Tempst, P. and Kingston, R.E. (1991). Isolation of a cDNA for HSF2: evidence for two heat shock factor genes in humans. Proc. Natl. Acad. Sci. USA , 88 , 6910–6915. Seufert, W. and Jentsch, S. (1990). Ubiquitin-conjugating enzymes UBC4 and UBC5 mediate selective degradation of short-lived and abnormal proteins . EMBO , 9 , 543– 550. Seufert, W., McGrath, J.P. and Jentsch, S. (1990). UBC1 encodes a novel member of an essential subfamily of yeast obiquitin-conjugating enzymes involved in protein degradation. EMBO J. , 9 , 4535–4541. Shamu, C.E. and Walter, P. (1996). Oligomerization and phosphorylation of the Ire1p kinase during intracellular signaling from the endoplasmic reticulum to the nucleus. EMBO J. , 15 , 3028–3039. Sherman, M. and Goldberg, A. (1992). Involvement of the chaperonin dnaK in the rapid degradation of a mutant protein in E. coli. EMBO J. , 11 , 71–77. Sherman, M.Y.S. and Goldberg, A.L., Involvement of molecular chaperones in intracellular protein breakdown. In Stress-Inducible Cellular Responses . U.Feige, R.I.Morimoto, I.Yahara and B.Polla, Eds.: 57–78. Birkhauser Verlag, Boston. Shi, Y., Kroeger, P.E. and Morimoto, R.I. (1995). The carboxyl-terminal transcription domain of heat shock factor 1 is negatively regulated and stress responsive. Mol. Cell. Bio. , 15 , 4309–4318. Singh, M.K. and Yu, J. (1984). Accumulation of a heat shock-like protein during differentiation of human erythroid cell line K562. Nature , 309 , 631–633. Sistonen, L., Sarge, K.D., Phillips, B., Abravaya, K. and Morimoto, R.I. (1992). Activation of heat shock factor 2 during hemin-induced differentiation of human erythroleukemia cells. Mol. Cell. Biol. , 12 , 4104–4111. Sistonen, L., Sarge, K.D. and Morimoto, R.I. (1994). Human heat shock factors 1 and 2 are differentially activated and can synergistically induce HSP70 gene transcription. Mol. Cell. Biol. , In press. Sorger, P.K. and Nelson, H.C.M. (1989). Trimerization of a yeast transcriptional activator via a coiled-coil motif. Cell , 59 , 807–813. Sorger, P.K. and Nelson, H.C.M. (1988). Yeast heat shock factor is an essential DNAbinding protein that exhibits temperature-dependent phosphorylation. Cell , 54 , 855– 864.
Inducible transcriptional regulation of the heat shock genes
57
Sorger, P.K, and Pelham, H.R.B. (1987). Purification and characterization of a heat shock element binding protein from yeast. EMBO J. , 6 , 3035–3041. Sorger, P.I., Lewis, M.J. and Pelham, H.R.B. (1987). Heat shock factor is regulated differently in yeast and HeLa cells . Nature , 329 , 81–84. Theodorakis, N.G., Zand, D.J., Kotzbauer, P.T., Williams, G.T. and Morimoto, R.I. (1989). Hemin-induced transcriptional activation of the hsp70 gene during erythroid maturation in K562 cells is due to a heat shock factor-mediated stress response. Mol. Cell. Biol. , 9 , 3166–3173. Ubeda, M., Wang, X.-Z., Zinszner, H., Wu, I., Habener, J. and Ron, D. (1996). Stressinduced binding of transcription factor CHOP to a novel DNA-control element. Mol. Cell. Biol. , 16 , 1479–1489. Vuister, G.W., Kim, S.J., Wu, C. and Bax, A. (1994a). NMR evidence for similarities between the DNA binding regions of Drosophila melanogaster heat shock factor and the helix-turn-helix and HNF-3/forkhead families of transcription factors. Biochemistry , 33 , 10–16. Vuister, G.W., Kim, S.J., Orosz, A., Marquardt, J., Wu, C. and Bax, A. (1994b). Solution structure of the DNA binding domain of Drosohpila heat shock transcription factor. Nature Struct. Biol. , 1 , 605–614. Wang, X.-Z. and Ron, D. (1996). Stress0induced phosphorylation and activation of the transcription factor CHOP (GADD153). by p38 MAP -kinase. Science , 272 , 1347– 1349. Watowich, S.S. and Morimoto, R.I (1988). Complex regulation of heat shock and glucose responsive genes in human cells. Mol Cell. Biol. , 8 , 393–405. Welihinda, A.A. and Kaufman, R.J. (1996). The unfolded protein response pathway in saccharomyces cerevisiae. J. Biol. Chem. , 271 , 18181–18187. Westwood, J.T., Clos, J. and Wu, C. (1991). Stress-induced oligomerization and chromosomal relocalization of heat-shock factor. Nature , 353 , 822–827. Westwood, J.T. and Wu, C. (1993). Activation of Drosophila heat shock factor: conformational change associated with a monomer-to-trimer transition. Mol. Cell. Biol. , 13 , 3481–3486. Wiederrecht, G., Seto, D. and Parker, C.S. (1988). Isolation of the gene encoding the S. cerevisiae heat shock transcription factor. Cell , 54 , 841–853. Williams, G.T. and Morimoto, R.I. (1990). Maximal stress-induced transcription from the human hsp70 promoter requires interactions with the basal promoter elements independent of rotational alignment . Mol. Cell. Biol. , 10 , 3125–3136. Wisniewski, J., Orosz, A., Allada, R. and Wu, C. (1996). The C-terminal region of drosophila heat shock factor (HSF). contains a constitutively functional transactivation domain. Nucleic Acids Res. , 24 , 367–374. Wu, C. (1980). The 5’ ends of Drosophila heat shock genes in chromatin are hypersensitive to DNase1. Nature , 286 , 854–860. Wu, C., Wilson, S., Walker, B., David, I., Paisley, T., Zimarino, V. and Ueda, H. (1987). Purification and properties of Drosophila heat shock activator protein. Science , 238 , 1247–1253. Wu, C. (1984). Two protein-binding sites in chromatin implicated in the activation of heat-shock genes. Nature , 309 , 229–234. Wu, C. (1995). Heat shock transcription factors: structure and regulation. Annu. Rev. Cell Dev. Biol. , 11 , 441–469. Wu. C. (1995). Heat stress transcription factors. Annu. Rev. Cell Biol. , 11 , 441–469. Xiao, H. and Lis, J.T. (1988). Germline transformation used to define key features of the
Molecular chaperones and folding catalysts
58
heat shock response element. Science , 239 , 1139–1142. Xiao, H., Perisic, O. and Lis, T.J. (1991). Cooperative binding of Drosophila heat shock factor to arrays of a conserved 5 bp unit. Cell , 64 , 585–593. Zimarino, V. and Wu, C. (1987). Induction of sequence-specific binding of Drosophila heat shock activator proteins without protein synthesis. Nature , 327 , 727–730. Zimarino, V., Tsai, C. and Wu., C (1990). Complex modes of heat shock factor activation. Mol. Cell. Biol. , 10 , 752–759. Zuo, J., Rungger, D. and Voellmy, R. (1995). Multiple layers of regulation of human heat shock transcription factor1. Mol. Cell Biol. , 15 , 4319–4330.
4. PROTEIN KINASE CASCADES INVOLVED IN HEAT SHOCK PROTEIN EXPRESSION AND FUNCTION OLIVIER BENSAUDE Génétique Moléculaire, Ecole Normale Supérieure, 46 rue d’Ulm, 75230 Paris Cedex 05, France
1. Introduction 2. Stress-activated MAP kinase cascades 2.1. Stress activation of the ERK type MAP kinases 2.2. Stress-activated MAP kinases related to the yeast osmosensing kinase, hog1 2.3. Multiple activating pathways 2.4. Negative regulation of MAP kinases 2.5. HSP interactions modulate protein kinase cascade activities 2.6. The MAP kinase cascades interfere with the cell survival after stress 3. HSP phosphorylation triggered by environmental stimuli 3.1. Protein kinases and phosphatases in the phosphorylation of small HSPs 3.2. The p38 MAP Kinase, an upstream activator of the small HSP Kinase 3.3. Phosphorylation of the small HSPs and cytoskeleton dynamics 4. Stress-induced phosphorylation of transcription factors 4.1. Transcription Factors Phosphorylated by MAP Kinases 4.2. Phosphorylation of the RNA polymerase II largest subunit by stressactivated CTD kinases 5. Involvement of MAP kinases in stress gene transcription 5.1. Heat shock element (HSE) dependent transcription 5.2. Stress response element (STRE) dependent transcription 5.3. GRP gene transcription 6. Control of protein synthesis 6.1. Dephosphorylation of ribosomal subunit S6 and eukaryotic initiation factor 4E (eIF-4E) 6.2. HSP depletion triggers the phosphorylation of the eukaryotic initiation factor 2 (eIF2 ) 7. Conclusion 8. Acknowledgements
Molecular chaperones and folding catalysts
60
9. References
1. INTRODUCTION The cell physiology is altered when cells are subjected to heat stress (reviewed in Laszlo, 1992; Nover, 1991; Welch, 1992). The cytoskeleton collapses, RNA splicing and protein synthesis are impaired, and the pattern of gene expression is extensively modified. The extent of the impairements depends on the severity of the stress and may lead to cell death. These alterations constitute the heat shock response and result in a reprogramming of central metabolic processes. The enhanced synthesis of heat shock proteins (HSPs), which contribute to repair and degradation of proteins damaged during stress, is a central manifestation of this reprogramming and contributes to the adaptation of the cells to the stress condition. Numerous posttranslational modifications of proteins occur in response to heat stress that may also be involved in preventing or repairing cellular damages. Changes in protein phosphorylation, glycosylation, methylation, acetylation, farnesylation and ubiquitination have been found to occur during stress (reviewed in Bensaude et al, 1996; Nover, 1991). Among these, protein phosphorylation is by far the best studied modification. Several HSPs are modified by phosphorylation in response to stress treatment. In E. coli for example, heat shock increases the proportion of phosphorylated GroEL and DnaK (Sherman and Goldberg, 1994; Sherman and Goldberg, 1993). But, apart from isolated reports on prokaryote protein phosphorylation, most studies concern the phosphorylation of the eukaryotic HSPs, and among these, phosphorylation of the “small” HSPs (23 to 28 kd) in response to a wide range of stimuli including growth factors, exposure to steroid hormones, cytokines and several stresses (oxidative, heat, osmotic, UV irradiation) (reviewed in Arrigo and Landry, 1994). Indeed, stress modulates the activity of various protein kinases signaling cascades. This chapter will focus on such cascades which relate to the HSPs in three aspects: (i) HSP levels control the activation of the cascades; (ii) HSPs are targets of the activated cascades; (iii) stress gene expression depends on the induction of protein kinases. To provide an accurate picture of the physiological importance of these regulatory circuits the discussion will be extended to stresses other than heat shock which trigger the same signaling pathways.
2. STRESS-ACTIVATED MAP KINASE CASCADES The mitogen activated protein kinases (MAP kinases or MAPK) belong to a network of signal transduction pathways (reviewed in Kyriakis and Avruch, 1996; Robinson and Cobb, 1997; Mahdani and Fink, 1998). Phosphorylation of these kinases is activated by MAP kinase kinases (MEK or MKK), which in turn are activated by MAP kinase kinase kinases (MAPKKK or MEKK). Many of these signal transduction pathways are induced by stress treatment that also induces synthesis of HSPs, and they are part of the cellular system that allows adaptation of metabolism to stress. On the one hand the activation of MAP kinases leads to the phosphorylation of the “small HSPs”, on the other hand it is
Protein kinase cascades involved in heat shock protein
61
involved in the regulation of stress gene expression. 2.1. Stress Activation of the ERK Type MAP Kinases The ERK type MAP kinases were first described as growth factor-activated kinases. Subsequent work established that in cultured mammalian cells, they were also activated during heat (Bendinelli et al., 1995; Venetianer et al., 1995; Chen et al., 1995) oxidative (Guyton et al., 1996) and osmotic (Matsuda et al., 1995) stress. In yeast, the activity of the ERK/MAP kinase homologue, encoded by the slt2/mpk1 gene, is modulated by the growth temperature and increases 170-fold when temperature is raised from 23 to 37°C (Kamada et al., 1995; Torres et al., 1991; Zarzov et al., 1996). 2.2. Stress-activated MAP Kinases Related to the Yeast hog1 Osmosensing Kinase A considerable interest has recently arised for two new types of MAP kinase strongly activated by UV irradiation, protein synthesis inhibitors, tumor necrosis factor, oxidative and heat stress (reviewed in Kyriakis and Avruch, 1996): the c-Jun kinases, JNKs or SAPK1 (stress-activated protein kinase) and the p38 MAP kinases. The latter are also designated by a wide variety of acronyms such as RK (Rouse et al., 1994), p40 MAP kinase (Freshney et al., 1994), p38 MAP kinase (Han et al., 1994), CSBP1 and CSBP2 (cytokine-suppressive anti-inflammatory drug—CSAID binding proteins) (Lee et al., 1994) or SAPK2 (Meier et al., 1996) (see Table 1).
Table 1 Mammalian mitogen-activated-protein kinases (MAPK) cascades: Acronyms MAP kinase kinase kinase kinases: Ste20, PAK1, p65PAK, GCK, SOK-1 MAP kinase kinase kinases: MEKKs, MAPKKKs, MKKKs, MUK, SAPKKKs, SEKKs, JNKKKs, RKKKs Raf-1, c-mos MAP kinase kinases: MEKs, MAPKKs, MKKs, SAPKKs, SEKs, JNKKs, RKKs MAP kinase repressers: Cdk inhibitor: p21 WAF1/CIP1/Sdi1 , p21 CIP1 , p21 WAF1 Dual specificity MAP kinase phosphatases: CL100, 3CH134, erp, MKP-1 B23, hVH-3 MAP kinases: Genuine mitogen-activated-protein kinases (MAPK) subclass:
Molecular chaperones and folding catalysts
62
ERK1, p44MAPK, p44mpk ERK2, p42MAPK, p42mpk Stress-activated protein kinase (SAPK) subclass of MAPK: JNK, SAP kinase, SAPK1; JNK1, SAPK-p46 I; JNK2, SAPK-p54 SAPKp54 I; SAPK-p46 I p38MAPK, p38HOG, CSBP1 and CSBP2, MPK2, SAPK2, RK
II; p54MAP-2 kinase,
MAP kinase activated kinases: MAPKAPK1, p90 ribosomal S6 kinase, p90rsk, RSK2 MAPKAPK2, HSP27 kinase, CREB kinase MAPKAPK3, HSP27 kinase, 3pK
Figure 1 Schematic representation of nuclear and cytoplasmic targets of the signaling network of the ERK, JNK, p38 MAP kinase pathways. PD098059 and SB203580 are highly selective inhibitors of respectively the MEK1/MEK2 kinases, which activate ERK1/2, and the p38 MAP kinase. The nucleus is schematized by a grey box.
The mammalian p38 MAP kinases are highly homologous to the Saccharomyces cerevisiae hog1 gene product that is involved in resistance to hyperosmotic stress. The
Protein kinase cascades involved in heat shock protein
63
human MAP kinase, CSBP1, complements yeast mutants carrying a deletion in hog1 (Han et al., 1994). Surprisingly, the CSBP2 kinase does not complement the hog1 deletion mutants despite its strong (96%) homology to CSBP1 (Kumar et al., 1995). Recently, a third member of the p38 MAP kinase family, p38 , has been characterized (Jiang et al., 1996). The p38 MAP kinases are furthermore closely related to the JNK. Similar to the CSBP1 gene, the mammalian JNK1 gene rescues a hog1 deficient yeast mutant (Galcheva-Gargova et al., 1994). However, although the JNK2 isoform shares 80% homology with JNK1, it does not rescue the yeast hog1 deficient yeast mutants (Sluss et al., 1994). The JNKs and the p38 MAP kinases are often presented as stress specific kinases and designated as SAPK1 and SAPK2 in contrast to the ERK MAP kinases which would rather be growth factor specific. This differentiation, however, seems not to be consistent with the recent findings described above and with the observation that stress specificity depends on the cell system investigated (Hu et al., 1997). 2.3. Multiple Activating Pathways The MAP kinases are activated by phosphorylation of threonine and tyrosine residues. Multiple stress-specific activators of p38 and JNK MAP kinases have been identified (Fanger et al., 1997). Depending on the inducer, different activating pathways are involved (Adler et al., 1995). Exposure of cells to N-acetylcysteine, an oxygen radical scavenger, or to low concentrations of Triton X-100 inhibits the UV-mediated, but not the heat shock-mediated JNK activation. Conversely, low concentrations of hydrogen peroxide inhibited heat shock-mediated but not UV-mediated JNK activation. Moreover, JNK activation by UV irradiation was impaired in cell lines derived from patients with Cockayne Syndrome of complementation group B (CS-B), whereas induction by oxidative stress and heat stress were normal (Dhar et al., 1996). Separate pathways have been demonstrated in the activation of the JNKs. One mechanism involves the c-Abl tyrosine kinase and is operative after treatment with ionizing radiation and genotoxic agents. Another pathway involving the tyrosine kinase, Pyk2, is required for activation by tumour-necrosis factor (TNF- ), UV and osmotic shock (Kharbanda et al., 1995; Tokiwa et al., 1996). Disruption of the MKK4 gene blocked JNK activation by anisomycin or heat shock but not so much by other stress such UV radiation or osmotic shock (Yang et al., 1997). Tyrosine kinase membrane receptors or the Ras small GTP protein in cooperation with the Raf-1 kinase are upstream activators of the ERK cascade. Upstream activators of the mammalian JNK and p38 cascades can be members of the family of p21-activated kinases, PAK This family includes the Ste20 kinase which controls mating in yeast (Sells and Chernoff, 1997) in cooperation with small GTP-binding proteins of the Rho subfamily, Rac1, Cdc42 (for reviews see (Nagata and Hall, 1996; Symons, 1996) and the human vav oncogene (Crespo et al., 1996). Independent signals, which emanate from distinct cell surface sensors, have been shown to activate the yeast Pbs2p MAP kinase (Posas et al., 1996). A phosphate residue is transferred from a histidine of the sensor kinase, Sln1p, to an aspartate of the response regulator, Ssk1p. Sln1p and Ssk1p are physically linked together by the Ypd1p. This
Molecular chaperones and folding catalysts
64
phosphorelay is initiated by the stress-induced autophosphorylation of Sln1p. Independently, Pbs2p can be activated by binding to the Src homology 3 (SH3) domain of the osmosensing protein, Sho1p. As a whole, most of the primary sensors which trigger the various MAP kinases cascades appear to be anchored in the plasma membrane where several members of the cascade can associate to each other and form molecular complexes (reviewed in Elion, 1995; Faux and Scott, 1996). It is tempting to speculate that upon heat shock treatment of the cells a temperature-induced increase in membrane fluidity might be the primary event that leads to their activation. Interestingly in yeast, the membrane lipid composition has been shown to be a determinant in the temperature set point of the heat shock response (Carratù et al., 1996). 2.4. Negative Regulation of MAP Kinases The MAP kinase cascades can be regulated at two steps, either phosphorylation or dephosphorylation of kinases (Burgering and Bos, 1995). Two types of protein phosphatases which dephosphorylate the critical threonine and tyrosine residues negatively regulate the MAP kinases (Groom et al., 1996; Keyse, 1995). The protein tyrosine phosphatase SHPTP1, which also downregulates the JNK, is activated by phosphorylation through the stress-activated c-Abl (Kharbanda et al., 1996). Another phosphatase, Pyst 1, which downregulates specifically the ERK type MAP kinases has also been identified (Groom et al., 1996). A serine phosphatase, protein phosphatase 2A, is responsible for the inactivation of the ERK type MAP kinases in some cell lines (Alessi et al., 1995). Although inducers of the heat shock response such as arsenite stimulate the JNK and p38 MAP kinases by inhibiting a dual specificity phosphatase (Cavigelli et al., 1996), the protein phosphatases are generally thought to provide a negative feed back control of the MAP kinase cascades as their expression is inducible by stress. Thus, expression of the S. pombe pyp2 gene encoding a phosphotyrosine phosphatase is stress-inducible and under the control of spc1 MAP kinase (Degols et al., 1996). A similar regulatory loop is also found in mammalian cells as the transcription of the dual specifity phosphatase genes, CL100 (MKP-1, 3CH134, erp) and B23/hVH-3 is rapidly enhanced after heat and oxidative stress and contributes to a downregulation of the MAP kinases of the ERK, p38 and JNK types (Chu et al., 1996b; Keyse, 1995; Sun et al., 1993) with some cell to cell variations (Alessi et al., 1995). Furthermore an anisomycin pretreatment selectively desensitizes components in the stress-kinase activation pathway (Hazzalin et al., 1998). Another negative regulation loop of MAP kinase cascades that relies on a mechanism distinct from phosphatases is mediated by protein p21 WAF1/CIP1/Sdil , a cyclin-dependentkinase inhibitor (reviewed in Harper and Elledge, 1996). This inhibitor also interacts with JNK1 and p38 MAP kinase, but not with ERK1 (Shim et al., 1996), thereby inhibiting the corresponding kinases. The p21 WAF1/CIPI/Sdil protein is induced two-fold by stress including heat shock (Ohnishi et al., 1996). This recently reported negative regulation loop illustrates a new connection between the MAP kinases and the cyclin-dependent kinases (cdk), which constitutes a distinct network of kinases controlling the cell cycle.
Protein kinase cascades involved in heat shock protein
65
2.5. HSP Interactions Modulate Protein Kinase Cascade Activities Interactions with HSP modulate the activation of protein kinases. For example Hsp70 overexpression prevents the stress-activation of the JNK and the p38 MAP kinases (Gabai et al., 1997). A novel member of the small heat shock protein family has recently been identified in muscle cells (Suzuki et al., 1998). This protein binds and activates a protein kinase the myotonic dystrophy protein kinase, MDPK Various protein kinases associate with Hsp90 and the cdc37 protein in their inactive form (reviewed in Hunter and Poon, 1997; Pratt, 1997). This is exemplified for tyrosine kinases such as src and for the Raf-1 protein kinase, an upstream activator of the ERK type MAP kinases (Morrison and Cutler, 1997). Raf-1 forms a heterocomplex with Hsp90 and p50Cdc37. Preventing the association of Raf-1 with Hsp90 by the antibiotic geldanamycin interferes with trafficking of the newly synthesized Raf-1 from the cytoplasm to the plasma membrane and with the association of Raf-1 with Ras. As a result, the Raf-1-MEK-MAP kinase signalling pathway is disrupted (Schulte et al., 1996; van der Straten et al., 1997). The src family tyrosine kinases are bound to Hsp90 when inactive. Phosphorylation of Hsp90 disrupts its complex with pp60v-src (Mimnaugh et al., 1995). In mammalian cells, Hsp90 chaperones are phosphoproteins for which the turnover of phosphate is increased during heat stress (Legagneux et al., 1991). It is interesting to note that upon heat shock tyrosine phosphorylation increases in cultured mammalian cells (Maher and Pasquale, 1989). In the fission yeast S. pombe, the wis1 and the spc1/sty1 MAP kinases are counteracted by overexpression of the phosphatase 2C (PP2C) encoded by the ptc1 gene (Shiozaki and Russell, 1995) and the phosphotyrosine phosphatases encoded by the pyp1 and the pyp2 genes (Degols et al., 1996). As the ptc1 gene is a multicopy suppressor of swo1–26, a temperature-sensitive mutation of the yeast hsp90 gene (Shiozaki et al., 1994), this observation also illustrates the importance of Hsp90 in the regulation of the MAP kinase cascade. 2.6. The MAP Kinase Cascades Interfere with the Cell Survival after Stress Stresses such as heat shock result in a variety of cellular responses that lead to increased stress resistance (see chapter by Li et al., this volume). Contribution of different stressactivated signal cascades to stress resistance will be presented. The S. cerevisiae slt2/mpk1 gene is essential for survival at high temperature (37°C) in normal medium (Kamada et al., 1995; Torres et al., 1991). Cells disrupted for the slt2/mpk1 gene lyse rapidly at 37°C with a phenotype similar to that of actin and polarized growth mutants (Mazzoni et al., 1993). The temperature sensitivity of the yeast cells disrupted for the slt2/mpk1 gene can be rescued by growing the cells at high osmolarity. The HOG cascade which mediates the response of cells to exposure to high osmolarity is not however, responsible for this rescue but the deletion of antioxidant genes controlled by the HOG cascade contributes to the lethality after heat shock in S. cerevisiae (Davidson et al., 1996). In mammalian cells, ERK and JNK-p38 MAP kinases show opposing effects on apoptotic death. Treatments that induce apoptosis, such as nerve growth factor
Molecular chaperones and folding catalysts
66
withdrawal from PC 12 pheochromocytoma cells in culture, activate the JNK-p38 MAP kinases and apoptosis was prevented by the expression of recombinant dominant negative p38 or JNK mutants (Xia et al., 1995). In contrast, constitutive activation of the ERK signalling pathway prevented apoptosis. Expression of a constitutive JNK activator (but not of an ERK constitutive activator) in Swiss 3T3 and REF52 fibroblasts enhanced the apoptotic response to UV irradiation (Johnson et al., 1996). Blockade of p38 MAP kinase and JNK activity prevented cell death in rat fibroblasts following UV irradiation, heat shock (Zanke et al., 1996) and exposure to elevated concentrations of sodium salicylate (Schwenger et al., 1997). However, inhibition of the p38 MAP kinase did not affect the toxic effects of tumour necrosis factor in murine cells (Beyaert et al., 1996).
3. HSP PHOSPHORYLATION TRIGGERED BY ENVIRONMENTAL STIMULI In the following paragraphs, we will review first the protein kinases pathways involved in small Hsp phosphorylation and second the biological implications proposed for this modification. 3.1. Protein Kinases and Phosphatases in the Phosphorylation of Small HSPs The phosphorylation status of a protein is determined by the balance between kinases and phosphatases. In mammalian cells, protein kinases which are activated to phosphorylate small HSPs in response to environmental stimuli have been identified (Guesdon et al., 1993; Stokoe et al., 1992b; Zhou et al., 1993). Two proteins of 45-kDa/54-kDa showing stress induced Hsp27 kinase activity have been characterized in hamster cells (Huot et al., 1995). In murine cells, two related 45-kDa/ 55-kDa kinases are activated by the protein synthesis inhibitor, anisomycin, and the epidermal growth factor (Cano et al., 1996). These kinases are easily detected by in-gel kinase assays. The small HSP kinases are activated in vitro by MAP kinases (see previous section). An Hsp27 kinase was identified as MAP kinase activated protein kinase 2 (MAPKAPK2) (Stokoe et al., 1992a). MAPKAPK2 phosphorylates human Hsp27 at serines 15, 78 and 82 (Landry et al., 1992) and murine Hsp25 at serines 15 and 86 (there is no serine at the position corresponding to 78 in the human protein) (Gaestel et al., 1991). Interestingly, these phosphorylation sites share with the MAP kinase activated protein kinase 1 (MAPKAPK1) the Arg-X-X-Ser motif. However, MAPKAPK2 differs from MAPKAPK1 in its requirement for a bulky hydrophobic residue in the -5 position Nterminal to the serine (Stokoe et al., 1993). Several MAPKAPK2 cDNAs have been cloned (Stokoe et al., 1992a; Zu et al., 1994) and recently a related but distinct cDNA has been isolated. This homologue is identical to 3pK, a kinase gene located in the gene region of the small cell lung cancer tumor suppressor (Sithanandam et al., 1996). The corresponding recombinant protein product shows Hsp27 kinase activity and is also designated as MAPKAPK3 (McLaughlin et al., 1996). The in vivo role for this kinase in phosphorylation of small HSPs remains unclear.
Protein kinase cascades involved in heat shock protein
67
Less is known with respect to the phosphatases controlling small HSP phosphorylation. In cell-free systems, the human small HSPs are dephosphorylated by the calcium/calmodulin-dependent protein phosphatase, calcineurin (PP2B) (Gaestel et al., 1992), and type A protein phosphatase (PP2A) (Cairns et al., 1994). In vivo, however, the dephosphorylation of small HSPs is insensitive to cyclosporin A, a potent inhibitor of PP2B, suggesting that this phosphatase is not involved. The involvement of PP2A in small HSP dephosphorylation in vivo remains to be determined. Taken together, it seems possible that both, kinase activation and phosphatase inactivation contribute to the increased phosphorylation of small HSPs in response to stress. 3.2. The p38 MAP Kinases, Upstream Activators of the Small HSP Kinase (MAPKAPK2) Although the MAP kinases of the ERK type are activated by inducers of the small HSP phosphorylation, the activation of MAPKAPK2 by ERK1/2 became questionable when it became apparent that MAPKAPK2 activation did not strictly correlate with ERK1/2 activation (Guesdon et al., 1993). The MAPKAPK2 is activated in vitro both by the ERK type and by the p38 type MAP kinases (Guay et al., 1997). In mammalian cells, the p38 MAP kinase are activated by a wide variety of stresses including heat shock, oxidative stress (Rouse et al., 1994), hyperosmolarity (Han et al., 1994), endotoxic lipopolysaccharide (LPS) (Han et al., 1994), TNF(Beyaert et al., 1996) and interleukin-1 (IL-1) (Freshney et al., 1994). All these treatments also lead to the phosphorylation of small HSPs (reviewed in Arrigo and Landry, 1994). Most fascinating is the discovery of drugs which inhibit specific MAP kinase cascades (Cohen, 1997). PD098059 is a specific inhibitor in vitro of MEK1 and MEK2, the ERK1/2 activating kinases. In vivo, this compound prevents ERK2 activation but had no influence on p38 MAP kinase and MAPKAPK2 activation. The anti-inflammatory drug, CSAID/SB203580, binds with high affinity to the p38 MAP kinase and inhibits its activity with an extremely high specificity. Stress activation of MAPKAPK2 and phosphorylation of Hsp27 is suppressed in mammalian cells exposed to micromolar concentrations of SB203580 (Ben-Levy et al., 1995; Cuenda et al., 1995; Guay et al., 1997). This result provides strong support for the involvement of a p38 MAP kinase and MAPKAPK2 in phosphorylation of the small HSP during stress. The B-crystallin is another small HSP phosphorylated in response to stress and depending on the stress, either the ERK or the p38 MAP kinase is involved (Ito et al., 1997). The phosphorylation of small HSPs induced by lymphokines is abolished in cells treated with two tyrosine kinase inhibitors, genistein and herbimycin A (Ahlers et al., 1994), suggesting an involvement of a tyrosine kinase upstream in the activation pathway. 3.3. Phosphorylation of the Small HSPs and Cytoskeleton Dynamics Phosphorylation of the small HSPs often parallels the formation of large protein aggregates. It has been suggested that phosphorylation contributes to the dissociation of these aggregates (Kato et al., 1994; Lavoie et al., 1995; Mehlen et al., 1995). But most significantly, phosphorylation of the small HSPs might modulate fluid phase pinocytosis and actin microfilament dynamics (reviewed in Landry and Huot, 1995). In vitro, the
Molecular chaperones and folding catalysts
68
unphosphorylated chicken small HSP inhibits actin polymerization (Miron et al., 1991), and phosphorylation of the murine Hsp25 abolishes its actin-depolymerizing activity (Benndorf et al., 1994). The human Hsp27 is accumu-lated in motile cell protrusions such as lamellipodia, filopodia, and membrane ruffles. These structures also stain for a number of known and suspected actininteracting proteins thought to regulate the dynamics and structure of F-actin. Collapse of the actin microfilaments occurs in response to growth factors, heat and oxidative stresses (Huot et al., 1995; Lavoie et al., 1995). Hsp27 overexpression in hamster cells enhances cell survival in the presence of the actin polymerization inhibitor, cytochalasin D (Lavoie et al., 1995). This protective effect does not occur with a triple phosphorylation site mutant of human Hsp27. Inhibiting the phosphorylation of Hsp27 with compound SB203580 (see below) suppresses the protective effects of Hsp27 overexpression (Guay et al., 1997). Oxidative stress induces both Hsp27 phosphorylation and severe fragmentation of the F-actin network, however the SB203580 compound blocked these responses in cells which express high levels of Hsp27 (Huot et al., 1997). Expression of small HSPs in murine cells was reported to confer resistance to TNF- and oxidative stress-induced necrotic (Huot et al., 1996; Mehlen et al., 1995) and apoptotic (Mehlen et al., 1996) cell death. Hamster cells stably transfected with DNA encoding human Hsp27 show an increased resistance to heat, oxidative stress and/or to actin depolymerizing agents but this effect is abolished upon replacement of the phosphorylated serines by glycines (Huot et al., 1996; Lavoie et al., 1995). Yet it should be mentioned that in transformed murine fibroblasts transiently transfected with a murine Hsp25 expression vector, replacement of the phosphorylated serines by alanines did not diminish the increased resistance to heat (Knauf et al., 1993) and negative results were obtained in other laboratories which used cell lines constitutively expressing small HSPs (Arata et al., 1995; Beyaert et al., 1996). The SB203580 compound had little effect on fibroblasts which do not spontaneously express Hsp27, however this influence was acquired if these cells were forced to permanently express high levels of Hsp27 after transfection (Huot et al., 1997). Differences in constitutive levels of Hsp27 in the various systems investigated may explain the conflicting results. Interestingly, the Rho family of GTPases which indirectly controls the actin cytoskeleton dynamics (reviewed in Symons, 1996; Zigmond, 1996) also contribute to the control of Hsp27 phosphorylation as upstream activators of the p38 MAPK kinase cascade. Stress-induced phosphorylation of non-heat shock proteins may also modulate the cytoskeleton dynamics. For example stathmin is phosphorylated on MAP kinase sites in response to a wide variety of treatments, including growth factors and stress such as heat shock (Beretta et al., 1995). Stathmin destabilizes microtubules in vitro and it has been suggested that phosphorylation controls this activity in vivo (Belmont and Mitchison, 1996). Thus, the environmentally-induced phosphorylation of the small HSPs and of stathmin may influence the cyto-skeleton dynamic.
Protein kinase cascades involved in heat shock protein
69
4. STRESS-INDUCED PHOSPHORYLATION OF TRANSCRIPTION FACTORS Numerous putative targets of the ERK/MAP kinases, p38 and JNK have been described. Some of them are involved in transcriptional regulation (reviewed in Cahill et al., 1996; Treisman, 1996). The following paragraphs will review the involvement of MAP kinases in the regulation of gene expression. 4.1. Transcription Factors Phosphorylated by MAP Kinases In mammals, the MAP kinases phosphorylate and modulate the activity of numerous transcription factors including c-Jun, ATF-2, Elk-1 (Raingeaud et al., 1996; Treisman, 1996; and references therein) and CHOP (also known as growth arrest and DNA damageinducible gene 153, GADD153) (Wang and Ron, 1996). Some of these factors are targets for specific kinases, e.g. c-Jun for the JNKs and CHOP for the p38 MAP kinase. Phosphorylation of c-Jun by JNK both targets c-Jun for ubiquitination and prolongs its half-life (Fuchs et al., 1996). The cyclic-AMP response element binding protein, CREB or CRE-BP1, and ATF-1 are phosphorylated downstream the p38 MAP kinase in response to stress or fibroblast growth factor (FGF). Both MAPKAPK1/RSK2 and MAPKAPK2 are CREB kinases in vitro (Tan et al., 1996; Xing et al., 1996). Other factors such as ATF-2, a member of the CREB subfamily of transcription factors but unresponsive to cyclic AMP, or Elk-1, a member of the Ets family, integrate signals from different kinase cascades (Cahill et al., 1996). In addition to transcription factors, chromatin components are MAP kinase targets as well. In particular, hyperphosphorylation of the Sir3p yeast protein through a MAP kinase cascade strengthens transcriptional silencing during mating, starvation and heat shock (Stone and Pillus, 1996). 4.2. Phosphorylation of the RNA Polymerase II Largest Subunit by Stressactivated CTD Kinases In many cells including Drosophila cells, mild heat shock causes dephosphorylation of the C-terminal domain (CTD) of the largest subunit of RNA polymerase II whereas upon severe heat shock, however, the RNA polymerase largest subunit accumulates in a form that remains multiphosphorylated on the CTD (Dubois et al., 1994; Venetianer et al., 1995) (reviewed in Dubois and Bensaude, 1998). Numerous studies have established that the CTD undergoes a cycle of phosphorylation/dephosphorylation in connection with gene transcription (Dahmus, 1996). RNA polymerase II with an unphosphorylated CTD assembles on the promoters and phosphorylation of the CTD appears to be required to release the interactions between the polymerase and the transcription factors sitting on the promoters. Likewise, RNA polymerase assemblies that are hypophosphorylated on their CTD bind to the 5' end of heat shock genes and pause after initiating transcription (O’Brien et al., 1994; Weeks et al., 1993). Phosphorylation occurs upon entry into
Molecular chaperones and folding catalysts
70
elongation of transcription and is next required to dock the various factors involved in mRNA maturation (splicing, polyadenylation) (Steinmetz, 1997). The CTD is phosphorylated within the preinitiation complex of transcription by a CTD kinase associated with the TFIIH general transcription factor (Seroz et al., 1995; Valay et al., 1995). Furthermore, TFIIH subunits accumulate within the coiled bodies which contain splicing factors (Grande et al., 1997; Jordan et al., 1997). The CTD dephosphorylation observed during mild heat stress relates to the impairement of TFIIH-associated kinase (Dubois et al., 1997). As the heat shock genes remain very efficiently transcribed in these conditions, they might differ from other cellular genes in their requirements for an active TFIIH kinase. Heat stress activates a CTD kinase which copurifies with ERK type MAP kinases. The stressactivated MAP kinases might rescue the impairement of the TFIIH-associated CTD kinase. As initiation of transcription requires an unphosphorylated CTD, the hyperphosphorylation of the CTD observed during severe stress may contribute to a general shut-off of transcription. The CTD phosphorylated in mammalian cells during severe stress or in vitro with purified MAP kinase lacks a particular epitope associated with splicing factors and generated in vitro by phosphorylation with TFIIH (Dubois et al., 1997). Data obtained with insect cells are also suggestive of abnormal phosphorylation of the CTD from polymerases engaged in heat shock gene transcription (Weeks et al., 1993). Thus the abnormal CTD phosphorylation occuring during heat shock may also relate to the impairement of splicing occuring during such stress. Noteworthy, the predominant heatinducible hsp70 genes have no introns in insects and vertebrates.
5. INVOLVEMENT OF MAP KINASES IN STRESS GENE TRANSCRIPTION 5.1. Heat Shock Element (HSE) Dependent Transcription A connection exists at least at some conditions between the MAP kinase cascades and the heat shock transcription factor (HSF)-dependent heat shock response. The stress activation of heat shock gene transcription is inhibited in yeast and rat cells in which the Ras pathway is constitutively active (Engelberg et al., 1994). Activation of ERK kinases has been suggest to inhibit heat shock response in murine fibroblasts (Mivechi and Giaccia, 1995). Indeed, the heat shock transcription factor, HSF-1, is an in vitro substrate for ERK MAP kinases and constitutive activation of these kinases represses heat shock gene transcription at controlled temperatures (Chu et al., 1996a; Knauf et al., 1996) (see chapter Morimoto). In constrast, p38 MAP kinase activation is required for osmotic stress induction of Hsp70 mRNA (but not for heat-shock induction) (Sheikh-Hamad et al., 1998). 5.2. Stress Response Element (STRE) Dependent Transcription The HOG kinase cascade activates the general stress response in S. cerevisiae (reviewed in Ruis and Schüller, 1995). This response is distinct from the HSF-dependent heat shock
Protein kinase cascades involved in heat shock protein
71
response and requires a regulatory DNA sequence, the stress response element (STRE) with the consensus sequence CCCCT. It mediates the stress induction of transcription of a set of genes including the cytosolic catalase T (ctt1) gene, the DNA damage responsive gene (ddr2) (Kobayashi and McEntee, 1993), the trehalose phosphate phosphatase (tps2) gene (Gounalaki and Thireos, 1994) and the glycerol-3-phosphate dehydrogenase (gpd1) gene. Furthermore, STRE elements are present in the promoters of heat shock genes containing heat shock elements such as ubi4, hsp12, hsp104 and a protein tyrosine phosphatase gene (ptp2) (Ruis and Schüller, 1995). It is interesting to note that many genes controlled by the HOG pathway are involved in either protection from or repair of damage. In S. pombe, the Hog1 kinase homologue, Spc1, mediates the stress induction of the gpd1 and tps2 genes as well as the pyp2 tyrosine phosphatase which inactivates spc1 (see above) (Degols et al., 1996). The Atf1 transcription factor, a homologue of the mammalian ATF-2 factor, mediates this general stress response (Shiozaki and Russell, 1996). In rat pheochromocytoma cells, arsenite treatment leads to enhanced expression of the C/EBP- and CHOP transcription factors (Fawcett et al., 1996). The CHOP factor has been postulated to act as a dominant-negative regulator of CCAAT/ enhancer-binding proteins (C/EBP) and may be involved in an autoregulatory loop. Heat and arsenite stress enhance the transcription of the c-fos proto-oncogene (Bukh et al., 1990; van Delft et al., 1993). This enhanced transcription is accompanied by a marked increase in c-fos mRNA stability (Andrews et al., 1987). The induction of c-fos is likely a consequence of JNK and p38 MAP kinase activation. Indeed, SB203580 suppresses the induction of c-Fos and c-Jun following anisomycin or UV treatments (Hazzalin et al., 1996). The c-fos mRNA accumulation might be responsible for heat-induced FGF gene expression (Erdos et al., 1995) and could lead, in some cases, to a stimulated cell cycle progression after mild stress (van Wijk et al., 1993). 5.3. GRP Gene Transcription Stress may occur in distinct organelles which poses the problem of signal communication between cellular compartments. Signal pathways sensing stress in the ER and transducing it to the nucleus have recently been discovered (see chapter by Morimoto). In yeast, one begins to understand the activation pathway to kar2 expression (the grp78/hsp70 homologue). The ire1/ern1 gene product is one of the sensors for unfolded proteins in the ER (Shamu, 1998; Welihinda et al., 1998). It is a protein kinase homologous to cdc28 and distinct from the MAP kinases. It is oligomerized and phosphorylated during signaling from the ER to the nucleus. The primary function of the Ire1 protein appears to stabilize the expression of the Hac1 transcription factor, that is homologous to the mammalian ATF andCREB transcription factors, by promoting appropriate HAC1 mRNA splicing. In mammalian cells, induction of Grp78 is inhibited by genistein, an inhibitor of tyrosine kinases such as c-Abl, suggesting an involvement of kinases in the signal transduction pathway. However the yeast Ire1p protein is inefficient in inducing the mammalian grp78 gene in an heterologous transfection assay (Cao et al., 1995). Signals from stressed ER leading to the induced synthesis of BiP/Grp78 are also efficient in inducing the transcription factor, CHOP (GADD153) (Wang et al., 1996) which is phosphorylated and activated in turn by p38 MAP kinase (Wang and Ron,
Molecular chaperones and folding catalysts
72
1996). This transcription factor binds to specific regulatory DNA elements, the CCAAT sequences. Interestingly, transcription of the gene encoding the mammalian BiP/Grp78 following calcium stress is itself transduced by a CCAAT element proximal to the site of transcription initiation (Roy and Lee, 1995). It is unclear at present whether a synergy exists between the different regulatory pathways.
6. CONTROL OF PROTEIN SYNTHESIS Protein synthesis is rapidly arrested in eukaryotic cells submitted to heat shock (reviewed in Duncan, 1996; Brostrom et al., 1998). During recovery from stress, a preferential translation of the HSP mRNAs is observed. The translation of some stress regulated mRNAs might be initiated by a ribosome jumping mechanism operating when eukaryotic initiation factors are deficient (Yueh and Schneider, 1996). It is an attractive possibility that changes in the phosphorylation state of various components of the translational apparatus contribute to these alterations. 6.1. Dephosphorylation of Ribosomal Subunit S6 and Eukaryotic Initiation Factor 4E (eIF-4E) Phosphorylation of the ribosomal S6 protein increases the translation efficiency of mRNAs coding for proteins of the translational apparatus such as the eukaryotic elongation factor, eEF-1 (Stewart and Thomas, 1994). Heat shock induces the rapid dephosphorylation of S6 in growing cells (reviewed in Nover, 1991) and, in sharp contrast, the phosphorylation of S6 in quiescent fibroblasts (Jurivich et al., 1991). Two kinases which phosphorylate S6 in vitro, p70s6k and p90rsk S6 kinases, are activated in vivo by heat shock as well as by mitogenic stimulation (Kozma and Thomas, 1994). The p90rsk which is activated by the ERK/MAP kinases (see above), is referred as MAPK activated protein kinase 1 (MAPKAPK1) or rRSK2. The p70s6k or RSK1 is turned on by a transducing pathway which is specifically inhibited by the immuno suppressant rapamycin (Proud, 1996). p70s6k is the major physiological S6 kinase in mammalian cells. Both S6 kinases have additional targets distinct from S6. One of the p70s6k targets is the protein synthesis represser 4E-BP1 (BP-1; PHAS-I) (von Manteuffel et al., 1996). Phosphorylation of 4E-BP1 would disrupt its interaction with protein synthesis initiation factors, e1F-4E, allowing it to interact with the mRNA cap binding complex and restore cap-dependent protein synthesis (Lawrence and Abraham, 1997). Phosphorylation of 4EBP1 decreases during heat shock, and this decrease is aggravated in the presence of rapamycin which blocks p70s6k activity (Feigenblum and Schneider, 1996). Heat shock also triggers the dephosphorylation of the capbinding factor, eIF-4E. Consequently, an increased association of 4E-BP1 with eIF-4E is observed (Vries et al., 1997). The resulting inactivation of eIF-4E along with the dephosphorylation of 4E-BP1 and phosphorylation of eIF-2 might contribute to the selective translation of heat shock mRNAs in stressed cells (Yueh and Schneider, 1996). Multiple signaling pathways mediate the phosphorylation of eIF-4E in response to growth factor activation (Morley et al., 1997). Rapamycin (inhibitor of p70s6k activation), P98059 (MEK inhibitor),
Protein kinase cascades involved in heat shock protein
73
SB203380 (p38 MAP kinase inhibitor) all block eIF-4E phosphorylation. Interestingly, they delay recovery of protein synthesis after heat-shock (our unpublished results). Hence, the MAP kinase cascades appear to play an important role in recovery of protein synthesis after stress. 6.2. HSP Depletion Triggers the Phosphorylation of the Eukaryotic Initiation Factor 2 (eIF2 ) Lysates prepared from heat-shocked human cells are inactive for translation in vitro. This inactivation may in part be due to phosphorylation of the a subunit of initiation factor 2 (eIF-2 ) resulting from the stress activation of the heme-regulated inhibitor (HRI) or hemin-controlled eIF-2a kinase (HCR) (De Benedetti and Baglioni, 1986; Farrell et al., 1977). HRI or HCR is a protein kinase, that is activated by hemin deficiency (reviewed in Chen and London, 1995). Phosphorylation of this factor on serine 51 inhibits initiation of translation (Samuel, 1993). Although phosphorylation of eIF-2 is not the sole mechanism responsible for the shut-off of protein synthesis under heat stress conditions, there is evidence for an involvement of such mechanism. Expression of a phosphorylation-resistant eIF-2 mutant (serine 51 is replaced by an alanine) partially protects cells from the inhibition of protein synthesis in response to heat treatment (Murtha-Riel et al., 1993). Noteworthy, eIF2 is itself a heat shock protein (Colbert et al., 1987) which may contribute to a quicker recovery of the cells from translational arrest during the shut off phase of the heat shock response. In contrast to the HCR, the double-stranded RNA-activated protein kinase, PKR, is an eIF-2 kinase which is impaired during heat stress (Dubois et al., 1991). This kinase contributes to the antiviral effects of interferon through the inhibition of protein synthesis. Some viruses such as the influenza virus avoid the translational inhibitory effects of PKR by activating a cellular inhibitory protein, P58IPK. Interestingly, this protein possesses a significant homology to the J domain of DnaJ (Hsp40), and is inhibited through an interaction with Hsp40 (Melville et al., 1997). Interaction with HSPs is also involved in the control of HCR activity. In reticulocyte lysates, the inactive HCR is bound to Hsp90 (Rose et al., 1987), Hsp70 and Hsp56 (Matts et al., 1992). Interaction with Hsp90 is obligatory for HCR to acquire and maintain an activable conformation (Uma et al., 1997). Hsp70 binding to HCR stabilizes the inactive HCR. Limiting the amount of Hsp70 in the lysates reduces the concentration of hemin required to prevent activation of HCR (Gross et al., 1994). Addition of denatured proteins to the lysate activates HCR (Matts et al., 1993). Denatured proteins may bind to Hsp70 present in the lysate and deplete the lysate from “free” Hsp70. The dissociation of Hsp70 from HCR is strongly correlated with the activation of HCR suggesting that sequestration of Hsp70 might be the signal that leads to activation of HCR in response to heat shock. Since Hsp70 chaperones assist the folding of nascent polypeptide chains (Beckmann et al., 1990), the cellular economy might justify the arrest of protein synthesis when the pool of “free” Hsp70 is depleted.
Molecular chaperones and folding catalysts
74
7. CONCLUSION The heat shock responses involve not only chaperones but also protein kinases, protein phosphatases and detoxifying enzymes, which all contribute to the cellular resistance to stress. The responses to stress activate various signaling cascades of protein kinases, yet the same cascades are activated by different stimuli generated by stress treatment as well as developmental processes. Different stimuli activate the same cascades through different pathways and each cascade targets a wide spectrum of functions including those affecting cell shape, protein synthesis and gene expression. A further level of complexity is brought about by the fact that a given protein may be targeted by different cascades and integrate their specific signals. As a whole, we begin to understand that the protein kinase cascades form an intricate network of interactions which regulates the cellular homeostasis. The signaling cascades represent an early step in the reaction of cells to an economy imbalance generated by stress.
8. ACKNOWLEDGMENTS I wish to thank Drs. A.P.Arrigo, M.-F.Dubois, M.Gaestel, J.Landry and A.Michels for critical reading of the manuscript.
9. REFERENCES Adler, V., Schaffer, A., Kim, J., Dolan, L. and Ronai, Z. (1995). UV irradiation and heat shock mediate JNK activation via alternate pathways. J. Biol Chem. , 270 , 26071– 26077. Ahlers, A., Engel, K., Sott, C., Gaestel, M., Herrmann, F. and Brach, M.A. (1994). Interleukin-3 and granulocyte-macrophage colony-stimulating factor induce activation of the MAPKAP kinase 2 resulting in in vitro serine phosphorylation of the small heat shock protein (hsp27). Blood , 83 , 1791–1798. Alessi, D.R., Gomez, N., Moorhead, G., Lewis, T., Keyse, S.M. and Cohen, P. (1995). Inactivation of p42 MAP kinase by protein phosphatase 2A and a protein-tyrosine phosphatase, but not CL100, in various cell lines. Curr. Biol. , 5 , 283–288. Andrews, G.K., Harding, M.A., Calvet, J.P. and Adamson, E.D. (1987). The heat shock response in HeLa cells is accompanied by elevated expression of the c-fos protooncogene. Mol. Cell. Biol. , 7 , 3452–3458. Arata, S., Hamaguchi, S. and Nose, K. (1995). Effects of the overexpression of the small heat shock protein, HSP27, on the sensitivity of human fibroblast cells exposed to oxidative stress. J. Cell. Physiol. , 163 , 458–465. Arrigo, A.P. and Landry, J. (1994). Expression and function of the low-molecular-weight heat shock proteins. In: The biology of heat shock proteins and molecular chaperones . Morimoto, R.I., Tissiéies, A. and Georgopoulos, C. eds. Cold Spring Harbor Laboratory Press, 335–373. Beckmann, R.P., Mizzen, L.A. and Welch, W.J. (1990). Interaction of hsp 70 with newly
Protein kinase cascades involved in heat shock protein
75
synthesized proteins: implications for protein folding and assembly. Science , 248 , 850– 854. Belmont, L. and Mitchison, T. (1996). Catastrophic revelations about Op18/stathmin. Trends Biochem. Sci. , 21 , 197–198. Ben-Levy, R., Leighton, I.A., Doza, Y.N., Attwood, P., Morrice, N., Marshall, C.J. and Cohen, P. (1995). Identification of novel phosphorylation sites required for activation of MAPKAP kinase-2. EMBO J. , 14 , 5920–5930. Bendinelli, P., Piccoletti, R., Maroni, P. and Bernelli-Zazzera, A. (1995). The liver response to in vivo heat shock involves the activation of MAP kinases and RAF and the tyrosine phosphorylation of She proteins. Bioch. Biophys. Res. Comm. , 216 , 54– 61. Benndorf, R., Hayeß, K., Ryazantsev, S., Wieske, M., Behlke, J. and Lutsch, G. (1994). Phosphorylation and supramolecular organization of murine small heat shock protein HSP25 abolish its actin polymerization-inhbiting activity. J. Biol. Chem. , 269 , 20780–20784. Bensaude, O., Bellier, S., Dubois, M.-F., Giannoni, F. and Nguyen, V.T. (1996). Heatshock induced protein modifications and modulation of enzyme activities. In: Stressinducible cellular responses . Feige, U., Morimoto, R.I., Yahara, I. and Polla, B., eds. Basel: Birkhaüser Verlag, 199–219. Beretta, L., Dubois, M.-F., Sobel, A. and Bensaude, O. (1995). Stathmin, a major substrate for mitogen-activated protein kinase during heat shock and chemical stress. Eur. J. Biochem. , 227 , 388–395. Beyaert, R., Cuenda, A., Vanden Berghe, W., Plaisance, S., Lee, J.C., Haegeman, G., Cohen, P. and Fiers, W. (1996). The p38/RK mitogen-activated protein kinase pathway regulates interleukin-6 synthesis in response to tumour necrosis factor. EMBO J. , 15 , 1914–1923. Brostrom, C.O. and Brostrom, M.A. (1998). Regulation of translational initiation during cellular responses to stress. Prog, in Nuc. Ac. Res. & Mol. Biol. , 58 , 79–125. Bukh, A., Martinez-Valdez, H., Freedman, J., Freedman, M.H. and Cohen, A. (1990). The expression of c-fos, c-jun, and c-myc genes is regulated in human lymphoid cells. J. Immunol. , 144 , 4835–4840. Burgering, B.M.T. and Bos, J.L. (1995). Regulation of Ras-mediated signaling: more than one way to skin a cat. Trends in Biochem. Sci. , 20 , 18–22. Cahill, M.A., Janknecht, R. and Nordheim, A. (1996). Signaling pathways: Jack of all cascades. Curr. Biol. , 6 , 16–19. Cairns, J., Qin, S., Philp, R., Tan, Y.H. and Guy, G.R. (1994). Dephosphorylation of the small heat shock protein hsp27, in vivo by protein phosphatase 2A. J. Biol. Chem. , 269 , 9176–9183. Cano, E., Doza, Y.N., Ben-Levy, R., Cohen, P. and Mahadevan, L.C. (1996). Identification of anisomycin-activated kinases p45 and p55 in murine cells as MAPKAP kinase-2. Oncogene , 12 , 805–812. Cao, X., Zhou, Y. and Lee, A.S. (1995). Requirement of tyrosine- and serine/threonine kinases in the transcriptional activation of the mammalian grp78/BiP promoter by thapsigargin. J. Biol. Chem. , 270 , 494–502. Carratù, L., Franceschelli, S., Pardini, C.L., Kobayashi, G.S., Horvath, I., Vigh, L. and Maresca, B. (1996). Membrane lipid perturbation modifies the set point of the temperature of heat shock response in yeast. Proc. Natl. Acad. Sci. USA , 93 , 3870– 3875. Cavigelli, M., Li, W.W., Lin, A., Su, B., Yoshioka, K. and Karin, M. (1996). The tumor
Molecular chaperones and folding catalysts
76
promoter arsenite stimulates AP-1 activity by inhibiting a JNK phosphatase. EMBO J. , 15 , 6269–6279. Chen, J.-J. and London, I.M. (1995). Regulation of protein synthesis by heme-regulated eIF-2 kinase. Trends in Biochem. Sci. , 20 , 105–108. Chen, F., Torres, M. and Duncan, R.F. (1995). Activation of mitogen-activated protein kinase by heat shock treatment in Drosophila. Biochem. J. , 312 , 341–349. Chu, B., Soncin, F., Price, B.D., Stevenson, M.A. and Calderwood, S.K. (1996a). Sequential phosphorylation by mitogen-activated protein kinase and glycogen synthase kinase 3 represses transcriptional activation by heat shock factor-1. J. Biol. Chem. , 271 , 30847–30857. Chu, Y., Solski, P.A., Khosravi-Far, R., der, C.J. and Kelly, K. (1996b). The mitogenactivated protein kinase phosphatases PAC-1, MKP-1, and MKP-2 have unique substrate specificities and reduced activity, in vivo toward the ERK2, sevenmaker mutation. J. Biol. Chem. , 271 , 6497–6501. Cohen, P. (1997). The search for physiological substrates of MAP and SAP kinases in mammalian cells. Trends in Cell Biology , 7 , 353–358. Colbert, R.A., Hucul, J.A., Scorsone, K.A. and Young, D.A. (1987). Subunit of eukaryotic translational initiation factor-2 is a heat-shock protein. J. Biol. Chem. , 262 , 16763–16766. Crespo, P., Bustelo, X.R., Aaronson, D.S., Coso, O.A., Lopez-Barahona, M., Barbacid, M. and Gutkind, J.S. (1996). Rac-1 dependent stimulation of the JNK/SAPK signaling pathway by Vav. Oncogene , 13 , 455–460. Cuenda, A., Rouse, J., Doza, Y.N., Meier, R., Cohen, P., Gallagher, T.F., Young, P.R. and Lee, J.C. (1995). SB 203580 is a specific inhibitor of a MAP kinase homologue which is stimulated by cellular stresses and interleukin-1. FEBS Letters , 364 , 229– 233. Dahmus, M.E. (1996). Reversible phosphorylation of the C-terminal domain of RNA polymerase II. J. Biol. Chem. , 271 , 19009–19012. Davidson, J.F., Whyte, B., Bissinger, P.H. and Schietl, R.H. (1996). Oxidative stress is involved in heat-induced cell death in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. (USA) , 93 , 5116–5121. De Benedetti, A. and Baglioni, C. (1986). Activation of Hemin-regulated initiation factor-2 kinase in heat-shocked HeLa cells. J. Biol. Chem. , 261 , 338–342. Degols, G., Shiozaki, K. and Russell, P. (1996). Activation and regulation of the Spc1 stress-activated protein kinase in Schizosaccharomyces pombe. Mol. Cell. Biol. , 16 , 2870–2877. Dhar, V., Adler, V., Lehman, A. and Ronai, Z. (1996). Impaired Jun-NH2-terminal kinase activation by ultraviolet irradiation in fibroblasts of patients with Cockayne Syndrome complementation group B. Cell Growth & Differentiation , 7 , 841–846. Dubois, M.-F., Bellier, S., Seo, S.-J. and Bensaude, O. (1994). Phosphorylation of the RNA polymerase II largest subunit during heat-shock and inhibition of transcription in HeLa cells. J. Cell. Physiol. , 158 , 417–426. Dubois, M.-F., Vincent, M., Adamczewski, J., Egly, J.M. and Bensaude, O. (1997). Heat shock inactivation of the TFIIH-associated kinase and change in the phosphorylation sites on RNA polymerase II largest subunit. Nucleic Acids Res. , 25 , 694–700. Dubois, M.-F., Hovanessian, A.G. and Bensaude, O. (1991). Heat-shock-induced denaturation of proteins. Characterization of the insolubilization of the interferoninduced p68 kinase. J. Biol. Chem. , 266 , 9707–9711. Dubois, M.-F. and Bensaude, O. (1998). Phosphorylation of RNA Polymerase II C-
Protein kinase cascades involved in heat shock protein
77
Terminal Domain (CTD): A new control for heat-shock gene expression? Cell Stress & Chaperones , in press. Duncan, R.F. (1996). Translational control during heat shock. In: Translational control . Cold Spring Harbor Laboratory press, 271–293. Elion, E.A. (1995). Ste5: a meeting place for MAP kinases and their associates. Trends in Cell Biol. , 5 , 322–325. Engelberg, D., Zandi, E., Parker, C.S. and Karin, M. (1994). The yeast and mammalian Ras pathways control transcription of heat shock genes independently of heat shock transcription factor . Mol. Cell. Biol. , 14 , 4929–4937. Erdos, G., Lee, Y.J., Cho, J.M. and Corry, P.M. (1995). Heat-induced FGF gene expression in the absence of heat shock element correlates with enhanced AP-1 binding activity. J. Cell. Physiol. , 164 , 404–413. Fanger, G.R., Gerwins, P., Widmann, C., Jarpe, M.B. and Johnson, G.L. (1997). MEKKs, GCKs, MLKs, PAKs, TAKs, and tpls: upstream regulators of the c-Jun amino-terminal kinases? Cur. Op. in Gen. of Dev. , 7 , 67–74. Farrell, P.J., Balkow, K., Hunt, T. and Jackson, R.J. (1977). Phosphorylation of initiation factor eIF-2 and the control of protein synthesis. Cell , 11 , 187–200. Faux, M.C. and Scott, J.D. (1996). Molecular glue: kinase anchoring and scaffolded proteins. Cell , 85 , 9–12. Fawcett, T.W., Eastman, H.B., Martindale, J.L. and Hollbrook, N.J. (1996). Physical and functional association between GADD153 and CCAAT/enhancer-binding protein ß during cellular stress. J. Biol. Chem. , 271 , 14285–14289. Feigenblum, D. and Schneider, R.J. (1996). Cap-binding protein (eukaryotic initiation factor 4E) and 4E-inactivating protein BP-1 independently regulate cap-dependent translation. Mol. Cell. Biol. , 16 , 5450–5457. Freshney, N.W., Rawlinson, L., Guesdon, F., Jones, E., Cowley, S., Hsuan, J. and Saklavatla, J. (1994). Interleukin-1 activates a novel protein kinase cascade that results in the phosphorylation of Hsp27. Cell , 78 , 1039–1049. Fuchs, S.Y., Dolan, L., Davis, R.J. and Ronai, Z. (1996). Phosphorylation-dependent targeting of c-Jun ubiquitination by Jun N-kinase. Oncogene, 13 , 1531–1535. Gabai V.L., Meriin, A.B., Mosser, D.D., Caron, A.W., Rits, S., Shifrin, V.I. and Sherman, M.Y. (1997). Hsp70 prevents activation of stress kinases. A novel pathway of cellular thermotolerance. J. Biol. Chem. , 272 , 18033–18037. Gaestel, M., Benndorf, R., Hayess, K., Priemer, E. and Engel, K. (1992). Dephosphorylation of the small heat shock protein hsp25 by calcium/calmodulindependent (type 2B) protein phosphatase. J. Biol. Chem. , 267 , 21607–21611. Gaestel, M., Schröder, W., Benndorf, R., Lippman, C., Buchner, K., Hucho, F., Erdmann, V.A. and Bielka, H. (1991). Identification of the phosphorylation sites of the murine small heat shock protein hsp25. J. Biol Chem. , 266 , 14721–14724. Galcheva-Gargova, Z., Dérijard, B., Wu, I.-H. and Davis, R.J. (1994). An osmosensing signal transduction pathway in mammalian cells. Science , 265 , 806–808. Gounalaki, N. and Thireos, G. (1994). Yap1, a yeast transcriptional activator that mediates multidrug resistance, regulates the metabolic stress response. EMBO J., 13 , 4036–4041. Grande, M.A., van der Kraan, I., de Jong, L. and van Driel, R. (1997). Nuclear distribution of transcription factors in relation to sites of transcription and RNA polymerase II. J. Cell Science , 110 , 1781–1791. Groom, L.A., Sneddon, A.A., Alessi, D., Dowd, S. and Keyse, S.M. (1996). Differential regulation of the MAP, SAP and RK/p38 kinases by Pyst1, a novel cytosolic dual-
Molecular chaperones and folding catalysts
78
specificity phoshatase. EMBO J. , 15 , 3621–3632. Gross, M., Olin, A., Hessefort, S. and Bender, S. (1994). Control of protein synthesis by hemin. Purification of a rabbit reticulocyte hsp70 and characterization of its regulation of the activation of the hemin-controlled eIF-2(a) kinase. J. Biol Chem. , 269 , 22738– 22748. Guay, J., Lambert, H., Gingras-Breton, G., Lavoie, J.N., Huot, J. and Landry, J. (1997). Regulation of actin filament dynamics by p38 MAP kinase-mediated phosphorylation of heat shock protein 27. J. Cell Science , 110 , 357–368. Guesdon, F., Freshney, N., Waller, R.J., Rawlinson, L. and Saklavatla, J. (1993). Interleukin1 and tumour necrosis factor stimulate two novel protein kinases that phosphorylate the heat shock protein hsp27 and ß-casein. J. Biol. Chem. , 268 , 4236– 4243. Guyton, K.Z., Liu, Y., Gorospe, M., Xu, Q. and Hollbrook, N.J. (1996). Activation of mitogen-activated protein kinase by H2O2. J. Biol. Chem. , 271 , 4138–4142. Han, J., Lee, J.-D., Bibbs, L. and Ulevitch, R.J. (1994). A MAP kinase targeted by endotoxin and hyperosmolarity in mammalian cells. Science , 265 , 808–811. Harper, J.W. and Elledge, S.J. (1996). Cdk inhibitors in development and cancer. Curr. Opin. in Genet. and Dev. , 6 , 56–64. Hazzalin, C.A., Le Panse, R., Cano, E. and Mahadevan, L.C. (1998). Anisyomycin selectively desensitizes signalling components involved in stress kinase activation and fos and jun induction. Mol. Cell Biol. , 18 , 1844–1854. Hu, Y., Metzler, B. and Xu, Q. (1997). Discordant activation of stress-activated protein kinases or c-Jun NH2-terminal protein kinases in tissues of heat-stressed mice. J. Biol. Chem. , 272 , 9113–9119. Hunter, T. and Poon, R.Y (1997). Cdc37: a protein kinase chaperone? Trends in Cell Biol. , 7 , 157–161. Huot, J., Houle, F., Marceau, F. and Landry, J. (1997). Oxidative stress-induced actin reorganization mediated by the p38 mitogen-activated protein kinase/heat shock protein 27 pathway in vascular endothelial cells. Circulation Res. , 80 , 383–392. Huot, J., Houle, F., Spitz, D.R. and Landry, J. (1996). HSP27 phosphorylation-mediated resistance against actin fragmentation and cell death induced by oxidative stress. Cancer Res. , 56 , 273–279. Huot, J., Lambert, H., Lavoie, J., Guimond, A., Houle, F. and Landry, J. (1995). Characterization of 45-kDa/54-kDa HSP27 kinase, a stress-sensitive kinase which may activate the phosphorylation-dependent protective function of mammalian 27-kDa heat-shock protein HSP27. Eur. J. Biochem. , 227 , 416–427. Ito, H., Okamoto, K., Nakayama, H., Isobe, T. and Kato, K. (1997). Phosphorylation of aB-cystallin in response to various types of stress. J. Biol. Chem. , 272 , 29934–29941. Jiang, Y., Chen, C., Li, Z., Guo, W., Gegner, J.A., Lin, S. and Han, J. (1996). Characterization of the structure and function of a new mitogen-activated protein kinase (p38ß). J. Biol. Chem. , 271 , 17920–17926. Johnson, N.L., Gardner, A.M., et al. (1996). Signal transduction pathways regulated by mitogen-activated/extracellular response kinase kinase kinase induce cell death. J. Biol. Chem. , 271 , 3229–3237. Jordan, P., Cunha, C. and Carmo-Fonseca, M. (1997). The cdk7-cyclin H-Matl complex associated with TFIIH is localized in coiled bodies. Mol. Biol. Cell. , 8 , 1207–1217. Jurivich, D.A., Chung, J. and Blenis, J. (1991). Heat shock induces two distinct S6 protein kinase activities in quiescent mammalian fibroblasts. J. Cell. Physiol. , 148 , 252–259.
Protein kinase cascades involved in heat shock protein
79
Kamada, Y., Sung Jung, U., Piotrowski, J. and Levin, D. (1995). The protein kinase Cactivated MAP kinase pathway of Saccharomyces cerevisiae mediates a novel aspect of the heat shock response. Genes & Development , 9 , 1559–1571. Kato, K., Hasegawa, K., Goto, S. and Inaguma, Y. (1994). Dissociation as a result of phosphorylation of an aggregated form of the small stress protein, hsp27. J. Biol. Chem. , 269 , 11274–11278. Keyse, S.M. (1995). An emerging family of dual specificity MAP kinase phosphatases. Biochim. Biophys. Acta , 1265 , 152–160. Kharbanda, S., Bharti, A., Pei, D., Wang, J., Pandey, P., Ren, R., Weichselbaum, R., Walsh, C.T. and Kufe, D. (1996). The stress response to ionizing radiation involves cAbl-dependent phosphorylation of SHPTP1. Proc. Natl. Acad. Sci. USA , 93 , 6898– 6901. Kharbanda, S., Ren, R., Pandey, P., Shafman, T.D., Feller, S.M., Weichselbaum, R.R. and Kufe, D.W. (1995). Activation of the c-Abl tyrosine kinase in the stress response to DNA-damaging agents. Nature , 376 , 785–788. Knauf, U., Jakob, U., Engel, K., Buchner, J. and Gaestel, M. (1993). Stress- and mitogeninduced phosphorylation of the small heat shock protein Hsp25 by MAPKAP kinase 2 is not essential for chaperone properties and cellular thermoresistance. EMBO J. , 13 , 54–60. Knauf, U., Newton, E.M., Kyriakis, J. and Kingston, R.E. (1996). Repression of human heat shock factor 1 activity at control temperature by phosphorylation. Genes & Dev. , 10 , 2782–2793. Kobayashi, N. and McEntee, K. (1993). Identification of, cis and, trans components of a novel heat shock stress regulatory pathway in Saccharomyces cerevisiae. Mol. Cell. Biol. , 13 , 248–256. Kozma, S.C. and Thomas, G. (1994). p70s6k/p85s6k: mechanism of activation and role in mitogenesis. Semin. Cancer Biol. , 5 , 255–260. Kumar, S., MacLaughlin, M.M., McDonnell, P.C., Lee, J.C., Livi, G.P. and Young, P.R. (1995). Human mitogen-activated protein kinase CSBP1, but not CSBP2, complements a hog1 deletion in yeast. J. Biol. Chem. , 270 , 29043–29046. Kyriakis, J.M. and Avruch, J. (1996). Protein kinase cascades activated by stress and inflammatory cytokines. BioEssays , 18 , 567–577. Landry, J. and Huot, J. (1995). Modulation of actin dynamics during stress and physiological stimulation by a signalling pathway involving p38 MAP kinase and heatshock protein 27. Biochem. Cell Biol. , 73 , 703–707. Landry, J., Lambert, H., Zhou, M., Lavoie, J., Hickey, E., Weber, L.A. and Anderson, C.W. (1992). Human HSP27 is phosphorylated at serines 78 and 82 by heat shock and mitogen-activated kinases that recognize the same amino acid motif as S6 kinase II. J. Biol. Chem. , 267 , 794–803. Laszlo, A. (1992). The effects of hyperthermia on mammalian cell structure and function. Cell Prolifer. , 25 , 59–87. Lavoie, J., Lambert, H., Hickey, E., Weber, L.A. and Landry, J. (1995). Modulation of cellular thermoresistance and actin filament stability accompanies phosphorylationinduced changes in the oligomeric structure of heat shock protein 27. Mol. Cell. Biol. , 15 , 505–516. Lawrence, J.C. and Abraham, R.T. (1997). PHAS/4E-BPs as regulators of mRNA translation and cell proliferation. Trends in Biochem. Sci. , 22 , 345–349. Lee, J.C., Laydon, J.T, et al. (1994). A protein kinase involved in the regulation of inflammatory cytokine biosynthesis. Nature , 372 , 739–746.
Molecular chaperones and folding catalysts
80
Legagneux, V., Morange, M. and Bensaude, O. (1991). Heat shock increases turnover of 90 kDa heat shock protein phosphate groups in HeLa cells. FEBS Lett. , 291 , 359–362. Mahdani, H.D. and Fink, G.R. (1998). The riddle of MAP kinase signaling specificity. Trends in Genet. , 14 , 151–155. Maher, P.A. and Pasquale, E.B. (1989). Heat shock induces protein tyrosine phosphorylation in cultured cells. J. Cell Biol. , 108 , 2029–2035. Matsuda, S., Kawasaki, H., Moriguchi, T., Gotoh, Y. and Nishida, E. (1995). Activation of protein kinase cascades by osmotic shock. J. Biol. Chem. , 270 , 12781–12786. Matts, R.L., Hurst, R. and Xu, Z. (1993). Denatured proteins inhibit translation in heminsupplemented rabbit reticulocyte lysate by inducing the activation of the hemeregulated eIF-2 kinase. Biochemistry , 32 , 7323–7328. Matts, R.L., Xu, Z., Pal, J.K. and Chen, J.-J. (1992). Interactions of the heme-regulated eIF-2 kinase with heat shock proteins in rabbit reticulocyte lysates. J. Biol. Chem. , 267 , 18160–18167. Mazzoni, C., Zarzov, P., Rambourg, A. and Mann, C. (1993). The SLT2 (MPK1) MAP kinase homolog is involved in polarized cell growth in Saccharomyces cerevisiae. J. Cell Biol. , 123 , 1821–1833. McLaughlin, M.M., Kumar, S., McDonnell, P.C., Van Horn, S., Lee, J.C., Livi, G.P. and Young, P.R. (1996). Identification of mitogen-activated protein (MAP) kinaseactivated protein kinase-3, a novel substrate of CSBP p38 MAP kinase . J. Biol. Chem. , 271 , 8488–8492. Mehlen, P., Préville, X., Chareyron, P., Briolay, J., Klemenz, R. and Arrigo, A.-P. (1995). Constitutive expression of human hsp27, or human B-crystallin confers resistance to TNF- and oxidative stress-induced cytotoxicity in stably transfected murine L929 fibroblasts. J. Immunol. , 154 , 363–374. Mehlen, P., Schulze-Osthoff, K. and Arrigo, A.-P. (1996). Small stress proteins as novel regulators of apoptosis. Heat shock protein 27 blocks Fas/APO-1- and staurosporineinduced cell death. J. Biol. Chem. , 271 , 16510–16514. Meier, R., Rouse, J., Cuenda, A., Nebreda, A.R. and Cohen, P. (1996). Cellular stresses and cytokines activate multiple mitogen-activated-protein kinase kinase homologues in PC12 and KB cells. Eur. J. Biochem. , 236 , 796–805. Melville, M.W., Hansen, W.J., Freeman, B.C., Welch, W.J. and Katze, M.G. (1997). The molecular chaperone hsp40 regulates the activity of P58IPK, the cellular inhibitor of PKR. Proc. Natl. Acad. Set. USA , 94 , 97–102. Mimnaugh, E.G., Worland, P.J., Whitesell, L. and Neckers, L.M. (1995). Possible role for serine/ threonine phosphorylation in the regulation of the heteroprotein complex between the hsp90 stress protein and the pp60v-src tyrosine kinase. J. Biol. Chem. , 270 , 28654–28659. Miron, T., Vancompernolle, K., Vandecherckhove, J., Wilchek, M. and Geiger, B. (1991). A 25-kD inhibitor of actin polymerization is a low molecular mass heat shock protein. J. Cell Biol. , 114 , 255–261. Morley, S.J. and McKendrick, L. (1997). Involvement of stress-activated protein kinase and p38/RK mitogen-activated protein kinase signalling pathways in the enhanced phosphorylation of initiation factor 4E in NIH3T3 cells. J. Biol. Chem. , 272 , 17887– 17893. Mivechi, N.F. and Giaccia, A.J. (1995). Mitogen-activated protein kinase acts as a negative regulator of the heat shock response in NIH3T3 cells. Cancer Res. , 55 , 5512–5519. Morley, S.J. and McKendrick, L. (1997). Involvement of stress-activated protein kinase
Protein kinase cascades involved in heat shock protein
81
and p38/RK mitogen-activated protein kinase signalling pathways in the enhanced phosphorylation of initiation factor 4E in NIH3T3 cells. J. Biol. Chem. , 272 , 17887– 17893. Morrison, D.K. and Cutler, R.E. (1997). The complexity of Raf-1 regulation. Cur. Op. in Cell Biol. , 9 , 174–179. Mosser, D.D., Caron, A.W., Bourget, L., Denis-Larose, D. and Massie, B. (1997). Role of the human heat shock protein hsp70 in protection against stress-induced apoptosis. Mol. Cell Biol. , 17 , 5317–5327. Murtha-Riel, P., Davies, M.V., Scherer, B.J., Choi, S.-Y, Hershey, J.W.B. and Kaufman, R.J. (1993). Expression of a phosphorylation-resistant eukaryotic initiation factor 2bsubunit mitigates heat shock inhibition of protein synthesis. J. Biol. Chem. , 268 , 12946–12951. Nagata, K.-i. and Hall, A. (1996). The Rho GTPase regulates protein kinase activity. BioEssays , 18 , 529–531. Nover, L. (1991). Heat Shock Response . (Boca Raton: CRC Press). O’ Brien, T., Hardin, S., Greenleaf, A. and Lis, J.T. (1994). Phosphorylation of RNA polymerase II C-terminal domain and transcriptional elongation. Nature , 370 , 75–77. Ohnishi, T., Wang, X., Ohnishi, K., Matsumoto, H. and Takahashi, A. (1996). p53dependent induction of WAF-1 by heat treatment in human glioblastoma cells. J. Biol. Chem. , 271 , 14510–14513. Posas, F., Wurgler-Murphy, S.M., Maeda, T., Witten, E.A., Thai, T.C. and Saito, H. (1996). Yeast HOG1 MAP kinase cascade is regulated by a multistep phosphorelay mechanism in the SLN1-YPD1-SSK1 “two-component” osmosensor. Cell , 86 , 865– 875. Pratt, W.B. (1997). The role of the hsp90-based chaperone system in signal transduction by nuclear receptors signaling via MAP kinase. Annu. Rev. of Pharmacol. & Toxicol. , 37 , 297–326. Proud, C.G. (1996). p70 S6 kinase: an enigma with variations. Trends in Biochem. Sci. , 21 , 181–185. Raingeaud, J., Whitmarsh, A.J., Barren, T., Dérijard, B. and Davis, R.J. (1996). MKK3and MKK6-regulated gene expression is mediated by the p38 mitogen-activated protein kinase signal transduction pathway. Mol. Cell. Biol. , 16 , 1247–1255. Robinson, M.J. and Cobb, M.H. (1997). Mitogen-activated protein kinase pathways. Cur. Op. in Cell Biol. , 9 , 180–186. Rose, D.W., Wettenhall, R.E.H., Kudlicki, W., Kramer, G. and Hardesty, B. (1987). The 90-kilodalton peptide of the heme-regulated eIF-2_ kinase has sequence similarity with the 90-kilodalton heat shock protein. Biochemistry , 26 , 6583–6587. Rouse, J., Cohen, P., Trigon, S., Morange, M., Alonso-Llamazares, A., Zamanillo, D., Hunt, T. and Nebreda, A.R. (1994). A novel kinase cascade triggered by stress and heat-shock that stimulates MAPKAP Kinase-2 and phosphorylation of the small heat shock proteins. Cell , 78 , 1027–1037. Roy, B. and Lee, A.S. (1995). Transduction of calcium stress through interaction of the human transcription factor CBF with the proximal CCAAT regulatory element of the grp78/BiP promoter. Mol. Cell. Biol. , 15 , 2263–2274. Ruis, H. and Schüller, C. (1995). Stress signaling in yeast. BioEssays , 17 , 959–965. Samuel, C.E. (1993). The eIF-2 protein kinases, regulators of translation in eukaryotes from yeasts to humans. J. Biol. Chem. , 268 , 7603–7606. Schulte, T.W., Blagosklonny, M.V., Romanova, L., Mushinski, J.F., Monia, B.P., Johnston, J.F., Nguyen, P., Trepel, J. and Neckers, L.M. (1996). Destabilization of
Molecular chaperones and folding catalysts
82
Raf-1 by geldanamycin leads to disruption of the Raf-1-MEK-Mitogen-Activated Protein kinase signalling pathway. Mol. Cell. Biol. , 16 , 5839–5845. Schwenger, P., Bellosta, P., Victor, I., Basilico, C., Skolnik, E.Y. and Vilcek, J. (1997). Sodium salicylate induces apoptosis via p38 mitogen-activated protein kinase but inhibits tumor necrosis factor-induced c-Jun N-terminal kinase/stress-activated protein kinaseactivation. Proc. Natl. Acad. Sci. , 94 , 2869–2873. Sells, M.A. and Chernoff, J. (1997). Emerging from the Pak: the p21-activated protein kinase family. Trends in Cell Biology , 7 , 162–167. Seroz, T., Hwang, J.R., Moncollin, V. and Egly, J.M. (1995). TFIIH: a link between transcription, DNA repair and cell cycle regulation. Cur. Op. Gen. Dev. , 5 , 217–221. Shamu, C.E. (1998). Splicing: HACking into the unfolded-protein response. Curr. Biol. , 8 , 121–123. Sheikh-Hamad, D., Di Mari, J., Suki, W.N., Safirstein, R., Watts, B.A. and Rouse, D. (1998). p38 kinase activity is essential for osmotic induction of mRNA for HSP70 and transporter for organic solute betaine in Madin-Darby canine kidney cells. J. Biol. Chem. , 273 , 1832–1837. Sherman, M. and Goldberg, A.L. (1994). Heat shock-induced phosphorylation of GroEL alters its binding and dissociation from unfolded proteins. J. Biol. Chem. , 269 , 31479–31483. Sherman, M.Y. and Goldberg, A.L. (1993). Heat shock of Escherichia coli increases binding of dnaK (the hsp70 homolog) to polypeptides by promoting its phosphorylation. Proc. Natl. Acad. Sci. USA , 90 , 8648–8652. Shim, J., Lee, H., Park, J., Kim, H. and Choi, E.-J. (1996). A non-enzymatic p21 protein inhibitor of stress-activated protein kinases. Nature , 381 , 804–807. Shiozaki, K., Akhavan-Niaki, H., McGowan, C. and Russell, P. (1994). Protein phosphatase 2C, encoded by, ptc1+ is important in the heat shock response of Schizosaccharomyces pombe. Mol. Cell. Biol. , 14, 3742–3751. Shiozaki, K. and Russell, P. (1995). Counteractive roles of protein phosphatase 2C (PP2C) and a MAP kinase homolog in the osmoregulation of fission yeast. EMBO J. , 14 , 492–502. Shiozaki, K. and Russell, P. (1996). Conjugation, meiosis, and the osmotic stress response are regulated by Spcl kinase through Atfl transcription factor in fission yeast. Genes & Dev. , 10 , 2276–2288. Sithanandam, G., et al. (1996). 3pK, a new mitogen-activated protein kinase-activated protein kinase located in the small cell lung cancer tumor suppressor gene region. Mol. Cell. Biol. , 16 , 868–876. Sluss, H.K., Barren, T., Dérijard, B. and Davis, R.J. (1994). Signal transduction by tumor necrosis factor mediated by JNK protein kinases. Mol Cell Biol. , 14 , 8376–8384. Steinmetz, E.J. (1997). Pre-mRNA processing and the CTD of RNA polymerase II: The tail that wags the dog? Cell , 89 , 491–494. Stewart, M.J. and Thomas, G. (1994). Mitogenesis and protein synthesis: A role for ribosomal protein S6 phosphorylation? BioEssays , 16 , 809–815. Stokoe, D., Campbell, D.G., Nakielny, S., Hidaka, H., S.J., L., Marshall, C. and Cohen, P. (1992a). MAPKAP kinase-2: A novel protein kinase activated by mitogen-activated protein kinase. EMBO J. , 11 , 3985–3994. Stokoe, D., Caudwell, B., Cohen, P.T.W. and Cohen, P. (1993). The substrate specificity and structure of mitogen-activated protein (MAP) kinase-activated kinase-2. Biochem. J. , 296 , 843–849. Stokoe, D., Engel, K., Campbell, D.G., Cohen, P. and Gaestel, M. (1992b). Identification
Protein kinase cascades involved in heat shock protein
83
of MAPKAP kinase 2 as a major enzyme responsible for the phosphorylation of the small mammalian heat shock proteins. FEBS Letters , 313 , 307–313. Stone, E.M. and Pillus, L. (1996). Activation of a MAP kinase cascade leads to Sir3p hyperphosphorylation and strengthens transcriptional silencing. J. Cell Biol. , 135 , 571–583. Sun, H., Charles, C.H., Lau, L.F. and Tonks, N. (1993). MKP-1 (3CH134) an immediate early gene product, is a dual specificity phosphatase that dephosphorylates MAP kinase in vivo. Cell , 75 , 487–493. Suzuki, A., Sugiyama, Y., Hayashi, Y., Nyu-i, N., Yoshida, M., Nonaka, I., Ishiura, S.-I., Arahata, K. and Ohno, S. (1998). MKBP, a novel member of the small heat shock protein kinase. J. Cell Biol. , 140 , 1113–1124. Symons, M. (1996). Rho family GTPases: the cytoskeleton and beyond. Trends in Biochem. Sci. , 21 , 178–185. Tan, Y., Rouse, J., Zhang, A., Cariati, S., Cohen, P. and Comb, M.J. (1996). FGF and stress regulate CREB and ATF-1 via a pathway involving p38 MAP kinase and MAPKAP kinase-2. EMBO J. , 15 , 4629–4642. Tokiwa, G., Dikic, I., Lev, S. and Schlessinger, J. (1996). Activation of Pyk2 by stress signals and coupling with JNK signaling pathway. Science , 273 , 792–794. Torres, L.H., Martin, M.I., Garcia-Saez, J., Arroyo, M., Molina, M., Sanchez, M. and Nombela, C. (1991). A protein kinase gene complements the lytic phenotype of Saccharomyces cerevisiae lyt2 mutants. Mol. Microbiol. , 5 , 2845–2854. Treisman, R. (1996). Regulation of transcription by MAP kinase cascades. Curr. Op. in Cell Biol , 8 , 205–215. Uma, S., Hartson, S.D., Chen, J.-J. and Matts, R.L. (1997). Hsp90 is obligatory for the heme-regulated eIF-2 kinase to acquire and maintain an activable conformation. J. Biol. Chem. , 272 , 11648–11656. Valay, J.-G., Simon, M., Dubois, M.-F, Bensaude, O., Facca, C. and Faye, G. (1995). The KIN28 gene is required both for RNA polymerase II mediated transcription and phosphorylation of the Rpb1p CTD. J. Mol. Biol. , 249 , 535–544. van Delft, S., Coffer, P., Kruijer, W. and van Wijk, R. (1993). C-fos induction by stress can be mediated by the SRE. Bioch. Biophys. Res. Comm. , 197 , 542–548. van der Straten, A., Rommel, C., Dickson, B. and Hafen, E. (1997). The heat shock protein 83 (Hsp83) is required for Raf-mediated signalling in Drosophila. EMBO J. , 16 , 1961–1969. van Wijk, R., Welters, M., Souren, J.E.M., Ovelgonne, H. and Wiegant, F.A. C. (1993). Serum-stimulated cell cycle progression and stress protein synthesis in C3H1OT1/2 fibroblasts treated with sodium arsenite. J. Cell Physiol , 155 , 265–272. Venetianer, A., Dubois, M.-F., Nguyen, V.T., Seo, S.J., Bellier, S. and Bensaude, O. (1995). Phosphorylation state of RNA Polymerase II C-Terminal Domain (CTD) in Heat-Shocked Cells. Possible involvement of the stress activated MAP kinases. Eur. J. Biochem. , 233 , 83–92. von Manteuffel, S.R., Gingras, A.-C., Ming, X.-F., Sonenberg, N. and Thomas, G. (1996). 4E-BP1 phosphorylation is mediated by the FRAP-p70s6k pathway and is independent of mitogen-activated protein kinase. Proc. Natl. Acad. Sci. USA , 93 , 4076–4080. Vries, R.G.J., Flynn, A., Patel, J.C., Wang, X., Denton, R.M. and Proud, C.G. (1997). Heat shock increases the association of binding protein-1 with initiation factor 4E. J. Biol. Chem. , 272 , 32779–32784. Wang, X.-Z., et al. (1996). Signals from the stressed endoplasmic reticulum induce
Molecular chaperones and folding catalysts
84
C/EBP-homologous protein (CHOP/GADD153). Mol. Cell. Biol. , 16 , 4273–4280. Wang, X.Z. and Ron, D. (1996). Stress-induced phosphorylation and activation of the transcription factor CHOP (GADD153) by p38 MAP kinase. Science , 272 , 1347– 1349. Weeks, J.R., Hardin, S.E., Shen, J., Lee, J.M. and Greenleaf, A. (1993). Locus-specific variation in phosphorylation state of RNA polymerase II in vivo: correlations with gene activity and transcript processing. Genes & Dev. , 7 , 2329–2344. Welch, W.J. (1992). Mammalian stress response: cell physiology, structure/function of stress proteins, and implications for medicine and disease. Physiol. Rev. , 72 , 1063– 1081. Welihinda, A.A. Tirasophon, W., Green, S.R. and Kaufman, R.J. (1998). Protein serine/threonine phosphatase Ptc2P negatively regulates the unfolded-protein response by dephosphorylating Ire1p kinase. Mol. Cell Biol. , 18 , 1967–1977. Xia, Z., Dickens, M., Raingeaud, J., Davis, R.J. and Greenberg, M.E. (1995). Opposing effects of ERK and JNK-p38 MAP kinases on apoptosis. Science, 270 , 1326–1331. Xing, J., Ginty, D.D. and Greenberg, M.E. (1996). Coupling of the RAS-MAPK pathway to gene activation by RSK2, a growth factor-regulated CREB kinase. Science , 273 , 959–963. Yang, D., Tournier, C., Wysk, M., Lu, H.T., Xu, J., Davis, R.J. and Flavell, R.A. (1997). Targeted disruption of the MKK4 gene causes embryonic death, inhibition of c-Jun NH2-terminal kinase activation, and defects in AP-1 transcriptional activity. Proc. Natl. Acad, Sci. USA , 94 , 3004–3009. Yueh, A. and Schneider, R.J. (1996). Selective translation initiation by ribosome jumping in adenovirus-infected and heat-shocked cells. Genes & Dev. , 10 , 1557–1567. Zanke, B.W., Boudreau, K., Rubie, E., Winnett, E., Tibbles, L.A., Zon, L., Kyriakis, J., Liu, F.-F. and Woodgett, J.R. (1996). The stress-activated protein kinase pathway mediates cell death following injury induced by cis-platinum, UV irradiation or heat. Curr. Biol. , 6 , 606–613. Zarzov, P., Mazzoni, C. and Mann, C. (1996). The SLT2 (MPK1) MAP kinase is activated during periods of polarized cell growth in yeast. EMBO J. , 15 , 83–91. Zhou, M., Lambert, H. and Landry, J. (1993). Transient activation of a distinct serine protein kinase is responsible for 27-kDa heat shock protein phosphorylation in mitogen-stimulated and heat-shocked cells. J. Biol. Chem. , 268 , 35–43. Zigmond, S.H. (1996). Signal transduction and actin filament organization. Curr. Op. in Cell Biol. , 8 , 66–73. Zu, Y.L., Wu, F., Gilchrist, A., Ai, Y., Lbadia, M.E. and Huang, C.-K. (1994). The primary structure of a human MAP kinase activated protein kinase 2. Biochem. Biophys. Res. Commun. , 200 , 1118–1124.
5. THERMOTOLERANCE AND STRESS RESPONSE: POSSIBLE INVOLVEMENT OF KU AUTOANTIGEN G.C.LI*, L.LI, D.KIM, A.NUSSENZWEIG, S.-H.YANG, P.BURGMAN H.OUYANG and C.C.LING Department of Medical Physics and Department of Radiation Oncology, Memorial Sloan-Kettering Cancer Center, New York, NY 10021
1. Introduction 2. Thermotolerance: a transient phenomenon 2.1. Induction and Development of Thermotolerance 2.2. Characterization of Thermotolerant Cells 3. Thermotolerance and heat shock proteins 3.1. Thermotolerance and heat shock protein Hsp70 3.2. Heat Shock Proteins and RNA Splicing 3.3. Heat Shock Proteins and Apoptosis 3.4. Thermotolerance and heat shock protein Hsp27 4. Regulation of heat shock response: possible involvement of Ku Autoantigen 5. Requirement for Ku80 in growth and V(D)J recombination 6. Conclusion 7. References
1. INTRODUCTION Mammalian cells, when exposed to a non-lethal heat shock, can acquire a transient resistance to one or more subsequent exposures at elevated temperatures. This phenomenon has been termed thermotolerance (Gerner 1983; Gerner, et al., 1975; Henle, et al., 1978; Henle, et al., 1976). On the molecular level, heat shock activates a specific set of genes, the so-called heat shock genes, resulting in the preferential synthesis of heat shock proteins (Lindquist 1986; Lindquist, et al., 1988; Morimoto, et al., 1990). *Corresponding author
Molecular chaperones and folding catalysts
86
There is much data supporting the hypothesis that heat shock proteins play a key role in modulating the cellular response to heat shock and other environmental stress, and are involved in the development of thermotolerance (Landry, et al., 1989; Laszlo, 1988; Laszlo, et al., 1985; Li, 1985, Li, et al., 1992a; Morimoto, et al., 1994). We begin this review by summarizing our current knowledge about the induction of thermotolerance by heat shock and other stresses. We will examine the biochemical and molecular mechanisms underlying the induction of thermotolerance and consider the role that heat shock proteins play in its development and decay. Finally, we will discuss the possible involvement of Ku protein in the regulation of heat shock and other stress response of mammalian cells.
2. THERMOTOLERANCE: A TRANSIENT PHENOMENON 2.1. Induction and Development of Thermotolerance Exposure of mammalian cells in culture to temperatures above 40°C often leads to reproductive death. Although mammalian cells vary appreciably in their intrinsic thermal sensitivity (Raaphorst et al., 1979), clonogenic survival curves, when plotted as a function of duration of heating, are characterized by an initial shoulder region and then an exponential decrease (Dewey et al., 1980), both of which are dependent on the applied temperature and the duration of exposure. The kinetics of inactivating a wide variety of mammalian cells by heat shock has been analyzed in terms of Arrhenius plots, where 1/Do is plotted against the inverse of the absolute temperature (Do is the time required to reduce the survival on the exponential region to 37% of its initial value). For most cell lines, the Arrhenius plots exhibit two straight segments, with a break at or near 43°C (Bauer et al., 1979; Dewey et al., 1977; Westra, et al., 1971). Above that temperature, the activation energy is between 110 and 150 kcal/mol, consistent with the view that protein damage is responsible for cell death. At lower temperatures, the activation energy is around 300 to 400 kcal/mol. This change in slope of Arrhenius plots below 43°C was interpreted as a manifestation of the ability of the cells to develop thermotolerance (Li et al., 1980; Sapareto et al., 1978). Henle and Leeper (Henle and Leeper, 1976) and Gerner and Schneider (Gerner and Schneider, 1975) first showed that cultured mammalian cells exposed to a nonlethal heat treatment have the ability to develop resistance to a subsequent heat challenge. Thermotolerance can be induced by a short initial heat treatment at temperatures above 43°C followed by a 37°C incubation before the second heat challenge. Thermotolerance can also be induced during continuous heating at temperatures below 43°C (Gerweck, 1977, Harisiadis et al., 1977; Palzer et al., 1973; Sapareto et al., 1978). The effect of thermotolerance can be dramatic, increases in survival levels can be several orders of magnitude (Field et al., 1982; Gerner, 1983; Henle and Dethlefsen, 1978). The mechanism underlying thermotolerance may also be responsible for the variation in thermosensitivity; permanent heat resistance may simply be a genetic alteration of constitutive levels of macromolecules transiently induced in thermotolerant cells. Thermal sensitivity is affected by a number of factors: the cells’ proliferative and
Thermotolerance and stress response
87
nutritional status (Gerweck, 1977, Goldin et al., 1981; Hahn 1982), the thermal history, the heat fractionation and sequencing, and the conditions during recovery (Gerner 1983, Henle and Dethlefsen, 1978). The effects of the temperature and the duration of the first heat treatment on the subsequent expression of thermotolerance were studied in detail by many investigators (Li et al., 1982). In general, if the temperature of the priming dose was below 43°C, thermotolerance was nearly fully expressed during the initial treatment. On the other hand, temperatures of 43°C or higher did not permit the development of thermotolerance during this first heat exposure; a subsequent incubation at 37°C was required. Li and Hahn (1980) proposed an operational model of thermotolerance based on these observations. The authors suggest that thermotolerance can be divided into three complementary and sometimes competing processes: an initial event (“trigger”); the expression of resistance (“development”), and the gradual disappearance of resistance (“decay”). In this model, thermotolerance develops in at least two steps. First, the initial event converts normal cells to the triggered state with a rate constant k1. This process very likely involves the activation of the heat shock transcription factor, HSF1 (Lis et al., 1993, Morimoto 1993). Second, these triggered cells are converted to thermotolerant cells with a rate constant kg. Above 42.5°C, k2=0, the triggered cells remain sensitive. When transferred to 37°C, kg > 0, and cells develop thermotolerance. This thermotolerant state is characterized by the elevated expression of heat shock proteins, enhanced protection and faster recovery from thermal damage. Finally, thermotolerant cells can revert to the sensitive state with a rate constant k3, which is generally less than k2. An Arrhenius plot of the induction of thermotolerance as measured by the clonogenic survival assay is shown in Figure 1A; shown for comparison are similar plots for the induced synthesis of one of the heat shock proteins (Hsp70) (Figure 1B), and for the activation of the heat shock transcription factor HSF1 (Figure 1C). The slopes of these plots yield
Figure 1 Phenomenological model of thermotolerance development and decay (upper panel), and Arrhenius plots for induction of thermotolerance,
Molecular chaperones and folding catalysts
88
induction of heat shock protein Hsp70, and heat shock factor (HSF1). activation (lower panels). (A) Rate at which thermotolerance is induced, as a function of the duration of the triggering, is plotted against the inverse of the absolute temperature of the appropriate treatment; (B) Similar to (A), except that the time required to induce a maximum amount of Hsp70 was used to determine the ordinate. (C) The rate of maximum HSF1 binding activity to the heat shock element (HSE). of the rat heat shock promoter, determined by quantifying bands of HSE-HSF1 complexes that were obtained using the gel-mobility shift assay (Kim et al., 1995).
Table 1 Characterization of the thermotolerant mammalian cell: molecules, cellular structures or functions that are protected after a mild heat shock
Macromolecules, cellular structure and function
References
• DNA Polymerase activity (a)
(Chu, et al., 1987, Dikomey, et al., 1987)
Synthesis
(Van Dongen, et al., 1984)
rRNA synthesis
(Burdon 1986, Nover, et al., 1986)
Splicing
(Yost and Lindquist 1986)
Synthesis
(Hahn, et al., 1985, Mizzen and Welch 1988, Sciandra, et al., 1984)
Denaturation
(Lepock, et al., 1990, Nguyen, et al., 1989, Pinto, et al, 1991)
Removal of aggregates
(Kampinga, et al., 1989, Wallen, et al., 1990)
Con A capping
(Stevenson, et al, 1981)
Na+/K+ ATPase (b)
(Anderson and Hahn 1985, Burdon, et al., 1982)
Permeability
(Maytin, et al., 1990)
Insulin receptors
(Calderwood, et al., 1983)
Cytoskeletal reorganization
(Wiegant, et al., 1987)
cAMP levels
(Calderwood, et al., 1985)
• RNA
• Protein
• Membrane
• Cytoplasm
Thermotolerance and stress response ATP levels (c)
89
(Lunee and Cresswell 1983)
• Cell cycle progression
(Van Dongen, et al., 1984)
(a) Not observed by (Jorritsma, et al., 1986, Kampinga, et al., 1985) (b) Anderson and Hahn, 1985, observed thermotolerance development for the ouabain binding capacity only, the heat sensitivity of the ATP hydrolyzing activity was unchanged in thermotolerant cells. (c) Only in L5178Y-S cells, no resistance was observed in Ehrlich ascites cells.
a value of 120 kcal/mole for the activation energy of the associated rate-limiting reaction, suggestive of protein unfolding or protein denaturation. The similarity between these graphs and the Arrhenius plot for cell killing at temperatures above 43°C (Dewey et al., 1977, Westra and Dewey, 1971) strongly suggests a common genesis for the inactivation of cells by heat, the induction of thermotolerance, the induction of Hsp70 synthesis and HSF1 activation. 2.2. Characterization of Thermotolerant Cells When cells are heat-shocked, energy is absorbed throughout the cell, damaging virtually all cellular structures/functions, resulting in cell killing (Laszlo, 1992a).
Table 2 Comparison of the damage and resistance induced in the membrane fraction (PF) and the nuclear fraction of HeLa S3 cells by different resistance inducing agents
TTR10 damage to membranes (by agent) C
enhanced resistance of membranes
damage to nuclei (by agent)
enhanced resistance of nuclei
(1.0)
HTT 2.3
yes
yes
yes
yes
ATT 1.8
yes (a)
yes
no
no
DTT 2.5
n.d.
no
yes
yes
ETT
no
no
yes
yes
2.3
As a measure for the induced thermotolerance, the thermotolerance ratio is given (TTR10: ratio of heating times required to reduce survival to 10%). Thermotolerance is induced as follows: HeLa S3 cells were pretreated with heat (15 min, 44°C+5 hr, 37°C: HTT) , sodium arsenite (1 hr, 100 mM+5 hr, 37°C: ATT), diamide (1hr, 500 M+5 hr, 37°C: DTT) or ethanol (1 hr, 6%+4 hr, 37° C: ETT) (Burgman, et al., 1993, Kampinga, et al., 1995). C: control, non-tolerant cells, n.d.=not determined, (a): taken from (Yih, et al., 1991).
Molecular chaperones and folding catalysts
90
In thermotolerant cells, many of these structures/functions are more resistant to heatinduced modifications. The data in Table 1 show that thermotolerance develops for nuclear, cytoplasmic and membrane components. Thus, there seems to be no spatial preference in the cell for the expression of heat-induced thermotolerance. However, an increasing amount of data suggest that the development of thermotolerance can be more specific than previously assumed, depending on the agents used for induction. For example, when thermotolerance is induced by sodium arsenite, diamide or ethanol, thermotolerance only develops in cellular fractions that are damaged during the thermotolerance-inducing treatment (Table 2). The subcellular localization of the induced resistance appears to correspond with the localization of the damage induced during the tolerance-inducing treatment. For instance, Lunec and Cresswell (Lunec et al., 1983) observed an enhanced resistance of the ATPsynthesis in thermotolerant cells, but only in a cell line (L5178Y-S) in which the ATPsynthesis was impaired by the tolerance-inducing heat treatment. In another cell line (Ehrlich ascites) this treatment had no effect on the ATP-synthesis, and in the thermotolerant cells no enhanced resistance of this process was observed. In firefly luciferase-transfected cells, Nguyen et al., (Nguyen et al., 1989) found that heat treatments (42°C) that decreased the luciferase activity but not protein synthesis in these cells, only induced resistance for the luciferase activity but not for protein synthesis. On the other hand, after a priming treatment at a higher temperature (45°C), both activities were impaired; and, in the thermotolerant cells resistance against a subsequent heat treatment was observed for both the luciferase activity and protein synthesis. Anderson and Hahn (Anderson et al., 1985) showed that the correlation between damage and induced resistance was even true for multidomain proteins in which the domains differ in heat sensitivity. These authors, working on Na+/K+ATPase, reported that thermotolerance could be induced in a heat sensitive domain of the protein (the ouabain binding domain) without changing the heat sensitivity of a more heat resistant domain (the ATPhydrolyzing domain). Such a direct correlation between damage and tolerance opens the possibility to study the role of inactivation of different proteins/structures in hyperthermic cell killing.
3. THERMOTOLERANCE AND HEAT SHOCK PROTEINS 3.1. Thermotolerance and Heat Shock Protein Hsp70 The mechanism underlying thermotolerance is not well understood, although many studies suggest that the heat shock proteins (HSP) are involved in its development (Landry et al., 1982; Li et al., 1982; Subjeck et al., 1982). Qualitative evidence exists for a causal relationship between HSP synthesis and thermotolerance (Landry et al., 1982; Laszlo and Li, 1985; Li, 1985; Li and Werb, 1982; Subjeck et al., 1982): (i) heat shock induces transiently enhanced synthesis of HSP that correlates temporally with the development of thermotolerance; (ii) the persistence of thermotolerance correlates well with the stability of HSP; (iii) agents known to induce HSP induce thermotolerance; (iv) conversely, agents known to induce thermotolerance induce HSP (Amici et al., 1993;
Thermotolerance and stress response
91
Boon-Niermeijer et al., 1988; Boon-Niermeijer et al., 1986; Burgman et al., 1993; Crete et al., 1990; Hahn et al., 1985; Haveman et al., 1986; Henle et al., 1986; Kampinga et al., 1992; Laszlo, 1988; Lee et al., 1987; Li, 1983; Li et al., 1978; Li et al., 1986; Ritossa, 1962; Ritossa, 1963); and (v) stable heat-resistant variant cells express high levels of HSP constitutively (Laszlo and Li, 1985). Quantitatively, of the many HSP preferentially synthesized after heat shock in mammalian cells, the concentration of the 70-kDa heat shock protein (Hsp70) appears to correlate best with heat resistance, either permanent or transient (Laszlo and Li 1985, Li 1985, Li and Werb 1982). However, a good correlation between a 27-kDa heat shock protein (Hsp27) and thermal resistance also has been reported (Landry et al., 1989). In mammalian cells, three types of experiments (e.g., microinjection of affinitypurified anti-Hsp70 antibodies or Hsp70 protein, amplification of hsp70 promoter sequence) were performed before 1990 to vary the intracellular concentration of Hsp70 and to correlate this change with thermal-stress response (Johnston et al., 1988; Li, 1989; Riabowol et al., 1988). More recently, the expression of hsp70 under heterologous promoters has yielded additional insight into its structure and function. Transient expression of Drosophila Hsp70 in monkey COS cells demonstrated that Hsp70 accelerates the recovery of cell nucleoli after heat shock (Munro et al., 1984). Using a retroviral-mediated gene transfer technique, Li et al., have generated rat cell lines stably and constitutively expressing a cloned human hsp70 gene (Li et al., 1992a). These cell lines provide a direct means of studying the effects of Hsp70 expression on cell survival after heat shock. It is clearly demonstrated (Figure 2A) that the clonogenic survival of five independent pools of MVH-infected Rat-1 cells (constitutively expressing human Hsp70)
Figure 2 (A) Expression of human Hsp70 gene confers thermal resistance to
Molecular chaperones and folding catalysts
92
Rat-1 cells. Survival at 45°C of pooled MVH-infected (expressing human Hsp70), MV6-infected (vector control), and uninfected Rat-1 cells. Survivals from five pooled populations of MVH-infected cells independently derived from separate infection experiments are shown (open symbols). Each pool is derived by pooling 200–600 colonies. Survival values after 30-min heating at 45°C are clustered around 10% for all cells and are, therefore, omitted for clarity (redrawn from Li et al. PNAS, 1992). (B) Cells expressing antisense Hsc70 (designated A12 and A13). show increased thermal sensitivity. Cellular survivals after heat shock at 45 °C were determined by the colony formation assay.
were approximately 100-fold higher (for a 60 or 75 min, 45°C heat treatment) than that of the parental Rat-1 or the MV6-infected Rat-1 cells (vector control). Similar results were obtained for individually isolated clones of Rat-1 cells expressing human Hsp70 (Li et al., 1992a). Furthermore, cells expressing higher levels of human Hsp70 generally survive thermal stress better than cells expressing lower levels, and there appears to be a good correlation between levels of exogenous human Hsp70 expressed and the degree of thermal resistance (Li et al., 1992a). These data provide direct evidence for a causal relation between expression of a functional form of mammalian Hsp70 and survival of cells at elevated temperatures. However, one should note that production of Hsp70 is only part of the program of protein biosynthesis initiated after heat shock, and other components of this response might also enhance cell survival (see chapter Lindquist et al., for role of Hsp104 in yeast). It is generally believed that Hsc70, a constitutively expressed member of Hsp70 family, binds to nascent polypeptide chains to prevent misfolding and aggregation (Hightower et al., 1994; see chapter Welch et al.). It functions cooperatively with Hsp40 and CCT in the folding of nascent polypeptide chains in a large assembly in the cytoplasm of eukaryotic cells (Frydman et al., 1994). Hsc70 also functions to chaperone proteins to organelles such as nuclei, mitochondria and endoplasmic reticulum (Brodsky et al., 1994; Langer et al., 1994). However, experimental evidence for these roles of Hsc70 has been derived mostly from in vitro studies; in vivo experiments using mammalian cells are lacking. In order to understand the functional roles of Hsc70 in mammalian cells, Li and Hightower (Li et al., 1995) used a dexamethasone (Dex)inducible antisense RNA expression vector to generate pseudomutants of NIH3T3 cells containing reduced levels of Hsc70, and to examine and identify phenotypic alteration in these cells. These investigators observed that in Dex-treated cultures, hsc70 antisense RNA blocked the induction of Hsp70 by heat shock. On the other hand, Hsc70 protein synthesis did not decrease in either Dex-treated or untreated cells. Taken together, these data suggest that antisense RNA was either less effective in reducing hsc70 RNA in Dexinduced cultures, or that cells compensated by producing more hsc70 RNA. Alternatively, it is plausible that newly synthesized RNA was targeted effectively before it became translationally active. To verify this hypothesis, Dex-treated cultures were made quiescent via serum starvation, and then restimulated with serum to induce RNA and protein synthesis. Consistent with the above hypothesis, Hsc70 synthesis was blocked in cultures expressing antisense RNA (Li and Hightower 1995).
Thermotolerance and stress response
93
The role of Hsp70/Hsc70 in thermotolerance development was also examined using a cell line that constitutively expresses antisense Hsc70 RNA (Nussenzweig et al., 1993). The antisense RNA was designed to also target Hsp70, which is > 75% homologous to Hsc70. Similar to the finding of Li and Hightower (Li and Hightower, 1995), heatinduced Hsp70 expression is significantly reduced and delayed in the transfected cells relative to the parental Rat-1 cells. For example, after a 45°C, 15 min heat shock, increased Hsp70 protein level is apparent within 4 hours in Rat-1 cells, 2 to 4 hours earlier than in the antisense RNA-transfected cells. In contrast, there were no significantly detectable changes in Hsc70 level in these cell lines. This is also in agreement with previous attempts to regulate gene expression by antisense RNA in which a large excess of antisense over sense transcripts is often necessary to produce significant results. Consistent with the above, the antisense transfected cells are more heat-sensitive than Rat-1 cells (Figure 2B), and show a profoundly reduced thermotolerance level as well as a delay in thermotolerance development. It is well established that heat shock inhibits RNA and protein synthesis. This inhibition is reversible, and the transcriptional and translational activity recovers gradually when heated cells are returned to 37°C. Thermotolerant cells subjected to heat shock exhibit less translational inhibition (Liu et al., 1992; Mizzen et al., 1988). The time required for the recovery of RNA and protein synthesis upon returning the heated cells to 37°C is also found to be considerably shorter for thermotolerant cells than for control nontolerant cells (Black et al., 1989; Laszlo, 1992b; Liu et al., 1992). The role of Hsp70 in these processes was examined using
Table 3 Comparison of heat resistance and recovery from heat-induced inhibition in RNA and protein synthesis relative to control cells
Schematic diagrams of human hsp70 a
Hsp70b Deletion
Heat c Resistance
Recovery after d Heat Shock RNA
protein 0.94 ±0.073
intact hsp70
none
++++
0.86±0.11
hsp70
Bgl
120–428
+++
0.16 ±0.03 0.22 ±0.065
hsp70
Sma-p 438–618
–
0.13 ±0.046
0.15 ±0.03
hsp70
S-C
438–478
+
0.12 ± 0.033
0.22±0.12
hsp70
C-S
478–618
+
0.15 ± 0.025
0.20±0.12
Schematic diagrams of human hsp70 gene and its deletion derivatives are shown on the left with relevant restriction sites. B, BamHI; H, Hind III; S, SmaI; C, ClaII; Bg, Bgl II; nucleotides
Molecular chaperones and folding catalysts
94
encoding a hexapeptide substance P epitope are represented by the dark box sub-P. The ATPbinding and nucleolar/nuclear localization domains are indicated as a checkerboard or a thinhatched bar, respectively. bDeletion is given in codons (all deletions are in-frame). c++++: 104-fold increase in survival,+++: 102-to 103-fold increase, +: 10- to 102-fold increase, and -: no increase in survival compared to control Rat-1 and MV6 cells. Heat treatment used: 45° C, 60 min. dRecovery of RNA and protein synthesis at 8 hr after 45°C, 25 min heat shock was determined. The relative rate of RNA and protein synthesis was normalized to the unheated control group. Average values ± SD are presented here. At 8 hr after heat shock, the relative RNA synthesis of Rat-1 and MV6 cells is 0.24±0.09 and 0.16±0.03, respectively; the relative protein synthesis of Rat-1 and MV6 cells is 0.35±0.13 and 0.12 ± 0.09, respectively. The differences observed between M21 cells and other cells are statistically significant (p<0.01). On the other hand, the differences among Bgl, Rat-1, Sma, C-S, S-C and MV6 cells are not statistically significant (p-values around 0.4).
Rat-1 fibroblasts expressing a cloned human hsp70 gene, designated M21 (Li et al., 1992a). The constitutive expression of the human hsp70 gene in Rat-1 cells confers heat resistance as evidenced by the enhanced survival of heat-treated cells and resistance against heat-induced translational inhibition (termed translational tolerance) (Liu et al., 1992). In addition, after a 45°C, 25-minute heat treatment, the time required for RNA and protein synthesis to recover was considerably shorter in M21 cells than in control Rat-1 cells. These data demonstrate that the expression of human hsp70 in Rat-1 cells confers heat resistance and translational tolerance, and enhances the ability of cells to recover from heat-induced translational and transcriptional inhibition (Table 3). The effects of mutant human Hsp70 on the cellular heat sensitivity, transcriptional/translational activity after heat shock were studied (Li et al., 1995; Table 3). Neither a 4-bp out-of-frame deletion ( 21) nor an in-frame deletion of the nucleolar localization domain ( Sma) of human Hsp70 significantly affected the cells’ intrinsic heat sensitivity, heat-induced transcriptional/translational inhibition, or the subsequent recovery at 37°C. In contrast, cells expressing a mutant human Hsp70 missing the ATPbinding domain ( Bgl) were found to be resistant to both heat killing and translational inhibition. However, when the recovery kinetics of protein synthesis were evaluated, no significant differences were observed between Rat-1 cells and cells expressing any of these mutated forms of Hsp70. These studies provide strong support for a link between the expression of functional Hsp70 and protection of cellular procedures at elevated temperature. They further suggest that ATP-binding and/or hydrolysis by Hsp70 are dispensable in the Hsp70-mediated protection against thermal killing and translational inhibition. Perhaps Hsp70 lacking its ATP-binding domain can still bind to cellular proteins, stabilizing them and preventing their aggregation at elevated temperatures. On the other hand, because the mutant Hsp70 Bgl does not facilitate cells’ recovery from translational inhibition, ATP-binding and/or hydrolysis may be important in enabling the dissociation of Hsp70 from its substrates, or in facilitating the dissociation of aggregated protein complexes to restore their functional integrity. The ability of Hsp70 to protect cellular processes and enzymatic activities against heat
Thermotolerance and stress response
95
stress has been demonstrated in a series of experiments. For example, in vitro studies by Skowyra et al., (Skowyra et al., 1990) have shown that the E. coli Hsp70 homologue, DnaK, can protect RNA polymerase from heat-induced inactivation, and that thermally inactivated RNA polymerase can be reactivated by DnaK in a process that requires ATP hydrolysis. Furthermore, the thermal protection efficiency of DnaK is enhanced by the action of its partner proteins DnaJ and GrpE (Hendrick et al., 1993). Similarly, mitochondrial Hsp60 has been shown to have thermal protective functions through its ability to prevent the thermal inactivation of dihydrofolate reductase imported into the mitochondria (Martin et al., 1992). It was also demonstrated that the in vivo enzymatic activities of luciferase and bgalactosidase were protected in thermotolerant mouse and Drosophila cells against a heat challenge (Nguyen et al., 1989). To extend these studies further to test the possibility that the intranuclear microenvironment may affect the stability of proteins under heat shock, Michels et al., (Michels et al., 1995) compared the in situ thermal stability of a reporter protein localized in the nucleus or in the cytoplasm, using recombinant firefly luciferase targeted to the cytoplasm (cyt-luciferase) or the nucleus (nuc-luciferase), respectively. In each case, decreased luciferase activity and solubility were found in lysates from heat-shocked cells, indicating thermal denaturation in situ. However, nucluciferase was more susceptible suggesting that the microenvironment of an intracellular compartment may modulate the thermal stability of proteins. In thermotolerant cells, the thermal inactivation of the recombinant luciferases by a second heat shock occurred at a slower rate with a lesser effect for the nuc-luciferase (threefold) than for the cytluciferase (sevenfold). The heat-inactivated luciferases were partially reactivated during recovery after stress, suggesting the capacity of both the cytoplasmic and nuclear compartments to reassemble proteins from an aggregated state. To further test the hypothesis that Hsp70 plays a role in protecting cells from thermal stress, the reporter gene firefly luciferase was constitutively expressed in mammalian Rat-1 cells or M21 cells (Rat-1 cells expressing high level of human Hsp70 via stable transfection). The effects of 43°C heat shock on luciferase activity in these cells were determined; in parallel, the solubility of luciferase in heat-shocked cells were monitored by immunoblot analysis of the soluble and insoluble protein fractions using antisera against luciferase, respectively. Our data demonstrate that in cells expressing a high level of intact human Hsp70, the heat inactivation of luciferase enzyme activity is attenuated (Li et al., 1992b). Similarly, a decrease in the insoluble fraction of luciferase molecules in these cells is observed (Li et al., 1992b). These data support the hypothesis that Hsp70 protects cells either by stabilization and prevention of thermal denaturation of normal proteins, and/or by facilitating the dissociation of protein aggregates. 3.2. Heat Shock Proteins and RNA Splicing Heat shock proteins have also been implicated in protecting the splicing machinery from disruption by heat shock (Yost et al., 1986; Yost et al., 1988; Yost et al., 1991): high temperatures are found to result in the accumulation of mRNA precursors due to the block in splicing. However, in thermotolerant cells, splicing is not disrupted and mature mRNAs accumulate. The mechanism by which this process is protected in thermotolerant
Molecular chaperones and folding catalysts
96
cells is unclear. 3.3. Heat Shock Proteins and Apoptosis Heat induces cell death both in terms of loss of clonogenicity and apoptosis. Recently, it has been shown that thermotolerance protects against heat-induced apoptosis in human acute lymphoblastic leukemic T-cells (Mosser et al., 1992), as evidenced by a reduction in DNA fragmentation and in the fraction of cells with morphological appearance of apoptosis. However, whether heat shock proteins have a direct role in the protection against apoptosis is yet unknown. We studied heat-induced apoptosis and loss of clonogenicity in Rat-1 fibroblasts, thermotolerant Rat-1 (TT Rat-1), Rat-1 transfected with the human Hsp70 gene (M21) and Rat-1 transfected with the human c-myc protooncogene (Rat-1: myc). A significant fraction of 44°C heat-treated rat fibroblasts exhibited nuclear chromatin condensation, cytoplasmic contraction, loss of adhesion, dye exclusion, and DNA laddering (Figure 3, A-D). Relative to Rat-1, TT Rat-1 and M21 cells are heat-resistant, but Rat-1: myc cells are heat-sensitive, both in terms of apoptosis induction and clonogenic survival (Figure 3, E and F). This study suggests that apoptosis is an important mechanism of heat-induced cell killing in some cell lines, which is suppressed by an overexpression of the human HspTO gene, but enhanced by c-myc expression. These data also suggest that the effects of these genes on thermal sensitivity is in part due to their modulation of apoptosis; although, other factors as yet unknown may also contribute. Vidair and Dewey reported that heated cells died by two distinct modes (Dewey 1989, Vidair et al., 1988). In the “slow” mode, postulated to be associated with chromosome aberration, multinucleation, and aberrant mitotic division, cells maintain their physiological activity for several days prior to the loss of reproductive capability
Thermotolerance and stress response
97
Figure 3 Induction of apoptosis in Rat-1, thermotolerant Rat-1 (TT Rat-1), and Rat-1 cells transfected with human Hsp70 (M21). or c-myc protooncogene (Rat-1: myc). In (A) and (B), cells were heated at 44°C for 30 min and studied 12 hours after heating. Cells were stained with 10 mg/ml acridine orange and studied under fluorescence microscope at 200X. (A) Detached cells: most cells are apoptotic as evidenced by chromatin condensation or nuclear fragmentation. (B) Attached cells: most cells are non-apoptotic. (C) DNA gel electrophoresis. Lane 1: Rat-1: myc, unheated control; lane 2: Rat-1: myc, attached cells, 12 hours after heating at 44°C for 30 min; and lane 3: Rat-1: myc, detached cells, same heat treatment as for lane 2. Equal amount of DNA (3 mg). was loaded per lane. (D) DNA gel electrophoresis of control and heated (24 hours after heating at 44°C for 40 min). Rat-1, TT Rat-1 and M21 cells. Lanes 1 and 2: heated M21 and Rat-1 cells; lanes 3 and 4: control M21 and Rat-1 cells; lanes 5 and 6: heated Rat1 and TT Rat-1 cells; lane 7: control TT Rat-1 cells. The attached and detached cells were pooled together for DNA gel electrophoresis
Molecular chaperones and folding catalysts
98
analysis (3 mg DNA was loaded per lane). (E) Surviving fraction of Rat-1, TT Rat-1, M21 and Rat-1: myc cells heated for various periods at 44°C. □-, Rat-1; -●-, TT Rat-1; -■-, M21; and -O-, Rat-1: myc. (F) Dose response of heat-induced apoptosis assayed 24 hours after heating for Rat-1, TT Rat-1, M21 and Rat-1: myc cells. -□-, Rat-1; -●-, TT Rat1; -■-, M21; and -O-, Rat-1: myc. As in (D), cells were pooled and used for the analysis.
(Dewey, 1989; Vidair and Dewey, 1988; Vidair et al., 1991). In the “rapid” mode, the underlying mechanism which is as yet unknown, cell death occurs within a few days after heating and is characterized by cell detachment from the culture surface. That the similar time course and the same loss of adhesion are observed for both this “rapid mode” and heat-induced apoptosis suggests further investigation to assess whether a common mechanism underlies these two phenomena. Recently, Strasser and Anderson have shown that both Bcl-2 gene expression and thermotolerance protect cells from heat-induced apoptosis and promote cell proliferation after thermal stress (Strasser et al., 1995). Based on these results, the authors suggested that heat shock induces apoptosis by two independent pathways, one inhibited by Bcl-2 and the other by the heat shock proteins. Our finding that Hsp70 overexpression reduces heat-induced apoptosis complements their findings. It has been reported that Hsp70 can interact with and alter the conformation of p53 protein (Hainaut et al., 1992), although the biological significance of p53-Hsp70 complexes is far from understood. Given the current belief that the induction of apoptosis requires p53 (Lowe et al., 1993, Lowe et al., 1993), the putative inactivation of p53 by Hsp70 may underly the observed decrease in apoptosis in M21 and IT Rat-1 cells. The protein product of the c-myc proto-oncogene is a transcription factor which has strong mitogenic effect. The induction of apoptosis by cmyc is believed to be due to a conflict of growth signals centered upon the G1/S transition (Evan et al., 1992; Shi et al., 1992), although the exact mechanism is largely unknown. It has also been reported that overexpression of c-myc gene enhances the induction of apoptosis by other insults, e.g. IL-3 deprivation or ionizing radiation (Askew et al., 1991; Clarke et al., 1993). Thus, our observation of the effect of cmyc in enhancing heat-induced apoptosis is consistent with the previous findings from other laboratories. 3.4. Thermotolerance and Heat Shock Protein Hsp27 In addition to Hsp70, the small heat shock protein, Hsp27, has been shown to have thermal protective functions. Landry and co-workers have overexpressed Hsp27 from various species into different cell lines and have observed that resistance to heat shock correlates with levels of Hsp27 (Huot et al., 1991; Lavoie et al., 1993). In addition it appears that phosphorylation of Hsp27 may play an important role in its thermal protective function. In support of this, Crete and Landry (Crete and Landry, 1990) observed that chemical agents such as cycloheximide, A23187 and EGTA, which induce phosphorylation but not accumulation of Hsp27, resulted in a significant degree of thermal protection. Furthermore, phosphorylation mutants of Hsp27 failed to protect cells from heat stress (Landry et al., 1993). However, this issue remains controversial because of recent evidence that the chaperone properties of the small heat shock protein
Thermotolerance and stress response
99
contributes to the increased cellular thermoresistance in a phosphorylation-independent manner (Knauf et al., 1994). This study suggests that the phosphorylation dependent function of Hsp27 is distinct from its thermoresistance-mediating functions.
4. REGULATION OF HEAT SHOCK RESPONSE: POSSIBLE INVOLVEMENT OF KU AUTOANTIGEN Although the importance of the heat shock transcription factor HSF1 in the regulation of mammalian heat shock genes is well-established (Morimoto et al., 1990; see chapter by R.Morimoto), recent data indicate that activation of HSF1, by itself, is not sufficient for the induction of Hsp70 mRNA synthesis, suggesting the existence of additional regulatory factors (Jurivich et al., 1992; Liu et al., 1993; Mathur et al., 1994). A constitutive heat shock element binding factor (CHBF) was recently characterized in rodent cells that exhibits a tight inverse correlation with the HSE-binding activity of HSF1 in vitro (Kim et al., 1995; Liu et al., 1993). Upon heat shock, a rapid increase in the level of HSF1-HSE binding activity correlates with a rapid decrease in CHBF-HSE binding activity. During post-heat shock-recovery, as rates of heat shock gene transcription return to their pre-heat-shock levels, the disappearance of HSF1-HSE binding activity parallels the reappearance of CHBF-HSE binding activity. Additionally, HSF1-HSE binding is not sufficient for heat shock gene induction; agents such as sodium salicylate (Jurivich et al., 1992) and arsenite (Liu et al., 1993) elicit considerable HSF1HSE binding activity, yet do not result in significant induction of Hsp70 mRNA synthesis. Interestingly, unlike heat shock stimuli, these agents do not inactivate the CHBF activity, which remains at the high levels observed in untreated cells (Liu et al., 1993). These findings suggest that CHBF may be involved in the regulation of heat shock gene expression, possibly by acting as a represser (Liu et al., 1993). Further biochemical analysis of CHBF revealed that it is identical or closely related to the Ku autoantigen (Kim et al., 1995). The Ku autoantigen is a heterodimer of 70-kDa (Ku70) and 86-kDa (Ku80) polypeptides (Mimori et al., 1981, Reeves et al., 1989). In vitro studies have shown that this abundant nuclear protein binds to the termini of doublestranded DNA and DNA ending in stem-loop structures (Mimori et al., 1986). By virtue of its DNA binding activity, Ku serves as a regulatory component of the mammalian DNA-dependent protein kinase, DNA-PK, which is activated by DNA ends (Gottlieb et al., 1993). The other component of DNA-PK is a 460-kDa catalytic subunit (DNA-PKcs). Various cellular roles for Ku have been suggested, including a role in transcription, recombination, replication, and DNA repair (Amabis et al., 1990; Bannister et al., 1993; Dvir et al., 1993; Falzon et al., 1993; Gottlieb and Jackson, 1993; Lees-Miller et al., 1991; Lees-Miller et al., 1990; Reeves, 1992; Yaneva et al., 1986). However, until recently there is a paucity of experimental data on the in vivo function of Ku. With the finding that Ku80 complements the DNA double-strand break rejoining deficiency and V (D)J recombination defect in radiation-sensitive CHO mutant cell lines, a role for Ku in the repair of DNA double strand breaks and V(D)J rearrangement is now established (Smider et al., 1994, Taccioli et al., 1994). In another report, the 460-kDa catalytic subunit of DNA-PK was linked to the V(D)J recombination and DNA repair machinery,
Molecular chaperones and folding catalysts
100
and shown to be the product of the murine gene SCID (Kirchgessner et al., 1995). To test whether the Ku-protein plays a role in the heat shock response, the intracellular concentration of Ku was modulated through the use of retroviral-mediated gene transfer technique. Coexpression of human Ku70 and Ku80 or expression of only the Ku70 subunit in rodent cells specifically inhibits the induction of Hsp70 upon heat shock (Li et al., 1995; Yang et al., 1996) (Figure 4). On the other hand, expression of human Ku80 alone does not have this effect. Thermal induction of other heat shock proteins in all of the Ku-overexpressing cell lines appears not to be significantly affected (Figure 4), nor the state of phosphorylation or the DNA-binding ability of HSF1 (Figure 5) (Yang et al., 1996). While Ku80 is involved in DNA repair, the 70-kDa component of the mammalian Ku-protein, but not the Ku80 subunit, appears to be involved in the regulation of Hsp70 gene expression. Thus, different subunits of Ku or different Ku-containing complexes may be involved in distinct biological processes. The fact that Ku is involved in gene regulation is not without precedent. Both the 70and 86-kDa subunits of Ku contain leucine/serine repeats which are reminiscent of similar “zipper” motifs found in a variety of transcription factors. Furthermore, Ku is a component of DNA-PK which phosphorylates several transcription factors, such as SP1 (Gottlieb and Jackson 1993), c-Jun (Bannister et al., 1993), p53 (Lees-Miller et al., 1990), c-Myc, Oct-1, and Oct-2 (Lees-Miller and Anderson 1991) in vitro. Studies also indicate that the Ku/DNA-PKcs complex modulates RNA polymerase I mediated transcription (Hoff et al., 1993; Knuth et al., 1990; Kuhn et al., 1995; Kuhn et al., 1993). Additionally, Ku has been found to be localized on certain transcriptionally active loci of chromosomal DNA (Amabis et al., 1990; Reeves, 1992). These data support a role for DNA-PK/Ku in transcription; specific genes regulated by DNA-PK/Ku need yet to be identified.
5. REQUIREMENT FOR KU80 IN GROWTH AND V(D)J RECOMBINATION It is well established that DNA double-strand break repair is important in maintaining cells’ genomic integrity. To determine the role of Ku80 in vivo, Ku80 was targeted in mice (Nussenzweig et al., 1996). Ku80 was inactivated by deleting 3.4 kilobases of the mouse Ku80 locus, including 100 base pairs of the promoter and the first two exons. Independently targeted embryonic stem (ES) cell clones were used to generate Ku80mutant mice. Heterozygous Ku80+/- mice were indistinguishable from their wild type littermates in all aspects, and 25% of the offspring born from Ku80+/- x Ku80+/- crosses were Ku80-/-. In contrast, new born Ku80-/- mice were significantly smaller than their wild type and heterozygous littermates (Figure 6). The growth retardation, as determined by weight difference, was first noted at embryonic day 15.5 (e15.5), and became statistically significant at e17.5. This difference increased to 40% at birth. During a 6month observation period, Ku80- /- mice grew and maintained body weights at 40–60% of controls, remaining proportional dwarfs. The reduced body weight in Ku80-/- mice does not appear to be a consequence of cell size, since comparisons of cell dimensions in kidney, muscle and liver sections showed no significant differences. Consistent with this
Thermotolerance and stress response
101
growth defect, fibroblasts derived from Ku80-/- embryos showed an early loss of proliferating cells, a longer doubling time, and intact cell-cycle checkpoints that prevented
Figure 4 (left) Overexpression of Ku70 or Ku70 and Ku80 jointly specifically
Molecular chaperones and folding catalysts
102
suppresses heat-induced Hsp70 expression. (A) Northern analysis of Hsp70 mRNA during post-heat-shock recovery at 37°C. Rat-1 cells, R7080– 6 cells (overexpressing human Ku-70 and Ku-80 protein jointly), and R70–15 cells (overexpressing human Ku-70 protein), were exposed to 45°C for 15 min, and returned to 37°C for 2–10 hr. Equal amounts of total RNA (10 mg). were loaded per lane, size fractionated on an agarose gel, transferred to Hybond-N membrane (Amersham), and probed with the human Hsp70 gene. C: no heat shock; 2, 4, 6, 8, 10, recovery time (hr). at 37°C. The heat-inducible Hsp70 mRNA is indicated by an arrowhead on the right margin. Note that the level of Hsp70 mRNA in heat-shocked R7080–6 and R70–15 cells is significantly reduced relative to that in Rat-1 cells. The same membrane was subsequently probed with a -actin gene. The levels of actin mRNA at all time points are not significantly different for all cell lines (see (B) below). The patterns of thermal induction of Hsp70 mRNA in R80–1 cells (overexpressing the human Ku-80 protein). are similar to the Rat-1 cells (data not presented). (B) Northern analysis of Hsc70, Hsp90, and Hsp27 mRNAs during post-heat-shock recovery. Rat-1 cells and R7080–6 cells were exposed to 45°C for 15 min, and returned to 37°C. Equal amounts of total RNA (10mg). were loaded per lane, size fractionated, and transferred to membranes as described in (A). The same membranes were probed sequentially with Hsp70, Hsc70, Hsp27, and Hsp90 and b-actin probes. C: no heat shock; H: heat-shocked. Total RNAs from control and heat-shocked cells (extracted 6–8 hr after heat shock), were shown. The levels of actin mRNA, used as a control, remain relatively constant after heat shock. The heat-induction of Hsc70 mRNA, Hsp27 mRNA and Hsp90 mRNA are clearly shown for both Rat-1 cells and R7080–6 cells. In contrast, the induction of Hsp70 mRNA in R7080–6 cells is significantly repressed.
cells with damaged DNA from entering the cell-cycle. This observation is in agreement with the correlation between senescence and defective DNA repair (Weirich-Schwaiger et al., 1994). Histologically, with the exception of lymphoid organs, Ku80−/− mice appear to be normal. However, spleen and lymph nodes were 5–10 fold smaller than controls and were devoid of lymphocytes. The thymus was also disproportionally small and had no corticalmedullary boundaries. Consistent with the histology, there was a complete absence of mature T and B cells. Both T and B lymphocyte development are arrested at early progenitor stages. Transient transfection experiments in mutant cell lines have shown that Ku80 is required for the formation of coding and signal joint during V(D)J recombination (Smider et al., 1994; Taccioli et al., 1994). To determine whether the KuSO−/− mutation affects rearrangements of antigen-receptor gene segments in T and B lymphocytes in vivo, immunoglobulin and T-cell antigen receptor (TCR) rearrangements were measured by polymerase chain reaction (PCR) (Carroll et al., 1991; Carroll et al., 1993; Costa et al., 1992; Schlissel et al., 1993). This analysis revealed that there is a profound deficiency in V(D)J rearrangement at either immunoglobulin or TCR loci in KuSO−/−
Thermotolerance and stress response
103
mice.
Figure 5 Analysis of the HSF1-HSE binding activity and HSF1 phosphorylation in Ku-overexpressing cells. Monolayers of Rat-1 cells and R7080–6 cells were exposed to 45°C for 15 min, and returned to 37°C for 0–10 hr. Equal amounts of cell extracts were subjected to gel mobility shift analysis (upper panel), or immunoblot analysis (lower panel), using HSF1-specific antiserum. Upper panel: The position of HSF1-HSE binding complex is indicated by an arrow (HSF1). The CHBF/Ku binding activity is indicated by an arrowhead, and is much greater in cells overexpressing human Ku-protein. Lower panel: the level of phosphorylation of HSF1 during heat shock and subsequent recovery was examined by Western analysis with antiserum specific to HSF1, taking advantage of the known reduction in the electrophoretic mobility of HSF1 in SDS-polyacrylamide gel upon its phosphorylation (Sarge et al., 1993). The molecular sizes are indicated in kilodaltons. The heat-induced HSE binding activity and hyperphosphorylation of HSF1 are unaffected by the expression of the human Ku protein.
6. CONCLUSION Research from the past decade has made it clear that heat shock proteins are required for the development of thermotolerance. The fact that mammalian cells become permanently thermoresistant when transfected with Hsp70, and conversely, that cells become thermal
Molecular chaperones and folding catalysts
104
sensitive when Hsp70 levels are reduced, directly demonstrates that Hsp70 plays a vital role in thermo tolerance. Hsp70 acts by stabilizing and preventing thermal denaturation of proteins, and by facilitating the dissociation of protein aggregates that are formed during conditions of stress. Other HSPs, such
Figure 6 Ku80 deficiency alters both pre- and post- natal growth (A) Photograph of 3 week old Ku80 −/− and Ku80 +/− mice (there is no significant difference in size between Ku80 +/− and Ku8 +/+ mice, not shown). (B) Growth of KuSO /− and Ku80 +/− embryos, as measured by weight. Pregnant females from KuSO +/− intercrosses were sacrificed at e13.5, e15.5, and e17.5 and 5–10 embryos of each genotype were weighed. The average weights of wild type (filled bars), heterozygous (hatched bars), and homozygous (white bars) embryos, and newborn mice are plotted. Error bars represent the standard deviation from the mean for each genotype. Statistical analysis using Student’s t test (p<0.05). showed significant differences in mean values between the three genotypes starting at e17.5, but not at earlier stages. (C) Post-natal growth of three female Ku80 +/− and three female Ku80−/− littermates. Weights of individual animals are plotted vs. age (in days). (D) Growth kinetics of primary fibroblasts. 105 passage two MEFs were plated in
Thermotolerance and stress response
105
replicate 60 mm dishes and individual dishes trypsinized and counted every 24 hours as indicated. Doubling time was calculated to be 41 hours for Ku80 /− compared to 23 hours for Ku80 +/−, based on the exponential portion of the growth curve. Ku80 −/− MEF’s reached a four-fold lower saturation density than controls (2.66×104/cm2 for Ku80 /− compared to 1.1×105/cm2 for Ku80 +/−. Wild type and Ku80 +/− MEFs were indistinguishable (not shown). Data redrawn from Nussenzweig et al., (Nussenzweig et al., 1996)
as Hsp27, may also have thermal protective functions, and there may even be cooperative action among members of the heat shock protein family during thermotolerance development, as has been shown for the Hsp70 and Hsp60 chaperones during the folding of denatured proteins (Hendrick and Hartl, 1993). Studies both in vitro and in vivo have shown that chaperones have the ability to recognize structures that are exposed in unfolded proteins, protect polypeptides from aggregation or premature folding, and promote the refolding of proteins after denaturation. These folding and unfolding reactions are critical both during normal cellular biogenesis and during metabolic stress. Under various conditions of stress the increased need to repair denatured proteins and protect vital structures is fulfilled by the enhanced and preferential synthesis of chaperones. The importance of heat shock factor HSF1 in the regulation of mammalian heat shock gene expression is well-established. However, recent data indicate that activation of HSF1, by itself, is not sufficient for the induction of Hsp70mRNA synthesis. We have previously proposed a dual control mechanism for the regulation of Hsp70 gene expression in mammalian cells: a positive control mechanism mediated by HSF1, and a negative control mechanism mediated by CHEF. CHBF was found to be identical to the Ku-autoantigen. In rodent cell lines that stably and constitutively overexpress the human Ku subunits Ku70 and Ku80 jointly, or Ku 70 alone, the induction of Hsp70 upon heat shock is specifically suppressed (Li et al., 1995; Yang et al., 1996). Thermal induction of the other heat shock proteins appears not to be affected, nor is the state of phosphorylation or the DNA-binding ability of HSF1 affected. These data suggest a role for Ku in regulating Hsp70 expression. Accumulating experimental evidence implicates that Ku plays an important role in modulating cellular response to stresses, e.g., heat shock and ionizing radiation. Under normal growth conditions, Ku is linked to V(D)J recombination, T and B lymphocyte development and growth control. A challenge for future studies will be to elucidate how Ku is functionally involved in these processes.
7. REFERENCES Amabis, J.M., Amabis, D.C., Kahuraki, J. and Stollar, B.O. (1990). The presence of an antigen reactive with a human autoantibody in Trichosia pubescens (Diptera: sciaridae). and its association with certain transcriptionally active regions of the genome. Chromosoma , 99 , 102–110. Amici, C., Palamara, T. and Santoro, M.G. (1993). Induction of thermotolerance by
Molecular chaperones and folding catalysts
106
prostaglandin A in human cells. Exp. Cell. Res. , 207 , 230–234. Anderson, R.L. and Hahn, G.M. (1985). Differential effects of hyperthermia on the Na+, K+-ATPase of Chinese hamster ovary cells. Radiat. Res. , 102 , 314–323. Askew, D.S., Ashmun, R.A., Simmons, B.C. and Clevelan, J.L. (1991). Constitutive cmyc expression in an IL-3-dependent myeloid cell line suppresses cell cycle arrest and accelerates apoptosis. Oncogene , 6 , 1915–1922. Bannister, A.J., Gottlieb, T.M., Kouzarides, T. and Jackson, S.P. (1993). c-Jun is phosphorylated by the DNA-dependent protein kinase in vitro; definition of the minimal kinase recognition motif. Nucl. Acids Res. , 21 , 1289–1295. Bauer, K.D. and Henle, K.J. (1979). Arrhenius analysis of heat survival curves from normal and thermotolerant CHO cells. Radiat. Res. , 78 , 251–263. Black, A.R. and Subjeck, J.R. (1989). Involvement of rRNA synthesis in the enhanced survival and recovery of protein synthesis seen in thermotolerance. J. Cell. Physiol , 138 , 439–449. Boon-Niermeijer, E.K., Souren, J.E.M., De Waal, A.M. and Van Wijk, R. (1988). Thermotolerance induced by heat and ethanol. Int. J. Hyperthermia , 4 , 211–222. Boon-Niermeijer, E.K., Tuyl, M. and Van der Scheur, H. (1986). Evidence for two states of thermotolerance. Int. J. Hyperthermia , 2 , 93–105. Brodsky, J.L. and Schekman, R. (1994). Heat shock cognate proteins and polypeptide translocation across the endoplasmic reticulum membrane. In R.I.Morimoto, A.Tissieres and C.Georgopoulos (eds.). The Biology of Heat Shock Proteins and Molecular Chaperones , Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, pp. 85–110. Burdon, R.H. (1986). Heat shock and the heat shock proteins. Biochem. J. , 240 , 313– 324. Burdon, R.H. and Cutmore, C.M.M. (1982). Human heat shock gene expression and the modulation of Na+K+-ATPase activity. FEBS Letters , 140 , 45–48. Burgman, P.W.J.J., Kampinga, H.H. and Konings, A.W.T. (1993). Possible role of localized protein denaturation in the induction of thermotolerance by heat, sodiumarsenite, and ethanol. Int. J. Hyperthermia , 9 , 151–162. Calderwood, S.K. and Hahn, G.M. (1983). Thermal sensitivity and resistance of insulinreceptor binding. Biochim. Biophys. Acta , 756 , 1–8. Calderwood, S.K., Stevenson, M.A. and Hahn, G.M. (1985). Cyclic AMP and the heat shock response in Chinese hamster ovary cells. Biochem. Biophys. Res. Commun. , 126 , 911–916. Carroll, A.M. and Bosma, M.J. (1991). T-lymphocyte development in scid mice is arrested shortly after the initiation of T-cell receptor d gene recombination. Genes Dev. , 5 , 1357–1366. Carroll, A.M., Slack, J.K. and Chang, W.-T. (1993). Biased T-cell receptor d element recombination in scid thymocytes. Mol. Cell. Biol. , 13 , 3623–3640. Chu, G.L. and Dewey, W.C. (1987). Effect of cycloheximide on heat-induced cell killing, radiosensitization, and loss of cellular DNA polymerase activities in Chinese hamster ovary cells. Radiat. Res. , 112 , 575–580. Clarke, A.R., Purdie, C.A., Harrison, D.J., Morris, R.G., Bird, C.C., Hooper, M.L. and Wylie, A.H. (1993). Thymocyte apoptosis induced by p53-dependent and independent pathways. Nature (London) , 362 , 849–852. Costa, T.E.F., Suh, H., Nussenzweig, M. (1992). Chromosomal position of rearranging gene segments influences allelic exclusion in transgenic mice. Proc. Natl. Acad. Sci. USA , 89 , 2205–2208.
Thermotolerance and stress response
107
Crete, P. and Landry, J. (1990). Induction of Hsp27 phosphorylation and thermoresistance in Chinese hamster cells by arsenite, cycloheximde, A23187, and EGTA Radiat. Res. , 121 , 320–327. Dewey, W.C. (1989). The search for critical cellular targets damaged by heat. Radiat. Res. , 120 , 191–204. Dewey, W.C., Freeman, M.L., Raaphorst, G.P., Clark, E.P., Wong, R.S., Highfield, D.P., Spiro, I.J., Tomasovic, S.P., Denman, D.L. and Coss, R.A. (1980). Cell biology of hyperthermia and radiation. In R.E.Meyn and H.R.Withers (eds.). Radiation Biology in Cancer Research , Raven Press, New York, pp. 589–623. Dewey, W.C., Hopwood, L.E., Sapareto, S.A. and Gerweck, L.E. (1977). Cellular responses to combination of hyperthermia and radiation. Radiology , 123 , 464–477. Dikomey, E., Becker, W. and Wielckens, K. (1987). Reduction of DNA-polymerase b activity of CHO cells by single and combined heat treatments. Int. J. Radiat. Biol. , 52 , 775–785. Dvir, A., Stein, L.Y., Calore, B.L. and Dynan, W.S. (1993). Purification and characterization of a template-associated protein kinase that phosphorylates RNA polymerase II. J. Biol. Chem. , 268 , 10440–10447. Evan, G.I., Wylie, A.H., Gilbert, C.S., Littlewood, T.D., Land, H., Brooks, M., Waters, C.M., Penn, L.Z. and Hancock, D.C. (1992). Induction of apoptosis in fibroblasts by cmyc protein. Cell , 69 , 119–128. Falzon, M., Fewell, J.W. and Kuff, E.L. (1993). EBP-80, a transcription factor closely resembling the human autoantigen Ku, recognizes single- to double-strand transitions in DNA. J. Biol. Chem. , 268 , 10546–10552. Field, S.B. and Anderson, R.L. (1982). Thermotolerance: a review of observations and possible mechanisms. Cancer therapy by hyperthermia, drugs and radiation. Nat. Cancer Inst. Monogr. , 61 , 193–201. Frydman, J., Nimmesgern, E., Ohtsuka, K. and Hartl, U. (1994). Folding of nascent polypeptide chains in a high molecular mass assembly with molecular chaperones. Nature (London) , 370 , 111–117. Gerner, E.W. (1983). Thermotolerance. In F.K.Storm (eds.). Hyperthermia in Cancer Therapy , G.K.Hall, Boston, pp. 141–162. Gerner, E.W. and Schneider, M.J. (1975). Induced thermal resistance in HeLa cells. Nature (London) , 256 , 500–502. Gerweck, L.E. (1977). Modification of cell lethality at elevated temperatures. The pH effect. Radiat. Res. , 70 , 224–235. Goldin, E.M. and Leeper, D.B. (1981). The effect of low pH on thermotolerance induction using fractionated 45°C hyperthermia. Radiat. Res. , 85 , 472–479. Gottlieb, T.M. and Jackson, S.P. (1993). The DNA-dependent protein kinase: requirement for DNA ends and association with Ku antigen. Cell , 72 , 131–142. Hahn, G.M. (1982) Hyperthermia and Cancer , Plenum Press, New York. Hahn, G.M., Shiu, E.G., West, B., Goldstein, L. and Li, G.C. (1985). Mechanistic implications of the induction of thermotolerance in Chinese hamster cells by organic solvents. Cancer Res. , 45 , 4138–4143. Hainaut, P. and Milner, J. (1992). Interaction of heat-shock protein 70 with p53 translated in vitro: evidence for interaction with dimeric p53 and for a role in the regulation of p53 conformation. Eur. Mol. Biol Organ. , 11 , 3513–3520. Harisiadis, L., Sung, D.I. and Hall, E.J. (1977). Thermal tolerance and repair of thermal damage by cultured cells. Radiology , 123 , 505–509. Haveman, J., Li, G.C., Mak, J.Y. and Kipp, J.B. (1986). Chemically induced resistance to
Molecular chaperones and folding catalysts
108
heat treatment and stress protein synthesis in cultured mammalian cells. Int. J. Radiat. Biol. , 50 , 51–64. Hendrick, J.P. and Hartl, F.-U. (1993). Molecular chaperone functions of heat-shock proteins. Annu. Rev. Biochem. , 62 , 349–384. Henle, K.J. and Dethlefsen, L.A. (1978). Heat fractionation and thermotolerance: a review. Cancer Res. , 38 , 1843–1851. Henle, K.J. and Leeper, D.B. (1976). Interaction of hyperthermia and radiation in CHO cells: recovery kinetics. Radiat. Res. , 66 , 505–518. Henle, K.J., Moss, A.J. and Nagle, W.A. (1986). Temperature-dependent induction of thermotolerance by ethanol. Radiat. Res. , 66 , 505–518. Hightower, L.E., Sadis, S.E. and Takenaka, I.M. (1994). Interactions of vertebrate Hsc70 and Hsp70 with unfolded proteins and peptides. In R.I.Morimoto, A.Tissieres and C.Georgopoulos (eds.). The Biology of Heat Shock Proteins and Molecular Chaperones , Cold Spring Harbor Laboratory Press, Plainview, pp. 179–207. Hoff, C.M. and Jacob, S.T. (1993). Characterization of the factor EIBF from a rat hepatoma that modulates ribosomal RNA gene transcription and its relationship to the human Ku autoantigen. Biochem. Biophys. Res. Commun. , 190 , 747–753. Huot, J., Roy, G., Lambert, H., Chretien, P. and Landry, J. (1991). Increased survival after treatments with anticancer agents of Chinese hamster cells expressing the human Mr 27,000 heat shock protein. Cancer Res. , 51 , 5245–5252. Johnston, R.N. and Kucey, B.L. (1988). Competitive inhibition of Hsp70 gene expression causes thermosensitivity. Science , 242 , 1551–1554. Jorritsma, J.B.M., Burgman, P., Kampinga, H.H. and Konings, A.W.T. (1986). DNA polymerase activity in heat killing and hyperthermic radiosensitization of mammalian cells as observed after fractionated heat treatments. Radiat. Res. , 105 , 307–319. Jurivich, D.A., Sistonen, L., Kroes, R.A. and Morimoto, R.I. (1992). Effect of sodium salicylate on the human heat shock response. Science , 255 , 1243–1245. Kampinga, H.H., Brunstig, J.F., Stege, G.J.J., Burgman, P.W.J.J. and Konings, A.W.T. (1995). Thermal protein denaturation and protein aggregation in cells made thermotolerant by various chemicals: Role of heat shock proteins. Exp. Cell. Res. , 219 , 536–546. Kampinga, H.H., Brunsting, J.F. and Konings, A.W.T. (1992). Acquisition of thermotolerance induced by heat and arsenite in HeLa cells: multiple pathways to induce tolerance? J. Cell. Physiol. , 150 , 406–415. Kampinga, H.H., Jorritsma, J.B.M., Konings, A.W.T. (1985). Heat-induced alterations in DNA polymerase activity of HeLa cells and of isolated nuclei. Relation to cell survival. Int. J. Radiat. Biol. , 47 , 29–40. Kampinga, H.H., Turkel-Uygur, N., Roti Roti, J.L. and Konings, A.W.T. (1989). The relationship of increased nuclear proteins content induced by hyperthermia to killing of HeLa S3 cells. Radiat. Res. , 117 , 511–522. Kim, D., Ouyang, H. and Li, G.C. (1995). Heat shock protein Hsp70 accelerates the recovery of heat-shocked mammalian cells through its modulation of heat shock transcription factor HSF1. Proc. Natl. Acad. Sci. USA , 92 , 2126–2130. Kim, D., Ouyang, H., Yang, S.-H., Nussenzweig, A., Burgman, P. and Li, G.C. (1995). A constitutive heat shock element-binding factor is immunologically identical to the Kuautoantigen. J. Biol. Chem. , 270 , 15277–15284. Kirchgessner, C.U., Patil, C.K., Evans, J.W., Cuomo, C.A., Fried, L.M., Carter, T., Oettinger, M.A. and Brown, J.M. (1995). DNA-dependent kinase (p350). as a candidate gene for the murine SCID defect. Science , 267 , 1178–1183.
Thermotolerance and stress response
109
Knauf, U., Jakob, U., Engel, K., Buchner, J. and Gaestel, M. (1994). Stress- and mitogeninduced phosphorylation of the small heat shock protein hsp25 by MAPKAP kinase 2 is not essential for chaperone properties and cellular thermoresistance. EMBO J., 13 , 54–60. Knuth, M.W., Gunderson, S.I., Thompson, N.E., Strasheim, L.A. and Burgess, R.R. (1990). Purification and characterization of proximal sequence element-binding protein 1, a transcription activating protein related to Ku and TREF that binds the proximal sequence element of the human U1 promoter. J. Biol. Chem. , 265 , 17911–17920. Kuhn, A., Gottlieb, T.M., Jackson, S.P. and Grummt, I. (1995). DNA-dependent protein kinase: a potent inhibitor of transcription by RNA polymerase I. Genes Dev. , 9 , 193– 203. Kuhn, A., Stefanovsky, V. and Grummt, I. (1993). The nucleolar transcription activator UBF relieves Ku-antigen-mediated repression of mouse ribosomal gene transcription. Nucl. Acid. Res. , 21 , 2057–2063. Landry, J., Bernier, D., Chretien, P., Nicole, L.M., Tanguay, R.M. and Marceau, N. (1982). Synthesis and degradation of heat shock proteins during development and decay of thermotolerance. Cancer Res. , 42 , 2457–2461. Landry, J., Chretien, P., Lambert, H., Hickey, E. and Weber, L.A. (1989). Heat shock resistance conferred by expression of the human HSP27 gene in rodent cells. J. Cell Biol. , 109 , 7–15. Landry, J., Lavoiem, J.N., Lambert, H., Hickey, E., Weber, L.A., Houle, F., Guay, J. and Guimond, A. (1993). Modulation of Hsp27 functions by phosphorylation, StressInduced Post-Translational Protein Modifications , Paris, France, 1993, O-40. Langer, T. and Neupert, W. (1994). Chaperoning mitochondrial biogenesis. In R.I.Morimoto, A. Tissieres and C.Georgopoulos (eds.). The Biology of Heat Shock Proteins and Molecular Chaperones , Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, pp. 53–84. Laszlo, A. (1988). The relationship of heat-shock proteins, thermotolerance, and protein synthesis. Exp. Cell. Res. , 178 , 401–414. Laszlo, A. (1992a). The effects on hyperthermia on mammalian cell structure and function. Cell Proliferation , 25 , 59–87. Laszlo, A. (1992b). The thermoresistant state: protection from initial damage or better repair? Exp. Cell. Res. , 202 , 519–531. Laszlo, A. and Li, G.C. (1985). Heat-resistant variants of Chinese hamster fibroblasts altered in expression of heat shock protein. Proc. Natl. Acad. Sci. USA , 82 , 8029– 8033. Lavoie, J.N., Bingras-Breton, G. and Landry, J. (1993). Induction of Chinese hamster Hsp27 gene expression in mouse cells confers resistance to heat shock. J. Biol. Chem. , 268 , 3420–3429. Lee, Y.J. and Dewey, W.C. (1987). Induction of heat shock proteins in Chinese hamster ovary cells and development of thermotolerance by intermediate concentrations of puromycin. J. Cell. Physiol. , 132 , 1–11. Lees-Miller, S.P. and Anderson, C.W. (1991). The DNA-activated protein kinase, DNAPK: a potential coordinator of nuclear events. Cancer Cells , 3 , 341–346. Lees-Miller, S.P., Chen, Y.-R. and Anderson, C.W. (1990). Human cells contain a DNAactivated protein kinase that phosphorylates simian virus 40 T antigen, mouse p53, and the human Ku autoantigen. Mol. Cell. Biol. , 10 , 6472–6481. Lepock, J.R., Frey, H.E., Heynen, M.P., NishioJ., Waters, B., Ritchie, K.P. and Kruuv, J. (1990). Increased thermostability of thermotolerant CHL V79 cells as determined by
Molecular chaperones and folding catalysts
110
differential scanning calorimetry. J. Cell Physiol. , 142 , 628–634. Li, G.C. (1983). Induction of thermotolerance and enhanced heat shock protein synthesis in Chinese hamster fibroblasts by sodium arsenite and by ethanol. J. Cell. Physiol. , 115 , 116–122. Li, G.C. (1985). Elevated levels of 70,000 dalton heat shock protein in transiently thermotolerant Chinese hamster fibroblasts and in their stable heat resistant variants. Int. J. Radiat. Oncol. Biol. Phys. , 11 , 165–177. Li, G.C. (1989). HSP 70 as an indicator of thermotolerance, 5th International Symposium on Hyperthermic Oncology , Kyoto, Japan, 1989, 256–260. Li, G.C., Fisher, G.A. and Hahn, G.M. (1982). Modification of the thermal response by D2O. II. Thermotolerance and the specific inhibition of development. Radiat. Res. , 92 , 541–551. Li, G.C. and Hahn, G.M. (1978). Ethanol-induced tolerance to heat and to adriamycin. Nature (London) , 274 , 699–701. Li, G.C. and Hahn, G.M. (1980). A proposed operational model of thermotolerance based on effects of nutrients and the initial treatment temperature. Cancer Res. , 40 , 4501– 4508. Li, G.C., Li, L., Liu, R.Y., Rehman, M. and Lee, W.M.F. (1992a). Heat shock protein Hsp70 protects cells from thermal stress even after deletion of its ATP-binding domain. Proc. Natl Acad. Set. USA , 89 , 2036–2040. Li, G.C., Liu, R.Y., Li, L., Shen, G. and Li, X. (1992b). Hsp70: Role in thermotolerance, 6th International Congress on Hyperthermic Oncology , Tucson, AZ, pp. 127–130. Li, G.C. and Mivechi, N.F. (1986). Thermotolerance in mammalian systems: a review. In L.J.Anghileri and J.Robert (eds.). Hyperthermia in Cancer Treatment , CRC Press, Inc., Boca Raton, pp. 59–77. Li, G.C. and Werb, Z. (1982). Correlation between synthesis of heat shock proteins and development of thermotolerance in Chinese hamster fibroblasts. Proc. Natl. Acad. Set. USA , 79 , 3218–3222. Li, G.C., Yang, S.-H., Kim, D., Nussenzweig, A., Ouyang, H., Wei, J., Burgman, P. and Li, L. (1995). Suppression of heat-induced Hsp70 expression by the 70-kDa subunit of the human Ku-autoantigen. Proc. Natl. Acad. Sci. USA , 92 , 4512–4516. Li, L., Shen, G. and Li, G.C. (1995). Effects of expressing human Hsp70 and its deletion derivatives on heat killing and RNA and protein synthesis. Exp. Cell. Res. , 217 , 460– 468. Li, T. and Hightower, L.E. (1995). Effects of dexamethasone, heat shock, and serum responses on the inhibition of Hsc70 synthesis by antisense RNA in NIH 3T3 cells. J. Cell. Physiol , 164 , 344–355. Lindquist, S. (1986). The heat-shock response. Annu. Rev. Biochem. , 55 , 1151–1191. Lindquist, S. and Craig, E.A. (1988). The heat shock proteins. Annu. Rev. Genet. , 22 , 631–677. Lis, J. and Wu, C. (1993). Protein traffic on the heat shock promoter: parking, stalling, and trucking along. Cell , 74 , 1–4. Liu, R.Y., Kim, D., Yang, S.-H. and Li, G.C. (1993). Dual control of heat shock response: Involvement of a constitutive heat shock element-binding factor. Proc. Natl. Acad. Sci. USA , 90 , 3078–3082. Liu, R.Y., Li, X., Li, L. and Li, G.C. (1992). Expression of human Hsp70 in rat fibroblasts enhances cell survival and facilitates recovery from translational and transcriptional inhibition following heat shock. Cancer Res. , 52 , 3667–3673. Lowe, S.W., Ruley, H.E., Jacks, T. and Housman, D.E. (1993). p53-dependent apoptosis
Thermotolerance and stress response
111
modulates the cytotoxicity of anticancer agents. Cell , 74 , 957–967. Lowe, S.W., Schmitt, E.M., Smith, S.W., Osborne, B.A. and Jacks, T. (1993). p53 is required for radiation-induced apoptosis in mouse thymocytes. Nature (London) , 362 , 847–849. Lunec, J. and Cresswell, S.R. (1983). Heat-induced thermotolerance expressed in the energy metabolism of mammalian cells. Radiat. Res. , 93 , 588–597. Martin, J., Horwich, A.L. and Haiti, F.-U. (1992). Prevention of protein denaturation under heat stress by the chaperonin Hsp60. Science , 258 , 995–998. Mathur, S.K., Sistonen, L., Brown, I.R., Murphy, S.P., Sarge, K.D. and Morimoto, R.I. (1994). Deficient induction of human Hsp70 heat shock gene transcription in Y79 retinoblastoma cells despite activation of heat shock factor 1. Proc. Natl Acad. Sci. USA , 91 , 8695–8699. Maytin, E.V., Wimberly, J.M. and Anderson, R.R. (1990). Thermotolerance and the heat shock response in normal human keratinocytes in culture. Journal of Investigative Dermatology , 95 , 635–642. Michels, A.A., Nguyen, V.T., Konings, A.W., Kampinga, H.H. and Bensaude, O. (1995). Thermostability of a nuclear-targeted luciferase expressed in mammalian cells. Destabilizing influence of the intranuclear microenvironment. Eur. J. Biochem. , 234 , 382–389. Mimori, T., Akizuki, M., Yamagata, H., Inada, S., Yoshida, S. and Homma, M. (1981). Characterization of a high molecular weight acidic nuclear protein recognized by autoantibodies in sera from patients with polymyositis-scleroderma overlap. J. Clin. Invest. , 68 , 611–620. Mimori, T. and Hardin, J.A. (1986). Mechanism of interaction between Ku protein and DNA. J. Biol. Chem. , 261 , 10375–10379. Mizzen, L.A. and Welch, W.J. (1988). Characterization of the thermotolerant cell. I. Effects on protein synthesis activity and the regulation of heat-shock protein 70 expression. J. Cell Biol. , 106 , 1105–1116. Morimoto, R.I. (1993). Cells in stress: Transcriptional activation of heat shock genes. Science , 259 , 1409–1410. Morimoto, R.I., Tissieres, A. and Georgopoulos, C. (1990). Stress Proteins in Biology and Medicine , Cold Spring Harbor Laboratory Press, Plainview. Morimoto, R.I., Tissieres, A. and Georgopoulos, C. (1994). The Biology of Heat Shock Proteins and Molecular Chaperones , Cold Spring Harbor Laboratory Press, Plainview. Mosser, D.D. and Martin, L.H. (1992). Induced thermotolerance to apoptosis in a human T lymphocyte cell line. J. Cell Physiol. , 151 , 561–570. Munro, S. and Pelham, H.R.B. (1984). Use of peptide tagging to detect proteins expressed from cloned genes: deletion mapping functional domains of Drosophila Hsp70. EMBO J. , 3 , 3087–3093. Nguyen, V.T., Morange, M. and Bensaude, O. (1989). Protein denaturation during heat shock and related stress. J. Biol. Chem. , 264 , 10487–10492. Nover, L., Munsche, D., Neumann, D., Ohme, K. and Scharf, K.D. (1986). Control of ribosome biosynthesis in plant cell cultures under heat-shock conditions. Ribosomal RNA. Eur. J. Biochem. , 160 , 297–304. Nussenzweig, A., Chen, C., da Costa Soares, V., Sanchez, M., Sokol, K., Nussenzweig, M.C. and Li, G.C. (1996). Requirement for Ku80 in growth and immunoglobulin V(D) J recombination. Nature (London) , 382 , 551–555. Nussenzweig, A. and Li, G.C. (1993). Alteration of the heat shock response in Hsc70
Molecular chaperones and folding catalysts
112
antisense transfected mammalian cells , 1st Advanced Workshop of the European Science Foundation , Paris, France, 1993, P-5. Palzer, R.J. and Heidelberger, C. (1973). Studies on the quantitative biology of hyperthermic killing of HeLa cells. Cancer Res. , 33 , 415–421. Pinto, M., Morange, M. and Bensaude, O. (1991). Denaturation of proteins during heat shock. J. Biol. Chem. , 266 , 13941–13946. Raaphorst, G.P., Romano, S.L., Mitchell, J.B., Bedford, J.S. and Dewey, W.C. (1979). Intrinsic differences in heat and/or X-ray sensitivity of seven mammalian cell lines cultured and treated under identical conditions. Cancer Res. , 39 , 396–401. Reeves, W.H. (1992). Antibodies to p70/p80 (Ku). antigens in systemic lupus erythematosus. Rheum. Dis. Clin. North Am. , 18 , 391–414. Reeves, W.H. and Sthoeger, Z.M. (1989). Molecular cloning of cDNA encoding the p70 (Ku). lupus autoantigen. J. Biol. Chem. , 264 , 5047–5052. Riabowol, K.T., Mizzen, L.A. and Welch, W.J. (1988). Heat shock is lethal to fibroblasts microinjected with antibodies against Hsp70 . Science , 242 , 433–436. Ritossa, F.M. (1962). A new puffing pattern induced by heat shock and DNP in Drosophila. Experientia , 18 , 571–573. Ritossa, F.M. (1963). New puffs induced by temperature shock, DNP and salicylate in salivary chromosomes of Drosophila melanogaster. Drosophila Inf. Service , 37 , 122– 123. Sapareto, S.A., Hopwood, L.E., Dewey, W.C., Raju, M.R. and Gray, J.W. (1978). Effects of hyperthermia on survival and progression of Chinese hamster ovary cells. Cancer Res. , 38 , 393–400. Sarge, K.D., Murphy, S.P. and Morimoto, R.I. (1993). Activation of heat shock gene transcription by heat shock factor 1 involves oligomerization, acquisition of DNAbinding activity, and nuclear localization and can occur in the absence of stress. Mol. Cell. Biol. , 13 , 1392–1407. Schlissel, M., Constantinescu, A., Morrow, T., Baxter, M. and Peng, A. (1993). Doublestrand signal sequence breaks in V(D)J recombination are blunt, 5'-phosphorylated, RAG-dependent, and cell cycle regulated. Genes Dev. , 7 , 2520–2532. Sciandra, J.J. and Subjeck, J.R. (1984). Heat shock proteins and protection of proliferation and translation in mammalian cells. Cancer Res. , 44 , 5188–5194. Shi, Y., Glynn, M., Guilbert, L.J., Cotter, T.G., Bissonnette, R.P. and Green, D.R. (1992). Role for c-myc activation-induced apoptotic cell death in T cell hybridomas. Science , 257 , 212–214. Skowyra, D., Georgopoulos, C. and Zylicz, M. (1990). The E. coli dnaK gene product, the Hsp70 homolog, can reactivate heat-inactivated RNA polymerase in an ATP hydrolysis-dependent manner. Cell , 62 , 939–944. Smider, V., Rathmell, W.K., Lieber, M.R. and Chu, G. (1994). Restoration of x-ray resistance and V(D)J recombination in mutant cells by Ku cDNA. Science , 266 , 288– 291. Stevenson, M.A., Minton, K.W. and Hahn, G.M. (1981). Survival and concavalin-A induced capping in CHO-fibroblasts after exposure to hyperthermia, ethanol and xirradiation. Radiat. Res. , 86 , 467–478. Strasser, A. and Anderson, R.L. (1995). Bcl-2 and thermotolerance cooperate in cell survival. Cell Gr. Diff. , 6 , 799–805. Subjeck, J.R., Sciandra, J.J. and Johnson, R.J. (1982). Heat shock proteins and thermotolerance: comparison of induction kinetics. Br. J. Radiol. , 55 , 579–584. Taccioli, G.E., Gottlieb, T.M., Blunt, T., Priestly, A., Demengeot, J., Mizuta, R.,
Thermotolerance and stress response
113
Lehmann, A.R., Alt, F.A., Jackson, S.P. and Jeggo, P.A. (1994). Ku80: product of the XRCC5 gene and its role in DNA repair and V(D)J recombination. Science , 265 , 1442–1445. Van Dongen, G., Van de Zande, L., Schamhart, D. and Van Wijk, R. (1984). Comparative studies on the heat-induced thermotolerance of protein synthesis and cell cycle division in synchronized mouse neuroblastoma cells. Int. J. Radiat. Biol. , 46 , 759–769. Vidair, C. and Dewey, W.C. (1988). Two distinct modes of hyperthermic cell death. Radiat. Res. , 116 , 157–171. Vidair, C. and Dewey, W.C. (1991). Division-associated and division-independent hyperthermic cell death: comparison with other cytotoxic agents. Int. J. Hyperthermia , 7 , 51–60. Wallen, C.A. and Landis, M. (1990). Removal of excess nuclear protein from cells heated in different physiological states. Int. J. Hyperthermia , 6 , 87–95. Weirich-Schwaiger, H., Weirich, H.G., Gruber, B., Schweiger, M. and Hirsch-Kauffman, M. (1994). Correlation between senescence and DNA repair in cells from young and old individuals and in premature aging syndromes. Mutation Research , 316 , 37–48. Westra, A. and Dewey, W.C. (1971). Variation in sensitivity to heat shock during the cell-cycle of Chinese hamster cells in vitro. Int. J. Radiat. Biol. , 19 , 467–477. Wiegant, F.A.C., Van Bergen en Henegouwen, P.M.P., Van Dongen, G. and Linnemans, W.A.M. (1987). Stress-induced thermotolerance of the cytoskeleton of mouse neuroblastoma N2A cells and rat Reuber H35 hepatoma cells. Cancer Res. , 47 , 1674– 1680. Yaneva, M. and Busch, H. (1986). A 10S particle released from deoxyribonucleasesensitive regions of HeLa cell nuclei contains the 86-kilodalton-70-kilodalton protein complex. Biochemistry , 25 , 5057–5063. Yang, S.-H., Nussenzweig, A., Li, L., Kim, D., Ouyang, H., Burgman, P. and Li, G.C. (1996). Modulation of thermal induction of Hsp70 expression by Ku autoantigen or its individual subunits. Mol. Cell. Biol. , 16 , 3799–3806. Yih, L.-H., Huang, H., Jan, K.Y. and Lee, T.-C. (1991). Sodium arsenite induces ATP depletion and mitochondrial damage in HeLa cells. Cell Biology International Reports , 46 , 253–264. Yost, H.J. and Lindquist, S. (1986). RNA splicing is interrupted by heat shock and is rescued by heat shock protein synthesis. Cell , 45 , 185–193. Yost, H.J. and Lindquist, S. (1988). Translation of unspliced transcripts after heat shock. Science , 242 , 1544–1548. Yost, H.J. and Lindquist, S. (1991). Heat shock proteins affect RNA processing during the heat shock response of Saccharomyces cerevisiae. Mol. Cell. Biol. , 11 , 1062– 1068.
III. CELLULAR FUNCTIONS
6. GENETIC EVIDENCE FOR THE ROLES OF MOLECULAR CHAPERONES IN ESCHERICHIA COLI METABOLISM WILLIAM F.BURKHOLDER and MAX E.GOTTESMAN* Department of Biochemistry and Molecular Biophysics and Institute of Cancer Research, Columbia University, New York, NY 10032
1. Introduction 2. Genetic Analysis of Chaperone Function: The Case of dnaK 2.1. dnaK, dnaJ, and grpE Mutations Affect Gene Expression 2.2. E. coli Homologs of DnaK and DnaJ 2.3. Dominant dnaK Mutants and the Requirement for grpE 3. Genetic Evidence for the Roles of Chaperones in E. coli Metabolism 3.1. Protein Folding 3.2. Ribosomes and RNA Metabolism 3.3. Protein Export 3.4. Proteolysis 3.5. DNA Replication 3.6. Cell Division 3.7. Acquired Stress Tolerance and Survival during Starvation 4. The Orphaned Chaperones of E. coli: Homologs in Search of Function 4.1. HtpG (Hsp90) 4.2. IbpA and IbpB (Small Hsps) 5. Chaperones and Adaptation: Comparing Genomes 6. Acknowledgments 7. References *Corresponding author
Genetic evidence for the roles of molecular chaperones
117
1. INTRODUCTION As the key regulated modulators of protein conformation in the cell, chaperones are likely to be involved in every aspect of cellular metabolism. A combination of genetic and biochemical analyses of chaperone function in E. coli has suggested possible roles for chaperones in a diverse range of cellular processes. Some of these roles are now fairly well understood on a mechanistic level, while others are still poorly explained. Efforts to unravel the involvement of chaperones in cellular metabolism are broadly motivated by three sets of goals: 1. to identify the cellular processes that chaperones participate in directly, the polypeptide substrates acted on, and the mechanisms of action of the chaperones involved, 2. to determine the ways in which chaperones are specialized, the extent to which their activities overlap, and how chaperones cooperate with each other, and 3. to characterize the regulatory effects of chaperones and any roles they may have in coordinating adaptive responses of the cell to changing environmental conditions. Genetic approaches can be very useful for revealing cellular processes influenced by chaperones and interactions between chaperones and their substrates. However, the road from a mutant phenotype to its underlying mechanism can be long and difficult: 1. A mutant phenotype may result from the disruption of a single cellular process, or from the simultaneous disruption of many processes. If several processes are affected, it may be difficult to identify specific mechanisms underlying the phenotype. 2. A chaperone mutation may affect a cellular process indirectly by altering gene regulation or protein expression. Most notable is the induction of heat shock proteins (hsps), many of which are chaperones or proteases, by mutations in dnaK, dnaJ, or grpE. 3. The role of a chaperone in a particular cellular process may go undetected because of the redundant or overlapping activity of other chaperones. 4. On the other hand, chaperones that appear to have redundant activities may be specialized in vivo due to different patterns of expression, interactions with different cofactors, or distinct substrate specificities. 5. Mutation or loss of one member of a chaperone complex may result in altered interactions of the complex with substrates, not just loss of complex function. Such mutations may therefore have a dominant phenotype. Section 2 of this review focuses on dnaK to illustrate the issues of regulatory effects of chaperone mutations, redundancy, and dominance, since these are particularly well understood in the case of dnaK These issues will be returned to with groEL in section 3.1 as we move on to survey the ways in which chaperones may participate in cellular metabolism. This review only covers the cytoplasmic chaperones of E. coli. Periplasmic chaperones and the specialized chaperones of pilus assembly are reviewed by Missiakis and Raina, and by Hultgren, et al. in this volume. Many chaperones, including DnaK and GroEL, function as dynamic, multimeric complexes with their cochaperones. The biochemistry of these complexes is reviewed in several chapters in this volume, by
Molecular chaperones and folding catalysts
118
Ranson and Clarke, Burston and Saibil, and Buchberger et al. For other recent reviews of the genetics of molecular chaperones in E. coli see Georgopoulos et al. (1994) and Gross (1996). Useful earlier reviews are cited therein.
2. GENETIC ANALYSIS OF CHAPERONE FUNCTION: THE CASE OF DNAK 2.1. dnaK, dnaJ, and grpE Mutations Affect Gene Expression A central role of DnaK and its cochaperones DnaJ and GrpE is to down-regulate the dependent heat shock response, , encoded by the rpoH gene, is one of the two principal a factors regulating heat shock gene expression in E. coli (reviewed by Connolly et al. in this volume; Bukau, 1993; Yura et al., 1996). -dependent gene expression is induced by a variety of stresses, including misfolded cytoplasmic proteins. The second factor, , is induced by extreme heat shock and misfolded proteins in the extracytoplasmic space, and may mediate the heat shock response for that compartment (reviewed by Connolly et al., and Missiakas and Raina, in this volume; Missiakas and Raina, 1997). Cellular levels of are limiting, and increasing the level of up-regulates hsp expression (Grossman et al., 1987; Straus et al., 1987; Tilly et al., 1989). Mutations in dnaK, dnaJ, or grpE elevate basal hsp levels and eliminate the down-regulation of hsp synthesis during recovery from heat shock (Tilly et al., 1983; Straus et al., 1990). Since dnaK, dnaJ and grpE are positively regulated by , they form part of a homeostatic negative-feedback loop that maintains hsp synthesis at levels sufficient to cope with the concentrations of unfolded proteins in the cytoplasm (Craig and Gross, 1991). The failure of dnaK, dnaJ, and grpE mutants to down-regulate hsp expression is harmful to cells. Null mutants of dnaK are inviable at 16° or 40°C and grow slowly and filament at 30°C (Pack and Walker, 1987; Bukau and Walker, 1989a; Bukau and Walker, 1989b; Kang and Craig, 1990; McCarty and Walker, 1994). Mutations in rpoH (sidB alleles) that reduce -dependent gene expression fully suppress the slow growth and filamentation phenotypes of a dnaK52 mutant at 30°C and partially suppress cold sensitivity (Bukau and Walker, 1990). Rapid growth at 16°C and growth at 42°C can be restored by additional uncharacterized suppressor mutations (Bukau and Walker, 1990; Kang and Craig, 1990). Although dispensable for growth, dnaK appears to be active in many cellular processes over a range of growth temperatures, and it is likely that loss of dnaK is compensated for by other chaperones. A redundant chaperone might possibly be induced by the sidB mutations. In addition to inducing hsp expression, mutations in dnaK up- or down-regulate the expression of many other cellular proteins (Spence et al., 1990). Hence, the slow growth of unsuppressed dnaK, dnaJ, or grpE mutants could reflect either the reduced activity of specific proteins required for normal growth or the harmful effects of induced proteins. Some proteins are degraded at faster rates in dnaK mutants. This is due to the induction of proteases regulated by and perhaps also to the loss of dnaK chaperone activity (Gottesman, 1996; see section 3.4).
Genetic evidence for the roles of molecular chaperones
119
Transcriptional regulators other than are regulated by DnaK Consequently, a variety of adaptive cellular responses are impaired in dnaK mutants. DnaK is required for induction of the stationary phase-specific factor, (Muffler et al., 1997; Rockabrand et al., 1998; see section 3.7). DnaK both increases the rate of translation and decreases its rate of degradation. One of the positive regulators of the capsular polysaccharide biosynthetic (cps) genes, RcsA, aggregates in dnaK and dnaJ mutants (Jubete et al. 1996; Jubete, Maurizi, and Gottesman, manuscript in preparation). The tendency of RcsA to aggregate in the absence of DnaK may explain the decrease in cps gene expression that occurs with increasing temperature. Titration of DnaK and DnaJ by competing substrates at higher temperatures may lead to the formation of inactive RcsA aggregates. As with the cps genes, flagellar gene expression is reduced in dnaK, dnaJ, and grpE mutants. The effect is due to reduced expression of the positive regulators (encoded by fliA) and FhlD/C, resulting in a nonmotile phenotype (Shi et al., 1992). Gene expression in dnaK mutants may also be affected by competition between and other subunits for RNA polymerase, although this does not appear to be the case with (Muffler et al., 1997). 2.2. E. coli Homologs of DnaK and DnaJ Though DnaK, DnaJ, and GrpE act together as a complex, the phenotypes of null mutations in their corresponding genes differ in severity. grpE null mutants are inviable at all temperatures in an otherwise wild-type strain background (Ang and Georgopoulos, 1989). dnaK null mutants are inviable below 16°C and above 40°C (Bukau and Walker, 1989a; Kang and Craig, 1990; McCarty and Walker, 1994). At 30°C, dnaK mutants grow slowly, have poor viability, frequently form long filaments, and are defective in chromosome segregation (Paek and Walker, 1987; Bukau and Walker, 1989a; Bukau and Walker, 1989b; Kang and Craig, 1990). dnaJ null mutants have the mildest phenotypes of the three. They are not cold sensitive and are inviable only at 43°C and above (Sell et al., 1990). Like dnaK mutants, dnaJ mutants grow slowly and filament at 30°C, though less markedly than dnaK mutants. In the absence of GrpE, DnaK and DnaJ may be toxic to cells (see section 2.3), and this may explain why grpE is an essential gene. On the other hand, dnaJ mutants may have less severe phenotypes than dnaK mutants because of the activity of a redundant dnaJ/homolog, cbpA. The phenotypes of cbpA dnaJ double mutants are similar to those of dnaK mutants (Ueguchi et al., 1994; Ueguchi et al., 1995; Wegrzyn et al., 1996). A cbpA null mutation confers no apparent phenotype in a dnaJ + strain. Overexpression of CbpA suppresses the phenotypes of dnaJ null and missense mutants. Two lines of genetic evidence suggest that CbpA replaces DnaJ in the complex with DnaK rather than acting in a parallel bypass pathway: (1) overexpression of CbpA does not rescue the growth defects of dnaK, grpE, groEL, groES, or rpoH missense mutants, and, (2) the phenotype of a dnaK mutant is not exacerbated by a second mutation in cbpA (Ueguchi et al., 1994). CbpA and DnaK have not yet been shown to interact functionally in vitro. Expression of cbpA is regulated by the stationary phase-specific sigma factor and is induced when cells enter stationary phase or during phosphate limitation (Yamashino et al., 1994; see section 3.7). Unlike dnaJ, cbpA is not induced by heat shock. CbpA is 39%
Molecular chaperones and folding catalysts
120
identical to DnaJ but lacks the DnaJ cysteine rich region (Ueguchi et al., 1994). This region appears to be required for interactions of DnaJ with some, but not all, polypeptide substrates, and for promoting the formation of stable DnaK-substrate complexes (Szabo et al., 1996; Banecki et al., 1996). In addition to cbpA, E. coli has two other dnaJ homologs, djlA and hscB (yfhE) (Clarke et al., 1996; Seaton and Vickery, 1994). DjlA and the protein encoded by hscB, Hsc20, are similar to DnaJ only in the conserved J-domain, which is required for the binding of DnaJ homologs to the specific hsp70s with which they function (Wall et al., 1994; Schlenstedt et al., 1995). DjlA is an integral membrane protein, with the J-domain at its cytoplasmic C-terminus (Clarke et al., 1996). DjlA thus belongs to a growing family of Jdomain-containing integral and membrane-associated proteins that are known or presumed to target hsp70s for membrane localized functions (Cyr et al., 1994). Overexpression of DjlA induces expression of the capsular polysaccharide biosynthetic (cps) genes (Kelley and Georgopoulos, 1997; Clarke et al., 1997). This phenotype requires specific sequences in the transmembrane domain of DjlA, a functional ‘Jdomain’, DnaK, and GrpE. The phenotype may thus result from DjlA acting as a cochaperone with DnaK and GrpE on a component of the signaling pathway regulating cps expression, such as the integral membrane sensor kinase RcsC. In this model, the transmembrane domain of DjlA may participate in substrate recognition (Clarke et al., 1997). Deletion of djlA has a modest effect on the timing of cps gene expression (Kelley and Georgopoulos, 1997). E. coli also has a second dnaK homolog, Hsc66, that is 42% identical to DnaK (Seaton and Vickery, 1994). The gene encoding Hsc66, hscA, is in the same operon as the dnaJ homolog hscB, encoding Hsc20. Hsc66 and Hsc20 appear to functionally interact in a manner analogous to DnaK and DnaJ (Vickery et al., 1997). A possible target of Hsc66 action is the E. coli nucleoid-binding protein H-NS (Kawula and Lelivelt, 1994), which modulates the expression of other environmentally regulated genes, including the stationary sigma factor, (Barth et al., 1995). Hsc66 may antagonize H-NS activity, since inactivation of hscA suppresses a putative hns promoter-down mutation, but has no effect on a hns mutant (Kawula and Lelivelt, 1994). The pattern of protein expression following cold shock is altered in hscA null mutants (Lelivelt and Kawula, 1995). This altered expression pattern maybe related to H-NS function, since H-NS is induced by cold shock. 2.3. Dominant dnaK Mutants and the Requirement for grpE The analysis of dominant negative dnaK mutants has helped clarify the basis of dnaK mutant phenotypes and provided insights into the mechanism of dnaK action. Of the dnaK mutations selected as failing to down-regulate heat shock gene expression, roughly 20% are dominant (Wild et al., 1992b). The two dnaK mutants exhibiting the strongest dominant phenotypes, GD229 and GD341, encode proteins defective in ATP-binding (Wild et al., 1992b; mutations are designated by single letter amino acid code and residue number). ATP binding is required for release of both bound substrate and GrpE from DnaK (Palleros et al., 1993; Zylicz et al., 1987) and is stimulated by GrpE-catalyzed nucleotide exchange (Liberek et al., 1991). This suggests two complementary
Genetic evidence for the roles of molecular chaperones
121
mechanisms of dominance: 1. The DnaK mutants bind polypeptide substrates but are slow to dissociate, thus blocking the spontaneous or chaperone-assisted folding or function of these substrates. Dominant DnaK mutations may therefore interfere with cellular processes unaffected by DnaK loss-of-function mutations (see section 3.3). 2. The mutant proteins are defective in releasing bound GrpE, and thus titrate GrpE away from wild-type DnaK. Without GrpE-catalyzed nucleotide exchange, wild-type DnaK would also be slow to dissociate from substrates. Consistent with the suggestion that defects in substrate release can result in a dominant negative phenotype, the dominant EK171 mutant is defective in the ATP-dependent release of peptide substrates (Buchberger et al., 1994). Similarly, a C-terminal fragment of DnaK (CTF) containing only the substrate binding domain is toxic when overexpressed in wild-type cells (Burkholder et al., 1996; see Buchberger et al., in this volume, for a description of the domain structure of DnaK). CTF binds and releases substrates in a manner similar to ADP-bound full-length DnaK (Gragerov et al., 1994). Mutations in the CTF that reduce substrate binding relieve toxicity (Burkholder et al., 1996). DnaK mutants may also inhibit cell growth by titrating GrpE, as suggested by mutants of an N-terminal fragment (NTF) of DnaK comprising only the ATP binding domain (Burkholder and Gottesman, unpublished). Three NTF mutants were isolated that were defective in the ATP-dependent release of GrpE, GE228, GS229 and GD341. The mutants are toxic to cells when overexpressed, but coexpression of GrpE suppressed the lethality of GE228. This suggests that the mutants inhibit E. coli growth by titrating GrpE. The mutations carried by these NTFs are similar or identical to previously identified dominant mutants of full-length DnaK (Wild et al., 1992b). The full-length DnaK mutants EK171 and DN201 are also defective in ATP-dependent release of GrpE (Kamath-Loeb et al., 1995). The GD229 and EK171 dominant negative mutants of DnaK impair cellular growth at 30°C much more severely when coexpressed with wild-type DnaK than when expressed by themselves. GD229 has the stronger synthetic phenotype. Expression of GD229 from a single chromosomal copy is lethal to duaK + merodiploids (Wild et al., 1992b), whereas EK171 only retards growth when overexpressed from a multicopy plasmid (Buchberger et al., 1994). Neither of the two models of dominance outlined above explain why the toxicity of GD229 and EK171 requires wild-type dnaK. It is possible that the mutant proteins titrate GrpE but are not lethal because they fail to form stable complexes with polypeptide substrates. Complex formation between DnaK and polypeptide substrates is stimulated by ATP in the presence of DnaJ (cf. Wawrzynów et al., 1995a). GD229 and EK171 have been shown to be defective in either ATP-binding or ATP hydrolysis stimulated by substrates and DnaJ (Wild et al., 1992b; Buchberger et al., 1994; KamathLoeb et al., 1995). These mutants may, therefore, be unable to form stable ATPdependent complexes. EK171 and DN201 have wild-type affinities for peptide substrates in the absence of ATP, but these interactions may be qualitatively different than forming stable ATP- and DnaJ-dependent complexes. Extragenic suppressor mutations that permit the deletion of grpE in a dnaK + strain
Molecular chaperones and folding catalysts
122
have been reported (Ang and Georgopoulos, 1989). grpE can also be deleted in cells expressing DnaK mutants that appear to bind substrates less stably. One such mutant, DnaK332, has an increased basal rate of nucleotide exchange. A carboxyterminal truncation of DnaK also permits deletion of grpE (described in Georgopoulos et al., 1994). The effect of the truncation may be to remove the lid that locks bound substrate into ADP-bound DnaK while preserving the substrate binding cleft itself (see Zhu et al., 1996). These suppressor mutations of dnaK lend further support to the model that GrpE is essential to dissociate stable complexes formed between DnaK and substrate polypeptides.
3. GENETIC EVIDENCE FOR THE ROLES OF CHAPERONES IN E. COLI METABOLISM 3.1. Protein Folding 3.1.1. Cooperative Action of the DnaK and GroEL Complexes In this section, we focus on genetic evidence for the involvement of chaperones in protein folding in E. coli. Other evidence for the role of chaperones in protein folding is reviewed extensively in other chapters of this volume. The cellular requirement for chaperone function increases with increasing temperature. This is due in part to the increased rates of growth and protein synthesis that occur at higher temperatures and also to the temperature sensitivity of the folding intermediates of many proteins (for a recent review, see King et al., 1996). Cells lacking rpoH are unable to induce hsp expression and cannot grow above 20°C (Zhou et al., 1988). Shifting to nonpermissive temperature does not immediately halt protein synthesis in rpoH mutants, but most newly synthesized polypeptides aggregate and form cytoplasmic inclusion bodies (Gragerov et al., 1991; Kucharczyk et al., 1991; Gragerov et al., 1992). Overexpression of either GroEL/ES or DnaK/ J raises the permissive temperature of rpoH null mutants and suppresses the aggregation of newly synthesized polypeptides (Kusukawa and Yura, 1988; Gragerov et al., 1992). However, the defects in growth and protein aggregation can be suppressed with much lower levels of these chaperones if both pairs are coexpressed. Similarly, dnaK756 groEL100 or dnaK756 groES30 double mutants display extensive aggregation at 42°C whereas single mutants do not (Gragerov et al., 1992). Thus, aggregation is more efficiently suppressed by the cooperative action of GroEL/ES and DnaK/J. How do GroEL/ES and DnaK/J function cooperatively? NMR studies indicate that DnaK binds unfolded peptide substrates, whereas GroEL binds substrates that have some secondary structure (Landry et al., 1992). These results are supported by two lines of evidence. First, the crystal structure of the substrate binding domain of DnaK shows that substrate polypeptides must contain fully unfolded regions to be encapsidated in the DnaK binding cleft (Zhu et al., 1996; see Ha et al., and Buchberger et al., this volume). Second, biochemical analyses reveal interactions of GroEL with partially folded and even fully folded substrate proteins (see Burston and Saibil, this volume). These data, together
Genetic evidence for the roles of molecular chaperones
123
with studies of protein folding in mitochondria and in reconstituted in vitro reactions, suggest that DnaK/J and GroEL/ES act sequentially in the folding pathway. In this model, DnaK and DnaJ bind nascent unfolded polypeptides emerging from the ribosome to block premature folding and aggregation, and perhaps facilitate polypeptide translocation (see Nelson et al., 1992; Welch et al., this volume). Then full-length unfolded or partially folded polypeptides are transferred to GroEL/ES where they fold to the native state in one or more cycles of binding and release. Consistent with a sequential folding pathway, DnaK and DnaJ, but not GroEL, are found to bind ribosome-associated nascent polypeptides (Gaitanaris et al., 1994). Further, pulse-chase experiments indicate that newly synthesized full-length CI associates transiently with GroEL/ES before entering the soluble fraction (Gaitanaris et al., 1994). The vectorial model has its limitations, however. If DnaK and GroEL work at separate stages of folding, how is it that each can substitute for the other to suppress aggregation when overexpressed? Along the same lines, why is a dnaK strain viable at 30°C? It is possible that other chaperones may substitute for DnaK or GroEL at either early or late stages of protein folding. It is also possible, however, that the pathway of chaperoneassisted folding is more dynamic and flexible, with substrates passing back and forth between different chaperones until properly folded or degraded (see, for example Buchberger et al., 1996; Burston et al., 1996; reviewed in Bukau et al., 1996; Gragerov, 1997). Although no global aggregation of mature proteins is seen in rpoH mutants at nonpermissive temperature (Gragerov et al., 1991), temperature sensitive proteins may still require chaperones to prevent aggregation at high temperature and to refold correctly when returned to permissive temperature. Thus, firefly luciferase heterologously expressed in E. coli is inactivated when cells are shifted to high temperature. Luciferase activity is restored in the absence of new protein synthesis upon returning cells to permissive temperature (Schröder et al., 1993). Activity is not recovered in dnaK, dnaJ, or grpE mutants, although wild-type levels of the luciferase protein are still present. The chaperone mutants are, therefore, defective in protecting luciferase from irreversible heat inactivation. In vitro, DnaK, DnaJ and GrpE are required to refold denatured luciferase at 30°C (Schröder et al., 1993; Szabo et al., 1994). Whereas denatured full-length luciferase requires DnaK for refolding in vitro and in vivo, newly synthesized luciferase folds readily into active form at 30°C in a dnaK52 strain (Hesterkamp and Bukau, personal communication). The nascent polypeptide folding pathway for luciferase thus appears to be distinct from the pathway stabilizing and refolding the heat-inactivated protein. 3.1.2. A Central Role for GroEL/ES groEL and groES are essential for growth at all temperatures (Fayet et al., 1989). Consistent with the homeostatic model of hsp regulation (Craig and Gross, 1991; Bukau, 1993), GroEL/ES is expressed at levels just sufficient to meet the needs of the cell. Reducing GroEL/ES expression to 25% of normal levels slows cell growth and induces -dependent hsps at 37°C (Kanemori et al., 1994). Without sufficient GroEL/ES hsp expression is presumably stimulated by accumulated unfolded polypeptides. When strains carrying the groEL201 (EK461) mutation are shifted to 37°C, the concentrations of
Molecular chaperones and folding catalysts
124
roughly 50% of the newly synthesized soluble proteins are reduced (Horwich et al., 1993). Many of these proteins aggregate, and the rest may become susceptible to proteolysis. Other proteins appear to be inactivated while remaining stable and soluble. Unlike rpoH mutants, massive aggregation of nascent polypeptides and formation of inclusion bodies is not seen in the groEL201 mutant. The protein species affected by groEL201 may indicate which proteins interact with GroEL to fold at 37°C. Roughly 50% of E. coli cytoplasmic proteins, when diluted from denaturant, bind GroEL (Viitanen et al., 1992; Horwich et al., 1993). The ability of several proteins to interact with GroEL in vitro correlates with their defective folding in groEL201 in vivo (Horwich et al., 1993). However, estimates of the proportion of nascent polypeptides that GroEL could interact with during growth at 37°C, given the abundance of GroEL, the rate of protein synthesis, and the rates of substrate binding and release by GroEL in vitro, suggest that GroEL may only be involved in the folding of 5–30% of cellular polypeptides (Lorimer, 1996). Efforts to identify particular in vivo substrates of GroEL/ES are discussed in the following sections. The first identified substrates of GroEL/ES are the capsid head proteins of bacteriophages and T4 and the tail of bacteriophage T5 (for review, see Georgopoulos et al., 1990; Zeilstra-Ryalls et al., 1991; Georgopoulos et al., 1994). groEL and groES mutants are resistant to infection, growth in the mutants is blocked at an early step of phage head morphogenesis, at the point waen the B protein assembles into the homomeric preconnector ring. Assembly of the bacteriophage T4 head and bacteriophage T5 tail are also blocked at early stages. GroEL appears to play a similar role in the assembly of certain endogenous protein complexes. Mutations in groEL and groES block the anaerobic processing and assembly of one of the three NiFe hydrogenases in E. coli, HYD3, at a step sensitive to the availability of Ni2+ (Rodrigue et al., 1996). The mutations also interfere with the activity of a second NiFe hydrogenease. Ni2+ appears to be incorporated into HYD3 during folding, since it is buried in the native protein structure of a related hydrogenase. GroEL may stabilize HYD3 until Ni2+ is bound and folding can proceed. GroEL binds the HYD3 precursor in vitro, consistent with a direct role in HYD3 processing and assembly. In what is probably a related reaction, precursors of the MoFe nitrogenase complex of Klebsiella pneumoniae transiently associate with GroEL during assembly in vivo (Govezensky et al., 1991). E. coli GroEL facilitates assembly of the complex when it is expressed in E. coli or when E. coli GroEL is overexpressed in Klebsiella pneumoniae. The requirement of chaperones for the assembly of cofactor-dependent complexes may turn out to be a general rule. Interestingly, the assembly of a molybdenum-dependent nitrate reductase complex in E. coli appears to depend in part on a dedicated chaperone, NarJ (Liu and DeMoss, 1997; Blasco et al., 1998). GroEL may play a direct role in gene regulation by promoting the folding of transcriptional activators. GroEL is required in vivo for activation of both the NodD and the NifA transcriptional activators from Rhizobium meliloti and Klebsiella pneumoniae, respectively. It physically interacts with both in vitro, supporting a direct role for the chaperone in their activation (Ogawa and Long, 1995; Govezensky et al., 1991; Govezensky et al., 1994). Since groEL and groES mutations are pleiotropic, it is difficult to identify specific
Genetic evidence for the roles of molecular chaperones
125
substrates of GroEL/ES that might be affected by the mutations. The problem has been approached genetically by identifying allele-specific interactions between mutations in groEL or groES and the gene encoding a putative substrate. For instance, single-stranded DNA binding protein (SSB) was identified as a possible target for GroEL when groEL411 was isolated as an allele-specific, dominant suppressor of the ssb-1 mutant allele (Ruben et al., 1988). Another ssb mutation, ssb-13, is specifically suppressed by the groEL46 mutation (Laine and Meyer, 1992). Importantly, neither ssb mutation is suppressed by overexpressing wild-type GroEL. Allele-specific interactions and suppressor analysis have also been used to define sites of interaction between GroEL, GroES, and substrates such as DnaA46 (Zeilstra-Ryalls et al., 1993; Zeilstra-Ryalls et al., 1994; Zeilstra-Ryalls et al., 1996). The systematic structure/function analyses of GroEL/ES in vitro have made rapid progress on the same questions (Fenton et al., 1994; see Burston and Saibil, this volume). A change in GroEL substrate specificity occurs following infection with bacteriophage T4. Phage T4 encodes a functional homolog of GroES, gp31, which interacts with GroEL and can replace GroES for many functions in E. coli (van der Vies et al., 1994; for a review of early work, see Zeilstra-Ryalls et al., 1991). gp31 and GroEL are essential for T4 growth, whereas GroES is dispensable. This suggests that the GroEL-gp31 complex uniquely supports the folding of the T4 substrate polypeptide, the capsid head protein gp23, while still retaining activity towards endogenous bacterial substrates. The threedimensional structure of gp31 is similar to that of GroES, but gp31 complexed to GroEL forms a larger and more hydrophilic cavity encapsulating protein folding intermediates than is formed by the GroEL-GroES complex (Hunt et al., 1997). 3.2. Ribosomes and RNA Metabolism A major E. coli protein associated with nascent polypeptide chains is trigger factor. Trigger factor is a ribosome-associated peptidyl prolyl cis-trans isomerase that can be cross-linked to nascent chains of both cytosolic and secreted polypeptides (Lill et al., 1988; Stoller et al., 1995; Valent et al., 1995; Hesterkamp et al., 1996; reviewed in Hesterkamp and Bukau, 1996; see Fischer and Schmid, this volume). Trigger factor stabilizes the outer membrane precursor protein, proOmpA, in a translocation competent state in vitro, suggesting that it can act as a chaperone (Crooke and Wickner, 1987; Crooke et al., 1988). However, depletion of trigger factor in vivo has no effect on proOmpA export, even in a secB null mutant (section 3.3). Trigger factor also associates with GroEL and stimulates its binding to denatured proteins, the unstable fusion protein CRAG, and other model polypeptide substrates (Kandror et al., 1995; Kandror et al., 1997; see section 3.4). Trigger factor is induced by cold shock and appears to increase cell viability at low temperatures (Kandror and Goldberg, 1997). Underexpression or overexpression of trigger factor induces cell filamentation (Guthrie and Wickner, 1990). It has not been demonstrated that the peptidyl prolyl isomerase and chaperone-like activities of trigger factor are related. However, isomerization of prolyl peptide bonds is a significant rate-limiting step for the folding of some proteins (reviewed in Galat and Metcalfe, 1995; Schmid, 1995). In addition to trigger factor, a growing list of cytosolic and periplasmic peptidyl prolyl isomerases has been identified in E. coli. These include a
Molecular chaperones and folding catalysts
126
cytosolic FK506-binding protein homolog, SlyD, mutants of which affect the growth of E. coli and the function of a phage encoded lysis protein (Roof et al., 1996; see Missiakas and Raina, this volume, for a review of the role of periplasmic peptidyl prolyl isomerases). A role for hsps in ribosome function is suggested by the activation of the ribosomeassociated enzyme RelA by dnaK and dna/mutations at nonpermissive temperature. RelA is a guanosine tetraphosphate (ppGpp) synthetase that is activated when the proportion of charged tRNAs is growth-limiting. High levels of ppGpp block stable RNA synthesis via the stringent response pathway (for a recent review, see Cashel et al., 1996). Heat shock raises ppGpp concentrations in dnaK and dnaJ mutants (Tanaka et al., 1985; Itikawa et al., 1986). Correspondingly, when dnaK, dnaJ, or grpE mutants are shifted to nonpermissive temperature, RNA synthesis rapidly shuts off while DNA and protein synthesis continue for some time (Itikawa and Ryu, 1979; Wada et al., 1982; Ang et al., 1986). The accumulation of ppGpp and the rapid inhibition of RNA synthesis are relAdependent (Tanaka et al., 1985; Itikawa et al., 1986). A transient burst of ppGpp synthesis also occurs in some, but not all, wild-type control strains following heat shock. In these cells, however, ppGpp concentrations decline to new basal levels with kinetics similar to the shutoff of hsp induction. It is unclear why ppGpp concentrations increase during heat shock and why dnaK and dnaJ mutants are defective in down-regulating ppGpp synthesis following heat shock. Charged tRNA levels may become limiting after temperature upshift as a result of an increased rate of polypeptide synthesis. Alternatively, an aminoacyl-tRNA synthetic pathway could be inhibited by heat shock. For instance, the metA gene product, homoserine transsuccinylase, an essential enzyme in the methionine biosynthesis pathway, is unstable at 42°C (reviewed in Ron et al., 1990). Phosphorylation of the glutaminyl- and threonyl-tRNA synthetases is reduced in dnaK and dnaJ mutants (Wada et al., 1986), although it has not been proven that the activity of the enzymes is affected by phosphorylation. Ribosome function might also be directly affected following heat shock. A role for hsp70s in ribosomal function in S. cerevisiae has been proposed (Nelson et al., 1992), and recent in vitro evidence suggests that chaperone-mediated folding of nascent polypeptides may be coupled to translation termination and polypeptide release in E. coli (Kudlicki et al., 1994). DnaK may participate in the assembly of 30S and 50S ribosomal subunits. Ribosomal precursors accumulate in dnaK missense mutants upon shift to nonpermissive temperature (Alix and Guérin, 1993). The accumulated precursors remain competent for assembly, and chase into mature ribosomes when the cells are returned to permissive temperature. The assembly defects of dnaK mutants are not suppressed by introducing rpoH or relA mutations, indicating that neither hsp overexpression nor induction of the stringent response is responsible for the defects. The pattern of precursor accumulation varies among different mutant alleles of dnaK, presumably reflecting differences in behavior of the mutant DnaK proteins. GroEL has been suggested to regulate ribosomal RNA processing and mRNA turnover. When a groEL mutant is shifted to nonpermissive temperature, the 9S precursor to ribosomal p5S RNA accumulates, consistent with reduced RNase E processing (Sohlberg et al., 1993). Conversely, overexpression of a fragment of GroEL suppresses a
Genetic evidence for the roles of molecular chaperones
127
temperature sensitive ams (RNase E) mutation (Chanda et al., 1985). GroEL copurifies with RNase E and binds a mutant RNase E protein, but direct binding to wild-type RNase E has not been observed (Sohlberg et al., 1993; Miczak et al., 1996). GroEL also forms stable complexes with polynucleotide phosphorylase, which itself associates with RNase E, suggesting that GroEL may be a component of protein complexes that process and degrade RNA (Ybarra and Horowitz, 1996). Recent work implies that a form of GroEL may possess an opposing activity, binding RNA and protecting it from degradation (Georgellis et al., 1995; Georgellis, Vytvytska, Täuber, Rezazadek, Brandsch, Hartl, and von Gabain, manuscript in preparation). A fraction of GroEL copurifies with an RNA binding activity that inhibits RNase digestion of 9S RNA in vitro. The binding activity is induced by anaerobic growth or phosphate limitation, conditions under which OmpA mRNA is stabilized in vivo. An unphosphorylated monomeric form of GroEL also predominates under these conditions, and may correspond to the RNA binding activity. These data suggest that GroEL may not only play a role in protein folding, but also in mRNA stability and in the regulation of translation. 3.3. Protein Export Most proteins destined for the periplasmic space or the outer membrane are translocated across the inner membrane by a machinery encoded by the sec genes (for reviews, see Pugsley, 1993; den Blaauwen and Driessen, 1996; Murphy and Beckwith, 1996). Once in the periplasmic space, other general or specialized chaperone activities are required for protein folding and transport (see Missiakis and Raina, this volume). The insertion of proteins into the inner membrane is less well understood, and the role of the sec-encoded export machinery in this process is not defined. Though some cotranslational protein export occurs in E. coli, most proteins are exported posttranslationally. These proteins are presumably maintained in a transport-competent unfolded or loosely folded conformation. SecB is a cytosolic chaperone that binds proteins destined for export and prevents their folding, aggregation, or degradation (for recent reviews, see Collier, 1993; Randall and Hardy, 1995; Welch et al., this volume). In addition, SecB may facilitate the targeting of precursor proteins to the translocation apparatus by binding to SecA, a cytosolic factor that cooperates directly with the SecY inner membrane translocase during export. Null secB mutants are defective in exporting some, but not all, proteins. Thus, efficient targeting of SecB-independent proteins to the secretory apparatus either does not require a chaperone or utilizes a chaperone other than SecB. The chaperone function of SecB appears to be partially redundant with DnaK. secB null mutants fail to grow in rich medium below 42°C and are defective for protein export in minimal medium (Kumamoto and Beckwith, 1985). These phenotypes are suppressed by overexpression of (Altman et al., 1991). Overexpression of DnaK and DnaJ is sufficient to suppress the growth and export defects of a secB null mutant, whereas overexpression of DnaK or GroEL/ES alone does not (Altman et al., 1991; Wild et al., 1992a). Conversely, reducing hsp expression prevents the growth of a secB null mutant, suggesting that -dependent genes compensate for the absence of secB (Altman et al., 1991). DnaK is necessary for this compensation, since the growth of a secB null mutant is
Molecular chaperones and folding catalysts
128
blocked by strongly defective dnaJ or dnaK mutations or by depletion of DnaK, DnaJ, or GrpE (Wild et al., 1992a; Wild et al., 1996). During depletion, the export of SecBindependent proteins is impaired and the export of SecB-dependent proteins is reduced below the levels seen in a secB single mutant. Since the export of SecB-independent proteins is only affected in the absence of both SecB and DnaK, the two chaperones appear to be fully redundant for exporting this class of proteins. However, one of the SecB-independent proteins affected by DnaK depletion, ribose binding protein (RBP), is not a substrate for SecB (reviewed in Collier, 1993). SecB is thought to recognize its substrates in part through kinetic partitioning, binding to slow folding substrates in preference to substrates that fold quickly (reviewed in Randall and Hardy, 1995). RBP folds rapidly in vitro, consistent with its not being a substrate of SecB. How, then, does RBP export become dependent on SecB in the absence of DnaK? If the folding of RBP is slower in vivo in the absence of DnaK, then depletion of DnaK may make RBP a suitable substrate for SecB. The export of slow-folding mutants of RBP is, in fact, SecB dependent (Teschke et al., 1991; Kim et al., 1992). Interactions between DnaK and specific exported proteins are suggested by the allelespecific effects of dnaK and dnaJ mutations on PhoA translocation (Wild et al., 1992a). Translocation of PhoA to the periplasmic space is SecB-dependent at 30°C but SecBindependent at higher temperatures. Two dominant mutations of dnaK, EK171 and GD229, and a dnaJ mutation, HQ33, inhibit PhoA export at 30°C without affecting export of several other SecB-dependent and SecB-independent proteins. The defect is not due to loss of functional DnaK, since depletion of DnaK, DnaJ, or GrpE does not inhibit PhoA export in secB + cells (Wild et al., 1996). The mutant proteins may instead interfere with export by binding irreversibly to PhoA precursors or some other factor specifically required for PhoA export (see section 2.3). Doubt is cast on this interpretation, however, by the finding that GrpE depletion, which globally stabilizes DnaK-substrate complexes, does not inhibit PhoA export. GroEL/ES may also participate in the export of particular proteins. Pre- -lactamase export is SecB-independent and is impaired in groEL or groES mutants (Kusukawa et al., 1989; Laminet et al., 1991). Consistent with a direct role for GroEL/ES in this reaction, GroEL interacts with pre- -lactamase in vitro to suppress aggregation, maintain translocation competence, or promote either folding or unfolding (Bochkareva et al., 1988; Laminet et al., 1990; Zahn and Plückthun, 1992). Pre- -lactamase export is also reduced by depleting DnaK, DnaJ, or GrpE in a secB null mutant, although not as dramatically as by groEL or groES mutations (Wild et al., 1992a; Wild et al., 1996). This suggests that GroEL/ES plays a primary role in pre- -lactamase export whereas DnaK and SecB play secondary roles. GroEL/ES also maintains newly translated lactose permease, an integral inner membrane protein, in a conformation competent for insertion into the cytosolic face of membrane vesicles in vitro (Bochkareva et al., 1996). Overexpression of GroEL/ES or DnaK/J facilitates the export of certain heterologously expressed proteins (see, for example, Phillips and Silhavy, 1990; Bergès et al., 1996). In general, however, chaperone overexpression has had only mixed success in promoting protein folding and secretion (for recent reviews, see Makrides, 1996; Thomas et al., 1997). Furthermore, suppression of protein export defects by increasing chaperone levels may be unrelated to chaperone function. Thus, the growth and export defects of several
Genetic evidence for the roles of molecular chaperones
129
cold-sensitive secretory pathway mutants are suppressed by overexpressing a variety of proteins, including GroEL/ ES or even inactive truncation mutants of GroEL (Danese et al., 1995). Similarly, suppression of a secY24 mutant by overexpressing HtpG may be a nonspecific effect of reduced growth rate (Shirai et al., 1996). Slow growth and lower rates of protein synthesis, conferred by mutations or by protein synthesis inhibitors, also suppress sec mutants (Lee and Beckwith, 1986; Schatz and Beckwith, 1990). 3.4. Proteolysis Proteases, like chaperones, adapt cells to changing environmental conditions and protect them from stress. Proteases down-regulate unneeded or inappropriate cell functions during adaptation. Proteins induced by carbon starvation, for example, are degraded after starving cells are resupplemented with glucose (Damerau and St. John, 1993; for reviews, see Gottesman, 1996; Sherman and Goldberg, 1996). Several E. coli ATP-dependent proteases are induced by heat shock. They presumably protect the cell by degrading misfolded or damaged proteins. To be accessible to proteases, substrates must be soluble. By blocking aggregation and perhaps by promoting unfolding, chaperones aid in the presentation of substrates to proteases (reviewed by Maurizi et al., this volume). Chaperone-assisted proteolysis of a substrate may compete with aggregation and chaperone-mediated protein folding. This may account for the varied effects of chaperone mutations on the degradation of specific proteins in vivo. Mutations in dnaK, dnaJ, grpE, or groEL stabilize some rapidly degraded proteins, such as unfolded or misfolded polypeptides induced by puromycin or canavinine, the unstable regulatory factor RcsA, and the mutant protein PhoA61 (Keller and Simon, 1988; Straus et al., 1988; Sherman and Goldberg, 1992; Jubete et al. 1996). Mutations in dnaK, dnaJ, grpE, or groEL may increase the rate of proteolysis of other proteins, however, including a temperaturesensitive mutant of the CI represser, the C protein of bacteriophage Mu, and the E. coli UmuC protein (Keller and Simon, 1988; Sand et al., 1995; Donnelly and Walker, 1989). UmuC, which forms a heterodimer with UmuD’, may be stabilized by GroEL/ES until the heterodimer is formed (Donnelly and Walker, 1992). The UmuC-UmuD’ complex is required for the error-prone repair of DNA damage. groEL and groES mutants are defective in UV-induced mutagenesis and the reactivation of UV-damaged bacteriophage, perhaps as a result of lowered UmuC levels (Donnelly and Walker, 1989; Liu and Tessman, 1990). The association of unstable proteins with chaperones may increase their rates of degradation. Missense mutations of dnaK or grpE that stimulate the binding of PhoA61 to DnaK enhance PhoA61 degradation. Conversely, a loss-of-function mutation in dnaJ that decreases the amount of PhoA61 associated with DnaK reduces PhoA61 degradation (Sherman and Goldberg, 1992). Similarly, conditions that enhance the association of the unstable fusion protein CRAG with GroEL increase its rate of degradation by the protease ClpP (Kandror et al., 1994). The chaperone trigger factor facilitates the binding of CRAG to GroEL and stimulates CRAG degradation (Kandror et al., 1995). The importance of chaperone activities for proteolysis is underscored by the recent finding that the ATPase subunits of the ClpP protease, ClpA and ClpX, are chaperones in their own right (Wickner et al., 1994; Levchenko et al., 1995; Wawrzynów et al., 1995b;
Molecular chaperones and folding catalysts
130
Kruklitis et al., 1996; reviewed in Schirmer et al., 1996; Wawrzynów et al., 1996; Gottesman et al., 1997; Maurizi et al., this volume). In vivo evidence for the ClpPindependent chaperone activity of ClpX is discussed in section 3.5. ClpA and ClpX, which are members of the hsp100 family of chaperones, were first identified as factors that target specific polypeptide substrates to the ClpP protease for degradation. The ATPase subunits are now thought to not only recognize and target substrates to the ClpP protease but also to actively unfold them for presen-tation to ClpP. Recognition of substrates by ClpA and ClpX is not sufficient for proteolysis, however, since a stable protein tagged with a sequence recognized and bound by ClpX is not degraded by ClpP (Laachouch et al., 1996). Another member of the E. coli hsp100 family, ClpB, solubilizes protein aggregates, which are then rapidly degraded (Inoue and Rechsteiner, 1994; Laskowska et al., 1996a). ClpB is the nearest E. coli homolog to the S. cerevisiae, chaperone Hsp104, which also solubilizes protein aggregates in vivo (Parsell et al., 1994). ClpB is not required to target solubilized proteins for degradation, since the soluble fraction of aggregation-prone proteins is degraded at the same rate in clpB and wild-type cells (Inoue and Rechsteiner, 1994). Mutants of clpB, the protease Lon, and the protease subunits ClpA and ClpP are all impaired in the rapid removal of protein aggregates that form transiently in heat shocked cells (Laskowska et al., 1996a). 3.5. DNA Replication Much of the insight into the mechanism of DnaK chaperone function has come from studying the role of DnaK in the initiation of DNA replication. DnaK is required for efficient initiation of replication by bacteriophage A, and plasmids P1 and F, and perhaps in the initiation of E. coli chromosome replication (for a recent review, see Chattoraj, 1995). E. coli strains carrying missense mutations in dnaK, dnaJ, and grpE were isolated originally as resistant to killing by bacteriophage (Georgopoulos and Herskowitz, 1971; Sunshine et al., 1977; Saito and Uchida, 1977). growth is blocked in these mutants at the stage of DNA replication initiation (Georgopoulos and Herskowitz, 1971; Georgopoulos, 1977; Sunshine et al., 1977; Saito and Uchida, 1977; Saito and Uchida, 1978). Specific mutations in the P protein, a bacteriophage-encoded replication initiation factor, restore growth on the mutant strains. DnaK, DnaJ, and GrpE proteins are required in vitro to remove P from the replication initiation complex assembled at the origin of replication (reviewed by Zylicz et al., this volume). Removal of P from the complex permits subsequent DNA unwinding by DnaB helicase, which is initially recruited to the origin by P. The P mutants that bypass the requirement for wild-type DnaK, DnaJ, and GrpE activities form less stable complexes with DnaB helicase than wild-type P (Konieczny and Marszalek, 1995). A similar role for the ClpX chaperone during bacteriophage replication has recently been shown. ClpX, which is required for bacteriophage Mu replication in vivo, modifies the conformation of the replication initiation complex in vitro to permit subsequent partial disassembly of the complex and initiation of replication (Mhammedi-Alaoui et al., 1994; Levchenko et al., 1995; Kruklitis et al., 1996). In addition to blocking replication, mutations in dnaK, dnaJ, and grpE reduce the
Genetic evidence for the roles of molecular chaperones
131
efficiency of replication of plasmids carrying a minimal P1 origin of replication (Tilly and Yarmolinsky, 1989; Bukau and Walker, 1989b) and interfere with the autoregulation of the P1 replication initiation protein, RepA (Tilly et al., 1990). DnaK and DnaJ activate RepA for DNA binding by dissociating inactive RepA dimers into monomers (Wickner et al., 1991a; Wickner et al., 1991b; Wickner et al., 1992). RepA mutants with increased activity in dnaK, dnaJ, and grpE mutant hosts appear to form less stable dimers in vitro (Mukhopadhyay et al., 1994). It has been suggested that DnaK not only dissociates RepA into monomers but also activates misfolded monomeric RepA (DasGupta et al., 1993). DnaK, DnaJ, and GrpE are required for mini-F plasmid replication in vivo, as well (Ezaki et al., 1989; Kawasaki et al., 1990). DnaK disassembles dimers of the F plasmid replication factor RepE into active monomers (Ishiai et al., 1994), analogous to its action on RepA. Unlike RepA dimers, RepE dimers bind DNA, but recognize different sites than those bound by the monomer. Like the corresponding RepA mutants, DnaKindependent RepE mutants are defective in dimerization in vitro. When mutants of dnaK or dnaJ are shifted to nonpermissive temperature, DNA synthesis declines after a short lag, whereas protein synthesis continues for a longer time (Saito and Uchida, 1977; Saito and Uchida, 1978; Itikawa and Ryu, 1979; Wada et al., 1982; Sakakibara, 1988). It was on the basis of this replication phenotype that the genes were designated dna, grouping them together with other genes required for host DNA synthesis. DNA synthesis is also reduced relative to protein synthesis in a grpE280 mutant at nonpermissive temperature (Ang et al., 1986). Replication of the E. coli chromosome is initiated at a single origin, oriC, by the ATPbound form of the initiation factor, DnaA. When dnaK mutants are shifted to nonpermissive temperature, chromosome replication continues to completion and then ceases (Sakakibara, 1988). Upon return to permissive temperature, a burst of replication at oriC occurs, indicating that the mutants had been blocked at the stage of replication initiation (Sakakibara, 1988; Ohki and Smith, 1989). Consistent with this interpretation, DNA replication persists for several hours at nonpermissive temperature in a dnaK rnh double mutant, in which replication initiation is DnaA-independent. The rnh mutation does not suppress the temperature sensitivity of dnaK mutants, however (Sakakibara, 1988; Bukau and Walker, 1989b), indicating that other essential cellular functions are defective as well (see section 3.6). DnaK may only be required for DnaA-dependent replication at elevated temperatures, since at 30°C a naK52 rpoH mutant grows well (Bukau and Walker, 1990; see section 2.1) and an onC-dependent plasmid is able to replicate in a dnaK756 mutant (Malki et al., 1991). In vitro evidence supports a role for DnaK in DnaA-dependent replication. DnaA forms inactive aggregates containing phospholipids that can be dissociated into active form by incubation with DnaK or phospholipase A2 (Hwang et al., 1990). In vitro, DnaK stimulates replication initiation at oriC by wild-type or mutant DnaA and may protect DnaA from heat inactivation (Zylicz et al., 1985; Hwang and Kaguni, 1991; Malki et al., 1991; Hupp and Kaguni, 1993a; Hupp and Kaguni, 1993b). GroEL/ES may also play a role in E. coli DNA replication. When temperature sensitive groEL and groES mutants are shifted to nonpermissive temperature, DNA synthesis ceases earlier than protein synthesis (Wada and Itikawa, 1984). Genetic evidence is consistent with mutant DnaA being a substrate of GroEL/ES. Two classes of
Molecular chaperones and folding catalysts
132
thermosensitive dnaA mutants have been identified. The first class, including dnaA46, recovers activity when returned to permissive temperature. The second is irreversibly inactivated. Overexpression of GroEL/ES suppresses the thermosensitivity of the first but not the second class of mutants (Jenkins et al., 1986; Fayet et al., 1986). Curiously, overexpressing GroEL/ES in either a dnaA46 mutant strain or a dnaA46/dnaA + merodiploid strain confers a cold-sensitive phenotype that is not seen when GroEL/ES is overexpressed in a dnaA + strain (Hansen et al., 1984, Fayet et al., 1986;Jenkins et al., 1986; Katayama and Nagata, 1991). Though the genetic data suggests a role for GroEL/ES in activating a DnaA mutant, this may simply represent the ability of the chaperone to bind improperly folded proteins rather than indicate a role for GroEL/ES in the function of wild-type DnaA. 3.6. Cell Division Mutants of dnaK, dnaJ, and grpE filament without septa at permissive temperatures. Filamentation is more extensive when the mutants are shifted to nonpermissive temperatures (Georgopoulos and Herskowitz, 1971; Paek and Walker, 1987; Bukau and Walker, 1989a; Kang and Craig, 1990; Sell et al., 1990). groEL and groES mutants also filament without septa at nonpermissive temperature (Georgopoulos and Eisen, 1974; Wada and Itikawa, 1984; Horwich et al., 1993), as do cells with limiting or overproduced trigger factor (Guthrie and Wickner, 1990). Since mutations in a large number of other genes confer filamentation phenotypes, it has been difficult to determine the cellular targets affected by the chaperone mutations. Filamentation of dnaK52 mutants can be fully suppressed by the sidB mutant alleles of rpoH, which lower hsp expression (see section 2.1). This suggests that filamentation results from the overexpression of hsps or the consequent down-regulation of other proteins. It also implies that dnaK either does not play a direct role in cell division or is compensated for by a redundant activity. It should be noted that rpoH mutations that lower hsp expression more than the sidB alleles induce filamentation at nonpermissive temperature (Tsuchido et al., 1986). Thus, inappropriately low levels of hsp expression also block cell division. Filamentation of dnaK mutants is also suppressed when cells are grown in minimal glucose medium instead of rich medium (Georgopoulos et al., 1994; Rockabrand et al., 1995). Perhaps as a consequence, dnaK mutants retain higher viability and are less likely to acquire suppressing mutations when cultured in minimal medium. A factor required for cell division may be limiting in dnaK mutants grown in rich medium due to faster growth or to differences in gene expression in rich vs. minimal media. Filamentation might result from induction of a specific inhibitor of cell division or from impaired chromosome partition or septum formation. The block in cell division in dnaK52 mutants does not result from induction of the cell division inhibitor SulA (Paek and Walker, 1987; Bukau and Walker, 1989a; Bukau and Walker, 1989b). Mutants of ftsZ, which encodes a major structural component of the septum, also filament without septa. One possibility is that FtsZ may be malfunctioning in dnaK52 mutants. Overexpressing FtsZ does suppress filamentation of a dnaK52 mutant at 30°C (Bukau and Walker, 1989a; Bukau and Walker, 1989b). The mutants still grow poorly, however, and acquire subpopulations of minicells and anucleate cells. Furthermore, ftsZmutants
Genetic evidence for the roles of molecular chaperones
133
differ from dnaK52 mutants in that they exhibit normal nucleoid segregation (Bukau and Walker, 1989b). Thus, although FtsZ may be malfunctioning in dnaK52 mutants, the mutants are probably also blocked at another stage of cell division. Donachie (1993) has noted that the nucleoid segregation defect of dnaK52 strains resembles that of mukB null and mukB106 missense mutants at nonpermissive temperature (Niki et al., 1991). MukB is a myosin-like protein that may play a role in chromosome segregation. The similarity of the dnaK and mukB phenotypes is strengthened by the finding that overexpressing DksA, which suppresses the temperaturesensitivity of a mukB106 mutant (Yamanaka et al., 1994), also partially suppresses the filamentation and temperature-sensitivity of dnaK, dnaJ, and grpE mutants (Kang and Craig, 1990).dksA is homologous to several other bacterial and bacteriophage genes of unknown function (Bass et al., 1996). In addition to suppressing mukB and dnaK mutations, dksA has been isolated as a multicopy or missense suppressor of two other kinds of mutations, both related to cell division: 1. Overexpressing DksA suppresses the temperature-sensitive and filamentation phenotypes of a null mutation in prc, which encodes a periplasmic protease that is involved in the proteolytic processing of peptidoglycan precursors and the essential cell division protein PBP3 (Bass et al., 1996). PBP3 is required for peptidoglycan synthesis at the septum. 2. A missense mutation in dksA suppresses the replication initiation defects of pSC101 plasmids carrying mutations at the origin of replication in the replication enhancer element (Ohkubo and Yamaguchi, 1997). The missense allele also suppresses the growth defects of a dnaK mutation. The allele appears to be a gain-of-function mutation, because a null mutation in dksA has no effect on pSC101 replication). The finding that dksA suppresses defects in replication initiation, chromosome segregation, and septation conferred by various mutations and also suppresses dnaK mutant growth phenotypes may indicate that all the mutations affect some common process at a key stage of cell division. Alternatively, dksA may be a general bypass suppressor of various cell division defects. This is unlikely, however, since overexpressing DksA has no effect on the growth and filamentation phenotypes of groEL or groES mutants (Kang and Craig, 1990). A null mutation in dksA has no apparent effect on cell growth in rich media but retards growth in minimal media (Kang and Craig, 1990). The regulation of peptidoglycan biosynthesis is intimately coupled with septum formation and cell division (Nanninga, 1991; Donachie, 1993). dksA establishes a tentative link between the cell division defects of dnaK mutants and the defects in peptidoglycan synthesis caused by prc mutants. A more direct connection between dnaK and peptidoglycan synthesis is made by a second multicopy suppressor of dnaK, dapE (msgB), which is an essential gene in the peptidoglycan biosynthetic pathway (Wu et al., 1992). Overexpressing DapE partially suppresses the growth defects of dnaK, dnaJ, and grpE mutants. As with dksA, dapE may compensate for a specific defect of dnaK or instead be a nonspecific bypass suppressor. Very little is known about the basis for the filamentation phenotypes of groEL or groES mutants. However, the role of GroEL/ES in activating the F plasmid encoded
Molecular chaperones and folding catalysts
134
inhibitor of cell division, LetD, has been characterized in some detail. LetD (ccdB) and a second F plasmid encoded factor, LetA (ccdA), form a suicide/antidote system that blocks division of cells that have lost the plasmid (for review, see Jensen and Gerdes, 1995). LetD is a stable protein that blocks chromosome partitioning by binding and inhibiting DNA gyrase subunit A. LetA is an unstable protein, degraded by the lon protease (Van Melderen et al., 1994), that inhibits LetD activity by dissociating LetD from DNA gyrase. LetD activity is inhibited by mutations in groEL, groES, or two other host genes, tldD and tldE (Miki et al., 1988; Miki et al., 1992; Murayama et al., 1996). The activity of LetD is restored in these mutants by inactivation of csrA. CsrA, which is induced upon entry into stationary phase, regulates carbon storage by repressing glycogen biosynthesis and gluconeogenesis. The simplest model for these genetic data is that csrA or another gene under its control inhibits LetD activity but is itself inhibited by tldD/E and groEL/ES. The function of GroEL/ES might be to activate TldD or TldE. Since single and double null mutants of tldD and tldE have no discernible effect on growth in cells lacking the F plasmid, it is not clear what role they normally play. Interestingly, a subset of hsps that includes DnaK and GroEL is induced by LetD and other inhibitors of DNA gyrase, such as nalidixic acid and novobiocin (Krueger and Walker, 1984; VanBogelen et al., 1987b; Kaneko et al., 1996). 3.7. Acquired Stress Tolerance and Survival during Starvation Cells develop resistance to otherwise lethal exposures of stresses by at least two mechanisms, prior exposure to sublethal levels of certain stresses during exponential growth (acquired stress tolerance) and starvation-induced stress tolerance. The mechanism and regulation of acquired stress tolerance are still poorly understood (reviewed by Parsell et al., 1994; Li et al., this volume), whereas the regulation of starvation-induced stress tolerance is somewhat better understood (for recent reviews, see Hengge-Aronis, 1996; Loewen and Hengge-Aronis, 1994; Nyström, 1995). The response to starvation is mediated principally by the alternative sigma factor, , which initiates the transition into stationary phase. In response to induction, starved cells synthesize trehalose, which protects against heat and osmotic stress, and undergo other global changes in macromolecular synthesis and cell morphology that are likely to protect against stress. Since hsps are required for basal stress tolerance and are induced by starvation and other stresses, they are likely to be important for both acquired and starvation-induced stress tolerance, but their roles are not yet clear. The -dependent induction of hsps following heat shock is necessary for cells to acquire thermotolerance (Yamamori and Yura, 1982). In contrast, hsps induced by overexpression of from a heterologous promoter do not confer thermotolerance (VanBogelen et al., 1987a). The patterns of hsp induction when is overexpressed differs from those seen during heat shock. This raises the possibility that is not sufficient to induce one or more hsps required for thermotolerance. A comparison of the hsps induced by stresses that do or do not confer thermotolerance failed to identify a set of hsps uniquely associated with acquired thermotolerance (VanBogelen et al., 1987a). Thus, proteins other than hsps are likely to be necessary for the establishment of thermotolerance. These might include proteins under the regulation of , a subset of
Genetic evidence for the roles of molecular chaperones
135
which are induced by heat shock (Muffler et al., 1997). Conflicting results have been reported, however, regarding the role of in acquired thermotolerance. Under similar conditions, has been found to be necessary (McCann et al., 1991) or unnecessary (Hengge-Aronis et al., 1991) for acquired thermotolerance. Deletion of the gene for , rpoE, has indicated that the regulon is not necessary for acquired thermotolerance (Rouvière et al., 1995). Conflicting results have also been reported for the role of dnaK in acquired stress tolerance. Mutants of dnaK that are temperature sensitive at 42°C fail to acquire thermotolerance induced by heat shock at 42°C (Taniguchi et al., 1989; Delaney, 1990; Wild et al., 1993; Rockabrand et al., 1995). In contrast, the dnaK756 mutant, which is temperature resistant below 43°C, still acquires thermotolerance following heat shock at 42°C (Ramsay, 1988). Perhaps these results are explained by the temperature sensitivities of the dnaK mutants during the initial heat shock. Some cellular function may be impaired in the more defective dnaK mutants at 42°C, blocking the acquisition of thermotolerance, though not reducing viability during the heat shock period. This would indicate some role for dnaK during the acquisition of thermotolerance. The role of dnaK in acquired resistance to other stresses, such as exposure to H2O2, is not clear. In one study, a nonsense mutant of dnaK did not develop H2O2-induced H2O2 resistance (Taniguchi et al., 1989), whereas a dnaK52 mutant did under similar conditions in another study (Rockabrand et al., 1995). In S. cerevisiae, Hsp104 plays a major role in mediating acquired stress tolerance (Sanchez and Lindquist, 1990; see Lindquist and Schirmer, this volume). ClpB is the E. coli member of the hsp100 family of chaperones most closely related to Hsp104. Like Hsp104, ClpB appears to solubilize cytosolic protein aggregates (see section 3.4). However, null mutants of clpB are not defective for acquired thermotolerance, although they grow more slowly at 44° C and lose viability somewhat faster than a wild-type strain at 50°C (Squires et al., 1991). A closely related clpB homolog in a strain of the cyanobacterium Synechococcus sp. contributes to acquired thermotolerance (Eriksson et al., 1996). Both Hsp104 and its mitochondrial homolog in S. cerevisiae, Hsp78, can partially compensate for reduced Hsp70 function (Sanchez et al., 1993; Schmitt et al., 1995). Whether a similar redundancy might exist in E. coli between DnaK and ClpB has not been explored. Small hsps also confer stress tolerance in some organisms. Two members of the small hsp family have been identified in E. coli, IbpA and IbpB, but no studies of their role in stress tolerance have been reported (see section 4.2). The regulation of starvation-induced stress tolerance is quite distinct from that of acquired stress tolerance, and the roles of hsps also appear to differ. -dependent induction of hsp expression is not required for short-term viability of cells during starvation or for development of starvation-induced thermotolerance and H2O2 resistance, although a subset of the -dependent hsps, including DnaK, GroEL/ ES, and HtpG, is induced by glucose starvation. These hsps are not induced by starvation in a rpoH mutant, but the mutant still acquires starvation-induced thermotolerance and H2O2 resistance. A rpoH mutant is fully viable for up to 20 days of glucose starvation, but thereafter loses viability more rapidly than its wild-type parent (Jenkins et al., 1991). In contrast to rpoH, dnaK mutations inhibit many of the adaptive changes that occur during starvation, including development of starvation-induced thermotolerance and
Molecular chaperones and folding catalysts
136
resistance to H2O2 (Rockabrand et al., 1995). Starving dnaK52 mutants do not undergo the shift from a rod-like to a spherical cell morphology exhibited by starving wild-type cells. The mutant cells shorten in length, indicating arrest at an early stage of the transition. These phenotypes resemble those of strains deleted for rpoS, the gene encoding the stationary phase-specific factor, . dnaK52 mutants are, in fact, defective in the induction of and -dependent genes (Muffler et al., 1997; Rockabrand et al., 1998). dnaK appears to be required both for the increased translation and the proteolytic stabilization of as cells enter stationary phase. The effects of dnaK mutants during starvation appear to extend beyond their role in regulation, however. Recent evidence suggests that starvation-induced thermotolerance is mediated both by a -dependent pathway, which is impaired in dnaK52 mutants, and by a separate DnaK-dependent/ -independent pathway (Rockabrand et al., 1998). The starvation responses of dnaK and rpoS mutants also differ in that dnaK mutants retain viability during brief periods of glucose starvation under certain conditions (Rockabrand et al., 1995) whereas rpoS mutants rapidly lose viability. The resistance of dnaK52 mutants to brief periods of glucose starvation is similar to rpoH strains and contrasts to the sensitivity of rpoS strains. Evidently, dnaK mutants are not defective in all aspects of the starvation response. The sensitivities of dnaK mutants to glucose starvation do vary under different conditions, however. dnaK52 mutants remain viable in liquid minimal medium when starved for glucose for up to three days (Rockabrand et al., 1995), whereas dnaK103 rapidly loses viability on minimal plates when starved for the same length of time (Spence et al., 1990). This could reflect differences between the mutant alleles, but it is also possible that high cell densities or drying on plates sensitizes dnaK mutants to carbon starvation. The role of GroEL/ES during stationary phase has not been examined closely. A groEL mutant strain that expresses a truncated form of GroEL lacking the last 16 residues recovers slowly from stationary phase if starved at 42°C (McLennan et al., 1993). The mutant otherwise behaves like wild-type when screened for a range of phenotypes common to groEL mutants. This impairment of stationary phase survival may be allelespecific, or it may have been missed in more defective groEL mutants.
4. THE ORPHANED CHAPERONES OF E. COLI: HOMOLOGS IN SEARCH OF FUNCTION 4.1. HtpG (Hsp90) htpG encodes a protein highly homologous to the eucaryotic hsp90 chaperone family. Expression of htpG is -dependent and is induced under a wide range of stress conditions (Bardwell and Craig, 1987; for review see Buchner, 1996). Null strains of htpG grow nearly as well as wild-type, are mildly temperature sensitive (Bardwell and Craig, 1988), and retain full viability during starvation (Spence et al., 1990). The strongest evidence for HtpG having chaperone activity comes from in vitro studies, where HtpG, like its eucaryotic counterparts, suppresses the aggregation of unstable
Genetic evidence for the roles of molecular chaperones
137
folding intermediates (Jakob et al., 1995). An in vivo role for htpG is suggested by its identification as a multicopy suppressor of the growth defects of strains with limiting FtsH (HflB) expression or function (Shirai et al., 1996). FtsH is an essential integral membrane ATP-dependent protease that degrades and thus negatively regulates both (Herman et al., 1995; Tomoyasu et al., 1995) and, under certain conditions, SecY, an integral membrane component of the protein export machinery (Kihara et al., 1995). Missense mutants of ftsH are defective in protein export and membrane protein localization (Akiyama et al., 1994). Overexpression of HtpG suppresses mislocalization, but has no effect on protein export. In contrast, overexpression of groEL/ES suppresses the export but not the mislocalization defects of ftsH mutants. The mechanisms underlying these phenotypes and the possible roles of HtpG and GroEL/ES remain to be determined. htpG was also previously identified as a multicopy suppressor of the temperature sensitivity of the secY24 mutation (Ueguchi and Ito, 1992), but this may have been an indirect effect (Shirai et al., 1996). The eucaryotic hsp90s act in concert with other cofactors to stabilize intermediates in signal transduction pathways and might participate with hsp70s in the general cytoplasmic folding pathway (for a recent reviews see Buchner, 1996; Johnson and Craig, 1997). HtpG might also work with yet-to-be-defined cofactors in similar roles. 4.2. IbpA and IbpB (Small Hsps) The 16 kd IbpA and IbpB proteins are members of the small hsp (hsp20) family. The two proteins are about 48% identical and share three conserved C-terminal motifs with eucaryotic small hsps (Allen et al., 1992; Chuang et al., 1993; Jakob and Buchner, 1994). IbpA and IbpB are coexpressed from a -dependent operon and are strongly induced by heat shock. They have recently been reported to accumulate in the outer membrane following heat shock (Laskowska et al., 1996b). Both associate with proteins that transiently aggregate in wild-type cells following heat shock and inclusion bodies formed by the overexpression of unstable proteins (Allen et al., 1992; Laskowska et al., 1996b). One role of IbpA/B maybe to protect cells during starvation. A small hsp induced during long-term starvation in Mycobacterium tuberculosis increases stationary-phase survival of M. tuberculosis (Yuan et al., 1996). An effect of starvation on IbpA/B expression in E. coli has not been reported.
5. Chaperones and adaptation: comparing genomes This review has focused only on the genetics of chaperones in E. coli, with little reference to studies in other bacteria. However, one way of trying to understand how the requirements for different chaperones may change with changing conditions is to ask how bacterial species that have adapted to very different lifestyles and growth conditions differ in the chaperones that they use and the ways that they use them. Among the first four bacterial genomes to be completely sequenced, only those of Haemophilus influenzae and Escherichia coli have representatives of all the major chaperone families (Fleischmann et al., 1995; Blattner et al., 1997). Mycoplasma
Molecular chaperones and folding catalysts
138
genitalium lacks hsp90 and so far has no recognizable small hsps (Fraser et al., 1995; Mushegian and Koonin, 1996). The archaea Methanococcus jannaschii and Archaeoglobus fulgidus appear to lack an hsp70 complex entirely, as well as several other chaperone families (Bult et al., 1996; Klenk et al., 1997; see Koonin et al., 1997). This is not characteristic of the archaea generally, since homologs of dnaK have been identified in other archaebacterial species (Macario et al., 1991; Gupta and Singh, 1992; Gupta and Singh, 1994; Smith et al., 1997). The number of members of each chaperone family represented in an organism is also not conserved. E. coli and H. influenzae each have two hsp70 members, homologs of DnaK and Hsc66. By contrast, M. genitalium has only one hsp70. With respect to the hsp60 family, the three bacteria each have only a single homolog, while the nitrogen fixing bacteria Rhizobium meliloti (Rusanganwa and Gupta, 1993; Ogawa and Long, 1995) and Bradyrhizobium japonicum (Fischer et al., 1993) carry four or more groEL genes. The increased number of groEL/groES paralogs in the nitrogen fixing bacteria Rhizobium meliloti and Bradyrhizobium japonicum is matched by an increased specialization in their regulation or function. Expression of one of the groE operons in Bradyrhizobium japonicum is coordinately regulated with the genes involved in nitrogen fixation (Fischer et al., 1993). Similarly, high levels of expression from one of the groE operons of Rhizobium meliloti is required for nitrogen fixation and normal root nodule formation (section 3.1). Other examples of specialized regulation or function of the conserved chaperones are being found in different bacteria. The expression and cellular localization of DnaK and GroEL are regulated in a cell-cycle-dependent manner during the growth and differentiation of Caulobacter crescentus (Reuter and Shapiro, 1987; Gomes et al., 1990; Avedissian and Lopes Gomes, 1996). Several developmental responses of Myxococcus xanthus mediated by cell-cell signaling require a constitutively expressed dnaK homolog, sglK, including social motility and fruiting body formation (Yang et al., 1998). Finally, novel solutions to common problems are turning up in different bacteria. One example of this is the regulation of heat shock gene expression in the Gram-positive bacterium Bacillus subtilis. The expression of dnaK and several other heat shock genes in B. subtilis are principally regulated by a transcriptional repressor, HrcA, rather than by a transcriptional activator as in E. coli (reviewed in Hecker et al., 1996). GroEL promotes the formation of active HrcA and thus down-regulates the heat shock response of B. subtilis, much as DnaK negatively regulates the heat shock response in E. coli (Mogk et al., 1997). Clearly, comparative studies of chaperone regulation and function in divergent species of bacteria are beginning to provide a new and fruitful approach to understanding how chaperones are integrated into the metabolic and adaptive pathways of the cell.
6. ACKNOWLEDGMENTS We gratefully acknowledge former lab members George Gaitanaris and Alexander Gragerov for many years of fruitful and instructive collaboration and discussion. We thank Alexander Gragerov, Art Horwich, Susan Lindquist, Christos Panagiotidis, Justina
Genetic evidence for the roles of molecular chaperones
139
Voulgaris, Randolph Watnick, and Joanne Burkholder for their comments on the manuscript, and Bernd Bukau, Alexander von Gabain, and Susan Gottesman for communicating results prior to publication. Work in our laboratory is supported by NIH grant GM37219.
7. REFERENCES Akiyama, Y., Ogura, T. and Ito, K. (1994). Involvement of FtsH in protein assembly into and through the membrane. I. Mutations that reduce retention efficiency of a cytoplasmic reporter. J. Biol. Chem. , 269 , 5218–5224. Alix, J.H. and Guérin, M.E (1993). Mutant DnaK chaperones cause ribosome assembly defects in Escherichia coli . Proc. Natl Acad. Sci. USA , 90 , 9725–9729. Allen, S.P., Polazzi, J.O., Gierse, J.K. and Easton, A.M. (1992). Two novel heat shock genes encoding proteins produced in response to heterologous protein expression in Escherichia coli. J. Bacterial. , 174 , 6938–6947. Altman, E., Kumamoto, C.A. and Emr, S.D. (1991). Heat-shock proteins can substitute for SecB function during protein export in Escherichia coli. EMBO J. , 10 , 239–245. Ang, D., Chandrasekhar, G.N., Zylicz, M. and Georgopoulos, C. (1986). Escherichia coli grpE gene codes for heat shock protein B25.3, essential for both DNA replication at all temperatures and host growth at high temperature. J. Bacterial. , 167 , 25–29. Ang, D. and Georgopoulos, C. (1989). The heat-shock-regulated grpE gene of Escherichia coli is required for bacterial growth at all temperatures but is dispensable in certain mutant backgrounds . J. Bacterial. , 171 , 2748–2755. Avedissian, M. and Lopes Gomes, S. (1996). Expression of the groESL operon is cellcycle controlled in Caulobacter crescentus. Mol. Microbiol. , 19 , 79–89. Banecki, B., Liberek, K, Wall, D., Wawrzynow, A., Georgopoulos, C., Bertoli, E., Tanfani, F. and Zylicz, M. (1996). Structure-function analysis of the zinc finger region of the DnaJ molecular chaperone. J. Biol. Chem. , 271 , 14840–14848. Bardwell, J.C.A. and Craig, E.A. (1987). Eukaryotic Mr 83,000 heat shock protein has a homologue in Escherichia coli. Proc. Natl Acad. Sci. USA , 84 , 5177–5181. Bardwell, J.C.A. and Craig, E.A. (1988). Ancient heat shock gene is dispensable. J. Bacterial. , 170 , 2977–2983. Barth, M., Marschall, C., Muffler, A., Fischer, D. and Hengge-Aronis, R. (1995). Role for the histone-like protein H-NS in growth phase-dependent and osmotic regulation of and many -dependent genes in Escherichia coli. J. Bacterial. , 177 , 3455–3464. Bass, S., Gu, Q. and Christen, A. (1996). Multicopy suppressors of Prc mutant Escherichia coli include two HtrA (DegP). protease homologs (HhoAB), DksA, and a truncated RlpA. J. Bacterial. , 178 , 1154–1161. Bergès, H., Joseph-Liauzun, E. and Fayet, O. (1996). Combined effects of the signal sequence and the major chaperone proteins on the export of human cytokines in Escherichia coli. Appl. Environ. Microbiol. , 62 , 55–60. Blasco, F., Dos Santos, J.-P, Magalon, A., Frixon, C., Guigliarelli, B., Santini, C.-L. and Giordano, G. (1998). NarJ is a specific chaperone required for molybdenum cofactor assembly in nitrate reductase A of Escherichia coli. Mol. Microbiol. , 28 , 435–448. Blattner, F.R., Plunkett III, G., Bloch, C.A., Perna, N.T., Burland, V., Riley, M., ColladoVides, J., Glasner, J.D., Rode, C.K., Mayhew, G.F. et al. (1997). The complete genome sequence of Escherichia coli K-12. Science , 277 , 1453–1474.
Molecular chaperones and folding catalysts
140
Bochkareva, E., Seluanov, A., Bibi, E. and Girshovich, A. (1996). Chaperonin-promoted post-translational membrane insertion of a multispanning membrane protein lactose permease. J. Biol. Chem. , 271 , 22256–22261. Bochkareva, E.S., Lissin, N.M. and Girshovich, A.S. (1988). Transient association of newly synthesized unfolded proteins with the heat-shock GroEL protein. Nature , 336 , 254–257. Buchberger, A., Schröder, H., Hesterkamp, T., Schönfeld, H.J. and Bukau, B. (1996). Substrate shuttling between the DnaK and GroEL systems indicates a chaperone network promoting protein folding. J. Mol. Biol. , 261 , 328–333. Buchberger, A., Valencia, A., McMacken, R., Sander, C. and Bukau, B. (1994). The chaperone function of DnaK requires the coupling of ATPase activity with substrate binding through residue E171. EMBO J. , 13 , 1687–1695. Buchner, J. (1996). Supervising the fold: functional principles of molecular chaperones. FASEB J. , 10 , 10–19. Bukau, B. (1993). Regulation of the Escherichia coli heat-shock response. Mol. Microbiol. , 9 , 671–680. Bukau, B., Hesterkamp, T. and Luirink, J. (1996). Growing up in a dangerous environment: a network of multiple targeting and folding pathways for nascent polypeptides in the cytosol. Trends Cell Biol. , 6 , 480–486. Bukau, B. and Walker, G.C. (1989a). Cellular defects caused by deletion of the Escherichia coli dnaK gene indicate roles for heat shock protein in normal metabolism. J. Bacterial. , 171 , 2337–2346. Bukau, B. and Walker, G.C. (1989b). dnaK52 mutants of Escherichia coli have defects in chromosome segregation and plasmid maintenance at normal growth temperatures. J. Bacterial. , 171 , 6030–6038. Bukau, B. and Walker, G.C. (1990). Mutations altering heat shock specific subunit of RNA polymerase suppress major cellular defects of E. coli mutants lacking the DnaK chaperone. EMBO J. , 9 , 4027–4036. Bult, C.J., White, O., Olsen, G.J., Zhou, L., Fleischmann, R.D., Sutton, G.G., Blake, J.A., FitzGerald, L.M., Clayton, R.A., Gocayne, J.D. et al., (1996). Complete genome sequence of the methanogenic archaeon, Methanococcus jannaschii. Science , 273 , 1058–1073. Burkholder, W.F., Zhao, X., Zhu, X., Hendrickson, W.A., Gragerov, A., and Gottesman, M.E. (1996). Mutations in the C-terminal fragment of DnaK affecting peptide binding. Proc. Natl. Acad. Sci. USA , 93 , 10632–10637. Burston, S.G., Weissman, J.S., Farr, G.W., Fenton, W.A. and Horwich, A.L. (1996). Release of both native and non-native proteins from a as-only GroEL ternary complex. Nature , 383 , 96–99. Cashel, M., Gentry, D.R., Hernandez, V.J. and Vinella, D. (1996). The stringent response. In F.C. Neidhardt, (ed.), Escherichia coli and Salmonella: Cellular and Molecular Biology , 2nd ed.,Vol. 1, ASM Press, Washington, B.C., pp. 1458–1496. Chanda, P.K., Ono, M., Kuwano, M. and Kung, H. (1985). Cloning, sequence analysis, and expression of alteration of the mRNA stability gene (ams + ) of Escherichia coli. J. Bacterial. , 161 , 446–449. Chattoraj, D.K. (1995). Role of molecular chaperones in the initiation of plasmid DNA replication. Genet. Eng. , 17, 81–98. Chuang, S.-E., Burland, V., Plunkett III, G., Daniels, D.L. and Blattner, F.R. (1993). Sequence analysis of four new heat-shock genes constituting the hslTS/ibpAB and hslVU operons in Escherichia coli. Gene , 134, 1–6.
Genetic evidence for the roles of molecular chaperones
141
Clarke, D.J., Jacq, A. and Holland, I.B. (1996). A novel DnaJ-like protein in Escherichia coli inserts into the cytoplasmic membrane with a type III topology. Mol. Microbiol. , 20 , 1273–1286. Clarke, D.J., Holland, I.B. and Jacq, A. (1997). Point mutations in the transmembrane domain of DjlA, a membrane-linked DnaJ-like protein, abolish its function in promoting colanic acid production via the Rcs signal transduction pathway. Mol. Microbiol. , 25 , 933–944. Collier, D.N. (1993). SecB: a molecular chaperone of Escherichia coli protein secretion pathway. Adv. Protein Chem. , 44 , 151–193. Craig, E.A. and Gross, C.A. (1991). Is hsp70 the cellular thermometer? Trends Biochem. Sci. , 16 , 135–140. Crooke, E., Guthrie, B., Lecker, S., Lill, R. and Wickner, W. (1988). ProOmpA is stabilized for membrane translocation by either purified E. coli trigger factor or canine signal recognition particle. Cell , 54 , 1003–1011. Crooke, E. and Wickner, W. (1987). Trigger factor: a soluble protein that folds proOmpA into a membrane-assembly-competent form. Proc. Natl. Acad. Sci. USA , 84 , 5216–5220. Cyr, D.M., Langer, T. and Douglas, M.G. (1994). DnaJ-like proteins: molecular chaperones and specific regulators of Hsp70. Trends Biochem. Sci. , 19 , 176–181. Damerau, K. and St. John, A.C. (1993). Role of Clp protease subunits in degradation of carbon starvation proteins in Escherichia coli. J. Bacterial. , 175 , 53–63. Danese, P.N., Murphy, C.K. and Silhavy, T.J. (1995). Multicopy suppression of coldsensitive sec mutations in Escherichia coli. J. Bacterial. , 177 , 4969–4973. DasGupta, S., Mukhopadhyay, G., Papp, P.P., Lewis, M.S. and Chattoraj, D.K. (1993). Activation of DNA binding by the monomeric form of the P1 replication initiator RepA by heat shock proteins DnaJ and DnaK. J. Mol. Biol. , 232 , 23–34. Delaney, J.M. (1990). Requirement of the Escherichia coli dnaK gene for thermotolerance and protection against H2O2. J. Gen. Microbiol. , 136 , 2113–2118. den Blaauwen, T. and Driessen, A.J.M. (1996). Sec-dependent preprotein translocation in bacteria. Arch. Microbiol. , 165 , 1–8. Donachie, W.D. (1993). The cell cycle of Escherichia coli. Annu. Rev. Microbiol. , 47 , 199–230. Donnelly, C.E. and Walker, G.C. (1989). groE mutants of Escherichia coli are defective in umuDC-dependent UV mutagenesis. J. Bacterial. , 171 , 6117–6125. Donnelly, C.E. and Walker, G.C. (1992). Coexpression of UmuD’ with UmuC suppresses the UV mutagenesis deficiency of groE mutants. J. Bacterial. , 174 , 3133– 3139. Eriksson, M.-J. and Clarke, A.K. (1996). The heat shock gene ClpB mediates the development of thermotolerance in the cyanobacterium Synechococcus sp. strain PCC 7942. J. Bacterial. , 178 , 4839–4846. Ezaki, B., Ogura, T., Mori, H., Niki, H. and Hiraga, S. (1989). Involvement of DnaK protein in mini-F plasmid replication: temperature-sensitive seg mutations are located in the dnaK gene. Mol. Gen. Genet. , 218 , 183–189. Fayet, O., Louarn, J.-M. and Georgopoulos, C. (1986). Suppression of the Escherichia coli dnaA46 mutation by amplification of the groES and groEL genes. Mol. Gen. Genet. , 202 , 435–445. Fayet, O., Ziegelhoffer, T. and Georgopoulos, C. (1989). The groES and groEL heat shock gene products of Escherichia coli are essential for bacterial growth at all temperatures. J. Bacterial , 171 , 1379–1385.
Molecular chaperones and folding catalysts
142
Fenton, W.A., Kashi, Y., Furtak, K. and Horwich, A.L. (1994). Residues in chaperonin GroEL required for polypeptide binding and release. Nature , 371 , 614–619. Fischer, H.M., Babst, M., Kaspar, T., Acuna, G., Arigoni, F. and Hennecke, H. (1993). One member of a groESL-like chaperonin multigene family in Bradyrhizobium japonicum is co-regulated with symbiotic nitrogen fixation genes. EMBO J. , 12 , 2901–2912. Fleischmann, R.D., Adams, M.D., White, O., Clayton, R.A., Kirkness, E.F., Kerlavage, A.R., Bult, C.J., Tomb, J.-F., Dougherty, B.A., Merrick, J.M., et al., (1995). Wholegenome random sequencing and assembly of Haemophilus influenzae Rd. Science , 269 , 496–512. Fraser, C.M., Gocayne, J.D., White, O., Adams, M.D., Clayton, R.A., Fleischmann, R.D., Bult, C.J., Kerlavage, A.R., Sutton, G., Kelley, J.M., et al., (1995). The minimal gene complement of Mycoplasma genitalium. Science , 270 , 397–403. Gaitanaris, G.A., Vysokanov, A., Hung, S.-C., Gottesman, M.E. and Gragerov, A. (1994). Successive action of Escherichia coli chaperones in vivo. Mol. Microbiol. , 14 , 861–869. Galat, A. and Metcalfe, S. (1995). Peptidylproline cis/trans isomerases. Prog. Biophys. Mol. Biol. , 63 , 67–118. Georgellis, D., Sohlberg, B., Hartl, F.U. and von Gabain, A. (1995). Identification of GroEL as a constituent of an mRNA-protection complex in Escherichia coli. Mol. Microbiol. , 16 , 1259–1268. Georgopoulos, C., Ang, D., Liberek, K. and Zylicz, M. (1990). Properties of the Escherichia coli heat shock proteins and their role in bacteriophage growth . In R.I.Morimoto, A.Tissières and C.Georgopoulos, (eds.), Stress Proteins in Biology and Medicine , Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 191– 221. Georgopoulos, C., Liberek, K., Zylicz, M. and Ang, D. (1994). Properties of the heat shock proteins of Escherichia coli and the autoregulation of the heat shock response. In R.I.Morimoto, A.Tissières and C.Georgopoulos, (eds.), The Biology of Heat Shock Proteins and Molecular Chaperones , Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 209–249. Georgopoulos, C.P. and Herskowitz, I. (1971). Escherichia coli mutants blocked in DNA synthesis. In A.D.Hershey, (ed.), The Bacteriophage Lambda , Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, pp. 553–564. Georgopoulos, C.P. (1977). A new bacterial gene (groPC). which affects DNA replication. Mol. Gen. Genet. , 151 , 35–39. Georgopoulos, C.P. and Eisen, H. (1974). Bacterial mutants which block phage assembly. J. Supramol. Struct. , 2 , 349–359. Gomes, S.L., Gober, J.W. and Shapiro, L. (1990). Expression of the Caulobacter heat shock gene dnaK is developmentally controlled during growth at normal temperatures. J. Bacterial. , 172 , 3051–3059. Gottesman, S. (1996). Proteases and their targets in Escherichia coli. Annu. Rev. Genet. , 30 , 465–506. Gottesman, S., Maurizi, M.R. and Wickner, S. (1997). Regulatory subunits of energydependent pro teases. Cell , 91 , 435–438. Govezensky, D., Bochkareva, E.S., Zamir, A. and Girshovich, A.S. (1994). Chaperonins as potential gene regulatory factors. In vitro interaction and solubilization of NifA, the nif transcriptional activator, with GroEL. J. Biol. Chem. , 269 , 14003–14006. Govezensky, D., Greener, T., Segal, G. and Zamir, A. (1991). Involvement of GroEL in
Genetic evidence for the roles of molecular chaperones
143
nif gene regulation and nitrogenase assembly. J. Bacterial. , 173 , 6339–6346. Gragerov, A. (1997). Major heat shock chaperones in E. coli protein folding and inclusion body formation. In A.Fink and Y.Goro (ed.), Molecular chaperones in the life cycle of proteins: Strutures, function, and mode of action . Marcel Dekker, New York, 1998. Gragerov, A., Zeng, L., Zhao, X., Burkholder, W. and Gottesman, M.E. (1994). Specificity of DnaK-peptide binding. J. Mol. Biol. , 235 , 848–854. Gragerov, A., Nudler, E., Komissarova, N., Gaitanaris, G.A., Gottesman, M.E. and Nikiforov, V. (1992). Cooperation of GroEL/GroES and DnaK/DnaJ heat shock proteins in preventing protein misfolding in Escherichia coli. Proc. Natl. Acad. Sci. USA , 89 , 10341–10344. Gragerov, A.I., Martin, E.S., Krupenko, M.A., Kashlev, M.V. and Nikiforov, V.G. (1991). Protein aggregation and inclusion body formation in Escherichia coli rpoH mutant defective in heat shock protein induction. FEBS Lett. , 291 , 222–224. Gross, C.A. (1996). Function and regulation of the heat shock proteins. In F.C.Neidhardt, (ed.), Escherichia coli and Salmonella: Cellular and Molecular Biology , 2nd ed.,Vol. 1, ASM Press, Washington, DC, pp. 1382–1399. Grossman, A.D., Straus, D.B., Walter, W.A. and Gross, C.A. (1987). synthesis can regulate the synthesis of heat shock proteins in Escherichia coli . Genes Dev. , 1 , 179– 184. Gupta, R.S. and Singh, B. (1992). Cloning of the HSP70 gene from Halobacterium marismortui: relatedness of archaebacterial HSP70 to its eubacterial homologs and a model for the evolution of the HSP70 gene. J. Bacterial. , 174 , 4594–4605. Gupta, R.S. and Singh, B. (1994). Phylogenetic analysis of 70 kD heat shock protein sequences suggests a chimeric origin for the eukaryotic cell nucleus. Curr. Biol. , 4 , 1104–1114. Guthrie, B. and Wickner, W. (1990). Trigger factor depletion or overproduction causes defective cell division but does not block protein export. J. Bacteriol. , 172 , 5555– 5562. Hansen, E.B., Atlung, T., Hansen, F.G., Skovgaard, O. and von Meyenburg, K. (1984). Fine structure genetic map and complementation analysis of mutations in the dnaA gene of Escherichia coli. Mol. Gen. Genet. , 196 , 387–396. Hecker, M., Schumann, W. and Volker, U. (1996). Heat-shock and general stress response in Bacillus subtilis . Mol. MicrobioL , 19 , 417–428. Hengge-Aronis, R. (1996). Regulation of gene expression during entry into stationary phase. In F.C. Neidhardt, (ed.), Escherichia coli and Salmonella: Cellular and Molecular Biology , 2nd ed., Vol. 1, ASM Press, Washington, DC, pp. 1497–1512. Hengge-Aronis, R., Klein, W., Lange, R., Rimmele, M. and Boos, W. (1991). Trehalose synthesis genes are controlled by the putative sigma factor encoded by rpoS and are involved in stationary-phase thermotolerance in Escherichia coli. J. Bacteriol. , 173 , 7918–7924. Herman, C., Thévenet, D., D’Ari, R. and Bouloc, P. (1995). Degradation of , the heat shock regulator in Escherichia coli, is governed by HflB. Proc. Natl. Acad. Set. USA , 92 , 3516–3520. Hesterkamp, T. and Bukau, B. (1996). The Escherichia coli trigger factor. FEBS Lett. , 389 , 32–34. Hesterkamp, T., Hauser, S., Lütcke, H. and Bukau, B. (1996). Escherichia coli trigger factor is a prolyl isomerase that associates with nascent polypeptide chains. Proc. Natl. Acad. Sci. USA , 93 , 4437–4441.
Molecular chaperones and folding catalysts
144
Horwich, A.L., Low, K.B., Fenton, W.A., Hirshfield, I.N. and Furtak, K. (1993). Folding in vivo of bacterial cytoplasmic proteins: role of GroEL. Cell , 74 , 909–917. Hunt, J.F., van der Vies, S.M., Henry, L. and Deisenhofer, J. (1997). Structural adaptations in the specialized bacteriophage T4 co-chaperonin Gp31 expand the size of the Anfinsen cage. Cell , 90 , 361–371. Hupp, T.R. and Kaguni, J.M. (1993a). Activation of DnaA5 protein by GrpE and DnaK heat shock proteins in initiation of DNA replication in Escherichia coli. J. Biol. Chem. , 268 , 13137–13142. Hupp, T.R. and Kaguni, J.M. (1993b). Activation of mutant forms of DnaA protein of Escherichia coli by DnaK and GrpE proteins occurs prior to DNA replication. J. Biol. Chem. , 268 , 13143–13150. Hwang, D.S., Crooke, E. and Kornberg, A. (1990). Aggregated dnaA protein is dissociated and activated for DNA replication by phospholipase or dnaK protein. J. Biol. Chem. , 265 , 19244–19248. Hwang, D.S. and Kaguni, J.M. (1991). dnaK protein stimulates a mutant form of dnaA protein in Escherichia coli DNA replication. J. Biol. Chem. , 266 , 7537–7541. Inoue, I. and Rechsteiner, M. (1994). On the relationship between the metabolic and thermodynamic stabilities of T4 lysozymes: measurements in Escherichia coli. J. Biol. Chem. , 269 , 29241–29246. Ishiai, M., Wada, C., Kawasaki, Y. and Yura, T. (1994). Replication initiator protein RepE of mini-F plasmid: functional differentiation between monomers (initiator), and dimers (autogenous represser). Proc. Natl. Acad. Sci. USA , 91 , 3839–3843. Itikawa, H., Fujita, H. and Wada, M. (1986). High temperature induction of a stringent response in the dnaK(Ts) and dnaJ(Ts) mutants of Escherichia coli. J. Biochem. , 99 , 1719–1724. Itikawa, H. and Ryu, J. (1979). Isolation and characterization of a temperature-sensitive dnaK mutant of Escherichia coli B. J. Bacteriol. , 138 , 339–344. Jakob, U. and Buchner, J. (1994). Assisting spontaneity: the role of Hsp90 and small Hsps as molecular chaperones. Trends Biochem. Sci. , 19, 205–211. Jakob, U., Lilie, H., Meyer, I. and Buchner, J. (1995). Transient interaction of Hsp90 with early unfolding intermediates of citrate synthase. Implications for heat shock in vivo. J. Biol. Chem. , 270 , 7288–7294. Jenkins, A.J., March, J.B., Oliver, I.R. and Masters, M. (1986). A DNA fragment containing the groE genes can suppress mutations in the Escherichia coli dnaA gene . Mol. Gen. Genet. , 202 , 446–454. Jenkins, D.E., Auger, E.A. and Matin, A. (1991). Role of RpoH, a heat shock regulator protein, in Escherichia coli carbon starvation protein synthesis and survival. J. Bacteriol , 173 , 1992–1996. Jensen, R.B. and Gerdes, K. (1995). Programmed cell death in bacteria: proteic plasmid stabilization systems. Mol. Microbiol. , 17 , 205–210. Johnson, J.L. and Craig, E.A. (1997). Protein folding in vivo: unraveling complex pathways. Cell , 90 , 201–204. Jubete, Y., Maurizi, M.R. and Gottesman, S. (1996). Role of the heat shock protein DnaJ in the lon-dependent degradation of naturally unstable proteins. J. Biol. Chem. , 271 , 30798–30803. Kamath-Loeb, A.S., Lu, C.Z., Suh, W.-C., Lonetto, M.A. and Gross, C.A. (1995). Analysis of three DnaK mutant proteins suggests that progression through the ATPase cycle requires conformational changes. J. Biol. Chem. , 270 , 30051–30059. Kandror, O. and Goldberg, A.L. (1997). Trigger factor is induced upon cold shock and
Genetic evidence for the roles of molecular chaperones
145
enhances viability of Escherichia coli at low temperatures . Proc. Natl. Acad. Set. USA , 94 , 4978–4981. Kandror, O., Busconi, L., Sherman, M. and Goldberg, A.L. (1994). Rapid degradation of an abnormal protein in Escherichia coli involves the chaperones GroEL and GroES. J. Biol. Chem. , 269 , 23575–23582. Kandror, O., Sherman, M., Rhode, M. and Goldberg, A.L. (1995). Trigger factor is involved in GroEL-dependent protein degradation in Escherichia coli and promotes binding of GroEL to unfolded proteins. EMBO J. , 14 , 6021–6027. Kandror, O., Sherman, M., Moerschell, R. and Goldberg, A.L. (1997). Trigger factor associates with GroEL in vivo and promotes its binding to certain polypeptides. J. Biol. Chem. , 272 , 1730–1734. Kaneko, T., Mizushima, T., Ohtsuka, Y., Kurokawa, K., Kataoka, K., Miki, T. and Sekimizu, K. (1996). Co-induction of DNA relaxation and synthesis of DnaK and GroEL proteins in Escherichia coli by expression of LetD (CcdB). protein, an inhibitor of DNA gyrase encoded by the F factor. Mol. Gen. Genet. , 250 , 593–600. Kanemori, M., Mori, H. and Yura, T. (1994). Effects of reduced levels of GroE chaperones on protein metabolism: enhanced synthesis of heat shock proteins during steady-state growth of Escherichia coli . J. Bacteriol. , 176 , 4235–4242. Kang, P.J. and Craig, E.A. (1990). Identification and characterization of a new Escherichia coli gene that is a dosage-dependent suppressor of a dnaK deletion mutation. J. Bacteriol. , 172 , 2055–2064. Katayama, T. and Nagata, T. (1991). Initiation of chromosomal DNA replication which is stimulated without oversupply of DnaA protein in Escherichia coli. Mol. Gen. Genet. , 226 , 491–502. Kawasaki, Y., Wada, C. and Yura, T. (1990). Roles of Escherichia coli heat shock proteins DnaK, DnaJ and GrpE in mini-F plasmid replication. Mol. Gen. Genet. , 220 , 277–282. Kawula, T.H. and Lelivelt, M.J. (1994). Mutations in a gene encoding a new Hsp70 suppress rapid DNA inversion and bgl activation, but not proU derepression, in hns-1 mutant Escherichia coli . J. Bacteriol. , 176, 610–619. Keller, J.A. and Simon, L.D. (1988). Divergent effects of a dnaK mutation on abnormal protein degradation in Escherichia coli. Mol. Microbiol. , 2 , 31–41. Kelley, W.L. and Georgopoulos, C. (1997). Positive control of the two-component RcsC/B signal transduction network by DjlA: a member of the DnaJ family of molecular chaperones in Escherichia coli. Mol. Microbiol. , 25 , 913–931. Kihara, A., Akiyama, Y. and Ito, K. (1995). FtsH is required for proteolytic elimination of uncomplexed forms of SecY, an essential protein translocase subunit. Proc. Natl. Acad. Sci. USA , 92, 4532–4536. Kim, J., Lee, Y., Kim, C. and Park, C. (1992). Involvement of SecB, a chaperone, in the export of ribose-binding protein. J. Bacteriol , 174 , 5219–5227. King, J., Haase-Pettingell, C., Robinson, A.S., Speed, M. and Mitraki, A. (1996). Thermolabile folding intermediates: inclusion body precursors and chaperonin substrates. FASEB J. , 10 , 57–66. Klenk, H.P., Clayton, R.A., Tomb, J.F., White, O., Nelson, K.E., Ketchum, K.A., Dodson, R.J., Gwinn, M., Hickey, E.K., Peterson, J.D. et al. (1997). The complete genome sequence of the hyperthermophilic, sulphate-reducing archaeon Archaeoglobus fulgidus. Nature , 390 , 364–370. Konieczny, I. and Marszalek, J. (1995). The requirement for molecular chaperones in DNA replication is reduced by the mutation in P gene, which weakens the
Molecular chaperones and folding catalysts
146
interaction between P protein and DnaB helicase. J. Biol. Chem. , 270 , 9792–9799. Koonin, E.V., Mushegian, A.R., Galperin, M.Y. and Walker, D.R. (1997). Comparison of archaeal and bacterial genomes: computer analysis of protein sequences predicts novel functions and suggests a chimeric origin for the archaea. Mol. Microbiol. , 25 , 619– 637. Krueger, J.H. and Walker, G.C, (1984). groEL and dnaK genes of Escherichia coli are induced by UV irradiation and nalidixic acid in an htpR+-dependent fashion. Proc. Natl. Acad. Sci. USA , 81 , 1499–1503. Kruklitis, R., Welty, D.J. and Nakai, H. (1996). ClpX protein of Escherichia coli activates bacteriophage Mu transposase in the strand transfer complex for initiation of Mu DNA synthesis. EMBO J. , 15 , 935–944. Kucharczyk, K., Laskowska, E. and Taylor, A. (1991). Response of Escherichia coli cell membranes to induction of d857 prophage by heat shock. Mol. Microbiol. , 5 , 2935–2945. Kudlicki, W., Odom, O.W., Kramer, G. and Hardesty, B. (1994). Activation and release of enzymatically inactive, full-length rhodanese that is bound to ribosomes as peptidyltRNA. J. Biol. Chem. , 269 , 16549–16553 . Kumamoto, C.A. and Beckwith, J. (1985). Evidence for specificity at an early step in protein export in Escherichia coli. J. Bacteriol. , 163 , 267–274. Kusukawa, N. and Yura, T. (1988). Heat shock protein GroE of Escherichia coli: key protective roles against thermal stress. Genes Dev. , 2 , 874–882. Kusukawa, N., Yura, T., Ueguchi, C., Akiyama, Y. and Ito, K. (1989). Effects of mutations in heat-shock genes groES and groEL on protein export in Escherichia coli. EMBO J. , 8 , 3517–3521. Laachouch, J.E., Desmet, L., Geuskens, V., Grimaud, R. and Toussaint, A. (1996). Bacteriophage Mu repressor as a target for the Escherichia coli ATP-dependent Clp pro tease. EMBO J. , 15 , 437–444. Laine, P.S. and Meyer, R.R. (1992). Interaction of the heat shock protein GroEL of Escherichia coli with single-stranded DNA-binding protein: suppression of ssb-113by groEL46 . J. Bacteriol. , 174 , 3204–3211. Laminet, A.A., Kumamoto, C.A. and Plückthun, A. (1991). Folding in vitro and transport in vivo of pre- -lactamase are SecB independent. Mol. Microbiol. , 5 , 117–122. Laminet, A.A., Ziegelhoffer, T., Georgopoulos, C. and Plückthun, A. (1990). The Escherichia coli heat shock proteins GroEL and GroES modulate the folding of the lactamase precursor. EMBO J. , 9 , 2315–2319. Landry, S J., Jordan, R., McMacken, R. and Gierasch, L.M. (1992). Different conformations for the same polypeptide bound to chaperones DnaK and GroEL. Nature , 355 , 455–457. Laskowska, E., Kuczynska-Wisnik, D., Skórko-Glonek, J. and Taylor, A. (1996a). Degradation by proteases Lon, Clp and HtrA, of Escherichia coli proteins aggregated in vivo by heat shock; HtrA pro tease action in vivo and in vitro. Mol. Microbiol. , 22 , 555–571. Laskowska, E., Wawrzynów, A. and Taylor, A. (1996b). IbpA and IbpB, the new heatshock proteins, bind to endogenous Escherichia coli proteins aggregated intracellularly by heat shock. Biochimie , 78 , 117–122. Lee, C.A. and Beckwith, J. (1986). Suppression of growth and protein secretion defects in Escherichia coli secA mutants by decreasing protein synthesis. J. Bacterial. , 166 , 878–883.
Genetic evidence for the roles of molecular chaperones
147
Lelivelt, M.J. and Kawula, T.H. (1995). Hsc66, an Hsp70 homolog in Escherichia coli, is induced by cold shock but not by heat shock. J. Bacterial. , 177 , 4900–4907. Levchenko, I., Luo, L. and Baker, T.A. (1995). Disassembly of the Mu transposase tetramer by the ClpX chaperone. Genes Dev , 9 , 2399–2408. Liberek, K., Marszalek, J., Ang, D., Georgopoulos, C. and Zylicz, M. (1991). Escherichia coli DnaJ and GrpE heat shock proteins jointly stimulate ATPase activity of DnaK Proc. Natl. Acad. Sci. USA , 88 , 2874–2878. Lill, R., Crooke, E., Guthrie, B. and Wickner, W. (1988). The “trigger factor cycle” includes ribosomes, presecretory proteins, and the plasma membrane. Cell , 54 , 1013– 1018. Liu, S.-K. and Tessman, I. (1990). groE genes affect SOS repair in Escherichia coli. J. Bacterial. , 172 , 6135–6138. Liu, X. and DeMoss, J.A. (1997). Characterization of NarJ, a system-specific chaperone required for nitrate reductase biogenesis in Escherichia coli. J. Biol. Chem. , 272 , 24266–24271. Loewen, P.C. and Hengge-Aronis, R. (1994). The role of the sigma factor (KatF) in bacterial global regulation. Annu. Rev. Microbiol. , 48 , 53–80. Lorimer, G.H. (1996). A quantitative assessment of the role of the chaperonin proteins in protein folding in vivo. FASEB J. , 10 , 5–9. Macario, A.J.L., Dugan, C.B. and Conway de Macario, E. (1991). A dnaK homolog in the archaebacterium Methanosarcina mazei S6. Gene , 108 , 133–137. Makrides, S.C. (1996). Strategies for achieving high-level expression of genes in Escherichia coli. Microbiol. Rev. , 60 , 512–538. Malki, A., Hughes, P. and Kohiyama, M. (1991). In vitro roles of Escherichia coli DnaJ and DnaK heat shock proteins in the replication of oriC plasmids. Mol. Gen. Genet. , 225 , 420–426. McCann, M.P., Kidwell, J.P. and Matin, A. (1991). The putative a factor KatF has a central role in development of starvation-mediated general resistance in Escherichia coli. J. Bacteriol. , 173 , 4188–4194. McCarty, J.S. and Walker, G.C. (1994). DnaK mutants defective in ATPase activity are defective in negative regulation of the heat shock response: expression of mutant DnaK proteins results in filamentation. J. Bacteriol. , 176 , 764–780. McLennan, N.F., Girshovich, A.S., Lissin, N.M., Charters, Y. and Masters, M. (1993). The strongly conserved carboxyl-terminus glycine-methionine motif of the Escherichia coli GroEL chaperonin is dispensable. Mol. Microbiol. , 7 , 49–58. Mhammedi-Alaoui, A., Pato, M., Gama, M.J. and Toussaint, A. (1994). A new component of bacteriophage Mu replicative transposition machinery: the Escherichia coli ClpX protein. Mol Microbiol. , 11 , 1109–1116. Miczak, A., Kaberdin, V.R., Wei, C.L. and Lin-Chao, S. (1996). Proteins associated with RNase E in a multicomponent ribonucleolytic complex. Proc. Natl. Acad. Sci. USA , 93 , 3865–3869. Miki, T., Orita, T., Furuno, M. and Horiuchi, T. (1988). Control of cell division by sex factor F in Escherichia coli. III. Participation of the groES (mopB) gene of the host bacteria. J. Mol. Biol. , 201 , 327–338. Miki, T., Park, J.A., Nagao, K., Murayama, N. and Horiuchi, T. (1992). Control of segregation of chromosomal DNA by sex factor F in Escherichia coli . Mutants of DNA gyrase subunit A suppress letD (ccdB) product growth inhibition. J. Mol Biol. , 225 , 39–52. Missiakas, D. and Raina, S. (1997). Protein misfolding in the cell envelope of Escherichia
Molecular chaperones and folding catalysts
148
coli: new signaling pathways. Trends Biochem. Sci. , 22 , 59–63. Mogk, A., Homuth, G., Scholz, C., Kim, L., Schmid, F.X. and Schumann, W. (1997). The GroE chaperonin machine is a major modulator of the CIRCE heat shock regulon of Bacillus subtilis. EMBO J. , 16 , 4579–4590. Muffler, A., Barth, M., Marschall, C. and Hengge-Aronis, R. (1997). Heat shock regulation of turnover: a role for DnaK and relationship between stress responses mediated by and in Escherichia coli. J. Bacteriol. , 179 , 445–452. Mukhopadhyay, G., Sozhamannan, S. and Chattoraj, D.K. (1994). Relaxation of replication control in chaperone-independent initiator mutants of plasmid P1. EMBO J. , 13 , 2089–2096. Murayama, N., Shimizu, H., Takiguchi, S., Baba, Y., Amino, H., Horiuchi, T., Sekimizu, K. and Miki, T. (1996). Evidence for involvement of Escherichia coli genes pmbA, csrA and a previously unrecognized gene tldD, in the control of DNA gyrase by letD (ccdB). of sex factor F. J. Mol Biol. , 256 , 483–502. Murphy, C.K. and Beckwith, J. (1996). Export of proteins to the cell envelope in Escherichia coli. In F.C. Neidhardt, (ed.), Escherichia coli and Salmonella: Cellular and Molecular Biology , 2nd ed., Vol. 1, ASM Press, Washington, DC, pp. 967–978. Mushegian, A.R. and Koonin, E.V. (1996). A minimal gene set for cellular life derived by comparison of complete bacterial genomes. Proc. Natl. Acad. Sci. USA , 93 , 10268–10273. Nanninga, N. (1991). Cell division and peptidoglycan assembly in Escherichia coli. Mol. Microbiol. , 5 , 791–795. Nelson, R.J., Ziegelhoffer, T., Nicolet, C., Werner-Washburne, M. and Craig, E.A. (1992). The translation machinery and 70 kd heat shock protein cooperate in protein synthesis. Cell , 71 , 97–105. Niki, H., Jaffé, A., Imamura, R., Ogura, T. and Hiraga, S. (1991). The new gene mukB codes for a 177 kd protein with coiled-coil domains involved in chromosome partitioning of E. coli . EMBO J. , 10 , 183–193. Nyström, T. (1995). The trials and tribulations of growth arrest. Trends Microbiol. , 3 , 131–136. Ogawa, J. and Long, S.R. (1995). The Rhizobium meliloti groELc locus is required for regulation of early nod genes by the transcription activator NodD. Genes Dev. , 9 , 714–729. Ohki, M. and Smith, C.L. (1989). Tracking bacterial DNA replication forks in vivo by pulsed field gel electrophoresis. Nucleic Acids Res , 17 , 3479–3490. Ohkubo, S. and Yamaguchi, K. (1997). A suppressor of mutations in the region adjacent to iterons of pSC101 ori. J. Bacteriol , 179 , 2089–2091. Paek, K.-H. and Walker, G.C. (1987). Escherichia coli dnaK null mutants are inviable at high temperature. J. Bacteriol. , 169 , 283–290. Palleros, D.R., Reid, K.L., Shi, L., Welch, W.J. and Fink, A.L. (1993). ATP-induced protein-Hsp70 complex dissociation requires K+ but not ATP hydrolysis. Nature , 365 , 664–666. Parsell, D.A. and Lindquist, S. (1994). Heat shock proteins and stress tolerance. In R.I.Morimoto, A. Tissières and C.Georgopoulos, (eds.), The Biology of Heat Shock Proteins and Molecular Chaperones , Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 457–494. Parsell, D.A., Kowal, A.S., Singer, M.A. and Lindquist, S. (1994). Protein disaggregation mediated by heat-shock protein Hsp104. Nature , 372 , 475–478. Phillips, G.J. and Silhavy, T.J. (1990). Heat-shock proteins DnaK and GroEL facilitate
Genetic evidence for the roles of molecular chaperones
149
export of LacZ hybrid proteins in E. coli Nature , 344 , 882–884. Pugsley, A.P. (1993). The complete general secretory pathway in gram-negative bacteria. Microbiol. Rev. , 57 , 50–108. Ramsay, N. (1988). A mutant in a major heat shock protein of Escherichia coli continues to show inducible thermotolerance. Mol Gen. Genet. , 211 , 332–334. Randall, L.L. and Hardy, S.J.S. (1995). High selectivity with low specificity: how SecB has solved the paradox of chaperone binding. Trends Biochem Sci , 20 , 65–69. Reuter, S.H. and Shapiro, L. (1987). Asymmetric segregation of heat-shock proteins upon cell division in Caulobacter crescentus. J. Mol. Biol. , 194 , 653–662. Rockabrand, D., Arthur, T., Korinek, G., Livers, K. and Blum, P. (1995). An essential role for the Escherichia coli DnaK protein in starvation-induced thermotolerance, H2O2 resistance, and reductive division. J. Bacteriol. , 177 , 3695–3703. Rockabrand, D., Livers, K., Austin, T., Kaiser, R., Jensen, D., Burgess, R. and Blum, P. (1998). Roles of DnaK and RpoS in starvation-induced thermotolerance of Escherichia coli. J. Bacteriol. , 180 , 846–854. Rodrigue, A., Batia, N., Müller, M., Fayet, O., Böhm, R., Mandrand-Berthelot, M.-A. and Wu, L.-F. (1996). Involvement of the GroE chaperonins in the nickel-dependent anaerobic biosynthesis of NiFe-hydrogenases of Escherichia coli. J. Bacteriol. , 178 , 4453–4460. Ron, E.Z., Alajem, S., Biran, D. and Grossman, N. (1990). Adaptation of Escherichia coli to elevated temperatures: the metA gene product is a heat shock protein. Antonie van Leeuwenhoek , 58 , 169–174. Roof, W.D., Fang, H.Q., Young, K.D., Sun, J. and Young, R. (1997). Mutational analysis of slyD, an Escherichia coli gene encoding a protein of the FKBP immunophilin family. Mol. Microbiol. , 25 , 1031–1046. Rouvière, P.E., De Las Penas, A., Mecsas, J., Lu, C.Z., Rudd, K.E. and Gross, C.A. (1995). rpoE, the gene encoding the second heat-shock sigma factor, , in Escherichia coli. EMBO J. , 14 , 1032–1042. Ruben, S.M., VanDenBrink-Webb, S.E., Rein, D.C. and Meyer, R.R. (1988). Suppression of the Escherichia coli ssb-1 mutation by an allele of groEL. Proc. Natl. Acad. Sci. USA , 85 , 3767–3771. Rusanganwa, E. and Gupta, R.S. (1993). Cloning and characterization of multiple groEL chaperonin-encoding genes in Rhizobium meliloti. Gene , 126 , 67–75. Saito, H. and Uchida, H. (1977). Initiation of the DNA replication of bacteriophage lambda in Escherichia coli K12. J. Mol. Biol. , 113 , 1–25. Saito, H. and Uchida, H. (1978). Organization and expression of the dnaJ and dnaK genes of Escherichia coli K12. Mol Gen. Genet. , 164 , 1–8. Sakakibara, Y. (1988). The dnaK gene of Escherichia coli functions in initiation of chromosome replication. J. Bacteriol. , 170 , 972–979. Sanchez, Y. and Lindquist, S.L. (1990). HSP104 required for induced thermotolerance. Science , 248 , 1112–1115. Sanchez, Y., Parsell, D.A., Taulien, J., Vogel, J.L., Craig, E.A. and Lindquist, S. (1993). Genetic evidence for a functional relationship between Hsp104 and Hsp70. J. Bacteriol , 175 , 6484–6491. Sand, O., Desmet, L., Toussaint, A. and Pato, M. (1995). The Escherichia coli DnaK chaperone machine and bacteriophage Mu late transcription. Mol. Microbiol. , 15 , 977–984. Schatz, P.J. and Beckwith, J. (1990). Genetic analysis of protein export in E. coli. Annu. Rev. Genet. , 24 , 215–248.
Molecular chaperones and folding catalysts
150
Schirmer, E.G., Glover, J.R., Singer, M.A. and Lindquist, S. (1996). HSP100/Clp proteins: a common mechanism explains diverse functions. Trends Biochem. Sci. , 21 , 289–296. Schlenstedt, G., Harris, S., Risse, B., Lill, R. and Silver, P.A. (1995). A yeast DnaJ homologue, Scj1p, can function in the endoplasmic reticulum with BiP/Kar2p via a conserved domain that specifies interactions with Hsp70s. J. Cell Biol. , 129 , 979– 988. Schmid, F.X. (1995). Prolyl isomerases join the fold. Curr Biol. , 5 , 993–994. Schmitt, M., Neupert, W. and Langer, T. (1995). Hsp78, a Clp homologue within mitochondria, can substitute for, chaperone functions of mt-hsp70. EMBO J. , 14 , 3434–3444. Schröder, H., Langer, T., Hartl, F.-U. and Bukau, B. (1993). DnaK, DnaJ and GrpE form a cellular chaperone machinery capable of repairing heat-induced protein damage. EMBO J. , 12 , 4137–4144. Seaton, B.L. and Vickery, L.E. (1994). A gene encoding a DnaK/hsp70 homolog in Escherichia coli. Proc. Natl. Acad. Sci. USA , 91 , 2066–2070. Sell, S.M., Eisen, C., Ang, D., Zylicz, M. and Georgopoulos, C. (1990). Isolation and characterization of dnaJ null mutants of Escherichia coli. J. Bacteriol. , 172 , 4827– 4835. Sherman, M.Y. and Goldberg, A.L. (1992). Involvement of the chaperonin dnaK in the rapid degradation of a mutant protein in Escherichia coli. EMBO J. , 11 , 71–77. Sherman, M.Y. and Goldberg, A.L. (1996). Involvement of molecular chaperones in intracellular protein breakdown. EXS , 77 , 57–78. Shi, W., Zhou, Y., Wild, J., Adler, J. and Gross, C.A. (1992). DnaK, DnaJ, and GrpE are required for flagellum synthesis in Escherichia coli. J. Bacteriol. , 174 , 6256–6263. Shirai, Y., Akiyama, Y. and Ito, K. (1996). Suppression of ftsH mutant phenotypes by overproduction of molecular chaperones. J. Bacteriol. , 178 , 1141–1145. Smith, D.R., Doucette-Stamm, L.A., Deloughery, C., Lee, H., Dubois, J., Aldredge, T., Bashirzadeh, R., Blakely, D., Cook, R., Gilbert, K. et al. (1997). Complete genome sequence of Methanobacterium thermoautotrophicum H: functional analysis and comparative genomics. J. Bacterial. , 179 , 7135–7155. Sohlberg, B., Lundberg, U., Hartl, E-U. and von Gabain, A. (1993). Functional interaction of heat shock protein GroEL with an RNase E-like activity in Escherichia coli. Proc. Natl. Acad. Sci. USA , 90 , 277–281. Spence, J., Cegielska, A. and Georgopoulos, C. (1990). Role of Escherichia coli heat shock proteins DnaK and HtpG (C62.5) in response to nutritional deprivation. J. Bacterial. , 172 , 7157–7166. Squires, C.L., Pedersen, S., Ross, B.M. and Squires, C. (1991). ClpB is the Escherichia coli heat shock protein F84.1. J. Bacterial. , 173 , 4254–4262. Stoller, G., Rücknagel, K.P., Nierhaus, K.H., Schmid, F.X., Fischer, G. and Rahfeld, J.U. (1995). A ribosome-associated peptidyl-prolyl cis/trans isomerase identified as the trigger factor. EMBO J. , 14 , 4939–4948. Straus, D., Walter, W. and Gross, C.A. (1990). DnaK, DnaJ, and GrpE heat shock proteins negatively regulate heat shock gene expression by controlling the synthesis and stability of . Genes Dev. , 4 , 2202–2209. Straus, D.B., Walter, W.A. and Gross, C.A. (1987). The heat shock response of E. coli is regulated by changes in the concentration of . Nature , 329 , 348–351. Straus, D.B., Walter, W.A. and Gross, C.A. (1988). Escherichia coli heat shock gene mutants are defective in proteolysis. Genes Dev. , 2 , 1851–1858.
Genetic evidence for the roles of molecular chaperones
151
Sunshine, M., Feiss, M., Stuart, J. and Yochem, J. (1977). A new host gene (groPC) necessary for DNA replication. Mol. Gen. Genet. , 151 , 27–34. Szabo, A., Korszun, R., Hartl, F.U. and Flanagan, J. (1996). A zinc finger-like domain of the molecular chaperone DnaJ is involved in binding to denatured protein substrates. EMBO J. , 15 , 408–417. Szabo, A., Langer, T., Schröder, H., Flanagan, J., Bukau, B. and Hartl, F.U. (1994). The ATP hydrolysis-dependent reaction cycle of the Escherichia coli Hsp70 system— DnaK, DnaJ, and GrpE. Proc. Natl. Acad. Sci. USA , 91 , 10345–10349. Tanaka, M., Wada, M. and Itikawa, H. (1985). Abortive relaxation of RNA synthesis for cellular DNA replication affected by dnaK(Ts) and dnaJ(Ts) mutations in Escherichia coli K-12. Jpn.J. Genet. , 60 , 27–38. Taniguchi, H., Tokida, T., Fujita, H. and Itikawa, H. (1989). High levels of manganesecontaining superoxide dismutase and thermally induced DNA disruption in a dnaK7 (Ts) mutant of Escherichia coli K12. Mol. Gen. Genet. , 217 , 317–323. Teschke, C.M., Kim, J., Song, T., Park, S., Park, C. and Randall, L.L. (1991). Mutations that affect the folding of ribose-binding protein selected as suppressors of a defect in export in Escherichia coli. J. Biol. Chem. , 266 , 11789–11796. Thomas, J.G., Ayling, A. and Baneyx, F. (1997). Molecular chaperones, folding catalysts, and the recovery of active recombinant proteins from E. coli: to fold or to refold. Appl. Biochem. Biotechnol. , 66 , 197–238. Tilly, K., McKittrick, N., Zylicz, M. and Georgopoulos, C. (1983). The dnaK protein modulates the heat-shock response of Escherichia coli. Cell , 34 , 641–646. Tilly, K., Sozhamannan, S. and Yarmolinsky, M. (1990). Participation of the Escherichia coli heat shock proteins DnaJ, DnaK, and GrpE in autorepression of the P1 plasmid repA promoter. New Biol. , 2 , 812–817. Tilly, K., Spence, J. and Georgopoulos, C. (1989). Modulation of stability of the Escherichia coli heat shock regulatory factor . J. Bacteriol. , 171 , 1585–1589. Tilly, K. and Yarmolinsky, M. (1989). Participation of Escherichia coli heat shock proteins DnaJ, DnaK, and GrpE in P1 plasmid replication. J. Bacteriol. , 171, 6025– 6029. Tomoyasu, T., Gamer, J., Bukau, B., Kanemori, M., Mori, H., Rutman, A.J., Oppenheim, A.B., Yura, T., Yamanaka, K., Niki, H. et al., (1995). Escherichia coli FtsH is a membrane-bound, ATP-dependent protease which degrades the heat-shock transcription factor . EMBO J. , 14 , 2551–2560. Tsuchido, T., VanBogelen, R.A. and Neidhardt, F.C. (1986). Heat shock response in Escherichia coli influences cell division. Proc. Natl. Acad. Sci. USA , 83 , 6959–6963. Ueguchi, C. and Ito, K. (1992). Multicopy suppression: an approach to understanding intracellular functioning of the protein export system. J. Bacteriol , 174 , 1454–1461. Ueguchi, C., Kakeda, M., Yamada, H. and Mizuno, T. (1994). An analogue of the DnaJ molecular chaperone in Escherichia coli . Proc. Natl Acad. Sci. USA , 91 , 1054–1058. Ueguchi, C., Shiozawa, T., Kakeda, M., Yamada, H. and Mizuno, T. (1995). A study of the double mutation of dnaJ and cbpA, whose gene products function as molecular chaperones in Escherichia coli . J. Bacteriol , 177 , 3894–3896. Valent, Q.A., Kendall, D.A., High, S., Kusters, R., Oudega, B. and Luirink, J. (1995). Early events in preprotein recognition in E. coli: interaction of SRP and trigger factor with nascent polypeptides. EMBO J. , 14 , 5494–5505. van der Vies, S.M., Gatenby, A.A. and Georgopoulos, C. (1994). Bacteriophage T4 encodes a cochaperonin that can substitute for Escherichia coli GroES in protein folding. Nature , 368 , 654–656.
Molecular chaperones and folding catalysts
152
Van Melderen, L., Bernard, P. and Couturier, M. (1994). Lon-dependent proteolysis of CcdA is the key control for activation of CcdB in plasmid-free segregant bacteria. Mol Microbiol. , 11 , 1151–1157. VanBogelen, R.A., Acton, M.A. and Neidhardt, F.C. (1987a). Induction of the heat shock regulon does not produce thermotolerance in Escherichia coli . Genes Dev. , 1 , 525– 531. VanBogelen, R.A., Kelley, P.M. and Neidhardt, F.C. (1987b). Differential induction of heat shock, SOS, and oxidation stress regulons and accumulation of nucleotides in Escherichia coli. J. Bacteriol , 169 , 26–32. Vickery, L.E., Silberg, J.J. and Ta, D.T. (1997). Hsc66 and Hsc20, a new heat shock cognate molecular chaperone system from Escherichia coli. Protein Sci. , 6 , 1047– 1056. Viitanen, P.V., Gatenby, A.A. and Lorimer, G.H. (1992). Purified chaperonin 60 (groEL) interacts with the nonnative states of a multitude of Escherichia coli proteins. Protein Sci. , 1 , 363–369. Wada, M., Kadokami, Y. and Itikawa, H. (1982). Thermosensitive synthesis of DNA and RNA in dnaJ mutants of Escherichia coli K-12. Jpn. J. Genet. , 57 , 407–413. Wada, M. and Itikawa, H. (1984). Participation of Escherichia coliK-12 groE gene products in the synthesis of cellular DNA and RNA. J. Bacteriol. , 157 , 694–696. Wada, M., Sekine, K. and Itikawa, H. (1986). Participation of the dnaK and dnaJ gene products in phosphorylation of glutaminyl-tRNA synthetase and threonyl-tRNA synthetase of Escherichia coli K-12. J. Bacterial , 168 , 213–220. Wall, D., Zylicz, M. and Georgopoulos, C. (1994). The NH2-terminal 108 amino acids of the Escherichia coli DnaJ protein stimulate the ATPase activity of DnaK and are sufficient for replication. J. Biol Chem. , 269 , 5446–5451. Wawrzynow, A., Banecki, B., Wall, D., Liberek, K., Georgopoulos, C. and Zylicz, M. (1995a). ATP hydrolysis is required for the DnaJ-dependent activation of DnaK chaperone for binding to both native and denatured protein substrates. J. Biol Chem. , 270 , 19307–19311. Wawrzynow, A., Wojtkowiak, D., Marszalek, J., Banecki, B., Jonsen, M., Graves, B., Georgopoulos, C. and Zylicz, M. (1995b). The ClpX heat-shock protein of Escherichia coli, the ATP-dependent substrate specificity component of the ClpP-ClpX protease, is a novel molecular chaperone. EMBO J. , 14 ,1867–1877. Wawrzynow, A., Banecki, B. and Zylicz, M. (1996). The Clp ATPases define a novel class of molecular chaperones. Mol Microbiol , 21 , 895–899. Wegrzyn, A., Taylor, K. and Wegrzyn, G. (1996). The cbpA chaperone gene function compensates for dnaJ in plasmid replication during amino acid starvation of Escherichia coli. J. Bacteriol. , 178 , 5847–5849. Wickner, S., Gottesman, S., Skowyra, D., Hoskins, J., McKenney, K, and Maurizi, M.R. (1994). A molecular chaperone, ClpA, functions like DnaK and DnaJ. Proc. Natl Acad. Sci. USA , 91 , 12218–12222. Wickner, S., Hoskins, J. and McKenney, K. (1991a). Function of DnaJ and DnaK as chaperones in originspecific DNA binding by RepA. Nature , 350 , 165–167. Wickner, S., Hoskins, J. and McKenney, K. (1991b). Monomerization of RepA dimers by heat shock proteins activates binding to DNA replication origin. Proc. Natl Acad. Sci. USA , 88 , 7903–7907. Wickner, S., Skowyra, D., Hoskins, J. and McKenney, K. (1992). DnaJ, DnaK, and GrpE heat shock proteins are required in onP1 DNA replication solely at the RepA
Genetic evidence for the roles of molecular chaperones
153
monomerization step. Proc. Natl. Acad. Sci. USA , 89 , 10345–10349. Wild, J., Altman, E., Yura, T. and Gross, C.A. (1992a). DnaK and DnaJ heat shock proteins participate in protein export in Escherichia coli. Genes Dev. , 6 , 1165–1172. Wild, J., Kamath-Loeb, A., Ziegelhoffer, E., Lonetto, M., Kawasaki, Y. and Gross, C.A. (1992b). Partial loss of function mutations in DnaK, the Escherichia coli homologue of the 70-kDa heat shock proteins, affect highly conserved amino acids implicated in ATP binding and hydrolysis. Proc. Natl. Acad. Sci. USA , 89 , 7139–7143. Wild, J., Rossmeissl, P., Walter, W.A. and Gross, C.A. (1996). Involvement of the DnaK-DnaJ-GrpE chaperone team in protein secretion in Escherichia coli. J Bacterial , 178 , 3608–3613. Wild, J., Walter, W.A., Gross, C.A. and Altman, E. (1993). Accumulation of secretory protein precursors in Escherichia coli induces the heat shock response. J. Bacteriol. , 175 , 3992–3997. Wu, B., Georgopoulos, C. and Ang, D. (1992). The essential Escherichia coli msgB gene, a multicopy suppressor of a temperature-sensitive allele of the heat shock gene grpE, is identical to dapE. J. Bacteriol. , 174 , 5258–5264. Yamamori, T. and Yura, T. (1982). Genetic control of heat-shock protein synthesis and its bearing on growth and thermal resistance in Escherichia coli K-12. Proc. Natl. Acad. Sci. USA , 79 , 860–864. Yamanaka, K., Mitani, T., Ogura, T., Niki, H. and Hiraga, S. (1994). Cloning, sequencing, and characterization of multicopy suppressors of a mukB mutation in Escherichia coli. Mol. Microbiol. , 13 , 301–312. Yamashino, T., Kakeda, M., Ueguchi, C. and Mizuno, T. (1994). An analogue of the DnaJ molecular chaperone whose expression is controlled by during the stationary phase and phosphate starvation in Escherichia coli. Mol. Microbiol. , 13 , 475–483. Yang, Z., Geng, Y. and Shi, W. (1998). A DnaK homolog in Myxococcus xanthus is involved in social motility and fruiting body formation. J. Bacteriol. , 180 , 218–224. Ybarra, J. and Horowitz, P. (1996). Nucleotides reveal polynucleotide phosphorylase activity from conventionally purified GroEL. J. Biol. Chem. , 271 , 25063–25066. Yuan, Y., Crane, D.D. and Barry, C.E.III. (1996). Stationary phase-associated protein expression in Mycobacterium tuberculosis: function of the mycobacterial -crystallin homolog. J. Bacteriol. , 178 , 4484–4492. Yura, T., Nakahigashi, K. and Kanemori, M. (1996). Transcriptional regulation of stressinducible genes in procaryotes. EXS , 77 , 165–181. Zahn, R. and Plückthun, A. (1992). GroE prevents the accumulation of early folding intermediates of pre- -lactamase without changing the folding pathway. Biochemistry , 31 , 3249–3255. Zeilstra-Ryalls, J., Fayet, O., Baird, L. and Georgopoulos, C. (1993). Sequence analysis and phenotypic characterization of groEL mutations that block and T4 bacteriophage growth. J. Bacteriol. , 175 , 1134–1143. Zeilstra-Ryalls, J., Fayet, O. and Georgopoulos, C. (1991). The universally conserved GroE (Hsp60). chaperonins. Annu. Rev. Microbiol. , 45 , 301–325. Zeilstra-Ryalls, J., Fayet, O. and Georgopoulos, C. (1994). Two classes of extragenic suppressor mutations identify functionally distinct regions of the GroEL chaperone of Escherichia coli. J. Bacteriol. , 176 , 6558–6565. Zeilstra-Ryalls, J., Fayet, O. and Georgopoulos, C. (1996). In vivo protein folding: suppressor analysis of mutations in the groES cochaperone gene of Escherichia coli . FASEB J. , 10 , 148–152. Zhou, Y.N., Kusukawa, N., EricksonJ.W., Gross, C.A. and Yura, T. (1988). Isolation and
Molecular chaperones and folding catalysts
154
characterization of Escherichia coli mutants that lack the heat shock sigma factor . J. Bacteriol. , 170 , 3640–3649. Zhu, X., Zhao, X., Burkholder, W.F., Gragerov, A., Ogata, C.M., Gottesman, M.E., and Hendrickson, W.A. (1996). Structural analysis of substrate binding by the molecular chaperone DnaK. Science , 272 , 1606–1614. Zylicz, M., Ang, D. and Georgopoulos, C. (1987). The grpE protein of Escherichia coli. purification and properties. J. Biol. Chem. , 262 , 17437–17442. Zylicz, M., Yamamoto, T., McKittrick, N., Sell, S. and Georgopoulos, C. (1985). Purification and properties of the dnaJ replication protein of Escherichia coli. J. Biol. Chem. , 260 , 7591–7598.
7. GENETIC DISSECTION OF THE Hsp70 CHAPERONE SYSTEM OF YEAST ELIZABETH CRAIG * , WEI YAN and PHILIP JAMES Department of Biomolecular Chemistry, University of Wisconsin, Madison, WI, USA 53706
1. Introduction 2. The Hsp70 Class 2.1 Hsp70s of the Cytosol/Nucleus 2.1.1. The Ssa Hsp70s 2.1.2. The Ssb Hsp70s 2.1.3. Basis of the Functional Differences Between Ssa and Ssb Hsp70s 2.1.4. The Sse Class of Hsp70s 2.1.5. Pdr13 2.2. Hsp70s of the Mitochondria 2.3. Hsp70s of the ER 3. The DnaJ Multigene Family of Yeast 3.1. DnaJs of the Cytosol/Nucleus: Ydj1, Sis1 and Zuo1 3.1.1. Ydj1 3.1.2. Sis1 3.1.3. Zuo1 3.2. DnaJ of the Mitochondria 3.3. DnaJs of the ER: Sec63 and Scj1 3.4. Other DnaJs: Xdj1 and Caj1 4. MGE1, a GrpE-related Gene of the Mitochondrial Matrix 5. Functional Interactions Among Classes of Chaperones in Protein Folding 6. Conclusions 7. References *Corresponding
author
Molecular chaperones and folding catalysts
156
1. INTRODUCTION It has been known for more than a decade that eucaryotes have multiple families of genes encoding molecular chaperones, particularly those of the Hsp70 and DnaJ classes. The completion of the sequence of the S. cerevisiae genome allows a unique opportunity to consider the entire array of molecular chaperones within a eucaryotic organism. Although some of the predicted proteins are yet to be analyzed, a great deal of useful information has been obtained for many members. Different, but related proteins are found in various cellular compartments: the cytosol/nucleus, ER and mitochondria. Each of these three cellular compartments has more than one type of Hsp70, while the ER and the cytosol have more than one DnaJ. The presence of multiple chaperones of both the Hsp70 and DnaJ classes raises the question of whether related proteins in the same compartment are functionally different. In addition, it will be important to dissect the interrelationship between the Hsp70s and proteins with which they interact, such as DnaJs, in order to understand the roles of chaperones in cellular processes. Below we outline the families of molecular chaperones and their functions within the cell as we understand them today, focusing on those of the Hsp70 class and the proteins that function with them.
2. THE Hsp70 CLASS Analysis of the complete genome sequence of S. cerevisiae revealed a total of 14 open reading frames predicted to encode proteins related to Hsp70s. Ten of these were identified and reported prior to the sequencing of the yeast genome. The sequence relationship among these 14 proteins is shown in Figure 1. The proteins in the top part of the tree, Ssa1 through Ssh1, share a high level of sequence identity with one another and with identified Hsp70s from other organisms. For example, the Ssa subfamily is most closely related to Hsp70s of the cytosol of other eucaryotes. Kar2 is most closely related to BiP of the mammalian ER, while Ssc1, Ssh1, and Ssj1 are most closely related to other mitochondrial Hsp70s and bacterial Hsp70s such as DnaK. The others, Pdr13, Ssi1, and Sse1, 2, are less related and may ultimately form individual Hsp70-related subgroups. The Sse1, 2 and Ssi1 proteins are considerably larger than members of the “core” Hsp70 group and are most obviously related to Hsp70s in the ATPase domain. Further functional information is required to determine whether these proteins have the characteristic ATPase and peptide binding activities of Hsp70s. Figure 1 also indicates that the cytosol, ER and mitochondria each contain more than one family of Hsp70. Our current state of knowledge is sufficient to establish that at least the Ssa and Ssb proteins are present in the cytosol, Kar2 and Ssi1 in the lumen of the ER and Ssc1, Ssh1 and Ssj1 in the matrix of the mitochondria. Below, the individual classes of Hsp70s are discussed, emphasizing the roles they play in cellular processes and the similarities and differences among them.
Genetic dissection of the Hsp70 chaperone system of yeast
157
2.1. Hsp70s of the Cytosol/Nucleus The yeast cytosol contains at least two Hsp70 subfamilies, the Ssas and Ssbs. Analysis of Hsp70s from a diverse set of organisms revealed that the SSA proteins are more closely related to the cytosolic Hsp70s from other organisms than to the SSB proteins (Boorstein et al., 1994). Therefore, the SSB proteins represent an ancient divergence from other known cytosolic Hsp70s. As discussed in more detail below, the SSA and SSB proteins are functionally distinct. The Ssa proteins function in the regulation of the heat shock response, the translocation of some proteins into organelles, and the folding of proteins in the cytosol. The Ssb proteins associate
Figure 1 Phylogenetic tree of the Saccharomyces cerevisiae Hsp70 proteins. Subcellular location and phenotypes of deletion mutants are indicated for each Hsp70 family. The tree was constructed using MegAlign software from DNASTAR (Madison, WI) and the clustal method. Amino acid sequences were truncated at the N-terminus to remove the leader sequences of Hsp70s localized in mitochondria or ER. The first amino acids included in the alignment are: 4 (Ssa1–4, Sse1–2), 9 (Ssb1–2), 32 (Ssc1), 52 (Kar2), 30 (Ssh1), 22 (Ssi1), 29 (Ssj1), 39 (Pdr13).
with translating ribosomes and ssb1 ssb2 mutants are sensitive to some translationinhibiting drugs. In the next three sections, we describe the Ssa and Ssb families independently, and then discuss experiments which delineate the functional differences between them. In addition, the Sse1, 2 and Pdr13 Hsp70s, which are
Molecular chaperones and folding catalysts
158
discussed briefly below, likely also reside in the cytosolic compartment since they contain no obvious organellar targeting sequence. 2.1.1. The Ssa Hsp70s The Ssa subfamily is essential and is composed of four members that share between 84 and 97% identity (Werner-Washburne et al., 1987). Under optimal growth conditions, only SSA1 and SSA2 are expressed, with the more distantly-related SSA3 and SSA4 being induced in response to a temperature upshift or other stress. While expression of SSA2 remains unchanged during stress conditions, expression of SSA1 is increased. Although strains carrying mutations in different combinations of SSA genes have different phenotypes, this is likely due to differences in expression levels; a combination of genetic experiments have provided no evidence of functional differences among members of the SSA Hsp70 family (Craig et al., 1995). That is, cells with the same level of any of the four individual Ssa proteins or a combination of Ssa proteins appear phenotypically the same, at least under conditions tested thus far. The complexity and essentiality of the SSA family of genes illustrates the central role of this class of Hsp70s in cellular function. Analysis of strains containing mutations of SSA genes have been used to begin to dissect the multiple roles played by the Ssa proteins. Strains lacking both Ssa1 and Ssa2, the two constitutively expressed members of this family, have been particularly useful, ssa1 ssa2 strains grow more slowly than wild-type strains at temperatures up to 35°C and are unable to form colonies at 37°C (Craig and Jacobsen, 1984). The viability of cells lacking Ssa1 and Ssa2, along with their constitutive thermotolerance, is indicative of the role of Ssa proteins in regulating the heat shock response (see chapter by Morimoto). Like cytosolic Hsp70s of other eucaryotes, Ssa proteins negatively regulate the activity of Hsf, the heat shock transcription factor. In ssa1 ssa2 strains the expression of a variety of heat shock proteins is induced even at low temperatures (Craig and Jacobsen, 1984; Boorstein and Craig, 1990a, b). Among the genes whose expression is enhanced are SSA3 and SSA4; it is this expression that allows for viability of ssa1 ssa2 cells. However the expression of Ssa3 and Ssa4 does not reach the levels seen for Ssa1 and Ssa2 under normal conditions, explaining the failed regulation of heat-inducible genes in these strains. If overexpressed, Ssa3 or Ssa4 can restore the expression of heat shock proteins to normal levels. Consistent with the role of SSA proteins regulating Hsf activity, when Ssa1 is overexpressed the induction of the SSA3 and SSA4 genes in response to a heat shock is greatly reduced (Stone and Craig, 1990). Extragenic suppressors of the slow growth defect of an ssa1 ssa2 strain have been isolated (Nelson et al., 1992). Analysis of these suppressors has provided additional insights into the importance of the Ssa proteins in regulating the expression of heat shock-inducible proteins. One such suppressor, named EXA3, rescued the slow growth phenotype of ssa1 ssa2 strains at temperatures such as 23° and 30°C. EXA3 was subsequently shown to be an allele of HSF1, which encodes the Hsf transcription factor (Halladay and Craig, 1995). The initial expectation was that a mutation in HSF1 able to suppress the growth defect of ssa1 ssa2 cells would cause increased Hsf activity and thus increased Ssa3 and Ssa4 expression. However, the induction of Hsf-driven transcription
Genetic dissection of the Hsp70 chaperone system of yeast
159
in response to a heat shock is significantly delayed in an EXA3–1 strain compared to wild-type. In addition, the basal level of Hsf-driven expression at all temperatures was several fold lower in the mutant than in the wild-type strain. This decreased expression suggests that the EXA3–1 allele encodes an Hsf with reduced activity. Consistent with this idea, the EXA3–1 allele of Hsf contains a single point mutation which causes the change of a proline to a glutamine in the DNA binding domain of Hsf (Halladay and Craig, 1995). As a result, the DNA binding activity of the mutant Hsf was reduced about 50 fold in in vitro gel shift assays. The affected proline is at position 214, which is absolutely conserved in Hsf homologs from all organisms examined and produces a prominent kink (Harrison et al., 1994) near the end of the first helix of the helix-turnhelix DNA binding motif. How can these phenotypes be explained? Under normal conditions, the Ssa1 and Ssa2 proteins serve as negative regulators of Hsf, repressing the transcription of heat shock proteins. In an ssa1 ssa2 strain, Hsf activity is increased and a wide variety of heat shock proteins are constitutively expressed. The constitutive expression of the normally heatinducible Ssa3 and Ssa4 proteins allows the cells to survive in the absence of Ssa1 and Ssa2, but at the same time, the constitutive expression of other heat-inducible genes appears to be detrimental. Consistent with this idea is the observation that an additional copy of Hsf is deleterious to the growth of ssa1 ssa2 cells (Halladay and Craig, 1995). The reduced activity of the EXA3–1 allele of Hsf apparently aids the growth of ssa1 ssa2 cells by decreasing the expression of these detrimental proteins. Thus in an ssa1 ssa2 strain a delicate balance must be met to allow sufficient expression of Ssa proteins to perform functions other than heat shock regulation, while limiting the detrimental expression of other Hsps. Evidence has also been found for a similar situation in prokaryotes (Bukau and Walker, 1990). In E. coli, the Hsp70 DnaK plays an important role in regulating the heat shock response by modulating the level of the subunit of RNA polymerase (see Chapter by Connolly et al.). This subunit is required for the initiation of transcription of heat-inducible genes (Georgopoulos et al., 1994). Bukau and Walker (Bukau and Walker, 1990) selected suppressors of the slow growth phenotype of a dnaK deletion strain and identified mutants of . Interestingly, these mutants had either decreased activity or decreased stability of , leading to the hypothesis that increased expression of Hsps was in part responsible for the growth defects of dnaK mutants. Though Ssa proteins are critical for regulating the heat shock response, they carry out other functions as well. The initial indication that Ssa proteins are involved in the translocation of proteins into the ER and mitochondria came from studies of an ssa1 ssa2 ssa4 mutant which contained an SSA 1 gene under the control of the regulatable promoter GAL1 (Deshaies et al., 1988). Shift from galactose- to glucose-based media caused the repression of SSA1 transcription. By three hours after the shift to glucose the amount of Ssa1 present was significantly reduced and the precursor forms of several organellar proteins had accumulated in the cytosol. However the requirement for SSA function in protein translocation in vivo is certainly not universal. Of three mitochondrial proteins tested, the processing of only one, the subunit of the F1F0 ATPase, was dramatically affected. Of six proteins destined for the ER, the translocation of only prepro- -factor and proteinase A was inhibited.
Molecular chaperones and folding catalysts
160
The time delay required for Ssa1 depletion in these experiments left open the possibility that the translocation defect could be an indirect effect of the loss of Ssa function. However, similar translocation defects were recently observed using a strain carrying a temperature-sensitive SSA1 allele, ssa1–45 ssa2 ssa3 ssa4 (Becker et al., 1996); in this case the inactivation of Ssa protein should be very rapid. The defect in prepro- -factor translocation appears within 2 minutes after shift to the nonpermissive temperature in the ssa1–45 strain, suggesting a direct requirement for Ssa in the process. Interestingly, in this mutant the block in translocation was not complete. More than 50% of the protein accumulated in the precursor form, with the remainder rapidly reaching the mature form. The accumulated precursor could not be chased through the secretory pathway however. This apparent dichotomy—either complete blockage or normal translocation—could be explained in two ways. It is likely that the translocation process utilizes the chaperone activity of Hsp70s, and that some protein precursors require an association with Hsp70 to prevent aggregation, which would block subsequent translocation. In one scenario, the ssa1–45 mutation could be leaky, providing sufficient activity for translocation of a portion of the affected precursors, but leaving the rest to aggregate. Consistent with this idea, Ssa has been found associated with prepro- -factor in in vitro translation extracts (Chirico, 1992). Alternatively, the affected proteins might enter the ER by two different pathways, one that is dependent on Ssa proteins, and thus is completely blocked in the mutant, and one that is Ssa-independent. If this is true, proteins that fail to complete the Ssa-dependent pathway must not be able to return to the Ssaindependent one, perhaps due to aggregation. Recently, evidence has been reported for two translocation pathways, one SRP-independent and one SRP-dependent (Ng et al., 1996). In addition to the Ssas, the DnaJ homolog Ydj1 is also involved in protein translocation, as a YDJ1 temperature sensitive mutant accumulates the precursor forms of certain proteins (Caplan et al., 1992). A comparison of the proteins whose translocation either was or was not affected in ssa1–45 and ydj1 mutant strains shows a striking similarity (Becker et al., 1996). This similarity suggests that Ydj1 and Ssa1 function together in the translocation process. Consistent with this idea, the ssa1–45 ssa2 ssa3 ssa4 mutations are synthetically lethal with a ydj1 null mutation. In addition, an ssa2 ssa3 ssa4 ydj1 strain contains a functional SSA1 gene, but is defective in prepro- -factor translocation under conditions where neither a ydjl nor an ssa2 ssa3 ssa4 strain show precursor accumulation. These results suggest a specific functional interaction between the Ssa proteins and Ydj1 in the translocation process. Significantly, ssb1 ssb2 ydj1 strains have no synthetic phenotypes. Finally, there is compelling in vitro evidence for a role of cytosolic Hsp70s in folding of denatured proteins (Levy et al., 1995; Freeman et al., 1995; Szabo et al., 1994). Ssa1 in combination with the DnaJ homologue Ydj1 facilitates the efficient refolding of denatured luciferase. This again appears to be an Ssa-specific function, since Ssb proteins do not function in this assay (Ziegelhoffer and Craig, unpublished result). Although strong in vivo evidence is yet to be obtained for a direct role in this process, Ssa proteins are likely to be intimately involved in protein folding in the cell.
Genetic dissection of the Hsp70 chaperone system of yeast
161
2.1.2. The Ssb Hsp70s The SSB family contains two proteins, which are 99.3% identical and are encoded by the SSB1 and SSB2 genes. Unlike other Hsp70s, their expression is dramatically decreased upon a temperature upshift (Craig and Jacobsen, 1985). The SSB genes are not essential; ssb1 ssb2 cells are cold-sensitive for growth. No phenotypes of single ssb1 or ssb2 mutants have been detected, suggesting that these closely related proteins are functionally identical. It has been hypothesized that SSB proteins are involved in protein translation (Nelson et al., 1992). In part, this idea is based upon the fact that ssb1 ssb2 cells are sensitive to certain drugs which inhibit protein synthesis, including verrucarin A and aminoglycosides such as hygromycin B. In addition, the bulk of Ssb protein is found to sediment with polysomes upon sucrose gradient fractionation of crude cell extracts. This interaction appears to be specific for actively translating ribosomes, as it is disrupted in the presence of puromycin, treatment which results in the release of nascent chains from ribosomes. Given the propensity of Hsp70s to bind to polypeptides which are in an extended conformation (Landry et al., 1992), it has been proposed that Ssb proteins bind to nascent chains as they emerge from the large ribosomal subunit, thereby preventing their premature folding (Nelson et al., 1992) (see chapter by Welch et al.). This hypothesis is supported by recent data demonstrating that Ssb can be crosslinked to the nascent chain on translating ribosomes (Pçund, Wiedmann and Craig, unpublished results). A genetic screen for multicopy suppressors of the cold-sensitive phenotype of ssb1 ssb2 strains resulted in the isolation of a novel homologue of the GTP-binding protein, EF1 , which catalyzes the binding of aminoacyl-transfer RNAs to the ribosome (Nelson et al., 1992). The function of this gene, called HBS1, is unknown; a strain lacking Hbs1 has no obvious phenotype. However, its similarity to EF1 is consistent with the role of Ssb proteins in translation. SSB1 itself has been isolated as a multicopy suppressor of two very different cellular defects. Increased expression of SSB1 suppresses a temperature-sensitive mutation in the Y7 subunit of the multicatalytic proteasome, which functions to degrade both ubiquitinconjugated and non-ubiquitinated substrates (Ohba, 1994). In addition, SSB1 has been identified as a multicopy suppressor of the temperature-sensitive dbf3–1 allele, which was originally characterized as having a cell cycle phenotype due to inhibition of DNA synthesis (Shea et al., 1994). DBF3 is also identical to PRP8, which encodes the U5 snRNP, an important component of the yeast splicosome. These genetic interactions are intriguing, but it is not clear whether they are indicative of specific Ssb functions in proteolysis and RNA splicing, or an indirect effect of the role of SSB proteins in translation. 2.1.3. Basis of the Functional Differences Between Ssa and Ssb Hsp 70s Because eucaryotes have such a large number of different Hsp70s, a basic question is whether these Hsp70s are functionally distinct or overlap in function. The Ssa and Ssb Hsp70s of the cytosol of S. cerevisiae have been studied most extensively. The SSA
Molecular chaperones and folding catalysts
162
proteins are quite similar to one another in function since any of the four are able to allow wild-type or nearly wild-type growth at all temperatures if present in sufficient quantities (i.e., similar to the normal amount of total SSA protein) (Craig and Jacobsen, 1985; Craig et al., 1995). Similarly, the Ssb1 and Ssb2 proteins are interchangeable. However, it is clear that the Ssas and Ssbs are functionally distinct. Not only, as described above, are the phenotypes of SSA and SSB mutants very different, but even when differences in regulation are accounted for, members of one family can not rescue mutants of the other. Thus expression of Ssb1 under the control of an SSA promoter cannot rescue the phenotypes of an SSA mutant strain; similarly, Ssa1 under the control of an SSB promoter cannot rescue an SSB mutant strain. Differences have been observed between Ssa and Ssb proteins at the biochemical level as well. Ssa protein can facilitate the uncoating of clathrin-coated vesicles (see chapter by Eisenberg and Greene), while Ssbs cannot (Gao et al., 1991). Also, as noted above, Ssas in conjunction with Ydj1 can facilitate the refolding of denatured luciferase (Levy et al., 1995), while Ssbs cannot (Ziegelhoffer and Craig, unpublished result). The basis for the functional differences between the SSA and SSB proteins has been analyzed using chimeric proteins (James et al., 1997). Chimeras were constructed that contain all combinations of the 44 kDa ATPase, the 18 kDa peptide binding and the 10 kDa variable domains characteristic of Hsp70s (see chapter by Ha et al.). While no chimeras were able to rescue the growth phenotypes of SSA mutants, several chimeras allowed growth of ssb1 ssb2 cells at 18°C and/or in the presence of the aminoglycoside hygromycin B. It has been hypothesized that the functional differences between Hsp70s are determined by differences in peptide-binding specificity. However, a chimera that contains the peptide binding domain of Ssa1 was able to rescue both the cold-sensitive and the hygromycin B-sensitive phenotypes of an ssb1 ssb2 strain. Thus peptide binding specificity is not the determinant of the functional differences between these two Hsp70 families. If differences in peptide binding specificity do not explain the functional distinction between these two classes, then what does? An analysis of additional chimeras (James et al., 1997) revealed that rescue of the cold-sensitive phenotype required only that the chimera contain the Ssb1 ATPase domain, while rescue of the hygromycin B sensitivity was more complex (see Table 1). No single domain of Ssb was sufficient for hygromycin B rescue, but also no single domain was required. Rather, a chimera must have two Ssb1 domains to confer hygromycin B resistance, but any two will suffice. In addition, the data in Table 1 demonstrate that the cold-sensitive and hygromycin B-sensitive phenotypes of ssb1 ssb2 strains are separable, since the BAA chimera rescues cold sensitivity but not hygromycin B sensitivity while the ABB chimera rescues hygromycin B sensitivity but not cold sensitivity. Therefore, the SSB proteins carry out two different functions in the cell. Since the ability of a chimera to rescue hygromycin B sensitivity correlates with its ability to associate with ribosomes (with one exception; see Table 1), at least this phenotype is likely to represent a function of Ssb proteins in the translation process. Results from the chimeras demonstrate that the peptide binding domain is dispensible for two different fuctions of the Ssb proteins, and thus is not the source of functional specificity. Those same results suggest a model in which functional specificity depends on interactions between several Hsp70 domains and other components of the cellular
Genetic dissection of the Hsp70 chaperone system of yeast
163
machinery. The specific Hsp70 domains involved can vary from one function to another. According to this model, an interaction with the ATPase domain is required for the function involved in the rescue of cold sensitivity. On the other hand, all three domains of Ssb appear to interact with the ribosomes or other components of the translational apparatus. These interactions must be additive and sufficiently strong that the loss of any one domain by the
Table 1 Rescue of ssb1 ssb2 mutant phenotypes by Ssa/Ssb chimeric proteins
ssb1 ssb2 Phenotype a 44kDa ATPase Domain
18kDa Peptide Binding Domain
10 kDa Variable Domain
Cold Sensitivity
Hygromycin B Sensitivity
Polysome Association
B
B
B
+
+
+
B
A
A
+
–
–
A
B
A
–
–
–
A
A
B
–
–
–
A
B
B
–
+
+
B
A
B
+
+
+
B
B
A
+
+
–
A
A
A
–
–
–
a. Growth at 18°C or in the presence of 70 µg/ml hygromycin B.
substitution of Ssa sequences does not disrupt the functional interaction. In this model the peptide binding activity of Hsp70 is considered to be a “generic” activity, much like the ATPase activity, which is ferried to particular locations by specific interactions involving other parts of the protein. Such interactions may serve to target an Hsp70 to, for example, the ribosome in the case of Ssb, or to associate with cohort proteins such as DnaJs. 2.1.4. The Sse Class of Hsp70s The Sse1 and Sse2 proteins are quite distantly related to other Hsp70s (approximately 13 to 28% identical) but are 76% identical to each other. Another protein distantly related to Hsp70s is the mammalian heat-inducible protein, Hsp110 (Subjeck et al., 1983). Hsp110 is present at basal levels in mammalian cells and is induced upon a heat shock. A portion of Hsp110 has been found associated with the nucleolus, but the function of Hsp110 is unknown. The Sse proteins and Hsp110 are at least as similar to each other as either is to yeast or mammalian Hsp70s (Lee-Yoon et al., 1995). The subcellular localization of the Sses has not been determined. However, their lack of obvious targeting sequences
Molecular chaperones and folding catalysts
164
suggests that they may be localized to the cytosol, or perhaps the nucleus. The Sse proteins were originally isolated based upon their calmodulin binding activity (Mukai et al., 1993). In addition, SSE1 was identified as a multicopy suppressor of the heat shock sensitive phenotype of an ira1 mutation, which causes hyperactivation of the RAS-cAMP pathway (Shirayama et al., 1993). The single sse2 mutant has no observable phenotype while sse1 mutant cells have a slow growth phenotype. The phenotype of the double mutant is indistinguishable from that of the sse1 mutant. Like Hsp110, there is no understanding of the functional role which Sses play in the cell. 2.1.5. Pdr13 A recently discovered Hsp70 encoded by PDR13 has been implicated in the regulation of genes involved in pleiotropic drug resistance. PDR13 was isolated in a screen for multicopy suppressors which elevates the resistance of cells to the mitochodrial ATPase inhibitor oligomycin; in addition, overexpression of PDR13 renders cells resistant to cycloheximide (Katzmann et al., 1995). The same phenotype results from an overexpression of YOR1, a gene encoding an ABC transporter. Interestingly, increased copy number of PDR13 leads to elevated expression of YOR1 and other genes involved in pleiotropic drug resistance (Hallstrom et al., 1998). These Pdr1 3 effects require the presence of the zinc finger transcription factor Pdr1, which is necessary for both expression of YOR1 (Balzi and Goffeau, 1995) and oligomycin resistance. Strains that lack Pdr1 are cold-sensitive for growth and hypersensitive to drugs. Together these results suggest that one function of Pdr1 3 may be to positively regulate the transcription factor Pdr1, and thus the expression of a set of genes involved in drug resistance, including ABC transporters. 2.2. Hsp70s of the Mitochondria The matrix of the yeast mitochondria contains three Hsp70s. The most abundant is the well-characterized SSC1 protein which is essential for translocation of many proteins across the mitochondrial membranes. The mechanism of action of this Hsp70 in the translocation process has been studied extensively. Ssc1 is tethered to the inner membrane via its interaction with Tim44, a component of the translocation complex. It binds to the mature part of the translocating polypeptide as well as the presequence. The mechanism by which Ssc1 provides the directionality to the movement of the polypeptide chain is a matter of debate. Ssc1 may harness the energy of Brownian motion by preventing “backsliding” of the polypeptide chain. On the other hand, Ssc1 may act as a force generating motor, with energy derived from a conformational change in Ssc1 as a result of the binding/hydrolysis of ATP. The chapter by Dekker and Pfanner discusses in detail the function of Ssc1 in translocation. Consistent with the apparent multiple roles of the cytosolic Ssa proteins, analysis of conditional SSC1 mutants has provided evidence that Ssc1 is involved in several other mitochondrial processes. These additional functions include the folding of both mitochondrially-encoded (Hermann et al., 1994) and translocated polypeptides. Degradation of matrix proteins by the PIM1 pro tease is also hampered in SSC1 mutants
Genetic dissection of the Hsp70 chaperone system of yeast
165
suggesting a role for Ssc1 in presenting proteins to the proteolytic apparatus in a conformation permissive for degradation (Wagner et al., 1994). In all of these roles Ssc1 is presumed to bind to partially unfolded protein substrates, acting in cycles of binding and release as do other members of the Hsp70 family. In a seemingly different capacity, Ssc1 is the noncatalytic subunit of the mitochondrial site-specific endonuclease Endo.Sce1 (Morishima et al., 1990), and is required for activity of the catalytic subunit. SSH1 and SSJ1, two genes uncovered during the genome sequencing project, have recently begun to be examined. Only a preliminary examination of SSJ1 has been done at this time. Ssj1 is 80% identical to Ssc1 and contains an apparent mitochondrial leader sequence. However, mitochondrial localization has not been demonstrated at this time. No phenotype of a disruption mutant has been observed (Forster, Schilke, Davis and Craig, unpublished result). Since Ssj1 appears to be expressed at much lower levels than Ssc1, and since the proteins are about 80% identical, one might expect that overexpression of Ssj1 could compensate for lack of Ssc1. However preliminary experiments in which expression of Ssj1 was driven by the SSC1 promoter, ssc1 mutant phenotypes were not suppressed. An understanding of the role of this Hsp70 therefore awaits more extensive investigation. Unlike SSJ1, disruption of SSH1, which encodes a protein 53% identical to Ssc1, is detrimental to yeast cells. SSH1 is not an essential gene; however, cells lacking Ssh1 are cold-sensitive, growing very poorly at temperatures of 23°C and below (Schilke et al., 1996). The functional relationship between Ssc1 and Ssh1 is complex. The level of Ssc1 in the matrix is much higher than that of Ssh1, perhaps as much as a hundred times. However, overexpression of SSC1 can suppress the cold-sensitive phenotype of a ssh1 deletion. In contrast, a fusion which expresses Ssh1 under the control of the SSC1 promoter does not suppress the phenotypes of an SSC1 temperature-sensitive mutation. These results suggest that Ssh1 is unable to carry out the essential roles of Ssc1 when expressed at the same levels as Ssc1. However, Ssc1 can carry out the functions of Ssh1 when present in several hundred fold excess. The function of Ssh1 in mitochondria is not known. However clues have come from the analysis of suppressors of the cold-sensitive phenotype (Schilke et al., 1996). These suppressors arose spontaneously at a high frequency. All were found to be petite mutants, that is, they were respiratory incompetent. 15 of the 16 analyzed were rho0 mutants that lacked mitochondrial DNA (mtDNA) completely, while the 16th contained reduced amounts of mtDNA. Therefore, absence or reduction in the amount of mtDNA overcomes the deleterious effects of the absence of Ssh1. This result suggests two possible roles for Ssh1 in mitochondrial physiology. First, Ssh1 might be involved in the folding of mitochondrially-encoded proteins. In this case the synthesis of proteins in the mitochondria might be deleterious because these proteins would be likely to misfold and aggregate, interfering with other mitochondrial functions. Elimination of the information encoding these proteins would eliminate this problem. However, mitochondrially encoded proteins synthesized in the absence of Ssh1 appear to be soluble. A second hypothesis is that Ssh1 is involved in mtDNA replication and that the absence of Ssh1 results in alterations in the DNA replication process that are deleterious to the cell. In this model, also, the absence of mtDNA would eliminate the problem. In theory these two models could be distinguished by the analysis of rho− mutants of
Molecular chaperones and folding catalysts
166
SSH1. rho− mutants have deletions of mtDNA which eliminate protein synthesis; however, the remaining mtDNA is amplified such that rho− cells have the same total amount of mtDNA as do wild-type rho+ cells (Dujon, 1981). If the deleterious effect caused by the absence of Ssh1 is due to its role in protein folding, rho− cells should be suppressed, since no mitochondrial protein synthesis occurs. However, if DNA replication is defective in ssh1 cells, rho− and rho+ cells should have the same phenotype, since both contain the same amount of mitochondrial DNA. However, even though rho− cells arise at a rate of 1–5% in a normal population, no rho− ssh1 cells were obtained upon analysis of 4500 cells. The failure to obtain rho− ssh1 cells suggests that the presence of rho- DNA in an ssh1 background is deleterious. These studies cannot distinguish definitively between the two most obvious roles for Ssh1. However, the failure to isolate rho- cells and the fact that the growth of ssh1 cells does not correlate with the amount of mitochondrial proteins synthesized, suggest an involvement of Ssh1 in DNA replication. Further studies will be required to define the cellular role of Ssh1 and to determine the basis of the functional differences between Ssh1 and Ssc1. 2.3. Hsp70s of theER The lumen of the ER of S. cerevisiae contains two members of the Hsp70 family, the essential KAR2 protein, which is the homologue of mammalian BiP, and a recently identified protein Ssil/Lhs1/Cer1 (Craven et al., 1996; Baxter et al., 1996; Hamilton and Flynn, 1996) which is 25% identical in sequence to Kar2 as well as other Hsp70s. Kar2 is important in the translocation of proteins from the cytosol into the lumen of the ER, as shown by both in vivo and in vitro experiments (Rose et al., 1989; Vogel et al., 1990). Kar2 interacts directly with translocating polypeptides and has been proposed to provide a driving force for the unidirectional movement of the polypeptide across the ER membrane. This role is analogous to that carried out by Ssc1 of the mitochondrial matrix to drive translocation into that organelle. This function of Kar2 and its homologue BiP in the translocation process is described in more detail in the chapter by Haas and Zimmermann. Kar2 appears to also be involved in the folding of proteins after their translocation. In a KAR2 mutant the vacuolar protein carboxypeptidase Y aggregates after translocation into the ER, and fails to be transported into the vacuole (Simons et al., 1995). Unlike KAR2, SSI1/LHS1/CER1 is not an essential gene. Strains lacking Ssi1 are coldsensitive for growth (Baxter et al., 1996; Craven et al., 1996; Hamilton and Flynn, 1996). However the severity of the defect varies with the strain background. ssi1 strains accumulate precursor forms of several proteins which are normally translocated into the ER. In one case the precursor has been shown to be on the cytosolic side of the ER membrane indicating a defect in translocation. However, consistent with the nonessential nature of Ssil, the block is not complete. Overexpression of Scj1, a DnaJ of the ER, suppresses the translocation defect of ssi1. In addition to the presumptive role in translocation, it has been suggested that Ssi1 functions in the folding of proteins in the lumen. This suggestion is based upon the fact that the expression of Ssil, like Kar2, is induced by the presence of incorrectly folded proteins in the ER. In addition, an SSI1 deletion displays synthetic lethality with a mutation in IRE1, which encodes a protein
Genetic dissection of the Hsp70 chaperone system of yeast
167
required for the transcriptional response to an increase in unfolded protein in the ER (Craven et al., 1996). Thus, like Kar2, Ssil is thought to be involved in both the translocation and subsequent folding of proteins in the ER lumen. Ssil has also been implicated in the refolding of denatured proteins. The ER has an ATP-dependent, heat-resistant machinery which is able to refold thermally denatured proteins in cells (Jamsa et al., 1995). In the absence of Ssi1, an in vivo heat-denatured and aggregated reporter enzyme failed to be resolubilized and reactivated, as in wild-type cells. Instead, it was slowly degraded (Saris et al., 1997). Genetic interactions between KAR2 and SSI1 provide some clues as to the relationship between these related proteins (Baxter et al., 1996; Craven et al., 1996; Hamilton and Flynn, 1996). A number of conditional KAR2 mutants exhibit different synthetic growth defects in the presence of the ssi1 deletion mutation. Conditional KAR2 alleles have been grouped into classes based on their phenotypes (Rose, Vogel and Scidmore, personal communication). Class I alleles such as kar2–159 cause a block in translocation at the nonpermissive temperature and rapid loss of viability. Class II alleles, which fail to grow but remain viable at the nonpermissive temperature, do not cause a block in translocation. However, cells having alleles of this class, such as kar2–1, have defects in a number of posttranslocational events. Class III alleles such as kar2–191 show both translocational and posttranslocational defects. The genetic interactions of an ssi1 deletion with the various KAR2 mutants can be grouped into the same classes (Baxter et al., 1996). Class I alleles show synthetic lethality with ssi1. For example, a ssi1 kar2–159 is inviable at all temperatures, including those at which both ssi1 and kar2–159 strains are viable. Class II alleles are not synthetically lethal but show impaired growth compared to both parental strains. On the other hand, a Class III allele, kar2–191, shows better growth in the presence of the SSI1 mutation than in its absence. In addition, overexpression of Ssi1 is deleterious in a kar2–191 background, but not in Class I or Class II kar2 strains tested. What is the explanation of these complex genetic interactions? Since ssi1 cells are impaired in the translocation process, as well as protein folding events, it is likely that Ssi1 can partially substitute for KAR2 function, thus explaining the synthetic lethality or growth defect of Class I and Class II mutant strains, respectively. The synthetic growth enhancement of the Class III allele is more complicated to explain. It has been previously shown that this allele is synthetically lethal with sec63–1, a missense mutation in the gene encoding a component of the translocation machinery homologous to DnaJ (see chapter, Haas and Zimmermann), even in the presence of wild-type KAR2 (Scidmore et al., 1993), suggesting dominant effects of this mutation. Perhaps kar2–191 protein interacts with the translocation apparatus in an inhibitory fashion and Ssi1 enhances this interaction. Such an interaction would lead to improved growth in the absence of Ssi1. Consistent with this idea, increased expression of Ssi1 in this genetic background is deleterious. More thorough analysis of the various alleles of KAR2 should prove very helpful in understanding the relationship between these two Hsp70s of the ER lumen.
Molecular chaperones and folding catalysts
168
3. THE DNAJ MULTIGENE FAMILY OF YEAST DnaJ has the ability to bind unfolded polypeptides and prevent their aggregation and therefore can be considered a chaperone in its own right. DnaJ, however, also interacts with DnaK, stimulating the hydrolysis of ATP (see chapter by Buchberger et al.). Using the defining motif of DnaJ-related proteins, the “J” domain, 16 potential DnaJs have been identified through screening of the yeast genome, only nine of which have been reported in the literature. Ydj1 is the best studied DnaJ-related protein of S. cerevisiae and has many of the same properties as E. coli DnaJ. Although the studies of other yeast DnaJs is not as advanced, it is likely that all interact with one or more Hsp70s and some, particularly those with an obvious cysteine-rich motif able to form a “zinc finger”, also bind unfolded proteins. Below the best studied DnaJs are discussed. 3.1. DnaJs of the Cytosol/Nucleus: Ydj1, Sis1 and Zuol 3.1.1. Ydj1 The best studied DnaJ of yeast is Ydj1 (Mas5). Ydj1 was identified independently by two groups based on its association with membranes (Caplan and Douglas, 1991) and its role in the facilitation of translocation into mitochondria (Atencio and Yaffe, 1992). Ydj1 is not essential, but cells lacking it grow very poorly at 23°C and are unable to form colonies at 37°C. The structural organization of Ydj1 is similar to that of DnaJ of E.coli, containing glycine- and cysteine-rich regions in addition to the “J” domain (Figure 2) (see chapter by Buchberger et al., for structure/function aspects of DnaJ proteins). In fact, expression of DnaJ can overcome some of the
Figure 2 Schematic diagram of DnaJ homologs from Saccharomyces
Genetic dissection of the Hsp70 chaperone system of yeast
169
cerevisiae. The protein structure of the eight DnaJ homologs of S. cerevisiae that have been studied to date are shown. Protein motifs discussed in the text are shown. Subcellular location and phenotypes of deletion mutants are indicated.
defects caused by the lack of Ydj1. Unlike other DnaJs of S. cerevisiae, Ydj1 contains a “CAAX” box at its C-terminus and is farnesylated (Caplan et al., 1992), explaining the observed association with membranes. While at 23°C about 15% of Ydj1 was found associated with the membrane fraction, after a shift to 37°C for 1 hour, more than 65% was found membrane-associated. Cells carrying a mutant form of Ydj1 having a substitution for the conserved cysteine in the “CAAX” box are temperature-sensitive for growth and have increased amounts of Ydj1 in the cytosol. Together these results suggest that a membrane-associated function of Ydj1 may be particularly important at high temperatures. What functions does Ydj1 carry out in the cell? Analysis of the temperature-sensitive ydj1-151 mutant strongly indicates that Ydj1 functions in the translocation of some proteins from the cytosol to the ER and mitochondria (Caplan et al., 1992). The ydj1-151 mutant is defective in translocation of prepro- -factor but not three other ER proteins tested, and in the translocation of the F1 , and subunits of the F1F0 ATPase into mitochondria. As discussed in more detail in section 6, Ydj1 has also been implicated in the folding and activation of several heterologous proteins which require Hsp90 function. Although there is no genetic evidence for the function of Ydj1 in the general protein folding pathway, biochemical experiments strongly suggest that possibility. Like DnaJ acting with DnaK, Ydj1 acting with Ssa1 can faciltate the refolding of denatured luciferase (Levy et al., 1995). Furthermore, by itself, Ydj1 can bind to denatured rhodanase and prevent its aggregation (Cyr, 1995). 3.1.2. Sis1 Sis1 is an essential DnaJ of the yeast cytosol (Luke et al., 1991). While Sis1 contains the “J” domain and the glycine-rich repeat, it does not possess the cysteine-rich region found in Ydj1 and DnaJ. However, it does possess a Met-rich region. Whether this region is analogous to the cysteine-rich region in allowing binding to polypeptide substrates is unknown. Compelling genetic and physiological experiments indicate a role for Sis1 in protein synthesis, specifically initiation (Zhong and Arndt, 1993). First, Sis1 comigrates with polysomes and is specifically enriched in fractions that contain only a few ribosomes per messenger. This biased distribution toward small polysomes suggests that Sis1 associates with ribosomes only during and shortly after initiation. In addition, 80S monosomes rapidly accumulate upon temperature upshift of a temperature-sensitive sis1 mutant, indicating a defect in initiation in the absence of functional Sis1. Extragenic suppressors of sis1 ts mutations were found to be in genes that encode ribosomal proteins of the 60S subunit. Although this result underscores a role for Sis1 in translation, how it might function in the initiation process remains unknown. It is likely that Sis1 is involved in either assembly or dissembly of a complex rather than in general protein folding, since it is apparently present only in the very early stages of translation. It should also be
Molecular chaperones and folding catalysts
170
remembered that Sis1 was first isolated as a multicopy suppressor of mutations in SIT4, which encodes a serine/threonine phosphatase (Luke et al., 1991). Whether this suppressing ability is an indirect effect of its role in translation or reflects an additional function remains to be defined. 3.1.3. Zuo1 Zuo1 was originally detected in nuclear extracts in an assay designed to isolate Z-DNAbinding proteins (Zhang et al., 1992). Independently, Zuo1 was isolated as a transfer RNA binding protein (Wilhelm et al., 1994). Isolation by these two independent nucleic acid binding assays strongly suggests that Zuo1 functions by binding either DNA or RNA, or both. Zuo1 is not an essential protein. However, zou1 strains grow slowly compared to wild-type strains. Unlike Ydj1, Zuo1 contains no cysteine rich domain. Zuo1 has yet to be purified in sufficient quantities to determine whether it can interact with any Hsp70s, prevent protein aggregation, or facilitate protein folding. 3.2. DnaJ of the Mitochondria A single DnaJ, Mdj1, has been identified in the matrix of the mitochondria (Rowley et al., 1994). Mdj1 is not an essential protein, however mdj1 mutants are unable to grow at 37°C. Interestingly, mdj1 mutants are unable to maintain mitochondrial DNA, that is they are rho0. It is unclear at this time whether Mdj1 is actually involved in DNA replication, since many mutations which disrupt the structure or function of mitochondria fail to maintain normal mitochondrial DNA (Piskur, 1994). Further analysis of strains lacking Mdj1 has indicated a requirement for Mdj1 in the efficient folding of proteins imported into mitochondria, as well as to prevent their aggregation (Rowley et al., 1994). Luciferase which has been imported into mitochondria aggregates after partial unfolding in response to a shift to 40°C. This aggregation occurs significantly more rapidly in mdj1 mitochondria compared to wild-type mitochondria (Prip-Buus et al., 1996). These results with Mdj1 are reminiscent of those obtained with DnaJ of E. coli. For example, DnaJ is able to maintain the soluble state of denatured luciferase (Szabo et al., 1994; Schröder et al., 1993). Perhaps these similarities are not surprising considering the symbiotic origin of mitochondria. However, more direct comparisons of DnaJ and Mdj1 await purification of the mitochondrial DnaJ. 3.3. DnaJs of the ER: Sec63, Scj1 and Jem1 Two DnaJs of the ER have been extensively studied. Sec63, an essential integral membrane protein with its “J” domain exposed to the lumenal side of the membrane and Scj1, a soluble protein of the lumen. Recently a third DnaJ of the ER, Jem1, has been identified (Nishikawa and Endo, 1997). Like Sec63, Jem1 is localized to the ER membrane facing the lumen. jem1 cells are defective in karyogamy, as are kar2 mutants, suggesting that Jem1 may function together with Kar2 in nuclear fusion. jem1 cells are temperature-sensitive in the presence of an scj1 deletion. Sec63 was originally identified in a screen for mutants defective in translocation of
Genetic dissection of the Hsp70 chaperone system of yeast
171
proteins from the cytosol across the ER membrane (Sadler et al., 1989; Sanders et al., 1992). Sec63 is related to DnaJ only in a region of about 70 amino acids. This “J” domain is situated between two membrane spanning domains and is exposed to the ER lumen. The Kar2:Sec63 interaction was established through analysis of Kar2-Sec63 complexes isolated from detergent-treated microsomes (Brodsky and Schekman, 1993; Scidmore et al., 1993). A mutation resulting in the alteration of a highly conserved alanine of the “J” domain to a threonine resulted in an unstable complex. In addition, the wild-type complex could be disrupted by the addition of an ATP analogue, indicative of a dynamic interaction modulated by nucleotide binding or hydrolysis. That the interaction between Sec63 and Kar2 is important is underscored by the finding that kar2–1 and sec63–1 mutants are synthetically lethal and that second-site suppressors of sec63–1 map to KAR2 (Scidmore et al., 1993). Scj1, which was initially believed to be a mitochondrial protein, is found almost exclusively in the lumen of the ER (Schlenstedt et al., 1995). Unlike Sec63, whose only homology with DnaJ is the “J” domain, Scj1 contains a cysteine-rich region and a glycine-rich region as well as a “J” domain. Scj1 was originally isolated as a gene which in multiple copies causes mislocalization of a protein that should have been targeted to the nucleus (Blumberg and Silver, 1991). However, strains lacking Scj1 grow as well as wild-type strains under a number of conditions tested. Additional genetic studies suggest that Scj1 interacts functionally with Kar2 (Schlenstedt et al., 1995). kar2–159 is synthetically lethal with scj1; interestingly, kar2–1 is synthetically lethal with sec63–1 but is viable in combination with an SCJ1 deletion. As a test of whether Scj1 interacts directly with Kar2, the “J” domain of Scj1 was swapped for the Sec63 “J” domain. The chimeric Sec63-Scj1 protein could substitute for the wild-type Sec63, maintaining viability of a SEC63 deletion strain. These results suggest that Kar2 interacts with two different DnaJ-type proteins in the ER. Proteins being translocated into the ER may first interact with Kar2 and Sec63, and then, after release from the translocation machinery, with Kar2 and Scj1. This analysis of DnaJ chimeras was extended by substituting the “J” domain of mitochondrial Mdj1 and cytosolic Sis1 into Sec63 (Schlenstedt et al., 1995). Neither chimeric protein was able to substitute for Sec63. However, changing only 3 nonconserved amino acids of the Sis1 “J” domain to those found in Sec63 allowed the Sis1:Sec63 chimera to function. The ability of the Scj1 “J” domain to function in place of that of Sec63, while the “J” domains of Sis1 and Mdj1 cannot, may reflect an evolution in the specificity of Hsp70/DnaJ pairs. The interactions between Hsp70s and DnaJs are believed to be specific, such that each Hsp70 will function only with a specific subset of DnaJs. It is likely that Hsp70/DnaJ pairs have co-evolved to maintain important functional interactions, while diverging from family members that are separated by cellular localization or functional differences. Analysis of the numerous DnaJs which have been identified in the genome, but have yet to be functionally analyzed may be required before the full range of interactions between Hsp70s and DnaJs is appreciated. 3.4. Other DnaJs: Xdj1 and Caj1 The basic characterization of two other genes encoding proteins with significant
Molecular chaperones and folding catalysts
172
similarity to the “J” domain, XDJ1 and CAJ1, has been reported. XDJ1, identified based on its sequence similarity to DnaJ, encodes a protein with both a glycine-rich and a cysteine-rich domain as well as the characteristic “J” domain (Schwarz et al., 1994). Strains lacking Xdj1 have no obvious phenotype and expression from XDJ1 was not detected at either the RNA or protein level. Therefore, it is likely that XDJ1 is expressed at an extremely low level or under a particular condition which was not analyzed in the study. Caj1 was isolated due to its propensity to bind calmodulin (Mukai et al., 1993); Ssb1, Sse1 and Sse2 were isolated in the same study. Caj1 contains no glycine-rich domain, nor a cysteine-rich domain. However, the middle portion of the protein contains a periodic repeat of leucine residues, potentially a leucine zipper. The growth properties of a CAJ1 disruption mutant were indistinguishable from wild-type.
4. Mge1, A GrpE-RELATED GENE OF THE MITOCHONDRIAL MATRIX In addition to its interaction with DnaJ, DnaK physically interacts with GrpE, a nucleotide exchange factor that promotes the release of bound ATP or ADP (Johnson et al., 1989; Buchberger et al., 1994) (see chapter by Buchberger et al.). Functionally, the ability of GrpE to promote ADP release is more important, as it allows the protein to exchange ADP for ATP, which is present in higher concentrations in the cell. Although there have been thorough searches for GrpE homologues in all cellular compartments of eucaryotes, such proteins have only been found in the matrix of mitochondria. It is possible that the fundamental properties of the cytosolic and ER Hsp70s obviate the need for such release factors. There are several pieces of evidence that suggest cytosolic Hsp70s may in fact be biochemically different from procaryotic or mitochondrial Hsp70s. For example, although the intrinsic ATPase activities of the cytosolic yeast Ssa1 and E. coli DnaK are similar, the ability of Ydj1 to stimulate the Ssa1-ATPase activity is significantly greater (Cyr and Douglas, 1994; Ziegelhoffer et al., 1995) than that reported for the DnaK-DnaJ system, in which GrpE is also required for maximal stimulation (Liberek et al., 1991). Furthermore, polypeptide substrates decrease the stability of nucleotide binding to cytosolic Hsp70 (Sadis and Hightower, 1992; Ziegelhoffer et al., 1995), an effect not observed with DnaK Interestingly, a protein termed Hip, which actually stabilizes the Hsc70-ADP complex, has been isolated in mammalian cells (Höhfeld et al., 1995; see chapters by Ha et al., and Buchberger et al.). Most compelling is the fact that cytosolic Hsp70s require only an appropriate DnaJ homologue for refolding activity, while DnaK requires both DnaJ and GrpE. Together this evidence supports a circumstantial case for the absence of a nucleotide exchange factor for Hsp70s in the cytosol. Alternatively, however, nucleotide exchange factors may exist, but be unrelated in sequence to GrpE, and thus have escaped detection up to this point. The best characterized mitochondrial GrpE-related protein is that of S. cerevisiae, called Mge1 or Yge1. MGE1 was first identified on the basis of its similarity in sequence to E. coli grpE (Bolliger et al., 1994; Laloraya et al., 1994), as well as its ability, in multiple copies, to promote staurosporine resistance (Ikeda et al., 1994). MGE1 is an essential gene encoding a protein that is 34% identical to GrpE. Temperature-sensitive
Genetic dissection of the Hsp70 chaperone system of yeast
173
mutations in MGE1 result in delayed translocation of precursors as well as impaired maturation of imported proteins (Laloraya et al., 1995; Westerman et al., 1995). It is quite certain that Mge1 and Ssc1 function together in the translocation process. Mge1 can be quantitatively co-immunoprecipitated with Ssc1 in isolated mitochondria (Ungermann et al., 1994), and both Mge1 and Ssc1 can be co-immunoprecipitated with a precursor protein that is trapped at the site of import (Voos et al., 1994). The role of Mge1 in this process is discussed in more detail in the chapter by Dekker and Pfanner. Using purified components, Mge1 has been shown to stimulate the release of both ATP and ADP from Ssc1 (Miao et al., 1996). Like GrpE’s interaction with DnaK, Mge1 stably associates with Ssc1 in vitro and this interaction is resistant to high salt, but disrupted by the addition of ATP. The temperature sensitive mutant ssc1–3 produces a protein that fails to stably interact with Mge1, and Mge1 cannot stimulate the release of ATP from this mutant protein. The change in the ssc1–3 protein lies very close to the position of an analogous loop structure on the surface of DnaK which has been implicated in its interaction with GrpE (Buchberger et al., 1994). Deletion of this loop in Ssc1 eliminates detectable interaction with Mge1 (Miao et al., 1996). Interestingly, two mutants with a single amino acid change in this loop do not eliminate the stable binding with Mge1, yet do not stimulate the release of nucleotide. Therefore this loop is likely to be important not only for binding to Mge1 but also for generating a conformational change necessary for Mge1-induced nucleotide release.
5. FUNCTIONAL INTERACTIONS AMONG CLASSES OF CHAPERONES IN PROTEIN FOLDING As discussed in some detail above, Hsp70s and DnaJs function together. Recent data support the idea that these two chaperones also function with other classes of chaperones as well. For example, both Hsp70 and DnaJ function with chaperones of the Hsp90 class (Bohen et al., 1995). It has been known for some time that in mammalian cells Hsp90s function in the maturation of transcription factors of the steroid hormone receptor class. In addition, Hsp90 associates with the oncogenic tyrosine kinase p60v-src during its maturation. More recently, both Hsp70s and DnaJs have also been shown to associate with steroid hormone receptors during the maturation process (Smith and Toft, 1993). Maturation of both receptors and p60v-src have been studied in a heterologous yeast system. Expression of active receptors in the heterologous yeast system requires the endogenous Hsp90 as well as Ydj1 (Picard et al., 1990; Bohen and Yamamoto, 1993). Also, expression of p60s-src in yeast cells results in lethality (Brugge et al., 1983; Dey et al., 1996; Kimura et al., 1995), but both HSP90 and YDJ1 mutants are resistant to the toxic effects of p60v-src (Xu and Lindquist, 1993). Genetic evidence indicates an overlap in function between the SSA proteins and Hsp104, a molecular chaperone of the Clp/Hsp100 class (see chapters by Maurizi et al., and Lindquist et al.). Hsp104 has been shown to be critical for the reactivation of denatured luciferase in vivo (Parsell et al., 1994), while Ssa proteins can carry out this function in vitro. Phenotypic analysis of strains containing mutations in both SSA genes and HSP104 have also been informative. For example, overexpression of Ssa1 in cells
Molecular chaperones and folding catalysts
174
lacking Hsp104 partially restores thermotolerance to hsp104 cells, while the survival of ssa1 ssa3 ssa4 hsp104 cells is 100 fold less than hsp104 cells after exposure to high temperatures (Sanchez et al., 1993). In addition, ssa1 ssa2 cells grow more slowly than wild-type but reasonably well at 30°C, while hsp104 mutants show no detectable growth defect. However, ssa1 ssa2 hsp104 cells are barely viable under the same conditions. A more detailed understanding of the functional relationship between the Ssas and Hsp104 awaits further experimentation.
6. CONCLUSIONS As described in this chapter, we have learned a great deal about the function of molecular chaperones of the Hsp70 and DnaJ classes since their first identification. We now know that the Hsp70s of yeast fall into multigene families, that each well-studied family participates in multiple cellular functions, and that they function with other classes of chaperone. Both biochemical and genetic experiments indicate that many of the Hsp70 and DnaJ chaperones within a cell are functionally distinct. The recent completion of the sequence of the yeast genome has allowed the identification of all members of the established classes of molecular chaperones in this organism. This information should provide the opportunity in the coming years to map out the complex web of proteinprotein interactions, cooperative roles, and cellular functions among the full complement of molecular chaperones in a single eukaryotic cell. Only then will we truly begin to understand the intricacies of the roles of chaperones in cell physiology.
7. REFERENCES Atencio, D. and Yaffe, M. (1992). MAS5, a yeast homolog of DnaJ involved in mitochondrial import. Mol. Cell. Biol. , 12 , 283–291. Balzi, E. and Goffeau, A. (1995). Yeast multidrug resistance: The PDR network. J. Bioenerg. Biomembr. , 27 , 71–76. Baxter, B., James, P., Evans, T. and Craig, E.A. (1996). SSI1 encodes a novel Hsp70 homolog of the Saccharomyces cerevisiae endoplasmic reticulum. Mol. Cell Biol. , 16 , 6444–6456. Becker, J., Yan, W. and Craig, E.A. (1996). Functional interaction of cytosolic Hsp70 and DnaJ-related protein, Ydj1p, in protein tranlocation in vivo . Mol Cell Biol. , 16 , 4378–4386. Blumberg, H. and Silver, P. (1991). A homologue of the bacterial heat-shock gene dnaJ that alters protein sorting in yeast. Nature , 349 , 627–629. Bohen, S.P., Kralli, A. and Yamamoto, K.R. (1995). Hold ‘em and fold ‘em: Chaperones and signal transduction. Science , 268 , 1303–1304. Bohen, S.P. and Yamamoto, K.R. (1993). Isolation of Hsp90 mutants by screening for decreased steroid receptor function. Proc. Natl. Acad. Set. USA , 90 , 11424–11428. Bolliger, L., Deloche, O., Glick, B., Georgopoulos, C., Jeno, P., Kronidou, N., Horst, M., Morishima, N. and Schatz, G. (1994). A mitochondrial homolog of bacterial GrpE interacts with mitochondrial hsp70 and is essential for viability. EMBO J. , 13 , 1998– 2006.
Genetic dissection of the Hsp70 chaperone system of yeast
175
Boorstein, W. and Craig, E.A. (1990a). Transcriptional regulation of SSA3, an HSP70 gene from Saccharomyces cerevisiae. Mol. Cell. Biol. , 10 , 3262–3267. Boorstein, W.R. and Craig, E.A. (1990b). Structure and Regulation of the SSA4 HSP70 Gene of Saccharomyces cerevisiae. J. Biol Chem. , 265 , 18912–18921. Boorstein, W.R., Ziegelhoffer, T. and Craig, E.A. (1994). Molecular evolution of the HSP70 multigene family . J. Mol. Evol , 38 , 1–17. Brodsky, J.L. and Schekman, R. (1993). A Sec63p-BiP complex from yeast is required for protein translocation in a reconstituted proteoliposome. J.Cell Biol. , 123 , 1355– 1363. Brugge, J., Jarosik, G., Andersen, J., Queral-Lustig, A., Fedor-Chaiken, M. and Broach, J. (1983). Expression of Rous sarcoma virus-transforming protein pp60v-src in Saccharomyces cerevisiae cells. Mol. Cell Biol. , 3 , 9–19. Buchberger, A., Schröder, H., Buttner, M., Valencia, A. and Bukau, B. (1994). A conserved loop in the ATPase domain of the DnaK chaperone is essential for stable binding of GrpE. Nature Struct. Biol. , 1 , 95–101. Bukau, B. and Walker, G. (1990). Mutations altering heat shock specific subunit of RNA polymerase suppress major cellular defects of E. coli mutants lacking the DnaK chaperone. EMBO J. , 9 , 4027–4036. Caplan, A., Cyr, D. and Douglas, M. (1992). YDJ1p facilitates polypeptide translocation across different intracellular membranes by a conserved mechanism. Cell , 71 , 1143– 1155. Caplan, A. and Douglas, M. (1991). Characterization of YDJ1: A yeast homologue of the bacterial dnaJ protein. J. Cell Biol. , 114 , 609–621. Caplan, A., Tsai, J., Casey, P. and Douglas, M. (1992). Farnesylation of YDJ1p is required for function at elevated growth temperatures in S. cerevisae. J. Biol. Chem. , 267 , 18890–18895. Chirico, W.J. (1992). Dissociation of complexes between 70 kDa stress proteins and presecretory proteins is faciliated by a cytosolic factor. Biochem. Biophys. Res. Comm. , 189 , 1150–1156. Craig, E.A. and Jacobsen, K. (1984). Mutations of the heat-inducible 70 kilodalton genes of yeast confer temperature-sensitive growth. Cell , 38 , 841–849. Craig, E.A. and Jacobsen, K. (1985). Mutations in cognate gene of Saccharomyces cerevisiae HSP70 result in reduced growth rates at low temperatures. Mol. Cell. Biol. , 5 , 3517–3524. Craig, E.A., Ziegelhoffer, T., Nelson, J., Laloray, S. and Halladay, J. (1995). Complex multigene family of functionally distinct Hsp70s of yeast. In (eds.), Cold Spring Harbor Symposia on Quantitative Biology: Protein Kinesis: The dynamics of protein trafficking and stability , Cold Spring Harbor Laboratory Press, Cold Spring Harbor, pp. 441–449. Craven, R.A., Egerton, M. and Stirling, C.J. (1996). A novel Hsp70 of the yeast ER lumen is required for the efficient translocation of a number of protein precursors. EMBO J. , 15, 2640–2650. Cyr, D. (1995). Cooperation of the molecular chaperone Ydj1 with specific Hsp70 homologs to suppress protein aggregation. FEBS Letters , 359 , 129–132. Cyr, D.M. and Douglas, M.G. (1994). Differential regulation of Hsp70 subfamilies by the eukaryotic DnaJ homologue YDJ1. J. Biol. Chem. , 269 , 9798–9804. Deshaies, R., Koch, B., Werner-Washburne, M., Craig, E. and Schekman, R. (1988). A subfamily of stress proteins faciliates translocation of secretory and mitochondrial precursor polypeptides. Nature , 332 , 800–805.
Molecular chaperones and folding catalysts
176
Dey, B., Caplan, A. and Boschelli, F. (1996). The Ydj1 molecular chaperone facilitates formation of active p60v-src in yeast. Mol. Biol. Cell , 7 , 91–100. Dujon, B. (1981). Mitochondrial genetics and function . Molecular biology of the yeast Saccharomyces: Life cycle and inheritance. Cold Spring Harbor, New York, Cold Spring Harbor Press. Freeman, B., Myers, M., Schumacher, R. and Morimoto, R. (1995). Identification of a regulatory motif in Hsp70 that affects ATPase activity, substrate binding and interaction with HDJ-1. EMBO J. , 14 , 2281–2292. Gao, B., Biosca, J., Craig, E.A., Greene, L.E. and Eisenberg, E. (1991). Uncoating of coated vesicles by Yeast hsp70 proteins. J. Biol. Chem. , 266 , 19565–19571. Georgopoulos, C., Liberek, K., Zylicz, M. and Ang, D. (1994). Properties of the heat shock proteins of Escherichia coli and the autoregulation of the heat shock response. In R.I.Morimoto, A.Tissieres and C.Georgopoulos, (eds.), The biology of heat shock proteins and molecular chaperones , Cold Spring Harbor Laboratory Press, Plainview, NY, pp. 209–250. Halladay, J. and Craig, E. (1995). A heat shock transcription factor with reduced activity suppresses a yeast HSP70 mutant. Mol Cell. Biol. , 15 , 4890–4897. Hallstrom, T.C., Katzmann, D.J., Torres, R.J., Sharp, W.J. and Moye-Rowley, W.S. (1998). Regulation of Transcription Factor Pdr1p Function by an Hsp70 Protein in Saccharomyes cerevisiae. Mol. Cell Biol , 18 , 1147–1155. Hamilton, T.G. and Flynn, G.C. (1996). Cer1p, a novel Hsp70-related protein required for posttranslational endoplasmic reticulum translocation in yeast. J. Biol Chem. , 272 , 30610–30613. Harrison, G., Bohm, A.A. and Nelson, H.C.M. (1994). Crystal structure of the DNA binding domain of the heat shock transcription factor. Science , 263 , 224–227. Hermann, J., Stuart, R., Craig, E. and Neupert, W. (1994). Mitochondrial heat shock protein 70, a molcular chaperone for proteins encoded by mitochondrial DNA. J. Cell Biol , 127 , 893–902. Höhfeld, J., Minami, Y. and Hartl, F.-U. (1995). Hip, a novel cochaperone involved in the eukaryotic Hsc70/Hsp40 reaction cycle. Cell , 83 , 589–598. Ikeda, E., Yoshida, S., Mitsuzawa, H., Uno, I. and Toh-e, A. (1994). YGE1 is a yeast homologue of Escherichia coli grpE and is required for maintenance of mitochondrial function. FEBS Letters , 339 , 265–268. James, P., Pfund, C. and Craig, E. (1997). Functional specificity among Hsp70 molecular chaperones. Science , 275 , 387–389. Jamsa, E., Vakula, N., Arffman, A., Kilpelainen, A. and Makarow, M. (1995). In vivo reactivation of heat-denatured protein in the endoplasmic reticulum. EMBO J. , 14 , 6028–6033. Johnson, C., Chandrasekhar, G.N. and Georgopoulos, C. (1989). Escherichia coli DnaK and GrpE heat shock proteins interact both in vivo and in vitro . J. Bacteriol. , 171 , 1590–1596. Katzmann, D., Hallstrom, T., Voet, M., Wysock, W., Golin, J., Volckaert, G. and MoyeRowley, W.S. (1995). Expression of an ATP-binding cassette transporter-encoding gene (YOR1). is required for oligomycin resistance in Saccharomyces cerevisiae . Mol. Cell Biol. , 15 , 6875–6883. Kimura, Y., Yahara, I. and Lindquist, S. (1995). Role of the protein chaperone YDJ1 in establishing Hsp90-mediated signal transduction pathways. Science , 268 , 1362–1365. Laloraya, S., Dekker, P., Voos, W., Craig, E. and Pfanner, N. (1995). Mitochondrial GrpE modulates the function of matrix Hsp70 in translocation and maturation of
Genetic dissection of the Hsp70 chaperone system of yeast
177
proteins. Mol. Cell. Biol. , 15 , 7098–7105. Laloraya, S., Gambill, B.D. and Craig, E.A. (1994). A role for a eukaryotic GrpE-related protein, Mge1p, in protein translocation. Proc. Natl. Acad. Sci. USA , 91 , 6481–6485. Landry, S.J., Jordan, R., McMacken, R. and Gierasch, L.M. (1992). Different conformations for the same polypeptide bound to chaperones DnaK and GroEL. Nature , 355 , 455–457. Lee-Yoon, D., Easton, D., Murawski, M., Burd, R. and Subjeck, J.R. (1995). Identification of a Major Subfamily of Large hsp70-like Proteins through the Cloning of the Mammalian 110-kDa Heat Shock Protein. J. Biol. Chem. , 270 , 15725–15733. Levy, E., McCarty, J., Bukau, B. and Chirico, W. (1995). Conserved ATPase and luciferase refolding activities between bacteria and yeast Hsp70 chaperones and modulators. FEBS Letters , 368 , 435–440. Liberek, K., Marszalek, J., Ang, D. and Georgopoulos, C. (1991). Escherichia coli: DnaJ and GrpE heat shock proteins jointly stimulate ATPase activity of DnaK Proc. Natl. Acad. Sci. USA , 88 , 2874–2878. Luke, M., Suttin, A. and Arndt, K. (1991). Characterization of SIS1, a Saccharomyces cerevisiae homologue of bacterial dnaJ proteins. J. Cell Biol , 114 , 623–638. Miao, B., Davis, J.E. and Craig, E.A. (1996). Mge1 functions as a nucleotide release factor for Ssc1, a mitochondrial Hsp70 of Saccharomyces cerevisiae . J. Mol. Biol , 265 , 541–552. Morishima, N., Nakagawa, K., Yamamoto, E. and Shibata, T. (1990). A subunit of yeast site-specific endonuclease Sce1 is a mitochondrial version of the 70-kDa heat shock protein. J. Biol. Chem. , 265 , 15189–15197. Mukai, H., Kuno, T., Tanaka, H., Hirata, D., Miyakawa, T. and Tanaka, C. (1993). Isolation and characterization of SSE1 and SSE2, new members of the yeast HSP70 multigene family. Gene , 132 , 57–66. Nelson, R.J., Heschl, M. and Craig, E.A. (1992). Isolation and characterization of extragenic suppressors of mutations in the SSA hsp70 genes of Saccharomyces cerevisiae . Genetics , 131 , 277–285. Nelson, R.J., Ziegelhoffer, T., Nicolet, C., Werner-Washburne, M. and Craig, E.A. (1992). The translation machinery and seventy kilodalton heat shock protein cooperate in protein synthesis. Cell , 71 , 97–105. Nishikawa, S. and Endo, T. (1997). The yeast JEM1p is a DnaJ-like protein of the endoplasmic reticulum membrane required for nuclear fusion. J. Biol. Chem. , 272 , 12889–12892. Ng, D.T.W., Brown, J.D. and Walter, P. (1996). Signal sequences specify the targeting route to the endoplamic reticulum membrane. J. Cell Biol. , 134 , 269–278. Ohba, M. (1994). A 70-kDa heat shock cognate protein suppresses the dfects caused by a proteasome mutation in Saccharomyces cerevisiae. FEBS Letters , 351 , 263–266. Parsell, D., Kowal, A., Singer, M. and Lindquist, S. (1994). Protein disaggregation mediated by heat-shock protein Hsp104. Nature , 372 , 475–478. Picard, D., Khursheed, B., Garabedian, M., Fortin, M., Lindquist, S. and Yamamoto, K. (1990). Reduced levels of hsp90 compromise steroid receptor action in vivo. Nature , 348 , 166–168. Piskur, J. (1994). Inheritance of the yeast mitochondrial genome. Plasmid , 31 , 229–241. Prip-Buus, C., Westermann, B., Schmitt, M., Langer, T., Neipert, W. and Schwarz, E. (1996). Role of the mitochondrial DnaJ homologue, Mdj1p, in the prevention of heatinduced protein aggregation. FEBS Letters , 390 , 142–146. Rose, M., Vogel, J., Scidmore, M. and Misra, L. (1989). KAR2, the yeast homologue of
Molecular chaperones and folding catalysts
178
mammalian BiP/ GRP78, is required for translocation of proteins into the endoplasmic reticulum. J. Cell. Biochem. , suppl. 13E, 12–273. Rowley, N., Prip-Buus, C., Westermann, B., Brown, C., Schwarz, E., Barrell, B. and Neupert, W. (1994). Mdj1p, a novel chaperone of the DnaJ family, is involved in mitochondrial biogenesis and protein folding. Cell , 77 , 249–259. Sadis, S. and Hightower, L.E. (1992). Unfolded proteins stimulate molecular chaperone Hsc70 ATPase by accelerating ADP/ATP exchange. Biochemistry , 31 , 9406–9412. Sadler, I., Chiang, A., Kurihara, T., Rothblatt, J., Way, J. and Silver, P. (1989). A yeast gene important for protein assembly into the endoplasmic reticulum and the nucleus has homology to DnaJ, an Escherichia coli heat shock protein. J. Cell Biol , 109 , 2665–2675. Sanchez, Y., Parells, D.A., Taulien, J., Vogel, J.L., Craig, E.A. and Lindquist, S. (1993). Genetic evidence for a functional relationship between Hsp104 and Hsp70. J. Bacterial , 175 , 6484–6491. Sanders, S.L., Whitfield, K.M., Vogel, J.P., Rose, M.D. and Schekman, R.W. (1992). Sec61p and BiP directly facilitate polypeptide translocation into the ER. Cell , 69 , 353–365. Saris, N., Holkeri, H., Craven, R.A., Stirling, C.J. and Makarow, M. (1997). The Hsp70 homologue Lhs1p is involved in a novel function of the yeast endoplasmic reticulum, refolding and stabilization of heat-denatured protein aggregates. J. Cell Biol. , 137 , 813–824. Schilke, B., Forster, J., Davis, J., James, P., Walter, W, Laloraya, S., Johnson, J., Miao, B. and Craig, E. (1996). Cold-sensitivity of an S. cerevisiae mutant lacking a newly identified mitochondrial Hsp70 is suppressed by loss of mitochondrial DNA. J. Cell Biol , 134 , 603–614. Schlenstedt, G., Harris, S., Risse, B., Lill, R. and Silver, P.A. (1995). A yeast DnaJ homologue, Scj1p, can function in the endoplasmic reticulum with BiP/Kar2p via a conserved domain that specifies interactions with Hsp70s. J. Cell Biol , 129 , 979–988. Schröder, H., Langer, T., Hartl, F.-U. and Bukau, B. (1993). DnaK, DnaJ and GrpE form a cellular chaperone machinery capable of repairing heat-induced protein damage. EMBO J. , 12 , 4137–4144. Schwarz, E., Westermann, B., Caplan, A., Ludwig, G. and Neupert, W. (1994). XDJ1, a gene encoding a novel non-essential DnaJ homologue from Saccharomyces cerevisiae . Gene , 145 , 121–124. Scidmore, M.A., Okamura, H.H. and Rose, M.D. (1993). Genetic interactions between KAR2 and SEC63, encoding eukaryotic homologues of DnaK and DnaJ in the endoplasmic reticulum. Mol Biol. Cell , 4 , 1145–1159. Shea, J.E., Toyn, J.H. and Johnston, L.H. (1994). The budding yeast U5 snRNP Prp8 is a highly conserved protein which links RNA splicing with cell cycle progression. Nucl Acid Res. , 22 , 5555–5564. Shirayama, M., Kawakami, K., Matsui, Y., Tanaka, K. and Toh-e, A. (1993). MSI3, a multicopy suppressor of mutants hyperactivated in the RAS-cAMP pathway, encodes a novel HSP70 protein of Saccharomyces cerevisiae. Mol Gen. Genet. , 240 , 323–332. Simons, J.F., Ferro-Novick, S., Rose, M.D. and Helenius, A. (1995). BiP/Kar2p serves as a molecular chaperone during carboxypeptidase Y folding in yeast. J. Cell Biol , 130 , 41–49. Smith, D.F. and Toft, D.O. (1993). Steroid receptors and their associated proteins. Mol Endocrinol , 7 , 4–11. Stone, D.E. and Craig, E.A. (1990). Self-regulation of 70-kilodalton heat shock proteins
Genetic dissection of the Hsp70 chaperone system of yeast
179
in Saccharomyces cerevisiae. Mol. Cell. Biol , 10 , 1622–1632. Subjeck, J., Shyy, T., Shen, J. and Johnson, R. (1983). Association between the mammalian 110,000-dalton heat-shock protein and nucleoli. J. Cell Biol , 97 , 1389– 1395. Szabo, A., Langer, T., Schroeder, H., Flanagan, J., Bukau, B. and Haiti, F.U. (1994). The ATP hydrolysis-dependent reaction cycle of the Escherichia coli Hsp70 system— DnaK, DnaJ and GrpE. Proc. Natl. Acad. Set. USA , 91 , 10245–10349. Ungermann, C., Neupert, W. and Cyr, D.M. (1994). The role of Hsp70 in conferring unidirectionality on protein translocation into mitochondria. Science , 266 , 1250– 1253. Vogel, J.P., Misra, L.M. and Rose, M.D. (1990). Loss of BiP/GRP78 function blocks translocation of secretory proteins in yeast. J. Cell Biol , 110 , 1885–1895. Voos, W., Gambill, B.D., Laloraya, S., Ang, D., Craig, E.A. and Pfanner, N. (1994). Mitochondria! GrpE is present in a complex with hsp70 and preproteins in transit across membranes. Mol Cell. Biol. , 14 , 6627–6634. Wagner, I., Arlt, H., van Dyck, L., Langer, T. and Neupert, T. (1994). Molecular chaperones cooperate with PIM1 protease in the degradation of misfolded protein in mitochondria. EMBO J. , 13 , 5135–5145. Werner-Washburne, M., Stone, D.E. and Craig, E.A. (1987). Complex interactions among members of an essential subfamily of hsp70 genes in Saccharomyces cerevisiae. Mol. Cell Biol , 7 , 2568–2577. Westerman, B., Prip-Buus, C., Neupert, W. and Schwarz, E. (1995). The role of the GrpE homologue, Mge1p, in mediating protein import and protein folding in mitochondria. EMBO J. , 13 , 1998–2006. Wilhelm, M.L., Reinholt, J., Gangloff, J., Dirheimer, G. and Wilhelm, F.X. (1994). Transfer RNA binding protein in the nucleus of Saccharomyces cerevisiae. FEBS Letters , 349 , 260–264. Xu, Y. and Lindquist, S. (1993). Heat-shock protein hsp90 governs the activity of pp60vsrc kinase. Proc. Natl. Acad. Sci. USA , 90 , 7074–7078. Zhang, S., Lockshin, C., Herbert, A., Winter, E. and Rich, A. (1992). Zuotin, a putative Z-DNA binding protein in Saccharomyces cerevisiae. EMBO J. , 11 , 3787–3796. Zhong, T. and Arndt, K.T. (1993). The yeast SIS1 protein, a DnaJ homolog, is required for initiation of translation. Cell , 73 , 1175–1186. Ziegelhoffer, T., Lopez-Buesa, P. and Craig, E.A. (1995). The dissociation of ATP from hsp70 of Saccharomyces cerevisiae is stimulated by both Ydj1p and peptide substrates. J. Biol. Chem. , 270 , 10412–10419.
8. FUNCTIONS IN DEVELOPMENT MICHEL MORANGE Unité de Génétique Moléculaire, Département de Biologie, Ecole Normale Supérieure, 46 rue d’Ulm, 75230 Paris Cedex 05, France
1. Introduction 2. The Stress-inducibility of Chaperones During Development 3. The Different Experimental Systems Used to Follow Chaperone Expression During Embryogenesis 4. Chaperones as actors in development 5. Mechanisms Controlling the Developmental Expression of Chaperones 6. Research Perspectives 7. Acknowledgements 8. References
1. INTRODUCTION The study of the variations in expression of chaperones during development, and the role of these variations, is at the same time trivial and difficult. Trivial because it is well known that the expression of chaperones differs widely from one adult tissue to another (Tanguay et al., 1993) and that these differences are established during development. Moreover, it has been extensively documented that the synthesis of chaperones is dependent upon metabolic conditions (Lanks, 1986) and is not uniform during the different phases of the cell cycle (Wu and Morimoto, 1985; Milarski and Morimoto, 1985; Jérome et al., 1993). Since the embryo is submitted during development to dramatic variations in metabolism, and given that the duration of the different cell cycle phases varies widely in embryonic cells, it can be expected that the expression of chaperones will vary during development. If one considers the functions of chaperones in protein folding (Hartl, 1996), translocation across membranes but also protein degradation (Hayes and Dice, 1996), it is also obvious that the concentrations of chaperones and heat shock proteins (HSPs) will vary during development, in parallel with protein synthesis, cell remodeling or secretion (as during the formation of the extracellular matrix). As we shall briefly show below, a large number of observations in different organisms provides evidence for variations in chaperone levels and synthesis during embryogenesis (Heikkila, 1993; Morange, 1997). However these observations do not satisfy
Functions in development
181
developmental biologists, who are not interested in variations in the levels of expression of housekeeping genes, but are looking for genes specifically involved in crucial steps of embryogenesis. As we shall see, only now are we beginning to have data clearly demonstrating that chaperones play a major structural and/or regulatory role at some specific steps of development. The numerous early studies describing the expression of chaperones during development were necessary steps to point out these crucial functions. These studies required the preliminary characterization of the chaperone-encoding genes in the different organisms studied. Due to the complexity of the genetic organization of heat shock genes, this has been a long and tedious task which is still not fully completed. We will rapidly present the main developmental systems used, outlining the specific interest of each and the most important observations which have emerged. We will then focus on three different systems which are closest to providing data showing the crucial requirement of chaperones in development. We will conclude with a few words about the mechanisms which might be responsible for the specific expression of chaperones during embryogenesis and the developments we can expect in the near future.
2. THE STRESS-INDUCIBILITY OF CHAPERONES DURING DEVELOPMENT The most meaningful observations for developmental biologists concern the spontaneous expression of chaperones during embryogenesis. However, chaperones are also stress proteins, able to participate in the repair of cell damage (Parsell and Lindquist, 1993; Freeman and Morimoto, 1996). Two major arguments support the study of variations in chaperone (HSP) inducibility during development: 1) Stress treatments lead to uneven alteration of normal embryonic development. Some tissues and embryological processes are more sensitive to stress than others: for instance, in mammals, neural tube closure can be perturbed by many different chemical or physical stresses, leading to serious abnormalities (Walsh et al., 1991; Walsh et al., 1997). This sensitivity is probably due to the perturbation in relative amounts of important regulatory proteins which follows stress (Welte et al., 1995). 2) The inducibility of the stress proteins is different in the different tissues of an adult, but also in the different tissues of an embryo and at different stages of development. As a general rule, the early embryo is poorly responsive to stress. This is true from plants (reviewed in Vierling, 1991) and Drosophila to mammals. In mammals, the fully grown oocyte is unresponsive to stress and heat shock (Curci et al., 1991), as is the early embryo before the blastocyst stage (Morange et al., 1984). The uninducibility of heat shock genes is also observed in some differentiated cell lines where it might be linked with methylation (Gorzowski et al., 1995). One major goal of embryologists is to try to relate these two groups of observations and to elucidate the role of induced HSPs and chaperones in the protection of the embryo. The answer is not straightforward. The presence of a strong HSP induction in adult or embryonic tissues does not prevent these tissues from being a target for the stress: everything happens as if the intensity of HSP induction was more a reflection of the damages supported by the cells than an adaptive response of these cells (Brown, 1994). In
Molecular chaperones and folding catalysts
182
some cases, such as in mammalian testis, it has even been suggested that the accumulation of HSPs might have deleterious effects, blocking cell division or differentiation (Sarge, 1995).
3. THE DIFFERENT EXPERIMENTAL SYSTEMS USED TO FOLLOW CHAPERONE EXPRESSION DURING EMBRYOGENESIS We will only briefly discuss HSP and chaperone expression during plant development. The abundance and diversity of small HSPs, but also of other stress proteins such as GRP78, the presence of members of each HSP family in the different cell compartments (Boston et al., 1996), the occurrence of different heat shock transcription factors, some of which are themselves HSPs (Nover et al., 1996), make this study particularly difficult. However, it is obvious that the expression of the different HSPs, as well as their intracellular distribution, both in the presence and absence of a stress, is highly variable during development and in the different tissues of the plant. It is remarkable how diverse physiological conditions (stress?) are able to alter HSP synthesis: light, dehydration or starvation, addition of plant hormones (abscisic and gibberellic acids), mechanical wounds, as well as the circadian rhythm. The mechanisms involved in these variations of expression appear as diverse as the conditions leading to them (regulation by translational control or modification of protein stability as well as by transcription). Even if it clearly appears that HSP synthesis is “part of the developmental program” of plants (Wehmeyer et al., 1996), it has not yet been demonstrated that HSPs are essential components of this developmental program. In animals, the best studied system is, for historical reasons, Drosophila. In this organism, there exist four genes coding for the small HSPs, whose inducibility and spontaneous expression vary dramatically during development (Haass et al., 1990). For instance, HSP23 is specifically expressed during embryogenesis in glial cells identified as Midline glial cells (Tanguay, 1989). Similar observations have been reported for the three other small hsps which are expressed in characteristic structures of the larva and adult, mainly in the central nervous system and gonads (Arrigo and Mehlen, 1994; Michaud et al., 1997). The most puzzling observations concern the localization of hsp83 mRNA during oogenesis and early embryogenesis. This mRNA is transferred from nurse cells to the oocyte. It is evenly distributed throughout the embryo during the initial divisions, and then remains concentrated at the posterior pole by a selective stabilization process. It is then taken up by the pole cells. hsp83 gene is also transcribed from the zygotic genome at the syncitial blastoderm stage in the anterior third of the embryo (Ding et al., 1993). Both maternal and zygotic expressions are under the strict control of important developmental genes of Drosophila, oskar for maternal gene expression and bicoid for the zygotic one, suggesting that the control of HSP83 synthesis is critically important at these early stages of embryogenesis. Another developmental system in which the expression of chaperones has been studied at the very early stages is the amphibian embryo (Heikkila, 1993; Herberts et al., 1993; Heikkila et al., 1997). The need to isolate specific probes and to prepare antibodies has so far prevented a full use of this system, which is particularly well suited to the study of the
Functions in development
183
first steps of embryogenesis. It has been clearly demonstrated that the expression and inducibility of heat shock genes is controlled during oogenesis and early embryonic development by chromatin state, as well as by specific activators (Landsberger et al., 1995a and b; Ali et al., 1993). Stabilization of mRNAs might also be important for the progressive appearance of the heat-inducible expression of the small HSPs (Ohan and Heikkila, 1995). One recent observation has underlined the spontaneous variations in the level of one small HSP (HSP30) (Helbing et al., 1996) which occur during development. Mouse embryo has also been studied at very early stages for a differential induced or spontaneous expression of HSPs and chaperones (Mezger et al., 1991). The preimplantation phase has been most closely examined and data on the expression of the major HSPs and chaperones during mouse post-implantation development is still incomplete. One of the earliest events associated with the onset of zygotic genome transcription (Schultz, 1993) is the transient expression of the hsp70 gene (Bensaude et al., 1983; Manejwala et al., 1991; Christians et al., 1997a). Both maternal and paternal genomes contribute to this synthesis as demonstrated with parthenogenotes and androgenotes (Barra and Renard, 1988). The hsc70 gene is also expressed but, in contrast to hsp70, its expression remains high during later development (Morange et al., 1984; Giebel et al., 1988). Transgenic animals harboring a hybrid hsp70.1-luciferasegene reproduce this transient expression (Christians et al., 1995), definitely demonstrating that this expression is not an artefact of manipulation. Two mechanisms have been proposed for this transient expression: activation by the maternal HSF1 factor, which is abundant and localized in the nucleus at this stage of development (Christians et al., 1997b), or the action of the basal promoter elements (Bevilacqua et al., 1995; Bevilacqua et al., 1997). The physiological significance of this expression remains unknown and the generalization of this phenomenon to other mammals (and vertebrates) is problematic. hsp70 is also spontaneously expressed later in development, first in the extra-embryonic tissues, and then in specific tissues of the embryo (Kothary et al., 1987; Loones et al., 1997). The precise nature of the tissues or cells concerned, and the meaning of these expressions, await further studies. The expression in these tissues contrasts with the very low constitutive expression of HSP70 in most cells. A transient expression of HSP70 during in vitro differentiation of the human cell line K562 has been described (Theodorakis et al., 1989; Sistonen et al., 1992 and 1994) and shown to also occur during normal hematopoiesis (Singh andYu, 1984). The generalization of this observation to other mammals (in particular mouse) is once again problematic. Both HSC70 and HSP90 are ubiquitously expressed, in the adult as well as in embryonic tissues of mice. These two proteins, as well as HSP90 and HSP60, are also expressed at a very high level at the blastocyst stage and during the period immediately following implantation of the embryo (Bensaude and Morange, 1983). A remarkably high level of HSC70 and HSP90 can be found during mid-gestation in the nervous system and in neural crest cells, as well as in the hypertrophic chondrocytes during bone morphogenesis (Loones, Figure 1; Loones et al., 1997). In the first case, this high expression persists in the adult (Itoh et al., 1993; Izumoto
Molecular chaperones and folding catalysts
184
Figure 1 HSP distribution in the ulna cartilage primordium of a day 15.5 mouse embryo, revealed with peroxidase-coupled second antibody, a) HSP90 is expressed all along the chondrocyte differentiation pathway, in particular in the hypertrophic chondrocytes. b) HSC70 expression during chondrogenesis is limited to hypertrophic chondrocytes. scale bar: 100 m
and Herbert, 1993) and might play a crucial role in the protection of the nervous system against aggressions (Nowak et al., 1994). Three observations on the mammalian embryo deserve further study and will be discussed below. 1) The existence of a Heat Shock Transcription Factor (HSF2) specifically expressed during most of embryogenesis, as well as during hematopoiesis or spermatogenesis. 2) The very specific expression of the small HSP/ -crystallins in the heart, muscle and some neural structures in the brain during development. 3) The highly specific expression of minor members of the HSP/HSC70 family during spermatogenesis. The mouse embryo surely has further surprises and discoveries in store since the potential of this developmental system (existence of multiple mutations in the developmental genes, possibility of gene disruption by homologous recombination, existence of many in vitro differentiation systems) is far from being fully exploited.
Functions in development
185
The last system to have provided some clues on the role of chaperones and HSPs during development is the zebrafish. This organism is presently the best adapted to the combined use of genetic and developmental techniques, a sort of hybrid of what is possible in the mouse and in the amphibian. The work is still in its initial phases and the characterization of the different HSP and chaperone genes is not yet complete. For instance, nothing is presently known about the structure and expression of the small HSPs in this organism. However by whole mount in situ hybridization analysis, Sass et al. (1996) have revealed the very specific localization of zebrafish hsp90 mRNA in muscles during development (see below). hsp90 mRNA does not show the same distribution, but is detectable within the central nervous system at control temperatures. Finally, despite its interest for developmental biology, C. elegans has not so far furnished outstanding results concerning the role of chaperones and HSPs in development. However, it has confirmed the heat-inducible tissue specific expression of small HSPs during development (Stringham et al., 1992), quite reminiscent of what is observed in Drosophila.
4. CHAPERONES AS ACTORS IN DEVELOPMENT Previous data clearly demonstrate the variations in the levels of HSP and chaperone gene expression during embryogenesis. We would now like to describe in more detail three selected situations in which additional experimental data support the hypothesis that HSPs and chaperones are not only parts of the developmental program, but important actors in these programs. The first concerns the expression of hsp90during zebrafish development (Sass et al., 1996). The level of expression of this gene is weak and limited to a small subset of cells within the pre-somitic mesoderm, later in the somites and pectoral fin buds of the developing embryo. The expression of the gene coding for one of the Myogenic Regulatory Factors (MyoD) coincides with the expression of hsp90 in the somites and pectoral fin buds. The expressions of hsp90 and myoD are down-regulated in adult muscle. HSP90 has been shown to enhance the DNA-binding activity of murine MyoD synthesized in vitro (Shaknovich et al., 1992), as well as of the MyoD/E12 heterodimer purified from bacterial cells (Shue and Kohtz, 1994), the most probable form of MyoD in -vivo (Lassar et al., 1991). Therefore, in addition to a potential protection against stress, HSP90 might control the activity of the Myogenic Regulatory Factor MyoD and indirectly the muscular differentiation pathway. The co-localization of HSP90 and MRFs is also observed in chick (Sass and Krone, 1997), but in mouse (Loones et al., 1997). This situation is perhaps not unique. HSP90 interacts also with the developmental basic helix-loop-helix transcription factor Sim in Drosophila (McGuire et al., 1995). However, in this system, it remains to be demonstrated that the level of active Sim/ Arnt transcription factor complex is controlled by HSP90. Other good candidates for an active role in differentiation are small HSP/ -crystallins (Arrigo and Landry, 1994). A and R-crystallins have a high degree of identity with small HSPs and their expression is not limited, despite their names, to the vertebrate lens
Molecular chaperones and folding catalysts
186
(Bhat and Nagineni, 1989; Dubin et al., 1989). As briefly noted above, these proteins have a very specific pattern of expression in Drosophila, amphibians and mammals (Kato et al., 1991a and b, Gernold et al., 1993; Klemenz et al., 1993) with, in two of these systems, a preferential, but not exclusive expression in the central nervous system. In addition, their expression is controlled by steroid hormones. In many differentiation systems, a transient increase or modification of small HSPs has been also observed (Shakoori et al., 1992; Mehlen et al., 1997). The overexpression of small HSPs does not appear to be a general phenomenon occurring when a cell enters into a differentiation pathway, but a specific modification associated with certain specific differentiation pathways (Minowada and Welch, 1995). Reduction in the level of small HSP by the expression of antisense constructs fully prevents differentiation of ES cells after LIF deprivation. Cells go on dividing and die by apoptosis instead of differentiating (Mehlen et al., 1997). Evidence in favor of an essential role of small HSP in differentiation has also been obtained in the human promyelocytic leukemia HL60 cells. A decrease in the level of HSP27 by the addition of antisense oligonucleotides does not block, but alters the morphological changes induced by retinoic acid treatment (Chaufour et al., 1996). Moreover the function of small HSPs remains largely unknown, the hypotheses put forward ranging from a generalized chaperone function, on involement in microfilament organization (Lavoie et al., 1993) to a role in the control of red/ox state and apoptosis (Mehlen et al., 1996a, b). At present we can only guess as to the role of these proteins at early steps of differentiation. The third system which merits a full description and in which we now have strong evidence for a crucial role of chaperones concerns not the major chaperones, but a minor form of the HSP70/HSC70 family which is specifically expressed during spermatogenesis in the rat or mouse testis. Two proteins have been described: HSC70t whose RNA is present in early round spermatids whereas the protein accumulates later in elongating spermatids (Matsumoto and Fujimoto, 1991 a; Matsumoto et al., 1991b), and HSP70–2 which is detectable earlier in spermatocytes, but also in round and elongating spermatids (Allen et al., 1988a, b; Zakeri et al., 1988; Krawczyk et al., 1988; Rosario et al., 1992; Raab et al., 1995). The expression of hsp70–2 is controlled at the transcriptional level (Wisniewski et al., 1993; Dix et al., 1996a). The HSP70–2 protein is associated with the synaptonemal complexes in the nucleus of meiotic spermatocytes. The hsp70–2 gene is also expressed in the central nervous system (Murashov and Wolgemuth, 1996a, b). Recently the corresponding gene has been knocked-out by homologous recombination (Dix et al., 1996b). Targeted hsp70–2 gene disruption leads to failed meiosis, germ cell apoptosis and male infertility. HSP70–2 acts as a chaperone on CDC2 kinase, enabling it to interact with cyclin B1 to trigger the transition from the G2 to M phase (Zhu et al., 1997). It is also required for the disassembly of the synaptonemal complex prior to the diplotene stage, either directly, or indirectly through its activation of CDC2 kinase (Dix et al., 1997). Despite the expression of the hsp70–2 gene in the central nervous system, no abnormalities have so far been described in this tissue.
Functions in development
187
5. MECHANISMS CONTROLLING THE DEVELOPMENTAL EXPRESSION OF CHAPERONES Two mechanisms can be considered for the developmental regulation of chaperones and HSPs. The first is the occurrence, in the promoters of the corresponding genes, of tissue and stage specific regulatory sequences to which specific transcription factors bind. Such appears to be the case for the small HSPs in Drosophila (Michaud et al., 1997) or acrystallins in mammals (Gopal-Srivastava and Piatigorsky, 1993 and 1994; GopalSrivastava et al., 1995; Haynes et al., 1995. Cvekl et al., 1995), with a major role for the PAX-6 protein in the lens (Cvekl and Piatigorsky, 1996). However, the sequences responsible for the specific expression of B-crystallins in the heart (Gopal-Srivastava et al., 1995) overlap a heat-shock element. In fact, the second mechanism is the developmental regulation of chaperone genes by the same upstream sequences as those involved in the stress response. The activation of the chaperone genes would result either from the “developmental” activation of HSF1, the factor responsible for the enhanced expression of heat shock genes during the stress response—such might be the case during the mouse Zygotic Genome Activation at the 2-cell stage—, or from the binding to these sequences of a factor, distinct from HSF1, whose activity would be controlled during differentiation and development. Such appears to be the case during spermatogenesis, where another heat shock factor, HSF2 (Sarge et al., 1991 and 1993), is actively synthesized in the spermatocytes and round spermatids (Sarge et al., 1994). It is localized in the nucleus and is spontaneously active for binding to HSE sequences. Two different isoforms, generated by alternative splicing, have recently been described for this factor (Fiorenza et al., 1995; Goodson et al., 1995), which differ by their transcriptional activity. HSF2 is a transcriptional activator, whereas HSF2 is less active or even behaves as an inhibitor (Leppä et al., 1997). The same HSF2 factor also is involved in the transient overexpression of HSP70 and HSP90 which accompanies the hemininduced differentiation of the human erythroleukemia cell line K562 (Theodorakis et al., 1989; Sistonen et al., 1992 and 1994). HSF2 (Leppä et al., 1997) is activated during this process, and has been shown by genomic footprinting to bind to the HSP70 promoter (Sistonen et al., 1992). However, it remains difficult to determine whether the developmental role of HSF2 can be extended beyond these two well established situations. HSF2 might be involved in the control of crystallin gene expression (Sax and Piatigorsky, 1994). The less active form of HSF2 , is expressed at the blastocyst stage (Mezger et al., 1994a and b) as well as in embryonal carcinoma and embryonic stem cells (Murphy et al., 1994; Mezger et al., 1994 b). It remains present during a major part of embryogenesis, but its relative level in the different tissues do not coincide with the expression of any chaperone or HSP (Rallu et al., 1997): on the contrary, in many tissues, it appears as if there is an inverse relation between the level of HSF2 and the rate of synthesis of HSPs and chaperones. Perhaps the two different forms of HSF2 play different roles. The transcriptionally active form would be involved in the direct developmental control of chaperone and gene expression, whereas the other form would facilitate chromatin reorganization, exert an inhibitory role on HSP expression, or participate in the regulation of genes other than heat shock genes:
Molecular chaperones and folding catalysts
188
the DNA-binding properties of HSF1 and HSF2 are slightly different (Kroeger et al., 1993; Kroeger and Morimoto, 1994), less “stringent” for HSF2 than for HSF1. Since the occurrence of at least two other developmentally-regulated HSFs has been described in the chicken (Nakai and Morimoto, 1993; Nakai et al., 1995) and humans (Nakai et al., 1997), the potential role of other HSFs in the control of HSP expression during development remains an open question.
6. RESEARCH PERSPECTIVES As we saw previously, there are now plenty of data from different systems demonstrating a specific expression of chaperones and HSPs at precise stages of embryogenesis. At least in one case, Hsp70–2, we have the crucial experiments which demonstrate that these proteins are active agents in the developmental program. During recent years, the work was delayed by the necessary description of heat shock and chaperone gene expression in different developmental systems, and the characterization in these different organisms of the corresponding genes and proteins. Our limited understanding of the functions of the less generalized chaperones still remains a handicap, since it prevents the construction of precise hypotheses to explain the involvement of HSPs in differentiation and development. The multiplicity of the genes coding for chaperones and HSPs and then possible functional redundancy is also an obstacle to obtaining rapid results by targeted gene disruption. Perhaps many researchers are also reluctant to undertake these long and difficult experiments, frightened that, due to the ubiquitous function of these proteins, the disruption of the corresponding genes will have dramatic effects, preventing any observation of their specific developmental functions. However, some molecular tools have not yet been fully exploited. In in vitro differentiation systems, addition of antisense oligonucleotides or transfection by antisense expression vectors would make it possible to test the specific role of HSPs in differentiation processes. The progressive improvements in the techniques of conditional and tissue-specific gene disruption will probably allow to bypass the present obstacles in the near future and to clearly demonstrate the role of chaperones and HSPs in development.
7. ACKNOWLEDGEMENTS We are indebted to E.Christians, M.-T.Loones, V.Mezger, E.Michel and M.Rallu for helpful discussions. This work was supported by grants from l’Association pour la Recherche sur le Cancer (ARC n° 6505), and from the Ministère de la Recherche (ACC).
8. REFERENCES Ali, A., Krone, P. and Heikkila, J.J. (1993). Expression of endogenous and microinjected hsp30 genes in early Xenopus laevis embryos. Dev. Genet ., 14 , 42–50. Allen, R.L., O’Brien, D.A. and Eddy, E.M. (1988). A novel hsp70-like protein (P70) is
Functions in development
189
present in mouse spermatogenic cells. Mol. Cell. Biol. , 8 , 828–832. Allen, R.L., O’Brien, D.A., Jones, C.C., Rockett, D.L. and Eddy, E.M. (1988). Expression of heat shock proteins by isolated mouse spermatogenic cells. Mol. Cell. Biol. , 8 , 3260–3266. Arrigo, A.-P. and Landry, J. (1994). Expression and function of the low molecular weight heat shock proteins In R.Morimoto, A.Tissières and C.Georgopoulos (eds.) Heat Shock Proteins: Structure, Function and Regulation , Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 335–373. Arrigo, A.-P. and Mehlen, P. (1994). Expression, cellular location and function of low molecular weight heat shock proteins (hsp20s). during development of the nervous system In J.Mayer and I.Brown (eds.) Heat Shock Proteins in the Nervous System , Academic Press, London, pp. 145–167. Barra, J. and Renard, J.-P. (1988). Diploid mouse embryos constructed at the late 2-cell stage from haploid parthenogenotes and androgenotes can develop to term. Development , 102 , 773–779. Bensaude, O., Babinet, C., Morange, M. and Jacob, F. (1983). Heat shock proteins, first major products of zygotic gene activity in mouse embryo. Nature , 305 , 331–333. Bensaude, O. and Morange, M. (1983). Spontaneous high expression of heat shock proteins in mouse embryonal carcinoma cells and ectoderm from day 8 mouse embryo. EMBO J. , 2 , 173–177. Bevilacqua, A., Kinnunen, L.H., Bevilacqua, S. and Mangia, F. (1995). Stage-specific regulation of murine Hsp68 gene promoter in preimplantation mouse embryos. Dev. Biol. , 170 , 467–478. Bevilacqua, A., Fiorenza, M.T. and Mangia, F. (1997). Developmental activation of an episomic hsp70 gene promoter in two-cell mouse embryos by transcription factor Sp1. Nucleic Acids Res. , 25 , 1333–1338. Bhat, S.P. and Nagineni, C.N. (1989). B-subunit of lens-specific -crystallin is present in other ocular and non-ocular tissues. Biochem. Biophys. Res. Comm. , 158 , 319–325. Boston, R.S., Viitanen, P.V. and Vierling, E. (1996). Molecular chaperones and protein folding in plants. Plant Mol Biol , 32 , 191–222. Brown, I.R. (1994). Induction of heat shock genes in the mammalian brain by hyperthermia and tissue injury In J.Mayer and I.Brown (eds.). Heat Shock Proteins in the Nervous System , Academic Press, London, pp. 32–54. Chaufour, S., Mehlen, P. and Arrigo, A.-P. (1996). Transient accumulation, phosphorylation and changes in the oligomerization of Hsp27 during retinoic acidinduced differentiation of HL-60 cells: possible role in the control of cellular growth and differentiation. Cell Stress Chap. , 1 , 225–235. Christians, E., Campion, E., Thompson, E.M. and Renard, J.-P. (1995). Expression of the HSP70–1 gene, a landmark of early zygotic activity in the mouse embryo, is restricted to the first burst of transcription. Development , 121 , 113–122. Christians, E., Michel, E. and Renard, J.-P. (1997a). Hsp70 genes and heat shock factors during preimplantation phase of mouse development. Cell. Mol. Life Sci. , 53 , 168– 178. Christians, E., Michel, E., Adenot, P., Mezger, V., Rallu, M., Morange, M. and Renard, J.-P. (1997b). Evidence for the involvement of mouse heat shock factor 1 in the atypical expression of the HSP70.1 heat shock gene during mouse zygotic genome activation. Mol. Cell. Biol. , 17 , 778–788. Curci, A., Bevilacqua, A., Fiorenza, M.T. and Mangia, F. (1991). Developmental
Molecular chaperones and folding catalysts
190
regulation of heat-shock response in mouse oogenesis: identification of differentially responsive oocyte classes during graafian follicle development. Dev. Biol , 144 , 362– 368. Cvekl, A, Kashanchi, F., Sax, C.M., Brady, J.N. and Piatigorsky, J. (1995). Transcriptional regulation of the mouse A-crystallin gene: activation dependent on a cyclic AMP-responsive element (DE1/CRE) and a Pax-6-binding site. Mol Cell. Biol , 15 , 653–660. Cvekl, A and Piatigorsky, J. (1996). Lens development and crystallin gene expression: many roles for Pax-6. Bioessays , 18 , 621–630. Ding, D., Parkhurst, S.M., Halsell, S.R. and Lipshitz, H.D. (1993). Dynamic Hsp83 RNA localization during Drosophila oogenesis and embryogenesis Mol. Cell Biol , 13 , 3773–3781. Dix, D.J., Rosario-Herrle, M., Gotoh, H., Mori, C., Goulding, E.H., Barrett, C.V. and Eddy, E.M. (1996a). Developmentally regulated expression of HSP70–2 and a HSP70– 2/lacZ transgene during spermatogenesis. Dev. Biol , 174 , 310–321. Dix, D.J., Allen, J.W., Collins, B.W., Mori, C., Nakamura, N., Poorman-Allen, P., Goulding, E.H. and Eddy, E.M. (1996b). Targeted gene disruption of Hsp70.2 results in failed meiosis, germ cell apoptosis, and male infertility. Proc. Natl Acad. Sci. USA , 93 , 3264–3268. Dix, D.J., Allen, J.W., Collins, B.W., Poorman-Allen, P., Mori, C., Blizard, D.R., Brown, P.R., Goulding, E.H., Strong, B.D. and Eddy, E.M. (1997). HSP70–2 is required for desynapsis of synaptonemal complexes during meiotic prophase in juvenile and adult mouse spermatocytes. Development , 124 , 4595–4603. Dubin, R.A., Wawrousek, E.F. and Piatigorsky, J. (1989). Expression of the murine Bcrystallin gene is not restricted to the lens. Mol. Cell. Biol. , 9 , 1083–1091. Fiorenza, M.T., Farkas, T., Dissing, M., Kolding, D. and Zimarino, V. (1995). Complex expression of murine heat shock transcription factors. Nucleic Acids Res. , 23 , 467– 474. Freeman, B.C. and Morimoto, R.I. (1996). The human cytosolic molecular chaperones hsp90, hsp70 (hsc70). and hdj-1 have distinct roles in recognition of a non-native protein and protein refolding. TheEMBO J. , 15 , 2969–2979. Gernold, M., Knauf, U., Gaestel, M., Stahl, J. and Kloetzel, P.-M. (1993). Development and tissue-specific distribution of mouse small heat shock protein hsp25. Dev. Genet. , 14 , 103–111. Giebel, L.B., Dworniczak, B.P. and Bautz, E.K. (1988). Developmental regulation of a constitutively expressed mouse mRNA encoding a 72-kDa heat shock-like protein. Dev. Biol. , 125 , 200–207. Goodson, M.L., Park-Sarge, O.-K. and Sarge, K.D. (1995). Tissue-dependent expression of heat shock factor 2 isoforms with distinct transcriptional activities. Mol. Cell. Biol. , 15 , 5288–5293. Gopal-Srivastava, R. and Piatigorsky, J. (1993). The murine B-crystallin/small heat shock protein enhancer: identification of BE-1, BE-2, BE-3, and MRF control elements. Mol. Cell. Biol. , 13 , 7144–7152. Gopal-Srivastava, R. and Piatigorsky, J. (1994). Identification of a lens-specific regulatory region (LSR). of the murine B-crystallin gene. Nucleic Acids Res. , 22 , 1281–1286. Gopal-Srivastava, R., Haynes, J.I. II and Piatigorsky, J. (1995). Regulation of the murine B-crystallin/ small heat shock protein gene in cardiac muscle. Mol. Cell. Biol. , 15 , 7081–7090.
Functions in development
191
Gorzowski, J.J., Eckerley, C.A., Halgren, R.G., Mangurten, A.B. and Phillips, B. (1995). Methylation-associated transcriptional silencing of the major histocompatibility complex-linked hsp70 genes in mouse cell lines. J. Biol. Chem. , 270 , 26940–26949. Haass, C., Klein, U. and Kloetzel, P.M. (1990). Developmental expression of Drosophila melanogaster small heat-shock proteins. J. Cell Sci. , 96 , 413–418. Hartl, F.U. (1996). Molecular chaperones in cellular protein folding. Nature , 381 , 571– 580. Hayes, S.A. and Dice, J.F. (1996). Roles of molecular chaperones in protein degradation. J. Cell Biol ., 132 , 255–258. Haynes, J.I. II, Gopal-Srivastava, R., Frederikse, P.H. and Piatigorsky, J. (1995). Differential use of the regulatory elements of the B-crystallin enhancer in cultured murine lung (MLg), lens ( TN4–1). and muscle (C2C12). cells. Gene , 155 , 151– 158. Heikkila, J.J. ed. (1993). Heat shock gene expression and development. Dev. Genet. , 14 , 1–158. Heikkila, J.J., Ohan, N., Tam, Y. and Ali, A. (1997). Heat shock protein gene expression during Xenopus development. Cell. Mol. Life Sci. , 53 , 114–121. Helbing, C., Gallimore, C. and Atkinson, B.G. (1996). Characterization of a Rana catesbeiana hsp30 gene and its expression in the liver of this amphibian during both spontaneous and thyroid hormone-induced metamorphosis. Dev. Genet. , 18 , 223–233. Herberts, C., Moreau, N. and Angelier, N. (1993). Immunolocalization of HSP70-related proteins constitutively expressed during Xenopus, laevis oogenesis and development”, Int. J. Dev. Biol. , 37 , 397–406. Itoh, H., Tashima, Y., Eishi, Y. and Okeda, R. (1993). Localization of HSP90 in rat brain. Int. J. Biochem. , 25 , 93–99. Izumoto, S. and Herbert, J. (1993). Widespread constitutive expression of hsp90 messenger RNA in rat brain. J. Neurosci. Res. , 35 , 20–28. Jérome, V., Vourc’h, C., Baulieu, E.E. and Catelli, M.G. (1993). Cell cycle regulation of the chicken hsp90 expression. Exp. Cell Res. , 205 , 44–51. Kato, K, Shinohara, H., Kurobe, N., Inaguma, Y., Shimizu, K. and Ohshima, K. (1991). Tissue distribution and developmental profiles of immunoreactive B crystallin in the rat determined with a sensitive immunoassay system. Biochim. Biophys. Acta , 1074 , 201–208. Kato, K., Shinohara, H., Kurobe, N., Goto, S., Inaguma, Y. and Ohshima, K. (1991). Immunoreactive A crystallin in rat non-lenticular tissues detected with a sensitive immunoassay method. Biochim. Biophys. Acta , 1080 , 173–180. Klemenz, R., Andres, A.-C., Fröhli, E., Schafer, R. and Aoyama, A. (1993). Expression of the murine small heat shock proteins hsp25 and B crystallin in the absence of stress. J. CellBiol. , 120 , 639–645. Kothary, R., Perry, M.D., Moran, L.A. and Rossant, J. (1987). Cell-lineage specific expression of the mouse hsp68 gene during embryogenesis. Dev. Biol. , 121 , 342–348. Krawczyk, Z., Mali, P. and Parvinen, M. (1988). Expression of a testis-specific hsp70 gene-related RNA in defined stages of rat seminiferous epithelium. J. Cell Biol. , 107 , 1317–1323. Kroeger, P.E., Sarge, K.D. and Morimoto, R.I. (1993). Mouse heat shock transcription factors 1 and 2 prefer a trimeric binding site but interact differently with the HSP70 heat shock element. Mol. Cell . Biol., 13 , 3370–3383. Kroeger, P.E. and Morimoto, R.I. (1994). Selection of new HSF1 and HSF2 DNAbinding sites reveals difference in trimer cooperativity. Mol. Cell. Biol. , 14 , 7592–
Molecular chaperones and folding catalysts
192
7603. Landsberger, N., Ranjan, M., Almouzni, G., Stump, D. and Wolffe, A.P. (1995a). The heat shock response in Xenopus oocytes, embryos and somatic cells: a regulatory role for chromatin. Dev. Biol. , 170 , 62–74. Landsberger, N. and Wolffe, A.P. (1995b). Role of chromatin and Xenopus laevis heat shock transcription factor in regulation of transcription from the X. laevis hsp70 promoter in vivo. Mol. Cell. Biol. , 15 , 6013–6024. Lanks, K.W. (1986). Modulators of the Eukaryotic heat shock response. Exp. Cell Res. , 165 , 1–10. Lassar, A.B., Davis, R.L., Wright, W.E., Kadesh, T., Murre, C., Voronova, A., Baltimore, D. and Weintraub, H. (1991). Functional activity of myogenic HLH proteins requires hetero-oligomerization with E12/ E47-like proteins in vivo. Cell , 66 , 305–315. Lavoie, J.N., Hickey, E., Weber, L.A. and Landry, J. (1993). Modulation of actin microfilament dynamics and fluid phase pinocytosis by phosphorylation of heat shock protein-27. J. Biol. Chem. , 268 , 24210–24214. Leppä, S., Pirkalla, L., Saarento, H., Sarge, K.D. and Sistonen, L. (1997). Overexpresion of HSF2-β inhibits hemin-induced heat shock gene expression and erythroid differentiation in K562 cells. J. Biol. Chem. , 272 , 15293–15298. Loones, M.-T., Rallu, M., Mezger, V. and Morange, M. (1997). HSP gene expression and HSF2 in mouse development. Cell. Mol. Life Sci. , 53 , 179–190. Manejwala, F.M., Logan, C.Y. and Schultz, R.M. (1991). Regulation of hsp70 mRNA levels during oocyte maturation and zygotic gene activation in the mouse. Dev. Biol. , 144 , 301–308. Matsumoto, M. and Fujimoto, H. (1990). Cloning of a hsp70-related gene expressed in mouse spermatids. Biochem. Biophys. Res. Comm. , 166 , 43–49. Matsumoto, M., Kurata, S., Fujimoto, H. and Hoshi, M. (1993). Haploid specific activation of protamine 1 and hsc70t genes in mouse spermatogenesis. Biochim. Biophys. Acta , 1174 , 274–278. McGuire, J., Coumailleau, P., Whitelaw, M.L., Gustafsson, J.A. and Poellinger, L. (1995). The basic helix-loop-helix/PAS factor Sim is associated with hsp90. Implications for regulation by interaction with partner factors. J. Biol. Chem. , 270 , 31353–31357. Mehlen, P., Kretz-Remy, C., Préville, X. and Arrigo, A.-P. (1996a). Human hsp27, Drosophila hsp27 and human aB-crystallin expression-mediated increase in glutathione is essential for the protective activity of these proteins against TNFαinduced cell death. The EMBO J. , 15 , 2695–2706. Mehlen, P., Schulze-Osthoff, K. and Arrigo, A.-P. (1996b). Small stress proteins as novel regulators of apoptosis. J. Biol. Chem. , 271 , 16510–16514. Mehlen, P., Mehlen, A., Godet, J. and Arrigo, A.-P. (1997). Hsp27 as a switch between differentiation and apoptosis in murine embryonic stem cells. J. Biol. Chem. , 272 , 31657–31665. Mezger, V., Legagneux, V., Babinet, C., Morange, M. and Bensaude, O. (1991). Heat shock protein synthesis in preimplantation mouse embryos and embryonal carcinoma cells In L.Hightower and L.Nover (eds.). Heat Shock and Development , Results and Problems in Cell Differentiation 17 , Springer-Verlag, Berlin, pp. 153–166. Mezger, V., Renard, J.-P, Christians, E. and Morange, M. (1994a). Detection of heat shock element-binding activities by gel shift assay during mouse preimplantation development. Dev.Biol. , 165 , 627–638. Mezger, V., Rallu, M., Morimoto, R.I., Morange, M. and Renard, J.-P. (1994b). Heat
Functions in development
193
shock factor 2-like activity in mouse blastocysts. Dev. Biol. , 166 , 819–822. Michaud, S., Marin, R. and Tanguay, R.M. (1997). Regulation of heat shock gene induction and expression during Drosophila development. Cell. Mol. Life Sci. , 53 , 104–113. Milarski, K.L. and Morimoto, R.I. (1986). Expression of human HSP70 during the synthetic phase of the cell cycle. Proc. Natl. Acad. Sci. USA , 83 , 9517–9521. Minowada, G. and Welch, W. (1995). Variation in the expression and/or phosphorylation of the human low molecular weight stress protein during in vitro cell differentiation. J. Biol. Chem. , 270 , 7047–7054. Morange, M., Diu, A., Bensaude, O. and Babinet, C. (1984). Altered expression of heat shock proteins in embryonal carcinoma and mouse early embryonic cells. Mol. Cell. Biol. , 4 , 730–735. Morange, M. ed. (1997). Developmental control of heat shock and chaperone gene expression. Cell. Mol. Life Sci. , 53 , 78–129 & 168–213. Murashov, A.K. and Wolgemuth, D.J. (1996). Distinct transcripts are recognized by sense and antisense riboprobes for a member of the murine HSP70 gene family, HSP70.2, in various reproductive tissues. Mol. Reprod. Dev. , 43 , 17–24. Murashov, A.K. and Wolgemuth, D.J. (1996). Sense and antisense transcripts of the developmentally regulated murine hsp70.2 gene are expressed in distinct and only partially overlapping areas in the adult brain. Mol Brain Res. , 37 , 85–95. Murphy, S.P., Gorzowski, J.J., Sarge, K.D. and Phillips, B. (1994). Characterization of constitutive HSF2 DNA-binding activity in mouse Embryonal Carcinoma Cells. Mol Cell Biol. , 14 , 5309–5317. Nakai, A. and Morimoto, R.I. (1993). Characterization of a novel chicken heat shock transcription factor, heat shock factor 3, suggests a new regulatory pathway. Mol Cell Biol. , 13 , 1983–1997. Nakai, A., Kawazoe, Y., Tanabe, M., Nagata, K. and Morimoto, R.I. (1995). The DNAbinding properties of two heat shock factors, HSF1 and HSF3, are induced in the avian erythroblast cell line HD6. Mol Cell. Biol. , 15 , 5268–5278. Nakai, A, Tanabe, M., Kawazoe, Y., Inazawa, J., Morimoto, R.I. and Nagata, K. (1997). HSF4, a new member of the human heat shock factor family which lacks properties of a transcriptional activator. Mol Cell Biol. , 17 , 469–481. Nover, L., Scharf, K.D., Gagliardi, D., Vergne, P., Czarnecka-Verner, E. and Gurley, W.B. (1996). The Hsf world: classification and properties of plant heat stress transcription factors. Cell Stress Chap. , 1 , 215–223. Nowak, T.S. Jr., Suga, S. and Saito, N. (1994). The heat shock response and gene expression in brain after ischaemia In J.Mayer and I.Brown (eds.). Heat Shock Proteins in the Nervous System , Academic Press, London, pp. 55–81. Ohan, N.W. and Heikkila, J.J. (1995). Involvement of differential gene expression and messenger RNA stability in the developmental regulation of the Hsp30 gene family in heat shocked Xenopus laevis embryos. Dev. Genet. , 17 , 176–184. Parsell, D.A. and Lindquist, S. (1993). The function of heat-shock proteins in stress tolerance: degradation and reactivation of damaged proteins. Annu. Rev. Genet. , 27 , 437–496. Raab, L.S., Polakoski, K.L., Hancock, L.W. and Hamilton, D.W. (1995). Characterization of the heat shock protein P70 in rat spermatogenic cells. Mol. Reprod. Dev. , 40 , 186–195. Rallu, M., Loones, M.-T., Lallemand, Y., Morimoto, R.I., Morange, M. and Mezger, V. (1997). Function and regulation of heat shock factor 2 during mouse embryogenesis.
Molecular chaperones and folding catalysts
194
Proc. Natl. Acad. Sci. USA , 94 , 2392–2397. Rosario, M.O., Perkins, S.L., O’Brien, D.A., Allen, R.L. and Eddy, E.M. (1992). Identification of the gene for the developmentally expressed 70 kDa heat-shock protein (P70). of mouse spermatogenic cells. Dev. Biol. , 150 , 1–11. Sarge, K.D., Zimarino, V., Holm, K., Wu, C. and Morimoto, R.I. (1991). Cloning and characterization of two mouse heat shock factors with distinct inducible and constitutive DNA-binding ability. Genes Dev. , 5 , 1902–1911. Sarge, K.D., Murphy, S.P and Morimoto, R.I. (1993). Activation of heat shock gene transcription by heat shock factor 1 involves oligomerization, acquisition of DNAbinding activity, and nuclear localization and can occur in the absence of stress. Mol Cell. Biol. , 13 , 1392–1407. Sarge, K.D., Park-Sarge, O.-K., Kirby, J.D., Mayo, K.E. and Morimoto, R.I. (1994). Expression of heat shock factor 2 in mouse testis: potential role as a regulator of heatshock protein gene expression during spermatogenesis. Biology of Reproduction , 50 , 1334–1343. Sarge, K.D. (1995). Male germ cell-specific alteration in temperature set point of the cellular stress response. J. Biol Chem. , 270 , 18745–18748. Sass, J.B., Weinberg, E.S. and Krone, P.H. (1996). Specific localization of zebrafish hsp90a mRNA to myoD-expressing cells suggests a role for hsp90a during normal muscle development. Mech. Dev. , 54 , 195–204. Sass, J.B. and Krone, P.H. (1997). HSP90 gene expression may be a conserved feature of vertebrate somitogenesis. Exp. Cell. Res. , 233 , 391–394. Sax, C.M. and Piatigorsky, J. (1994). Expression of a-crystallin/small heat shock protein/molecular chaperone genes in the lens and other tissues. Adv. Enzymol. Relat. Areas Mol. Biol. , 69 , 155–201. Schultz, R.M. (1993). Regulation of zygotic genome activation in mouse. Bioessays , 15 , 531–538. Shaknovich, R., Shue, G. and Kohtz, D.S. (1992). Conformational activation of a basic helix-loop-helix protein (myoD1) by the C-terminal region of murine hsp90 (hsp84). Mol. Cell. Biol. , 12 , 5059–5068. Shakoori, A.R., Oberdorf, A.M., Owen, T.A., Weber, L.A., Hickey, E., Stein, J.L., Lian, J.B. and Stein, G.S. (1992). Expression of heat shock genes during differentiation of mammalian osteoblasts and promyelocytic leukemia cells. J. Cell. Biochem. , 48 , 277– 287. Shue, G. and Kohtz, D.S. (1994). Structural and functional aspects of basic helix-loophelix protein folding by heat-shock protein 90. J. Biol. Chem. , 269 , 2707–2711. Singh, M.S. and Yu, J. (1984). Accumulation of a heat shock-like protein during differentiation of human erythroid cell line K562. Nature , 309 , 631–633. Sistonen, L., Sarge, K.D., Phillips, B., Abravaya, K. and Morimoto, R.I. (1992). Activation of heat shock factor 2 during hemin-induced differentiation of human erythroleukemia cells. Mol. Cell. Biol. , 12 , 4104–4111. Sistonen, L., Sarge, K.D. and Morimoto, R.I. (1994). Human heat shock factors 1 and 2 are differentially activated and can synergistically induce hsp70 gene transcription. Mol. Cell. Biol. , 14 , 2087–2099. Stringham, E.G., Dixon, D.K., Jones, D. and Candido, E.P.M. (1992). Temporal and spatial expression patterns of the small heat shock (hsp16). genes in transgenic Caenorhabditis elegans. Mol. Biol. Cell , 3 , 221–233. Tanguay, R.M. (1989). Localized expression of a small heat shock protein, hsp23, in specific cells of the central nervous system during early embryogenesis in Drosophila.
Functions in development
195
J. Cell Biol. , 109 , 155. Tanguay, R.M., Wu, Y. and Khandjian, E.W. (1993). Tissue-specific expression of heat shock proteins of the mouse in the absence of stress. Dev. Genet. , 14 , 112–118. Theodorakis, N.G., Zand, D.J., Kotzbauer, P.T., Williams, G.T. and Morimoto, R.I. (1989). Hemin-induced transcriptional activation of the HSP70 gene during erythroid maturation in K562 cells is due to a heat shock factor-mediated stress response. Mol. Cell Biol. , 9 , 3166–3173. Vierling, E. (1991). The roles of heat shock proteins in plants. Annu. Rev. Plant Physiol . Plant Mol. Biol. , 42 , 579–620. Walsh, D., Li, K., Crowther, C., Marsh, D. and Edwards, M. (1991). Thermotolerance and heat shock response during early development of the mammalian embryo In L.High tower and L.Nover (eds.). Heat Shock and Development , Results and problems in Cell Differentiation, 17, Springer-Verlag, Berlin. Walsh, D., Li, Z., Wu, Y. and Nagata, K. (1997). Heat shock and the role of the HSPs during neural plate induction in early mammalian CNS and brain development. Cell. Mol. Life Sci. , 53 , 198–211. Wehmeyer, N., Hernandez, L.D., Finkelstein, R.R. and Vierling, E. (1996). Synthesis of small heat shock proteins is part of the developmental program of late seed maturation. Plant Physiol. , 112 , 747–757. Welte, M.A., Duncan, I. and Lindquist, S. (1995). The basis for a heat-induced developmental defect: defining crucial lesions. Genes Dev. , 9 , 2240–2250. Wisniewski, J., Malezewski, M., Krawczyk, Z. and Gedamu, L. (1993). An upstream region of the rat spermatogenesis-specific heat-shock-like Hst70 gene confers testisspecific expression in transgenic mice. Eur. J. Biochem. , 212 , 137–143. Wu, B.J. and Morimoto, R.I. (1985). Transcription of the human hsp70 gene is induced by serum stimulation. Proc. Natl. Acad. Sci. USA , 82 , 6070–6074. Zakeri, Z.F., Wolgemuth, D.J. and Hunt, C.R. (1988). Identification and sequence analysis of a new member of the mouse HSP70 gene family and characterization of its unique cellular and developmental pattern of expression in the male germ line . Mol. Cell. Biol. , 8 , 2925–2932. Zhu, D., Dix, D.J. and Eddy, E.M. (1997). HSP70–2 is required for CDC2 kinase activity in meiosis I of mouse spermatocytes. Development , 124 , 3007–3014.
9. EARLY EVENTS IN THE SYNTHESIS AND MATURATION OF POLYPEPTIDES WILLIAM J.WELCH*, DARYL K.EGGERS, WILLIAM J.HANSEN and HIRSOHI NAGATA Departments of Medicine and Physiology, University of California, San Francisco, CA 94143, USA
1. Introduction 2. Assisted Protein Folding Helps to Overcome Problems Confronting Nascent Polypeptides 3. Experimental Approaches for the Identification of Nascent Chain Binding Components 4. Components Dedicated to Nascent Chain Interactions: Organellar Targeting Factors 4.1. Targeting Factors in Bacteria 4.2. Targeting Factors in Yeast and Animal Cells 4.2.1. Targeting to the ER 4.2.2. Targeting to the Mitochondria 4.2.3. Targeting to the Peroxisome 4.3. Relative Roles of Dedicated vs. Generic Chaperones in Protein Targeting 5. Pathways and Components Involved in the Maturation of Cytosolic Proteins 5.1. Folding in the Eukaryotic Cytosol 5.2. Folding in the Prokaryotic cytoplasm 5.3. Summarizing the results obtained from Eukaryotic and Prokaryotic Studies 6. Unresolved Questions, Controversial Issues, Future Directions 7. Acknowledgements 8. References *Corresponding author
Early events in the synthesis and maturation of polypeptides
197
1. INTRODUCTION Events involved in the earliest stages of protein synthesis and maturation only now are beginning to be understood. While considerable progress has been realized in the identification and characterization of the different machineries involved in protein synthesis (e.g. large and small ribosomal subunits, initiation, elongation and release factors), relatively little is known about the nature of a polypeptide while it is being synthesized on the ribosome. Whether folding of a nascent polypeptide occurs cotranslationally or post-translationally (or both), and whether other factors are involved in the folding process are questions that now are beginning to be addressed by different laboratories. Here, we review our current understanding of nascent chain biogenesis in the cytoplasm of both prokaryotes and eukaryotes. We begin by discussing some of the potential problems a nascent polypeptide might encounter during the different stages of its maturation. Next a description of cytosolic components which appear dedicated to interacting with nascent polypeptides destined for transport across cellular membranes will be presented. Finally, we review our current understanding of the maturation pathway of cytoplasmic proteins. Here particular attention will be paid to the description of cellular components which interact with nascent polypeptides and how these components influence the folding pathway.
2. ASSISTED PROTEIN FOLDING HELPS TO OVERCOME PROBLEMS CONFRONTING NASCENT POLYPEPTIDES Progress in understanding the steps in protein folding has been facilitated by the use of so-called Anfinsen-like protein folding reactions (Anfinsen, 1973). Here a purified protein first is denatured via its exposure to a protein chaotrope. Upon subsequent dilution of the chaotrope, the protein undergoes spontaneous refolding back into its native conformation. Both the rate and the efficiency of the re-folding reaction can vary significantly depending upon the size and composition of the polypeptide being studied. Nevertheless, from these types of experiments some of the rate-limiting steps in the overall pathway of protein folding have been realized. Extrapolating the results of in vitro folding reactions to protein folding inside the cell however, has limitations. First, the folding of a newly synthesized protein in vivo has obvious restrictions since polypeptide synthesis and/or translocation into organelles occurs as a vectorial process. In an Anfinsen folding reaction, all of the information for protein folding (i.e. the entire amino acid sequence) is already present at the time the protein folding event is initiated. In the cell however, synthesis or translocation into an organelle occurs as a sequence of vectorial steps; i.e. synthesis and translocation proceed from the amino- to the carboxyterminus of the polypeptide. Considering that individual protein domains usually are comprised of at least 100 or more amino acids (Berman, 1994), one wonders what the nascent chain looks like when there are not enough amino acids synthesized to allow for a productive folding reaction? Is the nascent chain undergoing different folding transitions, or might it be stabilized via an interaction with one or more molecular
Molecular chaperones and folding catalysts
198
chaperones? Another problem not addressed by the Anfinsen model of protein folding concerns the issue of molecular crowding inside the cell. In vitro folding reactions typically employ a single polypeptide species under relatively dilute conditions. Upon removal of the chaotrope the protein is free to initiate folding without the nuances of encountering a significantly different macromolecule. Even here however, various intermediates in the folding pathway of a single species of polypeptide potentially can result in misfolding and/or aggregation. In the cell, where many different species of polypeptides are being synthesized (or translocated into organelles) at any one time, the possibility of a particular nascent chain inadvert-ently interacting with another nascent polypeptide(s) or with another macromolecule in its local vicinity seems quite high. Again such inadvertent interactions seemingly would lead to errors in the folding pathway and the generation of protein aggregates. Finally, whether the kinetics of protein folding observed in vitro are at all relevant to the situation in vivo is not known and obviously is a difficult question to test. The rate of protein folding can be faster than the rate of protein synthesis. Therefore, it seems plausible that folding of the nascent chain can occur co-translationally. For example, in the case of multi-domain proteins, folding of the most amino terminal domain might commence once a suitable number of amino acids (e.g. 100–200 amino acids) have been synthesized and extruded out of the immediate vicinity of the ribosomal subunits. In the case of single domain proteins, one suspects that the final folding of the polypeptide likely will occur only after release of the polypeptide from the ribosome. For both singleand multi-domain proteins however, we suspect that the very earliest events of their biogenesis will be similar; both will require stabilization of their amino terminus until enough amino acids have been synthesized to initiate a productive folding reaction. Despite these caveats, the fidelity of protein folding reactions, either in the cell or using in vitro translation systems, appears to be quite high. As we discuss below, the high efficiency of protein folding in vivo likely is due to a group of macromolecules whose primary task is to monitor the very earliest stages of protein maturation (summarized in Tables 1 and 2). Some of these components appear dedicated to interacting with nascent chains which are to be targeted to specific organelles or transported out of the cell. The dedicated factors recognize their nascent chain targets by virtue of particular amino acid sequences present within the polypeptide chain. Yet other nascent chain binding proteins appear to bind to their targets in a more generic-like fashion, presumably via their ability to recognize “non-native” or thermodynamically unstable structures. Members of this family, most of which are known molecular chaperones, likely function in all aspects of protein maturation, be it folding, transport across membranes, or higher-ordered assembly.
3. EXPERIMENTAL APPROACHES FOR THE IDENTIFICATION OF NASCENT CHAIN BINDING COMPONENTS A number of experimental approaches have been employed to examine both the status of a nascent polypeptide and to identify cellular components which participate in its
Early events in the synthesis and maturation of polypeptides
199
synthesis and folding. When we refer to a nascent polypeptide, we mean one that is still under synthesis and, therefore, attached to the ribosome. When referring to a newly synthesized protein, we mean one whose synthesis has been completed and which has been released from the ribosome (but has not necessarily acquired its final folded state). Perhaps our first realization that the cell might contain factors needed to interact with nascent or newly synthesized proteins came from studies examining the mechanisms of polypeptide transport across intracellular membranes (see chapters of this volume by Haas and Zimmernann, Dekker
Table 1 Components interacting with nascent chains and/or newly synthesized proteins in prokaryotes
Factor
Composition
Comments
SecB
homo-tetramer (4×16 kDa) DEDICATED FACTOR: High affinity for signal sequences of precursor proteins, slightly lower affinity for unfolded proteins; Substrates apparently distinct from those of SRP; Null mutants viable only under “slow-growth conditions;” No apparent nucleotide dependence for substrate interaction/release.
Signal Recognition Particle (SRP)
Ribonucleoprotein consisting of a single 48 kDa protein and a single 4.5S RNA
DEDICATED FACTOR: High affinity for signal sequences of precursor proteins; Substrates appear to be distinct from those of SecB; genes encoding protein or RNA are essential in E. coli; Utilizes GTP for both signal sequence binding and membrane receptor docking.
Trigger Factor (TF)
48 kDa monomer
GENERIC FACTOR (?): Originally suggested to be dedicated factor for binding signal sequences of precursor proteins; Recently suggested to be a general nascent chain binding component; Exhibits peptidyl cis/trans prolyl isomerase activity; No reports of nucleotide dependence; TF null mutants exhibit slight defects in cell division, but do not exhibit secretion defects, nor large scale protein folding abnormalities.
DnaK Chaperone Machine
DnaK (68 kDa) DnaJ (40 kDa) GrpE (20 kDa)
GENERIC FACTOR: DnaK, and possibly DnaJ, capable of recognizing non-native polypeptides or small peptides; Only a few reports of interactions with nascent and/or newly synthesized proteins; Overexpression helps to rescue secretory defects in SecB null mutants; DnaK, DnaJ, or GrpE null mutants all exhibit similar pleiotropic effects and are heat sensitive; DnaK-substrate interactions/release require ATP/ADP.
Molecular chaperones and folding catalysts GroEL/ES Chaperone Machine
GroEL: homo-oligomer of 60 kDa subunits, consists of two-7 membered rings. GroES: homo-oligomer of 10 kDa subunits, consists of a single 7-membered ring.
200
GENERIC FACTOR: Important for the folding of some (but probably not all) newly synthesized proteins. No evidence for direct interaction with nascent polypeptides. Overexpression helps to rescue some secretory defects observed in SecB null mutants, and to generally suppress temp.-sensitive and missence mutations; Deletion of groE operon is lethal at all growth temperatures; Conditional mutants have pleiotropic effects including defective protein folding and accumulation of protein aggregates; Utilizes ATP in reaction cycle.
Table 2 Components interacting with nascent chains and/or newly synthesized proteins in eukaryotic cytosol
Factor
Composition
Comments
Signal Recognition Particle (SRP)
Ribonucleoprotein consisting of 6 polypeptides and 1 species of 7S RNA
DEDICATED FACTOR: For targeting precursor proteins (co-translationally) to the endoplasmic reticulum (ER); High affinity for both ER targeting signal sequences and receptor in ER membrane; Facilitates movement of polysomes to ER membrane and initiates formation of translocation channel; Genes encoding protein or RNA are essential for growth (except in S. cerevisiae); Utilizes GTP in reaction cycle.
Nascent Heterodimer ( subunit 33 kDa) Polypeptide Associated ( subunit 21 kDa) Complex (NAC)
GENERIC FACTOR (?): Proposed to be the first cytosolic component interacting with nascent polypeptides; Has lower affinity for ER targeting signal sequences than does SRP; Bound to ribosome and may prevent inappropriate association of ribosome to ER membrane; Interaction with nascent chain likely is transient; No report of nucleotide dependence for substrate interaction/release.
Targeting Factor 28 kDa
DEDICATED FACTOR(?): Isolated from rabbit reticulocyte lysate; Binds some mitochondrial signal sequences; Stimulation of mitochondrial precursor import in vitro; No reports on structure or function since 1990.
PresequenceBinding Factor (PBF)
50 kDa (homo-dimer?) DEDICATED FACTOR (?): Isolated from rabbit reticulocyte lysate; Binds mitochondrial signal sequences; Activity insensitive to NEM and independent of ATP; May work together with cytosolic hsp70; No reports on structure or function
Early events in the synthesis and maturation of polypeptides
201
since 1992. Mitochondrial Import Stimulation Factor (MSF)
Hetero-dimer (32 kDa and 30 kDa subunits)
DEDICATED FACTOR: Isolated from rat liver cytosol via affinity for mitochondrial signal sequences; May also function to solubilize precursor protein aggregates in an ATP-dependent manner; Mitochondrial import stimulation activity and ATPdependent protein unfolding activity likely facilitated by different subunits.
Hsp70 Chaperone Machine
Hsp70 Hsp40
GENERIC FACTOR: Hsp70 interacts with nascent chains and newly synthesized proteins: One report of Hsp40 also interacting with nascent chain; Hsp70 copurifies with polysomes, but is released upon puromycin treatment; Yeast depleted of hsp70 grow poorly, are hypersensitive to drugs which cause misreading of mRNA, and contain relatively few polysomes; Hsp70 interaction/release from substrates requires ATP/ADP.
Chaperonin Hetero-oligomer: Ring- GENERIC OR DEDICATED (?): Subunits exhibit Containing TCP- like particles, each ring weak homology to other chaperonins; No apparent 1 (CCT) consisting of 8 or 9 50– GroES-like co-factor; Heterologous composition 60 kDa subunits unlike other chaperonins; Levels low relative to other chaperonins; Unlike other chaperonins, not a heat shock protein; Conditional mutants of TCP-1 exhibit cytoskeletal and cell division abnormalities; Important for the folding of newly synthesized actin and tubulin; Other substrates in vivo? OTHERS hsp90, p48/Hip, cis/trans prolyl isomerases
GENERIC(?): Hsp90 involved in the maturation/regulation of some steroid hormone receptors and protein kinases; Hip reported to interact with hsp70 and participate with hsp70/hsp40 innascent chain interactions; A number of proteins exhibiting cis/trans prolyl isomerase activities may interact with nascent/newly synthesized proteins.
and Pfanner, Muckel and Soll). Here the development of in vitro translation approaches coupled with biochemical fractionation led to the identification and characterization of the eukaryotic signal recognition particle (SRP) (Walter and Blobel, 1980). SRP was purified to homogeneity and was shown to interact specifically with so-called “signal sequences” present within the nascent polypeptide chain, as well as with components present in the membrane of the endoplasmic reticulum. Depletion/reconstitution experiments provided proof of an essential role for SRP in protein translocation events (reviewed by Walter and Johnson, 1994). Over the past 10 years genetic approaches also have led to the further identification and characterization of components involved in protein biogenesis. Deletion of specific
Molecular chaperones and folding catalysts
202
genes, the development of temperature sensitive mutants, and the generation of suppressor mutations, all have made a tremendous impact on the identification and characterization of components involved in protein maturation. Significant advances in the development of in vitro systems established to mimic the in vivo situation of protein synthesis, protein folding, and the translocation of proteins into subcellular organelles have been achieved. These in vitro protein synthesis systems combined with different cross-linking methods have led to the identification of macromolecules which make direct contact with nascent chains, either during their synthesis or translocation into organelles. A particularly clever technique being employed recently by some labs to characterize nascent chains involves the use of “truncated mRNA’s” (reviewed by Gilmore et al., 1991). Here mRNAs are created that no longer contain their normal translational stop codon. When used to program an in vitro translation system, the ribosome initiates translation of the polypeptide and moves toward the 3' end of the message, but is unable to release the nascent chain owing to the lack of a stop codon. This technique allows one to effectively “freeze in time” a relatively homogenous population of a particular nascent chain, and thereby examine its physical properties as well as identify components which interact directly with the nascent polypeptide. Moreover, release of the nascent chain can be accomplished by the addition of the antibiotic puromycin, thereby allowing for the separation of the nascent chain from the ribosomal subunits. An obvious caveat of this approach is the possibility that the puromycin-released chain now might bind to components (including molecular chaperones) which do not normally interact with the ribosome-bound nascent polypeptide. In addition, those components normally bound to the ribosome-bound nascent chain conceivably might be lost upon the premature release of the nascent chain from the polysome. Indeed, even without the use of puromycin, nascent chains can release spontaneously from the ribosome (Gilmore et al., 1991, and WJH, unpublished observations).
4. COMPONENTS DEDICATED TO NASCENT CHAIN INTERACTIONS: ORGANELLAR TARGETING FACTORS Much of what we know regarding nascent chain binding components has come from those studies examining how proteins made in the cytoplasm are transferred across different membranes. A number of different pathways have been characterized in both prokaryotes and eukaryotes which function to deliver proteins to their final destination. These delivery systems transfer proteins either during their synthesis (i.e. cotranslationally) or shortly following their synthesis (post-translationally). Therefore, cytosolic components must exist which are capable of scanning newly synthesized proteins and identifying those which are to be transferred out of the cytoplasm into the periplasmic space (prokaryotes), or transferred into different subcellular organelles (eukaryotes). Rather than exhaustively surveying this vast field of research, we will try and highlight the most important concepts and discuss their relevance as it pertains to protein folding events in general. Most proteins which are to be transferred across a membrane initially are synthesized
Early events in the synthesis and maturation of polypeptides
203
with a short sequence of amino acids which usually is not part of the mature polypeptide (Blobel, 1980). These sequences, referred to as “signal or leader sequences”, initially were thought to serve only as a type of “post-office address”, dictating where in the cell the newly synthesized protein was to be delivered (reviewed in Hartl and Neupert, 1990; Verner and Schatz, 1988; Walter and Lingappa, 1986; Schatz and Dobberstein, 1996). The fact that the signal sequence was not found in the mature form of the protein indicated that it was not important in the final structure of the polypeptide. In the case of prokaryotic proteins destined for export into the periplasmic space, or eukaryotic proteins to be transferred into the lumen of the ER, the signal sequence usually is situated at the extreme aminoterminus and is rich in hydrophobic amino acids. Both the lack of a defined amino acid sequence and the overall hydrophobic character of the signal peptide led many scientists to conclude that the signal sequence might actually insert itself directly into the lipid bilayer and thereby function to initiate the translocation process (von Heijne and Blomberg, 1979; Wickner, 1979; Gierasch, 1989). While some signal sequences do appear to insert themselves into lipid bilayers in vitro, it is unlikely that such a scenario occurs in vivo. As is described in detail elsewhere in his volume (chapters by Dekker and Pfanner, Muckel and Soll), translocation of polypeptides across membranes usually occurs through an aqueous channel formed by a distinct set of polypeptides unique to the particular membrane (Crowley et al., 1993; Gilmore and Blobel, 1985; reviewed by Rapoport et al., 1996; Schatz and Dobberstein, 1996). Signal sequences present within proteins destined for the mitochondria are rich in charged amino acids, these obviously not being amenable for intercalation into and across a lipid bilayer. In addition to specifying where a polypeptide is to be delivered, another important role played by the signal sequence has been realized in recent years. This additional function of the signal sequence, we believe, illustrates a very important concept as it relates to understanding the earliest stages of protein folding in the cell. Specifically, the presence of a signal sequence, in some cases, can slow down or perhaps prevent the overall folding of the precursor form of the polypeptide (Park et al., 1988). Slowing down protein folding early during synthesis might help to prevent inappropriate folding reactions which could prevent the nascent chain from reaching its final native conformation. As is discussed later, we suggest that a similar function may be carried out by one or more of those components which interact with nascent polypeptides. 4.1. Targeting Factors in Bacteria Different macromolecules involved in the transfer of newly synthesized proteins across the inner membrane of gram negative bacteria have been described (reviewed by Pugsley, 1993). Two bacterial proteins were suggested some time ago to preferentially interact with newly synthesized, signal-sequence containing polypeptides which are transported into, or across the bacterial inner membrane. One of these, termed trigger factor (TF), was implicated in protein translocation via its interaction with a purified outer membrane protein still containing its signal sequence, proOmpA. Using an in vitro system, denatured pro-OmpA could be translocated into isolated membrane vesicles as a function of added TF (Crooke et al., 1988; Crooke and Wickner, 1987). Based on these observations, it was suggested that TF might interact with newly synthesized proteins
Molecular chaperones and folding catalysts
204
containing a signal sequence and thereby facilitate their presentation to the translocation machinery at the membrane. Subsequent genetic studies however, revealed that depletion of TF did not lead to any obvious pleiotropic effects in protein secretion (Guthrie and Wickner, 1990). Thus, the idea that TF might be a dedicated component for protein translocation events in bacteria fell out of favor. As will be discussed later, very recent studies have led to the idea that TF may interact with nascent polypeptides in general, a suggestion which has revived interest into the biology of TF. A second bacterial component known to participate in at least some protein translocation events is the SecB protein (reviewed by Randall and Hardy, 1995). SecB represents one of a number of gene products which have been shown to play important roles in bacterial protein secretion. In vivo, SecB has been observed to interact with a limited number of presecretory proteins, presumably escorting such proteins to the SecA protein in a relatively unfolded and translocation competent state. Interaction with the SecA protein is believed to represent the first stage in the movement of the presecretory protein into the membrane translocation channel. SecB is a homo-tetramer of 16 kDa subunits and exhibits interactions with a variety of unfolded polypeptides. Like other members of the molecular chaperone family, SecB appears capable of recognizing proteins in their non-native state with no strict requirement for any particular sequence of amino acids within the polypeptide target (Randall, 1992). Indeed some investigators have suggested that SecB recognizes unfolded polypeptides in general, while others believe SecB recognizes primarily signal sequences of presecretory proteins (Randall et al., 1990; Watanabe and Blobel, 1995). While it seems clear that SecB, at least in vitro, is capable of binding to signal sequences, as well as regions present within the mature form of a polypeptide, all of the available evidence indicate that in vivo it is dedicated to proteins destined for export. For example, SecB null mutants exhibit a reduced efficiency in the export of some, but not all presecretory proteins (Kumamoto and Beckwith, 1983). No other obvious abnormalities (which one might expect upon deleting a gene encoding for one of the more generic chaperones) were observed. At the biochemical level, immunoprecipitation experiments (with and without the use of chemical cross-linking) have identified SecB in complexes with only presecretory proteins (Kumamoto, 1989). Unlike the more classical chaperones (e.g. Hsp70 and chaperonins), SecB does not appear to utilize ATP, either for its binding to, or release from protein substrates. Thus, SecB is thought to be involved primarily in facilitating the targeting of a subset of presecretory proteins to the bacterial translocation channel. Whether SecB interacts with its substrates co-translationally or posttranslationally has not been firmly established in all cases. Another component interacting with newly synthesized polypeptides destined for transport across the bacterial membrane is the prokaryotic homologue of signal recognition particle (SRP) (reviewed by Luirink and Dobberstein, 1994). In contrast to the complexity of its animal cell counterpart (discussed later), bacterial SRP appears to be comprised of only a single polypeptide and one RNA species. The polypeptide, referred to as p48, exhibits sequence homology with the well characterized 54 kDa subunit of eukaryotic SRP (Römisch et al., 1989; Bernstein et al., 1989). Similarly, the 4.5S RNA species of bacterial SRP is homologous to the 7S RNA moiety of eukaryotic SRP (Poritz et al., 1988). Depletion of either component from E. coli adversely affects the export of
Early events in the synthesis and maturation of polypeptides
205
several secretory proteins (Phillips and Silhavy, 1992). Interestingly, those proteins most dependent on SRP for efficient export are the same ones that are largely unaffected when a functional SecB product is missing. While this observation emphasizes the fact that more than one pathway exists for protein translocation events in bacteria, the reason why different pathways are used to handle different proteins is not clear. Chemical crosslinking studies have shown that bacterial SRP engages with the signal sequence in a cotranslational manner (Luirink et al., 1992). Thus, similar to the situation in animal cells, bacterial SRP interacts with pre-secretory proteins co-translationally and facilitates their entry into the translocation channel. Support for a role of the more classical chaperones, both the DnaK machine (DnaK/DnaJ/GrpE) and the GroEL/ES chaperonin, in the export of proteins in bacteria has been demonstrated by a number of laboratories. Firstly, GroEL/ES has been reported to bind to newly synthesized precursor proteins in vitro and -lactamase in vivo (Bochkareva and Girshovich, 1992; Bochkareva et al., 1988; Lecker et al., 1989). Moreover, bacteria carrying mutant forms of both GroEL/ES appear defective for the correct processing of newly synthesized -lactamase (Kusakawa et al., 1989). Secondly, induction of the heat shock response and the corresponding increased synthesis of the bacterial heat shock proteins can compensate for the lack of SecB function (Altman et al., 1991). Overproduction of DnaK or DnaJ by other means also is effective in rescuing the poor growth activity observed for SecB null mutants (Wild et al., 1992). In a similar vein, cells expressing higher than normal levels of GroEL and DnaK are more competent in carrying out the secretion of a fusion protein consisting of -galactosidase containing an amino-terminal signal sequence, a construct which normally tends to jam the translocation apparatus (Phillips and Silhavy, 1990). Finally, bacteria lacking either DnaJ or DnaK exhibit reduced synthesis and/or processing of alkaline phosphatase (but not another SecB independent substrate, ribose-binding protein) (Wild et al., 1992). Hence, these observations argue that the more classical or generic members of the molecular chaperone family can in fact act as a type of “back up” system when bacterial components dedicated specifically to protein translocation are missing, overloaded with different substrates, or presented with substrates that pose a particular problem in their translocation across the membrane. As is discussed below, a similar scenario also appears to be operative in yeast and animal cells. 4.2. Targeting Factors in Yeast and Animal Cells In both yeast and animal cells the problem of protein targeting is further complicated owing to subcellular compartmentalization. In addition to targeting newly synthesized proteins into the secretory pathway, additional machineries are needed to direct proteins from their site of synthesis in the cytosol to the nucleus, mitochondria, and peroxisome. Perhaps not too surprisingly, eukaryotic cells contain a variety of targeting components (summarised in Table 2), some of which appear dedicated to the interaction with precursor proteins, while others appear to represent the more classical members of the molecular chaperone family. Note that in what follows, we have specifically excluded from our discussion the pathway of protein import into the nucleus. Furthermore, other chapters of this volume provide detailed descriptions of the systems translocating
Molecular chaperones and folding catalysts
206
proteins to and folding within the endoplasmic reticulum (ER) (Haas and Zimmermann), mitochondria (Dekker and Pfanner) and chloroplasts (Muckel and Soll). 4.2.1. Targeting to the ER Analogous to secretion events in bacteria, newly synthesized proteins targeted to a subcellular compartment like the endoplasmic reticulum, mitochondria or peroxisome are usually “marked” via a specific sequence of amino acids which oftentimes is removed after the polypeptide has reached its final destination. Signal sequences for targeting to the ER are characteristically rich in hydrophobic amino acids and serve as a recognition motif for the eukaryotic signal recognition particle (reviewed by Walter, 1986; Walter, 1994). SRP in eukaryotes is composed of six polypeptides and a 7S species of RNA. The 54 kDa subunit, homologous to p48 of bacterial SRP, recognizes and binds to the signal sequence present within the nascent chain. The SRP-nascent chain-polysome complex then is directed to the ER membrane where translocation across the membrane occurs cotranslationally. There is evidence that once engaged to the nascent chain, SRP somehow can slow down or even stall the rate of translational elongation, presumably to facilitate the timely movement of the polysome to the ER membrane. Still under intense study, the interaction of the ribosome-SRP complex with the ER membrane somehow initiates the formation of a translocation channel through which the nascent polypeptide is then extruded into the lumen of the ER (reviewed by Rapoport et al., 1996, Schatz and Dobberstein, 1996). In yeast targeting into the ER can occur in the absence of a functional SRP. For example, deletion of genes encoding either the SRP 54 kDa subunit or the 7S RNA moiety in S. cerevisiae does not result in lethality (Hann and Walter, 1991). Although protein translocation events into the ER were impaired overall, the translocation defects varied for different proteins. While some proteins exhibited a greater than 90% block, others showed no detectable translocation defect whatsoever. Interest-ingly, those precursor proteins least affected were shown in other studies to be translocated into the ER either late in their synthesis, or even after their release from the ribosome (Hansen et al., 1986; Hansen and Walter, 1988; Rothblatt et al., 1987; Waters and Blobel, 1986). These observations, in sum, are indicative that translocation into the ER can occur in the absence of SRP, likely in a post-translational fashion. However, none of these studies actually addressed the mechanism (s) by which precursor proteins are transported into the ER in the absence of SRP. A possible clue to the mechanism (s) follows from the results of two other studies. First, depletion of cytosolic Hsp70 levels in yeast via genetic means leads to the accumulation of precursor proteins destined for either the mitochondria and ER (Deshaies et al. 1988). In a complementary study using an in vitro system to study protein translocation events, cytosolic Hsp70, along with a second and still unidentified component, was shown to stimulate the post-translational translocation of proteins into isolated ER-derived microsomes (Chirico et al. 1988). Evidence for a direct interaction of cytosolic Hsp70 with the precursor forms of proteins destined for transport into the ER however, is rather limited. To our knowledge there is only one report of a secretory precursor protein, prepro-α factor, in a complex with cytosolic Hsp70 (Chirico, 1992). Thus, whether the Hsp70 chaperone machinery serves as the operative “back-up” system
Early events in the synthesis and maturation of polypeptides
207
for protein translocation into the ER in those yeast strains genetically depleted of SRP still requires further study. Recently, Wiedmann and colleagues have identified a new component which they suggest might play an important role in both nascent chain biogenesis and the targeting of proteins to the ER. Specifically, a protein complex referred to as “nascent polypeptide associated complex” (NAC) was identified based on its affinity for short nascent polypeptides (as determined by photo-activated cross-linking) (Wiedmann et al., 1994). NAC is composed of 2 polypeptides (a subunit of 33 kDa and subunit of 21 kDa) which together appear to bind to all nascent chains, be it ones which do or do not contain an organellar targeting signal sequence. In addition NAC apparently exhibits an affinity for the ribosome, in particular that site on the ribosome involved with docking at the ER membrane. Based on these and a number of other observations it has been suggested that NAC might act to prevent potentially promiscuous interactions of ribosomes with the membrane of the ER (Wang et al., 1995). For example, either non-translating ribosomes, or those translating polypeptides not containing a signal sequence, might be expected to bind to the ER membrane via the same membrane attachment site used for the docking of ribosomes translating a nascent chain containing an ER targeting signal sequence. Accordingly, NAC interaction with the ribosome, perhaps at the membrane attachment site, would prevent such promiscuous interactions and thereby ensure that proteins lacking a signal sequence are not inadvertently translocated into the ER lumen. Support for this idea followed from studies showing that in the absence of NAC, nascent polypeptides lacking an ER targeting signal sequence could be cross-linked to SRP and subsequently translocated (albeit inefficiently) across the ER membrane (Lauring et al., 1995a; Lauring et al., 1995b). Hence in their model, both SRP and NAC would compete for the binding to nascent polypeptides. In the case of those nascent chains harboring an ER targeting signal sequence SRP would be the one to bind due to its higher affinity for signal sequences. A very recent study however, has concluded that SRP does not bind to nascent chains lacking a signal sequence, but rather that it binds to the ribosome in a saltsensitive manner. Hence, it was suggested that NAC does not influence the binding of SRP to nascent chains carrying signal sequences. Rather, the investigators proposed that NAC somehow might modulate the interaction of SRP with the ribosome itself, perhaps via an interaction with that region of the ribosome involved in the docking to the ER membrane (Powers and Walter, 1996). At the present time, there are no published reports on the consequences of depleting those genes encoding the NAC proteins. Obviously more work is necessary to determine the exact role played by NAC in the early stages of nascent chain biogenesis. 4.2.2. Targeting to the Mitochondria Components dedicated to the targeting of newly synthesized proteins from the cytosol into the mitochondria have been identified only recently (reviewed by Lill et al., 1996; Miharaand Omura, 1996). In contrast to the ER, where the bulk of targeting appears to be dependent on dedicated factors like SRP, targeting to the mitochondria has been thought to be facilitated, in large part, by the cytosolic Hsp70 molecular chaperone machinery. Similar to the situation with targeting to the ER, proteins destined for import into
Molecular chaperones and folding catalysts
208
mitochondria are synthesized initially with a leader peptide usually containing a number of charged amino acids. These targeting sequences often are removed during the import reaction and presumably are not important for the folding of the mature protein once inside the organelle. Two distinct cytosolic factors, referred to as “targeting factor” and “presequence-binding factor” were identified in the early 1990’s based on their ability to bind to mitochondrial signal sequences (Murakami and Mori, 1990; Ono and Tuboi, 1988). Both components appeared to stimulate the import of model proteins into isolated mitochondria. Curiously, however, nothing else has been reported concerning these two factors over the past few years. Another factor dedicated to mitochondrial targeting, and which today is under intense study, is the mitochondrial import stimulation factor (MSF). Composed as a heterodimer of 30 and 32 kDa subunits, MSF stimulates the import into isolated mitochondria of all mitochondrial precursor proteins so far examined (reviewed by Mihara and Omura, 1996). It has been suggested that MSF can unfold mitochondrial destined precursor proteins in an ATP-dependent manner, thereby facilitating the early stages of their translocation into the organelle. Results from a number of experiments performed in vitro indicate that MSF recognizes sequences within the mature portion of the precursor protein, as well as the mitochondrial signal sequence itself. This apparent dual specificity in the interaction of MSF with its protein target appears reminiscent of the situation with the bacterial SecB protein. MSF belongs to the so-called 14–3–3 family of related proteins. Although the significance of such homology is not entirely clear, members of this family are involved in a wide variety of interesting functions including cell-cycle regulation, signal transduction events, and exocytosis (reviewed by Aitken, 1992). As was mentioned above, targeting of many proteins to the mitochondria is thought to depend heavily on the actions of cytosolic Hsp70, along with its co-factors, the family of eukaryotic DnaJ homologues (reviewed by Caplan et al. 1993; Mihara and Omura, 1996). This idea was first suggested based on the earlier mentioned studies where the depletion of cytosolic Hsp70 levels in yeast resulted in the accumulation of the precursor forms of proteins normally destined for import into the mitochondria (Deshaies et al. 1988). Subsequent studies where one of the cytosolic homologues of the DnaJ co-factor (Ydj1) was deleted also resulted in the large scale accumulation of mitochondrial precursor proteins (Atencio and Yaffe, 1992; Caplan et al. 1992). While these observations, all based on genetic approaches, indicate an important role for the Hsp70/DnaJ molecular chaperone machinery in mitochondrial protein import, exactly how these chaperones influence mitochondrial import has not been firmly established. One presumes that the chaperones interact with and stabilize newly synthesized mitochondrial precursor in a relatively unfolded, or translocation competent state. To our knowledge however, only one laboratory has reported an interaction of a mitochondrial precursor protein with cytosolic Hsp70. Specifically, using an in vitro translation system, a portion of newly synthesized aspartate aminotransferase (with its amino-terminal leader sequence still present) was reported to co-precipitate with Hsp70 (Lain et al., 1994; Lain et al., 1995). In a related study, the effects of Hsp70 depletion from a rabbit reticulocyte lysate programmed with mRNA for ornithine transcarbamylase was examined as it related to the import of the protein into isolated mitochondria (Terada et al., 1995). In the absence of the Hsp70 chaperone, translocation of the mitochondrial precursor protein was
Early events in the synthesis and maturation of polypeptides
209
severely inhibited. Interestingly, addition of Hsp70 after translation, but before the addition of isolated mitochondria, was ineffective in rescuing import activity. Only when Hsp70 was present during the translation of the mitochondrial precursor protein was maximal import observed. Thus, these results enforce the idea that Hsp70 is required to maintain newly synthesized mitochondrial proteins in a “translocation competent state”. Unfortunately, experimental data demonstrating a direct interaction of mitochondrial precursor proteins with Hsp70 or its DnaJ co-factors, and how such interactions maintain translocation competence of the substrate are rather limited. 4.2.3. Targeting to the Peroxisome Proteins synthesized in the cytosol and transported into peroxisomes also carry organellar targeting sequences, oftentimes located at the extreme carboxy terminus (e.g. SKL), and which are not removed from the mature form of the protein. A second targeting sequence composed of 9 amino acids, present either at the extreme amino-terminus or internally, also has been described. Components or “receptors” dedicated to the binding (and perhaps transport) of newly synthesized peroxisomal proteins carrying either of these two signals have only recently been identified, primarily by genetic means in yeast. Some of these are localized on the peroxisomal membrane facing the cytosol, while others appear to be largely cytosolic (recently reviewed by Subramani, 1996). Newly synthesized peroxisomal proteins, unlike other proteins destined for either the mitochondria or ER, do not have to be in an extended or unfolded conformation in order to be transported into the organelle. Already folded peroxisomal proteins present in the cytosol appear to be taken up into the organelle through either a large channel in the membrane, or perhaps by some sort of endocytotic process. Whatever the case, evidence for a role of the cytosolic Hsp70 chaperone in facilitating peroxisomal import has been reported. For example, antibodies to cytosolic Hsp70, either when injected into living cells or when added to permeabilized cells, effectively inhibited the import of some proteins into the peroxisome (Walton et al. 1994). Whether Hsp70 acts to facilitate conformational changes during the import process, or perhaps acts to present the peroxisomal targeting sequence to its appropriate receptor, is still not clear. 4.3. Relative Roles of Dedicated vs. Generic Chaperones in Protein Targeting As we have discussed, the targeting of newly synthesized proteins across intracellular membranes is mediated by specific components dedicated to a particular translocation pathway, as well as by the more classical members of the molecular chaperone family. Whether the two different classes of factors, dedicated and generic, compete for their substrates, or alternatively work together to facilitate protein translocation events, is not clear. We suspect that in the case of the more generic-like molecular chaperones (e.g. DnaK/Hsp70 family, chaperonins), their interaction with a precursor protein is very dynamic; i.e. the molecular chaperone undergoes multiple cycles of binding and release to the target substrate with the reaction cycle dependent on ATP usage. In contrast, those components dedicated to specific protein translocation pathways (e.g. SecB, SRP, MSF)
Molecular chaperones and folding catalysts
210
likely exhibit a more static interaction with their substrate protein; i.e. once bound to the precursor protein, they remain bound until a subsequent interaction with their appropriate organellar receptor (e.g. SRP with the SRP-receptor, SecB with SecA, MSF with the translocase present at the mitochondrial outer membrane). Interaction with the receptor then might represent the rate limiting step for the release of the substrate protein. Thus, when both classes of factors are present, those that function in a dedicated manner (i.e. substrate-specific) likely prevail and handle the majority of protein translocation events. Presumably, when these dedicated factors are missing or are nonfunctional, or when overburdened due to substrate excess, the more generic molecular chaperones now are called into play to help deal with protein translocation events.
5. PATHWAYS AND COMPONENTS INVOLVED IN THE MATURATION OF CYTOSOLIC PROTEINS Proteins which are transported across membranes appear to rely on the help of various escorts, be they generic or dedicated members of the molecular chaperone family. Whether the maturation of cytosolic proteins also is dependent upon the participation of molecular chaperones only recently has begun to be addressed. Below we discuss the current ideas regarding cellular pathways involved in the maturation of proteins which reside in the cytoplasm. We begin by summarizing some of the original observations which led to the suggestion that nascent polypeptides interact with the cytosolic Hsp70 chaperone during the course of their synthesis on the ribosome. We then review more recent studies examining the complexity of protein folding in both eukaryotes and prokaryotes, and discuss the different components thought to be involved in the interaction with nascent and newly synthesized proteins. 5.1. Folding in the Eukaryotic Cytosol In both animal cells and yeast there are multiple and highly related forms of Hsp70 present within all cellular compartments. Within the eukaryotic cytosol there are several forms of Hsp70, often delineated by their different mode of regulation. For example, in animal cells the predominant form of cytosolic/nuclear Hsp70 which is expressed under normal growth conditions is referred to as the heat shock cognate (Hsc70). In response to metabolic stress cells make another form referred to as Hsp70. For the most part, Hsp70 expression is restricted to the cell experiencing metabolic stress (note however, there are several exceptions to this rule). Hsc70 and Hsp70 have a high degree of sequence identity and appear to interact with one another in some cells (Brown et al., 1993). It is still not clear why eukaryotic cells produce a novel form of hsp 70 in response to heat shock and other metabolic stresses (and therefore below we will use Hsp70 and Hsc70 interchangeably). The rationale for examining the potential interaction of cytosolic Hsc70 with nascent polypeptides was based on earlier studies showing that the Grp78/BiP homolog interacted with at least some newly synthesized proteins entering into the lumen of the endoplasmic reticulum. For example, BiP had been shown to bind to the monomeric forms of IgG
Early events in the synthesis and maturation of polypeptides
211
heavy (H) and light (L) chains, and to orchestrate their assembly into the mature H2L2 structure (see Haas and Zimmermann, this volume). Consequently, our laboratory examined whether the cytosolic forms of Hsp70, like BiP, might interact with nascent or newly synthesized proteins in the eukaryotic cytosol. To address this issue, animal cells were subjected to metabolic pulse-chase radiolabeling, and the possible interaction of cytosolic Hsc70 with the radiolabeled nascent and/or newly synthesized proteins examined by immunoprecipitation. Here, the cells were solubilized under conditions chosen to minimize the destruction of relevant protein-protein interactions. Moreover, owing to the fact that the BiP interacted with its protein substrates in an ATP dependent manner, the ATP hydrolyzing enzyme apyrase was included in the buffer used to solubilize the cells. Immunoprecipitation of Hsc70 from the cells pulse-labeled with 35Smethionine resulted in the capture of Hsc70 along with a collection of newly synthesized and/or nascent polypeptides. Analysis of the material by SDS-PAGE revealed a “smear” of radioactive material migrating from the top to the bottom of the gel, indicative of a heterogeneous mixture of nascent polypeptides. In contrast, immunoprecipitation of Hsc70 from the cells pulse-labeled and then chased in the absence of the radiolabel resulted in the co-precipitation of only a few radiolabeled proteins. Consequently, Hsc70 appeared to interact only transiently with many newly synthesized proteins. The capture of nascent and newly synthesized proteins in a complex with the Hsc70 chaperone was dependent upon the rapid depletion of ATP levels. Simply adding excess ATP to the cell lysate was sufficient to elicit the release of Hsc70 from its nascent/newly synthesized protein targets. It was also shown that nascent polypeptides released prematurely from the ribosome, via treatment of cells with puromycin, could be captured in a complex with Hsc70. Finally, from highly purified preparations of polysomes, radiolabeled nascent chains again were observed to coprecipitate with Hsc70. This result indicated that the interaction of Hsc70 with the nascent chains was occurring cotranslationally, rather than taking place only after the chains had been released prematurely from the ribosome. Based on these observations it was suggested that Hsc70 interacted with nascent polypeptides being synthesized on the ribosome (Beckmann et al., 1990). Through such an interaction with Hsc70 the nascent chain would be maintained in a “protected conformation”, perhaps being restricted from interacting with other components in its vicinity which might lead to its misfolding. Only after a sufficient amount of information was present (i.e. a sufficient number of amino acids synthesized and free of the ribosome channel), would the nascent polypeptide be able to initiate the folding process. Such folding would commence upon the ATP-dependent release of the Hsc70 chaperone (Beckmann et al., 1992; Beckmann et al. 1990). Work over the past 6 years has supported and extended these observations. Craig and co-workers reported that a considerable portion of the yeast cytosolic Hsp70 homologues, Ssb1 and Ssb2, fractionated with translating ribosomes (Nelson et al., 1992). Prior treatment of the cells with puromycin (to elicit release of the nascent chains) resulted in a significant loss of Ssb1/2 from the polysomes, consistent with the idea that Ssb1/2 was present in the polysomes owing to its interaction with the nascent chains. Yeast strains depleted of cytosolic Ssb1/2 were found to grow slowly, were hypersensitive to drugs which cause errors in protein synthesis, and contained relatively low levels of active polysomes. A suppressor which appeared to restore at least some wild type growth
Molecular chaperones and folding catalysts
212
properties to the cells depleted of Ssb1/2 was isolated. Interestingly, the suppressor gene encoded a protein synthesis factor involved in translational elongation. These observations, in sum, provided further support for the idea that cytosolic Hsp70 played a role in the early stages of polypeptide maturation. The pathways of protein folding in the eukaryotic cytosol now are being examined primarily using in vitro translation systems. Here the identification of cytosolic proteins which interact with nascent polypeptides, as well as the physical properties of the nascent polypeptide itself, are being explored with different model proteins. Our own studies have shown that nascent chloramphenciol acetyl transferase (CAT) chains of less than 200 amino acids exhibited a very large native mass when released from the ribosome. Via chemical cross-linking we found that a significant portion of the nascent CAT chains were bound to the Hsc70 chaperone. Moreover, when in a complex with their Hsc70 chaperone, the nascent chains appeared to be protected from added protease (Hansen et al., 1994). Frydman et al., examined the early events surrounding the synthesis and folding of firefly luciferase (Frydman et al., 1994). Either full length mRNA encoding luciferase, or truncated mRNA’s lacking a stop codon, were translated in rabbit reticulocyte lysate and analyzed by various means including limited proteolysis, coprecipitation with antibodies to specific molecular chaperones, and assaying the specific activity of the newly synthesized luciferase enzyme. In addition to Hsc70, they reported two other cytosolic chaperones which interacted with the nascent luciferase polypeptide. One of these, Hsp40, represents one of many eukaryotic homologues of the bacterial DnaJ protein, a known co-factor for the DnaK (i.e. bacterial Hsp70) chaperone. The second was the TCP-1 containing cytosolic chaperonin complex (CCT). Cytosolic chaperonin is weakly related to both the bacterial GroEL complex and the mitochondrial/ chloroplast Hsp60 chaperonin (for a review of CCT, see Kubota et al., 1994). To examine the relevance of these different chaperones interacting with nascent luciferase, depletion experiments were performed. Interestingly no adverse effects on the overall rate, or extent of translation was observed following the depletion of the individual chaperones. Rather, removal of any one of the 3 different chaperone components resulted in a significant reduction in the specific activity of the newly synthesized luciferase enzyme. Re-addition of Hsc/Hsp70 to the depleted lysate was effective in restoring the efficient folding of the enzyme, provided that it was added before the start of translation. Luciferase folding could be not be rescued when Hsp70 was added post-translationally, consistent with the idea that Hsp70 is needed cotranslationally. In contrast, addition of purified CCT posttranslationally was partially effective in restoring the activity of the newly synthesized enzyme, indicative that the chaperonin could act on the enzyme late in the folding process. From these and other results, the investigators suggested a model in which the 3 different chaperones cooperated in sequential fashion to facilitate the folding of nascent luciferase (Frydman et al., 1994). As the nascent chain emerges from the ribosome, with as few as 40 amino acids exposed, it is recognized by and subsequently becomes bound to Hsp40. Targeting of the chain by Hsp40 leads to the subsequent recruitment of Hsp70 and formation of a ternary complex where both Hsp40 and Hsp70 now are bound to and presumably stabilizing the nascent chain. Only after approximately 250–300 amino acids have been synthesized would the nascent luciferase chain now contain enough information to allow
Early events in the synthesis and maturation of polypeptides
213
for the folding of its first domain. At this stage in its synthesis, the nascent luciferase chain would be transferred over to the TCP-1 containing chaperonin. In their model, subsequent folding of the nascent luciferase then would occur within the confines of the chaperonin particle. Additional studies over the last two years, many from the Hartl lab, have resulted in some modifications of the model described above (Frydman and Hartl, 1996; Hartl, 1996). For example, another component has been suggested to be involved in the nascent chain interaction, an approximate 45 kDa protein termed Hip (for Hsc704-interacting protein, Höhfeld et al., 1995). Sequencing of Hip revealed its homology to the so-called tetratricopeptide repeat (TPR) family of proteins, all of which contain multiple degenerate repeats of a 34 amino acid motif. One such protein is the Stil protein, a stressinducible protein of unknown function in S. cerivisiae (Nicolet and Craig, 1989). In animals a Stil homologue, termed p48, has been shown to interact with steroid hormone receptors in collaboration with both Hsc70 and Hsp90 (Prapapanich et al. 1996). Owing to its reported ability to interact with Hsc70, it was suggested that the tetrameric Hip protein might function to stabilize Hsc70 in its ADP-bound state, thereby favoring Hsc70’s interaction with a nascent chain (Höhfeld et al., 1995). Our own work (both published and unpublished) continues to support the idea that nascent chains, in general, enter into high molecular weight complexes during the early stages of their biosynthesis. When mRNA’s (either full length or truncated) encoding tubulin (a heterodimer), firefly luciferase (monomer), chloramphenicol acetyl transferase (CAT, a homo-triiner) and -galactosidase (a homo-tetramer) are translated in rabbit reticulocyte lysate, we find Hsc70 to represent the most prominent nascent chain binding component. In experiments where we reduce the levels of either Hsc70 or Hsp40 from the rabbit reticulocyte lysate, we observe a significant reduction in the overall extent of protein synthesis activities. These observations appear in concert with those of Matts and colleagues where the relative extent of protein synthesis activities in rabbit reticulocyte lysates was directly correlated with the levels of Hsc70 (Matts and Hurst, 1992). Simply lowering Hsc70 levels via the addition of an unfolded protein (to provide a target for Hsc70 binding) resulted in a marked diminishment of protein synthesis activity (Matts et al., 1993). A useful reagent we have developed recently for our studies is an antibody specific for the antibiotic puromycin. Following release of the nascent chains by puromycin, we can rapidly isolate the puromycin containing nascent chains and examine both their physical properties as well as associated proteins. Results from such experiments performed in vivo also support the idea that it is Hsc70 which represents the most prominent nascent chain binding protein. To date, we have not had much success in capturing nascent chains bound to either Hsp40 or Hip. Rather, the most prominent component we observe bound to the puromycin released chains continues to be the Hsc70 chaperone. While we do observe small amounts of the cytosolic CCT chaperonin bound to the puromycin released chains, the relative levels indicate to us that a co-translational interaction of CCT with nascent chains seems unlikely. Consistent with these observations are our findings that in isolated polysomes the levels of CCT (as well as both Hsp40 and p48/Hip) are extremely low, be it relative to the mass of the nascent chains or that of Hsc70 (Hansen et al., in prep; Nagata et al., in press.; Eggers et al., 1997).
Molecular chaperones and folding catalysts
214
The idea that the folding of at least some polypeptides can occur co-translationally is supported by some recent observations utilizing clever methodologies. First, the observation that radiolabeled hemin is found to co-sediment with polysomes carrying nascent and globin is consistent with the idea of globin chains undergoing cotranslational folding (Komar et al., 1993). In other studies, modified mRNA’s were prepared where additional amino acid coding information (but no information for a stop codon) was engineered into the 3′ ends of mRNA’s encoding either luciferase of rhodanese. The idea was to generate ribosome bound forms of the proteins via these carboxy-terminal extensions, and to determine whether the newly synthesized proteins were enzymatically active. In both cases, carboxy-terminal extensions of approximately 25 amino acids resulted in the ribosome bound forms of the proteins being enzymatically active (Kudlicki et al., 1995; Makeyev et al., 1996). Presumably, the carboxy-terminal extensions allowed for the entirety of the luciferase or rhodanese molecules to escape the ribosome tunnel and initiate folding, even though still bound to the ribosome. Finally, the concept of co-translational folding takes on a whole different level of complexity when one considers the results of very recent studies examining the folding of a particular viral protein. For example, the reovirus cell attachment protein is known to exist as a homotrimer, with its trimerization being driven both at the amino and carboxy termini. Using in vitro translation approaches, Gilmore et al. have presented data showing that trimerization of the reovirus attachment protein occurred as the protein was being synthesized on the ribosome (Gilmore et al. 1996). Presumably the nascent chains initiated their interaction with one another co-translationally (within the same polysome?) once a sufficient number of amino acids had been synthesized and extruded from the ribosome. 5.2. Folding in the Prokaryotic Cytoplasm Support for a role of one or more members of the bacterial chaperone systems in protein folding and maturation follows from studies showing that E. coli mutants deficient in the expression of either the DnaK/DnaJ/GrpE chaperone machine, or the GroEL/ES chaperonin fail to support the growth of bacteriophages and exhibit high levels of protein aggregates (reviewed by Georgopoulos et al., 1990; Georgopoulos and Welch, 1993) (see Burkholder and Gottesman, this volume). Overexpressing different members of these chaperone machines helped to reverse such defects and restore normal protein solubility (Van Dyk et al. 1989). Using a slightly different approach, Horwich et al. created a bacterial cell line expressing a temperature sensitive form of GroEL. At the nonpermissive temperature where GroEL presumably was no longer functional, the cells grew very poorly, exhibited significant reductions in their ability to support phage biogenesis, and showed a reduction in overall protein synthesis activities. At the nonpermissive temperature, many newly synthesized proteins apparently failed to reach their native state. Instead, a large fraction of the newly synthesized proteins produced in the absence of functional GroEL were found to partition into the detergent insoluble fraction, likely due to their misfolding and subsequent aggregation (Horwich et al., 1993). In vitro translation, along with the use of either photo-activated or chemical-based
Early events in the synthesis and maturation of polypeptides
215
cross-linking methods, has been used to identify bacterial proteins interacting with nascent and/or newly synthesized proteins. For example, using an initiator methionine containing a photoreactive cross-linker, Bochkareva et al., reported the interaction of newly synthesized lactamase (in its precursor form) and chloramphenicol acetyl transferase with the GroEL chaperonin (Bochkareva and Girshovich, 1992; Bochkareva et al., 1988). Hendrick et al., also employing cross-linking methods, reported that the bacterial DnaJ protein, when added to a rabbit reticulocyte lysate, could be cross-linked to both nascent luciferase and CAT (Hendrick et al., 1993). Here, addition of high levels of DnaJ to the lysate appeared to arrest the folding of the newly synthesized proteins, unless the bacterial cofactors of DnaJ, DnaK and GrpE, also were added. Results of more recent cross-linking experiments have led two groups to conclude that the most prominent nascent chain binding proteins in E. coli are the signal recognition particle (SRP) and trigger factor (TF). Valent et al., found that the majority of presecretory proteins made in vitro could be cross-linked to SRP (Valent et al., 1995). These investigators also reported that TF could be cross-linked to almost any nascent chain being synthesized in the in vitro system, be it proteins with or without signal sequences. The interaction of TF with nascent polypeptides no longer was observed when the nascent chains were released prematurely from the ribosomes via puromycin. In a similar study, Hesterkamp et al., concluded that nascent polypeptides could be efficiently crosslinked to TF in a bacterial extract (Hesterkamp et al., 1996). These investigators also noted that TF exhibits homology to and can function as a peptidyl prolyl-cis/trans isomerase (PPI). Studies by Stoller et al., have shown that TF is a PPIase (Stoller et al., 1995) which is, as described earlier, associated with the bacterial 50S ribosomal subunit (see Fischer and Schmid, this volume). It is still not clear as to why these different groups have arrived at somewhat different conclusions using cross-linking methods to identify nascent chain binding proteins. To reiterate the results, GroEL, DnaJ, SRP, and TF all were reported to cross-link with a nascent polypeptide (sometimes the same polypeptide substrate). In the discussion of their work Valent et al., prompted by the results of Hendrick et al., mentioned that they looked for, but could not find, evidence for an interaction of DnaJ with nascent polypeptides (Valent et al., 1995). Finally, in light of the observations in eukaryotes where Hsc70 appears to bind nascent polypeptides, it is somewhat surprising that none of the aforementioned studies in bacteria reported the cross-linking of DnaK with nascent or newly synthesized proteins (see, e.g. Hesterkamp et al., 1996). Yet, as is described below, other investigators have noted the presence of DnaK within isolated polysomes. In yet another approach to the question of components which might interact with nascent/newly synthesized proteins, Langer et al. diluted a chaotrope denatured substrate, rhodanese, into buffers containing different members of the bacterial chaperone system. They concluded that DnaK, DnaJ, and GroEL cooperated together in the re-folding of rhodanese (Langer et al., 1992). Results from order of addition experiments led the authors to conclude that DnaK was the first to bind to the unfolded protein, with DnaJ participating shortly thereafter. It was noted that for certain substrates DnaJ might interact first, followed by DnaK Final folding of rhodanese would involve the transfer of the protein from the DnaK machinery (i.e. DnaK/DnaJ/GrpE) over to the GroEL/ES chaperonin machinery. While the GroEL/ ES chaperonin machinery clearly has been
Molecular chaperones and folding catalysts
216
implicated as being important for the folding of some newly synthesized proteins, all of the available evidence indicate that its interactions with folding substrates occurs posttranslationally. Gaitanaris, analyzing the composition of polysomes prepared from E. coli, noted the presence of both DnaK and DnaJ, but not the GroEL chaperonin. While some newly synthesized proteins were found in a complex with GroEL, it was suggested that such interactions occurred only following release of the newly synthesized proteins from the ribosome (Gaitanaris et al., 1994). Vysokanov has concluded an intimate role for both DnaK and DnaJ in the early stages of CAT folding (Vysokanov, 1995). Here DnaK and DnaJ co-fractionated with polysomes, and like the situation in eukaryotes, both were lost from the peak polysome fraction upon prior treatment of the lysates with puromycin. Interestingly, in those lysates depleted of both DnaK and DnaJ, newly synthesized CAT (released from the ribosome) was found either as an inactive monomer or bound to the GroEL chaperonin. Using rhodanese as the test protein, Reid and Flynn have concluded only a post-translational interaction of the newly synthesized protein with GroEL (Reid and Flynn, 1996). Finally, Kudlicki et al., also have examined the protein folding pathway of rhodanese synthesized in vitro. These investigators were surprised to find that as much as 60% of the newly synthesized full length rhodanese remained bound to the ribosome, apparently in the form of a peptidyl-tRNA still engaged with the ribosome (Kudlicki et al., 1994). While this form of the protein appeared enzymatically inactive, addition of DnaK/DnaJ/GrpE and GroEL/ES resulted in the release of the ribosome bound rhodanese and its acquisition of biological activity. Curiously, addition of either DnaK or GroES promoted the release of the ribosome bound protein in a form that was enzymatically inactive. Based on these and other observations, Kudlicki et al. suggested that full length rhodanese accumulates on the ribosome as peptidyl-tRNA, and that chaperone mediated interactions are necessary both for translational termination and the release of the properly folded protein from the ribosome. 5.3. Summarizing the Results Obtained from Eukaryotic and Prokaryotic Studies It is clear from the discussion presented here that no consensus regarding the components which interact with nascent polypeptides, either in eukaryotic or prokaryotic systems, has been reached. Nor do we have any firm conclusions regarding the physical properties of a nascent polypeptide. Finally, whether nascent chains undergo folding either co- or posttranslationally (or both) still awaits further study. We are not too surprised by this lack of consensus considering the complexity of the problem, the different techniques and systems being employed to attack the problem, and the fact that these types of studies are relatively new. Indeed, when reading textbooks published within the last 10–20 years, little mention is even made regarding the status of the nascent chain, or the early events involving the folding of a newly synthesized protein. All of the available data indicate that there are a number of cytosolic components which are capable of interacting with nascent polypeptides. There is general agreement that dedicated components such as SecB and SRP are the primary participants in the interaction of nascent polypeptides destined for secretion out of the cell. In the case of proteins destined for import into the mitochondria, evidence implicating dedicated
Early events in the synthesis and maturation of polypeptides
217
factors, as well as the more generic molecular chaperones like Hsc70, has been presented. Here however, it is still not clear which components handle the bulk of the work. While many point to the Hsp70 chaperone machinery as being responsible for maintaining mitochondrial precursor proteins in a translocation competent state, relatively little experimental data demonstrating a direct interaction of Hsc70 with mitochondrial precursor proteins has been presented. Finally, for proteins whose final destination is the cytosol (and nucleus in eukaryotes), the identity and/or necessity of their particular escort (s) during synthesis and folding requires further investigation. As was mentioned above, results from eukaryotic cells continues to implicate the Hsp70 chaperone machinery (i.e. Hsp70 and DnaJ co-factors) as the primary nascent chain binding component. Yet in prokaryotic cells, support for an interaction of either the DnaK or DnaJ chaperones with nascent polypeptides is weak. Rather, results of recent studies have implicated trigger factor as being the generic nascent chain binding component in bacteria. Unlike the more classical chaperones however, TF does not appear to require a nucleotide for its binding to or release from a polypeptide substrates and it has an enzymatic activity as PPIase. Consequently, what might regulate the on/off reaction of TF with a nascent chain is not yet clear. While there may not be a consensus on what components interact with the majority of nascent polypeptides, there does seem to be general agreement that a nascent chain would require component(s) to chaperone the early stages of its folding. Specifically, it seems plausible that the nascent chain, lacking a suitable number of amino acids to initiate productive folding, would need some type of escort as it emerges from the ribosome. Such an escort would help to insure that the nascent chain does not initiate an interaction with another macromolecule in its vicinity which could interfere with the eventual folding of the nascent/newly synthesized protein. Whether this putative escort in eukaryotes (prokaryotes) is Hsp70 (DnaK), Hsp40 (DnaJ), or other factors (e.g. trigger factor) will necessitate further study. Whether all nascent polypeptides and or newly synthesized proteins require an escort/molecular chaperone for their efficient synthesis and folding is still open to debate. Indeed, a similar question as to the necessity of the chaperonins for protein folding inside both prokaryotes and eukaryotes recently has been presented. For example, Lorimer has published calculations from which he concludes that the levels of the GroEL/ES chaperonin machinery could accommodate no more than 5% of the newly synthesized proteins being made in bacteria under normal conditions (Lorimer, 1996). Similarly, Rospert et al., recently demonstrated that out of 4 different proteins imported into the mitochondria, only one apparently required the Hsp60 (GroEL) chaperonin for its efficient folding (Rospert et al., 1996). In the case of the eukaryotic cytosolic chaperonin, CCT, its levels appear even lower than that observed for either the GroEL or Hsp60 chaperonins. Consequently, most investigators suspect that CCT has a limited role in protein folding events, likely being restricted to a rather small number of polypeptide substrates, the most notable being actin and tubulin. Even here however, the CCT chaperonin may not be sufficient. Recent studies from Cowan and colleagues have shown that CCT is needed, along with 4 other factors, for the efficient folding of -tubulin (Tian et al., 1996) (see Willison, this volume).
Molecular chaperones and folding catalysts
218
6. UNRESOLVED QUESTIONS, CONTROVERSIAL ISSUES, FUTURE DIRECTIONS A picture of nascent chain biogenesis hopefully will become clearer over the next few years. Continued efforts, utilizing both genetic and biochemical approaches, will be needed to help identify some of the important components involved, as well as clarify the pathway by which nascent and newly synthesized proteins reach their biologically active conformation inside the cell. Some of the more pressing questions we anticipate that the field will begin to address in the near future include: (1) Do all nascent chains require an interaction with nascent chain binding factors for efficient folding? Do both single domain and multi-domain polypeptides utilize the same nascent chain binding components? Are molecular chaperones involved in all protein folding events, or might they be used primarily for proteins that have gone “off pathway?” (2) What is the extent of co- versus post-translational folding? Will this vary significantly depending upon the nature of the polypeptide species (e.g. amino acid sequence, number of protein domains, monomer versus oligomer, etc.)? (3) How many generic nascent chain binding proteins exist? Do different nascent chain binding components compete for the binding of a particular nascent chain? What is the interplay between generic and dedicated nascent chain binding components? For example, does the Hsp70/DnaK chaperone machine compete with signal recognition particle for the binding of proteins bearing an ER targeting signal sequence? Once a polysome translating a secretory protein has docked at the ER membrane, is the nascent chain ever exposed to the cytosol? If so, does the cytosolic Hsp70 chaperone ever interact with the nascent chain? (4) Is the rate of translation affected by the interaction (or lack thereof) of molecular chaperones with the nascent chain? Similar to the situation in which interactions of the nascent chain with SRP have been suggested to influence translational elongation, might other nascent chain binding components affect either the rate of translation, or the release of the polypeptide from the translation machinery? Could the obvious stalls in polypeptide elongation observed in vitro be connected to chaperone interactions (or lack thereof) with the nascent chain? Might the rate of translational initiation be regulated, at least in part, by the available levels of the molecular chaperones? Do molecular chaperones play a role in the recycling of initiation and elongation factors? Under conditions of metabolic stress where many of the molecular chaperones become increasingly bound to proteins undergoing denaturation, does the translation machinery shut down owing to the low levels of available chaperones? If so, what is the operative mechanism by which low levels of the chaperones result in the inhibition of protein synthesis? (5) Might nascent chains be stabilized in an extended or unfolded conformation to allow for other cellular components to get a look at the nature of the nascent polypeptide? For example, do other molecules like the nuclear localization sequence binding protein, peptidyl-prolyl cis/trans isomerases, ubiquitin modifying enzymes, protein
Early events in the synthesis and maturation of polypeptides
219
kinases, myristylation and or glycosylation enzymes etc., monitor or even have access to the nascent chain by virtue of the chain being in an extended or unfolded conformation via the actions of nascent chain binding components? (6) Do protein prosthetic groups interact with nascent chains and if so, are such interactions rate limiting as it pertains to co-translational folding? In the case of monomeric proteins destined for assembly into higher ordered oligomeric structures, what is the folding status of the monomeric protein when it is released from the ribosome? Are monomeric subunits maintained in an assembly competent form by their continued interaction with a nascent chain binding component, even after their release from the ribosome? This type of question is particularly relevant for proteins whose higher ordered assembly is driven primarily through hydrophobic interactions. Exactly where is the “hydrophobic domain” situated when the newly synthesized monomer is released from the ribosome? Based on thermodynamic considerations the hydrophobic region presumably would be buried somewhere in the interior of the folded polypeptide. If correct, how is this region subsequently exposed in order to drive the oligomeric assembly reaction? Alternatively, might the hydrophobic domain be situated or exposed on the surface of the monomer, but be stabilized or shielded via the action of one or more molecular chaperones?
7. ACKNOWLEDGEMENTS Our work has been supported by both the National Institute of Health (GM 33551) and the National Science Foundation (MCB-9421946).
8. REFERENCES Aitken, A. (1992). 14–3–3 proteins: a highly conserved, widespread family of eukaryotic proteins. Trends Biochem. Sci. , 17 , 498–501. Altman, E., Kumamoto, C.A. and Emr, S.D. (1991). Heat-shock proteins can substitute for SecB function during protein export in Escherichia coli. EMBO J. , 10 , 239–245. Anfinsen, C.B. (1973). Principles that govern the folding of protein chains. Science , 181 (96), 223–230. Atencio, D.P. and Yaffe, M.P. (1992). MAS5, a yeast homologue of DnaJ involved in mitochondrial import. Mol. Cell Biol. , 12 , 283–291. Beckmann, R.P., Lovett, M. and Welch, W.J. (1992). Examining the function and regulation of hsp70 in cells subjected to metabolic stress . J. Cell Biol , 117 , 1137– 1150. Beckmann, R.P., Mizzen, L.A. and Welch, W.J. (1990). Interaction of hsp70 with newly synthesized proteins: Implications for protein folding and assembly. Science , 248 , 850–854. Berman, A.L., Kolker, E. and Trifonov, E.N. (1994). Underlying order in protein sequence organization. Proc. Natl Acad. Sci. USA , 91 , 4044–4047. Bernstein, H., Poritz, M., Strub, K., Hobern, P., Brenner, S. and Walter, P. (1989). Model for signal sequence recognition from amino acid sequence of 54kDa subunit of signal recognition particle. Nature , 340 , 482–486.
Molecular chaperones and folding catalysts
220
Blobel, G. (1980). Intracellular protein topogenesis. Proc. Natl. Acad. Sci. USA , 77 , 1496–1500. Bochkareva, E.S. and Girshovich, A.S. (1992). A newly synthesized protein interacts with GroES on the surface of chaperonin GroEL. J. Biol. Chem. , 267(36), 25672– 25675. Bochkareva, E.S., Lissin, N.M. and Girshovich, A.S. (1988). Transient association of newly synthesized unfolded proteins with the heat-shock GroEL protein . Nature , 336 , 254–257. Brown, C.R., Martin, R.L., Hansen, W.J., Beckmann, R.P. and Welch, W.J. (1993). The constitutive and stress inducible forms of hsp70 exhibit functional similarities and interact with one another in an ATP-dependent fashion. J. Cell Biol , 120 , 1101–1111. Caplan, A.J., Cyr, D.M. and Douglas, M.G. (1992). YDJ1p facilitates polypeptide translocation across different intracellular membranes in a conserved mechanism. Cell , 71 , 1143–1155. Caplan, A.J., Cyr, D.M. and Douglas, M.G. (1993). Eukaryotic homologues of Escherichia coli dnaJ: A diverse protein family that functions with HSP70 stress proteins. Mol Biol Cell , 4 , 555–563. Chirico, W., Waters, G. and Blobel, G. (1988). 70K heat shock related proteins stimulate protein translocation into microsomes. Nature , 332 , 805–810. Chirico, W.J. (1992). Dissociation of complexes between 70kDa stress proteins and presecretory proteins is facilitated by a cytosolic factor. Biochem. Biophys. Res. Comm. , 189 , 1150–1156. Craig, E.A. (1990). Regulation and function of HSP70 multigene family of Saccharomyces cerevisiae . In Stress Proteins in Biology and Medicine . R.Morimoto, A.Tissieres and C.Georgopoulos (eds.), Cold Spring Harbor Press, NY, pp. 301–321. Crooke, E., Guthrie, B., Lecker, S., Lill, R. and Wickner, W. (1988). ProOmpA is stabilized for membrane translocation by either purified E. coli trigger factor or canine signal recognition particle. Cell , 54(7), 1003–1011. Crooke, E. and Wickner, W. (1987). Trigger Factor: a soluble protein that folds proOmpA into a membrane-assembly-competent form. Proc. Natl. Acad. Sci. USA , 84 , 5216–5220. Crowley, K.S., Reinhart, G.D. and Johnson, A.E. (1993). The signal sequence moves through a ribosomal tunnel into a noncytoplasmic aqueous environment at the ER membrane early in translocation. Cell , 73 , 1101–1115. Deshaies, R.J., Koch, B.D., Werner-Washburne, M., Craig, E.A. and Schekman, R. (1988). A subfamily of stress proteins facilitates translocation of secretory and mitochondrial precursor polypeptides. Nature , 332(6167), 800–805. Eggens, D.K., Welch, W.J. and Hansen, W.J. (1997). Complexes between nascent polypeptides and their molecular chaperones in the cytosol y mammalian cells. Mol Biol. Cell , 8 , 1555–1573. Frydman, J. and Hartl, F.U. (1996). Principles of chaperone-assisted protein folding: Differences between in vitro and in vivo mechanisms. Science , 272 , 1497–1501. Frydman, J., Nimmesgern, E., Ohtsuka, K. and Hartl, F.U. (1994). Folding of nascent polypeptide chains in a high molecular mass assembly with molecular chaperones. Nature , 370 , 111–117. Gaitanaris, G.A., Vysokanov, A., Hung, S.C., Gottesman, M.E. and Gragerov, A. (1994). Successive action of E. coli chaperones in vivo . Mol. Microbiol , 14(5), 861–869. Georgopoulos, C., Ang, D., Liberek, K. and Zylicz, M. (1990). Properties of the Escherichia coli heat shock proteins and their role in bacteriophage lambda growth, in
Early events in the synthesis and maturation of polypeptides
221
Stress Proteins in Biology and Medicine : R. Morimoto, A.Tissieres and C.Georgopoulos (eds.), Cold Spring Harbor Press, NY, pp. 191–222. Georgopoulos, C. and Welch, W. (1993). Role of the major heat shock proteins as molecular chaperones. Annu. Rev. Cell Biol , 9 , 601–634. Gierasch, L.M. (1989). Signal sequences. Biochem. , 28 , 923–930. Gilmore, R. and Blobel, G. (1985). Translocation of secretory proteins across the micrososmal membrane occurs through an environment accessible to aqueous perturbants. Cell , 42 , 497–505. Gilmore, R., Coffey, M.C., Leone, G., McLure, K. and Lee, P.W. (1996). Cotranslational trimerization of the reovirus cell attachment protein. EMBO J. , 15(11), 2651–2658. Gilmore, R., Collins, P., Johnson, J., Kellaris, K. and Rapiejko, P. (1991). Transcription of full-length and truncated mRNA transcripts to study protein translocation across the endoplasmic reticulum. In Methods in Cell Biology : Academic Press Inc., pp. 223– 239. Guthrie, B. and Wickner, W. (1990). Trigger factor depletion or overproduction causes defective cell division but does not block protein export. J. Bacteriology , 172(10), 5555–5562. Hann, B.C. and Walter, P. (1991). The signal recognition particle in S. cerivisiae. Cell , 67 , 131–144. Hansen, W., Garcia, P.D. and Walter, P. (1986). In vitro protein translocation across the yeast endoplasmic reticulum: ATP-dependent posttranslational translocation of the prepro-alpha-factor. Cell , 45(3), 397–406. Hansen, W. and Walter, P. (1988). Prepro-carboxypeptidase Y and a truncated form of pre-invertase, but not full-length pre-invertase, can be posttranslationally translocated across microsomal vesicle membranes from Saccharomyces cerevisiae. J. Cell Biol , 106(4), 1075–1081. Hansen, W.J., Lingappa, V.R. and Welch, W.J. (1994). Complex environment of nascent polypeptide chains. J. Biol. Chem. , 269 , 26610–26613. Hartl, F.U. (1996). Molecular chaperones in cellular protein folding. Nature , 381 , 571– 580. Hartl, F.U. and Neupert, W. (1990). Protein sorting to mitochondrial: Evolutionary conservations of folding and assembly. Science , 247 , 930–938. Hendrick, J.P., Langer, T., Davis, T.A., Hartl, F.U. and Wiedmann, M. (1993). Control of folding and membrane translocation by binding of the chaperone DnaJ to nascent polypeptides. Proc. Natl. Acad. Sci. USA , 90(21), 10216–10220. Hesterkamp, T., Hauser, S., Lutcke, H. and Bukau, B. (1996). E. coli trigger factor is a prolyl isomerase that associates with nascent polypeptide chains. Proc. Natl. Acad. Set. USA , 93 , 4437–4441. Höhfeld, J., Minami, Y. and Hartl, F.U. (1995). Hip, a novel cochaperone involved in the eukaryotic hsc70/hsp40 reaction cycle. Cell , 83 , 589–598. Horwich, A.L., Low, K.B., Fenton, W.A., Hirshfield, I.N. and Furtak, K. (1993). Folding in vivo of bacterial cytoplasmic proteins: role of GroEL. Cell , 74 , 909–917. Komar, A.A., Kommer, A., Krasheninnikov, I.A. and Spirin, A.S. (1993). Cotranslational heme binding to nascent globin chains. FEBS Lett. , 326 , 261–263. Kubota, H., Hynes, G., Carne, A., Ashworth, A. and Willison, K. (1994). Identification of six Tcp-1-related genes encoding divergent subunits of the TCP-1-containing chaperonin. Current Biology , 4(2), 89–99. Kudlicki, W., Chirgwin, J., Kramer, G. and Hardesty, B. (1995). Folding of an enzyme
Molecular chaperones and folding catalysts
222
into an active conformation while bound as peptidyl-tRNA to the ribosome. Biochemistry , 34 , 14284–14287. Kudlicki, W., Odom, O.W., Kramer, G. and Hardesty, B. (1994). Chaperone-dependent folding and activation of ribosome-bound nascent rhodanese. J. Mol. Biol. , 244 , 319– 331. Kumamoto, C.A. (1989). Escherichia coli SecB associates with exported protein precursors in vivo . Proc. Natl. Acad. Sci. USA , 86 , 5320–5324. Kumamoto, C.A. and Beckwith, J. (1983). Mutations in a new gene, secB, cause defective protein localization in Escherichia coli . J. Bacteriol , 154(1), 253–260. Kusakawa, N., Yura, T., Ueguchi, C., Akiyama, Y. and Ito, K. (1989). Effects of mutations in heat-shock genes groES and groEL on protein export in Escherichia coli . EMBO J. , 8 , 3517–3521. Lain, B., Iriarte, A., Mattingly, J.R., Moreno, J. and Martinez-Carrion, M. (1995). Structural features of the precursor to mitochondrial aspartate amino transferase responsible for binding to hsp70 . J. Biol. Chem. , 270(42), 24732–24739. Lain, B., Iriarte, A. and Martinez-Carrion, M. (1994). Dependence of the folding and import of the precursor to mitochondrial aspartate aminotransferase on the nature of the cell-free translation system. J. Biol Chem. , 269 , 15588–15596. Langer, T., Lu, C., Echols, H., Flanagan, J., Hayer, M.K. and Hartl, F.U. (1992). Successive action of DnaK, DnaJ and GroEL along the pathway of chaperonemediated protein folding. Nature , 356 , 683–689. Lauring, B., Kreibich, G. and Weidmann, M. (1995). The intrinsic ability of ribosomes to bind to endoplasmic reticulum membranes is regulated by signal recognition particle and nascent-polypeptide-associated complex. Proc. Natl. Acad. Sci. USA , 92(21), 9435–9439. Lauring, B., Sakai, H., Kreibich, G. and Wiedmann, M. (1995). Nascent polypeptideassociated complex protein prevents mistargeting of nascent chains to the endoplasmic reticulum. Proc. Natl. Acad. Sci. USA , 92 , 5411–5415. Lecker, S.H., Lill, R., Ziegelhoffer, T., Georgopoulos, C., Bassford, P.J., Kumamoto, C.A. and Wickner, W. (1989). Three pure chaperones of Escherichia coli-SecB, trigger factor and GroEL-form soluble complexes with precursor proteins in vitro. EMBO J. , 8 , 2703–2709. Lill, R., Nargang, F. and Neupert, W. (1996). Biogenesis of mitochondrial proteins. Curr. Op. Cell Biol , 8, 505–512. Lorimer, G.H. (1996). A quantitative assessment of the role of the chaperonin proteins in protein folding in vivo. FASEB J. , 10 , 1–5. Luirink, J. and Dobberstein, B. (1994). Mammalian and Escherichia coli signal recognition particles. Molecular Microbiology , 11(1), 9–13. Luirink, J., High, S., Wood, H., Giner, A., Tollervey, D. and Dobberstein, B. (1992). Signal-sequence recognition by an Escherichia coli ribonucleoprotein complex. Nature , 359 , 741–743. Makeyev, E.V., Kolb, V.A. and Spirin, A.S. (1996). Enzymatic activity of the ribosomebound nascent polypeptide. FEBS Lett. , 378 , 166–170. Matts, R.L. and Hurst, R. (1992). The relationship between protein synthesis and heat shock proteins levels in rabbit reticulocyte lysates. J. Biol. Chem. , 267 , 18168–18174. Matts, R.L., Hurst, R. and Xu, Z. (1993). Denatured proteins inhibit translation in heminsupplemented rabbit reticulocyte lysate by inducing the activation of the hemeregulated eIF-2a kinase. Biochem. , 32 , 7323–7328. Mihara, K. and Omura, T. (1996). Cytoplasmic chaperones in precursor targeting to
Early events in the synthesis and maturation of polypeptides
223
mitochondria: the role of MSF and HSP70. Trends in Cell Biol. , 6(3), 104–108. Murakami, K. and Mori, M. (1990). Purified presequence binding factor (PBF). forms an import-competent complex with a purified mitochondrial precursor protein. EMBO J. , 9 , 3201–3208. Nagata, H., Hansen, W.J., Freeman, B. and Welch, W.J. (1998). Mammalian cytosolic DNaJ homologues affect the hsp70 chaperone-substrate reaction cycle, but do not interact directly with nascent or newly synthesised proteins. Biochemistry , in press. Nelson, R.J., Ziegelhoffer, T., Nicolet, C., Werner-Washburne, M. and Craig, E.A. (1992). The translation machinery and 70 kD heat shock protein cooperate in protein synthesis. Cell , 71 , 97–105. Nicolet, C.M. and Craig, E.A. (1989). Isolation and chacterization of STI1, a stress inducible gene from Saccharomyces cerevisiae . Mol Cell Biol , 9 , 3638–3648. Ono, H. and Tuboi, S. (1988). The cytosolic factor required for import of precursors of mitochondrial proteins into mitochndria. J. Biol Chem. , 263(7), 3188–3193. Park, S., Liu, G., Topping, T.B., Cover, W. and Randall, L.L. (1988). Modulation of folding pathways of exported proteins by the leader sequence. Science , 239(4843), 1033–1035. Phillips, G.J. and Silhavy, T.J. (1990). Heat shock proteins DnaK and GroEL facilitate export of LacZ hybrid proteins in E. coli. Nature , 344 , 882–884. Phillips, G.J. and Silhavy, T.J. (1992). The E. coliffh gene is necessary for viability and efficient protein export. Nature , 359 , 744–746. Poritz, M.A., Strub, K. and Walter, P. (1988). Human SRP RNA and E. coli 4.5S RNA contain a highly homologous stuctural domain. Cell , 55 , 4–6. Powers, T. and Walter, P. (1996). The nascent polypeptide-associated complex modulates interactions between the signal recognition particle and the ribosome. Current Biology , 6 , 331–338. Prapapanich, V., Chen, S., Nair, S.C., Rimerman, R.A. and Smith, D.F. (1996). Molecular cloning of human p48, a transient component of progesterone receptor complexes and an hsp-70 binding protein. Mol Endo. , 10 , 420–431. Pugsley, A. (1993). The complete general secretory pathway in gram-negative bacteria. Microbiol. Rev. , 57(1), 50–108. Randall, L.L. (1992). Peptide binding by chaperone SecB: implications for recognition of nonnative structure. Science , 257(5067), 241–245. Randall, L.L. and Hardy, S.J. (1995). High selectivity with low specificity: how SecB has solved the paradox of chaperone binding. Trends in Biochemical Sciences , 20(2), 65– 69. Randall, L.L., Hardy, T.B. and Hardy, S.J. (1990). No specific recognition of leader peptide by SecB, a chaperone involved in protein export. Science , 248 , 860–863. Rapoport, T.A., Rolls, M.M. and Jungnickel, B. (1996). Approaching the mechanism of protein transport across the ER membrane . Curr. Op. Cell Biol . 8 , 499–504. Reid, B.G. and Flynn, G.C. (1996). GroEL binds to and unfolds rhodanese posttranslationally. J. Biol. Chem. , 271(12), 7212–7217. Rōmisch, K., Webb, J., Herz, J., Prehn, S., Frank, R., Vingron, M. and Dobberstein, B. (1989). Homology of the 54K protein of signal recognition particle, docking protein and two E.coli proteins with putative GTP-binding domains. Nature , 340 , 478–482. Rospert, S., Looser, R., Dubaquie, Y., Matouschek, A., Glick, B. and Schatz, G. (1996). Hsp60-independent protein folding in the matrix of yeast mitochondria. EMBO J. , 15 , 764–774. Rothblatt, J.A., Webb, J.R., Ammerer, G. and Meyer, D.I. (1987). Secretion in yeast:
Molecular chaperones and folding catalysts
224
structural features influencing the post-translational translocation of prepro- -factor in vitro . EMBO J. , 6(11), 3455–3463. Schatz, G. and Dobberstein, B. (1996). Common principles of protein translocation across membranes. Science , 271 , 1519–1526. Stoller, G., Rucknagel, K., Nierhaus, K.H., Schmid, F.X., Fischer, G. and Rahfeld, J.U. (1995). Aribosome-associated peptidyl-prolyl isomerase identified as the trigger factor. EMBO J. , 14(20), 4939–4948. Subramani, S. (1996). Convergence of model systems for peroxisomal biogenesis. Curr. Op. Cell Biol , 8 , 513–518. Terada, K., Ohtsuka, K., Imamoto, N., Yoneda, Y. and Mori, M. (1995). Role of heat shock cognate 70 protein in import of ornithine transcarbamylase precursor into mammalian mitochondria. Mol Cell. Biol , 15(7), 3708–3713. Tian, G., Huang, Y., Rommelaere, H., Vandekerckhove, J., Ampe, C. and Cowan, N.J. (1996). Pathway leading to correctly folded B-tubulin. Cell , 86 , 287–296. Valent, Q.A., Kendall, D.A., High, S., Kusters, R., Oudega, B. and Luirink, J. (1995). Early events in preprotein recognition in E. coli: Interaction of SRP and trigger factor with nascent polypeptides. EMBO J. , 14(22), 5494–5505. Van Dyk, T.K., Gatenby, A.A. and LaRossa, R.A. (1989). Demonstration by genetic suppression of interaction of GroE products with many proteins. Nature , 342 , 451– 453. Verner, K. and Schatz, G. (1988). Protein translocation across membranes. Science , 241 , 1307–1313. von Heijne, G. and Blomberg, C. (1979). Transmembrane translocation of protein. Eur. J. Biochem ., 97 , 175–181. Vysokanov, A.V. (1995). Synthesis of chloramphenicol acetyltransferase in a coupled transcription-translation in vitro system lacking the chaperones DnaK and DnaJ. FEBS Lett. , 375 , 211–214. Walter, P. and Blobel, G. (1980). Purification of membrane-associated protein complex required for protein translocation across the endoplasmic reticulum. Proc. Natl. Acad. Sci. USA , 77(12), 7112–7116. Walter, P. and Johnson, A.E. (1994). Signal sequence recognition and protein targeting to the endoplasmic reticulum. Annu. Rev. Cell Biol , 10 , 87–120. Walter, P. and Lingappa, V.R. (1986). Mechanism of protein translocation across the endoplasmic reticulum. Annu. Rev. Cell Biol., 2 , 499–516. Walton, P.A., Wendland M., Subramani, S., Rachubinski, R.A. and Welch, W.J. (1994). Involvement of 70-kDa heat-shock proteins in peroxisomal import. J. Cell Biol. , 125 , 1037–1046. Wang, S., Sakai, H. and Wiedmann, M. (1995). NAC covers ribosome-associated nascent chains thereby forming a protective environment for regions of nascent chains just emerging from the peptidyl transferase center. J. Cell Biol. , 130(3), 519–528. Watanabe, M. and Blobel, G. (1995). High-affinity binding of Escherichia coli SecB to the signal sequence region of a presecretory protein. Proc. Natl. Acad. Sci. USA , 92 , 10133–10136. Waters, M.G. and Blobel, G. (1986). Secretory protein translocation in a yeast cell-free system can occur posttranslationally and requires ATP hydrolysis. J. Cell Biol. , 102 (5), 1543–1550. Welch, W.J. (1991). The role of heat shock proteins as molecular chaperones. Curr. Opin. Cell Biol. , 3 , 1033–1038. Wickner, W. (1979). The assembly of proteins into biological membranes: the membrane
Early events in the synthesis and maturation of polypeptides
225
trigger hypothesis. Annu. Rev. Biochem. , 48 , 23–45. Wiedmann, B., Sakai, H., Davis, T.A. and Wiedmann, M. (1994). A protein complex required for signal-sequence-specific sorting and translocation. Nature , 370 , 434–440. Wild, J., Altman, E., Yura, T. and Gross, C.A. (1992). DnaK and DnaJ heat shock proteins participate in protein export in Escherichia coli. Genes and Development , 6 (7), 1165–1172.
10. PROTEIN TRANSPORT INTO AND FOLDING WITHIN THE ENDOPLASMIC RETICULUM INGRID G.HAAS1, * and RICHARD ZIMMERMANN2 1 Biochemie-Zentrum
Heidelberg (BZH), Im Neuenheimer Feld 328, D-69120 Heidelberg, Germany 2 Medizinische Biochemie, Universität des Saarlandes, D-66421 Hamburg, Germany
1. Introduction 2. Cytosolic Proteins Involved in Protein Transport into the ER 3. The Protein Translocase in the ER-Membrane 4. Lumenal Proteins Involved in Protein Transport into the ER 5. Unifying Hypothesis: Kinetic Partitioning? 6. Covalent Modification of Proteins in the ER 7. Protein Folding Within the ER 7.1. The Role of N-Glycans in Protein Folding 7.2. Molecular Chaperones in Protein Folding 8. Concluding Remarks 9. References
1. INTRODUCTION The decisive initial step in the biogenesis of most extracellular and many organellar proteins of the eukaryotic cell, i.e. constituent proteins of the endoplasmic reticu-lum, plasma membrane proteins, and proteins of the various other organelles involved in endocytosis and exocytosis, is their transport into the endoplasmic reticulum (ER). Typically, transport requires a signal sequence at the amino terminus of the respective precursor proteins. Transport includes various steps such as association of the precursor protein with the ER membrane, membrane insertion and completion of translocation. The signal sequence not only facilitates membrane specificity, but also helps to preserve the transport competent (non-native) state of the precursor proteins. *Corresponding author
Protein transport into and folding within the endoplasmic reticulum
227
The ER is the compartment where transported proteins reach their native state, i.e. irrespective of their functional location within or outside the cell. Thus the ER also plays a key role in covalent modifications and in protein folding and subunit assembly. To a certain extent there may be a mechanistic coupling between transport and folding of proteins, i.e. some molecular chaperones appear to be involved in both reactions. Eventually, native proteins leave the ER by way of vesicular transport while proteins which fail to reach their native state become degraded. The terms “quality control” and “architectural editing” have been coined to describe the ability of the sorting and packaging systems in the ER to distinguish between native and non-native proteins.
2. CYTOSOLIC PROTEINS INVOLVED IN PROTEIN TRANSPORT INTO THE ER It is well established that precursor proteins are not transported into the ER in their native (i.e. folded) state and that signal peptides in the precursor proteins are involved in preserving the transport competent (i.e. non-native) state as well as in facilitating membrane specificity (Perlman and Halvorson, 1983; von Heijne, 1981, 1983, 1984, 1985). Furthermore, it is generally accepted that there are two alternatively acting mechanisms involved in preserving transport competence in both mammalian and yeast cytosol (Zimmermann and Meyer, 1986; see below). In addition, there is a complex of two cytosolic proteins, termed nascent-polypeptide-associated complex (NAC), which indirectly affects transport by preventing the association of ribosome-associated nascent polypeptide chains which are not destined for the ER with the signal recognition particle (SRP) and the ER membrane (Wiedmann et al., 1994; Lauring et al., 1995a and b; see also Welch et al., this volume). In the first mechanism protein synthesis is slowed down (cotranslational or RNPdependent mechanism). This mechanism involves two ribonucleoprotein particles (RNPs) and their receptors on the microsomal surface. In the mammalian system, this cotranslational pathway has been analyzed in great detail (Table 1, Figure 1). It involves the ribosome (Adelmann et al., 1973; Borgese et al., 1974; Perara et al., 1986) and the SRP (Walter et al., 1981; Walter and Blobel, 1980, 1981a and b and 1982; Krieg et al., 1986; Kurzchalia et al., 1986; Wolin and Walter, 1988 and 1989; Zopf et al., 1990; Römisch et al., 1990; Lütcke et al., 1992), the latter slowing down or even inhibiting the elongation activity of the ribosome upon binding to the emerging signal peptide. The two RNPs make contact to the microsomal surface via an SRP receptor (docking protein) (Meyer and Dobberstein, 1980, a and b; Meyer et al., 1982; Gilmore et al., 1982, a and b; Lauffer et al., 1985; Tajima et al., 1986) and a ribosome receptor (Savitz and Meyer, 1990 and 1993; Wanker et al., 1995; Jungnickel and Rapoport, 1995). There is a requirement for GTP-hydrolysis in the cotranslational transport of precursor proteins (Connolly and Gilmore, 1986 and 1989; Connolly et al., 1991; Rapiejko and Gilmore, 1992; Bacher et al., 1996). This GTP-effect is related to the two GTP-binding proteins, the a-subunit of docking protein and the 54kDa subunit of SRP (Connolly and Gilmore, 1989; Rapiejko and Gilmore, 1992), and to a subunit of the ribosome (Bacher et al.,
Molecular chaperones and folding catalysts
228
1996). In yeast, the SRP54-protein was identified as homolog of mammalian SRP54protein (Poritz et al., 1988; Ribes et al., 1988; Hann et al., 1989; Hann and Walter, 1991; Amaya et al., 1990). The Sec65-protein was identified genetically as a transport component and according to sequence analysis contains a domain with striking similarity to mammalian SRP19-protein (Stirling and Hewitt, 1992; Stirling et al., 1992; Hann et al., 1992). The SRP54- and SRP19-proteins together with five additional subunits and an RNA form yeast SRP (Table 1). This RNP may to a certain extent be seen as a molecular chaperone for signal peptides. However, SRP in concert with its receptor has the additional function to target the nascent chain to the translocase. In the second mechanism protein folding and/or aggregation are slowed down (posttranslational or RNP-independent mechanism). This mechanism involves molecular chaperones which can apparently function as an alternative to the RNPs and their receptors. In yeast, the posttranslational pathway has been analyzed in great detail (Table 1, Figure 1) (Hansen et al., 1986; Waters and Blobel, 1986; Rothblatt and Meyer, 1986). Genetic and biochemical evidence demonstrated the involvement of the cis-acting chaperones Hsc70 (Sssa1p) and Hsp40 (Ydj1p) (Waters et al., 1986; Chirico et al., 1988; Deshaies et al., 1988; Caplan and Douglas, 1991; Caplan et al., 1992; Craig et al., this volume). Hsp40 may be identical to what originally was identified as a second cytosolic (NEM-sensitive) protein which is required besides Hsc70. One active form of Ydj1p was shown to be membrane integrated by way of isoprenylation but it remains unclear how this relates to its role in translocation (Caplan and Douglas, 1991). These molecular chaperones bind to fully synthesized precursor proteins thereby keeping them in a loosely folded conformation. The prevention of premature folding is in addition promoted by the signal sequence of the precursor. The posttranslational transport of precursor proteins requires ATP that is hydrolyzed by Hsc70. In the mammalian in vitro system, the cisacting chaperone has been characterized as Hsc70 and also collaborates with a second cytosolic (NEM-sensitive) protein (Wiech et al., 1987; Zimmermann et al., 1988; Wiech et al., 1993). The current working model proposes that Hsp40 may be the second protein (see above). It remains to be seen whether other proteins described to functionally interact with Hsc70, such as Hsp90 and Hip, are also involved in translocation (Höhfeld et al., 1995; Gross and Hessefort, 1996; Freeman et al., 1996). In the mammalian system, the chain length of the precursor protein is the decisive feature with respect to which of the two mechanisms is operative. The critical chain length was found to be around 70 amino acid residues, larger precursor proteins typically depended on SRP (Müller and Zimmermann, 1987; Schlenstedt and Zimmermann, 1987; Schlenstedt et al., 1990). In the yeast system, the hydrophobicity of the signal peptides appears to be the decisive feature of the precursor protein, i.e. precursor proteins with long hydrophobic core regions within their signal peptides showed a more pronounced requirement for SRP (Ng et al., 1996). It remains to be seen if this latter determining feature acts at the level of SRP or, alternatively, at the level of NAC and only indirectly at the level of SRP (NAC was found to stay away from signal peptides).
Protein transport into and folding within the endoplasmic reticulum
229
3. THE PROTEIN TRANSLOCASE IN THE ER MEMBRANE In mammals, it is assumed that the cotranslational and posttranslational pathways converge at the level of a putative signal peptide receptor which could be the Sec61 subunit of the protein translocase in the ER membrane (Table 1, Figure 1) (Wiedmann et al., 1987; Jungnickel and Rapoport, 1995; Klappa et al., 1994). The core constitu-
Table 1 Components involved in protein transport into and folding within the endoplasmic reticulum
Complex/ Protein
RNA and Protein Components Yeast Name (1)
SRP
Mammals
MW (kDa) XTPase Name (1)
MW (kDa) XTPase
(2)
(2)
scR1 RNA
7SL RNA
Srp72p
72p
SRP72-su
72p
Srp68p
68p
SRP68-su
68p
Srp54p
54p
GTPase SRP54-su
54p
Sec65p
30p
SRP19-su
19p
Srp21p
21p
Srp14p
14p
SRP14-su
14p
Srp7p
7p
SRP9-su
9p
SRP-
DP
-su
receptor
DP -su
70p
GTPase DP
30p
GTPase D -su
ribosome
-su
ERp180
GTPase
69p
GTPase
30p
GTPase
180p
receptor cis-acting
Ssa1p
70p
ATPase Hsc70
70p
chaperones
Ydj1p
40p
Hdj1-p
37p
translocase
Sec66/71p (Hss1p)
31.5gp
Sec67/72p
23p
Sec62p
30p
Sec62p (HTP1) 45p
Try1p/Try2p
35p
TRAM-p (mp39)
36gp
Sec61p/Ssh1p
41p
Sec61 (P37)
37p
-p
ATPase
Molecular chaperones and folding catalysts
230
Sbh1p/Sbh2p
16p
Sec61 -p
14p
Sss1p
9p
Sec61 -p
8p
Sec63p
73p
Mtj1-p
64p
signal
See11p
18p
SPC18/21-su
18/21p
peptidase
SPC25-su
25gp
SPC23-su
23gp
SPC13-su
13p
SPC12-su
12p
SPC20-su
20p
SPC25-su
25p
oligo-
-su
60–64gp
ribophorinI
65gp
saccharyl
-su (Swp1p)
30p
ribophorinII
63gp
transferase
-su (Wbp1p)
45gp
Wbp1-p (OST48)
48p
-su
34p
-su
16p
-su
9p
Complex/ Protein
RNA and Protein Components Yeast
Mammals
Name( 1)
MW XTPase Name( 1) (2) (kDa)
MW XTPase (kDa) (2)
trans-acting
Kar2p
78p
ATPase
BiP (Grp78)
78p
ATPase
chaperones
Lhs1p (Ssi1p)
120gp
ATPase
Grp170
160gp
ATPase
and folding
Scj1p/Scj2p
40p
catalysts
Cne1p (Cnx1p) 76p
Calnexin
88gp
(p90, p88) Calreticulin
59gp
(CaBP3, Erp60) Grp94
94gp
(CaBP4, ERp99)
yPDI (PDI1)
57gp
Hsp47
47p
PDI (ERp59)
59p
ERp61 (Grp58)
61p
ERp72 (CaBP2)
72p
PDIp
62gp
P5 (CaBP1)
59gp
ATPase
Protein transport into and folding within the endoplasmic reticulum
trimming
PDIR
57p
Eug1p
58gp
Mdp1p
36gp
Cyclophilin2
19p
CyclophilinB
21p
FKBP2
15p
FKBP13
16p
GlucosidaseI
95p
enzymes Kre5p -1, 2 mannosidase
155gp 63gp
-GlucosidaseI
85p
-GlucosidaseII
123gp
UDP-Glc: gp-glctransferase
130gp
-1, 2-
231
65gp
mannosidase
(1) synonymous names are given in parentheses. Abbreviations: gp, glycoprotein; Grp, glucose regulated protein; Hsp, heat shock protein; p, protein; su, subunit. (2) p and gp indicate unglycosylated and glycosylated proteins, respectively. Note that cis and trans refers to the cytosolic and lumenal, respectively, surface of the ER membrane.
ent of this translocase on the molecular level is the Sec61p complex, comprising Sec61 p (p37, imp34), Sec61 p and Sec6 p (High et al., 1991 and 1993; Görlich et al., 1992b; Kellaris et al., 1991; Görlich and Rapoport, 1993; Hartmann et al., 1994). Evidence has accumulated that an additional nucleoside triphosphate utilizing (i.e. azidoATP sensitive) subunit takes part in both transport mechanisms (Klappa et al., 1991; Zimmermann et al., 1991; Zimmerman and Walter, 1991; see below). Besides these proteins, the so called TRAM-protein (mp39) appears to be part of the translocase (Wiedmann et al., 1987; Krieg et al., 1989; Thrift et al., 1991; Görlich
Molecular chaperones and folding catalysts
232
Figure 1 Components and mechanisms involved in protein transport into and folding within the endoplasmic reticulum. 1, ribosome with nascent precursor polypeptide chain; 2, signal recognition particle (SRP), shown together with ribosome and nascent precursor polypeptide chain; 3, docking protein (SRP-receptor); 4, cytosolic Hsp70 (Hsc70) (circles), shown together with precursor polypeptide chain and soluble Hsp40 (with J-domain) (squares); 5, Hsp40 with isoprenylanchor; 6, translocating chain-associating membrane protein (TRAM
Protein transport into and folding within the endoplasmic reticulum
233
protein); 7, Sec61 complex; 8, Sec63 protein with J-domain; 9, Sec62/71/72 complex; 10, lumenal Hsp70-protein family member (Kar2p/BiP or Lhs1p/Grp170); 11, oligosaccharyl transferase; 12, signal peptidase; 13, calnexin; 14, protein disulfide isomerase; 15, lumenal Hsp40 with J-domain, shown together with mature polypeptide chain and lumenal Hsp70-protein family member; 16, calreticulin, shown together with mature polypeptide chain.
et al., 1992a). Furthermore, putative mammalian homologs of the Sec62 and Sec63 proteins were recently discovered at the DNA level (Daimon et al., 1997; Brightman et al., 1995) and can be expected to be components of the translocase (see below). With respect to TRAM-protein it is noteworthy that different precursor proteins showed a requirement for TRAM-protein to a different extent. TRAM-protein dependence appeared to be a function of signal peptide characteristics, in that precursor proteins with long amino-terminal plus long hydrophobic core regions within their signal peptides showed a less pronounced TRAM-protein dependence (Voigt et al., 1996). With respect to the actual functions of the individual subunits in protein translocation there are only preliminary insights available at this point (Gilmore and Blobel, 1985; Simon and Blobel, 1991; Crowley et al., 1993a and b; Martoglio et al., 1995; Do et al., 1996; Borel and Simon, 1996). In yeast, it is also assumed that the two pathways converge at the level of a putative signal peptide receptor which may be constituted by proteins Sec71 and Sec72 (Sanz and Meyer, 1989; see below). Originally, genetic evidence suggested that the membrane proteins Sec61, Sec62 and Sec63 (also termed Ptl1 or Npl1) are part of the translocase (Table 1, Figure 1) (Deshaies and Schekman, 1987 and 1989; Sadler et al., 1989; Toyn et al., 1988). Sec63 contains a lumenal domain with a striking similarity to bacterial DnaJ, a protein which functionally interacts with DnaK, the bacterial Hsp70 homologue (Sadler et al., 1989). Additional genetic evidence suggests that the proteins Sec71 and Sec72 are also involved in transport (Green et al., 1992). Biochemical evidence suggested that the Sec61-, Sec62- and Sec63-proteins transiently form complexes with a 31.5 kDaglycoprotein and a 23kDa-protein (Deshaies et al., 1991) which turned out to be identical to Sec71p and Sec72p. Furthermore, precursor proteins in transit can be crosslinked to Sec61p and Sec62p (Müsch et al., 1992; Sanders et al., 1992). Recently, TRAM protein(Try1p and Try2p), Sec61 protein- (Sbh1p and Sbh2p) and Sec61 protein-homologs (Sss1p) were discovered in yeast (Finke et al., 1996; Sommer and Hartmann, personal communication). Furthermore, three types of translocase complexes were detected in yeast, one being identical with the mammalian core complex (comprising Sec61p, Sbh1p, and Sss1p), a related one with two different but highly related gene products (comprising Ssh1p, Sbh2p, and Sss1p), and a third one comprising the core complex plus all the other Sec-proteins which were mentioned above (Panzner et al., 1995; Finke et al., 1996). Therefore, it was proposed that there are different translocases, two dedicated to cotranslational transport and one dedicated to posttranslational transport. However, this view is not undisputed (see below). It will be interesting to see if the requirements for SRP and TRAM-protein correlate in any way.
Molecular chaperones and folding catalysts
234
4. LUMENAL PROTEINS INVOLVED IN PROTEIN TRANSPORT INTO THE ER A direct participation of BiP (Kar2p) in protein translocation into the ER was observed by Vogel et al. (1990), Nguyen et al. (1991), and Schekman and coworkers (Sanders et al., 1992; Brodsky et al., 1993 and 1995). Investigating yeast microsomes, it was possible to crosslink BiP with a translocating polypeptide intermediate. Furthermore, temperaturesensitive mutations in the BiP-encoding gene led to translocation defects both in microsomes and in proteoliposomes derived from these mutants. Surprisingly, in two of these mutants an early step, i.e. the association of the translocating polypeptide with Sec61 protein (membrane insertion), was affected implicating that there is an additional trans-effect of lumenal BiP on protein insertion at the cytosolic membrane surface. In order to explain the transeffect of BiP it was suggested that Sec63 protein of the translocase forms a complex with BiP due to its DnaJ-domain and triggers a conformational change in the translocase (Sadler et al., 1989; Toyn et al., 1988; Rothblatt et al., 1989; Feldheim et al., 1992; Sanders et al., 1992). The functional interaction of Sec63 protein and BiP in mediating membrane insertion and completion of translocation was also deduced from additional yeast mutants. Mutations in Sec63 were isolated which show a decreased association of Sec61 protein with the precursor proteins or a defect in completion of translocation, phenotypes also found in different BiP mutants (Sanders et al., 1992; Lyman and Schekman, 1995). In yeast the same phenotype as described above for BiP mutants was observed for Lhs1p mutants (Craven et al., 1996) and there is evidence that Lhs1 protein, which represents another lumenal member of the Hsp70 protein family, and Sec63 protein also function together in a complex (Craven and Stirling, personal communication). The following model was put forward for the posttranslational protein transport into yeast microsomes (Lyman and Schekman, 1995). In the first phase, the precursor binds to the signal sequence receptor complex in an ATP-independent manner (membrane association). In the second, ATP-dependent phase, BiP (Kar2 protein) and Sec63 protein mediate the transfer of the precursor from the signal sequence receptor complex to the Sec61 protein complex (membrane insertion). In the third and final phase, BiP and Sec63 protein facilitate completion of translocation in an ATP-dependent manner. Although knowledge about the role(s) of Lhs1p in transport is scarce as compared to what is known about BiP, it appears that both molecular chaperones have overlapping functions in transport (Craven et al., 1996). There is good reason to believe that this model is also valid for cotranslational transport (Brodsky et al., 1995) which argues against the existence of strictly separated dedicated translocases (see above). In contrast to the data obtained with yeast microsomes, neither an association of BiP with a translocating polypeptide chain nor a direct role of BiP in protein transport could be demonstrated for mammalian microsomes (Bulleid and Freedman, 1988; Zimmerman and Walter, 1990; Klappa et al., 1991). However, Nicchitta and Blobel (1993) provided evidence that lumenal components are necessary for net transfer of secretory proteins into the mammalian ER. Furthermore, protein disulfide isomerases (PDIs) (see Freedman and Klappa, this volume) and a cyclophilin-type peptidyl prolyl cis/trans isomerase (PPIase)
Protein transport into and folding within the endoplasmic reticulum
235
(see Fischer and Schmid, this volume) have been found in contact with secretory proteins during late stages of translocation into mammalian microsomes. Therefore, it was suggested that proteins other than BiP can be involved in completion of translocation (Klappa et al., 1995; Volkmer et al., 1997). Furthermore, photoaffinity labeling experiments using azido-ATP (Klappa et al., 1991; Zimmermann et al., 1991; Zimmerman and Walter, 1991) led to the conclusion that an ATP-dependent component must be involved in both cotranslational and posttranslational transport into mammalian microsomes. The azido-ATP sensitive component was shown to act at the level of the translocase, i.e. in mediating membrane insertion. In addition, a fraction of microsomal ATP-binding proteins, containing two proteins which are present in the ER-lumen (i.e. BiP and Grp170) as the major constituents, appeared to be required for efficient cotranslational protein transport into proteoliposomes and appeared to be involved in interaction of the precursor with the Sec61 protein complex. Since BiP was unable to functionally substitute for the complete set of ATP-binding proteins it was suggested that Grp170 (Lin et al., 1993; Chen et al., 1996) is the active component and that the model which was put forward for protein transport into yeast microsomes is also valid for mammalian microsomes (Dierks et al., 1996). Grp170 is related to both the Hsp70 and Hsp110 protein families (Chen et al., 1996) and is structurally related to yeast Lhs1p (Dierks et al., 1996).
5. UNIFYING HYPOTHESIS: KINETIC PARTITIONING Various views have been presented over the years on the pathways that polypeptides follow towards their functional location and native state. We favour the concept of a “kinetic partitioning” which was first put forward by Hardy and Randall (1991) to explain the interaction of the bacterial chaperone SecB with its substrate precursor polypeptides, leading to protein export from bacterial cells. According to this concept, the pathway a polypeptide takes depends on the relative rates of folding and association with SecB. Proteins that rapidly fold are precluded from binding to SecB and, therefore, stay in the cytosol. In the case of precursor proteins that contain a signal sequence known to retard folding, association with SecB and, therefore, translocation across the inner membrane is favoured. Adapting this concept to the roles of chaperones and translocase components in protein transport into the ER and of ER-lumenal chaperones in protein folding within the ER we propose the following model. A presecretory polypeptide in the cytosol (or its nascent equivalent) and the corresponding mature polypeptide in the ERlumen (or its nascent equivalent) is “free” to interact with any additional protein(s) and/ or complex(es) listed in Table 1 and shown in Figure 1. According to the concept of kinetic partitioning, the degree of freedom is determined by the rates of spontaneous folding and association with a given soluble or membrane bound component. In the cytosol the rate of folding of a presecretory polypeptide (or its nascent equivalent) is retarded by the signal peptide and, in the case of a nascent precursor polypeptide, by the ribosome. Therefore, the association of a precursor polypeptide (or its nascent equivalent) with either SRP (which interacts with the signal peptide and results in subsequent requirements for the docking protein and the ribosome receptor) or a molecular
Molecular chaperones and folding catalysts
236
chaperone (which interacts with any part of the precursor polypeptide and may result in a subsequent requirement for membrane bound Hsp40) may be favoured. Precursors differ with respect to their partitioning between these components. Furthermore, for some precursor polypeptides (or their nascent equivalents) association with the signal peptide receptor at the microsomal surface may be even more favoured and may preclude association with SRP or a molecular chaperone. In analogy, in the ER-lumen the association of a mature polypeptide (or its translocating equivalent) with a certain molecular chaperone is less favoured compared to its precursor in the cytosol since there is no retardation of folding by the signal peptide. However, the concentration of molecular chaperones in the ER-lumen appears to exceed the concentration of molecular chaperones in the cytosol considerably, and therefore, may compensate for the lower affinity. Again, the partitioning between these components may vary between different polypeptides, i.e. some mature polypeptides may well fold spontaneously (Tyedmers et al., 1996). In further analogy, non-essential components of the protein translocase such as the TRAM protein may or may not interact with a certain precursor polypeptide. There are limits to this kinetic argument and they are related to the energetics of protein transport. A common view of protein transport into the ER is that there are two mechanisms, a cotranslational where the ribosome pushes the polypeptide in transit through the translocase, and a posttranslational where a lumenal Hsp70 pulls the polypeptide in transit through the translocase, in analogy to protein transport into the mitochondrial matrix (see Dekker and Pfanner, this volume) (Figure 2). However, the picture is far from being clear. Some data suggest that Kar2p is involved in co- and posttranslational transport, while others imply that Kar2p is only involved in posttranslational transport of some precursor proteins. Therefore, the questions arise whether Kar2p or Lhs1p are essential components of the protein translocase, acting at the level of co- and posttranslational membrane insertion, and whether their involvement in completion of translocation can be explained by the kinetic argument. An important observation on protein translocation into mammalian microsomes may provide an explanation for these conflicting data with respect to completion of translocation. Cotranslational translocation of the secretory protein apolipoprotein B is not continuous due to “pause transfer” sequences (Chuck and Lingappa, 1992; Hedge and Lingappa, 1996). They are responsible for this phenomenon and lead to transient exposure of parts of the nascent polypeptides to the cytosol as well as to rearrangements in the ribosome/ ribosome receptor complex and the protein translocase (Figure 2). It may well be that under conditions of interrupted transport, for which pause transfer sequences may be just one possible reason, cotranslational completion of translocation becomes dependent on trans-acting molecular chaperones (as defined in the legend to Figure 1), and possibly even cis-acting chaperones. Furthermore, even in the case of cotranslational transport, translocation of the last 60 amino acid residues cannot be driven by elongation since they are burried in the ribosome and the translocase, respectively, at the time of termination of translation. Therefore, even under cotranslational conditions, translocation of certain proteins may become dependent on trans-acting molecular chaperones.
Protein transport into and folding within the endoplasmic reticulum
237
6. COVALENT MODIFICATION OF PROTEINS IN THE ER Protein disulfide isomerases (PDI) play important roles for the folding of proteins in the relatively oxidizing milieu of the ER, as discussed in detail by Freedman and
Figure 2 Mechanisms involved in vectorial translocation of polypeptides across the endoplasmic reticulum membrane. The mature polypeptide chain is shown in transit through the translocase.
Molecular chaperones and folding catalysts
238
Klappa (this volume). However, many proteins which are translocated into the ER do not only depend on PDI but require additional covalent modifications to reach their native state. Often, cleavable signal peptides have to be removed from the precursor proteins in order to allow correct folding. Furthermore, many proteins entering the secretory pathway have to be glycosylated. In these cases glycosylation is intimately linked to folding of these proteins (see below). All these covalent modifications typically occur while the polypeptides are still in transit. For example, transfer of the N-glycan core (see below) occurs rapidly on the translocating chain when approximately 10 to 14 amino acid residues following the acceptor site have passed the membrane (Nilsson et al., 1993; Chen et al., 1995). Therefore, the machineries for signal peptide removal (the signal peptidase) and core glycosylation (the oligosaccharyl transferase) have to be in close neighbourhood with the protein translocase (Figure 1). The signal peptidase of higher and lower eukaryotic organisms contains five and four, respectively, different subunits and is not directly involved in protein transport (Baker and Lively, 1987; Greenburg et al., 1989; Shelness et al., 1988; Shelness and Blobel, 1990; Böhni et al., 1988; YaDeau and Blobel, 1989; YaDeau et al., 1991; Kalies and Hartmann, 1996). Two subunits in the mammalian complex have been shown to be highly similar to the yeast Sec11-protein (Table 1). In the mammalian microsomes, ribophorins I and II together with a third protein, OST48, were originally identified as being involved in oligosaccharyl transfer (Kelleher et al., 1992; Crimaudo et al., 1987). The third protein is identical to the yeast wheat germ agglutinin binding protein (Wbp1) that has been identified as the first essential component of oligosaccharyl transferase (te Heesen et al., 1991 and 1992). As of today, the yeast enzyme has been shown to contain six subunits, two of these beeing structurally related to ribophorins I and II (Table 1) (Silberstein et al., 1995, a and b). Asparagine(N)-linked glycosylation of proteins is a complex process extensively reviewed by Kornfeld and Kornfeld (1985). Briefly, this process involves three major steps. First, the oligomeric oligosaccharyl transferase catalyzes the transfer of a complex oligosaccharide (Gluc3-Mann9-GlcNAc2) from its dolichol carrier onto acceptor sites in the translocating polypeptide chain (Figure 3). Acceptor is an asparagine side chain in the N-glycosylation consensus sequence NXS/T (where X is any amino acid except proline). It has to be noted that not all potential N-glycosylation sites are used (see below). Second, the oligosaccharide is subsequently trimmed starting in the ER with the rapid sequential removal of glucose and mannose residues, catalyzed by glucosidase I (removal of the terminal glucose), glucosidase II (removal of the remaining two 1–3-linked glucose residues) and an ER-mannosidase (Table 1). Third, trimming proceeds on Mann8– 7GlcNAc2 structures within the Golgi-compartment where terminal glycosylation also occurs.
7. PROTEIN FOLDING WITHIN THE ER Like in other cellular compartments protein folding in the ER involves the transient interaction of folding polypeptides with a number of different chaperones and
Protein transport into and folding within the endoplasmic reticulum
239
Figure 3 Core oligosaccharide structure cotranslocationally transferred by the oligosaccharyl transferase onto the asparagine-side chain of the Nglycosylation motif N-X-S/T (where X is any amino acid except proline). onto a nascent polypeptide. The arrows indicate the sites of glucosidase action, which can be blocked by deoxynojirimycin (dNM) or castanospermine (CST). Tunicamycin prevents core glycosylation of the polypeptides.
folding catalysts. However, in addition to providing a particular folding environment, including high Ca2+-concentrations (Montero et al., 1995) and relatively oxidizing conditions (Hwang et al., 1992), the ER differs from other protein folding compartments in further respects. First, the ER-specific covalent modifications discussed above have an important impact on the protein folding process. Second, binding to ER-resident proteins prevents further transport of newly synthesized polypeptides unless they reached their mature conformation, a phenomenon described as architectural editing (Klausner, 1989) or quality control (Hurtley and Helenius, 1989). Third, no ER-components similar to the ring-shaped chaperonin subclass of chaperones have been identified so far. Thus, different from other compartments, there seems to be no need for chaperonin-mediated processes. 7.1. The Role of N-Glycans in Protein Folding In recent years, it became evident that N-linked oligosaccharides are important for glycoprotein folding (Helenius et al., 1994; Fiedler and Simons, 1995). One of the first indications for an in vivo role of N-glycans in glycoprotein folding came from experiments using the glycosylation inhibitor tunicamycin that prevents transfer of the oligosaccharide core onto the glycoprotein. This treatment led to aggregation and intracellular retention of non-glycosylated polypeptides (Olden et al., 1982). However, not all glycoproteins depend on their N-linked glycans in the same way. Whereas (to a variable extent) some glycoproteins depend on their glyco-moiety for maturation and ERexport, others are transported even in the absence of glycosylation (Varki, 1993; Helenius
Molecular chaperones and folding catalysts
240
et al., 1994). Apparently, certain glycoproteins depend more on the overall presence of an N-linked glycan than on its precise location within the polypeptide chain. This is supported by studies on the export competence (as an indicator of successful folding) of mutant vesicular stomatitis virus G (VSV-G) proteins: ER-export of VSV-G that lacks both natural N-glycosylation sites can be restored by the introduction of novel sites at distinct locations (Machamer and Rose, 1988a and b). Furthermore, modification of individual N-glycosylation sites in the simian virus 5 hemagglutinin-neuraminidase (HN) protein in combination with the analysis of the reactivity pattern of HN conformationspecific monoclonal antibodies with the mutant HN proteins indicated that the four sugar side chains play different roles. One specific carbohydrate chain is important for promoting the correct folding of HN whereas another is not essential for the initial folding of HN but plays a role in preventing the aggregation of HN oligomers (Ng et al., 1990). More recent data confirm the interdependence between N-linked glycosylation and glycoprotein folding. For instance, human chorionic gonatotropin beta chains that lack both N-linked glycans is retarded in the formation of the last three disulfide bonds (Feng et al., 1995). Conversely, the tissue plasminogen activator (tPA) used one of the three glycosylation sites more efficiently under conditions that prevented in vivo oxidation of thiol groups in the ER (Allen et al., 1995). In isolated microsomes, the same site was less efficiently used under oxidizing conditions, which allowed a more rapid formation of native tPA (Bulleid et al., 1992). These findings suggest that N-glycosylation depends on the accessibility of the acceptor site located in a portion of the nascent polypeptide that can rapidly assume a secondary structure. Studies on carboxypeptidase Y maturation in yeast, including the introduction of novel glycosylation sites within defined regions of this protein, support that N-glycosylation and folding are competitive events (Holst et al., 1996). As discussed by Holst and coworkers, the rapid formation of folding intermediates could render some N-glycosylation acceptor sites transiently inaccessible to the action of the oligosaccharyl transferase which acts for a limited time period during polypeptide chain translocation. As an example, the authors refer to ovalbumin which does not use an acceptor site for glycosylation despite the fact that this site should readily be accessible because it is located at the protein’s surface, as known from the three dimensional structure of the protein. A major key to the understanding of the role of N-glycans in glycoprotein folding was provided by the identification of two ER-resident proteins, UDP-Glc: glycoprotein glucosyltransferase and calnexin (Table 1). These two proteins seem to play a major role in the recognition, marking, and retention in the ER of at least some glycoproteins that have not completed their folding (Figure 4). UDP-Glc: glycoprotein glucosyltransferase is an enzyme capable of glucosylating a trimmed Man7–9-GlcNac2-species but only when this is attached to a polypeptide that is improperly folded (Sousa et al., 1995; Trombetta et al., 1991; Trombetta et al., 1989). The resulting monoglucosylated Nglycan (Glc1Man7–9GlcNAc2) is crucial for the immature glycoprotein to be recognized by the transmembrane protein calnexin (Ahluwalia et al., 1992; Galvin et al., 1992; Wada et al., 1991, Ware et al., 1995). Monoglucosylated
Protein transport into and folding within the endoplasmic reticulum
241
Figure 4 Quality control and retention of glycoproteins in the ER. The core oligosaccharide on the nascent polypeptide is trimmed by the consecutive action of glucosidase (glue.) I and II. The deglucosylated structure is either correctly folded and can proceed along the secretory pathway to the Golgi-compartment or is a substrate for the action of UDP-glc: glycoprotein glucosyltransferase
Molecular chaperones and folding catalysts
242
(UDPglc transferase) if folding is not completed. The resulting monoglusolyated structure is recognized by calnexin (or calreticulin). The glucosidase II-catalyzed removal of the terminal glucose residue could either occur on free substrates (pathway a) or on the calnexinbound form (pathway b). Depending on the state of maturation, the unglucosylated polypeptide will re-enter the reglucosylation/calnexin cycle or be transported out of the ER.
glycan structures associated with glycoproteins also result from the initial oligosaccharide trimming by glucosidase I and II occuring directly after polypeptide core glycosylation (Figure 4). Thus, calnexin substrates are created either by simple glucose trimming or by re-glucosylation of trimmed glycans explaining why calnexin binding to glycoproteins depends on glucose trimming (Hammond et al., 1994). Polypeptides that carry monoglucosylated N-glycan species are prevented from further transport (Labriola et al., 1995) most likely because of association with calnexin (Hebert et al., 1995; Chen et al., 1995). Calnexin was suggested to have chaperone-like functions in glycoprotein maturation (Degen et al., 1992; Degen and Williams, 1991; Hochstenbach et al., 1992; Ou et al., 1993). Once a calnexin/ glycoprotein complex is established, the glycan moiety can be cleaved in vitro by EndoH treatment without leading to dissociation of the complex, which indicates that the N-glycan is not required for maintenance of the complex, and implies that calnexin directly interacts with the polypeptide chain (Ware et al., 1995; Zhang et al., 1995). Given this finding, it seems unlikely that removal of the terminal glucose by glucosidase II is sufficient to induce the in vivo release of glycoproteins from calnexin interaction. So far, it is not clear how glycoproteins are released from calnexin binding and at which time point the terminal glucose residue is removed. If, however, at this stage, the glycoprotein is correctly folded, it could escape reglucosylation and calnexin-binding and proceed in its way to the next export compartment. In contrast, released protein that is still incorrectly folded will be recognized and re-glucosylated by UDP-Glc: glycoprotein glucosyltransferase, which acts as a ‘folding sensor’, and would re-offer the substrate for calnexin binding as was proposed by Helenius and colleagues (Hammond et al., 1994). This quality control mechanism could ensure that only correctly folded glycoproteins proceed in their way to the next export compartment. The soluble ER-protein calreticulin (Fliegel et al., 1989; Smith and Koch, 1989) is similar in structure to calnexin, interacts with glycoproteins in a similar glycandependent fashion, and could thus play a role like calnexin in the retention of monoglucosylated glycoproteins (Peterson et al., 1995). In fact, calreticulin has recently been demonstrated to have binding specificity for monoglucosylated N-glycan species (Spiro et al., 1996). In addition to calnexin and calreticulin, a third ER component was recently shown to specifically interact with glycoproteins in a glucose trimming dependent fashion. The thioredoxin domain-containing protein Erp61 could be crosslinked to glycoproteins and was coprecipitated with calnexin or calreticulin, indicating a combined action of lektinlike and thioredoxin-like chaperones on the same monoglucosylated substrate (Oliver et al., 1997). Interestingly, in Trypanosoma cruzi, the oligosaccharide core transferred onto nascent polypeptides is unglucosylated. Thus, monoglucosylated glycoproteins appearing in this
Protein transport into and folding within the endoplasmic reticulum
243
organism must have originated from glucosylation by UDP-Glc: glycoprotein glucosyltransferase. Indeed, a major proportion of normal proteins are transiently glucosylated (Ganan et al., 1991) and glucosidase inhibition leads to a delay in the secretion of glycoproteins with monoglucosylated N-glycan (Labriola et al., 1995). In mammalian cells, some misfolded proteins were shown to accumulate in a monoglucosylated form (Suh et al., 1989). 7.2. Molecular Chaperones in Protein Folding Protein glycosylation and protein folding in the ER appear to be directly linked processes. This is indicated by the genetic interaction of yeast Wbp1p (Table 1) and the BiP homolog Kar2p, detected in the processing of carboxypeptidase Y (CPY) (te Heesen and Aebi, 1994). The close interrelationship between the two processes is also evident from a phenomenon known as ‘the unfolded protein resonse’ (discussed in more detail in the chapter by Morimoto, see also (Shamu et al., 1994) for a review). The expression of BiP as well as that of other ER-chaperones, which also are known as glucose regulated proteins (Grps) (Lee, 1981; Shiu et al., 1977), is markedly increased e.g. under conditions where glycosylation is inhibited or diglucosylated glycans persist on nascent glycoproteins within the ER. Transcriptional upregulation of BiP expression was first described by Kozutsumi and colleagues as a consequence of the accumulation of misfolded proteins in the ER (Kozutsumi et al., 1988). It is now generally believed that this feed-back regulation mechanism allows the cell to sense misfolded proteins and regulate the amount of chaperones and folding catalysts required in the ER. BiP is involved in the folding of different polypeptides (Blount and Merlie, 1991; Bole et al., 1986; Earl et al., 1991; Hammond and Helenius, 1994; Kim et al., 1992; Knittler and Haas, 1992; Kuznetsov et al., 1994; Li et al., 1993; Machamer et al., 1990; Mulvey and Brown, 1995; Ng et al., 1989; Roux, 1990; Brunke et al., 1996), and functional BiP is required for maturation and export of carboxypeptidase Y (CPY) (Simons et al., 1995). The latter study investigated CPY maturation in a yeast strain that expresses a temperature-sensitive mutant form of Kar2p. Under non-permissive conditions, CPY failed to be oxidized, was aggregated, bound to mutant Kar2p, and remained in the ER. However, it is still unclear how BiP supports protein maturation in the ER. Most evidence was obtained for immunoglobulins. When antibody heavy (H) and light (L) chain polypeptides were screened for internal peptides that have a high propability to bind to BiP (Blond-Elguindi et al., 1993), some of the potential BiP binding sites involved sequences that participate in contact sites between H and L chain (Knarr et al., 1995). Such data would fit with a model in which the subunits complete their folding prior to assembly and bind to BiP until the hydrophobic contact sites are burried inside the assembled antibody (Hammond and Helenius, 1995). However, it appears that Ig chains do not completely fold prior to the assembly with an appropriate partner chain. Ig L chains, for example, are export-incompetent when not allowed to form homodimers or to pair with H chain into an antibody molecule and remain partially folded with only one of two possible internal disulfide bonds being formed (Knittler et al., 1995). Such partially folded Ig L chains are in a stoichiometric complex with BiP (Cremer et al., 1994; Knittler and Haas, 1992). Assembly with IgH chains leads to disulfide bond
Molecular chaperones and folding catalysts
244
formation in the second Ig L chain domain and our data suggest that it is the assembly event which is required for completion of Ig L chain folding, release from BiP binding, and subsequent secretion (Leitzgen et al., 1996). Accordingly, Ig L chains capable of forming homodimers interact with BiP only transiently (Knittler and Haas, 1992). The group of Hendershot established defined mammalian BiP mutants which differed in nucleotide binding properties, ATP hydrolyzing activities, and/or in the ability to undergo an ATP-induced conformational change (Wei et al., 1995). The in vivo analysis of a BiPATPase mutant (that retained its ability to undergo an ATP-induced conformational change) in a transient expression system revealed that co-expressed Ig L chains were prevented from complete folding and remained bound to mutant BiP instead (Hendershot et al., 1996). These findings suggest that the ATPase function of BiP is required for the in vivo release of the BiP-bound molecules and/or for the folding of at least one of the Ig L chain domains. Together with the data obtained on Ig H chains (Kaloff and Haas, 1995), our findings support a model, in which BiP binding preserves antibody subunits in an assembly competent conformation. In cases where polypeptide folding of multi-subunit proteins only proceeds upon successful assembly, BiP could control oligomerization and ER exit of completely assembled molecules by monitoring the folding state of the individual subunits. If degradation of non-secreted polypeptides residing in the ER-lumen also occurs in the cytosol, as suggested by a number of recent findings (Biederer et al., 1996; Hiller et al., 1996; McCracken and Brodsky, 1996; Ward et al., 1995), our data point to a possible role of BiP in mediating the re translocation event (Knittler et al., 1995; Skowronek, Hendershot, and Haas, submitted for publication). Only little is known on the role of additional ER chaperones such as Grp94 (Shiu et al., 1977), a member of the Hsp90 protein family (Mazzarella and Green, 1987; Sorger and Pelham, 1987) or Grp170 (Lin et al., 1993; Chen et al., 1996). Grp170 was found in association with Ig chains (Lin et al., 1993). Grp94 has been described to bind to different proteins, e.g. incompletely glycosylated forms of Herpes simplex virus glycoprotein B (Navarro et al., 1991), unassembled Ig chains (Melnick et al., 1992), MHC class II chains in the absence of the invariant chain (Schaiff et al., 1992), procollagen (Ferreira et al., 1994), or human chorionic gonatotropin subunit (Feng et al., 1995). Interestingly, many of these reports describe the additional interaction of further chaperones and folding catalysts, e.g. BiP, ERp72, a protein containing three thioredoxin-like domains (Mazzarella et al., 1990; Nguyen Van et al., 1989), or Hsp47, a collagen-specific chaperone (Nagata, 1996). It seems therefore reasonable to suggest that, depending on the requirements of the individual substrates, a set of different ERcomponents participates in supporting the folding of proteins in a process controlled by kinetic partitioning. Furthermore, several chaperone species might simultaneously bind to a substrate as suggested by the finding that antibodies against Grp170 and Grp94 reciprocally coprecipitate Grp94 and Grp170 as well as BiP (Lin et al., 1993). Grp94 can directly bind a protein substrate and is not only indirectly linked to it via binding to BiP that in turn binds the ligand. This is indicated by the finding that co-precipitation of Grp94 with Ig H chain does not diminish after ATP-induced dissociation of BiP from the substrate (Cremer et al., 1994). However, ATP did not completely abolish interaction of Grp94 with BiP, indicating the existance of direct BiP/Grp94 complexes. A sequential interaction of BiP and Grp94 in the folding of Ig L chains was proposed by Melnick and
Protein transport into and folding within the endoplasmic reticulum
245
colleagues (Melnick et al., 1994). They described the two chaperones to interact with Ig L chains with different kinetics whereby BiP preferentially binds to incompletely disulfide-bonded Ig L chain species whereas Grp94 exclusively binds to Ig L chains containing the two internal disulfide bonds. Since cytosolic Hsp90 is known to bind to PPIases (Ratajczak and Carrello, 1996; see Fischer and Schmid, this volume), one might speculate that Grp94 interaction with assembled Ig chains reflects the action of an associated PPIase present in the secretory pathway. This is supported by the in vitro finding that prolyl isomerization of Ig chains occurs after antibody assembly (Lilie et al., 1995).
8. CONCLUDING REMARKS Future experiments related to protein transport into the endoplasmic reticulum will have to take into account the quantitative aspects and putative regulatory aspects of transport. Based on the reconstituted system for protein transport into the ER (Nicchitta and Blobel, 1991; Nicchitta et al., 1991), analysis of protein folding within the ER will undoubtedly have to be studied in proteoliposomes which contain components of the folding machinery in addition to the transport machinery. As yet, most of the interactions between folding catalysts respective chaperones and folding polypeptides have been established at the level of native or chemically crosslinked complexes but it is not yet clear whether different chaperones can act simultaneously on the same folding polypeptide. To understand the possibly multiple roles of a given ER-resident protein, it will also be important to determine its in vivo binding parameters for different folding substrates, to analyze the effects of this interaction on the conformation of the substrate, and to determine the conditions that are required for dissociation of the respective chaperone-substrate complex.
9. REFERENCES Adelman, M.R., Sabatini, D.D. and Blobel, G. (1973). Ribosome-membrane interaction. Nondestructive disassembly of rat liver rough microsomes into ribosomes and membranous components. J. Cell Biol , 56 , 206–229. Ahluwalia, N., Bergeron, J.J., Wada, I., Degen, E. and Williams, D.B. (1992). The p88 molecular chaperone is identical to the endoplasmic reticulum membrane protein, calnexin. J. Biol. Chem. , 267 , 10914–10918. Allen, S., Naim, H.Y. and Bulleid, N.J. (1995). Intracellular folding of tissue-type plasminogen activator. J. Biol. Chem. , 270 , 4797–4804. Amaya, Y., Nakano, A., Ito, K. and Mori, M. (1990). Isolation of a yeat gene, SRH1, that encodes a homologue of the 54k subunit of mammalian signal recognition particle. J. Biochem. , 107 , 457–463. Bacher, G., Lütcke, H., Jungnickel, B., Rapoport, T.A. and Dobberstein, B. (1996). Regulation by the ribosome of the GTPase of the signal-recognition particle during protein targeting. Nature , 381 , 248–251. Baker, R.K. and Lively, M.O. (1987). Purification and characterization of hen oviduct microsomal signal peptidase. Biochemistry , 26 , 8561–8567.
Molecular chaperones and folding catalysts
246
Biederer, T., Volkwein, C. and Sommer, T. (1996). Degradation of subunits of the Sec61p complex, an integral component of the ER membrane, by the ubiquitin proteasome pathway. EMBO J. , 15 , 2069–2076. Blond-Elguindi, S., Cwirla, S.E., Dower, W.J., Lipshutz, R.J., Sprang, S.R., Sambrook, J.F. and Gething, M.-J. (1993). Affinity panning of a library of peptides displayed on bacteriophages reveals the binding specificity of BiP. Cell , 75 , 717–728. Blount, P. and Merlie, J.P. (1991). BIP associates with newly synthesized subunits of the mouse muscle nicotinic receptor. J. Cell Biol , 113 , 1125–1132. Böhni, P.C., Deshaies, R.J. and Schekman, R.W. (1988). Sec11 is required for signal peptide processing and yeast cell growth. J. Cell Biol , 106 , 1035–1042. Bole, D.G., Hendershot, L.M. and Kearney, J.F. (1986). Posttranslational association of immunoglobulin heavy chain binding protein with nascent heavy chains in nonsecreting and secreting hybridomas. J. Cell Biol. , 102 , 1558–1566. Borel, A.C. and Simon, S.M. (1996). Biogenesis of polytopic membrane proteins: membrane segments assemble within translocation channels prior to mebrane integration. Cell , 85 , 379–389. Borgese, N., Mok, W., Kreibich, G. and Sabatini, D.D. (1974). Ribosome-membrane interaction: In vitro binding of ribosomes to microsomal membranes. J. Mol. Biol. , 88 , 559–580. Brightman, S.E., Blatch, G.L. and Zetter, B.R. (1995). Isolation of a mouse cDNA encoding MTJ1, a new murine member of the DnaJ family of proteins. Gene , 153 , 249–254. Brodsky, J.L., Goeckeler, J. and Schekman, R. (1995). BiP and Sec63p are required for both co- and posttranslational protein translocation into the yeast endoplasmic reticulum. Proc. Natl. Acad. Sci. USA , 92 , 9643–9646. Brodsky, J.L., Hamamoto, S., Feldheim, D. and Schekman, R. (1993). Reconstitution of protein translocation from solubilized membranes reveals topologically distinct roles for BiP and cytosolic Hsc70. J. Cell Biol , 120 , 95–102. Brunke, M., Dierks, T., Schlotterhose, P., Escher, A., Schmidt, B., Szalay, A.A., Lechte, M., Sandholzer, U. und Zimmermann, R. (1996). Luciferase assembly after transport into mammalian microsomes involves molecular chaperones and peptidylprolyl cis/trans isomerases. J. Biol Chem. , 271 , 23487–23494. Bulleid, N.J. and Freedman, R.B. (1988). Defective co-translational formation of disulphide bonds in protein disulphide-isomerase-deficient microsomes. Nature , 335 , 649–651. Bulleid, N.J., Bassel-Duby, R.S., Freedman, R.B., Sambrook, J.F. and Gething, M.-J.H. (1992). Cell-free synthesis of enzymically active tissue-type plasminogen activator. Biochem. J. , 286 , 275–280. Caplan, A.J. and Douglas, M.G. (1991). Characterization of YDJ1: a yeast homologue of the bacterial dnaJ protein. J. Cell Biol , 114 , 609–621. Caplan, A.J., Cyr, D.M. and Douglas, M.G. (1992). YDJ1p facilitates polypeptide translocation across different intracellular membranes by a conserved mechanism. Cell , 71 , 1143–1155. Chen, W., Helenius, J., Braakman, I. and Helenius, A. (1995). Cotranslational folding and calnexin binding during glycoprotein synthesis. Proc. Natl Acad. Sci. USA , 92 , 6229–6233. Chen, X., Easton, D., Oh, H.-J., Lee-Yon, D.-S., Liu, X. and Subjeck, J. (1996). The 170 kDa glucose regulated stress protein is a large HSP70-, HSP110-like protein of the endoplasmic reticulum. FEBS Lett. , 380 , 68–72.
Protein transport into and folding within the endoplasmic reticulum
247
Chirico, W.J., Waters, G.M. and Blobel, G. (1988). 70K heat shock related proteins stimulate protein translocation into microsomes. Nature , 332 , 805–810. Chuck, S.L. and Lingappa, V.R. (1992). Pause transfer: a topogenic sequence in apolipoprotein B mediates stopping and restarting of Translocation. Cell , 68 , 9–21. Connolly, T. and Gilmore, R. (1986). Formation of a functional ribosome-membrane junction during translocation requires the participation of a GTP-binding protein. J. Cell Biol , 103 , 2253–2261. Connolly, T. and Gilmore, R. (1989). The signal recognition particle receptor mediates the GTP-dependent displacement of SRP from the signal sequence of the nascent polypeptide. Cell , 57 , 599–610. Connolly, T., Rapiejko, P.J. and Gilmore, R. (1991). Requirement of GTP hydrolysis for dissociation of the signal recognition particle from its receptor. Science , 252 , 1171– 1173. Craven, R.A., Egerton, M. and Stirling, C.J. (1996). A novel Hsp70 of the yeast ER lumen is required for the efficient translocation of a number of protein precursors. EMBO J. , 15 , 2640–2650. Cremer, A., Knittler, M.R. and Haas, I.G. (1994). Biosynthesis of antibodies and molecular chaperones. In F.Wieland and W.Reutter, (eds.), Glyco- and Cellbiology , Springer-Verlag Berlin Heidelberg New York, pp. 171–184. Crimaudo, C., Hortsch, M., Gausepohl, H., Meyer, D.I. (1987). Human ribophorins I and II: the primary structure and membrane topology of two highly conserved rough endoplasmic reticulum-specific glycoproteins. EMBO J. , 6 , 75–82. Crowley, K.S., Lia, S., Worrell, V.E., Reinhart, G.D. and Johnson, A.E. (1993a). Secretory proteins move through the endoplasmic reticulum membrane via an aqueous, gated pore. Cell , 78 , 461–471. Crowley, K.S., Reinhart, G.D. and Johnson, A.E. (1993b). The signal sequence moves through a ribosomal tunnel into a noncytoplasmic aqueous environment at the ER membrane early in translocation. Cell , 73 , 1101–1115. Daimon, M., Susa, S., Suzuki, K., Kato, K., Yamatani, K. and Sasaki, H. (1997). Identification of a human cDNA homologue to the Drosophila translocation protein 1 (Dtrp1). Biochem. Biophys. Res. Commun. , 230 , 100–104. Degen, E. and Williams, D.B. (1991). Participation of a novel 88-kD protein in the biogenesis of murine class I histocompatibility molecules. J. Cell Biol , 112 , 1099– 1115. Degen, E., Cohen-Doyle, M.F. and Williams, D.B. (1992). Efficient dissociation of the p88 chaperone from major histocompatibility complex class I molecules requires both beta 2-microglobulin and peptide. J. Exp. Med. , 175 , 1653–1661. Deshaies, R.J. and Schekman, R. (1987). A yeast mutant defective at an early stage in import of secretory protein precursors into the endoplasmic reticulum. J. Cell Biol. , 105 , 633–645. Deshaies, R.J. and Schekman, R. (1989). Sec62 encodes a putative membrane protein required for protein translocation into the yeast endoplasmic reticulum. J. Cell Biol. , 109 , 2653–2664. Deshaies, R.J., Koch, B.D., Werner-Washburne, M., Craig, E.A. and Schekman, R. (1988). A subfamily of stress proteins facilitates translocation of secretory and mitochondrial precursor polypeptides. Nature , 332 , 800–805. Deshaies, R.J., Sanders, S.L., Feldheim, D.A. and Schekman, R. (1991). Assembly of yeast Sec proteins involved in translocation into the endoplasmic reticulum into a membrane-bound multisubunit complex. Nature , 349 , 806–808.
Molecular chaperones and folding catalysts
248
Dierks, T., Volkmer, J., Schlenstedt, G., Jung, C., Sandholzer, U., Zachmann, K., Schlotterhose, P., Neifer, K., Schmidt, B. and Zimmermann, R. (1996). A microsomal ATP-binding protein involved in efficient protein transport into the mammalian endoplasmic reticulum. EMBO J. , 15 , 6931–6942. Do, H., Falcone, D., Lin, J., Andrews, D.W. and Johnson, A.E. (1996). The cotranslational integration of membrane proteins into the phospholipid bilayer is a multistep process. Cell , 85 , 369–378. Earl, P.L., Moss, B. and Doms, R.W. (1991). Folding, interaction with GRP78-BiP, assembly and transport of the human immunodeficiency virus type 1 envelope protein. J. Virol. , 65 , 2047–2055. Feldheim, D., Rothblatt, J. and Schekman, R. (1992). Topology and functional domains of sec63p, an endoplasmic reticulum membrane protein required for secretory protein translocation. Mol. Cell. Biol. , 12 , 3288–3296. Feng, W., Matzuk, M.M., Mountjoy, K., Bedows, E., Ruddon, R.W. and Boime, I. (1995). The asparaginelinked oligosaccharides of the human chorionic gonadotropin beta subunit facilitate correct disulfide bond pairing. J. Biol. Chem. , 270 , 11851– 11859. Ferreira, L.R., Norris, K., Smith, T., Hebert, C. and Sauk, J.J. (1994). Association of Hsp47, Grp78 and Grp94 with procollagen supports the successive or coupled action of molecular chaperones. J. Cell. Biochem. , 56 , 518–526. Fiedler, K. and Simons, K (1995). The Role of N-glycans in the secretory pathway. Cell , 81 , 309–312. Finke, K., Plath, K., Panzner, S., Prehn, S., Rapoport, T.A., Hartmann, E. and Sommer, T. (1996). A second trimeric complex containing homologs of the Sec61p complex functions in protein transport across the ER membrane of S. cerevisiae. EMBO J. , 15 , 1482–1494. Fliegel, L., Burns, K., MacLennan, D.H., Reithmeier, R.A. and Michalak, M. (1989). Molecular cloning of the high affinity calcium-binding protein (calreticulin) of skeletal muscle sarcoplasmic reticulum. J. Biol. Chem. , 264 , 21522–21528. Freeman, B.C. and Morimoto, R.I. (1996). The human cystosolic molecular chaperones hsp90, hsp70 (hsc70) and hdj-1 have distinct roles in recognition of a non-native protein and protein folding. EMBO J. , 15 , 2969–2979. Galvin, K., Krishna, S., Ponchel, F., Frohlich, M., Cummings, D.E., Carlson, R., Wands, J.R., Isselbacher, K.J., Pillai, S. and Ozturk, M. (1992). The major histocompatibility complex class I antigen-binding protein p88 is the product of the calnexin gene. Proc. Natl. Acad. Set. USA , 89 , 8452–8456. Ganan, S., Cazzulo, J.J. and Parodi, A.J. (1991). A major proportion of N-glycoproteins are transiently glucosylated in the endoplasmic reticulum. Biochemistry , 30 , 3098– 3104. Gilmore, R. and Blobel, G. (1985). Translocation of secretory proteins across the microsomal membrane occurs through an environment accessible to aqueous perturbants. Cell , 42 , 497–505. Gilmore, R., Blobel, G. and Walter, P. (1982a). Protein translocation across the endoplamsic reticulum. I. Detection in the microsomal membrane of a receptor for the signal recognition particle. J. Cell Biol. , 95 , 463–469. Görlich, D., Hartmann, E., Prehn, S. and Rapoport, T.A. (1992a). A protein of the endoplasmic reticulum involved early in polypeptide translocation. Nature , 357 , 47– 52. Gilmore, R., Walter, P. and Blobel, G. (1982b). Protein translocation across the
Protein transport into and folding within the endoplasmic reticulum
249
endoplasmic reticulum. II. Isolation and characterization of the signal recognition particle receptor. J. Cell Biol. , 95 , 470–477. Görlich, D. and Rapoport, T.A. (1993). Protein translocation into proteoliposomes reconstituted from purified components of the endoplasmic reticulum membrane. Cell , 75 , 615–630. Görlich, D., Prehn, S., Hartmann, E., Kalies, K.-U. and Rapoport, T.A. (1992b). A mammalian homolog of sec61p and secYp is associated with ribosomes and nascent polypeptides during translocation. Cell , 71 , 489–503. Green, N., Fang, H. and Walter, P. (1992). Mutants in three novel complementation groups inhibit membrane protein insertion into and soluble protein translocation across the endoplasmic reticulum membrane of Saccharomyces cerevisiae. J. Cell Biol. , 116 , 597–604. Greenburg, G., Shelness, G.S. and Blobel, G. (1989). A subunit of mammalian signal peptidase is homologous to yeast sec11 protein. J. Biol. Chem. , 264 , 15762–15765. Gross, M. and Hessefort, S. (1996). Purification and characterization of a 66-kDa protein from rabbit reticulocyte lysate which promotes the recycling of Hsp70. J. Biol. Chem. , 271 , 16833–16841. Hammond, C. and Helenius, A. (1994). Folding of VSV G protein: sequential interaction with BiP and calnexin. Science , 266 , 456–458. Hammond, C. and Helenius, A. (1995). Quality control in the secretory pathway. Curr. Opin. Cell Biol. , 7 , 523–529. Hammond, C., Braakman, I. and Helenius, A. (1994). Role of N-linked oligosaccharide recognition, glucose trimming and calnexin in glycoprotein folding and quality control. Proc. Natl Acad. Sci. USA , 91 , 913–917. Hann, B.C. and Walter, P. (1991). The signal recognition particle in S. cerevisiae. Cell , 67 , 131–144. Hann, B.C., Poritz, M.A. and Walter, P. (1989). Saccharomyces cerevisiae and Schizosaccharomyces pombe contain a homologue to the 54-kD subunit of the signal recognition particle that in S. cerevisiae is essential for growth. J. Cell Biol. , 109 , 3223–3235. Hann, B.C., Stirling, C.J. and Walter, P. (1992). Sec65 gene product is a subunit of the yeast signal recognition particle required for its integrity. Nature , 356 , 532–533. Hansen, W., Garcia, P.D. and Walter, P. (1986). In vitro protein translocation across the yeast endoplasmic reticulum: ATP-dependent post-translational translocation of the prepro-a-factor. Cell , 45 , 397–406. Hardy, S.J.S. and Randall, L.L. (1991). A kinetic partitioning model of selective binding of nonnative proteins by the bacterial chaperone SecB. Science , 251 , 439–443. Hartmann, E., Sommer, T., Prehn, S., Görlich, D., Jentsch, S. and Rapoport, T.A. (1994). Evolutionary conservation of components of the protein translocation complex. Nature , 367 , 654–657. Hebert, D.N., Foellmer, B. and Helenius, A. (1995). Glucose trimming and relucosylation determine glycoprotein association with calnexin in the endoplasmic reticulum. Cell , 81 , 425–433. Hedge, R.S. and Lingappa, V.R. (1996). Sequence-specific alteration of the ribosomemembrane junction exposes nascent secretory proteins to the cytosol. Cell , 85 , 217– 228. Helenius, A. (1994). How N-linked oligosaccharides affect glycoprotein folding in the endoplasmic reticulum. Mol. Biol. Cell , 5 , 253–265. Hendershot, L., Wei, J., Gaut, J., Melnick, J., Aviel, S. and Argon, Y. (1996). Inhibition
Molecular chaperones and folding catalysts
250
of immunoglobulin folding and secretion by dominant negative BiP ATPase mutants. Proc. Natl. Acad. Sci. USA , 93 , 5269–5274. High, S., Andersen, S.S.L., Görlich, D., Hartmann, E., Prehn, S., Rapoport, T.A. and Dobberstein, B. (1993). Sec61p is adjacent to nascent type I and type II signal-anchor proteins during their membrane insertion. J. Cell Biol. , 121 , 743–750. High, S., Görlich, D., Wiedmann, M., Rapoport, T.A. and Dobberstein, B. (1991). The identification of proteins in the proximity of signal-anchor sequences during their targeting to and insertion into the membrane of the ER. J. Cell Biol. , 113 , 35–44. Hiller, M.M., Finger, A., Schweiger, M. and Wolf, D.H. (1996). ER degradation of a misfolded luminal protein by the cytosolic ubiquitin-proteasome pathway. Science , 273 , 1725–1728. Hochstenbach, F., David, V., Watkins, S. and Brenner, M.B. (1992). Endoplasmic reticulum resident protein of 90 kilodaltons associates with the T- and B-cell antigen receptors and major histocompatibility complex antigens during their assembly. Proc. Natl. Acad. Sci. USA , 89 , 4734–4738. Höhfeld, J., Minami, Y. and Hartl, F.-U. (1995). Hip, a novel cochaperone involved in the eukaryotic Hsc70/Hsp40 reaction cycle. Cell , 83 , 589–598. Holst, B., Bruun, A.W., Kielland-Brandt, M.C. and Winther, J.R. (1996). Competition between folding and glycosylation in the endoplasmic reticulum. EMBO J. , 15 , 3538– 3546. Hurtley, S.M. and Helenius, A. (1989). Protein oligomerization in the endoplasmic reticulum. Annu. Rev. Cell Biol , 5 , 277–307. Hwang, C., Sinskey, A.J. and Lodish, H.F. (1992). Oxidized redox state of glutathione in the endoplasmic reticulum. Science , 257 , 1496–1502. Jungnickel, B. and Rapoport, T.A. (1995). A posttargeting signal sequence recognition event in the endoplasmic reticulum membrane. Cell , 82 , 261–270. Kalies, K.-U. and Hartmann, E. (1996). Membrane topology of the 12- and 25-kDa subunits of the mammalian signal peptidase complex. J. Biol Chem. , 271 , 3925–3929. Kaloff, C.R. and Haas, I.G. (1995). Coordination of immunoglobulin chain folding and immunoglobulin chain assembly is essential for the formation of functional IgG. Immunity , 2 , 629–637. Kellaris, K.V., Bowen, S. and Gilmore, R. (1991). ER translocation intermediates are adjacent to a nonglycosylated 34-kD integral membrane protein. J. Cell Biol. , 114 , 21–33. Kelleher, D.J., Kreibich, G. and Gilmore, R. (1992). Oligosaccharytransferase activity is associated with a protein complex composed of ribophorins I and II and a 48 kd protein. Cell , 69 , 55–65. Kim, P.S., Bole, D. and Arvan, P. (1992). Transient aggregation of nascent thyroglobulin in the endoplasmic reticulum: relationship to the molecular chaperone, BiP. J. Cell Biol. , 118 , 541–549. Klappa, P., Freedman, R. and Zimmermann, R. (1995). Protein disulfide isomerase and a lumenal cyclophilin-type peptidyl prolyl cis-trans isomerase are in transient contact with secretory proteins during late stages of translocation. Eur. J. Biochem. , 232 , 755–764. Klappa, P., Mayinger, P., Pipkorn, R., Zimmermann, M. and Zimmermann, R. (1991). A microsomal protein is involved in ATP-dependent transport of presecretory proteins into mammalian microsomes. EMBO J. , 10 , 2795–2803. Klappa, P., Zimmermann, M. and Zimmermann, R. (1994). The membrane proteins TRAMp and Sec61ap may be involved in posttransaltional transport of presecretory
Protein transport into and folding within the endoplasmic reticulum
251
proteins into mammalian microsomes. FEBS Lett. , 341 , 281–287. Klausner, R.D. (1989). Architectural editing: determining the fate of newly synthesized membrane proteins. New Biol. , 1 , 3–8. Knarr, G., Gething, M.J., Modrow, S. and Buchner, J. (1995). BiP binding sequences in antibodies. J. Biol. Chem. , 270 , 27589–27594. Knittler, M.R. and Haas, I.G. (1992). Interaction of BiP with newly synthesized immunoglobulin light chain molecules: cycles of sequential binding and release. EMBO J. , 11 , 1573–1581. Knittler, M.R., Dirks, S. and Haas, I.G. (1995). Molecular chaperones involved in protein degradation in the endoplasmic reticulum: Quantitative interaction of the heat shock cognate BiP with partially folded immunoglobulin light chains that are degraded in the endoplasmic reticulum. Proc. Natl Acad. Sci. USA , 92 , 1764–1768. Kornfeld, R. and Kornfeld, S. (1985). Assembly of asparagine-linked oligosaccharides. Ann. Rev. Biochem. , 54 , 631–664. Kozutsumi, Y., Segal, M., Normington, K., Gething, M.J. and Sambrook, J. (1988). The presence of malfolded proteins in the endoplasmic reticulum signals the induction of glucose-regulated proteins. Nature , 332 , 462–464. Krieg, U.C., Johnson, A.E. and Walter, P. (1989). Protein translocation across the endoplasmic reticulum membrane: Identification by photocross-linking of a 39-kD integral membrane glycoprotein as part of a putative translocation tunnel. J. Cell Biol , 109 , 2033–2043. Krieg, U.C., Walter, P. and Johnson, A.E. (1986). Photocrosslinking of the signal sequence of nascent preprolactin to the 54-kilodalton polypeptide of the signal recognition particle. Proc. Natl. Acad. Sci. USA , 83 , 8604–8608. Kurzchalia, T.V., Wiedmann, M., Girshovich, A.S., Bochkareva, E.S., Bielka, H. and Rapoport, T.A. (1986). The signal sequence of nascent preprolactin interacts with the 54K polypeptide of the signal recognition particle. Nature , 320 , 634–636. Kuznetsov, G., Chen, L.B. and Nigam, S.K. (1994). Several endoplasmic reticulum stress proteins, including ERp72, interact with thyroglobulin during its maturation. J. Biol. Chem. , 269 , 22990–22995. Labriola, C., Cazzulo, J.J. and Parodi, A.J. (1995). Retention of glucose units added by the UDPGlc: glycoprotein glucosyltransferase delays exit of glycoproteins from the endoplasmic reticulum. J. Cell Biol. , 130 , 771–779. Lauffer, L., Garcia, P.D., Harkins, R.N., Coussens, L., Ullrich, A. and Walter, P. (1985). Topology of signal recognition particle receptor in endoplasmic reticulum membrane. Nature , 318 , 334–338. Lauring, B., Kreibich, G. and Wiedmann, M. (1995a). The intrinsic ability of ribosomes to bind to endoplasmic reticulum membranes is regulated by signal recognition particle and nascent-polypeptide-associated complex. Proc. Natl. Acad. Sci. USA , 95 , 9435– 9439. Lauring, B., Sakai, H., Kreibich, G. and Wiedmann, M. (1995b). Nascent polypeptideassociated complex protein prevents mistargeting of nascent chains to the endoplasmic reticulum. Proc. Natl. Acad. Sci. USA , 92 , 5411–5415. Lee, A.S. (1981). The accumulation of three specific proteins related to glucose-regulated proteins in a temperature-sensitive hamster mutant cell line K12. J. Cell. Physiol , 106 , 119–125. Leitzgen, K., Knittler, M.R. and Haas, I.G. (1997). Assembly of immunoglobulin light chains as a prerequisite for secretion: A model for oligomerization-dependent subunit folding. J. Biol. Chem. , 272 , 3117–3123.
Molecular chaperones and folding catalysts
252
Li, X., Wu, Y., Zhang, D.Z., Gillikin, J.W., Boston, R.S., Franceschi, V.R. and Okita, T.W. (1993). Rice prolamine protein body biogenesis: a BiP-mediated process. Science , 262 , 1054–1056. Lilie, H., Rudolph, R. and Buchner, J. (1995). Association of antibody chains at different stages of folding: prolyl isomerization occurs after formation of quaternary structure . J. Mol. Biol. , 248 , 190–201. Lin, H.-Y., Masso-Welch, P., Di, Y.-P., Cai, J.-W., Shen, J.-W. and Subjeck, J.R. (1993). The 170 kDa glucose-regulated stress protein is an endoplasmic reticulum protein that binds immunoglobulin. Mol. Biol. Cell , 4 , 1109–1119. Lütcke, H., High, S., Romisch, K., Ashford, A.J. and Dobberstein, B. (1992). The methionine-rich domain of the 54 kDa subunit of signal recognition particle is sufficient for the interaction with signal sequences. EMBO J. , 11 , 1543–1551. Lyman, S.K. and Schekman, R. (1995). Interaction between BiP and Sec63p is required for the completion of protein translocation into the ER of Saccharomyces cerevisiae. J. Cell Biol. , 131 , 1163–1171. Machamer, C.E. and Rose, J.K. (1988a). Influence of new glycosylation sites on expression of the vesicular stomatitits virus G protein at the plasma membrane. J. Biol. Chem. , 263 , 5948–5954. Machamer, C.E. and Rose, J.K. (1988b). Vesicular stomatitis virus G proteins with altered glycosylation sites display temperature-sensitive intracellular transport and are subject to aberrant intermolecular disulfide bonding. J. Biol. Chem. , 263 , 5955–5960. Machamer, C.E., Doms, R.W., Bole, D.G., Helenius, A. and Rose, J.K (1990). Heavy chain binding protein recognizes incompletely disulfide-bonded forms of vesicular stomatitis virus G protein. J. Biol. Chem. , 265 , 6879–6883. Martoglio, B., Hofmann, M.W., Brunner, J. and Dobberstein, B. (1995). The proteinconducting channel in the membrane of the endoplasmic reticulum is open laterally toward the lipid bilayer. Cell , 81 , 207–214. Mazzarella, R.A. and Green, M. (1987). ERp99, an abundant, conserved glycoprotein of the endoplasmic reticulum, is homologous to the 90-kDa heat shock protein (hsp90) and the 94-kDa glucose regulated protein (GRP94). J. Biol. Chem. , 262 , 8875–8883. Mazzarella, R.A., Srinivasan, M., Haugejorden, S.M. and Green, M. (1990). ERp72, an abundant luminal endoplasmic reticulum protein, contains three copies of the active site sequences of protein disulfide isomerase. J. Biol. Chem. , 265 , 1094–1101. McCracken, A.A. and Brodsky, J.L. (1996). Assembly of ER-associated protein degradation in vitro: dependence on cytosol, calnexin and ATP. J. Cell Biol. , 132 , 291–298. Melnick, J., Aviel, S. and Argon, Y. (1992). The endoplasmic reticulum stress protein GRP94, in addition to BiP, associates with unassembled immunoglobulin chains. J. Biol. Chem. , 267 , 21303–21806. Melnick, J., Dul, J.L. and Argon, Y. (1994). Sequential interaction of the chaperones BiP and GRP94 with immunoglobulin chains in the endoplasmic reticulum. Nature , 370 , 373–375. Meyer, D.I. and Dobberstein, B. (1980a). A membrane component essential for vectorial translocation of nascent proteins across the endoplasmic reticulum: Requirements for its extraction and reassociation with the membrane. J. Cell Biol. , 87 , 498–502. Meyer, D.I. and Dobberstein, B. (1980b). Identification and characterization of a membrane component essential for the translocation of nascent proteins across the membrane of the endoplasmic reticulum. J. Cell Biol. , 87 , 503–508. Meyer, D.I., Krause, E. and Dobberstein, B. (1982). Secretory protein translocation
Protein transport into and folding within the endoplasmic reticulum
253
across membranes—the role of the ‘docking protein’. Nature , 297 , 647–650. Montero, M., Brini, M., Marsault, R., Alvarez, J., Sitia, R., Pozzan, T. and Rizzuto, R. (1995). Monitoring dynamic changes in free Ca2+ concentration in the endoplasmic reticulum of intact cells. EMBO J. , 14 , 5467–5475. Mūller, G. and Zimmermann, R. (1987). Import of honeybee prepromelittin into the endoplasmic reticulum: Structural basis for independence of SRP and docking protein. EMBO J. , 6 , 2099–2107. Müsch, A., Wiedmann, M. and Rapoport, T.A. (1992). Yeast Sec proteins interact with polypeptides traversing the endoplasmic reticulum membrane. Cell , 69 , 343–352. Mulvey, M. and Brown, D.T. (1995). Involvement of the molecular chaperone BiP in maturation of Sindbis virus envelope glycoproteins. J. Virol , 69 , 1621–1627. Nagata, K. (1996). Hsp47: a Collagen Specific Molecular Chaperone. Trends Biochem . Set., 21 , 23–26. Navarro, D., Qadri, I. and Pereira, L. (1991). A mutation in the ectodomain of herpes simplex virus 1 glycoprotein B causes defective processing and retention in the endoplasmic reticulum. Virology , 184 , 253–264. Ng, D.T.W., Brown, J.D. and Walter, P. (1996). Signal sequences specify the targeting route to the endoplasmic reticulum membrane. J. Cell Biol. , 134 , 269–278. Ng, D.T, Hiebert, S.W. and Lamb, R.A. (1990). Different roles of individual N-linked oligosaccharide chains in folding, assembly and transport of the simian virus 5 hemagglutinin-neuraminidase. Mol. Cell. Biol. , 10 , 1989–2001. Ng, D.T, Randall, R.E. and Lamb, R.A. (1989). Intracellular maturation and transport of the SV5 type II glycoprotein hemagglutinin-neuraminidase: specific and transient association with GRP78-BiP in the endoplasmic reticulum and extensive internalization from the cell surface. J. Cell Biol. , 109 , 3273–3289. Nguyen, T.H., Law, D.T. S. and Williams, D.B. (1991). Binding protein BiP is required for translocation of secretory proteins into the endoplasmic reticulum in Saccharomyces cerevisiae. Proc. Natl. Acad. Set. USA , 88 , 1565–1569. Nguyen Van, P., Peter, F. and Söling, H.-D. (1989). Four intracisternal Calcium-binding glycoproteins from rat liver microsomes with high affinity for Calcium. J. Biol. Chem. , 264 , 17494–17501. Nicchitta, C.V. and Blobel, G. (1990). Assembly of translocation-competent proteoliposomes from detergent-solubilized rough microsomes. Cell , 60 , 259–266. Nicchitta, C.V. and Blobel, G. (1993). Lumenal proteins of the mammalian endoplasmic reticulum are required to complete protein translocation. Cell , 73 , 989–998. Nicchitta, C.V., Migliaccio, G. and Blobel, G. (1991). Biochemical fractionation and assembly of the membrane components that mediate nascent chain targeting and translocation. Cell , 65 , 587–598. Nilsson, I.M. and von Heijne, G. (1993). Determination of the distance between the oligosaccharyltransferase active site and the endoplasmic reticulum membrane. J. Biol. Chem. , 268 , 5798–5801. Olden, K., Parent, J.B. and White, S.L. (1982). Carbohydrate moieties of glycoproteins a re-evaluation of their function. Biochem. Biophys. Acta , 650 , 209–232. Oliver, J.D., Vanderwal, F.J., Bulleid, N.J., and High, S. (1997). Interaction of the thiol dependent reductase Erp57 with nascent glycoproteins. Science , 275 , 86–88. Ou, W.J., Cameron, P.H., Thomas, D.Y. and Bergeron, J.J. (1993). Association of folding intermediates of glycoproteins with calnexin during protein maturation. Nature , 364 , 771–776. Panzner, S., Dreier, L., Hartmann, E., Kostka, S. and Rapoport, T.A. (1995).
Molecular chaperones and folding catalysts
254
Posttranslational protein transport in yeast reconstituted with a purified complex of Sec proteins and Kar2p. Cell , 81 , 561–570. Perara, E., Rothman, R.E. and Lingappa, V.R. (1986). Uncoupling translocation from translation: Implications for transport of proteins across membranes. Science , 232 , 348–352. Perlman, D. and Halvorson, H.O. (1983). A putative signal peptidase recognition site and sequence in eucaryotic and procaryotic signal peptides. J. Mol. Biol , 167 , 391–409. Peterson, J.R., Ora, A., Van, P.N. and Helenius, A. (1995). Transient, lectin like association of calreticulin with folding intermediates of cellular and viral glycoproteins. Mol. Biol. Cell , 6 , 1173–1184. Poritz, M.A., Siegel, V., Hansen, W. and Walter, P. (1988). Small ribonucleoproteins in Schizosaccharomyces pombe and Yarrowia lipolytica homologous to signal recognition particle. Proc. Natl. Acad. Sci. USA , 85 , 4315–4319. Rapiejko, P.J. and Gilmore, R. (1992). Protein translocation across the ER requires a functional GTP binding site in the a subunit of the signal recognition particle receptor. J. Cell Biol. , 117 , 493–503. Ratajczak, T. and Carrello, A. (1996). Cyclophilin 40 (CyP-40), mapping of its hsp90 binding domain and evidence that FKBP52 competes with CyP-40 for hsp90 binding . J. Biol. Chem. , 271 , 2961–2965. Ribes, V., Dehaux, P. and Tollervey, D. (1988). 7SL RNA from Schizosaccharomyces pombe is encoded by a single copy essential gene. EMBO J. , 7 , 231–237. Römisch, K., Webb, J., Lingelbach, K., Gausepohl, H. and Dobberstein, B. (1990). The 54-kD protein of signal recognition particle contains a methionine-rich RNA binding domain. J. Cell Biol. , 111 , 1793–1802. Rothblatt, J.A. and Meyer, D.I. (1986). Secretion in yeast: Translocation and glycosylation of prepro- -factor in vitro can occur via an ATP-dependent posttranslational mechanism. EMBO J. , 5 , 1031–1036. Rothblatt, J.A., Deshaies, R.J., Sanders, S.L., Daum, G. and Schekman, R. (1989). Multiple genes are required for proper insertion of secretory proteins into the endoplasmic reticulum. J. Cell Biol. , 109 , 2641–2652. Roux, L. (1990). Selective and transient association of Sendai virus HN glycoprotein with BiP. Virology , 175 , 161–166. Sadler, I., Chiang, A., Kurihara, T., Rothblatt, J., Way, J. and Silver, P. (1989). A yeast gene important for protein assembly into the endoplasmic reticulum and the nucleus has homology to DnaJ, an Escherichia coli heat shock protein. J. Cell Biol. , 109 , 2665–2675. Sanders, S.L., Whitfield, K.M., Vogel, J.P., Rose, M.D. and Schekman, R. (1992). Sec61p and BiP directly facilitate polypeptide translocation into the ER. Cell , 69 , 353–365. Sanz, P. and Meyer, D.I. (1989). Secretion in yeast: Preprotein binding to a membrane receptor and ATP-dependent translocation are sequential and separable events in vitro. J. Cell Biol. , 108 , 2101–2106. Savitz, A.J. and Meyer, D.I. (1990). Identification of a ribosome receptor in the rough endoplasmic reticulum. Nature , 346 , 540–544. Savitz, A.J. and Meyer, D.I. (1993). 180-kDa ribosome receptor is essential for both ribosome binding and protein translocation. J. Cell Biol. , 120 , 853–863. Schaiff, W.T., Hruska, K.J., McCourt, D.W., Green, M. and Schwartz, B.D. (1992). HLA-DR associates with specific stress proteins and is retained in the endoplasmic reticulum in invariant chain negative cells . J. Exp. Med. , 176 , 657–666.
Protein transport into and folding within the endoplasmic reticulum
255
Schlenstedt, G. and Zimmermann, R. (1987). Import of frog prepropeptide GLa into microsomes requires ATP but does not involve docking protein or ribosomes. EMBO J. , 6 , 699–703. Schlenstedt, G., Gudmundsson, G.H., Boman, H.G. and Zimmermann, R. (1990). A large presecretory protein translocates both cotranslationally, using signal recognition particle and ribosome and posttranslationally, without these ribonucleoprotein particles, when synthesized in the presence of mammalian microsomes. J. Biol. Chem. , 265 , 13960–13968. Shamu, C.E., Cox, J.S. and Walter, P. (1994). The unfolded protein-response pathway in yeast. Trends Cell Biol. , 4 , 56–60. Shelness, G. and Blobel, G. (1990). Two subunits of the canine signal peptidase complex are homologous to yeast sec 11 protein. J. Biol. Chem . 265 , 9512–9519. Shelness, G.S., Kanwar, Y.S. and Blobel, G. (1988). cDNA-derived primary structure of the glycoprotein component of canine microsomal signal peptidase complex. J. Biol. Chem. , 263 , 17063–17070. Shiu, R.P., Pouyssegur, J. and Pastan, I. (1977). Glucose depletion accounts for the induction of two transformation-sensitive membrane proteins in Rous sarcoma virustransformed chick embryo fibroblasts. Proc. Natl. Acad. Sci. USA , 74 , 3840–3844. Silberstein, S., Collins, P.G., Kelleher, D. and Gilmore, R. (1995a). The essential OST2 gene encodes the 16-kDa subunit of the yeast oligosaccharyltransferase, a highly conserved protein expressed in diverse eukaryotic organisms. J. Cell Biol. , 131 , 371– 383. Silberstein, S., Collins, P.G., Kelleher, D., Rapiejko, P.J. and Gilmore, R. (1995b). The a subunit of the Saccaromyces cerevisiae oligosaccharyltransferase complex is essential for vegetative growth of yeast and is homologous to mammalian ribophorin I. J. Cell Biol. , 128 , 525–536. Simon, S.M. and Blobel, G. (1991). A protein-conducting channel in the endoplasmic reticulum. Cell , 65 , 371–380. Simons, J.F., Ferro-Novick, S., Rose, M.D. and Helenius, A. (1995). BiP/Kar2p serves as a molecular chaperone during carboxypeptidase Y folding in yeast. J. Cell Biol. , 130 , 41–49. Smith, M.J. and Koch, G.L. (1989). Multiple zones in the sequence of calreticulin (CRP55, calregulin, HACBP), a major calcium binding ER/SR protein. EMBO J. , 8 , 3581–3589. Sorger, P.K, and Pelham, H.R. (1987). The glucose-regulated protein grp94 is related to heat shock protein hsp90. J. Mol. Biol. , 194 , 341–344. Sousa, M. and Parodi, A.J. (1995). The molecular basis for the recognition of misfolded glycoproteins by the UDP-Glc: glycoprotein glucosyltransferase. EMBO J. , 14 , 4196– 4203. Spiro, R.G., Zhu, Q., Bhoyroo, V. and Söling, H.D. (1996). Definition of the lectin-like properties of the molecular chaperone, calreticulin and demonstration of its copurification with endomannosidase from rat liver Golgi. J. Biol. Chem. , 271 , 11588–11594. Stirling, C.J. and Hewitt, E.W. (1992). The S. cerevisiae Sec65 gene encodes a component of yeast signal recognition particle with homology to human SRP19. Nature , 356 , 534–537. Stirling, C.J., Rothblatt, J., Hosobuchi, M., Deshaies, R. and Schekman, R. (1992). Protein translocation mutants defective in the insertion of integral membrane proteins into the endoplasmic reticulum. Mol. Biol. Cell , 3 , 129–142.
Molecular chaperones and folding catalysts
256
Suh, K., Bergmann, J.E. and Gabel, C.A. (1989). Selective retention of monoglucosylated high mannose oligosaccharides by a class of mutant vesicular stomatitis virus G proteins. J. Cell Biol. , 108 , 811–819. Tajima, S., Lauffer, L., Rath, V.L. and Walter, P. (1986). The signal recognition particle receptor is a complex that contains two distinct polypeptide chains. J. Cell Biol. , 103 , 1167–1178. te Heesen, S. and Aebi, M. (1994). The genetic interaction of kar2 and wbp1 mutations. Distinct functions of binding protein BiP and N-linked glycosylation in the processing pathway of secreted proteins in Saccharomyces cerevisiae. Eur. J. Biochem. , 222 , 631–637. te Heesen, S., Janetzky, B., Lehle, L. and Aebi, M. (1992). The yeast WBP1 is essential for oligosaccharyl transferase activity in vivo and in vitro. EMBO J. , 11 , 2071–2075. te Heesen, S., Rauhut, R., Aebersold, R., Abelson, J., Aebi, M. and Clark, M.W. (1991). An essential 45 kDa yeast transmembrane protein reacts with anti-nuclear pore antibodies: purification of the protein, immunolocalization and cloning of the gene. Eur. J. Cell Biol. , 56 , 8–18. Thrift, R.N., Andrews, D.W., Walter, P. and Johnson, A.E. (1991). A nascent membrane protein is located adjacent to ER membrane proteins throughout its integration and translation. J. Cell Biol. , 112 , 809–821. Toyn,J., Hibbs, A.R., Sanz, P., Crowe, J. and Meyer, D.I. (1988). In vivo and in vitro analysis of ptl1, a yeast ts mutant with a membrane-associated defect in protein translocation. EMBO J. , 7 , 4347–4353. Trombetta, S.E., Bosch, M. and Parodi, A.J. (1989). Glucosylation of glycoproteins by mammalian, plant, fungal and trypanosomatid protozoa microsomal membranes. Biochemistry , 28 , 8108–8116. Trombetta, S.E., Ganan, S.A. and Parodi, A.J. (1991). The UDP-Glc: glycoprotein glucosyltransferase is a soluble protein of the endoplasmic reticulum. Glycobiol. , 1 , 155–161. Tyedmers, J., Brunke, M., Lechte, M., Sandholzer, U., Dierks, T., Schlotterhose, P., Schmidt, B. and Zimmermann, R. (1996). Efficient folding of firefly luciferase after transport into mammalian microsomes in the absence of lumenal chaperones and folding catalysts. J. Biol. Chem. , 271 , 19509–19513. Varki, A. (1993). Biological roles of oligosaccharides: all of the theories are correct. Glycobiol. , 3 , 97–130. Vogel, J.P., Misra, L.M. and Rose, M.D. (1990). Loss of BiP/GRP78 function blocks translocation of secretory proteins in yeast. J. Cell Biol. , 110 , 1885–1895. Voigt, S., Jungnickel, B., Hartmann, E. and Rapoport, T.A. (1996). Signal sequencedependent function of the TRAM protein during early phases of protein transport across the endoplasmic reticulum membrane. J. Cell Biol. , 134 , 25–35. Volkmer, J., Guth, S., Nastainczyk, W., Knippel, P., Klappa, P., Gnau, V. and Zimmermann, R. (1997). Pancreas specific protein diusulfide isomerase, PDIp, is in transient contact with secretory proteins during late stages of translocation. FEBS Lett. , 406 , 291–295. von Heijne, G. (1981). On the hydrophobic nature of signal sequences. Eur. J. Biochem. , 116 , 419–422. von Heijne, G. (1983). Patterns of amino acids near signal sequence cleavage sites. Eur. J. Biochem. , 113 , 17–21. von Heijne, G. (1984). Analysis of the distribution of charged residues in the N-terminal region of signal sequences: implications for protein export in prokaryotic and
Protein transport into and folding within the endoplasmic reticulum
257
eukaryotic cells. EMBO J. , 3 , 2315–2318. von Heijne, G. (1985). Signal sequences. The limits of variation. J. Mol. Biol. , 184 , 99– 105. Wada, I., Rindress, D., Cameron, P.H., Ou, W.J., Doherty, J.J., Louvard, D., Bell, A.W., Dignard, D., Thomas, D.Y. and Bergeron, J.J. (1991). SSR alpha and associated calnexin are major calcium binding proteins of the endoplasmic reticulum membrane. J. Biol. Chem. , 266 , 19599–19610. Walter, P. and Blobel, G. (1980). Purification of a membrane-associated protein complex required for protein translocation across the endoplasmic reticulum. Proc. Natl. Acad. Sci. USA , 77 , 7112–7116. Walter, P. and Blobel, G. (1981a). Translocation of proteins across the endoplasmic reticulum. II. Signal recognition protein, SRP, mediates the selective binding to microsomal membranes of in-vitro-assembled polysomes synthesizing secretory protein. J. Cell Biol. , 91 , 551–556. Walter, P. and Blobel, G. (1981b). Translocation of proteins across the endoplasmic reticulum. III. Signal recognition protein, SRP, causes signal sequence-dependent and site-specific arrest of chain elongation that is released by microsomal membranes. J. Cell Biol. , 91 , 557–561. Walter, P. and Blobel, G. (1982). Signal recognition particle contains a 7S RNA essential for protein translocation across the endoplasmic reticulum. Nature , 299 , 691–698. Walter, P., Ibrahimi, I. and Blobel, G. (1981). Translocation of proteins across the endoplasmic reticulum. I. Signal recognition protein, SRP, binds to in-vitro-assembled polysomes synthesizing secretory protein. J. Cell Biol. , 91 , 545–550. Wanker, E.E., Sun, Y., Savitz, A.J. and Meyer, D.I. (1995). Functional characterization of the 180 kDa ribosome receptor in vivo. J. Cell Biol. , 130 , 29–39. Ward, C.L., Omura, S. and Kopito, R.R. (1995). Degradation of CFTR by the ubiquitinproteasome pathway. Cell , 83 , 121–127. Ware, F.E., Vassilakos, A., Peterson, P.A., Jackson, M.R., Lehrman, M.A. and Williams, D.B. (1995). The molecular chaperone calnexin binds Glc1Man9GlcNAc2 oligosaccharide as an initial step in recognizing unfolded glycoproteins. J. Biol. Chem. , 270 , 4697–4704. Waters, M.G. and Blobel, G. (1986). Secretory protein translocation in a yeast cell-free system can occur posttranslationally and requires ATP hydrolysis. J. Cell Biol. , 102 , 1543–1550. Waters, M.G., Chirico, W.J. and Blobel, G. (1986). Protein translocation across the yeast microsomal membrane is stimulated by a soluble factor. J. Cell Biol. , 103 , 2629– 2636. Wei, J.Y., Gaut, J.R. and Hendershot, L.M. (1995). In vitro dissociation of BiP peptide complexes requires a conformational change in BiP after ATP binding but does not require ATP hydrolysis. J. Biol. Chem. , 270 , 26677–26682. Wiech, H., Buchner, J., Zimmermann, M., Zimmermann, R. and Jakob, U. (1993). Hsc70, Immunoglobulin heavy chain binding protein and Hsp90 differ in their ability to stimulate transport of precursor proteins into mammalian microsomes. J. Biol. Chem. , 268 , 7414–7421. Wiech, H., Sagstetter, M., Müller, G. and Zimmermann, R. (1987). The ATP requiring step in assembly of M13 procoat protein into microsomes is related to preservation of transport competence of the precursor protein. EMBO J. , 6 , 1011–1016. Wiedmann, M., Kurzchalia, T.V., Hartmann, E. and Rapoport, T.A. (1987). A signal sequence receptor in the endoplasmic reticulum membrane. Nature , 328 , 830–833.
Molecular chaperones and folding catalysts
258
Wiedmann, B., Sakai, H., Davis, T.A. and Wiedmann, M. (1994). A protein complex required for signalsequence-specific sorting and translocation. Nature , 370 , 434–440. Wolin, S.L. and Walter, P. (1988). Ribosome pausing and stacking during translation of a eukaryotic mRNA. EMBO J. , 7 , 3559–3569. Wolin, S.L. and Walter, P. (1989). Signal recognition particle mediates a transient elongation arrest of preprolactin in reticulocyte lysate. J. Cell Biol. , 109 , 2617–2622. YaDeau, J.T. and Blobel, G. (1989). Solubilization and characterization of yeast signal peptidase. J. Biol. Chem. , 264 , 2928–2934. YaDeau, J.T., Klein, C. and Blobel, G. (1991). Yeast signal peptidase contains a glycoprotein and the Sec11 gene product. Proc. Natl. Acad. Sci. USA. , 88 , 517–521. Zhang, Q., Tector, M. and Salter, R.D. (1995). Calnexin recognizes carbohydrate and protein determinants of class I major histocompatibility complex molecules. J. Biol. Chem. , 270 , 3944–3948. Zimmerman, D.L. and Walter, P. (1990). Reconstitution of protein translocation activity from partially solubilized microsomal vesicles . J. Biol. Chem. , 265 , 4048–4053. Zimmerman, D.L. and Walter, P. (1991). An ATP-binding membrane protein is required for protein translocation across the endoplasmic reticulum membrane. Cell Regulation , 2 , 851–859. Zimmermann, R. and Meyer, D.I. (1986). 1986: A year of new insights into how proteins cross membranes. Trends Biochem . Sci., 11 , 512–515. Zimmermann, R., Sagstetter, M., Lewis, M.J. and Pelham, H.R.B. (1988). Seventy kilodalton heat shock proteins and an additional component from reticulocyte lysate stimulate import of M13 procoat protein into microsomes. EMBO J. , 7 , 2875–2880. Zimmermann, R., Zimmermann, M., Mayinger, P. and Klappa, P. (1991). Photoaffinity labeling of dog pancreas microsomes with 8-azido-ATP inhibits association of nascent preprolactin with the signal sequence receptor complex. FEBS Lett , 286 , 95–99. Zopf, D., Bernstein, H.D., Johnson, A.E. and Walter P. (1990). The methionine-rich domain of the 54 kd protein subunit of the signal recognition particle contains an RNA binding site and can be crosslinked to a signal sequence. EMBO J. , 9 , 4511–4517.
11. THE ROLE OF MOLECULAR CHAPERONES IN TRANSPORT AND FOLDING OF MITOCHONDRIAL PROTEINS PETER J.T.DEKKER and NIKOLAUS PFANNER * Institut für Biochemie und Molekularbiologie, Universität Freiburg, HermannHerder-Str. 7, D-79104 Freiburg, Germany
1. Introduction 2. Targeting 2.1. Cytosolic Factors 2.1.1. Cytosolic Hsp70 2.1.2. PBF 2.1.3. MSF 2.2. Recognition at the Outer Membrane 2.3. Substrate Specificity 3. Translocation Across the Mitochondrial Membranes 3.1. The Outer Mitochondrial Membrane 3.2 The Inner Mitochondrial Membrane 3.3. MtHsp70 in Protein Translocation 3.4. Brownian Ratchet and the Import Motor 3.4.1. The Brownian Ratchet 3.4.2. The Import Motor 3.5. Co-chaperones in Protein Import 3.6. A Model for Protein Translocation 4. Folding and Degradation of Mitochondrial Proteins 4.1. Folding of Mitochondrial Proteins 4.2. Prevention of Heat Denaturation 4.3. Protein Turnover in Mitochondria 5. Concluding Remarks 6. Acknowledgements 7. References
The role of molecular chaperones in transport and folding
261
1. INTRODUCTION Mitochondria are ubiquitous organelles housing enzyme complexes that make a decisive contribution to energy production and conservation in all eukaryotic cells using oxygen. Key enzymes of major metabolic routes also reside inside mitochondria. Only few mitochondrial proteins are produced inside the organelle; most are synthesized on cytoplasmic ribosomes and have to be imported into mitochondria. Additionally, the maintenance of the small mitochondrial genome and expression of its genes requires the synthesis and import of an amazingly large amount of mitochondrial proteins encoded by nuclear genes (Attardi and Schatz, 1988; Grivell, 1995). As a result, a large part of the genome of a eukaryotic cell encodes proteins that end up inside mitochondria. For the recently completely sequenced genome of the yeast Saccharomyces cerevisiae, that can live without carrying out oxidative phosphorylation, it can be calculated that almost 20% of its ~6000 genes will encode mitochondrial proteins (Johnston, 1996). All these proteins have to be recognized as being mitochondrial, imported into the organelle and have to end up in one of the four possible mitochondrial locations: the outer membrane, the intermembrane space, the inner membrane, or the mitochondrial matrix. This requires both promiscuity and specificity of the mitochondrial import apparatus. Proteins that are targeted to mitochondria contain targeting regions specifying their final location inside mitochondria. For many proteins that end up in the matrix or intermembrane space this region is located at the amino terminus of the protein, and is clipped off by a specific peptidase (MPP, mitochondrial processing peptidase) after this sequence has entered the matrix. Therefore, this targeting region is often termed a mitochondrial presequence and the targeted protein a preprotein. The mitochondrial presequence is specified by an abundance of arginine and hydroxyl-residues and the virtual absence of negative residues. The presequence is thought to fold into an amphiphilic -helix (von Heijne, 1986) that has the possibility to interact directly with membranes or receptors on the mitochondrial surface. Additionally, many intermembrane space proteins contain a second amino-terminal region that is reminiscent of the targeting sequence of secreted proteins of E. coli. The outside of the inner membrane contains a specialized peptidase (IMP, inner membrane protease) that removes this second targeting sequence. Some mitochondrial proteins, especially polytopic outer and inner membrane proteins, do not contain an obvious presequence. These proteins are probably targeted to their final location by internal sequences or structures that are not removed after import. Internal membrane proteins containing several membrane spanning regions might, therefore, accommodate multiple sequences that specify their final topology. The nature of these internal targeting sequences, however, has not yet been established. Interestingly, a survey of different inner membrane spanning proteins in yeast mitochondria indicates that glutamate residues preferentially reside in small loops at the outside of the inner membrane, while loops that are transported to the matrix do not show any sequence specificity (Gavel and von Heijne, 1992). This is in contrast to inner membrane proteins
Molecular chaperones and folding catalysts
262
that are synthesized inside mitochondria, which follow the “positive inside rule” for their topology. This might indicate that insertion of cytoplasmically synthesized proteins into the inner membrane follows a functionally different route than insertion of mitochondrially synthesized inner membrane proteins. Molecular chaperones are ideally suited to fulfil the indiscriminate nature of the import machinery. Indeed, cytoplasmic Hsp70 (70 kDa class of heat shock proteins, Ssa1p in yeast) is thought to bind to mitochondrial proteins as they are synthesized on cytoplasmic ribosomes, and to keep them from premature folding. As a result, the proteins are presented to mitochondria in an unfolded state, thereby facilitating the translocation reaction. Matrix Hsp70 (Ssc1p in yeast) binds to proteins that enter the matrix compartment and thereby stimulates the further translocation of the incoming protein. Both Hsp70s do not discriminate between mitochondrial and non-mitochondrial. Ssa1p is also involved in the transport of proteins to other organelles (e.g. proteins that are targeted to the endoplasmatic reticulum, Deshaies et al., 1988) (see chapters in this volume by Craig et al., Haas and Zimmermann, Welch et al.), while Ssc1p also drives the import of non-mitochondrial proteins (e.g. cytosolic mouse dihydrofolate reductase (DHFR)) into yeast mitochondria, if fused to a mitochondrial presequence. Most studies that address the targeting and translocation of mitochondrial precursor proteins are performed in vitro. After synthesis of precursor proteins in radio-chemical amounts in rabbit reticulocyte lysates, they are presented to isolated mitochondria of wild-type or mutant yeast or Neurospora crassa. In this review we will discuss how these pre-synthesized proteins end up in the correct mitochondrial compartment in a biologically active state, and especially on the role of molecular chaperones in this process.
2. TARGETING In order to assure selectivity of the mitochondrial import process, mitochondrial proteins that are synthesized on cytoplasmic ribosomes are targeted to mitochondria with the assistance of soluble cytoplasmic factors that specifically recognize these proteins as being mitochondrial. Transport of mitochondrial proteins across the outer and inner membrane has been shown to be most efficient if the preprotein is presented to mitochondria in an unfolded state (reviewed by Stuart et al., 1994a). Proteins that have bound antibodies or Fab fragments, or chimeric preproteins consisting of mitochondrial presequences and mouse DHFR stabilized by binding of a specific ligand, are very poor substrates for the mitochondrial import machinery (Schleyer and Neupert, 1985; Eilers and Schatz, 1986). Ideally, preproteins have to be presented to the import machinery in an extended state or, alternatively, have to be unfolded by the import machinery prior to or during the import reaction. 2.1. Cytosolic Factors The cytoplasm contains several protein factors that have the ability to recognize mitochondrial preproteins and preserve an unfolded state (reviewed by Mihara and
The role of molecular chaperones in transport and folding
263
Omura, 1996) (see also Welch et al., this volume). The most extensively studied factors are; (1) cytoplasmic Hsp70 that keeps preproteins in an import competent state, (2) PBF (presequence binding factor) that is thought to confer specificity to the targeting of presequence containing mitochondrial proteins, and (3) MSF (mitochondrial import stimulation factor) that specifically recognizes mitochondrial proteins and has the ability to restore import competence of aggregated precursor protein. 2.1.1. Cytosolic Hsp70 Hsp70s preferentially bind and release hydrophobic peptides concomitant with ATP binding and hydrolysis (reviewed by Gething and Sambrook, 1992; see chapter by Ha et al., Buchberger et al., this volume) and, therefore, are ideally suited to interact with (partly) unfolded polypeptides or nascent chains in order to prevent aggregation and folding. The requirement for cytosolic Hsp70s during post-translational translocation into mitochondria and into the endoplasmatic reticulum (ER) has been primarily elucidated from studies in yeast and relies on evidence obtained both in vivo and in vitro. In yeast, the cytosolic Hsp70s are encoded by the SSA1 to 4 genes (among others), of which the presence of at least one is required for cell growth (Craig et al., 1994; see Craig et al., this volume). In an experiment designed to elucidate the function of these cytoplasmic Hsp70s, all four SSA genes were deleted and the strain was rescued by expression of the SSA1 gene from the inducible GAL1 promoter (Deshaies et al., 1988). When the yeast strain is shifted from galactose to glucose containing media, it is depleted of Ssa protein (a mitochondrial protein) and prepro- and accumulates precursor protein for F1 factor (a secreted protein). It was concluded that cytosolic Hsp70 is involved in the transport of proteins into both mitochondria and the ER. Further indications for the role that cytoplasmic Hsp70 plays in the translocation of mitochondrial preproteins comes from an in vitro study where a mitochondrial preprotein was translated in a wheat germ extract. Mitochondrial precursor proteins synthesized in reticulocyte lysates are imported efficiently into isolated mitochondria, whereas those synthesized in wheat germ lysates are imported only poorly or not at all, suggesting that the wheat germ system lacks some factors necessary for mitochondrial import (Murakami et al., 1988). Addition of Ssa1p to the wheat germ lysates restored translocation competence when import was examined in isolated yeast mitochondria. The restoration of import by Ssa1p in these experiments relied upon the addition of an N-ethylmaleimide (NEM)-sensitive cytosolic factor whose identity remains obscure. Surprisingly, the Hsp70 dependent import was recently shown to be largely independent of external ATP (Komiya et al., 1996), although the translocation of many preproteins requires the presence of external ATP. It was suggested that Hsp70 might dissociate from preproteins at the outer membrane by a conformational change induced by the interaction with membrane receptors, although this has to be characterized further. Furthermore, it was recently shown, by the analysis of a ssa1 temperature-sensitive mutant, that the translocation of only a subset of preproteins depends on the action of cytosolic Hsp70 (Becker et al., 1996). In order to initiate translocation over the mitochondrial membranes, cytoplasmic Hsp70 has to release the bound polypeptides at the mitochondrial surface. Ydj1p, a DnaJ homologue that might be bound to membranes via a farnesyl lipid moiety at its C-
Molecular chaperones and folding catalysts
264
terminus (Caplan et al., 1992a), could be involved in this process. A conditional (temperature-sensitive) ydj1 mutant shows a reduction of translocation of the precursors of F1 and prepro- -factor at the restrictive temperature, reminiscent of the situation in Ssa depleted cells (Caplan et al., 1992b). The mutant protein has a reduced capacity to stimulate the ATPase activity of Ssa1p and might, as a result, impair the release of preproteins from cytoplasmic Hsp70. Ydj1p might, therefore, be involved in the stimulation of release of preproteins from Ssa1. Indeed, a functional interaction was detected between Ssa1p and Ydj1p during protein translocation (Becker et al., 1996). It was proposed that localization of DnaJlike proteins to membranes might help to concentrate the activity of soluble Hsp70 proteins to these surfaces (Cyr et al., 1994). Since the complex Ssa1p-Ydj1p has no specificity for mitochondrial preproteins, and also functions in transport to e.g. the ER, the chaperone is probably involved in the preservation of an import competent state, rather than targeting. 2.1.2. PBF The targeting specificity of an Hsp70-preprotein complex might be accomplished by factors that specifically recognize the mitochondrial presequences. In vitro import of the purified precursor protein of ornithine carbamoyltransferase (pOTC) depends on a mitochondrial presequence binding factor (PBF) that has been purified from rabbit reticulocyte lysates (Murakami et al., 1992). Binding of this 50 kDa factor to the precursor protein depends on the presence of the presequence and is inhibited by synthetic peptides corresponding to known mitochondrial presequences (Murakami and Mori, 1990). Interestingly, PBF-dependent import of pOTC was further stimulated by the addition of Hsp70. PBF binds to the presequence portion of mitochondrial preproteins and keeps them in an import competent state, probably in cooperation with cytosolic Hsp70. Taken together, PBF might function in the targeting of those preproteins which are held in an import competent state by cytosolic Hsp70, and could therefore supply the specificity to the Hsp70 dependent import reaction. 2.1.3. MSF Rat liver cytosol contains a NEM-sensitive activity that stimulates the import of wheat germ lysate-synthesized preproteins into mitochondria. This protein, termed mitochondrial import stimulation factor (MSF), displays properties that are reminiscent of the action of molecular chaperones by binding to unfolded proteins and restores import competence of aggregated preprotein (Hachiya et al., 1993; 1994). MSF is, however, specific for mitochondrial proteins, in contrast to Ssa1p. While the presequence seems to be a major target for MSF, it might also act on mitochondrial proteins that lack presequences, like porin and apo-cytochrome c (Hachiya et al., 1994). Thus, MSF recognizes several features of the preproteins, including the conformational state of the mature region, in addition to the mitochondrial targeting sequences. MSF consists of two subunits of 30 and 32 kDa that both belong to the 14–3–3 protein family (Alam et al., 1994), members of which have been associated with many diverse intracellular functions, such as cell-cycle regulation, signal transduction and exocytosis
The role of molecular chaperones in transport and folding
265
(Aitken et al., 1992). Homologs of the two MSF subunits have also been identified in the yeast cytoplasm, although it is unclear if these yeast proteins perform the same function. The unfolding function of MSF requires ATP hydrolysis by the chaperone and is stimulated by the addition of preproteins. The import stimulating activity of MSF is sensitive to NEM, while its ATPase activity seems to be largely insensitive to NEM. This chaperone might, therefore, have two functions; one in the preservation of an import competent state, and one in the targeting of mitochondrial proteins. Isolated mitochondrial outer membranes bind MSF and inhibit the ATPase activity of the chaperone (Hachiya et al., 1994). The degree of inhibition correlates well with the degree of binding of the MSF-preprotein complex to the membrane. MSF might be involved in the translocation of a specific set of preproteins, preventing them from aggregation and delivering them to the receptors located in the mitochondrial outer membrane. MSF is, therefore, a major player in post-translational import by targeting mitochondrial preproteins to the outer membrane translocation complex. Thus, the cytoplasm might possess two distinct protein targeting pathways; one NEMsensitive and ATP dependent pathway involving MSF, and one NEM-insensitive, ATP independent pathway involving Hsp70 and perhaps PBF (Komiya et al., 1996; Figure 1). Some preproteins, however, might directly interact with the mitochondrial translocation machinery. During MSF-dependent import, preproteins probably bind to the outer membrane via MSF, while during Hsp70 dependent import the preproteins bind directly to the receptors of the outer membrane (Komiya et al., 1996). It should, however, be kept in mind that all results dealing with the activity of specific chaperones like MSF are obtained in in vitro experiments, and do not address the question of the in vivo importance of this post-translational targeting pathway. 2.2. Recognition at the Outer Membrane Preprotein recognition and translocation across the mitochondrial outer membrane is facilitated by a multi-subunit protein complex termed the Tom complex (for translocase of the outer membrane, reviewed by Lill and Neupert, 1996; Pfanner et al., 1996). Functionally the Tom complex can be subdivided into the general insertion pore (GIP, with the components Tom40, Tom5, Tom6 and Tom7; numbers indicate the molecular weights in kDa) and the receptor complex (Tom20, Tom22, Tom37 and Tom70). The receptor components can be subdivided in two distinct but functionally partly overlapping subcomplexes, consisting of the Tom20-Tom22 and the Tom70-Tom37 partners (Lithgow et al., 1995). Recently it was shown that Tom22 is an integral part of a 400k Tom complex that aslo contains the components of the GIP (Dekker et al., 1996; Dietmeier et al., 1997) and functions in protein translocation across the outer mambrane. Removal of Tom22 even leads to complete dissociation of this complex, indicating that Tom22 has a structural function, besides its receptor activity (Dekker, unpublished). The “real” receptors Tom70, Tom37 and Tom20 are more loosely associated with the GIP. The receptors contain membrane anchors with which they are attached to the mitochondrial outer membrane, and cytosolic domains that can be removed by protease treatment of mitochondria (reviewed by Lill and Neupert, 1996). Initial experiments using protease treated mitochondria have suggested that these cytosolic domains are
Molecular chaperones and folding catalysts
266
required for attachment of precursor proteins to the mitochondrial surface. Inhibition of in vitro protein import by specific antibodies against the Tom70 or Tom20 receptors
Figure 1 Model for the targeting and translocation of mitochondrial preproteins across the outer membrane. Three different targeting pathways are identified: (A) preproteins that depend on external ATP for translocation are bound by MSF in the cytosol. The MSF-preprotein complex interacts with the Tom70-Tom37 receptors on the mitochondrial outer membrane. After MSF is released by a single round of ATP hydrolysis, the preprotein is delivered to the Tom22Tom20 receptor complex and translocates across the outer membrane. (B) preproteins that do not require cytosolic factors for translocation are probably directly bound to the Tom22-Tom20 receptors and translocated across the outer membrane. (C) preproteins interact with cytosolic Hsp70 that keeps them in an import competent state. PBF might interact with the presequence portion of the preprotein. At the outer membrane the preprotein is released from Hsp70 (maybe by the action of Ydj1p). and directly transferred to the Tom22-Tom20 receptor complex. 70: Tom70; 37: Tom37; 22: Tom22; 20: Tom20; TOM: translocase of the outer membrane; IMS: intermembrane space.
indicated that many preproteins have a preference for one of the receptor complexes. This predisposition is not absolute; except for TOM22 none of the genes encoding the receptors is essential in yeast (Hines et al, 1990; Steger et al, 1990; Ramage et al., 1993; Moczko et al., 1994; Gratzer et al., 1995; Hönlinger et al., 1995). The absence of both
The role of molecular chaperones in transport and folding
267
Tom70 and Tom20, often referred to as the major receptor components, is even not lethal to the cell when Tom22 is expressed in normal amounts (Lithgow et al., 1994; Hönlinger et al., 1995). In contrast, deletion of both Tom37 and Tom70 (Gratzer et al., 1995) or deletion of Tom22 (Hönlinger et al., 1995) is lethal, indicating that at least one of the components of both subcomplexes should be present. This indicates that the different import receptors can partly substitute for each others function but their presence is strictly required for the import process. 2.3. Substrate Specificity What is the basis of the substrate specificity of the individual receptors? A clue to the answer to this question comes from the recent work on the functioning of MSF (Hachiya et al., 1995: Komiya et al., 1996). Import studies with yeast mitochondria have shown that mitochondrial precursor proteins can be classified into two groups (Stuart et al., 1990; Wachter et al., 1994). One group is independent of extramitochondrial ATP for their import (containing precursors for cytochrome b 2, Hsp60, CoxIV, cytochrome c haem lyase, and proteins fused with dihydrofolate reductase), the other group requires external ATP (the -subunit of the F1-ATPase, the -subunit of mitochondrial processing peptidase, alcohol dehydrogenase III, cytochrome c 1 and the adeninenucleotide translocator). Interestingly, MSF seems to stimulate the import of those mitochondrial proteins that require extra-mitochondrial ATP for their translocation. It is precisely this group of preproteins that is most dependent on the outer membrane receptors Tom37-Tom70 (reviewed by Lithgow et al., 1995). Indeed, the MSF-preprotein complex seems to dock initially onto the Tom37-Tom70 subcomplex (Hachiya et al., 1995), and this event quenches the ATPase activity of MSF. Upon a single round of ATP hydrolysis, MSF dissociates, and leaves the preprotein attached to the Tom37-Tom70 receptor complex. The precursors presequence is then transferred to the Tom20-Tom22 receptor complex and further through the GIP in the outer membrane (Figure 1). In contrast, the Hsp70 dependent import pathway is largely independent of the Tom37Tom70 receptor complex and ATP (Komiya et al., 1996). IgG against rat Tom37 inhibits the import of MSF-targeted pre-adrenodoxin but does not have an effect on Hsp70targeted preprotein. Both targeting pathways are, however, inhibited by IgG against rat Tom20, indicating that Hsp70-bound precursors are directly targeted to the Tom20Tom22 receptor complex. Furthermore, Tom20 has also been reported to function as the receptor for a chemically pure, urea-denatured precursor (Becker et al., 1992) interacting with the mitochondrial targeting signal (Haucke et al., 1995). Denatured and Hsp70 associated preproteins, therefore, might be targeted preferentially to the Tom20-Tom22 receptors, while MSF-targeted preproteins require the presence of Tom37-Tom70. It will be interesting to see if the requirement for MSF and the Tom37-Tom70 receptors is absolute, that is, if a MSF null strain still requires the presence of these receptors. Unfortunately, most results concerning MSF have been obtained using in vitro experiments using the rat liver protein. When the yeast homologs of the MSF heterodimer have been characterized more extensively, we might get an answer to the in vivo importance of this fascinating targeting pathway.
Molecular chaperones and folding catalysts
268
3. TRANSLOCATION ACROSS THE MITOCHONDRIAL MEMBRANES 3.1. The Outer Mitochondrial Membrane After binding to the Tom20-Tom22 receptors (the cis-site), preproteins should initiate translocation across the mitochondrial outer membrane. At present the driving force for the initial insertion and translocation of the amino-terminal presequence is unknown. Isolated outer membrane vesicles are able to translocate (only) positively charged presequences to the trans side of the membrane (reviewed by Lill and Neupert, 1996). Translocation occurs apparently without any input of ATP and in the absence of a membrane potential, suggesting it might involve spontaneous diffusion of the presequence through the GIP. Alternative ideas take the amphiphilic nature and positive charge of the presequence into account. It has been suggested that the presequence is released from the cytosolic domain of the Tom22 receptor via the GIP component Tom5 and transferred to a negatively charged intermembrane space (IMS) domain of the same receptor (acid-chain hypothesis; Bolliger et al., 1995; Hönlinger et al., 1995; Dietmeier et al., 1997; Moczko et al., 1997). In this model, Tom22 has an important function in the translocation of preproteins over the mitochondrial outer membrane, concordant with the essential character of this receptor. The existence of chaperones in the IMS that might assist the translocation across the outer membrane cannot be ruled out, and awaits further study. It is generally thought that translocation of preproteins in mitochondria occurs via so called “contact sites”. This idea comes from electron-microscopic studies, where translocating chains could be localized in areas where both membranes are in close contact (reviewed by Attardi and Schatz, 1988). The proposed trans site for translocation over the outer membrane might, therefore, also be part of the inner membrane translocation machinery. Indeed, the inter membrane space domain of one of the subunits of the translocase of the inner membrane, Tim23, was able to contact an incoming polypeptide (Bauer et al., 1996). Contact sites should not be viewed as stable entities, but are dynamic structures that might only be formed when a polypeptide chain translocates through the membranes (Pfanner et al., 1992). Preproteins can also be imported into mitoplasts, mitochondria where the outer membrane is removed, indicating the dynamic behaviour of the two translocation pores. Using a chimeric protein, containing a mitochondrial preprotein fused to mouse DHFR, translocation over the membranes can be arrested by binding of the ligand methotrexate. In this way the minimal length of a translocating preprotein spanning the two membranes could be calculated. Approximately 50 amino acids were shown to be sufficient to span the contact sites, which is equivalent to the length of a polypeptide in an extended conformation to span two membranes (Rassow et al., 1990). This experiment shows that under in vitro translocation conditions, the polypeptide chain can span the membranes in an extended state, and that tightly folded domains do not become imported. This idea should not be seen as an absolute requirement for complete unfolding of the translocating chain; proteins attached to bulky (and highly negatively charged) DNA or RNA molecules still get translocated (Vestweber and Schatz, 1989; Tarassov et al., 1995). The efficiency of translocation is,
The role of molecular chaperones in transport and folding
269
on the other hand, dramatically stimulated for unfolded or loosely folded proteins and might, therefore, represent the preferred in vivo situation. 3.2. The Inner Mitochondrial Membrane Compared with the protein components involved in targeting to and translocation across the mitochondrial outer membrane, the translocase of the inner membrane (Tim) is currently poorly characterized. The central pore-forming elements are the integral membrane proteins Tim23 and Tim17 which interact functionally and physically with each other. Co-immune precipitation experiments have identified additional proteins of 55, 33, 20 and 14 kDa interacting with this complex, but these are not further characterized (Berthold et al., 1995; Blom et al., 1995). The peripheral membrane protein Tim44 is only loosely, yet functionally, attached to this complex. Matrix Hsp70 (Ssc1p in yeast) interacts ATP-dependent with Tim44 and this sub-complex has an essential role in the translocation of preproteins across the mitochondrial membranes (see the section on matrix Hsp70). Translocation of the presequence of preproteins across the mitochondrial inner membrane is driven by the electrochemical potential( ) component of the protonmotive force (Schleyer and Neupert, 1985), maintained by proton pumping by the respiratory chain or the F1Fo-ATPase in the inner membrane. Presequence transport, therefore, is affected by uncouplers or ion-channels like CCCP or valinomycin. It is generally thought that the (negative inside) drives the translocation of the positively charged presequences across the membrane by an electrophoretic mechanism (Martin et al., 1991). The membrane potential might also have an effect on dimerization of Tim23, although it is at present unclear what function this serves (Bauer et al., 1996). The influence of the membrane potential on the import of matrix loops of integral inner membrane proteins is less well understood. Generally, these loops do not have a clear positively charged character (Gavel and von Heijne, 1992), yet they depend on the membrane potential for their translocation, and therefore seem to challenge the electrophoretic hypothesis. Recent evidence suggests that these proteins are inserted into the inner membrane via a different Tim complex (Sirrenberg et al., 1996; Dekker et al., 1997; Kersher et al., 1997). 3.3. MtHsp70 in Protein Translocation After the presequence has translocated the inner membrane, it is often removed by a specific mitochondrial processing peptidase (MPP) in the mitochondrial matrix. This processing reaction leads to the accumulation of a smaller (mature) product, of which the occurrence in in vitro and in vivo import experiments is used as an indication that the presequence has completely entered the matrix space. Further translocation of the mature part of a preprotein into the matrix is independent of , yet requires ATP instead. Recent experiments indicate that the ATP requirement can largely be ascribed to the function of a matrix Hsp70 (Ssc1p in yeast; Kang et al., 1990; Scherer et al., 1990). Depletion of matrix ATP in in vitro import experiments leads to the accumulation of unprocessed precursor proteins that stay protease accessible and are thus not imported
Molecular chaperones and folding catalysts
270
(Cyr et al., 1993; Gambill et al., 1993). Notably, import of proteins that are targeted to the inner membrane or the intermembrane space do not require matrix ATP for import and maturation, provided that tightly folded domains in the preproteins are absent (Pfanner et al., 1987; Wachter et al., 1992; Stuart et al., 1994b). Denatured IMS targeted preproteins are even efficiently imported into mitochondria of a mtHsp70 mutant with disturbed ATP binding. A similar preprotein that lacks the IMS targeting information and as a result ends up in the mitochondrial matrix, is on the other hand fully dependent on the ATPase function of mtHsp70 (Voos et al., 1993). The exact function of mtHsp70 during the translocation reaction is currently a topic of debate in the field of mitochondrial protein import. It is generally thought that mtHsp70 interacts with incoming preproteins as soon as an available epitope enters the matrix space. An interesting observation in this regard is that preproteins that have exposed only their presequence to the matrix space are much less efficiently or stably bound by mtHsp70 then preproteins that expose also mature parts to the matrix space (Ungermann et al., 1996). Since Hsp70-like proteins mainly interact with hydrophobic residues (Blond-Elguindi et al., 1993), the positively charged presequence might be a poor substrate for the chaperone. Studies on the ATP requirements for polypeptide binding of the E. coli Hsp70 homologue DnaK have indicated that the ATP-bound form of the chaperone releases and binds to the substrate protein quickly, while the ADP-bound form can be regarded as the slow-binding, slow-release form (Schmid et al., 1994; McCarty et al., 1995; see Buchberger et al., this volume). Consequently, the ADP-bound form of DnaK is bound more stable to polypeptides and is predominantly found in steadystate situations. Indeed, mitochondrial Hsp70 was also found to bind more stable to preproteins when in the ADPform. Addition of Mg-ATP during lysis of the mitochondria even leads to complete dissociation of the preprotein-mtHsp70 complex (Gambill et al., 1993; Voos et al., 1993). Similar to the situation in E. coli, mtHsp70 in the ADP form might be stably bound to preproteins, while the chaperone is released upon ATP binding. Recently it was suggested that the situation might be reverse for preproteins that expose only their presequence to the matrix space (Ungermann et al., 1994; Ungermann et al., 1996). These translocation intermediates tend to slip back through the translocation channel (retrograde translocation), and even completely dissociate from mitochondria when mitochondria are depleted of ATP by treatment with oligomycin and apyrase. However, retrograde translocation should not automatically be seen as the result of diminished mtHsp70 binding to incoming preproteins (as was suggested), it might also be the cause of it. Additionally, prevention of retrograde translocation might not be a sole effect of binding of mtHsp70 to incoming preproteins. Dissipation of the membrane potential leads to a similar amount of release of partly translocated preproteins. Surprisingly, the retrograde translocation of a preprotein that exposes a larger area to the matrix space is also increased upon ATP depletion, although the binding of mtHsp70 is clearly more stable than in the presence of ATP (Ungermann et al., 1996), indicating that the amount of retrograde translocation does not directly reflect the amount of stably bound mtHsp70.
The role of molecular chaperones in transport and folding
271
3.4. Brownian Ratchet and the Import Motor Currently two models for the action of mtHsp70 during protein translocation prevail (reviewed by Pfanner and Meijer, 1995). In the first model, the Brownian ratchet, preprotein translocation occurs merely as the result of Brownian motion of the translocating protein chain (Schneider et al., 1994). MtHsp70 serves in this model as a ratchet by binding to incoming polypeptides and thereby prevents them from slipping back in the translocation channel (Figure 2A). As the polypeptide chain diffuses further into the mitochondrial matrix, additional mtHsp70 molecules bind to it and prevent retrograde translocation. Finally the complete preprotein is transported into the matrix space and can fold into its native conformation. In the second model, the import motor model, preprotein translocation is stimulated by a conformational change in mtHsp70 that actively pulls translocating chains into the matrix (Glick, 1995). Similar to the Brownian ratchet model, mtHsp70 interacts with incoming polypeptides as soon as the presequence has passed the inner membrane (Figure 2B). Subsequently, however, the chaperone changes its shape and as a result exerts a force on the protein chain. Essential prerequisite for the functioning of the import motor is that mtHsp70 should be locally anchored to the membrane in order to exert a unidirectional force on the polypeptide. After the conformational change the chaperone should dissociate from the polypeptide or the membrane (or both) in order to initiate another round of binding and pulling. Eventually the complete preprotein will be transported and folded in the matrix. 3.4.1. The Brownian Ratchet Both models rely on several assumptions to function properly. The Brownian ratchet model depends on spontaneous diffusion of the unfolded polypeptide chain through the translocation channels and trapping of incoming chains in the matrix by mtHsp70. Ideally, therefore, preproteins are presented to mitochondria in an unfolded conformation. Indeed, as discussed above, cytosolic chaperones like cytosolic Hsp70 or MSF bind to and prevent refolding of mitochondrial preproteins. However, these chaperones probably dissociate as soon as the import receptors at the outer membrane are encountered. Outer membrane lipids and receptors might possess an unfoldase activity which could preserve the unfolded state of a preprotein (Eilers and Schatz, 1988). Obviously, the import velocity will be limited by the rate of dissociation from chaperones or receptors. On the other hand, most preproteins are not fully unfolded during normal in vitro import experiments from reticulocyte lysate, since import can often be facilitated by chemical denaturation of the preprotein, i.e. with urea, before import. If preproteins are folded before translocation, the velocity of import will be restricted by the rate of unfolding of the structured domains. Although some structural domains are unstable and consequently unfold spontaneously, others are extremely stable, especially when complexed with ligands. For instance, the haem-binding domain of cytochrome b is complexed by haem in reticulocyte lysate before import, and spontaneous unfolding will be slow. Nevertheless, import of cytochrome b 2 occurs on a minute time scale. Yet, it should be stressed that translocation traps those polypeptides that are unfolded, and might effectively shift the equilibrium to the unfolded state. Recently it was shown in a kinetic
Molecular chaperones and folding catalysts
272
study that some preproteins have to be unfolded actively, suggesting that translocation cannot be the result of Brownian motion alone (Matouschek et al., 1997). A prerequisite for the Brownian ratchet to function is that mtHsp70 should not dissociate from the translocating chain, since this would lead to retrograde translocation (as suggested by Ungermann et al., 1994). It is, of course, especially important for the mtHsp70 that is bound closest to the membrane to have a stable interaction with the incoming preprotein. Surprisingly, it was recently shown that this membrane-localized mtHsp70 has a much lower affinity for the translocating
Figure 2 Two models for the translocation of preproteins across the mitochondrial inner membrane. (A) The Brownian ratchet model.
The role of molecular chaperones in transport and folding
273
Movement of the preprotein across the membrane is a result of Brownian motion. MtHsp70 interacts with the incoming protein on the matrix side and thereby prevents back-sliding through the translocation channel (retrograde translocation). Stable mtHsp70 binding is essential for the Brownian ratchet to function. Further inside movement of the translocating chain provides space for the interaction of additional mtHspTO molecules. (B) The import motor model. Movement of the preprotein across the inner membrane is driven by a conformational change of the interacting mtHsp70. To supply unidirectional movement, mtHspTO should be attached to the inner membrane during this change. Tim44 might serve as a membrane anchor for the chaperone. After force generation, mtHsp70 should dissociate from Tim44 or the preprotein to provide space for the interaction of additional mtHsp70 molecules. 70: mtHsp70; 44: Tim44; IM: inner mitochondrial membrane; IMS: intermembrane space; TIM: translocase of the inner membrane.
polypeptide then mtHsp70s that interact with fully translocated chains (Ungermann et al., 1996). Moreover, when the interaction of the membrane-localized mtHsp70 with preprotein was stabilized by preincubation of mitochondria with AMP-PNP (a nonhydrolyzable analogue of ATP), import of a membrane spanning preprotein was slowed, as judged from the reduced processing compared to the situation in energized mitochondria (Ungermann et al., 1996). Furthermore, mutant mtHsp70s (Ssc1–2p and Ssc1–3’p) still bind efficiently to incoming preproteins but fail to translocate them in some cases (Gambill et al., 1993; Voos et al., 1993; Voos et al., 1996). Apparently, the mere binding of mtHsp70 to incoming preproteins is not sufficient for translocation (see below). Recently it was found only substoichiometric amounts of mt Hsp70 interact with a translocating polypeptide chain, suggesting that another component of the translocation apparatus performs the ratchet function (Dekker et al., 1997). These observations have led to the proposal of a different model for the mechanism of translocation of mitochondrial preproteins across the inner membrane. 3.4.2. The Import Motor The import motor model is somewhat more complicated then the Brownian ratchet, and consequently relies on more, but testable, assumptions. First of all, in order to generate a force on the translocating polypeptide chain, mtHsp70 should be anchored to the membrane at the entry point of the presequence. Indeed, mtHsp70 has recently been shown to bind to the peripheral inner membrane protein Tim44 (Kronidou et al., 1994; Rassow et al., 1994; Schneider et al., 1994). Part of Tim44 is located in the matrix space, while its carboxy-terminus extrudes into the IMS (Maarse et al., 1992; Kronidou et al., 1994). Tim44 is shown to have a biochemical and genetic interaction with Tim23 and Timl7, integral membrane proteins that are thought to constitute part of the inner membrane translocation channel (Berthold et al., 1995; Blom et al., 1995). In mitochondria, most Tim44 is complexed with mtHsp70 in a 1:1 ratio. Therefore, Tim44 might operate as a membrane anchor for mtHsp70 at the preprotein entry site. The
Molecular chaperones and folding catalysts
274
interaction between mtHsp70 and Tim44 could serve in concentrating the chaperone at the membrane as well. Interestingly, also Tim17 interacts with mtHsp70 and might serve as an additional candidate for the anchoring of the chaperone to the inner membrane translocation channel (Bömer et al., 1997). If Tim44 functions as a membrane anchor, the preprotein and Tim44 binding sites on mtHsp70 should not be identical. The conditional ssc1–2 mutant has shed some light on this issue. It was shown that mutant mtHsp70 interacts efficiently with preproteins, while the binding of Tim44 was disturbed (von Ahsen et al., 1995). This indicates that the preprotein and Tim44 bind to different parts of the mtHsp70 molecule. Mitochondria of ssc1–2 are disturbed in the import of matrix proteins that have not been unfolded prior to the import reaction (Gambill et al., 1993; Voos et al., 1993; Voos et al., 1996). The interaction of mtHsp70 with Tim44, therefore, seems to be essential for force generation by mtHsp70. In order to generate a force on the preprotein, mtHsp70 should bind simultaneously to the preprotein and to its membrane anchor. The occurrence of this ternary complex has been especially difficult to show. While both mtHsp70 and Tim44 can be crosslinked efficiently to translocating preproteins (Scherer et al., 1992; Blom et al., 1993; Kübrich et al., 1994), immunoprecipitation experiments only show marginal binding of Tim44 to preproteins. The unequivocal detection of a ternary complex in isolated mitochondria might be complicated by the inherent instability of this complex. Indeed, both peptide and a permanently unfolded protein (reduced and carboxy-methylated lactalbumin, RCMLA) affects the stability of the Tim44/mtHsp70 interaction (Schneider et al., 1994). These experiments are further complicated by the fact that Tim44 itself seems to preferentially interact with the presequence part of preproteins. This interaction even occurs when presequence containing proteins are added to mitochondrial lysates, making the distinction between pre- and post-lysis binding difficult (Blom et al., 1993). Recently, however, Horst et al. (1996) were able to reconstitute this part of the Hsp70 cycle by binding of a synthetic peptide to isolated mtHsp70, which in turn could interact with isolated, immobilized Tim44. The analysis of binding of entire preproteins to the mtHsp70-Tim44 complex awaits further study. After formation of a ternary complex between Tim44, mtHsp70 and preprotein, the chaperone should change conformation to provide the power stroke for the translocation of the preprotein. The E.coli Hsp70 homologue DnaK and eukaryotic Hsp70 homologs were shown to change conformation upon interaction with ATP (Buchberger et al., 1995; Ha and McKay, 1995; Wei et al., 1995). Additionally, the ADP form of DnaK can adopt different conformational states (Banecki and Zylicz, 1996). For mtHsp70 it was recently shown that it also changes conformation upon ATP binding (von Ahsen et al., 1995). If this conformational change provides the energy for the translocation of preproteins remains to be elucidated. The results do, however, indicate that Hsp70s are flexible molecules and change their shape during the ATP hydrolysis cycle. An essential prerequisite for the import motor to function, in contrast to the Brownian ratchet, is that mtHsp70 should dissociate from Tim44 or the preprotein or both. In other words, mtHsp70 should recycle to provide space for the next interaction and power stroke. Indeed, the interaction of mtHsp70 with preproteins and Tim44 is clearly transient. When mitochondria are lysed in the presence of MgATP both interactions can
The role of molecular chaperones in transport and folding
275
not be detected (Kronidou et al., 1994; Rassow et al., 1994; Schneider et al., 1994). E. coli DnaK in the ATP form binds and releases its substrate polypeptide quickly (Schmid et al., 1994; McCarty et al., 1995). ATP binding, therefore, leads to dissociation of the substrate polypeptide. At present, there are no indications that mtHsp70 would react otherwise. Increased interaction, upon ATP addition, of mtHsp70 with preproteins that have just entered the matrix space (Ungermann et al., 1996) might reflect efficient binding of mtHsp70 in these in vitro experiments, while the interaction is stabilized by ATP hydrolysis. On the other hand, it was published that the interaction between mtHsp70 and Tim44 is disrupted upon hydrolysis of ATP (Schneider et al., 1994). This observation turned out to be incorrect, mtHsp70 already dissociates from Tim44 upon binding ATP (von Ahsen et al., 1995; Horst et al., 1996). A mtHsp70 mutant (ssc1–3 ′ ) that fails to release from Tim44 is defective in the import of preproteins, suggesting that the cycling of mtHsp70 is essential for the function of the chaperone during import (Voos et al., 1996). These results indicate that mtHsp70 in wild-type mitochondria binds and releases the translocating preprotein and Tim44 during the ATP hydrolysis cycle, concordant with the import motor model for translocating preproteins across the mitochondrial membranes. Concluding, preproteins that lack stable folded domains might be imported by a Brownian ratchet type mechanism, while preproteins with (partly) folded domains (probably most imported proteins) rely on force generation by the import motor. 3.5. Co-chaperones in Protein Import The E.coli Hsp70 homologue DnaK functions in cooperation with two co-chaperones that regulate the ATPase cycle of the chaperone (Liberek et al., 1991; Szabo et al., 1994). Binding of the substrate polypeptide and DnaJ stimulate ATP hydrolysis by DnaK, while GrpE stimulates the release of nucleotide. Mitochondria also possess DnaJ and GrpE homologs that regulate the mtHsp70 cycling. Mdj1p is a DnaJ homologue in yeast mitochondria and regulates the activity of Ssc1p in protein folding (Rowley et al., 1994). Mdj1p does not seem to affect the import of preproteins and the deletion of the gene encoding this co-chaperone is not lethal to the cell. This suggests that the activity of mtHsp70 in protein import might be regulated by another DnaJ homologue. Tim44 contains a small region that has some homology to the J-domain (Rassow et al., 1994) that is shared by all known DnaJ homologs, and is implicated in the interaction with Hsp70 homologs. The presence of this J-domain in Tim44 is essential for its function (Meijer, pers. comm.). It is speculated that this J-domain of Tim44 is involved in its interaction with mtHsp70, and that Tim44 might have a DnaJ-like function, although this has yet to be shown experimentally. A GrpE homologue in yeast mitochondria, Mge1p, is essential for growth (Bolliger et al., 1994; Ikeda et al., 1994; Laloraya et al., 1994). The import of preproteins in vivo and in vitro is diminished when cells are depleted of Mge1p, and in temperaturesensitive mge1 mutants at nonpermissive conditons (Laloraya et al., 1994; Nakai et al., 1994; Laloraya et al., 1995; Westermann et al., 1995). Mge1p interacts in ATP dependent fashion with mtHsp70 and forms a ternary complex including mtHsp70 and preproteins in transit across the mitochondrial membranes (Bolliger et al., 1994; Voos et al., 1994).
Molecular chaperones and folding catalysts
276
While E.coli GrpE is involved in the release of nucleotides from DnaK, the mitochondrial homologue might function in a comparable manner. Mge1p might influence the interaction of mtHsp70 with preproteins. Indeed, the binding of mtHsp70 to preproteins spanning the mitochondrial membranes was shown to be decreased in different mge1 mutants (Laloraya et al., 1995; Westermann et al., 1995). When preproteins were fully imported into the matrix space, this interaction is stabilized in mge1–100 mitochondria, suggesting that Mge1p regulates both binding and release of mtHsp70 from preproteins (Laloraya et al., 1995). Additionally, the interaction between mtHsp70 and Tim44 was shown to be stabilized in the mutant mge1–3 (Westermann et al., 1995). Since Mge1p might affect nucleotide release from mtHsp70, the chaperone presumably accumulates in the ADP form in mutant mitochondria, having a slow-binding/slow-release phenotype. Mge1p might regulate the cycling of mtHsp70 from Tim44 and preproteins during the translocation but also has an effect on protein folding inside mitochondria. 3.6. A Model for Protein Translocation A coherent model that explains all available results that deal with the role of ATP and mtHsp70 in the translocation of preproteins is lacking at present. Interpretation of the experiments has been complicated by partial conflicts in results and diversity of methods. Since in recent time additional observations have become available that shed light on some of the problems, we here would like to propose a scheme for protein translocation into mitochondria that includes elements of both the Brownian ratchet and the import motor models. Isolated mitochondria contain a 1:1 complex between mtHsp70 and Tim44 (and perhaps Mge1p) located at the translocation channel in the mitochondrial inner membrane. MtHsp70 is in the ADP form in this complex (von Ahsen et al., 1995). It is essential to note that this complex already exists in the absence of any protein import. While ADP-bound mtHsp70 is in the “slow-binding” form, it will only interact with incoming preproteins when the binding-epitope is available for a longer period of time. In vivo this will occur when (partly) folded domains of the preprotein are opposed against the mitochondrial outer membrane, prohibiting further diffusion of the polypeptide across the membranes. After Tim44-bound mtHsp70, ADP has interacted with the preprotein, the cycle might be initiated by Mge1p that stimulates the release of ADP from the complex. A conformational change in Tim44 bound mtHsp70 (maybe upon ATP binding, ADP or phosphate release) results in force generation on the preprotein that leads to the unfolding of the structured domain outside mitochondria. MtHsp70 dissociates from Tim44 and the preprotein upon ATP binding and is then available for the next step in translocation. Since the domain on the outside is now unfolded it can quickly diffuse into mitochondria by Brownian motion. Inside the matrix the preprotein is then bound to mtHsp70 in the ATP form (the fast-binding form), without the need for Tim44 function. Binding of the preprotein probably stimulates the ATPase activity of mtHsp70 leading to a stable interaction with the preprotein, thereby trapping it in the matrix space. Tim44 can also interact again with MtHsp70, and the anchored chaperone in the ADP-form will await the next folded domain. This model states that polypeptides can be translocated across the mitochondrial
The role of molecular chaperones in transport and folding
277
membranes by a Brownian ratchet mechanism as long as they do not contain any folded domains. In this situation the function of Tim44 is not required. The model also explains how the import apparatus senses the folding state of the preprotein on the outside of mitochondria and knows when to switch to the import motor. Tim44-bound mtHsp70, which is in the slow-binding state, will only come into action when diffusion of the preprotein through the import channel is slowed by physical obstruction on the outside, and a specific peptide-epitope is in the vicinity of the peptide-binding site of the chaperone for a longer time. Since Tim44-bound mtHsp70 is in the ADP form, it will only react when translocation is slowed. Although additional details of this model still have to be filled in (like in which step of the cycle the power stroke occurs), the general consensus might be that both the Brownian ratchet and the import motor model have validity and might function jointly in the translocation of most mitochondrial preproteins.
Figure 3 New model for the translocation of preproteins across the mitochondrial inner membrane. (1) The membrane potential across the mitochondrial inner membrane drives the translocation of the
Molecular chaperones and folding catalysts
278
positively charged presequence. MtHsp70*ADP (slow-binding form), awaits the preprotein bound to Tim44. (2) Further translocation of the preprotein is obstructed by a folded domain on the outside of the membrane. Tim44-bound mtHsp70 has sufficient time to interact with the preprotein on the matrix side. (3) The action of Mge1p leads to release of ADP from mtHsp70, rebinding of ATP, and a conformational change in the chaperone pulls in the preprotein and unfolds the structured domain on the opposite side of the membrane. Until now it is unclear at which stage the effective conformational change occurs. (4) After binding ATP, mtHsp70 releases Tim44 and the preprotein, and provides space for the interaction of the next ATP-bound mtHsp70 (fast-binding form). (5) The interaction of mtHsp70 with preprotein is stabilized by ATP hydrolysis. (6) Since the preprotein is now unfolded, it can freely cross the inner membrane by diffusion. This provides the space for the interaction of additional mtHsp70 molecules in the ATP-bound form. Tim44 is not required anymore during this part of the translocation cycle. (7) While the preprotein is further translocated, additional mtHsp70*ATP molecules might interact with Tim44, and (8) await the next folded domain in the ADP form. The fully translocated protein is still bound to mtHsp70*ADP and can now be folded and assembled with the aid of additional chaperones. 44: Tim44; E: Mge1p; IM; inner mitochondrial membrane; TIM: translocase of the inner membrane; : membrane potential; IMS: intermembrane space.
4. FOLDING AND DEGRADATION OF MITOCHONDRIAL PROTEINS After mitochondrial proteins are translocated across the membranes in an unfolded state, they have to regain their biological activity by folding and assembly into multisubunit protein complexes. Since the import reaction was assisted by mtHsp70, the unfolded proteins are, per definition, bound to this chaperone. MtHsp70 is, however, not the only chaperone that can assist protein folding inside the mitochondrial matrix. A large array of specific and general chaperones, proteases and other factors assisting folding and assembly of imported and mitochondrially synthesized proteins have been analysed. In yeast mitochondria, already more then 15 specific factors only involved in the assembly of the respiratory chain complexes have been identified (reviewed by Grivell, 1996). Discussion of the properties of all these factors is beyond the scope of this review, and we will focus our attention to the more general chaperones and proteases, often related to bacterial counterparts. 4.1. Folding of Mitochondrial Proteins Besides its essential role in protein translocation, mtHsp70 also triggers the folding of some imported and mitochondrially synthesized proteins (Kang et al., 1990; Herrmann et al., 1994). Recently, the completion of the sequence of the entire yeast genome has led to
The role of molecular chaperones in transport and folding
279
the identification of the genes of additional mitochondrial Hsp70 homologs (Schilke et al., 1996), of which the exact role in mitochondrial biogenesis remains to be established. Besides its role in protein import, the function of Mge1p in folding of proteins has been investigated in conditional mge1 mutants in yeast. In mutant mitochondria, imported proteins accumulate in a mtHsp70 bound form and, therefore, aggregation at elevated temperatures was diminished (Laloraya et al., 1995). In a different mutant the yield of folding was also diminished (Westermann et al., 1995). In analogy to the situation in bacteria, the GrpE homologue in mitochondria might also mediate the release of nucleotide from Hsp70 and thereby increasing ATP-cycling and folding of protein (Dekker and Pfanner, 1997; Miao et al., 1997). It has been speculated that the folding of newly imported mitochondrial proteins occurs via a distinct sequential route. After initial binding to mtHsp70 during the import reaction, mitochondrial proteins are released from this chaperone by the action of Mge1p and transferred to the mitochondrial chaperonin complex consisting of Hsp60 and mtcpn10, where completion of folding occurs (Manning-Krieg et al., 1991). It was shown that imported preproteins only interact transiently with mtHsp70. ATP-dependent release from mtHsp70 is insufficient to cause folding of some imported proteins and assembly of several proteins requires sequential, ATP dependent interactions with both mtHsp70 and Hsp60. This sequential action of mtHsp70 and Hsp60 has been extended to the behaviour of the homologs of these chaperones in E.coli (Langer et al., 1992). Some imported mitochondrial proteins fold, however, in the absence of functional mtHsp70 (Rospert et al., 1996), indicating that there is no strict sequential action of mtHsp70 and Hsp60 in the folding of all preproteins. The observed pathway should be seen more as the consequence of the function of mtHsp70 in protein import, and not so much as the preferred folding sequence. Mt Hsp70 was found to exist as two different complexes. The ADP conformation of mt Hsp70 favors formation of a complex on the inner membrane with Tim44 and Mge1p and might be involved in protein translocation. The ATP conformation favors formation of a complex in the matrix with Mdj1p and Mge1p and was suggested to be involved in protein folding (Horst et al., 1997). Compared to the elusive in vitro studies to the function and structure of E.coli GroEL and GroES, its mitochondrial counterparts Hsp60 and cpn10 are relatively poorly investigated. The advantage of the mitochondrial system is, however, that proteins can be imported in organello, and subsequent folding studies consequently mimic the in vivo situation more closely. Hsp60 is essential for viability in yeast, meaning that it performs an essential step in the biogenesis of some mitochondrial proteins. Folding studies have most successfully been performed with mitochondria containing temperature sensitive mutant forms of Hsp60 and cpn10 (Cheng et al., 1989; Höhfeld and Hartl, 1994). From these studies, Hsp60 has been viewed as the universal mediator of protein folding in the mitochondrial matrix. It should, however, be emphasised that folding of imported mitochondrial proteins does not in all cases require the function of Hsp60. Recently, the folding of two fusion proteins, Su9-DHFR and Su9-barnase (a matrix-targeting signal fused to mouse dihydrofolate reductase and barnase, respectively), and the authentic preprotein Cpr3 were found to fold efficiently independently of Hsp60 after import into mitochondria (Rospert et al., 1996). Interestingly, assembly of subunits of oligomeric complexes was affected in a cpn10 mutant (Höhfeld and Hartl, 1994) or in mitochondria
Molecular chaperones and folding catalysts
280
depleted of Hsp60 (Hallberg et al., 1993), while folding of monomeric DHFR was independent of chaperonin function. Apart from the action of molecular chaperones in the mitochondrial matrix, folding efficiency is increased by the action of mitochondrial cyclophilin, a peptidylprolyl cistrans isomerase. Inactivation of the gene encoding this enzyme retards the folding rate of imported Su9-DHFR (Matouschek et al., 1995; Rassow et al., 1995). Interestingly, folding intermediates reversibly accumulated at the molecular chaperones mtHsp70 and Hsp60 in the matrix when the function of cyclophilin was impaired. Mitochondrial cyclophilin, therefore, might promote protein folding in cooperation with chaperone proteins (Rassow et al., 1995; Rospert et al., 1996). Most likely, cyclophilin acts sequentially after release of the substrate protein from Hsp60 and has a complementary function in protein folding (von Ahsen et al., 1997). 4.2. Prevention of Heat Denaturation After heat denaturation of imported luciferase, the interaction of the protein with mtHsp70 transiently increases. The interaction of the chaperone with luciferase depends on the function of both Mge1p and Mdj1p (Prip-Buus et al., 1996). Aggregation of this protein is, however, not increased in mge1 and ssc1 mutants. Neither mtHsp70 nor Mge1p are, therefore, mandatory for transient prevention of heat-induced aggregation of luciferase. In contrast, Mdj1p plays a central role in this process. Apparently, Mdj1p can prevent aggregation already in the absence of functional Ssc1p or Mge1p (Prip-Buus et al., 1996), analogous to E.coli DnaJ which is able to maintain the soluble state of luciferase in vitro (Schröder et al., 1993). Since Mdj1p is not essential for growth at normal temperatures but becomes indispensable at elevated temperatures (Rowley et al., 1994), this role in the prevention of heat-induced aggregation might be the main function of this chaperone. Apart from Hsp60’s role in folding of newly imported proteins, the chaperonin probably functions in the prevention of protein aggregation when mitochondria are subjected to heat shock or other stress. The matrix processing peptidase aggregates in mitochondria when the chaperonin is inactivated (Glick et al., 1992). Surprisingly, matrix localized DHFR (whose folding was not affected by Hsp60 after import) unfolds and aggregates in Hsp60 deficient mitochondria at higher temperatures (Martin et al., 1992). Apparently, the inactivation of Hsp60 directly or indirectly affects the stability of these proteins under stress conditions. Mitochondrial Hsp78 is a member of the Hsp100/ClpB family, which is indicated in the disassembly of higher-order protein structures and aggregates (Schirmer et al., 1996). Deletion mutants of HSP78 show normal growth at all conditions (Leonhardt et al., 1993). When the HSP78 deletion is combined with mutants of mtHspTO, growth is impaired and mitochondrial DNA is lost. Hsp78 apparently functions in the prevention of aggregation of the mutant mtHsp70 molecules (Moczko et al., 1995). This feature of Hsp78 might explain the reported import defect of the double mutants (Schmitt et al., 1995). Hsp78 might have a homologous function to its yeast cytosolic counterpart Hsp104, by inducing thermotolerance and resolubilizing protein aggregates. Since additional chaperones might have overlapping functions, deletion of the HSP78 gene
The role of molecular chaperones in transport and folding
281
causes only problems when additional chaperones are missing. 4.3. Protein Turnover in Mitochondria Many organisms contain proteases that function in the ATP-dependent degradation of abnormal, misfolded, or incomplete polypeptides (reviewed by Goldberg, 1992). The yeast mitochondrial homolog of the E.coli Lon protease is encoded by the PIM1 gene (Suzuki et al., 1994; van Dyck et al., 1994). Pim1p, in collaboration with mtHsp70, is capable of degradation of misfolded imported proteins (Wagner et al., 1994), and is also somehow involved in the maintenance of wild type mitochondrial DNA. The respiratory chain complexes in mitochondria are built from subunits produced on both mitochondrial and cytoplasmic ribosomes. In the absence of the supply of one or the other subunit, non-assembled polypeptides are subject to rapid proteolysis (Kalnov et al., 1979). Recently a group of related ATPases (AAA-protein family) with a zincendopeptidase motif has been identified, that might function in the removal of nonassembled subunits (reviewed by Grivell, 1995). The three mitochondrial homologs of this family (Yme1p, Rca1p and Afg3p) are required for respiratory function (Thorsness et al., 1993; Guelin et al., 1994; Tauer et al., 1994; Tzagoloff et al., 1994). Interestingly, Rca1p and Afg3p (also termed Yta12 and Yta10, resp.) can, at high ATP levels, associate into a large membrane associated protein complex (Arlt et al., 1996). The activity of the protease depends on the formation of this complex, indicating that protease activity might be regulated by ATP hydrolysis. However, a mutation in the metalloprotease active site of Afg3p strongly reduces proteolysis, but does not affect growth, suggesting that its in vivo function is independent of its protease activity (Arlt et al., 1996; Guelin et al., 1996). Since the peculiar phenotypes of mutants in these genes (like increased escape of mitochondrial DNA to the nucleus; Thorsness et al., 1993) can not readily be explained by a function in protein degradation, it is suggested that these proteins might also have a function as chaperones (Grivell, 1995). Indeed, oligomerization of subunit 9 of the F1F0ATPase was shown to require the presence of the Yta10–12 complex (Artl et al., 1996). Analogous, a chaperone function has also been suggested for the AAA-protease of E.coli, FtsH (Akiyama et al., 1994). Further study on this fascinating new group of chaperones/proteases might provide new insides how the assembly and turnover of the mitochondrial membrane complexes is regulated.
5. CONCLUDING REMARKS Chaperones function in diverse processes during the biogenesis of mitochondria. Cytoplasmic Hsp70 preserves the unfolded conformation of precursor proteins after synthesis on cytoplasmic ribosomes, while other factors like MSF are involved in the targeting of preproteins to the receptors located at the mitochondrial outer membrane. Translocation of preproteins across the mitochondrial membranes is assisted by Hsp70 molecules of the mitochondrial matrix. A large array of matrix chaperones and other factors are required for the correct folding of imported proteins and assembly of the multi-protein complexes of the mitochondrial inner membrane. Although progress in the
Molecular chaperones and folding catalysts
282
understanding of these processes has been rapid for the past years, mechanistics details are still lacking. Perhaps most interesting is the question concerning the extra-ordinary function of mtHsp70 and other (co)-chaperones during the preprotein translocation process. Like before, the most effective method in studying these problems seems to be a combination of biochemistry and genetics in the yeast Saccharomyces cerevisiae. Since most major players during the translocation process are currently known, and conditional mutants of several of them are available, we now can fully concentrate on the elucidation of the mechanistics of these processes. Therefore, the coming years will undoubtedly provide additional information concerning the function of the diverse factors involved in mitochondrial biogenesis, and might yield exciting new views about the functioning of chaperones in mitochondrial biogenesis.
6. ACKNOWLEDGEMENTS We wish to thank Dr. M.Meijer for communicating results prior to publication. The work in the authors’ laboratory is supported by the Deutsche Forschungsgemeinschaft, the Sonderforschungsbereich 388 and the Fonds der Chemischen Industrie. P.J.T.D. was a recipient of a long term fellowship of the Human Frontier Science Program.
7. REFERENCES Aitken, A., Collinge, D.B., van Heusden, B.P.H., Isobe, T., Roseboom, P.H., Rosenfeld, G. and Soll, J. (1992). 14–3–3 proteins: a highly conserved, widespread family of eukaryotic proteins. Trends Biochem . Set., 17 , 498–501. Akiyama, Y., Shirai, Y. and Ito, K. (1994). Involvement of FtsH in protein assembly into and through the membrane. 2. Dominant mutations affecting FtsH functions. J. Biol. Chem. , 269 , 5225–5229. Alam, R., Hachiya, N., Sakaguchi, M., Kawabata, S., Iwanaga, S., Kitajima, M., Mihara, K. and Omura, T. (1994). cDNA cloning and characterization of mitochondrial import stimulation factor (MSF). purified from rat liver cytosol. J. Biochem. , 116 , 416–425. Arlt, H., Tauer, R., Feldmann, H., Neupert, W. and Langer, T. (1996). The YTA10–12 complex, an AAA protease with chaperone-like activity in the inner membrane of mitochondria. Cell , 85 , 875–885. Attardi, G. and Schatz, G. (1988). Biogenesis of mitochondria. Annu. Rev. Cell Biol , 4 , 289–333. Banecki, B. and Zylicz, M. (1996). Real time kinetics of the DnaK/DnaJ/GrpE molecular chaperone machine action. J. Biol Chem. , 271 , 6137–6143. Bauer, M.F., Sirrenberg, C., Neupert, W. and Brunner, M. (1996). Role of Tim23 as voltage sensor and presequence receptor in protein import into mitochondria. Cell , 87 , 33–41. Becker, K., Guiard, B., Rassow, J., Söllner, T. and Pfanner, N. (1992). Targeting of a chemically pure preprotein to mitochondria does not require the addition of a cytosolic signal recognition factor. J. Biol Chem. , 267 , 5637–5643. Becker, J., Walter, W., Yan, W. and Craig, E.A. (1996). Functional interaction of cytosolic hsp70 and a DnaJ-related protein, Ydj1p, in protein translocation in vivo. Mol
The role of molecular chaperones in transport and folding
283
Cell Biol , 16 , 4378–4386. Berthold, J., Bauer, M.F., Schneider, H.C., Klaus, C., Neupert, W. and Brunner, M. (1995). The MIM complex mediates preprotein translocation across the mitochondrial inner membrane and couples it to the mt-Hsp70/ATP driving system. Cell , 81 , 1085– 1094. Blom, J., Kübrich, M., Rassow, J., Voos, W., Dekker, P.J.T., Maarse, A.C., Meijer, M. and Pfanner, N. (1993). The essential yeast protein MIM44 (encoded by MPI1). is involved in an early step of preprotein translocation across the mitochondrial inner membrane. Mol Cell Biol , 13 , 7364–7371. Blom, J., Dekker, P.J.T. and Meijer, M. (1995). Functional and physical interactions of components of the yeast mitochondrial inner-membrane import machinery (MIM). Eur. J. Biochem. , 232 , 309–314. Blond-Elguindi, S., Cwirla, S.E., Dower, W.J., Lipshutz, R.J., Sprang, S.R., Sambrook, J.F. and Gething, M.J. (1993). Affinity panning of a library of peptides displayed on bacteriophages reveals the binding specificity of Bip, Cell , 75 , 717–728. Bolliger, L., Deloche, O., Glick, B.S., Georgopoulos, C., Jenö, P., Kronidou, N., Horst, M., Morishima, N. and Schatz, G. (1994). A mitochondrial homolog of bacterial GrpE interacts with mitochondrial hsp70 and is essential for viability, EMBO J. , 13 , 1998– 2006. Bolliger, L., Junne, T., Schatz, G. and Lithgow, T. (1995). Acidic receptor domains on both sides of the outer membrane mediate translocation of precursor proteins into yeast mitochondria. EMBO J. , 14 , 6318–6326. Bömer, U., Meijer, M., Maarse, A.C., Dekker, P.J.T, Pfanner, N. and Rassow, J. (1997). Multiple interactions of components mediating preprotein translocation across the inner mitochondrial membrane. EMBO J. , 16 , 2205–2216. Buchberger, A., Theyssen, H., Schröder, H., McCarty, J.S., Virgallita, G., Milkereit, P., Reinstein, J. and Bukau, B. (1995). Nucleotide-induced conformational changes in the ATPase and substrate binding domains of the DnaK chaperone provide evidence for interdomain communication. J. Biol. Chem. , 270 , 16903–16910. Caplan, A.J., Tsai, J., Casey, P.J. and Douglas, M.G. (1992a). Farnesylation of YDJ1p is required for function at elevated growth temperatures in S. cerevisiae. J. Biol. Chem. , 267 , 18890–18895. Caplan, A.J., Cyr, D.M. and Douglas, M.G. (1992b). YDJ1p facilitates polypeptide translocation across different intracellular membranes by a conserved mechanism. Cell , 71 , 1143–1155. Cheng, M.Y., Hartl, F.U., Martin, J., Pollock, R.A., Kalousek, F., Neupert, W., Hallberg, E.M., Hallberg, R.L. and Horwich, A.L. (1989). Mitochondrial heat-shock protein hsp60 is essential for assembly of proteins imported into yeast mitochondria, Nature , 337 , 620–625. Craig, E.A., Baxter, B.K., Becker, J., Halladay, J. and Ziegelhoffer (1994). Cytosolic hsp70s of Saccharomyces cerevisiae: roles in protein synthesis, protein translocation, proteolysis, and regulation. In R.I. Morimoto, A.Tissières, and C.Georgopoulos, (eds.), The biology of heat shock proteins and molecular chaperones , CSHL Press, NY, pp. 31–52. Cyr, D.M., Stuart, R.A. and Neupert, W. (1993). A matrix ATP requirement for presequence translocation across the inner membrane of mitochondria. J. Biol. Chem. , 268 , 23751–23754. Cyr, D.M., Langer, T. and Douglas, M.G. (1994). DnaJ-like proteins: molecular chaperones and specific regulators of hsp70. Trends Biochem. Sci. , 19 , 176–181.
Molecular chaperones and folding catalysts
284
Dekker, P.J.T., Müller, H., Rassow, J. and Pfanner, N. (1996). Characterization of the preprotein translocase of the outer mitochondrial membrane by blue native electrophoresis. Biol. Chem. , 377 , 535–538. Dekker, P.J.T., Martin, F., Maarse, A., Bömer, U., Müller, H., Guiard, B., Meijer, M., Rassow, J. and Pfanner, N. (1997). The Tim core complex defines the number of mitochondrial translocation contact sites and can hold arrested preproteins in the absence of matix Hsp70-Tim44 . EMBO J. , 16 , 5408–5419. Dekker, P.J.T. and Pfanner, N. (1997). Role of mitochondrial GrpE and phosphate in the ATPase cycle of matrix Hsp70. J. Mol. Biol. , 270 , 321–327. Deshaies, R.J., Koch, B.D., Werner-Washburne, M., Craig, A.E. and Schekman, R. (1988). A subfamily of stress proteins facilitates translocation of secretory and mitochondrial precursor polypeptides. Nature , 332 , 800–805 Dietmeier, K., Hönlinger, A., Bömer, U., Dekker, P.J.T, Eckerskorn, C., Lottspeich, F., Kübrich, M. and Pfanner, N. (1997). Tom5 frictionally links mitochondrial preprotein receptors to the general import pore. Nature , 388 , 195–200. Eilers, M. and Schatz, G. (1986). Binding of a specific ligand inhibits import of a purified precursor protein into mitochondria. Nature , 322 , 228–232. Eilers, M. and Schatz, G. (1988). Protein unfolding and the energetics of protein translocation across biological membranes. Cell , 52 , 481–483. Endo, T., Mitsui, S. and Roise, D. (1995). Mitochondrial presequences can induce aggregation of unfolded proteins. FEBS Lett. , 359 , 93–96. Gambill, B.D., Voos, W., Kang, P.J., Mao, B., Langer, T., Craig, E.A., Pfanner, N. (1993). A dual role for mitochondrial heat shock protein 70 in membrane translocation of preproteins. J. Cell Biol. , 123 , 109–117. Gavel, Y. and von Heijne, G. (1992). The distribution of charged amino acids in mitochondrial innermembrane proteins suggests different modes of membrane integration for nuclearly and mitochondrially encoded proteins. Eur. J. Biochem. , 205 , 1207–1215. Gething, M.J. and Sambrook, J. (1992). Protein folding in the cell. Nature , 355 , 33–45. Glick, B.S., Brandt, A., Cunningham, K., Müller, S., Hallberg, R.L. and Schatz, G. (1992). Cytochromesc1 and b2 are sorted to the intermembrane space of yeast mitochondria by a stop-transfer mechanism. Cell , 69 , 809–822. Glick, B.S. (1995). Can hsp70 proteins act as force-generating motors? Cell , 80 , 11–14. Goldberg, A.L. (1992). The mechanism and functions of ATP-dependent proteases in bacterial and animal cells. Eur. J. Biochem. , 203 , 9–23. Gratzer , S., Lithgow, T., Bauer, R.E., Lamping, E., Paltauf, F., Kohlwein, S.D., Haucke, V., Junne, T., Schatz, G. and Horst, M. (1995). Mas37p, a novel receptor subunit for protein import into mitochondria. J. Cell Biol. , 129 , 25–34. Grivell, L.A. (1995). Nucleo-mitochondrial interactions in mitochondrial gene expression. Crit. Rev. Biochem. Mol. Biol. , 30 , 121–164. Guélin, E., Rep, M. and Grivell, L.A. (1994). Sequence of the AFG3 gene encoding a new member of the FtsH/Yme1/Tma subfamily of the AAA-protein family. Yeast , 10 , 1389–1394. Guélin, E., Rep, M. and Grivell, L.A. (1996). Afg3p, a mitochondrial ATP-dependent metalloprotease, is involved in the degradation of mitochondrially-encoded Cox1, Cox3, Cob, Su6, Su8 and Su9 subunits of the inner membrane complexes III, IV and V. FEBS Lett . 381 , 42–46. Ha, J.-H. and McKay, D.B. (1995). Kinetics of nucleotide-induced changes in the tryptophan fluorescence of the molecular chaperone Hsc70 and its subfragments
The role of molecular chaperones in transport and folding
285
suggest the ATP-induced conformational change follows initial ATP binding. Biochemistry , 34 , 11635–11644. Hachiya, N., Alam, R., Sakasegawa, Y., Sakaguchi, M., Mihara, K. and Omura, T. (1993). A mitochondrial import factor purified from rat liver cytosol is an ATPdependent conformational modulator for precursor proteins. EMBO J. , 12 , 1579– 1586. Hachiya, N., Komiya, T., Alam, R., Iwahashi, J., Sakaguchi, M., Omura, T. and Mihara, K. (1994). MSF, a novel cytoplasmic chaperone which functions in precursor targeting to mitochondria. EMBO J. , 13 , 5146–5154. Hachiya, N., Mihara, K., Suda, K., Horst, M., Schatz, G. and Lithgow, T. (1995). Reconstitution of the initial steps of mitochondrial protein import. Nature , 376 , 705– 709. Hallberg, E.M., Shu, Y. and Hallberg, R.L. (1993). Loss of mitochondrial hsp60 function: nonequivalent effects on matrix-targeted and intermembrane-targeted proteins. Mol. Cell. Biol , 13 , 3050–3057. Haucke, V., Lithgow, T., Rospert, S., Hahne, K. and Schatz, G. (1995). The yeast mitochondrial protein import receptor Mas20p binds precursor proteins through electrostatic interaction with the positively charged presequence. J. Biol. Chem. , 270 , 5565–5570. Herrmann, J.M., Stuart, R.A., Craig, E.A. and Neupert, W. (1994). Mitochondrial heatshock protein 70, a molecular chaperone for proteins encoded by mitochondrial DNA. J. Cell Biol. , 127 , 893–902. Hines, V., Brandt, A., Griffiths, G., Horstmann, H., Brütsch, H. and Schatz, G. (1990). Protein import into mitochondria is accelerated by the outer membrane protein MAS70. EMBO J. , 9 , 3191–3200. Höhfeld, J. and Hartl, F.U. (1994). Role of chaperonin cofactor hsp10 in protein folding and sorting in yeast mitochondria. J. Cell Biol , 126 , 305–315. Hönlinger, A., Kübrich, M., Moczko, M., Gärtner, F., Mallet, L., Bussereau, F, Eckerskorn, C., Lottspeich, F, Dietmeier, K., Jacquet, M. and Pfanner, N. (1995). The mitochondrial receptor complex: MOM22 is essential for cell viability and directly interacts with preproteins. Mol. Cell. Biol. , 15 , 3382–3389. Horst, M., Oppliger, W., Feifel, B., Schatz, G. and Glick, B.S. (1996). The mitochondrial protein import motor: Dissociation of mitochondrial hsp70 from its membrane anchor requires ATP binding rather than ATP hydrolysis. Prot. Sci. , 5 , 759–767. Horst, M., Oppliger, W., Rospert, S., Schonfeld, H.J., Schatz, G. and Azem, A. (1997). Sequential action of two hsp70 complexes during protein import into mitochondria. EMBO J. , 16 , 1842–1849. Ikeda, E., Yoshida, S., Mitsuzawa, H., Uno, I. and Toh-e, A. (1994). YGE1 is a yeast homologue of Escherichia coli grpE and is required for maintenance of mitochondrial functions. FEBS Lett. , 339 , 265–268. Johnston, M. (1996). Genome sequencing: The complete code for a eukaryotic cell. Curr. Biol. , 6 , 500–503. Kalnov, S.L., Novikova, L.A., Zubatov, A.S. and Luzikov, V.N. (1979). Proteolysis of the products of mitochondrial protein synthesis in yeast mitochondria and submitochondrial particles. Biochem. J. , 182 , 195–202. Kang, P.J., Ostermann, J., Shilling, J., Neupert, W., Craig, E.A. and Pfanner, N. (1990). Requirement for hsp70 in the mitochondrial matrix for translocation and folding of precursor proteins. Nature , 348 , 137–143. Kerscher, O., Holder, J., Srinivasan, M., Leung, R.S. and Jensen, R.E. (1997). The
Molecular chaperones and folding catalysts
286
Tim54p-Tim22p complex mediates insertion of proteins into the mitochondrial inner membrane. J. Cell Biol. , 139 , 1663–1675. Komiya, T., Sakaguchi, M. and Mihara, K. (1996). Cytoplasmic chaperones determine the targeting pathway of precursor proteins to mitochondria. EMBO J. , 15 , 399–407. Kronidou, N.G., Oppliger, W., Bolliger, L., Hannavy, K., Glick, B.S., Schatz, G. and Horst, M. (1994). Dynamic interaction between Isp45 and mitochondrial hsp70 in the protein import system of the yeast mitochondrial inner membrane. Proc. Natl. Acad. Sci. USA , 91 , 12818–12822. Kübrich, M., Keil, P., Rassow, J., Dekker, P.J.T., Blom, J., Meijer, M. and Pfanner, N. (1994). The polytopic inner membrane proteins MIM17 and MIM23 operate at the same preprotein import site. EEBS Lett. , 349 , 222–228. Laloraya, S., Gambill, B.D. and Craig, E.A. (1994). A role for a eukaryotic GrpE-related protein, Mge1p, in protein translocation. Proc. Natl. Acad. Sci. USA , 91 , 6481–6485. Laloraya, S., Dekker, P.J.T., Voos, W., Craig, E.A. and Pfanner, N. (1995). Mitochondrial GrpE modulates the function of matrix Hsp70 in translocation and maturation of preproteins. Mol. Cell. Biol. , 15 , 7098–7105. Langer, T., Lu, C., Echols, H., Flanagan, J., Hayer, M.K. and Hartl, F.-U. (1992). Successive action of DnaK, DnaJ and GroEL along the pathway of chaperonemediated protein folding. Nature , 356 , 683–689. Leonhardt, S.A., Fearon, K., Danese, P.N. and Mason, T.L. (1993). HSP78 encodes a yeast mitochondrial heat shock protein in the Clp family of ATP-dependent proteases . Mol. Cell. Biol. , 13 , 6304–6313. Liberek, K., Marszalek, J., Ang, D., Georgopoulos, C. and Zylicz, M. (1991). Escherichia coli DnaJ and GrpE heat shock proteins jointly stimulate ATPase activity of DnaK. Proc. Natl. Acad. Sci. USA , 88 , 2874–2878. Lill, R. and Neupert, W. (1996). Mechanisms of protein import across the mitochondrial outer membrane. Trends Cell Biol. , 6 , 56–61. Lithgow, T., Junne, T., Wachter, C. and Schatz, G. (1994). Yeast mitochondria lacking the two import receptors Mas20p and Mas70p can efficiently and specifically import precursor proteins. J. Biol. Chem. , 269 , 15325–15330. Lithgow, T., Glick, B.S. and Schatz, G. (1995). The protein import receptor of mitochondria. Trends Biochem. Sci. , 20 , 98–101. Maarse, A.C., Blom, J., Grivell, L.A. and Meijer, M. (1992). MPI1, an essential gene encoding a mitochondrial membrane protein, is possibly involved in protein import into yeast mitochondria. EMBO J. , 11 , 3619–3628. Manning-Krieg, U.C., Scherer, P.E. and Schatz, G. (1991). Sequential action of mitochondrial chaperones in protein import into the matrix. EMBO J. , 10 , 3273–3280. Martin, J., Mahlke, K. and Pfanner, N. (1991). Role of an energized inner membrane in mitochondrial protein import: Dy drives the movement of presequences. J. Biol. Chem. , 266 , 18051–18057. Martin, J., Horwich, A.L. and Hartl, F.U. (1992). Prevention of protein denaturation under heat stress by the chaperonin hsp60. Science , 258 , 995–998. Matouschek, A., Rospert, S., Schmid, K., Glick, B.S. and Schatz, G. (1995). Cyclophilin catalyzes protein folding in yeast mitochondria. Proc. Natl. Acad. Sci. USA , 92 , 6319–6323. Matouschek, A., Azem, A., Ratliff, K., Glick, B.S., Schmid, K. and Schatz, G. (1997). Active unfolding of precursor proteins during mitochondrial protein import. EMBO J. , 16 , 6727–6736. McCarty, J.S., Buchberger, A., Reinstein, J. and Bukau, B. (1995). The role of ATP in
The role of molecular chaperones in transport and folding
287
the functional cycle of the DnaK chaperone system. J. Mol. Biol. , 249 , 126–137. Miao, B., Davis, J.E. and Craig, E.A. (1997). Mge1 functions as a nucleotide release factor for Ssc1, a mitochondrial Hsp70 of Saccharomyces cerevisiae. J. Mol. Biol , 265 , 541–552. Mihara, K. and Omura, T. (1996). Cytoplasmic chaperones in precursor targeting to mitochondria: the role of MSF and hsp70. Trends Cell Biol , 6 , 104–108. Moczko, M., Ehmann, B., Gärtner, F., Hönlinger, A., Schäfer, E. and Pfanner, N. (1994). Deletion of the receptor MOM19 strongly impairs import of cleavable preproteins into Saccharomyces cerevisiae mitochondria. J. Biol Chem. , 269 , 9045–9051. Moczko, M., Schönfisch, B., Voos, W., Pfanner, N. and Rassow, J. (1995). The mitochondrial ClpB homolog Hsp78 cooperates with matrix Hsp70 in maintenance of mitochondrial function. J. Mol Biol , 254 , 538–543. Moczko, M., Bömer, U., Kübrich, M., Zufall, N. Hönlinger, A. and Pfanner, N. (1997). The intermembrane space domain of mitochondrial Tom22 functions as a trans binding site for preproteins with N-terminal targeting sequences. Mol Cell Biol , 17 , 6574– 6584. Murakami, H., Pain, D. and Blobel, G. (1988). 70-kDa heat-shock related protein is one of at least two distinct cytosolic factors stimulating protein import into mitochondria. J. Cell Biol , 107 , 2051–2057. Murakami, K. and Mori, M. (1990). Purified presequence binding factor (PBF) forms an import-competent complex with a purified mitochondrial precursor protein. EMBO J. , 9 , 3201–3208. Murakami, K., Tanase, S., Morino, Y. and Mori, M. (1992). Presequence binding factordependent and -independent import of proteins into mitochondria. J. Biol Chem. , 257 , 13119–13122. Nakai, M., Kato, Y., Ikeda, E., Toh-e, A. and Endo, T. (1994). Yge1p, a eukaryotic GrpE homolog, is localized in the mitochondrial matrix and interacts with mitochondrial Hsp70. Biochem. Biophys. Res. Comm. , 200 , 435–442. Ostermann, J., Horwich, A.L., Neupert, W. and Hartl, F.-U. (1989). Protein folding in mitochondria requires complex formation with hsp60 and ATP hydrolysis. Nature , 341 , 125–129. Pfanner, N., Tropschug, M. and Neupert, W. (1987). Mitochondrial protein import: nucleoside triphosphates are involved in conferring import competence to precursors. Cell , 49 , 815–823. Pfanner, N., Rassow, J., van der Klei, I.J. and Neupert, W. (1992). A dynamic model of the mitochondrial protein import machinery. Cell , 68 , 999–1002. Pfanner, N. and Meijer, M. (1995). Pulling in the proteins. Curr. Biol , 5 , 132–135. Pfanner, N., Douglas, M.G., Endo, T., Hoogenraad, N.J., Jensen, R.E., Meijer, M., Neupert, W., Schatz, G., Schmitz, U.K. and Shore, G.C. (1996). Uniform nomenclature for the protein transport machinery of the mitochondrial membranes. Trends Biochem. Sci. , 21 , 51–52. Prip-Buus, C., Westermann, B., Schmitt, M., Langer, T., Neupert, W. and Schwarz, E. (1996). FEBS Lett. , 380 , 142–146. Ramage, L., Junne, K., Hahne, K., Lithgow, T. and Schatz, G. (1993). Functional cooperation of mitochondrial protein import receptors in yeast. EMBO J. , 12 , 4115– 4123. Rassow, J., Hartl, F.U., Guiard, B., Pfanner, N. and Neupert, W. (1990). Polypeptides traverse the mitochondrial envelope in an extended state. FEBS Lett. , 275 , 190–194. Rassow, J., Maarse, A.C., Krainer, E., Kübrich, M., Müller, H., Meijer, M., Craig, E.A.
Molecular chaperones and folding catalysts
288
and Pfanner, N. (1994) Mitochondrial protein import: biochemical and genetic evidence for interaction of matrix hsp70 and the inner membrane protein MIM44. J. Cell Biol. , 127 , 1547–1556. Rassow, J., Mohrs, K., Koidl, S., Barthelmess, I.E., Pfanner, N. and Tropschug, M. (1995). Cyclophilin 20 is involved in mitochondrial protein folding in cooperation with molecular chaperones hsp70 and hsp60 . Mol. Cell. Biol. , 15 , 2654–2662. Rospert, S., Looser, R., Dubaquié, Y., Matouschek, A., Glick, B.S. and Schatz, G. (1996). Hsp60-independent protein folding in the matrix of yeast mitochondria. EMBO J. , 15 , 764–774. Rowley, N., Prip-Buus, C., Westermann, B., Brown, C., Schwarz, E., Barrell, B. and Neupert, W. (1994). Mdj1p, a novel chaperone of the dnaJ faily, is involved in mitochondrial biogenesis and protein folding. Cell , 77 , 249–259. Scherer, P.E., Krieg, U.C., Hwang, S.T., Vestbeber, D. and Schatz, G. (1990). A precursor protein partly translocated into yeast mitochondria is bound to a 70 kDa mitochondrial stress protein. EMBO J. , 9 , 4315–4322. Scherer, P.E., Manning-Krieg, U.C., Jenö, P., Schatz, G. and Horst, M. (1992). Identification of a 45-kDa protein at the protein import site of the yeast mitochondrial inner membrane. Proc. Natl. Acad. Sci. USA , 89 , 11930–11934. Schilke, B., Forster, J., Davis, J., James, P., Walter, W., Laloraya, S., Johnson, J., Miao, B. and Craig, E. (1996). The cold sensitivity of a mutant of Saccharomyces cerevisiae lacking a mitochondrial heat shock protein 70 is suppressed by loss of mitochondrial DNA. J. Cell Biol. , 134 , 603–613. Schirmer, E.G., Glover, J.R., Singer, M.A. and Lindquist, S. (1996). Hsp100/Clp proteins: a common mechanism explains diverse functions. Trends Biochem. Sci. , 21 , 289–296. Schleyer, M. and Neupert, W. (1985). Transport of proteins into mitochondria: translocation intermediates spanning contact sites between outer and inner membranes. Cell , 43 , 330–350. Schmid, D., Baici, A., Gehring, H. and Christen, P. (1994). Kinetics of molecular chaperone action. Science , 263 , 971–973. Schmitt, M., Neupert, W. and Langer, T. (1995). Hsp78, a Clp homologue within mitochondria, can substitute for chaperone functions of mt-hsp70. EMBO J. , 14 , 3434–3444. Schneider, H.C., Berthold, J., Bauer, M.F., Dietmeier, K, Guiard, B., Brunner, M. and Neupert, W. (1994). Mitochondrial hsp70/MIM44 complex facilitates protein import. Nature , 371 , 768–774. Schröder, H., Langer, T., Hartl, F.U. and Bukau, B. (1993). DnaK, DnaJ and GrpE form a cellular chaperone machinery capable of repairing heat-induced protein damage. EMBO J. , 12 , 4137–4144. Sirrenberg, C., Bauer, M.F., Guiard, B., Neupert, W. and Brunner, M. (1996). Import of carrier proteins into the mitochondrial inner membrane mediated by Tim22. Nature , 384 , 582–585. Steger, H.F., Söllner, T., Kiebler, M., Dietmeier, K., Pfaller, R., Trülzsch, K.S., Tropschug, M., Neupert, W. and Pfanner, N. (1990). Import of ADP/ATP carrier into mitochondria: two receptors act in parallel. J. Cell Biol. , 111 , 2353–2363. Stuart, R.A., Nicholson, D.W. and Neupert, W. (1990). Early steps in mitochondrial protein import: receptor functions can be substituted by the membrane insertion activity of apocytochrome c . Cell , 60 , 31–43. Stuart, R.A., Cyr, D.M., Craig, E.A. and Neupert, W. (1994a). Mitochondrial molecular
The role of molecular chaperones in transport and folding
289
chaperones: their role in protein translocation. Trends Biochem. Sci. , 19 , 87–92. Stuart, R.A., Gruhler, A., van der Klei, I., Guiard, B., Koll, H. and Neupert, W. (1994b). The requirement of matrix ATP for the import of precursor proteins into the mitochondrial matrix and intermembrane space. Eur. J. Biochem. , 220 , 9–18. Suzuki, C.K., Suda, K., Wang, N. and Schatz, G. (1994). Requirement for the yeast gene LON in intramitochondrial proteolysis and maintenance of respiration. Science , 264 , 273–276. Szabo, A., Langer, T., Schröder, H., Flanagan, J., Bukau, B. and Hartl, F.U. (1994). The ATP hydrolysis-dependent reaction cycle of the Escherichia coli Hsp70 system-DnaK, DnaJ, and GrpE. Proc. Natl. Acad. Sci. USA , 91 , 10345–10349. Tarassov, I., Entelis, N. and Martin, R.P. (1995). Mitochondrial import of a cytoplasmic lysine-tRNA in yeast is mediated by cooperation of cytoplasmic and mitochondrial lysyl-tRNA synthetases. EMBO J . 14 , 3461–3471. Tauer, R., Mannhaupt, G., Schnall, R., Pajic, A., Langer, T. and Feldmann, H. (1994). Yta10, a member of a novel ATPase family in yeast, is essential for mitochondrial function. FEBS Lett. , 353 , 197–200. Thorsness, P.E., White, K.H. and Fox, T.D. (1993). Inactivation ofYME1, a member of the FtsH-SEC18-PAS1-CDC48 family of putative ATPase-encoding genes, causes increased escape of DNA from mitochondria in Saccharomyces cerevisiae. Mol. Cell. Biol. , 13 , 5418–5426. Tzagoloff, A., Yue, J., Jang, J. and Paul, M.F. (1994). A new member of a family of ATPases is essential for assembly of mitochondrial respiratory chain and ATP synthetase complexes in Saccharomyces cerevisiae . J. Biol. Chem. , 269 , 26144– 26151. Ungermann, C., Neupert, W. and Cyr, D.M. (1994). The role of Hsp70 in conferring unidirectionality on protein translocation into mitochondria. Science , 266 , 1250– 1253. Ungermann, C., Guiard, B., Neupert, W. and Cyr, D.M. (1996). The - and Hsp70/MIM44 dependent reaction cycle driving early steps of protein import into mitochondria. EMBO J. , 15 , 735–744. van Dyck, L., Pearce, D.A. and Sherman, F. (1994). PIM1 encodes a mitochondrial ATPdependent protease that is required for mitochondrial function in the yeast Saccharomyces cerevisiae. J. Biol. Chem. , 269 , 238–242. Vestweber, D. and Schatz, G. (1989). DNA-protein conjugates can enter mitochondria via the protein import pathway. Nature , 338 , 170–172. von Ahsen, O., Voos, W., Henninger, H. and Pfanner, N. (1995). The mitochondrial protein import machinery: role of ATP in dissociation of Hsp70-Mim44 complex. J. Biol. Chem. , 270 , 29848–29853. von Ahsen, O., Tropschug, M., Pfanner, N. and Rassow, J. (1997). The chaperonin cycle cannot substitute for prolyl isomerase activity, but GroEL alone promotes productive folding of a cyclophilin-sensitive substrate to a cyclophilin-resistant form. EMBO J. , 16 , 4568–4578. von Heijne, G. (1986). Mitochondrial targeting sequences may form amphiphilic helices. EMBO J. , 5 , 1335–1342. Voos, W., Gambill, B.D., Guiard, B., Pfanner, N. and Craig, E.A. (1993). Presequence and mature part of preproteins strongly influence the dependence of mitochondrial protein import on heat shock protein 70 in the matrix. J. Cell Biol , 123 , 119–126. Voos, W., Gambill, B.D., Laloraya, S., Ang, D., Craig, E.A. and Pfanner, N. (1994). Mitochondrial GrpE is present in a complex with hsp70 and preproteins in transit
Molecular chaperones and folding catalysts
290
across membranes. Mol. Cell. Biol , 14 , 6627–6634. Voos, W., von Ahsen, O., Müller, H., Guiard, B., Rassow, J. and Pfanner, N. (1996). Differential requirement for the mitochondrial Hsp70-Tim44 complex in the unfolding and translocation of preproteins. EMBO J. , 15 , 2668–2677. Wachter, C., Schatz, G. and Glick, B.S. (1992). Role of ATP in the intramitochondrial sorting of cytochrome c1 and the adenine nucleotide translocator. EMBO J. , 11 , 4787–4794. Wachter, C., Schatz, G. and Glick, B.S. (1994). Protein import into mitochondria: the requirement for external ATP is precursor-specific whereas intramitochondrial ATP is universally needed for translocation into the matrix. Mol. Cell. Biol , 5 , 465–474. Wagner, I., Arlt, H., van Dyck, L., Langer, T. and Neupert, W. (1994). Molecular chaperones cooperate with PIM1 protease in the degradation of misfolded proteins in mitochondria. EMBO J. , 13 , 5135–5145. Wei, J., Gaut, J.R. and Hendershot, L.M. (1995). In vitro dissociation of Bip-peptide complexes requires a conformational change in Bip after ATP binding but does not require ATP hydrolysis. J. Biol. Chem. , 270 , 26677–26682. Westermann, B., Prip-Buus, C., Neupert, W. and Schwarz, E. (1995). The role of the GrpE homologue, Mge1p, in mediating protein import and folding in mitochondria. EMBO J. , 14 , 3452–3460.
12. PROTEIN IMPORT INTO AND FOLDING WITHIN CHLOROPLASTS EVA MUCKEL andJÜRGEN SOLL* Botanisches Institut, Christian-Albrechts-Universität, Am Botanischen Garten 1–9, D-24118 Kiel, Federal Republic of Germany
1. Introduction 2. Cytosolic Factors Involved in the Translocation of Precursor Proteins into Chloroplasts 2.1. Conformation of Precursor Proteins and Cytosolic Factors 2.2. Conformational Changes of Precursor Proteins at the Chloroplast Envelope 3. Molecular Chaperones in the Translocation Machinery of the Outer Envelope of Chloroplasts 3.1. Components of the Translocation Machinery 3.2. Characteristics of HSP70 Homologues Present in the Outer Envelope of Chloroplasts 4. Stromal Chaperones 4.1. Stromal HSP70 Homologues 4.2. Chloroplast Chaperonins 5. Role of Stromal Factors in the Translocation of Proteins into or Across the Thylakoid Membrane 5.1. Targeting to the Thylakoid Membrane 5.2. Targeting to the Thylakoid Lumen 6. Molecular Chaperones in the Thylakoid Lumen 7. References
1. INTRODUCTION Molecular chaperones seem to accompany plastid destined precursor proteins from their origin in the cytosol until they reach the outmost corner of the organelle, e.g. the thylakoid lumen. By then they will have encountered cytosolic, envelope membrane bound, stromal and lumenal chaperones of either the HSP70, Cpn60/ Cpn10, Sec or SRP family. While none of these systems is unique to plastids, the wide variety which is present simultaneously is surprising. This review will therefore focus on the diversity of chaperones involved in protein translocation and trafficking in plastids and point to the
Molecular chaperones and folding catalysts
292
uniqueness of several of the involved systems in comparison to other organelles and organisms. Excellent reviews on the mechanistic background of chaperone functions have appeared very recently and the reader is referred to them for further details (Gatenby and Viitanen, 1994; Hartl, 1996; Boston et al., 1996; Gatenby, 1996).
2. CYTOSOLIC FACTORS INVOLVED IN THE TRANSLOCATION OF PRECURSOR PROTEINS INTO CHLOROPLASTS 2.1. Conformation of Precursor Proteins and Cytosolic Factors It is generally agreed that translocation of precursor proteins across biological membranes cannot occur in a folded state. This notion is mostly supported by the finding that translocation is blocked, when a mature folded conformation of precursor proteins is stabilized (Randall and Hardy, 1986; Eilers and Schatz, 1988). Accordingly the conformational state of precursor proteins in the cytosol is believed to be different from that of the native functional protein. A translocation competent conformation has been correlated with a certain “looseness” of structure, characterized for example by an increase in sensitivity to proteolytic digestion (for review see: Meyer, 1988). This has led to the discovery of cytosolic factors that are considered to be involved in maintaining a precursor protein in a transport competent conformation before the import (Deshaies et al., 1988 for review). These factors most probably function in preventing aggregation and folding to native structure of precursor proteins, therefore extending the life span of a translocation competent conformation. There is abundant evidence that chloroplast presequences are substantially, if not completely unfolded (von Heijne and Nishikawa, 1991; Theg and Geske, 1992; Pilon et al., 1992a; America et al., 1994; Creighton et al., 1995; Walker et al., 1996). For the only precursor protein to be studied in detail, ferredoxin, it was found that it is even the entire precursor protein that is relatively unfolded (Pilon et al., 1992a). This seems not to be due to an interaction with a cytosolic factor, as no stable interaction of the precursor of ferredoxin, synthesized in a wheat germ lysate, with a cytosolic factor was observed (Pilon et al., 1992b). In contrast, the precursors of 5-enolpyruvylshikimate-3-phosphate synthase (pEPSPS) and NADPH: protochlorophyllide oxidoreductase (pPOR) were found to be enzymatically active and thus presumably correctly folded after synthesis in an in vitro-transcription/translation system (Della-Cioppa et al., 1986; Reinbothe et al., 1995). Furthermore, protease sensitivity studies on the precursor of the 23 kDa thylakoid lumen protein of the oxygen evolving complex (Creighton et al., 1995) and on ricin A chain in a chimeric precursor protein (Walker et al., 1996) suggested that these proteins are comprised of an unfolded presequence together with a folded mature moiety. There are only few studies on the involvement of cytosolic factors in the translocation of precursor proteins into chloroplasts. Such studies are complicated by the fact that lysates used for in vitro transcription/translation of precursor proteins contain the complete set of proteins necessary to produce an import competent precursor protein (Zimmermann et al., 1988). Using the purified precursor for the light harvesting chlorophyll a/b-binding protein (pLHCP) overexpressed in E.coli it was shown that
Protein import into and folding within chloroplasts
293
translocation is dependent on the presence of soluble proteins from leaf extract (Waegemann et al., 1990). Purified HSP70 could only partially substitute for the leaf extract, thus, at least two components were assumed to be involved in maintaining an import competent conformation of pLHCP. Association of pLHCP with the soluble factor most likely resulted in an unfolding as the import competent precursor was more susceptible to protease treatment. In contrast, no requirement of cytosolic factors for the translocation of overexpressed precursor proteins into chloroplasts was observed in vitro in other studies (Pilon et al., 1992b; Yuan et al., 1993). Chemical denaturing of precursor proteins usually results in a transient translocation activity that does not require the involvement of any cytosolic factor. Refolding and the accompanying loss of translocation competence occurs after dilution of the denaturing agent and seems to be dependent on the individual protein and the experimental conditions. Addition of cytosolic factors during refolding can preserve import competence over a greater period of time. Thus, cytosolic chaperones do not seem to play a general role in translocation of precursor proteins into chloroplasts in vitro, but may be required to maintain an import competent conformation of highly hydrophobic precursor proteins that have a tendency to aggregate like LHCP (Yuan et al., 1993). The in vivo situation could be different from what is found in vitro. Specific factors like the mitochondrial import stimulation factor (MSF), that does not only unfold precursor proteins, but also targets unfolded precursor proteins to mitochondria (reviewed in Mihara and Omura, 1996), have at present not been identified in translocation of chloroplastic precursor proteins. It was recently shown that a plant specific cytosolic protein kinase phosphorylates chloroplast precursor proteins, but not mitochondrial or peroxisomal preproteins, indicating that highly specific cytosolic proteins exist, which are involved in protein translocation (Waegemann and Soll, 1996). If this protein kinase co-operates with other cytosolic or envelope bound factors, e.g. HSP70 homologues, remains to be established. 2.2. Conformational Changes of Precursor Proteins at the Chloroplast Envelope There is increasing evidence that mechanisms inducing conformational changes, e.g. unfolding, are associated with the cytosolic surface of chloroplast envelope membranes (Guera et al., 1993; Endo et al., 1994; Reinbothe et al., 1995; Walker et al., 1996). This possibility was also considered for the import of precursor proteins into mitochondria (Eilers and Schatz, 1988). However, in contrast to mitochondria, chimeric precursor proteins containing dihydrofolate reductase (DHFR) fused to a chloroplast transit peptide or a chloroplast precursor protein were not prevented from being imported even if the folded conformation was stabilized by binding of the substrate methotrexate (MTX) (Guera et al., 1993; Endo et al., 1994; America et al., 1994). The presence of MTX did not affect the binding of precursor proteins to chloroplasts but slowed down the subsequent import considerably (America et al., 1994). These results are consistent with the data reported on the import of pEPSPS into lettuce chloroplasts. Complexation of pEPSPS with the herbicide glyphosate and the shikimate-3-phosphate substrate, thus stabilizing a folded conformation of the precursor protein, resulted in a reduced rate but not a complete block of import (Della-Cioppa and Kishore, 1988). A similar observation
Molecular chaperones and folding catalysts
294
was made by Reinbothe et al. (1995) with the precursor of POR. In the presence of its substrate protochlorphyllide an import incompetent conformation was stabilized due to a masking of the transit peptide. However, if pPOR was prebound to chloroplasts and then complexed with substrate, the translocation process was slowed down but not inhibited. It should be emphasized that in the mentioned studies it was not determined whether the folded polypeptide chains were still import competent or whether import occurred by an incompletely folded precursor subpopulation. An unfolding of tightly folded precursor proteins occurring after binding to the chloroplast envelope is indicated by the finding that the MTX stabilized DHFR moiety of chimeric precursor proteins, that was found to be tightly folded in solution, became highly sensitive to protease treatment upon binding to the envelope (Guera et al., 1993; Endo et al., 1994; America et al., 1994). This conformational change leads to a release of bound MTX (America et al., 1994). An increased susceptibility to proteolytic digestion was also observed for ricin A chain in a chimeric precursor protein after binding to chloroplasts (Walker et al., 1996). Taken together these data indicate that conformational changes of precursor proteins take place after binding of the precursor protein to the chloroplast envelope but before translocation across the envelope membranes. The identity of the “unfoldase” remains to be established. It is noteworthy in this respect, that for other translocation systems molecular chaperones, especially of the HSP70 family, have been implicated in ensuring an import competent conformation of precursor proteins (Deshaies et al., 1988; Hendrick and Hartl, 1993; Becker and Craig, 1994 for review). It is interesting to speculate that the ATP hydrolysis required for stable binding of precursor proteins to chloroplasts may well be used to drive the unfolding process, possibly by an HSP70 homologue of the chloroplast envelope (see below). Another possibility is that unfolding is driven from the inside due to the pulling force generated by stromal molecular chaperones that are involved in the ATP dependent translocation across the membrane.
3. MOLECULAR CHAPERONES IN THE TRANSLOCATION MACHINERY OF THE OUTER ENVELOPE OF CHLOROPLASTS 3.1. Components of the Translocation Machinery The translocation of chloroplast precursor proteins across the double membrane can experimentally be divided into distinct steps on the basis of differences in the energy requirements which possibly reflect a sequential action of molecular chaperones. The interaction of precursor proteins with chloroplasts includes specific interactions with lipids and a proteinaceous receptor (de Boer and Weisbeek., 1991; Keegstra et al., 1995 for review). In the complete absence of ATP, the interaction of precursor proteins is reversible, but becomes irreversible, i.e. unidirectional, in the presence of M ATP (Perry and Keegstra, 1994). This ATP requiring step was localized to the outer chloroplast envelope and/or the intermembrane space (Flügge and Hinz, 1986; Schindler et al., 1987). Complete translocation of the precursor protein across both envelope membranes is achieved only at higher (mM) ATP concentrations as they exist in the
Protein import into and folding within chloroplasts
295
chloroplast stroma (Theg et al., 1989). After years of controversy, recent findings by independent groups have provided evidence for the involvement of at least four outer envelope membrane proteins (OEPs) with molecular masses of 86 kDa, 75 kDa, 70 kDa and 34 kDa and inner envelope membrane proteins (IEPs) with molecular masses of 100 kDa, 55 kDa and 44 kDa, in the translocation of precursor proteins into chloroplasts (Gray and Row, 1995; Schnell, 1995; Soll, 1995; Heins et al., 1998; Heins and Soll, 1998). Analysis of the primary structures together with biochemical data yielded information on the functions of the identified proteins and the mechanism of import of precursors into chloroplasts. OEP86 exposes a large N-terminal domain on the cytoplasmic surface of chloroplasts and is very sensitive to proteolytic digestion (Hirsch et al., 1994). Antibodies raised against OEP86 are able to inhibit binding and translocation of the precursor of the small subunit of Ribulose-1, 5bisphosphate Carboxylase/ Oxygenase (Hirsch et al., 1994). Crosslink studies reveal that OEP86 interacts with precursor proteins at an early stage of the translocation process (Perry and Keegstra, 1994). Thus, OEP86 most likely represents an import receptor. OEP75 interacts at a later stage with precursor proteins (Perry and Keegstra, 1994) and is resistant to proteolytic digestion, presumably because it is deeply imbedded in the outer envelope membrane, probably spanning it with several -sheets (Schnell et al., 1994). Using overexpressed OEP75, a pore forming activity can be measured in vitro (Hinnah et al., 1997). Thus, it is most likely that OEP75 forms the translocation pore. The GTPbinding protein OEP34 is in close proximity to OEP75 (Seedorf et al., 1995) and therefore a regulatory function is proposed for OEP34, e.g. by controlling the gating status of the translocation pore. IEP110 was suggested to be involved in the formation of translocation contact sites as a large portion of the protein protrudes into the intermembrane space (Lübeck et al., 1996). The function of IEP44 is unknown at present. Properties and proposed functions of the OEP’s and IEP’s are summarized in Table 1. The finding that two of the identified proteins, OEP86 and OEP34, represent new types of GTP binding proteins (Kessler et al., 1994; Seedorf et al., 1995) and are moreover highly phosphorylated by an envelope protein kinase (Soll, 1985) suggests that the translocation process into chloroplasts may be highly regulated and, unexpectedly, different from that reported for mitochondria. 3.2. Characteristics of HSP70 Homologues Present in the Outer Envelope of Chloroplasts OEP70 was immunologically identified as a HSP70 homologue (Waegemann and Soll, 1991; Schnell et al., 1994). It is most likely identical to the HSP70 homologue identified by Marshall et al. (1990) in pea chloroplasts. This HSP70 homologue is unusual in that it behaves like an integral membrane protein. It is resistant to treatment of chloroplasts with thermolysin, a protease that does not penetrate the outer envelope of chloroplasts (Cline et al., 1984), but it is digested by a treatment with trypsin and chymotrypsin (Marshall et al., 1990). This combination of proteases gains access to the intermembrane space (Marshall et al., 1990). Thus, this HSP70 homologue is most likely localized on the intermembrane face of the outer
Molecular chaperones and folding catalysts
296
Table 1 Proteins of the Chloroplast Outer and Inner Envelope Membranes which are most likely Involved in Protein Translocation (also see Schnell et al., 1997)
Component Location
Proposed function
Literature
OEP86
outer membrane (integral) Receptor
Waegemann and Soll, 1991; Hirsch et al., 1994; Perry and Keegstra, 1994; Schnell et al., 1994
OEP75
outer membrane (integral) Protein translocation channel
Waegemann and Soll, 1993; Schnell et al., 1994; Perry and Keegstra, 1994
OEP34
outer membrane (integral) Regulation
Waegemann and Soll, 1993; Kessler et al., 1994; Seedorf et al., 1995
OEP70
outer membrane (integral) Chaperone
Waegemann and Soll, 1991; Schnell et al., 1994
COM70
outer membrane (peripheral)
Chaperone
Wu et al., 1994
IEP100
inner membrane
unknown
Schnell et al., 1994; Wu et al., 1994; Lübeck et al., 1996
JEP55
inner membrane
Regulation
Caliebe et al., 1997
IEP44
inner membrane
unknown
Wu et al., 1994
envelope. A second HSP70 homologue, COM70, which might also be involved in the translocation of precursor proteins was identified by Wu et al. (1994). COM70 is localized on the cytoplasmic surface of chloroplasts as judged by its accessibility to thermolysin. Analysis of the cDNA clone for this HSP70 homologue (Ko et al., 1992) shows greater homology to eukaryotic cytoplasmic HSP70 than to DnaK, the E. coli HSP70 homologue. It is thus possible that chloroplastic precursor proteins interact with HSP70 on either side of the outer envelope of chloroplasts. COM70 may bind precursor proteins on the cytosolic face of the outer envelope, maintaining or achieving an unfolded, import competent conformation. OEP70 may bind to the precursor protein as it emerges from the protein conducting channel of the outer envelope membrane. This could prevent a refolding of the precursor protein before it engages the translocation machinery of the inner envelope membrane. Interaction with one of the HSP70 homologues may ensure
Protein import into and folding within chloroplasts
297
that binding of the precursor protein becomes irreversible and thus contribute to unidirectionality of the translocation process. Interaction of precursor proteins with OEP70 may also provide some of the thermodynamic driving force for the translocation in a brownian ratchet mechanism (Simon et al., 1992). In that case OEP70 would have a function similar to that proposed for mitochondrial HSP70, which acts on the trans side of the membrane in protein import into mitochondria (Hendrick and Hartl, 1993; Stuart et al., 1994; Hartl, 1996).
4. STROMAL CHAPERONES Newly imported proteins either stay in the stroma or are routed to their final destination by distinct pathways (see below). Proteins residing in the stroma undergo folding to monomers and, if required, are assembled into oligomers. In those cases where further sorting takes place, reactions preventing folding must occur (see below). Accordingly, members of all classes of molecular chaperones have been identified in the chloroplast stroma. 4.1. Stromal HSP70 Homologues Two HSP70 homologues with molecular masses of 75 kDa and 78 kDa that are localized in the stroma were identified in pea chloroplasts (Marshall et al., 1990). Using ATP agarose affinity chromatography a major soluble HSP70 from spinach chloroplasts was purified by Wang et al. (1993). Analysis of cDNA sequences obtained for the above proteins (Marshall and Keegstra, 1992; Wang et al., 1993) established that the major stromal HSP70 homologue is nuclear encoded and is more similar to DnaK from prokaryotes than to cytoplasmic HSP70 from eukaryotes. Interestingly, a sequence comparison of HSP70 homologues indicated that HSP70 fall into four groups, reflecting the cellular compartment in which the proteins are localized (Wang et al., 1993). The group C proteins are localized in the chloroplast stroma and, in addition, include HSP70 from Synechocystis. This is consistent with the assumption that chloroplasts have been derived from a procaryotic ancestor. Two additional heat shock proteins, DnaJ and GrpE, serve as regulatory factors that modulate the chaperone activity of DnaK in E.coli (Cyr et al., 1994 for review). Stromal DnaJ and GrpE homologues were recently detected (Schlicher and Soll, 1997) suggesting that the mechanism of HSP70 action is similar in the chloroplast stroma. There has been only limited investigation concerning the functions of HSP70 homologues in the chloroplast stroma. HSP70 homologues localized within the endoplasmic reticulum and mitochondria appear to assist the translocation of precursor proteins across these membranes (Hendrick and Hartl, 1993; Becker and Craig, 1994; Stuart et al., 1994) and are required for the completion of the translocation process. As the release of bound protein from HSP70 requires ATP, it is temptative to speculate that an analogous interaction may account for the requirement of stromal ATP in the translocation of proteins into chloroplasts. However, there is only limited evidence that stromal HSP70 interacts with newly imported proteins. An initial transient interaction
Molecular chaperones and folding catalysts
298
between ferredoxin-NADP+reductase (FNR) and stromal HSP70 was observed by Tsugeki and Nishimura (1993). Association of FNR with the HSP70 homologue preceded that with the stromal Cpn60 homologue. This is similar to examples of imported proteins in mitochondria, which first associate with HSP70 prior to a transfer to Cpn60 (Manning-Krieg et al, 1991). In contrast, the Rieske FeS protein was shown to interact with Cpn60 before forming a complex with HSP70 (Madueno et al, 1993). Thus, different precursor proteins seem to rely to different extent on molecular chaperones for functional folding and assembly. Recently, it was proposed that Hsp100 might also be involved in importing proteins into the stroma (Akita et al, 1997; Nielsen et al, 1997). 4.2. Chloroplast Chaperonins Chaperonins are a group of related proteins that are found in all bacteria, mitochondria and plastids of eucaryotic cells and that play a major role in protein folding and assembly. There are two types of sequence related chaperonins, that are referred to as chaperonin 60 (Cpn60), and chaperonin 10 (Cpn10) reflecting the molecular masses of their subunits. The stromal CpnGO was originally identified as an important component in the biogenesis of Ribulose-1, 5-bisphosphate Carboxylase/Oxygenase (Rubisco) (for review see: Ellis and van der Vies, 1988; Roy, 1989; Gutteridge and Gatenby, 1995). Also, the first report of a newly imported protein becoming associated with stromal Cpn60 was for a chimeric protein comprising the large subunit of Rubisco (Gatenby et al, 1988). Imported small subunits also associated with Cpn60 although to a lower extent (Gatenby et al, 1988). In addition to Rubisco subunits a variety of imported proteins becomes associated with stromal Cpn60 (Lubben et al, 1989; Madueno et al, 1993; Tsugeki and Nishimura, 1993), indicating a more general role of this chaperonin in folding and assembly of proteins in chloroplasts. For seven of nine proteins analysed by Lubben et al (1989), including subunits of oligomeric structures as well as monomeric proteins, stable interactions with Cpn60 that were disrupted by illumination or addition of ATP were observed. Only ferredoxin and superoxid dismutase did not form a complex with stromal Cpn60. In contrast to bacteria and mitochondria, both of which contain only a single type of Cpn60, two distinct Cpn60 polypeptides of 61 kDa and 60 kDa, termed and , respectively, are present in the chloroplast stroma (Hemmingsen and Ellis, 1986, Musgrove et al, 1987). The two stromal Cpn60 isoforms are highly divergent in their predicted amino acid sequence (Martel et al, 1990). They are only about 50 % identical to each other, no more than to GroEL, their bacterial counterpart (Viitanen et al, 1995). Both isoforms are present in roughly equal amounts (Hemmingsen and Ellis, 1986; Musgrove et al, 1987), but it is not known at present whether and subunits form homo- or heterooligomeric structures. Both the and subunit are constitutively expressed and nuclear encoded (Hemmingsen and Ellis, 1986; Musgrove et al, 1987). A co-chaperonin has also been identified in the chloroplast stroma, where it presumably functions like other co-chaperonins to effectively discharge target proteins bound to Cpn60. The stromal GroES homologue has an apparent size of 24 kDa and is therefore twice the size of bacterial and mitochondrial Cpn10 molecules. Sequence analysis of the isolated cDNA of the spinach stromal Cpn10 revealed that the stromal
Protein import into and folding within chloroplasts
299
Cpn10 is comprised of two distinct Cpn10 domains fused head to tail to give a double domain co-chaperonin structure (Bertsch et al., 1992). Like the stromal Cpn60, the stromal Cpn10 is constitutively expressed and synthesized as a nuclear encoded precursor protein with an N-terminal transit peptide (Bertsch et al., 1992). Identical amino acids that are highly conserved in all known Cpn10 molecules to date are present in both halves of the double domain stromal Cpn10 and have been shown to be important for the cochaperonin function (Bertsch and Soll, 1995). At present it is unclear whether the special composition of the stromal Cpn60/ Cpn10 system has functional implications. It has been demonstrated that stromal proteins can be folded by heterologous, non chloroplastic chaperonin systems (Bertsch and Soll, 1995; Baneyx et al., 1995). Furthermore, the unique binary structure of the stromal Cpn10 is not obligatory for the folding reaction as the stromal Cpn60 functions equally well with bacterial, mitochondrial or stromal Cpn10 isoforms (Viitanen et al., 1995). However, while both halves of the stromal Cpn10 are functionally active in an in vivo system, they are not active in vitro (Bertsch and Soll, 1995). The double domain stromal Cpn10 is present in a broad range of distantly related photosynthetic eucaryotes, including liverworts, mosses, club-mosses, gymnosperms, monocots and dicots (Baneyx et al., 1995) whereas photosynthetic cyanobacteria, the postulated progenitors of chloroplasts, possess the normal single domain type of Cpn10 (Lehel et al., 1993). Thus, it appears that the duplication of domains, either through gene fusion or gene duplication, occurred early after the endosymbiontic event that gave rise to chloroplasts and the fusion has been highly conserved, suggesting that retention of the double domain structure is somehow advantageous in chloroplasts. It remains to be determined if the and subunit of the stromal Cpn60 require different domain interactions for maximum activity. Despite their unusual structural features, the complexes that are formed by stromal Cpn60 and Cpn10 are of similar size and shape as GroEL/GroES complexes (Viitanen et al., 1995) and the available evidence suggests that folding reactions mediated by GroEL or the stromal Cpn60 are mechanistically similar (Boston et al., 1996).
5. ROLE OF STROMAL FACTORS IN THE TRANSLOCATION OF PROTEINS INTO OR ACROSS THE THYLAKOID MEMBRANE The light reaction of oxygenic photosynthesis involves five multi-subunit complexes located to the thylakoid membrane. About 50% of these proteins are nuclear encoded (Cohen et al., 1995 for review). In principle, targeting of nuclear encoded thylakoidal proteins occurs in a two step pathway. Proteins are synthesized in the cytosol as precursors either with a normal stroma directing transit peptide with residual thylakoid targeting information located in the mature protein, e.g. pLHCP (Lamppa, 1988) or the Rieske FeS protein (Madueno et al., 1992), or with a bipartite presequence, which contains an “envelope transit” signal and a “thylakoid transfer” domain in tandem, e.g. the 33 kDa subunit of the oxygen evolving complex (Ko and Cashmore, 1989) or plastocyanin (Hagemann et al., 1990). Proteins are imported into the organelle by the general envelope translocation machinery outlined above and targeted to their final
Molecular chaperones and folding catalysts
300
location including the thylakoid membrane by distinct pathways (for reviews: De Boer and Weisbeek, 1991; Theg and Scott, 1993; Cohen et al., 1995). In this section only those import pathways will be discussed that depend on the presence of stromal factors. It is generally assumed that proteins are transferred through the soluble stromal phase rather than by membrane flow from the inner envelope. However, recently a stromal factor, plastid fusion and/ or translocation factor (Pftf), that is involved in vesicle fusion was identified by Hugueney et al. (1995). Sequence analysis of the isolated cDNA clone revealed that mature Pftf has significant homology to the yeast and animal NSF protein that is involved in vesicle fusion (Hugueney et al., 1995). It is thus tempting to speculate that this protein is part of a novel import pathway into or across the thylakoid membrane that involves the membrane flow from the inner envelope. 5.1. Targeting to the Thylakoid Membrane Most of the work concerning the targeting requirements for integral proteins of the thylakoid membrane has been done with the precursors of apoproteins of the light harvesting chlorophyll a/b-binding proteins (LHCP). It is well established that LHCP is routed to the thylakoids through the soluble stromal phase. As a result, solubility and integration competence must be maintained in the stroma as well as in the cytosol. Indeed, early studies have shown that LHCP traverses the stroma in a soluble intermediate form (Cline et al., 1989; Reed et al., 1990). Translocation into the thylakoid membrane is completely dependent on the presence of ATP, the pH across the thylakoid membrane and a stromal factor (Chitnis et al., 1987; Fulson and Cline, 1988) that functions in conversion of LHCP to a soluble species of about 120 kDa (Payan and Cline, 1991).The stromal factor has a molecular mass of about 65 kDa and does not contain RNA (Fulson and Cline, 1988). The nature of this stromal factor has been subject of discussion. Yalowsky et al. (1992) reported that chloroplast HSP70 could substitute for stromal extract in the translocation of LHCP. However, neither stromal HSP70 nor Cpn60 were found to be a component of the soluble complex and immunodepletion of the stromal extract of HSP70 or Cpn60 did not result in a loss of activity (Yuan et al., 1993). A completely new possibility for the nature of the stromal factor has emerged from studies of Franklin and Hoffmann (1993). These workers have cloned a chloroplast homologue of the 54 kDa subunit of the mammalian signal recognition particle (54CP), that was subsequently shown to be present in the soluble stromal complex of LHCP (Li et al., 1995) consistent with the finding that GTP is more effective than ATP in the translocation of LHCP into the thylakoids (Hoffmann and Franklin, 1994). This pathway may be inherited from the ancestral cyanobacteria as a gene coding for a 54 kDa signal recognition particle homologue was recently identified in the cyanobacterial genome (Packer and Howe, 1996). However, the isolated complex still required stroma to complete the integration process (Fulson and Cline, 1988) suggesting the need for at least one more stroma-mediated reaction that has yet to be elucidated. 5.2. Targeting to the Thylakoid Lumen Information available on translocation requirements for an increasing number of proteins
Protein import into and folding within chloroplasts
301
located on the lumenal side of the thylakoids has revealed the existence of at least two distinct pathways with different requirements for the translocation (for review: Robinson et al., 1993; Robinson and Klösgen, 1994). Proteins of one group, including the 16 kDa and 23 kDa subunit of the oxygen evolving complex and subunit N of photosystem I, are transported by a mechanism which is absolutely dependent on the thylakoidal pH, but does not require ATP or soluble stromal factors. On contrary, translocation of the second group, e.g. the 33 kDa subunit of the oxygen evolving complex, plastocyanin, and subunit F of photosystem I, depends on the presence of protein factors and ATP, but pH is not a prerequisite. Evidence is emerging that translocation of proteins of the second group precedes via a Sec-dependent pathway. There are many similarities between the Secdependent pathway of protein translocation into the thylakoids and export of proteins in bacteria. Thylakoid transfer domains of bipartite presequences share common features with bacterial signal sequences (von Heijne et al., 1989) and have been shown to function in translocation of proteins across the bacterial plasma membrane (Seidler and Michel, 1990). Both mechanisms require the participation of soluble factors and ATP and are stimulated by ∆pH. Bacterial Sec-dependent export involves the participation of SecB, a cytosolic chaperone molecule with an unfolding function, and SecA, a translocation ATPase (reviewed in Wickner et al., 1991). Consistent with the assumption of a Secdependent translocation pathway in chloroplast, it has been shown that azide, a potent inhibitor of SecA, inhibits the transport of proteins using this pathway (Knott and Robinson, 1994; Henry et al., 1994) whereas translocation of proteins of the Secindependent pathway is unaffected by this agent. Recently SecA homologues were immunologically identified in pea chloroplasts (Yuan et al., 1994; Nakai et al., 1994a). The chloroplast SecA homologue has a molecular mass of 110 kDa and appears, like E.coli SecA, to function as a homodimer (Yuan et al., 1994; Nakai et al., 1994a). SecA and SecY genes have been identified in the genome of cyanobacteria (Nakai et al., 1992; Nakai et al., 1994b) and the plastid genome of cyanophytes (Scaramuzzi et al., 1992; Flachmann et al., 1993). cDNA clones were isolated for the stromal SecA homologue from pea (Nohara et al., 1995) and spinach (Berghöfer et al., 1995) and for a chloroplastic SecY homologue from Arabidopsis thaliana (Laidler et al., 1995). SecA and SecY homologues are nuclear encoded in higher plants and are more similar to genes from cyanobacteria than to E.coli SecA. This strongly indicates that higher plant chloroplasts most probably inherited the SecA-dependent translocation pathway from the ancestral endosymbiont. This notion is strengthened by the fact that proteins translocated by the SecA-dependent pathway are also found in the photosynthetic apparatus of cyanobacteria, whereas proteins translocated by the pH dependent pathway are so far only found in chloroplasts of higher plants. Thus, most likely, new components of the photosynthetic apparatus of higher plants and a translocation machinery for these proteins were acquired after the endosymbiontic event. Interestingly, a Sec-dependent pathway has at present not been found in mitochondria that are also of endosymbiontic origin. There are only few studies on the translocation and sorting of proteins encoded by the plastid genome. It will be interesting to see whether these proteins are also translocated by a Sec-dependent pathway. This has been suggested for cytochrome f by a genetic approach (Voelker and Barkan, 1995). However, studies of van Wijk et al. (1995) indicate that the SecA related mechanism is not widely used for targeting of chloroplast
Molecular chaperones and folding catalysts
302
encoded proteins.
6. MOLECULAR CHAPERONES IN THE THYLAKOID LUMEN Molecular chaperones have very recently been found in the thylakoid lumen (Schlicher and Soll, 1996; Fulgosi et al., 1998) consistent with the necessity of folding and assembly of proteins, e.g. subunits of the oxygen evolving complex, in this compartment. Homologues of HSP70, Cpn60 and Cpn10 were identified immunologically. Only one Cpn60 homologue was detected in the thylakoid lumen which is of similar size as the stromal isoform but clearly separated by two dimensional electrophoresis. The lumenal Cpn10 has a molecular weight of 10–12 kDa and thus seems to be distinct from the double domain stromal Cpn10 (see above). Thus, composition of the lumenal chaperonin system resembles that of the “normal” system known from other cellular compartments and is unlike the unique stromal Cpn60/Cpn10 system. The lumenal chaperonins are most likely coded for by nuclear genes and thus it will be interesting to determine the translocation mechanism that is involved in targeting of these proteins and in preventing premature interception in the stroma with their stromal relatives. In addition, two lumenal HSP70 homologues were detected that are distinct from the two stromal HSP70 homologues described by Marshall et al.. (1990) (see above). This system is complemented by a peptidyl-prolyl cis-trans isomerase (Fulgosi et al., 1998). Taken together these data indicate that chloroplasts contain the largest number of molecular chaperone isoforms present in one organelle.
7. REFERENCES Akita, M., Nielsen, E. and Keegstra, K. (1997). Identification of protein transport complexes in the chloroplastic envelope membranes via chemical cross-linking. J. Cell Biol , 136 , 983–994. America, T., Hageman, J., Guera, A., Rook, F., Archer, K., Keegstra, K. and Weisbeek, P. (1994). Methotrexate does not block import of a DHFR fusion protein into chloroplasts. Plant Mol Biol , 24 , 283–294. Baneyx, F., Bertsch, U., Kalbach, C.E., van der Vies, S.M., Soll., J. and Gatenby, A.A. (1995). Spinach chloroplast cpn21 co-chaperonin possesses two functional domains fused together in a toroidal structure and exhibits nucleotide-dependent binding to plastid chaperonin 60. J. Biol. Chem. , 270 , 10695–10702. Backer, J. and Craig, E.A. (1994). Heat shock proteins as molecular chaperones. Eur. J. Biochem. , 219 , 11–23. Berghöfer, J., Karnauchov, I., Herrmann, R.G. and Klösgen, R.B. (1995). Isolation and characterization of a cDNA encoding the SecA protein from spinach chloroplasts. J. Biol. Chem. , 270 , 18341–18346. Bertsch, U., Soll, J., Seetharam, R. and Viitanen, P.V. (1992). Identification, characterization and DNA sequence of a functional “double” groES-like chaperonin from chloroplasts of higher plants. Proc. Natl. Acad. Set. USA , 89 , 8696–8700. Bertsch, U. and Soll, J. (1995). Functional analysis of isolated Cpn10 domains and conserved amino acid residues in spinach chloroplast co-chaperonin by site-directed
Protein import into and folding within chloroplasts
303
mutagenesis. Plant Mol. Biol. , 29 , 1039–1055. Boston, R.S., Viitanen, P.V. and Vierling, E. (1996). Molecular chaperones and protein folding in plants. Plant Mol. Biol. , in press. Caliebe, A., Grimm, R., Kaiser, G., Lübeck, J., Soll, J. and Heins, L. (1997). The chloroplastic protein import machinery contains a Rieske-type iron-sulfur cluster and a mononuclear iron-binding protein. EMBO J. , 16 , 7342–7350. Chitnis, P.R., Nechushtai, R. and Thornber, J.P. (1987). Insertion of the precursor of the light-harvesting chlorophyll a/b-protein into the thylakoids requires the presence of a developmentally regulated stromal factor. Plant Mol. Biol. , 10 , 3–11. Cline, K., Werner-Washburne, M., Andrews, J. and Keegstra, K. (1984). Thermolysin is a suitable protease for probing the surface of intact pea chloroplasts. Plant Physiol. , 75 , 675–678. Cline, K., Fulsom, D.R. and Viitanen, P.V. (1989). An imported thylakoid protein accumulates in the stroma when insertion into thylakoids is inhibited. J. Biol. Chem. , 264 , 14225–14232. Cohen, Y., Yalowsky, S. and Nechushtai, R. (1995). Integration and assembly of photosynthetic protein complexes in chloroplast thylakoid membranes. Biochim. Biophys. Acta , 1241 , 1–30. Creighton, A.M., Hulford, A., Mant, A., Robinson, D. and Robinson, C. (1995). A monomeric, tightly folded stromal intermediate on the DpH-dependent thylakoidal protein transport pathway. J. Biol. Chem. , 270 , 1663–1669. Cyr, D.M., Langer, T. and Douglas, M.G. (1994). DnaJ-like proteins: molecular chaperones and specific regulators of Hsp70. Trends Biochem. Sci. , 19 , 176–181. De Boer, A.D. and Weisbeek, P.J. (1991). Chloroplast protein topogenesis: import, sorting and assembly. Biochim. Biophys. Acta , 1071 , 221–253. Della-Cioppa, G., Bauer, S.C., Klein, B.K., Shah, D.M., Fraley, R.T. and Kishore, G.M. (1986). Translocation of the precursor of 5-enolpyruvylshikimate-3-phosphate synthase into chloroplasts of higher plants in vitro . Proc. Natl. Acad. Sci. USA , 83 , 6873–6877. Della-Cioppa, G. and Kishore, G. (1988). Import of a precursor protein into chloroplasts is inhibited by the herbicide glyphosate. EMBO J. , 7 , 1299–1305. Deshaies, R.J., Koch, B.D. and Shekman, R. (1988). The role of stress proteins in membrane biogenesis. Trends Biochem. Sci. , 13 , 384–388. Eilers, M. and Schatz, G. (1988). Protein unfolding and the energetics of protein translocation across biological membranes. Cell , 52 , 481–483. Ellis, R.J. and van der Vies, S.M. (1988). The rubisco subunit binding protein. Photosyn. Res. , 16 , 101–115. Endo, T., Kawakami, M., Goto, A., America, T., Weisbeek, P. and Nakai, M. (1994). Chloroplast protein import. Chloroplast envelopes and thylakoids have different abilities to unfold proteins. Eur. J. Biochem. , 225 , 403–409. Flachmann, R., Michalowski, C.B., Loffelhardt, W. and Bohnert, H.J. (1993). SecY, an integral subunit of the bacterial preprotein translocase, is encoded by a plastid genome. J. Biol. Chem. , 268 , 7514–7519. Flügge, U.I and Hinz, G. (1986). Energy dependence of protein translocation into chloroplasts. Eur. J. Biochem. , 160 , 563–570. Franklin, A.E. and Hoffman, N.E. (1993). Characterization of a chloroplast homologue of the 54-kDa subunit of the signal recognition particle. J. Biol. Chem. , 268 , 22175– 22180. Fulgosi, H., Vener, A.V., Altschmidt, L., Herrmann, R.G. and Andersson, B. (1998). A
Molecular chaperones and folding catalysts
304
novel multifunctional chloroplast protein: identification of a 40 kDa immunophilin-like protein located in the thylakoid lumen. EMBO J. , 17 , 1577–1587. Fulson, D.R. and Cline, K. (1988). A soluble protein factor is required in vitro for membrane insertion of the thylakoid precursor protein, pLHCP. Plant Physiol. , 88 , 1146–1153. Gatenby, A.A., Lubben, T.H., Ahlquist, P. and Keegstra, K. (1988). Imported large subunits of ribulose bisphosphate carboxylase/oxygenase, but not imported b-ATP synthase subunits, are assembled into holoenzyme in isolated chloroplasts. EMBO J. , 7 , 1307–1314. Gatenby, A.A. and Viitanen, P.V. (1994). Structural and functional aspects of chaperonin-mediated protein folding. Annu. Rev. Plant Physiol. Plant Mol Biol , 45 , 469–491. Gatenby, A.A. (1996). Chaperonins of photosynthetic organisms. In: R.J.Ellis, (ed.), The chaperonins , Academic Press, London, pp 65–90. Gray, J.C. and Row, P.E. (1995). Protein translocation across chloroplast envelope membranes. Trends Cell Biol. , 5 , 243–247. Guera, A., America, T., van Waas, M. and Weisbeek, P.J. (1993). A strong protein unfolding activity is associated with the binding of precursor chloroplast proteins to chloroplast envelopes. Plant Mol. Biol. , 23 , 309–324. Gutteridge, S. and Gatenby, A.A. (1995). Rubisco synthesis, assembly, mechanism, and regulation. Plant Cell , 7 , 809–819. Hagemann, J., Baecke, C., Ebskamp, M., Pilon, R., Smeekens, S. and Weisbeek, P. (1990). Protein import into and sorting inside the chloroplast are independent processes. Plant Cell , 2 , 479–494. Hartl, F.U. (1996). Molecular chaperones in cellular protein folding. Nature , 381 , 571– 580. Heins, L., Collinson, I. and Soll, J. (1998). The protein translocation apparatus of chloroplast envelopes. Trends in Plant Science , 3 , 56–61. Heins, L. and Soll, J. (1998). Mixing the prokaryotic and the eukaryotic? Current Biology , 8 , 215–217. Hemmingsen, S.M and Ellis, R.J. (1986). Purification and properties of Ribulosebisphophate Carboxylase large subunit binding protein. Plant Physiol. , 80 , 269–276. Hendrick, J.P. and Hartl, F.-U. (1993). Molecular chaperone functions of heat-schock proteins. Annu. Rev. Biochem. , 62 , 349–384. Henry, R., Kapazoglu, A., McCaffery, M. and Cline, K. (1994). Differences between lumen targeting domains of chloroplast transit peptides determine pathway specificity for thylakoid transport. J. Biol. Chem. , 269 , 10189–10192. Hinnah, S.C., Hill, K., Wagner, R., Schlicher, T. and Soll, J. (1997). Reconstitution of a chloroplast protein import channel. EMBO J. , 16 , 7351–7360. Hirsch, S., Muckel., E., Heemeyer, F., von Heijne, G. and Soll, J. (1994). A receptor component of the chloroplast protein translocation machinery . Science , 266 , 1989– 1992. Hoffman, N.E. and Franklin, A.E. (1994). Evidence for a stromal GTP requirement for the integration of a chlorophyll a/b-binding polypeptide into thylakoid membranes. Plant Physiol. , 105 , 295–304. Hugueney, P., Bouvier, F., Badillo, A., D’Harlingue, A., Kuntz, M. and Camara, B. (1995). Identification of a plastid protein involved in vesicle fusion and/or membrane protein translocation. Proc. Natl. Acad. Sci. USA , 92 , 5630–5634.
Protein import into and folding within chloroplasts
305
Keegstra, K., Bruce, B., Hurley, M., Li, H.-M. and Perry, S. (1995). Targeting of proteins into chloroplasts. Physiol. Plant. , 93 , 157–162. Kessler, F., Blobel., G., Patel, H.A. and Schnell, D.J. (1994). Identification of two GTPbinding proteins in the chloroplast protein import machinery. Science , 266 , 1035– 1039. Knott, T.G. and Robinson, C. (1994). The SecA inhibitor, azide, reversibly blocks the translocation of a subset of proteins across the chloroplast thylakoid membrane. J. Biol. Chem. , 269 , 7843–7846. Ko, K. and Cashmore, A.R. (1989). Targeting of proteins to the thylakoid lumen by the bipartite transit peptide of the 33 kd oxygen-evolving protein. EMBO J. , 8 , 3187– 3194. Ko, K., Bornemisza, O., Kourtz, L., Ko, Z.W., Plaxton, W.C. and Cashmore, A.R. (1992). Isolation and characterization of a cDNA clone encoding a cognate 70-kDa heat shock protein of the chloroplast envelope. J. Biol. Chem. , 267 , 2986–2993. Laidler, V., Chaddock, A.M., Knott, T.G., Walker, D. and Robinson, C. (1995). A SecY homolog in Arabidopsis thaliana. Sequence of a full-length cDNA clone and import of the precursor protein into chloroplasts. J. Biol Chem. , 270 , 17664–17667. Lamppa, G.K. (1988). The chlorophyll a/b-binding protein inserts into the thylakoids independent of its cognate transit peptide. J. Biol. Chem. , 263 , 14996–14999. Lehel, C., Los, D., Wada, H., Gyêrgyei, J., Horvath, I., Kovacs, E., Murata, N. and Vigh, L. (1993). A second groEL-like gene, organized in a groESL operon is present in the genome of Synechocystis sp. PCC 6803. J. Biol. Chem. , 268 , 1799–1804. Li, X., Henry, R., Yuan, J., Cline, K. and Hoffman, N.E. (1995). A chloroplast homologue of the signal recognition particle subunit SRP54 is involved in the posttranslational integration of a protein into thylakoid membranes. Proc. Natl. Acad. Sci. USA , 92 , 3789–3793. Lubben, T.H., Donaldson, G.K., Viitanen, P. and Gatenby, A.A. (1989). Several proteins imported into chloroplasts form stable complexes with the GroEL related chloroplast molecular chaperone. Plant Cell , 1 , 1223–1230. Lübeck, J., Soll, J., Akita, M., Nielsen, E. and Keegstra, K. (1996). Topology of IEP110, a component of the chloroplastic protein import machinery present in the inner envelope membrane. EMBO J. , in press. Madueno, F., Napier, J.A., Cejudo, F.J. and Gray, J.C. (1992). Import and processing of the precursor of the Rieske FeS protein of tobacco chloroplasts. Plant Mol. Biol , 20 , 289–299. Madueno, F., Napier, J.A. and Gray, J.C. (1993). Newly imported Rieske Iron-Sulfur protein associates with both Cpn60 and Hsp70 in the chloroplast stroma. Plant Cell , 5 , 1865–1876. Manning-Krieg, U.C., Scherer, P.E. and Schatz, G. (1991). Sequential action of mitochondrial chaperones in protein import into the matrix. EMBO J. , 10 , 3273–3280. Marshall, J.S., DeRocher, A.E., Keegstra, K. and Vierling, E. (1990). Identification of heat shock hsp70 homologues in chloroplasts. Proc. Natl. Acad. Sci. USA , 87 , 374– 378. Marshall, J.S. and Keegstra, K. (1992). Isolation and characterization of a cDNA clone encoding the major Hsp70 of the pea chloroplastic stroma. Plant Physiol , 100 , 1048– 1054. Martel, R., Cloney, L.P., Pelcher, L.E. and Hemmingsen, S.M. (1990). Unique composition of plastid chaperonin-60: and polypeptide encoding genes are highly divergent. Gene , 94 , 181–187.
Molecular chaperones and folding catalysts
306
Meyer, D.I. (1988). Preprotein conformation: the year’s major theme in translocation studies. Trends Biochem. Sci. , 13 , 471–474. Mihara, K. and Omura, T. (1996). Cytoplasmic chaperones in precursor targeting to mitochondria: the role of MSF and HSP70. Trends Cell Biol. , 6 , 104–108. Musgrove, J.E. Johnson, R.A. and Ellis, R.J. (1987). Dissociation of the ribulosebisphosphate-carboxylase large-subunit binding protein into dissimilar subunits. Eur. J. Biochem. , 163 , 529–534. Nakai, M., Tanaka, A., Omata, T. and Endo, T. (1992). Cloning and characterization of the secY gene from the cyanobacterium Synechococcus PCC7942. Biochim. Biophys. Acta , 1171 , 113–116. Nakai, M., Goto, A., Nohara, T., Sugita, D. and Endo, T. (1994a). Identification of the SecA protein homolog in pea chloroplasts and its possible involvement in thylakoidal protein transport. J. Biol. Chem. , 269 , 31338–31341. Nakai, M., Nohara, T., Sugita, D. and Endo, T. (1994b). Identification and characterization of the secA protein homologue in the cyanobacterium Synechococcus PCC7942. Biochem. Biophys. Res. Commun. , 200 , 844–851. Nielsen, E., Akita, M., Davila-Aponte, J. and Keegstra, K. (1997). Stable association of chloroplastic precursors with protein translocation complexes that contain proteins from both envelope membranes and a stromal Hsp110 molecular chaperone. EMBO J. , 16 , 935–994. Nohara, T., Nakai, M., Goto, A. and Endo, T. (1995). Isolation and characterization of the cDNA for pea chloroplast SecA. Evolutionary conservation of the bacterial-typeSecA-dependent protein transport within chloroplasts. FEBS Lett. , 364 , 305–308. Packer, J.C.L. and Howe, C.J. (1996). The cyanobacterial genome contains a single copy of the ffh gene encoding a homologue of the 54 kDa subunit of the signal recognition particle. Plant Mol Biol , 30 , 659–665. Payan, L.A. and Cline, K. (1991). A stromal protein factor maintains the solubility and insertion competence of an imported thylakoid membrane protein. J. Cell Biol , 112 , 603–613. Perry, S.E. and Keegstra, K. (1994). Envelope membrane proteins that interact with chloroplastic precursor proteins. Plant Cell , 6 , 93–105. Pilon, M., Rietveld, A.G., Weisbeek, P.J. and de Kruijff, B. (1992a). Secondary structure and folding of a functional chloroplast precursor protein. J. Biol Chem. , 267 , 19907– 19913. Pilon, M., de Kruijff, B. and Weisbeek, P.J. (1992b). New insights into the import mechanism of the ferredoxin precursor into chloroplasts. J. Biol. Chem. , 267 , 2548– 2556. Randall, L.L. and Hardy, S.J.S. (1986). Correlation of competence for export with lack of tertiary structure of the mature species: A study in vivo of maltose-binding protein in E. coli Cell , 46 , 921–928. Reed, J.E., Cline, K., Stephens, L.C., Bacot, K.O. and Viitanen, P.V. (1990). Early events in the import/ assembly pathway of an integral thylakoid protein. Eur. J. Biochem. , 194 , 33–42. Reinbothe, S., Reinbothe, C., Runge, S. and Apel, K. (1995). Enzymatic product formation impairs both the chloroplast receptor-binding function as well as translocation competence of the NADPH: protochlorophyllide oxidoreductase, a nuclear-encoded plastid precursor protein. J. Cell. Biol. , 129 , 299–308. Robinson, C., Klösgen, R.B., Herrmann, R.G. and Shackleton, J.B. (1993). Protein translocation across the thylakoid membrane-a tale of two mechanisms. FEBS Lett. ,
Protein import into and folding within chloroplasts
307
325 , 67–69. Robinson, C. and Klösgen, R.B. (1994). Targeting of proteins into and across the thylakoid membranea multitude of mechanisms. Plant Mol. Biol , 26 , 15–24. Roy, H. (1989). Rubisco assembly: A model system for studying the mechanism of chaperonin action. Plant Cell , 1 , 1035–1042. Scaramuzzi, C.D., Hiller, R.G. and Stokes, H.W. (1992). Identification of a chloroplastencoded secA gene homologue in a chromophytic alga: possible role in chloroplasts protein translocation. Curr. Gen. , 22 , 421–427. Schindler, C., Hracky, R. and Soll, J. (1987). Protein transport into chloroplasts: ATP is a prerequisite. Z. Naturforsch. , 42c , 103–108. Schlicher, T. and Soll, J. (1996). Molecular chaperones are present in the thylakoid lumen of pea chloroplasts. FEBS Lett. , 379 , 302–304. Schlicher, T. and Soll, J. (1997). Chloroplastic isoforms of DnaJ and GrpE in pea. Plant Mol. Biol , 33 , 181–185. Schnell, D.J., Kessler, F. and Blobel, G. (1994). Isolation of components of the chloroplast protein import machinery. Science , 266 , 1007–1012. Schnell, D.J. (1995). Shedding light on the chloroplast protein import machinery. Cell , 83 , 521–524. Schnell, D.J., Blobel, G., Keegstra, K., Kessler, F., Ko, K. and Soll, J. (1997). A consensus nomenclature for the protein-import components of the chloroplast envelope. Trends in Cell Biol. , 7 , 303–304. Seedorf, M., Waegemann, K. and Soll, J. (1995). A constituent of the chloroplast import complex represents a new type of GTP-binding protein. Plant J. , 7 , 401–411. Seidler, A. and Michel, H. (1990). Expression in Escherichia coli of the PsbO gene encoding the 33 kD protein of the oxygen evolving complex from spinach. EMBO J. , 9 , 1743–1748. Simon, S.M., Peskin, C.S. and Oster, G.F. (1992). What drives the translocation of proteins? Proc. Natl Acad. Sci. USA , 89 , 3770–3774. Soll, J. (1985). Phosphoproteins and protein-kinase activity in isolated envelopes of pea (Pisum sativum L.). chloroplasts. Planta , 166 , 394–400. Soll, J. (1995). New insights into the protein import machinery of the chloroplast’s outer envelope. Bot. Acta , 108 , 277–282. Stuart, R.A., Cyr, D.M., Craig, E.A. and Neupert, W. (1994). Mitochondrial molecular chaperones: Their role in protein translocation. Trends Biochem. Sci. , 19 , 87–92. Theg, S.M., Bauerle, C., Olsen, L.J., Selman, B.R. and Keegstra, K. (1989). Internal ATP is the only energy requirement for the translocation of precursor proteins across chloroplastic membranes. J. Biol. Chem. , 264 , 6730–6736. Theg, S.M. and Geske, F.J. (1992). Biophysical characterization of a transit peptide directing chloroplast protein import. Biochemistry , 31 , 5053–5060. Theg, S.M. and Scott, S.V. (1993). Protein import into chloroplasts. Trends Cell. Biol. , 3 , 186–190. Tsugeki, R. and Nishimura, M. (1993). Interaction of homologues of Hsp70 and Cpn60 with ferredoxin-NADP+reductase upon its import into chloroplasts. FEBS Lett. , 320 ,, 198–202. Van Wijk, K.J., Knott, T.G. and Robinson, C. (1995). Evidence for a SecA- and DpHindependent insertion of D1 into thylakoids. FEBS Lett. , 368 , 263–266. Viitanen, P.V, Schmidt, M., Buchner, J., Suzuki, T., Vierling, E., Dickson, R., Lorimer, G.H., Gatenby, A.A. and Soll, J. (1995). Functional characterization of the higher plant chloroplast chaperonins. J. Biol. Chem. , 270 , 18158–18164.
Molecular chaperones and folding catalysts
308
Voelker, R. and Barkan, A. (1995). Two nuclear mutations disrupt distinct pathways for targeting proteins to the chloroplast thylakoid. EMBO J. , 14 , 3905–3914. Von Heijne, G., Steppuhn, J. and Herrmann, R.G. (1989). Domain structure of mitochondrial and chloroplast targeting peptides. Eur. J. Biochem. , 180 , 535–545. Von Heijne, G. and Nishikawa, K. (1991). Chloroplast transit peptides. The perfect random coil? FEBS Lett. , 278 , 1–3. Waegemann, K., Paulsen, H. and Soll, J. (1990). Translocation of proteins into isolated chloroplasts requires cytosolic factors to obtain import competence. FEBS Lett. , 261 , 89–92. Waegemann, K. and Soll, J. (1991). Characterization of the protein import apparatus in isolated outer envelopes of chloroplasts . Plant J. , 1 , 149–158. Waegemann, K. and Soll, J. (1993). Isolation and characterization of a functionally active protein translocation apparatus from chloroplast envelopes. In: D.J.Morre, E.Howell, and J.J.M Bergeron, (eds.), Molecular Mechanisms of Membrane Traffic . NATO ASI Series, Vol. H 74, Springer-Verlag, Berlin, Heidelberg, pp. 101–104. Waegemann, K. and Soll, J. (1996). Phosphorylation of the transit sequence of chloroplast precursor proteins. J. Biol. Chem. , 271 , 6545–6554. Walker, D., Chaddock, A.M., Chaddock, J.A., Roberts, L.M., Lord, J.M. and Robinson, C. (1996). Ricin A chain fused to a chloroplast-targeting signal is unfolded on the chloroplast surface prior to import across the envelope membranes. J. Biol. Chem. , 271 , 4082–4085. Wang, H., Goffreda, M. and Leustek, T. (1993). Characteristics of an Hsp70 homolog localized in higher plant chloroplasts that is similar to DnaK, the Hsp70 of prokaryotes. Plant Physiol , 102 , 843–850. Wickner, W., Driessen, A.J.M., Hartl, F.-U. (1991). The enzymology of protein translocation across the Escherichia coli plasma membrane . Annu. Rev. Biochem. , 60 , 101–124. Wu, C., Seibert, F.S. and Ko., K. (1994). Identification of chloroplast envelope proteins in close physical proximity to a partially translocated chimeric precursor protein. J. Biol. Chem. , 269 , 32264–32271. Yalovsky, S., Paulsen, H., Michaeli, D., Chitnis, P.R. and Nechushtai, R. (1992). Involvement of a chloroplast HSP70 heat shock protein in the integration of a protein (light-harvesting complex protein precursor), into the thylakoid membrane. Proc. Natl. Acad. Set. USA , 89 , 5616–5619. Yuan, J., Henry, R. and Cline, K. (1993). Stromal factor plays an essential role in protein integration into thylakoids that cannot be replaced by unfolding or by heat shock protein Hsp70. Proc. Natl. Acad. Set. USA , 90 , 8552–8556. Yuan, J., Henry, R., McCaffery, M. and Cline, K. (1994). SecA homolog in protein transport within chloroplasts: evidence for endosymbiont-derived sorting. Science , 266 , 796–798. Zimmermann, R., Sagstetter, M.J.L. and Pelham, H.R.B. (1988). Seventy-kilodalton heat shock proteins and an additional component from reticulocyte lysate stimulate import of M13 procoat protein into microsomes. EMBO J. , 7 , 2875–2880.
13. PROTEIN FOLDING IN THE PERIPLASM OF ESCHERICHIA COLI DOMINIQUE MISSIAKAS1, *, CLAIRE DARTIGALONGUE2 and SATISH RAINA2 1 University
of California Los Angeles, Department of Microbiology and Molecular Genetics, 609 Circle Drive East, Los Angeles CA 90024 2 Centre Médical Universitaire, Département de Biochimie Médicale, 1 Rue Michel-Servet, 1211 Genève 4, Switzerland
1. Introduction 2. Oxidation of Transported Proteins by the Dsb Proteins 2.1. Genetic Identification of the dsb Genes 2.2. The Chemistry of Thiol: disulfide Exchange Reactions 3. Biological Activity of the Dsb Proteins 3.1. The DsbA and DsbB Oxidizing System 3.2. Disulfide Isomerase and Thiol: disulfide Reductase Activities in the Periplasm 3.3. Biogenesis of c-type Cytochromes Requires the Dsb Proteins 4. The Other Folding Catalysts in the Periplasm 4.1. The Problem of Sensing Misfolded Proteins in the Cell Envelope 4.2. Folding Catalysts of the PPI Family 4.3. Folding of the Outer Membrane Proteins 5. Conclusion 6. References *Corresponding author
Protein folding in the periplasm of Escherichia coli
311
1. INTRODUCTION For a long time it has been debated whether secreted proteins need the assistance of molecular chaperones in the periplasm, as newly synthesized polypeptides emerging from the ribosomes do (Hartl, 1996; Hendrick and Hartl, 1993; Welch et al., this volume). The periplasmic space of gram-negative bacteria provides a special environment for protein folding for several reasons. First, it favors the oxidation of disulfide bonds, second it lacks ATP, a component essential for the activity of most cytoplasmic chaperones. The question whether periplasmic chaperones exist and associate with nascent polypeptides emerging at the translocation sites is related to the question of how these polypeptides can be protected against the attack by the many proteases which can rapidly degrade slow-folding newly translocated species. The DegP/HtrA protease in particular has been shown to degrade abnormal secreted polypeptides (Strauch and Beckwith, 1988) and its own synthesis is induced by stress, including elevated temperature (Lipinska et al., 1989). Furthermore, the presence of a porous outer membrane makes it difficult to control the physico-chemical conditions in the overall cell envelope of the bacterium. Changes in the osmolarity, pH, redox potential of the medium are likely to alter the native state of secreted proteins as well. Unlike the cytoplasmic proteins, many secreted proteins contain disulfide bonds. In most cases, such disulfide bonds contribute to the stabilization of folded proteins (Creighton, 1986). Reducing them, by breaking the covalent linkage, will induce the unfolding of the protein. Making up disulfide bonds appears to be a very fast reaction in vivo whereas it is a slower process in vitro that occurs randomly, often leading to the wrong pairing between two cysteines. Hence formation of disulfide bonds has to be controlled in a catalyzed manner. Recent genetic evidence have uncovered the existence of multiple thiol:disulfide oxido-reductases present in the periplasm and belonging to the protein disulfide isomerase family (PDI) (see also chapter by Freedman and Klappa). A second rate-limiting step in folding of several proteins is the cis to trans isomerization of the peptide backbone around the imide bond of prolines (see chapter by Fischer and Schmid). Here again, the rearrangement around the imide bond is a very slow reaction in vitro (Brandts et al., 1975) and is highly accelerated in vivo by the presence of specific enzymes known as peptidyl prolyl isomerases (PPI) (Fischer and Schmid, 1990). Such enzymes appear to play a key role in the folding of secreted proteins. This review will describe the discovery and role of these folding catalysts and folding helpers present in the periplasm. For mechanistic details of their activities the reader is referred to chapters by Fischer & Schmid and Freedman & Klappa.
2. OXIDATION OF TRANSPORTED PROTEINS BY THE DSB PROTEINS 2.1. Genetic Identification of the dsb Genes The first evidence that the bacterial periplasm contains a PDI-like protein was based on the biochemical purification of a thiol: disulfide oxido-reductase activity in E. coli osmotic shock fractions (Barth et al., 1988). The discovery of the first gene which might
Molecular chaperones and folding catalysts
312
encode such an activity was achieved independently. Genetic screens set up to isolate mutants impairing the proper insertion of inner membrane proteins identified a gene designated as dsbA (Bardwell et al., 1991). An hybrid between the membrane MalF protein and -galactosidase was used, and mutants were isolated on the base of a Lac+ phenotype, meaning that the hybrid failed to integrate in the membrane. In fact, in a dsbA mutant, -galactosidase probably remains reduced and as a consequence unfolded, allowing the hybrid MalF -gal to slide back to the cytoplasm. Mutations in dsbA were found to be highly pleiotropic and to impair the folding of many secreted disulfide-bond containing proteins. The dsb nomenclature, disulfide bond formation, was given after the deduced function of that gene (Bardwell et al., 1991). The same genetic approach later on, identified yet another gene dsbB, also encoding a redox protein located in the inner membrane with its active site facing the periplasm (Bardwell et al., 1993; Jander et al., 1994) The presence of genes encoding thiol: disulfide oxido-reductase activities in E. coli was also revealed in two independent genetic studies. Mutations in ppfA (dsbA) were isolated because they resulted in a loss of alkaline phosphatase activity, a periplasmic enzyme which contains two disulfide bonds in its native conformation. The defects in AP folding in such mutants was directly attributed to the lack of an oxidant in the periplasm (Kamitani et al., 1992). A third genetic search for redox proteins in E. coli was based on isolating mutants unable to cope with variations of the redox potential in the medium (Missiakas et al., 1993). It was found that addition of dithiothreitol (DTT) in the growth medium was particularly poorly tolerated in certain mutant bacteria. Mapping of these mutations identified again the dsbA and dsbB genes. Quite interesting was the finding that the ability to sustain growth on highly reductive medium (excess of DTT) could be achieved by cells which have multiple copies of the dsbB gene (Missiakas et al., 1993) or the dsbG gene (Anderson et al., 1997). This suggested that DsbB is a disulfide oxidant is a disulfide oxidant able to neutralize the toxic effect of DTT by re-oxidation. A search for suppressors able to reverse the DTT-sensitivity of dsbA or dsbB mutants identified three additional PDI-like activities in the periplasm: DsbC (Missiakas et al., 1994; Shevchick et al., 1994), DsbD and DsbE (Missiakas et al., 1995). dsbC was obtained as a multicopy suppressor of the hyper-sensitivity to DTT exhibited by dsbA mutants (Missiakas et al., 1994), whereas the identification of dsbD resulted from a search for null suppressors of dsbA sensitivity to DTT (Missiakas et al., 1995). DsbC is a periplasmic protein like DsbA. Its multicopy suppressing effect in a dsbA mutant suggests some overlap of functions between the two proteins. DsbD is a membrane bound protein with presumably the opposite activity of DsbA, that is a thiol: disulfide reductase activity. The suppression effect of a dsbD null mutation in dsbA null mutant bacteria is best explained if it arises by resetting a new oxidizing balance in the periplasm. Since DsbC accumulates under its oxidized form in such a background, it is presumed that it actively substitutes for the lack of DsbA (Missiakas et al., 1995). 2.2. The Chemistry of Thiol: disulfide Exchange Reactions Biochemical characterization of the thiol: disulfide exchange abilities of the known Dsb proteins has been best achieved with DsbA and DsbC, mostly because both of them are
Protein folding in the periplasm of Escherichia coli
313
soluble proteins whereas DsbB and DsbD are membrane proteins. All of these redox proteins contain a pair of cysteines in their active site separated by two amino acids (Figure 1). Despite the overall lack of sequence homology, DsbA contains a domain superimposable to the thioredoxin fold as demonstrated by the resolution of its threedimensional structure (Martin et al., 1993). From its biochemical properties, DsbC is likely to contain such a fold as well (Zapun et al., 1995). For all members of the thioredoxin super-family, a disulfide bond can be formed reversibly between the two sulfur atoms of the cysteines in the active site (chapter Freedman and Klappa). However, only the first N-terminal cysteine in the motif is exposed and reacts with the thiol groups of other proteins. Yet, the chemistry of such thiol: disulfide reversible active sites varies between all members. The stability of the disulfide bond in thioredoxin’s active site confers to the protein its property as a good hydrogen donor, and greatly stabilizes the overall conformation of the protein (Holmgren, 1995). In PDI, DsbA or DsbC, this disulfide bond is very unstable and therefore the reduced form of these proteins is more stable (Lundström
Figure 1 The thiol/disulfide active sites of members of the thioredoxin superfamily.
and Holmgren, 1993; Zapun et al., 1993, 1995; Wunderlich et al., 1993). Hence such redox proteins are quite good at transferring their disulfide bonds to other proteins. Although both DsbA and DsbC active sites are nearly as unstable, the reactivity of the cysteines is quite different. Conclusive evidence have shown that DsbA is a very potent disulfide oxidant (Zapun et al., 1993; Wunderlich et al., 1993) whereas DsbC possesses a better isomerase activity (Zapun et al., 1995). An isomerase activity becomes essential when wrong pairings between cysteines occur and no free cysteine is available in the protein to allow an intramolecular reshuffling. DsbC was shown to allow the reshuffling of disulfides in quasi-native, fully oxidized species of bovine pancreatic trypsin inhibitor (Zapun et al., 1995). Such an activity implies that DsbC is reduced, i.e. the first cysteine of the active site is free. This also suggests that unlike DsbA, DsbC has an unfoldase activity since it can recognize quasi-native proteins as its substrates (Zapun et al., 1995).
Molecular chaperones and folding catalysts
314
The high oxidizing capacity of DsbA has been explained by the finding that the two residues within the DsbA active site play a direct role in lowering the pKa of the most Nterminal cysteine in that motif (Nelson and Creighton, 1994; Grauschopf et al., 1995). It is not yet clear why DsbC behaves more like the eukaryotic PDI and possesses an isomerase activity, which allows both catalysts to undo and rebuilt wrongly paired cysteines in native like folding intermediates (Zapun et al., 1995). The fact that DsbC is a dimer, each carrying an active site, may confer such an isomerase activity, since monomeric PDI itself contains twothiol: disulfide active sites.
3. BIOLOGICAL ACTIVITY OF THE DSB PROTEINS 3.1. The DsbA and DsbB Oxidizing System The first evidence that DsbA and DsbB act in concert to oxidize various transported proteins consisted in the observation that DsbA accumulates mostly into the reduced form in the periplasm of dsbB mutant bacteria (Bardwell et al., 1993; Missiakas et al., 1994). Obviously, for its correct activity, DsbA needs to be re-oxidized once it has transferred its own disulfide to a substrate protein (Figure 2). Hence a first model predicted that DsbB could re-oxidize the reduced DsbA protein thereby re-cycling an active DsbA redox protein (Bardwell et al., 1993). Because of the thiol: disulfide chemistry, this model implies that a mixed disulfide between DsbA and DsbB is formed via the cysteine of their active sites. Such a mixed disulfide had already been isolated between glutaredoxin (mutant Cys14→Ser) and glutathione in the cytoplasm, provided that only the reactive, first cysteine of each di-thiol motif was present (Bushweller et al., 1992). Similarly a mixed disulfide between DsbA-Cys30: Cys104-DsbB could be readily trapped in vivo (Guilhot et al., 1995; Kishigami et al. 1995). DsbA contains only the two cysteines of its active site whereas DsbB contains six cysteines, out of which four facing the periplasm are essential for DsbB activity (Jander et al., 1994). Cys104 forms a reversible disulfide bond with Cys130 which is the one responsible for recycling DsbA (Kishigami and Ito, 1996). The Cys104: Cys130 linkage is re-oxidized intra-molecularly by the N-terminally located Cys41-Val-Leu-Cys44 thioredoxin-like motif of the DsbB protein (Kishigami and Ito, 1996). It appears that the respiratory electron transfer chain (Ubi-MenA) participates in the oxidation of DsbB (Kobayashi et al., 1997). 3.2. Disulfide Isomerase and Thiol: disulfide Reductase Activities in the Periplasm The finding that overexpression of DsbC functionally substitutes for the lack of DsbA in the periplasm, suggested that both enzymes have overlapping functions (Figure 2). Comparison of the chemical properties of their redox active sites suggested that under its reduced form, DsbC possesses a rather good isomerase activity which is not the case with DsbA. Since in a dsbD mutant both DsbA and DsbC accumulate into their oxidized forms, a situation which causes abnormal protein oxidation (Missiakas et al., 1995), it was suggested that DsbD acts as a thiol: disulfide reductant (Figure 2). Recent
Protein folding in the periplasm of Escherichia coli
315
biochemical analyses have confirmed that it is indeed the case since DsbD redox properties are much like those of thioredoxin (Raina and Creighton, unpublished data). Hence, DsbD could partly be responsible for maintaining a physiologically significant pool of reduced DsbC in the periplasm thereby ensuring its isomerase activity. The model presented in Figure 2 also suggests that transported proteins containing disulfide bonds in their final native conformation are maintained reduced in the cytoplasm. Two independent reports suggest that thioredoxin reductase (TrxB) is the likely candidate for doing so. First, trxB mutants have been found to be sensitive to DTT in the same genetic scheme which revealed the importance of DsbA and DsbB for the folding of periplasmic disulfide containing proteins (Missiakas et al., 1993). Second, mutations in trxB were found to allow disulfide bond formation in the cytoplasm (Derman et al., 1993). Finally, both trxA and trxB mutants, quite like dsbD mutants, lead to the accumulation of oxidized DsbC in the periplasm (Rietsch et al., 1997).
Figure 2 Disulfide bond formation catalyzed by the various Dsb enzymes in the periplasm of E. coli.
It is likely that DsbD is not the unique thiol: disulfide reductant present in the periplasm. Overexpressing another thiol: disulfide oxido-reductase, designated as DsbE, could suppress some phenotypic defects associated with dsbD null mutants. In particular, randomoxidation of proteins due to the accumulation of oxidized DsbA and DsbC in such mutants was avoided to a great extent when DsbE was overexpressed (Missiakas and Raina, unpublished). DsbE was a purified soluble periplasmic protein (Missiakas, unpublished data) with Cys-Pro-Thr-Cys being its active site. The equilibrium constant (Kox) for the formation of the active site disulfide bond was estimated to be 250 mM by measuring its interchange with glutathione (Raina and Creighton, unpublished data). This suggests that DsbE is a reductant (as a comparison Kox values of 10 M and 80 M have been obtained for thioredoxin and DsbA, respectively).
Molecular chaperones and folding catalysts
316
3.3. Biogenesis of c-type Cytochromes Requires the Dsb Proteins The periplasm of E. coli contains at least five c-type cytochromes which are synthesized during anaerobic growth, c-type cytochromes are electron transfer proteins being part of the respiratory chains and using as electron acceptors either nitrite, nitrate or trimethylamine-N-oxide (Iobbi-Nivol et al., 1994; Grove et al., 1996). c-type cytochromes differ from the other cytochromes since the haem moiety they contain is covalently liganded via the sulfur atoms of two cysteines. Although no evidence for a haem lyase exists, many mutants impairing apo-protein maturation have been isolated. Some of them have been shown to map in genes predicted to encode either components of ABC-type transporters which may be involved in haem transport (Crooke and Cole, 1995) or genes coding for thioredoxin-like proteins, like HelX in Rhodobacter capsulatus (Beckman and Kranz, 1993) and TlpA in Bradyrhizobium japonicum (Loferer et al., 1993). In E. coli, the dsbD and dsbA genes have also been shown to be essential for formate-dependent nitrite reductase (Nrf) activity and c-type cytochrome biogenesis (Crooke and Cole, 1995; Metheringham et al., 1995; dsbD is referred as dipZ in these studies). Both DsbD and DsbE which have been shown to behave as thiol: disulfide reductant in vitro (Kox=1.5 M for DsbD, Raina and Creighton, unpublished data), may be good candidates to maintain the cysteine residues of the apo-cytochrome into a reduced state and thereby allow the proper covalent linkage with the haem moiety. More recently, DsbB has also been shown to be required for the synthesis of c-type cytochromes (Metheringham et al., 1996). However and quite unlike dsbA mutants, dsbB mutant bacteria are also defective in periplasmic nitrate reductase activity (Metheringham et al., 1996). This argues that DsbA and DsbB may not always function together, and DsbB may be involved in the oxidation of some other proteins than DsbA as well.
4. THE OTHER FOLDING CATALYSTS IN THE PERIPLASM 4.1. The Problem of Sensing Misf olded Proteins in the Cell Envelope Protein misfolding occurring in the cell envelope, that is periplasm or outer membrane, has been shown to trigger a specific cell response (see also chapter Connolly et al.). In particular, such a phenomenon induces the transcription of the htrA (degP) gene which encodes a periplasmic protease efficient at removing abnormal proteins (Strauch and Beckwith, 1988). In E. coli there are at least two transduction pathways which can signal the presence of misfolded proteins to the regulatory components responsible for htrA increased transcription. The main pathway involves a sigma factor (Raina et al., 1995; Rouvière et al., 1995) shown to belong to the Extra-Cytoplasmic Factor sub-family (ECF). It has been suggested that ECFs regulate extracytoplasmic functions and respond to extracytoplasmic stimuli (Lonetto et al., 1994). The -dependent response has been shown to be induced by the accumulation of outer membrane proteins (OMPs) (Mecsas et al., 1993) as well as misfolded periplasmic proteins (Missiakas et al., 1996). Under normal conditions, is negatively regulated at the post-translational level by a direct protein-protein interaction with RseA. RseA is an inner membrane protein with an anti
Protein folding in the periplasm of Escherichia coli
317
activity (Missiakas et al., 1997; De Las Penas et al., 1997). The second pathway leading to increased htrA transcription utilizes the CpxA CpxR two-component regulatory system (Danese et al., 1995; Raina et al., 1995; Missiakas and Raina, 1997). CpxA is an inner membrane histidine kinase and CpxR a response regulator activated upon phosphorylation by CpxA. This pathway is also fine tuned at the post-translational level by the activity of a phosphoprotein phosphatase PrpA (Missiakas and Raina, 1997). Although global protein misfolding leads to both and CpxA CpxR activation, it is not yet clear how the stimuli are sensed at the molecular level by the periplasmic domains of RseA and CpxA (Figure 3). Recent data have revealed that if htrA is transcribed by both CpxA CpxR and , other periplasmic folding catalysts appear to belong to either one of these stress regulons. DsbA, RotA and PpiD, a thiol disulfide oxidant and two PPIs, respectively, are all regulated via the CpxA CpxR system (Pogliano et al., 1997; Danese and Silhavy, 1997; Dartigalongue and Raina, 1998), whereas FkpA belongs to the regulon (Danese and Silhavy, 1997) (Figure 3). 4.2. Folding Catalysts of the PPI Family Unlike the PDI-like enzymes, PPIs can be found in any cell compartment. These enzymes catalyze the slow trans to cis isomerization of prolyl peptide bonds such as Xaa-Pro (Fischer and Schmid, this volume). In eukaryotes, PPIs were first identified as being the proteins which are the targets of two immuno-suppressive agents cyclosporin A (CsA) and the macrolide FK506. Hence, two sub-classes of enzymes were first distinguished: the cyclophilins which bind CsA and the FKBP or FK506 binding proteins (there are other macrolides as well with various affinities for this sub-class of PPIs). It is only recently that a third class was revealed with the discovery of Parvulin in the cytoplasm of E. coli (Rahfeld et al., 1994). This last class is referred as PPIc (Rudd et al., 1995). The first report of a PPI activity in the periplasm of E, coli was the result of a biochemical analysis of a 190 amino acid open reading frame (Liu and Walsh, 1990) exhibiting 35% homology to human cyclophylin (Kawamukai et al., 1989). This protein is designated as RotA and despite the sequence homology, its PPI activity is only marginally inhibited by cyclosporin A (Liu and Walsch, 1990). No noticeable growth or folding defects have been observed in rotA mutant bacteria (Kleerebezem et al., 1995) and the proteins requiring RotA activity for their folding are still not known. There are at least three other PPIs in the periplasm: SurA, FkpA and PpiD (Missiakas et al., 1996; Dartigalongue and Raina, 1998). The surA gene had been identified as a gene whose product is required for survival of E. coli under stationary phase (Tormo et al., 1990). Sequence analysis of the gene showed that the SurA protein shares a high degree of homology with the small cytoplasmic protein Parvulin whose function was revealed later (Rahfeld et al., 1994; Rudd et al., 1995). In fact, SurA contains two repeats of Parvulin-like domains at the C-terminus. Roles for SurA as chaperone and folding catalyst have been proposed based on its ability to prevent protein misfolding/aggregation in the periplasm when present in multicopy (Missiakas et al., 1996). Hence, SurA was found to assist the folding of OMPs which was impaired in the absence of proper lipopolysaccharide. SurA in multicopy was also found to promote the folding of
Molecular chaperones and folding catalysts
318
otherwise unstable proteins such as Protein A -lactamase hybrid protein or aggregation prone proteins like mutant MalE31 and a truncated form of the periplasmic Asparaginase synthase B protein. Finally, SurA could restore to near wild type level the -dependent
Figure 3 Transduction pathways responding to protein misfolding in the cell envelope (for details see Connolly et al. this volume).
response otherwise constitutively induced by misfolded proteins accumulating in bacteria carrying simultaneous mutations in the htrA and one of the dsb genes. In vitro, a rather weak PPI activity has been measured with the purified SurA protein (Missiakas et al., 1996). Such activities are usually monitored by using model peptides and may not reflect the real biochemical activity of SurA. Looking for clones which prevented the induction of the -dependent response by misfolded proteins identified a second gene encoding the FkpA protein (Missiakas et al., 1996). The fkpA gene had already been reported but no function, and a wrong cellular location (an inner membrane location), had been assigned to the protein (Horne and Young, 1995). However, a similarity with the eukaryotic FK506 binding proteins (FKBPs) could be deduced from the predicted amino acid sequence encoded by fkpA. Hence, the proposed nomenclature FkpA was retained (Horne and Young, 1995; Missiakas et al., 1996). FkpA was found to be a periplasmic protein and the purified protein exhibits a rather good PPI activity in vitro (Missiakas et al., 1996), comparable to that of the FkpA homologue MipA from Legionella pneumophila (Fischer et al., 1992). Hence, overexpressing FkpA in conditions leading to protein misfolding in the periplasm, provides a mean to prevent further misfolding and thereby to tune down the dependent response. Recently, a fourth PPI, PpiD, has been isolated as a multicopy
Protein folding in the periplasm of Escherichia coli
319
suppressor of surA mutants, based on its ability to allow survival of surA deficient strains when grown on novobiocin containing plates (Dartigalongue and Raina, 1998). ppiD has been shown to be regulated by CpxA CpxR but also by the heat shock factor . The protein was found to be anchored to the inner membrane with its active site facing the periplasm. 4.3. Folding of the Outer Membrane Proteins As mentioned previously, SurA is able to assist the folding of OMPs. It is not yet clear whether the PPI activity of SurA is involved or rather a chaperone-like activity of SurA. In addition to the two Parvulin-like domains, SurA contains a large 30 kDa domain which may serve as a specific substrate binding pocket. Such a role of SurA in OMPs folding was deduced from the finding that multicopy SurA could reverse the lethal growth phenotype of rfaD (htrM) null mutants on bile salt containing media (Missiakas et al., 1996). Such mutants synthesize a heptose-less lipopolysaccharide (LPS) (Pegues et al., 1990; Raina and Georgopoulos, 1991). Proper LPS and lipidA have been shown to be necessary for the stability and folding of OMPs (Nikaido and Vaara, 1985; Kloser et al., 1998). Quite like rfaD, surA mutants are extremely permeable and sensitive to hydrophobic and bulky compounds, such as bile salts and novobiocin Such sensitivity can be suppressed by overproducing PpiD. PpiD contains one Parvulin-like domain and a double surA ppiD null is lethal for E. coli (Dartigalongue and Raina, 1998). Presumably, both SurA and PpiD accelerate an early folding process of OMPs which are otherwise rapidly degraded by the HtrA/ DegP protease (Missiakas et al., 1996; Figure 4). In agreement with these findings, surA and ppiD mutants contain a leaky outer membrane (Missiakas et al., 1996; Dartigalongue and Raina, 1998). It was independently reported that in a surA mutant, most of the OMPs are highly sensitive to trypsin degradation meaning that their folding is retarded or impaired (Lazar and Kolter, 1996). A prominent role for a third protein Skp/OmpH during OMP folding was also deduced based on similar phenotypic defects observed with both skp and surA mutant bacteria. Both skp and surA mutants exhibit an elevated -dependent response and defects in the protein composition of the outer membrane (Missiakas et al., 1996; Chen and Henning, 1996). Skp was also shown to bind selectively to certain OMPs suggesting a chaperonelike role (Chen and Henning, 1996). Given the fact that Skp/OmpH also co-purifies with LPS (Geyer et al., 1979), it has been proposed that Skp may act as an exchange factor that exchanges bound LPS for pro-trimeric OMPs (Missiakas et al., 1996; Figure 4). This exchange of LPS could trigger the insertion of OMPs into the outer membranes and thereby accelerate the final folding step of OMPs.
5. CONCLUSION It emerges from various recent works that protein transiting through the periplasm are readily folded in a catalyzed manner. E. coli synthesizes very potent disulfide oxidants such as DsbA and DsbB which play key roles in the oxidation of disulfide
Molecular chaperones and folding catalysts
320
Figure 4 Proposed model for the folding of trimeric porin proteins of the outer membrane.
bond containing proteins. The recent discoveries of thiol: disulfide reductants such as the DsbD and DsbE redox proteins are the first clues to our understanding of c-type cytochrome synthesis and possibly to the maintenance of the DsbC disulfide isomerase activity. In addition, the periplasm contains at least four PPIs, RotA, an FKBP homologue FkpA, and SurA and PpiD belonging to the PPIc family. Out of these, SurA, PpiD and Skp may have in addition chaperone-like properties, essential for the folding of the OMPs. Deciphering the folding status of the substrates interacting with these various catalysts will open new insights to our understanding of protein folding in vivo. Finally, problems arising upon protein misfolding in the cell envelope generate multiple cell responses. The molecular events leading to these responses as well as their nature are still unknown and their studies represent exciting prospects in molecular biology.
6. REFERENCES Andersen, C.L., Matthey-Dupraz, A., Missiakas, D. and Raina, S. (1997). A new Escherichia coligene, dsbG, encodes a periplasmic protein involved in disulphide bond formation, required for recycling DsbA/ DsbB and DsbC redox proteins. Mol. Microbiol. , 26 , 121–132. Bardwell, J.C.A., Lee, J.-O., Jander, G., Martin, N., Belin, D. and Beckwith, J. (1993). A pathway for disulfide bond formation in vivo. Proc. Natl. Acad. Sci. USA , 90 , 1038– 1042. Bardwell, J.C.A., McGovern, K. and Beckwith, J. (1991). Identification of a protein required for disulfide bond formation in vivo . Cell , 67 , 581–589. Barth, P.T., Bust, C., Hawkins, H.C. and Freedman, R.B. (1988). Protein disulfide isomerase activity in bacterial osmotic shock preparations . Biochem. Soc. Transact. , 16 , 57.
Protein folding in the periplasm of Escherichia coli
321
Beckman, D.L. and Kranz, R.G. (1993). Cytochromes c biogenesis in a photosynthetic bacterium requires a periplasmic thioredoxin-like protein. Proc. Natl. Acad. Sci. USA , 90 , 2179–2183. Brandts, J.F., Halvorson, H.R. and Brennan, M. (1975). Consideration of the possibility that the slow step in protein denaturation reactions is due to cis-trans isomerism of proline residues. Biochemistry , 14 , 4953–4963. Bushweller, J.H., Åslund, F., Wütrich, K. and Holmgren, A. (1992). Structural and functional characterization of the mutant Escherichia coli glutaredoxine (C14<arrow>S). and its mixed disulfide with glutathione. Biochemistry , 31 , 9288– 9293. Chen, R. and Henning, U. (1996). A periplasmic protein (Skp). of Escherichia coli selectively binds a class of outer membrane proteins. Mol. Microbiol , 19 , 1287–1294. Creighton, T.E. (1986). Detection of folding intermediates using urea-gradient electrophoresis. Methods Enzymol. , 131 , 156–172. Crooke, H. and Cole, J. (1995). The biogenesis of c-type cytochromes in Escherichia coli requires a membrane-bound protein, DipZ, with a protein disulphide isomerase-like domain. Mol. Microbiol. , 15 , 1139–1150. Danese, P. and Silhavy T.J. (1997). The sigma(E) and the Cpx signal transduction systems control the synthesis of periplasmic protein-folding enzymes in Escherichia coli. Genes Dev. , 11 , 1183–1193. Danese, P., Snyder, W.B., Cosma, C., Davis, L. and Silhavy, T. (1995). The Cpx twocomponent signal transduction pathway of Escherichia coli regulates transcription of the gene specifying the stressinducible periplasmic protease. Genes & Dev. , 9 , 387– 398. Dartigalongue, C. and Raina, S. (1998). PpiD is a new heat shock protein encoding a peptidyl prolyl isomerase. EMBO J. , in press. De Las Penas, A., Connolly, L. and Gross, C.A. (1997). The sigmaE-mediated response to extracytoplasmic stress in Escherichia coli is transduced by RseA and RseB, two negative regulators of sigmaE. Mol. Microbiol. , 24 , 373–385. Derman, A.I., Prinz, W.A., Belin, D. and Beckwith, J. (1993). Mutations that allow disulfide bond formation in the cytoplasm of Escherichia coli. Science , 262 , 1744– 1747. Fischer, G., Bang, H., Ludwig, B., Mann, K. and Hacker, J. (1992). Mip protein of Legionella pneumophila exhibits peptidyl-prolyl-cis/trans isomerase (PPIase). activity. Mol. Microbiol , 6 , 1375–1383. Fischer, G. and Schmid, F.X. (1990). The mechanism of protein folding. Implications of in vitro refolding models for de novo protein folding and translocation in the cell. Biochemistry , 29 , 2205–2212. Geyer, R., Galanos, C., Westphal, O. and Golecki, J.R. (1979). A Lipopolysaccharidebinding cell-surface protein from Salmonella minnesota. Isolation, partial characterization and occurrence in different Enterobacteriaceae. Eur. J. Biochem. , 98 , 27–38. Grauschopf, U., Winther, J.R., Korber, P., Zander, T., Dallinger, P. and Bardwell, J.C. (1995). Why is DsbA such an oxidizing disulfide catalyst? Cell , 83 , 947–955. Guilhot, C., Jander, G., Martin, N.L. and Beckwith, J. (1995). Evidence that the pathway of disulfide bond formation in Escherichia coli involves interactions between the cysteines of DsbB and DsbA. Proc. Natl Acad. Set. USA , 92 , 9895–9899. Grove, J., Tanapongpipat, S., Thomas, G., Griffiths, L., Crooke, H. and Cole, J. (1996). Escherichia coli K-12 genes essential for the synthesis of c-type cytochromes and a
Molecular chaperones and folding catalysts
322
third nitrate reductase located in the periplasm. Mol. Microbiol. , 19 , 467–481. Hartl, F.U. (1996). Molecular chaperones in cellular protein folding. Nature , 381 , 571– 580. Hendrick, J.P. and Hartl, F.-U. (1993). Molecular chaperone functions of heat-shock proteins. Ann. Rev. Biochem. , 62 , 349–384. Holmgren, A. (1995). Thioredoxin structure and mechanism: conformational changes on oxidation of the active-site sulfhydryls to a disulfide. Structure , 3 , 239–243. Horne, S.M. and Young, K.D. (1995). Escherichia coli and other species of Enterobacteriaceae encode a protein similar to the family of Mip-like FK506-binding proteins. Arch. Microbiol. , 163 , 357–365. Iobbi-Nivol, C, Crooke, H., Griffiths, L., Grove, J., Hussain, H., Pommier, J., Méjean, V. and Cole, J.A. (1994). A reassessment of the range of c-type cytochromes synthesized by Escherichia coli K-12. FEMS Microbiol. Lett. , 119 , 89–94. Jander, G., Martin, N.L. and Beckwith, J. (1994). Two cysteines in each periplasmic domain of the membrane protein DsbB are required for its function in protein disulfide bond formation. EMBO J. , 13 , 5121–5127. Kamitani, S., Akiyama, Y. and Ito, K. (1992). Identification and characterization of an Escherichia coli gene required for the formation of correctly folded alkaline phosphatase, a periplasmic enzyme. EMBO J. , 11 , 57–62. Kawamukai, M., Matsuda, H., Fujii, W., Utsumi, R. and Komano, T. (1989). Nucleotide sequences of fic and fic-1 genes involved in cell filamentation induced by cyclic AMP in Escherichia coli. J. Bacterial , 171 , 4525–4529. Kishigami, S. and Ito, K. (1996). Roles of cysteine residues of DsbB in its activity to reoxidize DsbA, the protein disulphide bond catalyst of Escherichia coli. Genes to cells , 1 , 201–208. Kishigami, S., Kanaya, E., Kikuchi, M. and Ito, K. (1995). DsbA-DsbB interaction through their active site cysteines. J. Biol. Chem. , 270 , 17072–17074. Kleerebezem, M., Heutink, M. and Tommassen, J. (1995). Characterization of an Escherichia coli rotA mutant, affected in periplasmic peptidyl-prolyl cis/trans isomerase. Mol. Microbiol , 18 , 313–320. Kloser, A., Laird, M., Deng, M. and Misra, R. (1998). Modulations in lipid A and phospholipid biosynthesis pathways influence outer membrane protein assembly in Escherichia coli K-12. Mol. Microbiol. , 27 , 1003–1008. Kobayashi, T., Kishigami, S., Sone, M., Inokuchi, H., Mogi, T. and Ito, K. (1997). Respiratory chain is required to maintain oxidized states of the DsbA-DsbB disulfide bond formation system in aerobically growing Escherichia coli cells. Proc. Natl. Acad. Sci. USA , 94 , 11857–11862 Lazar, S. and Kolter, R. (1996). SurA assists the folding of Escherichia coli outer membrane proteins. J. Bacterial. , 178 , 1770–1773 Lipinska, B., Fayet, O., Baird, L. and Georgopoulos, C. (1989). Identification, characterization, and mapping of the Escherichia coli htrA gene, whose product is essential for bacterial growth only at elevated temperatures. J. Bacterial. , 171 , 1574– 1584. Liu, J. and Walsh, C.T. (1990). Peptidyl-prolyl cis-trans-isomerase from Escherichia coli: a periplasmic homolog of cyclophilin that is not inhibited by cyclosporin A. Proc. Natl. Acad. Sci. USA , 87 , 4028–32 Loferer, H., Bott, M. and Hennecke, H. (1993). Bradyrhizobium japonicum TlpA, a novel membrane-anchored thioredoxin-like protein involved in the biogenesis of cytochrome aa3 and development of symbiosis. EMBO J. , 12 , 3373–3383.
Protein folding in the periplasm of Escherichia coli
323
Lonetto, M., Brown, K.L., Rudd, K.E. and Buttner, M.J. (1994). Analysis of the Streptomyces coelicolor sigmaE gene reveals the existence of a subfamily of Eubacterial RNA polymerase sigma factors involved in the regulation of extracytoplasmic functions. Proc. Natl. Acad. Sci. USA , 91 , 7573–7577. Lundstrom, J. and Holmgren, A. (1993). Determination of the reduction-oxidation potential of the thioredoxin-like domains of protein disulfide-isomerase from the equilibrium with glutathione and thioredoxin. Biochemistry , 32 , 6649–6655. Martin, J.L., Bardwell, J.C.A. and Kuriyan, J. (1993). Crystal structure of the DsbA protein required for disulphide bond formation in vivo. Nature , 365 , 464–468. Mecsas, J., Rouvière, P.E., Erickson, J.W., Donohue, T.J. and Gross, C.A. (1993). The activity of SE, an Escherichia coli heat-inducible sigma factor, is modulated by expression of outer membrane proteins. Genes & Dev. , 7 , 2618–2628. Metheringham, R., Tyson, K., Crooke, H., Missiakas, D., Raina, S. and Cole, J. (1996). Effects of mutations in genes for proteins involved in disulfide bond formation in the periplasm on the activities of anaerobically induced electron transfer chains in Escherichia coli K-12. Mol. Gen. Genet. , in press. Metheringham, R., Griffiths, L., Crooke, H., Forsythe, S. and Cole, J. (1995). An essential role for DsbA in cytochrome c synthesis and formate-dependent nitrite reduction by Escherichia coli K-12. Arch. Microbiol. , 164 , 301–307. Missiakas, D., Mayer, M.P., Lemaire, M., Georgopoulos, C. and Raina, S. (1997). Modulation of the Escherichia coli sigmaE (RpoE). heat-shock transcription-factor activity by the RseA, RseB and RseC proteins. Mol. Microbiol. , 24 , 355–371. Missiakas, D. and Raina, S. (1997). Signal transduction pathways in response to protein misfolding in the extra-cytoplasmic compartments of E. coli: Role of two phosphoprotein phosphatases. EMBO J. , 16 , 1670–1685. Missiakas, D., Betton, J.-M. and Raina, S. (1996). New components of protein folding in the extracytoplasmic compartments of Escherichia coli: SurA, FkpA and Skp/OmpH. Mol. Microbiol. , 21 , 871–884. Missiakas, D., Schwager, F. and Raina, S. (1995). Identification and characterization of a new disulfide-isomerase like protein (DsbD). in Escherichia coli. EMBO J. , 14 , 3415–3424. Missiakas, D., Georgopoulos, C. and Raina, S. (1994). The Escherichia coli dsbC (xprA). gene encodes a periplasmic protein involved in disulfide bond formation. EMBO J. , 13 , 2013–2020. Missiakas, D., Georgopoulos, C. and Raina, S. (1993). Identification and characterization of the Escherichia coli gene dsbB, whose product is involved in the formation of disulfide bonds in vivo. Proc, Natl. Acad. Sd. USA , 90 , 7084–7088. Nelson, J.W. and Creighton, T.E. (1994). Reactivity and ionization of the active site cysteine residues of DsbA, a protein required for disulfide bond formation in vivo . Biochemistry , 33 , 5974–5983. Nikaido, H. and Vaara, M. (1985). Molecular basis of bacterial outer membrane permeability. Microbiol. Rev. , 49 , 1–32. Pegues, J.C., Chen, L., Gordon, A.W., Ding, L. and Coleman, W.G., Jr. (1990). Cloning, expression and characterization of the Escherichia coli K-12 rfaD gene. J. Bacteriol , 172 , 4652–4660. Pogliano, J., Lynch, A.S., Belin, D., Lin, E.C. and Beckwith, J. (1997). Regulation of Escherichia coli cell envelope proteins involved in protein folding and degradation by the Cpx two-component system. Genes Dev. , 11 , 1169–1182. Rahfeld, J.-U., Schierhorn, A., Mann, K. and Fischer, G. (1994). A novel peptidyl-prolyl
Molecular chaperones and folding catalysts
324
cis/trans isomerase from Escherichia coli. FEBS Lett. , 343 , 65–69. Raina, S., Missiakas, D. and Georgopoulos, C. (1995). The rpoEE gene encoding the SE (s24). heat shock sigma factor of Escherichia coli. EMBO J. , 14 , 1043–1055. Raina, S. and Georgopoulos, C. (1991). The htrM gene, whose product is essential for Escherichia coli viability only at elevated temperatures, is identical to the rfaD gene. Nucleic Acids Res. , 19 , 3811–3819. Rietsch, A., Bessette, P., Georgiou, G. and Beckwith, J. (1997). Reduction of the periplasmic disulfide bond isomerase, DsbC, occurs by passage of electrons from cytoplasmic thioredoxin. J. Bacteriol. , 179 , 6602–6608. Rouvière, P., de las Penas, A., Mecsas, J., Lu, C.Z., Rudd, K.E. and Gross, C.A. (1995). rpoE, the gene encoding the second heat-shock sigma factor, SE, in Escherichia coli . EMBO J. , 14 , 1032–1042. Rudd, K.E., Sofia, H.J., Koonin, E.V., Plunkett, G., Lazar, S. and Rouvière, P.E. (1995). New family of peptidyl-prolyl isomerases. Trends Biochem. Sci. , 20 , 12–14. Shevchik, V.E., Condemine, G., and Robert-Baudouy, J. (1994). Characterization of DsbC, a peiplasmic protein of Erwinia chrysanthemi and Escherichia coli with disulfide isomerase activity. EMBO J. , 13 , 2007–2012. Strauch, K.L. and Beckwith, J. (1988). An Escherichia coli mutation preventing degradation of abnormal periplasmic proteins. Proc. Natl. Acad. Sci. USA , 85 , 1576– 1580. Tormo, A., Almiron, M. and Kolter, R. (1990). surA, an Escherichia coli gene essential for survival in stationary phase. J. Bacteriol. , 172 , 4339–4347. Wunderlich, M. and Glockshuber, R. (1993). Redox properties of protein disulfide isomerase (DsbA). from Escherichia coli. Protein Sci. , 2 , 717–726. Zapun, A., Missiakas, D., Raina, S. and Creighton, T.E. (1995). Structural and functional characterization of DsbC, a protein involved in disulfide bond formation in Escherichia coli. Biochemistry , 34 , 5075–5089. Zapun, A., Bardwell, J.C.A. and Creighton, T.E. (1993). The reactive and destabilizing disulfide bond of DsbA, a protein required for protein disulfide bond formation in vivo. Biochemistry , 32 , 5083–5092.
14. ROLE OF CHAPERONES IN REPLICATION OF BACTERIOPHAGE LAMBDA DNA M.ZYLICZ*, A.WAWRZYNOW, J.MARSZALEK, K.LIBERER, B.BANECKI, I.KONIECZNY, A.BLASZCZAK1, P.BARSKI, J.JAKOBKIEWICZ, M.GONCIARZ-SWIATEK, M.DUCHNIEWICZ, J.PUZEWICZ and J.KRZEWSKA Department of Molecular and Cellular Biology, Faculty of Biotechnology, University of Gdansk, 80–822 Gdansk, Kladki 24, Poland 1 Polish Academy of Sciences, Institute of Biochemistry and Biophysics, Laboratory of Molecular Biology affiliated to the University of Gdansk, 80–822 Gdansk, Poland
1. Introduction 2. Early DNA Replication 2.1. Assembly of the O-some Structure 2.2. Assembly of the Preprimosomal Structure 2.3. Activation of the Preprimosomal Structure 2.4. Uni- and Bidirectional 3. 4. 5. 6.
Switch from Early to Late Concluding Remarks Acknowledgments References
DNA Replication DNA Replication
1. INTRODUCTION Genetic and biochemical studies on the involvement of Escherichia coli molecular chaperones in bacteriophage DNA replication gave the first evidence for the synergistic and cooperative action of the DnaK/DnaJ/GrpE molecular chaperone machine (for review see Georgopoulos et al., 1990, 1994). *Corresponding author
Molecular chaperones and folding catalysts
326
This particular system is now a paradigm for our understanding of fundamental biochemical processes, such as protein folding, intracellular protein translocation, selective proteolysis and autoregulation of heat shock response (for reviews see Hendrick and Haiti, 1993; Hartl, 1996). The in vitro replication system was also the first functional assay for purification of active DnaK, DnaJ and GrpE proteins (Zylicz et al., 1983; Zylicz and Georgopoulos, 1984; Zylicz et al., 1985, 1987). The replication of bacteriophage DNA depends on interactions between phageencoded replication proteins and the bacterial host’s replication machinery. The single cycle of reproductive growth which generates approximately 100 copies of viral genome may be divided into an early and a late phase (for review, see Furth and Wickner, 1983; Taylor and Wegrzyn, 1995). In the early phase the linear double
Figure 1 Model of the DNA replication cycle of bacteriophage . The phage particle attaches to the host E. coli cell and injects its linear, duplex DNA molecule. (1) the DNA molecule circularizes by the base-pairing of complementary single-stranded ends. The resulting nicked circle is closed by DNA ligase. (2) DNA gyrase introduces supercoils. (3) Transcriptional activation is required for the proper assembly of the ori O P-DnaB preprimosomal structure. (4) The DnaK/DnaJ/GrpE chaperone machine, in an ATP-dependent reaction, releases a fraction of the P protein, thus triggering the unidirectional unwinding of double stranded DNA, catalyzed by the DnaB helicase. Transcription, when transversing the ori sequence dissociates the O-some from the DNA complex, thus allowing DnaB helicase to unwind double-stranded DNA in both directions. (5) In
Role of chaperones in replication of bacteriophage lambda DNA
327
the presence of single-stranded-DNA binding protein (Ssb) and DNA gyrase, the DnaB-dependent unwinding of double stranded DNA proceeds, thus allowing the DnaG primase to synthesize RNA primers, which are subsequently extended into DNA by the DNA polymerase III. At early time after infection replication of most DNA molecules proceeds bidirectionally to generate structure—early replication (theta mode). The rolling-circle DNA replication starts with (4') the blocking of bidirectional unwinding of the double-stranded DNA at the ari sequence by transient inhibition of transcription (and/ or ClpX chaperone activity). The transcriptional activation could be modulated in the presence of Cro repressor and E. coli DnaA (see text for details). (5') In the presence of primase, Ssb, DNA gyrase and DNA polymerase III unidirectional DNA replication proceeds. (61) The collision of a DnaB helicase/DNA polymerase III replication fork complex with an undissociated O-some structure, following the completion of a round of unidirectional DNA replication may be the key event that triggers the initiation of a rolling-circle sigma mode of DNA replication (see text for more details). The concatameric DNA produced by rollingcircle replication can be cut at the cos sides and packaged into phage heads. The addition of tails completes the maturation of phage particles which after cell lysis are ready for a new cycle of infection. The exact stoichiometry of the proteins and the protein subunits are not shown.
stranded bacteriophage DNA is injected into the bacterial cell, where it is rapidly circularized and supercoiled (Figure 1). Then DNA replication is initiated at a single site, ori , and proceeds bidirectionally according to the theta ( ) mode. In the late phase, replication of DNA proceeds by a rolling-circle mechanism (a mode). During this process multiple-length double stranded DNA molecules are generated and then subsequently encapsidated (Figure 1). Only two phage proteins, encoded by genes O and P, appear to participate directly in the initiation and/or propagation of the replication forks. A small fragment of the λ genome, carrying only cro, O, P, the replication origin ori (located inside the O gene) and the promoter pR can replicate autonomously as a plasmid called
dv. The pR promoter is required not only for
expression of O and P, but also for the transcriptional activation of the ori sequence, an event which regulates the frequency of DNA replication (for review see Learn et al., 1993; Taylor and Wegrzyn, 1995). It is believed that the early theta mode of bacteriophage DNA replication is mimicked by plasmid dv DNA (called also plasmid) replication. Most of the host genes required for DNA replication have been identified by challenging the phage to replicate at nonpermissive temperatures in E. coli mutants which are temperature-sensitive for replication of the bacterial chromosome (e.g. dnaB, dnaE, dnaG, dnaZ). The involvement of three other bacterial genes, namely dnaK, dnaJ and grpE were identified using a different approach (Georgopoulos and
Molecular chaperones and folding catalysts
328
Herskowitz, 1971; Saito and Uchida, 1977). These genes were originally discovered because a mutation in any of them blocked DNA replication at both permissive and nonpermissive temperatures. Moreover, mutations in the P gene ( / called also ) enable bacteriophage to replicate on hosts carrying these mutations. Subsequent studies established that any mutation (s) in these heat shock genes exerted global, pleiotropic effect on the host’s metabolism, including defects in DNA and RNA synthesis, proteolysis, cell division and overproduction of heat shock proteins (for review see Georgopoulos et al., 1990; Burkholder and Gottesman, this volume). 2. EARLY
DNA REPLICATION
The in vitro reconstitution of the dv plasmid DNA replication system using purified proteins has allowed the identification of intermediate reactions leading to the initiation of DNA replication (Mensa-Wilmot et al., 1989a; Alfano and McMacken, 1989a; Zylicz et al., 1989; Zylicz, 1993). The order of the events in the pathway of replication complex assembly reconstituted in vitro has recently been confirmed in vivo (Wegrzyn et al., 1995a). 2.1. Assembly of the O-some Structure Four dimers of the O initiation protein bind to the ori sequence at four repeating sequences (iterons) forming a large nucleosome-like structure called the O-some (Tsurimoto and Matsubara, 1981; Roberts and McMacken, 1983; Dodson et al., 1985; Liberek et al., 1988; Echols, 1990). Formation of the O-some complex changes the conformation of the ori sequence (Schnos et al., 1988; McMacken, unpublished results). On linear or relaxed DNA templates these topological changes are limited to iteron sequences. However, when in the presence of Mg2+ the O-some structure is formed on negatively supercoiled DNA, the alteration of DNA topology is transferred to the neighbouring 40 bp A/T-rich segment of ori . In this case the A/T-rich region does not become completely unwound, however it is significantly destabilized (McMacken, unpublished results). Such effects help probably to potentiate loading of DnaB helicase onto DNA within the onA sequence (see discussion below). The O protein, like several other DNA-binding replication proteins, has a tendency to aggregate (Wawrzynow et al., 1995a). The aggregated form of O protein is inert in DNA replication. It was shown in vitro that several bacterial molecular chaperones (ClpX, ClpA or DnaK/DnaJ/GrpE) can protect the O protein from aggregation and also dissociate the previously aggregated O (Wawrzynow et al., 1995a; Wawrzynow and Zylicz, unpublished results). Among them, the function of ClpX was studied in more detail. ClpX, the ATP-dependent substrate specificity component of ClpXP protease (Wojtkowiak et al., 1993; Gottesman et al., 1993), in the absence of the ClpP proteolytic subunit, possesses all the properties expected for a molecular chaperone (Wawrzynow et al., 1995a). The ClpX-dependent dissociation of O aggregates requires ATP hydrolysis catalyzed by ClpX ATPase. The chaperone effect of ClpX enhances the specific binding of O to ori , stimulating DNA replication in vitro. The biological role of the ClpX-
Role of chaperones in replication of bacteriophage lambda DNA
329
dependent enhancement of the ori O complex formation is still not clear (Wawrzynow et al., 1996). We cannot exclude the possibility that the chaperone function could be required very early, before O protein synthesis is finished. Assuming that the ClpX ATPase can perform a chaperone function in promoting the de novo folding of proteins or repairing active protein structure and in the same time is involved in protein degradation, it is important to determine the regulatory mechanism responsible for the decision to either repair or destroy a protein substrate. A hypothetical model, based on the stability of ClpX-protein substrate complex has been recently proposed to address this problem (Wawrzynow et al., 1996). In vivo, O protein is extremely unstable (half life at 40°C is 1.5 min; Lipinska et al., 1980; Gottesman et al., 1981; Bejarano et al., 1993). The ATP-dependent serine protease ClpXP was identified as being involved both in vivo and in vitro in efficient O degradation (Wojtkowiak et al., 1993; Gottesman et al., 1993). In any purified replication system (s) lacking the RNAP, the binding of O to ori sequence protects the O protein from ClpXP-dependent degradation (Wawrzynow and Zylicz, unpublished results). By contrast, in a crude in vitro replication system based on bacterial protein extract (supplemented with purified O and P) such effects were not observed (Wojtkowiak et al., 1993). Recently, we were able to show that RNA transcription proceeding through the O-some structure, makes O accessible to the ClpXP protease (Wawrzynow and Zylicz, unpublished results). Thus in vitro the stability of the O-some structure depends both on transcription and the activity of ClpXP protease. Genetic experiments have shown that mutations in the clpX gene abolish O degradation, suggesting that other Clp ATPases do not substitute for ClpX in this reaction (Gottesman et al., 1993; Szalewska et al., 1994). However, these findings may not extrapolate to the ClpX chaperone function, and any interpretation of in vivo results obtained with ClpX mutants ought to take into account the possibility that ClpX chaperone activity may be substituted by other chaperones. In particular, it is possible that a close ClpX homologue, ClpY, that has recently been identified in E. coli (Missiakas et al., 1996), can substitute for ClpX. 2.2. Assembly of the Preprimosomal Structure A second phage protein involved in DNA replication is P which is responsible for sequestering a bacterial helicase from the host replication system making it available to the ori sequence. In bacterial cells the DnaB helicase (LeBowitz and McMacken, 1986) is responsible for the propagation of the replication fork during synthesis of the bacterial chromosome (Kaguni and Kornberg, 1984). The DnaB is complexed with the DnaC initiation factor which results in attracting the helicase to the chromosome origin sequence (Marszalek and Kaguni, 1994). After phage infection, when the intracellular level of P initiation protein increases, P protein wins the competition with DnaC for DnaB binding (Mallory et al., 1990) and the P-DnaB complex interacts with the Osome structure to form an ori O P-DnaB preprimosomal complex (Figure 2) (Dodson et al., 1985; Liberek et al., 1988; Alfano and McMacken, 1989a; Zylicz et al., 1989). The stoichiometry of the ori - O P-DnaB preprimosomal complex has not yet been resolved. Probably two P-DnaB complexes (each composed of a multimer of P
Molecular chaperones and folding catalysts
330
and a hexamer of DnaB; Mallory et al., 1990) are positioned in the A/T-rich region, on the right side of the O-some. The asymmetry of the preprimosomal complex was suggested on the basis of the electron-microscopy studies (Dodson et al., 1985). The observed stability of ori O- P-DnaB complex assembled on the supercoiled DNA could be due to both specific P O protein-protein interactions (Tomizawa, 1971; Zylicz et al., 1984; Wickner and Zahn, 1986; Liberek et al., 1990) and the affinity of their protein components to single stranded DNA. In support of this it is not possible to isolate the stable O- P-DnaB complex in the absence of the DNA template (Zylicz, unpublished results). Futhermore, using UV cross-linking experiments it was possible to demonstrate that both O and P, but not DnaB, are in close contact with the ssDNA (Learn et al., 1997). Proper assembly of an ori O P-DnaB preprimosome has an absolute requirement for a negatively supercoiled DNA template (McMacken, unpublished results). Following the formation of the O-some structure, the DNA replication potential is rapidly lost upon complete removal of free negative DNA superhelical turns. By contrast, the ori O P-DnaB complex once preformed on the negatively supercoiled DNA retains some replication potential even several minutes after removal of free DNA supercoils (t1/2 30– 40 sec on linearized DNA, t1/2 3–4 min on HU coated DNA; McMacken, unpublished results). These results suggest that during DNA replication supercoiling is absolutely required only for proper formation of the ori O P-DnaB. McMacken and coworkers suggest that properly assembled ori - O P-DnaB nucleoprotein structure traps an “open” DNA conformation that is
Role of chaperones in replication of bacteriophage lambda DNA
331
Figure 2 Chaperone-dependent activation of the DnaB helicase in the preprimosomal complex. The DnaK/DnaJ/GrpE-dependent activation of preprimosomal complex (reaction 4; Figure 1) can be described by the following steps: (4a) DnaJ chaperone binds to P and DnaB thus attracting (4b) the DnaK*-ADP chaperone to the preprimosomal structure. In the absence of DnaJ, DnaK in the DnaK*-ADP conformation does not form a stable complex with P. The presence of DnaJ changes the conformation of DnaK*-ADP in such mode that more stable preprimosomal-DnaK-ADP complex is formed. (4c) Following the GrpE-dependent ADP/ATP exchange and ATP hydrolysis, a portion of P is released from the preprimosomal complex and DnaK is converted back to the DnaK*-ADP conformation which possesses limited affinity to P but recognizes efficiently the DnaJ protein. Those molecules of DnaB helicase
Molecular chaperones and folding catalysts
332
which are no longer associated with P proceed to unwind double-stranded DNA rightwards from the ori sequence. (4d) Transcriptional activation displaces or rearranges the O-some structure allowing the DnaB helicase to (4e) unwind DNA leftwards. The ori
iterons which are recognition sites for O are shown as black boxes. The (*) symbol represent the DnaK*-ADP conformation reached only after ATP hydrolysis, and not as a result of the simple preincubation of DnaK with ADP. The stoichiometry of the proteins and the protein subunits are not represented.
induced in the presence of Mg2+ at ori A/T-rich region upon the O-some formation on supercoiled DNA (McMacken, unpublished results). According to this interpretation DnaB helicase is not positioned between the partially unwound DNA strands. DnaK and DnaJ action is needed before DnaB can stably interact with DNA (Learn et al., 1997). Other researchers also speculate that the DnaK/DnaJ/GrpE chaperone machine could be involved in loading DnaB helicase between two partially unwound strands (Wegrzyn et al., 1996b). We suggest that transcriptional activation is required for the “rearrangement” of the protein-DNA complex at ori , localizing the DnaB helicase at the junction of double stranded and single stranded DNA. This hypothesis is supported by the fact that during the RNA polymerase action at least O, P as well as DnaB helicase must be present to initiate the ori -specific DNA replication (Yamamoto et al., 1987). Assembly of the active ori - O- P-DnaB preprimosome on relaxed or coated (by the histone-like protein, HU) DNA needs transcriptional activation (Mensa-Wilmot et al., 1989b). Only when the purified DNA replication system is devoid of HU protein and the DNA template is highly negatively supercoiled (- >0.045) is transcriptional activation not required for efficient DNA synthesis (Alfano and McMacken, 1989a; Zylicz et al., 1989). McMacken and co-workers suggested that transcription activates the DNA replication by counter-acting an inhibitory effect of HU protein, which constrains negatively supercoiled DNA upon binding. For example, an increase of the negative supercoiling behind the RNA polymerase during transcription influences the topology of the A/T-rich ori region in such a way that it helps O and P to trap partially unwound DNA strands. The idea that transcriptional activation of λ replication does not simply rely on dissociation of the HU protein from ori sequence thus allowing formation of the ori - O- P-DnaB complex is supported by the fact that it can emanate from promoters located either upstream or downstream from ori (Furth et al., 1982; McMacken, unpublished results). In the case of the properly assembled ori - O- P-DnaB preprimosome, transcribing RNA polymerase molecules which readily “bypass” the O-some structure, pause indefinitely at ori - O- P-DnaB complex (McMacken, unpublished results). Thus, the O protein remains in the preprimosome structure and is inaccessible to the ClpXP protease (Wawrzynow and Zylicz, unpublished results). Recent genetic studies support these in vitro findings (Wegrzyn et al., 1995a).
Role of chaperones in replication of bacteriophage lambda DNA
333
2.3. Activation of the Preprimosomal Structure The preprimosomal ori - O- P-DnaB complex assembled on supercoiled DNA is stable, since it can be isolated by gel filtration and subsequently demonstrated to be still active in DNA replication without addition of further O, P or DnaB (Alfano and McMacken, 1988;1989b; Zylicz et al., 1989; Zylicz, 1993). The strong P-DnaB interaction required for assembly of DnaB helicase at ori also inhibits the ATPase and helicase activity of the DnaB (Wickner, 1978; Mallory et al., 1990). Subsequent activation reactions are needed to liberate the DnaB helicase from the inhibitory effects of P. The concerted action of DnaK, DnaJ and GrpE is respon-sible for the “liberation” (Figure 2) (Liberek et al., 1988; Alfano and McMacken, 1989b; Dodson et al., 1989; Zylicz et al., 1989; Hoffmann et al., 1992; Zylicz, 1993). The DnaJ chaperone binds to the preprimosome and further stabilizes its structure, by interacting primarily with P and DnaB (Alfano and McMacken, 1989b; Hoffman et al., 1992; Osipiuk et al., 1993; Zylicz, 1993; Wawrzynow and Zylicz, 1995). The presence of DnaJ bound to P facilitates the binding of the DnaK chaperone to the preprimosomal complex (Wawrzynow et al., 1995b). The binding of DnaK and DnaJ to the preprimosome was shown by using electron microscopy of gold-tagged antibodies against these proteins (Dodson et al., 1989). DnaK has a weak intrinsic ATPase activity (Zylicz et al., 1983; Buchberger et al., this volume). Depending on whether it is in ATP or ADP bound form DnaK possesses a different conformation and protein substrate affinity (Palleros et al., 1992; Banecki et al., 1992; Wawrzynow and Zylicz, 1995; Banecki and Zylicz, 1996). It was suggested that DnaK in its DnaK-ATP form binds to the DnaJ-modified substrate and after ATP hydrolysis, accelerated by DnaJ, a DnaJ-substrate-DnaK-ADP complex is formed (Szabo et al., 1994; McCarty et al., 1995; Gamer et al., 1996; Buchberger et al., this volume). Recent real time kinetics data have suggested a different scenario (Banecki and Zylicz, 1996; Banecki et al., 1996). Using stopped-flow analysis of the fluorescence of a single tryptophan located near the ATP binding site of DnaK, we showed that, upon ATP hydrolysis, DnaK is converted into a so-called DnaK*-ADP conformation, which can interact only transiently with protein substrate. Such a conformation cannot be obtained by simple incubation of DnaK in the presence of ADP. The presence of DnaJ induces other conformational changes of DnaK*-ADP, leading to the formation of a stable substrate-DnaK-ADP complex (Figure 2) (Wawrzynow et al., 1995b; Banecki and Zylicz, 1996). The formation of such a complex is a rate limiting step and depends on the affinity of DnaJ to the particular protein substrate. If the affinity of DnaJ to protein substrate is high a stable DnaJ-(substrate-DnaK-ADP) complex is formed (Wawrzynow et al., 1995b). In summary, DnaJ protein first “targets” DnaK to the preprimosomal structure and subsequently changes DnaK’s conformation to stabilize the P-DnaK-ADP complex. It is important to stress that, according to our experimental results, DnaJ-dependent conformational changes in DnaK occur after ATP hydrolysis when DnaK is already in the DnaK*-ADP form (Banecki and Zylicz, 1996; Banecki et al., 1996; Buchberger et al., this volume, for alternative views). By way of kinetic analysis, it has been shown that
Molecular chaperones and folding catalysts
334
conformational changes of monomeric DnaK, the form required for DnaJ binding, correlates well with ATP hydrolysis (Banecki and Zylicz, 1996; Banecki, unpublished results) but not with ATP binding as was suggested by Theyssen et al. (1996) and Pierpaoli et al. (1997). Discrepancies in the interpretation of the results could have derived from procedures used by other laboratories, which inadvertently stripped bound nucleotites from DnaK protein while the same time inducing its oligomerization (Banecki, unpublished results). The presence of GrpE and ATP is required for recycling DnaK from the substrateDnaK-ADP complex (Zylicz et al., 1989, Alfano and McMacken, 1989b; Banecki and Zylicz, 1996; Wawrzynow and Zylicz, unpublished results). In the presence of ATP, GrpE accelerates the exchange of ADP for ATP complexed with DnaK, thus triggering the next round of ATP hydrolysis catalysed by DnaK (Figure 2) (Liberek et al., 1991; Banecki and Zylicz, 1996). The ATP hydrolysis changes the conformation of DnaK back to the DnaK*-ADP form which possesses a low affinity for protein substrate (Banecki and Zylicz, 1996). Therefore DnaK dissociates from the preprimosomal complex. It had been postulated previously that binding of ATP, but not ATP hydrolysis, was sufficient for the release of DnaK from protein substrate (Palleros et al., 1993; Schmid et al., 1994; Szabo et al., 1994; McCarty et al., 1995). We suggest that GrpE-dependent exchange of ADP to ATP complexed with DnaK could trigger the release of protein substrate from substrate-DnaK complex but could not recycle the DnaK. Only ATP hydrolysis converts DnaK to DnaK*-ADP conformation which can rebind the DnaJ modified substrate (Figure 2) (Banecki and Zylicz, 1996; Buchberger et al., this volume). Induction of conformational changes of DnaK triggered by ATP hydrolysis may also induce conformational changes in P, resulting in dissociation of P from the preprimosomal complex. The net result of this reaction cycle is the selective release of DnaK and P from the preprimosomal complex (Figure 2) (Liberek et al., 1988; Alfano and McMacken, 1989b; Dodson et al., 1989; Zylicz et al., 1989; Hoffman et al., 1992; Zylicz, 1993). In contrast, most of the O present in the O-some structure does not dissociate from the DNA template during the DnaK/DnaJ/GrpE chaperone dependent activation of the preprimosome (Dodson et al., 1986, 1989; Liberek et al., 1988; Wojtkowiak et al., 1993). The release of P from the DnaB complex is the crucial event leading to the initiation of DNA replication. A mutant of P, called P or , binds so weakly to the DnaB helicase that the requirement of molecular chaperones is significantly reduced (Konieczny and Marszalek, 1995). This observation explains the previous classical genetic experiments showing that DNA replication could proceed in the dnaK, dnaJ and grpE mutant backgrounds, provided that the P gene was mutated to P (Georgopoulos and Herskowitz, 1971; Saito and Uchida, 1977). Konieczny and Marszalek proposed that the phenotype of the P mutation could shed a new light on the evolution of bacteriophage DNA replication and the role of heat shock proteins in this process. According to this hypothesis the ancestor of modern bacteriophage possessed a P protein which resembles P such that heat shock proteins were not required for activation of the preprimosomal complex. The selective pressure for fast growth and development of A, phage produced modern P protein, which was more efficient in sequestering DnaB helicase. The side effect of this adaptation is the stability
Role of chaperones in replication of bacteriophage lambda DNA
335
of preprimosome complex and the need of heat shock proteins for activation of the preprimosomal complex (Konieczny and Marszalek, 1995). The release of P from the complex with DnaB helicase leads to the unwinding of DNA near the ori sequence (Figure 2) (Dodson et al., 1986, 1989; Learn et al., 1993). Efficient unwinding of the supercoiled DNA template requires the additional presence of the Ssb protein. Extensive unwinding by DnaB creates positively supercoiled parental DNA, and DNA gyrase action is needed to counteract this effect (Learn et al., 1993). The DnaG primase recognizes the proper DnaB-ssDNA complex and proceeds to synthesize RNA primers at apparently random sites. The DNA polymerase III holoenzyme extends the RNA primers into the DNA. 2.4. Uni- and Bidirectional λ DNA Replication As discussed above, the preprimosomal complex is probably asymmetric, with the Osome structure located within the iteron sequences and with one or more P-DnaB complexes bound to the A/T-rich region located on the right side of theori sequence (Figure 2). Consequently the activated DnaB hexamer can freely unwind DNA duplex rightwards from theori sequence. However, the assembled O-some structure could cause a physical barrier for the passage of DnaB helicase through the O-some complex (Dodson et al., 1986). Supporting this hypothesis, in the in vitro DNA replication system reconstituted with purified proteins, replication is mostly unidirectional with the replication fork moving rightwards from the ori sequence (Figure 2) (Learn et al., 1993). However, addition of purified RNA polymerase together with rNTPs, switches a percentage (up to 30%) of the unidirectionally replicating molecules to a bidirectional DNA replication mode (Learn et al., 1993). McMacken and colleagues postulated that the role of transcriptional activation could be to displace or rearrange the O-some structure, thus allowing the DnaB helicase to unwind DNA to the leftward (Learn et al., 1993). We propose that the presence of the ClpXP protease (or the ClpX chaperone alone) may support RNA polymerase action leading to a more efficient transition from uni- to bidirectional DNA replication (Wojtkowiak et al., 1993). For example, in the presence of ClpXP the rearrangement of O protein into another O-some structure will almost certainly be inhibited, since O will be efficiently degraded upon transcription-mediated disruption of the O-some (Figure 2). Another factor which can influence the directionality of DNA replication is the efficiency of activation of the preprimosomal complex. For example, the release of P from the preprimosomal complex may not be complete (Dodson et al., 1989; Zylicz, 1993). About 50 to 60% of P indeed remain bound to DNA after action of the DnaK/DnaJ/GrpE chaperones. It is unclear whether this is a factor that limits the efficiency of DNA replication. Taylor and Wegrzyn suggest that DnaK/DnaJ/ GrpE chaperone machine only rearranges the preprimosomal complex but does not partially disassemble it (Taylor and Wegrzyn, 1995). According to this hypothesis the DnaB helicase which unwinds DNA may still be complexed with O, P and probably other replication proteins, forming a mobile replication complex which could be inherited, and used by one of two daughter DNA molecules during the second round of DNA replication (Wegrzyn and Taylor, 1992; Wegrzyn and Taylor, unpublished results). This
Molecular chaperones and folding catalysts
336
hypothesis is based on the fact that in vivo a small fraction of O is not subjected to rapid degradation. The authors of these experiments assume that the stable fraction represents the O present in the replication complex (Wegrzyn et al., 1992). Interestingly, such a stable O fraction disappears after the heat shock induction of the groE operon coding for the GroEL/ES chaperone machine (Wegrzyn et al., 1996a). It is unclear how this chaperone machine affects the halflife of O. We postulate a different interpretation, according to which only the hexamers of DnaB helicase which have dissociated from P are involved in the unwinding of DNA. Release of P from only one of the helicase complexes (specifically distant from the Osome) might result in unidirectional DNA replication (Figure 2). Consequently the DnaB-λP complex proximal to the O-some structure might remain intact on the DNA template and may be inherited by one of two daughter strands. Another possibility is that the inherited replication complex represents a fraction of preprimosomal complexes which has become “frozen” at the ori O- P-DnaB stage. Supporting this interpretation, in the presence of transcription, the O protein bound in the ori O P-DnaB complex is resistant to ClpXP protease (Wawrzynow and Zylicz, unpublished results). 3. SWITCH FROM EARLY TO LATE
DNA REPLICATION
McMacken and colleagues have proposed that unidirectional DNA replication, after one round, could lead to strand separation, resulting in a switch from the early ( ) to the late ( ) mode of DNA replication (Figure 1) (Dodson et al., 1986; Learn et al., 1993). According to this hypothesis, the switch from the to the mode of A DNA replication could be correlated with the inhibition of early transcription (Learn et al., 1993). We propose that the ClpX molecular chaperone could also modulate this reaction. For example, the transient inhibition of ClpX activity could protect the intact O-some structure, thus favouring unidirectional DNA replication, and as a consequence initiates the switch to the rolling circle replication mode. The hypothesis that modulation of O degradation could regulate DNA replication is supported by the fact that the product of the rexB gene, the expression of which is increasing after initiation of DNA replication, prevents the O degradation (Schoulaker-Schwarz et al., 1991). The possibility that rex is involved in the switching from early to late phage DNA replication was suggested by Toothman and Herskowitz (1980). Under stress conditions, like heat shock, involvement of ClpX in other metabolic processes (e.g. proteolysis of host proteins) could also favour unidirectional DNA replication and, consequently, the switch to rolling circle replication. Similarly, mutation in the clpX gene might suppress the early bidirectional mode of DNA replication, therefore after the first round of unidirectional theta replication the switch to rolling circle mode might occur. This hypothesis could explain why mutation in clpX gene only slightly influences plaque formation (Szalewska et al., 1994). Recent findings suggesting that the elimination of dnaA function decreases transcriptional activation of DNA replication (Wegrzyn et al., 1995b) have led to the hypothesis that DnaA replication protein, a key initiation factor required for host DNA
Role of chaperones in replication of bacteriophage lambda DNA
337
replication is also involved in DNA replication. In wild-type cells the utilization of DnaA by rapidly replicating phage DNA may inhibit transcriptional activation and thus favour the switch from bidirectional theta to the unidirectional theta replication mode and later to the sigma mode (Wegrzyn et al., 1995c). 4. CONCLUDING REMARKS Molecular chaperones are involved in many biochemical pathways, thus it is not surprising that mutations in these genes have pleiotropic metabolic effects that often compromise DNA replication of the bacterial chromosome and other replicons which are propagated in this bacterium. Detailed biochemical analysis, using in vitro DNA replication systems with purified proteins, indicates that molecular chaperones could be involved in DNA replication at least at three different stages. First, molecular chaperones might function during folding of newly synthesized replication proteins, protecting them from aggregation. Several replication proteins like O, DnaA, protein subunits of DNA polymerase III holoenzyme or RNA polymerase, possess strong tendencies to aggregate. The presence of molecular chaperones like DnaK, DnaJ, ClpX, ClpA, or GroEL could prevent formation of the inactive aggregates and stabilizes the active form of replication proteins (Skowyra et al., 1990; Hwang et al., 1990; Ziemienowicz et al., 1993, 1995; Blaszczak et al., 1995). Another example of this general chaperone mechanism is the ability of molecular chaperones to activate mutant replication proteins (Fayet et al., 1986; Hupp and Kaguni, 1993). It is important to note that several molecular chaperones could substitute for each other in protection reactions. This is probably one of the reasons why in many cases mutation in one of the genes that encode a molecular chaperone does not result in a complete block of the DNA replication process. The second stage at which molecular chaperones could work is the activation of already synthesized replication proteins. Both in vivo and in vitro initiation of plasmid P1 DNA replication is dependent on the functional presence of the DnaK/ DnaJ/GrpE molecular chaperones (Bukau and Walker, 1989; Tilly andYarmolinsky, 1989; Wickner et al., 1992). It was proposed that these chaperones dissociate RepA dimers in an ATPdependent reaction, thus producing RepA monomers that are active to bind to oriP1 sequence (Wickner et al., 1991). Results by DasGupta et al. suggest that monomerization of RepA may not be the only mechanism responsible for activation of RepA protein since dilute solutions of RepA, where the protein is essentially in monomeric form, still require chaperone function for activation (DasGupta et al., 1993). In the case of P1 DNA replication, bacterial molecular chaperones act before nucleoprotein complex formation (RepA-P1 DNA), and the presence of molecular chaperones is not obligatory at any of the subsequent steps leading to the initiation of P1 DNA replication (Wickner et al., 1992). The chaperone-dependent monomerization step required for the activation of initiation proteins for binding to the origin sequences was postulated also for F and RK2 plasmid DNA replication (Kawasaki et al., 1990; Konieczny and Helinski, 1997). In the latter case the ClpX chaperone can dissociate dimers of the TrfA initiation protein in an ATP-dependent reaction thereby allowing efficient TrfA binding to the oriRK2 sequence (Konieczny and Helinski, 1997).
Molecular chaperones and folding catalysts
338
The third stage at which molecular chaperones could act is the specific initiation process. Initiation of DNA replication is not the only example of such a reaction. Recently it has been shown that ClpX is required for the lytic growth of bacteriophage Mu. The kinetics of the block in bacteriophage growth after induction of a clpX-deleted lysogen suggest that ClpX functions during the initiation of Mu DNA replication (Mhammedi-Alaoui et al., 1994). Bacteriophage Mu employs a transposition mechanism to replicate its DNA that involves the activity of the MuA transposase (Nakai and Kruklitis, 1995). After MuA-dependent catalysis of the strand transfer reaction (ends of the Mu DNA are cleaved and joined to the target DNA) a stable MuA-DNA complex is formed. Initiation of Mu DNA replication is triggered by release of MuA from such complex (Nakai and Kruklitis, 1995). It was shown in vitro that ClpX is required for the ATP-dependent release of MuA from such stable DNA replication intermediate complex (Levchenko et al., 1995; Kruklitis et al., 1996). The molecular mechanism of bacteriophage Mu DNA replication requires the high specificity of chaperone action. ClpX has to interact selectively with the protein substrate (MuA) bound within the replication complex and differentiate it from the other proteins present in this nucleoprotein structure. The analogy with the activation reaction of bacteriophage preprimosomal complex is obvious. The established functions of chaperones in and Mu DNA replication may be unique adaptations of parasites, allowing them to use the host’s DNA replication machinery. The involvement of molecular chaperones in the folding and activation of replication proteins before initiation of DNA replication, however, may represent a more general mechanism of the action of molecular chaperones in DNA replication. 5. ACKNOWLEDGMENTS We are grateful to R.McMacken, K.Taylor, G.Wegrzyn, D.Helinski, D.Missiakas and S.Raina for sharing with us unpublished results. We thank R.McMacken for discussions, critical comments and reading of this manuscript. We thank B.Ashraf for correcting this manuscript. This work was supported by grant 6PO4AO1712 from the Polish State Committee for Scientific Research. 6. REFERENCES Alfano, C. and McMacken, R. (1989a). Ordered assembly of nucleoprotein structures at the bacteriophage replication origin during the initiation of DNA replication. J. Biol. Chem. , 264 , 10699–10708. Alfano, C. and McMacken, R. (1989b). Heat shock protein-mediated disassembly of nucleoprotein structures is required for initiation of bacteriophage DNA replication. J. Biol. Chem. , 264 , 10709–10718. Alfano, C. and McMacken, R. (1988). The role of template superhelicity in the initiation of bacteriophage DNA replication. Nucl. Acids Res. , 16 , 9611–9630. Banecki, B. and Zylicz, M. (1996). Real time kinetics of the DnaK/DnaJ/GrpE molecular chaperone machine action. J. Biol. Chem. , 271 , 6137–6143.
Role of chaperones in replication of bacteriophage lambda DNA
339
Banecki, B., Liberek, K., Wall, D., Wawrzynow, A., Georgopoulos, C., Bertoli, E., Tanfani, F. and Zylicz, M. (1996). Structure-functional analysis of the zinc finger region of the DnaJ molecular chaperone. J. Biol. Chem. , 271 , 6137–6143. Banecki, B., Zylicz, M., Bertoli, E. and Tanfani, F. (1992). Structural and functional relationships in DnaK and DnaK756 heat-shock proteins from Escherichia coli. J. Biol. Chem. , 267 , 25051–25058. Bejarano, I. Klemes, Y., Schoulaker-Schwarz, R. and Engelberg-Kulka, H. (1993). Energy-dependent degradation of O protein in Escherichia coli. J. Bacterial. , 175 , 7720–7723. Blaszczak, A., Zylicz, M., Georgopoulos, C. and Liberek, K. (1995). Both ambient temperature and the DnaK chaperone machine modulate the heat shock response in Escherichia coli by regulating the switch between and factors assembled with RNA polymerase. EMBO J. , 14 , 5085–5093. Bukau, B. and Walker, G.C. (1989). ∆dnaK52 mutants of Escherichia coli have defects in chromosome segregation and plasmid maintenance at normal growth temperatures. J. Bacteriol. , 171 , 6030–6038. DasGupta, S., Mukhopadhyay, G., Papp, P.P., Lewis, M.S. and Chattoraj, O.K. (1993). Activation of DNA binding by the monomeric form of P1 replication initiator RepA by heat shock proteins DnaJ and DnaK. J. Mol. Bioi , 232 , 23–34. Dodson, M., Echols, H., Wickner, S., Alfano, R., Mensa-Wilmot, K, Gomes, B., LeBowitz, J.H., Roberts, J.D. and McMacken, R. (1986). Specialized nucleoprotein structures at the origin of replication of bacteriophage : Localized unwinding of duplex DNA by six-protein reaction. Proc. Natl. Acad. Set. USA , 83 , 7638–7642. Dodson, M., Roberts, J.D., McMacken, R. and Echols, H. (1985). Specialized nucleoprotein structures at the origin of the replication of bacteriophage : Complexes wih O, P and Escherichia coli DnaB proteins. Proc. Natl. Acad. Sci. USA , 82 , 4678–4682. Dodson, M., McMacken, R. and Echols, H. (1989). Specialized nucleoprotein structures at the origin of replication of bacteriophage : Protein association and disassociation reactions responsible for localized initiation of replication. J. Biol. Chem. , 264 , 10719–10725. Echols, H. (1990). Nucleoprotein structures initiating DNA replication, transcription, and site-specific recombination. J. Biol. Chem. , 265 , 14697–14700. Fayet, O.T., Louarn, J.-M., Georgopoulos, C. (1986). Suppression of the Escherichia coli dnaA46 mutation by amplification of the groES and groEL genes. Mol. Gen. Genet. , 202 , 435–445. Furth, M.E. and Wickner, S. (1983). Lambda DNA replication. In R.W.Hendrix, J.W., Roberts, F.W. Stahl and R.A.Weisberg (eds.), Lambda II , Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y., pp. 145–173. Furth, M.E., Dove, W.F., Meyer, B.J. (1982). Specificity determination for bacteriophage DNA replication. III. Activation of replication in ric mutants by transcription outside of on. J. Mol. Biol. , 154 , 65–80 Gamer, J., Multhaup, G., Tomoyasu, T., McCarty, J.S., Rudigre, S., Schonefeld, H.J., Schirra, C., Bujard, H., Bukau, B. (1996). A cycle of binding and release of the DnaK, DnaJ and GrpE chaperones regulates activity of the Escherichia coli heat shock
Molecular chaperones and folding catalysts
340
transcription factor . EMBO J. , 15 , 607–617. Georgopoulos, C., Liberek, K., Zylicz, M. and Ang, D. (1994). Properties of the heat shock proteins of Escherichia coli and the autoregulation of the heat shock response. In R.I., Morimoto, A., Tissieres, and C.Georgopoulos, (eds), The biology of the heat shock proteins and molecular chaperones . Cold Spring Harbor Laboratory Press , pp. 209–249. Georgopoulos, C., Ang. D., Liberek, K. and Zylicz, M. (1990). Properties of the Escherichia coli heat shock proteins and their role in bacteriophage growth. In R.I., Morimoto, A., Tissieres, and C.Georgopoulos, (eds), Stress protein in biology and medicine , Cold Spring Harbor Laboratory Press, pp. 191–221. Georgopoulos, C. and Herskowitz, I. (1971). Escherichia coli mutants blocked in A, DNA synthesis. In A.D. Hershey, (eds), The bacteriophage lambda , Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, pp. 553–560. Gottesman, S., Clark, W.P., de Crecy-Lagard, V. and Maurizi, M.R. (1993). ClpX, an alternative subunit for the ATP-dependent Clp protease of Escherichia coli. J. Biol. Chem. , 268 , 22618–22626. Gottesman, S., Gottesman, M., Show, J. and Pearson, M.L. (1981). Protein degradation in E. coli: the Ion mutation and bacteriophage N and cII protein stability. Cell , 24 , 225–233. Hartl, F.-U. (1996). Molecular chaperones in cellular protein folding. Nature , 381 , 571– 580. Hendrick, J.P. and Hartl, F.-U. (1993). Molecular chaperone functions of heat shock proteins. Annu. Rev. Biochem. , 62 , 349–384. Hoffmann, H.J., Lyman, S.K., Lu, C., Petit, M.A., Echols, H. (1992). Activity of Hsp70 chaperone complex-DnaK, DnaJ and GrpE -in initiating phage A.DNA replication by sequestering and releasing P protein. Proc. Natl. Acad. Sci. USA , 89 , 12108–12111. Hupp, T.R. and Kaguni, J.M. (1993). Activation of mutant forms of DnaA protein of Escherichia coli by DnaK and GrpE proteins occurs prior to DNA replication. J. Biol. Chem. , 268 , 13143–13150. Hwang, D.S., Crooke, E. and Kornberg, A. (1990). Aggregated DnaA protein is dissociated and activated for DNA replication by phospholipase or DnaK protein. J. Biol Chem. , 265 , 19244–19248. Jordan, R. and McMacken, R. (1995). Modulation of the ATPase activity of the molecular chaperone DnaK by peptides and the DnaJ and GrpE heat shock proteins. J. Biol. Chem. , 270 , 4563–4569. Kaguni, J.M. and Kornberg, A. (1984). Replication initiated at the origin (oriC). of the E. coli chromosome reconstituted with purified enzymes. Cell , 38 , 183–190. Kawasaki, Y., Wada, C. and Yura, T. (1990). Role of Escherichia coli heat shock proteins DnaK, DnaJ and GrpE in mini-F plasmid replication. Mol. Gen. Genet. , 220 , 277–282. Konieczny, I. and Marszalek, J. (1995). The requirement for molecular chaperones in DNA replication is reduced by the mutation in P gene which weakens the interaction between lambda P protein and DnaB helicase. J. Biol. Chem. , 270 , 9792– 9799. Konieczny, I. and Helinski, D.R. (1997). The replication initiation protein of the broad-
Role of chaperones in replication of bacteriophage lambda DNA
341
host-range plasmid RK2 is activated by the ClpX chaperone. Proc. Natl. Acad. Sci. USA , 94 , 14378–14382. Kruklitis, R., Welty, D.J. and Nakai, H. (1996). ClpX protein of Escherichia coli activates Mu transposase in the strand transfer complex for initiation of Mu DNA synthesis. EMBO J. , 15 , 953–944. Learn, B., Karzai, A.W. and McMacken, R. (1993). Transcription stimulates the establishment of bidirectional DNA replication in vitro . Cold Spring Harbor Symposia on Quantitative Biology , vol. LVIII. Cold Spring Harbor Laboratory Press, pp. 389–402. Learn, B.A., Um, S.J. and McMacken, R. (1997). Cryptic single-stranded-DNA binding activities of the phage Lambda P and Escherichia coli DnaC replication proteins facilitate the transfer of E. coli DnaB helicase onto DNA . Proc. Natl. Acad. Sci. USA , 94 , 1154–1159. LeBowitz, J.H. and McMacken, R. (1986). The Escherichia coliDnaB replication protein is a DNA helicase. J. Biol. Chem. , 261 , 4738–4748. Levchenko, I., Luo, L. and Baker, T.A. (1995). Disassembly of the Mu transposase tetramer by the ClpX chaperone. Genes Dev. , 9 , 2399–2408. Liberek, K., Marszalek, J., Ang, D., Georgopoulos, C. and Zylicz, M. (1991). Escherichia coli DnaJ and GrpE heat shock proteins jointly stimulate ATPase activity of DnaK, Proc. Natl. Acad. Sci.USA , 88 , 2874–2878. Liberek, K., Osipiuk, J., Zylicz, M., Ang, D., Skorko, J. and Georgopoulos, C. (1990). Physical interactions among bacteriophage in Escherichia coli proteins required for initiation of A DNA replication. J. Biol. Chem. , 265 , 3022–3029. Liberek, K., Georgopoulos, C. and Zylicz, M. (1988). Role of the Escherichia coli DnaK and DnaJ heat shock proteins in the initiation of bacteriophage A DNA replication. Proc. Natl. Acad. Sci. USA , 85 , 6632–6636. Lipinska, B., Podhajska, A. and Taylor, K (1980). Synthesis and decay of DNA replication proteins in minicells. Biochem. Biophys. Res. Commun. , 92 , 120–126. Mallory, J.B., Alfano, C. and McMacken, R. (1990). Host-virus interactions in the initiation of bacteriophage DNA replication. Recruitment of Escherichia coli DnaB helicase by P replication protein. J. Biol. Chem. , 265 , 13297–13307. McCarty, J.S., Buchberger, A., Reinstein, J. and Bukau, B. (1995). The role of ATP in the function cycle of the DnaK chaperone system. J. Mol. Biol. , 249 , 126–137. Marszalek, J. and Kaguni, J. (1995). DnaA protein directs the binding of DnaB protein in initiation of DNA replication in Escherichia coli. J. Biol. Chem. , 269 , 4883–4890. Mensa-Wilmot, K., Seaby, R., Alfano, C., Wold, M.S., Gomes, B. and McMacken, R. (1989a). Reconstitution of a nine-protein system that initiates bacteriophage DNA replication. J. Biol. Chem. , 264 , 2853–2861. Mensa-Wilmot, K., Carrol, K. and McMacken, R. (1989b). Transcriptional activation of bacteriophage DNA replication in vitro. Regulatory of histone-like protein HU of Escherichia coli. EMBO J. , 8 , 2393–2398. Mhammedi-Alaoui, A., Pato, M., Gama, M.-J. and Toussaint, A. (1994). A new component of bacteriophage Mu replicative transposition machinery: The Escherichia coli ClpX protein. Molecular Microbiology , 11 , 1109–1116. Missiakas, D., Schwager, F., Betton, J.-M., Georgopoulos, C. and Raina, S. (1996).
Molecular chaperones and folding catalysts
342
Identification and characterization of HslV HslU (ClpQ ClpY) proteins involved in overall proteolysis of misfolded proteins in Escherichia coli. EMBO J. , 15 , 6899– 6909. Nakai, H. and Kruklitis, R. (1995). Disassembly of the bacteriophage Mu transposase for the initiation of Mu DNA replication. J. Biol. Chem. , 270 , 19591–19598. Osipiuk, J., Georgopoulos, C., Zylicz, M. (1993). Initiation of DNA replication. J. Biol. Chem. , 268 , 4821–4827. Palleros, D.R., Reid, K.L., Shi, L., Welch, W.J., Fink, A.L. (1993). ATP- induced protein-Hsp70 complex dissociation requires K+ but not ATP hydrolysis. Nature , 365 , 664–666. Palleros, D.R., Reid, K.L., McCarty, J.S., Walker, G.C. and Fink, A.L. (1992). DnaK, hsp73, and their molten globules. J. Biol. Chem. , 267 , 5279–5285. Pierpaoli, E.V., Sandmeier, E., Baici, A., Schonfeld, H.J., Gisler, S. and Christen, P. (1997). The power stroke of the DnaK/DnaJ/GrpE molecular chaperone system. J. Mol Biol. , 269 , 757–768. Roberts, J.D. and McMacken, R. (1983). The bacteriophage O replication protein: Isolation and characterization of the amplified initiator. Nucl. Acid. Res. , 11 , 7435– 7452. Saito, H. and Uchida, H. (1977). Initiation of the DNA replication of bacteriophage in Escherichia coli K12. J. Mol. Biol , 113 , 1–25 Schoulaker-Schwarz, R., Dekel-Gorodetsky, L. and Engelberg-Kulka (1991). An additional function of bacteriophage rex: the rexB product prevents degradation of the O protein. Proc. Natl. Acad. Sci. USA , 88 , 4996–5000. Schmid, D., Baici, A., Gehring, H., Christen, P. (1994). Kinetics of molecular chaperone action. Science , 263 , 971–973. Schnos, M., Zahn, K., Inman, R.B., Blattner, F.R. (1988). Initiation protein induced helix destabilization at the origin: A prepriming step in DNA replication. Cell , 52 , 385– 395. Skowyra, D., Georgopoulos, C., Zylicz, M. (1990). The E. coli dnaK gene product, the hsp70 homolog, can reactivate heat-inactivated RNA polymerase in an ATP hydrolysis-dependent manner. Cell , 62 , 939–944. Szabo, A., Langer, T., Schroder, H., Flanagan, J., Bukau, B. and Hartl, F.U. (1994). The ATP hydrolysisdependent reaction cycle of the Escherichia coli Hsp70 system-DnaK, DnaJ, and GrpE. Proc. Natl. Acad. Sci. USA , 91 , 10345–10349. Szalewska, A., Wegrzyn, G., Taylor, K. (1994). Neither absence nor excess of O initiator-digesting ClpXP protease affects plasmid or phage replication in Escherichia coli. Mol. Microbiol , 13 , 469–474. Taylor, K. and Wegrzyn, G. (1995). Replication of coliphage lambda DNA. FEMS Microbiol. Reviews , 17 , 109–119. Theyssen H., Schuster, H.P., Bukau, B. and Reinstein, J. (1996). The second step of ATP binding to DnaK induces peptide release. J. Mol. Biol. , 263 , 657–670. Tilly, K. and Yarmolinsky, M. (1989). Participation of Escherichia coli heat shock proteins DnaJ, DnaK and GrpE in P1 plasmid replication. J. Bacteriol , 171 , 6025– 6029. Tomizawa, J. (1971). Function cooperation of genes O and P. In A.D.Hershey, (ed.), The
Role of chaperones in replication of bacteriophage lambda DNA
343
bacteriophage lambda , Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, pp. 549–552. Toothman, P. and Herskowitz, I. (1980). Rex-dependent exclusion of phages. Virology , 102 , 167–171. Tsurimoto, T. and Matsubara, K. (1981). Purified bacteriophage O protein binds to four repeating sequences at the replication origin. Nucl. Acids Res. , 9 , 1789–1799. Wawrzynow, A., Banecki, B. and Zylicz (1996). The Clp ATPases define a novel class of molecular chaperones. Molec. Microbiol. , 21 , 895–899. Wawrzynow, A., Wojtkowiak, D., Marszalek, J., Banecki, B., Jonsen, M., Graves, B., Georgopoulos, C. and Zylicz, M. (1995a). The ClpX heat-shock protein of Escherichia coli, the ATP-dependent substrate specificity component of ClpP-ClpX protease, is a novel molecular chaperone. EMBO J. , 14 , 1867–1877. Wawrzynow, A. and Zylicz, M. (1995). Divergent effect of ATP on the binding of the DnaK and DnaJ chaperones to each other, or to their various native and denatured protein substrates. J. Biol. Chem. , 270 , 19300–19306. Wawrzynow, A., Banecki, B., Wall, D., Liberek, K., Georgopoulos, C. and Zylicz, M. (1995b). ATP hydrolysis is required for the DnaJ-dependent activation of DnaK chaperone for binding to both native and denatured protein substrates. J. Biol. Chem. , 270 , 19307–9311. Wegrzyn, A., Wegrzyn, G., Taylor, K. (1996a). Disassembly of the coliphage replication complex due to the heat shock induction of the groE operone. Virology , 217 , 594–597. Wegrzyn, G., Wegrzyn, A., Pankiewicz, A. and Taylor, K. (1996b). Allele-specificity of the Escherichia coli dnaA gene in the replication of plasmid derived from phage . Mol. Gen. Genet. , 252 , 580–586. Wegrzyn, A., Wegrzyn, G. and Taylor, K. (1995a). Protection of coliphage O initiator protein from proteolysis in the assembly of the replication complex in vivo. Virology , 207 , 179–184. Wegrzyn, G., Szalewska-Palasz, A., Wegrzyn, A., Obuchowski, M. and Taylor, K. (1995b). Transcriptional activation of the origin of coliphage DNA replication is regulated by the host DnaA initiator function. Gene , 154 , 47–50. Wegrzyn, G., Wegrzyn, A., Konieczny, I., Bielawski, K., Konopa, G., Obuchowski, M. and Helinski, D.R. Taylor (1995c). Involvement of the host initiator function dnaA in the replication of coliphage . Genetics , 139 , 1469–1481. Wegrzyn, G., Pawlowicz, A. and Taylor, K. (1992). Stability of coliphage DNA replication initiator, the O protein. J. Mol. Biol , 226 , 675–680. Wegrzyn, G. and Taylor, K. (1992). Inheritance of the replication complex by one of two daughter copies during plasmid replication in Escherichia coli . J. Mol. Biol. , 226 , 681–688. Wickner, S., Skowyra, D., Hoskins, J. and McKenney, K. (1992). DnaJ, DnaK and GrpE heat shock proteins are required in oriP1 DNA replication solely at the RepA monomerization step. Proc. Natl. Acad. Sci. USA , 89 , 10345–10349. Wickner, S., Hoskins, J. and McKenney, K. (1991). Function of DnaJ and DnaK as chaperones in origin-specific DNA binding by RepA. Nature , 350 , 165–167. Wickner, S.H. and Zahn, K. (1986). Characterization of the DNA binding domain of
Molecular chaperones and folding catalysts
344
bacteriophage O protein. J. Biol. Chem. , 261 , 8537–8543. Wickner, S.H. (1978). DNA replication proteins of Escherichia coli and phage . Cold Spring Harbor Symp. Quant. Biol. , 43 , 303–310. Wojtkowiak, D., Georgopoulos, C. and Zylicz, M. (1993). Isolation and characterization of ClpX, new ATP-dependent specificity component of the Clp protease of Escherichia coli . J. Biol. Chem. , 268 , 22609–22617. Yamamoto, T., McIntyre, J., Sell., S.M., Georgopoulos, C., Skowyra, D. and Zylicz, M. (1987). Enzymology of the pre-priming steps in λdvDNA replication. J. Biol. Chem. , 262 , 7996–7999. Ziemienowicz, A., Zylicz, M., Floth, C. and Hubscher, U. (1995). Calf thymus Hsc70 protein protects and reactivates procaryotic and eucaryotic enzymes. J. Biol. Chem. , 270 , 15479–15484. Ziemienowicz, A., Skowyra, D., Zeilstra-Ryalls, J., Fayet, O., Georgopoulos, C. and Zylicz, M. (1993). Both the Escherichia coli chaperone systems, GroEL/GroES and DnaK/DnaJ/GrpE can reactivate heattreated RNA polymerase. J. Biol. Chem. , 268 , 25425–25431. Zylicz, M. (1993). The Escherichia coli chaperones involved in DNA replication. Phil. Trans. R. Soc. London. B , 339 , 271–278. Zylicz, M., Ang, D., Liberek, K. and Georgopoulos, C. (1989). Initiation of DNA replication with purified host- and bactriophage-encoded proteins: role of the DnaK, DnaJ and GrpE heat shock proteins. EMBO J. , 8 , 1601–1608. Zylicz, M., Ang, D., Georgopoulos, C. (1987). The grpE protein of Escherichia coli . J. Biol. Chem. , 262 , 17437–17442. Zylicz, M., Yamamoto, T., McKittrick, N., Sell, S. and Georgopoulos, C. (1985). Purification and properties of the DnaJ replication protein of Escherichia coli. J. Biol. Chem. , 260 , 7591–7598. Zylicz, M. and Georgopoulos, C. (1984). Purification and properties of the Escherichia coli dnaK replication protein. J. Biol. Chem. , 259 , 8820–8825. Zylicz, M., Gorska, I., Taylor, K. and Georgopoulos, C. (1984). Bacteriophage replication proteins: formation of a mixed oligomer and binding to the origin of DNA. Mol. Gen. Genet. , 196 , 401–406. Zylicz, M., LeBowitz, J. H., McMacken, R. and Georgopoulos, C. (1983). The dnaK protein of Escherichia coli possesses an ATPase and autophosporylating activity and is essential in an in vitroDNA replication system. Proc. Natl. Acad. Sci. USA , 80 , 6431– 6435.
15. CONTROL OF HORMONE RECEPTOR FUNCTION BY MOLECULAR CHAPERONES AND FOLDING CATALYSTS DAVID O.TOFT Department of Biochemistry and Molecular Biology, Mayo Clinic, Rochester, MN 55905, USA
1. Introduction 2. The Identification of Receptor-associated Proteins 3. The Assembly of Steroid Receptor Complexes 4. Hsp90 and its Associated Proteins Act on a Wide Variety of Cellular Regulatory Proteins 5. References 1. INTRODUCTION The steroid hormone receptors belong to a large family of transcriptional regulators which also includes receptors for vitamin D, retinoic acid, thyroid hormone and a growing number of orphan receptors for which regulatory ligands have not been identified (Mangelsdorf et al., 1995; Mangelsdorf and Evans 1995; Beato et al., 1995; Kastner et al., 1995). These are modular proteins that contain separable domains for binding a regulatory ligand, for binding DNA response elements, and for interaction with other transcription factors. While some members of this protein family can dimerize, bind DNA and perform some functions in the absence of ligand, the unoccupied steroid receptors remain sequestered in an inactive complex. The binding of steroid induces a conformational change in the receptor which results in its liberation from the complex. It can then dimerize, bind to DNA and fulfill the functions of an active receptor. While the concept of inactive and active structural states of steroid receptors was developed several years ago (Gorski et al., 1968; Jensen et al., 1968), the realization that these states relate directly to the activities of molecular chaperones is more recent (Sanchez et al., 1985; Schuh et al., 1985; Catelli et al., 1985). In the simplest terms, the inactive receptor interacts with Hsp70, Hsp90 and some additional proteins that might be classified as cochaperones. These interactions persist until hormone binding occurs and they are lost when the receptor assumes its active state. The study of this system has several
Control of hormone receptor function by molecular chaperones
347
advantages for the characterization of chaperone functions. It illustrates the cooperative activities of several chaperones, some which were first identified through studies on steroid receptors. Also, one can turn on or off the chaperoning event by the absence or presence of hormone. In addition, the hormone binding activity of the receptor is dependent on chaperones and provides an easily measured end-point for certain chaperone assisted events. The following pages provide a description of the roles of chaperones in this system and of the approaches that are being used to understand functional elements of the system. The likelihood that these studies relate to much more general roles of these chaperones in cell biology is also discussed. 2. THE IDENTIFICATION OF RECEPTOR-ASSOCIATED PROTEINS Studies performed 10 to 15 years ago indicated that the receptors for progesterone (PR) and glucocorticoids (GR) existed in large complexes containing several proteins. The additional proteins were initially thought by some investigators to be contaminants and this made the identification of the receptor proteins confusing. It is interesting that one initial report on the isolation of progesterone receptor had confused it with Hsp90 (Dougherty et al., 1982) and another report had it confused with Grp94 (Kulomaa et al., 1986). These mistakes were soon resolved and several years of study have now revealed nine proteins that interact with steroid receptors in their inactive states (see for review, Pratt and Toft, 1997; Smith, 1995). These are listed in Table 1. These have somewhat arbitrarily been classified as chaperones or co-chaperones. This functional relationship is now quite clear for Hsp70 and DnaJ homologs, but most of the other co-chaperones in Table 1 have been recognized because they bind to Hsp70 or Hsp90 and their functions are still unclear. Thus, the assembly of steroid receptor complexes provides a focus on the chaperoning functions of Hsp70 and Hsp90 with the participation of several proteins that may modulate or extend the functions of these two chaperones. Hip (Hsp70 interacting protein) is a protein that was isolated because it interacts with the ATPase domain of Hsp70 (Höhfeld et al., 1995). This interaction appears to be dependent on the presence of Hsp40 (DnaJ homolog) and ATP/ADP, and it has been proposed that Hip promotes the state for tight substrate binding (ADP state) of Hsp70 (see Buchberger et al., and Ha et al., this volume). Hip was Table 1 Receptor-associated proteins
Chaperone
Co-chaperones
Hsp70
Hsp40/DnaJ homolog Hip/p48 Hop/Sti 1 homolog
Hsp90
Hop/Sti 1 homolog Immunophilins: FKBP52 FKBP51
Molecular chaperones and folding catalysts
348
CyP-40 p23
independently identified as a component of progesterone receptor complexes and was called p48 according to its migration on gel electrophoresis (Prapapanich et al., 1996). It is found in complexes with Hsp70 plus Hsp90 and Hop, and it appears mainly in intermediate receptor complexes (see below). Hop is a 60 kDa homolog of the yeast heat shock protein Sti 1 (Honoré et al., 1992; Nicolet and Craig, 1989) and it was recognized as an early component in receptor complex formation (Smith et al., 1993b; Smith 1993). Sti 1 is not an essential protein in yeast and, although it has been shown to bind to Hsp90 (Chang and Lindquist, 1994), its function is unknown. In higher eukaryotes, Hop can bind to both Hsp70 and Hsp90 to form a ternary complex (Smith et al., 1993b; Chen et al., 1996) and it has been suggested that Hop (hsp organizing protein) acts to organize the functions of Hsp70 and Hsp90. Gross and Hesseforth (1996) have isolated a protein identical to Hop which they call RFHsp70 (recycling factor for Hsp70). They show that this protein promotes the dissociation of ADP from Hsp70 and its replacement by ATP, but this activity has been questioned by Johnson et al., 1998). Another puzzling group of Hsp90-binding proteins are the high molecular weight immunophilins FKBP51, FKBP52 and CyP-40. These are proteins that have peptidylprolyl-cis/trans isomerase activity and the capacity to bind the immunosuppressant drugs FK506 (FKBP52, FKBP51) or cyclosporin A (CyP-40) (see Fischer and Schmid, this volume). FKBP51 has also been called FKBP54 (Smith et al., 1993a) and FKBP52 has had several names including p59, FKBP59, Hsp56 and HBI (Hsp90-binding immunophilin). These three proteins are found in receptor complexes and they are also bound to Hsp90 in the absence of receptor (Radanyi et al., 1994; Czar et al., 1994; Hoffmann and Handschumacher, 1995). They do not bind together but appear to compete for a common site on Hsp90 (Owens-Grillo et al., 1995). All three proteins contain TPR domains or 34 residue repeating units that are required for Hsp90 binding (Radanyi et al., 1994; Owens-Grillo et al., 1996; Ratajczak and Carello, 1996). TPR domains have been observed in a wide variety of proteins including Hip and Hop and it has been proposed that these domains participate in protein-protein interactions (Lamb et al., 1995). Immunophilins have been implicated as chaperones in protein folding, in the control of cell signalling pathways and in targeting proteins to cellular locations, but their actual functions in receptor complexes are completely unknown. Like the immunophilins, p23 was identified from its binding to receptor complexes and it also binds to Hsp90 in the absence of receptor (Johnson et al., 1994; Johnson and Toft, 1994). p23 is not structurally related to immunophilins or to known heat shock proteins. Its binding to Hsp90 is dependent on the presence of ATP and it thus may be a modulator of Hsp90 activities. However, its precise function is unknown. Most of the p23 in cell extracts is in complex with Hsp90 plus FKBP51, FKBP52 and CyP-40, and this may represent a chaperoning complex needed at the later stages of receptor complex assembly (Johnson and Toft, 1995). The proteins in Table 1 were revealed by their co-purification during isolation of
Control of hormone receptor function by molecular chaperones
349
inactive steroid receptor complexes. The co-chaperones were then shown to copurify with Hsp70 or Hsp90 as relatively prominent complexes from cell extracts. For example, the abundance of p23 is only 10–20% of the Hsp90 level in cytosol extracts and the majority of p23 is bound to Hsp90 (Johnson et al., 1994). It should be noted that the steroid receptors exist at less than 1% of the Hsp90 level and most known Hsp90 substrates or targets are proteins of low abundance that would only occupy a small fraction of cellular Hsp90. DnaJ homologs, which will be termed collectively as Hsp40, are the only proteins in Table 1 that haven’t been clearly demonstrated in steroid receptor complexes. However, they have been implicated by genetic manipulation of yeast model systems. Mutation of the yeast Ydj-1 protein has been shown to dramatically alter the functional capacity of androgen (Caplan et al., 1995) and glucocorticoid receptors (Kimura et al., 1995). Yeast have also been used to demonstrate the requirement for Hsp90 in the functioning of steroid receptors (Picard et al., 1990; Bohen and Yamamoto, 1993; Kimura et al., 1994; Bohen, 1995a; Nathan and Lindquist, 1995). It will be important to test the significance of the other receptor-associated proteins by genetic manipulation of yeast or other cellular systems. 3. THE ASSEMBLY OF STEROID RECEPTOR COMPLEXES Early attempts to study the interaction of steroid receptors with their associated proteins were frustrated by the instability of these associations in vitro and the observations that receptor complexes don’t form easily by a simple combination of components. This led Pratt and co-workers to suggest that the complexes were formed during receptor synthesis, requiring interactions with nascent polypeptide chains. They and others were able to show that glucocorticoid receptor synthesized in rabbit reticulocyte lysate was in a native multiprotein complex, able to bind hormone (Dalmon et al., 1989; Denis and Gustafsson, 1989). However, subsequent studies showed that translation was not essential, but factors in the reticulocyte lysate provided optimal conditions for complex formation. Thus, native receptor isolated from target tissue could be stripped of its associated proteins by high salt treatment and then reconstituted into a complex by incubation in rabbit reticulocyte lysate (Smith et al., 1990). It has since been shown that other cellular extracts are also effective in this regard, but they must be quite concentrated and able to maintain an effective level of ATP (Stancato et al., 1996a). The formation of receptor complexes required incubation at elevated temperatures, magnesium, potassium, and the hydrolysis of ATP (Smith et al., 1992; Hutchison et al., 1992a). The microsomal fraction, normally present in reticulocyte lysate, was not required. Much new information has been acquired through use of the in vitro system which has led to models of the receptor assembly process. One such model, illustrated in Figure 1, is quite similar to a recent model by Prapapanich et al., (1996). Many liberties have been taken in this model since we are still very early in our understanding of this process. The assembly of the receptor complex is shown to progress through three stages. The first stage, complex 1, represents an initial interaction of Hsp70 with hydrophobic regions in the ligand binding domain of the receptor. This stage has not been described
Molecular chaperones and folding catalysts
350
experimentally and is simply based on the demonstrated ability of Hsp70 to bind hydrophobic domains in other systems and the fact that Hsp70 is an essential component in the assembly of receptor complexes (Hutchison et al., 1994). Hsp40 is introduced here because it is known as an Hsp70 co-chaperone and it has been implicated by genetic and biochemical studies (Caplan et al., 1995; Kimura et al., 1995; Dittmar et al., 1998). An involvement of Hip at this step is also suggested because of its known interaction with Hsp70. In addition, studies on the time course of complex formation with the PR have shown extensive binding of Hsp70, Hip and Hop at early times and a loss of these components at later times (Smith, 1993). Hop and Hsp90 are incorporated at the second stage of assembly to make an intermediate state. This complex has been observed under three conditions: at early times in the assembly of PR complexes (Smith, 1993), in the absence of an ATP regeneration system when the level of ATP is suboptimal (Smith et al., 1992) and when assembly is inhibited by geldanamycin, a reagent that blocks some functions of Hsp90 (Whitesell et al., 1994; Johnson and Toft, 1995; Smith et al., 1995). Thus, there is good evidence for the existence of this intermediate form. Toward the right in Figure 1, a ternary complex with Hsp90, Hop, and Hsp70 is indicated in brackets. This has been observed in cell extracts and it has been proposed that this is a precursor for receptor assembly (Smith et al., 1993b; Dittmar et al., 1997). Thus, it is possible that the receptor binds pre-formed complexes of these proteins instead of the individual proteins as shown in the model. While not indicated in the model, at stage 2, Hsp90 is underrepresented in comparison to stage 3. In other words, there is more Hsp70 than Hsp90 at stage 2, but little Hsp70 and about twice as much Hsp90 at stage 3 (Smith, 1993; Johnson and Toft, 1995; Smith et al., 1995). The conversion from stage 2 to 3 is the geldanamycin-sensitive step where more Hsp90 plus p23 and the immunophilins are incorporated into the complex. This stage utilizes additional ATP which might be required for a small sustained action of Hsp70 and/or an ATP-dependent activity of Hsp90. Whether or not Hsp90 interacts with ATP in a functional sense has been a controversial issue (Csermely et al., 1993; Jacob et al., 1996) however, recent structural and biochemical studies clearly reveal an ATP binding site near the N-terminus of hsp90 (Stebbin et al., 1997; Prodromou et al., 1997; Grenert et al., 1997). This site also binds the hsp90 inhibitor, geldanamycin. We have found that the binding of ATP to hsp90 promotes a conformational change that allows hsp90 to interact with p23 (Grenert et al., 1997). It is at this last step in Figure 1 when the receptor becomes competent to bind hormone since hormone binding activity, at least for some receptors, is blocked by the inhibitor geldanamycin (Smith et al., 1995). Thus, there is clearly a conformational change in the receptor at this stage. As mentioned earlier, Hsp90 can exist in a pre-formed complex with p23 and any one of the immunophilins FKBP51, FKBP52 and CyP-40 and it may be this complex rather than the individual proteins that interacts with the receptor. The formation of this complex of Hsp90 with p23 and immunophilin also requires ATP (for p23 binding, but not for immunophilin binding) and it is this process that is inhibited by geldanamycin (Johnson and Toft, 1995). Several experiments have indicated the dynamic nature of this system. When individual proteins are labelled with radioisotopes or antibody epitopes, these are rapidly exchangeable with unlabelled components within a few minutes at 30° (Smith, 1993;
Control of hormone receptor function by molecular chaperones
351
Smith et al., 1995). Thus, it is likely that the complexes are continu-
Figure 1 A model for the assembly of steroid receptor complexes. The receptor (R) is divided into three domains: an amino terminal trans-activation domain, an internal DNA-binding domain, and a carboxy-terminal ligand binding domain interacting initially with Hsp70. In some cases, the proteins are abbreviated by use of a number, eg. 60=Hop/p60. At the right of the figure, two multiprotein complexes are shown which may participate in receptor complex formation. The shaded triangle in the receptor at step 3 indicates the formation of the ligand binding site.
ously dissociating and re-assembling and that the process would be better represented as a cycle rather than a linear pathway. This model brings to mind several important questions that require further study. The largest gap in knowledge concerns function at all levels. Why are so many proteins needed in this process? What are the functions of the individual proteins and how are these functions coordinated among proteins? What are the roles of ATP/ADP and are there other factors or proteins of this process that have not yet been recognized? Efforts to answer these questions have only begun and there are, as yet, few answers. Yeast genetics has been employed to test the significance of some proteins. This is possible because several steroid receptors can be expressed in yeast where they respond
Molecular chaperones and folding catalysts
352
to hormone to induce expression of a reporter gene (Picard, 1997; Caplan, 1997). Several mutants of Hsp90 have been tested in this system. When Hsp90 is mutated or reduced to low levels, hormone responsiveness is lowered dramatically and this can be explained, in part, by a loss of hormone binding activity of the receptor (Picard et al., 1990; Bohen andYamamoto, 1993; Kimura et al., 1994; Bohen, 1995a; Nathans and Lindquist, 1995). Similar results have been observed with mutations of the DnaJ homologYdj-1 (Caplan et al., 1995; Kimura et al., 1995) indicating that both of these proteins are needed to process the receptor to a native, fully functional state. The yeast Hop homolog Sti 1, can be deleted since it is not essential for viability (Nicolett and Craig, 1989). However, studies on the activity of the glucocorticoid receptor in yeast show that the deletion of Sti1 reduces the activity of this receptor and the functions of Sti1 appear to be linked to those of hsp90 (Chang et al., 1997). Preliminary studies indicate that the yeast homolog of p23 can also be deleted without loss of yeast viability resulting in only subtle effects on glucocorticoid receptor function (Bohen, 1995b). These studies must be pursued further to study compensatory mechanisms and additive effects of various mutations. A second approach has been to study the in vitro reconstitution of receptor complexes with regard to intermediate states and the removal of individual proteins. The ability to reconstitute GR complexes in rabbit reticulocyte lysate is lost after depletion of Hsp70 from the lysate and is re-gained by adding back Hsp70 (Hutchison et al., 1994). Similar results have been found with the removal and replacement of p23 (Johnson and Toft, 1994). In an interesting study by Hutchison et al. (1995), GR was found to form complexes containing Hsp90 in wheat germ lysate, but the receptor lacked hormone binding activity unless exogenous p23 was added. This indicates the importance of p23 in the formation or stability of a native complex and suggests that p23 is lacking in wheat germ lysate or is not compatible with animal steroid receptor systems. The importance of Hop in in vitro receptor complex formation has been indicated by use of a Hop antibody which apparently neutralizes Hop function (Chen et al., 1996). In the presence of antibody, PR complexes accumulate in an intermediate complex similar to complex 2 in Figure 1 which lacks hormone binding activity. An antibody to Hsp70 has also been used to block its activity and the formation of PR complexes (Smith et al., 1992). There is very little information on the functions of FKBP52, FKBP51 or CyP-40. Treatment of reticulocyte lysate with FK506 or cyclosporin A does not alter the formation of steroid receptor complexes (Hutchison et al., 1993). These drugs block the peptidyl prolyl isomerase activity of the immunophilins, indicating that this activity is not essential to receptor complex formation. On the other hand, these immunosuppressant drugs have been shown to alter receptor responsiveness in cell culture systems (Ning and Sanchez, 1993; Tai et al., 1994; Milad et al., 1995; Renoir et al., 1995) and one report indicates an increased hormone binding affinity of PR after treatment with FK506 in vitro (Renoir et al., 1994). Greater efforts must be made to develop chemically defined systems for the formation and study of receptor complexes. This would tell us first of all whether all of the essential components have been identified. One could then add or subtract components, replace them with mutant forms, prepare partial or intermediate complexes, thereby allowing a
Control of hormone receptor function by molecular chaperones
353
much clearer description of this process and the study of functional contributions of various components. Along this line, Pratt and co-workers have recently succeeded in forming GR complexes using a combination of purified proteins including Hsp90, Hsp70, Hop, p23, and Hsp90 (Dittmar et al., 1996; Dittmar et al., 1998). Four of these proteins are essential for receptor complex formation and the fifth, p23, appears to stabilize the complex in a state that provides the receptor with hormone binding activity. Additional studies are needed on this minimal system to test the importance of Hip or immunophilins and the possible contribution of other factors. 4. HSP90 AND ITS ASSOCIATED PROTEINS ACT ON A WIDE VARIETY OF CELLULAR REGULATORY PROTEINS Hsp90 has been studied in relation to cell stress and it has recently been shown to prevent aggregation of denatured proteins in a manner that does not require ATP or additional protein components (Jacob et al., 1995; Freeman and Morimoto, 1996). However, most studies on Hsp90 have concerned its association with other cellular proteins in either persistent complexes, as with steroid receptors, or transient associations such as that with the tyrosine kinase pp60v-src. These studies have led to the idea that Hsp90 plays a specialized role in chaperoning the folding, processing or trafficking of a unique group of cellular proteins. There are now a wide variety of proteins that have been indicated to be substrates for Hsp90. Most of these are either regulators of transcription or protein kinases (see chapter by Bensaude, this volume). Table 2 lists a number of proteins in these two groups. Some of these have only been linked to Hsp90 through genetics studies, such as Wee 1 and sevenless tyrosine kinases, but most have been recognized in complexes containing Hsp90. Other components in these complexes have not been described in most cases. The complexes for dioxane receptor and for heme-regulated eIF-2a kinase have been described and they show close similarity to steroid receptor complexes (Chen and Perdew, 1994; Nair et al., 1996; Matts et al., 1992). They also contain Hsp70, FKBP52, and p23. These proteins have also been shown in complexes with fes tyrosine kinase and hsf-1 (Nair et al., 1996). In addition, complexes with pp60v-src, fes and Raf-1 contain a phosphoprotein, p50, that is not present in steroid receptor complexes (Brugge et al., 1981; Oppermann et al., 1981; Nair et al., 1996; Stancato et al., 1993; Wartmann and Davis, 1984). p50 also binds to Hsp90 in the absence of the kinase (Whitelaw et al., 1991). A recent study by Stepanova et al., (1996) has revealed the identity of p50 to be Cdc 37. This protein is essential in yeast. It is genetically linked to multiple protein kinases and it is required for the production of Cdc 28/cyclin complexes. Cdc 37 appears to target Hsp90 to the Cdk 4 kinase and the authors suggest that it has a general role in the assisted folding of intrinsically unstable protein kinases. The studies on Hsp90-kinase complexes suggest that some variations exist in the composition of Hsp90 complexes and it will be important to study more examples for comparison. The list of proteins in Table 2 suggests that the functions of Hsp90 might be of special importance to signalling pathways and regulation. However, it is also possible that Table 2 reflects the research bias of investigators. There may be a number of
Molecular chaperones and folding catalysts
354
Table 2 Hsp90-associated proteins
Associated Proteins
References
Transcription Factors Glucocorticoid receptor
Sanchez et al., 1985
Progesterone receptor
Schuh et al., 1985; Catelli et al., 1985
Estrogen receptor
Joab et al., 1984; Redeuilh et al., 1987
Androgen receptor
Joab et al., 1984; Veldscholte et al., 1992
Mineralocorticoid receptor
Rafestin-Oblin et al., 1989
v-erb A
Privalsky, 1991
Ah receptor
Perdew, 1988; Denis et al., 1988
Myo D1
Shue and Kohtz, 1994
Heat shock factor
Nadeau et al., 1993; Nair et al., 1996
CN-tumor promoter-specific binding protein
Hashimoto and Shudo, 1991
Sim
McGuire et al., 1996
E12
Shue and Kohtz, 1992
p53 tumor suppressor mutant
Sepehrnia et al., 1996; Blagosklonay et al., 1996
Hepatitis B virus reverse transcriptase Hu and Seeger, 1996 Protein Kinases v-src, c-src
Brugge et al., 1981; Oppermann et al., 1981; Hutchison et al., 1992b
v-fps
Lipsich et al., 1982
v-yes
Lipsich et al., 1982
v-fes
Ziemiecki et al., 1986; Nair et al., 1996
v-frg, c-frg
Ziemiecki et al., 1986; Hartson and Matts, 1994
lck
Hartson and Matts, 1994
Weel
Alique et al., 1994
Sevenless
Cutforth and Rubin, 1994
Heme-regulated eIF-2a kinase
Rose et al., 1987; Matts and Hurst, 1989;
eEF-2 kinase
Palmquist et al., 1994
Casein kinase II
Miyata and Yahara, 1992; Shi et al., 1994
v-Raf, c-Raf, Gag-Mil
Stancato et al., 1993; Wartmann and Davis, 1994; Lovric et al., 1994
Control of hormone receptor function by molecular chaperones MAP kinase kinase (Mek)
355
Stancato et al., 1997
proteins outside transcription factors and protein kinases that interact with Hsp90. A few known examples are actin (Koyasu et al., 1986), tubulin (Sanchez et al., 1988), proteasomes (Tsubuki et al., 1994; Wagner and Margolis, 1995), G protein -complexes (Inanobe et al., 1994) and calmodulin (Minami et al., 1993). These may also be regulatory interactions or they may simply be transient chaperone-assisted folding processes of a more general purpose. The tools now available should expand the scope of studies in this area. The drug geldanamycin has already provided much new information on processes involving Hsp90 since it can be used to inhibit Hsp90 in intact cells. Thus, geldanamycin has been shown to block the processing and activity of pp60v-src (Whitesell et al., 1994), Raf-1 (Schulte et al., 1995), mutant p53 (Blagosklonny et al., 1995), glucocorticoid receptor (Whitesell and Cook, 1996), progesterone receptor (Smith et al., 1995) and the reverse transcriptase of hepatitis B virus (Hu and Seeger, 1996). Further characterization of these diverse processes should provide a much better understanding of the components and the functional responsibilities of the chaperone pathway involving Hsp70 and Hsp90. 5. REFERENCES Aligue, R., Akhavan-Niak, H. and Russell, P. (1994). A role for Hsp90 in cell cycle control: Weel tyrosine kinase activity requires interaction with Hsp90. EMBO J. , 13 , 6099–6106. Beato, M., Herrlich, P. and Schütz, G. (1995). Steroid hormone receptors: Many actors in search of a plot. Cell , 83 , 851–857. Blagosklonny, M.V., Toretsky, J. and Neckers, L. (1995). Geldanamycin selectively destabilizes and conformationally alters mutated p53. Oncogene , 11 , 933–939. Blagosklonny, M.V., Toretsky, J., Bohen, S. and Neckers, L. (1996). Mutant conformation of p53 translated in vitro or in vivo requires functional Hsp90. Proc. Natl. Acad. Sci., USA , 93 , 8374–8383. Bohen, S.P. (1995a). Hsp90 mutants disrupt glucocorticoid receptor ligand binding and destabilize aporeceptor complexes. J. Biol. Chem. , 270 , 29433–29438. Bohen, S.P. (1995b). Pharmacological and genetic analysis of the roles of Hsp90 and p23 in signaling protein function. Mol. Biol. Cell , 6 , 6a. Bohen, S.P. and Yamamoto, K.R. (1993). Isolation of Hsp90 mutants by screening for decreased steroid receptor function. Proc. Natl. Acad. Sci. USA , 90 , 11424–11428. Brugge, J.S., Erikson, E. and Erikson, R.L. (1981). The specific interaction of the Rous sarcoma virus transforming protein, pp60src, with two cellular proteins. Cell , 25 , 363– 372. Caplan, A.J., Langley, E., Wilson, E.M. and Vidal, J. (1995). Hormone-dependent transactivation by the human androgen receptor is regulated by a dnaJ protein. J. Biol. Chem. , 270 , 5251–5257. Caplan, A.J. (1997). Yeast molecular chaperones and the mechanism of steroid hormone action. Trends Endocrinol. Metab. , 8 , 271–276.
Molecular chaperones and folding catalysts
356
Catelli, M.G., Binart, N., Jung-Testas, I., Renoir, J.M., Baulieu, E.E., Feramisco, J.R. and Welch, W.J. (1985). The common 90-kd protein component of non-transformed ‘8S’ steroid receptors is a heat-shock protein. EMBO J. , 4 , 3131–3135. Chang, H.-C.J. and Lindquist, S. (1994). Conservation of Hsp90 macromolecular complexes in Saccharomyces cerevisiae. J. Biol. Chem. , 269 , 24983–24988. Chang, H-C.J., Nathan, D.F. and Lindquist, S. (1997). In vivo analysis of the hsp90 cochaperone Sti1 (p60). Molec. Cell Biol. , 17 , 318–325. Chen, H.S. and Perdew, G.H. (1994). Subunit composition of the heteromeric cytosolic aryl hydrocarbon receptor complex. J. Biol. Chem . 269 , 27554–27558. Chen, S., Prapapanich, V., Rimerman, R.A., Honoré, B. and Smith, D.F. (1996). Interactions of p60, a mediator of progesterone receptor assembly, with heat shock proteins Hsp90 and Hsp70. Mol Endocrinol , 10 , 682–693. Csermely, P., Kajtar, J., Hollóis, M.,Jalsovaszky, G., Holly, S., Kahn, C.R., Gergely, Jr., P., Soti, C., Mihály, K, and Somogyi, J. (1993). ATP induces a conformational change of the 90-kDa heat shock protein (Hsp90). J. Biol Chem. , 268 , 1901–1907. Cutforth, T. and Rubin, G.M. (1994). Mutations in hsp83 and cdc37 impair signaling by the sevenless receptor tyrosine kinase in drosophila. Cell , 77 , 1027–1036. Czar, M.J., Owens-Grillo, J.K., Dittmar, K.D., Hutchison, K.A., Zacharek, A.M., Leach, K.L., Deibel, Jr., M.R. and Pratt, W.B. (1994). Characterization of the protein-protein interactions determining the heat shock protein (Hsp90-Hsp70-hsp56). heterocomplex. J. Biol Chem. , 269 , 11155–11161. Dalman, F.C., Bresnick, E.H., Patel, P.O., Perdew, G.H., Watson, S.J., Jr. and Pratt, W.B. (1989). Direct evidence that the glucocorticoid receptor binds to Hsp90 at or near the termination of receptor translation in vitro. J. Biol. Chem. , 264 , 19815–19821. Denis, M., Cuthill, S., Wikstrom, A.-C., Poellinger, L. and Gustafsson, J.-A. (1988). Association of the dioxin receptor with the Mr 90,000 heat shock protein: A structural kinship with the glucocorticoid receptor. Biochem. Biophys. Res. Comm. , 155 , 801– 807. Denis, M. and Gustafsson, J.-A. (1989). Translation of glucocorticoid receptor mRNA in vitro yields a nonactivated protein. J. Biol. Chem. , 264 , 6005–6008. Dittmar, K.D., Banach, M., Galigniana, M.D. and Pratt, W.B. (1998). The role of DnaJlike proteins in glucocorticoid receptor-hsp90 heterocomplex assembly by the reconstituted hsp90-p60-hsp70 foldosome complex. J. Biol. Chem. , 273 , 7358–7366. Dittmar, K.D., Hutchison, K.A., Owens-Grillo, J.K. and Pratt, W.B. (1996). Reconstitution of the steroid receptor Hsp90 heterocomplex assembly system of rabbit reticulocyte lysate. J. Biol. Chem. , 271 , 12833–12839. Dougherty, J.J., Puri, R.K, and Toft, D.O. (1982). Phosphorylation in vivo of chicken oviduct progesterone receptor. J. Biol. Chem. , 257 , 14226–14230. Freeman, B.C. and Morimoto, R.I. (1996). The human cytosolic molecular chaperones Hsp90, Hsp70 (hsc70). and hdj-1 have distinct roles in recognition of a non-native protein and protein refolding. EMBO J. , 15 , 2969–2979. Gorski, J., Toft, D., Shyamala, G., Smith, D. and Notides, A. (1968). Hormone receptors: Studies on the interaction of estrogen with the uterus. Rec. Prog. Horm. Res. , 24 , 45– 80. Grenert, J.P., Sullivan, W.P., Fadden, P., Haystead, T.A.J., Clark, J., Mimnaugh, E.,
Control of hormone receptor function by molecular chaperones
357
Krutzsch, H., Ochel, H-J., Schulte, T.W., Sausville, E., Neckers, L.M. and Toft, D.O. (1997). The amino-terminal domain of heat shock protein 90 (hsp90) that binds geldanamycin is an ATP/ADP switch domain that regulates hsp90 conformation. J. Biol. Chem. , 272 , 23843–23850. Gross, M. and Hessefort, S. (1996). Purification and characterization of a 66-kDa protein from rabbit reticulocyte lysate which promotes the recycling of Hsp70. J. Biol. Chem. , 271 , 16833–16841. Hartson, S.D. and Matts, R.L. (1994). Association of Hsp90 with cellular Src-family kinases in a cell-free system correlates with altered kinase structure and function. Biochemistry , 33 , 8912–8920. Hashimoto, Y. and Shudo, K. (1991). Cytosolic-nuclear tumor promoter-specific binding protein: Association with the 90 kDa heat shock protein and translocation into nuclei by treatment with 12-O-tetradecanoylphorbol 13-acetate. Jpn. J. Cancer Res. , 82 , 665–675. Hoffmann, K. and Handschumacher, R.E. (1995). Cyclophilin-40: Evidence for a dimeric complex with Hsp90. Biochem. J. , 307 , 5–8. Höhfeld, J., Minami, Y. and Hartl, F.-U. (1995). Hip, a novel cochaperone involved in the eukaryotic Hsc70/Hsp40 reaction cycle. Cell , 83 , 589–598. Honoré, B., Leffers, H., Madsen, P., Rasmussen, H.H., Vandekerckhove, J. and Celis, J.E. (1992). Molecular cloning and expression of a transformation-sensitive human protein containing the TPR motif and sharing identity to the stress-inducible yeast protein STI1. J. Biol. Chem. , 267 , 8485–8491. Hu, J. and Seeger, C. (1996). Hsp90 is required for the activity of a hepatitis B virus reverse transcriptase. Proc. Natl. Acad. Sci. USA , 93 , 1060–1064. Hutchison, K.A., Czar, M.J., Scherrer, L.C. and Pratt, W.B. (1992a). Monovalent cation selectivity for ATP-dependent association of the glucocorticoid receptor with Hsp70 and Hsp90. J. Biol. Chem. , 267 , 14047–14053. Hutchison, K.A., Brott, B.K., DeLeon, J.H., Perdew, G.H., Jove, R. and Pratt, W.B. (1992b). Reconstitution of the multiprotein complex of pp60src, Hsp90, and p50 in a cell-free system. J. Biol. Chem. , 267 , 2902–2908. Hutchison, K.A., Scherrer, L.C., Czar, M.J., Ning, Y., Sanchez, E.R., Leach, K.L., Deibel, M.R., Jr. and Pratt, W.B. (1993). FK506 binding to the 56-kilodalton immunophilin (Hsp56). in the glucocorticoid receptor heterocomplex has no effect on receptor folding or function . Biochemistry , 32 , 3953–3957. Hutchison, K.A., Dittmar, K.D., Czar, M.J. and Pratt, W.B. (1994). Proof that Hsp70 is required for assembly of the glucocorticoid receptor into a heterocomplex with Hsp90. J. Biol. Chem. , 269 , 5043–5049. Hutchison, K.A., Stancato, L.F., Owens-Grillo, J.K., Johnson, J.L., Krishna, P., Toft, D.O. and Pratt, W.B. (1995). The 23-kDa acidic protein in reticulocyte lysate is the weakly bound component of the hsp foldosome that is required for assembly of the glucocorticoid receptor into a functional heterocomplex with Hsp90. J. Biol. Chem. , 270 , 18841–18847. Inanobe, A., Takahashi, K. and Katada, T. (1994). Association of the -subunits of trimeric GTP-binding proteins with 90-kDa heat shock protein, Hsp90. J. Biochem. , 115 , 486–492.
Molecular chaperones and folding catalysts
358
Jakob, U., Lilie, H., Meyer, I. and Buchner, J. (1995). Transient interaction of Hsp90 with early unfolding intermediates of citrate synthase. J. Biol Chem. , 270 , 7288–7294. Jakob, U., Scheibel, T., Bose, S., Reinstein, J., Buchner, J. (1996). Assessment of the ATP binding properties of Hsp90. J. Biol. Chem. , 271 , 10035–10041. Jensen, E.V., Suzuki, T., Kawashima, T., Stumpf, W.E., Jungblut, P.W. and DeSombre, E.R. (1968). A two-step mechanism for the interaction of estradiol with rat uterus. Proc. Natl. Acad. Sci. USA , 59 , 632–638. Joab, I., Radanyi, C., Renoir, M., Buchou, T., Catelli, M.-C., Binart, N., Mester, J. and Baulieu, E.-E. (1984). Common non-hormone binding component in non-transformed chick oviduct receptors of four steroid hormones. Nature , 308 , 850–853. Johnson, B.D., Schumacher, R.J., Ross, E.D. and Toft, D.O. (1998). Hop modulates hsp70/hsp90 interactions in protein folding. J. Biol. Chem. , 273 , 3679–3686. Johnson, J.L. and Toft, D.O. (1994). A novel chaperone complex for steroid receptors involving heat shock proteins, immunophilins, and p23. J. Biol. Chem. , 269 , 24989– 24993. Johnson, J.L., Beito, T.G., Krco, C.J. and Toft, D.O. (1994). Characterization of a novel 23-kilodalton protein of unactive progesterone receptor complexes. Mol. Cell. Biol. , 14 , 1956–1963. Johnson, J.L. and Toft, D.O. (1995). Binding of p23 and Hsp90 during assembly with the progesterone receptor. Mol. Endocrinol , 9 , 670–678. Johnson, J., Corbisier, R., Stensgard, B. and Toft, D.O. (1996). The involvement of p23, Hsp90, and immunophilins in the assembly of progesterone receptor complexes. J. Steroid Biochem. Mol. Biol. , 56 , 31–37. Kastner, P., Mark, M. and Chambon, P. (1995). Nonsteroid nuclear receptors: What are genetic studies telling us about their role in real life? Cell , 83 , 859–869. Kimura, Y., Matsumoto, S. and Yahara, I. (1994). Temperature-sensitive mutants of hsp82 of the budding yeast Saccharomyces cerevisiae. Mol. Gen. Genet. , 242 , 517– 527. Kimura, Y., Yahara, I. and Lindquist, S. (1995). Role of the protein chaperone YDJ1 in establishing Hsp90-mediated signal transduction pathways. Science , 268 , 1362–1365. Koyasu, S., Nishida, E., Kadowaki, T., Matsuzaki, F., Iida, K., Harada, F., Kasuga, M., Sakai, H. and Yahara, I. (1986). Two mammalian heat shock proteins, Hsp90 and hsp100, are actin-binding proteins. Proc. Natl Acad. Sci. USA , 83 , 8054–8058. Kulomaa, M.S., Weigel, N.L., Kleinsek, D.A., Beattie, W.G., Conneely, O.M., March, C., Zarucki-Schulz, T., Schrader, W.T. and O’Malley, B.W. (1986). Amino acid sequence of a chicken heat shock protein derived from the complementary DNA nucleotide sequence. Biochemistry , 25 , 6244–6251. Lamb, J.R., Tugendreich, S. and Hieter, P. (1995). Tetratrico peptide repeat interactions: to TPR or not to TPR? TIBS , 20 , 257–259. Lipsich, L.A., CuttJ. and Brugge, J.S. (1982). Association of the transforming proteins of Rous, Fujinami and Y73 avian sarcoma viruses with the same two cellular proteins. Mol. Cell. Biol. , 2 , 875–880. Lovric, J., Bischof, O. and Moelling, K, (1994). Cell cycle-dependent association of GagMil and Hsp90. FEBS Letters , 343 , 15–21. Mangelsdorf, D.J. and Evans, R.M. (1995). The RXR heterodimers and orphan receptors.
Control of hormone receptor function by molecular chaperones
359
Cell , 83 , 841–850. Mangelsdorf, D.J., Thummel, C., Beato, M., Herrlich, P., Schütz, G., Umesono, K., Blumberg, B., Kastner, P., Mark, M., Chambon, P. and Evans, R.M. (1995). The nuclear receptor superfamily: The second decade. Cell , 83 , 835–839. Matts, R.L. and Hurst, R. (1989). Evidence for the association of the heme-regulated eIF2a kinase with the 90-kDa heat shock protein in rabbit reticulocyte lysate in situ. J. Biol. Chem. , 264 , 15542–15547. Matts, R.L., Xu, Z., Pal, J.K. and Chen, J.-J. (1992). Interactions of the heme-regulated eIF-2a kinase with heat shock proteins in rabbit reticulocyte lysates. J. Biol. Chem. , 267 , 18160–18167. McGuire, J., Coumailleau, P., Whitelaw, M.L., Gustafsson, J.-A. and Poellinger, L. (1996). The basic helix-loop-helix/PAS factor sim is associated with Hsp90. J. Biol. Chem. , 270 , 31353–31357. Milad, M., Sullivan, W., Diehl, E., Altmann, M., Nordeen, S., Edwards, D.P. and Toft, D.O. (1995). Interaction of the progesterone receptor with binding proteins for FK506 and cyclosporin A. Mol. Endocrinol. , 9 , 838–847. Minami, Y., Kawasaki, H., Suzuki, K. and Yahara, I. (1993). The calmodulin-binding domain of the mouse 90-kDa heat shock protein. J. Biol. Chem. , 268 , 9604–9610. Miyata, Y. and Yahara, I. (1992). The 90-kDa heat shock protein, Hsp90, binds and protects casein kinase II from self-aggregation and enhances its kinase activity. J. Biol. Chem. , 267 , 7042–7047. Nadeau, K., Das, A. and Walsh, C.T. (1993). Hsp90 chaperonins possess ATPase activity and bind heat shock transcription factors and peptidyl prolyl isomerases. J. Biol. Chem. , 268 , 1479–1487. Nair, S.C., Toran, E.J., Rimerman, R.A., Hjermstad, S., Smithgall, T.E. and Smith, D.F. (1996). A pathway of multi-chaperone interactions common to diverse regulatory proteins: Estrogen receptor, Fes tyrosine kinase, heat shock transcription factor HSF1, and the arylhydrocarbon receptor. Cell Stress and Chaperones , 1 , 237–250. Nathan, D.F. and Lindquist, S. (1995). Mutational analysis of Hsp90 function: Interactions with a steroid receptor and a protein kinase. Mol. Cell. Biol. , 15 , 3917– 3925. Nicolet, C.M. and Craig, E. (1989). Isolation and characterization of STI1, a stressinducible gene from Saccharomyces cerevisiae. Mol Cell Biol , 9 , 3638–3646. Ning, Y.-M. and Sanchez, E.R. (1993). Potentiation of glucocorticoid receptor-mediated gene expression by the immunophilin ligands FK506 and rapamycin. J. Biol Chem. , 268 , 6073–6076. Oppermann, H., Levinson, W. and Bishop, J.M. (1981). A cellular protein that associates with the transforming protein of Rous sarcoma virus is also a heat-shock protein. Proc. Natl Acad. Sci. USA , 78 , 1067–1071. Owens-Grillo, J.K., Hoffmann, K., Hutchison, K.A., Yem, A.W., Deibel, Jr., M.R., Handschumacher, R.E. and Pratt, W.B. (1995). The cyclosporin A-binding immunophilin CyP-40 and the FK506-binding immunophilin hsp56 bind to a common site on Hsp90 and exist in independent cytosolic heterocomplexes with the un transformed glucocorticoid receptor . J. Biol. Chem. , 270 , 20479–20484. Owens-Grillo, J.K., Czar, M.J., Hutchison, K.A., Hoffmann, K., Perdew, G.H. and Pratt,
Molecular chaperones and folding catalysts
360
W.B. (1996). A model of protein targeting mediated by immunophilins and other proteins that bind to Hsp90 via tetratricopeptide repeat domains. J. Biol. Chem. , 271 , 13468– 13475. Palmquist, K. Riis, B., Nilsson, A. and Nygard, O. (1994). Interaction of the calcium and calmodulin regulated eEF-2 kinase with heat shock protein 90. FEBS Letters , 349 , 239–242. Perdew, G.H. (1988). Association of the Ah receptor with the 90-kDa heat shock protein. J. Biol. Chem. , 263 , 13802–13805. Picard, D. (1997). The role of heat shock proteins in the regulation of steroid receptor function. In: The Molecular Biology of Steroid and Nuclear Hormone Receptors , L.P.Freedman, ed. Birkhäuser, Boston, pp. 1–24. Picard, D., Khursheed, B., Garabedian, M.J., Fortin, M.G., Lindquist, S. and Yamamoto, K.R. (1990). Reduced levels of Hsp90 compromise steroid receptor action in vivo. Nature , 348 , 166–168. Prapapanich, V., Chen, S., Nair, S.C., Rimerman, R.A. and Smith, D.F. (1996). Molecular cloning of human p48, a transient component of progesterone receptor complexes and an Hsp70-binding protein. Mol. Endocrinol , 10 , 420–431. Pratt, W.B. and Toft, D.O. (1997). Steroid receptor interactions with heat shock protein and immunophilin chaperones. Endocrine Reviews , 18 , 306–360. Privalsky, M.L. (1991). A subpopulation of the v-erbA oncogene protein, a derivative of a thyroid hormone receptor, associates with heat shock protein 90. J. Biol. Chem. , 266 , 1456–1462. Prodromou, C., Roe, S.M., O’Brien, R., Ladbury, J.E., Piper, P.W. and Pearl, L.H. (1997). Identification and structural characterization of the ATP/ADP-binding site in the hsp90 molecular chaperone. Cell , 90 , 65–75. Radanyi, C., Chambraud, B. and Baulieu, E.-E. (1994). The ability of the immunophilin FKBP59-HBI to interact with the 90-kDa heat shock protein is encoded by its tetratricopeptide repeat domain. Proc. Natl. Acad. Sci. USA , 91 , 11197–11201. Rafestin-Oblin, M.-E., Couette, B., Radanyi, C., Lombes, M. and Baulieu, E.-E. (1989). Mineralocorticosteroid receptor of the chick intestine. J. Biol. Chem. , 264 , 9304– 9309. Ratajczak, T. and Carrello, A. (1996). Cyclophilin 40 (CyP-40), mapping of its Hsp90 binding domain and evidence that FKBP52 competes with CyP-40 for Hsp90 binding. J. Biol. Chem. , 271 , 2961–2965. Redeuilh, G., Moncharmont, B., Secco, C. and Baulieu, E.-E. (1987). Subunit composition of the molybdate-stabilized “8–9 S” nontransformed estradiol receptor purified from calf uterus. J. Biol. Chem. , 262 , 6969–6975. Renoir, J.-M., Bihan, S.L., Mercier-Bodard, C., Gold, A., Arjomandi, M., Radanyi, C. and Baulieu, E.-E. (1994). Effects of immunosuppressants FK506 and rapamycin on the heterooligomeric form of the progesterone receptor. J. Steroid Biochem. Molec. Biol. , 48 , 101–110. Renoir, J.-M., Mercier-Bodard, C., Hoffmann, K., Bihan, S.L., Ning, Y.-M., Sanchez, E.R., Handschumacher, R.E. and Baulieu, E.-E. (1995). Cyclosporin A potentiates the dexamethasone-induced mouse mam-mary tumor virus-chloramphenicol acetyltransferase activity in LMCAT cells: A possible role for different heat shock
Control of hormone receptor function by molecular chaperones
361
protein-binding immunophilins in glucocorticosteroid receptor-mediated gene expression. Proc. Natl. Acad. Sci. USA , 92 , 4977–4981. Rose, D.W., Wettenhall, R.E.H., Kudlicki, W., Kramer, G. and Hardesty, B. (1987). The 90-kilodalton peptide of the heme-regulated eIF-2a kinase has sequence similarity with the 90-kilodalton heat shock protein. Biochemistry , 26 , 6583–6587. Sanchez, E.R., Toft, D.O., Schlesinger, M.J. and Pratt, W.B. (1985). Evidence that the 90-kDa phosphoprotein associated with the untransformed L-cell glucocorticoid receptor is a murine heat shock protein. J. Biol. Chem. , 260 , 12398–12401. Sanchez, E.R., Redmond, T., Scherrer, L.C., Bresnick, E.H., Welsh, M.J. and Pratt, W.B. (1988). Evidence that the 90-kDa heat shock protein is associated with tubulincontaining complexes in L cell cytosol and in intact PtK cells. Mol. Endocrinol , 2 , 756–760. Schuh, S., Yonemoto, W., Brugge, J., Bauer, V.J., Riehl, R.M., Sullivan, W.P. and Toft, D.O. (1985). A 90,000-dalton binding protein common to both steroid receptors and the Rous sarcoma virus transforming protein, pp60v-src. J. Biol. Chem. , 260 , 14292– 14296. Schulte, T.W., Blagosklonny, M.V., Ingui, C. and Neckers, L. (1995). Disruption of the Raf-1-Hsp90 molecular complex results in destabilization of Raf-1 and loss of Raf-1Ras association. J. Biol. Chem. , 270 , 24585–24588. Sepehrnia, B., Paz, I.B., Dasgupta, G. and Momand, J. (1996). Heat shock protein 84 forms a complex with mutant p53 protein predominantly within a cytoplasmic compartment of the cell. J. Biol. Chem. , 271 , 15084–15090. Shi, Y., Brown, E.D. and Walsh, C.T. (1994). Expression of recombinant human casein kinase II and recombinant heat shock protein 90 in Escherichia coli and characterization of their interactions. Proc. Natl. Acad. Sci. USA , 91 , 2767–2771. Shue, G. and Kohtz, D.S. (1994). Structural and functional aspects of basic helix-loophelix protein folding by heat-shock protein 90. J. Biol. Chem. , 269 , 2707–2711. Smith, D.F., Schowalter, D.B., Kost, S.L. and Toft, D.O. (1990). Reconstitution of progesterone receptor with heat shock proteins. Mol. Endocrinol. , 4 , 1704–1711. Smith, D.F., Stensgard, B.A., Welch, W.J. and Toft, D.O. (1992). Assembly of progesterone receptor with heat shock proteins and receptor activation are ATP mediated events. J. Biol. Chem. , 267 , 1350–1356. Smith, D.F. (1993). Dynamics of heat shock protein 90-progesterone receptor binding and the disactivation loop model for steroid receptor complexes. Mol. Endocrinol. , 7 , 1418–1429. Smith, D.F., Albers, M.W., Schreiber, S.L., Leach, K.L. and Deibel, Jr., M.R. (1993a). FKBP54, a novel FK506-binding protein in avian progesterone receptor complexes and HeLa extracts. J. Biol. Chem. , 268 , 24270–24273. Smith, D.F., Sullivan, W.P., Marion, T.N., Zaitsu, K., Madden, B., McCormick, D.J. and Toft, D.O. (1993b). Identification of a 60-kilodalton stress-related protein, p60, which interacts with Hsp90 and Hsp70. Mol. Cell. Biol. , 13 , 869–876. Smith, D.F. (1995). Steroid receptors and molecular chaperones. Scientific American Science and Medicine 2 , 38–47. Smith, D.F., Whitesell, L., Nair, S.C., Chen, S., Prapapanich, V. and Rimerman, R.A. (1995). Progesterone receptor structure and function altered by geldanamycin, an
Molecular chaperones and folding catalysts
362
Hsp90-binding agent. Mol. Cell. Biol , 15 , 6804–6812. Stancato, L.F., Chow, Y.-H., Hutchison, K.A., Perdew, G.H.,Jove, R. and Pratt, W.B. (1993). Raf exists in a native heterocomplex with Hsp90 and p50 that can be reconstituted in a cell-free system. J. Biol. Chem. , 268 , 21711–21716. Stancato, L.F., Hutchison, K.A., Krishna, P. and Pratt, W.B. (1996a). Animal and plant cell lysates share a conserved chaperone system that assembles the glucocorticoid receptor into a functional heterocomplex with Hsp90. Biochemistry , 35 , 554–561. Stancato, L.F., Silverstein, A.M., Owens-Grillo, J.K, Chow, Y.H., Jove, R. and Pratt, W.B. (1997). The Hsp90 binding antibiotic geldanamycin decreases Raf levels and epidermal growth factor signalling without disrupting formation of signalling complexes or reducing Raf kinase activity. J. Biol. Chem. , 272 , 4013–4020. Stebbins, C.E., Russo, A.A., Schneider, C., Rosen, N., Hartl, F.U. and Pavletich, N.P. (1997). Crystal structure of an hsp90-geldanamycin complex: Targeting of a protein chaperone by an antitumor agent. Cell , 89 , 239–250. Stepanova, L., Leng, X., Parker, S.B. and Harper, J.W. (1996). Mammalian p50Cdc37 is a protein kinase-targeting subunit of Hsp90 that binds and stabilizes Cdk4. Genes and Development , 10 , 1491–1502. Tai, P.-K.K., Albers, M.W., McDonnell, D.P., Chang, H., Schreiber, S.L. and Faber, L.E. (1994). Potentiation of progesterone receptor-mediated transcription by the immunosuppressant FK506. Biochemistry , 33 , 10666–10671. Tsubuki, S., Saito, Y. and Kawashima, S. (1994). Purification and characterization of an endogenous inhibitor specific to the Z-Leu-Leu-Leu-MCA degrading activity in proteasome and its identification as heat-shock protein 90. FEBS Letters , 344 , 229– 233. Veldscholte, J., Berrevoets, C.A., Zegers, N.B., van der Kwast, T.H., Grootegoed, J.A. and Mulder, E. (1992). Hormone-induced dissociation of the androgen receptor-heat shock protein complex: Use of a new monoclonal antibody to distinguish transformed from nontransformed receptors. Biochemistry 31 , 7422–7430. Wagner, B.J. and Margolis, J.W. (1995). Age-dependent association of isolated bovine lens multicatalytic proteinase complex (proteasome). with heat-shock protein 90, an endogenous inhibitor. Arch. Biochem. Biophys. , 323 , 455–462. Wartmann, M. and Davis, R.J. (1994). The native structure of the activated raf protein kinase is a membrane-bound multi-subunit complex. J. Biol. Chem. , 269 , 6695–6701. Whitelaw, M.L., Hutchison, K. and Perdew, G.H. (1991). A 50-kDa cytosolic protein complexed with the 90-kDa heat shock protein (Hsp90) is the same protein complexed with pp60v-src Hsp90 in cells transformed by the Rous sarcoma virus. J. Biol. Chem. , 266 , 16436–16440. Whitesell, L., Mimnaugh, E.G., De Costa, B., Myers, C.E. and Neckers, L.M. (1994). Inhibition of heat shock protein Hsp90-pp60v-src heteroprotein complex formation by benzoquinone ansamycins: Essential role for stress proteins in oncogenic transformation. Proc. Natl. Acad. Sci. USA 91 , 8324–8328. Whitesell, L. and Cook, P. (1996). Stable and specific binding of heat shock protein 90 by geldanamycin disrupts glucocorticoid receptor function in intact cells. Mol. Endocrinol . 10 , 705–712. Ziemiecki, A., Catelli, M.-G., Joab, I. and Moncharmont, B. (1986). Association of the
Control of hormone receptor function by molecular chaperones
363
heat shock protein Hsp90 with steroid hormone receptors and tyrosine kinase oncogene products. Biochem. Biophys. Res. Comm. , 138 , 1298–1307.
16. ROLE OF CHAPERONES IN UNCOATING OF CLATHRIN COATED VESICLES EVAN EISENBERG and LOIS GREENE* Laboratory of Cell Biology, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland 20892
1. Introduction 2. Overall Cycle of Uncoating by Hsc70 3. Role of Auxilin, a DnaJ Homolog, in Uncoating 3.1. Role of Clathrin Domains in Uncoating 3.2. Direct Interaction of Auxilin with Hsc70 3.3. Mechanism of Action of Auxilin 3.4. Contrast Between Substrates and Products of Hsc70 4. Model of Uncoating 5. Time Course of Uncoating 5.1. Initial Burst of Uncoating 5.2. Slow Steady-state Uncoating 6. Concluding Remarks 7. References 1. INTRODUCTION The Hsp70 class of molecular chaperones and the related constitutive proteins, the Hsc70s, have a widely diverse set of functions in the cell in addition to their role in the heat shock response (see reviews by Hendrick and Hartl, 1993; Rassow and Pfanner, 1995; Hartl, 1996). The Hsc70s are involved in both the folding of proteins after synthesis on ribosomes and in maintaining proteins in an unfolded state after synthesis (chapter Welch et al.). They are also involved in unfolding proteins prior to translocation through mitochondrial or endoplasmic reticulum membranes, in driving the translocation process itself, and then in refolding these proteins following their translocation through the membranes (chapters Haas and Zimmerman; Dekker and Pfanner). *Corresponding author: Building 3, Room B1–22, 3 Center Dr., MSC 0301.
Molecular chaperones and folding catalysts
366
Finally they are involved in the dissociation of protein complexes such as RepA dimers in E. coli (chapter Burkholder and Gottesman) and clathrin coats in mammalian cells and in the formation of protein complexes such as the complex between Hsp90 and steroid receptors in mammalian cells (chapter Toft). Presumably there is a common thread in the mechanism of action of Hsc70 that allows it to carry out such diverse functions. As yet, however, the basic mechanism of action that underlies these varied functions of Hsc70 remains unclear. It has been suggested that the key property of Hsc70 is its ability to bind unfolded regions of proteins in an ATP dependent fashion. However, since attachment and detachment of substrate can occur in an equilibrium system without ATP hydrolysis as well as in a steady-state system driven by ATP hydrolysis, the role of ATP in attachmentdetachment cycles of the various substrates and how it affects the various functions of Hsc70 remains unclear. One of the first constitutive functions discovered for Hsc70 was its ability to uncoat clathrin-coated vesicles in vitro (Schlossman et al., 1984). Monomeric clathrin triskelions with a molecular mass of about 600 kDa, are composed of three identical heavy chains of 190 kDa and three tightly associated light chains (Pearse, 1987; Brodsky, 1988). The triskelion can be structurally and functionally divided into distal and proximal leg sections separated by a flexible hinge region and a globular terminal domain. During receptor-mediated endocytosis clathrin triskelions polymerize and form coats on pits in the plasma membrane that later invaginate into the cell (Keen, 1990; Pearse and Robinson, 1990). Similar clathrin-coated pits also form on the trans-Golgi membrane. In addition to clathrin and receptors, these pits contain assembly proteins or adaptins that serve two functions (Robinson, 1992). They catalyze the polymerization of clathrin triskelions into the planar hexagonal arrays that make up the coat and they bind the receptors that are localized in the clathrin-coated pit. Three major clathrin-assembly proteins have been described (Keen, 1990; Robinson, 1992). AP2 and AP1, which are multimeric subunit complexes of about 270 kDa, occur on the plasma membrane and trans-Golgi membranes, respectively. AP180 is a neuronal specific assembly protein that consists of only one subunit of 120 kDa (Ahle and Ungewickell, 1986; Prasad and Lippoldt, 1988). Following their formation, clathrin-coated pits invaginate, a process that involves the transformation of planar hexagonal clathrin arrays into spherical clathrin coats (Keen et al., 1979; Heuser, 1980; Heuser and Kirchhausen, 1985). This in turn requires the transformation of a number of hexagonal clathrin arrays into pentagonal arrays so that the resulting clathrin-coat has a soccer ball-like structure. As yet it is not clear if clathrin is required for the invagination of the membrane or whether the clathrin coat simply changes its structure as part of the invagination process. In any event, since clathrin coats form spontaneously in vitro (Keen et al., 1979), Rothman and his associates reasoned that removing clathrin from coated vesicles would be an ATP-dependent process and in searching for the enzyme involved, discovered the uncoating ATPase, an enzyme that uncoats clathrin-coated vesicles and artificial clathrin baskets in an ATP-dependent fashion in vitro (Schlossman et al., 1984; Braell et al., 1984). The uncoating ATPase was later found to be a constitutive member of the Hsp70 family (Ungewickell, 1985; Chappell et al., 1986).
Role of chaperones in uncoating of clathrin coated vesicles
367
Although the Hsc70 class of proteins is involved in numerous functions in vivo, these functions are often difficult to study in purified systems in vitro. The uncoating reaction, on the other hand, is highly amenable to study in vitro and, since it is ATP dependent, deciphering its mechanism may lead to important information about the basic processes that underlie all of the ATP-dependent functions of the Hsc70 class of proteins. In addition, numerous questions remain about the role of Hsc70 in endocytosis. Although there is evidence that Hsc70 is involved in uncoating clathrin-coated vesicles in vivo (Honing et al., 1994), it is not clear why Hsc70 does not uncoat clathrin-coated pits (Heuser and Steer, 1989), or whether Hsc70 is involved in the formation and invagination of clathrin-coated pits. Therefore, investigation of the uncoating reaction may yield important information about the mechanism of endocytosis. 2. OVERALL CYCLE OF UNCOATING BY HSC70 The basic properties of the uncoating reaction have been studied using both clathrincoated vesicles and artificial clathrin baskets as a substrate for Hsc70; artificial clathrin baskets are prepared using purified clathrin and various types of assembly proteins. Two observations suggest that uncoating is a specific function of Hsc70. First, this process appears to be ATP-dependent (Schlossman et al., 1984; Schmid et al., 1985; Greene and Eisenberg, 1990), although there have been reports that ATP analogs such at ATPγS and AMP-PNP can also support uncoating (Heuser and Steer, 1989; Buxbaum and Woodman, 1995). Second, the uncoating reaction only occurs with certain Hsc70s, as would be expected for a specific reaction. DnaK shows a much reduced level of uncoating activity (Ungewickell et al. 1997), while there is negligible uncoating activity with BiP, the Hsc70 homolog present in the endoplasmic reticulum (unpublished data). The yeast cytoplasmic Hsc70s, SSA1 and SSA2, do carry out uncoating with the same time course as bovine brain Hsc70 but, particularly with SSA1, considerably more enzyme is required for the same amount of uncoating as occurs with bovine brain Hsc70 (Gao et al., 1991). Interestingly, a similar effect occurs with a truncated bovine brain Hsc70 from which the 10 kDa C-terminal domain has been deleted (Ungewickell et al., 1997). Although earlier studies suggested that this truncated Hsc70 cannot carry out uncoating (Chappell et al., 1987; Tsai and Wang, 1994), Ungewickell et al. (1997) recently demonstrated that it can carry out normal uncoating but only at relatively high concentration. The ATP dependence of the uncoating reaction suggests that Hsc70 is acting catalytically, using the energy of ATP hydrolysis to dissociate polymerized clathrin. This, in turn, suggests that a small amount of Hsc70 should rapidly uncoat a large amount of polymerized clathrin and, indeed, in their preliminary studies Rothman and his associates (Schlossman et al., 1984) reported such behavior. However, more detailed studies on the time course of uncoating of clathrin-coated vesicles showed that this time course is strikingly non-linear, consisting of a rapid initial burst of uncoating followed by very slow steady-state uncoating (Greene and Eisenberg, 1990). Furthermore, at pH 7 this initial burst of uncoating is nearly stoichiometric with three Hsc70s uncoating one clathrin triskelion or three clathrin heavy chains (Greene and Eisenberg, 1990). Later studies showed that the same time course and stoichiometry of uncoating occurs with clathrin baskets prepared with mixed assembly proteins, purified AP2, AP180 (open
Molecular chaperones and folding catalysts
368
triangles, Figure 1), and even myelin basic protein (which acts as an assembly protein) showing that the biphasic time course of uncoating is a general feature of the uncoating reaction (Greene and Eisenberg, 1990; Barouch et al., 1994; Prasad et al., 1994; Prasad et al., 1995). Based on the time course and stoichiometry of uncoating, we proposed a simple model of the uncoating process (Greene and Eisenberg, 1990). In this model Hsc70-
Figure 1 Effect of varying amounts of auxilin on uncoating of clathrin baskets by Hsc70. The time course of clathrin heavy chain (HC) release from baskets consisting of 0.5 M clathrin and 0.5 M AP180, by 0.6 M Hsc70 was measured in the presence of varying concentrations of auxilin: 0 M (●), 0.02 M (□), 0.05 M (■), or 0.10 M (∆) auxilin. Baskets composed of recombinant truncated auxilin consisting of 0.5 M clathrin and 1.5 M recombinant auxilin (▲) were also used as a substrate in the uncoating reaction.
ATP first binds stoichiometrically to the clathrin baskets or vesicles. Then, in association with rapid ATP hydrolysis, a clathrin triskelion complexed with Hsc70-ADP-Pi dissociates from the baskets or vesicles. This accounts for the initial burst of uncoating. The Hsc70-ADP-Pi remains complexed with the clathrin triskelion until the ADP at the active site slowly dissociates, which accounts for the slow steady-state uncoating. Finally ATP rebinds to the Hsc70-clathrin complex, and the monomeric clathrin triskelion rapidly dissociates from the Hsc70 before the bound ATP is hydrolyzed. Since this model was proposed, further studies have confirmed several of its major points, and have shown that these points also apply to the interaction of Hsc70 with other substrates. First, it is now clear that Hsc70 has only one binding site for ATP (Gao et al., 1993). Second, the observation that an initial burst of ATP hydrolysis accompanies the
Role of chaperones in uncoating of clathrin coated vesicles
369
initial burst of uncoating (Barouch et al., 1994) confirms that clathrin baskets and vesicles bind to the ATP form of the enzyme rather than the ADP form, and it is also clear that, in the absence of substrate or cofactors, most of the Hsc70 is in the ATP form (Gao et al., 1993). Similar observations have shown that other substrates of Hsc70 also bind to the ATP form of the enzyme (Prasad et al., 1994, Schmid et al., 1994; Greene et al., 1995; McCarty et al., 1995; Wei et al., 1995; Takeda and McKay, 1996). Third, following the initial burst of uncoating, a complex of Hsc70 and clathrin forms that is stable in ADP and Pi and can be isolated by column chromatography confirming that hydrolysis of ATP to ADP traps the dissociated clathrin triskelions on the enzyme in a complex that is slow to dissociate (Prasad et al., 1994). It has also been shown that other substrates dissociate very slowly from Hsc70-ADP (Palleros et al., 1993, Greene et al., 1995; McCarty et al., 1995; Takeda and McKay, 1996). Finally, it has been shown that both clathrin triskelions and other substrates dissociate very rapidly from Hsc70-ATP (Palleros et al., 1993; Prasad et al., 1994; Greene et al., 1995; McCarty et al., 1995; Takeda and McKay, 1996). Indeed, they dissociate from Hsc70-ATP much more rapidly than ATP hydrolysis occurs, confirming that it is ATP binding rather than ATP hydrolysis that causes dissociation of both clathrin triskelions and other Hsc70 substrates from Hsc70 (Palleros et al., 1993; Prasad et al., 1994; Schmid et al., 1994; Greene et al., 1995; McCarty et al., 1995). Although the simple model presented above explains many aspects of the uncoating reaction, it soon became clear that another factor must be involved in the uncoating reaction because clathrin baskets prepared from highly purified clathrin and assembly proteins cannot be uncoated by Hsc70 alone (Prasad et al., 1993; Barouch et al., 1994; Ungewickell et al., 1995). 3. ROLE OF AUXILIN, A DNAJ HOMOLOG, IN UNCOATING The first evidence that a cofactor is required for Hsc70 to uncoat clathrin-coated vesicles came from studies on the ATPase activity associated with uncoating; to avoid contaminant ATPases, these studies must be carried out with artificial clathrin baskets prepared with purified clathrin and purified assembly proteins (Barouch et al., 1994). When the assembly protein, AP2, was used to prepare AP2-clathrin baskets, it was found that these baskets could no longer be uncoated by Hsc70. Further studies revealed that fractions obtained during purification of AP2 contained a 100 kDa protein cofactor that was required for Hsc70 to carry out the uncoating reaction in ATP (Prasad et al., 1993). The 100 kDa cofactor was later identified as auxilin (Ungewickell et al., 1995), a minor neuronal specific assembly protein first isolated and then more recently cloned by Ungewickell and his associates (Ahle and Ungewickell, 1990; Schroder et al., 1995). The requirement for auxilin appears to be a general one, applying to all types of clathrin baskets, whether prepared with AP2, AP180 (Figure 1), or even myelin basic protein (Prasad et al., 1993; Barouch et al., 1994; Ungewickell et al., 1995; Prasad et al., 1995). In its ability to support uncoating by Hsc70, auxilin resembles members of the DnaJ family of proteins that have been shown to act as partner proteins with Hsc70 in a number of the activities carried out by Hsc70 including protein folding, protein translocation, and formation and dissociation of protein complexes (Cyr et al., 1994; Rassow et al., 1995).
Molecular chaperones and folding catalysts
370
The DnaJ family of proteins is characterized by having a specific J domain usually at the N-terminal region of the protein (Caplan et al., 1993). When the sequence of auxilin was examined, it was indeed found to have a J-domain but at its C-terminal rather than at its N-terminal end. Overall, the J-domain of auxilin is 50–60% similar to the J-domains of E. coli and yeast DnaJ proteins (Ungewickell et al., 1995). Further evidence that this Jdomain is crucial for the activity of auxilin came from studies using recombinant auxilin that lacked the J-domain. While intact recombinant auxilin acts exactly like auxilin isolated from bovine brain, the recombinant auxilin that lacks the J-domain no longer supports uncoating but does bind to clathrin baskets (Ungewickell et al., 1995; Holstein et al., 1996). Interestingly, the N-terminal domain of auxilin shares similarity with tensin (Schroder et al., 1995) and this domain can be deleted resulting in much higher expression of the truncated recombinant auxilin (Holstein et al., 1996). This truncated auxilin still carries out normal uncoating and if its J-domain is deleted, like auxilin without a J-domain, the truncated molecule still binds to clathrin baskets but does not carry out uncoating (Holstein et al., 1996). Therefore, auxilin appears to be a DnaJ homolog with three domains: a C-terminal J-domain, a clathrin-binding domain, and an N-terminal tensin domain (Schroder et al., 1995; Holstein et al., 1996). 3.1. Role of Clathrin Domains in Uncoating The observation that auxilin is required for Hsc70 to uncoat clathrin baskets shed new light on the mechanism of the uncoating reaction itself, in particular on the role of the clathrin light chains and the N-terminal domains of clathrin in the uncoating reaction. It had been postulated that both the N-terminal domains of clathrin and the clathrin light chains were required for Hsc70 to uncoat clathrin baskets (Schmid et al., 1984; Schmid and Rothman, 1985). However, earlier studies on auxilin had suggested that auxilin can bind to clathrin baskets in which the light chains had been removed (Lindner and Ungewickell, 1991). This led to a reinvestigation of the involvement of the clathrin light chains and terminal domains in the uncoating process. Surprisingly, it was found that, when auxilin was present, clathrin baskets prepared from clathrin that lacked light chains and/or terminal domains, were uncoated normally by Hsc70 (Ungewickell et al., 1995). These results strongly suggest that neither the light chains nor terminal domains of clathrin are required for uncoating by Hsc70. The fact that the terminal domains are not required confirms the earlier observation of Prasad et al., (1994) that, during uncoating, Hsc70 acts by binding to the vertex of clathrin. The fact that light chains are not required for uncoating is consistent with the experiments of Acton et al. (1993) in which they found that reduced levels of clathrin light chains in the cell did not affect endocytosis and secretion. There is, however, direct interaction between one of the classes of clathrin light chains (LCa) and Hsc70 as measured by the ability of high concentrations of LCa to activate the Hsc70 ATPase and to inhibit clathrin uncoating by Hsc70 (DeLuca-Flahtery et al., 1990). This interaction could perhaps subtly alter clathrin uncoating by Hsc70 (DeLuca-Flahtery et al., 1990; Anton et al., 1993).
Role of chaperones in uncoating of clathrin coated vesicles
371
3.2. Direct Interaction of Auxilin with Hsc70 Since auxilin appears to be a DnaJ homolog, it was of interest to determine whether it had the same effects on Hsc70 as several other DnaJ homologs that have been studied. These other DnaJ homologs activate the Hsc70 ATPase rate by increasing the rate of the ATP hydrolysis step but do not affect the rate of ADP release (McCarty et al., 1995; King et al., 1995b). They also catalytically induce polymerization of Hsc70 in an ATP dependent reaction (King et al., 1995a). Qualitatively, auxilin behaves similarly in its interaction with Hsc70. It causes a 5-fold increase in the Hsc70 ATPase activity and a corresponding increase in the level of bound ADP during steady-state ATP hydrolysis showing that it increases the rate of the ATP hydrolysis step, but does not increase the rate of ADP release from Hsc70 (Jiang et al., 1997). It also induces polymerization of Hsc70 in an ATP-dependent reaction that is reversed by hydrolysis of ATP to ADP (Jiang et al., 1997); the polymerized Hsc70 can be detected by column chromatography although it appears that it does not sediment (Holstein et al., 1996). Interestingly, the J-domain of auxilin also induces Hsc70 to polymerize in an ATP-dependent reaction (Jiang et al., 1997), supporting the view that auxilin is a DnaJ homolog that interacts with Hsc70 through its C-terminal J-domain. However, in contrast to other DnaJ homologs that bind to Hsc70 quite weakly and act catalytically (King et al., 1995a, King et al., 1997b), auxilin binds to the Hsc70 polymer in a 1 to 1 molar ratio, with a dissociation constant of 0.6 M. The observation that this binding is reversed with a half-life of about 4 min when ATP is hydrolyzed to ADP (Jiang et al., 1997) provides direct evidence for the view that DnaJ homologs only interact with Hsc70 in ATP. 3.3. Mechanism of Action of Auxilin Given the observation that auxilin behaves like other DnaJ homologs in its direct effects on Hsc70, it was of interest to determine whether auxilin acts by inducing the binding of clathrin-coated vesicles to Hsc70 just as other DnaJ homologs present substrates to Hsc70. It is difficult to investigate this question at pH 7 because Hsc70 becomes essentially inactive after a single round of uncoating. However, at pH 6, although uncoating is almost completely inhibited, an auxilin-dependent steady-state interaction occurs between Hsc70 and clathrin baskets. This was demonstrated in a study showing that clathrin baskets and auxilin together cause more than 100-fold activation of the steady-state Hsc70 ATPase activity at pH 6 while neither clathrin baskets nor auxilin alone significantly activate the Hsc70 ATPase activity (Barouch et al., 1997). At saturating auxilin, half-maximal Hsc70 ATPase activity occurs at 0.2 M clathrin baskets suggesting that, under these conditions, auxilin induces a strong complex to form between Hsc70 and clathrin baskets. This was confirmed by direct binding studies using sedimentation to measure complex formation between Hsc70 and clathrin baskets at pH 6. Formation of this complex is both auxilin and ATP dependent (Ungewickell et al., 1995; Barouch et al., 1997); it does not occur in ADP, and when the ATP is hydrolyzed to ADP the Hsc70 dissociates from the baskets with a half-life of about 6 min (Barouch et al., 1997). This latter observation shows that, following ATP hydrolysis, the Hsc70-clathrin basket complex is not
Molecular chaperones and folding catalysts
372
energetically stable; it is metastable since it spontaneously dissociates. Taken together the studies at pH 6 strongly suggest that auxilin acts by inducing clathrin-coated vesicles to bind to Hsc70 in an ATP-depend-ent reaction, and therefore auxilin functions like other DnaJ homologs that present substrates to Hsc70. It is not yet clear whether DnaJ homologs act catalytically or stoichiometrically when they present substrates to Hsc70. Although auxilin has been reported to bind to clathrin triskelions at a 3 to 1 molar ratio (Ahle and Ungewickell, 1990), studies of uncoating at pH 7 showed that both the rate and the magnitude of uncoating saturate at less than a 1 to 5 molar ratio of auxilin to both clathrin triskelions in the baskets and Hsc70 (Figure 1). Even when the auxilin is increased to a 3 to 1 molar ratio to clathrin by preparing clathrin baskets with auxilin, alone, as an assembly protein, neither the rate nor magnitude of uncoating increases (closed triangles, Figure 1). Therefore, auxilin acts catalytically rather than stoichiometrically at pH 7, which raises the question of whether auxilin also acts catalytically when it induces binding of Hsc70 to clathrin baskets at pH 6. Studies at pH 6 showed that auxilin preferentially binds to clathrin baskets rather than Hsc70. These studies also showed that auxilin acts substoichiometrically when it activates the steadystate Hsc70 ATPase activity with maximum ATPase activation occurring at a molar ratio of 1 auxilin per clathrin triskelion in the baskets (Barouch et al., 1997). Even less auxilin is required to induce maximal binding of Hsc70 to clathrin baskets. Maximal binding occurs at about a 1 to 3 molar ratio of auxilin to clathrin triskelions in the clathrin baskets at pH 6 (Barouch et al., 1997) and pH 6.5 (Ungewickell et al., 1995) confirming that auxilin acts catalytically when it induces binding of Hsc70 to clathrin baskets at pH 6. This, in turn, shows that auxilin must induce binding of Hsc70 to clathrin baskets without itself becoming part of the resulting complex. The question then arises as to whether other DnaJ homologs can also catalytically induce the binding of clathrin baskets to Hsc70. Investigation of this question showed that neither the J-domain of auxilin by itself (Holstein et al., 1996) or other DnaJ homologs such HSJ1, YDJ1, and HDJ1 (Cheetham et al., 1996; King et al., 1997b) support uncoating. In fact, these other DnaJ homologs inhibit uncoating, and they do so under conditions where they either do not induce Hsc70 to polymerize (Cheetham et al., 1996) or cause only partial polymerization of Hsc70 (King et al., 1997b) suggesting that they are not inhibiting uncoating simply by polymerizing Hsc70 before it can interact with clathrin baskets. Rather, these homologs may inhibit uncoating by inducing hydrolysis of ATP to ADP before Hsc70-ATP can bind to the auxilin-clathrin basket complex. Therefore, auxilin may specifically support uncoating because it juxtaposes the Hsc70 and clathrin baskets allowing the clathrin baskets to form a transient complex with Hsc70-ATP just before auxilin activates hydrolysis of ATP to ADP. This view of auxilin action is similar to a model of DnaJ action proposed by Bukau and his collaborators (McCarty et al., 1995). It is also supported by the observation that the activity of auxilin can be partially restored by cross-linking the J-domain with the clathrin-binding domain of auxilin, neither of which is active by itself (Holstein et al., 1996). These results suggest that if other DnaJ homologs act like auxilin they would also present only specific substrates to Hsc70, rather than non-specifically presenting all substrates to Hsc70.
Role of chaperones in uncoating of clathrin coated vesicles
373
3.4. Contrast Between Substrates and Products of Hsc70 Studies on the auxilin-induced binding of clathrin baskets to Hsc70 reveal major differences in the way Hsc70 interacts with the substrate of the uncoating reaction, clathrin baskets, and the product of the uncoating reaction, free clathrin triskelions. First, although auxilin is required for the binding of clathrin baskets to Hsc70 in ATP it is not required for the binding of free clathrin triskelions. Second, the auxilin-induced binding of clathrin baskets to Hsc70 in ATP is an order of magnitude stronger than the binding of clathrin triskelions (Barouch et al., 1997; Prasad et al. 1994). Third, auxilin only induces clathrin baskets to bind to Hsc70 in the presence of ATP (Ungewickell et al., 1995) while clathrin triske ions bind stably to both Hsc70-ADP (Prasad et al., 1994) and nucleotidefree Hsc70 (Gao et al., 1995a). Finally mutations of Hsc70 that have a marked effect on the auxilin-induced binding of clathrin baskets to Hsc70 have little effect on the binding of clathrin triskelions (Rajapandi et al., 1998). Therefore, Hsc70 interacts very differently with clathrin baskets, the substrate of the uncoating reaction, and free clathrin triskelions, the product of the uncoating reaction. The distinction between the way Hsc70 interacts with clathrin baskets and free clathrin triskelions may have major implications for studies on the interaction of other ligands with Hsc70. Generally, ligands such as peptides or reduced carboxymethylated lactalbumin whose interaction with Hsc70 is affected by ATP have been considered substrates of Hsc70 (Palleros et al., 1991; Palleros et al., 1993). This may be misleading, however, because in their ability to bind to Hsc70-ATP in the absence of DnaJ homologs and in their stable binding to Hsc70-ADP these ligands behave more like clathrin triskelions than clathrin baskets, that is more like products than substrates of Hsc70. It may be that only ligands that can be presented to Hsc70 by DnaJ homologs in ATP are true substrates of Hsc70. 4. MODEL OF UNCOATING Based on our current understanding of the mechanism of action of auxilin, we have developed a model for the uncoating reaction (Figure 2). In this model, the initial burst of uncoating starts with auxilin binding to clathrin baskets. The auxilin then induces Hsc70ATP to bind to clathrin baskets and this is immediately followed byauxilin inducing rapid hydrolysis of the ATP bound to Hsc70. Note that in this model, auxilin activates ATP hydrolysis by Hsc70 only after formation of the Hsc70-ATP-clathrin basket complex. If auxilin activated ATP hydrolysis before this complex formed, the resulting Hsc70-ADP complex would not bind to the clathrin baskets since auxilin does not induce binding of clathrin baskets to Hsc70-ADP. Our model suggests that, following ATP hydrolysis and dissociation of auxilin, three Hsc70-ADPs are bound to a clathrin triskelion in the clathrin basket. Studies at pH 6 suggest that this is a metastable complex that, with time, would completely dissociate (Barouch et al., 1997). On the other hand, Hsc70-ADP does form an energetically stable complex with dissociated clathrin triskelions (Prasad et al.,
Molecular chaperones and folding catalysts
374
Figure 2 Model for the uncoating of clathrin-coated vesicles by Hsc70. The ATP-bound and ADP-bound forms of Hsc70 are shown by the white and black rectangles, respectively.
1994). Therefore, the power stroke of the uncoating reaction may be a transformation of the metastable Hsc70-ADP-clathrin basket complex to the energetically stable Hsc70ADP-dissociated clathrin complex. More generally, a similar transformation from a metastable Hsc70-ADP-substrate complex to a stable Hsc70-ADP-product complex could act as the power stroke for other reactions involving DnaJ homologs such as the translocation of proteins through membranes and the unfolding of misfolded proteins. Since the clathrin lattice is much more stable at pH 6 than at pH 7 (Van Jaarsveld et al., 1981), the transformation from the metastable Hsc70-ADP-clathrin basket complex to the energetically stable Hsc70-ADP-dissociated clathrin complex may be unable to occur at pH 6, which may explain why uncoating is strongly inhibited at pH 6. However, at pH 7, once all three legs of a clathrin triskelion form stable complexes with Hsc70-ADP, the clathrin will dissociate from the clathrin lattice. Since three Hsc70s are required to dissociate a single clathrin triskelion in this model, the model implies that Hsc70 should act cooperatively in the uncoating process but evidence for such cooperativity has not yet been observed. Since Hsc70-ADP dissociates very slowly from dissociated clathrin triskelions, our model suggests that, in the last step of the uncoating cycle, ADP is released and ATP rebinds to the Hsc70s complexed with the dissociated clathrin triskelion, thereby freeing up the Hsc70 to carry out further cycles of uncoating. In our original model of uncoating we proposed that this slow step was the rate-limiting step in the uncoating reaction and could therefore account for the very slow steady-state uncoating (Greene and Eisenberg, 1990). However, in ATP the rate of dissociation of clathrin from the Hsc70-ADP-Pi-
Role of chaperones in uncoating of clathrin coated vesicles
375
clathrin complex is too fast to account for the very slow steady-state uncoating rate (Prasad et al., 1994). Hence, there must be some other explanation for the slow steadystate uncoating. 5. TIME COURSE OF UNCOATING Studies on the time course of uncoating were directed to answering two key questions that are not explained by our model of uncoating. First, we investigated why auxilin acts catalytically, that is why the rate of the initial burst of uncoating saturates at very low levels of auxilin. Second, we investigated the cause of the slow steady-state uncoating. 5.1. Initial Burst of Uncoating Using purified AP180-clathrin baskets plus auxilin, it has been demonstrated that an initial burst of ATP hydrolysis occurs with the same time course as the initial burst of uncoating with three ATPs hydrolyzed per clathrin triskelion uncoated (Barouch et al., 1994). It therefore appears that the initial burst of uncoating represents a single pre-steady-state turnover of ATP by Hsc70. If the rate-limiting step in this single turnover occurred after Hsc70 bound to the clathrin baskets, e.g. if it were the ATP hydrolysis step itself, it would be easy to explain the ability of auxilin to act catalytically since low levels of auxilin could induce Hsc70 to bind to the clathrin baskets much faster than Hsc70 could uncoat them. However, the observation that most of the Hsc70 remains free in solution during the initial burst of uncoating (unpublished data) suggests that it is binding of Hsc70 to clathrin baskets that is rate-limiting, not a step that occurs after binding of Hsc70. This view was confirmed by showing that the rate of the initial burst of uncoating linearly increases with an increase in the Hsc70 concentration even when the Hsc70 concentration is in excess over the clathrin basket concentration (unpublished data); this could not occur if the rate-limiting step in the uncoating reaction occurred after Hsc70 bound to the clathrin baskets. If the binding of Hsc70 to clathrin baskets is the rate-limiting step in the uncoating reaction, simple kinetics predicts that the rate of the initial burst of uncoating should not only linearly increase with an increase in Hsc70 concentration but should also linearly increase with an increase in clathrin basket concentration or the amount of auxilin bound to the clathrin baskets. However, as we pointed out above, the rate of uncoating levels off at very low auxilin concentration (Figure 1). Furthermore, recent experiments showed that the rate of uncoating also does not increase with increasing concentrations of clathrin baskets (unpublished data). One explanation for this unusual kinetic phenomenon is that the rate of the initial burst
Molecular chaperones and folding catalysts
376
Figure 3 Recent findings on the uncoating cycle. This model alters the model in Figure 2 to include, first, a rate-limiting conformational change in the Hsc70-ATP form, depicted by the transition from the white to the grey rectangle, that occurs prior to the binding of Hsc70 to the baskets; and, second, formation of a steady-state complex between Hsc70, clathrin and assembly proteins after the initial burst of uncoating, which inhibits further uncoating by Hsc70.
of uncoating is controlled by a conformational change that must occur in the Hsc70 before it can bind to the clathrin baskets as illustrated in Figure 3. In this case, increasing the auxilin concentration or the clathrin basket concentration would not increase the rate of uncoating because most of the Hsc70 would be unable to bind to the clathrin baskets until the Hsc70 underwent the rate-limiting conformational change. If, indeed, Hsc70 must undergo a conformational change before it interacts with auxilin and clathrin baskets, it could have major implications for the general mechansim of action of the Hsp70 class of proteins. 5.2. Slow Steady-state Uncoating Detailed studies have ruled out several explanations for the slow steady-state uncoating (Gao et al., 1995b). First, when fresh Hsc70 is added to partially uncoated baskets, a second round of uncoating takes place showing that the partially uncoated baskets are still capable of being uncoated. Second, radioactively labeled clathrin does not rebind to clathrin baskets during slow steady-estate uncoating, ruling out that clathrin is still being released during slow steady-state uncoating but because the clathrin exchanges with clathrin previously released during the initial burst of uncoating, no net uncoating occurs. Therefore, the slow steady-state uncoating is due to an inability of Hsc70 in the
Role of chaperones in uncoating of clathrin coated vesicles
377
supernatant to carry out further uncoating. However, this Hsc70 is not irreversibly altered; its uncoating activity can be completely restored by chromatographing the supernatant of the uncoating reaction on a Superose 6 or 12 FPLC column that separates Hsc70 from clathrin and assembly proteins (Gao et al., 1995b). Evidence that assembly proteins are involved in causing the slow steady-state uncoating comes from studies on highly purified clathrin baskets prepared in calcium in the absence of assembly proteins (Gao et al., 1995b). Auxilin is required for the uncoating of these baskets, although less auxilin than is required for the uncoating of clathrin baskets prepared with assembly proteins. Strikingly, however, the time course of uncoating of these pure clathrin baskets is not biphasic. Rather, a nearly constant rate of uncoating occurs until the clathrin baskets are completely uncoated. When assembly proteins are added to the pure clathrin baskets, the time course of uncoating again becomes biphasic confirming that assembly proteins are indeed involved in causing the slow steady-state uncoating. However, a simple mixture of depolymerized clathrin triskelions and assembly proteins does not inhibit uncoating (Jiang et al., 1998). One explanation for the slow steady-state uncoating is that clathrin and assembly proteins are released as a special complex from clathrin baskets during the uncoating reaction, and Hsc70 preferentially binds to this complex, thus preventing Hsc70 from carrying out further uncoating as illustrated in Figure 3. If indeed, Hsc70 interacts with uncoated complexes of free clathrin triskelia and assembly proteins, it may reflect a physiological role for Hsc70 either in preventing these complexes from inappropriately polymerizing in the cytosol or in rapidly catalyzing the formation of clathrin-coated pits from these complexes in nerve cells. In fact, we have recently obtained evidence that these complexes do indeed interact with uncoated vesicle membranes in vitro (Jiang et al., 1998). In addition, in vivo, a stable non-membranous Hsc70-clathrin complex forms in neurons and is transported along the axon (De Waegh and Brady, 1989; Black et al., 1991), and over-expression of auxilin in HeLa cells causes it to form large complexes with clathrin and Hsc70 (Zhao et al., 1998). Interestingly, this complex formation actually inhibits endocytosis in these cells, apparently because the clathrin is inappropriately sequestered away from normal clathrin-coated pits on the plasma membrane (Zhao et al., 1998). 6. CONCLUDING REMARKS The uncoating of clathrin-coated vesicles provides an excellent model system for studying the role of DnaJ homologs and ATP in the mechanism of action of mammalian Hsc70. In addition, the role of Hsc70 and auxilin in receptor-mediated endocytosis is of interest in its own right. It is now clear that uncoating requires the DnaJ homolog, auxilin, as well as Hsc70 and ATP. Auxilin contains both a J-domain and a clathrin-binding domain so it not only interacts directly with Hsc70 but also with clathrin-coated vesicles. Perhaps because other DnaJ homologs do not contain a clathrin binding domain, they cannot substitute for auxilin in inducing uncoating, suggesting that auxilin and perhaps other DnaJ homologs, as well, only present specific substrates to Hsc70 rather than nonspecifically presenting all substrates to Hsc70. There is good evidence that auxilin functions by first inducing formation of a complex
Molecular chaperones and folding catalysts
378
between Hsc70-ATP and clathrin-coated vesicles and then inducing rapid ATP hydrolysis. Since clathrin baskets do not spontaneously bind to Hsc70-ADP in the presence of auxilin, this leads to formation of a metastable Hsc70-ADP-clathrin basket complex. However, Hsc70-ADP does stably bind to dissociated clathrin triskelions. It is therefore possible that a transition from a metastable Hsc70-ADP-clathrin basket complex to a stable Hsc70-ADP-dissociated clathrin complex provides the power stroke for the uncoating reaction; a similar transition might provide the power stroke in other reactions where Hsc70s in combination with DnaJ homologs utilize the free energy of ATP hydrolysis to do work. This mechanism of action of Hsc70 is based on the observation that Hsc70 interacts very differently with the substrate of the uncoating reaction, clathrin baskets, and the product of the uncoating reaction, clathrin triskelions. Interestingly, other ligands such as peptides and reduced carboxymethylated lactalbumin resemble clathrin triskelions rather than clathrin baskets in their interaction with Hsc70 suggesting that they may act more like products than substrates of Hsc70. One of the key characteristics of the uncoating reaction is its biphasic time course, which consists of a rapid initial burst of uncoating followed by very slow steady-state uncoating. The initial burst of uncoating shows an unusual kinetic phenomenon in which the binding of Hsc70 to clathrin baskets is the rate-limiting step in the initial burst of uncoating yet this rate is independent of both the clathrin basket concentration and the auxilin concentration. A posible explanation for this kinetic phenomenon is that Hsc70 must undergo a conformational change before it binds to clathrin baskets. If so, it will be important to determine if Hsc70 undergoes a similar conformational change in other reactions that it carries out. As for the slow steady-state uncoating, current evidence suggests that it is due to an interaction of Hsc70 with special clathrin-assembly protein complexes produced only during uncoating; this interaction apparently prevents Hsc70 from carrying out further uncoating. It is possible that interaction of Hsc70 with uncoated clathrin-assembly protein complexes in vitro reflects a role for Hsc70 in regulating clathrin dynamics in vivo. Furthermore, since auxilin is a nerve-specific protein, such a phenomenon might play a role in the rapid endocytosis that De Camilli and Takei (1996) have suggested is associated with synaptic transmission. We are currently investigating if auxilin-like proteins support uncoating in nonnervous system tissues and are also attempting to determine why uncoating occurs with varying efficiencies with clathrin-coated vesicles from different tissues in vitro (Buxbaum and Woodman, 1995; Buxbaum and Woodman, 1996). Interestingly, a recently cloned protein, GAK, contains both an N-terminal kinase domain (Kanaoka et al., 1997) and a C-terminal domain highly homologous to auxilin. Recently we have determined that this latter domain acts exactly like auxilin in vitro in regard to its interaction with Hsc70 and clathrin baskets (Greener et al., 1998). Furthermore, unlike auxilin, this protein shows a wide tissue distribution raising the possibility that it could function like auxilin in non-neuronal tissues. However, this protein is also associated with both cyclin G and CDK5, and, therefore its full function in vivo is not yet understood (Kanaoka et al., 1997; Kimura et al., 1997). Future studies will be required to determine if this protein indeed acts like auxilin in non-neuronal tissues and also to determine the role of its kinase domain in vivo.
Role of chaperones in uncoating of clathrin coated vesicles
379
7. REFERENCES Able, S. and Ungewickell, E. (1990). Auxilin, a newly identified clathrin-associated protein in coated vesicles from bovine brain. J. Cell Biol. , 111 , 19–29. Able, S. and Ungewickell, E. (1986). Purification and properties of a new clathrin assembly protein. EMBO J. , 5 , 3143–3149. Acton, S., Wong, D. Parhma, P, Brodsky, F. and Jackson, A. (1993). Alteration of clathrin light chain expression by transfection and gene disruption. Mol. Biol. Cell , 4 , 647–660. Barouch, W., Prasad, K., Greene, L.E. and Eisenberg, E. (1994). ATPase activity associated with the uncoating of clathrin baskets by Hsp70. J. Biol. Chem. , 269 , 28563–28568. Barouch, W., Prasad, K., Greene, L. and Eisenberg, E. (1997). Auxilin induced interaction of the molecular chaperone Hsc70 with clathrin baskets. Biochemistry , 36 , 4304–4308. Black, M., Chestnut, M., Pleasure I. and Keen, J. (1991). Stable clathrin: uncoating protein complexes in intact neurons and their axonal transport. J. Neurosci. and , 11 , 1163–1172. Braell, W.A., Schlossman, D.M., Schmid, S.L. and Rothman, J.E. (1984). Dissociation of clathrin coats coupled to the hydrolysis of ATP: role of an uncoating ATPase. J. Cell Biol. , 99 , 734–741. Brodsky, F.M. (1988). Living with clathrin: its role in intracellular membrane traffic. Science , 242 , 1396–1402. Buxbaum, E. and Woodman, P.G. (1995). Selective action of uncoating ATPase towards clathrin-coated vesicles from brain. J. Cell Sci. , 108 , 1295–1306. Buxbaum, E. and Woodman, P. (1996). The speed of partial reactions of the uncoating ATPase Hsc70 depends on the source of coated vesicles. J. Cell Sci. , 109 , 705–711. Caplan A.J., Cyr, D.M. and Douglas, M.G. (1993). Eukaryotic homologues of Escherichia coli dnaJ: A diverse protein family that functions with Hsp70 stress proteins. Mol. Biol. Cell , 4 , 555–563. Chappell, T.G., Konforti B.B., Schmid S.L. and Rothman, J.E. (1987). The ATPase core of a clathrin uncoating protein. J. Biol. Chem. , 262 , 746–751. Chappell, T.G., Welch, W.J., Schlossman, D.M., Palter, K.B., Schlesinger, M.J. and Rothman, J.E. (1986). Uncoating ATPase is a member of the 70 kilodalton family of stress proteins. Cell , 45 , 3–13. Cheetham, M.E., Anderton, B.H. and Jackson, A.P. (1996). Inhibition of hsc70-catalysed clathrin uncoating by HSJ1 proteins. Biochem J. , 319 , 103–108. Cyr, D.M., Langer, T. and Douglas, M.G. (1994). DnaJ-like proteins: molecular chaperones and specific regulators of Hsp70. Trends Biochem. Sci. , 19 , 176–181. De Camilli, P. and Takei, M. (1996). Molecular mechanisms in synaptic vesicle endocytosis and recycling. Neuron , 16 , 481–486. DeLuca-Flaherty, C., McKay, D., Parham, P. and Hill, B. (1990). Uncoating protein (hsc70) binds a conformationally labile domain of clathrin light chain LCa to stimulate ATP hydrolysis. Cell , 62 , 857–887.
Molecular chaperones and folding catalysts
380
De Waegh, S. and Brady, S. (1989). Axonal transport of a clathrin uncoating ATPase: a role for Hsc70 in the modulation of coated vesicle assembly in vivo. J.Neurosci. Res. , 4 , 433–440. Gao, B., Biosca, J., Craig, E.A., Greene, L.E. and Eisenberg, E. (1991). Uncoating of coated vesicles by yeast Hsp70 proteins. J. Biol. Chem. , 266 , 19565–19571. Gao, B., Emoto, Y., Greene, L. and Eisenberg, E. (1993). Nucleotide binding properties of bovine brain uncoating ATPase. J. Biol. Chem. , 268 , 8507–8513. Gao, B., Eisenberg, E. and Greene, L. (1995a). Interaction of nucleotide-free Hsc70 with clathrin and peptide and effect of ATP analogues. Biochemistry , 34 , 11882–11888. Gao, B., Prasad, K., Greene, L. and Eisenberg, E., (1995b). Uncoating of clathrin baskets by Hsp70: cause of inhibition following one round of uncoating. Mol. Cell Biol. , 6 , 411a. Greene, L.E. and Eisenberg, E. (1990). Dissociation of clathrin from coated vesicles by the uncoating ATPase. J. Biol. Chem. , 265 , 6682–6687. Greene, L.E., Zinner, R., Naficy, S. and Eisenberg, E. (1995). Effect of nucleotide on the binding of peptides to 70-kDa heat shock protein. J. Biol. Chem. , 270 , 2967–2973. Greener, T., Nojima, H., Eisenberg, E. and Greene, L. (1998). Cyclin G associated kinase (GAK) has kinase and auxilin domains that are active in vitro. FASEB J. , 12 , in press. Hartl, F. (1996). Molecular chaperones in cellular protein folding. Nature , 381 , 571– 580. Hendrick, J.P. and Hartl, F.U. (1993). Molecular chaperone functions of heat-shock proteins. Annu. Rev. Biochem. , 62 , 349–384. Heuser, J. (1980). Three-dimensional visulazation of coated vesicle formation in fibroblasts. J. Cell Biol. , 84 , 560–583. Heuser, J. and Kirchhausen, T. (1985). Deep-etch views of clathrin assemblies. J. Ultrastruct. Res. , 92 , 1–27. Heuser, J. and Steer, C.J. (1989). Trimeric binding of the 70-kDa uncoating ATPase to the vertices of clathrin triskelia: a candidate intermediate in the vesicle uncoating reaction. J. Cell Biol. , 109 , 1457–1466. Holstein, S.E.H., Ungewickell, H. and Ungewickell, E. (1996). Mechanism of clathrin basket dissociation: separate functions of protein domains of the dnaJ homologue auxilin. J. Cell Biol. , 135 , 925–937. Honing, S., Kreimer, G., Robenek, H. and Jockusch, B. (1994). Receptor-mediated endocytosis is sensitive to antibodies. J. Cell Set. , 107 , 1185–1186. Jiang, R., Gao, B., Eisenberg, E. and Greene, L. (1998). An Hsc70-clathrin complex formed during uncoating may be involved in rebinding of clathrin to uncoated clathrin vesicles. FASEB J. , 12 , in press. Jiang, R.-F., Greener, T., Barouch, W., Greene, L. and Eisenberg, E. (1997). Interaction of auxilin with the molecular chaperone, Hsc70. J. Biol. Chem. , 272 , 6141–6145. Kanaoka, Y., Kimura, S., Okazaki, I.Ikeda, M. and Nojima, H. (1997). GAK: a cyclin G associated kinase contains a tensin/auxilin-like domain. FEBS Lett , 402 , 73–80. Keen, J.H., Willingham, M.C. and Pastan, I. (1979). Clathrin coated vesicles: Isolation dissociation and factor-dependent reassociation of clathrin baskets. Cell , 16 , 303–312. Keen, J.H. (1990). Clathrin and associated assembly and disassembly proteins. Annu. Rev. Biochem. , 59 , 415–438.
Role of chaperones in uncoating of clathrin coated vesicles
381
Kimura, S., Tsuruga, H. Yabuta, N. Endo, Y. and Nojima, H. (1997). Structure, expression and chromosomal localization of human GAK. Genomics , 44 , 179–187. King, C., Eisenberg, E. and Greene, L. (1995a). Polymerization of 70-kDa heat shock protein by yeast DnaJ in ATP. J. Biol. Chem. , 270 , 22535–22540. King, C., Eisenberg, E. and Greene, L. (1995b). Polymerization of Hsp70 by DnaJ in the presence of ATP. Mol. Cell Biol. , 67a , 391. King, C., Eisenberg, E. and Greene., L. (1997a). Interaction between Hsc70 and DnaJ homologs: evidence for weak binding of DnaJ homologs to Hsc70. Mol Bio. Cell , 8 , 96a. King, C., Eisenberg, E. and Greene, L. (1997b). Effect of yeast and human DnaJ homologs on clathrin uncoating by 70 kDa-Heat Shock Protein. Biochemistry , 36 , 4067–4073. Lindner, G. and Ungewickell, E. (1991). Light-chain-independent binding of adaptors, AP180 and auxilin to clathrin. Biochemistry , 30 , 9097–10101. McCarty, J.S., Buchberger, A., Reinstein, J. and Bukau, B. (1995). The role of ATP in the functional cycle of the DnaK chaperone system. J. Mol. Biol. , 249 , 126–137. Palleros, D.R., Welch, W.J. and Fink, A.L. (1991). Interaction of Hsp70 with unfolded proteins: effects of temperature and nucleotides on the kinetics of binding. Proc. Natl. Acad. Sci. USA , 88 , 5719–5723. Palleros, D.R., Reid, K.L., Shi, L., Welch, W.J. and Fink, A.L. (1993). ATP-induced protein-Hsp 70 complex dissociation requires K+ but not ATP hydrolysis. Nature , 365 , 664–666. Pearse, B.M.F. (1987). Clathrin and coated vesicles. EMBO J. , 6 , 2507–2512. Pearse, B.M. and Robinson, M.S. (1990). Clathrin, adaptors and sorting. Annu. Rev. Cell Biol , 6 , 151–171. Prasad, K., Barouch, W., Greene, L. and Eisenberg, E. (1993). A protein cofactor is required for uncoating of clathrin baskets by uncoating ATPase. J. Biol. Chem. , 268 , 23758–23761. Prasad, K., Barouch, W., Martin, B.M., Greene, L.E. and Eisenberg, E. (1995). Purification of a new clathrin assembly protein from bovine brain coated vesicles and its identification as myelin basic protein. J. Biol. Chem. , 270 , 30551–30556. Prasad, K., Heuser, J., Eisenberg, E. and Greene, L. (1994). Complex formation between clathrin and uncoating ATPase. J. Biol. Chem. , 269 , 6931–6939. Prasad, K. and Lippoldt, R. (1988). Molecular characterization of the AP180 assembly protein. Biochemistry , 27 , 6098–6104. Rajapandi, T., Wu, C., Eisenberg, E. and Greene, L. (1998). Characterization of DIOS and K71E mutants of human cytosolic Hsp70. Biochemistry , 37 , in press. Rassow, J. and Pfanner, N. (1995), Molecular chaperones and intracellular protein translocation. Rev. Physiol. Biochem. Pharmacol. , 126 , 199–264. Rassow, J., Voos, W. and Pfanner, N. (1995). Partner proteins determine multiple functions of Hsp70. Trends Cell Biol , 5 , 207–212. Robinson, M.S. (1992). Adaptins. Trends Cell Biol , 2 , 293–297. Schmid, D., Baici, A., Gehring, H. and Christen, P. (1994). Kinetics of molecular chaperone action. Set. , 263 , 971–973. Schmid, S.L., Braell, W.A. and Rothman, J.E. (1985). ATP catalyzes these questration of
Molecular chaperones and folding catalysts
382
clathrin during enzymatic uncoating. J. Biol. Chem. , 260 , 10057–10062. Schmid, S.L., Braell, W.A., Schlossman, D.M. and Rothman, J.E. (1984). A role for clathrin light chains in the recognition of clathrin cages by uncoating ATPase. Nature , 311 , 228–231. Schmid, S.L. and Rothman, J. (1985). Two classes of binding sites for uncoating protein in clathrin triskelions. J. Biol. Chem. , 260 , 10050–10056. Schlossman, D.M., Schmid, S.L., Braell, W.A. and Rothman, R.E. (1985). An enzyme that removes clathrin coats: purification of an uncoating ATPase. J. Cell Biol . 99 , 723–733. Schroder, S., Morris, S., Knorr, R., Plessmann, U., Weber, K., Vinh, N. and Ungewickell, E. (1995). Primary structure of the neuronal clathrin-associated protein auxilin and its expression in bacteria. Eur. J. Biochem. , 228 , 297–304. Takeda, S. and McKay, D.B. (1996). Kinetics of peptide binding to the bovine 70 kDa heat shock cognate protein, a molecular chaperone. Biochemistry , 35 , 4636–4644. Tsai, M. and Wang, C. (1994). Uncoupling of peptide-stimulated ATPase and clathrinuncoating activity in deletion mutation of hsc70. J. Biol. Chem. , 269 , 5958–5962. Ungewickell, E. (1985). The 70-kd mammalian heat shock proteins are structurally and functionally related to the uncoating protein that releases clathrin triskelia from coated vesicles. EMBO J. , 4 , 3385–3391. Ungewickell, E., Ungewickell, H., Holstein, S., Lindner, R., Prasad, K., Barouch, W., Martin, B., Greene, L. and Eisenberg, E. (1995). Role of auxilin in uncoating clathrincoated vesicles. Nature , 378 , 632–635. Ungewickell, E., Ungewickell, H. and Holstein, S., (1997). Functional interaction of the auxilin J domain with the nucleotide-and substrate-binding modules of hsc70 . J. Biol. Chem. , 272 , 19594–19600. Van Jaarsveld, P., Nandi, P., Lippoldt, R., Saroff, H. and Edelhoch, H. (1981). Polymerization of clathrin protomers into basket structures. Biochemistry , 20 , 4129– 4135. Wei, J., Gaut, J.R. and Hendershot, L.M. (1995). In vitro dissociation of BiP-peptide complexes requires a conformational change in BiP after ATP binding but does not require ATP hydrolysis. J. Biol. Chem. , 44 , 26677–26682. Zhao, X., Greener, T., Greene, L. and Eisenberg, E. (1998). Expression of the clathrin assembly protein, auxilin, in tissue culture cells inhibits endocytosis. FASEB J. , 12 , in press.
17. THE ROLE OF HSP104 IN STRESS TOLERANCE AND PRION MAINTENANCE SUSAN LINDQUIST* and ERIC C.SCHIRMER Howard Hughes Medical Institute and the Department of Molecular Genetics and Cell Biology, The University of Chicago, 5841 S. Maryland Avenue, MC 1028, Chicago, IL 60637, USA
1. Introduction 2. The HSP100/CLP Family of Proteins 2.1. A Large Family of Proteins with Diverse Biological Activities 2.2. Hsp104 Protein Structure 3. Biochemical Properties of Hsp104 3.1. Oligomerization 3.2. ATPase Activity 4. The Role of Hsp104 in Stress Tolerance 4.1. Thermotolerance and the Importance of Hsp104 4.2. The Thermotolerance Function of Hsp104 4.2.1. No Evidence for a Proteolysis Function 4.2.2. A Disassembling/resolubilizing Activity 4.3. Other Stresses 4.4. Combinatorial Stresses 4.5. Physiological Inductions of Tolerance 4.6. Conservation of Tolerance Functions in B-Type Proteins 4.7. The Relationship Between Hsp104 and Other Tolerance Factors 4.7.1. Other Tolerance Factors 4.7.2. Genetic Interactions Between Hsp70 and Hsp104 4.7.3. Genetic Interactions Between Trehalose Synthase and Hsp104 5. Hsp104’s Interaction with a Yeast Prion 5.1. A Role for Hsp104 During Normal Growth 5.2. The Prion Hypothesis 5.3. The Relationship Between [PSI+] And Sup35 5.4. The Role of Hsp104 in the Maintenance of [PSI+] 5.4.1. Overexpression of HSP104 Cures [PSI+] 5.4.2. Deletion of HSP104 Cures [PSI+] 5.4.3. Hsp104 and Sup35 Aggregation in Vivo
The role of Hsp104 in stress tolerance and prion maintenance
385
5.4.4. Seeded Polymerization of Sup35 5.4.5. Hsp104 and Sup35 Interaction in Vitro 5.5. A Model For Hsp104-[PSI+] Interactions 5.6. An Explanation For Previous Puzzles 5.6.1. Hsp104 Explains Puzzling [PSI+] phenomena 5.6.2. [PSI + ] Explains a Puzzling Thermotolerance Phenomenon 5.7. Broader Implications 5.8. Hsp104 and Mammalian Prions 6. Concluding Remarks 7. References
1. INTRODUCTION The gene encoding Hsp104 from Saccharomyces cerevisiae was first cloned in an effort to decipher the mechanisms that cells employ to survive a variety of toxic conditions (Sanchez and Lindquist, 1990). Hsp104 is expressed at a low level under most growth conditions, but it is strongly induced by a variety of environmental stresses (Figure 1 and Sanchez et al., 1992). The protein proved to play a key role in helping cells survive many of these stresses, particularly under the most extreme conditions (Sanchez and Lindquist, 1990; Sanchez et al., 1992; Lindquist et al., 1995). Recent work indicates that, unlike many other Hsp chaperones, Hsp104 has little or no capacity to prevent the aggregation of denatured proteins. Nor does it promote the forward folding of denatured proteins. Rather, Hsp 104 helps to disentangle partially aggregated, stress damaged proteins so that they can be refolded by other chaperone systems in the cell (Parsell et al., 1994b; Glover and Lindquist, 1998). Hsp104 was independently cloned in a series of experiments aimed at understanding the mysterious mechanism of inheritance of the genetic element known as [PSI + ] (Chernoff et al., 1995). Recent genetic and biochemical evidence indicates that [PSI + ] is a prion-like element (Wickner, 1994; Chernoff et al., 1995; Patino et al., 1996). The inheritance of the [PSI + ] phenotype is due to the inheritance of an altered physical state of the normal cellular protein, Sup35. Once the altered state has been established, it is self-perpetuating in wild-type cells. Hsp 104 was cloned in a screen for proteins whose overexpression suppresses the [PSI + ] phenotype. The role of Hsp104 in this remarkable and novel mechanism of inheritance is related to its role in thermotolerance—that is, its ability to alter the conformational state of other proteins. The relationship between Hsp 104 and thermotolerance, however, is simpler than the relationship between Hsp104 and [PSI + ]. As the concentration of Hsp104 increases, the ability of cells to survive high temperatures increases, in a direct relationship (Lindquist and Kim, 1996). In contrast, the ability of cells to inherit the [PSI + ] factor depends upon the ratio of Hsp 104 to Sup35. Cells are cured of [PSI + ] if they express either too much
Molecular chaperones and folding catalysts
386
or too little Hsp104 (Chernoff et al., 1995). Determining the precise molecular mechanisms that dictate the role of Hsp 104 in stress tolerance and prion maintenance represents a major challenge for future investigations.
Figure 1 Hsp104 expression patterns. (A) Patterns of protein synthesis in yeast cells. Cells grown to mid-log phase in glucose media at 25°C were labeled for 30 min with 3H-isoleucine at 25°C (C, control cells) or 15 min after a shift to 39°C (H, heatshocked cells). Total cellular proteins were separated by electrophoresis in SDS gels and label incorporation was visualized by fluorography. (B) Accumulation of Hsp104 in response to stress. Cells grown to mid-log phase in glucose media at 25°C were maintained at 25°C (C), heat shocked at 39°C for 30 min (H), or incubated at 25°C in the presence of 6% ethanol (EtOH) for 1 hr, 0.75M sodium arsenite for 1 hr (As), or 1.5 mM cadmium chloride for 30 min (Cd). Proteins were separated by electrophoresis in SDS gels, transferred to membranes, and reacted with an antibody directed against Hsp104. Immune complexes were visualized by incubation with 125I-labeled protein A followed by fluorography. (C) Accumulation of Hsp104 under specific, normal physiological conditions. Left: cells were grown to mid-log phase at 25°C in synthetic glucose medium (S-Glu) or synthetic acetate medium (S-Ac) and maintained at 25°C (C) or heat shocked at 39°C for 30 min. Middle: cells grown
The role of Hsp104 in stress tolerance and prion maintenance
387
at 25°C in glucose medium were collected at various stages in the growth curve, from mid-log phase to stationary phase. Right: diploid cells were sporulated in low nitrogen media treated with zymolyase and mature spores were collected on Percoll gradients. Hsp104 protein was visualized as in B. Data taken from (Sanchez et al., 1992).
2. THE HSP100/CLP FAMILY OF PROTEINS 2.1. A Large Family of Proteins with Diverse Biological Activities Hsp104 belongs to a newly discovered family of chaperones with diverse patterns of expression, subcellular localizations, and biological functions, known as the
Figure 2 Structural features of the HSP100 family. Thicker boxes indicate greater conservation and particularly conserved sequence motifs unique to the family are indicated I–V. Abrupt transitions in conservation demarcate the two ATP-binding domains, the aminoterminal, middle, and carboxyl-terminal regions. The X and Y proteins are more distantly related and their single ATP-binding domain is depicted as lighter in color. The locations of the Walkertype nucleotide-binding motifs, A and B, are indicated. (It is not clear if both B motifs in NBD1 are used.) Class 1 proteins contain two NBDs. Class 2 proteins contain only one NBD and the carboxylterminal region. The M and N subfamilies of Class 2 are distinguished from the X and Y subfamilies by much greater homology to the Class 1 proteins. To the right are listed the percent similarity between the ATP-binding domains of representatives of different subfamilies or the overall percent similarities. The numbers listed were determined using the DNAStar
Molecular chaperones and folding catalysts
388
program with the Clustal method and PAM250 residue weight table. For class 2 HSP100 proteins, just the NBD2 and C-terminal domains were used for the overall analysis. The proteins representative of each subfamily used in this analysis are (A) Escherichia coli ClpA, (B) E. coli ClpE, (C) Tomato ClpC, (D) Ambidopsis ERD1, (M) mouse SKD2, (N) Pseudomonas aeruginosa AmiB, (X) E. coli ClpX, (Y) E. coli ClpY. A list of sequence motifs characteristic of the different domains, together with database accession numbers for members of the family cloned as of July 1996, can be found in Schirmer et al., 1996; updates appear on Web site: “http://http.bsd.uchicago.edu/~hsplab/index.html”.
HSP100/Clp family. (For a more detailed review of the family, see Schirmer et al., 1996). The distinct subtypes of the HSP100 family characterized to date are depicted in Figure 2. Several members are induced by stress (the B, D, and X subtypes, and some members of the C subtype); others are constitutively expressed. In eukaryotic cells, different members are found in the nuclear/cytosolic compartment(s), in mitochondria, and in chloroplasts. At least four different subtypes (the A, B, X, and Y subtypes, see Figure 2) are found in Escherichia coli (Hwang et al., 1987; Katayama-Fujimura et al., 1987; Squires et al., 1991; Chuang et al., 1993; Gottesman et al., 1993; Wojtkowiak et al., 1993). The ClpA protein of E. coli was the first member of this broad family of proteins to have its biochemical function defined: it is an ATP-regulated subunit of the ClpP protease (Hwang et al., 1987; Katayama-Fujimura et al., 1987). (The name derives from caseinolytic protease). ClpX also functions in proteolysis in conjunction with ClpP (Gottesman et al., 1993; Wojtkowiak et al., 1993). The substrate specificity of ClpP is determined by its association with ClpA or ClpX (Gottesman et al., 1993; Wojtkowiak et al., 1993; Wawrzynow et al., 1995). In the absence of the ClpP protease, ClpX and ClpA alter the conformational states of their substrates. In particular, they promote the disassembly of these substrates from higher ordered structures (Wickner et al., 1994; Wawrzynow et al., 1995). These observations argue that the function of ClpA and ClpX in proteolysis is not simply to determine the substrate specificity of the protease, but also to change the conformational state of the substrate and make it more accessible to digestion by ClpP. The complete sequence of the Saccharomyces cerevisiae genome is known and three members of the HSP100 family are present. Two of these are of the B subtype: Hsp104, found in the nucleus and cytoplasm, and Hsp78, in mitochondria. The third is of the X subtype. This protein may be mitochondrial, but it has not been established. The absence of an obvious ClpP homolog in S. cerevisiae (determined by a homology search of the genome) suggests the yeast ClpX protein may function solely to alter the conformational state of its substrates, rather than to connect them to proteolysis, but its specific functions have not yet been assigned. Members of the HSP100 family have diverse biological functions, but recent work suggests that these functions are unified by a common biochemical mechanism (Schirmer et al., 1996). All of the proteins investigated at the biochemical level to date (including members of the A, B, and X subtypes) employ ATP to promote changes in the
The role of Hsp104 in stress tolerance and prion maintenance
389
conformations of other proteins and/or their oligomeric state. In some proteins, this function is coupled to proteolysis, in others to gene regulation, DNA transposition, prion maintenance or stress tolerance. Stress tolerance seems to be a particular specialization of proteins of the B subtype. Bproteins from a very broad range of organisms are induced by stress and function in stress tolerance. These include both B-pro teins of the fungi S. cerevisiae, Hsp104 and Hsp78 (Sanchez and Lindquist, 1990; Schmitt et al., 1996), the B proteins of the bacterium E. coli, ClpB (Squires et al., 1991), of the trypanosomatid Leishmania major, Hsp100 (Hubel et al., 1995), of the land plants Arabidopsis thaliana (Schirmer et al., 1994) and Glycine max (Lee et al., 1994), Hsp101, of the cyanobacterium Synechococcus sp PCC7942, ClpB (Eriksson and Clarke, 1996), and of the closely related C-subtype protein of Bacillus subtilis, ClpC (MecB) (Kruger et al., 1994). It is striking that both Band C-type HSP100s are encoded in the genome of the hyperthermophilic bacterium Aquifex aeolicus, which grows at temperatures up to 95°C (Deckert et al., 1998). Their roles in thermal adaptation, however, have not yet been tested. This review will concentrate on the functions of Hsp104, the only HSP100 B protein whose biological and biochemical functions have been explored in detail. The reader is directed to other recent reviews for a discussion of proteins of other HSP100 subtypes (Gottesman et al., 1995; Schirmer et al., 1996; Maurizi et al., this volume). 2.2. Hsp104 Protein Structure As shown in Figure 2, based on sequence analysis, the Hsp104 protein can be divided into 5 domains. Whether these will correspond to structural domains in the folded protein is currently under investigation. Two of the domains contain highly conserved nucleotide-binding motifs known as the Walker A (P-loop) and Walker B consensus elements. These elements are embedded within larger blocks of highly conserved sequence shared with other members of the family. Although each of these larger domains is itself highly conserved, the two domains have little sequence in common, except for the few residues that constitute the Walker motifs. Thus, the first nucleotide binding domain of Hsp104, NBD1, shares 60% amino-acid identity with the NBD1 of E. coli ClpB and the second nucleotide-binding domain of Hsp104, NBD2, shares 62% identity with the NBD2 of E. coli ClpB, but NBD1 and NBD2 of Hsp104 share only 22% identity. The NBDs of Hsp104 are separated by a middle domain of ~180 amino acids, previously termed the “spacer” region, because of high sequence and size variability in different subfamilies. However, recent analysis indicates that in the ClpB subfamily this region contains conserved sequence motifs (Schirmer et al., 1996). Moreover, dominant Hsp104 mutations cluster in this region (Schirmer and Lindquist, unpublished). These observations suggest that this middle domain has an important biological function. Like the middle region, the N-terminal and C-terminal domains of Hsp104 are less conserved than the NBDs. However, small stretches of common amino acids can be discerned and these are more highly conserved within members of a particular subfamily (Schirmer et al., 1996). Additionally, the C-terminal regions of HSP100 proteins show some similarity to PDZ domains (Levchenko et al., 1997), which function in other
Molecular chaperones and folding catalysts
390
proteins to mediate protein:: protein interactions. Sequence motifs IV and V (Figure 2) are encompassed within this PDZ-like region. A fragment of the E. coli ClpX protein containing these motifs bound a ClpX proteolytic substrate in an immobilized ELISA assay, indicating this region provides an important function in substrate recognition. It is unclear whether these motifs function in this manner in Hsp104 and other HSP100 proteins, but proteins carrying deletions in the C-terminal domain of Hsp104 do not function in yeast thermotolerance (Schirmer and Lindquist, unpublished). Thus, the Cterminal regions of Hsp104 and of HSP100 proteins in general have important and until recently unappreciated biological functions. 3. BIOCHEMICAL PROPERTIES OF HSP104 3.1. Oligomerization Hsp104 assembles into hexameric ring-shaped particles in the presence of adenine nucleotides, a property it shares with the E. coli ClpA protein (Parsell et al., 1994a; Kessel et al., 1995). In the absence of nucleotides, Hsp104 assembles into hexamers at high protein concentrations (Schirmer, Queitsch, Kowal and Lindquist, unpub-lished). Identical single amino-acid substitutions in consensus nucleotide-binding residues of the two ATP-binding domains have different effects on the ability of Hsp104 to form these particles. In all P-loop containing nucleotide-binding proteins with solved structures, including both ATP- and GTP-binding proteins, a conserved lysine contacts the and phosphates of bound nucleotide (Saraste et al., 1990). A threonine to lysine substitution in this residue of NBD1 (the K218T mutation) has little effect on oligomerization, but the same substitution in NBD2 (K620T) has a strong detrimental effect on oligomerization (Parsell et al., 1994a). 3.2. ATPase Activity Another property that Hsp104 shares with other HSP100 proteins (all tested members, including representatives of the A, B, N, and X subfamilies) is the ability to bind and hydrolyze ATP (Woo et al., 1992; Maurizi et al., 1994; Wawrzynow et al., 1995; Wilson et al., 1995). Hsp104 cleaves ATP with a Vmax of ~2 nmole Pi min-1 g-1 (~20 ATP molecular per hexamer s-1) and binds ATP with a Km of ~5 mM (Schirmer et al., 1998). As with oligomerization, the K218T and K620T mutations have different effects on this activity. The K218T substitution eliminates or severely reduces ATPase activity, while the K620T substitution reduces ATPase activity several fold. However, the ATPase activity of K620T returns to near wild-type levels when the oligomerization defect of the protein is overcome by high protein concentrations (Schirmer, Queitsch, Kowal and Lindquist, unpublished). Thus, it appears that NBD1 is responsible for most of the ATPase activity of the protein, while nucleotide interactions in NBD2 influence that activity by affecting the structural state of the protein (Schirmer et al., 1998). As is the case with several other HSP100 proteins, the ATPase activity of Hsp104 is stimulated by peptides. A few peptides such as -insulin stimulate ClpA, ClpB and Hsp104 (Hwang et al., 1988; Woo et al., 1992; Maurizi et al., 1994; Schirmer and
The role of Hsp104 in stress tolerance and prion maintenance
391
Lindquist, unpublished); however most of the peptides which stimulate Hsp104 are different from those that stimulate ClpA (Maurizi et al., 1994; Schirmer and Lindquist, unpublished). The strongest stimulation observed with Hsp104 to date is with polylysine, which increases ATP hydrolysis seven-fold. Interestingly, biologically relevant peptides with very different amino-acid compositions can either stimulate or inhibit ATP hydrolysis by Hsp104 (Schirmer and Lindquist, 1997). One possibility is that different sites for peptide interaction exist in Hsp104, at which peptide binding has different biochemical consequences. 4. THE ROLE OF HSP104 IN STRESS TOLERANCE 4.1. Thermotolerance and the Importance of Hsp104 In all organisms, exposure to mildly elevated temperatures induces tolerance to more severe temperatures (see Li et al., this volume). The magnitude of this effect can be astonishing. In Saccharomyces cerevisiae, for example, a 30 min pretreatment
Figure 3 The role of Hsp104 in stress tolerance.
Molecular chaperones and folding catalysts
392
(A) Wild-type cells and hsp104 mutants die at similar rapid rates when shifted directly from 25°C to 50°C. (B) Pre-treatments at 37°C for 30 min induce thermotolerance in both mutant and wild-type cells (compare with A), but after a few minutes at 50° C, hsp104 mutants die at ~100 times the rate of wild-type cells. After about 20 minutes at 50°C, the hsp104 mutants die at ~10,000 times the rate of wild-type cells. (C) Mild heat pre-treatments induce tolerance to ethanol but hsp104 mutants die at ~1,000 times the rate of wild-type cells. (D) Mild heat pre-treatments induce tolerance to arsenite, but hsp104 mutants die at similar rates to wild-type cells. (E) Mild heat pre-treatments induce tolerance to cadmium, but hsp104 mutants die at the same rate as wild-type cells. (F) Acetate-grown cells are more tolerant to a direct shift from 25°C to 50°C than glucose grown cells (compare with A), and much of this tolerance is lost in hsp104 mutants. (G) Spores are much more tolerant of heat than log phase cells (compare to A and F), but hsp104 spores are much less tolerant than wild-type spores. Method: Wild-type cells and hsp104 deletion mutants exposed to various stresses for the indicated times were plated on rich glucose media to determine the number of viable cells (colony forming units) remaining. Prior to the stress, cells were grown at 25°C to mid-log phase in glucose medium (A-E), acetate medium (F), or sporulated in low nitrogen medium also at 25°C (G). Data taken from (Sanchez et al., 1992).
The role of Hsp104 in stress tolerance and prion maintenance
393
Figure 4 Hsp104 expression is sufficient for tolerance. Cells exposed to 50°C for 4 min were serially diluted, 5-fold at each step, and spotted onto rich glucose agar to determine the number of colony forming units remaining. Left: wild-type cells were maintained at 25°C prior to the heat shock (C) or preheated at 37°C for 30 min (H). Right: hsp104 deletion mutants transformed with plasmids that allow different levels of Hsp104 induction in response to exogenously added hormone (Louvion et al., 1993) were incubated in -estradiol at the indicated concentrations for two hours prior to heat shock. Data taken from (Lindquist and Kim, 1996).
at 37°C increases the ability of cells to survive a severe 50°C heat shock by nearly 10,000-fold (Sanchez and Lindquist, 1990; Sanchez et al., 1992). When log-phase cells carrying hsp104 deletions are grown at 25°C and shifted to 50°C, they die at the same rapid rate as wild-type cells (Figure 3A). When pretreated at 37°C, however, they behave very differently from wild-type. Although the mutants display some tolerance at 50°C, it quickly dissipates. Within 10 min, mutant cells start to die at 100 to 1,000 times the rate of wild-type cells (Figure 3B). Thus, Hsp104 is not the only factor involved in thermotolerance, but it plays a vital role. When total yeast proteins were analyzed by high resolution 2-D gels, no differences could be detected in the accumulation of any proteins, other than Hsp104 itself, either at high or low temperatures (Parsell et al., 1993). This strongly suggests that Hsp104 affects thermotolerance directly, rather than by altering the expression of other tolerance factors. Hsp104’s role in thermotolerance is specific to extreme conditions. Hsp104 deletions grow as well as wild-type cells, not only at normal temperatures, but even at the highest temperatures in the yeast growth range (Sanchez and Lindquist, 1990; Sanchez et al., 1992). They also survive as well as wild-type cells when held for two days at
Molecular chaperones and folding catalysts
394
temperatures just beyond the growth range (Sanchez and Lindquist, 1990). At 44°C, where cells take many hours to die, mutant cells die at 10 times the rate of wild-type cells (Sanchez et al., 1992). At 55°C, where cells survive for only a few minutes, hsp104 mutants die at 1,000 to 10,000 times the rate of wild-type cells (Blake and Linquist, unpublished). Such experiments demonstrate that Hsp104 is important in stress tolerance, but give little indication of how central its role might be. Disrupting just one element in a web of 50 equally vital factors could produce a severe reduction in survival. The true importance of Hsp104 became apparent in experiments where its expression was divorced from that of other tolerance factors, by placing HSP104 coding sequences under the control of heterologous promoters (Lindquist and Kim, 1996). When Hsp104 is selectively induced to a modest level it increases survival several fold. More remarkably, when it is expressed at a high level, Hsp104 provides as much tolerance as a mild heat pretreatment, at least for a short severe heat shock (Figure 4 and Lindquist and Kim, 1996). 4.2. The Thermotolerance Function of Hsp 104 4.2.1. No Evidence for a Proteolysis Function The homology between Hsp 104 and ClpA, with its known proteolysis function, suggested that Hsp 104 might provide thermotolerance by promoting the proteolysis of irretrievably heat-denatured proteins. That is, it might increase survival by preventing protein aggregates from ‘clogging up’ the cell. Experiments using previously established assays for detecting proteolytic deficits in yeast suggest this is not the case. Hsp104 deletion mutants 1) release previously incorporated amino acids into the media at the same rate as wild-type cells, 2) show similar rates of -galactosidase digestion, using the unstable N-end rule variants engineered by the Varshavsky laboratory, 3) grow on plates containing amino-acid analogs at least as well as wild-type cells, and 4) show no detectable accumulation of novel proteins by high resolution 2-D gels (Parsell et al., 1993; Parsell and Lindquist, unpublished; Taulien and Lindquist, unpublished). Although negative evidence is never conclusive, it seems very unlikely that a 100- to 1,000-fold difference in survival after heat shock could be based on a level of proteolysis that is undetectable by these methods. 4.2.2. A Disassembling/Resolubilizing Activity Rather than promoting proteolysis, Hsp 104 seems to function in stress tolerance by promoting the resolubilization and reactivation of proteins that have been denatured by heat and have begun to aggregate. Such a function was unexpected; it was generally believed that aggregated proteins are irretrievably lost. As surprising as such a function for Hsp104 may be, it is now supported by several different lines of evidence. One line of evidence involves in vivo analysis of a specific heat-damaged substrate. The critical heat-sensitive lethal targets in yeast are unknown. However, a heterologous temperature-sensitive luciferase protein provides a model heat-sensitive substrate that denatures at temperatures which do not affect viability, even in the hsp104 mutants. Luciferase activity is lost at the same rate in mutant and wild-type cells, and in both cell
The role of Hsp104 in stress tolerance and prion maintenance
395
types the enzyme is found in an aggregated state. Remarkably, when cells are returned to normal temperatures, luciferase is resolubilized and reactivated in wild-type cells, but not in cells lacking Hsp104 (Parsell et al., 1994b). Another line of evidence involves direct visualization of heat-induced damage by electron microscopy. It takes several hours to kill yeast cells at 44°C. However, after just one hour, when both wild-type and hsp104 mutant cells retain greater than 95% viability, extensive damage is already apparent. Wild-type and hsp104 mutant cells are indistinguishable: both display large masses of electron-dense material (presumably protein aggregates) in their nuclei and cytoplasms (Parsell et al., 1994b). During two hours of recovery at normal temperatures, this material completely disappears from wildtype cells, but not from the mutants. Since there appear to be no differences in proteolysis in wild-type and mutant cells, the disappearance of these aggregates is most likely due to resolubilization. A third line of evidence involves RNA splicing, one of the most heat-sensitive cellular processes known (Yost and Lindquist, 1986; Yost and Lindquist, 1988). When whole cells are shifted directly to moderately high temperatures, splicing is disrupted while transcription and translation continue. Splicing is equally affected in wild-type and mutant cells. Moreover, mild pretreatments protect splicing from disruption as effectively in mutant cells as in the wild-type cells (Yost and Lindquist, 1991). However, when splicing is disrupted by a short, sub-lethal heat shock and cells are returned to normal temperatures for recovery, it is restored much more rapidly in wild-type than in the mutant cells. Thus, Hsp104 does not protect splicing from disruption, but helps to reactivate splicing once it has been disrupted. In vitro experiments also indicate Hsp104’s function is to repair damage rather than prevent it. When splicing-competent cell-free extracts are heated, they are rapidly inactivated. If Hsp104 is present during the heat treatment, it does not prevent inactivation. However, if purified Hsp104 is added after inactivation, it at least partially restores splicing (Vogel et al., 1995). Hsp104 also has no capacity to prevent the aggregation of chemically denatured enzymes upon dilution into aqueous buffers. Nor does it promote the refolding of these substrates directly. Rather, Hsp104 helps Ydj1 and Hsp70 promote the activation of these enzymes by an ATP-dependent mechanism (Glover and Lindquist, 1998). Taken together, these data indicate that the primary function of Hsp104 is to mediate the repair of substrates damaged by heat, rather than to protect those substrates from damage. Heat is a protein denaturant. Thus, Hsp104’s molecular function provides an explanation of its biological role in protecting cells from exposure to extreme heat— when the rate of protein denaturation outpaces the ability of other protective systems to prevent aggregation. The biochemical properties of Hsp104 described earlier are clearly related to these functions. The single-amino acid substitutions in NBD1 and NBD2 that inhibit ATP hydrolysis (Schirmer et al., 1998) or oligomerization (Parsell et al., 1994a) do not promote the resolubilization of heat-damaged luciferase in vitro (Glover and Lindquist, 1998) or in vivo (Parsell et al., 1994b), and are ineffective in promoting survival after a severe heat shock (Parsell et al., 1991).
Molecular chaperones and folding catalysts
396
4.3. Other Stresses A striking and universal feature of stress tolerance is that one type of mild stress induces tolerance to many others. Thus, mild heat treatments induce tolerance not only to higher temperatures, but also to ethanol, anoxia, heavy metal ions, free radicals, UV light, and a host of other damage-causing agents. (See Watson, 1990; Nover, 1991; Weber, 1992; Parsell and Lindquist, 1993 for review.) For many years it was not clear whether the same or different mechanisms were responsible for tolerance to all of these stresses. Genetic analysis of Hsp104 provided one of the first tools for dissecting this problem. As shown in Figure 1, Hsp104 is induced by a wide variety of stressful conditions. It is one of the most prominent proteins induced by heat (Figure 1A) and is induced to a similar level by ethanol, sodium arsenite, and cadmium (Figure 1B). Hsp104 is just as important in protecting cells from ethanol as from heat (Figure 3C and Sanchez et al., 1992). Indeed these stresses exhibit complete cross tolerance: mild heat pretreatments protect against damage by ethanol, mild ethanol pretreatments protect against heat shock (Sanchez et al., 1992). Moreover, as with heat, the importance of Hsp104 in ethanol tolerance is greater under more severe conditions. It has little effect on survival when cells are incubated with 12% ethanol, but promotes survival by 10,000-fold when they are incubated with 20% ethanol. The expression of Hsp104 gives cells a modest survival advantage upon exposure to sodium arsenite, but this is apparent only with arsenite pretreatments and not with heat pretreatments (Figure 3D and Sanchez et al., 1992). Moreover, Hsp104 provides no advantage upon exposure to cadmium (Figure 3E) or copper, over a broad range of toxic concentrations and, in some cases, actually reduces survival (Sanchez et al., 1992, and Taulien and Lindquist, unpublished), whether Hsp104 is pre-induced by heat (Figure 3E), or by a low concentration of the metal ion itself (Sanchez et al., 1992). In the light of what we now know about Hsp104’s molecular functions and the nature of cadmium and copper toxicity, this makes perfect sense. Heat and ethanol are protein denaturants, and Hsp104 helps to restore denatured proteins to the native state (Parsell et al., 1994b). Cadmium and copper, however, cause covalent oxidative damage that requires degradation of the damaged protein (Stohs and Bagchi, 1995). Therefore, Hsp104 is not involved. Curiously, although Hsp104 is induced by DNA damaging agents, such as UV and MMS, it consistently reduces the rate of survival during exposure to radiation (Boreham and Mitchel, 1994)—that is, hsp104 mutant cells survive better than wild-type cells. The molecular basis of this phenomenon is unknown. The complex and overlapping stressresponse networks of the cell induce a broad range of proteins in response to many stresses (Ruis and Schuller, 1995) but, clearly, some proteins are not always beneficial to the particular stress encountered. 4.4. Combinatorial Stresses The fact that the functions of Hsp104 in stress tolerance are most important under extreme conditions calls into question its general relevance. It seems unlikely that yeast cells will commonly find themselves exposed to 50°C in their natural environment.
The role of Hsp104 in stress tolerance and prion maintenance
397
However, it seems equally unlikely that cells will encounter a single type of stress while all other conditions remain optimal. Single stresses reduce the number of variables in laboratory experiments, but they are hardly a realistic mimic of natural conditions. The synergistic effects of combined stresses is apparent in cells exposed to two stresses which themselves cause no loss in viability, 6% ethanol and 40°C for 2 hours. In combination, these two stresses kill 99.99% of the cells, unless the cells have been given a mild pretreatment and contain a wild-type copy of the HSP104 gene (Lindquist et al., 1995). In that case, most of the cells survive. Thus, Hsp104 is a vital tolerance factor for lethal combined stresses, a situation more likely to be encountered in the natural environment. 4.5. Physiological Inductions of Tolerance In addition to stress, Hsp104 is induced in response to certain normal conditions of growth and development. For example, it is expressed at a high basal level when cells are grown on carbon sources that require high rates of respiration, such as acetate (Figure 1C). Such cells survive a direct shift from 25°C to 50°C much better than glucose-grown cells (Sanchez et al., 1992) and this basal tolerance is greatly reduced in hsp104 mutants (Figure 3F). Because yeast cells normally switch into respiratory metabolism when their preferred carbon source, glucose, has been fermented to ethanol, this induction may be part of a developmentally regulated program to ensure that Hsp104 is present to protect against ethanol toxicity. The protein is also induced in glucose-grown cells a few generations before stationary phase, at the time of the diauxic shift. Stationary phase cells exhibit a much higher level of stress tolerance than log-phase cells and, again, this is largely dependent upon Hsp104. Finally, Hsp104 is strongly induced during sporulation and is present in mature spores at a high concentration (Figure 1C). Yeast spores employ many mechanisms for surviving hostile environments. (Note the difference in time scale on the killing curves of Figures 3B and 3G.) Nevertheless, they remain dependent upon Hsp104. Once spores begin to die, hsp104 mutants die at ~200 times the rate of wild-type cells (Figure 3G). Long-term viability of spores stored at low temperatures also depends on Hsp104 (Sanchez et al., 1992). Apparently, sufficient structural fluctuations take place in the proteins of these dormant cells to constitute a threat to long-term survival. It is not clear whether the maintenance of viability in spores involves the continuous action of Hsp104 or a delayed repair function initiated on germination. 4.6. Conservation of Tolerance Functions in B-Type Proteins The tolerance functions of the B- and closely related C-type HSP100 proteins have been conserved throughout evolution. Mutations in the E. coli B protein compromise survival at high temperatures (Squires et al., 1991; see Burkholder and Gottesman, this volume); mutations in the B. subtilis C protein reduce survival at high temperatures, high concentrations of ethanol, and high salt (Kruger et al., 1994); and mutations in the cyanobacterial B protein reduce viability roughly 5-fold and severely reduce the ability of cells to remain photosynthetically active after heat shock (Eriksson and Clarke, 1996). The related cyanobacterial C protein is essential for growth and is induced by high light intensities or CO2 concentrations (Clarke and Eriksson, 1996). Although it is not
Molecular chaperones and folding catalysts
398
normally induced by heat, when the B gene is defective, the C protein becomes inducible, presumably fulfilling stress tolerance functions of the absent B protein (Eriksson and Clarke, 1996). Except for Hsp104 and, to a lesser degree, ClpB from E. coli, the biochemical functions of the B-type proteins have not been investigated. Given their sequence conservation and common roles in stress tolerance, it seems very likely that all B-type proteins function through a conserved molecular mechanism. Indeed, the Hsp104 homologs of vascular land plants, organisms separated from yeast by nearly a billion years of evolution, are able to partially restore thermotolerance to hsp104 mutants. The Arabidopsis protein increases survival ~100-fold (Schirmer et al., 1994) and the soybean protein ~25-fold (Lee et al., 1994). The ability of plant Hsp104 homologs to function in yeast thermotolerance strongly suggests they have a thermotolerance function in their native organisms, and preliminary data from Arabidopsis strains with altered HSP101 levels support this suggestion (Queitsch, Vierling and Lindquist, unpublished). In contrast, the E. coli B protein provides no thermotolerance in hsp104 mutants (Parsell et al., 1993). Assuming that the E. coli protein’s function in stress tolerance is similar, this observation indicates that the functions of HSP100 proteins are not autonomous, but require cooperation with other factors. In vitro studies (Section 4.2.2) suggest that these cooperating factors include the chaperones Ydj1 and Hsp70. Presumably, the Arabidopsis and soybean proteins are still capable of interacting with helper factors, but do so less effectively than Hsp104, while the E. coli protein does not interact with them productively. In eukaryotes, many Hsps have homologs in different cellular compartments, where they serve similar functions (see chapter by Bukau et al., this volume). The complete genomic sequence of S. cerevisiae indicates there is no Hsp104 homolog in the endoplasmic reticulum. There is, however, a B-type HSP100 protein in yeast mitochondria (Hsp78), and recent work suggests it functions in stress tolerance in that organelle. hsp78 mutations themselves have little phenotype (Leonhardt et al., 1993), but lead to defects in mitochondrial protein import and DNA maintenance when combined with certain Hsp70 mutations (Moczko et al., 1995; Schmitt et al., 1995). However, the recovery of mitochondrial protein synthesis after heat stress depends on Hsp78. This appears to involve a repair function rather than a protective function because Hsp78 does not prevent inactivation of mitochondrial protein synthesis by heat (Schmitt et al., 1996). When Hsp78 is expressed in the cytosol it can partially restore thermotolerance in hsp 104 deletion mutants (Schmitt et al., 1996). Thus far, three human relatives of Hsp 104 have been cloned, one type X, one type M, and one fragment that is too small to assign. The smaller human fragment is more homologous to class 1 HSP100 proteins than to class 2 proteins; however it is too early to predict whether a B-type Hsp 104 homolog exists in humans. Oddly, a 100 kd heatinducible hamster protein that cross reacts with an antibody directed against a C-terminal peptide of yeast Hsp104 (Parsell et al., 1991) is actually an unusual variant of the Hsp70 protein family, with no sequence homology to Hsp104 except for a few amino acids at the extreme C-terminus. In any case, the HSP100/Clp B-type proteins are clearly relevant to human biology and medicine. Many human parasites mount a strong heat-shock response when they transfer
The role of Hsp104 in stress tolerance and prion maintenance
399
from other environments to their human hosts and the induction of HSP100 proteins is often particularly intense. A deletion mutant of the Hsp104 homolog of Leischmania is less virulent than wild-type Leischmania cells (Hubel et al., 1997). This extends B-type protein functions in thermotolerance to yet another evolutionarily distinct group of organisms. If Homo sapiens should prove not to have a B-type HSP100 homolog, then this protein may prove a useful target for pharmaceutical intervention against human pathogens. 4.7. The Relationship Between Hsp104 and Other Tolerance Factors 4.7.1. Other Tolerance Factors Although Hsp104 plays a central role in stress tolerance, it is by no means the only factor involved. Other known tolerance mechanisms include the induction of osmolytes to alter the solvent properties of the fluid phase, the alteration of lipid composition to change the fluidity of membranes, the synthesis of metal-chelating agents and free-radical scavengers to sequester or eliminate toxins, and the induction of other Hsps (Parsell and Lindquist, 1993; Ruis and Schuller, 1995). The induction of other Hsps itself represents several different survival mechanisms, as each protein has a distinct biochemical function. Hsps do, however, have a common functional theme: in one way or another they help the cell cope with protein damage. The best-studied Hsps bind unfolded proteins and prevent them from aggregating (e.g., Hsp70, Hsp60, Hsp40) or help clear the cell of damaged proteins by ushering them along the degradation pathway (e.g., ubiquitin and certain ubiquitinconjugating enzymes; see Maurizi et al., this volume). The relationship between these different tolerance mechanisms and that of Hsp104 is just beginning to be explored, but in two specific cases informative stories are emerging. 4.7.2. Genetic Interactions Between Hsp70 and Hsp104 In many organisms, heat-inducible members of the HSP70 family play an important role in thermotolerance (Parsell and Lindquist, 1993; Schröder et al., 1993). Surprisingly, in S. cerevisiae, hsp70 mutations have little effect on thermotolerance (Craig and Jacobsen, 1984). The interpretation of this observation, however, must be tempered by the complexity of the problem. S. cerevisiae express four closely related HSP70 proteins in the cytosolic/nuclear compartment, the Ssa proteins, which form a single complementation unit (see Craig et al., this volume). Deletion of all four proteins is lethal. Deletion of a subset produces a strong temperaturerelated phenotype—failure to grow at the upper end of the normal temperature range—but does not impair the ability of cells to survive a short severe heat shock. In fact, these deletions increase survival compared to wild-type when cells are shifted directly from 25°C to 50°C (Craig and Jacobsen, 1985). This increase in basal thermotolerance is due to Hsp70’s role in regulating other stress-response proteins. Several Hsps, including Hsp104 and trehalose synthase (whose significance is explained below) are overexpressed in ssa mutants at normal temperatures (Hottiger et al., 1992; Sanchez et al., 1993). The essential character of the ssa proteins, and the effects of hsp70 mutants on the expression of other proteins, makes it difficult to unambiguously determine the role of Hsp70 in yeast
Molecular chaperones and folding catalysts
400
thermotolerance. Two other experiments, however, suggest Hsp70 functions complement those of Hsp104. First, as described earlier, when Hsp104 mutants are given a conditioning pretreatment and then shifted to 50°C, they initially exhibit some tolerance. This early tolerance is reduced when cells also carry mutations in heat-inducible ssa proteins (Sanchez et al., 1993). More compellingly, overexpression of Hsp70 partially restores thermotolerance in cells carrying hsp104 mutations (Sanchez et al., 1993). Thus, in the absence of Hsp104, Hsp70 clearly plays a role in thermotolerance. In a complementary fashion, when Hsp70 function is compromised by mutation, Hsp104 becomes important for normal growth (Sanchez et al., 1993). Cells carrying mutations in certain ssa proteins do not grow at high temperatures, but grow quite well at 25°C (Craig and Jacobsen, 1984). However, if an hsp104 deletion, which itself has no effect on growth, is introduced into these cells, their growth rate is severely reduced (Sanchez et al., 1993). The genetic complementarity between Hsp70 and Hsp104 reflects an underlying biochemical complementarity. Although Hsp70 can resolubilize certain special types of protein aggregates (Skowyra et al., 1990), it generally does not (Schröder et al., 1993; Wickner et al., 1994). Rather, it helps proteins that are already in an unfolded, or partially unfolded state, to remain soluble. As described elsewhere in this volume (Buchberger et al.; Ha et al.), Hsp70 accomplishes this by binding to small, hydrophobic, aggregationprone stretches of amino acids that would normally be buried in the hydrophobic core of a protein. Hsp104 does not prevent protein aggregation, but helps resolubilize aggregates once they occur. Presumably, when Hsp70 function is compromised, some aggregation occurs even at 25 °C and Hsp104 then becomes essential for normal growth. (Indeed electron microscopy of Ssa mutants reveals a low level of electron dense material, indicative of protein damage, at normal temperatures and this damage is exacerbated by hsp104 deletions; Lindquist et al., 1995). Conversely, when Hsp104 is not present to repair damaged proteins after heat shock, overexpression of Hsp70 compensates by reducing the rate at which protein aggregation occurs. While increased expression of Hsps is beneficial at high temperatures, it can be detrimental to growth at normal temperatures (Feder et al., 1992). The need to balance the beneficial and detrimental effects of these proteins in the context of an individual cell’s particular physiology is believed to account for differences in the patterns of Hsp synthesis observed in different organisms. A case in point is the absence of a stressinduced Hsp104 homolog in Drosophila cells. In this organism, the Hsp70 genes have been amplified, with five identical copies per haploid genome. Thus, yeast cells in which Hsp70 overexpression partially compensates for the loss of Hsp 104 in thermotolerance, mimic what appears to be a naturally-occuring adaptation of Drosophila. Only when we gain a more complete understanding of the beneficial and detrimental effects of Hsps will we understand the manner in which the expression of these proteins has changed in the course of evolution. 4.7.3. Genetic Interactions Between Trehalose Synthase and Hsp104 The induction of osmolytes is another common tolerance mechanism. Different
The role of Hsp104 in stress tolerance and prion maintenance
401
organisms produce different osmolytes in response to stress. For example, E. coli, S. cerevisae, and Artemia selina (brine shrimp) produce trehalose, whereas certain vascular plants produce sucrose, and mammalian kidney cells produce sorbitol and other solutes (Yancey et al., 1982; Wiemken, 1990; Somero et al., 1992). These compounds accumulate to extraordinary levels, both in response to stress and as part of the natural developmental adaptations for stress tolerance. For example, in yeast trehalose accumulates up to 0.5 M in response to heat shock (Hottiger et al., 1994), and in some fungi may accumulate to 15% of the dry weight of the cell during sporulation (Van Laere, 1989). Trehalose is the only osmolyte whose function in stress tolerance has been investigated genetically. In yeast, trehalose synthase mutants die more rapidly than wild-type cells during exposure to extreme heat (De Virgilio et al., 1994). One study suggests that, in tobacco, drought tolerance can be promoted by trehalose via expression of trehalose synthase (Holmstrom et al., 1996). By changing the properties of the fluid phase, trehalose preserves protein and membrane structure at high temperatures (Crowe et al., 1984; Colaco et al., 1992; Hottiger et al., 1994). Other osmolytes function in a similar manner in vitro and are likely to function in a similar manner in vivo (Yancey et al., 1982) To determine how protection afforded by the production of trehalose relates to that afforded by Hsp104, double mutants were examined. Each mutation alone reduced survival in conditioned cells on the order of 100-fold. In the double mutant, tolerance was reduced by 10,000-fold (Elliott et al., 1996; and Singer and Lindquist, 1998). Thus the genetic data confirm biochemical expectations. These two mechanisms operate on different critical points, trehalose reducing the likelihood that proteins will be denatured by heat, and Hsp104 acting to repair denaturation damage once it occurs. Together, they appear to provide the primary mechanisms by which yeast cells survive sudden exposure to extreme conditions. 5. HSP104’S INTERACTION WITH A YEAST PRION 5.1. A Role for Hsp 104 During Normal Growth An unexpected role for Hsp 104 in yeast cells at normal temperatures was discovered in a screen for factors that reduce nonsense suppression in strains carrying the omnipotent suppressor known as [PSI + ] (Chernoff et al., 1995). [PSI + ] proved to be a remarkable new type of genetic element composed only of protein, in which a heritable change in phenotype derives from a self-perpetuating change in protein structure. Thus, [PSI + ] is conceptually related to the remarkable infectious agents in mammals known as prions. The role of Hsp 104 in the propagation of this novel genetic element becomes intelligible in the context of its function in altering protein conformations. 5.2. The Prion Hypothesis Prions are unconventional infectious agents responsible for several devastating, invariably fatal neurodegenerative diseases, commonly termed the transmissible
Molecular chaperones and folding catalysts
402
spongiform encephalopathies (TSEs) (Prusiner, 1994; Prusiner and DeArmond, 1994; Weissman, 1994; Gajdusek, 1996; Prusiner, 1996). In humans these include CreutzfeldtJacob disease, Gerstmann-Straussler-Schienker syndrome, fatal familial insomnia and kuru (most famous for its transmission by the cannibalistic eating of infected human brains). In sheep, which were early subjects for experimental investigation, it is commonly known as scrapie; in cattle it is commonly known as “mad cow” disease. Although there are some differences in presentation, the TSEs are all characterized by dementia, motor neuron dysfunction, and an unusual “spongy” histology of brain tissue brought about by extensive vesiculation without lymphocytic infiltration. The devastating economic impact of “mad cow” disease in Great Britain, and the frightening prospect that a new, more virulent form of the disease is now appearing in humans (perhaps transmitted through contaminated beef products; Collinge et al., 1995) have heightened the urgency of the quest for effective medical intervention. Unlike all other known pathogens, prions are believed to consist entirely of protein— specifically, a normal cellular plasma-membrane protein that has undergone a change in conformation from its usual protease sensitive state, PrPC, to a pathogenic, proteaseresistant “scrapie” form, PrPSc. Pathogenicity is believed to derive from the ability of proteins in the PrPSc conformation to interact with normal PrP molecules and induce them to adopt the PrPSc structure. That is, a protein-conformational chain reaction both propagates the disease and generates new infectious material (Prusiner, 1994; Lansbury and Caughey, 1995). As the disease progresses, new infectious material is generated and PrPSc conformers are deposited in the brain and in certain peripheral tissues as dense, detergent-insoluble complexes. The TSEs are also unusual in that they can arise either from sporadic incidents, from the inheritance of mutated PrP proteins, or from infection. Rare sporadic cases may derive from rare, spontaneous misfolding events that produce PrPSc conformers. Inheritance is associated with mutations in PrP that are thought to destabilize the normal folded state, increasing the likelihood of obtaining the PrPSc conformation. Both the sporadic and the heritable diseases give rise to infectious prion particles that contain the same protein, PrPSc. Not surprisingly, the notion that an infectious agent can consist entirely of protein, with no associated nucleic acid, was greeted initially with great skepticism. Recently, however, the hypothesis has gained critical support from many different types of investigation, including biochemical analyses of the infectious material, in vitro conversions of PrPC into PrPSc-like conformers, and genetic analysis of transgenic mice. The prion hypothesis has now gained broad, though by no means universal, acceptance. Recent work in yeast suggests prions may be more common than initially envisioned. For nearly 30 years yeast geneticists have been tracing the inheritance of two mysterious cytoplasmic elements known as [PSI + ] and [URE3]. In biological terms [PSI + ] and [UEE3] are unrelated. [PSI + ] is a factor that reduces the fidelity of ribosome termination during protein synthesis, leading to the suppression of nonsense codons by ribosome read-through (Cox et al., 1988). [URE3] allows cells that are growing in the presence of ammonia to use additional nitrogen sources (Wickner et al., 1995). However, [PSI + ] and [URE3], share several unusual patterns of genetic behavior. Moreover, repeated attempts to link their dominant, cytoplasmic inheritance to any cytoplasmically
The role of Hsp104 in stress tolerance and prion maintenance
403
inherited nucleic acid have failed. Four years ago, Reed Wickner made the remarkable suggestion that the puzzling behavior of these elements could be explained through a prion-like model of protein conformation-based heredity (Wickner, 1994). In the short time since that suggestion, several lines of evidence have provided support—one of the most convincing being the central role played by Hsp104 in the inheritance of [PSI + ]. Because a role for Hsp104 has not been established for [URE3], it will not be considered further here. The reader is referred to Wickner’s excellent recent review for further information on this element (Wickner et al., 1995). 5.3. The Relationship Between [PSI + ] and Sup35 [PSI + ] is a dominant, cytoplasmically inherited trait. It is postulated to represent the inheritance of conformationally altered Sup35 protein. Sup35 in its normal conformation is a subunit (together with Sup45) of the translation-release factor that allows ribosomes to recognize termination signals (Stansfield et al., 1995). Just as mutations in PrP can produce heritable spongiform encephalopathies (Prusiner, 1994; Prusiner and DeArmond, 1994), mutations in Sup35 can suppress nonsense mutations by ribosome read-through (Tuite and Stansfield, 1994; Stansfield et al., 1995). Of course, [PSI + ] can not be due to a mutation in Sup35, because such mutations exhibit normal Mendelian inheritance (Tuite and Stansfield, 1994). That is, when a sup35 mutant is mated to a wild-type cell and the diploid is sporulated, the defect in translation termination is found in half of the haploid progeny. When a [PSI + ] cell is mated to a [psi - ] cell, the termination defect is observed in all of the progeny (Cox et al., 1988; Cox, 1994). Remarkably, overexpression of wild-type Sup35 can induce not only the [PSI + ] phenotype, but also new, heritable [PSI + ] elements. Even more remarkably, only transient overexpression of Sup35 is required. Once cells that overexpress Sup35 have converted to [PSI + ], this state is maintained in all of its progeny, even when the plasmid that confers overexpression is lost (Chernoff et al., 1993; Wickner, 1994). The gene encoding Sup35 has an unusual structure (Figure 5). The C-terminal domain encodes the translation-release factor and is essential. The N-terminal domain has an unusual amino acid composition and is not essential (Ter-Avanesyan et al., 1993). Overexpression of just the N-terminal domain is sufficient to convert a cell to [PSI + ]. Deletion of the N-terminal domain prevents cells from ever acquiring the [PSI + ] phenotype. Moreover, it causes the loss of [PSI + ] in cells that have already acquired it (Chernoff et al., 1993; Ter-Avanesyan et al., 1993). Physical analysis of the Sup35 protein demonstrates that it has a different conformational state in [PSI + ] and [psi - ] cells: 1) it pellets at 12,000×g in [PSI + ] cells, while it is mostly soluble in [psi - ] cells, 2) it is protease resistant in [PSI + ] cells while it is protease sensitive in [psi - ] cells, 3) newly synthesized proteins containing the
Molecular chaperones and folding catalysts
404
Figure 5 Sup35 domain structure and its relationship to [PSI + ]. A. Domain N has an unusual amino acid composition and contains several repeats ‘PQGGYQQYN’ that are similar to repeats found in the mammalian prion protein ‘PHGGGWGQ’. Domain M is highly charged and domain C, which is sufficient for Sup35’s translation termination function, contains four potential GTP-binding sites. A correlation was observed between the fiber forming propensity of domains and their ability to induce [PSI + ]. In cases where domains did not precisely match, the amino acid numbers are given in parentheses. B. The effects of deleting or overexpressing the NM and the C domains are shown. (Data from Chernoff et al., 1993, Ter-Avanesyan et al., 1993, Derkatch et al., 1996, and Glover et al., 1997.)
N-terminal region of Sup35 are rapidly sequestered into large particles in [PSI + ] cells, but remain freely distributed in [psi - ] cells (Patino et al., 1996; Paushkin et al., 1996). These data are consistent with the hypothesis depicted in Figure 6. The N-terminal domain of Sup35 converts cells from [psi - ] to [PSI + ] by promoting a conformational change that interferes with the ability of the C-terminal domain to function in translation termination, perhaps simply sequestering it in an aggregate that makes it inaccessible to ribosomes. The reduced concentration of functional termination factor leads to the readthrough of nonsense codons. The conformational alteration of Sup35, once achieved, is self-sustaining (with the help of Hsp104, see below). Therefore, it behaves as a dominant, cytosolically-inherited genetic element. 5.4. The Role of Hsp104 in the Maintenance of [PSI+] 5.4.1. Overexpression of HSP104 Cures [PSI + ] The relationship between Hsp104 and [PSI + ] was discovered by Yury Chernoff, who screened a yeast plasmid library for factors that would inhibit the [PSI + ] phenotype. The only plasmid uncovered in this initial screen proved to encode Hsp104 (Chernoff et al., 1995). Several criteria were used to demonstrate that Hsp104 does not affect translation directly, but simply eliminates the [PSI + ] element (Chernoff et al., 1995). Remarkably,
The role of Hsp104 in stress tolerance and prion maintenance
405
just as the transient overexpression of the N-terminal domain of Sup35 can cause the appearance of new, heritable [PSI + ] elements, the transient overexpression of Hsp104 causes the heritable loss of [PSI + ] (Chernoff et al., 1995;
Figure 6 Model of Sup35 sequestration effects in [PSI + ] cells. Sup35 is one subunit of the eukaryotic translation termination factor. The other is Sup45. In [PSI + ] cells, most Sup35 is found in a self-perpetuating, conformationally altered state (here depicted as an ordered aggregate) that prevents it from functioning in termination. The low level of functional Sup35 reduces the fidelity of protein synthesis, causing some ribosomes to read through nonsense codons. The presence of [PSI + ] is detected by the ability of cells carrying nonsense mutations in auxotrophic markers to grow without otherwise essential nutrients in the media.
Patino et al., 1996; Paushkin et al., 1996). When Hsp104 is placed under the control of a galactose-inducible promoter, a few hours of galactose induction is sufficient to eliminate [PSI + ]: 95% of the cells become [psi-] and pass this trait on to their progeny. The ability of a chaperone protein to cure cells of [PSI + ] strongly supports the hypothesis that [PSI + ] represents a heritable, altered conformation of Sup35. Once this conformation is eliminated, the strain no longer propagates [PSI + ]. 5.4.2. Deletion of HSP104 Cures [PSI + ] In most cases, when overexpession of a protein produces one phenotype, reduced expression produces the opposite. That is, deletion of Hsp104 might be expected to increase the stability of [PSI + ]. This is not the case. Deletion of Hsp104 cures cells of [PSI + ]. In fact, deletion of Hsp104 produces a psi-no-more (PNM) phenotype. When cells carrying an hsp104 deletion are mated to [PSI + ] cells, the diploid is [PSI + ]. However, when that diploid is sporulated, only half of the progeny, those that carry the wild-type HSP104 gene, remain [PSI + ] (Chernoff et al., 1995). 5.4.3. Hsp104 and Sup35 Aggregation in Vivo The unexpectedly similar effects of overexpressing and underexpressing Hsp104 on the inheritance of [PSI + ] provided an opportunity to test the relationship between [PSI + ]
Molecular chaperones and folding catalysts
406
and the altered, aggregated state of Sup35. If this altered state is responsible for the [PSI + ] phenotype, the aggregates should disappear when [PSI + ] is cured either by overexpression of Hsp104, or by deletion of Hsp104. Note that this is counter to expectations derived from the effects of Hsp104 on heat-induced aggregates: deletion of HSP104 causes proteins damaged by heat to remain in an aggregated state. The conformational state of Sup35 was examined in two ways, by differential sedimentation analysis of yeast lysates and by cytological visualization of Sup35 aggregates with Green Fluorescent Protein (Patino et al., 1996). Results of both analyses agreed. When cells are cured of [PSI + ], Sup35 returns to the soluble state, whether that cure is affected by overexpression of HSP104 or by deletion of HSP104. The effects of point mutations in consensus residues of the Hsp104 ATP-binding domains are also telling. When cells carrying a wild-type copy of HSP104 are transformed with plasmids encoding the lysine to threonine substitutions in NBD1 and NBD2 described in Section 3 (K218T, K620T), the [PSI+] phenotype disappears. Because cells are cured of [PSI + ] by either overexpression of Hsp104 or deletion of Hsp104, it is not possible to determine from these experiments whether these plasmids eliminate the [PSI+] phenotype by increasing Hsp104 function or by interfering with Hsp104 function. The effects of these mutations do, however, provide another means of testing the relationship between [PSI + ] and Sup35 aggregation. Remarkably, the K218T or K620T plasmids eliminate the [PSI + ] phenotype only as long as they are present in the cell. When the plasmids are lost by growth in non-selective media, [PSI + ] reappears. [PSI + ] must still be present in these cells, but somehow translation termination remains normal. Analysis of the Sup35 protein in these cells explains this cryptic behaviour. In cells carrying the individual substitutions, aggregates (intense fluorescent foci) of a Sup35-GFP fusion protein are still present, but a substantial fraction of Sup35 is diffusely distributed and soluble. Apparently, the soluble protein functions in translation termination, eliminating the [PSI + ] phenotype, while the aggregates promote the condensation of most newly synthesized Sup35 protein, restoring [PSI + ] when the plasmids are lost. These observations confirm that the aggregated state of Sup35 is intimately related to the inheritance of [PSI + ]. They also demonstrate that the relationship between Hsp104 and Sup35 aggregates is different from that which exists between Hsp104 and heatdamaged protein aggregates. Hsp104 is actually required for the formation of Sup35 aggregates. 5.4.4. Seeded Polymerization of Sup35 Recent biochemical investigations of Sup35 confirm that it can form aggregates with very special properties. When purified Sup35, produced in E. coli, is denatured in urea and diluted into aqueous buffer, it forms highly ordered, amyloid-like fibrils that bind Congo Red dye (Glover et al., 1997). These have a central, rod-like core of 10–11 nm, with more amorphous material splayed out on the sides. The C-terminal region (see Figure 5) and the M region show no capacity for fiber formation, but the N-terminal domain does, forming fibers that are short and 8 to 9 nm in diameter. Fibers formed by the N-terminal domain appear to be rich in -sheet structure, as are the fibrils formed by PrP and other
The role of Hsp104 in stress tolerance and prion maintenance
407
amyloids (Glover et al., 1997; King et al., 1997). The N and M regions together form long smooth fibers that look very much like the central core of the fibers formed by whole Sup35. Thus it would appear that fiber formation is directed by the glutamine and tyrosine-rich N domain,
Figure 7 Model of Hsp104’s function in [PSI+ ] conversion. Here, [PSI+ ] (right), is depicted as a special, ordered aggregate of Sup35 (left). (1) and (2) Hsp104 produces a conformational change in Sup35, either by unfolding Sup35 monomers, or by dissociating Sup35::Sup45 dimers. (3) The conformationally altered state of Sup35 produced by Hsp104 is unstable. In the absence of pre-existing [PSI + ] elements, the protein reverts to its normal, functional state. (4) When [PSI + ] elements are present, they capture the unstable conformers of Sup35. Presumably these elements fragment from time to time, giving rise to new elements that can be transmitted to the progeny of cells that contain them. Because [PSI + ] is transmitted to progeny with high fidelity (loss is less than 10−4 per generation), these elements must either be subject to frequent random fragmentation, or be partitioned to progeny in an orderly way. (5) In cells that do not already contain [PSI + ], overexpression of Sup35 will generate new elements. (6) Without Hsp104 (black bar) Sup35 can not acquire an aggregationprone state. Thus, Hsp104 deletion mutants have a psi-no-more phenotype. This is likely due to high local concentrations of the unstable, aggregation prone conformers of Sup35 in the vicinity of the individual Hsp104 molecules that produce them. Because Hsp104 is hexameric, it is possible that several Sup35 molecules are bound to the same Hsp104 particle at the same time, facilitating their coalescence into [PSI + ] elements. Once the [PSI + ] element has formed continuous over expression of Sup35 is not required, as the
Molecular chaperones and folding catalysts
408
cycle, once started, is self perpetuating. (7) The mechanism by which over expression of Hsp104 cures cells of [PSI + ] remains a mystery. There are several possibilities. For example, when it is present at high concentrations, Hsp104 might bind Sup35 aggregates and disaggregate them, in a manner analogous to the manner in which it promotes the disaggregation of heat-damaged proteins. Another possibility, is that high concentrations of Hsp104 reduce the local concentration of a limiting number of Sup35 proteins, so that they do not form postulated pre-aggregate subassemblies (depicted as trimers at the bottom of the figure) that might be precursors for association with [PSI + ]. Yet another possibility, among many, is that a limiting cofactor is required to activate Hsp104 in this process and it must associate with Hsp104 in a specific stoichiometry.
that the highly charged M domain packs into the fiber on the outside and the C-terminus remains amorphous. Most remarkably, the N and M domains, as well as the whole Sup35 protein, form fibers very slowly, over many hours of incubation. However, the addition of a very small fraction of preformed fibers greatly increases the rate of fiber formation in fresh protein (Glover et al., 1997). This seeded polymerization process provides a coherent molecular explanation for the inheritance of [PSI + ] (Glover et al., 1997), similar to that proposed previously for mammalian prions (Lansbury and Caughey, 1995). That is, the ordered aggregates of Sup35 reduce the population of free Sup35, reducing the efficiency of translation termination and causing the [PSI + ] phenotype. These are unlikely to form on their own in vivo, and thus [psi - ] cells remain [psi - ] for generation after generation. But once an aggregate has formed, it seeds the polymerization of new protein, which is passed on from mother cells to their daughters as heritable [PSI + ] elements. An important independent study by Paushkin et al., 1997 also supports such a model. Aggregated Sup35 protein purified from [PSI + ] cells causes the soluble Sup35 protein of [psi - ] cells to aggregate, and this aggregated material can be used to seed the aggregation of more soluble Sup35 from fresh [psi - ] lysates. 5.4.5. Hsp104 and Sup35 Interaction in Vitro The genetic interactions between Hsp104 and [PSI + ] and their influence on the aggregation state of Sup35 support the idea that a chaperone controls the conformational state of this yeast prion. But evidence for a direct physical interaction between Hsp104 and Sup35, which could mediate this transformation was wanting. Recent in vitro studies with purified Sup35 and Hsp104 proteins have provided this evidence (Schirmer and Lindquist, 1997). While stable Sup35-Hsp104 complexes have not been detected, three separate approaches indicate they interact directly. 1) Sup35 protein selectively inhibits the ATPase activity of Hsp104. This was surprising in that casein, a test-substrate of the ClpAP pro tease, stimulated the ATPase activity of ClpA (Hwang et al., 1988) and casein even weakly stimulated the ATPase activity of Hsp104 (Schirmer and Lindquist, unpublished). Perhaps the nature of Hsp104’s interaction with Sup35 requires more
The role of Hsp104 in stress tolerance and prion maintenance
409
concerted “work” from the chaperone, which reduces its basal rate of ATP hydrolysis. 2) Light scattering is increased in mixtures of purified Sup35 and Hsp104 proteins over the sum of the scattering by the proteins individually. 3) The circular dichroism (CD) spectrum observed for mixtures of the two proteins deviated significantly from that predicted by adding their individual spectra, contrary to the expected result if two proteins do not interact. This effect was dependent on the initial state of Sup35 which was controlled using different buffer conditions. In contrast, mixtures of Sup35 and control -lactalbumin, bovine proteins (aldolase, IgM, IgG, 2-mac-roglobulin, apoferritin, serum albumin, Hsp70) or of Hsp104 and (aldolase, IgM, IgG, 2-macroglobulin, apoferritin, -lactalbumin, bovine serum albumin) produced in each case a spectrum that matched the sum of the proteins individual spectra regardless of which buffers were employed. These data clearly indicate that a direct and specific physical interaction occurs between Hsp104 and Sup35 which changes the state of the protein(s). However, it is as yet unclear what is the exact nature of the change that occurs in the protein (s). In mixed solutions it is impossible to assign CD changes to individual proteins. Furthermore, the increase in light scattering suggests that some part of the spectral shift observed by CD is due to a loss of material from the solvent phase making analysis of the structural changes underlying the shift more complex. It is also unclear whether the interaction of Sup35 and Hsp104 alone in vitro will fully represent the role of Hsp104 in prion maintenance. In vivo, Sup35 is normally bound to a partner protein, Sup45, and other as yet unidentified cofactors might also be involved. Finally, in keeping with genetic observations, variations in the ratios of Hsp104 to these components need to be investigated in vitro. 5.5. A Model for Hsp104-[PSI + ] Interactions Clearly, Hsp104 plays a central role in the formation and maintenance of [PSI + ], The relationship between Hsp104 and heat-induced aggregates is different from that between Hsp104 and Sup35 aggregates. Nevertheless, the role Hsp104 plays in providing thermotolerance and the role it plays in the inheritance of [PSI + ] must be biochemically related by Hsp104’s capacity to alter the conformational state of other proteins. Although it must be taken as highly speculative, a model to explain the role of Hsp104 in the inheritance of [PSI + ] is presented in Figure 7 and its legend. 5.6. An Explanation for Previous Puzzles The remarkable interactions between Hsp104 and [PSI + ] provide an explanation for three long-standing puzzles concerning [PSI + ] and one more recent puzzle concerning thermotolerance. 5.6.1. Hsp104 Explains Puzzling [PSI + ] Phenomena The longest standing puzzle concerns the underlying genetic nature of the [PSI + ] element. Many types of evidence, including the effects of Hsp104, now support the hypothesis that the inheritance of [PSI + ] is based upon the inheritance of a conformationally altered Sup35 protein. This mechanism of inheritance is, of course,
Molecular chaperones and folding catalysts
410
novel and unexpected. For many years it had been assumed that the inheritance of [PSI + ] must be based upon the inheritance of a specific nucleic acid. The ability of mutagens such as UV light and MMS (methyl methanesulfonate) to cure cells of [PSI + ] seemed to provide strong support for this hypothesis (Tuite and Cox, 1980). How can this be reconciled with current data? DNA sequence analysis reveals the presence of two DNA damage-response elements in the HSP104 promoter. Indeed, Hsp104 is strongly and selectively induced by UV and MMS (Schmitt and McEntee, personal communication). The ability of Hsp104 to cure cells of [PSI + ] thus allows the effects of these mutagens to be accomodated within the proteinconformation based hypothesis of [PSI + ] inheritance. Another puzzle concerns the effects of guanidinium hydrochloride (GuHCl), a protein denaturant, on [PSI + ]. Growth in low concentrations of GuHCl cures virtually all cells of [PSI + ] (Tuite and Cox, 1980). This concentration seems too low to change the conformation of Sup35 directly. More likely, curing results from the induction of Hsp104 by low concentrations of GuHCl (Lindquist et al., 1995). Finally, heat shock also cures cells of [PSI + ]. This, of course, is likely due to the induction of Hsp104. Curiously, however, heat shock cures cells of [PSI + ] much less efficiently than growth in the presence of GuHCl (Tuite and Cox, 1980), and growth at high temperatures does not cure cells of [PSI + ] at any appreciable frequency (Tuite et al., 1981; Chang and Lindquist, unpublished). The reason may be that the induction of Hsp104 by GuHCl is far more selective than its induction by heat (Lindquist et al., 1995). For example, at the low concentrations of GuHCl required to cure [PSI + ] Hsp70 and Hsp26 are not induced (Lindquist et al., 1995). Apparently, in response to heat, yeast cells produce other factors that protect [PSI + ] from efficient Hsp104 curing, thereby protecting prion-based inheritance from minor vagaries of the environment. This protection of [PSI + ] suggests that [PSI + ] is not simply a disease of yeast cells, but represents a biologically advantageous mechanism for changing the fidelity of protein synthesis, and presumably the pattern of gene expression, in a heritable way. Indeed, the mere existence of the N-terminal domain of Sup35 suggests as much. Yeast cells rapidly rid themselves of disadvantageous markers, yet they retain the N-terminal domain of Sup35. Since this domain is dispensible for normal growth and apparently serves only to reduce the fidelity of translation when cells carry [PSI + ], it follows that reducing the fidelity of translation is likely to be of selective advantage under some circumstances. Perhaps [PSI + ] allows ribosomes to read through stop codons on certain natural messages, into new coding information, which might allow cells to exploit a specific new environmental niche. If so, the heritability of the element would ensure that this trait is stably maintained, but the low-level instability of the element would ensure that once a population has attained an appreciable size, [PSI + ] will be lost from some individuals, returning them to normal translation and fitness. 5.6.2. [PSI + ] Explains a Puzzling Thermotolerance Phenomenon The vital role played by Hsp104 in stress tolerance is clear. For several years, however, the observations of Smith and Jaffe (1991) on the effects of a temperature-sensitive heat shock transcription factor (HSF) provided a puzzling contradiction. This mutant can not
The role of Hsp104 in stress tolerance and prion maintenance
411
induce Hsps in response to stress, yet it is perfectly capable of acquiring thermotolerance. The existence of non Hsp-based mechanisms of tolerance, such as the synthesis of trehalose, provides a partial explanation. However, these HSF mutants acquire just as much thermotolerance as wild-type cells, which should not be possible in the absence of Hsp104 induction. Analysis of Hsp accumulation in these mutants with specific antibodies seemed to solve the problem. The mutants have high basal levels of Hsp 104 expression and induction of Hsp 104 by heat is not completely blocked. In addition to classic HSEs, heat shock elements, the Hsp104 promoter contains STREs, stress response elements that use a different transcription factor. It seemed a simple matter to prove that thermotolerance in the HSF mutant was due to Hsp104 expression. But, it proved impossible to delete the Hsp104 gene in this strain in order to test this directly. In fact, it was impossible to even add an extrachromosomal wild-type copy of HSP104 to these cells. Either overexpression or deletion of Hsp104 was lethal. This conundrum was solved when it was determined that the HSF mutation was not a missence mutation, but a nonsense mutation that should have been lethal, because HSF is essential in yeast. However, the mutation had been produced in a [PSI + ] strain. Readthrough of the nonsense codon apparently kept the cells alive. However the cells were fixed with a requirement for moderate expression of Hsp104. Either overexpression or deletion of Hsp104 would cure the cells of [PSI + ], uncovering the effect of the lethal nonsense mutation in HSF. Thus, these cells represent a new type of mutant in [PSI + ] phenomenology, PFM, or psi-forever-more (Lindquist and Kim, 1996). 5.7. Broader Implications The study of mammalian neurodegenerative diseases is labor-intensive, expensive, and requires the use of many animals. The study of yeast prion-like elements can now provide a valuable supplement to mammalian investigations that will speed our understanding of self-propagating conformational changes, and may lead to the development of novel approaches for therapeutic intervention in neurodegenerative diseases. The hypothesis that [PSI + ] represents a protein-conformation based mechanism of inheritance, however, has much broader implications. If prions exist in both mammals and yeast, it is likely that they are very broadly distributed and may be responsible for many different epigenetic phenomena in other organisms, which have thus far defied molecular explanation. Finally, because chaperone proteins are induced by stress, their ability to influence the phenotypic state of the cell in this manner provides an extraordinary mechanism for a cell to respond to its environment in a heritable way. 5.8. Hsp 104 and Mammalian Prions The role of Hsp104 in the maintenance of the yeast prion [PSI + ] suggests that chaperone proteins might also regulate the structural state of other prions. Indeed there have been previous indirect suggestions of such interactions (Telling et al., 1995; Kenward et al., 1996). Providing a “proof of principle”, two recent studies have demonstrated that Hsp 104 can affect the conformational state of the mammalian prion, PrP, believed to be the causative agent in the transmissible spongiform encephalopathies (TSEs; See section 5.2).
Molecular chaperones and folding catalysts
412
In the first, a bacterially expressed fragment of PrP corresponding to the proteaseresistant core that is characteristic of the TSE-associated form (rPrP) was mixed with Hsp104 in vitro. Changes in the ATPase activity of Hsp 104 and changes in the CD spectra of mixed proteins was assessed as for mixtures of Hsp104 and Sup35 (see section 5.4.4). Like Sup35, rPrP inhibited the ATPase activity of Hsp104. Also, mixtures of Hsp104 and rPrP exhibited a specific and highly reproducible shift in their circular dichroism spectra from that predicted by addition of each protein’s individual spectrum (Schirmer and Lindquist, 1997). This version of PrP has been extensively characterized (Mehlhorn et al., 1996; Zhang et al., 1997) and can be folded initially into either helical or -sheet forms. Just as the interaction between Hsp104 and Sup35 depended upon the initial state of Sup35, a spectral shift was observed when Hsp104 was mixed with the -sheet form of rPrP but not when Hsp104 was mixed with the -helical form of rPrP (Schirmer and Lindquist, 1997). Some of the control proteins used were also rich in -sheet, yet spectral shifts were not observed: thus this feature alone is insufficient to explain the shift. Another difference between the two rPrP forms is that the -sheet form sediments by ultracentrifugation at an average Mr which suggests it is in small aggregates containing ~7 proteins each, while the -helical form sediments as a monomer (Zhang et al., 1997). This is particularly interesting in light of the observation that in Hsp104’s in vitro interaction with luciferase, a thermotolerance test-substrate (section 4.2.2), the size of the most reactivation-competent aggregates were composed of 7 or more proteins (Glover and Lindquist, 1998). The second study employed an in vitro assay for the conformational conversion of PrP developed by Byron Caughey and colleagues. This assay uses preparations of the infectious, protease-resistant form of PrP, PrPSc, to seed the conversion of a small amount of radio-labeled protease-sensitive form, PrPC, to a PrPSc-like protease-resistant form. Although the infectious nature of the converted material has not been demonstrated, the conversion reaction mimics many essential features of PrPSc conversion, including the requirement for PrPSc, the production of strain-specific proteolytic fragments, and the species barrier (Kocisko et al., 1995; Raymond et al., 1997). Hsp104 and several other chaperones were added to this conversion reaction to determine if they could affect the rate of conversion. None of the chaperones could convert PrPC to the protease resistant form in the absence of added PrPSc, confirming that even in the presence of proteins that can alter the conformations of other proteins, a “templating” activity of PrPSc is required. In the presence of PrPSc, both GroEL and Hsp104 increased the rate of conversion (DebBurman et al., 1997). The initial state of the PrP was critical to the conversion process: Hsp104 only facilitated conversion if a partially denatured PrPSc seed was used. The post-translational modifications of PrP also influenced the process. A ~30–40% increase in conversion to protease-resistant material was observed when a modified PrP lacking its GPI anchor or a non-glycosylated form was used in the assay (DebBurman et al., 1997). While Hsp104 promoted conversion, osmolytes (chemical chaperones) inhibited it (DebBurman et al., 1997). Both Hsp104 and osmolytes protect cells from severe heat stress (see section 4.7.3). Osmolytes do so by preventing protein aggregation and Hsp104 does so by solubilizing proteins that have already aggregated. With PrP, the role of the osmolytes is consistent—preventing aggregation. In contrast, the role of Hsp104 appears to be to promote aggregation, contrary to its role in thermotolerance, but
The role of Hsp104 in stress tolerance and prion maintenance
413
consistent with one of its two effects on [PSI + ] and also consistent with the idea that Hsp104’s function with [PSI + ] differs from its function in thermotolerance (section 1). Hsp104 and PrP are from different organisms and different subcellular compartments. These observations, therefore, do not imply that Hsp104 plays a role in prion-mediated disease. However, the demonstration that a chaperone can affect the conformational transitions of PrP, emphasizes the importance of determining what natural chaperones might interact with PrP in vivo, as a possible means for therapeutic intervention. Perhaps the most striking outcome of these studies is the underlying structural similarities they suggest between the PrP protein and Sup35. At an amino-acid sequence level, these proteins have little in common and their biological functions are completely unrelated. Yet Hsp104, which shows no evidence of interaction with most other cellular proteins, interacts with both PrP and Sup35 in a highly specific manner to promote an unusual change in structure that has profound consequences. 6. CONCLUDING REMARKS All proteins that are produced in large quantities in response to heat, the Hsps, are now known to perform functions that address problems in protein folding. Some of the proteins, such as polyubiquitin and certain of the ubiquitin congugating enzymes, function to rid the cell of damaged proteins through proteolysis. Others, such as Hsp70, DnaJ, and GroEL (or Hsp60), bind unfolded proteins and prevent them from aggregating. Hsp104 complements these functions in a unique and unexpected way, helping to resolve aggregates once they have formed and to return the aggregated proteins to the soluble, functional state. This biochemical function fits the biological role of Hsp104 in stress tolerance. Hsp104 is most important under conditions of extreme stress, when the rate of protein aggregation is likely to outpace the capacity of other protective mechanisms to prevent aggregation. The role of Hsp104 in the inheritance of the novel, protein conformation-based genetic element known as [PSI + ] is more perplexing. In this case, Hsp104 is actually required, at a specific concentration, to maintain Sup35 aggregates. Hsp104’s ability to promote the dissolution of heat-damaged aggregates and to promote the formation of Sup35 aggregates are biochemically related. Mutations in conserved nucleotide-binding residues of Hsp104 have strong effects on thermotolerance and the maintenance of [PSI + ]. Homologous mutations in NBD-1 and NBD-2 have different effects on the two most distinctive biochemical properties of Hsp104, its ability to hydrolyze ATP and its capacity to oligomerize in response to nucleotide. In the future, a detailed understanding of these biochemical properties will no doubt illuminate Hsp104’s remarkable biological functions. A complete picture of Hsp104’s functions in thermotolerance and [PSI + ] maintenance will also require better understanding of the structural nature of Hsp104’s substrates. This presents a major challenge, as methods for studying protein aggregates are still poorly developed. Happily, the problem of protein aggregation is engaging an ever growing number of biochemists due, in part, to a new appreciation of the important role played by protein aggregation in several devastating human diseases (Thomas et al, 1995). Moreover, it is increasingly clear that chaperone proteins play an important role in the
Molecular chaperones and folding catalysts
414
metabolism of these aggregated proteins (Perutz et al., 1994; Thomas et al., 1995; Gajdusek, 1996; Prusiner, 1996; Wetzel, 1996). Lessons learned in studying the aggregated proteins that cause disease will no doubt provide important insights on Hsp104 functions. Hopefully, the ability to analyze the problem of protein aggregation in an organism as genetically tractable as yeast, will also provide insights of importance to human biology and medicine. The problem is also of great economic importance, not only in the context of human and animal diseases (e.g. mad cow disease), but also because heat and water stress are the greatest barriers to crop productivity. The role demonstrated for Hsp104 in thermotolerance and the functional conservation of its homologs in plants suggest that manipulation of HSP100 proteins will be of great advantage for the survival of crops during heat and drought stresses. 7. REFERENCES Boreham, D.R. and Mitchel, R.E. (1994). Regulation of heat and radiation stress responses in yeast by hsp-104. Radiation Rsch. , 137 , 190–195. Chernoff, Y.O., Derkach, I.L. and Inge-Vechtomov, S.G. (1993). Multicopy SUP35 gene induces de-novo appearance of psi-like factors in the yeast Saccharomyces cerevisiae. Curr. Genet. , 24 , 268–270. Chernoff, Y.O., Lindquist, S.L., Ono, B., Inge-Vechtomov, S.G. and Liebman, S.W. (1995). Role of the chaperone protein Hsp104 in propagation of the yeast prion-like factor [psi + ]. Science , 268 , 880–884. Chuang, S.E., Burland, V., Plunkett, G.D., Daniels, D.L. and Blattner, F.R. (1993). Sequence analysis of four new heat-shock genes constituting the hslTS/ibpAB and hslVU operons in Escherichia coli. Gene , 134 , 1–6. Clarke, A.K. and Eriksson, M.-J. (1996). The cyanobacterium Synechococcus sp. PCC7942 possesses a close homologue to the chloroplast ClpC protein of higher plants. Plant Molec. Biol , 31 , 721–730. Colaco, C., Sen, S., Thangavelu, M., Pinder, S., and Roser, B. (1992). Extraordinary stability of enzymes dried in trehalose: simplified molecular biology. Bio/Technology , 10 , 1007–1011. Collinge, J., Palmer, M.S., Sidle, K.C., Hill, A.E, Gowland, I., Meads, J., Asante, E., Bradley, R., Doey, L.J. and Lantos, P.L. (1995). Unaltered susceptibility to BSE in transgenic mice expressing human prion protein. Nature , 378 , 779–783. Cox, B. (1994). Cytoplasmic inheritance. Prion-like factors in yeast. Curr. Biology , 4 , 744–748. Cox, B.S., Tuite, M.F. and McLaughlin, C.S. (1988). The psi factor of yeast: a problem in inheritance. Yeast , 4 , 159–178. Craig, E.A. and Jacobsen, K. (1984). Mutations of the heat inducible 70 kilodalton genes of yeast confer temperature sensitive growth. Cell , 38 , 841–849. Craig, E.A. and Jacobsen, K. (1985). Mutations in cognate genes of Saccharomyces cerevisiae hsp70 result in reduced growth rates at low temperatures. Molec. Cell. Biol , 5 , 3517–3524. Crowe, J.H., Crowe, L.M., and Chapman, D. (1984). Preservation of membranes in anhydrobiotic organisms: the role of trehalose. Science , 223 , 701–703.
The role of Hsp104 in stress tolerance and prion maintenance
415
De Virgilio, C., Hottiger, T., Dominguez, J., Boller, T. and Wiemken, A. (1994). The role of trehalose synthesis for the acquisition of thermotolerance in yeast. I. Genetic evidence that trehalose is a thermoprotectant . Eur. J. Biochem. , 219 , 179–186. DebBurman, S.K., Raymond, G.J., Caughey, B. and Lindquist, S. (1997). Chaperonesupervised conversion of prion protein to its protease-resistant form. Proc. Natl. Acad. Set. USA , 94 , 13938–13943. Deckert, G., Warren, P.V., Gaasterland, T., Young, W.G., Lenox, A.L., Graham, D.E., Overbeek, R., Snead, M.A., Keller, M., Aujay, M., Huber, R., Feldman, R.A., Short, J.M., Olsen, G.J. and Swanson, R.V. (1998). The complete genome of the hyperthermophilic bacterium Aquifex aeolicus. Nature , 392 , 353–358. Derkatch, I.L., Chernoff, Y.O., Kushnirov, V.V., Inge-Vechtomov, S.G. and Liebman, S.W. (1996). Genesis and variability of [PSI] prion factors in Saccharomyces cerevisiae. Genetics , 144 , 1375–1386. Elliott, B., Haltiwanger, R.S. and Futcher, B. (1996). Synergy between trehalose and Hsp104 for thermotolerance in Saccharomyces cerevisiae. Genetics , 144 , 923–933. Eriksson, M.J. and Clarke, A.K. (1996). The heat shock protein ClpB mediates the development of thermotolerance in the cyanobacterium Synechococcus sp. strain PCC 7942. J. Bacteriology , 178 , 4839–4846. Feder, J.H., Rossi, J.M., Solomon, J., Solomon, N. and Lindquist, S. (1992). The consequences of expressing hsp70 in Drosophila cells at normal temperatures. Genes Devel. , 6 , 1402–1413. Gajdusek, D.C. (1996). Infectious amyloids: subacute spongiform encephalopathies as transmissible cerebral amyloidoses. In Fields, B.N., Knipe, D.M. and Howley, P.M. (ed.), Fields Virology , Third Edition. Lippencott-Raven Publishers, Philadelphia, pp. 2851–2900. Glover, J.R., Kowal, A.S., Schirmer, E.C., Patino, M.M., Liu, J.-J. and Lindquist, S. (1997). Self-seeded fibers formed by Sup35, the protein determinant of [PSI+], a heritable prion-like factor of S. cerevisiae. Cell , 89 , 811–819. Glover, J.R. and Lindquist, S. (1998). Hsp104, Hsp70 and Hsp40: a novel chaperone system that rescues previously aggregated proteins. Submitted. Gottesman, S., Clark, W.P., de Crecy-Lagard, V. and Maurizi, M.R. (1993). ClpX, an alternative subunit for the ATP-dependent ClpProtease of Escherichia coli. J. Biol. Chem. , 268 , 22618–22626. Gottesman, S., Wickner, S., Jubete, Y., Singh, S.K., Kessel, M. and Maurizi, M. (1995). Selective, energy-dependent proteolysis in Escherichia coli. In Protein Kinesis: The Dynamics of Protein Trafficking and Stability, Cold Spring Harbor Symposia on Quantitative Biology, Volume LX. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, pp. 533–548. Holmstrom, K-O., Mantyla, E., Welin, B., Mandal, A., Palva, E.T., Tunnela, O.E. and Londesborough, J. (1996). Drought tolerance in tobacco. Nature , 379 , 683–684. Hottiger, T., De Virgilio, C., Bell, W., Boller, T. and Wiemken, A. (1992). The 70kilodalton heat-shock proteins of the SSA subfamily negatively modulate heat-shockinduced accumulation of trehalose and promote recovery from heat stress in the yeast, Saccharomyces cerevisiae. Eur. J. Biochem. , 210 , 125–132. Hottiger, T., De Virgilio, C., Hall, M.N., Boller, T. and Wiemken, A. (1994). The role of
Molecular chaperones and folding catalysts
416
trehalose synthesis for the acquisition of thermotolerance in yeast. II. Physiological concentrations of trehalose increase the thermal stability of proteins in vitro. Eur. J. Biochem. , 219 , 187–193. Hubel, A., Brandau, S., Dresel, A. and Clos, J. (1995). A member of the ClpB family of stress proteins is expressed during heat shock in Leishmania spp. Molec. Biochem. Parasitology , 70 , 107–118. Hubel, A., Krobitsch, S., Horauf, A. and Clos, J. (1997). Leishmania major Hsp100 is required chiefly in the mammalian stage of the parasite. Mol. Cell. Biol , 17 , 5987– 5995. Hwang, B.J., Park, W.J., Chung, C.H. and Goldberg, A.L. (1987). Escherichia coli contains a soluble ATP-dependent protease (Ti). distinct from protease La. Proc. Natl. Acad. Sci. USA , 84 , 5550–5554. Hwang, B.J., Woo, K.M., Goldberg, A.L. and Chung, C.H. (1988). Protease Ti, a new ATP-dependent protease in Escherichia coli, contains protein-activated ATPase and proteolytic functions in distinct subunits. J. Biol. Chem. , 263 , 8727–8734. Katayama-Fujimura, Y., Gottesman, S. and Maurizi, M.R. (1987). A multiplecomponent, ATP-dependent protease from Escherichia coli. J. Biol. Chem. , 262 , 4477–4485. Kenward, N., Landon, M., Laszlo, L. and Mayer, R.J. (1996). Heat shock proteins, molecular chaperones and the prion encephalopathies. Cell Stress Chaperones , 1 , 18– 22. Kessel, M., Maurizi, M.R., Kim, B., Kocsis, E., Trus, B.L., Singh, S.K. and Steven, A.C. (1995). Homology in structural organization between E. coli ClpAP protease and the eukaryotic 26 S proteasome. J. Mol. Biol. , 250 , 587–594. King, C.Y., Tittmann, P., Gross, H., Gebert, R., Aebi, M. and Wuthrich, K. (1997). Prion-inducing domain 2–114 of yeast Sup35 protein transforms in vitro into amyloidlike filaments. Proc. Natl. Acad. Sci. USA , 94 , 6618–6622. Kocisko, D.A., Priola, S.A., Raymond, G.J., Chesebro, B., Lansbury, P.T. Jr. and Caughey, B. (1995). Species specificity in the cell-free conversion of prion protein to protease-resistant forms: a model for the scrapie species barrier. Proc. Natl. Acad. Sci. USA , 92 , 3923–3927. Kruger, E., Volker, U. and Hecker, M. (1994). Stress induction of clpC in Bacillus subtilis and its involvement in stress tolerance. J. Bacteriology , 176 , 3360–3367. Lansbury, P.T.J. and Caughey, B. (1995). The chemistry of scrapie infection: implications of the ‘ice 9’ metaphor. Chem. Biol. , 2 , 1–5. Lee, Y.R., Nagao, R.T. and Key, J.L. (1994). A soybean 101-kD heat shock protein complements a yeast HSP104 deletion mutant in acquiring thermotolerance. Plant Cell , 6 , 1889–1897. Leonhardt, S.A., Fearson, K., Danese, P.N. and Mason, T.L. (1993). HSP78 encodes a yeast mitochondrial heat shock protein in the Clp family of ATP-dependent proteases. Molec. Cell Biol , 13 , 6304–6313. Levchenko, I., Smith, C.K., Walsh, N.P., Sauer, R.T. and Baker, T.A. (1997). PDZ-like domains mediate binding specificity in the Clp/Hsp100 family of chaperones and protease regulatory subunits. Cell , 91 , 939–947. Lindquist, S. and Kim, G. (1996). Heat-shock protein 104 expression is sufficient for
The role of Hsp104 in stress tolerance and prion maintenance
417
thermotolerance in yeast. Proc. Natl. Acad. Sci. USA , 93 , 5301–5306. Lindquist, S., Patino, M.M., Chernoff, Y.O., Kowal, A.S., Singer, M.A., Liebman, S.W., Lee, K.-H. and Blake, T. (1995). The role of Hsp104 in stress tolerance and [PSI+] propagation in Saccharomyces cerevisiae. In Protein Kinesis: The Dynamics of Protein Trafficking and Stability, Cold Spring Harbor Symposia on Quantitative Biology, Volume LX. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, pp. 451–460. Louvion, J.-F., Havaux-Copf, B. and Picard, D. (1993). Fusion of GAL4-VP16 to a steroid-binding domain provides a tool for gratuitous induction of galactose-responsive genes in yeast. Gene , 131 , 129–134. Maurizi, M.R., Thompson, M.W., Singh, S.K. and Kim, S.-H. (1994). Endopeptidase Clp: ATP-dependent Clp protease from Escherichia coli. Meth. Enzymol. , 244 , 314– 331. Mehlhorn, I., Groth, D., Stockel, J., Moffat, B., Reilly, D., Yansura, D., Willett, W.S., Baldwin, M., Fletlerick, R., Cohen, F.E., Vandlen, R., Henner, D. and Prusiner, S.B. (1996). High-level expression and characterization of a purified 142-residue polypeptide of the prion protein. Biochemistry , 35 , 5528–5537. Moczko, M., Schonfisch, B., Voos, W., Pfanner, N. and Rassow, J. (1995). The mitochondrial ClpB homolog Hsp78 cooperates with matrix Hsp70 in maintenance of mitochondrial function. J. Molec. Biol , 254 , 538–543. Nover, L. (1991). HSFs and HSPs—a stressful program on transcription factors and chaperones. Stress Proteins and the Heat Shock Response, sponsored by Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, USA, April 29–May 2, 1991. New Biologist , 3 , 855–859. Parsell, D.A., Kowal, A.S. and Lindquist, S. (1994a). Saccharomyces cerevisiae Hsp104 protein. Purification and characterization of ATP-induced structural changes. J. Biol. Chem. , 269 , 4480–4487. Parsell, D.A., Kowal, A.S., Singer, M.A. and Lindquist, S. (1994b). Protein disaggregation mediated by heat-shock protein Hsp104. Nature , 372 , 475–8. Parsell, D.A. and Lindquist, S. (1993). The function of heat-shock proteins in stress tolerance: degradation and reactivation of damaged proteins. Annu. Rev. Genetics , 27 , 437–496. Parsell, D.A., Sanchez, Y., Stitzel, J.D. and Lindquist, S. (1991). Hsp104 is a highly conserved protein with two essential nucleotide-binding sites. Nature , 353 , 270–273. Parsell, D.A., Taulien, J. and Lindquist, S. (1993). The role of heat-shock proteins in thermotolerance. Philos. Trans. Royal Soc. London—Series B: Biological Sciences , 339 , 279–85. Patino, M.M., Liu, J.-J., Glover, J.R. and Lindquist, S. (1996). Support for the prion hypothesis for inheritance of a phenotypic trait in yeast. Science , 273 , 622–626. Paushkin, S.V., Kushnirov, V.V., Smirnov, V.N. and Ter-Avanesyan, M.D. (1996). Propagation of the yeast prion-like [psi+] determinant is mediated by oligomerization of the SUP35-encoded polypeptide chain release factor. EMBO J. , 15 , 3127–3134. Paushkin, S.V., Kushnirov, V.V., Smirnov, V.N. and Ter-Avanesyan, M.D. (1997). In vitro propagation of the prion-like state of yeast Sup35 protein. Science , 277 , 381– 383. Perutz, M.F., Johnson, T., Suzuki, M. and Finch, J.T. (1994). Glutamine repeats as polar
Molecular chaperones and folding catalysts
418
zippers: their possible role in inherited neurodegenerative diseases. Proc. Natl. Acad. Sci. USA , 91 , 5355–5358. Prusiner, S.B. (1994). Biology and genetics of prion diseases. Annual Review of Microbiology , 48 , 655–686. Prusiner, S.B. (1996). Prions. In Fields, B.N., Knipe, D.M. and Howley, P.M. (ed.), Fields Virology , Third Edition. Lippencott-Raven Publishers, Philadelphia, pp. 2901– 2950. Prusiner, S.B. and DeArmond, S.J. (1994). Prion diseases and neurodegeneration. Annu. Rev. Neurosci. , 17 , 311–339. Raymond, G.J. , Hope, J., Kocisko, D.A., Priola, S.A., Raymond, L.D., Bossers, A., Ironside, J., Will, R.G., Chen, S.G., Petersen, R.B., Gambetti, P., Rubenstein, R., Smits, M.A., Lansbury, P.T. Jr. and Caughey, B. (1997). Molecular assessment of the potential transmissibilities of BSE and scrapie to humans. Nature , 388 , 285–288. Ruis, H. and Schuller, C. (1995). Stress signaling in yeast. Bioessays , 17 , 959–965. Sanchez, Y. and Lindquist, S.L. (1990). HSP104 required for induced thermotolerance. Science , 248 , 1112–1115. Sanchez, Y., Parsell, D.A., Taulien, J., Vogel, J.L., Craig, E.A. and Lindquist, S. (1993). Genetic evidence for a functional relationship between Hsp104 and Hsp70. J. Bacteriology , 175 , 6484–6491. Sanchez, Y., Taulien, J., Borkovich, K.A. and Lindquist, S. (1992). Hsp104 is required for tolerance to many forms of stress. EMBO J. , 11 , 2357–2364. Saraste, M., Sibbald, P.R. and Wittinghofer, A. (1990). The P-loop—a common motif in ATP- and GTP-binding proteins. Trends Biochem . Sci., 15 , 430–434. Schirmer, E.C., Glover, J.R., Singer, M.A. and Lindquist, S. (1996). HSP100/Clp proteins: a common mechanism explains diverse functions. Trends Biochem, Sci. , 21 , 289–296. Schirmer, E.C., Lindquist, S. and Vierling, E. (1994). An Arabidopsis heat shock protein complements a thermotolerance defect in yeast. Plant Cell , 6 , 1899–909. Schirmer, E.C. and Lindquist, S. (1997). Interactions of the chaperone Hsp104 with yeast Sup35 and mammalian PrP. Proc. Natl. Acad. Sci. USA , 94 , 13932–13937. Schirmer, E.C., Queitsch, C., Kowal, A.S., Parsell, D.A. and Lindquist, S. (1998). The ATPase activity of Hsp104: effects of environmental conditions and mutations. In press. Schmitt, M., Neupert, W. and Langer, T. (1995). Hsp78, a clp homologue within mitochondria, can substitute for chaperone functions of mt-hsp70. EMBO J. , 14 , 3434–3444. Schmitt, M., Neupert, W. and Langer, T. (1996). The molecular chaperone Hsp78 confers compartment-specific thermotolerance to mitochondria. J. Cell Biol. , 134 , 1375– 1386. Schröder, H., Langer, T., Hartl, F.U. and Bukau, B. (1993). DnaK, DnaJ and GrpE form a cellular chaperone machinery capable of repairing heat-induced protein damage. EMBO J. , 12 , 4137–4144. Singer, M.A. and Lindquist, S. (1998). Multiple effects of trehalose on protein folding in vitro and in vivo. Modular Cell , 1 , 639–648. Skowyra, D., Georgopoulos, C. and Zylicz, M. (1990). The E. coli dnaK gene product,
The role of Hsp104 in stress tolerance and prion maintenance
419
the hsp70 homolog, can reactivate heat-inactivated RNA polymerase in an ATP hydrolysis-dependent manner. Cell , 62 , 939–944. Smith, B.J. and Yaffe, M.P. (1991). Uncoupling thermotolerance from the induction of heat shock proteins. Proc. Natl Acad. Sci. USA , 88 , 11091–11094. Somero, G.N., Osmond, C.B. and Bolis, C.L. (1992). Water and life , Springer-Verlag, Heidelberg. Squires, C.L., Pedersen, S., Ross, B.M. and Squires, C. (1991). ClpB is the Escherichia coli heat shock protein F84.1. J. Bacteriology , 173 , 4254–4262. Stansfield, I., Jones, K.M., Kushnirov, V.V., Dagkesamanskaya, A.R., Poznyakovski, A.I., Paushkin, S.V., Nierras, C.R., Cox, B.S., Ter-Avanesyan, M.D. and Tuite, M.F. (1995). The products of the SUP45 (eRF1). and SUP35 genes interact to mediate translation termination in Saccharomyces cerevisiae. EMBO J. , 14 , 4365–4373. Stohs, S.J. and Bagchi, D. (1995). Oxidative mechanisms in the toxicity of metal ions. Free Radical Biol. Med. , 18 , 321–336. Telling, G.C., Scott, M., Mastrianni, J., Gabizon, R., Torchia, M., Cohen, F.E., DeArmond, S.J. and Prusiner, S.B. (1995). Prion propagation in mice expressing human and chimeric PrP transgenes implicates the interaction of cellular PrP with another protein. Cell , 83 , 79–90. Ter-Avanesyan, M.D., Kushnirov, V.V., Dagkesamanskaya, A.R., Didichenko, S.A., Chernoff, Y.O., Inge-Vechtomov, S.G. and Smirnov, V.N. (1993). Deletion analysis of the SUP35 gene of the yeast Saccharomyces cerevisiae reveals two non-overlapping functional regions in the encoded protein. Mol. Microb. , 7 , 683–692. Thomas, P.J., Qu, B.-H. and Pedersen, P.L. (1995). Defective protein folding as a basis of human disease. Trends Biochem. Sci. , 20 , 456–459. Tuite, M.F. and Cox, B.S. (1980). Ultraviolet mutagenesis studies of [psi], a cytoplasmic determinant of Saccharomyces cerevisiae. Genetics , 95 , 611–630. Tuite, M.F., Mundy, C.R. and Cox, B.S. (1981). Agents that cause a high frequency of genetic change from [psi+] into [psi−] in Saccharomyces cerevisiae. Genetics , 98 , 691–711. Tuite, M.F. and Stansfield, I. (1994). Translation. Knowing when to stop. Nature , 372 , 614–615. Van Laere, A. (1989). Trehalose, reserve and/or stress metabolite? FEMS Microbiol. Rev. , 63 , 201–210. Vogel, J.L., Parsell, D.A. and Lindquist, S. (1995). Heat-shock proteins Hsp104 and Hsp70 reactivate mRNA splicing after heat inactivation. Curr. Biology , 5 , 306–317. Watson, K. (1990). Microbial stress proteins. Adv. Microbial Phys. , 31 , 183–223. Wawrzynow, A., Wojtkowiak, D., Marszalek, J., Banecki, B., Jonsen, M., Graves, B., Georgopoulos, C. and Zylicz, M. (1995). The ClpX heat-shock protein of Escherichia coli, the ATP-dependent substrate specificity component of the ClpP-ClpX protease, is a novel molecular chaperone. EMBO J. , 14 , 1867–1877. Weber, L.A. (1992). Relationship of heat shock proteins and induced thermal resistance. Cell Proliferation , 25 , 101–113. Weissman, C. (1994). Trends Cell Biol. , 4 , 10–14. Wetzel, R. (1996). For protein misassembly, it’s the “I” decade. Cell , 86 , 699–702. Wickner, R.B. (1994). [URE3] as an altered URE2 protein: evidence for a prion analog in
Molecular chaperones and folding catalysts
420
Saccharomyces cerevisiae. Science , 264 , 566–569. Wickner, R.B., Masison, B.C. and Edskes, H.K. (1995). [PSI] and [URE3] as yeast prions. Yeast , 11 , 1671–1685. Wickner, S., Gottesman, S., Skowyra, D., Hoskins, J., McKenney, K. and Maurizi, M.R. (1994). A molecular chaperone, ClpA, functions like DnaK and DnaJ. Proc. Natl. Acad. Sci. USA , 91 , 12218–12222. Wiemken, A. (1990). Trehalose in yeast, stress protectant rather than reserve carbohydrate. Antonie van Leeuwenhoek , 58 , 209–217. Wilson, S.A., Williams, R.J., Pearl, L.H. and Drew, R.E. (1995). Identification of two new genes in the Pseudomonas aeruginosa amidase operon, encoding an ATPase (AmiB). and a putative integral membrane protein (AmiS)*. J. Biol. Chem. , 270 , 18818–18824. Wojtkowiak, D., Georgopoulos, C. and Zylicz, M. (1993). Isolation and characterization of ClpX, a new ATP-dependent specificity component of the Clp protease of Escherichia coli. J. Biol. Chem. , 268 , 22609–22617. Woo, K.M., Kim, K.I., Goldberg, A.L., Ha, D.B. and Chung, C.H. (1992). The heatshock protein ClpB in Escherichia coli is a protein-activated ATPase. J. Biol. Chem. , 267 , 20429–20434. Yancey, P.H., Clark, M.E., Hand, S.C., Bowlus, R.D. and Somero, G.N. (1982). Living with water stress: evolution of osmolyte systems . Science , 217 , 1214–1222. Yost, H.J. and Lindquist, S. (1986). RNA splicing is interrupted by heat shock and is rescued by heat shock protein synthesis. Cell , 45 , 185–193. Yost, H.J. and Lindquist, S. (1988). Translation of unspliced transcripts after heat shock. Science , 242 , 1544–1548. Yost, H.J. and Lindquist, S. (1991). Heat shock proteins affect RNA processing during the heat shock response of Saccharomyces cerevisiae. Molec. Cell. Biology , 11 , 1062–1068. Zhang, H., Stockel, J., Mehlhorn, I., Groth, D., Baldwin, M.A., Prusiner, S.B., James, T.L. and Cohen, F.E. (1997). Physical studies of conformational plasticity in a recombinant prion protein. Biochemistry , 36 , 3543–3553.
18. CHAPERONES AND CHARONINS: PROTEIN UNFOLDING ENZYMES AND PROTEOLYSIS MICHAEL R. MAURIZI1,*, SUE WICKNER2 and SUSAN GOTTESMAN2 1 Laboratory
2 Laboratory
of Cell Biology and of Molecular Biology, National Cancer Institute, Bethesda, MD 20892, USA
1. Introduction 2. Chaperone-mediated Proteolysis by ATP-dependent Pro teases 2.1. Protein Degradation in vivo by ATP-dependent Proteases 2.1.1. Lon Proteases 2.1.2. Clp Proteases 2.1.3. FtsH 2.2. Structures and in vitro Activities of ATP-dependent Proteases 2.2.1. Lon Protease 2.2.2. Clp Proteases 2.2.3. FtsH 3. Chaperone-like Activities of ATP-dependent Proteases 3.1. General Considerations 3.2. ClpA Activation of RepA 3.3. ClpX Functions in Phage Replication and Transposition 3.2.1. Disassembly of the Mu Transpososome 3.2.2. Lambda O Protein 3.4. ClpB/Hsp100 Proteins: Disaggregation but no Degradation 3.5. Chaperone Activity of the Lon Protease ATPase Domain 3.6. Chaperone Activity of FtsH Homologs 4. Protein Unfolding as a Step in ATP-dependent Proteolysis 4.1. Requirement for ATP Hydrolysis 4.2. ATP Hydrolysis Drives Unfolding by Lon Protease 5. Involvement of General Chaperones in Protein Degradation 5.1. Decreased Protein Degradation in Chaperone Mutants 5.1.1. General Chaperone Mutants in E. coli 5.1.2. E. coli Chaperones Allow Protein Degradation by Preventing
Molecular chaperones and folding catalysts
422
Aggregation 5.1.3. Defects in Protein Degradation in Eukaryotic Chaperone Mutants 5.2. Pleiotropic Effects of Mutations in Chaperones 6. Substrate Selection by ATP-dependent Proteases and Chaperones 6.1. Degradation Motifs 6.2. Exposed Bonding Domains as Potential Recognition Motifs 6.3. Do Chaperones and Proteases Recognize Similar Features of Proteins? 7. Concluding Remarks
1. INTRODUCTION When protein structures become disrupted, for example in response to chemical damage or physical stress, they unfold and tend to aggregate. Similar unfolding and aggregating behavior is displayed by proteins that are not necessarily damaged but have unusual surface features or abnormal conformations, such as a few intrinsically unstable proteins, mutant proteins, unbound subunits of multimeric complexes, proteins that have lost stabilizing ligands, and proteins that are in a foreign intracellular milieu. Because unfolded proteins can interfere with the functioning of normal proteins, quality control over protein conformation and structure is important for cell growth. Cells respond to unfolded proteins by attempting to refold the proteins or by degrading them. In either case, the ability to prevent rapid aggregation and to disaggregate existing aggregates is necessary to allow refolding or degradation. What has become clear in recent years is that the two major systems that respond to misfolded proteins are closely integrated: enzymes whose primary role is protein folding are needed for rapid protein degradation, and proteins identified as integral components of pro teases can promote refolding as well as degradation. The best studied refolding enzymes are the molecular chaperones and chaperonins, which we call general chaperones. In E. coli, general chaperones include the DnaK system (see chapters Burkholder & Gottesman; Zylicz et al.; Buchberger et al., this volume) consisting of DnaK (a member of the Hsp70 family), DnaJ, and GrpE, as well as the GroE system (see chapters Burkholder & Gottesman; Burston & Saibil; Ranson & Clarke, this volume) consisting of GroEL (a member of the Hsc60 chaperonin family) and GroES. These general chaperones display a broad range of ATP-dependent protein folding activities and also promote assembly or disassembly of macromolecular complexes. Another general chaperone is ClpB/Hsp104 (Gottesman et al., 1990b; Parsell et al., 1991; Squires & Squires, 1992), which has protein disaggregating activity (Parsell et al., 1994b). ClpB is highly homologous to ClpA, a component of the ATP-dependent ClpAP pro tease (Gottesman et al., 1990b), but has not been shown to participate directly in protein degradation (Park et al., 1993; Woo et al., 1992). In addition to influencing protein folding and aggregation, general chaperones and other protein remodeling enzymes have large effects on protein degradation in vivo. While cells contain a multitude of proteases, most intracellular protein degradation is
Chaperones and charonins
423
carried out by a limited set of ATP-dependent proteases. Prototypes of all of the known ATP-dependent proteases are found in E. coli and include Lon protease, ClpAP, ClpXP, FtsH (or HflB), and ClpYQ (or HslVU). In eukaryotic cells, homologs of Lon and FtsH are found in mitochondria. The major cytosolic protease, the 26S proteasome, is structurally analogous to ClpAP (Kessel et al., 1995) and evolutionarily related to ClpYQ (Chuang et al., 1993). Sequence analysis as well as enzymatic and physical studies of the ATP-dependent proteases indicate that they are composed of two domains: a proteolytic domain and an ATPase domain that has been shown to have a protein unfolding or remodeling activity (Maurizi, 1992; Wickner et al., 1994). The ATP-dependent activities of this domain, which include unfolding the protein and facilitating its translocation to the proteolytic active sites, are essential for protein degradation. General chaperones and ATP-dependent proteases work on a common pool of substrates and may share a common mode of action in the initial steps of interaction with protein substrates. The distinction between general chaperones and the chaperone components of ATP-dependent proteases may thus be arbitrary, an accident of discovery or assay, although differences in specificities, binding affinities, and kinetics of interaction and unfolding may determine if particular chaperones participate predominantly in refolding or in degradation. Mutations in general chaperones can drastically affect degradation of proteins in vivo, but it has been difficult to demonstrate direct physical association between general chaperones and proteases or even to define a pathway in which chaperones and proteases act in an ordered fashion. The discovery that integral components of ATP-dependent proteases have molecular chaperone activity, however, has provided evidence that chaperone-assisted unfolding can be directly coupled to protein degradation. We have proposed that the chaperone components of ATP-dependent proteases be called “charonins’” after the mythological underworld creature who ferried souls across the river Styx. Chaperone components or domains of ATP-dependent proteases may in fact unfold proteins exclusively for degradation, or they could first attempt to release the protein so it can refold. Conversely, general chaperones may also have a “charonin” function, taking an active role in promoting degradation when a protein fails to refold. We do not know the extent to which chaperone and charonin functions overlap. The relative contributions of different chaperones and proteases to either refolding a particular protein or to elimination of an intransigent unfolded protein are likely to vary widely. In this chapter we will summarize current data about the properties and activities of general chaperones and of the chaperone components of ATP-dependent proteases with the following questions in mind: • Do general chaperones and protein remodeling enzymes act directly or indirectly in affecting proteolysis—that is, do proteases recognize and degrade the “chaperonebound” or the “released” form of unfolded proteins? • What chemical or structural motifs in unfolded proteins are recognized by chaperones and proteases and are they the same? • And perhaps the most critical question—what is the nature of the decision checkpoint between refolding and degradation? As will be seen, we have only begun to obtain answers to some of these questions, and
Molecular chaperones and folding catalysts
424
we are almost completely in the dark regarding mechanistic issues. However, our base of knowledge is now sufficiently broad and solid to provide context for obtaining more profound answers in the near future. 2. PROTEOLYTIC FUNCTIONS AND STRUCTURES OF ATP-DEPENDENT PROTEASES 2.1. Protein Degradation in vivo 2.1.1. Lon Proteases Lon protease (protease La) from E. coli was identified on the basis of its ability to degrade proteins in vitro in the presence of ATP and was the first ATP-dependent protease to be purified (Goldberg, 1992; Goldberg et al., 1994). E. coli Lon is the product of a single gene, lon, which is required in vivo for the degradation of proteins such as SulA, RcsA, N protein, CcdA, and Tn903 transposase (Gottesman & Maurizi, 1992). lon mutants display phenotypes, such as UV sensitivity, mucoidy, and susceptibility to post-segregational killing after loss of certain plasmids, as a consequence of defects in degradation of specific Lon substrates. Lon can also degrade certain mutant proteins in vivo (Sherman & Goldberg, 1992) and is responsible for = 50% of the degradation of abnormal proteins, such as puromycyl fragments and canavanine-containing proteins (Bukhari & Zipser, 1973; Maurizi et al., 1985). Lon degrades the N protein (Maurizi, 1987) and CcdA (Van Melderen et al., 1996) in a purified in vitro system, which demonstrates that no other proteins are essential for expression of Lon activity. Lon proteases have important functions during growth and development in other bacteria. In Myxococcus, two Lon proteases are found; one is essential during vegetative growth and the other plays an unknown role during stalk formation and myxospore maturation (Gill et al., 1993; Tojo et al., 1993a; Tojo et al., 1993b). In Caulobacter, lon mutations lead to developmental defects attributable in part to the failure to degrade CcrM, an adenine DNA methyltransferase involved in cell-cycle dependent variation in DNA methylation (Wright et al., 1996). Lon protease has been identified in eukaryotic cells where it is encoded in chromosomal DNA but is found, apparently exclusively, in the mitochondrial matrix (Suzuki et al., 1994; Van Dyck et al., 1994; Wang et al., 1993). In vitro properties of eukaryotic Lon suggest it is similar to E. coli Lon in promoting ATP-dependent protein degradation (Desautels & Goldberg, 1982; Wang et al., 1993; Watabe & Kimura, 1985). Yeast cells with mutations in LON (also called PIM1) accumulate deletions in mitochondrial DNA, are respiration deficient, cannot grow on non-fermentable carbon sources, and accumulate dense inclusion bodies within mitochondria (Suzuki et al., 1994; Van Dyck et al., 1994). Yeast LON mutations lead to stabilization of several mitochondrial proteins, including the subunit of the matrix processing peptidase and the , , and subunits of F1 ATPase (Suzuki et al., 1994).
Chaperones and charonins
425
2.1.2. Clp Proteases ClpAP (also referred to as protease Ti (Woo et al., 1989) was isolated from E. coli lon mutant cell extracts based on its ability to degrade casein in an ATP-dependent manner (Katayama-Fujimura et al., 1987; Maurizi, 1992). ClpAP is composed of two separate gene products, ClpA, an ATPase, and ClpP, a peptidase, which can be purified separately and reconstituted to obtain the active protease (Maurizi, 1992). In vivo, ClpAP appears to play a minor role in the degradation of abnormal proteins, such as canavanyl proteins (Gottesman et al., 1990a) but targets certain -galactosidase fusion proteins quite specifically (Gottesman et al., 1990a; Tobias et al., 1991). ClpAP is responsible for degradation of abnormal -galactosidase proteins with bulky hydrophobic amino terminal residues (Leu, Phe, and Trp) and thus appears to recognize the degradation motif displayed by those proteins whose in vivo stability follows the N-end rule (Tobias et al., 1991). The only natural substrate for ClpAP identified to date is MazE (Aizenman et al., 1996), a low molecular weight protein homologous to antidote proteins, such as PemI, which are part of plasmid post-segregational killing systems (Jensen & Gerdes, 1995). The function of MazE in E. coli and the consequence of its degradation are not known, although it appears to be regulated as part of the stringent response (Aizenman et al., 1996). ClpP is also activated in vivo and in vitro by ClpX (Gottesman et al., 1993; Wojtkowiak et al., 1993), a protein with sequence similarity to ClpA. The clpP and clpX genes are co-transcribed in vivo (Gottesman et al., 1993), but the two proteins are easily separated in cell extracts and must be added together in the presence of ATP to reconstitute the active protease (Wojtkowiak et al., 1993; Grimaud, et al., 1998). ClpXP does not degrade the same set of proteins degraded by ClpAP indicating that the ATPase component confers specificity to the protease. In vivo, ClpXP degrades the E. coli starvation sigma factor, RpoS (Schweder et al., 1996) several unidentified proteins made during carbon starvation (Damerau & St. John, 1993), and several phage proteins: the O protein (Wojtkowiak et al., 1993), PhD, an antidote protein made by phage P1 (Lehnherr & Yarmolinsky, 1995), and mutant forms of the Mu repressor (MhammediAlaoui et al., 1994). Another member of the Clp protease family, ClpYQ, was originally identified by sequence analysis (Chuang et al., 1993). ClpY is homologous to ClpX (Gottesman et al., 1993) and is encoded in an operon along with ClpQ, which is homologous to the subunit of the 20 S proteasome (Chuang et al., 1993). A mixture of purified ClpY and ClpQ has been shown to have peptidase activity (Rohrwild et al., 1996; Yoo et al., 1996) as well as ATP-dependent protein degrading activity (Kessel et al., 1996; Huang and Goldberg, 1997). In vivo, ClpYQ may affect the levels of the acute heat shock sigma factor, , but no other targets have been identified (Missiakas et al., 1996). 2.1.3. FtsH (HflB) The ftsH gene is essential in E. coli and encodes an ATP-dependent protease found in the cytoplasmic membrane (Tomoyasu et al., 1995). Mutants of ftsH are temperature sensitive for growth, possibly as a result of defective translocation of membrane spanning
Molecular chaperones and folding catalysts
426
proteins, and these mutations can be partially suppressed by overproduction of GroEL/GroES or Hsp90 (Shirai et al., 1996). ftsH was originally identified as hflB (high frequency lysogenization), because of its effect on the frequency of lambda lysogenization (Banuett et al., 1986). Mutations in ftsH lead to stabilization of two regulators of lambda lysogenization, CII (Banuett et al., 1986) and CIII (Herman et al., 1997). FtsH is responsible for the rate limiting step in degradation of (Herman et al., 1995; Tomoyasu et al., 1995) (see chapter Connolly et al., this volume) and for degradation of unassembled subunits of the membrane-associated secretory protein, SecY (Kihara et al., 1995). Homologs of FtsH have been identified in several organisms. In yeast, the FtsH homologs YTA-10 and YTA-12 can degrade newly synthesized proteins in isolated mitochondria, and this activity depends on an intact Zn-metalloprotease site (Arlt et al., 1996). 2.2. Structures and in vitro Activities of ATP-dependent Proteases 2.2.1. Lon Protease Lon (subunit Mr 89,000) is a soluble oligomeric protein whose exact structure remains to be definitively determined (Goldberg et al., 1994; Watabe & Kimura, 1985). From sequence analysis, Lon appears to be divided into two domains, an amino terminal ATPase and a carboxy terminal pro tease (Chin et al., 1988; Goldberg et al., 1994; Gottesman et al., 1995). The ATPase domain contains a Walker consensus ATP-binding motif, and mutations in that site eliminate the ATP-dependent protein degrading activity of Lon (Fischer & Glockshuber, 1994). The carboxy terminal region has no obvious sequence similarity to other pro teases, although Lon is inhibited by reagents that inactivate classical serine pro teases (Goldberg et al., 1994). A mutation in Ser679 eliminates protease and peptidase activity (Fischer & Glockshuber, 1993; Yu et al., 1991), but the mutant Lon-S679A retains ATPase as well as a chaperone-like activity (see below). Purified Lon protease can cleave low molecular weight peptides in the absence of nucleotide, and this function is activated by non-hydrolyzable analogs of ATP (Waxman & Goldberg, 1982; Waxman & Goldberg, 1986). Degradation of various partially denatured proteins in vitro shows a more complex response to nucleotide. For some denatured proteins, there is a stoichiometric coupling of two ATP molecules hydrolyzed per peptide bond cleaved (Goldberg & Waxman, 1985; Menon et al., 1987). The requirement for ATP hydrolysis, however, is dependent on a combination of the size and the structure of the protein substrate, such that ATP hydrolysis is absolutely required for degradation of some (generally higher molecular weight) proteins (Goldberg & Waxman, 1985; Menon et al., 1987), but non-hydrolyzable analogs can activate degradation of smaller (Mr=30,000) proteins (Maurizi, 1987; Van Melderen et al., 1996). Protein substrates bind to Lon and activate the ATPase activity (Waxman & Goldberg, 1986) but stimulation of ATP hydrolysis is not dependent on peptide bond cleavage, because ATPase activity of a mutant Lon lacking the catalytic serine residue was still activated by protein binding (Fischer & Glockshuber, 1993).
Chaperones and charonins
427
2.2.2. Clp Proteases ClpP shares no motifs common to the trypsin and subtilisin families but does contain an active site serine, histidine, and aspartate triad characteristic of serine proteases (Maurizi et al., 1990a; Maurizi et al., 1990b). ClpQ (HslV), which by sequence comparison is unrelated to ClpP, belongs to a newly discovered family of threonine proteases that includes the Archae and eukaryotic proteasomes (Seemuller et al., 1995). ClpP (subunit Mr 21,400) is composed of two axially aligned rings of seven subunits each. The ClpP structure is reminiscent of the β subunits of the proteasome, which has 14 active sites located on the interior surface of an aqueous cavity formed by the joining of two concave seven-membered rings (Lowe et al., 1995). ClpQ (subunit Mr 19,000) also is composed of two superimposed rings, but, despite its homology with the proteasome, ClpQ has only six subunits in each ring and the proteolytic cavity should then have 12 rather than 14 active sites (Kessel et al., 1996; Rohrwild et al., 1997). How this difference affects efficiency of cleavage or the size of the average peptide products generated by ClpQ and proteasomes should provide valuable information about the mechanism of these proteases. The ATPases that define the Clp family can be divided into two subfamilies, one represented by ClpA and ClpB, which possess two ATPase domains, and the other represented by ClpX and ClpY, which possess a single ATPase domain (Gottesman et al., 1995). ClpA and ClpB are highly homologous (64% overall similarity and 88% similarity in approximately 200 amino acids in each of the core ATPase regions) (Gottesman et al., 1990b). ClpX and ClpY are homologous to each other (55% overall similarity) but somewhat more distantly related to ClpA (~50% similarity in the ATPase region and Cterminal domain) (Gottesman et al., 1993; Gottesman et al., 1995). Despite its similarity to ClpA, ClpB has not been shown to activate proteolysis (Park et al., 1993; Woo et al., 1992) by ClpP or any other proteases. ClpA subunits (Mr 84,000) associate to form six-membered rings in response to binding of ATP or non-hydrolyzable analogs of ATP. The two ATPase domains of ClpA appear to form structurally distinct rings (Kessel et al., 1995). In the ClpAP complex, a hexamer of ClpA binds on each face of ClpP forming a barrel-like molecule. The axial channels through ClpA and ClpP into the proteolytic active sites appear constricted and may have to expand in response to protein binding or ATP hydrolysis in order to allow passage of unfolded proteins. ClpX (subunit Mr 46,000) is purified as a heterogeneous mixture of oligomers and forms stable hexameric rings in the presence of ATP or ATP S (Grimaud et al., 1998). Complexes of ClpX and ClPP resemble ClpAP, with a hexameric ring of ClpX bound to each face of ClpP. ClpX and ClpA have similar affinities for ClpP, and complexes with CIPA and ClpX bound to opposite faces of ClpP have been observed (Grimaud et al., 1998). ClpY (subunit Mr 49,000) associates in the presence of ATP to form two types of oligomeric rings (Kessel et al., 1996). Both six-membered and seven-membered rings of ClpY have been observed (Kessel et al., 1996; Rohrwild et al., 1997). The significance of the heterogeneity in ring structures is not known, but it might affect substrate selection. Since the central channels through the six- and seven-membered rings should have different diameters, the size of proteins that interact with or are translocated to ClpP may
Molecular chaperones and folding catalysts
428
be different. ClpP alone has activity against low molecular weight peptides, but its active site appears to be inaccessible to large or folded polypeptides and proteins (Thompson & Maurizi, 1994; Woo et al., 1989). In the presence of ClpA, ClpP can degrade longer peptides as well as proteins; the requirement for ATP hydrolysis depends on some combination of the size and structure of the substrate. Non-hydrolyzable analogs of ATP stimulate cleavage of some polypeptides (10–30 amino acids) at remarkably rapid rates (>11,000 peptide bonds per minute per tetradecamer of ClpP) (Thompson & Maurizi, 1994). Protein degradation by ClpP in the presence of ClpA requires ATP hydrolysis (Thompson et al., 1994; Woo et al., 1989). ClpA in the absence of ClpP has a substantial ATPase activity (~600 ATP per min per ClpA6) which is activated only 20–40% in the presence of protein substrates with or without ClpP (Hwang et al., 1988; Singh & Maurizi, 1994). Thus, ATP hydrolysis is not coupled in a concerted mechanistic manner to peptide bond cleavage, and the stoichiometric ratio of ATP hydrolysis to peptide bond cleavage (4–8 ATP per peptide bond) is variable (Maurizi et al., 1994). The peptide bond cleavage rate in protein substrates is 120–180 per minute per ClpAP complex, about 50 times slower than the fastest rates seen with peptides, indicating that the rate limiting step in protein degradation is not peptide bond cleavage but some reaction associated with the ATP hydrolysis. In the presence of ClpX, ClpP can degrade specific substrates (such a O protein) in vitro, and this activity also requires ATP hydrolysis (Wojtkowiak et al., 1993). ClpQ by itself can degrade small fluorogenic peptides (Wu, Gottesman, Maurizi, unpublished) but peptide cleavage can be activated by ClpY in the presence of ATP (Rohrwild et al., 1996; Yoo et al., 1996). ClpY and ClpQ can also degrade proteins, such as casein, when ATP hydrolysis is allowed but not in the presence of non-hydrolyzable analogs of ATP (Kessel et al., 1996; Huang and Goldberg, 1997). 2.2.3. FtsH FtsH is composed of a single polypeptide chain (Mr 71,000) that forms an oligomeric integral membrane protein (Tomoyasu et al., 1995). Sequence analysis predicts that FtsH has two membrane spanning regions and that the C-terminal 530 amino acids constitute a large cytoplasmic domain containing a 200-amino acid region conserved among the AAA family of ATPases (Tomoyasu et al., 1995) and a putative carboxy terminal Znmetallopeptidase domain. Purified detergent solubilized FtsH has ATPase activity and could degrade a specific protein substrate, , in the presence of ATP (Tomoyasu et al., 1995), suggesting that FtsH alone is capable of carrying out ATP-dependent protein degradation. 3. CHAPERONE-LIKE ACTIVITIES OF ATP-DEPENDENT PROTEASE COMPONENTS 3.1. General Considerations The evidence that components or domains of ATP-dependent proteases have activity
Chaperones and charonins
429
similar to that of molecular chaperones is now incontrovertible. Direct evidence for chaperone-like activities has been obtained in purified systems, and biological activity has been shown for ATPase domains in the absence of enzymatically active proteolytic domains. Both ClpA and ClpX are able to bind to thermally-unfolded proteins and protect them from irreversible denaturation, though neither protein has been shown to catalyze refolding of thermally unfolded proteins, as has been shown for general chaperones (Wawrzynow et al., 1995; Wickner et al., 1994). More convincing data has been obtained in studies showing that ClpA, ClpX, and proteolytically inactive mutants of Lon protease can help assemble or disassemble specific protein complexes. These studies show clearly that the ATPase components of proteases can remodel protein structure. 3.2. ClpA Activation of RepA ClpA interacts with an inactive dimeric form of the RepA protein of phage P1 and converts it to an active monomeric form that binds specific oriP1 DNA sequences (Wickner et al., 1994). The inactive dimer of RepA binds tightly to ClpA in the presence of non-hydrolyzable analogs of ATP, and activation and release of the active monomer requires ATP hydrolysis. RepA activation was originally shown to be catalyzed by the DnaJ/K chaperone system and to be accomplished non-enzymatically by dissociation of RepA dimers with mild denaturants such as urea (Wickner et al., 1991). ClpA thus has the activities of a molecular chaperone in binding and catalyzing partial unfolding of proteins. ClpA will also promote the ATP-dependent degradation of RepA if ClpP is present (Wickner et al., 1994). Do the same unfolding reactions that lead to activation of RepA also lead to degradation? The answer is not entirely clear but several bits of data suggest a close link. Dimers of RepA bind with the same stoichiometry to ClpA and to ClpAP complexes in the presence of non-hydrolyzable ATP analogs, and the dimer of RepA has a lower Km for degradation than the monomer, suggesting that ClpA acts on the same form of RepA for both reactions (Pak and Wickner, unpublished). Also, when oriP1 DNA, which protects active RepA from degradation, is present from the start of a reaction with ClpAP and RepA, some of the RepA is activated and binds DNA, but some of the RepA is rapidly degraded. Thus, some RepA unfolded by ClpA can be captured and degraded before it is released to refold into an active form. The timing of unfolding and release versus unfolding and capture is the key, still missing, to a more complete understanding of chaperone-mediated proteolysis. 3.3. ClpX Functions in Phage Replication and Transposition 3.3.1. Mu Transposase ClpX displays chaperone-like activity with several proteins in vitro and experiments suggest that, in vivo, ClpX may have activity, presumably as a chaperone, independently of ClpP (Mhammedi-Alaoui et al., 1994). Purified ClpX promotes disassembly of Mu transposase (MuA) from the highly stable strand transfer complex, which is an obligatory intermediate in Mu transposition (Kruklitis et al., 1996; Levchenko et al., 1995). Displacement of the MuA tetramer from the strand transfer complex requires ATP
Molecular chaperones and folding catalysts
430
hydrolysis and results in no covalent modification of the protein. The released MuA is active in a subsequent round of Mu recombination (Levchenko et al., 1995). In vivo, ClpX is essential for Mu growth, however ClpP is not, suggesting that it is the chaperone activity of ClpX that is required for Mu replication (Mhammedi-Alaoui et al., 1994). 3.3.2.
O Protein
Zylicz and co-workers demonstrated in a purified system that ClpX can protect O protein, which is a specific substrate for degradation by ClpXP in vivo and in vitro (Wawrzynow et al., 1995), from heat-induced aggregation and can disaggregate preformed aggregates of O protein. Binding of O protein stimulates ATP hydrolysis by ClpX, which is required for the disaggregation reaction. 3.4. ClpB/Hsp104: Disaggregation but no Stimulation of Degradation Purified ClpB, which has over 80% sequence similarity with ClpA, does not appear to interact with ClpP (Maurizi, unpublished), and biochemical studies have failed to show protein degradation mediated by E. coli ClpB (Woo et al., 1992) or the homologous yeast Hsp104 (Parsell et al., 1994a). E. coli ClpB mutants (Squires et al., 1991) and yeast Hsp104 (Parsell et al., 1994b) mutants tend to accumulate protein inclusion bodies, and Hsp104 mutants are sensitive to acute heat stress or to combinations of two or more moderate stresses (Parsell et al., 1994b; Sanchez et al., 1992). These and other data suggest that members of the ClpB subfamily catalyze the disaggregation of proteins (Schirmer et al., 1996). For example, activity of thermally unfolded firefly luciferase is recovered at low temperature in wild type yeast cells but not in HSP104 mutants (Parsell et al., 1994b), and purified Hsp104 can increase the recovery of RNA splicing activity when added back to heated yeast extracts lacking Hsp104 (Vogel et al., 1995). Hsp104 also appears to be necessary for the propagation of a prion-like particle in yeast cells (Paushkin et al., 1996) (for details see chapter Lindquist et al., this volume). Aggregates of PSI, a translation termination protein, are stable in yeast cells and are passed on to daughter cells, where they aggregate with newly synthesized PSI, thereby maintaining the steady state concentration of free PSI at a sub-optimal level which leads to read-through of stop codons. Moderate levels of Hsp104 are required to maintain this self-propagating aggregate of PSI, whereas aggregation is not initiated in the absence of Hsp104, and preexisting aggregates disappear when Hsp104 is in excess (Paushkin et al., 1996). A homolog of ClpB/Hsp104, HSP78, is found in mitochondria, where it does not appear to have an essential function (Schmitt et al., 1995). However, deletion of HSP78 in yeast cells carrying a temperature sensitive mt-Hsp70 results in a petite phenotype and impaired growth on non-fermentable carbon sources even at the permissive temperature. Overproduction of HSP78 suppresses a defect in import of mitochondrial proteins in the mt-Hsp70 strain but failed to repair a defect in degradation of abnormal proteins (Schmitt et al., 1995). HSP78 may have an auxiliary role in mitochondrial protein assembly and may be more restricted in its ability to recognize unfolded or unassembled proteins.
Chaperones and charonins
431
3.5. Chaperone Activity of the Lon Protease ATPase Domain An intriguing, although indirect, demonstration of chaperone-like activity for an ATPdependent protease is the ability of multicopy yeast Lon to complement mutations in yeast genes, AFG3 (YTA-10) and RCA1 (YTA-12), which encode membrane-associated protease/chaperones involved in assembly of heterologous inner membrane complexes in yeast mitochondria (Rep et al., 1996). Complementation did not depend on the proteolytic activity of the overexpressed Lon, because a mutant Lon in which the active site serine was mutated to alanine showed enhanced assembly of cytochrome c oxidase. Mutating the ATPase site of Lon did abolish its ability to promote assembly (Rep et al., 1996). These results suggest that Lon can interact with intermediates in an assembly pathway and possibly prevent them from aggregating before they can find their appropriate partners in the complex. Whether Lon plays a passive role by interacting with exposed bonding domains of the unassembled subunits or can promote remodeling of the subunit structure is not clear from the available data. These data also confirm the ability of ATP-dependent proteases to specifically bind substrates that are recognized by other molecular chaperones. 3.6. Chaperone Activity of FtsH Homologs Yeast YTA-10 (AFG3) and YTA-12 (RCA1), homologs of E. coli FtsH, form an 850 kDa membrane-bound complex with both proteolytic activity and chaperone-like activity in assembly of mitochondrial inner-membrane protein complexes (Arlt et al., 1996). Mutations in YTA-10 or YTA-12 cause defects in respiration and impair assembly of ATP synthase in mitochondria (Arlt et al., 1996). Activity could be reintroduced with plasmids encoding either gene, and mutants of YTA-10 or YTA-12, in which a critical histidine residue in the Zn-metalloprotease site was altered, were able to promote ATP synthase assembly and restore respiration. Unassembled subunits of membrane complexes were degraded in wild type cells, but were stable when the cells were mutated in YTA-10 or YTA-12. Thus, it appears that the YTA10–12 complex mediates not only assembly of the inner membrane complexes but also degradation of any unassembled subunits (Arlt et al., 1996). 4. PROTEIN UNFOLDING AS A STEP IN ATP-DEPENDENT PROTEOLYSIS 4.1. Requirement for ATP Hydrolysis Much enzymatic and structural data has suggested that the energy-dependent step in protein degradation involves an ATPase component of the protease acting on the protein prior to cleavage of the polypeptide chain. First, ATP hydrolysis and peptide bond cleavage are not coupled stoichiometrically or mechanistically, since the ATPase activity is activated by protein binding independently of peptide bond cleavage, and increasing the rate of peptide bond cleavage occurs upon nucleotide binding but does not require ATP hydrolysis (Fischer & Glockshuber, 1993; Thompson & Maurizi, 1994). Second, the
Molecular chaperones and folding catalysts
432
ATPase components confer specificity, which suggests that productive interaction with the substrate should first occur at a recognition site in the ATPase domain. Third, the structures of the proteases suggest that the openings to the proteolytic sites are blocked by the ATPases and only unwound polypeptide chains can use the limited access channels. With Lon, ClpAP, and the proteasome only relatively short peptides, which presumably can pass easily through the access channels, can be degraded without ATP hydrolysis. 4.2. ATP Hydrolysis Drives Unfolding by Lon Protease Recent studies have shown that the presence of secondary structure in a substrate determines whether ATP hydrolysis is needed for its degradation. CcdA, an F plasmidencoded protein that is a physiological substrate for Lon (Van Melderen et al., 1994), is a small (72 amino acid), dimeric protein with an extended structure composed of ~55% helix (Van Melderen et al., 1996). CcdA is degraded by purified Lon protease, and degradation at 25–37°C requires ATP hydrolysis (Van Melderen et al., 1996). A variant of CcdA, CcdA41, which is composed of the carboxy-terminal 41 amino acids, has some of the in vivo and in vitro activities of CcdA, but lacks stable secondary structure. CcdA41 is degraded by Lon at a similar rate and cleaved at the same peptide bonds as CcdA; however, CcdA41 degradation is activated by non-hydrolyzable analogs of ATP and ATP hydrolysis does not alter its rate of degradation. Furthermore, with CcdA at 50° C, which is the melting temperature for the secondary structure, degradation by Lon also proceeds with non-hydrolyzable analogs of ATP at =25% of the rate with ATP. These data suggest that ATP hydrolysis is used to drive disruption of the secondary structure of protein substrates. The ability of ATP-dependent proteases to disrupt protein structure is limited, and they can degrade only proteins that have less stable tertiary structures or have been “denatured” to some extent, for example, ccasein, which exists in a molten globule state. Proteins that have stable quaternary or tertiary structures are generally resistant to proteolysis even by ATP-dependent proteases. Susceptible proteins should have a structure that we can describe as being “deformable,” meaning that the energy barriers between various conformational states of the protein bound to the protease are relatively low. ATP hydrolysis is required to drive the transitions between states. Presumably the protease then has a mechanism for recognizing and trapping the more extended, unfolded states needed to allow the polypeptide chain to be translocated into the proteolytic sites. 5. INVOLVEMENT OF GENERAL CHAPERONES IN PROTEIN DEGRADATION 5.1. Decreased Protein Degradation in General Chaperone Mutants 5.1.1. General Chaperone Mutants in E. Coli In E. coli, mutations in the major heat shock regulator, , lead to lower expression of Hsps, including both chaperones and proteases, and result in defects in the degradation of “abnormal proteins” such as truncated polypeptides produced by disruption of translation
Chaperones and charonins
433
with puromycin and misfolded proteins formed by incorporation of the arginine analog, L-canavanine (Baker et al., 1984; Goff et al., 1984). These abnormal proteins, as well as several mutant proteins, are also degraded more slowly in E. coli mutants with defects in major chaperones such as DnaK, DnaJ, GrpE, or GroEL (Jubete et al., 1996; Kandror et al., 1994; Keller & Simon, 1988; Sherman & Goldberg, 1992; Straus et al., 1988). A significant portion of the proteins whose degradation is affected by chaperones are targeted by ATP-dependent proteases (Jubete et al., 1996; Straus et al., 1988). When degradation of individual canavanine-containing proteins was analyzed by pulse-chase labelling and one-dimensional SDS-PAGE electrophoresis, the same proteins were affected by both lon and dnaJ mutations (Jubete et al., 1996). Mutations in lon or chaperone genes affect degradation of specific mutant proteins as well. Lon-dependent degradation of the temperature sensitive PhoA61 protein was almost completely blocked by a mutation in dnaK (Sherman & Goldberg, 1992), and a mutation in groEL blocked degradation of an abnormal fusion protein, CRAG, which appears to be a substrate for ClpP (Kandror et al., 1994). Curiously, CRAG degradation in vivo is not affected by mutations in either ClpA or ClpX, suggesting either that this protein is degraded by ClpP itself or that degradation by ClpP can be activated by a more general chaperone (Kandror et al., 1994; Kandror et al., 1995). No in vitro data supports the latter suggestion however. CRAG degradation was also correlated with the intracellular levels of trigger factor (Kandror et al., 1995). These results demonstrate that various enzymes that catalyze protein remodeling can act in concert with ATP-dependent proteases to affect the rate of protein degradation in E. coli. 5.1.2. E. Coli Chaperones Allow Protein Degradation by Preventing Aggregation General chaperones could affect protein degradation by 1) promoting an active conformation of the pro tease or 2) affecting the accessibility or the conformation of the protein substrate. Purified Lon and Clp proteases are active in vitro (see above) in the absence of chaperones such as DnaK or GroEL (Goldberg et al., 1994; Maurizi et al., 1994). In vivo, natural substrates of Lon, such as SulA and N protein, were degraded at the same rates in wild type and in dnaJ mutant cells (Jubete et al., 1996), and a LacZ fusion protein targeted by ClpA was also degraded in a dnaJ mutant (Maurizi, unpublished). Thus, chaperones are not essential to promote active conformations of either of these proteases, which recognize at least some substrates that are not bound to and have not been modified by chaperones. Chaperones may facilitate protein degradation by preventing or reversing aggregation of potential protein targets. In E. coli cells, degradation of RcsA, a specific substrate of Lon protease, is dependent on a functional DnaJ (Jubete et al., 1996). In cell extracts from dna/mutants, RcsA and SulA are found in an insoluble fraction, whereas in dnaJ + cells, the major fraction of these proteins is soluble even when they accumulate to nonphysiological levels in lon mutant cells (Jubete et al., 1996). Thus, DnaJ contributes to RcsA and SulA degradation by maintaining the protein in a non-aggregated state. DnaJ’s effect on abnormal protein turnover may reflect a similar mechanism. A large fraction of abnormal canavanine-containing proteins is found in an insoluble form, presumably
Molecular chaperones and folding catalysts
434
inclusion bodies. In wild-type cells, degradation is accompanied by a redistribution of the canavanyl proteins from the insoluble to the soluble fraction. In dnaJ mutants there is reduced degradation and an increase in insoluble protein with time. ClpB/Hsp104, which has a protein disaggregating activity, can also affect degradation of proteins in vivo. T4 lysozyme expressed in E. coli cells is unstable but degradation is blocked in clpB mutant cells, and the accumulated lysozyme is found in inclusion bodies (Inoue & Rechsteiner, 1994). Since ClpB does not activate proteolysis directly, these data further indicate that chaperones can promote degradation of a protein by preventing its aggregation. 5.13. Defects in Protein Degradation in Eukaryotic Chaperone Mutants In eukaryotic cells, degradation in mitochondria, in the endoplasmic reticulum (ER), and in the cytosol is also dependent on molecular chaperones. In vitro synthesized proteins taken up by isolated mitochondria but mistargeted to the matrix are degraded by PIM1 protease, the yeast Lon homolog (Suzuki et al., 1994). Degradation of these abnormal proteins is reduced in the absence of Mdj1 or mtHsp70 (Wagner et al., 1994), yeast mitochondrial homologs of E. coli DnaJ and DnaK, respectively. Thus, there is a functional interaction between yeast mitochondrial chaperones and ATP-dependent proteases. Protein that accumulates in PIM1 mutants wild-type for the chaperones is found in the soluble fraction. However, in the chaperone mutants, the protein is found either bound to mt-Hsp70 or in insoluble aggregates (Wagner et al., 1994). Protein degradation is also defective in yeast mutants lacking the specific cytosolic chaperone, Ydj1 (Lee et al., 1996; Yaglom et al., 1996). Degradation rates of abnormal proteins containing azetidine carboxylate as well as ubiquitin-proline- -galactosidase and Gcn4- -galactosidase, which are subject to ubiquitin-dependent degradation, were reduced 50–80% at the non-permissive temperature in a temperature sensitive Ydj1 mutant (Lee et al., 1996). Degradation of other ubiquitin-dependent substrates (namely, the N-end rule substrates, arginine- or leucyl -galactosidase and the B-type cyclin Clb5 -galactosidase) were not affected, indicating that the chaperones act on specific substrates or on specific components of the degradative pathway. Ydj 1 mutants are also defective in degradation of native forms of the Gl cyclin, Cln3 as well as -galactosidase fusions to this protein (Yaglom et al., 1996). For Cln3, degradation is preceded by phosphorylation, which was also blocked in the YDJ1 mutant (Lee et al., 1996; Yaglom et al., 1996). Unassembled immunoglobulin (Ig L) light chains are degraded in an intracellular compartment, probably the ER (Beggah et al., 1996). Degradation of these chains is dependent on interaction with BiP, a DnaK/Hsp70 homolog found within the ER, and degradation of variants of the Ig L chains was inversely proportional to their rates of release from BiP (Beggah et al., 1996). 5.2. Pleiotropic Effects of Chaperone Mutations Several factors contribute to the extreme difficulty of interpreting the effects of general chaperones on protein degradation in vivo; these cautions should be kept in mind when considering the results of in vivo studies. 1) Chaperone mutants are generally sick and do not grow well, especially at temperatures above 32°C; consequently, extragenic
Chaperones and charonins
435
suppression of phenotypes is common, though poorly documented. 2) Unfolded protein accumulating in chaperone mutants leads to induction of other chaperones and also ATPdependent pro teases. 3) The accumulated abnormal protein in chaperone mutants can be expected to interfere with the activity of other chaperones and proteases. 4) Proteins targeted by chaperones and proteases are structurally heterogeneous and prone to aggregation, thus the stability of two different proteins may be affected oppositely by mutations in chaperones. Although chaperones appear to facilitate degradation of proteins, particularly abnormal ones, examples of chaperone mutations enhancing protein degradation have also been reported. In E. coli, a temperature sensitive mutant of LacI is degraded more rapidly in dnaK mutants than in dnaK + strains (Keller & Simon, 1988), and the catabolite regulatory protein (CRP) is degraded more rapidly in a dnaJ mutant than in dnaJ + cells (Ohki et al., 1992). It is possible that the chaperone mutations lead to induction of a protease that targets these proteins, but it is more likely that the chaperones are required for proper folding or oligomerization of the proteins and that in the absence of the chaperones, the misfolded or unassembled proteins are susceptible to proteolysis. 6. SUBSTRATE SELECTION BY ATP-DEPENDENT PROTEASES AND CHAPERONES 6.1. Degradation Motifs ATP-dependent proteases probably do not have unique specificity, although different proteases show a high degree of selectivity towards different substrates. The basis for selectivity may lie in sequence or structural motifs recognized by the proteases, or it may reflect extrinsic factors such as localization, the presence of competing substrates, or the participation of auxiliary proteins. Protein degradation in eukaryotic cells (other than that occurring in mitochondria) depends on ubiquitin conjugated proteins being uniquely recognized and degraded by the 26 S proteasome (Hershko & Ciechanover, 1992). However, the existence of the ubiquitin-dependent degradation simply redirects the question of selectivity to the ubiquitination system- how are proteins targeted by the ubiquitination enzymes? That question remains to be answered. In prokaryotic cells, proteases interact directly with substrate proteins. Two degradation motifs have been proposed: 1) bulky hydrophobic amino terminal residues (the N-end rule) (Varshavsky, 1992), and 2) alanine-rich hydrophobic carboxy terminal sequences (Keiler et al., 1996; Parsell et al., 1990). Other degradation signals, such as PEST sequences (Rogers et al., 1986) and “degradation boxes” (Glotzer et al., 1991) have not been shown to occur in rapidly degraded proteins or to influence degradation in prokaryotic cells. Proteins with the N-end degradation signal (degron) are degraded by ClpAP, but this signal is not obligatory in ClpAP substrates (Gottesman et al., 1990a; Thompson & Maurizi, 1994; Tobias et al., 1991). Conditions that give rise to proteins bearing the N-end degron are not known. Proteins with the alanine-rich tail degradation signal are targeted by several proteases, including ClpXP and ClpAP (Gottesman et al., 1998), but again the substrate range of these proteases is not limited to proteins carrying that motif. A specific alanine-rich tail (AANDENYALAA) is enzymatically linked to
Molecular chaperones and folding catalysts
436
nascent polypeptide chains whose synthesis cannot be completed because their messenger RNAs have been truncated (Keiler et al., 1996), which represents an elegant example of signal transduction in a degradative pathway. Further confounding this issue is the observation that the carboxy-terminal amino acids of MuA, a substrate for the ClpX chaperone activity, can be transferred to a heterologous protein causing it to become a substrate for ClpXP (Levchenko et al., 1997). Although some ClpX or ClpXP substrates have carboxy terminal regions similar to that of MuA, there are at least three dissimilar classes of ClpX substrates based on carboxy terminal sequences, suggesting that ClpX might have binding sites for several different motifs. Carboxy-terminal sequences could constitute part of or all of specific recognition motifs that target a limited number of proteins for degradation by ClpXP. Alternatively, high specificity of interaction is reserved for specialized functions (such as disassembly of the Mu strand transfer complex). Degenerate versions of these sequences might be found in various proteins, sometimes at the carboxy terminus or sometimes, as in RpoS, not. When exposed, these sequences may help to target proteins for degradation by ClpXP. 6.2. Exposed Bonding Domains as Potential Recognition Motifs Proteins targeted for rapid degradation generally have some structural perturbation. The truth of that statement for many unstable proteins (proteins containing amino acid analogs, proteins mistargeted to cells or organelles, various fusion proteins, thermally unfolded proteins) is self-evident. This generalization may apply equally to rapidly degraded normal proteins, which include many important cellular proteins whose degradation is an integral part of regulating their activity. All of the proteins known to be rapidly degraded in vivo function as parts of high molecular weight protein or proteinnucleic acid complexes. In a number of cases, the proteins are stabilized by interaction with their functional partners and degraded when they are free, or, when the protein is degraded from a complex, it is usually capable of forming several different complexes and only one such complex is targeted. Unsatisfied bonding domains in abnormal proteins and in the free subunits or in exposed regions of certain hetero-oligomeric complexes of normal proteins are the target sites for ATP-dependent proteases. Such bonding sites generally have an amphipathic character, but otherwise do not have a defined structural or chemical composition for all proteins. 6.3. Do Chaperones and Proteases Recognize Similar Features of Proteins? There is much in common between the unstructured proteins recognized by chaperones and proteases. Lon protease, GroEL, and DnaK can bind simultaneously to the same immobilized abnormal protein (Sherman & Goldberg, 1992), and immunoprecipitation has been used to show a soluble complex containing an unstable abnormal protein, DnaK, and Lon protease (Sherman & Goldberg, 1991). Since the chaperones could be released from the complex with ATP but not non-hydrolyzable analogs, the binding appears to reflect normal interactions with unfolded proteins (Sherman & Goldberg, 1991). Whether chaperones and proteases bind to exactly the same sites or whether they recognize different exposed regions in abnormally folded proteins is not known. Studies of chaperones binding to peptides of 6–13 amino acids has revealed some
Chaperones and charonins
437
positional preferences but no real consensus for high affinity binding. For E. coli DnaK and other Hsp70s, peptides with a core of 4–5 hydrophobic residues, with at least one (often central) leucine and basic residues at the ends, tend to bind best (Gierasch, 1994; Rüdiger et al., 1997b) (see chapter Buchberger et al., this volume). Crystal structure data on DnaK with a bound peptide indicate that the peptide backbone as well as side chains of the bound substrate make contact with the binding pocket (Zhu et al., 1996). Sequence comparisons indicate that the overall character of the binding pocket of DnaK is conserved among other Hsp70 proteins, particularly in the central hydrophobic binding position. However, as pointed out in a recent review (Rüdiger et al., 1997a), there is some variation in the disposition of specific residues that make up the subsites in the binding cavities and there are differences in the surface charge distribution at the sides of the binding pockets, which may account for differences in specificity of peptide binding between these proteins. Further thermodynamic and structural studies should bring the issue of substrate recognition by Hsp70 proteins into clearer focus. 7. CONCLUDING REMARKS The critical elements in intracellular proteolysis are 1) the ability to recognize abnormal structural features of proteins and to bind tightly to a region of the protein, probably exposed amphipathic stretches, and 2) the ability to unfold the bound protein to an extended conformation that can penetrate the access channel to the proteolytic site. The first element is common to both chaperones and ATP-dependent proteases, although there is a range of specificities, and consequently affinities, between different chaperones and proteases for particular structural or chemical features exposed in individual proteins. The unfolding activity is also common to both chaperones and proteases, but there is an additional property of the proteases that with a high frequency drastically alters the outcome of the unfolding reaction. Proteases have a means of holding on to proteins and directing them into the proteolytic active sites. Once the degradation pathway is selected, no significant amounts of partially cleaved protein are released and the protein is degraded into peptides of 5–15 amino acids. We can propose several models to explain how proteases might trap proteins for degradation. (1) A “flypaper” model. In this model, the protease has multiple interaction sites for proteins and peptides. A protein that interacts with the protease will bind at one or more of those sites depending on the extent to which it is unfolded and, with time, will interact with additional sites unless it can rapidly assume a stable folded structure. If the interaction sites line a channel leading to the proteolytic sites, the bound protein will move by a simple facilitated diffusion mechanism, similar to that proposed for transport of proteins through membrane channels. ATP hydrolysis by the protease could cause the “sticky” side to alternate with a “non-stick” side allowing the protein to migrate from one site to another. Such a change in binding character in response to ATP hydrolysis has been shown to occur in the protein binding cavity of GroEL (Roseman et al., 1996). (2) A “trapdoor” model. In a trapdoor model, a channel in the protease is lined with
Molecular chaperones and folding catalysts
438
flexible structures that can bend inward to allow passage of a portion of the protein. In response to nucleotide hydrolysis, these structures could close around narrow regions of the protein and, because they cannot be bent backwards, not allow the protein to withdraw. One major difference between this and the previous model is that it does not require interaction sites to line the access channel into the proteolytic sites. (3) A “clamping” model. The crystal structure of DnaK suggests that it can fold around the polypeptide chain of a substrate, essentially clamping the protein to the chaperone (Zhu et al., 1996) (see chapter Buchberger et al., this volume). The clamp might be formed in one of a few stable conformations of a chaperone/protease, and transitions from one state to another could be facilitated during nucleotide binding, hydrolysis, and release. A clamping mechanism might apply to proteases such as Lon, which undergoes substantial conformational change upon protein binding and for which protein binding results in a large activation of ATP hydrolysis. (4) A cork-screw mechanism. A substrate anchored to a point on the ATPase could be moved through the access channel in a kind of cork-screw motion by rotation or ratcheting of the ATPase subunits with respect to the protease subunits. The symmetry mismatch between six-membered rings of ClpA and seven-membered rings of ClpP allows only one pair of subunits to be in direct register at any given time, and conformational changes upon ATP hydrolysis could lead to different pairs of subunits coming into alignment with each catalytic cycle (Kessel et al., 1996). The protein binding properties of the channel would also have to change during the cycles to allow the protein to move through. A possible method of anchoring protein substrates to the protease would be by covalent attachment to a site on the ATPase. No evidence for such an attachment has been found, but it has not been ruled out experimentally. Whichever model applies, the attached protein must be unwound and translocated into the proteolytic active sites. The multiple active sites could serve as additional points of attachment for substrates, and interactions there could serve to reel the extended polypeptide chain into the proteolytic cavity. If ATP-dependent proteases have their own built-in chaperones, how is it that general chaperones also affect degradation? There is no evidence for direct interaction between general chaperones and ATP-dependent proteases, and we think it is likely that chaperones facilitate proteolysis primarily by binding misfolded proteins and preventing their aggregation. General chaperones, such as DnaK and GroEL, are present in much higher concentrations than the proteases and most likely encounter misfolded proteins more frequently than do the proteases. However, the available data do not differentiate between direct (the pro tease interacts with the substrate bound to the chaperone) or indirect (the pro tease interacts with the substrate after release from the chaperone) mechanisms for the subsequent steps. Chaperones and proteases probably recognize similar sites, and a direct mechanism would require at least two interaction sites on the substrate. For substrates that are released slowly from chaperones, proteases may interact with other exposed sites on the chaperone-bound protein and initiate degradation. For proteins that aggregate rapidly, maintaining the misfolded protein at a minimal concentration so it remains free and accessible to proteases may be the kinetically most significant factor. It may be that both mechanisms are at work with different substrates and the difference depends on the kinetics of release from the chaperone and on the
Chaperones and charonins
439
number of binding sites in the substrate. 8. REFERENCES Aizenman, E., Engelberg-Kulka, H. and Glaser, G. (1996). An Escherichia coli chromosomal “addiction module” regulated by ppGpp: a model for programmed bacterial cell death. Proc. Natl. Acad. Sci. USA , 93 , 6059–6063. Arlt, H., Tauer, R., Feldmann, H., Neupert, W. and Langer, T. (1996). The YTA10–12 complex, an AAA protease with chaperone-like activity in the inner membrane of mitochondria. Cell , 85 , 875–885. Baker, T.A., Grossman, A.D. and Gross, C.A. (1984). A gene regulating the heat shock response in Escherichia coli also affects proteolysis. Proc. Natl. Acad. Sci. USA , 81 , 6779–6783. Banuett, F., Hoyt, M.A., McFarlane, L., Echols, H. and Herskowitz, I. (1986). hflB, a new Escherichia coli locus regulating lysogeny and the level of bacteriophage lambda CII protein. J. Molec. Biol. , 187 , 213–224. Beggah, A., Mathews, P., Beguin, P. and Geering, K. (1996). Degradation and endoplasmic reticulum retention of unassembled alpha- and beta-subunits of Na, KATPase correlate with interaction of BiP. J. Biol. Chem., 271 , 20895–20902. Bukhari, A.I. and Zipser, D. (1973). Mutants of Escherichia coli with a defect in the degradation of nonsense fragments. Nature , 243 , 238–241. Chin, D.T., Goff, S.A., Webster, T., Smith, T. and Goldberg, A.L. (1988). Sequence of the lon gene in Escherichia coli: A heat-shock gene which encodes the ATP-dependent protease La. J. Biol. Chem. , 263 , 11718–11728. Chuang, S.-E., Burland, V., III, G.P., Daniels, D.L. and Blattner, F.R. (1993). Sequence analysis of four new heat shock genes constituting the hslTS/ibpAB and hslVU operons in Escherichia coli. Gene , 134 , 1–6. Damerau, K. and St. John, A.C. (1993). Role of Clp protease subunits in degradation of carbon starvation proteins in Escherichia coli. J. Bacterial. , 175 , 53–63. Desautels, M. and Goldberg, A.L. (1982). Liver mitochondria contain an ATP-dependent, vanadate-sensitive pathway for the degradation of proteins. Proc. Natl. Acad. Set. USA , 79 , 1869–1873. Fischer, H. and Glockshuber, R. (1993). ATP hydrolysis is not stoichiometrically linked with proteolysis in the ATP-dependent protease La from Escherichia coli. J. Biol. Chem. , 268 , 22502–22507. Fischer, H. and Glockshuber, R. (1994). A point mutation within the ATP-binding site inactivates both catalytic functions of the ATP-dependent protease La (Lon) from Escherichia coli. Febs Lett . 356 , 101–103. Gierasch, L.M. (1994). Molecular chaperones. Panning for chaperone-binding peptides. Curr. Biol. , 4 , 173–174. Gill, R.E., Karlok, M. and Benton, D. (1993). Myxococcus xanthus encodes an ATPdependent protease which is required for developmental gene transcription and intercellular signaling. J. Bacteriol. , 175 , 4538–4544. Glotzer, M., Murray, A.W. and Kirschner, M.W. (1991). Cyclin is degraded by the ubiquitin pathway. Nature , 349 , 132–138.
Molecular chaperones and folding catalysts
440
Goff, S.A., Casson, L.P. and Goldberg, A.L. (1984). Heat shock regulatory gene htpR influences rates of protein degradation and expression of the lon gene in Escherichia coli. Proc. Natl. Acad. Sci. USA , 81 , 6647–6651. Goldberg, A.L. (1992). The mechanism and functions of ATP-dependent proteases in bacterial and animal cells. Eur. J. Biochem. , 203 , 9–23. Goldberg, A.L., Moerschell, R.P., Chung, C.H. and Maurizi, M.R. (1994). ATPdependent protease La (Lon) from Escherichia coli. Methods Enzymol. , 244 , 350– 375. Goldberg, A.L. and Waxman, L. (1985). The role of ATP hydrolysis in the breakdown of proteins and peptides by protease La from Escherichia coli. J. Biol. Chem. , 260 , 12029–12034. Gottesman, S., Clark, W.P., Crecy-Lagard, V.D. and Maurizi, M.R. (1993). ClpX, an alternative subunit for the ATP-dependent Clp protease of Escherichia coli: sequence and in vivo activities. J. Biol. Chem. , 268 , 22618–22626. Gottesman, S., Clark, W.P. and Maurizi, M.R. (1990a). The ATP-dependent Clp protease of Escherichia coli: sequence of clpA and identification of a Clp-specific substrate. J. Biol. Chem. , 265 , 7886–7893. Gottesman, S. and Maurizi, M.R. (1992). Regulation by Proteolysis: Energy-Dependent Proteases and Their Targets. Microbiol. Rev. , 56 , 592–621. Gottesman, S., Roche, E., Zhou, Y.-N. and Sauer, R.T. (1998). The ClpXP and ClpAP proteases degrade proteins with C-terminal peptide tails added by the SsrA tagging system. Genes andDev. , 12, (in press). Gottesman, S., Squires, C., Pichersky, E., Carrington, M., Hobbs, M., Mattick, J.S., Dalrymple, B., Kuramitsu, H., Shiroza, T., Foster, T., Clark, W.P., Ross, B., Squires, C.L. and Maurizi, M.R. (1990b). Conservation of the regulatory subunit for the Clp ATP-dependent protease in prokaryotes and eukaryotes. Proc. Natl. Acad. Sci. USA , 87 , 3513–3517. Gottesman, S., Wickner, S.Jubete, Y., Singh, S.K., Kessel, M. and Maurizi, M.R. (1995). Selective energy-dependent proteolysis in Escherichia coli. Cold Spring Harbor Sympos. Quant. Biol. , Volume LX, 533–548. Grimaud, R., Kessel, M, Beuron, F., Steven, A.C. and Maurizi, M.R. (1998). Enzymatic and structural similarities between the E. coli ATP-dependent proteases, ClpXP and ClpAP. J. Biol. Chem. , 273 , (in press). Herman, C., Thevenet, D., D’Ari, R. and Bouloc, P. (1995). Degradation of sigma 32, the heat shock regulator in Escherichia coli, is governed by HflB. Proc. Natl. Acad. Sci. USA , 92 , 3516–3520. Herman, C., Thevenet, D., D’Ari, R. and Bouloc, P. (1997). The HflB protease of Escherichia coli; degrades its inhibitor lambda cIII. J. Bacteriol. , 179 , 358–363. Hershko, A. and Ciechanover, A. (1992). The Ubiquitin system for protein degradation. Annu. Rev. Biochem. , 61 , 761–807. Huang, H.-C. and Goldberg, A.L. (1997). Proteolytic activity of the ATP-dependent protease HslUV can be uncoupled from ATP hydrolysis. J. Biol. Chem. , 272 , 21364– 21372. Hwang, B.J., Woo, K.M., Goldberg, A.L. and Chung, C.H. (1988). Protease Ti, a new ATP-dependent protease in Escherichia coli contains protein-activated ATPase and
Chaperones and charonins
441
proteolytic functions in distinct subunits. J. Biol Chem. , 263 , 8727–8734. Inoue, I. and Rechsteiner, M. (1994). On the relationship between the metabolic and thermodynamic stabilities of T4 lysozymes; measurements in Escherichia coli. J. Biol. Chem. , 269 , 29241–29246. Jensen, R.B. and Gerdes, K. (1995). Programmed cell death in bacteria: proteic plasmid stabilization systems. Mol Microbiol. , 17 , 205–210. Jubete, Y., Maurizi, M.R. and Gottesman, S. (1996). Role of the heat shock protein, DnaJ, in the Lon-dependent degradation of naturally unstable proteins. J. Biol. Chem. , 271 , 30798–30803. Kandror, O., Busconi, L., Sherman, M. and Goldberg, A.L. (1994). Rapid degradation of an abnormal protein in Escherichia coli involves the chaperones GroEL and GroES. J. Biol. Chem. , 269 , 23575–23582. Kandror, O., Sherman, M., Rhode, M. and Goldberg, A.L. (1995). Trigger factor is involved in GroEL-dependent protein degradation in Escherichia coli and promotes binding of GroEL to unfolded proteins. EMBO J. , 14 , 6021–6027. Katayama-Fujimura, Y., Gottesman, S. and Maurizi, M.R. (1987). A multiple-component ATP-dependent protease from Escherichia coli. J. Biol. Chem. , 262 , 4477–4485. Keiler, K.C., Waller, P.R. and Sauer, R.T. (1996). Role of a peptide tagging system in degradation of proteins synthesized from damaged messenger RNA. Science , 271 , 990–993. Keller, J.A. and Simon, L.D. (1988). Divergent effects of a dnaK mutation on abnormal protein degradation in Escherichia coli. Mol. Microbiol. , 2 , 31–41. Kessel, M., Maurizi, M.R., Kim, B., Trus, B.L., Kocsis, E., Singh, S.K. and Steven, A.C. (1995). Homology in structural organization between E. coli ClpAP protease and the eukaryotic 86 S proteasome. J. Mol. Biol. , 250 , 587–594. Kessel, M., Wu, W.-F., Gottesman, S., Kocsis, E., Steven, A. and Maurizi, M.R. (1996). Six-fold rotational symmetry of ClpQ, the E. coli homolog of the 20 S proteasome, and its ATP-dependent activator, ClpY. EEBS Letters , 398 , 274–278. Kihara, A., Akiyama, Y. and Ito, K. (1995). FtsH is required for proteolytic elimination of uncomplexed forms of SecY, an essential protein translocase subunit. Proc Natl Acad Sci USA , 92 , 4532–4536. Kruklitis, R., Welty, D.J. and Nakai, H. (1996). ClpX protein of Escherichia coli activates bacteriophage Mu transposase in the strand transfer complex for initiation of Mu DNA synthesis. EMBO J. , 15 , 935–944. Lee, D.H., Sherman, M.Y. and Goldberg, A.L. (1996). Involvement of the molecular chaperone Ydj1 in the ubiquitin-dependent degradation of short-lived and abnormal proteins in Saccharomyces cerevisiae. Mol. Cell Biol. , 16 , 4773–4781. Lehnherr, H. and Yarmolinsky, M.B. (1995). Addiction protein Phd of plasmid prophage P1 is a substrate of the ClpXP serine protease of Escherichia coli. Proc. Natl. Acad. Sci. USA , 92 , 3274–3277. Levchenko, I., Luo, L. and Baker, T.A. (1995). Disassembly of the Mu transposase tetramer by the ClpX chaperone. Genes and Dev. , 9 , 2399–2408. Levchenko, I., Smith, C.K., Walsh, N.P., Sauer, R.T. and Baker, T.A. (1997). PDZ-like domains mediate specificity in the Clp/Hsp100 family of chaperones and protease regulatory subunits. Cell , 91 , 939–947.
Molecular chaperones and folding catalysts
442
Lowe, J., Stock, D., Jap, B., Zwickl, P., Baumeister, W. and Huber, R. (1995). Crystal structure of the 20 S proteasome from the Archaeon T. acidophilum at 3.4 A resolution. Science , 268 , 533–539. Maurizi, M.R. (1987). Degradation in vitro of bacteriophage lambda N protein by Lon protease from Escherichia coli. J. Biol. Chem. , 262 , 2696–2703. Maurizi, M.R. (1992). Proteases and protein degradation in Escherichia coli. Experientia , 48 , 178–201. Maurizi, M.R., Clark, W.P., Katayama, Y., Rudikoff, S., Pumphrey, J., Bowers, B. and Gottesman, S. (1990a). Sequence and structure of Clp P, the proteolytic component of the ATP-dependent Clp protease of Escherichia coli. J. Biol. Chem. , 265 , 12536– 12545. Maurizi, M.R., Clark, W.P., Kim, S.-H. and Gottesman, S. (1990b). ClpP represents a unique family of serine proteases. J. Biol. Chem. , 265 , 12546–12552. Maurizi, M.R., Thompson, M.W., Singh, S.K. and Kim, S.H. (1994). Endopeptidase Clp: ATP-dependent Clp protease from Escherichia coli. Methods Enzymol , 244 , 314–331. Maurizi, M.R., Trisler, P. and Gottesman, S. (1985). Insertional mutagenesis of the lon gene in Escherichia coli: lon is dispensable. J. Bacteriol. , 164 , 1124–1135. Menon, A.S., Waxman, L. and Goldberg, A.L. (1987). The energy utilized in protein breakdown by the ATP-dependent protease La from Escherichia coli. J. Biol. Chem. , 262 , 722–726. Mhammedi-Alaoui, A., Pato, M., Gama, M.J. and Toussaint, A. (1994). A new component of bacteriophage Mu replicative transposition machinery: the Escherichia coli ClpX protein. Mol. Microbiol. , 11 , 1109–1116. Missiakas, D., Schwager, F., Georgopoulos, C. and Raina, S. (1996). Isolation and characterization of HslV HslU (ClpQ ClpY) proteins involved in overall proteolysis of misfolded proteins in E. coli. EMBO J. , 15 , 6899–6909. Ohki, R., Kawamata, T., Katoh, Y., Hosoda, F. and Ohki, M. (1992). Escherichia coli dnaJ/deletion mutation results in loss of stability of a positive regulator, CRP. J. Bacteriol. , 267 , 13180–13184. Park, S.K., Kim, K.I., Woo, K.M., Seol, J.H., Tanaka, K., Ichihara, A., Ha, D.B. and Chung, C.H. (1993). Site-directed mutagenesis of the dual translational initiation sites of the clpB gene of Escherichia coli and characterization of its gene products. J. Biol Chem. , 268 , 20170–20174. Parsell, D.A., Kowal, A.S. and Lindquist, S. (1994a). Saccharomyces cerevisiae Hsp104 protein: purification and characterization of ATP-induced structural changes. J. Biol. Chem. , 269 , 4480–4487. Parsell, D.A., Kowal, A.S., Singer, M.A. and Lindquist, S. (1994b). Protein disaggregation mediated by heat-shock protein Hsp104. Nature , 372 , 475–478. Parsell, D.A., Sanchez, Y., Stitzel, J.D. and Lindquist, S. (1991). Hsp104 is a highly conserved protein with two essential nucleotide-binding sites. Nature , 353 , 270–273. Parsell, D.A., Silber, K.R. and Sauer, R.T. (1990). Carboxy-terminal determinants of intracellular protein degradation. Genes and Dev. , 4 , 277–286. Paushkin, S.V., Kushnirov, V.V., Smirnov, V.N. and Ter-Avanesyan, M.D. (1996). Propagation of the yeast prion-like [psi+] determinant is mediated by oligomerization of the SUP35-encoded polypeptide chain release factor. EMBO J. , 15 , 3127–3134.
Chaperones and charonins
443
Rep, M., van Dijl, J.M., Suda, K., Schatz, G., Grivell, L.A. and Suzuki, C.K. (1996). Promotion of mitochondrial membrane complex assembly by a proteolytically inactive yeast Lon. Science , 274 , 103–106. Rogers, S., Wells, R. and Rechsteiner, M. (1986). Amino acid sequences common to rapidly degraded proteins: the PEST hypothesis. Science , 234 , 364–368. Rohrwild, M., Coux, O., Huang, H.C., Moerschell, R.P., Yoo, S.J., Seol, J.H., Chung, C.H. and Goldberg, A.L. (1996). HslV-HslU: a novel ATP-dependent protease complex in Escherichia coli related to the eukaryotic proteasome. Proc. Natl Acad. Sci. USA , 93 , 5808–5813. Rohrwild, M., Pfeifer, G., Santarius, U., Muller, S.A., Huang, H.C., Engel, A., Baumeister, W. and Goldberg, A.L. (1997). The ATP-dependent HslVU protease from Escherichia coli is a four-ring structure resembling the proteasome. Nat. Struct. Biol. , 4 , 133–139. Roseman, A.M., Chen, S., White, H., Braig, K. and Saibil, H.R. (1996). The chaperonin ATPase cycle: mechanism of allosteric switching and movements of substrate-binding domains in GroEL. Cell , 18 , 241–251. Rüdiger, S., Buchberger, A. and Bukau, B. (1997a). Interaction of Hsp70 chaperones with substrates. Nat. Struct. Biol. , 4 , 342–349. Rüdiger, S., Germeroth, L., Schneider-Mergener, J. and Bukau, B. (1997b). Substrate specificity of the DnaK chaperone determined by screening cellulose-bound peptide libraries. EMBO J. , 16 , 1501–1507. Sanchez, Y., Taulien, J., Borkovich, K.A. and Lindquist, S. (1992). Hsp104 is required for tolerance to many forms of stress. EMBO J. , 11 , 2357–2364. Schirmer, E.G., Glover, J.R., Singer, M.A. and Lindquist, S. (1996). HSP100/Clp proteins: a common mechanism explains diverse functions . Trends Biochem. Sci. , 21 , 289–296. Schmitt, M., Neupert, W. and Langer, T. (1995). Hsp78, a Clp homologue within mitochondria, can substitute for chaperone functions of mt-hsp70. EMBO J. , 14 , 3434–3444. Schweder, T., Lee, K.H., Lomovskaya, O. and Matin, A. (1996). Regulation of Escherichia coli starvation sigma factor ( ) by ClpXP protease. J. Bacteriol., 178 , 470–476. Seemuller, E., Lupas, A., Stock, D., Lowe, J., Huber, R. and Baumeister, W. (1995). Proteasome from Thermoplasma acidophilum: a threonine protease. Science , 268 , 579–582. Sherman, M.Y. and Goldberg, A.L. (1991). Formation in vitro of complexes between an abnormal fusion protein and the heat shock proteins from Escherichia coli and yeast mitochondria. J. Bacteriol. , 173 , 7249–7256. Sherman, M.Y. and Goldberg, A.L. (1992). Involvement of the chaperonin dnaK in the rapid degradation of a mutant protein in Escherichia coli. EMBO J. , 11 , 71–77. Shirai, Y., Akiyama, Y. and Ito, K. (1996). Suppression of ftsH mutant phenotypes by overproduction of molecular chaperones. J. Bacteriol. , 178 , 1141–1145. Singh, S.K. and Maurizi, M.R. (1994). Mutational analysis demonstrates different functional roles for the two ATP-binding sites in ClpAP protease from Escherichia coli. J. Biol. Chem. , 269 , 29537–29545.
Molecular chaperones and folding catalysts
444
Squires, C. and Squires, C.L. (1992). The Clp proteins: Proteolysis regulators or molecular chaperones? J. Bacteriol. , 174 , 1081–1085. Squires, C.L., Pedersen, S., Ross, B.M. and Squires, C. (1991). ClpB is the Escherichia coli heat shock protein F84.1. J. Bacteriol. , 173 , 4254–4262. Straus, D.B., Walter, W.A. and Gross, C. (1988). Escherichia coli heat shock gene mutants are defective in proteolysis. Genes and Dev. , 2, 1851–1858. Suzuki, C.K., Suda, K., Wang, N. and Schatz, G. (1994). Requirement for the yeast gene LON in intramitochondrial proteolysis and maintenance of respiration [published erratum appears in Science (1994) 264 , 891]. Science , 264 , 273–276. Thompson, M.W. and Maurizi, M.R. (1994). Activity and specificity of Escherichia coli ClpAP protease in cleaving model peptide substrates. J. Biol. Chem. , 269 , 18201– 18208. Thompson, M.W., Singh, S.K. and Maurizi, M.R. (1994). Processive degradation of proteins by the ATP-dependent Clp protease from Escherichia coli: requirement for the multiple array of active sites in ClpP but not ATP hydrolysis. J. Biol. Chem. , 169 , 18209–18215. Tobias, J.W., Shrader, T.E., Rocap, G. and Varshavsky, A. (1991). The N-End rule in bacteria. Science , 254 , 1374–1376. Tojo, N., Inouye, S. and Komano, T. (1993a). Cloning and nucleotide sequence of the Myxococcus xanthus lon gene: indispensability of lon for vegetative growth. J. Bacteriol. , 175 , 2271–2277. Tojo, N., Inouye, S. and Komano, T. (1993b). The lonD gene is homologous to the lon gene encoding an ATP-dependent protease and is essential for the development of Myxococcus xanthus. J. Bacteriol. , 175 , 4545–4549. Tomoyasu, T., Gamer, J., Bukau, B., Kanemori, M., Mori, H., Rutman, A.J., Oppenheim, A.B., Yura, T., Yamanaka, K., Niki, H. and Ogura, T (1995). Escherichia coli FtsH is a membrane-bound, ATP-dependent protease which degrades the heat-shock transcription factor . EMBO J. , 14 , 2551–2560. Van Dyck, L., Pearce, D.A. and Sherman, F (1994). PIM1 encodes a mitochondrial ATPdependent protease that is required for mitochondrial function in the yeast Saccharomyces cerevisiae. J. Biol. Chem. , 269 , 238–242. Van Melderen, L., Bernard, P. and Couturier, M. (1994). Lon-dependent proteolysis of CcdA is the key control for activation of CcdB in plasmid-free segregant bacteria. Mol. Microbiol . 11 , 1151–1157. Van Melderen, L., Dao Thi, M.H., Lecchi, P., Gottesman, S., Couturier, M. and Maurizi, M.R. (1996). ATP-dependent degradation of CcdA by lon protease: effects of secondary structure and heterologous subunit interactions. J. Biol. Chem. , 271 , 27730–27738. Varshavsky, A. (1992). The N-end Rule. Cell , 69 , 725–735. Vogel, J.L., Parsell, D.A. and Lindquist, S. (1995). Heat-shock proteins Hsp104 and Hsp70 reactivate mRNA slicing after heat inactivation. Current Biology , 5 , 306–317. Wagner, I., Arlt, H., van Dyck, L., Langer, T. and Neupert, W. (1994). Molecular chaperones cooperate with PIM1 protease in the degradation of misfolded proteins in mitochondria. EMBO J. , 13 , 5135–5145. Wang, N., Gottesman, S., Willingham, M.C., Gottesman, M.M. and Maurizi, M.R.
Chaperones and charonins
445
(1993). A human mitochondrial ATP-dependent protease that is highly homologous to bacterial Lon protease. Proc. Natl. Acad. Sci. USA , 90 , 11247–11251. Watabe, S. and Kimura, T. (1985). Adrenal cortex mitochondrial enzyme with ATPdependent protease and protein-dependent ATPase activities. J. Biol. Chem. , 260 , 14498–14504. Wawrzynow, A., Wojtkowiak, D., Marszalek, J., Banecki, B., Jonsen, M., Graves, B., Georgopoulos, C. and Zylicz, M. (1995). The ClpX heat-shock protein of Escherichia coli, the ATP-dependent substrate specificity component of the ClpP-ClpX protease, is a novel molecular chaperone. EMBO J. 14 , 1867–1877. Waxman, L. and Goldberg, A.L. (1982). Protease La from Escherichia coli hydrolyzes ATP and proteins in a linked fashion. Proc. Natl. Acad. Sci. USA , 79 , 4883–4887. Waxman, L. and Goldberg, A.L. (1986). Selectivity of intracellular proteolysis: protein substrates activate the ATP-dependent protease (La). Science , 232 , 500–503. Wickner, S., Gottesman, S., Skowyra, D., Hoskins, J., McKenney, K. and Maurizi, M.R. (1994). A molecular chaperone, ClpA, functions like DnaK and DnaJ. Proc. Natl. Acad. Sci. USA , 91 , 12218–12222. Wickner, S., Hoskins, J. and McKenney, K. (1991). Monomerization of RepA dimers by heat shock proteins activates binding to DNA replication origin. Proc. Natl. Acad. Sci. USA , 88 , 7903–7907. Wojtkowiak, D., Georgopoulos, C. and Zylicz, M. (1993). Isolation and characterization of ClpX, a new ATP-dependent specificity component of the Clp protease of Escherichia coli. J. Biol. Chem. , 268 , 22609–22617. Woo, K.M., Chung, W.J., Ha, D.B., Goldberg, A.L. and Chung, C.H. (1989). Protease Ti from Escherichia coli requires ATP hydrolysis for protein breakdown but not for hydrolysis of small peptides. J. Biol. Chem. , 264 , 2088–2091. Woo, K.M., Kim, K.I., Goldberg, A.L., Hai, D.B. and Chung, C.H. (1992). The heatshock protein ClpB in Escherichia coli is a protein-activated ATPase. J. Biol. Chem. , 267 , 20429–20434. Wright, R., Stephens, C., Zweiger, G., Shapiro, L. and Alley, M.R. (1996). Caulobacter Lon protease has a critical role in cell-cycle control of DNA methylation. Genes and Dev. , 10 , 1532–1542. Yaglom, J.A., Goldberg, A.L., Finley, D. and Sherman, M.Y. (1996). The molecular chaperone Ydj1 is required for the p34/CDC28-dependent phosphorylation of the cyclin Cln3 that signals its degradation. Mol. Cell Biol. , 16 , 3679–3684. Yoo, S.J., Seol, J.H., Shin, D.H., Rohrwild, M., Kang, M.-S., Tanaka, K., Goldberg, A.L. and Chung, C.H. (1996). Purification and characterization of the heat shock proteins HslV and HslU that form a new ATP-dependent protease in Escherichia coli. J. Biol. Chem. , 271 , 14035–14040. Yu, A., Antonov, V.K., Gorbalenya, A.E., Kotova, S.A., Rotanova, T.V. and Shimbarevich, E.V. (1991). Site-directed mutagenesis of La protease. A catalytically active serine residue. FEBS Lett. , 287 , 211–214. Zhu, X., Zhao, X., Burkholder, W.F., Gragerov, A., Ogata, C.M., Gottesman, M.E. and Hendrickson, W.A. (1996). Structural analysis of substrate binding by the molecular chaperone DnaK. Science , 272 , 1606–1614.
IV. MECHANISMS
19. SPONTANEOUS VERSUS ASSISTED PROTEIN FOLDING RAINER JAENICKE * and ROBERT SECKLER Institut für Biophysik und Physikalische Biochemie, Universität Regensburg, D93040 Regensburg, Germany
1. Protein Structure and Protein Self-organization 1.1. The Protein Folding Problem 1.2. Native, Intermediate and Denatured States 1.3. Denaturation-Renaturation 2. Folding of Small Single-Domain Proteins 2.1. Rate-Determining Late Folding Events 2.2. Early Intermediates 2.3. Multiple Pathways 2.4. Protein Engineering Analysis of Protein Folding 2.5. Chymotrypsin Inhibitor 2 (CI-2) 3. Folding of Domain Proteins 4. Association 4.1. Sequential Folding and Association 4.2. Specificity of Association 4.3. Superstructures 5. Off-pathway Reactions 5.1. Thermodynamics Versus Kinetics 5.2. Productive Versus Non-productive Pathways 6. The in vitro vs. in vivo Issue 6.1. Cytosolic Solvent Parameters 6.2. Folding Catalysts 6.3. Chaperones 7. Practical Aspects 8. Recent Developments and Conclusions 9. References *Corresponding author
Spontaneous versus assisted protein folding
449
1. PROTEIN STRUCTURE AND PROTEIN SELF-ORGANIZATION It has almost become a truism for more than a generation of biochemists that the acquisition of the native functional state of proteins is a spontaneous and autonomous process depending solely on the amino acid sequence and its (quasi-) physiological environment. As has been demonstrated by Anson and his colleagues, who pioneered the field long before the three-dimensional structure of a single protein was known, the natural cellular environment can be mimicked by wellchosen buffers. Based on this discovery, the time course of protein folding and association has been investigated in great detail leading to plausible models describing sequential and multiple pathways of protein self-organization. 1.1. The Protein Folding Problem Because of the large size of the molecules and the unresolved problem how to summarize the weak attractive and repulsive interatomic forces in a unique potential function, theorists have been unsuccessful in solving the protein folding code which would allow to translate the one-dimensional amino acid sequence of a given protein into the corresponding unique three-dimensional structure of the polypeptide chain. Thus, the only way to “predict” protein structures still is homology modeling based on sequence alignments and known spatial structures of related proteins. Obviously, this computeraided, knowledge-based approach is postdictive rather than predictive, and the result commonly does not provide us with the high resolution required in order to work out mechanistic details, e.g., of enzyme function or stability. To reach this level, X-ray diffraction of protein crystals and/or NMR analysis in solution are indispensable. The question whether or not the physical state of a protein obtained from crystal data is compatible with its functional state in vivo has been extensively discussed since Kendrew, Perutz and Phillips came up with their first final analyses of the structures of myoglobin, hemoglobin and lysozyme. Hardly any scientific result has ever been challenged to such an extent. The final (fundamentally positive) answer came from comparative studies using model proteins crystallized in different space groups, on one hand, and NMR spectra, on the other. The accumulated evidence allows us to conclude that, under optimum conditions, both methods yield threedimensional structures at atomic resolution that complement each other to the extent that local interactions of functional groups, including single molecules of the surrounding solvent, can be defined. Regarding the formation of the native functional state of protein molecules, the situation is more complicated, because of fundamental differences between the situation in the cell, on the one hand, and optimum conditions of in vitro reconstitution experiments, on the other. One obvious discrepancy comparing in vivo and in vitro folding is that the nascent polypeptide chain may fold cotranslationally, i.e., in a vectorial fashion from its N- to its C-terminal end, whereas folding in vitro commonly starts from the integral protein in its denatured state. This differs from the native state in two ways: first, it represents an astronomically large ensemble of different configurations, and
Molecular chaperones and folding catalysts
450
second, it is solvated to a higher extent than the native state. Considering the amount of hydrophobic residues present in most soluble proteins, water is a poor solvent. Structure formation in the cytosol or in aqueous buffer systems is driven by this very fact. Making use of heteropolymers with approximately equal amounts of polar and unpolar residues, nature allows solvation to balance by exposing hydrophilic groups to the aqueous solvent, at the same time minimizing the hydrophobic surface (Richards, 1992). Solubilization of the inner core by mixed solvents or altered temperature leads to unfolding (Privalov, 1992). However, it is obvious from the balance of hydrophilic and hydrophobic amino acids in common proteins that complete solvation cannot be accomplished. For this reason it is doubtful whether a polypeptide chain will ever be “fully randomized” (Jaenicke, 1987). Even in the process of translation, the nascent protein is limited in its conformational space because the space-filling properties of the growing polypeptide and its side chains do not allow all / angles in the Ramachandran plot. 1.2. Native, Intermediate and Denatured States Because of the high internal flexibility of the molecules, the detailed structural characterization of unfolded or partially unfolded proteins is difficult. However, X-ray small angle scattering and multidimensional heteronuclear NMR revealed that even in extensively unfolded states, e.g., at high concentrations of chaotropic agents, considerable residual structure remains. This result confirms previous assumptions that local hydrophobic clusters with features of the native state are preserved, at least in equilibrium with less-structured states. There is no evidence for significant long-range structure in the denatured polypeptide chain. In contrast to the multitude of unfolded states, the native state of a given protein is commonly assumed to be well defined within the limits of the B-factors of X-ray analysis or the conformational dynamics calculated from multidimensional NMR. It is clear that there must be a certain range of flexibility for the obvious reason that in many cases proteins serve as multifunctional devices which have evolved, e.g., to bind, transform and release educts and products of metabolic reactions (Huber, 1988). The driving forces guiding the folding polypeptide chain to its final structure are the same next-neighbor and through-space short-range interactions that are responsible for protein stability: H-bonds, hydrophobic and Coulomb interactions, and van der Waals forces (Dill, 1990; Jaenicke, 1991). From the point of view of thermo-dynamics, the minimum of free energy of the native conformation corresponds to the state with maximum packing density connected with minimum hydrophobic surface area and cavity space (Richards, 1992; Matthews, 1995); the entropic driving force in both hydrophobic collapse and subunit assembly comes from water release involving nonpolar residues and ion pairs (Lauffer, 1975; Jaenicke, 1987). Considering the kinetics of the folding process, it is evident that the acquisition of the native state cannot be a stochastic process; rather, folding must proceed along a pathway or a limited number of alternative pathways (Levinthal, 1968; Baldwin & Eisenberg, 1987). Successive kinetic intermediates along the U N transition have been clearly established, indicating that folding is a hierarchical condensation reaction (Figure 1): next-neighbor interactions first form fluctuating native and non-native secondary structural elements; during this step (which occurs in the sub-milliseconds time range)
Spontaneous versus assisted protein folding
451
kinetic nuclei gain increasing stability. As soon as the intermolecular interactions surpass the thermal energy, the polypeptide chain collapses into a persistent native-like secondary structure which still lacks the close
Figure 1 Hierarchy of protein structure illustrating the levels of selforganization and stability of globular proteins.
packing characteristics and the low hydrophobic surface area of the native tertiary structure. The term “molten globule” has been used to describe these early, partially folded equilibrium intermediates on the folding pathway, which also form at low pH or medium denaturant concentration (Ptitsyn, 1995). Its properties may be summarized as follows: high content of secondary structure, overall compactness with highly mobile aromatic side chains, exposure of hydrophobic surface, tendency to aggregate, lack of highly cooperative (thermal) unfolding, and rapid equilibration with the unfolded state. Evidently, the given characteristics are not well defined. Attempts to distinguish between native-like structures or irregular collapsed unfolded forms have led to much confusion regarding various molten globule “states”; for a clarification of the terminology, see Baldwin (1991). At a later stage of folding, stable tertiary structure is formed by intramolecular shuffling. If the resulting tertiary structure still exposes excessive hydrophobic surface area or specific charge patterns, intermolecular quaternary contacts will finally cause assembly. 1.3. Denaturation-Renaturation The underlying principle that allows us to analyze single steps along the folding pathway of proteins is that most proteins undergo at least partial refolding after preceding denaturation, with the native structure as the final state of the cycle (with N, U, N* as native, unfolded and renatured states, respectively).
(1)
Molecular chaperones and folding catalysts
452
Experimental evidence proving that the native and renatured states are indistinguishable has been accumulated using all available physical, biochemical and immunochemical methods. Regarding the difference between directionality of translation from the N- to the C-termini, on the one hand, and integral refolding of the complete polypeptide after preceding denaturation, on the other, three sets of experiments may be taken as proof that vectorial and integral folding are compatible: (i) Merrifield synthesis of ribonuclease and other polypeptides from the C- to the N-terminal end has been shown to yield authentic, active proteins (Merrifield, 1993). (ii) Truncation experiments, as well as limited proteolysis of domain proteins and circular permutations of domains clearly indicate for a variety of systems that removing N- or C-terminal extensions, and altering the succession of domains may leave the overall structure (and function) unchanged (Jaenicke, 1987, 1996). In the case of the domain constructs, this conclusion is in accord with the concept of independent “folding-by-parts” (Wetlaufer, 1981) which has long been considered as one solution to Levinthal’s paradox (1968) that the self-organization of proteins would require an astronomical period of time if folding were the stochastic exploration of the conformational space for the global minimum of potential energy, (iii) Immobilization of proteins at their N- or C-terminal end has been used as another means to invert the direction of folding. Tagging polypeptide chains (with oligo-glutamic acid or oligoarginine extensions fixed at their natural N- or C-termini) to an ion-exchange resin has been shown to allow successful denaturation/renaturation of proteins both in the N→C and C→N direction; in both cases, the products of reconstitution were equally active, showing identical physical properties after release from the column (Stempfer et al., 1996). In summary, translation and denaturation/renaturation yield identical final products so that, as a working hypothesis, in vitro renaturation experiments may be assumed to mimic the cellular process of protein self-organization. There have been occasional reports of proteins which—for unknown reasons—are inaccessible to refolding. Examples are glutamate dehydrogenase (Müller & Jaenicke, 1980; West & Price, 1988), blood coagulation factor XIIIa (Rinas et al., 1990), and creatinase (Schumann & Jaenicke, 1993). For most proteins, optimization of in vitro renaturation has been shown to reach high yields (Jaenicke, 1987). In comparing these results, the protein in its initial state is used as 100% reference. In this context, it is commonly assumed that stucture formation in vivo yields exclusively native protein. However, a number of observations clearly indicate that there is misfolding and misassembly in the cell (Hurtley & Helenius, 1989). Obviously, protein secretion or translocation through the endoplasmic reticulum (ER), and subsequent specific degradation provide an inherent quality control, leading to an apparent yield of 100%. Under unbalanced physiological conditions, in vivo folding, like in vitro reconstitution, gives rise to wrong conformers which may reduce the yield
Spontaneous versus assisted protein folding
453
considerably, in limiting cases down to 0%. This holds true especially for overexpressed recombinant proteins which, as a consequence of kinetic partitioning between folding and aggregation, may be trapped at early stages of folding, ending up in inclusion bodies (Mitraki & King, 1989; Rudolph, 1990). The fact that proteins, due to their low free energy of stabilization, are basically close to their denaturation transitions, implies that the native state in vivo may occupy a whole set of substates where “misfits” are continuously removed by proteolysis. Thus, again, the yield of (re-)folding does not represent a significant difference between structure formation in vivo and in vitro. The following discussion focuses, first, on spontaneous folding, reviewing the various levels of the hierarchy of protein structure and self-organization. Second, rate-limiting steps and their enhancement by folding catalysts will be considered. Finally, we shall return to the question how far in vitro experiments do provide insight into the cellular processes of structure formation, and at which point the classical reductionist’s approach of physical biochemistry needs to be revised in order to cope with the complexity of the multicomponent multiparameter system of the living cell. 2. FOLDING OF SMALL SINGLE-DOMAIN PROTEINS The elucidation of the folding path of any well-chosen protein would require the complete description of the nascent (unfolded) and final (native) states, together with all intermediates along the Un→N transition: (2) where Un is the ensemble of unfolded conformations, Ii a series of intermediates in sequential order, and N the functional native state. Due to (i) the structural degeneracy of Un, (ii) the limited time resolution and (iii) the question whether there is one sequential pathway or multiple pathway funnelling in a complex energy landscape, this formidable task has not been solved for any protein so far. The most detailed mechanisms which have been worked out in the past, refer to small single-chain one domain proteins with and without cystine crossbridges. Considering the best-known standard systems: basic trypsin inhibitor (BPTI), ribonuclease (RNase A and RNase T1), cytochrome c, hen eggwhite lysozyme, chymotrypsin inhibitor CI-2, and barnase (Figure 2), each has been a role model for a particular facet of the folding problem. This indicates that there is no consensus pathway of protein folding, except for the above mentioned general idea of hierarchical condensation. Evolution seems to have selected individual solutions to the problem how a nascent polypeptide chain can reach the native state, facing the kinetic competition between folding, association, chemical modification and degradation. A couple of examples may illustrate the present state of the art. 2.1. Rate-determining Late Folding Events There are two slow reactions along the folding pathway which serve two quite different
Molecular chaperones and folding catalysts
454
functions: the oxidation of cysteine residues to form cystine crosslinks contributes to stability; the cis-trans isomerization of proline residues serves as a switch providing the necessary local flexibility of the polypetide chain in late shuffling processes during tertiary structure formation. Due to the enormous number
Figure 2 Architecture of model proteins used in protein folding studies. The regular secondary structure elements of globular proteins may consist entirely of -helices, like in myoglobin (a), or -strands, like in.the fast folding cold-shock protein from Bacillus subtilis (b). Globular proteins may be built from alternating helices and strands, like the / -barrel family of proteins (c, triose phosphate isomerase), but most of them contain helices and -sheets in less symmetrical
Spontaneous versus assisted protein folding
455
arrangements (d-k). Most of the small proteins depicted here (d, pancreatic trypsin inhibitor BPTI; e, ribonuclease A; f, chymotrypsin inhibitor 2; g, lysozyme; h, ribonuclease T1; i, barnase) have been model subjects of studies on protein folding and stability. A view on a subunit of a larger protein, lactate dehydrogenase (LDH, j), discloses its two-domain nature, with the nucleotide binding domain in the upper right, and the substrate binding domain in the lower left. The same subunit is shown in bold lines in k as part of the complete LDH tetramer. Figures were prepared with MolScript (Kraulis, 1991). using coordinates taken of protein data bank entries 1MBO, 1NMG, 1YPI, 4PTI, 1ROB, 2CI2, 1LZ1, 1RNT, 1RNB, and 1LDN.
of combinatorial alternatives in crossbridge formation, on one hand (Anfinsen & Scheraga, 1975), and the high activation energy of proline isomerization, on the other, both reactions are catalyzed in the cell (see below). Early experiments to probe the conformational transitions during stucture formation following disulfide cross-linking of bovine pancreatic trypsin inhibitor (BPTI), were the first successful approach to unravel the folding pathway of a small protein (Creighton, 1992). Optimizing the technique by acid quenching (to block SH/SS exchange reactions) and HPLC (to separate intermediates), it turned out that the intermediates accumulating in the folding pathway contain a subset of the correct cystine crosslinks (Weissman & Kim, 1991). X-ray crystallography, two-dimensional 1H-NMR, and circular dichroism showed that intermediates with one or two disulfides exhibit a compact conformation that is very similar to that of the native protein (Eigenbrot et al., 1990; van Mierlo et al., 1994). The redox system necessary to introduce the last crosslink via disulfide exchange can be provided by a cysteine residue in the pro-sequence of the BPTI precursor which takes care of the high local SH-concentration required for the reaction with the sterically inaccessible cysteine residues in the major two-disulfide intermediates (Weissman & Kim, 1992). In the cell, oxidative folding is strongly accelerated and the accumulation of BPTI folding intermediates is prevented by protein-disulfide isomerase (Creighton et al., 1993). Because of the required oxidation potential in the cell the reaction takes place only in the ER or in the periplasm. As a consequence, hardly any proteins with disulfide bonds are found in the cytosol. Using ribonuclease (RNase A) as a representative example for a single-chain protein with the size of an average domain, denaturation/reduction and subsequent controlled reoxidation experiments gave the first proof for the one-to-one relationship of the primary and tertiary structure of proteins. However, with 4 disulfide bonds, i.e., 105 possible combinations of SH groups and 7193 mixed disulfide intermediates, no folding mechanism could be established. Instead, the oxidized (native) enzyme was chosen as a paradigm in order to identify rate-determining slow steps on the folding pathway (Anfinsen & Scheraga, 1975). The primary aim was to explain the apparent discrepancy of thermodynamic two-state and kinetic multi-state behavior which clearly pointed to a kinetic scheme with at least three states:
Molecular chaperones and folding catalysts
456
(3a)
(3b) (with U=unfolded state, Us, UF=slow and fast folding species, and N=native state). The explanation on the basis of Brandts hypothesis that the isoenergetic cis-trans isomerization of proline residues might be involved in the two-state transition (Brandts et al., 1975), has been confirmed by a wide range of experiments (Schmid et al., 1993). With respect to spontaneous versus assisted folding, the essential point is that proline cistrans isomerization, next to disulfide crosslinking, is the second rate-limiting process in the overall folding reaction of single-chain proteins. Its significance is high as indicated by the findings that (i) the nascent trans-conformation of the peptide bond is isomerized to cis in approximately 7% of all prolyl
Figure 3 Protein folding kinetics studied by NMR spectroscopy. NMR spectra of -lactalbumin, a structural homolog of lysozyme, recorded at varied time increments after initiation of refolding. The positions of the peaks reflect the chemical environment of methyl and methylene protons. Reproduced from Balbach et al. (1995) with permission.
residues in native proteins (in RNase, two out of four prolyl residues are in cisconformation) and (ii) there is now clear evidence that proline cis-trans isomerization plays a role in protein folding in vivo (Fischer, 1994; Rassow & Pfanner, 1996). In connection with the question of how the acquisition of the native-like conformation and disulfide formation are interconnected, RNase T1 mutants were used to demonstrate
Spontaneous versus assisted protein folding
457
that low stability and high flexibility are essential for correct oxidative protein folding (Frech & Schmid, 1995). 2.2. Early Intermediates RNase has been the first paradigm in the elucidation of protein folding, applying pulsed hydrogen exchange in conjunction with NMR spectroscopy (Udgaonkar & Baldwin, 1990; Kim & Baldwin, 1990; cf. Figure 3). Present evidence shows that the -sheet is formed rapidly and cooperatively, shortly after the start of folding. In contrast to a simple sequential model, its stability increases over the first 0.4 s so that one must assume that “the early intermediate” actually represents a broad distribution of species which gradualy changes until side-chain interactions lock the fluctuating polypeptide chain in the stable -sheet conformation. Similar “funnelling” behavior has been observed for other proteins (see below). Based on the observation that the folding mechanism corresponds to the reverse unfolding reaction (Kiefhaber & Baldwin, 1995), the questions whether partially folded intermediates are significantly populated on the folding pathway and how they look like, has also been analyzed starting from the well-defined N-state of RNaseA rather than the degenerate Un states. Monitoring circular dichroism and chemical shift dispersion (pulsed hydrogen-deuterium exchange), the start of unfolding is accompanied by a wide-spread unlocking reaction of side chains which become free to rotate even though the peptide hydrogen bond network still remains intact. The faster exchanging protons correspond to “weak points” in the three-dimensional structure; they are found mainly at the ends of the -strands. Early intermediates on the folding path are difficult to characterize because of their short lifetime. Spectroscopic approaches (UV absorbance, fluorescence, CD), as well as protein-chemical methods (limited proteolyis, antibody binding, H-D or H-T exchange) are commonly restricted to the analysis of global structural changes with a relatively low time resolution, having in mind that the rates of helix-coil transition and -structure formation are in the s time range. However, initiating the folding reaction by a light pulse may shift the time resolution to the s or even ns range, measuring the change in absorption of an intrinsic chromophore to monitor the kinetics. Cytochrome c (Cyt c) has been the model case for this approach (Jones et al., 1993), using the ns photo-dissociation of the heme-CO complex to trigger the folding of the unliganded enzyme. By monitoring time-resolved absorption spectra, transient binding of native and non-native axial amino-acid ligands is observed in the s time range, i.e., before folding begins. Obviously, folding of Cyt c is preceded by the formation of transient loops of the polypeptide chain with non-native contacts between residues or clusters of residues. This is what has been hypothesized earlier as “collapsed form” of the folding protein which would subsequently relaxe or shuffle to reach the native tertiary structure. Another approach to unravel early folding events, with broader application but lower time resolution (ms), has been 2D NMR spectroscopy with rapid pulse labeling, using stopped-flow multimixing techniques combined with the trapping and subsequent identification of exchangeable amide protons (Udgaonkar & Baldwin, 1990; Roder &
Molecular chaperones and folding catalysts
458
Elöve, 1994). Since structure formation is accompanied by reduced exchange rates, the method provides a direct measure of secondary structure or core formation upon folding. Early results seemed to corroborate the above mentioned heterogeneity of protein fractions and folding kinetics, this time in the seconds time range. However, avoiding kinetic barriers along the folding path by using proper solvent conditions, the time constant for Cyt c folding, was found to be reduced to essentially the same rate as the molecular collapse (< 15 ms) in a single kinetically unresolved step, i.e., without populating observable intermediates (Sosnick et al., 1994). This observation is in contrast to the view that particular steps in protein folding (including the supposedly rate-limiting molten globule to native transition) are intrinsically slow. Instead, it appears that folding intermediates may be kinetically trapped by barriers that are optional rather than integral to the folding process. Major barriers may result from misfolding of the polypeptide chain in the initial collapse step. Fast folding is not unique to cytochrome c: chymotrypsin inhibitor-2, ubiquitin, RNaseA, the Ig binding domain of streptococcal protein G and the cold-shock protein from Bacillus subtilis have been shown to fold at similar rates (Baldwin, 1995; Jaenicke, 1996). Whether the fast rate in all these systems depends on the fact that the (re-)folding polypeptide chain, in collapsing to its early intermediate state, avoids non-native ligations needs further experimental verification. In the case of the refolding of cytochrome c, protection against NH proton exchange shows multiphasic kinetics with at least three phases. Taking the NH exchange rates as a measure, the N- and C-terminal helices are formed within ca. 20 ms, whereas most of the other NH protons follow in the 200 ms time range; the slowest exchange rates (10 s) are attributable to groups involved in tertiary contacts. It is important to note that not only the formation of the helices at the N- and C-terminal end is an early event, but also their specific docking. Being the most stable structural elements in the native molecule, this clearly contradicts the idea that cotranslational (“vectorial”) folding from the N- to the Cterminal end of the nascent polypeptide chain might be decisive for the folding pathway and/or the final structure of the functional protein. In contrast to helix formation in cytochrome c, in the case of ubiquitin and the cold-shock protein from B. subtilis, structure formation has been shown to be involved in very fast early folding events (Briggs & Roder, 1992). On the other hand, the all- interleukin 1 b refolds slowly, with a half time of about 20 min (at 4°C) (Varley et al., 1993). Evidently, there is no clear correlation between either the size or the stability or the structural type of a protein and its folding rate. 2.3. Multiple Pathways Egg lysozyme has been an ideal model for protein folding because it is similar to RNase regarding its size and crosslinking pattern, with two important differences: (i) lysozyme does not contain cis proline so that folding of the oxidized protein is not perturbed by the rate-determining late folding events; (ii) the enzyme shows certain properties of a twodomain protein, in spite of the fact that the two lobes that constitute the “active cleft” do not represent contiguous stretches along the amino acid sequence. Strong evidence that two independent “domains” can be differentiated came from X-ray data at elevated
Spontaneous versus assisted protein folding
459
pressure (Kundrot & Richards, 1987). In accordance with the topological “ - and -domains” both pulse labeling H-D exchange experiments and 2D NMR allow two regions of the enzyme that differ in their NH protection to be distinguished (Mirankar et al., 1991). Thus, the kinetic mechanism contains an intermediate in which one lobe has reached its compact native-like state, whereas the other is still unfolded or in the process of folding. A more detailed analysis (monitoring H-D exchange by mass spectrometry) indicated that the enzyme does not fold in a single cooperative event. Instead, different parts of the structure become stabilized with different kinetics, the a-helical domain folding faster than the -sheet domain. Furthermore, different populations of molecules fold by kinetically distinct pathways, so that the results force us to believe that folding of oxidized lysozyme is not a simple sequential reaction, but involves parallel alternative pathways (Evans & Radford, 1994). Comparing the kinetics of the reduced polypeptide chain, it has been shown that decreasing the structural constraints leads to a drastic decrease in the folding rate (Goldberg & Guillou, 1994). Thus, the question remains whether the cystine crosslinks may be important determinants in the above kinetic mechanism. In summarizing the previous results, there is clear evidence that folding intermediates with native-like structure accumulate. The fact that they are not well synchronized seems to contradict the classical sequential framework model (Kim & Baldwin, 1990). Theories of the folding process suggest that neither the folding pathway nor the set of folding intermediates is unique, and that folding intermediates only accumulate because they are trapped kinetically by partial misfolding. This view resembles the jigsaw puzzle model (Harrison & Durbin, 1985) which, in the past, has been neglected because experimental data give clear evidence that there are folding pathways with successive intermediates and hierarchical order. Further experiments are needed in order to distinguish between the classical and the new view. As has been pointed out, one important issue in this context is the consideration of secondary structural elements as a major factor determining the folding pathway. One may assume that robust and fast folding pathways have been selected through evolution. Thus, random misfolding in computer simulations may not necessarily apply to the folding behavior of real proteins (Baldwin, 1995). 2.4. Protein Engineering Analysis of Protein Folding With barnase, a 100 amino acid, single-domain extracellular RNase from Bacillus amyloliquefaciens, protein engineering has entered the analysis of protein folding (Fersht, 1995a). As the enzyme contains -helices and an antiparallel -sheet, but no cysteine and cis-proline, it offers itself as a paradigm for folding studies on a stable protein without covalent crosslinks. In using kinetic and equilibrium unfolding/refolding measurements on more than 60 mutants, structure formation in transition states and intermediates was mapped. The fundamental assumptions that the mutations neither perturb the structures of the folded and unfolded states nor the folding pathway, and that the target groups make no additional interactions with partners in their spatial environment, were studied by measuring the structure and stability increments of the standard states and the first significant transition states of unfolding. The structure of the transition state (resulting from urea denaturation) is that of the native-like enzyme, with
Molecular chaperones and folding catalysts
460
its hydrophobic core weakened and several of the tertiary interactions and loops lost, but with the majority of the secondary structure elements, including tight turns, maintained. This implies that the last events in folding must be the consolidation of the hydrophobic core, the closing of loops and slight rearrangements of tertiary contacts, without any indication for parallel pathways. The same approach as in the case of the transition state for unfolding was applied in order to characterize the structure of an intermediate on the refolding pathway (for details, cf. Jaenicke, 1996). This time, the order of events is: early formation of the N-terminal -helix, the -sheet, part of the core, and docking of the Cterminus to this “nucleus”; subsequent steps stabilize the rest of the core and the loops, with the tertiary contacts as the coda. The theory has been confirmed and supplemented by independent evidence taken from H-D exchange NMR studies. The general conclusions may be summarized as follows: There is a compulsory pathway of folding which is, at least in part, sequential. Secondary structure formation is driven by the local minimization of hydrophobic surface area; it precedes tertiary structure formation. Tertiary interactions become increasingly defined as water release consolidates the hydrophobic core. 2.5. Chymotrypsin Inhibitor 2 (CI-2) Chymotrypsin inhibitor 2 from barley is a 64-residue, single-domain protein of mixed secondary structure. Protein engineering of CI-2 has been employed to map the interactions characterizing the transition state of its folding reaction (Fersht, 1995a). This reaction proceeds rapidly and without a populated intermediate for those CI-2 molecules having all proline residues in the native trans conformation at the onset of refolding. For a large number of mutant proteins with amino acid substitutions scattered throughout the sequence, folding and unfolding rates were extrapolated to zero denaturant concentration and compared to the extrapolated free energy of folding determined in equilibrium unfolding experiments. In addition, amino-terminal CI-2 fragments comprising 5–63 of its 64 residues as well as a nicked form of the protein and its point mutants have been examined by equilibrium and kinetic methods. Major conclusions are the following: (i) the CI-2 folding reaction conforms to a nucleation-collapse model, where residues in an -helix form a local nucleation site, around which all of the secondary structure collapses in a single cooperative process; (ii) the transition state of CI-2 folding is like an expanded form of the native structure without fully formed secondary structure elements; (iii) the nascent CI-2 polypeptide chain has very little tendency to fold before completion of its synthesis (Fersht, 1995b). 3. FOLDING OF DOMAIN PROTEINS Beyond a certain limiting length of the polypeptide chain, proteins consist of domains which may be considered as independent folding units giving rise to separate phases in the process of folding/unfolding. In the simplest case of two-domain proteins, this may be described by a three-state model
Spontaneous versus assisted protein folding
461
(4) where I represents an intermediate with one domain still intact and the other unfolded (Figure 4A). If the whole molecule represents one single cooperative unit, the population of I remains undetectable, and Eqn. (3a) is sufficient to analyze the equilibrium transition.
Figure 4 Folding of multidomain and multisubunit proteins. A, equilibrium unfolding of bovine B-crystallin, as observed by fluorescence spectroscopy (◯) and analytical ultracentrifugation (●) at pH 2. The amino-terminal domain is stably folded whereas the carboxy-terminal domain is denatured in the intermediate populated around 3 M urea. B, reactivation of LDH after acid denaturation is speeding up with increased protein concentration, until a first-order folding reaction becomes rate-limiting in the uni-bimolecular reaction scheme. The addition of the coenzyme has no effect on the reactivation kinetics.
As taken from in vitro translation experiments, independent domain folding takes place also in the cell: using immunoglobulin and serumalbumin, it has been shown that intradomain cystine-bridges are formed sequentially during translation (Bergman & Kuehl, 1979; Peters & Davidson, 1982). Essentially, folding-by-parts must be considered a most significant acquisition of evolution for a number of reasons: (i) it enhances the folding rate by synchronous folding at multiple sites along the nascent polypeptide chain, (ii) it is a most efficient way to exclude wrong intramolecular interactions in the case of large protein molecules, (iii) it protects the nascent polypeptide chain from proteolysis, and (iv) it may be considered a simple mechanism to proceed from monomeric to multimeric proteins by domain swapping (Bennett et al., 1995; Schlunegger et al., 1997; see below). Regarding the folding mechanism, domain proteins may be considered as the sum of their constituent parts; this means that what has been discussed in connection with the sequence of consecutive steps in single-domain proteins, holds unchanged. In the overall kinetic scheme, the rate-determining reactions are of first order. Depending on whether the reaction is two-state or three-state, one or two pairs of rate constants will be detectable. There are examples where biological function requires the cooperation of domains,
Molecular chaperones and folding catalysts
462
e.g., in dehydrogenases which contain NAD- and substrate-binding sites residing in two separate domains (D1, D2). In such cases, domain folding may precede domain pairing as the rate-determining step, according to
(5) In cases where proline isomerization or other slow reactions participate in the overall mechanism, the kinetic mechanism will lead to schemes of even higher complexity (Garel, 1992). Using intrinsic markers (fluorophores, epitopes for monoclonal antibodies, ligands), sequential folding steps may be resolved on the time scale from a few milliseconds to seconds, in accordance with the sequential model. 4. ASSOCIATION In advancing from domain proteins to protein assemblies, we approach cellular substructures and finally the macroscopic world: surface layers, the microtrabecular lattice, the tubulin-dynein system, flagella, ribosomes, the extracellular matrix, muscle, are all selfassembly systems involving proteins or protein conjugates. From the point of view of the structural hierarchy of proteins, oligomerization corresponds to domain pairing, except that the docking process is dominated by non-covalent interactions. 4.1. Sequential Folding and Association Model reactions simulating quaternary structure formation made use of proteolytic fragments. In certain cases they were found to exhibit high specificity of subdomain or domain interactions which allow them to recognize and complement each other. In general, the association process is entropy-driven as a consequence of water release from the subunit interfaces. There may be a significant contribution to protein stability from quaternary stucture formation (Lauffer, 1975; Jaenicke, 1991). Complementation requires the correct recognition sites to be preformed; this means that fragments or domains that are expected to trigger the assembly process, must fold autonomously. The overall reconstitution can then be visualized as a sequential folding-association reaction, where folding provides the correct docking surfaces allowing the consecutive association reaction to take place. At low concentrations, association becomes rate-determining (Jaenicke, 1987). Thus, quaternary structure formation may affect both the stability of proteins and their rate of folding. In describing the complete association pathway, the steps preceding subunit docking are the same as in domain proteins. The overall mechanism consists of three stages: first, formation of elements of secondary and super-secondary structure, second, collapse to subdomains and domains, ending up with structured monomers, and third, association to form the correct stoichiometry and geometry of the native quaternary structure. Evidently, the “collision complex” of the structured monomers may still undergo intramolecular rearrangements in order to reach the state of maximum packing density
Spontaneous versus assisted protein folding
463
and minimum hydrophobic surface area. Thus, the uni-bimolecular folding/association mechanism may involve further first-order steps belonging to slow shuffling processes at the level of the native-like assembly. Focusing on the rate-limiting steps, in the simplest case for a dimer, the overall reaction would then obey a uni-bi-uni-molecular reaction according to
(6) with M, M as unfolded and structured monomers, N as the native dimer and k1, k2, as first- and second-order rate constants (Jaenicke, 1987; Garel,1992). How the single steps along the folding/association pathway can be monitored depends on the specific structure-function relationship for a given system (cf. Figure 4B). In most cases, biological function relates to the native quaternary structure such that the final ratedetermining step can be measured by the regain of activity. Preceding steps may be accessible to spectral analysis, cross-linking and a wealth of other methods (Rudolph et al., 1996). 4.2. Specificity of Association Considering the crowding of a great variety of different components in the cell, one important aspect of protein folding and association is the specificity of subunit recognition, i.e. the question of whether or not other proteins may interfere with the formation of the correct native quaternary structure. A qualitative criterion for the fidelity of subunit recognition was gained from renaturation experiments in crude mixtures where reactivation in the presence of excess foreign protein can be considered a direct measure of correct quaternary structure formation. For example, in refolding recombinant antibodies, no significant differences in the yield as well as the kinetics are observed in the homogenous system compared to the crude mixture obtained upon braking up the E. coli cell (Buchner & Rudolph, 1991). A quantitative investigation made use of pairs of topologically related enzymes. Applying strictly synchronized reactivation conditions, neither hybrid intermediates nor chimeras as endproduct were detected (Gerl et al., 1985). As has been mentioned, the same high degree of specificity holds at the level of domains (Wetlaufer, 1981; Opitz et al., 1987). Clear evidence that hybrid formation may occur comes from isoenzymes, (e.g. the five isoforms of lactate dehydrogenase, LDH), or from multifunctional enzymes where certain gene products have been found as subunits in different complexes (e.g. protein disulfide isomerase (PDI) in prolyl hydroxylase). LDH isoenzymes show exceedingly high structural homology leading to complementary subunit surfaces. In cases in which isoenzymes in different compartments are involved, both target sequences and different folding mechanisms may contribute to specificity. For example, in the case of dimeric mitochondrial and cytosolic malate dehydrogenases, the second-order subunit assembly of m-MDH is preceded by slow folding (which may even be retarded by the signal sequence), whereas for c-MDH first-order reconstitution indicates diffusioncontrolled association (Jaenicke, 1987). It might be attributed to these mechanistic
Molecular chaperones and folding catalysts
464
differences that all attempts to trap hybrid dimers of MDH, either in the process of reconstitution or in equilibrium experiments failed (Jaenicke, unpublished results). 4.3. Superstructures So far, the relationship of folding and association was discussed mainly considering oligomers. However, successful attempts to reconstitute highly complex biological systems such as phage or the ribosome or giant multienzyme complexes have clearly demonstrated that, in going from oligomeric to multimeric structures, no fundamental differences occur (Jaenicke, 1987; 1996). Regarding their structural analysis, unpredicted progress in these fields has been accomplished thanks to technical developments in the fundamental methods, i.e., electron microscopy, X-ray analysis and NMR. The greatest impact in this connection came from cryo-electron microscopy, synchrotron radiation, the perfection of the image plate and the analysis of tractable structural elements. There are good prospects to fit these elements so that insight into the structure-function relationship up to the level of complex macroassemblages may be gained. Evidently, increasing numbers of subunits lead to a wealth of topological variants which may require specific assembly programs for their regulation. Available data show that cells contain more or less all platonean bodies as compact core structures; in addition, there are hollow cylindrical or helical rods and shells and, at the other extreme, one- and two-dimensional assemblies. A wide variety of methods have been used to elucidate assembly pathways: X-ray crystallography; electrone microscopy; chemical cross-linking, hybridization, recombinant and mutant techniques etc. An example will be discussed in connection with the in vivo and in vitro assembly of the tailspike protein of Salmonella bacteriophage P22. 5. OFF-PATHWAY REACTIONS 5.1. Thermodynamics vs Kinetics The problem of off-pathway reactions in the process of protein folding relates to the questions whether the folded structure is determined by its thermodynamic stability, and how it is possible for a polypeptide chain to fold up along a pathway that allows rapid folding and, nevertheless, arrive at the thermodynamic most stable structure. The compromise may be a consequence of natural selection, insofar as evolution may have selected sequences that have the ability both to fold rapidly and to arrive at thermodynamically stable structures. However, there are two challenges to this hypothesis: First, proteins uncapable of reversible unfolding/refolding and, second, proregion dependent folding of (pre-)proproteins or proenzymes. To illustrate the first point, the kinetic partitioning between the assembly and aggregation of the tailspike protein (TSP) of Salmonella typhimurium bacteriophage P22 may serve as an example (cf. Figure 7). TSP is a homotrimer which is noncovalently bound to the neck of the virus capsid and essential for phage adsorption to the bacterial host. The protein has served as a model system for the folding and assembly of large
Spontaneous versus assisted protein folding
465
multi-subunit proteins. Its folding pathway comprises subunit folding, followed by the formation of a protrimer, in which the chains are stably associated but not fully folded, and a final shuffling reaction from the protrimer to native TSP as the rate-limiting step (Goldenberg & King, 1982; Fuchs et al., 1991). Along the whole reaction sequence, offpathway aggregation competes with proper folding and association. Numerous point mutants enhancing or suppressing aggregation at high temperature have been isolated, proving that single amino-acid substitutions may profoundly affect the partition ratio between the two competitive processes (Mitraki & King, 1992). The in vitro refolding and the thermal stability of a number of these mutant proteins have been characterized in detail, with the general result that the physicochemical observations precisely complement the genetic analysis (Danner & Seckler, 1993). On the basis of the in vitro results and the recent high-resolution structure of the N-terminally shortened protein (Steinbacher et al., 1994), the folding pathway and the high thermal stability of the tailspike protein, as well as the effects of the folding mutations can be rationalized (Beissinger et al., 1995). In connection with the effect of the pro-region of cymogens on the final structure of the processed molecules, the kinetic competition between on-pathway reactions leading to the native state, and non-productive pathways leading to aggregation gains importance. In this context, pro-regions could function either by increasing the rate of the forward folding reaction or by decreasing the rate of aggregation. Chaperones are known to suppress protein aggregation; in contrast, pro-regions function by directly increasing the rate of the forward folding reaction. They are required for folding under conditions in which off-pathway reactions are suppressed; in addition they interact strongly with the product of the folding reaction, i.e., the native state of the processed protein. For example, denatured a-lytic protease, subtilisin, carboxypeptidase A, among others, do not fold to the native state in the absence of the pro-region. Instead, upon removal of the denaturant, the protein folds to a stable intermediate with substantial secondary structure but little organized tertiary structure; upon addition of the pro-region, the intermediate is rapidly converted to the native state (Winther et al., 1994; Baker & Agard, 1994). 5.2. Productive vs Non-productive Pathways As has been mentioned, there are three stages where side reactions on the folding path may compete with proper folding and association of proteins: the hydrophobic collapse, the merging and swapping of domains and the docking of subunits (Figure 5). An example for the first was discussed in connection with the initial phase of cytochrome c folding where transient interactions of non-native and native amino-acid ligands to the heme iron were found to precede correct folding; cases illustrating the second and third are the domain swapping in crystallins, on one hand, and inclusion body formation, on the other. At all three levels, correct folding requires specific substructures to be preformed in order to proceed on the correct folding path. Collapse and domain merging involve intramolecular rearrangements. Due to the high local concentrations of the reacting groups, they are not significantly affected by neighboring molecules, i.e., they obey first-order kinetics with the slowest isomerization reaction determining the overall rate. In the case of domain proteins, the relative stabilities of the
Molecular chaperones and folding catalysts
466
domains and the contributions of the domain interactions to the overall stability are crucial. The significance of the linker peptide connecting two well-defined domains has been studied by grafting experiments, e.g., by mutually exchanging the linkers of - and -crystallins (Mayr et al., 1994; Trinkl et al., 1994; Jaenicke, 1994). In both cases, domain contacts dominate over
Figure 5 Side reactions of protein folding. Polypeptide chains may become trapped during folding in conformations representing local minima on the multidimensional conformational landscape (A, reproduced from Mirny et al. (1996) with permission). At high protein concentration, interactions between structural elements maybe formed intermolecularly rather than intramolecularly, leading to off-pathway aggregation (B).
subunit contacts. At concentrations up to 0.5 mM, the recombinant separate domains do not interact with each other, which stresses the above local concentration argument. In going from single-chain domain proteins to protein assemblies, kinetic competititon of first-order folding and second-order association becomes important as soon as the protein concentration reaches the level where folding becomes rate-limiting (Zettlmeissl et al., 1979) (Figure 6). As has been discussed, the reason for this is that subunit assembly requires the monomers to be close to their proper conformation before they coalesce to form the native quaternary structure. If folding intermediates expose wrong contact sites, they will give rise to aggregation instead of association, because the folding polypeptide chain does not distinguish between intra- and intermolecular interactions. Accordingly, not only a shift in the kinetic mechanism but also a decrease in yield of active protein will be observed. The occurrence of inclusion bodies rather than soluble protein in overexpressing recombinant genes illustrates the consequence. The underlying kinetic mechanism
(7)
Spontaneous versus assisted protein folding
467
resembles the previous sequential uni-bimolecular mechanism which represented the limiting case at high dilution. With increasing protein concentration, when k1
Figure 6 Kinetic competition between folding and aggregation. A, The reactivation yield of LDH (●) decreases with increasing protein concentration, concomitant with an increase in the amount of aggregated material ( ). There is little difference in the specific circular dichroism (□) between renatured LDH and small aggregates proving that the aggregates contain much secondary structure (compare Figure 5 B). B, The protein concentration dependence of LDH renaturation yields (●, data from A) can be modeled as the result of a kinetic competition between unimolecular folding and bimolecular aggregation, as indicated by the solid line (Kiefhaber et al., 1991).
determines the overall rate, the limiting value for k2 will correspond to diffusioncontrolled association. If this explanation is correct, and if the native protein is in a lower energy minimum than the aggregates, it should be possible to manipulate the kinetic competition by adding weakly destabilizing agents. In fact, it has been shown, that the yield of reconstitution may be improved, e.g., in the presence of moderate concentrations of arginine or urea (Rudolph, 1990). Considering Eqn. (7), three questions need to be answered: (i) what is the committed step in aggregate formation, i.e., at which stage along the sequential reaction are aggregates formed, (ii) when is the structured monomer committed to end up as the native protein, and (iii) what is known about the structure of aggregates and their constituent polypeptide chains? With respect to the first two problems, commitment to aggregation was shown to be a fast reaction, whereas the kinetics of the commitment to renaturation follows precisely the slow kinetics of the overall reaction (Goldberg et al.,
Molecular chaperones and folding catalysts
468
1991). This means that early, collapsed intermediates are much more sensitive to aggregation than later species, but only the native state is fully protected from misassembly. Regarding the structure of aggregates, electron microscopy and circular dichroism indicate that wrong subunit interactions give rise to irregular networks with a broad distribution of highly structured particles at least 10 times the size of the native proteins. They resemble the native protein in its spectral properties, as far as perturbations by turbidity allow a quantitative analysis (Figure 6). Inclusion bodies may represent surprisingly homogeneous precipitates, often with less than a dozen components. Whether disulfide bonds contribute to their compactness in the cell, or whether they are preparative artifacts is still controversial (Mitraki & King, 1989; Valax & Georgiou, 1993). 6. THE IN VITRO VS. IN VIVO ISSUE Aggregation in the cell causes inclusion body formation, because of the high local concentration of folding intermediates, i.e., for the same reasons that are responsible for aggregation in vitro. From this parallelism one may conclude that there is no difference between folding in vitro and folding in the cell, except that overexpression might be considered unphysiological and, therefore, atypical for the standard cell. In this context, it is important to note that there is misfolding and misassembly also in the cell (Mitraki & King, 1989; Helenius et al., 1992). Obviously, secretion or trafficking through the ER and Golgi, and subsequent specific degradation provide an inherent quality control, clearing the cell from “misfits”. Thus, the recovery of refolding does not represent a significant difference between structure formation in vitro and in vivo. As has been discussed before, the same holds for the directionality of protein biosynthesis in the cell which is in contrast to the integral folding of the complete polypeptide chain in vitro. As indicated by the successful “renativation” of proteins under a wide variety of conditions, cytosolic solvent conditions, cotranslational and posttranslational modification, transcriptional or translational control etc. do not play significant roles in the folding process. In addition, tertiary structure formation and subunit association have been shown to tolerate a wide range of variations with respect to sequence and chain connectivities: circular permutations of parts of the sequence, chain extensions, covalent joining of subunits, fragmentation, derivatization, hybridization by peptide interchange etc. (Jaenicke, 1993). Obviously, certain core regions of a protein determine the overall topology, whereas peripheral parts of the protein may be altered or even lacking (Privalov, 1994). Regarding the time requirements, nascent proteins acquire their native structure with half-times in the minutes range (at most), whereas in vitro refolding rates may vary in a wide range from seconds to days. One reason is that optimum conditions regarding the yield of refolding do not necessarily represent optimum conditions for the folding rate; in addition, in the cell, accessory proteins are involved in regulating and/or catalyzing the rates of folding and association. That they assist rather than direct folding is generally accepted. To date, protein folding in the cell has been inaccessible to a detailed analysis of its kinetic mechanism and structural intermediates on its folding pathway. Thus, it is still
Spontaneous versus assisted protein folding
469
unresolved whether refolding polypeptides in vitro proceed via the same pathway as nascent chains released from the ribosome within the cell. In the case of the tailspike endorhamnosidase from Salmonella bacteriophage P22, detailed in vitro refolding studies indicate that the acquisition of the native structure closely resembles the self-organization of the protein in vivo (Figure 7A). The properties of refolding intermediates resemble those of intermediates observed after pulse-labeling in vivo and the rate-determining folding reaction occurs with identical rates
Figure 7 Bacteriophage P22 tailspike endorhamnosidase. A, the folding and assembly reactions of phage P22 tailspike protein upon biosynthesis in vivo and upon denaturation by acid and urea in vitro comprise similar, if not identical, intermediates and occur with identical rates
Molecular chaperones and folding catalysts
470
under comparable conditions. B, folding yields of the tailpike protein decrease with increasing temperature both in vivo and in vitro and are affected by single amino-acid substitutions, which alter the thermal stability of thermolabile folding intermediates. C, sites of mutations that decrease (black dots) or increase (open circles) folding yields at high temperature are spread through most of the parallel β-helix (Steinbacher et al., 1994), suggesting that this main structural component of a tailspike subunit is formed early in the folding pathway.
both in the test tube and in the bacteriophage-infected cell (Fuchs et al., 1991; Danner et al., 1993). In vitro translation of firefly luciferase has shown that the ribosome-bound polypeptide chain is essentially inactive; activity appears within a few seconds after release of the enzyme from the ribosome. In contrast, the reactivation of chemically denatured luciferase in the cell-free translation system occurs with a half-time of 14 minutes (Kolb et al., 1994), and the reactivation of pure, completely unfolded luciferase in buffer takes many hours (Herbst et al., 1996). Although the reason for such variable refolding rates is still unknown, the results support the idea that the nascent polypeptide chain folds cotranslationally, corrobo-rating earlier results for immunoglobulin chains (Bergman & Kuehl, 1979; Peters & Davidson, 1982). In summary, the resulting three-dimensional structure is the same, in vitro as well as in vivo. Differences refer to the kinetics of folding and association and to the partitioning between folding/association on the one hand, and aggregation on the other. Shifts of the corresponding partition coefficient are caused by the presence of helper proteins or accessory proteins in the cell. They do not interfere with the “central dogma” that the three-dimensional structure of proteins is determined by their amino-acid sequence. Thus, there remain two questions: (i) are there differences in the mechanism of protein folding within and without the cell, and (ii) what are the effects of cytosolic solvent parameters, on one hand, and folding catalysts and chaperones, on the other? 6.1. Cytosolic Solvent Parameters The solvent conditions in common cells vary only in a narrow range, in spite of the fact that in the biosphere life faces a wide range of extremes of physical conditions. In extremophiles, they may require adaptation of the entire cell inventory to temperatures close to the boiling point of water and presssures ~1200 atmospheres, as well as water activities as low as 0.6. On the other hand, the intracellular pH is kept constant, close to neutrality, even in the case of extreme acido- or alkalophiles (Jaenicke, 1991). In connection with protein self-organization, expressing recombinant proteins from extremophiles in mesophilic hosts, or vice versa, provides information with respect to the impact of solvent parameters on protein folding and association. It turns out that alterations in the folding conditions often have surprisingly little effect. An example is the expression of active enzymes from hyperthermophilic bacteria in E. coli where temperature differences between host and guest may amount to = 60ºC (Jaenicke et al., 1996) After complete denaturation, e.g., of glyceraldehyde 3-phosphate dehydrogenase, renaturation at 5–100°C yields the fully active protein indistinguishable from its initial
Spontaneous versus assisted protein folding
471
native state. From this we may conclude that not only the native tetramer but also the intermediates on its folding pathway must be stable over the whole temperature range. Little is known with respect to other solvent parameters. In the case of halophiles, the few proteins that have been investigated require salt for folding and assembly (Jaenicke, 1991). For non-halophilic proteins, the ionic strength is not as critical, apart from stabilizing Hofmeister effects that may be used to optimize the folding conditions. What is crucial, is the effect of specific structurally or functionally essential ions, either in metallo-enzymes or in proteins stabilized by other ligands. Such individual requirements can easily be quantified in vitro, thus allowing the cellular situation to be mimicked (Jaenicke, 1987). Regarding effects of hydrostatic pressure, no systematic experiments have been performed so far. The same holds for viscosity effects in the crowded cytosol. 6.2. Folding Catalysts As has been mentioned, there are three possible rate-determining steps in the selforganization of proteins: disulfide shuffling, proline cis-trans isomerization and assembly. On the other hand, there are two enzymes, localized in the appropriate cellular compartments to catalyze the first two, and there is a whole “machinery” of chaperones to assist the third. The biological significance of all three reactions has been established, however, their molecular mechanism is still far from being understood. Protein disulfide isomerase (PDIs, in E. coli DsbA) and peptidyl prolyl cis-trans isomerases (PPIs) are abundant and ubiquitous in all organisms, suggesting that disulfide shuffling and proline isomerization are catalyzed both in eu- and prokaryotes. These catalysts are described in detail in other chapters of this volume (Fischer and Schmid; Missiakas and Raina; Freedman and Klappa). Here, we only briefly summarize their activities and discuss their roles in the assisted folding of proteins in a more general sense. PDIs such as DsbA catalyze the oxidative refolding of a number of small substrate proteins by the rapid unidirectional transfer of an active site disulfide to the substrate protein and subsequent disulfide isomerization:
(8)
The catalytic efficiency of the enzyme is evident: adding catalytic quantities of DsbA leads to rapid refolding under conditions where no spontaneous refolding can be achieved in its absence. With respect to substrate specificity, inhibition studies with peptides of various lengths and sequences suggest that PDIs show rather broad peptide-binding capacity. There is clear evidence that DsbA catalyzes protein folding in vitro. It must also act catalytically in the cell since the levels of oxidized substrates can exceed the level of DsbA by almost three orders of magnitude. Being oxidoreductases and isomerases at the same time, the catalytic cycle of PDIs requires reoxidation of the active-site disulfide, making use of low-molecular-weight compounds like oxidized glutathione or of a protein electron acceptor (Bardwell, 1994; Bader et al., 1998).
Molecular chaperones and folding catalysts
472
As for the isomerase activity, catalysis may be operative at two levels, folding of the nascent chain and rescue of misfolded (and aggregated) polypeptides. The latter has been shown to be involved in the dramatic acceleration of the isomerization reactions necessary for the completion of disulfide formation on the folding pathway of BPTI. Apparently, in certain proteins PDI is capable of gaining access to buried thiol groups after the protein substrate has acquired a substantial percentage of its tertiary structure. In the case of the oxidative renaturation of antibodies, this holds only for the first phase of the reaction (Lilie et al., 1993). If PDI can act late on the refolding pathway, it could help explain why the correct disulfide bonds are made: The proteins have already folded so as to place those cysteines that are to participate in disulfide-bond formation in close proximity, supporting the idea that PDIs are true folding catalysts in the sense that they enhance the rate of folding without determining the final structure. In this context they are involved in both formation and isomerization of disulfide bonds in their specific compartments in the cell. PPIs catalyze the rotation around the X-pro peptide bond. Translation yields the alltrans configuration of the polypeptide chain. In cases where the final structure of a protein contains cis-peptide bonds, the high activation energy EA of their isomerization may give rise to slow folding, depending on the neighboring groups. By far the most frequently occurring cis-rotamer in known protein structures involves proline residues:
(9)
In the case of essential prolines EA is of the order of 85 kJ/Mol, so that at room temperature and in the absence of a catalyst the reaction becomes rate-limiting in the overall folding process (Schmid et al., 1993). A variety of structurally unrelated families of enzymes have been discovered which are capable of catalyzing proline isomerization. Three of them, the cyclophilins (CyPs), FK 506 binding proteins (FKBPs) and trigger factor have been investigated in detail (Fischer, 1994). All three are abundant and ubiquitous proteins involved in a variety of cell biological phenomena. Whether their action is in all cases related to their specific catalytic function as PPIs is still unresolved. The catalytic efficiency of the PPIs depends very much on the accessibility of the Xprolyl bonds involved in the isomerization reaction. For small prolyl peptides, the catalytic efficiency approaches the diffusion-limited second-order rate constant for the association of a protein with a small molecule (Schmid et al., 1993). Little is known regarding the sequence specificity; as in the case of PDIs, the substrate specificity seems to be low. The significance of PPI in catalyzing the in vitro refolding of small model proteins containing essential proline residues is well-established. Concerning the biological significance of PPI catalysis, three observations support the idea that the in vitro results reflect cotranslational or posttranslational folding events in the cell: (i) the striking evolutionary conservation of PPI function, together with the proper loca-tion and ubiquity
Spontaneous versus assisted protein folding
473
of the enzyme in every organism and subcellular compartment; (ii) PPI has been shown to be involved in mitochondrial protein folding in cooperation with molecular chaperones (Rassow et al., 1995; Matouschek et al., 1995); (iii) in the case of the maturation of procollagen to form the mature collagen triple helix, PPI accelerates the rate by a factor of approximately three, whereas PPI inhibition leads to retardation (Bruckner et al., 1981; Bächinger, 1987). In transferring this kind of experiment into intact cells, it was shown that the time for half-completion of the triple helix in chicken embryo fibroblasts increased significantly upon PPI inhibition; the obvious explanation is that the ratelimiting prolyland hydroxyprolyl isomerizations during the in vivo folding of collagen are catalyzed by PPI. A PPIase associated with E. coli ribosomes has recently been identified as the trigger factor, a protein originally suggested to trigger secretory proteins to adopt a translocation-competent conformation, but recently found to interact with nascent chains of non-secretory as well as secretory proteins (see chapters by Welch et al. and Fischer and Schmid, this volume). Whether its main cellular function is to catalyze proline isomerization or to prevent premature folding and aggregation of newly synthesized polypeptides, remains to be determined. The given results suggest the following folding mechanism: The nascent all-trans polypeptide chain collapses rapidly to form a compact state with elements of secondary structure and sufficient stability to expel water from its hydrophobic interior; subsequent slow steps lead to the native-like state which finally undergoes trans→cis isomerization of proline residues to reach the biologically active, native conformation. Assuming this mechanism to hold in the cell, extrapolation from model peptides would predict protein folding to be much slower than estimated from in vivo studies. There are three explanations why proline isomerization does not necessarily limit in vivo protein folding: (i), trans prolines may be trapped in their initial configuration; (ii), the occurrence of proline residues in solvent-exposed turns causes their state of isomerization to be of minor functional significance, and (iii), constraints of the chain conformation may greatly decrease the energy barrier of proline isomerization, thus speeding up the reaction. These arguments do not consider the idea that PPIs may operate at the site of protein translation, thus enhancing the trans→cis isomerization at defined sites depending on the sequence specificity of the enzyme. Finally, it is worth mentioning that PDI and PPI catalysis occur simultaneously, in a specific and synergistic way (Schönbrunner & Schmid, 1992). Whether a similar synergism operates in the course of the de novo synthesis and folding in the cell remains to be shown. 6.3. Chaperones As has been shown in connection with side reactions competing with folding and association in vitro, the limit at which kinetic partitioning between folding and association becomes significant is commonly far below the average protein concentrations in vivo (Mitraki & King, 1989). Thus, in the cell, mechanisms must be effective that inhibit the unproductive aggregation and, at the same time, promote correct protein folding and association. Molecular chaperones, the components which serve this function, are a large family of unrelated proteins; they block the above side reaction, without being components of the final structures (Ellis & van der Vies, 1991). The
Molecular chaperones and folding catalysts
474
requirement for “folding helpers” differs, depending on the system; in general, multidomain oligomeric proteins seem to need chaperone assistance for unperturbed assembly. Apart from their involvement in folding, molecular chaperones also limit damage caused by stress conditions such as heat. Accordingly, their cellular level was found to be strongly enhanced under shock conditions; however, not all chaperones are heat shock proteins. It has become clear that molecular chaperones play significant roles in a variety of cellular processes such as protein targeting, translocation through membranes and compartmentation. Since the activities and mechanisms of action of molecular chaperones are subject of many chapters of this volume we will not further discuss their roles in protein folding here. 7. PRACTICAL ASPECTS As has been mentioned, aggregation in vitro and inclusion body formation in the cell correspond to each other. There are various strategies to cope with the problem. Making use of weaker promoters, the local concentration may be reduced. Discontinuous pulse dilution has been devised in order to optimize the yield of soluble protein: a certain amount of the denatured protein is subjected to dilution and reactivated at low concentration; after refolding, new portions of the denatured protein are added in a stepwise fashion until the whole batch is transferred. This way, the concentration of folding intermediates is always kept below the critical concentration of aggregation. Additives such as arginine may strongly increase the yield by shuffling aggregates back on the productive folding path. The repertoire of reconstitution methods has been extended by systematic experiments, applying folding catalysts and chaperone proteins (Rudolph et al., 1997). One might assume that overexpressing a specific protein together with chaperones and folding catalysts would finally yield 100% of the desired protein in its native state. Attempts toward this goal have so far been unsuccessful. Since sequestration of the folding polypeptide chain seems to be the underlying principle of chaperone action, renaturation in reverse micelles and immobilization on solid matrices might be promising alternatives. Successful developments in this direction have recently been reported (Jaenicke, 1995; Stempfer et al., 1996; Tuena de Gómez-Puyou and Gómez-Puyou, 1998). 8. RECENT DEVELOPMENTS AND CONCLUSIONS Fifty years after the protein folding problem was first formulated, there is still no way to predict the spatial structure of a given polypeptide chain from its sequence, nor has any protein been fully described regarding its detailed folding pathway. There are various reasons for that, apart from the insufficient size of the data set, and extrinsic factors not encoded in the primary structure. Practical applications of denaturation/renaturation placed high priority on the cellular aspects of folding. Anfinsen anticipated them at both levels, folding catalysis and folding on a template, i.e., chaperone-assisted folding (Epstein et al., 1963). However, before approaching the problem at the complex cellular
Spontaneous versus assisted protein folding
475
level, the problem had to be solved under ideal conditions, varying the relevant parameters in dilute solution. The outcome has been the mechanistic concept of multiplepathway sequential folding with very fast early events (hydrophobic collapse), middle events (local shuffling toward the native tertiary structure), and late events determining the rate of the overall reaction; the latter are proline cis-trans isomerization, disulfide cross-bridge formation and subunit assembly. As has been shown, all three are facilitated by accessory proteins, either enzymes or chaperones. From their ubiquity and abundance in all organisms, cells and cell-compartments, the two isomerases may be assumed to be essential for the formation of the native structure of proteins in the cell. In the case of the chaperones, the fundamental importance (beyond the stress response) has been clearly shown, (i) for the kinetic partioning between folding and association on the one hand, and aggregation on the other, (ii) for protein targeting, e.g., translocation of mitochondrial proteins across membranes, and (iii) for processes involved in morphogenesis, such as growth and self-assembly of phages. However, the mechanistic details of their action are still not fully understood. For a brief review on the present state of the art, see Beissinger and Buchner (1998). 9. REFERENCES Anfinsen, C.B. and Scheraga, H.A. (1975). Adv. Protein Chem. , 29 , 205–300. Bächinger, H.-P. (1987). J. Biol Chem. , 262 , 17144–17148. Bader, M., Muse, W., Zander, T. and Bardwell, J.C.A. (1998). J. Biol. Chem. , in press. Baker, D. and Agard, D.A. (1994). Biochemistry , 33 , 7505–7509. Balbach, J., Forge, V., van Nuland, N.A., Winder, S.L., Hore, P.J., and Dobson, C.M. (1995). Nature Struct. Biol. , 2 , 865–870. Baldwin, R.L. (1991). Chemtracts: Biochem. Mol Biol . 2 , 379–390. Baldwin, R.L. (1995). J. Biomol. NMR 5 , 103–109. Baldwin, R.L. and Eisenberg, D. (1987). In: Protein Engineering (Oxender, D.L. and Fox, C.F., eds.). Alan R.Liss, Inc., New York, pp. 127–148. Bardwell, J.C.A. (1994). Mol. Microbiol . 14 , 199–205. Beissinger, M. and Buchner, J. (1998). Biol Chem. , 379 , 245–259. Beissinger, M. Lee, S.C., Steinbacher, S., Reinemer, P., Huber, R., Yu, M.-H. and Seckler, R. (1995). J. Mol. Biol . 249 , 185–194. Bennett M.J., Schlunegger, M.P. and Eisenberg, D. (1995). Protein Sci. , 4 , 2455–2468. Bergman, L.W. and Kuehl, W.M. (1979b). J. Biol. Chem . 254 , 8869–8876. Brandts, J.F., Halvorson, H.R. and Brennan, M. (1975). Biochemistry , 14 , 4953–4963 Briggs, M. and Roder, H. (1992). Proc. Natl Acad. Sci. USA , 89 , 2017–2021. Bruckner, P., Eikenberry, E.F. and Prockop, D.J. (1981). Eur. J. Biochem . 118 , 607– 613. Buchner, J. and Rudolph, R. (1991). Bio/Technology 9 , 157–162. Creighton, T.E. (1992). In: Protein Folding (Creighton, T.E., Ed.), W.H.Freeman, New York, pp. 301–351. Creighton, T.E., Bagley, C.J., Cooper, L., Darby, N.J., Freedman, R.B., Kemmink, J. and Sheikh, A. (1993). J. Mol. Biol . 232 , 1176–1196. Danner, M. and Seckler, R. (1993). Protein Science , 2 , 1869–1881.
Molecular chaperones and folding catalysts
476
Danner, M., Fuchs, A., Miller, S. and Seckler, R. (1993). Eur. J. Biochem. , 215 , 653– 661. Dill, K.A. (1990). Biochemistry 29 , 7133–7155. Eigenbrot, C., Randal, M., and Kossiakoff, A.A., (1990). Protein Eng. , 3 , 591–598. Ellis, R.J. and van der Vies, S.M. (1991). Annu. Rev. Biochem. , 60 , 321–347. Epstein, C.J., Goldberger, R.F. and Anfinsen, C.B. (1963). Cold Spring Harber Symp. Quant. Biol 28 , 439–449. Evans, P.A. and Radford, S.E. (1994). Curr. Opin. Struct. Biol. , 4 , 100–106. Fersht, A.R. (1995a). Philos. Trans. R. Soc. (London). 348 , 11–15. Fersht, A.R. (1995b). Proc. Natl. Acad. Sci. USA , 92 , 10869–10873. Fischer, G. (1994). Angew. Chem. Int. Ed. Engl. , 33 , 1415–1436. Frech C. and Schmid, F.X. (1995). J. Mol. Biol. , 251 , 135–49. Fuchs, A., Seiderer, C. and Seckler, R. (1991). Biochemistry , 30 , 6598–6604. Garel, J.-R (1992). In: Protein Folding (Creighton, T.E., Ed.). W.H.Freeman New York, pp. 405–454. Gerl, M. Rudolph, R. and Jaenicke, R. (1985). Biol. Chem. Hoppe-Seyler , 366 , 447–454. Goldberg, M.E. and Guillou, Y. (1994). Protein Sci . 3 , 883–887. Goldberg, M.E. and Zetina, C.R. (1980). In: Protein Folding (Jaenicke, R., Ed.). Elsevier/North Holland, Amsterdam, pp. 469–484. Goldberg, M.E., Rudolph, R. and Jaenicke, R. (1991). Biochemistry , 30 , 2790–2797. Goldenberg, D.P. and King, J. (1982). Proc. Natl. Acad. Sci. USA , 79 , 3403–3407. Harrison, S.C. and Durbin, R. (1985). Proc. Natl. Acad. Sci. USA , 82 , 4028–4030. Helenius, A., Marquardt, T. and Braakman, I. (1992). Trends Cell Biol. , 2 , 227–231. Herbst, R., Schaefer, U. and Seckler, R. (1997). J. Biol. Chem. , 272 , 7099–7105. Huber, R. (1988). Angew. Chem. 100, 79–89; Angew. Chem. Int. Ed. Engl. , 27 , 79–88. Hurtley, S.M. and Helenius, A. (1989). Annu. Rev. Cell Biol , 5 , 277–307. Jaenicke, R. (1987). Progr. Biophys. Mol. Biol. , 49 , 127–237. Jaenicke, R. (1991). Eur. J. Biochem . 202 , 715–728. Jaenicke, R. (1993). Philos. Trans. R. Soc. (London) , 339 , 287–295. Jaenicke, R. (1994). Naturwissenschaften 81 , 423–429. Jaenicke, R. (1995). Philos. Trans. R. Soc. (London) , 348 , 97–105. Jaenicke, R. (1996). Curr. Topics Cell. Reg . 34 , 209–314. Jaenicke, R., Schurig, H., Beaucamp, N. and Ostendorp, R. (1996). Adv. Protein Chem. , 48 , 181–269. Jones, C.M., Henry, E.R., Hu, Y., Chan, C.-K., Luck, S.D., Bhuyan, A., Roder, H., Hofrichter, J. and Eaton, W.A. (1993). Proc. Natl. Acad. Sci. USA , 90 , 11860–11864. Kiefhaber, T. and Baldwin, R.L. (1995). Proc. Natl. Acad. Sci. USA , 92 , 2657–2661 Kiefhaber, T., Rudolph, R., Kohler, H.-H., and Buchner, J. (1991). Bio/Technol. , 9 , 825–829. Kraulis, P. (1991). J. Appl. Crystallogr. , 24 , 946–950. Kundrot, C.E. and Richards, F.M. (1987). J. Mol. Biol . 193 , 157–170. Lauffer, M.A. (1975). Entropy-driven Processes, Springer Verlag, Berlin, 264p. Levinthal, C. (1968). J. Chim. Phys. , 65 , 44–45. Lilie, H., Lang, K., Rudolph, R. and Buchner, J. (1993). Protein Science , 2 , 1490–1496. Matouschek, A., Rospert, S., Schmid, K., Glick, B.S. and Schatz, G. (1995). Proc. Natl.
Spontaneous versus assisted protein folding
477
Acad. Sci. USA , 92 , 6319–6323. Matthews, B.W. (1995). Adv. Prot. Chem. , 46 , 249–278. Mayr, E.-M., Jaenicke, R. and Glockshuber, R. (1994). J. Mol. Biol. , 235 , 84–88. Merrifield, R.B. (1993). In: Profiles, Pathways and Dreams (J.I.Seeman, ed.), ACS, Washington D.C. Mirankar, A., Radford, S.E., Karplus, M. and Dobson, C.M. (1991). Nature , 349 , 633– 636. Mirny, L.A., Abkevich, V. and Shakhnovich, E.I. (1996). Folding and Design 1 , 103– 116. Mitraki, A. and King, J. (1989). Bio/Technology , 7 , 690–697. Mitraki, A. and King, J. (1992). FEBS Lett. , 307 , 20–25. Müller, K. and Jaenicke, R. (1980). Z. Naturforsch. C . 35 , 222–228. Opitz, U., Rudolph, R., Ericsson, L. and Neurath, H. (1987). Biochemistry 26 , 1399– 1406. Peters, T., Jr. and Davidson, L.K. (1982). J. Biol Chem. , 257 , 8847–8853. Privalov, G. (1994). PhD Thesis, The Johns Hopkins University, Baltimore. Privalov, P.L. (1992). In: Protein Folding (Creighton, T.E., Ed.). W.H.Freeman New York, pp. 83–126. Ptitsyn, O.B. (1995). Adv. Protein Chem. , 47 , 83–229. Rassow, J. and Pfanner, N. (1996). Curr. Biol. , 6 , 115–118. Rassow, J., Mohrs, K., Koidl, S., Barthelmess, I.B., Pfanner, N. and Tropschug, M. (1995). Mol. Cell. Biol . 15 , 2654–2662. Richards, F.M. (1992). In: Protein Folding (Creighton, T.E., Ed.). W.H.Freeman New York, pp. 1–58. Rinas, U., Risse, B., Jaenicke, R., Abel, K.J. and Zettlmeissl, G. (1990). Biol Chem. Hoppe-Seyler. , 371 , 49–56 Roder, H. and Elöve, G.A. (1994). In: Mechanisms of Protein Folding (Pain, R.H., Ed.). Oxford Univ. Press, New York, pp. 26–54. Rudolph, R. (1990). In: Tschesche, H. (Ed.). Modern methods in protein and nucleic acid research. de Gruyter, Berlin, New York, pp. 149–171. Rudolph, R., Böhm, G., Lilie, H. and Jaenicke, R. (1997). In: Protein Function, a Practical Approach (Creighton, T.E., ed.). IRL Press, Oxford, UK., pp. 57–99. Schlunegger, M.P., Bennett, M.J. and Eisenberg, D. (1997). Adv. Protein Chem. , 50 , 61–122. Schmid, F.X., Mayr, L., Mücke, M. and Schönbrunner, E.R. (1993). Adv. Protein Chem. , 44 , 25–66. Schönbrunner, E.R. and Schmid, F.X. (1992). Proc. Natl. Acad. Sci. USA 89 , 4510– 4513. Schumann, J. and Jaenicke, R. (1993). Eur. J. Biochem . 213 , 1225–1233. Sinclair, J.F., Ziegler, M.M. and Baldwin, T.O. (1994). Nat. Struct. Biol. , 1 , 320–326. Sosnick, T.R., Mayne, L., Hiller, R. and Englander, S.W. (1994). Struct. Biol. , 1 , 149– 156. Steinbacher, S., Seckler, R., Miller, S., Steipe, B., Huber, R. and Reinemer, P. (1994). Science , 265 , 383–386. Stempfer, G., Höll-Neugebauer, B., and Rudolph, R. (1996). Nature Biotechnol , 14 ,
Molecular chaperones and folding catalysts
478
329–334. Trinkl, S., Glockshuber, R. and Jaenicke, R. (1994). Protein Science , 3 , 1392–1400. Tuena de Gómez-Puyou, M. and Gómez-Puyou, A. (1998). Crit. Rev. Biochem. Mol. Biol , 33 , 53–89. Udgaonkar, J.B. and Baldwin, R.L. (1990). Proc. Natl Acad. Sci. USA 87 , 8197–8201 Valax, P. and Georgiou, G. (1993). Biotechnol Prog. , 9 , 539–547. van Mierlo, C.P.M., Kemmink, J., Nauhaus, D., Darby, N.J. and Creighton, T.E. (1994). J. Mol Biol 235 , 1044–1061. Varley, P., Gronenborn, A.M., Christensen, H., Wingfield, P.T., Pain, R.H. and Clore, G.M. (1993). Science , 260 , 1110–1113. Weissman, J.S. and Kim, P.S. (1991). Science , 253 , 1386–1393. Weissman, J.S. and Kim, P.S. (1992). Proc. Natl Acad. Sci. USA , 89 , 8527–8531 West, S.M. and Price, N.C. (1988). Biochem. J. , 251 , 135–139. Wetlaufer, D.B. (1981). Adv. Protein Chem. , 34 , 61–92. Winther, J.R., Sörensen, P. and Kielland-Brandt, M.C. (1994). J. Biol Chem. , 269 , 1–7. Zettlmeissl, G., Rudolph, R. and Jaenicke, R. (1979). Biochemistry , 18 , 5567–5571.
20. PROTEIN DISULPHIDE-ISOMERASE: A CATALYST OF THIOL: DISULPHIDE INTERCHANGE AND ASSOCIATED PROTEIN FOLDING ROBERT B.FREEDMAN* and PETER KLAPPA Research School of Biosciences, University of Kent, Canterbury, Kent CT2 7NJ, UK
1. Introduction 2. Molecular Properties of PDI 2.1. Analysis of PDI Protein Sequences 2.2. Evidence for the Domain Organization of Mammalian PDI 2.3. PDI as a Member of the Thioredoxin Superfamily of Proteins 2.4. Chemical Properties of the Thioredoxin-like Domains within PDI 3. PDI as a Catalyst of Thiol:disulphide Interchange and Protein Folding 3.1. Activity Towards Simple Peptide Substrates 3.2. Activity Towards Well-Defined Protein Substrates 3.3. Activities of Mutant PDIs and Domains 3.4. Does PDI Catalyse Folding or Just the Disulphide-Isomerizations Associated with Folding? 4. Binding Properties of PDI 4.1. Interaction of PDI with Non-Peptides 4.1.1. Interaction of PDI with Thyroid Hormone 4.1.2. Interaction of PDI with Oestrogens 4.2. Interaction of PDI with Peptides and Proteins 4.2.1. Interaction of PDI with Peptides 4.2.2. Interaction of PDI with Proteins 4.2.3. Interaction of PDI with ER Proteins 5. Is PDI a Molecular Chaperone? 6. References *Corresponding author
Molecular chaperones and folding catalysts
480
1. INTRODUCTION Protein disulphide-isomerase (PDI), an enzyme found in the endoplasmic reticulum of eukaryotic cells, was the first catalyst of protein folding to be identified. In their initial studies on protein refolding in vitro, Anfinsen and colleagues recognised that the conditions required for successful refolding of reduced ribonuclease (high dilution, high pH etc.) were not physiological, and that the time-scales involved (hours to days) were much longer than those for protein folding in the cell (Epstein et al., 1963). They inferred that the process was catalysed in the cell and sought potential cellular catalysts. In the mid 1960s, they and other groups showed that refolding of reduced ribonuclease and other proteins could be catalysed by microsomal preparations from vertebrate secretory tissues such as liver, pancreas etc. (Goldberger et al., 1963), purified the enzyme responsible (DeLorenzo et al., 1966) and proposed a role for it in protein folding in the cell (Givol et al., 1964). The catalytic properties of the purified enzyme from mammalian liver were described over the next 10 years, and were consistent with its being involved in the catalysis of native disulphide bond formation during the folding of nascent secretory proteins. It was 20 years after this initial work on PDI before it was generally accepted that protein folding in the cell is a process mediated and facilitated by cellular factors. By that time, the evidence for the involvement of PDI in cellular catalysis of protein folding was stronger than before, but still circumstantial (Freedman, 1984). PDI was known to be most abundant in cells actively synthesizing and secreting disulphide-bonded proteins and to be located in the lumen of the endoplasmic reticulum, the compartment in which such proteins fold immediately after synthesis (Lambert and Freedman, 1985). Furthermore, in several physiological situations, the pattern of expression of PDI matched the pattern of expression of the disulphide-bonded proteins which are its potential targets (Freedman, 1984). Within the last 10 years, the role of PDI in cellular protein folding has been established beyond doubt by the classical methods of biochemistry, cell biology and molecular genetics. In whole cells, and in cell-free systems for in vitro translation and translocation, PDI can be cross-linked to nascent and newly translocated secretory proteins (Roth and Pierce, 1987; Klappa et al., 1995). In such cell-free systems, the presence of PDI is required for the rapid formation of native disulphide bonds; removal of lumenal resident proteins leads to defective folding which is repaired by reconstitution with purified PDI (Bulleid and Freedman, 1988). Finally, the gene encoding PDI in S. cerevisiae has been cloned and shown to be required for the proper folding, targeting and secretion of disulphide-bonded proteins (LaMantia and Lennarz, 1993; Dunn et al., 1995). More recently, this cellular role for PDI has been demonstrated by a further stringent test which also has practical application. In both eukaryotic and prokaryotic systems it has been shown that overexpression of recombinant PDI (in the ER or periplasm, respectively) can significantly increase production of functional recombinant proteins containing multiple disulphide bonds (Wittrup, 1995; Humphreys et al., 1995). In other words, there are cases where the yield of a correctly folded recombinant protein is limited by inefficient formation of the native disulphide bonds, and this limitation can be
Protein disulphide-isomerase
481
overcome by co-overexpression of PDI. This chapter reviews the structural, catalytic and ligand binding properties of PDI. It considers, in particular, recent work aimed at establishing the structural basis of how PDI acts as a catalyst of protein folding. 2. MOLECULAR PROPERTIES OF PDI 2.1. Analysis of PDI Protein Sequences The cloning and sequencing of rat PDI cDNA (Edman et al., 1985) immediately led to the recognition of internal sequence homologies within the protein and the
Figure 1 Domain structure of PDI. The border between the a' and c domain is not exactly known.
suggestion of a structural organization based on duplicated sequence modules (see Figure 1). The homologous a and a' sequence modules show significant sequence identity to thioredoxin, and on the basis of this homology, PDI was recognised as a member of a superfamily of proteins containing thioredoxin-like sequences (see 2.3 below). Since the internal sequence repetition does not include the whole of PDI, a more detailed analysis was undertaken (Freedman et al., 1998) in an attempt to define more precisely the sequence modules within PDI and their boundaries. This analysis drew on i) the pattern of sequence conservation in a large set of multiply-aligned PDI sequences from a range of species, ii) the pattern of hydrophobicity/ hydrophilicity in this sequence set, iii) the intron/exon structure of the human PDI gene, and iv) the claim that part of the PDI protein sequence was homologous to part of a steroid receptor sequence. The results of this analysis suggested the existence of up to 6 structural domains (Freedman et al., 1994), but did not define their boundaries unambiguously; the analysis provided the basis for experiments aimed at determining the domain organization of PDI (see 2.2 below), on the assumption that the sequence modules defined within PDI would correspond to structural domains. 2.2. Evidence for the Domain Organization of Mammalian PDI As yet, no high resolution structure has been determined for PDI from any source. Although crystals have been obtained by some groups over the years, they have not yet yielded high resolution x-ray diffraction data. In view of its molecular size—a homodimer of polypeptides each of over 55 kD—PDI remains beyond the range of structure determination by current NMR methods. In the absence of a high resolution
Molecular chaperones and folding catalysts
482
structure, even quite low resolution information would be helpful. Is PDI organized into clear structural domains? If so, how many are there and where are their boundaries? Can they be obtained as independent folded structures and subjected to high resolution structure determination, so that the overall structure of PDI can be built up domain-bydomain? In an attempt to answer these questions two complementary methods have been used recently by the laboratory of Creighton and by us. First, native purified PDI has been subjected to limited proteolysis, the fragments identified and the data interpreted on the argument that the preferential sites for proteolysis are likely to lie at the boundaries between structural domains. Secondly, putative domains inferred from sequence analysis (see 2.1 above), have been expressed as recombinant polypeptides and their structural properties characterised; the underlying assumption is that a fragment which adopts a stable conformation with secondary and tertiary structure, and cooperative unfolding and refolding behaviour, corresponds to a domain since it behaves as an independent folding unit. Limited proteolysis of bovine PDI by trypsin and by V8 protease indicated a number of preferential sites of proteolysis (Freedman et al., 1998). N-terminal sequencing of derived fragments demonstrated cleavage following R115, R328, E331 and K333 and mass determinations of fragments by mass spectrometry suggested that proteolysis also occurred at E222 and K349. The region from R328 to K349 is clearly highly susceptible to proteolysis and we propose that it is a disordered or exposed region linking the b′ and a′ domains. The proteolysis at E222 probably marks a boundary region between the b and b′ domains. No cleavage sites were found within the region putatively linking the a′ and c domains, suggesting that the acidic C-terminal region of PDI does not form a distinct structural domain. Similarly no sites of cleavage were found at either end of the putative e domain (despite the presence there of residues which are compatible with the cleavage specificity of the enzymes used). Instead, there was preferential proteolysis at R115, in the middle of the putative e domain, and fragments cleaved at this site were stable and could be recovered. This suggests that there is no e domain, and that the cleavage at R115 marks a boundary region between the a and b domains. The conclusions from proteolysis are supported by work on the recombinant expression of putative domains (Freedman et al., 1998). We have expressed each of the putative e, b and b′ domains as fusions with glutathione-S-transferase (GST), recovered the products and released the putative domain from the fusion partner. In parallel, we have generated the putative ‘linked domains’ e-b (101–226) and e-b-b′ (101–335) in the same way. None of the released fragments corresponding to putative individual domains showed significant secondary structure in solution or co-operative conformational change as a function of denaturant concentration. In contrast, after cleavage from the N-terminal fusion partner, both the ‘linked domain’ polypeptides had the properties of folded proteins. Interestingly, in both cases the N-terminal sequence of the polypeptide was not as expected, based on the junction with the fusion partner, but showed that the Nterminus had been trimmed to remove 2 linking residues and part of the putative e domain, leaving the N-terminal sequence corresponding to residues 115–117 in the mature sequence. This result implies that the residues prior to 115–117 do not form part of the folded domain constituted by the remainder of the fragment, and is consistent with
Protein disulphide-isomerase
483
the conclusions from proteolysis of authentic full-length PDI. A similar approach has been adopted, independently, by Creighton and colleagues (Darby et al., 1996). They found no evidence of structure in recombinant polypeptides corresponding to the e (100–157) or b (148–257) domains alone. They then expressed a polypeptide corresponding to the hypothetical linked e-b domains (residues 100–257) and deliberately subjected it to limited proteolysis by trypsin, V8 protease and thermolysin. Each of these enzymes trimmed the recombinant fragment substantially at both ends; analysis of the products suggests that a core b domain comprises residues 116–218 and the individual proteolysis products extend beyond this, based on the availability of sites consistent with the specificity of the proteases. On the basis of this analysis, a new putative b domain polypeptide was designed and expressed (residues 119–228) which in contrast to earlier constructs (100–157 and 148–257) showed CD, NMR and urea gradient gel electrophoresis properties consistent with being a folded polypeptide. Similarly, recombinant polypeptides corresponding to putative a and a' domains have been expressed and characterised. The results suggest that residues 1–120 and 348–462 both form folded structures when expressed in isolation (Darby and Creighton, 1995b), as do the shorter fragments 5–96 and 347–436 (Freedman et al., 1998). These recombinant polypeptides show some of the functional properties of PDI (see 3.4 below) and hence have been the target for more detailed structural studies. In summary, the work reviewed here reveals a 4-domain structure (a-b-b'-a'), with the inter-domain boundaries falling around residues 116, 220 and 335–350 respectively, plus an ill-defined region extending from residue 462 to the C-terminus. 2.3. PDI as a Member of the Thioredoxin Superfamily of Proteins As noted, sequence similarity between PDI and the small ubiquitous protein thioredoxin immediately suggested a family relationship between them. Thioredoxin is a cytosolic protein which functions as a dithiol reductant in a range of cellular functions. Over time, a substantial number of proteins have been discovered which have sequence and functional similarities to thioredoxin and are recognized as members of a thioredoxin superfamily (Table 1). A larger number of cDNA sequences or genomic open reading frames also are clearly related to this family, even though the corresponding protein has not been isolated or characterised. Some members of the superfamily (thioredoxins, glutaredoxins) are small and of known tertiary structure; they comprise a single domain with a characteristic fold which is also found in a number of enzymes involved in glutathione or sulphur metabolism (Martin, 1995). DsbA, a bacterial periplasmic protein which acts as a disulphide oxidant (see Missiakas and Raina, this volume), also has this fold, but is larger and contains an additional helical domain inserted within the thioredoxin fold (Martin et al., 1993). The structural data on the N-terminal 120 residues of PDI, corresponding to the a domain, confirm that it is a structural domain and has
Molecular chaperones and folding catalysts
484
Table 1 The thioredoxin/PDI superfamily
Protein
Mr
Structure
Sequence of active site motif
Thioredoxin
11kDa
thioredoxin domain
-CGPC-
Glutaredoxin 12kDa
thioredoxin domain
-CP(Y/F)C-
DsbA
22kDa
thioredoxin domain+helical insert
-CPHC
PDI
55kDa
2 thioredoxin domains+others
-CGHC-
PDI-like ER proteins
50–75kDa 2 or 3 thioredoxin domains+others
-CGHC-CGTC- (PDIp)
the classic thioredoxin fold; residues 1–3 and 117–120 were not fixed in the native state of this fragment confirming that residues 116–117 mark the limit of the a domain (Kemmink et al., 1995, 1996). More recent studies of the global fold of the b domain of PDI by NMR reveal that, like the a domain, the b domain shows the secondary structure pattern characteristic of the thioredoxin fold, despite showing no significant amino-acid sequence similarities to any members of the thioredoxin family. This observation, together with modelling and preliminary experimental studies on the a' and b' regions strongly implies that PDI comprises four sequential domains each with the thioredoxin fold (Kemmink et al., 1997). Interestingly, many cDNA sequences have been cloned from a wide variety of sources showing the existence of gene products which can be inferred to contain two (or three) thioredoxin-like domains. In some mammalian species, four or more of such sequences co-exist (Freedman et al., 1994). The significance of this multiplicity has not been established, but it is natural to assume that it relates either to differences in specificity, or to differently regulated patterns of expression, or to different cellular or sub-cellular location. For example, one member of the family (PDIp) has recently been claimed to be expressed specifically in the acinar cells of the pancreas (DeSilva et al., 1996, 1997), whereas another, ERp57, has been found to interact exclusively with glycosylated secretory proteins after their translocation into the ER (Oliver et al., 1997). It has been shown that this specificity is due to the interaction between ERp57 and calnexin, a glycoprotein-specific chaperone of the ER (Zapun et al., 1998). 2.4. Chemical Properties of the Thioredoxin-like Domains within PDI Early chemical modification studies with alkylating reagents, and inhibition studies by arsenicals and heavy metals, suggested that PDI contained an essential dithiol group which could be oxidised to a disulphide form in the isolated protein (Hillson and Freedman, 1980). Now it is clear that both the a and a' domains of PDI contain such a group with the local sequence…WCGHCK…and that similar vicinal dithiol/ disulphide groups are characteristic of the members of the thioredoxin superfamily. The structural properties of this active site dithiol/disulphide group are known in thioredoxin and DsbA,
Protein disulphide-isomerase
485
whose structures have been determined to high resolution by X-ray diffraction (Martin, 1995). The motif is found in an exposed turn linking a strand to an extended helix. The sulphur atom of the more N-terminal Cys residue is at the N-terminal pole of the a helix and is exposed at the surface of the molecule, while the sulphur atom of the more Cterminal Cys residue is buried behind it. The chemical properties of this group have been most thoroughly studied in human PDI and in thioredoxin and DsbA from E. coli (see Missiakas and Raina, this volume). The following features are characteristic both of DsbA and the a domain of PDI (Zapun et al., 1993; Nelson and Creighton, 1994; Darby and Creighton, 1995c): (i) the more N-terminal Cys residue of the dithiol group is unusually low in pK and is present exclusively as the -S- thiolate anion at neutral pH, (ii) this group is unusually reactive as a nucleophile both in alkylation reactions and in the formation of mixed disulphides by reaction with, for example, oxidised glutathione (GSSG),
(iii) mixed disulphides formed by the latter reaction are unusually unstable since they are prone to attack by the more C-terminal Cys residue of the active site, leading to displacement of reduced glutathione and formation of the disulphide form of the protein
(iv) the equilibrium constant for reaction between the dithiol/disulphide group of the protein and reduced and oxidised glutathione (GSH/GSSG), via the two partial reactions above, is in the range 1 mM or less, corresponding to a standard redox potential in the range –90 to –180 mV which is unusually high for a protein, indicating that the active site disulphide is unusually oxidising (i.e. weak, unstable). This can be contrasted with thioredoxin where the active site dithiol/disulphide is highly reducing and the equilibrium constant with GSH/GSSG is 10 M. (v) the active site disulphide in DsbA and PDI is so unstable that, unlike in most disulphide bonds in proteins, it actually destabilizes the protein, so that the reduced form of the protein is the more stable to denaturation by urea etc. These chemical properties are clearly interrelated and the fundamental property appears to be the unusually low pK of the exposed thiol group of the N-terminal Cys residue. The structural basis of this can be analysed by comparing the high resolution structures of DsbA and thioredoxin (Martin, 1995); regrettably, a comparable high resolution structure of PDI is not yet available. The steric properties of the active site disulphides in the two proteins are indistinguishable; there are no obvious indications of strain to account for the weakness of the disulphide in DsbA. The low pK must reflect the existence of electrostatic interactions which stabilize the thiolate form of the exposed Cys residue. In both DsbA and thioredoxin, the S atom of the N-terminal Cys is at the N-terminus of an
Molecular chaperones and folding catalysts
486
helix and could therefore interact electrostatically with the directed H-bonds along the axis of the helix, which generate an effective ‘helix dipole’. The positive pole of this dipole is at the N-terminus and hence can stabilize an anionic -S- group there, but this applies, in general terms, to both thioredoxin and DsbA and cannot simply account for the difference between them. A major difference occurs between the proteins in the residues within the active site motif (see Table 1); mutation studies in thioredoxin, DsbA and the recombinant a domain of PDI (Lundstrom et al., 1992; Grauschopf et al., 1995; Kortemme et al., 1996) demonstrate that the presence of a His residue at position…CxHC…(as in wild-type DsbA and PDI) has a major effect on the properties of the site, and other mutations in these residues also affect the chemistry and functional properties of the protein (Grauschopf et al., 1995). However, it is unlikely that this is the only source of difference (Warwicker and Gane, 1996). Our electrostatic analysis identified two other factors (Gane et al., 1995): i) the geometry of the helix following the active site is different between thioredoxin and DsbA, leading to a stronger electrostatic interaction with the exposed thiolate group in the latter case, and ii) the presence of the additional inserted domain in DsbA modifies the dielectric properties of the region close to the active site giving rise to stronger electrostatic interactions. It remains to be seen whether similar additional factors exist in PDI to account for the chemical properties of its active site disulphides, and indeed whether there are differences in such properties between the active sites of the a and a' domains. 3. PDI AS A CATALYST OF THIOL:DISULPHIDE INTERCHANGE AND PROTEIN FOLDING 3.1. Activity Towards Simple Peptide Substrates PDI was discovered on the basis of its ability to catalyse reactivation of reduced inactive ribonuclease and has been widely assayed using this and related reactions with protein substrates. However, the underlying processes here are complex, involving a mixture of protein disulphide formation, disulphide isomerization and conformational change. To clarify some aspects of PDI catalysis, it is useful to work with simpler and more defined catalysts. Darby et al. (1994) introduced a model 28-residue peptide containing 2 cysteine residues at positions 2 and 27, analysed thiol: disulphide interchange reactions between it and glutathione by HPLC, and characterised the catalysis of these reactions by PDI and DsbA (Darby et al., 1994; Darby and Creighton, 1995a). The model peptide has no defined secondary or tertiary structure in either the reduced dithiol form or after oxidation to form an intramolecular disulphide; as a result the reactions are ‘pure’ examples of thiol: disulphide interchange and oxidoreductions in a protein-like substrate without being associated with folding or extensive conformational change. In conditions of pH and glutathione concentrations comparable to those within the ER lumen, PDI catalysed (approximately 30-fold) reactions between the reduced peptide and oxidised GSSG leading to formation of both single mixed disulphide species (with glutathione bonded to residues 2 and 27 respectively) and it catalysed (approximately 100-fold) their subsequent disulphide isomerization to form an intramolecular peptide
Protein disulphide-isomerase
487
disulphide and release reduced GSH. There was little formation of the double mixed disulphide species in the conditions used, nor did PDI effectively catalyse transfer of the glutathione group between one cysteine residue of the peptide and the other. In addition to these reactions, PDI also catalysed the direct oxidation of the peptide to the disulphide state, in a process not directly involving glutathione, but requiring the GSSG-dependent recycling of the resultant reduced PDI. This latter mechanism obviously requires the involvement of the disulphide form of PDI, whereas the former mechanism can be represented as a series of thiol: disulphide interchange steps in which the exposed Cys residue interconverts between the thiol form, and mixed disulphides with glutathione and/ or with the peptide substrate, and the buried Cys residue of the active site is not directly involved. This model peptide system has been most instructive in analyzing basic features of the mechanism of action of both PDI and DsbA. However it is not convenient for routine assay. Recently Ruddock et al. (1996) introduced a related substrate, a small peptide containing two cysteine residues and a single tryptophan residue whose fluorescence emission is quenched when an intramolecular disulphide bond is formed. PDI and DsbA catalysis of glutathione-dependent oxidoreduction using this substrate are comparable to those with the model substrate of Darby et al. (1994) but the overall reaction can be monitored continuously by fluorimetry. 3.2. Activity Towards Well-Defined Protein Substrates PDI is active towards a continuously-growing list of protein substrates, but only a few have been studied in any detail. It was shown many years ago (Creighton et al., 1980) that PDI catalyses the net process of refolding of reduced bovine pancreatic trypsin inhibitor (BPTI) and specifically catalyses processes involving formation or isomerization of protein disulphide bonds together with associated conformational change. Since that time, the folding pathway of BPTI has been characterized in great detail so that the identities and structures of many intermediate species have been determined, and the kinetics and pathways of their interconversion have been established (Goldenberg, 1992; Weissmann and Kim, 1993; Creighton et al., 1995). Although there have been extensive studies with bovine pancreatic ribonuclease and other proteins, BPTI is by far the best-characterised protein system in which to study action of PDI on a protein substrate. In particular, the folding pathway of BPTI involves isomerizations between a number of species containing two of the three native disulphide bonds. The species containing the disulphides 30–51 and 5–55 (30–51, 5–55) is also referred to as N(SH)2 since it has almost complete native folding and rapidly forms the final native disulphide bond (14– 38). However the species N(SH)2 is not readily formed from species containing a single disulphide bond. The major species which accumulate under native conditions are (30– 51, 14–38) and, in some conditions (5–55, 14–38), both of which contain two native disulphide bonds and are extensively folded. These species (termed N' and N* respectively by Weissmann and Kim (1993)) require only the formation of one further native disulphide bond but they cannot do so directly or rapidly since to do so would require extensive unfolding and refolding. Instead they rearrange to form N(SH)2 via
Molecular chaperones and folding catalysts
488
species (30–51, 5–14) and (30–51, 5–38) and these latter species which contain nonnative disulphides (5–14 and 5–38, respectively) are obligatory intermediates in the folding pathway. These rearrangements provide a stringent test of the activity of PDI since they involve disulphide isomerization with associated conformational change, rather than net formation or breakage of protein disulphide bonds. They can be studied by isolating the intermediates N' and N* and, as suggested by the earlier work, they are strikingly (3500– 6000-fold) catalysed in the presence of excess PDI (Weissmann and Kim, 1993). The dramatic catalytic effect on these reactions found with PDI is not observed with DsbA (Darby et al., 1995a); this bacterial periplasmic relative of PDI appears to act primarily as a net oxidant, a donor of disulphide bonds, rather than as a catalyst of rearrangements. Another contrast is between the dramatic catalysis of the isomerizations between species containing two disulphides (N'→N(SH)2 and N*→N(SH)2), and the minimal catalysis by PDI (2-fold) of the further conversion of N(SH)2 to the native folded protein (Weissmann and Kim, 1993), a process which involves disulphide formation but limited conformational change. Other well-defined systems in which the catalytic activity of PDI can be studied include the oxidation of reduced pancreatic ribonuclease (Lyles and Gilbert, 1991) and the formation of native ribonuclease T1 from both the reduced state (with 4 free thiol groups) and the mixed disulphide state with four bound glutathione groups (Ruoppolo and Freedman, 1995; Ruoppolo et al., 1996). The standard assay for PDI with a protein substrate is its ability to catalyse the reactivation of ‘scrambled’ ribonuclease, a substrate in which the eight Cys residues of bovine ribonuclease have been oxidised essentially at random in denaturing conditions (Freedman et al., 1995). This reactivation clearly requires both disulphide isomerization and protein conformational change, but its details have not been established due to the heterogeneity of the substrate. 3.3. Activities of Mutant PDIs and Domains The possibility of designing and expressing mutant or truncated forms of PDI allows detailed insight into the structural basis of its activity. This is particularly significant in view of the presence of multiple domains and two apparently similar active sites in the a and a' domains respectively. Darby and Creighton (1995b) expressed fragments corresponding to residues 1–120 and 348–462 (the a and a' domains respectively) of human PDI and showed that the a domain comprised a well-folded protein with some disordered residues at each terminus (Kemmink et al., 1995). These constructs (Table 2) were relatively active in catalysing simple thiol: disulphide reactions with no protein folding component. Thus, in the glutathione-dependent conversion of the dithiol to disulphide form of the model 28residue peptide, each showed a specific activity approximately 50–60% that of native full-length PDI. Similarly in the conversion of reduced BPTI to the major two-disulphide intermediate (N'; 30–51, 14–38) each showed activity up to 70% that observed with the full-length enzyme. These results suggest that in these relatively simple reactions, the action of full-length PDI is simply the sum of roughly equivalent activities mediated by each of its a and a' domains.
Protein disulphide-isomerase
489
However, a different picture emerges in other assays. The isolated recombinant domains are completely inactive in catalysing the conversion of the N' form of BPTI to the native product, an intrinsically very slow process requiring disulphide isomerization and protein conformational change. Similarly, the recombinant domains only show approximately 10% of the activity of native PDI in catalysing the reactivation of ‘scrambled’ ribonuclease, another process dependent on both disulphide isomerizations and conformational change in the protein substrate. (A slightly lower Table 2 Relative activities of active site mutants of full length PDI and single domain constructs of PDI
Assay
Mutant
Activity (%)
Reference [Species]
sRNase/DTT (SH)
SGHC…CGHC
46
Vuori et al., (1992)
CGHC…SGHC
47
[Human]
SGHC…SGHC
2
CLHS…CGHC
60
LaMantia and Lennarz (1993)
CGHC…CIHS
60
[S. cerevisiae]
CLHS…CIHS
5
SGHC…SGHC
2
Laboissiere et al., (1995)
CGHS…CGHS
92
[Rat]
SGHC…SGHC
–
CGHS…CGHS
3
CGHS…CGHS
18–30
Lu et al., (1992)
SGHS…SGHS
–
Lyles and Gilbert (1994)
CGHC…SGHS
23–50
Walker et al., (1996)
CGHS…SGHS
3–8
[Rat]
SGHC…SGHS
–
SGHS…CGHC
23–27
SGHS…CGHS
9–18
SGHS…SGHC
–
rBPTI/DTT (SS)
sRNase/GSH/GSSG
rRNase/GSH/GSSG
rRNase/GSH/GSSG
Assay
a Domain a′ Domain Activity (%) Reference [Species]
sRNase/DTT (SH)
CGHC CGHC
14
Darby and Creighton
9
(1995b, c) [human]
rBPTI/GSSG/GSH to N′
CGHC
70 CGHC
70
Molecular chaperones and folding catalysts N′ to BPTI
CGHC
– CGHC
Peptide (SH)2 to Peptide (SS)
490
CGHC
– 62
CGHC
54
sRNase, ‘scrambled’ ribonuclease; rRNase, reduced ribonuclease; rBPTI, reduced BPTI; N′, native-like form of BPTI
figure of 3% was determined for this rearrangement activity by Freedman et al. (1998), using more extensively truncated forms of the a and a′ domains, residues 5–96 and 347– 436 respectively). These results strongly suggest that whereas PDI functions as the sum of independent a and a′ domains in simple reactions, its ability to catalyse complex disulphide isomerizations linked to protein conformational change either involves the other domains of the molecule, or depends on specific interactions or co-operation between the a and a′ domains in the intact enzyme (or both). Some light can be thrown on this question by studying mutant full-length PDI molecules in which one active site has been inactivated by mutation. Several groups have generated such mutations (see Table 2) and assayed the mutants in a variety of activities. The earliest work was that of Kivirikko’s group on mutant human PDIs (Vuori et al., 1992). They used the assay based on ‘scrambled’ ribonuclease and showed that the enzyme could be fully inactivated by mutating to serine the N-terminal Cys residues in the both active site motifs (i.e. in both the a and a′ domains of the full-length enzyme). When each of these mutations is made separately, leaving one wild-type active site and one mutant site, the resulting proteins are essentially 50% active in this assay. Thus a full-length protein with a single functional active site is 50% active in this rearrangement assay, whereas the individual active site domains, when present as isolated fragments are only 10% active (see paragraph above). This strongly implies a role for the other domains of the molecule in isomerization activity. The most extensive set of mutants is that of rat PDI generated by Gilbert and colleagues (Lyles and Gilbert, 1994; Walker et al., 1996) assayed in the reactivation of reduced ribonuclease. Mutants with one entirely wild-type active site (CGHC) and the other converted to SGHS are moderately active (up to 50% of the wild-type enzyme, consistent with the results of Vuori et al. (1992)), although detailed kinetic analysis suggests that the kinetic properties of the a and a′ active sites in the intact protein are distinct. In the other mutants, activity is only detected when the wild-type Cys residue is present in the N-terminal position of the active site motif in at least one of the active sites, confirming the crucial importance of the N-terminal Cys in each site. Interestingly, however, mutation of the more C-terminal Cys residue in each site has more limited effects; the mutant with this mutation in both active sites has 18–30% of the activity of the wild-type enzyme, suggesting that the enzyme can act entirely through thiol: disulphide interchange processes involving the exposed Cys residues only. In qualitative agreement with this result, Laboissiere et al. (1995) find that the double mutant with the sequence SGHC at both active sites is minimally active in rearranging ‘scrambled’ ribonuclease, whereas the double mutant with the sequence CGHS at both
Protein disulphide-isomerase
491
sites is almost fully active. In their hands, neither double mutant is active towards reduced ribonuclease, implying that activity towards this substrate requires net oxidation by the disulphide form of the PDI active site. The quantitative discrepancy between this result and the finding of Gilbert’s group may arise from differences in assay conditions and method, but has not been resolved. Similarly the result of LaMantia and Lennarz (1993) on the effect of mutation in the C-terminal Cys in each active site is not immediately consistent with the results of the other groups, but this work used yeast PDI and a different assay. Overall the fragment and mutant data point to two conclusions. 1) The full-length molecule is required for full activity in processes involving disulphide isomerization and conformational change, and full-length molecules with a single functional active site are 50% active in such assays. The individual isolated domains are much less active in these assays, but fully functional in simple thiol: disulphide interchange. This strongly suggests direct involvement of the other domains in the more complex activities. 2) Only the Nterminal Cys residue in each active-site motif is absolutely required for activity, although mutations in the more C-terminal Cys residue have some effect. This implies that the enzyme operates through a mixture of thiol: disulphide exchange and direct oxidation mechanisms, dependent on the substrate and conditions, and that species mutated in the C-terminal Cys are moderately effective in the former mechanism. 3.4. Does PDI Catalyse Folding or just the Disulphide-Isomerizations Associated with Folding? This question cannot be directly answered from the studies reviewed here. In the more complex activities studied to date (e.g. the interconversions of the 2-disulphide isomers of BPTI, and the reactivation of ‘scrambled’ ribonuclease) the disulphide isomerizations and conformational changes are intimately associated. However the fact that these reactions have more stringent requirements for catalysis than do processes simply involving thiol:disulphide chemistry, strongly implies that PDI is directly involved in facilitating conformational changes in its protein substrates. If so, it is difficult to imagine how this could be brought about except by direct contacts between PDI and its substrate proteins during folding, which are more extensive than the transient formation of mixed disulphides at the active site. In view of this argument, it is important to review what is known about the ability of PDI to interact with other proteins, and with peptides and polypeptides which might be models for its interactions with folding substrates. In the following section, we consider first some cases in which a specific ligand-binding activity has been proposed for PDI. Next we review studies on the interactions between PDI and peptides and polypeptides, including protein folding intermediates. Finally we examine the cases in which PDI forms long-term stable interactions with other polypeptides to generate species with other activities; detailed investigation of these “special cases” may illuminate several aspects of the interaction between PDI and protein folding intermediates.
Molecular chaperones and folding catalysts
492
4. BINDING PROPERTIES OF PDI 4.1. Interaction of PDI with Non-peptides 4.1.1. Interaction of PDI with Thyroid Hormone It has been shown that a bromoacetylated derivative of 3, 3', 5-triiodo-L-thyronine (T3) can specifically label a protein with a molecular mass of 55kDa in a variety of mammalian cell lines. The nucleotide sequence of this human cellular thyroid hormone binding protein, present in the endoplasmic reticulum, has been determined and the encoded protein was found to be identical to PDI (Cheng et al., 1987; Yamauchi et al., 1987). In another independent approach to identify the type I iodothyronine deiodinase (ID-I), a rat cDNA clone coding for rat PDI has been isolated (Boado et al., 1988). Here, a library was screened with polyclonal antibodies directed against solubilized rat liver microsomes with clones coding for ID-I being selected by an indirect enzyme-binding inhibition assay. However, Schoenmakers et al. showed that ID-I is not identical to PDI, since inhibitors and substrates for ID-I did not prevent the labeling of PDI with bromoacetylated [125I]T3. In contrast, a 27kDa microsomal protein with ID-I activity has been detected (Schoenmakers et al., 1989). They also showed that the binding of bromoacetylated [125I]T3 to PDI was not inhibited by unmodified T3 or 3, 3', 5, 5'tetraiodo-L-thyronine. Since PDI has a high reactivity towards alkylating agents (see above) like haloacetyl compounds the most likely explanation is that PDI reacts only with the bromoacetyl moiety. From the lack of competition between modified and unmodified T3 it is tempting to speculate that PDI has no specific binding affinity for thyroid hormone. 4.1.2. Interaction of PDI with Oestrogens PDI activity has been claimed to be selectively inhibited by certain oestrogens (Tsibris et al., 1989). In this approach the effect of oestrogens on the GSH-dependent degradation of insulin by partially purified PDI of different sources was tested with 17 -oestradiol (17 E2) and oestrone being the most potent inhibitors (Ki 1mM). Chemical crosslinking of [125I] insulin to PDI was partially inhibited in the presence of 17b-E2 suggesting that oestrogens can compete with peptides/proteins for the binding to PDI. Furthermore, it has been proposed that segments of PDI share some sequence homologies with human oestrogen receptor, since residues 101–144 (corresponding to exon 3) and residues 163– 211 of mature PDI are similar to residues 350–392 and 304–349 of human oestrogen receptor, respecively. However, the binding of radiolabeled oestradiol to PDI preparations was not saturable, thus the nature of interaction between oestrogens and PDI remains elusive. Furthermore, the quantitative interpretation of results obtained by chemical crosslinking can be very difficult, while this method provides a powerful tool to study the qualitative interaction between proteins. As for the homology between PDI and the human oestrogen receptor, it has been shown recently by NMR that it is unlikely that any of the PDI segments with similarities to the oestrogen receptor comprise individual
Protein disulphide-isomerase
493
domains (Kemmink et al., 1995: see above). Taken together, these findings suggest that PDI has no specific binding affinity for oestrogens and related compounds. 4.2. Interaction of PDI with Peptides and Proteins 4.2.1. Interaction of PDI with Peptides Morjana and Gilbert (1991) showed that peptides of various length and amino acid composition can inhibit the GSH-dependent degradation of insulin by PDI. The observed inhibition constants indicate that the affinity of PDI for peptides increases with the length of the peptide. Peptides containing cysteine residues are 4–8 fold better inhibitors than other peptides of comparable length. The Ki ranged from 46 M (20 amino acids containing a cysteine residue) to >10 mM (3 amino acids without a cysteine residue). Apart from the presence or absence of cysteine residues the binding affinity seemed to be determined by the length of the peptide while the amino acid composition was of minor significance. This result was confirmed by Noiva et al. (1991) by using a photo-crosslinking approach. It was found that PDI can be photo-crosslinked to glycosylatable peptides containing the sequence -Asn-Xaa-Ser/Thr-with the binding not restricted to peptides containing this N-glycosylation site. Furthermore, the presence of Bolton-Hunter-reagent (3-[4-hydroxy-5-iodophenyl]-propionate) in the peptides tested, seemed to increase the binding affinities. Subsequently, the binding site on PDI for the peptides was determined by using a photoreactive [125I] tripeptide probe which was photo-crosslinked to purified PDI (Noiva et al., 1993). After fragmentation of the crosslinking product the binding site could not be mapped to the thioredoxin-like active sites of PDI but was found in the Cterminal acidic segment of PDI. However, since the binding affinity of PDI for tripeptides is rather low (see above) the question is whether a tripeptide is a suitable probe. It is conceivable that a tripeptide is too short to target a specific binding site. As mentioned above, it can not be ruled out that the Bolton-Hunter-reagent, attached to a peptide, contributes significantly to the binding. Since this effect was observed to be more pronounced in small peptides, it is important to use longer peptides in order to define the complete peptide binding site of PDI. To characterise the substrate binding site of PDI, we investigated the binding of radiolabelled peptides to PDI by chemical cross-linking. We identified a specific interaction with a number of labelled oligopeptides of 10+ residues in length. The interaction was independent of the method of radiolabelling or cross-linking and appeared to be physiologically significant in that it was saturable, reversible and required native PDI; furthermore binding was competitively inhibited by unlabelled peptides and by misfolded or unfolded proteins, but not by native proteins (Klappa et al., 1997). To locate the site at which this binding occurred, we studied interactions of various recombinant fragments of human PDI, expressed in E.coli, with different radiolabelled model peptides. We observed that the b' domain of human PDI is essential and sufficient for the binding of small peptides. In the case of larger peptides, specifically a 28 amino acid fragment derived from bovine pancreatic trypsin inhibitor, or misfolded proteins, the
Molecular chaperones and folding catalysts
494
b' domain is essential but not sufficient for efficient binding, indicating that contributions from additional domains are required. Hence we propose that the different domains of PDI all contribute to the binding site with the b' domain forming the essential core (Klappa et al., 1998). 4.2.2. Interaction of PDI with Proteins By crosslinking, it has been demonstrated that PDI can interact in vivo with misfolded human lysozyme but not with correctly folded lysozyme (Otsu et al., 1994). This result clearly shows that i) PDI can interact not only with small peptides but also with proteins, ii) the interaction between PDI and proteins/peptides is not re-stricted to artifical in vitro conditions but can be detected in vivo as well, and iii) PDI can discriminate between correctly and incorrectly folded proteins. In an independent approach, nascent chains of secretory proteins after their translocation into the endoplasmic reticulum were chemically crosslinked to PDI, irrespective of whether the secretory protein contained a Cysteine residue or not (Klappa et al., 1995). The interaction between the probed secretory proteins and PDI was found to be transient and restricted to unfolded or partially folded proteins. These findings confirm the observations of Otsu et al. and indicate that PDI can bind to proteins without Cysteine residues. It is therefore tempting to speculate that PDI is not only involved in the formation and rearrangement of disulphide bonds but is also involved in other processes. This idea is supported by the observations of LaMantia and Lennarz (1993) that disruption of the gene coding for yeast PDI interfered with the viability of the yeast cells while mutations of the active site cysteine residues were not lethal but affected the formation of disulphide bonds. In addition to that, Hayano et al. (1995) could demonstrate that a PDI mutant devoid of its isomerase activity was able to accelerate intracellular folding of human lysozyme in vivo. Taken together, these results suggest that neither the active site dithiol of PDI nor a Cys residue on a protein are absolutely required for PDI/protein interaction. Therefore, PDI might be capable of taking part in the folding, prevention of aggregation, and assembly of proteins without disulphide bonds. This was confirmed by the findings that PDI was able to suppress aggregation of Dglyceraldehyde-3-phosphate dehydrogenase (GAPDH), which does not contain any disulphide bonds, during refolding in vitro (Cai et al., 1994). Similar results were obtained for rhodanese (Song and Wang, 1995). When the active site Cysteine residues were inactivated by alkylating agents, PDI lost its ability to form and rearrange disulphide bonds. However, this PDI was still active with respect to suppressing aggregation of GAPDH (Quan et al., 1995). In contrast, when PDI was incubated together with a peptide containing 28 amino acids without Cysteine residues, the efficiency of PDI to prevent aggregation of and refold GAPDH was remarkably reduced (Quan et al., 1995). The Ki of this peptide inhibitor is in the range of 200 µM, which is in excellent agreement to the numbers obtained by Morjana and Gilbert (1991). PDI therefore might interact with unfolded or partially folded proteins, thus preventing aggregation or incorrect interactions between the folding substrates. The site or sites on PDI involved in such interactions include the b' domain (see 4.2.1) but are not yet fully
Protein disulphide-isomerase
495
defined. It is also not clear whether this property is exploited in the cell in the folding of proteins lacking disulphide bonds, or whether it is simply an aspect of PDIs ability to catalyse folding in association with the generation of native disulphide bonds. 4.2.3. Interaction of PDI with ER Proteins (a) PDI as a subunit of the oligosaccharyl transferase complex It has been reported that a glycosylation site binding protein, which is a component of the oligosaccharyl transferase, is highly similar to PDI (Geetha-Habib et al., 1988). By using a photoreactive tripeptide containing the sequence -Asn-Lys-Thr- it was demonstrated that PDI can interact with glycosylatable peptides. In contrast, Bulleid and Freedman (1990) observed cotranslational glycosylation of proteins in systems depleted of PDI, which makes it unlikely that PDI is an essential component of the oligosaccharyl transferase. This result was extended by Noiva et al. (1991), who showed that the interaction between PDI and tripeptides is not limited to peptides containing Nglycosylation sites (see above). Purified PDI showed no significant preference for glycosylatable tripeptides. However, when intact microsomes were used, PDI was preferentially labeled by glycosylatable tripeptides. The most likely explanation for this is that by glycosylation these tripeptides were trapped within the microsomes, thus increasing the peptide concentration in the lumen while non-glycosylatable peptides might exit the microsomes by diffusion (Noiva et al., 1991). These findings indicate that PDI is neither a component of the oligosaccharyl transferase complex nor involved directly in the glycosylation of proteins. However, it can not be ruled out that PDI might have some indirect influence on the glycosylation of proteins, e.g. by making the glycosylation sites accessible for the oligosaccharyl transferase. (b) PDI interacting with calreticulin Recently it has been reported that PDI can interact in vitro with calreticulin, a molecular chaperone of the ER involved in the correct folding of glycoproteins (Baksh et al., 1995). Furthermore, it has been shown that interaction between calreticulin and PDI can modulate their activities: calreticulin does not bind Ca2+ with high affinity in the presence of PDI while PDI is inhibited in the presence of calreticulin with respect to folding of ‘scrambled’ ribonuclease. However, the biological significance of this interaction is far from clear, particularly in view of the finding (see 2.3) that another member of the PDI family, Erp57, interacts specifically with calnexin in vivo (Zapun et al., 1998). (c) PDI as a subunit of the microsomal triglyceride transferase PDI is an essential subunit of the microsomal triglyceride transfer protein (MTP) (Wetterau et al., 1991). This -heterodimer is obligatory for the assembly of apoBcontaining lipoproteins; the -subunit is identical with PDI. A 30 amino acid sequence at the C-terminus of the -subunit was found to be essential for the formation of the active heterodimer (Ricci et al., 1995). Upon dissociation of the heterodimer by treatment with chaotropic agents, low concentrations of denaturants or a nonionic detergent, the catalytic activity was lost completely. Furthermore, the free 88kDa -subunit showed high tendency to aggregate. Even under conditions when the -subunit did not aggregate
Molecular chaperones and folding catalysts
496
(chaotropic agents) no catalytic activity was expressed. Even when PDI was added to the -subunit after dissociation of the complex no catalytic activity was obtained, indicating that PDI must interact with the α-subunit cotranslationally. These results were confirmed by Lamberg et al. who overexpressed the MTP subunits in Sf9 cells. Expression of the asubunit alone led to insoluble aggregates with no catalytic activity while, upon coexpression of the / -subunits, soluble heterodimers were formed with MTP activity (Lamberg et al., 1996). By inactivating the active site cysteine residues of PDI/ -subunit it was demonstrated that the ability of PDI to form disulphide bonds is not essential for the formation of active MTP. However, it is still not clear whether PDI has an additional, more direct role in MTP activity. Taken together these results indicate that PDI might keep the -subunit in an active, non-aggregated conformation without the disulphide isomerase activity being involved. (d) PDI as a subunit of the prolyl 4-hydroxylase Prolyl-4-hydroxylase (P4H) is important in the posttranslational formation of 4hydroxyproline in procollagen in the ER (Kivirikko and Myllylä, 1987). This enzyme is formed by an heterodimer in C. elegans and an 2 2 tetramer in vertebrates with the -subunit identical with PDI (Pihlajaniemi et al., 1987; Vuori et al., 1992a; Veijola et al., 1994). P4H is very similar to MTP with respect to dissociation: once dissociated the subunits could not re-associate (Kirivikko et al., 1992). When the -subunit alone was synthesized in a baculovirus expression system no P4H activity was obtained, instead the -subunit had a remarked tendency to form aggregates. This result was confirmed in vitro using dog pancreas microsomes (John et al., 1993). However, upon co-expression of and -subunit aggregation of the -subunit was prevented and a functional 2 2 tetramer was formed (Vuori et al., 1992a, b). Site-directed mutagenesis of the subunit/PDI demonstrated that the active site cysteine residues of PDI were not essential for the assembly and activity of active 2 2 tetramer (Vuori et al., 1992b). Thus, as in the case of MTP (see above), these results indicate that the -subunit/ PDI is essential for keeping the -subunit in a non-aggregated, active conformation. Furthermore, the -subunit/PDI seems to be involved in the retention of P4H in the ER. Since the -subunit does not contain any retention signal, deletion of the -KDEL retention sequence of the -subunit led to substantial secretion of P4H and PDI. The interaction between PDI and P4H/MTP, respectively, might very well reflect the interaction between PDI and its substrates on which PDI acts to facilitate folding and/or disulphide bond formation. Thus, it might be useful to determine how PDI interacts with P4H/MTP in order to elucidate its function in protein folding and disulphide bond formation. 5. IS PDI A MOLECULAR CHAPERONE? “Molecular chaperones are defined as a functional class of unrelated families of protein that assist the correct non-covalent assembly of other polypeptide-containing structures in vivo, but are not components of these assembled structures when they are performing their normal biological function” (Ellis, 1993). This definition implies that molecular
Protein disulphide-isomerase
497
chaperones prevent or reverse the formation of incorrect interactions between transiently exposed surfaces of folding polypeptides. PDI clearly has the ability to suppress the aggregation of a variety of denatured proteins upon renaturation. It seems also clear that this is not exclusively related to the enzymatic function of PDI, i.e. formation and rearrangement of disulphide bonds, since PDI also suppresses the aggregation of proteins lacking disulphide bonds. It is now established that PDI has a specific binding site for oligopeptides and unfolded proteins to which the b' domain and other domains contribute, depending on the size of the ligand (Klappa et al., 1998). This interaction is clearly essential for PDI’s ability to catalyse disulphide-bond isomerization and folding in misfolded or partially folded protein substrates and could account for PDI’s ability to suppress aggregation of some non-disulphide-containing proteins during renaturation. However, all these results were obtained from in vitro experiments and there are only indirect data available suggesting that PDI is a molecular chaperone in vivo. As for PDI being identical to the -subunit of the prolyl 4-hydroxylase and of the microsomal triglyceride transferase respectively, PDI can not be regarded as a molecular chaperone in these contexts, since PDI is still a component of the assembled complexes. In summary, i) PDI certainly is an enzyme involved in the formation and rearrangement of disulphide bonds, ii) it seems certain that PDI has specific binding affinities for peptides and some unfolded proteins, which contribute to its ability to act as a catalyst of folding in addition to simple thiol-disulphide exchange, though further work is required to define these interactions, and iii) it is not yet known whether PDI indeed is a molecular chaperone in vivo. 6. REFERENCES Baksh, S., Burns, K., Andrin, C. and Michalak, M. (1995). Interaction of calreticulin with protein disulfide isomerase. J. Biol. Chem. , 270 , 31338–31344. Boado, R.J., Campbell, D.A. and Chopra, I.J. (1988). Nucleotide sequence of rat liver iodothyronine 5'-monodeiodinase (5’MD): its identity with the protein disulfide isomerase. Biochem. Biophys. Res. Com. , 155 , 1297–1304. Bulleid, N.J. and Freedman, R.B. (1988). Defective co-translational formation of disulphide bonds in protein disulphide isomerase-deficient microsomes. Nature , 335 , 649–651. Bulleid, N.J. and Freedman, R.B. (1990). Cotranslational glycosylation of proteins in systems depleted of protein disulphide isomerase. EMBO J. , 9 , 3527–3532. Cai, H., Wang, C. and Tsou, C. (1994). Chaperone-like activity of protein disulfide isomerase in the refolding of a protein with no disulfide bonds. J. Biol. Chem. , 269 , 24550–24552. Cheng., S., Gong, Q., Parkinson, C., Robinson, E.A., Appella, E., Merlino, G.T. and Pastan, I. (1987). The nucleotide sequence of a human cellular thyroid hormone binding protein present in endoplasmic reticulum. J. Biol. Chem. , 262 , 11221–11227. Creighton T.E., Hillson, D.A. and Freedman, R.B. (1980). Catalysis by proteindisulphide isomerase of the unfolding and refolding of proteins with disulphide bonds. J. Mol. Biol. , 142 , 43–62. Creighton T.E., Zapun, A. and Darby, N.J. (1995). Mechanisms and catalysts of
Molecular chaperones and folding catalysts
498
disulphide bond formation in proteins. TIBTECH , 13 , 18–23. Darby, N.J. and Creighton, T.E. (1995a). Catalytic mechanism of DsbA and its comparison with that of protein disulfide isomerase. Biochemistry , 34 , 3576–3587. Darby, N.J. and Creighton, T.E. (1995b). Functional properties of the individual thioredoxin-like domains of protein disulfide isomerase. Biochemistry , 34 , 11725– 11735. Darby, N.J. and Creighton, T.E. (1995c). Characterization of the active site cysteine residues of the thioredoxin-like domains of protein disulfide isomerase. Biochemistry , 34 , 16770–16780. Darby, N.J., Freedman, R.B. and Creighton, T.E. (1994). Dissecting the mechanism of protein disulfide isomerase: catalysis of disulfide bond formation in a model peptide. Biochemsitry , 33 , 7937–7947. Darby, N.J., Kemmink, J. and Creighton, T.E. (1996). Identifying and characterizing a structural domain of protein disulfide-isomerase. Biochemistry , 35 , 10517–10528. DeLorenzo, F., Goldberger, R.F., Steers, E.jr, Givol, D. and Anfinsen, C.B. (1966). Purification and properties of an enzyme from beef liver which catalyses sulfhydryldisulfide interchange in proteins. J. Biol. Chem. , 241 , 1562–1567. DeSilva, M.G., Lu, J., Donadel, G., Modi, W.S., Xie, H., Notkins, A.L. and Lan, M.S. (1996). Characterization and chromosomal localization of a new protein disulfide isomerase, PDIp, highly expressed in human pancreas. DNA and Cell Biol. , 15 , 9–16. DeSilva, M.G., Notkins, A.L. and Lan, M.S. (1997). Molecular characterization of a pancreas-specific protein disulfide isomerase, PDIp. DNA and Cell Biol. , 16 , 269– 274. Dunn, A., LuzJ.M., Natalia, D., Gamble, J.A., Freedman, R.B. and Tuite, M.F. (1994). Protein disulphide isomerase (PDI). is required for the secretion of a native disulphidebonded protein from Saccharomyces cerevisiae. Biochem. Soc. Trans. , 22 , 78. Edman, J.C., Ellis, L., Blacher, R.W., Roth, R.A. and Rutter, W.J. (1985). Sequence of protein disulphide isomerase and implications of its relationship to thioredoxin. Nature , 317 , 267–270. Ellis, R.J. (1993). The general concept of molecular chaperones. In Molecular Chaperones , (Ellis, R.J., Laskey, R.A., and Lorimer, G.H. eds.), Chapman and Hall, London, Glasgow, New York, Tokyo, Melbourne, Madras, 1–5. Epstein, C.J., Goldberger, R.F. and Anfinsen, C.B. (1963). The genetic control of tertiary protein structure: studies with model systems. Cold Spring Harbor Symp. Quant. Biol , 28 , 439–449. Freedman, R.B. (1984). Native disulphide bond formation in protein biosynthesis: evidence for the role of protein disulphide isomerase. TIBS , 9 , 438–441. Freedman, R.B., Hawkins, H.C. and McLaughlin, S.H. (1995). Protein disulfideisomerase. Methods Enzymol , 251 , 387–406. Freedman, R.B., Hirst, T.R. and Tuite, M.F. (1994). Protein disulphide isomerase: building bridges in protein folding. TIBS , 19 , 331–336. Freedman, R.B., Gane, P.J., Hawkins, H.C., Hlodan, R., McLaughlin, S.H. and Parry, J.W.L. (1998). Experimental and theoretical analyses of the domain architecture of mammalian protein disulphide-isomerase. Biol. Chem . 379 , in press. Gane, P.J., Freedman, R.B. and Warwicker, J. (1995). A molecular model for the redox
Protein disulphide-isomerase
499
potential difference between thioredoxin and DsbA, based on electrostatics calculations. J. Mol. Biol. , 249 , 376–387. Geetha-Habib, M., Noiva, R., Kaplan, H.A. and Lennarz, W.J. (1988). Glycosylation site binding protein, a component of oligosaccharyl transferase, is highly similar to three other 57 kd luminal proteins of the ER. Cell , 54 , 1053–1060. Givol, D., Goldberger, R.F. and Anfinsen, C.B. (1964). Oxidation and disulfide interchange in the reactivation of reduced ribonuclease. J. Biol. Chem. , 239 , PC3114PC3116. Goldberger, R.F., Epstein, C.J. and Anfinsen, C.B. (1963). Acceleration of reactivation of reduced bovine pancreatic ribonuclease by a microsomal system from rat liver. J. Biol. Chem. , 238 , 628–635. Goldenberg, D.P. (1992). Native and non-native intermediates in the BPTI folding pathway. TIBS , 17 , 257–261. Grauschopf, U., Winther, J.R., Korber, P., Zander, T., Dallinger, P. and Bardwell, J.C.A. (1995). Why is DsbA such an oxidizing disulfide catalyst? Cell , 83 , 947–955. Hayano, T., Hirose, M. and Kikuchi, M. (1995). Protein disulfide isomerase mutant lacking its isomerase activity accelerates protein folding in the cell. FEBS Lett. , 377 , 505–511. Hillson, D.A. and Freedman, R.B. (1980). Resolution of protein disulphide-isomerase and glutathione-insulin transhydrogenase activities by covalent chromatography. Biochem. J. , 191 , 373–388. Humphreys, D.P., Weir, N., Mountain, A. and Lund, P.A. (1995). Human protein disulfide isomerase functionally complements a DsbA mutation and enhances the yield of pectate lyase C in Escherichia coli. J. Biol. Chem. , 270 , 28210–28215. John, D.C.A., Grant, M.E. and Bulleid, N.J. (1993). Cell-free synthesis and assembly of prolyl 4-hydroxylase: the role of the b-subunit (PDI). in preventing misfolding and aggregation of the a-subunit. EMBO J. , 12 , 1587–1595. Kemmink, J., Darby, N.J., Dijkstra, K., Scheek, R.M. and Creighton, T.E. (1995). Nuclear magnetic resonance characterization of the N-terminal thioredoxin-like domain of protein disulfide isomerase. Protein Science , 4 , 2587–2593. Kemmink, J., Darby, N.J., Dijkstra, K., Nilges, M. and Creighton, T.E. (1996). Structure determination of the N-terminal thioredoxin-like domain of protein disulfide isomerase using multidimensional heteronuclear 13C/15N NMR spectroscopy. Biochemistry , 35 , 7684–7691. Kemmink, J., Darby, N.J., Dijkstra, K., Nilges, M. And Creighton, T.E. (1997). The folding catalyst protein disulfide isomerase is constructed of active and inactive thioredoxin modules. Current Biol , 7 , 239–245. Kivirikko, K.I. and Myllylä, R. (1987). Recent developments in posttranslational modification: intracellular processing. Methods Enzym. , 144 , 96–114. Kivirikko, K.I., Myllylä, R. and Pihlajaniemi, T. (1992). Hydroxylation of proline and lysine residues in collagens and other plant and animal proteins. In Post-Translational Modifications of Proteins , (Harding, J.J. and Crabbe, M.J.C. eds), CRC Press, Boca Raton, 1–51. Klappa, P., Freedman, R.B. and Zimmermann, R. (1995). Protein disulphide isomerase and a lumenal cyclophilin-type peptidyl prolyl cis-trans isomerase are in transient
Molecular chaperones and folding catalysts
500
contact with secretory proteins during late stages of translocation. Eur. J. Biochem. , 232 , 755–764. Klappa, P., Hawkins, C., and Freedman, R.B. (1997). Interaction between protein disulphide isomerase and peptides. Eur. J. Biochem. , 248 , 37–42. Klappa, P., Ruddock, L.W., Darby, N.J. and Freedman, R.B. (1998). The b domain provides the principal peptide-binding site of protein disulfide isomerase but all domains contribute to binding of misfolded proteins. EMBO J. , 17 , 927–35 Kortemme, T., Darby, N.J. and Creighton, T.E. (1996). Electrostatic Interactions in the Active-Site of the N-terminal thioredoxin-like domain of protein disulfide-Isomerase. Biochemistry , 35 , 14503–14511. Laboissiere, M.C.A., Sturley, S.L. and Raines, R.T. (1995). The essential function of protein-disulfide isomerase is to unscramble non-native disulfide-bonds. J. Biol. Chem. , 270 , 28006–28009. LaMantia, M. and Lennarz, W.J. (1993). The essential function of yeast protein disulfide isomerase does not reside in its isomerase activity. Cell , 74 , 899–908. Lamberg, A., Jauhiainen, M., Metso, J., Ehnholm, C., Shoulders, C., Scott, J., Pihlajaniemi, T. and Kivirikko, K.I. (1996). The role of protein disulphide isomerase in the microsomal triacylglycerol transfer protein does not reside in its isomerase activity. Biochem. J. , 315 , 533–536. Lambert, N. and Freedman, R.B. (1985). The latency of rat liver microsomal protein disulphide-isomerase. Biochem. J. , 228 , 635–645. Lu, X., Gilbert, H.F. and Harper, J.W. (1992). Conserved residues flanking the thiol/disulfide centers of protein disulfide isomerase are not essential for catalysis of thiol/disulfide exchange. Biochemistry , 31 , 4205–4210. Lundstrom J., Krause, G. and Holmgren, A. (1992). A Pro to His mutation in active site of thioredoxin increases its disulfide-isomerase activity 10-fold. J. Biol. Chem. , 267 , 9047–9052. Lyles, M.M. and Gilbert, H.F. (1991). Catalysis of the oxidative folding of ribonuclease A by protein disulfide isomerase: dependence of the rate on the composition of the redox buffer. Biochemistry , 30 , 613–619. Lyles, M.M. and Gilbert, H.F. (1994). Mutations in the thioredoxin sites of protein disulfide isomerase reveal functional nonequivalence of the N- and C-terminal domains. J. Biol.Chem. , 269 , 30946–30952. Martin, J.L. (1995). Thioredoxin: a fold for all reasons. Structure , 3 , 245–250. Martin, J.L., Bardwell, J.C.A. and Kuriyan, J. (1993). Crystal structure of the DsbA protein required for disulphide bond formation in vivo. Nature , 365 , 464–468. Morjana, N.A. and Gilbert, H.F. (1991). Effect of protein and peptide inhibitors on the activity of protein disulfide isomerase. Biochemistry , 30 , 4985–4990. Nelson, J.W. and Creighton, T.E. (1994). Reactivity and ionization of the active site cysteine residue of DsbA, a protein required for disulfide bond formation in vivo. Biochemsitry , 33 , 5974–5983. Noiva, R., Kimura, H., Roos, J. and Lennarz, W.J. (1991). Peptide binding by protein disulfide isomerase, a resident protein of the endoplasmic reticulum lumen. J. Biol. Chem. , 266 , 19645–19649. Noiva, R., Freedman, R.B. and Lennarz, W.J. (1993). Peptide binding to protein disulfide
Protein disulphide-isomerase
501
isomerase occurs at a site distinct from the active site. J. Biol. Chem. , 268 , 19210– 19217. Oliver, J.D., van der Wal, F.J., Bulleid, N.J. and High, S. (1997). Interaction of the thioldependent reductase Erp57 with nascent glycoproteins. Science , 275 , 86–88. Otsu, M., Omura, F., Yoshimori, T. and Kikuchi, M. (1994). Protein disulfide isomerase associates with misfolded human lysozyme in vivo. J. Biol. Chem. , 269 , 6874–6877. Pihlajaniemi, T., Helakoski, T., Tasanen, K., Myllyla, R., Huhtala, M.L., Koivu, J. and Kivirikko, K.I. (1987). Molecular cloning of the b-subunit of human prolyl-4hydroxylase. This product and protein disulphide isomerase are products of the same gene. EMBO J. , 6 , 643–649. Quan, H., Fan, G, and Wang, C. (1995). Independence of the chaperone activity of protein disulfide isomerase from its thioredoxin-like active site. J. Biol. Chem. , 270 , 17078–17080. Ricci, B., Sharp, D., O’Rourke, E., Kienzle, B., Blinderman, L., Gordon, D., SmithMonroy, C., Robinson, G., Gregg, R.E., Rader, D.J. and Wetterau, J.R. (1995). A 30amino acid truncation of the microsomal triglyceride transfer protein large subunit disrupts its interaction with protein disulfide-isomerase and causes abetalipoproteinemia. J. Biol. Chem. , 270 , 14281–14285. Roth, R.A. and Pierce S.B. (1987). In vivo cross-linking of protein disulfide isomerase to immunoglobulins. Biochemistry , 26 , 4179–4182 Ruddock, L.W., Hirst, T.R. and Freedman, R.B. (1996). pH-dependence of the dithioloxidizing activity of DsbA (a periplasmic protein thiol: disulphide oxidoreductase). and protein disulphide-isomerase: studies with a novel simple peptide substrate. Biochem. J. , 315 , 1001–1005. Ruoppolo, M. and Freedman, R.B. (1995). Refolding of disulfide isomerization: the mixed disulfide between ribonuclease T1 and glutathione as a model refolding substrate. Biochemistry , 34 , 9380–9388. Ruoppolo, M., Freedman, R.B., Pucci, P. and Marino, G. (1996). Glutathione-dependent pathways of refolding of Rnase T-1 by oxidation and disulfide isomerization— catalysis by protein disulfide-isomerase. Biochemistry , 35 , 13636–13646. Schoenmakers, C.H.H., Pigmans, I.G.A.J., Hawkins, H.C., Freedman, R.B. and Visser, T.J. (1989). Rat liver type I iodothyronine deiodinase is not identical to protein disulfide isomerase. Biochem. Biophys. Res. Com. , 162 , 857–868. Song, J. and Wang, C. (1995). Chaperone-like activity of protein disulfide-isomerase in the refolding of rhodanese. Eur. J. Biochem. , 231 , 312–316. Tsibris, J.C.M., Hunt, L.T., Ballejo, G., Barker, W.C., Toney, L.J. and Spellacy, W.N. (1989). Selective inhibitors of protein disulfide isomerase by estrogens. J. Biol. Chem. , 264 , 13967–13970. Veijola, J., Koivunen, P., Annunen, P., Pihlajaniemi, T. and Kivirikko, K.I. (1994). Cloning, baculovirus expression, and characterization of the a subunit of prolyl 4hydroxylase from the nematode Caenorhabditis elegans. J. Biol. Chem. , 269 , 26746– 26753. Vuori, K., Myllylä, R., Pihlajaniemi, T. and Kivirikko, K.I. (1992). Expression and sitedirected mutagenesis of human protein disulfide isomerase in Escherichia coli. J. Biol. Chem. , 267 , 7211–7214.
Molecular chaperones and folding catalysts
502
Vuori, K., Pihlajaniemi, T., Marttila, M. and Kivirikko, K.I. (1992a). Characterization of the human prolyl 4-hydroxylase tetramer and its multifunctional protein disulfideisomerase subunit synthesized in a baculovirus expression system. Proc. Natl. Acad. Sci. USA , 89 , 7467–7470. Vuori, K., Pihlajaniemi, T., Myllylä, R. and Kivirikko, K.I. (1992b). Site-directed mutagenesis of human protein disulphide isomerase: effect on the assembly, activity and endoplasmic reticulum retention of human prolyl 4-hydroxykase in Spondoptera frugiperda insect cells. EMBO J. , 11 , 4213–4217. Walker, K.W., Lyles, M.M. and Gilbert, H.F. (1996). Catalysis of oxidative protein folding by mutants of protein disulfide isomerase with a single active-site cysteine. Biochemistry , 35 , 1972–1980.. Warwicker, J. and Gane, P.J. (1996). Calculation of Cys DpKa’s and oxidising power for DsbA mutants. FEBS Letters , 385 , 105–108. Weissmann, J.S. and Kim P.S. (1993). Efficient catalysis of disulphide bond rearrangements by protein disulphide isomerase. Nature , 365 , 185–188. Wetterau, J.R., Combs, K.A., McLean, L.R., Spinner, S.N. and Aggerbeck, L.P. (1991). Protein disulfide isomerase appears necessary to maintain the catalytically active structure of the microsomal triglyceride transfer protein . Biochemistry , 30 , 9728– 9735. Wittrup, K.D. (1995). Disulfide bond formation and eukaryotic secretory productivity. Curr. Opin. Biotech. , 6 , 203–208. Yamauchi, K., Yamamoto, T., Hayashi, H., Koya, S., Takikawa, H., Toyoshima, K. and Horiuchi, R. (1987). Sequence of membrane-associated thyroid hormone binding protein from bovine liver: its identity with protein disulphide isomerase. Biochem. Biophys. Res. Com. , 146 , 1485–1492. Zapun, A., Bardwell, J.C.A. and Creighton, T.E. (1993). The reactive and destabilizing disulfide bond of DsbA, a protein required for protein disulfide bond formation in vivo. Biochemistry , 32 , 5083–5092. Zapun, A., Darby, N., Tessier, D.C., Michalak, M., Bergeron, J.J.M. and Thomas, D.Y. (1998). Enhanced catalysis of ribonuclease B folding by the interaction of calnexin or calreticulin with ERp57. J Biol Chem. , 273 , 6009–6012.
21. PEPTIDYL-PROLYL CIS/TRANS ISOMERASES GUNTER FISCHER1, * and FRANZ X.SCHMID2 1 Max-Planck-Gesellschaft,
Research Unit “Enzymology of Protein Folding”, D-06120 Halle/Saale, Germany 2 Laboratorium für Biochemie, Universität Bayreuth, D-95440 Bayreuth, Germany
1. Introduction 2. Prolyl Isomerization 2.1. Chemistry of Peptide Bond Isomerization 2.2. Prolyl Isomerizations as Slow Steps in Protein Folding 3. Enzymes Catalyzing Prolyl Isomerization 3.1. Activity Assays 3.2. Classification 3.3. Fundamental Properties 3.4. Cellular Interactions of PPIases 4. Catalysis of in vitro Protein Folding Reactions by PPIases 4.1. Acceleration of Proline-limited Folding 4.2. Ribonuclease T1 as a Model Protein for Investigating Catalyzed Protein Folding 4.3. Catalysis of RNase T1 Folding by Cyclophilins and FKBPs 4.4. Catalysis by Parvulins 4.5. Isomerization of Peptide Bonds not Preceding Proline 4.6. Autocatalytic Folding of a PPIase 5. The Trigger Factor as a Folding Enzyme 5.1. PPIase Activity 5.2. Enzyme Kinetics of Trigger Factor-catalyzed Folding 5.3. Chaperone Properties of the Trigger Factor 6. Catalysis of Prolyl Isomerization During de novo Protein Folding 7. Concluding Remarks 8. References *Corresponding author
Peptidyl-prolyl cis/trans isomerases
505
1. INTRODUCTION The complexity of protein folding originates in large part from the huge amount of rotations about single bonds that are possible in a polypeptide chain. Even when only the backbone is considered, rotations about two single bonds can occur at every Cα atom. The enormous number of conformations that are thus possible for the backbone of an unfolded protein have long been considered as a major problem for protein folding (Levinthal, 1968). At the same time the peptide bonds introduce considerable rigidity into the protein main chain. The partial double bond character of the carbon-nitrogen bond leads to the planarity of the peptide units, it creates a high barrier to rotation about this bond, and it allows a peptide bond to occur in only two conformations, cis or trans. It also leads to an enhanced stability toward nucleophilic attack. The individual properties of the amino acid preceding the prolyl bond may play a particular role for the height of the rotational barrier separating individual conformations as well as for general folding transitions. Unrestricted motions, acting cooperatively in structure-forming processes may appear kinetically uncoupled from such highly hindered rotations. Well-resolved slow folding reactions result which have a frequent occurrence even in the folding of small globular proteins. The molecular nature of the individual motion that will cause a slow, uncoupled folding event with a relaxation time > 100 ms is not always known. However, there is good evidence now that the slow isomerizations about proline imidic bonds are frequently the rate-determining events in folding, and thus they should be prime targets for a potential catalysis of protein folding. In the cell, folding should not be too slow and partially-folded intermediates should not be present for an extended time in order to minimize the risk of their aggregation. Indeed, several families of enzymes have been discovered which catalyze prolyl peptide bond isomerizations. These so-called peptidyl-prolyl cis/trans isomerases (PPIases) are the only known enzymes, evolved to stabilize a transition state that is separated from a ground state only by a difference in a torsional angle. They are very widespread and occur in virtually all tissues and organisms. As many as seven PPIases distinct in amino acid sequence, localization and enzymatic properties have been found in E. coli cells. Even more are known in yeast (Dozinski et al., 1997). PPIases are a unique class of enzymes, because they catalyze the interconversion of conformers by accelerating the rotation about a chemical bond. This is one of the simplest processes that can lead to products chemically distinct from reactants. Unlike chaperones, PPIases are classical enzymes. They do not require energy (e.g. in the form of ATP) and in their enzymatic properties they follow the simple MichaelisMenten relation. Also, they can perform unlimited numbers of catalytic cycles while leaving the product composition unaltered. In addition, a well-defined active site is located on the surface of the PPIases for binding of the substrates or inhibitors and for catalysis. Enzymes have two major functions in metabolism. They accelerate reactions which otherwise would occur at too low rate, and, in doing so, they convey specificity, because competing, unwanted reactions are not catalyzed and thus suppressed. These two functions may also be important for PPIase-catalyzed protein folding. High specificity
Molecular chaperones and folding catalysts
506
constants approaching kcat/Km≥107 M-1 s-1 were determined for substrates with the reactive bond freely accessible, indicating a perfectly evolved catalytic function. A proline ring does not suffice to activate the catalytic machinery of PPIases. Therefore, substrate reactivity must be viewed in the context of secondary binding sites within substrate chains which may be restricted by steric limitations of partially folded substrates. These preconditions create features of regiospecific enzyme actions (Kern et al., 1993).
Figure 1 (A) Representation of the prolyl isomers; (B) The family tree of peptidyl-prolyl cis/trans isomerases according to their amino acid sequence homologies.
2. PROLYL ISOMERIZATION 2.1. Chemistry of Peptide Bond Isomerization In folded proteins the peptide bonds occur only in two conformations, cis or trans, and the dihedral angles for the rotation about the CN bonds are tightly clustered around 0° (as) and 180° (trans) (Figure 1A) (Stewart et al., 1990). Peptide bonds not preceding proline are almost always trans in folded proteins, but 5.7% of all Xaa-Pro peptide bonds show the cis conformation in the proteins with known three-dimensional structure (MacArthur & Thornton, 1991). In the following we refer to Xaa-Pro peptide bonds simply as “prolyl bonds” and to their isomerizations as “prolyl isomerizations”. Prolyl bonds introduce considerable conformational heterogeneity into peptides. In principle, peptides or proteins with n proline residues can form 2n isomers unless structural constraints (such as in folded proteins) stabilize one isomer strongly relative to the other. Accordingly, as many as 16 cis/trans isomers could be identified by 1H-NMR
Peptidyl-prolyl cis/trans isomerases
507
spectroscopy for the octapeptide IFPPVPGPG, which is derived from the prolactine receptor (Oneal et al., 1996). In this case, the decrease in the fraction of isomers is not a monotone function of a decreasing number of trans peptide bonds in the molecule. Consequently, the most abundant species is the trans/trans/ cis/trans isomer having a cis conformation for the prolyl bond in penultimate position. For a prolyl bond the probability to occur in cis conformation in a polypeptide is in part determined by its local environment and, in particular, by the nature of the amino acid preceding proline. Extensive investigations on Ac-Ala-Xaa-Pro-Ala-Lys-amides, having all 20 amino acids in Xaa position, revealed a high percentage of cis isomer up to 40% when Xaa is an aromatic residue and low cis contents of about 6% when Xaa is an aliphatic residue (including proline itself). In this series of compounds the relaxation times for prolyl isomerization were found to vary about 3 fold around 1000 sec at 4°C (Reimer et al., 1998). With about 90 kJ/mol the activation barrier to prolyl isomerization is high and entirely enthalpic in nature. As a consequence, the rate of cis/trans isomerization is highly temperature dependent and decreases 3–4 fold when the temperature is decreased by 10°C (Stein, 1993). As many as four distinct transition state structures with a 90° twisted peptide bond have been suggested from theoretical calculations for the isomerization reaction. The calculated energy barriers for the isomerization leading to the syn orientation in the activated complex are always lower than those of the anti conformation (Fischer et al., 1994a). Because the cis/trans interconversion is relatively slow at low temperature and because cis and trans prolyl isomers can differ in their molecular shape, it is possible in favorable cases to separate proline peptides into different isomers by capillary zone electrophoresis (Meyer et al., 1994; Moore & Jorgenson, 1995; Ma et al., 1995) or by reversed phase chromatography in the cold (Kalman et al., 1996; Melander & Horvath, 1982). These results are straightforward for probing isomer specific recognition (Meyer et al., 1994) since the transient stability of the resolved isomers allows their separate application in chemical reactions. Contigent on the chemical and structural differences of the isomers, many examples exist for the initiation of conformational specific recognition by cis/trans isomerizations. It preferentially occurs when macromolecules are allowed to react with prolyl peptides. Well-known consequences for peptide bond hydrolysis have been reported due to isomer specificity of a wide range of proteases (Brandts & Lin, 1986; Fischer et al., 1983), even if the isomeric bond is remote from the scissile bond. Often complete lack of reactivity for the “wrong” prolyl isomer was obtained. Thus, fractions of proline-containing peptide hormones, like bradykinin, can be refractory to proteolytic inactivation even during passage through organs (Merker & Dawson, 1995). Conformational preferences in receptor binding is exemplified by the cis peptide bond preference for the binding of dermorphin analogues to the opioid receptor (Yamazaki et al., 1993). Surprisingly, selectivity in conformational recognition can already be achieved by simple, bidentate terephtaloyl amides which can bind preferentially to cis prolyl bond derivatives (Vicent et al., 1991). The rate constants of prolyl isomerizations depend on the solvent, and a change from water to organic solvents led to an about 20 fold increase in the isomerization rate of acetyl-Gly-Pro-methylester (Eberhardt et al., 1992). A similar acceleration of cis/trans isomerization was found for proline peptides which are incorporated into micelles and
Molecular chaperones and folding catalysts
508
phospholipid vesicles. In organic chemistry, micellar catalysis often follows the pseudophase model which quantitatively describes the dependence of the reaction rate on the detergent concentration (Fendler & Fendler, 1975). In fact, prolyl isomerization investigated in the presence of detergents differing in their ionic character also fits the pseudophase model. Interestingly, the rate constants kmic determined under saturating conditions do not vary greatly with respect to the charge of the micelle-forming detergent indicating close contact of the prolyl bond to the hydrocarbon chains buried in micelles (Kramer & Fischer, 1992). Under cellular conditions chemical catalysis of prolyl isomerization is difficult to obtain. The isomerization is independent of pH in the physiological range unless dissociable groups are present adjacent to proline. Acid catalysis requires protonation of the imide nitrogen, which show a pKa of about –7 and thus it can be observed only in very strong acids (Sigel & Martin, 1982; Steinberg et al., 1960; Schmid & Baldwin, 1978). Recently, in a five-membered metal chelate containing the nitrogen of the proline ring of acetyl-proline amides transiently coordinated to Cu(II) the rotational barrier to CN rotation was shown to decrease by up to 18 kJ mol-1. A catalytic rate enhancement by this Cu (II) complex of 10 fold was measured for the interconversion of all-trans polyproline to the all-cis form in CDCl3 (Cox et al., 1996). Moderate success was obtained by constructing antibodies that were complementary to the twisted peptide bond thought to mimic the transition state of PPIase catalysis (YliKauhaluoma et al., 1996). The a-dicarbonyl moiety of FK506 served as a model to design haptens for these experiments. However, the kcat values of the monoclonal antibodies produced were rather similar to those already found for the micellar catalysis described above. 2.2. Prolyl Isomerizations as Slow Steps in Protein Folding Ribonuclease A (RNase A) was the first protein for which a proline-limited folding reaction was observed. In 1973, Garel & Baldwin (1973) discovered that in its unfolded state RNase A consists of a kinetically heterogeneous mixture of fast-folding (UF) molecules, which could refold in less than a second, and slow-folding (Us) molecules, which refolded in the range of several minutes. Similar UF and Us species have since then been detected in the folding of many other proteins (Kim & Baldwin, 1982; Schmid, 1992; Schmid, 1993). Brandts and coworkers were the first to point out that this pronounced heterogeneity in the folding rates of UF and Us could be linked with prolyl isomerization. They suggested that the fast- and slow-folding molecules differ in the cis/trans isomeric state of one or more Xaa-Pro peptide bonds (cf. Figure 1) (Brandts et al., 1975). In a native protein (N) usually each prolyl peptide bond is in a defined conformation being either cis or trans in every molecule as defined by the ordered structure in the native state. After unfolding (N UF, Eq. 1), however, the prolyl bonds become free to isomerize slowly as in short oligopeptides (in the UF
Us i reaction, Eq. 1):
Peptidyl-prolyl cis/trans isomerases
509
(1) A mixture of unfolded species is thus created. It consists of a single unfolded form with correct prolyl isomers (UF) and one or more unfolded species with incorrect prolyl isomers (Us i). As outlined above, the UF molecules can refold rapidly to the native conformation. The refolding of the Us i molecules, however, is slow, because refolding is coupled with the re-isomerizations of the incorrect prolyl bonds. Of course, non-native isomers do not block the very first steps of refolding as suggested initially. Incorrect prolyl isomers can usually be well accommodated in partially folded, flexible intermediates, but not in native proteins. Thus a protein chain can begin to fold while some prolines are still in the nonnative state. At some point, however, correct prolines are required, and therefore the final steps of folding are necessarily limited in rate by their slow isomerizations (Cook et al., 1979; Schmid & Blaschek, 1981; Schmid, 1992; Schmid et al., 1993). The conformational folding with incorrect prolyl isomers can proceed to different extents, depending on the location of these non-native isomers in the structure and on the conditions used for refolding. Generally, incorrect prolyl bonds interfere least with conformational folding when they are located at the surface of the folding protein or in flexible chain regions. Solvent conditions that strongly stabilize folded proteins will also stabilize partially folded structure in intermediates with incorrect isomers. For RNase T1 it was found that protein chains with correct and with incorrect prolines folded with very similar rates of 1–5 s-1 (Mayr et al., 1996). Of course, the products of these folding reactions were different: native protein was formed in the folding of the chains with correct prolines, but in the presence of incorrect prolines intermediates were formed, which still had to undergo the very slow proline-limited final steps of folding. Conformational folding steps and prolyl isomerization are thus mutually interdependent. The presence of incorrect isomers in the chain can block folding at a certain stage, and at the same time the presence of folded structure can affect the equilibrium and the kinetic properties of the subsequent Xaa-Pro peptide bond isomerization. This close interrelationship between structure formation and prolyl peptide bond isomerization is a key feature of these slow folding steps and is of central importance for understanding the role of PPIases in these processes. 3. ENZYMES CATALYZING PROLYL ISOMERIZATION 3.1. Activity Assays There are only a few methods available to measure the enzymatic activity of PPIases. They are often based on methods monitoring the re-establishment of the fraction of isomers to actual reaction conditions following rapid disturbance of the cis/trans isomeric equilibrium. These preconditions always lead to coupled assays since a rapid jump in the conditions is followed by the relatively slow monitoring of an isomer sensitive signal. Recording of time courses must be performed up to completion of the cis/trans
Molecular chaperones and folding catalysts
510
isomerization. PPIase activities are calculated by determining the rate constants of the time-course traces. Based on these facts, isomer-specific proteolysis provides a useful method for designing sensitive PPIase assays (Fischer et al., 1984). Since chymotrypsin has a very high conformational specificity for a trans peptide bond at the P2 subsite, it cannot cleave even at high concentrations the approx. 8% fraction of cis conformer present in aqueous solution of the chromogenic oligopeptide Suc-Ala-Ala-ProPhe-4-nitroanilide. In contrast, under these conditions the anilide bond of the trans isomer is cleaved in much less than a second. Thus, after this rapid reaction, which is complete within the time of manual mixing, a slow first-order reaction is observed, which is limited in rate by the irreversible cis to trans interconversion of the Ala-Pro bond in the assay peptide. In the presence of PPIases, this slow process is accelerated and its firstorder rate constant increases. This rate constant can be used to calculate the enzymatic bimolecular rate constant kcat/Km. Evaluation of the full set of steady state enzyme constants, kcat and Km, requires specialized analysis (Kofron et al., 1991). An improved performance of the protease-coupled assay resulted from increasing the fraction of cis isomer up to 70% by starting the assay from a LiCl/trifluoroethanol stock solution of the assay peptide (Kofron et al., 1991). Some disadvantages of the protease-coupled assay such as the risk of proteolytic degradation of the PPIases under study and the limited substrate variability were avoided by using NMR-techniques (London et al., 1990; Justice et al., 1990; Kern et al., 1993; Kern et al., 1995) and proteolytically uncoupled fluorescence (GarciaEcheverria; et al., 1992) or absorbance techniques (Garcia-Echeverria; et al., 1993), but at the expense of a much lowered sensitivity and reduced variability in the conditions. 3.2. Classification In the course of identifying catalysts in pig kidney cortex for slow steps in protein folding, a proline-containing oligopeptide 4-nitroanilide was utilized to record kinetic traces typical of catalysis of prolyl isomerization. Catalysis resulted from a protein with a molecular mass lower than 20 kDa. It displayed properties typical of an enzyme thus dating the discovery of PPIases to 1984 (Fischer et al., 1984). In 1989 this enzyme was found to be identical to the already known CsA receptor cyclophilin of 18 kDa molecular mass (pig Cyp18) (Fischer et al. 1989; Takahashi et al., 1989). Subsequent research yielded a FK506-binding protein (FKBP) (see below) that represents a PPIase of a 11.8 kDa catalytic core not related to Cyp18 in its amino acid sequence (Siekierka et al., 1989; Harding et al., 1989). The currently known families of PPIases are depicted in Figure 1B. All enzymes are classified as isomerases under the EC number 5.2.1.8. The enzymatic properties and the biological functions of cyclophilins and FKBPs were reviewed recently (Fischer, 1994b; Galat & Metcalfe, 1995; Kay, 1996). A proposal for the nomenclature of the isoenzymes has been made there (Fischer, 1994b). The major cyclophilins and FKBPs tend to be high in their intracellular concentrations because kidney tubules and endothelial cells contain about 10 g of Cyp18 per mg of total protein (Ryffel et al., 1991). This is comparable to 20 M of FKBP12 present in the cytosol of lymphocyte cells (Siekierka et al., 1991). Immunosuppressive compounds, the cyclic undecapeptide cyclosporin A and the pipecolic acid-derived peptidomacrolides FK506 and rapamycin, can bind tightly (pM to
Peptidyl-prolyl cis/trans isomerases
511
M level of Ki values) to the respective active sites of cyclophilins and FKBPs, thereby inhibiting the enzymatic activity of the proteins in a reversible and competitive manner. Cross-inhibition does not occur. Formation of the PPIase/inhibitor complexes proved complicated in kinetics. Solvent-dependent conformational changes of the inhibitors as well as isomerizations of the initially formed enzyme/ inhibitor complexes can lead to lags in the development of the full inhibitory capacity of up to several hours (Kock et al., 1992; Kofron et al.; 1992; Zarnt et al.; 1995, Janowski et al., 1996). Due to the ability to bind specifically immunosuppressants, proteins of both PPIase families were alternatively designated as immunophilins (Standaert et al., 1990). However, several PPIases do not exhibit high affinity to the immunosuppressants, and other proteins devoid of PPIase activity can bind immunosuppressants (Fischer, 1994b). The inhibition of the PPIase activity of the putative cytosolic receptors of the immunosuppressant in T-lymphocytes, Cyp18 and FKBP12, seems to be insufficient to block signal transduction and to induce both immunosuppression and toxic effects of the inhibitors on cell growth (Tropschug et al., 1989; Schreiber et al., 1993). Further extension of the number of PPIase families has recently been obtained. In E. coli, the trigger factor was identified as a 48 kDa PPIase during screening the cellular protein synthesis apparatus for PPIases (Stoller et al., 1995). It is situated adjacent to nascent secretory and non-secretory polypeptide chains at the 50 S subunit of the ribosome (Valent et al., 1995; Hesterkamp et al., 1996a). Homologs of the trigger factor have been detected in other prokaryotes, but not yet in eukaryotes (Hesterkamp & Bukau, 1996b). In the complete genome of Mycoplasma genitalium, the smallest known genome for any free-living organism, a gene was identified coding for a trigger factor homologue (Fraser et al., 1995). It represents the only PPIase identified so far by homology search in this gene complement. There are no sequence similarities between the E. coli trigger factor and Cyp18. Weak similarities were found, however, with the FKBPs. By using a hydrophobic cluster analysis Callebaut & Mornon (1995) could identify a small, but significant overall identity (value of 28.3%) between FKBP12 and amino acid residues Gln148 to Thr249 of the trigger factor (Stoller et al., 1996; Hesterkamp & Bukau, 1996b). However, in consideration of the inability of high concentrations of FK506 to inactivate full-length E.coli trigger factor as well as its catalytic domain (Stoller et al., 1995; Stoller et al., 1996), positioning of these enzymes in a special subgroup of the FKBP family seems to be a reasonable choice (Figure 1B). In 1994, a cytoplasmic PPIase was identified in E. coli lacking any significant sequence similarity to the other families of PPIases. With only 92 amino acids, it forms the archetype of the parvulin family (from Latin, parvulus, very small) of PPIases (Rahfeld et al., 1994a). High concentrations of both cyclosporin A and FK506 do not affect the PPIase activity of parvulin (Rahfeld et al., 1994b). In a first approach, juglone, a chinoid walnut dye, was found to irreversibly inhibit E.coli parvulin and yeast ESS1/PTF1 leaving other PPIases completely unaffected (Hennig et al., 1998). Database searches with the protein sequence of parvulin showed highly significant similarity with a number of domains in larger bacterial, yeast, invertebrate and human proteins (Rahfeld et al., 1994a; Rudd et al. 1995; Hani et al., 1995). Several of the prokaryotic homologues were thought to be involved in maturation or transport of specific proteins (Lazar & Kolter, 1996; Kontinen & Sarvas, 1993).
Molecular chaperones and folding catalysts
512
The cellular functions of PPIases have been investigated genetically by gene disruption experiments. They usually revealed either dispensability or strong redundancy of PPIase functions. A notable exception was found in a conditional lethal phenotype for a mitochondrial Cyp20 in yeast that was identified genetically (Davis et al., 1992). Similarly, complementation of yeast cells defective in a parvulin homologue by the 18 kDa human parvulin Pin1 showed an essential function of this enzyme for normal cell growth. Pin1 variants with either a C-terminal truncation or a triple mutation of critical amino acids (both inactivating the PPIase activity) cannot complement the yeast mutant cells with respect to mitotic arrest followed by nuclear fragmentation (Lu et al., 1996). There is evidence that eukaryotic parvulins are interchangeable in terms of enzymatic function (Maleszka et al., 1996). 3.3. Fundamental Properties The catalytic activity of all PPIases seems to be independent of metal ions, cofactors or posttranslational modifications, unless sensitivity is conveyed by an extra domain in the sequence. Cyp 18 treated with Zn2+ was reported to acquire DNA binding properties but at the expense of losing its enzymatic activity (Krummrei et al., 1995). In addition, for this PPIase a Ca2+/Mg2+-dependent nuclease activity has been suggested (Montague et al., 1994). It seems likely that this additional catalytic function does not exist in the native state of Cyp 18 because a protein sample which was active in a nuclease assay did not bind CsA. While being of unclear biological significance this nuclease activity may serve as a probe for integrity of the Cyp18 structure (Schmidt et al., 1996). The specificity constant kcat/Km represents a very useful kinetic parameter for the assessment of the effectiveness of PPIase catalysis since non-productive substrate binding does not affect this parameter. Two microscopic specificity constants have to be considered, (kcat/Km)trans and (kcat/Km)cis, because catalysis occurs in both directions of the reversible isomerization. To calculate both constants it is essential to know whether or not an irreversible reaction, such as a folding step of the protein substrate, is coupled to the prolyl isomerization. The magnitude of (kcat/Km)cis can be more easily determined by an assay that is based on isomer specific proteolysis of the substrate since under the used conditions a trans to cis isomerization would never occur. Similar to many proteinases, PPIases have an extended binding site covering several peptide units of the substrate. To be recognized by Cyp18 a substrate peptide should contain at least three amino acids, including the proline (Fischer et al., 1994c). Accordingly, PPIases do not have affinity for an isolated proline ring. Although systematic investigations are still lacking, Cyp18 is believed to be maximally active toward substrates encompassing a sequence of 5 peptide bonds with the proline in the penultimate position. It is possible, however, that PPIases also bind to protein substrates at positions that are much further apart from the reactive prolyl bond (Schutkowski et al., 1996). Considering the high degree of conservation of the amino acid residues within the active cleft, which extends the binding site of the docked tetrapeptide (Zhao & Ke, 1996) in both directions of the substrate chain, enhanced affinity for large substrates is anticipated (Fischer, 1994b). The internal equilibrium constant of the Michaelis complexes Kint=Cyp 18* [trans]/
Peptidyl-prolyl cis/trans isomerases
513
Cyp18* [cis]=1.45 has been determined for Cyp 18 with a tetrapeptide substrate by line shape analysis of the 1H-NMR spectrum (Kern et al., 1995). The magnitude of Kint was shown to match the situation that an enzyme has to maintain in vivo. Enzymes evolved to operate near equilibrium have a Kint near unity (Burbaum et al., 1989). Thus, the experimental Kint for Cyp18 points to a cellular function of the enzyme for the catalysis in a reversible manner. Obviously this is at variance with a function in the de novo protein folding since this process has an equilibrium constant far from unity. The kinetic specificity of Cyp18 and FKBP12 towards peptide substrates is manifested essentially in the turnover numbers kcat, while Km values range invariantly around 0.6±0.5 mM (Decenzo et al., 1996; Kofron et al., 1991; Kern et al., 1995). Km still remains in this range for mutant FKBPs with strongly reduced (kcat/Km) cis values (Decenzo et al., 1996). Provided a favorable contribution of the amino acid in the position preceding proline, Cyp18 (kcat>104 s−1) predominates over FKBP12 in the turnover number by one or two orders of magnitude. According to these data, the proficiency of PPIases as catalysts is quite low. Defined by dividing the second-order rate constant by the rate constant k0 of the spontaneous reaction of a substrate, the proficiency factor (and the rate enhancement factor kcat/k0) of Cyp18 resides on the lower end of an enzyme scale (Radzicka &Wolfenden, 1995). The value of kcat/Km>107 M-1 s-1 that already approaches diffusion control for the bimolecular step of enzymatic catalysis indicates that the proficiency of PPIases will become physically limited. Thus enzymes catalyzing reactions with high spontaneous rates cannot evolve while maximizing proficiency. Except for cyclophilins which have a limited variability in kcat/Km within one order of magnitude, other families of PPIases prefer strongly substrates with hydrophobic residues in the position preceding proline (Stoller et al., 1995; Harrison & Stein, 1990; Bergsma et al., 1991; Albers et al., 1990). On the PPIase side, the amino acid residues that are thought to be important for catalysis seem to be highly conserved for the cyclophilins and parvulins, but less for the FKBPs (Figure 2). It was hypothesized that for cyclophilins the evolution of the intracellular compartimentalization occurred much later than the division of prokaryotes and eukaryotes took place. In contrast, distinct prokaryotic members of the FKBP family were used to diverge into specialized enzymes (Trandinh et al., 1992). Domains active as PPIases occur in many large proteins and in two cases proteolytically resistant, active FKBP and trigger factor cores could be isolated from larger proteins (Chambraud et al., 1993; Stoller et al., 1996; Hesterkamp & Bukau, 1996). Far from completeness, Figure 3 schematically describes the protein modules found in the different families of PPIases. Besides the catalytic domain most of the characteristic sequence patterns are known to direct compartimentalization of the protein. Interestingly, the presence of a WW domain in all eukaryotic parvulins with the consequent affinity for prolin-rich protein segments (Sudol et al., 1995), may confer additional specificity to the PPIase domain of parvulins.
Molecular chaperones and folding catalysts
514
Figure 2 Fully conserved amino acid residues through the entries of the PPIase database. Portions of the depicted residues can be used as signature sequences of the respective enzyme family. (A) cyclophilins; (B) FKBPs and (C) parvulines.
The three-dimensional structures of the parent parvulin and trigger factors are not yet known. However, there is a wealth of information concerning the high-resolution molecular structures of both FKBP12 and FKBP12/inhibitor complexes, and for cyclophilins complexed with cyclosporins or linear oligopeptides. Based on antiparallel -strands supplemented with short helical stretches, Cyp18 and FKBP12 show a protein fold resembling -lactoglobulin and retinol-binding protein or the pleckstrin-homology domain, respectively. The structure of Cyp18 was solved for the apoprotein as well as in several complexes with ligands. In general, the protein conformation did not change significantly upon complexation with substrates or inhibitors (Ke, 1992; Kallen & Walkinshaw, 1992; Ke at al., 1994; Zhao & Ke, 1996). A hydrophobic pocket on the surface of the eight-stranded -barrel structure forms the binding cleft. Bound oligopeptides are exclusively found with cis conformation of the respective prolyl bond. Reasoning that the minimal reaction pathway for PPIase catalysis would comprise Michaelis complexes of both the cis and trans isomer of the substrate, microscopic rate constants can be calculated by 1H-NMR based line shape consistent with preferred catalysis of trans to cis isomerization (Kern et al., 1995). Currently it is believed that the cyclophilins differ in the enzyme mechanism from the FKBPs (Fischer et al., 1993; Kramer & Fischer, 1997). For the FKBPs desolvation and substrate-assisted catalysis may be important factors in the mechanism of catalysis. For cyclophilin, an electrophilic assistance for example by Arg55 could in part account for its high activity. The positively charged side chain of this amino acid residue is located in hydrogen bonding distance perpendicularly to the plane of the proline ring (Zhao & Ke, 1996). It could possibly immobilize the lone electron
Peptidyl-prolyl cis/trans isomerases
515
Figure 3 The domain structure of several members of cyclophilins, FKBPs and parvulins. Numbers represent percentage of identical amino acids in the catalytic cores according to the BESTFIT sequence comparison program. Symbols characterize: ; transmembraine ; signal sequences ; WW domain ; metal binding ; ER targeting ; histone-like ; proline-rich ; catalytic domain.
pair at the nitrogen atom and thus, as in acid catalysis, abolish the partial double bond character of the amide bond. Indeed, the replacement of Arg55 by Ala decreases the activity of Cyp18 about 1000-fold (Zydowsky et al., 1992). In the refolding of dihydrofolate reductase a prolyl isomerization was suggested to be catalyzed intramolecularly also by an arginine (Texter et al., 1992). 3.4. Cellular Interactions of PPIases As stated above the major cytosolic cyclophilin Cyp18 may play a role in structure formation of cellular proteins that is not coupled to an irreversible folding step. Thus compartimentalized PPIases are likely to have pleiotropic biological effects. A promising approach to dissect the different functions is based on the identification of proteins intracellularly associated with PPIases. The complexes between cyclophilin and CsA and between FKBP and FK 506 or rapamycin have apparently important functions in immunosuppression. They have been reviewed recently (Dumont & Su, 1996; Cardenas et al., 1995) and will not be considered here. Copurification, affinity chromatography and the yeast two-hybrid screen provided valuable tools to characterize protein/PPIase complexes. In view of low stability of known Michaelis complexes and the lack of fairly stable intermediates of the enzyme
Molecular chaperones and folding catalysts
516
reaction it is difficult to say a priori whether or not a cellular binding protein will represent the original substrate of the PPIase within the cell. It is also uncertain whether the complex of a binding protein with a PPIase detected by the above methods will have any in vivo relevance. A useful probe may lie in complementation studies which make use of variants of the binding partners independently altered in their proteinaceous functions. Evidence of similar phenotypes will point to a common signalling pathway, and thus indicate concerted action of the proteins within the cell. Representative examples for the PPIase/protein complexes are given in the following. By using the yeast two-hybrid system, the zinc finger transcription factor YY1 was found to interact with both Cyp18 and FKBP12 (Yang et al., 1995). These interactions could be disrupted by CsA and FK506, respectively, suggesting that the PPIase active sites were involved in these interactions. The interaction of YY1 with FKBP12, but not with Cyp18, involves a C-terminal stretch of 83 amino acids. An enhanced cellular level of FKBP12, realized by overexpression, is able to competitively prevent binding of YY1 to its DNA target site. Evidence for a direct interaction between FKBP12 and YY1 in vitro has not yet been obtained. Several of the proteins associated with heterooligomeric steroid receptors in their inactive state were described as FKBPs and as a cyclophilin, each of high molecular mass (Ratajczak et al., 1993; Peattie et al., 1992) (see chapter by Toft, this volume). Conflicting reports exist about the influence of CsA and FK506 on the receptors (Milad et al., 1995). Affinity to Hsp90, another component of the receptors, was demonstrated for members of both families of PPIases, FKBP52 (Lebeau et al., 1992) and Cyp18 (Nadeau et al. 1993). For Cyp18, binding is dependent on the Mg2+/ATP concentration but does not require the hydrolysis of ATP and cannot be blocked by CsA. Obviously, in the inactive steroid receptor complexes PPIases interact with chaperones, but the function of this interaction is unknown. Another example for PPIase complexes with cellular proteins is the interaction of FKBP12 with the skeletal muscle ryanodine receptor. The first evidence came from the observation that the two proteins copurify (Collins, 1991). It now appears that most of the intracellular calcium release channels are associated with FKBP12. Involvement of the enzymatic function of FKBP12 in the regulation of the channel may be inferred from the effects of administration of the FKBP12 inhibitor rapamycin which causes immediate channel activation (Kaftan et al., 1996). Since a large distance has been found between the putative ion-conducting site and the FKBP12 binding site in the transmembrane assembly, only a long range interaction may be hypothesized for a possible involvement of FKBP12 in channel activation (Wagenknecht et al., 1996). An exchange procedure of receptor bound FKBP12 with 3 different FKBP12 variants nearly devoid of PPIase activity does not confirm the necessity of enzyme catalysis in the modulation of channel activity (Timerman et al., 1995). However, the activity of the FKBP12 variants was measured with artificial substrates. Considering the lack of knowledge about the efficiency toward natural substrates it remains difficult to predict the residual activity of the FKBP12 variants toward natural substrates. FKBP12 also forms a complex with the transforming growth factor- type 1 receptor. This interaction was identified by monitoring a rat heart cDNA library in the yeast twohybrid screen (Wang et al., 1994). It is suggested that the complex may play a role in
Peptidyl-prolyl cis/trans isomerases
517
type I receptor-mediated signalling. Co-immunoprecipitation of FKBP12 and a cytoplasmic domain of the receptor was used to confirm the interaction of both proteins in vitro. In contrast to the situation with the calcium release channel, the low activity D37G FKBP12 variant cannot maintain complex formation with the growth factor- type 1 receptor. Inactivating FKBP12 activity by rapamycin or FK506 abolished complexation as the D37G mutation did. Recently, it was shown that host cell Cyp18 is required for HIV-1 infectivity prior to reverse transcription of the virus RNA but subsequent to receptor binding and membrane fusion (Braaten et al., 1996). A two-hybrid screen identified a Cyp18/ pr55gag interaction (Franke et al., 1994; Thali et al., 1994). Incorporation of Cyp18 into HIV-1 virions was found to occur via contact with a proline-rich segment in the capsid domain of Pr55gag (Franke et al., 1994). This conserved array totaling four prolines occurs in the periodicity (P(Xaa)4P222(Xaa)2P(Xaa)5P). Mutant proteins of the HIV-1HXB2 having site-directed mutations, either P222A or G221A, failed to bind to a GST-Cyp18 fusion protein. Virions equipped with these mutant proteins cannot sequester Cyp18 into the released virions, emphasizing the importance of the Gly221-Pro222 bond for Pr55gag /Cyp18 complex formation. Analogously, small in-frame deletions in the capsid domain of Pr55ggag drastically decreased the efficiency of packaging Cyp18 into viral particles (Thali et al., 1994). It seems that the PPIase activity of Cyp18 is required for packaging, because there was a good correlation between the differential effects that several derivatives of CsA had on the enzymatic activity of Cyp18 and on the packaging process (Bartz et al., 1995). However, it is unclear whether or not catalysis of prolyl isomerization in Pr55gag contributes to the enhancement of HIV-1 infectivity mediated by the Pr55gag/Cyp18. In this respect, the level of affinity for the active site of Cyp18 is 6 fold higher for a 25-mer peptide encompassing all the relevant proline residues than it was found for a 10-mer peptide containig only 3 of the conserved prolines (Schutkowski et al., submitted). There is also convincing evidence for a specific cellular binding protein of a parvulin. The intracellular binding partner of the NIMA (never in mitosis) protein kinase, screened by the two hybrid system using the Aspergillus nidulans protein kinase with a human cDNA bank, resulted in the identification of the parvulin-like PPIase Pin1 (Lu et al., 1996). Colocalization in HeLa cells and co-immunoprecipitation experiments provided further evidence for this interaction. Pin1 specifically contacts the residues 280–396 (NID domain) of NIMA kinase known to mediate protein/ protein interactions, but does not function as substrate or inhibitor for the kinase. Taken together, these results suggested that binding of Pin1 to NID mediates inhibition of the mitotic function of NIMA protein kinase. In terms of PPIase/polypeptide interactions, a puzzling finding may be presented finally. It was shown by CsA sensitive cross-linking that the proline-free peptide CGYDVSTAQKIIAKL can react specifically with cytosolic Cyp 17 of yeast, indicating some unexpected affinity of the compounds to each other (McNew et al., 1993). Sequence specific, non-proline directed binding of oligopeptides to Cyp18 was also demonstrated in kinetic competition experiments (Fischer & Schutkowski, unpublished results). The above interaction may not cause a catalytic event given the inability of Cyp18 to catalyze the trans to cis isomerization of the stable non-proline cis peptide bond
Molecular chaperones and folding catalysts
518
Tyr38-Ala39 in a RNase T1 variant (see chapter 4.5). However, amino acid sequences lacking proline may adopt partial structural characteristics typical of proline, as it was already found for a seven residue stretch of cytochrom C551 (McArtur & Thornton, 1991). It is not unlikely that such segments could bind to PPIases. 4. CATALYSIS OF IN VITRO PROTEIN FOLDING REACTIONS BY PPIases 4.1. Acceleration of Proline-limited Folding The immunoglobulin light chain, porcine ribonuclease (RNase) and the S-protein fragment of bovine RNase A were the first proteins for which a catalysis of folding by a PPIase, porcine cytoplasmic cyclophilin18, could be demonstrated (Lang et al., 1987). In all these protein folding reactions, however, the activity of this PPIase was much lower than towards prolines in tetrapeptides. It is now clear that the rapid formation of folding intermediates and the concomitant decrease in accessibility are the major reasons for the decreased activity of the cyclophilins in protein folding. A catalysis of proline-limited steps by various PPIases was observed in the folding of many proteins, including the collagen triple helix (Bächinger et al., 1980; Bächinger. 1987; Davis et al., 1989), barnase (Matouschek et al., 1990), carbonic anhydrase (Freskgard et al., 1992; Kern et al., 1994), -lactamase (Lejeune, unpublished results), chymotrypsin inhibitor CI2 (Jackson & Fersht, 1991), yeast iso-2 cytochrome c (Veeraraghavan & Nall, 1994; Veeraraghavan et al., 1995), the immunoglobulin light chain (Lang et al., 1987; Lilie et al., 1995), and trp aporepressor (Mann et al., 1995). Most of these proteins are small, and the molecules with correct prolyl isomers refold in the time range of milliseconds. Both cis to trans and trans to cis isomerizations are catalyzed equally well in the folding of these proteins, which rules out that the catalytic function of the PPIases is restricted to the cis to trans direction. The importance of the accessibility of the prolines was demonstrated for iso-2 cytochrome c. In aqueous buffer the folding of this protein is barely catalyzed by cyclophilin. When, however, guanidinium chloride is added in increasing, but still non-denaturing concentrations to destabilize folding intermediates the prolyl bonds becomes better accessible, and the catalysis of folding is markedly improved (Veeraraghavan et al., 1995), even though the catalytic activities of the PPIases decrease with increasing denaturant concentration. Of course, not all slow steps in protein folding are prolyl isomerizations. The very slow refolding of large proteins is often limited in rate by other slow conformational rearrangements, by domain pairing reactions, or by subunit associations. 4.2. Ribonuclease T1 as a Model Protein for Investigating Catalyzed Protein Folding Much work on the mechanisms of proline-limited protein folding and of its catalysis by PPIases has been carried out with RNase T1 (Pace et al., 1991) as a model protein. Almost the entire folding of RNase T1 involves prolyl isomerizations (Kiefhaber et al., 1990b; Odefey et al., 1995), and it is a good substrate not only for PPIases, but also for other folding enzymes and chaperones (Schmid, 1993; Schmid et al., 1993).
Peptidyl-prolyl cis/trans isomerases
519
RNase T1 is a small single-domain protein of 104 amino acids (Heinemann & Saenger. 1982; Pace et al., 1991; Martinez-Oyanedel et al., 1991) with two disulfide bonds (Cys2Cys10 and Cys6-Cys103). It contains four prolyl peptide bonds; two are trans (Trp59Pro60 and Ser72-Pro73) and two are cis (Tyr38- Pro39 and Ser54-Pro55) in the native protein. RNase T1 is most stable near pH 5 and further stabilized when NaCl is added (Oobatake et al., 1979; Pace et al., 1988). Importantly, in the absence of the disulfide bonds, RNase T1 can still fold to a native-like conformation in 1 M NaCl. Variants of RNase T1 have been designed to create simple models for investigating particular aspects of protein folding. These variants differ in the conformational stability and in the number and location of cis prolyl bonds and of disulfide bonds (Kiefhaber et al., 1990; Mayr et al., 1994a; Mayr et al., 1994b). The folding mechanism of the wild-type protein with the two cis prolines and two disulfide bonds has been studied most extensively (Mayr et al., 1996). In unfolding of RNase T1, the rapid conformational unfolding step is followed by the isomerizations of the two cis prolyl peptide bonds at Pro39 and Pro55 until, at equilibrium, both Pro39 and Pro55 are 80–90% trans. Four unfolded species are thus formed, one UF species with the native isomers at Pro39 and Pro55 and three Us species with one or two incorrect isomers. The UF species with both Pro39 and Pro55 in the correct cis conformation refolds to the native state N in less than a second (Mayr et al., 1996). The three Us species with one or two incorrect prolyl isomers also regain most of their secondary structure and presumably part of their tertiary structure early in refolding. The products of these reactions are, however, partially folded intermediates that still contain the same incorrect prolines as the Us species from which these reactions were initiated. The reisomerizations of the incorrect prolyl isomers occur in the subsequent slow steps and are coupled with further folding. The tight coupling between conformational folding and the isomerizations of the prolyl peptide bonds is evident in the folding of RNase T1. As discussed above, very rapid partial folding is possible in the presence of nonnative proline isomers (Kiefhaber et al., 1990a; Kiefhaber et al., 1992), but the correct isomers are required to complete folding. The final events of folding are thus about 1000-fold decelerated. The protein chains therefore remain in a partially folded state for an extended time and could, in principle, undergo nonproductive reactions, such as aggregation. In RNase T1, partial folding has an additional unfavorable effect: it decelerates the trans to cis isomerization at Pro39 in the final step of folding (Kiefhaber et al., 1990a). In the folding of pancreatic RNase A, however, prolyl isomerization is accelerated in a folding intermediate (Cook et al., 1979; Schmid & Blaschek. 1981), and in the folding of dihydrofolate reductase intramolecular catalysis of a prolyl isomerization was proposed to occur (Texter et al., 1992). After breaking of the two disulfide bonds of RNase T1 by reduction and subsequent carboxymethylation, the resulting RCM-RNase T1 is unfolded in aqueous buffer, but folds reversibly to a native-like ordered conformation in 1 M NaCl (Mücke & Schmid, 1992). The RCM form of the S54G/P55N variant ((-P55)-RCM-T1) is a particularly valuable model protein for studying assisted protein folding. Its folding mechanism is very simple, because it involves a single trans to cis isomerization only (of Pro39) and because the access to this proline is not impaired by premature structure formation, as in
Molecular chaperones and folding catalysts
520
the presence of the disulfide bonds. Since all RCM forms are unfolded in aqueous buffer, spontaneous as well as assisted and catalyzed unfolding and refolding can be studied in the absence of denaturants, simply by varying the NaCl concentration (Mücke & Schmid, 1994a; Mücke & Schmid, 1994b). This is important because several PPIases are very sensitive to residual concentrations of denaturants, such as guanidinium chloride or urea. 4.3. Catalysis of RNase T1 Folding by Cyclophilins and FKBPs The slow reactions in the folding of RNase T1 (with intact disulfide bonds) are catalyzed by PPIases with varying efficiency. Pro55 is solvent-exposed in the native protein and presumably also in folding intermediates, and therefore catalysis at this proline is generally good. Pro39 is buried in the native protein, and its isomerization is not well catalyzed, because rapid conformational folding blocks the access for PPIases. The catalysis of Pro39 isomerization is improved when intermediate formation is inhibited by destabilizing mutations, or by breaking of the two disulfide bonds (as in the reduced and carboxymethylated RCM form of RNase T1). In these variants the PPIases are no longer hindered in their access to the refolding protein molecules (Mücke & Schmid, 1992). Many different PPIases of the cyclophilin and FKBP families from a variety of organisms catalyze the refolding of RNase T1 (Schmid et al., 1993). The best catalysis of Pro39 isomerization in the refolding of RNase T1 with intact disulfide bonds was observed with the cytoplasmic cyclophilin 18 from E. coli. In the presence of 29 M cyclophilin 18 the rate of this folding reaction was 300 fold increased (Schönbrunner et al., 1991). In protein folding the PPIases function as enzymes. They catalyze cis to trans isomerization in either direction and do not determine the isomeric states of the prolyl bonds in the protein substrates. Catalysis of prolyl isomerization in protein unfolding cannot be measured easily, because PPIases are inactivated under conditions, such as the presence of denaturants and elevated temperatures, which are normally used to denature proteins. These problems could be overcome by using the RCM forms of RNase T1 as substrates (Mücke & Schmid, 1992). As outlined above, these proteins are only folded in the presence of = 1 M NaCl and can be unfolded by a dilution to e.g. 0.4 M NaCl at pH 8 and 15°C. Under these conditions the PPIases are stable and retain their full enzymatic activity. Unfolding, or, to be more precise, the prolyl cis to trans isomerizations in the unfolding chains of RCM-RNase T1 were indeed catalyzed and the efficiency of Cyp 18 was the same in unfolding and refolding experiments under identical conditions near the midpoint of the unfolding transition. Analogous results were also obtained for the catalysis of folding by FKBP12 (Scholz et al., 1998). The extents of unfolding in the kinetic experiments with and without PPIases were found to be identical, indicating that these enzymes do not affect the equilibrium between native and unfolded protein. Clearly, as expected for enzymes, PPIases catalyze prolyl isomerization in protein folding in either direction, and they have no directional information for folding. The isomeric states of the prolines are determined only by the structure and the conformational stability of the folding protein under the employed conditions.
Peptidyl-prolyl cis/trans isomerases
521
4.4. Catalysis by Parvulins An additional family of PPIases, the parvulins, was discovered by Rahfeld et al. (1994). These small proteins show no sequence homologies with the cyclophilins and the FKBPs, and they are not inhibited by the respective inhibitors cyclosporin A and FK 506. Parvulin from E. coli also catalyzes the folding and unfolding of our model protein (P55)-RCM-T1 with an efficiency that is slightly higher than that of the FKBPs, but lower than that of the cyclophilins (Scholz et al., 1997a). 4.5. Isomerization of Peptide Bonds not Preceding Proline For normal peptide bonds (not preceding proline) the trans state is favored at least 100fold over cis (Ramachandran & Mitra, 1976; Jorgensen & Gao, 1988), which implies that in unfolded proteins on average one incorrect cis peptide bond occurs in every 100 residues. In native proteins such cis peptide bonds are infrequent, but still several of them have been found by X-ray crystallography. Therefore, non-prolyl cis/trans isomerizations should also occur in protein folding and might also be catalyzed by enzymes, possibly by the PPIases. Because of the strong preference for the trans state, it is virtually impossible to design peptide substrates with cis peptide bonds to search for such a function. Also, until recently no protein folding reactions were known which are limited in rate by the isomerization of normal peptide bonds. In our model protein RNase T1 we could generate a cis Tyr-Ala bond by replacing the cis-Pro39 of the wild-type protein by an alanine (Mayr et al., 1994). This mutation caused a major change in the folding mechanism of RNase T1. In the unfolded state virtually all molecules contain a trans Tyr38-Ala39 bond, and in refolding its very slow trans to cis isomerization determines the reaction rate (Odefey et al., 1995). The time constant of this reaction is about 500 s at 25°C. Because nascent proteins are synthesized probably as all-trans chains, the folding of the proteins which contain cis peptide bonds (Herzberg & Moult, 1991) in their native conformations should be very slow and limited in rate by trans to cis peptide bond isomerizations as in the folding of the Pro39Ala variant of RNase T1. None of the known PPIases (cyclophilin18, FKBP12, or parvulin) could catalyze this folding reaction, suggesting that normal peptide bonds might not be substrates for these enzymes (Scholz et al., 1998). It will be intriguing to learn, whether enzymes similar to the PPIases can be found, which can catalyze these isomerizations of normal peptide bonds. 4.6. Autocatalytic Folding of a PPIase As folding enzymes the PPIases can, in principle, catalyze their own folding. Indeed, for human cytosolic FKBP12 it could be shown that its folding is an autocatalytic process for the mature protein and, even more pronounced, for a variant with an aminoterminal extension of 16 residues (Veeraraghavan et al., 1995; Scholz et al., 1996). In the native form FKBP contains seven trans prolyl peptide bonds, and the cis to trans isomerizations of some or all of them determine the rate of its folding. The reaction product catalyzes its
Molecular chaperones and folding catalysts
522
own formation in an autocatalytic reaction, and therefore its rate increases with reactant concentration. Indeed, a more than 10-fold acceleration of folding was observed when the concentration of the FKBP 12 fusion protein was increased from 0.05 M to 10 M (Scholz et al., 1996). Different residues flank the seven trans prolyl bonds of FKBP12 (Siekierka et al., 1989), they probably remain not equally accessible during folding, and only two of them (Phe15-Pro16 and Ile91-Pro92) should be good substrates for the native FKBP molecules. Therefore, in the autocatalysis of its own folding, FKBP12 probably accelerates the prolyl isomerizations with different efficiencies. In the large proteins which contain FKBP-, cyclophilin- or parvulin-like domains (Fischer, 1994; Galat & Metcalfe, 1995; Rahfeld et al., 1994; Hani et al., 1995) the functions of these PPIase modules are not yet known. It is tempting to speculate that they could serve a role as intermolecular or even intramolecular catalysts of the folding or of conformational rearrangements which are necessary for the biological functions of these proteins. At present there is, however, no experimental evidence in support of such a role. 5. THE TRIGGER FACTOR AS A FOLDING ENZYME 5.1. PPIase Activity In the cells the trigger factor is located near the site of protein biosynthesis at the ribosome and can interact with newly formed protein chains. Therefore there is no doubt that it might be the prime candidate for a folding enzyme that accelerates prolyl isomerizations in the de novo folding of nascent proteins. Intriguingly, the trigger factor catalyzes the folding of RNase T1 much better than all other PPIases, which are known to date. For the cyclophilins and FKBPs the PPIase activities are much lower in protein folding than towards exposed prolyl peptide bonds in short tetrapeptides. Trigger factor, however, catalyzes prolyl isomerizations in folding proteins more efficiently than in peptides. The addition of as low as 2.5 nM trigger factor leads already to a doubling of the folding rate of (-P55)-RCM-T1, and in the presence of 20 nM trigger factor folding is 14-fold accelerated. This remarkable catalytic efficiency of the trigger factor as a folding enzyme is reflected in a specificity constant k cat /K m of 1.1×106 M-1s-1. This value is almost 100-fold higher than the respective value for human FKBP12 (Scholz et al., 1998). 5.2 Enzyme Kinetics of Trigger-factor-catalyzed Folding To determine the basis for the excellent catalysis of folding by the trigger factor, K M and k cat of catalyzed folding were determined separately (Scholz et al., 1997b). In these experiments the trigger factor behaved like a classical enzyme. It obeyed the MichaelisMenten equation in its catalysis of folding, and the initial rates of catalyzed folding showed saturation behavior. The analysis of the kinetic data gave values of 0.7 M for the K M value and 1.3 s-1 for the catalytic rate constant k cat of trigger-factor-catalyzed folding. The data can be compared with the K M of 220 M and the k cat value of 620 s-1, as measured for the catalysis by Cyp 18 of the trans to cis prolyl isomerization in a short
Peptidyl-prolyl cis/trans isomerases
523
tetrapeptide (Kern et al., 1995). This comparison indicates that the turnover number of the trigger factor in protein folding is very low, and that its high activity as a folding catalyst originates clearly from its tight binding to protein substrate, as reflected in the low K M value. For other PPIases such kinetic analyses cannot be performed, because their binding to protein or peptide substrates is very weak. 5.3. Chaperone Properties of the Trigger Factor The catalysis of RCM-T1 folding by the trigger factor is strongly inhibited by the reduced and carboxymethylated form of bovine -lactalbumin (RCM-La), a permanently unfolded, but soluble protein. RCM-La, which is also a good substrate for the chaperone GroEL (Okazaki et al., 1994; Hayerhartl et al., 1994), competes efficiently with RCM-T1 for binding to the trigger factor, and half-maximal inhibition is observed in the presence of 0.6 M RCM-La. In the presence of 1.0 M RCM-La the Michaelis-Menten kinetics of catalyzed folding are significantly shifted to higher substrate concentrations, as expected for competitive inhibition. This suggests that indeed the folding substrate RCMT1 and the non-folding inhibitor RCM-La compete for the same binding site on the trigger factor. This high-affinity binding site for unfolded proteins is probably distinct from the catalytic site of trigger factor. The isolated FKBP domain of the trigger factor is fully active as a PPIase towards a short tetrapeptide (Stoller et al., 1996), but in protein folding its activity is about 800-fold reduced, and, moreover, this residual activity of the FKBP domain is no longer inhibited by RCM-La. This suggests that the active site and the highaffinity binding site for unfolded substrate proteins reside on different domains of the intact trigger factor. It is possible that the additional protein binding site (s) on the intact trigger factor decelerates the dissociation of the protein substrate so strongly that a change in the ratelimiting step from bond rotation (in tetrapeptide substrates) to product dissociation (in protein substrates) occurs. Such a change might explain the observed low k cat value of 1.3 s-1 of the trigger factor in RCM-T1 folding. Additionally, some of the binding events should be non-productive, because the reactive prolyl peptide bonds are not positioned correctly within the PPIase site. Indeed, a lowering of both K M and k cat, as observed for the intact trigger factor, points to non-productive binding of a substrate to an enzyme (Fersht, 1985). The trigger factor thus seems to resemble a chaperone. The cooperation of protein binding and prolyl isomerization in catalyzed folding probably forms the basis for the very high catalytic efficiency of the trigger factor in protein folding. It is unknown at present, whether the chaperone properties of the trigger factor are also required for additional functions, such as accepting newly-synthesized protein chains at the ribosome. Prolyl isomerizations are presumed to be late steps in protein folding (Schmid, 1992), and therefore it seems surprising that a PPIase binds to a nascent protein chain very early, possibly cotranslationally, while it is still at the ribosome. The trigger factor might, however, remain associated with the folding chains and could even accompany the folding protein chain to other chaperone systems, such as GroEL. There is indeed evidence that the trigger factor can bind to GroEL in a substrate-dependent manner (Kandror et al., 1995). It is also possible that not all newly formed proteins require the
Molecular chaperones and folding catalysts
524
DnaK or GroE chaperone systems and consequently might fold rapidly to completion right after their synthesis at or near the ribosomes. 6. CATALYSIS OF PROLYL ISOMERIZATION DURING DE NOVO PROTEIN FOLDING Studies on the folding of collagen (Steinmann et al., 1991) provided the first evidence for an involvement of prolyl isomerization and of PPIases in a cellular protein folding reaction. The formation of the collagen triple helix is limited in rate by successive prolyl isomerizations both in vitro and in vivo. The maturation of collagen in chick embryo fibroblasts is retarded when CsA is added and Steinmann and colleagues (Steinmann et al., 1991) proposed that CsA inhibits a cyclophilin which catalyzes collagen folding in the endoplasmic reticulum. Similarly, in rabbit reticulocyte lysate the folding of luciferase was retarded by CsA (Kruse et al., 1995). Effects of CsA on proteins other than cyclophilins could of course not be ruled out in these experiments. Additional convincing evidence for a role of cyclophilins in cellular folding came from the work of Rassow et al. (1995) and Matouschek et al. (1995). Both groups used the mitochondrial protein import system to find a function of cyclophilins in protein folding. Proteins that are targeted to the mitochondrial matrix must unfold outside the mitochondria, cross the two mitochondrial membranes and then refold in the matrix. To investigate folding after transport through the membranes, both groups used a fusion protein between the presequence of subunit 9 of the Neurospora crassa F1F0-ATPase and mouse cytosolic dihydrofolate reductase (Su9-DHFR). This construct has been used before to study other aspects of protein import into isolated mitochondria. After import the refolding of DHFR in the matrix showed half times of about 5 min in yeast mitochondria at 30°C as well as in N. crassa mitochondria at 25°C. These folding reactions were about 5-fold decelerated when the mitochondria had been pretreated with 2.5–5 M CsA to inhibit the PPIase activity of mitochondrial cyclophilin. In vitro, the rates of folding of Su9-DHFR differed to a similar extent when purified cyclophilin from N. crassa mitochondria was either present or absent. The refolding of DHFR was also retarded in mitochondria from yeast and N. crassa mutants which lack a functional mitochondrial cyclophilin, and the folding of the model protein Su9-DHFR in the mutant mitochondria proceeded with similar kinetics as in wildtype mitochondria which were pretreated with CsA. This provided additional evidence that the mitochondrial cyclophilins catalyze proline-limited protein folding reactions in organello. 7. CONCLUDING REMARKS Due to the electronic nature of the prolyl peptide bond, a high rotational barrier exists in polypeptides for the dihedral angle ω which cannot be greatly lowered through low molecular weight substances within the cell. This effect gives rise to the occurrence of kinetic intermediates in protein folding being decelerated in adopting the native state as well as to native state isomerizations. The detection and characterization of several
Peptidyl-prolyl cis/trans isomerases
525
families of PPIases have been vital in advancing our understanding of such prolinelimited slow phases in in vitro protein folding. In their catalytic function these proteins act as Michaelis-Menten type enzymes. They exert catalytic effects in both directions, cis to trans and trans to cis, of the reversible prolyl isomerization. With regard to steady state catalytic constants some PPIases are reminiscent of perfectly evolved enzymes which have high turnover numbers k cat in conjunction with high Michaelis constants Km. Protein folding is catalyzed with variable efficiency in vitro. Among PPIases the E.coli trigger factor represents the most remarkable example of catalytic power when the trans to cis isomerization in the refolding of a RNase T1 variant is considered. Kinetic studies of the folding of FKBP12 gave rise to intermolecular autocatalysis in that the already folded molecules can catalyze isomerizations in molecules that are still unfolded. Several cellular binding proteins for PPIases have been identified by different techniques. Conclusions whether they would represent cellular substrates cannot yet be drawn unequivocally. Nevertheless, protein folding in cells assisted by PPIases was already detected in several cases. 8. REFERENCES Albers, M.W., Walsh, C.T. and Schreiber, S.L. (1990). Substrate specificity for human rotamase FKBP: a view of FK506 and rapamycin as leucine-(twisted amide)- proline mimic. J. Org. Chem. , 55 , 4986–4986. Bächinger, H.-P. (1987). The influence of peptidyl-prolyl cis/trans isomerase on the in vitro folding of type III collagen. J. Biol. Chem. , 262 , 17144–17148. Bächinger, H.P., Bruckner, P., Timpl, R., Prockop, D.J. and Engel, J. (1980). Folding mechanism of the triple helix in type-III collagen and type-II pN collagen. Eur. J. Biochem. , 106 , 619–632. Bartz, S.R., Hohenwalter, E., Hu, M.-K., Rich, D.H. and Malkovsky, M. (1995). Inhibition of human immunodeficiency virus replication by nonimmunosuppressive analogs of cyclosporin A. Proc. Natl. Acad. Sci. USA , 92 , 5381–5385. Braaten, D., Franke, E.A. and Luban, J. (1996). Cyclophilin A is required for an early step in the life cycle of human immunodeficiency virus type 1 before the initiation of reverse transcription. J. Virol. , 70 , 3551–3560. Brandts, J.F., Halvorson, H.R. and Brennan, M. (1975). Consideration of the possibility that the slow step in protein denaturation reactions is due to cis/trans isomerism of proline residues. Biochemistry , 14 , 4953–4963. Brandts, J.F. and Lin, L.-N. (1986). Proline isomerization studied with proteolytic enzymes. Methods Enzymol , 131 , 107–126. Bergsma, D.J., Eder, C., Gross, M., Kersten, H., Sylvester, D., Appelbaum, E., Cusimano, D., Livi, G.P., McLaughlin, M.M., Kasyan, K., Porter, T.G. Silverman, C., Dunnington, D., Hand, A. Pritchett, W.P. Bossard, M.J. Brandt, M. and Levy, M.A. (1991). The cyclophilin multigene family of peptidyl-prolyl isomerases. Characterization of three separate human isoforms. J. Biol. Chem. , 266 , 23204– 23214. Burbaum, J.J., Raines, R.T., Albery, W.J. and Knowles, J.R. (1989). Evolutionary
Molecular chaperones and folding catalysts
526
optimization of the catalytic effectiveness of an enzyme. Biochemistry , 28 , 9293–9305. Callebaut, I. and Mornon, J.P. (1995). Trigger factor, one of the Escherichia coli chaperone proteins, is an original member of the FKBP family. FEBS Lett. , 374 , 211– 215. Chambraud, B., Rouvierefourmy, N., Radanyi, C., Hsiao, K., Peattie, D.A., Livingston, D.J. and Baulieu, E.E. (1993). Overexpression of p59-HBI (FKBP59), full length and domains and characterization of PPlase activity. Biochem. Biophys. Res. Commun. , 196 , 160–166. Cardenas, M.E., Zhu, D. and Heitman, J. (1995). Molecular mechanism of immunosuppression by cyclosporine, FK506 and rapamycin. Curr. Opin. Nephrol. Hypertens. , 4 , 472–477. Collins, J.H. (1991). Sequence analysis of the ryanodine receptor: possible association with a 12KDa, FK506-binding immunophilin/protein kinase C inhibitor. Biochem. Biophys. Res. Commun. , 178 , 1288–1290. Cook, K.H., Schmid, F.X. and Baldwin, R.L. (1979). Role of proline isomerization in folding of ribonuclease A at low temperatures. Proc. Natl. Acad. Sci. USA , 76 , 6157– 6161. Cox, C., Ferraris, D., Murthy, N.N. and Lectka, T. (1996). Copper(II)-catalyzed amide isomerization -evidence for N-coordination. J. Am. Chem. Soc. , 118 , 5332–5333. Davis, E.S., Becker, A., Heitman, J., Hall, M.N. and Brennan, M.B. (1992). A yeast cyclophilin gene essential for lactate metabolism at high temperature. Proc. Natl. Acad. Sci. USA , 89 , 11169–11173. Davis, J.M., Boswell, B.A. and Bächinger, H.P. (1989). Thermal stability and folding of type IV procollagen and effect of peptidyl-prolyl cis/trans isomerase on the folding of the triple helix. J. Biol. Chem. , 264 , 8956–8962. Decenzo, M.T., Park, S.T., Jarrett, B.P., Aldape, R.A., Futer, O., Murcko, M.A. and Livingston, D.J. (1996). Fk506-binding protein mutational analysis—defining the active-site residue contributions to catalysis and the stability of ligand complexes. Protein Engineering 9 , 173–180. Dulinski, K., Muir, S., Cardenas, M. and Heitman, J. (1997). All cyclophilins and FKBPs are individually and collectively dispensible for viability in Saccharomyces cerevisiae. Proc Natl. Acad. Sci., USA , 94 , 13093–13098. Dumont, F.J. and Su, Q.X. (1996). Mechanism of action of the immunosuppressant rapamycin. Life Sci. , 58 , 373–395. Eberhardt, E.S., Loh, S.N., Hinck, A.P. and Raines, R.T. (1992). Solvent effects on the energetics of prolyl peptide bond isomerization . J. Am. Chem. Soc. , 114 , 5437–5439. Fendler, J.H. and Fendler, E.J. (1975). Catalysis in micellar and macromolecular systems. Academic Press, New York, pp. 86–103. Fersht, A. (1985). Enzyme structure and mechanism . New York: W.H.Freeman. Fischer, G., Heins, J. and Barth, A. (1983). The conformation around the peptide bond between the P1- and P2-positions is important for catalytic acivity of some prolinespecific proteases. Biochim. Biophys. Acta , 742 , 452–462. Fischer, G., Bang, H. and Mech, C. (1984). Determination of enzymatic catalysis for the cis/trans-isomerization of peptide binding in proline-containing peptides. Biomed. Biochem. Acta , 43 , 1101–1111.
Peptidyl-prolyl cis/trans isomerases
527
Fischer, G., Wittmann-Liebold, B., Lang, K., Kiefhaber, T. and Schmid, F.X. (1989). Cyclophilin and peptidyl-prolyl cis/trans isomerase are probably identical proteins. Nature , 337 , 476–478. Fischer, S., Michnick, S. and Karplus, M. (1993). A mechanism for rotamase catalysis by the FK506 binding protein (FKBP) . Biochemistry , 32 , 13830–13837. Fischer, S., Dunbrack, R.L. and Karplus, M. (1994a). Cis/trans imide isomerization of the proline dipeptide. J. Am. Chem. Soc. , 116 , 11931–11937. Fischer, G. (1994b). Peptidyl-prolyl cis/trans isomerases and their effectors. Angew. Chem. Int. Ed. Engl. , 33 , 1415–1436. Fischer, G., Wöllner, S., Schönbrunner, R. and Scherer, G. (1994c). Inhibitors for peptidyl prolyl cis/ trans isomerases. Proc. 5th Akabori Conference , Max-Planck Society, Dresden, pp. 142–145. Franke, E.K., Yuan, H.E.H. and Luban, J. (1994). Specific incorporation of cyclophilin A into HIV-1 virions. Nature , 372 , 359–362. Fraser, C.M., Gocayne, J.D., Whithe, O., Adams, M.D., Clayton, R.A., Fleischmann, R.D. Bult, C.J. Kerlavage, A.R. Sutton, G., Kelley, J.M. et al. (1995). The minimal gene complement of Mycoplasma genitalum. Science , 270 , 397–403. Freskgard, P.O., Bergenhem, N., Jonsson, B.H., Svensson, M. and Carlsson, U. (1992). Isomerase and Chaperone Activity of Prolyl Isomerase in the Folding of Carbonic Anhydrase. Science , 258 , 466–468. Galat, A. and Metcalfe, S.M. (1995). Peptidylproline cis/trans isomerases. Progr. Biophys. Mol. Biol. , 63 , 67–118. Garcia-Echeverria, C., Kofron, J.L., Kuzmic, P., Kishore, V. and Rich, D.H. (1992). Continuous fluorimetric direct (uncoupled), assay for peptidyl prolyl cis/trans isomerases. J. Am. Chem. Soc. , 114 , 2758–2759. Garcia-Echeverria, C., Kofron, J.L., Kuzmic, P. and Rich, D.H. (1993). A continuous spectrophotometric direct assay for peptidyl prolyl cis/trans isomerases. Biochem. Biophys. Res. Commun. , 191 , 70–75. Garel, J.R. and Baldwin, R.L. (1973). Both the fast and slow refolding reactions of ribonuclease A yield native enzyme. Proc. Natl. Acad. Sci. USA 70 , 3347–3351. Hani, J., Stumpf, G. and Domdey, H. (1995). PTF1 encodes an essential protein in Saccharomyces cerevisiae, which shows strong homology with a new putative family of PPIases. FEBS Lett. , 365 , 198–202. Harding, M.W., Galat, A., Uehling, D.E. and Schreiber, S.L. (1989). A receptor for the immunosuppressant FK506 is a cis/trans peptidyl-prolyl isomerase. Nature 341 , 758– 760. Harrison, R.K. and Stein, R.L. (1990). Substrate specificities of the peptidyl prolyl cis/trans isomerase activities of cyclophilin and FK-506 binding protein: evidence for the existence of a family of distinct enzymes. Biochemistry , 29 , 3813–3816. Hayerhartl, M.K., Ewbank, J.J., Creighton, T.E. and Hartl, F.U. (1994). Conformational specificity of the chaperonin GroEL for the compact folding intermediates of lactalbumin. EMBO J. , 13 , 3192–3202. Heinemann, U. and Saenger, W. (1982). Specific protein-nucleic acid recognition in ribonuclease T1–2′-guanylic acid complex: an X-ray structure. Nature , 299 , 27–31. Hennig, L., Christner, C., Kipping, M., Schelbert, B., Rucknagel, P., Kullertz, G. and
Molecular chaperones and folding catalysts
528
Fischer, G. (1998). Selective inactivation of parvulins. Biochemistry , in press. Herzberg, O. and Moult, J. (1991). Analysis of the steric strain in the polypeptide backbone of protein molecules. Proteins: Structure Function and Genetics , 11 , 223– 229. Hesterkamp, T., Hauser, S., Lutcke, H. and Bukau, B. (1996a). Escherichia coli trigger factor is a prolyl isomerase that associates with nascent polypeptide chains. Proc. Natl. Acad. Sci. USA , 93 , 4437–4441. Hesterkamp, T. and Bukau, B. (1996b). Identification of the prolyl isomerase domain of Escherichia coli trigger factor. FEBS Lett. , 385 , 67–71. Jackson, S.E. and Fersht, A.R. (1991). Folding of chymotrypsin inhibitor-2 .2. Influence of proline isomerization on the folding kinetics and thermodynamic characterization of the transition state of folding. Biochemistry , 30 , 10436–10443. Janowski, B. and Fischer, G. (1997). Modulation of cyclosporin A/cyclophilin interactions by drug vehicles. Bioorg. Med. Chem. , 5 , 179–186. Jorgensen, W.L. and Gao, J. (1988). Cis/trans energy difference for the peptide bond in the gas phase and in aqueous solution. J. Am. Chem. Soc. , 110 , 4212–4216. Justice, R.M. Jr., Kline, A.D., Sluka, J.P., Roeder, W.D., Rodgers, G.H., Roehm, N. and Mynderse, J.S. (1990). The detection of proline isomerase activity in FK506-binding protein by two-dimensional 1H NMR exchange spectroscopy. Biochem. Biophys. Res. Commun. , 171 , 445–450. Kaftan, E., Marks, A.R. and Ehrlich, B.E. (1996). Effects of rapamycin on ryanodine receptor Ca2+-release channels from cardiac muscle. Circulation Res. , 78 , 990–997. Kallen, J. and Walkinshaw, M.D. (1992). The X-ray structure of a tetrapeptide bound to the active site of human cyclophilin-A. FEBS Lett. , 300 , 286–290. Kalman, A., Thunecke, F., Schmidt, R., Schiller, P.W. and Horvath, C. (1996). Isolation and idendification of peptide conformation by reversed-phase high performance liquid chromatography and nmr at low temperature. J. Chromatogr. A , 729 , 155–171. Kandror, O., Sherman, M., Rhode, M. and Goldberg, A.L. (1995). Trigger factor is involved in GroEL-dependent protein degradation in Escherichia coli and promotes binding of GroEL to unfolded proteins. EMBO J . 14 , 6021–6027. Kay., J.E. (1996). Structure-function relationships in the FK506-binding protein (FKBP). family of peptidylprolyl cis/trans isomerases. Biochem. J. , 314 , 361–385. Ke, H. (1992). Similarities and differences between human cyclophilin A and other barrel structures. Structural refinement at 1.63 A resolution. J. Mol Biol. , 228 , 539– 550. Ke, H.M., Mayrose, D., Belshaw, P.J., Alberg, D.G., Schreiber, S.L., Chang, Z.Y., Etzkorn, F.A., Ho, S. and Walsh, C.T. (1994). Crystal structures of cyclophilin a complexed with cyclosporin A and N-methyl-4-[(E)-2-butenyl]-4, 4-dimethylthreonine cyclosporin A. Structure , 2 , 33–44. Kern, D., Drakenberg, T., Wikstrom, M., Forsen, S., Bang, H. and Fischer, G. (1993). The cis/trans interconversion of the calcium regulating hormone calcitonin is catalyzed by cyclophilin. FEBS Letters , 323 , 198–202. Kern, D., Kern, G., Scherer, G., Fischer, G. and Drakenberg, T. (1995). Kinetic analysis of cyclophilin-catalyzed prolyl cis/trans isomerization by dynamic NMR spectroscopy. Biochemistry , 34 , 13594–13602.
Peptidyl-prolyl cis/trans isomerases
529
Kern, G., Kern, D., Schmid, F.X. and Fischer, G. (1994). Reassessment of the putative chaperone function of prolyl-cis/trans-isomerases. FEBS Letters , 348 , 145–148. Kiefhaber, T., Grunert, H.-P., Hahn, U., Schmid, F.X. (1990). Replacement of a cis proline simplifies the mechanism of ribonuclease T1 folding. Biochemistry , 29 , 6475– 6480. Kiefhaber, T., Quaas, R., Hahn, U. and Schmid, F.X. (1990a). Folding of ribonuclease T1. 2. Kinetic models for the folding and unfolding reactions. Biochemistry , 29 , 3061–3070. Kiefhaber, T., Quaas, R., Hahn, U. and Schmid, F.X. (1990b). Folding of Ribonuclease T1 1. Existence of multiple unfolded states created by proline isomerization. Biochemistry , 29 , 3053–3061. Kiefhaber, T., Schmid, F.X., Willaert, K., Engelborghs, Y. and Chaffotte, A. (1992). Structure of a rapidly formed intermediate in ribonuclease-T1 folding. Protein Science , 1 , 1162–1172. Kim, P.S. and Baldwin, R.L. (1982). Specific intermediates in the folding reactions of small proteins and the mechanism of protein folding. Annu. Rev. Biochem. , 51 , 459– 489. Kock, M., Kessler, H., Seebach, D. and Thaler, A. (1992). Novel backbone conformation of cyclosporin-A—the complex with lithium chloride. J. Am. Chem. Soc , 114 , 2676– 2686. Kofron, J.L., Kuzmic, P., Kishore, V., Colon Bonilla, E. and Rich, D.H. (1991). Determination of kinetic constants for peptidyl prolyl cis/trans isomerases by an improved spectrophotometric assay. Biochemistry , 30 , 6127–6134. Kofron, J.L., Kuzmic, P., Kishore, V., Gemmecker, G., Fesik, S.W. and Rich, D.H. (1992). Lithium chloride perturbation of cis/trans peptide bond equilibria—effect on conformational equilibria in cyclosporin-A and on time-dependent inhibition of cyclophilin. J. Am. Chem. Soc. , 114 , 2670–2675. Kontinen, V.P. and Sarvas, M. (1993). The PrsA lipoprotein is essential for protein secretion in Bacillus subtilis and sets a limit for high-level secretion. Mol. Microbiol . 8 , 727–737. Kramer, M.L. and Fischer, G. (1997). FKBP-like catalysis of peptidyl-prolyl bond isomerization by micelles and membranes . Biopolymers , 42 , 49–60. Krummrei, U., Bang, R., Schmidtchen, R., Brune, K. and Bang, H. (1995). Cyclophilin-A is a zinc-dependent DNA binding protein in macrophages. FEBS Lett. , 371 , 47–51. Kruse, M., Brunke, M., Escher, A., Szalay, A.A., Tropschug, M. and Zimmermann, R. (1995). Enzyme assembly after de novo synthesis in rabbit reticulocyte lysate involves molecular chaperones and immunophilins. J. Biol. Chem. , 270 , 2588–2594. Lang, K., Schmid, F.X. and Fischer, G. (1987). Catalysis of protein folding by prolyl isomerase. Nature , 329, 268–270. Levinthal, C. (1968). Are there pathways of folding. J. Chim. Phys. , 65 , 44–45. Lilie, H., Rudolph, R. and Buchner, J. (1995). Association of antibody chains at different stages of folding: Prolyl isomerization occurs after formation of quaternary structure. J. Mol. Biol. , 248 , 190–201 . Lazar, S.W. and Kolter, R. (1996). SurA assists the folding of Escherichia coli outer membrane proteins. J. Bacterial. , 178 , 1770–1773.
Molecular chaperones and folding catalysts
530
Lebeau, M.C., Massol, N., Herrick, J., Faber, L.E., Renoir, J.M., Radanyi, C. and Baulieu, E.E. (1992). P59, an hsp 90-binding protein. Cloning and sequencing of its cDNA and preparation of a peptide-directed polyclonal antibody. J. Biol. Chem. , 267 , 4281–4284. London, R.E., Davis, D.G., Vavrek, R.J., Stewart, J.M. and Handschumacher, R.E. (1990). Bradykinin and its Gly6 analogue are substrates of cyclophilin: a fluorine-19 magnetization transfer study. Biochemistry , 29 , 10298–10302. Lu, K.P., Hanes, S.D. and Hunter, T. (1996). A human peptidyl-prolyl isomerase essential for regulation of mitosis. Nature 380 , 544–547. Ma, S., Kalman, F., Kalman, A., Thunecke, F. and Horvath, C. (1995). Capillary zone electrophoresis at subzero temperatures—Separation of the cis and trans conformers of small peptides. J. Chromatogr. A , 716 , 167–182. MacArthur, M.W. and Thornton, J.M. (1991). Influence of proline residues on protein conformation. J. Mol. Biol. , 218 , 397–412. Maleszka, R., Hanes, S.D., Hackett, R.L., deCouet, H.G. and Miklos, G.L.G. (1996). The Drosophila melanogaster dodo (dod) gene, conserved in humans, is functionally interchangeable with the ESS1 cell division gene of Saccharomyces cerevisiae. Proc. Natl Acad. Sci. USA , 93 , 447–451. Mann, C.J., Shao, X. and Matthews, C.R. (1995). Characterization of the slow folding reactions of trp aporepressor from Escherichia coli by mutational analysis of prolines and catalysis by a peptidyl-prolyl isomerase. Biochemistry , 34 , 14573–14580. Martinez-Oyanedel, J., Choe, H.-W., Heinemann, U. and Saenger, W. (1991). Ribonuclease T1 with free recognition and catalytic site: crystal structure analysis at 1.5 A resolution. J. Mol. Biol. , 222 , 335–352. Matouschek, A., Kellis, J.T., Serrano, L., Bycroft, M. and Fersht, A.R. (1990). Transient folding intermediates characterized by protein engineering. Nature , 346 , 440–445. Matouschek, A., Rospert, S., Schmid, K., Glick, B.S. and Schatz, G. (1995). Cyclophilin catalyzes protein folding in yeast mitochondria. Proc. Natl. Acad. Sci. USA , 92 , 6319–6323. McNew, J.A., Sykes, K. and Goodman, J.M. (1993). Specific cross-linking of the proline isomerase cyclophilin to a non-proline containing peptide. Mol. Biol. Cell , 4 , 223– 232 . Melander, W.R. and Horvath, C. (1982). Effect of molecular structure and conformational change of proline-containing dipeptides in reversed phase chromatography. J. Chromatgr. , 234 , 269–276. Merker, M.P. and Dawson, C.A. (1995). Cyclophilin-facilitated bradykinin inactivation in the perfused rat lung. Biochem. Pharmacol. , 50 , 2085–2091. Meyer, S., Jabs, A., Schutkowski, M. and Fischer, G. (1994). Separation of cis/trans isomers of a prolyl peptide bond by capillary zone electrophoresis. Electrophoresis , 15 , 1151–1157. Mayr, L.M., Willbold, D., Landt, O. and Schmid, F.X. (1994a). Role of the Cys 2-Cys 10 disulfide bond for the structure, stability, and folding kinetics of ribonuclease T1. Protein Sci. , 3 , 227–239. Mayr, L.M., Willbold, D., Rösch, P. and Schmid, F.X. (1994b). Generation of a nonprolyl cis peptide bond in ribonuclease T-1. J. Mol. Biol. , 240 , 288–293.
Peptidyl-prolyl cis/trans isomerases
531
Mayr, L.M., Odefey, C., Schutkowski, M. and Schmid, F.X. (1996). Kinetic analysis of the unfolding and refolding of ribonuclease T1 by a stopped-flow double-mixing technique. Biochemistry , 35 , 5550–5561. Milad, M., Sullivan, W., Diehl, E., Altmann, M., Nordeen, S., Edwards, D.P. and Toft, D.O. (1996). Interaction of the progesterone receptor with binding proteins for FK506 and cyclosporin A. Mol. Endocrin . 9 , 838–847. Montague, J.W., Gaido, M.L., Frye, C. and Cidlowski, J.A. (1994). A calcium-dependent nuclease from apoptotic rat thymocytes is homologous with cyclophilin—Recombinant cyclophilins A, B, and C have nuclease activity. J. Biol. Chem. , 269 , 18877–18880. Moore, A.W. and Jorgenson, J.W. (1995). Resolution of as and trans isomers of peptides containing proline using capillary zone electrophoresis. Anal. Chem. , 67 , 3464–3475. Mücke, M. and Schmid, F.X. (1992). Enzymatic catalysis of prolyl isomerization in an unfolding protein. Biochemistry , 31 , 7848–7854. Mücke, M. and Schmid, F.X. (1994a). Folding mechanism of ribonuclease T1 in the absence of the disulfide bonds. Biochemistry , 33 , 14608–14619. Mücke, M. and Schmid, F.X. (1994b). Intact disulfide bonds decelerate the folding of ribonuclease T1. J. Mol. Biol. , 239, 713–725. Nadeau, K., Das, A. and Walsh, C.T. (1993). Hsp90 chaperonins possess ATPase activity and bind heat shock transcription factors and peptidyl prolyl isomerases. J. Biol. Chem. , 268, 1479–1487. Odefey, C., Mayr, L.M. and Schmid, F.X. (1995). Non-prolyl cis/trans peptide bond isomerization as a rate-determining step in protein unfolding and refolding. J. Mol. Biol. , 245, 69–78. Okazaki, A., Ikura, T., Nikaido, K. and Kuwajima, K. (1994). The chaperonin GroEL does not recognize apo-α-lactalbumin in the molten globule state. Nature Struct. Biol. , 1, 439–446. Oobatake, M., Takahashi, S. and Ooi, T. (1979). Conformational stability of ribonuclease T1. I. Thermal denaturation and effects of salts. J. Biochem. , 86, 55–63. Oneal, K.D., Chari, M.V., Mcdonald, C.H., Cook, R.G., Yulee, L.Y., Morrisett, J.D. and Shearer, W.T. (1996). Multiple cis/trans conformers of the prolactin receptor prolinerich motif (PRM). peptide detected by reverse-phase HPLC, CD and NMR spectroscopy. Biochem. J. , 315, 833–844. Pace, C.N., Heinemann, U., Hahn, U. and Saenger, W. (1991). Ribonuclease T1: structure, function, and stability. Angew. Chem. Int. Ed. Engl. , 30, 343–360. Pace, C.N., Grimsley, G.R., Thomson, J.A. and Barnett, B.J. (1988). Conformational stability and activity of ribonuclease T1 with zero, one, and two intact disulfide bonds. J. Biol. Chem. , 263, 11820–11825. Peattie, D.A., Harding, M.W., Fleming, M.A., Decenzo, M.T., Lippke, J.A., Livingston, D.J. and Benasutti, M. (1992). Expression and characterization of human FKBP52, an immunophilin that associates with the 90-kDa heat shock protein and is a component of steroid receptor complexes. Proc. Natl. Acad. Sci. USA , 89, 10974–10978. Radzicka, A. and Wolfenden, R. (1995). A proficient enzyme. Science , 267, 90–93. Rahfeld, J.U., Rucknagel, K.P., Schelbert, B., Ludwig, B., Hacker, J., Mann, K. and Fischer, G. (1994a). Confirmation of the existence of a third family among peptidylprolyl cis/trans isomerases—Amino acid sequence and recombinant production of
Molecular chaperones and folding catalysts
532
parvulin. FEBS Letters , 352, 180–184. Rahfeld, J.U., Schierhorn, A., Mann, K. and Fischer, G. (1994b). A novel peptidyl-prolyl cis/trans isomerase from Escherichia coli . FEBS Letters , 343, 65–69. Ramachandran, G.N. and Mitra, A.K. (1976). An explanation for the rare occurrence of cis peptide units in proteins and polypeptides. J. Mol. Biol. , 85–92. Rassow, J., Mohrs, K., Koidl, S., Barthelmess, I.B., Pfanner, N. and Tropschug, M. (1995). Cyclophilin 20 is involved in mitochondrial protein folding in cooperation with molecular chaperones Hsp70 and Hsp60. Mol. Cell. Biol. , 15, 2654–2662. Ratajczak, T., Carrello, A., Mark, P.J., Warner, B.J., Simpson, R.J., Moritz, R.L. and House, A.K. (1993). The cyclophilin component of the unactivated estrogen receptor contains a tetratricopeptide repeat domain and shares identity with p59 (FKBP59). J. Biol. Chem. , 268, 13187–13192. Reimer, U., Scherer, G., Drewello, S., Kruber, S., Schutkowski, M. and Fischer, G. (1998). Side chain effects on peptidylprolyl cis/trans isomerization. J. Mol. Biol. , in press. Rudd, K.E., Sofia, H.J., Koonin, E.V., Plunkett, G., Lazar, S. and Rouviere, P.E. (1995). A new family of peptidyl prolyl isomerases. Trends Biochem. Sci. , 20, 12–14. Ryffel, B., Woerly, G., Greiner, B., Haendler, B., Mihatsch, M.J. and Foxwell, B.M. (1991). Distribution of the cyclosporine binding protein cyclophilin in human tissues. Immunology , 7(1), 399–404. Siekierka, J.J., Hung, S.H., Poe, M., Lin, C.S. and Sigal, N.H. (1989). A cytosolic binding protein for the immunosuppressant FK506 has peptidyl-prolyl isomerase activity but is distinct from cyclophilin. Nature , 341, 755–757. Siekierka, J.J., Wiederrecht, G., Cryan, J., Hung, S.H.Y., Comisky, M. and Sigal, N.H. (1991). Potential roles of other FK-506-Binding proteins in mediating the effects of FK-506. Tranplant. Proc. , 23, 2720–2721. Schreiber, S.L., Albers, M.W. and Brown, E.J. (1993). The cell cycle, signal transduction, and immunophilin ligand complexes. Acc. Chem. Res. , 26, 412–420. Schmid, F.X. and Baldwin, R.L. (1978). Acid catalysis of the formation of the slowfolding species of RNase A: evidence that the reaction is proline isomerization. Proc. Natl. Acad. Sci. USA , 75, 4764–4768. Schmid, F.X. and Blaschek, H. (1981). A native-like intermediate on the ribonuclease A folding pathway. 2. Comparison of its properties to native ribonuclease A. Eur. J. Biochem. , 114, 111–117. Schmid, F.X. (1992). Kinetics of unfolding and refolding of single-domain proteins. In T.E.Creighton (Ed.), Protein folding , (pp. 197–241). New York: Freeman. Schmid, F.X. (1993). Prolyl isomerase—enzymatic catalysis of slow protein-folding reactions. Annual Review of Biophysics and Biomolecular Structure , 22, 123–143. Schmid, F.X., Mayr, L.M., Mêcke, M. and Schönbrunner, E.R. (1993). Prolyl isomerases: Role in protein folding. Adv. Protein Chem. , 44, 25–66. Schmidt, B., Tradler, T., Rahfeld, J.-U., Ludwig, B. Jain, B. Mann, K. Rücknagel, K.P., Janowski, B., Schierhorn, A., Küllertz, G., Hacker J. and Fischer, G. (1996). A cyclophilin-like peptidyl-prolyl cis/ trans isomerase from Legionella pneumophila— characterisation, molecular cloning and overexpression. Mol Microbiol. , 21, 1147– 1160.
Peptidyl-prolyl cis/trans isomerases
533
Schönbrunner, E.R., Mayer, S., Tropschug, M., Fischer, G., Takahashi, N. and Schmid, F.X. (1991). Catalysis of protein folding by cyclophilins from different species. J. Biol. Chem. , 266, 3630–3635. Scholz, C., Scherer, G., Mayr, L.M., Schindler, T., Fischer, G. and Schmid, F.X. (1998). Prolyl isomerases do not catalyze isomerization of non-prolyl peptide bonds. Biol. Chem. , 329, 361–365. Scholz, C., Mücke, M., Rape, M., Pecht, A., Pahl, A., Bang, H. and Schmid, F.X. (1998). Recognition of protein substrates by the prolyl isomerase trigger factor is independent of proline residues. J. Mol. Biol. , 277, 723–732. Scholz, C., Rahfeld, J., Fischer, G. and Schmid, F.X. (1997a). Catalysis of protein folding by parvulin. J. Mol Biol , 273, 752–762. Scholz, C., Stoller, G., Zarnt, T., Fischer, G. and Schmid, F.X. (1997b). Cooperation of enzymatic and chaperone functions of trigger factor in the catalysis of protein folding. EMBO J. , 16, 54–58. Scholz, C., Zarnt, T., Kern, G., Lang, K., Burtscher, H., Fischer, G. and Schmid, F.X. (1996). Autocatalytic folding of the folding catalyst FKBP12. J. Biol Chem. , 271, 12703–12707. Schutkowski, M., Drewello, M., Wollner, S., Jakob, M., Reimer, U., Scherer, G., Schierhorn, A. and Fischer, G. (1996). Extended binding sites of cyclo-pholin as revealed by the interaction with HIV1 Gag poly protein-derived oligo-peptides. FEBS Lett. , 394, 289–294. Sigel, H. and Martin, B.R. (1982). Coordinating properties of the amide bond stability and structure of metal ion complexes of peptides and related ligands. Chem. Rev. , 82 385–426. Standaert, R.F., Galat, A., Verdine, G.L. and Schreiber, S.L. (1990). Molecular cloning and overexpression of the human FK506-binding protein FKBP. Nature , 346, 671– 674. Stein, R.L. (1993). Mechanism of enzymic and non-enzymic prolyl cis/trans isomerization. Adv. Protein Chem. , 44, 1–24. Steinberg, I.Z., Harrington, W.F., Berger, A., Sela, M. and Katchalski, E. (1960). The configurational changes of poly-L-proline in solution. J. Am. Chem. Soc. , 82, 5263– 5279. Steinmann, B., Bruckner, P. and Supertifurga, A. (1991). Cyclosporin-A slows collagen triple-helix formation in vivo—Indirect Evidence for a Physiologic Role of PeptidylProlyl cis/trans-Isomerase. J. Biol. Chem. , 266, 1299–1303. Stewart, D.E., Sarkar, A. and Wampler, J.E. (1990). Occurrence and role of cis peptide bonds in protein structures. J. Mol. Biol. , 214, 253–260. Stoller, G., Rucknagel, K.P., Nierhaus, K.H., Schmid, F.X., Fischer, G. and Rahfeld, J.U. (1995). A ribosome-associated peptidyl-prolyl cis/trans isomerase identified as the trigger factor. EMBO J. , 14, 4939–4948. Stoller, G., Tradler, T., Rucknagel, K.P., Rahfeld, J.U. and Fischer, G. (1996). An 11.8 kda proteolytic fragment of the E. coli trigger factor represents the domain carrying the peptidyl-prolyl cis/trans isomerase activity. FEBS Letters , 384, 117–122. Sudol, M., Chen, H.I., Bougeret, C., Einbond, A. and Bork, P. (1995). Characterization of a novel protein-binding module—The WW domain. FEBS Letters , 369, 67–71.
Molecular chaperones and folding catalysts
534
Takahashi, N., Hayano, T. and Suzuki, M. (1989). Peptidyl-prolyl cis/trans isomerase is the cyclosporin A-binding protein cyclophilin. Nature , 337, 473–475. Texter, F.L., Spencer, D.B., Rosenstein, R. and Matthews, C.R. (1992). Intramolecular catalysis of a proline isomerization reaction in the folding of dihydrofolate reductase. Biochemistry 31, 5687–5691. Thali, M., Bukovsky, A., Kondo, E., Rosenwirth, B., Walsh, C.T., Sodroski, J. and Gottlinger, H.G. (1994). Functional association of cyclophilin A with HIV-1 virions. Nature , 372, 363–365. Timerman, A.P., Wiederrecht, G., Marcy, A. and Fleischer, S. (1995). Characterization of an exchange reaction between soluble FKBP-12 and the FKBP ryanodine receptor complex—Modulation by FKBP mutants deficient in peptidyl-prolyl isomerase activity. J. Biol. Chem. , 270, 2451–2459. Trandinh, C.C., Pao, G.M. and Saier, M.H. (1992). Structural and evolutionary relationships among the immunophilins-2 ubiquitous families of peptidyl-prolyl cis/trans isomerases. FASEB J. , 6, 3410–3420. Tropschug, M., Barthelmess, I.B. and Neupert, W. (1989). Sensitivity to cyclosporin A is mediated by cyclophilin in Neurospora crassa and Saccharomyces cerevisiae . Nature , 342, 953–955. Valent, Q.A., Kendall, D.A., High, S., Kusters, R., Oudega, B. and Luirink, J. (1995). Early events in preprotein recognition in E-coli: Interaction of SRP and trigger factor with nascent polypeptides. EMBO J. , 14, 5494–5505 Veeraraghavan, S., Rodriguez-Gdiharpour, S., MacKinnon, C., Mcgee, W.A., Pierce, M.M. and Nall, B.T. (1995). Prolyl isomerase as a probe of stability of slow-folding intermediates. Biochemistry , 34, 12892–12902 Veeraraghavan, S., Nall, B.T. and Holzman, T.F. (1995). Protein Science , 4, Suppl. 2, 72. Vicent, C., Hirst, S.C., Garciatellado, F. and Hamilton, A.D. (1991). Conformational selectivity in molecular recognition—the influence of artificial receptors on the cis/trans isomerization of acylprolines. J. Am. Chem. Soc. , 113 , 5466–5467. Wagenknecht, T., Grassucci, R., Berkowitz, J., Wiederrecht, G.J., Xin, H.B. and Fleischer, S. (1996). Cryoelectron microscopy resolves FK506-binding protein sites on the skeletal muscle ryanodine receptor. Biophys. J. , 70, 1709–1715. Wang, T.W., Donahoe, P.K and Zervos, A.S. (1994). Specific interaction of type I receptors of the TGF- family with the immunophilin FKBP-12. Science 265, 674– 676. Yamazaki, T., Ro, M., Goodman, M., Chung, N.N. and Schiller, P.W. (1993). A topochemical approach to explain morphiceptin bioactivity. J. Med. Chem. , 36 , 708– 719. Yang, W.M., Inouye, C.J. and Seto, E. (1995). Cyclophilin A and FKBP12 interact with YY1 and alter its transcriptional activity. J. Biol. Chem. , 270, 15187–15193. Yli-Kauhaluoma, J.T., Ashley, J.A., Lo, C.H.L., Coakley, J., Wirsching, P. and Janda, K.D. (1996). Catalytic antibodies with peptidyl-prolyl cis/trans isomerase activity. J. Am. Chem. Soc. , 118 , 5496–5497. Zarnt, T., Lang, K., Burtscher, H. and Fischer, G. (1995). Time-dependent inhibition of peptidylprolyl cis/trans-isomerases by FK506 is probably due to cis/trans
Peptidyl-prolyl cis/trans isomerases
535
isomerization of the inhibitor’s imide bond. Biochem. J. , 305, 159–164. Zhao, Y.D. and Ke, H.M. (1996). Crystal structure implies that cyclophilin predominantly catalyzes the trans to cis isomerization. Biochemistry , 35 , 7356–7361. Zydowsky, L.D., Etzkorn, F.A., Chang, H.Y., Ferguson, S.B., Stolz, L.A., Ho, S.I. and Walsh, C.T. (1992). Active site mutants of human cyclophilin—a separate peptidylprolyl isomerase activity from cyclosporin-A binding and calcineurin inhibition. Protein Sci. , 1, 1092–1099.
22. THE ATPase CYCLE OF THE GroE MOLECULAR CHAPERONES NEIL A.RANSON1 and ANTHONY R.CLARKE* Department of Biochemistry, School of Medical Sciences, University of Bristol, Bristol, BS8 1TD, UK 1Present address: Department of Crystallography, Birkbeck College London, Malet Street, London, WC1E 7HX, UK
1. Introduction 2. GroEL-protein Interactions: Some Basic Considerations 3. ATP-induced Rearrangements 3.1. Equilibrium Properties and Transient Kinetics of the ATP-induced Conformational Switch 3.2. Negative Co-operativity and the Establishment of Asymmetry in the GroEL Toroids 4. The Binding of ADP 4.1. Weak and Tight Interactions—A Further Manifestation of Negative Cooperativity 5. The GroEL ATPase Cycle in the Absence of GroES 6. GroES-GroEL Interactions and Their Effects on the Hydrolysis of ATP 6.1. Binding of GroES to the ADP and ATP States of GroEL 6.2. ATP Hydrolysis in the Presence of GroES 6.3. The Dissociation of GroES During the ATPase Cycle 6.4. Bullets and Footballs 7. The Coupling of Protein and Nucleotide Binding Affinities 7.1. The Effects of ATP, ADP & AMP-PNP on Protein Binding Affinity 7.2. The Effect of Unfolded Protein on Co-operativity Within the GroEL Oligomer 8. The Coupling of the ATPase Cycle to Chaperonin-assisted Protein Folding 8.1. Models of Chaperonin Action 8.1.1. Encapsulation and Folding 8.1.2. Unfolding and Release 8.2. Towards a Global Model for GroE-mediated Folding 9. Concluding Remarks 10. Acknowledgements 11. References
Molecular chaperones and folding catalysts
538
Abbreviations ATP—
adenosine 5' triphosphate
ADP—
adenosine 5' diphosphate
*Corresponding author
AMP-PNP—
adenylylimidodiphosphate
LDH—
bacillus stearothermophilus lactate dehydrogenase
DHFR—
dihydrofolate reductase
GroEL—
Escherichia coli chaperonin-60
GroES—
Escherichia coli chaperonin-10
mMDH—
porcine mitochondrial malate dehydrogenase
rubisco—
ribulose bisphosphate carboxylase oxgenase
1. INTRODUCTION One of the central tenets of biogenesis is that the primary sequence of a protein contains all of the information necessary to specify the precise, three-dimensional structure of the native state. At its simplest, protein folding can be viewed as a directed conformational search which locates the kinetically accessible state of lowest free energy. Given that the principle function of the GroE proteins is to enhance the efficiency of the folding process, it is somewhat curious that energy derived from ATP hydrolysis is required to facilitate a reaction which is intrinsically thermodynamically favourable. Where does this chemically derived energy go and how is it harnessed? Any attempt to address this intriguing and novel mechanistic question inevitably requires us to elucidate the steps in the cycle of ATP binding, hydrolysis and product release. Both cryo-electron microscopy and recent crystallographic data reveal profound changes in the structure of the GroE complex as this chemistry proceeds, i.e. energy liberated in the cleavage of orthophosphate from ATP is used to drive rearrangements of GroEL which, in turn are coupled to the folding of substrate proteins (see Burston and Saibil, this volume). In this chapter we concentrate on four aspects of this energytransducing system: (1) the kinetics of binding, hydrolysis and dissociation of nucleotides, (2) the behavioural features of GroE conformers which arise during these steps, (3) the elaborate system of conformational communication between GroEL subunits in response to nucleotide binding and (4) the dynamics of GroES and substrate interactions in the ATPase cycle. GroEL, and its co-protein GroES, were initially identified as the products of E. coli genes in which mutations could block the assembly of bacteriophage (Georgopoulos et
The ATPase cycle of the GroE molecular chaperones
539
al., 1973), but the proteins are expressed and required for cell viability at all temperatures (Fayet et al., 1989). GroEL is a tetradecamer of identical 57,259 Da. subunits arranged as two stacked heptameric rings, and GroES is a heptamer of identical 10,497 Da. subunits which associate to form a single ring (for a comprehensive review of the structures of GroEL, GroES and their complexes see the chapter by Burston and Saibil). Each GroEL subunit has a weak, Mg2+-dependent (or Mn2+-dependent, Diamant et al., 1995) ATPase activity (Hendrix, 1979), which is inhibited by approximately 50% in the presence of GroES (Chandrasekhar et al., 1986). The binding affinity of ATP is also highly dependent on the concentration of K+ ions (Viitanen et al., 1990; Todd et al., 1993). The detailed mechanistic study of the GroE proteins and the part they play in protein folding was initiated by the work of Goloubinoff et al. (1989), who recon-stituted the chaperonin system in vitro from purified components. The essential finding was that GroEL alone formed an extremely stable complex with unfolded bacterial rubisco, thus entirely arresting the folding reaction. On addition of both Mg-ATP and GroES rubisco was discharged from the complex and went on to fold with a final yield of around 80%. This ATP-dependent enhancement of folding yield has since been demonstrated and kinetically examined in vitro with a range of substrates including DHFR and rhodanese (Martin et al., 1991), LDH (Badcoe et al., 1991), glutamine synthetase (Fisher, 1992), citrate synthase (Buchner et al., 1991), mMDH (Miller et al., 1993; Staniforth et al., 1994b). There are four, general findings from this type of experiment which cast light on the coupling of the ATPase cycle to the folding reaction. Firstly, apo-GroEL is a strong inhibitor of folding owing to the high stability of the GroEL: substrate complex; hence some nucleotide-bound state arising in the ATPase cycle must be responsible for substrate dissociation. Secondly, while some substrates can be discharged by the addition of ATP alone, others require the additional impetus provided by GroES. Thirdly, the timescale of folding is usually much longer than that of the hydrolytic reaction ( ~25 seconds per subunit), implying multiple cycles of ATP turnover during chaperoninassisted folding reactions. Fourthly, under optimal conditions, assisted folding reactions give very large enhancements of yield at folding rates comparable with those measured in the absence of the chaperone. 2. GroEL-PROTEIN INTERACTIONS: SOME BASIC CONSIDERATIONS The last observation is mechanistically more interesting than it first appears. It can be argued, a priori, that the yield of the folding reaction can be enhanced by a passive process in which the unfolded protein chains are tightly bound to the surface of GroEL. This lowers the concentration of unfolded chains which are free in solution and will enhance the yield of folding, given that two criteria are satisfied: (i) the effective stability of the interaction between GroEL and the unfolded substrate is not larger than the free energy of folding of the polypeptide chain (∆G(F-U)) i.e.
Molecular chaperones and folding catalysts
540
and (ii) the loss of yield in the spontaneous folding reaction results from the aggregation of unfolded chains. In these circumstances the folding yield (Y), at the outset of the folding reaction, can be described by the competition between the velocities of folding (Vf) and aggregation (Va). This can be represented as follows:
where kf and ka are the unimolecular and multi-molecular rate constants describing these processes, [U] is the concentration of free, unfolded substrate and N is the molecularity of the aggregation process. Hence, any reduction in [U] will increase Y. The price to be paid for this energy-independent process is a reduction in folding rate. If folding can occur while the protein substrate is bound to the GroEL surface (Gray & Fersht, 1993; Itzhaki et al., 1995), then the reduction in rate is defined by the difference in binding energy between the chaperone and the unfolded state, and the chaperone and the transition state for folding (∆GB(U-t)) such that the observed rate (kobs) becomes:
If the unfolded substrate must dissociate before folding and the binding equilibrium is relatively rapid then:
Either way, the observed folding rate for a protein which binds to GroEL in the unfolded state will be slower than the intrinsic folding rate. While this passive mechanism for the enhancement of folding yield—with a concomitant reduction in folding rate—is observed when bacterial LDH folds in the presence of GroEL alone (Badcoe et al., 1991), it is violated in all reported systems when ATP and GroES are included. Hence, however nucleotides modify the binding affinity between GroEL and the protein substrate, the simple binding schemes outlined above cannot explain the action of the GroE proteins during folding. 3. ATP-INDUCED REARRANGEMENTS 3.1. Equilibrium Properties and Transient Kinetics of the ATP-induced Conformational Switch The observation that substrate proteins can be released from GroEL by adding hydrolysable Mg-ATP to a highly stable binary complex does not reveal which GroEL: nucleotide state is responsible for altering the binding properties. However, the
The ATPase cycle of the GroE molecular chaperones
541
subsequent observation that the non-hydrolysable analogue Mg-AMP-PNP is able to discharge protein substrates (Badcoe et al., 1991), identifies ATP binding, rather than any post-hydrolytic process, as the critical step in the substrate release mechanism. This finding was reinforced by transient kinetic experiments in which the binary complex between unfolded barnase and GroEL was challenged with ATP and GroES (Corrales & Fersht, 1996b). The formation of the nucleotide-containing complex initiated folding of the enzyme which proceeded at a rate which is faster than ATP hydrolysis. In addition, the fact that ADP is scarcely able to promote substrate dissociation (Staniforth et al., 1994) is entirely consistent with this conclusion. This led to the minimal hypothesis that one of the roles of the ATPase activity of GroEL was to switch the protein between two states: an ATP-stabilized conformation which has a low affinity for substrate proteins and an apo- or ADP-stabilized state with a high substrate affinity. Nucleotide binding and hydrolysis would then drive the system cyclically between these forms, leading to a continual tendency to bind and release unfolded polypeptide chains (Jackson et al., 1993; Todd et al., 1994). The early observation that the steady-state rate of ATP hydrolysis showed a sigmoidal saturation curve established not only that the complex can be converted between two conformational states but that the switch was positively co-operative. Steady-state results also reveal that the affinity of GroEL for ATP is high (K1/2~ 10–20 M) and that GroES is able to increase the co-operativity of the system by shifting the Hill constant from around 2.5 to 4.0 (Gray & Fersht, 1991; Yifrach & Horovitz, 1994). The labelling of GroEL with a conformationally sensitive fluorescent dye marked a considerable step forward in detecting and characterizing the binding of nucleotides. If the GroEL oligomer is labelled with a single pyrene maleimide group the function of the protein is not detectably altered but the fluorescence intensity is highly sensitive to rearrangements induced by nucleotide binding (Jackson et al., 1993). Owing to the slow turnover of ATP, the extrinsic fluorescence enhancement on binding of the ligand can be recorded prior to the hydrolytic step. These experiments showed that binding rather than hydrolysis of ATP is co-operative, i.e. the binding curves were remarkably similar to the rate versus [ATP] curves recorded by Gray & Fersht and Yifrach & Horovitz. If, in the classical terminology of Monod, Wyman & Changeux (Monod et al., 1965) GroEL exists in a T and an R conformation, with ATP-binding stabilizing the latter, then the pyrene fluorescence probe ought to be capable of reporting the dynamics of the cooperative rearrangement of the oligomer. The transient kinetics of this interaction have been studied using stopped-flow fluorescence spectroscopy (Jackson et al., 1993) by mixing labelled apo-GroEL (T-conformation) with differing concentrations of ATP and measuring the first-order rate constant for the formation of the R-conformer. The data show that the observed rate constant for this transition reaches a maximum of 180 s−1, and that the half-maximal rate is achieved at an ATP concentration of 4mM. The simplest kinetic mechanism which is consistent with these results can be written:
Molecular chaperones and folding catalysts
542
and the observed rate constant (kobs) for formation of the hyper-fluorescent Rconformation is given by:
Figure 1 A minimal model of ATP binding and hydrolysis by GroEL. The initial binding of ATP is weak (K1/2=4 µM) and is followed by a conformational rearrangement occurring at a maximal rate of 180 s−1, which converts the chaperonin to a state which binds ATP tightly (Kd=10 M). Hydrolysis of the bound nucleotide then converts the chaperonin back to a weak nucleotide binding state.
These transient data show the initial formation of a weak collision complex between ATP and the predominant T-conformer, with a dissociation constant of 4mM, followed by a rearrangement of the complex to the R-conformation at a rate of 180 s−1. Thus in the terminology of Monod, Wyman & Changeux (Monod et al., 1965), KT=4mM. This rather loose interaction is in stark contrast to the tight equilibrium binding of ATP (K1/2=10 µM) in which the nucleotide is ‘pulled on’ to GroEL by virtue of the T→R transition. It is clear from the transient kinetic data that the rate of the R→T transition (i.e. the value of k2, the intercept on the vertical axis) is very small, showing that the equilibrium between T and R conformations lies heavily toward the latter when the complex is occupied by ATP. The transient kinetics of this process have also been measured using an intrinsic, genetically introduced tryptophan probe at residue 485 (Tyr→Trp) and are essentially identical to those revealed by the pyrene label (Kad and Clarke., unpublished data). In the light of these results from transient experiments which define KT, the equilibrium binding data can be fully interpreted in accordance with the MWC model which envisages a concerted switch of the protein from a state in which all subunits are in
The ATPase cycle of the GroE molecular chaperones
543
the T- to one in which all are in the R-conformation (see Figure 2). In an equilibrated system, the proportion of oligomers in the R-confor-
Figure 2 Positive co-operativity in GroEL. GroEL exists in at least two states, a T-conformer with (square complex) with a weak affinity for nucleotide, and an R-conformer (circular complex) with a tight affinity. The binding of ATP is represented as promoting a T→R transition according to the MWC model of cooperatively as discussed in the text.
mation (PR) as a function of the concentration of ATP is defined by:
where c=Kr/KR (the ratio of the dissociation constants of ATP for the T- and R-
Molecular chaperones and folding catalysts
544
conformers, α=[ATP]/KR, L is the ratio of T- to R-conformers ([T]/[R]) in the unliganded state and n is the number of binding sites for ATP in the co-operative unit. The value of each of these constants is given in the scheme in Figure 2. From a biological standpoint it is important to consider why such a co-operative mechanism should have evolved in GroEL. The most compelling reason is that the binding site for a single, respectably sized protein substrate is effectively constructed from seven binding surfaces on seven individual subunits within a ring. If GroEL must bind and then displace such a substrate during the ATPase cycle, the only way to achieve this efficiently would be to coordinate the behaviour of the subunits in a ring such that all are ‘on’, or all are ‘off’ at a given instant. Hence, a concerted switch of all subunits is central to the mechanism. Such an argument is consistent with the positively co-operative unit comprising seven rather than fourteen subunits (Horovitz et al., 1994; 1995; Bochkareva & Girshovic, 1994; Burston et al., 1995). A further feature of the cooperativity parameters is the relative values of L and of c. Since the equilibrium distribution of T-rings and R-rings (T7/R7) is determined by the value of L.cn (where n is the number of ATP molecules bound), then the binding of a single ligand is sufficient to balance the equilibrium. The binding of seven ATP molecules leads to a predominance of the R7 state of 3×1017 to 1. This massive stabilization of the R-conformation when fully liganded with ATP reflects a second mechanistic property, the very tight interaction of substrate proteins with the Tconformer. When a ring is loaded with a substrate protein, the effective value of L will be much larger, hence the transition to the R-conformation—which is required for displacement of the substrate—becomes more difficult (Yifrach & Horovitz, 1996). Hence, the large change in ATP binding energy between the T and R confomers is required to break the GroEL-substrate interaction and discharge the polypeptide from the binding surface of GroEL. The co-operative, ATP-induced rearrangement of GroEL has been visualised by both negative stain and cryo-electron microscopy (Saibil et al., 1991; Chen et al., 1994; Roseman et al., 1996). Reassuringly these structures offer a clear physical explanation for the effect of ATP binding on the affinity of GroEL for protein substrates. In the ATPconformation (the R7 state) the apical domains are all rotated such that their proteinbinding surfaces are occluded and less available for interaction with the polypeptide. The structural aspects of these conformational rearrangements and their relationship with the functional properties of GroEL are discussed more extensively in the chapter by Burston and Saibil. 3.2. Negative Co-operativity and the Establishment of Asymmetry in the GroEL Toroids The basic description of ATP binding given above views GroEL as behaving like a single positively co-operative unit comprising seven sites. It was recognized at an early stage that there must be some element of negative co-operativity in the 14-mer complex, since the binding of GroES to GroEL is seen to be asymmetric in most conditions (Roseman et al., 1996; Chen et al., 1994; Xu et al., 1997). That is, the binding of the GroES co-protein to one ring reduces its ability to bind to the other. As GroES can only associate with
The ATPase cycle of the GroE molecular chaperones
545
GroEL after the latter has bound nucleotide, this poses the question of whether this asymmetry is imposed by the association of the co-protein or is established prior to this step, i.e. by nucleotide binding itself.
Figure 3 The biphasic dependence of the initial velocity of the GroEL (R196→A) ATPase upon the ATP concentration (Adapted from Yifrach & Horovitz, 1994). Data were fitted to the equation V°= (0.5*Vmax1*[S]+Vmax2*L2*[S]*(1+[S]/K)7)/(K+[S]+K*L2(1+[S]/K) 8), where S is the substrate (ATP) concentration, K is the dissociation constant of the R-state for ATP (12.5 M), L2 is an equilibrium constant derived from [RR]/[TR] (L2=0.0096), and the subscripts 1 & 2 denote the two rings of GroEL. This analysis assumes that the ATP binds exclusively to the R-conformer, that in these conditions the TT state can be ignored, and that the affinity of ATP to a ring in the Rstate is similar for both RR and TR complexes.
Initially, two lines of evidence emerged which suggested that nucleotide binding was sufficient to enforce asymmetry upon the chaperonin oligomer. Horovitz and colleagues determined the concentration dependence of the initial rate of ATP hydrolysis by a GroEL mutant Arg196→Ala (shown in Figure 3), an amino acid substitution designed to modify co-operative contacts between subunits within a ring. If all of the fourteen available GroEL sites were able to bind and hydrolyse ATP at the same time, then this experiment should give a sigmoidal plot (reflecting the positively co-operative binding of ATP by GroEL) which reaches a plateau at high ATP concentrations as the ATPase sites of GroEL become saturated with ligand. However, the concentration-dependence of the initial rate of ATP hydrolysis is strongly biphasic, showing a rapid increase in initial rate followed by an inhibition of the rate at higher ATP concentrations. They proposed that these two phases correspond to the increase in ATPase activity as one of the two rings of GroEL is saturated with ATP, and the inhibition of the ATPase rate as the second ring
Molecular chaperones and folding catalysts
546
becomes occupied at higher substrate concentrations. The authors concluded that negative co-operativity would have to exist between the GroEL rings in order that binding of nucleotide to one ring weakens the affinity of the second ring. The inhibition of the ATPase rate by ADP (Bochkareva & Girshovic, 1994) is also biphasic and points to the same conclusion.
Figure 4 Initial velocity of the GroEL ATPase as a function of GroEL concentration. ATPase velocity was estimated by measuring release of 32Pi. after four seconds, and increases linearly until only half of the available ATPase sites are occupied.
This question has also been examined by determining the stoichiometry of ATP binding to GroEL (Burston et al., 1995). Shown in Figure 4 is a plot of initial velocity as a function of protein concentration; a standard method for determining the number of active sites per protein in classical enzyme kinetics. In this case, the initial velocity of the GroEL ATPase was determined as a function of the GroEL concentration. A saturating concentration of ATP (175 µM; K1/2=10 µM) was included in each reaction, and this was mixed with varying concentrations of GroEL (expressed for simplicity in this instance as the GroEL subunit concentration) to initiate the reactions. The initial velocity of the ATPase reaction increases linearly until a GroEL concentration of 350 µM is reached. In other words, the velocity increases until there are two GroEL subunits for every ATP molecule, after which it flattens. This indicates that at any given time, only half of the GroEL ATPase sites are active, presumably the GroEL subunits constituting one of the two rings of the chaperonin. In support of this conclusion, addition of 7 ATPs per pyrenelabelled GroEL oligomer gives the maximal fluorescence enhancement after which the signal decays with a single exponential rate of 0.12 s-1 as the ATP is hydrolysed (see Figure 5). Addition of any greater quantity of ATP leads to a steady state phase of
The ATPase cycle of the GroE molecular chaperones
547
hydrolysis prior to the exponential phase in which the last 7 ATPs are turned over. This confirms the half-sites reactivity of GroEL and, in addition, shows that only 7 ATPs are required to generate the hyper-fluorescent R-conformation. Thus asymmetry is enforced at the level of negative co-operativity in nucleotide binding and the ‘one-sided’ association of GroES results from this. In summary,
Figure 5 Time resolved fluorescence of ATP binding to pyrenyl-GroEL. 100 M pyrenyl-GroEL (subunits) was present in each reaction, and either 50 M (A) or 100 M (B) ATP was added. Addition of ATP to pyrenyl-GroEL causes a 60% enhancement in fluorescence (λex=340 nm, λem=355nm), which decays as the nucleotide is hydrolysed.
seven sites on a single ring of GroEL are filled as a positively co-operative unit and the acquisition of the R-conformation in this ring inhibits ATP binding and hydrolysis on the other, i.e. negative co-operativity exists between rings preventing the formation of a symmetrical complex. The more recent work of Yifrach, Horovitz and colleagues has furthered our understanding of the GroEL ATPase through the development of a mathematical description of the co-operativity (Yifrach & Horovitz, 1995). This “nested cooperativity” model describes GroEL in terms of both the Monod-Wyman-Changeux (MWC) and Koshland-Nemethy-Filmer (KNF) models of co-operativity (see Figure 6). Each heptameric ring is in equilibrium between the T and R state, and binding of ATP to a single ring promotes the T→R transition within that ring by a concerted MWC mechanism. The two rings of GroEL however, undergo sequential transitions between TT (both rings in the T-state; i.e. with a low affinity for ATP), TR (ATP bound to one ring) and RR (ATP bound to both rings) states via a KNF mechanism where mixed oligomers are allowed. For wild type GroEL, the positively co-operative binding of ATP to one ring occurs with half-saturation at 16 µM, forming a TR complex with high ATPase activity. Weaker binding of ATP to the second ring produces the RR state which has a lower ATPase activity.
Molecular chaperones and folding catalysts
548
4. THE BINDING OF ADP 4.1. Weak and Tight Interactions—A Further Manifestation of Negative Co-operativity The interaction of ADP with GroEL has also been studied in some depth, both in the presence and in the absence of GroES. In its absence, pyrene-labelled GroEL
Figure 6 A nested model of co-operativity in the GroEL ATPase (Adapted from Yifrach & Horovitz, 1995). As discussed in the text.
detects a weak and positively co-operative binding of ADP which occurs with a K1/2 of 2.3 M Jackson et al., 1993). However, the ATPase activity of GroEL is strongly modified by the inclusion of only 5 M ADP in steady-state reactions (Yifrach & Horovitz, 1996) and, as mentioned previously, the biphasic inhibition of the GroEL ATPase by ADP has been used as evidence for asymmetric binding of nucleotides (Bochkareva & Girshovic, 1994). These experiments imply that ADP, like ATP, can exhibit positive co-operativity in binding within a ring, but once the first ring is occupied negative co-operativity inhibits binding to the second. The stoichiometry of the tight phase of ADP binding has been determined by ultrafiltration which gives a value of seven tight sites per GroEL oligomer in accordance with the implications described above (Burston et al., 1995). The K1/2 for the high affinity occupation of the first ring has been crudely estimated to be 5 (±3) M using an extremely small fluorescence quench (0.25%) reported by pyrene-labelled GroEL (Burston et al., 1995); this occurs before the onset of the large (12%) enhancement caused by the weak binding to the second ring at high ADP concentrations. This phenomenon of negative co-operativity provides a mechanism for the biphasic inhibition of the initial ATPase rate observed by Bochkareva and Girshovic (1994) since the first phase may be caused by tight binding of ADP to the ring distal to that occupied by ATP and the second by weak competition with ATP binding. Recent data using a single-tryptophan mutant of GroEL (Tyr485→Trp), which acts as a fluorescence reporter, also shows tight equilibrium binding of ADP with a dissociation constant of 70 M (Kad and Clarke, unpublished data). Moreover, the transient kinetics shows that binding occurs in two steps: an initial weak collision complex is formed (Kd=650 µM) which triggers a rapid rearrangement, at a rate of 720 s−1, to a conformation which is very different from the ATP-stabilized R-conformer with respect to its kinetic behaviour. While ATP drives the conformational equilibrium heavily in favour of the rearranged state, the ADP-stabilized conformer remains in a rapid and
The ATPase cycle of the GroE molecular chaperones
549
poised equilibrium with the initial collision state, with a reverse conformational rate of 60 s−1. Hence, the binding-rearrangement reaction for the binding of the first 7 ADPs can be described:
The ADP-stabilized conformer is shown by cryo-electron microscopy to be different in structure from that generated by the association of ATP (Roseman et al., 1996) and retains a high affinity for unfolded protein substrates (Staniforth et al., 1994a). 5. THE GroEL ATPase CYCLE IN THE ABSENCE OF GroES In the absence of GroES the steady-steady rate of ATP turnover by GroEL is 0.04–0.06 s1 per subunit or 0.08–0.12 s-1 per active subunit (Viitanen et al., 1990; Jackson et al., 1993; Todd et al., 1993; Horovitz et al., 1993). The rate-determining step in the steadystate ATPase cycle was found to be the hydrolytic reaction by the following reasoning, (i) The single exponential rate constant for a single turnover reaction (7 ATPs per complex) is identical to the steady-state rate constant per active site and there is no burst of hydrolysis preceding the steady-state phase. These observations demonstrate that the release of products is non-limiting, (ii) Conformational changes induced by ATP binding are rapid in comparison to the steady-state rate, showing that steps preceding bond breakage cannot be limiting (Jackson et al., 1993; Todd et al., 1994, Burston et al., 1995). The above observations of nucleotide binding, conformational switches and hydrolysis suggest the following model for the ATPase cycle of GroEL shown in Figure 7. Asymmetric association of ATP to one of the two rings of GroEL occurs in a positively co-operative manner (Gray & Fersht, 1991; Bochkareva et al., 1992; Jackson et al., 1993), and promotes a conformational shift from a T-state with low affinity for ATP, to an R-state with a high affinity (Jackson et al., 1993; Yifrach & Horovitz, 1994 & 1995). Subsequent hydrolysis occurs at 0.12 s-1 per active subunit (Burston et al., 1995), the slowest step in the reaction cycle. The resulting 7 molecules of ADP are stably bound (Burston et al., 1995) and a second seven molecules of ATP bind to the unoccupied ring. If ATP can hydrolyse on one ring of GroEL while the other is occupied by ADP, then the next round of hydrolysis creates a symmetrical GroEL14: ADP14 complex which is unstable (Jackson et al., 1993) owing to strong inter-ring negative co-operativity, and promotes dissociation of 7 ADPs from the complex. However, recent results show unambiguously that ATP may only turn over upon dissociation of ADP from the other ring. This phenomenon is demonstrated in experiments in which the hydrolysis of ATP on one ring is shown to be completely blocked by the binding of ADP to the other. ADP does not prevent binding of ATP to the opposite ring, but it does prevent turnover, i.e. the mode of inhibition is asymmetric and non-competitive (Kad et al., 1998). Such a mechanism provides the basis of a reciprocating, ring-switching mechanism as depicted in Figure 7, where ATP hydrolysis is carried out alternately by each GroEL ring.
Molecular chaperones and folding catalysts
550
This series of reactions imposes an obligatory change in the conformation of each of the GroEL rings during a revolution of the ATPase cycle, i.e. each GroEL ring is driven reciprocally between states which have a high and a low affinity for substrate proteins.
Figure 7 The GroEL ATPase cycle. Starting at the bottom right, ATP binds to a GroEL: ADP ring to form an asymmetric ADP7: GroEL14: ATP7
The ATPase cycle of the GroE molecular chaperones
551
complex. ADP must dissociate before hydrolysis can proceed. This series of reactions enforces ATP hydrolysis on alternating GroEL rings. Broken rectangles represent unliganded, and solid rectangles represent ADP-liganded GroEL rings respectively. Both species have a high affinity for non-native protein substrate. ATP liganded rings with low affinity for non-native protein substrate are represented by ovals. Note that this cycle represents the system at low ATP concentrations where the possibility of generating an ATP14 complex is avoided.
6. GROES-GROEL INTERACTIONS AND THEIR EFFECTS ON THE HYDROLYSIS OF ATP 6.1. Binding of GroES to the ADP and ATP States of GroEL While apo-GroEL and GroES do not interact, the addition of adenine nucleotides generates two types of GroEL: GroES complex observable by electron microscopy. Depending on the reaction conditions, the presence of ATP can create either an asymmetric structure with GroES bound to only one of the two GroEL toroids (Saibil et al., 1993; Chen et al., 1994; Roseman et al., 1996), or a symmetrical complex with GroES bound at each end of the GroEL tetradecamer (Llorca et al., 1994; Azem et al., 1995; Llorca et al., 1997; Behlke et al., 1997). With ADP only the former, asymmetric complex is formed (Roseman et al., 1996; Xu et al., 1997). In an attempt to understand the mechanistic features of GroEL/GroES interactions and to elucidate their functional relevance, the dynamic and equilibrium properties of such complexes, formed in the presence of each nucleotide, have been studied in considerable detail. The association of GroES with pyrene-labelled GroEL in the presence of ADP leads to a large enhancement of the extrinsic fluorescence which can be used to report the formation of the asymmetric (1:1) protein-protein complex (Jackson et al., 1993). This process occurs slowly, yielding a bimolecular rate constant of 1×105 M-1 s-1. Incubation of the components at micromolar ADP concentrations requires many hours to allow the GroEL-GroES interaction to reach equilibrium. In the equilibrated system the binding affinity of ADP in the GroEL: GroES complex is extremely high with a dissociation constant below 70 nM and a stoichiometry of 7. Thus, although the GroEL: ADP7: GroES complex forms very slowly, it is extremely stable. This conclusion is confirmed by the low rate of dissociation of this complex (4×10-3 s-1) (Burston et al., 1995) and the low rate of exchange of radio-labelled ADP and GroES (3×10-5 s-1) (Todd et al., 1994). However, in the presence of ATP the formation of the GroEL: GroES complex is much faster. Shown in Figure 8 is a stopped-flow fluorescence experiment in which a subsaturating concentration of ATP (20 M) is mixed with pyrene-labelled GroEL in the presence and absence of GroES. With no GroES in the system, the transient obtained is described by a single rate constant of 2.8 s-1 representing the rate of rearrangement of GroEL to the R-conformation at low ATP concentration. Addition of either 0.375 or 3 M GroES introduces a new phase into the fluorescence enhancement which occurs at 16s-1, i.e. it is rapid and
Molecular chaperones and folding catalysts
552
insensitive to the concentration of the co-chaperonin. This result is compatible with the following reaction scheme:
Figure 8 Stopped flow fluorescence of the binding of ATP and GroES to pyrenyl-GroEL. Sub-saturating concentrations of ATP (20 M) cause a sub-maximal fluorescence enhancement in pyrenyl-GroEL. Subsequent binding of GroES is then reported by a greater enhancement of fluorescence as discussed in the text.
GroES must bind to the GroEL14: ATP7 complex very rapidly and before the transition to the R-conformation to yield GroEL14: ADP7 GroES7. The rate of rearrangement of this complex, to give the high fluorescence GroEL (R)14: ATP7: GroES7 form, is faster than in the absence of GroES and is not limited by the prior association of GroES, since increasing its concentration does not alter the kinetics. The co-protein must associate with the GroEL14: ATP7 complex at a rate faster than 16 s-1 in these conditions. This corresponds to a collision rate constant in excess of 4×107 M-1 s−1 (i.e. >16 s−1 / (0.375×10−6 M)) which approaches the upper limit for protein diffusion in aqueous solution and is at least two orders of magnitude faster than the formation of the GroEL14: ADP7: GroES7 complex. Although these kinetic observations appear somewhat arcane, they may have considerable bearing on the dynamics of the ATPase cycle and its relationship to the folding mechanism. Firstly they show that GroES has a kinetic preference for binding to an ATP ring, and secondly that the co-protein can bind before the rearrangement of subunits which leads to the low protein affinity R-conformation. With respect to the overall chaperonin mechanism this provides a crucial ‘trapping’ reaction in which a
The ATPase cycle of the GroE molecular chaperones
553
stably bound protein substrate could be encapsulated by GroES before being displaced into the cavity by the ATP-driven rearrangement of the GroEL apical domains. Although kinetic experiments shed no light on whether the toroid structure associated with the formation of the binary complex (GroEL14: ATP7) is different from that found in the ternary species (GroEL14: ATP7: GroES7), it should be noted that the structure of GroEL in these two complexes appears different under the electron microscope (Chen et al., 1994; Roseman et al., 1996) (see Burston and Saibil, this volume). In keeping with the ability of GroES to bind to GroEL prior to the hydrolysis of ATP and to influence conformational transitions, GroES has significant effects on the measured ATP binding properties of GroEL as shown by enhanced Hill coefficients in both steady-state and equilibrium binding studies (Gray & Fersht, 1991; Jackson et al., 1993; Kovalenko et al., 1994). 6.2. ATP Hydrolysis in the Presence of GroES When GroEL, GroES and ATP are mixed and the rate of formation of total ADP/ Pi. (both free and bound) is measured, there is a first-order ‘burst’ of ATP hydrolysis occuring at a rate of 0.12 s-1 with an amplitude equal to 7 sites per GroEL oligomer, followed by a steady-state phase which proceeds at rate of 0.042 s-1 per active site, i.e. half the sites in the oligomer (Todd et al., 1994; Burston et al., 1995). Thus, despite the presence of GroES in the complex, hydrolysis initially occurs at the same rate as is seen in the absence of co-protein, implying that a new, non-hydrolytic ratelimiting step has been introduced into the reaction cycle. To confirm this, isolated single-turnover reactions were designed to examine the next hydrolytic step. The transient phase of the above experiment describes the rate of ATP hydrolysis by a GroEL ring to which GroES is bound; it thus reports the rate of conversion of the GroEL14: ATP7: GroES7 complex to produce GroEL14: ADP7: GroES7. When this product complex was challenged with a further 7 ATP molecules to form the ATP7: GroEL14: ADP7: GroES7 species, hydrolysis of ATP on the ring distal to GroES occurred in a single-exponential phase with a rate constant of 0.12 s-1 (Burston et al., 1995). Hence, the hydrolytic step in the presence of GroES, whether on the ring proximal or distal to the co-chaperonin, occurs at a rate of 0.12 s-1. This rate is identical to the rate of hydrolysis in GroEL alone, yet the steady-state rate of the hydrolytic cycle in the GroEL: GroES system is two-to-three times slower. The fact that the rate-determining step is not observed when the isolated turnover reactions are carried out, shows that either a product release step or a structural rearrangement is the slowest process in steady-state hydrolysis in the presence of the co-protein. The relationship of this step to the rate of dissociation of GroES is discussed below. 6.3. The Dissociation of GroES During the ATPase Cycle Lorimer and colleagues have carried out a comprehensive study of the GroEL ATPase cycle (Todd et al., 1994) which provides extensive information on the lifetimes of a number of complexes containing GroEL, GroES and nucleotide. They show, by exchange of radio-labelled ADP and GroES, that the asymmetric GroEL14: ADP7: GroES7 complex has a half-life of approximately five hours. However, the addition of ATP reduces its
Molecular chaperones and folding catalysts
554
half-life to about 15 seconds, commensurate with the turnover of a single ring of ATP. Addition of the non-hydrolysable ATP analogues ATP S or AMP-PNP also promotes complex dissociation, although somewhat more slowly (t1/2 of ~60 seconds). Their results show that the dissociation of GroES
Figure 9 The rate of decay of GroEL: GroES complexes followed by pyrene fluorescence. Decay is promoted by either ATP hydrolysis (k=0.042 s−1) or by addition of alkaline phosphatase (k=0.004 s−1). For discussion see text. Adapted from Burston et al., 1995.
and ADP from a GroEL ring is linked to the binding and/or hydrolysis of ATP on the opposite ring. Further, if ATP is bound on the GroES-associated ring of GroEL it is committed to hydrolysis, i.e. the rate of cleavage is faster than the rate of dissociation, and the GroES remains tightly associated through the hydrolytic step. Similarly, the rate of dissociation of the chaperonin complex has been determined directly by pyrene fluorescence (Burston et al., 1995). The addition of a sub-saturating concentration of ATP to a solution of pyrene-labelled GroEL leads to an increase in fluorescence intensity which is further enhanced by the addition of GroES (see Figure 9). The ATPase reaction then reaches a steady state with a high-fluorescence GroEL: GroES: nucleotide complex predominating. If a large excess of unlabelled GroEL is then added, any GroES dissociating from the labelled protein will rebind to the unlabelled GroEL which then serves as an optically silent GroES trap. As shown in Figure 9, the rate of loss of fluorescence fits to a single-exponential decay at a rate of ~0.04 s−1, and the fluorescence returns to that of the GroEL: ATP complex. This rate is the same as the steady state velocity of ATP hydrolysis in the presence of GroES. If, instead of adding the unlabelled trap, all free and rapidly dissociating (uncommitted) ATP and ADP is destroyed by the addition of a large concentration of alkaline phosphatase, the GroEL: GroES complex decays very slowly. Hence, in the steady state reaction, GroES obligatorily dissociates once per cycle at a rate coincident with the rate-limiting step and
The ATPase cycle of the GroE molecular chaperones
555
this requires binding of ATP to the ring distal to GroES. 6.4. Bullets and Footballs One of the more controversial questions concerning the GroE hydrolytic cycle is the nature of the GroES exchange intermediate, i.e. it is indisputable that GroES binds to and dissociates from GroEL on each turnover of the steady-state ATPase reaction, but it is an attractive proposition that this process is mediated by a quasi-symmetrical GroES7: GroEL14: GroES7 complex in the following way:
In such a mechanism GroES is forced to exchange from one ring of GroEL to the other through an intermediate which is symmetrical with respect to the co-chaperonin but not the nucleotide. This process would open and close each cavity on successive rounds of ATP turnover. Evidence from electron microscopy shows that two types of GroEL: GroES are formed. The asymmetrical “bullet shaped” complex with GroES bound to one of the two available GroEL rings was observed in the first electron micrographs of GroE complexes (Hendrix, 1979), and has since been extensively characterized (Chen et al., 1994; Roseman et al., 1996; Xu et al., 1997). More recently, a symmetrical complex shaped like an American football with GroES bound to each end of the GroEL tetradecamer (Llorca et al., 1994; Schmidt et al., 1994; Llorca et al., 1997) has been imaged. Hence, the fact that both complexes exist is not in question, but the significance of the “football” intermediate has been the subject of intense debate, with evidence both for and against the symmetrical GroES7: GroEL14: GroES7 complex being an important intermediate in the biologically functional GroE reaction cycle. Two lines of evidence have been cited in favour of a symmetrical intermediate. Firstly, an analysis of the rates of refolding of several chaperonin dependent substrate proteins as a function of the GroEL: GroES stoichiometry has been completed (Azem et al., 1995). This study concludes that the rate of refolding is maximal in conditions where the symmetrical complex is found, as judged by negative stain electron microscopy and glutaraldehyde cross-linking studies. Secondly, a comprehensive study of the kinetics of the chaperonin-assisted refolding of barnase shows that GroES: GroEL stoichiometries in excess of 1 introduce a faster phase of folding which rises to 50% of the total folding amplitude at a 2:1 stoichiometry (Corrales & Fersht, 1996a). Against this, when the dissociation of GroES in steady-state hydrolysis is measured using a GroES trap (Burston et al., 1995) the rate of 0.04 s-1 is unaltered when the GroES: GroEL stoichiometry is reduced below 1:1. Additionally, in this experiment all free GroES is rapidly absorbed onto the trap, thus preventing rebinding to the labelled GroEL. Similarly, surface plasmon resonance studies by Hartl and colleagues (HayerHartl et al., 1995) indicate that if a stable GroEL14: ADP7: GroES7 complex is formed, addition of ATP causes dissociation of the complex via hydrolysis on the GroEL ring to which GroES is not bound, without binding of a second molecule of GroES. The latter pair of experiments show that the formation of a “football” intermediate is not an obligatory step in discharging a GroES ring during the ATPase cycle, i.e. they
Molecular chaperones and folding catalysts
556
demonstrate the existence of the following reaction:
However, the results do not show that such a structure is not formed as an exchange intermediate when GroES is in excess. In conclusion, the mechanistic significance of the symmetrical GroES7: GroEL14: GroES7 intermediate remains somewhat unclear, and further experiments are required if its role in assisted protein folding is to be fully elucidated. In particular it is necessary to examine the mechanism in reaction conditions which mimic the ATP: ADP and GroEL: GroES ratios and concentrations within the cell. Recent experiments with a GroEL mutant in which hydrolysis of ATP occurs very slowly have clarified the nature of the step in which GroES is released (Rye et al., 1997). When the ADP stabilized GroEL-GroES complex (GroEL: ADP7: GroES) is challenged with ATP, the co-protein is released at a much faster rate than the ATP can be hydrolysed. Hence GroES dissociation is driven by the binding of ATP to the opposite ring and is mediated by the consequent structural rearrangement of GroEL. Given the experimental results described in the above sections, it is tempting to propose a concensus model for the ATPase reaction mechanism at the macromolecular level. Such a model is forced to account for the following phenomena: (1) Asymmetry of nucleotide interactions. (2) The action of ADP as a non-competitive inhibitor of ATP hydrolysis. (3) The obligatory dissociation and reassociation of GroES in the hydrolytic cycle. (4) The role of GroES in committing bound ATP to undergo hydrolysis. (5) The ability of ATP to drive GroES dissociation from the opposite ring prior to its hydrolysis. (6) The kinetic preference for GroES to bind to an ATP-occupied ring. The reaction scheme shown in Figure 10 represents a minimal model for the reaction steps which constitute the GroE reaction cycle. 7. THE COUPLING OF PROTEIN AND NUCLEOTIDE BINDING AFFINITIES 7.1. The Effects of ATP, ADP & AMP-PNP on Protein Binding Affinity The affinity of GroEL for protein substrates in different stages of folding has long been a matter of intense interest. An early study concluded that apo-GroEL had
The ATPase cycle of the GroE molecular chaperones
557
Figure 10 The GroEL ATPase cycle in the presence of GroES. Binding of ATP (T) to an unliganded GroEL ring (rectangle) promotes a conformational shift to a state with low affinity for non-native protein (oval) and promotes GroES dissociation from the opposite ring. Binding of GroES occurs rapidly to a GroEL: ATP ring creating the likely acceptor state for non-native protein. A round of ‘committed’ hydrolysis occurs to yield a stable GroEL: ADP7: GroES complex. Subsequent binding of ATP promotes dissociation of GroES from the distal ring and rebinding of GroES to the ATP occupied ring would lead to encapsulation of the protein substrate. Hydrolysis of the
Molecular chaperones and folding catalysts
558
bound ATP and rebinding of ATP to the opposite ring would lead to substrate release. As in Figure 7 this represents the system at low ATP concentrations. It is possible that at high ATP concentrations GroES: GroEL: ATP7 complex should be added to the group of species (shown in square brackets) which accumulates before the hydrolytic step.
the highest affinity for the most unfolded intermediate on the refolding pathway of an oligomeric enzyme, and a much lower affinity for the late, compact, monomeric folding intermediate which immediately precedes dimerization (Badcoe et al., 1991). A considerable body of evidence has now been accumulated which suggests that GroEL recognizes a reasonably well-structured and rapidly formed folding intermediate often referred to as the “molten globule” state (e.g. Martin et al., 1991; Hayer-Hartl et al., 1994). However, the definition of which protein states are bound tightest may be somewhat arbitrary and case-dependent, since an incisive study of the conformation of barnase bound to GroEL using proton-deuterium amide exchange kinetics measured by NMR (Zahn et al., 1996) concludes that both the fully unfolded and the partially structured intermediate state interact tightly. Most would agree that any conformational state which has extensive exposure of hydrophobic surface will bind with high affinity. Estimates of the dissociation constants for complexes between non-native proteins and apo-GroEL vary from 10–12 M for large polypeptides up to 10−8 M for small ones (Badcoe et al., 1991; Todd et al., 1994; Staniforth et al., 1994; Lilie & Buchner, 1995; Sparrer et al. 1996; Zahn et al., 1996). Regardless of which type of pre-native state is recognised by GroEL, the effect which nucleotide binding has on the protein binding affinity has been studied extensively. One of the continuing themes of this discussion of the GroEL ATPase cycle has been the cooperative binding of ATP triggering a shift from a T-state with a low affinity for ATP, to an R-state with a high affinity. This rearrangement causes an accompanying change in the protein-binding affinity of GroEL, with the T-state having high, and the R-state having low affinity for the unfolded protein (Badcoe et al., 1991; Jackson et al., 1993). This proposal has been examined by measuring the affinity of GroEL for lactate dehydrogenase (LDH) through the retardation of the substrate folding rate (Staniforth et al., 1994a). The affinity of GroEL for unfolded LDH is tightest to the apo and the ADP states, is weaker to GroEL: AMPPNP and weakest to GroEL: ATP (see Table 1). This is entirely in accordance with the proposal that the affinity of GroEL for protein and ATP are reciprocally coupled. Accordingly, the GroEL ATPase cycle will enforce a switch between weak and tight protein binding state. The conclusion that ATP triggers a weakening of substrate binding by two or three orders of magnitude has also been made by Buchner and colleagues using surface plasmon resonance as a probe for complex formation (Lilie & Buchner, 1995; Sparrer et al. 1996). The presence of GroES further weakens substrate binding, thus potentiating the effect of ATP (Staniforth et al., 1994; Schmidt et al., 1994).
The ATPase cycle of the GroE molecular chaperones
559
Table 1 Role of nucleotide and GroES on the affinity of GroEL for unfolded LDH
Nucleotide None
−GroES
+GroES 5 nM
5 nM
ADP (4 mM)
18 nM
25 nM
AMP-PNP (4 mM)
53 nM
83 nM
ATP
118nM
439 nM
The practical consequences of this nucleotide-induced modulation of protein binding affinity is to facilitate a cycle of binding and release of protein substrates from GroE complexes. The release of substrate has been characterized by using mutant GroEL protein traps (Fenton et al., 1994) which bind but do not release non-native proteins, even in the presence of ATP and GroES. These experiments show that rhodanese and ornithine transcarbamylase (OTC) are expelled from the chaperonin in a non-native state (Weissman et al., 1994) and that many cycles of binding and dissociation must occur before the folded, chaperonin-independent state is acquired. Similar findings have been made by monitoring the distribution of isotopically labelled rubisco after a single round of ATP hydrolysis (Todd et al., 1994), by using chemically modified GroEL (Taguchi & Yoshida, 1995; Ranson et al., 1997), and by analysis of the kinetics of rhodanese refolding (Smith & Fisher, 1995). 7.2. The Effect of Unfolded Protein on Co-operativity Within the GroEL Oligomer The analysis of co-operativity in the GroEL ATPase by Horovitz and co-workers has been advanced still further by directly examining the effect of unfolded protein on these co-operative transitions (Yifrach and Horovitz, 1996). The stimulation of the GroEL ATPase by addition of non-native protein has been observed previously (Martin et al., 1991; Jackson et al., 1993). However, these measurements were made by addition of non-native protein which then proceeded to refold, and resulted in a highly complicated kinetic system. In experiments using a stably unfolded protein substrate, the effect of protein binding on the steady-state ATPase activity of GroEL has been determined. If the binding of one ligand (e.g. ATP) to a protein, in a co-operative manner, affects the affinity of the protein for a second ligand (unfolded protein), then the reverse should also be true. This is precisely the outcome they observed; stably unfolded -lactalbumin (in a reduced, calcium depleted form) binds preferentially to the T state and pushes the equilibrium back
Molecular chaperones and folding catalysts
560
from the RR state towards the TR and TT states. The fact that T state rings have a higher ATPase activity means that an unfolded substrate will enhance the rate of turnover. This is a possible mechanism for adapting the GroEL molecule to a role in responding to cellular stress; as the concentration of misfolded protein increases, binding to GroEL is made more likely. 8. THE COUPLING OF THE ATPase CYCLE TO CHAPERONIN-ASSISTED PROTEIN FOLDING So far, we have concentrated on the kinetic properties of the ATPase cycle and the characteristic complexes which arise therein. In this section we consider potential mechanisms for the coupling of ATP turnover to protein folding in a way which will enhance the folding yield. Rapid progress has recently been made in elucidating how the nucleotide-driven switches in affinity, for both GroES and unfolded substrates, may account for the ability of GroE to force protein substrates to fold productively and two main models of the GroE mechanism now predominate. 8.1. Models of Chaperonin Action 8.1.1. Encapsulation and Folding In an encapsulation model the unfolded substrate protein enters the central cavity of GroEL and remains there until committed to fold. It was originally proposed that DHFR and rhodanese reached the native state while associated with the GroEL-GroES complex owing to the inability of casein to block refolding when the GroES capping protein was included in the reaction (if substrate proteins were leaving GroEL in the reaction cycle, then the presence of casein should block refolding by blocking the rebinding of substrate protein to the binding sites of GroEL) (Martin et al., 1991). Several experiments localised the binding of substrate protein to the central cavity of GroEL (Langer et al., 1992; Braig et al., 1993) and further experiments led to the proposal of a detailed model of chaperonin action which involved rebinding and release within this cavity until the native state or near-native state was attained, after which the polypeptide was released into bulk solution (Martin et al., 1993). An electron microscopy study visualised an expansive GroES-capped GroEL cavity and introduced the term “Anfinsen cage” to describe protein folding within this protected space in which the substrate is prevented from making incorrect intermolecular interactions which lead to irreversible aggregation (Martin et al.,
The ATPase cycle of the GroE molecular chaperones
561
1991; Saibil et al., 1993). 8.1.2. Unfolding and Release The unfolding model arose from the argument that GroEL has its highest affinity for unfolded states of proteins and that protein-protein binding energy may be used to disrupt non-native interactions (Hubbard & Sander, 1991; Jackson et al., 1993; Todd et al. 1994; Ranson et al., 1995; Chan & Dill, 1996, Sfatos et al., 1996). In this hypothesis, an enhancement of folding yield could be achieved by actively unfolding misfolded proteins in a tight binding phase, while in the weak phase they are released to allow them to refold spontaneously. Hence, if a population of unfolded molecules is transferred to folding conditions, a proportion will fold to a highly organised native-like state which does not interact tightly with GroEL and goes on to fold spontaneously. The rest becomes trapped in unproductive, misfolded conformations and interact tightly with GroEL by virtue of greater exposure of hydrophobic residues. Having been bound they are unfolded, and then released in an ‘untrapped’ state and given a second chance to acquire a native-like structure. Repetition of this cycle will deliver all molecules to their native structure. Several lines of evidence support this proposal, including the switch in affinity between tight and weak protein binding states (Jackson et al., 1993), the release of non-native protein by GroEL (Weissman et al., 1994; Todd et al., 1994; Taguchi & Yoshida, 1995; Burston et al., 1996; Ranson et al., 1997) and the degree of enhancement of amide exchange in GroEL: substrate complexes (Zahn et al., 1994; Zahn et al., 1996). However, both proposals are attractive in that they account for the absence of substrate specificity in GroEL and explain the enhancement of yield without the penalty of slowing down the folding reaction unduly (see section 2). Also, both mechanisms require an input from ATP binding and hydrolysis; in the encapsulation mechanism it provides the energy to capture the substrate and open and close a GroES-capped cavity while in the unfolding/release hypothesis this energy is required to bind, unfold and discharge. 8.2. Towards a Global Model for GroE-mediated Folding Recent developments have begun to tie together these seemingly disparate models, and produce a more unified view of the action of GroE (for reviews see Ellis & Hartl, 1996; Clarke, 1996; Hartl, 1996). By a series of order-of-addition and protease digestion/protection experiments, the productive complex in chaperonin-assisted protein folding has been isolated as being a as complex of GroEL, GroES and substrate protein, with the GroES and substrate protein located on the same side of the GroEL tetradecamer (Weissman et al., 1995). They have also subsequently shown that substrate protein can reach the native state in this complex, by using mutant GroEL molecules which cannot release GroES once it has bound (Weissman et al., 1996). These results have been mirrored by those of Hartl and colleagues who have shown that a monomeric substrate protein can reach the native state whilst covalently tethered within the GroEL cavity (Mayhew et al., 1996). This conclusion is underlined by a kinetic analysis of the refolding of mMDH at a range of concentrations of GroEL: GroES. These results show that the GroEL: GroES complex to which unfolded mMDH binds in the assisted, ATP-driven reaction has a high
Molecular chaperones and folding catalysts
562
substrate affinity with an effective KM of ~10-8 M and a stoichiometry of 1:1. At an mMDH concentration of 1 M and a GroEL: GroES concentration in the range 0 to 1 µM the observed rate of folding increases owing to the disruption of misfolded enzyme in the form of low molecular weight aggregates (Ranson et al., 1995). The rate of folding then remains constant up to 40 M GroEL: GroES. At these elevated concentrations of chaperone, only 0.025% of mMDH is in free solution yet there is no diminution of the observed folding rate (Ranson et al., 1997). Recent work (Rye et al., 1997) has confirmed that both mMDH and bacterial rubisco reach a native-like state within the chaperonin cage, and oligomerise upon release to form the native, active enzyme complex. These observations provide convincing evidence that (1) the substrate protein can reach the folded, monomeric state whilst associated with GroEL and (2) the nature of the association allows rapid folding i.e. the substrate must be encapsulated rather than bound, so that folding is not inhibited by tight interactions between GroEL and the unfolded state. The previously reported release of non-native protein is also seen in the mMDH system with a dissociation rate of 0.04 s−1. That is, the complex ejects both GroES and the substrate protein once every cycle of ATP hydrolysis (Burson et al., 1996; Ranson et al., 1997). At first sight, it appears somewhat paradoxical that mMDH can fold to a committed, monomeric state in association with the GroE complex, yet is continually being ejected into bulk solution to be rebound by GroE to go through further cycles of encapsulation and obligatory ejection. However, such a mechanism will prevent covalently ‘damaged’ substrates (which are unable to fold) from saturating the GroE apparatus by remaining stably bound, rather than being degraded through proteolysis (Weissman et al. 1996; Mayhew et al., 1996; Ranson et al., 1997). The general dynamics of the binding, encapsulated folding and ejection cycle with respect to the renaturation of mMDH (Ranson et al., 1997) are shown in Figure 11. GroEL, therefore, appears to bind its substrate proteins in trans to GroES within a high-affinity ring; the presence of GroES would occlude the binding site for substrate protein on the cis ring (Chen et al., 1994; Fenton et al., 1994). This binding step will afford an opportunity for incorrect interactions to be reversed (Jackson et al., 1993; Todd et al., 1994; Ranson et al., 1995; Todd et al., 1996; Walter et al., 1996; Corrales & Fersht, 1996; Zahn et al., 1996a&b). Subsequent ATP binding leads to a reduction in substrate binding affinity and the rapid association of GroES. This displaces the substrate from its binding surfaces and encapsulates it in a cis complex (Weissman et al., 1995). The encapsulation phase may proceed through a symmetrical intermediate with GroES bound to each end of the GroEL oligomer (Llorca et al., 1994; Schmidt et al., 1994; Todd et al., 1994; Azem et al., 1995) or via ATP-driven dissociation and rebinding of GroES (Burston et al., 1995; Hayer-Hartl et al., 1995)—see Figure 10. The substrate will then dwell in the cavity and have an opportunity to fold without inhibitory contacts with hydrophobic surfaces on the apical domains (Weissman et al., 1995; 1996; Mayhew et al., 1996; Burston et al., 1996; Ranson et al., 1997). The proportion of molecules able to fold per ‘visit’ will be dependent solely on the ratio of the lifetime of the cis complex and the rate of protein folding. In the case of mMDH ~3% of polypeptide chains fold to a committed, GroE-independent state per cycle. For fast folding subunits this percentage may become quite large. For example, the subunits
The ATPase cycle of the GroE molecular chaperones
563
which constitute the glutamine synthetase dodecamer com-mit to the folded state at a rate of ~0.09 s−1 (Fisher & Yuan, 1994), so that 70% will fold in a single cycle. Further turnover of ATP forces the cavity to open with the expulsion of GroES, ADP and the protein substrate whether committed or not (Burston et al., 1996; Ranson et al., 1997; Rye et al., 1997). The possible fates of ejected, non-native protein are diverse. It can rebind to another (or the same) chaperonin molecule to attempt another round of encapsulated folding. It also has an opportunity to fold in free solution depending on the time between encounters with GroE. At physiological
Figure 11 The dynamics of mMDH binding, encapsulation and folding by GroEL.
concentrations of GroE, however, the transit time between one encounter with GroE and the next is much shorter than the dwell time in the encapsulated state. Thus owing to the fact that sponatneous folding occurs at an intrinsic and fixed rate in either environment, a minority of molecules will reach their native states when free in solution. Thirdly, the release of non-native substrate protein ensures that damaged proteins which can never refold can be released from GroEL into the bulk solvent, rather than rapidly “outtitrating” the cell’s chaperonin capacity upon cellular stress. Once in solution, they can partition to the proteolytic apparatus of the cell, a process in which GroEL is been implicated in vivo (Kandror et al., 1994). 9. CONCLUDING REMARKS This discussion of the GroEL ATPase cycle and its connection with assisted protein folding began with a question: What is the role of ATP binding and hydrolysis in aiding a thermodynamically favourable process? The answer lies in the enforced rearrangements of GroEL which are imposed by its differential interaction with ATP and ADP. These
Molecular chaperones and folding catalysts
564
serve two purposes: firstly, the binding and hydrolysis of ATP provides energy to drive formation and disruption of the chaperonin cage in which assisted-folding takes place (as summarized in Figure 12). Secondly, it provides energy to disrupt the tight interaction between substrate protein and GroEL which
Figure 12 Coupling of the ATPase cycle with substrate binding and ejection. The substrate is bound by an open end of an asymmetric GroEL: GroES complex (step (I)) where the acceptor ring is most likely to be in a tight binding, ADP-induced conformation. ATP hydrolysis on the opposite ring (step (ii)) then allows ATP to associate with the substrate-containing site. This has three consequences (step (iii)); it forces dissociation of ADP and GroES from the opposite ring, it weakens interactions with the substrate and allows reassociation of GroES on the substrate side to displace the unfolded protein molecule into the cavity where productive folding can occur. ATP is then hydrolysed (step (iv)) a process which then allows ATP to associate with the trans ring and eject GroES, ADP and the protein substrate from the opposite ring to complete the cycle.
The ATPase cycle of the GroE molecular chaperones
565
is used to reverse incorrect interactions. GroES has three major roles. It co-ordinates ATP hydrolysis in GroEL, strengthening co-operativity and therefore potentiating the conformational shift between states with a high and low affinity for substrate protein. Secondly it commits ATP to hydrolysis and serves to maintain a reciprocating cycle of reactions. Thirdly, its binding to GroEL displaces substrate protein into a cavity, which GroES itself is essential in forming. The question of whether the major function of GroE in the cell is to disrupt misfolded states or to encapsulate sticky folding intermediates, thereby preventing aggregation, is difficult to answer unequivocally. Both activities have been demonstrated with different substrates and both activities may be inevitable consequences of the mechanism described here. 10. ACKNOWLEDGEMENTS The authors would like to thank the Wellcome Trust for Programme Grant Support for their work. A.R.Clarke is a Lister Institute Research Fellow. 11. REFERENCES Azem, A., Diamant, S., Kessel, M., Weiss, C. and Goloubinoff, P. (1995). The protein refolding activity of chaperonins correlates with the symmetrical GroEL14: (GroES7)2 heteroligomer. Proc. Natl. Acad. Sci. USA. , 92 , 12021–12025. Badcoe, I.G., Smith, C.J., Wood, S., Halsall, D.J., Holbrook, J.J., Lund, P. and Clarke, A.R. (1991). Binding of a chaperonin to the folding intermediates of lactate dehydrogenase. Biochemistry , 30, 9195–9200. Behlke, J., Ristau, O. and Schonfeld, H.J. (1997). Nucleotide dependent complex formation between the Escherichia coli chaperonins GroEL and GroES studied under equilibrium conditions. Biochemistry , 36, 5149–5156. Bochkareva, E.S. and Girshovich, A.S. (1994). ATP induces non-identity of two rings in chaperonin GroEL. J. Biol. Chem ., 269 , 23869–23971. Bochkareva, E.S. Lissin, N.M., Flynn, G.C., Rothman, J.E. and Girshovic, A.S. (1992). Positive cooperativity in the functioning of chaperonin GroEL. J. Biol. Chem ., 267 , 11637–11644. Braig, K., Otwinowski, Z., Hegde, R., Boisvert, D.C.Joachimiak, A., Horwich, A.L. and Sigler, P.B. (1994). The crystal structure of the bacterial chaperonin GroEL at 2.8Å. Nature , 371, 578–586. Braig, K., Simon, M., Furuya, F., Hainfeld, J.F. and Horwich, A.L. (1993). A polypeptide bound by the chaperonin GroEL is localized within a central cavity. Proc. Natl. Acad. Sci. USA , 90 , 3978–3982. Buchner, J., Schmidt, M. Fuchs, M. Jaenicke, R., Rudolph, R., Schmid, F.X. and Kiefhaber, T. (1991). GroE facilitates refolding of citrate synthase by supressing aggregation. Biochemistry , 30, 1586–1591. Burston, S.G., Weissman, J.S., Farr, G.W., Fenton, W.A. and Horwich, A.L. (1996). Release of both native and non-native proteins from a as-only GroEL ternary complex.
Molecular chaperones and folding catalysts
566
Nature , 383, 96–99. Burston, S.G., Ranson, N.A. and Clarke, A.R. (1995) The origins and consequences of asymmetry in the chaperonin reaction cycle. J. Mol. Biol. , 249 , 138–152. Chan, H.S. and Dill, K.A. (1996). A simple model of chaperonin-mediated proteinfolding. Prot. Struc. Func. Genet. , 24 , 345–351. Chandrasekhar, G.N., Tilley, K., Woolford, C., Hendrix, R. and Georgopoulos, C. (1986). Purification and properties of the groES morphogenetic protein of Escherichia coli. J. Biol. Chem. , 261 , 12414–12419. Chen, S., Roseman, A.M., Hunter, A.S., Wood, S.P., Burston, S.G., Ranson, N.A., Clarke, A.R. and Saibil, H.R. (1994). Location of a folding protein and shape changes in GroEL-GroES complexes imaged by cryo-electron microscopy. Nature. , 371, 261– 264. Clarke, A.R. (1996). Molecular chaperones in protein folding and translocation. Curr. Opin. Struct. Biol. , 6, 43–50. Corrales, F.J. and Fersht, A.R. (1996a). Kinetic significance of GroEL14. (GroES7)2 complexes in molecular chaperone activity. Folding and Design , 1, 265–273. Corrales, F.J. and Fersht, A.R. (1996b). Towards a mechanism for GroEL: GroES chaperone activity, an ATPase-gated and ATPase-pulsed folding and annealing cage. Proc. Natl. Acad. Sci. USA , 93, 4509–4512. Diamant, S., Azem, A., Weiss, C. and Goloubinoff, P. (1995). Effect of free and bound magnesium and manganese ions on the ATPase activity of chaperonin GroEL14. Biochemistry , 34 , 273–277. Ellis, R.J. and Hartl, F.U. (1996). Protein folding in the cell: competing models of chaperonin function. FASEB. J. , 10, 20–26. Fayet, O., Ziegelhoffer, T. and Georgopoulos, C. (1989). The GroES and GroEL heat shock gene products of E. coli are essential for bacterial growth at all temperatures. J. Bacteriol. , 171, 1379–1385. Fenton, W.A., Kashi, Y., Furtak, K. and Horwich, A.L. (1994). Residues in chaperonin GroEL required for polypeptide binding and release. Nature , 371, 614–619. Fisher, M.T. and Yuan, X. (1994). The rates of commitment to renaturation of rhodanese and glutamine synthetase in the presence of the GroE chaperonins. J. Biol. Chem. , 269, 29598–29601. Fisher, M.T. (1992). Promotion of the in vitro renaturation of dodecameric glutamine synthetase from Eschericia coli in the presence of GroEL (chaperonin-60) and ATP. Biochemistry , 31 , 3955–3963. Georgopoulos, C.P., Hendrix, R.W., Casjens, S.R. and Kaiser, A.D. (1973). Host participation in bacteriophage head assembly. J. Mol. Biol. , 76, 45–60. Goloubinoff, P., Christeller, J.P., Gatenby, A.A. and Lorimer, G.H. (1989). Reconstitution of active dimeric ribulose bisphosphate carboxylase from an unfolded state depends on two chaperonin proteins and ATP. Nature , 342, 884–889. Gray, T.E. and Fersht, A.R. (1993). Refolding of barnase in the presence of GroE. J. Mol. Biol , 232, 1197–1207. Gray, T.E. and Fersht, A.R. (1991). Co-operativity in ATP hydrolysis by GroEL is increased in the resence of GroES. FEBS Lett ., 292, 254–258. Hartl, F.U. (1996). Molecular chaperones in cellular protein folding. Nature , 381, 571–
The ATPase cycle of the GroE molecular chaperones
567
580. Hayer-Hartl, M.K., Martin, J. and Hartl, F.U. (1995). Asymmetrical interaction of GroEL and GroES in the ATPase cycle of assisted protein folding. Science , 269, 836–841. Hayer-Hartl, M.K., Ewbank, J.J., Creighton, T.E. and Hartl, F.U. (1994). Conformational specificity of the chaperonin GroEL for the compact folding intermediates of lactalbumin . EMBO J. , 3, 3192–3202. Hendrix, R.W. (1979). Purification and properties of groE, a host protein involved in bacteriophage assembly. J. Mol. Biol. , 129, 375–392. Horovitz, A., Bochkareva, E.S., Kovalenko, O. and Girshovic, A.S. (1993). Mutation Ala2ÆSer estabilizes intersubunit interactions in the molecular chaperone GroEL. J. Mol. Biol. , 231, 58–4. Hubbard, T.J.P. and Sander, C. (1991). The role of heat-shock and chaperone proteins in protein folding—possible molecular mechanisms. Prot. Eng. , 4, 711–717. Itzhaki, L.S., Otzen, D.E. and Fersht, A.R. (1995). Nature and consequences of GroELprotein interactions. Biochemistry , 34 , 14581–14587. Jackson, G.S., Staniforth, R.A., Halsall, D.J., Atkinson, T., Holbrook, J.J., Clarke, A.R. and Burston, S.G. (1993). Binding and hydrolysis of nucleotides in the chaperonin catalytic cycle: Implications for the mechanism of assisted protein folding. Biochemistry , 32, 2554–2563. Kad, N.M., Ranson, N.A., Cliff, M. and Clarke, A.R. (1998). Asymmetry, commitment and inhibition in GroE ATPase cycle impose alternating functions upon the two GroEL rings. J. Mol. Biol. , 278, 267–278. Kandror, O., Busconi, L., Sherman, M. and Goldberg, A.L. (1994). Rapid degradation of an abnormal protein in Escherichia coliinvovles the chaperones GroEL and GroES. J. Biol. Chem. , 269, 23575–23582. Langer, T., Pfeifer, G., Martin, J., Baumeister, W. and Hartl, F.U. (1992). Chaperoninmediated protein folding—GroES binds to one end of the GroEL cylinder, which accomodates the protein substrate within its central cavity. EMBO J. , 11 , 4757–4765. Llorca, O., Marco, S., Carrascosa, J.L. and Valpuesta, J.M. (1994). The formation of symmetrical GroEL-GroES complexes in the presence of ATP. FEBS Lett. , 345 , 181– 186. Llorca, O., Marco, S., Carrascosa, J.L. and Valpuesta, J.M. (1997). Symmetric GroELGroES complexes can contain substrate in both GroEL rings. FEBS Lett. , 405 , 195– 199. Martin, J., Mayhew, M., Langer, T. and Hartl, F.U. (1993). The reaction cycle of GroEL and GroES in chaperonin-assisted protein folding. Nature , 366 , 228–233. Martin, J., Langer, T., Boteva, R., Schramel, A., Horwich, A.L. and Hartl, F.U. (1991). Chaperonin-mediated protein folding at the surface of GroEL through a molten globule-like intermediate. Nature , 352, 36–42. Miller, A.D., Maghlaoui, K., Albanese, G., Kleinjan, D.A. and Smith, C. (1993). Eschericia coli chaperonins cpn60 (GroEL) and cpn10 (GroES) do not catalyse the refolding of mitochondrial malate dehydrogenase. Biochem. J. , 291, 139–144. Monod, J., Wyman, J., Changeux, J.P. (1965). On the nature of allosteric transitions: a plausible model. J. Mol. Biol , 12, 88–118. Okazaki, A., Ikura, T., Nikaido, K. and Kuwajima, K. (1994). The chaperonin GroEL
Molecular chaperones and folding catalysts
568
does not recognize apo-α-lactalbumin in the molten globule state. Nature Struct. Biol. , 1, 439–446. Ranson, N.A., Burston, S.G. and Clarke A.R. (1997). Binding, encapsulation and ejection; substrate dynamics in a chaperonin-assisted folding reaction. J. Mol. Biol. , 266, 565–664. Ranson, N.A., Dunster, N.J., Burston, S.G. and Clarke, A.R. (1995). Chaperonins can catalyse the reversal of early aggregation steps when a protein misfolds. J. Mol. Biol. , 250, 581–586. Rye, H.S., Burston, S.G., Fenton, W.A., Beechem, J.M., Xu, Z.H., Sigler, P.B. and Horwich, A.L. (1997). Distinct actions of cis and trans ATP within the double ring of the chaperonin GroEL. Nature , 388, 792–798. Saibil, H.R., Zheng, D., Roseman, A.M., Hunter, A.S., Watson, G.M.F., Chen, S., auf der Mauer, A., O’ Hara, B.P., Wood, S.P., Mann, N.H., Barnett, L.K. and Ellis, R.J. (1993). ATP induces large quaternary rearrangements in a cage-like chaperonin structure. Curr. Biol. , 3, 265–273. Schmidt, M., Rutkat, K., Rachel, R., Pfeifer, GL., Jaenicke, R., Viitanen, P.V., Lorimer, G.H. and Buchner, J. (1994). Symmetric complexes of GroE chaperonins as part of the functional cycle. Science , 265, 656–659. Sfatos, C.D., Gutin, A.M., Abkevich, V.I. and Shakhnovich, E.I. (1996). Simulations of chaperone-assisted protein folding. Biochemistry , 35, 334–339. Smith, K.E. and Fisher, M.T. (1995). Interactions between the GroE chaperonins and rhodanese—multiple intermediates and release and binding. J. Biol. Chem. , 270, 21517–21523. Staniforth, R.A., Burston, S.G. Atkinson, T. and Clarke, A.R. (1994a). Affinity of chaperonin-60 for a protein substrate and its modulation by nucleotides and chaperonin-10. Biochem. J. , 300, 651–658. Staniforth, R.A., Cortés, A., Burston, S.G., Atkinson, T., Holbrook, J.J. and Clarke, A.R. (1994b) The stability and hydrophobicity of cytosolic and mitochondrial malate dehydrogenases and their relation to chaperonin-assisted folding. FEBS Lett. , 344, 129–135 Taguchi, H. and Yoshida, M. (1995). Chaperonin releases the substrate protein in a form with a tendency to aggregate and ability to rebind to chaperonin. FEBS Lett. , 359, 195–198. Todd, M.J. and Lorimer, G.H. (1995). Stability of the asymmetric Escherichia coli chaperonin complex—guanidine chloride causes rapid dissociation. J. Biol. Chem. , 270, 5388–5394. Todd, M.J., Viitanen, P.V. and Lorimer, G.H. (1994) Dynamics of the chaperonin ATPase cycle: implications for facilitated protein folding. Science. , 265, 659–666. Todd, M.J., Viitanen, P.V. and Lorimer, G.H. (1993). Hydrolysis of adenosine-5'triphosphate by Escherichia coli GroEL: effects of GroES and potassium ion. Biochemistry , 32, 8560–8567. Viitanen, P.V., Lubben, T.H., Reed, J., Goloubinoff, P., O’ Keefe, O.P. and Lorimer, G.H. (1990). Chaperonin-facilitated refolding of ribulosebisphosphate carboxylase and ATP hydrolysis by chaperonin-60 (GroEL) are K+ dependent. Biochemistry , 29, 5665–5671.
The ATPase cycle of the GroE molecular chaperones
569
Walter, S., Lorimer, G.H. and Schmid, F.X. (1996). A thermodynamic coupling mechanism for GroEL-mediated unfolding. Proc. Natl. Acad. Set. USA , 93, 9425– 9430 Weissman, J.S., Kashi, Y., Fenton, W.A. and Horwich, A.L. (1994) GroEL-mediated protein folding proceeds by multiple rounds of binding and release of non-native forms. Cell. , 78, 693–702. Weissman, J.S., Hohl, C.M., Kovalenko, O., Kashi, Y., Chen, S., Braig, K., Saibil, H.R., Fenton, W.A. and Horwich, A.L. (1995). Mechanism of GroEL action: productive release of polypeptide from a sequestered position under GroES. Cell. , 83, 577–587. Weissman, J.S., Rye, H.S., Fenton, W.A., Beechem, J.M. and Horwich, A.L. (1996). Characterisation of the active intermediate of a GroEL-GroES mediated protein folding reaction. Cell , 84, 481–490. Xu, Z.H., Horwich, A.L. and Sigler, P.B. (1997). The crystal structure of the asymmetric GroEL-GroES-(ADP)7 chaperonin complex. Nature. , 388, 741–750. Yifrach, O. and Horovitz, A. (1994). Two line of allosteric communications in the oligomeric chaperonin GroEL are revealed by the single mutation Arg196<arrow>Ala. J. Mol. Biol. , 243, 397–401. Yifrach, O. and Horovitz, A. (1995). Nested co-operativity in the ATPase activity of the oligomeric chaperonin GroEL. Biochemistry , 34, 5303–5308. Yifrach, O. and Horovitz, A. (1996). Allosteric control by ATP of non-folded protein binding to GroEL. J. Mol. Biol. , 255 , 356–361. Zahn, R., Perrett, S., Stenberg, G. and Fersht, A.R. (1996a). Catalysis of amide protonexchange by the molecular chaperones GroEL and SecB. Science , 271 , 642–645. Zahn, R., Perrett, S. and Fersht, A.R. (1996b). Conformational states bound by the molecular chaperones GroEL and SecB—a hidden unfolding (annealing) activity. J. Mol. Biol. , 261 , 43–61. Zahn, R., Spitzfaden, C, Ottiger, M., Wuthrick, K. and Pluckthun, A. (1994). Destabilization of the complete protein secondary structure on binding to the chaperone GroEL. Nature , 368, 261–265.
23. THE RELATIONSHIP BETWEEN CHAPERONIN STRUCTURE AND FUNCTION STEVEN G.BURSTON1 and HELEN R.SAIBIL2 , * 1 Department
of Genetics Boyer Center for Molecular Medicine, Yale University School of Medicine, 295 Congress Avenue, New Haven, CT 06510, USA 2 Department of Crystallography, Birkbeck College, Malet Street, London WC1E 7HX, UK
1. Introduction 2. Architecture of the Oligomeric GroEL 2.1. Overall Structure of the GroEL Oligomer 2.2. Domain Structure of GroEL 2.3. The GroEL Central Cavity 3. Nucleotide Binding to GroEL 3.1. The Nucleotide Binding Site on GroEL 3.2. Cryo-electron Microscopy of Nucleotide-bound GroEL Structures 4. The Structure of the Co-chaperonin, GroES 4.1. GroES Monomer Structure 4.2. Structural and Chemical Properties of the GroES Heptamer 4.3. Subunit-subunit Interface 5. Protein-protein Interactions 5.1 . The Binding of Polypeptide Substrate to GroEL 5.2. The Interaction Between GroEL and GroES 6. Chaperonins from Archaebacteria and Mammalian Cytosol: TF55 Thermosome and CCT 7. Summary of the Functional Consequences of Chaperonin Structure 8. Acknowledgments 9. References *Corresponding author
The relationship between chaperonin structure and function
571
1. INTRODUCTION The intriguing structure of chaperonins, first indicated by electron microscopy (EM) studies in the late 1970’s, suggested that they would have a novel and fascinating mode of action. The interest of structural biologists and biochemists alike was ensured by the subsequent demonstration of their essential biological function of stabilizing unfolded, or partially folded states of a wide range of polypeptides, thus preventing irreversible loss through aggregation, and assisting their folding to a native conformation (Ellis & van der Vies, 1991; Jaenicke, 1993; Clarke, 1996a; Hartl, 1996) (see chapters by Welch et al. and Burkholder and Gottesman). However, the way in which the structural complexity of these macromolecules (approximate molecular mass ~lMDa) is organized to transduce the energy of ATP hydrolysis into the refolding process has presented an enormous challenge to researchers. This chapter will focus on the variety of structural data which has been collected, as well as discussing some of the site-directed mutagenesis which has identified critical regions in the chaperonin architecture. We will then attempt to place this structural data in the context of the mechanistic biochemistry which has elucidated many features of chaperonin function (see chapter by Ranson and Clarke, this volume). 2. ARCHITECTURE OF THE OLIGOMERIC GroEL GroEL (the Escherichia coli chaperonin-60 homologue) has a subunit molecular weight of 58.6kDa and was initially identified as a protein involved in bacteriophage assembly (Georgopoulos et al., 1973). It was later shown to interact transiently with newly synthesized proteins in vivo (Bochkareva et al., 1988; Horwich et al., 1993) as well as improve the refolding efficiencies of a wide variety of unfolded polypeptides in vitro (Goloubinoff et al., 1989; Martin et al., 1991; Buchner et al., 1991; Badcoe et al., 1991; Fisher, 1992). 2.1. Overall Structure of the GroEL Oligomer Negative stain EM revealed the architecture of the GroEL oligomer (Hendrix, 1979; Hohn et al., 1979), characterized by rectangular side views with four stripes of electron density and circular end views with seven-fold symmetry. The correct relationship between these two views was elucidated by Hutchinson et al., (1989) who collected EM images at different tilt angles and showed that GroEL was arranged as two double rings of seven subunits with a two-fold axis of symmetry perpendicular to the seven-fold axis. The presence of two layers of density in each ring suggested that each subunit was divided into two major domains. Three-dimensional reconstruction from negative stain electron micrographs of the GroEL homologue from Rhodobacter spheroides revealed the two major domains to be linked by a small bridge of density on the outside of the cylindrical oligomer, creating a cage-like structure with internal cavities (Saibil et al., 1993).
Molecular chaperones and folding catalysts
572
However, negative stain EM suffers from the limitation that the sample is dehydrated and thus flattened, and the stain used to contrast the structure does not allow imaging of internal features. These limitations have been overcome by the use of cryo-electron microscopy (cryo EM) in which a thin layer of solution on the grid is rapidly plunged into liquid ethane at liquid nitrogen temperature. The water surrounding the structure becomes vitrified instead of forming ice crystals, thus preserving the native protein without distortion, in a stable, frozen-hydrated state. The grid can then be viewed in the EM provided the grid temperature is maintained below –150°C (Dubochet et al., 1988). The low contrast, and hence signal-to-noise ratio, of these cryo-EM images is improved by averaging large numbers and combining them by tomographic reconstruction to provide three-dimensional (3D) density maps. This technique has been used to provide 3D reconstructions of apo-GroEL, GroEL bound to its nucleotide, unfolded polypeptide and co-chaperonin ligands and also the R197A mutant (Chen et al., 1994; Roseman et al., 1996; White et al., 1997). The initial hurdle in the X-ray determination of any structure is obtaining good crystals. Early trials revealed that the wild-type protein was difficult to crystallize. However, a double mutant of GroEL with Arg-13 changed to Gly and Ala-26 changed to Val produced the crystals suitable for X-ray diffraction (Braig et al., 1994). The crystals were orthorhombic with space group C2221 and had a single heptameric ring in the asymmetric unit. The molecular two-fold axis corresponded to the lattice twofold with the sevenfold perpendicular to this axis. Initial phases were obtained from a single isomorphous derivative to yield a 6 A map. Sevenfold averaging was then used to extend the phasing further to 2.7 Å. Some regions of disordered structure could not be wellresolved using these techniques because of the deviation from seven-fold symmetry in the crystal lattice. The X-ray structure (Braig et al., 1994) revealed the tetradecamer to be a porous cylinder of 146 A in length and 137 A in diameter confirming the earlier EM studies. Refinement of the structure without sevenfold symmetry improved the features, but some regions were still poorly defined because of disorder (Braig et al., 1995). The overall shape of the oligomer is shown in Figure 1(a). 2.2. Domain Structure of GroEL The subunits of GroEL are folded into three distinct domains as shown in Figure 1 (b). The large equatorial domain (residues 6–133 and 409–523) is predominantly -helical and accounts for all of the contacts between the two heptameric rings and most of the inter-subunit contacts within a ring. Also part of the equatorial domain are the mobile Nterminal and C-terminal residues which are crystallographically too disordered to resolve but appear to project into the central channel, probably providing a barrier between the central cavities of the two rings. The small intermediate domain (residues 134–190 and 377–408) consists of antiparallel chains linking the equatorial domain in the midplane of the GroEL cylinder to the apical domain at the end of the cylinder. An exposed region capable of hinge rotation is present at each end of the intermediate domain, at the boundaries with the other two domains (Figure 1b). Rotations about these hinges provide the large conformational change observed by EM during the functional cycle (see section 3 below; Roseman et al., 1996). Another notable feature is the diagonal projection of the intermediate domain
The relationship between chaperonin structure and function
573
relative to the equatorial plane which provides a site of contact between the intermediate domain of one subunit with the apical domain of the adjacent subunit on the right when viewed from the side. The contact contains a salt bridge between Arg-197 of the intermediate domains and Glu-386 on the apical domain of the adjacent subunit. This was revealed to be an important site of allosteric communication by characterization of the mutant Arg-197→Ala. This mutant GroEL showed a reduction in both positive and negative cooperativity of ATP binding and hydrolysis (Yifrach & Horovitz, 1994).
Figure 1 Left: Structure of the GroEL 14-mer. This image was created from the atomic structure of Braig et al. (1995) displayed as surface contoured electron density filtered to 25 Å resolution. One subunit is outlined, and the two inter-ring contacts are numbered. Right: Structure of a GroEL subunit in approximately the same orientation as the one outlined in (a). The domain structure, ATP binding site, contacts and hinge regions are indicated. Charged residues in the inter-ring contacts are shown in space-filling form, red for negatively charged and blue for positively charged residues. The space-filling groups shown in yellow are hydrophobic residues whose mutation abolishes substrate and GroES binding, and those in cyan affect GroES binding. The pink residue is Arg 197, which contacts Glu 386 in the intermediate domain of the neighbouring subunit. Mutation of any of the blue residues shown in ball and stick form in the intermediate domain causes global defects in function. Adapted from Figure 1 of Roseman et al., (1996) by permission of Cell Press.
The apical domain (residues 191–376) surrounds the entrance to the central cavity and
Molecular chaperones and folding catalysts
574
has been shown by electron microscopy (Chen et al., 1994) and site-directed mutagenesis studies (Fenton et al., 1994) to contain the binding site for both unfolded polypeptide chain and the co-chaperonin, GroES. This domain contains both and structures and has a great deal of flexibility, indicated by its high crystallographic temperature factors (B factors), particularly on top and in the regions lining the central cavity (Braig et al., 1995). 2.3. The GroEL Central Cavity Each heptameric ring of GroEL contains a central cavity within the cage-like struc-ture formed by the seven subunits of the ring. The diameter of the entrance to the cavity formed by the seven apical domains is 45 A and widens at the level of the intermediate domains. It appears to be continuous through the 146 A length of the cylinder. However, as mentioned above, the disordered N-terminal and C-terminal residues are likely to project from the equatorial domains into the channel thus creating a block between the rings. The N-terminus has also been shown to be crucial for subunit assembly (Horovitz et al., 1993). It has been estimated that a non-native protein such as a folding intermediate of Mr~20,000 could fit into the central cavity of GroEL (Braig et al., 1994), although it should be noted that binding of GroES and nucleotide results in an expanded but enclosed central cavity as described below (section 5.2). A recent cryo-EM structure of the unliganded GroEL showed asymmetry in which the central cavity of one ring was significantly more open than that of the opposite ring (Figure 3; Roseman et al., 1996). Another interesting feature of GroEL is the presence of portals on the side of the central channel (Figure 1a). These side windows are elliptical and are formed by the gap between the top of the equatorial domain, the crossing of the intermediate domain diagonal to the equatorial plane and the apical domain of the neighboring subunit to the right. They are 36 Å×13 Å at the widest point and 20 Å×10 at the narrowest (Braig et al., 1994). These allow solvent molecules and small ligands, such as nucleotides, access to the interior of the central channel even when GroES is bound to that ring. 3. NUCLEOTIDE BINDING TO GroEL The binding and hydrolysis of ATP by GroEL, and the resultant conformational changes which take place during this hydrolytic cycle are essential in order to drive the assisted folding reaction (Martin et al., 1991 & 1993; Jackson et al., 1993; Todd et al., 1994; Weissman et al., 1994 & 1996). The binding and hydrolysis of ATP by GroEL were observed to be positively co-operative within a heptameric ring (Gray & Fersht, 1991; Bochkareva et al., 1992; Jackson et al., 1993; Todd et al., 1993) and negatively cooperative between the two heptameric rings (Yifrach & Horovitz, 1994; Bochkareva & Girshovich, 1994; Burston et al., 1995). These two forms of cooperativity within the same protein complex have been described in a nested model of co-operativity (Yifrach & Horovitz, 1994 & 1995) in which the binding of ATP within a ring occurs in an “allor-none” fashion described by the Monod-Wyman-Changeux model of co-operativity (Monod et al., 1965), while the negative co-operativity between the rings is best described as a sequential Koshland-Nemethy-Filmer model (Koshland et al., 1966).
The relationship between chaperonin structure and function
575
Binding of ATP to a GroEL oligomer results in a conformational change in GroEL from the unliganded conformation which has a high affinity for unfolded protein substrates to a conformation with low substrate affinity (Jackson et al., 1993; Martin et al., 1991; Staniforth et al., 1994a; Sparrer et al., 1996; Yifrach & Horovitz, 1996) pointing to significant allosteric communication within the GroEL oligomer which is essential for its function. ADP appears to bind less cooperatively than ATP within a heptameric ring but with a high degree of negative cooperativity between the rings (Burston et al. 1995). 3.1. The Nucleotide Binding Site on GroEL The X-ray structure of the R13G/A26V GroEL double mutant with ATP S bound to each subunit has been solved to a resolution of 2.4 A (Boisvert et al., 1996). The overall architecture of the GroEL-ATPT S structure is essentially the same as that of unliganded GroEL in that it contains two heptameric rings sharing seven-fold rotational symmetry and a two-fold axis perpendicular to the seven-fold. The ATP-binding site is located at the top of the equatorial domain facing the central cavity. Figure 2 shows the amino acid residues in the equatorial domain which interact with the nucleotide analogue. Residues 87–91 (Asp-Gly-Thr-Thr-Thr) interact with the - and -phosphates of ATP (Boisvert et al., 1996) and lie on a highly conserved loop region (Kim et al., 1994). The specificity for adenosine is imposed via hydrogen bonds from adenine-N6 to the side-chain oxygen of Asn-479 and the backbone amide of Ala 480 to adenine-N1. ATP S also makes two interactions with GroEL mediated via metal ions. A magnesium ion chelates the non-bridging oxygen from each of the three phosphates and completes the octahedral coordination via the carboxylate of Asp-87 and two water molecules. A second metal ion also links the -phosphate to the carbonyl of Thr-30 and Lys-51. This is assumed to be either a K+ ion which decreases the cooperativity of ATP binding (Viitanen et al., 1990; Todd et al., 1993) or Ca2+, a component of the crystallization buffer. The peptide nitrogen of Gly-32 makes a hydrogen bond with both the 5'- -phosphate bridging oxygen and non-bridging oxygen of ATP S. Residues Gly-32, Thr-30 and Lys-51 are near the base of a stem-loop formed by -strands 2 and 3 ( 2/ 3 stem-loop) which interacts with the adjacent subunit by forming an extended b-sheet with the N- and C-terminal regions( -strands 1 and 18) of the neighboring subunit within the ring. Differences between the GroEL-ATP S crystal structure and the unliganded GroEL crystal structure are very small in contrast to the change in substrate polypeptide binding properties and large movements seen using cryo-EM (Chen et al., 1994; Roseman et al., 1996) and detected by fluorescence changes (Jackson et al., 1993; Burston et al., 1995). The reasons for the discrepancy are not known but may be a result of the high protein concentration in the crystal lattice. 3.2. Cryo-electron Microscopy of Nucleotide-bound GroEL Structures Negative stain EM of the ATP-free and ATP-bound states of GroEL (Saibil et al., 1993) revealed a conformational change upon binding ATP in agreement with observations made using an extrinsic fluorescence probe (Jackson et al., 1993). Subsequent cryo-EM studies showed that this structural change was due to a opening out of the apical domains
Molecular chaperones and folding catalysts
576
about 5–10° relative to the equatorial axis causing a slight elongation of the GroEL cylinder (Chen et al., 1994). More recent cryo-EM image reconstructions (Roseman et al., 1996) are shown in Figure 3. The nucleotide-bound structures show an opening and twisting of the apical domains and elongation compared to the unliganded structure. Interestingly, they also show significant deviation-from two-fold symmetry, providing a structural correlation for the biochemical evidence pointing to asymmetry in nucleotide binding and
Figure 2 The ATP binding site from the X-ray structure of Boisvert et al. (1996). The GroEL subunit is viewed from the opposite side from that shown in Figure 1 (b). The protein backbone is represented as a light blue ribbon with the phosphate binding loop (residues 87–91) in dark blue. The non-hydrolyzable analogue ATP S is shown as a balland-stick model using the following atom colours: blue (nitrogen); red (oxygen); yellow (sulphur); purple (phosphorus). and green (carbon). The magnesium ion which coordinates the phosphate oxygens as well as the carboxylate of Asp-87 is shown in gold. Some important amino acid residues involved in the binding of ATP are shown as yellow ball-and-stick side chains. Asp-87 and Thr-91 are at the extreme ends of the phosphate binding loop. Mutation of Asp-87 blocks ATP hydrolysis (Fenton et al., (1994) while Thr-91 is at the start of a helix which extends downwards to the inter-ring interface and is likely to be involved in allosteric communication between the rings. Asn-479 is involved in determining the specificity of adenosine as the preferred base while Ile-493 provides a hydrophobic locality for the purine. Figure produced using InsightII (BioSym Technology).
The relationship between chaperonin structure and function
577
hydrolysis (Yifrach & Horovitz, 1994; Bochkareva & Girshovich, 1994; Burston et al., 1995). The ATP-bound structure, from samples vitrified within 4 seconds of mixing, is indistinguishable from the steady-state ATP structure. (The rate-limiting step is the hydrolytic step (Jackson et al., 1993)). These results refine the previous negative stain (Saibil et al., 1993) and cryo-EM observations on GroEL-ATP (Chen et al., 1994) by providing more accurate and higher resolution three-dimensional information. The apical domains of the top ring are opened up in the presence of ATP, but the subunits of the lower ring are rotated inwards, showing the asymmetry between the rings resulting from negative cooperativity (Figure 3). The ADP-bound structure
Figure 3 3-D cryo-EM reconstructions of GroEL complexes. The top row shows the GroEL, GroEL-ADP and GroEL-ATP complexes. Both of the nucleotide bound structures are more extended than GroEL with the upper apical domains (see Figure 1) twisted outwards. The
Molecular chaperones and folding catalysts
578
bottom row shows the “bullet”-shaped GroEL-GroES-ADP, GroEl-GroESATP and the “American football”-shaped GroEL-GroES-AMP-PNP which can be observed under certain conditions. Compared with the top row it can be seen that GroEL adopts a completely different structure when bound to GroES, with the twist of the subunits reversing handedness. Additionally a large cavity is opened up underneath GroES in which protein substrates can be encapsulated. The contour levels for each structure was chosen to enclose the correct molecular volume. Each reconstruction was produced from an EM data set of approximately 1000 images. Figure adapted from Roseman et al. (1996) by permission of Cell Press.
shows the largest opening of the rings and is generally a more expanded structure. Superposition of unliganded GroEL, GroEL-ADP and GroEL-ATP shows a successive clockwise twisting of the apical domains which would cause the substrate binding site to be rotated out of the central cavity and become buried in the intersubunit interface (Figure 4a). These observations explain how nucleotides control the binding affinity of GroEL for non-native substrates, by controlling the accessibility of the hydrophobic binding sites (Roseman et al., 1996). A recent study of the GroEL mutant R197A has resolved the TR (ATP bound to one ring) and RR (ATP bound to both rings) states (White et al., 1997). The asymmetric GroEL-ATP complex in Figure 3 is in the TR state, and the RR state (not shown) has both rings in a very open conformation. Comparison of the EM reconstructions of GroEL-ATP with GroEL or GroEL-ADP showed a loss of density in the region of the inter-ring contacts (Figure 5; Roseman et al., 1996). This difference was found to be statistically signficant by comparing the data sets with a student’s t test. By superimposing the EM reconstruction over the X-ray structures, the altered region was identified as the inter-ring contact containing Lys-105 and Glu434. Since Lys-105 is directly connected via an -helix to the nucleotide phosphate binding residues 87–91, Roseman et al. (1996) propose that during ATP binding and hydrolysis small displacements of this helix may withdraw the lysine residue, thus weakening the interaction energy across the contact region. The other inter-ring contact, between Glu-461 and Arg-452, shows significant variation between GroEL-GroES-ADP and GroEL-GroES-ATP complexes. Moreover, mutation of Glu-461 to lysine blocks polypeptide release and decreases the stability of the GroEL-GroES interaction (Fenton et al., 1994). This contact may mediate inter-ring effects of nucleotide binding upon the GroES and polypeptide substrate binding sites. 4. THE STRUCTURE OF THE CO-CHAPERONIN, GroES GroES, a member of the Hsp10, or chaperonin-10 class of chaperones, is a heptamer of identical ~10 kDa subunits which acts as a co-protein in the functional cycle of GroEL (Chandrasekhar et al., 1986). The expression of GroES is essential for cell viability and has been shown in vitro to be critical for the folding of certain substrate polypeptides under conditions where little or no spontaneous folding is observed (Goloubinoff et al., 1989; Martin et al., 1991; Schmidt et al., 1994a). It has been observed in EM studies to
The relationship between chaperonin structure and function
579
bind tightly to one end of the GroEL cylinder in the presence of adenine nucleotides (Saibil et al., 1991; Langer et al., 1992; Chen et al., 1994). A second GroES binds weakly to the opposite ring of GroEL (Schmidt et al., 1994b; Llorca et al., 1994) in the presence of ATP ot AMP-PNP (see section 5.2). The X-ray crystallographic structure of GroES has been determined to a resolution of 2.8 Å (Hunt et al., 1996; Figure 6a). The protein crystallized in the orthorhombic P212121 space group with one heptamer in the asymmetric unit. Two heavy atom derivatives, one of which clusters to the acidic roof of the GroES dome (see below), were used to create an isomorphous replacement map which in turn
Figure 4 (a) Outlines of the top view of apical domains in GroEL, GroEL-ADP and GroEL-ATP. The grey bar shows the position of the hydrophobia binding site facing the central cavity, and the dark grey and black bars show how ADP and ATP binding respectively twist the domain so that the site becomes progressively occluded. (b) 3D reconstruction of GroEL-GroES-ATP with outlines of one subunit in GroEL, GroEL-ATP and in the GroES complex. The hinge rotation causing the apical domain twist is around 100° (compare to Figure 8a). Figures adapted from Roseman et al. (1996) by permission of Cell Press.
Molecular chaperones and folding catalysts
580
allowed the structure to be solved making use of seven-fold non-crystallographic symmetry averaging. It was then refined without seven-fold averaging. The structure of a second chaperonin-10 from Mycobacterium leprae has also been solved to 3.5 Å with seven-fold averaging (Mande et al., 1996) and shows essentially the same topology and physical characteristics as the E. coli version described by Hunt et al. (1996). 4.1. GroES Monomer Structure Each subunit is composed of an irregular anti-parallel -barrel which encompasses a hydrophobic core with a -hairpin which extends outwards at the top towards the
Figure 5 (a) GroEL-ADP and (b) GroEL-ATP, aligned with the atomic structure of the interring contact regions of GroEL. The EM reconstructions are shown as wire frame surfaces (white) to reveal the atomic models inside. One inter-ring contact is at the front centre of each structure and is seen as solid density in the ADP reconstruction and as a hole in the ATP reconstruction. The positively and negatively charged residues in the contacts are shown in blue and red, respectively. The helix (green) connecting the central contact (Lys105) to the ATP binding site (Thr-91) extends diagonally outward from the contact, with an ATP molecule (purple) bound in each site. (Reproduced from Roseman et al., (1996) by permission of Cell Press.)
center of the ring. Electron density corresponding to a region termed the “mobile loop” region (residues 16–33), which had been shown previously by NMR studies to interact with GroEL in the presence of ADP (Landry et al., 1993), was only visible in one of the seven subunits. The mobile loop density of this one ring formed a -hairpin sandwiched between the wall of one GroES heptamer and the roof of another in the crystal lattice. Lys-34 (Lys-36 in M. leprae numbering) which has been shown to modulate allosteric transitions in GroEL (Kovalenko et al., 1994) lies adjacent to this mobile loop region and
The relationship between chaperonin structure and function
581
is likely to interact directly with GroEL (Mande et al., 1996). 4.2. Structural and Chemical Properties of the GroES Heptamer The GroES heptamer resembles a dome-shaped cap which is 30Å high and 70–80 Å in diameter. The inside of the dome is approximately 20 Å high and 30 Å in diameter (Hunt et al., 1996). Subunits are arranged such that the -barrels of each monomer are parallel to each other and to the seven-fold axis. The inter-subunit contacts are between the first -strand in each subunit and the -barrel of the adjacent subunit. The -hairpin structures of each monomer which extend towards
Molecular chaperones and folding catalysts
582
Figure 6 (a) A side view of GroES with the backbone displayed in ribbon form. The mobile loop is shown for the subunit in which electron density for this region could be seen. Electron density could not be
The relationship between chaperonin structure and function
583
seen for the mobile loop region (residues 16–32) in the other subunits; therefore the discontinuity at residue 15 and residue 33 is indicated with an (L) on the foremost subunit. The N- and C-terminus are also marked on two subunits to distinguish them from the ends of mobile loop. (b) A top view of GroES in ribbon format. The one mobile loop region is again marked as the N- and C-terminus of one subunit. The -hairpin of each subunit which extends toward the top centre can be clearly seen. Both figures were produced using the program RIBBONS (Carson, 1987).
the top and center of the dome contain two glutamate residues (Glu-50 and Glu-53) at the tip of the hairpin, resulting in a dense region of negative charge (Figure 7a). A significant amount of disorder was observed in this region, presumably due to charge repulsion of these clustered acidic side chains. It is not clear how these interactions are stabilized at physiological pH and ionic strength in the bacterial cell. It has been postulated that this flexibility at the roof of the dome may allow non-native proteins to leave the central cavity of GroEL without dissociation of GroES (Hunt et al., 1996). However, biochemical evidence to support such a hypothesis has not been reported. The opening facing the GroEL cavity on the underside of the dome contains the highly conserved residue Tyr-71, forming a region of hydrophobicity near the point at which GroES interacts with GroEL (Figure 7b). These seven tyrosine residues could be available for interaction with the unfolded polypeptide substrate when sequestered in the cis cavity (i.e. on the same ring of GroEL as GroES). The crystal structure indicates a high degree of conformational flexibility in these residues, consistent with a possible role in solvating a dynamic polypeptide folding within the central cavity of GroEL. 4.3. Subunit-subunit Interface Substantial deviation of the GroES heptamer from seven-fold symmetry was observed in the crystal structure (Hunt et al., 1996). Although the hydrophobic core of the -barrel deviated little from seven-fold symmetry, the packing of adjacent subunits was highly irregular due to structural plasticity at the interface between subunits. The principal interaction across the interface is between the first -strand of one subunit with the final -strand of the adjacent subunit. If seven C-terminal residues are cleaved from GroES by limited proteolysis in urea, the cleaved monomers are unable to oligomerize (Scale & Horowitz, 1995). The long, charged side chains of Lys-74, Glu-76 and Arg-37 also serve to maintain a flexible subunit-subunit interface by weakening the interaction between subunits. It has been observed that GroES heptamers dissociate into monomers at ~10-8 M oligomers (Zondlo et al., 1995) corresponding to a binding energy of about –7 kcal mol-1 of interface. This flexibility may be present to allow conformational changes in GroES induced by
Molecular chaperones and folding catalysts
584
Figure 7 A surface map of the GroES oligmer produced by GRASP (Nicholls et al., 1991). The positive and negative surface potential is represented by blue and red, respectively. The 8 A diameter orifice at the top of the oligomer can be seen surrounded by a high degree of negative charge provided by the two glutamate residues on the hairpin of each subunit.
movements of the GroEL domains to which it is bound, and which may be critical to the folding of a protein substrate in the cis ring of the GroEL-GroES complex. 5. PROTEIN-PROTEIN INTERACTIONS 5.1. The Binding of Polypeptide Substrate to GroEL Central to a structural description of the chaperonin-assisted folding mechanism is the identification of the site and physical nature of the interaction between GroEL and unfolded polypeptide substrate. Since an X-ray crystallographic approach is hampered by the fact that unfolded polypeptides do not have seven-fold symmetry and may be disordered in a crystal lattice, efforts have focused upon electron microscopy and sitedirected mutagenesis. The first attempt to locate the site of the bound polypeptide was made using negative
The relationship between chaperonin structure and function
585
stain electron microscopy to study GroEL bound to either unfolded rhodanese or unfolded alcohol dehydrogenase (Langer et al., 1992). End views of an average of ~650 images revealed some extra, poorly defined electron density in the region of the opening of the central cavity which was assigned to bound, unfolded polypeptide. However, because of the difficulty in identifying a bound polypeptide chain using this method, Braig et al. (1993) took the approach of covalently labelling dihydrofolate reductase (DHFR) with colloidal gold particles in order to localize the protein substrate in EM images. When the unfolded, gold-labelled DHFR was bound to GroEL, gold particles were observed in the centres of end views, corroborating the notion that GroEL binds substrates in its central cavity. Side views showed that some GroEL molecules could bind a gold-labelled DHFR inside the central cavity of both rings simultaneously, indicating that unliganded GroEL contains a potential binding site for polypeptide on each of its two rings. Binding of IPMDH was observed in the trans ring (opposite GroES) by labelling with an antibody to the substrate (Ishii et al., 1994). A more direct view of the bound substrate density was obtained using cryo-EM of mitochondrial malate dehydrogenase (mMDH) bound to GroEL (Chen et al., 1994). This study showed a region of additional electron density which could be assigned to the MDH located in one end of the GroEL cylinder between the apical domains and bulging outwards into the solvent. The location and physical characteristics of the interaction between polypeptide chain and GroEL has been examined using site-directed mutagenesis (Fenton et al., 1994). A large number of site-directed mutants of GroEL were screened for their effects upon the binding and refolding of unfolded ornithine transcarbamylase (OTC), a substrate that requires GroES to fold efficiently (Zheng et al., 1993). The mutants were also screened for GroES binding, ATPase activity and ability to rescue E. coli cells where expression of the chromosomal GroEL had been totally repressed. It was noted that a number of mutants abolished the ability to bind unfolded OTC. These mutants all mapped to hydrophobic residues in the apical domain lining the central cavity (e.g. Y199E, Y203E, F204E, L234E, L237E—see Figure 1b). Subsequently the interaction between GroEL and either a mutant of subtilisin, BPN′, which is unable to fold, or -casein, a disordered protein, has been analyzed by scanning microcalorimetry (Lin et al., 1995). Both proteins were able to bind to GroEL with micromolar affinity and showed a negative heat capacity upon binding which are characteristic of a hydrophobic interaction. The conformation of GroEL-polypeptide binary complexes has been probed using amide hydrogen/deuterium exchange, NMR and electrospray mass spectroscopy. Studies using the scrambled three-disulphide -lactalbumin (Robinson et al., 1994) and DHFR (Gross et al., 1996; Goldberg et al., 1997) suggested that the bound polypeptide was highly labile but showed a small degree of protection from amide hydrogen exchange in regions of the polypeptide with a high degree of protection in the folded structure. Although these studies indicated the possible existence of native-like secondary structure when bound to GroEL it was impossible draw a definitive conclusion. More recently the crystal structure of an apical domain (residues 191–376) with a 17 residue N-terminal peptide tag has been solved (Buckle et al., 1997), in which the N-terminal tag appears to be mimicking an unfolded polypeptide substrate by binding to the peptide binding site of a neighbouring apical domain in the crystal lattice. The binding surface maps to the same
Molecular chaperones and folding catalysts
586
region as those identified by site-directed mutagenesis (Fenton et al., 1994), specifically Leu-234, Leu-237 and V-264, and would appear to extend back to residues Tyr-199, Tyr203 and Phe-204. The interactions with the apical domain are predominantly hydrophobic in nature although there is some hydrogen bonding between the main chain of the tag and side chains of the apical domain. Whilst the tag is essentially in an extended conformation there appears to be significant flexibility in the polypeptide binding site in order to accommodate any general hydrophobic surface, such as that of molten globuletype non-native proteins. 5.2. The Interaction Between GroEL and GroES GroES is able to form a stable interaction with GroEL only in the presence of adenine nucleotides (Chandrasekhar et al., 1986). When the complex was viewed in the electron microscope GroES appeared to cap one end of the GroEL cylinder (Saibil et al., 1991; Langer et al., 1992; Ishii et al., 1992). It was also noted that the presence of GroES inhibited the steady-state ATP hydrolytic rate of GroEL by approximately 40–50% in the presence of K+ ions (Gray & Fersht, 1991; Jackson et al., 1993; Todd et al., 1993). This was originally thought to result from half-sites reactivity due to inhibition of the ATPase in the ring bound to GroES. However, the finding that ATP binding and hydrolysis are asymmetric in GroEL alone (Yifrach & Horovitz, 1994; Bochkareva & Girshovich, 1994) led to the conclusion that the presence of GroES is in fact altering the rate-limiting step in the ATP hydrolytic cycle of GroEL (Burston et al., 1995). The 1:1 stoichiometry of GroES to GroEL in the presence of ADP has been confirmed by protease protection (Langer et al., 1992) and fluorescence studies (Jackson et al., 1993). However, it was noted by negative stain EM that in the presence of ATP and analogues a significant number of GroEL cylinders were found with a GroES oligomer bound to each end in the shape of an American football (Schmidt et al., 1994). The functional significance of these “football” structures remains a mystery (Lorimer, 1997), although their presence as a transient species in the GroEL functional cycle has been detected (Corrales & Fersht, 1996). The site of interaction between GroES and GroEL has been visualised by cryo-EM. The GroEL ring bound to GroES shows an approximately 60° opening between the apical and equatorial domains such that GroES is able to interact with the ends of the apical domains that were previously oriented towards the central cavity (Chen et al., 1994). The most striking feature of this complex is that the central cavity becomes enlarged approximately three-fold compared to that in GroEL alone, and capped, such that a protein of molecular weight < 60,000 could be encapsulated (Clarke, 1996b). Secondly, the site of interaction between GroEL and GroES was in the same region of GroEL to which unfolded MDH was bound (Chen et al., 1994). Close coupling of sites for substrate and GroES was demonstrated by the construction of site-directed mutants in the apical domain (e.g. Y203E, L237E, V263S) which affected the binding of both GroES and unfolded OTC (Fenton et al., 1994; Figure 1b). The region of GroES which interacts with GroEL has also been identified by NMR. Landry et al. (1993) found that a one-dimensional 1H-NMR spectrum of GroES contained sharp peaks on a background of broad resonances. The broad resonances were
The relationship between chaperonin structure and function
587
expected since GroES is an oligomer of 70 kDa and has too slow a rotational correlation time in free solution to give rise to sharp signals. The sharp peaks indicated therefore that GroES contained some regions of polypeptide which are more mobile than the rest of the structure. They assigned the sharp peaks to residues 17–32 (VETKSAGGIVLTGSAA) of GroES, which they termed the “mobile loop”, and showed that it was susceptible to proteolysis. When GroES was associated with GroEL in the presence of ATP the linewidths of the mobile loop were broadened and the loop was no longer protease sensitive. Given that a number of mutants of GroES which were unable to support growth of bacteriophage 1 also mapped to this region (Georgopoulos et al., 1973), this provided strong evidence that the mobile loop region interacts directly with GroEL. A peptide corresponding to this region was synthesized and shown to bind to GroEL. Since the exchange rate of this peptide between free and GroEL-bound forms was sufficiently rapid then transfer Nuclear Overhauser Effects (trNOEs) could be used to monitor binding and conformation of the bound state. The conformation of the peptide was determined to be a hairpin loop and could be competed off with intact GroES. The overall conformational changes of GroEL-GroES complexes during the hydrolytic cycle have recently been examined (Figure 3; Roseman et al., 1996). Thin bridges of electron density were observed connecting GroES and GroEL, suggesting that the seven mobile loops of the GroES heptamer may be responsible for most of the interaction with the GroEL apical domains. The GroES-GroEL complexes show a dramatic prolongation of the apical domain twisting observed in the GroEL-ATP cryo-EM structure. Indeed the change in twist is so large that the subunit is reversed in handedness, suggesting that the exposed surface in the central cavity which interacts with bound polypeptide substrate may be very different (Figure 4b). In the GroEL-GroES complex formed with ADP, the during apical domains have a slightly different tilt compared to that formed in the presence of ATP. This shifts the site of interaction with GroES slightly and distorts the GroES subunit interfaces. The apical domains of the open trans-ring in the presence of ADP show a twisting outwards when compared to those in complexes formed with ATP (Figure 3). Changes in the exposure of amino acids in the interior of the cavity upon GroES binding and nucleotide binding and hydrolysis may well be critical structural changes during the functional cycle of GroE. The crystal structure of the asymmetric GroEL7•GroEL7•ADP7•GroES7 has recently been determined (Xu et al., 1997; Figure 8a). The conformation of the cis ring forming the enclosed cavity reveals details of the domain movements generally predicted from the cryo-EM (Roseman et al., 1996). The interface between GroES and GroEL occurs via residues identified by previous mutational analysis (Fenton et al., 1994) and NMR (Landry et al., 1993). In particular, Leu-234, Leu-237, Leu-259, Val-263 and Val-264 of GroEL, which lie on two helices which have shifted to the top of the subunit in comparison to the unliganded structure, and interact with residues Ile-25, Val-26 and Leu-27 of GroES. The twisting and raising of the GroEL apical domain to form the GroEL•GroES interaction site dramatically changes the interaction surface available to the polypeptide. In addition to residues 234, 237 and 264 being recruited to interact with the co-chaperonin, six other residues (Tyr-
Molecular chaperones and folding catalysts
588
The relationship between chaperonin structure and function
589
Figure 8 (a) Side view of GroEL-GroES-ADP (Xu et al., 1997). Domains are coloured red (equatorial), green (intermediate) and yellow (apical). The GroES co-chaperonin (green) is bound to the top ring of GroEL. An equatorial (blue), intermediate (light blue) and apical domain (violet) have been highlighted for comparison with the side view of the thermosome (Figure 9a). Reproduced from Ditzel et al. (1998) by permission of Cell Press. (b) Schematic drawing of the chaperonin ATP-mechanism based upon the thermosome structure (Ditzel et al., 1998). The thermosome numbering is used and the equivalent GroEL residue number is shown in parentheses in the following. The -phosphate of ATP is stabilized via hydrogen bonds to the conserved Thr-96 (Thr-89) and 97 (Thr-90) and coordination to the magnesium ion. The carboxylate side chains of Asp-63 (Asp-52) and Asp-390 (Asp-398) polarize the water molecule which makes a nucleophilic attack on the phosphorus atom. The resultant pentavalent intermediate subsequently collapses to liberate the orthophosphate and ADP products. Figure reproduced from Ditzel et al., (1998) by permission of Cell Press. (c) The modified nucleotide binding pocket in the GroEL-GroES-(ADP)7 crystal structure. The protein backbone is represented as a light blue ribbon with the phosphate-binding loop (residues 87–91) in dark blue. The bound ADP is shown as a ball-and-stick model using the following atom colours: blue (nitrogen); red (oxygen); purple (phosphorus) and green (carbon). The magnesium ion which coordinates the nucleotide phosphate oxygens as well as the carboxylate side chains of Asp-87 and Asp398 is shown in gold. In
Molecular chaperones and folding catalysts
590
comparison with Figure 2 (from the GroEL-ATP S structure) the insertion of Asp398 from the intermediate domain into the nucleotide binding pocket can be seen. Mutation of this residue to alanine resulted in a drastic reduction of ATP hydrolytic activity (Rye et al., 1997).
199, Ser-201, Tyr-203, Phe-204, Leu-259, and Val-263) become buried either between apical domains or along the interface with GroEL. This removal of the polypeptide binding site upon ATP and GroES binding provides a means of releasing the polypeptide into the GroEL central cavity. The residues which now line the wall of the central cavity are polar in nature, thus providing an environment which strongly favours intra-molecular interactions within the polypeptide rather than a hydrophobic interaction between the non-native polypeptide and the chaperonin. A further important aspect of the conformational change upon binding GroES and nucleotide, not detected in the low resolution cryo-EM, was observed. The intermediate domain rotates downwards to interact with the equatorial domain, thus effectively locking the nucleotide into its binding pocket. As a consequence of this movement Asp-398 moves down into the nucleotide binding site such that the carboxylate of the aspartate side chain coordinates with the Mg2+ ion (Figure 8c). Replacement of this aspartate residue with alanine results in the 60-fold reduction in ATP hydrolytic activity (Rye et al., 1997) demonstrating the importance of this domain movement for the hydrolysis of ATP. Xu et al. (1997) also propose that the allosteric communication across the ring-ring interface is due to an en bloc movement of the trans equatorial domains in a complimentary fashion to tilting movements of the cis equatorial domains, thus preserving the integrity of interactions across the ring-ring interface and avoiding steric clashes. 6. CHAPERONINS FROM ARCHAEBACTERIA AND MAMMALIAN CYTOSOL: TF55 THERMOSOME AND CCT A second subfamily of chaperonins, weakly related to GroEL, has been discovered in archaebacteria (e.g. Sulfolobus shibitae, Sulfolobus solfataricus), namely TF55 (Trent et al., 1991) and its homologue the thermosome (Phipps et al., 1991 & 1993), and in the eukaryotic cytosol, CCT (for reviews see Willison & Kubota, 1994; Willison & Horwich, 1996; chapter Willison in this volume). By negative stain EM it could be seen that TF55 had a similar overall architecture to that of GroEL with two stacked rings (Trent et al., 1991; Marco et al., 1994a; Knapp et al., 1994). However, in contrast to GroEL, TF55 contains two nine-membered rings, while the related thermosome from Pyrodictium, Thermoplasma and Archaeoglobus all have eight subunits per ring (Phipps et al., 1991). Additionally, both TF55 and the thermosome have two distinct, although closely related subunits (Knapp et al., 1994; Kagawa et al., 1995). They bind unfolded polypeptides in vitro and have ATPase activity similar to GroEL. However, no GroES homologue has been found to function with this class of chaperonins. Although TF55 and the thermosome are only very weakly related to GroEL they do show ~40% homology with subunits of CCT and it has been hypothesized that the thermosome is the evolutionary ancestor of CCT (Willison & Horwich, 1996).
The relationship between chaperonin structure and function
591
The function of the eukaryotic cytosolic chaperonin CCT, appears to be more specific than that of either GroEL or TF55. It has been found to mediate the folding of - and tubulin, actin, and Ga-transducin in vitro and in vivo (Yaffe et al., 1992; Lewis et al., 1992; Gao et al., 1992; Sternlicht et al., 1993; Farr et al., 1997) and can assist the folding of firefly luciferase in vitro (Frydman et al., 1992). At the level of primary structure CCT is significantly more complex than either GroEL or TF55 having up to 9 different polypeptide subunits. Each complex consists of two stacked hetero-oligomeric rings and top views show 8-fold symmetry by negative stain EM (Lewis et al., 1992; Marco et al., 1994b). Electron microscopy of CCT suggests that it has the same overall subunit architecture as GroEL. Side views show four stripes corresponding to the apical domains and equatorial domains of each ring (Lewis et al., 1992; Gao et al., 1992; Frydman et al., 1992). The central cavity of CCT appears larger than that of GroEL, measuring 60 Å in diameter as opposed to 45 Å. Electron micrographs of CCT complexed with bound actin revealed the substrate localized to the central cavity (Marco et al., 1994b). The primary amino acid sequences of GroEL and CCT have been compared and show a great deal of similarity at the N- and C-termini (Kim et al., 1994), especially among the conserved residues of the ATP binding pocket (Boisvert et al., 1996). These regions make up the equatorial domain of GroEL (Braig et al., 1994) suggesting that the equatorial domains of CCT may be very similar to those of GroEL. The location of the ATP binding site on CCT subunits has been supported by the results of mutagenesis in this region which lead to tubulin deficiency and temperature sensitivity (Miklos et al., 1994; Ursic et al., 1994). However, comparison of the sequences corresponding to the apical domains shows little or no homology indicating that the divergent properties of CCT and GroEL result largely from the differences in these domains. Like TF55, CCT performs its function without requirement of a GroES homologue although a number of co-factors have been identified (Gao et al., 1993; Tian et al., 1996). Like GroELs, CCT requires ATP hydrolysis to refold its substrates and substrates undergo cycles of binding and release as they do with GroEL (Tian et al., 1995; Farr et al., 1997). 6.1 High-resolution Structural Studies of the Thermosome Recently, X-ray structures of an isolated apical domain of the -subunit of the thermosome (Klumpp et al., 1997) and the complete thermosome (Ditzel et al., 1998) from the archaeon Thermoplasma acidophilum have been determined. The overall thermosome structure consists of two stacked octameric rings which are composed of alternating and subunits in a ( )4 arrangement and encloses a central cavity similar to GroEL (Figures 9a and b). However, the overall shape of the oligomer is more spherical than the GroEL cylinder, with a height of 158 A along the pseudo 8-fold axis and a diameter of 164 Å along the equatorial 2-fold axes. The sequence identity between the and subunits is 60% and they share a 46% sequence similarity with the eubacterial chaperonin GroEL. Each subunit adopts the same basic fold as a GroEL subunit with an equatorial, intermediate and apical domain. The main difference between the evolutionary diverse structures lies in the apical domain, as might be expected given that the group II chaperonins function with no GroES homologue. Compared to the GroEL subunit structure the thermosome apical domains
Molecular chaperones and folding catalysts show a 35 residue deletion which comprises helices
592
The relationship between chaperonin structure and function
593
Figure 9 Thermosome structure, (a) Side view of the thermosome. The domains are coloured in red (equatorial), green (intermediate) and yellow (apical). The 16 molecules of bound ADP are also shown bound to the equatorial domains in yellow. The stacking of subunits in register forming and pairs across the inter-ring interface can be seen. (b) Top view of the thermosome. -subunits are coloured red or violet and the -subunits are coloured yellow. The two strands of each subunit which lie in the centre and the N-terminal half of the helix which projects away from the centre back into the main core of the apical domain comprise the lid segment. Movements of the lid segment upon twisting of the apical domains during the thermosome functional cycle may open and close the entrance to the central cavity. (c) Comparison of the configuration of the lid segments from the structures of the thermosome (red; Ditzel et al., 1998) and the isolated apical domain (green; Klumpp et al., 1997). It has been speculated that in the polypeptide acceptor state could be generated by opening of the entrance to the central cavity, and that the lid segment may adopt the same fold as that found in the isolated apical domain. This would result in a clustering of hydrophobic residues (whose side chains are shown) which may form a polypeptide binding site.
Molecular chaperones and folding catalysts
594
(a) and (b) reproduced from Ditzel et al. (1998) by permission of Cell Press, (c) reproduced from Horwich & Saibil (1998) by permission of McMillan.
K and L in GroEL and an insertion of 28 residues which introduces two -strands and extends the N-terminus of helix H. This insertion in the thermosome subunits protrudes towards the pseudo 8-fold axis at the ends of the complex providing an obstruction to the entrance to the central cavity termed the lid domain (Ditzel et al., 1998). The structure of each subunit within the complex resembles that of a subunit in the cis-ring of the GroELGroES-ADP complex (Xu et al., 1997) in that a segment of the intermediate domain has closed down to form the upper part of the nucleotide binding site. The intra-ring contacts of the equatorial domains are conserved between the group I and group II chaperonins. However, the inter-ring contacts are different. The thermosome has its two rings in register producing pairs and pairs (Figure 9a), in contrast to the staggered arrangement of rings in GroEL (see Figure 8a). Of considerable interest is the structure of the lid domain which arches over to block the entrance to the central cavity of each ring. The X-ray structure of the isolated apical domain (residues 214–365) region (Klumpp et al., 1997) shows the same core structure as that in the intact thermosome (Figure 9c). However, the lid segment in the isolated domain has a helix-turn-helix structure whereas in the intact thermosome the same region is strand-strand-turn-helix (Figure 9c). The lid protrusions in the intact thermosome have their 8 hydrophobic side chains buried in side-to-side packing contacts. However, the helix-turn-helix structure of the isolated domain clusters these hydrophobic residues together. It has been speculated (Ditzel et al., 1998) that in a polypeptide-acceptor conformation of the thermosome a counter-clockwise rotation of the apical domains relative to their position in the crystal structure (believed to be the ‘folding-active’ state) could open the lid and give access to the binding sites, perhaps adopting the same conformation as that seen in the isolated domain structure. This could perhaps generate a polypeptide binding site via the clustered hydrophobic residues. Ditzel et al. (1998) have also performed experiments in which they co-crystallized or soaked in various nucleotides. The nucleotide-binding site was found to be located on top of the equatorial domain in the same position as that found in GroEL (compare Figures 8a and 9a and also Figures 8b and 8c). The Mg-ADP-bound crystal structure showed domain motions about the hinge regions between the three domains compared to the unliganded structure. The presence of the nucleotide phosphates displaces Asp-94 from its position in the triphosphate-binding site which it occupies in the unliganded structure. Residues 89–98 are shifted towards the neighbouring intermediate domain forcing the residues 85–88 to undergo a transition from relaxed loop to a helix. However, movements in the 2/ 3 stem-loop noted in the GroEL-ATP S structure (Boisvert et al., 1996) could not be seen in the thermosome-ADP structure (Ditzel et al., 1998). In contrast to GroEL-GroES-ADP where the nucleotide is not exchangeable with free nucleotide (Todd et al., 1994; Xu et al., 1997), the nucleotide-binding site in the thermosome crystal structure is still accessible to solvent, even when the intermediate domain is partly clamped down onto the equatorial domain. Soaking experiments using the ATP transition state analogue Mg-ADP-AIF3 were also performed providing some structural evidence
The relationship between chaperonin structure and function
595
on the mechanism of ATP hydrolysis (Ditzel et al., 1998). One fluorine was coordinated to the Mg2+ ion while the remaining two fluorine atoms form hydrogen bonds to threonines 96 and 97. A water molecule forms hydrogen bonds with Asp-390 (equivalent to Asp-398 in GroEL) and Asp-63 (Asp-52). The carboxylate side chains of the aspartate polarize the water molecule such that a nucleophilic attack is made on the phosphorus atom. The resulting pentavalent intermediate can subsequently collapse yielding a free orthophosphate and ADP. The role of Asp-390 is strongly supported by evidence that mutation of Asp-398 to Ala in GroEL resulted in a 60-fold reduction in the rate of ATP hydrolysis (Rye et al., 1997). 7. SUMMARY OF THE FUNCTIONAL CONSEQUENCES OF CHAPERONIN STRUCTURE The challenge remains today to explain how chaperonins are able to harness the energy of ATP hydrolysis in order to improve the efficiency of protein folding in vivo. However, the structural work described above, together with the extensive biochemical and biophysical studies described in the chapter by Ranson and Clarke, provide insights into the way in which chaperonins achieve their cellular task. The earliest stages of protein folding involve the collapse of a polypeptide chain around a hydrophobic core and are usually coincident with the formation of some elements of secondary structure (for review see Dobson et al., 1995; Dill et al., 1995). For many small proteins this occurs rapidly and efficiently, before the rate-limiting step late in the folding pathway. However, it has been observed for a number of proteins, especially larger oligomeric proteins, that a significant degree of misfolding occurs during these early stages (Sosnick et al., 1994; Jaenicke, 1995; Ranson et al., 1995). This ensemble of misfolded conformers contains a high degree of exposed hydrophobic surface and they are thus prone to irreversible aggregation. While some of these misfolded polypeptides can slowly isomerize, by reversing incorrect interactions, to a conformation that is able to rapidly fold to the native state, this process has to compete directly against collision with other misfolded states leading to aggregation. Thus the rate-limiting step of refolding for many large proteins may be the partial unfolding of kinetically trapped conformations, or even dissociation of small, low-order aggregates, to a fast-folding species. The chaperonin, therefore, has the task of preventing these occurrences and/or actively reversing them in order to improve the efficiency of refolding. GroEL has a number of unique physicochemical features in order to perform this function. The initial step is the “recognition” of polypeptide substrates. Identification of the binding site by electron microscopy and mutagenesis (Fenton et al., 1994), coupled with the high resolution X-ray structure has located the site of interaction to the apical domains (Chen et al., 1994; Braig et al., 1994; Buckle et al., 1997). Site-directed mutagenesis of hydrophobic residues in this region strongly affects both polypeptide and GroES binding (Fenton et al., 1994), suggesting that the binding is due primarily to hydrophobic interaction between the apical domains and substrate polypeptide. This has determined directly using microcalorimetry (Lin et al., 1995) and binding studies (Hutchinson et al., 1997). Since early folding intermediates and misfolded protein chains
Molecular chaperones and folding catalysts
596
expose hydrophobic surface which would otherwise be buried, this provides a means by which chaperonins can “recognize” a wide range of substrate polypeptides, without requiring any specific structural feature which would vary from substrate to substrate. It has also been proposed that the binding energy may be used to reverse incorrect interactions which have formed early in the folding/misfolding pathway (Jackson et al., 1993; Todd et al., 1994; Weissman et al., 1994; Ranson et al., 1995; Todd et al.., 1996; Zahn et al., 1996). The binding of ATP to the GroEL-polypeptide complex causes rotations of the two hinge regions between GroEL domains, raising and twisting the apical domains, and thus with-drawing part of the hydrophobic peptide binding site (Roseman et al., 1996; Xu et al., 1997). These movements permit the rapid association of GroES (Burston et al., 1995) to the remaining hydrophobic surface. Binding of ATP and GroES results in the rapid displacement of the polypeptide into the central cavity (Rye et al., 1997) as the hydrophobic binding site becomes buried (Roseman et al., 1996; Xu et al., 1997). The central cavity is now lined by a relatively polar surface (Xu et al., 1997) which provides an environment in which the protein may fold to a native or near-native conformation (Weissman et al., 1995 & 1996; Mayhew et al., 1996; Rye et al., 1997). This “folding active” complex has a defined lifetime regulated by the rate of ATP hydrolysis within the cis ring which weakens the interaction between the cis ring and GroES (Rye et al., 1997) due to the loss of interactions between the ?-phospate of the nucleotide (Xu et al., 1997). Subsequent ATP binding to the trans ring (Rye et al., 1997) is communicated to the cis ring via movements of the trans equatorial domains (Xu et al., 1997) and inter-ring contacts (Roseman et al., 1996) evicting GroES and both native and non-native polypeptides from the enclosed cavity (Todd et al., 1994; Weissman et al., 1994; Smith & Fisher, 1995; Burston et al., 1996; Ranson et al., 1997). This discharge of polypeptide from the cis cavity allows kinetic partitioning between the native state, other molecular chaperones, the same chaperonin or proteolytic machinery in vivo (Kandror et al., 1994; Buchberger et al., 1996). Iterative cycles of binding, encapsulation and release can be used to actively reverse misfolding/low-order aggregation and thus increase the yield of the native form (Jackson et al., 1993; Todd et al., 1994; Weissman et al., 1994; Ranson et al., 1995; Corrales & Fersht, 1996; Ranson et al., 1997). In a similar manner, although via a different structural mechanism, substrates for CCT in the eukaryotic cytosol also undergo rounds of binding and release (Tian et al., 1995; Farr et al., 1997). Further characterization of the group II chaperonins should yield more details of the similarities and more interestingly the differences between two groups of chaperonins. 8. ACKNOWLEDGMENTS S.G.Burston is a Wellcome Trust International Travelling Fellow. H.R.Saibil thanks the Wellcome Trust and the Biotechnology and Biological Sciences Research Council (UK) for support. The authors thank Art Horwich and Paul Sigler (Yale) for helpful discussions, and Lars Ditzel (Martinsried) for figures of the thermosome and GroELGroES-ADP structures.
The relationship between chaperonin structure and function
597
9. REFERENCES Aharoni, A. and Horovitz, A. (1996). Inter-ring communication is disrupted in the GroEL mutant Arg-13->Gly, Ala-126->Val with known crystal structure. J. Mol. Biol. , 258, 732–735. Azem, A., Diamant, S., Kessel, M., Weiss, C. and Goloubinoff, P. (1995). The proteinfolding activity of chaperonins correlates with the symmetric GroEL14 (GroES7)2 heterooligomer. Proc. Natl. Acad. Sci. USA , 92, 12021–12025. Badcoe, I.G., Smith, C.J., Wood, S., Halsall, D.J., Holbrook, J.J., Lund, P. and Clarke, A.R. (1991). Binding of a chaperonin to the folding intermediates of lactate dehydrogenase. Biochemistry , 30, 9195–9200. Bochkareva, E.S. and Girshovich, A.S. (1994). ATP induces non-identity of two rings in chaperonin GroEL. J. Biol Chem. , 269, 23869–23871. Bochkareva, E.S., Lissin, N.M. and Girshovich, A.S. (1988). Transient association of newly synthesized unfolded proteins with the heat-shock GroEL protein. Nature (London) , 336, 254–257. Bochkareva, E.S., Lissin, N.M., Flynn, G.C., Rothman, J.E. and Girshovich, A.S. (1992). Positive cooperativity in the functioning of molecular chaperone GroEL. J. Biol. Chem. , 267, 6796–6800. Boisvert, D.C., Wang, J., Otwinowski, Z., Horwich, A.L. and Sigler, P.B. (1996). The 2.4 Å crystal structure of the bacterial chaperonin GroEL complexed with ATP S. Nature Struct. Biol. , 3, 170–177. Braig, K,, Simon, M., Furuya, F., Hainfeld, J.F. and Horwich, A.L. (1993). A polypeptide bound by the chaperonin GroEL is localized within a central cavity. Proc. Natl. Acad. Sci. USA. , 90, 3978–3982. Braig, K., Otwinowski, Z., Hegde, R., Boisvert, D.C., Joachimiak, A., Horwich, A.L. and Sigler, P.B. (1994). The crystal structure of the bacterial chaperonin GroEL at 2.8Å. Nature (London) , 371, 578–586. Braig, K., Adams, P.D. and Brünger, A.T. (1995). Conformational variability in the refined structure of the chaperonin GroEL at 2.8Å resolution. Nature. Struct. Biol. , 2, 1083–1094. Buchner, J., Schmidt, M., Fuchs, M., Jaenicke, R., Rudolph, R., Schmid, F.X. and Kiefhaber, T. (1991). GroE facilitates folding of citrate synthase by suppressing aggregation. Biochemistry , 30, 1586–1591. Buchberger, A., Schröder, H., Hesterkamp, T., Schönfeld, H.J. and Bukau, B. (1996). Substrate shuttling between the DnaK and GroEL systems indicates a chaperone network promoting protein folding. J. Mol. Biol. , 261, 328–333. Buckle, A.M., Zahn, R. and Fersht, A.R. (1997). A structural model for GroELpolypeptide recognition. Proc. Natl. Acad. Sci. USA , 94, 3571–3575. Burston, S.G., Ranson, N.A. and Clarke, A.R. (1995). The origins and consequences of asymmetry in the chaperonin reaction cycle. J. Mol. Biol. , 249, 138–152. Burston, S.G., Weissman, J.S., Farr, G.W., Fenton, W.A. and Horwich, A.L. (1996). Native and non-native forms of a protein substrate are released from a “cis-only” form of GroEL. Nature (London) , 383, 96–99.
Molecular chaperones and folding catalysts
598
Carson, M. (1991). Ribbons 2.0 J. Appl. Cryst. , 24, 958–961. Chandrasekhar, G.N., Tilly, K., Woolford, C., Hendrix, R. and Georgopoulos, C. (1986). Purification and properties of the groES morphogenetic protein of Escherichia coli . J. Biol. Chem. , 261, 12414–12419. Chen, S., Roseman, A.M., Hunter, A.S., Wood, S.P., Burston, S.G., Ranson, N.A., Clarke, A.R. (1994). Location of a folding protein and shape changes in GroEL-GroES complexes imaged by cryo-electron microscopy. Nature (London) , 371, 261–264. Clarke, A.R. (1996a). Molecular chaperones in protein folding and translocation. Curr. Opin. Struct. Biol. , 6, 43–50. Clarke, A.R. (1996b). Kinetic and energetic aspects of chaperonin function. In “The Chaperonins” ed. R.J.Ellis, Academic Press Inc., San Diego, U.S.A. Corrales, F.J. and Fersht, A.R. (1996). Kinetic significance of GroEL14: (GroES7)2 complexes in molecular chaperone activity. Folding and Design , 1, 265–273. Dill, K.A., Bromberg, S., Yue, K., Fiebig, K.M., Yee, O.P., Thomas, P.D. and Chan, H.S. (1995). Principles of protein folding—A perspective from simple exact models. Protein Sci. , 4, 561–602. Ditzel, L., Löwe, J., Stock, D., Stetter, K.-O., Huber, H., Huber, R. and Steinbacher, S. (1998). Crystal structure of the thermosome, the archael chaperonin and homolog of CCT. Cell , 93, 125–138. Dobson, C.M., Evans, P.A. and Radford, S.E. (1995). Understanding how proteins fold: the lysozyme story so far. Trend Biochem. Sci. , 19, 31–37. Dubochet, J., Adrian, M., Chang, J.-J., Homo, J.-C., Lepault, J., McDowell, A.W. and Schultz, P. (1988). Cryo-electron microscopy of vitrified specimens. Q. Rev. Biophys. , 21, 129–228. Ellis, R.J. and Van der Vies, S.M. (1991). Molecular chaperones. Annu. Rev. Biochem. , 60, 321–347. Farr, G.W., Scharl, E.C., Schumacher, R.J., Sondek, S. and Horwich, A.L. (1997). Chaperonin-mediated folding in the eukaryotic cytosol proceeds through rounds of release of native and non-native forms. Cell , 89, 927–937. Fenton, W.A. and Horwich, A.L. (1997). GroEL-mediated protein folding. Protein Sci. , 6, 743–760. Fenton, W.A., Kashi, Y., Furtak, K. and Horwich, A.L. (1994). Residues in chaperonin GroEL required for polypeptide binding and release. Nature (London) , 371, 614–619. Fisher, M.T. (1992). Promotion of the in vitro folding of dodecameric glutamine synthetase from Escherichia coli in the presence of GroEL (chaperonin-60). and ATP. Biochemistry , 31, 3955–3963. Frydman, J., Nimmesgern, E., Erdjument-Bromage, H., Wall, J.S., Tempst, P. and Hartl, F.U. (1992). Function in protein folding of TRiC, cytosolic ring complex containing TCP-1 and structurally related subunits. EMBO J., 11, 4767–4778. Gao, Y., Thomas, J.O., Chow, R.L., Lee, G.H. and Cowan, N.J. (1992). A cytoplasmic chaperonin that catalyze b-actin folding. Cell , 69, 1043–1050. Gao, Y., Vainberg, I.E., Chow, R.I. and Cowan, N.J. (1993). Two co-factors and cytoplasmic chaperonin are required for the folding of - and -tubulin . Mol. Cell Biol. , 13, 2478–2485.
The relationship between chaperonin structure and function
599
Georgopoulos, C.P., Hendrix, R.W., Casjens, S.R. and Kaiser, A.D. (1973). Host participation in bacteriophage head assembly. J. Mol. Biol. , 76, 45–60. Goldberg, M.S., Zhang, J., Sondek, S., Matthews, C.R., Fox, R.O. and Horwich, A.L. (1997). Native-like structure of a protein-folding intermediate bound to the chaperonin GroEL. Proc. Natl. Acad. Sci. USA , 94, 1080–1085. Goloubinoff, P., Christeller, J.P., Gatenby, A.A. and Lorimer, G.H. (1989). Reconstitution of active dimeric ribulose bisphosphate carboxylase from an unfolded state depends on two chaperonin proteins and ATP. Nature (London) , 342, 884–889. Gorovits, B.M. and Horowitz, P.M. (1995). The molecular chaperonin cpn60 displays local flexibility that is reduced after binding with an unfolded protein. J. Biol. Chem. , 270, 13057–13062. Gray, T.E. and Fersht, A.R. (1991). Cooperativity in ATP hydrolysis by GroEL is increased by GroES. FEBS Lett. , 292, 254–258. GroB, M., Robinson, C.V., Mayhew, M., Hartl, F.U., Radford, S.E. (1996). Protein Sci. , 5, 2506–2513. Hartl, F.U. (1996). Molecular chaperones in cellular protein folding. Nature (London) , 381, 571–579. Hendrix, R. (1979). Purification and properties of GroE, a host protein involved in bacteriophage assembly. J. Mol. Biol. , 129, 375–392. Hohn, T., Hohn, B., Engel, A., Wurtz, M. and Smith, P.R. (1979). Isolation and characterization of the host protein GroE involved in bacteriophage lambda assembly. J. Mol. Biol. , 129, 359–373. Horovitz, A., Bochkareva, E.S. and Girshovich, A.S. (1993). The N-terminus of the molecular chaperonin GroEL is a crucial structural element for its assembly. J. Biol. Chem. , 268, 9957–9959. Horwich, A.L., Brooks Low, K., Fenton, W.A., Hirshfield, I.N. and Furtak, K. (1993). Folding in vivo of bacterial proteins: Role of GroEL. Cell , 74, 909–917. Hunt, J.F., Weaver, A.J., Landry, S.J., Gierasch, L. and Deisenhofer, J. (1996). The crystal-structure of the GroES co-chaperonin at 2.8 Å resolution. Nature (London) , 379, 37–45. Hutchinson, E.G., Tichelaar, W, Hofhaus, G., Weiss, H. and Leonard, K. (1989). Identification and electron microscope analysis of a chaperonin oligomer from Neurospora crassa mitochondria. EMBO J. , 8, 1485–1490. Hutchinson, J.P., Oldham, T.C., El-Thaher, T.S. H. and Miller, A.D. (1997). Electrostatic as well as hydrophobic interactions are important for the association of Cpn60(groEL). with peptides. J. Chem. Soc., Perkin Trans. , 2, 279–288. Ishii, N., Taguchi, H., Sumi, M. and Yoshida, M. (1992). Structure of holo-chaperonin studied with electron-microscopy—oligomeric cpn10 on top of two layers of cpn60 with two stripes each. FEBS Lett. , 299, 169–174. Ishii, N., Taguchi, H., Sasabe, H. and Yoshida, M. (1994). Folding intermediate binds to the bottom of bullet-shaped holo-chaperonin and is readily accessible to antibody. J. Mol. Biol. , 236, 691–696. Jackson, G.S., Staniforth, R.A., Halsall, D.J., Atkinson, T., Holbrook, J.J., Clarke, A.R. and Burston, S.G. (1993). The binding and hydrolysis of nucleotides in the chaperonin catalytic cycle: Implications for the mechanism of assisted protein folding.
Molecular chaperones and folding catalysts
600
Biochemistry , 32, 2554–2563. Jaenicke, R. (1993). Role of accessory proteins in protein folding. Curr. Opin. Struct. Biol. , 3, 104–112. Jaenicke, R. (1995). Folding and association versus misfolding and aggregation of proteins. Phil. Trans. Roy. Soc. Lon. B. , 348, 97–105. Kagawa, H.K., Osipiuk, J., Maltsev, N., Overbeek, R., Quaite-Randall, E., Joachimiak, A. and Trent, J. D. (1995). The 60 kDa heat shock proteins in the hyperthermophilic archaeon Sulfolobus shibatae . J. Mol. Biol. , 253, 712–725. Kandror, O., Busconi, I., Sherman, M. and Goldberg, A.L. (1994). Rapid degradation of an abnormal protein in Escherichia coli involves the chaperones GroEL and GroES. J. Biol. Chem. , 269 , 23575–23582. Kim, P.S. and Baldwin, R.L. (1990). Intermediates in the folding reactions of small proteins. Annu. Rev. Biochem. , 59 , 631–660. Kim, S., Willison, K. and Horwich, A.L. (1994). Cytosolic chaperonin subunits have a conserved ATPase domain but diverged polypeptide-binding domains. Trend Biochem. Sci. , 19 , 543–548. Klumpp, M., Baumeister, W. and Essen, L.-O. (1997). Structure of the substrate binding domain of the thermosome, an archael group II chaperonin. Cell , 91 , 263–270. Knapp, S., Schmidt-Krey, I., Herbert, H., Bergman, T., Jornvall, H. and Ladenstein, R. (1994). The molecular chaperonin TF55 from Archaeon Sulfolobus solfataricus . J. Mol. Biol. , 242 , 397–407. Koshland, D.E., Jr, Nemethy, G. and Filmer, D. (1966). Comparison of experimental binding data and theoretical models in proteins containing subunits. Biochemistry , 5, 365–385. Kovalenko, O., Yifrach, O. and Horovitz, A. (1994). Residue Lysine-34 in GroES modulates allosteric transitions in GroEL. Biochemistry , 33 , 14974–14978. Landry, S.J., Zeilstra-Ryalls, J., Fayet, O., Georgopoulos, C. and Gierasch, L.M. (1993). Characterization of a functionally important mobile domain of GroES. Nature (London) , 364, 255–258. Langer, T., Pfeifer, G., Martin, J., Baumeister, W. and Hartl, F.U. (1992). Chaperoninmediated protein folding: GroES binds to one end of the GroEL cylinder, which accommodates the protein substrate within its central cavity. EMBO J. , 11, 4757– 4765. Lewis, V.A., Hynes, G.M., Zheng, D., Saibil, H. and Willison, K. (1992). T-complex polypeptide-1 is a subunit of a heteromeric particle in the eukaryotic cytosol. Nature (London) , 358, 249–252. Lin, Z., Schwarz, F.P. and Eisenstein, E. (1995). The hydrophobic nature of GroELsubstrate binding. J. Biol. Chem. , 270, 1011–1014. Llorca, O., Marco, S., Carrascosa, J.L. and Valpuesta, J.M. (1994). The formation of symmetrical GroEL-GroES complexes in the presence of ATP. FEBS Lett. , 345, 181– 186. Lorimer, G.H. (1997). Folding with a two-stroke motor. Nature (London) , 388, 720–723. Mande, S.C., Mehra, V., Bloom, B.R. and Hol, W.G.J. (1996). Structure of the heat shock protein chaperonin-10 of Mycobacterium leprae . Science , 271, 203–207. Marco, S., Urena, D., Carrascossa, J.L., Waldmann, T., Peters, J., Hegerl, R., Pfeifer, G.,
The relationship between chaperonin structure and function
601
Sackkongehl, H. and Baumeister, W. (1994a). The molecular chaperone TF55: Assessment of symmetry. FEBS Lett. , 341, 152–155. Marco, S., Carrascossa, J.L. and Valpuesta, J.M. (1994b). Reversible interaction of actin along the channel of the TCP-1 cytoplasmic chaperonin. Biophys. J. , 67, 364– 368. Martin, J., Langer, T., Boteva, R., Schramel, A., Horwich, A.L. and Hartl, F.U. (1991). Chaperonin-mediated folding at the surface of GroEL through a “molten globule”-like intermediate. Nature (London) , 352, 36–42. Martin, J., Mayhew, M., Langer, T. and Hartl, F.U. (1993). The reaction cycle of GroEL and GroES in chaperonin-assisted protein folding. Nature (London) , 366, 228–233. Mayhew, M., DaSilva, A.C.R., Martin, J., Erdjument-Bromage, H., Tempst, P. and Hartl, F.U. (1996). Protein folding in the central cavity of the GroEL-GroES chaperonin complex. Nature (London) , 379, 420–426. Miklos, D., Caplan, S., Martens, D., Hynes, G., Pitluk, Z., Brown, C., Barrell, B., Horwich, A.L. and Willison, K. (1994). Primary structure and function of a second essential member of heterooligomeric TCP1 chaperonin complex of yeast., TCP1 . Proc. Natl. Acad. Set. USA , 91, 2743–2747. Miranker, A., Robinson, C., Radford, S. and Dobson, C.M. (1993). Detection of transient protein folding populations by mass spectrometry. Science , 262 , 896–900. Monod, J., Wyman, J. and Changeux, J.-P. (1965). On the nature of allosteric transitions: a plausible model. J. Mol. Biol , 12, 88–118. Nicholls, A. (1992). GRASP: Graphical representation and Analysis Surface Properties (Columbia University, New York). Phipps, B.M., Hoffman, A., Stetter, K.O. and Baumeister, W. (1991). A novel ATPase complex selectively accumulated upon heat shock is a major cellular component of thermophilic archaebacteria. EMBO J. , 10, 1711–1722. Radford, S.E., Dobson, C.M. and Evans, P.A. (1992). The folding of hen lysozyme involves partially structured intermediates and multiple pathways. Nature (London) , 358, 302–307. Ranson, N.A., Dunster, N.J., Burston, S.G. and Clarke, A.R. (1995). Chaperonins can catalyze the reversal of early aggregation steps when a protein misfolds. J. Mol. Biol. , 250, 581–586. Ranson, N.A., Burston, S.G. and Clarke, A.R. (1997). Binding, encapsulation and ejection: substrate dynamics during a chaperonin-assisted folding reaction. J. Mol. Biol. , 266, 656–664. Roseman, A.M., Chen, S., White, H., Braig, K. and Saibil, H.R. (1996). The chaperonin ATPase cycle: mechanism of allosteric switching and movements of substrate-binding domains in GroEL. Cell , 87, 241–251. Rye, H.S., Burston, S.G., Fenton, W.A., Beechem, J.M., Xu, Z., Sigler, P.B. and Horwich, A.L. (1997). Direct action of cis and trans ATP within the double ring of the chaperonin GroEL. Nature (London) , 388, 792–798. Saibil, H.R., Zheng, D., Wood, S.P. and auf der Mauer, A. (1991). Binding of chaperonins. Nature (London) , 353, 25–26. Saibil, H.R., Zheng, D., Roseman, A.M., Hunter, A.S., Watson, G.M.F., Chen, S., auf der Mauer, A., O’Hara, B.P., Wood, S.P, Mann, N.H., Barnett, L.K. and Ellis, R.J. (1993).
Molecular chaperones and folding catalysts
602
ATP induces large quaternary rearrangements in a cage-like chaperonin structure. Curr. Biol. , 3, 265–273. Scale, J.W. and Horowitz, P.M. (1995). The C-terminal sequence of the chaperonin GroES is required for oligomerization. J. Biol. Chem. , 270, 30268–30270. Schmidt, M., Buchner, J., Todd, M.J., Lorimer, G.H. and Viitanen, P.V. (1994a). On the role of GroES in the chaperonin-assisted folding reaction. J. Biol. Chem. , 267, 10304– 10311. Schmidt, M., Rutkat, K., Rachel, R., Pfeiffer, G., Jaenicke, R., Viitanen, P.V., Lorimer, G.H., and Buchner, J. (1994b). Symmetric complexes of GroE chaperonins as part of the functional cycle. Science , 265, 656–659. Sosnick, T.R., Mayne, L., Hiller, R. and Englander, S.W. (1994). The barriers in proteins folding. Nature Struct. Biol. , 1, 149–156. Sosnick, T.R., Mayne, L. and Englander, S.W. (1996). Molecular collapse: The ratelimiting step in two-state cytochrome c folding. Proteins: Struct., Funct. and Genet. , 24, 413–426. Sparrer, H., Lilie, H. and Buchner, J. (1996). Dynamics of the GroEL-protein complex: effects of nucleotides and folding mutants. J. Mol. Biol. , 258, 74–87. Staniforth, R.A., Burston, S.G., Atkinson, T. and Clarke, A.R. (1994a). Affinity of chaperonin-60 for a protein substrate and its modulation by nucleotides and chaperonin-10. Biochem. J. , 300, 651–658. Staniforth, R.A., Cortés, A., Burston, S.G., Atkinson, T, Holbrook, J.J. and Clarke, A.R. (1994b). The stability and hydrophobicity of cytosolic and mitochondrial malate dehydrogenases and their relation to chaperonin-assisted protein folding. FEBS Lett. , 344, 129–135. Sternlicht, H., Farr, G.W., Sternlicht, M.L., Driscoll, J.K., Willison, K. and Yaffe, M.B. (1993). The t-complex polypeptide-1 complex is a chaperonin for tubulin and actin in vivo . Proc. Natl. Acad. Sci. USA , 90, 9422–9426. Tian, G., Vainberg, I.E., Tap, W.D., Lewis, S.A. and Cowan, N.J. (1995). Specificity in chaperonin-mediated protein folding. Nature (London) , 375, 250–253. Tian, G.L., Huang, Y., Rommelaere, H., Van de Kerkhove, J., Ampe, C and Cowan, N.J. (1996). Pathway leading to correctly folded -tubulin. Cell , 86, 287–296. Trent, J.D., Nimmesgern, E., Wall, J.S., Hartl, F.U. and Horwich, A.L. (1991). A molecular chaperone from a thermophilic archaebacterium is related to the eukaryotic protein t-complex polypeptide-1. Nature (London) , 354, 490–493. Todd, M.J., Viitanen, P.V. and Lorimer, G.H. (1993). Hydrolysis of adenosine 5'triphosphate by Escherichia coli GroEL: effects of GroES and potassium ion. Biochemistry , 32, 8560–8567. Todd, M.J., Viitanen, P.V. and Lorimer, G.H. (1994). Dynamics of the chaperonin ATPase cycle—implications for facilitated protein folding. Science , 265, 659–666. Todd, M.J., Lorimer, G.H. and Thirumalai, D. (1996). Chaperonin-facilitated protein folding—opti-mization of rate and yield by an iterative annealing mechanism. Proc. Natl. Acad. Sci. USA , 93, 4030–4035. Ursic, D., Sedbrook, J.C., Himmel, K.L. and Culbertson, M.R. (1994). The essential yeast Tcp1 protein affects actin and microtubules. Mol. Cell Biol. , 11, 2629–2640. Viitanen, P.V, Lubben, T.H., Reed, J., Goloubinoff, P., O’Keefe, D.P. and Lorimer, G.H.
The relationship between chaperonin structure and function
603
(1990). Chaperonin-facilitated refolding of ribulose-bisphosphate carboxylase and ATP hydrolysis by chaperonin-60 (GroEL). are K+ dependent. Biochemistry , 29, 5665– 5671. Weissman, J.S., Kashi, Y., Fenton, W.A. and Horwich, A.L. (1994). GroEL-mediated protein folding proceeds by multiple rounds of binding and release of non-native forms. Cell , 78, 693–702. Weissman, J.S., Hohl, C.M., Kovalenko, O., Kashi, Y., Chen, S., Braig, K., Saibil, H.R., Fenton, W.A. and Horwich, A.L. (1995). Mechanism of GroEL action—productive release from a sequestered position under GroES. Cell , 83, 577–587. Weissman, J.S., Rye, H.S., Fenton, W.A., Beechem, J.M. and Horwich, A.L. (1996). Characterization of the active intermediate of a GroEL-GroES-mediated protein folding reaction. Cell , 84, 481–490. White, H.E., Chen, S., Roseman, A.M., Yifrach, O., Horovitz, A. and Saibil, H.R. (1997). Structural basis of allosteric changes in the GroEL mutant Arg197->Ala. Nature Struct. Biol. , 4, 690–694. Willison, K.R. and Kubota, H. (1994). The structure, function and genetics of the chaperonin containing TCP-1 (CCT). in eukaryotic cytosol. In “The Biology of Heat Shock Proteins and Molecular Chaperones” (R.I.Morimoto, A.Tissieres and C.Georgopoulos, eds.). pp. 299–312. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY Willison, K.R. and Horwich, A.L. (1996). Chaperonins in archaebacteria and eukaryotic cytosol. In “The Chaperonins” (R.J.Ellis, ed.). pp 107–136. Academic Press, San Diego, CA. Xu, Z., Horwich, A.L. and Sigler, P.B. (1997). The crystal structure of the asymmetric GroEL•GroES• (ADP)7 chaperonin complex. Nature (London) , 388, 741–750. Yaffe, M.B., Farr, G.W., Miklos, D., Horwich, A.L., Sternlicht, M.L. and Sternlicht, H. 91992). TCP1 complex is a molecular chaperone in tubulin biogenesis. Nature (London) , 358, 245–248. Yifrach, O. and Horovitz, A. (1994). Two lines of allosteric communication in the oligomeric chaperonin GroEL are revealed by the single mutation Arg196->Ala. J. Mol. Biol. , 243, 397–401. Yifrach, O. and Horovitz, A. (1995). Nested cooperativity in the ATPase activity of the oligomeric chaperonin GroEL. Biochemistry , 34, 5303–5308. Yifrach, O. and Horovitz, A. (1996). Allosteric control by ATP of a non-folded proteinbinding to GroEL. J. Mol. Biol. , 255, 356–361. Zahn, R., Perrett, S., Stenberg, G. and Fersht, A.R. (1996). Catalysis of amide protonexchange by the molecular chaperones GroEL and SecB. Science , 271, 642–645. Zheng, X.X., Rosenberg, L.E., Kalousek, F. and Fenton, W.A. (1993). GroEL, GroES and ATP-dependent folding and spontaneous assembly of ornithine transcarbamylase. J. Biol. Chem. , 268, 7489–7493. Zondlo, J., Fisher, K.E., Lin, Z., Ducote, K.R. and Eisenstein, E. (1995). Monomerheptamer equilibrium of Escherichia coli chaperonin GroES. Biochemistry , 34, 10334–10339.
24. COMPOSITION AND FUNCTION OF THE EUKARYOTIC CYTOSOLIC CHAPERONIN-CONTAINING TCP-1 KEITH R.WILLISON Institute of Cancer Research, Cancer Research Campaign Centre for Cell and Molecular Biology, Chester Beatty Laboratories, 237 Fulham Road, London SW3 6JB, UK
1. Introduction 2. The Component Parts of CCT 2.1. CCT Gene Family 2.2. CCT is a Group II Chaperonin 2.3. CCT Gene Linkage in Lower Eukaryotes 2.4. Co-chaperones of CCT 2.5. Polypeptide Binding Sites on CCT 3. CCT Mediated Folding 3.1. Spectrum of Substrates 3.2. Tubulin Folding 4. CCT Function 4.1. CCT Functional Cycle 4.2. Co-translational Folding 4.3. Cell Biology 5. Conclusions 6. Acknowledgments 7. References 1. INTRODUCTION Since the discovery and early characterization of the eukaryotic cytosolic chaperonin containing TCP-1 (CCT) in 1992 (Lewis et al., 1992; Gao et al., 1992; Yaffe et al., 1992; Frydman et al., 1992), we have reviewed developments in this field annually (Horwich and Willison, 1993; Willison and Kubota, 1994; Kubota, Hynes and Willison, 1995a; Willison and Horwich, 1996). In contrast to the huge efforts being expended on the structure and reaction mechanism of the GroEL chaperonin of E. coli, there is still surprisingly little work being conducted on CCT. This may be partially due to a
Molecular chaperones and folding catalysts
606
widespread perception that CCT is merely the eukaryotic equivalent of GroEL and that they must be very similar protein machines, sharing reaction mechanisms and functions in assisting general protein folding. However, evidence is accumulating that CCT is different to GroEL in many respects, particularly in its substrate specificity and interaction with co-chaperones. A stumbling block to the study of CCT is the absence of a high resolution structure, without which we will be unable to understand the significance of its multi-subunit nature and why 8 separate genes evolved to encode the core complex. The purpose of this chapter is not to re-review the early literature on CCT, but to give an up-to-date opinion on the composition and genetic structure of CCT, its natural substrates and mechanism of action. If, as we suspect from present data, CCT is a sequence specific chaperonin for the actins and tubulins, it may be possible in the future to determine the structures of their folding intermediates trapped on CCT. If this experimental approach proves successful, it should provide insight into the general nature of chaperone action and inform us how some partially folded proteins are truly perceived in the eukaryotic cytosol. 2. THE COMPONENT PARTS OF CCT 2.1. CCT Gene Family The studies of Kubota et al. (1994, 1995a, 1995b, 1997a, and 1997b) established the correspondences between the protein subunits of CCT and the genes encoding them. The obtention of the complete sequences of 9 members of the mouse CCT gene family provided the ground work for studies in all other eukaryotic species. It is now clear, both from the molecular genetics of the CCT gene family and from our own studies on the protein composition of mouse and human CCT, using a combination of protein sequencing, mass spectrometry and antibody characterization (Hynes et al., 1995, 1996a, 1996b), and from the protein sequencing studies of Rommelaere et al., (1993) on rabbit reticulocyte cytosolic chaperonin, that there are 8 core components of CCT, and that they are encoded by 8 separate genes. The “extra” CCT subunits in preparations isolated from mouse or bovine testis are now known to be the consequence of the presence of tissue specific CCT genes; mouse contains aCCT related gene, CCT 2, apparently only expressed in testis (Kubota et al., 1997a). A particularly gratifying result, a consequence of the completion of the Saccharomyces cerevisiae genome sequencing project in 1996, was the confirmation of the presence of 8 CCT genes in this organism (Stoldt et al., 1996), each one being clearly the orthologue of one of the 8 core murine genes isolated in our laboratory (Kubota et al., 1997b). It is likely that all eukaryotes possess these 8 CCT genes, since orthologues of one or more CCT genes have been discovered in amphibians (Sun et al., 1995, Dunn and Mercola, 1996), C. elegans (Leroux and Candido, 1995), Tetrahymena (Soares et al., 1994), Drosophila (Ursic and Ganetzky, 1988) and plants (Mori et al., 1992), for example. 2.2. CCT is a Group II Chaperonin There exist two distinct families of ring-shaped chaperonins composed of similar sized
Composition and function of the eukaryotic cytosolic
607
subunits around 550 amino acid residues in length. Group I chaperonins, found in eubacteria, mitochondria and chloroplasts, have 7-fold rotational symme-try and E. coli GroEL is the best studied member of this group (see chapters by Burston & Saibil, and Ranson & Clarke, this volume). The structures of GroEL (Braig et al., 1994) and its ATP bound form (Boisvert et al., 1996) and its co-chaperonin GroES (Hunt et al., 1996) and the asymmetric GroEL-GroES-(ADP)7 complex (Xu et al., 1997) have been determined. Group II chaperonins are found in Archaebacteria, TF55 (Trent et al., 1991, 1994) and the thermosome (Phipps et al., 1991, 1993; Waldmann et al., 1995a,b) and the eukaryotic cytosol, CCT (Kubota et al., 1994) also known as TRiC (Frydman et al., 1992) or cytosolic chaperonin (Gao et al., 1992). The lines of evolution of the two groups of chaperonins are consistent with the view that most components of the eukaryotic cytosol are more closely related to archaebacterial relatives and that the organelles of eukaryotic cytosol, mitochondria and chloroplasts, are derived from eubacterial ancestors by symbiosis. The Archaebacterial chaperonins comprise two symmetry groups, with TF55 from Sulfolobus species having 9 subunits per ring (Knapp et al., 1994, Marco et al., 1994b, Kagawa et al., 1995) and the Thermosome from Pyrodictium, Thermoplasma and Methanococcus species all having 8 subunits per ring (Phipps et al., 1991, 1993; Waldmann et al., 1995b; Andra et al., 1996). CCT has 8 subunits per ring (Marco et al., 1994a; Waldmann et al., 1995a; Saibil 1996; Llorca et al., 1998) and we have proposed a model whereby each position in the ring is occupied by one of the 8 constitutively expressed CCT subunit species (Liou and Willison, 1997). We have suggested that it is likely that CCT evolved from a thermosome having 8-fold symmetry (Willison and Horwich, 1996) and recently Nitsch et al. (1997) found the 8-membered rings of the thermosome of thermoplasma acidophilum to contain two subunit species in alternating positions in each ring. Ditzel et al. (1998) have determined the crystal structure of this )4 arrangement with 42-point thermosome and shown it to have an( )4 ( symmetry. Some such thermosome type was the likely immediate precursor of CCT. However, there is so far no evidence from sequence comparisons to indicate that any particular Archaebacterial subunit is more closely related to any of the CCT subunits than any other. Indeed, all the Group II amino acid sequences are about 30–40% identical to each other in pairwise comparisons (Kubota et al., 1994). A most instructive way to examine the evolutionary relatedness of all the chaperonins is to compare the blocks of sequences which constitute their equatorial, nucleotide binding, domains based on the structure of GroEL (Braig et al., 1994; Boisvert et al., 1996) and the sequence analysis of Kim et al., (1994). Figure 1 shows an example of this type of phylogenetic analysis as first performed by Kubota et al., (1994, 1995a). The separateness of the two groups is apparent, as is the more homogeneous nature of Group I. The remarkable result though is the fact that the eukaryotic CCT subunits diverged from each other approximately 2 billion years ago, around the time of the origin of the eukaryotic cell. Depending upon the number of chaperonin genes present in the precursor of the eukaryotic cell, probably one or two as found in present day Archaebacteria, there may have been two or three rapid gene duplication events to provide the sequences from which the 8 modern CCT genes could evolve. A further fascinating question concerning the evolution of CCT is how it managed to evolve into a machine which could bind actins and tubulins which are
Molecular chaperones and folding catalysts
608
unrelated in sequence, structure and evolutionary origin. The complete sequencing of Methanococcus Table 1 Correspondence between subunit proteins and genes of CCT Protein Source
CCT
CCT
GCT
CCT
CCT
CCT 1
CCT 2 CC
Mouse testis
S3
S4
S5
S9
S2
S6/S7
S10
S8
Bovine testis
P4
P1
P5
P3
—
—
—
P2
6b
4
6a
3
1
—
2
Ccta
Cctb
Cctg
Cctd
Ccte
Cctz1
Cctz2
Cct
Yeast/Worm
(CCT1)
(CCT2)
CCT3)
(CCT4)
(CCT5)
(CCT6)
—
(CC
Yeast/ORFs
YDR212w YIL142w YJL014w YDL143w YJR064w YDR188w —
YJL
Worm/ORFs
CE 07941 CE 25632
—
Rabbit reticulo 5 -cycte Gene Mouse
—
CE 25697 CE 25698
pT 2
pT 2
pT 5
u38846
D43950* L27706
Mouse cDNAs Tcp-1
pT 2
Human cDNAs x52882
w72593* x74801
—
—
pT 12
pT 2.2 pC D78333 N2
The gene and protein nomenclatures for CCT subunits have been described in Kubota et al. (1994) genes for CCT -CCT proteins, respectively); Kubota et al. (1995b) (Criq for CCT ); and Kubot (Cct l and Cct 2 for CCT l and CCT 2 respectively). Numbers are shown for each mammalian C protein prepared from each source; mouse testis (Kubota et al., 1994, 1997a). Bovine testis (Frydm and rabbit reticulocyte (Rommelaere et al., 1993). Clone names are shown for Cct genes of mouse a accession numbers for human sequences. The gene nomenclatures shown in parentheses (CCT1-CC recommended for budding yeast (S. cerevisiae) genes (Stoldt et al., 1996) and C. elegans genes (Le Candido, 1995). The worm ORFs correspond to the EMBL entries for the complete protein sequenc CE 07941 for CCT1. Some human cDNAs are asterisked indicating partial cDNA sequences and no of human cDNA CCT partial sequences are available at NCBI UniGene (http://www.ncbi.nlm.nih.g search using T-COMPLEX as query. This table is modified from Kubota et al., (1995a).
Composition and function of the eukaryotic cytosolic
609
Figure 1 Phylogenetic analysis of chaperonins based on the amino acid sequences of their putative ATPase domains. The chaperonin family proteins were aligned over the four conserved putative ATPase domains which have been suggested to be involved in ATP binding and hydrolysis (Kim et al., 1994) based upon the crystal structure of GroEL (Braig et al., 1994, Boisvert et al., 1996). The evolutionary tree is based on amino-acid substitutions in the conserved domains and the actual alignments used for this analysis can be found in Kubota et al., (1995a, b). The tree was constructed by the neighbourjoining method (Saitou and Nei, 1987). Rubisco subunit binding protein (RBP) and TF55 have two subunit species and CCT has eight subunit species. The other chaperonins are homo-oligomeric. MM, Mus musculus (Mouse); SC, Saccharomyces cerevisiae (yeast); SS, Sulflobus shibatae (archaebacterium); EC, Escherichia coli (eubacterium); ML, Mycobacterium leprae (eubacterium); BN, Brassica napus (plant). References of sequences are as follows: CCT (Willison et al., 1986; Kirchhoff and Willison, 1990); CCT , CCT , CCT , CCT , CCT and CCT , (Kubota et al., 1994); CCTq, (Kubota et al., 1995b); TF55 (Trent et al., 1991): E. coli GroEL (Hemmingsen et al., 1988); M leprae homologue of GroEL (Mehra et al., 1986); M.musculus Hsp60 (Venner and Gupta, 1990); S. cerevisiae (Hsp60 (Reading et al., 1989); B.napus RBPa and RBPb (Martel et al., 1990). The lengths of branches represent genetic distances between chaperonin proteins determined by the numerical output of the phylogenetic computer analysis.
jannaschii (Bult et al., 1996) shows no actin/hexokinase/Hsp70 ATP binding fold (Flaherty et al., 1991) in this organism although it does contain two tubulin related genes (ORFs 0370 and 0622) and Archaeoglobus fulgidus (Klenk et al., 1997) is similar; no DnaK genes and two Fts-Z tubulin related genes (ORFs 0535 and 0570). However eubacteria do contain the actin/hexokinase/Hsp70 fold and tubulin related genes, as do modern eukaryotic cells, so it is possible that the actin/hexokinase/Hsp70 fold evolved in
Molecular chaperones and folding catalysts
610
a separate lineage to thermophilic archaebacteria and that a fusion event (Lake, 1991; Doolittle, 1996) brought the precursors of the two cytoskeletal families together in the precursor of the eukaryotic cell. Perhaps when we understand how CCT interacts with folding intermediates of actins and tubulins we will be able to speculate further how the elements of the modern eukaryotic cytoskeleton suddenly evolved and then became virtually fixed in sequence over the subsequent 2 billion years (Willison and Horwich, 1996).
Figure 2 CCT gene linkage in the yeast and worm. The chromosomal locations of all 8 CCT genes in Saccharomyces cerevisiae and in Caenorhabditis elegans are shown representationally. The exact sequences, orientations and locations of the yeast CCT genes can be found at http://speedy.mips.biochem.mpg.de/mips/yeast/. The more precise locations of the worm CCT genes can be found in the C. elegans ACeDB database (Richard Durbin and Jean Thierry-Meig) as described by Leroux and Candido, (1995) and cosmids containing the CCT genes can be accessed by running TBLASTN with cct sequences at www.sanger.ac.uk/projects/c_elegans/blastserver.shtml.See Table 1 for the yeast and worm CCT ORFs. Cosmids for all 8 worm CCT genes have been isolated. Five yeast CCT genes are found in two linkage groups; CCT 3, 8, 5 within 150,000 bps on chromosome X and CCT 6 and 1, 50,000 bp apart on chromosome IV. Three worm CCT genes are closely linked on chromosome II; CCT 1, 4 and 2.
2.3. Gene Linkage in Lower Eukaryotes The CCT genes of mice and humans are distributed across the autosomes and there is no
Composition and function of the eukaryotic cytosolic
611
linkage of any pairs of CCT genes (Ashworth et al., 1994; Kubota et al., in preparation). This result is not unexpected for a gene family which expanded 2 billion years ago, since there have occurred around 300 chromosomal rearrangements between mouse and human in the 80 million years since they last shared a common ancestor. However, there are strong suggestions of extant ancestral linkage groups in yeast and the worm (Figure 2). The linkage data hints at the existence of two subgroups of CCT genes and, given further data from other species in the future, it may prove possible to discern the order of the duplications and, possibly, the ancestral linkage groups in the original eukaryotic cell and/or its precursor. 2.4. Co-chaperones of CCT The GroES co-chaperonin plays a central role in the mechanism of action of GroEL particularly in helping to form the expanded central cavity in the cis ring and also causing release of bound substrate from the hydrophobic peptide-binding residues into the cavity (Xu et al., 1997) (Chapters Ranson and Clarke; Burston and Saibil, this volume). However, no ring-shaped GroES-like co-chaperonins have been discovered for any of the Group II family and there is no GroES homologous sequence in the Methanococcus jannaschii genome which has a single gene encoding a Group II TF55-like chaperonin (Open reading frame: MJ 0999) (Bult et al., 1996) nor in the Archaeoglobus fulgidus genome (Klenk et al., 1997) which has two Group II thermosome-like chaperonins (Open reading frames AF2238 and AF1451). As we argued in Section 2.2, the GroEL-like Group I chaperonins are a homogeneous, later evolving group of proteins and all cells and organelles containing them also appear to possess a GroES type ring co-chaperonin. It is likely that GroES is a later addition in the evolution of Group I chaperonin function, since the GroES genes seem to be able to duplicate, amplify and diverge; T4 phage even encodes a Gp31 protein which is unrelated in sequence to GroES but nevertheless is able to substitute functionally for it (Hunt et al., 1997). Recently Klumpp et al. (1997) determined the crystal structure of the apical polypeptide binding domain of the thermoplasma acidophilum thermosome -subunit and discovered the presence of a large helical protrusion which they suggest may control access to the central cavity of this chaperonin. The complete structure of the 16-mer confirms that these helices close off the central cavity by forming a lid (Ditzel et al., 1998). Recent negative stain EM analysis of CCT suggests that ATP induces assymmetric movements of the apicol domains in opposite rings and these events could be related to access of the cavity by substrates (Llorca et al., 1998). 2.5. Polypeptide Binding Sites on CCT Electron microscopic studies of GroEL (Chen, 1994; Saibil, 1996; chapter Burston and Saibil, this volume) and group II chaperonins (Phipps et al., 1991, 1993; Marco et al., 1994b) and CCT (Marco et al., 1994a; Waldman et al., 1995a; Saibil, 1996; Llorca et al., 1998) suggest they have the same overall architecture despite the variation in the number of subunits per ring. However, the side views of all the group II chaperonin cylinders suggest that the putative apical polypeptide binding domains are rather smaller than that of GroEL (Braig et al., 1994). Also, CCT seems to have a larger central channel with a
Molecular chaperones and folding catalysts
612
diameter of 60Å compared with 45 A for GroEL. We have previously discussed comparisons between GroEL and CCT based on sequence similarities and molecular modelling (Kim et al., 1994; Willison and Horwich, 1996) and argued that the ATPase, equatorial domain is highly conserved in all chaperonins, whereas the putative apical polypeptide binding domains of all the CCT subunits are very divergent and Waldmann et al. (1995b) came to similar conclusions. It is suggested that each CCT subunit has specific binding sites for specific regions of its substrates. This suggestion is presently unproven, but testing each CCT apical domain for substrate binding specificity is a high priority in the field. The determination of the structure of the isolated thermosome apical domain by Klumpp et al. (1997) has been a major advance and shows that the core of the apical domain does in fact resemble GroEL but that the hydrophobic substrate binding regions found in GroEL are absent (chapter Burston and Saibil, this volume). However, as mentioned previously in section 2.4, a helical projection, displaying a novel helix-turnhelix motif, is found in the new structure which may contain the substrate binding sites. In the light of this structure, the amino acid sequences of the thermosome apical domain may be aligned very well with the equivalent regions in the 8 CCT subunits (see Figure 1, Klumpp et al., 1997). There is indication that some of the short stretches of about 10 residues which are conserved between mouse and yeast CCT subunit apical domains (Kim et al., 1994), but which vary between the 8 different CCT subunits within a species, could be potential substrate binding regions (Counsell and Willison, 1998; submitted for publication). Interestingly these regions contain a high proportion of charged residues which leads Klumpp et al. (1997) to discount them as substrate binding sites based on the GroEL model of hydrophobic substrate interaction sites. However, if CCT is a sequencespecific chaperonin it is possible that the subunits may have rather different substrate interaction sites to GroEL. Nevertheless Dobrzynski et al. (1996) identified residues 150– 350 in -tubulin which interact wich CCT (TriC) and which contains a hydrophobic-rich domain. The identification of the specific sequences in substrates which interact with CCT is an urgent problem to solve. 3. CCT-MEDIATED FOLDING 3.1. Spectrum of Substrates In contrast to the large number of proteins which interact with GroEL in vitro and in vivo (Hartl, 1996), a restricted range of substrates has been identified so far for CCT, notably actins, tubulins and their homologues. CCT can capture actins and tubulins upon dilution from denaturants (Gao et al., 1992; Frydman et al., 1992) or upon in vitro translation in reticulocyte lysate (Yaffe et al., 1992; Frydman et al., 1996; Liou et al., 1998). CCT can be recovered from cells with actin and tubulins remaining tightly bound (Sternlicht et al., 1993; Hynes et al., 1995, 1996a) and incubation with ATP at 37°C can cause release of these substrates (Willison et al., submitted). So far, all temperature sensitive mutations in yeast CCT cause defects in actin and/or tubulin behaviour or localization (Ursic and Culbertson, 1991; Miklos et al., 1994; Chen et al., 1994; Vinh and Drubin, 1994; Ursic et al., 1994). Thus, it seems fairly certain that actins and tubulins are bona fide substrates of CCT. There are other CCT binding proteins including luciferase (Frydman et al., 1992),
Composition and function of the eukaryotic cytosolic
613
neurofilament fragments (Roobol and Garden, 1993), hepatitis capsids (Lingappa et al., 1994), Gα transducin (Farr et al., 1997) and various unidentified polypeptides (Hynes et al., 1995, 1996a; Willison et al., submitted). The spectrum of proteins which can bind CCT could be as few as some tens of actin and tubulin family members or, as many as hundreds of different proteins, but one has the impression that there are not as many substrates as suggested for GroEL (Hartl, 1996). The identification of the true substrates will be revealed once we know the motifs on the apical polypeptide chain binding regions of CCT subunits which are responsible for binding unfolded substrate proteins. 3.2. Tubulin Folding Gao et al. (1992) discovered the requirement for co-factor proteins in CCT mediated folding of tubulins. It appears that it is not possible to produce assembly competent tubulin through the action of CCT alone (Lewis et al., 1996). Initial studies from Cowan’s laboratory suggested interaction of co-factor A directly with CCT to increase its ATPase activity during folding of -tubulin (Gao et al., 1993, 1994), but subsequent studies showed requirement for co-factor A in sequestration of -tubulin subunits downstream of their release from CCT (Melki et al., 1996). A yeast homologue of cofactor A, Rb12, rescues cells overexpressing toxic levels of -tubulin (Archer et al., 1995). Campo et al. (1994) independently discovered co-factor A and showed that it caused release of -tubulin monomers from C300 complexes. Cowan’s laboratory have now examined the -tubulin pathway in great detail, and a remarkable view of the tubulin folding pathway is emerging in which -tubulin folding requires CCT and four co-factors (FA, FD, FE, FC). -tubulin folding intermediates pass from co-factor to cofactor before attaining an assembly competent state, that is, the ability to form heterodimers with -tubulin (Tian et al., 1996). The 3-D structures of the various, proposed -tubulin intermediates are unknown and, hence it is not yet clear whether the role of the co-factors is to facilitate completion of folding or to stablize aggregationprone tubulin monomers. Tian et al. (1997) now argue that the folding pathways of and -tubulins merge through their interaction together in a complex with FC, FD and FE from which / -tubulin dimers emerge and that tubulin subunits are maintained in activated conformational states by the set of co-factors. However, in addition to co-factor A, several of the co-factors behave operationally as MAPs as pointed out by Tian et al. (1996) for co-factors C and E. Hirata et al. (1998) have just shown an essential role for the S.pombe homologue of FD, Alp1, in microtubule assembly and that Alp1 is associated with microtubules in vivo. It seems that the major roles for the co-factors may not be in folding per se but rather in tubulin monomer and heterodimer sequestration and flux into and out of microtubules. There is no evidence to suggest that any of the four tubulin co-factors identified by Cowan and coworkers interact directly with CCT. In conclusion, it seems clear that tubulin is not released from CCT in a native state, although a portion of -tubulin molecules can reach a quasi-native state without leaving the chaperonin (Tian et al., 1995b; Farr et al., 1997). If, as we think, tubulins are suspended on CCT via multiple contacts, it may be the case that release is ordered and progressive. Once the CCTmediated aspects of tubulin folding are understood it will be easier to comprehend the
Molecular chaperones and folding catalysts
614
roles of the co-factors in tubulin biogenesis. 4. CCT FUNCTION 4.1. CCT Functional Cycle The nature of the functional cycle of CCT is poorly understood. In vitro, CCT can bind actins, tubulins and luciferase, diluted from denaturant or upon in vitro translation of these substrate proteins in rabbit reticulocyte lysate. Presently there are conflicting models as to the mechanism by which CCT binds actin and tubulin substrates and subsequently uses ATP hydrolysis to release them. There is also conflict concerning the conformation of the released substrates; are they fully native or are they still folding intermediates? Hartl’s laboratory claim that human -actin interacts with TRiC whilst still bound to the ribosome and that folding of actin is facilitated to its native state in a single round of interaction (Frydman et al., 1994, 1996). Cowan’s laboratory suggest that tubulins are released from CCT in a nonnative state and in a series of hand-over reactions proceed to competence for / -tubulin heterodimer formation (Tian et al., 1995a, 1995b, 1996, 1997; Lewis et al., 1996). Horwich’s laboratory have used GroEL and CCT traps to show that, in single rounds of folding, CCT can release a proportion of nonnative tubulins which can be rebound by another chaperonin; simply put, CCT works like GroEL (Farr et al., 1997). Liou et al. (1998) have shown a single-ring mediated assembly and disassembly cycle of CCT in reticulocyte lysate which occurs constitutively at 30°C. This activity could be linked to CCT replication or foliling but its significance for the reaction mechanism of CCT is presently unknown. Much further work is required to resolve the issues. One clear fact is that substrate release is ATP dependent. CCT (Chromobindin A from bovine adrenal medulla and TRiC from bovine testis) has an ATP hydrolysis rate of 14–18 molecules of ATP hydrolyzed/ molecule CCT/min at 37°C (Martin and Creutz, 1987; Frydman et al., 1992). This ATPase activity is relatively slow and, indeed, various aspects of CCT behaviour seem slow. Tian et al. (1995a) showed that actin takes some minutes to attain a conformation which is capturable by CCT, and actins and tubulins interact with CCT for several minutes at 30°C in rabbit reticulocyte lysate (Liou et al., 1998). Sternlicht et al. (1993) measured interactions of actins and tubulins in vivo in pulse-labelled CHO cells and found that these were occurring for several minutes, up to 10 minutes in the case of tubulin. Formation of stable binary complexes between CCT and substrate is favoured in the presence of ADP, suggesting that ADP-bound CCT is the in vivo acceptor state (Melki and Cowan, 1994). Lin and Sherman (1997) have examined genetic interactions between mutations in the ATP binding sites of four CCT subunits in Saccharomyces cerevisiae, Cct 1, 2, 3 and 6, and proposed a simple sequential KNF model of co-operative ATP binding progressing in one direction around the CCT ring (Figure 3). They also suggest a functional hierarchy among CCT subunits in their interaction with nucleotide and that nucleotide binding may initiate at a fixed position in a CCT ring. These are important observations because they raise the possibility that CCT could facilitate ordered capture and release of a substrate bound by two or more subunits as suggested by Liou and Willison (1997) and hence add informational content to a folding reaction which might not be realized by simple
Composition and function of the eukaryotic cytosolic
615
diffusion and collision reactions onto and off the chaperonin.
Figure 3 Models for subunit order within a single CCT ring. The top diagram shows a model proposing an unique structure for the CCT ring derived from the examination of CCT microcomplexes, present at low abundance in mouse testis and human cells, which are composed of pairs of CCT subunits. Liou and Willison (1997) used these association patterns to suggest a single solution to the topological arrangement of the 8 CCT subunits in the ring. Independently Lin and Sherman (1997) investigated genetic interactions between various mutations in the conserved ATP binding motifs of the Cct1p, Cct2p,
Molecular chaperones and folding catalysts
616
Cct3p and Cct6p subunits of Saccharomyces cervisiae CCT and discovered a functional hierarchy amongst the subunits. In wild-type CCT the following sequence of co-operative ATP binding/hydrolysis is proposed; ATP→→Cctlp→→Cct3p→→Cct2p→Cct6p. Based upon the nature and strength of the interactions Cct2p and Cct6p are suggested to be neighbouring subunits whereas the number of intervening subunits between Cct1p and Cct3p, Cct3p and Cct2p and Cct6p and Cct1p is undetermined. The two models are mutually consistent.
4.2. Co-translational Folding The role of molecular chaperones in folding of newly synthesized proteins in the eukaryotic cytosol is discussed in detail in the chapter by Welch et al., this volume. We therefore restrict our discussion to the specific role of CCT in this process. As already mentioned above, Frydman and Hartl have presented a model for early events in the CCT (TRiC) folding pathway (Frydman et al., 1994; Frydman and Hartl, 1996; Hartl, 1996) suggesting that TRiC captures and folds actin co-translationally and that there may be functional coupling between the Hsp70/ Hsp40 chaperone machine and TRiC. Our laboratory also found Hsp70 co-purifying with CCT (Lewis et al., 1992; Kubota et al., 1994) and it is possible that Hsp70 bound substrates could occupy CCT. Frydman and Hartl (1996) show that -actin cannot bind TRiC until approximately 150–200 amino acid residues have been translated; therefore, the co-translational interaction can only last as long as it takes to synthesize the remaining C-terminal portion of -actin, say 10–20 seconds, depending upon translation elongation rates. As we discussed in Section 4.1, CCT and substrates remain in contact for several minutes in vitro and in vivo, so it seems that a significant part of CCT function must occur post-translationally. Eggars et al. (1997) have challenged the model of co-translational folding on TRiC through the use of anti-puromycin antibodies to precipitate nascent polypeptide chains, finding Hsp70 associated with them but not any CCT subunits. It is important that TRiC be shown to associate with polysomes in vivo by other techniques hopefully including electron microscopy. 4.3. Cell Biology In addition to its role in the folding of newly synthesized proteins, CCT might have other important functions in cell metabolism. Knowledge of these functions, however, is only emerging. Brown et al. (1996) have shown that antibodies to CCT subunits bind to centrosomes and can inhibit centrosome function after microinjection into living cells. This result implicates CCT in microtubule assembly processes, possibly via interactions with -tubulin, a known CCT binding protein (Melki et al., 1994; Moudjou et al., 1996). Studies in neuronal cells suggest roles for subsets of CCT proteins in neurites (Roobol et al., 1995), possibly in transport of actin or tubulin monomers (Campenot et al., 1996). The human CCTd subunit has appeared as a factor involved in the stimulation of HIV1 TAR RNA (Wu-Baer et al., 1996) suggesting roles for individual CCT subunits in the life cycle of HIV.
Composition and function of the eukaryotic cytosolic
617
5. CONCLUSIONS CCT appears to be an important component of the protein folding machinery in the eukaryotic cytosol. CCT is made from eight, essential, gene products and present data suggest it is rather specific for folding of actins and tubulins, although there are strong hints of other activities and functions. Hopefully, progress on the struc-tural front will be made over the next few years and this will illuminate the mechanism of CCT activity. Perhaps GroEL and CCT are as similar and different as the Type I and Type II restriction enzymes of bacteria, rather unspecific versus site-specific proteins. 6. ACKNOWLEDGMENTS This work is funded by the Cancer Research Campaign. I thank Hiroshi Kubota for help with Figure 1 and Michel Leroux and Jonathan Hodgkin for help with Figure 2 and Sylvia Holt for help with the manuscript. 7. REFERENCES Andra, S., Frey, G., Nitsch, M., Baumeister, W. and Stetter, K.O. (1996). Purification and structural characterization of the thermosome from the hyperthermophilic archaeum Methanopyrus kandleri. FEBS Lett. , 379 , 127–131. Archer, J.E., Vega, L.R. and Solomon, F. (1995). Rbl2, a yeast protein that binds to tubulin and participates in micro tubule function in vivo. Cell , 82 , 425–434. Ashworth, A., Malik, A.N., Walkley, N.A., Kubota, H., and Willison, K.R. (1994). The Tcp-1 related gene, Cctg, maps to mouse Chromosome 3. Mammalian Genome , 5 , 509–510. Boisvert, B.C., Wang, J., Otwinowski, Z., Horwich, A.L. and Sigler, P.B. (1996). The 2.4 A crystal structure of the bacterial chaperonin GroEL complexed with ATP S. Nature Struct. Biol. , 3(2), 170–177. Braig K., Otwinowski, Z., Hegde, R., Boisvert, D.C.Joachimiak A., Horwich, A.L. and Sigler, P.B. (1994). The crystal structure of the bacterial chaperonin GroEL at 2.8 Å. Nature , 371 , 578–586. Brown, C.R., Doxsey, S.J., Hongbrown, L.Q., Martin, R.L. and Welch, W.J. (1996). Molecular Chaperones and the Centrosome—a Role For TCP-1 In Microtubule Nucleatin. J. Biol. Chem. , 271 , 824–832. Bult, C.J. plus 40 authors (1996). Complete Genome Sequence of the Methanogenic Archaeon, Methanococcus jannaschii. Science , 273 , 1058–1073. Campenot, R.B., Lund, K. and Senger, D.L. (1996). Delivery of Newly Synthesized Tubulin to Rapidly Growing Distal Axons of Rat Sympathetic Neurons in Compartmented Cultures. J. Cell Biol. , 135(3), 701–709. Campo, R., Fontalba, A., Sanchez, L.M. and Zabala, J C. (1994). A 14 kDa release factor is involved in GTP-dependent -tubulin folding. FEBS Lett. , 353 , 162–166 Chen, X., Sullivan, D.S., and Huffaker, T.C. (1994). Two yeast genes with similarity to
Molecular chaperones and folding catalysts
618
TCP-1 are required for microtubule and actin function in vivo. Proc. Natl. Acad. Sci. USA , 91 , 9111–9115. Chen, S., Roseman, A.M., Hunter, A.S., Wood, S.P., Burston, S.G., Ranson, N.A., Clarke, A.R., and Saibil, H.R. (1994). Location of a folding protein and shape changes in GroEL-GroES complexes imaged by cryo-electron microscopy. Nature , 371 , 261– 264. Counsell, D.J. and Willison, K.R. (1998). Bioinformatic prediction of substrate binding sites on Group II chaperonins, submitted for publication. Ditzel, L., Lowe, J., Stock, D., Stetter, K.-O., Huber, H., Huber, R. and Steinbacher, S. (1998). Crystal Structure of the Thermosome, the Archaeal Chaperonin and Homologue of CCT. In press. Dobrzynski, J.K., Sternlicth, M.L., Farr, G.W. and Sternlicht, H. (1996) NewlySynthesized -Tubulin Demonstrates Domain-Specific Interactions with the Cytosolic Chaperonin. Biochem . 35 , 15870–15882. Doolittle, W.F. (1996). At the core of the Archaea. Proc. Natl Acad. Sci. USA , 93 , 8797–8799. Dunn, M.K. and Mercola, M. (1996). Cloning and expression of Xenopus CCT , a chaperonin subunit developmentally regulated in neural-derived and myogenic lineages. Developmental Dynamics , 205 , 387–394. Eggars, D.K., Welch, W.J. and Hansen, W.J. (1997). Complexes between Nascent Polypeptides and their Molecular Chaperones in the Cytosol of Mammalian Cells. Mol. Biol. of the Cell , 8 , 1559–1573. Farr, G.W., Schar, I.E-C., Schumacher, R.J., Sandek, S. and Horwich, A.L. (1997) Chaperonin-mediated folding in the eukaryotic cytosol proceeds of release of native and nonnative forms. Cell , 89 , 927–937. Flaherty, K.M., McKay, D.B., Kabsch, W. and Holmes, K.C. (1991). Similarity of the three-dimensional structures of actin and the ATPase fragment of a 70-kDa heat shock cognate protein. Proc. Natl. Acad. Set. USA , 88 , 5041–5045. Frydman, J., Nimmesgern, E., Erdjument-Bromage, H., Wall, J.S., Tempst, P., and Hartl, F.-U. (1992). Function in protein folding of TRiC, a cytosolic ring complex containing TCP-1 and structurally related subunits. EMBO J. , 11 , 4767–4778. Frydman, J., Nimmesgern, E., Ohtsuka, K., and Hartl, F.-U (1994) Folding of nascent polypeptide chains in a high molecular mass assembly with molecular chaperones. Nature , 370 , 111–117 Frydman, J. and Hartl, F.-U. (1996). Principles of Chaperone-Assisted Protein Folding: Differences between in vitro and in vivo mechanisms. Science , 272 , 1497–1502. Gao, Y., Thomas, J.O., Chow, R.L., Lee, G.-H., and Cowan, N.J. (1992). A cytoplasmic chaperonin that catalyzes β-actin folding. Cell , 69 , 1043–1050. Gao, Y., Vainberg, I.E., Chow, R.L. and Cowan, N.J. (1993). Two cofactors and cytoplasmic chaperonin are required for the folding of - and -tubulin. Mol. Cell. Biol. , 13 , 2478–2485. Gao, Y., Melki, R., Walden, P.D., Lewis, S.A., Ampe, C., Rommelaere, H., Vandekerckhove, J., and Cowan, N.J. (1994). A novel cochaperonin that modulates the ATPase activity of cytoplasmic chaperonin. J. Cell Biol. , 125 , 989–996. Hartl, F.-U. (1996). Molecular chaperones in cellular protein folding. Nature , 381 , 571–
Composition and function of the eukaryotic cytosolic
619
580. Hemmingsen, S.M., Woolford, C., van der Vies, S.M., Tilly, K., Kennis, D.T., Georgopoulos, C.P., Hendrix, R.W. and Ellis, R.J. (1988). Homologous plant and bacterial proteins chaperone oligomeric protein assembly. Nature , 333 , 330–334. Hirata, D., Masuda, H., Eddison, M. and Toda, T. (1998). Essential role of tubulinfolding cofactor D in microtubule assembly and its association with microtubules in fission yeast. EMBO J. , 17 , 658–666. Horwich, A.L. and Willison, K.R. (1993). Protein folding in the cell: functions of two families of molecular chaperone hsp 60 and TF55-TCP1. Phil. Trans. R. Soc. Lond. , 339 , 313–325. Hunt, J.F., Weaver, A.J., Landry, S.J., Gierasch, L., and Deisenhofer, J. (1996). The crystal structure of the GroES co-chaperonin at 2.8 Å resolution. Nature , 379 , 37–45. Hunt, J.F., Saskia, M., van der Vies, L.H. and Deisenhofer, J. (1997). Structural Adaptations in the Specialized Bacteriophage T4 Co-Chaperonin Gp31 Expand the Size of the Anfinsen Cage. Cell , 90 , 361–371. Hynes, G., Kubota, H. and Willison, K.R. (1995). Antibody characterisation of two distinct conformations of the chaperonin containing TCP-1 from mouse testis. FEBS Lett. , 358 , 129–132. Hynes, G., Sutton, C.W., U, S., Willison, K.R. (1996a). Peptide mass fingerprinting of chaperonin-containing TCP-1 (CCT) and copurifying proteins. FASEB Journal 10(1), 137–147. Hynes, Gillian., Celis, Julio E., Lewis, Victoria, A., Carne, Alexander., U, Sally., Lauridsen, Jette B. and Willison, Keith R. (1996b). Analysis of chaperonin-containing TCP-1 subunits in the human keratinocyte two-dimensional protein database: Further characterisation of antibodies to individual subunits. Electrophoresis , 17 , 1720–1727. Kagawa, H.K., Osipiuk, J., Maltsev, N., Overbeek, R., Quate-Randall, E., Joachimiak, A., and Trent, J.D. (1995). The 60kDa Heat Shock Proteins in the Hyperthermophilic Archaeon Sulfolobus shibatae. J. Mol. Biol. , 253 , 712–725. Kim, S., Willison, K.R. and Horwich, A.L. (1994). Cytosolic chaperonin subunits have a conserved ATPase domain but diverged polypeptide-binding domains. Trend. Biochem. Set. , 19 , 543–548. Kirchhoff, C. and Willison, K. (1990). Nucleotide and amino-acid sequence of human testis-derived TCP1. Nucleic Acids Res. , 18 , 4247. Klenk, H.-P. and 50 authors, (1997). The complete genome sequence of the hyperthermophilic, sulphate-reducing archaeon Archaeoglobus fulgidus. Nature , 390 , 364–370. Klumpp, M., Baumeister, W. and Essen, L-O. (1997). Structure of the Substrate Binding Domain of the Thermosome, an Archael Group II Chaperonin. Cell , 91 , 263–270. Knapp, S., Schmidt, K.I., Hebert, H., Bergman, T., Jörnvall, H. and Ladenstein, R. (1994). The molecular chaperonin TF55 from the thermophilic archaeon Sulfolobus solfataricus: A biochemical and structural characterization. J. Mol. Biol. , 242 , 397– 407. Kubota, H., Hynes, G., Carne, A., Ashworth, A. and Willison, K. (1994). Identification of six Tcp-1-related genes encoding divergent subunits of the TCP-1-containing chaperonin. Curr. Biol . 4 , 89–99.
Molecular chaperones and folding catalysts
620
Kubota, H., Hynes, G. and Willison, K. (1995a). The chaperonin containing t-complex polypeptide 1 (TCP-1): Multisubunit machinery assisting in protein folding and assembly in the eukaryotic cytosol. Eur. J. Biochem. , 230 , 3–16. Kubota, H., Hynes, G., and Willison, K. (1995b). The eighth Cct gene, Cctq, encoding the theta subunit of the cytosolic chaperonin that contains TCP-1. Gene , 154 , 231– 236. Kubota, H., Hynes, G.M., Kerr, S.M. and Willison, K.R. (1997a). Tissue-specific subunit of the mouse cytosolic chaperonin-containing TCP-1. FEBS Letters , in press. Kubota, H. and Willison, K.R. (1997b). Introduction and 8 sections on CCT for Guidebook to Molecular Chaperones. (Ed.) M.-J.Gething, Sambrook/Tooze publications at Oxford University Press, Oxford. Lake, J.A. (1991) Tracing origins with molecular sequences: metazoan and eukaryotic beginnings. TIBS , 16 , 46–50. Leroux, M.R. and Candido, E.P.M. (1995). Characterization of Four New tcp-1-Related cct Genes from the Nematode Caenorhabditis elegans. DNA and Cell Biology , 14(11), 951–960. Lewis, V.A., Hynes, G.M., Zheng, D., Saibil, H. and Willison, K. (1992). T-complex polypeptide-1 is a subunit of a heteromeric particle in the eukaryotic cytosol. Nature , 358 , 249–252. Lewis, S.A., Tian, G., Vainberg, I.E. and Cowan, N.J. (1996). Chaperonin-mediated folding of actin and tubulin. J. Cell Biol. , 132 , 1–4. Lin, P. and Sherman, F. (1997) The unique hetero-oligomeric nature of the subunits in the catalytic cooperativity of the yeast Cct chaperonin complex. Proc. Natl. Acad. Sci. USA , 94 , 10780–10785. Lingappa, J.R., Martin, R.L., Wong, M.L., Gamen, D., Welch, W.J. and Lingappa, V.R. (1994). A eukaryotic cytosolic chaperonin is associated with a high molecular weight intermediate in the assembly of hepatitis B virus capsid, a multimeric particle. J. Cell Biol. , 125 , 99–111. Liou, K.F., McCormack, E.A. and Willison, K.R. (1998). The Chaperonin containing TCP-1 (CCT). displays a single-ring mediated disassembly and reassembly cycle (in press). Biological Chem. Hoppe-Seyler. Liou, A.K.F. and Willison, K.R. (1997) Elucidation of the subunit orientation in CCT (chaperonin containing TCP1) from the subunit composition of CCT micro-complexes. EMBO J. , 16 , 4311–4316. Llorca, O., Smyth, M.G., Marco, S., Carrascosa, J.L., Willison, K.R. and Valpuesta, J.M. (1998). ATP binding induces large conformational changes in the apical and equatorial domains of the eukaryotic chaperonin containing TCP-1 complex. J. Biol. Chem. , 273 , 1–4. Marco, S., Carrascossa, J.L. and Valpuesta, J.M. (1994a). Reversible interaction of actin along the channel of the TCP-1 cytoplasmic chaperonin. Biophys. J. , 67 , 364– 368. Marco, S., Urena, D., Carrascosa, J.L., Waldmann, T., Peters, J., Hegerl, R., Pfeifer, G., Sack, K.H. and Baumeister, W. (1994b). The molecular chaperone TF55: Assessment of symmetry. FEBS Lett. , 341 , 152–155. Martel, R., Cloney, L.P., Pelcher, L.E. and Hemmingsen, S.M. (1990). Unique
Composition and function of the eukaryotic cytosolic
621
composition of plastid chaperonin-60 and b polypeptide-encoding genes are highly divergent. Gene , 94 , 181–187. Martin, W.H. and Creutz, C.E. (1987). Chromobindin A: a Ca2+ and ATP regulated chromaffin granule binding protein. J. Biol. Chem. , 262 , 2803–2810. Mehra, V., Sweetser, D. and Young, R.A. (1986). Efficient mapping of protein antigenic determinants. Proc. Natl. Acad. Sci. USA , 83 , 7013–7017. Melki, R. and Cowan, N.J. (1994). Facilitated folding of actins and tubulins occurs via a nucleotide-dependent interaction between cytoplasmic chaperonin and distinctive folding intermediates. Mol. Cell Biol. , 14 , 2895–2904. Melki, R., Rommelaere, H., Leguy, R., Vandekerckhove, J. and Ampe, C. (1996) Cofactor A Is a Molecular Chaperone Required for -Tubulin Folding: Functional and Structural Characterization. Biochemistry , 35(32), 10422–10435. Miklos, D., Caplan, S., Martens, D., Hynes, G., Pitluk, Z., Brown, C., Barrell, B., Horwich, A.L. and Willison, K. (1994). Primary structure and function of a second essential member of heterooligomeric TCP1 chaperonin complex of yeast, TCP1 . Proc. Natl. Acad. Sci. USA , 91 , 2743–2747. Mori, M., Murata, K., Kubota, H., Yamamoto, A., Matsushiro, A. and Morita, T. (1992). Cloning of a cDNA encoding the Tcp-1 (t complex polypeptide 1) homologue of Arabidopsis thaliana. Gene , 122 , 381–382. Moudjou, M., Bordes, N., Paintrand, M. and Bornens, M. (1996). -Tubulin in mammalian cells: the centrosomal and the cytosolic forms. J. of Cell Science , 109 , 875–887. Nitsch, M., Klumpp, M., Lupas, A. and Baumeister, W. (1997) The Thermosome: Alternating a and -Subunits Within the Chaperonin of the Archaeon Thermoplasma acidophilum. J. Mol. Biol , 267 , 142–149. Phipps, B., Hoffmann A., Stetter K.O. and Baumeister, W. (1991). A novel ATPase complex selectively accumulated upon heat shock is a major cellular component of thermophilic archaebacteria. EMBO J. , 10 , 1711–1722. Phipps, B.M., Typke, D., Hegerl, R., Volker, S., Hoffmann, A., Stetter, K.O. and Baumeister, W. (1993). Structure of a molecular chaperone from a thermophilic archaebacterium. Nature , 361 , 475–477. Reading, D.A., Hallberg, R.L. and Myers, A.M. (1989). Characterization of the yeast HSP60 gene coding for a mitochondrial assembly factor. Nature , 337 , 655–659. Rommelaere, H., van Troys, M., Gao, Y., Melki, R., Cowan, N.J., Vandekerckhove, J. and Ampe, C. (1993). Eukaryotic cytosolic chaperonin contains t-complex polypeptide 1 and seven related subunits. Proc. Natn. Acad. Sci. USA , 90 , 11975–11979. Roobol, A. and Garden, M.J. (1993). Identification of chaperonin particles in mammalian brain cytosol and t-complex polypeptide 1 as one of their components. J. Neurochem. , 60 , 2327–2330. Roobol, A., Holmes, F.E., Hayes, N.V.L., Baines, A.J. and Garden, A.J. (1995) Cytoplasmic chaperonin complex enter neurites developing in vitro and differ in subunit composition within single cells. J. Cell Sci . 107 , 1477–1488. Saibil, H.R. (1996) Chaperonin Structure and Conformational Changes. In The Chaperonins , Academic Press Inc. 107–135 Saitou, N. and Nei, M. (1987). Neighbor-joining method: A new method for
Molecular chaperones and folding catalysts
622
reconstructing phylogenetic trees. Mol Biol. Evol. , 4 , 406–425. Soares, H., Penque, D., Mouta, C. and Rodrigues-Pousada, C. (1994). A Tetrahymena orthologue of the mouse chaperonin subunit CCT and its coexpression with tubulin during cilia recovery. J. Biol. Chem. , 269 , 29299–29307. Sternlicht, H., Farr, G.W., Sternlicht, M.L., Driscoll, J.K., Willison, K. and Yaffe, M.B. (1993). The tcomplex polypeptide 1 complex is a chaperonin for tubulin and actin in vivo. Proc. Natl. Acad. Sci. USA , 90 , 9422–9426. Stoldt, V., Rademacher, F., Kehren, V., Ernst, J.F., Pearce, D.A. and Sherman, F. Review (1996): The Cct Eukaryotic Chaperonin Subunits of Saccharomyces cerevisiae and Other Yeasts. Yeast , 12 , 523–529. Sun, H.B., Neff, A.W., Mescher, A.L. and Malacinski, G.M. (1995). Expression of the axolotl homologue of mouse chaperonin t-complex protein 1 during early development. Biochem. Biophys. Acta , 1260 , 157–166. Tian, G., Lewis, S.A., Feierbach, B., Stearns, T., Rommelaere, H., Ampe, C. and Cowan, N.J. (1997). Tubulin subunits exist in an activated conformational state generated and maintained by protein cofactors. J. Cell Biol. , 138 , 821–832. Tian, G., Vainberg, I.E., Tap, W.D., Lewis, S.A. and Cowan, N.J. (1995a). Specificity in chaperonin-mediated protein folding. Nature , 375 , 250–253. Tian, G., Vainberg, I.E., Tap, W.D., Lewis, S.A. and Cowan, N.J. (1995b). Quasi-native chaperonin-bound intermediates in facilitated protein folding. J. Biol. Chem. , 270 , 23910–23913. Tian, G., Huang, Y., Rommelaere, H., Vandekerckhove, J., Ampe, C. and Cowan, N.J. (1996). Pathway Leading to Correctly Folded -Tubulin. Cell 86 , 287–296. Trent, J.D., Nimmesgern, E., Wall, J.S., Hartl, F.-U. and Horwich, A.L. (1991). A molecular chaperone from a thermophilic archaebacterium is related to the eukaryotic protein t-complex polypeptide-1. Nature , 354 , 490–493. Trent, J.D., Gabrielsen, M., Jensen, B., Neuhard, J. and Olsen, J. (1994). Acquired thermotolerance and heat shock proteins in thermophiles from the three phylogenetic domains. J. Bacteriol. , 176 , 6148–6152. Ursic, D. and Ganetzky, B. (1988). A Drosophila melanogaster gene encodes a protein homologus to the mouse t complex polypeptide 1. Gene , 68 , 267–274. Ursic, D. and Culbertson, M.R. (1991). The yeast homolog to mouse Tcp-1 affects microtubule-mediated processes. Mol. Cell Biol. , 11 , 2629–2640. Ursic, D., Sedbrook, J.C., Himmel, K.L. and Culbertson, M.R. (1994). The essential yeast Tcp1 affects actin and microtubules. Mol. Biol. Cell , 5 , 1065–1080. Venner, T.J. and Gupta, R.S. (1990). Nucleotide sequence of mouse HSP60 (chaperonin, GroEL homolog) cDNA. Biophys. Acta , 1087. Vinh, D.B.-N. and Drubin, D.G. (1994). A yeast TCP-1 like protein is required for actin function in vivo. Proc. Natl. Acad. Sci. USA , 91 , 9116–9120. Waldmann, T., Nimmesgern, E., Nitsch, M., Peters, J., Pfeifer, G., Muller, S., Kellermann, J., Engel, A., Hartl, F.-U. and Baumeister, W. (1995a). The thermosome of Thermoplasma acidophilum and its relationship to the eukaryotic chaperonin TRiC. Eur. J. Biochem. , 227 , 848–856. Waldmann, T., Lupas, A., Kellerman, J., Peters, J. and Baumeister, W. (1995b). Primary structure of the thermosome from Thermoplasma acidophilum. Biol. Chem. Hoppe-
Composition and function of the eukaryotic cytosolic
623
Seyler , 376 , 119–126. Willison, K.R. and Kubota, H. (1994). The structure, function and genetics of the chaperonin containing TCP-1 (CCT) in eukaryotic cytosol. In The Biology of Heat Shock Proteins and Molecular Chaperones. R. I.Morimoto, A.Tissieres and C.Georgopoulos eds., (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press), pp. 299–312. Willison, K.R. and Horwich, A.L. (1996). Structure and Function of Chaperonins in Archaebacteria and Eukaryotic Cytosol. In The Chaperonins , Academic Press Inc. 107–135. Wu-Baer, F., Lane, W.S. and Gaynor, R.B. (1996) Identification of a group of cellular cofactors that stimulate the binding of RNA polymerase II and TRP-185 to human immunodeficiency virus 1 TAR RNA. J. Biol. Chem. , 271 , 4201–4208. Xu, Z., Horwich, A.L. and Sigler, P.B. (1997) The crystal structure of the asymmetric GroEL-GroES-(ADP) 7 chaperonin complex. Nature , 388 , 741–750. Yaffe, M.B., Farr, G.W., Miklos, D., Horwich, A.L., Sternlicht, M.L. and Sternlicht, H. (1992). TCP1 complex is a molecular chaperone in tubulin biogenesis. Nature , 358 , 245–248.
25. STRUCTURE AND MECHANISM OF HSP70 PROTEINS JEUNG-HOI HA, ERIC R.JOHNSON, DAVID B.McKAY*, MARCELO C.SOUSA, SHIGEKI TAKEDA and SIGURD M.WILBANKS Department of Structural Biology, Stanford University School of Medicine, Stanford, California 94305–5400, USA
1. Introduction 2. Structure 3. ATPase Activity 4. Coupling of the ATPase Activity to the Peptide Binding Actitivity 5. Modulation of Hsp70 Activity 5.1. Self-association 5.2. Posttranslational Modification 5.3. Accessory Proteins 6. Epilogue 7. References 1. INTRODUCTION The 70 kDa heat shock related proteins (referred to generically as “Hsp70s” in this chapter—taken to include the endoplasmic reticulum-resident BiPs, cytoplasmic/ nuclear Hsc70s, prokaryotic DnaKs and others) comprise a family of molecular chaperones that are essential for cell viability. Their biochemical and biological functions are discussed in depth in several chapters of this volume, and hence they will not be elaborated extensively here. In brief, the Hsp70s cyclically bind and release “unstructured” segments of polypeptides; the binding presumably suppresses aggregation or misfolding of the peptide segments, while their release gives them the opportunity to coalesce into a correctly folded tertiary structure. The kinetics of the binding and release cycle is modulated by ATP/ADP; in the presence of ATP, peptide release is relatively rapid, while in the presence of ADP, Hsp70-peptide complexes are stable and long-lived. In vivo, unstructured segments of polypeptide may present themselves in many different contexts ranging from nascent polypeptides emerging from ribosomes during translation (Beckmann et al., 1990; Hartl, 1996; Welch et al., this volume), *Corresponding author
Molecular chaperones and folding catalysts
626
polypeptides translocating across membranes into intracellular organelles (Scherer et al., 1990; Schatz and Dobberstein, 1996; chapter in this volume by Welch et al.; Haas and Zimmermann; Dekker and Pfanner; Muckel and Soll) and protein denaturation (e.g. heatdenatured proteins arising from thermal stress of cells (Skowyra et al., 1990; Li et al., this volume). The focus of this chapter is the structure and biochemical mechanism of the Hsp70 proteins. In this context, it is convenient to delineate the in vitro activities of the Hsp70 proteins as (i) an ATPase activity; (ii) a peptide binding activity (discussed in detail by Buchberger et al. in this volume); (iii) a mechanism of coupling binding and release of peptides to the enzymatic ATPase cycle; and (iv) mechanisms of modulating the Hsp70 activities through interaction through self-association, posttranslational modification and interaction with accessory proteins. For a detailed description of the functional cycle of the procaryotic DnaK homolog, and in particular the role of the accessory proteins including DnaJ the reader is referred to the chapter by Buchberger et al. 2. STRUCTURE Approximately 100 amino acid sequences (derived from DNA sequences) are currently known for Hsp70-related proteins (Swiss-prot release 34.0, 4/96). The mature proteins are typically ~640–650 amino acid residues in length. The sequences show a high degree of homology, with the most widely separated prokaryotic and eukaryotic sequences sharing ~50% amino acid identity. The ATPase activity is localized in the amino-terminal ~385 residues, while the peptide binding activity resides in ~160 residues following the ATPase domain (Figure 1). The carboxy terminal regions of the proteins are the most variable in sequence; the eukaryotic Hsp70s have a glycine, proline rich segment of variable length and with an approximate (GGMP)n repeat whose function is still unknown. Additionally, the non-organellar eukaryotic Hsp70s end with a conserved EEVD sequence whose presence is important for stable binding of model denatured polypeptide substrates by Hsp70s (Freeman et al., 1995). The X-ray crystallographic structure of the ATPase domain of bovine Hsc70 is known (Flaherty et al., 1990) (Figure 2). It consists of two lobes with a cleft between them; nucleotide is bound at the base of the cleft and is almost entirely inaccessible to solvent. The tertiary structure of the ATPase domain is highly similar to that of DnaK (Harrison, et al., 1997) and human Hsp70 (Sriram, et al., 1997). The first available structure of an Hsp70 ATPase domain clearly showed a striking similarity to that of actin
Figure 1 Schematic delineation of functional domains within the primary
Structure and mechanism of Hsp70 proteins
627
sequence of Hsp70 proteins.
Figure 2 Structure of the ATPase domain, (a) Schematic drawing of the crystallographic structure of bovine Hsc70 ATPase domain with MgADP, Pi and two K+ ions. Separate structural domains are shown
Molecular chaperones and folding catalysts
628
in separate colors. Mg2+ ion is yellow, K+ ions are blue. Drawing made with Molscript (Kraulis, 1991). (b) Folding topology, with domains in same relative orientation as in (a) Secondary structure assignment determined with the program DSSP (Kabsch et al., 1983); numbering referenced to bovine Hsc70 primary sequence.
(Kabsch et al., 1990); further, the structural domains of Hsc70 and actin that are involved in nucleotide binding have identical folding topology and a similar tertiary structure to the equivalent regions in hexokinase (Fletterick et al., 1975) and glycerol kinase (Hurley et al., 1993). Although the overall sequence similarity between actin and Hsc70 is relatively low, many of the residues that interact with MgADP and MgATP are identical, raising the possibility of parallels between the two families of proteins in the manner in which they utilize ATP in their respective functions (Flaherty et al., 1991). A proposal was made several years ago, based on a weak amino acid sequence similarity, that the structure of the peptide binding domain of Hsp70s would be similar to that of the peptide binding portion of major histocompatibiltiy complex (MHC) proteins (Flajnik et al., 1991; Rippmann et al., 1991). This model was appealing since it rationalized the interaction of Hsp70 proteins with unstructured segments of polypeptides. A shadow of doubt first fell on this model when NMR studies of an Hsc70 peptide binding domain led to a secondary structure assignment and folding topology that was fundamentally different from that of MHC peptide binding domains (Morshauser et al., 1995). Recently, the X-ray crystallographic structure of the DnaK peptide binding domain (residues 389–607) complexed with a seven residue peptide (Asn-Arg-Leu-LeuLeu-Thr-Gly), has revealed a novel tertiary fold that is unrelated to the MHC structure (Zhu et al., 1996) (Figure 3). The peptide binding moiety is composed of an 8-stranded antiparallel structure which forms a crevice with a single, deep pocket that accommodates the central leucine of the bound peptide. Although the secondary structure diagram depicts a pair of four-stranded sheets, only one of these segments (strands 3 and 6–8, Figure 3) forms a continuous sheet. In the other segment, the first two strands (strands 1 and 2, Figure 3) arc over the top of the peptide, while the other two strands (strands 4 and 5, Figure 3) run under it. Although biochemical studies suggested that a minimum of seven residues was required for optimal peptide affinity (Flynn et al., 1991), only the central five residues of the peptide make substantial interaction with the peptide binding domain in the crystallographic structure. One residue appears to be “buried” in a binding pocket of the protein; it is the central leucine of the peptide, which is sandwiched between the two pairs of beta strands, and whose side chain protrudes downward into the crevice between the strands in the view shown in Figure 3. A long alpha helix lies across the crevice, but does not actually contact the peptide; it has been suggested that this helix acts as a “latch” to stabilize peptide complexes and to control the kinetics of peptide binding and release. This suggestion implies that the crystal structure shows the stable peptide binding form and that the transient binding form has yet to be observed. The functional imperatives of Hsp70 proteins to recognize a diversity of unstructured polypeptides and to modulate their peptide binding and release rates is consonant with the structure of the DnaK peptide binding domain. The hydrophobic binding pocket for the peptide rationalizes the preference for aliphatic hydrophobic amino acids in peptides bound by Hsp70s (Flynn et al., 1991; Rüdiger et al., 1997a, b; see chapter Buchberger et
Structure and mechanism of Hsp70 proteins
629
al., this volume). The long helix that lies over the top of the peptide binding crevice suggests a mechanism whereby peptide binding and release would be relatively slow under circumstances where the helix
Figure 3 Structure of the peptide binding domain, (a) Schematic Molscript drawing of the crystallographic structure of E. coli; DnaK peptide binding domain. Peptide is shown as a ball-and-stick model, with
Molecular chaperones and folding catalysts
630
oxygen, red; nitrogen, blue; carbon, gray, (b) Folding topology. Secondary structure assignment determined with the program DSSP; numbering referenced to E. coli DnaK primary sequence.
was latched over the crevice, but relatively rapid if the helix swung outward to allow unhindered access to the crevice. The helix is bound to the peptide binding groove through the polypeptide backbone at the N-terminus and by salt bridges at the Cterminus. Disruption of one of these salt bridges destabilizes peptide binding (Ha, et al., 1997), supporting the hypothesis that the helix serves as a latch controlling access of the peptide to the crevice. 3. ATPase ACTIVITY The steady state ATPase turnover rate (Table 1) and nucleotide affinity (Table 2) have been reported for several Hsp70-related proteins, as well as for the aminoterminal ATPase fragment of bovine and rat Hsc70 (Table 3). The general consensus that emerges from these studies is, (i) the steady state ATPase turnover rate is relatively slow; reported kcat values range 0.02 min-1 to 1.0 min-1, with the faster rates generally measured under conditions of peptide stimulation of the ATPase activity; steady state Km values for MgATP range ~0.1–1 M; (ii) within the enzymatic cycle, ATP hydrolysis is also relatively slow and under many conditions, it appears to be the rate-limiting step of the cycle (Gao et al., 1993; McCarty et al., 1995; Theyssen et al., 1996) (although, by contrast, Pi release is rate-limiting for the isolated ATPase domain of Hsc70 (Ha et al., 1994)); (iii) affinities of both MgATP and MgADP for Hsp70 proteins is submicomolar, with reported values ranging 0.01–1 M; in absence of divalent ion, the affinities are significantly weaker, ~1–10 M; (iv) following ATP hydrolysis, at least for Hsc70 product release is ordered, with Pi preceeding ADP (Ha et al., 1994). The ATPase activity requires Mg2+; the K0.5 for the Mg2+-dependence of the activity is ~1–10 M (Wilbanks et al., 1994). The ATPase activity is also dependent on monovalent ion with a K0.5 20–50 M; K+ (ionic radius=1.33 Å) induces maximal activity, while significantly smaller (e.g. Na+ or Li+ with ionic radii 0.97 Å and 0.68 Å respectively) or larger (e.g. Cs+, ionic radius=1.67 Å) monovalent ions give a substantially lower ATPase activity (O’Brien et al., 1995). A pathway for chemical hydrolysis of ATP has been suggested (Figure 4), based on both the effects of mutagenesis of residues in the nucleotide binding cleft (Table 3) and the comparison of crystallographic structures of the Hsc70 ATPase fragment that have ATP or slowly hydrolyzable analogs bound (AMPPNP bound to wild type protein; ATP bound to several different mutant proteins) to those that have ADP bound (Flaherty et al., 1994). In the suggested scenario, MgATP would initially bind in a conformation that was not correctly aligned for hydrolysis, possibly with an H2O molecule intercalated between the Mg2+ ion and the phosphate of the nucleotide, as observed crystallographically. Then, the phosphate would rearrange to form a “tight” , -bidentate complex with the Mg2+ ion. This is proposed to align the terminal phosphate of the nucleotide for in-line attack by an H2O molecule or OH ion, ultimately yielding ADP and Pi as products. Model-building of a hypothetical transition state of the hydrolysis reaction suggested an
Structure and mechanism of Hsp70 proteins
631
experimentally observed H2O molecule that is hydrogen bonded to the -amino group of the side chain of Lys71 in the Hsc70 ATPase fragment as a candidate catalytic H2O molecule. Lys71 could then act either to accept a proton from the H2O molecule or to stabilize an OHTable 1 Kinetic parameters of unstimulated and substrate-stimulated steady state ATPase activities of wild-type Hsp70s derived from data in the literature. Sequence of peptide A: KRQIYTDLEMNRLGK: a-pp peptide: CALLOSRLLLSAPRAAATARA: peptide C: KLIGVLSSLFRPK: peptide A7: RRLIEDAETAARG: S peptide: KETAAAKFEROHMDSSTSAA: cytochrome C peptide: IFAGIKKKAERADLIAYLKQATAK; f-AF1 fluorescen-FYQLALT: Mpp: MALLQSRLLLSAPRAAATARA.
# Protein
Ligand (conc.) (1)
k cat K M ( Assay (min-1) M) condition
1 Bovine Hsc70 natural
0.5 mg/ml triton-washed clathrin cages
1
0.7
20 mM (Braell et al., HEPES 1984) 25 mM KC1 10 mM (NH4) 2SO4 2mMMg (OAc)2 0.8 mM DTT pH 7.0, 37°C
2 Bovine Hsc70 natural
vesicular stomatitis virus glycoprotein peptides
»0.2
nd
(Flynn et al., 20 mM 1989) HEPES 20 mM NaCl 2 mM MgC12 1% sodium cholate 8 M ATP pH 7.0, 37°C
3 Bovine Hsc70 natural
clathrin cages & various peptides derived from clathrin
1
nd
40 mM HEPES 75 mM KC1 4.5 mM Mg (OAc)2 550 M ATP pH 7.0, 37°C
4 Bovine Hsc70
apo cyt C:
natural
0
g/ml
Reference
(DeLuca-Flaherty et al., 1990)
20 mM (Sadis et al., HEPES-KOH 1992) 0.075
1.37
25 mM KC1 10 mM (NH4) 2SO4
Molecular chaperones and folding catalysts 400
# Protein
Ligand (conc.) (1)
5 Bovine Hsc70
peptide C:
natural
0.19
g/ml
k cat (min1)
1.44
K M( M)
632
2mMMg (OAc)2 0.1 mM EDTA 1mM DTT pH 7.0, 37°C
Assay condition
Reference
Reportedly same as (Wang et al., 1993)
no peptide
0.028
nd
1.2mM
0.29
nd
6 Bovine Hsc70 natural
none
0.037±0.002 nd
20 mM imidazole (Gao et al., 25 mM KCl 1994) 10 mM (NH4)2SO4 2 mM Mg(OAc)2 1mM DTT 0.1mM ATP pH 7.0, 25°C
7 Bovine Hsc70
no peptide
»0.02
20 mM imidazole
natural
8 Bovine Hsc70 natural
9 Bovine Hsc70 natural
nd
#3 above.
1 mM of:
25 mM KCl
cyt C peptide
»0.12
10 mM (NH4)2SO4
peptide C
»0.08
2mM Mg (OAc)2
S peptide
»0.04
1mM DTT
KFERQ
»0.03
0.2 mM ATP pH 7.0, 25°C
faf1
nd
10 mM MOPS
no peptide
0.027±0.013
150 mM KCl 4.5 mM Mg(OAc)2
83 M
0.051±0.007 Vmax
pH 7.0, 25°C
faf1
nd
10 mM MOPS
no peptide
0.32±0.02
150 mM KCl 4.5 mM Mg(OAc)2
83 M
0.20±0.08 Vmax
pH 7.0, 25°C 1:1 cysteine string
(Greene et al., 1995)
(Braun et al., 1996)
(Braun, et al., 1996)
Structure and mechanism of Hsp70 proteins
633
protein
# Protein
Ligand (conc.) (1)
10 Bovine Hsc70
faf1
natural
11 Bovine Hsc70 natural
12 Bovine Hsc70
nd
(Ha, et al., 1997)
150 mM KCl 4.5 mM Mg(OAc)2
80 M
0.2±0.04
50 g/ml bovine serum albumin
152 M
0.29±0.01 khyd, single turnover
2.3 M Hsc70 10 nM ATP pH 7.0, 37°C
faf1
nd
40 mM HEPES
(Ha et al., 1997)
no peptide
0.029±0.0006
150 mM NaCl 4.5mM Mg (OAc)2
152 M
0.06±0.006 khyd, single turnover
50 g/ml bovine serum albumin 2.3 M Hsc70 10 nM ATP pH 7.0, 37°C
faf1
Ligand (conc.) (1)
nd
40 mM HEPES
(Ha et al., 1997)
0.016 to 0.072
150 mM KCl
0.1 to 0.52
4.5 mM Mg (OAc)2
khyd, single turnover
50 g/ml bovine serum albumin 2.3 M Hsc70 10 nM ATP pH 7.0, 37°C
kcat (min- 1 )
13 Bovine Hsc70 faf1 no peptide
Reference
40 mM HEPES
0.03±0.006
152 M
recombinant
K M ( Assay condition M)
no peptide
recombinant no peptide
# Protein
k cat (min- 1 )
K M ( Assay M) condition nd
0.006±0.002
40 mM HEPES 150 mM NaCl 4.5 mM Mg (OAc)2
Reference (Ha et al., 1997)
Molecular chaperones and folding catalysts 152 M
14 Rat Hsc70 recombinant
0.017±0.004 khyd, single turnover
S-peptide:
634
50 g/ml BSA 2.3 M Hsc70 10 nM ATP pH 7.0, 37°C nd
40 mM HEPES
(Wang et al., 1993)
no peptide
0.16
75 mM KCl 4.5 mM Mg (OAc)2
1.2 M
0.38
pH 7.0, 37°C
15 Canine BiP natural
none
0.34
0.11
50 mM MES (Kassenbrock et 25 mM NaCl al., 1989) 1 mM MgCl2 1 mg/ml BSA pH 6.0, 37°C
16 Bovine BiP
no peptide
0.02
nd
20 mM HEPES
natural
20 mM NaCl various peptides
17 Bov. BiP natural
(Flynn et al., 1989)
0.08–0.16
peptide A:
2 mM MgCl2 1 % sodium cholate 8 M ATP pH 7.0, 37°C 20 mM HEPES
no peptide
0.027
20 mM KCl 10 mM (NH4)2SO4
1 mM
0.15
2 mM Mg (OAc)2 0.5 mM DTT 0.25 mM PMSF pH 7.0, 37°C
# Protein
Ligand kcat (min- K M ( (conc.) (1) 1 ) M)
Assay condition
18 Murine BiP
peptide A
20 mM HEPES
(Blond-Elguindi et al., 1993)
Reference (Blond-Elguindi et al., 1993)
Structure and mechanism of Hsp70 proteins recombinant
no peptide: 0.1
0.4
635
20 mM KCl 10 mM (NH4) 2SO4 2 mM Mg(OAc)2 0.5 mM DTT 0.25 mM PMSF pH 7.0, 37°C
1 mM
0.15
19 Hamster BiP recombinant
none
0.4±0.1
1.5±0.2
20 mM HEPES 25 mM KCl 2 mM MgCl2 0.1 mM EDTA 0.5 mM DTT pH 7.0, 37°C
(Wei et al., 1995b)
20 E. coli DnaK overproduced
none
0.15
nd
100 mM HEPES 10 % (v/v) glycerol 35 mM KCl 5 mM MgCl2
(McCarty et al., 1991)
5 mM mercaptoethanol 70 M ATP pH 8.1, 37°C 21 E. coli DnaK overproduced
none
1
200
50 mM Tris-HCl 2 mM MgCl2 0.2 mg/ml BSA 5 mM mercaptoethanol 5% (v/v) glycerol pH 8.0, 30°C
(Zylicz et al, 1983)
# Protein
Ligand (conc.) (1)
K M( kcat 1 (min- ) M)
Assay condition
22 E. coli DnaK
none
0.23
20
30 mM HEPES (Liberek et al., 40 mM KCl 1991a) 50 mM NaCl 7mM Mg (OAc)2 2mM DTT 0.29 mg/ml BSA pH 7.6, 30°C
23 E. coli DnaK
a-pp peptide:
nd
25 mM HEPES (Schmid et al., 1994)
Reference
Molecular chaperones and folding catalysts
636
overproduced
no peptide
0.13
100 mM KCl 5 M DnaK
5 M
0.13
24 E. coli DnaK
peptide C:
overproduced
no peptide
0.035
50 mM KCl 11 mM Mg (OAc)2
500 mM
0.24
pH 7.6, 30°C
pH 7.0, 30°C 3.9±0.7
25 mM HEPES-KOH
25 E. coli DnaK
peptide C:
overproduced
no peptide
0.045
50 mM KCl 10mM Mg (OAc)2
200 M
0.41
200 M ATP pH 7.6, 30°C
no peptide
0.03
26 E. coli DnaK overproduced
# Protein
nd
25 mM HEPES-KOH
nd
50 mM HEPES-KOH 50 mM KCl
0.5 mM pep. 0.12 A
10 mM Mg (OAc)2 0.2 mM ATP
0.5 mM pep. 0.21 C
pH 7.6, 30°C
Ligand kcat (min- 1 ) (conc.)
KM(
M)
(Buchberger et al., 1994b)
(McCarty et al., 1995)
(Kamath-Loeb et al., 1995)
Assay condition
Reference
(Theyssen et al., 1996)
(1)
27 E. coli DnaK
none
overproduced
28 E. coli DnaK
peptide Mpp
overproduced
no peptide
single turnover
0.073
50 mM Tris-HCl 100mM KCl
0.09±0.006
Kd
5 mM MgCl2
steady state
2 mM EDTA
0.09±0.01
2 mM dithioerythritol pH 7.5, 25°C nd
0.02 (NaCl) 0.05 (KCl)
25 mM HEPES 100 mM KCl or NaCl 5 mM MgCl2
(Feifel et al., 1996)
Structure and mechanism of Hsp70 proteins 50 M 29 E. coli DnaK overproduced
30 E. coli DnaK overproduced
# Protein
0.02 (NaCl) 0.15 (KCl)
peptide Mpp
0.4 mM ATP pH 7.0, 25°C nd
25 mM HEPES
(Feifel et al., 1996)
none
0.018 -cofactor 0.036+grpE 0.19+DnaJ 2.9+DnaJ+grpE
50 M
0.03 -cofactor 0.048+grpE 0.20+DnaJ 2.9+DnaJ+grpE
none
0.019+0.001 steady state
0.019±0.002
0.018±0.004 khyd
0.001±0.0003 11mM Mg (OAc) Kd 2 50 g/ml BSA pH 7.6, 25°C
Ligand (conc.)
100 mM KCl orNaCl 5 mM MgCl2 1 M DnaK 0.1mM ATP pH 7.0, 25°C
25 mM HEPES 200 mM potassium glutamate
(Russell et al., 1998)
kcat (min1)
K M( M)
Assay condition
Reference
nd
25 mM HEPES 50 mM KCl 11 mM MgCl2, 5 mM mercaptoethanol pH 7.0, 37°C
(Pierpaoli et al., 1997)
50 mM Tris-HCl
(Ziegelhoffer et al., 1995)
(1)
31 E. coli DnaK overproduced
none
0.084±0.006 khyd
32 S. cerevisiae
none
2.5 mM KCl
Ssa1P natural
637
0.028±0.003
4.2±1.0
various [KCl] 2 mM MgCl2
15 mM KCl
0.68 ± 0.006
5 mM mercaptoethaol
0.046±0.003
0.19 ± 0.07
pH 7.5, 33°C
200 mM KCl 0.045±0.007
Molecular chaperones and folding catalysts 33 S. cerevisiae Ssa1P natural
peptide A7:
50 mM Tris-HCl
no peptide 5 mM KCl
200 M
638 (Ziegelhoffer et al., 1995)
various [KCl]
0.01
2 mM MgCl2
50 mM KCl
5 mM mercaptoethaol
0.04
10
5 mM KCl
pH 7.5, 33°C
M ATP
0.01 50 mM KCl 0.1 (1)Sequence
of peptide A: KRQIYTDLEMNRLGK; a-pp peptide: CALLQSRLLLSAPRAAATARA; peptide C: KLIGVLSSLFRPK; peptide A7: RRLIEDAETAARG; S peptide: KETAAAKFERQHMDSSTSAA; cytochrome C peptide: IFAGIKKKAERADLIAYLKQATAK.
Table 2 Binding affinities of adenine nucleotides for Hsp70 proteins
Protein
Nucleotide K d (
Bovine Hsc70 natural
ATP ADP
Bovine Hsc70 natural Bovine Hsc70 natural
M)
Assay conditions
Method/comments Reference
0.7 1.35
40 mM HEPES 75 mM KCl 4.5 mM Mg (OAc)2 0.8 mM DTT, 5% BSA pH 7.0, 37°C
Nucleotide binding proceeded for 5'; then, free nucleotide removed by charcoal.
ATP ADP
9.5±3.9 1.6±0.5
20 mM Na phosphate 200 mM KCl pH 6.5, 4°C
Equilibrium dialysis. (Palleros et al., 1991)
ATP ADP AMPPNP dATP
0.012 0.018 2.8 23
20 mM Imidazole, 25 mM KCl 10 mM (NH4) 2SO4 2 mM Mg (OAc)2 1 mM DTT
Equilibrium dialysis: (Gao et al., AMPPNP and dATP 1994) measured directly; ATP pH 7.0, 25°C and ADP measured by competition with AMPPNP.
(Schmid et al., 1985)
Structure and mechanism of Hsp70 proteins
639
Bovine ATP Hsc70 ADP E543K recombinant
0.042±0.007 0.11±0.02
40 mM HEPES Equilibrium binding. (Ha et al., 75 mM KCl (by filter binding) 1994) 4.5 mM Mg (OAc)2 pH 7.0, 25°C
Bovine ATP Hsc70 E543K recombinant
Two-step binding; K1=1.5 M K2=0.03
40 mM HEPES 75 mM KCl 4.5 mM Mg (OAc)2 pH 7.0, 25°C
ADP
0.020±0.002
Kinetics of change in (Ha et al., tryptophan 1995) fluorescence when nucleotides bind.
Protein
Nucleotide K d (mM)
Assay conditions
Method/comments Referenc
Bovine Hsc70
ATP
20 mM HEPES
Kinetics of nucleotide
E543K recombinant
no peptide
0.011±0.005 150 mM KCl with peptide 3 mM Mg(OAc)2 0.022±0.001 pH 7.0, 25°C
(Takeda e al., 1996)
binding, (by flourescence).
ADP
no peptide (peptide=fluorescein 0.048±0.005 -FYQLALT) with peptide 0.099±0.014
Bovine Hsc70 44 kDa ATPase fragment
ADP
0.094±0.012 40 mM HEPES 75 mM KCl 4.5 mM Mg(OAc)2 pH 7.0, 25°C
Equilibrium binding (Ha et al. and kinetics of 1994) nucleotide binding.
Rat Hsc70 recombinant
ATP/ADP
0.2–0.3
25 mM Tris-HCl 120 mM NaCl 5 mM MgCl2 pH 7.0, room temp.
Binding of (Wang et nucleotide to al., 1993) protein immobilized on Sepharose.
Rat Hsc70 recombinant WT, D10N
ATP/ADP
0.3
40 mM HEPES 75 mM KCl 5 mM Mg (OAc)2 pH 7.0, 4°C
Binding of (Huang e nucleotide to al., 1993) protein immobilized on Sepharose.
1.1±0.1 1,400±700
20 mM HEPES Filter binding 25 mM KCl 10 mM (NH4)2SO4 1 mM Mg(OAc)2 pH 7.0, r. t.
Mung bean ATP site 1 Hsc70 ATP site 2 (pig brain was found to be indistinguishable)
Protein
Nucleotide K d (mM)
Assay conditions
(Buxbaum et al., 1996)
Method/comments Reference
Molecular chaperones and folding catalysts E. coli DnaK MABAoverproduced ADP*
640
0.09±0.01
50 mM Tris/HCl Kinetics of change 100 mM KCl in MABA* flourescence;
AMP
17.6±2.5
5 mM MgCl2
Competition with
ADP
0.13±0.03
2 mM EDTA
MABA-ADP* at
AMPPNP
1.6±0.1
2 mM dithioerythritol
equilibrium
ATPgS
0.044±0.002
pH 7.5, 25°C
GDP
90±14
ATP
0.007
(Theyssen et al., 1996)
Kinteics of change in tryptophan fluoresence
E. coli DnaK MABAoverproduced ADP*
20
0 mM Tris/HCl 100 mM KCl 5 mM MgCl2 2 mM EDTA 2 mM dithioerythritol 1:1 grpE pH 7.5, 25°C
E. coli DnaK ATP
0.001±0.0003 25 mM HEPES
overproduced ADP
0.0012±0.0003 200 mM potassium glutamate 11 mM Mg (OAc)2 5 mM PO4 50 g/ml BSA pH 7.6, 25°C
Kinteics of change in MABA* flourescence
(Packschies et al., 1997)
Filter binding
(Russell et al., 1998)
*N8-[4-[(N’-methyl(anthraniloyl)-amino]butyl]-8-aminoadenosine 5’-diphosphate
Protein
Nucleotide K d (mM)
Assay conditions
Method/comments Reference
E. coli DnaK
ATP
20 mM HEPES 100 mM NaCl 100 mM KCl ~10 nM Mg2+, controlled with 20 mM
Equilibrium dialysis.
0.5
(Skowyra et al., 1995)
Structure and mechanism of Hsp70 proteins
641
EDTA 0.1 mg/ml BSA 5% glycerol 5 mM DTT pH 7.6, 4°C Hamster BiP
ATP
0.2±0.1
20 mM HEPES
recombinant ADP
0.29±0.12
25 mM KCl
T229G
ATP
0.11±0.1
2 mM MgCl2
ADP
0.15±0.11
0.1 mM EDTA
ATP
0.12±0.11
0.5 mM DTT
ADP
0.16±0.12
pH 7.0, 4°C
T37G
Equilibrium dialysis.
(Wei et al., 1995b)
G227, 226D ATP, ADP not detectable
Table 3 Mutants in the isolated ATPase domain of Hsc70 derived from data in the literature(1)
Protein Mutation E. coli k cat ( min−1 ) DnaK equivalent residue Bovine w.t.
0.8±0.1
Rat*
0.28
w.t.
k cat KM( (mutant) M) k cat (w.t.) 0.7±0.1
KM Reference (mutant) KM (w.t.) (O’Brien et al., 1993) (Huang et al., 1993)
Bovine w.t.
0.53±0.07
0.49±0.26
(Wilbanks et al., 1994)
Bovine w.t. (25° C)
0.14±0.012
0.37±0.09
(Ha et al., 1994)
Rat*
D10N
D8
0
0
(Huang et al., 1993)
Bovine D10N
D8
0.021±0.008
0.04 10.1±4.3
20 (Wilbanks et al., 1994)
Bovine D10S
D8
0.009±0.003
0.02 5.9±2.3
12 (Wilbanks et al., 1994)
Molecular chaperones and folding catalysts
642
Bovine C17K
C15
0.02±0.0003*
0.037 0.04±0.01
0.07 (Wilbanks et al., 1998)
Bovine K71A
K70
<0.003
<0.01
(O’Brien et al., 1996)
Bovine K71E
K70
<0.003
<0.01
(O’Brien et al., 1996)
Bovine K71M
K70
<0.003
<0.01
(O’Brien et al., 1996)
Bovine E175Q
E171
0.066±0.022
0.12 46.5±17.6
95 (Wilbanks et al., 1994)
Bovine E175S
E171
0.014±0.010
0.03 16.2±3.1
33 (Wilbanks et al., 1994)
Bovine D199N
D194
0.011±0.002
0.02 0.73±0.49
1.5 (Wilbanks et al., 1994)
Bovine D199S
D194
0.008±0.11
0.02 0.45±0.19
0.82 (Wilbanks et al., 1994)
Bovine T204E
T199
1.9±0.3
2.4 93±14
133 (O’Brien et al., 1993)
Bovine T204V
T199
1.9±0.2
2.4 89±12
127 (O’Brien et al., 1993)
Bovine D206K
D201
0.0044±0.0004*
Bovine D206N
D201
Bovine D206S
D201
0.008 0.27±0.11
0.45 (Wilbank et al., 1998)
0.064±0.012
0.12 0.27±0.25
0.55 (Wilbanks et al., 1994)
0.076±0.030
0.14 0.27±0.25
0.55 (Wilbanks et al., 1994)
(1)Assays done in 40mM
HEPES, 75mM KCl, 4.5mM Mg (CH, C00)2, pH 7.0; or, (*) 50 mM MES, 5 mM MgSO4, 10 mM CaCl2, 1 mM DTT, pH 6.5. Assay temperature was 37°C except where noted. Bovine constructs include residues 1–386 ( =387–650); rat constructs include residues 1–384 ( =385–646). The amino acid sequences for bovine and rat Hsc70 are identical in the ATPase domain.
Structure and mechanism of Hsp70 proteins
643
Figure 4 Suggested pathway of ATP hydrolysis by Hsp70 proteins. Only a few amino acid side chains are included, for clarity. See text for discussion of details.
ion. Subsequent studies have shown that mutagenesis of Lys71 completely abolishes ATPase activity, lending plausibility to the proposed scenario (O’Brien et al., 1996). The imperative of the triphosphate moiety of ATP forming a specific , -bidentate complex with Mg2+ or similar divalent ion for hydrolysis rationalizes the requirement for Mg2+ in the ATPase reaction. Additionally, crystallographic studies have identified two specific monovalent ion binding sites in the nucleotide binding cleft of the ATPase fragment (Figure 5); these monovalent ions influence the exact geometry and electrostatic interactions of the protein-metal-nucleotide complex (Wilbanks et al., 1995b). Substitution of one monovalent ion with another of significantly
Molecular chaperones and folding catalysts
644
Figure 5 Drawing showing selected features of of the active site regions of the ATPase fragment, including MgADP, Pi, two K+ ions, neighboring amino acid residues and H2O molecules. Oxygen (except H2O), red; nitrogen, blue; carbon, gray; H2O molecules, green.
different size (e.g. Na+ for K+) has the potential to perturb the geometry and electrostatic interactions of the catalytic transition state and to affect the ATP hydrolysis reaction rate, leading to the effects that have been observed. 4. COUPLING OF THE ATPase ACTIVITY TO THE PEPTIDE BINDING ACTIVITY Several lines of evidence (proteolytic digestion patterns of E. coli DnaK (Buchberger et al., 1995; Kamath-Loeb et al., 1995; Liberek et al., 1991b) yeast Ssa1 (Fung, et al., 1996) and hamster BiP (Wei et al., 1995b); changes in intrinsic tryptophan fluorescence of DnaK (Banecki et al., 1996; Palleros et al., 1992; Theyssen et al., 1996; Buchberger et
Structure and mechanism of Hsp70 proteins
645
al., 1995) and bovine Hsc70 (Ha et al., 1995); changes in infrared spectra of DnaK (Banecki et al, 1992); solution small-angle X-ray scattering of DnaK (Shi et al., 1996) and bovine Hsc70 (Wilbanks et al., 1995a)) indicate that Hsp70 proteins access at least two significantly different conformations in a nucleotide-dependent manner. Additionally, this conformational change is observed in 60 kDa fragments lacking ~100 carboxy-terminal amino acids. These fragments, which terminate at or near the end of the first long helix of the peptide binding domain, have been isolated for both DnaK (from a genetic screen) (Buchberger et al., 1995) and bovine Hsc70 (by rational design) (Wilbanks et al., 1995a). The two predominant conformations appear to correspond to (i) a state which forms long-lived, stable Hsp70-peptide complexes, often referred to as the “peptide binding” state, and (ii) a state in which Hsp70-peptide complexes are relatively transient, or labile, often called the “peptide release” state. It should be emphasized that the ATPase and peptide binding cycles are not stoichiometrically coupled in the sense of one ATP molecule inducing the release of one peptide; rather, the nucleotide can be thought to regulate the kinetics of peptide binding and release (Schmid et al., 1994). It should also be emphasized that describing the Hsp70 proteins as “two state” systems is a first, initial approximation that will likely be supplanted at some time in the future by a more complex scheme, as may arise when Hsp70 proteins interact with accessory proteins (Wawrzynow et al., 1995). The question of how the ATPase cycle is coupled to peptide binding and release can then be split into two considerations: (i) how does nucleotide drive the conformational switch, and (ii) how does peptide stimulate ATP hydrolysis. Little is known about the second consideration. The first can be phrased, which steps of the ATPase cycle switch Hsp70 proteins from one conformation to the other, and what is the molecular switching mechanism? In early experiments it was established in several different contexts that addition of ATP to Hsp70-(poly) peptide complexes facilitated peptide dissociation (i.e. switched Hsp70 from its peptide binding state to its peptide release state), while nonhydrolyzable ATP analogs (AMPPNP; -S-ATP) failed to do so (Liberek et al., 1991b; Palleros et al., 1991; Schmid et al., 1985). These data were taken as evidence that ATP hydrolysis was required for peptide release, and by implication, for the conformational transition that induced release of peptides. However, it was subsequently shown that ATP binding prior to or (for a T199A mutant of DnaK) in absence of significant ATP hydrolysis, induces peptide release (Palleros et al., 1993; Theyssen et al., 1996; McCarty et al., 1995). Nonhydrolyzable ATP analogs fail to do so; further, the change is discriminatory with respect to monovalent ion: Hsp70 proteins will undergo the conformational change to the ATP-induced form in the presence of K+ ions, but not in the presence of Na+, although ATP binds tightly in the presence of either ion. Hence, ATP binding, and binding in a specific nucleotide-metal complex, rather than hydrolysis, triggers a conformational change. Once it was established at the phenomenological level that ATP binding induced a conformational change, possibly in a two-step process with the major change in conformation coming about as a second step predicated on initial ATP binding (see also Buchberger et al., this volume), the question of how this is accomplished at the molecular level arose. Mutagenesis studies have identified residues in the
Molecular chaperones and folding catalysts
646
Figure 6 Schematic drawing summarizing the effects of mutagenesis of residues in the nucleotide binding cleft of Hsp70 proteins. Refer to Tables 3 and 4 for details. “ATPase” refers to the ATPase fragment of bovine Hsc70; “Hsc70” refers to rat Hsc70; “DnaK” refers to E. coli DnaK; “BiP” refers to hamster BiP.
nucleotide binding cleft, as seen in the isolated ATPase fragment, whose alteration abolishes the ability of Hsp70 proteins to undergo an ATP-induced conformational change (Table 4; Figure 6). In hamster BiP, mutagenesis of Gly226 or Gly227 (corresponding to Gly201 and Gly202 respectively in bovine Hsc70) to aspartic acid results in a protein which has a substantially lower ATP hydrolysis rate and which does not undergo an ATP-induced transition, as monitored by proteolytic digestion patterns (Wei et al., 1995b). These results can be rationalized by suggesting that introduction of a large side chain at a position normally occupied by glycine results in a clash between the side chain and ATP when it binds. Mutagenesis of Thr37 in BiP (corresponding to Thr13 of Hsc70) to glycine also yields a protein that will not undergo the normal conformational change, suggesting that the side chain (possibly the hydroxyl) of this residue is essential for the coupling mechanism (Wei et al., 1995b). In contrast, some residues have been mutated without causing a loss of coupling; these include Thr199 (McCarty et al., 1991) and E171 (Schuster et al., 1995) in DnaK (corresponding to T204 and E175 in Figure 6). Combining these data with the observations showing that the coupling mechanism can be defeated by relatively subtle changes in the nucleotide (substitution of nonhydrolyzable
Structure and mechanism of Hsp70 proteins
647
Table 4 ATPase activity of missense mutants of Hsp70 proteins(1)
Protein
Mutation Bov. k cat (min1) (Origin) Hsc70 equivalent
Rat Hsc70 D10N (sd)
D10
k cat (mutant)
K M( M)
k cat (w.t.)
“extremely low”
“extremely low”
0.32 (Vmax)
2.2
KM Phenotype/ (mutant) vitro effects KM (w.t.)
0.3 (Kd)
0.3 (Kd)
Reduced abili autophosphor
E. coli DnaK
28–33 (sd)
Hamster BiP
T37G (sd)
T13
0.077±0.038 0.19
Hamster BiP
T37G (sd)
T13
not reported 0.022±0.049
No ATPdependent rel of IgG
Hamster BiP
T37G (sd)
T13
not reported 0.025±0.010
No autokinas activity
Hamster BiP
T37S (sd) T13
not reported 0.094±0.029
3.8% w.t. autokinase activity
Hamster BiP
175– 201 (sd)
Y159–175 not reported 0.002±0.005
Hamster BiP
E201G (sd)
E175
not reported 0.026±0.020
No ATPdependent rel of IgG
Hamster BiP
E201G (sd)
E175
not reported 0.025±0.008
65% w.t. autokinase activity
E. coli DnaK
E171K (sel)
E175
0.4
4.1
0.12±0.11 0.6 (Kd)
ATP-induced conformation change undectable by limited proteolysis. B Peptide C, no released+ATP ATPase not stimulated by peptide C
30 & 175 30 & 175 Dominant (Two Km does not supp A, replication observed)
Molecular chaperones and folding catalysts
648 % activity in F’ replication viable up to 3 ATPase activ 44-fold lower absence of acetate. ATPa activity insensitive to peptides.
Protein Mutation Bov. k cat (min k cat -1 ) (mutant) (Origin) Hsc70 equivalent k cat (w.t)
K M( M)
KM Phenotype/ (mutant) in vitro effects KM (w.t.)
R
E. coli DnaK
E171A (sd)
E175
extremely extremely low low
Dominant; defective ATP hydrolysis; ATP releases peptide.
(S al B et 19
E. coli DnaK
E171L (sd)
E175
0.09
2.6
48±10
12
Peptide C binding does not stimulate ATPase; ATP does not release peptide.
(B et 19
E. coli DnaK
E171K (sd)
E175
1.0
28
70±7
18
Peptide C binding does not stimulate ATPase; ATP does not release peptide.
(B et 19
E. coli DnaK
A174T (sel)
A178
0.014
0.46
1.5 & 23 1.5 & 23 Recessive ( (K (Two Km {does not L support observed) 19 replication} ≤1% activity in mini F’ replicati viable up to 40°C. ATPase stimulated by
Structure and mechanism of Hsp70 proteins
649 peptide.
Hamster G226D BiP (sd)
G201
0.13±0.01 0.31
97±14
65
ATP-induced (W conformational 19 change undetectable by limited proteolysis. Binds Peptide C, not released when ATP added.
Hamster G227D BiP (sd)
G202
0.11±0.01 0.28
93±17
62
ATP-induced (W conformational 19 change undetectable by limited proteolysis. Binds Peptide C, not released when ATP added.
Hamster T229G BiP (sd)
T204
not reported
0.037±0.004
No ATP(G 19 dependent release of IgG
k cat k cat Protein Mutation Bov. 1 (Origin) Hsc70 (min- ) (mutant) equivalent k cat (w.t.)
KM (mM)
Hamster T229G BiP (sd)
T204
none detected
0.11±0.10 0.55 (Kd reported)
Hamster T229G BiP (sd)
T204
not reported
0.061±0.001
KM Phenotype/ (mutant) in vitro effects KM (w.t.)
R
ATP-induced (W conformational a change detectable by limited, proteolysis. Binds Peptide C, and releases in an ATP dependent fashion No autokinase (G activity a
Molecular chaperones and folding catalysts
650
Hamster T229D BiP (sd)
T204
not reported
0.156±0.024
No autokinase (G activity a
Hamster T229S BiP (sd)
T204
not reported
0.217±0.026
11.6% w.t. autokinase activity
E. coli DnaK
T199A (sd)
T204
0.003– 0.007
0.03
No autokinase (M et activity 1
E. coli DnaK
T199D (sd)
T204
0.009– 0.011
0.07
No autokinase (M activity et 1
E. coli DnaK
T199V (sd)
T204
0.001
0.007
No autokinase (M activity et 1
(G a
Bovine 0 T204E, Hsc70 E543K (sd/ca)
1.5±0.3
97±30
Reference to (O Hsc70 et E543K mutant 1 below.
Bovine 0 T204V, Hsc70 E543K (sd/ca)
0.07±0.02
78±34
Reference to (O Hsc70 et E543K mutant 1 below.
Protein Mutation Bov. k cat (min 1) (Origin) Hsc70 equivalent
E. coli DnaK
D201N (sel)
D206
0.31
k cat KM KM Phenotype/ Refere (mutant) (mM) (mutant) in vitro effects k K cat
M
(w.t.)
(w.t.)
10
8
8 Dominant does not support A, replication ≤1% activity in mini F’ replicatic viable up to 34°C. ATPase insensitive to peptide.
(Kama Loeb et al., 1995)
Structure and mechanism of Hsp70 proteins E. coli G32D, Dnak756 G455D, G468D (sel)
G34D, G457D, G470D
12–18
Hamster K446M, Bip K447M (sd)
K423, Q424
“w.t. levels” “w.t. levels”
Bovine Hsc70
E543K (ca)
0.072±0.001 (T=25°C)
Bovine Hsc70
E543K (ca)
0.2±0.1
Bovine Hsc70
E543K (ca)
0.13±0.01 (T=25°C)
~90
temperature (Libere sensitive al., 19 does not support replication (Gaut al., 19
0.7±0.5
~4
651
Authors were not aware of mutation.
(Ha et 1994)
Authors were not aware of mutation.
(O’Bri et al., 1993)
khyd=0.22 (Ha et 1997) with 152 M peptide faf1 (ratio to w.t.=0.8)
(1) See Table 1
for assay conditions and comparable w.t. rates. Origin: sd=site-directed; sel=selecte genetic screen; ca=cloning artifact.
analogs for ATP) or monovalent ion (e.g. substitution of Na+ for K+), it can be suggested that MgATP must bind Hsp70 proteins in a complex with very precise geometric constraints in order to trigger a conformational change, and that small perturbations of this complex defeat the coupling mechanism. The specific nature of this complex is not yet well-defined, since structures of representative fragments that undergo the change (either a full-length Hsp70 or a 60 kDa fragment) are not available, and the isolated ATPase and peptide binding fragments each manifest only a single overall conformation. One possibility is that the coupling mechanism involves , -bidentate MgATP complex that is a reaction intermediate state on the pathway of ATP hydrolysis, similar to the proposed intermediate shown in Figure 4. Alternatively, it is possible that a conformation not on the reaction pathway is involved. In the presence of ADP, Hsp70 proteins form high affinity complexes with peptides. Hence, some step in the enzymatic ATPase cycle following ATP binding must be responsible for the conformational transition to the peptide binding state. Which step is responsible for this transition has not been resolved definitively; evidence has been presented that the transition is predicated on release of either Pi or ADP after ATP hydrolysis (Wilbanks et al., 1995a).
Molecular chaperones and folding catalysts
652
5. MODULATION OF Hsp70 ACTIVITY 5.1. Self-association It has been known for some time that Hsp70 proteins self-associate to form dimers and other low-order oligomers, and that these oligomers can be induced to dissociate to monomers by addition of ATP, but not by addition of nonhydrolyzable ATP analogs (Ladjimi et al., 1995; Schmid et al., 1985). Additionally, Hsp70 proteins form high molecular weight aggregates which cannot be dissociated by ATP; these appear to be inactive, dead-end aggregates (Palleros et al., 1991). The self-association parallels the interactions of Hsp70 proteins with target polypeptides and denatured proteins in many regards. It can be reversed by addition of short peptides (suggesting competition for the same peptide binding site) or by addition of ATP. In addition, stable complexes are formed in the presence of ADP. The region of Hsp70 proteins that is responsible for intramolecular association is not known with certainty; one candidate is the glycine-rich carboxy terminal 30–50 amino acid residues. Consistent with this suggestion is the observation that the 60 kDa fragments of Hsc70, which lack this carboxy-terminal region, are substantially less prone to intramolecular aggregation (Wilbanks et al., 1995a). The significance of the self-peptide binding activity of Hsp70 proteins for their overall function is not yet clear; it may be a mechanism of self-regulating polypeptide binding activity in vivo. 5.2. Posttranslational Modification Several posttranslational modifications have been documented for Hsp70 proteins. In vivo phosphorylation of serine and threonine has been seen in both prokaryotic DnaK (Rieul et al., 1987) and eukaryotic Hsp70’s (Freiden et al., 1992; Gaut et al., 1993; Loomis et al., 1982). The level of phosphorylation increases under conditions of cell stress or (for BiP) glucose starvation. In vitro Ca2+-dependent phosphorylation of Thr199 of DnaK (and the structurally equivalent residue of BiP) has been reported (Dalie et al., 1990; Leustek et al., 1992; McCarty et al., 1991; Panagiotidis et al., 1994). However, this targets a different residue from the in vivo phosphorylation and appears to be a distinct activity. ADP ribosylation of BiP has been documented; this modification has not been seen with other Hsp70 proteins (Carlsson et al., 1983; Freiden et al., 1992). In vivo, ADP ribosylated and phosphorylated BiP is localized in the oligomeric pools of the protein (Freiden et al., 1992). Methylation of lysine and arginine has been observed for eukaryotic Hsp70s, including both BiP and cytoplasmic proteins (Wang et al., 1992). Differences between natural bovine BiP and recombinant mouse BiP in their propensity to self-aggregate, as well as in their basal ATPase activities, have been reported; the native bovine protein oligomerizes readily and has a lower basal ATPase rate, while the recombinant mouse protein is predominantly monomeric in vitro and has a higher basal ATPase rate (Blond-Elguindi et al., 1993). Posttranslational modification of the natural protein has been suggested as the source of the differences in behavior between the two forms of the proteins in this case. Difference in the activities of natural and recombinant
Structure and mechanism of Hsp70 proteins
653
bovine Hsc70 observed in our lab were traced to a missense mutation introduced in the course of cloning (E543K (Ha, et al., 1997)). No changes in activity could be attributed to the tri-methylation of lysine561 which occurs in vivo. Although there appear to be intriguing correlations between different levels of posttranslational modification on the one hand, and different oligomeric states of the Hsp70 proteins or different cellular circumstances (stressed versus nonstressed) on the other hand, the specific functional consequences of the known modifications are not understood at this time. 5.3. Accessory Proteins Modulation of the in vitro activities of E. coli DnaK by DnaJ and GrpE is discussed in depth by Buchberger et al. in this volume. Here, we will only mention briefly that DnaJ stimulates nucleotide hydrolysis and GrpE accelerates nucleotide release (Liberek et al., 1991a). Acting in concert, DnaJ and GrpE accelerate the ATPase activity as much as 190fold over its basal rate (Liberek et al., 1991; McCarty et al., 1995). DnaJ has a modular structure, with a “J domain” which mimics some of the effects of the full-length protein localized in the amino-terminal residues 1–108 (Wall et al., 1994). Homologous J domains have been identified in both organellar and cytoplasmic proteins in eukaryotes. For example, auxilin, which has recently been shown to facilitate Hsc70dependent clathrin disassembly by recruiting Hsc70 to clathrin lattices, has a carboxyterminal J-domain (Ungewickell et al., 1995). In addition, the vesicle associated cysteine string protein has been shown to associate with and activate Hsc70 in vitro (Braun, et al., 1996, Chamberlain, et al., 1997). In contrast, the search for GrpE homologs in eukaryotes has so far revealed them only in mitochondria (Bolliger et al., 1994; Laloraya et al., 1994; Nakai et al., 1994). An alternative cytoplasmic candidate, selected for its ability to interact with an Hsc70 ATPase domain in a yeast two-hybrid screen, has been identified (Höhfeld et al., 1995). This protein, named HiP (for Hsc70 interacting protein), is unrelated in sequence to the prokaryotic GrpE proteins; the primary sequence of the 41 kDa protomer has a region of approximate GGMP repeats, similar to those found in the carboxy terminal region of eukaryotic Hsp70s, and three repeats of a 34-residue motif called the tetratricopeptide repeat (TPR), a motif whose function is unknown, but which has been seen in a variety of proteins. Although HiP appears to mimic GrpE in its ability to interact with the ATPase domain of Hsc70, it will not associate with full-length Hsc70 alone in vitro, but will bind to a preformed Hsp70-Hsp40 complex (the eukaryotic analog of an E. coli DnaK-DnaJ complex). HiP does not stimulate the ATPase activity of the Hsp70-Hsp40 complex but, in contrast to GrpE, stabilizes Hsp70 in the ADP state (Höhfeld et al., 1995). It has been suggested that the function of HiP may be to stabilize a relatively labile Hsp70-Hsp40 complex, thereby prolonging the time during which Hsc70 would interact in a stable complex with unfolded polypeptides; future studies can be expected to further clarify the participation of the HiP protein in the molecular chaperone machinery. 6. EPILOGUE The activities with which Hsp70 proteins were first associated—disassembly of a
Molecular chaperones and folding catalysts
654
bacteriophage replication initiation complex by DnaK (Liberek et al., 1988), disassembly of clathrin cages into triskelions by Hsc70 (Chappell et al., 1986; Schlossman et al., 1984), facilitation of transmembrane translocation by Hsc70 (Chirico et al., 1988; Deshaies et al., 1988), binding to immunoglobulin heavy chains (the consequence of which was nebulous) by BiP (Munro et al., 1986), renaturation of heatdenatured RNA polymerase by DnaK (Skowyra et al., 1990)—were so diverse that it seemed inevitable that the Hsp70 proteins would be complex, multifunctional entities. It now appears that these activites and others that have been subsequently identified for Hsp70 proteins can all be reconciled with a relatively simple, unifying molecular mechanism of (i) binding unstructured segments of polypeptides with a single peptide binding site, and (ii) regulating the kinetics of peptide binding and release through interactions with ATP/ADP at a single nucleotide binding site, accomplished by (iii) nucleotide-induced switching between (to a first approximation) two conformations, one of which forms stable, long-lived complexes with peptides, and the other of which binds peptides only transiently. The nucleotide appears to serve a regulatory role in a passive chaperone process rather than providing a source of free energy for an active protein folding process; it is probable that in vivo, the ATP: ADP ratio is an important modulator of Hsp70 activity. Since ATP binding (a rapid step in the ATPase cycle) switches Hsp70 proteins into a “peptide release” conformation, and since hydrolysis (a slow step) is a rate-limiting step in the ATPase cycle, it can be suggested that under cellular conditions of high ATP: ADP ratio, Hsp70s will exist predominantly in an ATP-induced, low peptide affinity state, and unstructured segments of polypeptides will be bound only transiently. On the other hand, under conditions of cell stress, which would lead to a high ADP: ATP ratio, Hsp70 proteins will switch to their tight binding conformation and “lock onto” unstructured segments of polypeptide. However, describing the mechanism of Hsp70 proteins with such an elementary model, by itself, fails to do justice to the diversity of functions ascribed to this protein family. Many questions still remain—what are the roles of the accessory proteins, particularly the DnaJ homologs, in modulating and/or targeting Hsp70 activities? What are the roles of posttranslational modifications? Why are several different Hsp70 proteins needed in the cytoplasm/nucleus of eukaryotic cells, each having its own characteristic pattern of regulation of expression—do they have subtle differences in function that we do not yet appreciate? We seem to have clarified the basic theme of activity of isolated Hsp70 proteins in vitro, and in so doing, to have poised ourselves for studies of the variations on the theme reflected in the proteins’ complex and multifaceted in vivo activities. 7. REFERENCES Banecki, B. and Zylicz, M. (1996). Real time kinetics of the DnaK-DnaJ-GrpE molecular chaperone machine action. J. Biol. Chem. , 271 , 6137–6143. Banecki, B., Zylicz, M., Bertoli, E. and Tanfani, F. (1992). Structural and functional relationships in DnaK and DnaK756 heat-shock proteins from Escherichia coli. J. Biol. Chem. , 267 , 25051–25058. Beckmann, R.P., Mizzen, L.A. and Welch, W.J. (1990). Interaction of Hsp 70 with newly synthesized proteins: implications for protein folding and assembly. Science , 248 ,
Structure and mechanism of Hsp70 proteins
655
850–854. Blond-Elguindi, S., Fourie, A.M., Sambrook, J.F. and Gething, M.J. (1993). Peptidedependent stimulation of the ATPase activity of the molecular chaperone BiP is the result of conversion of oligomers to active monomers. J. Biol. Chem. , 268 , 12730– 12735. Bolliger, L., Deloche, O., Glick, B.S., Georgopoulos, C., Jeno, P., Kronidou, N., Horst, M., Morishima, N. and Schatz, G. (1994). A mitochondrial homolog of bacterial GrpE interacts with mitochondria! hsp70 and is essential for viability. Embo J. , 13 , 1998– 2006. Braell, W.A., Schlossman, D.M., Schmid, S.L. and Rothman, J.E. (1984). Dissociation of clathrin coats coupled to the hydrolysis of ATP: role of an uncoating ATPase. J. Cell Biol , 99 , 734–741. Braun, J., Wilbanks, S.M. and Scheller, R.H. (1996). The cysteine string secretory vesicle protein activates Hsc70 ATPase. J. Biol. Chem. , 271 , 25989–25993. Buchberger, A., Schröder, H., Buttner, M., Valencia, A. and Bukau, B. (1994a). A Conserved Loop in the ATPase Domain of the DnaK Chaperone is Essential for Stable Binding of GrpE. Stuct. Biol. , 1 , 95–101. Buchberger, A., Theyssen, H., Schroder, H., McCarty, J.S., Virgallita, G., Milkereit, P., Reinstein, J. and Bukau, B. (1995). Nucleotide-induced conformational changes in the ATPase and substrate binding domains of the DnaK chaperone provide evidence for interdomain communication. J. Biol. Chem. , 270 , 16903–16910. Buchberger, A., Valencia, A., McMacken, R., Sander, C. and Bukau, B. (1994b). The chaperone function of DnaK requires the coupling of ATPase activity with substrate binding through residue E1 71. Embo J. , 13 , 1687–1695. Buxbaum, E. and Woodman, P.G. (1996). Binding of ATP and ATP analogues to the uncoating ATPase Hsc70 (70 kDa heat-shock cognate protein). Biochem. J. , 923–929. Carlsson, L. and Lazarides, E. (1983). ADP-Ribosylation of the Mr 83,000 stressinducible and glucose-regulated protein in avian and mammalian cells: modulation by heat shock and glucose starvation. Proc. Natl. Acad. Sci. USA , 80 , 4664–4668. Chamberlain, L.H. and Burgoyne, R.D. (1997). Activation of the ATPase activity of heatshock proteins Hsc70/Hsp70 by cysteine-string protein. Biochem J , 322 , 853–858. Chappell, T.G., Welch, W.J., Schlossman, D.M., Palter, K.B., Schlesinger, M.J. and Rothman, J.E. (1986). Uncoating ATPase is a member of the 70 kilodalton family of stress proteins. Cell , 45 , 3–13. Chirico, W.J., Waters, M.G. and Blobel, G. (1988). 70K Heat shock related proteins stimulate protein translocation into microsomes. Nature , 332 , 805–810. Dalie, B.L., Skaleris, D.A., Kohle, K., Weissbach, H. and Brot, N. (1990). Interaction of DnaK with ATP: binding, hydrolysis and Ca+2-stimulated autophosphorylation. Biochem. Biophys. Res. Comm. , 166 , 1284–1292. DeLuca-Flaherty, C., McKay, D.B., Parham, P. and Hill, B.L. (1990). Uncoating protein (hsc70) binds a conformationally labile domain of clathrin light chain LCa to stimulate ATP hydrolysis. Cell , 62 , 875–887. Deshaies, T.J., Koch, B.D., Werner-Washburne, M., Craig, E.A. and Schekman, R. (1988). A subfamily of stress proteins facilitates translocation of secretory and mitochondrial precursor polypeptides. Nature , 332 , 800–805.
Molecular chaperones and folding catalysts
656
Feifel, B., Sanmeier, E., Schonfeld, H.J. and Christen, P. (1996). Potassium ions and the molecular-chaperone activity of DnaK. Eur. J. Biochem. , 237 , 318–321. Flaherty, K.M., DeLuca-Flaherty, C. and McKay, D.B. (1990). Three-dimensional structure of the ATPase fragment of a 70 kilodalton heat shock cognate protein. Nature , 346 , 623–628. Flaherty, K.M., McKay, D.B., Kabsch, W. and Holmes, K.C. (1991). Similarity of the three-dimensional structures of actin and the ATPase fragment of a 70-kDa heat shock cognate protein. Proc. Natl. Acad. Sci. USA , 88 , 5041–5045. Flaherty, K.M., Wilbanks, S.M., DeLuca-Flaherty, C. and McKay, D.B. (1994). Structural basis of the Hsc70 ATP hydrolytic activity. II. Structure of the active site with ADP or ATP bound to wild-type and mutant ATPase fragment. J. Biol. Chem. , 269 , 12899–12907. Flajnik, M.F., Canel, C., Kramer, J. and Kasahara, M. (1991). Which came first, MHC class I or class H? Immunogenetics , 33 , 295–300. Fletterick, R.J., Bates, D.J. and Steitz, T.A. (1975). The structure of a yeast hexokinase monomer and its complexes with substrates at 2.7 Å resolution. Proc. Natl. Acad. Sci. USA , 72 , 38–42. Flynn, G.C., Chappell, T.G. and Rothman, J.E. (1989). Peptide binding and release by proteins implicated as catalysts of protein assembly. Science , 245 , 385–390. Flynn, G.C., Pohl, J., Flocco, M.T. and Rothman, J.E. (1991). Peptide-binding specificity of the molecular chaperone BiP. Nature , 353 , 726–730. Freeman, B.C., Myers, M.P., Schumacher, R. and Morimoto, R.I. (1995). Identification of a regulatory motif in Hsp70 that affects ATPase activity, substrate binding and interaction with HDJ-1. Embo J. , 14 , 2281–2292. Freiden, P.J., Gaut, J.R. and Hendershot, L.M. (1992). Interconversion of three differentially modified and assembled forms of BiP. EMBO J. , 11 , 63–70. Fung, K.L., Hilgenberg, L., Wang, N.M. and Chirico, W.J. (1996). Conformations of the nucleotide and polypeptide binding domains of a cytosolic Hsp70 molecular chaperone are coupled. J. Biol. Chem. , 271 , 21559–21565. Gao, B., Emoto, Y., Greene, L. and Eisenberg, E. (1993). Nucleotide binding properties of bovine brain uncoating ATPase. J. Biol. Chem. , 268 , 8507–8513. Gao, B., Greene, L. and Eisenberg, E. (1994). Characterization of nucleotide-free uncoating ATPase and its binding to ATP, ADP, and ATP analogues. Biochemistry , 33 , 2048–2054. Gaut, J.R. and Hendershot, L.M. (1993a). Mutations within the nucleotide binding site of immunoglobulin-binding protein inhibit ATPase activity and interfere with release of immunoglobulin heavy chain. J. Biol. Chem. , 268 , 7248–7255. Gaut, J.R. and Hendershot, L.M. (1993b). The immunoglobulin-binding protein in vitro autophosphorylation site maps to a threonine within the ATP binding cleft but is not a detectable site of in vivo phosphorylation. J. Biol. Chem. , 268 , 12691–12698. Greene, L.E., Zinner, R., Naficy, S. and Eisenberg, E. (1995). Effect of nucleotide on the binding of peptides to 70-kDa heat shock protein. J. Biol. Chem. , 270 , 2967–2973. Ha, J.-H., Hellman, U., Johnson, E.R., Li, L., McKay, D.B., Sousa, M.C., Takeda, S., Wernstedt, C. and Wilbanks, S.M. (1997). Destabilization of peptide binding and interdomain communication by an E543K mutation in the seventy kilodalton bovine heat
Structure and mechanism of Hsp70 proteins
657
shock cognate protein, a molecular chaperone. J. Biol. Chem. , 272 , 27796–27803. Ha, J.-H. and McKay, D.B. (1994). ATPase kinetics of recombinant bovine 70 kilodalton heat shock cognate protein and its amino-terminal ATPase domain. Biochemistry , 33 , 14625–14635. Ha, J.-H. and McKay, D.B. (1995). Kinetics of nucleotide-induced changes in the tryptophan fluorescence of the molecular chaperone Hsc70 and its subfragments suggest the ATP-induced conformational change follows initial ATP binding. Biochemistry , 34 , 11635–11644. Harrison, C.J., Hayer-Hartl, M., Di Liberto, M., Hartl, F. and Kuriyan.J. (1997). Crystal structure of the nucleotide exchange factor GrpE bound to the ATPase domain of the molecular chaperone DnaK. Science , 276 , 431–435. Hartl, F.U. (1996). Molecular chaperones in cellular protein folding. Nature , 381 , 571– 580. Höhfeld, J., Minami, Y. and Hartl, F.U. (1995). Hip, a novel cochaperone involved in the eukaryotic Hsc70/Hsp40 reaction cycle. Cell , 83 , 589–598. Huang, S., Tsai, M.-Y., Tzou, Y.-M., Wu, W. and Wang, C. (1993). Aspartyl residue 10 is essential for ATPase activity of rat hsc70. J. Biol Chem. , 268 , 2063–2068. Hurley, J.H., Faber, H.R., Worthylake, D., Meadow, N.D., Roseman, S., Pettigrew, D.W. and Remington, S.J. (1993). Structure of the regulatory complex of Escherichia coli IIIGlc with glycerol kinase. Science , 259 , 673–677. Kabsch, W., Mannherz, H.G., Suck, D., Pai, E.F. and Holmes, K.C. (1990). Atomic structure of the actin: DNase I complex. Nature , 347 , 37–44. Kabsch, W. and Sander, C. (1983). Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features. Biopolymers , 22 , 2577– 637. Kamath-Loeb, A.S., Lu, C.Z., Suh, W.C., Lonetto, M.A. and Gross, C.A. (1995). Analysis of three DnaK mutant proteins suggests that progression through the ATPase cycle requires conformational changes. J. Biol Chem. , 270 , 30051–30059. Kassenbrock, C.K. and Kelly, R.B. (1989). Interaction of heavy chain binding protein (BiP/GRP78). with adenine nucleotides. EMBO J. , 8 , 1461–1467. Kraulis, P. (1991). MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures. J. Appl. Cryst. , 24 , 946–950. Ladjimi, M.M., Benaroudj, N., Batelier, G. and Triniolles, F. (1995). Self-association of the molecular chaperone HSC70 as assessed by analytical ultracentrifugation. Prog. Colloid Polym . Sci., 99 , 1–6. Laloraya, S., Gambill, B.D. and Craig, E.A. (1994). A role for a eukaryotic GrpE-related protein, Mge1p, in protein translocation. Proc Natl Acad Sci USA , 91 , 6481–5. Leustek, T., Amir, S.D., Toledo, H., Brot, N. and Weissbach, H. (1992). Autophosphorylation of 70 kDa heat shock proteins. Cell. Mol. Biol. , 38 , 1–10. Liberek, K., Georgopoulos, C. and Zylicz, M. (1988). Role of the Escherichia coli DnaK and DnaJ heat shock proteins in the initiation of bacteriophage DNA replication. Proc. Natl. Acad. Sci. USA , 85 , 6632–6636. Liberek, K., Marszalek, J., Ang, D., Georgopoulos, C. and Zylicz, M. (1991a). Escherichia coli DnaJ and GrpE heat shock proteins jointly stimulate ATPase activity of DnaK. Proc. Natl. Acad. Sci. USA , 88 , 2874–2878.
Molecular chaperones and folding catalysts
658
Liberek, K., Skowyra, D., Zylicz, M., Johnson, C. and Georgopoulos, C. (1991b). The Escherichia coli DnaK chaperone, the 70-kDa heat shock protein eukaryotic equivalent, changes conformation upon ATP hydrolysis, thus triggering its dissociation from a bound target protein. J. Biol. Chem. , 266 , 14491–14496. Loomis, W., Wheeler, S. and Schmidt, J. (1982). Phosphorylation of the major heat shock protein of Dictyostelium discoideum. Mol. Cell. Biol. , 2 , 484–489. McCarty, J.S., Buchberger, A., Reinstein, J. and Bukau, B. (1995). The role of ATP in the functional cycle of the DnaK chaperone system. J. Mol. Biol. , 249 , 126–137. McCarty, J.S. and Walker, G.C. (1991). DnaK as a thermometer: threonine-199 is site of autophosphorylation and is critical for ATPase activity. Proc. Natl. Acad. Sci. USA , 88 , 9513–9517. Morshauser, R.C., Wang, H., Flynn, G.C. and Zuiderweg, E.R.P. (1995). The peptidebinding domain of the chaperone protein Hsc70 has an unusual secondary structure topology. Biochemistry , 34 , 6261–6266. Munro, S. and Pelham, H.R. (1986). An Hsp70-like protein in the ER: identity with the 78 kd glucose-regulated protein and immunoglobulin heavy chain binding protein. Cell , 46 , 291–300. Nakai, M., Kato, Y., Ikeda, E., Toh, E.A. and Endo, T. (1994). Yge1p, a eukaryotic GrpE homolog, is localized in the mitochondrial matrix and interacts with mitochondrial Hsp70. Biochem. Biophys. Res. Comm. , 200 , 435–442. O’Brien, M.C., Flaherty, K.M. and McKay, D.B. (1996). Lysine 71 of the chaperone protein Hsc70 is essential for ATP hydrolysis. J. Biol. Chem. , 271 , 15874–15878. O’Brien, M.C. and McKay, D.B. (1993). Threonine 204 of the chaperone protein hsc70 influences the structure of the active site but is not essential for ATP hydrolysis. J. Biol. Chem. , 268 , 24323–24329. O’Brien, M.C. and McKay, D.B. (1995). How potassium affects the activity of the molecular chaperone Hsc70: I. Potassium is required for optimal ATPase activity. J. Biol. Chem. , 270 , 2247–2250. Packschies, L., Theyssen, H., Buchberger, A., Bukau, B., Goody, R.S. and Reinstein, J. (1997). GrpE accelerates nucleotide exchange of the molecular chaperone DnaK with an associative displacement mechanism. Biochemistry , 36 , 3417–3422. Palleros, D.R., Reid, K.L., McCarty, J.S., Walker, G.C. and Fink, A.L. (1992). DnaK, hsp73, and their molten globules. Two different ways heat shock proteins respond to heat. J. Biol. Chem. , 267 , 5279–5285. Palleros, D.R., Reid, K.L., Shi, L., Welch, W.J. and Fink, A.L. (1993). ATP-induced protein-Hsp70 complex dissociation requires K+ but not ATP hydrolysis. Nature , 365 , 664–666. Palleros, D.R., Welch, W.J. and Fink, A.L. (1991). Interaction of hsp70 with unfolded proteins: effects of temperature and nucleotides on the kinetics of binding. Proc. Natl. Acad. Sci. USA , 88 , 5719–5723. Panagiotidis, C.A., Burkholder, W.F., Gaitanaris, G.A., Gragerov, A., Gottesman, M.E. and Silverstein, S.J. (1994). Inhibition of DnaK autophosphorylation by heat shock proteins and polypeptide substrates. J Biol Chem , 269 , 16643–7. Pierpaoli, E.V., Sandmeier, E., Baici, A., Schonfeld, H.J., Gisler, S. and Christen, P. (1997). The power stroke of the DnaK/DnaJ/GrpE molecular chaperone system. J. Mol
Structure and mechanism of Hsp70 proteins
659
Biol , 269 , 757–768. Rieul, C., Cortay, J.C., Bleicher, F. and Cozzone, A.J. (1987). Effect of bacteriophage M13 infection on phosphorylation of dnaK protein and other Escherichia coli proteins. Eur. J. Biochem. , 168 , 621–627. Rippmann, F., Taylor, W.R., Rothbard, J.B. and Green, N.M. (1991). A hypothetical model for the peptide binding domain of hsp70 based on the peptide binding domain of HLA. Embo J. , 10 , 1053–1059. Rüdiger, S., Germeroth, L., Schneider-Mergener, J. and Bukau, B. (1997a). Substrate specificity of the DnaK chaperone determined by screening cellulose-bound peptide libraries. Embo J . 16 , 1501–1507. Rüdiger, S., Buchberger, A. and Bukau, B. (1997b). Principles governing the substrate specificity of Hsp70 chaperones. Nature Struct. Biol . 4 , 342–349. Russell, R., Jordan, R. and McMacken, R. (1998). Kinetic characterization of the ATPase cycle of the DnaK molecular chaperone. Biochemistry , 37 , 596–607. Sadis, S. and Hightower, L.E. (1992). Unfolded proteins stimulate molecular chaperone Hsc70 ATPase by accelerating ADP/ATP exchange. Biochem. , 31 , 9406–9412. Schatz, G. and Dobberstein, B. (1996). Common principles of protein translocation across membranes. Science , 271 , 1519–1526. Scherer, P.E., Krieg, U.C., Hwang, S.T., Vestweber, D. and Schatz, G. (1990). A precursor protein partly translocated into yeast mitochondria is bound to a 70 kd mitochondrial stress protein. EMBO J. , 9 , 4315–4322. Schlossman, D.M., Schmid, S.L., Braell, W.A. and Rothman, J.E. (1984). An enzyme that removes clathrin coats: purification of an uncoating ATPase. J. Cell Biol. , 99 , 723–733. Schmid, D., Baici, A., Gehring, H. and Christen, P. (1994). Kinetics of molecular chaperone action. Science , 263 , 971–973. Schmid, S.L., Braell, W.A. and Rothman, J.E. (1985). ATP catalyzes the sequestration of clathrin during enzymatic uncoating. J. Biol. Chem. , 260 , 10057–10062. Schmid, S.L. and Rothman, J.E. (1985). Enzymatic dissociation of clathrin cages in a two-stage process. J. Biol. Chem. , 260 , 10044–10049. Schuster, H.-P., Theyssen, H., Buchberger, A., Bukau, B., Reinstein, J. (1995). The mutation E171A of the molevular chaperone DnaK is deficient in ATP hydrolysis but still couples the binding of ATP to peptide release. Conference on Biology of Molecular Chaperones, Aghia Pelaghia, Greece , Poster abstract, A90. Shi, L., Kataoka, M. and Fink, A.L. (1996). Conformational characterization of DnaK and its complexes by small-angle X-ray scattering. Biochemistry , 35 , 3297–3308. Skowyra, D., Georgopoulos, C. and Zylicz, M. (1990). The E. coli dnaK gene product, the Hsp70 homolog, can reactivate heat-inactivated RNA polymerase in an ATP hydrolysis-dependent manner. Cell , 62 , 939–944. Skowyra, D. and Wickner, S. (1995). GrpE alters the affinity of DnaK for ATP and Mg2+. Implications for the mechanism of nucleotide exchange. J. Biol. Chem. , 270 , 26282–26285. Sriram, M., Osipiuk, J., Freeman, B., Morimoto, R. and Joachimiak, A. (1997). Human Hsp70 molecular chaperone binds two calcium ions within the ATPase domain. Structure , 5 , 403–414.
Molecular chaperones and folding catalysts
660
Takeda, S. and McKay, D.B. (1996). Kinetics of peptide binding to the bovine 70 kDa heat shock cognate protein, a molecular chaperone. Biochemistry , 35 , 4636–4644. Theyssen, H., Schuster, H.P., Packschies, L., Bukau, B. and ReinsteinJ. (1996). The second step of ATP binding to DnaK induces peptide release. J. Mol. Biol , 263 , 657– 670. Theyssen, H., Schuster, H.-P., Packschies, L., Bukau, B. and Reinstein, J. (1996). The Second Step of ATP Binding to DnaK Induces Peptide Release. J. Mol. Biol. , 263 , 657–670. Ungewickell, E., Ungewickell, H., Holstein, S.E., Lindner, R., Prasad, K., Barouch, W., Martin, B., Greene, L.E. and Eisenberg, E. (1995). Role of auxilin in uncoating clathrin-coated vesicles. Nature , 378 , 632–635. Wall, D., Zylicz, M. and Georgopoulos, C. (1994). The NH2-terminal 108 amino acids of the Escherichia coli DnaJ protein stimulate the ATPase activity of DnaK and are sufficient for replication. J. Biol. Chem. , 269 , 5446–5451. Wang, C. and Lee, M.R. (1993). High-level expression of soluble rat hsc70 in Escherichia coli: purification and characterization of the cloned enzyme. Biochem J , 69–77. Wang, C., Lin, J.M. and Lazarides, E. (1992). Methylations of 70,000-Da heat shock proteins in 3T3 cells: alterations by arsenite treatment, by different stages of growth and by virus transformation. Arch Biochem Biophys , 297 , 169–175. Wawrzynów, A. and Zylicz, M. (1995). Divergent effects of ATP on the binding of the DnaK and DnaJ chaperones to each other, or to their various native and denatured protein substrates. J. Biol. Chem. , 270 , 19300–19306. Wei, J., Gaut, J.R. and Hendershot, L.M. (1995). In vitro dissociation of BiP-peptide complexes requires a conformational change in BiP after ATP binding but does not require ATP hydrolysis. J. Biol. Chem. , 270 , 26677–26682. Wei, J. and Hendershot, L.M. (1995a). Characterization of the nucleotide binding properties and ATPase activity of recombinant hamster BiP purified from bacteria. J. Biol. Chem. , 270 , 26670–26676. Wei, J., Gaut, J.R. and Hendershot, L.M. (1995b). In vitro dissociation of BiP-peptide complexes requires a conformational change in BiP after ATP binding but does not require ATP hydrolysis. J. Biol. Chem. , 270 , 26677–26682. Wilbanks, S.M. and B, M.D. (1998). Structural replacement of active site monovalent cations by the ε-amino group of lysine in the ATPase fragment of bovine. Biochemistry , in press. Wilbanks, S.M., Chen, L., Tsuruta, H., Hodgson, K.O. and McKay, D.B. (1995a). Solution small-angle X-ray scattering study of the molecular chaperone Hsc70 and its subfragments. Biochemistry , 34 , 12095–12106. Wilbanks, S.M., DeLuca-Flaherty, C. and McKay, D.B. (1994). Structural basis of the Hsc70 ATP hydrolytic activity. I. Kinetic analyses of active site mutants. J. Biol Chem. , 269 , 12893–12898. Wilbanks, S.M. and McKay, D.B. (1995b). How potassium affects the activity of the molecular chaperone Hsc70: II. Potassium binds specifically in the ATPase active site. J. Biol Chem. , 270 , 2251–2257. Zhu, X., Zhao, X., Burkholder, W.F., Gragerov, A., Ogata, C.M., Gottesman, M.E. and
Structure and mechanism of Hsp70 proteins
661
Hendrickson, W.A. (1996). Structural analysis of substrate binding by the molecular chaperone DnaK. Science , 272 , 1606–1614. Ziegelhoffer, T., Lopez, B.P. and Craig, E.A. (1995). The dissociation of ATP from hsp70 of Saccharomyces cerevisiae is stimulated by both Ydj1p and peptide substrates. J Biol Chem , 270 , 10412–9. Zylicz, M., LeBowitz, J.H., McMacken, R. and Georgopoulos, C. (1983). The DnaK protein of Escherichia coli possesses an ATPase and autophosphorylating activity and is essential in an in vitro DNA replication system. Proc. Natl. Acad. Sci. USA , 80 , 6431–6435.
26. THE DnaK CHAPERONE SYSTEM: MECHANISM AND COMPARISON WITH OTHER HSP70 SYSTEMS ALEXANDER BUCHBERGER1, JOCHEN REINSTEIN2 and BERND BUKAU3, * 1 Centre
for Protein Engineering, Medical Research Council Centre, Hills Road, Cambridge CB2 2QH, UK 2 Abteilung Physikalische Biochemie, Max Planck-Institut für Molekulare Physiologie, Rheinlanddamm 201, D-44139 Dortmund, Germany 3 Institut für Biochemie und Molekularbiologie, Universität Freiburg, HermannHerder-Str. 7, D-79104 Freiburg, Germany
1. Introduction 2. Components of the DnaK Chaperone System 2.1. DnaK 2.1.1. Structure 2.1.2. Substrate Binding and its Coupling to the Nucleotide Status of DnaK 2.2. DnaJ 2.2.1. Structure 2.2.2. Structural Basis for the Interaction with DnaK 2.2.3. Substrate Binding 2.3. GrpE 3. The ATPase Cycle of DnaK 3.1. Binding of ATP 3.2. Switch 1: ATP Hydrolysis and the Role of DnaJ 3.3. Switch 2: Product Release and the Role of GrpE 3.4. Balance of Rate Constants 4. The Chaperone Cycle 5. Other Hsp70 Systems 6. Perspectives 7. Acknowledgements 8. References * Corresponding author
Molecular chaperones and folding catalysts
664
1. INTRODUCTION The ubiquitous Hsp70 chaperones and their cofactors are central components of the cellular system for folding, repair and degradation of proteins. They cooperate with other chaperones to assist protein folding reactions in a large variety of metabolic processes in most cellular compartments. In this chaperone network, Hsp70 proteins have unique and essential functions which rely on their ability to associate with short hydrophobic segments of polypeptides in an ATP-dependent manner (Haiti, 1996; Rüdiger et al., 1997a). Such association can prevent aggregation of substrates, by shielding exposed hydrophobic segments of polypeptides, and assist refolding, probably by decreasing the concentration of aggregation-prone folding intermediates. The elucidation of the mechanism of Hsp70 activity is an important aim of current research in the chaperone field. Substantial progress has been made particularly with the prokaryotic homolog, the DnaK heat shock protein of E. coli, and its cochaperones DnaJ and GrpE. The DnaK system assists protein folding processes including (i) prevention of aggregation and assistance of refolding of denatured proteins (Gaitanaris et al., 1990; Skowyra et al., 1990; Langer et al., 1992; Schröder et al., 1993; Ziemienowicz et al., 1993; Szabo et al., 1994), (ii) disassembly of protein oligomers (Echols, 1986; Alfano and McMacken, 1989; Zylicz et al., 1989), (iii) degradation of unstable proteins (Straus et al., 1988; Sherman and Goldberg, 1992), (iv) translocation of secretory proteins (Wild et al., 1992; Wild et al., 1996), and (v) control of activity of folded proteins (Wickner, 1990; Liberek and Georgopoulos, 1993; Gamer et al., 1996). The interaction of DnaK with substrates is tightly regulated by the status of the bound nucleotide. DnaJ and GrpE control the checkpoints of the ATPase cycle of DnaK by stimulating ATP hydrolysis and nucleotide exchange, respectively. DnaJ in addition associates with substrates and thereby acts as a targeting factor for DnaK This review focusses on mechanistic aspects of the DnaK chaperone system. We will summarize first the structural and functional features of the individual components, and then the functional cycle of the system. Finally we will discuss the mechanistic differences and similarities of the DnaK system with respect to other Hsp70 systems. The cellular roles of the DnaK system are described in more detail in other chapters of this volume (Connolly et al.; Burkholder and Gottesman; Welch et al.; Zylicz et al.; Maurizi et al.). 2. COMPONENTS OF THE DnaK CHAPERONE SYSTEM 2.1. DnaK 2.1.1. Structure DnaK is a 69 kDa protein that tends to form dimers and higher order oligomers (Schönfeld et al., 1995a). ATP and peptide substrates dissociate DnaK oligomers
The DnaK chaperone system
665
suggesting that the monomer of DnaK is the active species (Shi et al., 1996). DnaK shares approx. 50% amino acid identity with eukaryotic Hsp70 proteins (Bardwell and Craig, 1984). Conservation is highest for the N-terminal 540 residues of DnaK while the C-terminal 98 residues are less well conserved. The overall structure of DnaK consists of a N-terminal ATPase domain of 44 kDa and a C-terminal segment of 25 kDa which is further divided into a substrate binding domain of approx. 15 kDa (Wang et al., 1993) and a C-terminal domain of 10 kDa with unclear function (Figure 1). Attempts to solve the crystal structure of DnaK were successful for the ATPase domain in complex with a truncated form of GrpE (Harrison et al., 1997)
Figure 1 Domain organisation of DnaK, DnaJ and GrpE. Individual domains/modules are represented by differentially shaded boxes. The residues defining the approximate domain borders, known structural features and functions of domains are indicated. The definition of domains is based on 3D-structures where known and sequence alignments using standard algorithms. DnaK: residues 386 to 392 constitute a linker between the ATPase and the substrate binding domain. GrpE: residues 86 to 88 constitute a break of the N-terminal long -helix in the GrpE monomer that interacts with DnaK.
(see section 2.3 and Figure 6), and for an internal fragment of DnaK with substrate binding activity (V389-A607) in complex with a peptide substrate (Zhu et al., 1996) (Figure 2). The structure of the ATPase domain of DnaK is almost identical to that of bovine Hsc70 (Flaherty et al., 1990; Harrison et al., 1997; see Ha et al. in this volume) and, as first realized for Hsc70, shares high similarity with actin and hexokinase (Kabsch et al., 1990; Flaherty et al., 1991; Holmes et al., 1993). It consists of two subdomains that are separated by a deep central cleft. Two crossed -helices connect the subdomains and divide the cleft into an upper cleft, at the bottom of which nucleotide and Mg2+ are bound, and a lower cleft. The structural basis for the ATPase activity is discussed in
Molecular chaperones and folding catalysts
666
detail by Ha et al. in this volume. The adjacent substrate binding domain is a rather flat structure that also consists of two subdomains (Zhu et al., 1996) (Figure 2). The structure of the N-terminal domain is conserved within the Hsp70 protein family. The N-terminal subdomain (D393—K502) forms a -sandwich composed of 8 antiparallel -strands and two helices (A and B) that are packed onto the -sandwich. The 4 upper strands of the sandwich form an unusual sheet with two strands turned up by 20° relative to the neighbouring strands and four connecting loops which protrude upwards from the -sandwich. Helix B interacts with the four protruding loops thereby stabilizing
Figure 2 Crystal structure of the substrate binding domain of DnaK in complex with a peptide substrate. Ribbon diagram of the substrate binding domain in standard view (A) and a view rotated by 90° counterclockwise around the vertical axis of the standard view (B). The peptide substrate is shown in blue, the strands( 1, 2, 4, 5). and loops (L1,2, L3,4, L4,5, L5,6) of DnaK constituting the upper -
The DnaK chaperone system
667
sheet in green, the hinge-forming residues Arg536 to Gln538 of helix B, which may allow ATP-dependent opening of the helical lid, in orange. Helices, strands and loops are labeled, large N and C indicate the NH2- and COOH-termini of the DnaK fragment (A) and the substrate peptide (B). (C) The buried nature of the bound peptide substrate is shown in the space-filling representation of the standard view. The peptide is shown in blue, DnaK in green, side chains of residues which contribute to the interaction of helix B with the loops of the upper -sheet are pink. (D), (E) and (F) Details of the peptide binding cavity of DnaK. The cavity is shown from standard view (D) and after vertical rotation (E). The side chains of the peptide (blue) are shown as yellow sticks. (F) The central position (defined as position “0”). of the substrate binding pocket viewed from the bottom. Peptide side chains and backbone are shown as yellow and blue space filling representation, the side chains of DnaK forming pocket 0 as red space filling representation (radius corresponding to van der Waals contact surfaces). All illustrations were generated with INSIGHT II, Biosym. (Modified from Rüdiger et al., 1997a).
them and contributing to substrate binding (see below). Helix B in addition constitutes a linking region to the helices C to E which form the separate C-terminal subdomain (Q538-A607) of unknown function. The peptide substrate is bound in a hydrophobic cavity formed by two of the protruding loops and the upper part of the -sandwich (Figure 2). DnaK interacts with five consecutive residues of the bound peptide. The strongest interaction occurs at a central position within the binding site which forms a deep hydrophobic pocket tailored to bind Leu (Figure 2). Helix B constitutes a lid that closes the substrate binding cavity but does not contact directly the bound peptide. The C-terminal segment of DnaK beyond A607 appears less well structured (Buchberger et al., 1995). A DnaK mutant protein lacking the C-terminal 94 residues exhibits normal intrinsic ATPase activity but shows decreased interaction with DnaJ, suggesting a role for the C-terminus in DnaJ binding (Wawrzynów and Zylicz, 1995). However, a DnaK mutant protein lacking 101 C-terminal residues has an ATPase that can still be stimulated efficiently by DnaJ and has retained at least some chaperone activity (Schröder and Bukau, unpublished results). In view of these conflicting results it is clear that further investigations are required to define the role of the C-terminus of DnaK 2.1.2. Substrate Binding and its Coupling to the Nucleotide Status of DnaK The control of substrate binding by the nucleotide status of DnaK is the central part of the mechanism that drives the chaperone activity of DnaK In the absence of ATP, substrates stably associate with DnaK with dissociation constants (Kd) of the complexes of typically 50 nM–5 M (Buchberger et al., 1994b; Schmid et al., 1994; Burkholder et al., 1996; Gamer et al., 1996; McCarty et al., 1996; Theyssen et al., 1996). These interactions are characterized by low association (kon) and dissociation (koff) rate constants (Schmid et al., 1994; Gamer et al., 1996; Theyssen et al., 1996). Binding of ATP leads to rapid dissociation of bound substrate, due to an increase in the koff by 2 to 3 orders of magnitude (Palleros et al., 1993; Schmid et al., 1994; McCarty et al., 1995; Theyssen et
Molecular chaperones and folding catalysts
668
al., 1996). ATP binding also increases the kon for substrate binding (Schmid et al., 1994; Theyssen et al., 1996; Pierpaoli et al., 1997) which is proposed to have a profound effect on the reactivity of DnaK towards substrates (see section 4). This switch in binding kinetics is best explained by an ATP-induced opening of the substrate binding pocket of DnaK, which may include a displacement of the helical lid that closes the pocket (Zhu et al., 1996) (Figures 2 and 3). ATP-induced conformational changes in DnaK have indeed been observed using a variety of approaches including Fourier transform infrared and tryptophan fluorescence spectroscopy, partial tryptic degradation, and small angle X-ray scattering analysis (Liberek et al., 1991; Banecki et al., 1992; Palleros et al., 1993; Buchberger et al., 1995; Shi et al., 1996; Theyssen et al., 1996). Major nucleotide dependent conformational changes occur within the substrate binding domain (Buchberger et al., 1995). Further work is needed to clarify the molecular basis for the coupling mechanism (see also Ha et al. in this volume) (Rüdiger et al., 1997a).
Figure 3 Model for the mechanism of domain coupling in DnaK DnaK is drawn schematically, with the grayscale for the different structural elements adopted from Figure 1. The nucleotide binding site is depicted as box in the centre of the ATPase domain (two half circles), the substrate as small circle. In the ADP-bound state (top), substrate is tightly bound under a lid formed by -helix B of the substrate binding domain. In the ATP-bound state (bottom), the two ATPase subdomains adopt a slightly different conformation leading to an altered interaction between the ATPase and substrate binding domains. This interaction induces structural alterations in the substrate binding domain and, consequently, the opening of the lid helix and substrate release. (Taken from Rüdiger et al., 1997a).
DnaK binds hydrophobic stretches of peptides (Fourie et al., 1994; Gragerov et al., 1994;
The DnaK chaperone system
669
Rüdiger et al., 1997b) that are in extended conformation (Landry et al., 1992; Zhu et al., 1996). Bound peptides are unlikely to be in helical conformation since helices do not fit into the substrate binding pocket. The association of the peptide substrate with DnaK in the crystal structure involves side chain contacts as well as hydrogen bonding and van der Waals interactions with the peptide backbone (Zhu et al., 1996). This interaction mode probably prevents DnaK from binding to those sites in protein substrates where a stretch of hydrophobic side chains, but not the corresponding peptide backbone, is exposed at the surface. Furthermore, the bound peptide segment is completely enclosed by the substrate binding cavity and the lid (Figure 2). This suggests that binding of a polypeptide to DnaK requires separation of the interacting peptide segment from the remainder of the substrate by approx. 10 Å and therefore substantial local unfolding. DnaK binding sites within protein sequences and the consensus binding motif have been identified by screening a library of cellulose-bound peptides representing 37 complete sequences of proteins for DnaK binding (Rüdiger et al., 1997a, b)
Figure 4 The substrate binding motif of DnaK. DnaK binding sites in protein sequences were identified by screening of peptide scans for DnaK binding (Rüdiger et al., 1997b). The DnaK binding motif in substrates consists of a core (Hy) of up to five large hydrophobic or aromatic residues and flanking regions (+) enriched in basic residues which are of decreasing importance with increasing distance from the core.
(Figure 4). DnaK binding sites are frequent within protein sequences, occuring statistically every 36th residue and localizing preferentially to -strands of the corresponding folded proteins. The consensus binding motif is composed of a hydrophobic core of 4–5 residues that is flanked by basic segments which most likely electrostatically interact with the negatively charged surroundings of the substrate binding pocket. The hydrophobic cores of individual binding sites are built up in most cases by 2–4 hydrophobic residues, with Leu being by far the most enriched residue found in about 90% of the binding peptides. This strong enrichement of Leu is probably explained by the existence of a hydrophobic pocket within the substrate binding cavity of DnaK that allows for tight binding of Leu (Zhu et al., 1996) (Figure 2). Acidic residues are excluded from the cores and disfavored in the flanking regions of peptide substrates. Based on the results of this study an algorithm has been established that predicts DnaK binding sites in sequences with high accuracy (Rüdiger et al., 1997b). Together, these characteristics of the substrate binding motif fit well with the hydrophobic character and size as well as the electrostatic surrounding of the binding cavity. This motif also defines the protein targets of the chaperone activity of DnaK since it is typically found in the hydrophobic cores of folded proteins and thus only available for DnaK binding when the protein is unfolded to a significant extent at least locally. Such unfolded states might e.g. result from misfolding upon stress treatment of cells, or occur during de novo synthesis and transport of proteins across membranes.
Molecular chaperones and folding catalysts
670
2.2. DnaJ Since the discovery of E. coli DnaJ more than 140 DnaJ homologs (also referred to as Hsp40) have been identified in prokaryotic and eukaryotic cells and viruses, with the exception of specific archaebacteria lacking the entire DnaK chaperone system (Laufen et al., 1997). DnaJ’s are modular multidomain proteins (Figure 1) that share the highly conserved J-domain of approximately 78 amino acids (Bork et al., 1992; Silver and Way, 1993; Cyr et al., 1994). This domain is required for the functional interaction of DnaJ proteins with Hsp70 partner proteins. There is increasing evidence that the cooperation between DnaJ and Hsp70 proteins is a conserved, obligatory step for the chaperone activity of Hsp70. On the other hand, DnaJ proteins exhibit strong variability with respect to their other domains and are mosaic proteins composed of modules with diverse functions (Laufen et al., 1997). This diversity of DnaJ proteins is an important determinant for the observed diversity of cellular functions of Hsp70 chaperone systems. 2.2.1. Structure E. coli DnaJ is a heat shock protein with a molecular mass of 41 kDa (Zylicz et al., 1985). It was proposed to be a dimer on the basis of glycerol gradient centrifugation experiments (Zylicz et al., 1985). Analytical ultracentrifugation under equilibrium sedimentation conditions, however, revealed a pH-dependent broad distribution of oligomers, ranging from monomers to 20 mers at pH 7.7, which probably result from a rapid equilibrium between different oligomeric forms (Schönfeld et al., 1995b). The functionally active form of DnaJ is currently unknown. E. coli DnaJ is composed of 4 modules as indicated by sequence alignment with DnaJ homologs (Bork et al., 1992; Caplan and Douglas, 1991; Luke et al., 1991; Laufen et al., 1997) and analysis of DnaJ fragments (Wall et al., 1994; Wall et al., 1995; Banecki et al., 1996; Karzai and McMacken, 1996; Szabo et al., 1996). These domains are the highly conserved J domain (residues 2–78), the Gly/Phe-rich region (residues 79–106), the Zinc binding domain (residues 144–200) and a low homology region at the C-terminus (residues 201–376) (Figure 1). The activity of DnaJ fragments in DnaJ assisted folding reactions requires the complete structure of DnaJ (Wall et al., 1994; Karzai and McMacken, 1996; Szabo et al., 1996). The roles of the individual domains is discussed in the following. 2.2.2. Structural Basis for the Interaction with DnaK The J domain is essential but not sufficient for the DnaJ-mediated stimulation of the DnaK ATPase activity. This is indicated by the findings that fragments of DnaJ comprising the J domain only (1–75) or lacking the J domain ( 1–73) as well as mutant proteins with single amino acid alterations within the J domain (see below) fail to stimulate the ATPase activity (Wall et al., 1994; Karzai and McMacken, 1996). Efficient, though perhaps not maximal, stimulation of the DnaK ATPase requires the J domain together with the Gly/Phe-rich region (1–106) (Wall et al., 1994; Karzai and McMacken, 1996; Szabo et al., 1996). It is an important observation that the J domain is capable of
The DnaK chaperone system
671
stimulating the ATPase activity in the presence of a peptide substrate for DnaK (Karzai and McMacken, 1996). This led to the proposal that the Gly/Phe-rich region constitutes a second determinant for the interaction with DnaK by mimicking a peptide substrate. Accordingly, the maximal stimulation of the ATPase activity of DnaK involves a “bipartite signaling mechanism” composed of two interactions: the J-domain with an unknown binding site on DnaK, and the Gly/Phe-rich region with the substrate binding pocket of DnaK (Karzai and McMacken, 1996). The interaction of DnaK with the Gly/Phe-rich region allows DnaK to associate with a protein substrate bound to DnaJ. The failure of a DnaJ mutant protein lacking the Gly/Phe-rich region ( 77–107) to target DnaK to substrates in presence of ATP (Wall et al., 1995) is in agreement with this model.
Figure 5 NMR-structure of the J domain of DnaJ. NMR structure “1” of the Brookhaven data base entry 1XBL (Pellecchia et al., 1996) comprising residues 2–108 is shown as ribbon (top) and space filling (bottom), representation in standard view and rotated rightwards by 90º and 180º. Helices are numbered according to Pellecchia et al. (1996). The conserved HPD motif is shown as sticks (top; side, chains in black, backbone in gray) or as black spheres (bottom). Residues 78–108 are poorly structured and not defined in the NMR
Molecular chaperones and folding catalysts
672
structure (Taken from Laufen et al. (1997)).
An alternative model which incorporates these findings and extends existing models is presented in section 4. The NMR structure of a fragment of DnaJ (residues 2–108) comprising the J domain and the Gly/Phe-rich region has been solved (Szyperski et al., 1994; Pellecchia et al., 1996) (Figure 5). The J domain consists of four helices, of which the amphipathic antiparallel helices 2 and 3 form a coiled coil connected by a flexible turn, and the helices 1 and 4 are oriented perpendicular to it. The hydrophobic core formed by helices 2 and 3 is further stabilized by hydrophobic contacts of these helices with helix 1. The loop connecting helices 2 and 3 is highly mobile which may be important for the association of DnaJ with DnaK (see below). Helix 4 is structurally flexible and its position is therefore less clear. Even more unstructured is the Gly/Phe-rich region located adjacent to the J domain. This region may serve as flexible linker connecting the J domain with the adjacent Zn binding domain. The highest degree of conservation within the J domain is found for residues forming helix 2 and the mobile loop. This loop contains the entirely conserved sequence motif His33-Pro34-Asp35 (Bork et al., 1992) (Figure 5), referred to as HPD motif, which plays a key role in the interaction DnaJ with DnaK leading to stimulation of the ATPase activity. This role is indicated by the finding that alterations in the loop strongly affect the activity of DnaJ proteins. For instance, the His33Gln exchange in DnaJ259 abolishes the ability of this mutant protein to stimulate the ATPase activity of DnaK and to cooperate with DnaK in chaperone functions (Wall et al., 1994). 2.2.3. Substrate Binding The association of DnaJ with substrates is believed to be a key step in the functional cycle of the DnaK chaperone system (Langer et al., 1992; Schröder et al., 1993; Szabo et al., 1994; Liberek et al., 1995; Gamer et al., 1996). Interactions of DnaJ with substrates have been demonstrated for folded protein substrates including λP (Alfano and McMacken, 1989; Liberek et al., 1990), RepA (Wickner et al., 1991) and (Gamer et al., 1992; Liberek and Georgopoulos, 1993; Gamer et al., 1996), and for nonnative conformers of protein substrates including rhodanese (Langer et al., 1992) and firefly luciferase (Schröder et al., 1993),as well as carboxymethylated -lactalbumin (Wawrzynów and Zylicz, 1995) and casein (Langer et al., 1992). Substrate binding occurs through the C-terminal half of DnaJ. The ability to bind to unfolded rhodanese is preserved in fragments comprising either the J domain, Gly/Phe rich region and Zinc binding domain (residues 1–209) or the Zinc binding domain and Cterminal region (residues 121–376) (Szabo et al., 1996). Furthermore, an internal deletion of the Zinc binding domain ( 144–200) decreases the affinity of DnaJ for most, but not all substrates (Banecki et al., 1996). These data are consistent with a role of the region between residues 121 and 209 in substrate binding. This region contains eight cysteines in a spacing (twofold repetition of C-X2-C-X10–13-C-X2-C) similar to that of cysteines in Zinc finger domains (Cyr et al., 1994). X-ray absorption fine structure spectroscopy revealed that its structure indeed ressembles that of Zinc finger proteins (Szabo et al., 1996), and two Zn2+ ions were found to be bound per DnaJ monomer (Banecki et al.,
The DnaK chaperone system
673
1996; Szabo et al., 1996). Site-specific mutational alteration of individual cysteine residues of DnaJ decreases the number of Zn atomes bound to DnaJ and the activity of the DnaJ mutant proteins in substrate binding. Furthermore, denatured luciferase can be crosslinked to DnaJ by using sulfhydryl-specific crosslinking reagents bound to cysteine residues of the cysteine clusters (Szabo et al., 1996). These findings indicate an involvement of the Zinc binding domain in substrate recognition. It is interesting, though, that DnaJ homologs with predicted or demonstrated substrate binding activity, including E. coli CbpA which functionally replaces DnaJ in vivo (Ueguchi et al., 1994; Ueguchi et al., 1995; Wegrzyn et al., 1996) and Thermus thermophilis DnaJ (Motohashi et al., 1996), lack the Zinc binding domain (Laufen et al., 1997). In particular the case of CbpA raises the possibility that it is another domain which mediates substrate binding. Further work is thus needed to precisely localize the substrate binding site of DnaJ. The kinetics of DnaJ association with a substrate have been determined for by plasmon surface resonance spectroscopy (Gamer et al., 1996). The kon and koff values for DnaJ complexes are 3.3 x 105 M-1s-1 and 6.2 x 10–3 s-1, respectively. DnaJ thus interacts with high affinity (Kd=20 nM) and fast with this substrate. It is possible that the rate constant for substrate release from DnaJ is affected upon interaction with DnaK (see model for chaperone cycle). The interaction of DnaJ with unfolded luciferase is fast as well, as indicated by the ability of DnaJ to partially suppress its aggregation upon dilution from denaturant (Schröder et al., 1993). In contrast, DnaK alone cannot prevent aggregation of luciferase but requires the cooperation with DnaJ to efficiently prevent aggregation and, with further help by GrpE, to refold luciferase. Together these data provide evidence that DnaJ rapidly associates with substrates, a feature that is proposed to be essential for the functional cycle of the DnaK system as will be discussed later (section 4). The consensus binding motif for DnaJ has been investigated by screening of peptide libraries (Rüdiger, Germeroth, Schneider-Mergener and Bukau, in preparation). DnaJ shows a similar pattern of peptide binding as DnaK and recognizes similar sequences, although some differences exist. Statistical analysis of the bound peptides revealed that DnaJ prefers clusters of hydrophobic and especially aromatic residues, but in contrast to DnaK does not favor a single side chain. Furthermore, in comparison to DnaK the binding of DnaJ to peptides is less strongly inhibited by negatively charged residues. 2.3. GrpE GrpE acts as nucleotide exchange factor for DnaK GrpE homologs with 25 to 70% overall sequence identity exist in various bacteria, eukaryotic mitochondria and chloroplasts, but not in the eukaryotic cytosol and endoplasmic reticulum. Hsp70 chaperone systems thus appear to differ with respect to the requirement for a nucleotide exchange factor. Cytosolic Hsp70 chaperones of yeast and man indeed require only an appropriate DnaJ homolog as co-chaperone for their activity in assisting protein folding in vitro (Levy et al., 1995). In contrast, E. coli DnaK requires GrpE for most, if not all, chaperone activities in vivo (see Burkholder and Gottesman, this volume) and in vitro (Alfano and McMacken, 1989; Wyman et al., 1993; Zylicz et al., 1989; Langer et al., 1992; Schröder et al., 1993; Skowyra and Wickner, 1993).
Molecular chaperones and folding catalysts
674
GrpE (Mr=22 kDa) forms a dimer of elongated shape that interacts stably with DnaK in the absence of ATP (Zylicz et al., 1987; Buchberger et al., 1994a; Schönfeld et al., 1995a; Wu et al., 1996), with a dissociation constant of 1 nM for the complex (Packschies et al., 1997). GrpE dissociates from DnaK upon addition of ATP (Zylicz et al., 1987; Schönfeld et al., 1995a) and, through thermodynamic coupling, accelerates the dissociation of DnaK-bound ADP and ATP (Zylicz et al., 1987; Liberek et al., 1991; McCarty et al., 1995; Packschies et al., 1997) (see section 3 for stimulation
Figure 6 Crystal structure of the complex between GrpE and the ATPase domain of DnaK (PDB entry number 1DKG). The structure of the complex was solved by Harrison et al. (1997) at a resolution of 2.8 Å. The DnaK ATPase domain is shown in yellow and the GrpE dimer in dark (DnaK distal monomer) and light (DnaK proximal monomer) blue. All contacts between GrpE and DnaK are mediated by the DnaK proximal monomer of GrpE. This interaction induces a bending in GrpE resulting in an asymmetry. Side chains of residues involved in major contacts between both molecules are shown in red and major contact regions are indicated by letters A-E.
factors). This activity of GrpE allows for rapid ATP-dependent release of substrates bound to DnaKADP. The crystal structure of a complex between the nucleotide-free ATPase domain of DnaK and a GrpE dimer has been solved at 2.8 Å resolution (Harrison et al., 1997) (Figure 6). The mutant GrpE protein used for crystallisation lacks the N-terminal 33 residues and in addition carries a Gly to Asp exchange at residue 122. The N-terminal part of the truncated GrpE forms a long α-helix of 100 Å length which accounts for the extraordinarily elongated shape of the molecule. This helix is connected by a rather disordered loop to two short α-helices and a small β-sheet domain constituted by six short -strands. The interface of the GrpE dimer is constituted by the N-terminal -helix that
The DnaK chaperone system
675
lies nearly in the same plane with the a-helix of the other monomer without forming a classical coiled coil, and the two short helices which form a four helix bundle with the corresponding helices of the other monomer. There are several major contact regions between the ATPase domain of DnaK and the β-sheet domain and the long helix of GrpE (numbered A-E in Figure 6). The contacts include an exposed loop of the ATPase domain identified earlier as essential for the stable binding of GrpE (Buchberger et al., 1994a). Furthermore, the crystal structure data are in pleasing congruence with genetic and biochemical results indicating roles for the conserved N-terminal region of GrpE (residues 40 to 85) in dimerization and the C-terminal region (residues 100 to 197) in binding to and modulation of DnaK function (Wu et al., 1996). GrpE mediates nucleotide exchange by inducing an opening of the nucleotide binding cleft of DnaK GrpE binding causes a 14º rotation of the IIB subdomain of the DnaK ATPase domain compared to the ADP-bound structure of the Hsc70 ATPase domain (Flaherty et al., 1990). This rotation displaces residues Ser274, Lys270 and Glu267 of DnaK by 2 to 3 Å. The corresponding residues of bovine Hsc70 provide hydrogen bond contacts to the adenine and ribose rings of bound ADP. GrpE thus disrupts the nucleotide binding site without directly interacting with it. This associative displacement mechanism has been inferred kinetically from the finding that nucleotide binds to DnaK GrpE complexes with a rate close to the expected diffusion controlled limit (Packschies et al., 1997). In the bound state, GrpE is bent toward DnaK thus generating asymmetry in the dimer structure. In this bent conformation the binding of a second DnaK molecule to the same GrpE dimer would require structural distortions and, as experimentally verified (Schönfeld et al., 1995a), does not occur. The dimer structure of GrpE appears primarily to be required for stabilization of the long N-terminal -helix. This helix has a rather unclear role in GrpE activity since it makes only a minor contribution to the interaction with the ATPase domain of DnaK It has been speculated that its extraordinary length allows GrpE to have an additional, yet undetected, contact with the substrate binding site of DnaK that influences substrate association with DnaK This possibility is supported by the observation that GrpE decreases the efficiency of formation of complexes between DnaK and reduced carboxymethylated -lactalbumin (Harrison et al., 1997). 3. THE ATPase CYCLE OF DnaK This section presents a stepwise kinetic dissection of the ATPase cycle (Figure 7 and Table 1), followed by an integration of the kinetic parameters into a model of the chaperone cycle of the DnaK system (Figure 8). 3.1. Binding of ATP The binding of ATP to DnaK is a rapid process involving at least two steps, the rapid (kinetically unresolved) formation of an initial weak complex (Kd=7 M), followed by a slower isomerisation (kiso=1.5 s-1) (Theyssen et al., 1996) (Figure 7 and Table 1). The rapid, first step of ATP binding was recently further resolved into two sub-steps (Pierpaoli et al., 1997; Slepenkov and Witt, 1998), which were not included in Figure 7B
Molecular chaperones and folding catalysts
676
and Table 1 for the sake of clarity. For bovine Hsc70, binding of ATP involves also two steps, and it was achieved to kinetically resolve the first step with k+1=7×105 M-1 s-1 and k-1=1.1 s-1 (Ha and McKay, 1995). The secons step of ATP binding is central to the chaperone activity of DnaK and Hsc70 since it is kinetically correlated to the ATPinduced switch of the enzyme from the high to the low affinity state for pep tide binding and consequently to peptide release. The switch of affinity states relies on a conformational change in the substrate binding domain (see 2.1.2.) that is induced by the second step of ATP binding. This change is not observed with AMPPNP or ATP S (Palleros et al., 1993; Buchberger et al., 1995; Theyssen et al., 1996) and in absence of Mg2+ and K+ (Banecki et al., 1992; Palleros et al., 1993; Feifel et al., 1996). This indicates that the binding of ATP is highly specific and the precise location of nucleotide and ions in the active site is crucial for the coupling with substrate binding. In the case of bovine Hsc70, there is direct structural evidence for specific coordination of Mg2+ and 2 K+ ions at the active site (Wilbanks and McKay, 1995) (see Ha et al., this volume). 3.2. Switch 1: ATP Hydrolysis and the Role of DnaJ The hydrolysis of ATP is the rate limiting step of the unstimulated ATPase cycle of DnaK (McCarty et al., 1995; Karzai and McMacken, 1996; Theyssen et al., 1996; Pierpaoli et al., 1997; Russell et al., 1998) with a khyd of approx. 0.0015 s-1 (McCarty et al., 1995; Theyssen et al., 1996) that seems to vary slightly with different experimental conditions (Pierpaoli et al., 1997; Russell et al., 1998) (Figure 7). At steady state most of the enzyme is therefore in the ATP bound form which is the low affinity state for peptide binding (McCarty et al., 1995). DnaK mutant proteins deficient in ATP hydrolysis, such as mutant proteins carrying Thr199Ala and Glu171Ala changes, remain constitutively in the low affinity state for peptide binding even in presence of ATP (Kamath-Loeb et al., 1995; McCarty et al., 1995; Schuster et al., 1998). This inability to switch to the high affinity state for substrate binding causes a deficiency in chaperone function (KamathLoeb and Gross, 1991; McCarty and Walker, 1994; Buchberger et al., 1994b). DnaJ controls khvd and thereby regulates the amount of DnaK present in the DnaKADPPi form (Liberek et al., 1991; McCarty et al., 1995; Wawrzynów et al., 1995; Theyssen et al., 1996; Pierpaoli et al., 1997). DnaJ was reported to stimulate the ATPase reaction at least 200-fold in single turnover experiments (Karzai and McMacken, 1996). In recent experiments using single turnover quench flow experiments DnaJ stimulated ATP hydrolysis even more, 550-fold at 1 M DnaK, 0.8 M ATP and saturating DnaJ (Laufen et al., in preparation). As introduced in section 2.2.2., the stimulation of ATP hydrolysis by DnaJ appears to require the presence of a chaperone substrate as additional signal. In the ATPase experiments discussed above such additional signal may be provided by other DnaK or DnaJ molecules present in the test solution. It should be emphasized that these experiments were performed under conditions where DnaJ oligomerizes which may increase the availability of such signal. 3.3. Switch 2: Product Release and the Role of GrpE For several energy transducing systems the release of the hydrolysis products, ADP and
The DnaK chaperone system
677
Pi, are key steps in the functional cycle. One example is actin/myosin where these steps are ordered and Pi release constitutes the power stroke of this system. For DnaK it appears that product release, while not being the power stroke itself, constitutes the ratelimiting step in the DnaJ/peptide-stimulated cycle and provides a second switch that is under tight and essential control by GrpE (Figure 7). For DnaK and bovine Hsc70 it was observed that 2 mM Pi reduces the dissociation rate constant of ADP several fold (Burkholder et al., 1994; Ha and McKay, 1994; Theyssen et al., 1996). For Hsc70 this inhibition was interpreted to suggest a mechanism of ordered product release where Pi is released prior to ADP (Ha and McKay, 1994). However, at least for DnaK these experiments do n ot rigorously clarify whether the binding of Pi and ADP is ordered or random since the rate of ADP release would also be slowed down upon addition of Pi if the b; nding of these two ligands were synergistic. Furthermore, the binding of Pi to the DnaKADP-complex seems to differ from the binding of Pi to the Hsc70 ADP complex. The dissociation rate constant of Pi from Hsc70 ADP Pi was determined to be 3.8×10-3 s-1, comparable to the rate of hydrolysis (khyd=3×10-3 s-1). In contrast, the dissociation rate constant of Pi from the DnaK ADP Pi-complex is at least 10-times faster than the rate of hydrolysis (khvd=1.5×10-3 s-1) (Theyssen et al., 1996) since otherwise the rate of hydrolysis determined under steadystate conditions would not match the khvd determined under single turnover conditions. The question about the order of product release was investigated more thoroughly recently (Russell et al., 1998) where it was clearly shown that even saturating amounts of Pi do not inhibit ADP release completely thus ruling out a mechanism where Pi has to be released first in a strictly ordered mechanism. The binding of ADP to DnaK is fairly tight with a Kd of 0.13 M (Theyssen et al., 1996). Upon stimulation of the DnaK ATPase activity by DnaJ, ADP release becomes the rate limiting step of the overall cycle. The intrinsic dissociation rate constant for ADP (koff) is 35×10-3 s-1 and thus approx. 50 fold slower than the ATP binding step (kiso) that is prerequisite for switch 1. This rate constant is dramatically affected by GrpE. The association of GrpE with a DnaK nucleotide complex reduces the affinity for nucleotide by 200-fold and accelerates nucleotide exchange by 5000-fold (Neuhofen et al., 1996; Packschies et al., 1997). 3.4. Balance of Rate Constants The ATPase cycle of DnaK establishes a delicate balance between at least two major enzyme forms, DnaKATP (K*ATP) and DnaKADPPi (KADPPi), that alternate in their affinity and exchange rates for peptide (Figure 7). Population of these conformational states of DnaK depends mainly on the two slowest steps of the overall cycle, the rate limiting ATP hydrolysis (khvd=1.5×10–3 s-1) followed by product release (koff=35×10–3 s1). Because of the similarity of the rate constants by which KATP and KADPPi are converted, these two enzyme states are well suited to be regulated by cofactors. The binding of DnaJ to the DnaKATP complex accelerates the switch of DnaK from the low affinity, fast exchange form (K* ATP) to the high affinity, slow exchange form (KADP Pi). GrpE facilitates the recycling of DnaK to the ATP-form by accelerating nucleotide exchange. The maximal stimulation of
Molecular chaperones and folding catalysts
678
Figure 7 The functional cycle of DnaK. (A) Basic features. DnaK alternates between two major states. The ADP bound state exhibits high affinity for substrates but slow exchange rates; the ATP bound state exhibits low affinity but fast exchange rates. The conversion of the ATP bound state to the ADP bound state is stimulated by substrate and,
The DnaK chaperone system
679
much more efficiently, by DnaJ which itself can associate with substrates. The conversion from the ADP bound state to the ATP bound state requires nucleotide exchange that is stimulated by GrpE. (B) Kinetic model. This model dissects kinetically the functional cycle of DnaK (K). The lengths of the double arrows indicate the equilibrium between the different states. The kinetic parameters are shown as rate constants and listed in Table 1; not determined. The cycle at the front (K indicated with bold letters), is the ATPase cycle in absence of substrate (Pep); the cycle in the back is in presence of Pep. For simplicity, the (unordered) dissociation of KADPPi is represented by a single step with the dissociation (k+4) and association (k–4) rate constants as determined for ADP. See text for a detailed discussion. GrpE (E) acts as nucleotide exchange factor for KADPPi with or without bound Pep. The ability of E to also dissociate ATP bound to DnaK is considered less important given the 200-fold lower affinity of E for KATP. DnaJ (J) stimulates the hydrolysis of K bound ATP. It is important to note that J itself can associate with substrate and may require the presence of chaperone substrate for stimulation of the ATPase of K (see Figure 8). This ability of J may prevent futile ATPase cycles in absence of substrate.
Table 1 Rate constants of the ATPase and substrate binding activities of DnaK as shown in Figure 7B
Step
Unstimulated
k−1
3 15
k+1
3 2 µM-1 s-1a
k+2
1.5 s-1a
k−2
≈ 1.510-3 s-1a
k+3
1.510-3 s-1a
k−3
no ATP synth. observeda
k+4
≈ 0.025 s
k−4
-1 -1a ≈ 0.2 µM s
k+5
0.1−0.4 µM-1 s-1a,d,e
k−5
1−4 s-1a,d
k+6
0.00002−0.01 µM-1 s-1a,d,e
k−6
0.001 s-1a,d,e
k+7
0.007−0.02 s-1a
k−7
not determined
Stimulated
s-1a 15 s-1 (by peptide)2
-1a
-1 c ≈ 0.8 s (DnaJ)
130 s-1 (GrpE)b 3 50
µM-1 s-1b
References: aTheyssen et al., 1996; bPackschies et al., 1997; cLaufen et al., in preparation;
Molecular chaperones and folding catalysts dSchmid et al.,
680
1994; eFarr et al., 1995
nucleotide exchange by GrpE is 5000-fold, and maximal stimulation of ATP hydrolysis is 550-fold. It thus appears that the two counteracting activities of ATP hydrolysis and ADP release are balanced in that their maximal factors of stimulation by DnaJ and GrpE are comparable. DnaJ and GrpE are essential for the DnaK-dependent refolding of denatured firefly luciferase (Schröder et al., 1993; Buchberger et al., 1994b; Szabo et al., 1994). Interestingly, the yield of refolded luciferase drops at GrpE: DnaJ ratios of greater than 5:1 (Packschies et al., 1997) demonstrating that the balance of ATPase stimulation and acceleration of nucleotide exchange by DnaJ and GrpE, respectively, is of crucial importance for DnaK-assisted protein folding. It is therefore not surprising that in vivo the relative levels of DnaK, DnaJ and GrpE are kept similar at a variety of growth conditions due to the common regulation of expression of their genes through the E. coli heat shock transcription factor, (Bukau, 1993). 4. THE CHAPERONE CYCLE A key question of the functional cycle of the DnaK system is the mechanism by which the DnaJ and GrpE controlled ATPase cycle is coupled to the productive association of DnaK with chaperone substrates. This coupling is required to prevent “futile” ATPase cycles that waste ATP in the absence of substrates. In the following our current view of the functional cycle of the DnaK chaperone system is summarized and an answer to this key question is suggested. The presented model is based on recent findings (Laufen, Paal and Bukau, unpublished results) and extends and modifies earlier models proposed by several laboratories including ours (Szabo et al., 1994; Gamer et al., 1996; Karzai and McMacken, 1996; Packschies et al., 1997). The kinetic dissection of substrate interactions of DnaK revealed that in the high affinity ADP state the association of DnaK with substrates is too slow to be of physiological relevance (Schmid et al., 1994; Gamer et al., 1996). This is exemplified with the association of DnaK with heat denatured luciferase in absence of ATP that is too slow to prevent its aggregation (Schröder et al., 1993). The ATP-state of DnaK, on the other hand, allows the rapid association with substrates but excludes the formation of stable complexes. Productive interaction of DnaK with substrates therefore requires its rapid initial association with substrates in the DnaKATP form followed by stabilization, or “locking-in”, of the DnaK substrate complex through conversion to the DnaKADPPi form. According to our model DnaJ is essential for this locking-in mechanism through its dual role as ATPase activating factor for DnaK and chaperone on its own right. The proposed chaperone cycle (Figure 8) starts with the transient association of DnaJ with substrates. This is followed by the rapid, high affinity interaction of DnaKATP with the DnaJ substrate complex in a process consisting of at least two steps of unclear order. One step is the interaction of DnaKATP with the J domain of DnaJ. This interaction is transient, since DnaK: DnaJ complexes were found to be unstable (Wawrzynow and Zylicz, 1995; Laufen and Bakau, unpublished results), and is not sufficient to stimulate
The DnaK chaperone system
681
ATP hydrolysis. A second step is the loading of DnaJ-bound substrate onto DnaK through interaction with the opened substrate
Figure 8 Model of the chaperone cycle of the DnaK system. Substrate (closed circle) is initially bound to DnaJ (J) via its C-terminal domains. Subsequently, the J-substrate complex associates with DnaK (K) in the ATP bound state resulting in a transient ternary KJ substrate complex. One important step during formation of this complex is the interaction of the J domain with DnaK (step 1). Within the ternary complex, J bound substrate dissociates and associates with the opened substrate binding pocket of KATP (step 2). The order of steps 1 and 2 is unclear. Both steps are required to lead to efficient and rapid stimulation of hydrolysis of K bound ATP (step 3). The ADP state of K has low affinity for J which consequently leads to dissociation of J from the K-substrate complex (step 4). GrpE (E) stimulates ADP dissociation from K. Spontaneous nucleotide release in absence of E is considered less important since E is essential for virtually all chaperone functions of the K system. Binding of ATP induces dissociation of the substrate from K The dissociated substrate partitions between productive folding (N) and rebinding to J. Although we propose that this model represents the functional cycle of the K system operating in most chaperone reactions it cannot be excluded that the cycle can also start with the association of KATP with substrate, followed by association of J with the KATP substrate complex.
Molecular chaperones and folding catalysts
682
binding pocket of DnaK ATP. Both steps together are required to provide the signal for efficient stimulation of ATP hydrolysis which then results in tight binding of substrates. This mechanism allows DnaJ to tightly couple the hydrolysis of DnaK-bound ATP with the association of DnaK with substrate. Furthermore, DnaJ may scan substrates in an ATP-independent fashion, presenting to DnaK only those substrates which show repeated and/or prolonged interaction with DnaJ. Such preselection may constitute a safeguard mechanism to restrict the ATP-consuming activity of DnaK. However, we cannot exclude at present that in some reactions the cycle starts by association of DnaKATP with a substrate, followed by the locking-in of DnaK through the ATPase stimulating activity of DnaJ (see also alternative model proposed by Zylicz et al., this volume). Since DnaJ and DnaK have similar substrate specificities (Rüdiger et al., 1997a, b; Rüdiger and Bukau, unpublished results), DnaK may bind to the same segment of a substrate polypeptide as DnaJ. A high koff of substrate-DnaJ complexes, possibly increased by binding of DnaKATP, and a high kon of formation of substrate DnaKATP complexes would allow DnaK to bind to this segment right after dissociation of DnaJ. It is tempting to speculate that the transfer of substrate from DnaJ to DnaK is facilitated by positioning of the substrate binding pockets of DnaK and DnaJ in close proximity to each other within the ternary DnaJ-substrate DnaKATP complex. DnaJ is predicted to leave the complex after dissociation from the substrate and conversion of DnaK to the ADP form. This prediction is consistent with the observations that ternary complexes are unstable in presence of ATP and competing substrate (Gamer et al., 1996) and DnaJ acts catalytically in the targeting of DnaK to substrates (Liberek et al., 1995). The reported ability to isolate ternary complexes in the absence of competitor substrates (Szabo et al., 1994; Liberek et al., 1995; Gamer et al., 1996) may result from a rapid equilibrium between the complex and its dissociated components. A final event in the functional cycle of the DnaK system is the binding of GrpE to the DnaK ADP (DnaJ)substrate complex which triggers ADP release thereby allowing the rapid rebinding of ATP and the release of substrate. Released substrate may either proceed along its folding pathway, or be rebound by DnaJ/DnaK. The regulated cycle of substrate binding and release bears several implications for the chaperone activity of the DnaK system. First, binding of aggregation-prone protein to DnaK decreases its free concentration and thus its probability to aggregate. The chaperone activity of DnaK therefore underlies the principle of kinetic partitioning. Second, there is so far no evidence that substrates fold while bound to DnaK They consequently fold most likely in free solution. Third, interactions of DnaK with substrates are repetitive thus allowing multiple folding attempts for substrates. For unfolded firefly luciferase it has been shown that repeated binding/ release cycles are indeed required for efficient DnaK-dependent refolding (Szabo et al., 1994; Buchberger et al., 1996). Fourth, the repetitive nature of the cycle provides the basis for a functional network with other chaperones in which substrates released from DnaK might freely partition between different chaperone systems. A cooperation in folding processes between the Hsp70 system and other chaperone systems has indeed been reported (Langer et al., 1992; Buchberger et al., 1996; Freeman and Morimoto, 1996; Freeman et al., 1996). Such cooperation between chaperone systems is likely to be essential for the cell’s ability to efficiently assist protein folding processes under normal as well as under stress
The DnaK chaperone system
683
conditions. 5. OTHER Hsp70 SYSTEMS Compared to the E. coli DnaK system, a less detailed understanding of the functional cycle is available for other Hsp70 systems. This is mainly due to the limited number of systematic kinetic analyses and to significant heterogeneity with regard to experimental conditions and the source of Hsp70. The available data show that the Hsp70 system, while offering room for functional and mechanistic diversity throughout evolution, seems to have conserved mechanistic key features. These include the rapid and high affinity two-step binding of ATP (Ha and McKay, 1995); the slow hydrolysis of ATP as ratelimiting step (Gao et al., 1993; Cheetham et al., 1994; Ha and McKay, 1994; Ziegelhoffer et al., 1995); the relatively fast dissociation of ADP compared to ATP (Ha and McKay, 1994) and the ability of DnaJ homologs to stimulate the ATPase activity of Hsp70 proteins (Laufen et al., 1997). The outlined key role of DnaJ in the activation of DnaK for substrate binding may also be conserved within the Hsp70 family and may provide a basis for the functional diversity of Hsp70 systems. Individual cells and cellular compartments usually contain several DnaJ homologs, e.g. four homologs in the E. coli cytosol and at least 17 homologs in S. cerevisiae, which have distinct functions in cell metabolism (Laufen et al., 1997) (see chapters in this volume by Burkholder and Gottesman; Craig et al.; Eisenberg and Greene; Dekker and Pfanner; Haas and Zimmermann). It seems plausible that these DnaJ homologs ensure that Hsp70 partner proteins are reliably targeted to distinct substrates. Differences between the E. coli DnaK system and eukaryotic Hsp70 systems include the significantly slower Pi release for bovine Hsc70 compared to DnaK (Ha and McKay, 1994). Furthermore, there is evidence that eukaryotic cytosolic DnaJ homologs, in addition to stimulating γ-phosphate cleavage, also speed up ADP (and Pi?) release (Ziegelhoffer et al., 1995). These data raise the possibility that the second checkpoint of the ATPase cycle, nucleotide exchange, is differently regulated in eukaryotic cytosolic Hsp70’s. In order to maintain productive substrate interactions, the stabilization of the ADP state, rather than the stimulation of nucleotide exchange by a GrpE homolog, might be required in some folding reactions. Indeed, the recently identified cytosolic cochaperone Hip has been proposed to serve exactly this function (Höhfeld et al., 1995). Several laboratories have made the intriguing discovery that in addition to Hip there are further binding proteins for eukaryotic cytosolic Hsp70 homologs. These proteins share the ability to associate with the ATPase domain of Hsp70 or Hsc70 and affect their chaperone activity, and it is tempting to speculate that they also interfere with the ATPase activity of their partner Hsp70. One newly discovered binding protein for human Hsc70 and Hsp70 (but not DnaK) is BAG-1 (RAP46) (Takayama et al., 1997; Höhfeld and Jentsch, 1977; Zeiner et al., 1997). BAG-1 is a multifunctional protein that suppresses the apoptotic pathway through binding to Bcl-2 and stimulates the activities of the serine/threonine-specific protein kinase Raf-1 and the plasma membrane-associated tyrosine kinase growth factor receptors for hepatocyte growth factor (HGF) and plateletderived growth factor (PDGF). BAG-1 is also found in association with steroid hormone receptors although the significance of this interaction remains unknown. BAG-1 inhibits
Molecular chaperones and folding catalysts
684
the Hsc70 and Hsp70 dependent refolding of denatured -galactosidase and thus may be a novel factor connecting chaperone activity of the Hsp70 system with cell signaling and apoptosis (Takayama et al., 1997). BAG-1 has recently been shown to stimulate exchange of nucleotide bound to Hsc70 and therefore to have a GrpE-like activity (Höhfeld and Jentsch, 1997). Together, these recent findings indicate that the chaperone activity of eukaryotic Hsp70 systems is modulated and targeted to several cellular processes through the specific association of adaptor molecules. It is premature, however, to mechanistically define the roles of these adaptors. Differences in the functional cycle may also exist within the prokaryotic world. Analysis of the Thermus thermophilus DnaK chaperone system revealed the existence of an apparently stable complex consisting of DnaK, DnaJ and a cofactor, DafA, that is required for assembly and stability of this complex (Motohashi et al., 1994; Motohashi et al., 1996). DafA is encoded by a gene that is part of an operon encoding DnaK, DnaJ and GrpE. Recent microcalorimetry studies on the isolated DnaK, DnaJ and DafA. components from T. thermophiles revealed that DafA is essential to mediate DnaK-DnaJ interactions (Klostermeier, Seidel, Reinstein, unpublished results). Furthermore, there is an apparent lack in coupling of ATP binding and peptide release of this DnaK homolog and no observable stimulation of the DnaK ATPase activity by the DnaJ homolog (Klostermeier et al., 1998). These results raise the question whether the T. thermophilus system functions with the same underlying principles as observed with the E. coli system. 6. PERSPECTIVES Significant progress has been made in recent years in dissecting the molecular basis of the chaperone activity of the DnaK/Hsp70 chaperone system. However, central aspects of its activity remain unclear and require further investigation. A more complete understanding of the interactions involved in the functional cycle, both between chaperones and substrates and between the individual chaperone components, is desirable. It should be kept in mind that the known crystal structures of components of the DnaK chaperone system provide only snapshots of the highly dynamic chaperonesubstrate interaction cycle. Of particular interest would be to solve the puzzle of how DnaK and DnaJ cooperate in the functional cycle, and to elucidate the coupling mechanism that allows control of substrate binding by nucleotides and co-chaperones. Furthermore, it is important to explore the degree of mechanistic diversity existing within the family of Hsp70 chaperones and their increasing number of cofactors. 7. ACKNOWLEDGEMENTS We thank C.Gässler and T.Laufen for preparation of figures. 8. REFERENCES Alfano, C. and McMacken, R. (1989). Ordered assembly of nucleoprotein structures at the bacteriophage replication origin during the initiation of DNA replication. J. Biol.
The DnaK chaperone system
685
Chem. , 264 , 10699–10708. Banecki, B., Liberek, K., Wall, D., Wawrzynów, A., Georgopoulos, C., Bertoli, E., Tanfani, F. and Zylicz, M. (1996). Structure-function analysis of the zinc finger region of the DnaJ molecular chaperone. J. Biol. Chem. , 271 , 14840–14848. Banecki, B., Zylicz, M., Bertoli, E. and Tanfani, F. (1992). Structural and functional relationships in DnaK and DnaK756 heat-shock proteins from Escherichia coli. J. Biol. Chem. , 267 , 25051–25058. Bardwell, J.C.A. and Craig, E.A. (1984). Major heat shock gene of Drosophila and the Escherichia coli heat-inducible dnaK gene are homologous. Proc. Natl. Acad. Sci. USA , 81 , 848–852. Bork, P., Sander, C., Valencia, A. and Bukau, B. (1992). A module of the DnaJ heat shock proteins found in malaria parasites. Trends Biochem. Sci. , 17 , 129. Buchberger, A., Schröder, H., Büttner, M., Valencia, A. and Bukau, B. (1994a). A conserved loop in the ATPase domain of the DnaK chaperone is essential for stable binding of GrpE. Nature Struct. Biol , 1 , 95–101. Buchberger, A., Valencia, A., McMacken, R., Sander, C. and Bukau, B. (1994b). The chaperone function of DnaK requires the coupling of ATPase activity with substrate binding through residue E1 71. EMBO J. , 13 , 1687–1695. Buchberger, A., Schröder, H., Hesterkamp, T., Schönfeld, H.-J. and Bukau, B. (1996). Substrate shuttling between the DnaK and GroEL systems indicates a chaperone network promoting protein folding. J. Mol. Biol. , 261 , 328–333. Buchberger, A., Theyssen, H., Schröder, H., McCarty, J.S., Virgallita, G., Milkereit, P., Reinstein, J. and Bukau, B. (1995). Nucleotide-induced conformational changes in the ATPase and substrate binding domains of the DnaK chaperone provide evidence for interdomain communication. J. Biol. Chem. , 270 , 16903–16910. Bukau, B. (1993). Regulation of the E. coli heat shock response. Molec. Microbiol. , 9 , 671–680. Burkholder, W.F., Panagiotidis, C.A., Silverstein, S.J., Cegielska, A., Gottesman, M.E. and Gaitanaris, G.A. (1994). Isolation and characterization of an Escherichia coli dnaK mutant with impaired ATPase activity. J. Mol. Biol. , 242 , 364–377. Burkholder, W.F., Zhao, X., Zhu, X. and Hendrickson, W.A. (1996). Mutations in the Cterminal fragment of DnaK affecting peptide binding. Proc. Nat. Acad. Sci. USA , 93 , 10632–10637. Cheetham, M.E., Jackson, A.P. and Anderton, B.H. (1994). Regulation of 70-kDa heatshock-protein ATPase activity and substrate binding by human DnaJ-like proteins, HSJ1a and HSJ1b. Eur. J. Biochem. , 226 , 99–107. Cyr, D., Langer, T. and Douglas, M. (1994). DnaJ-like Proteins: Molecular Chaperones and specific Regulators of Hsp70. Trends Biochem. Sci. , 19 , 176–181. Echols, H. (1986). Multiple DNA-protein interactions governing high-precision DNA transactions. Science , 233 , 1050–1056. Farr, C.D., Galiano, F.J. and Witt, S.N. (1995). Large activation-energy barriers to chaperone peptide complex-formation and dissociation. Biochemistry , 34 , 15574– 15582. Feifel, B., Sandmeier, E., Schönfeld, H.J. and Christen, P. (1996). Potassium ions and the molecular chaperone activity of DnaK. Eur. J. Biochem. , 237 , 318–321.
Molecular chaperones and folding catalysts
686
Flaherty, K.M., Deluca-Flaherty, C. and McKay, D.B. (1990). Three-dimensional structure of the ATPase fragment of a 70K heat-shock cognate protein. Nature (London) , 346 , 623–628. Flaherty, K.M., McKay, D.B., Kabsch, W. and Holmes, K.C. (1991). Similarity of the three-dimensional structures of actin and the ATPase fragment of a 70K heat shock cognate protein. Proc. Natl. Acad. Sci. USA , 88 , 5041–5045. Fourie, A.M., Sambrook, J.F. and Gething, M.J.H. (1994). Common and divergent peptide binding specificities of Hsp70 molecular chaperones. J. Biol. Chem. , 269 , 30470–30478. Freeman, B.C. and Morimoto, R.I. (1996). The human cytosolic molecular chaperones hsp90, hsp70 (hsc70). and hdj-1 have distinct roles in recognition of a non-native protein and protein refolding. EMBO J. , 15 , 2969–2979. Freeman, B.C., Toft, D.O. and Morimoto, R.I. (1996). Molecular chaperone machines: Chaperone activities of the cyclophilin Cyp-40 and the steroid aporeceptor-associated protein p23. Science , 274 , 1718–1720. Gaitanaris, G.A., Rubock, A.G.P., Silverstein, S.J. and Gottesman, M.E. (1990). Renaturation of denaturated represser requires heat shock proteins. Cell , 61 , 1013– 1020. Gamer, J., Bujard, H. and Bukau, B. (1992). Physical interaction between heat shock proteins DnaK, DnaJ, GrpE and the bacterial heat shock transcriptional factor . Cell , 69 , 833–842. Gamer, J., Multhaup, G., Tomoyasu, T., McCarty, J.S., Rüdiger, S., Schönfeld, H.-J., Schirra, C., Bujard, H. and Bukau, B. (1996). A cycle of binding and release of the DnaK, DnaJ and GrpE chaperones regulates activity of the E. coli heat shock transcription factor . EMBO J. , 15 , 607–617. Gao, B., Yumiko, E., Greene, L. and Eisenberg, E. (1993). Nucleotide binding properties of bovine brain uncoating ATPase. J. Biol. Chem. , 268 , 8507–8513. Gragerov, A., Zeng, L., Zhao, X., Burkholder, W. and Gottesman, M.E. (1994). Specificity of DnaK-peptide binding. J. Mol Biol , 235 , 848–854. Ha, J.-H. and McKay, D.B. (1994). ATPase kinetics of recombinant bovine 70 kDa heat shock cognate protein and its amino-terminal ATPase domain. Biochemistry , 33 , 14625–14635. Ha, J.-H. and McKay, D.B. (1995). Kinetics of Nucleotide-Induced Changes in the Tryptophane Fluo-rescence of the Molecular Chaperone Hsc70 and Its Subfragments Suggest the ATP-Induced Conformational Change Follows Initial ATP Binding. Biochemistry , 34 , 11635–11644. Harrison, C.J., Hayer-Hartl, M., Di Liberto, M., Hartl, F.-U. and Kuriyan, J. (1997). Crystal structure of the nucleotide exchange factor GrpE bound to the ATPase domain of the molecular chaperone DnaK. Science , 276 , 431–435. Hartl, F.U. (1996). Molecular chaperones in cellular protein folding. Nature , 381 , 571– 580. Höhfeld, J., Minami, Y. and Hartl, F.U. (1995). Hip, a novel cochaperone involved in the eukaryotic Hsc70/Hsp40 reaction cycle. Cell , 83 , 589–598. Höhfeld, J. and Jentsch, S. (1997). GrpE-like regulation of the hsc70 chaperone by the anti-apoptotic protein BAG-1. EMBO J. , 16 , 6209–6216.
The DnaK chaperone system
687
Holmes, K.C., Sander, C. and Valencia, A. (1993). A new ATP-binding fold in actin, hexokinase and Hsc70. Trends Cell Biol , 3 , 53–59. Jordan, R. and McMacken, R. (1995). Modulation of the ATPase Activity of the Molecular Chaperone DnaK by Peptides and the DnaJ and GrpE Heat Shock Proteins. J. Biol. Chem. , 270 , 4563–4569. Kabsch, W., Mannherz, H.G., Suck, D., Pai, E.F. and Holmes, K.C. (1990). Atomic structure of the actin: DNase I complex. Nature (London) , 347 , 37–44. Kamath-Loeb, AS. and Gross, C.A. (1991). Translational regulation of synthesis: requirement for an internal control element. J. Bacteriol , 173 , 3904–3906. Kamath-Loeb, AS., Lu, S.Z., Suh, W.-C., Lonetto, M.A. and Gross, C.A. (1995). Analysis of Three DnaK Mutant Proteins Suggests that Progression through the ATPase Cycle Requires Conformational Changes. J. Biol Chem. , 270 , 30051–30059. Karzai, A.W. and McMacken, R. (1996). A bipartite signaling mechanism involved in DnaJ-mediated activation of the Escherichia coli DnaK protein. J. Biol Chem. , 271 , 11236–11246. Klostermeier, D., Seidel, R. and Reinstein, J. (1998). Functional properties of the molecular chaperone DnaK from Thermus thermophilus. J. Mol Biol , in press. Landry, S.J., Jordan, R., McMacken, R. and Gierasch, L.M. (1992). Different conformations for the same polypeptide bound to chaperones DnaK and GroEL. Nature , 355 , 455–457. Langer, T., Lu, C., Echols, H., Flanagan, J., Hayer, M.K. and Hartl, F.U. (1992). Successive action of DnaK, DnaJ and GroEL along the pathway of chaperonemediated protein folding. Nature , 356 , 683–689. Laufen, T., Zuber, U., Buchberger, A. and Bukau, B. (1997). DnaJ proteins. In Molecular chaperones in proteins: Structure, function, and mode of action (Fink, A. and Goto, ed.), pp. 241–274, Marcel Dekker, New York. Levy, E.J., McCarty, J., Bukau, B. and Chirico, W.J. (1995). Conserved ATPase and luciferase refolding activities between bacteria and yeast Hsp70 chaperones and modulator. FEBS Lett. , 368 , 435–440. Liberek, K. and Georgopoulos, C. (1993). Autoregulation of the Escherichia coli heat shock response by the DnaK and DnaJ heat shock proteins. Proc. Natl Acad. Sci. USA , 90 , 11019–11023. Liberek, K., Marszalek, J., Ang, D., Georgopoulos, C. and Zylicz, M. (1991). Escherichia coli DnaJ and GrpE heat shock proteins jointly stimulate ATPase activity of DnaK. Proc. Natl. Acad. Sci. USA , 88 , 2874–2878. Liberek, K., Osipiuk, J., Zylicz, M., Ang, D., Skorko, J. and Georgopoulos, C. (1990). Physical Interactions between bacteriophage and Escherichia coli proteins required for initation of DNA replication. J. Biol. Chem. , 265 , 3022–3029. Liberek, K., Skowyra, D., Zylicz, M., Johnson, C. and Georgopoulos, C. (1991). The Escherichia coli DnaK chaperone, the 70-kDa heat shock protein eukaryotic equivalent, changes conformation upon ATP hydrolysis, thus triggering its dissociation from a bound target protein. J. Biol. Chem. , 266 , 14491–14496. Liberek, K., Wall, D. and Georgopoulos, C. (1995). The DnaJ chaperone catalytically activates the DnaK chaperone to preferentially bind the sigma 32 heat shock transcriptional regulator. Proc. Natl. Acad. Sci. USA , 92 , 6224–6228.
Molecular chaperones and folding catalysts
688
McCarty, J. and Walker, G. (1994). DnaK Mutants defective in ATPase Activity are defective in negative Regulation of the Heat Shock Response: Expression of mutant DnaK Proteins results in Filamentation. J. Bacterial. , 176 , 764–780. McCarty, J.S., Buchberger, A., Reinstein, J. and Bukau, B. (1995). The role of ATP in the functional cycle of the DnaK chaperone system. J. Mol. Biol. , 249 , 126–137. McCarty, J.S., Rüdiger, S., Schönfeld, H.-J., Schneider-Mergener, J., Nakahigashi, K., Yura, T. and Bukau, B. (1996). Regulatory region C of the E. coli heat shock transcription factor, , constitutes a DnaK binding site and is conserved among eubacteria. J. Mol. Biol. , 256 , 829–837. Motohashi, K., Taguchi, H., Ishii, N. and Yoshida, M. (1994). Isolation of the stable hexameric DnaK-DnaJ complex from Thermus thermophilus. J. Biol. Chem. , 269 , 27074–27079. Motohashi, K., Yohda, M., Endo, I. and Yoshida, M. (1996). A novel factor required for the assembly of the DnaK and DnaJ chaperones of Thermus thermophilus. J. Biol. Chem. , 271 , 17343–17348. Neuhofen, S., Theyssen, H., Reinstein, J., Trommer, W.E. and Vogel, P. (1996). Nucleotide binding to the heat shock protein DnaK as studied by ESR spectroscopy. Eur. J. Biochem. , 240 , 78–82. Packschies, L., Theyssen, H., Buchberger, A., Bukau, B., Goody, R.S. and Reinstein, J. (1997). GrpE accelerates nucleotide exchange of the molecular chaperone DnaK with an associative displacement mechanism. Biochemistry , 36 , 3417–3422. Palleros, D.R., Reid, K.L., Shi, L., Welch, W.J. and Fink, A.L. (1993). ATP-induced protein-Hsp70 complex dissociation requires K+ but not ATP hydrolysis. Nature , 365 , 664–666. Pellecchia, M., Szyperski, T., Wall, D., Georgopoulos, C. and Wüthrich, K. (1996). NMR structure of the J-domain and the Gly/Phe-rich region of the Escherichia coli DnaJ chaperone. J. Mol. Biol. , 260 , 236–250. Pierpaoli, E.V., Sandmeier, E., Baici, A., Schönfeld, H.-J., Gisler, S. and Christen, P. (1997). The power stroke of the DnaK/DnaJ/GrpE molecular chaperone system. J. Mol. Biol. , 269 , 757–768. Rüdiger, S., Buchberger, A. and Bukau, B. (1997a). Interaction of Hsp70 chaperones with substrates. Nature Struct. Biol. , 4 , 342–349. Rüdiger, S., Germeroth, L., Schneider-Mergener, J. and Bukau, B. (1997b). Substrate specificity of the DnaK chaperone determined by screening cellulose-bound peptide libraries. EMBO J. , 16 , 1501–1507. Russell, R., Jordan, R. and McMacken, R. (1998). Kinetic characterization of the ATPase cycle of the DnaK molecular chaperone. Biochemistry , 37 , 596–607. Schmid, D., Baici, A., Gehring, H. and Christen, P. (1994). Kinetics of molecular chaperone action. Science , 263 , 971–973. Schönfeld, H.-J., Schmidt, D., Schröder, H. and Bukau, B. (1995a). The DnaK chaperone system of Escherichia coli: quaternary structures and interactions of the DnaK and GrpE components. J. Biol. Chem. , 270 , 2183–2189. Schönfeld, H.J., Schmidt, D., and Zulauf, M. (1995b). Investigation of the molecular chaperone DnaJ by analytical centrifugation. Progr. Colloid Polym. Sci. , 99 , 7–10. Schröder, H., Langer, T., Hartl, F.-U. and Bukau, B. (1993). DnaK, DnaJ, GrpE form a
The DnaK chaperone system
689
cellular chaperone machinery capable of repairing heat-induced protein damage. EMBO J. , 12 , 4137–4144. Schuster, H.-P., Buchberger, A., Theyssen, H., Bukau, B. and Reinstein, J. (1998). The mutation E171A of the molecular chaperone DnaK is deficient in ATP hydrolysis but still couples the binding of ATP to peptide release. Manucscript in preparation. Sherman, M. and Goldberg, A.L. (1992). Involvement of the chaperon DnaK in the rapid degradation of a mutant protein in Escherichia coli. EMBO J. , 11 , 71–77. Shi, L., Kataoka, M. and Fink, A.L. (1996). Conformational characterization of DnaK and its complexes by small-angle X-ray scattering. Biochemistry , 35 , 3297–3308. Silver, P.A. and Way, J.C. (1993). Eucaryotic DnaJ Homologs and the specificity of Hsp70 activity. Cell , 74 , 5–6. Skowyra, D., Georgopoulos, C. and Zylicz, M. (1990). The E. coli dnaK gene product, the Hsp70 homolog, can reactivate heat-inactivated RNA polymerase in an ATP hydrolysis-dependent manner. Cell , 62 , 939–944. Skowyra, D. and Wickner, S. (1993). The interplay of the GrpE heat shock protein and Mg2+ in RepA monomerization by DnaJ and DnaK. J. Biol Chem. , 268 , 25296– 25301. Slepenkov, S.V. and Witt, S.N. (1998). Kinetics of the reactions of the Escherichia coli molecular chaperone DnaK with ATP: Evidence that a three-step reaction precedes ATP hydrolysis. Biochemistry , 37 , 1015–1024. Straus, D.B., Walter, W.A. and Gross, C.A. (1988). Escherichia coli heat shock gene mutants are defective in proteolysis. Genes & Dev. , 2 , 1851–1858. Szabo, A., Korzun, R., Hartl, F.U. and Flanagan, J. (1996). A zinc finger-like domain of the molecular chaperone DnaJ is involved in binding to denatured protein substrates. EMBO J. , 15 , 408–417. Szabo, A., Langer, T., Schröder, H., Flanagan, J., Bukau, B. and Hartl, F.U. (1994). The ATP hydrolysis-dependent reaction cycle of the Escherichia coli Hsp70 system-DnaK, DnaJ and GrpE. Proc. Natl. Acad. Sci. USA , 91 , 10345–10349. Szyperski, T., Pellecchia, M., Wall, D., Georgopoulos, C. and Wüthrich, K, (1994). NMR structure determination of the Escherichia coli DnaJ molelcular chaperone: secondary structure and backbone fold of the N-terminal region (residues 2–108). containing the highly conserved J domain. Proc. Natl. Acad. Sci. USA 91 , 11343–11347. Takayama, S., Mimston, D.N., Matsuzawa, S., Freeman, B., Aime-Sempe, C., Xie, Z., Morimoto, R.I., Reed, J.C. (1997). BAG-1 modulates the chaperone activity of Hsp70/Hsc70. EMBO J. , 16 , 4887–4896. Theyssen, H., Schuster, H.-R, Bukau, B. and Reinstein, J. (1996). The second step of ATP binding to DnaK induces peptide release. J. Mol. Biol. , 263 , 657–670. Ueguchi, C., Kakeda, M., Yamada, H. and Mizuno, T. (1994). An analogue of the DnaJ molecular chaperone in Escherichia coli. Proc. Natl. Acad. Sci. USA , 91 , 1054–1058. Ueguchi, C., Shiozawa, T., Kakeda, M., Yamada, H. and Mizuno, T. (1995). A study of the double mutation of dnaJ and cbpA, whose gene products function as molecular chaperones in Escherichia coli. J. Bacteriol. , 177 , 3894–3896. Wall, D., Zylicz, M. and Georgopoulos, C. (1994). The NH2-terminal 108 amino acids of the Escherichia coli DnaJ protein stimulate the ATPase activity of DnaK and are sufficient for 1 replication. J. Biol. Chem. , 269 , 5446–5451.
Molecular chaperones and folding catalysts
690
Wall, D., Zylicz, M. and Georgopoulos, C. (1995). The conserved G/F motif of the DnaJ chaperone is necessary for the activation of the substrate binding properties of the DnaK chaperone. J. Biol. Chem. , 270 , 2139–2144. Wang, T.-F., Chang, J. and Wang, C. (1993). Identification of the peptide binding domain of Hsc70. J. Biol. Chem. , 268 , 26049–26051. Wawrzynów, A., Banecki, B., Wall, D., Liberek, K, Georgopoulos, C. and Zylicz, M. (1995). ATP hydrolysis is required for the DnaJ-dependent activation of DnaK chaperone for binding to both native and denatured protein substrates. J. Biol. Chem. , 270 , 19307–19311. Wawrzynów, A. and Zylicz, M. (1995). Divergent effects of ATP on the binding of the DnaK and DnaJ chaperones to each other, or to their various native and denatured protein substrates. J. Biol. Chem. , 270 , 19300–19306. Wegrzyn, A., Taylor, K. and Wegrzyn, G. (1996). The cbpA chaperone gene function compensates for dnaJ in lambda plasmid replication during amino acid starvation of Escherichia coli. J. Bacteriol. , 178 , 5847–5849. Wickner, S., Hoskins, J. and McKenny, K. (1991). Function of DnaJ and DnaK as chaperones of origin specific DNA binding by RepA. Nature (London) , 350 , 165– 167. Wickner, S.H. (1990). Three Escherichia coli heat shock proteins are required for P1 plasmid replication: formation of an active complex between E. coli DnaJ protein and the P1 initiator protein. Proc. Natl. Acad. Sci. USA , 87 , 2690–2694. Wilbanks, S.M. and McKay, D.B. (1995). How Potassium Affects the Activity of the Molecular Chaperone Hsc70. II. Potassium Binds Specifically in the ATPase Active Site. J. Biol. Chem. , 270 , 2251–2257. Wild, J., Altman, E., Yura, T. and Gross, C.A. (1992). DnaK and DnaJ heat shock proteins participate in protein export in Escherichia coli. Genes Dev. , 6 , 1165–1172. Wild, J., Rossmeissl, P., Walter, W.A. and Gross, C.A. (1996). Involvement of the DnaK-DnaJ-GrpE chaperone team in protein secretion in Escherichia coli. J. Bacteriol. , 178 , 3608–3613. Wu, B., Wawrzynów, A., Zylicz, M. and Georgopoulos, C. (1996). Structure-function analysis of the Escherichia coli GrpE heat-shock protein. EMBO J. , 15 , 4806–4816. Zeiner, M., Gebauer, M. and Gehring, U. (1997). Mammalian protein RAP46: an interaction partner and modulator of 70 kDa heat shock proteins. EMBO J. , 16 , 5483– 5490. Zhu, X., Zhao, X., Burkholder, W.F., Gragerov, A., Ogata, C.M., Gottesman, M. and Hendrickson, W.A. (1996). Structural analysis of substrate binding by the molecular chaperone DnaK, Science , 272 , 1606–1614. Ziegelhoffer, T., Lopez-Buesa, P. and Craig, E. (1995). The dissociation of ATP from hsp70 of Saccharomyces cerevisiae is stimulated by both Ydj1p and peptide substrates. J. Biol. Chem. , 270 , 10412–10419. Ziemienowicz, A., Skowyra, D., Zeilstra-Ryalls, J., Fayet, O., Georgopoulos, C. and Zylicz, M. (1993). Both the Escherichia coli Chaperone Systems GroEL/Gro/ES and DnaK/DnaJ/GrpE, can activate heat treated RNA Polymerase. Different mechanisms for the same Activity. J. Biol. Chem. , 268 , 25425–25431. Zylicz, M., Ang, D. and Georgopoulos, C. (1987). The grpE protein of Escherichia coli.
The DnaK chaperone system
691
Purification and properties. J. Biol. Chem. , 262 , 17437–17442. Zylicz, M., Ang, D., Liberek, K, and Georgopoulos, C. (1989). Initiation of DNA replication with purified host- and bacteriophage-encoded proteins: the role of the DnaK, DnaJ and GrpE heat shock proteins. EMBO J. , 8 , 1601–1608. Zylicz, M., Yamamoto, T., McKittrick, N., Sell, S. and Georgopoulos, C. (1985). Purification and properties of the dnaJ replication protein of Escherichia coli. J. Biol. Chem. , 260 , 7591–7598.
27. MECHANISMS OF ATP-INDEPENDENT VS. ATP-DEPENDENT CHAPERONES SUCHIRA BOSE1, MONIKA EHRNSPERGER and JOHANNES BUCHNER* 1Department of Biochemistry, School of Medical Sciences, University of Bristol, University Walk, Bristol BS8 7TD, United Kingdom Institut für Biophysik and Physikalische Biochemie, Universität Regensburg, 93040 Regensburg, Germany
1. Introduction 2. Heat Shock Protein 47 2.1. Molecular Properties of Hsp47 2.2. The pH Dependent Interaction of Hsp47 with Collagen 3. Small Heat Shock Proteins 3.1. Molecular Properties of sHsps 3.2. sHsps Chaperone Protein Folding in Cooperation with Hsp70 4. Heat Shock Protein 90 4.1. Molecular Properties of Hsp90 4.2. Mechanism of Hsp90 Chaperone Action 5. ATP function in Chaperone Action 6. Acknowledgements 7. References 1. INTRODUCTION Heat shock proteins (Hsps) are synthesised in response to external stresses such as a sudden increase in temperature (Nover, 1991). The predominant classes of stress proteins including GroE, Hsp70, Hsp90 and small Hsps (sHsps) have been implicated in protein folding as molecular chaperones (Morimoto et al., 1994; Jakob and Buchner, 1994; Buchner, 1996). While the precise molecular mechanisms of these chaperones are still under extensive investigation, it recently has become increasingly clear that chaperones can be divided into two groups: those that act ATP-dependently and those that function independent of ATP. Here, we will outline the properties of ATP-independent chaperones with a view to comparing their differences to ATP-dependent chaperones. *Corresponding author
Molecular chaperones and folding catalysts
694
Chaperones that require ATP hydrolysis to mediate protein folding reactions include members of the GroE and Hsp70 families. Both the ATPase activity of these proteins and their influence in assisted protein folding have been analysed in detail (Morimoto et al., 1994; McKay et al., 1994; Todd et al., 1994; Jackson et al., 1993; chapters by Burston and Saibil and Ranson and Clarke for GroE; chapters by Ha et al. and Buchberger et al. for Hsp70, this volume). Those chaperones that function independent of ATP include sHsps (Jakob et al., 1993), DnaJ (Langer et al., 1992), SecB (Randall & Hardy, 1995), calnexin (Helenius, 1994), Hsp47 (Saga et al., 1987, El-Thaher et al., 1996), and the Hsp90 partner proteins p23 and FKBP52 or cyclophilin40 (Cyp40) (Bose et al., 1996, Freeman et al., 1996). For Hsp90, ATP-dependent and ATPindependent functions have been described (Thomas Scheibel and Buchner, unpublished). In this chapter, we will concentrate on the mechanisms of Hsp47, sHsps, Hsp90 and its partner proteins. 2. HEAT SHOCK PROTEIN 47 Hsp47, also known as colligin (Jain et al., 1994) or J6 protein (Wang, 1994) was identified in 1986 as the major collagen-binding heat-inducible glycoprotein in fibroblasts (Nagata et al., 1986; Nagata & Yamada, 1986). The protein is overexpressed in response to environmental stresses such as heat. It is a resident endoplasmic reticulum (ER) protein that binds specifically to various types of collagens and procollagens (Nagata et al., 1986; Takechi et al., 1992; for a review, see Nagata, 1996). Hsp47 plays a major role in collagen processing and quality control under stress conditions by preventing the secretion of procollagen with abnormal conformation (Nakai et al., 1992; Nagata, 1996).
Mechanisms of ATP-independent vs. ATP-dependent chaperones
695
2.1. Molecular Properties of Hsp47 Cloning from various species (Hirayoshi et al., 1991; Takechi et al., 1992) has revealed that Hsp47 (a) has a N-terminal signal sequence and two glycosylation sites, both of which are fully glycosylated, (b) has an RDEL (Arg-Asp-Glu-Leu) sequence at its Cterminus, which is assumed to act as an ER retention signal (Andres et al., 1990), and (c) is a member of the serine protease inhibitor (serpin) superfamily (Clarke et al., 1991; Clarke et al., 1993). The localization of Hsp47 to the ER was confirmed by immunofluorescence and immunoelectron microscopic techniques (Saga et al., 1987; Nakai et al., 1990). The direct association of Hsp47 with procollagen in vivo was shown by immunoprecipitation using anti-Hsp47 antibody after treatment of cells with a membrane-permeable cross-linking agent (Nakai et al., 1992). Pulse-chase experiments together with immunoprecipitation revealed that Hsp47 transiently binds to and dissociates from procollagen prior to procollagen secretion. However, when cells are under stress conditions, the conformationally abnormal procollagen species formed were shown to be bound to Hsp47 and retained in the ER for extended periods of time (Nakai et al., 1992). Similar observations have been reported for BiP (Haas, 1994; Wei and Hendershot, 1996), the ER homologue of Hsp70, and Calnexin (Bergeron et al., 1994) with a view that when correct folding of newly synthesized polypeptides is inhibited, either by preventing glycosylation or by introduction of folding-incompetent mutants, the misfolded proteins are bound to chaperones for longer periods and thus retained in the ER. Hence, Hsp47 has been postulated to play a similar role in collagen processing and quality control under stress conditions. In this context, it has been demonstrated that the synthesis of both Hsp47 and type IV collagen increases drastically during differentiation of the mouse tetratocarcinoma cell line F9 (Kurkinen et al., 1984; Takechi et al., 1992). There is also a strong correlation in the regulation of Hsp47 synthesis under pathological conditions in vivo as depicted by the dramatic increase in the synthesis of Hsp47 as well as type I and III collagens during the progression of rat liver fibrosis induced by carbon tetrachloride (Masuda et al., 1994). Interestingly, Hsp47 synthesis is not observed in cells where collagen is not produced (Nagata et al., 1991). In vitro, the binding of Hsp47 to collagen has been studied by surface plasmon resonance. The association and dissociation rate constants, and the dissociation constant have been determined using purified collagens (types I–V) and recombinant Hsp47 (Natsume et al., 1994). Relatively low dissociation constants (in the range 10–6-10–7M) are explained by the rapid dissociation rate constant (Kd>10-2s-l) and the high association rate constant (Ka∼2×104M-1s-1). This is in agreement with the transient binding of Hsp47 to procollagen observed in vivo (see above). In this context, it has been postulated that the rapid decrease in the free Hsp47 concentration could account for its dissociation from procollagen in the cis-Golgi (Natsume et al., 1994; Satoh et al., 1996). However, changes in pH may also be critical for dissociation of Hsp47 from procollagen, as discussed below.
Molecular chaperones and folding catalysts
696
2.2. The pH Dependent Interaction of Hsp47 with Collagen The binding of Hsp47 to procollagen and collagen (Types I–V) is ATP independent (Nakai et al., 1992). Release seems to be regulated by solvent conditions, especially by intraorganellar pH (Saga et al., 1987; Nakai et al., 1992; Natsume et al., 1994). Saga and coworkers demonstrated that Hsp47 dissociates from gelatin-Sepharose at pH lower than 6.3 (Saga et al., 1987). Later, it was shown that procollagen coprecipitated with Hsp47 is dissociated from Hsp47 after washing with a low pH buffer, and that procollagen is not crosslinked in vivo at low intracellular pH (Nakai et al., 1992). This has led to the proposal that the release mechanism may involve a lowering of pH (acidification) (see below). Recently, the effect of pH on the secondary and tertiary structure of Hsp47 was determined by circular dichroism (CD) and fluorescence spectroscopy (El-Thaher et al., 1996). Solutions of the protein were titrated from pH 7.5 to 5.0 and then back to pH 7.5. Secondary structure changes were observed by far UV CD in the downward titration. The observed changes could be grouped into three phases: I, small secondary structure changes were observed between pH 7.5 and 6.4; II, pronounced changes in secondary structure occured between pH 6.3 and 5.8; III, between pH 5.7 and 5.0, no further structural alteration was observed. The data suggest that the protein undergoes pHinduced structural changes, going from a high pH (or alkali) state, to a low pH (acid) state via an intermediate state. The transition was associated by a decrease in -helical contents and an increase in -sheet character. Tertiary structural changes also indicated that the acid state was significantly different from that of the alkali and intermediate states. These changes were found to be fully reversible. Taken together, these data suggest that the interaction of Hsp47 with procollagen and collagen could be modulated by pH through reversible conformational changes in Hsp47. Although there is no direct evidence to demonstrate that the conformational changes are occuring in the cell, the interior of endocytic vesicles, lysosomes, portion of the transGolgi apparatus and certain secretory vesicles are known to be acidic (Anderson & Orci, 1988; Wilson et al., 1993; Kim et al., 1996). This acidification is attributed to the action of the vesicular proton ATPase (see Anderson & Orci, 1988, for a review). The interaction of the murine T cell receptor (TCR) with TRAP, an ER resident protein, has also been shown to be disrupted at acidic pH (Bonifacino et al., 1988). The dissociation kinetics of TRAP seem to follow a similar pattern to that of Hsp47. Acidification in the Golgi compartments may therefore regulate the dissociation of procollagen from Hsp47, allowing for the secretion of procollagen out of the ER (Figure 1). Recently, the intracellular interaction of Hsp47 with procollagen was examined by coimmunoprecipitation in combination with pulse-chase experiments with a series of protein secretion inhibitors (Satoh et al., 1996; Nagata, 1996). If the secretion of procollagen is blocked by , í-dipyridyl, an iron chelator that inhibits procollagen triple helix formation, or by the presence of brefeldin A, which inhibits protein transport between the ER and the Golgi, procollagen was found associated with Hsp47 during the chase period in the intermediate compartment. However, when inhibitors of the post-cisGolgi were used, dissociation of Hsp47 from procollagen readily occured, suggesting that the interaction between Hsp47 and procollagen is disrupted in the cis-Golgi compartment
Mechanisms of ATP-independent vs. ATP-dependent chaperones
697
(Satoh et al., 1996). This dissociation could be attributed to the intracellular pH of the cis-Golgi, which is believed to be about 6.4 (Wilson et al., 1993; Kim et al., 1996). Thus, under the slightly acidic conditions of the cis-Golgi, a conformational change in Hsp47 could lead to its dissociation from procollagen (Figure 1). However, it remains unclear whether the decrease in concentration of free Hsp47 or the acidification in the cis-Golgi is the critical factor bringing about dissociation. 3. SMALL HEAT SHOCK PROTEINS sHsps are an abundant and ubiquitous family of stress proteins that are present in all organisms studied to date. The number of sHsps can vary within species. In Drosophila and yeast, several members have been identified, only one is known to exist in mammals, while plants contain up to 20 members (Arrigo and Landry, 1994). Interestingly, Bcrystallin, which is not restricted to the eye lens but is present in several non-lenticular tissues as well (Bhat and Nagineni, 1989), was shown to be homologous and structurally and functionally related to the sHsps (Ingolia & Craig, 1982; Klemenz et al., 1991; Horwitz, 1992, Merck et al., 1993; Jakob et al., 1993). The prokaryotic homologues, called inclusion body proteins (IbpA and B) were found in association with aggregates of recombinant protein in E. coli (Allen et al., 1992; Chuang et al., 1993). sHsps have been identified in several cellular compartments. In plants, they are localised in the chloroplast, mitochondria, ER
Figure 1 Model for the Chaperone Function of Hsp47. The individual procollagen chains are synthesized on membrane-bound ribosomes. Hsp47 binds to the procollagen precursors, called pro chains, during translation into the endoplasmic reticulum (ER). Within the lumen of the ER, selected proline and lysine residues are hydroxylated to form hydroxyproline and hydroxylysine,
Molecular chaperones and folding catalysts
698
respectively. The triple helix is formed by the combination of pro chains; for type I collagen, two 1(I) and one 2(I) chain start to combine at the carboxy termini via inter-chain disulphide bonds. Triple helix formation proceeds from the carboxy termini to the amino termini during the assembly of the procollagen molecule. Binding of Hsp47 during these steps allows for correct folding and assembly of the procollagen molecules. The triple helix is then transported to the cisGolgi compartment where dissociation of Hsp47 occurs, possibly by conformational changes in Hsp47 under the acidic conditions of the cis-Golgi, or, the low concentration of Hsp47 in this compartment favours dissociation. Hsp47 molecules are cycled back to the ER, in piggy-back fashion, via the KDEL receptor. The procollagen continues to move through the Golgi complex until it is secreted at the cell surface. (Model according to Nagata, 1996).
and the cytosol (cf. Waters et al., 1996) whereas in other eukaryotes, mainly cytosolic representatives have been characterised. Both B-crystallin and sHsps have been shown to act as molecular chaperones in protein folding and unfolding reactions (Jakob et al., 1993; Horwitz, 1992; Ehrnsperger et al., 1997; Lee et al., 1997). 3.1. Molecular Properties of sHsps In contrast to other Hsp families, the homology between different sHsps is relatively low (Lindquist and Craig, 1988; Jakob and Buchner, 1994). However, they are grouped together according to (a) regions of homology in the C-terminal part of the protein, including a flexible extension of variable length (Wistow, 1985; Plesofsky et al., 1992; Carver et al., 1992; 1995), (b) their increased expression upon heat shock (Lindquist et al.,1982, Inaguma et al., 1992; Klemenz et al., 1993; Chuang et al., 1993), and (c) their relatively small monomeric size (15–30 kDa). sHsps exist within the cell as large oligomeric complexes comprising of 9 to 50 subunits with an average molecular mass of 300–800 kDa (Chiesa et al., 1990; Behlke et al., 1991; Lee et al., 1995; Chang et al., 1996; cf. Groenen et al., 1994). With exposure to differing stress factors, sHsps rapidly become phosphorylated (Kim et al., 1984; Landry et al., 1992; Mehlen and Arrigo, 1994; Kantorow and Piatigorsky, 1994; see chapter by Bensaude, this volume). However, the function of phosphorylation is still largely unknown. In this context, mammalian Hsp27 seems to be involved in actin binding and microfilament dynamics in a phosphorylation dependent manner (Miron et al., 1988; Lovoie et al., 1993; Benndorf et al., 1994; Loktionova et al., 1996). Upon phosphorylation, the oligomeric size undergoes significant changes resulting in smaller or larger complexes (Arrigo et al., 1988; Collier et al., 1988; Nover et al., 1989; Kato et al., 1994; Mehlen et al., 1995). These findings, together with the fact that sHsp expression is modulated by hormone stimuli, indicate that the proteins may be involved in signal transduction processes (Arrigo and Landry, 1994; Loktionova et al., 1996). Under physiological conditions, the amounts of sHsps varies according to cell type and organism studied (Ciocca et al., 1993). In addition, the expression of sHsps is dependent upon development, growth cycle, differentiation and the oncogenic status of the cell (Bond and Schlessinger, 1987; Gaestel et al., 1989; Cràte and Landry, 1990; Pauli et al.,
Mechanisms of ATP-independent vs. ATP-dependent chaperones
699
1990; Ciocca et al., 1993; Klemenz et al., 1993). Under heat shock and other stress conditions, sHsp expression is induced 10–20 fold, finally amounting to as much as 1 % of the total cellular protein (Arrigo & Landry, 1994). Overexpression of sHsps has not only been observed after heat shock but is connected to a variety of other stress factors including oxidative stress, cytotoxic agents or hormone stimuli (Mehlen et al., 1995; Huot et al., 1996; Lavoie et al., 1995; Ciocca et al., 1993). In mammalian cells, overexpression of sHsps was shown to convey a significant degree of thermoresistance to cells (Landry et al., 1989; Knauf et al., 1992; Aoyama et al., 1993; van den Jissel et al., 1994). The proposed functions of sHsps are manifold, ranging from processes such as RNA stabilization (Nover et al., 1989), elastase inhibition (Voorter et al., 1994) to interaction with the cytoskeleton (Miron et al., 1991; Lavoie et al., 1993; Benndorf et al., 1994; Nicholl and Quinlan, 1994). Furthermore, overexpression of sHsps has been repeatedly reported in connection with several diseases including cancer and neurodegenerative disorders such as Alexander disease or Alzheimer’s disease (Ciocca et al., 1993; Fuller et al., 1994; Iwaki et al., 1993; Thomas et al., 1995). However, the function of sHsps in these disorders remains unclear. Many of the agents and conditions leading to enhanced expression of sHsps, especially heat shock, are known to affect the conformation and therefore the function of cellular proteins (Arrigo and Landry, 1994; Raman et al., 1995). sHsps are heat shock proteins and thus the question arose whether sHsps are chaperones. Experiments utilising well established in vitro assay systems to analyse chaperone function revealed that sHsps possess properties similar to other well-established chaperones (Jakob & Buchner, 1994; Buchner, 1996). sHsps as well as a-crystallin, which was previously thought to play only a structural role, exhibit protective functions, including suppression of aggregation of thermally unfolding proteins and increasing reactivation yields during refolding processes in vitro (Horwitz, 1992; Jakob et al., 1993; Jinn et al., 1989; Merck et al., 1993; Rao et al., 1993; Raman et al., 1995; Das et al., 1996). However, the chaperone function of sHsps is not only observed in connection with heat shock situations. Members of the sHsp family were discovered to also increase reactivation yields of chemically denatured protein (Horwitz, 1992; Jakob et al., 1993; Merck et al., 1993), to prevent aggregation of mellitin and insulin at room temperature and to protect cellular enzymes from modification-induced inactivation (Ganea and Harding, 1995; 1996). In contrast to other chaperones such as Hsp90 or Hsp70 (Jakob et al., 1995; Stege et al., 1994), sHsps do not alter the inactivation kinetics of substrate proteins (Horwitz, 1992; Rao et al., 1993; Ehrnsperger et al., 1997). These findings are in good agreement with in vivo data showing that overexpression of Hsp27 in rat cell lines does not alter the inactivation of model substrates but recovery is strongly accelerated in the presence of the sHsp (Kampinga et al., 1995; Stege et al., 1995). In Arabidopsis cell culture overexpression of the Hsp17–6 class I protein lead to maintainence of firefly luciferase during heat shock and to increased recovery after stress (Forreiter et al., 1997). The chaperone activity of sHsps has never been found to be associated with ATP-binding or hydrolysis (Horwitz, 1992; Jakob et al., 1993; Merck et al., 1993).
Molecular chaperones and folding catalysts
700
3.2. sHsps Chaperone Protein Folding in Cooperation with Hsp70 Until recently, only stable, seemingly non-productive complexes between substrate molecules and sHsps have been described under heat shock and other stress conditions (Horwitz, 1992; Rao et al., 1993; Lee et al., 1995; Das et al., 1996) leading to the conclusion that protein bound to sHsps may be destined for degradation (Lee et al., 1995). However, such a mechanism would imply a substantial loss of proteins after stress situations. To address this question, the interaction of murine Hsp25 with thermally unfolding proteins was investigated under heat shock conditions in vitro. Hsp25 has been shown to form complexes with unfolding intermediates of citrate synthase (CS) (Ehrnsperger et al., 1997). However, it does not alter the inactivation kinetics of CS, indicating that a stable complex is formed with the denatured enzyme. Even after returning to physiological temperatures the nonnative protein stays bound to Hsp25, thus providing a pool of folding competent protein. Similar results were obtained with pea Hsp18.1. Upon regain of permissive temperatures, the model substrate was found to refold in the presence of rabbit reticulocyte lysate or wheat germ extract in the presence of ATP (Lee et al., 1997). Thus, the question arose as to what energy-dependent factor was responsible for this refolding. A likely candidate was Hsp70 as it is an abundant, ATP-dependent, cytosolic chaperone with promiscuous binding properties. Thus upon addition of Hsp70 to the preformed Hsp25: substrate complexes, release and subsequent reactivation of the bound
Figure 2 Model for the Chaperone Function of sHsps under Heat Shock Conditions. The thermal unfolding of proteins leads to the formation of nonnative intermediates that are sensitive to irreversible side reactions such as aggregation. Hsp25 binds tightly to such intermediates and thus prevents aggregation and insolubilization. Even when the temperature is lowered to permissive conditions, only a small amount of bound substrate is released and folds to the native state. However, when Hsp70 is present during the recovery phase, the nonnative protein is released from Hsp25 and reactivation occurs in
Mechanisms of ATP-independent vs. ATP-dependent chaperones
701
an ATP-dependent manner. The amount of refolding observed in the presence of Hsp70 does not account for all the substrate that was initially bound to Hsp25. Thus, it is likely that in the cell, co-factors of Hsp70 such as Hdj-1, may be necessary for the optimal function of Hsp70. N: native state; I: nonnative intermediate.
nonnative protein could be triggered in the presence of ATP (Ehrnsperger et al., 1997). Taken together, the data presented suggest that sHsps act as general traps for unfolding proteins preventing them from irreversible side reactions such as aggregation and holding them in a folding competent state until permissive conditions for refolding are restored and other, more active, ATP-dependent chaperones can complete the folding process. The combination of elevated temperatures and the resulting energy limitation (Patriarca & Moresca, 1990) may hinder the refolding of nonnative protein under heat stress conditions. In this context, sHsps may provide a pool of protected protein for refolding by ATP-dependent chaperones, after returning to permissive conditions (Figure 2). 4. HEAT SHOCK PROTEIN 90 Hsp90 is one of the most prominent and conserved heat shock protein of eukaryotic cells (Jakob & Buchner, 1994; Buchner, 1996). It has been shown to regulate the activity of proteins such as kinases and steroid receptors that participate in signal transduction pathways (Brugge, 1986; Pratt, 1993; Xu & Lindquist, 1993; Cutforth & Rubin, 1994; Aligue et al., 1994; Smith, 1995; Pratt and Toft, 1997). Furthermore, in vitro studies highlight a general role for Hsp90 in chaperoning protein folding and unfolding reactions (Wiech et al., 1992; Jakob et al., 1995; Freeman & Morimoto, 1996). Hsp90 performs at least part of its function in complex with a specific set of partner proteins which include immunophilins, p23 and Hop/p60. This complex has been suggested to act as a “superchaperone” complex within the cytosol of eukaryotic cells (Buchner, 1996; Bose et al., 1996). 4.1. Molecular Properties of Hsp90 Hsp90 is a highly expressed heat shock protein, even at physiological conditions, in the cytosol of eukaryotic cells (Borkovich et al., 1989). Homologues have been identified in the ER of higher eukaryotes (Grp94/endoplasmin/ERp99/Hsp108/ gp96) (Koch et al., 1986; Mazzarella Se Green, 1987; Sargan et al., 1986; Li & Srivastava, 1993) and in prokaryotes (HtpG) (Bardwell & Craig, 1987). Hsp90 seems to be a dimer in vivo, and in higher eukaryotes, the cytosolic protein is known to exist in two homologous isoforms, and , which are mainly present as - and - homodimers (Minami et al., 1991; Nemoto et al., 1995; Meng et al., 1996). Deletion mutants of Hsp90 have identified amino acids in the C-terminal region to be important for dimerization (Minami et al., 1994; Meng et al., 1996). Conflicting evidence exists concerning the importance of dimerization on the function of Hsp90. Deletions that disrupt dimer formation on the one hand result in the inability to rescue an Hsp90 deficient yeast strain (Minami et al., 1994; Meng et al., 1996).
Molecular chaperones and folding catalysts
702
The specific functions of Hsp90 at physiological conditions remain enigmatic. In yeast, Hsp90 is an essential protein at all temperatures (Parsell & Lindquist, 1994), while it is not essential in prokaryotes (Bardwell & Craig, 1988). Large quantities of Hsp90 are required for growth in yeast at increased temperatures; however, in contrast to some mammalian cells (Bansal et al., 1991), it does not convey thermotolerance (Borkovich et al., 1989; Kimura et al., 1994; Parsell & Lindquist, 1994) although it is able to protect certain proteins against thermal unfolding in vivo (Nathan et al., 1997). While Hsp90 is an abundant cytosolic protein in eukaryotic cells, there seems to exist only a limited, but growing set of examples of “substrate proteins” in the literature (Table 1). Under physiological conditions, Hsp90 has been found in association with cytosolic and nuclear proteins involved in cell signalling, including steroid hormone receptors and some kinases (see Table 1). While little is known about the interaction of Hsp90 with the cytoskeletal proteins and calmodulin, substantial progress has been made regarding the association with steroid receptors and kinases (for reviews, see Pratt, 1993; Bohem & Yamamoto, 1994; Smith, 1995; Pratt and Toft, 1997). Steroid hormone receptors are soluble intracellular proteins that act as transcription factors shuttling between the cytosol and nucleus. Receptors, in their hormone-free state are predominantly associated with Hsp90 and a number of other proteins (see below). The emerging theme concerning the function of the Hsp90 Table 1 Overview of Hsp90 substrate proteins
Protein
Reference
in vivo steroid hormone receptors (including progesterone, estrogen, androgen, glucocorticoid and mineralocorticoid receptors)
Jaob et al., 1984; Catelli et al., 1985; Sanchez et al., 1985; Pratt, 1993; Bohen &Yamamoto, 1994; Smith, 1995
tyrosine kinases (including pp60src, wee 1, sevenless receptor, and cellular src-family kinases)
Opermann et al., 1981; Brugge et al., 1986; Lindquist & Craig, 1988; Xu & Lindquist, 1993; Aligue et al., 1994; Cutforth & Rubin, 1994; Hartson & Matts, 1994
serine-threonine kinases (including eIF-2α, casein kinase II, raf- Rose et al., 1987; Matts & 1 and p50) Hurst, 1989; Miyata & Yahara, 1992; Stancato et al., 1993; Schulte et al., 1995 reverse transcriptase
Hu & Seeger, 1996
Mechanisms of ATP-independent vs. ATP-dependent chaperones
703
tumor necrosis factor
Song et al., 1995
p53
Blagosklonny et al., 1995; Blagosklonny et al., 1996
tubulin
Sanchez et al., 1988
in vitro actin
Koyasu et al., 1986
MyoD
Shaknovich et al., 1992
casein kinase II
Miyata & Yahara, 1992
calmodulin
Minami et al., 1993
luciferase
Schumacher et al., 1994
dihydrofolate reductase (DHFR)
Yonehara et al., 1996
-galactosidase
Freeman & Morimoto, 1996
citrate synthase
Wiech et al., 1992; Jakob et al., 1995
Insulin
Scheibel et al., 1998
complex seems to be that a multi-membered chaperone machinery allows to keep aporeceptors in an inactive yet activatable state. The three dimensional structure of the ligand binding domain of a steroid hormone receptor in the absence of the steroid hormone (Renaud et al., 1995) suggests that the ligand is part of the hydrophobic core of the protein. We propose that the function of Hsp90 is to ensure an opened conformation maintaining the hydrophobic cavity required for hormone binding—a task requiring the fine-tuned interplay of a number of chaperone components. The involvement of Hsp90 complexes has been analysed in most detail in the case of the progesterone aporeceptor (PR) (Smith et al., 1993; Pratt & Welsch, 1994; Johnson and Toft, 1994; cf. Smith and Toft, 1993; Smith, 1995) utilising a cell free lysate system that allows in vitro assembly of PR. Soon after their synthesis, aporeceptors become complexed with Hsp90 and its partner proteins (Pratt, 1993; Bohen and Yamamoto, 1994). It has been proposed that Hsp90 and PR constantly associate and dissociate, thus holding the receptors in an activatable state competent for ligand binding (Smith et al., 1993; 1995; Smith, 1995). Ligand binding would lead to a conformational change in the receptor, thus preventing Hsp90 from rebinding and simultaneously allowing receptor dimerization (Smith et al., 1993; 1995; Smith, 1995). Although no specific binding motif has been identified on receptors mutational analysis of Hsp90 suggests that the carboxy terminal part of the protein is important in this context (Sullivan & Toft, 1993; Cadepond et al., 1993; Bohen & Yamamoto, 1993). Similar studies with protein kinases indicate that newly synthesized kinases are bound to Hsp90 (and partners) until the protein becomes attached to the plasma membrane following myristoylation (Opermann et al., 1981; Brugge, 1986; Lindquist and Craig, 1988; Xu & Lindquist, 1993; Aligue et al., 1994; Cutforth & Rubin, 1994; Hartson & Matts, 1994). While these in vivo experiments suggest that Hsp90 is a specific chaperone, regulating the structure and activity of
Molecular chaperones and folding catalysts
704
proteins involved in signal transduction (Xu & Lindquist, 1993), in vitro experiments demonstrate general chaperone properties of Hsp90 in protein folding under physiological and heat shock conditions (cf. Jakob & Buchner, 1994; Freeman & Morimoto, 1996; Buchner, 1996). 4.2. Mechanism of Hsp90 Chaperone Action Studies on the renaturation of completely unfolded proteins reveal that stoichiometric amounts of Hsp90 suppress unspecific side reactions (aggregation) thereby promoting functional refolding to the native state (Wiech et al., 1992). Similar data were obtained with MyoD, a helix-loop-helix transcription factor. Hsp90 was found to interact transiently with nonnative MyoD resulting in increased DNA-binding activity (Shaknovich et al., 1992). Furthermore, evidence for the chaperone role of Hsp90 came from the in vitro studies on casein kinase II, a physiological substrate of Hsp90. Hsp90 was shown to protect casein kinase II from unfolding and subsequent aggregation (Miyata & Yahara, 1992). More recently, Freeman and Morimoto (1996) have examined the role of Hsp90, in conjunction with Hsp70 and its cofactors on the fate of chemically denatured ß-galactosidase. The data suggest that Hsp90 can maintain non-native -galactosidase in a folding-competent state, independent of ATP. The mechanism of Hsp90 chaperone action has been further characterized utilising an in vitro thermal unfolding assay (Jakob et al., 1995). The assay allows the quantitative analysis of the number of intermediates formed in the unfolding pathway and the distribution between native and aggregated forms of citrate synthase (CS), the model substrate. The data indicate that Hsp90 is able to stabilize unfolding proteins which are still structured and dimeric by transient interactions, in an ATP-independent manner. Conflicting evidence existed concerning the ATP-hydrolysis and ATP-binding of Hsp90 (for review, see Jakob & Buchner, 1994). Binding of ATP to Hsp90 had been reported (Csermely & Kahn, 1991). However, tight binding could be excluded and a potential hydrolytic activity was in the range of background values (Jakob et al., 1996). Furthermore, Hsp90 has been shown to chaperone in vitro protein folding reactions independent of ATP (Wiech et al., 1992; Shaknovich et al., 1992; Shue & Kohtz, 1994; Schumacher et al., 1994; Jakob et al., 1995; Miyata & Yahara, 1995; Freeman & Morimoto, 1996). Recently, however, it could be demonstrated by ESR spectrometry that Hsp90 binds spin-labeled ATP weakly with a Kd around 400 M (Scheibel et al., 1997). This implies that under cellular conditions the major part of Hsp90 is saturated with ATP (Scheibel et al., 1997). In addition, the crystal structures of two N-terminal fragments of Hsp90 have been solved (Prodromou et al., 1997a, Stebbins et al., 1997) and an ATP binding site has been identified (Prodromou et al., 1997b; Stebbins et al., 1997). Interestingly, this site is also the binding site for the anti-tumor drug Geldanamycin, which is a specific inhibitor of Hsp90 function in vivo (De Boer et al., 1970; Prodromau et al., 1997b). The functional consequences of ATP binding to Hsp90 are still under intensive investigation. Recently, in vitro experiments have shown that Hsp90 possesses two chaperone sites located in Nand C-terminal domains (Young et al., 1997; Scheibel et al., 1998). The C-terminal site reflects the promiscuous, ATP-independent chaperone activity previously detected for wild type protein. The N-terminal domain contains a peptide-binding site, which is the
Mechanisms of ATP-independent vs. ATP-dependent chaperones
705
target of ATP and geldanamycin (Scheibel et al., 1998). Nine components of the steroid receptor-chaperone complex have been identified to date, including Hsp90, Hsp70, p60/Hop, p48/Hip, Hdj1, p23 and one of three large immunophilins: FKBP52 (p59, Hsp56, HBI), FKBP51 (FKBP54) or cyclophilin 40 (Cyp40) (Pratt & Welsch, 1994; Smith et al., 1995; Smith, 1995). Most of these proteins have been initially identified in mammals in complexes with steroid receptors (as described above). The initial complex that binds to the PR contains Hsp90, Hsp70, Hip and Hop. These early complexes are displaced by a mature complex that consists of Hsp90, p23 and one of three large immunophilins (Smith et al., 1995; Smith, 1995; Chen et al., 1996). These receptor and chaperone complexes dissociate spontaneously and its components re-enter the assembly pathway, thus allowing a dynamic turnover of receptor complexes. However, it seems that the complexes have more general functions since they are conserved throughout eukaryotes (Chang & Lindquist, 1994). Recently, the functional properties of three of the Hsp90 partner proteins, Hop/ p60, p23 and one of the large immunophilins, FKBP52 or Cyp40 have been analyzed (Bose et al., 1996; Freeman et al., 1996). Using a thermally denatured substrate (CS) (Bose et al., 1996) and a chemically denatured substrate ( -galactosidase) (Freeman et al., 1996) it was demonstrated that three of the Hsp90 partner proteins, p23, FKBP52 and Cyp40, have chaperone activity. In contrast, Hop did not exhibit properties of a molecular chaperone, which correlates with the observation that Hop’s function is primarily to mediate the association of Hsp90 and Hsp70 during the maturation of steroid receptors and kinases (Smith et al., 1995; Chen et al., 1996). Fundamental differences in the mechanism of chaperone action were found
Molecular chaperones and folding catalysts
706
Figure 3 Model for the Chaperone Function of the Hsp90 Complex. Two major complexes containing Hsp90 are involved in protein folding in the cell, as has been demonstrated by Smith and co-workers (Smith et al., 1995). These complexes seem to be able to interact with partially folded proteins formed during de novo protein folding or during unfolding under stress conditions. Both complexes contain several components capable of interacting productively with partially folded proteins. The current results integrate p23, Cyp40 and FKBP52 into this category. In contrast, Hop/p60 seems to be important in mediating the association of Hsp70 and Hsp90 (Smith et al., 1995). Irreversible aggregation is effectively suppressed by association with the Hsp90 chaperone complexes. N: native protein; I: folding intermediate; U: unfolded protein.
to exist between p23 and FKBP52, suggesting that the two proteins have unique roles in assisting protein folding. While FKBP52 was found in stable association with the nonnative protein, p23 was found to interact transiently with the nonnative substrate. Interestingly, both p23 and FKBP52 or Cyp40 were able to chaperone protein folding and unfolding reactions, independent of ATP. Previously, immunophilins have only been shown to be catalysts of prolyl isomerisation (Schmid, 1993; Fischer, 1994; see chapter on prolyl isomerases by Fischer & Schmid). The recent finding that the large immunophilins possess chaperone functions suggests that, in general, they may have
Mechanisms of ATP-independent vs. ATP-dependent chaperones
707
acquired additional domains which are responsible for their chaperone action (Bose et al., 1996). While the prolyl isomerase activity of immunophilins had already suggested their participation in assisted folding processes, the function of p23 had remained completely enigmatic until recently (Pratt, 1993; Smith, 1995). The demonstration of chaperone function of these proteins suggests the existence of a super-chaperone complex in the cytosol of eucaryotic cells in which several functionally distinct chaperone components are capable of interacting productively and selectively with nonnative proteins. One can envisage a chaperone assembly line whereby each of the partner proteins, in conjunction with Hsp90, is able to assist and maintain the unfolded protein in slightly differing manner, until the final conformation is achieved (Figure 3). 5. ATP FUNCTION IN CHAPERONE ACTION Here, we have given examples of chaperones that operate (at least to some extent) independent of ATP. In this context, the general question whether ATP hydrolysis is an essential prerequisite for efficiently chaperoning protein folding arises. The large number of chaperones, whose function does not require ATP argues strongly against this. As described above, the functional mechanism seems to vary between different members of these ATP-independent chaperones; nevertheless, they all have the ability to recognize nonnative polypeptides and the capacity to bind them in the absence of ATP. To understand the role of ATP independent chaperones in protein folding, it is important to bear in mind how ATP affects the mechanism of ATP dependent chaperones such as GroE and Hsp70/DnaK. One has to differentiate between binding and release of nonnative proteins in the presence and absence of ATP. For the DnaK and GroE chaperone systems, it has been demonstrated that binding and release of nonnative proteins occur both in the presence and absence of ATP (Schmid et al., 1994; Sparrer et al., 1996). However, in the presence of ATP the overall release reaction seems to be favoured, implying that the binding and hydrolysis of ATP induces a cycle of conformational changes that influences the interaction with nonnative proteins significantly. In the case of GroE, it is well established that ATP and the co-chaperone GroES are required for the efficient release of nonnative proteins from GroEL under nonpermissive conditions. Recently, the necessity of ATP-induced conformational changes for effective protein release has been elegantly demonstrated with a GroE mutant which was artificially trapped in one conformational state (Murai et al., 1996). These conformational changes have been most convincingly visualized for the GroE system using cryoelectron microscopy (Roseman et al., 1996) (see chapter Burston and Saibil, this volume). In order to determine the microscopic rate constants of binding and release, a reversibly folding protein has to be used in the presence of a competitor to prevent rebinding. Such a system has been established for GroE using maltose binding protein (MBP) and ribose binding protein (RBP) as competing nonnative substrates. The data derived from this experimental system suggest that ATP binding is sufficient to change the binding characteristics of the chaperonin so that efficient release from GroEL is achieved (Sparrer et al., 1996) (see chapter by Ranson and Clarke, this volume). Interestingly, an analysis of the partial reaction revealed that almost exclusively the on-rate is affected by ATP
Molecular chaperones and folding catalysts
708
binding while the off-rate is not altered significantly in the presence of ATP. This implies that the nonnative protein is not expelled from GroEL but rather that its rebinding is prevented. For DnaK, the E.coli homologue of Hsp70, the influence of ATP was investigated by stopped-flow spectrofluorimetry using fluorophore-labeled peptide ligands (Schmid et al., 1994) which allowed binding to be measured in real time. The data suggest that the binding-release cycle of DnaK is too fast to be stoichiometrically coupled with the hydrolysis of ATP. Here, release of protein is achieved mainly by increasing the off-rate in the presence of ATP (Schmid et al., 1994; Theyssen et al., 1996) (see chapter by Buchberger et al., this volume). Thus, different mechanisms seem to be used by the two ATP-dependent chaperone systems to achieve release of bound protein. While release for ATP-dependent chaperones can, in some instances, be achieved in the absence of ATP, efficient folding under nonpermissive folding conditions seems to necessitate cycles of conformational changes which are driven by ATP hydrolysis. In the case of GroE, this allows to unfold kinetically trapped protein (Todd et al., 1994). The growing body of results on the chaperone function of an increasing number of proteins shows that ATP-independent interactions with nonnative proteins are important. The dissection of the functional mechanism of different classes of chaperones has demonstrated that in all cases recognition and binding of nonnative proteins occurs independent of ATP. Thus recognition seems to depend on the specific positioning of “sticky” interactive surfaces on the respective chaperone protein. The arrangement of these surfaces determines the “specificity” of the chaperone. For example, the binding site of Hsp70 is constructed to accomodate an extended peptide segment of an unfolded protein (Zhu et al., 1996) and the inner surface of the GroEL cavity is lined with a sevenfold repeated pattern of hydrophobic residues (Braig et al., 1994) well suited for multiple interactions with proteins at different stages of folding. For these two ATPdependent chaperones the energy of ATP hydrolysis is used to perform a cycle of conformational changes which, intermittently, changes the three-dimensional structure of the binding site so that efficient release of the substrate protein occurs and finally to complete the cycle by re-establishing the initial binding conformation. In the case of the GroE complex an additional level of complexity is introduced by the interaction of the cofactor GroES with GroEL which allows to perform folding under otherwise nonpermissive folding conditions (Schmidt et al., 1994). In the case of ATP-independent chaperones these carefully controlled modulations of the release reactions are of course lacking. Therefore, one wonders what the beneficial effects of these chaperones except for a mere “buffering effect” for aggregation-prone intermediates may be. The idea of a “maintenance” activity of Hsp90 introduced by Freeman & Morimoto (1996) expands this scheme. Here Hsp90 is thought to bind unfolded protein until other ATP-dependent chaperones such as Hsp70 take over and allow efficient refolding. Recently, however, with the identification of weak ATP binding to Hsp90 (Scheibel et al., 1997) and the identification of a defined ATP-binding site (Prodromou et al., 1997b), the situation has changed. At present it seems that Hsp90 combines both ATP-dependent and ATP-independent chaperone properties (Scheibel and Buchner, unpublished). In the cell, the situation seems to be even more complex, since some of the Hsp90
Mechanisms of ATP-independent vs. ATP-dependent chaperones
709
partner proteins exhibit chaperone activities themselves (Bose et al., 1996; Freeman et al., 1996). It is hard to imagine that for a rather passive “holding” function several functionally distinct chaperones organized in a defined machinery may be required. Thus, at present it seems that for Hsp90, we have just caught a first glimpse of one of the most complex folding machineries of the cell. Another emerging theme is the concept of chaperone networks, impyling that chaperones do not act on their own or in a strict sequential and hierarchic order, but that folding intermediates are released and rebound by any chaperone around at a given time. Only under stress conditions, where the concerted action of several chaperones are required to guarantee survival, the function of individual chaperones may become apparent. In this context, the chaperone repertoire becomes flooded with nonnative aggregation-prone client proteins requiring specific job sharing, especially given that the energy levels in these cells may be depleted. Here, ATP-independent chaperones, and especially small Hsps, could be the most important players in promoting fast and efficient trapping of unfolded proteins. Separated in time, the Hsp70 machinery could, upon restoration of physiological conditions, work on the refolding of these proteins stored in an intermediate conformation. In this case, holding and folding may be really separately performed by two synergistically interacting chaperone machineries. Thus, ATP-independent chaperone functions seem to have evolved either to assure rapid binding and trapping of unfolding proteins, as has been proposed for sHsps, with subsequent folding steps being performed in synergistic cooperation with an ATPdependent chaperone system. Or, the ATP-independent chaperones are part of a multichaperone complex in which different modes of interactions with nonnative proteins may be sufficient to allow folding. However, while considerable progress has been achieved in dissecting the functional cycles of GroE and the Hsp70 machinery, the mechanisms of ATP-independent chaperones are just beginning to be understood—and new members of this family are being identified. At present we may just be seeing the tip of the iceberg concerning the number, function and mechanism of energy-independent folding helpers. 6. ACKNOWLEDGEMENTS We thank members of our lab for ongoing support. Special thanks to Christian Mayr for help with illustrations and to Tina Weikl, Hans Bügl and Thomas Scheibel for their scientific contribution. 7. REFERENCES Aligue, R., Akhavan-Niak, H. and Russell, P. (1994). A Role for Hsp90 in Cell Cycle Control: Weel Tyrosine Kinase Activity Requires Interaction with Hsp90. EMBO J. , 13 , 6099–6106. Allen, S.P., Polazzi, J.O., Gierse, J.K. and Easton, A.M. (1992). Two Novel Heat Shock Genes Encoding Proteins Produced in Response to Heterologous Protein Expression in Escherichia coli. J. Bacterial. , 174 , 6938–6947. Anderson, R.G.W. and Orci, L. (1988). A View of Acidic Intracellular Compartments. J.
Molecular chaperones and folding catalysts
710
Cell Biol. , 106 , 539–543. Andres, D.F., Dickerson, I.M. and DixonJ.E. (1990). Variants of the Carboxyl-terminal KDEL Sequence Direct Intracellular Retention. J. Biol Chem. , 265 , 5952–5955. Aoyama, A., Froehli, E., Schaefer, R. and Klemenz, R. (1993). B-crystallin Expression in Mouse NIH 3T3 Fibroblasts: Glucocorticoid Responsiveness and Involvement in Thermal Protection. Mol. Cell. Biol , 13 , 1824–1835. Arrigo, A.-P. and Landry, J. (1994). Expression and Function of the Low-molecularweight Heat Shock Proteins. In Morimoto, R.I., Tissiäres, A and Georgopoulos, C. (eds.), The Biology of Heat shock Proteins and Molecular Chaperones , Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Arrigo, A.-P., Suhan, J. and Welch, W.J. (1988). Dynamic Changes in the Structure and Intracellular Locale of the Mammalian Low-molecular-weight Heat Shock Protein. Mol. Cell. Biol , 8 , 5059–5071. Bansal, G.S., Norton, P.M. and Latchman, D.S. (1991). The 90kDa Heat Shock Protein Protects Mammalian Cells From Thermal Stress but not From Viral Infection. Exp. Cell Res. , 195 , 303–306. Bardwell, J.C.A. and Craig, E.A. (1987). Eukaryotic Mr 83,000 Heat Shock Protein has a Homologue in Escherichia coli. Proc. Natl Acad. Set. , 84 , 5177–5181. Bardwell, J.C.A. and Craig, E.A. (1988). Ancient Heat Shock Protein is Dispensible. J. Bacterial , 170 , 2977–2983. Behlke, J., Lutsch, G., Gaestel, M. and Bielka, H. (1991). Supramolecular Structure of the Recombinant Murine Small Heat Shock Protein Hsp25. FEBS-Lett. , 288 , 119– 122. Benndorf, R., Hayess, K., Ryazantsev, S., Wieske, M., Behlke, J. and Lutsch, G. (1994). Phosphorylation and Supermolecular Organization of Murine Small Heat Shock Protein Hsp25 Abolish its Actin Polymerization-Inhibiting Activity. J. Biol. Chem. , 269 , 230780–20784. Bergeron, J.J., Brenner, M.B., Thomas, D.Y. and Williams, D.B. (1994). Calnexin: A Membrane-Bound Chaperone of the Endoplasmic Reticulum. Trends Biochem. Set. , 19 , 124–128. Bhat, S.P. and Nagineni, C.N. (1989). B Subunit of Lens-specific Protein -crystallin is Present in Other Ocular and Non-ocular Tissues. Biochem. Biophys. Res. Commun. , 158 , 319–325. Blagosklonny, M.V., Toretsky, J. and Neckers, L. (1995). Geldanamycin Selectively Destabilizes and Conformationally Alters Mutated p53. Oncogene , 11 , 933–939. Blagosklonny, M.V., Toretsky, J., Bohen, S. and Neckers, L. (1996). Mutant Conformation of p53 Translated in vitro or in vivo Requires Functional Hsp90. Proc. Natl. Acad. Sci. , 93 , 8379–8383. Bohen, S.P. and Yamamoto, K.R. (1993). Isolation of Hsp90 Mutants by Screening for Decreased Steroid Receptor Function. Proc. Natl. Acad. Set. , 90 , 11424–11428. Bohen, S.P. and Yamamoto, K.R. (1994). Modulation of Steroid Receptor Signal Transduction by Heat Shock Proteins. In Morimoto, R.I., Tissiäres, A. and Georgopoulos, C. (eds.), The Biology of Heat shock Proteins and Molecular Chaperones , Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Bond, U. and Schlesinger, M.J. (1987). Heat-shock Proteins and Development. Adv.
Mechanisms of ATP-independent vs. ATP-dependent chaperones
711
Genet. , 24 , 1–29. Bonifacino, J.S., Lippincott-Schwartz, J., Chen, C., Antusch, D., Samelson, L.E. and Klausner, R.D. (1988). Association and Dissociation of the Murine T Cell Receptor Associated Protein (TRAP). Early Events in the Biosynthesis of a Multisubunit Receptor. J. Biol. Chem. , 263 , 8965–8971. Borkovich, K.A., Farrelly, F.W., Finkelstein, D.B., Taulien, J. and Lindquist, S. (1989). Hsp82 is an Essential Protein that is Required in Higher Concentrations for Growth of Cells at Higher Temperatures. Mol. Cell. Biol. , 9 , 3919–3930. Bose, S., Weikl, T., Bügl, H. and Buchner, J. (1996). Chaperone Function of Hsp90Associated Proteins. Science , 274 , 1715–1717. Braig, K, Otwinowski, Z., Hegde, R., Boisvert, D.C., Joachimiak, A., Horwich, A.L. and Sigler, P.B. (1994). The Crystal, Structure of the Bacterial Chaperonin GroEL at 2.8è. Nature , 371 , 578–586. Brugge, J.S. (1986). Interaction of the Rous Sarcoma Virus Protein pp60src with the Cellular Proteins pp50 and pp90. Curr.Top. Microbiol. Immunol , 123 , 1–22. Buchner, J. (1996). Supervising the Fold: Functional Principles of Molecular Chaperones. FASEB J. , 10 , 10–19. Cadepond, F., Binart, N., Chambraud, B., Jibard, N., Schweizer-Goyer, G., SegardMaurel, I. and Baulieu, E.E. (1993). Interaction of Glucocorticoid Receptor and WildType or Mutated 90-kDa Heat Shock Protein Coexpressed in Baculovirus-infected Sf9 Cells. Proc. Natl Acad. Sci. , 90 , 10434–10438. Carver, J.A., Aquilina, J.A., Truscott, R.J.W. and Ralston, G.B. (1992). Identification by 1H NMR Spectroscopy of Flexible C-terminal Extensions in Bovine Lens -crystallin. FEBS Lett. , 311 , 143–149. Carver, J.A., Esposito, G., Schwedersky, G. and Gaestel, M. (1995). 1H NMR Spectroscopy Reveals that Mouse Hsp25 has a Flexible C-terminal Extension of 18 Amino Acids. FEBS Lett. , 369 , 305–310. Catelli, M.-G., Binart, N., Jung Testas, I., Renoir, J.M., Baulieu, E.-E. and Feramisco, J.R. (1985). The Common 90-kD Protein Component of Non-transformed, 8S’ Steroid Receptors is a Heat Shock Protein. EMBO J. , 4 , 3131–3135. Chang, H.-C.J. and Lindquist, S. (1994). Conservation of Hsp90 Macromolecular Complexes in Saccharomyces cerevisiae. J. Biol. Chem. , 269 , 24983–24988. Chang, Z., Primm, T.P., Jakana, J., Lee, I.H., Serysheva, I., Chiu, W., Gilbert, H.F. and Quiocho, F.A. (1996). Mycobacterium tuberculosis 16-kDA Antigen (Hsp 16.3). Functions as an Oligomeric Structure in vitro to Suppress Thermal Aggregation. J. Biol. Chem. , 271 , 7218–7223. Chen, S., Prapapanich, V., Rimerman, R.A., HonorÇ, B. and Smith, D.F. (1996). Interactions of p60, a Mediator of Progesterone Receptor Assembly, with Heat Shock Proteins Hsp90 and HSp70. Mol. Endocrinol , 10 , 682–693. Chiesa, R., McDermott, M.J. and Spector, A. (1990). The Apparent Molecular Size of Native -crystallin B in Non-lenticular Tissues. FEBS-Lett. , 268 , 222–226. Chuang, S.E., Burland, V., Plunkett, G., Daniels, D.L. and Blattner, F.R. (1993). Sequence Analysis of Four New Heat-Shock Genes Constituting the hsITS/ibpAB and hsIVU operons in Escherichia coli. Gene , 134 , 1–6. Ciocca, D.R., Oesterreich, S., Chamness, G.C., McGuire, W.L. and Fuqua, S.A. (1993).
Molecular chaperones and folding catalysts
712
Biological and Clinical Implications of Heat Shock Protein 27,000 (Hsp27): A Review. J. Natl. Cancer Inst. , 85 , 1558–1570. Clarke, E.P., Cates, G.A., Ball, E.H. and Sanwal, B.D. (1991). A Collagen Binding Protein in the Endoplasmic Reticulum of Myoblasts Exhibit Relationship with Serine Protease Inhibitors. J. Biol. Chem. , 266 , 17230–17235. Clarke, E.P., Jain, N., Brickenden, A., Lorimer, I.A. and Sanwal, B.D. (1993). Parallel Regulation of Procollagen I and Colligin, a collagen-binding protein and a Member of the Serine Protease Inhibitor Family. J. Cell Biol. , 121 , 193–199. Créte, P. and Landry, J. (1990). Induction of Hsp27 Phosphorylation and Thermoresistance in Chinese Hamster Cells by Arsenite, Cyclohexamide, A23187, and EGTA Radiat. Res. , 121 , 320–327. Cutforth, T. and Rubin, G.M. (1994). Mutations in Hsp83 and cdc37 Impair Signaling by the Sevenless Receptor Tyrosine Kinase in Drosophila. Cell , 77 , 1027–1036. Das, K.P., Petrash, J.M. and Surewicz, W.K. (1996). Conformational Properties of Substrate Proteins Bound to a Molecular Chaperone a-crystallin. J. Biol. Chem. , 271 , 10449–10452. DeBoer, C., Meulman, P.A., Wnuk, R.J. and Peterson, D.H. (1970). Geldanamycin, a new antibiotic. J. Antibiot. (Tokyo) , 23 , 442–447. Ehrnsperger, M., Gräber, S., Gaestel, M. and Buchner, J. (1997). Binding of Nonnative Protein to Hsp25 during Heat Shock Creates a Reservoir of Folding Intermediates for Reactivation. EMBO J., (in press) . El-Thaher, T.S.H., Drake, A.F., Yokota, S., Nakai, A., Nagata, K., and Miller, A.D. (1996). The pH-Dependent, ATP-Independent Interaction of Collagen Specific Serpin/Stress Protein Hsp47. Protein Peptide Letters , 3 , 1–8. Fischer, G. (1994). Peptidyl-Prolyl cis/trans Isomerases and their Effectors. Angew. Chem. Int. Ed. Engl , 33 , 1415–1436. Forreiter, C., Kirschner, M. and Nover, L. (1997). Stable transformation of an Arabidopsis cell suspension culture with firefly luciferase providing a cellular system for analysis of chaperone activity in vivo. Plant Cell , 9 , 2171–2181. Freeman, B.C. and Morimoto, R.I. (1996). The Human Cytosolic Molecular Chaperones Hsp90, Hsp70 (Hsc70). and Hdj-1 have Distinct Roles in Recognition of a Non-Native Protein and Protein Refolding. EMBO J. , 15 , 2969–2979. Freeman, B.C., Toft, D.O. and Morimoto, R.I. (1996). Molecular Chaperone Machines: Protein Folding Activities of the Cyclophilin Cyp-40 and the Steroid Aporeceptor Associated Protein, p23. Science , 274 , 1718–1720. Fuller, K.J., Issels, R.D., Slosman, D.O., Guillet, J.- G., Soussi, T. and Polla, B.S. (1994). Cancer and the Heat Shock Response. Eur. J. Cancer , 30 , 1884–1891. Gaestel, M., Gross, B., Benndorf, R., Strauss, M., Schunk, W.H., Kraft, R., Otto, A., Boehm, H., Stahl, J., Drabsch, H., et al., (1989). Molecular Cloning, Sequencing and Expression in Escherichia coli of the 25 kDa Growth-Related Protein of Ehrlich Ascites Tumor and its Homology to Mammalian Stress Proteins. Eur. J. Biochem. , 179 , 209–213. Ganea, E. and Harding, J.J. (1995). Molecular Chaperones Protect Against Glycationinduced Inactivation of Glucose-6-phosphate Dehydrogenase. Eur. J. Biochem. , 231 , 181–185.
Mechanisms of ATP-independent vs. ATP-dependent chaperones
713
Ganea, E. and Harding, J.J. (1996). Inhibition of 6-Phosphogluconate Dehydrogenase by Carbamylation and Protection by -crystallin, A Chaperone Like Protein. Biochem. Biophys. Res. Comm. , 222 , 626–631. Groenen, P.J.T.A., Merck, K.B., de Jong, W. and Bloemendal, H. (1994). Structure and Modifications of the Junior Chaperone -crystallin: From Lens Transparency to Molecular Pathology. Eur. J. Biochem. , 225 , 1–19. Haas, I.G. (1994). BiP (Grp78), An Essential Hsp70 Resident Protein in the Endoplasmic Reticulum. Experientia , 50 , 1012–1020. Hartson, S.D. and Matts, R.L. (1994). Association of Hsp90 with Cellular Src-Family Kinases in a Cell-Free System Correlates with Altered Kinase Structure and Function. Biochemistry , 33 , 8912–8920. Helenius, A. (1994). How N-linked Oligosaccharides Affect Glycoprotein Folding in the Endoplasmic Reticulum. Mol. Biol. Cell , 5 , 253–265. Hirayoshi, K., Kudo, H., Takechi, H., Nakai, A, Iwamatsu, A., Yamada, K.M. and Nagata, K. (1991). Hsp47: A Tissue-Specific, Transformation-Sensitive, CollagenBinding Heat Shock Protein of Chicken Embryo Fibroblasts. Mol. Cell. Biol. , 11 , 4036–4044. Horwitz, J. (1992). -crystallin Can Function as a Molecular Chaperone. Proc. Natl. Acad. Sci. , 89 , 10449–10453. Hu, J. and Seeger, C. (1996). Hsp90 is Required For the Activity od a Hepatitis B Virus Reverse Transcriptase. Proc. Natl. Acad. Sci. , 93 , 1060–1064. Huot, J., Houle, F., Spitz, D.R. and Landry, J. (1996). Hsp27 Phosphorylation-mediated Resistance Against Actin Fragmentation and Cell Death Induced by Oxidative Stress. Cancer Res. , 56 , 273–279. Inaguma, Y., Shinohara, H., Goto, S. and Kato, K. (1992). Translocation and Induction of B Crystallin by Heat Shock in rat glioma (GA-1). Cells. Biochem. Biophys. Res. Commun. , 182 , 844–850. Ingolia, T.D. and Craig, E.A. (1982). Four Small Drosophila Heat Shock Proteins are Related to each other and to Mammalian -crystallin. Proc. Natl. Acad. Sci. , 79 , 2360–2364. Iwaki, T., Iwaki, A., Tateishi, J., Sakaki, Y. and Goldman, J.E. (1993). B-crystallin and 27kD Heat Shock Protein are Regulated by Stress Conditions in the Central Nervous System and Accumulate in Rosenthal Fibers. Am. J. Pathol , 143 , 487–495. Jackson, G.S., Staniforth, R.A., Halsall, D.J., Atkinson, T., Holbrook, J.J., Clarke, A.R., Burston, S.G. (1993). Binding and Hydrolysis of Nucleotides in the Chaperonin Catalytic Cycle: Implications for the Mechanism of Assisted Protein Folding. Biochemistry , 32 , 2554–2563. Jain, N., Brickenden, A., Lorimer, I., Ball E.H. and Sanwal, B.D. (1994). Interaction of Procollagen I and other collagens with Colligin. Biochem. J. , 304 , 61–68. Jakob, U. and Buchner, J. (1994). Assisting Spontaniety: The Role of Hsp90 and Small Hsps as Molecular Chaperones. Trends. Biochem. Sci. , 19 , 205–211. Jakob, U., Gaestel, M., Engel, K. and Buchner, J. (1993). Small Heat Shock Proteins are Molecular Chaperones. J. Biol. Chem. , 268 , 1517–1520. Jakob, U., Lilie, H., Meyer, I. and Buchner, J. (1995). Transient interaction of Hsp90 with Early Unfolding Intermediates of Citrate Synthase. Implications for Heat Shock In vivo. J. Biol. Chem. , 270 , 7288–7294.
Molecular chaperones and folding catalysts
714
Jakob, U., Scheibel, T., Bose, S., Reinstein, J. and Buchner J. (1996). Assesment of the ATP Binding Properties of Hsp90. J. Biol. Chem. , 271 , 10035–10041. Jaob, I., Radanyi, C., Renoir, M., Buchou, T., Catelli, M.-G., Binart, N. and Mester, J. (1984). Common Non-Hormone Binding Component in Non-transformed Chick Oviduct. Nature , 308 , 850–853. Jinn, T.L., Yeh, Y.C., Chen, Y.M. and Lin, C.Y (1989). Stabilization of Soluble Proteins in vitro by Heat Shock Proteins-enriched Ammonium Sulfate Fraction from Soybean Seedlings. Plant Cell. Physiol , 30 , 463–469. Johnson, J.L. and Toft, D.O. (1994). A Novel Chaperone Complex for Steroid Receptors Involving Heat Shock Protein, Immunophilins and p23. J. Biol. Chem. , 269 , 24989– 24993. Kampinga, H.H., Brunstig, J.F., Stege, G.J.J., Burgman, P.W. and Konings, A.W. (1995). Thermal Protein Denaturation and Protein Aggregation in Cells made Thermotolerant by Various Chemicals: The Role of Heat Shock Proteins. Exp. Cell Res. , 219 , 536– 546. Kantorow, M. and Piatigorsky, J. (1994). -crystallin/Small Heat Shock Protein has Autokinase Activity. Proc. Natl. Acad. Sci. , 91 , 3112–3116. Kato, K., Hasegawa, K., Goto, S. and Inaguma, Y. (1994). Dissociation as a Result of Phosphorylation of an Aggregated Form of the Small Stress Protein Hsp27. J. Biol. Chem. , 269 , 11274–11278. Kim, J.H., Lingwood, C.A., Williams, D.B., Furuya, W., Manolson, M.F. and Grinstein, S. (1996). Dynamic Measurement of the pH of the Golgi Complex in Living Cells Using Retrograde Transport of the Verotoxin Receptor. J. Cell Biol. , 134 , 1387–1399. Kim, Y.J., Shuman, J., Sette, M. and Przybyla, A. (1984). Nuclear Localization and Phosphorylation of Three 25-kilodalton Rat Stress Proteins. Mol. Cell Biol. , 4 , 468– 474. Kimura, Y., Matsumoto, S. and Yahara, I. (1994). Temperature-sensitive Mutants of Hsp82 of the Budding Yeast, Saccharomyces cerevisiae. Mol. Cell. Genet. , 242 , 517– 527. Klemenz, R., Froehli, E., Steiger, R.H., Schaefer, R. and Aoyama, A. (1993). Expression of the Murine Small Heat Shock Protein Hsp25 and B Crystallin in the Absence of Stress. J. Cell Biol , 120 , 639–645. Klemenz, R., Froehli, E., Steiger, R., Schaefer, R. and Aoyama, A. (1991). Bcrystallin is a Small Heat Shock Protein. Proc. Natl. Acad. Sci. , 88 , 3652–3656. Knauf, U., Bielka, H. and Gaestel, M. (1992). Overexpression of the Small Heat Shock Protein Hsp25 Inhibits Growth of Ehrlich Ascites Tumor Cells. FEBS Lett. , 309 , 297– 302. Koch, G., Smith, M., Macer, D., Webster, P. and Mortara, R. (1986). Endoplsamic Reticulum Contains a Common, Abundant Calcium-Binding Glycoprotein, Endoplasmin. J. Cell Sci. , 86 , 217–232. Koyasu, S., Nishida, E., Kadowaki, T., Matsuzaki, F., Iida, K., Harada, F., Kasuga, M., Sakai, H. and Yahara, I. (1986). Two Mammalian Heat Shock Proteins, Hsp90 and Hsp100, are Actin-Binding Proteins. Proc. Natl. Acad. Sci. , 83 , 8054–8058. Kurkinen, M., Taylor, A., Garrels, J.I. and Hogan., B.L. (1984). Cell Surface-Associated Proteins which Bind Native Type IV collgen or gelatin. J. Biol. Chem. , 259 , 5915–
Mechanisms of ATP-independent vs. ATP-dependent chaperones
715
5922. Landry, J., Lambert, H., Zhou, M., Lavoie, J.N., Hickey, E., Weber, L.A. and Anderson, C.W. (1992). Human Hsp27 is Phosphorylated at Serines 78 and 82 by Heat Shock and Mitogen-activated Kinases that Recognize the same Amino Acid Motif as S6 Kinase. J. Biol. Chem. , 267 , 794–803. Landry; J., ChrÇtien, P., Lambert, H., Hickey, E. and Weber, L.A. (1989). Heat Shock Resisitance Confered by Expression of the Human Hsp27 Gene in Rodent Cells. J. Cell Biol. , 109 , 7–15. Langer, T., Lu, C., Echols, H., Flanagan, J., Hayer, M.K. and Hartl, F.U. (1992). Succesive Action of DnaK, DnaJ, and GroEL Along the Pathway of ChaperoneMediated Protein Folding. Nature , 356 , 683–689. Lavoie, J.N., Hickey, E., Weber, L.A. and Landry, J. (1993). Modulation of Actin Microfilament Dynamics and Fluid Phase Pinocytosis by Phosphorylation of Heat Shock Protein 27 . J. Biol. Chem. , 268 , 24210–24214. Lavoie, J.N., Lambert, H., Hickey, E., Weber, L.A. and Landry, J. (1995). Modulation of Cellular Thermoresistance and Actin Filament Stability Accompanies Phosphorylationinduced Changes in the Oligomeric Structure of Heat Shock Protein 27. Mol. Cell Biol. , 15 , 505–516. Lee, G.A., Roseman, A.M., Saibil, H.R. and Vierling, E. (1997). A Small Heat Shock Protein Stably Binds Heat-Denatured Model Substrates and can Maintain A Substrate in a Folding Competent State. EMBO J. , (in press). Lee, J.G., Pokala, N. and Vierling, E. (1995). Structure and In Vitro Molecular Chaperone Activity of Cytosolic Small Heat Shock Proteins from Pea. J. Biol. Chem. , 270 , 10432–10438. Li, Z. and Srivastava, P.K (1993). Tumor Rejection Antigen Gp96/Grp94 is an ATPase: Implications For Protein Folding and Antigen Presentation. EMBO J. , 12 , 3143–3151. Lindquist, S. and Craig, E.A. (1988). The Heat-Shock Proteins. Annu. Rev. Genet. , 22 , 631–677. Lindquist, S., DiDomenico, B.J., Bugaisky, G.E., Kurtz, S. and Petko, L. (1982). Regulation of the Heat Shock Response in Drosophila and Yeast. In Schlessinger, M.J., Asburner, M. and Tissiäres, A., (eds.), Heat Shock from Bacteria to Man , Cold Spring Harbor, NY., 167–176. Loktionova, S.A.Ilyinskaya, O.P., Gabai, V.L. and Kabakov, A.E. (1996). Distinct Effects of Heat Shock and ATP Depletion on Distribution and Isoform Patterns of Human Hsp27 in Endothelial Cells. FEBS Lett. , 392 , 100–104. Masuda, H., Fukumoto, M., Hirayoshi, K. and Nagata, K. (1994). Coexpression of the Collgen-Binding Stress Protein Hsp47 gene and the 1 (I). and 1(III). Collgen Genes in Carbon Tetrachloride-induced Rat Liver Fibrosis. J. Clin. Invest. , 94 , 2481– 2488. Matts, R.L. and Hurst, R. (1989). Evidence for the Association of the Heme-regulated eIF-2 Kinase with the 90-kDa Heat Shock Protein in Rabbit Reticulocyte Lysate in situ. J. Biol. Chem. , 264 , 15542–15547. Mazzarella, R.A. and Green, M. (1987). ERp99, An Abundant, Conserved Glycoprotein of the Endoplasmic Reticulum, is Homologous to the 90kDa Glucose Regulated Protein (Grp94). J. Biol. Chem. , 262 , 8875–8883.
Molecular chaperones and folding catalysts
716
McKay, D.B., Wilbanks, S.M., Flaherty, K.M., Ha, J.-H., O’Brien, M.C. and Shirvanee, L.L. (1994). Stress-70 Proteins and Their Interaction with Nucleotides. In Morimoto, R.I., Tissiäres, A and Georgopoulos, C. (eds.), The Biology of Heat shock Proteins and Molecular Chaperones , Cold Spring Harbor Laboratory, Cold Spring Harbor, NY Mehlen, P. and Arrigo, A.-P. (1994). The Serum Induced Phosphorylation of Mammalian Hsp27 Correlates with Changes in its Intracellular Localization and Levels of Oligomerization. Eur. J. Biochem. , 221 , 327–334. Mehlen, P., Mehlen, A., Guillet, D., Preville, X. and Arrigo, A.P. (1995). Tumor Necrosis Factor-alpha Induces Changes in the Phosphorylation, Cellular Localisation and Oligomerisation of Human Hsp27, A Stress Protein that Confers Cellular Resistance to this Cytokine. J. Cell. Biochem. , 58 , 248–259. Meng, X., Devin, J., Sullivan, W.P., Toft, D.O., Baulieu, E.-E. and Catelli, M.-G. (1996). Mutational Analysis of Hsp90a Dimerization and Subcellular Localization: Dimer Disruption Does Not Impede In vivo Interaction with Estrogen Receptor. J. Cell Sci. , 109 , 1677–1687. Merck, K.B., Groenen, P.J., Voorter, C.E., de Haard-Hoekman, W.A., Horwitz, J., Bloemendal, H. and de Jong, W.W. (1993). Structural and Functional Similarities of Bovine -crystallin and Mouse Small Heat-shock Protein. A Family of Chaperones. J. Biol. Chem. , 268 , 1046–1052. Minami, Y., Kawasaki, H., Miyata, Y., Suzuki, K. and Yahara, I. (1991). Analysis of Native Forms and Isoform Compositions of the Mouse 90-kDa Heat Shock Protein, Hsp90. J. Biol. Chem. , 266 , 10099–10103. Minami, Y., Kawasaki, H., Suzuki, K. and Yahara, I. (1993). The Calmodulin-binding Domain of Mouse 90-kDa Heat Shock Protein. J. Biol. Chem. , 268 , 9604–9610. Minami, Y., Kinura, Y., Kawasaki, H., Suzuki, K. and Yahara, I. (1994). The Carboxyterminal Region of Mammalian Hsp90 is Required for its Dimerization and Function in vivo. Mol. Cell. Biol. , 14 , 1459–1464. Miron, T., Vancompernolle, J., Vandekerckhove, J., Wilchek, M. and Geiger, B. (1991). A 25-kD Inhibitor of Actin Polymerization is a Low Molecular Mass Heat Shock Protein. J. Cell Biol. , 114 , 255–261. Miron, T., wichek, M. and Geiger, B. (1988). Characterization of an Inhibitor of Actin Polymerization in Vinculin-rich Fraction of Turkey Gizzard Smoothe Muscle. Eur. J. Biochem. , 178 , 543–553. Miyata, Y. and Yahara, I. (1992). The 90-kDa Heat Shock Protein, Hsp90, Binds and Protects Casein Kinase II from Self-aggregation and Enhances its Kinase activity. J. Biol. Chem. , 267 , 7042–7047. Miyata, Y. and Yahara, I. (1995). Interaction Between Casein Kinase II and the 90-kDa Stress Protein, Hsp90. Biochemistry , 34 , 8123–8129. Morimoto, R.I., Tissiäres, A and Georgopoulos, C. (1994). The Biology of Heat Shock Proteins and Molecular Chaperones , Cold Spring Harbor Laboratory, Cold Spring Harbor, NY Murai, N., Makino, Y. and Yoshida, M. (1996). GroEL Locked in a ‘Closed’ Conformation by an Interdomain Cross Link can Bind ATP and Polypeptide, but Cannot Process Further Reaction Steps. J. Biol. Chem. , (in press). Nagata, K. (1996). Hsp47: A Collagen-specific Molecular Chaperone. Trends Biochem.
Mechanisms of ATP-independent vs. ATP-dependent chaperones
717
Sci. , 21 , 23–26. Nagata, K. and Yamada, K.M. (1986). Phosphorylation and Transformation Sensitivity of a Major Collagen-binding Protein of Fibroblasts. J. Biol. Chem. , 261 , 7531–7536. Nagata, K., Nakai, A., Hosokawa, N., Kudo, H., Takechi, H., Satoh, M. and Hirayoshi, K. (1991). Interaction of Hsp47 with Newly Synthesized Procollagen, and Regulation of HSP Expression. In Maresca, B. and Lindquist, S. (eds.), Heat Shock , SpringerVerlag, New York, 105–110. Nagata, K., Saga, S. and Yamada, K.M. (1986). A Major Collagen-binding Protein of Chick Embryo Fibroblasts is a Novel Heat Shock Protein. J. Cell Biol. , 103 , 223–229. Nakai, A., Satoh, M., Hirayoshi, K. and Nagata, K. (1992). Involvement of the Stress Protein Hsp47 in Procollagen Processing in the Endoplasmic Reticulum. J. Cell Biol , 117 , 903–914. Nardai, G., Schnaider, T., Sìti, C., Ryan, M.T., Hoj, P.B.Somogyi, J. and Csermely, P. (1996). Characterization of the 90 kDa Heat Shock Protein-(Hsp90)-Associated ATP/GTP-ase. J. Bioscience , 21 , 179–190. Nathan, D.F., Vos, M.H. and Lindquist, S. (1997). In vivo functions of the Saccharomyces cerevisiae Hsp90 chaperone. Proc. Natl Acad Sci. , 94 , 12949–12956. Natsume, T., Koide, T., Yokota, S.-L, Hirayoshi, K. and Nagata, K. (1994). Interaction Between Collagenbinding Stress Protein Hsp47 and Collagen. J. Biol. Chem. , 269 , 31224–31228. Nemoto, T., Ohara-Nemoto, Y., Ota, M., Takagi, T. and Yokoyama, K. (1995). Mechanism of Dimer Formation of the 90-kDa Heat-Shock Protein. Eur. J. Biochem. , 233 , 1–8. Nicholl, I.D. and Quinlan, R.A. (1994). Chaperone Activity of -Crystallin Modulates Intermediate Filament Assembly. EMBO J. , 13 , 945–953. Nover, L. (1991). The Heat Shock Response , CRC Press, Boca Raton, FL. Nover, L., Sharf, K.D. and Neumann, D. (1989). Cytoplasmic Heat Shock Granules are Formed from Precursor Particles and are Associated with a Specific Set of mRNAs. Mol. Cell. Biol. , 9 , 1298–1308. Opermann, H., Levinson, W. and Bishop, J.M. (1981). A Cellular Protein that Associates with the Transforming Protein of Rous Sarcoma Virus is also a Heat-Shock Protein. Proc. Natl. Acad. Sci. , 78 , 1067–1071. Parsell, D.A. and Lindquist, S. (1994). Heat Shock Proteins and Stress Tolerance. In Morimoto, R.I., Tissiäres, A and Georgopoulos, C. (eds.), The Biology of Heat shock Proteins and Molecular Chaperones , Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Patriarca, E.J. and Maresca, B. (1990). Acquired Thermotolerance Following Heat Shock Protein Synthesis Prevents Impairment of Mitochondrial ATPase Activity at Elevated Temperatures in Saccharomyces cervisiae. Exp. Cell. Res. , 190 , 57–64. Pauli, D., Tonka, C.H., Tissieräs, A. and Arrigo, A.-P. (1990). Tissue-specific Expression of the Heat Shock Protein Hsp27 During Drosophila Melanogaster Development. J. Cell Biol. , 111 , 817–828. Plesofsky; V.N., Vig, J. and Brambl, R. (1992). Phylogeny of the -crystallin-related Heat-shock Proteins. J. Mol. Evol. , 35 , 537–545. Pratt, W.B. (1993). The Role of Heat Shock Proteins in Regulating the Function, Folding
Molecular chaperones and folding catalysts
718
and Trafficking of the Glucocorticoid Receptor. J. Biol Chem. , 268 , 21455–21458. Pratt, W.B. and Welsch, M.J. (1994). Chaperone Functions of the Heat Shock Proteins Associated with Steroid Receptors. Seminars in Cell Biol , 5 , 83–93. Pratt, W.B. and Toft, D.O. (1997). Steroid receptors interactions with heat shock protein and immunophilin chaperones. Endocrin. Rev. , 18 , 306–360. Prodromou, C., Roe, S.M., O’Brien, R., Ladbury, J.E., Piper, P.W. and Pearl, L.H. (1997). Identification and structural characterization of the ATP/ADP-binding site in the Hsp90 molecular chaperone. Cell , 90 , 65–75. Prodromou, C., Roe, S.M., Piper, P.W. and Pearl, L.H. (1997). A molecular clamp in the crystal structure of the N-terminal domain of the yeast Hsp90 chaperone. Nat. Struct. Biol. , 4 , 477–482. Raman, B., Ramakrishna, T. and Rao, Ch.M. (1995). Temperature Dependent Chaperone Activity of -crystallin. FEBS Lett. , 365 , 133–136. Randall, L.L. and Hardy, S.J. (1995). High Selectivity with Low Specificity: How SecB Has Solved the Paradox of Chaperone Binding. Trends. Biochem. Sci. , 20 , 65–69. Rao, V., Horwitz, J. and Zigler, J.S. (1993). -crystallin, A Molecular Chaperone, Forms a Stable Complex with Carbonic Anhydrase Upon Heat Denaturation. Biochem. Biophys. Res. Comm. , 190 , 786–793. Renaud, J.-P, Rochel, N., Ruff, M., Vivat, V., Chambon, P., Gronemeyer, H. and Moras, D. (1995). Crystal Structure of the RAR- Ligand-Binding Domain Bound to All-trans Retinoic Acid. Nature , 378 , 681–689. Rose, D.W., Wettenhall, R.E.H., Kudlicki, W., Kramer, G. and Hardesty, B. (1987). The 90-kilodalton Peptide of the Heme-regulated eIF-2 Kinase has Sequence Similarity with the 90-kilodalton Heat Shock Protein. Biochemistry , 26 , 6583–6587. Roseman, A.M., Chen, S., White, H., Braig, K. and Saibil, H.R. (1996). The Chaperonin ATPase Cycle: Mechanism of Allosteric Switching and Movements of SubstrateBinding Domains in GroEL. Cell , 87 , 241–251. Saga, S., Nagata, K., Chen, W.-T. and Yamada, K.M. (1987). pH-Dependent Function, Purification, and Intracellular Location of a Major Collagen-binding Glycoprotein. J. Cell. Biol. , 105 , 517–527. Sanchez, E.R., Redmond, T., Scherrer, L.C., Bresnick, E.H., Welsh, M.J. and Pratt, W.B. (1988). Evidence that the 90-kilodalton Heat Shock Protein is Associated with Tubulin-containing Complexes in L Cell Cytosol and in Intact PtK Cells. Mol. Endocrinol. , 2 , 756–760. Sanchez, E.R., Toft, D.O., Schlesinger, M.J. and Pratt, W.B. (1985). Evidence that the 90-kDa Phospho-protein Associated with the Untransformed L-cell Glucocorticoid Receptor is a Murine Heat Shock Protein. J. Biol. Chem. , 260 , 12398–12401. Sargan, D.R., Tsai, M.J. and O’Malley, B.W. (1986). Hsp108, A Novel Heat Shock Inducible Protein of Chicken. Biochemistry , 25 , 6252–6258. Satoh, M., Hirayoshi, K., Yokota, S., Hosokawa, N. and Nagata, K. (1996). Intracellular Interaction of Collagen-specific Stress Protein Hsp47 with Newly Synthesized Procollagen. J. Cell Biol. , 133 , 469–483. Scheibel, T., Neuhofen, S., Weikl, T., Mayr, C., Reinstein, J., Vogel, P.D. and Buchner, J. (1997). ATP binding properties of human Hsp90. J. Biol. Chem. , 272 18608–18613. Scheibel, T., Weikl, T. and Buchner, J. (1998). Two chaperone sites in Hsp90 differing in
Mechanisms of ATP-independent vs. ATP-dependent chaperones
719
substrate specificity and ATP dependence. Proc. Natl. Acad. Sci. , 95 , 1495–1499. Schmid, D., Baici, A., Gehring, H. and Christen, P. (1994). Kinetics of Molecular Chaperone Action. Science , 263 , 971–973. Schmid, F.X. (1993). Prolyl Isomerase: Enzymatic Catalysis of Slow Protein Folding Reactions. Annu. Rev. Biophys. Biomol. Struct. , 22 , 123–142. Schmidt, M., Buchner, J., Todd, M., Lorimer, G.H. and Vitanen, P. (1994). On the Role of GroES in the Chaperonin-Assisted Folding Reaction: Three Case studies. J. Biol. Chem. , 269 , 10304–10311. Schulte, T.W., Blagosklonny, M.V., Ingui, C. and Neckers, L. (1995). Disruption of the Raf-1-Hsp90 Molecular Complex Results in Destabilization of Raf-1 and Loss of Raf1-Ras Association. J. Biol. Chem. , 270 , 24585–24588. Schumacher, R.J., Hurst, R., Sullivan, W.P., McMahon, N.J., Toft, D.O. and Matts, R.L. (1994). ATP-Dependent Chaperoning Activity of Reticulocyte Lysate. J. Biol. Chem. , 269 , 9493–9499. Shaknovich, R., Shue, G. and Kohtz, D.S. (1992). Conformational Activation of a Basic Helix-loop-helix protein (MyoD1). by the C-terminal Region of Murine Hsp90 (hsp84). Mol. Cell. Biol. , 12 , 5059–5068. Shue, G. and Kohtz, D.S. (1994). Structural and Functional Aspects of Basic Helix-LoopHelix Protein Folding by Heat Shock Protein 90. J. Biol. Chem. , 269 , 2707–2711. Smith, D.F. (1995). Steroid Receptors and Molecular Chaperones. Sci. Med. , 2 , 38–47. Smith, D.F. and Toft, D.O. (1993). Steroid Receptors and their Associated Proteins. Mol. Endocrinol. , 7 , 4–11. Smith, D.F., Sullivan, W.P., Marion, T.N., Zaitsu, K., Madden, B., McCormick, D.J. and Toft, D.O. (1993). Identification of a 60-kilodalton Stress-related Protein, p60, which Interacts with Hsp90 and Hsp70. Mol. Cell. Biol. , 13 , 869–876. Smith, D.F., Whitesell, L., Nair, S.C., Chen, S., Prapapanich, V. and Rimerman, R.A. (1995). Progesterone Receptor Structure and Function Altered by Geldanamycin, an Hsp90-Binding Agent. Mol. Cell. Biol. , 15 , 6804–6812. Song, H.Y., Dunbar, J.D., Zhang, Y.X., Guo, D. and Donner, D.B. (1995). Identification of a Protein with Homology to Hsp90 That Binds the Type 1 Tumor Necrosis Factor Receptor. J. Biol. Chem. , 270 , 3574–3581. Sparrer, H., Lilie, H. and Buchner, J. (1996). Dynamics of the GroEL-Protein Complex: Effects of Nucleotides and Folding Mutants. J. Mol. Biol. , 258 , 74–87. Stancato, L.F., Chow, Y.-H., Hutchison, K.A., Perdew, G.H., Jove, R. and Pratt, W.B. (1993). Raf Exists in a Native Heterocomplex with Hsp90 and p50 that can be Reconstituted in a Cell-free System. J. Biol. Chem. , 268 , 21711–21716. Stebbins, C.E., Russo, A.A., Schneider, C., Rosen, N., Hartl, F.U. and Pavletich, N.P. (1997). Crystal structure of an Hsp90-geldanamycin complex: targeting of a protein chaperone by an antitumor agent. Cell , 89 239–250 Stege, G.J.J., Brunsting, J.F, Kampinga, H.H. and Konings, A.W. (1995). Thermotolerance and Nuclear Protein Aggregation: Protection Against Initial Damage or Better Recovery? J. Cell Physiol. , 164 , 573–586. Stege, G.J.J., Li, G.C., Li, L., Kampina, H.H. and Konings, A.W. (1994). On the Role of Hsp27 in Heat Induced Intranuclear Protein Aggregation. Int. J. Hyperthermia , 10 , 659–674.
Molecular chaperones and folding catalysts
720
Sullivan, W.P. and Toft, D.O., (1993). Mutational Analysis of Hsp90 Binding to the Progesterone Receptor. J. Biol. Chem. , 268 , 20373–20379. Takechi, H., Hirayoshi, K., Nakai, A., Kudo, H., Saga, S. and Nagata, K., (1992). Molecular Cloning of a Mouse 47-kDa Heat-Shock Protein, and its Expression During the Differentiation of F9 Tetratocarcinoma cells. Eur. J. Biochem. , 206 , 323–329. Theyssen, H., Schuster, H.-P., Packschies, L., Bukau, B. and Reinstein, J. (1996). The Second Step of ATP Binding to DnaK Induces Peptide Release. J. Mol. Biol , 263 , 657–670. Thomas, P.J., Qu, B.-H. and Pedersen, P.L. (1995). Defective Protein Folding as a Basis of Human Disease. Trends Biochem. Sci. , 20 , 456–459. Todd, M.J., Vitanen, P.V. and Lorrimer, G.H. (1994). Dynamics of the Chaperonin ATPase Cycle: Implications for Facilitated Protein Folding. Science , 265 , 659–666. van den Jissel, P.R.L.A., Overcamp, P., Knauf, U., Gaestel, M. and de Jong, W. (1994). A-crystallin Confers Cellular Thermoresistance. FEBS Lett. , 355 , 54–56. Voorter, C.E., de Haard-Hoekman, W., Merck, K.B., Bloemendal, H. and deJong, W.W. (1994). Elastase Inhibition by the C-terminal Domains of -crystallin and Small Heat Shock Protein. Biochem. Biophys. Acta , 1204 , 43–47. Wang, S.-Y. (1994). A Retinoic Acid-inducible GATA-binding Protein Binds to the Regulatory Region of J6 Serpin Gene. J. Biol. Chem. , 269 , 607–613. Waters, E.R., Lee, J.L. and Vierling, E. (1996). Evolution, Structure and Function of the Small Heat Shock Proteins in Plants. J. Exp. Bot. , 47 , 325–338. Wei, J. and Hendershot, L.M. (1996). Protein Folding and Assembly in the Endoplasmic Reticulum. In Morimoto, R.I. and Polla, B. (eds.), Stress-Inducible Cellular Response , Birkhäuser, Boston, MA. Welch, W.J. (1985). Phorbol Ester, Calcium Ionophore, or Serum Added to Quiescent Rat Embryo Fibroblast Cells Result in the Elevated Phosphorylation of Two 28,000dalton Mammalian Stress Proteins. J. Biol. Chem. , 260 , 3058–3062. Wiech, H., Buchner, J., Zimmermann, M., Zimmermann, R. and Jakob, U. (1993). Hsc70, Immunoglobulin Heavy Chain Binding Proein, and Hsp90 Differ in their Ability to Stimulate Transport of Precursor Proteins into Mammaliam Microsomes. J. Biol. Chem. , 268 , 7414–7421. Wiech, H., Buchner, J., Zimmermann, R. and Jakob, U. (1992). Hsp90 Chaperones Protein Folding In Vitro. Nature , 358 , 169–170. Wilson, D.W., Lewis, M.J. and Pelham, H.R.B. (1993). pH-Dependent Binding of KDEL to Its Receptor in vitro. J. Biol. Chem. , 268 , 7465–7468. Wistow, G.J. (1985). Domain Structure and Evolution in -Crystallins and Small HeatShock Proteins. FEBS Lett. , 181 , 1–6. Xu, Y. and Lindquist, S. (1993). Heat-shock Protein Hsp90 Governs the Activity of pp60v-src Kinase. Proc. Natl. Acad. Sci. , 90 , 7074–7078. Yonehara, M., Minami, Y., Kawata, Y., Nagai, J. and Yahara, I. (1996). Heat-Induced Chaperone Activity of Hsp90. J. Biol. Chem. , 271 , 2641–2645. Young, J.C., Schneider, C. and Hartl, F.U. (1997). In vitro evidence that hsp90 contains two independent chaperone sites. FEBS Lett. , 418 , 139–143. Zhu, X., Zhao, X., Burkholder, W.F., Gragerov, A., Ogata, C.M., Gottesman, M.E. and Hendrickson, W.A. (1996). Structural Analysis of Substrate Binding by the Molecular
Mechanisms of ATP-independent vs. ATP-dependent chaperones Chaperone DnaK. Science , 272 , 1606–1614.
721
28. STRUCTURE AND FUNCTION OF PERIPLASMIC PAPD-LIKE CHAPERONES INVOLVED IN ASSEMBLY OF BACTERIAL P PILI SCOTT J.HULTGREN1, *, DANIELLE L.HUNG1 C.HAL JONES1 and STEFAN KNIGHT2 1 Department
of Molecular Microbiology, Washington University, 660 S. Euclid Ave., Box 8230, St. Louis, MO 63110 2S wedish University of Agricultural Sciences, Uppsala Biomedical Center, Department of Molecular Biology, PO Box 590, S-751 24 Uppsala, Sweden
1. Introduction 2. Three Dimensional Structure of PapD 2.1. Structural Basis of Pilus Subunit Recognition by PapD 3. Pilus Assembly Depends on Chaperone Uncapping of a Beta Zipper 4. Conserved Structural Features Delineate Two Chaperone Subfamilies 4.1. Structural Conservation in the Chaperone Superfamily 4.2. FGS and FGL Chaperones 4.3. Structural Features Unique to Each Subfamily 4.3.1 . A Variation in the Beta Zipper Motif Recognized by FGS and FGL Chaperones 5. Role of PapD in Import and Folding of Pilus Subunit Proteins 6. Concluding Remarks 7. References 1. INTRODUCTION Gram negative bacteria express adhesins that mediate microbial attachment by binding to receptors present in host tissues. Microbial attachment is thought to be a key event in the early stages of most diseases as it allows bacteria to gain a foothold in the host and colonize epithelial surfaces (Hultgren et al., 1996). *Corresponding author
Structure and function of periplasmic
723
Adhesins can be assembled onto the surfaces of bacteria as monomers, simple oligomers or as part of supramolecular fibers (de Graaf et al., 1981; de Graaf and Mooi 1986; Duguid et al., 1979; Duguid and Old, 1980; Duguid et al., 1955; Hultgren et al., 1993b; Hultgren et al., 1996). The most well-characterized adhesive fibers are known as pili or fimbriae which are rod-like organelles. Adhesins can also assemble into fibers that are much thinner than pili and tend to collapse into an amorphous mass on the surface of bacteria. Over 30 different adhesive organelles are known to require immunoglobulin-like periplasmic chaperones for their assembly (Hung et al., 1996). The model system discussed in this chapter is the P pilus system. Uropathogenic strains of E. coli express P pili that bind to receptors present in the urinary tract (Hull et al., 1981; Hull et al., 1984; Johanson et al., 1992) and are associated with virulence in pyelonephritic strains of E. coli (Roberts et al., 1984; Roberts et al., 1994). P pili contain an adhesin called PapG that binds to the Gala (1–4)Gal moiety present in the globoseries of glycolipids (Stromberg et al., 1992). The expression of P pili requires the pap gene cluster which encodes eleven proteins, each of which has been studied by extensive genetic and biochemical analyses (Baga et al., 1987; Baga et al., 1984; Hultgren et al., 1991b; Lindberg et al., 1989; Lindberg et al., 1986; Lund et al., 1987; Nilsson and Uhlin, 1991; Svanborg-Eden et al., 1983). Briefly, the pilus is made up of two distinct subassemblies, a thin tip fibrillum joined to the distal end of a thick pilus rod (Hultgren et al., 1991b; Jacob-Dubuisson 1993; Kuehn et al., 1992). The rod is made up of repeating PapA subunits arranged in a right-handed helical cylinder and the tip fibrillum is made up mostly of PapE subunits arranged in an open helical configuration (Baga et al., 1984; Kuehn et al., 1992; Lindberg et al., 1986). The PapG adhesin is joined to the distal end of the fibrillum via the PapF adaptor and the fibrillum is joined to the rod by the PapK adaptor (Jacob-Dubuisson et al., 1993; Kuehn et al., 1992). PapH anchors the pilus to the cell wall by an unknown mechanism (Baga et al., 1987). PapD is the periplasmic chaperone required for the assembly of P pili. All adhesive organelles that require a PapD-like chaperone for their assembly also require an outer membrane protein for their assembly called an usher. PapC is the outer membrane usher required for the assembly of P pili (Thanassi et al., 1998; Dodson et al., 1993; Hultgren et al., 1991b; Lindberg et al., 1989; Nilsson and Uhlin, 1991). The role of PapJ is unclear but has been suggested to function as a co-chaperone (Svanborg-Eden et al., 1978). The structure and function of each of the pap gene products has been discussed extensively in other reviews (Hultgren et al., 1993a; Hultgren et al., 1996). This chapter will focus on the structural basis of how PapD may control pilus assembly. 2. THREE DIMENSIONAL STRUCTURE OF PapD The three dimensional structure of the PapD periplasmic chaperone has been solved by Holmgren and Brändén (Holmgren et al., 1992). The molecule consists of two immunoglobulin-like domains at approximately right angles to each other (Figure 1), with a wide and deep cleft between the two domains. There are seven strands A1-G1 in the first, N-terminal, domain, and eight strands A2-H2 in the second, C-terminal, domain. In both of the domains the core of the -barrel structure is formed by strands E and B in the upper sheet and strands C, F and G in the lower sheet (Figure 2). The second domain
Molecular chaperones and folding catalysts
724
has the same topology as the second domain of the HIV receptor, CD4 (Ryu et al., 1990; Wang et al., 1990) where strand D belongs to the lower sheet in contrast to the situation in a classical immunoglobulin domain where this strand belongs to the upper sheet (Holmgren et al., 1992). The additional eighth H2 strand in the C-terminal domain of PapD
Figure 1 Ribbon model of the PapD-PapG peptide complex. Domain 1 is on the left with strands in the upper sheet (strands D′, E, B, A′). colored pink and strands in the lower sheet (strands D′′, C, F, G, A′′; also referred to as the conserved sheet) colored red. In domain 2, strands in the upper sheet (strands E, B, A). are colored light green, and strands in the lower sheet (D, C, F, G, H). are shown in blue. Also shown as an orange coil is the position of the PapG peptide bound in the cleft and along the G1 strand of PapD. The anchoring residues Arg 8 and Lys 112 in the cleft, as well as three hydrophobic residues at positions 103, 105 and 107 that are part of the beta zipper pattern in the G1 strand are indicated as ball-and-stick models. The flexible F1-G1 loop is highlighted in green. The position of the conserved DRES motif that might be important for domain orientation and uncapping of subunits is indicated in purple. This figure, as well as figures 3, 4 and 5, was produced using MolScript (Kraulis, 1991) and Raster3D (Bacon and Anderson, 1988; Merritt and Murphy, 1994).
extends the lower sheet and is linked to strand G2 by a disulfide bond. In the N-terminal domain of PapD, both the A1 and D1 strands belong to both of the two sheets. The Nterminal domain is thus more similar to an immunoglobulin V domain where strand A is
Structure and function of periplasmic
725
shared between the two sheets in the barrel. The strand switching of strand A1 in PapD is critical for PapD function since it places the conserved residue Arg-8 in the cleft of the chaperone, next to the conserved residue Lys-112 in the G1 edge strand. These two residues play a central role in binding peptides and subunits in the PapD cleft as will be described below. The switch in strand D1 may also be of functional significance due to its role in maintaining the
Figure 2 Strand order in PapD and other immunoglobulin domains. Strands are seen endon and indicated by triangles. The direction of strands along the polypeptide chain is indicated by upward pointing triangles for strands running away from the reader and downward pointing arrows for strands coming out of the plane of the paper. Core strands that are found in all immunoglobulin-like domains are shown in red.
conserved hydrophobic patch across the back of the N-terminal domain (Hung et al., 1996). 2.1. Structural Basis of Pilus Subunit Recognition by PapD One hallmark of all pilus proteins assembled by PapD-like chaperones is the presence of
Molecular chaperones and folding catalysts
726
a conserved carboxyl terminal motif. This motif is defined as a penultimate tyrosine, alternating hydrophobic residues at positions 4, 6 and 8, and a glycine at
Figure 3 Anchoring of the PapG peptide in the cleft of PapD. The C-terminal proline of the PapG peptide bound to PapD is anchored in the cleft of the chaperone via hydrogen bonds to two invariant positively charged residues, Arg 8 and Lys 112. The proline rings packs with van der Waals contacts in a shallow pocket defined by residues Thr 152, Ile 154, Thr 170 and Ile 194 in the second domain of PapD. A short stretch of the peptide has been included in the picture as an orange coil to indicate the direction of the peptide.
position 14 from the COOH-terminus (Kuehn et al., 1993). Studies analyzing PapDPapG complex formation demonstrated that deletions and point mutations in the conserved carboxyl terminal motif abolished PapD-PapG complex formation in vivo (Hultgren et al., 1989) and abolished the ability of the mutant PapG proteins to be incorporated into pili. Based on this information, the conserved COOH terminus was hypothesized to form part of a surface recognized by PapD. To test this hypothesis, 19mer peptides were synthesized corresponding to the carboxyl termini of the P pilus subunits and investigated for their ability to bind to PapD (Kuehn et al., 1993). PapD bound best to the COOH-terminal PapG peptide but also bound moderately well to PapF, PapE, and PapK COOH-terminal peptides. PapD did not bind to the PapH COOHterminal peptide nor to a totally unrelated, hydrophobic control peptide (Kuehn et al., 1993). These results supported the hypothesis that the conserved carboxyl terminal motif of pilus proteins formed part of a surface recognized by the chaperone. The molecular basis of the PapD-peptide interactions was investigated by cocrystallizing PapD with the COOH terminal PapG peptide (Kuehn et al., 1993). The
Structure and function of periplasmic
727
COOH terminal carboxylate group of the peptide bound in the cleft of the chaperone via hydrogen bonding to the invariant Arg-8 and Lys-112 residues (see Figures 1 and 3). These residues probably have a critical subunit-binding function in all of the PapD-like chaperones. Site-directed mutations in Arg-8 and Lys-112 abolish the ability of PapD to bind to pilus subunits and to mediate their assembly into pili arguing that the PapDpeptide crystal structure is a reflection of PapD-subunit interactions (Kuehn et al., 1993). The subsequent positioning of the peptide along the exposed edge of PapD’s G1 strand was mostly the result of backbone hydrogen bonds forming a ß-sheet structure between the chaperone and peptide that was defined as being a “beta zipper” (Kuehn et al., 1993). This relatively non-specific beta zippering interaction may explain in part the ability of PapD to bind to a number of different polypeptides in a relatively sequence-independent manner. In the complex, 690 Å2, or 57%, of the surface area of the extended peptide is buried by the contact with PapD. The largest contribution by far to the interface comes from the ultimate proline and penultimate phenylalanine residues in the peptide which together donate 250 Å2 to the contact area. These two residues thus make up 36% of the total peptide area buried in the PapD-peptide interface. The molecular details of what determines the specificity of PapD-subunit interactions is unknown. In the PapDPapGpeptide crystal structure the alternating hydrophobic residues in the peptide were in register with conserved alternating hydrophobic amino acids present at positions 103, 105 and 107 in the G1 beta strand of PapD. The appropriate alignment of these hydrophobic amino acids may contribute to the stability and the specificity of the complex. However, the polar residues located between these hydrophobic residues might also be important for specific binding since in general they contribute as much to the interface area as the alternating hydrophobic residues. Additional genetic studies have identified a second site of interaction between PapD and PapG in a region between residues 175 and 198 of PapG (Xu et al., 1995). Peptides were synthesized corresponding to the region between residues 159 to 195 and tested for their ability to bind PapD in an ELISA assay. PapD bound to one peptide that corresponded to residues 175 to 190 but the structural basis of the interaction has not yet been elucidated. 3. PILUS ASSEMBLY DEPENDS ON CHAPERONE UNCAPPING OF A BETA ZIPPER Pilus subunits possess complementary interactive surfaces that drive their assembly into the fiber. PapA subunits associate into right handed helical rods while PapE subunits associate into linear open helical fibers. PapK is thought to adapt each pilus rod to a tip fibrillum which presumably involves an interaction with the most distal PapA of the rod and the most proximal PapE of the tip fibrillum (Jacob-Dubuisson et al., 1993). Similarly, PapF is thought to join the PapG adhesin to the distal end of the tip fibrillum and thus presumably has a structure that facilitates the appropriate interactions with the PapG and PapE subunits. The function of PapD is to prevent premature subunit associations at the wrong time and the wrong place in the cell (Baga et al., 1985). PapD binds to subunits as (or immediately after) they emerge from the cytoplasmic membrane forming periplasmic
Molecular chaperones and folding catalysts
728
chaperone-subunit complexes (Hultgren et al., 1996, Kuehn et al., 1991). Thus, the subunits are partitioned into chaperone-subunit complexes to prevent subunit-subunit interactions in the periplasm. Electron microscopic analyses of purified PapD-PapA, PapD-PapE, PapD-PapK and PapD-PapG complexes does not reveal any oligomeric subassemblies. However, under in vitro freeze-thaw conditions that facilitate the dissociation of the chaperone, rod and tip subassemblies were reconstituted from purified PapD-PapA and PapD-PapE complexes, respectively, demonstrating that dissociation of the chaperone triggers self-assembly of the corresponding fibers (Bullitt et al., 1996). However, the presence of excess PapD blocked the formation of the fibers arguing that PapD binds competitively to a self-associative surface on the subunit. From the structural studies described above, we know that PapD binds in part to the carboxyl terminus of subunits via a beta zippering interaction. Mutagenesis studies have revealed that this same motif participates in subunit-subunit interactions after chaperone uncapping (Bullitt et al., 1996). This finding explains the molecular basis of how the PapD chaperone controls the development of pilus fibers: PapD caps a critical assembly surface at the carboxyl terminus of the subunit to prevent premature subunit interactions. This surface is exposed following chaperone uncapping and provides an assembly template for the controlled incorporation of the next subunit. In vivo, chaperone-subunit complexes are targeted to the PapC usher, where the chaperone is dissociated. Dissociation of the chaperone uncaps the carboxyl terminal interactive surface of the subunit which drives its assembly into the pilus. The mechanism by which PapC facilitates chaperone uncapping is currently under investigation. Pilus biogenesis depends on the targeting of chaperone-subunit complexes to the usher (Dodson et al., 1993). Dissociation of the chaperone from the subunit exposes the surface on the subunit required for its assembly into the pilus structure. Very little is known about the molecular basis of the interactions that ushers make with chaperone-subunit complexes. In the absence of PapC, chaperone-subunit complexes accumulate in the periplasmic space arguing that the usher is required to facilitate the dissociation of the chaperone from subunits allowing their assembly into pili (Dodson et al., 1993, Normark et al., 1988). Binding of PapD to the PapG COOH terminal peptide was shown to induce a 13° jaw-closing or hinge-bending motion making the angle between the two domains more acute (Kuehn et al., 1993). Therefore, dissociation of the chaperone may be triggered by a rearrangement of chaperone domains upon chaperonesubunit-usher interactions that induces its release from the subunit. Insight into the mechanism of how these myriad of protein-protein interactions may trigger chaperone dissociation has come from an inspection of conserved features of the domain-domain interface as described below. There exists a charge-charge/hydrogen-bond network at the domain interface involving two carboxylate side chains (Glu-83 in domain 1 and Asp-196 in domain 2) bridged by an arginine side chain (Arg-116) that is invariant in all except one of the chaperones (Figure 4). This bridge is thought to be important for orienting the domains to give a functional subunit binding cleft (Holmgren et al., 1992). The glutamic acid Glu-83 side chain is part of the conserved surface exposed DRES motif at the end of the E1-F1 loop which is located near the hinge of the molecule. Part of this loop packs against the hinge region connecting the two domains. It is
Structure and function of periplasmic
729
Figure 4 Conserved DRES motif (red coil) at the domain interface of PapD. The DRES motif packs against the polypeptide segment joining the two domains (shown as a green coil at the bottom of the picture). The positively charged side chain of Arg 116 in the linker region bridges the two domains by hydrogen bonds to the carboxylate side chains of Glu 83 in the DRES motif, and Asp 196 in the second domain. The Asp 81 and Arg 82 residues in the DRES motif are solvent exposed, and it is conceivable that interactions between these residues and other proteins during uncapping at the usher could cause conformational changes that would result in a reorientation of the domains.
conceivable that interactions with other proteins (e.g. subunits or ushers) in this region could cause conformational changes that would be transmitted to the interdomain region, possibly disrupting the conserved charge-charge/hydrogenbond network and triggering or facilitating a reorientation of the domains. We therefore propose that incoming chaperone-subunit complexes may interact directly with the solvent exposed residues in the DRES motif of the chaperone that is part of the chaperone-subunit-usher complex, which could trigger the release of the chaperone via a conformational change. In this context it is interesting to note that a similar motif (DREA) is present at the sequentially and topologically equivalent position in another class of immunoglobulinlike molecules called cadherins. Cadherins are a class of eukaryotic cell surface molecules that have tandemly repeated immunoglobulin-like domains in their extracellular regions (Shapiro et al., 1995) and that are important for controlling the
Molecular chaperones and folding catalysts
730
development and maintenance of tissues. In cadherins the DREA motif is involved in calcium ion binding which orients successive cadherin domains in a rigid conformation on the cell surface where they participate in cell-cell interac-tions. The glutamic acid side chain of the DREA motif, together with negatively charged side chains from other parts of the cadherin domain, forms a calcium binding site at the prospective interface between successive cadherin domains (Shapiro et al., 1995). In chaperones, instead of binding calcium, Glu-83 interacts with the positively charged side chain of Arg-116. Thus a similar motif, at the same location in both cadherins and periplasmic chaperones, may be involved in a similar function. In both cases the glutamic acid side chain is part of a binding site for a positively charged group that bridges successive domains, a calcium ion in cadherins and an arginine side chain in the chaperones. In both classes of molecules these bridging interactions are thought to be important for proper orientation of successive immunoglobulin-like domains. 4. CONSERVED STRUCTURAL FEATURES DELINEATE TWO CHAPERONE SUBFAMILIES The chaperone/usher pathway is used to assemble a wide range of adhesive organelles in gram negative bacteria (Table 1). Essentially, the different fibers assembled by the chaperone/usher pathway can be divided into two different groups which both contain adhesins: pilus fibers and nonpilus fibers. By typical transmission electron microscopy, pili appear to be rod-like structures that range in diameter from 2–10 nm and are typically 5–7 micrometers in length. However, high resolution electron microscopy reveals that pili typically contain a very thin and short projection at their distal ends, the tip fibrillum, that often contains the adhesive component of the pilus. Pilus rods are comprised of subunits that can be arranged in a number of different helical symmetries depending on the pilus type. For example P pilus rods are 68 A in diameter and are comprised of PapA subunits arranged in a right handed helical cylinder having 3.28 subunits per turn (Gong and Makowski, 1992). The pilus rod also has a 15 A helical cavity winding through the rod and communicating with the external environment by a set of radial channels. The fibrillum of the P pilus is ~20 Å diameter and is comprised mostly of repeating PapE subunits arranged in an open helical configuration (Kuehn et al., 1992). Joined to the distal end of the P pilus tip fibrillum is the PapG adhesin. The nonpilus fibers have dramatically different molecular architectures. They are not associated with pilus rods but instead consist only of a thin linear fiber that contains an adhesin. In some cases the adhesin is presented on the surface of the organism as a monomer or simple oligomer similar to a tip fibrillum (Ahrens et al., 1993; Garcia et al., 1994; Goldhar et al., 1987; Le Bouguenec et al., 1993). In other bacteria, the nonpilus fibers are very thin (< 20 A) long and flexible and often coil up into an amorphous mass on the bacterial surface (Clouthier et al., 1993; Iriarte et al., 1993; Karlyshev et al., 1992; Knutton et al., 1989; Levine et al., 1984; Lindler and Tall 1993; Muller et al., 1991; Savarino et al., 1994; Wolf et al., 1989). Fine molecular details of the architecture of nonpilus fibers is lacking. The immunoglobulin-like superfamily of chaperones can be divided into two subfamilies based on structural differences conserved within each subgroup (Hung et al., 1996). Interestingly, one subfamily of chaperones is specific for the assembly of pili whereas the
Structure and function of periplasmic
731
other subfamily is specific for the Table 1 Superfamily of Immunoglobulin-like chaperones
Chaperone % Identity to PapD*
Organism
FGS:
Surface Structure Assembled
Reference
Pili:
+PapD
–
E. coli
P pili
Lindberg et al., 1989 Kuehn et al., 1992
+SfaE
33
E. coli
S Fimbriae
Schmoll et al., 1990 Hacker et al., 1994
+FimC
34
E. coli
Type 1 Pili
Jones et al., 1993 Klemm and Krogfelt, 1994
+HifB
32
H. influenzae
H. influenzae Fimbriae
Stull et al., 1984 NIH sequence data base
MrpD
56
P. mirabilis
MR/P Fimbriae
Bahrani et al., 1991 Bahrani et al., 1994
FocC
33
E. coli
FIC Fimbria
Klemm et al., 1994 Klemm et al., 1995
FimB
31
B. pertussis
Type 2 and 3 Fimbriae
Steven et al., 1986 Willems et al., 1992
PefD
32
S. typhimurium PEF
Friedrich et al., 1993
PmfD
47
P. mirabilis
Bahrani et al., 1993 Massad et al., 1994
LpfB
33
S. typhimurium Long Polar Fimbriae
Baumler et al., 1995
FasB
28
E. coli
987P Fimbriae
Edwards et al., 1996* Cao et al., 1995*
HafB
28
H. influenzae
H. influenzae biogroup aegyptius fimbrium
Read et al., 1996*
AftB
36
P. Mirabilis
Ambient-temperature fimbriae
Massad et al., 1994* Massad et al., 1996*
FanE
31
E. coli
K99 Fimbriae
Duchet-Suchaux et al., 1988 Bakker et al., 1991
FaeE
28
E. coli
K88 Fimbriae
Stirm et al., 1967 Foged et al., 1986 Bakker et al., 1991
F17D
30
E. coli
F1 7 Fimbriae
Lintermans et al.,
PMF Fimbriae
Molecular chaperones and folding catalysts
732 1988 Holmgren et al., 1992
MrkB
34
K. pneumonia
Type 3 Fimbriae
Allen et al., 1991
Unknown Structure: EcpD
35
E. coli
?
Raina et al., 1993
YehC
33
?
?
NIH sequence data base
YraI
35
E. coli
?
NIH sequence data base*
RalE
28
Rabbit EPEC
?
NIH sequence data base*
FGL: Nonpilus Fibers: ClpE
29
E. coli
CS31A Capsule -Like Protein
Girardeau et al., 1988 Bertin et al., 1993
CssC
27
E. coli
Antigen CS6
Knutton et al., 1989 Wolf et al., 1989 NIH sequence data base
MyfB
31
Y. enterocolitica
Myf Fimbriae
Iriarte et al., 1993
PsaB
29
Y. pestis
pH6 Antigen
Lindler et al., 1993
CS3–1
25
E. coli
CS3 Pili
Levine et al., 1984 Jalajakumari et al., 1989
Surface
Chaperone
% Identity to PapD* Organism
Caf1 M
25
Y. pestis
Envelope Antigen F1 Galyov et al., 1991 Karlyshev et al., 1992
NfaE
32
E. coli
Nonfimbrial Adhesins Goldhar et al., 1987 I Ahrens et al., 1993
SefB
25
S. enteritidis SEF14 Fimbriae
Muller et al., 1991 Clouthier et al., 1993
AggD
31
E. coli
Aggregative Adherence Fimbria I
Savarino et al., 1994
AfaB
32
E. coli
AFA-III
Le Bovguenec et al., 1993 Garcia et al., 1994
Structure Assembled
Reference
Structure and function of periplasmic
733
*Each protein was aligned to PapD, the prototype member of the chaperone superfamily, from amino acid 22 to the end of the protein using the program GAP (version 7; Genetics Computer Group). +Consists of two subassemblies: a thick pilus rod joined to a thin, flexible tip fibrillum.
assembly of non-fimbrial adhesins. An analysis of the fine molecular details that distinguishes these two subgroups of chaperones and the subunits that they assemble has given insight into the molecular basis of chaperone-assisted assembly of diverse adhesive organelles in pathogenic gram negative bacteria (Hung et al., 1996). 4.1. Structural Conservation in the Chaperone Superfamily All of the periplasmic chaperones have a conserved hydrophobic core that maintains the overall immunoglobulin-like features of the two domains. The two immunoglobulin-like domains are oriented at approximately right angles creating a cleft between the domains. The subunit binding site in the chaperone cleft is conserved amongst the entire superfamily. All chaperones contain the invariant Arg-8 and Lys-112 residues in the crevice of the cleft as well as conserved hydrophobic residues at the exposed edge of the G1 beta strand at positions 103, 105 and 107. It is known that the exposed edge of the G1 beta strand of PapD and the crevice of the cleft form a critical part of the subunit binding site. All of the chaperones probably utilize these conserved features to recognize subunits and form chaperone-subunit complexes. PapD will functionally substitute for FimC in the assembly of type 1 pili. Mutations in Arg-8 or Lys-112 of PapD abolish it’s ability to assemble both P and type 1 pili (Jones et al., 1993) arguing that the subunit binding surfaces are conserved between PapD and FimC and probably throughout the whole chaperone superfamily. These conserved residues most likely participate in an anchoring and -zippering subunit recognition paradigm used by all of the chaperones. In addition to the hydrophobic residues at positions 103, 105, and 107 there are eight other surface exposed residues that are conserved in character in the chaperone superfamily: five more are hydrophobic, one is positively charged, and two are polar (Hung et al., 1996). Many of these conserved surface exposed residues are located on the -sheet formed by strands A1'’, G1, F1, C1 and D1'’ in domain 1 that we have defined as the conserved sheet (shown in red in Figure 1). Almost all of the conserved surface exposed residues are in domain 1, either in the cleft, at the side of the domain defined by the conserved sheet, or at the back of the domain. 4.2. FGS and FGL Chaperones The chaperone sequences fall into two groups differentiated by conserved differences in the conserved sheet of domain 1 and by the number of amino acids in the loop between -strands F1 and G1, termed the F1-G1 loop (Hung et al., 1996). We have denoted the twenty-two chaperones that have a short F1-G1 loop ranging in size from 10 to 20 (only one having 20) residues as part of the FGS subfamily, deriving their name from the loop length (F1-G1 Short). The other nine chaperones have a long F1-G1 loop that ranges in size from 21 to 29 (only one has 21 residues). We have classified these chaperones as
Molecular chaperones and folding catalysts
734
part of the FGL subfamily (F1-G1 Long). Interestingly, FGS chaperones specifically assemble pili while FGL chaperones specifically assemble nonpilus fibers (See Table 1). 4.3. Structural Features Unique to each Subfamily Conserved differences between the two subfamilies are concentrated in the conserved sheet of domain 1 in the immediate vicinity of the subunit binding Arg-8 and Lys-112 residues (Hung et al., 1996). In the FGS subfamily, the conserved sheet contains invariant residues Trp-36 and Asn-89 and position 110 is always occupied by a positive charge. These residues are all adjacent to one another in a row across the beta strands of the conserved sheet. In the FGL subfamily, position 36 is never a tryptophan but instead is occupied by either a polar or charged residue and positions 89 and 110 are invariant cysteine residues that by modeling are predicted to form a disulfide bond. Given the increased number of residues between the F1 and G1 beta strands in the FGL chaperones, the two cysteines may play a role in stabilizing the structure. The F1-G1 loop occurs at the same relative position in a chaperone molecule as the hypervariable region three (CDR3) in an immunoglobulin molecule which forms part of the antibody combining site (Amit et al., 1986; Holmgren et al., 1992; Hultgren et al., 1993b). CDR3 has the highest degree of variability in its amino acid composition and thus provides for much of the specificity in an antibody-antigen interaction (Amit et al., 1986). The F1-G1 loop is also variable in length and sequence and may provide part of the specificity in chaperonesubunit interactions (Holmgren et al., 1992; Hultgren et al., 1993b). 4.3.1. A Variation in the Beta Zipper Motif Recognized by FGS and FGL Chaperones All subunits assembled by the FGS chaperones contain the conserved carboxyl terminal motif known as the beta zipper (Hung et al., 1996). This motif is bounded by a conserved penultimate tyrosine and a conserved glycine residue positioned 14 residues from the COOH terminus. In between these two residues is a conserved pattern of alternating hydrophobic residues. The subunits assembled by the FGL chaperones contain a variation of this beta zipper motif. The majority of these subunits contain a tyrosine at position 3 from the COOH-terminus, hydrophobic residues at positions 6 and 8, a tyrosine at position 12, and a glycine at position 14 from the COOH terminus. FGL chaperones contain the invariant Arg-8 and Lys112 residues as well as the conserved alternating hydrophobic residues in the G1 beta strand that are known to form a critical part of the subunit binding site of FGS chaperones. This conservation argues that FGL chaperones utilize a similar subunit binding paradigm. In this context it is interesting to note that K110, along with Q108, makes the largest individual contribution to the PapD-peptide interface in the PapD-PapGpeptide crystal structure. In the structure, the K110 side chain takes part in a hydrogen bond with the side chain of Asn 89, and forms van der Waals interactions with Ser 3' and Val 5' in the peptide. Thus, it would not be surprising if changes in these two residues could affect peptide, and subunit, binding. As described above, the beta zipper motif is a surface that docks to two different proteins during pilus biogenesis. Prior to the assembly of a subunit into the pilus this surface is capped by a periplasmic chaperone. However, at the outer membrane assembly
Structure and function of periplasmic
735
site where the chaperone is uncapped, the beta zipper motif exchanges it’s interaction with PapD to associate with an unidentified surface on a pilus subunit present at the growing base of the fiber. If the altered beta zipper motif, present in subunits assembled by FGL chaperones, also participates in subunit-subunit interactions, the structural alteration of the beta zipper motif may result in the subunits being assembled into linear fibers as opposed to helical rods. The variation of the beta zipper motif may evoke a requirement for a different type of subunit-subunit interaction that has important ramifications on the molecular architecture of the fiber that is formed. 5. ROLE OF PapD IN IMPORT AND FOLDING OF PILUS SUBUNIT PROTEINS Pilus formation requires the periplasmic chaperone. In the absence of the chaperone, pilus subunits are subject to proteolytic degradation. This seemingly simple result raises many intriguing questions concerning the role of the chaperone in pilus biogenesis. Is degradation the result of misfolding of the subunits? If so, does this mean that the chaperone plays a role in directing the folding process? How early in the pathway does PapD intervene to block competing off pathway reactions? Insight into some of these questions was gained by investigating the fate of pilus subunits when expressed in the absence of PapD (Jones et al., 1997). It was reasoned that a thorough understanding of the off pathway reaction would give insight into the mechanism of action of PapD. These experiments were facilitated by the use of a degP strain (Strauch and Beckwith, 1988; Strauch et al., 1989). DegP is the periplasmic pro tease responsible for the degradation of pilus subunits (Jones et al., 1997). Interestingly, pilin expression in a strain lacking the DegP protease is toxic; bacteria fail to grow (Jones et al., 1997). Thus, the addition of IPTG to a degP culture containing a subunit cloned downstream of the tac promoter is toxic: the cells fail to grow in contrast to the vector control. However, if the culture is plated out on media lacking IPTG, growth is restored. These results suggest that the toxic effect associated with the expression of a subunit in the absence of a chaperone is not lethal but instead activates a growth checkpoint control mechanism until inducer is removed. Coexpression of the chaperone with the subunit or complementation with the DegP protease suppresses the toxic effect. Pilus subunit proteins have typical leader sequences and are thought to cross the inner membrane via the Sec machinery (Hultgren et al., 1993b; Hultgren and Normack 1991a; Hultgren et al., 1991b). If the leader peptide encoding sequence is deleted from the pilus subunit, expression of the leaderless subunit is no longer toxic (Jones et al., 1997). This suggests that the toxic effect is associated with off-pathway reactions engaged in by the subunits following their export across the cytoplasmic membrane. As explained previously in this chapter, the role of PapD is to cap interactive surfaces on subunits to prevent their premature aggregation at the wrong time and in the wrong place in the cells. Therefore, the toxic effect of subunit expression in the absence of PapD is probably related in some way to premature aggregation of the subunits. Subcellular localization of pilus subunits expressed in the presence and absence of PapD revealed a dramatic effect. In the presence of PapD, subunits are localized in the periplasmic space as soluble chaperone-subunit complexes. However, in the absence of PapD, periplasmic
Molecular chaperones and folding catalysts
736
localization of the subunits is no longer efficient. Instead, the majority of subunit proteins are associated with the inner membrane fraction (Jones et al., 1997). We therefore propose that the SecYEG translocase is not sufficient for efficient release of subunits into the periplasm. Expression of PapG does not, however, block translocation and maturation of MBP (Jones et al., 1997) arguing that PapG is not “jamming” the translocation pore in a manner similar to several hybrid proteins described previously (Snyder and Silhavy, 1992). The toxic effect must therefore be associated with reactions that occur after the export of the subunits through the SecYEG translocase. Thus, the chaperone facilitates the partitioning of subunits from a membrane-tethered state to soluble periplasmic chaperone-subunit complexes. It remains to be determined whether PapG is associated with SecD, another Sec component or a novel site in the inner membrane prior to it’s interaction with PapD. Jones et al., (Jones et al., 1997) developed a spheroplast assay to further investigate the chaperone-assisted import reaction. PapG was expressed in radiolabeled spheroplasts. Following a chase, the spheroplasts were centrifuged and the supernatent was examined for the presence of PapG by immunoprecipitation with anti-PapG antibodies. Using this assay, we discovered that the addition of purified PapD to the spheroplasts stimulated a 30 fold increase in the amount of PapG released into the supernatent. The chaperoneassisted release of the subunits occurred whether PapD was added at the beginning of the chase or 30 minutes later. Taken together these results argue for the following model. PapG is exported through the Sec translocase and folds into a membrane-associated stable intermediate. PapD recognizes a surface on the membrane-associated PapG which results in the release of the adhesin into the supernatent as a soluble PapD-PapG complex. The structural basis of the chaperone-assisted import reaction was investigated by analyzing the effect of mutations in both PapD and PapG on the reaction (Jones et al., 1997). Mutations in the invariant Arg-8 residue in the cleft of PapD greatly reduced the ability of the purified mutant protein to release PapG from the surface of spheroplasts arguing that hydrogen bonding of the terminal carboxylate group of PapG to the Arg-8 residue of PapD is critical for the import reaction. The role of beta zipper formation in import was investigated by mutagenizing the COOH terminal beta zipper motif of PapG and by deleting it. Mutations in the COOH terminal beta zipper of PapG that blocked PapD-PapG complex formation, also blocked the ability of PapD to facilitate the release of the mutant PapG from the spheroplast surface (Jones et al., 1997). However, deletion of the carboxyl terminal 100 amino acids of PapG resulted in a truncate, PapG- l00, that was efficiently released into the supernatant even without the addition of purified PapD. These results suggest that the presence of the beta zipper motif on a subunit predicates the need for a periplasmic chaperone to facilitate its import into the periplasmic space. The structural basis for the import reaction involves the anchoring of the terminal carboxylate group of a subunit into the cleft of the chaperone and the formation of a beta zipper between the COOH terminus of the subunit and the exposed edge of the G1 beta strand of PapD. Based on studies with the PapA protein (Bullitt et al., 1996), this same COOH-terminal beta zipper motif forms part of a surface involved in subunit-subunit interactions. If this surface is not capped by the chaperone there is inefficient import into the periplasm and the majority of the subunits remain associated with the membrane.
Structure and function of periplasmic
737
Mutations in the COOH terminal motif of PapA that reduce or abolish PapA-PapA interactions also reduce or eliminate the toxic effect associated with the expression of PapA in the absence of PapD in a degP strain (Jones et al., 1997). Unlike wild type subunits that aggregate when released into the periplasm in the absence of PapD, these mutant PapA subunits are relatively non-aggregative. This finding suggests that the Cterminal beta zipper may contribute to inner membrane retention and may also be a cause of the observed toxicity. The association of this surface with the membrane may help to keep this surface capped until the subunit is bound by PapD and released into the periplasm as a chaperone-subunit complex. In this context, the membrane may provide a chaperone-like function to prevent premature aggregation of the subunits prior to their interaction with PapD. This hypothesis is consistent with the finding that PapD can facilitate the import of subunits up to 30 minutes after their expression in spheroplasts. Interestingly, the chaperone-assisted import of subunits by PapD is reminiscent of the role reported for Hsp70 chaperones which are thought to facilitate the vectorial transport of proteins into intracellular compartments as well as to assist in the folding process (Kang et al., 1990; Nicchitta and Blobel, 1993; Schneider et al., 1994; Ungermann et al., 1994). The studies described above have brought to light some of the most interesting questions regarding the mechanism of action of PapD. What is the state of a subunit that is recognized by PapD and does the chaperone-subunit interaction prevent off-pathway misfolding reactions? What are the molecular details by which chaperone intervention may predispose subunits down a productive folding pathway? When complexed to PapD, subunits are thought to exist in native-like conformations. This is best exemplified by PapG which possesses its intramolecular disulfide bonds and is capable of binding its Gal (1–4) Gal receptor when in a complex with PapD. Thus, one possibility is that the offpathway reactions that are observed in the absence of the chaperone would occur after the correct folding of the subunits. In other words, in this model, the chaperone intervention would occur only after the subunits fold into native conformations. However, the final folding of the subunits most likely does not occur until they are released from the membrane and since subunit-subunit associations (self-assembly) can occur instantaneously between folded subunits in solution, it is difficult to explain how the chaperone would efficiently prevent premature subunit-subunit associations in this model. An alternative model would be that the subunits are maintained in a stable intermediate in association with the periplasmic side of the cytoplasmic membrane. In part, the subunit is retained by the membrane due to interactions involving the C-terminal beta zipper. Chaperone binding to the beta zipper releases the subunit from the membrane and provides some sort of template for predisposing subunits down the correct folding pathway. Pilus subunits are thought to contain a high degree of beta sheet structure based on their sequences. By maintaining the COOH terminus of subunits in an extended conformation, the chaperone may facilitate beta sheet formation in the subunit. Based on the spheroplast assay, we would propose that PapD binds to a stable intermediate. This intermediate could be molten globule-like, containing a significant amount of tertiary structure. The interaction with PapD may facilitate the shuffling of secondary and/or tertiary structures into their native-like states or possibly provide a tertiary context for the formation of native structure. In this model, chaperone
Molecular chaperones and folding catalysts
738
intervention could be coupled with translocation of subunits across the cytoplasmic membrane and conformational changes or folding reactions that occur on the chaperone template could facilitate the import of the subunits into the periplasmic space as soluble chaperone-subunit complexes. Additional structural alterations in the subunit may occur upon it’s assembly into the pilus which could be a factor in driving the process. In the crystal structure of the PapD-PapG peptide complex, the juxtaposed hydrophobic residues in the peptide and in the G1 strand of PapD form a relatively large continuous hydrophobic bed on the surface of the chaperone (Figure 5). Due to its size, it appears likely that this hydrophobic bed would not be solvent exposed in a PapD-subunit complex. The hydrophobic effect is thought to be one of the main driving forces in protein folding. Non-assisted protein folding is thought to begin with the formation of secondary structure, and there is also accumulating evidence that beta strand formation is at least to some extent context dependent (Hamada et al., 1996; Kelly, 1996; Minor and Kim, 1994). If the anchoring of a pilus subunit C-terminus in the cleft of a periplasmic chaperone, and formation of the beta zipper, is an early event in the interaction between a chaperone and a subunit, the chaperone could provide a predefined context in the form of a template for initial beta-strand formation. The ensuing hydrophobic bed could then function as a hydrophobic nucleus to drive the folding of the subunit. A nucleation/condensation folding mechanism has been argued for a number of small proteins (Baldwin, 1995; Fersht, 1995; Itzhaki et al., 1995; Shakhnovich, 1996). In the context driven nucleation/condensation model for pilus subunit folding presented here,
Figure 5 GRASP-generated (Nicholls, Sharp and Honig, (1991). Protein folding and association: Insights from the interfacial and thermodynamic properties of hydrocarbons. Proteins II, 281–96) solvent accessible surface of the PapD-PapG peptide complex with surface exposed hydrophobic groups coloured yellow. The Nterminal domain with the bound peptide is to the right. The alternating hydrophobic residues in the beta-zipper interaction
Structure and function of periplasmic
739
between the PapG peptide and the G1 strand of PapD form a continuous hydrophobic bed on the surface of the complex that could function as a hydrophobic nucleus for subunit folding in a developing PapDsubunit complex.
periplasmic chaperones would act as folding catalysts by facilitating the rate limiting nucleation step. Jones et al., (Jones et al., 1997), also recently discovered that the “off pathway” products of misassembly are sensed by the bacterium and trigger the expression of a panoply of new gene products that are recruited to the periplasmic space (see Missiakas and Raina, this volume). The presence of “unchaperoned” pilin subunits that are misfolded and/or aggregated in the periplasm is sensed by one of two different signal transduction pathways recently described: the sigma E modulation pathway and the Cpx pathway (Cosma et al., 1995; Danese et al., 1995; Mecsas et al., 1993). The two pathways that have been described respond to different signals and activate the degP promoter via different intermediates. Jones et al., demonstrated that PapG expression stimulated degP transcription via both pathways in the absence of PapD. PapD coexpression prevented the activation of the promoters. The signal sensed in the periplasm may be the misfolding or aggregation of PapG or possibly a membrane perturbation due to PapG association with the inner membrane. The ability of the bacteria to sense misassembly up-regulates the synthesis of proteins involved in folding, such as DsbA (Paul, Danese and Thomas Silhavy, personal communication), and proteases, like DegP that destroy misfolded proteins (see Missiakas and Raina, this volume). The Cpx and sigma E modulation pathways may function to monitor pilus assembly, to ward off the toxic effects of off-pathway reactions. 6. CONCLUDING REMARKS Pilus biogenesis is an excellent model for the investigation of a panoply of basic issues in molecular biology. It represents a unique opportunity to blend an advanced genetic system with X-ray crystallography, protein chemistry, biophysical chemistry, high resolution electron microscopy, and cell biology to study a defined macromolecular system: the development of adhesive surface fibers called pili in pathogenic bacteria. In addition, the function of these fibers in mediating interactions at the host-pathogen interface and the molecular consequences of these interactions in pathogenesis is also of great interest. Genes important in P pili biogenesis, papA-K, are linked in an operon and their expression coordinately controlled. Pilus subunits are exported across the cytoplasmic membrane via the Sec machinery. However, the release of the subunit from the membrane into the periplasm requires a beta zippering interaction with an immunoglobulin-like chaperone. The membrane-associated subunits most likely exist in a molten globule-like pro tease sensitive state. The interaction with the chaperone releases the subunits from the membrane as protease resistant chaperone-subunit complexes. By maintaining the carboxyl terminus in an extended conformation the interaction of the subunit with the chaperone may predispose it’s folding into an assembly competent conformation. The crystal structure of PapD bound to a PapG peptide has provided a
Molecular chaperones and folding catalysts
740
molecular snapshot of this fundamental process. In the related type 1 pilus system, FimH is the adhesin and FimC is the chaperone. We have crystallized the FimC-FimH complex (Knight et al., 1997) and are currently working on solving it’s three dimensional structure. This information should provide great insight into the function of the chaperone in interacting with subunits and it’s mechanism of action. Chaperone-subunit complexes are targeted to specific outer membrane assembly sites comprised of proteins called molecular ushers. The interaction of the chap-erone-subunit complexes with the usher results in dissociation of the chaperone, exposing interactive surfaces on the subunits that drive their assembly into pili on the surface of the bacteria. Efforts are now underway to define the structural and functional characteristics of the PapC usher. PapC has been incorporated into multilamellar liposomes and demonstrated to form a pore (Thanassi et al., 1998), suggesting that the usher may act as a channel in the outer membrane for the pilus subunits to pass through. A carboxyl terminal truncate of PapC retained the ability to make a pore, but was unable to assemble pili in vivo. This indicates that pore formation alone is not sufficient for usher function and that other activities, such as proper interaction with the chaperone-subunit complexes, are important. High molecular weight complexes of PapC were detected by gel electrophoresis and crosslinking of PapC also yielded high molecular weight species, suggesting that the active form of PapC is an oligomer. High resolution electron microscopy of purified PapC revealed the first images of an usher: roughly circular complexes containing a central pore of sufficient diameter to allow passage of a pilus subunit. The PapC usher protein is therefore likely to facilitate the assembly of pili by acting as an oligomeric transport channel in the outer membrane. The long term goals of these efforts are to be able to reconstitute pilus assembly in vitro and to solve the structural basis of the assembly process. Understanding the molecular details of the biogenesis of these virulence organelles will facilitate the design of novel antimicrobial therapeutics and vaccines. 7. REFERENCES Ahrens, R., M.Ott, A.Ritter, H.Hoschutzky, T.Buhler, F.Lottspeich, G.J.Boulnois, K.Jann, and J.Hacker (1993). Genetic analysis of the gene cluster encoding nonfimbrial adhesin I from an Escherichia coli uropathogen. Infection and Immunity , 61, 2505– 2512. Allen, B.L., G.F.Gerlach, and S.Clegg (1991). Nucleotide sequence and functions of mrk determinants necessary for expression of type 3 fimbriae in Klebsiella pneumoniae . Journal of Bacteriology , 173, 916–920. Amit, A.G., R.A.Marrizzua, S.E.Phillips, and R.J.Poljak (1986). Three dimensional structure of an antibody-antigen complex at 2.8 Å resolution . Science , 233, 747–753. Bacon, D.J. and W.F.Anderson (1988). A fast algorithm for rendering space-filling molecule pictures. Journal of Molecular Graphics , 6, 219–220. Baga, M., M.Goransson, S.Normark, and B.E.Uhlin (1985). Transcriptional activation of a Pap pilus virulence operon from uropathogenic Escherichia coli . EMBO Journal , 4, 3887–3893. Baga, M., M.Norgren, and S.Normark (1987). Biogenesis of E. coliPap pili: PapH, a minor pilin subunit involved in cell anchoring and length modulation. Cell , 49, 241–
Structure and function of periplasmic
741
251. Baga, M., S.Normark, J.Hardy, P.O’Hanley, D.Lark, O.Olsson, G.Schoolnik, and S.Falkow (1984). Nucleotide sequence of the gene encoding the papA pilus subunit of human uropathogenic Escherichia coli . Journal of Bacteriology , 157, 330–333. Baldwin, R.L. (1995). The nature of protein folding pathways: The classical versus the new view . Journal of Biochemical NMR, 5, 103–109. Bullitt, E., C.H.Jones, R.Striker, G.Soto, F.Jacob-Dubuisson, J.Pinkner, M.-J.Wick, L.Makowski, and S.J.Hultgren (1996). Development of pilus organelle subassemblies in vitro depends on chaperone uncapping of a beta zipper. Proc. Natl. Acad. Sci. USA , 93, 12890–12895. Cao, J., Khan, A.S., Bayer, M. and Schifferli, D.M. (1995). Ordered translocation of 987P fimbrial subunits through the outer membrane of Escherichia coli . Journal of Bacteriology , 177, 3704–3713. Clouthier, S.C., K.H.Muller, J.L.Doran, S.K.Collinson, and W.W.Kay (1993). Characterization of three fimbrial genes, sefABC, of Salmonella enteriditis . Journal of Bacteriology , 175, 2523–2533. Cosma, C.L., P.N.Danese, J.H.Carlson, T.J.Silhavy, and W.B.Snyder (1995). Mutational activation of the Cpx signal transduction pathway of Escherichia coli suppresses the toxicity conferred by certain envelope-associated stresses. Molecular Microbiology , 18, 491–505. Danese, P.N., W.B.Snyder, C.L.Cosma, L.J.B.Davis, and T.J.Silhavy (1995). The Cpx two-component signal transduction pathway of Escherichia coli regulates degP transcription. Genes and Development , 9, 387–398. de Graaf, F.K., P.Klemm, and W.Gaastra (1981). Purification, characterization and partial covalent structure of Escherchia coli adhesive antigen K99. Infection and Immunity , 33, 877–883. de Graaf, F.K. and F.R.Mooi (1986). The fimbrial adhesins of Escherichia coli . Advances in Microbial Physiology , 28, 65–143. Dodson, K.W., F.Jacob-Dubuisson, R.T.Striker, and S.J.Hultgren (1993). Outer membrane PapC usher discriminately recognizes periplasmic chaperone-pilus subunit complexes. Proc. Natl. Acad. Sci. USA , 90, 3670–3674. Duguid, J.P., S.Clegg, and M.I.Wilson (1979). The fimbrial and non-fimbrial haemagglutinins of Escherichia coli . Journal of Medical Microbiology , 12, 213–227. Duguid, J.P. and D.C.Old (1980) Adhesive properties of Enterobacteriacae , in E.H.Beachy, Editor, Bacterial Adherence Receptors and Recognition , Chapman & Hall: London, pp. 186–217. Duguid, J.P., I.W.Smith, G.Dempster, and P.N.Edmunds (1955). Non-flagellar filamentous appendages (“fimbriae”). and hemagglutinating activity in bacterium coli . Journal of Pathological Bacteriology , 70, 335–348. Edwards, R.A., Cao, J. and Schifferli, D.M. (1996). Identification of major and minor chaperone proteins involved in the export of 987P fimbriae. Journal of Bacteriology , 178, 3426–3433. Fersht, A.R. (1995). Optimization of rats of protein folding: The nucleation-condensation mechanism and its implications. Proc. Natl. Acad. Sci. USA , 92, 10869–10873. Garcia, M.I., A.Labigne, and C.LeBouguenec (1994). Nucleotide sequence of the
Molecular chaperones and folding catalysts
742
afimbrial-adhesin-encoding afa-3 gene cluster and its translocation via flanking IS 1 insertion sequences. Journal of Bacteriology , 176, 7601–7613. Goldhar, J., R.Perry, J.R.Golecki, H.Hoschutzky, B.Jann, and K.Jann (1987). Nonfimbrial, mannose-resistant adhesins from uropathogenic Escherichia coli O83: K1: H4 and O14: K?: H11. Infection and Immunity , 55, 1837–1842. Gong, M. and L.Makowski (1992). Helical Structure of Pap Adhesion Pili from Escherichia Coli . Journal of Molecular Biology , 228, 735–742. Hamada, D., S.Segawa, and Y.Goto (1996). Non-native a-helical intermediate in the refolding of -lactoglobulin, a predominantly b-sheet protein. Nature Structural Biology , 3, 868–873. Holmgren, A., M.J.Kuehn, C.-I.Brändén, and S.J.Hultgren (1992). Conserved imunoglobulin-like features in a family of periplasmic pilus chaperones in bacteria. EMBO Journal , 11, 1617–1622. Hull, R.A., R.E.Gill, P.Hsu, B.H.Minshaw, and S.Falkow (1981). Construction and expression of recombinant plasmids encoding type 1 and D-mannose-resistant pili from a urinary tract infection Escherichia coli isolate. Infection and Immunity , 33, 933–938. Hull, R.A., S.I.Hull, and S.Falkow (1984). Frequency of gene sequences necessary for pyelonephritis-associated pili expression among isolates of Enterobacteriaceae from human extraintestinal infections. Infection and Immunity , 43, 1064–1067. Hultgren, S.J., S.N.Abraham, M.G.Caparon, P.Falk, J.W. St. Geme III, and S.Normark (1993a). Pilus and non-pilus bacterial adhesins: assembly and function in cell recognition. Cell , 73, 887–901. Hultgren, S.J., F.Jacob-Dubuisson, C.H.Jones, and C.-I.Brāndén (1993b). PapD and superfamily of periplasmic immunoglobulin-like pilus chaperones , in G.Lorimer, Editor, Advances in Protein Chemistry , Academic Press: San Diego, pp. 99–123. Hultgren, S.J., C.H.Jones, and S.N.Normark (1996). Bacterial Adhesins and Their Assembly , in F.C. Neidhardt, Editor, Escherichia Coli and Salmonella; Cellular and Molecular Biology , ASM Press: Washington D.C., pp. 2730–2756. Hultgren, S.J., F.Lindberg, G.Magnusson, J.Kihlberg, J.M.Tennent, and S.Normark (1989). The PapG adhesin of uropathogenic Escherichia coli contains separate regions for receptor binding and for the incorporation into the pilus. Proc. Natl. Acad. Sci. USA , 86, 4357–4361. Hultgren, S.J. and S.Normark (1991a). Biogenesis of the bacterial pilus. Current Opinion in Genetics and Development , 1, 313–318. Hultgren, S.J., S.Normark, and S.N.Abraham (1991b). Chaperone-assisted assembly and molecular architecture of adhesive pili. Annual Review Of Microbiology , 45, 383–415. Hung, D.L., S.D.Knight, R.M.Woods, J.S.Pinkner, and S.J.Hultgren (1996). Molecular basis of two subfamilies of immunoglobulin-like chaperones. EMBO J. , 15, 3792– 3805. Iriarte, M., C.Vanooteghem, I.Delor, R.Diaz, S.Knutton, and G.R.Cornelis (1993). The Myf fibrillae of Yersinia enterocolitica . Molecular Microbiology , 9, 507–520. Itzhaki, L.S., D.E.Otzen, and A.R.Fersht (1995). The structure of the transition state for folding of chymotrypsin inhibitor 2 analyzed by protein engineering methods: Evidence for a nucleation-condensation mechanism for protein folding. Journal of
Structure and function of periplasmic
743
Molecular Biology , 254, 260–288. Jacob-Dubuisson, F., J.Heuser, K.Dodson, S.Normark, and S.J.Hultgren (1993). Initiation of assembly and association of the structural elements of a bacterial pilus depend on two specialized tip proteins. EMBO Journal , 12, 837–847. Johanson, I., R.Lindstedt, and C.Svanborg (1992). Roles of the pap- and prs-encoded adhesins in Escherichia coli to human uroepithelial cells. Infection and Immunity , 60 , 3416–3422. Jones, C.H., Danese, P.N., Pinkner, J.S., Silhavy, T.J. and Hultgren, S.J. (1997). The chaperone-assisted membrane release and folding pathway is sensed by two signal transduction systems. EMBO J. , 16 , 6394–6406. Jones, C.H., J.S.Pinkner, A.V.Nicholes, L.N.Slonim, S.N.Abraham, and S.J.Hultgren (1993). FimC is a periplasmic PapD-like chaperone that directs assembly of type 1 pili in bacteria. Proc. Natl Acad. Sci. USA , 90, 8397–8401. Kang, P.J., J.Ostermann, J.Shilling, W.Neupert, E.A.Craig, and N.Pfanner (1990). Requirement for hsp70 in the mitochondrial matrix for translocation and folding of precursor proteins. Nature , 348, 137–143. Karlyshev, A.V., E.E.Galyov, O.Y.Smirnov, A.P.Guzayev, V.M.Abramov, and V.P.Zav’yalov (1992). A new gene of the f1 operon of Y. pestis involved in the capsule biogenesis. FEBS Letters , 297 , 77–80. Kelly, J.M. (1996). Alternative conformations of amloidogenic proteins govern their behavior. Current Opinion in Structural Biology , 6, 11–17. Knight, S., M.Mulvey, and J.Pinkner (1997). Crystallization and preliminary X-ray diffraction studies of the FimH-FimC chaperone-adhesin complex from Escherichia coli. Acta Cryst D53 . 207–210. Knutton, S., M.M.McConnel, B.Rowe, and A.S.McNeish (1989). Adhesion and ultrastructural properties of human enterotoxigenic Escherichia coli producing colonization factor antigens III and IV. Infection and Immunity , 57, 3364–3371. Kraulis, P.J. (1991). MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures. Journal of Applied Crystallography , 24, 946–950. Kuehn, M.J., J.Heuser, S.Normark, and S.J.Hultgren (1992). P pili in uropathogenic E. coli are composite fibers with distinct fibrillar adhesive tips. Nature , 356 , 252–255. Kuehn, M.J., S.Normark., and S.J.Hultgren (1991). Immunoglobulin-like PapD chaperone caps and uncaps interactive surfaces of nascently translocated pilus subunits, Proc. Natl. Acad. Sci. USA , 88, 10586–10590. Kuehn, M.J., D.J.Ogg, J.Kihlberg, L.N.Slonim, K.Flemmer, T.Bergfors, and S.J.Hultgren (1993). Structural Basis of Pilus Subunit Recognition by the PapD Chaperone. Science , 262 , 1234–1241. Le Bouguenec, C., M.I.Garcia, Ouin, V., J.-M.Desperrier, P.Gounon, and A.Labigne (1993). Characterization of plasmid bourne afa-3 gene clusters encoding afimbrial adhesins expressed by Escherichia colistrains associated with intestinal or urinary tract infections. Infection and Immunity , 61 , 5106. Levine, M.M., P.Ristaino, G.Marley, C.Smyth, S.Knutton, E.Boedeker, R.Black, C.Young, M.L. Clements, C.Cheney, and R.Patnaik (1984). Coli surface antigens 1 and 3 of colonization factor antigen II-positive enteritoxigenic Escherichia coli: morphology, purification, and immune responses in humans. Infection and Immunity ,
Molecular chaperones and folding catalysts
744
44, 409–429. Lindberg, F., J.M.Tennent, S.J.Hultgren, B.Lund, and S.Normark (1989). PapD, a periplasmic transport protein in P-pilus biogenesis, Journal of Bacteriology , 171 , 6052–6058. Lindberg, F.P., B.Lund, and S.Normark (1986). Gene products specifying adhesion of uropathogenic Escherichia coli are minor components of pili. Proc. Natl. Acad. Sci. USA , 83, 1891–1895. Lindler, L.E. and B.D.Tall (1993). Yersinia pestis pH 6 antigen forms fimbriae and is induced by intracellular association with macrophages. Molecular Microbiology , 8, 311–324. Lund, B., F.Lindberg, B.I.Marklund, and S.Normark (1987). The PapG protein is the alpha-D-galactopyranosyl-(1–4)- -D-galactopyranose-binding adhesin of uropathogenic Escherichia coli . Proc. Natl. Acad. Sci. USA , 84, 5898–902. Massad, G., Fulkerson, J., J.F., Watson, D.C. and Mobley, H.L.T. (1996). Proteus mirabilis ambient-temperature fimbriae: cloning and nucleotide sequence of the aftgene cluster. Infection and Immmunity , 64, 4390–4395. Massad, G., Lockatell, C.V., Johnson, D.E. and Mobley, H.L.T. (1994). Proteus mirabilis fimbriae: construction of an isogenic pmfA mutant and analysis of virulence in a CBA mouse model of ascending urinary tract infection. Infection and Immmunity , 62 , 536– 542. Mecsas, J., P.E.Rouviere, J.W.Erickson, T.J.Donohue, and C.A.Gross (1993). The acivity of SE, an Escherichia coli heat-inducible a factor, is modulated by outer membrane proteins . Genes and Development , 7, 2618–2628. Merritt, E.A. and M.E.P.Murphy (1994). Raster3D Version 2.0—A program for photorealistic molecular graphics. Acta Crystallography , D50 , 869–873. Minor, D.L.J. and P.S.Kim (1994). Context is a major determinant of beta-sheet propensity. Nature , 371, 264–267. Muller, K.H., S.K.Collinson, T.J.Trust, and W.W.Kay (1991). Type 1 fimbriae of Salmonella enteriditis . Journal of Bacteriology , 173, 454–457. Nicchitta, C.V. and G.Blobel (1993). Lumenal proteins of the mammalian endoplasmic reticulum are required to complete protein translocation. Cell , 73, 989–998. Nilsson, P. and B.E.Uhlin (1991). Differential decay of a polycistronic Escherichia coli transcript is initiated by RNaseE-dependent endonucleolytic processing. Molecular Microbiology , 5, 1791–1799. Normark, S., S.Hultgren, B.I.Marklund, N.Stromberg, and J.Tennent (1988). Biogenesis of pili adhesins associated with urinary tract infectious Escherichia coli. Antonie Van Leeuwenhoek , 54, 405–409. Read, T.D., Dowdell, M., Satola, S.W. and Farley, M.M. (1996). Duplication of pilus gene complexes of Haemophilus influenzae biogroup aegyptius. Journal of Bacteriology , 178, 6564–6570. Roberts, J.A., B.Kaack, G.Kallenius, R.Mollby, J.Winberg, and S.B.Svenson (1984). Receptors for pyelonephritogenic Escherichia coli in primates. Journal of Urology , 131, 163–168. Roberts, J.A., B.-I.Marklund, D.Ilver, D.Haslam, M.B.Kaack, G.Baskin, M.Louis, R.Mollby, J.Winberg, and S.Normark (1994). The Gal (1–4). Gal-specific tip
Structure and function of periplasmic
745
adhesin of Escherichia coli P-fimbriae is needed for pyelonephritis to occur in the normal urinary tract. Proc. Natl. Acad. Sci. USA , 91, 11889–11893. Ryu, S.E., P.D.Kwong, A.Truneh, T.G.Porter, J.Arthos, M.Rosenberg, X.Dai, N.H.Xuong, R.Axel, R.W.Sweet, and W.A.Hendrickson (1990). Crystal structure of an HIV-binding recombinant fragment of human CD4. Nature , 348, 419–426. Saulino, E.T., Thanassi, D.G., Pinkner, J. and Hultgren, S.J. (1998). Ramifications of kinetic partitioning on usher-mediated pilus biogenesis. EMBO J. , in press. Savarino, S., P.Fox, D.Yikang, and J.P.Nataro (1994). Identification and characterization of a gene cluster mediating enteroaggregative Escherichia coli aggregative adherence fimbriae I biogenesis. Journal of Bacteriology , 176, 4949–4957. Schneider, H.-C., J.Berthold, M.F.Bauer, K.Dietmeier, B.Guiard, M.Brunner, and W.Neupert (1994). Mitochondrial Hsp70/MIM44 complex facilitates protein import. Nature , 371, 768–774. Shakhnovich, E.I. (1996). Modeling protein folding: the beauty and power of simplicity. Folding and Design , 1, R50-R54. Shapiro, L., A.M.Fannon, P.D.Kwong, A.Thompson, M.S.Lehmann, G.Grubel, J.F.Legrand, J. Als-Nielsen, D.R.Colman, and W.A.Hendrickson (1995). Structural basis of cell-cell adhesion by cadherins. Nature , 374, 327–337. Snyder, W.B. and T.J.Silhavy (1992). Enhanced export of -galactosidase fusion proteins on prlF mutants is lon dependent. Journal of Bacteriology , 174, 5661–5668. Strauch, K.L. and J.Beckwith (1988). An Escherichia commutation preventing degradation of abnormal periplasmic proteins. Proc. Natl. Acad. Sci. USA , 85, 1576– 1580. Strauch, K.L., K.Johnson, and J.Beckwith (1989). Characterization of degP, a gene required for proteolysis in the cell envelope Escherichia coli at high temperature. Journal of Bacteriology , 171, 2689–2696. Stromberg, N., S.J.Hultgren, D.G.Russell, and S.Normark (1992). Microbial Attachment, Molecular Mechanisms , in J.Lederberg, (ed.), Encyclopedia of Microbiology , Academic Press: New York, pp. 1–16. Svanborg-Eden, C., B.Andersson, L.Hagberg, L.A.Hanson, H.Leffler, G.Magnusson, G.Noori, J. Dahmen, and T.Soderstrom (1983). Receptor analogues and anti-pili antibodies as inhibitors of bacterial attachment in vivo and in vitro . Annals of the New York Academy of Sciences , 409, 580–591. Svanborg-Eden, C., B.Eriksson, L.A.Hanson, U.Jodal, B.Kaijser, G.Lidin-Janson, O.Lindberg, and S.Olling (1978). Adhesion to normal human uroepithelial cells of Escherichia coli from children with various forms of urinary tract infection. Journal of Pediatrics , 93, 398. Thanassi, D.G., Saulino, E.T., Lombardo, M.-J., Roth, R., Heuser, J. and Hultgren, S.J. (????). The PapC usher forms an oligomeric channel: Implications for pilus biognesis across the outer membrane. Proceedings of the National Academy of Sciences , 95, 3146–3151. Ungermann, C., W.Neupert, and D.M.Cyr (1994). The role of Hsp70 in conferring unidirectionality on protein translocation into mitochondria. Science , 266 , 1250– 1253.
Molecular chaperones and folding catalysts
746
Wang, J., Y.Yan, T.P.Garret, J.Liu, D.W.Rodgers, R.L.Garlick, G.E.Tarr, Y.Husain, E.L.Reinherz, and S.C.Harrison (1990). Atomic structure of a fragment of human CD4 containing two immu-noglobulin-like domains. Nature , 348, 411–418. Wolf, M.K., G.P.Andrews, B.D.Tall, M.M.McConnell, M.M.Levine, and E.G.Boedeker (1989). Characterization of CS4 and CS6 antigenic components of PCF8775, a putative colonization factor complex from enterotoxigenic Escherichia coli E8775. Infection and Immunity , 57, 164–173. Xu, Z., C.H.Jones, D.Haslam, K.Dodson, J.Kihlberg, and S.J.Hultgren (1995). Molecular dissection of PapD interaction with PapG reveals two chaperone binding sites. Molecular Microbiology , 16 , 1011–1020.
INDEX Acquired stress tolerance 137 Actin 614, 617, 619, 620, 631, 676, 701 Actin microfilament dynamics 72 Adhesin 719, 726 Alpha A-crystallin 191 Alpha B-crystallin 191 Alpha-casein 593 Amyloidose 4 Anfinsen 203, 573 AP180 379 AP1 379 AP2 379, 382 Apoptosis 100 Archaeoglobus fulgidus 141 Arsenite tolerance 409 ASPK2 67 Auxilin 382 –6 Auxilin, mechanism of action 384 Auxilin/Hsc70 interaction 383 Bacillus subtilis 34, 142 Bag 1 9 Beta-galactosidase 701 Blood coagulation factor XIIIa 466 Bovine pancreatic trypsin inhibitor (BPTI) 469, 502 Bradyrhizobium japonicum 34, 142, 324 Brownian ratchet 278 –9 Cad1 175 Cadherin 725 Cadmium tolerance 409 Calmodulin 700 Calnexin 7, 249, 693 Calreticulin 7, 249 Capsular polysaccharide biosynthetic (cps) gene 120, 122 Carboxypeptidase Y (CPY) 250 Casein kinase II 700, 701 Castanospermine (CST) 246 Catabolite regulatory protein (CRP) 448 Caulobacter crescentus 142 CbpA 674
Index
748
CCT 7, 96, 207, 597, 610, 616–8, 620, 621 CCT co-chaperones 615 CCT functional cycle 618 CCT gene family 611 CCT-mediated folding 616 CDR3 730 Cell survival after stress 69 Chaperonin 7, 8 Chaperonin 10 (Cpn 10) 305 Chaperonin 60 (Cpn 60) 305 Charonin 436 Chloramphenciol acetyl transferase (CAT) 217 Chloroplast chaperonins 305 CHOP 73 CHOP (C/EBP homologous protein) 55, 75 Chymotrypsin inhibitor 2 (CI-2) 474 CIRCE (controlling inverted repeat for chaperone expression) 34 Citrate synthase (CS) 701 C-Jun 73 C-Jun kinase 66 Clathrin coated vesicles 378 Clathrin triskelion 379 Clostridium acetobutylicum 34 ClpA 401, 435, 449, 440, 441 ClpA activation of RepA 442 ClpAP 438 ClpB 16 ClpB/Hsp104 435, 443 ClpP 16, 401, 438, 440 ClpQ 440 ClpX 16, 401, 438, 440, 441 ClpX functions in phage replication and transposition 442 ClpY (HtpI, HslU) 17, 440 Cockayne syndrome 68 Collagen 694 COM70 303 Constitutive heat shock element binding factor (CHBF) 102 Cpx 33 Creatinase 466 Creutzfeldt-Jacob disease 415 CsA 532 CSAID 72 CSBP2 (cytokine-suppressive anti-inflammatory drug—CSAID binding protein) 66 C-type cytochrome 324 Cyclic-AMP response element binding protein, CREB 73 Cyclophilin (CyP) 21, 240, 487, 531, 532, 534, 536, 540 Cyclophilin 40 (Cyp40) 362, 366, 693 Cyclosporin A (CsA) 325, 527
Index
749
Cyp18 528, 532 Cytochrome c 472 D. melanogaster 46 DegP (HtrA) 18, 33, 731 Degradation motifs 448 Deoxynojirimycin (dNM) 246 D-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) 509 Dihydrofolate reductase (DHFR) 269, 300, 592, 701 Dj1 A 122 DnaB 340 DnaC 340 DnaJ 6, 7, 14, 16, 52, 65, 119, 120, 172, 174, 211, 334, 344, 361, 449, 450, 656, 667, 669, 671 DnaK 16, 119, 120, 211, 334, 344, 435, 631, 655, 666, 670 DnaK structure 666 DnaK-DnaJ 14, 52, 65, 449, 450, 655, 667, 668, 669, 671 DnaK-DnaJ-GrpE 23, 666, 667, 672, 673, 675 DREA motif 724, 725 DRES motif 724 DsbA 5, 32, 320–2, 498–500, 503 DsbB 320 –2 DsbC 5, 320 –2 DsbD 320 –2 DsbE 321, 323 DsbG 321 EJ54 15 Embryogenesis 187 Encapsulation model 572 Endoplasmic reticulum (ER) 32, 55, 75, 211, 232, 270, 447, 466, 693, 696 ERp57 499 Escherichia coli 14, 65, 124, 141, 319, 324, 334, 401 EsE 32 Ethanol tolerance 409 Eukaryotic elongation factor, eEF-1 alpha 76 F1-G1 loop 729 Fatal familial insomnia 415 Ferredoxin 299 FGL chaperone 729, 730 FGS chaperone 729, 730 Fibroblast growth factor (FGF) 73 FimC 729 FK 566 531, 532 FK506 binding protein (FKBP) 326, 487, 526, 536 FKBP12 532 FKBP51 362, 366 FKBP52 362, 366, 693
Index
750
FkpA 32, 325 Fluid phase pinocytosis 72 FtsH 439, 441, 444 FtsJ 16 GapA 16 Gerstmann-Straussler-Schienker syndrome 415 Glucocorticoid receptor (GR) 361 Glucose regulated protein (Grp) 250 Glutamate dehydrogenase 466 Glutaredoxin 498 Glutathione-S-transferase (GST) 497 Glycosylation site binding protein 509 Gp31 9 GroEL 7, 8, 15, 52, 65, 205, 211, 426, 435, 450, 554, 582, 588, 615, 621 GroEL ATPase co-operativity 571 GroEL ATPase cycle 563, 564, 570 GroEL nucleotide binding 584 GroEL structure 582, 583, 585, 618 GroES 9, 17, 52, 435, 554, 588, 589, 612, 615 Growth factor-activated kinase 65 Growth factors 44 GRP gene transcription 75 Grp170 251 GRP78 187 Grp94 251, 361 GrpE 6, 9, 16, 52, 119, 120, 334, 344, 435, 656 Haemophilus influenzae 141 Hdj-1 52 Heat shock cognate (Hsc70) 216 Heat shock element (HSE) 45 Heat shock element (HSE) dependent transcription 74 Heat shock proteins and apoptosis 100 Heat shock proteins and RNA splicing 100 Heat shock response (hsr) 13, 65 Heat shock transcription factor (HSF) 45, 74, 92, 423 Heme-regulated inhibitor (HRI) 77 Hemin 52 Hemin-controlled eIF-2alpha kinase (HCR) 77 Hepatocyte growth factor (HGF) 683 Hexokinase 614 HflB 17, 23 Hip 9, 219, 364, 703 Hip (Hsp70 interacting protein) 218, 361, 656 HIV-1 533 Hop 362, 366, 703 HrtC 15
Index
751
HSBP1 (heat shock factor binding protein) 50 Hsc20 122 Hsc66 122 Hsc70 96, 188, 216, 234, 378, 380, 386, 629, 631, 632 HSF1 47, 48, 49 HSF2 48, 50, 52, 192 HSF3 54 HSF4 54 HslA 16 HslC 16 HslE 18 HslF 18 HslG 18 HslI (HtpH) 18 HslJ 18 HslK 16 HslM 18 HslO 17 HslP 17 HslQ 18 HslR 18 HslV (HtpO) 17 HslW 17 HslX 17 HsIY 15 HslZ 17 Hsp10 6 Hsp20 141 Hsp23 187 Hsp25 698 Hsp27 72, 95, 101 Hsp30 187 Hsp40 9, 96, 234, 361, 363, 364, 382, 435, 655, 666, 667, 671 Hsp47 6, 8, 693 Hsp47 chaperone function 696 Hsp47 interactin with collagen 694 Hsp47 molecular properties 693 Hsp60 see GroEL Hsp70 6, 7–, 46, 49, 51, 52, 92, 95, 138, 160, 187, 212, 269, 270, 299, 303, 360, 361, 363, 364, 378, 450, 614, 629, 630, 631, 648, 649, 654, 656, 665–7, 671 Hsp78 138, 401, 443 Hsp83 187 Hsp90 6, 8, 46, 51, 52, 187, 190, 360, 361, 364, 532 Hsp90 chaperone action 701 Hsp90 molecular properties 699 Hsp100 6 Hsp Phosphorylation 70 Hsp100/Clp family 8, 400, 440
Index Hsp104 8, 138, 397, 400, 425, 440, 443 Hsp104 biochemical properties 402 Hsp104 and mammalian prion 424 Hsp104 and other tolerance factors 412 Hsp104 and PSI 417 Hsp104 and Sup35 421 Hsp104 thermotolerance 403, 407 Hsp110 167 HtpG 16 HtpG (Hsp90) 139, 367 HtpK 17 HtpT 18 HtpX 16 HtpY 16 HtrM (RfaD) 17 IbpA (HtpN, HslT) 17, 141 IbpB (HtpE, Hsls) 17, 141 ID-I 507 IEP100 302 IEP110 302 IEP44 302, 303 IMP (inner membrane protease) 268 Import motor model 278 –81 Influenza virus 78 Insulin 701 Jem1 175 JEP55 303 K. lactis 48 Koshland-Nemethy-Filmer (KNF) model 562, 585 Ku autoantigen 102 Ku70 103 Ku80 103 Ku80 in growth and V (D) J recombination 103 Kuru 415 Lactacystin 52 LamB 28, 33 Lambda DNA replication, uni-and bidirectional 347 Lambda O protein 443 Legionella pneumophila 327 LHCP 307 Lon 15, 437, 439, 444, 450 Luciferase 619, 701 LysU 19
752
Index
753
Mad cow disease 415 Major histocompatibility complex (MHC) 631 Maltose binding protein (MBP) 704 MAP kinase 65, 66 MAP kinase activated protein kinase 1 (MAPKAPK1) 67, 77 MAP kinase activated kinase 2 (MAPKAPK2) 66, 70 MAP kinase kinase (MEK or MKK) 65, 66 MAP kinase kinase kinase (MAPKKK or MEKK) 65, 66 MAP kinase kinase kinase kinase 66 MAP kinase represser 66 Methanococcus jannaschii 141 Methotrexate (MTX) 301 Mge1 176 Microsomal triglyceride transfer protein (MTP) 510 Mitochondrial import stimulation factor (MSF) 206, 213, 271, 300 Mitochondrial malate dehydrogenase (mMDH) 592 Mitochondrial processing peptidase (MPP) 276 MMDH binding 575 Molecular chaperone, definition 5 Monod-Wyman-Changeux (MWC) model 562 Mu transposase (MuA) 442, 449 MucA 29 MucB 29 Mycoplasma genitalium 141 Myogenic regulatory factors (MyoD) 190, 700, 701 Myosin 677 Myotonic dystrophy protein kinase, MDPK 68 Myxococcus xanthus 142 Nascent polypeptide associated complex (NAC) 206, 213, 233 N-Glycans in protein folding 246 NIMA (never in mitosis) 534 Nonpilus fiber 725 OEP34 302, 303 OEP70 302, 303 OEP75 302, 303 OEP86 302, 303 Oligosaccharyl transferase 236, 510 Organellar targeting factor 207 Ornithine carbamoyltransferase (pOTC) 271 Ornithine transcarbamylase (OTC) 593 O-some 336 OST48 245 Outer envelope membrane protein (OEP) 302 Outer membrane protein (OMP) 27, 324, 327
Index
754
Oxidative stress 44 P pili 720 P21-activated kinase, PAK 67 P23 362, 364, 693, 703 P38 MAP kinase 66, 71 P48 362 P53 701 PapA 720, 723, 726 PapC 720, 724 PapD 7, 9, 720, 722, 723, 728, 730, 731 –3 PapD pilus subunit recognition 721 PapD three dimensional structure 720 PapD-PapE complex 723 PapD-PapG complex 721, 722, 723, 733, 734 PapE 720, 722 PapF 720, 722 PapG 720, 722, 723, 726, 731, 732, 733, 734 PapH 722 PapJ 720 PapK 720, 722 Parvulin 528, 537 PBF (presequence binding factor) 269 PDI 5, 240, 243, 319, 321, 486–8, 494, 499, 500 PDI and thioredoxin 498 PDI catalysis 501 PDI protein sequences 495 PDI, domain organisation 496 PDI, molecular properties 495 Pdr13 168 PEPSPS 300 Peptide bond isomerization 522 Peptidyl prolyl cis-trans isomerases (PPI) 5, 252, 319, 325, 486–8, 521, 538 Pilus fiber 725 PKI proteolysis 497 Platelet-derived growth factor (PDGF) 683 PpiD 325, 327 Preprimosomal structure 339, 343 Presequence binding factor (PBF) 206, 271 Prion 4, 415 Pro39 535, 536 Pro55 535 Procollagen 694 Progesterone (PR) 361 Progesterone aporeceptor (PR) 701 Prolyl-4-hydroxylase (P4H) 511 ProOmpA 128 Protein disulfide isomerase see PDI
Index Protein kinase, HSP Interaction 68 Protochloro-phyllide oxidoreductase (pPOR) 299 PrP 425 PrPsc 415, 426 Pseudomonas aeruginosa 35 PSI 398, 415, 416 PSI and Sup35 416 Psp 15 PspA 18 Rapamycin 532 RcsA 121 RepA 442 Reverse transcriptase 701 Rhizobium meliloti 142 Rhodobacter capsulatus 324 Rhodobacter spheroides 582 Ribonuclease A (Rnase A) 524 Ribonucleoprotein particle (RNP) 233 Ribophorin I 245 Ribophorin II 245 Ribose binding protein (RBP) 129, 704 Ribosome receptor 235 Rnase T1 535, 536 Rnase T1 folding 536 RseA 29 RseB 29 S. pombe 69 Saccharomyces cerevisiae 46, 138, 160, 268, 398 Salmonella typhimurium 479 SAPK1 (stress-activated protein kinase) 66, 67 Scj1 175 Scrapie 415 Sec 71 236 Sec 72 236 Sec61 236 Sec62 236 Sec63 175, 236 SecA 210, 308 SecB 7, 130, 205, 210, 308, 693 Serine-threonine kinase 700 Sigma 32 14, 18, 23, 49, 119, 120, 137, 445 Sigma 32, translational regulation 19 Sigma 32-dependent gene expression 120 Sigma 70 16, 18 Sigma E 14, 18, 27, 32, 324 Sigma E, regulation 27
755
Index
756
Sigma S 14, 120, 137 Signal peptidase 235 Signal recognition particle (SRP) 205–7, 210, 211, 221, 233, 235, 236 Sis1 173 Skp 32 Small heat shock protein, (sHsp) 6, 8, 692, 693, 695, 696, 698 Spermatogenesis 192 SRP receptor (docking protein) 233, 235 Ssa 161, 269 Ssb 164 Ssc1p 269 Sse Hsp70 167 Starvation-induced stress tolerance 137 Steroid hormone receptor 360, 363, 700 Sti 1 362, 366 Stress response element (STRE) 74 Stress tolerance 137, 401 Subtilisin 593 Sulfolobus shibitae 597 Sulfolobus solfataricus 597 Sup35 398, 419 SurA 28, 32, 325 Tailspike protein (TSP) 478 T-cell antigen receptor (TCR) 105 Tetratricopeptide repeat (TPR) 656 Thermoplasma acidophilum 597 Thermosome 597 Thermosome structure 598, 599 Thermotolerance 90, 398, 403, 412 Thiol: disulfide exchange reaction chemistry 320, 501 Thioredoxin 321, 498, 499, 500 Thioredoxin reductase (TrxB) 322 TkpA 18 Tom complex (translocase of the outer membrane) 272 Tom20 272 Tom22 272 Tom37 272 Tom40 272 Tom5 272 Tom6 272 Tom7 272 Tom70 272 Transfer nuclear overhauser effect (trNOE) 594 Transmissible spongiform encephalopathy (TSE) 415 Trehalose synthase and Hsp 104 genetic interaction 414 TRiC 612, 619, 620
Index Trigger factor (TF) 127, 205, 209, 221, 487, 538 Trigger factor catalyzed folding 539 Trypanosoma cruzi 250 Tubulin 614, 617–8, 701 Tumor necrosis factor 67, 700 Tyrosine kinase 700 Ubiquitin 52 Unfolded protein response (UPR) 32, 55 Unfolding protein response element (UPRE) 55 URE3 415 Xdj1 176 Ydj-1 (DnaJ homologue) 52, 172, 214 Zuol 174
757
