Protein Export and Membrane Biogenesis, Volume 4 (Advances in Cellular and Molecular Biology of Membranes and Organelles)

ADVANCES IN CELL AND MOLECULAR BIOLOGY OF MEMBRANES AND ORGANELLES Volume 4

9 1995

PROTEIN EXPORT AND MEMBRANE BIOGENESIS

This Page Intentionally Left Blank

ADVANCES IN CELL AND MOLECULAR BIOLOGY OF MEMBRANES AND ORGANELLES PROTEIN EXPORT AND MEMBRANE BIOGENESIS

Series Editor: ALAN M. TARTAKOFF Institute of Pathology Case Western Reserve University Editor: ROSS E. DALBEY Department of Chemistry The Ohio State University

VOLUME 4

91995

@ Greenwich, Connecticut

JAI PRESS INC.

London, England

Copyright 91995 by JAI PRESSINC. 55 Old Post Road, No. 2 Greenwich, Connecticut 06836 JAI PRESSLTD. The Courtyard 28 High Street Hampton Hill, Middlesex TWI 2 1PD England All rights reserved. No part of this publication may be reproduced, stored on a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, filming, recording, or otherwise, without prior permission in writing from the publisher. ISBN: 1-55938-924-9

Transferred to Digital Printing 2006

T LIST OF CONTRIBUTORS

VII

INTRODUCTION TO THE SERIES

Alan M. Tartakoff

ix

PREFACE

Ross E. Dalbey

xi

MEMBRANE PROTEIN ASSEMBLY

Paul Whitley and Gunnar von Heijne

MEMBRANE INSERTION OF SMALL PROTEINS: EVOLUTIONARY AND FUNCTIONAL ASPECTS

Dorothee Kiefer and Andreas Kuhn

PROTEIN TRANSLOCATION GENETICS

Koreaki Ito

BIOCHEMICAL ANALYSES OF COMPONENTS COMPRISING THE PROTEIN TRANSLOCATION MACHINERY OF ESCHERICHIA COLI Shin-ichi Matsuyama and Shoji Mizushima

17

35

61

PIGMENT-PROTEIN COMPLEX ASSEMBLY IN RHODOBACTER SPHAEROIDESAND

RHODOBACTER CAPSULATUS Amy R. Varga and Samuel Kaplan

85

IDENTIFICATION AND RECONSTITUTION OF ANION EXCHANGE MECHANISMS IN BACTERIA

Atul Varadhachary and Peter C. Maloney

105

CONTENTS HELIX PACKING IN THE C-TERMINAL HALF OF LACTOSE PERMEASE H. Ronald Kaback, Kirsten]ung, Heinrich ]ung, ]ianhua Wu, Gilbert G. PrivY, and Kevin Zen

129

EXPORT AND ASSEMBLY OF OUTER MEMBRANE PROTEINS IN E. COL/ Jan Tommassenand Hans de Cock

145

STRUCTURE-FUNCTION RELATIONSHIPS IN THE MEMBRANE CHANNEL PORIN Georg E. Schulz

175

ROLE OF PHOSPHOLIPIDS IN ESCHERICHIACOL/ CELL FUNCTION William Dowhan

189

MECHANISM OF TRANSMEMBRANE SIGNALING IN OSMOREGULATION Arfaan A. Rampersaud

219

INDEX

263

LIST OF CONTRIBUTORS Hans de Cock

Department of Molecular Cell Biology Institute of Biomembranes Utrecht University Utrecht, The Netherlands

William Dowhan

Department of Biochemistry and, Molecular Biology The University of Texas-Houston Medical School Houston, Texas

Koreaki Ito

Department of Cell Biology Institute for Virus Research Kyoto University Kyoto, Japan

Heinrich Jung

Departments of Physiology and Microbiology and Molecular Genetics University of California Los Angeles, California

Kirsten Jung

Departments of Physiology and Microbiology and Molecular Genetics University of California Los Angeles, California

H. Ronald Kaback

Departments of Physiology and Microbiology and Molecular Genetics University of California Los Angeles, California

Samuel Kaplan

Department of Microbiology and Molecular Genetics University of Texas Health Science Center at Houston Houston, Texas oo

VII

viii

LIST OF CONTRIBUTORS

Dorothee Kiefer

Angewandte Mikrobiologie Universit~t Karlsruhe (TH) Karlsruhe, Germany

Andreas Kuhn

Angewandte Mikrobiologie Universit~t Karlsruhe (TH) Karlsruhe, Germany

Peter C. Maloney

Department of Physiology Johns Hopkins University School of Medicine Baltimore, Maryland

Shin-ichi Matsuyama

Institute of Molecular and Cellular Biosciences University of Tokyo Tokyo, Japan

Shoji Mizushima

Tokyo College of Pharmacy Tokyo, Japan

Gilbert G. Priv~

Departments of Physiology and Microbiology and Molecular Genetics University of California Los Angeles, California

Affaan A. Rampersaud

Department of Pathology The Ohio State University Columbus, Ohio

Geor8 E. Schulz

InstitOt for Organische Chemie und Biochmie A Ibert-Ludwigs-Universit~it Freiburg, Germany

Jan Tommassen

Department of Molecular Cell Biology Institute of Biomembranes Utrecht University Utrecht, The Netherlands

Atul Varadhachary

Department of Biological Chemistry Johns Hopkins University School of Medicine Baltimore, Maryland

Lbt of Contributors Amy R. Varga

Chemical and Agricultural Products Division Abbott Laboratories North Chicago, Illinois

Gunnar von Heijne

Department of Biochemistry Arrhenius Laboratory Stockholm University Stockholm, Sweden

Paul Whitley

Department of Biochemistry Arrhenius Laboratory Stockholm University Stockholm, Sweden

Jianhua Wu

Departments of Physiology and Microbiology and Molecular Genetics University of California Los Angeles, California

Kevin Zen

Departments of Physiology and Microbiology and Molecular Genetics University of California Los Angeles, California

This Page Intentionally Left Blank

INTRODUCTION TO THE SERIES The remarkable vigor and central importance of cell biology result from the realization that emphasis on structure/function relations at the cellular and subcellular levels is essential for a rigorous and satisfactorily complete understanding. Unlike many subdivisions of biomedical science, cell biology is not linked to any one methodology. It often emphasizes topological or topographic questions, and it is concerned with the structure, biogenesis, and turnover of macromolecular structures; however, there is no limit to the techniques and conceptual approaches which it brings to bear on these issues. Indeed, this has even been true since the term cell biology was first used. Certain scientists trace its origins back to ultrastructure and histology; others consider E.B. Wilson's The Cell in Development and Heredity to epitomize the foundations of cell biology; while others consider cell biology as an outgrowth of somatic cell genetics or the extension of biophysics to objects of increasingly large size. This varied and often changing identity makes it easier to specify what cell biology is not, rather than to describe it in a positive sense. It also suggests the rich intellectual mix which underlies today's successes in cell biology research. This series aims to match the continuing evolution of cell biology---with particular emphasis on cell membranes--by treating coherent areas in multi authored volumes This approach allows a multifaceted coverage of topics which benefits from the unity of vision of the volume editor(s), but does not rely on one individual to synthesize an entire subject. Hopefully, the result will be more readable and intrinsically richer than lengthy review chapters which attempt to be all encompassing. Alan M. Tartakoff Series Editor xi

This Page Intentionally Left Blank

PREFACE The incentive for putting together Volume 4 of this series was to review the wealth of new information that has become available in prokaryotic organisms in protein export and membrane biogenesis. Just in the last several years, protein translocation has now been efficiently reconstituted using defined components and the mechanism by which proteins are moved across membrane bilayers is now being examined at a higher resolution. In addition, because of a new technical breakthrough using osmolytes, it is now possible to reconstitute a number of channel proteins, ATPase, receptors, and transporters. In many cases, it is possible to successfully predict the membrane topology of these types of proteins using both "hydrophobicity analysis" and the "positive inside" rule. In this volume, two chapters focus on protein translocation across membranes

(Biochemical Analyses of Components Comprising the Protein Translocation Machinery orE. coil; Protein Translocation Genetics), while several others on how proteins assemble into the inner membrane of E. coil (Membrane Protein Assembly, Membrane Insertion of Small Proteins: Evolutionary and Functional Aspects; Pigment-Protein Complex Assembly in Rhodobacter sphaeroides and Rhodobac. ter capsulatus). Other sections review recent progress on transporters (Identification and Reconstitution of Anion Exchange Mechanisms in Bacteria; Helix Packing in the C-Terminal Half of Lactose Permease) and signal transduction (Mechanism of Transmembrane Signaling in Osmoregulation) as well as the assembly of porins into the outer membrane (Export and Assembly of Outer Membrane Proteins in E. coil) and their structures (Structure-Function Relationships in the Membrane Channel Porin). Although the emphasis of the book is on proteins, the role of xiii

xiv

PREFACE

phospholipids in controlling various cell surface processes is reviewed (Role of Phospholipids in Escherichia coil Cell Function). I should point out the reason for the rapid progress in bacteria research is because of the possibility to apply biochemistry and genetics to this organism. I would like to thank the contributors to this volume who gave their time and expertise and who made this task enjoyable. Ross E. Dalbey

Editor

MEMBRANE PROTEIN ASSEMBLY

Paul Whitley and Gunnar yon Heijne

I. II. III.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Membrane Protein Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . Membrane Protein Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Experimental Determination of Topology . . . . . . . . . . . . . . . . . . B. Prediction and Experimental Manipulation of Topology . . . . . . . . . . . IV. Helix Packing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Mechanism of Membrane Insertion . . . . . . . . . . . . . . . . . . . . . . . . A. Translocation of Periplasmic Loops . . . . . . . . . . . . . . . . . . . . . . B. Insertion of Transmembrane Segments . . . . . . . . . . . . . . . . . . . . VI. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 2 3 3 4 5 6 6 10 11

I. I N T R O D U C T I O N Cells are made up of multip!e membrane-bound compartments, each containing a unique set of proteins. The sorting and insertion of proteins across and into the many membrane systems of eukaryotes and prokaryotes has been the subject of a vast amount of research. In this chapter we will deal with only one area of membrane

Advances in Cell and Molecular Biology of Membranes and Organelles Volume 4, pages 1-16. Copyright 9 1995 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-924-9

PAUL WHITLEY and GUNNAR VON HEIJNE

protein biogenesis, namely, the assembly of proteins into the cytoplasmic (inner) membrane of prokaryotes, notably Escherichia coll. We believe that many of the basic aspects of membrane protein assembly in E. coli are relevant to the assembly of proteins into other membrane systems such as the endoplasmic reticular (ER) membrane of eukaryotes, the thylakoid membrane of chloroplasts, and the inner mitochondrial membrane. Knowledge from E. coil can often illuminate the analogous processes in these other systems. As a background to the discussion on the membrane assembly processes per se, we will first give a brief introduction to membrane protein structure. This is followed by sections on membrane protein topology and helix packing in a lipid environment. Finally, the mechanisms of membrane insertion in E. coil are reviewed.

II. MEMBRANE PROTEIN STRUCTURE Although membrane proteins have diverse functional roles in the cell such as ion channels, transporters, and receptors, they appear to be limited in the types of structure they can adopt, due to the constraints imposed by the hydrophobic environment of their membrane-spanning regions. The hydrophobic interior of the lipid bilayer has for a long time been considered as energetically unfavorable for peptide bonds unless the polar groups are hydrogen bonded (32). Furthermore, amino acids with polar and potentially charged side chains are not favored energetically in a hydrophobic environment. Thus a-helices built mainly from amino acids with nonpolar side chains are structures likely to be found in membrane-spanning regions. Spectroscopic studies on a number of membrane proteins indeed have provided evidence that they are comprised largely of a-helical structure (34, 97). However, there is an embarrassing lack of high-resolution structural information for membrane proteins to confirm existing theories. The X-ray structure of the photosynthetic reaction center from Rhodopseudomonas viridis (25) and the structure of bacteriorhodopsin from Halobacterium halobium derived from electron microscopy studies (43), are among the few exceptions that confirm the predicted a-helical nature of membrane-spanning segments. a-Helices are not the only structures that can span the lipid bilayer. The crystal structures of bacterial outer membrane porins have recently been solved and shown to consist of large 16-stranded 13-barrelstructures in the membrane-spanning region (21, 98). Such closed ~-sheets fulfill the hydrogen-bonding criterion but do not possess the long hydrophobic stretches of amino acids characteristic of the helical bundle membrane proteins. Very little is known about the mechanism of insertion of the ~-barrel t y ~ membrane proteins. We therefore limit the following discussion to the assembly of proteins of the a-helical variety.

Membrane Protein Assembly Due to the difficulties in obtaining high-resolution structural information for membrane proteins by classical structural techniques, such as X-ray crystallography and NMR, it is necessary to turn to other low resolution techniques for looking at their structures. In the following sections we will discuss some of the theoretical and experimental methods that have been used to map or predict the topology of membrane proteins in the cytoplasmic membrane of bacteria.

III. MEMBRANE PROTEIN TOPOLOGY A. ExperimentalDetermination of Topology In order to establish accurately the topology of a membrane protein, we would ultimately like to know for each amino acid whether or not it is buried in the lipid bilayer, and towards which side of the lipid bilayer it is situated. One way to determine this is by so-called vectorial-labeling methods (46). Vectorial labeling requires a sealed membrane preparation of known orientation containing the protein of interest and a labeling reagent that only has access to one side of the membrane. Only those reactive amino acid residues that are accessible to the labeling reagent will be modified, and by analyzing the pattern of modification, their membrane disposition can be determined. Modifications can range from the covalent attachment of small probes, such as fluorescent, radioactive, and spin-labeled molecules, to binding of antibodies or proteolytic cleavages. The most frequently used methods to determine membrane protein topology in E. coli, however, rely on gene-fusion technology to produce hybrid proteins. Fusions to the reporter proteins PhoA (alkaline phosphatase), LacZ ([3-galactosidase), and 13-1actamase are the most commonly used. All of these fusions are useful in determining the topology of membrane proteins due to different activities of these proteins when located in the cytoplasm or periplasm. For example, PhoA is only active when it is periplasmic (68), and PhoA fusions to a portion of a membrane protein that is normally periplasmic likewise have a high activity. If fused to a portion of the protein that is normally cytoplasmic, the PhoA activity will be low (65). Another feature of PhoA that makes it useful in topological studies is the high protease resistance when situated in the periplasm, but not when located in the cytoplasm (79). The case is similar for 13-1actamase fusions that confer ampicillin resistance and give high [3-1actamase activity only when the [3-1actamase moiety is present in the periplasm (16). LacZ fusions, in contrast, are only active when fused to cytoplasmic domains (36, 38). Results from fusion protein analysis should, however, be treated with some caution. PhoA fusions may in certain situations show substantial activity, even when fused to normally cytoplasmic regions of the polypeptide chain. This phenomenon was observed when PhoA was fused to the beginning of proposed cytoplasmic domains of MalF (36), presumably because these fusions were translocated, albeit

PAUL WHITLEY and GUNNAR VON HEIJNE

slowly, across the cytoplasmic membrane (86). Fusions towards the C-terminal ends of the cytoplasmic and periplasmic regions are thus to be preferred (15). LacZ fusions sometimes exhibit low activity for reasons other than being attached to an export signal; for example, if the fusion is out of frame, there is often a low but detectable ~-galactosidase activity (41). Nevertheless, a combination of LacZ with PhoA and/or I~-lactamase fusions at different places within a polytopic membrane protein has been used successfully to determine the topology of many inner membrane proteins in E. coil [see (95) for a list of examples].

B. Prediction and Experimental Manipulation of Topology Although the powerful fusion protein techniques make topological mappings relatively straightforward, much can be deemed from the primary amino acid sequence. Apolar membrane-spanning helices are often easy to predict using hydrophobicity analysis. Essentially, a window of 15-20 residues is moved along the sequence, and the average hydrophobicity of. the residues inside the window is calculated for each position. Peaks on the hydrophobicity plot above a certain cutoff level correspond to candidate membrane-spanning segments and allow a tentative topological model that can serve as a basis for planning further fusion protein studies. However, although many transmembrane segments have a very high average hydrophobicity and are easy to spot, others have a more moderate hydrophobicity and can be difficult to distinguish from nonmembrane segments. In such cases, a considerably improved prediction can be obtained if the so-called positive-inside rule is taken into account. This rule is based on the observation that nontranslocated segments of inner membrane proteins have a 3- to 4-fold higher content of positively charged lysines and arginines than translocated segments (91) and that any predicted topology should thus be expected to show a similar bias. Ifa particular stretch is of moderate hydrophobicity, and thus difficult to assign unambiguously as a transmembrane segment, two tentative topologies may be constructed: one with and one without the doubtful transmembrane segment. The one with the highest Lys+Arg bias is chosen as the best prediction (95). In practice, this simple recipe gives a substantial improvement in prediction accuracy: For bacterial inner membrane proteins, one can expect to predict the correct topology about 9 times out of 10. Even more importantly, experimental studies have shown that the topology of an inner membrane protein can be manipulated by simply redistributing positively charged residues from one end of a transmembrane stretch to the other. Such topology-engineering was first demonstrated for the leader peptidase enzyme (93), where, in certain constructs, as few as one lysine residue can make the difference between Nin and Noet orientations (70). How could the positively charged residues influence topology? One possibility is that the electrical potential across the membrane (positive periplasmic) works

Membrane Protein Assembly

against the translocation of positively charged residues. Other possibilities are that negatively charged phospholipid head groups on the cytoplasmic face of the inner membrane may interact with the positively charged amino acid residues or that proteins of the secretory machinery such as SecA may bind preferentially to the positively charged end of a transmembrane segment (1).

IV. HELIX PACKING The actual membrane insertion process appears to follow a two-stage model (75), where independently stable a-helices are first inserted across the lipid bilayer to determine the topology of the protein. In the next stage, the membrane-spanning ct-helices pack together to form functional transmembrane structures (helix bundies). From the available data, it appears that interactions between transmembrane ct-helices can contribute substantially towards structural stabilization, and in some cases, the oligomerization of membrane proteins (20, 55, 66). Evidence for the importance of interactions between pre-inserted membranespanning helices that supports the two-stage model of protein folding comes from the observations that helix-containing fragments of membrane proteins can reassemble into native-like structures. B acteriorhodopsin can be cleaved by chymotrypsin to give two fragments containing 5 and 2 transmembrane helices, respectively. These two fragments, when reconstituted separately into lipid vesicles, were shown by circular dichroism to be ct-helical. When the two populations of vesicles were fused, the resulting protein complex acquired the ability to bind retinal and regained the spectroscopic properties of native bacteriorhodopsin (74). In fact, the same trick can be performed using two fragments of one transmembrane segment each, plus a third fragment containing the remaining five (48, 49). Further evidence comes from in vivo experiments where co-expression of fragments of lactose permease were shown to restore lactose transport in E. coli, whereas the fragments were rapidly degraded when expressed alone (9, 104). The types of interactions that take place between membrane o~-helices have been under investigation recently. Structures of membrane proteins known to sufficiently high resolution have been analyzed; it was found that residues that are buried in the interior of these proteins, on average, tend to be more hydrophilic than those in contact with the lipid and that the helix-helix interfaces are generally close-packed (78). There is no direct evidence for salt bridges in the known structures. However, there is indirect evidence for salt bridges between residues in different helices of lactose permease (26, 50, 59, 81). Interactions between potentially charged amino acids in transmembrane regions of eukaryotic proteins have also been proposed to be involved in oligomerization (20, 66), organellar retention (13), and degradation

(51). Not all helix-helix interactions appear to be driven by polar interactions. Giycophorin A from human erythrocytes has a single transmembrane domain and forms

PAUL WHITLEY and GUNNAR VON HEIJNE

very stable homodimers. Dimerization is driven by interactions between the transmembrane helices (14), and saturation mutagenesis of the transmembrane regions has shown the dimerization interface to be largely apolar (60). A general approach for analyzing helical packing that may be applicable to a wide variety of membrane proteins is to look for the formation of disulfide bridges between pairs of cysteines engineered into membrane-spanning helices. This approach was first used to look at the dimerization interface in transmembrane regions of the E. coil aspartate sensory receptor (Tar receptor) (63). The analysis was extended to include intrachain interactions between the two transmembrane helices in Tar receptor monomers (72). Recently, the interface between the two transmembrane helices of E. coil leader peptidase has been mapped by a similar approach (100). One problem with the disulfide approach is that it is not possible to say how the native structure of the membrane protein is perturbed by the formation of disulfide bonds between pairs of genetically engineered cysteines, and a fairly large number of mutants have to be analyzed to get a consistent picture. A couple of computational methods have been put forward to predict how helices may pack against one another (27, 28, 87, 88). However, without access to a sufficiently large number of 3-D structures, it is difficult to validate and refine predictive methods. It should be noted that some oligomerization of membrane proteins almost certainly takes place because of interactions between polar portions of the protein exposed outside the bilayer (89, 106). Interactions between polar domains within a membrane protein may also influence the packing of helices and will, especially in the case of short loops, impose constraints as to which helices can pack against each other. V. M E C H A N I S M O F M E M B R A N E I N S E R T I O N The assembly of a protein into the inner membrane of E. coli requires some portions of the polypeptide chain to pass completely through the lipid bilayer, some to partition into the lipid bilayer, and some to remain cytoplasmic. In previous sections we have dealt with features of the polypeptide chains of membrane proteins that contribute toward these topological decisions, but we have not discussed the mechanisms by which they are achieved. In what follows, we will focus on how periplasmic loops are translocated across, and how transmembrane segments are inserted into, the inner membrane.

A. Translocationof Periplasmic Loops One of the key questions in the study of membrane protein biogenesis is: How do hydrophilic segments of the polypeptide chain pass through the hydrophobic core of the lipid bilayer? This problem also arises when secretory proteins are to

Membrane Protein Assembly

be completely translocated across the lipid bilayer into the periplasmic space. The great majority of secreted proteins in E. coil (and also in eukaryotes) are synthesized with amino terminal extensions known as signal (or leader) sequences that are generally around 20 amino acid residues in length and are removed from the mature polypeptide on the periplasmic side of the cytoplasmic membrane by the leader peptidase enzyme. Despite the fact that signal sequences of different proteins do not appear to have any primary sequence homology, they do share common features (90): an amino terminal positively charged region (n-region, 1-5 residues long), a central hydrophobic part (h-region, 7-15 residues long) thought to be the key region in signal peptide function (37), and a more polar carboxy terminal domain (c-region, 3-7 residues) that contains the cleavage recognition site for leader peptidase (92). It is well established that the presence of a functional signal sequence is essential for efficient secretion, and mutations that impede export of a protein and result in the accumulation of its precursor in the cytoplasm nearly always map to the hydrophobic core of the signal sequence of the protein (30, 33). In general, signal sequences can be exchanged between proteins and still correctly direct export (8), [although this is not always the case (57)], which suggests the importance of secondary structure rather than specific sequence for correct function. Proposals as to how signal peptides function to direct the transfer of polypeptide chains across the lipid bilayer are diverse and are based on data from experimental techniques ranging from classical genetics to biophysics (see (37, 94)). It has thus been proposed that signal peptides interact directly with iipids and direct translocation through the lipid bilayer; in support of this idea, experiments with synthetic peptides corresponding to signal sequences have shown them to have substantial affinity for phospholipid mono- and bilayers and to have an o~-helical conformation in the lipid environment (7, 39). On the other hand, it has also been suggested that hydrophilic stretches of polypeptide must pass through the membrane in a polar proteinaceous channel (12, 84). This is supported by the observation that in electrophysiological experiments signal peptides can open aqueous channels when added to the cytoplasmic side of E. coil planar lipid bilayers (83). Also, translocating polypeptides can be crosslinked to protein components in the E.coli inner membrane (47) and the ER membrane (4, 40, 52). In vivo, it appears that interactions of signal peptides with both proteins (I0) and lipids (24) are important for efficient secretion. Components of the E. coil secretion machinery have been identified by both genetic and biochemical means (see chapters by Ito and by Matsuyama & Mizushima). At least six proteins are known to be involved; namely, SecA, SecB, SecD, SecE, SecE and SecY. In addition, a homologue of the eukaryotic SRP 54-kDa protein, called Ffh, has recently been implicated in the process (62, 73). A recent model as to how the secretion machinery functions in E. coli has been proposed by Wickner (102). Unlike translocation of proteins across the endoplasmic reticulum in eukaryotes (77), translocation across the cytoplasmic membrane

PAUL WHITLEY and GUNNAR VON HEIJNE

of E. coil is not obligatorily co-translational (76). After synthesis, some preproteins are thought to associate with secB, which acts as a molecular chaperone (2, 19) and keeps preproteins in an export competent state. SecB is not an essential protein; other molecular chaperones, such as GroEL and GroES, can substitute for secB function (3) and are, in fact, necessary for the efficient secretion of some proteins (56). SecA is a peripheral membrane protein (22) that probably binds with high affinity to SecB/preprotein complexes by virtue of its affinities for SecB (42) and both the signal (23) and mature (61) domains of the precursor proteins. This affinity for the secretory preprotein complex is greatly enhanced if SecA is in association with membranes containing acidic phospholipids (44) and the integral SecY/E protein complex (42). Thus SecA provides a pivotal link between the presecretory protein and the membrane components of the secretory machinery. SecA has an ATPase activity (the translocation ATPase activity) that is induced when it is in contact with the elements mentioned above (61). The binding of ATP to SecA without hydrolysis driveg limited translocation of the polypeptide chain (82). Hydrolysis of ATP then probably releases the precursor protein from SecA and proton motive force (AI~H+) dependent translocation ensues, although exactly how AI~H+ drives translocation is not known. Subsequent rounds of SecA binding to the preprotein, ATP binding to SecA, hydrolysis of ATP, release of the preprotein, and ApH + driven translocation drive the preprotein through the membrane until it is fully translocated. Since the problem of transferring hydrophilic stretches of polypeptide chain across the lipid bilayer is the same for membrane proteins as it is for secreted proteins, it would seem reasonable to assume that membrane proteins may also use the sec machinery. Additionally, signal sequences broadly resemble membranespanning sequences in their overall properties. In fact, it has been shown that a signal sequence can be converted into a membrane anchor (signal anchor) simply by increasing the length of the hydrophobic core (17) or by removing the consensus site for cleavage by leader peptidase (58). The converse is also true: A signal-anchor sequence can be converted to a signal sequence simply by inserting a recognition site for leader peptidase in the appropriate position (71). At least one membrane protein, leader peptidase, is known to require the function of the SecA and SecY proteins for its correct membrane assembly (103). Other proteins such as MI3 coat protein (101), fl coat protein (80), and lactose permease (105) do not appear to require the sec gene products. A possible distinction between the sec-independent and sec.dependent modes of insertion was first suggested by statistical analysis of inner membrane proteins (91, 96) and concerns the length of the translocated segment. Short periplasmic proteins or short periplasmic loops or tails in membrane proteins were postulated not to need the sec machinery--and indeed all of the sec-independent proteins mentioned above have only short translocated stretches of amino acids--whereas, longer pieces of chain, such as the large periplasmic domain of leader peptidase, were proposed to depend on the sec machinery for translocation.

Membrane Protein Assembly The first experimental evidence for this proposed correlation between length and

sec-dependence was the observation that when the periplasmic domain of M 13 procoat protein, which is sec-independent for assembly, had its length increased from 20 to 118 residues, it became sec-dependent (.54). Recently, a more systematic approach has shown that there is a linear relationship between the length of a translocated loop situated between two membrane-spanning segments and its degree of dependence on functional SecA and SecY proteins for lengths up to -60 residues (5). This observation suggests that there is no abrupt transition from a see-independent to a sec-dependent mode of translocation, but rather that there is a gradual increase with length in the need for functional SecA and SecY proteins. It should be noted, however, that results seemingly in conflict with this simple relationship between length and sec-dependence have been reported for certain MalF-PhoA fusions (67). It is not entirely clear what the see-independent mode of insertion is like in mechanistic terms, and proposals range from a simple direct insertion into the lipid bilayer (.53) to a partial utilization of the sec machinery (.5).

B. Insertion of Transmembrane Segments It has been proposed that polytopic membrane proteins insert into the membrane using an alternating series of uncleaved signal sequences (signal anchors) initiating translocation followed by stop transfer sequences that prevent the translocation of downstream amino acids (11). Evidence to support this mechanism comes from observations that internal hydrophobic sequences in membrane proteins can act as uncleaved signal sequences (6, 35). Furthermore, signal sequences can sometimes act as stop transfer sequences when they are artificially introduced downstream of another signal sequence (18, 64). According to this model, the integration of the first membrane-spanning sequence of a multispanning membrane protein determines the topology of the whole molecule. Thus, if the first hydrophobic membrane segment inserts in an Nin orientation, the second must insert in the opposite Nout orientation, the third again in the Nin orientation, etc. Some evidence for this alternating insertion process comes from an analysis of artificial polytopic membrane protein constructs made from tandem repeats of the human asialoglycoprotein (ASGP) receptor, which normally contains only one hydrophobic signal anchor (99). Constructs with up to four repeats of ASGP receptor were shown to be integrated into dog pancreas microsomes as polytopic membrane proteins. The first copy of the signal anchor indeed acted as a signal anchor, the second integrated into the membrane in the opposite orientation to the first and thus acted as a stop transfer sequence, and the third again acted as a signal anchor, etc. It was concluded that the most N-terminal of the signal anchors inserted first and determined the topology of the whole molecule. Because translocation across the ER membrane of eukaryotes is coupled to protein synthesis, it seems likely that the most N-terminal hydrophobic stretch integrates first as more downstream ones may not yet be synthesized when assembly

PAUL WHITLEY and GUNNAR VON HEIJNE

10

into the membrane begins. In prokaryotes, however, translocation is not mechanisticaUy coupled to protein synthesis, and therefore, it is not obvious that the most N-terminal hydrophobic segment necessarily integrates first. Thus, the amino terminal transmembrane segment in the MalF protein can be deleted with no apparent effects on the membrane integration and topology of the remainder of the molecule (29). Furthermore, the fact that polytopic bacterial inner membrane proteins have high numbers of positively charged residues in all of their cytoplasmic loops, whether they are at the amino or caboxy terminal ends of the protein (95), suggests that each transmembrane segment (or pairs of segments) inserts into the membrane independently of the rest of the molecule. In eukaryotes, the high number of positive charges in cytoplasmic loops is more pronounced towards the N-terminus of the protein (85), which may reflect the fact that if the N-terminus is inserted correctly then the rest of the molecule will follow. As has been mentioned above, the most characteristic feature of the membranespanning portions of a polypeptide chain is their high hydrophobicity which, in principle, could allow them to spontaneously partition into the lipid bilayer (31, 45, 69). However, if a polypeptide chain is translocating through an aqueous channel (a translocon) as is now the preferred model, interactions between the amino acid side chains of a hydrophobic stop transfer sequence and the core of the lipid bilayer cannot account for termination of translocation. Rather, proteins in the translocon must somehow recognize the hydrophobic stretch of amino acids as a stop transfer signal, and the translocon must presumably disassemble, or in some other way facilitate release of the membrane-spanning sequence (e.g., by letting it slip between helices of the translocon), in order for it to become integrated into the lipid bilayer. How this is achieved is still completely unknown.

VI.

CONCLUSIONS

Over the past few years, there have been major advances in our understanding of membrane protein biogenesis in bacteria. Thus, the components of the secretory machinery are basically known, and their functional roles have at least been partly defined. The most important characteristics of the substrates for the secretory machinery--the nascent membrane proteins--have also been clarified with the discovery that positively charged amino acids act as the major topogenic determinants during membrane integration. As a measure of our understanding, we can now predict the topology of an inner membrane protein from its amino acid sequence and expect to be fight 9 times out of 10. Nevertheless, a number of issues need to be addressed in the future. The obvious ones are of course related to the mechanism of translocation: Which proteins make the channel (if there is one)? What are the properties of the channel? How are signal sequence and stop transfer sequences recognized?

Membrane Protein Assembly

11

Other questions concern the nascent polypeptide: Can charge-pairs in or between t r a n s m e m b r a n e segments play a role during m e m b r a n e insertion? H o w should we approach the study of helix-helix packing? Can we design new m e m b r a n e proteins o f p r e d e t e r m i n e d topology and structure? Problems such as these should keep us busy for some time to come.

REFERENCES 1. Akita, M., Sasaki, S., Matsuyama, S., & Mizushima, S. (1990). SecA interacts with secretory proteins by recognizing the positive charge at the amino terminus of the signal peptide in Escherichia coli. J. Biol. Chem. 265, 8164-8169. 2. Aitman, E., Emr, S. D., & Kumamoto, C. A. (1990). The presence of both the signal sequence and a region of mature lamb protein is required for the interaction of lamb with the export factor secb. J. Biol. Chem. 265, 18154-18160. 3. Altman, E., Kumamoto, C. A., & Emr, S. D. (1991). Heat-shock proteins can substitute for secb function during protein export in Escherichia coli. EMBO J. 10, 239-245. 4. Anderson, L., & Denny, J. B. (1992). Protein translocation in the endoplasmic reticulumn ultraviolet light induces the noncovalent association of nascent pepfides with translocon proteins. J. Biol. Chem. 267, 23916-23921. 5. Andersson, H., & yon Heijne, G. (1993). Sec-dependent and sec-independent assembly of E. coil inner membrane proteins--the topological rules depend on chain length. EMBO J. 12, 683-691. 6. Audigier, Y., Friedlander M., & Blobel, G. (1987). Multiple topogenic sequences in bovine opsin. Proc. Natl. Acad. Sci. USA 84, 5783-5787. 7. Batenburg, A. M., Demel, R. A., Verkleij, A. J., & de Kruijff, B. (1988). Penetration of the signal sequence of Escherichia coil PhoE protein into phospholipid model membranes leads to lipidspecific changes in signal peptide structure and alterations of lipid organization. Biochemistr), 27, 5678-5685. 8. Benson, S. A., Hall, M. N., & Silhavy, T. J. (1985). Genetic analysis of protein export in Escherichia coli KI2. Ann. Rev. Biochem. 54, 101-134. 9. Bibi, E., & Kaback, H. (1990). In vivo expression of the lacy gene in two segments leads to functional iac permease. Proc. Natl. Acad. Sci. USA 87, 4325-4329. 10. Bieker-Brady, K., & Silhavy, T. J. (1992). Suppressor analysis suggests a multistep, cyclic mechanism for protein secretion in Escherichia coli. EMBO J. 11, 3165-3174. 11. BIobel, G. (1980). Intraceilular protein topogenesis. Proc. Natl. Acad. Sci. USA 77, 1469-1500. 12. Blobel, G., & Dobherstein, B. (1975). Transfer to proteins across membranes. II. Reconstitution of functional rough microsomes from heterologous components. J. Cell. Biol. 67, 852-862. 13. Bonifacino, J. S., Cosson, P., Shah, N., & Klausner, R. D. (1991). Role of potentially charged transmembrane residues in targeting proteins for retention and degradation within the endoplasmic reticulum. EMBO J. 10, 2783-2793. 14. Bormann, B.-J., Knowles, W. J., & Marchesi, V. T. (1989). Synthetic peptides mimic the assembly of transmembrane glycoproteins. J. Biol. Chem. 264, 4033--4037. 15. Boyd, D., Traxler, B., & Beckwith, J. (1993). Analysis of the topology of a membrane protein by using a minimum number of alkaline phosphatase fusions. J. Bacteriol. 175, 553-556. 16. Broorne-Smith, J. K., & Spratt, B. G. (1986). A vector for the construction of translational fusions to TEM [$-lactamase and the analysis of protein export signals and membrane protein topology. Gene 49, 341-349. 17. Chou, M. M., & Kendall, D. A. (1990). Polymeric sequences reveal a functional interrelationship between hydrophobicity and length of signal peptides. J. Biol. Chem. 265, 2873-2880.

12

PAUL WHITLEY and GUNNAR VON HEIJNE

18. Coleman, J., Inukai, M., & Inouye, M. (1985). Dual functions of the signal peptide in protein transfer across the membrane. Cell 43, 351-360. 19. Collier, D., Bankaitis, V., Weiss, J., & Bassford, P. (1988). The antifolding activity of SecB promotes the export of the E. coli maltose-binding protein. Cell 53, 273-283. 20. Cosson" P., Lankford, S. P., Bonifacino, J. S., & Klausner, R. D. (1991). Membrane protein association by potential intramembrane charge pairs. Nature 351,414-416. 21. Cowan, S. W.,'Schirmer, T., Rummel, G., Steiert, M., & Ghosh, R. (1992). Crystal structures explain functional properties of two E. coil porins. Nature 358, 727-733. 22. Cunningham, K., Lill, R., Crooke, E., Rice, M., & Moore, K. (1989). SecA protein, a peripheral protein of the Escherichia coli plasma membrane, is essential for the functional binding and translocation of proOmpA. EMBO J. 8, 955-959. 23. Cunningham, K., & Wickncr, W. (1989). Specific recognition of the leader region of precursor proteins is required for the activation of translocation ATPase of Escherichia coll. Proc. Natl. Acad. Sci. USA 86, 8630-8634. 24. de Vrije, T., de Swart, R., Dowhan, W., Tonenassen, J., & de Kruijff, B. (1988). Phosphatidylglycerol is involved in protein translocation across Escherichia coli inner membranes. Nature 334, 173--175. 25. Deisenhofer, J., Epp, O., Mild, K, Huher, R., & Michel, H. (1985). Structure of the protein subunits in the photosynthetic reaction centre of Rhofl.opseudomonas viridis at 3,1. resolution. Nature 318, 618--624. 26. Dunten, R. L., Sahintoth, M., & Kaback, H. R. (1993). Role of the charge pair aspartic acid-237-1ysine-358 in the lactose permease of Escherichia coli. Biochemistry 32, 3139-3145. 27. Efremov, R. G., Gulyaev, D. I., & Modyanov, N. N. (1992). Application of 3-dimensional molecular hydrophobicity potential to the analysis of spatial organization of membrane protein domains. 2. Optimization of hydroplmbic contacts in transmembrane hairpin structures of Nat-, K+-ATPase. J. Protein Chem. 1I, 699-708. 28. Efremov, R. G., Gulyaev, D. I., Vergoten, G., & Modyanov, N. N. (1992). Application of 3-dimensional molecular hydrophobicity potential to the analysis of spatial organization of membrane domains in proteins. !. Hydrophobic Woperties of transmembrane segments of Ha+, K+-ATPase. J. Protein Chem. I 1,665--675. 29. Ehrmann M., & Beckwith, J. (1991). Proper insertion of a complex membrane protein in the absence of its amino-terminal export signal. J. Biol. Chem. 266, 16530-16533. 30. Emr, S., Hedgpeth, J., ClCtment,J.-M., Silhavy, T., & Hofnung, M. (1980). Sequence analysis of mutations that prevent export of lambda receptor, an Escherichia coli outer membrane protein. Nature 285, 82-85. 31. Engelman, D. M., & Steitz, T. A. (1981). The spontaneous insertion of proteins into and across membranes: the helical hairpin hypothesis. Cell 23, 411-422. 32. Engelman, D. M., Steilz, T. A., & Goldman, A. (1986). Identifying nonpolar transbilayer helices in amino acid sequences of membrane provdns.Aml. Rev. BhTphys. Biophys. Chem. 15, 321-353. 33. Fikes, J., Bankaitis, V., Ryan, J., & Bassford, P. (1987). Mutational alterations affecting the export competence of a truncated but fully functional maltose-binding protein signal peptide. J. BacterioL 169, 2345-2351. 34. Foster, D. L., Buoblik, M., k Kaback, H. R. (1983). Structure of the lac carrier protein of Escherichia coll. J. Biol. Chem. 258, 31-34. 35. Friedlander, M., & BIobel, G. (1985). Bovine opsin has more than one signal sequence. Nature 318,338--343. 36. Froshauer, S., Green, G. N., Boyd, D., McGovem, K., & Beckwith, J. (1988). Genetic analysis of the membrane insertion and topology of MaIF, a cytoplasmic membrane protein of Escherichia coli. J. Mol. Biol. 200, 501-511. 37. Gennity, J., Goldstein, J., & Inouye, M. (1990). Signal peptide mutants of Escherichia coll. J. Bioenerg. Biomembr. 22, 233-69.

Membrane Protein Assembly

13

38. Georgiou, C. D., Dueweke, T. J., & Gennis, R. B. (1988). 13-galactosidasegene fusions as probes

39. 40.

41.

42.

43.

4.

45.

6,

47. 48.

9.

50. 51. 52.

53. 54. 55. 56.

57.

for the cytoplasmic regions of subunits I and II of the membrane-bound cytochrome d terminal oxidase from Escherichio coll. J. Biol. Chem. 263, i 3130-13137. Gierasch, L. M. (1989). Signal sequences. BiochemistD' 28, 923-930. GOrlich, D., Prehn, S., Hartmann, E., Kalies, K. U., & Rapoport, T. A. (1992). A mammalian homolog of SEC61p and SECYp is associated with ribosomes and nascent polypeptides during translocation. Cell 71,489-503. GOtt, P., & Boos, W. (1988). The transmembrane topology of the sn-glycerol-3-phosphate permease of Escherichia coil analysed by phoA and LacZ protein fusions. Moi. Microbiol. 2, 655-663. Hard, E U., Lecker, S., Schiebel, E., Hendrick, J. P., & Wickner, W. (1990). The binding cascade of Secb to Seca to Secy/E mediates preprotein targeting to the E-coil plasma membrane. Cell 63, 269-279. Henderson, R., Baldwin, J. M., Ceska, T. A., Zemlin, F., Beckmann, E., & Downing, K. H. (1990). A model for the structure of bacteriorhodopsin based on high resolution electron cryo-microscopy. J. Mol. Biol. 213, 899-929. Hendrick, J. P., & Wickner, W. (1991). SecA protein needs both acidic phospholipids and secy/e protein for functional high-affinity binding to the Escherichia-coli plasma membrane. J. Biol. Chem. 266, 24596-24600. Jacobs, R., & White, S. (1989). The nature of the hydrophobic binding of small peptides at the bilayer interface: implications for the insertion of transbilayer helices. Biochemistry 28, 34213437. Jennings, M. L. (1989). Topography of membrane proteins. Ann. Rev 8iochem. 58, 999-1027. Joly, J. C., & Wickner, W. (1993). The secA and secY subunits of translocase are the nearest neighbors of a transiocafing preprotein, shielding it from phospholipids. EMBO J. 12, 255-263. Kahn, T., & Engelman, D. (1992). Bacteriorhodopsin can be refolded from two independently stable transmembrane helices and the complementary five-helix fragment. Biochemist~' 31, 6144--6151. Kataoka, M., Kahn, T. W., Tsujiuchi, Y., Engelman, D. M., & Tokunaga, E (1992). Bacteriorhodopsin reconstituted from 2 individual helices and the complementary 5-helix fragment is photoacfive. Photochem. Photobiol. 56, 895-901. King, S. C.,.Hansen, C. L., & Wilson, T. H. (1991). The interaction between aspartic acid 237 and lysine 358 in the lactose carrier of Escherichia coll. Biochim. Biophys. Acta 1062, 177-186. Kiausner, R. D., & Sitia, R. (1990). Protein degradation in the endoplasmic reticulum. Cell 62, 611--614. Krieg, U. C., Johnson, A. E., & 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. Kuhn, A. (1987). Bacteriophage MI3 procoat protein inserts into the plasma membrane as a loop structure. Science 238, 1413-1415. Kuhn, A. (1988). Alterations in the extracellular domain of MI3 procoat protein make its membrane insertion dependent on secA and secY. Eur. J. 8iochem. 177, 267-271. Kurosaki, T., & Ravetch, J. V. (1989). A single amino acid in the glycosyl phosphatidylinositol attachment domain determines the membrane topology of fc-gamma-riii. Nature 342, 805-807. Kusukawa, N., Yura, T., Ueguchi, C., Akiyama, Y., & Ito, K. (1989). Effects of mutations in heat-shock genes GroES and GroEL on protein export in Escherichia coll. EMBO J. 8, 35173521. Laforet, G. A., Kaiser, E. T., & Kendall, D. A. (1989). Signal peptide subsegments are not always functionally interchangeable. M I3 hydrophobic core fails to transport alkaline phosphatase in Escherichia coll. J. Biol. Chem. 264, 14478-14485.

14

PAUL WHITLEY and GUNNAR VON HEIJNE

58. Laforet, G. A., & Kendall, D. A. (1991). Functional limits of conformation, hydrophobicity, and

steric constraints in wokaryotic signal peptide cleavage regions--wild type transport by a simple polymeric signal sequence. J. Biol. Chem. 266, 1326-1334. 59. Lee, J.-l., Hwang, P. P., Hansen, C., & Wilson, T. H. (1992). Possible salt bridges between transmembrane a-hefices of the lactose carrier of Escherichia coll. J. Biol. Chem. 267, 207582O764. M. A., Flanagan, J. M., Treutlein, H. R., Zhang, J., & Engeiman, D. M. (1992). 0. Sequence specificity in the dimerization of transmembrane alpha-helices. Biochemistry 31, 12719-12725. 61. Lill, R., Dowhan, W., & Wickncr, W. (1990). The ATPase activity of secA is regulated by acidic phospl~lipids, secY, and the leader and mature domains of precursor proteins. Cell 60, 271-280. 20 Luirink, J., High, S., Wood, H., Giner, A., Toilervey, D., & Dobberstein, B. (1992). Signalsequence recognition by an Escherichia coli ribonucleoprotein complex. Nature 359, 741-743. 63. Lynch, B., & Koshland, D. (1991). Disulfide cross-finking studies of the transmembrane region of the aspartate sensory receptor of Escherichia coll. Proc. Natl. Acad. Sci. USA 88,10402-10406. 4. Madntyre, S., Freudi, R., Eschbach, M.-L., & Henning, U. (1988). An artificial hydrophobic sequence functions as either an anchor or a signal sequence at only one of two positions within the Escherichia coil outer membrane protein OmpA. J. Biol. Chem. 263, 19053-19059. 5. Manoil, C., Mekalanos, J. J., & Beckwith, J. (1990). Alkaline phosphatase fusions---sensors of subcellular location. J. Bact. 172, 515-518. 6. Manolios, N., Bonifacino, J. S., & Klausner, R. D. (1990). Transmembrane helical interactions and the assembly of the t-cell receptor complex. Science 249, 274-277. 67. McGovern, K., & Beckwith, J. (1991). Membrane insertion of the Escherichia coli malF protein in cells with impaired secretion machinery. J. Biol. Chem. 266, 20870-20876. 68. Michaelis, S., Inouye, H., Oliver, D., & Beckwith, J. (1983). Mutations that alter the signal sequence of alkaline phosphatase in Escherichia coli. J. Bacteriol. 154, 366-374. 9. Milik, M., & Skolnick, J. (1992). Spontaneous insertion of polypeptide chains into membranes--a Monte-Carlo Model. Proc. Natl. Acad. Sci. USA 89, 9391-9395. 70. Niisson, I. M., & yon Heijne, G. (1990). Fine-tuning the topology of a polytopic membrane Wotein. Role of positively and negatively charged residues. Cell 62, 1135-1141. 71. Nilsson, I. M., & yon Heijne, G. (1991). A de nov, designed signal peptide cleavage cassette functions in vivo. J. Biol. Chem. 266, 3408--3410. 72. Pakula, A. A., & Simon, M. I. (1992). Determination of transmembrane protein structure by disulfide cross-linking--the Escherichia-coli tar receptor. Proc. Natl. Acad. Sci. USA 89, 41444148. 73. Phillips, G. J., & Silhavy, T. J. (1992). The E. coil ffh-Gene is necessary for viability and efficient proteinexport.Nature 359, 744-746. 4. Pop,t, J.-L.,Gerchman, S.-E.,& Engelman, D. M. (I987). Refolding of bacteriorhodopsinin lipidbilayers.A thennodynmnically controlledtwo-stage process.J. Mol, Biol. 198, 655--676. 75. Pop,t, J. L., & Engelman, D. M. (1990). Membrane proteinfolding and oligomerization--4he 2-stage model. Biochemistry 29, 4031-.4037. 76. Randall, L. L. (1983). Translocation of domains of nascent periplasmic proteins across the cytoplasmic membrane isindependent of elongation.Cell 33, 231-240. 77. Rap.port, T. A. (1992). Transport of proteins across the endoplasmic reticulum membrane. Science 258, 931-936. 78. Rees, D. C., DeAntonio, L., & Eisenberg, D. (1989). Hy~ophobic organizationof membrane proteins.Science 245, 510-513. 79. Roberts,C. H., & Chlebowski, J.E (1984).PhoA+T X- 100+trypsin.J.Biol.Chem. 259, 729-733.

Membrane Protein Assembly

15

0. Rohrer, J., & Kuhn, A. (1990). The function of a leader peptide in translocating charged amino acyl residues across a membrane. Science 250, 1418-1421. 81. Sahin-T6th, M., Dunten, R. L., Gonzales, A., & Kaback, H. R. (1992). Functional interactions between putative intramembrane charged residues in the lactose permease of Escherichia coil Proc. Natl. Acad. Sci. USA 89, 10547-10551. 82. Schiebel, E., Driessen, A. J. M., Hartl, E-U., & Wickner, W. (199 l). ApH~ and ATP function at different steps of the catalytic cycle of preprotein translocase. Cell 64, 927-939. 83. Simon, S. M., & Blobel, G. (1992). Signal peptides open protein-conducting channels in E-coli. Cell 69, 677-684. 84. Singer, S., Maher, P., & Yaffe, M. (1987). On the translocation of proteins across membranes. Proc. Natl. Acad. Sci. USA 84, 1015-1019. 85. Sipos, L., & van Heijne, G. (1993). Predicting the topology of eukaryotic membrane proteins. Eur. J. Biochem. 213, 1333-1340. 86. Traxler, B., Lee, C., Boyd, D., & Beckwith, J. (1992). The dynamics of assembly of a cytoplasmic membrane protein in Escherichia-coti. 1. Biol. Chem. 267, 5339-5345.

87. Treutlein, H. R., Lemmon, M. A., Engelman, D. M., & Brunger, A. T. (1992). The glycophorin A transmembrane domain dimer: sequence-specific propensity for a right-handed supercoii of helices. Biochemistry 31, 12726-12732. 88. TufteD', P., & LaveD', R. (1993). Packing and recognition of protein structural elements--a new approach applied to the 4-helix bundle of myohemeD'thrin. Proteins Struct. Funct. Genet. 15, 413-425. 89. Verrall, S., & Hall, Z. W. (1992). The n-terminal domains of acetylcholine receptor subunits contain recognition signals for the initial steps of receptor assembly. Cell 68, 23-31. 0. van Heijne, G. (1985). Signal sequences. The limits of variation. 1. Mol. Biol. 184, 99-105. 91. van Heijne, G. (1986). The distribution of positively charged residues in bacterial inner membrane proteins correlates with the trans-membrane topology. EMBO 1. 5, 3021-3027. 92. van Heijne, G. (1988). Transcending the impenetrable: how proteins come to terms with membranes. Biochim. Biophys. Acta 947, 307-333. 93. van Heijne, G. (1989). Control of topology and mode of assembly of a polytopic membrane protein by positively charged residues. Nature 341,456-458. 94. van Heijne, G. (1990). The signal peptide. J. Membr. Biol. 115, 195-201. 95. van Heijne, G. (1992). Membrane protein structure prediction--hydrophobicity analysis and the positive-inside rule. J. Mol. Biol. 225, 487-494. 6. van Heijne, G., & Gavel, Y. (1988). Topogenic signals in integral membrane proteins. Eur. J. Biochem. 174, 671-678. 97. Wallace, B. A., Cascio, M., & Mielke, D. L. (1986). Evaluation of methods for the prediction of membrane protein secondary structures. Proc. Natl. Acad. Sci. USA 83, 9423-9427. 98. Weiss, M. S., Kreusch, A., Schiltz, E., Nestel, U., Welte, W. (1991). The structure of porin from Rhodobacter capsulata at 1.8A, resolution. FEBS Len. 280, 379-382. 9. Wessels, H. P., & Spiess, M. (1988). Insertion of a multispanning membrane protein occurs sequentially and requires only one signal sequence. Cell 55, 61-70. 100. Whitley, P., Nilsson, L., & yon Heijne, G. (1993). A 3D model for the membrane domain of Escherichia coli leader peptidase based on disulfide mapping. Biochemistry (in press). 101. Wickner, W. (1988). Mechanisms of membrane assembly: general lessons from the study ofM 13 coat protein and Escherichia coli leader peptidase. Biochemistry 27, 1081-1086. 102. Wickner, W., Driessen, A. J. M., & Hartl, E U. (1991). The enzymology of protein translocation across the Escherichia coil plasma membrane. Ann. Rev. Biochem. 60, 101-124. 103. Wolfe, P. B., Rice, M., & Wickner, W. (1985). Effects of two sec genes on protein assembly into the plasma membrane of Escherichia coli. J. Biol. Chem. 260, 1836-184 I.

16

PAUL WHITLEY and GUNNAR VON HEIJN[

104. Wrubel, W., Stochaj, U., Sotmewald, U., Theres, C., & Ehring, R. (1990). Reconstitution of an active lactose carrier in vivo by simultaneous synthesis of two complementary protein fragments. J. Bacteriol. 172, 5374--5381. 105. Yamato, I. (1992). Membrane assembly of lactose permease of Escherichia-coli. J. Biochem. Tokyo 111,444-450. 106. Yu, X.-M., & Hall, Z. W. (199 I). Extracellular domains mediating e subunit interactions of muscle acetylcholinerecelXOr.Nature 352, 64--67.

MEMBRANE INSERTION OF SMALL PROTEINS: EVOLUTIONARY AND FUNCTIONAL ASPECTS

Dorothee Kiefer and Andreas Kuhn

I. II. III. IV. V. VI. VII.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Aminoterminal versus Carboxylterminal Translocation . . . . . . . . . . . . . Determination of Membrane Protein Orientation . . . . . . . . . . . . . . . . Transl~ Translocation . . . . . . . . . . . . . . . . . . . . . Insertion of Small Proteins into the Encloplasmic Reticulum . . . . . . . . . . Requirement of Chaperones . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

17 18 20 21 23 26 27 28 28

I. I N T R O D U C T I O N Each protein in a living s y s t e m - - p r o k a r y o t i c or eukaryotic---resides in a characteristic cellular c o m p a r t m e n t to fulfill its biological function. With the increasing

Advances in Cell and Molecular Biology of Membranes and Organelles Volume 4, pages 17-35. Copyright 9 1995 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-$$938-924-9 17

18

DOROTHEE KIEFERand ANDREASKUHN

complexity during evolution, enzymatic mechanisms have been developed that guarantee the specific locations of proteins within certain compartments that are usually separated from each other by biological membranes. Since many translocared proteins have a rather complex conformation, often protein domains cannot cross or insert into the membranes spontaneously to reach their final destination. Therefore, a cellular export machinery evolved that allows the transport of more complex protein domains across a bilayer and that also prevents premature folding of these domains prior to the translocation event. In prokaryotes, as in gram-negative bacteria, the only membraneous system is the cell wall consisting of the cytoplasmic membrane and the outer membrane. Proteins synthesized in the cytoplasm, therefore, do not have to select between different membrane systems. Eukaryotic cells, however, possess several defined compartments and organeUes, such as the nucleus, the endoplasmic reticulum and the Golgi apparatus, as well as mitochondria, chloroplasts, and peroxisomes. It is obvious that, due to this differentiated organization, eukaryotes had to evolve sophisticated targeting systems to determine a protein's final destination. Both in eukaryotes and in prokaryotes, proteins that are not resident in the cytoplasm face the problem of translocating through a highly hydrophobic environment. Translocation machineries composed of mainly proteinaceous components were developed during evolution to facilitate this process. However, there are some proteins known to translocate through biological membranes independently of the transport components. Most of these proteins are small with simple secondary structures, and the mechanism of their membrane insertion may reflect an ancient situation before a complex translocation apparatus was evolved. II.

A M I N O T E R M I N A L VERSUS CARBOXYLTERMINAL TRANSLOCATION

Since membrane proteins are synthesized in the cytoplasm, they insert their hydrophobic portion from the cis side into the membrane. In many cases, this initial process leads to the formation of a structure with both hydrophilic regions that flank the hydrophobic region on the cis side (Figure 1). To achieve a transmembrane orientation either the aminoterminal or the carboxylterminal hydrophilic portion of the protein has to translocate across the bilayer. In nature we find that the majority of single membrane-spanning proteins are translocating their C-terminal regions across the membrane. In addition, most proteins that have their N-terminal portion at the trans side of the membrane use cleavable signal or leader peptides and leave the N-terminus of the translocating precursor at the cis side. After membrane translocation the signal peptides are proteolytically processed resulting in a new aminoterminus at the trans side. It is therefore possible that a cleavable signal sequence has evolved to initiate the translocation of the N-terminal domain of a protein. Multispanning proteins are

Membrane Insertion of Small Proteins

19

periplasm ~

.......... . .... ' .......

.

"

9

.......

~..

/

f

,

I

C N-terminal translocation

C

I

L

_

~

N C-terminal trarmlocat.lon

Figure 1. Amino- versus carboxylterminal membrane insertion of a bitopic membrane protein.

thought to insert as pairs of hydrophobic stretches synergistically forming single loops as initial insertion structures. The most simple case for such a synergistic insertion mode is the M 13 procoat protein, which forms a loop structure during membrane insertion (30). There are a few single membrane-spanning proteins in bacteria known to translocare their N-terminal region directly to the trans side. The most simple case is the coat protein of the Pseudomonas aeruginosa phage Pf3. When this protein of 44 amino acid residues is expressed in E. coli, the first N-terminal 18 hydrophilic residues are facing the periplasm (34). The N-terminal translocation of the Pf3 coat protein is limited to only simple protein regions. When the N-terminal region was replaced by one that contains 3 additional charged residues, membrane insertion was impaired (62). The same charged region, however, was translocated with an additional signal peptide derived from the coat protein of the bacteriophage M 13. Although the two phage proteins are functionally related, as they are both coat proteins of filamentous phages assembling in the bacterial plasma membrane, only the M 13 coat protein is synthesized with a cleavable signal sequence. Little is known regarding the translocation of aminoterminal regions since no involvement of the Sec components has been observed. It has been suggested that the N-terminal region of the Pf3 protein has an amphiphilic character and buries information to support an extracellular location (63). However, there is no direct evidence to support export by amphiphilic sequences in bacteria. Remarkably, when the 18-amino-acid long N-terminal region of the Pf3 coat protein is fused to the aminoterminus of leader peptidase, the Pf3 portion of the hybrid protein is clearly translocated to the periplasm (46). Since the region flanking the hydrophobic stretch C-terminally is highly charged, translocation of the C-terminal region is prevented in favor of the N-terminal. Consequently, Lee et al. (46) deleted the cluster of 7 charged residues and observed both species, either with the amino or

20

DOROTHEE KIEFERand ANDREAS KUHN

with the carboxy terminus on the trans side. The most likely molecular mechanism is the partitioning of the hydrophobic region into the bilayer, which results in a horizontal orientation to the membrane surface (Figure 1). Either hydrophilic region is then passively translocated in favor of an or-helical conformation of the transmembrane region then parallel to the membrane normal.

III. DETERMINATION OF MEMBRANE PROTEIN ORIENTATION A question arises when proposing a single membrane-spanning protein: Which of the two hydrophilic regions will be translocated across the membrane and which region will remain at the cis side? As predicted from the "positive-inside rule" (79, 81), the region bearing more positively charged residues remains at the cis side of the membrane. This was clearly demonstrated with a number of mutants of the leader peptidase leading to an inverted orientation (51, 80). Moreover, it has been shown for a number of proteins that the introduction of positively charged residues close to the hydrophobic region inhibits the translocation process (4, 45, 47, 92). These results might reflect a sensitivity of the Sec-components towards positively charged residues. However, the studies by von Heijne with the leader peptidase are confused by the fact that the insertion mode is altered from a Sec-dependent (wild-type leader peptidase) (17) to a Sec-independent mechanism (inverted leader peptidase) (6, 51, 80). Another reason for the influence of positive charges might be an electrophoretic contribution in the translocation process itself (22). There are, however, a number of exceptions to the "positive-inside rule." The subunit H of the photosynthetic reaction center of Rhodobacter capsulatus and R. sphaeroides has a string of 5 negatively charged residues flanking the hydrophobic core at the cis side of the membrane (1, 93). The same is true for the two small [l-subunits of the antenna complexes of R. capsulatus (75). Moreover, in the M13 coat protein, the reversion of the negatively charged periplasmic region to a 4-fold positively charged one does not affect the orientation, but does affect the velocity of the translocation process (35). These results question simple models proposing, for example, that the electrochemical potential orients horizontally inserted proteins or that charge interactions between the membrane surface and hydrophilic protein regions are the only factors that determine the orientation of the helix. The question of orientation of an inserting polypeptide has also been extensively studied in the eukaryotic system. As mentioned above, single-spanning (bitopic) membrane proteins can achieve two orientations, type H proteins leaving the aminoterminus at the cytoplasmic side of the membrane, while type I and type HI proteins translocate the aminoterminus to the luminal side. Type I proteins achieve their topology indirectly by a cleavable signal sequence creating a new aminoterminus. Because in this case the orientation is mainly directed by the signal peptide, only membrane inserting proteins with a single hydrophobic region (signal anchor)

Membrane Insertion of Small Proteins

21

are considered here. From a number of studies on different bitopic membrane proteins, it became clear that the properties of the regions flanking the hydrophobic core of the signal-anchor sequence are crucial (24, 41). Most of the work was done with type II bitopic membrane proteins, and it was shown that positively charged amino acid residues preceding the signal-anchor were important for the orientation of the transmembrane segment. For the asialoglycoprotein receptor the replacement of these N-terminal positive charges by negative ones resulted in an inverted orientation (9). Similarly, the orientation of paramyxovirus hemagglutinin-neuraminidase (HN) was inverted 75% by the introduction of charged residues in the flanking sides (54). Lamb and coworkers also showed that the orientation of the type III membrane protein M2 of the influenza A virus is not altered when the hydrophobic membrane-spanning region is replaced by one of the SH proteins of simian virus 5, a type II membrane protein (55). This clearly suggests that the orientation is not determined by the membrane anchor, but by the regions flanking the anchor. However, in further studies on P450, Interleukin-2 and the Na-K-ATPase, it was demonstrated that the hydrophobicity and length of the transmembrane segment can influence membrane orientation (42, 58, 64, 65). Conclusively, these results show that membrane orientation and functionality of topogenic signals are determined by a balance between the length of the hydrophobic domain and the charge distribution of its flanking regions. It should be mentioned that there are also a few eukaryotic proteins that do not assume the predicted "positive-inside" orientation, such as some mutants of a preprolactin and IgM fusion protein (7) or the adenovirus E3 14.5-kDa protein (29). Similar to prokaryotes, a simple explanation (e.g., an electrostatic effect between the inserting protein and the membrane surface) seems, therefore, not to be satisfactory. It has been suggested that the positive charges at either side of the hydrophobic region may interfere with components of the translocation channel (70). Eukaryotic proteins might need additional features to determine their correct transmembrane orientation. The components involved in this process might include SRP, ribosomes, GTP, and membrane bound transport components since most membrane proteins require these factors as do secreted proteins (26).

IV. TRANSLOCASE-INDEPENDENT TRANSLOCATION The translocation of small and simple protein regions (N-terminal or C-terminal to the transmembrane domain) might occur spontaneously, presumably driven by conformational changes generated by the transition from a hydrophilic to a hydrophobic environment (84). In E. coil, several small proteins have been investigated that insert into the membrane in the absence of a functional transiocase. In the Pf3 coat protein, the N-terminal 18 residues, of which two are negatively charged, are localized in the periplasm (34, 62). The translocation of this Pf3 region occurs without the help of the Sec proteins and does not always require the membrane

22

DOROTHEE KIEFERand ANDREAS KUHN

potential (46, 62). This is also observed very clearly in a Pf'3 leader peptidase fusion protein, since the Sec- and potential dependent translocation of the leader peptidase moiety at the C-terminus serves as an intramolecular control. The Sec-independent and potential-independent translocation of the Pf3 region, representing the N-terminus of the fusion protein, was blocked by the addition of two charged residues at positions 17 and 18 (46), whereas an extension at the N-terminus by 20 uncharged residues was not inhibitory (13b). The possession of a signal sequence does not necessarily determine the involvement of Sec proteins in the membrane insertion process, although all current models of the translocase mechanism predict a central role for the signal sequence, particularly in the early stages. Membrane translocation of M 13 procoat protein does not depend on the presence of SecB (36), nor on functional SecA or SecY (90), even though it has a signal sequence. In addition, no basic differences were found between a signal sequence of a Sec-independent and a Sec-dependent protein because both were interchangeable (33). Rather, the complexity of that part of the protein that is transported across a membrane calls for a support by the Sec components. It remains unclear, however, how the Sec components recognize their substrates and which regions actually interact. Two proteins have been extensively studied to decipher the Sec-requiring elements, the M 13 procoat protein and the inverted leader peptidase. Extension of the periplasmic loop of M 13 procoat by 98 amino acid residues derived from the mature region of the OmpA protein causes a See-requirement for membrane insertion (31). Similarly, the extension of the periplasmic loop of the inverted leader peptidase from 25 amino acid residues to 54 residues made membrane insertion dependent on SecA and SecY (80). For both proteins a study of a series of loop regions of different lengths showed that intermediate lengths resulted in a loss of translocation efficiency (.5). Presumably, protein segments of about 40 residues are already too complex for spontaneous insertion and too small for efficient interaction with the Sec components. Both hydrophobic regions are required in a synergistic way for the Sec-independent translocation of M 13 procoat. Substitution of valine 30 for an arginine in the center of the mature hydrophobic region results in a membrane-associated, but not translocated, species (32). The same substitution did not inhibit translocation when introduced into the Sec-dependent M 13 derivative with an extended periplasmic loop (13a), suggesting that the synergistic insertion mode requires a close proximity to the two hydrophobic regions. There are only a few examples of other small bitopic membrane proteins in bacteria other than the Pf3 and M 13 coat. Recently, some work has been done with the light-harvesting pigment-binding polypeptides of phototrophic purple bacteria. The o~- and the 13-subunits of the light-harvesting complex I of Rhodobacter capsulatus are 59 and 48 amino acid residues in length, respectively (73, 74). Their central domains of about 20 hydrophobic amino acid residues are proposed to form an 0~-helix (12). The N-termini of both subunits are located in the cytoplasm, whereas the C-termini are exposed to the periplasmic space (75). In accordance

Membrane Insertion of Small Proteins

23

with the "positive-insideomle," (81) the N-terminus of the (x-subunit is positively charged as it is known for various photosynthetic bacteria (99). However, the I~-subunit of R. capsulatus has multiple negatively charged residues at its N-terminus, although located at the cis side of the membrane. It was suggested that the opposite charges of the N-termini of the two subunits may stabilize the formation of the oligomeric complex. Recent work by DOrge et al. (21) with simultaneously exchanged N-terminal charges of the a- and [3-subunits, however, resulted in a destabilization of the antenna complex. Interestingly, the formation of the complex was strongly impaired in mutants with charge exchanges very close to the hydrophobic core; mutations in more distal positions had little or no effect on complex assembly as revealed by the respective absorption spectra. In a mutant with an (x-subunit carrying a negatively charged N-terminus no assembly of the light-harvesting complex was observed, whereas the mutation of the [3-subunit to a positively charged N-terminus led to a functional antenna, although with a decreased stability (72). In a more detailed analysis, it was shown that the subunits do not insert into the membrane independently of one another. The qx-subunit is absent in mutant membranes lacking the I~-protein, whereas the lS-subunit inserts independently of the a-protein (60). Remarkably, in a heterologous in vitro system, the oc-subunit inserts into E. coil membranes or mutant R. capsulatus membranes efficiently only if it was cosynthesized with the [3-subunit (Kiefer, unpublished results). Some highly conserved amino acid residues such as Trp8 or Pro 13 in the hydrophobic region of the o~-subunit seem to be important for a correct membrane assembly of the complex (8, 61). From the work with the light-harvesting proteins, it is obvious that one has to account for the way that the different subunits of a heterooligomeric membrane complex may insert when compared with the single polypeptide chains. It may well be that subunits of membrane-located oligomeric protein complexes insert in a synergistic manner, as proposed for monomeric multispanning integral proteins. The involvement of a translocase in these processes is yet not well characterized. Almost nothing is known about a protein export machinery in Rhodobacter, but it was suggested that cytosolic as well as membrane-bound transport components exist (77, 89). Recently, a signal peptidase has been partially characterized in this organism with similar specificities as the E. coil enzyme (88), but there is no evidence for components functionally homologous to the Sec proteins in E. coil (89). For the heat shock proteins GroEL and GroES, however, which are known to be involved in the export of l~-lactamase in E. coil, homologs in Rhodobacter sphaeroides have been characterized (76).

V. INSERTION OF SMALL PROTEINS INTO THE

ENDOPLASMIC RETICULUM

Protein transport into the endoplasmic reticulum (ER) of eukaryotic cells is in many respects similar to the mechanisms of the prokaryotic process (Figure 2). There also

DOROTHEE KIEFERand ANDREAS KUHN

24

ER-lumen

periplasm

il

J

| o

T

4.

4,

4.

'~

,IF

1 2

3

4

5

6

prokaryotlc translocatlon pathways

SRP

eukaryotictransloeatlon pathways

Figure 2. Comparative models of prokaryotic and eukaryotic insertion pathways. The prokaryotic soluble components are (B) SecB and GroE, the membrane-bound components are (A) SecA, (G) SecG, (E) SecE, (Y) SecY, (D) Sec D and F (SecF). The postulated pathways describe insertion mechanisms of (1) SecB-dependent proteins e.g. proOmpA, preMBP or preLamB; (2) SecB-independent proteins (e.g., leader peptidase or preRBP); (3) prel3-1actamasewhich uses GroE instead of SecB as a cis acting chaperone; and (4) translocase- and Sec-independent proteins M13 procoat, Pf3 coat or inverted leader peptidase. The components of the eukaryotic translocation machinery are the (SRP)signal recognition particle, (DP) docking protein, (Sec61) the Sec61 oc~7complex, (TRAM) translocating chain-associating membrane protein, (nO NEM-sensitive factor, (a0 azido-ATP-sensitive factor, (RR) ribosome receptor, (HspT0) and (Hspg0) heat shock proteins, and (BiP) the trans-acting chaperone. The postulated pathways describe the insertion mechanisms of (5) SRP-independent proteins (e.g., preprocecropin A or M13 procoat) and (6) SRP-dependent proteins (e.g., preproodactor or preprolactin). The TRAM protein is only required for the translocation of a subset of proteins and its role is not fully understood (22a). The E. coli leader peptidases and the ER signal peptidase are not depicted because they do not participate in the actual insertion or translocation process. exists a proteinaceous machinery consisting of soluble cytoplasmic (39, 52), as well as membrane-located components (22a, 25, 28, 65a). As is the case for the exported proteins in bacteria, the vast majority of ER-proteins carries a cleavable signal sequence that is recognized cotranslationally by the SRP transported to the cis side of the membrane where the precursor-SRP complex is recognized by the above mentioned membrane-bound components. During translocation across the membrane, the signal sequence is cleaved off by the signal peptidase, which in eukaryotes is a multimeric enzyme complex (18, 49). Interestingly, some ERoproteins are translocated in the absence of the central components of the export machinery. Usually these SRP-independent proteins

Membrane Imertion of Small Proteins

25

have, similar to the situation in prokaryotes, simple secondary structures and are small in size. Most of them are only 50 to 70 amino acid residues in length, are synthesized with a cleavable signal sequence, and usually are secreted from the cell-like the frog skin peptide GLa (66) or some insect venoms, discussed in detail in the following section. To our knowledge there exists only one exception of a non-secreted protein, the rat liver cytochrome bs. It is a 16-kDa integral protein of the ER-membrane and is synthesized without a cleavable signal sequence. Its membrane insertion does not depend on the presence of SRP (3). The viral M2 protein with 97 amino acid residues is an integral membrane protein in infected cells and does not contain a cleavable signal sequence. In contrast to the cytochrome, however, its membrane insertion requires SRP. It was shown that deletion of two residues in the hydrophobic domain strongly impaired translocation, while deletion of six residues completely abolished membrane insertion. As M2 requires the signal recognition particle, it was suggested that the deletion mutants were not recognized by SRP (27). One of the best studied small ER-proteins is melittin, the main component of the honeybee venom. It is synthesized as a precursor molecule (prepromelittin) of 70 amino acids with a cleavable signal sequence of 21 residues. The cleaved form is secreted as the inactive promelittin, which is subsequently processed to the mature toxin consisting of the carboxyterminai 26 amino acids (78), of which the basic C-terminus plays an important role in membrane binding allowing lysis to occur (53). Translocation of prepromelittin into dog pancreas microsomes does not require SRP or the docking protein and can also occur posttranslationally (50). Accordingly, if expressed in a prokaryotic in vitro system, prepromelittin is translocated into E. coil inner membrane vesicles in a SecA/SecY-independent fashion, as has been observed for M13 procoat in vivo (15, 31). However, for the mammalian system, treatment of the microsomes with N-ethylmaleimide (NEM) and trypsin, showed that a membrane-located proteinaceous component other than the docking protein, or any known membrane protein, is involved in prepromelittin and preprocecropin A sequestration (95, 97). Preprocecropin A is a small antibacterial peptide from the moth Hyalophora cecropia with 64 amino acid residues including the cleavable signal peptide of 26 amino acid residues (10, 28). Klappa et al. (28) discovered a novel microsomal component dependent on ATP hydrolysis that is required for membrane transport of preprocecropin A. This was shown by photoaffinity inactivation of the microsomal component and, consequently, the inhibition of precursor processing by azido-ATP. Interestingly, this membrane-bound transport component is also required for the translocation of SRP-dependent proteins such as preprolactin or yeast preproafactor (ppa) (28, 98). A striking feature of some very small secreted proteins in eukaryotes is that they are synthesized and translocated into the ER as longer precursors containing several copies of the same polypeptide. The proteins achieve their functional size by proteolytic processing. The precursor proteins are either a multiplication of a small peptide, as in the case of pp~ factor in yeast (40), or contain spacer sequences connecting the mature

26

DOROTHEE KIEFERand ANDREAS KUHN

peptides, as in the case of the thyrotropin releasing hormone in Xenopus (59). It is conceivable that very small peptides are not able to translocate through a lipid bilayer spontaneously because electrostatic and hydrophobic interactions are too weak. On the other hand, the interaction of the translocation machinery probably requires a minimal length of the secreted protein.

Vl. REQUIREMENTOF CHAPERONES The involvement of cis-acting chaperones in membrane transport has been established in different prokaryotic and eukaryotic systems. In E. coli, the best characterized systems are the GroE and the SecB chaperonins. The GroE complex consists of a 14mer of a 57-kDa protein (GroEL) and a heptarner of a 10-kDa protein (GroES). The complex appears as two stacked rings where the bound substrate is located as a molten globule in a central cavity of GroEL (!1, 48). In addition to the binding of GroE to a number of cytoplasmic proteins (71, 9d), it has been shown that GroE is essential for the efficient export of the periplasmic protein ]3-1actamase by maintaining the precursor in a loosely folded, transport-competent conformation (38, 43). A transport-competent folding of certain exported proteins is also catalyzed by the interaction with the SecB protein, which in contrast to GroE is not a heat shock protein (2). It has been shown that SecB forms a tetramer consisting of four identical subunits of 16 kDa (83). The majority of exported proteins in E. coil needs SecB for efficient translocation, although it is not essential for viability (37). SecB binds to the mature moiety of precursor proteins, probably by hydrophilic as well as hydrophobic interactions (16, '37), but it is not known what features specify a protein SecB-dependency or dependency on other chaperones. There are a number of heat shock proteins with chaperone activity, such as GroE, DnaK, DnaJ, and GrpE, that might interact with the newly synthesized proteins in a complex, cascade-like manner (23, 44). An open question is how the different chaperone proteins recognize their substrates and how they are passed on in the pathway. Nothing is known about the interaction of chaperones with small, Sec-independent membrane proteins. It might be that the small proteins like M 13 procoat bypass the stabilization by chaperones and insert into the membrane spontaneously. Alternatively, other translocation factors might be used, such as Ffh, a component of a ribonucleoparticle in E. coll. Strikingly, this protein is required for the efficient transport of the SecB-independent pre[l-lactamase and preRBP (56) and it was recently suggested that the SRP-like Ffh-ribonucleoparticle uses the FtsY protein as a receptor, a homologue of the eukaryotic docking protein (48a). Members of the heat shock protein 70 group are also required for efficient protein translocation in the ER from mammalian as well as from yeast cells (14, 19, 20, 39, 86). Strikingly, the transport of small, SRP-independent precursor proteins into mammalian microsomes is dependent on the presence of cytosolic Hsc?0 and ATP, as well as at least one additional NEM-sensitive soluble factor (85, 96). Unlike

Membrane Insertion of Small Proteins

27

preprocecropin A, the membrane transport of M 13 procoat does not require the azido-ATP sensitive membrane component described by Zimmennann and coworkers (28). In a heterologous in vitro system involving bacterial cellular extracts and dog pancreas microsomes, M 13 procoat is only processed to its mature form when the mixture has been supplemented with rabbit reticulocyte lysate or purified Hsc70 protein (87). The presence of the Hsc70 protein seems to be irrevocable for the transport process, since it could not be substituted by other chaperone-like GroE, Hsp90, Hsp60, BiP or DnaK in Mammalia (87) or BiP in yeast (13). This is in contrast to results with a prokaryotic in vitro system involving E. coli membrane vesicles, where M 13 procoat is translocated independently of the Sec components. In this system Hsc70 has no stimulating effect on membrane insertion. Hsp90 and GroE can bind M 13 procoat as shown by competition experiments, but obviously are not capable in mediating export competence of the precursor. Similarly, SecB does not support membrane insertion of M 13 procoat in E. coil and is incapable of binding (Kuhn et al., unpublished results). Addressing the problem of defining the features of SRP-dependence and -independence, respectively, it is first of all obvious that SRP-independent proteins are usually small in size. As shown by Siegel and Walter (69), there exists a critical size of a polypeptide (approximately 140 amino acid residues) that is necessary to interact with the SRP. The SRP cannot bind smaller precursor proteins probably because they are buried in the ribosome until their synthesis is completed. Thus, these short proteins have to take an alternate (-posttranslational-) pathway for membrane insertion not yet well characterized. It was shown that prepromelittin translocates into E. coil inner membrane vesicles in a SecA and SecY independent manner, whereas a longer derivative fused to dihydrofolate reductase (DHFR) needed these two components for export (15). Schlenstedt et al. (68) could demonstrate that preprocecropin A derivatives of up to 85 residues retained their SRP-independence, but longer hybrid proteins between preprocecropin A and DHFR (252 amino acid residues) were translocated in a cotranslational, SRP-dependent manner into dog pancreas microsomes. However, this construct could be translocated in a posttranslationai, SRP-independent way in the presence of the folate analogue methotrexate (67). To explain this, the authors propose that the substrate analogue leads to a looser, export-competent conformation of the fusion protein as indicated by its protease sensitivity. One could hypothesize that in an alternative export pathway, the chaperoning effect of SRP via a translational delay of precursor proteins (82, 91) can be substituted by other cytosolic chaperone such as the above mentioned Hsc70.

VII. CONCLUSIONS Spontaneous insertion of proteins into membranes seems to be limited to simple and short proteins. It involves a short stretch of hydrophobic amino acids that is

flanked by two hydrophilic regions. These two flanking regions are involved in orienting the protein in the membrane. Translocase-independent insertion has been observed in E. coli for small proteins and leads to the translocation of either the aminoterminal or the carboxyltenninal flanking region. So far, translocase-independent protein insertion has not been described in eukaryotes, although striking homologies between the eukaryotic and prokaryotic translocas systems exit (20a, 25a).

During evolution a number of different systems were developed that enable more complex proteins to transloeate across the membrane. Translocation machineries support the carboxylterminal translocation process of long polypeptide chains. To date, however, it is unknown whether a transloease enzyme complex can also support aminoterminal transloeation. To allow a topology with an extracellular location of the aminoterminus, cleavable signal sequences were added in front of the mature protein sequence. In the cytoplasm sophisticated chaperone systems assure a loose folding of the polypeptide chain that keeps the protein in an insertion-competent condition. And finally, chaperones in the periplasm control the folding of the secreted proteins and, most likely, membrane proteins as well.

ACKNOWLEDGMENTS This work was supported by the Deutsche Forschungsgemeinschaft (Ku 749/1-1 and Ku749/1-2).

REFERENCES 1. Allen, J. P., Feher,G., Yeatcs,T. O., Komiya,H., &Rccs, D. C. (1987). Structure of the reaction center from Rhodobacter sphaeroides R-26: the protein subunits. Proc. Natl. Acad. Sci. USA 84, 6162-6166. 2. Alunan,E., Kumarnoto,C. A., & Farnr,S. D. (1991). Heat-shockproteins can substitute for SecB function during protein export in Escherichia coll. EMBO J. 10, 239-245. 3. Anderson, D., Mostov, K. E., & Blebel, G. (1983). Mechanismsof integration of de novo-synthesized polypeptidesinto membranes:signal-recognitionparticleis requiredfor integrationinto microsomal membranes of calcium ATPase and of lens MP26 but not of cytochromeI)5. Proc. Natl. Acad. Sci. USA 80, 7249-7253. 4. Anderssm, H., &von Heijne, G. (1991). A 30 residue long "export initiation domain" adjacent to the signal sequence is critical for protein translocation across the inner membrane of Escherichia coll. Proc. Natl. Acad. Sci. USA 88, 9751-9754. 5. Andersson,H., & yon Heijne, G. (1993). Sec dependent and sec independentassemblyof E. coli inner membraneproteins: the topologicalrules depend on chain length. EMBO J. 12, 683-691. 6. Andersson, H., Bakker, E., & yon Heijne, G. (1992). Different positively charged amino acids have similareffects on the topology of a polytopic transmembraneprotein in Escherichia coll. J. Biol. Chem. 267, 1491-1495. 7. Andrews, D. W., Young,J. C., Mirels, L. E, & Czarnota, G. J. (1992). The role of the N region in signal sequence and signal-anchorfunction. J. Biol. Chem. 267, 7761-7769.

Membrane Insertion of Small Proteins

29

8. Babst, M., Albrecht, H., Wegmann, I., Brunisholz, R., & Zuber, H. (1991). Single amino acid substitutions in the B870 a and 13light-harvesting polypeptides of Rhodobacter capsulatus. Eur. J. Biochem. 202, 277-284. 9. Beltzer, J. F., Fiedler, K., Fuhrer, C., Geffen, I., Handschin, C., Wessels, H. E, & Spiess, M. (1991). Charged residues are major determinants of the transmembrane orientation of a signal-anchor sequence. J. Biol. Chem. 266, 973-978. 10. Boman, H. G., Boman, I. A., Andreu, D., Li, Z., Merrifield, R. B., Schlenstedt, G., & Zimmermann, R. (1989). Chemical synthesis and enzymic processing of precursor forms of cecropins A and B. J. Biol. Chem. 264, 5852-5860. 11. Braig, K., Simon, M., Furuya, F., Hainfeld, J. F., & 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. 12. Breton, J., & Nabedryk, E. (1984). Transmembrane orientation of a-helices and the organization of chlorophylls in photosynthetic pigment-protein complexes. FEBS Lett. 176, 355-359. 13. Brodsky, J., Hamamoto, S., Feldheim, D., & Schekman, R. (1993). Reconstitution of protein translocation from solubilized yeast membranes reveals topologically distinct roles for BiP and cytosolic Hsc70. J. Cell Biol. 120, 95-102. 13a.Cao, G., Cheng, S., Whitley, P., yon Heijne, G., Kuhn, A., & Dalbey, R. E. (1994). Synergistic insertion of two hydrophobic regions drives Sec-independent membrane protein assembly. J. Biol. Chem. 269, 26898-26903. 13b.Cao, G., & Dalbey, R. E. (1994). Translocation of N-terminal tails across the plasma membrane. EMBO J. 13, 4662--4669. 14. Chirico, W. J., Waters, M. G., & Biobel, G. (1988). 70 k heat shock related proteins stimulate protein transiocation into microsomes. Nature 332, 805-810. 15. Cobet, W. W. E., Mollay, C., M011er, G., & Zimmermann, R. (1989). Export of honeybee prepromelittin in E. coli depends on the membrane potential but does not depend on proteins SecA or SecY. J. Biol. Chem. 264, 10169-10176. 16. Collier, D. N., Strobel, S. M., & Bassford, P. J. (1990). SecB-independent export of Escherichia coli Ribose-Binding Protein (RBP): some comparisons with export of maltose-binding protein (MBP) and studies with RBP-MBP hybrid proteins. J. Bacteriol. 172, 6875--6884. 17. Dalbey, R. E. (1991). Leader peptidase. Mol. Microbiol. 5, 2855-2860. 18. Dalbey, R. E., & yon Heijne, G. (1992). Signal peptidases in prokaryotes and eukaryotes--a new protease family. Trends Biochem. Sci. 17, 474-478. 19. Deshaies, R. J., Koch, B. D., & Schekman, R. (1988). The role of stress proteins in membrane biogenesis. Trends Biochem. Sci. 13, 384-388. 20. Deshaies, R. J., Koch, B. D., Werner-Washburne, M., Craig, E., & Schekman, R. (1988). A subfamily of stress proteins facilitates translocation of secretory and mitochondrial precursor polypeptides. Nature 332, 800-805. 20a.Dobberstein, B. (1994). On the beaten pathway. Nature 367, 599--600. 21. Dtrge, B., Klug, G., Gad'on, N., Cohen, S. N., & Drews, G. (1990). Effects on the formation of antenna complex B870 of Rhodobacter capsulatus by exchange of charged amino acids in the N-terminal domain of the {xand 13pigment-binding proteins. Biochemistry 29, 7754-7758. 22. Geller, B., Zhu, H. Y., Cheng, S., Kuhn, A., & Dalbey, R. E. (1993). Charged residues render proOmpA potential-dependent for initiation of membrane translocation. J. Biol. Chem. 268, 9442-9447. 22a.GOrlich, D., & Rapoport, T. A. (1993). Protein translocation into proteoliposomes reconstituted from purified components of the endoplasmic reticulum membrane. Cell 75, 615-630. 23. Gragerov, A., Nudler, E., Komissarova, N., Gaitanaris, G. A., Gottesmann, M. E., & Nikiforov (1992). Cooperation of GroEL/GroES and DnaK/ aJ heat shock proteins in preventing protein misfolding in Escherichia coli. Proc. Natl. Acad. Sci. USA 89, 10341-10344.

30

DOROTHEE KIEFERand ANDREAS KUHN

24. Haeuptle, M.-T., Flint, N., Gongh, N. M., & Dobberstein, B. (1989). A tripartite structure of the signals that determine protein insertion into the endoplasmic reticulum membrane. J. Cell Biol. 108, 1227-1236. 25. Harunann, E., & R ~ T. A. (1992). Ttanslocation of proteins through the endoplasmic membrane: investigation of their molecular environment by cross-linking. In W. Neupert, and R. Lill, (Eds.), Membrane Biogenesis and Protein Targeting (pp. 119-127). Amsterdam: Elsevier Science. 25a.Hartmann" E., Sommer, T., Prelm, S., GOrlich, D., Jentsch, S., & Rapopo~, T. A. (1994). Evolutionary conservation of components of the protein translocation complex. Nature 367, 654--657. 26. High, S., Flint, N., & Dobberstein, B. (1991). Requirements for the membrane insertion of signal-anchor type proteins. J. Cell Biol. 113, 25-34. 27. Hull, J. D., Gilmore, R., & Lamb, R. A. (1988). Integration of a small integral membrane protein, M2, of influenza virus into the r reticulum: analysis of the internal signal-anchor domain of a protein with an ectoplasmic NH2 terminus. J. Cell Biol. 106, 1489-1498. 28. Klappa, P., Mayinger, P., Pipkom, R., Zimmermann, M., & Zimmermann, R. (1991). A microsomal protein is involved in ATP-dependent transport of presecretory proteins into mammalian microsomes. EMBO J. 10, 2795-2803. 29. Krajcsi, P., Toilefson, A. E., Anderson, C. W., & Wold, W. M. (1992). The adenovirus E3 14.5-kilodalton protein, which is required for down-regulation of the epidermal growth factor receptor and prevention of tumor necrosis factor cytolysis, is an integral membrane protein oriented with its C terminus in the cytoplasm. J. l~rol. 66, 1665-1673. 30. Kuhn, A. (1987). Bacteriophage MI3 procoat protein inserts into the plasma membrane as a loop structure. Science 238, 1413-1415. 31. Kuhn, A. (1988). Alterations in the extracellular domain of MI3 procoat make its membrane insertion dependent on secA and secY. Eur. J. Biochem. 177, 267-271. 32. Kuhn, A., Kreil, G., & Wickner, W. (1986). Both hydrophobic domains ofMl3 procoat initiate membrane insertion. EMBO J. 5, 3681-3685. 33. Kuhn" A., Kreil, G., & Wickner, W. (1987). Recombinant forms ofMl3 procoat with an OmpA leader sequence or a large c a r b o x y t ~ a i extension retain their independence of secY function. EMBO J. 6, 501-505. 34. Kulm, A., Rohrer, J., & Gallusser, A. (1990). Bacteriophage MI3 and Pf3 tell us how proteins insert into the membrane. J. Struct. Biol. 104, 38-43. 35. Kuhn, A., Zhu, H.-Y., & Daibey, R. E. (1990). Efficient translocation of positively charged residues of MI 3 procoat protein across the membrane excludes electrophoresis as the primary force for membrane insertion. EMBO J. 9, 2385-2388, 2429. 36. Kumarnoto, C. A. (1991). Molecular chaperones and protein transiocation across the Escherichia coil inner membrane. Mol. Microbial. 5, 19-22. 37. Kumamoto, C. A., & Beckwith, J. (1985). Evidence for specificity at an early step in protein export in Escherichia coli. J. Bacterial. 163, 253-260. 38. Kusukawa, N. T., Yura, C., Ueguchi, C., Akiyama, Y., & Ito, K. (1989). Effects of mutations in heat shock genes groES and groEL on protein export in Escherichia coll. EMBO J. 8, 3517-3521. 39. Kurihara, T., & Silver, P. (1992). DnaJ homologs and protein transport. In W. Neupert, and R. Lill (Eds.), Membrane Biogenesis and Protein Targeting (pp. 309--327). Amsterdam: Elsevier Science. 40. Kurjan, J., & Herskowitz, I. (1982). Structure of a yeast pheromone gene (MFtz): a putative tz-factor precursor contains four tandem copies of mature or-factor. Cell 30, 933-943. 41. Kuroiwa, T., Sakaguchi, M., Mihara, K., & Omura, T. (1990). Structural requirements for interruption of protein translocation across rough endoplasmic reticulum membrane. J. Biochem. 108, 829-834.

Membrane Insertion of Small Proteins

31

42. Kuroiwa, T.,Sakaguchi, M., Mihara, K., & Omura, T. (199 I).Systematic analysisof stop-transfer sequence for microsomal membrane. J. Biol. Chem. 266, 9251-9255. 43. Laminet, A. A., Ziegelhoffer,T.,Georgopoulos, C., & Plilckthun,A. (1990).The Eacherichiacoli heat shock proteinsGroEL and GroES modulate the foldingof the ~lactanm~ precursor.E M B O J. 9, 2315-2319. 44. Langer, T.,Lu, C., Echols, H., Flanagan, J.,Hayer, M. K., & Hard, F.-U.(1992). Successive action of DnaK, DnaJ & GruEL along the pathway of chaperone-mediated proteinfolding.Nature 356, 683-689. 45. Laws, J.,& Dalbey, R. E. (1989). Positivecharges in the cytoplasmic domain of Escherichiacoli leader peptidase prevent an apolar domain from functioning as a signal.E M B O J. 8, 2095-2099. 46. Lee J.,Kuhn, A., & Dalbey, R. E. (1992). Distinctdomains of an oligotopicmembrane protein are Scc-dependent and Sec-independent for membrane insertion.J. Biol.Chem. 267, 938--943. 47. Li, P.,Bcckwith, J.,& Inouye, H. (I988). Alterationof the amino terminus of the mature sequence of a periplasmic proteincan severely affectproteinexport in Escherichiacoll.Proc. Natl.Acad. Sci. USA 85, 7685-7689. 48. Martin, J., Langer, T., Botcva, R., Schramel, A., Horwich, A. L. & Hanl, F.-U. (1991). Chaperonin-mediated protein folding at the surface of GroEL through a "moltcn-globule" intermediate.Nature 352, 36-42. 48a.Miller, J. D., Bemstein, H. D., & Walter, P. (1994). Interactionof E. coilFfh/4.5 S ribonuclcoprotein and FtsY mimics thatof mammalian signalrecognition particleand itsreceptor.Nature 367, 657-659. 49. MUller, M. (1992). Proteolysisin proteinimport and export: signalpeptide processing in eu- and prokaryotes. E.rperientia48, 118-129. 50. MUller, G., & Zimmermann, R. (1987). Import of honeybee prepromelittininto the endoplasmic reticulum: structuralbasisfor independence of SRP and docking protein.E M B O J. 6, 2099-2107. 5 I. Nilsson,I.& yon Heijnc, G. (I990). Fine-tuning the topology of a polytopic membrane protein. Cei162, 1135-1141. 52. Ogg, S. C., Nunnari, J. M., Miller,J.D., & Walter,P. (1992).The role of GTP in protein targeting to the endoplasmic reticulum. In W. Neupert, and R. Lill (Eds.),Membrane Biogenesis and Protein Targeting(pp. 129-136). Amsterdam: ElsevierScience. 53. Otoda, K., Kimura, S.,& Imanishi,Y. (1992). Interactionof melittinderivativeswith lipidbilayer membrane. Role of basic residues at the C-terminal and theirreplacement with lactose.Biochim. Biophys. Acta. 1112, I--6. 54. Parks, G. D., & Lamb, R. A. (1991). Topology of eukaryotic type II membrane proteins: Importance of N-terminal positivelycharged residuesflankingthe hydrophobic domain. Cell64, 777-787. 55. Parks, G. D., Hull, J. D., & Lamb, R. A. (1989). Transposition of domains between the M2 and HN viralmembrane proteins resultsin polypeptides which can adopt more than one membrane orientation.J. Cell Biol. 109, 2023-2032. 56. Phillips,G.J., & Silhavy,T. J.(1992). The E. colifjhgene isnecessary for viabilityand efficient protein export. Nature 359, 744-746. 57. Randall, L. L., (1992). Peptide binding by chaperone SccB: implications for recognition of nonnative structure.Science 257, 241-245. 58. Rcnaud, K. J., Inman, E. M., & Fambrough, D. M. (1991). Cytoplasmic and transmembrane domain deletions of Na, K-ATPase ~-subunit. Effects on subunit assembly and intracellular transport.J. Biol.Chem. 266, 20491-20497. 59. Richter,K., Kawashima, E., Egger, R., & Krcil,G. (1984). Biosynthesis of thyrotropinreleasing hormone in the skin of Xenopus laevis:partialsequence of the precursor deduced from cloned DNA. E M B O J. 3, 617-62 I. 60. Richter,P.,& Drews, G. (1991). Incorporationof light-harvestingcomplex I a and 15polypeptides into the intracytoplasmicmembrane of Rhodobacter capsulatus.J. Bacteriol.173, 5336-5345.

32

DOROTHEE KIEFERand ANDREAS KUHN

61. Richter, P., Cortez, N., & Drews, G. (1991). Possible role of the highly conserved animo acids

Trp-8 and Pro-13 in the N-terminal segment of the pigment-binding polypeptide LHI r of Rhodobacter capsulatus. FEB$ Lett. 285, 80-84. 62. Rohrer J., & Kuhn, A. (1990). The function of a leader peptide in translocating charged amino acyl residues across a membrane. Science. 250, 1418-1421. 63. Saier, M. H., Wemer, P. K., & M011er,M. (1989). Insertion of proteins into bacterial membranes: Mechanism, characteristics, and con~t~'~xisonwith the eukaryotic process. Microbiol. Rev. 53, 333-366. 64. Sakaguchi, M., Tomiyoshi, R., Kuroiwa, T., & Mihara, K. (1992). Functions of signal and signal-anchor sequences are determined by the balance between the hydrophobic segment and the N-terminal charge. Proc. Natl. Acad. Sci. USA 89, 16-19. 65. Sato, T., Sakaguchi, M., Mihera, K., & Omura, T. (1990). The amino-terminal structures that determine topological orientation of cytochrome P-450 in microsomal membrane. EMBO J. 9, 2391-2397. 65a.Savitz, A. J., & Meyer,D. I. (1990). Identificationof a ribosome receptor in the rough endoplasmic reficulum. Nature 346, 540-544. 66. Schlenstedt, G., & Zimmermann, R. (I 987). Import of frog prewopeptide GLa into microsomes requires ATP but does not involve docking protein or ribosomes. EMBO J. 6, 699-703. 67. Schlenstedt, G., G n d m u ~ o n , G. H., Boman, H. G., & Zimmermann, R. (1990). A large presecretory protein mmslocates both cocanslationagy, using signal recognition particle and ribosome, and posttramlafionally, without these ribonucleopardcles, when synthesized in the presence of mammalian ndcrosomes. J. Biol. Chem. 265, 13960-13968. 68. Schlenstedt, G., Gudmundsson, G. H., Boman, H. G., & Zimmermann, R. (1992). Structural requirements for transport of ~ e c r o p i n A and related presecretory proteins into mammalian microsomes. J. Biol. Chem. 267, 24328-24332. 69. Siegel, V., & Walter, P. (1988). The affinity of signal recognition particle for presecretory proteins is dependent on nascem chain length. EMBO J. 7, 1769-1775. 70. Simon, S. M., & BIobel, G. (1991). A proteiwconducfing channel in the endoplasmic reticulum. Cell 65, 371-380. 71. Sohlberg, B., Lundberg, U., Hartl, E-U., & yon Gabain, A. (1993). Functional interaction of heat shock protein GroEL with an RNase E-like activity in Escherichia coll. Proc. Natl. Acad. $ci. USA 90, 277-281. 72. Stiehle, H., Cortez, N., Kiug, G., & Drews, G. (1990). A negatively charged N terminus in the ct polypeptide inhibits formation of light-harvesting complex in Rhodobacter capsulatus. J. Bacteriol. 172, 7131-7137. 73. Tadros, M. H., Suter, E, Seydewitz, H. H., Win, I., Zuber, H., & Drews, G. (1984). Isolation and complete amino-acid sequence of the small polypeptide from light-harvesting pigment-protein complex I (B870) of Rhodopseudomonas capsulato~ Fur. J. Biochem. 138, 209-212. 74. Tadros, M. H., Frank. G., Zuber, H., & Drews, G. (1985). The complete amino acid sequence of the large bacteriochlorophyll-binding polylgl~ide B870 r from the light-harvesting complex B870 of Rhodopseudomonas capsulata. FEBS Lett. 190, 41-44. 75. Tadros, M. H., Frank, R., Dikge, B., Gad'on, N., Takemoto, J. Y., & Drews, G. (1987). Orientation of the B800-850, B870, and reaction center polypeptides on the cytoplasmic and periplasmic surfaces of Rhodobacter capsulatus membranes. Biochemistry 26, 7680-7687. 76. Terlesky, K. C., & Tabita, E R. (1991). Purification and characterization of the chaperonin 10 and chaperonin 60 proteins from Rhodobacter sphaeroides. Biochemistry 30, 8181-8186. 77. Troschel, D., & Milller, M. (1990). Developmem of a cell-flee system to study the membrane assembly of photosynthetic proteins of Rhodobacter capsulotus. J. Cell Biol. 111, 87-94. 78. Vlasak. R., Unger-Ullmann, C., Kreil, G., & Frischanf, A.-M. (1983). Nucleotide sequence of cloned cDNA coding for honeybee prewomelittin. Fur. J. Biochem. 135, 123-126.

Membrane Insertion of Small Proteins

33

79. yon Heijne, G. (I 986). The distribution of positively charged residues in bacterial inner membrane woteins correlates with the trans-membrane topology. EMBO J. 5, 3021-3027. 0. yon Heijne, G. (1989). Control of topology and mode of assembly of a polytopic membrane protein by positively charged residues. Nature 341,456-458. 81. yon Heijne, G., & Gavel, Y. (1988). Tolx~enic signals in integral membrane proteins. Fur./. Biochem. 174, 671-678. 82. Walter, P., & Blobel, G. (1981). 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. 83. Watanabe, M., & Blobel, G. (1989). Cytosolic factor purified from Escherichia coli is necessary and sufficient for the export of a preprotein and is a homotetramer of SecB. Proc. Natl. Acad. Sci. USA 86, 2728-2732. 84. Wickner, W. T. (1980). Assembly of proteins into membranes. Science 210, 861-868. 85. Wiech, H., Sagstetter, M., Mtiller, G., & Zimmermann, R. (1987). The ATP requiring step in assembly of M I3 procoat protein into microsomes is related to preservation of transport competence of the precursor protein. EMBO J. 6, 1011-1016. 6. Wiech, H., Smart, R., & Zimmermann, R. (1990). Role of cytosolic factors in the transport of proteins across membranes. Sere. Cell Biol. 1, 55-63. 87. Wiech, H., Buchner, J., Zimmermann, M., Zimmermann, R., & Jakob, U. (1993). HscT0, immunoglobulin heavy chain binding protein, and Hsp~ differ in their ability to stimulate transport of precursor proteins into mammalian cells. J. Biol. Chem. 268, 7414-7421. 88. Wieseler, B., Schiltz, E., & MUller, M. (1992). Identification and solubilization of a signal peptidase from the phototrophic bacterium Rhodobacter capsulatus. FEBS Lett. 298, 273-276. 89. Wieseler, B., & M011er, M. (1993). Translocation of precytochrome c2 into intracytoplasmic membrane vesicles of Rhodobacter capsulatus requires a peripheral membrane protein. Mol. Microbiol. 7, 167-176. 0. Wolfe, P. B., Rice, M., & Wickner, W. (1985). Effects of two sec genes on protein assembly into the plasma membrane of Escherichia coli. J. Biol. Chem. 260, 1836-1841. 91. Wolin, S. L., & Walter, P. (1989). Signal recognition particle mediates a transient elongation arrest of preprolactin in reticulocyte lysate. J. Cell Biol. 109, 2617-2622. 92. Yamane, K., & Mizushima, S. (1988). Introduction of basic amino acid residues after the signal peptide inhibits protein translocation across the cytoplasmic membrane of Escherichia coli. Relation to the orientation of membrane proteins. J. Biol. Chem. 263, 19690-19696. 93. Youvan, D. C., Bylina, E. J., Alberti, M., Begusch, H., & Hearst, J. E. (1984). Nucleotide and deduced polypepdde sequences of the photosynthetic reaction-center, B870 antenna, and the flanking polypeptides from R. capsulata. Cell 37, 949-957. 4. Zeiistra-Ryalls, J., Fayet, O., & Georgopoulos, C. (1991). The universally conserved GroE (Hspt0) chaperonins. Annu. Rev. Microbiol. 45, 301-325. 95. Zimmermann, R., & Mollay, C. (1986). Import of honeybee prepromelittin into the endoplasmic reticulum. J. Biol. Chem. 27, 12889-12895. . Zimmermann, R., Sagstetter, M., Lewis, M. J., & 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. 97. Zimmennann, R., Sagstetter, M., & Schlenstedt, G. (1990). Ribonucleoparticle-independent import of proteins into mammalian microsomes involves a membrane protein which is sensitive to chemical aikylation. Biochim. 72, 95-101. 98. Zimmermann, R., Zimmermann, M., Mayinger, P., & Klappa, E (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. 9. Zuber, H. (2986). Structure of light-harvesting antenna complexes of photosynthetic bacteria, cyanobacteria and red algae. Trends Biochem. Sci. 11,414--419.

This Page Intentionally Left Blank

PROTEIN TRANSLOCATION GENETICS

Koreaki Ito

I. ll. Ill.

IV.

V.

VI.

Vll.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Identification of Escherichia Coil Genes Involved in Protein Export . . . . . . Cytoplasmic Factors, Precursor Targeting, and Anti-Folding . . . . . . . . . . A. SecB and Other Chaperones . . . . . . . . . . . . . . . . . . . . . . . . B. SRP Homologs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SecA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Membrane-Embedded Components of the Translocator . . . . . . . . . . . . . A. SecY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. SecE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. SecD and SecF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Recognition of Signal Peptide and Early Part of Mature Sequence by the Translocator . . . . . . . . . . . . . . . . . . . . . . A. The prlA Mutations in SecY . . . . . . . . . . . . . . . . . . . . . . . . B. prig and Suppressor-Directed Inactivation . . . . . . . . . . . . . . . . . Sec Factor Interaction and lntramembrane Translocation Pathway . . . . . . . A. Intramembrane Steps of Translocation . . . . . . . . . . . . . . . . . . . B. Bieker-Silhavy Model of Translocation Complexes . . . . . . . . . . . . C. Insights into the SecY-SecE Interaction . . . . . . . . . . . . . . . . . . .

Advances in Cell and Molecular Biology of Membranes and Organdies Volume 4, pages 35-60. Copyright 9 1995 by JM ~ Inc. All rights of reproduction in any form reserved. ISBN: I-$$938-924-9 35

36 36 38 38 39 40

41 41 43 44 44 44 45 47 47 48 49

KOREAKI ITO

36

VIII. AdditionalFactors Possibly Involved in Translocation . . . . . . . . . . . . . . A. Band I andPl2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

B. C. D. E.

PspA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Skp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ydr . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Orfl2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E FtsH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G. Are There Translocation Factors in the Periplasm? . . . . . . . . . . . . . IX. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

51 51 51 51 51 52 52 52 53 53 54

I. INTRODUCTION Biological membranes are restrictive to the passage of small and macromolecules and, hence, maintain the integrity of the cell as well as its intracellular compartments. At the same time, membranes are surprisingly dynamic, asymmetrically interfacing the communication between aqueous milieus on both sides. A class of proteins that must localize out of the cytoplasm (the compartment fed with the genetic message) experience, once in their lives, a critical process of translocation across the membrane. While extracytoplasmic proteins of prokaryotic cells and eukaryotic cells cross different membranes (the plasma [cytoplasmic] membrane vs the endoplasmic reticulum membrane), their mechanisms of translocation appear to be well conserved. The conservation includes the exchangeability of substrate secretory proteins and their signal sequences (1) as well as the occurrence of the SecY/SEC61p family of integral membrane "translocator" components (2). In addition, at least some elements of the signal recognition particle exist and seem to function in bacterial cells. Thus, elucidation of the translocation mechanisms in any system will positively contribute to the understanding of similar processes in other organisms. This review is intended to update the progress in studies of the protein translocation system in E. coli, with emphasis placed on genetic analyses of the translocation machinery. More general accounts of the secretory pathway in gramnegative bacteria have recently been reviewed (3).

II. IDENTIFICATION OF ESCHERICHIA COLI GENES INVOLVED IN PROTEIN EXPORT Because the selection and screening systems that have been successfully designed for obtaining mutants with altered activities of protein export have been reviewed extensively (4), only outlines of these approaches will be summarized. Fusion proteins between the N-terminal regions of exported proteins, such as MalE, LamB, and the cytoplasmic ~-galactosidase (LacZ), play central roles in the

Protein Translocation Genetics

37

genetics of protein export. Such fusion proteins can initiate translocation but cannot complete it because of the presence of the translocation-incompatible sequence or conformation within the LacZ part (5); they jam up the translocation machinery when induced by maltose, rendering the cell carrying them maltose-sensitive. ~l-galactosidase activity of the fusion proteins is low, presumably because the membrane localization does not effectively allow the formation of the active tetramer structure. The maltose-sensitive phenotype provided conditions for the initial isolation of the signal sequence mutations that conferred maltose-resistance (4). The Lac- phenotype allowed the isolation of not only the signal sequence mutations on the fusion gene, but also the secA and secB mutations that conferred lactose utilizing ability. Similar Lac+ selection using the phoA-lacZ fusion allowed the isolation of the secD and secF mutations (4). More recently, a Lac + selection using a malF-lacZ fusion, in which PhoA was attached to a periplasmic domain of the MaIF membrane protein, allowed the isolation of mutations in the dsbA and dsbB genes, whose products are involved in disulfide bond formation of proteins after export (6, 7). A clear and simple indication of an export defect is the accumulation of precursor proteins with unprocessed signal peptide. Direct screening for precursor-accumulating mutants, without involving the bias imposed by a particular selection, yielded mutations in the secY gene (8--10). Another "brute force" screening for plasmids overexpressing the leader (signal) peptidase activity allowed the identification of the lep gene (11); (note that this enzyme [12] will not be discussed further because its activity is not essential for the protein translocation reaction itself [13]). The most powerful method of isolating export-impaired mutants is to use the peculiar regulatory property of the secA gene whose level of expression elevates in response to an export defect (other than that caused by a defect in cytoplasmic factors such as SecB). Using cells carrying the secA-lacZ gene fusion, mutants with increased expression of the fusion were selected or screened for by virtue of their increased ~-galactosidase activity (14). Mutants of the secE, sec u secDlsecF, and secA genes were thus isolated. The wide spectrum of the sec mutations isolated by the secA-lacZ approach raised the possibility that the "sec" genes had been exhaustively identified in E. coil (4, 15). It should be noted, however, that the secA-lacZ method does not unravel defects in the cytoplasmic events, and the earlier Lac § method did not yield a secY mutant. It seems safe to say that each selection method contains inherent bias and that not all the genes for protein export have been discovered. Although secB is dispensable (at least in synthetic media), other sec gene functions appear essential for cell viability (as has conclusively been shown for secE; [16]), since their mutational alterations can give conditionally lethal phenotypes. Strikingly, most conditional mutations are cold-sensitive (14). Temperaturesensitive export mutations are only known for secA (17) and secY (9, I 0). Pogliano and Beckwith (18) proposed that the E. coil export pathway includes some intdn-

38

KOREAKIITO

sicaUy cold-sensitive steps, and lowered activity at a step following the cold-sensitive step invariably leads to the cold-sensitive export phenotypes. Another successful approach to identifying export components was the prl selection in which mutations that suppress export defects caused by signal sequence mutations were selected (19). The prlA, prlD, and prig mutations that were isolated are alleles of secY, secA and secE, respectively (19-21). A more specific genetic strategy for dissecting the translocation mechanisms and additional components that possibly participate in the system will be discussed in the following sections.

III. CYTOPLASMIC FACTORS, PRECURSOR TARGETING, AND ANTI-FOLDING A. SecBand Other Chaperones Protein export in E. coil occurs both posttranslationally and before polypeptide chain completion. The latter mode of translocation could be called cotranslational but it occurs after a substantial proportion (some 80%) has been synthesized; the polypeptide chain does not move through the membrane residue-by-residue as it elongates (22). The cellular number of the translocation machinery in the membrane (about 500 per cell) (23) appears to be too small to account for the strictly cotranslational biogenesis of cell envelope proteins. In other words, such a small number of translocators should translocate envelope proteins in periods shorter than those required for their translation (23). Tightly folded precursor or its segments cannot effectively enter or complete the translocation pathway (24, 25). Thus, E. coil cells possess cytoplasmic factors that prevent premature folding of secretory precursors (26). In the secB mutant cells, a subset of envelope proteins (OmpA, MalE, PhoE, LamB, and OmpF) are retarded in their export. The secB mutational effects are epistatic or additive to phenotypes of many other export mutations (27, 28), placing its action point early in the export pathway. Purified SecB protein binds to MalE and other precursor proteins, after their in vitro synthesis or denaturation, and keep them competent for subsequent translocation across inverted membrane vesicles (29). It also retards refolding of urea-denatured proteins (30). In vivo, mutations in the mature part of exported proteins that destabilize the protein conformation exhibit decreased dependence on secB (31). These observations established the notion that SecB is an anti-folding factor and important for the posttranslational mode oftranslocation. However, SecB is, in fact, cotranslationally associated with incomplete nascent chains of some exported proteins and found in the polysomal fraction (32). Biochemical studies indicate that SecB binds rather indiscriminately to unfolded proteins, which can be in kinetic competition between SecB-binding and folding (30). Closer examinations revealed that SecB, under low ionic strength, has some specificity to polypeptides enriched in positively charged residues, whereas hydro-

Protein Translocation Genetics

39

phobic interactions are probably more important at physiological ionic conditions (33). Since signal peptides retard the folding of the mature part (34), the apparent specificity of SecB for signal sequences (35) could be secondary. The "translocation-competent" states of precursors could have substantial tertiary structure (36) or could be more unfolded (30, 33), depending upon the precursors examined. Nascent chains bound to SecB in vivo are mainly secretory proteins such as MalE, LamB, OmpE and OmpA (32). Signal sequence exchange experiments between MalE (secB-dependent) and ribose-binding protein (secB-independent) showed that the mature portions are solely responsible for their SecB dependence/independence (37). However, such discrimination is not absolute because ribose-binding protein becomes partially secB-dependent when its export efficiency is lowered

(28). The fact that the secB nul mutation (secB::Tn5) does not confer the Lac § phenotype to MalE-LacZ fusion-bearing cells (38) suggests that some other factors might substitute for the simple loss of function of SecB. In other words, the original secB missense mutations as well as those isolated recently have some dominant characters (38). The latter mutations have been shown to increase SecB's affinity for unfolded proteins (38). SecB's ability to release secretory precursors should be important for targeting the membrane-associated components of the export system; SecB is believed to have a targeting role in the cell (26, 35). Cell-free experiments showed that SecB can bind to SecA (39), the peripheral component of the translocator, although genetic evidence for the SecB-SecA interaction is still to be provided. The observation that a mutation in the secA locus that creates a duplication of most of the SecA sequence partially compensating for the loss of SecB function (40) is consistent with the notion that SecAreceives precursors from SecB. Cytoplasmic accumulation of secretory precursors under secB-deficient conditions induces heat shock proteins (41). Further induction of heat shock proteins, DnaK and DnaJ in particular, compensates partially for the loss of SecB, whereas the secB nul and the dnaK/dnaJ mutations are synthetically lethal (42, 43). Thus, the DnaK/DnaJ chaperone can substitute for the SecB function to a certain extent. Export of PhoA may naturally be dependent on DnaK/DnaJ (42), while it also depends on SecB at low temperature (44). In contrast, ~-lactamase (Bla) is totally independent of secB in vivo and in vitro (45). The export of Bla is retarded in the groEL or groES Ts mutants (44), as well as under the conditions of GroE depletion (46). Taken together with the fact that Bla precursor is a preferred substrate of GroEL (47), it is indicated that Bla utilizes this chaperone as the cytoplasmic escort factor.

B. SRP Homologs Prokaryotic cells possess homologs of some components of the signal recognition particle that mediates cotranslational attachment of the translocation complex containing nascent secretory precursors to the endoplasmir reticulum membrane.

40

KOREAKI ITO

The 4.5 S RNA of E. coli (thefts gene product) has a sequence homology to the 7 SL RNA (48) of SRP, and the./~ gene product is homologous to an SRP subunit, SRP54 (49). Also, the ftsY gene product is homologous to the SRP receptor r subunit (50). The C-terminal "methionine-rich" domain of SRP54 may interact with a signal peptide while its N-terminal "GTPase" domain engages in the SRP-SRP receptor interaction (2). While evidence indicates that the 4.5 S RNA is involved in general protein synthesis (51), it also associates with the Ffh protein in a ribonucleoprotein complex that interacts with the signal sequence part of nascent preprolactin in a heterologous cell free system (52). Depletion or overexpression offfs causes retardation of Bla export but not of other exported proteins (53, 54). Depletion of Ffh causes growth arrest and export retardation of not only Bla, which is most severely blocked, but also OmpA, MalE, and PhoA (55). The Ffh protein can substitute for SRP54 in a mixed reconstitution with the eukaryotic SRP components with respect to the particle formation and the translation arrest activities (56). However, the chimeric particle cannot interact with the eukaryotic SRP receptor and cannot facilitate translocation activity. These results suggest that the SRP homologs in E. coli have a role in preprotein export, possibly in the targeting process. It is interesting to note that, while the eukaryotic SRP is believed to act in the cotranslational pathway of translocation, Bla that is most clearly affected by the ffs.ffh mutations shows the slowest (posttranslational) export kinetics in E. coli (57). The relationship between the SRP pathway and the Sec pathway is not clear at present. A similar situation also holds for protein translocation in Saccharomyces cerevisiae. In these cases, different proteins show different dependence on the two systems (58). The SRP pathway may not be unique in its timing of action but may be unique in that it recognizes the signal peptide segment of the precursor protein for targeting to the membrane. Thus, it could be complementary to the chaperone pathway that recognizes the mature part.

IV. SECA SecA has recently been reviewed by Oliver (59) and detailed discussion on its biochemical properties has been provided. SecA does not contain any segment that is sufficiently hydrophobic to traverse the membrane, but a fraction of it is associated with the membrane. While its membrane association under certain conditions is strong enough to be regarded as "integral" (60-62), a major fraction of SecA is in the cytoplasm or only loosely bound to the membrane (60). SecA is essential for the in vitro translocation reaction (63). It possesses an ATPase activity (translocation ATPase) that is activated by anionic phospholipids as well as vesicles containing functional SecY/SecE proteins (64). SecA translocation ATPase is inhibited by azide. The identification of the azide-resistance gene with secA indicates that secA is the primary target of low azide concentration in E. coli (65).

Protein Translocation Genetics

41

SecA binds to precursor molecules probably by recognizing both the N-terminal basic region and the hydrophobic core of the signal peptide as well as the mature part (64, 66, 67). Genetically, the occurrence of prlD mutations is consistent with SecA's ability to recognize the signal sequence. Biochemical results and "intracistronic" complementation suggest that SecA acts as a dimer (60, 68). ATP- and preprotein-binding sites are located at the N-terminal region (69, 70). Model experiments in vitro show that binding of ATP to SecA stimulates the SecA-preprotein binding (71), allowing subsequent penetration of some 20 residues of the preprotein (72). ATP hydrolysis then allows release of the preprotein from SecA. Translocation of the preprotein can be continued by repeating the cycles of SecA-binding, translocation of the limited segments, and release. If proton motive force is imposed, more efficient translocation occurs upon the ATP hydrolysis-dependent discharge of precursor from SecA (72, 73). There is a likely possibility that SecA itself penetrates into the membrane along with the preprotein (60-62, 74), although in the model in vitro system ATP (presumably its hydrolysis) weakens the membrane binding of SecA (60, 62). Biochemical results suggest that SecA binding sites on the membrane include SecY or its vicinity (39). Genetic evidence, such as specific suppression, remains to be provided in support of the SecA interaction with SecY (75). SecA (of B. subtilis) with a mutation in the ATP-binding site sequence is inactive or somewhat inhibitory and binds strongly to the membrane (76, 77). The SecA52 mutant protein also partitions mostly to the membrane (60). The ATPase mutant is defective in release of bound preprotein and hence in the proton motive force-dependent transiocation (77). Normally, SecA should undergo dynamic transitions among different conformations, localizations, and degrees of membrane penetration. ATP-induced conformational changes should be central to this problem. Mutational "freezing" of SecA into one state is destructive. The existence of different isoforms of SecA provides the cell with a means to monitor its export states and regulate the synthesis of SecA itself. Thus, the cytoplasmic form of SecA can bind to the secA mRNA and presumably shut down its own synthesis (78).

V. MEMBRANE-EMBEDDED COMPONENTS OF THE

TRANSLOCATOR A. SecY

SecY (PrlA) is defined by the prlA suppressor mutations and export-defective, conditionally lethal secY mutations. It encodes the first identified integral membrane translocation factor among prokaryotes and eukaryotes (79, 80). SecY homologs are now identified in gram-positive bacteria (81), plastid (82), chloroplasts (83), yeast (the SEC61 gene product; 84), and mammalian cells (85). E. coil SecY contains 10 transmembrane segments, 6 cytoplasmic domains, and 5 perip-

Figure 1. Mutations in the secY gene. prl mutations suppress signal sequence mutations, while other mutations (except for s e c Y l 2 l and secY161) retard export; most of the latter are Cs, but secY24 and secYlOO are Ts. secY121 and secY161 are mutations that cannot support export of staphylokinase ( 103). Note that prlA4 and secYlOO cause more than one amino acid alteration.

Protein Translocation Genetics

43

lasmic regions (Figure 1; [86]). The estimated cellular concentration of SecY (and also of SecE) is about 300--500 molecules per cell (23). Mutations that impair SecY function include an amber mutation in rplO that lowers the SecY amount (8), and Ts and Cs missense mutations (9, 10, 14, 75, 87). Recently, we isolated 10 new export-defective secY alleles (Figure 1; [87]). Most of them are Cs and the one Ts mutant contained the same mutation in the cytoplasmic domain 4 (C4; see Figure 1) as the classical secY24 mutant (9). Many Cs mutations are in the cytoplasmic domains, C3, C5, and C6, but some can be within transmembrane segments, TM2 and TM9, as well as in the periplasmic domain 1 (P l). The CS-TM9 region is the preferred site of Cs export-defective mutations (Figure 1). The occurrence of export-defective mutations all over the SecY sequence contrasts to the secE case where defective mutations affect only the expression levels (16; see below). Because SecY probably functions in a complex with other components, its "dominant-negative" variants could be obtained. For instance, "active site" mutations can be dominant-negative if they do not abolish the SecY's ability to complex with other component(s); the mutant protein will sequester the interacting components in an inactive complex and hence will compete with wild-type SecY for the formation of the functional translocation complex. An internal deletion (AArg372Leu-Thr) of SecY (Figure 1; secY -a 1) is dominant negative when expressed from a multicopy plasmid (88). Of the secY Cs mutations, three mutations in C5 and one in C6 also dominantly interfere with export when cloned onto multicopy plasmids (87). Evidence suggests that these dominant effects are mainly due to sequestration of SecE (88, 89; see below). These results with the Cs and the dominant mutations suggest that the carboxy-terminal (C5, TM9 and C6) region will be important for the "catalytic activity" of SecY. As discussed in a later section, the central region of SecY is important for its interaction with other factors, such as SecE. B. SecE

The gene secE (priG) is defined by mutations that lower the expression level of the gene and retard protein export and growth at low temperature (15, 16), as well as by missense prig mutations (21). secE is essential for the viability of the E. coli cell (16). Reconstitution studies indicate that SecE and SecY are the principal membrane-embedded components for the in vitro translocation activity (90, 91). SecE contains three transmembrane segments, but the most C-terminal one (TM3) plus some surrounding sequences can function in vivo (16) and in vitro (92). Recently, a B. subtilis gene encoding a 59-amino acid peptide was shown to complement the E. coli secE mutant (93). The general feature of this peptide resembles that of TM3 and the preceding cytoplasmic region of SecE; the SecE counterparts in B. subtilis and B. licheniformis naturally contain only the equivalent of the C-terminal half of SecE (93).

44

KOREAKI ITO C. SecD and SecF

Also defined by Cs export.defective mutations (94), secD and secF constitute an operon and encode proteins containing both a large periplasmic domain and six transmembrane segments (95). No prl mutations are known for this pair of genes. The large periplasmic domains of these gene products are speculated to play a role in the late stage of translocation (95). Although purified SecD and SecF do not stimulate reconstitution of translocation activity as assayed by sequestration of precursor molecules into the proteoliposomes (23), experiments using semi-intact cells showed that antibodies against SecD inhibited precursor protein release to the aqueous milieu when added outside of the intact spheroplasts (96). Thus, complete translocation may require SecD (and SecF?). The role of SecD-F may be to facilitate either protein release from the membrane or post-translocational folding.

VI. RECOGNITION OF SIGNAL PEPTIDE AND EARLY PART OF MATURE SEQUENCE BY THE TRANSLOCATOR A. The prlA Mutations in SecY The existence of priG and prlA mutations suggests that the two integral membrane proteins, SecE and SecY, interact somehow with the signal peptide. Recent evidence indicates that the signal peptide and the following 30 or so residues of the mature part comprise a special domain called export initiation domain (97). The mature part of this domain was defined as the segment in which artificial placement of positively charged residues is not tolerated for efficient translocation. The prl mutations are not specific to the signal peptide since the translocation-incompatible existence of basic amino acids in the early mature part can be overcome by some prl mutations (98). Interestingly, one such mutation, prlA666, alters a residue of the P1 domain of SecY (Fig. 1). The same residue can be mutated to another prlA allele, prlA3. Other prlA mutations are located either in TM2, TM7, or TM 10 (Figure 1, 99, 100). The lack of clear allele specificity is sometimes interpreted to mean that the prl suppression does not reflect a change in specific protein-protein interaction (101). However, it must be remembered that the protein-protein interaction here is special in that it is an interaction between a protein factor and its "substrate." Such an interaction should be very dynamic. Signal sequences themselves are only conserved in their general features, such as hydrophobicity and charge balance flanking the hydrophobic core, but not in their specific sequences. Even an artificially designed sequence, MKQSTLIoA6, can function perfectly well as a signal sequence in vivo (102). If a common apparatus recognizes signal sequences of different exported proteins, the recognition cannot be sequence specific to the extent that the classical concept of allele specificity would predict. Flexible protein-protein inter-

Protein Translocation Genetics

45

actions could be a key feature in the action of protein translocation factors, which should deal with polypeptide segments of diverse sequences that are even moving. In reality, there is some specificity for the prlA suppression. Some prlA mutations preferentially suppress mutations of the hydrophobic core while others can suppress those reversing the charge balance (98). Of the Met to Arg changes at the 18th residue and the 19th residue of the MalE signal sequence, the prlAlO01 mutation at TM2 prefers the former, while many other prlA mutations prefer the latter (I 00). Some of the prlA mutations as well as other mutations in TM7 render the translocation system refractory to some heterologous proteins, such as staphylokinase (99, 103) and streptokinase (104). Overcoming mutations can then occur in the signal peptide region of these proteins (104, 105). Taken together, these observations suggest that P l, TM2, TM7, and TM 10 of SecY somehow recognize the signal peptide and the region immediately following the signal. Membrane insertion of this region should be a critical early phase event in translocation, in which the precursor polypeptide will assume a loop-like configuration (Figure 2). Such a configuration of the translocation initiation domain may explain the periplasmic localization of some prlA mutations. The Pl domain could be involved in acceptance/rejection of the early mature part (89; Figure 2). The existence of two cold-sensitive mutations (secY125 and secY124) and three prlA mutations (prlA3,prlA666,andprlAI OO1)in the P I-TM2 segment (Figure 1) points to the importance of this segment (100). It may be speculated that the P I-TM2, TM7, and TM l0 regions together recognize the early part of the precursor allowing its insertion. The TM3 region of SecE could also participate in this hypothetical channel (Figure 2). in addition, the CS-TM9 region of SecY, the importance of which has been suggested by the occurrence of a number of Cs and dominant-negative mutations, may additionally participate and confer mid- to late-phase translocation activities on the channel (Figure 2). The notion that theprlA mutations alter the SecY recognition of signal sequences has been questioned by the fact that some of the prlA mutations allow slow export of periplasmic proteins (PhoA and MalE) whose signal sequence had been completely eliminated. (106) However, these signalless proteins behaved anomalously in cell fractionation (106). The mature portion of these proteins appears to contain some elements that target them to the periphery of the prlA or the wild-type cells (106). It is possible that SecY recognizes such an element (in the mature portion of the translocation initiation domain?), and the prlA mutation enhances the acceptance of the signal-less precursor molecules for further translocation.

B. priG and Suppressor-Directed Inactivation Silhavy and his colleagues used a LamB-LacZ fusion protein with a mutation in the signal peptide of the LamB part to selectively inactivate the prl mutated component of translocation machinery (SDI or suppressor-directed inactivation; 107). They reported that the LamB-LacZ fusion protein with the LamB signal

KOREAKI ITO

46 K

i P1 ~~1"1 .:i M2-~ ~'M7 TM9

NH2 o4 I *

cs

Figure 2. A hypothetical pathway of preprotein entry and transit. On the basis of the existence and locations of the prl (stars) and export-defective (or interfering) sec (asterisks) mutations, transmembrane segments of SecE and SecY are supposed to provide a pathway for both an early phase of translocation in which the amino terminal part of the preprotein inserts, and a late phase in which the bulk of the mature part moves across the membrane. Not all the sec mutations (see Figure 1) are shown. Also, only the prlGl mutation is shown for SecE (16). The hatched region of the bold line represents the signal sequence part of the preprotein and the filled part represents the mature part. See the text for more detailed discussions. sequence mutation assumes a different location in the cell, depending upon whether it is present in a strain carrying theprlAd ortheprlG1 mutation. The signal sequence has been processed for the prlA-suppressed fusion protein, but not for the priG-suppressed one. Thus, it was proposed that the former (termed translocation complex) is in the transmembrane configuration, but the latter (termed pretranslocation complex) is only associated with the cytoplasmic surface of the membrane. This provided a basis for the proposal that SecE acts prior to SecY (or more exactly SecY-SecE complex; see Section VII.B). However, the location of theprlGl mutation close to the periplasmic side of TM3 of SecE (16) poses an apparent difficulty. As already pointed out, we consider the possibility that TM3 of SecE, along with some of the SecY TM segments (TM2, TM7 and TM I0), forms the entry pathway (Figure 2). Because the lack of signal sequence cleavage does not necessarily exclude a transmembrane (but leader

Protein Translocation Genetics

47

peptidase-inaccessible) configuration of the pr/G-suppressed fusion protein, establishment of its disposition by independent methods will be crucially important. Curiously, the fusion protein with wild-type signal peptide is subject to rapid cleavage of the signal (107). If the pretransloeation complex and the translocation complex indeed represent sequential events in translocation, why does not the "wild-type" fusion protein jam at the former step? Also, why does aprlA suppressor work without having a prig mutation (see 108 and 89, for further discussion)? The priG1 mutation actually exacerbates the processing of the LamB-LacZ fusion protein with mutated signal peptide (107). These peculiar properties of the prlG l mutation could have "mutation-specifically" contributed to the SDI properties. A more fundamental complication in the SDI analysis is that it relies on the prl-specific signal recognition on one hand and jamming by LacZ on the other. Simply stated, these are separate events only linked with some translocation process itself occurring between them. One cannot assume a priori that LacZ jams the very component that recognized the signal. Pogliano and Beckwith (18) proposed that the initial insertion of the signal peptide is intrinsically cold-sensitive and it subsequently interacts with the integral membrane components. The proposal that two kinds of molecules accumulate in different signal sequence mutants is interesting in their relationship to the pretranslocation and translocation complexes.

VII. SEC FACTOR INTERACTION AND INTRAMEMBRANE TRANSLOCATION PATHWAY A. Intramembrane Steps of Translocation In a model in vitro experiment, an engineered OmpA-dihydrofolate reductase (DHFR) fusion protein was forced to be stuck in the membrane because of the folded DHFR domain. This transmembrane intermediate could be photo-crosslinked with SecA and SecY, but not with phospholipids, via a modified cysteine residue introduced close to the DHFR moiety (67). This result is consistent with the notion that translocating polypeptides move through the proteinaceous translocator, but not directly through the hydrocarbon phase of the membrane, from the early through the late phases. If mutations in the translocation components cause conditional arrest in the intramembrane translocation reactions, similar transmembrane intermediates could be detected in vivo. Indeed, late phase intermediates of MalE can be detected in some secY and secA mutant cells (109) as well as under conditions of dissipated proton motive force across the membrane (110). They ("the processed immature forms") have already been processed for the signal peptide cleavage but remain associated with the membrane and unfolded. A similar intermediate of B la has been detected at low temperature (111).

48

KOREAKI ITO

Pogliano and Beckwith (18) proposed that a number of Cs sec mutants are non-specifically defective in the translocation step and accumulate metabolically unstable transmembrane intermediates. However, taken together with the results of our screening of a number of Cs secF mutants (87), it should be stated that no mutation has been fully characterized that accumulates a transmembrane intermediate as a major product. The small number (about 500 per cell) of membraneembedded translocators (23) will limit the upper number of translocation intermediates that remain in direct contact with the translocator within the translocation pathway. Thus, one of the most interesting classes of mutations that primarily affects an intramembrane process might be difficult to distinguish from mutations affecting other processes; a translocation defect will rapidly back up the system and lead to the secondary accumulation of earlier step intermediates.

B. Bieker-SilhavyModel of Translocation Complexes Bieker and Silhavy (107) proposed that the pretranslocation complex and the translocation complex, as defined by the SDI analysis, have different compositions. The former contains SecA in addition to SecE (PriG), whereas the latter contains SecA, SecE, SecD, and SecF, in addition to SecY (PrlA). Thus, two different molecular assemblies sequentially participate in translocation, and SecE cycles between association with and dissociation from the other integral membrane components. This model is based on "Sec titration" experiments (107, 112). When the LamB-LacZ fusion protein with the signal sequence mutation is induced in the priG or prlA mutant cells, it inhibits general protein export and cell growth. Growth of partially diploid cells containing both the prl and the prF alleles is not affected. This result sounds reasonable, because the wild type (pr/+) gene product that cannot recognize the signal-mutated fusion protein should not be interfered with directly by the fusion protein. However, it is also inferred from the diploid analysis results that SDI at neither Prig (SecE) nor PrlA (SecY) functionally titrates out other Sex gene products that participate in the system. In other words, it seems either that the interaction is only transient (or nonexistent), or that the interacting components exist in excess. According to Bieker and Silhavy, the "Sec titration" experiments discriminate between the above possibilities. Thus, they combined a series of export-defective sec mutations with the prlGl/prlG + or the prlA4/prlA + partial diploid construction. Growth inhibition by the induced fusion protein was examined under "semi-permissive" conditions with lowered functional levels of the mutated Sec protein. The prIGl/prlG +diploid cell was not sensitized to the fusion protein by either secY, secD, or SecF mutations, indicating that SexY, SecD, and SecF are not stably contained in the "pretranslocation complex." In contrast, the prlA4/prlA § diploid cell became fusion protein-sensitive when it received the secE, secD, or secF mutation. Thus, the "translocation complex" does contain SecE, SecD, and SecF, but SecE, SeeD, and SecF might normally be present

Protein Translocation Genetics

49

in excess over SecY (112). Interactions between SecY and SecD, as well as between SecY and SecF, were also supported by our observations on the multicopy suppressing or exacerbating effects of SecD and SecF on the phenotypes of specific secY mutations (87). The revised experiments of Bicker and Silhavy showed that the secA mutant affected both of the SDI systems (112). Thus, they predicted that precursors first bind to the membrane via the SecA-SecE complex forming the pretranslocation complex to which SecY, SeeD, and SecF subsequently associate to form the translocation complex. Upon completion of translocation, the translocation complex is disassembled. Although the Bieker-Silhavy model provides a very important guide to studies of the intracellular pathway of protein translocation, it should not be accepted without critical examination. In addition to the points already raised about the use of the LacZ fusion protein as the secluding molecules and the disposition of the priG-suppressed fusion protein, the following questions about the See titration experiments are addressed. (i) Since the translocation capacity of the wild-type cell is not conclusively estimated, the apparent recessiveness of SDI does not necessarily indicate the excess presence of the interacting components over the one affected by SDI. (ii) Although the Sec titration experiment concerns the number, but not quality, of the Sec factor, only the secE501 mutation was defined as reducing the quantity of SecE (16). Can the results of other undefined sec mutations be compared meaningfully with each other or with secESOl? (iii) As far as SecY and SecE are concerned, mutational reduction of the SecE level destabilizes SecY and leads to a reduced accumulation of the latter (see below), whereas, SecE is stable by itself. Such asymmetric mutational effects can complicate the interpretation; the secE501 effects on the phenotypes of the SDI at PrlA (SecY) should have been affected by the instability of SecY.

C. Insights into the SecY-SecE Interaction When solubilized membrane components of the wild-type cells were fractionated and assayed for reconstitution of proteoliposomes with SecA-dependent translocation activity, a complex containing SecY, SecE, and Band 1 was obtained (90). Although SecY and SecE can be isolated separately from the overproducing cells and can then be reconstituted for active translocation (91), they are in close interaction in vivo. Although prlA and prig mutations generally do not affect cell growth, one particular combination, prlA4 and prlGl, gives a synthetically lethal phenotype, providing genetic evidence for the interaction between SecY and SecE (107). SecY can only be overproduced by overexpressing both the secY and the secE genes (113). This is due to rapid (half life about 2 min) degradation of excess (uncomplexed) SecY (114); simultaneous overproduction of SecE completely

SO

KOREAKI ITO

stabilizes the excessive SecY (113, 114). SecE is the primary determinant of the stability of SecY in the wild-type cell, because the secE501 mutational reduction (by about 50%) of the secE expression level results in destabilization of a fraction (again about 50%) of SecY (114); one of the roles played by SecE should be to stabilize SecY. In the secE501 mutant cells, about 50% of the newly synthesized SecY molecules are immediately committed to rapid degradation, while the remaining 50% remain completely stable (114). The existence of two distinct populations of SecY when availability of the partner SecE molecules is limited implies the following. First, newly synthesized SecY immediately associates with SecE in the wild-type cell. Second, the SecY-SecE complex, once formed, does not dissociate to the extent that dissociated SecY is attacked by the SecY-degradation system. This seems to invalidate a simple version of the Bieker-Silhavy model. The dominant negative secF mutation, secF'dl, provided a useful system for genetic analysis of SecY-SecE interaction (88). Our working hypothesis is that the mutation inactivates an essential SecY activity without abolishing its binding to SecE, and the mutant protein sequesters SecE. In accordance with this hypothesis, SecE overproduction effectively overcame the export interference. The dominant effects of a series of SecY-PhoA fusion proteins focuses the region necessary for the dominant export interference to the central part of SecY (P3 to TM7 interval; 88). To define more precisely the region of SecY important for the interaction with SecE, second site insertion/deletion mutations that suppressed secF-dl but did not knock out the synthesis of SecY itself were isolated using a derivative of the sec Fdl gene to which the lacZa sequence had been attached in frame to the 3' end (115). The intragenic suppressor mutations were found to reside mostly in the C4 domain. The secF24 mutation, a single amino acid substitution in this region, also suppressed the export interference by secF-dl. These suppressor mutations do not restore the translocation activity of the SecF-dl protein itself, but impairs its interaction with SecE. It was found that the SecY24 mutant protein, when it was overexpressed, was no longer stabilized by the overexpression of SecE, although stabilization by Ydr (see below) was unchanged. While SecY-SecE complexes in the wild-type cell survived the immunoprecipitation experiments under appropriate solubilization conditions, only uncomplexed SecY24 protein was found in the anti-SecY immunoprecipitates from secu mutant cells. In addition, the basal level SecY24 protein was destabilized at 42 C, the nonpermissive temperature for this Ts mutant. Thus, the Gly to Asp change in the C4 domain weakens the SecY's interaction with SecE (115). These results establish genetically that the SecY-SecE interaction is crucial for the translocator function and suggest that the C4 domain of SecY is important for this interaction. How the C4 domain facilitates actual binding between SecY and SecE, that could well involve TM-TM interaction (see the preceding sections), remains to be determined.

51

Protein Translocation Genetics

VIII.

A D D I T I O N A L FACTORS POSSIBLY I N V O L V E D IN TRANSLOCATION A. Band 1 and P12

These factors have been identified by biochemical studies. They appear distinct and their in vivo roles remain to be established. Band I is found in the active fraction purified as membrane-embedded components of"translocase" (90, 116). Chromatographic and immunoprecipitation behaviors suggest that it is associated with SecY/E. The association is labile at ambient temperature. The translocase preparation reconstituted with Band 1 is as active as native membrane vesicles (117). PI2 is a membrane protein which confers markedly higher translocation activity when present during reconstitution of proteoliposomes from purified SecY and SecE (118).

B. PspA This protein is induced when abnormal secretory proteins, such as mutant forms of PhoE and MalE-LacZ fusion protein, enter the translocation pathway and are stuck in the membrane, as well as upon infection by filamentous phages (phage shock protein). ApspA mutant cell shows export retardation of several cell surface proteins including Male and B la (119).

C. Skp This protein stimulates protein translocation in vitro when added to membrane vesicles prepared from secA mutant cells. Although it was purified from a ribosomal fraction, its amino terminal sequence puzzlingly proved to be identical with the mature part of an outer membrane protein (120). D. Ydr Ydr has been identified as a multicopy suppressor of the dominant negative

secY-dl mutation (89, 121). It is a 181-residue hydrophilic protein with its C-terminal region potentially forming an amphiphilic helix. Ydr is loosely associated with the cytoplasmic membrane in a saturable fashion, possibly via SecY. Its overproduction can partially stabilize the oversynthesized SecY that is otherwise unstable. While its overproduction improves protein export in the presence of secy-dl on the one hand, it strikingly exacerbates export and viability of the secY24 mutant cell on the other hand. As already discussed, the sec}24 mutation weakens the SecY-SecE interaction. Taken together with the observation that ydr can be disrupted without appreciable phenotypes, one hypothetical role of Ydr is to control

KOREAKI ITO

52

the formation of active translocator by negatively intervening in the SecY-SecE interaction.

E. Orfl2 This gene, located in the upstream of secD, constitutes an operon with secD and secF (95,122). Its overexpression, like that ofydr, suppresses secF'al and stabilizes oversynthesized SecY. Its disruption causes synthetic growth impairment with the sec F39 mutation, but is otherwise silent. The deduced amino acid sequence suggests that the Off12 product is a membrane protein. F. FIsH

In an attempt to identify genes involved in membrane protein anchoring and assembly, we devised a screening system that allowed the isolation of mutations that enhance translocation of a PhoA reporter attached to a cytoplasmic domain of a membrane protein (123). A mutation thus obtained, std101 (std for stop transfer defective), proved to be an allele offisH, that encodes a putative membrane-bound ATPase with significant sequence homology with the members of an eukaryotic ATPase family such as SEC 18p, NSF, and PAS I p (124, 125). A temperaturesensitive ftsH mutation (ftsH1), depletion of FtsH, and expression of either a C-terminally truncated form of FtsH or an ATP-binding sequence mutant of FtsH, lead to the stop transfer defective phenotype for the PhoA fusion protein and translocation-defective phenotypes for several exported proteins (123, 126, 127). Thus, impaired FtsH function leads to translocation enhancement of normally anchored proteins and translocation inhibition of normally translocated proteins. Among the exported proteins tested, Bla was most affected by the ftsH mutations (126), but OmpA export is also significantly retarded by the fish depletion (123) or by dominant mutations offish (127). FtsH could functionally interact with the translocation machinery or facilitate correct assembly of membrane proteins, including SecY, SecE, or other translocator components.

G. Are There Translocation Factors in the Periplasm? Molecular chaperones in the trans side of the membrane, such as the mitochondrial and the endoplasmic reticulum HSF70 homologs, can actively participate in the translocation reaction (128,129). They either drive translocation by binding the incoming polypeptide chain or facilitate assembly or recycling of translocator components. In E. coli, Skp may reside outside the cytoplasn,ic membrane, but its role has not been established. SecD (and SecF?), that contains a large periplasmic domain and possesses a release/folding facilitating role (96), can be regarded as a periplasmic translocation factor. A well-defined folding factor in the periplasm is DsbA (6, 130) that, in concert with the membrane-bound DsbB (7), facilitates the

Protein Translocation Genetics

53

formation of disulfide bonds in a set of envelope proteins after membrane translocation (131). In the absence of DsbA, some proteins such as PhoA remain on the periplasmic surface of the membrane in the reduced and unfolded form (130). It remains to be established whether polypeptide-binding chaperones and/or folding enhancers (including the DsbA system) in the periplasm have roles in the late phase translocation reactions.

IX. CONCLUSION The significance of genetic analysis is at least two-fold. First, it is useful to identify actors that participate in the system. Second, it is a powerful tool in assigning their intracellular roles. Protein translocation genetics in E. coli has been successful in both of these aspects. In particular, elaborate genetic dissection, such as the SDI/Sec titration and the dominant negative analyses, now approaches the central question: How proteinaceous components in the membrane interact with one another in providing the pathway through which movement of substrate polypeptide is facilitated. Although, as discussed in detail, each approach has its own limitations or uncertainties, use of complementary biochemical experiments will be of great assistance in the search for a solution. For instance, mutational defects in the membrane traversing reaction might effectively be studied only in vitro. Recent and remarkable progress in reconstitution and energetics of the in vitro translocation reaction and its substeps (132) indicates that it is now feasible to blend biochemistry and genetics to study molecular events that occur in the intracellular pathway of protein topogenesis.

ACKNOWLEDGMENTS The author would like to thank the membersof his laboratory,especially Yoshinori Akiyama, Tohru Yoshihisa, Tetsuya Taura, Tadashi Baba, Takashi Shimoike, and Satoshi Kishigami for fruitful collaborations, and Kurt Cannon for reading the manuscript. The work from the author's laboratory has been supported by grants from the Ministry of Education, Science and Culture, Japan; the Naito Foundation, Takara Shuzo, Mitsubishi-Kasei Corporation, and Yamada Science Foundation.

NOTE ADDED IN PROOF Band I and P12 are identical and given the name of Sec G. Ydr is now called "Syd."

54

KOREAKI ITO

REFERENCES 1. Beckwith, J., & Ferro-Novick, S. (1986). Genetic studies on protein export in bacteria. Curr. Top. Microbiol. Immunol. 125, 5-27. 2. Rapoport, T. A. (1992). Transport of proteins across the endoplasmic reticulum membrane. Science 258, 931-936. 3. Pugsley, A. P. (1993). The complete general secretory pathway in gram-negative bacteria. Microbiol. Rev. 57, 50-108. 4. Schatz, P. J., & Beckwith, J. (1990). Genetic analysis of protein export in Escherichia coli. Annu. R~: Genet. 24, 215-248. 5. Lee, C., Li, P., Inouye, H., Brickman, E., & Beckwith, J. (1989). Genetic studies on the inability of ~galactosidase to be translocated across the Escherichia coil cytoplasmic membrane. J. Bacterioi. 171, 4609-4616.

6. Bardwell, J. C. A., McGovern, K., & Beckwith, J. (1991). Identification of a protein required for disulfide bond formation in vivo. Cell 65, 581-589. 7. Bardwell, J. C. A., Lee, J.-O., Jander, G., Martin, N., Belin, D., & Beckwith, J. (1993). A pathway for disulfide bond formation in vivo. Proc. Natl. Acad. Sci. USA 90, 1038--1042. 8. Ito, K., Wittekind, M., Nomura, M., Shiba, K., Yura, T., Miura, A., & Nashimoto, H. (1983). A temperature-sensitive mutant of E. coil exhibiting slow processing of exported proteins. Cell 32, 789-797. 9. Shiba, K., Ito, K., Yura, T., & Cerretti, D. P. (1984). A defined mutation in the protein export gene within the spc ribosomal protein operon of Escherichia coil: Isolation and characterization of a new temperature-sensitive secY mutant. EMBO J. 3, 631-635. 10. Ito, K., Hirota, Y., & Akiyama, Y. (1989). Temperature-sensitive sec mutants of Escherichia coil: inhibition of protein export at the permissive temperature. J. Bacteriol. 171, 1742-1743. 11. Date, T., & Wickner, W. (1981). Isolation of Escherichia coil leader peptidase gene and effects of leader peptidase overproduction in vivo. Proc. Natl. Acad. Sci. USA 78, 6106- 6110. 12. Dalbey, R. E., & yon Heijne, G. (1992). Signal peptidases in prokaryotes and eukaryotes- a new protease family. Trends Biochem. ScL 17, 474- 478. 13. Dalbey, R. E., & Wickner, W. (1985). Leader peptidase catalyzes the release of exported proteins from the outer surface of the Escherichia coil plasma membrane. J. Biol. Chem. 260, 1592515931. 14. Riggs, P. D., Derman, A. I., & Beckwith, J. (1988). A mutation affecting the regulation of a secA-lacZ fusion defines a new sec gene. Genetics I 18, 571-579. 15. Schatz, P. J., Riggs, P. D., Jacq, A., Fath, M. J., & Beckwith, J. (1989). The secE gene encodes an integral membrane protein required for protein export in Escherichia coll. Genes Dev. 3,10351044. 16. Schatz, P., Bicker, K. L., Ottemann, K. M., Silhavy, T. J., & Beckwith, J. (1991). One of three transmembrane stretches is sufficient for the functioning of the SecE protein, a membrane component of the E. coli secretion machinery. EMBO J. 10, 1749-1757. 17. Oliver, D. B., & Beckwith, J. (1981). E. coli mutant pleiotropically defective in the export of secreted proteins. Cell 25, 765-772. 18. Pogliano, K. J., & Beckwith, J. (1993). The Cs sec mutants of Escherichia coli reflect the cold sensitivity of protein export itself. Genetics 133, 763--773. 19. Bieker, K. L., Phillips, G. J., & Silhavy, T. J. (1990). The sec andprl genes of Escherichia eoli. J. Bioenerg. Biomembr. 22, 291-310. 20. Fikes, J. D., & Bassford, P. J. (1989). Novel secA alleles improve export of maltose-binding protein synthesized with a defective signal peptide. J. Bacteriol. 171,402-.409. 21. Stader, J., Gansheroff, L. J., & Silhavy, T. J. (1989). New suppressors of signal-sequence mutations, priG, are linked tightly to the secE gene of Escherichia coll. Genes Dev. 3, 10451052.

Protein Translocation Genetics

55

22. Randall, L. L., & Hardy, S. J. S. (1984). Export of protein in bacteria. Microbial. Rev. 48, 290298. 23. Matsuyama, S., Fujita, Y., Sagara, K., & Mizushima, S. (1992). Overproduction, purification and characterization of SecD and SecF, integral membrane components of the protein transiocation machinery of Escherichia coli. Biochim. Biophys. Acta. 1122, 77- 84. 24. Randall, L. L., & Hardy, S. J. S. (I 986). Correlation of competence for export with lack of tertiary structure of the mature species: a study in viva of maltose-binding protein in E. coli. Cell 46, 921928. 25. Arkowitz, R. A., Joly, J. C. and Wickner, W. (1992). Translocation can drive the unfolding of a preprotein domain. EMBO J. 12, 243-253. 26. Kumamoto, C. A. (1991). Molecular chaperones and protein translocation across the Escherichia coil inner membrane. Mol. Microbial. 5, 19-22. 27. Trun, N. J., Stader, J., Lupas, A., Kumamoto, C., & Silhavy, T. J. (1988). Two cellular components, PriA and SecB, that recognize different sequence determinants are required for efficient protein export. J. Bacterial. 170, 5928-5930. 28. Kim, J., Lee, Y., Kim, C., & Park, C. (1992). Involvement of SecB, a chaperone, in the export of ribose-binding protein. J. Bacterial. 174, 5219-5227. 29. Weiss, J. B., Ray, P. H., & Bassford P. J. (1988). Purified SecB protein of Escherichia coli retards folding and promotes membrane transiocation of the maltose-binding protein in vitro. Proc. Natl. Acad. Sci. USA 85, 8978-8982. 30. Hardy, S. J. S., & Randall, L. L. (1991). A kinetic partitioning model of selective binding of normative proteins by the bacterial chaperone SecB. Science 251,439-.443. 31. Collier, D. N., Bankaitis, V. A., Weiss, J. B., & Bassford, P. J. (1988). The antifolding activity of SecB promotes the export of the E. coli maltose-binding protein. Cell 53, 273-283. 32. Kumamoto, C. A., & Francetic O. (1993). Highly selective binding of nascent polypeptides by an Escherichia coli chaperone protein in viva. J. Bacterial. 175, 2184-2188. 33. Randall, L. L. (1992). Peptide binding by chaperone SecB: Implications for recognition of normative structure. Science 257, 241-245. 34. Randall, L. L., & Hardy, S. J. S. (1989). Unity in function in the absence of consensus in sequence: Role of leader peptides in export. Science 243, 1156-1159. 35. Watanabe, M., & Blobel, G. (1989). SecB functions as a cytosolic signal recognition factor for protein export in E. coli. Cell 58, 695-705. 36. Lecker, S. H., Driessen, A. J. M., & Wickner, W. (I 990). ProOmpA contains secondary and tertiary structure and is shielded from aggregation by association with SecB. EMBO J. 9, 2309-2314. 37. Collier, D. N., Strobel, S. M., & Bassford. P. J. (1990). SecB-independent export of Escherichia coli ribose-binding protein (RBP): some comparison with export of maltose-binding protein (MBP) and studies with RBP-MBP hybrid proteins. J. Bacterial. 172, 6875-6884. 38. Gannon, P. M., & Kumamoto, C. A. (1993). Mutations of the molecular chaperone protein SecB which alter the interaction between SecB and maltose-binding protein. J. Biol. Chem. 268, 1590-1595. 39. Hartl, E, Lecker, S., Schiebel, E., Hendrick, J. P., & Wickner, W. (1990). The binding cascade of SecB to SecA to SecY/E mediates preprotein targeting to the E. coli plasma membrane. Cell 63, 269-279. 0. McFarland, L., Francetic, O., & Kumamoto, C. A. (1993). A mutation of Escherichia coli SecA protein that partially compensates for the absence of SecB. J. Bacterial. 175, 2255-2262. 41. Wild, J., Walter, W. A., Gross, C., & Airman, E. (1993). Accumulation of secretory precursors in Escherichia coli induces the heat shock response. J. Bacterial. 175, 3992-3997. 42. Wild, J., Altman, E., Yura, T., & Gross, C. A. (1992). The DnaK and DnaJ heat shock proteins participate in protein export in E. coli. Genes Dev. 6, 1165-1172. 43. All,nan, E., Kumamoto, C. A., & Emr, S. (1991). Heat-shock proteins can substitute for SecB function during protein export in Escherichia coli. EMBO J. 10, 239-245.

56

KOREAKI ITO

44. Kusukawa, N., Yura, T., Ueguchi, C., Akiyama, Y., & Ira, K. (1989). Effects of mutations in heat-shock genes gmES and groEL on protein export in Escherichia coll. EMBO J. 8, 3517-3521. 45. Laminet, A. A., Kumamoto, C. A., & Pliickthun, A. (1991). Folding in vitro and transport in viw of pre-lS-lactarnase are SecB independent. Mol. Microbial. 5, 117-122. 6. Akiyama, Y., Kanemori, M., Shirai, Y., Yura, T., & Ira, K., unpublished results. 47. Bochkareva, E. S., Lissin, N. M., & Girshovich, A. S. (1988). Transient association of newly synthesized unfolded proteins with heat shock GroEL protein. Nature 336, 254-257. 48. Poritz, M. A., Strub, K., & Walter, P. (1988). Human SRP RNA and E. coil 4.5S RNA contain a highly homologous structural domain. Cell 55, 4--6. 9. R6misch, K., Webb, J., Herz, J., Prehn, S., Frank, R., Vingron, M., & Dobberstein, B. (1989). Homology of 54K protein of signal-recognition particle, docking protein and two E. coil proteins with putative GTP-binding domains. Nature 340, 478-482. 50. Gill, D. R., Hatfull, G. F., & Salmond, G. P. C. (1986). A new cell division operon in Escherichia coll. Mol. Gen. Genet. 205, 134-145. 51. Brown, S. (1991). 4. 5S RNA, does form predict function? New Biol. 3, 430--438. 52. Luirink, J., High, S., Wood, H., Giner, A., Tollervey, D., & Dobberstein, B. (1992). Signal-sequence recognition by an Escherichia coil ribonucleoprotein complex. Nature 359, 741-743. 53. Poritz, M. A., Bernstein, H. D,, Strub, K., Zopf, D., Wilhelm, H., & Walter, P. (1990). An E. coil ribonucleoprotein containing 4. 5S RNA resembles mammalian signal recognition particle. Science 250, 1111-1117. 54. Ribes, V., Romisch, K., Giner, A., Dobberstein, B., & Tollervey, D. (1990). E. coil 4. 5S RNA is part of a ribonucleoprotein particles that has properties related to signal recognition particle. Cell 63,591-600. 55. Phillips, G. J., & Silhavy, T. J. (1992). The E. coilfib gene is necessary for viability and efficient protein exlxa'L Nature 359, 744-746. 56. Bernstein, H. D., Zopf, D., Freymann, D. M., & Walter, P. (1993). Functional substitution of the signal recognition particle 54-kDa subunit by its Escherichia coil homolog. Proc. Natl. Acad. Sci. USA 90, 5221-5233. 57. Josefsson, L. G., & Randall, L. L. ( 1981). Different exported proteins in E. coil show differences in the temporal mode of processing in viva. Cell 25, 151-157. 58. Harm, B. C., & Walter, P. ( 1991). The signal recognition particle of S. cerevisiae. Cell 67,131-144. 59. Oliver, D. B. (1993). SecA protein: autoregulated ATPase catalysing preprotein insertion and translocation across the Escherichia coil inner membrane. Mol. Microbial 7, 159--165. 0. Cabeili, R. J., Dolan, K. M., Qian, L., & Oliver, D. B. (1991). Characterization of membrane-associated and soluble states of SecA protein from wild-type and secASl (TS) mutant strains of Escherichia coll. I. Biol. Chem. 266, 24420-24427. 61. Ulbrandt, N. D., London, E., & Oliver, D. B. (1992). Deep penetration of a portion of Escherichia coil SecA protein into model membranes is promoted by anionic phospholipids and by partial unfolding. J. Biol. Chem. 267, 15184-15192. 21 Breukink, E., Demel, R. A., de Korte-Kool, G., & de Kruijff, B. (1992). SecA insertion into phospholipids is stimulated by negatively charged lipids and inhibited by ATP: a monolayer study. Biochemistry 31, 1119-1124. 63. Cabelli, R. J., Chen, L., Tai, E-C., & Oliver, B. (1988). SecA protein is required for secretory protein translocation into E. coli membrane vesicles. Cell 55, 683-692. 4. Lill, R., Dowhan, W., & Wickner, W. (I 990). The ATPase activity of SecA is regulated by acidic phospholipids, SecY, and the leader and mature domains of precmsor proteins. Cell 60, 271-280. 65. Oliver, D. B., Cabelli, R. J., Dolan, K. M., & Jarosik, G. P. (1990). Azide-resistant mutants of Escherichia coil alter the SecA protein, an azide-sensitive component of the protein export machinery. Proc. Natl. Acad. Sci. USA 87, 8227-8231.

Protein Translocation Genetics

57

66. Akita, M., Sasaki, S., Matsuyama, S., & Mizushima, S. (1990). SecA interacts with secretory proteins by recognizing the positive charge at the amino terminus of the signal peptide in Escherichia coil J. Biol. Chem. 265, 8164-8169. 67. Joly, J. C., & Wickner, W. (1993). The SecA and SecY subunits of translocase are the nearest neighbors of a translocating preprotein, shielding it from phospholipids. EMBO J. 12, 255-263. 68. Akita, M., Shinkai, A., Matsuyama, S., & Mizushima, S. (1991). SecA, an essential component of the secretory machinery of Escherichia coli, exists as homodimer. Biochem. Biophys. Res. Commun. 174, 211-216. 69. Kimura E., Akita, M., Matsuyama, S., & Mizushima, S. (1991). Determination of a region in SecA that interacts with presecretory proteins in Escherichia coil J. Biol. Chem. 266, 6600-6606. 70. Matsuyama, S., Kimura, E., & Mizushima, S. (1990). Complementation of two overlapping fragments of SecA, a protein translocation ATPase of Escherichia coil, allows ATP binding to its amino-terminal region. J. Biol. Chem. 265, 8760-8765. 71. Shinkai, A., Mei, L. H., Tokuda, H., & Mizushima, S. (1991). The conformation of SecA, as revealed by its protease sensitivity, is altered upon interaction with ATP, Dresecretory proteins, everted membrane vesicles, and phospholipids. J. Biol. Chem. 266, 5827-5833. 72. Schiebel, E., Driessen, A. J. M., Hartl, E-U., & Wickner, W. (1991). AI,tH§ and ATP function at different steps of the catalytic cycle of preprotein translocase. Cell 64, 927-939. 73. Driessen, A. J. M. (1992). Precursor protein translocation by the Escherichia coil translocase is directed by the protonmotive force. EMBO J. 11,847-853. 74. Wickner, W., personal communication. 75. Baba, T., Jacq, A., Brickman, E., Beckwith, J., Taura, T., Ueguchi, C., Akiyama, Y., & Ira, K. (1990). Characterization of cold-sensitive secY mutants of Escherichia coll. J. Bacterial. 172, 7005-7010. 76. Klose, M., Schimz, K., van der Walk, J., Driessen, A. J. M., & Freudl, R. (1993). Lysine 106of the putative catalytic ATP-binding site of the Bacillus subtilis SecA protein is required for functional complementation of Escherichia coil secA mutants in viva. J. Biol. Chem. 268, 4504--4510. 77. van der Walk, J., Klose, M., Breukink, E., Demel, R. A., de Kruijff, B., Freudl, R., & Driessen, A. J. M. (1993). Characterization of a Bacillus subtilis SecA mutant protein deficient in translocation ATPase and release from the membrane. Mol. Microbial. 8, 31-42. 78. Dolan, K. M., & Oliver, D. B. (1991). Characterization of Escherichia coli SecA protein binding to a site on its mRNA involved in autoregulation. J. Biol. Chem. 266, 23329-23333. 79. Ira, K. (1984). Identification of the secY (prlA) gene product involved in protein export in Escherichia coll. Moi. Gen. Genet. 197, 204-208. 80. Akiyama, Y., & Ito, K. (1985). The SecY membrane component of the bacterial protein export machinery, Analysis by new electrophoretic methods for integral membrane proteins. EMBO J. 4, 3351-3356. 81. Tschauder, S., Driessen, A. J. M., & Freudl, R. (1992). Cloning and molecular characterization of the sec Y genes from Bacillus Iicheniformis and Staphylococcal carnosus, comparative analysis of nine members of the SecY family. Mol. Gen. Genet. 235, 147-152. 82. Douglas, S. E. (1992). A secY monologue is found in the plastid genome of Co'ptomonas 4). FEBS Lett. 298, 93-96. 83. Scaramuzzi, C. D., Stokes, H. W., & Hiller, R. G. (1992). Characterisation of a chloroplast-encoded secY homologue and atpH from a chromophytic alga. Evidence for a novel chloroplast genome organisation. FEBS Lett. 304, 119-123. 84. Stifling, C. J., Rothblatt, J., Hosobuchi, M., Deshaies, R., & Schekman, R. (1992). Protein translocation mutants defective in the insertion of integral membrane proteins into the endoplasmic reticulum. Moi. Biol. Cell 3, 129-142.

58

KOREAKI ITO

85. Gorlich, D., Prehn, S., Harunann, E., Kalies, K.-U., & Rapoport, T. A. (1992). A mammalian

homolog of SEC61p and SecYp is associated with ribosomes and nascent polypeptides during translocation. Cell 71,489-503. 86. Akiyama, Y., & lto, K. (1987). Topology analysis of the SecY protein, an integral membrane protein involved in protein export in Escherichia coli. EMBO J. 6, 3465--3470. 87. Taura, T., Akiyama, Y., & Ira, K. (1994). Genetic analysis of SecY." additional export-defective mutations and factors affecting their phenotypes. Mol. Gen. Genet. 243, 261-269. 88. Shimoike, T., Akiyama, Y., Baba, T., Taura, T., & Ira, K. (1992). SecY variants that interfere with Escherichia coli protein export in the presence of normal secY. Mol. Microbial. 6, 1205-1210. 89. Ira, K. (1992). SecY and integral membrane components of the Escherichia coil protein translocation system. Mol. Microbial. 6, 2423-2428. 0. Brundage, L., Hendrick, J. P., Schiebel, E., Driessen, A. J. M., & Wickner, W. (1990). The purified E. coil integral membrane protein SecY/SecE is sufficient for reconstitution of SecA-dependent precursor protein translocation. Cell 62, 649--657. 91. Akimaru, J., Matsuyama, S., Tokuda, H., & Mizushima, S. (1991). Reconstitution of a protein translocation system containing purified SecY, SecE, and SecA from Escherichia coll. Prec. Natl. Acad. Sci. USA 88, 6545--6549. 92. Nishiyama, K., Mizushima, S., & Tokuda, H. (1992). The carboxyl-terminal region of SecE interacts with SecY and is functional in the reconstitution of protein translocation activity in Escherichia coli. J. Biol. Chem. 267, 7170-7176. 93. Jeong, S. M., Yoshikawa, H., & Takahashi, H. (1993). Isolation and characterization of the secE homologue of Bacillus subtilis. Mol. Microbial., in press. 94. Gardel, C., Benson, S., Hunt, J., Michaelis, S., & Beckwith, J. (1987). secD, a new gene involved in protein export in Escherichia coll. J. Bacterial. 169, 1286-1290. 95. Gardel, C., Johnson, K., Jacq, A., & Beckwith, J. (1990). The secD locus of E. coli codes for two membrane proteins required for protein export. EMBO J. 9, 3209-3216. 96. Matsuyama, S., Fujita, Y., & Mizushima, S. (1993). SecD is involved in the release oftranslocated secretory proteins from the cytoplasmic membrane of Escl~richia coli. EMBO J. 12, 265-270. 97. Andersson, H., & van Heijne, G. (1991). A 30-residue-long "export initiation domain" adjacent to the signal sequence is critical for protein translocation across the inner membrane of Escherichia coli. Proc. Natl. Acad. Sci. USA 88, 9751-9754. 98. Puziss, J. W., Strobel, S. M., & Bassford, P. J. (1992). Export of maltose-binding protein species with altered charge distribution surrounding the signal peptide hydrophobic core in Escherichia coli cells harboring pd suppressor mutations. J. Bacterial. 174, 92-101. 99. Sako, T., & lino, T. (1988). Distinct mutation sites in prlA suppressor mutant strains of Escherichia coli respond either to suppression of signal peptide mutations or to blockage of staphylokinase processing. J. Bacterial. 170, 5389-5391. 100. Francetic, O., Hanson, M. P., & Kumamoto C. A. (1993). prlA suppression of defective export of maltose-binding protein in secB mutants of Escherichia coll, J. Bacterial. 175, 4036--4044. 101. Randall, L. L., Hardy, S. J. S., & Thorn, J. R. (1987). Export of protein, a biochemical view.Annu. Rev. Microbial. 41,507-54 I. 102. Laforet, G. A., & Kendall, D. A. (1991). Functional limits of conformation, hydrophobicity, and steric constraints in prokaryotic signal peptide cleavage regions. J. Biol. Chem. 266, 1326-1334. 103. Sako, T. (1991). Novel prM alleles defective in supporting staphylokinase processing in Escherichia coll. J. Bacterial. 173, 2289-2296. 104. Mtiller, J., Reinert, H., & Malke, H. (1989). Streptokinase mutations relieving Escherichia coil K- 12 (prlA4) of detriments caused by the wild-type skc gene. J. Bacterial. 171, 2202-2208. 105. lino, T., & Sako, T. (1988). Inhibition and resumption of processing of the staphylokinase in some Escherichia coil prlA suppressor mutants. J. Biol. Chem. 263, 19077-19082. 106. Derman, A. I., Puziss, J. W., Bassford, P. J., & Beckwith, J. (1993). A signal sequence is not required for protein export in prlA mutants of Escherichia coll. EMBO J. 12, 879-888.

Protein Translocation Genetics

59

107. Bieker, K. L., & Silhavy, T. J. (1990). PrlA (SecY) and PriG (SecE) interact directly and function sequentially during protein translocation in E. coll. Cell 61,833-842. 108. Bieker, K. L., & Silhavy. T. J. (1990). The genetics of protein secretion in E. coli. Trends Genet. 6, 329-334. 109. Ueguchi, C., & Ira, K. (1990). Escherichia coli sec mutants accumulate a processed immature form of maltose-binding protein (MBP), a late-phase intermediate in MBP export. J. Bacterial. 172, 5643-5649. 110. Geiler, B. L. (1990). Electrochemical potential releases a membrane-bound secretion intermediate of maltose-binding protein in Escherichia coll. J. Bacterial. 172, 4870-4876. 111. Minsky, A., Summers, R. G., & Knowles, J. R. (1986). Secretion of ~lactamase into the periplasm of Escherichia coil: Evidence for a distinct release step associated with a conformational change. Proc. Natl. Acad. Sci. USA 83, 4180-4184. 112. Bieker-Brady, K., & Silhavy, T. J. (1992). Suppressor analysis suggests a multistep, cyclic mechanism for protein secretion in Escherichia coli. EMBO J. 11, 3165-3174. 113. Matsuyama, S., Akimaru, J., & Mizushima, S. (1990). SecE-depcndent overproduction of SecY in Escherichia coll. FEBS Len. 269, 96-100. 114. Taura, T., Baba, T., Akiyama, Y., & Ito, K. (1993). Determinants of the quantity of the stable SecY complex in the Escherichia coli cell. J. Bacterial. 175, 7771-7775. I15. Baba, T., Taura, T, Shimoike, T., Akiyarna, Y., Yoshihisa, T., & Ira, K. (1994). A cytoplasmic domain is important for the formation of a SecY-SecE translocator complex. Proc. Natl. Acad. Sci. USA 91, 4539-4543. 116. Brundage, L., Fimmei, C. J., Mizushima, S., & Wickner, W. (1992). SecY, SecE, and Band I form the membrane-embedded domain of Escherichia coil preprotein transiocase. J. Biol. Chem. 267, 4166-4170. 117. Bassilana, M., & Wickner, W. (1993). Purified Escherichia coil preprotein translocase catalyzes multiple cycles of precursor protein translocation. Biochemistry 32, 2626-2630. 118. Nishiyama, K., Mizushima, S., & Tokuda, H. (1993). A novel membrane protein involved in protein translocation across the cytoplasmic membrane of Escherichia coli. EMBO J. 12, 3409-3415. 119. Kleerebezem, M., & Tomassen, J. (1993). Expression of the pspA gene stimulates efficient protein export in Escherichia coil. Mol. Microbial. 7, 947-956. 120. Thome, B. M., Hoffschulte, H. K., Schiltz, E., & Mtiller, M. (1990). A protein with sequence identity to Skp (FirA) supports protein translocation into plasma membrane vesicles of Escherichia coli. FEBS Lett. 269, 113-116. 121. Shimoike, T., Taura, T., Kihara, A., Yoshihisa, T., Akiyama, Y., Cannon, K., & Ira, K. (1995). Product of a new gene, s)~/, functionally interacts with SecY when overproduced in Escherichia coll. J. Biol. Chem., in press. 122. Reuter, K., Slany, R., Ullrich, E, & Kersten, H. (1991). Structure and organization of Escherichia coil genes involved in biosynthesis of the deazaguanine derivative queuine, a nutrient factor for eukaryotes. J. Bacterial. 173, 2256-2264. 123. Akiyama, Y., Ogura, T., & 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. 124. Tomoyasu, T., Yuki, T., Morimura, S., Mori, H., Yamanaka, K., Niki, H., Hiraga, S. and Ogura, T. (1993). The Escherichia coli FtsH protein is a prokaryotic member of a protein family of putative ATPase involved in membrane function, cell cycle control, and gene expression. J. Bacterial. 175, 1344-1351. 125. Tomoyasu, T., Yamanaka, K., Murata, K., Suzaki, T., Bouloc, P., Kato, A., Niki, H., Hiraga, S., & Ogura, T. (1993). Topology and subcellular localization of FtsH protein in Escherichia coli. J. Bacterial. 175, 1352-1357.

60

KOREAKI ITO

126. Tomoyasu, T., Yamanaka, K., Niki, H., Hiraga, S., & Ogura, T. Effects of a thermosensitiveftsH mutation on localization and assembly of cell envelope proteins in Escherichia coll. Personal communication. 127. Akiyama, Y., Shirai, Y., & lto, K. (1994). Involvement of FIsH in protein assembly into and through the membfme. II. Do~nant mutations affecting FtsH functions. J. Biol. Chem. 269, 5225-5229. 128. Kang, P. J., Ostenmn, J., Shilfing, J., Neupert, W., Craig, E. A., & Pfanner, N. (1990). Requirement for hsp70 in the mitochondrial matrix for translocation and folding of precursor proteins. Nature 348, 137-143. 129. Sanders, S. L., Whitfield, K. M., Voge, J. P., Rose, M., & Schelunan, R. W. (1992). Sec61p and BiP directly facilitate polypeptide Iranslocation into the ER. Cell 69, 353-365. 130. Kamitani, S., Akiyama, Y., & Ito, K. (1992). Identification and characterization of an Escherichia coli gene required for the fommtion of correctly folded alkaline phosphatase, a periplasmic enzyme. EMBO J. i 1, 57-62. 131. Akiyama, Y., Kamitani, S., Kusukawa, N., & Ito, K. (1992). In vitro catalysis of oxidative folding of disulfide-bonded proteins by Escherichia coli dsbA (ppfA) gene product. J. Biol. Chem. 267, 22440-22445. 132. Wickner, W. Driessen, A. J. M., & Hard, E-U. (1991). The enzymology of protein translocation across the Escherichia coli plasma membrane. Annu. Rev. Biochem. 60, 101-124.

BIOCHEMICAL ANALYSES OF COMPONENTS COMPRISING THE PROTEIN TRANSLOCATION MACHINERY OF ESCHERICHIA COL/

Shin-ichi Matsuyama and Shoji Mizushima

I. II. III. IV. V.

Vl.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Components Involved in Protein Translocation . . . . . . . . . . . . . . . . . Overproduction and Purification of Sec Proteins . . . . . . . . . . . . . . . . Numbers of Molecules of Sec Proteins in One E. coli Cell . . . . . . . . . . . Functions of Sec and Related Proteins . . . . . . . . . . . . . . . . . . . . . . A. SecB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. SecA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. SecY and SecE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. SecD and SecF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. SecG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Signal Peptidases and Signal Peptide Peptidases . . . . . . . . . . . . . . . . A. Signal Peptidases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Signal Peptide Peptidases . . . . . . . . . . . . . . . . . . . . . . . . . .

Advances in CeU and Molecular Biology of Membranes and Organelles Volume 4, pages 61-84. Copyright 9 1995 by JAI Press Inc. All fights of reproduction in any form reserved. ISBN: 1-55938-924-9 61

...

62 62 65 65 67 67 68 69 71 72 72 72 73

SHIN-ICHI MATSUYAMAand SHOJIMIZUSHIMA

62

VII. Other Factors Involved in Protein Translocation . . . . . . . . . . . . . . . . . VIII. Model for Protein Translocation . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I.

73 75 77 78

INTRODUCTION

Proteins synthesized on ribosomes in the cytosol have to be targeted to their final destinations for them to exhibit their intrinsic functions. Cells of gram-negative bacteria, such as Escherichia coil, comprise four compartments: the cytoplasm, cytoplasmic membrane, periplasm, and outer membrane. Proteins of the outer membrane and of the periplasm have to be translocated across the cytoplasmic membrane to reach their destinations as genuine secretory proteins do. Proteins are usually translocated across the cytoplasmic membrane in a precursor form (preprotein) that possesses an extra polypeptide, called the signal peptide, at the N-terminus. The translocation shares common features with that of the endoplasmic reticulum membrane of eukaryotes. The protein translocation in both cases requires ATP (25, 38, 44,103,128,132) and a signal peptide, which comprise the N-terminal positively charged region and a long hydrophobic stretch (54, 91). Some prokaryotic preproteins can be translocated across the endoplasmic reticulum membrane and some eukaryotic ones across the bacterial cytoplasmic inembrane (74, 117). Homologues of E. coli SecY, one of the membrane components of the translocation machinery, have been identified in the endoplasmic reticulum membranes of yeasts (114) and mammals (40). Homologues of the mammalian signal recognition panicle, which is involved in the targeting of preproteins to the endoplasmic reticulum membrane, have been found in E. coli (13, 95, 96, 100, 102) and yeasts (9, lO, 42, 43). These facts suggest that the mechanisms underlying the translocation of presecretory proteins are similar in all organisms. In this chapter, we summarize recent progress in biochemical studies on protein translocation across the cytoplasmic membrane of E. coil, focusing on our own biochemical studies, including reconstitution ones. We will, of course, mention relevant work done in other laboratories.

II. COMPONENTS INVOLVED IN PROTEIN TRANSLOCATION Protein translocation across membranes is believed to be catalyzed by a specific proteinaceous complex. Extensive genetic studies on E. coil have revealed that.at least six gene products (SecA, SecB, SecD, SecE, SecF, and SecY proteins) are involved in the general protein translocation across the cytoplasmic membrane (14). In addition to these gene products, several factors have been reported to participate in the translocation process. They are summarized in Table 1 together with Sec

Table 1. Components Participating in the Protein Translocation across the Cytoplasmic Membrane in Escherichia coli Component

Molecular Weight

CellularLocalization

SecB

18 K(subunit)

Cytoplasm

SecA

102 K (Subunit) 204 K (homodimer)

CytopCytoplasmic mtmbrane (WPM)

SecY

Cytoplasmic membrane (integral)

SecE

Cytoplasmic mcmbrane (Integral) Cytoplasmic membrane (with a large periplasmic domain)

SecD

SecF Sed.3 ( ~ 1 2 )

67 K

Cytoplasmic membrane (with a large paiplasmic domain) Cytoplasmic membrane (in@@)

Function

Translocation-specificchaperone Interact with SecA General molecular chaperone General molecular chaperone Essential component for translocation TranslocationATPase Interact with preprotein (signal peptide), SecB, SecY. and SecE Essmtial component for translocation Interact with preprotein (signal peptide), SecA. SecE SecD, SccF, p12, and Ydr Essential component for translocation Interact with signal peptide, SecA, and SecY ihcntial component for translocation Interact with SecY and SecF Involved in release of secretory proteins from the cytoplasmic membrane Component required for hanslocation Interact with SecY and SecD Interact with SecY

Table 1. (continued) Molecular Wig&

Co??pnent

Ydr

19 K

Signal pcptidasc I

36 K

Signal peptidasc 11

18 K

Signal pcptide peptidase

34 K

Cellulor Localization

Cytoplasmic membrane (@PM) Cytoplasmic membrane (integral) cytoplasmic manbnme

Ffh

g 4.5s RNA P~PA

cyt0pb

26 K

Cytoplasmic mmbnmc

Fvnction Intaact with SecY

Cleavage of signal peptides of pnprotcins except prolipoproteins Cleavage of signal peptides of prolipoproteins Rotease 1V:Hydrolysisof cleaved signal pcptih Homologue of 54 K subunit of mammalian SRP Form a complex with 4.5s RNA Homologue of 7s RNA of mammalian SRP Fonn a complex with Ffh Stimulate translocation InmmoIecular disufide bound formation in

Protein Translocation of E. Coli

65

proteins. In the following sections, the roles of these components will be discussed, focusing on Sec proteins from the biochemical point of view.

III. OVERPRODUCTION AND PURIFICATION OF SEC PROTEINS The cellular contents of Sec proteins in E. coil are not high. To obtain large quantities of purified Sec proteins for biochemical studies, including reconstitution ones, the overproduction of Sec proteins by means of gene manipulation was carried out. Since the overproduction of proteins, especially membrane proteins, often causes cell death, the sec genes encoding Sec proteins were placed under a controllable expression promoter on a plasmid. The tac and T7 promoters were judged to be suitable for this purpose. The construction of plasmids carrying the sec genes controlled by these promoters and overproduction of all Sec proteins were successfully achieved (23, 60, 77, 79, 129). SecA, SecB, SecE, and SecF could be overproduced alone, whereas the overproduction of SecD or SecY alone resulted in its rapid breakdown (4, I03a). The overproduction of SecD and SecY was achieved with the simultaneous overproduction of SecF and SecE. (77, I03a), suggesting the occurrence of interactions between SecD and SecF, and SecY and SecE. Interaction between SecY and SecE has also been suggested genetically (15) and biochemically (22). In addition, the SecF-dependent overproduction of SecY was observed, suggesting interaction between SecF and SecY (K. Sagara, S. Matsuyama, and S. Mizushima, unpublished data). All the overproduced Sec proteins have been purified (Figure 1). SecA was purified from the cytosol fraction, which contained most of this protein in the overproducing cells (2, 60). During the purification more than 50% of the SecA molecules lost the eight amino acid residues comprising the N-terminus (82).. Intact SecA was only obtained when the purification was performed in the presence of six protease inhibitors (82). SecD, SecE, SecF, and SecY were purified from the detergent-solubilized membrane fraction containing the overproduced amount of the individual Sec proteins (1, 79, 120). Sarcosyl was used for the purification of SecF and octylglucoside for the other Sec proteins.

IV. NUMBERS OF MOLECULES OF SEC PROTEINS IN ONE E. C O L I CELL The amounts of SecA, SecD, SecY, SecE, and SecF in overproducing cells were determined densitometrically after SDS-gel electrophoresis of the cellular fractions and staining with Coomassie Brilliant Blue. The overproduction (fold)of individual Sec proteins was then determined by means of immunoblot analysis of the cellular fractions of normal and overproducing cells. From these values, the numbers of individual Sec proteins in one normal E. coli cell were estimated (Table 2) (79).

66

SHIN-ICHI MATSUYAMA and SHOJI MIZUSHIMA

SecA SecE SecY SecD SecF p12

F~ure 1. SDS-Polyacrylamide gel profile of purified Sec proteins, pl 2 is now called

SecG (86).

SecE, SecY, SecD, and SecF are cytoplasmic membrane proteins, whereas only 10% of SecA is usually located in the cytoplasmic membrane of normal cells (79). Thus, numbers of molecules of Sec proteins, except SecF, in the cytoplasmic membrane of one E. coli cell were assumed to be approximately 500. The number of molecules of SecG (p 12), a recently found membrane protein involved in protein translocation (86), is similar. This figure may represent the upper limit for the amount of protein translocation machinery in one E. coli cell. It is unclear how many SecE, $ecY, SecD, and SecG molecules exist in one such unit of machinery. SecA was found to exist as a dimer (3). The smaller number of SecF molecules suggests that the role of this protein is different from those of the other Sec proteins.

Protein Translocation of E. Coli

67

Table 2. Number of Molecules of Sec Proteins in one E. coli cell Protein

SecA SecE SecY SecD SecF SecG ( p 1 2 )

Protein Content (%)

Overproduction(fold)

25.0/WC 3.4/M 5.3/M 6.9/M 16.5/M 15.7/M

50.-100 80-160 50-100 20-40 1500-3000 200-400

Number of Molecules per One Normal Cell 2500-5000 300-.6(g) 206..400 450-900 30-60 650-.1300

Note: WC: Wholecell;M: Membrane

V. FUNCTIONS OF SEC AND RELATED PROTEINS A. SecB

SecB, a cytosolic protein, is a homotetramer of a 16-kDa subunit (127) and is required for the translocation of a subset of secretory proteins, including maltosebinding protein (MBP) and LamB, in vivo (65, 66) and in vitro (127, 129). The conformation of preproteins is often crucial for their translocation across membranes (./2, 98). Preproteins folded into a tight conformation cannot be translocated. SecB binds directly to preproteins to inhibit their folding into a translocation-incompetent conformation (26, 64, 71, 126), and the signal peptide was assumed to play a role in this interaction (8, 126). However, this does not necessarily exclude the participation of the mature domain in the binding. In fact, SecB-dependent in vivo protein translocation was inhibited by the overproduction of mature MBP lacking the signal peptide (11, 26). The exchange of signal peptides between MBP and alkaline phosphatase, of which the translocation is SecB-dependent and SecBindependent, respectively, has no effect on their SecB-dependence (35). Finally, SecB forms complexes not only with preproteins but also with denatured mature proteins (45, 71). It is suggested, therefore, that SecB can bind to the mature domain of preproteins and that the signal peptide is not absolutely required for the recognition by SecB. In fact, genetic analyses have revealed that SecB-binding regions are located between amino acid residues 151 and 186 (26) and 72 and 200 (35) of the mature domain of MBR and between 320 and 380 of that of LamB (7). No characteristic features common to these regions have been found, however. Randall proposed that SecB recognizes a positively charged region of proteins but not a specific amino acid sequence (97). Her and another group proposed that signal peptides play a role in the retardation of the folding of the mature domain (45, 70,

99). A factor that exhibits antifolding activity, such as SecB, is called a molecular chaperone (39). Some other chaperones have also been reported to be involved in

68

SHIN-ICHI MATSUYAMAand SHOJl MIZUSHIMA

protein translocation: GroEL is essential for the secretion of ~-lactamase in vivo (69) and in vitro (18). Overproduced amounts of GroEL and DnaK improved the export of LamB-LacZ fusion proteins (92). It is probable that several molecular chaperones play similar roles in keeping the conformation of preproteins translocation-competent. The trigger factor was originally identified as a factor that interacts with proOmpA to facilitate in vitro translocation across the cytoplasmic membrane (27). In vivo genetic analyses revealed, however, that the translocation of proOmpA is normal in trigger-factor-deficient cells (41). B. SecA

SecA is a homodimer of a 102-kDa subunit (3, 108) and is located both in the cytoplasm and on the inner surface of the cytoplasmic membrane (28, 89). The direct involvement of SecA in protein translocation has been demonstrated both in

vitro (23, 60) and in vivo (88, 89). The interaction between SecA and preproteins was revealed by a chemical cmss-linldng study (2). The interaction was signal-peptide-dependent, and an increase in the N-terminal positive charges of the signal peptide resulted in an increase in the efficiency of the cross-linking of preproteins to SecA. This result, together with the fact that the cross-linking was inhibited under high salt conditions, indicates that an electrostatic interaction is important for the SecA-preprotein interaction. SecA is negatively charged as a whole. Very recently evidence for recognition by SecA of the signal hydrophobic region was also presented (H. Mori and S. Mizushima et al., unpublished data). SecA alone exhibits weak ATPase activity (72). The activity was enhanced in the presence of prepmteins and membrane vesicles that contain acidic phospholipids and SecY/SecE (29, 72, 73). The enhanced ATPase activity was termed translocation ATPase activity. Both the translocation ATPase activity and the SecA-dependent protein translocation were inhibited by sodium azide (90), and azide-resistant mutations have been mapped to the secA gene (34, 90), suggesting that SecA is the primary target of this inhibitor. The regions in the SecA molecule that interact with ATP and preproteins have been determined by us (62, 80). The ATP-binding region of SecA is located within the first 217 amino acid residues from the H-terminus. A study involving 8-azido ATP by Lill et al. (72) demonstrated three ATP-binding sites in SecA, although their locations were not determined. The ATP-binding region determined by us is presumably the high-affinity one of the three. The preprotein-binding region of SecA was determined to be located between amino acid residues 267 and 340 (62), and the N-terminal half of SecA was reported to be involved in the interaction with the membrane (24). All the secA(ts) mutations so far reported have been mapped within the first 200 amino acid residues (108). Thus, all these results support the importance of the H-terminal region for the SecA function.

Protein Translocation orE. Coil

69

The ATP-binding region is adjacent to the preprotein-binding region in the SecA molecule (62), suggesting that these two regions may be functionally related. Consistent with this view, (i) ADP/ATP on the SecA molecule was released upon the addition of preproteins (110); (ii) ATP enhanced the cross-linking of preproteins to SecA (2); and (iii) a protease-sensitivity test revealed that SecA takes on different conformations in the presence of ATP and preproteins and that the conformation taken on in the presence of ATP disappeared upon the addition of preprotein (110). Furthermore, evened membrane vesicles caused SecA to take on another conformation. These results further suggest that SecA takes on different conformations upon interaction with different components in the process of protein translocation. Since SecA is the only protein of the secretory machinery that is known to interact with ATP, it is likely that such changes in the SecA conformation constitute the driving force for the transmembrane movement of preproteins. Phosphatidylglycerol and carcliolipin, acidic phospholipids of E. coli, stimulated protein translocation and SecA translocation ATPase activity (30, 68, 73, 94). These phospholipids made SecA sensitive to V8 protease (110), suggesting that they also modulated the conformation of SecA. The insertion of SecA into a phospholipid monolayer was enhanced when the layer contained these acidic phospholipids (19). Several lines of evidence support the view that preproteins first interact with the soluble SecA to initiate the translocation reaction: (1) Preproteins can directly bind to soluble SecA in a manner that is dependent on the strength of the positive charge at the N-terminus of the signal peptide, as is the translocation reaction (2); (2) In vitro protein translocation is stimulated by externally added soluble SecA in an amount that is much more than that which can be retained by the membrane (131); and (3) Different preproteins require different concentrations of SecA for efficient in vitro translocation (48). If the membrane-bound SecA is the initial site of preprotein binding, such a difference in SecA requirement would not be observed. It is likely, therefore, that SecA molecules alternate between the two forms in the process of protein translocation. The following evidence suggests that ATP, and probably its hydrolysis, are involved in the SecA alternation: (1) Exchange between membrane bound B. subtilis SecA and soluble E. coli SecA was only observed in the presence of ATP (122); (2) SecA (Ts) proteins, which possess mutations within the ATP-binding region, were preferably localized in the cytoplasmic membrane (24); and (3) ATR but not nonhydrolyzable ATP analogues, inhibited the insertion of SecA into the phospholipid monolayer (19).

C. SecYand SecE SecY and SecE are integral membrane proteins that span the cytoplasmic membrane ten and three times, respectively (.5, 105). Mutations on the secu and secE genes confer general protein export defects (56, 101), suggesting the direct participation of these proteins in protein translocation. Re,constitution studies have shown this directly. Proteoliposomes reconstituted from purified SecY, SecE, and E. coli

70

SHIN-ICHI MATSUYAMAand SHOJl MIZUSHIMA

phospholipids exhibited protein translocation activity that was dependent on SecA and ATP (1). Omission of either one of these Sec proteins resulted in the complete loss of the activity, suggesting that SecY, SecK and SecA are indispensable proteinaceous components of the protein translocation machinery. Reconstitution of translocation-active proteoliposomes from a complex composed of SexY, SecE, and another protein called band 1 was also performed (22). Recently band 1 was found to be identical with SecG (30a). The SecE molecule has three transmembrane segments. Genetic and reconstitution studies have revealed that a truncated fragment containing only the C-terminal segment was 50% as active as intact SecE (85,104). This fragment stabilized SecY in the membrane, as does intact SecE, suggesting that the third transmembrane segment directly interacts with Sexy (85). A gene in Bacillus subtilis corresponding to E. coil secE has been identified and found to complement the E. coil secE (cs) mutation (57). The deduced amino acid sequence suggests that 8. subtilis SecE has only one transmembrane segment (57). The homology between this segment and the E. coil C-terminal segment is quite low, however. What are the functions of SecY and SecE in the translocation reaction? Since the primary structure of the mature domains of preproteins that can pass through the cytoplasmic membrane is highly hydrophilic (81), it is unlikely that secretory proteins can be translocated directly through the lipid phase of the membrane. The interaction of SecA with SecY and SecY/SecE was demonstrated (15, 22, 33, 72, 73, 77, 106). High affinity binding of SecA to the membrane required SecY/SecE (46). These facts suggest that SecY/SecE functions as a receptor for SecA, enabling the transfer of preproteins from SecA to the SecY/SecE-containing translocation machinery in the membrane. It is assumed that protein translocation takes place through a specific channel. SecY possesses ten membrane-spanning segments, all of which are seemingly required for the translocation reaction (Nishiyama et al., unpublished data), whereas only one such segment is required for SecE to function (85, 104). It is likely that SecY constitutes the major portion of the channel. In this respect, it is interesting that translocation intermediates arrested in the membrane were directly crosslinked to SecY but only limited cross-linking to phospholipids (58). It should be noted, however, that SecY is highly hydrophobic. It is unclear how SecY can be responsible for the formation of a hydrophilic pore, if one exists. Simon and Blobel electrophysiologically observed translocation-dependent channel activity by fusing E. coil membrane vesicles to a planar lipid bilayer (112). The channel activity, observed for ion (K+) passage, was signal.peptide.dependent. Schiebel and Wickner (107) demonstrated that SecA- and AlP-dependent protein translocation increases the membrane permeability to halide anions. Kawasaki et al. (61), on the other hand, observed the countermovement of protons during the protein translocation into membrane vesicles that contain overproduce~ amounts of SecY and SecE. The proton effiux depended on a presecretory protein, ATP, SecA, and SecY. The addition of the mixture of signal peptide and mature protein

Protein Translocation of E. Coli

71

did not cause the proton efflux in this case. It is likely that the translocation reaction is coupled with the opening up of a channel, of which the major constituent is SecY. It is still unclear, however, whether the channel thus formed is responsible for the protein passage. D. SecD and SecF

SecD and SecF are integral membrane proteins that appear to span the cytoplasmic membrane six times (36). Genetic studies have suggested that both proteins are important for in vivo protein translocation (36). Extensive biochemical analyses including reconstitution studies, however, have failed to demonstrate the involvement of SecD and SecF in protein translocation (79). It has been suggested that SecD and SecF each possess a large periplasmic domain (36). Furthermore, the prl mutations, which suppress signal peptide defects, have not been mapped to the secD and SecF genes (36, 113). These facts suggest that SecD and SecF play a role in a late step of protein translocation, when the signal peptide is no longer required. If this is the case, the functions of SecD and SecF may not be detectable with the current in vitro assay method, in which translocation activity is measured as the resistance of preproteins to externally added protease, namely, the entry of preproreins into membrane vesicles. Recently, the role of SecD in a late step of protein translocation was demonstrated by employing E. coli spheroplasts (78). The spheroplasts secreted authentic periplasmic and outer membrane proteins into the medium. When the spheroplasts were pretreated with anti-SecD antibody, the secretion of MBE a periplasmic protein, and OmpA, an outer membrane protein, into the medium was inhibited, and accumulation of their precursors in the spheroplasts was observed. Interestingly, a time-course experiment revealed that the anti-SecD antibody treatment resulted in simultaneous accumulation of the mature form of MBR which has been processed for the signal peptide and localized on the outer surface of the spheroplasts. These results indicate that SecD is involved in a late step of protein translocation and probably plays a role in the release of translocated secretory proteins from the cytoplasmic membrane. It is unclear, however, whether SecD is directly involved in the release from the membrane or the final stage of the translocation reaction (i.e., export of the C-terminus of secretory proteins). The secD and secF genes constitute an operon (36). The membrane topologies of SecD and SecF are thought to be similar, and their primary sequences are partially homologous, suggesting that the two proteins perform related functions (36, 78). However, no biochemical studies, including ones involving anti-SecF antibody, have demonstrated the functioning of SecF in the translocation reaction. Very recently, Sagara et al. (103a) of our group showed in vivo that SecF stabilizes not only SecD but also SecY. It is probable, therefore, that SecF plays a role in the translocation event by interacting with both SecD and SecY.

72

SHIN-ICHI MATSUYAMAand SHOJI MIZUSHIMA

E. SecG All the components of the protein translocation machinery described so far were originally identified through genetic studies. Recently, a new proteinaceous factor for protein translocation was discovered using a reconstitution system and named p 12 (86) and later on SecG (86a). SecG is a 12-kDa cytoplasmic membrane protein. When purified SecG was added to the reconstitution mixture together with SecY and SecE, the translocation activity of the reconsfitmed proteoliposomes was enhanced more than 20-fold. The translocation thus enhanced was ATP- and SecA-dependent. Furthermore, anti-SecG antibody inhibited the translocation of proteins into everted membrane vesicles. The gene encoding SecG was identified and sequenced, and the amino acid sequence of SecG was deduced. This protein possesses two or three possible membrane-spanning domains. The cloning of the gene enabled the overproduction of SecG. The overproduction of SecG supported the simultaneous overproduction of SecY, as SecE does, suggesting that this protein also interacts with SecY in the membrane (K. Nishiyama, S. Mizushima, and H. Tokuda, unpublished data). A secG::kan null mutant was cold-sensitive for protein export and growth (86a). VI.

SIGNAL PEPTIDASES A N D SIGNAL PEPTIDE PEPTIDASES

A. Signal Peptidases E. coil has two types of signal peptidases that cleave the signal peptides of preproteins during protein translocation. Signal peptidase II (lipoprotein signal peptidase) cleaves the signal peptides ofglyceride-modified prolipoproteins. Signal peptidase I (leader peptidase) apparently cleaves the signal peptides of all preproteins except prolipoproteins. Signal peptidase H is an integral membrane protein that spans the cytoplasmic membrane four times (83). The amino acid sequence, Leu-(Ala Ser)-(Gly Ala)-Cys, surrounding the cleavage site, is conserved in all prolipoproteins, suggesting this domain to be the site of recognition by the enzyme (47). Glyceride modification of the cysteine residue immediately after the cleavage site is a prerequisite for the proteolytic processing (51). This processing is specifically inhibited by the antibiotic, globomycin (50). Signal peptidase I is a cytoplasmic membrane protein that possesses a large periplasmic region functioning as the catalytic domain (17). There are some common features around the cleavage site of signal peptides, the target domain of this enzyme: Small apolar amino acids (Ala, Ser and Gly) at- 1 and-3 are important for the processing by signal peptidase I (91, 109, 123). Pro and Thr at +1 strongly inhibit the cleavage of signal peptides (75,109). The {3-turn structure at the cleavage

Protein TranslocationorE. Coli

73

site for recognition by signal peptidase I was purported to be important (31, 91). No protease inhibitor has been found to inhibit signal peptidase I activity. Recently, the Ser90 residue was determined to be essential for the function of this peptidase (115), suggesting that signal peptidase I belongs to a novel class of serine proteases. The precise structure of the catalytic domain remains unclear, however. About 500 signal peptidase I molecules exist in one E. coil cell (130). This number is close to that estimated for See proteins comprising the translocation machinery (79), suggesting that signal peptidase may be part of the translocation machinery together with Sec proteins. It should be noted, however, that the cleavage of signal peptides is not essential for protein translocation in vivo (53) or in vitro (133), although the release of translocated proteins from the membrane requires cleavage. Purified signal peptidases actively cleave signal peptide, even in the absence of See proteins (130). Furthermore, the sec mutation has not been mapped to the signal-peptidase-encoding genes. It is likely, therefore, that the function of signal peptidases is not directly involved in the translocation event, although the cleavage of signal peptides is usually tightly coupled with protein translocation.

B. Signal Peptide Peptidases Signal peptides cleaved from preproteins by signal peptidases are rapidly degraded in the membrane. Protease IV, a cytoplasmic membrane protease named signal peptide peptidase, was found to specifically hydrolyze cleaved signal peptides (52). Signal peptides attached to the mature domain were not attacked by this enzyme. The disruption of the protease IV-encoding gene, spp, on the chromosome did not result in accumulation of signal peptides (116), indicating that other peptidases could also be involved in the signal peptide hydrolysis. Two cytoplasmic enzymes were found to digest cleaved signal peptides (87).

VIi. OTHER FACTORS INVOLVED IN PROTEIN TRANSLOCATION Ffh Protein and 4.5S RNA

Thesignal recognition particle (SRP) is a mammalian cytosolic factor which transfers nascent presecretory proteins from the cytosol to the endoplasmic reticulure membrane (124). SRP consists of six distinct proteins and one 7S RNA (125). The 54-kDa subunit, one of the protein components of SRP, directly interacts with signal peptides (67). Recently, the E. coli homologue of the 54-kDa subunit and the 7S RNA of SRP were identified. They were identified as the 48-kDa Ffh protein (13, 102) and 4.5S RNA (95, 96, 100), respectively. Ffh and 4.5S RNA form a complex (95, 100), which interacts with the signal peptide domain of nascent

74

SHIN-ICHI MATSUYAMA and SHOJl MIZUSHIMA

preprolactin and E. coil prolipoprotein (76), suggesting that the complex may be the E. coli homologue of SRP. Genetic analyses have revealed that the genes encoding Ffh and 4.5S RNA are essential for cell growth (21, 93). Precursors of 13-1actamase, alkaline phosphatase, and ribose-binding protein accumulated in the Ffh-depleted cells (93), whereas only the I]-lactamase precursor did so in the cells lacking 4.5S RNA (100). Translocation of these three preproteins was SecB-independent (65, 66). On the other hand, depletion of Ffh or 4.5S RNA did not have a significant effect on the translocation of MBP and LamB, whose export is SecB-dependent (65, 66). The Ffh-4.5S RNA complex may act as a molecular chaperone for a group of preproteins, as SecB does for another group. Components corresponding to Ffh and 4.5S RNA of E. coli have also been identified in Bacillus subtilis (49, 84). They are required for cell growth and the depletion of either B. subtilis Ffh or scRNA caused defective secretion of o~-amylase.

PspAProtein PspA, a peripheral cytoplasmic membrane protein, was originally identified as a protein induced by fl phage infection (20). This protein has a stimulative effect on protein translocation both in vitro and in vivo, although it is not absolutely required (63).

DsbA (PpfA) Protein This is a periplasmic protein which catalyzes intramolecular disulfide bond formation (12, 59). It plays an essential role in the folding of disulfide bond-possessing proteins in the periplasm (6, 59). However its function may not be directly involved in the translocation event.

Ydr Protein A gene of which overexpression suppresses the translocation defect caused by the secFamutation, a dominant negative mutation of the secY gene, has been identified (55). The gene, ydr, seems to encode a 19-kDa peripheral membrane protein. The overproduction of Ydr stabilized the overproduced level of SecY, similar to SecE, suggesting an interaction between Ydr and SecY. However, the role of Ydr in protein translocation remains unknown.

75

Protein Translocation orE. Coli

Chaperone--.-.--)~Preprotein ~ Signalpeptide

,~ ~

'~ATP

....

...

"l'l~"SecE

~'Signa,

SecE [ sock, p12]

gnal l]I ] Si,e~

~-~--__.._d

-

Figure 2. Model for protein translocation. Although SecD, SecF, signal peptidase and signal peptide peptidase are depicted only at a certain stages of translocation, they are assumed to be in the membrane throughout the translocation reaction, p12 = SecG; PMF; proton motive force.

VIII. MODEL FOR PROTEIN TRANSLOCATION Our current model for the translocation of preproteins across the cytoplasmic membrane of E. coli is shown in Figure 2. The first step is the binding of molecular chaperones to nascent polypeptide chains of preproteins being synthesized on ribosomes. This binding event prevents the polypeptide chains from folding into a translocation-incompetent conformation (26, 64, 71, 126). SecB, GroEL, and probably Ffh.4.SS RNA, function as

76

SHIN-ICHI MATSUYAMAand SHOJl MIZUSHIMA

chaperones. There is no evidence to suggest a tight coupling of the translation reaction to the translocation reaction in prokaryotes. Preproteins then interact with SecA with the aid of the signal peptide (2). ATP stimulates SecA-preprotein interaction, possibly by making the SecAconformation suitable for the binding (62). The preprotein binding in turn causes the release of ATP/ADP from SecA (110). We propose that cytosolic SecA is primarily responsible for the interaction with preproteins (82). The signal peptide domain is essential for this interaction, and the positively charged amino acid residues at its N-terminus are involved in the interaction with SecA (2, 48, Moil, H. et al., unpublished data). The initiation complex, comprising a preprotein, SecA, and probably a chaperone, is then moved onto the secretory machinery in the membrane. It is proposed that the preproteins are transferred to SecA on the membrane in a complex with a chaperone (46). It remains unknown when chaperones dissociate from preproteins. SeeA, SexY, SecE, SecG, and SecD occur in the membrane roughly in an equimolar ratio, suggesting that they may constitute the secretory machinery. Interactions between these See proteins have been demonstrated in vivo (15, 16, 77, 79, 85, 103a: K. Nishiyama, S. Mizushima, and H. Tokuda, unpublished data) and in vitro (22, 23, 72, 73) (Figure 3). On the other hand, the results ofstoichiometric studies on SexY, SeeE, and SecG involving reconstitution suggest that these proteins may not exist as a tight complex in the membrane. They may form the complex machinery only in a preprotein-dependent manner. Joly et al. reported recently that SecE and SecY form a stable complex, while band 1 (SecG) is exchangeable (58a). It is unclear how preproteins are transferred from the cytosolic SecA to the machinery in the membrane. They may be transferred from the cytosolic SeeA to the membrane-bound SecA. Alternatively, the preprotein-SecA complex may replace SeeA pre-existing on the membrane. In any event, the transloeation complex comprising a preprotein, SecA, SecY, SecE, SeeD, and SecG can be formed. For the translocation of the first small domain (20--30 amino acid residues) of the preprotein, ATP, but not its hydrolysis, is apparently required, because the signal peptide cleavage that represents an early stage of the translocation reaction can be observed in the presence of AMP-PNP (106). The subsequent process, the translocation of the polypeptide chain of the mature domain, requires both ATP hydrolysis and protonmotive force (PMF) (106, 118). It is proposed that ATP hydrolysis is needed for SecA to dissociate from the translocation complex (106) and PMF is needed to release ADP subsequently formed from SeeA (II1) and to induce transmembrane movement of the protein that has been released from SecA (106). Subsequently, SecA rebinds to the translocation intermediate to allow further translocation. Thus, stepwise translocation can be assumed (106). Stepwise translocation by about 30 amino acid residues was recently observed using mutant proOmpAs possessing a looped structure in different positions of the mature domain (K. Uchida and S. Mizushima, unpublished data). It is also interesting in this respect that SecA was reported to undergo membrane insertion and deinsertion

Protein Translocation orE. Coli

Preprotein~. ~ ~.....

77

~

~ Translocatedprotein

Figure 3. Interaction map of components comprising the protein translocation machinery, pl 2 = SecG

during the translocation reaction (31a, 61a). Although ATP is essentially required for the translocation of the polypeptide chain to a certain stage, a later stage can proceed without ATP hydrolysis when PMF is supplied as an energy source (106,

118,119). Finally, proteins are released from the membrane with the aid of SecD (78) and PMF (37, 121). The precise molecular mechanism of the SecD function remains unknown. It may be involved in the pulling out of the C-terminal domain from the membrane, or simply the release of the completely translocated proteins from the membrane. The interactions between components comprising the translocation machinery are summarized in Figure 3. The interaction map is based on genetic and biochemical evidence. As described above, SecA, SecE, SecY, and p 12 are likely involved in a relatively early step in the translocation process, and interaction among these proteins has been suggested. On the other hand, SecD was shown to participate in a late stage of the translocation reaction (78). The sec titration study also suggested that SecD/SecF follows SecY in the translocation reaction (16). It is interesting in this respect that a stabilization test suggested that SecF interacts with both SecY and SecD (103a). It is probable that SecF connects the early and late steps of the translocation reaction. However, no biochemical evidence supports this view.

ACKNOWLEDGMENTS

We thank Dr. Hajime Tokuda for helpful comments and Ms. Chikako Motoyama for the secretarial support. The work from the authors' laboratories has been supported by grants from the Ministry of Education, Science and Culture of Japan; the Human Frontier Science Program Organization; the Nisshin Foundation; and the Naito Memorial Foundation.

78

SHIN-ICHI MATSUYAMAand SHOJl MIZUSHIMA

REFERENCES 1. Akimaru, J., Matsuyama, S., Tokuda, H., & Mizushima, S. (1991). Reconstitution of a protein translocation system containing purified SecY, SecE and SecA from Escherichia coil Proc. Natl. Acad. Sci. USA 88, 6545--6549. 2. Akita, M., Sasaki, S., Matsuyama, S., & Mizushima, S. (1990). SecA interacts with secretory proteins by recognizing the positive charge at the amino terminus of the signal peptide in Escherichia coli.J. Biol. Chem. 265, 8164-8169. 3. Akita, H., Shinkai, A., Matsuyama, S., & Mizushima, S. (1991). SecA, an essential comlxment of the secretory machinery of Escherichia coil, exists as homodinter. Biochem. Biophys. Res. Commun. 174, 211-216. 4. Akiyama, Y., & Ira, K. (1986). Overproduction, isolation and determination ofthe amino-terminal sequence of the SecY protein, a membrane protein involved in protein export in Escherichia coll. Eur. J. Biochem. 159, 263-266. 5. Akiyama, Y., & Ira, K. (1987). Topology analysis of the SecY protein, an integral membrane protein involved in protein export in Escherichia coll. EMBO I. 6, 3465-3470. 6. Akiyama, Y., & Ira, K. (1993). Folding and assembly of bacterial alkaline phosphatase in vitro and in viva. J. Biol. Chem. 268, 8146-8150. 7. Airman, E., Bankaitis, V. A., & Emr, S. D. (1990). Characterization of a region in mature LamB protein that interacts with a component of the export machinery of Escherichia coll. J. Biol. Chem. 265, 18146-18153. 8. Airman, E., Emr, S. D., & Kunmmoto, C. A. (1990). The presence of both the signal sequence and a region of mature Lamb protein is required for the interaction of LamB with the export factor SecB. J. Biol. Chem. 265, 18154-18160. 9. Amaya, Y., & Nakano, A. (1991). SRI-ll protein, the yeast hcm~logue of the 54 kDa subunit of signal recognition particle, is involved in ER translocation of secretory proteins. FEBS Lett. 283, 325-328. 10. Amaya, Y., Nakano, A., Ira, K., & Marl, M. (1990). Isolation of a yeast gene, SRHI, that encodes a homologue of the 54 K subunit of mammalian signal recognition particle. J. Biochem. 107, 457-463. 11. Bankaitis, V. A., & Bassford, P. J. Jr. (1984). The synthesis of export-defective proteins can interfere with normal protein export in Esclwrichia coil J. Biol. Chem. 259, 12193-12200. 12. Bardwell, J. C. A., McGovem, K., & Beckwith, J. (1991). Identification of a protein required for disulfide bond formation in viva. Cell 65, 581-589. 13. Bernstein, H. D., Poritz, M. A., Strub, K., Hoben, P. J., Brenner, S., & Walter, P. (1989). Model for signal sequence recognition from amino-acid sequence of 54 K subunit of signal recognition particle. Nature 340, 482-486. 14. Bieker, K. L., Phillips, G. J., & Silhavy, T. J. (1990). The sec and pd genes of Escherichia coll. J. Bioenerg. Biomembr. 22, 291-310. 15. Bieker, K. L., & Silhavy, T. J. (1990). PrlA (SecY) and PrIG (SecE) interact directly and function sequentially during protein translocation in E. coll. Cell 61,833-842. 16. Bieker-Brady, K., & Silhavy, T. J. (1992). Suppressor analysis suggests a multistep, cyclic mechanism for protein secretion in Escherichia coli. EMBO J. I 1, 3165-3174. 17. Bilgin, N., Lee, J. I., Zhu, H.-Y., Dalbey, R., & yon Heijne, G. (1990). Mapping of catalytically important domains in Escherichia coli leader peptidase. EMBO J. 9, 2712-2722. 18. Bochkareva, E. S., Lissin, N. M., & Gishovich, A. S. (1988). Transient association of newly synthesized unfolded proteins with heat shock GroEL protein. Nature 336, 254-257. 19. Breukink, E., Demel, R. A., de Korte-Kool, G., & de Kruijff, B. (1992). SecA insertion into phospholipids is stimulated by negatively charged lipids and inhibited by ATP: a monolayer study. Biochemistry 31, 1119-1124.

Protein Translocation of E. Col i

79

20. Brissette, J. L., Russei, M., Weiner, L., & Model, P. (1990). Phage shock protein, a stress protein of Escherichia coli. Proc. Natl. Acad. Sci. USA 87, 862-866. 21. Brown, S., & Fournier, M. J. (1984). The 4.5S RNA gene of Escherichia coli is essential for cell growth.'J. Mol. Biol. 178, 533-550. 22. Brundage, L., Hendrick, J. P., Schiebel, E., Driessen, A. J. M., & Wickner, W. (1990). The purified E. coli integral membrane protein SecY/E is sufficient for reconstitution of SecA-dependent precursor protein translocation. Cell 62, 649-657. 23. Cabelli, R. J., Chen, L., Tai, P.-C., & Oliver, D. B. (1988). SecA protein is required for secretory protein translocation into E. coli membrane vesicles. Cell 55, 683-692. 24. Cabelli, R. J., Dolan, K. M., Qian, L., & Oliver, D. B. (1991). Characterization of membrane-associated and soluble states of SecA protein from wild-type and SecA51(TS) mutant strains of Escherichia coli. J. Biol. Chem. 266, 24420-24427. 25. Chen, L., & Tai, P.-C. (1985). ATP is essential for protein translocation into Escherichia coil membrane vesicles. Proc. Natl. Acad. Sci. USA 82, 4384--4388. 26. Collier, D. N., Bankaitis, V. A., Weiss, J. B., & Bassford, P. J. Jr. (1988). The anfifolding activity of SecB promotes the export of the E. coli maltose-binding protein. Cell 53, 273-283. 27. Crooke, E., & Wickner, W. (1987). Trigger factor: a soluble protein that folds pro-OmpA into a membrane-assembly-competent form. Proc. Natl. Acad. Sci. USA 84, 5216-5220. 28. Cunningham, K., Lill, R., Crooke, E., Rice, M., Moore, K., Wickner, W., & Oliver, D. (1989). SecA protein, a peripheral protein of the Escherichia coli plasma membrane, is essential for the functional binding and translocation of proOmpA. EMBO J. 8, 955-959. 29. Cunningham, K., & Wickner, W. (1989). Specific recognition of the leader region of precursor proteins is required for the activation of translocation ATPase of Escherichia coll. Proc. Natl. Acad. Sci. USA 86, 8630-8634. 30. de Vrije, T., de Swat, R. L., Dowhan, W., Tommassen, J., & de Kruijff, B. (1988). Phosphatidylglycerol is involved in protein translocation across Escherichia coil inner membranes. Nature 334, 173-175. 30a.Douville, K., Leonard, M., Brundage, L., Nishiyarna, K., Tokuda, H., Mizushima, S., & Wickner, W. (1994). Band 1 subunit of Escherichia coli preprotein translocase and integral membrane export factor p12 are the same protein. J. Biol. Chem. 269, 18705-18787. 31. Duffaud, G., & Inouye, M. (1988). Signal peptidases recognize a structural feature at the cleavage site of secretary proteins. J. Biol. Chem. 263, 10224-10228. 31a.Economou, A., & Wickner, W. (1994). SecA promotes preprotein translocation by undergoing ATP-driven cycles of membrane insertion and deinsertion. Cell 78, 835-843. 32. Eilers, M., & Schatz, G. (I 986). Binding of a specific ligand inhibits import of a purified precursor protein into mitochondria. Nature 322, 228-234. 33. Fandl, J. P., Cabelli, R., Oliver, D., & Tai, P.-C. (1988). SecA suppresses the temperature-sensitive SecY24 defect in protein translocation in Escherichia coli membrane vesicles. Proc. Natl. Acad. Sci. USA 85, 8953-8957. 34. Fortin, Y., Phoenix, P., & Drapeau, G. R. (1990). Mutations conferring resistance to azide in Escherichia coil occur primarily in the secA gene. J. Bacterioi. 172, 6607-6610. 35. Gannon, P. M., Li, P., & Kumamoto, C. A. (1989). The mature part of Escherichia coil maltose-binding protein (MBP) determines the dependence of MBP on SecB for export. J. Bacteriol. 171, 813-818. 36. Gardel, C., Johnson, K., Jacq, A., & Beckwith, J. (1990). The secD locus of E. coli codes for two membrane proteins required for protein export. EMBO J. 9, 3209-3216. 37. Geller, B. L. (1990). Electrochemical potential releases a membrane-bound secretion intermediate of maltose-binding protein in Escherichia coll. J. Bacteriol. 172, 4870-4876. 38. Geller, B. L., Movva, N. R., & Wickner, W. (1986). Both ATP and the electrochemical potential are required for optimal assembly of pro-OmpA into Escherichia coil inner membrane vesicles. Proc. Natl. Acad. Sci. USA 83, 4219-4222.

80

SHIN-ICHI MATSUYAMAand SHOJl MIZUSHIMA

39. Gething, M.-J., & Sambrook, J. (1992). Protein folding in the ceil. Nature 355, 33--45. 40. G0rlich, D., Prehn, S., Hartmann, E., Kalies, K.-U., & Rapolx~ T. A. (1992). A mammalian homoiog of SEC61p and SECYp is associated with ribosomes and nascent polypeptides during translocation. Cell 71,489-503. 41. Guthrie, B., & Wickner, W. (1990). Trigger factor depletion or overproduction causes defective cell division but does not block protein export. J. Bacteriol. 172, 5555-5562. 20 Harm, B. C., Poritz, M. A., & Walter, P. (1989).Saccharomyces cerevisiae and Schizosaccharomyces pombe contain a homologue to the 54 K subunit of the signal recognition particle that in S. cerevisiae is essential for growth. J. Cell. Biol. 109, 3223--3230. 43. Harm, B. C., & Walter, P. (1991). The signal recognition particle inS. cerevisioe. Cell 67,131-144. 44. Hans, n, W., Garcia, P. D., & Walter, P. (1986). In vitro protein translocation across the yeast endoplasmic reticulum: ATP-del~ndent posttranslocational translocation of the prepro.cz-factor. Cell 45, 397--406. 5. Hardy, S. J. S., & Randall, L. L. (1991). A kinetic partitioning model of selective binding of nonnative proteins by the bacterial chaperone SecB. $cience 251,439-443. 68 Hartl, E-U., Lecker, S., Schiehel, E., Hendricl~ J. P., & Wicimer, W. (1990). The binding cascade of SecB to SecA to SecY/E mediates prewotein targeting to the E. coil plasma membrane. Cell 63, 269-279. in bacteria. J. Bioenerg. Biomembr. 22, 451-472. 7. Hayashi, S., & Wu, H. C. (19c)0). L i ~ e i n s 48. Hikita, C., & Mizushima, S. (1992). The requirement of a positive charge at the amino terminus can be compensated for by a longer central hydrophobic stretch in the functioning of signal peptides. J. Biol. Chem. 267, 12375-12379. 90 Honda, K., Nakamma, K., Nishiguchi, M., & Yamane, K. (1993). Cloning and characterization of a Bacillus subtilis gene encoding a homolog of the 54-Kilodalton subunit of mammalian signal recognition particle and Escherichia coli Ffh. J. Bacteriol. 175, 4885-4894. 0e Hussain, M., lchihara, S., & Mizushima, S. (1980). Accumulation of glyceride-containing precursor of the outer membrane lipoprotein in the cytoplasmic membrane of Escherichia coil treated with giobomycin..I. Biol. Chem. 255, 3707-3712. 51. Hussain, M., Ichihara, S., & Mizushima, S. (1982). Mechanism of signal peptide cleavage in the biosynthesis of the major lipowotein of the Escherichia coli outer membrane. J. Biol. Chem. 257, 5177-5182. 52. Ichihara, S., Beppu, N., & Mizushinm, S. (1984). Protease IV, a cytoplasmic membrane protein of Escherichia coil, has signal peptide peptidase activity. J. Biol. Chem. 259, 9853-9857. 53. Ichihara, S., Hussain, M., & Mizushima, S. (1982). Mechanism of export of outer membrane lipowoteins through the cytoplasmic membrane in Escherichia coli. J. Biol. Chem. 257, 495-500. 4 . lnouye, M., & Halegoua, S. (1980). Secretion and membrane localization of proteins in Escherichia coll. CRC Crit. Rev. Biochem. 7, 339-371. 55. lto, K. (1992). SecY and integral membrane compm~nts of the Escherichia coli wotein translocation system. Mol. Microbiol. 6, 2423-2428. 56. lto, K., Witterkind, M., Nomura, M., Shiba, K., Yura, T., Miura, A., & Nashimoto, H. (1983). A temperature-sensitive mutant of E. coli exhibiting slow wocessing of exported proteins. Cell 32, 789-797. 57. Jeong, S. M., Yoshikawa, H., & Takxhashi, H. (1993). Isolation and characterization of the secE homologne gene of Bacillus subtilis. Mol. Microbiol. 10, 133-142. 58. Joly, J. C., & Wickner, W. (1993). The SecA and Sexy subunits of transiocase are the nearest neighbors of a translocating preprotein, shielding it from phospholipids. EMBO J. 12, 255-263. 58a.Joly, J. C., Leonard, M. R., & Wickner, W. T. (1994). Subunit dynamics in Escherichia coil preprotein transiocase. Proc. Natl. Acad. $ci. USA 91, 4703-4707. 59. Kamitani, S., Akiyama, Y., & Ito, K. (1992). Identification and characterization of an Escherichia folded alkaline phosphatase, a periplasmic coli gene required for the formation of ~ l y enzyme. EMBO J. 1I, 57-62.

Protein Translocation of E. Col i

81

60. Kawasaki, H., Matsuyama, S., Sasaki, S., Akita, M., & Mizushima, S. (1989). SecA protein is directly involved in protein secretion in Escherichia coli. FEBS Lett. 242, 431.-434. 61. Kawasaki, S., Mizushima, S., & Tokuda, H. (1993). Membrane vesicles containing overproduced SecY and SecE exhibit high translocation ATPase activity and countermovement of protons in a SecA- and presecretory protein-dependent manner. J. Biol. Chem. 268, 8193-9198. 61a.Kim, Y. J., Rajapandi, T., & Oliver, D. (1994). SecA protein is exposed to the periplasmic surface of the E. coil inner membrane in its active state. Cell 78, 845-853. 62. Kimura, E., Akita, M., Matsuyama, S., & Mizushima, S. (1991). Determination of a region of SecA that interacts with secretory proteins in Escherichia coll. J. Biol. Chem. 266, 6600-.6606. 63. Kleerebezern, M., & Tommassen, J. (1993). Expression of the pspA gene stimulates efficient protein export in Escherichia coli. Mol. Microbial 7, 947-956. 64. Kumamoto, C. A. (1989). Escherichia coli SecB associates with exported protein precursors in viva. Proc. Natl. Acad. Sci. USA 86, 5320-5324. 65. Kumamoto, C. A., & Beckwith, J. (1983). Mutations in a new gene, secB, cause defective protein localization in Escherichia coll. J. Bacterial. 154, 254-260. 66. Kumamoto, C. A., & Beckwith, J. (1985). Evidence for specificity at an early stage in protein export in Escherichia coll. J. Bacterial 163, 267-274. 6"1. Kurzchalia, T. V., Wiedman, M., Girshovich, A. S., Bochkareva, E. S., Kielka, H., & Rapoport, T. A. (1986). The signal sequence of nascent preprolactin interacts with the 54 K polypeptide of the signal recognition particle. Nature 320, 634-636. 68. Kusters, R., Dowhan, W., & de Kruijff, B. (1991). Negatively charged phospholipids restore prePhoE translocation across phosphatidylglyceroi-depleted Escherichia coil inner membrane vesicles. J. Biol. Chem. 266, 8659-8662. 69. Kusukawa, N., Yura, T., Ueguchi, C., Akiyama, Y., & Ira, K. (1989). Effects of mutations in heat-shock genes gmES and groEL on protein export in Escherichia coil EMBO J. 8, 3517-3521. 70. Laminet, A. A., & Pliichthum, A. (1989). The precursor of [~-Iactamase: purification, properties and folding kinetics. EMBO J. 8, 1469-1477. 71. Lecker, S. H., Lill, R., Ziegelhoffer, T., Georgopoulos, C., Bassford, P. J. Jr., Kumamoto, C. A., & Wickner, W. (1989). Three pure chaperones of Escherichia colim SecB, trigger factor and GroEL-form soluble complexes with precursor protein in vitro. EMBO. J. 8, 2703-2709. 72. Lill, R., Cunningham, K., Brundage, L. A., Ira, K., Oliver, D., & Wickner, W. (1989). SecA protein hydrolyzes ATP and is an essential component of the protein translocation ATPase of Escherichia coil EMBO J. 8, 961-966. 73. Liil, R., Dowhan, W., & Wickner, W. (1990). The ATPase activity of SecA is regulated by acidic phospholipids, SecY, and the leader and mature domains of precursor proteins. Cell 60, 271-280. 74. Lingappa, V. R., Chaidez, J., Yost, C. P., & Hedgpeth, J. (1984). Determinants for protein localization [~-Iactamase signal sequence directs globin across microsomal membranes. Proc. Natl. Acad. ScL USA 8 I, 456--460. 75. Lu, H.-M., Yamada, H., & Mizushima, S. (1991). A praline residue near the amino terminus of the mature domain of secretory proteins lowers the level of the proton motive force required for translocation. J. Biol. Chem. 266, 9977-9982. 76. Luirink, J., High, S., Wood, H., Giner, A., Tollervey, D., & Dobberstein, B. (1992). Signal-sequence recognition by an Eschericllia coli ribonucleoprotein complex. Nature 359, 741-743. 77. Matsuyama, S., Akimaru, J., & Mizushima, S. (1990). SecE-dependent overproduction of SecY in Escherichia coil Evidence for interaction between two components of the secretory machinery. FEBS Len. 269, 69-100. 78. Matsuyama, S., Fujita, Y., & Mizushima, S. (1993). SecD is involved in the release oftranslocated secretory proteins from the cytoplasmic membrane of Escherichia coll. EMBO J. 12, 265-270. 79. Matsuyama, S., Fujita, Y., Sagara, K., & Mizushima, S. (1992). Overproduction, purification and characterization of SecD and SecF, integral membrane components of the protein translocation machinery of Escherichia coll. Biochim. Biophys. Acta. 1122, 77-84.

SHIN-ICHI MATSUYAMA and SHOJI MIZUSHIMA

82

~0Q Matsuyama, S., Kimura, E., & Mizushima, S. (1990). Complementation of two overlapping fragments of SecA, a protein translocation ATPase of Escherichia coil, allows ATP binding to its amino-terminal region. J. Biol. Chem. 265, 8760--8765. 81. Mizushima, S. (1986). Assembly of membrane proteins. In: lnouye, M. (Ed.), Bacterial outer membranes as model systems. New York: John Wiley & Sons, Inc., pp. 163-185. 82. Mizushima, S., Tokuda, H., & Matsuyama, S. (1992). Molecular characterization of Sec proteins comprising the protein secretory machinery ofEscherichia coli. In W. Neupert, and R. Lill, (Eds.), Membrane biogenesis and protein targeting New York: Elsevier Science. (pp. 21-32). 83. Munoa, F. J., Miller, K. W., Beers, R., Graham, M., & Wu, H. C. (1991). Membrane topology of Escherichia coli prolipoprotein signal peptidase (signal peptidase II). J. Biol. Chem. 266,

17667-17672. 84. Nakamura, K., lmai, Y., Nakamura, A., & Yamane, K. (1992). Small cytoplasmic RNA of Bacillus subtilis: Functional relationship with human signal recognition particle 7S RNA and Escherichia coli 4.5S RNA. J. Bacterial. 174, 2185--2192. 85. Nishiyama, K., Mizushima, S., & Tokuda, H. (1992). The carboxyl-terminal region of SecE

interacts with SecY and is functional in the reconstitution of protein translocation activity in Escherichia coll. J. Biol. Chem. 267, 7170-7177. 86. Nishiyama, K., Mizushima, S., & Tokuda, H. (1993). A novel membrane protein involved in protein translocation across the cytoplasmic membrane of Escherichia coli. EMBO J. 12, 3409-3415. 86a.Nishiyama, K., Hanada, M., & Tokyda, 14. (1994). Disruption of the gene encoding p12 (SecG) reveals the direct involvement and important function of SecG in the protein translocation of Escherichia co/i at low temperature. EMBO J. ! 3, 3272-3277. 87. Novak, E, Ray, P. H., & Dev, I. K. (1986). Localization anti purification of two signal peptide hydrolases from Escherichia coll. J. Biol. Chem. 261,420--427. 88. Oliver, D. B., & Beckwith, J. (1981). E. coli mutant pleiotropically defective in the export of secreted woteins. Cell 25, 765-772. 89. Oliver, D. B., & Beckwith, J. (1982). Regulation of a membrane component required for protein secretion in Escherichia coli. Cell 30, 311-319. 90. Oliver, D. B., Cabeili, R. J., Dolan, K. M., & Jarosik, G. P. (1990). Azide-resistant mutants of Escherichia coli alter the SecA protein, an azide-sensitive component of the protein export machinery. Proc. Natl. Acad. Sci. USA 87, 8227-8231. 91. Perlman, D., & Halvorson, H. O. (1983). A putative signal peptidase recognition site and sequence in eukaryotic and prokaryotic signal peptides. 3. Mol. Biol. 167, 391--409. 92. Phifipis, G. J., & Silhavy, T. J. (1990). Heat shock proteins DnaK and GroEL facilitate export of LacZ hybrid proteins in E. coli. Nature 344, 882-884. 93. Phillips, G. J. & Silhavy, T. J. (1993). The E. coliJ~z gene is necessary for viability and efficient protein export. Nature 359, 744-746. 94. Phoenix, D. A., Kusters, R., Hikita, C., Mizushima, S., & de Kruijff, B. (1993). OmpF-Lpp signal sequence mutants with varying charge hydrophobicity ratios provide evidence for a phosphatidylglycerol-signal sequence interaction during protein translocation across the Escherichia coil inner membrane. J. Biol. Chem. 268, 17069-17073. 95. Poritz, M. A., Bernstein, H. D., Strub, K., Zopf, D., Wilhelm, H., & Walter, P. (1990). An E. coli ribonucleoprotein containing 4.5S RNA resembles mammalian signal recognition particle. Science 250, 1111-1117. 96. Poritz, M. A., Strub, K., and Walter, P. (1988). Human SRP RNA and E. coli 4.5S RNA contain a highly homologous structural domain. Cell 55, 4-6. 97. Randall, L. L. (1992). Peptide binding by chaperone SecB: implications for recognition of normative structure. Science 257, 241-245.

Protein Translocation of E. Coli

83

98. Randall, L. L., & 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. coil. Cell 46, 921-928. 99. Randall., L. L., & Hardy, S.J.S. (1989). Unity in function in the absence of consensus in sequence: role of leader peptides in export. Science 243, 1156-1159. 100. Ribes, V., ROmisch, K., Giner, A., Dobberstein, B., & Tollervey, D. (1990). E. coli 4.5S RNA is part of a ribonucleoprotein particle that has properties related to signal recognition particle. Cell 63,591-600. 101. Riggs, P. D., Derman, A. J., & Beckwith, J. (1988). A mutation affecting the regulation of a secA-iacZ fusion defines a new sec gene. Genetics 118, 571-579. 102. RSmisch, K., Webb, J., Herz, J., Prehn, S., Frank, R., Vingron, M., & Dobberstein, B. (1989). Homology of 54K protein of signal-recognition panicle, docking protein and two E. coil proteins with putative GTP-binding domains. Nature 340, 478-482. 103. Rothblatt, J. A., & Meyer, D. I. (1986). Secretion in yeast" reconstitution of the translocation and glycosylation of a-factor and invertase in a homologous cell-free system. Cell 44, 619--628. 103a.Sagara, K., Matsuyama, S., & Mizushima, S. (1994). SecF stabilizes SecD and SecY, components of the protein translocation machinery of the Escherichia coil cytoplasmic membrane. J. Bacteriol. 176, 4111-4 116. 104. Schatz, E J., Bleker, K. L., Ottemann, K. M., Silhavy, T. J., & Beckwith, J. (1991). One of the three transrnembrane stretches is sufficient for the functioning of the SecE protein, a membrane component of the E. coil secretion machinery. EMBO J. 10, 1749-1757. 105. Schatz, P. J., Riggs, P. D., Jacq, A., Fath, M. J., & Beckwith, J. (1989). The secE gene encodes an integral membrane protein required for protein export in Escherichia coll. Genes Dee 3, 1035-1044. 106. Schiebel, E., Driessen, A. J. M., Hartl, E-U., & Wickner, W. (1991). z~ttH+ and ATP function at different steps of the catalytic cycle of preprotein translocase. Cell 64, 927-939. 107. Schiebel, E., & Wickner, W. (1992). Preprotein translocation creates a halide anion permeability in the Escherichio coli plasma membrane. ). Biol. Chem. 267, 7505-7510. 108. Schmidt, M. G., Roilo, E. E., Grodberg, J., & Oliver, D. B. (1988). Nucleotide sequence of the secA gene and secA(Ts) mutations preventing protein export in Escherichia coll. J. Bacteriol. 170, 3404-3414. 109. Shun, L. M., Lee, J.-l., Cheng, S., JuRe, H., Kuhn, A., & Dalbey, R. E. ( 1991). Use of site-directed mutagenesis to define the limits of sequence variation tolerated for processing of the M 13 procoat protein by the Escherichio coil leader peptidase. Biochemistry 30, 11775-I 1781. 110. Shinkai, A., Lu, H.-M., Tokuda, H., & Mizushima, S. (1991). The conformation of SecA, as revealed by its protease sensitivity, is altered upon interaction with ATP, presecretory proteins, everted membrane vesicles, and phospholipids. J. Biol. Chem. 266, 5827-5833. 111. Shiozuka, K., Tani, K., Mizushima, S., & Tokuda, H. (1990). The proton motive force lowers the level of ATP required for in vitro translocation of secretory proteins in Escherichia coll. J. Biol. Chem. 265, 18843-18847. i 12. Simon, S. M., Blobel, G. (1992). Signal peptides open protein-conducting channels in E. coll. Cell 69, 677-684. 113. Stader, J., Gansheroff, L. J., & Silhavy, T. J. (1989). New suppressors of signal-sequence mutations, pdG, are linked tightly to the secE gene of Escherichia coll. Genes Dee 3,1 045-1052. 114. Stirling, C. J., Rothblatt, J., Hosobuchi, M., Deshaies, R., & Schekman, R. (1992). Protein translocation mutants defective in the insertion of integral membrane proteins into the endoplasmic reticulum. Mol. Biol. Cell 3, 129-142. 115. Sung, M., & Dalbey, R. E. (1992). Identification of protential active-site residues in the Escherichia coil leader peptidase. J. Biol. Chem. 267, 13154-13159.

84

SHIN-ICHI MATSUYAMAand SHOJl MIZUSHIMA

116. Suzuki, T., Itoh, A., lchihara, S., & Mizushima, S. (1987). Characterization of the sppA gene coding for protease IV, a signal peptide peptidase of Escherichia coli. J. Bacterial. 169, 2523-2528. 117. Talmrdge, K., Stahl, S., & Gilbert, W. (1980). Eukaryotic signal sequence transports insulin antigen in Escherichia coil Proc. Natl. Acad. Sci. USA 77, 3369-3373. 118. Tani, K., Shiozuka, K., Tokuda, H., & Miznshima, S. (1989). In vitro analysis of the process of transiocation of OmpA across the Escherichia coli cytoplasmic membrane. A translocation intermediate accumulates transiently in the absence of the proton motive force. J. Biol. Chem. 264, 18582-18588. 119. Tani, K., Tokuda, H., & Mizushima, S. (1990). Translocation of ProOmpA possessing an intramolecular disulfide bridge into memlmme vesicles of Escherichia coll. Effect of membrane energization. J. Biol. Chem. 265, 17341-17347. 120. Tokuda, H., Akimaru, J., Matsuyama, S., Nishiyama, K., & Mizushima, S. (1991). Purification of SecE and reconstitution of SecE-dependent protein translocation activity. FEBS Left. 279, 233-236. 121. Ueguchi, C., & Ira, K. (1990). Escherichia call sec mutants accumulate a processed immature form of maltose-binding protein (MBP), a late phase intermediate in MBP export. J. Bacterial. 172, 5643-5649. 122. van der Walk, J., Klose, M., Breukink, E., Demei, R. A., de Kmijff, B., Freudl, R., & Driessen, A. J. M. (1993). Characterization of a Bacillus subtilis SecA mutant protein deficient in translocation ATPase and release from the membrane. Mol. Microbial. 8, 31-42. 123. van Heijne, G. (1985). Structural and thermodynamic aspects of the transfer of proteins into and across membranes. Curt. Top. Memb. Transp. 24, 15 I- 179. 124. Walter, P., & 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. 125. Walter, O., & Blobel, G. (1982). Signal recognition particle contains a 7SL RNA essential for protein translocation across the endoplasmic reticulum. Nature 299, 691-698. 126. Watanabe, M., & Blobel, G. (1989). SecB functions as a cytosolic signal recognition factor for protein export in E. coil Cell 58, 695-705. 127. Watanabe, M., & Blobel, G. (1989). Cytosolic factor purified from Escheridtia coil is necessary and sufficient for the export of a preprotein and is a heterotetramer of SecB. Proc. Natl. Acad. Sci. USA 86, 2728-2732. 128. Waters, M. G., & 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. 129. Weiss, J. B., Ray, P. H., & Bassford, E J. Jr. (1988). Purified SecB protein of Escherichia coil retards folding and Wcmmtes membrane translocation of the maltose-binding protein in vitro. Proc. Natl. Acad. Sci. USA 85, 8978-8982. 130. Wolfe, P. B., Silver, P., & Wickner, W. (1982). The isolation of homogeneous leader peptidase from a strain of Escherichia call which overproduces the enzyme. J. Biol. Chem. 257, 7879-7902. 131. Yamada, H., Matsuyama, S., Tokuda, H., & Mizushima, S. (1989). A high concentration of SecA allows proton motive force-independent translocation of a model secretory protein into Escherichia coil membrane vesicles. J. Biol. Chem. 264, 18577-18581. 132. Yamane, K., Ichihara, S., & Mizushima, S. (1989). In vitro translocation of protein across Escherichia coil membrane vesicles requires both the proton motive force and ATP.J. Biol. Chem. 262, 2358--2362. 133. Yamane, K., Matsuyama, S., & Mizushima, S. (1988). Efficiem in vitro translocation into Escherichia coil membrane vesicles of a protein carrying an uncleavable signal peplide. J. Biol. Chem. 263, 5368-5372.

PIGMENT-PROTEIN COMPLEX ASSEMBLY IN RHODOBACTER $PHAEROIDES AND RHODOBACTER CAPSULATUS

Amy R. Varga and Samuel Kaplan

I. II.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Reaction Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Composition and Structure . . . . . . . . . . . . . . . . . . . . . . . . . B. Role of the RC-H Subunit . . . . . . . . . . . . . . . . . . . . . . . . . . C. Role of Bacteriochlorophyll . . . . . . . . . . . . . . . . . . . . . . . . . D. Other Gene Products Affecting RC Assembly . . . . . . . . . . . . . . . III. The LHI Complex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Composition and Structure . . . . . . . . . . . . . . . . . . . . . . . B. Membrane Insertion and Assembly . . . . . . . . . . . . . . . . . . . . . C. Other Gene Products Affecting LHI Assembly . . . . . . . . . . . . . . . IV. The LHII Complex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Composition and Assembly . . . . . . . . . . . . . . . . . . . . . . . . . B. Other Gene Products Affecting LHII Assembly . . . . . . . . . . . . . .

.L

,

.

.

.

.

Advances in Cell and Molecular Biology of Membranes and OrganeIles Volume 4, pages 85-104. Copyright 9 1995 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-924-9 85

86 87 87 88 89 91 93 . . 93 93 94 96 96 97

AMY R. VARGA and SAMUEL KAPLAN

86

V. The Assembly Model for Pigment-Protein Complex.Abundance and R a t i o . . . VI. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . ". . . . . . . . . . . . . . . . . . . . . . . .

97 98 99 99

I. I N T R O D U C T I O N The anoxygenic photosynthetic bacteria provide a powerful system for the study of membrane biogenesis, protein targeting, and membrane protein complex assembly in prokaryotes. The primary factor in the utility of these organisms lies in the existence of an inducible membrane system of unique composition, termed the intracytoplasmic membrane (ICM). The synthesis of this membrane system is controlled by growth conditions, primarily oxygen tension and light intensity. Under aerobic growth conditions, the synthesis of the ICM is repressed and the cells exhibit a typical Gram-negative bacterial cell structure (Figure 1). Under

.' ., .

,.~,,~

.

-"

..~ ~ .

.

.

.

,,

,,

,~.~,

~

................ ."

D +-,+, '

.

,,

~

.

~,

+,

.

~2~.

99

/

+

,,~..

~, .. ~ ' . .

,~

. . . .

. ...., +. ;:,! ~::,"~:: f+,.,+.::,,.~+. ...... .~,,,.+~++ ............. ~ 9CM . ~ .' :. "," '-"~+-:~:~-~'---~

, +++, :

.

g .+'~'

+++~;~ . ,++,+,, ,:.+..++++~.~

,

"+~.,';,':,~!" . ~ ',".

~, ,,.,,..+, .~++7;++,,., -.+++ r',.,, + ++ ~ ~ .

.

"

++~+::+~ ,

.',.

~.."

., '++,.:~.~'+'.:

"

.

+,

++~+.~; ,..

Figure 1. Thin-section electron micrographs of R. sphaeroides. Panel A shows a cell grown under aerobic conditions. The outer membrane (OM) and cell membrane (CM) are clearly seen. Bar = 0.5 pm. Panel B shows a cell grown under photosynthetic conditions. The intracytoplasmic membrane (ICM) vesicles can be seen in this cell. Arrowheads indicate points at which ICM vesicles are connected to the CM. Bar = 0.5 pm.

Assemblyof Pigment-Protein Complexes

87

anaerobic growth conditions, the synthesis of the ICM is induced, forming by invagination from the cell membrane (Figure 1). The ICM remains attached to the cell membrane, but the composition of the ICM is distinct from the cell membrane in that the ICM houses the pigment-protein complexes and electron transport chain components involved in anoxygenic photosynthesis. The pigment-protein complexes that carry out light-harvesting and primary photochemistry are the photochemical reaction center (RC), the primary light-harvesting complex (LHI or B875), and the accessory light-harvesting complex (LHII or B800-850), and they are found exclusively in the ICM. The most intensely studied species within this group of bacteria in recent years have been Rhodobacter (R) sphaeroides and R. capsulatus, and this review will focus on the assembly of pigment-protein complexes in these two species (see 24, 25, 37, 41, 43, 59; for reviews on ICM structure and biogenesis).

II. THE REACTION CENTER A. Compositionand Structure The photochemical reaction center (RC) absorbs light energy directly or via the light-harvesting complexes, performs primary charge separation, and initiates photosynthetic electron transport (51). The RC is a highly ordered integral membrane complex consisting of three polypeptides and 10 noncovalently bound cofactors (51). The RC polypeptides are designated H, M, and Ldue to their relative mobilities in sodium dodecylsulfate polyacrylamide gels (49, 50). The genes for the RC-L and -M polypeptides reside in a multicomponent operon termed the puf operon, and these genes have been cloned and sequenced in both R. sphaeroides and R. capsulatus (74, 75, 77). The third polypeptide of the RC, RC-H, is encoded by an unlinked operon, puhA, that has also been cloned and sequenced (23, 77). The RC-L and -M polypeptides are very hydrophobic, and each has five transmembrane helices, whereas the RC-H subunit has only a single amino-terminal membrane helix and a large cytoplasmic domain (4, 73). There is a high degree of amino acid homology between the R. sphaeroides and R. capsulatus RC-L (78%, 73), RC-M (77%, 73, 77), and RC-H subunits (64%, 73), indicating the close evolutionary relationship between these species. The transcriptional expression of the RC polypeptide-encoding genes is controlled by oxygen (14, 80, 81). All of the cofactors comprising four bacteriochlorophylls (Bchl), two bacteriophephytins (Bphe), two quinones, one carotenoid, and one non-heine iron atom are noncovalently bound by the RC-L and -M subunits (29, 49), and the primary charge separation is carded out by a Bchl dimer referred to as the "special pair" (51). Cross-linking studies have shown that the protein components of the RC are closely associated with H-M-L, H-M, H-L, and M-L cross-linked species detected (34).

88

AMY R. VARGA and SAMUEL KAPLAN

Studies on the RC have been greatly facilitated by the availability of the crystal structure of the RC complex from R. sphaeroides (3, 4) and a related species, Rhodopseudomonas viridis (21, 22). The detailed structural determinations now available have confirmed the positions of the cofactors relative to the RC-L and -M subunits and indicate that the RC-H subunit does not bind any of the cofactors (5, 76).

B. Role of the RC-H Subunit The role of the RC-H polypeptide in RC structure and function has not been well defined. RC-H does not bind RC cofactors and, in vitro, the RC-H subunit can be removed from the isolated RC, yielding the L-M complex termed RC~ (29), with little attendant loss in photochemical efficiency (2, 18), although RC-H may participate in the formation of the quinone-binding pocket and/or protonation of quinones (3, 18). Regions of RC-H exhibiting conserved amino acid sequence homology among R. sphaeroides, R. capsulatus, and R. viridis subunits (73) (residues 30-42 on the cytoplasmic side of the transmembrane helix and 229-244 near the carboxyl terminus) appear to be involved in the formation of salt bridges between RC-H and -L and RC-H and -M, as well as within the RC-H subunit (4, 11), which may be needed to stabilize the RC complex. Deletion of the carboxyterminal 26 residues of the R. sphaeroides RC-H subunit results in a photosynthetically incompetent phenotype (Lee and Kaplan, unpublished results), indicating that this region may be important in formation of the RC. The RC-H subunit differs from many other integral membrane proteins in that it does not follow the "inside positive" rule as proposed by von Heijne (72), who observed that most transmembrane helices are oriented with the most positively charged amino acids toward the cytoplasmic side of the membrane. The single amino-terminal transmembrane helix of the RC-H polypeptide has several negatively charged amino acids in the conserved region on the cytoplasmic side of the helix (73). This group of negatively charged residues does not appear to be critical for the proper membrane orientation of the helix since they can be deleted, and the helix orientation is maintained (12). In fact, the orientation of the RC-H transmembrane helix can be reversed by inserting positively charged amino acids at the amino terminus (12). The RC-H subunit is detected in the membrane of aerobic R. sphaeroides cells and is found in molar excess in membranes of photosynthetic cells (13). Recently, the synthesis of RC-H and RC-M was examined and RC-H was synthesized in excess relative to RC-M (and -L), resulting in a 3:1 ratio of H:M (71). When the ratio of RC-H to RC-M is disrupted by a mutation that affects the oxygen-regulation of the expression of the RC polypeptides, yielding a RC-H to -M ratio of 1:1.4, loss of excess RC-M is clearly seen (71), indicating that the normal excess of RC-H is important in RC assembly. The amount of stable RC-M, that is, RC-M incorporated into RCs, is most likely limited by the amount of available RC-H polypeptide, and when RC-H is in limiting amounts, the normal productive association between RC-M (and -L) and RC-H is disrupted and excess RC-M is turned over (71).

Assemblyof Pigment-ProteinComplexes

89

An R. sphaeroides mutant lacking the RC-H subunit is photosynthetically incompetent (58). This mutant, termed PUHAI, synthesizes wild-type levels of mRNA encoding the RC-L and -M subunits, abundant Bchl, LHII complexes, and ICM (58), and thus may assemble RC* complexes. In fact, RC* complexes are detected in PUHAI membranes but at levels at least 150-fold lower than RC complexes in wild-type membranes and in amounts insufficient to sustain photosynthetic growth (36). These data on PUHA1 raised the questions of whether RC-H is required for efficient translation of puf mRNA encoding RC-L and -M and whether RC-H is needed for efficient targeting to or assembly of RC-L and -M with cofactors in the ICM. Recent studies indicate that RC-M, and presumably RC-L, is synthesized at essentially wild-type levels in the RC-H-deficient mutant and that the RC-M polypeptide is efficiently targeted to the membranes of PUHA1 (71). The major difference in the behavior of the cofactor-binding RC-M polypeptide in the absence of the RC-H polypeptide is its stability. In PUHAI, RC-M is turned over rapidly with a hf2 of about 30 rain once it reaches the membrane, regardless of the presence of Bchl and ICM (Figure 2), whereas, in wild-type R. sphaeroides 2.4.1. the RC-M subunit in the RC is very stable with a tit2 of greater than 4 hours (71). Conversely, the RC-H polypeptide is not stable in the absence of the RC-L and -M polypeptides, regardless of the presence of Bchl and ICM, and shows a tlf2 of only 6-8 min (71), indicating that specific interactions among all of the RC polypeptides are required for assembly and stability of the RC complex in vivo. Photosynthetic competence of PUHA l can be restored by supplying the puhA gene in trans (58). This complementation also restores the stability of the RC-M and -H polypeptides in photosynthetic cells (Figure 2), although both polypeptides are unstable under aerobic growth conditions as is the case in wild-type cells (Figure 2) (71). The requirement for RC-H in the RC complex appears to be at the level of efficient assembly, promoting interactions between RC polypeptides and cofactors to form the functional complex.

C. Role of Bacteriochlorophyll The oxygen-regulated synthesis of Bchl was observed over 30 years ago (16), when it was demonstrated that high levels of oxygen repressed synthesis of Bchl. The synthesis of Bchl-binding polypeptides is coupled to Bchl synthesis (9, 13, 38, 65), so it was assumed that Bchl-binding proteins were not synthesized in the absence of Bchl synthesis. More recent studies have shown that mRNA encoding Bchl-binding proteins is present in aerobic cells (25, 40, 80, 81) and that the RC-H polypeptide is present in aerobic R. sphaeroides cell membranes (13). Utilizing sensitive radiolabeling techniques, the presence of additional RC polypeptides has been noted in the absence of Bchl, either under highly aerobic conditions in R. sphaeroides (71) or in Bchl- mutants of both R. sphaemides (65) and R. capsulatus (20). The common feature in all of these studies is that RC

AMY R. VARGA and SAMUEL KAPLAN

90

9

.! ' .

.....

aD

QD am

oB

mD

..

. . . . . . . . 9

C

D

E

gmb allll~ a ~

~m, - , ,

.............

4NINIi~ qlllll~ tlNIII~ tUlIIID 41nlllnl ~gllllll~ ~

luonD ~

amllD qnul~ imnllD 41NIP qUlil~ qllNl~ aunmb qmllnlb a 0

2

5

10

20

.

.

30

~0

gO

'~20

4mlul, ~mnnn~ '1.50

180

Chase lime (rain)

Figure 2. Autoradiographs of immunoprecipitated RC-M and -H polypeptides from pulse-chase 3SS-methionine labeled R. sphaeroides wild-type and mutant cells. (A) RC-M immunoprecipitated from PUHA1 grown under anaerobic/dimethyl sulfoxide conditions to induce Bchl synthesisand ICM formation. RC-M is not stable under these conditions. (B) RC-M immunoprecipitated from wild-type cells grown under photosynthetic conditions to induce Bchl synthesisand ICM formation. RC-M is stable under these conditions. (C) RC-H immunoprecipitated from PUHA1 containing the puhA gene in trans grown aerobically so Bchl and ICM synthesis are repressed. RC-H is not stable under these conditions. (D) RC-H immunoprecipitated from Tla cells grown aerobically. This mutant strain produces Bchl under these conditions, and RC-H is stable. (E) RC-H immunoprecipitated from PUHA1 containing the puhA gene in trans and grown under photosynthetic conditions so Bchl and ICM synthesis are induced. RC-H is stable under these conditions. polypeptides are synthesized but are very unstable in the absence of Bchl; RC-H exhibits a tl~ of 5 rain (Figure 2) and RC-M a tn~ of about 30 rain in aerobic R. sphaeroides cells (71). Under photosynthetic growth conditions the turnover of the polypeptides is measured in hours (Figure 2). In contrast, mutant strains that synthesize Bchl under aerobic conditions show significant levels of RC polypep-

Assemblyof Pigment-Protein Complexes

91

tides (strain TA-R; 65), and these polypeptides are stable (Figure 2) (strain Tla; 71), indicating that Bchl binding in the RC complex is critical for complex stability. Studies utilizing site-specific mutagenesis of Bchl-binding sites in the RC of R. capsulatus have shown that significant alterations can be made in the PC, such as substitution of Bphe for one Bchl of the Bchl2 special pair, without severely affecting the cells' ability to grow photosynthetically (10). In most cases the RC is still found in the membrane, with the exception of an RC-Lglul~ mutation (10), although it was not determined if the RC-L polypeptide was inserted into the membrane or if this polypeptide or the RC complex as a whole was unstable due to the mutation.

D. Other Gene Products Affecting RC Assembly The assembly of a structure as complex as the RC might be expected to involve assembly proteins and chaperonins, so regions genetically linked to the pufoperon have been examined to determine their possible roles in RC assembly. The Q open reading frame lies immediately upstream of the pufgenes encoding the LHI structural polypeptides (pufBA) that precede the genes for the RC-L and -M polypeptides (pujLM) (26, 77). Mutants with deletions or interruptions in orfQ show significantly lower levels of Bchl biosynthesis (17, 79). Studies in R. capsulatus indicate that orfQ is part of the puf operon and have shown that the absence of the pufQ gene product does not affect transcription of Bchl biosynthetic enzymes genes (7) nor is it absolutely required for Bchl biosynthesis (1, 52). These results suggest that PufQ may be a membrane-bound Bchl binding protein or catalytically involved in Bchl-protein interactions. There is significant amino acid homology between PufQ and the quinone-binding regions of the RC-L and -M polypeptides (7), implying that PufQ may be involved in redox sensing in the cell and coordinating Bchl with the RC and LHI complexes (7, 52). Site-specific mutagenesis studies in R. sphaeroides indicate that the orfQ gene product is possibly involved in assembly of pigment-protein complexes. A mutation removing the start codon for orfQ yields a photosynthesis-minus phenotype and very reduced levels of Bchl (33). A number of amino acid substitutions in the regions of OrfQ which show homology to RC-L and -M, exhibit specific variations in levels of light-harvesting complexes, although all remain photosynthetically competent. One such mutant, QalaT~ does not contain detectable LHI complex and shows an eight-fold increase in puc-encoded LHII complexes, although no change in puc-specific mRNA is seen (33). These data have led to speculation on the role of OffQ in the heirarchy of pigment-protein assembly, with priority order being RC>LHI>LHII. OrfQ, in conjunction with complex-specific assembly factors (33), may play a role in Bchl binding in pigment-protein complexes. The role of orfQ in light-harvesting complex formation is discussed below. The puff( open reading frame lies immediately downstream of the final structural gene in the pufoperon (pufM) (26, 77). A deletion ofpufX in R. sphaeroides yields

92

AMY R. VARGA and SAMUEL KAPLAN

a photosynthetically incompetent phenotype, although puf transcription is not blocked and RC polypeptides are synthesized and assembled with cofactors into photobleachable RC complexes (28). Although the PufX polypeptide appears to copurify with the RC-LHI photosynthetic unit, it is not proposed to have a chaperonin function. Rather, the PufX-phenotype is attributed to a block in cyclic electron transfer resulting in impaired generation oftransmembrane potential (27). It appears, therefore, that PufX is not involved in RC assembly but that OrfQ may be involved. More detailed studies of these and other protein-protein interactions, such as those involving RC-H, as well as critical protein-cofactor interactions, will help to determine the order of events involved in RC assembly. In light of the fact that we have demonstrated the presence of the RC-H and RC-M polypeptides in the membranes of aerobically growing R. sphaeroides, albeit with shortened half-lives (71), we would extrapolate and suggest that all of the apopolypeptides corresponding to the three pigment-protein complexes are synthesized under aerobic conditions. As in the case of RC-H and RC-M, the cognate mRNA species for all pigment-protein structural polypeptides are present at low levels in aerobic cells (14, 39, 80), and thus the cell is "primed" to undergo rapid ICM synthesis and photosynthetic growth under appropriate conditions. We also know from studies of the oxygen-regulatory mutant T la (45), which makes RCs and LHII complexes and presumably ICM under highly aerobic growth conditions, that there is no intrinsic block to aerobic assembly of pigment-protein complexes (45, 71). In wild-type R. sphaeroides no Bchl is produced under aerobic conditions despite the presence of low levels of mRNA for Bchl biosynthetic enzymes (33). However, the orfQ gene-specific mRNA is not detected in aerobic cells. Therefore, we envision that the orfQ gene product in its role as an assembly factor is critical to the entire process and in its absence, at the posttranslational level, Bchl would not be inserted into each "pro-complex," and thus polypeptide components would turn over rapidly. Furthernmre, it is possible that the absence of OrfQ and the ability to insert Bchl into nascent complexes leads to an inhibition of the Bchi branch of the tetrapyrrole biosynthetic pathway (l 6). Thus, the analogy can be drawn between the anoxygenic photosynthetic bacteria and green plant chloroplast development. In the latter case, several of the protein components for photosynthetic membrane complexes are present in dark-grown, nonpigmented proplastids, but turn over rapidly in the absence of chlorophyll (Chl) (47). When exposed to light, the final steps in chlorophyll biosynthesis take place and pigment-protein complexes are assembled and remain stable in chloroplast thylakoid membranes (47). Therefore, to extend this analogy, it is not unreasonable to suggest the existence of key anaerobically.induced complex-specific assembly factors for bacteria and light-induced factors for green plants that serve to insert (B)Chl into the developing pigment-protein complexes.

Assemblyof Pigment-ProteinComplexes

93

III. THE LHI COMPLEX A. Composition and Structure The primary light-harvesting complex, LHI or B875, is in close association with the RC and transfers absorbed light energy to the RC (reviewed in 60). LHI is an integral membrane pigment-protein complex, and the minimal unit consists of oligomers of two low molecular weight polypeptides, termed o~and 13(63, 64, 66, 67), noncovalently bound with two Bchl and two carotenoid pigment cofactors (8). The genes encoding the LHI 0~and 13polypeptides are encoded by the pufoperon, proximal to the genes for the RC L and M polypeptides (39, 77) and are similarly regulated by oxygen at the transcriptional level (14, 81). The LHI complex has not yielded sufficiently ordered crystals for analysis, so the structural details analagous to the RC are not yet available for this pigment-protein complex.

B. Membrane Insertion and Assembly The membrane insertion of the o~ and 13polypeptides and assembly of the LHI complex have been the subjects of several studies in R. capsulatus. In vitro studies have shown that cotranslational membrane association of the LHI a and 13polypeptides is more efficient than post-translational association, suggesting that synthesis of these polypeptides occurs on membrane-bound ribosomes (70). It has been observed that the amino.-temfinal regions of LHI r and 13appear to be involved with membrane insertion of the polypeptides. The amino-termini of LHI r and 13are located on the cytoplasmic side of the ICM (62) and are oppositely charged (63, 64), which is an evolutionarily conserved feature of the LHI polypeptides (Figure 3) (61, 83). The LHI ~ polypeptide has several conserved features that appear to be involved in the proper membrane insertion and interaction of the LHI r and 13polypeptides. The amino terminus of LHI ~ has a structure typical of an uncleaved signal sequence of bacterial membrane proteins (56); typical characteristics include positively charged residues, a short hydrophobic region, and a helix-breaking residue adjacent to the membrane-spanning region of the protein (63). The aromatic residue Trp-8 is highly conserved in LHI o~polypeptides (Figure 3) and appears to be important in insertion of the LHI a polypeptide into the membrane (54). The helix-breaking Pro-13 residue is also highly conserved (Figure 3) and may be required for keeping the LHI 13 partner polypeptide in the membrane (54). The deletion of the charged residues Arg-I 4 and Arg- 15 leads to LHI oc being undetectable in the membranes, as does the deletion of the conserved Trp-8 plus the preceding two residues (53). Alteration of the hydrophobic stretch of amino acids LHI r 7-11 by deletion or insertion of amino acids had some effect on LHI stability, but had a more severe effect on the interaction between LHI o~and [3 such that efficient o~-[3formation was prevented, leading to rapid degradation of LHI [$ and lack of LHI spectral complexes (53). Deletion of the highly conserved LHIo~

94 R.c. LHI ~: R.s. LHI ~:

AMY R. VARGA and SAMUEL KAPLAN I I ADKNDLSFTGLTDEQAQELI~VYMSGLSAFIAVAVL/~LAVMIWRPWF

ADKSDLGYTGLTDEQAQELHSVYMSGLWPFSAVAVLAHLAVYIWRPWF

R.c. LHII~:MTDDK--AGPSGLSLKEAEEIIISYLIDGTRVFGAMALVAHILSAIATPWLG

R.s. LHII~: R.c. LHI ~

R.s. LHI a

R.c. LHIIa

R.s. LHIIQ

DDLNKVWPSGLTVAEAEEVHKQLILGTRVFGGMALIAHFLAAAATPWLG l I MSKFYKIWLVFDPRRVFVAQGVFLFLLAVLIHLILLSTPAFNWLTVATAHGYVAAAQ

MSKFYKINMIFDPRRVFVAQGVFLFLLAVMIHLILLSTPAFNWLEISAAKYNRVAVAE

MNNA-KIWTVVKIPSTGIPLILGAVAVAALIVliAGLLTNTT--WFANYWNGNPMATVVAVAPAQ MTN-GKIWLWKIPTVGVPLFLSAAFIASWILqAVLTTTT--WL-PAYYQGSAAVAAE

F~gure 3. Amino acid sequences of LHI and LHII (x and I~ polypeptides from R. capsulatus (R.c.) and R. sphaeroides (R.s.). Sequences are aligned to show highly

conserved amino acids (boldface type).

Asp- 12, Pro- 13 resulted in LHI a and [$being inserted into the membrane, however, [5 turns over rapidly yielding a LHI-phenotype. The assembly of LHI a and [~ in the membrane appears to depend upon specific interaction and contacts between the amino-termini, particularly LHIaPro-13, which is needed to fix LHI J5 in the membrane; changes in the relative positions of these critical amino acids, or their deletion, affect the tertiary structure of the polypeptides, and proper interaction between o~and [3 cannot take place (55). Just as with the RC polypeptides, the stability of the LHI polypeptides is interdependent; in the absence of one of the polypeptides, the remaining polypeptide is unstable (55). In a mutant that does not synthesize LHI ~, LHI ~ is incorporated into the membranes but is not stable; likewise, in a LHI [$- mutant, LHI a is hardly detectable in the membranes (55). In wild-type cells, LHI ~ appears in the membranes within 10 s of pulse-labeling with radiolabeled amino acids, followed at 40-50 s by the appearance of LHI u, leading to the speculation that LHI [5 performs a chaperone function for LHI a (55).

C. Other Gene Products Affecting LH! Assembly The region upstream of the puhA operon appears to encode a gene product involved in the assembly of the LHI complex in both R. capsulatus and R. sphaeroides. An interruption of this region in R. capsulatus, an open reading frame termed F1696 (Figure 4) (77), yields a mutant that is photosynthetically competent and has the LHII spectral complex but shows a 67% reduction of the LHI complex, although the pufoperon genes encoding the LHI structural polypeptides are intact and puf mRNA levels are normal (6). In R. sphaeroides a homologous region is found upstream ofpuhA (Figure 4). Deletion of the 3' portion of this open reading frame results in an LHI- phenotype that can be complemented by providing the complete open reading frame in trans (58). Additional R. sphaeroides LHI- mutants, RS 103 and Tla, are also complemented by supplying this puhA upstream region in trans (45, 58). Thus, it was suggested that this region upstream of puhA encodes an assembly factor specific for the LHI complex (58).

Assembly of Pigment-Protein Complexes

95

R.s.PucC: M S R I A E H L V R I G P R F L P F A D A A S D Q L P L R K L L R L S L F Q V A V G M A I V L L V G R.C.PucC: R A F A L K N L A R H A P K Y L P F A D V ~ E E V I P L S R L L R L S L F Q I T V G M T L T L L A G R.c.F1696:MLLSRRMIGSLAMTWLPFADA~ETLPLRQLLRLSLFQVSVGMAQVLLLG

50

R.s.PucC: T L N R V M I V E L K V P A S V V G I M A S L P L L F A P F R A L I G F K S D T H V S A L G W R R R R.c.PucC: T L N R V I 4 Z V E L A ~ A S L L S V M L A M P M L F A ~ F R T L I G F K S D T H K S A L G L R R A R.c.FI696:TLMRVMILELGVPALVVAAMISIPVLVAPFRAILGHRSDTYRSALGWKR-

100

R.s.PucC: V P W I Y R G T L A L W G G F A I M P F A L I V L G G Q G Y A E G Q P F W L G V S S A A L A F ~ H V R.c.PucC: - P W I W K G T I Y Q F G G F A I M P F A L L V L S G F G E S V D A P R W I G M S A A A L A F L L V R.c.FI696:VPYLWFGSLWQMGGLALMPFSLILLS--GDQTMGPAWAGEAFAGVAFLMA

150

R,s.PucC: G G G V M T I Q T V G L A L A T D L A P R E D Q P K V V G L M 2 ~ / V L L I S M I F A S I G F G W L L R.c.PucC: G A G V X I V Q T A G L A L A T D L V A E E D Q P K V V G L M T V M L L F G M V I S A L V Y G A L L R.c.F1696:OVGML~4TQTAGLALAADRATEETRPQWALL~'VMFLIGMGISAVIVGWLL

200

R.s.PucC: D P Y Y D A Q L Z K V I S G V A V A V F F L N M I A L W K M E P - - R N R A F T V K P E K E P E F G D H R.c.PucC: A D Y T P G R L Z Q V I Q G T A L A S W L N M A ~ Q E A V S R D R A R Q M E T A E H P T F K E A R.c.F1696:RDFDQITLZRWQGCGAMTLVLNVIALWKQEVM-RPMTKAEREAPRQSFREA

250

R.s.PucC: W R E F I S R E N A L H G L I V I G L G T L G F G M A D V I L E P Y G G E V L S H T V A E T T R L T R.c.PucC: F G L L M G R P G M L A L L T V I A L G T F G F G M A D V L L E P ~ G G Q A L H L T V G E T T K L T R.c. F 1 6 9 6 : W G L L A A E T G A L R L L A T V M V G T L A F S M O D V L L E P Y G G Q V L G L K V G Q T T W L T

300

R.s.PucC: ATFAGGGLVGFWLASWVLGRGFDPLRM-AIrLGAAAGLPGFFAI-HGATEH-TN R.c.PucC: A L F A L G T L A G F G T A S R V L G N G A R P M R W S A G C T D R V - - P G F V A I : M S S L I S Q D G R.c.FI696:AGWAFGALVGFIWSARRLSQGAVAHRVAARGLLVGI-VAFTAVLFSPLFGSKV

350

R.s.PucC: V W V F L L G T L V V G F G G G L F S H G T L T A T M R L A P K E Q V G L A L G A W G A V Q A T A A R,c.PucC: I W L F L A G T F A V G L G I G L F G H A T L T A T M R T A P A D R I G L A L G A W G A V Q A T A A R.c.F1696:--LFFASAMGIOLGSGMFGIATLTVAMMVVVRGHSGIALGAWGAAQATAA

400

R. s. PucC : G V A I A G A G V L R D - - I L Q A M - P D L S G Y G - - P G A P T V A V F A L E A G F L F L T H I V ! L P L 450 R. c. PucC : G L G V A L A G V V R D - - G L V A L - P G T F G S G - - W G P Y N T V F A I E A L ILIVAIAFAVPL R. c. F 1696 :GLAVF I G G A T R D L V A H A A A - A G Y L G S L H S P A L G Y T W Y V T E IGLLF ITLAVLGPL R. s. F1696 : AGKAGHLGALQD-G IGYSS-GTLE IGLLFATL IVLGPL R. s.PucC : LRSALAARRL. R.c.PucC : LKRGG. R. c. F1696 :VRPGSLFPKKPEAGEARIGLAEFPT. R. s. F1696 :VRTT I LSSERP -AGGTRVGLADFPT.

Figure 4. Amino acid sequences of PucC proteins encoded downstream of LHII a and ~ poly~ptides from R. sphaeroides (R.s.) and R. capsulatus (R.c.) aligned with the open reading frame upstream of puhA designated F1696 of R. capsulatus. The

carboxy-terminal portion of the R. sphaeroides F1696 ORF is also shown. Highly

conserved amino acids are shown in boldface type:

The results of Sockett et al. (58) point to the existence of an LHl-specific assembly factor upstream of puhA in R. sphaeroides, and a role for the OrfQ gene product was noted as well. It was shown that the LHI" phenotype of the PUHA1 mutant was complemented in trans by the orfQ gene, although the RC was not restored because the pultA gene was not present (58). These results, coupled with the observations on the effects of single amino acid replacements within OrfQ on the relative abundance of LHI and LHII complexes (see above and 33), point to a role for the orfQ gene product in LHI assembly.

96

AMY R. VARGAand SAMUELKAPLAN IV. THE LHII C O M P L E X

A. Compositionand Assembly The LHII complex is a secondary light-harvesting complex that passes absorbed light energy to LHI and ultimately to the RC (60). The minimal unit of the LHII consists of oligomers of two low molecular weight polypeptides, termed a and ~, noncovalently bound to three Bchl and one carotenoid (15, 57); the isolated R. capsulatus LHII complex also contains a non-pigment-binding polypeptide, designated LHII ~ (30, 31). The genes for the pigment-binding a and ~ polypeptides are encoded by the pucBA operon and have been cloned and sequenced (40, 78), and the gene for the LHII 6, pucE, has also been cloned and sequenced from R. capsulatus (69). There is significant sequence and structural homology between the LHI and LHII a and [3polypeptides (61, 83), indicating that the origin of LHII may have been via gene duplication of the LHI genes. Expression of the puc operon at the transcriptional level is regulated by both oxygen and light (40, 44, 80). Very little is known about the assembly of the LHII complex at the molecular level. It is known that mutations that eliminate carotenoid biosynthesis lead to an LHII- phenotype (42), but the mechanism is unknown. However, it has led to the hypothesis that, unlike the RC or LHI complexes, the LHII complex requires carotenoids for assembly. Mutations that limit levels of Bchl appear to limit the level of LHII complex, and control occurs at both the level of transcription of the puc operon and assembly of the complex (48), but again the mechanism(s) involved is unknown. There is significant sequence and structural homology between LHI and LHII a and ~) polypeptides (Figure 3) (61, 83). The general structure of the amino-terminus of the LHII a polypeptide is homologous to the LHI a polypeptide (Figure 3) so the same rules governing membrane insertion and assembly of the LHI complex may also be applicable to LHII complex assembly. Wood and Kaplan (manuscript in preparation) used a series of water-soluble benzodiazapines to inhibit pigment-protein complex formation in R. sphaeroides. Fluorazapam concentrations of 10-50 gg/ml were able to increasingly inhibit the formation of the LHII complex, and the complex was virtually absent at the higher concentration. Although there is a three-fold effect upon transcription of the puc operon, it is clear that the drug inhibits assembly of the LHII complex; however, it appears to have little effect on growth rate or LHI levels. Other benzodiazapine derivatives affect both LHI and LHII complexes, with the greater effect seen on LHII complex formation. Given the structure of fluorazapam and related compounds and the fact that their generalized effects appear to be minimal, we suggest that these compounds inhibit or interfere with the action of complex~specific assembly factors. For the LHI complex this interference would minimally be the F1696 gene product upstream of puhA (see above) and for the LHII complex this would be pucC gene product (see below).

Assemblyof Pigment-ProteinComplexes

97

B. Other Gene Products Affecting LHII Assembly The region downstream of the pucBA genes encoding the structural polypeptides of the LHII complex appears to be important in the expression and assembly of the LHII complex in both R. capsulatus and R. sphaeroides (32, 46, 68, 69). An open reading frame immediately downstream of pucA has been designated pucC and encodes a large hydrophobic polypeptide (32, 69), which is highly homologous to the open reading frame F1696 found upstream of puhA and involved in the assembly of the LHI complex (Figure 4) (see above; 6, 69) and which also shows homology to a chlorophyll-binding protein from plants (Moore and Kaplan, unpublished; 24). Deletion of pucC in R. capsulatus shows an LHII- phenotype with a six-fold decrease in transcription from the puc operon, although this transcriptional down-regulation is insufficient to explain the loss of the LHII complex (68). The authors speculate that the pucC gene product participates in the LHII assembly process via a feedback mechanism ensuring that LHII complex synthesis takes place only if all components are present, and the loss of any component of that assembly process, such as PucC, leads to a down regulation in puc expression (68). Additional downstream open reading frames, pucD and pucE (encoding the 8 subunit of the LHII complex), appear to encode gene products that are involved in the stabilization and assembly of the LHII complex, because deletions in these genes lead to destabilization of the LHII complex, in the membrane of R. capsulatus (68). In contrast to R. capsulatus, the R. sphaeroides LHII complex does not contain a 8 subunit, although a low molecular weight polypeptide designated 15A is always absent in LHII- mutants (42, 46); the relationship between this polypeptide and the ~isubunit in R. capsulatus is unknown. The region downstream ofpucBA is required for LHII expression and assembly in R. sphaeroides because when it is interrupted the cells no longer assemble LHII complexes, although the transcript for the puc structural polypeptides is present (46). The R. sphaeroides DNA sequence downstream ofpucBA shows an open reading frame highly homologous to the pucC of R. capsulatus, which is also designated pucC in R. sphaeroides (Figure 4), although no open reading frames homologous to the R. capsulatus pucD and E were found in R. sphaeroides (32). Since the amino acid sequence homology between the R. capsulatus and R. sphaeroides PucC protein is high (56% identity; Figure 4), the function of these gene products in LHII complex assembly is presumed to be similar (32), and perhaps functionally homologous to the F1696 LHI assembly factor found upstream ofpuhA (Figure 4).

V. THE ASSEMBLYMODEL FOR PIGMENT-PROTEIN COMPLEX ABUNDANCE AND RATIO In photosynthetically grown cells, the ratio of LHI to the RC is fixed at approximately 12:1 (82), and this has been referred to as the fixed photosynthetic unit (FPU;

98

AMY R. VARGA and SAMUEL KAPLAN

25). On the other hand, the level of LHII can vary widely with respect to the FPU and is referred to as the variable photosynthetic unit (VPU; 25). Numerous independent experiments have revealed that in photosynthetically growing cells, the levels of mRNA encoding the apoproteins of the pigment-protein complexes is, in general, in excess to the actual amount of complexes present (19, 33, 44, 48). It appears that the cellular levels of their cognate apoproteins are in excess of the level of assembled complexes present, and thus unused apoproteins are turned over (71). Therefore, the availability of Bchl ultimately determines the cellular abundance of the three pigment-protein complexes, and this may be mediated through the OrfQ protein with the priority of RC>LHI>LHII. Given the above observations and the loose coupling between pigment-protein complex abundance and their cognate mRNA species and apoproteins, we can provide a model that helps to explain these varied observations. The striking resemblance between the pucC and F1696 gene products and their resemblance to green plant chlorophyll-binding proteins (24) suggest that these proteins have similar but not identical roles. Each protein has been implicated in the assembly of a specific complex: F1696 for LHI and pucC for LHII. Therefore, it is evident that the orfQ gene product is involved in both LHI and LHII assembly. We will assume that OrfQ is also involved in RC assembly and that one or more RC-specific assembly factors are present. Given the experimental evidence cited, we propose that OrfQ is capable of interacting with each of the complex-specific assembly factors to assist in the assembly process by inserting Bchl into the "pro-complex." We also assume that the affinity of OrfQ for each assembly complex-specific factor represents a hierarchy of levels with the RC factor(s) at the top and LHII at the bottom. Thus, with limiting Bchl (48), the RC complex will be present, followed by LHI and then LHII complexes. Because of the similarities of the complex-specific factors, they may compensate for one another when participants in the assembly sequence are either over- or under-expressed, depending upon their role in the assembly process. Thus, we imagine that the mutant OrfQ proteins described earlier are altered in amino acids residues that affect their affinitY for one or more complex-specific assembly factors. Therefore, the abundance of each complex and the ratio of one complex to another is broadly fixed at the level of transcription, but it is the posttranslational assembly processes that provides the high level of specificity. The OrfQ protein is clearly critical in this process, and its presence or absence may represent a switch between aerobic and anaerobic expression of the pigment-protein complexes and ICM.

VI. SUMMARY The process of the assembly of the pigment-protein complexes in R. sphaeroides and R. capsulatus is only beginning to be revealed. The role of the RC-H polypeptide as a component critical for the efficiency of association of the pigment-binding

Assemblyof Pigment-ProteinComplexes

99

RC-L and-M polypeptides with the RC cofactors has been elucidated. The structural role of Bchi in RC complex stability is also being refined because of the availability of the RC crystal structure and the recently developed techniques of site-specific mutagenesis to alter critical Bchl-binding amino acids. The role of assembly proteins in RC assembly is, at present, established at the genetic level, but the gene products that are encoded by these genetic elements are yet to be characterized, and the order of events remains to be demonstrated. The amino acid similarity of both the LHI and LHII pigment-binding polypeptides and, as recently discovered, their putative assembly factors, indicates the close evolutionary relationship of these two light-harvesting complexes, both at the structural level and at the level of assembly. Likewise, the hypothetical existence of RC-specific assembly factors remains to be demonstrated, but in light of the evidence presented, their existence appears quite plausible. Finally, the orfQgene product and its pivotal role in assembly of pigment-protein complexes and subsequent ICM development is clearly testable, and this system is ideal for future exploitation. The genetics and molecular genetic methodologies are established, the protein and cofactor components are readily available. Thus, researchers intent upon an understanding of membrane complex assembly processes have a wealth of potential at their disposal.

ACKNOWLEDGMENT This work supported by USPHS grant GM15590 to S.K.

REFERENCES 1. Adams, C. W., Forres~ M. E., Cohen, S. N., & Beatty, J. T. (1989). Transcriptional control of the Rhodobacter capsulatus put operon: A structural and functional analysis. J. Bacteriol. 171, 473-482. 2. Agalidis, I., Nuijs, A. M., & Reiss-Husson, E (1987). Characterization of an LM unit purified by affinity chromatography from Rhodobacter sphaeroides reaction centers and interaction with the H subunit. Biochim. Biophys. Acta 890, 242-250. 3. Allen, J. P., Feher, G., Yeates, T. O., Komiya, H., & Rees, D. C. (1987). Structure of the reaction center from Rhodobacter sphaeroides R-26: the cofactors. Proc. Natl. Acad. Sci. USA 84, 5730-5734. 4. Allen, J. P., Feher, G., Yeates, T. O., Komiya, H., & Rees, D. C. (1987). Structure of the reaction center from Rhodobocter sphaeroides R-26: the protein subunits. Proc. Natl. Acad. Sci. USA 84, 6162-6166. 5. Allen, J. P., Feher, G., Yeates, T. O., Komiya, H., & Rees, D. C. (1988). Structure of the reaction center from Rhodobocter sphaeroides R-26: protein-cofactor (quinones and Fe 2§ interactions. Proc. Natl. Acad. Sci. USA 85, 8587-8491. 6. Bauer, C. E., Buggy, J. J., Yang, Z., & Man's, B. L. (1991). The supeml~ronal organization of genes for pigment-protein biosynthesis and reaction center proteins is a conserved feature in Rhodobacter capsulatus: analysis of overlapping bchB and puhA transcripts. Mol. Gen. Genet. 228, 433-444.

100

AMY R. VARGA and SAMUEL KAPLAN

7. Bauer, C. E., & Marrs, B. L. (1988). Rhodobacter capsulatus puf operon encodes a regulatory protein (PufQ) for bacteriochlorphyll biosynthesis. Proc. Natl. Acad. Sci USA 85, 7074-7078. 8. Broglie, R. M., Hunter, C. N., Delepelaire, P., Niederman" R. A., Chua, N.-H., & Clayton, R. K. (1980). Isolation and characterization of the pigment-protein complexes of Rhodopseudomonas sphaemides by lithium d~lsulfate/polyacrylamide gel electrophoresis. Proc. Natl. Acad. Sci. USA 77, 87-91. 9. Brown, A. E., Eiserling, E A., & Lascelles, J. (1972). Bacteriochlorcq~hyll synthesis and the ultrastructure of wild type and mutant strains of Rhodopseudomonas sphaeroides. Plant Physiol. 50, 743-746. 10. Bylina, E. J, & Youvan, D. C. (1988). Directed mutations affecting spectroscopic and electron transfer im31mties of the primary donor in the photosynthetic reaction center. Proc. Natl. Acad. Sci. USA 85, 7226-7230. 11. Chang, C.-H., EI-Kabbani, O., Tiede, D., Norris, J., & Schiffer, M. (1991). Structure of the membrane-bound protein photosynthetic reaction center from Rhodobacter splmemides. Biochemistry 30, 5352-5360. 12. Cheng, E (1992). Ph.D. Thesis. The Ohio State University. 13. Chary, J., Donohue, T. J., Varga, A. R., Slaehelin, L. A., & Kaplan, S. (1984). Induction of the photosynthetic membranes of Rhodopseudomonas sphaeroides: biochemical and morphological studies. J. Bacterial. 159, 540-554. 14. Clark, W. G., Davidsm, E., & Marts, B. L. (1984). Variation of levels of mRNA coding for antenna and reaction center polypeptides in Rhodopseudomonas capsulata in response to changes in oxygen concentration. J. Bacterial. 157, 945--948. 15. Cogdell, R. J., & Crofts, A. R. (1978). Analysis of the pigment-protein content of an antenna pigment-protein complex from three strains of Rhodopseudomonas sphaeroides. Biochim. Biophys. Acta 502, 409--416. 16. Cohen-Bazire, G., Sistrmn, W. R., & Stanier, R. Y. (1957). Kinetic studies of pigment synthesis by non-sulfur purple bacteria. J. Cell. Camp. Physiol. 49, 25--68. 17. Davis, J., Donohue, T. J., & Kaplan" S. (1988). Construction, characterization, and complemenration of a Puf- mutant of Rhodobacter sphaeroides. J. Bacterial. 170, 320-329. 18. Debus, R. J., Feher, G., & Okamura, M. Y. (1985). LM complex of reaction centers from Rhodopseudomonas sphaemides R-26: characterization and reconstitution with the H subunit. Biochemistry 24, 2488-2500. 19. DeHoff, B. S., Lee, J. K., Donohue, T. J., Gumport, R. I., & Kaplan, S. (1988). In viva analysis of puf operon expression in Rhodobacter sphaemides after deletion of a putative intercistronic transcription terminator. J. Bacterial. 170, 4681--4692. 20. Dierstein" R. (1983). Biosynthesis of pigment-protein complex polypeptides in bacteriochlorophyll-less mutant cells of Rhodopseudomonas capsulata YS. FEBS Left. 160, 281-286. 21. Diesenhofer, J., Epp, O., Miki, K., Huher, R., & Michel, H. (1984). X-ray structure analysis of a membrane protein complex. Electron density map at 3A resolution and a model of the chrmnophores of the photosynthetic reaction center from Rhodopseudomonas viridis. J. Mol. Biol. 180, 385-398. 20 Diesenhofer, J., Epp, O., Miki, K., Huber, R., & Michel. H. (1985). Structure of the protein subunits if the photosynthetic reaction centre of Rhodopseudomonas viridis at 3~ resolution. Nature (London) 318, 19--26. 23. Donohue, T. J., Hoger, J. H., & Kaplan, S. (1986). Cloning and expression of the Rhodobacter sphaemides reaction center H gene. J. Bacterial. 168, 953-961. 24. Donohue, T. J., & Kaplan, S. (1993). Genetic analysis of photosynthetic membrane biogenesis in Rhodobacter sphaeroides. In J. Diesenhofer (Ed.) The photosynthetic reaction center (pp. 101-131 ). New York: Academic Press 25. Donohue, T. J., & Kaplan, S. (1986). Synthesis and assembly of bacterial photosynthetic membranes. In L. A. Staehelin & C. J. Amtzen (Eds.). Encyclopedia of plant physiology new

Assembly of Pigment-Protein Complexes

26. 27.

28.

29.

30.

31. 32.

33. 34. 35.

36. 37. 38. 39.

40. 41. 42. 43. 44.

101

Series (vol. 19). Photosynthesis IIL Photosynthetic membranes and light-harvesting systems (pp. 632-639). Berlin: Springer-Verlag Donohue, T. J., Kiley, P. J., & Kaplan, S. (1988). The puf operon region of Rhodobacter sphaeroides. Photosynth. Res. 19, 39--61. Farchaus, J. W., Barz, W. P., Grunberg, H., & Oesterhelt, D. (1992). Studies on the expression of the pufX polypeptide and its requirement for phozoheterotrophic growth in Rhodobacter sphaeroides. EMBO J. 1I, 2779-2788. Farchaus, J. W., Gruenberg, H., & Oesterhelt, D. (1990). Complementation of a reaction center-deficient Rhodobacter sphaeroides pufLMX deletion strain in trans with pufBALM does not restore the photosythesis-positive phenotypr J. Bacteriol. 172, 977-985. Feller, G., & Okamura, M. Y. (1978). Chemical composition and properties of reaction centers. In R. K. Clayton & W. R. Sistrom (Eds.) The Photosynthetic bacteria (pp. 349-386). New York: Plenum Press. Feick, R., & Drews, G. (1978). Isolation and characterization of light-harvesting bacteriochlorophyll-protein complexes from Rhodopseudomonas capsulata. Biochim. Biophys. Acta 501, 499-513. Feick, R., & Drews, G. (1979). Protein subunits of bacteriochlorophylls B802 and B855 of the light-harvesting complex II of Rhodopseudomonas capsulata. Z. Naturforsch. (C) 34, 196-199. Gibson, L. C. D., McGlynn, P., Chaudhri, M., & Hunter, C. N. (1992). A putative anaerobic coproporphyrinogen III oxidase in Rhodobacter sphaeroides. II. Analysis of a region of the genome encoding hemF and the puc operon. Mol. Microbioi. 6, 3171-3186. Gong, L., & Kaplan, S. (1993). Abstract HI34. The role of gene Q in the expression of the puf operon in Rhodobacter sphaeroides. 93rd Annual Meeting American Society for Microbiology. Hoger, J. H., & Kaplan, S. (1985). Topology and neighbor analysis of the photosynthetic reaction center from Rhodopseudomonas sphaeroides. J. Biol. Chem. 260, 6932-6937. Hunter, C. N., & Coomher, S. A. (1988). Cloning and oxygen-regulated expression of the bacteriochlorophyll biosynthesis genes bchE, B, A, and C of Rhodobacter sphaeroides. J. Gen. Microbiol. 134, 1491-1497. Kaplan, S. (1990). Post-transcriptional control of the expression of photosynthetic complex formation in Rhodobacter sphaeroides In G. Drews & E. A. Dawes (Eds.) Molecular biology of membrane-bound complexes in phototrophic bacteria (pp. 10.5--113). New York: Plenum Press. Kaplan, S., & Amtzen, C. J. (1982). Photosynthetic membrane structure and function. In Govindjee (Ed.). Photosynthesis: energy conversion by plants and bacteria (vol. 1) (pp. 65--151). New York: Academic Press. Kaplan, S., Cain, B. D., Donohue, T. J., Shepherd, W. D., & Yen, G. S. L. (1983). Biosynthesis of the photosynthetic membranes of Rhodopseudomonas sphaeroides. J. Cell. Biochem. 22, 15-29. Kiley, P. J., Donohue, T. J., Havelka, W. A., & Kaplan, S. (1986). DNA sequence and in vitro expression of the B875 light-harvesting polypeptides of Rhodobacter sphaeroides. J. Bacteriol. 169, 742-750. Kiley, P. J., & Kaplan, S. (1987). Cloning, DNA sequence, and expression of the Rhodobacter sphaeroides light-harvesting BS00-850-cx and B800-850-[~ genes. J. Bacteriol. 169, 3268-3275. Kiley, P. J., & Kaplan, S. (1988). Molecular genetics of photosynthetic membrane biosynthesis in Rhodobacter sphaeroides. Mi'crobiol. Rev. 52, 50-69. Kiley, P. J., Varga, A., & Kaplan, S. (1988). Physiological and structural analysis of light-harvesting mutants of Rhodobacter sphaeroides. J. Bacteriol. 170, 1103-1115. Klug, G. (1993). Regulation of expression of photosynthesis genes in anoxygenic photosynthetic bacteria. Arch. Microbiol. 159, 397-404. Lee, J. K., & Kaplan, S. (1992). cis-Acting regulatory elements involved in oxygen and light-control of puc operon transcril~ion in Rhodobacter sphaeroides. J. Bacteriol. 174, 11461157.

102

AMY R. VARGA and SAMUEL KAPLAN

4~. Lee, J. K., & Kaplan, S. (1992). Isolation and characterization of trans-acting mutations involved

in oxygen regulation ofpuc operon transcription in Rhodobacter sphaeroides J. Bacteriol. 174, 1158-1171. 6. Lee, J. K., Kiley, P. J., & Kaplan, S. (1989). Posttranscriptional control ofpuc operon expression of B800-850 light-harvesting complex formation in Rhodobacter sphaeroides. J. Bacteriol. 171, 3391-3405. 47. Mullet. J. E., Klein, P. G., & Klein, R. R. (1990). Chlorophyll regulates accumulation of the plastid-encoded chlorophyll apoproteins CP43 and DI by increasing apoprotein stability. Proc. Natl. Acad. Sci. USA 87, 4038--4042. 48. Neidle, E. L., & Kaplan, S. (1993). 5-Aminolevulinic acid availability and control of spectral complex formation in HemA and HemT mutants of Rhodobacter sphaeroides. J. Bacterioi. 175, 2304-2313. 9Q Okamura, M. Y., Feher, G., & Nelson, N. (1982). Reaction centers. In Govindjee (Ed.). Photosynthesis: energy conversion by plants and bacteria (vol. I). (pp. 195-274). New York: Academic Press. 00 Okamura, M. Y., Steiner, L. A., & Feher, G. (1974). Characterization of reaction centers from photosynthetic bacteria. I. Subunit structure of the protein mediating the primary photochemistD, in Rhodopseudomonas sphaeroides R-26. Biochemistry 13, 1394-1403. 51. Parson, W. W. (1991). Electron transfer in reaction centers. In H. Scheer (Ed.). Chlorophylls. (pp. 1153-1180). Boca Raton, FL: CRC Press. 52. Richards, W. R., & Fidai, S. (1990). Studies on the function of the pufQ gene product in bacteriochlorophyll biosynthesis, in G. Drews & E. A. Dawes (Eds.). Molecular biology of membrane-bound complexes in phototrophic bacteria. New York: Plenum Press. 53. Richter, P., Brand, M., & Drews, G. (1992). Characterization of LHI- and LHI~ Rhodobacter capsulatus pufA mutants. J. Bacteriol. 174, 3030-3041. 4. Richter, P., Cortez, N., & Drews, G. (1991). Possible role of the highly conserved amino acids trp-8 and pro-13 in the N-terminal segment of the pigment-binding polypeptide LHIcz of RHodobacter capsulatus. FEBS Lett. 285, 80-84. 55. Richter, P., & Drews, G. (1991). Incorporation of light-harvesting complex I u and ~ polypeptides into the intracytoplasmic membrane of Rhodobacter capsulatus. J. Bacteriol. 173, 5336-5345. 6. Saier, M. H., Werner, P. K., & Muller, M. (1989). Insertion of proteins into bacterial membranes: Mechanism. characteristics and comparison with the eukaryotic process. Microbiol. Rev. 53, 333-366.

57. Shiozawa, J. A., Welte, W., Hodapp, N., & Drews, G. (1982). Studies on the size and composition of the isolated light-harvesting B800-850 pigment-wotein complex of Rhodopseudomonas capsulata. Arch. Biochem. Biophys. 213, 473-485. 58. Sockett, R. E., Dmmhue, T. J., Varga, A. R., & Kaplan, S. (1989). Control of photosynthetic membrane assembly in Rhodobacter sphaeroides mediated by puhA and flanking sequences. J. Bacteriol. 171, 43r A.A.~i. 59. Sprague, S. G., & Varga, A. R. (1986). Membrane architecture of anoxygenic photosynthetic bacteria. In L. A. Staehelin & C. J. Arntzen (Eds.). Encyclopedia ofplant physiology. New Series

vol 19. Photosynthesis !!1. Photosynthetic membranes and light-harvesting systems. (pp. 603619). Berlin: Springer-Verlag 00 Sundstrom, V., & van Gxondeile, R. (1991). Dynamics of excitation energy transfer in photosynthetic bacteria, in H. Scheer (Ed.). Chorophylls. (pp. 1097-1124). Boca Raton, FL: CRC Press. 61. Tadros, M. H., & Drews, G. (1990). Pigment-proteins of antenna complexes from purple non-sulfur bacteria: localization in the membrane, alignments of primary structure and structural predictions. In G. Drews & E. A. Dawes (Eds.). Molecular biology of membrane-bound complexes in phototrophic bacteria (pp. 181-192). New York: Plenum Press.

Assemblyof Pigment-ProteinComplexes

103

62. Tadros, M. H., Frank, R., Dorge, B., Gad'on, N., Takemoto, J. Y., &:Drews, G. (1987). Orientation of the B800-850, B870, and reaction center polypeptides on the cytoplasmic and periplasmic surfaces of RHodobacter capsulatus membranes. Biochemistry 26, 7680-7687. 63. Tadros, M. H., Frank, G., Zuber, H., & Drews, G. (1985). The complete amino acid sequence of the large bacteriochlorophyll-binding polypeptide B870r from the light-harvesting complex B870 of Rhodopseudomonas capsulata. FEBS Lett. 190, 41-44. 4. Tadros, M. H., Suter, F., Seydewitz, H. H., Wilt, I., Zuber, H., & Drews, G. (1984). Isolation and complete amino acid sequence of the small p01ypeptide from light-harvesting pigment-protein complex I (B870) of Rhodopseudomonas capsulata. Fur. ). Biochem. 138, 209-212. 65. Takemoto, J., & Lascelles, J. (1973). Coupling between bacteriochlorophyli and membrane protein synthesis in Rhodopseudomonas sphaeroides. Proc. Natl. Acad. ScL USA 70, 700-803. 6. Theiler, R., Surer, F., Pennoyer, J. D., Zuber, H., & Niederman, R. A. (1985). Complete amino acid sequence of the B875 light-harvesting proteins of Rhodopseudomonas splmeroides strain 2.4.1. FEBS Lett. 184, 231-236. 67. Theiler, R., Suter, E, Wiemken, V., & Zuber, H. (1984). The light-harvesting polypeptides of Rhodopseudomonas sphaeroides R-26.1 I. Isolation, purification and sequence analysis. HoppeSeyler 7.. PhysioL Chem. 365, 703-719. 68. Tichy, H.-V., Aibien, K.-U., Gad'on, N., & Drews, G..(1991). Analysis of the Rhodobocter capsulatus puc operon: the pucC gene plays a central role in the regulation of LHII (B800-850 complex) expression. EMBO J. 10, 2949-2955. 69. Tichy, H. V., Oberle, B., Stiehle, H., Schlitz, E., & Drews, G. (1989). Genes downstream from pucB and pucA are essential for formation of the B800-850 complex of Rhodobocter capsulatus. ). Bacterial. 17 I, 4914--4922. 70. Troschel, D., & Muller, M. (1990). Development of a ceil-free system to study the membrane assembly of photosynthetic proteins of Rhodobacter capsulatus. J. Cell Biol. 111, 87-94. 71. Varga, A. R., & Kaplan, S. (1993). Synthesis and stability of reaction center polypeptides and implications for reaction center assembly in Rhodobacter sphaeroides. J. Biol. Chem. 268, 19842-19850. 72. van Heijne, G. (1992). Membrane protein slructure prediction. Hydrophobicity analysis and the inside positive rule. J. MoL Biol. 225, 487-494. 73. Williams, J. C., Steiner, L. A., & Feher, G. (1986). Primary structure of the reaction center from Rhodopseudomonas sphaeroides. Proteins 1, 312-325. 74. Williams, J. C., Steiner, L. A., Feher, G., & Simon, M. I. (1984). Primary structure of the L subunit of the reaction center from Rhodopseudomonas sphaeroides. Proc. Natl. Acod. ScL USA 81, 7303-7307. 75. Williams, J. C., Steiner, L. A., Ogden, R. C., Simon, M. I., & Feher, G. (1983). Primary structure of the M subunit of the reaction center from Rhodopseudomonas sphaeroides. Proc. Natl. Acad. Sci. USA 80, 6505-6509. 76. Yeates, T. O., Komiya, H., Chirino, A., Rees, D. C., Allen, J. P., & Feher, G. (1988). Structure of the reaction center from Rhodobacter sphaeroides R-26 and 2.4.1.: protein-cofactor (bacteriachlorophyll, pbacteriopheophytin, and carotenoid) interactions. Proc. Natl. Acad. Sci. USA 85, 7993-7997. 77. Youvan, D. C., Bylina, E. J., Alberti, M., Begusch, H., & Hearst, J. E. (1984). Nucleotide and deduced polypeptide sequences of the photosynthetic reaction center, B870 antenna and flanking sequences from Rhodopseudomonas capsulata. Cell 37, 949-957. 78. Youvan, D. C., & Ismail, S. (1985). Light-harvesting II (B800-850 complex) structural genes from Rhodopseudomonas capsulata. Proc. Natl. Acad. Sci. USA 82, 58--62. 79. Youvan, D. C., lsmail, S., & Bylina, E. J. (1985). Chromosomal deletion and plasmid complementation of the photosynthetic reaction center and light-harvesting genes from Rhodopseudomonas capsulata. Gene 38, 19-30.

104

AMY R. VARGA and SAMUEL KAPLAN

80. Zhu, 3/'.S., & Hearst, J. E. (1986). Regulation of expression of genes for light-harvesting antenna

proteins LHI and LHII; reacdon center poylpeptides RC-L, RC-M, and RC-H; and enzymes of bacteriochlorophyll and carotenoid biosynthesis in Rhodobacter capsulatus during the shift from anaerobic 9 to aerobic gowOt. 1. Bacteriol. 168, 1180-1188. 81. Zhu, Y. S., & Kaplan, S. (1985). Effects of light, oxygen and substrates on steady state levels of mRNA coding for ribulose 1,5-bisphosphate carboxylase and light-harvesting and reaction center polypeptides in Rhodopseudomonos sphaeroides. J. Bocteriol. 162, 925-932. 82. Zhu, Y. S., Kiley, P. J., Donohue, T. J., & Kaplan, S. (1986). Origin of the mRNA stoichiometry of the puf operon in Rhodobacter splu~eroides. J. Biol. Chem. 261, 10366-10374. 83. Zuber, H., & Brunisholz, R. A. (1991). Structure and function of antenna polypeptides and chlorophyll-woteins complexes: principles and variability, in H. Scheer (Ed.). Chlorophylls. (pp. 627-703). Boca Raton, FL: CRC Press.

IDENTIFICATION AND RECONSTITUTION OF ANION EXCHANGE MECHANISMS IN BACTERIA

Atul Varadhachary and Peter C. Maloney

I.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Chemiosmotic Circuits . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Mechanisms of Carrier-Mediated Transport . . . . . . . . . . . . . . . . C. Anion-Linked Exchange in Bacteria . . . . . . . . . . . . . . . . . . . . II. Pi-Linked Exchange Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . A. The History of Pi Self-Exchange . . . . . . . . . . . . . . . . . . . . . . B. Current Status of the Pi-Linked Exchange Family . . . . . . . . . . . . . III. Two Key Experiments with Pi-Exchange . . . . . . . . . . . . . . . . . . . . A. Choosing between Antiport and Symport Mechanisms . . . . . . . . . . . B. Stoichiometry and Selectivity . . . . . . . . . . . . . . . . . . . . . . . . IV. Reconstitution of Pi-Linked Exchange . . . . . . . . . . . . . . . . . . . . . . V. Carboxylate-Linked Anion Exchange . . . . . . . . . . . . . . . . . . . . . . A. Oxalobacter formigenes . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Other Indirect Proton Pumps . . . . . . . . . . . . . . . . . . . . . . . . Advances in Cell and Molecular Biology of Membranes and Organeiles Volume 4, pages 105--128. Copyright 9 1995 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-$5938-924..9

105

106 106 107 108 109 109 111 113 113 114 116 119 119 121

106

ATUL VARADHACHARY and PETERC. MALONEY

Vl. Other Bacterial Anion Exchange Systems . . . . . . . . . . . . . . . . . . . . VII. Common Themes and Structural Rhythms . . . . . . . . . . . . . . . . . . . VIII. Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

122 123 124 125 125

I. I N T R O D U C T I O N A. Chemiosmotic Circuits Active transport at cellular and organellar membranes is driven by "pumps" or by "carders." Systems in the former category, such as the ion-motive pumps, move their substrates against an electrochemical gradient by linking the transport step to a chemical (or photochemical) transformation. On the other hand, carriers allow substrate movement along an electrochemical gradient, so that if one substrate is taken "uphill," a second substrate must at the same time move in the "downhill" direction. It is the indirect coupling between such primary (pumps) and secondary (carrier) reactions that forms the basis for much of the solute transport observed at bacterial (and other) membranes. In such cases, it is most common to find that a primary proton pump extrudes H +, generating a cytoplasm that is both alkaline and electrically negative relative to the external phase (34, 46), and that these gradients, which together comprise the "proton-motive force," support the activity of a variety of proton-linked carriers. Similar "chemiosmotic circuits" can be built on the movement ofNa + (or other) ions (13, 14, 58), but those that depend on H + seem to be the most common in bacteria and in the major eukaryotic organelles (34, 55). And while in most instances a primary proton pump is responsible for generating a proton-motive force, other arrangements can on occasion serve the same function. In fact, as described later, one can even organize events so that a carder itself serves in this capacity (6, 54). As shown by Figure 1, a surprisingly complex network of membrane transport events can be generated when chemiosmotic coupling ties together the activity of primary and secondary systems. The general principle, of course, extends well beyond this simple illustration, and we now appreciate that the ionic circuits described by Mitchell's chemiosmotic theory drive many different processes, ranging from substrate accumulation to cell motility (42). The availability of such a diverse repertory of driven events, each with its own substrate and kinetic specificity, confers a remarkable flexibility to the cell in its dealings with the external environment, and it is probable that "invention" of this principle was essential to the evolutionary transition from an acellular to a cellular mode of life

(38).

Anion Exchange Mechanisms in Bacteria

107

\ f

2HG6pI" S1

,

S2 / H *

H"

9\ ATP

H 9

~Na"

S3

,"

-

/ / G6P 2"

G6p2-

Figure 1. Chemiosmotic circuits at the bacterial membrane. All membranes have a collection of pumps and carriers. Those shown here can be found in anaerobes or in facultative organisms growing under anaerobic conditions, where the proton circulation is usually initiated by a proton-translocating ATPase. (Taken from ref. 34)

B. Mechanisms of Carrier-Mediated Transport Carrier-mediated events are traditionally placed in one of three mechanistic groups: (1) that of uniport (also termed facilitated diffusion); (2) symport (or co-transport); or (3) antiport (exchange). Examples of each of these are given in Figure 1. Thus, systems that catalyze uniport (carrier #1 in the figure) allow substrates to move freely along their electrochemical gradient and movement becomes independent of a chemiosmotic circuit, except insofar as the substrate may be charged or behave as a weak acid or base. Uniport systems are well represented in mammalian cells, but are less frequently found in the prokaryotic world, although a few examples have been described (34). Symporters (#2, #4) mediate a reaction in which two or more substrates move in tandem across the membrane; symport is widely spread among bacteria, and this mechanism is discussed in detail with reference to the lactose permease elsewhere in this volume. The remaining mechanistic category is that of antiport. Antiporters move substrates in opposite directions, and in the context of Figure 1 we might distinguish two different roles for such transporters. On the one hand, the activity of proton:cation exchange can effectively "transduce" a proton-motive force into an equivalent cation-motive force. In this way, for example, the presence of a Na§ § antiporter can extrude the Na + that enters by Na+-linked symport, allowing a subsidiary Na + circuit to coexist within the dominant H + circulation (Figure 1). In this sense, then, antiport is used to extrude a substrate. One might alternatively view an exchange system as a way to accumulate desired substrates. The known bacterial anion exchange mechanisms, which are the topic of this chapter, fall into this latter group.

108

ATUL VARADHACHARY and PETERC. MALONEY

C. Anion-Linked Exchange in Bacteria In bacteria, there are now two well-defined families of anion exchange reactions. The Pi-linked exchange family comprises a recently described group of carriers, each one of which accepts inorganic phosphate as a low affinity substrate and some specific organic phosphate as the high affinity ligand. The example, Figure 1 (carrier #5) depicts the activity of UhpT, a Pi-linked antiporter whose physiological action is directed to the transport of glucose 6-phosphate (G6P). This protein catalyzes electrically neutral exchange involving substrate(s) carrying two negative charges. The unusual feature of this protein is that these anionic equivalents can be taken as either a pair of monovalent substrates or as a single divalent species. Such plasticity has (as one might expect) real benefit in the context of bacterial cell biology, since the pH gradient is normally oriented as alkaline inside (see above). Thus, because monovalent sugar phosphate anions would dominate in the relatively acidic external phase, a pair of monoanions would move inward during the first half-turnover. Once within the cell, they would be stripped of their accompanying protons in the relatively alkaline interior, generating a pair of divalent anions. Since G6P is the high affinity substrate, and since two negative charges must move outward during the next half-turnover, it is likely that this latter step is preferentially accomplished by outflow of a single divalent G6P anion (rather than, say, a pair of monovalent Pi anions), leaving behind divalent G6P as a source of carbon and phosphorous. The net reaction, then, is import of 2H § and 1G6P2-, so that the antiport reaction effectively mimics a proton-linked symport, and sugar phosphate accumulation is driven by the pH gradient. This physiological model satisfactorily accounts for the available biochemical data, and it will be of great interest to describe the molecular basis of the rules governing such substrate selectivity and stoichiometry. Since very recent work (64) has identified a portion of the translocation pathway in UhpT, such information may be accessible in the near future. The second main family of bacterial anion exchange is exemplified by the antiport of oxalate with formate, as found in the gram-negative anaerobe, Oxalobacter formigenes. This exchange is catalyzed by a carrier known as OxlT (for oxalate transport) (54) and is of particular interest in the present context, because its activity is part of an "indirect" proton pump (6, .54). OxlT mediates the uptake of divalent oxalate (-OOC-COO-) against monovalent formate (HCOO-), its metabolic derivative. Since the one-for-one "vectorial" exchange carries negative charge inward, and since a "scalar" internal proton is consumed as formate is produced, the ensemble system acts as a proton pump. The antiporter, OxlT, plays a central role in this process and serves as the first of several examples in which a secondary carrier is incorporated into an "indirect" proton pump. The number of such systems is likely to grow much larger as we learn more about how bacteria extract energy from simple organic molecules. In the material that follows, we will review the evidence supporting these views of Pi-linked and carboxylate-linked exchanges, citing experiments that exploit

Anion Exchange Mechanisms in Bacteria

109

intact cells, membrane vesicles, and even proteoliposomes. A molecular analysis of these systems is now underway, and early results indicate that such exchangers can continue to add important information to mechanisms of antiport and membrane transport.

II. PI-LINKED EXCHANGE MECHANISMS A. The History of Pi Self-Exchange The study of Pi-linked exchanges began in the early 1950s, but it was not until recently that we realized these antiporters could carry both inorganic and organic phosphates. As a result, historical comments must allude to transport of both kinds of substrates. In 1953, Mitchell found that in Staphylococcus aureus (then called Micrococcus pyogenes), inorganic phosphate readily moved between the internal and external pools (41, 43). Full characterization of this exchange represented the first study of bacterial transport, and in doing this, Mitchell accomplished two important tasks. First, he showed that bacteria have an internal aqueous phase separated from the outside world by a lipoidal barrier that excludes large molecules--that is, bacteria have a membrane-bound cytoplasm, just as other cells do. Second, Mitchell exploited the fact that the Pi self-exchange reaction was highly sensitive to heavy metals, such as mercury and mercurials. By a series of clever inhibition experiments, he proved that the number of Pi molecules undergoing transport greatly exceeded the number of Hg atoms needed to inhibit the process. Accordingly, he concluded that the exchange was catalytic and that internalized Pi was not bound to a receptor, directly refuting the "exchange-adsorption" hypothesis for solute accumulation in bacteria. Mitchell also observed that Pi self-exchange was specific for the monovalent anion, HePO41-. This finding had no particular impact at the time, but it later provided our own studies with an important intellectual and experimental tool with which to complete analysis of this system. Figure 2 gives an example of the kind of Pi self-exchange reaction Mitchell studied in S. aureus. (The experiment is actually from our own work with Streptococcus lactis.) Cells harvested from the stationary phase of growth were suspended in a medium containing about 50 I~M Pi. Such cells have an internal Pi pool of 50 raM, but lack internal metabolizable reserves. As a result, ATP levels are low and the proton-translocating ATPase of this anaerobe (e.g., Figure 1) is unable to sustain a proton-motive force. Nevertheless, one can demonstrate membrane transport. Tracer amounts of 32pi are readily taken up by the cell, and at the steady state as much as 50-70% of the label has been internalized, where it is found in the pool of free inorganic phosphate (note that the label is readily chased by excess unlabeled Pi). Moreover, this process requires no metabolic energy; that is, it occurs in the absence of glycolysis, and is insensitive to a protonophore (FCCP) and to an

ATUL VARADHACHARY and PETERC. MALONEY

110 I.O

-

* " IC O ~ I + FCCP

0.8

& +OCCD

_

a

-'r-/

o.,,

~~ 0.6 o.D

o co

O.4 -

0.2

&

~..~+E.xces$ Unlabeled Phosphate

I

0

........

0

I

. . . . . . . . . . . . . . . . . . . . . . .

60 Minutes

I

120

.

9 9 I F

180

Figure 2. The Pi self-exchange reaction in Streptococcus lactis. Accumulation of 32pi was monitored after its addition to cells suspended in 300 mM KCI/20 mM MOP$/K (pH 7). At the arrow, part of the control suspension was given 4 mM Pi; where indicated, parallel suspensions had been treated with 10 pM FCEP (p-trifluoromethoxyphenylhydrazone) or I mM DCCD (N,N'-dicydohexylcarbodiimide). (Taken from ref. 39) inhibitor (DCCD) of the proton-translocating ATPase. These findings therefore establish a kind of paradox. The cell membrane is clearly permeable to Pi, which is free to rapidly move inward or outward, as traced by 32pi. Yet at the same time the membrane sustains a large chemical gradient for Pi; an inside/outside ratio of about 1000:1. How can this gradient be maintained across a permeable membrane in the absence of energy? The only satisfactory answer is that there is a simple one-for-one exchange of internal and external Pi. Mitchell also noted that, while Pi self-exchange was evident in resting cells, in metabolizing cells there was a net accumulation of Pi. From this observation, he hypothesized that the Pi exchange carrier was designed to operate as an energy-dependent Pi uptake system, but that it could cycle idly in the resting state. Some years later, Harold and coworkers continued the study of bacterial Pi transport in another gram-positive cell, Streptococcusfaecalis (23). They observed the expected energydependent system for net Pi transport, but could not demonstrate its exchange component. This apparent conflict was not resolved until much later, when experiments with Escherichia coil made it obvious that bacteria have many different systems that handle Pi. For example, we now know there are systems dedicated to the transport of phosphate alone; in E. coil these are exemplified by the multicomportent, ATP-dependent system known as Pst and by the H+/Pi symporter known as

Anion Exchange Mechanisms in Bacteria

111

Pit (10, 52, 53, 61). E. coli has two other systems that take Pi with low affinity and sugar phosphates with relatively high affinity (24, 48). These latter systems are now known to mediate the homologous Pi self-exchange reaction, while the Pi-specific porters (Pst, Pit) do not. In retrospect, then, it seems clear that Harold studied one of the dedicated Pi systems, while Mitchell worked with one of the less discriminating sugar-phosphate porters. These sugar phosphate transport systems have also had a long history. It had been known since 1950 (51) that E. coli could use sugar phosphates as sources of carbon and phosphorous, and in 1962, Lin et al., (32) demonstrated that G3P could enter cells without prior hydrolysis. Similar observations were soon made for G6P (19, 22, 48), and the relevant transport systems came to be known as GIpT (for glycerol phosphate transport) and UhpT (describing the uptake of hexose phosphate). In the 1970s, as it became clear that the chemiosmotic theory correctly described membrane events in bacteria, further work seemed to confirm that these transporters fit into the traditional mold as proton symporters (17, 63), and it was not until 1984 that their exchange nature became evident. At that point, independent studies in Streptococcus lactis (e.g. Figure 2) suggested that at least one of these sugar phosphate transporters mediated a Pi self-exchange of the sort described by Mitchell 30 years earlier (39). This idea carried with it the implication that other such systems might also use antiport as their mechanistic base. Further biochemical analysis of the streptococcal system, later extended to E. coli and to S. aureus, now provides definitive evidence supporting the idea of antiport as the operative mechanism of all Pi-linked exchangers (2, 40).

B. Current Status of the Pi-Linked Exchange Family As their name implies, all members of the Pi-linked exchange family take inorganic phosphate as a low affinity substrate, showing Michaelis constants for transport (Kt) on the order of 0.5 mM to 10 raM. In addition, each family member has one or more higher affinity sugar-phosphate substrates with binding constants in the range of 10 I.tM to 100 ~M (40). This family now includes seven known examples, each of which has a different pattern of preferred substrates (Table 1). The examples known in gram-positive cells use hexose phosphates as their natural substrate. The four gram-negative examples take either hexose or triose phosphates, depending on the specific example. In all cases, there is sufficient biochemical work to justify grouping these systems into a single family, but a structural relatedness has been verified only for the gram-negative cases, all of which show significant amino acid sequence homology to one another (__.30-35% amino acid identity) (I 5,

20, 21, 26). A diagnostic feature of this family is the Pi self-exchange reaction described earlier (Figure 2), a reaction that can be documented at all levels of biochemical analysis--in the intact cell, with membrane vesicles and with proteoliposomes. This exchange is relatively specific. No inorganic cations move with Pi during

ATUL VARADHACHARY and PETERC. MALONEY

112

Table 1. Pi-Linked Anion Exchange Systemsa Cell l)~e

Geneb

Primary Organic Substrate

V,~g Ratioc

E. coil

uhpT slpT uhpT pgtP

G6P, M6P G3P G6P PEP G3P G6P, G3P G6P, M6P

5.5

S. lyphimurium

S. aureus 8. iactis

6.5 4.2 4.7

Notes: 'Reviewed in (40). The reactions in S. oureusand S. lactisate presumed to reflect the activity of single systems, since transport-negativemutants show neither homologous nor heterologoes ~change (56, unpublished work). bSystems described by fonr-lemer genetic symbols show close sequence homology (15, 20,

2L 26).

CRatio of maximal velocities for homologous and hetetologons exchanges (2, 3, .56, 60, unpublished work with UhpT).

exchange, and of the inorganic anions tested, only arsenate is known as an alternative substrate; in fact, it appears that these systems do not distinguish between the phosphate and arsenate anions. This tight specificity is accompanied by an equally tight selectivity--in the two examples studied carefully ($. aureus and S. lactis), only monovalent phosphate (arsenate) partakes in exchange (39, 43). More relevant to physiological function is the heterologous exchange of Pi for a sugar phosphate. The identity of the favored sugar phosphate varies with the individual carrier, as does its Michaelis constant for transport; as noted above, it seems generally true that Pi is of relatively low affinity, while the preferred organic phosphate is of much higher affinity. In addition, for reasons that are not clear, the heterologous exchange is slower, usually by about fivefold, than the homologous exchange reaction (Table 1). But perhaps most important, in the cases examined so far, this heterologous exchange proceeds with no net transfer of charge--that is, the exchange is electroneutral. (This behavior contrasts strikingly with that of the carboxylate-linked exchange described later.) Note that in principle, each exchange carrier should also display a sugar-phosphate self-exchange, so that in thinking about the biochemistry and cell biology of these systems, one has to integrate the expected behavior of all three possible modes of operation. In addition, since each substrate (Pi, sugar phosphate) is a weak acid with a pK in the physiological range, and since these carriers might, in principle, select some or all salt forms of its substrates (e.g., monovalent Pi, or both divalent and monovalent sugar phosphate), the analysis of any particular situation can become complex (see below). Fortunately, with adequate control over some of these variables, one may simplify the analysis considerably, and to illustrate this, we have chosen to discuss two key experiments in the next section.

Anion Exchange Mechanisms in Bacteria

113

III. TWO KEY EXPERIMENTS WITH PI-EXCHANGE A. Choosing between Antiport and Symport Mechanisms

Work with both gram-positive (S. lactis, S. aureus) and gram-negative (E. coil) cells has contributed importantly to the biochemfcal analysis of Pi-linked exchange and to our understanding of mechanism. Among the most important of these experiments have been those that exclude the possibility of symport and positively support the idea of antiport. Recall that for some time sugar phosphate transporters were assumed to operate as symport mechanisms. Work with the lactose carrier of E. coli had established that symporters could display a variety of partial reactions, including a prominent substrate-substrate exchange, in the absence of a driving ion-motive force. Therefore, it was feasible (at least in principle) that the exchanges associated with the UhpT transporters orS. aureus, S. lactis, and E. coil were similar partial reactions on the part of a 2H+/G6P 2- symporter. To evaluate this alternative directly, right-side-out membrane vesicles from E. coli were prepared using sulfate rather than phosphate as the main internal anion. When presented with an oxidizable

-A

4IB

tm

UhpT § UhpT-

0 0.

c

3

:"t~

_

No Additions +PMS/Asc

*PMS/Asc/FCCP Q.

~o~

O~ m 0 0 =3

II

E! 9

3

O

9

X" 2

2

1 -

-

9

.c

"~

1

u

(3 J

~

6 .............. 12

0

i5-

20

25

Minutes

F~,ure 3. Sugar phosphate transport by Escherichia coli requires internal Pi. Membrane vesicles from wild-type or UhpT-negative strains were loaded with MOPS(K)/SO4 as the internal anions, and, as indicated, suspended at pH 6 in the absence or presence of PMS/ascorbate as the oxidizable energy source. (A) [14C]G6P was added to 100 ~M. (B) Vesicles were preincubated with 100 pM Pi for 15 min before adding [ 14C]G6P. (Taken from ref. 57)

114

ATUL VARADHACHARY and PETERC. MALONEY

energy source, these vesicles accumulated Pi and other substrates by the expected proton-linked reactions (e.g., via Pit, and other carriers), but they could not accumulate G6P (Figure 3A)--even though tests of the cells used to make vesicles had shown vigorous UhpT activity. However, when G6P was added after vesicles had taken up Pi, the capacity to transport G6P was recovered (Figure 3B) (57). This kind of experiment, which can be replicated in S. aureus (56), excludes the model of symport for G6P transport and strongly favors the idea of antiport.

B. Stoichiometry and Selectivity

Variable Stoichiometry Measurement of exchange stoichiometry has also proven instructive--perhaps even essential--to the complete description of Pi-linked exchange. These data provided the basis for both a satisfactory biochemical model of antiport and a way to account for some otherwise puzzling aspects of cell biology. A direct approach to this issue was best done with the UhpT cartier of S. lactis, since in this cell the main sugar phosphatase is completely blocked by low concentrations of the phosphatase inhibitor, orthovanadate. Nevertheless, the conclusions drawn from these studies almost certainly characterize the sugar phosphate carrier proteins of S. attreus and E. coli. When assayed at pH 7, a measurement of net fluxes shows that in S. lactis the macroscopic stoichiometry of heterologous exchange by UhpT is 2:1, Pi:G6P. Since the reaction is electrically neutral, and since Pi-self exchange selects monovalent Pi (in both S. lactis and S. aureus (39, 43), this bulk stoichiometry suggests that at a molecular level the protein moves 2H2PO41- against I G6P 2- (4). In turn, this suggests the protein can bind either a pair of monovalent anions or a single divalent anion. Having established that divalent G6P was a substrate, it was next important to ask whether monovalent G6P could be accepted--bacteria often find themselves in relatively acidic environments where monovalent G6P would dominate (pK = 6.1 ). Accordingly, two relevant parameters were examined at pH 5.2, where HG6P lis enriched. First, kinetic tests revealed a considerably lower rate of exchange but no substantial change in the affinity (IG) of the system for G6P. For this reason, it was concluded that both divalent and monovalent forms of G6P were substrates of UhpT (4). The second test involved determination of the net stoichiometry of heterologous exchange at pH 5.2. Surprisingly, at pH 5.2 the exchange ratio had a value of 1:1 (Pi:G6P), rather than the 2:1 value measured at pH 7 (4) (Figure 4); an intermediate value of stoichiometry was found at an intermediate pH (1:1.5 at pH 6.1). These findings seemed most easily understood after recalling that both monovalent Pi and monovalent G6P anions are acceptable substrates, at least as determined by kinetic tests. Therefore, the 1"1 stoichiometry at acid pH likely reflects the exchange of monovalent salts, and to be consistent with the earlier

Anion ExchangeMechanisms in Bacteria

115

b

200

200

U 0

c

I 4,. Q.

e 150 E

150 "i

.,,,. 0

E c

I00

'O 0 c 0

IIX) i

0

9

50

50

,oc

50

0 ,c

o8

0

15

30

45

0

5 0 Minutes

I0

15

20

0

0

50

I00

2-Deoxyglucose 6 - Phosphate Lost (nmol/mQ protein )

Figure 4. Variable stoichiometry of Pi-linked exchange in Streptococcus lactis. (A) Membrane vesicles loaded with 50 mM KPi were suspended in 125 mM K2SO4/20 mM MOPS/I( (pH 7) and exposed to 180 laM [14CIG6P, with 0.25 mM Na3VO4 added to block phosphatase activity. (B) After 45 rain, vesicles were isolated by centrifugation, resuspended, and distributed to tubes with assay buffer at pH 7 (v ,T), pH 6.1 (I,c3) or pH 5.2 (A, zx).After a further 10 rain, 3 mM 32pi was added (arrow) and samples taken to evaluate 32p gained (open symbols) and 14C lost (closed symbols). (C) Phosphate gained is correlated with sugar phosphate lost, using symbols as in B. The lines drawn have slopes of 1 (pH 5.2), 1.5 (pH 6.1) and 2 (pH 7). (Taken from ref. 4) finding of two monoanion binding sites (at pH 7), one must only imagine that the bulk ratio of l: 1 (pH 5.2) corresponds to a molecular exchange ratio of 2:2. Clearly, these results suggest that UhpT has an unusual pH-dependent variable stoichiometry, at least with respect to its sugar phosphate substrate. We believe the interpretation of this phenomenon is correctly embodied by the ionophore, A23187, a simple molecule that also shows variability in exchange stoichiometry. A23187 has a pair of carboxyl groups, each of which must be paired with a positively charged substrate before transport can occur. This is accomplished by the binding of either a pair of protons (2H+), or, after a change of conformation, the binding of a single divalent cation (e.g., Ca2+). The ionophore, then, shows a pH-dependent stoichiometry---either 2:2 (2H+:2H+), 2:1 (2H+:Ca2+), or 1:1 (Ca2+:Ca2§ according to the concentrations of each substrate in either aqueous compartment. In the same way, we suggest that UhpT has a pair of binding sites (perhaps a pair of arginine residues) that accept two monovalent anions, and that each of these sites (arginines?) can contribute to a more complex binding pocket that accepts divalent

116

ATUL VARADHACHARY and PETERC. MALONEY

substrate. In this way, the phenomenon of variable stoichiometry for UhpT is attributed to a pH-dependent change in the protonation state of its substrates and not to any special property of the protein itself (4, 40).

The Phenotypeof Proton-LinkedSymport These arguments are relevant to understanding the cell biology of UhpT, for we can now appreciate why these proteins were categorized as proton-linked symporters in the past (17, 63). The analysis of variable stoichiometry has led to the conclusion that UhpT can accept either a pair of monovalent sugar phosphate anions or a single divalent anion, depending on which is the more prevalent (i.e., depending on pH). Therefore, when a pH gradient exists, it is clear that the neutral self-exchange of sugar phosphate will be in a 2:1 ratio, as discussed earlier (Figure 1), and a protein whose mechanistic base is antiport will appear at the macroscopic level to operate as a 2H§ 2- symporter. Note also that the model of symport carries risks that antiport does not. If E. coil growing on G6P were to consume its substrate, a symport mechanism would allow net efflux of internal G6P, whereas an antiport mechanism would not. (This in no way points to antiport as a necessary mechanism, but it does make one feel more comfortable with the general idea.)

Setting the Carbon:PhosphorousRatio Despite its prominence in the discussions outlined above, the self-exchange based on sugar phosphate cannot be the only mechanism used by UhpT. For if that were the case, growth on G6P would seriously distort the normal carbon:phosphorous ratio of40:1 (33). It seems reasonable, therefore, to suggest a concomitant use of the heterologous 2Pi:G6P exchange, allowing the import of sugar to proceed against the simultaneous export of Pi--the net result of 2Pi:G6P antiport. It is not necessary to assume unusual regulatory events to set the balance of these exchange modalities (although such regulation may well occur), for it is easy to see this occurring by mass action alone, as the internal pH and G6P and Pi pools fluctuate in response to the substrate fluxes.

IV. RECONSTITUTION OF PI-LINKED EXCHANGE Any rigorous analysis of antiport systems requires that one have experimental control over the composition of internal and external compartments, and this is most easily accomplished in liposomal systems containing only membrane proteins and simple salts. For that reason, it was important to develop adequate methods of reconstitution for the study of Pi-linked exchange. The techniques we devised for this purpose have greatly aided the study of Pi-linked events and have also had a

Anion ExchangeMechanisms in Bacteria

117

Table 2. Effect of Osmolytes on Reconstitution of Pi-Linked Exchange a

Test Compound Valine (7%) Methanol (10--20%) None Ethylene glycol (10-15%)

32Pi Transport nmol/mg Protein 15 27 + 5 36 + 9 42 + 16

Proline (12%) Glycine (8%) Glucose (20%)

305 410 425

Glycerol (20%)

530 :k98

Note: =Proteoliposomes loaded with 100 mM KPi were suspended in sulfatebased MOPS-buffered medium, and the Pi self-exchange reaction w a s measured at steady state after a 75-rain incubation with 50 IJM 32Pi.(Data from a larger survey reported in ref. 3.)

positive impact on the more general use of reconstitution as a tool in membrane biology. The general methods we adopted were chosen for their simplicity of use. Thus, we have most often relied on the technique of octylglucoside-dilution to form liposomes. This approach, introduced by Thompson and by Racker and his colleagues (50), exploits the water-soluble non-ionic detergent, octylglucoside. After solubilization by octylglucoside in the presence of exogenous lipid (see below), proteins are incorporated into a liposomal membrane by a dilution step which lowers detergent to a level below its critical micellar concentration (about 25 mM for octylglucoside). The insoluble lipid and protein form small (ca. 15 nm diameter), very stable, unilameilar liposomes with an internal aqueous volume of about 1 ~l/mg phospholipid (3). Our experiments have also drawn on the work of Newman & Wilson, who noted the beneficial effect when solubilization was performed in the presence of excess lipid (45). Despite the general success of reconstitution in other systems, our initial attempts to apply these methods to the Pi self-exchange reaction were disappointing; the activity recovered in proteoliposomes reflected only a small part of that present in the parent membrane vesicles, indicating a significant inactivation. Pronounced inactivation is, in fact, the usual timing at the beginnings of such a study, and one normally expends considerable time and labor in searching for conditions compatible with retention of function. These searches may now be easier, because the solution to this problem for Pi-linked exchange has uncovered an approach of general value. Analysis of this issue for the antiporter of S. lactis led us to conclude that the protein was subjected to an unpredictable all-or-none inactivation during solubilization. On the other hand, when protein was finally reincorporated into a liposomai membrane, after the dilution step, transporters that had survived were

118

ATUL VARADHACHARY and PETERC. MALONEY

able to operate with their normal kinetic properties (3). This prompted us tO search for materials that might protect solubilized membrane protein, and that survey soon led to the discovery that a class of compounds known as "osmolytes" (65) have a particularly beneficial action. We also noted that inactivation of the streptococcal carrier was largely restricted to a brief interval, literally during solubilization, so that these protein stabilants were effective only if added before the detergent (3). The positive effect of osmolytes is described in Table 2. Note that in Pi-loaded liposomes prepared without such intervention, the steady level of 32Pi accumulation is about 20-d0 nmol/mg protein. Traditional stabilants such as methanol, or ethyleneglycol have no effect, but high concentrations of osmolytes, including certain amino acids (proline, glycine, not valine), sugars, and glycerol or higher polyols, lead to a 10 to 20-fold increased recovery of activity. This remarkable effect is not restricted to the Pi-linked exchange proteins, nor is it confined to bacterial systems. Instead, it appears to be a general response on the part of many membrane proteins, prokaryotic and eukaryotic, including ATP-driven pumps, various secondary carriers, and several membrane-bound enzymes (37). As noted earlier, detergent solubilization of many membrane proteins, including the Pi-linked exchangers, yields largely inactive protein, and this presents a major impediment to the characterization and purification of such systems. Fortunately, in many cases this inactivation appears to be largely overcome by appropriate use of osmolytes. Therefore, even though the mechanism by which osmolytes exert their effect remains conjectural (3, 8), the intervention can have a substantial practical value. An example showing the benefits of this approach to reconstitution comes from our attempts to streamline the protocol. Traditionally, membrane proteins are solubilized from membrane vesicles, but because of the labor involved in making vesicles, proteoliposomes are used less frequently than their advantages might warrant. This is especially true if a number of strains are to be screened, since vesicles would have to be prepared for each strain, with the method often having to be optimized for each strain. In an attempt to circumvent this difficulty, we developed a rapid method for reconstitution of membrane proteins directly from intact cells (59, 60). In this technique, cells are digested with lysozyme, osmotically lysed, and the resultant ghosts are then solubilized in the presence of an osmolyte, as already outlined. With this technique, protein can be solubilized from cell cultures as small as a few milliliters, and many different strains can be examined in a single day. Proteoliposomes made in this way are suitable for both quantitative and qualitative studies of antiport. In particular, there is no evidence for parallel reconstitution of the bacterial porins, possibly because of their generally hydrophilic nature, so that the proteoliposomal membrane is capable of sustaining a membrane potential (59). We anticipate that such an approach can become a useful tool for the screening and characterization of large numbers of strains, both wild-type and mutant.

Anion Exchange Mechanisms in Bacteria

119

V. CARBOXYLATE-LINKED ANION EXCHANGE A. Oxalobacter formigenes As noted in the previous section, the application of reconstitution to the study of antiport systems provides the investigator with an explicit control over internal and external milieux. Perhaps more generally important, the success of osmolytemediated techniques encourages one to use reconstitution as an analytical rather than as a preparative tool. Indeed, this approach was essential in the description and analysis of carboxylate-linked exchange, the first example of which was described using protein reconstituted from membranes of the gram-negative anaerobe, O. formigenes. Early studies of this cell had demonstrated that it uses the decarboxylation of oxalic acid, the simplest dicarboxylic acid, as its sole source of metabolic energy (l, 9). Thus, ifonly for its role in transport of this growth substrate, it seemed that studies of oxalate transport might be of interest. Guided by experiments of reconstitution, we were led to the discovery and eventual purification of a membrane carrier--OxlTwthat contributes in an entirely unexpected way to membrane biology in this cell (6, 54). The carrier is also of interest for biochemical reasons, since the purified material has a specific activity during oxalate self-exchange in excess of 500-700 lamol/min per mg protein, and a turnover number on the order of> 1000Is (6). Among carriers that handle organic molecules, this places OxiT at the head of the line in terms of velocity. Two simple observations form the basis of the model of how OxlT contributes to cell biology. First, work using oxalate-loaded proteoliposomes shows that purified OxlT mediates an oxalate self-exchange, a reaction that proceeds independently of co-ions and that reflects the strict one-for-one antiport of internal and external substrate. Second, this same carrier can accept (with somewhat lower affinity) formate as an alternative substrate. Most important, the heterologous exchange of oxalate and formate is electrogenic, with negative charge moving in the same direction as oxalate. The electrogenic nature of this exchange (which contrasts with the electroneutral behavior of Pi-linked events) is readily demonstrated in a reconstituted system, as shown by Figure 5. In that experiment, proteoliposomes containing purified OxlT were loaded with either the potassium or N-methylglucamine salt of formate, and then suspended in N-methylglucamine or potassium sulfate, respectively, in the presence of labeled oxalate. Of themselves, these maneuvers had no effect on the accumulation of oxalate, but when valinomytin was also present, to impose a membrane potential negative or positive inside, oxalate accumulation was either inhibited (internally negative potential) or accelerated (internally positive potential). These responses are precisely those expected for the transport of divalent oxalate (-O(~-COO-) against monovalentformate (HCCO-). It is the electrogenic nature of the heterologous exchange of oxalate and formate that is essential to the role ascribed to OxlT. Figure 6 depicts this in simplified fashion by showing how the metabolic coupling between OxlT and an internal

ATUL VARADHACHARY and PETERC. MALONEY

120 2O

15

o3

~o t~Q.

! A v

00

A v

, 2

A v

n 4

A

n 6

,

I 8

10

Minutes

Figure 5. OxlT catalyzes the electrogenic exchange of oxalate and formate. During

reconstitution of purified OxlT, proteoliposomes were loaded with 100 mM NMG formate or 200 mM potassium formate. To start each assay, proteoliposomes were diluted from their concentrated stocks into a potassium- or NMG-based buffer, respectively, so that in one case there was an inwardly directed potassium gradient (o,=), while in the other case this gradient was directed outward (o,o). Experimental tubes (solid symbols) received 1 FM valinomycin; control tubes (open symbols) received the ethanol carrier. The reaction was started by addition of 1O0 I~M [14C]oxalate. (Taken from ref. 54)

decarboxylating system allows O.formigenes to sustain a proton-motive force. The complete system has three components---(1 ) OxlT, which catalyzes the electrogenic exchange of divalent oxalate and monovalent formate; (2) a formylCoA transferase (6), which uses formyICoA as a CoA donor for incoming oxalate, thereby releasing the formate whose efflux returns OxlT to its original state; and (3) an oxalylCoA decarboxylase, which acts to decarboxylate oxalylCoA, generating formylCoA and carbon dioxide. (The catalytic role of the formylCoA transferase is not shown in Figure 6.) Since the one-for-one exchange of oxalate and formate carries a single negative charge into the cell, the membrane potential becomes internally negative, and since the decarboxylation reaction consumes a single scalar proton, an internal alkalinity is sustained. In aformal sense, then, a single turnover of the OxlT-transferase-decarboxylase system acts to pump a single proton (H+) out of the cell! Three turnovers of this "pump" move three protons outward, and as these reenter via the proton-translocating ATPase, they provide sufficient energy to synthesize ATP. Therefore, this model also offers a solution to the thermodynamic problem that a decarboxylation of oxalate cannot of itself yield sufficient free energy to make ATP.

Anion ExchangeMechanisms in Bacteria

121

,§

Oxalate2~..

ANTIPORT --.---eOxalate2- " ~ DIECARBOXYLASE FormatelACID

§ .~. +

~ ALKALINE

CO2

2

3H+

F~ure 6. The indirect proton pump of Oxalobacter formigenes. A proton-motive force arises after entry of divalent oxalate, its internal decarboxylation and exit of monovalent formate. The net entry of one negative charge and the disappearance of one internal proton is equivalent to the outward movement of one positively charged proton. Three such cycles, each with a stoichiometry of 1H+/turnover, would support synthesis of ATP. (Taken from ref. 6) The biochemistry of O. formigenes ensures that this system is always poised for a flux of inwardly flowing oxalate and outwardly moving formate. Such directionality is maintained in two ways. On the one hand, dissemination of the gaseous carbon dioxide limits the availability of product required for the back-reaction. In addition, O. formigenes can express very high levels of both OxlT and the decarboxylase (and, presumably, the CoA transferase). For example, under extreme conditions, when cells are grown with limiting oxalate, so that external substrate is reduced to micromolar levels, OxlT comes to represent at least 5% of membrane protein (54), and the decarboxylase alone can be as much as 10% of cytoplasmic protein (9). Inasmuch as these proteins are central to the lifestyle of this organism (Figure 6), such overexpression would seem to be a sensible investment.

B. Other Indirect Proton Pumps

O. formigenes represents the first explicit demonstration of how vectorial and scalar reactions might be organized so as to construct a new kind of proton pump. But the principle is a general one, and it is sure to emerge in other circumstances.

ATULVARADHACHARYand PETERC. MALONEY

122

Even now, other examples are coming to the surface. For example, we now appreciate that growth by malolactic fermentation (malate + H + - , lactate + CO2) can fit the paradigm, and several lines of evidence suggest there is the required electrogenic net exchange of malate and lactate (47, 49). Similarly, the decarboxylation of histidine to yield histamine follows the same general rules (44), as does decarboxylation of aspartate to yield alanine (Abe and Maloney, unpublished results). Of course, this economical design need not be restricted to simple decarboxylation reactions, and it is possible to imagine the same principle at work in cases where complex intracellular processing is responsible for the scalar consumption of protons. In this context, Archaebacterial methanogenesis might serve as an interesting test case. In such organisms, energy is derived from the metabolism of simple anions (acetate, formate, bicarbonate) to yield even simpler products (methane, water, carbon dioxide, hydrogen gas) (27). If these substrates enter the cell as anions, that is as charged species, one could view the totality of their remaining biochemistry as an ensemble of reactions designed only to consume a single scalar proton for each charge that has entered. This gives the required coupling between vectorial and scalar events, with the end result of an indirect proton pump. While such a suggestion is entirely speculative, it does illustrate the potential utility of the general principle when extended beyond the immediate examples.

VI.

OTHER BACTERIAL A N I O N EXCHANGE SYSTEMS

Despite the common occurrence of anion exchange in eukaryotes, there are still relatively few documented examples of anion exchange in prokaryotes. And while this chapter has emphasized the Pi-linked and carboxylate-linked exchangers, a few others are known. A relatively well-studied example is that of the ATP-ADP translocase of Rickettsia prowazeki. This carrier, which is used by the parasite to gain access to the host adenine nucleotide pool, has now been cloned, sequenced and expressed in E. coil (30). The amino acid sequence of this protein supports its inclusion in the general class of secondary carriers, but does not suggest it belongs to the Pi-linked family described above. A phenotypically similar transporter, catalyzing the exchange of ADP and Pi, has also been described in Rhodobacter capsulatus (11), but this latter example is as yet known only at a biochemical level. And in the same way, biochemical but not genetic evidence has been given for the presence of a dicarboxylate exchange in anaerobically grown E. coli (16). Finally, we note that there are many anion exchange events associated with mitochondria and chloroplasts. This, and the widespread existence of eukaryotic anion exchange, suggests there are equal numbers of prokaryotic examples awaiting identification.

Anion Exchange Mechanisms in Bacteria

123

VII. C O M M O N THEMES AND STRUCTURAL RHYTHMS While the study of anion exchange proteins has its intrinsic value, recent discoveries in the larger field of membrane transport suggest that work on any one of these secondary carriers can have relevance to the entire group. Despite differences in the kinetic and biochemical behavior of these various proteins, an underlying structural theme is emerging. The most convincing evidence to this point arises from theoretical and practical studies of how these proteins are arranged within the membrane. Topological predictions of a large number of secondary carriers, both eukaryotic and prokaryotic, show that their most characteristic feature is the presence of transmembrane segments in multiples of 5-6 (35, 36). In cases where the functioning carrier is known, 10-12 segments appear to comprise the operational unit, whether this is as a dimer of subunits containing 5-6 segments, as in the chloroplast and mitochondrial carriers (7, !8, 29), or as a monomer containing 10-12 segments, as for the UhpT (5) and LacY (12) proteins of E. coli (Figure 7). Some time ago it was argued that transport proteins should function as oligomeric proteins, probably dimers, with the translocation pathway defined as the volume enclosed by the interface between subunits (28, 31). Early work with the mitochondrial exchange proteins supports the idea, but demonstration that LacY and UhpT (and other transporters) can function as monomers appears contradictory. But this conflict can be easily resolved if the larger (monomeric) carriers are viewed as having a substructure allowing them to behave as functional dimers. And, in fact, the argument for an intramolecular dimer is supported (albeit, indirectly) by the finding of sequence homology between the two N- and C-terminal halves of some carders (25). If only for the purpose of designing the next generation of experiments, we offer the UhpT antiporter of bacteria as providing a unique perspective on this issue. As noted before, UhpT should contain a pair of symmetrical binding sites so that the protein can accept either two monovalent substrates or single divalent passenger. Is it unreasonable to suggest that each half of the molecule contains one of these binding sites and that substrate passes by the interface between the N-terminal and C-terminal halves (Figure 7)? Perhaps not. Preliminary experiments have convincingly demonstrated that the seventh transmembrane segment of UhpT comprises part of the transport path (64), and our next experiments will attack this issue more fully by using molecular biology to define the arrangement of those parts of the molecule surrounding this substrate translocation pathway. If this at the same time clearly delineates the region between the N- and C-terminal halves, we will take this as evidence to support the view that UhpT is constructed as, in effect, a covalent heterodimer.

124

ATUL VARADHACHARY and PETERC. MALONEY

NH2

J

COOH CYTOPLASM

i

m PERIPLASM

/

Figure 7. A low resolution view of UhpT: generic structure of a membrane carrier. The schematic shows the topology of UhpT derived by Kadner and his colleagues from hydropathy analysis and gene fusion experiments (26). The N- and C-termini, as well as the central loop, are expected to be exposed to the cytoplasm. Transmembrane segments are indicated by the shaded columns labeled I-XII. New data (64) show that cysteine-265 (circled lies on the pathway taken by substrates moving through Uhpt. This is with the idea that the translocation pathway is defined by the apposition of the N- and C-terminal domains of a covalent heterodimer. VIII.

CHAPTER SUMMARY

Even though anion exchange in bacteria was described as early as 1953, it is only in the last decade that the underlying mechanism has been characterized in a satisfactory way. Work with vesicles and proteoliposomes shows that the chemiosmotic transport of sugar phosphates occurs, not by proton-symport, but by anion exchange, in both gram-positive and gram-negative bacteria. Further analysis shows the existence of a variable stoichiometry that allows homologous sugar phosphate exchange to mediate an unexpected net uptake of substrate. Since this exchange causes the influx of protons along with the sugar phosphate, continued transport depends on maintenance of a pH gradient (alkaline inside), thereby ensuring that the Pi-linked exchangers take part in the overall chemiosmotic circuitry of the cell.

Anion Exchange Mechanisms in Bacteria

125

The study of Pi-linked exchange has defined not only their physiological and biochemical properties, but has also been the vehicle for development of a particularly simple set of protocols designed to study such proteins in an artificial liposomal system. Such methods, which take advantage of the stabilizing effects of osmolytes (e.g., glycerol), have made it possible to use reconstitution in an analytical fashion, not just as a preparative tool. The value of this methodological strategy is well-illustrated by studies of OxlT, the membrane carrier responsible for the electrogenic exchange of oxalate and formate in O. formigenes. Because this cell had not been subjected to earlier biochemical or genetic tests, it is unlikely that traditional approaches would have so readily led to an appreciation of the central role OxIT plays in establishing the proton-motive force in this anaerobe. This finding is perhaps the major contribution of such work, for OxlT has now become the first of many examples in which generation of a proton-motive force relies, not on a primary proton pump, but on the indirect coupling between secondary carriers and internal metabolism, an organizational scheme that may be of unsuspected antiquity (38). We have also touched on what seems to be an emerging commonality to all membrane carriers, one reflected in the 10-12 transmembrane segments that form the functional unit of UhpT. This strengthens the view that examining bacterial anion exchange might reveal facts and principles more broadly applicable to membrane transport systems.

ACKNOWLEDGMENTS The study of anion exchange in this laboratory is supported by grants from the Public Health

Service (GM24195) and the National Science Foundation (MCB9220823).

REFERENCES 1. Allison, M. J., Dawson, K. A., Mayberry, W. R., & Foss, J. G. (1985). Oxalobacterformigenes gen. nov., sp. nov.: oxalate-degrading anaerobes that inhabit the gastrointestinal tract. Arch. Microbiol. 141, 1-7. 2. Ambudkar, S. V., & Maloney, P. C. (1984). Characterization of phosphate:hexose 6-phosphate antiport in membrane vesicles of Streptococcus lactis. J. Biol. Chem. 259, 12576-12585. 3. Ambudkar, S. V. & Malone),, P. C. (1986). Bacterial anion exchange. Use of osmolytes during solubilization and reconstitution of phosphate-linked antiport from Streptococcus lactis. J. Biol. Chem. 26 I, 10079-10086. 4. Ambudkar, S. V., Sonna, L. A., & Maloney, P. C. (1986). Variable stoichiometry of phosphatelinked anion exchange in Streptococcus lactis: implications for the mechanism of sugar phosphate transport by bacteria. Proc. Natl. Acad. Sci. USA 83, 280-284. 5. Ambudkar, S. V., Anantharam, V., & Maloney, P. C. (1990). UhpT, the sugar phosphate antiporter of Escherichia coil, functions as a monomer. J. Biol. Chem. 265, 12287-12292. 6. Anantharam, V., Allison, M. J., & Maloney, P. C. (1989). Oxalate:formate exchange: the basis for energy coupling in Oxalobacter. J. Biol. Chem. 264, 7244-7250. 7. Aquila, H., Link, T., & Klingenberg, M. (1987). Solute carriers involved in energy transfer of mitochondria form a homologous protein family. FEBS Lett. 212, 1-9.

126

ATUL VARADHACHARY and PETERC. MALONEY

8. Arakawa, T., & Timasheff, S. N. (1985). The stabilization of proteins by osmolytes. Biophys. J. 47, 411--414. 9. Baetz, A. L., & Allison, M. J. (1989). Purification and characterization of oxalyI-CoA-decarboxylase from Oxalobacterformigenes. J. Bacteriol. 171, 2605--2608. 10. Bennett, R. L., & Malamy, M. H. (1970). Arsenate resistant mutants of Escherichia coli and phosphate transport. Bioch. Biophys. Res. Commun. 40, 496-503. 11. Carmeli, C., & Lifshitz, Y. (1989). Nucleotide transport in Rhodobacter capsulatus. J. Bacteriol. 171, 6251--6255. 12 Costello, M. J., Escaig, J., Matsushita, K., Viitanen" P., Menick, D. R., & Kaback, H. R. (1987). Purified lac permease and cytochrome o oxidase are functional as monomers. J. Biol. Chem. 262, 17072-17082. 13. Dimroth, P. (1980). A new sodium-transport system energized by the decarboxylation of oxaloacetate. FEBS Le,. 122, 234--236. 14. Dimroth, P. (1990). Mechanisms of sodium transport in bacteria. Phil. Trans. R. Soc. London. B. 326, 465--477. 15. Eiglmeier, K., Boos, W., & Cole, S. T. (1987). Nucleotide sequence and transcriptional start point of the glpT gene of Escherichia coil: extensive sequence homology of the glycerol-3-phosphate transport protein with components of the he xose-6-phosphate transport system. Molec. Microbiol. 1, 251-258. 16. Engel, P., Kramer, R., & Unden, G. (1992). Anaerobic fumarate transport in Escherichia coil by anfnr-dependent decarboxylate uptake system which is different from the aerobic dicarboxylate uptake system. J. Bacteriol. 174, 5333-5539. 17. Essenberg, R. C., & Kornberg, H. L. (1975). Energy coupling in the uptake of hexose phosphates by Escherichia coll. J. Biol. Chem. 250, 939-945. 18. Flugge, U. I., & Helot, H. W. (I 986). Chloroplast phosphate-triose phosphate phosphoglycerate translocator: its identification, isolation and reconstitution. Methods Enzymol. 125, 716-730. 19. Fraenkel, D. G., Falcoz-Kelly, E, & Horecker, B. L. (1964). The utilization of glucose 6-phosphate by glucokinaseless and wild-type strains of Escherichia coll. Proc. Natl. Acad. Sci. USA 52, 1207-1213. 20. Freidrich, M. J., & Kadner, R. J. (I 987). The nucleofide sequence of the uhp region of Escherichia coll. J. Bacteriol. 169, 3556-3563. 21. Goldrick, D., Yu, G. L. Q., Jiang, S. Q., & Hong, J. S. (1988). Nucleotide sequence and transcription start point of the phosphoglycerate transporter gene of Salmonella typhimurium. J. Bacteriol. 179, 3421-3426. 22. Grover, W. H., & Winkler, H. H. (1974). Evidence that glucose 6-phosphate is transported intact in Escherichia coll. Biochim. Biophys. Acta 363, 428--430. 23. Harold, E M., Harold, R. L., & Abrams, A. (1965). A mutant of Streptococcusfaecalis defective in phosphate transport. J. Biol. Chem. 240, 3145-3153. 24. Hayashi, S.-I., Koch, J. P., & Lin, E. C. C. (1964). Active transport ofL-a glycerolphosphate in Escherichia coll. J. Biol. Chem. 239, 3098-3105. 25. Henderson, E J. F. (1990). The homologous glucose transport proteins of prokaryotes and eukaryotes. Res. Micwbiol. 141,316-328. 26. Island, M. D., Wei, B. Y., & Kadner, R. J. (1992). Structure and function of the uhp genes for the sugar phosphate transport system in Escherichia coll. J. BacterioL 174, 2754-2762. 27. Jones, W. J., Nagle, D. P., Jr., & Whitman, W. B. (1987). Methanogens and the diversity of archaebacteria. Microbiol. Rev. 51, 135-177. 28. Klingenberg, M. (1981). Membrane protein oligomeric structure and transport function. Nature 290, 449-454. 29. Klingenberg, M., Hackenherg, H., Kramer, R., Lin, C. S., & Aquila, H. (1980). Two transport proteins from mitochondria. I. Mechanistic aspects of asymmetry of the ADP/ATP transiocator. II. The uncoupling protein of brown adipose tissue nfitochondria.Ann, iV. Y.Acad. Sci. 358, 83-95.

Anion Exchange Mechanisms in Bacteria

127

30. Krause, D. C., Winkler, H. H., & Wood, D. O. (1985). Cloning and expression of the Rickettsia prowazekii ADP/ATP translocator in Escherichia coli. Proc. Natl. Acad. Sci. USA 82, 3015-3019. 31. Kyte, J. (1981). Molecular considerations relevant to the mechanism of active transport. Nature 292, 201-204. 32. Lin, E. C. C., Koch, J. P., Chusen, T. M., & Jorgensen, S. E. (1962). Utilization of L-alpha-glycerophosphate by Escherichia call without hydrolysis. Proc. Natl. Acad. Sci. USA 48, 2145-2150. 33. Luria, S. E. (1960). The bacterial protoplasm: composition and organization. In The bacteria, vol. 1 (I. C. Gunsalus & R. Y. Stainer (Eds.) (pp. 1-34). New York: Academic Press. 34. Maloney, P. C. (1987). Coupling to an energized membrane: the role of ion-motive gradients in the transduction of metabolic energy. In Escherichia coil and Salmonella typhimurium: cellular and molecular biology (E C. Neidhardt, J. L. Ingraham, B. K. Low, B. Magasanik, M. Schaechter, & H. E. Umbarger (Eds.) (pp. 222-243). Washington, DC: American Society for Microbiology. 35. Maloney, P. C. (1990). Resolution and reconstitution of anion exchange reactions. Phil. Trans. R. Sac. London. B. 326, 437--454. 36. Maloney, P. C. (1990). A consensus structure for membrane transport. Res. Microbial. 141, 374-383. 37. Maloney, P. C., & Ambudkar, S. V. (1989). Functional reconstitution of prokaryote and eukaryote membrane proteins. Arch. Biochem. Biophys. 269, 1-10. 38. Maloney, P. C., & Wilson, T. H. (1993). Evolution of membrane carriers. In Molecular biology and function of carrier proteins (L. Reuss, J. R. Russell, Jr., & M. L. Jennings (Eds.) (pp. 147-160). New York: Rockefeller University Press. 39. Maloney, P. C., Ambudkar, S. V., Thomas, J., & Schiller, L. (1984). Phosphate: hexose 6-phosphate antiport in Streptococcus lactis. J. Bacterial. 158, 238-245. 40. Maloney, P. C., Ambudkar, S. V., Anantharam, V., Sonna, L. A., & Varadhachary, A. (1990). Anion exchange mechanisms in bacteria. Microbial. Rev. 54, 1-17. 41. Mitchell, P. (1954). Transport of phosphate across the osmotic barrier of Micrococcus p)~genes: specificity and kinetics. 3: Gen. Microbial. 11, 73-82. 42. Mitchell, P. (1979). Compartmentation and communication in living systems. Ligand conduction: a general catalytic principle in chemical, osmotic and chemiosmofic reaction systems. Eur. 3. Biochem. 95, 1-20. 43. Mitchell, P., & Moyle, J. M. (1953). Paths of phosphate transfer in Micrococcus pyogenes: phosphate turnover in nucleic acids and other fractions. J. Gen. Microbial. 9, 257-272. 44. Molenaar, D., Bosscher, J. S., ten Brink, B., Driessen, A. J. M., & Konings, W. N. (1993). Generation of a proton-motive force by histidine decarboxylation and electrogenic histidine/histamine antiport in Lactobacillus buchneri. J. Bacterial. 175, 2864-2870. 45. Newman, M. J., & Wilson, T. H. (1980). Solubilization and reconstitution of the lactose transport system from Escherichia coll. J. Biol. Chem. 255, 10583-10586. 46. Nicholis, D. G. (1982). Bioenergetics. An introduction to the chemiosmotic theory. Orlando, FL: Academic Press. 47. Olsen, E. B., Russell, J. B., & Henick-Kling, T. (1991). Electrogenic L-malate transport by Lactobacillus plantarum: a basis for energy derivation from malolactic fermentation. J. Bacterial. 173, 6199-6206. 48. Pogell, B. M., Malty, B. R., Frumkin, S., & Shapiro, S. (1966). Induction of an active transport system for glucose 6-phosphate in Escherichia coil. Arch. Biochem. Biophys. 116, 406--415. 49. Poolman, B., Molenaar, D., Smid, E. J., Ubbink, T., Abee, T., Renault, P. P., & Konings, W .N. (1993). Malolactic fermentation: electrogenic malate uptake and malate/iactate antiport generate metabolic energy. J. Bacterial. 170, 6030-6037. 50. Racker, E. (1985). Reconstitution of transporter, receptors, and potlwlogical states. Orlando, FL: Academic Press. 51. Roberts, R. B., & Roberts, I. Z. (1950). Potassium metabolism in Escherichia coll. I11.Imerrelationship of potassium and phosphorous metabolism. J. Cell. & Compar. Physiol. 36, 15-39.

128

ATUL VARADHACHARY and PETERC. MALONEY

52. Rosenberg, H., Gerdes, R. G., & Chegwidden, K. (1977). Two systems for the uptake of phosphate by Escherichia coil J. Bacteriol. 131,505-511. 53. Rosenberg, H., Gerdes, R. G., & Harold, E M. (1979). Energy coupling of inorganic phosphate in Escherichia coil KI2. Biochem. J. 178, 133-137. 54. Ruan, Z-S., Anantharam, V., Crawford, I. T., Ambudkar, S. V., Seung, Y. R., Allison, M. J., & Maloney, P. C. (1992). Identification, purification and recomfitution of OxlT, the oxalate:formate antiport protein of Oxalobacterformigenes. J. Biol. Chem. 267, 10537-10543. 55. Rudnick, G. (1986). ATP-driven 14+ pumping into intracellular organelles. Annu. Rev. Physiol. 48, 403-413. 56. Sonna, L. A., & Malone),, P. C. (1988). Identification and functional reconstitufion of phosphate:sugar phosphate antiport from Staphylococcus aureus. J. Membr. Biol. 101,267-274. 57. Sonna, L. A., Ambudkar, S. V., & Malone),, P. C. (1988). The mechanism of glucose 6-phosphate transport by Escherichia coil J. Biol. Chem. 263, 6625-6630. 58. Tokuda, H., & Unemoto, T. (1982). Characterization of the respiration-dependent Na+ pump in the marine bacterium Hbrio alginolyticus. J. Biol. Chem. 257, 10007-10014. 59. Varadhachary, A., & Maloney, P. C. (1990). A rapid method for reconstitution of bacterial membrane proteins. Mol. Microbioi. 4, 1407-1411. 0. Varadhachary, A., & Maloney, P. C. (1991). Reconstitution of the phosphoglycerate transport protein of Salmonella typhimurium. J. Biol. Chem. 266, 130-135. 61. WiUsky, G. R., & Malamy, M. H. (1980). Characterization of two genetically separable inorganic phosphate transport systems in Escherichia coll. J. Bacteriol. 144, 366-374. 62. Winkler, H. H. (1973). Energy coupling of the hexose phosphate transport system in Escherichia coli. J. Bacteriol. 116, 203-209. 63. Winkler, H. H. (1973). Distribution of an inducible hexose-phosphate transport system among various bacteria. J. Bacteriol. 116, 1079-1081. 40 Yan, R.-T., & Maloney, P. C. (1993). Identification of a residue in the translocation pathway of a membrane carrier. Cell 75, 37--44. 65. Yancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., & Somero, G. N. (1982). Living with water stress: evolution of osmolyte systems. Science 217, 1214-1222.

HELIX PACKING IN THE C-TERMINAL HALF OF LACTOSE PERMEASE

H. Ronald Kaback, Kirsten Jung, Heinrich Jung, Jianhua Wu, Gilbert G. PrivY, and Kevin Zen

I. Introduction . II. III. IV. V.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Secondary Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Functional Interactions Between Putative Intramembrane Charged Residues Use of Site-Directed Fluorescence labeling to Obtain Proximity Relationships S u m m a r y and Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

129 131 .134 . 138 141 143

I. I N T R O D U C T I O N Although the driving force for a variety of seemingly unrelated phenomena (e.g., secondary active transport, oxidative phosphorylation, and rotation of the bacterial flagellar motor) is a bulk-phase, transmembrane electrochemical ion gradient, the molecular mechanism(s) by which free energy stored in such gradients is transduced into work or into chemical energy remains enigmatic. Nonetheless, gene

Advances in Cell and Molecular Biology of Membranes and Organelles Volume 4, pages 129-144. Copyright 9 1995 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-$$938-924-9 129

KABACK, )UNG, )UNG, WU, PRIVI~, and ZEN

130

H§

A

B

c

F~ure 1. H§ symport in E. coll. (A) Lactose accumulation in response to A~H+ (interior negative and alkaline) generated either by respiration or ATP hydrolysis. (B) Uphill H + transport in response to an inwardly directed lactose gradient. (C) Uphill H + transport in response to an outwardly directed lactose gradient.

sequencing and analyses of deduced amino acid sequences indicate that many biological machines involved in energy transduction, secondary transport proteins in particular (24, 43), fall into large families encompassing proteins from archaebacteria to the mammalian central nervous system, thereby suggesting that the family members have common basic structural features and mechanisms of action. This contribution will concentrate on recent observations with the lactose (lac) ~ permease of Escherichia coli as a paradigm for secondary transport proteins that catalyze ion-gradient coupled active transport. Accumulation of [3ogalactosides against a concentration gradient in E. coil is catalyzed by lac permease, a hydrophobic polytopic cytoplasmic membrane protein that carries out the coupled translocation of a single [3-galactoside with a single H § (i.e., 13-galactoside/H§ symport or cotransport) (see 31, 32, 33 for reviews). Physiologically, the proton electrochemical gradient across the cytoplasmic membrane (A~x§ is interior negative and/or alkaline, and the permease utilizes free energy

Helix Packing of Lac Permease

131

released from downhill translocation of H + to drive accumulation of [3-galactosides against a concentration gradient (Figure 1). In the absence of (A~H+), lac permease catalyzes the converse reaction, utilizing free energy released from downhill translocation of [$-galactosides to drive uphill translocation of H + with generation of a A~H+ the polarity of which depends upon the direction of the substrate concentration gradient. Lac permease is encoded by the lacy gene, the second structural gene in the lac operon, which has been cloned into a recombinant plasmid (66) and sequenced (6). By combining overexpression of lacY with the use of a highly specific photoaffinity probe (34) and reconstitution of transport activity in artificial phospholipid vesicles (i.e., proteoliposomes) (50), the permease was solubilized from the membrane, purified to homogeneity (51, 22, 70, 72) and shown to catalyze all the translocation reactions typical of the ~3-galactoside transport system in vivo with comparable turnover numbers (46, 69). Therefore, the product of lacY gene is solely responsible for all of the translocation reactions catalyzed by the [3-galactoside transport system.

II. SECONDARY STRUCTURE Circular dichroic measurements on purified lac permease demonstrate that the protein is about 80% helical, an estimate consistent with the hydropathy profile of the permease, which suggests that approximately 70% of its 417 amino acid residues are found in hydrophobic domains with a mean length of 24 :i: 4 residues (21). Based on these findings, it was proposed that the permease is composed of a hydrophilic N-terminus followed by ] 2 hydrophobic segments in a-helical conformation that traverse the membrane in zigzag fashion connected by hydrophilic domains (loops) with a 17-residue C-terminal hydrophilic tail (Figure 2). Support for general features of the model and evidence that both the N and C termini are exposed to the cytoplasmic face of the membrane were then obtained from laser Raman spectroscopy (71), immunological studies (11, 12, 9, 62, 63, 61, 26, 17), limited proteolysis (23, 65) and chemical modification (53). However, none of these approaches differentiates between the 12-helix motif and other models containing l0 (71) or 13 (5) transmembrane domains. Calamia and Manoil (7) have provided elegant, unequivocal support for the topological predictions of the 12-helix model by analyzing an extensive series of lac permease-alkaline phosphatase (lacY-phoA) fusion proteins. Under normal conditions, alkaline phosphatase is synthesized as an inactive precursor in the cytoplasm of E. coli with an N-terminal signal sequence that directs its secretion into the periplasm where it dimerizes to form active enzyme. If the signal sequence is deleted, the enzyme remains in the cytoplasm in an inactive form. When alkaline phosphatase devoid of the signal sequence is fused to the C-termini of fragments of a cytoplasmic membrane protein #1 vivo, enzyme activity reflects the ability of the N-terminal portions of the fusions to translocate the enzyme to the outer surface

1.

K

N

,vwl

'I'

N

..s 'I'

V

O

1;

v

OH

K

1,

I d S

1.


I:

Figure2. Secondary-structureof lac permease. The model is basedprimarily on hydropathy analysis (Foster et at., 1983). The single-letter amino-acid code is used, and Asp237, Asp240, Glu269, Arg302, Lys319, His322, Glu325 and Lys358 are highlighted. Hydrophobic transmembrane helices are shown in boxes, and the topology of helix VII was modified according to results obtained from a series of lacy-phoA fusions in this domain, (M. L. Ujwal and H. R. Kaback, unpublishedobservations).

Helix Packing of Lac Permease

133

of the membrane (34). Alkaline phosphatase activity in cells independently expressing each of 36 lacY-phoA fusions exclusively supports the topological predictions of the 12 transmembrane domain model. In addition, Calamia and Manoil (7) showed that the alkaline phosphatase activity of fusions engineered at every third amino acid residue in putative helices III and V (Figure 2) increases abruptly as the fusion junction proceeds from the 8th to the 1l th residue. Thus, approximately half a transmembrane domain is needed to translocate alkaline phosphatase through the membrane to the external surface. When fusions are constructed at each amino acid residue in putative helix X, the data are in good agreement with the model. The alkaline phosphatase activity of the fusions increases sharply as the junction proceeds from Phe320 to His322 (Figure 2; M. L. Ujwal, E. B ibi, C. Manoil and H. R. Kaback, unpublished information), suggesting that these residues are located at or near the middle of helix X. The lacY gene has been restricted into two approximately equal-size fragments that were subcloned individually or together under separate lac operator/promoters (4). Under these conditions, lac permease is expressed in two portions: (1) the N terminus, the first 6 putative transmembrane helices and most of putative loop 7; and (2) the last 6 putative transmembrane helices and the C terminus. Cells expressing both fragments transport lactose at about 30% the rate of cells expressing intact permease to a comparable steady-state level of accumulation. In contrast, cells expressing either half of the permease independently do not transport lactose. [35S]Methionine labeling and immunoblotting experiments demonstrate that intact permease is completely absent from the membrane of cells expressing lacy fragments either individually or together. Thus, transport activity must result from an association between independently synthesized portions oflac permease. When the gene fragments are expressed individually, the N-terminal portion of the permease is observed sporadically and the C-terminal portion is not observed. When the gene fragments are expressed together, polypeptides identified as the N- and C-terminal moieties of the permease are found in the membrane. The results indicate that the N- or C-terminal halves of lac permease are proteolyzed when synthesized independently and that association between the two complementing polypeptides leads to a more stable, catalytically-active complex. Additional experiments demonstrate that co-expression of independently cloned fragments of the lacY gene encoding N2 and Cl0 (73), Nl and C~l or N7 and C5 (K. Zen, E. McKenna, D. Hardy, E. Bibi and H. R. Kaback, in preparation) also form stable molecules in the membrane that interact to form functional permease, while expression of the fragments by themselves yields polypeptides that are relatively unstable and exhibit no transport activity. Thus, lacy gene fragments encoding duplex permeases split in either periplasmic or cytoplasmic loops are able to form functional complexes. In contrast, lacY gene fragments encoding duplex permeases split in putative transmembrane domains III or VII (Figure 2) are unable to form functional complexes, implying that the "split permease approach" might be useful for approximating helical

134

KABACK,JUNG,JUNG,WU, PRIVY,and ZEN

boundaries. In addition, the demonstration that polypeptides corresponding to N j and C li form a relatively stable, functional complex makes it unlikely that the N terminus of the permease inserts into the membrane as a helical hairpin. Purified lac permease reconstituted into proteoliposomes exhibits a cleft (14, ! 5), an observation also documented by Li and Tooth (42). The presence of a solvent~ filled cleft in the molecule may be important with regard to the mechanism of [~-galactoside/H§ symport, as the barrier within the permease may be thinner than the full thickness of the membrane. Therefore, the number of amino acid residues directly involved in translocation may be fewer than required for lactose and H § to traverse the entire thickness of the membrane. A solvent-filled cleft may also r~resent a caveat with respect to interpretation of spectroscopic experiments designed to test the accessibility of specific regions of the protein to the aqueous phase. The remainder of this discussion deals with recent experiments that utilize site-directed mutagenesis and protein engineering combined with fluorescence spectroscopy to obtain information on the tertiary structure of the C-terminal half of lac permease. Other aspects of structure and function will not be covered, but a noteworthy and surprising conclusion derived from extensive site-directed mutagenesis studies (reviewed in 33) is that very few of the amino acid residues in the permease are directly involved in the mechanism of secondary active transport.

III. FUNCTIONAL INTERACTIONS BETWEEN PUTATIVE INTRAMEMBRANE CHARGED RESIDUES In 1991, King, Hansen, and Wilson (35) found that lac permease mutants with neutral amino acid substitutions for Lys358 or Asp237 (Thr or Ash, respectively) do not catalyze active transport. Second-site suppressor mutations of K358T2 exhibit neutral amino acid replacements for Asp237 (Ash, Gly or Tyr), while suppressors of D237N have Gin in place of Lys358. It was proposed that Asp237 and Lys358 interact via a salt bridge, thereby neutralizing each other. Presumably, neutral replacement of either charged residue individually causes a functional defect because of the remaining unpaired charge, while neutral substitutions for both residues do not inactivate because the unpaired charge is removed. Consequently, the secondary-structure model proposed for the permease (21) was altered to accommodate a putative salt bridge between Asp237 and Lys358 in the low dielectric of the membrane by moving Asp23? from the hydrophilic domain between helices VII and VIII (35) into the middle of transmembrane helix VII. As discussed below, spectroscopic studies and a series of lacF-phoA fusions in helix VII suggest that Asp237, as well as Asp240, may be close to the membrane-water interface. Therefore, these residues are placed near the N terminus of putative helix VII in Figure 2. In any event, the presence of this interaction clearly necessitates

Helix Packing of Lac Permease

135

modification of the secondary-structure model as predicted from hydropathy profiling. As part of an extensive site-directed mutagenesis study with an engineered form of lac permease that is functional but devoid of Cys residues (C-less permease; 68), putative intramembrane residues Asp237, Asp240, Glu269, Arg302, Lys319, His322, Glu325 and Lys358 were systematically replaced with Cys (Figure 2). Individual replacement of any of these residues essentially abolishes lactose accumulation against a concentration gradient (58). Starting with the single-Cys mutants D237C and K358C, a double-Cys mutant was constructed containing Cys replacements for both Asp237 and Lys358 in the same molecule. D237C/K35 8C transports lactose at about half the rate of C-less permease to virtually the same steady-state. Remarkably, replacement of Asp237 and Lys358, respectively, with Ala and Cys or Cys and Ala or even interchanging Asp237 with Lys358 causes relatively little change in activity. Subsequently, the side chains of Asp237 and/or Lys358 were extended by replacement with Glu and/or Arg or by site-specific derivatization of single-Cys replacement mutants (19, 59). Iodoacetic acid was used to carboxymethylate Cys or methanethiosulfonate derivatives (1) were used to attach negatively charged ethylsulfonate or positively charged ethylammonium groups. Replacement of Asp237 with Glu, carboxymethyl-Cys or sulfonylethylthio-Cys yields active permease with Lys or Arg at position 358. Similarly, the permease tolerates replacement of Lys358 with Arg or ammoniumethylthio-Cys with Asp or Glu at position 237. Moreover, permease with Lys, Arg, or ammoniumethylthioCys in place of Asp237 is highly active when Lys358 is replaced with Asp or Glu, confirming the conclusion that the polarity of the charge interaction can be reversed without loss of activity and demonstrating that the distance between positions 237 and 358 can be extended by up to five bond-lengths (59). Therefore, neither Asp237, Lys358 nor the interaction between these residues is important for permease activity, and Asp237 and Lys358 must interact in a highly "flexible" manner to form a salt bridge. Lac permease mutants in the charge pair Asp237-Lys358 are inserted into the membrane at wild-type levels if the charge pair is maintained with either polarity. On the other hand, disruption of the interacting pair often causes a marked decrease in the amount of protein inserted into the membrane, suggesting a role for the salt bridge in permease folding and/or stability (19, 59). Significantly, as opposed to certain C-terminal truncation mutants that are proteolyzed after insertion (57, 47, 48), these permease mutants may be degraded prior to or during insertion into the membrane, as the mutant proteins are inserted into the membrane in a stable form if they are overproduced at a high rate from the T7 promoter (19). In any case, the observations raise the possibility that Asp237 and Lys358 may interact in a folding intermediate, but not in the mature molecule. As discussed above, however, inactive single mutants with Cys in place of Asp237 or Lys358 regain full activity upon carboxymethylation or treatment with methanethiosulfonate ethylammoniun, respectively, which restores a negative charge at position 237 or a positive charge at

136

KABACK,JUNG,JUNG, WU, PRIVI~,and ZEN

position 358, and similar results are obtained when the charges are reversed. Therefore, it seems very likely that although neither residue nor the putative charge pair is important for activity, the interaction between Asp237 and Lys358 plays a role in folding/stability, and the residues maintain proximity in the mature permease. The results are also interesting because they suggest that there must be a step(s) between translation of the permease and its insertion into the membrane. Furthermore, the findings raise the possibility that the C-terminal half of the permease may be inserted post translationally into the membrane (i.e., helix VII with Asp237 may have to wait for helix XI with Lys358 before insertion can take place). To test the possibility that other charged residues in transmembrane helices are neutralized by charge-pairing ("charge-pair neutralization"), 13 additional double mutants were constructed in which all possible interhelical combinations of negative and positively charged residues were replaced pair-wise with Cys (58). Out of all the combinations of double-Cys mutants, only D240C/K319C exhibits significant transport activity. However, the functional interaction between Asp240 and Lys319 is different phenomenologically from Asp237-Lys358. Thus, D24(~/K319C catalyzes lactose transport at about half the rate of C-less to a steady-state level of accumulation that is about 25-30% of the control. Moreover, although significant activity is observed with the double-Ala mutant or with the two possible Ala-Cys combinations, interchanging Asp240 and Lys319 completely abolishes active transport. In addition, replacement of Asp240 with Glu abolishes lactose transport, and permease with carboxymethyl-Cys at position 240 is inactive when paired with Lys319, but exhibits significant activity with Arg319. Sulfonylethylthio-Cys substitution for Asp also results in significant transport activity. Permease with Arg or ammoniumethylthio-Cys in place of Lys319 exhibits high activity with Asp240 as the negative counterion, but no lactose transport is observed when either of these modifications is paired with Glu240. Finally, mutations in Asp240-Lys319 do not effect insertion of the permease into the membrane. Therefore, although neither Asp240 nor Lys319 per se is mandatory for active transport, the interaction between this pair of charged residues exhibits more stringent properties than Asp237-Lys358, and the polarity of the interaction appears to be important for activity. Lee, Hwang, Hansen, and Wilson (41) have also provided evidence that there is an interaction between Asp240 and Lys319 by using a different experimental approach. These workers replaced Asp240 with Ala by site-directed mutagenesis and found little or no active sugar transport. Two second-site revertants were then isolated, one with Gin in place ofLys319 and the other with Val in place of Gly268. The double mutants exhibit little or no accumulation of sugar, but manifest significant rates of lactose entry down a concentration gradient. Although suppression of D240A by G268V was not explained (see below), the properties of the double mutant D240A/K319Q are consistent with the observations of Sahin-T6th, Dunten, Gonzalez, and Kaback (58), and Sahin-T6th and Kaback (59). However,

Figure 3. Helical wheel model of putative helices VII-IX in lac permease viewed from the periplasmic surface. In addition to the paired residues described in the text (i-e., R302 and E325, E325 and H322, H322 and E269, D240 and K319, 0237 and K358), also shown are G268 and G262 which are second-site suppressors for 0 2 4 0 and E325, respectively, and V315.

138

KABACK,lUNG, lUNG, WU, PRiVI~,and ZEN

the suggestion that Asp240-Lys319 may play a direct role in H + translocation (41) is inconsistent with the observation that various combinations of Cys-Ala replacements at these positions yields permease that is able to concentrate lactose against a concentration gradient (58). The charge-pair neutralization approach is dependent upon permease activity, and it should be stressed that the approach will not reveal charge-paired residues if they are essential for activity. In this context, it is noteworthy that double-Cys mutants involving residues suggested to be H-bonded and directly involved in lactose-coupled H + translocation and/or substrate recognition [i.e., Arg302, His322 and Glu325 (32, 33)], as well as Glu269, which has also been shown to be an important residue (27, 67), are defective with respect to active lactose transport. On the other hand, as discussed below, it is likely that certain pairs of these residues are in close proximity in the tertiary structure of the permease and probably interact. Despite the indication that Asp237-Lys358 and Asp240-Lys319 may participate in salt bridges, the evidence for the interactions is indirect. Therefore, other approaches are required to determine the location of the residues relative to the plane of the membrane and to demonstrate directly that the pairs are in close proximity. C-less permease mutants containing double-Cys or paired Cys-Ala replacements are particularly useful in this respect. Efforts to estimate the accessibility of Cys residues at positions 237 and 358 with water- or lipid-soluble sulfhydryl reagents suggest that Cys residues at the two positions are accessible to both types of reagents, although the lipid-soluble reagents are relatively more effective (19). Other preliminary experiments with permease mutants specifically labeled at positions 237 or 358 with paramagnetic or fluorescent probes (29), as well as a series of alkaline phosphatase fusions in helix VII (M. L. Ujwal, H. Jung andH. R. Kaback, unpublished observations), are consistent with placement of both residues near the interface at the external surface of the membrane rather than in the middle of helix VII (Figure 2). Efforts to demonstrate disulfide bond formation directly by oxidation of appropriate double-Cys mutants are in progress. In summary, Asp237-Lys358 and Asp240-Lys319 interact functionally, and it is reasonable to suggest that both pairs of residues are in close proximity. It follows that putative helix VII (Asp 237 and Asp240) may neighbor helices X (Lys319) and XI (Lys358) in the tertiary structure of the permease (Figure 3).

IV. USE OF SITE-DIRECTED FLUORESCENCE LABELING TO OBTAIN PROXIMITY RELATIONSHIPS Since hydrophobic membrane proteins are notoriously difficult to crystallize (18), a high-resolution structure of lac permease is not available, and development of alternative methods for obtaining structural information is essential. With this objective in mind, pairs of charged residues in putative transmembrane helices of C-less permease were replaced with Cys in order to provide specific sites for

Helix Packing of Lac Permease

139

labeling with N-(l-pyrene)maleimide (pyrene maleimide) (30). Pyrene maleimide was chosen as a fluorophore because two pyrene moieties can form an excited-state dimer (excimer) that exhibits a unique emission maximum at approximately 470 nm if the conjugated ring systems are within about 3.5 ,/~ of each other and in the correct orientation (39) and because this fluorophore has been used previously to study proximity relationships between Cys residues (e.g., 3, 28, 43, 64). Initially, Arg302 (putative helix IX), His322 (helix X) and Glu325 (helix X) were studied because these residues are important for activity and postulated to interact (54, 8, 10, 55, 56, 40, 60). To test the proximity of the residues, the double-Cys mutants H322C/E325C, R302C/H322C and R302C/E325C and the corresponding single-Cys mutants were constructed with the biotinylation domain from a Klebsiella pneumoniae oxalacetate decarboxylase (16) in the middle cytoplasmic loop, purified by avidin-affinity chromatography (13), labeled with pyrene maleimide, reconstituted into proteoliposomes and subjected to fluorescence studies

(30). With the double mutant H322C/E325C, a typical pyrene excimer fluorescence emission band is observed at about 470 nm. The observation is consistent with the postulate that His322 and Glu325 are in a portion of the permease that is 0t-helical, since the residues are predicted to lie on the same face of an or-helix. Mutant R302C/E325C labeled with pyrene maleimide also exhibits strong excimer emission band. However, insignificant excimer fluorescence is observed with mutant R302C/H322C. As a whole, the data indicate that helix IX is in close proximity to helix X, but Arg302 appears to be close to Glu325 rather than His322 (see 49). Since hydrophobic proteins have a strong tendency to aggregate, it is important to determine whether the excimer fluorescence observed with lac pennease results from an intramolecular interaction within single molecules or from an intermolecular interaction between two permease molecules. The following experiments were performed to test these alternatives: (1) Each single-Cys mutant was analyzed, and no excimer fluorescence is observed at 470 nm. (2) Single-Cys mutants R302C and E325C were purified separately, labeled, mixed, and reconstituted into proteoliposomes; no excimer fluorescence is observed, and the fluorescence emission spectrum is identical to those obtained from the unmixed single-Cys mutants. (3) If excimer fluorescence results from an intermolecular interaction, intensity at 470 nm should be inversely related to lipid:protein ratio. Therefore, pyrene maleimidelabeled R302C/H322C permease was reconstituted at lipid:protein ratios of 128:1, 385:1 and 10~:1 (w/w). All three samples exhibit no excimer fluorescence. Based on these three control experiments, it seems highly likely that excimer fluorescence observed with E325C/H322C and R302C/E325C permeases results from intramolecular interactions between pyrene molecules attached to Cys residues within single molecules. As discussed above, studies on second-site suppressor mutations (35, 41) and site-directed mutagenesis studies on C-less permease (58, 19, 59) indicate that helix VII (Asp237 and Asp240) is close to helices X (Lys319) and XI (Lys358).

140

KABACK, JUNG, JUNG, WU, PRIVY,and ZEN

Combining the results obtained from pyrene fluorescence with the previous studies leads to the model shown in Figure 3 where helices VII, IX, X and XI are shown to interact via ion pairs between R302 and E325, K319 and D240, and K358 and D237. Although not shown, a side-view projection of the helices reveals that these residues are at ~ x i m a t e l y the same level with respect to depth in the membrane (30). In addition to neutral substitutions for Lys319, as mentioned above, a second-site suppressor mutant for Asp240 has been described (41) with Val in place of Gly268. Also, recent experiments (J. Wu and H. R. Kaback, unpublished information) demonstrate that the phenotype of E325C is suppressed by substitution of Ser for Gly262. These observations suggest that putative helix VIII may interact with helices VII and X so as to bring Glu269 into proximity with His322 (Figure 3). Regarding Glu269, by carrying out cassette mutagenesis of/acY DNA encoding putative helix VIII, Hinlde, Hinkle, and Kalack (27) identified mutants that retain the ability to catalyze lactose accumulation. A stripe of residues, largely on one side of helix VIII opposite Glu269, was shown to tolerate mutations with relatively little effect on activity, suggesting that this mutable strip of low information content is probably in contact with the membrane phospholipids. No active mutants in Glu269 were identified, however, and this residue was subjected to site-directed mutagenesis (67). Permease with Cys or Gin in place of Glu269 is completely inactive, while E269D permease is completely defective with respect to lactose transport, but transports [3,D-galactopyranosyl l-thio-~),D-galactopyranoside (TDG) reasonably well. In addition, as noted previously, paired double mutants containing E269C and Cys replacements for each of the other charged residues in transmembrane domains are inactive (58). In brief, therefore, Glu269 plays an important role in transport, but the charge-pair neutralization approach gives no indication as to whether or not Glu269 interacts with another residue. To test the idea that Glu269 interacts with His322, the double-Cys mutant E269C/H322C was constructed (30). The mutant and the corresponding single-Cys mutants containing biotinylation domains in the middle cytoplasmic loop were purified, labeled with pyrene maleimide, and reconstituted into proteoliposomes. Strikingly, pyrene maleimide-labeled E269C/H322C permease exhibits distinct excimer emission, and importantly, no excimer is observed with pyrene maleimidelabeled E269C or H322C nor when the single-Cys mutants are labeled and mixed prior to reconstitution. The results provide a strong indication that Glu269 is close to His322 and imply that helix VIII is close to helix X (Figure 3). Additional support for the model comes from Cys-scanning mutagenesis of helix XI in C-less permease (20). When each residue in helix XI (from Ala347 to Ser366) is replaced with Cys, most of the mutants exhibit highly significant activity, and the singleoCys mutants that are strongly inhibited by N-ethylmaleimide fall on the same face of helix XI (Figure 3). Similarly, Cys-scanning mutagenesis of helix X (60) reveals that with the exception of Va1331, the N-ethylmaleimide-inhibitable Cys-replacement mutants are present on the same face of helix X as Val315, Lys319, His322 and Glu325 (Figure 3). Finally, although highly speculative, Baldwin (2) has

Helix Packingof Lac Permease

141

independently put forward a model of the erythrocyte glucose facilitator in which the arrangement of the helices in the C-terminal half of the molecule are virtually identical to that discussed here for lac pennease. If this is the fact, it would suggest that these two transport proteins are likely to have a similar tertiary structure. Other experiments (K. Jung, H. Jung and H. R. Kaback, in preparation) indicate that excimer fluorescence between pyrene maleimide-labeled Cys residues in transmembrane domains can be used to study certain dynamic aspects of permease folding. The excimer observed with reconstituted pyrene maleimide-labeled R302C/E325C or E269C/H322C permease is markedly diminished by increasing concentrations of sodium dodecylsulfate (up to 0.6%), while excimer fluorescence with the H322C/E325C mutant is unaffected. Apparently, the detergent disrupts tertiary interactions within the permease with little effect on secondary structure. Consistently, the double mutants E269C/H322C and R302C/E325C do not exhibit excimer fluorescence after labeling with pyrene maleimide in octyl-[~,D-glucopyranoside, but do so after reconstitution into proteoliposomes. Although individual replacement of each of the charged residues in transmembrane domains inactivates active lactose transport (52, 54, 56, 8, 10, 55, 40, 36, 37, 38, 58, 60, 41, 20, 67), some of the constructs exhibit ligand-induced conformational alterations (K. Jung, H. Jung and H. R. Kaback, in preparation). Excimer fluorescence in proteoliposomes containing pyrene maleimide-labeled E269C/H322C permease is quenched by TI§ and the effect is strongly attenuated by 5 mM TDG (i.e., the quenching constant for TI+ is decreased from 27 to 10 M-I in the presence of 5 mM TDG). It has also been shown (60) that C-less permease with a single Cys residue in place of Val315 [presumably the N-terminal residue in helix X (Figure 2) which is on the same face as His322 and E325 (Figure 3)] is inactivated by N-ethylmaleimide much more rapidly in the presence of TDG or A~H+. Recently (H. Jung, M. Sahin-T6th and H. R. Kaback, in preparation), V315C permease containing the biotin acceptor domain in the middle cytoplasmic loop (13) has been purified, and the kinetics of pyrene maleimide labeling has been studied in the presence and absence of TDG. The studies confirm the observations described with right-side-out membrane vesicles and imply that ligand binding or A~H+ induces a change in tertiary structure without effecting secondary structure.

V. SUMMARY AND CONCLUDING REMARKS The lac permease of E. coli is providing a paradigm for secondary active transporters that transduce the free energy stored in electrochemical ion gradients into work in the form of a concentration gradient. This hydrophobic, polytopic, cytoplasmic membrane protein catalyzes the coupled, stoichiometric translocation of [3-galactosides and H+, and it has been solubilized, purified, reconstituted into artificial phospholipid vesicles and shown to be solely responsible for ~-galactoside transport. The lac Y gene that encodes the permease has been cloned and sequenced,

142

KABACK,lUNG, IUNG, WU, PRIVI~,and ZEN

and based on spectroscopic analyses of the purified protein and hydropathy profiling of its amino acid sequence, a secondary structure has been proposed in which the protein has 12 transmembrane domains in r configuration that traverse the membrane in zigzag fashion connected by hydrophilic loops with the N and C termini on the cytoplasmic face of the membrane. Unequivocal support for the topological predictions of the 12-helix model has been obtained by analyzing a large number of lac permease--alkaline phosphatase (lacY.phoA) fusions. Secondsite suppressor analysis and application of site-directed mutagenesis and chemical modification to a functional permease devoid of Cys residues (C-less permease) have provided evidence that helix VII is probably close to helices X and XI in the tertiary structure of the permease. Recent experiments in which paired Cys replacements in C-less p e ~ were labeled with pyrene, a fluorophore that exhibits excimer fluorescence when two of the unconjugated ring systems are in close approximation, indicate that His322 and Glu325 are probably located in an 0~-helical region of the p e ~ and that helix IX is probably close to helix X. Based on these and the previous results, helix VIII (Glu269) is suggested to be close to helix X (His322), and site-directed pyrene maleimide labeling experiments are consistent with the postulate. Other findings indicate that conformational changes in the permease can be detected as a result of either ligand binding or imposition of AI~H+. Since many membrane transporters appear to have similar secondary structures based on hydropathy profiling, it seems likely that the basic tertiary structure and mechanism of action of these proteins have been conserved throughout evolution. Therefore, studies on bacterial transport systems, which are considerably easier to manipulate than their eukaryotic counterparts, have important relevance to transporters in higher order systems, particularly with respect to the development of new approaches to stngture-function relationships. Although it is now possible to manipulate membrane proteins to an extent that was unimaginable only a few years ago, it is unlikely that transport mechanisms can be defined on a molecular level without high-resolution structural information. In addition to structure, however, dynamic information is required at high resolution, and as suggested by some of the experiments discussed here, a combination of site-directed mutagenesis and fluorescence spectroscopy may be particularly useful in this regard.

NOTES 1. Abbreviations:lac, lactose; ~q~H+, the Wotonelectrochemicalgradient across the membrane; pyrene maleimide,N-(I-pyrene)maleimide;C-less permease, functionallac pennease devoid of Cys residues; TDG, ~,D-galacto-pyranosyll-thio-~,D-galacto-pyranoside. 2. Site-directedmutantsare designatedas follows:the one-letteramino-acidcode is usedfollowed by a numberindicatingthe positionof the residuein wild-typelac permease.The sequenceis followed by a second letterdenotingthe amino-acidreplacementat this position.

Helix Packin8 of Lac Permease

143

REFERENCES

10. 11. 12. 13. 14. 15. 16. 17.

18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.

Akabas, M. H., Stauffer, D. A., Xu, M., & Karlin, A. (I 992). Science 258, 307-310. Baldwin, S. A. (1993). Biochim. Biophys. Acta 1154, 17-49. Betcher-Lange, S. L., & Lehrer, S. S. (1978). J. Biol. Chem. 253, 3757-3760. Bibi, E., & Kaback, H. R. (1990). Proc. Natl. Acad. Sci. USA 89, 4325--4329. Bieseler, B., Heinrich, P., & Beyreuther, C. (1985). Ann. N Y Acad. Sci. 456, 309-325. BUchel, D. E., Gronenborn, B., & Miller-Hill, B. (1980). Nature (London) 283, 541-545. Calamia, J., & Manoil, C. (1990). Proc. Natl. Acad. Sci. USA 87, 4937-4941. Carrasco, N., Antes, L. M., Poonian, M. S., & Kaback, H. R. (1986). Biochemistry 25, 4486-4488. Carrasco, N., Herzlinger, D., Mitchell, R., DeChiara, S., Danho, W., Gabriel, T. E, & Kaback, H. R. (1984b). Proc. Natl. Acad. Sci. USA 81, 4672-4676. Carrasco, N., Pilttner, I. B., Antes, L. M., Lee, J. A., Larigan, J. D., Lolkema, J. S., Roepe, P. D., & KabacL H. R. (1989). Biochemistt3, 28, 2533-2439. Carrasco, N., Tahara, S. M., Patel, L., Goldkom, T., & Kaback, H. R. (1982). Proc. Natl. Acad. Sci. USA 79, 6894-6898. Carrasco, N., Viitanen, P., Herzlinger, D., & Kaback, H. R. (1984a). Biochemistry 23, 3681-3687. Consler, T. G., Persson, B. L., Jung, H., Zen, K. H., Jung, K., Priv6, G. G., Verner, G. E., & Kaback, H. R. (1993). Prec. Natl. Acad. Sci. USA 90, 6934-6938. Costello, M. J., Escaig, J., Matsushita, K., Viitanen, P. V., Menick, D. R., & Kaback, H. R. (1987). J. Biol. Chem. 262, 17072-17082. Costello, M. J., Viitanen, P., Carrasco, N., Foster, D. L., & KabacL H. R. (1984). J. Biol. Chem. 259, 15570-15586. Cronan, J. E. (1990). J. Biol. Chem. 265, 10327-10333. Danho, W., Makofske, R., Humiec, E, Gabriel, T. F., Carrasco, N., & Kaback, H. R. (1985). In Peptides: structure & function C. M. Deber, V. J. Hruby, and K. D. Kopple (Eds.). (pp. 59--62). Pierce Chem. Co. Deisenhofer, J., & Michel, H. (1989). Science 245, 1463-1473. Dunten, R. L., Sahin-T6th, M., & Kaback, H. R. (1993a). Biochemistry 32, 3139-3145. Dunten, R. L., Sahin-T6th, M., & KabacL H. R. (1993b). Biochemistry 32, 12644-12650. Foster, D. L., Boublik, M., & Kaback, H. R. (1983). J. Biol. Chem. 258, 31-34. Foster, D. L., Garcia, M. L., Newman, M. J., Patel, L., & Kaback, H. R. (1982). Biochemistry 21, 5634-5638. Goldkorn, T., Rimon, G., & Kaback, H. R. (1983). Prec. Natl. Acad. Sci. USA 80, 3322-3326. Henderson, P. J. F. (1990). Journal of Bioenergeu'cs and Biomembranes 22, 525-569. Herzlinger, D., Carrasco, N., & Kaback, H. R. (1985). Biochemistry 24, 221-229. Herzlinger, D., Viitanen, P., Carrasco, N., & KabacL H. R. (1984). Biochemistt3, 23, 3688-3693. Hinkle, P. C., Hinkle, P. V., & Kaback, H. R. (1990). Biochemistry 29, 10989-10994. lshii, Y., & Lehrer, S. S. (1987). Biochemistry 26, 4922-4925. Jung, H., Lopez, C., Altenbach, C., Hubbell, W. L., & Kaback, H. R. (1993a). Biophys. J. 64, A 14 (Abs. MI'MH6). Jung, K., Jung, H., Wu, J., Priv6, G. G., & KabacL H. R. (1993b).Biochemistry 32,12273-12278. Kaback, H. R. (1983). J. Membr. Biol. 76, 95-112. KabacL H. R. (1989). Harvey Lectures 83, 77-105. KabacL H. R. (1992). International Review of Cytology 137A K. W. Jean & M. Friedlander, (Eds.), Academic Press, New York, N.Y. pp. 97-125. Kaczorowski, G. J., Leblanc, G., & Kaback, H. R. (1980). Prec. Natl. Acad. Sci. USA 77, 6319-6323. King, S. C., Hansen, C. L., & Wilson, T. H. (1991). Biochim. Biophys. Acta 1062, 177-186. King, S. C., & Wilson, T. H. (1989a). J. Biol. Chem. 264, 7390-7394. King, S. C., & Wilson, T. H. (1989b). Biochim. Biophys. Acta 982, 253-264.

144

KABACK,JUNG,JUNG, WU, PRIVY,and ZEN

38. King, S. C., & Wilson, T. H. (1990). J. Biol. Chem. 265, 3153-3160. 39. Kinnunen, P. K. J., Koiv, A., & Mustonen, P. (1993). In Fluorescence Spectroscopy. O. S. Wolfbeis, (Ed.). New York: Springer Verlag, pp. 159. 0. Lee, J. A., P0ttner, I. B., Carrasco, N., Antes, L. M., & Kaback, H. R. (1989). Biochemistry 28, 2540--2544. 41. Lee, J.-I., Hwang, P. P., Hansen, C., & Wilson, T. H. (1992). J. Biol. Chem. 267, 20758--20764. 42. Li, J., & Tooth, P. (1987). Biochemistry 26, 4816--4823. 43. L0di, H., & Hasselbach, W. (1988). Biophys. J. 51,513-519. 44. Manoil, C., & Beckwith, J. (1986). Science 233, 1403-1408. 45. Marger, M. D., & Saier, M. H., Jr. (1993). Trends in Biochemical Sciences 18, 13-20. 46. Matsushita, K., Patel, L., Gennis, R. B., & Kaback, H. R. (1983). Proc. Natl. Acad. Sci. USA 80, 4889-4893. 7. McKenna, E., Hardy, D., Pastore, J. C., & Kaback, H. R. (1991). Proc. Natl. Acad. Sci. USA 88, 2969-2973. 48. McKenna, E., Hardy, D., & Kaback, H. R. (1992)../. Biol. Chem. 267, 6471-6474. 49. Menick, D. R., Carrasco, N., Antes, L. M., Patel, L., & Kaback, H. K. (1987). Biochemistry 26, 6638-6644. 50. Newman, M. J., & Wilson, T. H. (1980). J. Biol. Chem. 255, 10583-10586. 51. Newman, M. J., Foster, D. L., Wilson, T. H., & Kaback, H. R. (1981 ). J. Biol. Chem. 256, 11804. 952. Padan, E., Sarkar, H. K., Viitanen, P. V., Poonian, M. S., & Kaback, H. R. (1985). Proc. Natl. Acad. Sci. USA 82, 6765-6768. 53. Page, M. G. P., & Rosenbusch, J. P. (1988)../. Biol. Chem. 263, 15906-15914. 54. Pilttner, I. B., Sarkar, H. K., Poonian, M. S., & Kaback, H. R. (1986). Biochemistry 25, 4483--4485. 55. Ptittner, I. B., & Kaback, H. R. (1988). Proc. Natl. Acad. Sci. USA 85, 1467-1471. 56. Piltmer, I. B., Sarkar, H. K.o Padan, E., Lolkema, J. S., & Kaback, H. R. (1989). Biochemistry 28, 2525-2533. 57. Roepe, P. D., Zbar, R., Sarkar, H. K., & Kaback, H. R. (1989). Proc. Natl. Acad. $ci. USA 86, 3992-3996. 5 8 . Sahin-T6th, M., Dunten, R. L., Gonzalez, A., & Kaback, H. R. (1992). Proc. Natl. Acad. Sci. USA 89, 10547-10551. 59. Sahin-T6th, M., & Kaback, H. R. (1993a). Biochemistry 32, 10027-10035. 60. Sahin-T6th, M., & Kaback, H. R. (1993b). Protein Science 2, 1024-1033. 61. Seckler, R., & Wright, J. K. (1984). Fur. J. Biochem. 142, 269-279. 62. Seckler, R., M0r6y, T., Wright, J. K., & Overath, P. (1986). Biochemistry 25, 2403-2409. 63. Seckler, R., Wright, J. K., & Overath, E (1983). J. Biol. Chem. 258, 10817-10820. 64. Sen, A. C., & Chakrabarti, B. (1990)../. Biol. Chem. 265, 14277-14284. 65. Stochaj, V., Bieseler, B., & Ehring, R. (1986). Fur. J. Biochem. 158, 423-428. 66. "leather, R. M., Miiller-Hiil, B., Abmtsch, U., Aichele, G., & Overath, P: (1978). Mol. Gen. Genet. 159, 239-248. 7. Ujwal, M. L., Sahin-T6th, M., Persson, B., & Kaback, H. R. (1994).Mol. Membr. Biol. 1, 9-16. 6 8 . van lwaarden, P. R., Pastore, J. C., Konings, W. N., & Kaback, H. R. (1991). Biochemistry 30, 9595-9600. 9. Viitanen" P., Garcia, M. L., & Kaback, H. R. (1984). Proc. Natl. Acad. Sci. USA 81, 1629-1633. 70. Viitanen, E, Newman, M. J., Foster, D. L., Wilson, T. H., & Kaback, H. R. (1986). Methods Enzymol. 125, 429. 71. Vogel, H., Wright, J. K., & Jiflmig, E (1985). EMBO J. 4, 3625-3631. 72. Wright, J. K., SecHer, R., & Overath, P. (1986). Ann. Rev. Biochem. 55, 255. C., & Ehring, R. (1990). J. Bacterial. 172, 73. Wrubel, W., Stochaj, U., Sonnewald, U., ~ , 5374-5381.

EXPORT AND ASSEMBLY OF OUTER MEMBRANE PROTEINS IN E. COLt

Jan Tommassen and Hans de Cock

I. II.

Ill.

IV. V.

VI.

......

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Outer Membrane Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Function of Outer Membrane Proteins . . . . . . . . . . . . . . . . . . . B. Structure of Outer Membrane Proteins . . . . . . . . . . . . . . . . . . . Transport Across the Inner Membrane . . . . . . . . . . . . . . . . . . . . . . A. The Export Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Role of SecB in Protein Export . . . . . . . . . . . . . . . . . . . . . . . C. lntragenic Export Information . . . . . . . . . . . . . . . . . . . . . . . Pathway to the Outer Membrane . . . . . . . . . . . . . . . . . . . . . . . . . Insertion and Assembly in the Outer Membrane . . . . . . . . . . . . . . . . . A. Assembly Intermediates, Detected in vivo . . . . . . . . . . . . . . . . . B. Role of LPS and Lipid Biosynthesis . . . . . . . . . . . . . . . . . . . . C. In vitro Reconstitution of the Insertion and Assembly Process . . . . . . . D. Sorting Signals in Outer Membrane Proteins . . . . . . . . . . . . . . . . Conclusions and Future Prospects . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

,,

,

Advances in Cell and Molecular Biology of Membranes and Organeiles Volume 4, pages 145-173. Copyright 9 1995 by JAI Prms Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-924-9 145

146 146 146 148 150 150 150 152 154 155 155 158 159 162 165 166

146

JAN TOMMASSEN and HANS DE COCK I.

INTRODUCTION

The outer membrane of Escherichia coil is an extremely asymmetric bilayer containing phospholipids and lipopolysaccharides (LPS) only in the inner and outer monolayer, respectively (77). It protects the cell by forming a barrier for harmful compounds, like bile salts and enzymes, and prevents periplasmic proteins leaking into the medium. The outer membrane also contains a number of proteins (OMPs). After their synthesis in the cytoplasm, OMPs have to pass through the inner membrane before reaching their destination. The fact that these proteins don't end up in the inner membrane suggests either that these proteins have a structure that is incompatible with stable insertion into this membrane or that they somehow bypass this membrane on their way out. Indeed, OMPs have a completely different structure as compared with inner membrane proteins. In this chapter, we will review the current knowledge on the biogenesis of OMPs. First, we will introduce the leading characters in this process and discuss their structure within the outer membrane, since this is important to understand the assembly process. For a more detailed description of the structure of outer membrane porins, we refer to the chapter of G. Schulz in this book. For the transport across the inner membrane, OMPs follow the same pathway as periplasmic proteins (59). This pathway will be discussed here only briefly, with only some special attention to the role of SecB protein, which is of special importance in the export of OMPs. For more details we refer to the chapters of K. Ito and of S. Mizushima. The focus of this chapter will be on the folding process of OMPs and on their insertion and assembly into the outer membrane. We will restrict the discussion to integral outer membrane proteins. For the biogenesis of lipoproteins, the reader is referred to a recent review (55).

II.

OUTER MEMBRANE PROTEINS

A. Function of Outer Membrane Proteins Most OMPs of E. coil are involved in the uptake of nutrients (Table 1). The porins, OmpF, OmpC and PhoE, form general diffusion pores in the outer membrane through which hydrophilic solutes with molecular masses up to about 600 Da can pass (10, 77, 90). These proteins, which exhibit extensive sequence similarity (85), form homo- or heterotrimers (46). Expression of the OmpF and OmpC proteins is regulated by the osmolarity of the growth medium (130), whereas the synthesis of PhoE is induced by phosphate-starvation (97). In contrast to the OmpF and OmpC pores, which are cation-selective, the PhoE pores are anion-selective (11). In addition to the general porins, the outer membrane contains proteins that form specific channels (90). These proteins have no sequence similarity to the porins. They also function as general diffusion pores, but in addition, they contain a specific

Outer Membrane Proteins in E. coli

147

Table 1. Survey of E. coli Outer Membrane Proteins and Their Functions a Protein

Function

OmpF

general pore, cation-selective

OmpC PhoE LamB

general pore, cation-selective general pore, anion-selective specific channel for maltoseand maltodextrins specific channel for nucleosides receptor for vitaminB 12 receptor for Fe3+-ferrichrome receptor for Fe3+-enterochelin pore; integrityof outer membrane phospholipase protease secretion of pill secretion of hemolysin

Tsx BtuB FhuA FepA OmpA PldA OmpT PapC TolC Note: ~

list is not intended to be complete, but only mentions the proteins described in this chapter.

binding site for a certain solute. An example of a protein that forms specific channels is LamB protein. LamB is a trimeric protein in the outer membrane, like the porins, and contains a binding site for maltodextrins (12). It facilitates the diffusion of maltose and maltodextrins and of some other sugars (64, 123). Another example is Tsx protein, which has a binding site for nucleosides (81). Another category of transport proteins in the outer membrane consists of receptors that bind their ligands with high affinity. The subsequent transport of the ligands across the outer membrane requires cytoplasmic membrane energy, which is coupled to the transport process via TonB protein (98). Except for BtuB, which is required for the transport of vitamin B 12 (61), all other known TonB-dependent receptors of E. coli are induced under iron-limitation, and they are involved in the uptake of iron-siderophore complexes (98). Examples are FhuA and FepA, the receptors for ferrichrome and enterochelin, respectively. The physiological function of OmpA protein, a major monomeric outer membrane protein, has remained an enigma for a long time. Together with Braun's lipoprotein, it is involved in maintaining the structural integrity of the outer membrane (115). It was recently demonstrated that this protein has pore properties as well (106, 120). Finally, the outer membrane of E. coli contains a few enzymes. The PldA protein has phospholipase A1 and A2 activities (94, 107), and OmpT is a protease (48). The exact physiological function of these enzymes is not known.

148

JAN TOMMASSEN and HANS DE COCK

B. Structure of Outer Membrane Proteins

Integral inner membrane proteins contain hydrophobic stretches of approximately 20 amino acid residues that span the bilayer in an o~-helical configuration. Inspection of the primary sequences of OMPs does not reveal the presence of such hydrophobic stretches and suggests that these proteins are very hydrophilic. This is probably related to the fact that OMPs have to be transported across the inner membrane before they reach their destination, since hydrophobic stretches would act as stop-transfer sequences (26). Indeed, it has been shown that the introduction of hydrophobic sequences in OMPs interferes with their transport across the inner membrane (2, 5, 80). Much effort has been invested in the understanding of how these apparently hydrophilic OMPs can be accommodated in the hydrophobic environment of the outer membrane. A variety of biophysical measurements, including infrared absorption, high-angle X-ray diffraction, circular dichroism measurements, and Raman spectroscopy, have revealed that the porins, OmpA and LamB, have a predominant 13-structure with many antiparallel ~strands in the membrane (63, 134). Further information on the topology of OMPs has been derived from genetic investigations. These experiments were directed to identify cell-surface-exposed amino acid residues, involved in the binding of bacteriophages (69, 87) or monoclonal antibodies (mAbs) (33, 132). Alternative approaches consisted of insertion of foreign antigenic determinants (2, 22) or protease-sensitive sites (67) into OMPs and determining if these epitopes/sites became accessible in intact cells, which would prove insertion in an exposed loop of the OMPs. After the identification of a few cell-surface-exposed amino acids in PhoE, we have proposed a model for the topology of this protein in the membrane, postulating 16 amphipathic, antiparallel [3-strands spanning the membrane, thereby exposing eight regions at the cell surface (126, 132). This model was largely confirmed by the resolution of the crystal structures of OmpF and PhoE (25), which showed that each monomer of these porins forms a 16-stranded, antipamllel 13-barrel (Figure 1A), with hydrophobic amino acids exposed to the lipids and to the subunit interface. Therefore, it is possible to postulate topology models for OMPs with a fair amount of accuracy on the basis of only a few genetic experiments or from sequence data only. A model, similar to the PhoE model, has been proposed for LamB (23). For the much larger FhuA protein, a 32-stranded [3-barrel has been predicted (67). OmpA consists of two domains, an N-terminal membrane-embedded domain and a C-terminal periplasmic tail (Figure IB) (18). The N-terminal domain is proposed to contain eight antiparallel amphipathic ~3-strands (87, 134). In conclusion, it seems that the majority of OMPs have a similar type of organization within the membrane, with multiple membrane-spanning amphipathic ~-strands. The hydrophobic side of these [3-strands is exposed to the lipidic environment of the membrane, and, in the case of multimeric proteins, to the subunit interface.

Figure 1. (A) 3-D structure of porin (25).Arrows, labeled 1-1 6, represent membranespanning &strands. Cell surface exposed loops are denoted L1-L8 and turns at the periplasmic side of the outer membrane TI -T8. Loop L3 folds inside the f3-barrel. (0) Topology model of the OmpA protein (87, 134).The first 170 amino acids mainly consist of eight &strands travening the outer membrane (OM). The C-terminal phenylalanine (amino acid 170 in OmpA) in the last membrane spanning &strand, conserved in outer membrane proteins, is indicated. The C-terminal 155 amino acids of the OmpA protein are located in the periplasm

(PI.

150

JAN TOMMASSENand HANSDE COCK Iil.

TRANSPORT ACROSS THE INNER MEMBRANE A. The Export Apparatus

At least the initial steps in the export of OMPs and periplasmic proteins appear to be identical. OMPs as well as periplasmic proteins are synthesized as precursors with N-terminal signal sequences, which are essential for export. The signal sequences of OMPs and periplasmic proteins are functionally exchangeable (129). Furthermore, it has been observed that export of hybrid proteins consisting of an N-terminal part of the periplasmic maltose-binding protein (MBP) and the cytoplasmic enzyme [3-galactosidase was initiated but not completed, presumably because the ~-galactosidase moiety of the hybrid protein interferes with the passage of the polypeptide through the inner membrane (9). Expression of such a hybrid protein blocked the export apparatus, resulting in the accumulation of the precursors of both OMPs and periplasmic proteins. Also, the mutational distortion of components of the export apparatus, e.g., by secA mutations (96), simultaneously affected the export of OMPs and periplasmic proteins. Genetic approaches have been used to identify components of the export apparatus. A variety of selection procedures were developed to select for export mutants (for a review, see ref. 108). In this way, six genes, designated secA (prlD), secB, secD, secE secE (priG), and sec Y (prlA ), were identified that encode components of the export apparatus. A multisubunit enzyme designated translocase is central in the export apparatus. Protein purification and reconstitution of the export process showed that SecE and SecY are essential integral inner membrane components of the translocase (20, 124). SecA is a peripheral component with an intrinsic ATPase activity that is stimulated by the binding of a precursor and by membrane vesicles (75). ATP hydrolysis as well as the proton motive force are required for efficient transport (35, 47). SecD and SecF are also integral inner membrane proteins; they contain, however, a large periplasmic domain and were postulated to have a role late in the export process (45). Such a role could not be established in the in vitro reconstitution of the export process (75, 124). Antibodies directed against SecD inhibited the secretion of processed envelope proteins from spheroplasts into the medium (82), suggesting that SecD is required for the release of translocated proteins from the inner membrane.

B. Role of SecB in Protein Export In contrast to the central components of the export apparatus, SecB is required for the export of only a subset of envelope proteins (71). Very interestingly, all OMPs examined thus far are dependent on SecB for efficient export, whereas periplasmic proteins, with the exception of MBP, are not. SecB is a cytoplasmic protein that binds to the mature domains of precursor proteins (24, 30, 70,101). Its exact role in the export process is not yet resolved. Two different but not mutually

Outer Membrane Proteins in E. coli

151

exclusive functions have been proposed, for instance, (1) it might stabilize the precursors in an export-competent conformation (23, 141), and (2) it might be required for proper targeting of the precursors to the membrane components of the export apparatus (122, 140). It is essential that a precursor is maintained in an export-competent (i.e., an unfolded or loosely folded conformation), since folding into a stable tertiary structure prevents its translocation from the cytoplasm (100). Indeed, it has been demonstrated that SecB retards the folding of the MBPprecursor into the native structure of the protein (24, 51). However, it should be noted that MBP might be an exceptional case in being the only periplasmic protein known so far that is dependent on SecB for efficient export. The situation might be different in the case of OMPs, which are unlikely to fold into their native structure in the cytoplasm in the absence of other outer membrane components. Initially, some evidence was published that the precursors of OMPs are also stabilized in an export-competent conformation by SecB. In in vitro translocation experiments, the purified precursor of PhoE protein (prePhoE) remained translocation-competent in the presence of SecB with a functional half-life of several hours, whereas it rapidly lost its translocation competency in the absence of SecB (72). SecB had little, if any, effect on the adoption of secondary and tertiary structure of prePhoE as measured by circular dichroism and fluorescence spectroscopy (! 9). However, it prevented the purified precursor from aggregating after dilution from urea (19), which might fully explain its effect on maintaining the translocation competent state of the precursor. Similar effects of SecB on the folding of the precursor of OmpA have been reported (73). It seems extremely unlikely that in vivo the concentration of the precursors of OMPs in a cell will ever increase to levels that support aggregation, unless in genetically manipulated overproducing systems. Furthermore, the in vitro translocation systems that employ purified precursors diluted from urea are somewhat suspicious, since proteins that are not involved in protein export in vivo have been reported to stabilize the export-competent state of a precursor protein (49). Therefore, the data referenced so far do not allow the conclusion that the role of SecB in the translocation of the precursors of OMPs is to stabilize their export-competent state. Recently, we employed a different in vitro system to study the role of SecB in prePhoE translocation (32). Radioactively labeled prePhoE was synthesized in vitro in an E. coli cell extract, and its posttranslational translocation into inverted inner membrane vesicles was studied in the presence or absence of SecB. The binding of SecB to prePhoE was studied by coimmunoprecipitation. The translocation of prePhoE into the membrane vesicles was much more efficient in the presence of SecB, confirming that SecB is required for efficient transiocation of this OMP. However, the translocation-competency ofprePhoE was rapidly lost in the presence of SecB with a functional half-life of only 14 min, even though the coimmunoprecipitation experiments did not reveal the dissociation of the SecB-prePhoE complexes during the incubations. Apparently, SecB-binding did not stabilize the export-competent state of prePhoE, prePhoE was translocated in the absence of

JAN TOMMASSENand HANSDE COCK

152

. ~ i ~ . ~ ~ ; . : . ~ : - - P r o OmpA .........

:..,~.

,~

9

. ....... ~ . . . .

"

'

"

SOC B +

.:

~ ~ ~ : ~ ' ;:..... : ~" ~ " : : : :

Z p OmpA

Sec B"

O' 1' 2' 5' 10'

Figure 2. Autoradiogram of a gel containing [3SS]-methionine labeled proteins from pulse-labeled cells of secBmutant strain CE1343 (secB-) and wild-type strain CE1344 (secB+). Cells were pulse-labeled for 30 s. The pulse was followed by chase periods of 0, 1, 2, 5 and 10 rain. OmpA proteins were immunoprecipitated and analyzed by SDS-PAGE and autoradiography. The positions of precursor (p) and mature (m) OmpA protein are indicated.

SecB, although less efficiently. Under these conditions, the export competency was lost with the same kinetics (i.e., with a functional half-life of 14 min) as occurred in the presence of SecB (32). From these experiments it can be concluded that SecB does not stabilize the export-competent state of prePhoE, but is probably required for efficient targeting to the membrane components of the export apparatus. This conclusion might be extrapolated to the precursors of other OMPs. It is not yet clear how this targeting function operates. However, direct recognition of SecB by SecA has been demonstrated (52). Thus, it seems likely that a SecB-precursor complex will be more readily bound by SecA than the precursor alone. It should be noted that in vivo experiments underscore the conclusions described above. If the role of SecB is to stabilize the export-competent state of the precursors of outer membrane proteins, one would expect that export-incompetent precursors accumulate in a secB mutant. This is not the case (except under conditions of overproduction). Pulso-chase experiments showed that the processing (and therefore probably export) of the precursor of an outer membrane protein is retarded in a secB mutant, but eventually all the precursors are processed (Figure 2).

C. Intragenic Export Information OMPs and periplasmic proteins are synthesized as precursors with an N-terminal signal sequence. Signal sequences are in general 20-25 amino acid residues long and consist of three domains: (1) an amino-terminal positively charged region, (2) a central hydrophobic part, and (3) a more polar carboxy-terminal region containing the signal peptidase recognition site (136). Signal sequences probably interact via their positively charged N-terminal region with the polar headgroups of anionic

Outer Membrane Proteins in E. coil

153

phospholipids in the membrane, whereas the hydrophobic core inserts between the fatty acyl chains and thus initiates translocation (34). In addition, the signal sequence may interact with the proteinaceous components of the export apparatus, like SecA (5) or the prokaryotic analogue of the eukaryotic signal-recognition particle (78). Whereas a signal sequence is essential for efficient export, it is not always sufficient to mediate this process. For instance, [$-galactosidase was not exported out of the cytoplasm when fused to the signal sequence of LamB (86). It is now generally assumed that [3-galactosidase rapidly adopts its stable tertiary structure in its native environment, the cytoplasm, which prevents it from being exported. However, an alternative or additional interpretation of this result is that the mature domain of a precursor contains export information. The latter hypothesis was tested by creating a series of overlapping deletions coveting the complete mature domains ofprePhoE (14,16) or proOmpA (42). All mutant proteins were exported, showing that no essential export signals are present in the mature domains of these OMPs. However, the efficiency of the in vitro translocation of many mutant prePhoE molecules into inverted inner membrane vesicles appeared to be reduced (30). This result suggests that the mature domain contains information that contributes to the efficiency of the export process. The nature of this information is not yet completely clear, but it probably involves the binding sites for components of the export apparatus, like SecB. The mutational removal of such a binding site has no drastic effect on export, probably because there are multiple SecB-binding sites present in a precursor; for example, four SecB-binding sites have been mapped in the mature domain of prePhoE (30). Alternatively, such deletions prevent the folding of the precursors into a stable, export-incompetent configuration, thus alleviating the need for rapid targeting to the translocase in the membrane by SecB. It should be noted that SecA recognizes, in addition to the signal sequence, elements in the mature domain of a precursor (76). However, such a binding site has not yet been defined in any precursor protein. Possibly, it is located directly adjacent to the signal sequence in the N-terminal region of the mature protein. Mutations in this region had an effect on the efficiency of the in vitro translocation of PhoE without disturbing SecB-binding (30). Furthermore, deletions in the corresponding region of LamB had an effect on export kinetics in vivo (102). The charge of the N-terminal region of the mature domain of a precursor seems to be of importance. A database analysis showed that this region is in general negatively charged (135). Especially position +2 was found to be enriched in acidic residues, whereas a basic residue was never found in this position in wild-type precursors. The introduction of basic residues in this region had a drastic effect on export efficiency (7, 74, 79, 116, 143). How these positively charged residues interfere with translocation is not known. Possibly they prevent proper interaction of the precursor with SecA or the phospholipids in the membrane or they prevent the proper orientation of the signal sequence with respect to the electrochemical gradient across the membrane. The observation that the strength of the inhibitory

JAN TOMMASSEN and HANS DE COCK

154

effect depended on the pKa of the inserted residues (116) suggests that basic residues in this region have to be deprotonated to become translocated.

IV. PATHWAY TO THE OUTER MEMBRANE At some stage, the export pathways of OMPs and periplasmic proteins must diverge. In the simplest model (125), OMPs are completely translocated across the inner membrane like periplasmic proteins. The subsequent insertion into the outer membrane would occur in a separate step (Figure 3). Other models have been proposed according to which OMPs do not actually pass all the way through the cytoplasmic membrane to reach their final location, but rather move by some form of membrane-membrane contact (1/3). The latter type of models is supported by electron microscopic studies (114), which revealed that newly synthesized porins appeared at the cell surface in patches located above fusion sites between the two membranes. However, the nature and even the existence of these fusion sites, the so-called Bayer bridges, has recently been challenged on the basis of improved electron microscopy methods (62). Furthermore, this type of model was initially supported by the results obtained with LamB/~galactosidase hybrid proteins.

--OM PP Sec PMF ~

......

IM

1

.__.J

Figure 3. Model for the biogenesis of outer membrane proteins. (1) Precursor proteins are co- or post-translationally targeted to the inner membrane (IM)-Iocated translocation apparatus (See). Translocation across the IM requires ATP and the protonmotive force. The signal sequence is removed during or after translocation. (2) After release of the mature protein in the periplasm (PP), it will fold into an insertion-competent conformation. (3) Insertion requires hydrophobic interactions between protein and outer membrane components. After insertion into the outer membrane (OM), proteins fold into their final 3-D structure. Porins have to assemble into their trimeric configuration probably after insertion of their folded monomers into the OM.

Outer Membrane Proteins in E. coli

155

Whereas, as discussed in Section IIIA, MBP/~l-galactosidase fusions were not exported to the periplasm because the 13-galactosidase moiety cannot pass the inner membrane (9), similar fusions containing an N-terminal part of the OMP Lamb instead of MBP fractionated with the outer membranes, suggesting that they were exported (50). This result was interpreted to mean that the C-terminus of the fusion proteins, and consequently of OMPs, does not have to be translocated across the inner membrane in order to reach the outer membrane, and therefore favored transport models that implicated a role of membrane contact sites. However, immunoelectron microscopy suggested that the co-fractionation of the fusion proteins with the outer membranes was actually artificial; the fusion proteins were mostly found in the cytoplasm as aggregates and also associated with the inner membrane and to intracytoplasmic membranes, which appeared upon induction of the synthesis of these hybrid proteins (127, 128, 137). Clearly, both the aggregate formation and the appearance of intracytoplasmic membranes may have deceived the interpretation of standard cell fractionations. Consequently, there is no solid experimental evidence that supports the transport of OMPs via membrane contact sites, and therefore, we favor the two-step transport model depicted in Figure 3. Several lines of evidence support this periplasmic pathway: 1. mutations that prevented the assembly of PhoE (14) or OmpA (42) into the outer membrane, resulted in the pedplasmic accumulation of the mutant proteins 2. pulse-chase experiments suggested that FhuA protein passed through a soluble, membrane-free pool after cleavage of the signal sequence and before insertion into the outer membrane (60) 3. after the removal of the outer membrane by EDTA/lysozyme treatment, newly synthesized OmpF porin was secreted by the spheroplasts into the medium (83) 4. OmpE secreted by spheroplasts, could be inserted in vitro into isolated outer membranes (111), suggesting that they were true assembly intermediates 5. treatment of spheroplasts with antibodies directed against SecD prevented the release of processed MBP and OmpA (82) suggesting that the function of SecD is to release periplasmic proteins as well as OMPs from the inner membrane after their translocation.

V. INSERTION AND ASSEMBLY IN THE OUTER MEMBRANE A. Assembly Intermediates, Detected in vivo After their linear translocation across the inner membrane, OMPs have to adopt tertiary structure, insert into the outer membrane and, in many cases, oligomerize (Figure 3). The identification of assembly intermediates is very beneficial to gain

156

JAN TOMMASSEN and HANS DE COCK

insight in this process. Correctly assembled OMPs have several distinct biochemical characteristics and the successive acquirement of these characteristics can be used to monitor the kinetics of the assembly process. In contrast to inner membrane proteins, OMPs are not readily solubilized in detergents like Sarkosyl (39) or Triton X-100 (I 09). Many OMPs, including the porins and LamB, are trimeric proteins that are highly resistant to denaturation. Even in 2% SDS, temperatures above approximately 70 ~ are required to denature these trimers into monomers (see also Figure 5, p. 000). Monomeric OMPs like OmpA (57) and PldA (94) display a particular electrophoretic behavior, designated heat-modifiability. When cell envelope-containing samples are not heated above 60 ~ before electrophoresis, these proteins migrate much faster in SDS-polyacrylamide gels than the fully denatured proteins. Furthermore, OMPs are highly resistant to proteases when they are correctly assembled into the outer membrane, being either totally protected against digestion (88) or yielding few, well-defined degradation products like the membrane-embedded part of OmpA (1119). Finally, monocional antibodies (mAbs) raised against native OMPs recognize often conformational epitopes that are not present in the denatured proteins (131). Pulse-chase experiments have been performed to study the assembly of OmpA (43). An assembly intermediate was detected, designated imp-OmpA (immature processed), which had already been processed by signal peptidase, but had not yet acquired the characteristic heat-modifiability and protease-resistance of native OmpA. This imp-OmpA fractionated like an inner membrane protein. A protein form with properties similar to imp-OmpA accumulated in cells overproducing OmpA, probably because of the limited capacity of the outer membrane to take up the protein. Immuno-electron microscopy revealed that imp-OmpA was localized in the periplasm under these conditions. The imp-OmpAcould be converted in vitro into the heat-modifiable form and displayed phage receptor activity upon addition of LPS, suggesting that imp-OmpA is a true assembly intermediate. Similarly, the assembly of LamB has been studied in pulse-chase experiments (138). Newly synthesized Lamb monomers had a half-life of about 20 seconds and were converted into trimers. However, these trimers were denatured in SDS at temperatures between 60 and 70 ~ and were therefore designated "metastable trimers." The metastable trimers were converted slowly, with a half-life of about 5.7 minutes, into stable trimers, possibly by interaction with other outer membrane components like LPS. These results show that the assembly of LamB into trimers requires multiple steps. Newly synthesized LamB was detected at the cell surface with a delay of 30-50 seconds (I 38), indicating that the metastable trimer intermediate is already inserted into the outer membrane. A metastable trimer was also detected by Fourel et al. (44) as an intermediate in the assembly of OmpE These authors used a large panel of mAbs, all recognizing conformational epitopes, and studied the kinetics of the appearance of the corresponding epitopes in pulse-chase experiments. The different epitopes were exposed on the assembling protein in four distinct stages. The earliest epitopes were detected

Outer Membrane Proteins in E. coli

157

immediately after processing and they are probably present on a folded monomeric assembly intermediate. This intermediate could be immunoprecipitated from the periplasmic fraction, shortly after the pulse, consistent with the periplasmic pathway for OMP biogenesis (Figure 3). This early intermediate was converted successively into metastable trimers, stable trimers, and finally to native trimers, with each step being defined by the appearance of additional epitopes. A metastable trimeric assembly intermediate of OmpF was not detected by Reid et al. (103). However, these authors detected a novel intermediate with an electrophoretic mobility in between those of denatured monomers and trimers. This assembly intermediate (see also Figure 5) was identified as a dimer. The data discussed above show that the assembly of OMPs into their native structures is a multistep process. Unfortunately, the subcellular localization of the intermediates may be prone to artefacts. Standard cell fractionation methods have been established for fully assembled native proteins. Fractionation artefacts have previously been reported for genetically manipulated proteins (125). Even bona fide E. coli proteins may be prone to fractionation artefacts; for instance, the A and B subunits of heat-labile enterotoxin are periplasmic proteins, but they form aggregates in the Tris-EDTA buffers normally used for membrane isolation, and consequently, they pellet with the membranes (58). It is conceivable that assembly intermediates become membrane-associated or aggregated during cell disruption, or that they pass through a membranous compartment that behaves differently during fractionations than the bulk of inner or outer membranes. Probably the most reliable method for the subcellular localization of proteins is immunoelectron microscopy. However, this method is not very sensitive and therefore inappropriate to detect low amounts of assembly intermediates. A possible solution to this problem is the use of mutants with thermosensitive assembly defects. Such a mutant has been described for LamB (84). However, the assembly intermediate was rapidly degraded at the restrictive temperature. Degradation was reduced in a degP mutant strain, lacking a periplasmic protease. After a short incubation at the restrictive temperature, the accumulated assembly intermediate could be chased into trimers by a shift-down to the permissive temperature. However, after a longer incubation at the restrictive temperature, which would be required to accumulate sufficient assembly intermediates for detection by immunoelectron microscopy, the protein lost its assembly competence and could therefore no longer be considered as a true intermediate. Presently, only those intermediates that are already inserted into the outer membrane can reliably be localized. Whereas most OMPs are resistant to proteases, a protease-sensitive site can be created by inserting a short stretch of additional amino acid residues in an external loop (e.g., refs. 2, 41). These insertions do not generally interfere with the correct assembly of the proteins into the outer membrane. Thus, the outer membrane localization of assembly intermediates can be assessed by determining their sensitivity to externally added proteases.

158

JAN TOMMASSEN and HANS DE COCK

B. Role of LPS and Lipid Biosynthesis An important question to answer is why OMPs selectively insert into the outer membrane, and not in the cytoplasmic membrane. LPS could be important in this respect, since it is a lipidic component, specific to the outer membrane. Indeed, a tight interaction between several OMPs and LPS has been reported (36, 105, 142). LPS consists of a complex phosphorylated heteropolysaccharide that is linked to a glucosamine-containing lipid, lipidA (104). The polysaccharide portion has been divided into two major regions, an internal core region and the peripheral O-antigen, which shows a high degree of variability among different strains. Because of this variability, and also because many commonly used laboratory strains, including E. coli K-12, are lacking the O-antigen, one would not expect this part of the LPS to be important in the assembly of OMPs. Nevertheless, it has been observed that OmpF preferentially interacts with O-antigen-containing LPS (36). Mutants affected in the core region of the LPS, most notably heptose-deficient strains, are markedly decreased in the proportion of protein recovered in the outer membrane fraction (6, 68), indicating a strong correlation between OMP biogenesis and LPS structure. Not all OMPs were affected to the same extent by the LPS mutations (e.g., whereas OmpF was barely detectable, OmpA was only moderately and OmpC hardly, if at all, affected in heptose-deficient mutants [53]). This result suggests that different OMPs depend to different degrees on LPS structure or that they interact with different regions of the LPS molecule. The exclusive location of LPS in the outer leaflet of the outer membrane may be difficult to reconcile with a role as recognition site for the selective insertion of OMPs into the outer membrane. However, one should realize that LPS also has to be transported from its site of synthesis in the inner membrane to the outer membrane. Thus, assembly intermediates of the OMPs might interact with nascent LPS, and consequently, the biogenesis of the molecules might be coupled. This supposition was underscored by the observation that inhibition of fatty acid synthesis, and consequently of LPS, by the drug cerulenin interfered with the assembly of OmpF into trimers (13). Later experiments suggested that the conversion of metastable into stable trimers was inhibited by the drug (44). The assembly intermediates that accumulated during treatment with the drug were degraded in the cell (13). Furthermore, a dramatic decrease in the level of synthesis of OmpF was observed when its assembly was inhibited by the drug. The synthesis of several other OMPs was also found to be affected by cerulenin treatment, and this inhibition probably takes place at the translational level (89). These results suggest that there is a feedback inhibition of OMP synthesis by assembly intermediates. The observation that expression of an altered OmpC protein, lacking two amino acid residues from a transmembrane segment inhibited the expression of several other major OMPs (21), might be explained by the same feedback inhibition mechanism. In conclusion, there appears to be a direct coupling between the synthesis of lipid (most likely LPS) and the assembly of OMPs.

Outer Membrane Proteins in E. coli

159

C. In vitro Reconstitution of the Insertion and Assembly Process

Several laboratories have reported on the development of cell-free in vitro systems to study the molecular details of OMP assembly As mentioned above (Section V.A), an assembly intermediate of OmpA, designated imp-OmpA, accumulates in cells overproducing this protein. Addition of LPS to imp-OmpA resulted in conversion to the heat-modifiable form and in the acquisition of phage receptor activity (43). Earlier, it was described that the denatured OmpA, isolated from outer membranes, could be renatured by the addition of LPS but not by E. coil phospholipids or dimyristoylphosphatidylcholine (DMPC) (110). The renatured protein was heat-modifiable, displayed phage receptor activity, and its N-terminal part, which is normally embedded in the outer membrane (see Figure I B), became protease-resistant. The lipid A moiety of LPS appeared to be as effective as complete LPS in the renaturation experiments (I10). These results underscore the postulated involvement of LPS in the assembly of OMPs. However, it was demonstrated more recently that SDS-denatured OmpA could be refolded by adding the nonionic detergent octyiglucoside (OG), and could subsequently be reconstituted into DMPC lipid membranes (37). Hence, LPS is neither required for proper folding of OmpA nor for incorporation into a lipid membrane. Furthermore, urea-denatured OmpA was shown to refold and insert in the absence of detergent when diluted in a dispersion of small preformed DMPC vesicles (121). Large vesicles were not effective, unless in the presence of low concentrations of OG. The precise function of OG in the latter case is not clear. It seems unlikely that it induced the refolding of OmpA and allowed the subsequent insertion into the large vesicles, since concentrations below the critical micelle concentration were effective. Possibly, the detergent-induced small distortions in the large vesicles might be required to allow the insertion of OmpA. In small vesicles, the lipid bilayer is highly curved, and consequently, the lipid packing is not optimal. This feature may allow insertion in the absence of detergent. Interestingly, freeze-fracturing electron microscopy has revealed that the outer fracture face of the outer membrane is densely occupied with particles (133). Consequently, the periplasmic face of the outer membrane is covered with highly-curved micelle-like structures (schematically depicted in Figure 4). We suggest that these particles enable the insertion of OMPs, similar to the small DMPC vesicles in the in vitro reconstituted system for OmpA assembly. Moreover, the number of particles observed was drastically reduced in mutants with defects in the core region of the LPS, for example, in heptose-deficient mutants (133). Hence, the effect of LPS mutations on the biogenesis of OMPs (see section V.B) might be explained by the reduced number of particles (i.e., the reduced number of insertion sites). Recently, we have developed a system to study the assembly of PhoE protein in vitro (27, 28, 29). Radioactively labeled PhoE protein was synthesized in vitro in an E. coil lysate and its folding was probed in immunoprecipitation experiments with mAbs that recognize conformational epitopes. Since the protein could be

JAN TOMMASSEN and HANS DE COCK

160

.,

.

.

.

.

.

9 ..

i!iillii~iil]~~ ..... . . . . . . .

~ ........ ~.~

'i :

i i!= ii .~".........

~ ! i ) ! i ! , . . ~ ' !:::}~ ....' s = .~ :~Y:' ..

,

:'i:....

, 9

"_

.

...

..

F~ure 4. Schematic representation of a transversal section of an outer membrane particle with corresponding pit. The figure shows the hemimicellelike appearance of the outer membrane particles and the wedge-shaped organization of the LPS molecules within these particles. Proteins and divalent cations, which are implicated in the formation of the particles, are not shown. precipitated with these mAbs (29), it apparently adopts in vitro a native-like configuration. When prePhoE was synthesized in vitro, the efficiency of the immunoprecipitations was much lower, consistent with the idea that one function of the signal sequence is to maintain the translocation competent (i.e., in the non-native state of the mature domain). The radiolabeled PhoE that was immunoprecipitated displayed a particular electrophoretic behavior, reminiscent of the hot-modifiability of OmpA, that is, in the unheated samples the protein migrated much faster in SDS-polyacrylamide gels than the denatured form (Figure 5). Therefore, the immunoprecipitated protein probably represents a "folded monomer" and might be an intermediate in the assembly process. Consistent with the latter supposition is the recent detection of the folded monomer in vivo in pulsechase experiments (van Gelder and Tommassen, manuscript in preparation). Dimers and trimers were observed after adding isolated outer membranes to the in vitro synthesized PhoE (Figure 5) (29). Neither inner membrane vesicles nor liposomes consisting of phospholipids, LPS, or both could effectively induce trimerization. These results suggest that the outer membrane contains a factor other than LPS that is required for trimerization. Alternatively, the LPS-containing liposomes are incompetent in inducing trimerization by the lack of structural elements like the outer membrane particles. Directly or indirectly, LPS seems to be involved in the trimerization process, since outer membranes from mutant strains

.>~:..,..,....:.;$.:

.~.~:..~.

tri P h o E di P h o E -

w

.:--.

m PhoE -- ~

111

~:~~/~

m*PhoE-q=~

1 2 3 4 5

"C

TL

0 23 56100

Figure 5. Different forms of PhoE, detected after its in vitro synthesis. Plasmid piP370 was used to direct the in vitro synthesis of a quasimature PhoE protein, which contains at the N-terminus only two amino acids instead of the signal sequence (28). Lane 1 shows the translation products. The major band below PhoE representschloramphenicol acetyltransferase, which is also encoded by the plasmid. After translation, outer membranes were added to induce trimerization (29), and PhoE was immunoprecipitated with a mAb, recognizing a conformational epitope in PhoE. The immunoprecipitates were incubated at the indicated temperatures in sample buffer before electrophoresis, to assessthe stability of the different PhoE conformations (lanes 2-5). In addition to some denatured PhoE (mPhoE), trimers (triPhoE), dimers (di PhoE) and folded monomers (m" PhoE) can be detected on the autoradiogram. The sample in lane I was boiled before electrophoresis.

161

162

JAN TOMMASSEN and HANS DE COCK

with a defect in the core region of the LPS were less efficient in inducing trimerization (27). Furthermore, the trimers seemed to be associated with LPS, since their mobility during electrophoresis was slightly different, depending on the chemotype of the LPS in the outer membranes used to induce trimerization. The trimers obtained were highly resistant to denaturation in SDS (Figure 5), like the native PhoE produced in vivo. However, they were not inserted into the membranes since they remained in the supernatant after pelleting of the membranes by centrifugation (28). Furthermore, they were not protected against proteases. Insertion in a protease-resistant configuration was observed when small amounts of a detergent (0.06% Triton X-100) were present during trimerization (27). This observation is reminiscent of the results obtained for the insertion of OmpA in large DMPC vesicles, discussed above. Similar to the PhoE porin, the assembly of the related OmpF porin, obtained either by synthesis in an E. coli lysate (112) or by secretion from spheroplasts (111), has been studied in vitro. The results obtained were largely consistent with those described for PhoE. The major discrepancy is that trimerization of the OmpF proteins could not only be induced by outer membranes, but also by LPS preparations, and even by a mixture of an anionic and a nonionic detergent. Similarly, the renaturation of OmpF, isolated from the outer membrane and denatured in guanidium hydrochloride, by a combination of SDS and octyl-pentaoxyethylene, has been reported (38). An explanation for this discrepancy is not obvious at the moment. Anyhow, these results suggest that LPS is neither required for the folding nor for the trimerization of the OmpF molecules. However, it is still possible that the LPS plays an important role in the in vivo situation and that SDS functions as the substitute for LPS in vitro. Furthermore, the OmpF secreted by spheroplasts was reported to trimerize more efficiently than the OmpF or PhoE proteins synthesized in vitro in an E. coli lysate (112). There are several possible explanations for this difference: (1.) the in vitro synthesized proteins contained short N-terminal extensions of a few amino acid residues that may have interfered with the trimerization process, (2.) trimerization may have been inhibited by binding of cytoplasmic chaperones like SecB (30), which were present in the lysates used to direct the in vitro synthesis of these proteins, and (3.) secretion across the cytoplasmic membrane might be accompanied by conformational changes favoring trimerization. Indeed, the spheroplast-secreted OmpF porin appeared to be different from the in vitro synthesized OmpF in its reactivity with antibodies recognizing conformational epitopes (112).

D.

Sorting Signals in Outer Membrane Proteins

In general, the membrane-spanning segments of a polytopic inner membrane protein are hydrophobic and can insert into the membrane, independent of the other membrane-spanning segments. In contrast, the membrane-spanning segments of OMPs are ~-strands with only one hydrophobic side (see Section II.B). Conse-

Outer Membrane Proteins in E. coli

163

quently, OMPs are compatible with the lipidic environment of the membrane only after the formation of the entire ~barrel structure containing a hydrophobic exterior. It is, therefore, not surprising that large deletions (removing > 30 amino acid residues) in the mature domain of PhoE prevented the normal incorporation of the mutant proteins into the outer membrane and led to their periplasmic accumulation (14). Such large deletions can be expected to interfere with the folding of the protein into the [3-barrel configuration. Similarly, large deletions in the membrane-embedded N-terminal part of OmpA prevented the normal incorporation of this protein in the membrane (66), whereas deletions in the C-terminal periplasmic tail of this protein (see Figure IB) were tolerated (18). Although the entire conformation of OMPs is apparently required for their correct localization, some portions of these proteins might be more important to the correct conformation than others, and in addition, specific sorting signals might still be present to allow interaction with the appropriate membrane. A priori, one could expect that the J-strands are more important for the assembly of the proteins than the surface exposed loops, since there are restrictions on both the [3-structure and on the hydrophobicity of residues in these segments. Comparison of the primary structures of the porins OmpF, OmpC, and PhoE (8.5) show that the sequences of the membrane-spanning segments are much better conserved than those of the exposed loops. A similar observation was made when the primary structures of the OmpA proteins of different Enterobacteriaceae were compared (I 7). Furthermore, insertions (2, 4) and deletions (/) were well tolerated in the exposed loops of PhoE, as well as of other OMPs (e.g., refs. 22, 41, 67). However, the [3-strands also tolerated mutations. In the case of PhoE, hydrophobic residues in the J-strands that are exposed to the fatty acyl chains of the lipids or to the subunit interface of the trimers could be replaced by hydrophilic or even charged residues without affecting the assembly of the protein (119). However, two of such substitutions in a single [3-strand dramatically affected the efficiency of outer membrane incorporation. Furthermore, the first membrane-spanning segment of PhoE could be completely deleted (e.g., mutant protein A3-30, in ref. 16) and still, the mutant protein was correctly inserted in the membrane as evidenced by the binding of PhoE-specific bacteriophage and mAbs to intact cells. However, part of the total amount of the mutant protein produced accumulated in the periplasm, showing that the efficiency of outer membrane insertion was reduced. The first membrane-spanning J-strand of PhoE does not face the lipids, but is located at the subunit interface within the trimers (2.5). Consistently, no trimers of this mutant protein were detected, indicating that either the stability of the trimers is drastically reduced or that the protein inserts into the membrane as a monomer. According to the model for the localization of OMPs, depicted in Figure 3, these proteins adopt their tertiary structure in the periplasm before inserting into the outer membrane. Consequently, a putative sorting signal might be a conformational one and therefore difficult to detect by comparison of the primary structures of diverse OMPs. Nevertheless, a few short stretches of amino acid residues with vague

164

JAN TOMMASSEN and HANS DE COCK

sequence homology were detected in the primary structures of the porins, LamB and OmpA (92). Probably, most of these sequences do not correspond to sorting signals since the homologies became insignificant when more OMP sequences became available, or because deletion analysis in OmpA or PhoE showed that they were not involved in outer membrane localization. For the most pronounced similarity region, sequence comparisons have recently been updated (91), and it appeared that only a single glycine residue, corresponding to Gly- 144 in PhoE, was completely conserved in this segment. Gly-144 is located in a periplasmic turn in the PhoE structure (25). In the OmpA model (134), the corresponding residue is located in an entirely different position near the cell surface, which makes a common function for these residues unlikely. Gly-144 of PhoE was replaced by a leucine residue (31). Indeed, outer membrane localization was somewhat defective, especially at higher growth temperatures. However, in the in vitro assembly system for PhoE (described in Section V.C), the mutant protein appeared to be defective in folding (31). Therefore, we believe that Gly-144 of PhoE is not part of a sorting signal, but is important for protein folding. A more significant similarity was recently detected at the C-termini of OMPs (118). Comparison of the last ten amino acid residues of diverse OMPs revealed the presence of potential amphipathic ~strands with hydrophobic residues at positions 1 (Phe), 3 (preferentially Tyr), 5, 7, and 9 from the C-terminus. In the porins this segment corresponds to the last membrane-spanning segment (25), and its deletion in PhoE had a detrimental effect on assembly in the membrane (15). This consensus sequence was found in the vast majority of bacterial OMPs, including E. coli proteins with widely diverse functions and OMPs of other Gram-negative bacteria (118). In a few cases, including LamB, Trp instead of Phe was found at the C-terminal position. However, like Phe, Trp is a hydrophobic aromatic residue and is therefore expected to functionally replace the Phe. In OmpA and its analogs of other bacteria, the consensus sequence is not found at the ultimate C-terminus, which is extending into the periplasm (Figure 1B), but at the C-terminus of the membrane embedded part (i.e., residues 161-170). The consensus sequence was not detected at the C-termini of OMPs that are involved in the secretion of macromolecular compounds like PapC, which is involved in the secretion of pill subunits (95), or TolC (93) known to be involved in the secretion of o~-haemolysin (139). Interestingly, the consensus sequence is located in a region in OmpA that was earlier implicated in sorting to the outer membrane (66). As mentioned in the beginning of this section, large deletions in OmpA prevented the correct incorporation of the protein into the outer membrane. However, immunoelectron microscopy revealed that the mutant proteins were still associated with the outer membrane, unless the deletions covered an area between residues 154 and 180. When this region was deleted, the mutant proteins accumulated in the periplasm (66). These results suggested that the area between residues 154 and 180 (including the consensus sequence between residues 161-170) contains an outer membrane

Outer Membrane Proteins in E. coli

165

sorting signal. Moreover, of many missense mutations created in ompA, only a mutant protein containing two amino acid substitutions in the consensus sequence was totally blocked in outer membrane assembly (65). Since the C-terminal Phe is the most pronounced feature of the consensus sequence, this residue was used as a target for extensive site-directed mutagenesis in PhoE (118). High-level expression of all mutant proteins obtained was lethal to the cells, and the efficiency of outer membrane incorporation was dramatically affected. The mutational effect was the least severe when the Phe was replaced by another aromatic residue and was most severe when this residue was deleted. In the latter case, hardly any PhoE could be detected in the outer membrane, and the protein accumulated as aggregates in the periplasm (117). In vitro the folding and the trimerization of this mutant protein were virtually unaffected, but the insertion in isolated outer membranes was drastically reduced. When the expression level of the mutant proteins carrying substitutions for or deletion of the C-terminal Phe was reduced, the detrimental effects on growth were alleviated and more mutant protein was correctly assembled into the outer membrane (117). This result suggests that the C-terminal Phe is not absolutely required for the correct assembly of the protein, but does contribute to the efficiency of the process. Apparently, there is a kinetic partitioning between outer membrane incorporation and aggregation of periplasmic assembly intermediates. At high levels of expression, more protein will aggregate and become assembly-incompetent, whereas at low expression levels the tendency to aggregate will be reduced, resulting in increased outer membrane incorporation. The C-terminal Trp in Lamb may play a similar role as the Phe in PhoE, since deletion of the last two amino acid residues of LamB has been reported to drastically affect outer membrane incorporation (56).

VI.

CONCLUSIONS A N D FUTURE PROSPECTS

Whereas the molecular details of the mechanism of the transport of proteins across the cytoplasmic membrane are beginning to emerge, even the basic concepts of the last steps in the biogenesis of OMPs are far from clear. The proteins probably fold into their tertiary structure after passage through the inner membrane (Figure 3). After folding into a [~-barrel, the proteins have a hydrophobic exterior, compatible with insertion in the lipidic environment of the membrane. It is not yet clear where folding takes place. The process probably does not occur free in the periplasm, but is associated at the periplasmic face of one of the membranes, since amphiphiles were usually observed to induce folding in vitro. Although refolding of denatured OMPs was observed in the absence of any other proteins, this does not exclude the possibility that chaperones or enzymes are involved in this Process in vivo. Several other proteins have been reported to renature spontaneously in vitro, whereas their folding is guided by other proteins in vivo (e.g., ref. 40). The possible involvement of periplasmic proteins in the folding of OMPs will be an important topic of future

166

JAN TOMMASSEN and HANS DE COCK

research. Many OMPs, including the porins, are devoid of cysteine residues, and in these cases the involvement of protein disulfide isomerases can be excluded. Nevertheless, the expression of OmpF was affected at the transcriptional level in a dsbA mutant lacking a periplasmic disulfide isomerase (99). On the other hand, native OmpA does contain a disulfide bond and its formation was apparently retarded in a dsbA mutant (8). However, since the two cysteines of OmpA are located in the C-terminal periplasmic extension, no effect on outer membrane localization would be expected. The cis-trans isomerization of peptidylpropyl bonds is frequently a rate-limiting step in protein folding. An enzyme catalyzing this reaction has been detected in the periplasm of E. coli (54), but the involvement of this enzyme and of putative chaperones in the folding of OMPs remains to be investigated. Another important aspect for future research will be to gain insight in the mechanism of selective insertion of OMPs in the outer membrane and not in the inner membrane. There are some indications that LPS is involved in this process, but the effect of LPS could be indirect. The role of the observed outer membrane particles and of lipid packing will have to be investigated. Furthermore, if the C-terminal consensus sequence (118) indeed represents a sorting signal, a receptor for this signal can be expected to exist and should be identified.

REFERENCES 1. Agterberg, M., Adriaanse, H., Tijhaar, E., Resink, A., & Tommassen, J. (!989). Role of the cell surface-exposed regions of outer membrane protein PhoE of Escherichia coil KI2 in the biogenesis of the protein. Fur. J. Biochem. 185, 365-370. 2. Agterberg, M., Adrisanse, H., & Towanassen, J. (1987). Use of outer membrane protein PhoE as a carrier for the transport of a foreign antigenic determinant to the cell surface of Escherichia coli K-12. Gene 59, 145-150. 3. Agterberg, M., Adriaanse, H., van Bruggen, A., Karperien, M., & Tommassen, J. (1990). Outer-membrane PhoE protein of Escherichia coli K-12 as an exposure vector: Possibilities and limitations. Gene 88, 37--45. 4. Agterberg, M., Benz, R., & Tommassen, J. (1987). Insertion mutagenesis on a cell-surface-exposed region of outer membrane protein PhoE of Escherichia coli K-12. Fur. J. Biochem. 169, 65-71. 5. Akita, M., Sasaki, S., Matsuyama, S.-i., & Mizushima, S. (1990). SecA interacts with secretory proteins by recognizing the positive charge at the amino terminus of the signal peptide in Escherichia coil J. Biol. Chem. 265, 8164-8169. 6. Ames, G. E, Spudich, E. N., & NikaJdo, H. (1974). Protein composition of the outer membrane of Salmonella ophimurium: effect of lipopolysaccharide mutations. J. Bacterial. 117, 406--416. 7. Andersson, H., & volt Heijne, G. (1991). A 30-residue-long "export initiation domain" adjacent to the signal sequence is critical for protein translocation across the inner membrane of Escherichia coll. Proc. Natl. Acad. Sci. USA 88, 9751-9754. 8. Bardwell, J. C. A., McGovern, K., & Beckwith, J. (1991). Identification of a protein required for disulfide bond formation in viva. Cell 67, 581-589. 9. Bassford, P. J. Jr., Silhavy, T. J., & Beckwith, J. R. (1979). Use of gene fusions to study secretion of maltose-binding protein into Escherichia coli periplasm..L Bacterial. 139, 19-31.

Outer Membrane Proteins in E. coli

167

10. Benz, R., & Bauer, K. (1988). Permeation of hydrophilic molecules through the outer membrane of gram-negative bacteria. Review on bacterial porins. Eur. J. Biochem. 176, 1-19. 11. Benz, R., Schmid, A., & Hancock, R. E. W. (1985). Ion selectivity of Gram-negative bacterial porins. J. Bocteriol. 162, 722-727. 12. Benz, R., Schmid, A., Nakae, T., & Vos-Scheperkeuter, G. H. (1986). Pore formation by LamB of Escherichia coil in lipid bilayer membranes. J. Bacteriol. 165, 978--986. 13. Boll& J.-M., Lazdunski, C., & Pages, J.-M. (1988). The assembly of the major outer membrane protein OmpF of Escherichia coli depends on lipid synthesis. EMBO J. 7, 3595-3599. 14. Bosch, D., Leunissen, J., Verbakel, J., de Jong, M., van Erp, H., & Tommassen, J. (1986). Periplasmic accumulation of truncated forms of outer-membrane PhoE protein of Escherichia coli K-12. J. Mol. Biol. 189, 449--455. 15. Bosch, D., Scholten, M., Verhagen, C., & Tommassen, J. (1989). The role of the carboxy-terminal membrane-spanning fragment in the biogenesis of Escherichia coil K 12 outer membrane protein PhoE. Mol. Gen. Genet. 216, 144-148. 16. Bosch, D., Voorhout, W., & Tommassen, J. (1988). Export and localization of N-terminally truncated derivatives of Escherichia coli K- 12 outer membrane protein PhoE. J. Biol. Chem. 263, 9952-9957. 17. Braun, G., & Cole, S. T. (1984). DNA sequence analysis of the Serrotia marcescens ompA gene: implications for the organization of an enterobacteriai outer membrane protein. Mol. Gen. Genet. 195, 321-328. 18. Bremer, E., Cole, S. T., Hindennach, I., Henning, U., Beck, E., Kurz, C., & Schaller, H. (1982). Export of a protein into the outer membrane of Escherichio coil K 12. Stable incorporation of the OmpA protein requires less than 193 amino-terminal amino-acid residues. Eur. J. Biochem. 122, 223-231. 19. Breukink, E., Kusters, R., & de Kruijff, B. (1992). In vitro studies on the folding characteristics of the Escherichia coil precursor protein prePhoE. Evidence that SecB prevents the precursor from aggregating by forming a functional complex. Eur. J. Biochem. 208, 419--425. 20. Brundage, L., Hendrick, J. P., Schiebel, E., Driessen, A. J. M., & Wickner, W. (1990). The purified E. coli integral membrane protein SecY/E is sufficient for reconstitution of SecA-dependent precursor protein translocation. Cell 62, 649-657. 21. Catron, K. M., & Schnaitman, C. A. (1987). Export of protein in Escherichia coli: a novel mutation in ompC affects expression of other major outer membrane proteins. J. Bacteriol. 169, 4327--4334. 22. Charbit, A., Boulain, J. C., Ryter, A., & Hofnung, M. (1986). Probing the topology of a bacterial membrane protein by genetic insertion of a foreign epitope; expression at the cell surface. EMBO J. 5, 3029-3037. 23. Charbit, A., Cl6ment, J.-M., & Hofnung, M. (1984). Further sequence analysis of the phage lambda receptor site. Possible implications for the organization of the LamB protein in E. coli KI2. J. Mol. Biol. 175, 395-401. 24. Collier, D. N., Bankaitis, V. A., Weiss, J. B., & Bassford, E J. Jr. (1988). The antifolding activity of SecB promotes the export of the E. coil maltose-binding protein. Cell 53, 273-283. 25. Cowan, S. W., Schirmer, T., Rummel G., Steiert, M., Ghosh, R., Paupfit, R. A., Jansonius, J. N., & Rosenbusch, J. P. (1992). Crystal structures explain functional properties of two E. coil porins. Nature 358, 727-733. 26. Davis, N. G., & Model, P. (1985). An artificial anchor domain: hydrophobicity suffices to stop transfer. Cell 41,607-614. 27. De Cock, H., Blokland, S., & Tommassen, J. (1995). In vitro assembly and insertion of PhoE protein of E. coil K-12 into the outer membrane. (Manuscript submitted for publication.) 28. De Cock' H., Hekstra, D., & Tommassen, J. (1990). In vitro trimerization of outer membrane protein PhoE. Biochimie 72, 177-182.

168

JAN TOMMASSEN and HANS DE COCK

29. De Cock, H., Hendriks, R., de Vrije, T., & Tommassen, J. (1990). Assembly of an in vitro synthesized Escherichia coli outer membrane porin into its stable trimeric configuration. J. Biol. Chem. 265, 4646--4651. 30. De Cock, H., Overeem, W., & Tommassen, J. (1992). Biogenesis of outer membrane protein PhoE of Escherichia coll. Evidence for multiple SecB-binding sites in the mature portion of the PhoE protein. J. Mol. Biol. 244, 369-379. 31. De Cock, H., Quaedvlieg, N., Bosch, D., Schoiten, M., & Tommassen, J. (1991). Glycine-144 is required for efficient folding of outer membrane protein PhoE of E. coil K-12. FEBS Lett. 279, 285-288. 32. De Cock, H., & Tommassen, J. (1992). SecB-binding does not maintain the translocation-competent state of prePhoE. Mol. Microbiol. 6, 599-604. 33. Desaymard, C., D~barbouill6, M., Jolit, M., & Schwartz, M. (1986). Mutations affecting antigenic determinants of an outer membrane protein of Escherichio coll. EMBO J. 5, 1383-1388. 34. De Vrije, G. J., Batenburg, A. M., Killian, J. A., & de Kmijff, B. (1990). Lipid involvement in protein translocation in Escherichia coll. Mol. Microbiol. 4, 143-150. 35. De Vrije, T., Tommassen, J., & de Kruijff, B. (1987). Optimal posttranslationai transiocation of the precursor of PhoE protein across Escherichia coil membrane vesicles requires both ATP and the protonmotive force. Biochim. Biophys. Acto 900, 63-72. 6. Diedrich, D. L., Stein, M. A., & Schnaitman, C. A. (I 990). Associations of Escherichia coil K- 12 OmpF trimers with rough and smooth iipolx)lysaccharides. J. Bacteriol. 172, 5307-5311. 37. Dornmair, K., Kiefer, H., & Jltlmig, E (1990). Refoiding of an integral membrane protein. OmpA of Escherichia coll. J. Biol. Chem. 265, 18907-18911. 38. Eisele, J.-L., & Rosenbusch, J. P. (1990). In vitro folding and oligomerization of a membrane protein. Transition of a bacterial porin from random coil to native conformation. J. Biol. Chem. 265, 10217-10220. 39. Fifip, C., Fletcher, G., Wulff, J. L., & Earhart, C. E (1973). Solubilization of the cytoplasmic membrane of Escherichia coli by the ionic detergent sodiumolauryl sarcosinate. J. Bocter~ol. 115, 717-722. 0. Frenken, L. G. J., de Gwot, A., Tommassen, J., & Verrips, C. T. (1993). Role of the lipB gene product in the folding of the secreted lipase of Pseudomonas glumae. Mol. Microbiol. 9, 591-599. 41. Freudl, R., Maclntyre, S., Degen, M., & Hemming,U. (1986). Cell surface exposure of the outer membrane protein OmpA of Escherichia coli K-12. J. MoL Biol. 188, 491-.494. 20 Freudl, R., Schwarz, H., Klose, M., Mowa, N. R., & Henning, U. (1985). The nature of information, required for export and sorting, present within the outer membrane protein OmpA of Escher~chia coil K-12. EMBO J. 4, 3593-3598. 30 Freudl, R., Schwarz, H., Stierhof, Y.-D., Gamon, K., Hindennach, I., & Helming, U. (I 986). An outer membrane protein (OmpA) of Escherichia coil K-12 undergoes a confmmational change during expm~ J. Biol. Chem. 261, 11355-11361. 4 O Fourel, D., Mizushima, S., & Pages, J.-M. (1992). Dynamics of the exposure of epitopes on OmpF, an outer membrane protein of Escherichia coll. Fur. J. Biochem. 206, 109-114. 5. Gardel, C., Johnson, K., Jacq, A., & Beckwith, J. (1990). The secD locus of E. coli codes for two membrane proteins required for protein export. EMBO J. 9, 3209-3216. Gehring, K. B., & Nikaido, H. (1989). Existence and purification of porin heterotrimers of Escherichio coli K12 OmpC, OmpF and PhoE proteins. J. Biol. Chem. 264, 2810-2815. Geller, B. L., Movva, N. R., & Wickner, W. (1986). Both ATP and the electrochemical potential are required for optimal assembly of pro-OmpA into Escherichia coil inner membrane vesicles. Proc. Natl. Acod. Sci. USA 83, 4219--4222. 4J~. Grodberg, J., & Dunn, J. J. (1988). ,rapT encodes the Escherichia coli outer membrane protease that cleaves "1"7RNA polymerase during purification. J. Bacteriol. 170, 1245-1253. 9. Guthrie, B., & Wickner, W. (1990). Trigger factor depletion or overproduction causes defective cell division but does not block protein export. J. Bacteriol. 172, 5555-5562.

Outer Membrane Proteins in E. coli

169

0. Hall, M. N., Schwartz, M., & Silhavy, T. J. (1982). Sequence information within the lamB gene

51. 52.

53. 4.

55. 56. 57.

58.

59. 0.

61. 62. 63. 4. 65.

6.

67.

68.

69.

70.

is required for proper routing of the bacteriophage ~. receptor protein to the outer membrane of Escherichia coil K-12. J. Moi. Biol. 156, 93-112. Hardy, S. J. S., & Randall, L. L. (1991). A kinetic partitioning model of specific binding of normative proteins by the bacterial chaperone SecB. Science 25 I, 439-443. Harti, E-U., Lecker, S., Schiebel, E., Hendrick, J. P., & Wickner, W. (1990). The binding cascade of SecB to SecA to SecY/E mediates preprotein targeting to the E. coli plasma membrane. Cell 63, 269-279. Havekes, L. M., Lugtenberg, B. J. J., & Hoekstra, W. P. M. (1976). Conjugation deficient E. coli K 12 F mutants with heptose-less lipopolysaccharide. Mol. Gen. Genet. 146, 43-50. Hayano, T., Takahashi, N., Kato, S., Maid, N., & Suziki, M. (1991). Two distinct forms of peptidylprolyl-cis-trans-isomerase are expressed separately in periplasmic and cytoplasmic compartments of Escherichia coli cells. Biochemistry 30, 3041-3048. Hayashi, S., & Wu, H. C. (1990). Lipoproteins in bacteria. J. Bioenerg. Biomembr. 22, 451--471. Heine, H.-G., Francis, G., Lee, K.-s., & Ferenci, T. (1988). Genetic analysis of sequences in maltoporin that contribute to binding domains and pore structure. J. Bacterial. 170, 1730-1738. Heller, K. B. (1978). Apparent molecular weights of a heat-modifiable protein from the outer membrane of Eschericlda coil in gels with different acrylamide concentrations. J. Bacterial. 134, 1181-1183. Hirst, 1". R., Randall, L. L, & Hardy, S. J. S. (1984). Cellular location of heat-labile enterotoxin in Escherichia coli. J. Bacterial. 157, 637-642. Ira, K., Bassford, P. J. Jr., & Beck-with,J. (1981). Protein localization in E. coli: Is there a common step in the secretion of periplasmic and outer-membrane proteins? Cell 24, 707-717. Jackson, M. E., Pratt. J. M., & Holland, I. B. (1986). Intermediates in the assembly of the TonA polypeptide into the outer membrane of Escherichia coli KI2. J. Mol. Biol. 189, 477-486. Kadner, R. J. (1990). Vitamin Bi2 transport in Escherichia coil: energy coupling between membranes. Mol. Microbial. 4, 2027-2033. Kellenberger, A. (1990). The "bayer bridges" confronted with results from improved electron microscopy methods. Mol. Microbial. 4, 697-705. Kleffel, B., Garavito, R. M., Baumeister, W., & Rosenbusch, J. P. (1985). Secondary su,ucture of a channel-forming wotein: Porin from E. coli outer membranes. EMBO J. 4, 1589-1592. Klein, W., & Boos, W. (1993). Induction of the ;~ receptor is essential for effective uptake of trehalose in Escherichia coll. J. Bacterial. 175, 1682-1686. Klose, M., Maclntyre, S., Schwm2, H., & Helming, U. (1988). The influence of amino substitutions within the mature part of an Escherichia coil outer membrane protein (OmpA) on assembly of the polypeptide into its membrane. J. Biol. Chem. 263, 13297-13302. Klose, M., Schwarz, H., Maclntyre, S., Freudl, R., Eschbach, M.-L., & Henning, U. (1988). Internal deletions in the gene for an Escherichia coil outer membrane protein define an area possibly important for recognition of the outer membrane by this polypeptide. J. Biol. Chem. 263, 13291-13296. Koebnik, R., & Braun, V. (1993). Insertion derivatives containing segments of up to 16 amino acids identify surface- and periplasm-exposed regions of the FhuA outer membrane receptor of Escherichia coli K-12. J. Bacterial. 175, 826-839. Koplow, J., & Goldfine, H. (1974). Alterations in the outer membrane of the cell envelope of heptose-deficient mutants of Escherichia coll. J. Bacterial. 117, 527-543. Korteland, J., Overbeeke, N., de Gruff, P., Overduin, P., & Lugtenberg, B. (1985). Role of the Arg 158residue of the outer membrane PhoE pore protein of Escherichia coil K 12 in bacteriophage TC45 recognition and in channel characteristics. Fur. J. Biochem. 152, 691-697. Kumamoto, C. A. (1989). Escherichia coil SecB protein associates with exported protein precursors in viva. Proc. Natl. Acad. Sci. USA 86, 5320-5324.

170

JAN TOMMASSEN and HANS DE COCK

71. Kumamoto, C. A., & Beckwith, J. (1985). Evidence for specificity at an early step in protein export in Escherichia coil J. Bacterial. 163, 267-274. 72. Kusters, R., de Vrije, T., Breukink, E., & de Kruijff, B. (1989). SecB protein stabilizes a translocation-competent state of purified prePhoE wotein. J. Biol. Chem. 264, 20827-20830. 73. Lecker, S. H., Driessen, A. J. M., & Wickner, W. (1990). ProOmpAcontains secondary and tertiary structure prior to translocation and is shielded from aggregation by association with SecB protein. EMBO I. 9, 2309-2314. 74. Li, P., Beckwith, J., & Inouye, H. (1988). Alteration of the amino terminus of the mature sequence of a periplasmic protein can severely affect protein export in Escherichia coll. Proc. Natl. Acad. Sci. USA 85, 7685-7689. 75. Lill, R., Cunningham, K., Bmndage, L., lto, K., Oliver, D., & Wickner, W. (1989). The SecA protein hydrolyzes ATP and is an essential component of the protein translocation ATPase of E. coil EMBO J. 8, 961-966. 76. Lill, R., Dowhan, W., & Wickner, W. (1990). The ATPase activity of SecA is regulated by acidic phospholipids, SecY, and the leader and mature domains of precursor proteins. Cell 60, 271-280. 77. Lugtenberg, B., & van A I ~ , L. (1983). Molecular architecture and functioning of the outer membrane of Escherichia coil and other Gram-negative bacteria. Biochim. Biophys. Acta 737, 51-115. 78. Luirink, J., High, S., Wood, H., Giner, A., Tollervey, D., & Dobberstein, B. (1992). Signal-sequence recognition by an Escherichia coil ribonucleoprotein complex. Nature 359, 741-743. 79. Maclntyre, S., Eschbach, M.-L., & Mutschler, B.(1990). Export incompatibility of N-terminal basic residues in a mature polypeptide of Escherichia coli can be alleviated by optimising the signal peptide. MOl. Gen. Genet. 22 I, 466-474. 80. Maclntyre, S., Freudl, R., Eschbach, M.-L., & Henning, U. (1988). An artificial hydrophobic sequence functions as either an anchor or a signal sequence at only one of two positions within the Escherichia coli outer membrane protein OmpA. J. Biol. Chem. 263, 19053--19059. 81. Maier, C., Bremer, E., Schmid, A., & Benz, R. (1988). Pore-forming activity of the Tsx protein from the outer membrane of Escherichia coll. Demonstration of a nucleoside-specific binding site. J. Biol. Chem. 263, 2493-2499. 82. Matsuyama, S., Fujita, Y., & Mizushima, S. (1993). SecD is involved in the release of translocated secretory proteins from the cytoplasmic membrane of Escherichia coil EMBO J. 12, 265-270. 83. Metcalfe, M., & Holland, I. B. (1980). Synthesis of a major outer membrane porin by Escherichia call spheroplasts. FEMS Microbial. Lett. 7, 11I-114. 84. Misra, R., Peterson, A., Ferenci, T., & Silhavy, T. J. (1991). A genetic approach for analyzing the pathway of LamB assembly into the outer membrane of Escherichia coll. J. Biol. Chem. 266, 13592-13597. 85. Mizuno, T., Chou, M.-Y., & Inouye, M. (1983). A comparative study on the genes for three porins of the Escherichia coil outer membrane. DNA sequence of the osmoregulated ompC gene. J. Biol. Chem. 258, 6932--6940. 86. Moreno, E, Fowler, A. V., Hall, M., Silhavy, T. J., Zabin, I., & Schwartz, M. (1980). A signal sequence is not sufficient to lead ~galactosidase out of the cytoplasm. Nature 286, 356-359. 87. Morona, R., Klose, M., & Henning, U. (1984). Escherichia coil K-12 outer membrane wotein (OmpA) as a bacteriophage receptor: analysis of mutant genes expressing altered proteins. J. Bacterial. 159, 570--578. 88. Morona, R., Tommassen, J., & Henning, U. (1985). Demonstration of a bacteriophage receptor site on the Escherichia coil KI2 outer-membrane protein OmpC by the use of a protease. Eur. J. Biochem. 150, 161-169. 89. Murgier, M., Pages, C., Lazdunski, C., & Lazdunski, A. (1982). Translational control of ompF, ompC and lamb genetic expression during lipid synthesis inhibition of Escherichia coll. FEMS Microbial. Lett. 13, 307-31 I.

Outer Membrane Proteins in E. coli

171

0, Nikaido, H. (1992). Porins and specific channels of bacterial outer membranes. Mol. Microbiol.

6, 435--442. 91. Nikaido, H., & Reid, J. (1990). Biogenesis of prokaryotic pores. Experientia 46, 174-180. 92. Nikaido, H., & Wu, H. C. P. (1984). Amino acid sequence homology among the major outer membrane proteins of Escherichia coli. Proc. Natl. Acad. Sci. USA 81, 1048--1052. 93. Niki, H., Imamura, R., Ogura, T., & Hiraga, S. (1990). Nucleotide sequence of the toIC gene of Escherichia coil Nucleic Acids Res. 18, 5547. 94. Nishijima, M., Nakaike, S., Tamori, Y., & Nojima, S. (1977). Detergent-resistant phospholipase A of Escherichia coli K-12. Purification and properties. Eur. J. Biochem. 73, 115-124. 95. Norgren, M., B/tga, M., Tennent, J. M., & Normark, S. (1987). Nucleotide sequence, regulation and functional analysis of the papC gene required for cell surface localization of Pap pili of uropathogenic Escherichia coil Mol. Microbiol. I, 169-178. 6. Oliver, D. B., & Beckwith, J. (1981). E. coli mutant pleiotropically defective in the export of secreted proteins. Cell 25, 765-772. 97. Overbeeke, N., & Lugtenberg, B. (1980). Expression of outer membrane protein e of Escherichia coil K-12 by phosphate limitation. FEBS Lett. 112, 229-232. 98. Postle, K. (1990). TonB and the Gram-negative dilemma. Mol. Microbiol. 4, 2019-2025. 99. Pugsley, A. P. (1993). A mutation in the dsbA gene coding for periplasmic disulfide oxidoreductase reduces transcription of the Escherichia co/i ompF gene. Mol. Gen. Genet. 237, 407--41 I. 100. Randall, L. L., & 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. coil. Cell 46, 921-928. 101. Randall, L. L., Topping, T. B., & Hardy, S. J. S. (1990). No specific recognition of leader peptide by SecB, a chaperone involved in protein export. Science 248, 860-863. 102. Rasmussen, B. A., & Silhavy, T. J. (1987). The first 28 amino acids of mature Lamb are required for rapid and efficient export from the cytoplasm. Genes Develop. !, 185-196. 103. Reid, J., Fung, H., Gehring, K., Klebba, P. E., & Nikaido, H. (1988). Targeting of porin to the outer membrane of Escherichia coil Rate of trimer assembly and identification of a dimer intermediate. J. Biol. Chem. 263, 7753-7759. 104. Rick, P. D. (1987). Lipopolysaccharide biosynthesis. In E C. Neidhardt, J. L. Ingraham, K. Brooks Low, B. Magasanik, M. Schaechter, and H. E. Umbarger (Eds.) Escherichia coil and Salmonella t)phimurium. Cellular and molecldar biology (pp. 648--662). Washington, DC: American Society for Microbiology. 105. Rocque, W. J., Coughlin, R. T., & McGroarty, E. J. (1987). Lipopolysaccharide tightly bound to porin monomers and trimers from Escherichia coli K-12. J. Bacteriol. 169, 4003-4010. 106. Saint, N., De, E., Julien, S., Orange, N., & Molle, G. (1993). Ionophore properties of OmpA of Escherichia coll. Biochim. Biophys. Acta 1145, 119-123. 107. Scandella, C. J., & Kornberg, A. (1971). A membrane-bound phospholipase A! purified from Escherichia coli. Biochemistry 10, 4447--4456. 108. Schatz, P. J., & Beckwith, J. (1990). Genetic analysis of protein export in Escherichia coll. Annu. Rev. Genet. 24, 215-248. 109. Schnaitman, C. A. (1971). Solubilization of the cytoplasmic membrane of Escherichia coil by Triton X-100. J. Bacterioi. 108, 545-552. I10. Schweizer, M., Hindennach, I., Garten, W., & Henning, U. (1978). Major proteins of the Escherichia coli outer cell envelope membrane. Interaction of protein II* with lipopolysaccharide. Eur. J. Biochem. 82, 21 !-2 ! 7. 111. Sen, K., & Nikaido, H. (1990). In vitro trimerization of OmpF porin secreted by spheroplasts of Escherichia coli. Proc. Natl. Acad. Sci. USA 87, 743-747. 112. Sen, K., & Nikaido, H. (1991). Trimerization of an in vitro synthesized OmpF porin of Escherichia coil outer membrane. J. Biol. Chem. 266, 11295-11300.

172

JAN TOMMASSEN and HANS DE COCK

113. Silhavy, T. J., Bassford, R J. Jr., & Beckwith~ J. R. (1979). A genetic approach to the study of protein localization in Escherichia coll. In M. Inouye (Ed.) Bacterial outer membranes. Biogenesis and functions (pp. 203-254). New York: John Wiley & Sons. 114. Stair, J., & Nikaido, H. (1978). Outer membrane of Gram-negative bacteria. XVIII. Electron microscopic studies on porin insertion sites and growth of cell surface of Salmonella typhimurium. J. Bacteriol. 135, 687-702. 115. Sonntag, I., Schwarz, H., Hirota, Y., & Henning, U. (1978). Cell envelope and shape of Escherichia coil: multiple mutants missing the outer membrane i i ~ e i n and other major outer membrane proteins. J. Bacteriol. 136, 280-285. 116. Struyv6, M., Bosch, D., Visser, J., & Tomnmssen, J. (1993). Effect of different positively charged amino acids, C-terminally of the signal peptidase cleavage site, on the translocation kinetics of a precursor protein in Escherichia coil K-12. FEMS Microbiol. Lett. 109, 173-178. 117. Struyv6, M., Heutink' M., Kleerebezem, M., van der Krift T., de Cock, H., & Tommassen, J. (1993). Role of the cerlmxy-terminal phenylalanine in the biogenesis of outer membrane protein PhoE of Escherichia coli K-12. (Manuscript submitted for publication.) 118. Struyv~, M., Moons, M., & Tommassen, J. (1991). Carboxy-terminal phenylalanine is essential for the correct assembly of a bacterial oetermembrane protein. J. Mol. Biol. 218, 141-148. 119. Struyv6, M., Nouwen, N.., Veldschoiten, M., Heutink, M., de Cock' H., & Tommassen, J. (1993). Mutagenesis of the exterior hydroplmbic region of the ~-barrel of outer membrane porin PhoE. (Manuscript submitted for publication.) 120. Sugarawa, E., & Nikaido, H. (1992). P o r e - f m n g activity of OmpA protein of Escherichia coll. J. Biol. Chem. 267, 2507-2511. 121. Surrey, T., & Jtihnig, E (1992). Refolding and oriented insertion of a membrane protein into a lipid bilayer. Proc. Natl. Acad. ScL USA 89, 7457-746 I. 122. Swidersky, U. A., Hoffschulte, H. K., & Mailer, M. (1990). Determinants of membrane targeting and transmembrane Iranslocation during bacterial protein export. EMBO J. 9, 1777-1785. 123. Szmelcman, S., & Hofnung, M. (1975). Maltose transpoN in Escherichia coil K-12: involvement of the bacteriophage lambda receptor. J. Bacteriol. 124, 112-118. 124. Tokuda, H., Akimaru, J., Matsuyama, S.-i., Nishiyama, K.-i., & Mizushima, S. (1991). Purification of SecE and reconstitution of SecE-dependent im3tein translocation activity. FEBS Len. 279, 233-236. 125. Tommassen, J. (1986). Fallacies of E. coil cell fractionatiom and consequences thereof for protein export models. Microb. Pathogen. 1, 225-228. 126. Tommassen, J. (1988). Biogenesis and membrane topology of outer membrane proteins in Escherichia coil In J. A. E Op den Kamp (Ed.) Membrane biogenesis, NATO ASI Series Vol. HI6, (pp. 351-373). Berlin: Springer Verlag. 127. Tommassen, J., & de Kmon, T. (1987). Subcellular localization of a PhoE-LacZ fusion protein in E. coil by ptotease accessibility experiments reveals an inner-membrane-spanning form of the protein. FEB$ Left. 221,226--230. 128. Tommassen, J., Leunissen, J., van Denmm-Jongsten, M., & Overduin, E (1985). Failure of E. coil K-12 to transport PhoE-LacZ hybrid proteins out of the cytoplasm. EMBO J. 4, 1041-1047. 129. Tommassen, J., van 1"ol, H., & Lugtenberg, B. (1983). The ultimate localization of an outer membrane protein of Escherichia coil K-12 is not determined by the signal sequence. EMBO J. 2, 1275-1279. 130. Van Alphen, W., & Lugtenberg, B. (1977). Influence of osmolarity of the growth medium on the outer membrane protein pattern of Escherichia coll. J. Bacteriol. 131,623-630. 131. Van der Ley, P., Amesz, H., Tommassen, J., & Lugtenberg, B. (1985). Monoclonal antibodies directed against the cell-surface-exposed part of PhoE pore protein of the Escherichia coli K- 12 outer membrane. Fur. J. Biochem. 147, 401-407.

Outer Membrane Proteins in E. coli

173

132. Van der Ley, P., Struyv6, M., & Tommassen, J. (1986). Topology of outer membrane pore protein PhoE of Escherichia coli. Identification of cell surface-exposed amino acids with the aid of monoclonal antibodies. J. Biol. Chem. 261, 12222-12225. 133. Verldeij, A., van Alphen, L., Bijvelt, J., & Lugtenberg, B. (1977), Architecture of the outer membrane of Escherichia coli K I2. II. Freeze fracture morphology of wild type and mutant swains. Biochim. Biophys. Acta 466, 269-282. 134. Vogel, H., & Jithnig, E (1986). Models for the structure of outer-membrane proteins of Escherichia coil derived from Raman spectroscopy and prediction methods. J. Moi. Biol. 190, 191-199. 135. Von Heijne, G. (1986). Net N-C charge imbalance may be important for signal sequence function in bacteria. J. Mol. Biol. 192, 287-290. 136. Von Heijne, G. (1990). The signal peptide. J. Membr. Biol. 115, 195-201. 137. Voorhout, W., de Kroon, T., Leunissen-Bijvelt, J., Verkleij, A., & Tommassen, J. (1988). Accumulation of LamB-LacZ hybrid proteins in intracytoplasmic membrane-like structures in Escherichia coil KI2. J. Gen. Microbiol. 134, 599-604. 138. Vos-Scheperkeuter, G. H., & Witholt, B. (1984). Assembly pathway of newly synthesized LamB protein an outer membrane protein of Escherichia coli K-12. J. Mol. Biol. 175, 511-528. 139. Wandersman, C., & Delepelaire, P. (1990). TolC, an Escherichia coli outer membrane protein required for hemolysin secretion. Proc. Natl. Acad. Sci. USA 87, 4776-4780. 140. Watanabe, M., & Biobel, G. (1989). SecB functions as a cytosolic signal recognition factor for protein export in E. coli. Cell 58, 695-705. 141. Weiss, J. B., & Bassford, P. J. Jr. (1990). The folding properties of the Escherichia coil maltose-binding protein influence its interaction with SecB in vitro. J. Bacteriol. 172, 3023-3029. 142. Yamada, H., & Mizushima, S. (1980). Interaction between major outer membrane protein (0-8) and lipopolysaccharide in Escherichia coil KI2. Eur. J. Biochem. 103, 209-218. 143. Yamaha, K., & Mizushima, S. (1988). Introduction of basic amino acid residues after the signal peptide inhibits protein translocation across the cytoplasmic membrane of Escherichia coll. Relation to the orientation of membrane proteins. J. Biol. Chem. 263, 19690-19696.

This Page Intentionally Left Blank

STRUCTU RE-F U NCTION RELATIONSHIPS IN THE MEMBRANE CHANNEL PORIN

Georg E. Schulz

I. Porin Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II. Porin Location . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Orientation in the M e m b r a n e . . . . . . . . . . . . . . . . . . . . . . . . B. Packing in the Crystal . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. The Aromatic Girdles . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Folding Pathway . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III. Permeation Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. The Pore Eyelet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. A n Electric Separator . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Voltage Dependence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Porin Specificity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments ................................ References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Advances in Cell and Molecular Biology of Membranes and Organelles Volume 4, pages 175-187.

Copyright 91995 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-924.9

175

176 177 177 180 I g0 181 183 183 183 185 185 186 186

176

GEORG E. SCHULZ I.

P O R I N STRUCTURE

Protozoa have to protect themselves against adverse environments. For this purpose, Gram-negative bacteria have an outer membrane containing channels that are permeable to nutrients. These channels are f o r n ~ by porins that are usually hemotrimeric proteins with subunit sizes ranging from 30 to 50kDa and solute exclusion limits of around 600 Da (1, 2). The channels are quite permeable for polar solutes, but exclude nonpolar molecules of comparable sizes. Porins have been subdivided into specific and nonspecific types. For a particular solute, specific porins show a comparatively large diffusion rate at low concentrations and saturation effects at high concentrations. In contrast, nonspecific porins act like inert holes, their diffusion rates being proportional to the solute concentration. A general description of the porin architecture was derived from electron microscopy studies (3). The first porin crystals suitable for X-ray analysis were obtained from Escherichia coil (4). The first atomic structure was that of the major porin of Rhodobacter capsulatus (5-9). An analysis of other crystals using molecular replacement with the R. capsulatus porin structure showed that the 16-stranded [~-barrel fold is present in a number of porins (I0). Three more porin structures are now known in atomic detail (11); they confirm the general picture derived from the first structure characterization. After careful protein preparation the major porin of R. capsulatus yielded particularly good crystals (5). As shown by amino acid sequence analysis (6), one subunit consists of 301 amino acids. The crystal structure has been solved at 1.8 resolution and refined to a crystallographic R-factor of 18.6% using a 97% complete data set (9). The model of one subunit contains all polypeptide atoms, 274 water molecules, 3 calcium ions, 3 detergent molecules (octyltetraoxyethylene, CsE4), and 1 bound ligand. Since this ligand could not be identified, it was modeled as C8E4. Each porin subunit consists of a very large 16-stranded antiparallel [~-barrel and 3 short r All ~strands are connected to their next neighbors. The loops at the bottom end of the barrel are short and smooth. In contrast, the top end of the barrel is rough and contains much longer loops. The longest loop has 44 residues and runs into the interior of the barrel where it constricts the pore size to a small eyelet (9). The chain fold of trimeric porin (Figure 1) shows that the [~-barrel height in the trimer center is lower than the barrel height facing the membrane. This center forms a three-pronged star composed of all three subunits resembling the hub and spokes of a wheel. The mobilities of the atoms in this central star are the lowest of the entire molecule as determined by the crystallographic temperature factors (9). The interior of this star is nonpolar, 18 phenylalanines (6 from each subunit) interdigitate tightly, while the surface is polar. The star contains all six chain termini paired in salt bridges. Trimeric porin can thus be described as a rigid central core constructed like

The Membrane Channel Porin

177

4

Figure 1. Stereoview of the Ca-backbone chain fold of trimeric porin taken from ref. (7). The view is from the external medium.

a water-soluble protein that is surrounded by three ~-barrel walls fencing off the membrane.

II. PORIN LOCATION A. Orientation in the Membrane A cut through the center of a pore shows the shape of the channel as illustrated in Figure 2. The aggregation to trimers connects the rear ~-barrel walls at the height levels indicated by dashed lines. The central part is low, giving rise to a common channel formed by all three subunits in the upper half. The three pore eyelets limiting the diffusion are located between the barrel wall I close to the molecular threefold-axis and the 44-residue loop 3 inside the barrel. The longitudinal position of the porin in the membrane is clearly defined by the nonpolar ring with a height of 24 A that surrounds the trimer and fits the nonpolar moiety of the membrane. The rough upper end of the J3-barrel (Figure 1) contains the larger loops with numerous charged side chains, whereas the smooth lower barrel end has only small loops with few polar residues. As shown by binding studies with antibodies and phages, the large polar loops face the external medium (12). Accordingly, the external medium is at the top of Figures 1 through 5. This orientation agrees well with other features of the outer surface of the trimer (Figure 3). While the nonpolar surface moiety faces the nonpolar interior of the lipid bilayer, the upper polar part with its numerous ionogenic side chains is most

178

GEORG E. SCHULZ

10,~'

l

Figure 2. Shapeof a porin subunit represented by a sliced 6 ,~ resolution density map calculated from the high resolution model (8). The density level is at 1 o; the distance from the viewer is indicated by shading. The sectional areas 1,2,3,4,5 and 6 belong to the ~-barrel wall at the trimer interface, the ~barrel wall facing the membrane, the pore size-defining 44-residue loop inside the I~-barrel, a smallish domain facing the external medium, the nonpolar moiety of the membrane, and the LPS core, respectively. As indicated by bars, there is no free space between sectional areas 2 and 3. likely connected to the lip~olysacchaddes (LPS) forming the external layer of the bacterial outer membrane. Presumably, the numerous Asp and Glu side chains are glued by divalent cations (like calcium and magnesium) to the carboxylate groups of the LPS cores. This network integrates the podn efficiently in the tough bacterial protection layer of crosslinked LPS molecules, avoiding a fragile protein-membrane interface. Furthermore, Figure 3 shows two girdles of aromatic residues along the upper and lower border lines between polar and nonpolar residues. The upper girdle contains mostly tyrosines pointing with their hydroxyl groups to the upper polar moiety of the membrane, while the lower girdle has phenylalanines pointing to the nonpolar membrane interior and tyrosines pointing to the polar part of the periplasmic membrane layer. These patterns are obviously significant. They were also observed in three recently elucidated porin structures (11). Moreover, there are four type-If' ~turns at the lower smooth end of the barrel. These four turns point with their peptide amides toward the membrane where they can form hydrogen bonds to the phosphodiesters of the pedplasmic layer of the bacterial outer membrane.

r e 3. Projection of the outer surface of the &barrel onto a cylinder as taken from ref. (9. All Imps at the periplasmic side are indicated. S u l c e areas A,B,C, and D contain the ionogenic side chains connecting to the LPS core, the nonpolar side chains facing the nonpolar interior of the membrane, the lefthand parts of the interface and the righthand parts, respectively. lonogenic atoms are emphasized by squares and polar atoms by circles. The aromatic girdles are obvious.

180

GEORG E. SCHULZ B. Packing in the Crystal

The exceptionally well ordered crystals of the R. capsulatus porin (9) have a packing s,;heme shown in Figure 4. It resembles the crystal packing of the photoreaction center (13). The polypeptides form only polar contacts. Apart from the subunit interface, there exists merely one contact type (i.e., between head and tail) giving rise to a trimer arrangement like that in a cubic closest packing. This contact contains five hydrogen bonds and one salt bridge and is obviously strong. The closest lateral distance between trimers is 15 ,~, and it occurs between nonpolar surfaces. Any conceivable contact through detergent molecules should therefore be very weak and should provide only a minor contribution to the crystal packing energy. The molecular packing in the crystal corresponds to a stack of lipid bilayers containing porins, suggesting that such bilayers can also be formed in the crystal. This hypothesis could be tested by measuring the electric conductivity of the crystal along and perpendicular to the threefold rotation axis. Such measurements can also be done after changing the detergent in the crystal (possibly to lipids) or after crystallizing with other detergents, which would yield information about the detergent(lipid)-protein contact.

C. The Aromatic Girdles Since the Tyr(Trp)-Phe-pattern (Figure 3) has been observed in all of the four known porin structures (9,11), it most likely reflects a function. I suggest that these bulky, aromatic side chains prevent conformational damage of the protein on

.

.

.

.

.

.

I

Figure 4. Crystal packing arrangement of trimeric porin molecules, the molecular threefold axes (&) are crystallographic. The space group R3 requires only one contact type to form a three-dimensional network. This contact (,-) is head to tail and polar. There is no lateral contact between the nonpolar surfaces (cross-hatche~, which are presumably covered by detergent.

181

The Membrane Channel Porin

.

.

.

.

~/ _

Jo,

i'-

.

"

!

~ . _. p o l a r

,

Figure S. Sketch describing the suggested shielding r61e of phenylalanines and tyrosines during relative movements of protein and membrane.

mechanical movements in the membrane as indicated in Figure 5. Any transversal wave in the membrane or any knocking at a porin trimer immersed in the membrane causes rocking movements that would expose nonpolar protein surface to a polar membrane layer and polar protein surface to the nonpolar membrane interior. Both contact types give rise to a large surface tension, which is likely to scramble the polypeptide conformation. Since the aromatic side chains rotate around their Ca-Cp-bonds much faster than the trimer can rock, they should be able to shield the respective surfaces against the wrong counterpart (Figure 5).

D. Folding Pathway Considering the chain fold of the homotrimer with the central rigid star, the prongs of which are connected by the three ~-barrel walls, a folding pathway can be suggested: At first, the three-pronged star folds like a water soluble protein. Since this star contains all chain termini, the remaining chain parts form three large loops of about 200 amino acid residues, each suspended between the three prongs. On contact with the nonpolar membrane interior, the three 200-residue loops arrange themselves to the actually observed most simple ~-barrel topology with all strands antiparallel and connected to their next neighbors. This folding process is straight~ forward as it requires no chain crossing over, or wrapping around. On folding, the nonpolar side chains of the 15-barrel residues orient themselves to the outside and face the nonpolar moiety of the membrane. Subsequently, the large 44-residue loop is inserted into the barrel. This loop supports the barrel at its center against the membrane pressure, and it defines the pore size.

Figure 6. Stereo view of a pore eyelet as taken from ref. (9).The pore eyelet is lined by negatively charged side chains at its upper rim and positively charged ones at the lower rim. There exists a strong transversal electric field between these charges as indicated by the rigidity ofthe neighboringarginines at the lower rim (see text). Water molecules (x)and bound ca2+(e)are given. The molecular threefold axis is indicated (A).

183

The Membrane Channel Porin

III.

P E R M E A T I O N PROPERTIES

A. The Pore Eyelet The eyelet of the channel is lined by ionogenic groups that segregate into positively and negatively charged rims (Figure 6). The positive rim is closer to the trimer center and consists of half a dozen Arg, Lys and His side chains. The negative rim is further to the circumference of the trimer; it contains about a dozen Asp and Glu residues located mostly in the 44-residue loop that is inserted into the ~-barrel. These abundant negative charges are partially compensated by two bound Ca 2+ions that tighten the rim structure appreciably. As a consequence, the removal of these Ca 2§ should change permeabilities. Actually, it was previously observed that the analyzed porin permits the diffusion of ATP only after the bacterium has been treated with EDTA (I 4).

B. An Electric Separator The juxtaposed positive and negative rims cause a transversal electric field across the eyelet. The field strength is best estimated from two arginines participating in the positive rim. These arginines are at van der Waals distance in good electron density, demonstrating that they are rigidly positioned in spite of their repelling positive charges (the crystals are at pH 7.2, the pK of Arg is 12.4). Such an arrangement seems only possible if the charges are fixed by a strong electric field. An X-ray structure analysis at 1.8 A resolution allows the assignment of reasonably well fixed water molecules (9). The eyelet contains quite a number of them (Figure 6). They orient their dipoles in the electric field (Figure 7). In the center of the eyelet there are mobile water molecules confined to a rather small cross section of about 4 A by 5 A. They form a hole through which a molecule with the cross section of an alky! chain should be able to permeate. The diffusion of a small polar solute is illustrated in Figure 8A. The solute is oriented by the transversal electric field formed by the charged rims of the eyelet and will remain oriented over the whole diffusion distance. The solute is oriented in the pore eyelet like a substrate on an enzyme. This reduces the entropy of the diffusion process in a fashion similar to that for a chemical reaction by binding to an enzyme surface. Without tumbling, the activation energy barrier for permeation is lowered and the diffusion of polar solutes is appreciably accelerated. Although the eyelet cross section that is free of fixed water molecules suffices for the penetration of alkyl chains, these still cannot permeate because they are blocked by the strong transversal electric field. The eyelet with the opposing charges at its rim like a charged capacitor that stores an electric energy, which is proportional to the high dielectric constant of 80 of oriented water molecules (Figure 8B) multiplied by the volume. Any incoming molecule with a lower dielectric constant (the respective value for an alkyl chain is about 2), that is, with

184

GEORG E. SCHULZ

/

Figure 7. Sketch of the pore eyelet indicating a ring of bound water molecules that are oriented in the transversal electric field. The rim of negatively charged side chains is strengthened by two strongly (Ca) and one weakly (Ca') bound Caz+. The depth of the pore eyelet is about 6 A. For a hydrated ion, always half the hydration shell has an energetically unfavorable orientation in the electric field. a smaller dipole than water, causes a decrease of the stored capacitor energy and therefore generates a force expelling the intruder. Consequently, the transversal field acts as an electric separator, facilitating the permeation of polar solutes while blocking off nonpolar ones. A special situation seems to arise for ions. When diffusing through the pore, half their hydration shell enters into an energetically unfavorable orientation (Figure 7). In the R. capsulatus porin, them exists a possible separate pathway for positive ions at one corner of the eyelet. This pathway is outside the electric field and completely

A

B

\

/

-c

Figure 8. Diffusion through the pore eyelet. (A) Sketch of a permeating zwitter ion ~alanine that remains oriented in the field. Its permeation is facilitated by entropy reduction. (B) Sketch of the pore eyelet as a capacitor explaining the high activation energy barrier for the permeation of nonpolar solutes.

The Membrane Channel Porin

185

lined by carbonyl and carboxyl oxygens. It may explain the observed, although low, cation selectivity of this porin. Since this channel is blocked by a weakly bound Ca 2+ ion, its importance can be tested by measuring cation currents as a function of Ca 2+ .concentration.

C. VoltageDependence There are observations indicating that some porins incorporated into black lipid films reduce their electric conductivity upon application of a voltage of approximately 100 mV across this film (2) (i.e., voltage closure occurs). Such a behavior can be understood when considering the shape of the pore (Figure 2). The membrane channel has a rather large cross section over its whole length except for the short distance of about 6 A at the diffusion-defining eyelet at the center. This implies that any voltage across the membrane (say 100 mV) lies essentially across the eyelet with its small cross section (certainly more than 90 mV). This is a strong longitudinal electric field. Since the energy gained by moving a single charge (e.g., the protonated e-amino group ofa lysine)over 100 mV is 10 kJ/mol and thus equivalent to the energy of a hydrogen bond, the strong longitudinal electric field along the eyelet can tear off any charged group that can be moved over the 6 ,~, length of the eyelet (which is possible for lysine). The applied voltage thus disrupts the eyelet structure and diminishes the permeability. In conclusion, the pores are vulnerable to such high voltages. This does not exclude, however, that some porins possess lids that move onto and close the pore in a longitudinal electric field.

IV. PORIN SPECIFICITY Porins have been subdivided into 2 classes, specific and nonspecific ones. The structurally known R. capsulatus porin has been previously classified as nonspecific. In the crystal structure, however, this porin is ligated by a small molecule (Figure 9). Moreover, efficient binding of tetrapyrrols to this porin has been reported (15). In the crystal, the bound ligand cannot be identified from the density. The binding site is formed between the 44-residue loop inserted into the barrel (where it forms the eyelet) and a smallish domain protruding to the external medium. Facing the external medium this binding site can pick up solutes at very low concentrations. After being bound close to the eyelet (Figure 9), the solute may subsequently dissociate and diffuse through the pore at a considerable rate. Binding and permeation should follow Michaelis-Menten kinetics and should show saturation effects as observed within the class of specific porins. Accordingly, the R. capsulatus porin previously assigned to the class of nonspecific porins is actually both specific and nonspecific, depending on the solute. Generalizing this observation, it is suggested that many porins belong to both classes, but that the specific solutes have been detected in only a very few cases.

186

GEORG E. SCHULZ

<

Figure 9. Sketch of the Cot backbone of one subunit with the pore eyelet and the observed solute binding site within the pore close to the eyelet. Note that the solute binds at the external side of the eyelet.

ACKNOWLEDGMENTS I thank A. Kreusch, E. Schiltz, J. Weckesser, M. S. Weiss and W. Welte for their essential contributions to the structure analysis of the R. capsulatus porin.

REFERENCES 1. (a) Benz, R., & Bauer, K. (1988). Permeation of hydrophilic molecules through the outer membrane of Gram-negative bacteria. Fur. J. Biochem. 176, 1-19. (b) Nikaido, H. (1992). Pofim and specific channels of bacterial outer membranes. Molec. Microbiol. 6, 435-442. 2. Jap, B. K., & Walian, P. J. (1990). Biophysics of the structure and function of porins. Quart. Reu Biophys. 23, 367--403. 3. (a) Engel, A., Massalski, A., Schindler, H., Dorset, D. L., & Rosenbusch, J. E (1985). Porin channel triplets merge into single outlets in Escherichia coli outer membranes. Nature 317, 643--645. (b) Jap, B. K., Walian, P. J., & Gehring, K. (1991). Structural architecture of an outer membrane channel as determined by electron crystallography. Nature 350, 167-170. 4. Garavito, R. M., & Rosenbusch, J. P. (1980). Three dimensional crystals of an integral membrane protein: An initial X-ray analysis. J. Cell. Biol. 86, 327-329. 5. Kreusch, A., Weiss, M. S., Welte, W., Weckesser, J., & Schulz, G. E. ( 1991). Crystals of an integral membrane protein diffracting to 1.8 A resolution. J. Mol. Biol. 217, 9-10. 6. Schiltz, E., Kreusch, A., Nestel, U., & Schulz G. E. (1991). Primary slructure of porin from Rhodobacter capsulatus. Fur. J. Blochem. 199, 587-594.

The Membrane Channel Porin

187

7. Weiss, M. S., Kreusch, A., Schiltz, E., Nestel, U., Welte, W., Weckesser, J., & Schulz, G. E. (199 I). The structure of porin from Rhodobacter capsulatus at 1.8 ,/~resolution. FEBS Lett. 280, 379-382. 8. Weiss, M. S., Abele, U., Weckesser, J., Welte, W., Schlitz, E., & Schulz, G. E. (1991). Molecular architecture and electrostatic properties of a bacterial porin. Science 254, 1627-1630. 9. Weiss, M. S., & Schulz, G. E. (1992). Structure of porin refined at 1.8 A resolution. J. Mol. Biol. 227, 493-509. 10. Pauptit, R. A., Schirmer, T., Jansonius, J. N., Rosenbusch, J. P., Parker, M. W., Tucker, A. C., Tsernoglou, D., Weiss, M. S., & Schulz, G. E. (1991). A common channel-forming motif in evolutionarily distant porins. J. Struct. Biol. 107, 136-145. 11. (a) Cowan, S. W., Schirmer, T., Rummel, G., Steiert, M., Ghosh, R., Pauptit, R. A., Jansonius, J. N., & Rosenbusch, J. P. (1992). Crystal structures explain functional properties of two E. coli porins. Nature 358,727-733. (b) Kreusch, A., Neubfiser, A., Weckesser, J., & Schulz, G. E. (1994). Structure of the membrane channel porin from Rhodopseudomonas blastica at 2.0 A resolution. Protein Sci. 3, 58-63. 12. Tommassen, J. (1988). Biogenesis and membrane topology of outer membrane proteins in Escherichia coli. In Membrane Biogenesis (NATO ASI series, Op den Karnp, J. A. E, ed.), vol. H 16, pp. 351-373, Berlin: Springer-Verlag. 13. Deisenhofer, J., & Michel, H. (1989). The photosynthetic reaction center from the purple bacterium Rhodopseudomonas viridis. Science 245, 1463-1473. 14. Carmeli, C., & Lifshitz, Y. (1989). Nucleotide transport in Rhodobacter capsulatus. J. Bacteriol. 17 I, 6521-6525. 15. Bollivar, D. W., & Bauer, C. E. (1992). Association of tetrapyrrole intermediates in the bacteriochlorophyll a biosynthetic pathway with the major outer-membrane porin protein of Rhodobacter capsulatus. Biochem. J. 282, 471-476.

This Page Intentionally Left Blank

ROLE OF PHOSPHOLIPIDS IN ESCHERICHIA COL/CELL FU NCTION

William Dowhan

I. II.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemical Nature of the Phospholipids . . . . . . . . . . . . . . . . . . . . . . A. Acyl Chain Diversity . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Head Group Diversity . . . . . . . . . . . . . . . . . . . . . . . . . . . . III. Biosynthesis of Phospholipids . . . . . . . . . . . . . . . . . . . . . . . . . . A. Synthesis of Phosphatidic Acid . . . . . . . . . . . . . . . . . . . . . . . B. Synthesis of CDP-Diacylglycerol . . . . . . . . . . . . . . . . . . . . . . C. Synthesis of Phosphatidylethanolamine . . . . . . . . . . . . . . . . . . D. Synthesis of Phosphatidylglycerol . . . . . . . . . . . . . . . . . . . . . E. Synthesis of Cardiolipin . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Role of Phospholipids in Cell Functions . . . . . . . . . . . . . . . . . . . . . A. Protein Translocation Across the Inner Membrane . . . . . . . . . . . . . B. Colicin Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Initiation of DNA Replication . . . . . . . . . . . . . . . . . . . . . . . . D. Phospholipid Polymorphism and Membrane Structure . . . . . . . . . . . E. Lactose Permease Function . . . . . . . . . . . . . . . . . . . . . . . . . F. Cell Motility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G. Energy Transduction Systems . . . . . . . . . . . . . . . . . . . . . . . .

Advances In Cell and Molecular Biology of Membranes and OrganeHes Volume 4, pages 189..217. Copyright 9 1995 by JAI Press Inc. -All rights of reproduction in any form reserved. ISBN: 1-55938-924-9

189

190 191 191 193 195 195 197 198 200 202 203 204 205 206 207 209 209 210

190

WILLIAM DOWHAN

H. LipopolysaccharideSynthesis . . . . . . . . . . . . . . . . . . . . . . . V. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I.

211 211 212 212

INTRODUCTION

Our understanding of the complex roles of phospholipids in normal cellular processes has increased dramatically in the past 20 years. Phospholipids were once thought of as static structural components of cell and organelle membranes. The physical chemical principles underlying the concept of the hydrophobic effect (1). which is the major force determining membrane structure, and the importam conceptualization of membrane structure detailed in the fluid mosaic model (2) are summarized elsewhere. Although these ideas sparked new interest in the study o! membrane related phenomena and lead to an appreciation of phospholipids as e more dynamic component of the membrane, phospholipids were still religated tc mostly a structural role. This structural role depends on the collective amphipathic properties of phospholipids that allows formation in aqueous environments oJ bilayer structures with a hydrophobic core composed of the fatty acid side chain~ and two surfaces made up by the hydrophilic charged head groups of the phos. pholipids; the major contributors to the charge properties of the membrane surface are the combination of the acidic and zwitterionic head groups. Such models of membrane structure emphasize the mobility of phospholipi& and the gross physical properties of these molecules as a group, but largely ignore the role of structural diversity of the members of this group of molecules and th~ importance of minor phospholipid species as metabolic intermediates, regulator) molecules and cellular signals. Considering the possible range of fatty acids ant hydrophilic head groups, bacteria such as Escherichia coil have over 100 phos. pholipid species while a eukaryotic cell contains an order of magnitude greate~ diversity (3). Since a relatively simple mixture of phospholipid species can form bilayer, why is there such structural diversity among phospholipids? Certainly thi., diversity plays an important role in determining the collective physical chemica properties of the membrane bilayer, which has a direct influence on membrane stability and on the catalytic properties of the enzymes embedded in and associatec with the membrane. On the other hand many of the minor species must have role; other than simply serving as structural components. In recent years several minor phospholipids have been identified as importan biologically active molecules, which has stimulated great interest in phospholipi& and their function. In particular, the role of phosphatidylinositols as metaboli~ signals (4) and membrane anchors (5) has received attention. Other metabolite~ derived from phospholipids such as diacylglycerol (6), arachidonic acid (7), anc platelet activating factor (8) also play important regulatory roles. In Escherichk

Role of Phospholipids in Escherichiacoli Cell Function

191

coli there are several examples of phospholipids as dynamic metabolic intermediates for important membrane components rather than being merely static components of the membrane. Components of both phosphatidylethanolamine and phosphatidylglycerol become integral parts of several outer membrane lipoproteins (9, 10) and of the membrane derived oligosaccharides (11, 12) located in the periplasmic space. The former molecules have both structural and catalytic roles. The latter molecules are synthesized in inverse proportion to the osmoladty of the growth medium and result in the formation of diacylglycerol from phosphatidylglycerol (13). This diacylglycerol-phosphatidylglycerol cycle may serve as a metabolic signal for transmitting changes outside the cell to the interior much as the diacylglycerol/phosphatidylinositol cycle in eukaxyotic cells (3). Given the great diversity among the major and minor phospholipid classes, there must certainly be many more unrecognized structural and regulatory roles for phospholipids. One of the major problems of determining such roles is the lack of an assay for phospholipid function. Determination of the importance of individual phospholipids or of a particular phospholipid composition for membrane structure has relied heavily on in vitro physical chemical studies using model membrane systems. Such studies have also focused on the affect of the phospholipid environ, ment on the properties of membrane associated processes and sought to optimize function, as defined by catalytic rate, by manipulating the phospholipid environment. Due to the difficulty of changing the phospholipid composition of the membrane in living cells, there has been little in vivo verification of the importance of composition to function, and the optimal conditions as defined by in vitro model systems may have little in common with the in vivo situation. Therefore, the involvement of individual lipids as regulatory molecules or as precursors along a biosynthetic pathway has been determined incidental to the study of a particular cellular process rather than by direct study of a particular lipid. The use of molecular genetic approaches coupled with biochemical studies has great potential for determining the role of individual phospholipids in normal cell physiology. There is a wealth of genetic and biochemical information relating to phospholipids in E. coli, which is summarized in several recent reviews (3, 14-16). The current review will only present sufficient genetic and biochemical information about phospholipid metabolism in E. coli to serve as a background introduction to recent approaches aimed at more precisely defining the role that phospholipids play in cellular processes.

II. CHEMICAL NATURE OF THE PHOSPHOLIPIDS A. Acyl Chain Diversity All of the glycerol-based phospholipids of E. coli have sn-glycerol-3-phosphate as their backbone to which is esterified fatty acids at the one and two positions (see

WILLIAM DOWHAN

192 1)

GLYCEROL-3-P

pIBa

~

FairyBcvt-ACP

I-ACYL-GLYCEROL-3-P 2) plBC ~ Fattyacyt-ACP

4,

o

o c.,oc., o

CH~-O-P-X OH PHOSPHATIDICACID D) 3) c~A CDP-DIACYLGLYCEflOL " ~ Gh/ce~-3-P L'Serif4 ~A ) pEEA g) ptrsA ~ PHOSPHATIDYLGLYCEROL-3-P PHOSPHATIDYLSERINE s) p.d

FCO2

PHOSPHATIDYLETHANOLAMINE

PHOSPHATIDYLGLYCEROL 8) e~.

~~- Phoq)ha~lglycerol v

CARDIOUPIN+ GLYCEROL

Figure 1. Pathway of phospholipid metabolism in E. coil and the associated genes. The moeity X attached to phosphatidic acid (Step 2) is defined for the product of each enzyme, respectively. The name of each gene is listed with the respective step catalyzed by the following enzymatic activities: (1) sn-Glycerol-3-phosphate acyltransferase (2) 1-Acyl-sn-glycerol-3-phosphate acyltransferase (X = OH); (3) CDPdiacylglycerol synthase (X = CMP); (4) Phosphatidylserine synthase (X = serine); (5) Phosphatidylserine decarboxylase (X = ethanolamine); (6) Phosphatidylglycerophosphate synthase (X = sn.glycerol-3-phosphate); (7) Phosphatidylglycerophosphate phosphatase (X = glycerol); (8) Cardiolipin synthase (X = phosphatidylglycerol); (9) CDP-diacylglycerol hydrolase.

Figure 1). E. coil also contain glucosamine-based phospholipids that make up the lipid A membrane anchor of the outer membrane lipopolysaccharide of Gramnegative bacteria; these phospholipids have been reviewed extensively elsewhere (I 7). Under normal growth conditions virtually all of the fatty acids in the cell are in ester linkage to the two types of phospholipids with trace amounts decorating other cellular components. For the glycerol-based phospholipids (hereafter referred to as phospholipids) the primary saturated fatty acid is palmitic acid

Role of Phospholipids in Escherichia coli Cell Function

193

(16:0) and the two unsaturated fatty acids are palmitoleic acid (cis-A 9'1~ 16:1) and cis-vaccenic acid (cis-All'12-18:1 ); myristic acid (14:0) and D-3-hydroxymyristic acid are significant components of E. coil but are restricted to the glucosaminebased phospholipids. Under some growth conditions E. coil phospholipids contain cyclopropane-containing fatty acids (cis-9,10-methylene hexadecanoic acid (17:0) and cis-I 1,12-methylene octadecanoic acid (19:0)) that arise from direct addition of a methylene residue across the double bonds of the above unsaturated fatty acids in existing ester linkage in phospholipids. The steady state fatty acid composition (chain length and saturation) of the phospholipids of E. coil is affected by the growth conditions (stage of growth, growth temperature, and composition of growth medium) through a complex interaction between the make-up of the available acyl-ACP (acyl carrier protein) pool, the specificity of the acyltransferases responsible for synthesizing phosphatidic acid, and the postsynthetic remodeling of phospholipids. These subjects are reviewed more extensively elsewhere (15).

B. Head Group Diversity The minimal phospholipid structure is phosphatidic acid that is simply snglycerol-3-phosphate esterified at the 1 and 2 positions with fatty acids (see Figure 1, Step 2). Over 95% of the moieties in phosphate ester linkage to the phosphate at the sn-3 position of the glycerol backbone of phosphatidic acid are ethanolamine (phosphatidylethanolamine [PE]), glycerol (phosphatidylglycerol [PG]), and another phosphatidylglycerol moiety (cardiolipin [CL]). Although there is some variance in the relative proportion of these three phospholipid classes from strain to strain and under different growth conditions, they range from 70-80% for PE, 15-25% for PG and 5-10% for CL (3). The best recognized change in composition is an increase in CL at the expense of PG as cells enter stationary phase of growth (18). Virtually all of the phospholipids are associated with the cell envelope fraction as components of the bilayer of the inner membrane and of the inner monolayer of the outer membrane; the outer monolayer of the latter membrane is made up primarily of the glucosamine-based phospholipids of lipid A (I 9). Limited studies have been reported on differences in head group composition between the inner and outer membranes. The outer membrane relative to the inner membrane phospholipid composition is enriched with respect to zwitterionic phospholipid, PE (83% vs. 67%, respectively) in comparison to the acidic phospholipids, PG plus CL (12% vs. 31%, respectively); the remainder of the phospholipids in each case is primarily lysophosphatidylethanolamine, which skews the proportion of zwitterionic phospholipids in the outer membrane even further (20). Since the outer leaflet of the outer membrane appears to be a monolayer of lipid A molecules (which are negatively charged), there is a large charge difference across this membrane. Association of calcium ion with the outer membrane is important for maintaining

WILLIAM DOWHAN

194

9~ !

,'

~,, 9

,

CDP-DG? ,,,

Figure 2. Autoradiograms of two-dimensional thin-layer Chromatograms of chloroform extracts of 32pO4-radiolabeled E. coil Panels B and D are of material derived from a strain carrying a null allele of the pssA gene and therefore does not contain phosphatidylethanolamine. Panels A and C are of material derived from a wild-type strain for all phospholipid metabolic enzymes. PanelsA and B are the result of a 5-hour exposure to X-ray film while panels C and D are the result of a 72-hour exposure. (From DeChavigny et al. (22)).

integrity of this structure, which suggests a charge neutralization role (19). Mono-, di- and triacylglycerols represent about 1% of the lipid content of E. coil, but can reach levels of near 10% (primarily diacylglycerol) in mutants lacking diglyceride kinase (21). Further inspection of the nature of the phosphate-containing, chloroform-soluble fraction of E. coil reveals a vast heterogeneity in the remaining 5% of the phospholipids (defined as phosphate-containing, chloroform-soluble material). Long-term exposure to X-ray film of the radiolabeled phospholipid fraction of E. coil after separation by thin layer chromatography reveals the proportion of the diversity of this minor fraction (22) (see Figure 2). Particularly noteworthy is the large number of species missing from extracts of cells that cannot make PE relative to those that can make PE (Figure 2C vs. 2D). Pursuing the nature of these minor lipid components is still an understudied area and can be fruitful as demonstrated by Raetz and coworkers in defining the biosynthetic pathway for lipid A (17).

Role of Phospholipids in Escherichia coli Cell Function

195

III. BIOSYNTHESIS OF PHOSPHOLIPIDS The biosynthetic pathway leading to the major phospholipid classes of E. coli and the known genes encoding the enzymes involved are shown in Figure 1 (23). Contrary to the organization of the genes encoding the enzymes for many biosynthetic pathways in E. coil, there is no apparent clustering or coordinate organization and expression of the genes responsible for phospholipid biosynthesis (16) (Figure 3). Mutations have been isolated in each step of the pathway, all the respective genes have been cloned, the isolated genes have been sequenced in all cases except for plsC and cls, and all the enzymes have been purified to homogeneity and extensively studied except for those enzymes encoded by the plsC, pgpA and pgpB genes (23).

A. Synthesisof Phosphatidic Acid The committed step to the synthesis of the phospholipids of E. coil is catalyzed by a single enzyme, sn-glycerol-3-phosphate acyltransferase (Figure 1, Step 1) that

dgk R d

pls B

psd

cds A !PX B rnh

pls

B

pss A

pgs A

Figure 3. Relative location of genes on the 100-min circular genome of E. coll. The genes listed in the text and in Figure 1 are shown.

196

WILLIAM DOWHAN

is encoded by the plsB gene (24). This reaction results in the formation of l-acyl-sn-glycerol-3-phosphate. Fatty acyl derivatives of both CoA and ACP can serve as the acyl donors in vitro (2~, but the in vivo donor of newly synthesized fatty acids is acyl-ACP (26) while the donor of exogenously added fatty acids appears to be acyl-CoA (27). The enzyme has a preference for saturated fatty acids, which is consistent with the distribution of the fatty acids between the 1 and 2 positions of the phospholipids. However, the mixture of fatty acids available to the enzyme in vivo may be of equal or greater importance in determining fatty acid composition. Whether the enzyme itself is rate limiting is not clear since overproduction of the enzyme through gene amplification leads to a sequestering of excess enzyme in tubular phospholipid organeiles, which in itself would prevent an increase in the phospholipid to protein ratio of the membrane (28); however, even when this sequestration is largely prevented, high overproduction of enzymatic activity has little effect on phospholipid metabolism suggesting that regulation at the enzyme level is more important than at the level of gene expression. The only available mutant in the plsB locus results is a glycerol auxotrophy presumably due to a 10-fold increase in the Km for glycerophosphate of the mutant enzyme. Although the Km change is due to the mutation in the plsB gene the phenotype is dependent on a mutation in theplsX locus for which there is no known function (29). Therefore, precise conclusions based on the phenotype of this mutant must be made with caution, work with this mutant has established a relationship between cell growth and phospholipid metabolism. When starved for glycerol, this mutant stops growing when its phospholipid to protein ratio is reduced by half of wild-type levels (30). After addition of glycerol there is a lag in reinitiation of growth until the phospholipid to protein ratio returns to normal. How the cell senses this ratio and the mechanism by which it responds is still not known. Mutants in which the level of this enzyme could be systematically regulated to limiting levels would be useful in determining the factors that govern the protein to phospholipid ratio of the membrane. Addition of a second fatty acid by the l-acyl-sn-glycerol-3-phosphate acyltransferase (plsC gene product) results in the formation of phosphatidic acid (Figure 1, Step 2). Although there have been several reports indicating that mutations in the gene encoding this activity had been isolated, analysis of the mutations have been hampered by their complex phenotype (15). Coleman has reported convincing evidence of isolating a mutant in this activity and for cloning the gene encoding this activity (31). A temperature sensitive mutant was serendipidously isolated from a collection of mutants defective in membrane biosynthesis. At the restrictive temperature this mutant accumulated 1-acyl-sn-glycerol-3-phosphate, and no 1o acyl-snoglycerol-3-phosphate acyltransferase activity was detectable in this mutant under any in vitro conditions. Growth arrest occurred 90 min after a shift to the restrictive temperature with rapid loss of viability although there was little evidence of cell lysis. By complementation of the temperature sensitive phenotype, the plsC gene was cloned. Cells carrying multiple copies of the cloned gene have amplified

Role of Phospholipicls in Escherichia coli Cell Function

197

acyltransferase activity. The ability to overproduce the enzyme should make possible its eventual isolation. The available clone will make possible the construction of additional mutants at this locus, which should further our understanding of the early steps of phospholipid biosynthesis, particularly the verification of suggestive reports of molecular interaction between the two acyltransferases (Figure 1, Steps 1 and 2) (15). A secondary pathway for the synthesis of phosphatidic acid utilizing diacylglycerol kinase exists in E. coli that appears to be a salvage mechanism for diacylglycerol. A major source of this neutral lipid results from the transfer of sn-glycerol-3-phosphate from PG and ethanolamine phosphate from PE to the oligosaccharide precursor of membrane derived oligosaccharide (MDO) (11, 12). MDO levels are inversely proportional to the osmolarity of the growth medium, and in dgkA mutants the level of diacylglycerol can reach 8% (up from less than 1%) of the total lipid when cells are grown in medium of low osmoladty (21); this appears to be toxic to E. coli since under these conditions mutants in MDO biosynthesis can readily be isolated (32). Potential exists for coordinated regulation of the genes responsible for phosphatidic acid synthesis. The plsB and dgkA genes are transcribed in opposite directions from a region between the two open reading frames which must share overlapping but divergent promoters (33). A second locus termed dgkR when mutated results in a sevenfold increase in the transcription of the dgkA locus and a 30-50% decrease in the activity of the plsB gene product (34). Although little else is known about the function of the dgkR locus, it clearly has a reciprocal affect on the expression of these two loci responsible for phosphatidic acid synthesis.

B. Synthesisof CDP-Diacylglycerol The liponucleotide pool of E. coil comprises about 0.04% of the total phospholipid with the ratio of the dCDP/CDP form being about 0.9 (35). This pool is the immediate precursor to all the phospholipids. The relative rate of utilization of this pool by the next two enzymes in the pathway must have a dominant influence on the final zwitterionic and acidic character of the membrane phospholipid mixture. The (d)CDP-diacylglycerol synthase (Figure 1, Step 3) utilizes either CTP or dCTP with equal efficiency and appears to be the only activity to catalyze these reactions in vivo (36). The only mutant in this gene confers a conditional lethal sensitivity to growth in medium above pH 8 (35). The mutant enzyme has a pH optimum 0.5 units lower than the wild type enzyme and a greatly reduced activity in mutant extracts at all pHs tested. Growth of the mutants at elevated pH's results in accumulation of phosphatidic acid to levels as high as 30% of the total phospholipid. A similar observation, accompanied with growth arrest, is seen when cytidine auxotrophs are starved for cytidine, consistent with the importance of this enzyme to phospholipid synthesis (35); such mutants are a natural source for phosphatidic acid enriched membranes, which can be used as a starting point for

198

WILLIAM DOWHAN

the study of phospholipid synthesis in vitro (37). An unexplained phenotype of cds mutants is an increased resistance to the antibiotic erthyromycin, which may be related to increased levels of phosphatidic acid, even when grown under permissive conditions (35). Clearly, deeper understanding of the phenotypes and the effect of CDP-diacyiglycerol synthesis on cellular properties will require isolation of additional mutants in the cds gene. Although (d)CDP-diacylglycerol is the precursor to all the remaining phospholipids, it is also the substrate of a specific hydrolase (cdh gene) of unknown function (Figure 1, Step 9) (38). AMP is a strong inhibitor of the enzyme that distinguishes this activity from the hydrolytic activity of the phosphatidylserine synthase (see below). The role this activity plays in phospholipid metabolism or the function of this enzyme is unclear since a null allele of this gene confers no distinguishing phenotype on cells (39). However, the mode of hydrolysis carried out by the hydrolase suggests the enzyme catalyzes transfer of CMP to an acceptor molecule rather than hydrolysis (40). All other enzymes utilizing the liponucleotide as substrate transfer the phosphatidyl moiety to either water (in nonproductive hydrolytic reactions) or the hydroxyl of a substrate acceptor with the subsequent release of CMP. In the case of the hydrolase, the leaving group is the phosphatidyl moiety with the CMP being transferred to water or possibly another substrate acceptor.

C. Synthesisof Phosphatidylethanolamine The committed step (Figure I, Step 4) to the synthesis of the major phospholipid of E. coil and the only zwitterionic phosp.holipid in this organism is catalyzed by the phosphatidylserine (PS) synthase (41) (pssA gene). Since there is little turnover of PE in E. coil (42), the primary regulation of PE biosynthesis must occur at the level of the PS synthase. This is the only enzyme in this pathway that is not found tightly associated with the inner membrane fraction of the cell (43). In cell free extracts the enzyme is found tightly associated with the ribosomal fraction from which it can be dissociated with high salt or its liponucleotide substrate under optimal in vitro assay conditions (44). Membrane association can be induced by supplementation of the membranes with acidic phospholipids such as phosphatidylglycerol or cardiolipin (45). The most specific and effective membrane association is induced by the liponucleotide substrate. Catalysis in vitro is dependent on lipid substrate association with an inert hydrophobic-hydrophilic interface such as a nonionic detergent; physical association of the enzyme with nonionic detergents is also dependent on the presence of the lipid substrate (46). The deduced amino acid sequence of this enzyme shows a high concentration of positively charged amino acids at both the amino and carboxyl ends of the protein (22). These facts, taken together, suggest a model (47) for the association of the enzyme with the membrane surface during catalysis, which is dependent on its acidic lipid substrate. The general affinity of the enzyme for negatively charged surfaces, potentially through

Role of Phospholipids in Escherichia coli Cell Function

199

these same positive domains, suggests a possible regulatory role involving translocation of the enzyme to the membrane surface in response to an increase in the level of acidic phospholipids, resulting from the other branch of the biosynthetic pathway. Growth at the restrictive temperature of temperature sensitive mutants in this enzyme results in a reduction of the level of PE relative to PG and CL, which eventually leads to arrest of cell growth when PE drops to less than 40% of the total phospholipid (20, 48). The growth arrest phenotype can be suppressed without suppressing the reduced PE levels by adding 20 mM Mg 2+ to the growth medium. The critical questions of whether a second enzyme is responsible for this residual PE and whether PE itself is essential to E. coil were answered by isolating a null allele of the pssA gene incapable of synthesizing convincingly detectable levels of PE (22). This mutation is lethal under normal growth conditions. However, the requirement for PE can be suppressed by supplementing the growth medium with 5-10 mM divalent metal ions such as Ca 2+, Mg 2+, or Sr2+; Ba 2+ or molar concentrations of monovalent cations will not support the growth of this mutant. Under these conditions cells grow fairly normally with no PE, have an unaltered fatty acid composition, and demonstrate a normal ratio of membrane protein to phospholipid; the cells tend to filament, especially at reduced divalent metal ion concentrations. The fact that neither Ca 2+nor Sr2+reaches levels above micromolar in the cytoplasm of E. coil, due to an active antiport system for these ions (49), while Mg 2+ approaches 100 mM in the cytoplasm (50), suggests that the suppression of the growth phenotype by divalent metal ions is due to a correction of a structural defect outside of the inner cytoplasmic membrane. Due to the importance of divalent metal ions in stabilizing outer membrane structure, a likely site of this divalent metal ion requirement is in maintaining outer membrane integrity (19). The ability to eliminate the major phospholipid of this organism without losing cell viability was completely unexpected but has provided a vehicle to investigate the role of PE in normal, but not absolutely essential, functions of the cell (see below). Phosphatidylserine is immediately decarboxylated to form PE, which suggests no regulatory role exists for the phosphatidylserine decarboxylase (Figure 1, Step 5). Temperature-sensitive, conditional-lethal mutants have been isolated in the psd gene (51). Divalent metal ions also suppress the growth phenotype with accumulation of large amounts of PS without significantly altering the ratio of [PS + PE] to [PG + CL]. This result suggests that PS cannot replace the functions of PE and that the regulation of the balance between the zwitterionic phospholipids and acidic phospholipids does not involve PE directly, but may be a more subtle sensing of the ratio of the total phospholipids synthesized by each branch of the pathway. A null allele of this gene has not been isolated because construction of such a mutant is complicated by the psd gene apparently being the first gene of an operon containing at least one other essential gene (52). The synthesis and formation of the prosthetic group of the PS decarboxylase has several unique features. The enzyme is made in a proenzyme form (n-subunit) that

200

WILLIAM DOWHAN

is convened by a posttranslational event to a complex containing an unknown number of heterodimers of a and ~ subunits resulting from cleavage between Gly253 and Ser254 and the conversion of Ser254 to the pyruvate prosthetic group covalently attached to the amino terminus of the o~subunit (53, 54). This mechanism of posttranslational activation is followed by all o~-amino acid decarboxylases containing a pyruvoyl prosthetic group (55) including the PS decarboxylases from Chinese hamster ovary cells (56) and yeast (57).

D. Synthesisof Phosphatidylglycerol Since as much as 30-40% of the head group of PG can turnover per generation orE. coli, primarily in the synthesis of membrane derived oligosaccharide (58), the steady-state level of PG is much more dependent on a balance between synthesis and turnover than is the steady-state level of the head group of PE, which only turns over about 5% per generation (42). Other important but more minor turnover routes for PG are the synthesis of cardiolipin (59) and the major outer membrane lipoprotein, which is covalently attached to the peptidoglycan, as well as several minor membrane-associated lipoproteins (10). The phosphatidylglycerophosphate (PGP) synthase (60) catalyzes the committed step to the synthesis of the two major acidic phospholipids (PG and CL) in E. coil (Figure 1, Step 6) and therefore, must be involved in the regulation of the net charge on the E. coli membrane. PGP does not accumulate and is immediately converted to PG by a yet uncharacterized PGP phosphatase. The products of the pgpA and pgpB genes do dephosphorylate PGP in vitro and apparently to some extent in vivo (61), but a null mutant lacking both genes still contains considerable in vitro PGP-specific phosphatase activity and sufficient specific PGP phosphatase activity in vivo to maintain proper PG levels without accumulation of significant levels of PGP (62); therefore, the biosynthetic PGP phosphatase remains uncharacterized. The question of whether acidic phospholipids are essential to E. coil was first addressed by isolation of point mutants in the pgsA gene. The first sets of mutants isolated showed varying degrees of in vitro PGP synthase activity, which was inconsistent with the minor affect on in vivo PG levels (63). There are numerous examples in phospholipid metabolism where there is little correlation between residual levels of mutant enzyme activity detected in vitro and the level of the conversion of substrate to product observed in vivo. Such "silent mutants" result from a difference in mutant enzyme stability and function under in vitro versus in vivo conditions. Therefore, lack of appreciable in vitro activity in the face of little or no change in in vivo phospholipid composition cannot be taken as evidence for a large excess of catalytic capacity in vivo. Further mutagenesis of a partially defective pgsA mutant produced a temperature sensitive mutant that stopped growing and had significantly reduced PG levels at the restrictive temperature (63). However, this phenotype is a result of a second mutation at what later proved to be the lpxB locus that is involved in an early step

Role of Phospholipids in Escherichia coli Cell Function

201

in lipid A biosynthesis (17). It is still not clear how either dysfunction of the lpxB gene product or the accumulation of precursors to lipid A also results in inhibition of a mutant form of the pgsA gene product. Finally, Miyazaki et al. (64) were able to isolate a mutant (pgsA3 allele) that has barely detectable PGP synthase activity and displays a 10-fold reduction in the content of PG plus CL. Surprisingly, this mutant grows normally and displays no dramatic phenotypes. In light of the findings of Miyazaki et al. (64), Heacock and Dowhan (65) approached the establishment of the requirement for PG in essential cellular processes by constructing a null allele of the pgsA gene. First, a plasmid copy of the pgsA gene was inactivated in vitro by insertion of a drag marker. The null gene was then moved into the chromosome of a strain that carried a functional copy of the gene on a plasmid that was temperature sensitive for replication (23). Such a strain was viable at the permissive temperature for plasmid replication, but at the nonpermissive temperature daughter cells, which could not inherit a copy of the plasmid, stopped growing after existing PGP synthase and acidic phospholipid were depleted. No conditions have yet been found to suppress the growth phenotype of this null allele, but a suppressor of"leaky" mutations such as the pgsA3 allele has been found (66). The original background in which this latter allele was constructed also carried a mutation in the Ipp locus that made the cell defective in the synthesis of the major outer membrane lipoprotein and apparently now capable of growth even with a reduced ability to make PG. The link between the pgsA and lpp genes is through PG, which is the donor of diacylglycerol in a posttranslational modification event leading to the lipid moiety of the lpp gene product (10); the remaining fatty acids esterified to the amino terminus of the protein is derived by transacylation from existing phospholipids (9); the lpp gene product is not absolutely required by E. coli. The number of lpp molecules in the cell is on the order of the PG molecules when this lipid is reduced to less than 2% (19, 67). The high demand on the limited PG pool (or more likely the reduced capacity to synthesize PG) in a strain with the pgsA3 allele and an otherwise wild type background appears to be the basis for the arrest of cell growth and the suppressor affect of an Ipp mutation. However, an lpp mutation will not suppress the pgsA null allele which supports the essential role of PG (65). The suppressor role of the lpp mutation was further verified in another genetic background that also provided insight into some of the possible roles of PG in membrane structure and function. Heacock and Dowhan (68) introduced a #p[lacOP-pgsA] fusion in single copy into a pgsA null background so that PG catalytic capability could be regulated by controlling expression of a functional pgsA gene, via the lac promoter, using the gratuitous inducer IPTG. In such cells growth rate, enzyme level, and PG + CL content is proportional to the IPTG concentration in the growth medium. When IPTG is removed from growing cultures, PG content begins to fall when in vitro measured enzymatic activity drops to about 20% of normal wild type levels suggesting there is at least a fivefold

WILLIAM DOWHAN

202

catalytic excess of the activity in wild type cells. This conclusion is consistent with the observation of Jackson et ai. (69) that increasing the flux through the PG pool sevenfold by supplying an exogenous analog of membrane derived oligosaccharide did not significantly reduce the steady state proportion of PG in the total phospholipid pool. In the absence of IPTG, growth eventually arrests when the total PGP synthase activity is reduced to about 3% of wild type levels and the PG + CL pool is at 2-4% of total phospholipid. However, introduction of an lpp mutation in this strain suppresses the dependence on IPTG of both growth and growth rate but does not affect the low PC; content or the basal level of enzymatic activity expressed by the repressed lac promoter (70). Most interesting was the observation that growth arrest in the absence of IPTG is reversible (by addition of IPTG) for several hours after cessation of growth suggesting that acidic phospholipids are not important at their normal higher levels in maintaining membrane integrity. This mutant has been successfully used to investigate the role of PG in several cell functions (see below) and has good potential for analyzing the properties of other suppressors of reduced acidic phospholipid content.

E. Synthesisof Cardiolipin Synthesis of CL (Figure 1, Step 8) occurs by phosphatidyl transfer from one PC; molecule to the sn-1-hydroxyl of the unacylated glycerol of another PG molecule with the release of glycerol from the donor molecule (59). The CL synthase has been extensively purified (71), but has not been well characterized. Although the cloned gene is available (72) and a point mutant exists in this step (73), the gene sequence has not been reported. Furthermore, a null allele of the gene (74) has been made that demonstrates that cardiolipin may not be essential for E. coli. The major phenotype of such a mutant is incompatibility with either null alleles of the pssA gene (22) (even in the presence of divalent metal ions) or conditional mutants of the pssA gene grown under conditions approaching the restrictive conditions (75). There are several possibilities for this incompatibility. As discussed below, E. coli may require large amounts of either PE or CL for structural reasons. On the other hand, much lower concentrations of both of these phospholipids may be necessary for support of essential functions. Mutants carrying null alleles of either gene still contain trace amounts of the phospholipid normally synthesized by their respective gene products. In the case of a cls point mutant (74, 75), the residual level of CL decreased with increasing growth temperature in a pssl temperature sensitive mutant. This result suggests that the residual level of CL may be made by a side reaction of the PS synthase wherein the phosphatidyl acceptor would be PG rather than serine. This proposal has merit since CL in mitochondria is synthesized by such a mechanism (76) and the PS synthase has been shown to transfer the phosphatidyl moiety of CDP-diacylglycerol in vitro to other acceptors such as water and glycerol (41), although no transfer has been demonstrated to PG. There are also trace levels of what appears to be PE in pssA null mutants (22). However, the CL

Role of Phospholipids in Escherichia coli Cell Function

203

synthase reaction is reversible in vivo (77), so that low levels of PS and eventually PE could be made if serine rather than glycerol were the attacking group in the reverse reaction (see Figure l, Step 8). Therefore, if CL and PE were essential in trace amounts, then elimination of both the pssA an cls genes would be lethal as has been observed (22). Further investigation of this possibility may uncover presently unknown functions for these two lipids.

IV. ROLE OF PHOSPHOLIPIDS IN CELL FUNCTIONS As noted earlier, phospholipids form a bulk phase hydrophobic permeability barrier that defines the cell membrane, but this phase is also the matrix in which many enzymes reside and the surface which supports many cellular functions. Because of its influence on membrane surface charge, the nature of the ionic head group of membrane phospholipids must have profound dynamic and static effects on the properties of cellular processes associated with the cell membrane. Since phospholipids have no inherent catalytic activity, it has been difficult to systematically determine in vivo functions for phospholipids based on studying these molecules alone in vitro. Therefore, many proposed nonstructural roles for phospholipids rely on serendipitous in vitro observations of the affect of phospholipids on enzymatic reactions. Given the empirical nature of reconstituting phospholipids and proteins, many artifacts can be introduced into these in vitro systems. Verification of in vitro observations in vivo is also fraught with difficulty since, until recently, there have been few systems in which the affect of phospholipids on in vivo processes could be monitored. Mutants cannot be made directly in phospholipids but must be made in the biosynthetic enzymes which in vivo may result in pleiotropic effects due to the combined general structural role and specific functional role phospholipids play. In spite of these difficulties a combined in vivo molecular genetic and in vitro biochemical approach has led to recent advances in our understanding of phospholipid function. Mutants have been designed in which specific and controlled alterations in membrane phospholipid composition can be made without resulting in cell death (23). Phenotypic and biochemical characterization of these mutants can be carried out to make direct correlations between changes in cellular processes in vivo and changes in phospholipid composition. Once a direct biochemical or physiological relationship between a cellular process and a phospholipid alteration is established in vivo, in vitro biochemical approaches can be used to define the molecular basis for the involvement of phospholipids. Not only can new processes involving phospholipids be uncovered but in vivo verification of processes believed to involve phospholipids can be obtained.

204

WILLIAM DOWHAN

A. Protein Translocation Across the Inner Membrane

Outer membrane proteins of E. coil are synthesized in the cytoplasm in a preprotein form that must be translocated across the inner membrane prior to assembly in the outer membrane. The general process for many such proteins is described as follows (78). Translocation requires delivery of a partially unfolded preprotein to the membrane surface via a chaperone protein. The preprotein interacts or forms a complex with the membrane associated SecY/SecE/Bandl proteins and the cytoplasmic SecA protein. Complex formation activates latent SecA protein ATPase activity, and in conjunction with other membrane associated energy generating processes, the preprotein is translocated across the membrane, its amino terminal leader sequence is removed, and the protein is assembled in the outer membrane. The translocation process can be reconstituted in vitro using isolated inverted membrane vesicles and the required cytoplasmic factors. Given the steps involved in the above .process, it is not surprising that membrane phospholipid composition might influence the process. DeVrije et al. (79) first noted that in strains carrying the pgsA3 mutation, precursors to outer membrane proteins accumulated in the cytoplasm due to a reduction in their rate of translocation suggesting that a defect in the translocation process may be related to lack of PC; and/or CL. Using strains (68) in which the level of acidic phospholipids, PG and CL, could be regulated by IPTG levels in the growth medium, Lill et al. (80) established a direct involvement of PC; and possibly CL in protein translocation. The cytoplasmic SecA protein exhibited low ATPase activity, but its latent activity was uncovered when it associated with phospholipid vesicles containing PG or CL. Similar association and activation occurred with membranes isolated from the above mutant grown in the presence of IPTG (normal acidic phospholipid content) but not with membranes from the mutant grown in the absence of IPTG (depleted of acidic phospholipids). Binding of SecA and activation occurred in the absence of other cytoplasmic factors but was with the highest activity and stability when a complex was formed between the SecY/SecE/Band I protein complex in inverted inner membrane vesicles containing normal levels of PG, cytoplasmic SecA protein, and the outer membrane preprotein. Membrane translocation occurs upon the addition of ATI' to the complex associated with membranes from mutant cells grown with IPTG, but not with membranes derived from cells grown without linG. Both Hendrick and Wickner (81) and Kusters et al. (82) have shown that the translocation incompetent membranes can be made competent by in vitro supplementation with PC; or CL. The latter group also demonstrated a dose-response relationship for normal in vivo and for reconstituted in vitro translocation dependent on the acidic phospholipid content of membranes as adjusted by growth of the mutant at various levels of IFFG. They also showed that in vitro reconstitution occurred with all natural and several unnatural acidic phospholipids indicating that the primary requirement Was negative surface charge and not the structure of the acidic head group. These latter experiments also explain why protein translocation

Role of Phospholipids in Escherichia coli Cell Function

205

does not completely cease when PG + CL are reduced to 2% in leaky pgsA mutants in which the growth arrest phenotype is suppressed by an lpp mutation (79, 82). Under these conditions the precursors, phosphatidic acid and CDP-diacylglycerol, accumulate to 5% of total lipid (68) and must be able to substitute for the reduced PG. The above demonstration of the involvement of acidic phospholipids in protein translocation has lead to increased studies of SecA protein-lipid interactions. Breukink et al. (83) showed that SecA protein, which is itself acidic, will insert into phospholipid monolayers dependent on the presence of the acidic phospholipids PG or CL. This insertion was not inhibited by the presence of salt which actually increased the rate of monolayer association further confirming association through hydrophobic interactions with the monolayer; the insertion event is accompanied by a partial unfolding of the protein (84). The addition of ATP, ADP, or non-hydrolyzable analogs of ATP had differing affects on the affinity of SecA protein for these monolayers, indicating differences in the conformation of the protein as a function of the nucleotide bound. These results are consistent with the increased protease sensitivity of SecA protein when associated with acidic phospholipids and the resistance to protease in the presence of ATP (85). Now that an in vivo functional role for SecA protein interaction with acidic phospholipids has been established, further physical chemical studies are justified to establish the molecular details of this interaction. B. Colicin Assembly

Colicin A is a pore-forming bacterial toxin that kills cells by introducing a voltage-gated potassium channel in the cytoplasmic membrane (86). Effiux of cytoplasmic potassium dissipates the membrane electrochemical gradient resulting in cell death. Insertion of colicin A into planar lipid bilayers and liposomes in vitro has been shown to require acidic phospholipids and a pH of 5.5. A detailed model for this requirement proposes that a basic domain of the colicin interacts with acidic phospholipids that is a required step in assembling the pore in the cytoplasmic membrane (87). Using a similar mutant as above in which IPTG can be used to regulate acidic phospholipid content of membranes without affecting cell viability, van der Goot et al. (88) have demonstrated a definite requirement of acidic phospholipids for optimal functioning of colicin A in vivo. They first established that general energy metabolism was not affected by the level of PG in the mutant. The major effects of reduced PG content on the target cell membranes were an increase in the lag time for observing potassium effiux and a reduction in the potassium effiux rate. These results suggest that PG has an effect on the rate of insertion into the inner membrane and the properties of the channel, respectively. However, assembly ofcolicin A in vivo is a more complicated process than assembly in vitro and requires several outer membrane proteins (89). Due to the effect of

206

WILLIAM DOWHAN

reduced PG on the assembly of outer membrane proteins noted above, the results observed might be due to compromising the outer membrane assembly machinery. This pos.sible explanation of the results was ruled out by studying colicin N that parallels colicin A in its properties. Colicin N has a similar in vitro requirement for PG in assembly as does colicin A (90). However, colicin N has a higher isoelectric point (10.2 vs. 5.8) than colicin A, which should make it much less dependent for insertion on the acidic phospholipid content of the membrane (88), but still dependent on the outer membrane assembly machinery. If the PG effect were unrelated to alterations in the outer membrane assembly machinery, then colicin N would be expected to be insensitive to the level of acidic phospholipids, which was the result found (88). Therefore, the in vitro observations suggesting an involvement of acidic phospholipids in colicin assembly and function has relevance to the in vivo situation.

C. Initiation of DNA Replication It has been thought for some time that DNA replication occurs on a membrane surface, and evidence exists from in vitro experiments that initiation of DNA replication in E. coli may require acidic phospholipids. Initiation of DNA replication (91) occurs at the oriC locus of the E. coli chromosome and requires the organization of a multienzyme complex made up of individual cytoplasmic proteins including DnaAprotein (dnaA gene product) that initiates replication by promoting strand opening. The active form of the DnaA protein requires tightly bound ATP that turns over to ADP; either nucleotide form of the protein has high affinity for the specific binding sites at oriC, but only the former form is active in initiation. Acidic phospholipids have been shown to stimulate initiation of DNA replication in vitro by facilitating exchange between ADP with ATP to activate the inactive ADP-DnaA protein for new rounds of initiation (92, 93). Initially it was thought that CL was specific for this activation, but later it was found that PC; and other acidic phospholipids were equally effective as long as they contained unsaturated fatty acids, which is consistent with earlier findings of a positive link between membrane fluidity and initiation of DNA replication (94). Association of DnaA protein with oriC was required for rejuvenation of the inactive form of the protein (95) since treatment of free DnaA protein (even in the presence of either nucleotide) with acidic phospholipids inhibits subsequent binding to oriC. The latter phospholipid-inactivated form of the protein could be rejuvenated by either the DnaK protein (a heat shock/chaperone protein) or phospholipase treatment (96). Finally, cessation and reinitiation of total phospholipid synthesis using the plsB-glycerol auxotroph was found to be coupled to macromolecular synthesis and, in the case of DNA synthesis, specifically to the initiation step (30, 97). These observations are highly suggestive of a central role for acidic phospholipids in either organization or function of the DnaA protein-dependent initiation complex and fit the scenario seen for organization of the protein translocation machinery in that a

Role of PhospholipMs in Escherichia coli Cell Function

207

cytoplasmic ATPase is brought into contact with the membrane through binding to acidic phospholipids, followed by organization of a membrane associated complex to carry out function. What is lacking is an in vivo confirmation of the involvement of acidic phospholipids in this process. The following genetic approach should begin to provide the in vivo test of the above model for acidic phospholipid involvement. DnaA protein-dependent initiation of replication can be by passed by utilization of the alternate oriK initiation sites on the E. coli chromosome (98-100). Normally, these sites are suppressed because the RNA primer necessary for initiation of replication is degraded by the RNase H protein (rnh gene) at all sites other than at oriC where the primer is protected by the DnaAprotein; therefore, rnh mutants no longer require either DnaA protein or oriC for initiation and viability. If acidic phospholipids are limiting for DnaA protein function in acidic phospholipid mutants, then rnh mutants that eliminate the need for DnaA protein should be suppressors of pgsA mutants. Xia and Dowhan (119) have shown that introduction of an rnh mutation (null allele) into a strain that is dependent on IPTG for both acidic phospholipid synthesis and growth suppresses the IPTG dependence for growth without correcting the alteration in phospholipid composition. Furthermore, DnaA protein-independent initiation is dependent on RecA protein and therefore will not occur in a recA mutant (101). The above suppression of the IPTG-dependent pgsA mutant by an rnh mutation was also found to be dependent on the RecA protein. Finally, the retention as a function of the number of cell divisions of plasmids, which rely on an oriC-origin for replication, was dramatically reduced in this pgsA mutant when IPTG was removed from the growth medium; plasmids with a ColEl origin, which is not dependent on DnaA protein for initiation (98), were stable in such cells, even in the absence of IFTG. The above results with protein translocation across membranes and initiation of DNA replication suggest an important previously unrecognized role for acidic phospholipids, which are the minor phospholipids in most eukaryotic and prokaryotic membranes. These lipids may be integral components of many attachment or organization sites for complexes composed of both membrane associated and cytoplasmic components. Given the lack of significant changes in head group composition in normal E. coli, these lipids may only play a structural role in complex assembly. However, a more dynamic role involving regulation of formation of these complexes may be possible if the rate of formation of acidic phospholipids could be sensed by the cell. Many signal transduction processes involving transient interaction between membrane associated and cytoplasmic proteins potentially could be influenced by or require the involvement of acidic phospholipids.

D. PhospholipidPolymorphismand Membrane Structure Phospholipids assume a bilayer configuration in natural membranes that is essential for membrane integrity and barrier function. However, functions such as

208

WILLIAM DOWHAN

the interface between phospholipid and protein, membrane fusion and cell division, insertion of macromolecules into the membrane, and the passage of macromolecules through a membrane might be better served by a non-bilayer configuration (102), at least in local areas of the membrane. PE has long been recognized as a non-bilayer forming lipid. Depending on the fatty acid composition and temperature this phospholipid can assume an inverted hexagonal phase (Ha), which is essentially an inverted micelle structure where the head groups are sequestered inside and the nonpolar fatty acids are on the outside. In addition, CL can assume the Hn structure in the presence of the divalent metal ions Ca 2+ > Mg 2+ > Sr2+, in the indicated order of effectiveness (103). This type of structure within a lipid bilayer would introduce discontinuity that might be favorable for the above processes. Mutants of E. coli lacking PE and dependent on the above three divalent metal ions for viability (22) were investigated for correlative evidence between viability and the ability of dispersions of their phospholipids to assume non-bilayer structures (104). The mutant cells grow optimally in Ca2§ and Mg 2§ concentrations from 20-100 raM, but the minimum concentration that will sustain growth is 2 mM and 5 raM, respectively, for these ions. Sr2§ supports growth optimally at 15 raM, but not at 20 raM; Bae§ which does not promote the non-bilayer structure for CL, does not support growth at any concentration. The major phospholipids (over 90%) under all growth conditions for this mutant are PG and CL, but cells grown in the presence Ca 2§ which is more effective than the other ions in inducing non-bilayer structures for CL, has nearly half (27% versus 43%) the amount of CL compared to cells grown in the presence of the other two ions. Phospholipids extracted from these mutants grown at 37~ under various conditions were analyzed using 3~p NMR for their ability to undergo temperature-dependent transitions from the bilayer to the non-bilayer Hn phase. The mid-point for this transition for phospholipid dispersions from cells with a normal PE content was in a narrow range about 10~ above the growth temperature independent of divalent metal ions. For mutant cells grown in each divalent metal ion, the extracted phospholipids showed almost identical transition profiles to phospholipids extracted from wild-type cells provided they were suspended in the same divalent metal ion concentration used to support growth. However, ifphospholipids from Mg2+-grown cells (high in CL) are suspended in Ca2+, the mid point of the transition was below the growth temperature that in vivo might be expected to be lethal. Conversely, ifphospholipids from Ca2+-grown cells (low in CL) were suspended in Mg2§ the mid point of the transition was shifted to about 30~ above the growth temperature that in vivo might result in no capacity to form non-bilayer structures. In the case of Sr2+-grown cells the transition resembled the wild-type transition in the presence of 12-15 mM Sr2§ but did not occur at either 0 or 20 mM Sr2§ neither of which support growth. Ba2§ did not induce a phase transition with lipid dispersions made from cells grown in the presence of any of the above ions. These results suggest that wild-type cells have evolved to adjust their level ofnon-bilayer forming phospholipids so that some

Role of Phospholipid$ in Escherichia coli Cell Function

209

potential to form such structures near the growth temperature does exist, but that excess potential, which might result in loss of barrier integrity, is avoided. These studies have not established a precise molecular description for non-bilayer structures in vivo but do provide in vivo support for their importance and also points to oneof the functions of PE in normal cells, which might be replaced by the divalent metal ion induced non-bilayer properties of CL; this conclusion is supported by the loss of viability of a pss cls double null mutant (22).

E. Lactose Permease Function PE has been implicated in the functioning of the lactose (105) and proline (106) transport systems of E. coli. In both cases in vitro functional reconstitution of the carriers into liposomes of defined phospholipid composition has shown an absolute requirement for PE. There is now in vivo evidence for a role of PE in the functioning of the lactose permease (lacy gene product) provided through studies utilizing cells lacking PE (22, 107). The lacy gene product was present in the membranes of and expressed to the same level in both wild-type membranes and membranes from the mutant lacking PE even when the permease was overexpressed from a multicopy number plasmid. However, the Vmax for transport of lactose and its analogs was reduced 10- to 20-fold when the permease was associated with membranes devoid of PE. When wild-type inverted E. coil membrane vesicles loaded with thiomethyigalactoside (an analog of lactose,) were energized by adding ATR the analog rapidly exited in response to coupling of the carrier to the electrochemical gradient generated by ATP hydrolysis. No such ATP-stimulated exit was observed with PE-depleted membrane vesicles loaded with analog. These results are consistent with the in vitro data that PE is required for the proper functioning of the lactose carrier and also suggest that the carrier in the absence of PE cannot undergo the transitions necessary for coupling transport with energy to allow accumulation of substrate against a concentration gradient. However, both isolated membranes and whole cells from the mutant are not defective in generating and maintaining a similar electrochemical potential across their membranes (107, 108). Preliminary evidence using phoA-lacY gene fusions suggests that several of the transmembrane-spanning regions of the lactose carrier are misoriented in membranes from cells lacking PE (Bogdanov and Dowhan, unpublished data).

F. Cell Motility Cell motility and chemotaxis rely on complex, multicomponent membraneassociated components and signalling systems (109). Therefore, some or all of these systems might be expected to be affected by large changes in phospholipid composition. Mutants lacking PE (22) are nonmotile at all growth temperatures and therefore cannot respond via chemotaxis (110). At what point in these complex systems does the lack of PE have its affect? A mutant temperature sensitive in its

210

WILLIAM DOWHAN

ability to make PE (20) (i.e., pssts mutant) showed reduced motility at the permission temperature and no mobility at the normally restrictive growth temperature; under the latter conditions this mutant could grow in the presence of 20 mM Mg 2+ that had no effect on motility at the permissive temperature. A similar result was found with a temperature sensitive psd mutant (51) that can make PS but not PE. All of these mutants have a decrease in flagella under permissive conditions and no flagella under conditions restrictive for PE synthesis. Expression of several genes necessary for flagenar synthesis, motility, and chemotaxis were also repressed in parallel to the loss of PE, most notably theflD gene, which is the master controller of all of the transcriptional responses relating to motility, chemotaxis, and flagella synthesis. In general these systems are repressed under conditions of poor nutrition or stress, which indicates that this regulatory mechanism can also sense changes in membrane structure. The exact mechanism of how lack of PE is transmitted to the transcriptional machinery remains to be elucidated.

G. EnergyTransductionSystems Energy transduction is associated with membrane systems due to the generation and utilization of the energy in the form of a concentration gradient across a membrane barrier. Since in E. coli energy transduction is confined to the inner membrane, this process might be expected to be influenced by changes in phospholipid composition (111). Electrons from NADH are fed into the electron transport chain by two dehydrogenases and passed on to the cytochromes via ubiquinone and eventually to molecular oxygen (whole chain oxidase activity) (112). NADH dehydrogenase I has not been extensively studied, but it appears to be similar to complex I of mitochondria in that it is a multisubunit enzyme containing iron-sulfur centers and couples oxidase activity directly to proton gradient formation. NADH dehydrogenase II is a simpler single subunit enzyme that does not couple oxidation directly to proton gradient formation. Electrons also enter the chain via succinate and lactate dehydrogenases coupled to ubiquinone reduction. When wild-type membranes are compared to mutant membranes lacking PE (22), the level of all the above dehydrogenase and ubiquinone are the same (108). Both membranes, as noted above, have similar properties with respect to generation and utilization of membrane potentials. However, when the rate of whole chain-dependent oxidation of substrates is investigated, the oxidation of NADH dependent on dehydrogenase II is reduced four- to fivefold; dehydrogenase I-dependent oxidation was not investigated because this enzyme is unstable in isolated membranes from both mutant and wild type cells. Since the level of dehydrogenase II and ubiquinone were unaffected and the rate of electron flow via the other dehydrogenase (which follow the same path) was also unaffected, there appears to be a specific effect on the interaction of the dehydrogenase II with ubiquinone. The purified dehydrogenase II has associated with it 0.6 to 1.0 mole equivalents of ubiquinone and significant phospholipid, suggesting that the enzyme has strong

Role of Phospholipids in Escherichia coli Cell Function

211

affinity for both its cofactor and the surrounding phospholipid environment (113); this interaction certainly could be perturbed by changes in phospholipid composition. Demonstration of an in vivo requirement of PE for optimal functioning of the dehydrogenase II justifies further detailed in vitro studies on the role of phospholipids in the functioning of this enzyme and should lead to a better understanding of the role of phospholipid environment in the function of electron transport.

H. Lipopolysaccharide Synthesis The synthesis of lipopolysaccharide (LPS) of E. coli may involve PE in both its structure and in influencing the enzymes responsible for its synthesis. Ethanolamine-phosphate and -pyrophosphate are found covalently linked to the sugars of the inner core of LPS and from in vivo labeling studies the latter pyrophosphate derivative appears to be formed by direct transfer of ethanolamine-phosphate from PE (114). Presumably LPS from mutants lacking PE would not be decorated with ethanolamines, which are the only positively charged residues on LPS. These residues may be involved in salt-bridge crosslinks that may stabilize close packing of the chains on the outer surface of the cell. Disruption of this tight packing in mutants with reduced or no PE may explain their hypersensitivity to many antibiotics (22, 115). Divalent metal ions may supply the necessary links between neighboring chains to stabilize these structures in the absence of a source for ethanolamine. In vitro evidence also suggests that some of the sugar transferases responsible for synthesis of the inner core of LPS have a strong preference for PE (116-118). A detailed analysis of LPS from a PE deficient mutant should supply in vivo verification of the role of PE in both biosynthesis of LPS and its structure.

V. S U M M A R Y The above examples demonstrate the utility of a combined molecular genetic and biochemical approach to acquire both the in vivo and in vitro data necessary to document the functions of phospholipids in cells. A similar approach can be taken for any membrane related process either using the existing mutants ofphospholipid metabolism or designing similar mutants for specific purposes (23). This approach can certainly be generalized to eukaryotic microorganisms such as yeast. The most straightforward approach would be in studying the role of specific phospholipids in mitochondrial function since, many times, cell viability does require functional mitochondria. Although possibly a more complex problem, construction of regulatable genes that encode phospholipid biosynthetic enzymes which supply lipids to all the membranes of yeast may also be possible. Certainly, the concept Of specific phospholipids being integral parts of systems organized on the membrane surface

WILLIAM DOWHAN

212

can be generalized to all cell types and utilized in probing potential functions o f phospholipids in more complex systems.

ACKNOWLEDGMENTS The work from the author's laboratory was supported in part by grant GM 20487 from the National Institutes of Health.

REFERENCES I. Tanford, C. (1973). In The Hydrophobic Effecl:Formation of Micellesand BiologicalMembranes.New York: J. Wiley and Sons. 2. Singer,S.J.,& Nicolson,G.L.(1972). The fluidmosaic model of the smJcture ofcellmembranes. Science 175, 720-731. 3. Raetz,C. R. H., & Dowhan, W. (1990). Biosynthesis and function ofphospholipids in Escherichia coli. J. Biol. Chem. 265, 1235-1238. 4. Rhee, S. G., & Choi, K. D. (1992). Regulation of inositol phospholipid-specific phospholipase C isozymes. J. Biol. Chem. 267, 12393-12396. 5. Englund, P. 1". (1993). The slructure and biosynthesis of glycosyl phosphatidylinositol protein anchors. Annu. Rev. Biochem. 62, 121-138. 6. Bell, R. M., & Burns, D. J. (1991). Lipid activation of protein Hnase C. J. Biol. Chem. 266, 4661-4664. 7. Curtis-Prior, P. B. (1988). In Prostaglandins: Biology and Chemistry of Prostaglandins and Related Eicosanoids. Ely, UK: Cambridge Research Institute. 8. Prescott, S. M., Zimmennan, G. A., & Mclntyre, T. M. (1990). Platelet-activating factor. J. Biol. Chem. 265, 17381-17384. 9. Gupta, S. D., Dowhan, W., & Wu, H. C. (1991). Phosphatidylethanolamine is not essential for the N-acylation of a p o l i ~ e i n in Escherichia coll. J. Biol. Chem. 266, 9983-9986. 10. Sankaran, K., & Wu, H. C. (1994). Lipid modification of bacterial prolipoprotein: transfer of diacylglyerol moiety frownphosphatidylglyceroi.J. Biol. Chem. 269, 19701-19706. 11. Jackson, B. J., & Kennedy, E. P. (1983). The biosynthesis of membrane-derivedoligosaocharides. J. Biol. Chem. 258, 2394-2398. 12. Miller, K. J., & Kennedy, E. P. (1987). Transfer of phosphoethanolamineresidues from phosphatidylethanolamine to the membrane-derived oligosaccharides of Escherichia coll. J. Bacteriol. 169, 682-686. 13. Kennedy, E. P. (1982). Osmotic regulation and the biosynthesis of membrane-derived oligosaccharides in Escherichia coli. Proc. Natl. Acad. Sci. USA 79, 1092-1095. 14. Raetz, C. R. H. (1986). Molecular genetics of membrane phospl~olipid synthesis. Anna. Rev. Genet. 20, 253-295. 15. Vanden Boom, T., & C'ronan, J. E., Jr. (1989). Genetics and regulation of bacterial lipid metabolism. Anna. Rev. Microbiol. 43, 317-343. 16. Shibuya, I. (1992). Metabolic regulations and biological functions of phospholipids in Escherichia coil Prog. Lipid. Res. 31,245-299. 17. Raetz, C. R. H. (1990). Biochemistry of endotoxins.Annu. Rev. Biochem. 59, 129-170. 18. Tunaitis, E., & Cronan, J. E., Jr. (1973). Characterization of the cardiolipin synthetase activity of Escherichia coli cell envelopes. Arch. Biochem. Biophys. 155, 420-427. 19. Nikaido, H., & Vaara, M. (1987). Outer membrane. In Escherichia coli and Salmonella lyphimurium: Cellular and Molecular Biology. Vol. 1, E C. Neidhardt, J. L. lngraham, K. B. Low,

Role of Phospholipids in Escherichia coli Cell Function

213

B. Magasanik, and M. Schaechter, (FAs.). (pp. 7-22). Washington, DC: American Society for Microbiology. 20. Raetz, C. R. H., Kantor, G. D., Nishijima, M., & Newman, K. F. (1979). Cardiolipin accumulation in the inner and outer membranes of Escherichia coli mutams defective in phosphatidylserine synthetase. J. Bacteriol. 139, 544-551. 21. Rotering, H., & Raetz, C. R. H. (1983). Appearance of monoglyceride and triglyceride in the cell envelope of Escherichia coil mutants defective in diglyceride kinase. J. Biol. Chem. 258, 8068-8073. 22. DeChavigny, A., Heacock, P. N., & Dowhan, W. (1991). Phosphatidylethanolamine may not he essential for the viability of Escherichia coli. J. Biol. Chem. 266, 5323-5332. 23. Dowhan, W. (1992). Sn'ategies for generating and utilizing phospholipid synthesis mutants in Escherichia coil In Methods in Enzymology Vol. 209 E. A. Dennis, and D. E. Vance, (Eds.). (pp. 7-20). San Diego: Academic Press, Inc. 24. Bell, R. M. (1975). Mutants of Escherichia coli defective in membrane phospholipid synthesis. Properties of wild-type and Km defective sn-glycerol-3-phosphate aeyltransferase activities. J. Biol. Chem. 250, 7147-7152. 25. Green, P. R., Merrill, A. H., Jr., & Bell, R. M. (1981). Membrane phospholipid synthesis in Escherichia coil J. Biol. Chem. 256, 11151-11159. 26. Rock, C. O., & Jackowsld, S. (1982). Regulation of ~ p h o l i p i d synthesis in Escherichia coll. J. Biol. Chem. 257, 10759-10765. 27. Cronan, J. E., Jr. (1984). Evidence that incorporation of exogenous fatty acids into the phospholipids of Escherichia coli does not require acyl carrier protein. J. BacterioL 159, 773-775. 28. Wilkison, W. O., & Bell, R. M. (1988). sn-Glycerol-3-phosphate acyltransferase tubule formation is dependent upon heat shock proteins (htpR). J. Biol. Chem. 263, 14505-14510. 29. Oh, W., & Larson, T. J. (1992). Physical locations of genes in the rne (ams)-rpmF-plsX-fab region of the Escherichia coli K-12 chromosome. J. Bacteriol. 174, 7873-7874. 30. Mclntyre, T. M., Chamberlain, B. K., Webster, R. E., & Bell, R. M. (1977). Mutants of Escherichia coli defective in membrane phospholipid synthesis. Effects of cessation and reinitiation of phospholipid synthesis on macromolecular synthesis and phospholipid turnover. J. Biol. Chem. 255, 4487-4493. 3 I. Coleman, J. (1990). Characterization of Escherichia coil cells deficient in l-acyl-sn.glycerol.3-P acyltransferase activity. J. Biol. Chem. 265, 17215-17221. 32. Weissborn, A. C., Rumley, M. K., & Kennedy, E. P. (1992). Isolation and characterization of Escherichia coli mutants blocked in production of membrane-derived oligosaccharides. J. Bacteriol. 174, 4856--4859. 33. Lightner, V. A., Larson, T. J., Tailleur, P., Kantor, G. D., Raetz, C. R. H., Bell, R. M., & Modrich, P. (1980). Membrane phospholipid synthesis in Escherichia coil J. Biol. Chem. 255, 9413-9420. 34. Walsh, J. P., Loomis, C. R., & Bell, R. M. (1986). Regulation of diacylglycerol kinase biosynthesis in Escherichia coli: a ~ans-acting dgkR mutation increases transcription of the structural gene. J. Biol. Chem. 261, 11021-11027. 35. Ganong, B. R., & Raetz, C. R. H. (1982). Massive accumulation of phosphatidic acid in conditionally lethal CDP-diglyceride synthetase mutams and cytidine auxotrophs of Escherichia coil J. Biol. Chem. 257, 389-394. 36. Sparrow, C. P., & Raetz, C. R. H. (1985). Purification and properties of the membrane-bound CDP-diglyceride synthetase from Escherichia coil J. Biol. Chem. 260, 12084-12091. 37. Sparrow, C. P., Ganong, B. R., & Raetz, C. R. H. (1984). Escherichia coli membrane vesigles with elevated phosphatidic acid levels. A detergent-free system for in vitro phospholipid synthesis. Biochim. Biophys. Acta 796, 373-383. 38. Raetz, C. R. H., Hirschberg, C. B., Dowhan, W., Wickner, W., & Kennedy, E. P. (1972). A membrane bound pyrophospha1~se catalyzing the hydrolysis of CDP-diglyceride. J. Biol. Chem. 247, 2245-2247.

214

WILLIAM DOWHAN

39. Bulawa, C. E., & Raetz, C. R. H. (1984). Isolation and characterization of Escherichia coli strains defective in CDP-diglyceride hydrolase. J. Biol. Chem. 259, 11257-11264. 0. Bulawa, C. E., Hermes, J. D., & Raetz, C. R. H. (1983). Chloroform-soluble nucleotide$ in Escherichia coli. J. Biol. Chem. 258, 14974-14980. 41. Larson, T. J., & Dowhan, W. (1976). Ribosomal-associated phosphatidylsefine synthetase from Escherichia coil: purification by substrate-specific elufion from phosphocellulose using cytidine 5'-diphospho- 1,2-diacyl-sn-glycerol. Biochemistry 15, 5212-5218. 2. Kanfer, J., & Kennedy, E. P. (1963). Metabolism and function of bacterial iipids. J. Biol. Chem. 238, 2919-2922. 43. Raetz, C. R. H., & Kennedy, E. P. (1972). The association of phosphatidylserine synthetase with ribosomes in extracts of Escherichia coll. J. Biol. Chem. 247, 2008--2014. 40 Louie, K, & Dowhan, W. (1980). Investigations on the association ofphosphatidylserine synthase with the ribosomal c ~ n t from Escherichia coli. J. Biol. Chem. 255, 1124-1127. 50 Louie, K., Chen, Y.-C., & Dowhan, W. (1986). Substrate-induced membrane association of phosphatidylserine synthase from Escherichia coil J. Bocteriol. 165, 80.5-812. 6. Carman, G. M., & Dowhan, W. (1979). Phosphatidylserine synthase from Escherichia coil: the role of Triton x-100 in catalysis. J. Biol. Chem. 254, 8391-8397. 4~. Dowhan, W. (1992). Phosphatidylserine synthase from Escherichia coil In Methods in Enzymology. Vol. 209 E. A. Dennis and D. E. Vance, (Eds.) (pp. 287-298). San Diego: Academic Press, Inc. 48, Ohta, A., & Shibuya, I. (1977). Membrane phospholipid synthesis and phenotypic correlation of an Escherichia coli pss mutant. J. Bacteriol. 132, 434-443. 90 Gangola, P., & Rosen, B. P. (1987). Maintenance of intracellular cakium levels in Escherichia coli. J. Biol. Chem. 262, 12570--12574. 50. Chang, C.-F., Shuman, H., & Sorely,, A. P. (1986). Electron wobe analysis, X-ray mapping and electron energy-loss spectroscopy of calcium, magnesium, and monovalent ions in log-phase and in dividing Escherichia coil B cells. J. Bacteriol. 167, 93.5-939. 51. Hawrot, E., & Kennedy, E. P. (1978). Phospholipid composition and membrane function in phosphatidylserine decarboxylase mutants of Escherichia coil J. Biol. Chem. 253, 8213-8220. 52. Dowhan, W., & Li, Q.-X. (1990). Mechanism of formation of the pyruvate prosthetic group of phosphatidylserine decarboxylase of Escherichia coil (pp. 429-436). In 8th International Symposium on Vitamin B6 and Carbonyl Catalysis. T. Fukui, H. Kagamiyama, K. Soda, and H. Wada, (Eds.). Osaka, Japan: Pergamon Press. 53. Li, Q.-X., & Dowhan, W. (1988). Structural characterization of Escherichia coil phosphafidylserine decarboxylase. J. Biol. Chem. 263, 11516-11522. 4, Li, Q.-X., & Dowhan, W. (1990). Studies on the mechanism of formation of the pyruvate prosthetic group of phosphatidylserine decarboxylase from Escherichia coli. J. Biol. Chem. 265, 4111--4115. 55. van Poelje, P. D., & Snell, E. E. (1990). Pyruvoyl-dependent enzymes.Annu. Rev. Biochem. 59, 29-59. 6Q Kuge, O., Nishijima, M., & Akamatsu, Y. (1991). A cloned gene encoding phosphatidylserine decarboxylase complements the phosphatidylserine biosynthetic defect of a Chinese hamster ovary cell mutant. J. Biol. Chem. 266, 6370-6376. 57. Clancey, C. J., Chang, S.-C., & Dowhan, W. (1993). Cloning of a gene (psdl) encoding phosphatidylserine decarboxylase from Saccharomyces cerevisiae by complementafion of an Escherichia coli mutant. J. Biol. Chem. 268, 24580-24590. 58. Schulman, H., & Kennedy, E. P. (1977). Relation of tun~ver of membrane phospholipids to synthesis of membrane-derived oligosaccharides of Escherichia coli. J. Biol. Chem. 252, 42504255. 59. Hirschberg, C. B., & Kennedy, E. P. (1972). Mechanism of the enzymatic synthesis ofcardiolipin in Escherichia coll. Proc. Natl. Acod. Sci. USA 69, 648--651.

Role of Phospholipids in Escherichia coli Cell Function

215

0. Hirabayashi, T., Larson, T. J., & Dowhan, W. (1976). Membrane-associated phosphatidylglycerophosphate synthetase from Escherichia coil: purification by substrate affinity chromatography on cytidine 5'-diphospho- 1,2-diacyl-sn-glycerol Sepharose. Biochemistry 15, 5205-521 I. 61. Icho, T., & Raetz, C. R. H. (1983). Multiple genes for membrane-bound phosphatases in Escherichia coli and their action on phospholipid precursors. J. Bacterial. 153, 722-730. 2. Funk, C. R., Zimniak, L., & Dowhan, W. (1992). The pgpA and pgpB genes of Escherichia coil are not essential: evidence for a third phosphatidylglycerophosphate phosphatase. J. Bacterial.

174, 205-213. 63. Nishijima, M., & Raetz, C. R. H. (1979). Membrane lipid biogenesis in Escherichia coli: identification of genetic loci for phosphatidylglycerophosphate synthetase and construction of mutants lacking phosphatidylglycerol. J. Biol. Chem. 254, 7837-7844. 4. Miyazaki, C., Kuroda, M., Ohta, A., & Shibuya, I. (1985). Genetic manipulation of membrane phospholipid composition in Escherichia coli: pgsA mutants defective in phosphatidylglycerol synthesis. Proc. Natl. Acad. Sci. USA 82, 7530-7534. 65. Heacock, P. N., & Dowhan, W. (1987). Construction of a lethal mutation in the synthesis of the major acidic phospholipids of Escherichia coll. J. Biol. Chem. 262, 130'~,-13049. 6. Ased, Y., Katayose, Y., Hikita, C., Ohm, A., & Shibuya, I. (1989). Suppression of the lethal effect of acidic-phospholipid deficiency by defective formation of the major outer membrane lipopro, rein in Escheric~ia coli. J. Bacterial. 171, 6867-6869. 7. Neidhardt, F. C. (1987). Chemical composition of Escherichia coll. In Escherichia coil and Salmonella ~phimurium. Cellular and Molecular Biology. Vol. 1, E C. Neidhardt, J. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, & H. E. Umbarger, (eds.). (pp. 3-6). Washington DC: American Society for Microbiology. 68. Heacock, P. N., & Dowhan, W. (1989). Alterations of the phospholipid composition of Escherichia coli through genetic manipulation. J. Biol. Chem. 264, 14972-14977. 9. Jackson, B. J., Gennity, J. M., & Kennedy, E. P. (1986). Regulation of the balanced synthesis of membrane phospholipids: experimental test of models for regulation in Escherichia coll. J. Biol. Chem. 261, 13464-13468. 70. Dowhan, W. (1991). Role of phospholipids in cell function. In NATO ASI Series: Dynamics of Membrane Assembly. J. A. F. Opden Kamp, (Ed.). (pp. 11-32). Corsica, France: Springer-Verlag. 71. Hiraoka, S., Nukui, K., Uetake, N., Ohta, A., & Shibuya, I. (1991). Amplification and substantial purification of cardiolipin synthase of Escherichia coll. J. Biochem. 110, 443-449. 72. Ohta, A., Obara, T., Asami, Y., & Shibuya, I. (1985). Molecular cloning of the cls gene responsible for cardiolipin synthesis in Escherichia coil and phenotypic consequences of its amplification. J. Bacterial. 163, 506-514. 73. Pluschke, G., Hirota, Y., & Overath, P. (1978). Function of phospholipids in Escherichia coil: characterization of a mutant deficient in cardiolipin synthesis. J. Biol. Chem. 253, 5048-5055. 74. Nishijima, S., Asami, Y., Uetake, N., Yamogoe, S., Ohia, A., & Shibuya, I. (1988). Disruption of the Escherichia coil cls gene responsible for cardiolipin synthesis. J. Bacterial. 170, 775-780. 75. Shibuya, I., Miyazaki, C., & Ohta, A. (1985). Alteration of phospholipid composition by combined defects in phosphatidylserine and cardiolipin synthases and physiological consequences in Escherichia coll. J. Bacterial. 161, 1086-1092. 76. Tamed, K. T., & Greenberg, M. L. (1990). Biochemical characterization and regulation of cardiolipin synthase in Saccharomyces cerevisiae. Biochim. Biophys. Acta 1046, 214-222. 77. Shibuya, I., Yamogoe, S., Miyazaki, C., Matsuzaki, H., & Ohta, A. (1985). Biosynthesis of novel acidic phospholipid analogs in Escherichia coll. J. Bacterial. 161,473-477. 78. Joly, J. C., & Wickner, W. (1993). The SecA and SecY subunits of transiocase are the nearest neighbors of a translocating preprotein, shielding it from phospholipids. EMBO J. 12, 255-263. 79. de Vrije, T., de Swart, R. L., Dowhan, W., Tommassen, J., & de Kruijff, B. (1988). Phosphatidylglycerol is involved in protein translocation across Escherichia coil inner membranes. Nature (London) 334, 173-175.

216

WILLIAM DOWHAN

80. Lill, R., Dowhan, W., & Wickner, W. (1990). The ATPase of SecA is regulated by acidic ph~pholipids, SecY, and the leader and mature domains of precursor proteins. Cell 60, 271-280. 81. Hendrick, J. P., & Wickner, W. (1991). SecA protein needs both acidic phospholipids and SecY/E protein for functional high-affinity binding to the Escherichia coli plasma membrane. J. Biol. Chem. 266, 24596-24600. 82. Kusters, R., Dowhan, W., & de Kmijff, B. (1991). Negatively charged phospholipids restore prePhoE translocation across phosphatidylglyceroi depletedEscherichia coil membranes../. Biol. Chem. 266, 8659-8662. 83. Breukink, E., Demel, R. A., de Korte-Kool, G., & de Kruijff, B. (1992). SecA insertion into phosplmlipids is stimulated by negatively charged lipids and inhibited by ATP: a monolayer study. Biochemistry 31, 1119-1124. 84. Ulbrandt, N. D., London, E., & Oliver, D. B. (1992). Deep membrane penetration of SecA involves a partial unfolding event promoted by high temperature and anionic lipids. FASEB J. 6, ABS 481. 85. Shinkai, A., Mei, L. H., Tokuda, H., & Mizushima, S. (1991). The conformation of SecA, as revealed by its protease sensitivity, is altered upon interaction with ATP, presecretory proteins, evened membrane vesicles, and phospholipids. J. Biol. Chem. 266, 5827-5833. 86. Bonrdineaud, J. P., Boulanger, P., Lazdunski, C., & Letellier, L. (1990). in viva properties of colicin A: channel activity is voltage dependent but translocation may be voltage independent. Proc. Natl. Acad. Sci. USA 87, 1037-1041. 87. Pattus, F., Massotte, D., Wilmsen, H. U., Lakey, J., Tsemoglou, D., Tucker, A., & Parker, M. W. (1990). Colicins: prokaryotic killer-pores. Experientia 46, 180-192. 88. van der Goat, E G., Didat, N., Pattus, E, Dowhan, W., & Letellier, L. (1993). Role of acidic lipids in the translocation and channel activity of colicins A and N in Escherichia coil cells. Fur. J. Biochem. 213, 217-221. 89. Lazdunski, C., Baty, D., Geli, V., Cavard, D., Morion, J., Lloub~s, R., Howard, S. E, Knibiehler, M., Chartier, M., Varenne, S., Frenette, M., Dasseux, J.-L., & Pattus, E (1988). The membrane channel-forming colicin A: synthesis, secretion, structure, action and immunity. Biochim. Biophys. Acta 947, 445-464. 90. Wilmsen, H. U., Pugsley, A. P., & Pattus, E (1990). Colicin N forms voltage- and pH-dependent channels in planar lipid bilayer membranes. Fur. Biophys. J. 18, 149-158. 91. Sekimizu, K., Bramhill, D., & Kornberg, A. (1988). Sequential early stages in the in vitro initiation of replication at the origin of the Escherichia coil chromosome. J. Biol. Chem. 263, 7124-7130. 92. Yung, B. Y.-M., & Kornberg, A. (1988). Membrane attachment activates DnaA protein, the initiation protein of chromosome replication in Escherichia coil Proc. Natl. Acad. Sci. USA 85, 7202-7205. 93. Sekimizu, K., & gornberg, A. (1988). Cardiolipin activation of DnaA protein, the initiation protein of replication in Escherichia coil J. Biol. Chem. 263, 7131-7135. 94. Fralick, J. A., & Lark, K. G. (1973). Evidence for the involvement of unsaturated fatty acids in initiating chromosome replication in Escherichia coil J. Mol. Biol. 80, 459-475. 95. Crooke, E., Castuma, C. E., & Kornberg, A. (1992). The chromosome origin of Escherichia coil stabilizes DnaA protein during rejuvenation by phospholipids. J. Biol. Chem. 267,16779-16782. 96. Hwang, D. S., Crooke, E., & Kornberg, A. (1990). Aggregated DnaA protein is dissociated and activated for DNA replication by phospholipase or DnaK protein. J. Biol. Chem. 265, 1924419248. 97. Pierucci, O., & Rickert, M. (1985). Duplication of Escherichia coil during inhibition of net phospholipid synthesis. J. Bacterial. 162, 374-382. 98. Horiuchi, T., Maki, H., & Sekiguchi, M. (1984). RNase H-defective mutants of Escherichia coil: a possible discriminatory role of RNase H in initiation of DNA replication. Mol. Gen. Genet. 195, 17-22.

Role of Phospholipids in Escherichia coli Cell Function

217

9. Ogawa, T., Pickett.G. G., Kogoma, T., & Komberg, A. (1984). RNasr H confers specificityin the dnaA-dependent initiation of replication at the unique origin of the Escherichia coil chromosome in vivo and in vitro. Proc. Natl. Acad. Sci. USA 81, 1040-1044.

I00. yon Meyenburg, K., Boye, E., Skarstad, K., Koppes, L., & Kogoma, T. (1987). Mode of initiation of constitutive stable DNA replication in RNase H-defective mutants of Escherichia coil K-12. J. Bacteriol. 169, 2650-2658. I01. Kogoma, T., Skarstad, K., Boye, E., yon Meyenburg, K., & Steen, H. B. (1985). RecA protein acts at the initiation of stable DNA replication in rnh mutants of Escherichia coil K-12. J. Bacteriol. 163, 439 ~A. 102. Cuilis, P. R., & de Kru~ff, B. (1979). Lipid polymorphism and the functional roles of lipids in biological membranes. Biochim. Biophys. Acta 559, 399--420. 103. Lindblom, G., & Rilfors, L. (1989). Cubic phases and is, tropic structures formed by membrane iipids- possible biological relevance. Biochim. Biophys. Acta 988, 221-256. 104. Rietveld, A. G., Killian, J. A., Dowhan, W., & de Kruijff, B. (1993). Polymorphic regulation of membrane phospholipid composition in Escherichia coll. J. Biol. Chem. 268, 12427-12433. 105. Chen, C.-C., & Wilson, T. H. (1984). The phospholipid requirement for activity of the lactose carrier of Escl~richia coll. J. Biol. Chem. 259, 10150- ! 0158. 106. Chen, C.-C., & Wilson, T. H. (1986). Solubilization and functional reconstitution of the proline transport system of Escherichia coll. J. Biol. Chem. 261, 2599-2604. 107. Bogdanov, M., & Dowhan, W. (1995). Phosphatidylethanolamine is required for in vivo function of the membrane associated lactose permease of Escherichia coil. J. Biol. Chem. 270, 732-739. 108. Mileykovskaya, E., & Dowhan, W. (1993). Alterations in the electron transfer chain of Escherichia coil lacking phosphatidylethanolamine. J. Biol. Chem. 268, 24824-24831. 109. Macnab, R. M. (1992). Genetics and biogenesis of bacterial flagella. Annu. Rev. Genet. 26, 131-158. 110. Shi, W., Bogdanov, M., Dowhan, W., & Zusman, D. R. (1993). The pss and pds genes are required for motility and chemotaxis in Escherichia coll. J. Bacteirol. 175, 7711-7714. 111. Hoch, E L. (1992). Cardiolipins and hi.membrane function. Biochim. Biophys. Acta 1113, 71-133. 112. Yagi, T. (1991). Bacterial NADH-quinone oxidoreductases. J. Bioenerg. and Biomem. 23, 211-225. 113. Jaworowski, A., Mayo, G., Shaw, D. C., Campbell, H. D., & Young, I. G. (198 I). Characterization of the respiratory NADH dehydrogenase of Escherichia coil and reconstitution of NADH oxidase in ndh mutant membrane vesicles. Biochemistry 20, 3621-3628. 114. Hasin, M., & Kennedy, E. P. (1982). Role of phosphatidylethanolamine in the biosynthesis of pyrophosphoethanolamine residues in the LPS of Escherichia coll. J. Biol. Chem. 257, 1247512477. 115. Raetz, C. R. H., & Foulds, J. (1977). Envelope composition and antibiotic hypersensitivity of Escherichia coil mutants defective in phosphatidylserine synthetase. J. Biol. Chem. 252, 59115915. 116. End,, A., & Rothfield, L. (1969). Studies of a phospholipid-requiring bacterial enzyme. I. Purification and properties of uridine diphosphate galactose:lipopolysaccharide r transferase. Biochemistry 8, 3500-3507. 117. End,, A., & Rothfield, L. (1969). Studies of a phospholipid-requiring bacterial enzyme. II. The role of phospholipid in the uridine diphosphate galactose:lipopolysaccharide r transferase reaction. Biochemistry 8, 3508-3515. 118. Milller, E., Hinckley, A., & Rothfield, L. (1972). Studies of phospholipid-requiring bacterial enzymes. IH. Purification and properties of uridine diphosphate glucose:lipopolysaccharide glucosyltransferase I. J. Biol. Chem. 247, 2614-2622. 119. Xia, W., & Dowhan, W. (1995). In vivo evidence for the involvement of anionic phospholipids in initiation of DNA replication in Escherichia coli. Proc. Natl. Acad. Sci. USA 92, 783-787.

This Page Intentionally Left Blank

MECHANISM OF TRANSMEMBRANE SIGNALING IN OSMOREGULATION

Arfaan A. Rampersaud

I. II. Ill.

Introduction: Signal Transduction and Osmoregulation . . . . . . . . . . . . . Regulation of Porin Gene Expression by EnvZ and OmpR . . . . . . . . . . . General Studies of the ompB Operon; the ompR and envZ Genes . . . . . . . A. A Model for Osmoregulation by EnvZ and OmpR . . . . . . . . . . . . . B. Organization and Transcription of ompB . . . . . . . . . . . . . . . . . . C. Related Genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Expression of ompB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Well Characterized ompB Mutants . . . . . . . . . . . . . . . . . . . . . IV. Structure of OmpR AND EnvZ . . . . . . . . . . . . . . . . . . . . . . . . . A. OmpR Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Organization of EnvZ . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Relationships with Other Proteins . . . . . . . . . . . . . . . . . . . . . . V. Phosphorylation Studies of OmpR and EnvZ . . . . . . . . . . . . . . . . . . A. Phosphate Transfer Reactions . . . . . . . . . . . . . . . . . . . . . . . . B. Phosphorylation Mutants of EnvZ and OmpR . . . . . . . . . . . . . . . C. Forms of OmpR-P . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Importance of a Phosphate Relay Cycle for Porin Gene Expression . . . .

Advances in Cell and Molecular Biology of Membranes and OrganeHes Volume 4, pages 219-262. Copyright 9 1995 by JAI Press Inc. All rights of reproduction in any forumreserved. ISBN: 1-55938-924-9

219

220 221 222 223 225 227 228 228 234 234 235 236 237 237 238 239 240

220

ARFAAN A. RAMPERSAUD

E. Alternative Routes of OmpR Phosphorylation . . . . . . . . . . . . . . . VI. EnvZ Signaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Osmosensing by EnvZ . . . . . . . . . . . . . . . . . . . . . . . . . . . B. The Amino Terminus as the Sensor Domain . . . . . . . . . . . . . . . C. The Hybrid Tar-EnvZ Sensor . . . . . . . . . . . . . . . . . . . . . . . D. Activation of EnvZ by Local Anesthetics . . . . . . . . . . . . . . . . . VII. Multifunetional Properties of OmpR . . . . . . . . . . . . . . . . . . . . . . A. Conformational Changes in OmpR . . . . . . . . . . . . . . . . . B. Multimerization of OmpR . . . . . . . . . . . . . . . . . . . . . . . . . C. Additional OmpR Mutants . . . . . . . . . . . . . . . . . . . . . . . . . VIII. DNA-Binding and Transcriptional Properties of OmpR . . . . . . . . . . . . A. Binding to the ompF Promoter . . . . . . . . . . . . . . . . . . . . . . B. Binding to the ompC Promoter . . . . . . . . . . . . . . . . . . . . . . . C. Transcriptional Activation at ompF and ompC . . . . . . . . . . . . . . . IX. Other Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. The r Subunit of RNA-polymerase . . . . . . . . . . . . . . . . . . . . B. Integration Host Factor . . . . . . . . . . . . . . . . . . . . . . . . . . . X. Future Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

241 242 242 243 243 244 245 . . . 246 247 247 248 248 251 252 252 252 253 254 254 254

I. INTRODUCTION: SIGNAL TRANSDUCTION A N D OSMOREGULATION Escherichia coli are highly responsive to their external environment and have adaptive mechanisms that allow them to adjust their metabolism or behavior to suit particular environmental circumstances (.5, 23, 24, 98, 108, 130, 131). In many instances the adaptive mechanism involves signal-transmitting protein pairs, consisting of a sensor protein and a response regulator. The sensor detects an environmental stimulus and relays the information to the response regulator. In turn, the response regulator initiates changes in cellular metabolism or behavior (5, 101, 108, 130, 131). As sensors are often integral membrane proteins, transmembrane signaling can be an important part of the signal transduction process. A large number of sensor/response regulator protein pairs have been reported, and the broad aspects of their signaling mechanism have been reviewed (5, 14, 33, 101, 108, 130, 131). As a group they are related through their common mechanism of protein phosphorylation in which phosphate is transferred between a histidine residue of the sensor and an aspartate residue of the response regulator. The sensor class of proteins share segmented amino acid sequence similarities around a highly conserved histidine residue that represents their site of phosphorylation. Response regulators are often DNA-binding proteins, and as a class, share amino acid similarities at their N-terminus. A highly conserved aspartate residue in this same region represents the phosphate acceptor site.

Transmembrane Signaling in Osmoregulation

221

This chapter focuses on the EnvZ and OmpR couple that is a particularly well characterized sensor/response regulator pair in E. coli K- 12. They act in concert to differentially regulate two major outer membrane porin (omp) proteins, OmpF and OmpC, according to changes in external osmolarity (33, 45, 56, 86, 89, 114). EnvZ is a membrane-bound sensor while OmpR represents the response regulator. The goal of this chapter is to discuss the signaling mechanism for these two proteins and how this results in osmoregulation of porin genes.

II. REGULATION OF PORIN GENE EXPRESSION BY ENVZ AND OMPR The OmpF and OmpC proteins form homotrimedc pores in the outer membrane for the passive diffusion of small hydrophilic solutes (less than 600 Daltons) from the external environment into the periI~iasm (99, I00). They are major constituents of the bacterial outer membrane and can represent as much as 2% of the total protein content in the cell (99, 100). Due to their abundance, as well as their non-selective diffusion properties, the pot'in channels have a key role in determining the permeability properties of the outer membrane (99, 100). OmpF and OmpC are differentially expressed according to a variety of environmental conditions such as medium osmolarity (70, 144), carbon source (123), temperature (6), pH (50, 51), or the presence of membrane perturbants (27, 41, 107). Under conditions of low osmoladty, neutral pH, or at 30 ~ OmpF is mainly produced; increases in osmolarity or temperature, decreases in pH, or the presence of membrane perturbants cause OmpC levels to increase with a dramatic decrease in OmpF production. Despite fluctuations in the amounts of OmpF and OmpC protein, the total amount of porin in the cell remains constant (99, 100). Medium osmolarity is the main environmental condition used to study reciprocal changes in OmpF and OmpC. The OmpF and OmpC proteins are of the same size, are functionally similar, and show considerable sequence homology at both the protein and DNA levels (61% and 69%, respectively) (92, 99, 100). They are not essential proteins, and their expression patterns can be switched without adversely affecting cell growth (82, 99, 100). Given these data it is curious that an E. coli cell chooses to differentially regulate these highly similar proteins using a sensitive signal transduction system. The pore formed by OmpF is, however, slightly larger than OmpC and provides for a greater diffusion rate through this channel (99, 100). Thus, the relative amounts of OmpF and OmpC can influence the amount and/or type of solutes that diffuse across the outer membrane, and this is probably important to the cell. Osmoregulation of the ompF and ompC porin genes occurs at the transcriptional level through the combined action of the OmpR and EnvZ proteins. OmpR and EnvZ are encoded at the ompR and envZ genes, respectively, which together form the ompB operon (33, 43, 45). The ompB operon maps to 74 minutes on the E. coli

222

ARFAAN A. RAMPERSAUD

chromosome (43) and is unlinked to either ompF or ompC loci, which map to 21 minutes and 48 minutes, respectively, on the chromosome (9). In addition to regulation at the transcriptional level, OmpF mRNA is further regulated at the translational level by the micFgene (6, 7). The micF gene is located upstream to the ompC gene and is transcribed in the opposite direction (6, 7). It produces an RNA transcript complementary to the Shine-Dalgarno and initiation codon on the ompF mRNA. It is proposed that micF hybridizes with the ompF mRNA and blocks further translation of the message (6, 7). This type of regulation occurs primarily for thermal adaptation and does not appear to be a significant factor for osmoregulation (22). The micF gene will not be discussed any further in this chapter.

III. GENERAL STUDIES OF THE ompBOPERON; THE ompR AND envZGENES The ompB operon was identified as a regulatory locus in which mutations affected both OmpF and OmpC expression (122, 145, 147, 148). Clarification of the regulatory role of ompB came about by the creation and analysis of ompC-lacZ and ompF-lacZ operon fusion strains (37, 42-44). Such strains expressed 13-galactosidase in the same osmoregulatory fashion as the corresponding porin protein with their expression pattern being dependent on ompB. For example, strains that do not make functional OmpR protein or have lost the entire ompB operon do not express either ompF-lacZ or ompC-lacZoperon fusions (37, 42-44). In an envZ nonsense mutant (envZ22), expression of both operon fusions is reduced but not completely eliminated (36, 146). These studies establish some of the features of ompF and ompC regulation. First, the absolute level of OmpF and OmpC is controlled at the transcriptional level by ompB (43). Second, ompB is primarily responsible for osmoregulation of ompF and ompC transcription and coordinates this so that total podn content in the outer membrane remains constant (43). Finally, expression of ompF and ompC is absolutely dependent on OmpR and exhibits a strong but not strict requirement for EnvZ

(42-44, 37, 146). The ompB operon not only controls ompF and ompC expression but is also involved with regulation of microcin B- 17 biosynthesis (mcrB), expresgion of the gene for tripeptide permease (tppB), as well as regulation of an outer membrane protease (opr) (23, 39, 47). These additional regulatory processes are not completely understood, but indicate that ompB may have a more general regulatory role than previously suspected.

Transmembrane Signaling in Osmoregulation

223

A. A Model for Osmoregulation by EnvZ and OmpR Originally the OmpR protein was proposed to be a bifunctional regulatory molecule that shifted between two different states, one causing ompF expression and the other causing ompC expression (44). EnvZ was believed to be an envelope protein that modulated the two states of OmpR depending on environmental conditions (44). Extensive biochemical and genetic analyses have added further details to this model. One important result has been the demonstration that OmpR binds to recognition sequences in both ompF and ompC promoters and therefore is a transcription factor for both genes (87, 103, 142). In addition, EnvZ has been shown to phosphorylate and dephosphorylate OmpR and therefore is both a kinase as well as a phosphatase for OmpR (1, 30, 53-55). Finally, increasing OmpR-phosphate levels correlate closely with increasing OmpC and decreasing OmpF levels (33, 56, 86, 89). This has led to a general agreement that the relative phosphorylated state of OmpR determines reciprocal expression of ompF and ompC, possibly through differential binding of OmpR to high and low affinity target sites in their promoter regions (3, 4, II 8, 126). Based on the above points, a simple model of EnvZ/OmpR action can be created and is shown in Figure 1A. EnvZ is an inner-membrane protein having a putative sensor domain and a known signaling domain on opposite sides of the membrane. The cytoplasmic signaling domain undergoes phosphorylation and transfers phosphate to and from OmpR. At low osmolarity conditions EnvZ produces low levels of OmpR-phosphate (OmpR-P) through the combined action of its kinase and phosphatase activity. The amount of OmpR-P produced is sufficient for ompF but not ompC expression, and accordingly an OmpF +, OmpC" (F+C-) porin pattern is produced (Figure 1B). In this model, non-phosphorylated OmpR is considered to be nonfunctional. Environmental (osmotic) signals are believed to stimulate EnvZ so as to increase the amounts of OmpR-E At intermediate osmolarity enough OmpR-P is produced to allow activation of both ompF and ompC genes. At high osmolarity EnvZ further increases the amount of OmpR-E and when a particular level is reached, OmpR-P actively represses the ompF gene (Figure l A and I B). The ompC gene remains activated and thus produces an OmpF-OmpC + (F-C +) outer membrane phenotype (Figure 1B). While signaling by EnvZ involves changes in kinase and phosphatase activities, it is not clear whether this involves reduced phosphatase activity or elevated kinase activity. At present all that can be said is that these activities are not equivalent or that the ratio of these activities is not I. Environmental (osmotic) signals modulate these two activities with the signal thought to be received at the amino terminus of EnvZ. This information may be transmitted across the membrane to affect the cytoplasmic signaling domain. For OmpR, phosphorylation may cause conformational changes necessary for DNA-binding. These different conformations are not well understood and are best

ElzVZ

A.

AcP

~]amuto

,Mj)p

=..

low

~ P (low]~eb)

o==l=T~ p_h,x,p=~,.a Idu=~

JI

inte~

lB.

+

4.

"-

---

4.

4.

--4. MalE"

Figure 1. General model for EnvZ and OmpR signal transduction and the relationship between phosphorylated OmpR and porin expression. Panel A shows how phosphatase and kinase activities of EnvZ affect the amounts of non-phosphorylated OmpR (OmpR) and phosphorylated OmpR (OmpR-P). Phosphorylation of EnvZ (P) occurs on Histidine-243 at the expense of ATE Phosphate is transferred to Aspartate-55 on OmpR to form OmpR-P. Environmental conditions (low and high osmolarity) determine the relative kinase and phosphatase activities of EnvZ. The downward bold line on the far right indicates increased OmpR-P production. Panel B shows a diagram relating phosphorylated OmpR (OmpR-P) levels to osmolarity conditions and porin phenotypes. Non-phosphorylated (OmpR) is believed to be nonfunctional. A plus sign (+) indicates expression and a minus sign (-) indicates little or no expression. Pleiotropic forms of EnvZ are shown to greatly increase OmpR-P levels and cause reduced production of PhoA and MalE proteins. This aspect may be the result of a qualitative rather than a quantitative change in OmpR-P levels. 224

Transmembrane Signaling in Osmoregulation

225

described as low- and high-osmolarity forms. Whether OmpR is phosphorylated at one or more sites is also not clear, but current evidence suggests that EnvZ phosphorylates OmpR at a single site (26). In certain circumstances a form of OmpR-P is produced that is characterized by an F-C + phenotype and by its pleiotropic effects on other proteins such as Male and PhoA (19, 77, 145, 147). This is shown in Figure 1B as a highly phosphorylated form, since these OmpRs are closely associated with pleiotropic EnvZ mutants that accumulate OmpR-P (2, 136, 137).

B. Organization and Transcription of ompB The genes for ompR and envZ have been cloned and sequenced (20, 96). The ompR gene encodes the 239 amino acid OmpR protein (29.5 kD) while envZ encodes the 450 amino acid EnvZ protein (50 kD). Their amino acid sequences as well as other details are shown in Figures 2 and 3. The genes overlap by four base

Met Gin Glu Am Tyr L3m= lie I ~ Olu Aq[ T ~

Leu Thr Glu Gin Gly Phe Gin VM ~

Leu~*Leu ~

Asp ~

Val Val~ Asp* Asp* Asp Met Arlg* Leu* Arg Ak

,.

~

~ . : .~-

8er Val Ale Asn Ala* Glu Gin Met-40*

VM Gly* Leu Glu* lie

Gly Ala Asp-100

..

Asp Tyr* He P r o i 4 m Pro, Phe Ash Pro ~ R

Leu-~0

Arll Gin Ber* ])he His Leu Met VMt Leu Asp" Leu Met Leu Pro Oly-60

Met Val* *Thr Ale Lye Oly Glu* Giu Vsl* Asp* Arg He ,

L~

9

Arg Any Gin AIa Gly* Lys Phe L ~ Leu Thr 8er* g~ly 8er Arx*~Asp Lys ne Asp Val* Gin Tyr Be Gin Thr

Olu* Leu Leu Ale Arg* He

Arg Ale# Ved Leu-120

Pro* 8er Glu,,Me,,t AJ,~t ,,Leu Oly* Arg Met Val Val Phe

Val Gu ,ArK Met Ala Ser

-

Asn Leu, Glu Leu lie Vsl

Glu A~m Phe Met Set Trp

Leu Leu Ale* Am Arg Gly

Pro Gly Vsl Leu Leu Leu

Gly* A l p Thr Arg, Leu L35 Ala* Arg* A~ Arg Gly Tyr

Gin phe Val Olu* Olu Va]

Glu ArK Set Tyr Glu Pro

Giu Giu His Set Asp Asp

Ala ,, Asp Pro, AIs Pro Gly

Ile Pro Glu Olu Hi, Lys

/ida Phe-140 Met Pro-160 ..Pro, ,Leu-,180 ~ Ser-200 Pro Arg-320* AIm-239

Figure 2. Sequence of OmpR and proposed secondary structure at the amino terminus. The amino acid sequence of OmpR (239 residues) was taken from reference 20. Superimposed above the amino terminus is a diagram of alternating a and 13 secondary structures (from ref. 131). The boundaries of these structures are not absolute assignments. Within the sequence, point mutations are indicated with a symbol (.). Other symbols correspond to amino acid differences between OmpR from E. coli B (u) or S. typhi (#) or to a three amino acid insertion (f) which replaces Val-53 with the sequence Ala-Leu-Glu. The amino acid residues that are deleted in the OmpR101 molecule are enclosed within the box. For further details about these amino acid replacements see Table 2 and the text.

ARFAAN A. RAMPERSAUD

226

I'1~243 Am-,~17 [~ , DXGXG GXG

.

.

.

m

.

I

m

Periplssmic domsin

Met Ar8 Tbr Leu Pro'Set Lye Leu Tyr Arl[ Trp Ale Glu Vsl Pro Asn Phe ArE ne Gin lie Pro Ash _His V.al" 8er" Olu Gin lle G]u Leu Am .Thr. AIm VsI Ala GI.y Thr Glu (]In Gly Thr Glu Leu Thr Arg

Art Leu Leu Gin Olu (]In Ar~ lie Tyr ASh

I

I

...........

. . . . . . .

m

m

m

I__

L

|'~

,

,,I

O ~ e doaudn

Leu Art Phe 8er Pro ArS 8er Set Phe AIm Arw Thr Leu P h o A l a 8er Ldm Vsl Thr Thr Tyr L e u V s l Val L e s * A m GInOGIn Pho Ash Lys Vsl Lea .Ms Tyr GIn Vsl Arg Met Leu Glu Asp Oly Thr Gin Leu Val Vsl Pro Pro AIa Phe Leu Gly Ue 8er Leu Tyr 8er Ash Olu AIs AIs Olu Glu His Tyr Glu Phe Leu Set His GIn Met Ala Gin Gin Leu Vsi Glu Vsl Am Lys Set 8er Pro Vsl Vsl Trp Leu Lys Trp Vsl Art Vsl Pro Leu Thr Glu lie His (]In Gly Asp Thr Leu Aia Be Met Leu ~ Ale Re Gly Gly Ala Trp A q Pro" Leu Val Asp Leu Glu His AIm Aims Leu Gin Val Pro Pro Leu Arg Olu Tyr Gly Ale 8er Glue Vnl Arl[ 8er Vid Met ,,A]a Als Oly Val Lye Gin Leu Ale Asp Asp Arg Thr Leu l i b " Asp. Leu...An~ Thr" Pro" Leu -Thr JAr~' lie Ant Leu Ala Thr Asp Gly Tyr Leu Ala Glu 8er lie Ash Lys Asp He Glu Glu Gin Phe lie Asp ~ Leu ~ Thr Gly Gin Glu Met Pro Met Ai. Vsl ~ Gly Glu Vsl lie Ala AIs Glu Set .Oly Tyr. Glu Leu ~yr Pro Gly 8er lie_ Glu Vsl Lys Met His [Pro . . . . . Leu . 8er ASh Met Vsl Vsl AsneAls JAIs At[ ~ O l y , . . . A s h G l y Trp lle GIu...Pr0 Ash Ars AI. Trp phe Qln JVsl Glu Asp Asp GIy Pro Arl[ Lys His Leu Phe Gin Pro !~h,, Vsd A r t GIy &.so 8er &Is Gly Leu,,Gly . Leu . . . Ala lie Vsl [-(]ln ArE He Vsi Asp ASh His Gly Thr Set Giu Arg Gly Giy Leu Set lie Arg .Ms Trp Leu Ala Gin Oly Thr Thr bys Olu Giy*450

Leu Pho Leu Arl AI8 Oly Thr Phe Leu Gly Thr Leu Olu Cys Glu ArK lie_ L~. Gly Ar~ Ash Pro

L e u ' l l e VJd-Z} Ala IDle Leu-40 Met Thr Asp-60 A ~ Glu lie.SO Oly Leu Arg-100 Oly Pro Thr-120 Trp Leu 8er-140 Set Pros Leu-160 Phe lie Ar~-180" Lye Gly lie-200 Arl[ Ala* Phe-220 Met AIs'IGIy-240* Met Met 8er-~N30 Ash Ale Ile-280 Met AIs Asp-S00 Glu lie Glu-.~0 Lye ArE "'Aia-340 Vsi 8 e r - S e r - 3 ~ He AI. Pro-380 / Thr He 8er-400 I Giy Met Leu-420 Vsl Pro Vsl-440

Figure 3. Organization and amino acid sequence of the EnvZ protein. Panel A shows the major structural features of EnvZ (450 residues), with the amino acid sequence shown at the bottom. The sequence was taken from reference 20. In the top diagram, the filled boxes (a) refer to transmembrane regions 1 and 2 (TM1 and TM2, respectively) while the periplasmic and cytoplasmic domains refer to portions of the molecule localized to the same compartments. The stippled boxes (1) correspond to the three conserved domains, Regions I, II and III and are labeled as I, II and III, respectively. Conserved sequences DXGXG and GXG or conserved residues His-243 and Asn-347 are indicated. In the amino acid sequence the transmembrane domains are shown in bold, and the Region I, II and III domains are boxed. Relevant amino acid residues in these regions are highlighted in bold and point mutations are indicated with a symbol (,). Point mutations correlate with those in Table 1.

pairs and are regulated by a promoter region in front of ompR (20). Both proteins are made from a long polycistronic mRNA that includes a 100- to 123-bp untranslated region (150). Transcription of ompB initiates at a number of start sites within the 5' regulatory region but occurs primarily at two sites, designated T1 and T2 (49). They are reciprocally regulated by cyclic AMP-cyclic AMP receptor protein (cAMP-CRP) complexes that increase the'l"2 transcript and has an overall positive effect on ompB expression (49). Production of the TI transcript is favored by integration host factor (IHF), which has a negative effect on ompBexpression (140). The molecular details of how these transcripts regulate ompB expression are not yet clear, but it has been

Transmembrane Signaling in Osmoregulation

227

proposed that the different transcripts may specify altered ratios of EnvZ and OmpR in the cell, which in turn affect porin production (49, 140). Altered porin patterns in CRP- and IHF-strains seem to support this idea (49, 123, 140).

C. Related Genes OmpB or highly similar ompB-like DNA sequences have been identified in E. coil B, S. typhimurium, Shigella sp., E. cloacae, K. pneumoniae, S. marcesens, (11, 72-74, 91, 128) and X. nematophilus (S. Forst, personal communication). The ompR and envZ genes from S. typhimurium and X. nematophilus have been cloned and sequenced, as well as the ompR gene from E. coil B (74, 91, and S. Forst, personal communication). Comparisons between deduced amino acid sequences reveals that OmpR from E. coli B differs from the E. coli K- 12 OmpR at positions 6 and 130, while the S. typhimurium OmpR differs from the E. coli K-12 OmpR at position 118 (Table 1). There are 21 amino acid differences between E. coli and S. typhimurium envZ genes with most of the alterations occurring in the cytoplasmic signaling region (Table 1). The changes do not occur in any of the amino acid

table 1. Amino Acid Differences between EnvZ and OmpR Proteins from E. coli K-12, E. coli B, and S. typhimurium LT-2a OmpRComparisons E. coil K-12 Lys-6 Ala-ll8 Aia-130

E. coliB

S. t)phi.

Ash-6 Ala-ll8 Thr-130

Lys-6 Pro-liB Ala-130

EnvZ Comparisons E. coil K-12

S. typhi.

E. coil K-12

S. ryphi

Leu-4 Ala-25 Ser-90 Arg-206 Set-260 Gin-262 Ala-303 Glu-320 Tyr-324 Pro-325 Glu-329

Met-4 Val-25 Thr-90 Leu-206 Gly-260 Glu-262 Ser-303 Asn-320 Gin-324 Aia-325 Gin-329

Pro-364 Asn-365 Ala-379 Thr-398 Ile-399 Val-413 Leu-422 Thr-441 Ala-443 Gly-450

Ser-364 His-365 Lys-379 Ser-398 Thr-399 Lys-413 I!e-422 Ala-441 Val-443 Ala-450

Note: aData taken from references20, 74.9 I.

ARFAAN A. RAMPERSAUD

228

residues thought to be important for EnvZ function. Most are found outside of the Region I, If, and 111domains that characterize the sensor class of proteins (see below). S. typhimurium and E. coil B need ompB for porin production. Both OmpF and OmpC are made in S. typhimurium (72-74), while E. coil B only makes OmpF, due in part to deletion of a portion of the ompC gene (106). The ompR gene from S. typhimurium complements an ompR deletion strain of E. coli K-12, indicating that the two OmpR proteins are functionally similar (72-74). Complementation experiments indicate that the E. coil B OmpR is not the same as the E. coli K-12 OmpR

(91). D. Expressionof ompB While OmpR and EnvZ are produced from the same mRNA their expression levels differ considerably (20, 74). Analysis of ompR-lacZ and envZ-lacZ expression in S. typhimurium indicates that EnvZ is produced at levels 10-fold lower than OmpR (74). In both E. coli K- 12 and S. typhimurium, the termination codon for the OmpR protein (TGA) overlaps the initiation codon (ATG) for EnvZ in the sequence ATGA resulting in a 1-bp shift in the reading frame of the envZ gene (20, 74). This arrangement, as well as the absence of a typical ribosome binding site immediately before the envZ gene, probably accounts for the reduced level of EnvZ expression

(20, 74). E. Well Characterized

ompBMutants

Much of our knowledge about EnvZ and OmpR has come from mutant analysis. Tables 2 and 3 show most, if not all, reported ompR and envZ mutants at the time this chapter was prepared and include their amino acid replacements and relative porin phenotypes. It should be emphasized that for OmpR mutants, phenotypes are based on either protein or lacZ expression levels and may not be directly comparable (lacZ determinations are much more sensitive). Interested readers should consult the primary reference for further details.

OmpR Mutants

OmpR mutants have been categorized into three genetic classes depending on their outer membrane phenotypes, ompRl, ompR2 and ompR3 (33, 44, 45). The ompRl class causes an F-C- phenotype; the ompR2 class produces an F+Cphenotype, while the ompR3 class causes an F'C + phenotype. Two other ompR groups have also been reported: ompR20, which produces OmpF but not OmpC at high osmolarity, and ompR40, which produces substantial amounts of OmpC at low osmolarity (96, 97). At least one allele for each class of mutants has been cloned and sequenced and, most are point mutations that alter a single amino acid (see Table 2 and Figure 2).

TransmembraneSignaling in Osmoregulation

229

Table 2. O m p R Mutations and Phenotypes at Low and High Osmolarities Phenotype Mutation a

LOWb

Lys-6 to Asn (E. coli B)

F+C-

Val-lO to lie*

F+C:!:

PhenoO'pe

Highf

Ref. g

F+C-

91

Giy-94 to Asp

F-C-

F,I,CI'

F+C+

F,I,CI"

F-C-

118 26

F-C1"

16

Glu-96 to Ala. (ompR96A )

F-C-

F+C+

Tyr- 102 to Cys. (ompRl02C)

F+C+

26

F"CI" FI"c-

F+C-

Pro-106 to Leu

F+C+

Glu-lll to Lys

F-C +

Leu- 16 to Gin (ompR77)

F+CFJ,CT (d)envZ l l sup

Ala-37 to Thr

F+C+

16 66

Arg-115 to Set. (ompRllSS)

F-C-

F-C +

96

(d)

F,I,c l '

83 I 18

FJ,CI'

118

F-Cn.d. Dn~P

16

F-C-

Gly- 129 to Asp

F+C-

Ala- 130 to Thr (E. coli B)

F+C-

57

F-C-

F~C~

Pro-131to Ser (ompR307)

F+C~

F+C+

Val-82 to Met

F~C~

Gly 141 to Asp Arg- 150 to Cys (ompR20)

F-C-

95

F+cl "

118

F+C-

91

F+cl "

118

F+cl "

118

Arg- 150 to His

F+c 9

F1"r

96

(c) 118

(c) Set-163 to Asn

F+C~

F-C-

16 66

AGly- 164 to Arg182, (ompRlOl)

F-C-

F~C1'

26

Aia-167 to Val

F~C -

F,I,CT

97

Arg- 182 to Cys

F-C-

F+cl "

118

F-C= (e) (c)

96

F'rc'r

118

F=C-

18

F+cl '

118

(d)

(e) FI"cT

(c)

F-C-

(r

(d)

Arg-71 to Cys (ompR40)

18

(r

(d)

Asp-55 to Ala

F-C-

(d)

(d) F-C=

118

(r

(c)

OmpRX6, 3 a. a. insertion at Vai-53 Asp-55 to Gin

67

(d)

(c)

Ser-48 to Phe

FJ,cT

(c)

Arg 15 to Cys (ompR36)

F+C+

95

F,[,C'I'

F-C-

(d)

Met-40 to lie

F-C-

(d)

(d) Asp- 12 to Val

18

(c) F,I,CI"

F-C-

F-C-

(d), endZind

(d) Asp- 12 to Gin

Ref.8

(d)

(d) Asp- 12 to Ala

Highf (d)

(d) Asp- I 1 to Asn

Lowb

(d) (c)

Asp- ! I to Ala

Mutationa

118

Ala-189 to Val

F+C(c)

(continued)

ARFAAN A. RAMPERSAUD

230

Table 2. (continued) PhenoOpe

Phenotype Mutation a

Lowb

Thr-83 to Ala

F+C + F,I,c T (d)env2~ nd

Glu-87 to Lys

High f

F• +

F,I,CT

R~ g

17

Mutation a

Low b

F• " F-C"

118

Gly-191 to Ser

~c-

F+C +

~8

Pc+

118

(r FTCT

118

FTCT

I 18

F,I,CI"

17

(c) Gly-94 to Ser

~c1'

(c)

Glu- 193 to Lys

Pc+

(ompR472)

Val-203 to Met,

F+C " F+C (e)envZll r

Val-203 to Gin

F+C -

(r Asp-90 to Ash

Ref. s

Arg- 190 to Cys

(r Val-89 to Met

High f

(d)e,wz ~d

(e)

F+C "

46

FTC~

]]8

(d)envZll s

Arg-220 to Cys (ompR324) and S. o'phi. .

.

.

F+C(e)

97

75

.

Holes: qVlutations are shown using standard three letter abbreviations and col'respond to those shown in Figure 2. The original residue is shown first and alleles are indicated. * OmpRV 101 also has a point mutation at the ribosome binding site (118). ~henotypes were taken from (c) measurements of [3-galactosidase of ompF-lacZ or ompC-iacZ operon fusions; (d) outer membrane porin patterns: or (e) both. /Relative phenotypes at low and high osmolarity conditions are indicated by a plus sign (+) for moderate to high levels of expression; a (• sign for low levels of expression or: a negative sign (-) for very low levels of exwession or none at all. Relative changes in ompF and ompC expression levels are indicated by a (,[) or (~) sign to indicate overall decreases or increases, respectively. Other abbreviations are envZll wp, em,ZIi suppressor: em,Z~, em,Z independent: D I I ~ upsuppressor of OmpR(D 11N); en~,Zllr insensitive to envZll and em,Zl l s sensitive to envZi l. # Numbers refer to references at the end of the chapter.

The ompRlO1 allele is a member of the ompR1 class of mutants (33, 44, 45). Its porin-minus phenotype is due to the lack of functional OmpR (31, 96). DNA sequencing of the ompRlOl gene reveals a loss of nucleotides between positions 547-789 that corresponds to an in-frame deletion of Gly164 to Arg182 (96). In diploid analysis, ompRlOl is recessive to most ompR mutants (44, 126), as would be expected for a null mutant. However, ompRlOl negatively compliments an ompR4 allele (a member of the ompR2 class of mutants), indicating that ompRlOl is not completely without function (44). This curious interference characteristic has also been shown for ompRlOl genes carried on multicopy plasmids in wild type cells (97). The ompR472gene is the best characterized ompR2mutant and produces an WC" porin pattern regardless of medium osmolarity (44, 126). In diploid studies ompR472 is recessive to wild-type ompR as well as to ompR3 (120, 126), but is not

Transmembrane Signaling in Osmoregulation

231

Table 3. EnvZ Point Mutations and Deletions EnvZ point Mutations Leu-18 to Phc Leu-35 to Gin,

Comment/ Phenotype

Ref.

EnvZ point Mutations

FCC-

136

Thr-247 to Arg,

Comment/ Phenotype

R~

F"C'c

83 136

(envZll)

F-C+

83

Pro-248 to Ser

F"C c

Pro-41 to Ser

F-C c

136

Arg-253 to His

F-C"c

Pro-41 to Leu

F-C c

136

Gin-44 to stop,

low levels of F and C

*

Tyr-351 to Ser, (envZ30)

Pro-159 to Set, (envZ250)

F-C-

118

* Sequence from S. Forst (unpublisheddata).

Arg- 180 to Cys

FCC-

136

Truncated E n v Z

Pro-185 to Leu

F-C c

136, 46

Ala-193 to Val

F'-Cc

46

(envZl60)

(envZ22)

Glu-212 to Lys

~c c

46

Aia-219 to Val,

low levels o f f

136, 149

(envZ3) Ala-239 to Thr

and C F-'C-

non-functional ompR36 suppressor

46 65

Comment

Ref.

envZllS(411 a.a.) From Ala-39 to

53

Gly-450 EnvZ* ('369 a.a.)

From Tyr-81 to Gly-450

EnvZc (270 a.a)

From Arg- 180 to Gly-450

lib

1 30

Gly-240 to Glu

F-C c

16

Val-241 to Gly,

F-C c

149

Asn-381 to D, Q, A, T, C or H

12

Ser-242 to Asp

F-C c

136

GI: G409A, G411A and A413S

152

His-243 to Val

envZ22-1ike

30

152

His-243 to Arg

envZ22-1ike

66

G2: G437A,G439A, A441G and 1442L

(envZ473)

Tar-EnvZ Mutants

EnvZ hlternal Deletions

gel.

Comment

envZAB, (412 a.a.). From Ala-38 to

Ref. 137

Ile-80

envZAC, (428 a.a.) From lie-80 to

137

Glu-106

affected by strong OmpC producing envZ mutants such as envZll or envZ473 (46, 126). Other R2 alleles such as ompR321, and ompR307 are, to varying degrees, sensitive to envZ473, indicating that not all ompR2 mutants are equivalent (46, 120, 126). An OmpR2-1ike phenotypr is also generated by an OmpR-LacZ fusion in which all but the last amino acid of OmpR is fused to the C-terminal portion of the ~-galactosidase enzyme (10, 126).

232

ARFAAN A. RAMPERSAUD

Genetic data indicate that ompR472 has lost the ability to negatively regulate ompF expression (126). The OmpR472 protein has a Val-to-Met replacement at position 203 in its amino acid sequence (V203M) (96) and has been shown to only bind activator sites in the ompF promoter (87, 142). A DNA-binding defect explains why ompR472 is not affected by envZll or envZ473, since envZmutants indirectly influence DNA-binding activity by affecting intracellular OmpR-P levels (2, 32, 126, 149). It is not clear whether the DNA-binding defect of OmpR472 is due to an altered protein-DNA interaction or an inability to adopt an appropriate DNA-binding conformation (i.e., a high osmolarity form) (46, 126). Based on genetic studies, OmpR472 requires EnvZ for its phenotype, indicating that the protein is not fixed into a conformation that acts independently of EnvZ (126). OmpR472 does appear to interact with the ompC promoter since in certain genetic backgrounds an ompR472 allele can activate ompC expression (139). Additionally, an OmpR mutant having a glutamine instead of a methionine substitution at Val-203 produces a strong !72 phenotype in a wild-type envZ background, but unlike OmpR472 is sensitive to an envZll niutant (46). For the OmpRV203Q mutant, an inability to adopt a (DNA-binding) conformation for ompC activation is thought to be the underlying reason for its R2 phenotype (46). The ompRl07, ompRllO and ompR36 alleles are members of the ompR3 class of mutants (phenotypically F-C +) and are dominant to virtually all other ompR alleles (44, 97, 126). In particular, ompRl07 and ompRllO maintain their OmpF-minus phenotypes even in an ompR472 background (126), and this led to the conclusion that OmpR actively represses ompF expression (126). The OmpR36 mutant has an Arg-15-to-Cys replacement (96), which seems to allow the protein to be phosphorylated but not dephosphorylated by EnvZ (2). Phosphorylated OmpR36 is thought to accumulate in an ompR36 strain and ultimately cause an F-C + phenotype (2). This is one of several studies that has correlated elevated levels of OmpR-P with ompC expression. At low osmolarity ompRl07, ompRllO strains require envZ for their phenotype (126). They can also produce the same porin patterns in the absence of envZ provided that cells are grown at high osmolarity (126). This may be explained on the basis of an EnvZ-independent phosphorylation (crosstalk) pathway that operates at high osmolarity to phosphorylate OmpR3 mutant proteins. The porin phenotypes of ompR3 mutants are very similar to those caused by envZll or envZ473 mutants (discussed next). However, in an envZll or envZ473 strain, OmpR causes pleiotropic effects on other genes (77, 147, 148) while ompRl07 and ompRllO alleles have no such effects (126). Thus, an ompR3 mutant is distinct from the pleiotropic form of OmpR. EnvZ Mutants

Table 3 shows most EnvZ missense mutants and their resulting porin phenotypes. The best characterized mutants are envZll and envZ473 (33, 44). Both produce an

TransmembraneSignaling in Osmoregulation

233

F-C + porin pattern, have similar phenotypes in diploid analysis, and are pleiotropic on the same subset of genes (33). They are dominant to most other envZ alleles as well as a number of ompR mutants (44, 126). The amino acid replacements in EnvZ11 and EnvZA73, (Thr-247 to Arg and Val-241 to Gly, respectively) are close to each other (83, 149), suggesting their common phenotypes are due to similar biochemical defects. In this regard, the EnvZ11 protein is known to lack phosphatase activity yet retains its kinase activity (2). A similar defect may also apply for EnvZ473 because high intracellular levels of OmpR-P are found in both envZll as well as envZ473 strains (32, 149). Once again a relationship can be drawn between elevated OmpR-P levels and an F-C + phenotype. The envZll or envZ473 alleles characteristically depress the levels of a number of proteins that are not normally osmoregulated by EnvZ (19, 77, 147, 148). The affected proteins include the PhoA and PhoE proteins that constitute part of the pho regulon (148); MalE, MalT, and LamB proteins that are part of the real regulon (19); and several iron regulated proteins (77, 147). The pleiotropic effects of envZ473 are mediated through OmpR (124). This suggests that EnvZA73 (and EnvZll) modifies the DNA-binding properties of OmpR in a way that interferes with the normal transcription of other genes. In an envZll mutant, transcription of the malT gene, encoding the positive regulator MalT, is decreased (19), and this subsequently decreases the transcription of male and lamb genes. Reduced PhoA levels are not due to decreased phoA transcription but rather are caused by a post-transcriptional defect (I 48). On the other hand, rpoA mutants (encoding the 0~subunit of RNA polymerase) suppress the PhoA- phenotype of envZ473 (35, 125), presumably by affecting OmpR. One way of explaining these rather complicated results may be to consider that pleiotropic forms of OmpR interfere with the transcription of genes encoding posttranslational processing enzymes. For example, the dsbA gene, encoding the processing enzyme disulfide oxidoreductase (10a), has been shown to influence OmpF as well as PhoA production (lOa, 110), and its transcription could be affected by pleiotropic OmpR. A number of EnvZ missense mutants such as EnvZ(P41S) (136), EnvZ(PI85L) (136, 46), EnvZ(212K) (46), and EnvZ(G240E) (16) behave identically to EnvZ 11. All produce an F'C + porin phenotype, and most have been shown either directly or indirectly to lack phosphatase but have kinase activity (16, 46, 136). In some cases they have been shown to cause pleiotropic effects on Male production as well (136). One ompR mutant, ompR77, has been isolated for envZll; it not only restores the F+C+ phenotype but also suppresses pleiotropic effects (83). OmpR77 is allele-specific since it has no effect on wild-type envZ or another envZ mutant, envZl60, that produces a F'C + porin pattern (83). This was one of the first studies to indicate a functional interaction between EnvZ and OmpR. Sequencing of the ompR77 gene identified a point mutation that created a Gin replacement at Leu-16 in the protein (83). In vitro, OmpR77 protein is poorly phosphorylated by EnvZ11 but appears to be normally phosphorylated and dephosphorylated by wild-type EnvZ (2). The results suggest that OmpR77 is a poor in vivo substrate for EnvZl 1 (2).

234

ARFAAN A. RAMPERSAUD

The envZ250 and envZ247 alleles represent a new class of mutants that are characterized by their F-C- outer membrane phenotypes (118). They are not null mutants since they are codominant with either wild-type envZ or an envZ473 allele (118); null mutants such as envZ22 are completely recessive to all envZ alleles (36). Furthermore, strains missing the EnvZ protein (such as envZ22 or AenvZ) still produce low levels of OmpF and OmpC due to EnvZ-independent phosphorylation mechanisms (see below). Both EnvZ250 and EnvZ247 proteins have been shown to lack kinase but not phosphatase activity, and this activity results in the continuous dephosphorylation of OmpR. The lack of OmpR-P in these cells prevents expression of either porin gene (118). This is the one of the prime reasons for considering nonphosphorylated OmpR as nonfunctional.

IV. STRUCTURE OF OMPR AND ENVZ EnvZ and OmpR are related to other proteins involved with adaptive processes. These include the (sensor/response regulator) CheA/CheY proteins that facilitate chemotaxis; the NtrB/NtrC pair that mediates adaptation to nitrogen availability, and PhoR/PhoB proteins that mediate adaptations to phosphate availability (.5, 14, 101, 108, 131). There are residues or amino acid sequences that are common to most members of the sensor or response regulator classes. These are pointed out below along with current ideas about their potential function.

A. OmpR Structure Wild-type and mutant OmpRs as well as several N-terminal and C-terminal fragments have been purified to homogeneity (26, 63, 69, 103, 133) and the full length non-phosphorylated protein shown to be monomeric (63). OmpR has a modular organization with an N-terminal phosphorylation domain and a C-terminal DNA-binding domain (26, 69, 133, 142). In general, this modular organization is a recurring theme among response regulators (71, 101, 108, 131). A truncated OmpR protein, containing the first 122 amino acids (Met-1 to Arg-122), has been created through molecular cloning techniques and shown to be phosphorylated as well as dephosphorylated by EnvZ (69). This not only localizes the phosphorylation region to the amino terminus of OmpR but also shows that this domain folds independently of its C-terminus into a structure that interacts with EnvZ. N-terminal OmpR cannot directly regulate porin genes, because it does not have a DNA-binding domain, but it may do so indirectly. In an envZll strain, overproduction of N-terminal OmpR changes porin patterns from an F-C* to an W'C§ phenotype (94). The truncated OmpR is thought to compete with intact OmpR for phosphoryl groups on EnvZ11 (94) and thereby reduce the amount of full-length OmpR-P in the cell. Due to decreased levels of intact OmpR-P, ompF expression

TransmembraneSignaling in Osmoregulation

235

is subsequently restored in the envZll strain. A similar reasoning could explain the curious negative complementation data observed between ompRlOl and ompR4 (44, 75, 97). Since the ompRlOl gene product is probably only a truncated N-terminal molecule, it may compete nonproductively with OmpR4 for phosphoryl groups on EnvZ and reduce the levels of phosphorylated OmpR4 protein. The C-terminus of OmpR contains the DNA-binding domain (69, 133, 142). LacZ-OmpR fusions and purified C-terminal fragments inherently bind to OmpR target sites in ompF and ompC promoters (69, 133, 142). Sequential shortening of LacZ-OmpR fusions indicates that the DNA-binding domain lies between residues 123 and 239 of OmpR (117 amino acids) (142). In comparison to the full length OmpR, in vitro DNA binding by a purified C-terminal fragment is relatively weak (69, 133). Additionally, expression of LacZ-OmpR or C-terminal versions in cells do not activate either porin gene (94, 142). These studies indicate that the amino terminus of OmpR is needed for full activity and probably provides additional functions that not only enhance DNA binding activity but also stimulates gene expression.

B. Organization of EnvZ EnvZ is a transmembrane protein of the inner membrane (28). Its topology consists of a periplasmic (115 amino acids) and cytoplasmic domain (271 residues) separated by two transmembrane segments, TM 1 (approximately 32 amino acids) and TM2 (approximately 17 amino acids) (33). There is also a short 15 amino acid signal peptide-like region at the amino terminus (33). Membrane-bound forms of EnvZ have been studied in total or as purified inner membranes preparations (117, 136, 138, 151, 152), but purification of the intact protein and reconstitution of its catalytic activity has not yet been reported. Several truncated forms of EnvZ, lacking portions of the amino terminus, have been purified for biochemical analysis (1, 30, 54, 55). Recent studies show that the cytoplasmic domain (amino acid residues from Glu-106 to Gly-450) purifies as stable dimer (S. Forst, personal communication). The C-terminus of EnvZ undergoes phosphorylation and phosphate transfer to and from OmpR and is therefore considered to be the signaling domain (I, 30, 55). Overall, the topology of EnvZ is very similar to that of the bacterial chemoreceptor proteins Tar, Tsr, and Trg that have sensing and signaling functions at their periplasmic and cytoplasmic domains, respectively (28, 33). By analogy to these proteins, the osmosensing region of EnvZ might be located at the periplasmic region with the transmembrane region possibly assisting in the transduction of information from one side of the membrane to the other. However, the osmosensing role of the periplasmic region has not been clearly established.

236

ARFAAN A. RAMPERSAUD

C. Relationshipswith Other Proteins The N-terminal phosphorylation domain of OmpR (the first 125 amino acids) shows sequence similarities to other response regulators including the CheY protein that mediates chemotaxis (101, 108, 129-131). The three-dimensional structure of CheY (129 amino acids) has been elucidated (129) and reveals a doubly wound a/IS protein in which a core of 5 parallel ~-sheets are surrounded by 5 a-helices (129-131). The CheY phosphorylation site, Asp-57 (121), is in an acidic pocket formed at the top of the molecule where short loops connect the C-terminus of [}-sheets to the N-terminus of the following a-helices (129-131). Two additional aspartates, Aspl2 and 13, located at the C-terminal end of the ~51 strand, help form the acidic pocket and are involved in essential metal binding (129-131). A lysine residue at the C-terminal end of the ~5 strand Lys 109 may be close to Asp-57 and is thought to be displaced from its position by phosphorylation of Asp-S7. The movement may propagate or cause conformational transitions important to the function of CheY (108, 130). Virtually all response regulators, including OmpR, have aspartate and lysine residues that closely correspond to those mentioned in the CheY molecule (101, 108, 131). This is shown for OmpR in Figure 2 where the secondary structure of CheY (a pattern of alternating ~-sheets and o~-helices) is shown on top of the amino acid sequence of OmpR. In OmpR, Asp-S5 is the site of phosphorylation (16, 17, 26, 66) and lies at the C-terminal end of a putative ~3 strand (as in Asp-S7 of CheY). Three aspartate residues, Asp- 11, Asp- 12, and Asp- 13, are at the C-terminal end of the [31 strand and may help in forming an acidic pocket. Based on site-directed mutagenesis, Asp-11 and Asp-12 correspond to the relevant aspartate residues in CheY (Asp- 12 and Asp- 13) (5, 16, 17, 26, 66). Another highly conserved residue is Lys-105 (5, 101, 108, 130, 131) that, according to Figure 2, would be somewhere near the C-terminal end of the [55 strand. As will be discussed below, OmpR probably also undergoes conformational changes at its amino terminus upon phosphorylation, and this change might involve the Lys-105 residue. While the structure of CheY has provided insight into how the N-terminus of OmpR may look, the rest of OmpR is completely unknown. This makes it difficult to predict how phosphorylation and conformational transitions would influence other properties of OmpR such as, DNA-binding, OmpR subunit interactions, and RNA polymerase interactions. The DNA-binding region of OmpR does not have structural motifs characteristic of many transcription factors (helix-turn-helix or zinc fingers motifs) but the domain is related to a subset of response regulators having somewhat related DNA-binding regions (5, 108). These proteins include the PhoB and VirG proteins (for phosphate regulation and plant virulence, respectively) (5, 108). Based on the related sequence of their target promoters, they are proposed to share a common mechanism of transcriptional activation (8).

TransmembraneSignaling in Osmoregulation

237

EnvZ and other sensor/kinase proteins are related to each other through their signaling domains (5, 101, 108, 131). Segmented sequence similarities are found within a 200 amino acid stretch at the C-terminus and are separated by regions of poor homology (5, 101, 108, 131). These are shown in Figure 3 as Regions I, II, and III. Region I is short and contains a histidine residue that is totally conserved among all members of the kinase family. This is believed to be a site of phosphorylation, and in EnvZ the histidine residue occurs at position 243 (5, I01, 108, 131). It is interesting to note that many missense mutations in EnvZ occur near or within Region I (see Figure 3 and Table 3). The two remaining segments of sequence similarity are found near the C-terminal portion of EnvZ (Figure 3). In many sensor/kinase molecules Region II is located approximately 100 residues away from Region i and is characterized by at least one asparagine residue (101, 108, 131). In EnvZ the conserved residue is Asn-347, and in a chimeric signal transducer, Tar-EnvZ, it occurs at position 381 (see below). This amino acid has been extensively altered in Tar-EnvZ (see Table 3) where it has been shown to be essential for kinase activity but not for phosphatase activity (152). The Region IIl segment is a long glycine-rich stretch of residues characterized by the sequence motifs DXGXG and GXG (101, 108, 131). They may represent a potential nucleotide binding motif (101, 108, 131). These sequences have also been studied in the Tar-EnvZ molecule where they have been designated as G 1 and G2, respectively (Figure 3). They are also essential for kinase activity, but have different effects on phosphatase activity (152).

V. PHOSPHORYLATION STUDIES OF OMPR AND ENVZ Phosphorylation of OmpR is essential for porin gene expression. It occurs in both an EnvZ-dependent as well as EnvZ-independent fashion (crosstalk). EnvZ-dependent phosphorylation is the primary pathway for generating OmpR-P and is comprised of both kinase as well as phosphatase reactions. The ratio of these two activities is crucial for porin gene expression. Details of this process as well as crosstalk mechanisms are discussed below.

A. PhosphateTransfer Reactions EnvZ incorporates the terminal phosphate from ATP to form a stable phosphoprotein (1, 30, 55). This has been shown for several purified molecules lacking portions of their amino terminus as well as intact, membrane-bound forms of the enzyme (1, 30, 54, 55, 117, 138, 151, 152). Autophosphorylation of EnvZ has been reported for the truncated molecules (1, 53), while membrane-bound EnvZ and the related protein Tar-EnvZ undergo transphosphorylation reactions (151). The importance of auto- versus transphosphorylation reactions for in vivo function of EnvZ remains to be clarified.

238

ARFAAN A. RAMPERSAUD

Site-directed mutagenesis indicates EnvZ is phosphorylated at the conserved histidine residue, His-243, in Region I (30, 66). Phospho-EnvZ is reasonably stable with the phosphorylated amino acid residue having a pH-stability profile typical of a phosphoramidate linkage (30). More recently, the phospho-histidine residue has been directly identified by chemical methods (J. Delgado and M.I. Inouye, unpublished results, S. Forst, personal communication). Incubating phospho-EnvZ with purified OmpR results in the transfer of phosphate from EnvZ to OmpR (1, 26, 30, 54). The overall reaction is fast, exhibits a metal-dependence, and is affected by monovalent as well as divalent ions (I, 30, 54, 117, 138). The site of OmpR phosphorylation (by EnvZ) is Asp-55 (26). This has been shown by pH-stability studies, which characterized the linkage as an acyl phosphate bond (30), and site directed mutagenesis of the Asp-55 residue (see below). Newer studies employing HPLC of tryptic peptides fragments derived from [3H]borohydride-reduced phospho-OmpR and subsequent peptide sequencing demonstrates that the majority of phosphate is located on Asp-55 (26). Phospho-OmpR is labile but has a longer half life than most other phosphorylated response regulators (1, 54, 101). When incubated with either ATP or EnvZ, there is little loss of phosphate from OmpR-P (1, 54). However, when both EnvZ and ATP are present, OmpR-P is rapidly dephosphorylated with the subsequent release of inorganic phosphate (I, 54). ATP can be substituted with adenosine nucleotides such as ATP-7-S, AMP-PNP, and AMP-PCP without inhibiting the dephosphorylation reaction (I, 54). Thus, hydrolysis of nucleotide is not a mandatory step during the reaction. It has been proposed that the nucleotide component may serve as an allosteric effector for dephosphorylation (1, 54). While it is not yet clear whether dephosphorylation activity resides solely in EnvZ or occurs through some sort of EnvZ-OmpR interaction, it seems unlikely that dephosphorylation of OmpR is the reverse of the phosphorylation step. The reactions are mechanistically different since phosphorylation requires a phosphorylated EnvZ intermediate, while dephosphorylation liberates inorganic phosphate without the formation of a (stable) EnvZ-intermediate (1, 54, 65). Several EnvZ mutants have been isolated that have lost one but not both activities, implying that different portions of the molecule (different catalytic sites) are involved in the two reactions (2, 118, 46, 136). Some amino acids may participate in both reactions since it is possible to obtain EnvZ mutants that have lost both activities, such as EnvZ(H243R) and EnvZ(T35 IS) (65, 66). Thus, the regions for kinase and phosphatase activities may overlap to a certain degree.

B. PhosphorylationMutants of EnvZ and OmpR The EnvZ(H243V) and EnvZ(H243R) have amino acid replacements at His-243 and are not phosphorylated by ATP (30, 66). Subsequently they do not phosphorylate OmpR (30, 66). In the case of the His-to-Arg replacement, there is also a loss of OmpR-dephosphorylation activity (66). The outer membrane porin phenotype

TransmembraneSignaling in Osmoregulation

239

produced by both mutants closely resembles that of an envZ null mutant in that it produces low levels of OmpF at low osmolarity and a small amount of OmpC at high osmolarity (66; A. Rampersaud, unpublished observations). The results can be interpreted in a fashion similar to an envZ null mutant (see below). That is, while EnvZ is important for producing OmpR-P, it is not the only means by which OmpR-P is generated. The three aspartate residues that form the putative phosphorylation domain of OmpR, Asp- 11, Asp- 12, and Asp-55 have been studied by site-directed mutagenesis (16, 17, 66). Mutants having an asparagine substitution at Asp-I 1 (D 1IN) or glutamine substitution at Asp-55 (D55Q, the phosphate acceptor site) do not produce significant amounts of OmpF or OmpC in the outer membrane and cannot be phosphorylated by EnvZ in vitro (16, 66). They do not bind OmpR-target sites in the ompF promoter indicating that interactions with the ompF promoter requires OmpR-P (16). A different situation applies to ompC since it appears that both mutants bind to the ompC promoter (16). In this case OmpR-P may be needed for transcription of the ompC gene. Valine or glutamine replacements at Asp- 12 (D 12V and D12Q, respectively) are somewhat variable in their effects on porin gene expression, with the D 12V mutant shown to be phosphorylated (albeit poorly) by EnvZ (16, 66). The three aspartates have also been changed to alanine residues (26) as its smaller side chain avoids distortions of the acidic pocket that would otherwise be introduced by asparagine or glutamine residues. In vitro results with these mutants support the view that EnvZ phosphorylates OmpR at Asp-55, but in vivo results are somewhat different. In particular, an OmpR(D55A) mutant was found to produce both porins while a double mutant, OmpR(D55A, D 11A), produced none (26). Low but significant levels of phospho-OmpR(D55A) were also found in an in vivo phosphorylation experiment (26), and this led to the conclusion that Asp-11 is another site for OmpR phosphorylation. Presumably, this site is phosphorylated through an EnvZ-independent mechanism (crosstalk) (26).

C. Formsof OmpR-P Biochemical, genetic, and physiochemical studies demonstrate that EnvZ phosphorylates OmpR for ompF and ompC expression and that OmpR-P mediates ompF repression (16, 17, 26, 66, 126). Thus, porin genes are regulated through one or more phosphorylated species of OmpR rather than through an on/off switching mechanism between non-phosphorylated and phosphorylated forms (118). Two different models can explain the overall process: A quantitative one, involving changes in the total amount of a single phosphorylated form, or a qualitative model involving multiple phosphorylations on individual OmpR molecules (118, 126). In vivo phosphorylation studies have shown a significant increase in OmpR-P levels during an osmotic stress, and this is consistent with a quantitative model (32). However, multiple phosphorylations on OmpR could also produce the same results.

240

ARFAAN A. RAMPERSAUD

Mathematical modeling has led to the conclusion that a single species of OmpR-P can account for porin gene expression, provided that OmpR-P differentially binds to DNA target sites in ompFandompCpromoters (126). That is, there must be highaffinity activator and low-affinity repressor sites in ompF to rationalize how low levels of OmpR-P activate ompF while high levels of OmpR-P repress it. Intermediate affinity or similar low-affinity sites are needed in ompC to account for ompC expression at intermediate levels of OmpR-P (126). So far, DNA-binding studies, at least with the ompFpromoter, are consistent with this proposal (A. Rampersaud, S. Harlocker and M.I. Inouye, manuscript in preparation). This does not eliminate the possibility that qualitative changes in OmpR occur, and in fact, multiple phosphorylations of OmpR may explain pleiotropic effects (see below). The particular amounts of OmpR-P needed for a certain porin phenotype are unknown. According to the diagram shown in Figure 1B, the total amount of OmpR in the cell remains constant while the ratio of OmpR-P to OmpR varies. Upper limits of OmpR-P levels can be set that separate one phenotype from another. Between these boundaries, levels of OmpR-P can vary without significantly altering phenotype. Whether such threshold levels actually exist remains to be determined, but the graph provides a useful picture of the dynamic fluctuations in OmpR-P levels.

D. Importance of a Phosphate Relay Cycle for Porin Gene Expression The relative phosphorylated state of OmpR is determined by EnvZ through a combination of its kinase and phosphatase activities (89, 118, 126). These opposing activities form a phosphate relay cycle in which both activities seem to operate simultaneously. Phosphatase activity prevents the build-up of excessive OmpR-P levels, while kinase activity provides a continuous supply of OmpR-P. Low levels of OmpR-P are produced when these activities differ slightly. Larger differences in the ratio of kinase to phosphatase activities cause more substantial changes in OmpR-P levels (89, 118, 126). Several EnvZ mutants have already been mentioned that are disrupted for a portion of their phosphate relay cycle. The EnvZ11 mutant does not have phosphatase but retains kinase activity (2), while the gene products of the envZ250 and envZ247 mutants, EnvZ(Pl59S) and EnvZ(A239T), respectively, are both defective in their kinase but not phosphatase activity (126). These mutants represent the two extremes of the cycle where the kinase/phosphatase ratio is either very high (EnvZ l 1) or very low [EnvZ(P159S) and EnvZ(A239T)]. The corresponding porin phenotypes are F-C+ for EnvZl I and F C - for the other two. The low osmolarity F'C- porin pattern is absent as an extreme phenotype indicating that it cannot be made by one activity alone. Thus, the phosphate relay cycle may be particularly important for sustaining low enough levels of OmpR-P so as to activate ompF expression without activating ompC.

Transmembrane Signaling in Osmoregulation

241

In addition to EnvZl 1, OmpR36 is also disrupted in the dephosphorylation part of the cycle (see the earlier section on ompB mutants) and suppressors for both mutants, ompR77 for envZll (2) and envZ30 for ompR36 (65), have been isolated. The suppressors compensate for lost dephosphorylation activity by reducing the accumulation of OmpR-P and thereby restore OmpF expression as well as a certain degree of osmoregulation (2, 65). OmpR77 has been shown to be a poor substrate for EnvZ11 while EnvZ30 is completely nonfunctional, leaving OmpR36 protein to be phosphorylated through crosstalk mechanisms (2, 65).

E. Alternative Routes of OmpR Phosphorylation EnvZ-minus strains such as envZ22 and AenvZ make low amounts of OmpF and OmpC in the outer membrane and elevate OmpC levels further at high osmolarity conditions (29, 36, 88, 146). In an in vivo phosphorylation study, low but significant amounts of OmpR-P were found in the envZ22 strain with further increases in OmpR-P levels occurring at high osmolarity conditions (32). These results are explained on the basis of an EnvZ-independent phosphorylation mechanism that produces enough OmpR-P for porin gene expression. While the responsible components of this pathway have not been identified, related crosstalk or cross-regulatory phosphorylation pathways have been reported for other response regulators such as CheY and PhoB (102, 148a). One way OmpR can be phosphorylated in the absence of EnvZ is through related sensor proteins. In vitro studies have shown that OmpR can be phosphorylated by the CheA protein, which is the kinase for the CheY molecule (54). Furthermore, a gene designated barA, has been isolated, which in multiple copies complements an envZ deletion strain (93). The BarA protein has sequence similarities to conserved regions in both EnvZ as well as OmpR and is suspected to use its kinase activity to phosphorylate OmpR (93). While it seems clear that OmpR can be phosphorylated by related sensors, the reactions require high levels of either OmpR or the kinase (54, 93). This raises the question of whether the cellular levels of these kinases are high enough to be physiologically relevant for OmpR. An alternative and more attractive means of OmpR phosphorylation may be through the direct abstraction of phosphate from a high energy phosphate donor. The related proteins CheY and CheB can catalyze their own phosphorylation when provided with high energy substrates such as phosphoramidate, acetyl phosphate, or carbamoyl phosphate (76). In this same study OmpR was mentioned as being capable of similar phosphorylations. Regardless of how it happens, studies of the OmpR(D55A) mutant indicate that phosphorylation through crosstalk pathways takes place on Asp-I 1 (26). Under normal circumstances this phosphate group may be removed by the phosphatase activity of EnvZ. Evidence for this is indirect and is based on the envZ250 and envZ47 alleles, which code for mutants having phosphatase but not kinase activity.

242

ARFAAN A. RAMPERSAUD

Crosstalk does not occur in these strains presumably because the mutants are removing phosphate at Asp-11. It is intriguing to ask whether crosstalk contributes to the pleiotropic effects observed in EnvZ phosphatase mutants (e.g., EnvZ11 or EnvZ473). Such mutants may be unable to dephosphorylate any aspartyl-phosphate residue on OmpR. Hence, OmpR could acquire two phosphate groups: one at Asp-55 from the EnvZ mutant and the other at Asp- 11 through crosstalk. The double phosphorylation event would then convert OmpR into a pleiotropic regulator. VI.

ENVZ S I G N A L I N G

While the role of EnvZ as a kinase/phosphatase for OmpR is well established, evidence that would associate these activities with environmental sensing remains elusive. This link is essential to the model of osmoregulation since environmental conditions are proposed to be responsible for fluctuations in OmpR-P levels. In this section, information concerning the role of EnvZ as a sensory molecule is summarized.

A. Osmosensingby EnvZ A portion of the amino terminus of EnvZ projects into the periplasmic space and may receive an environmental signal. While the nature of the signal is unknown, one model suggests that environmental stimuli cause conformational changes at the periplasmic domain, which then modulates the cytoplasmic signaling domain (137). According to this model, the two transmembrane segments TM1 and TM2 would be important for transmitting information from one side of the membrane to the other (137). EnvZ has been compared to hypothetical osmosensor proteins such as turgor pressure sensors, wall stretch receptors, and stretch receptors in order to more clearly define its function (24). It was pointed out that model osmosensors detect transient osmotic signals, such as ion fluxes, pressure differentials, cell wall or membrane stretching, for short term activation events, while porin gene regulation continues well after cells have adapted to an osmotic stress (24, 64). It therefore seems unlikely that transient physiochemical events activate EnvZ for signaling (24). Another way that EnvZ might work would be by indirectly detecting osmotic conditions, perhaps by responding to a long lived signal that is itself sensitive to osmolarity. In this regard the levels ofperiplasmic membrane~erived oligosaccharides (MDO) change dramatically during an osmotic stress, and these have been considered as potential signals for EnvZ (27a). However, recent studies have challenged this idea by showing that OmpF and OmpC production as well as osmoregulation are not significantly affected in strains having an mdoA mutation (which encodes one of the MDO biosynthetic enzymes) (38). Thus, it is unlikely that MDOs are signals for EnvZ.

TransmembraneSignaling in Osmoregulation

243

B. The Amino Terminus as the Sensor Domain

While osmosensing activity of EnvZ remains unclear, the general importance of the amino terminus for proper osmoregulation has been established through several indirect studies. For example, truncated EnvZ fragments missing substantial portions of the first transmembrane segment (TMI) (53) or regions further into the N-terminus (88) produce a constitutive F+C+ phenotype. While these C-terminal domains can still transmit information to OmpR, their unregulated porin phenotype indicates that they have lost their ability to respond to osmotic changes. This suggests that a C-terminal fragment of EnvZ by itself cannot account for all of the functions of the intact protein and implies that the amino terminus of EnvZ contains a region necessary for osmoregulation. EnvZ mutants having amino acid replacements in TM 1 and at positions flanking the second transmembrane segment (TM2) are shown in Table 3 and provide insight into the role of the transmembrane regions for signaling. Amino acid replacements in TM1 include LI8F, L35Q, P41S, and P41L while two mutations, P159 and R180C, immediately flank TM2 (83, 118, 136). The EnvZproteins that carry these missense mutations exhibit a variety of altered osmoregulatory porin phenotypes with several of them showing significant changes in their phosphorylation or dephosphorylation properties towards OmpR (136). These experiments demonstrate that mutations in or near the transmembrane region of EnvZ affect the cytoplasmic signaling domain. Point mutations in the periplasmic region proper that affect EnvZ signaling have not been reported, and this has made it difficult to assign a functional role to this domain. However, EnvZ molecules deleted for portions of their periplasmic domain have been created and analyzed (see Table 3). They produce an OmpC constitutive phenotype and in biochemical tests have greatly reduced dephosphorylation activity (137). Furthermore, they become pleiotropic and decrease the production of Male and PhoA proteins (137). While this is consistent with the view that alterations at the N-terminus affect signaling at the C-terminus, the deletions are somewhat difficult to clearly interpret. They could alter the conformation of the periplasmic region, or alternatively, affect the orientation of transmembrane segments. The latter possibility could be rationalized by considering that a shortened periplasmic region might return to the membrane sooner than the full periplasmic region. This may disrupt an alignment of transmembrane segments, and in turn would disrupt transmembrane signaling.

C. The Hybrid Tar-EnvZ Sensor A significant amount of knowledge has come from the analysis of a chimeric molecule in which the C-terminus of EnvZ was fused to the amino terminus of the chemoreceptor protein Tar (143). Tar is one of the chemoreceptors mentioned earlier and binds the ligand aspartate at its periplasmic domain to activate its

244

ARFAAN A. RAMPERSAUD

cytoplasmic signaling domain (130). In the chimeric Tar-EnvZ receptor, binding of an aspartate signal by the Tar portion of the molecule transduces a signal across the membrane to affect the cytoplasmic EnvZ domain. In turn, this domain phosphorylates OmpR for ompC expression (143). These studies indicate that EnvZ and chemoreceptors share similar features of transmembrane signaling and, as compared with truncated EnvZ molecules, demonstrate how the cytoplasmic signaling domain of EnvZ can be brought under direct control by a receptor domain (143). Novel intermolecular complementation studies have been shown with the TarEnvZ receptor (151, 152). Truncated Tar-EnvZ mutants lacking portions of the C-terminal signaling domain, and fuU-length mutants having point mutations at critical residues in EnvZ, are defective in OmpR phosphorylation or dephosphorylation (151, 152) (Table 3). Yet, when particular combinations of mutants are coexpressed in the same cell, certain aspects of signaling are restored. The complementation approach has provided evidence for transphosphorylation of EnvZ, presumably through the association of EnvZ as multimers (152). Other experiments indicate that ligand-dependent (aspartate) regulation of the signaling domain requires both kinase as well as phosphatase activities. It was proposed that in the absence of aspartate, the kinase/phosphatase ratio of Tar-EnvZ favors phosphatase activity (152). Aspartate binding is thought to decrease phosphatase without affecting kinase activity (152).

D. Activation of EnvZ by Local Anesthetics Another aspect of EnvZ signaling relates to how the molecule responds to membrane perturbants. At low concentrations, agents such as procaine or phenethyl alcohol act at the transcriptional level to decrease OmpF and increase OmpC (36, 41, 107, 111, 135). This is similar to high osmolarity conditions, but procaine and phenethyl alcohol also reduce the levels of PhoA and Male proteins (19, 107, 146). Thus, their phenotypes are more closely related to that produced by the pleiotropic envZll and envZ473 mutants (19). Two envZ missense mutants have been identified, env7_3 and envZ6, that are resistant to the effects of procaine for changes in porin patterns (135). Additionally, envZ null strains (envZ22 and AenvZ) do not respond to procaine for either changes in porin profiles or pleiotropic effects on other genes (115, 146). These results demonstrate that EnvZ directly or indirectly mediates the effects of procaine. Procaine prevents the in vitro phosphatase activity of membrane-bound EnvZ but is reported not to influence dephosphorylation activity of a C-terminal EnvZ molecule (138). The anesthetic also has no effect on the signaling properties of the Tar-EnvZ protein that lacks the amino terminus of EnvZ (115). These experiments suggest that procaine acts somewhere at the amino terminus. As procaine affects the fluidity of the bacterial membrane (41a, 62) it is reasonable to suspect that its mechanism of action involves some sort of disruption of the transmembrane segments. Interestingly, lowering of lipid fluidity has been shown

Transmembrane Signaling in Osmoregulation

245

to affect porin production in the outer membrane (27, 41). These data indicate that the physiochemical properties of the inner membrane could have a profound influence on the signaling activity of EnvZ.

VII. MULTIFUNCTIONAL PROPERTIESOF OMPR Mutations in OmpR occur throughout the molecule with a number of them localized to the N-terminus (Figure 2). Relative to the structure shown in Figure 2, N-terminal amino acid replacements are frequently found between the C-terminal portion of 13-sheets and the N-terminal portion of a-helices. They are spatially within the vicinity of the phosphorylation domain, with many of them affecting the production of OmpR-P. Obvious phosphorylation mutants are the various Asp-55 and Asp-I 1 mutants (16, 17, 26, 66). The L 16Q (ompR77) and R 15C (ompR36) replacements also affect phosphorylation by altering EnvZ-OmpR interactions (2, 83, 96, 97). Mutants tentatively proposed to affect interactions between OmpR subunits or interactions with RNA-polymerase are also located within the N-terminus (18, 57, 95, 120). Other amino acid replacements are thought to change OmpR conformation, with three occurring away from the phosphorylation domain between the N-terminal portion of 13-sheets and the C-terminal portion of co-helices (16, 17, 67). Within the C-terminal region, several mutations appear to specifically affect DNA-bincling (96, 120), but it is not known whether they do this through conformational defects or through defective interactions between amino acid side chains and base pairs.

A. Conformational Changes in OmpR Phosphorylation of OmpR may drive conformational changes in the molecule necessary for porin gene expression. Alternatively, the phosphate group could participate in an essential interaction between OmpR subunits, DNA target sites, or RNA-polymerase interactions. A phosphorylation mutant, OmpR(D55Q), has been overproduced from a high copy number vector and shown to osmoregulate porin genes, due to its latent DNA-binding activity (67). This result along with the mutants discussed below make it unlikely that the phosphate group is directly involved with protein-protein or protein-DNA contacts. Three second-site suppressors of OmpR(D55Q) (expressed from pBR322 or low copy number plasmids) have been isolated that allow porin gene expression despite the D55Q replacement (17, 67). They contain either a T83A, G94S, or TI02C replacement (! 7, 67). While they do not restore phosphorylation activity, they allow DNA-binding to both ompF and ompC promoters (17, 67). The suppressors are independent of the original D55Q mutation and regulate porin production even in the absence of the envZ gene (17, 67). The most likely explanation for them is that they permit OmpR to adopt a structure or Conformation that mimics the phospho-

246

ARFAAN A. RAMPERSAUD

rylated state of the protein (17). This would be consistent with the view that phosphorylation drives some sort of conformational change in OmpR structure that is important for gene expression. The three mutations cluster in the middle of the OmpR molecule with T102C lying close to Lys- 105. As mentioned earlier, Lys- 105 is a highly conserved residue thought to be involved with conformational transitions in the CheY protein (108, 130). It is possible that the T I02C substitution may influence this lysine residue. Another amino acid replacement PI06L is adjacent to Lys-105, but is still subject to regulation by envZ (120). An intriguing aspect of D55Q mutants containing T83A, G94S, or T102C suppressors is their osmoregulation of OmpF and OmpC porins (17, 67). EnvZindependent means of osmoregulation might explain these results, and these would be different than the previously mentioned crosStalk mechanisms. Examples of these would include osmotically regulated changes in DNA topology and/or the participation of general transcription factors, such as integration host factor or the histone-like protein H-NS (12, 17, 40, 67). A somewhat different conformational mutant of OmpR is a $48F replacement (16). In combination with a DI 1N mutation, $48F partially restores porin expression and osmoregulation that was lost by D 11N (16). It has been proposed that the conformation of OmpR is distorted by D 1IN, resulting in poor interaction and/or phosphate transfer with EnvZ (16). The $48F mutation allows the protein to serve as a (weak) substrate for EnvZ and thus restores some of its structure. As the $48F replacement by itself produces an F-C-phenotype, the two mutations essentially complement each other.

B. Multimerization of OmpR Multimers of OmpR were originally proposed as part of a regulatory mechanism for ompB (44, 45). This was supported by negative complementation results between an ompR4 strain and an ompRl01 allele (10, 44, 75) (although another explanation for OmpR 101 was mentioned in an earlier section of this chapter). However, when OmpR (the non-phosphorylated form) was purified in later studies, evidence for a multimeric state of OmpR was not found (63). Size exclusion chromatography of the purified protein showed it to be a monomer, which did not form multimers even when treated with the crosslinking agent, dimethyl subermidate (63). Recently, experiments have shown that if OmpR is phosphorylated by EnvZ and ATP, oligomers corresponding to OmpR dimers and trimers are readily detected by dimethyl suberimidate crosslinking (95). Two OmpR mutants, OmpR(E96A) and OmpR(R115S), that do not form phosphorylation induced multimers in crosslinking experiments have been isolated and show a dramatic reduction in their phosphorylation enhanced DNA-binding activity (95). As neither mutant produces any porin protein in the outer membrane, OmpR multimerization may be important for gene expression (9.$). It is possible

Transmembrane Signaling in Osmoregulation

247

that phosphorylation brings about changes in OmpR, followed by oligomerization of subunits. This may be important for DNA-binding and subsequent regulation of porin gene expression. Two other mutants, OmpR(G94D) and OmpR(E 111K), have amino acid replacements close to Glu-96 and Arg-115, respectively, and are thought to affect OmpR multimerization (18). In a wild-type ompR background, they prevent the expression of both ompF and ompC genes (18). OmpRX6 produces a similar dominant negative phenotype (57) and may represent yet another such mutant.

C. Additional OmpR Mutants Analysis of a collection of mutants broadly classified as ompR2 (low OmpC and normal OmpF levels) indicate several functions of OmpR can be disrupted independently of one another (120). One set, exemplified by OmpR(V203M) and OmpR(R220C) (products of the ompR472 and ompR324 alleles, respectively), are apparent DNA-binding mutants that prevent both ompF repression and ompC expression (120). Others interfere with repression of ompF but not with the activation of ompC. These include OmpR(GI29D) and OmpR(P131S) and are thought to bind to repressor sites in the ompF promoter but do not proceed further in the repression mechanism (120). Interestingly, an Ala-to-Thr replacement at position 130 (between Gly- 129 and Pro- 131) is one of the two differences between the ompR genes of E. coli B and E. coli K12 (91). A repression-type mutation in OmpR of E. coil B could explain why these cells make OmpF constitutively. A final set of mutants bind normally to both promoters but reduce expression of both genes proportionately (120). The amino acid replacements cluster in two regions of the molecule, around positions 90 and 190. Representative examples of this class include the OmpR(E87K), OmpR(G 19 IS), and OmpR(E193K). They are thought to specifically affect gene activation, possibly by influencing OmpROmpR or OmpR-RNA polymerase interactions (120).

VIII. DNA-BINDING AND TRANSCRIPTIONAL PROPERTIES OF OMPR OmpR binds to the upstream region of both ompFand ompCgenes to regulate their expression. Most, if not all, of the major OmpR binding sites in ompF and ompC have been identified. In this section the involvement of these sites for both activation and repression are discussed, as is the role of phosphorylated OmpR for these processes.

248

ARFAAN A. RAMPERSAUD A.

Binding

to the

ompFP r o m o t e r

Figure 4 shows several features of the ompF regulatory region that includes at least three OmpR binding sites, two binding sites for IHF, a-35 and- 10 region, and a transcriptional start site 100-bp upstream to the initiation codon. The binding sites

9lot)

-80

o

-00

O

.4O

O

o

TTTTA(,-I-rrrGGTTACATA']TrrI't*C-rrrrk'GAAACCAAA~ATCTt'rGTAGCACTTTC pl

-380 o

- mo

T

TGTTACGG~ATATTACATTGCAA

- m -10 +I

[

,

I> Mer (Omp~

- Im licit

.100

-m

O

.00

O

.40

O

O

CATTTACATTTTG~CATCTATAO CGATAAATGAAACATCTTAAAA~AG'rATCATAT a

i s

~.~. . . . _ _

"-

I

. . . . I ..... I

188~-178

'

'

' o i

. . . . . .

T

'~-

-

cm

.......

".

-? -I0 +Im-i

-

.

.

.

.

.

(omit)

FiBure4. Organization of OmpR-binding sites in the upstream regulatory region of ompF and ompC. The region upstream to both ompF(A) and ompC (B) are dia-

grammed. The diagram is not to scale. OmpR binding sites in both promoters are shown as filled boxes, (me)with the DNA sequences shown above. The half arrowhead ( ) indicates potential OmpR-binding sites as 20-bp motifs, (FI, FII, Fill, CI, CII, and CIII) while the boxes below show OmpR binding sites as 10-bp motifs, (F- and C-boxes). The unfilled boxes (D) represent high affinity OmpR target sites while the stippled boxes (n) represent target sites of lower affinity. IHF binding sites (IHF), the -35, -10 regions, transcriptional start site (), and the first residue of both proteins (Met) are indicated.

TransmembraneSignaling in Osmoregulation

249

for OmpR and IHF are positioned upstream to the transcriptional start site (+1) between -380 and -40. Deletion analysis of plasmid-borne ompF-lacZ fusions shows that DNA sequences essential for ompF activation are located around -100 (relative to the transcriptional start site) (59). Downstream deletions past-100 severely depress ompF expression (68, 116). Specialized transducing phage carrying ompF-lacZ operon fusions have demonstrated that osmoregulation (i.e., repression) requires a further upstream region, between -1400 and -240 (104). Based on DNase I footprinting experiments, OmpR recognition sites are between -100 and -40 and between -360 and -380. The OmpR binding site between -100 and -70 is the activator site in ompF (68, 116) and is large enough to accommodate at least two OmpR molecules (87, 103). This 30-bp element can be divided into three juxtaposed lO-bp sequences, termed F-boxes, having the consensus sequences TITAC(AfF)TIT (103, 142). Three larger 20-bp elements have also been found between - 100 and -40 and define larger OmpR recognition sequences (shown in Figure 4 as FI, FH, and FHI) (81). While it is not clear which of these DNA motifs represent actual OmpR-binding sites, an in vivo chemical footprinting study noted that a 10-bp sequence between -51 to -42 was more like OmpR-binding sequences in the ompC promoter than F-boxes (142). It was designated Cd-box to call attention to this relationship (142). The F- and C-boxes differ in terms of their DNA-binding affinities (31). Mobility shift experiments using short sequences representing the -100 to -70 or-52 to -40 regions demonstrate that the -100 to -70 region (F-boxes) is readily recognized by OmpR in crude cellular extracts (31). In contrast, the region between -51 to -42 (Cd-box) is poorly bound unless the F-boxes are also present (31). These results demonstrate that OmpR binds with greater affinity to the -100 to -70 region than the -51 to -42 sequence. This can be explained on the basis of different nucleotide sequences in the F- versus C-boxes, the placement of multiple target sites close to each other (F-boxes), or a combination of these features. In vitro and in vivo footprinting experiments have shown that the OmpR472 protein binds the F-box region but not the Cd-box sequence (and also fails to recognize similar C-boxes in the ompC promoter) (87, 142). Recent studies have also shown that OmpR472 binds poorly to the upstream repressor site between -380 to -360 (A. Rampersaud, S. Harlocker & M.I. Inouye, manuscript in preparation). The inability of this mutant to bind to these sequences can be directly correlated with the F+C- phenotype of an ompR472 strain. A segment of the ompF promoter migrates anomalously in polyacrylamide gels and led to the discovery of an intrinsic curvature or bend in the ompF regulatory region (85). The major determinants for the DNA bending are four short runs of peri•193193 dA" dT tracts, designated T 1 through T4. The tracts overlap the OmpR recognition site with T1 and T2 tracts occurring between -111 and-92, while T3 and T4 tracts are between -83 and -70 (68, 85). Nucleotide replacements in one tract alone do not significantly affect OmpR binding but do change ompF expres-

250

ARFAAN A. RAMPERSAUD

sion, as well as the characteristic mobility of the promoter fragment (68). A severe reduction in OmpR-DNA interactions is observed when two tracts are simultaneously mutated (68). Deletion of a T residue within the T3 or T4 tract also affects ompF expression, as well as the migration of ompF promoter fragments (127). These results suggest that curvature or structure of the ompF promoter is important for proper ompF regulation. Most of the early DNA binding studies with purified OmpR probably used the dephosphorylated form of the protein, with the results representing specific but latent DNA-binding activity (63, 103, 116).Later experiments employing OmpR-P have shown that, relative to the non-phosphorylated species, about 10-fold less OmpR-P is needed for optimal DNA-binding at both ompF and ompC promoters (4). Thus, it can be concluded that phosphorylation of OmpR increases its affinity for DNA target sites. Recently completed studies have further clarified the relationship between OmpR-P levels and DNA-binding at the ompF promoter (A. Rampersaud, S. Harlocker and M.I. Inouye, manuscript in preparation). The results indicate that when OmpR-P levels are low, DNA-binding occurs preferentially at the high at~nity site between -100 and -70 (F-boxes). As OmpR-P levels increase, there is increased occupancy of the -40 to -50 region (Cd-box), which is closely followed by binding at the -360 to -380 repressor site. The overall process appears cooperative; OmpR interactions at the high affinity site promote (or stabilize) binding interactions at lower affinity sites. These results are consistent with the current model of osmoregulation and fulfill the expectations of a quantitative model for OmpR-P. At low osmolarity, low levels of OmpR-phosphate stimulate ompF expression by binding to the high-affinity activator site. At high osmolarity, the buildup of OmpR-P allows binding at low-affinity sites and eventually leads to ompF repression. The -360 to-380 repressor site is distantly positioned from the -100 to -40 regulatory sites (104, 127). In order for it to act as a repressor site, it needs to be brought close to the other sites. A DNA looping model has been proposed as one way to achieve this (127). Interestingly, genetic studies suggest that the actual block in transcription by such a repression loop may not be due to steric effects (e.g., occlusion ofRNA polymerase from its binding site) (35, 84, 125). Rather, a direct interaction may occur between OmpR and RNA polymerase (125). Details of the repressor loop are not yet clear, but several biochemical results are entirely consistent with such a repression mechanism. Cooperative binding and OmpR-OmpR interactions (OmpR multimerization) would be important for stabilizing OrnpR bound at low-affinity sites, and both have already been mentioned. The inherent bend in the ompF promoter also supports a repression loop model as the DNA needs to bend in order for target sites to come close to each other. Indeed, putative bending mutations have been found in the - 100 to -70 region that prevent repression of ompF (127).

TransmembraneSignaling in Osmoregulation

B. Binding to the

251

ompCPromoter

The ompC promoter has a somewhat different organization than ompF (Figure 4B). Upstream deletions of plasmid-based ompC-lacZ fusions initially identified sequences starting from -100 as essential for OmpR-dependent regulation (90). Soclium-bisulfite-generated point mutations of similar fusions identified sets of base pairs between - 100 and -40 that reduced LacZ production (90). These mapped into three 10-bp elements designated sequences a, b, and c and were shown later to be OmpR-binding sites (63, 78, 80, 81, 90, 103, 142). Sequences a and b are identical to each other, and all are separated by multiples of 10 or 11 bp (90, 142). Thus, they occur on the same side of the DNA helix (90). In other studies, sequences a, b, and c have also been designated Ca-, Cb-, and Cc-boxes, respectively (Figure 4B) (142). A consensus sequence can be defined for all C-boxes, TG(A/I')NCATNT, and this is different from that of the F-boxes (142). As in ompF, they are within three large 20-bp regions designated CI, CII, and CIII (Figure 413) (81). For the remainder of this section the C-box notation will be used to refer to OmpR-binding sites in the ompC promoter. In vitro and in vivo studies clearly demonstrate that OmpR binds to the Ca, Cb, and Cc boxes (63, 78, 80, 81, 90, 103, 142). Another OmpR binding site has also been identified and termed the Fd-box (103, 142). This is a 10-bp region that overlaps the Ca box and shows similarities to the F-boxes in the ompF promoter (142). In vivo dimethyl-sulfate-protection studies reveal that OmpR protects several guanine residues within the Fd- and C-box regions (142). Not surprisingly, the protected bases are in register with each other by multiples of 10- or ll-bp, indicating that DNA-binding occurs on the same side of the DNA helix (142). OmpR binding to the ompC promoter appears to be cooperative, with the highest affinity site being Ca (80). Protection of the Ca region occurs at a lower OmpR concentration than the downstream sequence Cb or Cc regions (80). Although the Ca and Cb boxes are identical in sequence, the Fd box overlaps the Ca box and probably contributes to the enhanced OmpR-binding (142). Removal of the Ca sequence also causes significant loss in ompC-lacZ expression and reduces binding to the remaining regions (80). More recently, On~R has been shown to directly bind DNA sequences containing just the Ca as well as flanking sequences, but does not bind the other two elements (81). This region also drives ompC.lacZ expression in the absence of the Cb and Cc regions (81). The fact that the Ca, Cb, and Cc boxes are in phase by integrals of 10 or 1l-bp suggests that the positioning of OmpR may be important for transcription (78, 79). Insertion of different lengths of DNA oligomers between Cc and the -35 region causes periodic variations in ompC-lacZ expression, with insertions of half-integral turns having the largest reduction in lacZ expression (79). Similar results have also been reported for the ompF promoter, but in this case there is an additional contribution of DNA curvature (132). The region between -107 and-35 can also be

ARFAAN A. RAMPERSAUD

252

completely inverted without affecting ompC expression, but requires correct phasing of the binding sites on the DNA-helix (78). C. Transcriptional Activation at

ompFand ompC

The -10 and -35 regions of both ompF and ompC promoters are poorly related to their canonical consensus sequences, indicating that RNA polymerase by itself does not efficiently recognize these promoters (25, 90, 105, 141). Point mutations have been identified in the -35 region of both promoters, which make it more similar to the consensus sequence, and these subsequently allow OmpR-independent expression of both genes (25, 90, 105). In a related study, OmpR binding sites were fused to other poor -35 and -10 regions and shown to enhance transcription of a lacZ reporter gene in an OmpR-dependent fashion (141). These studies support the view that OmpR facilitates ompF and ompC transcription by improving or enhancing recognition of ~e -35 and -10 sequences by RNA polymerase (141). In vitro, both ompF and ompC transcripts can be made using OmpR and RNA polymerase holoenzyme as sole protein components (3, 15, 48, 53, 103, 113). Both linear (3, 15, 48, 53, 103, 113) as well as supercoiled templates (48) have been used for this purpose, but the phosphorylated state of OmpR was not reported in all cases (48, 103, 113). A few studies have directly examined the involvement of OmpR-P for in vitro transcription and have shown significant stimulation by OmpR-P (3, 48). In one set of experiments, differential transcription of ompF and ompC promoters was studied by providing both promoters in single round transcription experiments (3). OmpF transcription was favored at low OmpR-P, while high OmpR-P favored ompC transcription over ompF (3). This is entirely consistent with the current model of osmoregulation. The OmpR phosphorylation requirement for in vitro transcription is not absolute. As mentioned earlier T83A and G94S replacements in OmpR suppress the phosphorylation defect caused by D55Q. The double mutants OmpR(D55Q, T83A) and OmpR(D55Q, G94S) are fully active for the in vitro transcription of the ompF promoter (15). This means that the phosphate on OmpR is not intimately involved with interactions between OmpR and RNA p o l y p . The results add further supl~rt to the view that the primary role of phosphorylation is for a conformational change.

IX. OTHER FACTORS A. The (xSubunit of RNA-polymerase The a subunit of RNA polymerase is encoded at the rpoA gene and is important for proper assembly of the core RNA polymerase enzyme (a2[3[Y)(60, 119). Genetic studies indicate that OmpR interacts with the a-subunit for porin gene expression (35, 84, 125) and is part of a growing body of evidence indicating an impc~ant role

TransmembraneSignaling in Osmoregulation

253

of this factor for proper interactions with positive-acting transcriptional regulators (60, 119). In the case of OmpR, the a-subunit is suspected to influence both positive and negative regulation. Genetic evidence for an a-subunit-OmpR interaction came originally from the isolation of the rpoA77 mutant, which suppressed the porin phenotype of an envZll/ompR77 double mutant (84) and rpoA85, which suppressed the pleiotropic phenotype caused by envZ473 (35). In later studies, localized mutagenesis of the rpoA gene led to the identification of four additional rpoA mutants designated rpoASO, rpoA52, rpoA53, and rpoA54 (125). These four mutants were shown to not only affect porin gene expression but also the pleiotropic phenotype of envZ473

(125).

In general, rpoA mutants are remarkably limited in the types of genes they affect (60, 119); the ones changing porin production or suppressing pleiotropic envZ mutants seem to be specific for OmpR (125). A given rpoA allele is also strain dependent for its action, and when different rpoA alleles are tested in the same strain background, such as envZ473, they cause different porin and/or pleiotropic phenotypes. The complexity is thought to reflect the highly specific nature of OmpR-~ interactions with variable phenotypes attributed to different states or functional forms of OmpR (125). The amino acid replacements of several rpoA mutants cluster at the C-terminus of the a-subunit and are thought to represent potential sites for OmpR interaction (125). Support for an OmpR-a-subunit interaction has come from functional studies of a-subunits missing portions of their carboxyl-terminus (52). Truncated a-subunits allow assembly of a 2 ~ ' RNA core particles when provided with o 7~ facilitate transcription of the lacUV5 promoter (52). However, the modified RNA polymerase holoenzyme does not provide for OmpR-dependent (OmpR-P) transcription of the ompC promoter (52).

B. Integration Host Factor Integration host factor (IHF) is a dimeric molecule composed of subunits encoded at the himA and hip (hireD) genes (34). It has an important role in site specific recombination events, DNA bending, and DNA packaging (34) and was mentioned earlier to affect ompB expression (49, 140). IHF also binds to two sites in the ompF promoter (between -199 and -159 and between -80 and -42) and one site in the ompC promoter (between - 193 and - 158). In the ompF promoter the - 199 and - 159 region is a high-affinity site for IHF, as well as a DNA bending center for the promoter (113). It is not clear how this bending region is related to that reported between -100 and -70 (85). IHF negatively regulates both ompF and ompC promoters since purified protein prevents OmpR-dependent transcription of ompF as well as RNA polymerase mediated transcription of ompC (48, 113, 139). Both himA and hired mutants increase the amount of OmpF protein produced at high osmolarity (113). One

ARFAAN A. RAMPERSAUD

254

explanation for this result may be that IHF helps form the negative repression loop in the ompF promoter through its DNA-bending activity. The loss of bending activity in IHF mutants may make it more difficult to adopt the looped structure and thus allows for continued expression of ompF. An IHF mutant also allows OmpC expression in an ompR472 strain (139). Possibly, IHF acts as a negative regulatory at ompC or influences the DNA binding properties of OmpR. As IHF also has general pleiotropic effects on a host of other genes, it is also possible that the effects of IHF mutants are indirect.

X. FUTURE STUDIES Over the last few years extensive studies have shed new light on the mechanism of porin gene regulation by OmpR and EnvZ. As a result, several new features can be added to the model discussed at the beginning of this chapter. These include potential dimerization of EnvZ, interactions between OmpR subunits, cooperative DNA-binding by OmpR, formation of a negative repression DNA-loop in the ompF promoter, and interactions between OmpR and the 0c-subunit of RNA polymerase. This new information also raises new questions as to the molecular details of osmosensing, pleiotropic effects, transcriptional activation, and the role of other proteins in the regulatory process. Several aspects regarding EnvZ and OmpR function remain to be determined. Details concerning the discrete DNA-binding mechanism of OmpR need to be clarified, as well as how various protein-protein interactions contribute to transcriptional regulation. This information should provide new insight into the molecular events occurring at both ompF and ompC promoters and contribute to our present understanding of gene regulation. Deducing the signals for EnvZ and how they modulate the kinase and phosphatase activities is also necessary to clearly define the sensing mechanism. Structural studies with OmpR and EnvZ would also be extremely helpful in elucidating the nature of these molecules.

ACKNOWLEDGMENTS I am grateful to Drs. M.I. Inouye, T.J. Silhavy,T. Mizuno, and S. Forst for providing reprints and preprints of their work. I thank S. Forst and S. Harlocker for reviewing the manuscript and for their helpful comments. I also owe a special thanks to E. Borem and J. Hutchison for their excellent secretarial assistance.

REFERENCES 1. Aiba, H., Mizuno, T., & Mizushima,S. (1989). 'rramfer of phosphorylgroups between two regulatory proteins involved in osmoregulatoryexpression of the ompFand ompCgenes in Escherichiacoli.J. Biol. Chem.264,8563-8567.

TransmembraneSignaling in Osmoregulation

255

2. Aiba, H., Nakasai, F., Mizushima, S., & Mizuno, T. (1989). Evidence for the physiological i ~ e of the phosphotransfer between the two regulatory compo~nts, EnvZ and OmpR in Escherichio coil J. Biol. Chem. 264, 14090-14094. 3. Aiba, H., & Mizuno, T. (1990). Phosphorylation of a bacterial activator protein, On~R, by a protein kinase, EnvZ, stimulates the transcription of the ompF and ompC genes in E$cherichia coli. FEBS Lett. 261, 19-22. 4. Aiba, H., Nakasai, E, Mizushima, S., & Mizuno,'T. (1989). Phosphorylation of a bacterial activator protein, OmpR, by a protein kinase, EnvZ, results in stimulation of its DNA-binding ability. J. Biochem. 106, 5-7. 5. Albright, L. M., Huala, E., & Ausbel, E M. (1989). Prokaryotic signal transduction mediated by sensor and regulator protein pairs. Annu. Reg Genet. 23, 311-336. 6. Anderson, J., Forst, S. A., Zhao, K., & Inouye, M. I. (1989). The function of micF RNA: micF PdqA is a major factor in the thermal regulation of OmpF protein in Escherichia coil J. Biol. Chem. 264, 17961-17970. 7. Anderson, J., Delihas, N., Ikenaka, K., Green, P. J., Pines, O., llercil, O., & Inouye, M. I. (1987). The isolation and characterization of RNA coded by the micF gene in Escherichia coli. Nucleic Acids. Res. 15, 2089-2101. 8. Aoyama, T., & Oka, A. (1990). A c o ~ mechanism of transcriptional activation by the three positive regulators, VirG, PhoB, and OmpR. FEBS Lett. 263, 1-4. 9. Bachmann, B. J. Linkage map of Escherichia coll. In E C. Neidhardt, J. L. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, & H. E. Umbarger (Eds.) Escherichia coli and salmonella t)phimurium: cellular and molecular biology, vol. 2 (pp. 807-876). Washington, DC: American Society for Microbiology. 10a.Bardwell, J. C., McGovern, K., & Beckwith J. (1991). Identification of a protein required for disulfide bond formation in viva. Cell. 67, 581-589. 10. Berman, M. L., & Jackson, D. E. (1984). Selection ofloc gene fusions in viva: ompR-lacZ fusions that define a functional domain of the ompR gene product. J. Bacterial. 159, 750-756. 11. Bernardini, M. L., Fontaine, A., & Sansonetti, P. J. (1990). The two-component regulatory system OmpR-EnvZ controls the virulence of Shigellafle.meri. J. Bacterial. 172, 6274-6281. 12. Bhriain, N. N., Dorman, C. J., & Higgins, C. E (1989). An overlap between osmotic and anaerobic stress responses: a potential role for DNA supercoiling in the coordinate regulation of gene expression. MoL Microbial. 3, 933-942. 14. Bourret, R. B., Hess, J. E, Borkovich, K. A., Pakula, A. A., & Simon, M. I. (1989). Protein phosphorylation in chemotaxis and two-component regulatory systems of bacteria. J. Biol. Chem. 264, 7085-7088. 15. Bowfin, V., Brissette, R. E., & Inouye, M. I. (1992). Two transcriptionally active OmpR mutants that do not require phosphorylation by EnvZ in an Escherichia coli cell-free system. J. Bacterial. 174, 6685--6687. 16. Bfissette, R. E., ~ung, K., & Inouye, M. I. (1991). Suppression of a mutation in OmpR at the putative phosphorylation center by a mutant EnvZ protein in Escherichia coil J. Bacterial. 173, 601-608. 17. Brissette, R. E., Tsung, K., & lnouye, M. I. (1991). Intramolecular second-site revertants to the phosphorylation site mutation in OmpR, a kinase-dependent transcriptional activator in Escherichio coil J. Bacterial 173, 3749-3755. 18. Bfissette, R. E., Tsung, K., & lnouye, M. I. (1992). Mutations in a central highly conserved non-DNA binding region of OmpR, an Eschericlda coil transcriptional activator, influence its DNA-binding ability. J. Bacterial. 174, 4907-4912. 19. Case, C. C., Bukau, B., Villarejo, M., & Boos, W. (1986). Contrasting mechanisms of envZ control of real and pho regulon genes in Escherichio coil. J. Bacterial. 166, 706-712.

256

ARFAAN A. RAMPERSAUD

20. Comeau, D. E., Ikenaka, K., Tsung, K., & lnouye, M. I. (1985). Primary characterization of the protein products of the Escherichia coli ompB locus: structure and regulation of synthesis of the OmpR and EnvZ proteins../. Bacteriol. 164, 578-584. 25. Collado-Vides, J., Magasanik, B., & Gralla, J. D. (1991). Control site location and transcriptional regulation in Escherichia coll. Microbiol. Reu 55, 371-394. 20 Coyer, J., Anderson, J., Forst, S. A., Inouye, M. I., & Delihas, N. (1990). micF RNA in ompB mutants of Escherichia coli: different pathways regulate micF RNA levels in response to osmolarity and temperature change. J. Bacteriol. 172, 4143-4150. 23. Csonka, L. N. (! 989). Physiological and genetic responses of bacteria to osmotic stress. Microbiol. Rev. 53, 121-147. 24. Csonka, L. N. (1991). ~ o t i c osmoregulation: genetics and physiology. Annu. Rev. Microbiol. 45, 569--606. 25. Dairi, T., lnokuchi, K., Mizuno, T., & Mizushima, S. (1985). Positive control of transcription initiation in Escherichia coil A base substitution at the Pribnow box renders ompF expression independent of a positive regulator. J. biol. Biol. 184, 1--6. 6Q Delgado, J., Forst, S. A., Harlocker, S., & lnouye, M. I. (1993) Identification of a phosphorylation site and functional analysis of conserved aspartic acid residues of OmpR, a transcriptional activator for ompF and ompC in Escherichia coil Mol. Microbiol. (in press). 27. DiRienzo, J. M., and lnouye, M. I. (1979). Lipid fluidity-dependent biosynthesis and assembly of the outer membrane prc~ins of E. coll. Cell 17, 155-161. 27a.Fiedler, W., & Rottering, M. (1988). Properties of Escherichia coil mutants lacking membranederived oligosaccharides. J. Biol. Chem. 263, 14684-14689. 28. Forst, S. A., Comcath D. E., Norioka, S., & Inouye, M. I. (1987). Localization and membrane topology of EnvZ, a protein involved in osmoregulation of OmpF and OmpC in Escherichia coll. J. Biol. Chem. 262, 16433-16438. 29. Forst, S. A., Delgado, J., Ramakrishnan, G. R., & Inouye, M. I. (1988). Regulation of ompC and ompF expression in Escherichia coil in the absence of envZ. I. Bacteriol. 170, 5080-5085. 30. Forst, S. A., Delgado, J., & lnouye, M. I. (1989). Phosphorylation of OmpR by the osmosensor EnvZ modulates the expression of the ompF and ompC genes of Escherichia coll. Proc. Natl. Acad. Sci. USA 86, 6052--6056. 31. Forst, S. A., Delgado, J., & Inonye, M. I. (1989). DNA-binding properties of the transcription activator (OmpR) for the upstream sequences of ompF in Escherichia coil are altered by envZ mutations and medium osmolarity. J. Bacteriol. 171, 2949-2955. 32. Font, S. A., Delgudo, J., Rampersaud, A., & lnouye, M. I. (1990). In vivo phosphorylation of OmpR, the transcriptional activator of the ompF and ompC genes in Escherichia coll. J. Bacteriol. 172, 3473-3477. 33. Foist, S. A., & lnouye M. I. (1988). Environng~tally regulated gene expression for membrane proteins in Escherichia coll. Annu. Rev. Cell Biol. 4, 21-42. 34. Freundlich, M., Ramani, N., Matthew, E., Sirko, A., & Tsui, P. (1992). The role of integration host factor in gene expression in Eschedchia coll. Mol. Microbiol. 6, 2557-2563. 35. Garrett, S., & Silhavy, T. J. (1987). Isolation of mutations in the a operon of Escherichia coil that suppress the transcriptional defect conferred by a mutation in the porin regulatory gene envZ. J. Bacteriol. 169, 1379-1385. 36. Garrett' S., Taylor, R. K., & Silhavy, T. J. (1983). Isolation and characterization of chain-terminating nonsense mutations in a porin regulator gene, em,Z. J. Bacteriol. 156, 62-69. 37. Garrett' S., Taylor, R. K., Silhavy, T. J., & Berman, M. L. (1985). Isolation and characterization of AompB strains of Escherichia coli by a general method based on gene fusions. J. BacterioL 162, 840-844. 38. Geiger, O., Russo, E, Silhavy, T. J., & Kennedy, E. P. (1992). Membrane-derived oligosaccharides affect porin osmoregulation only in media of low ionic strength. J. Bacteriol. 174, 1410--1413.

Transmembrane Signaling in Osmoregulation

257

39. Gibson, M. M., Ellis, E. M., Graeme-Cook, K. A., & Higgins, C. E (1987). OmpR and EnvZ are pleiotropic regulatory proteins: positive regulation of the tripeptide permease (rppB) of Salmonella typhimurium. Mol. Gen. Genet. 207, 120-129. 40. Graeme-Cook, K. A., May, G., Bremmer, E., & Higgins, C. E (1989). Osmotic regulation of pot'in gene expression: a role for DNA supercoifing. Mol. Microbiol. 3, 1287-1294. 41. Granett, S., & Villarejo, M. (1982). Regulation of gene expression in Escherichia coli by the local anesthetic procaine. J. Mol. Biol. 160, 363-367. 41a.Halegoua, S., & Inouye, M. I. (1979). Translocation and assembly of outer membrane proteins of Escherichia coil. Selective accumulation of precursors and novel assembly intermediates caused by phenethyi alcohol. J. Mol. Biol. 130, 39-61. 42. Hall, M. N., & Silhavy, T. J. (1979). Transcriptional regulation of EschericMa coil K-12 mjor outer membrane protein lb. J. BocterioL 140, 342-350. 43. Hail, M. N., & Silhavy T. J. (1981). The ornpB locus and the regulation of the major outer membrane porin proteins of Escherichia coli K-12. J. Mol. Biol. 146, 23--43. 44. Hail, M. N., and Silhavy T. J. (1981). Genetic analysis of the ompB locus in Escherichia coli K-12. J. Mol. Biol. 151, 1-15. 45. Hall, M. N., & Silhavy T. J. (1981). Genetic analysis of the major outer membrane proteins of Escherichia coil. Annu. Rev. Genet. 15, 91-142. 46. Harlocker, S., Rampersaud, A., Yang, W-P., & Inouye, M. I. (1993). Phenotypic reverent mutations of a new OmpR2 mutant (V203Q) of Escherichia coli lie in the envZ gene which encodes the OmpR kinase. J. Bacteriol. 175, 1956--1960. 47. Hemandez-Chico, C., San Millan, J. L., Koiter, R., & Moreno, E (1986). Growth phase and OmpR regulation of transcription of Microcin BIT genes. J. Bocteriol. 167, 1058-1065. 48. Huang, L., Tsui, P., & Freundlich, M. (1990). Integration host factor is a negative effector of in vivo and in vitro expression of ompC in Escherichia coll. J. Bacteriol. 172, 5293-5298. 49. Huang, L., Tsui, P., & Freundlich, M. (1992). Positive and negative control ofompB transcription in Eschet~chia coli by cyclic AMP and the cyclic AMP receptor protein. J. Bacteriol. 174, 664-67O. 50. Heyde, M., Coil, J-L., & Portafier, R. (1991). Identification of Escherichia coli genes whose expression increases as a function of external pH. Mol. Gen. Genet. 229, 197-205. 51. Heyde, M., & Portalier, R. (1987). Regulation of major outer membrane porin proteins of Escherichio coli KI2 by pH. Mol Gen. Genet. 208, 511-517. 52. Igarashi, K., Hanamura, A., Makino, K., Aiba, H., Mizuno, T., Nakata, A., & lshihama, A. (1991). Functional map of the tx subunit of Escherichia coli RNA polymerase: Two modes of transcription activation by positive factors. Proc. Natl. Acad. Sci. USA 88, 8958--8962. 53. Igo, M. M., Ninfa, A. J., & Silhavy, T. J. (1989). A bacterial environmental sensor that functions as a protein kinase and stimulates transcriptional activation. Genes Dev. 3, 598--605. 54. Igo, M. M., Ninfa, A. J., Stock, J. B., & Silhavy, T. J. (1989). Phosphorylation and dephosphorylation of a bacterial transcriptional activator by a transmembrane receptor. Genes Dev. 3, 1725-1734. 55. Igo, M. M., and Silhavy, T. J. (1988). EnvZ, a transmembrane environmental sensor of Escherichia coli, is phosphorylated in vitro. J. Bacteriol. 170, 5971-5973. 56. Igo, M. M., Slauch, J. M., & Silhavy, T. J. (1990). Signal transduction in bacteria: kinases that control gene expression. Nature New Biol. 2, 5-9. 57. Ikenaka, K., Tsung, K., Comeau, D., & Inouye M. I. (1988). A dominant mutation in Escherichia coil OmpR lies within a domain which is highly conserved in a large family of bacterial regulatery proteins. Mol. Gen. Genet. 21 I, 538-540. 58. Inokuchi, K., Furukawa, H., Nakamura, K., & Mizushima, S. (1984). Characterization by deletion mutagenesis in vitro of the promoter region of ompF, a positively regulated gene of Escherichia. J. MoL Biol. 178; 653--668.

258

ARFAAN A. RAMPEI~AUD

9. Inokuchi, K., Itoh, M., & Mizushima, S. (1985). Domains involved in osmoregulation of the

ompF gene in Escherichia coil J. Bacteriol. 164, 585-590. 60. Ishihama, A. (1992). Role of the RNA polymerase a subunit in transcription activation. Mol. Microbiol. 6, 3283-3288. 61. Jamieson, D. J., & Higgins, C. E (I 984). Anaerobic and leucine-dependentexpression of a peptide transport gene in Salmonella typhimurium. J. Bacteriol. 160, 131-136. 20 Janoff, A. S., Haug, A., & M ~ , E. J. (I 980). Anesthetics alter outer membrane architectme and temlg~ture range growth of Eschen'chio coll. Biochem. Biophys. Re$. Comm. 95, 1364-1371. 3. Jo, Y., Nara, F., Ichihara, S., Mizuno, T., & Mizushima, S. (1986). Purification and characterization of the OmpR protein, a positive regulator involved in osmoregulatory expression of the ompF and ompC genes in Escherichia coll. J. Biol. Chem. 261, 15252-15256. 4. Jovanovich. S. B., Martin, U, M., Record, M. T., & Burgess R. R. (1988). Rapid response to osmotic upshift by osmoregulated genes in Escherichia coil and Salmonella typhimurium. J. Bacteriol. 170, 534-539. Kanamaru, K., Aiba, H., Mizushima, S., & Mizuno, T. (1989). Signal transduction and osmoregulation in Escherichia coil, a single amino acid change in the protein kinase, EnvZ, results in loss of its phosphorylation and dephosphorylation abilities with respect to the activator protein, OmpR. J. Biol. Chem. 264, 21633--21637. 60 Kanama~m,K., Aiba, H., & Mizuno, T. (1990). Trammembrane signal transduction and osmoreguiation in Escherichia coil: 1. Analysis by site directed mutagenesis of the amino acid residues involved in phosphotransfer between the two regulatory components EnvZ and OmpR. J. Biochem. 108, 483--487. 67. Kanamaru, K., & Mizuno, T. (1992). Signal transduction and osrnoregulation in Escherichia coli: a novel mutant of the positive regulator OmpR that functions in a ~phorylation-independent manner. J. Biochem. I I I, 4 ~ 3 0 . 68. Kato, M., Aiba, H., & Mizuno, T. (1989). Molecular analysis by deletion and site-directed mutagenesis of the cis-acting upstream sequence involved in activation of the ompF promoter in Escherichio coll. J. Biochem. 105, 341-347. 90 Kato, M., Aiba, H., Tat,, S-I., Nishimura, Y., & Mizuno, T. (1989). Location of phosphorylation site and DNA-binding site of a positive regulator, OmpR, involved in activation of the osmoregulatory genes of Escherichia coll. FEBS Lett. 249, 168-172. 70. Kawaji, H., Miztmo, T., & Mizushima, S. (1979). Influence of molecular size and osmolarity of sugars and dextrans on the synthesis of outer membrane proteins O-8 and O-9 of Escherichia coll. J. Bacteriol. 140, 843-847. 71. Kofoid, E. C., & Parkinson, J. S. (1988). Transmitter and receiver modules in bacterial signaling proteJns. Proc. Natl. Acad. Sci. USA 85, 4981--4985. 72. LiljestrOm, P., Mlittlnen, P. L., & Palva, E. T. (1982). Cloning of the regulatory locus ompB of Salmonella t3,phimwium LT-Z I. Isolation of the ompR gene and identification of its gene product. Mol. Gen. Genet. 188, 184--189. 73. Liijestr0m, P., MiUitttnen, P. L.,& Palva, E. T. (1982). Cloning of the regulatory locus ompB of Salmonella typhimurium LT-2. ii. Identification of the envZ gene product, a protein involved in the expression of the porin wc~ins. MoL Gen. Genet. 188, 190-194. 4Q Liljestr0m, P., Laamanen, I., Pairs, E. T. (1988). Structure and expression of the ompB operon, the regulatory locus for the outer membrane porin regulon in Salmonella typhimurium LT-2. J. Mol. Biol. 201,663-673. 75. Liljestr0m, P., Luokkangtki, M., & Palva, E. T. (1987). Isolation and characterization of a substitution mutation in the ompR gone of Salmonella typhimurium. J. Bacteriol. 169, 438--441. 76. Lukat, G., McCleary, W. R., Stock, A. M., & Stock, J. B. (1992). Phosphorylation of bacterial response regulator proteins by low molecular weight phospho-dono~. Proc. Natl. Acad. Sci. USA 89, 718-722.

TransmembraneSignaling in Osmoregulation

259

77. Lundrigan, M., & Earhart, C. E (1981). Reduction in three iron-regulated outer membrane proteins and protein a by the Escherichia coil K-12 perA mutation. J. Bocteriol. 146, 804-807. 78. Maeda, S., & Mizuno, T. (1988). Activationof the ompC gene by the OmpR protein in Escherichia coli . J. Biol. Chem. 263, 14629-14633. 79. Maeda, S., Ozawa, Y., Mizuno, T., & Mizushima, S. (1988). Stereospecific positioning of the c/s-acting sequence with respect to the canonical Womoteris required for activation of the o m ~ gene by a positive regulator, OmpR, in Escherichia coll. J. Mol. Biol. 202, 433-441. 80. Maeda, S., & Mizuno, T. (1990). Evidence for multiple OmpR-binding sites in the upstream activation sequence of the ompC promoter in Escherichia coil: a single OmpR-binding site is capable of activating the promoter. J. Bacteriol. 172, 501-503. 81. Maeda, S., Takayanagi, K., Nishimura, Y., Maruyama, T., Sato, K., & Mizuno, T. (1991). Activation of the osmoregulated ompC gene by the OmpR protein in Escherichia coil: a study involving synthetic OmpR-binding sequences. J. Biochem. 110, 324--327. 82. Matsuyama, S-I., Mizuno, T., & Mizushima, S. (1984). Promoter exchange between ompF and ompC, genes for osmoregulated major outer membrane proteins of Escherichia coil K-12. J. Bacterioi. 158, 1041-1047. 83. Matsuyama, S-I., Mizuno, T., & Mizushima, S. (1986). Interactions between two regulatory proteins in osmoregulatory expression of ompF and ompC genes in Escherichia coli: a novel ompR mutation suppresses pleiotropic defects caused by an envZ mutation. J. Bacteriol. 168, 1309-1314. 84. Matsuyama, S-I., & Mizushima, S. (1987). Novel rpoA mutation that interferes with the function of OmpR and EnvZ, positive regulators of the ompF and ompC genes that code for outer-membrane woteins in Escherichia coil KI2. J. Mol. Biol. 195, 847-853. 85. Mizuno, T. (1987). Static bend of DNA helix at the activatorrecognition site of the ompF promoter in Escherichia coll. Gene. 54, 57-64. 6. Mizuno, T. (1991). Control of envelope protein synthesis in bacteria. In A. lshihama & H. Yoshikawa (Eds.) Control ofcell growth and division (pp. 141-159). Tokyo: Japan Sci. Soc. Press. 87. Mizuno, T., Kato, M., Jo, Y-L., & Mizushima, S. (1988). InteractionofOmpR, a positive regulator, with the osmoregulated ompC and ompF genes of Escherichia coll. J. Biol. Chem. 263, 1008-1988. 88. Mizuno, T., & Mizushima, S. (1987). Isolation and characterization of deletion mutants OfompR and envZ, regulatory genes for expression of the outer membrane proteins O n ~ and OmpF in Escherichia coll. J. Biochem. 101,387-396. 89. Mizuno, T., & Mizushima, S. (1990). Signal transduction and gene regulation through the phosphorylation of two regulatory components: the molecular basis for the osmotic regulation of the porin genes. Mol. Micmbiol. 4, 1077-1082. 0. Mizuno, T., & Mizushima, S. (1986). Characterization by deletion and localized mutagenesis in vitro of the pron~ter region of the Escherichia coil ompC gene and imlx~ance of the upstream DNA domain in positive regulation by the OmpR prc~in. J. Bacteriol. 168, 86-95. 91. Mizuno, T., Shinkai, A., Matsui, K., & Mizushima, S. (1990). Osmoregulatory expression of podn genes in Escherichia coil: a comparative study on strains B and K-12. FEMS MicmbioL Lett. 68, 289-294. 2. Mizuno, T., Chou, M-Y., & Inouye, M. I. (1983). A comparative study on the genes for three porins of the Escherichia coli outer membrane. DNA sequence of the osmoregulated ompC gene. J. Biol. Chem. 258, 6932-6940. 93. Nagasawa, S., Tokishita, S., Aiba, H., & Mizuno, T. (1992). A novel sensor-regulatorprotein that belongs to the homologous family of signal-transduction proteim involved in adaptive responses in Escherichia coll. Mol. Microbiol. 6, 799-807. 94, Nakashima, K., Kanamaru, K., Aiba, H., & Mizuno, T. (1991). Osmoregulatoryexpression of the porin genes in Escherichia coil: Evidence for signal titration in the signal transduction through EnvZ-OmpR phosphotransfer. FEMS Microbiol. Left. 82, 43-48.

260

ARFAAN A. RAMPERSAUD

95. Nakashima, K., Kanamaru, K., Aiba, H., & Mizuno, T. (1991). Signal transduction and osmoregu-

lation in Escherichia coll. A novel type of mutation in the phosphorylation domain of the activator protein. OmpR, results in a defect in its phosphorylation-dependent DNA binding. J. Biol. Chem. 266, 10775--10780. 6. Nara, E, Matsuyama S-I., Mizuno, T., & Mizushima, S. (1986). Molecular analysis of mutant ompR genes exhibiting different phenotypes as to osmoregulation of the ompF and ompC genes of Escherichia coll. MOl. Gen. Genet. 202, 194-199. 97. Nara, E, Mizuno, T., & Mizushima, S. (1986). Complementation analysis of the wild-type and mutant ompR genes exhibiting different phenotyig~ of osmoregulation of the ompF and ompC genes of Escherichia coil MOl Ge~ Genet. 205, 51-55. 98. Neidhardt, E C. (1987). Multigene systems and regulom. In E C. Neidhardt, J. L. Ingraham, K. B. Low, B. Magasanik, M. Schaechter & H. E. Umbarger (Eds.) Escherichia coli and salmonella typhimurium: Cellular and molecular biology, vol. 2 (pp. 1313-1317). Washington, DC: American Society for Microbiology. 9. Nikaido, H., & Vaara, M. (1985). Molecular basis of bacterial outer membrane permeability. Microbial. Reu 49, 1-32. 100. Nikaido, H., & Vaara, M. (1987). Outer membranes. In E C. Neidhardt, J. L. lngraham, K. B. Low, B. Magasanik, M. Scbaechter and H. E. Umbarger (Eds.) Escherichia coli and salmonella tsphimurium: Cellular and molecular biology, vol. 1 (pp. 7-22). Washington, DC: American Society for Microbiology. 101. Ninfa, A. (1991). Protein phosplgxylatiou and the regulation of cellular processes by the homologous two-comgonent regulatory systems of bacteria. In J. K. Setiow (Ed.) Genetic engineering (pp. 39--72), New York: Plenum Press. 102. Ninfa, A. J., Ninfa, E. G., Lupas, A. N., Stock. A., Magasanik, B., & Stock, J. (1988). Crosstalk between bacterial chemotaxis signal transduction proteins and regulators of Iran~ription of the Ntr regulon: evidence that nitrogen assimilation and chemotaxis are controlled by a common phos~fer mechamsm. Proc. Natl. Acad. Sci. USA 85, 5492-5496. 103. Noriokg S., Ramak~shnan. G., lkenaka, K., & Inoeye, M. I. (1986). Interaction of a transcriptional activator, OmplL with reciprocally osmoregulated genes, ompF and ompCo of Escherichia coli. J. Biol. Chem. 261, 17113-17119. 104. Ostrow, K. S., Silbavy, T. J., & Garrett, S. (1986). c/s-Acting sites required for osmoregulation of ompF expression in Escherichia coli K-12. J. Bacteriol. 168, 1165-1171. 105. Ozawa, Y., Mimno, T., & Mizushima, S. (1987). Roles of the Pribnow box in posRive regulation of the ompC and ompF genes in Escherichia coil J. Bacterial. 169, 1331-1334. 106. Ozawa, Y., Mizmhima, S., & Mizuno, T. (1990). Osmon~gulatoryexpression of the ompC gene in Escherichia coli K-12; ISI insertion in the upstream regulatory region results in constitutive activation of the promoter. FEMS Microbial. Lett. 68, 295-300. 107. Pages, J-M., & Lazdunski, C. (1982). Transcriptional regulation of ompF and lamB genetic expression by local anesthetics. FEM$ Microbial. Lett. 15, 153-157. 108. Parkimon, J. S., & Kofoid, E. C. (1992). Cmnmunication modules in bacterial signaling proteins. Annu. Rev. Genet. 26, 71-112. 109. Pidamen. M., Saarilahti, H. T, Kurkela,S., & Palva, E. T. (1986).In viva cloning and ~mracterization of abe regulatory locus ompR of Escherichia coll. MOl. C,eat Genet. 203, 520-523. II0. Pugsley, A. P. (1993). Ammtion in the dsbA gene coding for periplasmir disulfide oxidoreductase reduces t r a n s c r i ~ of the Escherichia call ompF gene. MOl. Gen. Genet. 237, 407-411. 111. Pugsley, A. P., Conrad, D. J., Sr C. A., & Gregg, T. I. (1980). In viva effects of local anesthetics on the production of major outer membrane proteins by Escherichia coli. Biochim. Biophys. Acta 599, 1-12. 113. Ramani, N., Huang, L., & Freundlich, M. (1992). In vilro interactions of integration host factor with the ompF Immmter-regulatory region of Escherichia coll. Mol. GetL Genet. 231,248-255.

TransmembraneSignaling in Osmoregulation

261

114. Ramakrishnan G., Comeau, D., lkenaka, K., & Inouye, M. I. (1986). Transcriptional control of gene expression: osmoregulation of porin protein synthesis. In M. Inouye (Ed.) Bacterial outer membranes (pp. 3-I 5). New York: John Wiley and Sons, Inc. 115. Rampersaud, A., & Inouye, M. I. (1991). Procaine, a local anesthetic, signals through the EnvZ receptor to change the DNA-binding affinity of the transcriptional activator OmpR. J. Bacteriol. 173, 6882-6888. 116. Rampersaud, A., Norioka, S., & Inouye, M. I. (1989). Characterization of OmpR binding sequences in the upstream region of the ompF promoter essential for transcriptional activation. J. Biol. Chem. 264, 18693-18700. 117. Rampersaud, A., Utsumi, R., Deigado, J., Forst, S., & Inouye, M. I. (1991). Ca2+-enhanced phosphorylation of a chimeric protein kinase involved with bacterial signal transduction. J. Biol. Chem. 266, 7633-7637. 118. Russo, E D., & Silhavy, T. J. (1991). EnvZ controls the concentration of phosphorylated OmpR to mediate o s ~ g u l a t i o n of porin genes. J. Mol. Biol. 222, 567-580. 119. Russo, E D., & Silhavy, T. J. (1992). Alpha: the Cinderella subunit ofRNA polymerase. J. Biol. Chem. 267, 14515-14518. 120. Russo, E D., Slauch, J. M., & Silhavy, T. J. (1993). Mutations that affect separate functions of OmpR, the phosphorylated regulator of porin transcription in Escherichia coil J. Mol. Biol. 231(2), 261-273. 121. Sanders, D. A., Castro, B. L., Stock, A. M., Burlingame, A. L., & Koshland, D. E. (1989) Identification of the site of phosphorylation of the chemotaxis response regulator, CheY. J. Biol. Chem. 264, 21770--21778. 122. Sarma V., & Reeves, P. (1977). Genetic locus (ompB) affecting a major outer-membrane protein in Escherichia coli. J. Bacteriol. 132, 23-27. 123. Scott, N. W., & Harwood, C. R. (1980). Studies on the influence of the cyclic AMP system on major outer membrane proteins of Escherichia coli K 12. FEMS Microbiol. Letts. 9, 95-98. 124. Slauch, J. M., Garrett, S., Jackson, D. E., & Silhavy, T. J. (1988). EnvZ functions through OmpR to control porin gene expression in Escherichia coil K-12. J. Bacteriol. 170, 439-441. 125. Slauch, J. M., Russo, E D., & Silhavy, T. J. (1991). Suppresser mutations in rpoA suggest that OmpR controls transcription by direct interaction with the ot subunit of RNA polymerase. J. Bocterioi. 173, 7501-7510. 126. Slauch, J. M., & Silhavy, T. J. (1989). Genetic analysis of the switch that controls porin gene expression in Escherichio coli K-12. J. Mol. Biol. 210, 281-292. 127. Slauch, J. M., & Silhavy, T. J. (1991). cis-acting ompF mutations that result in OmpR-dependent constitutive expression. J. Bacteriol. 173, 4039-4048. 128. Silhavy, T. J., Berman, M. L., & Enquist, L. W. (1984). Experiments with genefusions. Cold Swing Harbor, NY: Cold Swing Harbor Laboratory. 129. Stock, A. M., Mottonen, J. M., Stock, J. B., & Schutt, C. E. (1989). Three-dimensional structure of CheY, the response regulator of bacterial chemotaxis. Nature 337, 745-749. 130. Stock, J. B., & Lukat, G. S. (1991). Bacterial chemotaxis and the molecular logic of intracellular signal transduction networks.Annu. R~. Biophys. 20, 109-136. 131. Stock, J. B., Ninfa A. J., & Stock, A. M. (1989). Protein phosplggylation and regulation of adaptive responses in bacteria. Microbiol. Rev. 53, 450-490. 132. Takayanagi, K., & Mizuno, T. (1992). Activation of the osmoregulated ompF and ompC genes by the OmpR protein in Escherichia coli: A study involving chimeric Wonders. J. Biochem. 112, 1-6. 133. Tate, S-I., Kato, M., Nishimura, Y., Anita, Y., & Mizuno, T. (1988). Location of DNA-binding segment of a positive regulator, OmpR, involved in activation of the ompF and ompC genes of Escherichia coli. FEBS Len. 242, 27-30. 134. Taylor, R. K., Garrett, S., Sodergren, E., & Silhavy, T. J. (1985). Mutations that define the promoter of ompE a gene specifying a major outer membrane porin protein. J. Bacteriol. 162, 1054--1060.

262

ARFAAN A. RAMPERSAUD

135. Taylor, R. K., Hall, M. N., & Silhavy, T. J. (1983). Isolation and characterization of mutations altering expression of the major outer membrane porin proteins using the local anesthetic procaine. J. MOI. Biol. 166, 273-282. 136. Tokishita, S-l., Kojirna, A., & Mizuno, T. (1992). Transmembrane signal transduction and osmoregulation in Escherichia coil Functional imcotxm~e of the transmembrane regions of membrane-located protein kinase, EnvZ. J. Biochm. I I I, 707-713. 137. Tokishita, S-I., Kojirrm, A., Aiba, H., & Mizuno, T. (1991). Transmembrane signal transduction and osmoregulation in Escherichia coil Functional importance of the periplasmic region of membrane-located protein kinase, EnvZ. J. Biol. Chem. 266, 6780-6785. 138. Tokishita, S-I., Yamada, H. A., Aiba, H., & Mizuno, T. (I 990). Transmembrane signal transduction and osmoregulmion in Escherichia coli: II. The osmotic sensor, EnvZ, located in the isolated cytoplasmic membrane displays its phosphorylation and dephosphorylation abilities as to the activator protein, OmpR. J. Biochem. 108, 488-493. 139. Tsui, P., Helu, V., & Freundlich, M. (1988). Altered osmoregulation of ompF in integration host factor mutants of Escherichia coil J. Bacteriol. 170, 4950-4953. 140. Tsui, P., Huang, L., & Freundlich, M. (1991). Integration host factor binds specifically to multiple sites in the ompB promoter of Escherichia coli and inhibits transcription. J. Bacterial. 173, 5800--5807. 141. Tsung, K., Brissette, R. E., & Inouye, M. I. (1990). Enhancement ofRNA polymerase binding to promoters by a transcriptional activator, OmpR, in Esclnerichia coil: its positive and negative effects on ~ranscription. Proc. Natl. Acad. Sci. USA 87, 5940-5944. 142. Tsung, K., Brissette, R. E., & Inouye, M. I. (1989). Identification of the DNA-binding domain of the OmpR protein required for transcriptional activation of the ompF and ompC genes of Escherichia coli by in vivo DNA footprinfing. J. Biol. Chem. 264, 10104-10109. 143. Utsumi, R., Brissette R. E., Rmnpersaud, A., Forst, S. A., Oosawa, K., and Inouye, M. I. (I 989). Activation of a bacterial porin gene expression by a chimeric signal transducer in response to aspartate. Science 245, 1246-1249. 144. Van Aiphen, W., & Lugtenberg, B. (1977). Influence of osmolarity of the growth medium on the outer membrane protein pattern of Escherichia coll. J. Bacteriol. 131,623-630. 145. Verhoef, C., Lugtenberg, B., van Boxtel, R., deGraaff, P., & Verheij, H. (1979). Genetics and biochemistry of the peptidoglycan-associated proteins b and c of Escherichia coli K-12. Mol. Gen. Genet. 169, 137-1383. 146. Villarejo, M., & Case, C. (1984). envZ mediates transcriptional control by local anesthetics but is not required for osmoregulation in E. coll. J. Bacteriol. 159, 883-887. 147. Wandersman, C., Moreno, E, & Schwartz, M. (1980). Pleiotropic mutations rendering Es. cherichia coli K-12 resistant to bacteriophage TPl. J. Bacteriol. 143, 1374-1383. 148. Wanner, B. L., Smhy, A., & Beckwith, J. (1979). Escherichia coli pleiotropic mutant that reduces amounts of several periplasmic and outer membrane protmns. J. Bacteriol. 140, 229-239. 148a.Warmer, B. L. (1992). Is cross regulation by phosp~lation of two-component response regulator proteins important in bacteria? J. Bacteriol. 174, 2053-2058. 149. Waukau, J., & Forst, S. A. (1992). Molecular analysis of the signaling pathway between EnvZ and OmpR in Escherichia coli. J. Bacteriol. 174, 1522-1527. 1.50. Wurtzei, E. T., Chon, M-Y., & Inouye, M. I. (1982). Osmoregulation of gene expression, I. DNA sequence of the ompR gene of the ompB ~ of Escherichia coil and characterization of its gene product. J. Biol. Chem. 257, 13685-13691. 151. Yang, Y., & Inouye M. I. (1991). Intermolecular cmnplementation between two defective mutant signal-transducing receptors of Escherichia coli. Proc. Natl. Acad. Sci. USA 88, 11057-11061. 152. Yang, Y., & Inouye M. I. (1993). Requirement of hoth Idnase and phosphatase activities of an Escherichia coli receptor (Tazl) for ligand-dependent signal transduction. J. MOl. Biol. 231 (2), 261-273.

INDEX

Acidic phospholipids, role of in initiation of DNA replication, 2O7 Alkaline phosphatase, 3 Amino terminus of EnvZ, 243 Anion exchange mechanisms in bacteria, identification and reconstitution of, 105-128 carboxylate-linked exchange, 119= 122 electrogenic nature of exchange, 119-120 histamine, 122 histidine, 122 malolactic fermentation, 122 methanogenesis, 122 oxalic acid, 119 Oxalobacterf ormigenes, 119121 OxlT, 108 proton pumps, other indirect, 121-122 common themes and structural rhythms, 123-124 translocation pathway, 124 transmembrane segments, 123, 124 transport path, 123 UhpT antiporter of bacteria, 123-125 263

experiments with Pi-cxchange, 113-116 antiport and symport mechanisms, choice between, 113114 carbon-phosphorus ratio, 116 proton-linked symport, phenotype of, 116 stoichiometry and selectivity, 114-116 introduction, 106-109 anion-linked exchange in bacteria, 108-109 antiport, 107 carboxylate-linked exchanges, 108, 119-122 carrier-mediated transport, mechanisms of, 107 carriers, 106 chemiosmotic circuits, 106-107 co-transport, 107 facilitated diffusion, 107 G6P anion, 108 H +, dependence on 106 "indirect" proton pump, 108 Mitchell's theory, 106-107, 109= 111 OxlT, 108 Oxalobacterf ormigenes, 108 Pi-linked exchange, 108

INDEX

264

proton-motive force, 106, 107 pumps, ion-motive, 106 symport, 107 UhpT, 108, 111 uniport, 107 Pi-linked exchange mechanisms, 109-112, 124-125 current status, 111-112 electroneutral exchange, 112 "exchange-adsorption" hypothesis, refutation of, 109

GpT,

heterologous exchange of, 112 history, 109-111 Escherichia coil, 110-111, I 13 Micrococcus pyogenes, 109 Pi self-exchange, 111 Pit, 111 Pst, 110-111 Staphylococcus aureus, 109, 113-114 Streptococcus faecalis, 110 Streptococcus lactis, 109-111, 114-116 sugar-phosphate self-exchange, 112 sugar-phosphate transport systems, 111,113 UhpT, 111, 113, 123-125 reconstitution of Pi-linked exchange, 116-118 octylglucoside dilution, technique of, 117 osmolytes, 117-118 summary, 124-125 systems, other, 122 ATP-ADP translocase of Rickettsia prowazeki, 122 dicarboxylate exchange in E.coli, 122 in eukaryotes primarily, 122 in prokaryotes, 122 Asialoglycoprotein receptor, 21

Asp237 and Asp240, 134-138 ATP activity in SecA molecule, 6869 Azide, sodium, inhibition of ATPase activity in SecA, 68 Bacteriochlorophyll, role of in Rhodobacter sphaeroides, 91 BarA protein, 241 ~-barrel type membrane proteins, 2 Bayer bridges, 154 Bieker-Silhavy model of translocation complexes, 48-49 Bitopic membrane proteins, 20-21 Cardiolipin, 193 biosynthesis of, 202-203 c/s point mutant, 202 CDP-diacylglycerol, biosynthesis of, 197-198 cds mutants, 198 precursor to remaining phospholipids, 198 cds mutants, 198 Cecropin A, 25 Chaperones, molecular, 8 CheY protein, 236 (see also "OmpR") c/s-vaccenic acid, 193 Colicin A assembly, 205-206 Colicin N, 206 Crosstalk, 241-242 Crystal packing in porins, 180 DNA replication, initiation of, 206207 acidic phospholipids, 207 genetic approach, 207 oriC locus, 206 RecA protein, 207 rnk gene, 207 DnaJ, 39

Index

DnaK, 39, 206 in protein translocationin E. Coli, 63 in SecB, 68 DsbA, 52-53 in protein translocationin E. Coli, 64,74 DsbB, 52-53

265

osmosensing by, 242 Tar-EnvZ sensor, hybrid, 243244 structure, 235 Tar.EnvZ, 237 sensor, 243-244 Escherichia coli, 1-16 (see also "Membrane protein assembly") and Pi-linked exchange mechanisms, 110-111 protein translocation machinery of, 61-84 (see also "Protein translocation...") requirement of, 26-27 (see also "Membrane insertion...") secretion machinery, 7-8, 10 translocation machinery, genetic analysis of, 36-60 (see also "Genetics...") Export initiation domain, 44

Electrophoresis, 20 Endoplasmic reticulum, insertion of small proteins into,23-26 EnvZ, 219-262 (see also "Osmoregulation...") missense mutants, 233, 237 ompR77, 233-234 organization, 235 and amino acid sequence of, 226 T M I and TM2, 235 phosphorylation studiesof, 237-242 acctyl phosphate, 241 alternate routes of OmpR phos- Ffh, 7,26 phorylation, 241 as molecular chaperone, 74 autophosphorylation of, 237 in protein translocation in E. Coli, barA, 241 64, 73-74 crosstalk, 241 SRP, 73 dephosphorylation activity, 238 Flagella synthesis, 210 mutants, 238-239 flD gene, 210 phosphate relay cycle, 240-241 FtsH, 52 phosphate transfer reactions, fls Y gene, 40 237-238 Ftsy protein, 26 pleiotropic effects, 233 Fusion protein analysis, 3-4 point mutations and deletions, in genetics of protein export, 37 231,232-234 relationship with other proteins, /~-galactosidase, 3, 130 237 in genetics of protein export, 36-38 signaling, 242-245 Genetics of protein translocation, 35activation of by local anesthet6O ics, 244-245 anti-folding, 38-40 amino terminus as sensor SecB, 38-39 domain, 243 of secretory precursors, 38 membrane perturbants, signal sequences, 38-39 response to, 244-245 conclusion, 53

266

INDEX

cytoplasmic factors, 38-40 precursor targeting, 38-40 (see also "...cytoplasmic...") DnaK/DnaJ, 39 Sec factor interaction and intra4.5S RNA, 39-40 membrane translocation fts Y gene, 40 GroE depletion, 39 pathway, 47-50 heat shock proteins, 39 Bieker-Silhavy model of translocation complexes, 48-49 SecB chaperones, 38-39 intramembrane steps of transloS RP homologs, 39-40 cation,47-48 targeting, 39 Sec titration experiments, 48-49 E.coli genes involved in protein SecY-SecE interaction, 49-50 export, identification of, 36(see also "SecY...") 38 translocation intermediates, 48 cold-sensitive steps, 37-38 SecA, 40-41 export defect, 37 and azide concentration, 40 fusion proteins, 36-37 translocation ATPase, 40-41 9 ~-galactosidase, 36-37 signal peptide and early part of lep gene, 37 mature sequence, recogniprl selection, 38 tion of, 44-47 secA-lacZ gene fusion, 37 export initiation domain, 44 sec Y gene, 37 preprotein entry and transit, signal sequence mutations, 37 hypothetical pathway of, 46 factors, other possible, in translopretranslocation complex, 46, cation, 51-53 48-49 Band 1, 51 priG and suppressor-directed DsbA and DsbB, 52-53 inactivation, 45-47 FtsH, 52 SecY, prlA mutations in, 44-45 molecular chaperones, 52 translocation complex, 46 Off12, 52 translocation factors, other possiP12, 51 ble, 51-53 (see also in periplasm, 52-53 "...factors...") PspA, 51 Globomycin, 72 SecD, 52 GroE chaperonins, 26-27, 39 Skp, 51 GroEL, 23, 26 Ydr, 51-52 in protein translocation in introduction, 36 E. Coli, 63 membrane-embedded components in SecB, 68 of translocator, 41-44 GroES, 23, 26 prlA suppressor mutations, 4142, 44 SecD, 44 Hn, 208 Heat shock proteins, 39 SecE, 43 Helical packing in membrane proS~F, 44 teins, 5-6 SecY, 41-43 (seea/so "SecY gene")

Index

267

a-helical structure of membrane proteins, 2 Helix packing in C-terminal half of lactose permease, 129-144 (see also "Lactose...") Hydrophobicity analysis, 4 "Inside-positive" rule of von Heijne, 88 Integration host factor (IHF), 253254 Interleukin-2, 21 Intracytoplasmic membrane (ICM), 86 (see also "Rhodobacter sphaeroides. . .")

IPTG inducer, 201-202 Lactose permease, helix packing in C-terminal half of, 129-144 fluorescence labeling, site-directed, use of, 138-141 excimer, 139 pyrene maleimide, 139-141 intramembrane charged residues, functional interactions between putative, 134-138 amino acid substitutions, neutral, 134 Asp237 and Asp240, 134-138 C-less permease, 135, 138 "charge-pair neutralization," 136 helical wheel model of putative helices, 137 salt bridges, 5, 134, 138 second-site suppressor mutations, 134, 139-140 site-directed mutagenesis, 135 introduction, 129-131 /~-galactosides, 130 H§ symport in E. cog, 130-131 lac Y gene, 131, 133

transmembrane electrochemical ion gradient, 129 secondary structure, 131-134 complementation, 133-134 lac permease-alkaline phosphatase (lacY-phoA) fusion proteins, 131 N terminus and C terminus, 133-134 summary and concluding remarks, 141-142 lac Y gene, 131, 133, 209 lactose permeasr function, 209 LacZ fusions, 3-4 /3-1actamase, 3, 26 LamB protein, 147 assembly in pulse-chase experiments, 156 //]-galactosidase hybrid proteins, 153-154 Leader peptidase, 4, 8, 19, 20, 72-73 (see also "Signal peptidase...~) LHI complexes, 93-95 (see also "Rhodobacter sphaeroides. . .")

Lipopolysaccharide synthesis, 211 Lipopolysaccharides (LPS), 146 (see also "Outer membrane proteins...") connection of porins to, 178 Lipoprotein signal peptidase, 64, 7273 (see also "Signal peptidase...") LPS, insertion in outer membrane and, 158-166 (see also "Outer membrane proteins...") MI3 procoat protein, 19, 22, 25 Maltose-binding protein (MBP), 150-152 (see also "Outer membrane proteins...")

INDEX

268

Melittin, 25 Membrane insertion of small proteins, 17-33 amino- vs carboxylterminal translocation, 18o20 amphiphilic character, 19 leader peptides, 18 loop structure, 19 M 13 procoat protein, 19, 25, 26 partitioning of hydrophobic region, 20 Pt3, 19

chaperones, requirement of, 26-27 dihydrofolate reductase (DHFR), 27 DnaJ and DnaK, 26, 39 Ftsy protein, 26 GroE and S 9 26 GrpE, 26 Hsc70, 26-27 //-lactamase, 26 M 13 procoat, 26, 27 melittin, 27 methotrexate, 27 SRP, 26-27 conclusions, 27-28 aminoterminal translocation, 28 complex proteins, translocation of, 28 translocase-independent insertion not described in eukaryotes, 28 into endoplasmic reticulum (ER), 23-26 cecropin A, 25, 26 cytochrome bs, 25 docking protein, 25 frog skin peptide, 25 M2 protein, 25 M13 procoat, 25, 26 melittin, 25, 27 NEM, 25, 26 preproafactor, 25

preprolactin, 25 prepromelittin, 25 prokaryotic and eukaryotic pathways, comparison of models of, 24 proteolytic processing, 25 Sec components, 24 signal sequence, 25 SRP, 24 Xenopus, 26 introduction, 17-18 compartmentalization, 18 evolution, 18 member protein orientation, deter~ mination of, 20-21 asialoglycoprotein receptor, 21 bitopic, 20-21 electrophoresis, 20 GTP, 21 Interleuldn-2, 21 leader peptidase, 20 NaoK-ATPase, 21 P450, 21 paramyxovirus, 21 positive-inside rule, 20, 23 Sec components, 20, 22 signal anchor, 20-21 signal peptide, 20-21 SRP, 21, 24 topogenic signals, 21 translocation, 20 translocase-independent translocation, 21-23 antenna complex, 23 GroEL and GroES, 23 fight-harvesting complex, 22-23 OmpA protein, 22 P~, 21-22 positive-inside rule, 23 R. capsulatus, 22-23 Sec proteins, 22 signal sequence, 22

Index

Membrane protein assembly, 1-16 conclusions, 10-11 helix packing, 5-6 disulfide bridges between pairs of cysteines, 6 oligomerization of membrane proteins, 5-6 salt bridges, 5 introduction, 1-2 Eacherichia coli, 2 mechanism of membrane insertion, 6-10 ATPase activity, 8 chaperones, molecular, 8 E.coli secretion machinery, 7-8, 10 Ffh, 7 leader peptidase enzyme, 4, 8 periplasmic loops, translocation of, 6-9 sec-dependent or independent, 9 signal sequences, 7-9 signal-anchor sequence, 8, 9 stop transfer sequences, 9 translocon, I0 transmembrane segments, insertion of, 9-10 structure, 2-3 a-helical, 2 /t-barrel, 2 hydrogen bonding criterion, 2 hydrophobic interior, 2 porins, 2(see also "Porins") topology, 3-5 alkaline phosphatase, 3 experimental determination of, 3-4 fusion protein analysis, 3-4 //-galactosidase, 3 gene-fusion technology, 3 gydrophobicity analysis, 4 LacZ, 3-4 ~-lactamase, 3

269

leader peptidase enzyme, 4, 8 PhoA, 3-4 positive-inside rule, 4-5 prediction and experimental manipulation of, 4-5 topology-engineering, 4-5 vectorial-labeling methods, 3 micF gene, 222 Mitchell, chemiosmotic theory of, 106-107, 109-111 Molecular chaperone, SecB as, 67-68 Myristic acid, 193 Na-K-ATPase, 21 NADH dehydrogenase II, 210-211 NEM, 25 Oligomerization of membrane proteins, 5-6 OMPs, 145-173 (see also "Outer membrane proteins...") OmpB oberon, 222-234 amino acid differences between EnvZ and OmpR proteins, 227 EnvZ, organization and amino sequence of, 226 expression of, 228 genes, related, 227-228 model, 223-225 mutants, 228-232 ompR, 228-232 (see also "OmpR") OmpR, sequence of, 225 organization and transcription of, 225-227 regulatory role of, 222 OmpC and OmpF, 221-222 ompC promoter, 251-252 ompF-lacZ fusions, 249, 251 OmpR, 219-262 (see also "Osmoregulation...")

270

INDEX

relationships with other proteins, DNA-bninding and transcrip236 tional properties of, 247-252 CheY protein, 236 Fd box, 251 structure, 234-235 looping model, 250 DNA-binding region of, 236 ompC promoter, binding to, Orfl2, 52 251-252 Osmolytes, 118 (see also "Anion ompF-lacZ fusions, 249, 251 exchange...") ompF promoter, binding to, Osmoregulation, mechanism of 248-250 transmembrane signaling in, ompF and ompC, transcrip219-262 tional activation at, 252 EnvZ, organization of, 235 factors, other, 252-254 relationship with other proteins, integration host factor (IHF), 237 253-254 signaling, 242-245 (see also a subunit of RNA-polymerase, =EnvZ") 252-253 factors, other 252-254 mutants, 228-232 integration host factor (IHF), classes, three, 228 253-254 ompR77, 233-234 a subunit of RNA-polymerase, ompR101, 230 252-253 ompR472, 230-232 future studies, 254 phosphorylation studies of, 237introduction: signal transduction 242 and osmoregulation, 220acetyl phosphate, 241 221 alternate routes of, 241-242 DNA-binding proteins, 220 barA, 241 EnvZ, 221-222 crosstalk, 241-242 OmpR, 221-222 dephosphorylation of, 238 response regulator, 220 domain, 239 sensor protein, 220 mutants of, 238-239 ompB oberon, general studies of, OmpR-P, forms of, 239-240 222-234 (see also ~ompB phosphate relay cycle, 240-241 oberon" phosphate transfer reactions, OmpR, DNA-binding and trans237-238 criptional properties of, 247properties of, multifunctional, 252 (see also =OmpR'~ 245-247 OmpR, multifunctional properties conformational changes in, 245of, 245-247 (see also 246 "OmpR") D55Q, 245-246, 252 OmpR structure, 234-235 multimerization of, 246-247 dephosphorylation of, 238 relationship with other proteins, mutants, additional, 247 236 $48F, 246

Index

osmosensing by EnvZ, 242-245 (see also "EnvZ") MDOs, 242 phosphorylation studies of OmpR and EnvZ, 237-242 (see also "OmpR" and "EnvZ") porin gene expression, regulation of by EnvZ and OmpR, 221-222 micF gene, 222 OmpF and OmpC, 221-222 Outer membrane proteins in E.coli, export and assembly of, 145-173 conclusions and future prospects, 165-166 function of, 146-147 LamB protein, 147 OmpA protein, 147, 149 porins, 146, 149, 154 TonB protein, 147 Tsx protein, 147 inner membrane, transport across, 150-154 ATP hydrolysis,150 export apparatus, 150 intragenicexport information, 152-154 maltose-bindingprotein(MBP), 150-152 prcPhoE translocation,SecB in, 151-152 SecB, role of in protein export, 150-152 SecD and SecF, 150 insertion and assembly in, 155-165 assembly intermediates, detected in vivo, 155-157 cerulenin, 158 LamB assembly, 156 LPS and lipid biosynthesis, 158 metastable trimers, 156-157 pulse-chase experiments, 156

271

reconstitution of, in vitro, 159162 sorting signals in, 162-165 introduction, 146 lipopolysaccharides(LPS), 146 OMPs, 146 phospholipids, 146 pathway to outer membrane, 154155 Bayer bridges, 154 biogenesis, model for, 154 LamB/~-galactosidase hybrid proteins, 154 SecD, 155 structure, 148-149 P12, 51 functions of, 72 in protein translocation in E. Coil, 63 P450, 21 Palmitic acid, 192-193 Palmitoleic acid, 193 Pefiplasmic loops in memberane proteins, 6-9 signal sequences, 7 Pf3, 19, 21-22 PhoA fusions, 3-4 Phosphatidic acid, 192, 193 (see also "Phospholipids...") biosynthesis of, 195-197 acyl-ACP, 196 dgkA mutants, 197 diacylglycerol kinase, 197 membrane derived oligosaccharide (MDO), 197 pbB and plsC genes, 196-197 sn=glycerol-3=phosphate acyltransferase, 195-196 Phosphatidylethanolamine, biosynthesis of, 198-200 phosphatidylserine synthase (PS), 198-199

272

psd gene, 199 pssA gene, 198-199 pyruvate prosthetic group, 200 Phosphatidylglycerol, biosynthesis of, 200-202 IPTG inducer, 201-202 lipoprotein, 200 lpx B locus, 200-201 lpp mutation, 201 pggpA and pgpB genes, 200 pgsA gene, 200-201 phosphatidylglycerophosphate (PGP) synthase, 200 Phospholipids, role of in E.coli cell function, 189-217 bimynthesis of, 195-203 (see also under particular head) of cardiolipin, 202-203 of CDP-diacylglycerol, 197-198 of phosphatidic acid, 195-197 of phosphatidylethanolamine, 198-200 of phosphatidylglycerol, 200-202 chemical nature of, 191-194 acidic, 190, 193, 197 acyl chain diversity, 191-193 cardiolipin, 193 c/s-vacc~nic acid, 193 cyclopropane-containing fatty acids, 193 head group diversity, 193-194 metabolism, pathway of, 192 myristic acid, 193 palmitic acid, 192-193 palmitoleic acid, 193 phosphatidic acid, 192, 193 zwitterionic, 190, 193, 197 introduction, 190-191 assay for function of, lack of, 191 phosphatidylinositols as metabolic signals and membrane anchors, 190 structural diversity of, 190

INDEX

phospholipids, role of in cell functions, 203-211 acidic phospholipids, 205 cell motility, 209-210 chemotaxis, 209-210 colicin assembly, 205-206 (see also "Colicin assembly") DNA replication, initiation of, 206-207 (see also "DNA replication...") energy transduction systems, 210-211 IPTG levels, regulation by, 204 lactose permease function, 209 lipopolysacchafide synthesis, 211 pgsA3 mutation, 204 phosphofipid polymorphism, 207-209 protein translocation across inner membrane, 204-205 Sec proteins, 204-205 in vitro systems, 203 summary, 211-212 Pi-finked exchange mechanism, 105128 (see also "Anion exchange mechanisms...") PMF, 76 Pore eyelets, 177, 182, 183-185, 1869 see also "Porins...") longitudinal electric field of, 185 Porins, 2, 146-165 (see also "Outer membrane proteins...") /3-barrel structures, 2 gene expression, regulation of by EnvZ and OmpR, 221-222 (see also "Osmoregulation..,") phosphate relay cycle, importance of, 240-241 Porins, structure-function relationships in, 175-187 location, 177-182 aromatic girdles, 178, 179, 180181

Index

~8-barrel fold, 176, 181 crystal packing in, 180 folding pathway, 181 lipopolysaccharides (LPS), connection to, 178 orientation in membrane, 177179 polar contacts only, 180 pore eyelets, 177, 182, 183-185, 186 shape of channel, 177-178 permeation properties, 183-185 as charged capacitator, 183 diffusion, 183-184 electric separator, 183-18 longitudinal electric field, 185 permeability, 185 pore eyelet, 177, 182, 183-185, 186 voltage dependence, 185 specificity, 185 specific and nonspecific, 185 tetrapyrrois, 185 structure, 176-177 amino acid sequence analysis, 176 ~-barrel fold, 176, 181 chain fold, 176-177 mobilities of atoms, 176 outer membrane in Gramnegative bacteria, 176 trimeric, 176-180 X-ray analysis, 176 Positive-inside rule, 4-5, 20, 23 PpfA protein, translocation and, 74 prlA suppressor mutations, 41-42, 44 not mapped to SecD and SecF genes, 71 in SecY, 44-45 prig mutation, 47 Protease IV (signal peptide peptidase), 73

273 Protein assembly in membranes, 116 (see also "Membrane protein assembly") Protein translocation machinery of E.coli, 61-84 components involved in, 62-65 DnaK, 63 DsbA, 64 Ffh, 64 GroEL, 63 P12, 63 PpfA, 64 PspA, 64 4.5S RNA, 64 Sec proteins, 62-84 signal peptidase I, II, 64 signal peptide peptidase, 64 Ydr, 64 factors, other, 73-74 (see also under particular head) DsbA protein, 74 Ffh protein, 73-74 PpfA protein, 74 PspA protein, 74 4.5S RNA, 73-74 Ydr protein, 74 functions of Sec and related proteins, 67-72 (see also under particular head) SecA, 68-69 SecB, 67-68 SecD, 71 SecE, 69-71 SecF, 71 SecG (p12), 72 SecY, 69-71 interaction map of components, 77 introduction, 62 compartments of, four, 62 preprotein, 62 model for, 74-77 ATP, 76

274

interaction map of components, 77 preprotein binding, 76 protonmotive force (PMF), 76 Sec proteins, 76 secretory machinery, 76 stepwise translocation, 76 translocation complex, 76 molecules, numbers of in one E. Coli cell, 65-67 table, 67 overproduction and purification of Sec proteins, 65 cytosol fraction, 65 octylglucoside, 65 sarcosyl, 65 T7 promoters, 65 tac, 65 signal peptidases, 72-72-3 signal peptide peptidases, 73 Proteins, small, membrane insertion of, 17-33 (see also "Membrane insertion...") Protonmotive force (PMF), 76 psd gene, 199 mutant, 210 PspA, 51 in protein translocation in s Coli, 64, 74 pssA gene, 198, 202 mutant, 210 Pyrene maleimide, 139-141 RC-H, 87, 88-92 "insider positive" rule, 88 PUHA 1, 89 and RC-M, 88 RecA protein, 207 Reconstitution of anion exchange mechanisms in bacteria, 105-128 (see also "Anion exchange...")

INDEX

Rhoddocater sphaeroides and Rhodobacter capsulatus, pigment-protein complex in, 85-104 introduction, 86-87 electron micrographs, 86 intracytoplasmic membrane (ICM), 86 LHI complex, 91, 93-95 a and/3, 94 amino acid sequences, 94, 95 composition and structure, 93 F1696 assembly factor, 94-95 gene products, other, 94-95 membrane insertion and assembly, 93-94 PucC proteins, 95 puf operon, 93 LHII complex, 96-97 a and/] polypeptides, 96 carotenoids in assembly of, 96 composition and assembly, 96 F1696 LHI assembly factor, 97 fluorazapam, 96 gene duplication of LHI genes, 96 gene products, other, 97 pucC gene, 97 model, assembly, for, and ratio, 97-98 fixed photosynthetic unit (FPU), 97-98 OrfQ proteins, 98 pucC and F1696 gene products, resemblance between, 98 variable photosynthetic unit (VPU), 98 reaction center, 87-92 bacteriochlorophyU, role of, 8991 composition and structure, 8788

Index

gene products, other, 91-92 H, M, and L polypeptides, 87 LHI complexes, 91, 93-95, 9697 orfQ, 91-92 puf operon, 87, 89, 92, 93

pufQ, 91 puhA, 87 RC-H subunit, role of, 87, 8892 (see also "RC-H") Rhodopseudomonas viridis, 88 summary, 98-99 4.5S RNA in protein translocation in E. Coil, 64, 73-74 as molecular chaperone, 74 SRP, 73 Salt bridges, 5, 134, 138 Sec proteins involved in translocation, 62-84 (see also "Protein translocation...") functions of, 67-72 See titration experiments, 48-49 SccA, 40-41 functions of, 68-69 ATPase activity, weak, 68 azide, inhibition by, 68 cardiolipin, 69 negatively charged, 68 phosphatidylglycerol, 69 phospholipids, 68, 69 prcproteins, 68 signal-peptide dependent, 68 translocation ATPase, 40-41, 68 SecB chaperonins, 26-27, 38-39 functions of, 67-68 molecular chaperone, 67-68 preproteins, conformation of, 67 signal peptides, 67 trigger factor, 68 protein export, role in, 150-152

275 SeeD, 44 functions of, 71 MBP, 71 operon, 71 periplasmic domain, large, 71 spheroplasts, 71 as inner membrane proteins, 150 as translocation factor, 52 late stage of, participant in, 77 SecE, 43 functions of, 69-71 Bacillus subtilis, 70 channel, 70 proteoliposomes, 69-70 transmembrane segments, three, 70 -SecY interaction, 49-50 Band 1, 51 -degradation system, 50 SecF, 44 functions of, 71 operon, 71 periplasmic domain, large, 71 as stabilizer of SecD and SecY, 71, 77 as inner membrane proteins, 150 SecG (p12): functions of, 72 Sec Y gene, 37, 40-43 -degradation system, 50 dominant-negative variants, 43 export initiation domain, 44 functions of, 69-71 channel, major portion of, 70 hydrophobic, 70 membrane-spanning segments, ten, 70 proteoliposomes, 69-70 mutations in, 42 prlA suppressor mutations, 41-42, 44-45 -SecE interaction, 49-50 Band 1, 51

276

Signal sequences in membrane proteins, 7-9 Signal-anchor sequence, 8, 9, 20-21 Signal peptide, 20-21 peptidase in protein translocation in E.Coli, 64, 73 Protease IV, 73 Signal peptidase I and II in protein translocation in E. Coil, 64, 72-73 globomycin, 72 numbers in one E. Coli cell, 73 prolipoproteins, 72 Signal recognition particle (SRP), 73 Skp protein, 51 Spheroplasts, 71 SRP, 73 Stop transfer sequences, 9

INDEX

Tar-EnvZ sensor, 243-244 Tetrapyrrols, 185 (see also "Porins') TonB protein, 147 Topology-engineering, 4 Translocon, 10 Trimeric porins, 176-180 (see also "Porins") Tsx protein, 147 Ubiquinone, 210-211 UhpT, 108, 111, 113, 123-124 Vectorial-labeling methods, 3 Ydr, 51-52 in protein translocation in/~ Coil, 63, 74

Advances in Molecular and Cell Biology Edited by E. Edward Blttar, Department of Physiology, University of Wisconsin-Madison Volume 8, Organelles In Vlvo 1994, 191 pp. ISBN 1-55938-636-3

$97.50

Edited by I.an Bo Chen, Dana-Farber Cancer Institute, Harvard Medical School CONTENTS: Confocal Redox Imaging of Cells, Barry R. Mas. ters, Walter Reed Army Medical Center. Calcium Channels and Vasodilation, Alison M. "Gumey and Lucie H. Clapp, UMDS, St. Thomas Hospital, London. Changes in DNA Su-

percoiling Status of Cells Treated with Antineoplastic Drugs, William D. Wright and J.L. Roti Roti, Washington University School of Medicine. Functional Morphology of the Golgi Region: A Lectino-Electromicroscopic Exploration, Margit Pavelka and Adolf Ellinger, Universit~it Innsbruk. Receptor Mediated Endocytosis of Plasminogen Activators, Gouujun Bu, Phillipa A. Morton, and Alan L. Schwartz, Washington University School of Medicine. Lipid Trafficking in Hepatocytes: Relevance to Biliary Lipid Secretion, Kristien J.M. Zaal, Jan Willem Kok, Folkert Kuipers, and Dick Hoekstra, University of Gro;ningen, Faculty of Medicine. Identification and Characterization of Functional Secretory Cells: Advantages of Multiparameter Flow Cytometry Kinetics, Elizabeth R. Simons and Theresa A. Davies, Boston University. An Outline of Neurosecretion, Jane Somsel Rodman, Tufts University School of Medicine. Volume 9. Homing Mechanisms and Cellular Targeting

1994, 303 pp. ISBN 1-55938-686-X

$97.50

Edited by Bruce R. Zetter, Harvard Medical School and Childrens Hospital CONTENTS: Introduction, Bruce R. Zetter. Directed Cell Migration in Embryonic Blood Vessel Assembly, Thomas J. Poole, SUNY Health Science Center at Syracuse. Leukocyte

Interaction with Endothelium and Extracellular Matrix: The Role of Selectins and CO44, Ivan Stamenkovic, Massachusetts General Hospital and Harvard Medical School. Oligosacchadde-dependent Mechanisms of Leukocyte Adhesion, John B. Lowe, Univeslty of Michigan Medical School. Cell Adhesion Molecules: Novel Therapeutic Targets for Chronic Inflammatory Diseases of the Central Nervous System, Gregory N. Dietsch, Gary M. Peterman and W. Michael Gallatin. Molecular Mechanism of Targeting of Hemopoietic Stem Cells to the Bone Marrow After Intravenous Transplantation, Mehdi

J A l P R E S S

J A l

Tavassoli, University of Mississippi School of Medicine. Tumor Cell Adhesion and Growth in Organ Preference of Tumor Metastasis, Gareth L. Nicolson, The University of Texas M.D. Anderson Cancer Center. Cancer Cell Chemotaxis: Mechanisms and Influence on Site-Specific Tumor Metastasis, F. William Orr, McMaster University. Experimental Orthotopic Models of Organ-specific Metastasis by Human Neoplasms, Isaiah J. Fidler, The University of Texas, M.D. Anderson Cancer Center. Inter and Intracellular Targeting of Drugs, Srnadar Cohen, Ben Gurion University of the Negev and Robert Langer, Massachusetts Institute of Technology. Chimeric Molecules Constructed with Endogenous Substances, Gregory T. Lautenslager and Lance L. Simpson, Jefferson Medical Col. lege. Organ-Specific Targeting of Synthetic and Natural Drug Carders, S. Moein Moghimi, Llebeth Ilium and Stanley S. Davis, University of Nottingham. Volume 10, Molecular Processes of Photosynthuls 1994, 437 pp. $97.50

ISBN 1-55938-710-6

P R E S S

Edited by J. Barber, Imperial College of Science, Technology and Medicine, London CONTENTS: Organisation and Dynamics of Thylakold Membranes, B. Andersson, Stockholm University and J. Berber Wolfson Laboratories, England. Antenna Pigment-Protein Complexes of Higher Plants and Purple Bacteria, J.P. Thornher, University of California, R. Cogdell, The University of Glasgow, Scotland, P. Chitnis, D.T. Morishige, G.F. Peter, S. Gomez, S. Anandan, S. Preiss, B. Welty, A. Lee, T. Takeuchi, C. Keffield, Kansas State University. Adaptive Variations in Physobilisome Structure, A.N. Glazer, University of California. Photoprotection and Photoinhibiton/ Damage, W.S. Chow, CSIRO Division of Plant Indutry, Australia. Molecular Genetic Manipulation and Characterization of Mutant Photosynthetic Reaction Centres from Purple Non-Sulfur Bacteria, E. Talcshashi, C.A. Wraight, University of Illinois. PROTON-TRANSLATING NAD(P)-H Transhydrogenase and NADH Dehydrogenase in Photosynthetic Membranes, J.B. Jackson, University of Birmingham and A. McEwan, University of East Anglia, England. Structural Elements Involved in the Assembly and Mechanims of Action of Rubisco, S. Gutteridge, Dupont Company, T. Lundqvist, Swedish University of Agriculture. The FerredoxiNThioredoxin System: Update on its Role in the Regulation of Oxygenic Photosynthesis, R. Buchanan, University of California. Identification, Cellular Localization and Participation of Chaperonins in Protein Folding, A.A. Gatanby, P. Viitanen,E.L duPont de Nemours & Co., V. Speth, Institut fur Biologle, R. Grimm, Hewlett-Packard GmbH. Translocation of Proteins Across Chloroplast Membranes, B. Bruce, K. Keegstra, University of Wisconsin, Madison.

J A I P R E S S

Volume 11, In preparation, Summer 1995 ISBN 1-55938-844-7

Approx. $97.50

CONTENTS: List of Contributors. Preface, Kevin M. Brindle. Metabolic Channeling in Organized Enzyme Systems: Experiments and Models, Pedro Mendes, Douglas B. Kell, and G. Rickey Welch. Metabolic Control Analysis in Theory and Practice, Athel Comish-Bowden. Experimental Approaches to Studying Enzymes in Vivo: The Application of Nuclear Magnetic Resonance Methods to Genetically Manipulated Organisms, Simon-Peter Williams, Alexandra M. Fulton, and Kevin M. Brindle. Glycolysis in Vivo: Fluorescence Microscopy as a Tool for Studying Enzyme Organization in Living Cells, Len Pagliaro. The cooperative Behavior of Krebs Tdcarboxylic Acid Cycle Enzymes, Paul A. Srere, Craig R. Matloy, A. Dean Sherry, and Balazs Sumegi. NMR Studies of Erythrocyte Metabolism, Hilary A. Berthon and Philip W. Kuchel. Studies of Physiological Control of ATP Synthesis, K. F. LaNoue and C. Doumen. Hepatic High Energy Phosphate Metabolism in Transgenic Livers Expressing Creatine Kinase as Reveled by 31p NMR, Alan P. Koretsky, Kenneth R. Miller and Jessica M. Halow. Index. Also Available: Volumes 1-7 (1987-1993)

$97.50 each

FACULTY/PROFESSIONAL discounts are available in the U.S. and Canada at a rate of 40% off the list price when prepaid by personal check or credit card and ordered directly from the publisher.

JAI PRESS INC.

55 Old Post Road # 2 - P.O. Box1678 Greenwich, Connecticut 06836-1678 Tel: (203) 661- 7602 Fax:(203) 661-0792

J A 1 P R E S S

Advances In Cell and Molecular Biology of Membranes and Organelles (Previosly published as Advances In Cell and Molecular Biology of Membranes) Edited by Alan M. Tartakoff, Institute of Pathology, Case Western Reserve University Volumes of Advances in Biochemistry and the Biology of Cells are intended to present interrelated reviews concerned with structure and function of membranes of prokaryotes and eukaryotes. Topics may include membrane protein and lipid biosynthesis, membrane-cytosketal relations, membrane permeability, signal transduction, etc. Individual volumes will have a common theme. Volume 1, Endosomes and Lysosomes: A Dynamic Relationship 1993, 434 pp. ISBN 1-55938-362-3

$97.50

Edited by Brian Storrle, Department of Biochemistry and Nutrition, Hrginia Polytechnic Institute and State University and Robert R Murphy, Department of Biological Sciences, Carnegie Mellon University CONTENTS: Preface, Brian Stonfe, Virginia Polytechnic Institute and State University and Robert F. Murphy, Carnegie Mellon University. Models of Endosome and Lysosome Traffic, Robert F. Murphy, Carnegie Mellon University. Endocytic Receptors, Michael G. Roth, University of Texas Southwestem. Functions of the Mannose 6-Phosphate Receptors, Bernard Hoflack, European Molecular Biology Laboratory, Germany and Peter Lobel, University of Medicine and Dentistry of New Jersey. Chemistry of Lysosomal Cysteine Proteinases, Robert W. Mason, Virginia Polytechnic Institute and State University and Donna Wilcon, Washington University School of Medicine. Mechanism and Regulation of Autophagic Degradation of Cellular Proteins, II~lliarn A. Dunn, Jr., University of Rorida. Cell-Free Systems for Endocytosis, William A. Braell, Harvard Medical School. Genetic Analysis of Merebrans Traffic in Mammalian Cells, Penelope A. Colbaugh and Rockford K. Draper, The University of Texas at Dallas. Plasma Membrane Lipid Transport in Cultured Cells: Studies Using Upid Analogs and Model Systems, Michael Koval, Washington University School of Medicine. Endosomes, Lysosomes, and Trans-Golgi-Related Systems in Conventional Neurons and the Grof Retina: Shards and Suppositions, Eric Holtzman, Eliene Augenbraun, Robert St. Jules, and Mafia Santa-Hernandez, Columbia University. The Role of Endocy-

tosis in Epidermal Growth Factor Signaling, Bryan K. McCune, Johns Hopkins University School of Medicine, William R. Huckle, and H. Shelton Earp, University of North Carolina at Chapel Hill. Membrane Traffic Through the Late Stages of the Yeast Secretory, Eric A. Whitters, Henry B. Skinner, and Vytas A. Bankaitis, University of Alabama at Birmingham. Regulation of Lysosomal Trafficking and Function During Growth and Development of Dictyostelium Discoideum, James A. Cardelli, Louisiana State University Medical Center. Towards an Understanding of the Inheritance of Mammalian Lysosomes and Yeast Vacuoles, Brian Storrie, Virginia Polytechnic Institute and State University. Volume 2, Membrane Transport In Protozoa

1993,483 pp. 2 Part Set Set ISBN 1-55938-628-2

J A

$195.00

Edited by Helmut Plattner, Fakult4t for Biologie, Universit4t Konstanz PART A - CONTENTS: Preface. Involvement of the TransGolgi Network, Coated Vesicles, Vesicle Fusion and Secretory Product Condensation in the Biogenesis of Pseudomicrothorax Tdchocysts, Robert K. Peck, Barbara Swiderski and Anne-Marie Tourmel, Geneva, Switzerland. Early Steps of the Secretory Pathway in Paramecium: Ultra-structural, Immunocyto-Chemical and Genetic Analysis of Tdchocyst Blogenesis, Nicole Garreau de Loubresse, Gff-sur-Yvette, France. Calcium and Tdchocyst Exocytosls in Paramecium: Genetic and Physiological Studies, Jean Cohen and Daniel Kerboeuf, Gifsur-Yvette, France. Exocytotic Events During Cell Invasion by Apicomplexa, Jean Francois Dubremeta and Rolf Entzeroth, Villeneuve dAsca, France~Bonn, FRG. Pathways of Lysosomal Enzyme Secretion in Tetrahymena, Amo Tiedtke, Thomas Kiy, Christian VosskOhler and Left Rasmussen, MOnster, FRG/Odense, Denmark. Synchronization of Different Steps of the Secretory Cycle in Paramecium Tetraurelia: Tdchocyst Exocytosis, Exocytosis-Coupled Endocytosis and Intracellular Transport, Helmut Plattner, Gerd Knoll, and Regina Pape, Konstanz, FRG. The Ciliary Membrane and its Engagement In Conjugation, Jason Wolfe, Middletown, CT. Ciliary and Plasma Membrane Proteins in Paramecium: Description, Localization and IntraceUular Transit, Yvonne Capdeville, Ren6e Charret, Claude Anthony, Julienne Delorme, Pierre Nahon and Andr6. Adoutte, Orsay, Francea. PART B - CONTENTS: Endocytosis and Intracellular Transport of Variant Surface Glycoproteins in Trypanosomes, Michael Duszenko and Andreas Seyfang, TObingen, FRG. A Comparative Survey on Phagosome Formation in Protozoa, Klaus Hausmann and Renate Radek, Berlin, FRG. Endosomal Membrance Traffic of Ciliates, Richard D. Allen and Agnes K. Fok, Honolulu, HI. Membrane Flow in the Digestive Cycle in Paramecium, Agnes K. Fok and Richard D. Allen, Honolulu, HI. Signal Coupling During Endocytosis in Amoeba Proteus,

P R E S S

J A 1 P R E S S

Robert D. Prusch, Spokane, WA. Membrane Recycling and Tumover in Large, Free-Living Amoebae, Kwang W. Jeon, Knoxville, TN. Food Uptake and Digestion in Amoebae, Wil. helm Stockem and Melpo Christofidou-Solomidou, Bonn, FRG. The Lysosomal System in Malaria Parasites, Christian Slomianny, Villeneuve d'Ascq, France. Membrane and Microtubule Dynamics in Heliozoa, Toshinobu Suzaki and Yoshinobu Shigenaka, Hiroshima, Japan. The Host-SymbiontInterface in Ciliate-Algae Associations: Inhibition of Membrane Fusion, Wemer Reisser, G6ttingen, FRG. Lipid Composition of Membranes Involved in Membrane Traffic in Tetrahymena, Shigenobu Umeki and Yoshinori Nozawa, Okayama, Japan. Volume 3 Signal Transduction Through Growth Factor Receptors 1994, 223 pp. ISBN 1-55938-344-5

Approx: $97.50

Edited by: Yasuo Kitagawa, BioSclences Center, Laboratory of Organogenesis, Nagoya University and Ryuzo Sasald, Faculty of Agriculture, Department of Food Science and Technology,, Kyoto Universi~ CONTENTS: The Hepatocyte Growth Factor/c-MET Signaling Pathway, D. P. Bottario, A. M.-L. Chan, J. S. Rubin, E. Gak, E. Fortney, J. , Schindler, M. Chedid, and S., A. Aaronson. Insulin Receptor, Y. Ebina, H. Hayashi, F. Kanai, S. Kamohara, and Y. Nishioka. Intedeukin-3 Receptor: Structure and Signal Transduction, T. Kitamura, and A. Miyajima. Interleukin-5 Receptor, K. Takatsu. Intedeukin-6 Receptor and Signal Transduction, T. Matsuda, T. Nakajima, T. Kaisho, K. Nakajima, and T. Hirano. Receptor for Granulocyte ColonyStimulating Factor, S. Nagata, & R. Fukunaga. Receptor for Granulocyte/Macrophage Colony-stimulating Factor, K. Kurata, T. Yokota, A. Miyajima, & K. Arai. Perspectives On The Structure And Mechanisms Of Signal Transduction By The Erythropoietin Receptor, S. S. Jones. Intedeukin-1 Signal Transduction, J. E. Sims, T. A. Bird, J. G. Giri, and K. S. Dower FACULTY/PROFESSIONAL discounts are available in the U.S. and Canada at a rate of 40% off the list price when prepaid by personal check or credit card and ordered directly from the publisher.

JAI PRESS INC.

55 Old Post Road # 2 - P.O. Box 1678 Greenwich, Connecticut 06836-1678 Tel: (203) 661- 7602 Fax:(203) 661-0792