Glycoproteins II, Volume 29 (New Comprehensive Biochemistry)

GLYCOPROTEINS I1

New Comprehensive Biochemistry

Volume 29B

General Editors

A. NEUBERGERt L.L.M. van DEENENt

ELSEVIER Amsterdam . Lausanne . New York . Oxford . Shannon . Singapore . Tokyo

Editors

J. Montreuil Universiti des Sciences et Technologies de Lille, Laboratoire de Chimie Biologique (UMR no. 111 du CNRS), 59655 Villeneuve d’Asq Cedex, France

J.F.G. Vliegenthart Bijvoet Center for Biomolecular Research, Department of Bio-organic Chemistry, RO. Box 80.075, 3508 TB Utrecht, The Netherlands

H. Schachter Department of Biochemical Research, Hospital for Sick Children, 555 University Avenue, Toronto, Ont. M5G 1x8, Canada

1997 ELSEVIER Amsterdam . Lausanne . New York . Oxford . Shannon . Singapore . Tokyo

Elsevier Science B.V PO. Box 21 1 1000 AE Amsterdam The Netherlands

L i b r a r y of C o n g r e s s Cataloging-in-Publication

Data

G l y c o p r o t e i n s I1 e d i t o r s , J. M o n t r e u i l , J.F.G. V l i e g e n t h a r t . H. Schachter. p. cm. -- ( N e w c o m p r e h e n s i v e b i o c h e m i s t r y ; v . 2 9 b ) I n c l u d e s b i b l i o g r a p h i c a l r e f e r e n c e s a n d index. ISBN 0-444-82393-X 1. G l y c o p r o t e i n s . I. M o n t r e u i l , J e a n , 192011. V l i e g e n t h a r t . J. F. G. 111. S c h a c h t e r . H. ( H a r r y ) IV. S e r i e s . [ D N L M : 1. G l y c o p r o t e i n s . W 1 N E 3 7 2 2 F v.29b 1997 / QU 55 G5682 19971 Q D 4 1 5 . N 4 8 vol. 2 9 b [QP552.G591 572 s--dc21 [572'.S81 97-40500

CIP ISBN 0 444 82393 x ISBN 0 444 80303 3 (series) 01997 Elsevier Science B.V

All rights reserved

No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher, Elsevier Science B.V., Copyright and Permissions Department, P.O. Box 521, 1000 AM Amsterdam, the Netherlands. No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of the rapid advances in the medical sciences, the publisher recommends that independent verification of diagnoses and drug dosages should be made. Special regulations for readers in the USA - This publication has been registered with the Copyright Clearance Center Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923. Information can he obtained from the CCC about conditions under which photocopies of parts of this publication may be made in the USA. All other copyright questions, including photocopying outside the USA, should be referred to the publisher.

Printed on acid-free paper Printed in the Netherlands

Preface Glycoproteins I/ completes a three volume contribution to the New Comprehensive Biochemistry series of Elsevier; the two previous volumes were entitled Glycoproteins (Vol. 29a, 1995) and Glycoproteins and Disease (Vol. 30, 1996). The Editors hope that these volumes follow in the footsteps of the 1965 and 1972 works entitled Glycoproteins edited by Alfred Gottschalk and published in Elsevier’s B.B.A. Library series. Of course only you, the reader, can decide whether we have succeeded in matching those magnificent milestones in glycobiology literature. In 1973, one year after the appearance of the Second Edition of Gottschalk’s Glycoproteins, one of us (JM) organized an international glycoprotein symposium in Lille, France, that was subsequently named “The Second International Symposium on Glycoconjugates”. The proceedings were published (J. Montreuil, Methodologie de la structure et du mktabolisme des glycoconjugues: glycoproteines et glycolipides, Editions du CNRS, 1974, 2 volumes). Gottschalk attended that conference and passed away a few months later. It is quite remarkable that Gottschalk’s book and his death ended an era in glycoprotein research and that the Second International Symposium was the clear beginning of a new and explosive period in the field. Gottschalk showed an uncanny prescience with the following remark at that conference: “We are not at the end of all progress but at the beginning; we have but reached the shores of a great unexplored continent.” These two eras are reviewed by one of us (JM) in Chapter 1 of Vol. 29a of the present series. The contents of Parts A and B of Gottschalk’s Second Edition filled 1378 pages, indicating that there was no shortage of excellent science at that time. The emphasis was on the physical and analytical chemistry of individual glycoproteins and their constituent sugars. There were only a few chapters with discussions of biological function such as the role of sialic acid in influenza virus cellular receptors (co-written by Gottschalk, an expert in the area), the degradation of glycoproteins by lysosomal enzymes and plasma glycoprotein turnover. The sections on glycoprotein biosynthesis reported work on subcellular fractionation and in uitro glycosyltransferase assays using crude extracts. Gottschalk concluded the book with a discussion of the “present state of glycoprotein research”. He marvelled at the progress in the seven years between his two editions and indeed there had been striking advances. However, one can only imagine the exultation that Gottschalk would have expressed had he been able to review progress in our field since he wrote his little chapter! These rapid advances are the result ‘of many different factors: techniques such as lectin chromatography, high resolution nuclear magnetic resonance spectrometry, mass spectrometry and X-ray crystallography; entirely new approaches to glycoprotein research permitted by advances in synthetic carbohydrate chemistry, molecular biology, genetics, cell biology, immunology and various other disciplines; the analysis of human diseases (accidents of nature), of experimentally produced somatic cell mutations, and of transgenic mice and mice with ‘‘null’’ mutations. We have tried in these three volumes to V

v1

cover many of these exciting aspects of glycoprotein research. Not all topics have been included and coverage of included topics is not always comprehensive. The time has long since passed when comprehensive coverage of the glycoprotein field could be achieved in a single work. Perhaps Gottschalk’s books were the last such tomes. We hope, however, that we have presented a significant portion of the excitement of modern glycobiology and that the reader will be stimulated by these books to delve further into the field and contribute to its continued growth and success.

J. Montreuil J.F.G. Vliegenthart H. Schachter

In Memoriam To our great sorrow, we have to report the death of Professor Albert Neuberger CBE FRS on August 14th 1996. Albert Neuberger was born in 1908 in Northern Bavaria. He studied Medicine in the University of Wuerzburg being awarded an MD summa cum laude, and also studied in Berlin where he was able to work in the biochemistry laboratories of Peter Rona. When Hitler came to power in 1933 he moved to London. He gained a PhD in 1936 for his work at University College Hospital Medical School and remained there for six years. After a period in Cambridge, in 1942 he moved to the National Institute for Medical Research where he remained until 1954. He then became Professor of Chemical Pathology at St Mary’s Hospital Medical School, London, where he remained for 18 years. On retirement he continued his research in the Biochemistry Department of the Charing Cross Hospital Medical School. During these periods he held many other important posts including the Principalship of the Wright-Fleming Institute and the Chairmanship of the Lister Institute. Albert Neuberger was a member of the editorial board of Biochimica et Biophysica Acta from 1968 to 1981 and associate managing editor for the last 13 years of this period. He was also one of the editors of Comprehensive Biochemistry and of New Comprehensive Biochemisq. Throughout his life Albert Neuberger showed a very broad interest in biochemistry and medicine. He made highly significant contributions in the areas of the chemistry and biochemistry of amino acids (particularly of glycine, serine, tryptophan and hydroxyproline), nutrition (with particular regard to amino acids and proteins), porphyrin biosynthesis, the chemistry of sugars (particularly amino sugars), lysozymes, lectins and particularly glycoproteins. His pioneering work on glycoproteins started in 1936. Up until the mid-1950’s most biochemists felt that, with the possible exception of mucins, the carbohydrate which could be detected in protein preparations was just an impurity. Neuberger felt that some of these proteins could have covalently attached sugars and chose to investigate hen eggwhite albumin. Before chromatography had been developed, the only possible purification method was by repeated ( x 7) crystallization until the preparation had a constant ratio of protein to carbohydrate. By proteolytic digestion he was able to isolate a compound with composition Man 4 : GlcNAc 2 : unknown amino acid 1 (suspected asparagine or glutamine). This problem was taken up again in 1955 with Marshall and Johansen and they were able to prove that the linkage between the protein and carbohydrate moieties was between asparagine and N-acetylglucosamine. This has subsequently been shown to be a conserved linkage structure in animals, plants, protozoa and other organisms. Albert Neuberger was an outstanding scientist with an impressive intellect and a range of interests outside science. He was very well liked and respected by his colleagues for his tolerance, good humour and high scientific standards. vii

...

Vlll

Albert Neuberger, 1908-1 996

He was also a devoted family man and we extend our sympathy to his widow and family. London, December 1996 Anthony K. Allen

List of contributors* Merton Bernfield 1 Harvard Medical School, and Joint Program in Neonatology, 300 Longwood Ave., Enders 9, Boston, MA 02115, USA Kenneth J. Clemetson 173 Theodor Kocher Institute, University of Berne, Freiestrasse I , 3012 Berne, Switzerland Robert N. Cole 69 University of Alabama at Birmingham, Schools of Medicine and Dentistry, Department of Biochemistry and Molecular Genetics, 404 Basic Health Sciences Building, I918 University Boulevard, Birmingham, AL 35294-0005, USA Ten Feizi 571 The Glycosciences Laboratory, Northwick Park Hospital, Watford Road, Harrow, Mid&. HA1 3UJ UK. Jukka Finne 55 Department of Medical Biochemistry, University of Turku, Kiinamyllynkatu 10, FIN-20520 Turku, Finland Hudson H. Freeze 89 The Burnham Institute, 10901 North Torrey Pines Road, La Jolla, CA 92037, USA Irwin J. Goldstein 403 Department of Biological Chemistq University of Michigan Medical School, 1301 Catherine Rd., Ann Arbol; MI 48109-0624, USA Kenneth D. Greis 33 Department of Biochemistry and Molecular Genetics, UAB School of Medicine and Dentistry, UAB Station, Birmingham, AL 35294, USA Gerald W. Hart 33, 69 University of Alabama at Birmingham, Schools of Medicine and Dentistry, Department of Biochemistry and Molecular Genetics, 404 Basic Health Sciences Building, I918 University Boulevard, Birmingham, AL 35294-0005, USA * Authors’ names are followed by the starting page number@) of their contribution(s).

ix

X

R. Colin Hughes 507 National Institute for Medical Research, The Ridgeway, Mill Hill, London NW7 IAA. UK. Sadako Inoue 143 School of Pharmaceutical Sciences, Showa University, Hatanodai-I, Tokyo 142, Japan Yasuo Inoue 143 Department of Biophysics and Biochemistry, Graduate School of Science, University of Tokyo, Hongo-7, Tokyo 113, Japan Johannis P. Kamerling 123, 243 Bijvoet Center, Department of Bio-Organic Chemistry, Utrecht University, PO. Box 80,075, 3508 TB Utrecht, Netherlands Robert Kokenyesi 1 Edith Nourse Rogers Memorial Veterans Administration Medical Center, Building 70, 200 Springs Road, Bedford, MA 01730, USA; Department of Rheumatology/lmmunology, Brigham and Women b Hospital, and Joint Program in Neonatology, 300 Longwood Ave., Enders 9, Boston, MA 02115, USA Reiko T. Lee 601 Department of Biology, Johns Hopkins University, 3400 N. Charles St., Baltimore, MD 21218, USA Yuan C. Lee 601 Department of Biology, Johns Hopkins University, 3400 N Charles St., Baltimore, MD 21218, USA Halina Lis 475 Department of Membrane Research and Biophysics, The Weizmann Institute of Science, Rehovot 76100, Israel Joel Mazurier 203 Universitk des Sciences et Technologies de Lille, Laboratoire de Chimie Biologique (UMR no. 111 du CNRS), Bdtiment C9, 59655 Villeneuve dilscq Cedex, France Jean Montreuil 203 Universitk des Sciences et Technologies de Lille, Laboratoire de Chimie Biologique (UMR no. 111 du CNRS), Bdtiment C9, 59655 Villeneuve d’Ascq Cedex, France Ronald D. Poretz 403 Department of Biochemistry and Microbiology, Rutgers University, New Brunswick, NJ 08903, USA

XI

Roland Schauer 243 Biochemisches Institut, Christian-Albrechts- Universitat zu Kiel, Olshausenstr 40, 24098 Kiel, Germany Nathan Sharon 475 Department of Membrane Research and Biophysics, The Weizrnann Institute of Science, Rehovot 76100, Israel Jeremiah E. Silbert 1 Edith Nourse Rogers Memorial Veterans Administration Medical Centel; Building 70, 200 Springs Road, Bedford, MA 01 730, USA; Department of Rheumatology/lmmunology, Brigham and Womenb Hospital, and Harvard Medical School, 300 Longwood Aue., Enders 9, Boston, MA 02115, USA Genevikve Spik 203 Universiti des Sciences et Technologies de Lille. Laboratoire de Chimie Biologique (UMR no. I l l du CNRS), BGtiment C9, 59655 Villeneuve d 'Ascq Cedex, France Gerard Strecker 163 Universiti des Sciences et Technologies de Lille, Laboratoire de Chimie Biologique (UMR no. 111 du CNRS), BBtiment C9, 59655 Villeneuve d 'Ascq Cedex, France Johannes F.G. Vliegenthart 123 Bijvoet Centel; Department of Bio-Organic Chemistry, Utrecht University, PO. Box 80.075, 3508 TB Utrecht, Netherlands Harry C. Winter 403 Department of Biological Chemistry, University of Michigan Medical School, 1301 Catherine Rd., Ann Arbor, MI 48109-0624, USA Jean-Pierre Zanetta 587 Centre National de la Recherche Scientijique, Center of Neurochemistry, 5,rue Blaise Pascal, 67000 Strasbourg, France

This Page Intentionally Left Blank

Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

V

In Memoriam Albert Neuberger . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

vii

List of contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix

Other volumes in the series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xxv

Chapter 1. Proteoglycans: a special class of glycoproteins Jeremiah E . Silbert. Merton Bernjeld and Robert Kokenyesi . . . . . . . . . . . . .

1

Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Glycosaminoglycans . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Linkage region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1. Chondroitin sulfate, dermatan sulfate, heparan sulfate, heparin . . . . . . . . . . . 2.2.2. Keratan sulfate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 . Biosynthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 . Degradation and turnover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Classification, distribution, and function of proteoglycan . ................... 5.1. Matrix proteoglycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.1. Aggrecadversican family . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.2. Leucine-rich core protein family . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.3. Perlecan family . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.4. Part-time proteoglycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Cell surface proteoglycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.1. Syndecan family . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.2. Glypican family . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.3. Part-time proteoglycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. Intracellular proteoglycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.1. Serglycin family . . . . . . . . ........................ 5.3.2. Other proteoglycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.3. Part-time proteoglycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 . Fine structure/function relations of glycosaminoglycans . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

20 21 21 21 22 22 22 25

Chapter 2. Nuclear and cytoplasmic glycoproteins Kenneth D. Greis and Gerald W Hart . . . . . . . . . . . . . . . . . . . . . . . . . . . .

33

Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

33 33

...

XI11

1

1

3 3 6 6 7 7 12 15 16 16 17 18 19 19 19

XIV

2 . 0-linked N-acetylglucosamine modified proteins . . . . . . . . . . . . . . . . . . . . . 2.1, The enzymes of 0-GlcNAc cycling . . ................ 2.2. Nuclear proteins that contain 0-GlcNAc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1. Nuclear pore proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2. Chromatin proteins, transcription factors and RNA polymerase I1 . . . . . . . . . 2.2.3. Estrogen receptor, Aplasia 83 kDa protein and autoantigen p43 . . . . . . . . . . 2.3. Cytoplasmic proteins that contain 0-GlcNAc . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. I . Cytoskeletal glycoprotein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.2. Eukaryotic initiation factor 2-associated 67 kDa polypeptide ......... 2.4. Viral proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Other cytosolic glycoproteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Phosphoglucomutase and parafusin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Glycogenin . . ............................ 3.3. Cytosolic fucosylation . ........................... 3.4. Cytosolic proteoglycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. N-linked GlcNAc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 . Other nuclear glycoproteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Glycosaminoglycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. High mobility group proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. N-linked glycoproteins of the nucleus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

34 36 38 38 39 40 40 41 41 42 43 43 44 45 46 47 48 48 48 49 50 50 51

Chapter 3. Carbohydrate units of nervous tissue glycoproteins Jukka Finne . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

55

Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . ................... 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 . Core structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Terminal sequences of N-linked glycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 . Classical 0-linked oligosaccharides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 . 0-mannose linked oligosaccharides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 . Poly-N-acetyllactosamine glycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 . Polysialic acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8. Other structures . . . . . . . . ..................... ..... .... 9 . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

55 55 55 56 58 59 60 62 64 65 66

Chapter 4. Glycosyl-phosphatidylinositol anchors: structure. biosynthesis and function Robert N Cole and Gerald K! Hart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

69

Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 . Structure of GPI anchors . . . . . . . . . . . . . . . . . . . . ................... 2.1. Common core structure of GPI anchors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Structural diversity of GPI anchors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Identification of a GPI-anchored protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Biosynthesis of GPI-anchored proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Biosynthesis of GPI anchors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

69 69 70 70 70 72 72 74

xv 3.2. GPI anchor remodeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. GPI anchor attachment to protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Topology of GPI anchor biosynthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. Defects in GPI anchor biosynthesis: paroxysmal nocturnal hemoglobinuria . . . . . . . . 4 . Proposed functions of GPI anchors . . . . . . . . . ........................ 4.1. Lateral mobility . . . . . . . . . . . . . . . . . . . . . ................... 4.2. Protein release . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. Protein targeting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. Endocytosis and protein turnover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5. Signal transduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References (Literature review completed as per 29 May 1995) . . . . . . . . . . . . . .....

78 78 79 80 80 80 81 82 83 84

Chapter 5. Dictyostelium discoideum glycoproteins: using a model system for organismic glycobiology Hudson H. Freeze . . . . . . . . . . . ...... ...................

89

1 . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 2. The life cycle of Dictyostelium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Dictyostelium as a single celled organism . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Development in Dictyostelium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Types of protein glycosylation in Dictyostelium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. N-linked oligosaccharides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 . 1. Biosynthesis of the lipid precursor and processing . . . . . . . . . . . . . . . . . . 3.1.2. Phosphorylation and sulfation of N-linked oligosaccharides . . . . . . . . . . . . 3.2. Two unusual types of 0-linked protein glycosylation in Dictyostelium . . . . . . . . . . . 3.2.1, Phosphoglycosylation in Dictyostelium . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2. Another type of 0-linked .............. 3.3. Glycophospholipid anchors . . . ................... 3.4. Cytoplasmic glycosylation in Dic ..... 4 . Antibodies against glycans and mutants in glycosylation . . . . . . . . . . . . . . . . . . . . . . . 4.1, Determinants found on N-linked oligosaccharides . . . . . . . . . . . . . . . . . . . . . . . 4.2. Mutants in 0-linked oligosaccharides-modB . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. Antibodies against fucose-mAb 83.5 and MUD62 . . . . . . . . . . . . . . . . . . . . . . 4.4. mAbs 81.8, 40.1 and MUD9 . . . . . . . . . . . . 5 . Glycoproteins in specific aspects of the Dictyostelium li 5.1. Mating types in Dictyostelium . . . . . . . . . . . 5.2. Getting around . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. Preparing for development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4. Glycoconjugates in cell adhesion during development . . . . . . . . . . . . . . . . . . . . . 5.4.1. General features in aggregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4.2. The EDTA-sensitive adhesion molecules . . . . . . . . . . . . . . . . . . . . . . . . 5.4.3. The EDTA-resistant adhesion molecule . . . . . . . . . . . . . . . . . . . . . . . . . 5.4.4. Post-aggregation adhesion system . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4.5. The surface sheath . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4.5.1. Cellulose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4.5.2. Glycoantigens and glycoproteins . . . . . . . . . . . . . . . . . . . . . . . . 5.4.6. Pre-spore vesicles and the spore coat . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4.6.1. Pre-spore vesicles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4.6.2. Spore coat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 . Molecular glycobiology in Dictyostelium discoideum . . . . . . . . . . . . . . . . . . . . . . . . .

77 77

89 89 89 89

92 92 92 93 97 97 100 101

102 102 104 104 105 106 106 106 106 107 107 107 108

109 110 111

111

112 112 113 115 115

XVI

Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Chapter 6. Hemocyanins Johannis I? Kamerling and Johannes FG. Vliegenthart

..

123

Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 . Carbohydrate parameters of arthropod hemocyanins . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Androctonus australis hemocyanin glycans . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Aslacus leptodactylus hemocyanin glycans . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Panulirus interruptus hemocyanin glycans . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 . Carbohydrate parameters of mollusc hemocyanins . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Lymnaea stagnalis hemocyanin glycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Helix pomatia hemocyanin glycans . . . . . . . . . . ..... ..... 4 . Synthesis and conformational analysis of xylose-containing elements of mollusc hemocyanin glycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 . Biosynthesis of Lymnaea stagnalis hemocyanin glycans . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

123 123 124 126 127 127 129 129 132

Chapter 7. Fish glycoproteins Sadako Inoue and Yasuo Inoue

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

List of abbreviations used . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 . Cortical alveolus glycoproteins (hyosophorins) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. First isolation of polysialoglycoproteins (PSGP) from rainbow trout eggs and their ubiquitous occurrence in salmonid fish . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Occurrence of a deaminoneuraminic acid residue (Kdn) at the non-reducing end of oligo/poly-Sia chains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Diversity in oligo/poly-Sia chains of salmonid egg PSGP . . . . . . . . . . . . . . . . . . 2.4. Fish egg PSGP is a cortical alveolar component . . ..... 2.5. Polyprotein nature of apo-PSGP and the molecular mechanism of fertilization-associated depolymerization . . . . . . . . . . . . . . ......................... 2.6. Molecular cloning of apo-PSGP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.7. Biosynthesis of polysialyl glycan chains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.8. Hyosophonns bearing bulky multi-antennary N-glycan chains . . . . . . . . . . . . . . . . 2.9. Biological function of hyosophorin and future perspective . . . . . . . . . . . . . . . . . . 2.9.1. Formation and function of the penvitelline fluid . . . . . . . . . . . . . . . . . . . 2.9.2. Sperm agglutinating properties of hyosophorins . ......... ... 2.9.3. Calcium ion binding properties of hyosophorins . . . . . . . ... 2.9.4. De-N-glycosylation of hyosophorins . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 . Mucin-type glycoproteins found in the vitelline envelope and ovarian fluid of salmonid fish . 3.1, Isolation and glycan structures of Kdn-gp and Sia-gp . . . . . . . . . . . . . . . . . . . . . 3.2. Biosynthesis and possible functions of Kdn-(Sia-)gps . . . . . . . . . . . . . . . . . . . . . 4 . Glycoproteins related to vitellogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 . Epilogue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

136 137 140 140

143 143 143 144 144 145 146 146

146 148 148 149 151 152 152 153 153 154 154 156 156 157 I59

xvii

Chapter 8. Amphibian glycoproteins Gerard Strecker . . . . . . . . . . . . . .

....

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

Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Key for glycosyltranferase activities, examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 . Role of the oviducal secretions in mediation of gamete fusion in amphibians . . . . . . . . . . 3 . Carbohydrate chains of egg jelly coat glycoproteins . . . . . . . . . . . . . . . . . . . . . . . . . 4 . Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Chapter 9. Blood glycoproteins Kenneth 1 Clemetson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Abbreviations . . . . . . . . . . . . ................................. 1 . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 . Plasma proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1, al -Acid glycoprotein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Antithrombin 111 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. a 1-Antitrypsin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Apolipoproteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5. Ceruloplasmin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6. C1 inhibitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.7. Complement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.7.1. Complement C3 .................................... 2.8. Factor J complement inhibitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.9. Factor V . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.10. Factor VII . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 1 . Factor VIIl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.12. Factor IX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.13. Factor X . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.14. Factor XI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.15. Factor XI1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.16. a-Fetoprotein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.17. Fibrinogen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 8. Fibronectin . . . . . . . ............................... 2.19. Hemopexin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.20. a2-HS-glycoprotein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.21. a2-Leucine-rich glycoprotein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.22. a2-Macroglobulin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.23. Plasminogen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.24. a 1 -Proteinase inhibitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.25. Protein C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.26. Protein S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.27. Prothrombin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.28. Transferrin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.29. Vitronectin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.30. Von Willebrand factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Platelets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1, Glycoprotein Ib-V-IX complex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.1. Glycoprotein Iba/glycocalicin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

163 163 163 163 164 165 169 170

173 173 173 174 174 176 177 178 179 179 180 180 180 180

181 181 183 184 185 185 185 185

186 187 187 188 188 189 189 190 190 190 190 190 191 192 192 192

xviii 3.1.2. Glycoprotein Ibfi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.3. Glycoprotein IX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.4. Glycoprotein V . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.4.1. Leucine-rich domains . . . . . . . . . . . . . . . . . . . . . . . 3.2. Glycoprotein IIb-IIIa (all&) . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1, Glycoprotein IIb . . . . . . . . . . . . ........................ 3.2.2. Glycoprotein IIIa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. fil Integrin family . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. CD36 (GPIIIb, GPIV) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. PECAM-I (CD31) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6. Thrombospondin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References

Chapter I0. Transferrin superfamily. An outstanding model for studying biochemical evolution Jean Montreuil. Geneuidue Spik and Joel Mazurier . . . . . . . . . . . . . . . . . . . . 1 . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 . 1 . Definition of transferrin superfamily . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.2. Biological importance of iron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. The transferrin superfamily . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1, The saga of transferrins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1, Ovotransferrins (conalbumins) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.2. Serotransferrins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.3. Lactotransferrins (lactoferrins) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.4. Melanotransferrin (human melanoma-associated antigen p97) . . . . . . . . . . . 2.2. Comparative study of transferrin peptide chains . . . . . . . . . . . . . . . . . . . . . . . . 2.2. 1 . Primary and three-dimensional structure . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2. Location of glycosylation sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2.1. Serotransferrins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2.2. Ovotransferrins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2.3. Lactotransferrins . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2.4. Melanotransferrin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2.5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Role of transferrins and of their receptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.1. Serotransferrin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. I . 1. Role of serotransferrin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.1.2. Serotransferrin receptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.2. Lactotransferrin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.2. 1 . Role of lactotransferrin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.2.2. Lactotransferrin receptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.3, Ovotransferrin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.4. Melanotransferrin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Comparative study of transferrin glycan primary structures . . . . . . . . . . . . . . . . . . . . . 3 . I . Normal transferrin glycans . . . . . . . . . . . . . . ................... 3.1.1. Serotransferrin glycans . . . . . . . . . . . . . . . . . . . . . . 3.1.2. Lactotransferrin glycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.3. Ovotransferrin glycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.4. Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

i94 194 194 194 195 195 196 196 196 198 198 198 198

203 203 203 204 204 204 205 205 206 207 207 207 210 210 212 213 214 214 214 214 214 215 216 216 218 218 219 219 219 219 222 223 226

XIX

3.2. Physiopathological modifications of transferrin glycan primary structure . . . . . . . . . 3.2.1.Physiological modifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1.1. Pregnancy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1.2. Embryogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1.3. Tissue-dependent glycosylation . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2.Serotransferrin and disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2.1. Liver diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2.2. Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2.3. CDG syndromes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2.4. Serotransferrin in HEMPAS . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Three-dimensional structure of transferrins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Three-dimensional structure of glycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Three-dimensional structure of transferrins as glycoproteins . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Chapter 11. Chemistry. biochemistry and biology of sialic acids Roland Schauer and Johannis l? Kumerling ..... ............ List of Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. General characteristics of sialic acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Occurrence of sialic acids in biomolecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Screening of biological materials for the presence of sialic acid . . . . . . . . . . . . . . . . . . 5. Isolation and analysis of sialic acids . ........................ 5.1. Liberation . . . . . . . . . . . . . . . . . . . . ..................... 5.2. Colorimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. Chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.1. Thin-layer chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.2.High-performance liquid chromatography . . . . . . . . . . . . . . . . . . . . . . . 5.3.3. Gas-liquid chromatography combined with mass spectrometry . . . . . . . . . . . 5.3.4.Fast atom bombardment mass spectrometry . . . . . . . . . . . . . . . . . . . . . . 5.3.5. H NMR spectroscopy . . . . . . . . ........................ 6. Chemo-enzymatic highlights in sialic acid chemistry . . . . . . . . . . . . . . . . . . . . . . . . . 6.1. Free sialic acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2. Glycosides of sialic acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3. Sialo-oligosaccharides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. Conformational aspects of sialic acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8. Biosynthesis of sialic acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1. General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2. Biosynthesis of CMP-sialic acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3. Transfer of sialic acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4. Enzymatic modification of sialic acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4.1.Biosynthesis and functions of N-glycolylneuraminic acid . . . . . . . . . . . . . . 8.4.2.Biosynthesis and functions of 0-acetylated sialic acids . . . . . . . 8.4.3.Biosynthesis of 9-0-lactylated sialic acid . . . . . . . .............. 8.4.4.Biosynthesis of 8-0-methylated sialic acids . . . . . . . . . . . . . . . . . . . . . . 8.4.5.Biosynthesis of 5-N-acetyl-2-deoxy-2,3-didehydr o-neuraminic acid . . . . . . . . 8.4.6.Biosynthesis of 5-N-acetyl-2,7-anhydr o-neuraminic acid . . . . . . . . . . . . . . 8.4.7.Occurrence of 2-keto-3-deoxynononic acid (Kdn) . . . . . . . . . . . . . . . . . . .

'

227 227 227 228 229 230 230 230 231 231 231 232 233 234 234

243 243 244 244 251 262 264 264 268 269 269 270 275

281 282 289 290 299 304 309 311 311 311 314 320 320 324 328 328 328 328 329

xx 9 . Catabolism of sialic acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 . I . Sialate-O-acetylesterases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2. Sialidases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.1, Types of sialidases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.2. Primary structures of sialidases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.3. Trans-sialidases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.4. Pathophysiological significance of sialidases and trans-sialidases . . . . . . . . . 9.2.4.1. Eukaryotic (trans-)sialidases . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.4.2. Bacterial sialidases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.4.3. Viral sialidases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.5. Sialidase and trans-sialidase inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . 9.3. Sialate-pyruvate lyase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.4. Sialic acid permease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10. Physiological and pathobiochemical significance of sialic acids . . . . . . . . . . . . . . . . . . 10.1. General physico-chemical effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2. Sialic acids masking biological recognition sites . . . . . . . . . . . . . . . . . . . . . . . . 10.3. Sialic acids representing biological recognition sites . . . . . . . . . . . . . . . . . . . . . 10.3.1.Sialic acid receptors of microorganisms, plants and lower animals . . . . . . . . 10.3.2. Sialic acid receptors of vertebrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.4. Do sialic acids have “specific” functions? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.5. Medical significance of sialic acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

329 329 331 331 333 337 339 339 342 344 344 347 349 352 352 354 360 360 364 368 370 372 372

Chapter 12. Plant lectins: tools for the study of complex carbohydrates Irwin 1 Goldstein. Harry C. Winter and Ronald D. Poretz . . . . . . . . . . . . . .

403

1 . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 . Isolation. purification and characterization of plant lectins . . . . . . . . . . . . . . . . . . . . . . 3. Structure and carbohydrate-binding specificity of lectins . . . . . . . . . . . . . . . . . . . . . . . 3. I . Mannoseiglucose-binding lectins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 .1. Concanavalin A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.2. Pea lectin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.3. Lentil lectin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. I .4. Favin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Monocotyledonous mannose-binding lectins . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. N-acetyl-D-glucosamine-bindinglectins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1. Wheat germ agglutinin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.2. Tomato lectin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.3. Potato lectin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.4. Cytisus sessilijiolius . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.5. Datura stramonium lectin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.6. Griffonia (Bandeiraea) simplicifolia I1 lectin . . . . . . . . . . . . . . . . . . . . . . 3.4. N-acetylgalactosamine/galactose-bindinglectins . . . . . . . . . . . . . . . . . . . . . . . . 3.4.1. Dolichos bij4orus lectin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.2. Lima bean lectin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.3. Soybean agglutinin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.4. Erythrina lectins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.5. Peanut lectin . . . . . . . . . . . . . . ........................ 3.4.6. Maclura pomifera lectin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.7. Winged bean lectin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

403 403 404 411 413 414 414 414 415 416 417 417 418 418 418 418 419 419 419 420 420 421 421 421

XXI

3.4.8. Jack fruit lectin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.9. Castor bean lectins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.10. Griffonia (Bandeiraea) simplicifolia I lectin . . . . . . . . . . . . . . . . . . . . . . 3.5. L-Fucose-binding lectins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5.1, Asparagus pea lectin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5.2. Ulex europaeus lectins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5.3. Trichosanthes japonica lectin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5.4. Griffonia simplicifolia IV lectin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6. Lectins with complex binding sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6.1. Red kidney bean lectin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.7. Sialic-acid-binding lectins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.7.1, Sambucus nigra I lectin . . . . . . . . . . . . . . . . . . . . . . . . . . 3.7.2. Maackia amurensis leukagglutinin . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.7.3. Trichosanthes japonica lectin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 . Lectin-based approaches for the isolation and characterization of glycoconjugates . . . . . . . 4 . I . Lectin-based reagents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Soluble glycoconjugates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. 1 . Lectin precipitation analysis of glycans . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.2. Lectin affinity adsorptiodchromatography . . . . . . . . . . . . . . . . . . . . . . . 4.2.3. Lectin-based blot analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.4. Lectin affinity electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.5. Integration of lectin-based methodologies for soluble glycoconjugates . . . . . . 4.3, Cell-bound glycoconjugates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.1. Agglutination analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.2. Lectin-coated magnetic beads and flasks . . . . . . . . . . . . . . . . 4.3.3. Fluorescence activated lectin-based flow cytometry and cell sorting . . . . . . . . 4.3.4. Lectin histochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . ...........................................

422 422 422 423 423 423 424 424 424 424 425 425 426 426 426 427 427 429 431 434 435 436 438 438 442 443 444 445 455 470

Chapter 13. Microbial lectins and their glycoprotein receptors Nathan Sharon and Halina Lis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

475

1 . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Viruses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Sialic-acid-specific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1. Influenza virus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Other specificities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Bacteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Mannose-specific (type I fimbriae) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 . I .I . Enterobacteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Sialic-acid-specific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1. Escherichia coli . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2. Streptococcus suis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.3. Streptococcus sanguis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.4. Helicobacter pylori . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.5. Mycoplasma pneumoniae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Gal and GalNAc-specific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1. Escherichia coli . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.2. Pseudomonas aeruginosa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

475 476 476 476 480 480 480 480 483 483 484 485 485 485 486 486 486

xxii

3.3.3. Actinomyces species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.4. Rhizobia . . . . . . . . . . . . . . ......... ..... 3.3.5. Myxobacteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Fucose-specific . . . . . . . . . . . . . . . ........................ 3.4.1. Helicobacter pylori . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.2. Pseudomonas aeruginosa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.3, Vibrio cholerae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.4. Others . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. Multiple specificities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. I . Bordetella pertussis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 . Fungi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 . Protozoa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1, Gal and GalNAc-specific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. I . Entamoeba histolytica . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Sialic-acid-specific . . . . . . . . ............................... 5.2.1. Plasmodium faleiparum ............................... 5.3. N-Acetylglucosamine- and chitooligosaccharide-specific . . . . . . . . . . . . . . . . . . . 5.3.1. Entamoeba histolytica . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.2. Plasmodium falciparum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4. Mannose-6-phosphate-specific . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4.1. Giardia lamblia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Biological roles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1. Infection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1.1. Viruses . . . . . . . . . . . . . . . . . . . . . . ................... 6.1.2. Bacteria . . . . . . . . . . . . . . . . . . . . . . ................... 6.1.3. Fungi . . . . . . . . . . . . . . . . . . . . . . . ................... 6.1.4. Protozoa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2. Non-opsonic phagocytosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

487 488 488 488 488 489 489 489 489 489 492 492 492 492 493 493 494 494 495 495 495 495 495 495 496 498 499 500 501

Chapter 14. Adhesive glycoproteins and receptors R . Colin Hughes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

507

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 . Cadherin family of adhesive glycoproteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Classical cadherins . . . ............................ 2.2. Desmosomal glycoproteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Immunoglobulin superfamily of adhesive glycoproteins . . . . . . . . . . . . . . . . . . . . . . . 3.1. Nerve cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. 1. N-CAM: the prototype adhesion molecule of the immunoglobulin superfamily . 3.1.2. L1 glycoprotein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.3. Myelin glycoproteins MAG and PO . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.4. The L2/HNK-I carbohydrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Non-neuronal cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. I . ICAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2. VCAM . . . . . .................................... 3.2.3. PECAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 . Matrix glycoproteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Fibronectin . . . . . . . .................................... 4.2. Laminin . . . . . . . . . .................................... 4.3. Nidogedentactin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

507 510 510 514 517 517 517 52 1 523 524 525 525 527 528 528 529 535 54 1

xxiii

4.4. Tenascin . . . . . . . . . . . . . . . . . . . 4.5. SPARC/osteonectin . . . . . . . . . . . .

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

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

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

5. Integrins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1, Structure and specificity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Biosynthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. Roles of carbohydrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Chapter 15. Carbohydrate differentiation antigens Ii. SSEA- 1 (Le") and related structures. Prototype mammalian carbohydrate antigens that serve as ligands in molecular recognition Ten Feizi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 . The I and i antigens and their sialyl forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Biochemical nature of Ii antigens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Anti4 and -i antibodies as reagents in studies of cell differentiation . . . . . . . . . . . . 2.4. Anti-I and -i antibodies as immunosequencing reagents for glycoprotein oligosaccharides 2.5. Roles of Ii and related sequences as ligands . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.1. Sialyl Ii as host-cell ligands for a pathogen Mycoplasma pneumoniae . . . . . . 2.5.2. Ii-type sequences as ligands for endogenous carbohydrate-binding proteins? . . 3 . Stage-specific embryonic antigen-], SSEA-I (CDI 5/LeX/L5) and the sialyl and sulfated analogues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

........................................... 3.2. Biochemical nature of SSEA-1 and related antigens in mouse and human . . . . . . . . 3.3. LeX, sialyl-Le' and related sequences as ligands for endogenous carbohydrate-binding proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 . Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Chapter 16. Cell adhesion and recognition mechanisms in the nervous tissue Jean-Pierre Zanetta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 . Mechanisms of neuronal migration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 . Mechanisms of axonal growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 . Mechanisms of synaptogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Mechanisms of glial wrapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Conclusions . . . . . . . . . . . . . . . . ................... References . . . . . . . . . . . . . . . . . . . ...................

541 544 545 545 545 550 554 559

571 571 571 573 573 573 575 575 577 577 578 579 579 580 581 582 582 583

587 587 587 587 590 591 593 596 597

Chapter I 7. Neoglycoproteins Reiko T Lee and Yuan C. Lee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

601

List of Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 . History and definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

601 601

xxiv

.................................. 2. Preparation of neoglycoproteins . 2.1. Modification of primary ami roups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1. Reductive amination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. I .2. Amidination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.3. Acylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.4. p-Isothiocyanato-phenyl glycosides . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Modification of carboxyl groups . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1. Modification with glycosylamine derivatives . . . . . . . . . . . . . . 2.2.2. Glycamine derivatives . . . . . ........................... 2.2.3. o-Aminoalkyl glycosides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Modification of tyrosyl group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Modification of cysteinyl group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5. Conjugation of polysaccharides to proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6. Enzymatic methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.1, Use of glycosyl-transferases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.2. Use of glycosidases in transglycosylation . . . . . . . . . . . . . . . . . . . . . . . . 2.6.3. Use of transglutaminase . . . . . . . . . . . . . . . . . . . . . . . . . . 2.7. Glycoproteins of non-covalent attachment . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Synthetic glycopeptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Applications of neoglycoproteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. I . Probing carbohydrate-protein interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Use in isolation of carbohydrate-binding proteins . . . . . . . . . . . . . . . . . . . . . . . 4.3. Cytochemical markers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. Neoglycoenzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5. Biomedical applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Advantages of neoglycoproteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Other neoglycoconjugates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

60 I 602 602 605 605 605 605 606 607 608 608 608 609 610 610 612 613 61 3 613 614 614 614 614 615 615 615 617 618 618 61 8

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

621

Other volumes in the series Volume 1.

Membrane Structure (1982) J.B. Finean and R.H. Michell (Eds.)

Volume 2.

Membrane Transport (1982) S.L. Bonting and J.J.H.H.M. de Pont (Eds.)

Volume 3.

Stereochemistry (1982) C . Tamm (Ed.)

Volume 4.

Phospholipids (1982) J.N. Hawthorne and G.B. Ansell (Eds.)

Volume 5.

Prostaglandins and Related Substances ( 1983) C . Pace-Asciak and E. Granstrom (Eds.)

Volume 6.

The Chemistry of Enzyme Action (1984) M.I. Page (Ed.)

Volume 7.

Fatty Acid Metabolism and its Regulation (1984) S. Numa (Ed.)

Volume 8.

Separation Methods (1984) Z. Deyl (Ed.)

Volume 9.

Bioenergetics (1985) L. Ernster (Ed.)

Volume 10.

Glycolipids (1985) H. Wiegandt (Ed.)

Volume 1 la. Modern Physical Methods in Biochemistry, Part A (1985) A. Neuberger and L.L.M. van Deenen (Eds.) Volume 1 lb. Modern Physical Methods in Biochemistry, Part B (1988) A. Neuberger and L.L.M. van Deenen (Eds.) Volume 12.

Sterols and Bile Acids (1985) H. Danielsson and J. Sjovall (Eds.)

Volume 13.

Blood Coagulation (1986) R.F.A. Zwaal and H.C. Hemker (Eds.)

Volume 14.

Plasma Lipoproteins (1987) A.M. Gotto Jr. (Ed.)

Volume 16.

Hydrolytic Enzymes (1987) A. Neuberger and K. Brocklehurst (Eds.)

Volume 17.

Molecular Genetics of Immunoglobulin ( 1987) F. Calabi and M.S. Neuberger (Eds.) xxv

xxvi

Volume 18a. Hormones and Their Actions, Part I (1988) B.A. Cooke, R.J.B. King and H.J. van der Molen (Eds.) Volume 18b. Hormones and Their Actions, Part 2 - Specific Action of Protein Hormones (1988) B.A. Cooke, R.J.B. King and H.J. van der Molen (Eds.) Volume 19. Biosynthesis of Tetrapyrroles (1991) P.M. Jordan (Ed.) Volume 20. Biochemistry of Lipids, Lipoproteins and Membranes (1991) D.E. Vance and J. Vance (Eds.)

Molecular Aspects of Transport Proteins (1992) J.J. de Pont (Ed.) Volume 22. Membrane Biogenesis and Protein Targeting (1992) W. Neupert and R. Lill (Eds.) Volume 23. Molecular Mechanisms in Bioenergetics (1992) L. Ernster (Ed.) Volume 24. Neurotransmitter Receptors ( 1993) F. Hucho (Ed.) Volume 25. Protein Lipid Interactions (1993) A. Watts (Ed.) Volume 26. The Biochemistry of Archaea ( I 993) M. Kates, D. Kushner and A. Matheson (Eds.) Volume 27. Bacterial Cell Wall ( 1994) J. Ghuysen and R. Hakenbeck (Eds.) Volume 28. Free Radical Damage and its Control (1994) C. Rice-Evans and R.H. Burdon (Eds.) Volume 29a. Glycoproteins (1995) J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.) Volume 30. Glycoproteins and Disease (1 996) J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.) Volume 2 1.

J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins I1 Elsevier Science B.V

CHAPTER I

Proteoglycans: a special class of glycoproteins Jeremiah E. S i l b e ~ - t ~Merton i ~ ? ~ , Bernfield3>4,and Robert

K~kenyesi’,~,~

’Edith Nourse Rogers Memorial Veterans Administration Medical Center; Building 70, 200 Springs Road, Bedford, MA 01730, USA Department of Rheumatology/lmmunology. Brigham and Womenk Hospital, Haruard Medical School, and Joint Program in Neonatology, 300 Longwood Aue.. Enders 9,Boston, MA 02115, USA

Abbreviations Standard abbreuiations

Other abbreviations

Gal Xyl

galactose xylose

GlcN GlcNS GlcNAc GalN GalNAc

glucosamine N-sulfonylglucosamine N-acetylglucosamine galactosamine N-acetylgalactosamine

PAPS RER ER EGF TGFO aFGF bFGF

GlcA IduA

glucuronic acid iduronic acid

3’-phosphoadenylyl, 5’-phosphosulfate rough endoplasmic reticulum endoplasmic reticulum epidermal growth factor transforming growth factor acidic fibroblast growth factor basic fibroblast growth factor

The abbreviations “PG” and “GAG” are frequently used for proteoglycan and glycosaminoglycan, respectively, but these have not as yet been accepted as standard abbreviations by the major journals or by the terminology committee of the International Union of Biochemistry. Similarly the abbreviations “HA” for hyaluronan, “CS” for chondroitin sulfate, “DS” for dermatan sulfate, “HS” for heparan sulfate and “KS” for keratan sulfate are frequently used but have not as yet been accepted as standard abbreviations.

1. Introduction Proteoglycans consist of complex carbohydrates covalently linked to protein, and can therefore be classified as glycoproteins. Nevertheless, for historical reasons and because they differ substantially from other glycoproteins, they are usually regarded as a separate group of substances. There are excellent detailed articles concerning structure, metabolism, distribution, and function of proteoglycans [l-171, so this chapter is not intended to be a comprehensive review. Rather, the aim is to provide an overall background and then describe and emphasize the glycosaminoglycan portion which is the primary characteristic defining this class of substances. The glycosaminoglycans have structures, metabolism, and actions quite distinct from those of the oligosaccharide portions of other glycoproteins. While some of these distinctions are obvious, there are others which are more subtle and have not received much attention.

Originally proteoglycans/glycosaminoglycans were called “mucopolysaccharides” in order to describe their polysaccharide structure and a viscous or mucinous (from the Latin mucus, “slimy”) characteristic. Only the structure of the polysaccharide portions of these molecules was recognized initially, with no realization that the polysaccharides were covalently linked to protein. Because of this, it is understandable that the “mucopolysaccharides” were considered to constitute a separate group of compounds with no relationship to the glycoproteins. Subsequently it became clear that the “mucopolysaccharides” with the exception of hyaluronan (hyaluronic acid) included a covalently linked protein, and were therefore a special class of glycoproteins. Since the term “mucopolysaccharide” did not reflect the presence of protein, it was considered to be misleading and consequently was replaced. Nevertheless, because of previous widespread use, “mucopolysaccharide” is still occasionally used as an all-inclusive name in referring to the polysaccharide portion alone, and sometimes to the entire polysaccharide-protein. The term is not rigorously defined, and is only rarely found in the more recent literature with the exception of its use for the “mucopolysaccharidoses”, a group of disorders in glycosaminoglycan degradation. The term “glycosaminoglycan” was introduced to be more descriptive of the polysaccharide structure and to clarify some confusion over nomenclature. Thus “glycosaminoglycan” is descriptive of a repeating polysaccharide (glycan) structure containing hexosamines (glycosamino) and does not include the protein portion of the entire molecule. The trivial names for the naturally occurring glycosaminoglycans are, hyaluronan (hyaluronic acid), chondroitin 4-sulfate (formerly called chondroitin sulfate A), chondroitin 6-sulfate (formerly called chondroitin sulfate C), dermatan’ sulfate (formerly called chondroitin sulfate B and earlier called b-heparin), heparin, heparan sulfate (formerly called heparitin sulfate), keratan (poly-N-acetyllactosamine), and keratan sulfate. The non-sulfated analogs, chondroitin, dermatan, and heparan, are generally not found in biological materials except at early stages during biosynthesis or when cultured cells are grown in the presence of high concentrations of chlorate [ 191 which prevents the formation of the sulfate donor, PAPS. Nevertheless, non-sulfated analogs have been produced by enzymatic means and by chemical desulfation, and their use has been important in examination of the intermediary metabolism of the glycosaminoglycans. Chondroitin sulfate and heparan sulfate are produced to a variable extent by most if not all vertebrate cells, while heparin, dermatan sulfate, and keratadkeratan sulfate are less widespread. The above named glycosaminoglycans are the only substances found in vertebrates that satisfy the structural criteria for this designation. There are also substances in lower animals and bacteria that fulfill the structural requirements for being a glycosaminoglycan. These compounds may be distinct from the vertebrate compounds by virtue of different function, location, and so on. The term “proteoglycan” was introduced to describe an entire molecule consisting of one or more glycosaminoglycan chains attached to a core protein by means of a linkage oligosaccharide. For example the general term for an undefined proteoglycan

’

It has been suggested by J.E. Scott [I81 that the term “dermochondran or dermochondan” be used instead of dermatan, since dermatan sulfate is derived from chondroitin sulfate and always contains chondroitin-like regions with glucuronic acid rather than iduronic acid.

3

containing a specific glycosaminoglycan would be described as “chondroitin sulfate proteoglycan” or “proteochondroitin sulfate” when the entire molecule is intact, while the term “chondroitin sulfate” would be reserved for the glycosaminoglycan portion alone or the glycosaminoglycan plus the characteristic linkage oligosaccharide that glycosidically attaches the glycosaminoglycan to the core protein. In addition to the glycosaminoglycan chains, some of the proteoglycans contain oligosaccharide structures identical to those of the general class of glycoproteins. The term proteoglycan is used, independent of the presence of these “glycoprotein” portions. All of the glycosaminoglycans with the exception of hyaluronan are synthesized on a core protein and usually appear in the tissues as intact proteoglycans. However, free glycosaminoglycan chains also appear as products of proteoglycan processing or degradation. Since hyaluronan is synthesized without a core protein and appears only as the glycosaminoglycan, the term proteoglycan should not be applied to this substance. During the last few years the core proteins of many classes of proteoglycans have been characterized through the techniques of molecular biology, so that families with similar core protein structure have now been identified. New terms’ for these families of proteoglycans have been and are still being introduced by those investigators who initially describe the core proteins. These names have generally reflected their structure, or the tissue in which they were initially found, or what is presumed to be a major characteristic or function. In addition there are other proteins which can be considered as “part-time” proteoglycans since they may be found either with or without glycosaminoglycan chains.

2. Structure Proteoglycans are defined by their structure as a type of glycoprotein that has covalently linked large polysaccharide (glycosaminoglycan) chains composed of repeating identical or similar disaccharides. The glycosaminoglycans are all glycosidically attached at their reducing end to specific linkage oligosaccharides which in turn are attached covalently to the core protein. There is no characteristic protein structure that defines the proteoglycan core proteins as a class distinct from that of the protein portion of the general class of glycoproteins. The proteins are widely varied in structure and are only defined as core proteins for proteoglycans because they become substituted with the glycosaminoglycans. Thus only the presence of a glycosaminoglycan moiety provides the separate classification for this special type of glycoprotein. Structures of the various core proteins will be discussed in section 5 together with distribution and functions of the proteoglycans.

2.1. Glycosaminoglycans The glycosaminoglycans are distinguished from the oligosaccharide structures found in all other classes of glycoproteins (with the exception of those that contain poly-

*

Terms for proteoglycan families where the core proteins have been sequenced have become widely used and accepted. However it has been suggested [18] that these terms might be reserved to refer to the core protein alone and not the entire proteoglycan, since there may be considerable variation in the glycosaminoglycans attached to the same protein. For example, the term “serglycin” could be used to describe the core protein without glycosaminoglycans, or after removal of glycosaminoglycans, while “serglycin, H” would be used to refer to heparin-containing serglycin and “serglycin, CS’would be used for serglycin with chondroitin sulfate.

4

DERMATAN SULFATE (Chondroitin Sulfate B)

HYALURONIC ACID

[\+gj--p

r ,

7

FH,OH

HNCOCH,

n

/3 1,3Linkage

CHONDROlTlN 6-SULFATE (Chondroitin Sulfate C)

r

GOOH

0

CH20SqH

HNCOCH3

_In

KERATAN SULFATE

1

OH OH

L

OH

L

HNCOCH,

-

H NCOCH:,

J I I

Pi,4 Linkage

n

Fig. 1. Alternating sugars of hyaluronic acid, chondroitin/dermatan sulfate and keratan sulfate glycosaminoglycans.

N-acetyllactosamine regions) by virtue of the glycan heteropolymer consisting of hexosamine residues (either GlcN or GalN) alternating with another sugar (Gal, GlcA, or IduA) (Fig. 1). The GlcN is either N-sulfated (GlcNS) or N-acetylated (GlcNAc) with rare if any residues containing free amino groups, while the GalN is always N-acetylated (GalNAc). There may be 0-linked sulfate substituents on the hexosamine and/or the uronic acid or Gal. Because the hexosamines alternate with the other sugars, it is convenient to think of the glycosaminoglycans as polymers consisting of repeating identical or similar disaccharide units. Although an individual chain may contain both GlcA and IduA, no glycosaminoglycan chain has been found to contain both GlcN and GalN or Gal and uronic acid as part of the repeating portions of the same polymer. All the glycosaminoglycans are anionic by virtue of their uronic acids and especially their sulfate groups, with heparin being the most anionic organic substance found in living tissues. This extensive sulfation is unique to the glycosaminoglycans so that the high degree of negative charge is an important feature distinguishing proteoglycans from other glycoproteins. Glycosaminoglycan chains on proteoglycans range in size from -15 disaccharide units to several hundred, while hyaluronan may be as much as several thousand disaccharides in length. The chains of all the glycosaminoglycans are linear (unbranched) and have a non-specific termination, so that some chains may end with uronic acid and some with hexosamine [20]. Hyaluronan (hyaluronic acid) is the simplest glycosaminoglycan, since it contains identical alternating GlcA and GlcNAc saccharides throughout, fi- 1,3- and 6- 1,4-linked respectively, has no sulfate substituents, and is not linked covalently to protein. Other

5

glycosaminoglycans are sulfated (with the exception of poly-N-acetyllactosamine) and are synthesized bound to specific core proteins. Chondroitin sulfate glycosaminoglycans consist of alternating GlcA and GalNAc3 also 8-1,3- and p-1 ,Clinked respectively, with variable amounts and location of sulfation, including non-sulfated GalNAc, GalNAc 4-sulfate, GalNAc 6-sulfate, GlcA 2-sulfate, or combinations of sulfate substitutions on the same saccharides. Depending upon the animal and the tissue source, chondroitin sulfates exhibit a wide range of sulfation and differ in the amounts of 6-sulfate and 4-sulfate in the same glycosaminoglycan chain. In general, the sulfation on a single chain is mostly or entirely 6-sulfate or 4-sulfate, so that chondroitin sulfate is not ordinarily found with equal amounts of both types of sulfate in the same chain. Usually a single GalNAc residue will have only one sulfate, either 4 or 6, but disulfated 4,6 GalNAc residues are found. In addition, 2-sulfated GlcA is found alternating with sulfated or non-sulfated GalNAc. Dermatan sulfates are more complex than the chondroitin sulfates. The term defines a glycosaminoglycan similar to chondroitin sulfate but containing va$is amounts of IduA rather than having GlcA as the only uronic acid. This is an important concept since the formation of dermatan residues proceeds from chondroitin residues as precursors when the appropriate uronosyl epimerase is present to convert GlcA to IduA. This results in modified glycosaminoglycans with the same protein core as the precursor proteochondroitin sulfate. The IduA residues are ordinarily only found adjacent to 4-sulfated GalNAc and not adjacent to 6-sulfated or non-sulfated GalNAc, although non-sulfated or 6-sulfated GalNAc residues may be found adjacent to GlcA in the same glycosaminoglycan chain. In addition, dermatan sulfate frequently has IduA 2-sulfate in fairly high amounts. Heparin and heparan sulfate glycosaminoglycans are even more complex. These contain two main types of disaccharide residues, although the proportions differ greatly between heparin and heparan sulfate (Fig. 2). Thus heparan sulfate contains areas of sulfated IduA-GlcNS repeating disaccharides, and areas of non-sulfated GlcA-GlcNAc repeating disaccharides [22]. The sugars are linked @-1,4-(or a-1,4- if the uronic acid is I ~ u A and ) ~ a-1,Clinked respectively (in contrast to the 8-1,3- and B-1,4-linkages of hyaluronan and chondroitin. In addition to the N-sulfation of the GlcN, there are variable amounts of 6-0-sulfate on the GlcNS and 2-0-sulfate on the IduA. Heparan sulfate chains usually contain more non-sulfated disaccharides than sulfated disaccharides, resulting in an overall degree of sulfation that averages less than one sulfate per repeating disaccharide unit. Heparin differs from heparan sulfate by its higher content of IduA, N-sulfate and 0-sulfate with most of the disaccharide units containing 2-sulfated IduA alternating with GlcNS which is partially 6-0-sulfated as well. The overall sulfation averages two to Although the conventional description of the disaccharide repeating units is GlcA and GalNAc, (3-1,3and p-l,4-linked, respectively, chondroitin is actually a modified polylactose structure so that it has been suggested [21] that it would be better to describe the disaccharide repeating unit as GalNAc and GlcA, 8-1,4and p-I ,3-linked, respectively. The IduA is defined as L- rather than D- since the epimerization of GlcA occurs at the 5’ position. The linkage of the lduA in dermatan is therefore actually L-IduA, a-i,3-linked rather than the precursor D-GlcA, b-1,3-linked; in heparin and heparan sulfate it is L-IduA, a-I,4-linked rather than D-GlcA, p-1,4-linked. However, it should be understood that only the terminology has changed while the linkages remain the same.

6

HEPARAN (Heparitin) SULFATE

r

1

HEPARIN

r

1 0-

osofl

n 011,4 Linkage Fig. 2. Typical oligosaccharide structures of heparan sulfate and heparin.

two and a half sulfates per disaccharide. In addition to 6-0-sulfation of GlcNS there is also 3-0-sulfation of occasional GlcNS on a fraction of the glycosaminoglycan chains (as well as on occasional GlcNS of some heparan sulfate). This is highly important because the 3-0-sulfate substituent is necessary for the anti-coagulant activity of heparin and heparan sulfate [23]. The similarity of the heparin and heparan sulfate structures has caused some confusion, so that heparan sulfate has occasionally been described incorrectly as “cell surface heparin”, etc. However, as proteoglycans they are not similar, since heparin is only found attached to the specific intracellular serglycin protein or as the free glycosaminoglycan chain split from this proteoglycan, while heparan sulfate is found on several families of extracellular and cell surface core proteins unrelated to serglycin. Keratan sulfates are polymers composed of alternating Gal and GlcNAc residues p- 1,4and @-1,3-linked,respectively’. The degree of sulfation may vary along the keratan chains, with the first one or two reducing end disaccharides being non-sulfated, the next few monosulfated, while the remainder of the chain consists of disulfated disaccharides [24, 251.

2.2. Linkage region

2.2.1. Chondroitin sulfate, dermatan sulfate, heparan sulfate, heparin These glycosaminoglycans are all attached to a core protein at their reducing end through a tetrasaccharide region consisting of GlcA-Gal-Gal-Xyl[3,5,7] (Fig. 3) with As with chondroitin, keratan is actually a modified polylactose.

7

z

W

Glucuronic Acid

Galactose

Galactose

Xylose

Serine

I-

g0 I

NH 0-CHz-C

’0

I I

c=o OH

OH

OH

I E

W

I-

$ a Fig. 3. Structure of oligosaccharide linkage region.

the Xyl glycosidically linked to the hydroxyl of Ser, usually adjacent to a Gly in the protein core. The first sugar of the glycosaminoglycan chain (linked to the Gal) always is GlcA, but is considered to be a part of the linkage oligosaccharide, since its addition to the linkage region appears to be catalyzed by a specific enzyme different from the enzyme that is involved in the incorporation of GlcA into the rest of the glycosaminoglycan [26]. The Xyl may be phosphorylated [27], and one or both Gal residues may be sulfated [28]. In addition to the glycosaminoglycan substituents on proteoglycans, there may be 0-linked and N-linked oligosaccharides similar or identical to those of various species of glycoproteins.

2.2.2. Keratan sulfate Skeletal keratan sulfate and corneal keratan sulfate chains are attached to core protein through 0-linked oligosaccharides [29] and N-linked oligosaccharides [30,3 11 respectively, identical to 0-linked and N-linked oligosaccharides of the general class of glycoproteins.

3. Biosynthesis Cell-free translation of chondrocyte mRNA from a variety of sources [32-341 has been reported. More recently substantial details concerning the molecular size and the amino acid sequence for core proteins from a great number of different sources (see section 4) have been determined by use of conventional cloning techniques. In a few instances recombinant core proteins such as decorin [35], and syndecans [36-381 have been examined as receptors for the glycosylation reactions. Biosynthesis of the core protein does not appear to differ from that of other proteins except for possible post-translational modifications such as that of the large aggrecan protein core [39,40] which may proceed through an even larger “pro-protein” which then is reduced in size during or after formation of the glycosaminoglycan chains. Although post-translational proteolytic trimming of core protein has also been noted for decorin [41], neurocan [42] and serglycin [43,44], little is known about details, nor is there any knowledge whether or not this is general for all core proteins.

8

Post-translational modifications such as phosphorylation of serine residues on decorin [ 151 and attachment of myristoyl and palmitoyl moieties to perlecan [45] have been described, but these modifications are not distinguishable from similar reactions occurring with the general class of glycoproteins. Moreover the core proteins of many proteoglycans contain N- and/or 0-linked oligosaccharides identical to those of the general class of glycoproteins. It can be assumed that biosynthesis of these oligosaccharides is identical to that of other glycoproteins. Compared to the general class of glycoproteins, which may also contain SerGly dipeptides, only a small number of proteins become substituted with glycosaminoglycans. The mechanism of selection of these proteins for subsequent glycosaminoglycan attachment is unclear, but may be related to some as yet undetermined common structural feature. In contrast to the apparent general nature of the protein core biosynthesis, and some post-translational modifications, formation of the glycosaminoglycans follows pathways distinctly different from those of the oligosaccharides of the general class of glycoproteins. Thus, following formation of the precursor sugar nucleotides and synthesis of the protein core in the RER, some of the next steps in biosynthesis are unique to proteoglycans. These consist of: (A) formation of a specific oligosaccharide linkage region by (1) addition of Xyl to Ser moieties of the core protein followed by addition of two Gal residues and a GlcA residue for chondroitiddermatan, heparan, heparin or (2) addition (not unique) of glycoprotein-like N-Asn, 0-Thr, or 0-Ser linked oligosaccharides for keratadkeratan sulfate, (B) formation of the alternating saccharides of the glycosaminoglycan chains and (C) sulfation and/or epimerization of the glycosaminoglycan chains. A schematic presentation is shown in Fig. 4. Of these steps, only the sugar activation, protein formation, and keratan sulfate linkage formation are shared with the formation of other glycoproteins. In the case of hyaluronan, only the activation of sugars and formation of alternating saccharides are applicable, since hyaluronan has only the glycosaminoglycan component and is not sulfated. In addition to Gal, GalNAc, and GlcNAc, which are found in the general class of glycoproteins, the proteoglycans contain, GlcNS, GlcA, IduA, and Xyl, sugars which are rare or have not been found at all in other glycoproteins. The precursor activation pathway for GlcA is by dehydrogenation of UDP-Glc to UDP-GlcA which can in turn be decarboxylated to form UDP-Xyl. The activated form of sulfate, 3’-phosphoadenylyl 5’-phosphosulfate (PAPS), is formed from sulfate and ATF’. The enzyme involved in formation of UDP-Xyl from UDP-GlcA is membrane-bound (ER or Golgi) [46], while the activation of all the other sugars and sulfate take place in the cytosol with soluble enzymes. The main reactions involved in the biosynthesis of the glycosaminoglycan portions of the proteoglycans have been defined. Formation of the GlcA-Gal-Gal-Xyl linkage appears to take place sequentially in different locations as the core protein moves along the smooth ER and Golgi membrane with identical reactions for the formation of chondroitin sulfate, heparan sulfate, and heparin. Polysaccharide chains are initiated by direct transfer of Xyl from UDP-Xyl to specific Ser moieties of the protein core, followed by transfer of each Gal residue and GlcA with no apparent dolichol or other intermediate [47]. Thus there does not appear to be any pathway (except with keratan sulfate) similar to that for the earlier steps in formation of the N-linked glycoprotein oligosaccharides. Unlike protein

9

GLUCOSE

Jp.P GlcN-6-P<

G/ufomiie

I

Fru-6-P

e G l c - 6 - P \\

1k

UDPGlc

I

GICUA

G ~ N S\-

Idh

GalNAc

\

~ici

I

SERl NE

f

Gal NAc

Fig. 4.Biosynthesis of the glycosaminoglycan portions of proteoglycans.

synthesis, the formation of the oligosaccharide linkage region and the glycosaminoglycan chains does not appear to involve a molecular template mechanism, but no mechanisms have been found to account for the structural variations, such as localization of sulfate and IduA. A mutant CHO cell line deficient in xylosyl transferase has been used to demonstrate that only one xylosyl transferase is involved in the initiation of both chondroitin sulfate and heparan sulfate glycosaminoglycans [48]. Studies with different cellular fractions and immunocytochemistry [49], have suggested that xylosyl transferase activity is greater in RER than in smooth membranes, but other information concerning the timing of incorporation in intact cells has suggested that the transfer of Xyl to Ser takes place in the smooth ER and Golgi [50,51]. The membrane-bound enzyme involved in the formation of UDP-Xyl from UDP-GlcA is inhibited by UDP-GlcA [52], and it has recently been reported that the formation and control of the amounts of UDP-Xyl plus the xylosylation of core protein take place in the same subcellular fractions [53]. This is distinctive from the cytosol localization for synthesis of all the other sugar nucleotides, which then must be transported across the Golgi membranes in order to reach their sites of action. Such transport has been examined in some detail [54].

10

Since xylosylation of Ser residues is the first step in the sequence of reactions leading to the biosynthesis of glycosaminoglycan chains, it is likely that Xyl transfer is an important mechanism of regulation. The presence of phosphate groups on some of the Xyl residues of proteoglycans suggests that control might also be related to this substituent. The transfer of the two Gal residues to b-D-xylosides by cell-free systems has been demonstrated [55,56], and it has been shown that this takes place in separate earlier Golgi compartments from that of the Gal incorporation into glycoproteins [57,58]. No information has been presented concerning the enzymes involved or the subcellular localization in any sulfation of either the first or the second Gal residue. The presence of sulfate in the linkage region does however suggest that sulfation or de-sulfation of these residues may play a role in control and trafficking. The first GlcA is transferred by a glucuronosyl transferase that appears to be distinct from that involved in the formation of the repeating disaccharide units [26,59]. The first GalNAc of chondroitin [60] and the first GlcNAc of heparan [61] also have been reported to be transferred by enzymes different from the polymerases. UDP-GlcA and UDP-GlcNAc are the precursors for hyaluronan, heparin and heparan [62,63]; UDP-GlcA and UDP-GalNAc are the precursors for chondroitin and dermatan [63]. UDP-Gal and UDP-GlcNAc are presumed to be the precursors for keratan, but synthesis has not been achieved with cell-free systems. In polymerization of chondroitiddermatan and heparinheparan, the transfer of alternating saccharide units to the non-reducing end of the growing membrane-bound nascent proteoglycan primer occurs in a highly organized fashion so that both N-acetylhexosaminyl transferase and glucuronyl transferase act rapidly in concert on individual chains [62,64,65]. Exogenously added oligosaccharides can also serve as substrates for the addition of one or a few sugars in cell-free systems, but do not serve as primers for polymerization to any great extent [66], thus supporting the concept of a membranebound nascent proteoglycan substrate. The synthesis of hyaluronan is an exception to this in that the glycosaminoglycan, not linked to protein, is polymerized by the alternating transfer of sugar units to the reducing end of the growing chain rather than to the non-reducing end [67]. The IduA found in heparin, heparan sulfate, and dermatan sulfate is formed by epimerization of GlcA after it has been incorporated into the glycosaminoglycan chain [63,65,68] and not from an IduA nucleotide. The enzymes involved in polymerization have not been purified to homogeneity but the heparin GlcA transferase and GlcNAc transferase have been purified sufficiently to indicate that the activities apparently are on the same 70 kDa protein [69]. None of the polymerases have been sequenced. Since extensive cell-free polymerization has been demonstrated with nascent proteoglycan but not with oligosaccharide acceptors, there is a strong indication that the membrane-bound glycosaminyl and GlcA transferase enzymes are adjacent to the membrane-bound nascent proteoglycans to form enzyme-substrate complexes that provide an ordered reaction of alternating sugar placement. Thus rapid polymerization occurs with no measurable amounts of product containing the addition of a single sugar. This organization is quite different from that of the oligosaccharide formation in other glycoproteins where synthesis takes place sequentially in different locations as

the dolichol oligosaccharides or the nascent glycoproteins move from the site of one membrane-bound enzyme to another. Sulfation occurs with the direct transfer of sulfate groups from PAPS to appropriate sites on the glycosaminoglycans [63,65] during or after polymerization [70]. In chondroitin sulfate synthesis, the same Golgi fractions that are involved in glycosaminoglycan polymerization contain the enzymes for incorporation of sulfate [57], and separation of sulfotransferases from the site of polymerization results in a lower efficiency of sulfation. Specific 4- and 6-0-sulfotransferases are involved for transfer to each type of location for each receptor in a rapid “all or nothing” fashion so that chains rapidly become highly sulfated or are not sulfated at all [71]. This is indicative of a high degree of organization of the sulfotransferases with the membrane-bound nascent proteoglycan substrate. It is not known whether the occurrence of sulfation at both the 4 and 6 position in the same chain or on a single GalNAc residue is random or is programmed in some as yet unknown fashion. The epimerization of GlcA to IduA in the formation of dermatan sulfate is dependent upon the 4-sulfation of the GalNAc adjacent to the epimerized GlcA [72] indicating an interdependence of the epimerase with the 4-sulfotransferase, so that undersulfation results in chondroitin disaccharide units for the unsulfated portions rather than the dermatan disaccharide units for the sulfated portions. Thus the degree of sulfation could relate to the great range seen in the amounts of IduA relative to GlcA found in dermatan sulfate. There is great heterogeneity in the amount and positions of sulfate in heparin and heparan sulfate. As described in section 1, these glycosaminoglycans contain varying amounts of N-sulfate and 0-sulfate appearing mainly on the 6 position of GlcN and the 2 position of IduA [65] and specific small amounts of sulfate at the 3 position of some GlcN which is related to the anticoagulant capability of these molecules [23,65]. The sulfation of heparin and heparan sulfate, which has been so well demonstrated by Lindahl and his colleagues, takes place in an ordered sequence with N-deacetylation and N-sulfation of GlcNAc, then with GlcA epimerization to IduA coupled with 2-0-sulfation of IduA residues, followed by 6-U-sulfation and 3-0-sulfation of GlcNS residues [65]. The N-deacetylation and N-sulfation of the GlcN residues are linked reactions [73] with both enzyme activities residing on the same protein [74]. Thus N-deacetylation does not ordinarily occur to a marked degree in the absence of N-sulfation. 0-Sulfation does not appear to take place unless the GlcN has already been N-deacetylated and N-sulfated [65]. Clearly, from this information and by analogy with chondroitiddermatan sulfate, the biosynthesis of heparidheparan sulfate also involves concerted and ordered reactions organized around the interactions of membrane-bound nascent proteoglycan and membrane-bound enzymes. Although the coordinated action of all the enzymes mentioned above appears to be essential for the biosynthesis of the glycosaminoglycan portions of the proteoglycans, details regarding the localization and sub-cellular organization of these enzymes in intact cells is rudimentary. However, it can be stated with confidence at least for chondroitin sulfate, that the enzymes involved in this assembly are bound to the medial, trans, and trans-Golgi network portions of the Golgi complex [57,75]. The membrane localization for the biosynthesis of hyaluronan is different, since synthesis of this glycosaminoglycan appears to occur at the inner surface of cell membranes rather than in the Golgi [76].

Little is known concerning details of how the glycosaminoglycans or proteoglycans are processed and moved into the extracellular space.

4. Degradation and turnover A “matrixin” family of extracellular zinc-requiring metalloproteases play a prominent role in degradation of connective tissue proteins including the core proteins of proteoglycans [77,78]. In addition there are cathepsins and trypsin-like enzymes as well as other matrix proteases involved in this degradation. Intracellular degradation also takes place in lysosomes following the transport of intact or partially degraded proteoglycans to these organelles. Proteolysis and removal of proteoglycans from tissues has been demonstrated most graphically by injection of the proteolytic enzyme, papain, into rabbit ears [79] resulting in a rapid loss of cartilage rigidity accompanied by mobilization and excretion of glycosaminoglycans in the urine. Similarly, the injection of proteolytic enzymes into intervertebral discs has been utilized in many parts of the world as a technique to replace surgical removal of disc material following herniation. Since there apparently is nothing specific to the degradation of core proteins that sets them aside from the general class of glycoproteins, discussion of the proteolytic enzymes is beyond the scope of this chapter. A major component in the degradation and turnover of the glycosaminoglycan portion of the proteoglycans in vertebrate tissues is the result of action by specific endoglycosidases which cleave within the glycosaminoglycan chains [63,80-821. This contrasts to the degradation and turnover of the oligosaccharide portions of the general class of glycoproteins which is accomplished solely by sequential exoglycosidase activity from the non-reducing ends of the oligosaccharides. Endoglycosidases capable of degrading glycosaminoglycans have also been found in bee and snake venom, and in leeches. These enzymes have varying substrate specificities and degrade the glycosaminoglycans to several different products. The first glycosaminoglycan-degrading activity to be described was initially called a “spreading factor” [83] since the injection of India ink or similar material into skin during or after the injection of the substance increased diffusibility or “spreading” of the ink. The discovery of spreading factors and their substrates was a major stimulus leading to the initial description and characterization of glycosaminoglycans, primarily in the laboratory of Karl Meyer (see ref. [84] for a history of the fundamental accomplishments of this pioneering scientist). It is likely that the spreading following the enzymatic degradation of hyaluronan and the glycosaminoglycan portions of proteoglycans plays a role in the pathogenicity and spread of some lyase-producing bacteria and in the local toxicity of the various venoms. However, the studies with enzymes from invertebrate sources have not aided substantially in understanding the mechanism of physiologic degradation or turnover of proteoglycans. The best-defined vertebrate glycosaminoglycan-degrading enzyme is the hyaluronidase that is found in testicular tissue as well as in many other mammalian tissues [63,81,82], where it may play a role in connective tissue turnover, remodelling, and growth[85]. This endoglycosidase cleaves within the glycosaminoglycan chain to result in a family of even-numbered oligosaccharides ranging in size from tetrasaccharide to larger, with

13

I

Testicular Hyaluronidose

FOOH

CHzOShH

OH

HNCOCH3

CiOOH

$H20S03H

HNCOCH,

Fig. 5 . Degradation by testicular hyaluronidase.

GlcNAc at the reducing end (Fig. 5). The enzyme has an optimum activity at acid pH, suggesting a possible lysosomal origin. Chondroitin 4-sulfate and chondroitin 6-sulfate can also be degraded (more slowly) by the testicular enzyme so that it might be more appropriate to use the term glycosaminoglycanase or glycosaminoglycan hydrolase rather than hyaluronidase. Testicular hyaluronidase will not degrade the IduA-containing portions of dermatan sulfate, although there will be some depolymerization due to the cleavage that takes place in the chain wherever there is GlcA. Thus the degree of epimerization to IduA may have considerable importance in protecting the dermatan sulfate from this enzymatic activity. The enzyme will not degrade heparin or heparan sulfate. Older studies with radioactive precursors utilized in uiuo have shown a half-life for skin hyaluronan of 2 to 5 days and for skin chondroitin sulfate of 7 to 14 days [86,87]. More recent work with cartilage indicates half-lives of approximately 7 and 50 days for large and small proteochondroitin sulfates, respectively [go]. Studies on the turnover of heparin, heparan sulfate, and dermatan sulfate indicate that degradation of the glycosaminoglycan components of these compounds in uiuo occurs in mammalian systems. There have been descriptions of an enzyme from platelets [88] and other cells [89,90] which is capable of cleaving some linkages in heparin and heparan sulfate resulting in shorter glycosaminoglycans or large oligosaccharides. This enzyme is an endouronidase leaving uronic acid at the reducing end of the shortened products, and probably accounts for the appearance of heparin glycosaminoglycan fragments [9 11 in vertebrate tissues. However, extensive degradation of heparin and heparan sulfate by endoglycosidases in a manner similar to that of hyaluronan and chondroitin sulfate has not been described. Endocytosis of certain matrix proteoglycans and transport to lysosomes has been reviewed

14

1

Chondroitinase

T{20sh.

t~o~cH20so~~oH

HO

OH

HNCOCH3

Di -6s

H

HNCOCH,

ADi-6S

Fig. 6 . Degradation by bacterial glycosaminoglycan lyases

[ 151, as has the intracellular partial degradation and transport of certain proteoglycans to lysosomes [92]. Glycosaminoglycans and the oligosaccharide products of endoglycosidase action are acted upon by lysosomal or other exoglycosidases and sulfatases found in many tissues. These only act at the non-reducing end of an intact glycosaminoglycan or oligosaccharide. Thus it has been shown that lysosomal enzymes degrade an oligosaccharide or even a small entire glycosaminoglycan chain by sequentially removing sulfate and sugars from the non-reducing end. This is not significantly different from the degradation of the oligosaccharide components of the glycoproteins. Extensive investigations have indicated that exoglycosidases and exosulfatases provide the main mechanism for final degradation of heparin, heparan sulfate, and keratan sulfate, and that deficiencies of these various enzymes are the defects in the mucopolysaccharidoses group of inborn metabolic errors [93-951. Further discussion of the steps in degradation with endocytosis and intracellular movement of glycosaminoglycans and glycosaminoglycanderived oligosaccharides to lysosomes is beyond the scope of this chapter. Bacterial endoenzymes from a variety of sources have been used extensively for degradation of glycosaminoglycans. Their substrate specificities and products have been well described [96], originally in the laboratories of A. Linker and S. Suzuki, and their usage has been instrumental for identification and characterization of proteoglycans. These enzymes all work as eliminases (lyases) rather than glycosidases (hydrolases), producing disaccharides with hexosamine at the reducing end and uronic acid with a C4-C5 double bond at the non-reducing end (Fig. 6).

15

5. Classification, distribution, and function of proteoglycan Proteoglycans can best be classified by their occurrence in (A) extracellular matrix, (B) on cell surface, or (C) intracellularly, since their structures and functions are specifically related to these sites. In addition there are a number of “part-time” proteoglycans that are found in various locations. During the last few years a large number of core protein cDNAs have been isolated and sequenced [97,98], helping to determine and distinguish a great variety of primary structures of these compounds. Core protein sequences from evolutionarily distant animals have been shown to contain some structural features that have been preserved through millions of years of evolution indicating that the core proteins contain functionally important structures in a variety of organisms. The available sequence information allows the grouping of almost all those core proteins so far examined into gene families based on similarity of structures which have features that have evolved to carry covalently bound glycosaminoglycan chains. The core proteins may have functions in addition to the bearing of glycosaminoglycans, but essentially all of the molecules carry these chains. Each family appears to be specific for its tissue localization. Details concerning the structures of the families of core proteins are beyond the scope of this chapter, and good reviews have appeared during the last few years describing the various types of core proteins [3-5,10,13,14,98] with excellent diagrammatic representations. Size of the core proteins range from as small as 10 kDa to as large as 400 kDa, and overall size of the proteoglycans from as small as 80 kDa to as large as 3500kDa. The functions of most of the proteoglycans have not been precisely defined, but there are strong presumptions of function which derive from the structures and the known interactions of these compounds with other substances. Often, but not always, the glycosaminoglycan portions of the proteoglycans appear to be the “business ends” which provide the main functional aspects, while the core proteins direct the intracellular and extracellular trafficking and placement during synthesis of the glycosaminoglycan and positioning in the appropriate location with orientation for function. The glycosaminoglycans may act as receptors or as recognition sites for active agents or may be the directly active agents. In some cases this is due to the highly charged nature of these glycosaminoglycans, while in other cases it appears to be due to the specific order of substituents on the glycosaminoglycans. Thus differences in placement of sulfate, uronic acid epimerization, and/or N-deacetylation-sulfation have the potential to provide myriads of specific structures enabling highly specific interactions or modifications of the actions of other substances. The potential variations in structure are much more numerous than those of the shorter oligosaccharides with non-sulfated sugars found in the general group of glycoproteins. See section 6 for examples of function related to the specific fine structure of the glycosaminoglycans. There are a number of proteins that may bear glycosaminoglycan chains when isolated, but these proteins do not appear to be members of any of the proteoglycan gene families because their homologs do not bear glycosaminoglycans and thus their structures do not appear to have evolved as those of other proteoglycans. Moreover, a proportion of most of these proteins can also be isolated without any bound glycosaminoglycans, and thus

16

may be considered as “part-time” proteoglycans. They do not resemble one another in any consistent manner regarding their structure as proteoglycans, their locations, or their presumptive functions .

5.I . Matrix proteoglycans These proteoglycans are secreted from the cells after completion of biosynthesis, and are not found attached to cell surfaces. 5.I . 1. Aggrecanhersican family As well as aggrecan and versican, this family includes neurocan and brevican. These are large proteoglycans with typical core protein sizes of 220 kDa [99], 265 kDa [ 1001, 139kDa [42], and 99 kDa [loll, respectively, as deduced from DNA sequences. Aggrecan is the major proteoglycan of cartilage, and has received more attention than any other core protein or proteoglycan [102]. It is the largest of all proteoglycans, typically containing as many as 100 chondroitin sulfate chains of 2&60kDa, about 30 shorter keratan sulfate chains, and a small number of N-linked oligosaccharides, all attached in a “bottle brush” configuration. The entire core protein primary structures from several sources have been established [99,103-1051. The chondroitin sulfate chains are concentrated in a central polypeptide region of the core protein where there may be more than 100 SerGly dipeptides for attachment. The lesser number of keratan sulfate chains are concentrated on a region immediately N-terminal to the chondroitin sulfate attachment sites, and the smaller number of glycoprotein-like N-linked oligosaccharides are scattered on less well-defined areas of the core protein. There are globular N-terminal GI and G2 domains adjacent to the keratan sulfate-binding region, capable of binding non-covalently to hyaluronan allowing as many as 100 proteoglycan molecules to form an aggregate on a single hyaluronan molecule constituting a total molecular mass as high as 100,00CL200,000kDa. The term “aggrecan” was applied because of this well-defined aggregate formation. The binding is stabilized by a specific glycoprotein (similar in sequence to the N-terminal domain of aggrecan) which interacts with the hyaluronan and the binding region of the core protein. The aggregate can be dissociated to proteoglycan monomers by high salt concentration and will reaggregate when salt concentrations are lowered. The C-terminal end of aggrecan contains conserved lectin-like sequences, and alternatively spliced epidermal-growth-factor-like sequences [ 1061, and complement-regulatory-protein-like sequences [ 1031. Aggrecan is made by chondrocytes and is mainly found as the aggregate in cartilagineous tissues where it is immobilized in the extracellular matrix by the type I1 collagen meshwork. It is generally accepted that these largest of all proteoglycans function as a cushion to external pressure. The gigantic highly polyanionic structure confines a large “domain” within its boundaries so that a volume of water equal to as much as 1000 times the volume of the proteoglycan itself can be contained within the external limits of the molecule. In this domain, small molecules move freely, but large molecules such as proteins are excluded by the nature of the highly charged glycosaminoglycan chains. Under pressure, there is an efflux of water with a concomitant slow decrease in the hydrated volume of the chondroitin sulfate chains, while release of the pressure results in a

17

gradual rehydration of the glycosaminoglycan with a reconstitution of the original volume. Thus, this large form of proteoglycan provides an elastic cushion against mechanical stress on the cartilage. The highly anionic nature of aggrecan can also serve as a barrier or filter of charged molecules. Aggrecan is a prominent example of the “business end” functions of glycosaminoglycans. Versican (so named because of its versatile, complex structure) [90] is another prominent member of this family [107,108]. It was first found in skin, but is a major component in most extracellular matrices. Versican core proteins are larger than those of aggrecan but there are many fewer central region SerGly dipeptides for attachment of glycosaminoglycans. As a result versican may have as few as 12-15 chondroitin chains leading to an overall size considerably smaller than that of aggrecan. There is little or no attachment of keratan sulfate to versican, and there is only a single globular N-terminal domain, with a sequence similar to link protein. The C-terminal globular domain is similar to that of aggrecan [loo]. Although aggregation with hyaluronan has been described, it does not appear to be mediated by the same link protein mechanism as that of aggrecan [ 1071. Since versican is capable of aggregate formation with hyaluronan [ 1091, these proteoglycans may function somewhat in a capacity similar to aggrecan. However with many fewer glycosaminoglycan chains, these proteoglycans are not capable of holding water to the same degree. Versican appears to function in ion filtration, hydration of the extracellular matrix of the central nervous system [ 1101 and the multiple glycosaminoglycans are apparently involved with cell-matrix interactions. Neurocan (so named because it was found in nerve tissue) has N- and C-terminal domains similar to those of aggrecan and versican [42], but the sequences situated between these domains are shorter. Neurocan from the early postnatal brain is able to aggregate with hyaluronan via a link protein-assisted mechanism [ 1 113. Based on its restricted localization, neurocan is thought to play a role in delineating pathways for migrating axons in the central nervous system. Another protein isolated from brain has similarities to neurocan. It has been found to be substituted with chondroitin sulfate and also without chondroitin sulfate, so it can be considered to be a “part-time” proteoglycan. It has been called brevican[lOl] because of its shorter structure. The function of the protein or proteoglycan has not been addressed to any degree. 5.1.2. Leucine-rich core protein family The proteoglycans in this family, which includes decorin, biglycan, lumican, and fibromodulin, are major components of the interstitial matrix produced by fibroblasts and other cells. The core proteins are small (3745kDa) and have several leucine-rich motifs [ 1 121 with similarity to the LH-CG receptor, thyrotropin receptor, and Drosophilu proteins chaoptin and toll. Core proteins of this family characteristically undergo proteolytic processing following synthesis, with removal of an additional small peptide from the N-terminus. Decorin (formerly designated as “PG-11” or “PG-40”) is a ubiquitous connective tissue matrix proteoglycan. Its name derives from its characteristic binding to collagen where it “decorates” the collagen surfaces. The protein core is approximately 38 kDa, to which are attached one to three N-linked and/or 0-linked oligosaccharides, and a single dermatan

18

sulfate glycosaminoglycan chain as large as 80 kDa [ 1 131 attached at Ser-4 [ 1 141. It has been demonstrated that the non-covalent binding of decorin to collagen fibrils inhibits collagen fibrillogenesis in vitro [ 1151. Subsequently it has been shown that a decorin core protein will bind to a specific region on the surface of a type I collagen fibril while the single dermatan sulfate chain will interact with the dermatan sulfate chain of another decorin molecule bound to a neighboring collagen fibril [ 1 161. The interaction between these two dermatan sulfate chains is thought to provide adherence between collagen fibrils and to regulate the spacing between them, thus affecting connective tissue integrity in skin and other tissues. Decorin core proteins also bind transforming growth factor (3 (TGF(3)[ 117,1181 and may regulate the effective concentration of TGFP available to the signal-transducing receptors. In this regard, the administration of recombinant decorin to rats was found to alleviate the excessive extracellular matrix production caused by elevated levels of TGF(3 in a model of glomerulonephritis [I 191. It is of note that it is the core protein and not the glycosaminoglycan chains that interact with the TGFP. Biglycan (formerly called “PG-I”) is another ubiquitous component of connective tissue matrix [ 1201, and was so named because it has two chondroitin sulfate or dermatan sulfate chains. It binds to TGFP and to other proteins through the core protein rather than the glycosaminoglycan chains [ 1 181, but this binding has not been examined to the same degree as that of decorin. The role of IduA in the function of biglycan is also not known. Fibromodulin (so named because the proteoglycan affects collagen fibrillogenesis [7,121]) is a keratan sulfate-containing proteoglycan that has been found in cartilage, tendon, and sclera[l4]. It has been reported to attach to TGFP by means of its core protein [ 1 181. Lumican was so named because it was found in cornea where it appeared to be related to transparency [122]. In particular, the keratan sulfate structure is considered to be important to the function, since opacities in corneal macular dystrophies were found to correlate with a deficiency in keratan sulfate or a lack of sulfation of keratan [123]. Furthermore, corneal transparency in developing chick embryos was shown to correlate with the sulfation of keratan [ 1241. Collagen fibrillogenesis was shown to be inhibited by lumican core protein [125]. Lumican from aorta has been reported to contain only non-sulfated keratan [3 I].

5.I . 3. Perlecan family Perlecan is so named because rotary shadowing micrographs of the purified proteoglycan obtained from basement membrane resemble a string of pearls. It is the largest core protein (400kDa) as yet described, and is composed of five domains. Domain I has a unique sequence, but domains 11, 111, IV, and V are similar to the low density lipoprotein (LDL) receptor, short arm of the laminin A chain, the neuronal cell adhesion molecule (N-CAM), and the globular C-terminal region of the laminin A chain, respectively [ 126-1281, These likely provide the core protein with interaction sites for other matrix molecules. Mouse perlecan has binding sites for cell surface receptors such as 81 integrins on tumor cells [129], and PI and (33 integrins on endothelial cells[130]. The core protein appears to have 3 glycosaminoglycan chains attached to one end of the core protein (most likely domain I) [ 131,1321. It can carry either heparan sulfate, or (as in placenta) both heparan sulfate and dermatan sulfate [133]. In human

19

colon carcinoma cells, perlecan can be modified by covalently linked myristate and palmitate [45] which enable it to attach to plasma membranes. Perlecan appears to be the major proteoglycan of basement membranes, but other less well-defined proteoglycans containing chondroitin sulfate are also found in this matrix [ 1341. Recently perlecan has been shown to be produced by fibroblasts and to be deposited into the interstitial matrix of several tissues [ 1351. The heparan sulfate chains on basement membrane perlecan apparently play an important part in filtration of charged molecules. For example it has been shown that removal of heparan sulfate from glomerular basement membrane by use of degradative enzymes results in loss of a barrier to anionic substances such as anionic ferritin [ 1361. Perlecan in other basement membranes has also been shown to bind basic fibroblast growth factor (bFGF) and interferon gamma [137,138], with the heparan sulfate chains as the active factors in this binding. Since the binding of growth factors to heparan sulfate may protect them from proteolytic degradation, this may be a mechanism for storage. Proliferation of arterial smooth muscle cells can be inhibited by the heparan sulfate chains on perlecan, suggesting that perlecan may be involved in the regulation of smooth muscle growth [ 1391. This regulation may involve the endocytosis and nuclear transport of the heparan sulfate chains [ 1401. 5.1.4. Part-time proteoglycans

Part-time extracellular proteoglycans include variants of well-characterized extracellular glycoproteins such as type IX collagen [141], C l q [142], colony stimulating factor [143], amyloid precursor protein [ 1441, fibronectin [ 1451, and brain-specific receptor-type tyrosine phosphatase fi [146]. 5.2. Cell suYface proteoglycans

These are either integral membrane proteoglycans or are linked to the membrane via a phosphatidylinositol moiety. They appear to serve as receptors for growth factors and other components of the extracellular matrix, for cell-matrix and cell-cell interactions, and as receptors for other cell-cell interaction molecules. Usually it is the heparan sulfate glycosaminoglycan that is the interactive agent, although chondroitin sulfate and the core proteins have also been shown to function in this manner under some conditions. In addition there are reports of free glycosaminoglycan chains found in cell surfaces [ 1471. Their source and functions are unclear. 5.2.1. Syndecan family These proteoglycans show tissue selectivity [ 1481: syndecan-1 (so named because of a presumptive “binding together” of extracellular matrix and cytoskeleton [ 1491) is most abundant on epithelial cells; syndecan-2 (originally called fibroglycan because it was found in fibroblasts [ 1501) is most abundant on endothelial cells; syndecan-3 (also called N-syndecan because it is found on cells of the nervous system [151]); and ubiquitously expressed syndecan-4 (also called amphiglycan and ryudocan because of its domain structure [152] and because it appears to function as an anticoagulant [ 153,1541). The syndecans are 3 1, 20, 38, and 20 kDa in size, respectively [98]. However they migrate on polyacrylamide gels as spuriously large proteins of 69, 48, 120, and 30 kDa, respectively,

20

presumably due to an extended structural configuration imparted by their high proline content. As proteoglycans, they insert into the plasma membrane via a highly conserved hydrophobic transmembrane domain, leaving a small domain in the cytoplasm [98]. The extracellular domain has an extended configuration with attachment sites for heparan sulfate near the N-terminus away from the cell surface, andor attachment sites for chondroitin sulfate near the cell surface. Syndecan-1 and -4 are found with both heparan sulfate and chondroitin sulfate on the same core protein, while syndecan-2 and -3 have only heparan sulfate. Each core protein shows one or more basic amino acids adjacent to the transmembrane domain, thought to be the site for protease action that releases the extracellular domains from the cell surface [98]. Syndecan-3 contains a long extracellular domain segment rich in threonine, serine and proline residues that may contain attachment sites for short 0-linked oligosaccharides [151]. A syndecan cloned from Drosophila was shown to carry heparan sulfate and to have transmembrane and cytoplasmic domains as well as glycosaminoglycan attachment sites with extensive homology to vertebrate syndecan [ 1551. The heparan sulfate chains on syndecans are attached to sites near the N-terminus of the core proteins. Since this is pointing away from the cell surface, the structure appears to be optimal for presenting the glycosaminoglycans to the extracellular space. Syndecans are thought to function as receptors for growth factors or as receptors for components of extracellular matrix. Syndecan-1 from epithelia binds to type I, 111, and V collagens, fibronectin, thrombospondin and bFGF [98], while syndecan- 1 from embryonic mesenchyme also binds tenascin [ 1561. Syndecan-3 from Schwann cells can bind to bFGF, but not to collagen types I or V or to fibronectin [157]. Only heparan sulfate chains mediate the interactions; chondroitin sulfate in the hybrid proteoglycan does not appear to affect ligand binding [98]. The heparan sulfate chains of syndecan-1 and syndecan-4 have been identified as the substances interacting with antithrombin I11 on the surface of endothelial cells [153] (see section 6). Syndecan-4 is found within fibroblast focal contacts [ 1581, the cellular specialization which mediates adhesion of fibroblasts to tissue culture substrata.

5.2.2. Glypican family This family includes glypican, cerebroglycan, and OCI-5. The core proteins of glypican (so named because it was the first proteoglycan found to be attached to a phosphatidylinositol moiety) and cerebroglycan (so named because it was found in brain) are 64 and 59kDa, respectively. They are cysteine-rich, implying that they have a rigid tertiary structure, and are anchored in cell membranes by a C-terminus covalently attached to glycophosphatidylinositol [ 159,1601similar to the post-translational glycophosphatidylinositol modifications of many glycoproteins. This anchor enables the proteoglycans to be released from the cell surface by the action of specific phospholipases. Glypican and cerebroglycan carry only heparan sulfate chains which appear to be attached, in the case of cerebroglycan, to sites near the N-terminus [161]. Glypican has been found on a variety of cell types [162] while cerebroglycan expression is restricted to cells of the central nervous system [161]. OCI-5 is a transcript that is present in rat intestinal cell lines [ 1631, and whose expression appears to be developmentally regulated in the intact animal. The deduced amino acid sequence of OCI-5 shows a high degree of

21

similarity to that of glypican and cerebroglycan, but it is currently unknown whether or not the protein encoded by OCI-5 carries glycosaminoglycan chains. 5.2.3. Part-time proteoglycans Betaglycan (so called because it is the TGFP type I11 receptor) is a part-time chondroitin sulfate and heparan sulfate proteoglycan found mostly on fibroblasts where it binds TGFP with low affinity[164]. It is a transmembrane cell surface protein, showing a limited similarity to endoglin. In contrast to most other proteoglycans, the core protein of betaglycan and not its glycosaminoglycan chains bind and present the growth factor to its high affinity receptors via the formation of a multiprotein complex [165]. Betaglycan itself does not apparently participate in the intracellular signaling process. NG2 is a part-time chondroitin sulfate-containing transmembrane cell surface protein, which shows limited similarity to N-cadherin [166]. It can bind type VI collagen [ 1671, and this binding is thought to anchor cells to the matrix [168]. Other transmembrane part-time proteoglycans include the hematopoietic and epithelial splice variants of CD44 (Hermes antigen) [ 1691, thrombomodulin [ 1701, fibroblast growth factor receptor-2 (FGFR-2) [171], and the transferrin receptor [172]. 5.3. Intracellular proteoglycans Synthesis of all proteoglycans is intracellular, taking place on ER and Golgi membranes, and much of the degradation of glycosaminoglycans occurs after intracellular transport to lysosomes. However, other than these transient intracellular locations for biosynthesis and degradation, there are a few proteoglycans that are destined for final intracellular secretory granule locations, or for less well-described locations in nuclei, and possibly other organelles. 5.3.1. Serglycin family The core protein of serglycin, a distinctive intracellular granule proteoglycan, was the first core protein to have its sequence obtained by using cloned cDNA [173], and the entire gene has recently been sequenced [ 1741. It is the smallest (1&15 kDa) of all core proteins described to date. Following synthesis, the core protein is proteolytically clipped at the N-terminus to give rise to the mature protein [43,44]. It carries 7-15 glycosaminoglycan chains, 100 or more disaccharides in length, attached to a continuous stretch of 924 SerGly repeats which provide the name “serglycin”. The SerGly repeat portion of serglycin is exceptionally resistant to degradation by proteases [ 175,1761, presumably due to the close grouping of the highly anionic glycosaminoglycans. Serglycin is found within the secretory granules of mast cells, basophils, and natural killer cells where it functions to bind serine proteases and vasoactive amines via its glycosaminoglycan chains. This results in effective packaging of the proteases, prevents their self-digestion [ 177,1781, and reduces the diffusion of low molecular weight amines following degranulation of these cells. Serglycin containing covalently bound heparin is found in connective tissue mast cells [ 176-1 781, while a similar number of oversulfated chondroitin sulfate chains are found in place of heparin on the serglycin of mucosal mast cells [ 177,1781. The serine proteases in mucosal mast cells differ from those in

22

connective tissue mast cells, with each group of proteases interacting maximally with heparin or chondroitin sulfate respectively [ 1791. In oioo, heparin is not generally found as the proteoglycan, since glycosaminoglycan fragments are split from the serglycin by a specific endoglycosidase [91]. 5.3.2. Other proteoglycans Chromaffin granules of adrenal medullary cells have been shown to contain significant amounts of proteoglycans [ 1801. Heparan sulfate and/or chondroitin sulfate glycosaminoglycans have also been reported to be found in cell nuclei of melanoma cells [181], and granulosa cells [ 1821, as well as chondrocytes and hepatocytes [ 1831, but neither the core proteins associated with these nor the significance of this location have been elucidated. The nuclear heparan sulfate apparently originates from phosphatidylinositol-linked cell surface proteoglycans and has been shown to contain unusual 2-sulfated GlcA [ 1831. In addition, there are some suggestions of proteoglycan localization to mitochondria of brain [184] and chondrocytes [185], as well as other organelles in other cells. None of these proteoglycans have been well defined. 5.3.3. Part-time proteoglycans Invariant chain is a transmembrane protein which temporarily associates in the ER and the Golgi apparatus with class I1 gene products of the major histocompatibility complex. A small portion of the invariant chain molecules carry a chondroitin sulfate chain [ 1861. The proteoglycan form of the invariant chain was shown to play a role in the stimulation of T cell response [ 1871.

6. Fine structure/function relations of glycosaminoglycans Earlier work suggested that the glycosaminoglycan portions of some proteoglycans appeared to function in a variety of ways such as, angiogenesis [l88], cell attachment [ 189,1901, cell regulation [ 1911, and particularly through cell surfaceimatrix glycosaminoglycan interactions in morphogenesis and remodeling [85]. Concurrently, development of the techniques for sequencing and classifying core proteins has permitted identification and characterization of the proteoglycans involved in these functions. There also has been some limited fine structure information about the specific placement of sulfate substituents and positions of uronic acid epimerization within individual glycosaminoglycan molecules. Thus there has developed an extensive list of structure/function relationships for the glycosaminoglycans as they are presented on specific matrix or cell surface proteoglycans. Heparin and heparan sulfate bind to a large variety of other proteins (Table l), apparently due to the clustering of highly acidic IduA-GlcNS disaccharide residues. It is clear that the binding is not solely due to ionic interactions since other highly anionic polysaccharides, such as chondroitin sulfate or pentosan polysulfate, may not interfere with their binding. It appears that the presence of IduA enables the glycosaminoglycan chains to be more flexible because IduA can assume a greater number of conformational states than GlcA [176]. A variety of amino acid sequences and secondary protein

23

Table 1 Binding interactions of hepadheparan sulfate with proteins (incomplete listing) Category ~~~

Protein

~

Matrix components

Collagen types I, Ill, IV, V Fibronectin Laminin Pleiotropin Tenascin Thrombospondin Vitronectin

Growth factors

wnt-l Fibroblast growth factor family Heparin binding epidermal growth factor Hepatocyte growth factodscatter factor Interferon gamma Platelet derived growth factor Schwannoma derived growth factor Vascular endothelial growth factor

Cell adhesion molecules

CD45 L-selectin

Mac-I N-CAM PECAM CD31 Enzymes

Acetylcholinesterase Cholesterol esterase Extracellular superoxide dismutase Hepatic and pancreatic triglyceride lipases Lipoprotein lipase Thrombin Tissue plasminogen activator

Lipoproteins

apoB

Protease inhibitors

apoE Antithrombin 111 Heparin cofactor II Leuserpin Plasminogen activator inhibitor- 1 Protease nexin 1

Nuclear proteins

c-fos c-jun RNA and DNA polymerases Steroid receptors

Reference(s)

24

Table 1, continued Category Viral coat proteins

Parasites

Protein

Reference(s)

gC and gB of herpes simplex virus

[ I 11

gC-I1 of cytomegalovims gp 120 of human immunodeficiency virus

~091

Borrelia

[2101

Chlamydia

[2111

Leishmania Malaria circumsporozooite

[2121

[Ill

Trypanosoma

structures have been found to interact with heparin and heparan sulfate [ 11,194,2 16,2171 suggesting that distinct structures within the heparin or heparan sulfate chains might be important for specific interactions. Heparin is the best example of glycosaminoglycan function or action based upon a specific oligosaccharide sequence. The well-known and extensively utilized anticoagulant activity of this glycosaminoglycan has been shown to be based upon the binding and activation of antithrombin I11 [2 18,2191. The activity requires a specific pentasaccharide containing an unusual 3-0-sulfated GlcNS present in small numbers in a minority of the glycosaminoglycan chains [23]. The sequence-selective binding accelerates the formation of a stable and inactive complex with the proteolytic enzyme thrombin, and is the basis for the anticoagulant activity. As mentioned in section 5.2.1, the sequence is also found in heparan sulfate of the luminal surfaces of vascular endothelia, apparently on syndecan-1 and on syndecan-4 [152]. This heparan sulfate on endothelial or other cell surfaces is probably the physiological anticoagulant, while heparin from mast cells probably does not ordinarily function in uiuo in this capacity. Growth factors also appear to show interactions with oligosaccharide sequences in heparan sulfate. While there is not yet a consensus on the precise structures, bFGF binds to at least a specific pentasaccharide [220] and possibly a heptasaccharide [221]. A dodecasaccharide appears to be involved in forming a ternary complex with bFGF and its signal-transducing receptor. Other growth factors, such as acidic fibroblast growth factor (aFGF), FGF-4, hepatocyte growth factor/scatter factor, and heparin-binding epidermal growth factor (EGF) also may require specific structural motifs in heparan sulfate [222-2241. Recently a decasaccharide sequence from cell surface heparan sulfate was found to interact with lipoprotein lipase [225], an enzyme previously shown to bind to heparan sulfate on the surface of endothelial cells [226]. The binding of proteins to specific sequences in structurally heterogeneous heparan sulfate suggests that cells possess mechanisms for generating these sequences in a programmed rather than random manner, and recent evidence appears to support this. For example, the heparan sulfate on cell surface syndecan-1 from distinct cell types was

25

found to differ in fine structure in a consistent and reproducible manner that correlated with differences in ligand binding [227]. In this case the structure of the heparan sulfate depended upon the cell type rather than upon the nature of the core protein. Thus cell differentiation appears to result in specific heparan sulfate structures, but the process responsible for the differentiated characteristics are unknown. Interactions of chondroitin sulfate and dermatan sulfate with ligands have not been found to as great an extent as the interactions of heparan sulfate. The best described example is the specific hexasaccharide from dermatan sulfate [228] that can bind and activate heparin cofactor I1 at a site distinct from the interaction of this protein with heparin [229]. Another example is the presence of IduA rather than GlcA on decorin which provides a better conformation for interaction of two glycosaminoglycan chains with each other [116] (see section 5.1) in collagen fibril spacing. It could be postulated that a deficiency in the amount of IduA in the dermatan sulfate might result in a loss of stability in the collagen matrix.

References [I] Wight, T.N. and Mecham, R.D. (Eds.) (1987) Biology of Proteoglycans. Academic Press, Orlando, 400 pp. [2] Jolles, I? (Ed.) (1987) Proteoglycans. Birkhauser Verlag AG, Basel, 280 pp. [3] Beaty, N.B. and Mello, R.J. (1987) J. Chromatogr. 418, 187-222. [4] Fransson, L.-A. (1987) Trends Biochem. Sci. 12, 4 0 6 4 1 I . [5] Ruoslahti, E. (1988) Annu. Rev. Cell Biol. 4, 229-255. [6] Gallagher, J.T. (1989) Curr. Opin. Cell Biol. 1 , 1201-1218. [7] Heinegird, D. and Oldberg, A. (1989) FASEB J. 3, 2042-2051. [8] Kolset, S.O. and Gallagher, J.T. (1990) Biochim. Biophys. Acta 1032, 191-21 I . [9] Hascall, V.C., Heinegird, D.K. and Wight, T.N. (1991) In: E.D. Hay (Ed.), Cell Biology of Extracellular Matrix. Plenum Press, New York, pp. 149-175. [lo] Wight, T.N., Heinegird, D.K. and Hascall, VC. (1991) In: E.D. Hay (Ed.), Cell Biology of Extracellular Matrix. Plenum Press, New York, pp. 45-78. [I I] Jackson, R.L., Busch, S.J. and Cardin, A.D. (1991) Physiol. Rev. 71, 481-539. [I21 Kjelltn, L. and Lindahl, U. (1991) Annu. Rev. Biochem. 60, 443475. [I31 Hardingham, T.E. and Fosang, A.J. (1992) FASEB J. 6, 861-870. [I41 Kreis, T. and Vale, R. (1993) Guidebook to the Extracellular Matrix and Adhesion Proteins. Oxford Univ. Press, New York, 176 pp. [I51 Kresse, H., Hausser, H. and Schonherr, E. (1993) Experientia 49, 403416. [I61 David, G. (1993) FASEB J. 7, 1023-1030. [I71 lozzo, R.V., Cohen, I.R., Grassel, S. and Murdoch, A.D. (1994) Biochem. J. 302, 625-639. [I81 Scott, J.E. (1993) Glycoconj. J. 10, 419-421. [I91 Humphries, D.E., Sugumaran, G. and Silbert, J.E. (1989) Methods Enzymol. 179, 428434. [20] Silbert, J.E. (1978) J. Biol. Chem. 253, 6888-6892. [21] Scott, J.E. (1994) Biochem. J. 298, 221-222. [22] Turnbull, J.E. and Gallagher, J.T. (1990) Biochem. J. 265, 715-724. [23] Lindahl, U., Backstrom, G., Thunberg, L. and Leder, I.G. (1980) Proc. Natl. Acad. Sci. USA 77, 65516555. [24] Stuhlsatz, H.W., Keller, R., Becker, G., Oeben, M., Lennartz, L., Fischer, D.C. and Greiling, H. (1989) In: H. Greiling and J.E. Scott (Eds.), Keratan Sulphate: Chemistry, Biology, Chemical Pathology. The Biochemical Society, London, pp. 1-1 1 . [25] Greiling, H. (1994) In: I? Jolles (Ed.), Proteoglycans. Birkhauser Verlag, Basel, pp. 101-122.

26 [26] Helting, T. and Roden, L. (1969) J. Biol. Chem. 244, 2799-2805. [27] Oegema, T.R., Kraft, E.L. Jourdian, G.W. and VanValen, T.R. (1984) J. Biol. Chem. 259, 1720-1726. [28] Sugahara, K., Ohi, Y., Harada, T., DeWaard, P. and Vliegenthart, F.G. (1992) J. Biol. Chem. 267, 60276035. [29] Bray, B.A., Lieberman, R. and Meyer, K. (1967) J. Biol. Chem. 242, 3373-3380. [30] Baker, J.R., Cifonelli, J.A. and Roden, L. (1975) Connect. Tissue Res. 3, 149-156. [31] Funderburgh, J.L., Funderburgh, M.L., Mann, M.M. and Conrad, G.W. (1991) J. Biol. Chem. 266, 24773-24777. [32] Upholt, W.B., Vertel, B.M. and Dorfman, A. (1979) Proc. Natl. Acad. Sci. USA 76, 48474851. [33] Treadwell, B.V, Mankin, D.P., Ho, P.K. and Mankin, H.J. (1980) Biochemistry 19, 2269-2275. [34] Sandell, L.J. (1987) In: T.N. Wight and R.P. Mecham (Eds.), Biology of Proteoglycans. Academic Press, Orlando, pp. 27-57. [35] Mann, D.M., Yamaguchi, Y., Bourdon, M.A. and Ruoslahti, E. (1990) J. Biol. Chem. 265, 5317-5323. [36] Kokenyesi, R. and Bernfield, M. (1994) J. Biol. Chem. 269, 12304-12309. [37] Zhang, L. and Esko, J.D. (1994) J. Biol. Chem. 269, 19295-19299. [38] Shworak, N.W., Shirakawa, M., Mulligan, R.C. and Rosenberg, R.D. (1994) J. Biol. Chem. 269, 2120421214. [39] Faltynek, C.R. and Silbert, J.E. (1981) J. Biol. Chem. 256, 7202-7206. [40] Sandy, J.D. (1992) In: K.E. Kuettner, R. Schleyerbach, J.G. Peyron and VC. Hascall (Eds.), Articular Cartilage and Osteoarthritis. Raven Press, New York, pp. 21-33. [41] Sawhney, R.S., Hering, T.M. and Sandell, L.J. (1991) J. Biol. Chem. 266, 9231-9240. [42] Rauch, U., Karthikeyan, L., Maurel, P.,Margolis, R.U. and Margolis, R.K. (1992) J. Biol. Chem. 267, 19536-19547. [43] Bourdon, M.A., Shiga, M. and Ruoslahti, E. (1986) J. Biol. Chem. 261, 12534-12537. [44] Avraham, S., Stevens. R.L., Nicodemus, C.F., Gartner, M.C., Austen, K.F. and Weis, J.H. (1989) Proc. Natl. Acad. Sci. USA 86, 3763-3767. [45] lozzo, R.V, Kovalszky, I., Hacobian, N., Schick, P.K., Ellingson, J.S. and Dodge, G.H. (1990) J. Biol. Chem. 265, 19980-19989. [46] Silbert, J.E. and DeLuca, S. (1967) Biochim. Biophys. Acta 141, 193-196. [47] Faltynek, C.R, Silbert, J.E. and Hof, L. (1982) J. Biol. Chem. 257, 5490-5495. [48] Esko, J.D., Stewart, T.E. and Taylor, W.H. (1985) Proc. Natl. Acad. Sci. USA 82, 3197-3201. [49] Hoffmann, H.P., Schwartz, N.B., Roden, L. and Prockop, D.J. (1984) Connect. Tissue Res. 12, 151-163. [50] Kimura, J.H., Lohmander, L.S. and Hascall, VC. (1984) J. Cell. Biochem. 26, 261-278. [51] Vertel, B.M. Walters, L.M., Flay, N., Kearns, A.E. and Schwartz, N.B. (1993) J. Biol. Chem. 268, I 1 105-1 11 12. [52] Neufeld, E.F. and Hall, C.W. (1965) Biochem. Biophys. Res. Commun. 19, 4 5 6 4 6 1 . 1531 Kearns, A.E., Vertel, B.M. and Schwartz, N.B. (1993) J. Biol. Chem. 268, 11097-1 1104. [54] Hirschberg, C.B. and Snider, M.D. (1987) Annu. Rev. Biochem. 56, 63-87. [55] Helting, T. and Roden, L. (1969) J. Biol. Chem. 244, 2790-2798. [56] Helting, T. (1971) J. Biol. Chem. 246, 815-822. [57] Sugumaran, G. and Silbert, J.E. (1991) J. Biol. Chem. 266, 9565-9569. [58] Sugumaran, G., Katsman, M. and Silbert, J.E. (1992) J. Biol. Chem. 267, 8802-8806. [59] Helting, T. (1972) J. Biol. Chem. 247, 43274332. [60] Rohrmann, K., Niemann, R. and Buddecke, E. (1985) Eur. J. Biochem. 148, 4 6 3 4 6 9 . [61] Fritz, T.A., Gabb, M.M., Wei, G. and Esko, J.D. (1994) J. Biol. Chem. 269, 28808-28814. [62] Silbert, J.E. (1963) J. Biol. Chem. 238, 3542-3546. [63] Silbert, J.E. (1987) In: J. Uitto and A.J. Perejda (Eds.), Connective Tissue Disease: Molecular Pathology of the Extracellular Matrix. Marcel Dekker, New York, pp. 83-98. [64] Richmond. M.E., DeLuca, S. and Silbert, J.E. (1973) Biochemistry 12, 3904-3910. [65] Lindahl, U., Feingold, D.S. and Rodin, L. (1986) Trends Biochem. Sci. 11, 221-225. [66] Silbert, J.E. and Reppucci Jr., A.C. (1976) J. Biol. Chem. 251, 3942-3947. [67] Prehm, P. (1983) Biochem. J. 211, 191-198. [68] Malmstrom, A., Fransson, L.-A., Hook, M. and Lindahl, U. (1975) J. Biol. Chem. 250, 3419-3425.

27 Lind, T., Lindahl, U. and Lidholt, K. (1993) J. Biol. Chem. 268, 20705-20708. Sugumaran, G. and Silbert, J.E. (1990) J. Biol. Chem. 265, 18284-18288. DeLuca, S., Richmond, M.E. and Silbert, J.E. (1973) Biochemistry 12, 391 1-3915. Silbert, J.E., Palmer, M.E., Humphries, D.E. andsilbert, C.K. (1986) J. Biol. Chem. 261, 13397-13400. Silbert, J.E. (1967) J. Biol. Chem. 242, 5153-5157. Wei, Z., Swiedler, S.J., Ishihara, M., Orellana, A. and Hirschberg, C.B. (1993) Proc. Natl. Acad. Sci. USA 90, 3885-3888. [75] Silbert, J.E. and Freilich, L.S. (1980) Biochem. J. 190, 307-313. [76] Philipson, L.H. and Schwartz, N.B. (1984) J. Biol. Chem. 259, 5017-5023. [77] Woessner Jr., J.F. (1991) FASEB J. 5 , 2145-2154. [78] Alexander, C.M. and Werb, Z. (1991) In: E.D. Hay (Ed.), Cell Biology of Extracellular Matrix. Plenum Press, New York, pp. 255-302. [79] Thomas, L. (1964) Biophys. J. 4(Suppl.), 207-213. [80] Handley, C.J. and Campbell, M.A. (1987) Methods Enzymol. 144, 4 1 2 4 1 9 . [8 I ] Kresse, H. and Glossl, J. (1 987) Adv. Enzymol. 60, 2 17-3 1 I . [82] Roden, L., Campbell, P., Fraser, J.R.E., Laurent, T.E., Pertoft, H. and Thompson, J.N. (1989) In: The Biology of Hyaluronan. Ciba Foundation Symp. 143, 60-86. [83] Hoffman, D.C. and Duran-Reynals, F. (1931) J. Exp. Med. 53, 387-398. [84] Bhavanandan, VP. and Sampson, I? (1992) Trends Glycosci. Glycotech. 4, 390401. [85] Toole, B.P. (1991) In: E.D. Hay (Ed.), Cell Biology of Extracellular Matrix. Plenum Press, New York, pp. 305-341. [86] Schiller. S., Mathews, M.B., Cifonelli, J.A. and Dorfman, A. (1956) J. Biol. Chem. 218, 139-145. [87] Davidson, E.A. and Small, W. (1963) Biochim. Biophys. Acta 69, 453458. [88] Oosta, G.M., Favreau, L.V, Beeler, D.L. and Rosenberg, R.D. (1982) J. Biol. Chem. 257, 11249-1 1255. I891 Gallagher, J.T., Walker, A,, Lyon, M. and Evans, W.H. (1988) Biochem. J. 250, 719-726. [90] Bame, K.J. (1993) J. Biol. Chem. 268, 19956-19964. [91] Young, E. and Horner, A.A. (1979) Biochem. J. 180, 587-596. [92] Yanagishita, M. and Hascall, VC. (1992) J. Biol. Chem. 267, 9451-9454. [93] McKusick, VA. (Ed.) (1988) Mendelian Inheritance in Man, 8th ed. Johns Hopkins Press, Baltimore, pp. 1072-1 080. [94] Hopwood, J.J. and Morris, C.P. (1990) Mol. Biol. Med. 7, 381404. [95] Neufeld, E.F. (1991) Annu. Rev. Biochem. 60, 257-280. [96] Linhardt, R.J., Galliher, P.M. and Cooney, C.L. (1986) Appl. Biochem. Biotechnol. 12, 135-176. [97] Esko, J.D. (1991) Curr. Opin. Cell Biol. 3, 805-816. [98] Bernfield, M., Kokenyesi, R., Kato, M., Hinkes, M.T., Spring, J., Gallo, R.L. and Lose, E.J. (1992) Annu. Rev. Cell Biol. 8, 365-393. [99] Doege, K.J., Sasaki, M., Kimura, T. and Yamada, Y. (1991) J. Biol. Chem. 266, 894-902. [ 1001 Zimmermann, D.R. and Ruoslahti, E. (1 989) EMBO J. 8, 2975-298 1. [IOI] Yamada, H., Watanabe, K., Shimonaka, M. and Yamaguchi, Y. (1994) J. Biol. Chem. 269, 10119-10126. [I021 Neame, P.J. (1993) In: J.F. Woessner Jr. and D.S. Howell (Eds.), Joint Cartilage Degradation: Basic and Clinical Aspects. Marcel Dekker, New York, pp. 109-138. [lo31 Doege, K., Sasaki, M., Horigan, E., Hassell, J.R. and Yamada, Y. (1987) J. Biol. Chem. 262, 1775717767. [I041 Chandrasekaran, L. and Tanzer, M.L. (1992) Biochem. J. 288, 903-910. [I051 Walcz, E., Deak, F., Erhardt, P., Coulter, S.N., Fulop, C., Horvath, P.,Doege, K.J. and Glant, T.T. (1994) Genomics 22, 364-37 1. [I061 Baldwin, C.T., Reginato, A.M. and Prockop, D.W. (1989) J. Biol. Chem. 264, 15747-15750. [I071 Schafer, I.A., Sitabkha, L. and Pandy, M. (1984 ) J. Biol. Chem. 259, 2321-2330. [I081 Schmidtchen, A,, Carlstedt, I., Malmstrom, A. and Fransson, L.-A. (1990) Biochem. J. 265, 289-300. [I091 LeBaron, R.G., Zimmermann, D.R. and Ruoslahti, E. (1992) J. Biol. Chem. 267, 10003-10010. [IIO] Bignami, A,, Perides, G. and Rahemtulla, F. (1993) J. Neurosci. Res. 34, 97-106. [ I 1 I] Rauch, U., Gao, P., Janetzko, A., Flaccus, A,, Hilgenberg, L., Tekotte, H., Margolis, R.K. and Margolis, R.U. (1991) J. Biol. Chem. 266, 14785-14801, [69] [70] [71] [72] [73] [74]

28 [I121 [I131 [I141 [I151 [ 1 I61

Malvar, T., Biron, R.W., Kaback, D.B. and Denis C.L. (1992) Genetics 132, 951-962. Glossl, J., Beck, M. and Kresse, H. (1984) J. Biol. Chem. 259, 14144-14150. Chopra, R.K., Pearson, C., Pringle, G., Fackkre, D., Scott, P. (1985) Biochem. J. 232, 277-279. Scott, J.E. (1988) Biochem. J. 252, 313-323. Scott, J.E. (1993) In: J.E. Scott (Ed.), Dermatan Sulphate Proteoglycans: Chemistry, Biology, Chemical Pathology. Portland Press, London, pp. 165-1 8 1. [I171 Yamaguchi, Y., Mann, D.M. and Ruoslahti, E. (1990) Nature 346, 281-284. [I181 Hildebrand, A,, Romaris, M., Rasmussen, L.M., Heinegird, D., Twardzik, D.R., Border, W.A. and Ruoslahti, E. (1994) Biochem. J. 302, 527-534. [I191 Border, W.A. and Noble, N.A. (1994) New Eng. J. Med. 331, 1286-1292. [I201 Fisher, L.W., Heegaard, A.-M., Vetter, U., Vogel, W., Just, W., Termine, J.D. and Young, M.F. (1991) J. Biol. Chem. 266, 14371-14377. [I211 Oldberg, A,,Antonsson, P., Lindblom, K. and Heinegird, D. (1989) EMBO J. 6, 2601-2604. [I221 Blochberger, T.C., Vergnes, J.-P., Hempel, J. and Hassell, J.R. (1992) J. Biol. Chem. 267, 347-352. [123] Midura, R.J., Hascall, V.C., MacCallum, D.K., Meyer, R.F., Thonar, E.J.-M.A., Hassell, J.R., Smith, C.F. and Klintworth, G.K. (1990) J. Biol. Chem. 265, 15947-15955. [ 1241 Cornuet, P.K., Blochberger, T.C. and Hassell, J.R. (1994) Invest. Ophthalmol. Vis. Sci. 35, 870-877. [I251 Rada, J.A., Cornuet, P.K. and Hassell, J.R. (1993) Exp. Eye Res. 56, 635-648. [I261 Noonan, D.M., Fulle, A,, Valente, P., Cai, S., Horigan, E., Sasaki, M., Yamada, Y. and Hassell, J.R. (1991) J. Biol. Chem. 266, 22939-22947. [I271 Kallunki, P. and Tryggvason, K. (1992) J. Cell Biol. 116, 559-571. [I281 Murdoch, A.D., Dodge, G.R., Cohen, I., Tuan. R.S. and Iozzo, R.V. (1992) J. Biol. Chem. 267, 85448557. [I291 Battaglia, C., Aumailley, M., Mann, K., Mayer, U. and Timpl, R. (1993) Eur. J. Cell Biol. 61, 92-99. [I301 Hayashi, K., Madri, J.A. and Yurchenco, P.D. (1992) J. Cell Biol. 119, 945-959. [I311 Yurchenco, P.D., Cheng, Y.-S. and Ruben, G.C. (1987) J. Biol. Chem. 262, 17668-17676. [132] Laurie, G.W., Inoue, S., Bing, J.T. and Hassell, J.R. (1988) Am. J. Anat. 181, 320-326. [I331 Isemura, M., Sato, N., Yamaguchi, Y., Aikawa, J., Munakata, H., Hayashi, N., Yosizawa, Z., Nakamura, T., Kubota, A,, Arakawa, M. and Hsu, C.-C. (1987) J. Biol. Chem. 262, 8926-8933. [I341 McCarthy, K.J., Horiguchi, Y., Couchman, J.R. and Fine, J.-D. (1990) Arch. Dermatol. Res. 282, 397401. [I351 Iozzo, R.V: (1994) Matrix Biol. 14, 203-208. [I361 Farquhar, M.G. (1991) In: E.D. Hay (Ed.), Cell Biology of Extracellular Matrix. Plenum Press, New York, pp. 365418. [I371 Vigny, M., Ollier-Hartmann, M.P., Lavigne, M., Fayein, N., Jeanny, J.C., Laurent, M. and Courtois, Y. (1988) J. Cell. Physiol. 137, 321-328. [I381 Lortat-Jacob, H., Kleinman, H.K. and Grimaud, J.-A. (1991) J. Clin. Invest. 87, 878-883. [I391 Benitz, W.E., Kelley, R.T. Anderson, C.M., Lorant, D.E. and Bernfield, M. (1990) Am. J. Respir. Cell Mol. Biol. 2, 13-24. [I401 Castellot, J.J., Jr., Wong, K., Herman, B., Hoover, R.L., Albertini, D.F., Wright, T.C., Caleb, B.L. and Karnovsky, M.J. (1985) J. Cell. Physiol. 124, 13-20. [I411 Bruckner, P., Vaughan, L. and Winterhalter, K.H. (1985) Proc. Natl. Acad. Sci. USA 82, 2608-2612. [I421 Galanakis, D.K. and Ghebrehiwet, B. (1994) J. Clin. Invest. 93, 303-310. [I431 Price, L.K.H., Choi, H.U., Rosenberg, L. and Stanley, E.S. (1992) J. Biol. Chem. 267, 2190-2199. [I441 Shioi, J. Anderson, J.P., Ripellino, J.A. and Robakis, N.K. (1992) J. Biol. Chem. 267, 13819-13822. [I451 Cossu, G. and Warren, L. (1983) J. Biol. Chem. 258, 5603-5607. [I461 Maurel, P., Rauch, U., Flad, M., Margolis, R.K. and Margolis, R.U. (1994) Proc. Natl. Acad. Sci. USA 91, 2512-2516. [I471 Piepkom, M., Fleckman, P., Carney, H., Hovingh, P. and Linker, A. (1990) J. Invest. Dermatol. 94, 107-1 13. [I481 Kim, C.W., Goldberger, O.A., Gallo, R.L. and Bernfield, M. (1994) Mol. Biol. Cell 5, 797-805. [I491 Saunders, S., Jalkanen, M., O’Farell, S. and Bernfield, M. (1989) J. Cell Biol. 108, 154771556, [I501 Marynen, P., Cassiman, J., Van den Berghe, H. and David, G. (1989) J. Biol. Chem. 264, 7017-7024.

29 [I511 Carey, D.J., Evans, D.M., Stahl, R.C., Asundi, VK., Comer, K.J., Garbes, P. and Cizmeci-Smith, G. (1992) J. Cell Biol. 117, 191-201. [I521 David, G., Van der Schueren, B., Marynen, I?, Cassiman, J.-J. and Van den Berghe, H. (1992) J. Cell Biol. 118, 961-969. [I531 Kojima, T., Leone, C.W., Marchildon, G.A., Marcum, J.A. and Rosenberg, R.D. (1992) J. Biol. Chem. 267, 48594869. [I541 Kojima, T., Shworak, N.W. and Rosenberg, R.D. (1992) J. Biol. Chem. 267, 48704877. [155] Spring, J., Paine-Saunders, S.E., Hynes, R.O. and Bernfield, M. (1994) Proc. Natl. Acad. Sci. USA 91, 3 334-33 38. [I561 Salmivirta, M., Elenius, K., Vainio, S., Hofer, U., Chiquet-Ehrismann, R., Thesleff, I . and Jalkanen, M. (1991) J. Biol. Chem. 266, 7733-7739. [I571 Chernousov, M A . and Carey, D.J. (1993) J. Biol. Chem. 268, 16810-16814. [I581 Woods, A. and Couchman, J.R. (1994) Mol. Biol. Cell 5, 183-192. [I591 Ishihara, M., Fedarko, N.S and Conrad, H.E. (1987) J. Biol. Chem. 262, 47084726. [I601 David, G., Lories, V; Decock, B., Marynen, P., Cassiman, J.-J. and Van den Berghe, H. (1990) J. Cell Biol. 1 1 1, 3 165-3 176. [I611 Stipp, C.S., Litwack, E.D. and Lander, A.D. (1994) J. Cell Biol. 124, 149-160. [162] Lories, V;, Cassiman, J.-J., Van den Berghe, H. and David, G. (1992) J. Biol. Chem. 267, 1 1 16-1 122. [I631 Filmus, J., Church, J.G. and Buick, R.N. (1988) Mol. Cell Biol. 81, 42334249. [I641 Lopez-Casillas, F,, Cheifetz, S., Doody, J. Andres, J.L., Lane, W.S. and Massague, J. (1991) Cell 67, 785-795. [I651 Lopez-Casillas, F., Wrans, J.L. and Massague, J. (1993) Cell 73, 1435-1444. [I661 Nishiyama, A., Dahlin, K.J., Prince J.T., Johnstone, S.R. and Stallcup, W.B. (1991) J. Cell Biol. 114, 359-371. [I671 Stallcup, W.B., Dahlin, K. and Healy, P. (1990) J. Cell Biol. 1 1 1 , 3177-3 188. [I681 Nishiyama, A. and Stallcup, W.B. (1993) Mol. Biol. Cell 4, 1097-1 108. [I691 Brown, T.A., Bouchard,T., St. John, T., Wayner, E. andcarter, W.G. (1991) J. Cell Biol. 113,207-221. [I701 Wen, D.Z., Dittman, W.A., Ye, R.D., Deaven, L.L., Majerus, P.W. and Sadler, J.E. (1987) Biochemistry 26, 43504357. [I711 Sakaguchi, K., Yanagishita, M., Takeuchi, Y. and Aurbach, G.D. (1991) J. Biol. Chem. 266,7270-7278. [I721 Fransson, L.-A., Carlstedt, I., Coster, L. and Malmstrom, A. (1984) Proc. Natl. Acad. Sci. USA 81, 5657-5661. [I731 Bourdon, MA., Oldberg, A,, Pierschbacher, M. and Ruoslahti, E. (1985) Proc. Natl. Acad. Sci. USA 82, 1321-1325. [I741 Humphries, D.E., Nicodemus, C.F., Schiller, V and Stevens, R.L. (1992) J. Biol. Chem. 267, 1355813563. [I751 Homer, A.A. (1971) J. Biol. Chem. 246, 231-239. [176] Yurt, R.W., Leid, R.W., Austen, K.F. and Silbert, J.E. (1977) J. Biol. Chem. 252, 518-521. [I771 Stevens, R.L. (1987) In: T.N. Wight and R.P. Mecham (Eds.), Biology of Proteoglycans. Academic Press, Orlando, pp. 367-387. [I781 Stevens, R.L. and Austen, K.E (1989) Immunol. Today 10, 381-386. [I791 Sali, A,, Matsumoto, R.T., McNeil, H.P., Karplus, M. and Stevens, R.L. (1993) J. Biol. Chem. 268, 9023-9034. [I801 Kiang, W.-L., Krusius, T., Finne, J., Margolis, R.U. and Margolis, R.K. (1982) J. Biol. Chem. 257, I65 1-1 659. [181] Bhavanandan, W? and Davidson, E.A. (1975) Proc. Natl. Acad. Sci. USA 72, 2032-2036. [182] Hiscock, D.R.R., Yanagishita, M. and Hascall, VC. (1994) J. Biol. Chem. 269, 45394546. [I831 Fedarko, N.S. and Conrad, H.E. (1986) J. Cell Biol. 102, 587-599. [184] Margolis, R.K., Margolis, R.U., Preti, C. and Lai, D. (1975) Biochemistry 14, 47974804. [I851 Ratcliffe, A,, Fryer, P.R. and Hardingham, T.E. (1985) J. Cell Biol. 101, 235552365, [I861 Sant, A.J., Cullen, S.E., Giacoletto, K.S. and Schwartz, B.D. (1985) J. Exp. Med. 162, 1916-1934. [187] Naujokas, M.F., Morin, M., Anderson, M.S., Peterson, M. and Miller, J. (1993) Cell 74, 257-268.

30 [I881 Folkman, J.M. and Shing, Y. (1992) In: D. Lane and U. Lindahl (Eds.), Heparin and Related Saccharides. Plenum Press, London, pp. 355-364. [I891 Laterra, J., Silbert. J.E. and Culp, L.A. (1983) J. Cell Biol. 96, 112-123. [I901 Gill, PJ., Silbert, C.K. and Silbert, J.E. (1986) Biochemistry 25, 405-410. [I911 Ruoslahti, E. (1989) J. Biol. Chem. 264, 13369-13372. [I921 Koda, J.E., Rapraeger, A. and Bernfield, M. (1985) J. Biol. Chem. 260, 815778162, [I931 Tsilibary, E.C., Koliakos, G.G., Charonis, A.S., Vogel, A.M., Reger, L.A. and Furcht, L.T. (1988) J. Biol. Chem. 263, 19112-19228. [I941 Cardin, A.D. and Weintraub, H.J.R. (1989) Arteriosclerosis 9, 21-32. [I951 Merenmies, J. and Rauvala, H. (1990) J. Biol. Chem. 265, 16721-16724. [I961 Bradley, R.S. and Brown, A.M. (1990) EMBO J. 9, 1569-1575. [I971 Higashiyama, S., Abraham, J.A., Miller, J., Fiddes, F.C. and Klagsbrun, M. (1991) Science 251, 936939. [I981 Tashiro, K., Hagiya, M., Nishizawa, T., Seki, T., Shimonishi, M., Shimizu, S. and Nakamura, T. (1990) Proc. Natl. Acad. Sci. USA 87, 3200-3204. [I991 Raines, E.W. and Ross, R. (1992) J. Cell Biol. 116, 533-543. [200] Kimura, H., Fischer, W.H. and Schubert, D. (1990) Nature 348, 257-260. [201] Ferrara, N. and Henzel, W.J. (1989) Biochem. Biophys. Res. Commun. 161, 851-858. [202] Coombe, D.R., Watt, S.M. and Parish, C.R. (1994) Blood 84, 739-752. [203] Norgard-Sumnicht, K.E., Varki, N.M. and Varki, A. (1993) Science 261, 4 8 0 4 8 3 . [204] Watt, S.M., Williamson, J., Genevier, H., Fawcett, J., Simmons, D.L., Hatzfeld, A,, Nesbitt, S.A. and Coombe, D.R. (1993) Blood 82, 2649-2663. [205] Brandan, E., Maldonado, M., Garrido, J. and Inestrosa, N.C. (1985) J. Cell Biol. 101, 985-982. [206] Andrade-Gordon, P. and Strickland, S. (1990) Proc. Natl. Acad. Sci. USA 87, 1865-1869. [207] Ehrlich, H.J., Keijer, J., Preissner, K.T., Gebbink, R.K. and Pannekoek, H. (1991) Biochemistry 30, 102 1-1028. [208] Farooqui, A.A. and Horrocks, L.A. (1984) Adv. Chromatogr. 23, 127-148. [209] Kari, B. and Gehrz, R. (1992) J. Virol. 66, 1761-1764. [210] Isaacs, R.D. (1994) J. Clin. Invest. 93, 809-819. [211] Zhang, J.P. and Stephens, R.S. (1992) Cell 69, 861-869. [212] Love, D.C., Esko, J.D. and Mosser, D.M. (1993) J. Cell Biol. 123, 759-766. [213] Frevert, U., Sinnis, P., Cerami, C., Shreffler. W., Takacs, B. and Nussenzweig, V. (1993) J. Exp. Med. 177, 1287-1298. [214] Ortega-Barria, E. and Pereira, E.A. (1991) Cell 67, 41 1 4 2 1 . [215] Casu, B., Petitou, M., Provasoli, M. and Sinay, P. (1988) Trends Biochem. Sci. 13, 221-225. [216] Margalit, H., Fischer, N. and Ben-Sasson, S.A. (1993) J. Biol. Chem. 268, 19228-19231. [217] San Antonio, J.D., Lander, A.D., Karnovsky, M.J. and Slayter, H.S. (1994) J. Cell Biol. 125, 1179-1 188. [218] Rosenberg, R.D. and Damus, P.S. (1973) J. Biol. Chem. 248, 6490-6505. [219] Bourin, M.-C. and Lindahl, U. (1993) Biochem. J. 289, 313-330. [220] Maccarana, M., Casu, B. and Lindahl, U. (1993) J. Biol. Chem. 268, 23898-23905. [221] Turnbull, J.E., Femig, D.G., Ke, Y., Wilkinson, M.C. and Gallagher, J.T. (1992) J. Biol. Chem. 267, 10337-10341. [222] Guimond, S., Maccarana, M., Olwin, B.B., Lindahl, U. and Rapraeger, A.C. (1993) J. Biol. Chem. 268, 23906-239 14. [223] Lyon, M.. Deakin, J.A., Mizuno, K., Nakamura, T. and Gallagher, J.T. (1994) J. Biol. Chem. 269, 1121611223. [224] Spivak-Kroizman, T., Lemmon, M.A., Dikic, I., Ladbury, J.E., Pinchasi, D., Huang, F., Jaye, M., Crumley, G., Schlessinger, J. and Lax, 1. (1994) Cell 79, 1015-1024. [225] Parthasarathy, N., Goldberg, I.J., Sivaram, P, Mulloy, B., Flory, D.M. and Wagner, W.D. (1994) J. Biol. Chem. 269, 22391-22396. [226] Shimada, K., Gill, PJ., Silbert, J.E., Douglas, W.H.J. and Fanburg, B.L. (1981) J. Clin. Invest. 68, 9951002.

31 [227] Kato, M., Wang, H., Bernfield, M., Gallagher, J.T. and Turnbull, J.E. (1994) J. Biol. Chem. 269, 1888118890. [228] Maimone, M.M. and Tollefsen, D.M. (1990) J. Biol. Chem. 265, 18263-18271. [229] Blinder, M.A., Anderson, T.R., Abildgaard, U. and Tollefsen, D.M. (1989) J. Biol. Chem. 264, 51285133.

This Page Intentionally Left Blank

J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins II 0 1997 Elsevier Science B.V. All rights reserved CHAPTER 2

Nuclear and cytoplasmic glycoproteins Kenneth D. Greis* and Gerald W. Hart Department of Biochemistry and Molecular Genetics. UAB School of Medicine and Dentistry, UAB Station, Birmingham, AL 35294, USA

Abbreviations 0-GlcNAc

0-linked N-acetylglucosamine

PNGase F

GT bovine milk galactosyltransferase RNA Pol I1 catalytic subunit of RNA polymerase I1 from calf thymus CTD

carboxyl-terminal domain of RNA Pol I1

WGA

wheat germ agglutinin plant lectin

PUGNAC

LOGNAC

endo-P-N-acetylglucosaminidaseF [0-(2-acetamino-2-deoxyD-glucopyranosy1idene)-aminoN-phenylcarbamate] [2-acetamino-2-deoxyD-gluconhydroxime-1,5-lactone]

I . Introduction The structure, biosynthesis and function of secretory, cell surface, and lumenal oriented glycoproteins have been the primary focus of glycobiology for many years (for reviews see refs. [l-31 and Chapters 2 and 3 of Volume 29A of this series). Many studies have reported the existence of glycoproteins in the cytosol and the nucleus of cells via lectin binding studies, by direct monosaccharide compositional analysis, or by metabolic labeling of subcellular fractions [4]. However, these studies were often ignored due to the lack of complete structural analysis of the glycans, the inability to rule out contamination from other subcellular compartments, and probably most notably, there was no direct biochemical evidence to explain transportation or synthesis of glycoproteins in these compartments. Now, significant biochemical evidence has been presented over the last 10 years to confirm that numerous glycoproteins reside in both the nucleus and the cytoplasm of the cell. These types of glycoproteins include: (i) Single 0-linked N-acetylglucosamine monosaccharide residues attached to serines or threonines of nuclear and cytosolic proteins [5,6]. (ii) A unique 0-linked glycosyl-phosphomannose glycoprotein in the cytosol[7]. (iii) An a-glucosyl moiety attached to tyrosine of glycogenin, the cytosolic protein primer for glycogen biosynthesis [S]. (iv) Cytosolic 0-linked oligosaccharides containing fucose found in the cytosol of slime molds[9]. (v) 0-linked mannose proteoglycans in the cytosol [ 10,111. (vi) Nuclear localized glycosaminoglycans [ 12,131. In this chapter we will provide a critical review of the biochemical studies that have characterized these nuclear and cytosolic glycoproteins with some added emphasis on the abundance, diversity, dynamics and functional implications of 0-linked GlcNAc. In addition we will review some evidence for other possible forms of nuclear and cytosolic glycosylation that clearly need to be characterized further before they can be safely included in the list above. * Kenneth D. Greis is a UAB/Monsanto Glycoscience Postdoctoral Fellow. Present address: Parke-Davis Pharm. Research, 2800 Plymouth Rd., Ann Arbor, MI 48105, USA.

33

34

2. 0-linked N-acetylglucosamine modified proteins In studies probing the surface glycoconjugates on mouse lymphocytes with bovine milk galactosyltransferase and UDP-[3H]Gal, it was noted that the labeling of proteins was greatly increased when sufficient detergent was added to disrupt the plasma membrane, thereby exposing the cytosolic proteins to the enzyme probe. Further investigation of these cytosolic proteins demonstrated that they were all modified with an N-acetylglucosamine monosaccharide that was 0-glycosidically linked to serine or threonine, thus describing a new form of protein glycosylation known as 0-linked N-acetylglucosamine (0-GlcNAc) [14]. Galactosyltransferase labeling of subcellular fractions from rat liver demonstrated that the 0-GlcNAc modified proteins are greatly enriched in both the nuclear and cytosolic fractions [15], while studies of purified mouse lymphocytes demonstrated the abundance of 0-GlcNAc (1.5 x lo8 molecules/cell in lymphocytes) and refined the localization to include cytosolic and nucleoplasmic compartments exclusively [ 161. A survey of species diversity has shown that 0-GlcNAc modified proteins are present in all eukaryotic cells that were tested, but not in bacteria [4,17]. Since the initial observation and characterization of 0-GlcNAc, only a small portion of the several hundred proteins (Fig. 1) with this modification have been identified (Table 1) and in a few cases, the sites of glycosylation have been mapped (Table 2). These proteins represent a diverse array of functions including: cytoskeletal proteins, enzymes, regulatory proteins, transcription factors and oncogenes, viral proteins, nuclear pore proteins, chromatin proteins, and RNA Polymerase 11. However, all of these proteins share a few notable features including: (i) All 0-GlcNAc proteins are also known phosphoproteins. (ii) All 0-GlcNAc proteins exist as part of a reversible multimeric protein complex based on the state of phosphorylation, thus suggesting that the state of glycosylation may also be involved with complex assembly/disassembly. Several lines of evidence suggest that the addition and removal of 0-GlcNAc to glycoproteins, unlike classical glycosylation pathways in the endoplasmic reticulum and golgi, are dynamic processes much more akin to phosphorylation [18-211. First, mitogen activation of mouse T-lymphocytes results in a rapid and transient change in the levels of 0-GlcNAc on various proteins [21]. The kinetics of these changes are consistent with the hypothesis that 0-GlcNAc additiodremoval may be key in the early stages of T-cell activation. Secondly, in pulse-chase experiments, the turnover rate of 0-GlcNAc on human cytokeratins has been reported to be much faster than the turnover rate of the proteins [18]. Furthermore, recent evidence shows that there is an increase in 0-GlcNAc labeling of keratins 8 and 18 during mitotic arrest, thus suggesting a cell-cycle dependent turnover of 0-GlcNAc additionhemoval [ 191. In another cell-cycle study, the levels of 0-GlcNAc on various cytosolic proteins and known nuclear pore proteins were shown to change dramatically in a cell-cycle dependent manner [20]. Again, the dynamics of 0-GlcNAc during the cell cycle and in response to external stimuli make it a prime candidate as a regulatory modification. A survey of the known sites of attachment of 0-GlcNAc (Table 2) indicates that no clear consensus sequence is necessary for attachment for this modification although many of the sites contain a proline and a valine in close proximity to the glycosylated serine or threonine. However, many of these sites are nearly indistinguishable from those used by

OH-

-

35

NEPHGE GELH+

10 % SDSPAGE

16 h

-3Days

7 Days

30 Days

Fig. I.Two-dimensional NEPHGE/SDS-PAGE Gel of GT-labeled mouse liver nuclei. 5 pg of total mouse liver nuclear proteins were labeled with excess GT and UDP-[3H]Gal as described by Whiteheart et al. [30]. The labeled glycoproteins were then separated by a combination of non-equilibrium pH gel electrophoresis followed by SDS-PAGE as described by O’Farrell [31]. The gels were impregnated with sodium salicylate, dried and the labeled glycoproteins were visualized by fluorography for the indicated times. Protein molecular weight markers are as indicated.

various protein kinase families such as the proline-specific kinases [22,23], thus indicating that 0-GlcNAc may act as a regulatory modification by competing for phosphorylation sites. In fact, some examples studied, such as the CTD of RNA Pol I1 [24] and the human cytokeratins [ 18,191 reciprocal relationships between glycosylation and phosphorylation have been clearly documented. Furthermore, 0-GlcNAc attachment sites exhibit high “PEST” scores[25,26] which may mean that 0-GlcNAc is involved in regulating the targeting of proteins for degradation. Finally, for 0-GlcNAc to be considered a regulatory modification similar in abundance and dynamics as phosphorylation, the necessary enzymes that represent the counterparts of kinases and phosphatases must be present in the appropriate compartment of the cell. In fact, this has been demonstrated to be the case. By using synthetic peptides based on some of the known sites of 0-GlcNAc attachment (Table 2), a cytosolic UDPG1cNAc:polypeptide N-acetylglucosaminyltransferase (0-GlcNAc transferase) activity has been identified [ 171 and purified [27] from rat liver. Furthermore, a cytosolic, neutral B-glucosaminidase activity, distinct from lysosomal hexosaminidases, has also been identified [28] and purified [29] from rat spleen. These two enzymes (and probably many other isoforms not yet described) are likely to represent the analogs of kinases and phosphatases for 0-GlcNAc as a regulatory modification. While direct evidence for 0-GlcNAc as a regulatory modification is still lacking, functional studies of the known 0-GlcNAc-modified proteins and further characterization of the GlcNAc transferase and

36 Table I ldentified proteins modified by 0-GlcNAc Proteins

Reference(s)

Nuclear pore proteins RNA polymerase I1 catalytic subunit Many RNA Pol I1 transcription factors 65 kDa Nuclear tyrosine phosphatase Aplasia neuron proteins

v-Erb-a oncoprotein c-myc oncoprotein Estrogen receptor (murine, bovine, human) p43/hnRNP G canine autoantigen Adenovirus fiber protein HCMV UL32 (basic phosphoprotein) tegument protein Many chromatin proteins of Drosophila

Proteins

Reference(s)

Human erythrocyte band 4.1 & 65 kDa protein Cytokeratins 13, 8, 18 Neurofilaments H, M & L Rotavirus NS26 67 kDa RBC kinase Synapsin I & I1 92 kDa smooth ER protein Bovine Lens a-crystallins (small heat shock proteins) Many schistosome proteins Talin Baculovirus gp4 1 tegument protein Many trypanosome proteins

[Kelly and Hart, unpublished]

glucosaminidase are beginning to address the regulatory functions of this abundant and highly dynamic modification of cytosolic and nuclear proteins. In the following sections a survey and discussion of present literature on 0-GlcNAc is presented in an attempt to show the diversity and some potential functional roles of this modification. 2.1. The enqmes of 0-GlcNAc cycling

By using synthetic peptides based on some of the known sites of 0-GlcNAc-modification as acceptors (Table 2), a UDP-G1cNAc:polypeptide N-acetylglucosaminyltransferase (0-GlcNAc transferase) activity has been identified [ 171 and purified to apparent homogeneity [27] from the cytosol of rat liver. Some of the properties of this enzyme include: (i) The active site of the enzyme has been localized to the cytosol based on latency studies. (ii) UDP-GlcNAc has been demonstrated to be the nucleotide sugar donor with a Km of about 545nM, while UDP, UTP and UDP-GlcNAc, and to a 100-fold lesser extent, UMP and UDP-GalNAc have been shown to inhibit the transfer from UDP-[3H]GlcNAc. (iii) The enzyme appears to contain 2 subunits with an M , of 110 (a-subunit) and 78 kDa (fi-subunit) after purification of greater than 30 000 fold. (iv) The holoenzyme is very large, based on gel filtration and sedimentation data, with an apparent molecular weight of 340kDa, thus suggesting a heterotrimer of a2@ configuration. (v) Photoaffinity labeling studies with 4-[f~~~P]thio-UDP indicated that the

37 Table 2 Identified sites of 0-GlcNAc modification Protein

Reference(s)

Glycosylated peptide a

With a PV(S/T) motif (underlined)

Human erythrocyte 65 kDa Bovine Lens a-A-crystallin Human Serum Response factor Rat Neurofilament (NF-L) HCMV (UL32) BPP Rhesus monkey a-B-crystallin Neurofilament (NF-M) from rat spinal cord

. . . DmSQPSLVGSK . . .

. . . 15gDIpVSREEK166.. . . . 3 0 2 Y L A P V S A S V S m A 3 1.8 . .43YSAPVSSSLSVR54. . . 9'4PPSVPVSGSAPGR927. . ... ' 6 4 E E K w A A P K ' 7 4 . . . . .44GSPSTVSSSYK54.. .; . . .427QpSVTISSK435 .. '

'

Without PV(S/T) motif

Talin from chicken gizzardb RNA Pol I1 from calf thymusC Human Serum Response factor Rat Nuclear Pore p62 Human Erythrocyte Band 4.1 Rat Neurofilament (NF-L) HCMV (UL32) BPP a

. . .'475MAXQNLVDPAXTQ1488.. .; ' . .'886NQLTNDYGQLAQQ1889'.' . (S/T)P(S/T)SP . . . TPTSPN . . . SPTSPT . . . .; . . .269VTNLPGTTSTIQTAPSTSTT289. . . .280TQTSSSGTVTLPATIM395 .. . . . . MAGGPADTSDPL . . AQTITSETPSSTT . . . ' . . '8YVETPRVHISSVR30 ' . . . .935STTPTYPAVTTVYPPSSTAK955 .. '

' '

'

Sites of 0-GlcNAc attachment are indicated by the boldface S/T. Numbers correspond to mouse sequence. Glycosylation is spread throughout the repeat sequence of the CTD.

a-subunit was likely to contain the active site. Specificity of binding by 4-thio-UDP to the active site was confirmed by competition with cold UDP. While this is the only 0-GlcNAc transferase enzyme characterized to date, the numerous variations of attachment sites on known proteins (Table 2) and the appearance of other minor pools of activity from the purification protocol [27] suggest that other isoforms of this enzyme may be present in the cytosol and perhaps the nucleoplasm. Clearly, molecular biological approaches can address this question once the gene for the 0-GlcNAc transferase is identified. A neutral 0-GlcNAc-specific, b-D-N-acetylglucosaminidase (0-GlcNAcase) activity from rat spleen cytosol has also been identified [28] and recently purified and characterized [29]. This 0-GlcNAcase activity was distinguished from lysosomal hexosaminidases by its neutral pH optimum (pH 6.4) and by the inability to inhibit this enzyme with GalNAc or GalNAc analogs. Some of the other properties of this enzyme include: (i) A cytosolic active site as judged by subcellular fractionation and latency studies. (ii) The enzyme appears to consist of 2 subunits of M , 54 kDa (a-subunit) and M , 5 1 kDa @-subunit) of equal mass after purification of over 22 000-fold from rat spleen cytosol. (iii) The native enzyme activity sediments at M , 106 kDa in sucrose gradients suggesting a heterodimer holoenzyme complex in an a0 configuration. (iv) A photoaffinity analog

38

of GlcNAc specifically labels the a-subunit and can be competed away with l-aminoGlcNAc, thus suggesting that the catalytic domain of this enzyme resides on the a-subunit. (v) The enzyme activity can specifically remove 0-GlcNAc from peptide substrates that had been glycosylated with the 0-GlcNAc transferase with a 6-fold higher relative activity than Diplococcus pneumoniae hexosaminidase. (vi) The enzyme activity can also be effectively inhibited by various analogs of GlcNAc including l-aminoGlcNAc, 1-azido-GlcNAc, LOGNAC (ki = 1.7 pM) and PUGNAC ( K , = 52 nM). It is likely that the excellent inhibitory activities of these analogs will be important tools when studying the dynamics of 0-GlcNAc on various cellular proteins. 2.2. Nuclear proteins that contain 0-GlcNAc

Many nuclear proteins have been demonstrated to be modified with 0-GlcNAc (Table 1). However, in a typical mouse nuclear extract, labeled with galactosyltransferase and UDP[3H]Gal (Fig. I), many hundreds of unidentified proteins are modified by 0-GlcNAc. In fact, after a long exposure to detect the lower abundance proteins, the entire two-dimensional gel is covered with labeled proteins. The following is a description of some of the known nuclear proteins that bear 0-GlcNAc with a discussion of how 0-GlcNAc may be involved with these various protein functions.

2.2.1. Nuclear pore proteins Galactosyltransferase (GT) labeling of nuclear fractions of rat hepatocytes has demonstrated that the nuclear envelope is particularly enriched with 0-GlcNAc bearing proteins [ 151. Further studies with monoclonal antibodies that recognize a distinct group of nuclear pore proteins (nucleoporins) [32] has demonstrated that these protein were indistinguishable from GT labeled nuclear envelope proteins as judged by comparing western blots and autoradiographs [33]. Interestingly, removal of the 0-GlcNAc from these nucleoporins with hexosaminidase or extension of the monosaccharide with saturating amounts of GT and UDP-Gal, resulted in complete abolition of the binding of all of the anti-nucleoporin monoclonal antibod-ies, thus indicating that the 0-GlcNAc was a necessary constituent of the epitope for these antibodies. Nuclear pore reconstitution studies have demonstrated that nucleoporins are required for pore-mediated nuclear transport (for reviews see refs. [34,35]). The functional significance of the nuclear pore glycoproteins was suggested in microinjection studies which showed that wheat germ agglutinin (WGA), a plant lectin with specificity to terminal GlcNAc residues (includes 0-GlcNAc) could prevent import of proteins to the nucleus [36]. Furthermore, antibodies to the nucleoporins that had 0-GlcNAc as part of their epitope [33,37] when microinjected into Xenopus oocytes, could also prevent both import of proteins and export of RNA from the nucleus [38]. Steric problems associate with such a large lectin (or antibody) were ruled out since WGA appeared to block only protein translocation with little effect on binding of proteins to the pore complex [39,40]. In addition, when the 0-GlcNAc modified nucleoporin proteins are removed from Xenopus oocyte extracts by WGA-affinity chromatography, transport of proteins through the reconstituted pores was abolished even though the morphology of the pore complexes appeared unchanged [41]. Remarkably, addition of

39

rat nuclear pore glycoproteins to the glycoprotein-depleted Xenopus oocyte extracts resulted in the reconstitution of nuclear pores that were completely transport competent. Furthermore, the 0-GlcNAc modifications of the WGA-extracts could be covered by Gal with GT with no observed effect on nuclear pore assembly or nuclear transport, thus arguing against a lectin-like interaction involved in nuclear pore assembly [42]. Unfortunately, these reconstitution experiments did not include a pretreatment of the isolated glycoproteins with hexosamindase to remove the 0-GlcNAc to determine directly whether the 0-GlcNAc modification on these nucleoporins was necessary to restore transport activity to the glycoprotein-depleted nuclear extracts. These experiments do, however, demonstrate that the 0-GlcNAc modified proteins were needed for poremediated transport and that these nucleoporins are functionally conserved between amphibians and mammals. The most abundant and best characterized of the nucleoporins is p62. This glycoprotein has been cloned from various sources[4345] and shown to be modified by at least ten 0-GlcNAc monosaccharides [33]. In addition, the glycosylation of p62 has been localized in clusters in the amino terminal half of the protein [45] and several sites of glycosylation have been mapped (Table 2) [43,45]. Recently, p62 was reported to be sequentially glycosylated by an 0-GlcNAc transferase in reticulocyte lysates [46]. By using recombinant p62 and polypeptide fragments of p62 expressed in E. coli in an in oitro glycosylation system (reticulocyte lysates), it was reported that the region between amino acids 248-341 of the mouse p62 were preferentially glycosylated followed by low affinity glycosylation over the remaining amino terminal region [46]. Whether a similar high affinityllow affinity addition of 0-GlcNAc occurs in vivo remains to be addressed. Finally, it remains unclear whether 0-GlcNAc is directly involved with regulation of nuclear transport. However, in light of recent evidence that the levels of 0-GlcNAc on several nucleoporin proteins change dramatically in a cell cycle dependent manner (particularly during mitosis when the nuclear envelope disassembles and reassembles) [20], 0-GlcNAc is likely to play a role in this highly regulated process.

2.2.2. Chromatin proteins, transcription factors and RNA polymerase II While it is evident that the nuclear pore proteins are highly modified by 0-GlcNAc, far more glycosylated glycoproteins are found in chromatin [ 151. In fact, WGA binding and GT labeling studies of Drosophila embryo polytene chromosomes have shown that 0-GlcNAc is highly abundant along the entire length of the chromosomes [47]. Furthermore, the 0-GlcNAc appears to be particularly concentrated in condensed regions of the chromatin resulting in a banded pattern, while little 0-GlcNAc appeared to be present in "puff" regions which have been associated with areas of active transcription. The involvement of 0-GlcNAc modifications in transcriptional activity becomes even more intriguing in light of the fact that RNA polymerase I1 (RNA Pol 110) [24] and all of its transcription factors thus far investigated [48-511 have been shown to be modified by 0-GlcNAc. While no direct evidence has shown that 0-GlcNAc regulates transcriptional activity, glycosylated transcription factor Sp 1 has been reported to be more transcriptionally active than the non-glycosylated form [52]. Glycosylated Sp 1 appears to play a role in insulin-dependent stimulation of growth factor TGFa [53] and Spl appears to be differentially glycosylated in a cell-type specific manner [54].

40

The catalytic subunit of RNA Pol I1 contains a highly conserved domain at the carboxyl-terminus consisting of up to 52 repeat units with consensus sequence: -(TyrSer-Pro-Thr-Ser-Pro-Ser)- (for a review see [55]). Although this region of the enzyme does not contain the RNA polymerase activity, genetic analysis has demonstrated that it is required for cell viability. Furthermore, various forms of the RNA Pol I1 subunit have been demonstrated in vivo based on the phosphorylation state of the CTD and on mobility by SDS-PAGE. These include 11, ( M , 240kDa), which is highly phosphorylated over the entire repeat region of the CTD and is associated with the transcription complex during the transition from initiation to elongation [56], and 11, ( M , 215 m a ) , the nonphosphorylated form found in the preinitiation complex [57].Recently, the CTD of II,, but not of II,, has been shown to be extensively modified by 0-GlcNAc over the entire conserved repeat region [24]. Thus, the phosphorylation and glycosylation of the CTD appear to be mutually exclusive events suggesting a high degree of regulation. It appears likely that the 0-GlcNAc modified form (Ha) may be involved in the formation of the initiation complex, possibly by oligomerization with 0-GlcNAc modified transcription factors via lectin-like proteins, followed by deglycosylation and phosphorylation (11,) to begin elongation. These intriguing possibilities are presently under investigation. Recent evidence has shown that transcription factor-like oncogenes including c-myc [58] and v-Erb-a[59] are also modified by 0-GlcNAc. While this is not surprising in light of the fact that all other RNA Pol I1 transcription factors tested have been shown to contain 0-GlcNAc, but it does lead into the interesting questions of how 0-GlcNAc may be involved in cancer development and gene regulation.

2.2.3. Estrogen receptor, Aplasia 83 kDa protein and autoantigen p43 Several other nuclear proteins have been demonstrated to contain 0-GlcNAc modifications (Table l), including bovine and mouse estrogen receptor [60], Aplasia neuron 83 kDa protein [6 11 and canine autoantigen p43 [62,63]. For the estrogen receptor, considerable evidence is now available that phosphorylation may be crucial to the regulation of estrogen-responsive promoters [64]. Here again, like other 0-GlcNAc bearing proteins, the glycosylation may act as an antagonist to phosphorylation and thereby also have a role in gene expression. It is also worth noting that the 83 kDa protein from Aplasia neurons is prominent in both the axon and the nucleus [61]. Given that a pathway was recently discovered in neurons that can transport proteins from the axon to the nucleus [65], it will be interesting to know whether this glycoprotein is involved with the transport mechanism. However, to date there is no direct evidence to support such an idea.

2.3. Cytoplasmic proteins that contain 0-GlcNAc

The cytosol of eukaryotic cells has been shown to be enriched in proteins modified by 0-GlcNAc [15]. However, to date only a few of these proteins have been identified (Table 1) and characterized. In this section, a survey of some of the cytosolic proteins that contain 0-GlcNAc is presented along with some interesting potential functional aspects.

41

2.3.1. Cytoskeletal glycoprotein Many of the cytosolic 0-GlcNAc modified proteins that have been identified are components of the cellular cytoskeleton. The first to be characterized was human erythrocyte Band 4.1 [66]. This protein is involved in maintaining the unique shape of erythrocytes by anchoring actin and spectrin to the cytoplasmic tail of glycophorin. While preliminary studies suggested that only the glycosylated forms of Band 4.1 binds to the cytoskeleton, additional controlled studies are needed to confirm this finding. Human cytokeratins are a class of intermediate filaments found primarily in epithelial cells (for a review, see ref. [67]). An interesting feature of cytokeratins is their tissuespecific expression of unique polypeptide pairs that together form the intermediate filaments. Recently, cytokeratins 13 [68], 8 and 18 [18] were shown to contain 0-GlcNAc. Both phosphorylation and 0-GlcNAc modification also appeared to be enriched during mitotic arrest [ 191. The exact function of these intermediate filaments, or their 0-GlcNAc for that matter, remains unclear. However, there is evidence that these filaments assemble and disassemble based on their level of phosphorylation [69]. Thus the relationship between phosphorylation and glycosylation could provide some insight into the function of these intermediate filaments. In another study, neurofilaments from rat and mouse spinal cord have been shown to be multiply glycosylated. Interestingly, the 0-GlcNAc modifications were localized primarily in the head domain of both neurofilaments M and L [70]. The head domain is a region of the proteins that had already been implicated as being required for proper neurofilament assembly by deletion analysis [71,72]. Site directed mutagenesis of the sites of 0-GlcNAc modification are presently underway to address whether the glycosylation is required for assembly of these filaments. Finally, talin has also been reported to be modified by 0-GlcNAc [73]. This important cytosolic protein appears to provide a bridge between the cytoplasmic domain of integrins and the cytoskeleton by interaction with another cytosolic protein, vinculin. Interestingly, non-glycosylated talin derived from platelets does not interact with vinculin, suggesting that 0-GlcNAc may be necessary for the interaction of these cytoskeletal components [731. 2.3.2. Eukaryotic initiation factor 2-associated 67 kDa polypeptide (p") Protein synthesis can be inhibited at the initiation step by phosphorylation of the a-subunit of eukaryotic initiation factor 2 (eIF-2) by one or more eIF-2 kinases (for reviews, see refs. [74,75]). These kinases are believed to be activated in response to the physiological states of the cell (for example, in the absence of hemin under starvation conditions or the presence of double-stranded RNA during viral infection). Activation of the kinases would then ultimately result in the inhibition to protein synthesis. Recently a 67kDa 0-GlcNAc-modified glycoprotein has been described that can bind to eIF-2 and protect it from phosphorylation by eIF-2 kinases and thereby maintain protein synthesis initiation [76,77]. When p67 was removed from the cell extracts with WGA (a lectin which binds 0-GlcNAc) or with antibodies to p67, phosphorylation was no longer blocked and the kinases could readily transfer phosphate to the a-subunit of eIF-2 [78]. By using polyclonal antibodies to detect all p67 proteins, and a monoclonal antibody that reacts only with glycosylated p67, they showed that during serum starvation of hepatoma cells,

42

p67 was first deglycosylated within 10 h, then degraded after 16 h of serum starvation, while the levels of the a- and (J-subunits of eIF-2 remained constant. They concluded that p67 protected the a-subunit from phosphorylation under normal conditions, but during starvation, p67 is rapidly deglycosylated and degraded, allowing the eIF-2 kinases to phosphorylate the eIF-2 and prevent protein synthesis initiation [78]. Recently, this group reported that reticulocyte lysates contain a “deglycosylase” (0-GlcNAcase) that remains in a latent form in the presence of hemin; however, when hemin is absent, the deglycosylase is activated to remove 0-GlcNAc from p67 and begins the cascade toward inhibition of protein synthesis [79]. This study therefore provides evidence for a direct role of 0-GlcNAc in protein synthesis initiation. There are several possible mechanisms by which p67 might regulate phosphorylation of eIF-2. One possibility is that the O-GlcNAcmodified p67 binds to eIF-2 via some interaction with the 0-GlcNAc to prevent eIF-2 kinase activity. Upon activation of the deglycosylase, the 0-GlcNAcs are removed and the binding to eIF-2 is disrupted. p67 could then be targeted for degradation by kinases that could not phosphorylate p67 when it was glycosylated and bound to eIF-2. Clearly, continued studies of this cascade involving an 0-GlcNAc-modified protein may lead to a better understanding of the functions of 0-GlcNAc on this and other proteins. 2.4. Viral proteins

0-GlcNAc-modifications have been demonstrated on a number of viral proteins (Table 1). The function of 0-GlcNAc on these proteins is not yet known; however, the fiber proteins of adenovirus are known to form mature trimeric structures that are involved in virus attachment to the host cell surface [80]. Interestingly the 0-GlcNAc in the mature trimeric structures is inaccessible to labeling with GT unless the fibers are denatured with detergents indicating that the 0-GlcNAc moieties are buried in the trimeric structures [81]. A similar observation has been made for neurofilament assembly (see section 2.3.1), suggesting that 0-GlcNAc may also be involved with the formation of these multimeric fiber structures. Another interesting finding is that 0-GlcNAc is found on the major tegument proteins (the region between the viral capsid and the viral envelope) of human cytomegalovirus (HCMV) [82] and baculovirus [83]. The function of the basic phosphoprotein (BPP, UL32) of HCMV is presently unknown, but its location in the tegument region of the virus suggests that it might act as a signal for final envelopment of the capsid. In this capacity, 0-GlcNAc could be the signal for oligomerization of the BPP or the means of attachment to the cellular compartment used for viral envelopment. The demonstration that gp41 of baculovirus contains 0-GlcNAc [83], made it clear that insect cells are fully capable of adding this modification to proteins. Furthermore, this finding has opened the door to using the baculovirus overexpression system to study the glycosylation of many low abundance proteins that contain 0-GlcNAc including transcription factors and oncogenes. Interestingly, the BPP from HCMV appears to be glycosylated at the same sites when expressed in insect cells as a recombinant baculovirus or when isolated from HCMV virions, although the stoichiometry at each site appeared to be slightly different [ 1511. These results were also consistent with those found for human cytokeratins 8 and 18 expressed in baculovirus [84].

43

3. Other cytosolic glycoproteins 3.1. Phosphoglucomutase and parafusin

Recently, Marchase and colleagues have demonstrated the existence of an 0-linked mannose specific a-glucose- 1-phosphotransferase in rat liver homogenates [96]. The predominant acceptor for this transferase in rat liver is a 62 kDa glycoprotein [97]; however, acceptors of similar apparent molecular weight have also been found in a variety of vertebrate tissues and species [97,98], in Paramecium tetraurelia [99], and in yeast [ 1001, but not in bacteria [96]. The Glc-phosphotransferase activity was localized to microsomal membranes by subcellular fractionation; however, the active site of this enzyme was shown to be on the cytosolic face of this membranous fraction by the following criteria: (i) Unlike lumenal enzyme markers (mannose-6-phosphate and ~-1,4-galactosyItransferases)the maximal Glc-phosphotransferase activity was present in the absence of detergent disruption of the vesicles. (ii) The Glc-phosphotransferase activity could also be degraded by exogenous trypsin or pronase while the lumenal marker enzymes required detergent solubilization for protease degradation. (iii) Endogenous labeling of the 62 kDa acceptor protein with the 35S-labeled 0-phosphorothioate analog of UDP-Glc showed that 85% of the acceptor was found in the high speed supernatant while in the same preparation 94% of the CMP-[3H]NeuAc-labeled glycoproteins remained in the microsomal pellet [96]. Additional studies of the cytosolic acceptor molecule in yeast [loo] and rat liver [7] identified the 62 kDa proteins as phosphoglucomutase - a key enzyme in the maintenance of the equilibrium between glucose utilization for energy and synthesis of glycogen. Interestingly, the yeast transferase activity could modify the liver acceptor and vice versa [ 1001, thus indicating evolutionary conservation of this enzyme and acceptor system. Working independently, Satir’s group reported the existence of a 63 kDa cytosolic phosphoglycoprotein termed parafusin isolated from the ciliated protozoan Paramecium tetraurelia [ 1011. This protein, like phosphoglucomutase, was shown to be the primary acceptor for the Glc-phosphotransferase enzymes from both Paramecium and rat liver [99]. Previous studies of parafusin have demonstrated a very rapid (less than 1 s) and transient dephosphorylatiodrephosphorylation cycle during an external stimulusdependent, cytosolic calcium-mediated stimulation of secretion [ 102,1031, thus suggesting regulation by phosphatases and kinases. However, recent evidence suggests that the phosphate turnover is due to removal of the a-glucose- 1-phosphate from parafusin and the rapid replacement by Glc-phosphotransferase and UDP-Glc [99,104]. Consistent with a regulatory glucose-phosphate turnover system, the presence of a soluble glucose- 1-phosphate phosphodiesterase activity was reported and characterized in rat liver [ 1051. This enzyme was shown to selectively remove a-glucose-I-phosphate from phosphoglucomutase, but exhibited no phosphodiesterase activity on UDP-Glc or glucosylphosphoryldolichol. Finally, studies of the calcium dependent turnover of glucose- 1-phosphate on phosphoglucomutase in rat cortical synaptosomes and in PC- 12 cells suggest that the incorporation of phosphodiester-linked Glc is dependent on intracellular calcium levels [ 1061. Depolar-

44

ization of synaptosomes with potassium in the presence of 2mM calcium resulted in a two-fold increase in the incorporation of [I4C]-Glc into phosphoglucomutase within 5 s. The level of glucose-P incorporation returned to baseline levels within 25 s. Similarly, cells “loaded” with the 35S-labeled 6-phosphorothioate analog of UDP-Glc by freezelthawing in the presence of cryoprotectants showed an increased incorporation of Glc-P into phosphoglucomutase upon depolarization with potassium. These findings are consistent with those reported in Paramecium [99,104] and describe another cytosolic glycosylation event, like 0-linked GlcNAc, that is dynamic and responsive to external stimuli [ 18,211. While the function of the Glc-P transfer remains unresolved, such rapid changes in glycosylation in response to external stimuli may be analogous to other post-translational modifications such as phosphorylation that regulate the function of the acceptor proteins. 3.2. Glycogenin

Glycogen is the major macromolecular storage structure of glucose in many cells. Since the first description of glycogen phosphorylase in 1939 [I071 and glycogen synthase in 1957 [log], it had been postulated that biosynthesis of glycogen likely required a “primer” molecule to initiate glucose oligomerization. Krisman and Barrengo [ 1091 first detailed a working model of an acceptor protein for glycogen synthesis that resembled a proteoglycan in structure with multiple glycogen molecules extending from a protein backbone. Assays to determine whether purified rabbit skeletal muscle glycogen contained any protein backbone by extensive enzymatic removal of the glycan with a-amylase and amyloglucosidase [ 1101 or by chemical removal in anhydrous HF [I 1 I] demonstrated the existence of a 37-38 kDa protein termed glycogenin [ 1 1 I]. Glycogenin from rabbit skeletal muscle has been shown to exist as a heterodimer with glycogen synthase in a 1 : 1 molar ratio [ 1121 and to contain a glucosyl-I-0-tyrosyl linkage at tyrosine-194 [I 13,1141. Recent studies have demonstrated a priming mechanism of glycogenin necessary for glycogen biosynthesis with the following properties (for a review, see ref. [8]): (1) Glycogenin is glucosylated only at tyrosine-I94 by an unknown glucosyltransferase. (2) Glucosylation of Tyr-I94 is extended by an autoa- 1,4-glucosyltransferase activity in the presence of UDP-Glc and Mn2+ resulting in an average of 5.5mol of Glc (with a maximum of 8) per mole of glycogenin. (3) The “primed” glycogenin can then be elongated by glycogen synthase to produce glycogen. To address whether the first glucose residue added to Tyr- 194 was an autoglucosylation event, Roach and coworkers isolated the cDNA of rabbit muscle glycogenin [I 151 and expressed it in E. coli [ 1161. However the bacterial expressed recombinant proteins already contained glucose attached to Tyr- 194, thus indicating that either the bacteria contain the necessary glucosyltransferase or the first glucose and subsequent “priming” glucoses are added by autoglucosylation. Expression and purification of recombinant proteins with Tyr-194 changed to either Phe or Thr resulted in non-glucosylated glycogenin which unlike the wild type recombinant protein could not autoincorporate Glc or act as a substrate for glycogen synthase [ 1 171. While these studies were inconclusive concerning the addition of the initial glucose to glycogenin, they did confirm that Tyr-194 was the only site of glucosylation and that a minimal chain length of glucose

45

(about 8) generated by autoglucosylation was necessary for glycogenin to be a substrate for glycogen synthase. In studies to determine the role of glycogenin in the regulation of glycogen synthesis, Roach and colleagues transiently coexpressed glycogenin and glycogen synthase in COS cells and assayed for function and association at various extracellular glucose concentrations [ 1181. First, they provided some evidence for a redistribution of glycogen synthase into the soluble fraction when glycogenin was available as a substrate. This redistribution of the enzyme could be an early initiation step in glycogen synthesis. They further reported that at low Glc concentrations (0-1 5 mM), glycogenin was not fully primed with glucose and hence could not be elongated by glycogen synthase, while cells grown in the presence of more than 20mM Glc, expressed glycogenin that was fully primed and elongated by glycogen synthase. These results may indicate that the priming of glycogenin is regulated by glucose levels as a means of preventing glycogen synthesis when glucose is not readily available. Since it appeared that glucose levels might be an important factor in the autoglucosylation of glycogenin, the effect of phosphorylase, another known regulatory protein in glycogen metabolisdcatabolism, was assayed for its effect on glycogenin function [ 1 161. Purified recombinant glycogenin was incubated with UDP-[14C]glucose to allow autoglucosylation to occur, thus forming the primed glycogenin. However, in the presence of purified rabbit muscle phosphorylase a and inorganic phosphate, up to 70% of the I4C was converted to a single product of Glc-l-phosphate while the remaining glycogenin was a much less effective substrate for glycogen synthase. They concluded that phosphorylase a removed glucose from the primed glycogenin by phosphorolysis in a manner similar to that described for the degradation of the a-1,4-glucose units of glycogen (for reviews, see refs. [I 19,1201). These results along with those demonstrating a role for glucose levels on glycogenin priming and degradation, depict a whole new level of regulation of glycogen metabolism in the cell. Clearly this newly characterized cytosolic glycoprotein will require further studies to elucidate the origin of the tyrosine specific glucosyltransferase activity and the effects of other characterized regulatory cascades of glycogen synthesis (such as insulin and glucagon) on the priming and removal of the glucose units from glycogenin. 3.3. Cytosolic fucosylation

Recently a rather intriguing report described a novel cytosolic fucosylation pathway in Dictyostelium discoideum [9] in which both a 21 kDa acceptor protein and a fucosyltransferase activity were localized to a cytosolic fraction. In order to increase metabolic labeling, this group took advantage of a mutant strain of Dictyostelium (HL250) which was defective in the ability to convert GDP-Man to GDP-fucose and therefore required fucose supplemented into the medium for any cellular fucosylation to be detected. Metabolic labeling of wild type or HL250 mutant cells with [3H]-fucose followed by subcellular fractionation of these cells produced a major labeled glycoprotein with an apparent molecular weight of about 21 kDa (designated FP21) in the 100 OOOx g supernatant. Lysosomal and Golgi contamination of the cytosolic fraction were judged to be less than 7% by acid phosphatase activity and about 5% by GlcNAc transferase activity,

46

respectively. Western blots of spore coat proteins confirmed the integrity of secretory vesicles since these proteins were resistant to proteolysis to Proteinase K until 0.1 % Triton X-100 was added to the membranes. Most convincingly however, was the inability to detect significant amounts (less than 0.7%) of labeled FP21 in the microsomal fraction after sonication to disrupt the integrity of the microsomal vesicles. The fucosylated glycan of FP21 was shown to have the following properties: (i) The glycan was resistant to release by PNGase F but could be completely release by alkaline @-eliminationthus suggesting an 0-linked structure. (ii) The released glycan eluted from a Bio-Gel P-4 gel filtration column at a position of 4.8Glc units. (iii) Anion exchange chromatography on QAE-Sephadex indicated an acidic glycan with 2 negative charges that was resistant to bacterial alkaline phophatase. (iv) The labeled fucosyl residue(s) could be partially released by bovine kidney a-L-fucosidase. The origin of the fucosyltransferase activity was assayed in the cytosolic fractions in the presence of GDP-[3H] fucose and endogenous FP21 protein from the HL250 mutants cells grown in the absence of fucose. A fucosyltransferase activity was detected that could label endogenous FP2 1 in a time, concentration, and divalent cation-dependent fashion with a Km for GDP-fucose of 1.4pM. This activity could also be detected in wild type cells but only in the presence of exogenous, non-fucosylated FP21 from mutant cells thus suggesting that the acceptor protein in wild type cells was already fully fucosylated. Other properties of this enzyme included a broad pH optimum of 6.58.0 with GDP-fucose as the donor and the ability to use Gal(@l-3)GlcNAc@l-Rbut not Gal(~l-4)GlcNAc~l-R, GalPl-R (where R is 8-methoxycarbonyloctyl-) nor other a or @ monosaccharides as a substrate. While this data is intriguing, additional studies and characterization of the structure of the acceptor glycan, the transferase activity and the diversity of this type of modification in other species are needed to begin to understand its functional significance. Other reports of fucosylation have suggested the existence of proteins that contain fucose residues attached directly to a serine or threonine. This linkage was first described by the presence of amino acid fucosides in human urine [ 1211 then as components of rat tissue extracts [ 1221 and rat glycoproteins [ 1231. Release of the glycans by alkaline @-eliminationrevealed the presence of both fucitol and Glc(@1-3)fucitol[1231. Further examination of these fucosylated glycoproteins in soluble and membrane fractions from rat cells suggests the existence of both cytosolic and membrane associated 0-linked fucosylated proteins [ 1241, but the purity of the cytosolic fractions was never examined. Recently, many reports have demonstrated the existence of 0-linked fucose on epidermal growth factor (EGF)-like domains on many plasma proteins (for a review, see ref. [125]) as well as in CHO cell Lecl mutants [126,127]. However, to date none of these modifications have been conclusively demonstrated for cytosolic or nuclear glycoproteins. 3.4. Cytosolic proteoglycans

About twenty years ago, the Margolis group reported the existence of soluble proteoglycans and glycosaminoglycans consisting primarily of chondroitin sulfate in the nervous tissue [ 128-1 301. This material was distinguished from extracellular matrix or

41

cell surface material since it was resistant to protease digestion of intact cells and was not detected during the isolation of the cells. Subsequent studies of cytosolic and particulate fractions via hypotonic lysis of rat neurons showed that while only 20% of the cellular proteins were present in the cytosolic fraction (as judged by release of 90% of the lactate dehydrogenase activity), 82, 55 and 25% of the chondroitin sulfates, heparan sulfates and hyaluronate, respectively, were soluble [ 1311, thus suggesting a cytosolic localization. Structural characterization of these proteoglycans showed that as much as 50% of the oligosaccharides were linked to the protein by a novel O-linked mannose and that some of the species were sialylated [ 10,l 1,1321. Interestingly, the proportion of cytosolic to extracellular chondroitin sulfate proteoglycans in rat neurons increased during brain development [ 1331. At 7 days postnatal development all of the chondroitin sulfates were extracellular as judged by immunohistochemical staining. By day 2 1, significant cytosolic staining was evident and by 33 days the distribution was indistinguishable from that of adult brain neurons. Studies of hyaluronic acids in developing rat cerebellum demonstrated a similar developmental change from an extracellular to a cytosolic localization [ 1341. Here, all of the detectable hyaluronic acid proteoglycans were extracellular until 2 1 d postnatal development, when significant amounts could be detected in the cytoplasm. From all of the above studies, however, it remains unclear whether the proteoglycans were redistributed or resynthesized during development. Current models of proteoglycan biosynthesis (for a review, see ref. [ 1351) would suggest a redistribution of extracellular matrix material; however, this does not preclude the possibility of a cytosolic biosynthesis pathway in brain cells. Furthermore, the functional significance of these changes or redistribution during development remains obscure.

3.5. N-linked GlcNAc In studies of the glycosylation of the a-subunit of the plasma membrane sodiudpotassium ATPase in right-side-out vesicles from canine kidneys, Kaplan and colleagues [ 1361 have reported the presence of a PNGase F sensitive oligosaccharide that was only accessible to labeling with galactosyltransferase and UDP-[3H]Gal in the presence of detergent, thus suggesting a cytoplasmic orientation of an N-linked glycan with a terminal GlcNAc. The authors note that the vesicle orientation and integrity was at least 85% rightside-out as judged by latency of Na+,K+-ATPase activity, but the galactosyltransferase labeling of the a-subunit increased by 4-%fold in the presence of detergent. This cytosolically oriented glycan was distinguished from O-linked GlcNAc (see section 2 of this chapter) by sensitivity to PNGase F digestion and resistance to chemical hydrolysis by alkaline &elimination. Further characterization of the galactosyltransferase labeled, PNGase F-released glycan demonstrated the existence of a disaccharide that comigrated with lactosamine on a Bio-Gel P2 column and by thin layer chromatography [137]. The authors therefore conclude that this modification is most likely a single N-linked GlcNAc residue attached to an asparagine on a cytosolic loop of the a-subunit. However, further characterization of this glycan by HPAE-HPLC or high-voltage paper electrophoresis with all the appropriate disaccharide standards is needed to demonstrate that it is indeed [3H]Ga1(@1-4)GlcNAc.Furthermore, confirmation of the site(s) of glycosylation and

48

demonstration that this region of the proteins is indeed located in the cytosol (not just predicted by primary sequence) of intact cells are a prerequisite to future studies directed toward understanding the origin and function of this novel glycosylation event.

4. Other nuclear glycoproteins 4. I . Glycosaminoglycans

Many studies over the past twenty years have suggested the existence of proteoglycans and glycosaminoglycans in the nucleus (for a review, see ref. [4]). However many of these studies did not effectively address the presence of contaminating subcellular fractions in their nuclear preparations. In one particular study there was unambiguous evidence for the presence of unique heparan sulfate (HS) glycosaminoglycans isolated from the purified nuclei of a rat hepatocyte cell line [138]. The amount of this nuclear HS appeared to be dependent on the growth state of the cells and was shown to account for about 5 4 % of the total cellular incorporation of [35S]04 into all cell-associated HS. It has recently been reported that the accumulation of these nuclear HS fragments may be important in growth regulation, DNA synthesis and arrest in the GI phase of the cell cycle [139,140]. The extent of contamination of the purified nuclear fractions was addressed by adding labeled proteoglycans and glycosaminoglycan fragments to non-labeled samples during various stages of the purification of the nuclei to determine whether any [35S] became associated with the nuclear fractions [ 1381. In these studies, they consistently showed a nuclear recovery of <0.5% of the total [35S] added to the preparations, thus indicating minimal contamination to their purified nuclei. Interestingly, the structural features of the nuclear HS were distinct from those found on the cell surface or in other subcellular compartments, including an unusual sulfated glucuronosyl residue instead of a sulfated iduronic acid. Furthermore, pulse-chase studies indicated that the nuclear HS originated from a small fraction of the cell-surface proteoglycans by degradation of the proteoglycans and transport of HS to the nucleus via a lysosomal-independent pathway, since this process was not inhibited at 16°C or in the presence of sufficient ammonium chloride to disrupt lysosomal function on transport [13]. These studies were unable to address the mechanism by which the sulfated glucuronosyl residue was generated. In contrast to the above studies, the internalization and metabolism of cell-surface HS proteoglycans in rat ovarian granulosa cell cultures has been well documented and shown to be entirely within the lysosomal compartment, resulting in complete digestion of the HS glycosaminoglycans into their individual monosaccharides (for a review, see ref. [ 1411). Consistent with these findings, nuclei from rat ovarian granulosa cell cultures appear to contain no HS glycosaminoglycans but rather appear to contain dermatan sulfate (DS) glycosaminoglycan fragments from cell surface DS-proteoglycans [ 121. It remains unclear whether these nuclear DS-glycosaminoglycan fragments, like HS in hepatoma cell lines, are involved in growth regulation or DNA synthesis. 4.2. High mobility group proteins

High mobility group (HMG) proteins, the largest class of nonhistone proteins found in the nucleus, are often found in association with regions of active transcription in chromatin

49

(for reviews, see refs. [142,143]). These proteins have been reported to contain various post-translational modifications, including methylation, acetylation, phosphorylation, ADP-ribosylation and glycosylation. Direct compositional analysis of purified HMG 14 and HMG 17 demonstrated the presence of N-acetylglucosamine, mannose, glucose, galactose and fucose [144]. Further evidence for fucosylation was provided by the ability to bind the fucose-specific lectin UEA-I from Ulex europeus [144]. The glycans from HMG proteins metabolically labeled with various monosaccharides were also resistant to release by mild alkali, thus suggesting N-linked glycans. In addition, a functional role for this glycosylation was suggested since the removal of the glycans by digestion with mixed glycosidases resulted in the loss of binding to nuclear matrix [145]. A recent study which reexamined glycosylation on HMG proteins [ 1461 reported the following: (i) Galactosyltransferase labeling of HMG proteins was unable to detect 0-linked GlcNAc - an abundant form of glycosylation of nuclear proteins. (ii) Binding of HMG14 to the fucose-specific lectin UEA-I from Ulex europeus could not be competed with excess L-fucose, thus raising doubt concerning the validity of the lectin studies. (iii) Acid hydrolysis of HMG proteins transferred to PVDF membranes suggest the presence of galactose and mannose as described above. In another report, porcine thymus HMG17 was isolated and characterized [ 1471. Elucidation of the entire primary sequence by gas phase sequencing of proteolytic fragments revealed no methylation or glycosylation. Therefore, it is clear that the glycosylation of HMG proteins is far from being understood. Additional studies may result in the discovery of novel forms of nuclear glycosylation and a function for these glycoproteins in transcriptional activation. 4.3. N-linked glycoproteins of the nucleus

Evidence for the presence of N-linked glycoproteins in the nucleus by compositional analysis, lectin binding studies, metabolic labeling and other methods has been reported for many years (for a review, see ref. [4]). Recently Letoublon and Frot-Coutaz have reported and partially characterized N-acetylglucosaminyltransferase activities in rat liver nuclei that can transfer N-acetylglucosamine [ 1481 and N,N’-diacetylchitobiose [ 1491 to nuclear glycoprotein acceptors in oitro. They demonstrate that the oligosaccharides of the acceptor proteins are N-linked rather than 0-linked by cleavage with PNGase F and resistance to alkaline induced B-elimination. Kinetic analysis of the incorporation of labeled N-acetylglucosamine showed initial labeling of a lipophilic intermediate within 15 min followed by a linear increase in transfer to acceptor proteins and a decrease in the level of the labeled-lipophilic intermediate over the next 60 min. They concluded that the lipophilic intermediate was dolichol since it ran like Dol-PP-sugar by TLC and since the incorporation of labeled N-acetylglucosamine could be inhibited by 95% with 1 pM tunicamycin. However, in both of these studies and in many of those reviewed by Hart et a]. [4] the probability of contamination of the nuclear fraction with other subcellular compartments, particularly the endoplasmic reticulum and the plasma membrane, was not appropriately addressed or not critically documented. Clearly, many of the observations of Letoublon [ 1481 and Frot-Coutaz [ 1491 could be accounted for by a minor contamination of their nuclei with ER fragments. Their observations nonetheless could represent another novel pathway of glycoprotein biosynthesis.

50

In another report, N-linked glycosylation of nuclear proteins was investigated in highly purified membrane-depleted nuclei from HeLa cells [ 1501. Nuclei were judged to be free of extranuclear contaminants by light and electron microscopy and by mixing experiments where labeled extranuclear cell extracts were added to non-labeled cells prior to isolation of the nuclei. These control tests resulted in “no significant level of radioactivity” detected in the membrane depleted nuclei. Glycoproteins were extracted with increasing concentrations of NaCl from the membrane-depleted nuclei of cells that had been metabolically labeled with [3H]fucose. Nearly all of the labeled glycoproteins were extracted at 0.4 M NaCl corresponding to the nonhistone proteins. Treatment of these glycoproteins with PNGase F released about 80% of the fucose-labeled oligosaccharides while alkaline fl-elimination was unable to release the other 20%. Initial characterization of the PNGase F released material demonstrated oligosaccharides with the following properties: (i) Size fractionation showed oligosaccharide ranging in size from M , 2400 to 10 000. (ii) More than 50% of the oligosaccharides could bind to QAESephadex but binding was abolished by pretreatment of the glycans with Vibrio cholerea sialidase, thus demonstrating the presence of sialic acid. (iii) Pretreatment of some [3H]glucosamine/[~4C]fucose, dual-labeled oligosaccharides with endo-fl-galactosidase released about 75% of the [3H] and [14C] into a heterogeneous mixture of smaller oligosaccharides leading the authors to conclude that these glycans are likely to contain some N-linked polylactosamine chains. In this study and in all of those addressing nuclear glycosylation, identification of known nuclear proteins as acceptor for the N-linked glycan will be critical to further prove these observations and thus minimize the ever-present criticism of contaminating extranuclear material.

5. Conclusions Over the past 10 years significant biochemical evidence has confirmed the existence of many nuclear and cytosolic glycoproteins. Many of these glycoproteins, unlike the “classical” endoplasmic reticulum and Golgi processed glycoproteins, appear to be glycosylated and deglycosylated in a regulated and dynamic fashion. While little is known about the function of these posttranslational modifications, evidence is certainly beginning to address how the monosaccharides and oligosaccharides are added and removed from these proteins by cytosolic and nuclear glycosyltransferases and glycosidases. Over the next several years, these, and possibly other forms of nuclear and cytosolic glycosylation not yet described, may be the keys to understanding some of the complex regulatory processes of cells.

Acknowledgments The original research from the Hart Laboratory is supported by NIH grants HD13563 & CA42486 and a Monsanto Corporation Glycosciences Postdoctoral Fellowship to KDG.

51

References [I] [2] [3] [4] [5]

Hart, G.W. (1992) Curr. Opin. Cell. Biol. 4, 1017-1023. Kornfeld, R. and Kornfeld, S. (1985) Annu. Rev. Biochem. 54, 631-664. Strous, G.J. and Dekker, J. (1992) CRC Crit. Rev. Biochem. Mol. Biol. 27, 57-92. Hart, G.W., Haltiwanger, R.S., Holt, G.D. and Kelly, W.G. (1989) Annu. Rev. Biochem. 58, 841-874. Hart, G.W., Kelly, W.G., Blomberg, M.A., Roquemore, E.P., Dong, D.L.-Y., Kreppel, L., Chou, T.-Y., Snow, D. and Greis, K.D. (1993) In: F. Wieland and W. Reutter (Eds.), 44 Mosbach Colloquium: DNA Replication and the Cell Cycle, Springer, Berlin, pp. 9-103. [6] Haltiwanger, R.S., Kelly, W.G., Roquemore, E.P., Blomberg, M.A., Dong, L.-Y.D., Kreppel L., Chou T.-Y. and Hart, G.W. (1992) Biochem. SOC.Trans. 20, 264-269. [7] Auger, D., Bounelis, P. and Marchase, R.B. (1993) In: D.J. Morre, J.J.M. Bergeron and K.M. Howell (Eds.), Molecular Mechanisms of Membrane Traffic, Springer, New York, pp. 289-292. [8] Smythe, C. and Cohen, P. (1991) Eur. J. Biochem. 200, 625-631. [9] Gonzalez-Yanes, B., Cicero, J.M., Brown, R.D. and West, C.M. (1992) J. Biol. Chem. 267,9595-9605. [lo] Krusius, T., Reinhold, VN., Margolis, R.K. and Margolis, R.U. (1987) Biochem. J. 245, 229-234. [ I l l Krusius, T., Finne, J., Margolis, R.K. and Margolis, R.U. (1986) J. Biol. Chem. 261, 8237-8242. [I21 Hiscock, D.R.R., Yanagishita, M. and Hascall, VC. (1994) J. Biol. Chem. 269, 45394546. [I31 Ishihara, M., Fedarko, N.S. and Conrad, H.E. (1986) J. Biol. Chem. 261, 13575-13580. [I41 Torres, C-R. and Hart, G.W. (1984) J. Biol. Chem. 259, 3308-3317. [I51 Holt, G.D. and Hart, G.W. (1986) J. Biol. Chem. 261, 8049-8057. [I61 Kearse, K.P. and Hart, G.W. (1991) Arch. Biochem. Biophys. 290, 543-548. [I71 Haltiwanger, R.S., Holt, G.D. and Hart, G.W. (1990) J. Biol. Chem. 265, 2563-2568. [18] Chou, C.-F., Smith, A.J. and Omary, M.B. (1992) J. Biol. Chem. 267, 3901-3906. [I91 Chou, C.-F. and Omary, M.B. (1993) J. Biol. Chem. 268, 44654472. [20] Kelly, W.G., Roquemore, E.P. and Hart, G.W. (1994) submitted. [21] Kearse, K.P. and Hart, G.W. (1991) Proc. Natl. Acad. Sci. USA 88, 1701-1705. [22] Roach, P.J. (1991) J. Biol. Chem. 266, 14139-14142. [23] Taylor, S.S. and Adams, J.A. (1992) Curr. Opin. Struct. Biol. 2, 743-748. [24] Kelly, W.G., Dahmus, M.E. and Hart, G.W. (1993) J. Biol. Chem. 268, 10416-10424. [25] Rogers, S., Wells R. and Rechsteiner, M. (1986) Science 234, 364-368. [26] Hirai, S., Kawasaki, H., Yaniv, M. and Suzuki. K. (1991) FEBS Lett. 287, 57-61. [27] Haltiwanger, R.S., Blomberg, M.A. and Hart, G.W. (1992) J. Biol. Chem. 267, 9005-9013. [28] Dong, L.-Y.D. and Hart, G.W. (1992) Glycoconjugate J. 8, 211-212. [29] Dong, L.-Y.D. and Hart, G.W. (1994) submitted. [30] Whiteheart, S.W., Passaniti, A,, Reichner, J.S., Holt, G.D., Haltiwanger, R.S. and Hart, G.W. (1989) Methods Enzymol. 179, 82-95. [3 11 O’Farrell, P.H. ( 1 977) J. Biol. Chem. 250, 40074021. [32] Snow, C.M., Senior, A and Gerace, L. (1987) J. Cell Biol. 104, 1143-1156. [33] Holt, G.D., Snow, C.M., Senior, A,, Haltiwanger, R. S., Gerace, L. and Hart G.W. (1987) J. Cell Biol. 104, 1157-1 164. [34] Forbes, D.J. (1992) Annu. Rev. Cell Biol. 8, 495-527. [35] Hanover, J.A. (1992) FASEB J. 6, 2288-2295. [36] Wolff, B., Willingham, M.C. and Hanover, J.A. (1988) Exp. Cell. Res. 178, 318-334. [37] Park, M.K., D’Onofrio, M., Willingham, M.C. and Hanover, J.A. (1987) Proc. Natl. Acad. Sci. USA. 84, 6462-6466. [38] Featherstone, C., Darby, M.K. and Gerace, L. (1988) J. Cell. Biol. 107, 1289-1297. [39] Akey, C.W. and Goldfarb, D.S. (1989) J. Cell. Biol. 109, 971-982. [40] Richardson, W.D., Mills, A.D., Dilworth, R.A. and Dingwall, C. (1988) Cell 52, 655-664. [41] Finlay, D.R. and Forbes, D.J. (1990) Cell 60, 17-29. [42] Miller, M.W. and Hanover, J.A. (1994) J. Biol. Chem. 269, 9289-9297. [43] D’Onofrio, M., Starr, C.M., Park, M.K., Holt, G.D., Haltiwanger, R.S., Hart, G.W. and Hanover, J.A. (1988) Proc. Natl. Acad. Sci. USA 85, 959559599, [44] Starr, C.M., D’Onofrio, M., Park, M.K. and Hanover, J.A. (1990) J. Cell Biol. 110, 1861-1871.

52 Cordes, V., Waizenegger, I. and Krohne, G. (1991) Eur. J. Cell Biol. 55, 3 1 4 7 . Cordes, V. and Krohne, G. (1993) Eur. J. Cell Biol. 60, 185-195. Kelly, W.G. and Hart, G.W. (1989) Cell 57, 243-251. Jackson, S.P. and Tjian, R. (1988) Cell 55, 125-133. Jackson, S.P. and Tjian, R. (1989) Proc. Natl. Acad. Sci. USA 86, 1781-1785. Lichsteiner, S. and Schibler, U. (1989) Cell 57, 1179-1 187. Reason, A.J., Morns, H.R., Panico, M., Marais, R., Treisman, R.H., Haltiwanger, R.S., Hart, G.W., Kelly, W.G. and Dell, A. (1992) J. Biol. Chem. 267, 16911-16921. [52] Kadonaga, J.T., Courey, A.J., Ladika, J. and Tjian, R. (1988) Science 242, 1566-1570. [53] McClain, D.A., Paterson, A.J., Roos, M.D., Wei, X. and Kudlow, J.E. (1992) Proc. Natl. Acad. Sci. USA 89, 8150-8154. [54] Schaufele, F., West, B.L., Reudelhuber, T.L. (1990) J. Biol. Chem. 265, 17189-17196. [55] Corden, J.L. (1990) Trends Biochem. Sci. 15, 347-351. [56] Cadena, D.L. and Dahmus, M.E. (1987) J. Biol. Chem. 262, 12468-12474. [57] Layboum, P.J. and Dahmus, M.E. (1989) J. Biol. Chem. 264, 6693-6698. [58] Chou, T.-Y., Dang, C.V., Eiden, J. and Hart, G.W. (1993) Glycobiology 3, 523. [59] Pnvalsky, M.L. (1990) J. Virol. 64, 463466. [60] Jiang, M.-S. and Hart, G.W. (1994) J. Cell. Biochem. Supplement 188, 344 [61] Elliot, S.P., Schmied, R., Gabel, C.A. and Ambron, R.T. (1993) J. Neurosci. 13, 2424-2429. [62] Soulard, M., Barque, J.-P., Valle, V.D., Hernandez-Verdun, D., Masson, C., Danon, F. and Larsen, C.-J. (1991) Exp. Cell Res. 193, 59-71. [63] Soulard, M., Valle, V.D., Siomi, M.C., Pinol-Roma, S., Codogno, I? et al. (1993) Nucl. Acids Res. 21, 42 10-42 17. [64] Le Goff, P., Montano, M.M., Schodin, D.J. and Katzenellenbogen, B.S. (1994) J. Biol. Chem. 269, 44584466. [65] Ambron, R.T., Schmied, R., Huang, C.C., Smedman, M. (1992) J. Neurosci. 12, 2813-2818. 1661 Holt, G.W., Haltiwanger, R.S., Torres, C-R. and Hart, G.W.. (1987) J. Biol. Chem. 262, 14847-14850. [67] Steinert, P.M. and Roop, D.R. (1988) Annu. Rev. Biochem. 57, 593-625. [68] King, LA. and Hounsell, E.F. (1989) J. Biol. Chem. 264, 14022-14028. [69] Chou, Y.-H., Rosevear, E. and Goldman, R.D. (1989) Proc. Natl. Acad. Sci. USA 86, 1885-1889. [70] Dong, L.-Y.D., Xu, Z.-S., Chevrier, M.R., Cotter, R.J., Cleveland, D.W. and Hart, G.W. (1993) J. Biol. Chem. 268, 16679-16687. [71] Wong, P.C. and Cleveland, D.W. (1990) J. Cell Biol. 111, 1987-2003. [72] Gill, S.R., Wong, P.C., Monteiro, M.J. and Cleveland, D.W. (1990) J. Cell Biol. 11 I , 2005-2019. [73] Hagmann, J., Grob, M. and Burger, M.M. (1992) J. Biol. Chem. 267, 14424-14428. [74] Gupta, N.K., Ahmad, M.F., Chakrabarti, D. and Nasrin, N. (1987) In: J. Ilan (Ed), Translation Regulation in Gene Expression, Plenum, New York, pp. 287-334. [75] Hershey, J.W.B. (1991) Annu. Rev. Biochem. 60, 717-755. [76] Datta, B., Chakrabarti, D., Roy, A.L. and Gupta, N.K. (1988) Proc. Natl. Acad. Sci. USA 85, 33243328. [77] Datta, B., Ray, M.K., Chakrabarti, D., Wylie, D.E. and Gupta, N.K. (1989) J. Biol. Chem. 264, 2062020624. [78] Ray, M.K., Datta, B., Chakraborty, A,, Chattopadhyay, A,, Meza-Keuthen, S. and Gupta, N.K. (1992) Proc. Natl. Acad. Sci. USA 89, 539-543. [79] Gupta, N.K. (1993) Presented at the 22nd annual meeting of the Society for Complex Carbohydrates, San Juan, PR, November 1993. [80] Ishibashi, M. and Maizel, J.V. (1974) Virology 58, 345-361. [8l] Mullis, K.G., Haltiwanger, R.S., Hart, G.W., Marchase, R.B. and Engler, J.A. (1990) J. Virol. 64, 53175323. [82] Benko, D.M., Haltiwanger, R.S., Hart, G.W. and Gibson, W. (1988) Proc. Natl. Acad. Sci. USA 85, 2573-2577. [83] Whitford, M. and Faulkner, P. (1992) J. Virol. 66, 3324-3329. [84] Ku, N.-0. and Omary. M.B. (1994) Exp. Cell Res. 21 1, 24-35.

[45] [46] [47] [48] [49] [50] [51]

53 [85] Davis, L.1. and Blobel, G. (1987) Proc. Natl. Acad. Sci. USA 84, 7552-7556. Hanover, J.A., Cohen, C.K., Willingham, M.C. and Park, M.K. (1987) J. Biol. Chem. 262, 9887-9894. Schindler, M., Hogan, M., Miller, R. and DeGaetano, D. (1987) J. Biol. Chem. 262, 1254-1260. Meikrantz, W., Smith, D.M., Sladicka, M.M. and Schlegel, R.A. (1991) J. Cell Sci. 98, 303-307. Gabel, C.A., Den, H. and Ambron, R.T. (1989) J. Neurobiol. 20, 530-548. Ambron R.T., Protic, J., Den, H., Gabel, C.A. (1989) J. Neurobiol. 20, 549-568. Abeijon, C. and Hirschberg, C.B. (1988) Proc. Natl. Acad. Sci. USA 85, 1010-1014. Gonzalez, S.A. and Burrone, O.R. (1992) Virology 182, 8-16. Luthi, T., Haltiwanger, R.S., Greengard, P. and Bahler, M. (1991) J. Neurochem. 56, 1493-1498. Roquemore, E.P., Dell, A,, Morris, H.R., Panico, M., Reason, A.J., Savoy, L.-A,, Wistow, G.J., Zigler, Jr., J.S., Earles, B.J. and Hart, G.W. (1992) J. Biol. Chem. 267, 555-563. Nyame, K., Cummings, R.D. and Damian, R.T. (1987) J. Biol. Chem. 262, 7990-7995. Srisomsap, C., Richardson, K.L., Jay, J.C. and Marchase, R.B. (1988) J. Biol. Chem. 263, 17792-17797. Marchase, R.B., Hiller, A.M., Rivera, A.A. and Cook, J.M. (1987) Biochem. Biophys. Acta. 916, 157162. Hiller, A.M., Koro, L.A. and Marchase, R.B. (1978) J. Biol. Chem. 262, 43774381, Satir, B.H., Srisomsap, C., Reichman, M. and Marchase, R.B. (1990) J. Cell Biol. 1 1 1, 901-907. Marchase, R.B., Bounelis, P., Brumley, L.M., Dey, N., Browne, B., Auger, D., Fritz, T.A., Kulesza, P. and Bedwell, D.M. (1993) J. Biol. Chem. 268, 8341-8349. Murtaugh, T.J., Gilligan, D.M. and Satir, B.H. (1987) J. Biol. Chem. 262, 15724-15739. Gilligan, D.M. and Satir, B.H. (1982) J. Biol. Chem. 257, 13903-13906. Zieseniss, E. and Plattner, H. (1985) J. Cell Biol. 101, 2028-2035. Subramanian, S.V. and Satir, B.H. (1992) Proc. Natl. Acad. Sci. USA 89, 11297-1 1301. Srisomsap, C., Richardson, K.L., Jay, J.C. and Marchase, R.B. (1989) J. Biol. Chem. 264,20540-20546. Veyna, N.A., Jay, J.C., Srisomsap, C., Bounelis, I? and Marchase, R.B. (1994) J. Neurochem. 62, 456464. Cori, C.F. and Cori, G.T. (1939) J. Biol. Chem. 131, 397-398. Leloir, L.F. and Cardini, C.E. (1957) J. Am. Chem. SOC.79, 6340. Krisman, C.R. and Barrengo, R. (1975) Eur. J. Biochem. 156, 163-169. Kennedy, L.D., Kirkmsn, B.R., Lomako, J., Rodriguez, I.R. and Whelan, W.J. (1985) In: M.C. Berman, W. Geyers and L.H. Opie (Eds.), Membranes and Muscle, ICSU Press/IRL Press, Oxford, pp. 65-84. Rodriguez, I.R. and Whelan, W.J. (1985) Biochem. Biophys. Res. Commun. 132, 829-836. Smythe, C., Watt, P. and Cohen, P. (1990) Eur. J. Biochem. 189, 199-204. Campbell, D.G. and Cohen, P. (1989) Eur. J. Biochem. 185, 119-125. Smythe, C., Cauldwell, F.B., Ferguson, M. and Cohen, P. (1988) EMBO J. 7, 2681-2686. Viskupic, E., Cao, Y., Zhang, W., Cheng, C., Depaoli-Roach, A.A. and Roach, P.J. (1992) J. Biol. Chem. 267, 25759-25763. Cao, Y., Skurat, A.V., Depaoli-Roach, A.A. and Roach, I?J. (1993) J. Biol. Chem. 268, 21717-21721. Cao, Y., Mahrenholz, A.M., Depaoli-Roach, A.A. and Roach, P.J. (1993) J. Biol. Chem. 268, 1468714693. Skurat, A.V., Cao, Y. and Roach, I?J. (1993) J. Biol. Chem. 268, 14701-14707. Helmreich, E.J. (1992) Biofactors 3, 159-172. Caldor, P.C. (1991) Int. J. Biochem. 23, 133551352, Hallgren, E.F., Lundblad, A. and Svensson, S. (1975) J. Biol. Chem. 250, 5312-5314. Steiner, S., Via, D.P., Klinger, M., Larriba, G., Sramek, S. and Laine, R. (1978) In: E.F. Walborg Jr. (Ed.), Glycoproproteins and Glycolipids in Disease Processes, American Chemical Society, Washington, pp. 378403. [I231 Klinger, M.M., Laine, R.A. and Steiner, S.M. (1981) J. Biol. Chem. 256, 7932-7935. [I241 Morton, P.A. and Steiner, S.M. (1985) Biochem. J. 225, 5 9 4 5 . [I251 Harris, R.J. and Spellman, M.W. (1993) Glycobiology 3, 219-224. 11261 Lin, A.I., Polk, C., Xiang, W.-K., Phillipsberg, G.A. and Haltiwanger, R. S. (1993) Glycobiology 3, 524. 11271 Stults, N.L. and Cummings, R.D. (1993) Glycobiology 3, 589-596.

54

[I281 [I291 [I301 [I311 [I321 [I331 [I341 [I351 [I361 [I371 [I381 [I391 [I401 [I411 [I421 [I431 [I441 [I451 [I461 [I471 [I481 [I491 [I501 [I511

Margolis, R.U., Margolis, R.K. and Atherton, D.M., (1972) J. Neurochem. 19, 2317-2324. Margolis, R.K., Margolis, R.U., Preti, C. and Lai, D. (1975) Biochemistry 14, 4797-3804. Kiang, W.L., Crockett, C.P., Margolis, R.K., Margolis, R.U. (1978) Biochemistry 17, 3841-3848. Margolis, R.K., Thomas, M.D., Crockett, C.P., Margolis, R.U. (1979) Proc. Natl. Acad. Sci. USA 76, 171 1-1715. Finne, J., Krusius, T., Margolis, R.K., Margolis, R.U. (1979) J. Biol. Chem. 254, 10295-10300. Aquino, D. A,, Margolis, R.U., Margolis, R.K. (1984) J. Cell. Biol. 99, 1130-1139. Ripellino, J.A., Bailo, M., Margolis, R.U., Margolis, R.K. (1988) J. Cell. Biol. 106, 845-855. Hardingham, T.E. and Fosang, A.L. (1992) FASEB J. 6, 861-870. Pedemonte, C.H., Sachs, G. and Kaplan, J.H. (1990) Proc. Natl. Acad. Sci. USA 87, 9789-9793. Pedemonte, C.H. and Kaplan, J.H. (1992) Biochemistry 31, 10465-10470. Fedarko, N.S. and Conrad, H.E. (1986) J. Cell Biol. 102, 587-599. Ishihara, M. and Conrad, H.E. (1989) J. Cell. Phys. 138, 467476. Fedarko, N.S., Ishihara, M. and Conrad, H.E. (1989) J. Cell. Phys. 139, 287-294. Yanagishita, M. and Hascall, VC. (1992) J. Biol. Chem. 267, 9451-9454. Goodwin, G.H. and Mathew, C.G.P. (1982) In: E.W. Johns (Ed.), The HMG Chromosomal Proteins, Academic Press, New York, pp. 193-222. Reeves, R. (1984) Biochem. Biochys. Acta 782, 343-393. Reeves, R., Chang, D. and Chung, S.-C. (1981) Proc. Natl. Acad. Sci. USA 78, 6704-6708. Reeves, R. and Chang, D. (1983) J. Biol. Chem. 258, 679-687. Medina-Vera, L. and Haltiwanger, R.S. (1993) Glycobiology 3, 523. Boumba, VA., Tsolas, O., Choli-Papadopoulou, D. and Seferiadis, K. (1993) Arch. Biochem. Biophys. 303, 436442. Letoublon, R., Merit, X. and Frot-Croutaz, J. (1991) Biochem. Int. 23, 221-230. Frot-Coutaz, J., Degiuli, A,, Martel, M.-B., Letoublon, R. (1992) Biochem. Cell. Biol. 70, 677-683. Codogno, P., Bauvy, C., Seve, A.-P., Hubert, M., Ogier-Denis, E., Aubery, M., Hubert, J. (1992) J. Cell. Biochem. 50, 93-1 02. Greis, K.D., Gibson, W., Hart, G.W. (1994) J. Virol. 68, 8339-8349.

J. Montreuil, J.F.G. Vliegenthart and H . Schachter (Eds.), Glycoproteins 11 0 1997 Elsevier Science B.V. All rights reserved CHAPTER 3

Carbohydrate units of nervous tissue glycoproteins Jukka Finne Department of Medical Biochemistry, University of Turku, FIN-20520 Turku, Finland

Abbreviations N-CAM neural cell adhesion molecule

1. Introduction Cell surface carbohydrates, due to their complex structure and abundant presence at the cell surface, have long been regarded as candidate molecules participating in the specialized interactions of nervous tissue cells. Nervous tissue has been one of the first targets for attempts to characterize membrane bound tissue glycoproteins [ 11. It has become clear that the glycans of these glycoproteins differ in many respects from plasma glycoproteins which have been the subject of most structural studies on glycoproteins [2]. It is therefore not surprising that a number of novel structural features of glycoproteins were reported for the first time from studies on nervous tissue glycoproteins (reviewed in ref. [3]) and that the picture of the common properties of protein-bound carbohydrates has consequently somewhat changed since the early studies on plasma glycoproteins. Of the novel structures discovered in nervous tissue glycoproteins many have been later found to also exist in other tissues, but some are clearly enriched in nervous tissue and thus suggest involvement in biological interactions characteristic for this tissue. The knowledge accumulated on the structural features of nervous tissue glycoproteins has been summarized in a number of reviews [1,3-51. While much of the work on nervous tissue glycoprotein carbohydrates has focused on their structural properties, more recent observations have revealed indications of their involvement in diverse biological interactions. On the whole, the biological roles of nervous tissue glycoproteins, like those in other tissues, are still unresolved. The present paper is an overview on the main structural features of protein-bound glycans of nervous tissue with some indications on their possible involvement in some biological interactions. Most of the structural information on nervous tissue glycans has come from studies on whole tissues, cells or subcellular fractions, but information on the glycosylation patterns of individual glycoproteins is starting to accumulate. The carbohydrate patterns of individual neural glycoproteins and the involvement of neural glycoproteins in cell adhesion have been reviewed elsewhere [6-81.

2. Core structures The analysis of the highly complex mixture of glycans released from tissues by extensive proteolytic degradation was for a long time complicated by the lack of 55

56

fractionation methods that could resolve these heterogeneous glycan preparations into clear-cut fractions of different core structures. One of the break-throughs in the analysis of tissue-derived glycopeptides was the development of a fractionation method based on concanavalin A affinity chromatography and the selective release of the 0-linked glycans as reduced oligosaccharides [9]. The method has been further refined by including additional lectins in the fractionation procedure [lo], and the development of high-performance liquid chromatography has significantly improved the separation methods of protein-bound oligosaccharides [ 1 1-13]. The N-linked glycans are resolved into three main fractions by concanavalin A affinity chromatography: a fraction not bound to the lectin representing mainly tri- and tetraantennary glycans, a fraction of weakly bound glycans consisting of diantennary glycans, and a fraction of strongly bound high-mannose glycans [9,14]. It should be noticed that although the diantennary glycans were long regarded as the prototype of glycoprotein carbohydrates, in the brain like in many other cell and tissue sources the multi-antennary glycans form quantitatively the most prominent fraction (Fig. 1) [15-171. Owing to their smaller size in many tissues, the 0-linked glycans can be separated from the multi-antennary N-linked glycopeptides after mild alkaline borohydride treatment by gel filtration [9,15,18]. Quantitatively, they comprise a minor fraction of the proteinbound carbohydrate in brain tissue, but due to their small size they account mole-wise for about one third of all glycans. The majority of the 0-linked glycans are of the classical GalNAc-linked type, but a minor part of the glycans contain mannose and are apparently linked through Man-Ser/Thr linkages (Fig. 1) [ 19,201.

3. Terminal sequences of N-linked glycans The terminal sequences of the N-linked glycans (Fig. 2) include units that are encountered in glycoproteins of many tissues. Most of the structures are based on the disaccharide Gal(B14)GlcNAc, but about one fifth of the disaccharide units in rat brain glycoproteins are Gal((31-3)GlcNAc[21]. Like in many other tissue and cell sources the most common sialic acid substitution on the Gal@14)GlcNAc disaccharide is a-2,3 to galactose [ 15,17,21], in contrast to plasma glycoproteins, in which the a-2,6-sialic acid linkage predominates [22]. The sialylated sugar termini also predominate in parental tissues like liver and kidney [ 151, in contrast to some other tissue sources like the human small intestinal epithelium, where sialic acid is almost totally lacking as a terminal unit of N-glycans and is replaced by fucose [23]. The non-sialylated galactose-terminated glycans are potential ligands for the endogenous galactose-binding lectins of nervous tissue [24,25]. The sialyl LeX sequence with the a-1,3-bound fucose was first discovered as a constituent of N-glycans in rat brain[21], but is now known to occur in many other tissues as well [2]. This sugar sequence is suggested to function as a ligand in selectin mediated cell adhesion [26,27]. In addition to sialic acid, a negative charge is given to the N-glycans of nervous tissue by sulfate [4], which also represents potential for selectin interaction [28]. However, the exact patterns of sulfate substitution in nervous tissue glycoproteins have not yet been resolved in detail.

57

Multiantennary

GalNAc-linked

Mannose-linked

GlcNAc

Man

Ratio (mol YO):

35

30

5

ConA-Sepharose:

not bound

not bound

not bound

Biantennary

J

GlcNAc

GlcNAc

High-mannose

J

Man

Man

Ratio (rnol Yo):

15

15

ConA-Sepharose:

bound weakly

bound strongly

Man

Fig. I . Core structures of the carbohydrate units of nervous tissue glycoproteins. The structures are based on analytical data on rat brain glycoproteins and on assumptions of structural similarity with glycan cores from other sources. The main positions of variable or incomplete glycosylation are indicated by arrows. The approximate molar proportions of the glycans in rat brain and the mode of interaction with concanavalin ASepharose are indicated (the bisecting GlcNAc residue affects the interaction of the diantennary glycans with concanavalin A) [9].

58

NeuNAc (a2-6)

I GIcNAc (PI-

Gal (pl-4)GIcNAc (pl-

Gal (PI-4) GIcNAc (PI-

NeuNAc (a2-3) Gal (pl-4)GlcNAc ( P I -

-

Gal (Pl-3)GlcNAc -

Fig. 2. Terminal sugar sequences of the N-linked glycans of nervous tissue glycoproteins. For references, see ref. [3].

The high-mannose glycans are carriers of the L3 monoclonal antibody epitope shared by some neural cell adhesion molecules [29]. One potentially important interaction of the high-mannose glycans of the L1 glycoprotein on the neural cell surface has been suggested to be their involvement in the cis-binding of this glycoprotein to the neural cell adhesion molecule N-CAM, thus forming a complex of these two molecules [30].

4. Classical 0-linked oligosaccharides The most common 0-linked glycan of brain tissue is the tetrasaccharide consisting of the disaccharide GaI((31-3)GalNAc and two sialic acid residues (Fig. 3) [IS]. The amounts of the disaccharide containing 0, 1 or 2 sialic acid units are very similar in the brains of hen, rat and rabbit [3 11. Similar structures are also present in other derivatives of neural tissue like in the adrenal medulla, where an oligosaccharide with an a-2,8-sialyl linkage was also discovered [32]. In these oligosaccharides, like in the N-linked glycans, the sialic acid units are variably present as the N-acetyl or N-glycolyl forms depending on the animal species. The possible 0-acetyl and other substituents of the sialic acid units of nervous tissue glycoproteins have not been thoroughly investigated.

59 NeuNAc (a2-6)

NeuNAc(a2-6)

I

I Gal (al-3)GalNAc-

Gal (pl-3)GdNAC-

NeuNAc (a2-3)Gal (pl-3)GalNAc-

NeuNAc (a2-8)NeuNAcW-6)

I Gal (pl-3)GalNAC-

NeuNAc(a2-31-l

(Pl-3)GalNAc-

Gal (pI-3)GalNAC-

Fig. 3. Structures of the classical 0-linked glycans of nervous tissue glycoproteins [3,18,31,32].

A disaccharide unit with an a-galactose moiety, Gal(a 1-3)GalNAc has been discovered as an 0-linked unit in the brains of the rat, rabbit, hen, frog and fish [3 11. It seems to be concentrated in nervous tissue, because it was not found in five other tissues of the rat. Other sources where the disaccharide has been discovered are the muscle and the adrenal medulla [32] and human teratocarcinoma cells [33]. Its enrichment in brain tissue together with its occurrence in roughly similar amounts in different animal species may suggest a role important for nervous tissue function. However, no information of its possible role or occurrence in specified glycoprotein molecules is yet available.

5. 0-mannose linked oligosaccharides A series of mannitol-terminated oligosaccharides have been isolated after mild alkaline borohydride treatment of a brain chondroitin sulfate proteoglycan fraction (Fig. 4). The structures range from the monosaccharide mannitol to neutral and sialic acid containing oligosaccharides with GlcNAc-Man-ol as a common proximal disaccharide unit, and to sialic acid and/or sulfate containing acidic oligosaccharides [ 19,201. The largest units represent keratan sulfate-like polymers. Mannitol-containing oligosaccharides are present in rat, rabbit and hen brain. The linkage between GlcNAc and Man-ol was originally deduced to be b1-3 on the basis of a deuterium shift in the mass spectrum, but recent results unexpectedly indicate a similar fragmentation for GlcNAc(b 1-2)Man-o1[34], which suggests that the linkage may in fact be b1-2. The presence of mannitol at the proximal end of the oligosaccharides and the identification of the modified derivatives of serine and threonine in the peptide moieties Man-

NeuNAc (a2-3) Gal (pl-4)GlcNAc (PI-) Mon-

GlcNAc (pi-) Man-

Gal (pi-4)GlcNAc (pi-) Man-

(SO,-), (NeuNAc (a2-),

(SO;),

(-3)Gol (pl-4)GlcNAc (pl-),

-

-

- Man-

-

-

I Fuc (al-3)

Fig. 4. Structures of 0-mannose-linked glycans of the nervous tissue. For references, see refs. [19,20].

Man-

60

after alkaline borohydride treatment suggests that the oligosaccharides are bound through Man-0-Ser/Thr linkages [20]. However, the presence of an unrecognized alkali-labile unit between mannose and serine/threonine cannot be ruled out. 0-linked mannose has previously been found in the yeast but not in animal glycoproteins. Proteoglycans containing these oligosaccharides include at least neurocan and phosphacan [35]. The neural chondroitin sulfate proteoglycans have been implicated in various molecular interactions, but the role of the oligosaccharides has not been resolved. Potential ligands of importance present in the proteoglycans include the Lex-like structure and the L2/HNK-I epitope (see below).

6. Poly-N-acetyllactosamine glycans The poly-N-acetyllactosamine glycans were first discovered as constituents of N-glycans in erythrocytes [36] and teratocarcinoma cells [37], in which they represent a major fraction of the cell surface carbohydrates. In contrast, analysis of whole brain tissue does not indicate the presence of major amounts of poly-N-acetyllactosamine chains [ 151. That poly-N-acetyllactosamine glycans do exist in brain tissue is suggested by their accumulation in GM1-gangliosidosis [38]. Small amounts are also present in the form of keratan sulfate as 0-mannose linked chains (see above). Analysis of glycopeptides of different cell type in culture has revealed that at least small amounts of poly-N-acetyllactosamine chains are present in all cell lines analyzed [39]. Interestingly, in the cells of neural origin the poly-N-acetyllactosamine glycans were mainly of the linear type, whereas those of other cell types were mainly branched.

10

20

30 40 FRACTION NUMBER

50

60

Fig. 5. Induction of poly-N-acetyllactosamine expression in differentiating mouse neuroblastoma cells. NI E-I15 neuroblastoma cells were grown in the undifferentiated form or induced to differentiate by serum starvation for ten days. The glycopeptides metabolically labelled with 3H-glucosamine were prepared by pronase digestion and fractionated by gel filtration on a column of Bio-Gel P-l 00. The poly-N-acetyllactosamine glycopeptides (poly-LacNAc) were eluted in the fractions indicated by the bar [39].

61

-0

R-GlcNAc-Gal+ GlcNAc

endo-P-galactosidase ['4C]Gal-UDP + GlcNAc galactosyltransferase [''C]Gal-GlcNAc

4

-0

Fig. 6 . Principle of the poly-N-acetyllactosamine labelling method. In the first step, cells are treated with endo-p-galactosidase, which cleaves only poly-N-acetyllactosamine glycans. In the second step, the terminal N-acetylglucosamine residues are labelled by the transfer of radiolabelled galactose from UDP-galactose with the aid of galactosyltransferase [40,41].

Whether this represents a true neural property would need to be substantiated by the analysis of glycoproteins of natural tissue origin. In contrast to most other glycans the poly-N-acetyllactosamine glycans undergo major variation in quantity during induced differentiation in cell culture (Fig. 5). In order to identify the glycoproteins which carry poly-N-acetyllactosamine glycans a cell surface labelling method has been developed [40,41]. The method is based on the specificity of endo-fi-galactosidase which only cleaves poly-N-acetyllactosamine chains (Fig. 6). The remaining part of the glycans contains after endo-fi-galactosidase cleavage a terminal N-acetylglucosamine unit, a rare constituent of unmodified cell surface glycoproteins. The terminal N-acetylglucosamine residues are subsequently labelled with radioactive UDP-galactose in a galactosyltransferase-catalyzed reaction. Labelling of the cell surface of PC12 pheochromocytoma cells for the presence of poly-N-acetyllactosamine chains identifies a restricted number of glycoproteins as carriers of these glycans (Fig. 7). Induced neuronal differentiation of the cells in culture

-++-

Soc

++++

GT

205

-

916

-

el

-

86

-

45

-

28

-

fmnl

-

Ic

*und,f,

c1

dif,

Fig. 7. Poly-N-acetyllactosamine containing glycoproteins of PC12 pheochromocytoma cells. Undifferentiated cells and cells differentiated by growth in the presence of nerve growth factor for six days were labelled with the poly-N-acetyllactosamine labelling method using endo-P-galactosidase (EpG) and galactosyltransferase (GT) as indicated, and the labelled glycoproteins were subjected to SDS-polyacrylamide gel electrophoresis and fluorography [39]. (Reproduced with the permission of Raven Press, Ltd./New York from D. Spillmann and J. Finne, PolyN-acetyllactosamine glycans of cellular glycoproteins: predominance of linear chains in mouse neuroblastoma and rat pheochromocytoma cell lines, Journal of Neurochemistry 49 (1987) 874-883.)

62

causes a decrease of the labelling of some of the major bands and the appearance of a new high-molecular mass band. The latter represents the nerve-growth-factor-inducible glycoprotein NILE, a molecule related to the cell adhesion glycoprotein LI and the neuron-glia adhesion molecule Ng-CAM [42]. The poly-N-acetyllactosamine glycans have been suggested to modulate negatively the binding of fibronectin to gelatin [43], but their possible role in nervous tissue glycoproteins is unknown.

7. Polysialic acid The occurrence of a-2,8-linked N-acetylneuraminic acid units in brain glycoproteins was reported in 1977 [44], and they were subsequently found to be constituents of a novel structural unit of N-linked glycoproteins, polysialic acid [45]. Similar polymers of N-acetylneuraminic acid are known to occur as constituents of the capsular polysaccharides of meningitis-producing bacteria, Escherichia coli K1 and group B meningococci [46], and were also found as constituents of 0-linked oligosaccharides of salmonid fish eggs [47]. The polysialic acid units are linear polymers of up to at least 55 or more sialic acid residues [48]. They appear to be bound by an a-2,3-linkage to a galactose residue in an N-linked glycan (Fig. 2) [45]. A complete core structure has not yet been determined, but the structural analyses performed suggest N-linked structures, mainly multi- (tri- and tetra-) antennary structures [45]. A significant amount (10% or more) of all protein-bound sialic acid in developing brain occurs in polysialic-acid-containing glycans [45]. Polysialic acid units undergo a significant developmental decrease during brain development, with only small amounts being present in adult brain tissue (Fig. 8). Although small amounts of polysialic acid are

2001 YOUNG

N

ADULT

1

300 300

P 0 a

2 200 100 4

ffl

POLYSIALIC ACID

100

0

10

20

30 40 FRACTION NUMBER

50

I

Fig. 8. Developmental decrease of polysialic-acid-containing glycans in rat brain. Glycopeptides were prepared by pronase digestion from 8-day-old and adult rat brains, treated with alkaline borohydride, and fractionated by gel filtration on a column of Bio-Gel P-100. The polysialic-acid-containing glycopeptides, normal N-linked glycopeptides, and 0-glycosidic oligosaccharides were eluted in the fractions indicated [45].

63

NEWBORN

ADULT

Fig. 9. Polysialic-acid-containing glycoproteins in newborn and adult rat tissues. Tissue homogenates of brain (B), liver (L), kidney (K), spleen (S), gluteal muscle (M) and heart (H) were subjected to SDS-polyacrylamide gel electrophoresis and polysialic acid was detected using immunoblotting with a monoclonal antibody to the capsular polysaccharides of Escherichia coli KI and group B meningococci [49]. (From J. Immunol. 138, June 15, 1987, 44024407. Copyright 1987, The Journal of Immunology.)

present in many extraneural tissues during development, little is found in adult extraneural tissues [49]. Polysialic acid is present in many animal species including elasmobranchs [50] and Drosophila [51]. The main carrier of polysialic acid units is the neural cell adhesion molecule N-CAM [52]. In adult rat brain, the sodium channel glycoprotein has also been found to be a carrier of polysialic acid[53]. Owing to its identical structure with some bacterial polysaccharides, brain polysialic acid can be identified with the aid of antibodies directed against these polysaccharides (Fig. 9) [49]. Polysialic acid is an unusually weak immunogen, and the polysialic acid seems thus to give protection to the bacteria from immunological defence mechanisms [54]. The mode of interaction of polysialic acid with its antibodies is unusual as it requires usually a minimum of about 8-10 sialyl residues for interaction [55-571. Similarly, efficient cleavage of polysialic acid by a bacteriophage endosialidase requires a minimum chain length of 8 sialyl residues (Fig. 10) [55,58]. A likely explanation is that the interacting molecules are recognizing a conformational epitope of the polysialic acid chain [59]. The biological role of polysialic acid is not known. However, several observations suggest that polysialic acid may be a negative modulator of the interactions mediated by the protein part of N-CAM [60,61]. Alternatively, polysialic acid may, due to its huge size, form a negatively charged physical barrier at the cell surface, thus influencing not only N-CAM interactions, but also interactions of other cell surface molecules [62].

64

Antibodies

I,

Neu Ac

I,

Neu Ac

I u2-8

minimum size 8-1 0 residues

NeuNAc u2-8 NeuNAc

I

1I

Neu Ac u2-8 NeuNAc az-8 NeuNAc a2-8 NeuNAc u2-8

I I I

Endosialidases

I

u2-8 NeuNAc

bI

Neu Ac u2-8 a2-8 NeuNAc

minimum substrate 8 residues

+

distal fragment 3 residues

I I I

a-8 NeuNAc u2-8 NeuNAc uz-8 NeuNAc

cI

Ncu Ac uz-8 NeuNAc u2-8 NeuNAc 02-8

I I

Fig. 10. Requirement of a long oligosaccharide segment in molecular interactions of polysialic acid. A minimum size of approximately 8-10 residues is required for the binding of polysialic acid to antibodies [55,57]. A minimum size of 8 sialic acid residues is needed for cleavage by bacteriophage-associated endosialidases, which cleave the oligosaccharide recognized into a distal fragment containing a minimum of 3 and a proximal fragment containing a minimum of 5 sialyl residues [55,58].

8. Other structures The L2/HNK-1 epitope is a sulfated glucuronic-acid-containingoligosaccharide structure of glycolipids [63]. As indicated by the reactivity with the L2 monoclonal antibody,

65

a similar epitope may be present in a number of neural glycoproteins (N-CAM, L1, MAG, Po, tenascin, janusin, TAG- 1, some integrins and ependymins) which all have been implicated in cell adhesion [64]. The L2/HNK-1 carbohydrate is suggested to mediate the binding of neural cells to laminin [64]. L5, another epitope specified by a monoclonal antibody involved in neural cell interactions appears to be a carbohydrate shared by a number of glycoproteins and proteoglycans [65], but its structure has not yet been reported. Glycopeptide fractions isolated from developing brain contain a glucose polymer, which rapidly decreases during development [66]. It is not known whether it represents a true glycoprotein constituent.

9. Discussion Although many of the novel structural features of glycoproteins that were discovered in the nervous tissue were later found in many other tissues as well, it seems that there are features that are enriched in the nervous tissue. These include polysialic acid, the a-galactose containing 0-linked disaccharide, and possibly the 0-mannose linked oligosaccharides, the sialyl LeX structure, and the linear form of poly-N-acetyllactosamine glycans. Another approach to study tissue-specific glycosylation is to investigate the glycosylation patterns of individual glycoproteins occurring in different tissues. Such results are now starting to accumulate and the observations so far made support the idea of tissue-specified glycosylation patterns [7]. Thus, the glycosylation patterns of the Thy-1 glycoproteins of rat brain and thymocytes were found to be different, whereas the glycosylation patterns of mouse and rat brain Thy-1 were indicated to be very similar [7]. Of interest are the changes that occur during differentiation of tissues and cells. Although several structures undergo changes in quantity during differentiation, most of these are relatively small. There are two major exceptions to this, polysialic acid and poly-N-acetyllactosamine glycans which both undergo major changes during differentiation [39,45]. It remains to be established whether their differential behavior is related to their large size which might provide them with biological roles different from those of the smaller glycans. The biological roles of the carbohydrate units of nervous tissue glycoproteins are still for the most part unknown. At the molecular level at least two different interaction mechanisms are indicated. One is the functioning of the carbohydrates as mediators of molecular interactions by serving as ligands recognized by carbohydrate-binding proteins, like the selectins or other endogenous carbohydrate-binding proteins of the nervous system [8,67,68]. Another is the participation of carbohydrates in the interactions as modulators of the binding strength. Examples of the latter are the negative modulation of N-CAM to N-CAM binding by polysialic acid [60,61] or the stabilization of N-CAM by heparan sulfate [69]. Apart from their intrinsic role in tissue functioning, the carbohydrate units participate in other types of biological interactions such as the pathogenesis of microbial diseases. As indicated above for polysialic acid some pathogens may mimic host tissue oligosaccharide structures in order to escape immunological defence. In addition, microbes

66

like meningitis-inducing bacteria may use specific carbohydrate structures of host tissue glycoconjugates as their binding receptors, in order to attach to host tissues and cells. Some of the receptors identified [70,7 11 represent structures abundantly present in glycoproteins of neural and other tissues (Figs. 2 and 3). These interactions also serve as excellent models of protein-carbohydrate interactions. It has become clear that no single function can be ascribed to the carbohydrate units of nervous tissue glycoproteins or glycoproteins in general. The structural heterogeneity rather suggests that there are many different biological roles. A detailed knowledge of the structural properties is a prerequisite for the elucidation of the diverse biological roles the carbohydrate units of nervous tissue glycoproteins may be involved in.

References [I] Brunngraber, E.G. (1972) In: A.N. Davison, P. Mandel and I.G. Morgan (Eds.), Functional and Structural Proteins of the Nervous System. Plenum Press, New York, pp. 109-133. [2] Montreuil, J. and Vliegenthart, J.F.G. (1995) In: J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins, New Comprehensive Biochemistry, Vol. 29A. Elsevier, Amsterdam, Chapter 2. [3] Finne, J. (1990) In: G.M. Edelman, B.A. Cunningham and J.P. Thiery (Eds.), Morphoregulatory Molecules. Wiley, New York, pp. 81-1 16. [4] Margolis, R.K. and Margolis, R.U. (1979) In: R.U. Margolis and R.K. Margolis (Eds.), Complex Carbohydrates of Nervous Tissue. Plenum Press, New York, pp. 45-73. [5] Finne, J. (1989) Ciba Found. Symp. 145, 173-188. [6] Rougon, G., NCdelec, J., Malapert, P., Goridis, C. and Chesselet, M.F. (1990) Acta Histochem. (Jena) 89 (SUPPI. 38), 51-57. [7] Wing, D.R. (1993) Glycoconjugate J. 10, 190-194. [8] Zanetta, J.P., This volume, Chapter 16. [9] Finne, J. and Krusius, T. (1982) Methods Enzymol. 83, 269-277. [lo] Cummings, R.D. (1994) Methods Enzymol. 230, 66-86, [ I l l Hardy, M.R. and Townsend, R.R. (1994) Methods Enzymol. 230, 208-225. [I21 Hase, S. (1994) Methods Enzymol. 230, 225-237. [I31 Baenziger, J.U. (1994) Methods Enzymol. 230, 237-249. [I41 Krusius, T., Finne, J. and Rauvala, H. (1976) FEBS Lett. 72, 117-120. [I51 Krusius, T.and Finne, J. (1977) Eur. J. Biochem. 78, 369-379. [I61 Krusius, T., Finne, J. and Rauvala, H. (1978) Eur. J. Biochem. 92, 289-300. [I71 Finne, J., Tao, T.-W. and Burger, M.M. (1980) Cancer Res. 40, 2580-2587. [I81 Finne, J. (1975) Biochim. Biophys. Acta 412, 317-325. [I91 Finne, J., Krusius, T., Margolis, R.K. and Margolis, R.U. (1979) J. Biol. Chem. 254, 10295-10300. [20] Krusius, T., Finne, J., Margolis, R.K. and Margolis, R.U. (1986) J. Biol. Chem. 261, 8237-8242. [21] Krusius, T. and Finne, J. (1978) Eur. J. Biochem. 84, 3 9 5 4 0 3 . [22] Finne, J. and Krusius, T. (1979) Eur. J. Biochem. 102, 583-588. [23] Finne, J., Breimer, M.E., Hansson, G.C., Karlsson, K.-A,, LefRer, H., Vliegenthart, J.F.G. and Van Halbeek, H. (1989) J. Biol. Chem. 264, 5720-5735. [24] Hynes, M.A., Gitt, M., Barondes, S.H., Jessell, T.M. and Buck, L.B. (1990) J. Neurosci. 10, 1004-1013. [25] Zanetta, J.-P., Kuchler, S., Lehmann, S., Badache, A,, Maschke, S., Marschal, P., Dufourcq, P. and Vincendon, G. (1992) Int. Rev. Cytol. 135, 123-154. [26] Hakomori, S.A. (1992) Histochem. J. 24, 771-776. [27] Feizi, T. (1993) Curr. Opin. Struct. Biol. 3, 701-710. [28] Yuen, C.-T., Bezouska, K., O’Bnen, J., Stoll, M., Lemoine, R., Lubineau, A., Kiso, M., Hasegawa, A., Bockovich, N.J., Nicolaou, K.C. and Feizi, T. (1994) J. Biol. Chem. 269, 1595-1 598. [29] Kucherer, A., Faissner, A. and Schachner, M. (1987) J. Cell Biol. 104, 1597-1602.

67 [30] Horstkorte, R., Schachner, M., Magyar, J.P., Vorherr, T. and Schmitz, B. (1993) J. Cell Biol. 121, 14091421. [31] Finne, J. and Krusius, T. (1976) FEBS Lett. 66, 94-97. [32] Kiang, W.-L., Krusius, T., Finne, J., Margolis, R.U. and Margolis, R.K. (1982) J. Biol. Chem. 257, 16511659. [33] Leppanen, A., Korvuo, K., Puro, K. and Renkonen, 0. (1986) Carbohydr. Res. 153, 87-95. [34] Yuen, C.-T., Chai, W., Loveless, R.W., Lawson, A.M., Margolis, R.U. and Feizi, T. (1997) J. Biol. Chem. 272, 8 9 2 4 8 9 3 1. [35] Margolis, R.K. and Margolis, R.U. (1993) Experientia 49, 4 2 9 4 4 6 . [36] Finne, J., Krusius, T., Rauvala, H., Kekomaki, R. and Myllyla, G. (1978) FEBS Lett. 89, 1 1 1-1 15. [37] Muramatsu, T., Gachelin, G., Nicolas, J.F., Condamine, H., Jakob, H. and Jacob, F. (1978) Proc. Natl. Acad. Sci. USA 75, 2315-2319. [38] Berra, B., De Gasperi, R., Rapelli, S., Okada, S., Li, S.-C. and Li, Y.-T. (1986) Neurochem. Pathol. 4, 107-1 17. [39] Spillmann, D. and Finne, J. (1987) J. Neurochem. 49, 874-883. [40] Viitala, J. and Finne, J. (1984) Eur. J. Biochem. 138, 393-397. [41] Spillmann, D. and Finne, J. (1989) Methods Enzymol. 179, 270-274. [42] Friedlander, D.R., Milev, P., Karthikeyan, L., Margolis, R.K., Margolis, R.U. and Grumet, M. (1994) J. Cell Biol. 125. 6 6 9 4 8 0 . [43] Zhu, B.C.R., Laine, R.A. and Barkley, M.D. (1990) Eur. J. Biochem. 189, 509-516. [44] Finne, J., Krusius, T. and Rauvala, H. (1977) Biochem. Biophys. Res. Commun. 74, 405410. [45] Finne, J. (1982) J. Biol. Chem. 257, 11966-1 1970. [46] Jennings, H.J. (1990) Curr. Top. Microbiol. Immunol. 1.50, 97-128. [47] Inoue, S. and Iwasaki, M. (1978) Biochem. Biophys. Res. Commun. 83, 1018-1023. [48] Livingston, B.D., Jacobs, J.L., Click, M.C. and Troy, F.A. (1988) J. Biol. Chem. 263, 9443-9448. [49] Finne, J., Bitter-Suermann, D., Goridis, C. and Finne, U. (1987) J. Immunol. 138, 44024407. [SO] Edelman, G.M. (1985) Annu. Rev. Biochem. 54, 135-169. [51] Roth, J., Kempf, A,, Reuter, G., Schauer, R. and Gehring, W.J. (1992) Science 256, 6 7 3 4 7 5 . [52] Finne, J., Finne, U., Deagostini-Bazin, H. and Goridis, C. (1983) Biochem. Biophys. Res. Commun. 112, 482487. [53] Zuber, C., Lackie, P.M., Catterall, W.A. and Roth, J. (1992) J. Biol. Chem. 267, 9965-9971. [54] Finne, J., Leinonen, M. and Makela, P.H. (1983) Lancet 2, 355-358. [SS] Finne, J. and Makela, P.H. (1985) J. Biol. Chem. 260, 1265-1270. [56] Jennings, H.J., Roy, R. and Michon, F. (1985) J. Immunol. 134, 2651-2657. [57] Hayrinen, J., Bitter-Suermann, D. and Finne, J. (1989) Mol. Immunol. 26, 523-529. [58] Pelkonen, S., Pelkonen, J. and Finne, J. (1989) J. Virol. 63, 4409-4416. [59] Brisson, J.-R., Baumann, H., Imberty, A,, Perez, S. and Jennings, H.J. (1992) Biochemistry 31, 49965004. [60] Sadoul, R., Him, M., Deagostini-Bazin, H., Rougon, G. and Goridis, C. (1983) Nature 304, 347-349. [61] Hoffman, S. and Edelman, G.M. (1983) Proc. Natl. Acad. Sci. USA 80, 5762-5766. [62] Yang, P., Yin, X. and Rutishauser, U. (1992) J. Cell Biol. 116, 148771496, [63] Chou, D.K.H., Ilyas, A.A., Evans, J.E., Costello, C., Quarks, R.H. and Jungalwala, F.B. (1986) J. Biol. Chem. 261, 11717-11725. [64] Hall, H., Liu, L., Schachner, M. and Schmitz, B. (1993) Eur. J. Neurosci. 5, 3 4 4 2 . [65] Streit, A., Faissner, A,, Gehrig, B. and Schachner, M. (1990) J. Neurochem. 55, 1494-1506, [66] Krusius, T., Finne, J., Karkkainen, J. and Jarnefelt, J. (1974) Biochim. Biophys. Acta 365, 80-92. [67] Jessell, T.M., Hynes, M.A. and Dodd, J. (1990) Annu. Rev. Neurosci. 13, 227-255. [68] Mohan, P.S., Laitinen, J., Merenmies, J., Rauvala, H. and Jungalwala, EB. (1992) Biochem. Biophys. Res. Commun. 182, 689-696. [69] Cole, G.J., Loewy, A. and Glaser, L. (1986) Nature 320, 445447. [70] Parkkinen, J., Rogers, G.N., Korhonen, T., Dahr, W. and Finne, J. (1986) Infect. Immun. 54, 3 7 4 2 . [7 1 ] Liukkonen, J., Haataja, S., Tikkanen, K., Kelm, S. and Finne, J. (1992) J. Biol. Chem. 267,2 I 105-2 I I 1 I .

This Page Intentionally Left Blank

J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins II 0 1997 Elsevier Science B.V. All rights reserved CHAPTER 4

Glycosyl-phosphatidylinositol anchors: structure, biosynthesis and function Robert N. Cole and Gerald W. Hart Uniuersity of Alabama at Birmingham, Schools of Medicine and Dentistry, Department of Biochemistry and Molecular Genetics. 404 Basic Health Sciences Building, 1918 Uniuersity Bouleuard, Birmingham, AL 35294-0005, USA

Abbreviations CDP

cystidine 5’-diphosphate

IL-2

interleukin-2

Con A

concanavalin A cross-reacting determinant decay-accelerating factor

Ino

inositol

IPG

inositol phosphate glycan

LPG

CRD DAF

EthN

phosphoethanolamine

Man MDCK

FRT

Fischer rat thyroid

N-CAM

GDP GlPL

guanosine diphosphate

NEM 0-11

GPI-PLC

glycoinositolphospholipid glucosamine N-acetylglucosamine glycosyl-phosphatidylinositol GPI-specific phospholipase C

PIG-A

GPI-PLD

GPI-specific phospholipase D

PNH

lipophosphoglycan lipopeptidophosphoglygan mannose Madden-Darby canine kidney neural cell adhesion molecule N-ethylmaleimide 10-propoxydecanoic acid phosphatidylinositol phosphatidylinositol-specific phospholipase C phosphatidylinositol glycan Class A paroxysmal nocturnal hemoglobinuria

GTP

guanosine 5’-triphosphate

UDP

uridine 5’-diphosphate

GTPyS

guanosine 5’-0-(3-thiotriphosphate)

VSG

variant surface glycoprotein

IL-1

interleukin-I

Dol-P

dolichyl-phosphate Dol-P-Man dolichyl-phosphoryl-mannose

GlcN GlcNAc GPI

LPPG

PI PI-PLC

I . Introduction Residing in the outer leaflet of the plasma membrane of all eukaryotic cells are a class of glycolipids termed glycosyl-phosphatidylinositols (GPIs). These glycolipids anchor proteins with a wide variety of structures and functions to the external surface of the plasma membrane. Lipid anchors for membrane proteins were initially demonstrated by the ability of phosphatidylinositol-specific phospholipase C to solubilize alkaline phosphatase from various mammalian tissues [ 1,2]. Later, the observation that fatty acids and ethanolamine were attached to the C-terminal of rat Thy-1 provided direct evidence for a glycolipid anchor being covalently attached to protein [3]. Since then numerous GPI-linked membrane proteins have been identified and the structure of several GPI anchors have been determined, the first being that on the trypanosome variant surface 69

70

glycoprotein [4]. Many details concerning the biosynthetic pathway of GPI anchors are known and have been extensively reviewed[5-241, yet most of the enzymes remain unidentified. In this chapter, we describe (1) the core structure that appears to be common to all GPI anchors, indicating points of variation; (2) the biosynthetic pathway identified for trypansomal and mammalian GPI anchors; and (3) discuss the many proposed functions for GPI anchors ranging from membrane mobility to signal transduction.

2. Structure of GPI anchors 2.1. Common core structure of GPI anchors The core glycan of all characterized GPI anchors share a common structure (Fig. 1). Phosphatidylinositol is glycosidically linked through carbon 6 of the inositol ring to the reducing end of one glucosamine followed by three mannose residues in a l - 4 , a l 6, a1-2 giycosidic linkages. The third mannose at the non-reducing end of this linear tetrasaccharide is coupled through carbon 6 to phosphoethanolamine in a phosphodiester bond. The C-terminal end of the protein is then attached in an amide linkage to the phosphoethanolamine. This complete structure was first described for the trypanosome variant surface glycoprotein (VSG) [4] thanks to the discovery that the cell surface of trypanosomes is covered with 10 million copies of VSG, thus enabling the purification of milligrams of GPI anchors [25,26] for structural analysis. 2.2. Structural diversity of GPI anchors Substitutions on this common core structure provide the heterogeneity in GPI anchors (Fig. 1). The fatty acid chains attached to the phosphoinositol vary in chain length and saturation, consisting of diacylglycerol (e.g. VSG [4] or Torpedo acetylcholinesterase [27]), alkyl-acylglycerol (e.g. human erythrocyte acetylcholinesterase [28], decay accelerating factor [29] or foliate-binding protein [30]), stearoyl-lysoglycerol (e.g. trypanosome procyclic acidic repetitive protein [3 l]), or ceramide (e.g. slime mold and yeast GPIs [32,33]). Furthermore, the inositol may be acylated with an additional fatty acid, usually palmitate [28,29,3 I], presumably at positions 2 or 3 of the inositol ring [34]. Further modifications of GPI anchors arise from additions to the glycan core. The first mannose may contain a branched chain of a-linked galactoses (e.g. VSG [4]) or a (3-linked N-acetylgalactosamine (e.g. Thy- 1 [35]). The first and second mannoses may carry additional phosphoethanolamines (e.g. human erythrocyte acetylcholinesterase [36] and decay accelerating factor [29]). The non-reducing third mannose can accommodate an additional a-linked mannose, thus extending the linear tetrasaccharide core by one residue (e.g. Thy- 1 [35] and trypanosome IG7 [37]). Moreover, GPI core glycans substitutions may possess terminal sialic acids (ref. [38]; e.g. trypanosome procyclic acidic repetitive protein [39]). Clearly the structural diversity of GPI anchors has only just begun to be described. Other GPI-related lipids that do not anchor proteins are found on the surface of Leishmania and TYypanosoma parasites. These include lipophosphoglycans (LPGs), glycoinositolglycerolipids that anchor phosphorylated polysaccharides [40,4 13; low molecular mass

71

0

II

H,N - P r O t e i M - N H 2

I y

2

7"

'

- 0-P=OP

Phosphoethanolamine

I

Man(a1-6)

H

O

H

T

Man(a1-4) GlcN(a1-6)

0

R1, R2 = Fatty aciL or Cerarnide

- 0-P=OI

= f Fatty acid at Cz or C3 R3 R4,R7 = f Phosphoethanolarnine

RS

R6 Rg

-

f Gal(ui - ~ ) G a l (-e)[Gal(ai ~i -z)lGal(=i - 3 )

= f GalNAc(p1-4)

I

n

0

I

v)

CHZCH-CY

I I

R1

0

c 0

%

= f Man(ai-2)

Fig. 1. The structure of GPI anchors. All characterized GPl anchors share a common core consisting of ethanolamine-P04-6Man(a1-2)Man(a 14)Man(a1-4)GlcN(a 1d)myo-Ino- 1-PO4-lipid. Heterogeneity in GPI anchors is derived from various substitutions of this core structure and are represented as R groups. R1 and R2 may be long chain fatty acyl or alkyl groups or ceramide. R, is frequently palmitate positioned at either C2 or C3 of the inositol ring. R4 and R7 are additional phosphoethanolamines. R5,R6 and R8 may be various glycan substitutions. The few known substituted glycan structures are indicated. (Redrawn from ref. [ 181.)

glycoinositolphospholipids (GIPLs), GPI anchors not linked to protein or polysaccharides [4 11; and lipopeptidophosphoglygans (LPPGs), glycophosphosphingolipidswith the mannose-glucosamine-phosphoinositolstructure of the GPI glycan core that also do not anchor protein or polysaccharides [42,43]. Expression of these glycolipids is stage- and

12

species-specific [44-47] and is thought to aid in the parasite’s avoidance of its host immune system.

2.3. Identification of a GPI-anchored protein Methods used in analyzing the structure and biosynthesis of GPI anchors (reviewed in refs. [48-SO]) are also useful in detecting GPI-linked proteins. GPI-anchored proteins can be distinguished from integral or peripheral membrane proteins by their solubilization upon specific enzymatic or chemical cleavage in conjunction with detergent partitioning, antibody recognition and metabolic labeling (Fig. 2). Bacterial phosphatidylinositol-specific phospholipase C (PI-PLC) (reviewed in ref. [511) or trypanosomal GPI-specific phospholipase C (GPI-PLC) [52-54] releases GPIanchored proteins from the cell surface leaving behind diacylglycerol. This cleavage not only solubilizes the protein, but also exposes a phosphoinositol-containingcryptic epitope termed cross-reacting determinant [55]. Reactivity of polyclonal antibodies derived against the cross-reacting determinant (0) provide additional evidence for the presence of a GPI anchor. Both PI-PLC and GPI-PLC, however, will not cleave GPI anchors with an additional fatty acid in an ester linkage to the inositol ring unless the acyl chain is first removed by mild alkali treatment (reviewed in ref. [ 161). Nevertheless, these phospholipase C resistant GPI anchors are cleaved by GPI-specific phospholipase D (GPI-PLD) [56]. GPI-PLD is found in mammalian serum, potentially released by keratinocytes in stratified squamous epithelium[S7], and splits the bond between the inositol ring and phosphatidic acid. Hydrofluoric acid (HF) also cleaves GPI anchors between the inositol ring and phosphatidic acid, as well as releases phosphoethanolamines attached to any of the mannose residues [SO]. Another chemical cleaving agent, nitrous acid, breaks the glycosidic bond between the glucosamine and inositol releasing the protein from phosphatidylinositol[50]. Any of the above enzymatic or chemical cleaving agents can be employed to alter the partitioning behavior of a GPI-anchored protein when extracted with Triton X-114. A Triton X-114 solution raised above 20°C separates into a detergent phase and an aqueous phase. GPI-anchored proteins will partition into the detergent phase unless released from the anchor[S8] or unless some characteristics of the protein results in anomalous partitioning behavior [59-611. Finally, GPI-anchored proteins can be metabolically radiolabeled using radiolabeled precursors for the GPI anchor biosynthesis. Lipids, myo-inositol, ethanolamine, glucosamine or mannose have all been used to incorporate a radiolabeled tag into GPI anchors [48].

3. Biosynthesis of GPI-anchored proteins Analysis of GPI anchor biosynthesis was greatly facilitated by the development of a cell free system in trypanosomes [62]. The convenience of this system can not be overstated. It is a highly stable system which can be stored frozen for long periods

73

0

II

HzN- P r o t e i H - N H z

I y

2

y

2

0

I

- O-P=O

-&

HO

HF

HO

NaOH (NH3) I

R6

I

- 0-P=O I

PI-PLC or GPI-PLC NaOH (NH3)

--- --- - - - -I- k 0

-I-

CHZCH-CH2 R l R2

Fig. 2. Enzymatic and chemical cleavage sites of GPI anchors useful in identifying GPI anchored membrane proteins. GPI-PLC, GPI-specific phospholipase C; GPI-PLD, GPI-specific phospholipase D; HF, hydrogen fluoride; HONO, nitrous acid; NaOH (NH,), mild alkali treatment; PI-PLC, phosphatidylinositol-specific phospholipase C. (Redrawn from ref. [178].)

without significant loss in activity and where radiolabeled precursors, that normally are impermeable to living cells, can be used in the absence of the corresponding pools of soluble native precursors. Cell free systems also have been developed in Tonoplasma [63], yeast and mammalian cells [64-661. Although variations occur, such as in the addition

74

of acyl chains and phosphoethanolamines, in all cases the basic sequence of constructing the GPI core and its donors appears to be the same. 3. I . Biosynthesis of GPI anchors The first step in the synthesis of GPI anchors is the transfer of N-acetylglucosamine (GlcNAc) from UDP-GlcNAc to phosphatidylinositol (PI) (Fig. 3). This step appears to be regulated by at least three genes. At present, little is known concerning the enzymes catalyzing GPI anchor biosynthesis. Nevertheless, a panel of eight complementation Phosphatidylinositol (PI) G; ; ; I c NA c

GICNAc(a1-6)PI

GlcN(at-6)PI

k

Dol-P-Man

Man(al-4)GlcN(a1-6)PI

b

Dol-P-Man

Man(a1-6)Man(ar -4)GICN(al-6)PI

k k

Dol-P-Man

Man(al-z)Man(al-6)Man(al-4)GICN(al-6)PI

Phosphatidylethanolarnine

EthN-P-GMan(ai - Z ) M ~ ~ ( ~ I - ~ ) M ~ ~ ( ~ I - -6)PI ~)GICN(~I

J

Fatty

Acid

Remodeling

GPI Anchor Fig. 3. Linear pathway for biosynthesis of trypanosomal GPI anchors. Dol-P-Man, dolichyl-phosphorylmannose; EthN, phosphoethanolamine; GlcN, glucosamine; GlcNAc, N-acetylglucosamine; Man, mannose, OAc-, acetyl; UDF', uridine 5'-diphosphate. (Redrawn from ref. [22].)

75

classes of T-cell mutants deficient in expressing GPI-linked Thy-1 [67] provide some insight. Six of the eight classes are defective at distinct stages of GPI anchor biosynthesis. Of these six classes, three (identified as Class A, C, and H) do not synthesize GlcNAc-PI, but can complement each other in expressing GPI-linked Thy-1 [67]. Thus, three different genes may be required for the synthesis of GlcNAc-PI. Two of these genes have been cloned. The Class A gene consists of 3589 base pairs and codes for a predicted protein of 484 amino acids[68]. The Class H gene, on the other hand, contains 60 base pairs and encodes for a predicted 188 amino acid protein with a molecular mass of 21 kDa [69]. Recently, a third cDNA clone, isolated from a temperature-sensitive yeast mutant deficient in GlcNAc-PI synthesis, is reported to restore GlcNAc-PI synthesis [70]. It is not known, however, if any of these genes possess glycosyltransferase activity. In trypanosome cell lysates the N-acetylglucosaminyltransferase is inactivated by sulfhydryl alkylating reagents such as N-ethylmaleimide (NEM) [7 11. Inactivation by NEM is specifically blocked by UDP-GlcNAc, suggesting a sulfhydryl group near or at the active site. Dithiothreitol stimulates the synthesis of GlcNAc-PI (Cole, unpublished). This enzyme activity also appears to be metal independent (Doering and Cole, unpublished) which is unusual for nucleotide binding proteins. The next step in GPI anchor biosynthesis is the rapid deacetylation of GlcNAc-PI to form glucosamine-PI [72]. Incorporation of a non-acetylated glucosamine is unique to GPI anchors. The deacetylation of GlcNAc-PI in lymphoma cell lysates and microsomes is stimulated by GTP and this stimulation is blocked by GTPyS, suggesting that hydrolysis of GTP is required for the stimulatory effect [73]. Thy-1 mutant complement Classes A, C and H are able to deacetylate exogenous GlcNAc-PI [73] even though they are unable to synthesize GlcNAc-PI. Thus, it appears that deacetylation is not coupled to GlcNAc transferase activity. After deacetylation, either the linear string of mannoses are attached, as in blood stream trypanosomes [62,74], or the inositol ring is acylated followed by the mannose additions, as in yeast [33,75] and mammalian cells [76,77]. All three mannoses appear to be transferred from dolichyl-phosphoryl-mannose (Dol-P-Man). Incorporation of mannose from exogenously added GDP-mannose is blocked by amphomycin [78], which inhibits Dol-P-Man synthesis [79], and is stimulated by addition of Dol-P [63]. Dol-P-Man synthase is absent from Class E Thy-1 mutants [66], however, GPI anchor synthesis can be restored by transfecting the yeast gene for Dol-P-Man synthase into this mutant cell line [SO]. These results suggest that mannose is transferred from GDP-mannose to Dol-P before transfer to the glucosaminyl GPI. Mannosamine inhibits GPI anchor biosynthesis [8 11. Mannosamine incorporates almost exclusively into the second mannose position of the GPI[82] and prevents further elongation of the chain, thus, acting as a chain terminator [83,84]. Addition of phosphoethanolamine to the third mannose is the final step in the biosynthesis of common core structure of trypanosomal GPI anchors [62]. In mammalian GPI anchors, however, all three mannose residues may contain a phosphoethanolamine [85-871. Although the sequence of mannose and phosphoethanolamine incorporation remains unclear in mammalian cells, there appears to be a ladder arrangement between two parallel pathways (Fig. 4). One rail of the ladder proceeds

76 Phosphatidylinositol

(PI)

t GlcNAc(o1-6)PI

+

GlcN(o1-6)PI

t

GlcN(a1 -6)Pl(acyl)

J Man(oi-4)GIcN(ai-6)Pl(acyl)

i

+

Man(a1-4)GlcN(al -6)Pl(acyl) EP-2

I

t

Man(a1-6) Man(a1-4)GlcN(al -6)Pl(acyl) EP-z

I

Fig. 4. Proposed branched pathway for biosynthesis of mammalian GPI anchors. Abbreviations as in Fig. 3. Acylation of inositol ring is indicated by (acyl). EP, phosphoethanolamine. (Redrawn from ref. [ 8 5 ] . )

with the mannose additions prior to incorporation of phosphoethanolamine, whereas the parallel rail begins instead with a phosphoethanolamine attached to the first mannose [85]. Precursors from the former pathway may enter the latter pathway at the corresponding rungs of mannose or phosphoethanolamine additions, resulting in a GPI anchor with multiple phosphoethanolamines. The most likely candidate for the phosphoethanolamine donor is phosphatidylethanolamine. Metabolic radiolabeling studies demonstrate that [3H]ethanolamine is incorporated into CDP-phosphoethanolamine before it is transferred to phosphatidylethanolamine and attached to the GPI core glycan. CDP-ethanolamine, however, is not a donor since it is not required for, nor does it affect, GPI anchor biosynthesis [62]. Consistent with this observation is the lack of [3H]glucosamine incorporation into GPI anchors in yeast mutants that do not synthesize phosphatidylethanolamine from CDP-ethanolamine yet can construct GPI anchors [88]. Direct evidence for a phosphatidylethanolamine donor is still lacking. As with the other enzymes in the GPI anchor biosynthetic pathway, little is known concerning the phosphoethanolamine transferase. However, the gene responsible for the defect in phosphoethanolamine transfer in the Class F Thy-1 mutants has been cloned [89]. The 917 base pair sequence encodes for a predicted hydrophobic protein of 2 19 amino acids. Phosphoethanolamine transferase activity in trypanosome cell lysates is inhibited by active site serine-directed inhibitors, such as phenylmethylsulfonyl fluoride and diisopropylfluorophosphate [90].

I1

3.2. GPI anchor remodeling Prior to protein attachment, the fatty acids of the GPI anchor may be replaced in a conversion process termed fatty acid remodeling [91]. Although myristate is the sole fatty acid component in mature trypanosome GPI anchors, earlier GPI intermediates contain more hydrophobic stearate fatty acids. These longer fatty acid chains are replaced by an alternating sequence of removal and replacement of a fatty acid from each position on the glycerol [91]. Lipid remodeling may not be unique to trypanosome GPI anchor biosynthesis since mature yeast GPI anchors contain ceramide, whereas an early yeast GPI intermediate has a diacylglycerol species [33]. Interestingly, the requirement for myristate in bloodstream trypanosome GPI anchors is highly restrictive. Substitution with a closely related analog, 10-propoxydecanoic acid (0-1 l), is highly toxic to the trypanosomes [92]. Because trypanosomes neither make or store myristate[93] but rely on this relatively rare fatty acid in mammalian serum [94], and that analogs, such as 0-11, are selectively toxic to trypanosomes, new anti-trypanosomal chemotherapeutic strategies may be developed [92]. Further substitutions of the glycan core, such as additions of GlcNAc and Gal(a12)Gal(a1-6)[Gal(al-2)]Gal(a1-3) (Fig. l), appear to occur after the protein is attached to the GPI anchor [95,96], presumably in the Golgi.

3.3. GPI anchor attachment to protein Proteins destined to be GPI-linked are synthesized in the endoplasmic reticulum and attached to preassembled GPI anchors within 1 min after the protein’s synthesis [97,98]. The newly synthesized C-terminal sequence is rapidly replaced by the GPI anchor in what appears to be a transamidase reaction [99]. ATP and GTP stimulate processing of the C-terminal peptide in an in vitro microsome assay system [loo], suggesting that a nucleotide-dependent step occurs prior to transamidation. However, it is not clear what role this nucleotide-dependent step would have in attaching a GPI anchor. Two peptide sequences are necessary for GPI anchor addition, an N-terminal signal peptide directing the nascent protein into the endoplasmic reticulum and a C-terminal signal peptide directing GPI anchor attachment. The signal for GPI anchor attachment has been localized to the C-terminal region of GPI-linked proteins [ 101-1031. Yet, comparison of cDNA deduced amino acid sequences of C-terminal regions from GPI-anchored proteins does not identify a clear consensus sequence [5,104-1061. More detailed inspections of the C-terminal sequences of nascent GPI-linked proteins have been carried out to elucidate the signal for cleavage and attachment of the GPI anchor. Based on deletion mutations of decay-accelerating factor, the residues on the N-terminal side of the cleavage/attachment site do not appear to be involved in signaling GPI anchor attachment [ 1071. Using site-directed mutagenesis of placental alkaline phosphatase [ 105,1081, or synthetic constructs of signal sequences attached to non-GPI-linked CD46 [ 1091 has led to the proposal of consensus rules concerning the structure and amino acid composition of the GPI anchor attachment signal. The C-terminal region, cleaved during GPI anchor addition, appears to consist of three distinct regions: a cleavage/attachment site, a spacer domain, and a hydrophobic domain.

78

The cleavage/attachmentsite consists of 3 amino acids. The first amino acid is the site

of attachment and can be glycine, alanine, serine, cysteine, aspartic acid, or asparagine. Only these small amino acids are permissible in chimeric studies [ 105,108,109] and are the only ones reported in the C-terminal tails of sequenced GPI-anchored proteins [5,1041061. Serine is most favorable [ 1091 and is present in over half of the known GPI anchor proteins. The second amino acid, on the C-terminal side, can be any amino acid except proline. In contrast, the third amino acid is restricted to only small amino acids, preferably glycine or alanine, although serine and to a lesser extent cysteine, threonine and valine are possible [105,108]. The spacer domain between the cleavage/attachment tripeptide and the hydrophobic domain is non-specific in amino acid composition but appears to require a length of 812 residues [109]. The hydrophobic domain also does not have a specific amino acid requirement. However, it must have an overall hydrophobic character and a length of at least 1 1 residues [ 109,l lo]. 3.4. Topology of GPI anchor biosynthesis

Evidence that protein addition to GPI anchors occurs in the endoplasmic reticulum stems from kinetic studies on GPI anchor addition to newly synthesized proteins [97,98], the accumulation of unprocessed precursor proteins in the endoplasmic reticulum of yeast and mammalian mutants in GPI anchor biosynthesis [ 11 1,1121, and the use of a microsomal assay system to analyze the C-terminal processing of GPI-linked proteins [ 1131. The biosynthesis of GPI anchors also is assumed to be localized to the endoplasmic reticulum. This assumption is well founded since GlcNAc-PI transferase and deacetylase activity co-fractionate with markers for the endoplasmic reticulum [ 1141. Therefore, at least the initial steps in GPI anchor biosynthesis occur in the endoplasmic reticulum. The topology of GPI anchor biosynthesis has been addressed only recently. Synthesis of GlcNAc-PI and its subsequent deacetylation appears to occur on the cytoplasmic leaflet of the endoplasmic reticulum [ 1141. Radiolabeled GlcNAc-PI and GlcN-PI on intact microsomes can be released by treatment with PI-PLC. Membrane permeabilization with detergent or pores only slightly increased the amount of PI-PLC released radioactivity and PI-PLC treatment alone does not disrupt membrane integrity. Provided that these intact microsomes are properly oriented as to their lumenal and cytoplasmic surfaces, these data strongly indicate that the first steps in GPI anchor biosynthesis take place on the cytoplasmic surface of the endoplasmic reticulum. The surface location of the subsequent additions to these and other GPI intermediates, as well as when they translocate to the lumenal side of the endoplasmic reticulum, remains unknown.

3.5. Defects in GPI anchor biosynthesis: paroxysmal nocturnal hemoglobinuria Paroxysmal nocturnal hemoglobinuria (PNH) is an acquired hemolytic anemia resulting from a somatic mutation in a multipotent hemopoietic stem cell (reviewed in refs. [ 115,1161). The abnormal stem cells produce subpopulations of blood cell types that lack all GPI-anchored proteins. The absence of decay-accelerating factor (DAF) and

79

CD59 renders blood cells susceptible to complement-mediated lysis. Patients with PNH are at risk of developing venous thrombosis, leukemia and aplastic anemia. The biochemical lesion responsible for PNH is in a gene regulating the synthesis of GPI anchor intermediate, GlcNAc-PI [ 117-1 191. Homogeneous cell lines or Epstein-Barr virus immortalized clones of lymphocytes from 16 PNH patients with different heritages are all able to incorporate my~[~H]inositol into phosphatidylinositol but are unable to produce GlcNAc-PI. Thus, the defect appears to be in the synthesis of GlcNAc-PI in these patients. A recent study, however, reported the normal synthesis of GlcNAc-PI and its deacetylation in granulocytes isolated from two other PNH patients [85], suggesting an additional lesion in the GPI biosynthetic pathway associated with PNH. This is unlikely, since only a small proportion (5%) of normal cells contaminating their population of granulocytes would be sufficient to restore normal levels of GlcNAc-PI [ 1181. Restoration of GPI anchor biosynthesis in homogeneous cell lines or immortalized clones of PNH positive lymphocytes is achieved by the hybridization of PNH cells with any of the complementation classes of the Thy-1 deficient lymphoma mutants except Class A [ 117,1191. As previously stated (see section 3.1), the Class A mutants do not synthesize GlcNAc-PI and the lacking gene has been cloned [68]. Transfecting the cDNA of the Class A gene into PNH positive B lymphoblastoid cell lines corrects the defect and restores the expression of GPI-anchored proteins [ 1201. Taken together, these data clearly establishes that a mutation in the Class A (or PIG-A, phosphatidylinositol glycan-Class A) gene is responsible for PNH. Comparison of normal and aberrant PIG-A mRNA identifies a 207 base pair deletion, representing 69 amino acids in the PIG-A gene product [ 1201. I n situ hybridization has localized the PIG-A gene to a p22.1 position on the X chromosome [ 1201.

4. Proposed functions of GPI anchors GPI anchors undoubtedly serve to anchor proteins to the extracellular surface of plasma membranes. Beyond this obvious function, there is much discussion over whether this additional, highly conserved, multiple step, complex mechanism for anchoring membrane proteins might have further functions (reviewed in refs. [ 18,20,22]). Identifying a unique function for GPI anchors, however, is complicated both by not knowing the functions of many of the GPI-anchored proteins and by not being able to establish a single category for those functions which are known. Therefore, any additional functions GPI anchors may play most likely depends on the protein it anchors, including the possibility that they simply anchor the protein to a cell membrane. GPI anchors may provide subtle functions, such as influencing the overall characteristics of the cell membrane. The fatty acid content of the anchor contributes to the lipid composition of the membrane and can determine the membrane packing characteristics of the protein. Membrane-anchored proteins that do not require transmembrane or cytoplasmic domains will, by default, reduce “protein clutter” by not interfering with other molecules in these regions [8,20]. Nevertheless, GPI anchors have been proposed to: (1) allow proteins an increased lateral mobility; (2) mediate the release or secretion of proteins; (3) target proteins to

80

apical surfaces; (4) mediate endocytosis or protein turnover; and ( 5 ) take part in signal transduction of receptor-mediated events. 4.1. Lateral mobility

Without a transmembrane domain, the lateral movement of GPI-anchored proteins is not restricted by interactions with the cytoskeleton. Many GPI-anchored proteins are receptors or adhesion molecules and freedom of movement in the membrane may be advantageous for the interactions with their ligands. GPI anchors, then, may impart an increased lateral mobility to their linked proteins. As attractive as this hypothesis is, there is no strong evidence supporting increased lateral mobility due to a GPI anchor. Although GPI-anchored Thy-1, alkaline phosphatase, and DAF [ 104,12 1,1221 all have lateral diffusion coefficients higher than the average calculated for transmembrane proteins [ 1231, GPI-anchored proteins, trypanosomal VSG [52] and sperm adhesion protein PH-20 [124,125] have lateral diffusion coefficients lower than the average transmembrane protein. Direct comparison of the lateral mobility of chimeric proteins, where the GPI anchor is replaced with a transmembrane domain, demonstrates little change in their mobility [ 126,1271. Removal of asparagine-linked oligosaccharides has a greater effect on the lateral movement of the transmembrane protein H-2L [ 1281 than removal of its cytoplasmic tail [ 1291. Moreover, the lateral mobility of the sperm protein PH-20 changes during development [ 1241. Consequently, GPI anchors may provide increased lateral movement to some proteins, but interactions with other portions of these membrane proteins may have greater effects on their lateral mobility [127].

4.2. Protein release Cleavage of GPI anchors by highly specific phospholipases suggests a potential mechanism for protein release or secretion mediated by GPI anchors [5 1,561. This hypothesis is supported by the presence of soluble and GPI-anchored forms of proteins [ 130,1311, by the shedding of GPI-linked proteins from the surface of cells in culture [132] and by the presence of GPI-PLD in mammalian serum [%I. In addition, it appears that pancreatic granule membrane protein, GP2, is synthesized as a GPI-anchored protein but is secreted anchorless [133]. It is unclear, however, to which extent GPI-linked protein shedding occurs in oiuo, and whether it generally involves proteolytic cleavage or lipase activity. In some cases, proteins may be released from cells without the removal of their anchors [134]. To ascertain whether GPI anchors are directly involved in protein secretion, anchor removal mediated by lipase activity must be demonstrated. 4.3. Protein targeting Localization of GPI-anchored proteins to apical surfaces of epithelial cells suggests that GPI-anchored proteins contain an apical sorting or targeting signal [ 135-1371. Two lines

81

of evidence strongly support this hypothesis. First, GPI-anchored proteins from other cells are targeted to the apical surface when transfected into epithelial cells in culture [ 138-140] or in transgenic mice [141]. And secondly, targeting of GPI-linked and transmembrane proteins is altered by replacement of their membrane anchors. For example, replacing the transmembrane and cytoplasmic regions of vesicular stomatitis virus glycoprotein G [ 1381 or herpes simplex glycoprotein D [ 1391 with a GPI anchor re-routes the proteins from a normally basolateral location to a new apical position. By contrast apical proteins, such as placental alkaline phosphatase, can be re-routed to basolateral locations by the addition of a transmembrane domain [ 1381. And, finally transfection of naturally occurring GPI-linked or transmembrane-bound isoforms of the neural cell adhesion molecule, N-CAM, into MDCK epithelial cells results in the basolateral expression of the transmembrane form, whereas the GPI anchor form is apically targeted [ 1421. Recent evidence indicates that the signal possessed by GPI-anchored proteins may not be specific solely for apical surfaces but for polarized surfaces in general. GPI-anchored Thy- 1 is expressed specifically on axonal membranes rather than evenly distributed on the cell [ 1431. More striking is the demonstration that GPI-anchored proteins previously demonstrated to be targeted to the apical surface of MDCK cells are directed to the basolateral surface of Fischer rat thyroid (FRT) epithelial cells [ 1441. Expression of transfected chimeric fusion proteins in FRT cells indicates that transmembrane anchored proteins are directed to the apical surface, whereas GPI-anchored proteins are targeted to the basolateral surfaces. In these epithelial cells the GPI-anchored proteins appear to have a basolateral signal. Thus, the possession of a GPI anchor may signal the sorting of the protein into cell-specific pathways leading to a polarized surface. The signal for targeting is unknown. It may be associated with the signal for cleavage/attachment of the GPI anchor [ 1451. The sorting of GPI-linked protein also may involve the co-clustering of GPI anchors with apical glycolipids [146,147]. This mechanism has gained support by the observation that GPI-anchored proteins form insoluble complexes with glycosphingolipids, major components of apical membranes, during processing in the Golgi [148]. GPI-anchored proteins can arrive at the cell membrane in these clusters [ 1491. A transmembrane protein is postulated to be necessary for sorting of GPI-anchored proteins, because it is likely that the glycolipid moiety of GPI-anchored proteins can not interact directly with the cytoplasmic sorting machinery. In addition, clustering alone is not sufficient for targeting, since GPI-anchored proteins also cluster in Con A resistant MDCK cells [149] where GPI anchor proteins are not efficiently sorted, and yet other proteins are sorted normally [ 1501. 4.4. Endocytosis and protein turnover

GPI-anchored proteins play a role in a specialized form of endocytosis termed potocytosis. Potocytosis involves the capture and import of scarce extracellular molecules or ions against their concentration gradient through membrane invaginations, called caveolae, independent of the lysosomal pathway (reviewed in ref. [ 1511). Caveolae are 50 nm wide, flask-shaped structures coated with the 22 kDa transmembrane protein, caveolin [ 1521. Caveolae contain clusters of GPI-anchored proteins, most notably, the

82

folate receptor [ 1531. However, there is also evidence for the GPI-anchored proteins alkaline phosphatase [ 1541, Thy- 1 and prion PrP(C) [ 1551 being located in caveolae. High density clusters (30 000 molecules/pm*) of a mixed population of GPI-anchored proteins can reside in caveolae [155]. The structural integrity of these clusters and that of caveolae is mediated by the interactions between GPI anchors and cholesterol. Agents depleting cholesterol levels in the membrane also abolish caveolae and disassociate the clusters [ 1561. A recent study suggests that caveolae may be assembled intracellularly [ 1571. In MDCK cells, a heterogeneous population of transfected GPI-anchored proteins cluster and complex with caveolin within 5 min of caveolin synthesis [157]. By contrast, GPI-anchored proteins are poorly represented in clathirn coated pits [ 1581, the main pathway for receptor-mediated endocytosis. The cytoplasmic tail of most cell suiface receptors mediates their entry into coated pits where they are subsequently internalized and degraded or recycled (reviewed in ref. [ 1591). The lack of a cytoplasmic tail may preclude GPI anchors from entering coated pits, resulting in an increased resident time on the cell surface. Ninety percent of lymphoma Thy-1 is at the cell surface and has a turnover rate an order of magnitude slower than the transmembrane H-2 antigen in the same cells [ 1581. Internalization rates of chimeric GPI-anchored proteins transfected into MDCK cells can be as low as 2% of the internalization rate for the entire cell surface [ 1601. GPI anchors may extend the half-life of cell surface proteins whose functions do not involve internalization [ 181. This is consistent with their presence in caveolae. Ligands are bound by GPI-anchored receptors in open caveolae (reviewed in ref. [ 1511). Then caveolae close and the ligand is released enzymatically or by low pH. A large concentration gradient results from the small volume of in the caveolae. The trapped molecules or ions flow down their concentration gradient into the cytoplasm through membrane carriers or transporters. GPI-anchored proteins are not internalized and when caveolae re-open they are presented for the next round of potocytosis. 4.5. Signal transduction

Activation of lymphocytes may be mediated by GPI anchors. T-cells are activated normally by antigen receptors binding to antigenic peptides presented in association with major histocompatibility proteins. Antibodies to GPI-anchored proteins on T-cells mimic T-cell activation by inducing cell proliferation, IL- 1 and IL-2 production, and other metabolic changes in T-cells (reviewed in ref. [ 151). By contrast, most antibodies to other membrane components of T-cells do not activate the cells. Moreover, pretreatment of lymphocytes with PI-PLC, thereby, releasing GPI-anchored proteins, reduces the response of T-cells to antibody mitogens [1611. Fusion proteins have been engineered to assay the importance of GPI anchors to T-cell activation. Transgenic mice were produced to express normal GPI-anchored histocompatible antigen, Qa-2, or Qa-2 containing the membrane spanning region of histocompatible antigen, H-2 [ 1621. Conversely, transgenic mice also were produced to express a GPI-anchored derivative of H-2. Under normal conditions, antibodies to Qa-2 are mitogenic, whereas antibodies to H-2 are not. T-cell activation is induced by Qa-2 or H-2 specific antibodies binding to the GPI-anchored forms of Qa-2 or H-2, respectively,

83

and not by binding to the transmembrane forms. Likewise, the normal GPI-anchored form of Ly-6E, transiently transfected into lymphocyte cell lines, mediates T-cell activation, while the transmembrane form does not [163]. The signal transduction mechanism involving GPI-anchored proteins that lack intracellular domains is not known. GPI anchor degradation products, inositol phosphate glycan (IPG) and diacylglycerol, derived from phospholipase activity have been implicated in mediating the action of hormones such as insulin, insulin-like growth-factor- 1, nerve growth factor, interleukin-2 and thyroid-stimulating hormone (reviewed in refs. [ 1641671). The hormone-sensitive glycolipids have chemical compositions similar to GPIs. Moreover, inositol glycans derived from trypanosomal GPI mimic metabolic actions of insulin [ 1671, whereas anti-inositol glycan antibodies block the actions of insulin [ 1681. Insulin-sensitive GPIs also appear to mediate T-cell activation in T-cell mutants that are unable to link proteins to GPI anchors [169]. The actions of the released IPG may be subsequently mediated by G-proteins [ 170,1711. Although GPI anchor degradation is a possible mechanism for signal transduction in T-cell activation, it appears unlikely since PI-PLC treatment of lymphocytes does not stimulate, but rather, reduces T-cell activation [ 1613. Alternatively, the GPI-anchored protein may associate with other molecules that can transduce the signal into the cell. An early step in T-cell activation is stimulation of tyrosine kinase activity [172]. Interestingly, protein tyrosine kinases co-immunoprecipitate with antibodies against GPI-anchored proteins [ 173,1741 and co-localize with GPI-anchored proteins in large non-covalent complexes [ 1751. These studies suggest that protein tyrosine kinases are part of the signal transduction mechanism by which GPI-anchored proteins mediate T-cell activation. Contrary to the above results is the recent finding that endotoxin binding to either the GPI-anchored or transmembrane form of CD14 leads to T-cell activation [176]. The authors suggest that this normally GPI-linked protein is not primarily involved in signal transduction but instead is the ligand binding subunit of a membrane-bound receptor complex. They point out that the signal transduction mediated by GPI-anchored proteins may not always be dependent on the GPI anchor itself. In fact, no study has addressed whether the GPI anchor or its attached protein directly interacts with the signal transduction machinery. It has been shown, however, that transmembrane and GPI-anchored forms of the same protein possess functionally different structures [ 1771. Structural changes imposed by GPI anchors on their attached proteins, then, must be considered in assessing the differences between transmembrane and GPI-anchored proteins in T-cell activation.

5. Summary GPI anchors are highly conserved complex structures designed to attach proteins to the external surfaces of all eukaryotic organisms. Much is known about GPI anchor structure and biosynthesis. All characterized GPI anchors share a common core consisting of ethanolamine-P04-6Man(a 1-2)Man(a 14)Man(a14)GlcN(a 1-6)myo-Ino1-P04-lipid. Diversity in GPI anchors is derived from the various substitutions of the fatty

84

acids or additions to the inositol ring and the linear tetrasaccharide core. Biosynthesis of GPI anchors occurs by sequential addition of sugars and phosphoethanolamines to phosphatidylinositol, culminating in its en bloc transfer to protein shortly after the protein is synthesized. Additional glycosylations may occur during processing in the endoplasmic reticulum and Golgi, where most GPI-anchored proteins enter pathways targeting them to specific regions of the cell surface. The challenge for the future is to understand the regulation of GPI anchor biosynthesis. The foremost objective is to identify and characterize the catalytic enzymes directing GPI anchor biosynthesis. There is no question that GPIs serve to anchor proteins onto the plasma membrane. It seems reasonable, however, to assume that such a complex, highly conserved, alternative mechanism to attach proteins to membrane surfaces would have additional functions. The information accumulated in addressing the potential other functions of GPI anchors has yielded new insights to exciting areas of cell biology, including protein sorting and targeting, potocytosis, and receptor-mediated signal transduction. Clarification of the role of GPI anchors in these and other systems is left to the future. The enormous variety in structure and function of proteins possessing GPI anchors indicates that the additional functions of GPI anchors likely will be equally diverse.

References (Literature review completed as per 29 May 1995) [ l ] Ikezawa, H., Yamanegi, M., Taguchi, R., Miyashita, T. and Ohyabu, T. (1976) Biochim. Biophys. Acta 450, 154164. [2] Low, M.G. and Finean, J.B. (1977) Biochem. J. 167, 281-284. [3] Campbell, D.G., Gagnon, J., Reid, K.B.M. and Williams, A.F. (1981) Biochem. J. 195, 15-30. [4] Ferguson, M.A.J., Homans, S.W., Dwek, R.A. and Rademacher, T.W. (1988) Science 239, 753-759. [5] Ferguson, M.A.J. and Williams, A.F. (1988) Ann. Rev. Biochem. 57, 285-320. [6] Low, M.G. and Saltiel, A.R. (1988) Science 239, 268-275. [7] Low, M.G. (1989) Biochim. Biophys. Acta 988, 427454. [8] Cross, G.A.M. (1990) Ann. Rev. Cell Biol. 6, 1-39. [9] Doering, T.L., Masterson, W.J., Hart, G.W. and Englund, P.T. (1990) J. Biol. Chem. 265, 61 1-614. [lo] Lisanti, M.P., Rodriguez-Boulan, E. and Saltiel, A.R. (1990) J. Membr. Biol. 117, 1-10, [I 13 Thomas, J.R., Dwek, R.A. and Rademacher, T.W. (1990) Biochemistry 29, 5413-5422. [ 121 Ferguson, M.A.J. (1 991) Curr. Opin. Struct. Biol. 1, 522-529. [I31 Field, M.C. and Menon, A.K. (1991) Trends Glycosci. Glycotech. 3, 107-1 15. [I41 Mensa-Wilmot, K., Doering, T.L., Raper, J., Buxbaum, L., Hart, G.W. and Englund, P.T. (1991) In: C.C. Wang (Ed.), Molecular and Immunological Aspects of Parasitism. AAAS, Washington, DC, pp. 43-54. [I51 Robinson, P.J. (1991) Imunol. Today 12, 3 5 4 1 . [16] Rosenberry, T.L. (1991) Cell Biol. Int. Rep. 15, 113-1149. [17] Brown, D.A. (1992) Trends Cell Biol. 2, 338-343. [18] Ferguson, M.A.J. (1992) Biochem. SOC.Trans. 20, 243-256. [I91 Hooper, N.M. (1992) Curr. Biol. 2, 617-619. [20] Lublin, D.M. (1992) Curr. Top. Microbiol. Immunol. 178, 141-162. [21] Tartakoff, A.M. and Singh, N. (1992) Trends Biochem. Sci. 17, 470474. [22] Doering, T.L. (1993) In: D.D. Roberts and R.P. Mecham (Eds.), Cell Surface and Extracellular Glycoconjugates: Structure and Function. Academic Press, San Diego. [23] Raper, J., Doering, T.L., Buxbaum, L.U. and Englund, P.T. (1993) Exp. Parasitol. 76, 216-220. [24] Ferguson, M.A.J. (1 994) Parasitol. Today 10, 48-52. [25] Cross, G.A.M. (1975) Parasitology 71, 393417.

85 [26] Ferguson, M.A.J., Low, M.G. and Cross, G.A.M. (1985) J. Biol. Chem. 260, 14547-14555. [27] Butikofer, P., Kuypers, F.A., Shackleton, C., Brodbeck, U. and Steiger, S. (1990) J. Biol. Chem. 265, 18983-1 8987. [28] Roberts, W.L., Myher, J.J., Kuksis, A,, Low, M.G. and Rosenberry, T.L. (1988) J. Biol. Chem. 263, 18766-1 8775. [29] Walter, B.I., Roberts, W.L., Rosenberry, T.I., Ratnoff, W.D. and Medof, M.E. (1990) J. Immunol. 144, 103G1036. [30] Luhrs, C.A. and Slomiany, B.L. (1989) J. Biol. Chem. 264, 21446-21449. [31] Field, M.C., Menon, A.K. and Cross, G.A.M. (1991) EMBO J. 10, 2731-2739. [32] Stadler, J., Keenan, T.W., Bauer, G . and Gerish, G. (1989) EMBO J. 8, 371-377. [33] Conzelmann, A,, Puoti, A,, Lester, R.L. and Desponds, C. (1992) EMBO J. 1 I, 457466. [34] Ferguson, M.A.J. (1992) Biochem. J. 284, 297-300. [35] Homans, S.W., Ferguson, M.A.J., Dwek, R.A., Rademacher, T.W., Anand, R. and Williams, A.F. (1988) Nature 333, 269-272. [36] Deeg, M.A., Humphrey, D.R., Yang, S.H., Ferguson, T.R., Reinhold, VN. and Rosenberry, T. (1992) J. Biol. Chem. 267, 18573-18580. [37] Giither, M.L.S., de Almeida, M.L.C., Yoshida, N. and Ferguson, M.A.J. (1992) J. Biol. Chem. 267, 68204828. [38] Stahl, N. and Prusiner, S.B. (1991) FASEB J. 5, 2799-2807. [39] Ferguson, M.A.J., Murray, P, Rutherford, H. and McConville, N.J. (1993) Biochem. J. 291, 51-55. [40] Orlandi, PA., Jr. and Turco, S.J. (1987) J. Biol. Chem. 262, 10384-10391. [41] McConville, M.J. (1991) Cell Biol. Int. Rep. 15, 779-799. [42] Previato, J.O., Gorin, P.A.J., Mazurek, M., Xavier, M.T., Fournet, B., Wieruszesk, J.M. and MendoniqaPreviato, L. (1990) J. Biol. Chem. 265, 2518-2526. [43] Lederkremer, R.M., Lima, C., Ramirez, M.I., Ferguson, M.A.J., Homans, S.W. and Thomas-Oates, J. (1991) J. Biol. Chem. 266, 23670-23675. [44] Medina-Acosta, E., Karess, R.E., Schwartz, H. and Russell, D.G. (1989) Mol. Biochem. Parasitol. 37, 263-274. [45] McConville, M.J. and Blackwell, J.M. (1991) J. Biol. Chem. 266, 15170-15179. [46] Bahr, V, Stierhof, Y.-D., Ilg, T., Demar, M. and Overath, P (1993) Mol. Biochem. Parasitol. 58, 107122. [47] Schneider, P., Rost, J.-P., Ransijn, A,, Ferguson, M.A.J. and McConville, M.J. (1993) Biochem. I. 295, 555-564. [48] Doering, T.L., Raper, J., Buxabaum, L.U., Hart, G.W. and Englund, PT. (1990) Methods I , 288-296. [49] Mayor, S. and Manon, A.K. (1990) Methods 3, 297-305. [50] Ferguson, M.A.J. (1992) In: A.J. Turner and N.M. Hooper (Eds.), Lipid Modifications of Proteins: A Practical Approach. IRL Press, Oxford, pp. 191-231. [51] Ikezawa, H.C. (1991) Cell Biol. Int. Rep. 15, 1115-1 131. [52] Bulow, R. and Overath, P (1986) J. Biol. Chem. 261, 11918-1 1923. [53] Fox, J.A., Duszenko, M., Ferguson, M.A., Low, M.G. and Cross, G.A.M. (1986) J. Biol. Chem. 261, 15767-15771. [54] Hereld, D.,Krakow, J.L., Bangs, J.D., Hart, G.W. and Englund, PT. (1986) J. Biol. Chem. 261, 1381313819. [55] Holder, A.A. and Cross, G.A.M. (1981) Mol. Biochem. Parasitol. 2, 135-150. [56] Metz, C.N. (1991) Cell Biol. Int. Rep. 15, 875-882. [57] Xie, M., Sesko, A.M. and Low, M.G. (1993) Am. J. Physiol. (Cell Physiol.) 265, CI 1 5 6 1 168. [58] Bordier, C. (1981) J. Biol. Chem. 256, 16041607. [59] Clemetson, K.J., Bienz, D., Zahno, M.L. and Liischer, E.F. (1984) Biochim. Biophys. Acta 778, 463469. [60] Maher, PA. and Singer, S.J. (1985) Proc. Natl. Acad. Sci. USA 82, 958-962. [61] Pryde, J.G. and Phillips, J.H. (1986) Biochem. J. 133, 525-533. [62] Masterson, W.J., Doering, T.L., Hart, G.W. and Englund, PT. (1989) Cell 56, 793-800. [63] Tomavo, S., Dubremetz, J.-F. and Schwartz, R.T. (1992) J. Biol. Chem. 267, 2144621458.

86 [64] Hirose, S., Ravi, L., Hazra, S. and Medor, M.E. (1991) Proc. Natl. Acad. Sci. USA 88, 3762-3766. [65] Stevens, VL. and Raetz, C.R.H. (1991) J. Biol. Chem. 266, 10039-10042. [66] Sugiyama, E., DeGasperi, R., Urakaze, M., Chang, H., Thomas, J.J., Hyman, R., Warren, C.D. and Yeh, E.T.H. (1991) J. Biol. Chem. 266, 12119-12122. [67] Hyman, R. (1988) Trends Genet. 4, 5-8. [68] Miyata, T., Takeda, J., Iida, Y., Yamada, N., Inoue, N., Takahashi, M., Maeda, K., Kitani, T. and Kinoshita, T. (1993) Science 259, 1318-1 320. [69] Kamitani, T., Change, H.-M., Rollins, C., Waneck, G.L. and Yeh, E.T.H. (1993) J. Biol. Chem. 268, 20733-20736. [70] Leidich, S.D., Drapp, D.A. and Orlean, P. (1994) J. Biol. Chem. 269, 10193-10196. [71] Milne, K.G., Ferguson, M.A.J. and Masterson, W.J. (1992) Eur. J. Biochem. 208, 309-314. [72] Doering, T.L., Masterson, W.J., Englund, P.T. and Hart, G.W. (1989) J. Biol. Chem. 264, 11 168-1 1173. [73] Stevens, VL. (1993) J. Biol. Chem. 268, 9718-9724. [74] Menon, A.K., Schwartz, R.T., Mayor, S. and Cross, G.A.M. (1990) J. Biol. Chem. 265, 9033-9042. [75] Costello, L.C. and Orlean, P. (1992) J. Biol. Chem. 267, 8599-8603. [76] Hirose, S., Ravi, L., Prince, G.M., Rosenfeld, M.G., Silber, R. Andresen, S.W., Hazra, S.V. and Medor, M.E. (1992) Proc. Natl. Acad. Sci. USA 89, 6025-6029. I771 Urakaze, M., Kamitani, T., DeGasperi, R., Sugiyama, E., Chang, H. and Yeh, E.T.H. (1992) J. Biol. Chem. 267, 64594462. [78] Menon, A.K., Major, S. and Schwarz, R.T. (1990) EMBO J. 9, 42494258. [79] Kang, M.S., Spencer, J.P. and Elbein, A.D. (1978) J. Biol. Chem. 253, 8860-8866. [80] DeGespari, R., Thomas, L.J., Sugiyama, E., Chang, H.M., Beck, P.J., Orlean, P., Albright, C., Waneck, G., Sambrook, J.F., Warren, C.D. and Yeh, E.T.H. (1990) Science 250, 988-991. [81] Lisanti, M.P., Field, M.C., Caras, I.W., Menon, A.K. and Rodriguez-Boulan, E. (1991) EMBO J. 10, 1969-1 977. [82] Ralton, J.E., Milne, K.G., Guther, M.L.S., Field, R.A. and Ferguson, M.A.J. (1993) J. Biol. Chem. 268, 241 83-24189. [83] Pan, Y.-T., Kamitani, T., Bhuvaneswaran, C., Hallaq, Y., Warren, C.D., Yeh, E.T.H. and Elbein, A.D. (1992) J. Biol. Chem. 267, 21250-21255. [84] Field, M.C., Medina-Acosta, E. and Cross, G.A.M. (1993) J. Biol. Chem. 268, 9570-9577. [85] Hirose, S., Prince, G.M., Sevlever, D., Ravi, L., Rosenberry, T.L., Ueda, E. and Medof, M.E. (1992) J. Biol. Chem. 267, 16968-16974. [86] Kamitani, T., Menon, A.K., Hallaq, Y.,Warren, C.D. and Yeh, E.T.H. (1992) J. Biol. Chem. 267, 2461124619. [87] Ueda, E., Sevlever, D., Prince, G.M., Rosenberry, T.L., Hirose, S. and Medof, M.E. (1993) J. Biol. Chem. 268, 9998-10002. [88] Menon, A.K. and Stevens, V.L. (1992) J. Biol. Chem. 267, 15277-15280. [89] Inoue, N., Kinoshita, T., Orii, T. and Takeda, J. (1993) J. Biol. Chem. 268, 68824885. [90] Masterson, W.J. and Ferguson, M.A.J. (1991) EMBO J. 10, 2041-2045. [91] Masterson, W.J., Raper, J., Doering, T.L., Hart, G.W. and Englund, P.T. (1990) Cell 62, 73-80. [92] Doering, T.L., Raper, J., Buxbaum, L.U., Adams, S.P., Gordon, J.I., Hart, G.W. and Englund, P.T. (1991) Science 252, 185 1-1 854. [93] Mellors, A. and Samad, A. (1989) Parasitol. Today 5, 239-244. [94] Dixon, H. (1967) Trans. R. SOC.Trop. Med. Hyg. 61, 12-13. [95] Bangs, J.D. Andrews, N.W., Hart, G.W. and Englund, P.T. (1986) J. Cell Biol. 103, 255-263. [96] Bangs, J.D., Doering, T.L., Englund, P. and Hart, G.W. (1988) J. Biol. Chem. 263, 17697-17705. [97] Bangs, J.D., Hereld, D., Krakow, J.L., Hart, G.W. and Englund, P.T. (1985) Proc. Natl. Acad. Sci. USA 82, 3207-321 I . [98] Ferguson, M.A., Duszenko, M., Lamont, G.S., Overath, P. and Cross, G.A.M. (1986) J. Biol. Chem. 261, 356-362. [99] Boothroyd, J.C., Paynter, C.A., Cross, G.A.M., Bernards, A. and Borst, P. (1981) Nucl. Acids Res. 9, 47354743.

87

[IOO] Amthauer, R., Kodukula, K., Brink, L. and Udenfriend, S. (1992) Proc. Natl. Acad. Sci. USA 89, 6124-61 28. [I011 Caras, I.W., Weddell, G.N., Davitz, M.A., Nussenzweig, V; and Martin Jr., D.W. (1987) Science 238, 1280-1283. [lo21 Tykocinski, M.L., Shu, H.-K., Ayers, D.J., Walter, E.I., Getty, R.R., Groger, R.K., Hauer, C.A. and Medor, M.E. (1988) Proc. Natl. Acad. Sci. USA 85, 3555-3559. [ 1031 Waneck, G.L., Sherman, D.H., Kincade, P.W., Low, M.G. and Flavell, R.A. (1 988) Proc. Natl. Acad. Sci. USA 85, 577-58 1. [I041 Ishihara, A,, Hou, Y. and Jacobson, K. (1987) Proc. Natl. Acad. Sci. USA 84, 1290-1293. [I051 Gerber, L.D., Kodukula, K. and Udenfriend, S. (1992) J. Biol. Chem. 267, 12168-12173. [I061 Englund, F! (1993) Ann. Rev. Biochem. 62, 121-138. [I071 Caras, I.W., Weddell, G.N. and Williams, S.R. (1989) J. Cell Biol. 108, 1387-1396. [I081 Kodukula, K., Gerber, L.D., Amthauer, R., Brink, L. and Udenfriend, S. (1993) J. Cell Biol. 120, 657664. [I091 Coyne, K.E., Crisci, A. and Lublin, D.M. (1993) J. Biol. Chem. 268, 6689-6693. [ I 101 Berger, J., Howard, A.D., Brink, L., Gerber, L., Hauber, J., Cullen, B.R. and Udenfriend, S. (1988) J. Biol. Chem. 263, 1001610021. [I 1 I ] Conzelmann, A,, Riezman, H., Desponds, C. and Bron, C. (1988) EMBO J. 7, 2233-2240. [112] Delahunty, M.D., Stafford, F.J., Yuan, L.C., Shaz, D. and Bonifacino, J.S. (1993) J. Biol. Chem. 268, 12017-12027. [I 131 Kodukula, K., Amthauer, R., Cines, D., Yeh, E.T.H., Brink, L., Thomas, L.J. and Udenfriend, S. (1992) Proc. Natl. Acad. Sci. USA 89, 49824985. [ I 141 Vidugiriene, J. and Menon, A.K. (1993) J. Cell Biol. 121, 987-996. [I 151 Rosse, W.F. (1992) In: Current Topics in Microbiology and Immunology, Vol. 178. Springer, Berlin, pp. 163-173. [I161 Yomtovian, R., Prince, G.M. and Medor, M.E. (1993) Transfusion 33, 852-873. [I171 Armstrong, C., Schubert, J., Ueda, E., Knez, J.J., Gelperin, D., Hirose, S.. Silber, R., Hollan, S., Schmidt, R.E. and Medof, M.E. (1992) J. Biol. Chem. 267, 25347-25351. [I 181 Hillmen, P., Bessler, M., Mason, PJ., Watkins, W.M. and Luzzatto, L. (1993) Proc. Natl. Acad. Sci. USA 90, 5272-5276. [I 191 Takahashi, M., Takeda, J., Hirose, S., Hyman. R., Inoue, N., Miyata, T., Ueda, E., Kitani, T., Medof, M.E. and Kinoshita, T. (1993) J. Exp. Med. 177, 517-521. [I201 Takeda, J., Miyata, T., Kawagoe, K., Iida, Y., Endo, Y., Fujita, T., Takahashi, M., Kitani, T. and Kinoshita, T. (1993) Cell 73, 703-711. [I211 Noda, M., Yoon, K., Rodan, G.A. and Koppel, D.E. (1987) J. Cell Biol. 105, 1671-1677. [I221 Thomas, J. and Webb, W. (1987) Biophys. J. 51, 522a. [I231 McCloskey, M.A. and Poo, M. (1986) J. Cell Biol. 102, 88-96. [ 1241 Cowan, A.E., Myles, D.G. and Koppel, D.E. (1 987) J. Cell Biol. 104, 9 17-923. [I251 Phelps, B.M., Primakoff, P., Koppel, D.E., Low,M.G. andMyIes, D.G. (1988) Science240, 1780-1782. [I261 Edidin, M. and Stroynowski, 1. (1991) J. Cell Biol. 112, 1143-1 150. [I271 Zhang, T., Crise, B., Su, B., Hou, Y., Rose, J.K., Bothwell, A. and Jacobson, K. (1991) J. Cell Biol. 115, 75-84. [I281 Wier, M. and Edidin, M. (1988) Science 242, 412414. [129] Edidin, M. and Zuniga, M. (1984) J. Cell Biol. 99, 2333-2335. I1301 Medof, M.E., Walter, E.I., Rutger, J.L., Knowles, D.M. and Nussenzweig, V (1987) J. Exp. Med. 165, 848-864. [131] Lebel, D. and Beattie, M. (1988) Biochem. Biophys. Res. Commun. 154, 818-823. [132] Carey, D.J., Stahl, R.C., Asundi, V.K. and Tucker, B. (1993) Exp. Cell Res. 208, 10-18. [I331 Paul, E., Leblond, F.A. and Lebel, D. (1991) Biochem. J. 277, 879-881. [I341 Bublitz, R., Armesto, J., Hoffmann-Blume, E., Schuize, M., Rhode, H., Horn, A., Aulwurm, S., Hannappel, E. and Fischer, W. (1993) Eur. J. Biochem. 217, 199-207. [I351 Lisanti, M.P., Sargiacomo, M., Graeve, L., Saltiel, A.R. and Rodriguez-Boulan, E. (1988) Proc. Natl. Acad. Sci. USA 85, 9557-9561.

88 [I361 [I371 [I381 [I391 [I401 [I411 [I421 [I431 [I441

Ali, A. and Evans, W.H. (1990) Biochem. J. 271, 193-199. Lisanti, M.P., Bivic, A.L., Saltiel, A.R. and Rodriguez-Boulan, E. (1990) J. Membr. Biol. 113, 155-167. Brown, D.A., Brise, B. and Rose, J.K. (1989) Science 245, 1499-1501. Lisanti, M.P., Caras, I.W., Davitz, M.A. and Rodriguez-Boulan, E. (1989) J. Cell Biol. 109, 2145-2156. Wilson, J.M., Fasel, N. and Kraehenbuhl, J.-P. (1990) J. Cell Sci. 96, 143-149. Kollias, G., Evans, D.J., Ritter, M., Beech, J., Morris, R. and Grosveld, F, (1987) Cell 51, 21-31. Powell, S.K., Cunningham, B.A., Edelman, G.M. and Rodriguez-Boulan, E. (1991) Nature 353, 76-77. Dotti, C.G., Parton, R.G. and Simons, K. (1991) Nature 349, 158-161. Zurzolo, C., Lisanti, M.P., Cara, I.W., Nitsch, L. and Rodriguez-Boulan, E. (1993) J. Cell Biol. 121, 103 1-1 039. [I451 Lisanti et al. (1991) J. Cell Sci. 99, 637-640. [I461 Simons, K. and van Meer, G. (1988) Biochemistry 27, 6197-6202. [I471 Simons, K. and Wandinger-Ness, A. (1990) Cell 62, 207-210. [I481 Brown, D.A. and Rose, J.K. (1992) Cell 68, 533-544. [I491 Hannan, L.A., Lisanti, M.P., Rodriguez-Boulan, E. and Edidin, M. (1993) J. Cell Biol. 120, 353-358. [I501 Lisanti, M.P. and Rodriquez-Boulan, E. (1991) Cell Biol. Int. Rep. 15, 1023-1049. [I511 Anderson, R.G.W. (1992) Trends Cell Biol. 3, 69-72. 11521 Rothberg, K.G.C.O., Heuser, J.D., Donzell, W.C., Ying, Y.S., Glenney, J.R. andAnderson, R.G.W. (1992) Cell 68, 673-682. [I531 Rothberg, K.G., Ying, Y.S., Kolhouse, J.F., Kamen, B.A. and Anderson, R.G.W. (1990) J. Cell Biol. 110, 637-649. [154] Jemmerson, R., Klier, F.G. and Fishman, W.H. (1985) J. Histochem. Cytochem. 33, 1227-1234. [I551 Ying, Y . 4 . Anderson, R.G.W. and Rothberg, K.G. (1992) In: Cold Spring Harbor Symp. on Quantum Biology, Vol. 57. Cold Spring Harbor Press, Cold Spring Harbor, pp. 593-604. [I561 Rothberg, K.G., Ying, Y.S., Kamen, B.A. and Anderson, R.G.W. (1990) J. Cell Biol. 111, 2931-2938. [I571 Lisanti, M.P., Tang, Z. and Sargiacomo, M. (1993) J. Cell Biol. 123, 595-604. [I581 Lemansky, P., Fatemi, S.H., Gorcian, B., Meyale, S., Rossero, R. and Tartakoff, A.M. (1990) J. Cell Biol. 110, 1525-1531. [I591 Brodsky, EM., Hill, B.L., Acton, S.L., Nathke, I., Wong, D.H., Ponnambalam, S. and Parham, P. (1991) Trends Biochem. Sci. 16, 208-213. [I601 Lisanti, M.P., Caras, I.W., Gilbert, J., Hanzel, D. and Rodriguez-Boulan, E. (1990) Proc. Natl. Acad. Sci. USA 87, 7419-7423. [161] Stiernberger, J., Low, M.G., Flaherty, L. and Kincade, P.W. (1987) J. Immunol. 138, 3877-3884. [I621 Robinson, P.J., Millrain, M., Antoniou, J., Simpson, E. and Mellor, A.L. (1989) Nature 342, 85-87. [163] Su, B., Waneck, G.L., Flavell, R.A. and Bothwell, A.L.M. (1991) J. Cell Biol. 112, 377-384. [I641 Saltiel, A.R. and Cuatrecasas, P. (1988) Am. J. Physiol. 255, CI-CII. [I651 Jacquemin, C. (1991) Biochemie 73, 3 7 4 0 . [I661 Saltiel, A.R. (1991) J. Bioenerg. Biomembr. 23, 2 9 4 1 . [I671 Misek, D.E. and Saltiel, A.R. (1992) J. Biol. Chem. 267, 1626616273. [I681 Romer, G., Gamez, G., Huang, L.C., Lilley, K. and Lutrell, L. (1990) Proc. Natl. Acad. Sci. USA 87, 1476-1480. [I691 Avila, M.A., Clemente, R. and Varela-Nieto, I. (1992) Biochem. J. 282, 681-686. [I701 Martiny, L., Dib, K., Haye, B., Correze, C., Jacquemin, C. and Lambert, B. (1991) FEBS Lett. 286, 105-109. [I711 Kilgour, E., Larner, J. and Romero, J. (1992) Biochem. Biophys. Res. Commun. 186, 1151-1157. [I721 Klausner, R.D. and Samelson, L.E. (1991) Cell 64, 875-878. [I731 Stefanova, I., Horejsi, V., Ansotegui, I.J., Knapp, W. and Stockinger, H. (1991) Science 254, 1016-1019. [I741 Thomas, P.M. and Samelson, L.E. (1992) J. Biol. Chem. 267, 12317-12322. [I751 Cinek, T. and Horejsi, V. (1992) J. Immunol. 149, 2262-2270. [I761 Lee, J.-D., Kravchenko, V, Kirkland, R.N., Han, J., Mackman, N., Moriarty, A., Leturcq, D., Tobias, P.S. and Ulevitch, R.J. (1993) Proc. Natl. Acad. Sci. USA 90, 9930-9934. [I771 Kemble, G.W., Henis, Y.1. and White, J.M. (1993) J. Cell Biol. 122, 1253-1265. [I781 Puoti, A. and Conzelmann, A. (1992) J. Biol. Chem. 267, 22673.

J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins II 0 1997 Elsevier Science B.V. All rights reserved CHAPTER 5

Dictyostelium discoideum glycoproteins: using a model system for organismic glycobiology Hudson H. Freeze, Ph.D. The Burnham Institute, 10901 North Torrey Pines Road, La Jolla. CA 92037, USA

I . Introduction Dictyostelium discoideum is a haploid, eukaryotic amoeba that normally lives a solitary life, but it can be induced to form a multicellular organism. This, together with a unique combination of biochemical, genetic and molecular tools make it a versatile model for cellular and organismic glycobiology. This chapter will review the distribution, structure, and biosynthesis of carbohydrates in Dictyostelium discoideum glycoproteins. It will also describe carbohydrate-specific antibodies that are used to probe sugar functions, and how they may be used for isolating glycosyl transferase genes. The well-characterized classical genetic system has produced scores of haploid and diploid mutants altered in every conceivable aspect of Dictyostelium discoideum growth and development. Newly developed technology for targeted gene knockouts or random gene tagging offers a rational and systematic way to alter glycan structure. Like yeast, this system offers strong biochemical, genetic and molecular tools with the added benefit of a multicellular stage. Before presenting this, it is important to introduce the organism more fully.

2. The life cycle of Dictyostelium 2. I . Dictyostelium as a single celled organism

In nature, free-living amoebae phagocytose bacteria and divide every 4 h . In the laboratory, the amoebae can live on bacterial lawns, suspensions, in a simple axenic broth and even on a completely defined medium. It is quite easy to grow >lo" cells per day for biochemical analysis or development. When cells deplete the local food supply and reach a critical density, they begin a complex developmental program [ 11. 2.2. Development in Dictyostelium

Many recent comprehensive reviews describe the molecular basis of some aspects of development [2-6]. Vegetative haploid or diploid amoebae complete a well-defined and synchronous morphogenesis program in 24 h. Each 105-cell organism undergoes a dozen morphological transitions to make at least 4 different cell types including spores. 89

90

Dictyostelium discoideum has many of the classical hallmarks of development in higher systems, such as cell migrations, differential cell adhesions, temporal- and positionalspecific gene induction, and multiple extracellular matrices. All of these aspects are now studied by state-of-the-art methods [7-91. About 300 genes are activated during development including ones for multiple cAMP receptors, signal transduction pathways, cytoskeletal complexes, and novel glycoconjugates [2,3,5,6,10,11]. Some of the genes are activated by cAMP stimulation while others require a threshold amount of a secreted glycoprotein factor. Multicellular development begins when cAMP stimulates starving amoebae to stream into clusters of 104-105 cells [5,6,10,12]. This is accompanied by the expression of at least two separate cell-cell adhesion systems. After streaming into the aggregation centers for several hours, the entire mass becomes encased in an extracellular surface sheath containing glycoproteins and cellulose. This and all subsequent distinct morphological stages of development are shown in Fig. 1, along with the developmental time. Within a few hours, the cells form an apical tip (10h) that elongates vertically until a sheathenshrouded worm-like slug is standing erect (1 1-1 3 h). One or more additional adhesion proteins are now involved in binding the aggregated cells together. The slug-like structure topples over onto the substratum and phototactically migrates (toward the right side of the 14 h panel) for up to several days. The slug continually makes and migrates through the extracellular sheath matrix, leaving it behind as a collapsed tube. This stage of development is convenient for analysis since the slug has positional information that determines cell fate[24]. About 15% of cells at the slug’s anterior tip transform into a 1-2mm stalk by the end of development, while most of the rest of the cells in the slug transform into spores. During the slug-stage, the pre-spore cells are recognizable prior to their full morphological differentiation because they synthesize many spore-specific proteins and store them in a pre-spore vesicle (PSV) along with a polysaccharide that contains Gal, GalNAc and GalA. Later, these components are jettisoned and assembled into a cellulose-containing spore coat [13]. The slug is large enough to be easily dissected for biochemical analysis of pre-stalk and pre-spore cells or for expression of their gene products detected by specific antisera or fi-galactosidase fusion constructs [9,10]. In addition to these two major cell types, some cells resemble the anterior-like pre-stalk cells, but are seemingly misplaced; they are scattered throughout the slug or concentrated at its extreme posterior end [3,4]. A slug stops migrating when exposed to overhead light. It rears back on itself (1 5 h) forming a hemisphere that resembles a “Mexican hat” (16-18 h) as a preparatory step for terminal differentiation. A small collar of selected anterior cells in the center of the aggregate then secretes other matrix glycoproteins, synthesizes cellulose and assembles these components into a stalk tube that extends downward to the substratum (1 8 h). More anterior cells then move inside the tube, much like a fountain running in reverse. The anterior cells elongate, evacuolate and construct more of the tube forming a tapering column that lifts the spore mass off the ground (22 h). Finally, this rigid structure supports the remaining cells of the aggregate (22-24 h). The previously dispersed anterior-like cells (ALC) also assemble into three different groups at the base and top of the spore mass and at the base of the stalk. Each pre-spore cell jettisons its load of PSV components into a common pool of precursors that self-assemble into at least three distinct layers of the

91

Fig. 1. 'Stages in the development of Dicfyosrelium discoideum. Synchronous development begins when the food supply is removed. After only a few hours, the cells become adhesive and chemotactically stream into aggregates. Each frame shows a side view of one of the morphological stages of development after all cells have been recruited into an aggregate. The number in the upper right hand comer refers to the time in hours after the induction of development (0 h), and the bar in the upper left indicates the size of the organism in mm. The optional migrating slug-stage (14 h) can be prolonged for days if the cells are developed in the dark (from ref. [197], with permission) or it can be totally by-passed. In the final frame, about 85% of the cells have been transformed into spores which are resting on the tapering stalk. The spores are encased in a protein and cellulose spore coat and the stalk consists of the evacuolated remains of stalk cells surrounded by cellulose tube made by the stalk cells.

spore coat to seal the spores. Desiccated spores are viable for decades. Hydration activates a cellulase that splits the spore coat to liberate the revived amoebae and complete the life cycle.

92

3. Types of protein glycosylation in Dictyostelium 3.1. N-linked oligosaccharides 3.1.1. Biosynthesis of the lipid precursor and processing N-linked sugar chains are synthesized by the common pathway used by other eukaryotic cells beginning with the well-established dolichol-linked Glc3Man9GlcNAc2 oligosaccharide [ 11,141. A set of a-glucosidases removes the Glc residues as in other systems, but after this, oligosaccharide processing is more limited than in animal cells. Most of the oligosaccharides remain as high-mannose-type chains losing only 1 or 2 Man residues. About 12 h into development, some of the N-linked oligosaccharides are processed by a set of two neutral pH processing a-mannosidases [ 151. The enzymes are distinguished by their developmental kinetics, substrate specificity, and sensitivity to 1-deoxymannojirimycin (dMM) and swainsonine (SW). The first enzyme has maximal expression at about 12 h of development, removes Man residues from Man9GlcNAc, but not from pNp a-Man, and is sensitive to dMM, but not to SW. It appears to preferentially cleave a- 1,2-1inked Man residues [Freeze, unpublished]. The second a-mannosidase appears at about 18 h of development, cleaves pNp a-Man and is inhibited by SW, but not by dMM. It generates oligosaccharides of Man3-Man5 size. Although these changes in N-linked oligosaccharide structure are striking, their importance in normal development has not been conclusively shown. Adding inhibitors during development leads to the expected unprocessed sugar chains, but there are no obvious morphological effects on development [ 161. 0-GlcNAc residues are found on neutral and anionic oligosaccharides in an “intersecting” position shown in Fig. 2 [ 17,181. A membrane bound GlcNAc transferase [ 191 can add GlcNAc to the intersecting position of large (bMan5) oligosaccharides or to the more familiar bisecting position of smaller (< Man5) acceptors [ 15,191. a-Fuc residues are found on some neutral oligosaccharides in at least two different locations [20]. One of these is in the core region, probably bound to the reducing GlcNAc, and the other is in a peripheral location(s). The core localization is based on resistance of glycopeptides to EndoH digestion which is abolished by prior mild acid treatment. Some of the N-linked oligosaccharides are also resistant to PNGaseF digestion. Although this enzyme digests nearly all known N-linked oligosaccharides, a Fuc(a 1-3)GlcNAc linkage in the core makes the chains resistant to digestion [21]. Such linkages are found in plants and in a few insect glycoproteins, but they have not been proven to occur in Dictyostelium [22]. Peripheral localization of Fuc is assumed because some fucosylated glycopeptides are sensitive to EndoH digestion without acid treatment. Peripheral Fuc increases at about 8 h of development and continues to rise until culmination, while core Fuc increases slightly during early development and then drops rapidly, decreasing to 10% of that found in vegetative cells. Most of the fucosylated glycans found in the pre-spore and spore regions later in development are probably not on N-linked chains. Compositional analysis of some N-linked chains from membrane glycoproteins in Dictyostelium also show the presence of xylose residues [23]. Carbohydrate-specific monoclonal antibodies against several developmentally regulated glycoproteins are competed by horseradish peroxidase which is known to carry both 8-Xyl and a-Fuc residues

93

A

Man(al+2)Man(al+6)

\

Man(al+6)

Man(a 1+2)Man(a1-13) Glc(a1+2)Glc(a 1+3)Glc(al+3)Man(a

\

/

1+2)Man(a 1+Z)Man(a

Man(p lA)GlcNAc(P 14)GlcNAc

/ 1+3)

Man(al+6)

B Man(alj3)

/

\

Man@ 14)GlcNAc(pl4)GlcNAc

/ Man(al-+2)Man(al-+2)Man(al-13)

C

(MeOP+6)Man(a1+2)Man(al+6)

\ GlcNAc(P 1-+4)Man(al+6)

/

Man(al+2)Man(a 1 4 3 )

\

Man(p 14)GlcNAc

/ (MeOP+6)Man(al+Z)Man(al+3)

Fig. 2. Structures of glycans found in Dictyosfelium discoideum. Structures of known carbohydrate components in Dictyosfelium discoideum. The structures have been determined by physicalkhemical methods or by the analysis of radiolabeled materials. (A) Lipid-linked oligosaccharide precursor of N-linked oligosaccharides [ 141; (B) truncated lipid-linked oligosaccharide precursor found in mutant strains HL241 and HL243 [ 141; (C)oligosaccharide with GlcNAc residues at the “intersecting” position and locations of the methylphosphate esters [16].

on N-linked chains [24]. Sensitivity of some of these Dictyostelium oligosaccharides to EndoF digestion shows that they are also N-linked. There is little solid evidence for additional processing of N-linked chains. One cell adhesion molecule that appears during the early stages of development has been reported to contain typical sialylated diantennary chains [25,26]. Even though the structure was carefully documented, there was no quantitative analysis or proof that the sugar chains were actually derived only from the protein itself and not from extraneous contaminants. This consideration is important since none of the later typical processing enzymes such as N-acetylglucosaminyl, galactosyl, or sialyl transferase activities used in the synthesis of highly processed chains of higher organisms has been identified in Dictyostelium discoideum.

3.1.2. Phospholylation and sulfation of N-linked oligosaccharides In mammalian cells, mannose-6-phosphate (Man-6-P) residues are used for targeting lysosomal enzymes [27,28]. Many of the N-linked oligosaccharides on lysosomal enzymes and secreted glycoproteins in Dictyostelium contain Man-6-P residues [ 1 1,17,29331. These proteins are recognized by the cation-independent Man-6-P receptor (which

94

is also the insulin-like growth factor I1 receptor), but not by the cation-dependent Man-6-P receptor. This differential binding and their easy purification explains why crude mixtures of Dictyostelium enzymes are widely used to distinguish and separate the two receptors [34]. The reason for differential binding is discussed below. The Dictyostelium lysosomal enzymes are very efficiently endocytosed into mammalian cells through the cation-independent receptor [35,361. In mammalian cells, phosphorylation of Man residues is selective for the oligosaccharides on lysosomal enzymes. No phosphomannosyl receptor has been found in Dictyostelium, but the phosphorylation of Man residues appears to be carried out by a similar transferase [37,38]. In mammalian cells, addition of Man-6-P occurs in a two-step pathway [27,28]. In the first step, GlcNAc-1-P is transferred from UDP-GlcNAc to the 6-position of Man residues forming an acid-labile phosphodiester. A phosphodiester a-N-acetylglucosaminidase cleaves the GlcNAc generating Man-6-P. Dictyostelium also has a GlcNAc- 1-P transferase that recognizes similar acceptor oligosaccharides as the enzyme from rat liver and another amoeba, Acanthamoeba castelloni. All require terminal a-1,2-linked Man residues [38] but the different enzymes do not all selectively recognize the protein component. Mammalian and amoebae transferases preferentially phosphorylate the sugar chains on mammalian lysosomal enzymes, but the Dictyostelium discoideum enzyme does not, implying that the specific protein recognition binding site for mammalian lysosomal enzymes is missing in the latter [37,38]. Dictyostelium lysosomal enzymes have not been tested. The second enzyme in the pathway, GlcNAc- 1-P phosphodiester a-N-acetylglucosaminidase (uncovering enzyme) is also present in Dictyostelium [Freeze and Ichikawa, unpublished]; however, very little Man-6-P is found in the phosphomonoester form [29,3 1,321. Instead, Man-6-P occurs as a phosphomethyldiester (Fig. 2) [3 13. This modification still permits binding to the cation-independent receptor, but not to the cation-dependent receptor. The methyl group is donated by a Golgi-associated, S-adenosylmethionine-dependentmethyl transferase [39]. It can use free Man-6-P ( K , =4.3mM), but prefers oligosaccharides with terminal Man-6-P ( K , < 0.2mM) in the a-1,2 linkage to the underlying Man residue [40]. This specificity is very similar to that of GlcNAc-1-P transferase in the first step. This enzyme presumably accounts for the synthesis of the acid-stable methyldiester in Dictyostelium oligosaccharides. A comparison of the Man-6-P biosynthetic pathways in mammalian cells and in Dictyostelium is shown in Fig. 3. Even though Man-6-P occurs in Dictyostelium, it is not known if it is used to target lysosomal enzymes [41,42] since no Man-6-P receptor has been identified. M. Lammertz has recently isolated a group of mutants that are partially (3040%) deficient in GlcNAc- 1-P transferase activity, and also have considerably reduced activities for several lysosomal enzymes [unpublished results]. The mutants show abnormalities in the early part of development, and they can only grow on bacteria, not in liquid medium. The restricted growth condition explains why such mutants were not detected previously. More detailed analysis should help to resolve whether Man-6-P is actually involved in lysosomal enzyme targeting. Present evidence suggests that proteolysis of the precursor form of lysosomal enzymes is critical for targeting [43,44]. Studies using portions of the cloned Dictyostelium 6-hexosaminidase [45] fused to yeast invertase as a reporter have been unable to determine

95

Man-6-P biosynthesis

Receptor binding MPR~'/IGFII MPRCD

Y

1

]P-¤

ti

D-P[

pY1p 1 3

1

1

D-UDP

OCH,-P

[

+

+

AdoMet-CH,

G.f]

P-OCH,

1

+

Dictyostelium

Fig. 3. Biosynthesis of Man-6-P N-linked oligosaccharides and interaction with Man-6-P receptors. This figure compares the biosynthetic pathways of Man-6-P on lysosomal enzymes in mammalian cells and in Dicfyosfelium discoideum. It also shows their binding to the cation-independentlinsulin-like growth factor I1 receptor (Man-6-P/IGFII) and to the cation-dependent mammalian Man-6-P receptor. Enzymes are (1) GlcNAc- 1-P transferase, (2) a-N-acetylglucosaminidae, and (3) M-6-P phosphate methyltransferase. Solid circles, Man; solid squares, GlcNAc.

96

GlcNAc-1-P Trsnsferase

--hAR/P*R time (h) Stage

P

Fig. 4. Developmental regulation of Man-6-P biosynthetic enzymes. The specific activities of the three enzymes used for Man-6-P biosynthesis in Dicwostelium discoideum during development are shown. GlcNAc- 1-P transferase activity is measured using UDP-[3H]GlcNAc and 100mM a-methylmannoside (ManaMe) as an acceptor. Uncovering enzyme, GlcNAcI-PaN-acetylglucosaminidase, was measured using 3HGlcNAc-I -P-ManaMe as a substrate, and Man-6-P methyl transferase activity was measured with S-aden~syl[~H-methyl] methionine and Man(a1-2)Man-a-octyl. The changing levels of the enzymes may reflect differences in the processing of the phosphorylated oligosaccharides.

the region of b-hexosaminidase (@-Hex)needed for lysosomal targeting. However, as little as 15% P-Hex peptide allows the fusion protein to remain in the cell without going to the lysosome [46]. The total amount of Man-6-P containing oligosaccharides varies less than two-fold during development based on the binding of proteins to the cation-independent Man-6-P receptor [47]. Pulse-labeling studies of lysosomal P-glucosidase synthesized during different times in development show that Man-6-P is not added to oligosaccharides synthesized later in development [48], when this enzyme is secreted from the cells. It is possible that the smaller size of the highly processed oligosaccharides generally synthesized later in development cannot serve as appropriate acceptors for phosphorylation. However, the developmental regulation of the GlcNAc- 1-P phosphotransferase, uncovering enzyme, and methyl transferase are distinctly different (Fig. 4) suggesting that the state of esterification may vary during development. As discussed below, some of the phosphate in N-linked chains from later in development is sensitive to phosphatase digestion, indicating that not all of it is in a phosphodiester linkage [24]. Although Man-6-P has been identified on the lysosomal enzymes, total acid hydrolysates of plasma membranes also show a large amount of Man-6-P [23]. The majority of this material is probably in glycophospholipid-type molecules rather than on N-linked chains. Individual lysosomal enzymes and secreted glycoproteins are also rich in sulfate esters, and most of this is found as Man-6-SO4 which was first identified in Dictyostelium [30, 331. Loss of sulfate esters in a mutant strain, HL244, does not affect lysosomal enzyme processing or targeting, but it does reduce secretion of some lysosomal enzymes [49,50]. Phosphorylation precedes sulfation of lysosomal enzymes [39,48,5 I]. Sulfated N-linked oligosaccharides also occur on the cell adhesion molecule gp80, but no structural studies have been performed on them [52]. Davis reported that other types of sulfate esters may be present in various glycans of vegetative cells, but these also have not been characterized

97

[53]. Adding selenate to growing cells arrests growth and blocks sulfation leading to accumulation of unsulfated precursors [54,55]. As in most systems, 3’-phosphoadenosine 5’-phosphosulfate (PAPS) is the activated sulfate donor and Golgi-enriched membrane preparations faithfully carry out in oitro sulfation of endogenous acceptors including lysosomal enzymes and gp80 [5 1,521. Most of the products are released by N-glycanase digestion (85%) and have 1 4 negative charges, and some of them also contain phosphomethyl diesters. Based on the kinetics of acid hydrolysis, the great majority of 35S04 is incorporated into the primary hydroxyl groups of the sugars. Total sulfation of N-linked oligosaccharides decreases steadily during development [20]. Since no specific sulfotransferases have been identified or characterized, it is difficult to know what this means for specific structures.

3.2. Two unusual types of 0-linked protein glycosylation in Dictyostelium Mammalian cells have two major types of 0-linked glycosylation. One is the GalNAc(a1O)Ser/Thr type and the other is Xyl((31-0)Ser which is typical of glycosaminoglycan chains. Dictyostelium also has at least two types of 0-glycosylation, but neither appears to be identical to those of mammalian cells. The first, called phosphoglycosylation, begins with the addition of GlcNAc a l - P to serine in a phosphodiester linkage. In the second type of 0-linked glycosylation, GlcNAc is bound to Thr residues. The glycans occur almost exclusively in the pre-spore and spore cells on secreted, soluble, and glycophospholipid anchored proteins, including many in the extracellular matrix. No complete structures have been solved and no glycosyl transferases have been characterized. Monoclonal antibodies against this type of glycan were used to identify mutant strains that lacked the epitope(s) which are collectively called “modB-dependent” mutants.

3.2.1. Phosphoglycosylation in Dictyostelium Phosphoglycosylation was first reported for a cysteine proteinase called proteinase I which was purified from stationary phase vegetative cells where it accounts for about one percent of the cell protein [57]. It has an acid pH optimum and appears to co-fractionate with lysosomal enzyme markers in phago-lysosomes [56]. A portion also occurs in the cytoplasm where it is tightly bound to an inhibitor [58,59]. GlcNAcl-P was identified by chemical degradations and 3 1 P NMR analysis confirmed the presence of GlcNAc-P-Ser [60,61], comprising -20% of the mass of the protein. Most of the GlcNAc and phosphate in the molecule appear to reside only in the diester linkage. Although previous analysis suggested that no other sugars were present, more recent ones show that the protein has 1 or 2 N-linked chains that lack Man-6-P. Fucose is also present, but not on the N-linked chains. The GlcNAc residue can be removed by mild acid hydrolysis conditions typical for phosphodiesters and by the mammalian phosphodiester glycosidase that converts GlcNAc-P-Man containing phosphodiesters to Man-6-P in cells [62]. Cysteine proteinases are highly conserved in all eukaryotes [64]. The single peptide chain of the mature protein usually has around 200 amino acids and about 50% of them are highly conserved and distributed into multiple domains throughout the protein. The presence of a large number of GlcNAcl-P-Ser residues in proteinase I is inconsistent with

98 --LVLAN---AR----SRPSFHPVSDELVN L-LGVPV--C GA---AELSV NSLEKFHFKS --LGIAS--------ATLTFDHSLEAQWTK L- --VCSSA---VA----QLHK DPTLDHHWHL STLLILSLAF N W a R T N DE-VKAHYES HGLSFGDFSI VGYSQNDLTS TERLIQLFES --TVFVS--SR----GIPP EE--QSQFLE

CATHEPSIN B HUMAN CATHEPSIN H HUMAN CATHEPSIN L HUMAN CATHEPSIN S HUMAN ACTINIDIN PAPAIN CP1 DICTYOSTELIUM

CP5 DICTYOSTELIM Consensus

.. M .------L . .--b.L.. --. . . .

CATHEPSIN B HUMAN CATHEPSIN H HUMAN CATHEPSIN L HUMAN CATHEPSIN S HUMAN ACTINIDIN PAPAIN CP1 DICTYOSTELIUM CP2 DICTYOSTELIUM CP4 DICTYOSTELIW CPS DICTYOSTELIVII

---vEDLKLpAS-------S T K S N Y L RGT------G GFQNRKPRKG KVFQEPLFYE -QWQ=YK SNP-----NR -pMTINSNRY EPR----FGQ S T E L S Y E EVLN-D-GDV -1FTDDLWA DYLDDE-FIN -HSYNGYDGR EVLNVE-DLQ -SALIGTEEE KIF-----S -SSLIGTQEE KVHT-----

Consensus

-

CATHEPSIN B HUMAN CATHEPSIN H HUMAN CATHEPSIN L HUMAN CATHEPSIN S HUMAN ACTINIDIN PAPAIN CP1 DICTYOSTELIUM CP2 DICTYOSTELIUM CP4 DICTYOSTELILM CPS DICTYOSTELI’W Consensus

..

.........

..

..

. . . . . . . . . . .F.. W. . K .. u Y .

9

QGQCGSCWSF ST

....---..

--

DRR- NKG---ILR NVGGAGIKR GTG-NSY GK----SK DRK---SK DR-----

YSSLINPPAF

SR NRD----

...R . . . - - - - N . M . ..i

Fig. 5 . Comparison of conserved and novel sequences of cysteine proteinases in Dictyostelium discoideum. Sequences of four Dictyostelium cysteine proteases and others from mammals and plants deduced from their cDNAs are shown. CP1 and CP2 are found only during development. CP4 and CP5 are produced only during vegetative growth and disappear with the onset of development. All of them show the highly conserved (boxed areas) and non-conserved regions (dashed lines) typical of all eukaryotic cysteine proteases. The arrowheads show the conserved active site Cys and His residues. Potential N-glycosylation sites are double underlined.

99 MF--_----KYLWSEPQL

LTSSLRVPSTYLGFTSGSKYTGS I A G L YYLNNK-EATYLGTRVNATYLGTPFDGTYLGTKFDA-

74 101 93 99 110 115 98 103 96 96

.YL......-

120

______ QW

170 202 199 201 212 216 211 208 199 195 240 283 286 281 283 293 297 291 291 282 277 360

294 297 296 294 304 304 307 339 4 02 307 480 339 334 333 331 379 345 343 376 442 344 557

Fig. 5 (continued). CP4 is very unusual in having an insert of >I00 amino acids distributed into three types of contiguous, re-iterated domains. One domain shows a continuous stretch of Ser residues (SSSS), another contains several repeats of SGQ (SGQ) and finally a GSGS repeat (GSGS). One or more of these regions is probably the site for addition of GlcNAc-1-P residues. CP5 also has shorter versions of each region. All four genes have been mapped to different yeast artificial chromosome fragments in Dicvvostelium. The gene for proteinase I has not been identified, but it probably resembles CP4.

100

the known amino acid composition of these proteins. Two cysteine proteinase genes have recently been cloned from a vegetative cell cDNA library in Dictyostelium [63]. They are named cysteine proteases (CP) 4 and 5. The amino acid sequences of both gene products are shown in Fig. 5 and compared with two other Dictyostelium cysteine proteinases, CPl and CP2, that appear only in development along with several plant and animal cysteine proteinases. All four of the Dictyostelium enzymes share the conserved regions seen in other eukaryotic cysteine proteinases, but both CP4 and CP5 are substantially different from the others. CP4 contains a 1 15 amino acid insert that is rich in Ser residues distributed into several distinct, contiguous domains. CP5 contains the same series of Ser-rich regions, but each is much smaller. It is likely that one of the serine-containing regions would be the sites for addition of GlcNAcl-P, and expression cloning of both CP4 and CP5 confirms that they carry GlcNAcl-P [63]. GlcNAcl-P transferase uses a SGSG-containing peptide as an acceptor [65]. Based on protein sequencing, neither CP4 nor CP5 is the gene coding for proteinase I, but the amino acid composition of proteinase I is similar to that deduced for CP4 [56]. A spore coat protein called SP96 also reacts with rabbit antibodies against GlcNAcl-P and also has contiguous poly-Ser regions. Monoclonal antibodies against fucose also recognize both this protein and proteinase I, however it is not known if the Fuc is bound to GlcNAc in proteinase I. Riley et al. have shown that a small oligosaccharide containing GlcNAc and Fuc can be isolated from SP96 following p-elimination [66]. Carbohydrate-specific rabbit antibodies have been purified by affinity chromatography of the antiserum using immobilized UDP-GlcNAc as a ligand [67]. More recently, monoclonal antibodies against GlcNAc 1-P have also been prepared [Mehta and Freeze, unpublished]. Proteinase I is maintained at a high level during the first six hours of development and then appears to be lost. The function of GlcNAcl-P in proteinase I is not known. It must await the isolation of chemically induced or gene knockout mutants. Studies in our laboratory indicate that Man-6-P and GlcNAcl-P are found on mutually exclusive sets of proteins in the lysosomes [G. Souza, unpublished]. Moreover, immunofluorescence confocal microscopy shows that the two modifications are found in separate vesicles. This raises the exciting possibility that the proteins are sorted via their mutually exclusive carbohydrate modifications.

3.2.2. Another type of 0-linked glycosylation No complete structure has been determined for this type of 0-linked oligosaccharide in Dictyostelium. These carbohydrate chains are resistant to @-elimination, suggesting that they are not the standard type of HexNAcaIpThrlSer linkage. This type of modification has been defined by a panel of monoclonal antibodies. Reactivity to this antibody is lost in a series of mutant strains which are collectively said to be mutated at the modB locus. Although many proteins contain the modB carbohydrate, most of the attention has been on the analysis of two of them. One is a cell surface glycoprotein, called PsA (a.k.a. SP29), which has been cloned and sequenced. The carbohydrate modifications are located on a repeated motif, PTVT as shown by Edman degradation [68]. This sequence is typical of many of those in mammalian cell proteins which also contain 0-linked oligosaccharides [69]. Several allelic variants of this protein from related strains of Dictyostelium discoideum contain 3-5 of these glycosylated repeats [70]. This protein

101

PsA-C=O I NH-CHz-CH2-0 I

o=P-0 I

OH OH

0

I I +/- Man(a 1-2)Man(a 1-2)Man(al-2)Man(a 1-2)GlcNH,( a 1-6)Ino- 1-PO,-CH,-CH-CH-CH I I I 0I NH (CH,),, I I I O=P-0 6

I

0 I CHZ

%:I

O=!

cH3

I

Fig. 6 . Structure of the glycophospholipid anchor on PsA in Dicfyosfelium discoideum. (Adapted from ref. [721.)

has a glycophospholipid anchor which is discussed more fully below. The other molecule that has received most of the attention is a cell adhesion protein called by various names including contact site-A (csA), gp80, or the EDTA-resistant cell adhesion molecule [7 11. This developmentally regulated protein is also glycophospholipid anchored and has a PTVT repeat sequence where the 0-linked chain is added [70]. Other studies on PsA suggest that the 0-linked modB oligosaccharides may be a heterogeneous mixture of several species. A glycopeptide shows the presence of equimolar (-1 1) amounts of GlcNAc and organic phosphate that roughly correspond to the amount of Thr (-15) in the analysis. In addition, smaller amounts of Fuc (-3), Gal and Man (-1 each) were found [72]. These bound sugars are all resistant to B-elimination and not part of the glycophospholipid anchor. The presence of phosphate on the GlcNAc may make the sugar resistant to p-elimination or its resistance may be due to the close packing of the carbohydrates chains. Riley et al. determined that a small glycopeptide generated by pronase digestion of slug stage proteins, contains GlcNAc and Fuc; however, it is released by mild base hydrolysis suggesting that it is not the modB glycan [66]. Further progress on understanding modB glycans requires structural analysis.

3.3. Glycophospholipid anchors The cell adhesion molecules gp80, gp130/138 and PsA (SP29) all contain glycophospholipid (GPI) anchors [71,72]. Another, called ponticulin, has both an anchor plus a series of transmembrane regions [73]. The anchor from the pre-spore protein PsA has been partially characterized and the structure is shown in Fig. 6 [72]. It has structural features of anchors found in yeast, protozoa and higher eukaryotes [74]. The lipid moiety in PsA is an inositolphosphoceramide with a C18:O phytosphingosine and a mixture of fatty acids, the most common being C18:l unsaturated fatty acid. The lipid can be removed by a bacterial phosphatidyl-inositol-specificphospholipase C (PI-PLC). Exoglycosidase

102

digestions and nitrous acid deamination showed that the GlcNH2 is bound to the inositol ring. The anchor containing proteins are probably segregated from those in the rest of the plasma membrane, and this in turn influences their turnover rate [75]. 3.4. Cytoplasmic glycosylation in Dictyostelium Cytoplasmic glycosylation is now well established in mammalian cells. The most common type is a single O-linked P-GlcNAc residue which is found on a large variety of proteins, especially those involved in gene regulation and cell growth [76]. This type of glycosylation has not yet been seen in Dictyostelium, but another more complex type was recently identified. The cytosol of vegetative and developing amoebae contains a single protein called FP21 (fucoprotein of 21 kDa) that accounts for most, if not all, of this type of glycosylation. It was first detected by labeling cells with 3H-fucose followed by subcellular fractionation [77]. The protein appears to have one small (-5 glucose equivalents) anionic O-linked oligosaccharide composed of one residue each of fucose, xylose and two residues of galactose [78]. FP2 1 has recently been cloned. The protein lacks a typical signal sequence and MADLI-TOF-MS analysis of the protein shows that it is about 1 kDa larger than the predicted mass of about 19 kDa based on cDNA sequence [78]. The sugar at the reducing terminus has not been identified, and the nature of the anionic group on the oligosaccharide is unknown. When the protein is isolated from normal cells, it is fully fucosylated since it cannot be an acceptor for in oitro fucosylation. However, in a fucosedeficient mutant strain (HL250) that cannot convert GDP-Man to GDP-Fuc, FP21 lacks fucose residues and serves as a substrate for fucose addition [77]. A novel fucosyl transferase appears to originate from the cytoplasm and not be derived from proteolyzed or damaged membrane vesicles. The K , for GDP-Fuc is -0.35yM, which is considerably lower than most glycosyl transferases found in the Golgi [78]. This lower K , probably allows it to effectively compete with the sugar nucleotide transporters that deliver GDP-fucose to the Golgi lumen. The cytosolic fucosyltransferase has an apparent mass of 95 kDa and can be photoaffinity labeled with GDP-hexanolaminyl-azido-['251]salicylate.Paranitrophenyl-lacto-N-bioside (pNP-LNB, K , = 0.6mM) is preferred over a wide variety of other acceptors and is converted into Fuc(al,2)Gal(P 1,3)GlcNAcb-pNP, although this oligosaccharide has not been detected in Dictyostelium. Fucosylation of FP2 1 is inhibited by Fuc(a 1,2)Gal(P1,3)GlcNAc~-pNP but not by recombinant FP21, suggesting that the substrate specificity is based primarily on carbohydrate recognition [C.M. West, unpublished]. This is the most complex form of cytoplasmic glycosylation yet reported. Since both the native acceptor protein and the fucosyl transferase are novel, studying this pair in Dictyostelium should provide functional insights and the necessary tools to probe for counterparts in higher organisms.

4. Antibodies against glycans and mutants in glycosylation Many of the sugar structures in Dictyostelium are quite unusual compared to those found in higher organisms. Not surprisingly, they are highly immunogenic. Several laboratories

103 Table 1 Examples of mAb against carbohydrate determinants Antibodies

N-linked CA 1

Determinant recognized

Mutant strains

Comments

Multiple Man-6-S04

HL240, HL241

Mostly in lysosomes

Ref(s)

[14,79,82,198] [47,48,80,83]

HL242, HL243 HL244

-

~ 3 1

-

Phosphatase digestion increases recognition by antibody

~ 4 1

Same

-

Same

Same

-

Same

CAB14

Same

CAB1 1

9

-

Same Destroyed by phosphatase digestion

Probably a series of related carbohydrate epitopes

[IOI]

d-4 1

-

DLI 18, DLI 19, HL220, HL216 Same

Same

[201,202]

16.1

-

Same

Same

54.1 32.1

-

Same Same

Same

[ 107,108]

-

Same

[ 105,1641

MUD50

-

Same

Same

40-62-5

GlcNAc/Maltose/aMan GlcNAc

Same Same

Same

?

-

>1 O6 binding siteslcell Same

w11

83.5

Fuc

HL250

Defective in GDP-Fucose production

[ 107,1081

MUD62 81.8

GlcNAc?

Same Bound to WGA

[ I 101 [ 107,1081

40.1

-

-

Bound to WGA, not Con A

[I 121

-

-

Found in trails

[ 105,1641

-

-

GlcNAc-6-SOd (?)

-

CA3

?

-

CAB4

XylIFuc?

CAB10 CAB13

CA2

mods E48D8

40- 178-3

[97,98,200]

Same

Adhesion blocking d-47 d-48

Others

MUD9 MUD52 MUD54

~

4

1

104

have generated antibodies and a diverse and elaborate nomenclature to identify them. A few cross-comparisons of specificity have been made as shown in Table 1. The best evidence for relatedness or identity of different antibodies is based on the loss of the antigens in one or several mutant strains. 4.1. Determinants found on N-linked oligosaccharides

The best characterized immunogenic carbohydrate epitope is called common antigen 1 (CAI) [11,18,50,79,80] which reacts with a cluster of Man-6-SO4 residues on N-linked oligosaccharides [14,81]. Two mutants (HL241 and HL243) that lack the determinant have incomplete, under-sulfated N-linked chains resulting from the synthesis of a truncated lipid-linked oligosaccharide precursor [ 141. The loss of specific Man residues leads to loss of the determinant. These mutant strains target their lysosomal enzymes normally, but more slowly [82]. The primary lesions in the two mutants appear to be different from each other, and the loss of CAI itself does not prevent development [79,198]. Another mutant strain that lacks CA1 (HL244) is about 90% deficient in sulfation and has been discussed. CA1 is not synthesized later in development, but CAI antigens are retained and are about equally distributed between pre-spore and pre-stalk regions of the slug [47]. Two other monoclonal antibodies define CA2 and CA3 on N-linked oligosaccharides, but these are less characterized than CAI [83]. Competition studies suggest that CA2 may recognize GlcNAc-6-SO4 residues. CA3 shows little preference for inhibition by different sugars. Another series of monoclonal antibodies recognize determinants found on N-linked oligosaccharides derived from slug stage plasma membranes on pre-spore or pre-stalk cells [24]. There was no consistent variation in the amount of the antigens during development. Alkaline phosphatase digestion of the antigens either increased or decreased binding of the antibodies, suggesting that phosphate was part of the determinant or partially covered it. Several of the antibodies bound to horseradish peroxidase which has fi-xylose and core a-fucose on N-linked chains [84,85]. The first post-translational mutation found in Dictyostelium discoideum was not identified by screening with antibodies. It was found as a partial deficiency in lysosomal a-mannosidase activity [86,87]. This recessive mutation defines the modA locus that determines the activity of the N-linked oligosaccharide processing a 1,3-glucosidase [88]. Because of this, these mutants fail to add the normal complement of phosphate and sulfate residues to the branch containing the unprocessed glucose residues [891. This results in under-secretion of several lysosomal enzymes and their prolonged residence time in the Golgi compartment [90]. This is different than in mammalian cells where the retention of Glc residues causes retention in the ER [91]. One explanation for this difference is that one of the modifications on this branch is important for efficient trafficking of the enzymes through the system. 4.2. Mutants in 0-linked oligosaccharides - modB

One of the most studied carbohydrate determinants is defined by mutations at the modB locus, which is recognized by a panel of monoclonal antibodies (Table 1).

I05

Mutants lacking the epitope have been used to examine its role in EDTA-resistant cell adhesion [92-951. The structure is not known, but based on the compositional analysis of the glycopeptide from PsA, it is likely to be a heterogeneous group of structures having GlcNAc, Fuc, Man and Gal residues [72]. Proteins that contain the determinant bind to WGA, probably through GlcNAc residues. The modB epitope is 0-linked to Thr residues in a restricted domain containing several repeats of the sequence Pr*Thr-Val-Thr near the carboxyl-terminus [96]. Competition experiments with various sugars show that some of the modB antibodies are competed by GlcNAc, maltose and a-methyl mannoside while others are inhibited only by GlcNAc [97]. This panel of antibodies probably recognize an overlapping group of determinants containing GlcNAc, Man and Glc [97,98]. The compositional analysis of PsA also suggests that the modB-type sugar chains could contain GlcNAc, Man, Fuc residues [72]. Mutant strains have been isolated that lack the modB determinant [96,98]. Although the loss of the determinant results in proteolysis of gp80[71,93,95] and its lowered accumulation at the membrane, several other proteins that normally contain the determinant still accumulate and localize to the plasma membrane [ 1001. The mutants proceed through development and make smaller fruiting bodies that have fewer spores than wild-type [ l o l l . Mutant slugs show reduced traction on their substratum [102], and their spore coat proteins appear to be less firmly integrated into the coat since they can be more easily extracted from the spores [103]. These strains also have several other, but more subtle abnormalities. For instance, when wild-type and mutant cells are mixed and allowed to form chimeric slugs, the mutant cells tend to sort out to the anterior portion of the slug [ 1041. Consequently, they preferentially become stalk cells. This sorting out may be physiologically significant since pre-stalk cells from the wild-type strain also show reduced intercellular adhesion and expression of the determinant compared to spore cells. Even though the role of modB determinants in cell adhesion is not direct, it may be involved in other aspects of cell-cell recognition within the aggregate. Monoclonal antibody MUD50 also recognizes a determinant that is absent in the modB mutants; extensive studies on the proteins recognized by this antibody indicate that they are also found preferentially, but not exclusively, in pre-spore cells [ 1051. 4.3. Antibodies against fucose -mAb 83.5 and MUD62

Antibody 83.5 and MUD62 recognize a group of developmentally regulated proteins that carry Fuc [ 106-1 101. The proteins are highly enriched in pre-spore cells and in proteins found in the PSVs [ 1081. A mutant strain, HL250, fails to react with this antibody [ 1 1 11. In keeping with the earlier nomenclature, this defect is defined by the modC locus. The primary lesion in this strain is a failure to synthesize sufficient GDP-Fuc from GDP-Man [l 1 11. The block in antigen synthesis is alleviated simply by adding exogenous Fuc to developing cells. Two other mutant strains at the modD and modE loci also have defects in adding Fuc residues. Preliminary evidence suggests that strains with modD mutation could be defective in GDP-Fuc transport into the Golgi and those with a modE defect are likely to be missing a specific fucosyl transferase activity, since only some fucosylated proteins are affected [ 1 111. In each case, the absence of the determinant has significant

106

consequences for survival. Spores isolated from this strain appear to have more porous coats and show decreased viability upon storage. 4.4. mAbs 81.8, 40.1 and MUD9 These antibodies react with components that bind to WGA and may recognize GlcNAc residues [107,108]. 81.8 is expressed in vegetative cells and is found on several lysosomal enzymes. The glycans recognized by antibody 40.1 are found in vegetative cells and do not bind to concanavalin A. On Western blots, the reactive material runs as a highly diffuse, pronase resistant smear with a PI of about 3.5. Labeling with sugar precursors shows that it contains GlcNH2 and fucose, is extracted into butanol-saturated water and its binding to phenyl-Sepharose is destroyed by nitrous acid deamination [Freeze, unpublished]. These results suggest that it may be glycolipid, perhaps similar to the protein-free glycophospholipid inositols made by protozoa such as Leishmania [73]. 40.1 and MUD9 binding material is absent from spores and low in the pre-spore region of the slugs. It is enriched in pre-stalk cells, but is not part of the stalk tube [112].

5. Glycoproteins in specijic aspects of the Dictyostelium life cycle 5.1. Mating types in Dictyostelium

Dictyostelium has different mating types that undergo fusion to produce diploid cells. Two of these mating types are HMI and NC-4. A 70 kDa membrane protein of HMl cells that binds to Con A was partially purified on immobilized membrane proteins of NC-4. When these cells become fusion-competent, a specific, 138 kDa protein, appears on the cell surface of both NC-4 and HMl strains [I 131. The protein binds to WGA, Con A, and LCA. Fab fragments of an antiserum against the protein(s) cause complete inhibition of sexual fusion between NC-4 and HMI cells. Two genes called gp138A and gp138B, have been cloned. The amino terminal regions of both are very similar and the C-terminal region is highly hydrophobic, proline-rich and has homology to an analogous portion of gp80 and PsA and is glycophospholipid anchored [114]. The mRNA of gp138A is expressed at the time cells acquire fusion competence and antisense mRNA inactivation of gp138 decreases sexual cell fusion. gp138B is also expressed during growth and may be one of the proteins used for adhesion in early development [71]. 5.2. Getting around

The vegetative amoebae produce an unusual transmembrane protein called ponticulin that binds F-actin and nucleates actin assembly [73,115-1 181. It is present during vegetative growth and during the first 8 h of development as cells stream into the fledgling aggregates. The deduced protein sequence contains no typical a-helical membrane spanning regions, but it has several hydrophobic membrane spanning regions. Also, immunological and actin binding studies show that portions of the protein are present on both sides of the membrane. In addition, it is one of the few known transmembrane

107

proteins that also has a C-terminal glycophospholipid anchor [73]. Deletion of the single ponticulin gene decreases actin nucleation by ten-fold, but the cells still grow and pinocytose normally. However during aggregation, cells lacking ponticulin form aggregates faster than normal and then proceed into an asynchronous morphogenesis, suggesting that ponticulin is involved in cell pattern formation.

5.3. Preparing for development Many genes are activated when development begins. cAMP is essential for activation of some of these, but others require a threshold cell density for their expression. The second type of regulation is probably used by cells to determine when they are actually part of an aggregate, rather than in route toward it. A series of protein factors are secreted by developing cells at very low rates. These factors can also accumulate in the medium of starving cells and are called conditioned medium factors (CMFs). One of these is an 80 kDa glycoprotein which is sequestered in vegetative cells and secreted during early development [I 19-1221. Smaller CMFs are also seen, and the specific activity of the 80kDa CMF increases roughly 100-fold upon proteolysis. The large CMF is reported to contain both N- and 0-linked glycans based on PNGaseF and 0-glycanase digestions. This is the only report of an 0-glycanase sensitive glycan in Dictyostelium [ 1211. Because of the strict specificity of this enzyme, it may mean that Dictyostelium makes typical Gal(P1-3)GalNAc-0 glycans. This needs to be confirmed by more specific analytical methods. The gene coding for the 80kDa CMF has been cloned and found to have three potential N-glycosylation sites within its 57 1 amino acid length. However, complete activity is retained when the complete protein or an 88 amino acid fragment is expressed in E. coli, showing that glycosylation is not required for activity but it may influence stability or proteolytic processing of the 80 kDa CMF. 5.4. Glycoconjugates in cell adhesion during development The developmental stages use at least three distinct adhesion systems. The first appears within 2-3 h following the initiation of development. It is blocked by EDTA and is mediated, in part, by a set of closely related, non-glycosylated proteins which are collectively called gp24. A glycoprotein gp130/138 also appears to be involved in this type of adhesion. This is the same gp138B mentioned above. The second system begins to appear at about 6 h. It is resistant to EDTA and is mediated by the homophilic association of a glycoprotein called gp80, contact site A (csA), or sometimes antigen 117. It contains both N- and 0-linked sugar chains and a glycophospholipid anchor. This molecule is turned over and replaced by another adhesion system that appears beyond 12 h of development during slug migration. This is mediated by a glycoprotein called gp 150 which can be genetically and immunologically distinguished from the others. Other glycoproteins have also been implicated in mediating adhesion at this stage of development. 5.4.1. Generalfeatures in aggregation Tunicamycin inhibits chemotactic aggregation when added during the first few hours of development [123-1251. Since these effects can be overcome by adding cAMP [123] it

108

suggests that the loss of N-linked chains is not from a cell adhesion molecule itself, but they may be needed for the CAMP chemotactic signaling system. Other unidentified carbohydrates may be important in aggregation. Bozzaro used monosaccharides linked to polyacrylamide gels to examine cell binding, movement and aggregation [ 126,1271. Gels derivatized with glucose, maltose or cellobiose firmly bind the cells, but those with GlcNAc or Man bind cells to a lesser extent. Binding to GlcNAcand Man-containing gels could not be competed by these sugars, but mild trypsin digestion of the cells decreased binding; the opposite was found for binding to Glc gels. As cells proceed beyond aggregation binding to GlcNAc is lost. The interpretation of these results is that there are three receptors, one for each sugar. Further studies showed that when Glc, maltose or cellobiose, but not Man or GlcNAc is linked to the gels, development is arrested at the aggregation stage in a curious way. The cells initially form streams and aggregates, but just prior to the transformation into tight aggregates, they rapidly disperse into single cells, only to reform into aggregates once again. This futile cycle could continue for up to 30 rounds over a 24 h period [127]. Already formed tight aggregates also disperse when transferred to Glc-derivatized gels, but tipped (sheath enclosed) or slug-stage aggregates do not disperse. The mechanism underlying this unusual behavior is not known, but it clearly suggests that the cell surface must interact with a form of Glc and that this is able to affect further development at a specific point. One suggested explanation for the results, is that glucose is being recognized as a signal that coaxes the aggregating cells again into a feeding mode. Since the Glc is metabolically unavailable, they soon return to the starvation-induced developmental program. Two types of mutant cell lines that do not bind to the sugar-derivatized acrylamide gels have been isolated [128]. One type fails to bind to both Glc- and GlcNAc-derivatized gels, and the other fails to bind only to GlcNAc-derivatized gels. Three out of four of those in the first group aggregated only to the mound stage and the other formed aberrant fruiting bodies. Those in the second group mostly (10/13) formed minute fruiting bodies, tipless mounds, or did not aggregate at all. In addition, those that lack binding to both Glc and GlcNAc or severe deficiencies to GlcNAc alone showed reduced EDTAsensitive adhesion and a lower rate of bacterial phagocytosis. Although the structures of these glycans are not known, the results suggest that glycans with these sugars are physiologically important. One report suggests that secreted polysaccharides may play roles in cell adhesion [ 1291, and a variety of animal and plant polyanionic polysaccharides can partially inhibit aggregation at mg/ml concentration [ 1301. Since none of these molecules has been identified in Dictyostelium, the effects are difficult to interpret in molecular terms.

5.4.2. The EDTA-sensitive adhesion molecules A group of proteins collectively called gp24 is responsible for a portion of the EDTAsensitive adhesion (contact site B) and allows further development [ 131-1341. gp24 is not found in vegetative cells, but appears on the cell surface at the onset of aggregation and remains throughout development [ 1321. Cell adhesion is accomplished by a calcium-dependent homophilic interaction of gp24 molecules [ 1341. Four related and adjacent genes code for the different proteins whose molecular weights are approximately 12kDa [134] [W.F. Loomis, personal communication]. Thus, the name gp24 is

I09

somewhat inaccurate. The deduced amino acid sequences show no signal sequence or N-linked glycosylation sites. It is not anchored by glycophospholipid, so the mechanism by which the protein associates with the cell surface is not clear. Much of the protein is found in the cytoplasm and is probably secreted by a mechanism distinct from the endoplasmic reticulum-Golgi route. No mutant strains have been found that lack gp24. This is not surprising since all four genes are active, and the chances of inactivating all of them would be negligible. Production of antisense mRNA for gp24 under the control of a promoter that is active in vegetative cells until about 6 h of development delays cell adhesion and the appearance of gp24 for two hours [ 1351. Thereafter, transcription of the antisense mRNA ceases and the accumulation of gp24 and normal adhesion resume. Deletion of all the genes will be required to finally determine the functional significance of gp24, and other components in the earliest adhesion system. Another glycoprotein in vegetative cells, gp 130, has been implicated in EDTA-sensitive adhesion. It has the same sequence as gp138B, appears to be N- and 0-glycosylated and glycophospholipid anchored in the membrane [7 I]. Its carbohydrates closely resemble those of csA described below, but have been much less thoroughly studied than csA. Besides the gp24 and gp1301138 group, another protein may also be involved in the EDTA-sensitive adhesion system. Based on the historical nomenclature, this molecule is called contact site C. It differs from contact site B, since it has slightly different developmental regulation compared to gp24 [ 1361, because the inhibition is not seen using EGTA, and EDTA inhibition is Mg2+-reversible.The protein has not yet been identified.

5.4.3. The EDTA-resistant adhesion molecule Contact site A (csA) refers to EDTA-insensitive adhesion which is mediated by a glycoprotein called gp80 [137]. It appears at about 6 h of development, persists until 12 h, disappears, and is found once again on the surface of a subset of cells late in development [138]. gp80 is concentrated in filopodia in regions of cell-cell contact [139]. Increased level of gp80 expression appears to regulate the slug size, since individual cells are more adhesive [140]. The protein has 495 amino acids after removal of the 19 amino acid leader sequence and contains N-linked and 0-linked chains together with a glycophospholipid anchor. The N-linked chain is sulfated and fucosylated while the 0-linked carbohydrate has the modB antigenic determinant, and binds to WGA [71,92,93,96,141,142]. These chains are required for the transport and expression at the cell surface. When they are absent, the cell surface molecules are clipped by a surface protease with the release of large fragments into the medium [71,92]. The 0-linked chains are located near the glycophospholipid anchor and mutation of the protease results in normal accumulation of the protein at the surface and leads to normal cell adhesion [71]. The peptide, not the carbohydrate portion, of gp80 mediates adhesion. An eight amino acid peptide inhibits both cell-gp80 and gp8Sgp80 interaction, and antibodies against the peptide inhibit reassociation of previously dissociated cells [ 143,1441. Site directed mutagenesis to remove all of the N-linked chains does not alter adhesion [71]. Studies by Henderson and associates present another view of adhesion. Their results show that stage-specific EndoH-resistant membrane-derived pronase glycopeptides can partially block EDTA-resistant adhesion [ 1451. The glycopeptides could be derived from

110

another protein involved in some aspect of the EDTA-resistant adhesion. In another study by the same group, two temperature-sensitive mutant strains have reduced adhesion at the restrictive temperature of 27"C, but a revertant strain aggregates normally [ 1461. Cell membrane glycopeptides prepared from wild-type cells -and mutants grown at the permissive temperature partially block adhesion, but those isolated from mutant cells grown at the restrictive temperature do not [ 1471. The structure of the inhibitory glycopeptides has not been established. gp80 is glycophospholipid anchored [96,148,149]. PI-PLC digestion does not release the lipid from this protein, and strong base hydrolysis is needed to release free palmitic acid, suggesting that the lipid is in an amide linkage [149]. Ceramide glycans with mannose and glucosamine were described in Dictyostelium discoideum over 20 years ago [150], before such anchors were discovered. Structures similar to those of gp80 have also been reported in yeast, mycobacteria and plants [ 1491. Substituting the anchor domain of gp80 with a transmembrane domain does not alter its ability to mediate adhesion; however, the normally long-lived gp80 turns over much more rapidly when it lacks an anchor. Normally, the gp80 anchor probably excludes the protein from clathrinmediated pinocytosis, and when the anchor is removed it mingles with the other cell surface proteins and is degraded at a much higher rate [75]. Several studies by Klein and co-workers [138,151-1531 also showed that gp80 (called antigen 1 17 in their reports) is insensitive to PI-PLC digestion, but about 25% of non-lipid bearing protein containing ethanolamine can be released from the cell surface by an endogenous membrane-associated non-proteolytic activity [ 1531. When the protein is released, it exposes epitopes that are recognized by antibodies against the GPI anchor of the variant surface glycoprotein of T brucei. These antibodies will recognize the cryptic determinant only after it is cleaved from the lipid portion by PI-PLC, and suggest that the activity which releases antigen 117 from the surface may also be a phospholipase-type enzyme. These data argue against this release being merely a proteolytic clip as seen in the strains that carry the modB mutation. The fact that only a portion of the antigen 117 is released from the membranes could mean that there are distinct classes of the protein, some being cleavable and others not. The difference may be a matter of availability of the enzymes or may involve differently anchored forms of the protein. Deletion of the gp80 gene [71,199] leads to loss of EDTA-resistant adhesion, but does not abolish development. The other two (or more) adhesion systems can apparently mediate adhesion well enough under laboratory conditions in the absence of shear forces to produce mature fruiting bodies. 5.4.4. Post-aggregation adhesion system

After gp80 has disappeared, a third adhesion system is found at the slug-stage and beyond. Disaggregated slug-stage cells still adhere to each other in the presence of EDTA even when gp80 is not expressed and gp24 is blocked by antibody [ 1321. Antisera prepared against cells developed to this stage do not block the adhesion of cells prepared at earlier times in development, but it will block the third adhesion system [154]. A glycoprotein called gp150, appears to mediate the late adhesion system because antibodies against it can block cell reassociation of post-aggregation cells [ 155,1561. It has also been proposed that this molecule may be important in the differential adhesion of pre-stalk and pre-spore

111

cells to themselves, since anti-gp 150 can inhibit preferential self-association when a disaggregated mixture of cell types is allowed to reassociate. Purified gp 150 neutralized the effects of a rabbit antiserum raised against gel-purified gp 150; '*'I-labeled gp 150 shows saturable binding to intact post-aggregation cells suggesting specific cell surface binding sites [ 1551. Soluble gpl50 blocks the reassociation of dissociated cell aggregates, but it shows no effect on cells at the early aggregation stage. Although gp150 is an important post-aggregation adhesion molecule, other results show that a monoclonal antibody which blocks adhesion recognizes carbohydrates on other glycoproteins of 95, 90, 35 and 30kDa (PsA) [157]. This suggests that the carbohydrate may be involved in this adhesion. In summary, the results of two decades of cell adhesion studies still suggest that unidentified carbohydrates play a part in cell adhesion. The extensive studies on csA prove that its partially characterized glycans are not among these players. The elusive connection cannot be resolved until more decisive structural analysis is done on Dictyostelium glycans. 5.4.5. The surface sheath The surface sheath is required for slug migration and complete morphogenesis. Differential gene expression of stalk and spore cell components can occur in disaggregated cells under the proper environmental conditions, but it is the sheath that allows Dictyostelium discoideum to function as a coordinated organism. It must possess considerable resilience and strength to accommodate all of the complex movements a slug makes over various terrains. Analysis of the surface sheath has been challenging for several reasons. Firstly, it is extremely thin and does not provide much material for biochemical analysis. Secondly, although the sheath can be harvested as trails left behind the migrating slugs, the preparation can be contaminated by sluffed cells or the debris of lysed cells. These points have made it difficult to confidently identify bona fide sheath components. The sheath is stationary relative to the substratum and the cells move over and through it leaving the sheath behind as a collapsed tube. Thus, each region of the slug may make its individual contribution to the sheath as the cells pass by, and this probably explains why the sheath appears to be thicker and have additional components at the posterior of the slug compared to the tip.

5.4.5.1. Cellulose. The surface sheath surrounds the developing aggregate just prior to tip formation and continues to envelope it until culmination. A major structural component of the sheath is cellulose [ 1581, and like the sheath itself, it accumulates only at the surface and not in the area between the cells. Cellulose is much more dense on the ventral side of the slug and it may be deposited on the substratum when the tip of the migrating slug touches down [ 1591. The cellulose appears to mature during development as shown by the increase in crystallinity from sheath isolated from early aggregates compared to the trails left behind migrating slugs [ 1601. This change is probably physiologically important since high crystallinity is correlated with increased rigidity and tensile strength of the cellulose fibrils. How this maturation occurs is not known, but it would appear to be regulated, since cellulose is less crystalline in several mutant strains that lack the single lysosomal enzyme, b-N-acetylhexosaminidase [ 1611. Initially aggregates of the mutant strains form

112

normal-sized slugs, but they are unstable and quickly break up into smaller, stable slugs with lower surface to volume ratios. The low crystalline cellulose apparently does not provide enough strength to withstand the forces generated by a larger migrating slug. Since b-hexosaminidase can be secreted during development, it may play some role in controlling the crystallinity of the cellulose. Particulate membrane fractions prepared from the tight aggregate stage (10 h) onward are active in cellulose synthesis [162]. The product made from UDP[’4C]glucose was shown to be cellulose by its solubility properties, periodate oxidation and methylation analysis pattern. Cellulose can also be made in monolayer cultures by cells induced with CAMP and DIF, a potent inducer of stalk cell properties specific genes [24,163]. 5.4.5.2. Glycoantigens and glycoproteins. Immunoelectron microscopy shows that several monoclonal antibodies against carbohydrates react with components of the surface sheath. These include antibodies 8 1.1, 83.5, 40.1 , MUDSO, MUD62 [ 107-1 09,164-1 661. Immunofluorescence studies show that there is a regular pattern of hexagonal “cell-prints’’ outlining the borders of the ventral cells in contact with the substratum [ 102,159,1671, Coincident patterns can be seen using antibodies MUD50 and Calcofluor, a fluorescent tag for cellulose [ 1591. Cell prints are not seen using antibodies MUD1, MUD3, MUD62 or those against SP29 (PsA), SP70, or SP96, so the pattern is not due to cell lysis. Slugs from mutant cells lacking MUD50 epitope appear to “slip” in their forward migration, suggesting that the components bearing the MUD50 determinant are involved in traction and slug locomotion [ 105,159,1661. Two highly acidic pre-stalk-specific proteins ST3 10 and ST430 [ 1681 appear to be totally extracellular. ST430 occurs in the sheath and ST310 is found in the stalk tube and between the stalk cells [ 1691. Since they are extracellular matrix proteins, their genes are called ecmA (ST430) and ecmB (ST310)[170,171]. The ecmA gene has a series of 23 cysteine-rich repeats with a total of 4-6 N-glycosylation sites [ 1701. ecmB has at least 41 repeats of a similar cysteine-rich 24 amino acid segment with a total of 26 N-glycosylation sites! [ 1711 High cysteine content in re-iterated segments are characteristic of other extracellular proteins in lower eukaryotes [ 172,1731. Homologous disruption of the ecmB gene does not alter development, but disruption of the ecmA gene leads to formation of abnormally long and thin standing slugs [174]. Apparently, sheath that lacks ecmA is weak and slugs lose a large portion of the cells as they migrate. If development is arranged to omit the migrating slug-stage, loss of ecmA gives normal fruiting bodies. 5.4.6. Pre-spore vesicles and the spore coat

Ultimately, Dictyostelium discoideum undergoes development to ensure that a portion of the starving population survives. To accomplish this, the stalk cells provide a tower from which the spores can be dispersed over a wide area, but the long-term survival of the spores, and the species, is dependent upon their success in forming an environmentally resistant casing called the spore coat. Preparations begin early in the slug-stage when most of the coat components are beginning to accumulate in PSVs. At the appropriate time, the contents are jettisoned from the cell to generate a pool of precursors for the assembly of the spore coat. The regulation and synthesis of several spore coat glycoproteins and

113

cellulose have received most of the attention. Several of the proteins have been cloned and their regulation investigated [3]. 5.4.6.1. Pre-spore vesicles. Pre-spore vesicles (PSVs) contain all of the known major

components that will later form the spore coat, except for cellulose (Table 2) [106]. PSVs are found only in pre-spore cells [175] and appear to be derived from the Golgi, which is quite prominent in those cells; however, a typical Golgi is essentially undetectable in vegetative and pre-stalk cells using similar techniques [176]. It can be detected in vegetative cells by fluorescence microscopy using tagged WGA lectin [177] and a light membrane fraction has been biochemically characterized as the site of several glycoprotein processing events [39,4 1,5 1,521. PSVs have a characteristic density and can be separated from other organelles including lysosomes, and yet surprisingly, these PSVs also contain several typical lysosomal enzymes [ 178-1 801 (Table 2). It has been suggested that a portion of the lysosomes may transform into PSVs or, alternatively, that lysosomal enzymes are also routed to the PSV because they function within it or later in the assembly of the spore coat. The question is not yet resolved since recent work suggests that the PSVs exist as a tubular network which is continuous with the Golgi [106]. The spore coat proteins SP60, SP70 and SP96 have been cloned and their expression is coordinately regulated [ 181-1 831. They all have typical signal peptides and SP70 and SP96 have potential N-glycosylation sites. SP75 is also modified by Table 2 Components of pre-spore vesicles and spore coatsa Component

PSV

Location in spore coat Inner Middle Outer

mAb

Reference(s)

83.5/MUD62, MUD3

[109,191]

16.1/MUD102, MUDS0

[ 1091

7

83.5/MUD62

[lo91

83.5/MUD62

Soluble matrix

Proteins SP96/p 1 12

+

SP8S/PsB

+

+

SP80 SP75

-

-

SP70

?

?

+ +

?

[ 194,1951

SP60

?

?

+

7

[ 195,1961

Enzymes a-Mannosidase

+

+

Phosphatase

+

+

Polysaccharides Cellulose GaliGalNAc a

Compiled from West and Erdos [ 1061

[I091

114

Fig. 7. Probes for pre-spore vesicle components show a non-random distribution. Pre-spore vesicles have the galactose-containing polysaccharide that binds to lectin RCA-1 (large particles) and the spore coat protein SPWPsB binds to a rnodB-dependent carbohydrate (small particles). The probes clearly show their accumulation in separate regions of the vesicle. x 71 500. Another antibody that recognizes fucose (mAb 83.5) also displays particles in distinct regions of the PSV (From ref. [log], with permission.)

N-linked oligosaccharides since its molecular weight is altered in a strain carrying the modA mutation [ 1061. SP96, SP70 and SP75 are also phosphorylated [ 1841 on serine residues, [ 1851 and fucose containing oligosaccharides are released from SP96 and SP75 by ($elimination [106]. The released glycan runs as a hexasaccharide by gel filtration and probably contains the epitope recognized by antibodies 83.5 and MUD62. SP80 and SP85/PsB contain the carbohydrate epitopes in the modB series [107]. There is no direct evidence that SP60 is glycosylated and no N-linked sites occur, but the calculated molecular weight of the unmodified protein without the signal peptide is only about 47 kDa suggesting post-translational modifications may exist [ 1821. How the components are targeted to or retained by the PSV is not known, but a preformed complex of six proteins is found within the PSV [ 1861. The components are associated by both disulfide and non-covalent forces and include SP85/PsB and SP96/pll2. The composition of the complex changes during secretion and spore coat maturation [ 1871. The PSVs also contain polysaccharide composed of galactose, galactosamine and galacturonic acid that can be recognized by pre-spore-specific antisera [ 1881. Lectins RCA-1 and SBA that preferentially recognize Gal and GalNAc residues also bind to this polysaccharide [106]. The distribution of the various components in the PSV is not homogeneous, instead they tend to be clustered in separate regions of the vesicle [109]. This is shown in Fig. 7. The associations are non-random, but the spore coat is not a preassembled complex since the association of the components within the PSV is different than that in the mature spore coat [106,109].

I15

Since the PSV may be a tubular network rather than a series of distinct vesicles, secretion of the components may not require extensive vesicle fusion with the plasma membrane. Secretion parallels loss of UDP-Gal polysaccharide galactosyl transferase that is localized in the Golgi [ 189,1901 suggesting that multiple secretory pathways may be coordinated to form the spore coat.

5.4.6.2. Spore coat. Table 2 [lo91 lists the known spore coat components from either post-germination spore coats or by immunolocalization of spore preparations prior to germination. The various glycoproteins are recognized by several of the carbohydrateor peptide-specific antibodies or by lectins RCA-I or SBA to identify the galactosecontaining polysaccharide. Cellulose can be detected by the binding of colloidal goldderivatized cellulolytic enzymes. The spore coat has three distinct layers and the spores in the sori are bathed in a soluble “matrix”. The outer layer contains SP75, SP80, and SP96, as detected by antibody 83.5 and shows that they are neatly arranged along either side of the electron dense zone. The inner layer is close to the plasma membrane and contains SP85(PsB), a-mannosidase and the galactose polysaccharide. Cellulose is found only in the middle layer. The soluble matrix is lost during preparation of spore coats, but prior cetylpyridinium chloride treatment preserves it. Analysis of various washes of the sori or of secretions produced by submerged developing aggregates also suggests that the lysosomal enzymes and another group of proteins recognized by antibodies 8 1.8 and 85.2. are also found in the matrix. Although single spores can encapsulate individually, in the natural environment of the maturing sorus the assembly is not cell autonomous. The proteins appear to be part of a pool of precursors for spore coat assembly. In favor of this view are the results of West and Erdos [lo51 who show that mixed development of wild-type and mutant HL250 cells that do not produce antigen 83.5 results in the equal distribution of this determinant among all of the spores. The assembly of the coat seems to begin in the outer layer, but the mechanism is unknown. It is likely that the associations of some of the proteins in the outer layer are autonomous, and there is no evidence that membrane receptors guide the assembly of the coat.

6. Molecular glycobiology in Dictyostelium discoideum Since deleting entire classes of glycans is not lethal to single cells, most of the specific functions of sugar chains are probably occurring at the multicellular level [191]. Previously, our understanding of glycan biosynthesis and function in multicellular organisms relied on the chance discovery of rare mutations showing clear and dramatic phenotypes. The arrival of molecular approaches has provided much more systematic and rational methods. The most popular of these is making gene knockouts in transgenic mice, but the results of total deletions can be quite unpredictable. Sometimes the phenotypes are so subtle that they are undetectable during embryonic development or adult life, while deletion of others arrests early embryogenesis [191]. Useful information is derived between these extremes. It is difficult to predict what the outcome of any one gene deletion will be, especially when deleting glycosyl transferases that service many organ systems,

116

tissues and cell types. Even knowing the specific glycosyl transferase defect does not lead to accurate prediction of the phenotype or the severity of the defects [ 191,1921. Recent advances in targeted gene disruption and insertional mutagenesis in Dictyostelium offer an opportunity to study molecular aspects of glycobiology at the multicellular organismic level [7,9]. Restriction enzyme mediated integration (REMI) is a method to make random gene disruptions in Dictyostelium [8,9,193]. A linearized plasmid with a selectable marker is introduced into a recipient strain and random insertion of the marker disrupts an endogenous gene. Re-isolation of the marker together with flanking regions of the disrupted gene can be used for its isolation. To date most of the transformants have been screened for developmental abnormalities with subsequent characterization of the disrupted gene [7,9]. However, this same technique can be adapted to searching for disrupted glycosyl transferase genes by screening with monoclonal antibodies against the glycans. Identification of the mutated gene could then be correlated with an altered structure of the glycan and the loss of enzyme activity. Once isolated, these genes could be used to complement (correct) the mutant strains or to produce antisense oligonucleotides driven by some of the previously characterized temporal- or cell-type-specific promoters [3,4,7]. Using these promoters, one would be able to fine tune the abrogation of transferases rather than insisting only on a total knockout in all cells. The effects of each complete or selective gene knockout could then be monitored for its consequences during growth, aggregation, morphological development, slug migration, tissue proportioning, spore coat structure, and spore viability. Figure 8 shows a scheme using these methods to study glycan function in Dictyostelium.

I

mutant

I \

I

I overexDression I /I

mutant

Or

I

T I ' Clone gene \

Cell- or temporal-

spccific

Fig. 8. Systematic glycobiology in Dictyostelium. This figure shows how various features and technologies in Dictyostelium can be exploited to investigate glycobiology. Antibodies ( I ) against glycans can be used to screen for biosynthetic mutants created by insertional mutagenesis (REMI) (2). The disrupted genes can be identified using the REMI technology (3) and then be ablated in selected cell types or at different times in development by using antisense mRNA controlled by cell-type- or temporal-specific promoters (4). The glycans can be purified from gram quantities of vegetative or developing cells and used as affinity ligands to isolate putative receptors ( 5 ) or in some cases they could be added exogenously to disrupt normal development (6). In some cases the molecular deletions should alter development in ways that are already predicted by analysis of chemically mutagenized cells identified using these antibodies.

117

In this system, we already know that alterations in sugar chains affect cell growth, adhesion, slug migration, and a variety of developmental aspects. Systematic analysis of these glycoconjugates will ultimately yield useful information and strong encouragement to clone Dictyostelium glycosyl transferase genes.

Acknowledgments Supported by US Public Health Service grants GM32485 and GM94096. The author is indebted to members of his laboratory and to Dr. Chris West and Dr. Keith Williams for sharing their unpublished data and to Dr. Glaucia Souza and Marion Lammertz for critical comments on the manuscript.

References Literature review was completed December 1994. [ I ] Sussman, M. (1987) In: J.A. Spudich (Ed.), Methods in Cell Biology. Academic Press, New York, pp. 9-28. [2] Gross, J.D. (1994) Microbiol. Rev. 58, 330-351. [3] Loomis, W.F. (1993) Curr. Top. Dev. Biol. 28, 1 4 6 . [4] Williams, J. and Morrison, A. (1994) Prog. Nucl. Acid Res. Mol. Biol. 47, 1-27. [5] Devreotes, P.N. (1994) Neuron 12, 235-241. [6] Condeelis, J., Jones, J. and Segall, J.E. (1992) Cancer Metastasis Rev. 11, 55-68. [7] Kuspa, A. and Loomis, W.F. (1994) Methods Mot. Genet. 3, 3-21. [8] Kuspa, A. and Loomis, W.F. (1992) Proc. Natl. Acad. Sci. USA 89, 8803-8807. [9] Dynes, J.L., Clark, A.M., Shaulsky, G., Kuspa, A., Loomis, W.F. and Firtel, R.A. (1994) Genes Dev. 8, 948-958. [lo] Hereld, D. and Devreotes, P.N. (1992) Int. Rev. Cytol. 137B, 3 5 4 7 . [ I I] Freeze, H.H. (1992) Cell Surface Carbohydrates and Cell Development. CRC Press, Boca Raton, FL, pp. 285-317. [I21 Van Haastert, P.J., Janssens, P.M. and Erneux, C. (1991) Eur. J. Biochem. 195, 289-303. [I31 West, C.M. and Erdos, G.W. (1990) Dev. Genet. 1 1 , 492499. [14] Freeze, H.H., Willies, L., Hamilton, S. and Koza-Taylor, P. (1989) J. Biol. Chem. 264, 5653. [I51 Sharkey, D.J. and Kornfeld, R. (1991) J. Biol. Chem. 266, 18477-18484. [I61 Sharkey, D.J. and Kornfeld, R. (1991) J. Biol. Chem. 266, 18485-18497. [I71 Couso, R., Van Halbeek, H., Reinhold, V. and Kornfeld, S. (1987) J. Biol. Chem. 262, 45214527. [I81 Freeze, H.H. (1986) Mol. Cell Biochem. 72, 47-65. [19] Sharkey, D.J. and Kornfeld, R. (1989) J. Biol. Chem. 264, 10411. [20] Ivatt, R.L., Das, O.P., Henderson, E.J. and Robbins, P.W. (1984) Cell 38, 561-567. [21] Tretter, T., Altmann, F. and Man, L. (1991) J. Biochem. 199, 647452. [22] Kalada, P.M., Goldberg, R. and Poretz, R.D. (1983) Mol. Immunol. 20, 727-735. [23] Amatayakul-Chantler, S., Ferguson, M.A.J., Dwek, R.A., Rademacher, T.W., Parekh, R.B., Grandall, I.E. and Newell, P.C. (1991) J. Cell Sci. 99, 485495. [24] Crandall, I.E. and Newell, P.C. (1989) Development 107, 87-94. [25] Yoshida, M., Fuse, G., Matsui, T. and Ouchi, S. (1992) Biochem. Biophys. Res. Commun. 188, 794798. [26] Yoshida, M., Matsui, T., Fuse, G. and Ouchi, S. (1993) FEBS Lett. 318, 305-309. [27] Kornfeld, S. (1986) J. Clin. Invest. 77, 14. [28] Kornfeld, S. and Mellman, 1. (1989) Annu. Rev. Cell Biol. 5, 483-525. [29] Freeze, H.H., Yeh, R., Miller, A.L. and Kornfeld, S. (1983) J. Biol. Chem. 258, 14874-14879.

118 [30] Freeze, H.H. (1985) Arch. Biochem. Biphys. 243, 690-693. [31] Gabel, C.A., Costello, C.E., Reinhold, VN., Kurz, L. and Kornfeld, S. (1984) J. Biol. Chem. 259, 13762-13769. [32] Cardelli, J.A., Bush, J.M., Ebert, D. and Freeze, H.H. (1990) J. Biol. Chem. 265, 884778853, [33] Freeze, H.H. and Wolgast, D. (1986) J. Biol. Chem. 261, 127-134. [34] Kornfeld, S. (1992) Annu. Rev. Biochem. 61, 307-330. [35] Freeze, H., Miller, A.L. and Kaplan, A. (1980) J. Biol. Chem. 255, 11081-1 1084. [36] Freeze, H.H. (1985) J. Biol. Chem. 260, 8857-8864. [37] Lang, L., Couso, R. and Kornfeld, S. (1986) J. Biol. Chem. 261, 63204325. [38] Couso, R., Lang, L., Roberts, R.M. and Kornfeld, S. (1986) J. Biol. Chem. 261, 632643231. [39] Freeze, H.H. and Wolgast, D. (1986) J. Biol. Chem. 261, 135-141. [40] Freeze, H.H., Hindsgaul, 0. and Ichikawa, M. (1992) J. Biol. Chem. 267, 44314439. [41] Cardelli, J.A. and Dimond, R.L. (1988) Transport and Targeting of Lysosomal Enzymes in Dictyostelium discoideum in Protein Transfer and Organelle Biogenesis. Academic Press, New York, p. 363. [42] Cardelli, J.A. (1993) Adv. Cell Mol. Biol. Membranes I , 341-390. [43] Cardelli, J.A., Richardson, J.R. and Miears, D. (1989) J. Biol. Chem. 264, 3454-3463. [44] Richardson, J.M, Woychik, N.A., Ebert, D.L., Dimond, R.L. and Cardelli, J.A. (1988) J. Cell. Biol. 107, 2097-2107. [45] Graham, T.R., Zassenhaus, H.P. and Kaplan, A. (1988) J. Biol. Chem. 263, 16823-16829. [46] Lacoste, C.H., Graham, T. and Kaplan, A. (1992) J. Biol. Chem. 267, 5942-5948. [47] Judelson, H.S. and Dimond, R.L. (1988) Differentiation 37, 7-13. [48] Judelson, H.S. and Dimond, R.L. (1988) Arch. Biochem. Biophys. 267, 151-157. [49] Cardelli, J.A., Bush, J.M., Ebert, D. and Freeze, H.H. (1990) J. Biol. Chem. 265, 8847-8853. [50] Freeze, H.H., Bush, J.M. and Cardelli, J. (1990) Dev. Genet. 11, 463472. [51] Lacoste, C.H., Freeze, H.H., Jones, J.A. and Kaplan, A. (1989) Arch. Biochem. Biophys. 273,505-515. [52] Hohmann, H.P., Gerisch, G., Lee, R.W. and Huttner, W.B. (1985) J. Biol. Chem. 260, 13869-13678. [53] Davis, S.J. (1988) J. Cell Biochem. 38, 77-86. [54] Davis, S.J. and Wheldrake, J.F. (1985) FEMS Microbiol. Lett. 30, 353-358. [55] Schmidt, S. and Wheldrake, J.F. (1993) Mol. Cell. Biochem. 126, 109-114. [56] Mehta, D.P., Etchison, J.R. and Freeze, H.H. (1995) Arch. Biochem. Biophys. 321, 191-198. [57] Gustafson, G.L. and Thon, L.A. (1979) J. Biol. Chem. 254, 12471-12478. [58] Korth, M.J., Finn, D.J. and Gustafson, G.L. (1988) Anal. Biochem. 169, 181-184. [59] Gustafson, G.L., Finn, D.J. and Moin, K. (1988) Anal. Biochem. 169, 185-188. [60] Gustafson, G.L. and Milner, L.A. (1980) J. Biol. Chem. 255, 7208-7210. [61] Gustafson, G.L. and Gander, J.E. (1984) Methods Enzym. 107, 172-183. [62] Varki, A. and Kornfeld, S. (1981) J. Biol. Chem. 256, 9937-9943. [63] Souza, G.M., Hirai, J. and Freeze, H.H. (1995) J. Biol. Chem. 270, 28983-28945. I641 North, M.J. (1992) Hoppe Seyler Biol. Chem. 373, 401406. [65] Freeze, H.H. and Ichikawa, M. (1995) Biochem. Biophys. Res. Commun. 208, 384-389. [66] Riley, G.R., West, C.M. and Henderson, E.J. (1993) Glycobiology 3, 165-177. [67] Finn, D.J. and Gustafson, G.L. (1987) Biochem. Biophys. Res. Commun. 148, 834-837. [68] Gooley, A.A., Classon, B.J., Marschalek, R. and Williams, K.L. (1991) Biochem. Biophys. Res. Commun. 178, 11941201. [69] Gooley, A. and Williams, K.L. (1994), Glycobiology 4, 413417. [70] Gooley, A.A., Marshchalek, R. and Williams, K.L. (1992) Genetics 130, 749-756. [71] Gerisch, G., Faix, J., Wallraff, E., Noegel, A.A., Barth, A,, Lutzelschwab, R., Westphal, M., Langanger, G. and Francis, D. (1993). Mutational Analysis of Carbohydrate and Phospholipid Modifications of a Cell Adhesion Protein, Vol. 44. Springer, Berlin, pp. 131-144. [72] Haynes, P.A., Gooley, A.A., Ferguson, M.A.J., Redmond, J.W. and Williams, K.L. (1993) Eur. J. Biochem. 216, 729-737. [73] Hitt, A.L., Lu, T.H. and Luna, E.J. (1994) J. Cell Biol. 126, 1421-1431. [74] McConville, M.J. and Ferguson, M.A.J. (1993) Biochem. J. 294, 305-324. [75] Barth, A,, Muller-Taubenberger, A,, Taranto, P. and Gerisch, G. (1994) J. Cell Biol. 124, 20-215.

1 I9

[76] Hart, G.W. (1992)Curr. Opin. Cell Biol. 4,1017-1023. [77]Gonzalez-Yanes, B., Cicero, J.M., Brown Jr., R.D. and West, C.M. (1992)J. Biol. Chem. 267,95959605. [78]Kozarov, E., van der Wel, H., Field, M., Gritzali, M., Brown Jr., R.D. and West, C.M. (1995)J. Biol. Chem. 270,3022-3030. [79] Knecht, D.A., Dimond, R.L., Wheeler, S. and Loomis, W.F. (1984)J. Biol. Chem. 259,10633-10640. [80]Knecht, D.A. and Dirnond, R.L. (1981)J. Biol. Chem. 256,35643575. Mierendorf, R.C., Wunderlich, R. and Dirnond, R.L. (1984)J. Biol. Chern. 259,10641[81]Freeze, H.H., 10643. [82]Freeze, H.H., Koza-Taylor, P., Sanders, A. and Cardelli, J.A. (1989)J. Biol. Chern. 264,19278-19286. [83]Judelson, H.S., Freeze, H.H. and Dimond, R.L. (1987)Arch. Biochem. Biophys. 253,305-314. [84]McManus, M.T., McKeating, J., Secher, D.S., Osborne, D.J., Ashford, D., Dwek, R.A. and Rademacher, T.W. (1988)Planta 175,506. [85] Faye, L., Gomord, V., Fitchette-Lain, A.C. and Chrispeels, M.J. (1993)Anal. Biochem. 209,104-108. [86] Free, S., Schimke, R.T., Freeze, H. and Loomis, W.F. (1978)J. Biol. Chem. 253,41024106. [87] Free, S.J. and Schimke, R.T. (1978)J. Biol. Chem. 253,41074111. [88] Freeze, H.H., Yeh, R., Miller, A.L. and Kornfeld, S. (1983)J. Biol. Chem. 258, 14874-14879. [89] Freeze, H. and Miller, A.L. (1981)Mol. Cell Biochem. 35, 17-27. [90] Ebert, D.L., Bush, J.M., Dimond, R.L. and Cardelli, J.A. (1989)Arch. Biochem. Biophys. 273,479490. [91]Helenius, A. (1994)Mol. Biol. Cell 5 , 253-265. [92] Hohmann, H.P., Bozzaro, S., Merkl, R., Wallraff, E., Yoshida, M., Weinhart, U. and Gerisch, G. (1987) EMBO J. 6,3663-3671. [93] Hohmann, H.P., Bozzaro, S., Yoshida, M., Merkl, R. and Gerisch, G. (1987)J. Biol. Chem. 262,1661816624. [94] Gerisch, G., Faix, J., Wallraff, E., Wicke, U., Adrian, I. and Noegel, A.A. (1990)In: Developmental and Molecular Biology of Dictyosteliurn discoideurn. EMBOMSF Workshop, p. 67. [95] Siu, C.-H. and Lam, T.Y. (1988)Exp. Cell Res. 177,338-346. [96] Stadler, J., Keenan, T.W., Bauer, G. and Gerisch, G. (1989)EMBO J. 8,371-377. [97] Bozzaro, S. and Merkl, R. (1985)Cell Differ. 17,83-94. [98] Bertholdt, G., Stadler, J., Bozzaro, S., Fichtner, B. and Gerisch, G. (1985)Cell Differ. 16,187-202. [99]Murray, B.A., Wheeler, S., Jongens, T. and Loomis, W.F., (1984)Mol. Cell Biol. 4,514-519. [IOO] West, C.M. and Loomis, W.F. (1985)J. Biol. Chem. 260,13803-13809. [loll Loomis, W.F., Wheeler, S.A., Springer, W.R. and Barondes, S.H. (1985)Dev. Biol. 109,Ill-117. [I021 Breen, E.J., Vardy, P.H. and Williams, K.L. (1987)Development 101, 313-322. [I031 Aparicio, J.G., Erdos, G.W. and West, C.M. (1990)J. Cell Biochem. 42,255-266. [I041 Houle, J., Balthazar, J. and West, C.M. (1989)Proc. Natl. Acad. Sci. USA 86,3679-3683. [IOS] Alexander, S., Smith, E., Davis, L., Gooley, A,, Por, S.B., Browne, L. and Williams, K.L. (1988) Differentiation 38, 82-90. [I061 West, C.M. and Erdos, G.W. (1990)Dev. Genet. 1 1 , 492-506. [lo71 West, C.M., Erdos, G.W. and Davis, R. (1986)Mol. Cell. Biochem. 72,121-140. [I081 West, C.M. and Erdos, G.W. (1988)Cell Differ. 23, 1-16. [I091 Erdos, G.W. and West, C.M. (1989)Exp. Mycol. 13, 169-182. [ I I O ] Champion, A,, Griffiths, K., Gooley, A.A., Gonzalez, B.Y., Gritzali, M., West, C.M. and Williams, K.L. (1995)Microbiology 141,785-797. [I 1 I] Gonzalez-Yanes, B., Mandell, R.B., Girard, M., Henry, S., Aparicio, O., Gritzali, M., Brown, R.D., Erdos, G.W. and West, C.M. (1989)Dev. Biol. 133, 576-587. [ 1 121 Kreft, M., Voet, L., Gregg, J.H. and Williams, K.L. ( 1 985)J. Embryol. Exp. Morph. 88,15-24. [ I 131 Suzuki, K. and Yanagisawa, K . (1990)Cell. Differ. Dev. 30,3542. [I 141 Fang, H., Higa, M., Suzuki, K., Aiba, K., Urushihara, H. and Yanagisawa, K. (1993)Dev. Biol. 156,

201-208. [I151 Chia, C.P., Shariff, A,, Savage, S.A. and Luna, E.J. (1993)J. Cell Biol. 120,909-922. [I161 Chia, C.P., Hitt, A.L. and Luna, E.J. (1991)Cell. Motil. Cytoskeleton 18, 164-179. [I171 Hitt, A.L., Hartwig, J.H. and Luna, E.J. (1994)J. Cell Biol. 126,143344.

120 Hitt, A.L., Lu, T.H. and Luna, E.J. (1994) J. Cell Biol. 126, 1421-1431. Jain, R. and Gomer, R.H. (1994) J. Biol. Chem. 269, 9128-9136. Taphouse, C.R. and Gomer, R.H. (1992) Genes Dev. 6, 390400. Jain, R., Yuen, IS., Yuen, I S . , Taphouse, C., Halfant, K.A. and Gomer, R.H. (1991) Development 113, 1375-1385. Gomer, R.H., Yuen, I.S. and Firtel, R.A. (1991) Development 112, 269-278. Sadeghi, H. and Klein, C. (1988) Differentiation 38, 99-103. Lam, T.Y. and Siu, C.H. (1982) Dev. Biol. 92, 398407. Hirano, T., Yamada, H. and Miyazaki, T. (1983) J. Biochem. 93, 1249-1257. Bozzaro, S. and Roseman, S. (1983) J. Biol. Chem. 258, 13882-13889. Bozzaro, S. and Roseman, S. (1983) J. Biol. Chem. 258, 13890-13897. Ceccarelli, A. and Bozzaro, S. (1992) Antm. Biol. I , 59-68. Springer, W.R. and Barondes, S.H. (1982) Proc. Natl. Acad. Sci. USA 79, 6561-6565. Eitle, E., Keller, T., Parish, C.R. and Parish, R.W. (1993) Exp. Cell Res. 205, 374-382. Knecht, D.A., Fuller, D.L. and Loomis, W.F. (1987) Dev. Biol. 121, 277-283. Loomis, W.F. (1988) Dev. Genet. 9, 549-559. Loomis, W.F. and Fuller, D.L. (1990) Proc. Natl. Acad. Sci. USA 87, 886-890. Brar, S.K. and Siu, C.H. (1993) J. Biol. Chem. 268, 24902-24909. Loomis, F., Kuspa, A. and Fuller, D.L. (1990) In: Developmental and Molecular Biology ofDictyosfeliurn discoideum. EMBO/NSF Workshop, p. 66. [I361 Fontana, D.R. (1993) Differentiation 53, 139-147. [I371 Muller, K. and Gerisch, G. (1978) Nature 274, 445449. [I381 Browne, L.H., Sadeghi, H., Blumberg, D., Williams, K.L. and Klein, C. (1989) Development 105, 657664. [I391 Choi, A.H. and Siu, C.H. (1987) J. Cell Biol. 104, 1375-1387. [I401 Kamboj, R.K., Lam, T.Y. and Siu, C.H. (1990) Cell. Regul. I , 715-729. [I411 Mueller, K, Gerisch, G., Fromme, I., Mayer, H. and Tsugita, A. (1979) Eur. J. Biochem. 99, 419426. [I421 Yoshida, M., Stadler, J., Bertholdt, G. and Gerisch, G. (1984) EMBO J. 3, 2663-2670. [I431 Siu, C.H. (1990) BioEssays 12, 357-362. [I441 Kamboj, R.K., Gariepy, J. and Siu, C.H. (1989) Cell 59, 615-625. [I451 Ziska, S.E. and Henderson, E.J. (1988) Proc. Natl. Acad. Sci. USA 85, 817-821. [I461 Boose, J.A. and Henderson, E.J. (1991) Glycobiology I , 295-305. [I471 Boose, J.A., Ziska, S.E. and Henderson, E.J. (1988) Dev. Genet. 9, 569-578. [I481 Stadler, J., Bauer, G. and Gerisch, G. (1984) FEBS Lett. 172, 338. [I491 Stadler, J., Keenan, T.W., Bauer, G. and Gerisch, G. (1989) EMBO J. 8, 371. [I501 Wilhelms, 0.-H., Luderitz, O., Westphal, 0. and Gerisch, G. (1974) Eur. J. Biochem. 48, 89-101. [I511 Sadeghi, H., Williams, K. and Klein, C. (1987) J. Biol. Chem. 262, 16294-16299. [I521 Sadeghi, H., da Silva, A.M. and Klein, C. (1988) Proc. Natl. Acad. Sci. USA 85, 5512-5515. [I531 Da Silva, A.M. and Klein, C. (1989) Exp. Cell Res. 185, 464472. [I541 Steinemann, C. and Parish, R.W. (1980) Nature 286, 621-623. [155] Gao, E.N., Shier, P and Siu, C.H. (1992) J. Biol. Chem. 267, 9409-9415. [156] Siu, C.H., Des Roches, B. and Lam, T.Y. (1983) Proc. Natl. Acad. Sci USA 80, 6596-6600. [I571 Keller, T., Eitle, E., Balding, K., Corrick, C. and Parish, R.W. (1994) FEBS Lett. 339, 119-123. [I581 Gezelius, K. and Rinby, B.G. (1957) Exp. Cell. Res. 12, 265-289. [I591 Williams, K.L., Vardy, PH., Breen, E.J., Eliott, S. and Fuchs, M. (1990) In: Developmental and Molecular Biology o f Dictyosteliurn discoideurn. EMBO/NSF Workshop, p. 1 . [I601 Freeze, H. and Loomis, W. (1977) J. Biol. Chem. 252, 820-824. [I611 Freeze, H. and Loomis, W.F. (1977) Dev. Biol. 56, 184-194. [ 1621 Blanton, R.L. and Northcote, D.H. (1 990) Planta 180, 324-332. [I631 Blanton. R.L. (1993) Development 119, 703-71 0. [I641 Grant, W. and Williams, K. (1983) EMBO J. 2, 935-940. [I651 Grant, W.N., Welker, D.L. and Williams, K.L. (1985) Mol. Cell. Biol. 5, 2559-2566. [I661 Breen, E.J. and Williams, K.L. (1988) Dev. Genet. 9, 539-548. [167] Vardy, P.H., Fisher, L.R., Smith, E. and Williams, K.L. (1986) Nature 320, 526-529.

[I181 [I191 [I201 [I211 [122] [I231 [I241 [I251 [I261 [I271 [I281 [I291 [I301 [I311 [I321 [I331 [I341 [I351

121 [I681 Morrissey, J.H., Devine, K.M. and Loomis, W.F. (1984) Dev. Biol. 103, 414424. [I691 Early, A., McRobbie, S.J., Duffy, K.T., Jermyn, K.A., Tilly, R., Ceccarelli, A. and Williams, J.G. (1988) Dev. Genet. 9, 383402. [I701 Williams, J.G., Ceccarelli, A., McRobbie, S., Mahbubani, H., Kay, R.R., Early, A,, Berks, M. and Jermyn, K.A. (1987) Cell 49, 185-192. [I711 Ceccarelli, A., McRobbie, S.J., Jermyn, K.A., Duffy, K. and Williams, J.G. (1987) Nucl. Acids Res. 15, 7463-7466. [I721 Ozaki, L.S., Svec, P., Nussenzweig, R.S., Nussenmeig, V. and Godson, G.N. (1983) Cell 34, 815-822. [I731 Prat, A., Katinka, M., Caron, F. and Meyer, E. (1986) J. Mol. Biol. 189, 47-60. [I741 Morrison, A., Blanton, R.L., Grimson, M., Fuchs, M., Williams, K. and Williams, J. (1994) Dev. Biol. 163, 457466. [I751 Ikeda, T. and Takeuchi, 1. (1971) Dev. Growth Differ. 13, 221-229. [I761 Oyama, M., Maeda, Y. and Takeuchi, I. (1984) Protoplasma 123, 152. [I771 Wuestehube, L.J., Chia, C.P. and Luna, E.J. (1989) Cell Motil. 13, 245-263. [I781 Lenhard, J.M., Kasperek, E., Moore, B.R. and Free, S.J. (1989) Exp. Cell Res. 182, 242-255. [I791 Lenhard, J.M., Siegel, A. and Free, S.J. (1989) J. Cell Biol. 109, 2761-2769. [I801 Seshadri, J., Cotter, D. and Dimond, R. (1986) Exp. Mycol. 10, 131-143. [I811 Fosnaugh, K.L. and Loomis, W.F. (1989) Nucl. Acids Res. 17, 9489-9489. [I821 Fosnaugh, K.L. and Loomis, W.F. (1989) Mol. Cell Biol. 9, 5215-5218. [I831 Fosnaugh, K.L. and Loomis, W.F. (1991) Dev. Genet. 12, 123-132. [I841 Devine, K., Morrissey, J. and Loomis, W. (1982) Proc. Natl. Acad. Sci. USA 79, 7361-7365. [I851 Akalehiywot, T. and Siu, C.H. (1983) Can. J. Biochem. Cell Biol. 61, 966-1001. [I861 Watson, N., Williams, K.L. and Alexander, S. (1993) J. Biol. Chem. 268, 22634-22641. [I871 Watson, N., McGuire, V. and Alexander, S. (1994) J. Cell Sci. 107, 2567-2569. [188] Takemoto, K., Yamamoto, A. and Takeuchi, I. (1985) J. Cell Sci. 77, 93-108. [I891 Sussman, M. and Lovgren, N. (1965) Exp. Cell Res. 38, 97-105. [ 1901 Ikeda, T. (1 98 1) Biochim. Biophys. Acta 675, 69-76. [I911 Varki, A. (1993) Glycobiology 3, 97-130. [I921 Varki, A. and Marth, J. (1995) Semin. Dev. Biol. 6, 127-138. [I931 Marx, J. (1992) Science 258, 402403. [194] Devine, K., Bergmann, J. and Loomis, W. (1983) Dev. Biol. 99, 437446. [I951 Cooper, D.N.W. and Barondes, S.H. (1984) Dev. Biol. 105, 59-70. [I961 Gomer, R., Datta, S. and Firtel, R. (1986) J. Cell Biol. 103, 1999-2015. [ 1971 Loomis, W.F. (1 975) Dictyosfelium discoideum: A Development System. Academic Press, New York. [I981 Freeze, H.H., Koza-Taylor, P., Jones, J.A. and Loomis, W.F. (1990) J. Cell Biochem. 43, 2 7 4 2 . [I991 Harloff, C., Gerisch, G. and Noegel, A.A. (1989) Genet. Dev. 3, 201 1-2019. [200] Bozzaro, S. (1985) Cell Differ. 17, 67-82. [201] Springer, W.R. and Barondes, S.H. (1980) J. Biol. Chem. 258, 4698. [202] Springer, W.R. and Barondes, S.H. (1985) Dev. Biol. 109, 102.

This Page Intentionally Left Blank

J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins I1 0 1997 Elsevier Science B.V. All rights reserved CHAPTER 6

Hemocyanins Johannis P. Kamerling and Johannes F.G. Vliegenthart Bijvoet Center. Department of Bio-Organic Chemistry3 Utrecht University. Netherlands

Abbreviations Ara

arabinose

Arg CROSREL

arginine

LYs Man

CROSS RELaxation

MD

molecular dynamics

Fuc

fucose

Me-T

methy ltransferase

a2Fuc-T

a- 1,2-fucosyltransferase a- 1,3-fucosyltransferase

3MeMan

3-0-methylmannose

a3Fuc-T

3MeGal

3-0-methylgalactose

Gal

galactose

4MeGal

4-0-methylgalactose

fi3Gal-T

fi- 1,3-galactosyltransferae

NMR

nuclear magnetic resonance

GalNAc

N-acetylgalactosamine

PNGase-F

PGalNAc-T

p- 1,4-N -acetylgalactosaminyl-

peptide-N4-(N-acetylp-glucosaminy1)asparagine amidase F

transferase

lysine mannose

GDP

guanosine diphosphate

Pro

proline

Glc

glucose

ROESY

GlcNAc

N-acetylglucosamine

rotating-frame nuclear Overhauser enhancement spectroscopy

p2GlcNAc-T

p-I ,2-N-acetylglucosaminyl-

UDP

uridine diphosphate

transferase

Xaa

unidentified amino acid

Hc

hemocyanin

HSEA

hard sphere exoanomeric

XYl 02Xyl-T

0-1,2-xylosyltransferase

xylose

1. Introduction In nature, three types of oxygen-transporting proteins can be distinguished on the basis of the active site, namely, hemoglobins, hemerythrins, and hemocyanins (Hcs) [ 11. The latter type is found in the majority of the arthropod and mollusc species. In contrast to hemoglobin and hemerythrin, wherein Fe(I1) is essential for oxygen binding, Hc has a binuclear Cu(1) active site. The binding of oxygen in Hc is dependent on the pH and is accompanied by the formal valence change from Cu(1) to Cu(II), thereby giving the oxygenated Hc its characteristic blue color. Hemocyanin does not occur in blood cells, but is found freely dissolved in the hemolymph. It forms the major protein constituent (9&98%) of this fluid[2] in concentrations up to 120mg/ml [3,4], dependent on species, age and season. The high 123

124

molecular mass of the protein (molecular mass 4.5 x lo2-9x lo3 kDa) is of importance for the generation of a low osmotic pressure in the hemolymph. The protein structures of Hcs have been studied along different routes, including primary structure analysis (e.g. ref. [5]) and three-dimensional analysis using electron microscopy (e.g. refs. [6,7]) and X-ray crystallography (e.g. refs. [8,9]). In terms of quaternary structure, native Hc is built up from a number of structurally related subunits, and the oxygen binding is regulated by the allosteric behavior of Hc, based on complex interactions between subunits [4]. From the various studies, it can be concluded that the protein structures of arthropod and mollusc Hc differ specifically from each other, and an impression of the different Hc sizes is presented in Table 1 [ 101. Arthropod Hc consists of subunits with a molecular mass of about 75 kDa [ 1 11; the subunits have kidney-shaped structures and are built up from three domains[5,8]. Each subunit forms a functional unit, defined as a part of the Hc molecule that contains two Cu atoms and is capable of binding one oxygen molecule, and the subunits are arranged as hexamers (1 x 6) or multihexamers (2x6, 4x6, 6 x 6 or 8~6)[2,12,13].On the other hand, mollusc Hc is built up from subunits with a molecular mass of about 400 kDa [13], arranged as large cylindrical molecules (10 to 20 subunits). In general, each subunit is composed of eight functional units. Under specific conditions the cylinders (with collars) can dissociate into defined smaller molecules. In view of the pronounced differences between the structures of the Hc from both phyla, there is an ongoing discussion with respect to their origin, i.e., have they evolved independently or are they derived from one common ancestor (see e.g. refs. [4,14]). Most Hcs have been shown to be glycoproteins, but so far solid information with respect to the significance of the partly unusual glycosylation is not available. It has been found that there exist remarkable differences in carbohydrate content and monosaccharide composition between Hc from arthropods and Hc from molluscs [lo]. Detailed information, mainly based on monosaccharide analysis, methylation analysis, and 500 or 600 MHz 'H NMR spectroscopy, has become available on the primary structure of the carbohydrate chains of the Hcs of the scorpion Androctonus australis[151, the freshwater crayfish Astacus leptodactylus [ 161, the spiny lobster Panulirus interruptus[17,181, the terrestrial snail Helix pomatia [ 19-21], and the freshwater snail Lymnaea stagnalis[ 19,22,23]. Here, a survey is given of our present-day knowledge concerning the carbohydrate composition of Hcs.

2. Carbohydrate parameters of arthropod hemocyanins In Table 2 a survey is presented of the carbohydrate content and the monosaccharide composition of Hcs from typical species belonging to the four arthropod classes Merostomata, Arachnida, Crustacea and Chilopoda [lo]. As can be deduced from this table, the carbohydrate content of arthropod Hc is usually low. For those species containing > 0.1% carbohydrate, only D-Man and D-GIcNAc are observed, suggesting the occurrence of asparagine-linked glycans, dominated by the oligomannose type [24]. In view of our finding that the carbohydrate content of Eurypelma californicum is < 0.1 %, the reported data for the Hcs from the Araneae E. californicum, Eurypelma helluo

I25

Table 1 Survey of different types of hemocyanin distributed over the various species [ 101 Species

Number of subunits

Schematic structure

Molecular mass (kDa)

n"

4.5~10

Arthropod hemocyanin

Lobster

Crab

l

L

Spider

Scorpion

0

12 (2x6)

G

i

no

24 (4x6)

Centipede

36 (6x6)

Horseshoe crab

48 (8x6)

UU

2

9x10'

3

13x10

Mollusc hemocyanin Octopus, cuttle fish, squid; chiton

10 (2x5) (80 functional units)

% Snail (slug)

w e

20 (2x10) (160 functional units)

a Square represents 6 subunits (functional units):

Section represents 2 subunits (2x8 functional units):

00 =

9x1O3

126 Table 2 Carbohydrate content (“h,w/w) and monosaccharide composition (nmol sugar/mg protein) of Hcs from various arthropod species [ 101

Class

Order

Species

Monosaccharide Man

GlcNAc

%

(w/w)

Merostomata

Xiphosura

Limulus polyphemus (horseshoe crab)

Arachnida

Scorpiones

Androctonus australis (scorpion)

Aranea

Eurypelma calijornicum (tarantula)

Decapoda

Panulirus interruptus (subunit type a) (spiny lobster)

29.6

11.3

0.8

Panulirus interruptus (subunit type b)

39.7

16.1

1 .o

Panulirus interruptus (subunit type c)

35.9

15.2

1 .o

Astacus leptodactylus ( 1 x 6-mer) (freshwater crayfish)

31.9

22.0

1.1

Astacus leptodactylus (2 x 6-mer)

6.0

1.9

0.2

Scutigera coleoptrata (centipede)

182.0

78.0

4.9

Crustacea

Chilopoda

Scutigeromorpha

<0.1

34.8

7.9

0.8 <0.1

and Cupiennius s a k i (carbohydrate content: 1.5-2%; monosaccharide composition in mol sugar per 70000g Hc: Ara (I), Fuc (l), Man (1-2), Glc ( 5 - 6 ) and GlcNAc (23)) [25] have to be considered with care. In addition to the data summarized in Table 2, for the Decapoda Homarus americanus (lobster) a carbohydrate content of 0.9% has been found, whereas the monosaccharide analysis yielded 2-3 mol Man and 0.5 mol GlcNAc per 75000g Hc [26]. The carbohydrate content of the 1 x 6-mer Hc of Astacus leptodactylus is essentially the same as that of the (1 x 6-mer) Hc of Panulirus interruptus, but the A. leptodactylus 2 x6-mer Hc contains considerably less carbohydrate. This is in agreement with the proposal [27] that the carbohydrate chains prevent 2x6-mer formation. The near absence of carbohydrate in the large Hcs from Limulus polyphemus (8x6-mer) and E. californicum (4 x 6-mer) is in accordance with this theory; the presence of carbohydrate in Androctonus australis (4x6-mer) Hc, however, is in contrast with it. The centipede Scutigera coleoptrata has an unusually high carbohydrate content (4.9%) for an arthropod Hc. The deviation of the chilopod Hc from the other arthropod Hcs is also reflected by the three-dimensional structure, which is composed of six hexamers [ 131.

2.1. Androctonus australis hemocyanin glycans The Hc of the scorpion A. australis is a 4x6-mer glycoprotein built up from eight different subunit types (a-h) of polypeptide chains, each with a molecular mass of about 75 kDa [28-301. The positions of the different subunit types have been determined in the Hc quaternary structure by electron microscopy [3 11. Detailed analysis of the carbo-

127 Scheme 1 N-linked carbohydrate chain from Androclonus australis hemocyanin [IS] Man(a I-2)Man(a 1-6)

\

Man(al-6) Man(a 1-2)Man(a 1-3) I

\ Man(fl14)GlcNAc(fl 14)GlcNAc

Man(a 1-2)Man(a l-2)Man(al-3)

I

hydrate moiety of Hc, isolated via the hydrazinolysis procedure, i.e. anhydrous hydrazine treatment and re-N-acetylation, and converted into the corresponding oligosaccharidealditol, demonstrated the presence of MangGlcNAc2 only (Scheme 1) [15]. With a carbohydrate content of 0.8%, the number of chains per molecule Hc amounts to 8.

2.2. Astacus leptodactylus hemocyanin glycans The freshwater crayfish A. leptodactylus has 1x 6-mer and 2 x 6-mer Hcs. The molecular mass of the 2 x 6-mer glycoprotein is about 900 kDa and the carbohydrate content 0.2% [ 10,161. Of the different subunits, a partial amino acid sequence is to date only known for subunit b [32]. Only one of the subunits is glycosylated, but this is not subunit b [5,16]. Characterization of the carbohydrate chains of Hc (2 x 6-mer), enzymatically released by PNGase-F, shows the occurrence of a series of six oligomannose-type structures varying from Man9GlcNAcz to Ma~GlcNAc2.Man6GlcNAc2 represents the most abundant chain with 57% of all oligosaccharides (Scheme 2) [16]. 2.3. Panulirus interruptus hemocyanin glycans The Hc molecules of the spiny lobster I! interruptus are present as hexamers, built up from four different subunits, denoted a, b, b’ and c, each with a molecular mass of about 75 kDa. Sequence analysis data of the various subunits have been reported in refs. [33-361. The three-dimensional structure of the Hc has been determined by X-ray diffraction, with crystals consisting of a mixture of subunits a and b in roughly equal amounts[8,37]. The glycosylation site of subunit a, present in the first domain, has been identified as Asn-167 [34]; the same holds for the structurally closely related subunit b [35]. In subunit c, the glycosylation site is at Asn-476 in the third domain [36]. The X-ray studies demonstrated that the glycan moiety protrudes into a solvent region, and, except for the asparagine-bound GlcNAc unit, the glycan does not make contacts with the amino acid side chains [8]. The carbohydrate chains of native Hc were analyzed after isolation via the hydrazinolysis procedure and conversion into their oligosaccharide-alditols, revealing the occurrence of Man~GlcNAc2,GlcNAcMan3GlcNAc2 and sulfated GlcNAcMan3GlcNAcf (molar ratio, 4 5 : 1; Scheme 3) [ 17,181.

128 Scheme 2 N-linked carbohydrate chains from Asfucus leptoducfylus hemocyanin [ 161 Man(al -6)

\

Man(al4) \ M an(a 1-3) I Man@ I 4 ) G l c N A c ( ~ I 4 ) G l c N A c

Man(a 1-6) M an(a 1-3) I Man(al-2)Man(al-3)

\ Man(fil4)GlcNAc(@14)GlcNAc I

Man(a 1-6)

\

Man(a 1 4 ) I \ Man(a 1-3) Man(pl4)GlcNAc(fi14)GlcNAc I Man(a 1-2)Man(a 1-3)

Man(a I-2)Man(alP6)

\

Man(al-6) \ Man(a 1-3) I Man(fil4)GlcNAc(~14)GlcNAc I Man(a l-2)Man(al-3)

Man(a I-2)Man(a 1-6)

\

Man(al-6) Man(a1-3) I Man(a I-2)Man(a 1-2)Man(al-3)

\ Man(fi14)GlcNAc(fi 14)GlcNAc I

Man(a l-2)Man(a 1-6)

\

Man(a 1-6) \ Man(a l-2)Man(al-3) I Man(B1 4)GlcNAc(fi14)GlcNAc Man(a 1 -2)Man(a 1 -2)Man(al-3)

I

129

Scheme 3 N-linked carbohydrate chains from Punulirus interruprus hemocyanin [ 17,181 Man(a 1-6) \ Man(a 1-6) I

Man(a1-3f

1

\

Man(fil-4)GlcNAc(fi 1 -4)GlcNAc I

Man(a1-3) Man(a 1-6)

\ Man(fil-4)GlcNAc(~1-4)GlcNAc I

GlcNAc(P 1-2)Man(a 1-3) (S04-6)Man(al 4)

\ Man(fiI4)GlcNAc(fiI-4)GlcNAc I

GlcNAc(fi1-2)Man(a 1-3)

3. Carbohydrate parameters of mollusc hemocyanins In Table 3 a summary is compiled of the carbohydrate content and the monosaccharide analysis data of Hcs from typical species belonging to the four mollusc classes Gastropoda, Cephalopoda, Bivalvia, and Amphineura [lo]. The carbohydrate content of mollusc Hc is higher than that of most arthropods investigated so far. It is remarkable that the monosaccharide composition of mollusc Hc exhibits more variation than that of arthropods. All species contain D-Man, D-Gal, D-GalNAc, and D-GlcNAc. Additionally, the Hcs of Lymnaea stagnalis and Helix pomatia contain L-FUCand a series of monosaccharides unusual for animal glycoproteins, namely, D-XYI, 3-O-methyl-~-Man [3MeMan] and 3-O-methyl-~-Gal [3MeGal]. It should be mentioned that in nature Xyl is a typical constituent of proteoglycans and of plant glycoproteins (for a literature survey, see ref. [38]). 3MeMan was also detected in the Hcs of Acila castrensis and Stenoplex conspicua. As will be discussed below, in the case of the Hcs of L. stagnalis and H. pomatia only asparagine-linked carbohydrate chains could be demonstrated to occur, indicating that GalNAc is a constituent of the latter type of chains. It has to be noted that the literature data available for the Hcs of Busycon canaliculatum (Man, GlcNAc) [26] and Octopus uulgaris (Fuc, Man, GlcNAc) [39] do not fit the monosaccharide analysis data reported in Table 3. 3.I . Lymnaea stagnalis hemocyanin glycans The Hc of the freshwater snail L. stagnalis has a molecular mass of about 9 x lo3 kDa and a carbohydrate content of 3% (Table 3). So far, no detailed amino acid sequence data are available. Analysis of the glycosylation pattern of the Hc (hydrazinolysis procedure; oligosaccharide-alditols) afforded a series of 8 oligosaccharide structures (Scheme 4) [22,23]. The low-molecular-mass oligosaccharide is a xylosylated core

Table 3 Carbohydrate content (%, d w ) and monosaccharide composition (nmol sugarimg protein) of Hcs from various mollusc species [ 101 Class

Gastropda

Order

Species

Monosaccharide

%

Fuc

Xyl

3MeMan

3McGal

Man

Gal

GalNAc GlcNAc (w/w)

-

115.0

136.4

23.6

47.3

69.1

9.0

4.2

42.4

18.3

17.0

33.9

3.0

Stylommatophora

Helix pomatia (roman snail)

40.9

42.7

Bassomatophora

Lymnaea sfagnalis (pond snail)

15.5

9.9

Archaeogastropoda

Megathura crenulata (keyhole limpet)

36.9

-

-

-

63.3

49.3

34.0

44.3

4.3

Stenoglossa

Rusycon cnrico (whelk)

23.1

-

-

-

116.4

6.1

12.8

44.7

3.9

Buccinum undatum (whelk)a

1.1

-

-

-

5.1

0.5

0.9

3.3

4.0

Neptunica antiqua a

0.8

-

-

-

5.1

0.8

0.8

3.1

4.0

Colus gracilus a

1.1

-

-

-

5.8

0.7

2.7

1.1

4.0

Loligo Jurbesi (squid)

1.2

-

-

-

63.0

9.8

9.0

45.0

2.5

Sepia oficinalis (cuttle fish)

8.5

-

-

-

77.7

8.5

5.5

39.2

2.1

Octopus uulgaris (octopus)

-

-

-

-

89.4

8.3

2.8

26.0

2.3

Cephalopoda Dihranchia

15.5

Bivalvia

Protobranchia

Acila castrensis

28.9

-

37.0

-

49.5

63.4

21.1

41.8

5.2

Amphineura

Chitonida

Mopaliu muscosa (chiton)

15.1

-

-

-

126.4

22.1

4.8

36.5

3.8

Nuttallina frwra (chiton)

31.8

-

-

-

116.0

33.2

29.5

53.4

5.0

Stenoplex conspicua (chiton)

25.6

-

9.9

-

70.2

9.4

5.0

28.0

2.4

a

mol Carbohydrate/functional unit (50000 8); these data have been taken from [40].

131 Scheme 4 N-linked carbohydrate chains from Lymnaea stagnalis hemocyanin [22,23] 3MeMan(a 1-6)

\

Man(fil4)GlcNAc(fi14)GlcNAc 3MeMan(a1-3)

/I

XYKfil-2) 3 MeGal(0 1-3)GalNAc(fi 14)GlcNAc(fiI -2)Man(a 1-6)

\

Man@ 1 4)GlcNAc(fi14)GlcNAc 3MeMan(al-3)

/I

XYKfi1-2) Man(a 1-6)

\

Man@ 14)GlcNAc(fi14)GlcNAc 3MeGal(fi 1-3)GalNAc(fi 14)GlcNAc(fi1-2)Man(a 1-3)

/I

XYl(fi1-2) 1-6) Fuc(a 1 -2)Gal(fi 1-3)GalNAc(fil4)GlcNAc(fil-2)Man(a

\

Man(fiI4)GlcNAc(fiI4)GlcNAc 3MeMan(al-3)

/I

XYW 1-21 Man(a1-6)

\

Man@ 14)GlcNAc(fi14)GlcNAc Fuc(a 1 -2)Gal(~1-3)GalNAc(fil4)GlcNAc(fi1-2)Man(a 1-3)

/I

XYl(fi1-2) 3 MeGal(fi 1-3)GalNAc(fi 14)GlcNAc(fiI -2)Man(a 1-6)

\

3MeGal(fi1-3)GalNAc(fi 14)GlcNAc(fiI -2)Man(a 1-3)

Man@ 14)GlcNAc(fiI4)GlcNAc I

I

xYi(pI -2 j 3MeGal(@1 -3)GalNAc(fi 14)GlcNAc(fi1-2)Man(a 1-6) \

\

Man(fiI4)GlcNAc(fil4)GlcNAc Fuc(aI-2)Gal(fi1-3)GalNAc(fiI4)GlcNAc(fi1-2)Man(a1-3)

/I

XYKfiI-2)

continued on next page

132

Scheme 4,continued Fuc(a 1-2)Gal(b 1 -3)GalNAc(fi 14)GlcNAc(fi1-2)Man(a 1-6) \

Man((314)GlcNAc(fiI4)GlcNAc

Fuc(a 1-2)Gal(P 1-3)GalNAc(P 14)GlcNAc(fi 1-2)Man(a 1-3)

/I

XYl(B1-2)

structure, of which both terminal Man residues are 3-0-methylated. The Xyl(b12)Man(b14) element is frequently found in plant glycoproteins, but is highly unusual for animal glycoproteins. It should be noted that structures containing a (3-1,2-linked D-XYI residue attached at the P-D-Man unit of the carbohydrate core are highly immunogenic in mammalian species [4 1,421. The high-molecular-mass oligosaccharides are also xylosylated, and are all of the N,N’-diacetyllactosediamine type. The Fuc(a12)GaI(fil-3)GalNAc(~l- element is known as the blood group H type 4 determinant. In addition to the antennary structures depicted in Scheme 4, there is some evidence that also a diantennary compound is present having Fuc(a 1-2)GaI(~I-3)GalNAc(~14)GlcNAc@ 1-2)Man(a 1-6) and 3MeGal(P 1-3)GalNAc(b 14)GlcNAc(b l-2)Man(a 1-3) extensions. Although in principle both aMan residues can be methylated, as is evident from the structure of the low-molecular-mass oligosaccharide, in the extended mono-antennary structures only a terminal a- 1,3-linked Man unit is methylated. 3.2. Helix pomatia hemocyanin glycans

For the Hc of the terrestrial (vineyard) snail H. pomatia (molecular mass -9x lo3 kDa) two types occur, designated a ( a D + aN) and P (bc). Under certain conditions aD-Hc can be dissociated into 1M-molecules, in contrast to aN-HC and Pc-Hc [43]. The amino acid compositions of both forms are very similar [44]. Of the 8 functional units a-h of Pc-Hc, for functional unit d the complete amino acid sequence, including the glycosylation sites at Asn-253 and Asn-387, has been determined [45,46]. In the case of functional unit g, a partial amino acid sequence has been reported, comprising also three N-glycosylation sites [46]. Monosaccharide analyses have been carried out on Hc and on functional units of Hc. aD-Hc (Table 3) [lo] and ~ N - H give c rise to similar monosaccharide ratios [46], but &Hc showed higher values for 3MeCa1, GlcNAc and GalNAc [46]. For monosaccharide analysis data of the eight functional units of Pc-Hc, see ref. [47]. The aD-HC, investigated for its glycosylation pattern, has a carbohydrate content of 9% (Table 3). The carbohydrate chains of aD-Hc have been released along chemical (hydrazine) and enzymatic (PNGase-F) routes, and oligosaccharide-alditols as well as free oligosaccharides have been analyzed. A survey of the various structures found so far (24 oligosaccharide chains) is presented in Scheme 5 [20,21]. Inspection of these glycan chains shows that the antennae are of a higher complexity than those of L. stagnalis Hc. Of interest are the repetition of 3MeGal residues at 3,6-branched GalNAc of the NN’-diacetyllactosediamine element, and the detection of the previously missed 4-0-methyl-~-galactose(4MeGal) residue. Besides the xylosylated core structure

133 Scheme 5 N-linked carbohydrate chains from Helix pomatia hemocyanin [20,2 1 ] Man(a1-6)

\ Man(/? 14)GlcNAc(fi14)GlcNAc /

Man(a 1-3) Man(a 1-6)

\

Man(fil4)GlcNAc(~14)GlcNAc Man(a 1-3)

/I

XYW-2) Man(a 1-6)

Fuc(a 1-6)

\

\

Man(fiI4)GlcNAc(fi14)GlcNAc

Man(a1-3)

/I

XYUfi1-2) 3MeGal(fi1-6) 3MeGal(fl14)

\

\

3MeGal(fi1-3)GalNAc(fll4)GlcNAc(fil-2)Man(a 14)

Fuc(a1-6)

\

\

Man(fil4)GlcNAc(fil 4)GlcNAc Man(a1-3)

/I

XYKfi1-2) 3MeGal(fiI -6) \

3MeGal(fi1-6) 3MeGal(fi14)

\

\

3MeGal(fil-3)GalNAc(~l4)GlcNAc(fil-2)Man(al-6)

Fuc(a 1-6)

\

\

Man@ 1-4)GlcNAc(fi 1-4)GlcNAc Man(a1-3)

/I

XYW-2) 3MeCal(fi1-6)

\

3MeGal(fi1-6) 3MeGal(fi1-6) 5

\

5

\

Fuc(a 1-6)

3MeGal(fi 1-3)GalNAc(~l4)GlcNAc(fil-2)Man(a1-6)

\

\

Man(fil4)GlcNAc(fi14)GlcNAc Man(a 1-3)

/I

XYUfi1-2)

continued on next page

134

Scheme 5 , continued

\ 4MeGal(fil-3)GalNAc(fil4)GlcNAc(fiI-2)Man(al-6)

Fuc(al-6)

\

\

Man@14)GlcNAc(fi1-4)GlcNAc M an(a 1-3) I / XYUfiI -2)

Man(a 1-6)

\ 3MeGal(fiI -6) 3MeGal(pl-6)

\

Man@ 1 4)GlcNAc(fi14)GlcNAc

\

3MeGal(fi1-3)GalNAc(fi 14)GlcNAc(fi 1-2)Man(a 1-3)

I

Man(a1-5) 3MeGal(fil-6) 3MeGal(fiI -6)

\

\

Man@14)GlcNAc(fi14)GlcNAc

\

3MeGal(fi1 -3)GalNAc(fi 14)GlcNAc(fi1-2)Man(a 1-3)

/I

XYKfi1-2) Man(a 1-6)

Fuc(a 1-6)

\

\

3MeGal(fi1-6) 3MeGal(fi1-6) Man@ I ~ ) G l c N A c ( ~ I 4 ) G l c N A c \ \ 3MeGal(fiI -3)GalNAc(fiI -4)GlcNAc(fil -2)Man(a 1-3)

/I

Xyl(fi 1-2) Man(a 14)

3MeGal(fi1 -6)

\ 3MeGal(fi1-6)

3MeGal(fi1-6)

\

\

Man@ 14)GlcNAc(fi14)GlcNAc

\

3MeGal(fil-3)GalNAc(fiI4)GlcNAc(fiI -2)Man(a 1-3) I /

XYl(fi1-2) 3MeGal(fi1-6)

Man(a 1-6)

\

\

3MeGal(fi1-6) 3MeGal(fi1-6)

\

\

3MeGal(fiI -3)GalNAc(fiI -4)GlcNAc(fi1-2)Man(a 1-3)

Man@ I-4)GlcNAc(fil-4)GlcNAc I

Man(a 14)

3MeGal(fi 1-6)

\ 3MeGal(fi1-6) 3MeGal(fiI 4) \ \

\

Man@ 1-4)GlcNAc(fi 1-4)GlcNAc

3MeGaI(fi 1-3)GalNAc(fi 1-4)GlcNAc(fi 1-2)Man(a 1-3) I / XYKPI-2)

Man(a 1 4 )

3MeGal(fi1-6)

\

\

Fuc(al-6)

\

3MeGal(fi1-6) 3MeGal(fi14) Man(fiI-4)GlcNAc(fiI4)GlcNAc \ \ 3MeGal(fiI -3)GalNAc(fil4)GlcNAc(fi1-2)Man(a1-3)

/I

XYKfiI-2)

conlinued on next page

135

Scheme 5, continued Man(a 1 4 )

\

3MeGal(fi14)

Man@ 14)GlcNAc(fi 14)GlcNAc

\

4MeGal(fi1-3)GalNAc(fiI4)GlcNAc(fi1-2)Man(a 1-3)

/I

XYW-2) 3MeGal(@1-6) 3MeGal(@l4)

\

\

3MeGal(fi 1-3)GalNAc(@ 14)GlcNAc(fi 1-2)Man(a 14) 3MeGal(fi1-6) 3MeGal@I4)

\

\

Man(@14)GlcNAc(p1 4)GlcNAc

\

3MeGal(fi 1-3)GalNAc(b 14)GlcNAc(fi1-2)Man(a 1-3)

/I

XY 1(fi 1-2) 3MeGal(fi1-6) 3MeGal(fiI 4)

\

\

Fuc(a1-6)

3MeGal(fi1-3)GalNAc(fiI 4)GlcNAc(fi1 -2)Man(al4)

\

\

3MeGal(pl-6) 3MeGal(fild) Man@ 14)GlcNAc(fi14)GlcNAc \ \ 3MeGal(fi 1-3)GalNAc(fi1 4)GlcNAc(@1-2)Man(a 1-3)

/I

XY@-2) 3MeGal(fi1-6)

\

3MeGal(fi1-6) 3MeGal(fil4)

\

\

Fuc(a1-6)

3MeGal(fi 1-3)GalNAc(p 14)GlcNAc(@1-2)Man(a 1 4 )

\

3MeGal(fi1-6) 3MeGal(fi1 4 )

\

Man@14)GlcNAc(P 14)GlcNAc

\

\

3MeGal(@1-3)GalNAc(fil4)GlcNAc(fi 1-2)Man(a 1-3)

/I

XY@-2) 3MeGal(fi1-6)

\

3MeGal(fi1-6) 3MeGal(fi14)

\

\

3MeGal(fi1-3)GalNAc(fiI -4)GlcNAc(fi1-2)Man(al-6) 3MeGal(p1-6) 3MeGal(fild)

\

\

Man(fi14)GlcNAc(fi14)GlcNAc

\

3MeGal(fi 1-3)GalNAc(fi 14)GlcNAc(fi1-2)Man(a 1-3)

/I

XY Kfi 1-21 3MeGal(fi1-6)

\

3MeGal(fi1-6) 3MeGal(fil4)

\

\

Fuc(a1-6)

3MeGal(fi1 -3)GalNAc(fi 14)GlcNAc(fi 1-2)Man(a 1 4 ) 3MeGal(fil-6) 3MeGal(fil4)

\

\

\

\

Man(fi14)GlcNAc(fil4)GlcNAc

3MeGal(fi1-3)GalNAc(fiI4)GlcNAc(fiI -2)Man(al-3)

/I

XYUfiI -2)

continued on next page

136

Scheme 5, continued 3MeGal(fild) 3MeCa1((314)

\

3MeGa1(fi1-6)

\

Fuc(a 1-6)

3MeGal(fi1-3)GalNAc(fi 14)GlcNAc(fi 1 -2)Man(a 1-6)

\

\

3MeGal(fil-6) 3MeGal(fil-6)

\

Man(fiIM)GlcNAc(fi I -4)GlcNAc

\

\

3MeGal(fi1-3)GalNAc(fiI 4)GlcNAc(fil -2)Man(a 1-3)

I / XYl(B1-2) 3MeGal(fi1-6) 3MeGal(fi14)

\

\

3MeGal(fi1-3)GalNAc(fi 1-4)GlcNAc(fi 1-2)Man(a 1-6) 3MeGal(fi1-6)

\

\

Man(fi14)GlcNAc(~I4)GlcNAc

\

3MeGal(fi1-6) 3MeGal(fi1-6)

Fuc(a 1-6)

\ \

3MeGal(fi I -3)GalNAc(fi 1-4)GlcNAc(fi 1-2)Man(a 1-3)

'i

XYl(fi 1-21 3MeGal(f314)

\ 4MeGal(fi1-3)GalNAc(fiI 4)GlcNAc(fil-2)Man(al-6) 3MeGal(fi1-6) 3MeGal(fi14)

\

Man@ 1-4)GlcNAc(P 14)GlcNAc

\

\

Fuc(a 1-6)

\

3MeGal(fi1-3)GalNAc(fi 14)GlcNAc(fiI-2)Man(a 1-3)

/I

XYKfil-2) 3MeGal(fi1-6) 3MeCal(fi14)

\

\

3MeGal(fil-3)GalNAc(fil-4)GlcNAc(fil-2)Man(al-6) 3MeGal(fi1-6)

\

\

Fuc(a 1-6)

\

Man(fi1-4)GlcNAc(fi14)GlcNAc

4MeGal(fi 1-3)GalNAc(fi14)GlcNAc(fi 1-2)Man(a 1-3)

'I

XYW 1-21

(XylMan3GlcNAc2) also non-xylosylated and a-l,6-fucosylated core structures occur. It has to be mentioned that in plant glycoproteins the frequently found xylosylated trimannosyl-N,N'-diacetylchitobiosyl core structure can occur with a 1,3-1inked in stead of a 1,6-linked aFuc residue, attached at the asparagine-bound GlcNAc unit (for a review, see ref. [38]).

4. Synthesis and conformational analysis of xylose-containing elements of mollusc hemocyanin glycans In the framework of the primary structural analysis and the biosynthesis of mollusc hemocyanin glycans, attention has also been paid to the organic synthesis of structural

I37 Table 4 Survey of synthetic oligosaccharides related to mollusc hemocyanin glycans Compound

Ref.

1. Xyl(P1-2)Man(Bl-O)Me

2. Man(a1-3)[Xyl(fi1-2)]Man(fil-O)Me 3. Man(a 1-6)[Xyl@-2)]Man(fi I -0)Me 4. Man(al4)[Man(al-3)][Xyl(fiI -Z)]Man(fiI-O)Me 5. 3MeMan(al-6)[Man(aI-3)][Xyl(fi1-2)]Man(~l-O)Me

6. Man(a14)[3MeMan(a1-3)][XyI(fi1-2)]Man(fi1-O)Me 7. 3MeMan(al-6)[3MeMan(al-3)][Xyl(fil-2)]Man(fil-O)Me

8. Man(al-6)[Man(al-3)][Xyl(fi1-2)]Man(~1~)GlcNAc(~l4)GlcNAc(~l-O)Me 9. Man(al-6)[Man(a 1-3)][Xyl(fi1 -2)]Man(fiI 4)GlcNAc(BI 4 ) [ F u c ( a I4)]GlcNAc(fiI-0)Me

elements of xylose-containing carbohydrate chains. The various synthetically obtained carbohydrate chains are summarized in Table 4 [38,48-501. With respect to the conformational analysis of some of these compounds, ROESY studies and HSEA calculations have been carried out on compounds 1-4 [S 11, and Molecular Dynamics (MD) simulations in water in combination with ROESY studies on compounds 8 and 9 [52]. In the latter case the CROSREL program was applied to calculate theoretical ROESY cross-peak intensities from models obtained from the MD simulations [53]. It could be deduced [52] that for the flexible a-1,6 linkages, as well as for the Man(a1-3)Man linkage, ensembles of conformations are likely to exist. The Xyl(PI-2)Man(~1~)GlcNAc(~l-4)GlcNAc fragment is found in one rigid conformation. No significant differences were found for the corresponding structural elements in either of the two methyl glycosides. For conformational information on the xylose-containing glycan Xyl(P l-2)[Man(a 1-6)]Man(P 14)GlcNAc((31-4)[Fuc(a 1-3)IGlcNAc as part of the plant glycoprotein bromelain and of a glycopeptide derived from bromelain, see refs. [54,55].

5. Biosynthesis of Lymnaea stagnalis hemocyanin glycans One of the most conspicuous cell types in the connective tissue of gastropods is the pore cell. Several different functions have been attributed to these cells, one of these being the production and secretion of blood pigments into the hemolymph. It has been suggested that the pore cells of the freshwater snail L. stagnalis synthesize and store Hc[56]. Recently, a number of studies have appeared focusing on snail connective tissue glycosyltransferases involved in the biosynthesis of N-glycoprotein glycans [57601. To this end microsomal suspensions of this tissue were used as enzyme source. A large series of relevant substrates were tested for each glycosyltransferase activity. It is assumed that these enzymes are also involved in the biosynthesis of the N-glycans of L. stagnalis Hc. Based on the substrate specificities of the studied connective tissue glycosyltransferases (see below) and the structures of the Hc glycans (Scheme 4),

138

starting from MansGlcNAc2[Xaa,]AsnXaam a hypothetical biosynthetic scheme for the N-glycosylation in snail connective tissue has been published (Fig. 1) [59]. With respect to Xyl, a connective tissue microsomal suspension turned out to be capable of transferring Xyl from UDP-Xyl to R”-GlcNAc((3I-2)Man(a1-3)[R’-]Man((3l-R, where R” may be H or Gal((31-4), R’may be H, Man(a1-6) or GlcNAc((31-2)Man(a1-6), and R may be 4)GlcNAc, 4)GlcNAc((3I4)GlcNAc, 4)GlcNAc((314)GlcNAc((31-N)[Xaa,]AsnXaa,, or O)(CH&CH3, thereby forming [59] R”-GlcNAc(CJ 1-2)Man(a1-3)[Xy1((31-2)][R’-]Man(CJl-R. The detected enzyme activity was characterized as UDP-Xyl: GlcNAc((3I-2)Man(aI-3)Man((3 I-R (Xyl to Man@ 6- 1,2-~ylosyltransferase(02Xyl-T). The CJ2Xyl-T is not active on acceptors lacking a GlcNAc residue on the Man(a13) arm of the trimannosyl part of the core structure. This means that for the formation of low-molecular-mass compounds having a Xyl residue, but no GlcNAc(CJ12)Man(a 1-3) sequence (e.g. Man(a l-3)[Man(a 1-6)] [Xyl((31-2)]Man(CJ 14)GlcNAcz), several enzymatic steps are needed. Starting from MansGlcNAc2 [Xaa,]AsnXaa,,, first, (3- 1,2 N-acetylglucosaminyltransferase I ((32GlcNAc-T I) must introduce a GlcNAc residue at the Man(a1-3) arm. Then, after trimming with a-mannosidase 11, CJ2Xyl-T can introduce the Xyl residue. As is evident from Fig. 1, in principle this xylosylation can also occur after the introduction of a second GlcNAc residue at the Man(a1-6) arm, using (3- 1,2 N-acetylglucosaminyltransferase I1 (CJ2GlcNAc-T 11). Both (32GlcNAc-T I and I1 activities have been found in the connective tissue, whereas BGlcNAc-T 111-VI activities were absent [59]. It should be noted that the presence of a Xyl residue at CJMan totally abolished the (32GlcNAc-T I activity. Finally, non-reducing terminal GlcNAc residues are removed by (3-N-acetylhexosaminidase, a glycosidase which has been demonstrated to be present in connective tissue [59]. As already mentioned in section 3.1, xylose-containing N-glycans frequently occur in plant glycoproteins. In view of the results on Hc, it can be stated that the substrate specificity of the snail connective tissue (32Xyl-T, when focusing on the minimal structure requirements, is similar to that of the plant (32Xyl-T in Phuseolus oulguris cotyledons [6 I ] and Acer pseudoplatanus sycamore cells [62]. Snail connective tissue also contained the required UDP-GalNAc:GlcNAc@ I-R (3- 1,4-N-acetylgalactosaminyltransferase(CJ4GalNAc-T) activity, capable of transferring GalNAc from UDP-GalNAc to GlcNAc(CJI-O)R, yielding GalNAc((314)GlcNAc((31O)R [58]. The connective tissue (34GalNAc-T showed a broad specificity towards acceptors having non-reducing terminal (3GlcNAc residues, including saccharides, glycopeptides and glycoproteins with and without a Pro-Xaa-Arg/Lys-specific tripeptide motif. Even free GlcNAc can serve as an acceptor. This means that the snail fl4GalNAc-T is different from other (34GalNAc-T activities involved in the biosynthesis of N,N’-diacetyllactosediamine elements, namely, the Pro-Xaa-Arg/Lys-tripeptide-motifdependent (34GalNAc-T from mammalian pituitary glands [63], and the Pr+XaaArg/Lys-tripeptide-motif-independent(transferrin-specific) CJ4GalNAc-T from different animal cells [64] (both acting on glycoproteins only). A similar CJ4GalNAc-T activity as found in connective tissue was also detected in the albumen gland of the snail, and it has been suggested that both activities are probably from one enzyme [58]. The connective tissue UDP-Gal:GalNAc(fl 14)GlcNAc(CJ1-R (3- 1,3-galactosyltransferase ((33Gal-T) is responsible for transferring Gal from UDP-Gal to GalNAc(6 1-4)GlcNAc((31-O)R, affording Gal@ l-3)GalNAc( p 1-4)GlcNAc(p 1-0)R [571. This 03Gal-T

330 L-FUC

P2GlcNAc-T I

I

ManU

1

I

I

mono-antennary

D-Gal

0 n-GalNAc 0 D-GIcNAc o-Man

m methyl

-

n-

n

P2Xyl-T

BZGlcNAc-TIl

1

P2Xyl-T -----c

I

L

P2GlcNAc-T I1

I ~GaNAc-T i32Xyl-T

I

h4e-T

P-hexosaminidase I B-hexosaminidase 1, Me-T

--b

P2Xyl-T

- -c

a2Fuc-T

1 P2Xyl-T

- --

Fig. 1. Proposal for the biosynthesis of N-linked oligosaccharides in connective tissue from Lymnaea stugnalis [59]. The compounds are represented by short-hand symbolic notation as explained in the figure. Solid arrows indicate steps verified by experiments and dotted arrows indicate probable steps [ 5 7 4 0 ] . Man I1 represents a-mannosidase 11; abbreviations of other enzymes are explained in the text. Boxed structures are presented as formulae in Scheme 4, and have been isolated from L. stagnalis Hc [22,23].

140

differs clearly in substrate specificity from the UDP-Gal:GalNAc(a 1-R B3Gal-T in porcine submaxillary gland microsomes [65], and from the UDP-Gal:Gal(P 1-R B3Gal-T in snail albumen glands [66]. In additional experiments it was shown that the albumen gland contains both a UDP-Gal:GalNAc(~1-4)GlcNAc(~l -R P3Gal-T activity and a UDP-Gal:Gal(Bl-R p3Gal-T activity [57]. Finally, two distinct fucosyltransferase activities were detected in connective tissue [60]. One is the required GDP-Fuc:Gal@ 1-3)GalNAc (Fuc to Gal) a-1,2-fucosyltransferase (a2Fuc-T), capable of transferring Fuc from GDP-Fuc to Gal(B1-3)GalNAc acceptors, yielding Fuc(aI-2)Gal(P 1-3)GalNAc sequences. The other is a GDP-Fuc:Gal(P 1-4)GlcNAc (Fuc to GlcNAc) a-1,3-fucosyltransferase (a3Fuc-T), acting on Gal(B1-4)GlcNAc type acceptors, affording the Lewis x determinant Gal(B1-4)[Fuc(aI -3)IGlcNAc. The substrate specificity of the connective tissue a2Fuc-T is different from that of the other a2Fuc-Ts so far characterized, because it prefers Gal@ 1-3)GalNAc acceptors over Gal(P1-3)GlcNAc and Gal(B1-4)GlcNAc acceptors. It should be noted that in connective tissue neither B4Gal-T activity nor a-1,3-fucosylated structures have been detected so far. Therefore, the function of the snail a3Fuc-T is not yet known. Similar a2Fuc-T and a3Fuc-T activities were also demonstrated to occur in the snail albumen gland [60]. In earlier work, the origin of the 3-0-methylated monosaccharides has been studied[67]. These monosaccharides do not stem from the diet, but are biosynthetically formed in the cell. It was found that only very small amounts of injected 3-0-[3H]-methylD-Man were incorporated into the Hc carbohydrate chains. However, injection of ~-[methyl-'~C]methionineled to the incorporation of the labelled methyl group into 3MeGal of the Hc oligosaccharides. So far, information on the specific methyltransferase(s) (Me-T) involved is not available.

Acknowledgements The hemocyanin work from the authors' laboratory has been supported by the Netherlands Foundation for Chemical Research (SON) with financial aid from the Netherlands Organization for Scientific Research (NWO), and is part of the doctoral theses of Dr. J.A. van Kuik (1987), Dr. J.B. Bouwstra (1989), Dr. J.G.M. van der Ven (1993), Dr. J.P.M. Lommerse (1994), and Dr. H. Mulder (1 995) at Utrecht University.

References [I] [2] [3] [4] [5]

Van Holde, K.E. and Miller, K.I. (1982) Q. Rev. Biophys. 15, 1-129. Ellerton, H.D., Ellerton, N.F. and Robinson, H.A. (1983) Prog. Biophys. Mol. Biol. 41, 143-248. Mangum, C.P. (1985) Am. J. Physiol. 248, R505-R514. Markl, J. (1986) Biol. Bull. 171, 90-115. Linen, B., Soeter, N.M., Riggs, A.F., Schneider, H.-J., Schartau, W., Moore, M.D., Yokota, E., Behrens, PQ., Nakashima, H., Takagi, T., Nemoto, T., Vereijken, J.M., Bak, H.J., Beintema, J.J., Volbeda, A,, Gaykema, W.P.J. and Hol, W.G.J. (1985) Science 229, 519-524. [6] Van Bruggen, E.F.J. (1983) Life Chem. Rep. Suppl. 1, 1-14. [7] Van Bruggen, E.F.J. (1986) Micron Microscop. Acta 17, 167-173.

141 [8] Gaykema, W.P.J., Hol, W.G.J., Vereijken, J.M., Soeter, N.M., Bak, H.J. and Beintema, J.J. (1984) Nature 309, 23-29. [9] Gaykema, W.P.J., Volbeda, A. and Hol, W.G.J. (1985) J. Mol. Biol. 187, 255-275. [ 101 Van Kuik, J.A. (1 987) Thesis Utrecht University, The Netherlands. [ l l ] Rochu, D. and Fine, J.M. (1984) Comp. Biochem. Physiol. 79B, 4 1 4 5 . [12] Ellerton, N.F. and Ellerton, H.D. (1982) Biochem. Biophys. Res. Commun. 108, 1383-1387. [13] Mangum, C.P., Scott, J.L., Black, R.E.L., Miller, K.I. and van Holde, K.E. (1985) Proc. Natl. Acad. Sci. USA 82, 3721-3725. [I41 Lerch, K., Huber, M., Schneider, H.-J., Drexel, R. andLinzen, B. (1986) J. Inorg. Biochem. 26,213-217. [I51 Debeire, P., Montreuil, J., Goyffon, M., van Kuik, J.A., van Halbeek, H. and Vliegenthart, J.F.G. (1986) Carbohydr. Res. 151, 305-3 10. [I61 Tseneklidou-Stoeter, D., Gerwig, G.J., Kamerling, J.P. and Spindler, K.-D. (1995) Biol. Chem. HoppeSeyler 376, 531-537. [I71 Van Kuik, J.A., van Halbeek, H., Kamerling, J.P. and Vliegenthart, J.F.G. (1986) Eur. J. Biochem. 159, 297-30 1. [I81 Van Kuik, J.A., Breg, J., Kolsteeg, C.E.M., Kamerling, J.P. and Vliegenthart, J.F.G. (1987) FEBS Lett. 221, 150-154. [19] Hall, R.L., Wood, E.J., Kamerling, J.P., Gerwig, G.J. and Vliegenthart, J.F.G. (1977) Biochem. J. 165, 173-1 76. [20] Van Kuik, J.A., van Halbeek, H., Kamerling, J.P. and Vliegenthart, J.F.G. (1985) J. Biol. Chem. 260, 13984-13988. [21] Lommerse, J.P.M., Thomas-Oates, J.E., Gielens, C., Preaux, G., Kamerling, J.P. and Vliegenthart, J.F.G. (1997) Eur. J. Biochem., in press. [22] Van Kuik, J.A., Sijbesma, R.P., Kamerling, J.P., Vliegenthart, J.F.G. and Wood, E.J. (1986) Eur. J. Biochem. 160, 621425. [23] Van Kuik, J.A., Sijbesma, R.P., Kamerling, J.P., Vliegenthart, J.F.G. and Wood, E.J. (1987) Eur. J. Biochem. 169, 3 9 9 4 1 1. [24] Kamerling, J.P (1994) Pure Appl. Chem. 66, 2235-2238. [25] Markl, J., Schmid, R., Czichos-Tiedt, S. and Linzen, B. (1976) Hoppe-Seyler’s 2. Physiol. Chem. 357, 1713-1725. [26] Waxman, L. (1975) J. Biol. Chem. 250, 3796-3806. [27] Nakashima, H., Behrens, P.Q., Moore, M.D., Yokota, E. and Riggs, A.F. (1986) J. Biol. Chem. 261, 10.526-10533. [28] Lamy, J., Lamy, J. and Weill, J. (1979) Arch. Biochem. Biophys. 193, 140-149. [29] Jolles, J., Jolles, P., Lamy, J. and Lamy, J. (1979) FEBS Lett. 106, 289-291. [30] Sizaret, P-Y., Frank, J., Lamy, J., Weill, J. and Lamy, J.N. (1982) Eur. J. Biochem. 127, 501-506. [31] Lamy, J., Bijlholt, M.M.C., Sizaret, P-Y., Lamy, J. and van Bruggen, E.F.J. (1981) Biochemistry 20, 1849-1856. [32] Schneider, H.-J., Voll, W., Lehmann, L., Grisshammer, R., Goettgens, A. and Linzen, B. (1986) In: B. Linzen (Ed.), Invertebrate Oxygen Carriers. Springer, Berlin, pp. 172-176. [33] Soeter, N.M., Jekel, P.A., Beintema, J.J., Volbeda, A. and Hol, W.G.J. (1987) Eur. J. Biochem. 169, 323-332. [34] Bak, H.J. and Beintema, J.J. (1987) Eur. J. Biochem. 169, 333-348. [35] Jekel, PA., Bak, H.J., Soeter, N.M., Vereijken, J.M. and Beintema, J.J. (1988) Eur. 3. Biochem. 178, 403-4 12. [36] Neuteboom, B., Jekel, P.A. and Beintema, J.J. (1992) Eur. J. Biochem. 206, 243-249. [37] Volbeda, A. and Hol, W.G.J. (1989) J. Mol. Biol. 209, 249-279. [38] Kerekgyartb, J., Kamerling, J.P., Bouwstra, J.B., Vliegenthart, J.F.G. and Liptak, A. (1989) Carbohydr. Res. 186, 5 1 4 2 . [39] Albergoni, V, Cassini, A. and Salvato, B. (1972) Comp. Biochem. Physiol. 41B, 445451. [40] Hall, R.L. and Wood, E.J. (1976) Biochem. SOC.Trans. 4, 307-309. [41] Kaladas, P.M., Goldberg, R. and Poretz, R.D. (1983) Mol. Immunol. 20, 727-735. [42] Faye, L. and Chrispeels, M.J. (1988) Glycoconjugate J. 5, 245-256.

142 [43] Heirwegh, K., Borginon, H. and Lontie, R. (1961) Biochim. Biophys. Acta 48, 517-526. [44] Van Holde, K.E. and Cohen, L.B. (1965) Biochemistry 3, 1803-1808. [45] Drexel, R., Siegmund, S., Schneider, H.-J., Linzen, B., Gielens, C., Prkaux, G., Lontie, R., Kellermann, J. and Lottspeich, F. (1987) Biol. Chem. Hoppe-Seyler 368, 617435. [46] Vanhoegaerden, R. (1988) Thesis Catholic University of Louvain, Belgium. [47] Wood, E.J., Chaplin, M.F., Gielens, C., De Sadeleer, J., Preaux, G. and Lontie, R. (1985) Comp. Biochem. Physiol. 82B, 179-1 86. [48] KerkkgyLrtt6, J., van der Ven, J.G.M., Kamerling, J.P., LiptLk, A. and Vliegenthart, J.F.G. (1993) Carbohydr. Res. 238, 135-145. [49] Van der Ven, J.G.M., Wijkmans, J.C.H.M., Kamerling, J.P. and Vliegenthart, J.F.G. (1994) Carbohydr. Res. 253, 121-139. [SO] Van der Ven, J.G.M., Kerekgyirtb, J., Kamerling, J.P., LiptLk, A. and Vliegenthart, J.F.G. (1994) Carbohydr. Res. 264, 4 5 4 2 . [51 ] Leeflang, B.R., Bouwstra, J.B., Kerekgyirto, J., Kamerling, J.P. and Vliegenthart, J.F.G. (1990) Carbohydr. Res. 208, 117-126. [52] Lommerse, J.P.M., van Rooyen, J.J.M., Kroon-Batenburg, L.M.J., Kamerling, J.P. and Vliegenthart, J.F.G. (1997) (manuscript in preparation). [53] Leeflang, B.R. and Kroon-Batenburg, L.M.J. (1992) J. Biol. NMR 2, 495-518. [54] Lommerse, J.P.M., Kroon-Batenburg, L.M.J., Kroon, J., Kamerling, J.P. and Vliegenthart, J.F.G. (1995) J. Biol. NMR 5, 79-94. [SS] Lommerse, J.P.M., Kroon-Batenburg, L.M.J., Kamerling, J.P. and Vliegenthart, J.F.G. (1995) Biochemistry 34, 8196-8206. [56] Sminia, T. (1972) Z. Zellforsch. 130, 497-526. [57] Mulder, H., Schachter, H., de Jong-Brink, M., van der Ven, J.G.M., Kamerling, J.P. and Vliegenthart, J.F.G. (1991) Eur. J. Biochem. 201, 459465. [58] Mulder, H., Spronk, B.A., Schachter, H., Neeleman, A.P., van den Eijnden, D.H., de Jong-Brink, M., Kamerling, J.P. and Vliegenthart, J.F.G. (1995) Eur. J. Biochem. 227, 175-185. [59] Mulder, H., Dideberg, F., Schachter, H., Spronk, B.A., de Jong-Brink, M., Kamerling, J.P. and Vliegenthart, J.F.G. (1995) Eur. J. Biochem. 232, 272-283. [60] Mulder, H., Schachter, H., Thomas, J.R., Halkes, K.M., Kamerling, J.P. and Vliegenthart, J.F.G. (1996) Glycoconjugate J. 13, 1-7. [61] Johnson, K.D. and Chrispeels, M.J. (1987) Plant Physiol. 84, 1301-1308. [62] Tezuka, K., Hayashi, M., Ishihara, H., Akazawa, T. and Takahashi, N. (1992) Eur. J. Biochem. 203, 401413. [63] Smith, P.L. and Baenziger, J.U. (1992) Proc. Natl. Acad. Sci. USA 89, 329-333. [64] Dharmesh, S.M., Skelton, T.P. and Baenziger, J.U. (1993) J. Biol. Chem. 268, 17096-17102. [65] Schachter, H., McGuire, E.J. and Roseman, S. (1971) J. Biol. Chem. 246, 5321-5328. [66] Joziasse, D.H., Damen, H.C.M., de Jong-Brink, M., Edzes, H.T. and van den Eijnden, D.H. (1987) FEBS Lett. 221, 139-144. [67] Chaplin, M.F., Corfield, G.C. and Wood, E.J. (1983) Comp. Biochem. Physiol. 75B, 331-334.

J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins I1 0 1997 Elsevier Science B.V. All rights reserved CHAPTER 7

Fish glycoproteins Sadako Inoue and Yasuo Inoue Institute of Biological Chemistry, Academia Sinica. Nankang, Taipei I 1 5, Taiwan

List of abbreviations used CD CMP-Neu5Ac GalNAcol H-PSGP highSia-PSGP

circular dichroism cytidine 5'-monophospho Neu5Ac N-acetyl-o-galactosaminitol high molecular mass PSGP isolated from unfertilized eggs PSGP of high sialic acid content produced at later stages of oogenesis with glycan chains containing a series of oligo/poly-Sia groups, ranging from DP = 2 to >20 ((DP), = 6) H.46 polyclonal antibody specific to poly(Neu5Ac) 3-deoxy-o-glycero-o-galacto-nononic acid Kdn Kdn-rich glycoprotein isolated from vitelline envelope andor female ovarian fluid of Kdn-gp rainbow trout low molecular mass PSGP isolated from fertilized eggs L-PSGP a precursor of PSGP (i.e. PSGP of low sialic acid content formed at early stages of IowSia-PSGP oogenesis, each glycan chain contains mostly a disialyl group) NeuSAc N-acetylneuraminic acid Neu5Gc N-glycolylneuraminic acid NeuSAcyl N-acylneuraminic acid oligo/poly-Sia a-2,l-linked oligomers or polymers of sialic acid poly-Sia a-2,l-linked polysialic acid poly(Neu5Ac) a-2,l-linked polymers of NeuSAc poly(Neu5Gc) a-2,l-linked polymers of Neu5Gc poly(NeuSAc,NeuSGc) a-2,8-linked random copolymers of Neu5Ac and NeuSGc peptide:N-glycanase or peptide:N4-[N-acetyl-~-o-glucosarninyl] asparagine amidase PNGase (EC 3.5.1.52) polysialoglycoproteins or glycoproteins having polysialyl groups PSGP specific proteinase which is responsible for depolymerization of glycopolyprotein PSGPase (H-PSGP) into glycopeptide (L-PSGP) Sia sialic acid(s) mucin-type sialoglycoproteins isolated from vitelline envelope of cherry salmon Sia-gp containing Neu5Gc instead of Kdn ST sialyltransferase activity CMP-Sia:a-N-acetylgalactosaminide [R-Gal(@1-3)GalNAc(a 1-0-SerlThr] a-2,6-ST a-2,6-sialyltransferase a-2,l-ST CMP-Sia:a-2,6-sialoside Sia(a2-6)[R-Gal(@I-3)]GalNAc(a I-0-SerlThr] a-2,l-sialyltransferase a-2,l-polyST CMP-Sia:a-2,8(6)-sialoside a-2,l-sialyltransferase

1. Introduction The topics treated in this chapter do not cover all the glycoproteins and related substances, including enzymes, so far isolated from fish but are restricted to new findings obtained 143

144

through our 20 year studies on teleost eggs. We have been studying fish egg glycoproteins not to find something special to fish but in the belief that what we find in fish may also occur in higher vertebrates including humans. Fish eggs are obtainable with relative ease and in large amounts in Japan and we found them to be a rich source of glycoproteins.

2. Cortical alveolus glycoproteins (hyosophorins) Cortical vesicles are specialized Golgi-derived secretory organelles found in the peripheral cytoplasm of mature eggs of almost all animal species including humans [I]. Upon fertilization or parthenogenetic activation of the egg, the cortical vesicles fuse with the plasma membrane of the egg and release their contents into the perivitelline space; this process is a prerequisite for normal development of the embryos under natural conditions. These vesicles are about ten times larger ( 2 4 0 y m ) in fish eggs, than those in other animal eggs and are often called cortical alveoli rather than cortical granules, a term used for homologous vesicles found in sea urchin and mammalian eggs. Our studies revealed that each species of fish had homologous cortical alveolar glycoproteins which we named “hyosophorins” after hyosopho (Japanese), meaning cortical alveolus [2,3]. Hyosophorins are a family of glycoproteins ubiquitously found in the cortical alveoli of fish eggs and which obey the following criteria: (i) high carbohydrate content (8&90%, w/w); (ii) apo-hyosophorin comprises tandem repetitions of an identical peptide sequence; (iii) the protein is completely cleaved into repeating units when the dormant unfertilized eggs commence development in response to sperm fusion or a parthenogenetic stimulus.

2.1. First isolation of polysialoglycoproteins (PSGP) from rainbow trout eggs and their ubiquitous occurrence in salmonid fish In 1978, we isolated a new family of highly acidic 200 kDa sialoglycoproteins, which we later designated as polysialoglycoprotein (PSGP), from mature unfertilized eggs of rainbow trout [4,5]. The most striking feature of these glycoproteins was their high content (> 50%, by weight) of sialic acid that has been identified as Neu5Gc. All of the glycan chains were 0-glycosidically linked; the major carbohydrate components were Gal and GalNAc with smaller amounts of Fuc. Sialic acid usually occurs as a single non-reducing terminal residue in glycoproteins and in 1978, no poly-Sia structures were known to occur in glycoproteins [6]. The presence of a-2,g-linked oligo(Neu5Gc) in rainbow trout PSGP was established by methylation analysis coupled with glc-ms [7-91. The presence of polySia units with variable chain lengths up to 20 residues was shown by anion-exchange separation of the oligo/poly-sialylglycan units and the structural analysis of each chain [ 10,l I]. PSGP similar to that isolated from rainbow trout eggs was ubiquitously found in the eggs of other Sulmonidue fish species. We have examined 8 fish species from 3 of the 4 major genera of Salmonidae: Oncorhynchus (0. keta, 0. masou ishikawui, 0. mykiss, and 0. nerka udonis), Salmo (Salmo trutta furio), and Suluelinus (S.fontinalis, S. leucomaenispluuius, and S. namuycush) [12-141. All of the salmonid egg PSGP have 5 common oligosaccharide core glycan structures (a-e in Fig. l), of which 2 different

R( -6)

\

GalNAc(al4)SerRhr

/ Gal@1-3) R( -4

\

GalNAc(a I - O ) S e r m

/

Gal@14)Gal(P 1-3)

R( -6)

\

GalNAc(a l-O)Ser/Thr

/ GalNAc(P1-3)Gal(P 1-4)Gal( pl-3) R( -6)

\

GalNAc(a 1 - O ) S e r m

/ Fuc(al-t3)GalNAc(~l-3)Gal(p1-4)Gal(P 1-3)

GalNAc(Pl-4)

Kdn(a2-

\ GalNAc(al-O)SerRhr /

\ GalNAc(pl-3)Gal(pl-4)Gal(pl-3) / ) or NeuSAcyl(a2-3)

R = Kdn(a2-8)[NeuSAcyl(a2-

)I,

Fig. 1 . Structures of the five distinct types of carbohydrate units of salmonid fish egg PSGP Neu5Acyl indicates Neu5Ac and/or Neu5Gc. R = Kdn(a2-[8NeuSAcyl(a2-)1,.

kinds of core structures (d and e) are the biosynthetically matured forms [ 15-18]. The relative abundance of each core type varies depending on the individual rather than on the species of fish. 2.2. Occurrence of a deaminoneuraminic acid residue (Kdn) at the non-reducing end of oligo/poly-Sia chains

In 1986, we reported the first natural occurrence of a deaminoneuraminic acid, 3-deoxyD-glycero-D-galacto-nononicacid (Kdn), in rainbow trout egg PSGP [ 191. Our subsequent studies on the occurrence and biosynthesis of this residue have established that Kdn is a member of the sialic acid family [2&22]. Kdn has been found in all the PSGPs so far isolated from the eggs of various fish species. In fish egg PSGP, Kdn occurs only at the non-reducing terminal position of oligo/poly-Sia chains. Kdn is also found directly

146

linked to the proximal and penultimate GalNAc residues at the positions where Neu5Acyl residues are usually found (see Fig. 1). The Kdn residues are resistant to the action of the commercially available sialidases so far examined.

2.3. Diversity in oligo/poly-Sia chains of salmonid egg PSGP A species-specific structural diversity has been revealed in oligo/poly-Sia chains of salmonid egg PSGPs. For example, the PSGPs from Oncorhynchus (0.mykiss, 0. keta, 0. masou ishikawai, and 0. nerka adonis) contain exclusively NeuSGc, whereas those from Salmo and Salvelinus species contain both Neu5Ac and Neu5Gc. In these species poly-Sia chains can be either poly(NeuSAc), poly(Neu5Gc) or a hybrid type, poiy(NeuSAc,NeuSGc) [ 18,241. While O-acetyl substitution was not significant in most species, it was extensive in some species and occurred at C-4, C-7, and C-9 [25]. O-lactyl substitution also occurred in one species [24]. In species in which O-acetylation of NeuSAcyl was extensive, O-acetylation was also found at C-9 of Kdn [25]. Only the a2%linkage has been identified in poly-Sia chains of fish egg PSGP, although other linkage types have been reported for oligo/poly-Sia chains of bacteria and lower animals [26,27]. 2.4. Fish egg PSGP is a cortical alveolar component Because of its unusually low antigenicity, no antibody against fish egg PSGP has been obtained. The lake trout (Salvelinus fontinalis) PSGP contained poly(Neu5Ac) chains and was reactive with H.46 antiserum raised against colominic acid [28]. By the indirect immunofluorescence staining method using H.46 and cortical alveoli isolated from lake trout eggs, we showed that PSGP was localized in the cortical alveoli [29]. Furthermore, a cortical alveolus-rich fraction was separated from the rainbow trout egg and 200kDa PSGP was shown to be enriched in this fraction as a water-soluble component. 2.5. Polyprotein nature of apo-PSGP and the molecular mechanism of fertilizationassociated depolymerization

We found that PSGP isolated from fertilized rainbow trout eggs had a molecular mass of 9 kDa (L-PSGP), whereas the average molecular mass of PSGP isolated from unfertilized eggs was 200kDa (H-PSGP). Both the amino acid and carbohydrate compositions of H- and L-PSGP were identical. Apo-L-PSGP was a single tridecapeptide which was sequenced as DDAT*S*EAAT*GPSG (* indicates the glycosylated residue) [30]. Next, we found that H-PSGP was made up of a tandem repeat of the L-PSGP glycotridecapeptide [3 I]. No amino acid or peptide was inserted between the glycotridecapeptide repeats. It is apparent that fertilization-associated depolymerization occurs by proteolytic cleavage of the glycoprotein into glycotridecapeptide repeating units. Insemination was not necessary for this depolymerization. In salmonids, egg activation was induced by immersing the eggs in water without insemination. The depolymerization of H-PSGP following egg activation was catalyzed by a specific proteinase (PSGPase) present in the cortical region of the eggs, which becomes activated following egg activation. The enzyme

147

Lake trout (Salvelinus namaycush) L-PSGP(Sn) Asp-Ala-Thr*-Ser*-G1u-Ala-Ala-Thr*-Gly-Pro-Ser-Asp Asp-Asp-Ala-Thr*-Ser*-Glu-Ala-Ala-Thr*-Gly-Pro-Ser-Ser

Brook trout ( 5 . fontinalis) L-PSGP(Sf) Asp-Ala-Thr*-Ser*-Glu-Ala-Ala-Thr*-Gly-Pro-Ser-Asp Asp-Asp-Ala-Thr*-Ser*-Glu-Ala-Ala-Thr*-Gly-Pro-Ser-Ser

Japanese common char ( S . leucomaenis pluvius) L-PSGP (Slp) Asp-Ala-Thr*-Ser*-Glu-Ala-Ala-Thr*-Gly-Pro-Ser-Asp

Brown trout (Salmo trutta falio) L-PSGP(Stf) Asp-Ala-Thr*-Ser*-Glu-Ala-Ala-Thr*-Gly-Pro-Ser-Asp Asp-Asp-Ala-Thr*-Ser*-Glu-Ala-Ala-Thr*-Gly-Pro-Ser-Ser

Chum salmon (Oncorhynchus keta) L-PSGP(0k) Asp-Asp-Ala-Thr*-Ser*-Glu-Ala-Ala-Thr*-Gly-Pro-Ser-Ser

Land-locked cherry salmon (0. masou ishikawai) L-PSGP(0mi) Asp-Asp-Ala-Thr*-Ser*-Glu-Ala-Ala-Thr*-Gly-Pro-Ser-Ser

Rainbow trout (0. mykiss) L-PSGP(0m) Asp-Asp-Ala-Thr*-Ser*-Glu-Ala-Ala-Thr*-Gly-Pro-Ser-Gly

Kokanee salmon (0. nerka adonis) L-PSGP(0n) Asp-Ala-Thr*-Ser*-Asp-Ala-Ala-Thr*-Gly-Pro-Ser-Asp Asp-Asp-Ala-Thr*-Ser*-Asp-Ala-Ala-Thr*-Gly-Pro-Ser-Gly

Fig. 2. Amino acid sequences of L-PSGPs isolated from fertilized eggs of salmonid fish species where * indicates the 0-glycosylation sites. H-PSGPs isolated from unfertilized eggs are polymerized forms of L-PSGP

is active only at NaCl concentrations below 40 mM and the optimal temperature is about 16°C [32]. In natural fertilization, salmonid eggs are spawned and inseminated in fresh water so that the activation of the eggs occurs immediately after insemination. Under these conditions, water flows into the perivitelline space into which PSGP and PSGPase are translocated by exocytosis from the cortical alveoli. The estimated salt concentration within the perivitelline space is 10-20mM. We compared the structures of unfertilized and fertilized egg apo-PSGP from 8 Salmonidae species [24,30,3 1,33,34] (Fig. 2). The amino acid sequence of PSGP was

I48

highly homologous among the species. The occurrence of tandem repeats of dodeca- or tridecapeptide was found in apo-H-PSGP from all species. Sequence analysis of L-PSGP from each species (Fig. 2) revealed that proteolytic cleavages occur at the position two residues C-terminal to the Pro residue, -P-S-X-D (X is either G, S, or D). 2.6. Molecular cloning of apo-PSGP The PSGP molecule is initially synthesized as pre-pro-apo-PSGP and then posttranslationally modified in the Golgi apparatus by proteolytic cleavages of N- and C-termini, and by core glycosylations. cDNAs encoding apo-PSGP of rainbow trout eggs have been cloned [35]. Nucleotide sequence analysis showed that apo-PSGP mRNA contains tandem repeats, each composed of a 39-base oligoribonucleotide encoding a tridecapeptide, and that the sequences of the repeating units are completely preserved at the nucleotide level. Multiple mRNA species are present that are transcribed from multiple genes for apo-PSGP with different numbers of the repetitive sequence [36]. Expression of mRNAs is stage-specific: they are expressed only in immature ovaries but not in the mature eggs, nor in any other organs. 2.7. Biosynthesis of polysialyl glycan chains In a recent study [37], we showed developmental changes in the level and composition of PSGP during oogenesis of rainbow trout. A sialoglycoprotein family designated IowSiaPSGP was identified in ovaries 6 months prior to ovulation when the ovary weight was as low as 1.2 mg. This stage coincided with the appearance of detectable mRNA for apoPSGP. A second more highly sialylated glycoprotein (highSia-PSGP) appeared 3 months later (3 months prior to ovulation). Compositional analyses showed that the two forms of apo-PSGP differed only in their sialic acid contents while both amino acid and other carbohydrate compositions were identical. Structural studies confirmed that lowSia-PSGP contained mostly disialyl [Sia(a2-8)Sia(a2-6)-] chains whereas highSia-PSGP contained a-2J-linked oligo/poly-Sia chains ranging in length from 2 to 20 sialic acid residues. Biosynthetic studies using CMP-[ 14C]NeuSAc indicated that there are three sialyltransferase activities responsible for the assembly of the polysialylglycan chains of PSGP: (i) a-N-acetylgalactosaminide a-2,6-sialyltransferase (a-2,6-ST), which catalyzes the formation of the Sia residues a-2,6-linked to the proximal GalNAc residues in asialo-PSGP; (ii) a-2,6-sialoside a-2,8-sialyltransferase (a-2,8-ST), which catalyzes transfer of the first a-2,8-Sia residue to the a-2,6-linked Sia residue; and (iii) an a-2,8-polysialyltransferase (a-2,8-polyST), responsible for the synthesis of the a-2J-linked oligo/poly-Sia chains in highSia-PSGP (Fig. 3). These sialyltransferases were located in the cortical alveolus fraction obtained from the immature oocytes and mature eggs of rainbow trout. It was concluded that the core region of the glycan units of PSGP is assembled in the REWGolgi apparatus of oocytes giving rise to asialo-PSGPs, which are then transported from the Golgi apparatus into the cortical alveoli where sialylation occurs. Thus, asialo-PSGP molecules are packaged in immature cortical alveoli along with the sialyltransferases. The fact that both the Golgi complex and the endoplasmic reticulum play important roles in the formation of cortical

149

GalNAc(al-O)Ser/lk

+ CMP-Sia'

\GalNAc(al-O)SerlIhr

+ CMP-Sia*

a-2*wT*

Sia'(aZ-6)

Sia*(aZ-8)Sia(aZ-6)

/ R(4)Gd@1-3)

c R(-%4!31-3)

\ GalNAc(al-O)Ser/lh /

(ii)

Fig. 3. A summary of the results from studies on the biosynthesis of the polysialyl chains in salmonid fish egg PSGP. Three sialyltransferase activities, which are responsible for assembly of the polysialylglycan chains of PSGP, were identified: (i) a-N-acetylgalactosaminide a-2,6-sialyltransferase (a-2,6-ST), which catalyzes formation of the Sia residues a-2,6-linked to the proximal GalNAc residues in asialo-PSGP; (ii) a-2,6-sialoside a-2,8-sialyltransferase (a-2,8-ST), which catalyzes transfer of the first a-2,8-Sia residue to the a-2,6-linked Sia residue; (iii) an a-2,8-polysialyltransferase(a-2,8-polyST), responsible for synthesis of the a-2,8-linked oligo/poly-Sia chains in high Sia-PSGP. R = H, Gal(BIL, GalNAc(B1-3)Gal(f3-, Fuc(a1Sia' represents the 3)GalNAc(P 1-3)Gal(PI-, or Kdn(a2/NeuSGc(a2-3)[GalNAc(~1-4)]GalNAc(~1-3)Gal(~l-. sialic acid residue(s) newly incorporated from CMP-Sia' by the sialyltransferase in question.

vesicles and that cortical vesicles grow or mature during oogenesis of invertebrates and vertebrates have been documented by earlier morphological studies [38,39]. Our study first showed that these biosynthetic reactions occurred continuously in the cortical vesicles until ovulation. Both the a-2,6-ST and the a-2,8-ST can use CMP-NeuSAc and CMP-NeuSGc as activated sialyl donors. This is compatible with our finding that poly(NeuSAc), poly(Neu5Ac,NeuSGc), and poly (NeuSGc) chains were expressed in PSGPs from Saluelinus species such as lake trout and a Japanese common char [ 18,241. We have recently demonstrated the enzyme activity that catalyzes the transfer of Kdn from CMP-Kdn to PSGP in the cortical alveolus fraction of the rainbow trout ovary. Analyses of the reaction product showed that this Kdn-transferase could use the nonreducing terminal NeuSAcyl of poly-Sia chains of rainbow trout egg PSGP as an acceptor substrate but could not transfer either Kdn or NeuSAc from the respective activated donor to the terminal Kdn residues on the poly-Sia chains [40]. These observations coincide with our previous view that Kdn occurs as the capping residue of the poly-Sia chains of rainbow trout egg PSGP [17]. 2.8. Hyosophorins bearing bulky multi-antennary N-glycan chains

Our studies on the distribution of PSGP among fish species revealed that fish belonging to species other than Salmonidae have major N-glycosidic glycoprotein components that satisfy the criteria of hyosophorin. The first N-glycan-type hyosophorin was isolated from the eggs of medaka fish, Olyzias latipes[2]. By using a polyclonal antibody against H-hyosophorin and an indirect immunofluorescence staining technique, H-hyosophorin was localized in the peripheral region of the cortical alveoli. L-hyosophorin is a

150

Flounder (Paralichthys olivaceus)

glycononapeptide: AspAla-Ala-Ser-Asn’-Glu-Th-Val-Ser, where * indicates the glycosylation site. It has to be noted that a single N-glycan chain is attached to every repeating unit, in contrast to other PSGPs where three 0-glycan chains are attached to the repeating unit. The complete structure of a large pentaantennary glycan unit (molecular weight -6000) was recently established by H-NMR spectroscopy and FAB-MS spectrometry [41] (Fig. 4). Each of the antennae has a poly-N-acetyllactosaminyl core highly branched at both GlcNAc and Gal residues. The poly-N-acetyllactosaminyl core was completely resistant to endo-(3-galactosidase. The presence of a (3-galactosylated Lewis x antigenic epitope Gal@1-4)Gal(~1~)[Fuc(al-3)]GlcNAc((31-), and its sialylated form, Gal((3I-4)[Neu5Ac(a2-3)]Gal((31~)[Fuc(al-3)]GlcNAc(~l-), the presence of Gal((31-

’

4)Gal(fk and G a l ( p 1 4 ) G a l ( ~ ) G a l ( ~ l -and , the presence of branched Gal residues, -

4)GlcNAc(~l-3)[Gal(~l4)]Gal(~1are novel and unique features. Prior to the elucidation of this structure, we determined the structure of a tetraantennary sialoglycan unit attached to hyosophorin of Indian medaka, Oryzias melastigma [42]. Interestingly, while the amino acid sequence of hyosophorin is completely conserved between these species, their glycan structures show marked differences. The structures of the peripheral portions that contain mostly digalactosyl (0.latipes) and trigalactosylated (0. melastigma) structures are different (Fig. 4). Moreover, a portion of 0. latipes hyosophorin is fucosylated while 0. melastigma hyosophorin is not. In contrast to these inter-species differences, the glycans of hyosophorins are homogeneous with respect to antenna and branching structures within a species. Thus, hyosophorin molecules are uniquely N-glycosylated in uiuo with no heterogeneity in antenna number while most of the animal glycoproteins show heterogeneity with respect to the antenna structures (di-, tri-, and tetraantennary) of their glycan chains. Despite the large differences in the structures of glycan chains, these N-glycan-type hyosophorins and PSGP are both highly acidic, poly-anionic polymers. We have isolated similar N-glycan-type hyosophorins from other species of freshwater fish, Plecoglossus altiuelis, Tribolodon hakonensis, and Cyprinus carpio [43]. Although the species-specific difference was extensive in the peripheral structures, all of them have a poly-N-acetyllactosaminyl core structure attached to penta- (Plecoglossus) or tetra- (Tribolodon and Cyprinus) antennary glycan units. The presence of sulfate groups in addition to sialic acid makes the structural elucidation difficult. It is noted that in all of these fish eggs, PSGP molecules similar to those isolated from Salmonidae are also expressed as a minor component (less than one twentieth of N-glycosidic-type). Hyosophorin isolated from a sea-water fish, flounder (Paralichthys oliuaceus), had neutral fucosylated pentaantennary glycan chains [3]. This non-anionic nature of flounder hyosophorin may account for the poor histochemical staining of the cortical alveoli of this fish, in contrast to intense toluidine blue staining of cortical alveoli of the eggs of salmonid fish, medaka fish, and Plecoglossus. The salinity of water in which the fish spawn and the ionic properties of hyosophorins are not related as we first anticipated, since our studies revealed that the hyosophorin glycans of some seawater spawning fish, herring (Clupea pallasii) (S. Inoue, unpublished observation) and Fundulus heteroclitus [831, are sialylated. 2.9. Biological function of hyosophorin and future perspective Formation and maturation of Golgi-derived secretory vesicles (cortical vesicles) by developing oocytes of many animal species, and the dramatic discharge of their contents at the moment of fertilization just prior to the initiation of embryonic development have for a long time attracted the attention and interest of many biologists. It is very interesting that polysaccharides or mucopolysaccharides have been implicated as a major component of cortical vesicles of most animals ranging from echinoderms to mammals in these earlier works. We have now isolated and determined the chemical structures of cortical alveolar glycoproteins ubiquitously found in fish eggs. Although our study is still far from complete, the following findings may be relevant to the function of hyosophorin.

152

2.9.1. Formation and function of the perivitelline fluid In teleost fish, the spawned eggs are bathed in a hypotonic (fresh water) or a hypertonic (marine) external medium. The embryo enclosed by the plasma membrane is bathed in the perivitelline fluid formed in the space between the plasma membrane and the outermost envelope (vitelline envelope or fertilization envelope). At fertilization, cortical vesicle exocytosis occurs, and the contents of the vesicles are transported into the perivitelline space. Due to the impermeability of the vitelline envelope to high molecular mass substances such as hyosophorins, they are confined to the perivitelline space and cause intake of water (and small ions) across the envelope from the external medium. The tension of the envelope resulting from the internal hydrostatic pressure created by the osmotic properties of the perivitelline space fluid is thought to produce a cushioned environment for the embryo against external deformation and a decrease in the permeability of the plasma membrane. Being poly-anions, hyosophorins of freshwater fish are, under physiological conditions, negatively charged, and may accumulate cations (Na+, K+, Ca2+, and Mg2+) in excess of their concentrations in the external medium. Osmotic and ionic regulatory functions of the perivitelline fluid are supported by physiological measurements [44]. At the present stage, we cannot answer the question why H-hyosophorins undergo depolymerization to L-hyosophorins at fertilization. For polyanions such as PSGP, an increase in terms of molecular concentration by 20 times does not result in an increase of the calculated osmotic pressure due to the large contribution from the Donnan term. Increase in diffusion coefficient by H- to L-transformation may be significant in allowing a response to rapid formation of the perivitelline fluid. In marine species the mechanism of osmoregulation between the egg and the external medium is more complex. However, the observed values of zero to positive perivitelline potentials obtained for eggs of marine species [44] may suggest the difference in the ionic properties of colloids (may be hyosophorins) in the perivitelline fluid. These results are in accord with our results that showed hyosophorin from a marine species (flounder) bore only neutral carbohydrate units. 2.9.2. Sperm agglutfnating properties of hyosophorins One of the important roles of mammalian cortical granules may be to block polyspermy in the perivitelline space [45]. In fish, fertilizing sperm enters through a special opening in the egg envelope called micropyle. The diameter of the micropyle in most fish is usually just large enough to allow a single sperm. Moreover as soon as the first sperm enters into the egg, an electron-microscopically observable plug is formed at the bottom of the micropyle to prevent the entrance of a second sperm. There may be no chance for sperm to be directly in contact with the perivitelline fluid and therefore L-hyosophorins. And indeed, we found that L-hyosophorins, the forms actually present in the perivitelline space, have only weak, if any, sperm agglutinating activity. Though H-hyosophorins agglutinate sperm, this is a rather general property shown by large poly-anions. On the basis of electron-microscopic observation, some workers have claimed that the cortical vesicle contents of fish eggs are partially translocated to the outside surface of the fertilization envelope and there they function to prevent polyspermy. Our careful analysis could not detect any PSGP outside the fertilization envelope or in the extract of the isolated

153

fertilization envelope of rainbow trout. L-PSGP was found in the perivitelline space fluid throughout embryonic development of rainbow trout until hatching. 2.9.3. Calcium ion binding properties of hyosophorins Recently, we investigated the Ca2+ binding of H- and L-PSGP by equilibrium dialysis and circular dichroism methods [46]. Calcium binding to H- and L-PSGP occurred with apparent binding constants (K,) of 2 . 9 8 ~ lo3, and 1 . 0 0 lo3 ~ M-I, respectively. These values were lower than the value (13.9 x 1O3 M-') for colominic acid (DP = 24) but much higher than the reported values of 121 and 193 M-' for Ca2+-Neu5Ac [47] and Ca2+NeuSGc [48], respectively. Ca2+ binding affinity for L-PSGP was one-third of that for H-PSGP, whereas the number of binding sites (n) did not change on going from H-PSGP to L-PSGP. CD measurements also indicated that H-PSGP was more strongly and preferentially complexed with Ca2+ion than L-PSGP. The removal of oligo/poly(Neu5Gc) from PSGPs by sialidase treatment abolished the Ca2+ binding capacities as examined both by equilibrium dialysis and CD methods. The difference of Ca2+binding affinities between H- and L-PSGP has implications for the calcium release at fertilization, i.e. the Ca2+ions bound to H-PSGP represent a reservoir of calcium for liberation at fertilization. The calcium binding capacity of each sialic acid residue of H- and L-PSGP was larger than that of colominic acid. On average, one calcium ion is bound to two NeuSGc residues of PSGP and three NeuSAc residues of colominic acid. Although no information on the Ca2+ concentration in a cortical alveolus is available and it may be difficult to determine it directly in such a fragile vesicle as the cortical alveolus, now we can estimate the concentration to be 93mM on the basis of the average values of egg and cortical alveolus diameters and the amount of H-PSGP present in an egg with a value of Neu5Gc/Ca2+= 2: 1. The derived value of the Ca2+ concentration in cortical alveoli of rainbow trout compared well with those (30 and 95 mM) for two different species of the sea urchin determined by X-ray micro-analysis [49]. Using the estimated volume of 6.7 x ml, the Ca2+ and NeuSGc concentration in the perivitelline fluid of rainbow trout egg was 4.7 and 9.4mM, respectively. One of the morphological changes after fertilization is the deformation from a soft vitelline envelope to a tough fertilization envelope. It is known that this process called hardening is enzyme-catalyzed and requires Ca2+ions. We have observed that at least 1OmM Ca2+is necessary for hardening of the fertilization envelope of rainbow trout. Thus, Ca2+ions transported from the egg into the perivitelline space with H-PSGP, and there released by depolymerization of H-PSGP into L-PSGP may be used in this process. 2.9.4. De-N-glycosylution of hyosophorins We found that N-glycan chains apparently detached from hyosophorins become accumulated in the embryos of two different species of fish, flounder and medaka [3,50]. In both species hyosophorin-derived free glycans were found neither in unfertilized eggs nor in eggs which were fertilized but which failed to continue embryonic development. We have identified alkaline peptide:N-glycanase (PNGase) (optimal pH 7-9) in the blastodiscs isolated from the blastula stage of medaka embryos (A. Seko, in preparation). The expression of alkaline PNGase activity and the accumulation of hyosophorin-derived free

154

glycan increased in parallel and both reached the maximum level at stage 12. These results strongly suggest the physiological importance of de-N-glycosylation of L-hyosophorins for continuation of embryonic development.

3. Mucin-type glycoproteins found in the vitelline envelope and ovarian fluid of salmonidfish 3.1. Isolation and glycan structures of Kdn-gp and Sia-gp The morphology, chemical nature and nomenclature of the egg coats that enclose ovulated eggs vary in different animal groups and largely reflect the mode by which the egg achieves successful fertilization and embryonic development under a variable environment [51]. The following functions have been attributed to the egg coats: (i) to protect the eggs and assist their movement during ovulation and spawning; (ii) to assist species-specific fertilization; (iii) to prevent polyspermy; (iv) to become a chemically and mechanically tough covering for the embryo; and in some species (v) to fix the embryo to the substratum. The second function includes species-specific egg-sperm recognition and induction of the acrosome reaction for some animal groups. Glycoproteins are reported to be major components of the egg coat in many animal groups and their carbohydrate chains are believed to be functionally important. In fish, the egg coats consist of several layers called the vitelline envelope (before fertilization) which, upon fertilization, is transformed into the fertilization envelope. This transformation involves enzyme-catalyzed reformation of various macromolecule(s). An unusual family of glycoproteins containing -50% (w/w) Kdn and no N-acylneuraminic acid was first isolated by us in 1988 from the vitelline envelope of rainbow trout eggs and designated as Kdn-gp [52]. Kdn-gp contains -15% protein and -85% carbohydrate (Gal/GalNAc/Kdn = 1:2:av.5) which is 0-linked to ThdSer. Apo-Kdn-gp is also unusual in that threonine and alanine account for 40 and 27 residue %, respectively. More than 80% of the threonine residues are involved in a carbohydrate linkage. Kdn-gp is isolated as a family of polydisperse molecules ranging over 700-4000kDa (the peak at -3000 kDa) as estimated by gel filtration chromatography. The structure of oligoKdn-containing 0-glycan chains of Kdn-gp is given in Fig. 5. Only one type of core structure was found in Kdn-gp 0-glycans, though diversity was found in the size of the oligoKdn chains ( n= 1-7) [53]. More recently, Kdn-containing di- and triantennary N-glycan chains were identified in Kdn-gp isolated from rainbow trout vitelline envelope[54] (see Fig. 5). Based on composition and the yield of the carbohydrate chains, about 1000 0-glycans and 60 Nglycans are linked to a 3000 kDa Kdn-gp. While Kdn-gp is a major glycoprotein component of the vitelline envelope of rainbow trout, chum salmon, and kokanee salmon, the vitelline envelope of cherry salmon contains little Kdn-gp. An analogous family of mucin-type glycoproteins (Sia-gp) that contain Neu5Gc instead of Kdn is the major glycoprotein component in this salmon (Fig. 5) [55]. That Sia-gp is a functionally analogous molecule to Kdn-gp is supported by the similar molecular mass and amino acid composition (Thr + Ala M 70 residue %).

155

(a) 0- and N-linked Kdn-glycan chains present in Kdn-gp Kdn(a2-8)Kdn(a2- ) -( -8)Kdn(a2-6)

\

GalNAc(a 1-0)SerRhr

/ Kdn(a2-3)Gal(~l-3)GdNAc(al-3)

Kdn(a2-3)Gal(P 14)GlcNAc(P 1-2)Man(a 1-6)

\

*GlcNAc(Pl4)-Man(P

rtFuc(a 1-6)

I

14)GlcNAc(~l-I)GlcNAc(P1-N)Asn

I

Kdn(a2-3)Gal(pl4)GlcNAc(~l-2)Man(al-3)

Kdn(a2-3)Gal( pl4)GlcNAc(P 1-6), Man(al-6)

\

Kdn(a2-3)Gal(P 1-4)GlcNAc(fi1-2)’

@uc(a 1 4 ) I rtGlcNAc(P 14)-Man(Pl4)GlcNAc(P 14)GlcNAc(P 1-N)Asn

f

I

Kdn(a2-3)Gal(P 14)GlcNAc(P 1-2)Man( a 1-3)

(b) 0-linked NeuGc-glycan chains present in Sia-gp Kdn(a2-6)

\

GalNAc(a 1-0)SerRhr

/ NeuGc(a2-3)Gal(pl-3)GalNAc(al-3) Fig. 5. Structures of (a) Kdn-containing 0- and N-linked glycan chains; (b) Neu5Gc-containing 0-glycan chains of Kdn-gp and Sia-gp isolated from the vitelline envelope of salmonid fish species.

Immunohistochemical methods revealed that Kdn-gp is localized in the second layer of the outer surface of the vitelline envelope [56]. We have found that Kdn-gp was also a component of ovarian fluid of this and other salmonid fish (Kdn-gp-OF) [57].In salmonids, eggs are first ovulated in the body cavity of female fish, and wait for the stimulus for spawning while bathed in ovarian fluid (or body cavity fluid). The ovarian fluid is believed to be originated from the extraovarian cells in the ovary. No difference was found in the chemical nature (including the glycan structure) between Kdn-gp-VE and Kdn-gp-OF. The amount of Kdn-gps isolated from a single 2-year-old female rainbow trout (spawning 100g egg) was 10mg of Kdn-gp-VE and

-

-

156

10-20mg of Kdn-gp-OF. Sia-gp-OF was isolated from those fish species which synthesize Sia-gp-VE. 3.2. Biosynthesis and possible functions of Kdn-(Sia-)gps Kdn-gp appeared to be synthesized during relatively later stages of oogenesis, since it was undetectable in the oocyte 3 months prior to ovulation. Although the cell types that synthesize Kdn-gp have not been identified, it is most likely synthesized under hormonal control in some extraoocyte cells (i.e. follicle cells), secreted and partly incorporated into the second outermost layer of the vitelline envelope just before ovulation. Kdn-gp may thus be a molecule homologous to the oviduct glycoproteins of mammals that are reported to be secreted and partly incorporated into the egg surface [5X-601. Recently, we showed a strong Ca2+ ion binding property for Kdn-gp [46]. Being polyanionic, Kdn-gp may control cation and H+ concentrations in the ovarian fluid in which the ovulated eggs bath. The presence of clusters of anionic carbohydrate chains make Kdn-gp resistant to non-specific proteases such as actinase E. Kdn-gp, like other cell surface mucins, may function in facilitating the movement of cells and in protecting cells from proteolytic degradation and from bacterial invasion.

4. Glycoproteins related to vitellogenesis It is well established that in oviparous vertebrates, vitellogenin is synthesized in the liver under hormonal control, transported into the blood stream, and sequestered via a receptor into the growing oocyte [61]. The vitellogenin polypeptide (-200 kDa) was reported to be glycosylated and phosphorylated at the site of its synthesis. Vitellogenin molecules found in female blood also contain a large percentage of lipid. As soon as vitellogenin is incorporated into the oocyte, it is cleaved into lipovitellins and phosvitins and has not been isolated in its uncleaved state from the oocyte. Phosvitins are highly phosphorylated, relatively small proteins (2-30 m a ) and are often glycosylated. The vitellogenin receptor has been partly characterized in the chicken oocyte. The importance of N-glycosylation of vitellogenin on its uptake by the oocyte has been implicated in frog [62]. The system by which fish vitellogenin is sequestered and processed into yolk is poorly understood. Involvement of the lysosomal system in this process has been shown by electron-microscopic and cytochemical studies [63]. Recently, we found accumulation of relatively large amounts of complex-type free sialoglycans in the unfertilized mature eggs of two species of freshwater fish, Plecoglossus altivelis and Tribolodon hakonensis [64,65]. All of these free glycans possessed typical di-, tri-, or tetraantennary structures and the di-N-acetylchitobiose structure at their reducing termini. We originally speculated that these free glycans may be derived from vitellogenin, because of the large amount -5 x mol, and -25 x mol per g fresh egg of Plecoglossus and Tribolodon, respectively. These amounts are comparable to the amount of vitellogenin-derived proteins in the eggs. We also hypothesized that the enzyme involved may be peptide:N-glycanase, an enzyme not previously reported in an animal system. The speculation that free N-glycan chains in the unfertilized fish eggs may

157

originate from vitellogenin molecules was partly substantiated by our recent isolation of glycophosphoproteins having species-specific N-linked glycans previously found as free glycan chains [66]. In addition to the above fish species, the occurrence of free N-glycan chains and the phosphoprotein bearing the same N-linked unit as the free glycan was also demonstrated in medaka fish (Fig. 6). Most recently, we identified and partially purified a peptide:N-glycanase (PNGase) activity in the early embryos of medaka fish [67]. This was the first report of PNGase from an animal source. This PNGase from a medaka embryo had a low optimal pH of 3.7 (acid PNGase). Acid PNGase activity was also found in the ovary egg; spawned, fertilized eggs are usually used as the enzyme source because of the ease in collecting large amounts of material. Acid PNGase may participate in de-N-glycosylation of vitellogenin during vitellogenesis. N-glycosylation of vitellogenin may be needed in some process of its transportation, recognition by the receptor, or in specifying the site of proteolytic cleavages. De-N-glycosylation, in turn, may be necessary for recycling of the receptor and further processing and/or utilization of yolk proteins. This latter idea is supported by our finding that in the mature eggs of Tribolodon, in which the free N-glycan pool is large, non-glycosylated small phosphopeptides (molecular mass -2000) were found in excess of the amount of the glycophosphoprotein (S. Inoue, in preparation).

5. Epilogue Several new constituents of glycoproteins such as poly-Sia and Kdn have been discovered in fish egg glycoproteins. The occurrence of Kdn, a new member of the sialic acid family, has been reported in one bacterial strain [68], two amphibian species [69,70] and various mammalian tissues [7 1,84-861. Use of monoclonal antibodies [7 1,721 and specific enzymes [73-751 that cleave Kdn linkages may facilitate new discoveries not only in the area of fundamental biology but also in oncology and pathology. We have also described a new catabolic pathway for glycoproteins in animal cells. In eukaryotic cells, the biosynthetic mechanism of protein N-glycosylation has been well established. However, the necessity for removing specific N-glycan chain(s) from certain glycoproteins as a possible prerequisite for intracellular transport and/or for acquisition of a functional structure has yet to be demonstrated in all systems. We identified and partially purified peptide:N-glycanase (PNGase) in the fish egg. PNGase had been previously reported only in plants and bacteria [76-78]. Although the purified PNGases from plant and bacterial sources have been used as tools in structural and functional studies of glycoproteins, almost no attention has been paid to the physiological function of this enzyme in living cells. Subsequent to our first finding of PNGase in fish, we have demonstrated a wide occurrence of PNGase in animals, including mammalian species, and their physiological significance [87,88]. Our discovery of PNGase in fish eggs was based on the finding of an unusually large pool of free N-glycan chains identical to parts of the N-glycan units attached to the glycoproteins, including the di-N ,N'-acetylchitobiosyl structure at the reducing termini. If PNGase-catalyzed de-N-glycosylation occurs in other animal cells, the accumulation of free oligosaccharides would not be observed since they are transported to the lysosomal

Oryzius htipes NeuSAc(a2-3)Gal(P 1 +Gal@ 1-4)GlcNAc@ l-‘L)Man(a14)

\

Man($]-)R

/

Neu5Ac(a2-3)Gal(~14)Gal(P1-4)GlcNAc(pl-2)Man(a1-3)

Tribolodon hukonensis Sia(a2-3)Gal(P 14)GlcNAc(p 1-Z)Man(a1-6)

\

Man(pl-)R

/

Sia(a2-3)Gal(~l-4)GlcNAc(~l-2)Man(al-3) Sia(a2-3)[Gal(pl-4)]Gal(~l~)GlcNAc(pl-Z)Man(a1~/3)

\

Plecoglossus altivelis Neu5Ac(a2-3)Gal(~l~)GlcNAc(P1-Z)Man(a 1-6)

\ /Man(P1-)R Neu5Ac(a2-3)Gal(~l4)GlcNAc(~1-2)Man(al-3) Neu5Ac(&3)Gal(~14)GlcNAc(~I-Z)Man(a 14)

\

NeuSAc(a2-3)Gal(P 1-4)GlcNAc(P 14),

Man@-)R

/

Sia(a2-3)Gal(P 14)GlcNAc(p 1-2)Man(al-3/6)

Sia(a2-3)[Gal(~l-4)]Gal(~l4)GlcNAc(pl-2)Man(a1-6)

\ Sia(a2-3)[Gal(P14)]Gal(P14)GlcNAc(pl-2)Man(a1-3)

Man(pl-)R

/

Man(a1-3) Neu5Ac(a2-3)Gal(pl4)GlcNAc(pl-2)’

Neu5Ac(a2-3)Gal(p I4)GlcNAc(~I-6) ,, Man(a 14) \ Neu5Ac(a2-3)Gal(~l~)GlcNAc(~l-2)’ Man(Pl-)R Neu5Ac(aZ-3)Gal(~l-4)GlcNAc(~l-Z)Man(al-3)’

R = 4GlcNAcpl+ 4GlcNAc Sia = NeuSAc or Neu4,5Ac2

Fig. 6.Structures of free sialoglycans accumulated in the unfertilized eggs of Otyzius lofipes, Plecoglossus ulfiuelis, and Tribolodon hukonensis and N-linked glycan structures determined for glycophosphoproteins isolated from the corresponding fish species.

159

system immediately after liberation from the parent glycoproteins and undergo further degradation. As anticipated in our previous report [79], we have identified PNGase activities in mammalian-derived cells and tissues [80,81] and purified the enzyme from CH3 mouse-derived L-929 fibroblast cells [82]. These results indicate that PNGase may participate in important biological processes. Fertilization and early embryonic development may be an area in which the participation of specific carbohydrate units of glycoproteins in recognition and activation phenomena manifests itself. We believe the fish egg provides a good system for work aimed at solving the molecular mechanism of such phenomena.

References [l] [2] [3] [4] [5] [6] [7] [8] [9] [lo] [I I] [I21 [I31 [I41 [15] [16] [I71 [I81 [I91 [20] [21] [22] [23] [24] [25] [26] [27]

(281 [29] [30]

Guraya, S.S. (1982) Int. Rev. Cytol. 78, 257-360. Kitajima, K, Inoue, S. and Inoue, Y. (1989) Dev. Biol. 132, 544-533. Seko, A,, Kitajima, K., Iwasaki, M., Inoue, S. and Inoue, Y. (1989) J. Biol. Chem. 264, 15922-15929. Inoue, S. and Iwasaki, M. (1978) Biochem. Biophys. Res. Commun. 83, 1018-1023. Inoue, Y. (1993) In: J. Roth, U. Rutishauser and F.A. Troy (Eds.), Polysialic Acid. Birkhauser Verlag, Basel, pp. 171-181. Troy, F.A. (1992) Glycobiology 2, 5-23. Inoue, S. and Matsumura, G. (1979) Carbohydr. Res. 74, 361-368. Inoue, S. and Matsumura, G. (1980) FEBS Lett. 121, 33-36. Inoue, S., Matsumura, G. and Inoue, Y. (1982) Anal. Biochem. 125, 118-124. Inoue, S. and Iwasaki, M. (1980) Biochem. Biophys. Res. Commun. 93, 162-165. Nomoto, H., Iwasaki, M., Endo, T., Inoue, S., Inoue, Y. and Matsumura, G. (1984) Arch. Biochem. Biophys. 218, 335-341. Shimamura, M., Endo, T., Inoue, Y. and Inoue, S. (1983) Biochemistry 22, 959-963. Iwasaki, M., Inoue, S. and Matsumura, G. (1983) Proc. 7th Int. Symp. Glycoconjugates. pp. 205-206. Iwasaki, M., Inoue, S., Tazawa, 1. and Inoue, Y. (1985) Proc. 8th Int. Symp. Glycoconjugates. pp. 51-52. Iwasaki, M., Inoue, S., Kitajima, K., Nomoto, H. and Inoue, Y. (1984) Biochemistry 23, 305-310. Kitajima, K., Nomoto, H., Inoue, Y., Iwasaki, M. and Inoue, S. (1984) Biochemistry 23, 310-316. Iwasaki, M., Nomoto, H., Kitajima, K., Inoue, S. and Inoue, Y. (1984) Biochem. Int. 8, 573-579. Iwasaki, M. and Inoue, S. (1985) Glycoconjugate J. 2, 209-228. Nadano, D., Iwasaki, M., Endo, S., Kitajima, K., Inoue, S. and Inoue, Y. (1986) J. Biol. Chem. 261, 11550-1 1557. Iwasaki, M., Inoue, S., Nadano, D. and Inoue, Y. (1987) Biochemistry 26, 1452-1457. Yu Song, Kitajima, K., Inoue, S. and Inoue, Y. (1991) J. Biol. Chem. 266, 21929-21935. Yu Song, Kitajima, K., Inoue, S., Muto, Y., Kasama, T., Handa, S. and Inoue, Y. (1993) Biochemistry 32, 922 1-9229. Terada, T., Kitazume, S., Kitajima, K., Inoue, S., Ito, F., Troy, FA. and Inoue, Y. (1993) J. Biol. Chem. 268, 2640-2648. Sato, C., Kitajima, K., Tazawa, I., Inoue, Y., Inoue, S. and Troy, F.A. (1993) J. Biol. Chem. 268, 2367523684. Iwasaki, M., Inoue, S. and Troy, F.A. (1990) J. Biol. Chem. 265, 2596-2602. Battacharjee, A.K., Jennings, H.J., Kenny, C.P., Martin, A. and Smith, I.C.P. (1975) J. Biol. Chem. 250, 1926-1 93 2. Kitazume, S., Kitajima, K., Inoue, S., Troy, F.A., Cho, J.-W., Lennarz, W.J. and Inoue, Y. (1994) J. Biol. Chem. 269, 22712-22718. Kitajima, K., Inoue, S., Inoue, Y. and Troy, F.A. (1988) J. Biol. Chem. 263, 18269-18276. Inoue, S., Kitajima, K., Inoue, Y. and Kudo, S. (1987) Dev. Biol. 123, 442454. Inoue, S. and Inoue, Y. (1986) J. Biol. Chem. 261, 5256-5261.

160 [31] [32] [33] [34] [35]

Kitajima, K., Inoue, S. and Inoue, Y. (1986) J. Biol. Chem. 261, 5262-5269. Kitajima, K. and Inoue, S. (1988) Dev. Biol. 129, 270-274. Kitajima, K., Sorimachi, H., Inoue, S. and Inoue, Y.(1988) Biochemistry 27, 7141-7145. Yu Song, Kitajima, K. and Inoue, Y. (1990) Arch. Biochem. Biophys. 283, 167-172. Sorimachi, H., Emori, Y.,Kawasaki, H., Kitajima, K., Inoue, S., Suzuki, K. and Inoue, Y. (1988) J. Biol. Chem. 263, 17678-17684. [36] Sorimachi, H., Emori, Y.,Kawasaki, H., Suzuki, K. and Inoue, Y.(1990) J. Mol. Biol. 211, 35-48. [37] Kitazume, S., Kitajima, K., Inoue, S., Inoue, Y. and Troy, F.A. (1994) J. Biol. Chem. 269, 10330-10340. 1381 Anderson, E. (1968) J. Morphol. 125, 23-60. [39] Selman, K. and Anderson, E. (1975) J. Morphol. 147, 251-274. [40] Angata, T., Kitazume, S., Terada, T., Kitajima, K., lnoue, S., Troy, F.A. and Inoue, Y. (1994) Glycoconjugate J. 11, 493499. [41] Taguchi, T., Seko, A,, Kitajima, K., Muto, Y., Inoue, S., Khoo, K.-H., Morris, H.R., Dell, A. and Inoue, Y. (1994) J. Biol. Chem. 269, 8762-8771. [42] Taguchi, T., Seko, A,, Kitajima, K., Inoue, S., Iwamastu, T., Khoo K.-H., Morris, H.R., Dell, A. and lnoue, Y.(1993) J. Biol. Chem. 268, 2353-2362. [43] Iwasaki, M., Inoue, S., Kitajima, K., Ishii, K. and Inoue, Y. (1991) Glycoconjugate J. 8, 190-191. [44] Alderdice, D.F. (1988) In: W.S. Hoar and D.J. Randall (Eds.), Fish Physiology, Vol. 1 IA. Ch. 3, pp. 163251. [45] Gulyas, B.J. (1980) Int. Rev. Cytol. 63, 357-392. [46] Shimoda, Y., Kitajima, K., Inoue, S. and Inoue, Y. (1994) Biochemistry 33, 1202-1208. [47] Jaques, L.W., Brown, E.B., Barrett, J.M., Brey Jr, W.S., and Weltner Jr., W. (1977) J. Biol. Chem. 252, 4533-4539. [48] Jaques, L.W., Riesco, B.F. and Weltner Jr., W. (1980) Carbohydr. Res. 83, 21-32. [49] Gillot, I., Ciapa, B., Payan, P. and Sardet, C. (1991) Dev. Biol. 146, 396405. [50] Seko, A,, Kitajima, K., Inoue, S. and Inoue, Y. (1991) Biochem. Biophys. Res. Commun. 180, 11651171. [SI] Dumont, J.N. and Brummett, A.R. (1985) In: L.W. Browder (Ed.), Developmental Biology, A Comprehensive Synthesis, Vol. I . Plenum Press, New York, pp. 235-288. [52] Inoue, S., Kanamori, A,, Kitajima, K. and Inoue, Y. (1988) Biochem. Biophys. Res. Commun. 153, 172-176. [53] Knamori, A., Inoue, S., Iwasaki, M., Kitajima, K., Kawai, G., Yokoyama, S. and Inoue, Y. (1990) J. Biol. Chem. 265, 21811-21819. [54] Tezuka, T., Taguchi, T., Kanamori, A., Muto, Y., Kitajima, K., Inoue, Y. and Inoue, S. (1994) Biochemistry 33, 6495-6502. [55] Inoue, S., Iwasaki, M., Kanamori, A,, Terada, T., Kitajima, K. and Inoue, Y. (1991) Glycoconjugate J. 8, 223. [56] Kanamori, A,, Kitajima, K., Inoue, Y. and Inoue, S. (1991) Glycoconjugate J. 8, 222-223. [57] Kanamori, A., Kitajima, K., Inoue, S. and Inoue, Y. (1989) Biochem. Biophys. Res. Commun. 164, 744-749. [58] Kapur, R.P. and Johnson, L.V. (1985) Dev. Biol. 112, 89-93. [59] Oikawa, T., Sendai, Y.,Kurata, S.J. and Yanagimachi, R. (1988) Gamete Res. 19, 113-122. [60] Wegner, C.C. and Killian, G.J. (1991) Mol. Reproduct. Develop. 29, 77-84. [61] Wallace, R.A. (1985) In: L.W. Browder (Ed.), Developmental Biology, Vol. 1. Plenum Press, New York, pp. 127-177. [62] Lane, C.D., Champion, J., Colman, A,, James, T.C. and Applebaum, S.W. (1983) Eur. J. Biochem. 130, 529-53 5. [63] Busson-Mabillot, S. (1984) Biol. Cell 51, 53-66. [64] Ishii, K., Iwasaki, M., Inoue, S., Kenny, P.T.M., Komura, H. and lnoue, Y. (1989) J. Biol. Chem. 264, 1623-1630. [65] Inoue, S., Iwasaki, M., Ishii, K., Kitajima, K. and Inoue, Y. (1989) J. Biol. Chem. 264, 18520-18526. [66] Iwasaki, M., Seko, A,, Kitajima, K., Inoue, Y. and lnoue, S. (1992) J. Biol. Chem. 267, 24287-24296. [67] Seko, A., Kitajima, K., Inoue, Y.and Inoue, S. (1991) J. Biol. Chem. 266, 221 10-221 14.

161 [68] Knirel, Y.A., Kocharova, N.A., Shaskov, A.S., Kochetkov, N.K., Mamontova, VA. and Solov’eva, T.F. (1989) Carbohydr. Res. 188, 145-155. [69] Strecker, G., Wieruszeski, J.-M., Michalski, J.-C., Alonso, C., Boilly, B. and Montreuil, J. (1992) FEBS Lett. 298, 39-43. [70] Strecker, G., Wieruszeski, J.-M., Michalski, J.-C., Alonso, C., Leroy, Y., Boilly, B. and Montreuil, J. (1992) Eur. J. Biochem. 207, 995-1002. [71] Kanamori, A,, Inoue, S., Xulei, Z., Zuber, C., Roth, J., Kitajima, K., Ye, J., Troy, F.A. and Inoue, Y. (1994) J. Histochem. 101, 333-340. [72] Yu Song, Kitajima, K. and Inoue, Y. (1993) Glycobiology 3, 31-36. [73] Angata, T., Kitajima, K., Inoue, S., Chang, J., Warner, T.G., Troy, F.A. and Inoue, Y. (1994) Glycobiology 4, 517-523. [74] Li, Y.-T., Yuzik, J.A., Li, S.-C., Nematalla, A,, Hasegawa, A., Kimura, M. and Nakagawa, H. (1993) Arch. Biochem. Biophys. 310, 243-246. [75] Kitajima, K., Kuroyanagi, H., Inoue, S., Ye, J., Troy, F.A. and Inoue, Y. (1994) J. Biol. Chem. 269, 21415-21429. [76] Takahashi, N. (1977) Biochem. Biophys. Res. Commun. 76, 1194-1201. [77] Sugiyama, K., Ishihara, H., Tejima, S. and Takahashi, N. (1983) Biochem. Biophys. Res. Commun. 112, 155-160. [78] Plummer Jr, T.H., Elder, J.H., Alexander, S., Phelan, A.W. and Tarentino, A.L. (1984) J. Biol. Chem. 259, 10700-10704. [79] Inoue, S. (1990) Trends Glycosci. Glycotechnol. 2, 225-234. [SO] Suzuki, T., Seko, A,, Kitajima, K., Inoue, Y. and Inoue, S. (1993) Biochem. Biophys. Res. Commun. 194, 1124-1 130. [81] Suzuki, T., Seko, A,, Kitajima, K., Inoue, Y. and Inoue, S. (1993) Glycoconjugate J. 10, 223. [82] Suzuki, T., Seko, A,, Kitajima, K., Inoue, Y. and Inoue, S. (1994) J. Biol. Chem. 269, 17611-17618. [83] Taguchi, T., Kitajima, K., Muto, Y., Inoue, S., Khoo, K.-H., Morns, H.R., Dell, A,, Wallace, R.A., Selman, K. and Inoue, Y.(1995) Glycobiology 5, 61 1-624. [84] Ziak, M., Qu, B., Zuo, X., Zuber, C., Kanamori, A., Kitajima, K., Inoue, S., Inoue, Y. and Roth, J. (1996) Proc. Natl. Acad. Sci. USA 93, 2759-2763. [85] Qu, B., Ziak, M., Zuber, C. and Roth, J. (1996) Proc. Natl. Acad. Sci. USA 93, 8995-8998. [86] Inoue, S., Kitajima, K. and Inoue, Y. (1996) J. Biol. Chem. 271, 24341-24344. [87] Suzuki, T., Kitajima, K., Inoue, S. and Inoue, Y. (1997) In: H.-J. Gabius and S. Gabius (Eds.), Glycosciences. Chapman and Hall, Weinheim, pp. 121-13 1. [88] Suzuki, T., Kitajima, K., Emori, Y., Inoue, Y. and Inoue, S. (1997) Proc. Natl. Acad. Sci. USA 94, 62444249.

This Page Intentionally Left Blank

J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins I1 Elsevier Science B.V. CHAPTER 8

Amphibian glycoproteins Gerard Strecker Laboratoire de Chimie Biologique, UMR no. 111 du CNRS, Universiti des Sciences et Technologies de Lille. 59655 Villeneuve d ’Ascq, France

Abbreviations GalNAc-ol N-acetylgalactosaminitol

Gal

galactose

NeuAc

N-acetylneuraminic acid

GlcNAc

N-acetylglucosamine

Kdn

3-deoxy-D-glycero-D-gaiacto-nonu~osonic GalNAc acid GlcA fucose

Fuc

N-acetylgalactosamine glucuronic acid

Key for glycosyltranferase activities, examples Kdn:(a 14)FucT: a-1,4-fucosyltransferase with Kdn as acceptor. Fuc(a l-4)Kdn: ( a 1-3)FucT:a- 1,3-fucosyltransferase with Fuc(a 1-4)Kdn as acceptor.

1. Introduction When we started these studies, it was our aim to find new and easily accessible sources of natural carbohydrates of high biological interest. Indeed, some rare carbohydrate structures, such as the core GalNAc(al-3)GalNAc(a 1-O)Ser/Thr found in human rectal adenocarcinoma glycoproteins [ 11 and in meconium [2,3], or the terminal sequence GlcNAc(a14)Gal(B 14)GlcNAc present in human meconium but never detected in adult organism, constitute major sugar components in salivary gland mucins of the Chinese swiflet [4] and pig stomachal mucin [5], respectively. We suggested the possibility that some carbohydrate units, which play a fundamental role during tissue differentiation, cellular recognition and malignant growth, could be isolated, in high quantities, from some lower-placed species of the animal kingdom. This hypothesis was based upon the “theory of recapitulation”, proposed by E.H. Heckel (1834-1919): Ontogenesis recapitulates phylogenesis. Pleurodeles waltl and Ambystoma mexicanum (axolotl), bred in many laboratories, were chosen because of the abundance of mucins which surround their eggs. The first analyses led to the isolation of Lewis’ and LewisY antigenic determinants, which are known to be related to human tumor-associated antigens. For these reasons, the jelly coats of amphibian eggs could represent a valuable model to examine 0-linked carbohydrate structures during evolution, associated with a possible function involving specific markers for the fertilization process. 163

164

Table 1 Content (YO) of carbohydrate in various egg jelly coatsa Species

Fuc

Gal

GalNAc

GlcNAc

NeuAc

Kdn

Total 50. I

Bufo bufo

6.4

14.6

15.1

6.2

6.8

0

Rana temporaria

4.3

16.5

10.1

4.4

0

2.7

38.0

10.1

6.1

12.1

7.6

0

6.7

42.6

8.9

8.9

10.2

3.8

0

3.4

35.2 49.1

Ambystoma mexicanum Ambysfoma tigrinum

10.0

5.8

18.6

10.7

0

6.0

Pleurodeles waltl

7.0

6.6

20.2

8.4

0

9.4

5 1.6

Xenopus laevis

8.3

9.1

13.5

8.1

0.9

1.2

41.1

-

Rana palustris

+ +

Bufo japonicus

+

-

Bufo americanus Rana ufricularia

-

+

+

-

Ambystoma maculatum

Rana dalmafina

a

-

Unpublished results, except for Bufo japonicus [ 101. Also contains sulfate and glucuronic acid. Data obtained with a pool of Xenopus clutches (see text).

2. Role of the ooiducal secretions in mediation of gamete fusion in amphibians The requirement of oviducal secretion surrounding amphibian eggs for fertilization is known since the last century. Although amphibians have been often considered as models for studying the mechanism of gamete interaction, the molecular basis for the jelly requirement has been puzzling. The important literature devoted to this subject (for general reviews, see refs. [6-191) can be summarized as follows. The oviducal jellies deposited around amphibian eggs are secreted by the tubular gland cells lining the oviduct. The number of jelly coats varies among species from as many as six in Rana pipiens to only two for Lepidobatrachus laevis [6]. Chemical analysis of amphibian jelly revealed the presence of glycoproteins as major components [7-91 and the low amount of mannose, as well as the high proportion of N-acetylgalactosamine, serine and threonine allowed to predicate that they are typical mucin-type glycoproteins (Table 1). The most documented reports concerning these glycoproteins describe a high content of carbohydrates (about 80%) and a polydisperse state (MM ranging from lo& 4000 KDa) [lo]. It is generally accepted that the biological functions of the amphibian egg jelly coats include (1) sperm binding, (2) induction of the sperm acrosome reaction, (3) forming a barrier for the sperm penetration (4) a block to polyspermy (anurans) and ( 5 ) provision of a protective environment for the developing embryo (for general reviews see refs. [ 11-13]). Many reports have shown that coelomic eggs cannot be fertilized, but become fertilizable when inseminated in the presence of an appropriate jelly preparation [ 14-1 61. The primary role of jelly layers in fertilization may be related to the induction of sperm

165

acrosome reaction. The role of inner jelly layers in amphibian acrosome reaction may be similar to that of fluids from mammalian oviduct of the uterus, which have been also reported to stimulate introduction of sperm capacitation. Soluble factor(s) obtained by dialysis or after ashing at 600°C retained the biological activity, an observation that excluded the involvement of organic substances [ 171. The finding that Ca2+ and/or Mg2+ are essential factors for supporting fertilization of dejellied eggs, followed by determination of the Ca2+ binding capacity of the egg jelly glycoproteins, showed finally that the jelly keeps the ionic environment required for successful fertilization. These observations argued against the proposal that specific molecular interactions operating between jelly and sperm surface underlie the essential role of egg jelly in the fertilization process. The absence of a species-specificity occurring during these experiments performed in uitro may be explained by these properties of egg jelly factors. Although intergenic fertilization is possible in vivo [ 121, cross-fertilization remains impossible in most of the cases. This blockade of sperm migration observed in heterologous jelly is likely due to the incompatibility of the physical architecture of the egg jelly, as argued by Katagari [ 121. The role of carbohydrate determinants in amphibian egg vitelline envelope (which corresponds to the zona pellucida of mammalians) has been clearly demonstrated in the case of Xenopus laeois. The monospermy observed among anurans, but not in urodeles, involves the reaction of a lectin from egg cortical granules with components of the vitelline envelope and the innermost jelly coat JI surrounding the egg [18]. The action of cortical granule N-acetyl-b-D-glucosaminidase on egg integuments may function as a block to polyspermy at fertilization [ 191.

3. Carbohydrate chains of egg jelly coat glycoproteins Glycoproteins in amphibians have not been yet extensively analyzed, although it is well known since longtime that egg jellies or skin mucus are carbohydrate-rich materials. The integumentary mucins from Xenopus laeois have been investigated by molecular cloning in order to examine the organization of the polypeptide backbone [20-221. Analysis of carbohydrates in various egg jelly coats (Table 1) showed the presence of Fuc, Gal, GalNAc and GlcNAc, in variable amounts. Whereas NeuSAc occurs in Bufo bufo (personal observation) and Bufo japonicus japonicus [lo], Kdn which has been characterized for the first time in fishes by Inoue [23] (see chapter 7, this volume) appears to be highly representative for amphibians. In Ambystoma and Xenopus, Kdn and Kdn 9-0-Ac are occurring in the molar ratio 1:l (unpublished results). Traces of Kdn 7-0-Ac and Kdn 7,9-0-Ac were found in Pleurodeles (A. Klein, personal communication). The carbohydrate content averages between 35 and 50% of the mass of the crude material, but is in the order of 80% in a pure glycoprotein fraction [lo]. The carbohydrate composition strongly suggested that glycans consist mainly of 0-linked structures. They were released by reductive 6 elimination and analyzed by NMR spectroscopy [10,24-331. As shown in Figs. 1-8, a species-specific structural diversity has been observed in carbohydrate chains of amphibian egg jelly.

166

Kdn(a2-6)

Kdn(a24)

\

\

/ GlcNAc( p1-3)

/ GalNAc(a 1-3)Gal( p 14)GlcNAc(P1-3)

GalNAc-01

GalNAc-ol Fuc(al-3)/

Kdn(a24)

\

Kdn(a2-6)

GalNAc-01

\

/ Gal(p1-3)

GalNAc-ol

/ Fuc(a 1-2)Gal(P 14)GlcNAc(P1-3)

Fuc(a1-2)/ Kdn(a2-6)

Fuc(a1-3)’

\

GalNAc-ol

GalNAc(a1-3),

/ Gal@-3)

Kdn(a2-6) GalNAc(a1-3),

Fuc(a1-2)’

Fuc(a 1-2)/

Kdn(a24)

\ GalNAc-01 /

Gal(pl4)Gl~NA~(pl-3) Fuc(a 1-3f

\

GalNAc-01

/ Gal@1-4)GlcNAc(pl-3) Fuc(al-3)/ Fig. I . Oligosaccharide-alditols released from the jelly coat of Pleurodeles waltl[24]. GlcNAc(P 1-3)GalNAc4 Fuc(al-2)Gal(

Kdn(a24)

\

P1-4)GlcNAc(p 1-3)GalNAc-ol

Gal(a 1-4)Gal(P14)GlcNAc(P1-3)GalNAc-1 Fuc(al-2)/ Fuc(a 1-4)Kdn(a2-6) Fuc(a 1-3/

GalNAc-ol

/ Gal(al4)Gal(pl-4)GlcNAc(pl-3) Fuc(a 1-2)/ Fuc(a 1-4)Kdn(a24)

\

\

/

/

GalNAc-ol

GalNAc-ol

GlcNAc(P1-3)

Gal(a 14)Gal(P1-4)GlcNAc(P1-3f Fuc(al-2)/

Fuc(alA)Kdn(a24)

Fuc(al-3)Fuc(al4)Kdn(a24)

\

Fuc(a 1-3)/

G~NAc-~~ Gal(pl4)GlcNAc(pl-3) Fuc(al-2)/

/

\

GalNAc-ol

/ Gal(al4)Gal(~l-4)GlcNAc(~l-3) Fuc(a 1-2f

Fig. 2. Oligosaccharide-alditols released from the jelly coat of Ambystoma rnexicanum [25]

The mucin of Pleurodeles waltl (Fig. 1) appeared to be an abundant source of LeX, Ley and A Ley antigenic determinants, which can be prepared on a large scale, after removing Kdn by mild acidolysis. The much of Ambystoma maculatum (Fig. 3) contained essentially the determinant GalNAc(P14)[Fuc(a 1-3)]GlcNAc, which has also

167 GalNAc(P 14)GlcNAc(P1-3)GalNAc-d Fuc(al-3)/ Gal(plA)GlcNAc(P 1-3)GalNAc-ol Fuc(al-3)/ GalNAc(pl4)GlcNAc(P 1-6) Fuc(a 1-3)' Gal@1-3)

\

/""""'

GalNAc(PlA)GlcNAc(pl-6) \ Fuc(al-3)/ Fuc(al-3)Fuc(al4)Kdn(a2-3)Gal@l-3)

Fuc(a1-3) \ Fuc(alA)Kdn(a2-6) Fuc(a1-2)' \ Gal@14)GlcNAc(/31-3)

Gal(~l4)GlcNAc(pl-6) Fuc(al-3)/ \ GaWAc-ol Gal(P1-3) GalNAc(Pl4)GlcNAc(P 1-6) Fuc(al-3)/ Kdn(a2-3)Gal(P 1-3)

/

Fuc(al-3)/ Fuc(a1-3) \ Fuc(al4)Kdn(a2-6) \ Fuc(a1-2) / GalNAc-ol GalNAc(P14)GlcNAc(P1-3) Fuc(al-3)/

\

7"""

Fuc(a li3f

Fuc(a 1-2)Fuc(a 14)Kdn(a24)

\

Fuc(a1-3).

/

Fuc(a1-2) /

GaWAc-ol

GalNAc(p14)GlcNAc(Pl-3) Fuc(al-3)/

/

GalNAc-01

/

Fuc(al4)Kdn(a2-3)Gal(~1-3f

Fig. 3 . Oligosaccharide-alditols released from the jelly coat of Ambystoma maculatum [26,27].

been found in Schistosoma mansoni membranes [34], human urokinase [35], sea squirt H allergenic antigen [36] and recombinant glycoproteins expressed in human kidney 293 cells [37]. These data confirmed the hypothesis above stated about the presence of particularly rare human carbohydrate chains in lower animal species. Nevertheless, the glycans observed in the other amphibian species represent novel sequences which reflect a natural polymorphism. The mucin of Ambystoma mexicanum was characterized by the presence of new sequences such as Gal(al4)[Fuc(al2)]Gal(P 14)GlcNAc (hybrid P 1/H blood group determinant) and Fuc(a 1-3)Fuc(a 14)Kdn (Fig. 2). Such a substituted Kdn unit has been also observed in A . maculatum (Fig. 3) and X . laeois (Fig. 5). More surprising was the finding of disubstituted Kdn in A . tigrinum (Fig. 4). New polyfucosyl sequences like the trimer Fuc(a1-2)[Fuc(al3)]Fuc(al4) were also observed in A . maculatum (Fig. 4). Other unusual sequences, namely the dimer H epitope Fuc(a 1-2)Ga1(8 1-3)[Fuc(a 1-2)]Gal, and blood group A determinant terminated with an additional a-1,3-linked GlcNAc unit, were found in X.Iaevis (Fig. 5 ) . The first analyses performed on urodeles species samples pointed to the exclusive presence of Kdn, indicating that this sugar could be considered as the potential marker of this subclass of amphibians. But it was rapidly shown that the presence of Kdn or Neu5Ac among urodeles or anurans was not correlated with one of these amphibian subclasses. X . laeois seems to be an exception among the species analyzed so far, since the analysis

168 Fuc(al-4)Kdn(a2-6)

Fuc(al-2)Gal(pl-3)

\ GalNAc-ol /

Fuc(a1-3) \ Gal(PI4)GlcNAc(P 1-6)

\

Fuc(al-5)\ ICdn(a2-3)Gal(pl-3) Fuc(al-4) /

Fuc(al-4)Kdn(a2-6)

Gal(a1-3)

\

7""""'

Fuc(a1-5) \ Kdn(a2-6) \ Fuc(al-4) / GalNAc-ol / Gal(a1-3f

GaINAc(a1-3) \ Gal(pl-4)GlcNAc(pl-6) Fuc(a1-2) / \ GalNAc-ol Fuc(a1-5) \ / Kdn(al-4)Gal(p1-3) Fuc(al-4) Fuc(al-3)\ Fuc(al-2)Gal(p 14)GlcNAc(p1-6)

\ Fuc(al-5), Fuc(al-%)Kdn(a2-6)

\

Kdn(al-4)Gal(pl-3)

Fuc(ul-4)

'

Fuc(al-Z)Gal(al-3) Fuc(al-5)\ Fuc(a1-5)

Fuc(al-4)

Kdnfa2-6)

\ Kdn(a2-3)Gal(P 1-3)GaINAc-ol /

Fuc(al-4)/

~

'\GalNAc-ol /

Fuc(alJ), Kdn(a2-3)Gal(pl-3f Fuc(al-4)'

Fuc(al-Z)Gal(p1-4)GlcNAc(~1-6)

\\ Fuc(a1-5) \ Kdn(a%3)Gal(pl-3) Fuc(al-4)/

Fig. 4. Oligosaccharide-alditolsreleased from the jelly coat of Ambystoma tigrinurn [28,29].

of a series of clutches showed the presence of NeuSAc, while other samples, produced by other X. laeois specimen contained exclusively Kdn. These discrepancies require further experiments, based on the analysis of clutches obtained under controlled physiological conditions. Perhaps the Xenopus genus, which can display the hybridism phenomenon among numerous species, needs to be better defined from a taxonomic point of view. In Rana temporaria, the anionic charge is carried by Kdn, and also by sulfate and glucuronic acid (Fig. 8). In Rana utricularia, traces of NeuSAc were characterized and the presence of glucuronic acid and sulfate has been confirmed. On the basis of these results, it remains impossible to correlate the presence of these acidic sugars with the taxonomy and phylogeny of amphibians.

I69 GlcNAc(P1-6)

\ GalNAool / Gal(P1-3)

GlcNAc(P1-6)

\

Gah‘Ac-ol

Gal(al-4)Gal(pl-3)Gal(~l-3) / Fuc(al-2)’Fuc(al-2)’

GlcNac(P1-6)

Fuc(al-2)Gal(al-3)

\

NeuAc(aZ4)

\

GalNAc-ol

/GalNAc-ol Fuc(al-Z)Gal(p 1-3)Gal(pl-3) fic(al-2)/

/

Gal(al-4)Gal(pl-3)GalNAc4 Fuc(a1-2)

/

’

Fuc(al-4)Kdn(a2-6)

Fuc(a1-3)

GalNAwl

GlcNAc(al-3)Gah‘Ac(aI-3)Gal(~1-3)GalNAc-ol Fuc(a 1-2)

GlcNac(p1-3)

/

Fuc(al-I)Kdn(a2-6) GlcNAc(P1-6)

\ / \

GalNAc-01

\

GalNAc(~l-3)Gal(~l4)GlcNAc(~l-3) / Fuc(al-2)/

Fuc(al-2)Gal(pl-3)Gal(~l-3)

/ Fuc(al-3)Fuc(al~)Kdn(~-6)

Fuc(a1-2)

\

Fuc(al-2)Gal(pl-3)Gal(~l-3)GalNAc-ol Fuc(a1-2) /

GalNAc-01

/ GaINAc(pl-3)Gal(P I-4)GlcNAc(p1-3) Fuc(al-Z)/

Fig. 5. Oligosaccharide-alditols released from the jelly coat of Xenopus laeuis [30,3 11

NeuAc(a2-6)

GlcNAc(P14)

\

\

/ Gal(pl-3)

/ F~c(~tl-2)Gal( pl-3)

GalNAc-ol

NeuAc(a2-6)

\

GalNAc-ol

Fuc(al-2)Gal( p 1-3 j

GalNAc-ol

Fuc(a 1-2)Gal( pl-3)GalNAc-ol

/

Fig. 6 . Oligosaccharide-alditols released from the jelly coat of Bufo japonicus japonicus [ 101

4. Concluding remarks The starting hypothesis that amphibian egg jellies could be a good source of carbohydrates has been shown to be correct. Moreover, the tissues of the animals and in particular the oviducts could also be the source of new and specific glycosyltransferases. The structures listed in Figs. 1-8 allow us to speculate about the activity levels and specificities of transferases in the glycosylation process of oviducal mucins. In fact, referring only to the fucosyltransferase activities, a minimum of six enzymes such as Kdn:(a I-4)FucT, Fuc(a 14)Kdn:(a l-2)FucT, Fuc(a 1-4)Kdn:(a I-3)FucT, Kdn:(a 1-

170

Fig. 7. Oligosaccharide-alditols released from the jelly coat of Bufo bufo [32].

S)FucT, Gal(a 1-3)Gal:(a I-2)FucT and Gal(a l-3)[Fuc(a l-2)]Gal:(a 1-2)FucT can be postulated. As a further attempt to correlate the structure of the enzyme with the specific activity, such a model should be investigated by molecular cloning of this family of fucosyltransferases. Until now, no experimental results support the hypothesis of a carbohydrate-mediated species-specific sperm binding for amphibians, similar to that demonstrated for mammalians. Further studies will probably indicate if the pattern of glycan structures found in each amphibian species represent the support for a specific recognition of the ovule by the spermatozoon. At least, these findings create a field for the hypothesis, that carbohydrate structures could be used as a specific taxonomic marker. However further confirmation should come through the study of other species. The data also point to the diversity of novel glycosyltransferase activities which remain to be studied with regard to animal evolution.

References [ I ] Kurosaka, A,, Nakajima, H., Funakoski, I., Matsuyama, M., Nagayo, T. and Yamashina, 1. (1983) J. Biol. Chem. 258, 11594-1 1598. [2] Hounsell, E.F., Lawson, A.M., Feeney, J., Gooi, H.C., Pickering, N.J., Stoll, M.S., Lui, S.C. and Feizi, T. (1985) Eur. J. Biochem. 148, 367-377. [3] Capon, C., Leroy, Y., Wieruszeski, J.M., Ricart, G., Strecker, G., Montreuil, J. and Fournet, B. (1989) Eur. J. Biochem. 182, 139-1 52. [4] Wieruszeski, J.M., Michalski, J.C., Montreuil, J., Strecker, G., Peter-Katalinic, J., van Halbeek, H., Mutsears, J.H.G.M. and Vliegenthart, J.F.G. (1987) J. Biol. Chem. 262, 6650-6657.

GalNAc-ol

GalNAc-ol

HS03(6)

/ Gal(p1-3)

Fuc(a1-3) /GlcNAc(P14)

HS03(3)/ Gal(P1-3)

Gal(p1-4)Gal(p1-3) /GalNAcq1

/

Kdn(a2-6)

\

1

Gal(p1-3)/ Fuc(a1-2)

HS03(3)/ Gal@14)Gal(p1-3) / / HS03(3) Fuc(a1-2)

GalNAc-ol

Kdn(a2-6)

GalNAc-ol Gal(pl-4)S;al(pl-3)

\

GalNAc-ol

Gal(a1-3) Kdn(a2-6)

/

/ Gal(P1-3) /GalNAC-O1

I

HS03(3)/Gal(pl-3) Fuc(a1-2)

1

GalNAc-ol Gal(pl-4)Gal(Pl-3) HSO3(3)’Gal@1-3) /

\

1

/

GlcA(P1-3)/

Gal(p1-3{ F W a 1-2)

1

Fuc(a1-2)

I Fuc(a1-2) Fuc(a1-2)

Gal(p1-3)

Fuc(a1-2) /GlcNAc@14)

HS03(6)

\

Kdn(a2-6)

GlcNAc(P 1-6)

\

Gal(a1-3) Fuc(a1-2)

\

GalNAc-ol

GalNAc-ol /

\

Gal(P1-3)

GalNAc-ol Gal(a1-3)

\

/

Gal(p1-3)

Fuc(a1-2)

/

/

Fig. 8. Oligosaccharide-alditols released from the jelly coat of Rana temporaria [33].

Gal(P1-3) Gal(Pl-3)/ / GlcA(P1-3) Fuc(a1-2)

1

1

Fuc(a1-2)

\

172 [5] van Halbeek, H., Dorland, L., Vliegenthart, J.F.G., Kochetkov, N.K., Arbatsky, N.F! and Derevitskaya, V.A. (1982) Eur. J. Biochem. 127, 21-28. [6] Caroll, E.J., Wei, S.H., Nagel, G.M. and Ruibal, R. (1991) Develop. Growth Differ. 33, 3 7 4 3 . [7] Bolognani, L., Bolognani-Fantin, A.M., Lusignani, R. and Tonta, L. (1966) Experientia 22, 601. [8] Jego, P. (1974) Comput. Biochem. Physiol. 47, 435446. [9] Freeman, S.B. (1968) Biol. Bull. 135, 501-513. [lo] Shimoda, Y., Kitajima, K., Inoue, S. and Inoue, Y. (1994) Eur. J. Biochem. 223, 223-231. [ l l ] Jego, P., Jolly, J. and Boisseau, C. (1980) Reproduct. Nutr. Develop. 20, 557-567. [I21 Katagari, C. (1987) Zool. Sci. 4, 1-14. [I31 Hedrick, J.L. and Nishihara, T. (1991) J. Electron Microsc. Tech. 17, 319-335. [14] Katagari, C. (1966) Embryologia 9, 159-169. [15] Barbieri, ED. and Raisman, J.S. (1969) Embryologia, 10, 363-372. [16] Elinson, R.P. (1971) J. Exp. Zool. 176, 415428. [I71 Ishihara, K., Hosono, J., Kanatani, H. and Katagari, C. (1984) Develop. Biol. 105, 435442. [l8] Wyrick, R.E., Nishihara, T. and Hedrick, J.L. (1974) Proc. Natl. Acad. Sci. USA 71, 2067-2071. [19] Prody, G.A., Greve, G.A. and Hedrick, J.L. (1985) J. Exp. Zool. 235, 335-340. [20] Hoffman, W. (1988) J. Biol. Chem. 263, 63104316. [21] Probst, J.C., Hauser, F., Joba, W. and Hoffman, W. (1990) Biochemistry 29, 6240-6244. [22] Hoffman, W. and Joba, W. (1995) Biochem. SOC.Trans. 23, 805-810. [23] Inoue, Y. and Inoue, S. (1997) ch. 7, this volume. [24] Strecker, G., Wieruszeski, J.M., Alonso, C., Michalski, J.C., Boilly, B. and Montreuil, J. (1992) FEBS Lett. 298, 3 9 4 3 . [25] Strecker, G., Wieruszeski, J.M., Michalski, J.C., Alonso, C., Leroy, Y., Boilly, B. and Montreuil, J. (1992) Eur. J. Biochem. 207, 995-1002. [26] Strecker, G., Wieruszeski, J.M., Fontaine, M.D. and Plancke, Y. (1994) Glycobiology 4, 604-609. [27] Fontaine, M.D., Wieruszeski, J.M., Plancke, Y., Delplace, F. and Strecker, G. (1995) Eur. J. Biochem. 231, 424-433, [28] Maes, E., Wieruszeski, J.M., Plancke, Y. and Strecker, G. (1995) FEBS Lett. 358, 205-210. [29] Maes, E., Plancke, Y., Delplace, F. and Strecker, G. (1995) Eur. J. Biochem. 230, 146-156. [30] Strecker, G., Wieruszeski, J.M., Plancke, Y. and Boilly, B. (1995) Glycobiology 5, 137-146. [31] Plancke, Y., Wieruszeski, J.M., Alonso, C., Boilly, B. and Strecker, G . (1995) Eur. J. Biochem. 231, 434439. [32] MoreIle, W. and Strecker, G.(1997) Glycobiology, in press. [33] Maes, E., Florea, D., Delplace, F., Lemoine, J., Plancke, Y. and Strecker, G. (1996) Glycoconjugate J. 14, 127-146. [34] Srivatsan, J., Smith, D.F. and Cummings, R.D. (1992) Glycobiology 2, 445452. [35] BergwerR, A.A., Thomas-Oates, J.E., van Oostrum, J., Kamerling, J.P. and Vliegenthart, J.F.G. (1992) FEBS Lett. 314, 389-394. [36] Ohta, M., Matsuura, F., Kobayashi, Y., Shigata, S., Ono, K. and Oka, S. (1991) Arch. Biophys. Biochem. 290, 474483. [37] Yan, S.B., Chao, Y.B. and van Halbeek, H. (1993) Glycobiology 3, 597-609.

J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins II

0 1997 Elsevier Science B.V. All rights reserved CHAPTER 9

Blood glycoproteins* Kenneth J. Clemetson Theodor Kocher Institute, University of Berne, Berne, Switzerland

Abbreviations AAT

a1-antitrypsin

AGP apo ATIII

a l -acid glycoprotein apolipoprotein antithrombin 111

HPLC LCA

Lens culinaris agglutinin

LCAT MAA

1ecithin:cholesterol acyltransferase Maackia amurensis agglutinin

AZ

Alzheimer’s disease

MCTD mixed connective tissue disease

CEA

carcinoembryonic antigen

high performance liquid chromatography

cs

consensus sequence

Onf PM

CSF

cerebrospinal fluid

RA

Con A DSA EGF ESIMS FJ Fn GBF GP HP

concanavalin A

SCL SDS SF

synovial fluid

SLE SNA

Sambucus nigra agglutinin

Datura stramonium agglutinin

epidermal growth factor electrospray ionization mass spectrometry factor J fibronectin

ss

oncofetal polymyositis rheumatoid arthritis scleroderma sodium dodecyl sulfate systemic lupus erythematosus Sjogren’s syndrome wheat germ agglutinin

gelatin-binding fragment

WGA

glycoprotein

WHHL Watanabe heritable hyperlipidemic

hepatopathies

1. Introduction Blood consists of a liquid, plasma, in which a variety of cells circulate. It comes into contact with the lining of the vasculature, consisting of endothelial cells which normally provide a non-thrombotic surface. The plasma contains a very large number of different proteins of which only one, albumin, is known not to be glycosylated. The cells circulating in blood are erythrocytes, responsible for oxygen and carbon dioxide transport, platelets, with a basic responsibility for haemostasis and maintaining the integrity and function of the vascular system and with less well-defined roles in inflammation and defence against parasites, and leukocytes; these last are divided into several categories: lymphocytes, neutrophils, basophils, eosinophils and monocytes. The leukocytes are all critical in the body’s defence against bacteria, viruses and parasites. The endothelial cells provide the non-thrombogenic coating to the vasculature, secrete factors acting on the other blood cells and have an important repair function. As with other cells, all these blood cells have a plasma membrane, containing glycoproteins with a wide variety of functions essential for a normal physiology. These include adhesive receptors, agonist receptors and molecules with protective or housekeeping functions. ~

~~~

This chapter is dedicated to Prof. R.U. Lemieux who played a major role in awakening a whole generation to the importance of carbohydrate structure in biology. 173

114

Gal(fi14)GlcNAc(fi14)

\

Gal@14)GlcNAc(fi 1-2)Man(a 1-6) Fuc(a 1-3),-3

\

FuW+,

Man(BI4)GlcNAc(fi14)GlcNAc(fiI-N)Asn (B) I

Gal@14)GlcNAc(fi1-2)Man(a 1-3) I Gal(fi14)GlcNAc(fi1-6)

2. Plasma proteins Plasma contains minor amounts of a wide range of glycoproteins, involved in immunology, protease inhibition, transport and other functions, and mostly synthesized in the liver [I]. Some major components are listed below.

2. I . al-Acid glycoprotein al-Acid glycoprotein (AGP), also called orosomucoid, is one of the major plasma proteins and has been shown to bind a number of basic drugs. Nevertheless, its function remains obscure. Despite preparations with confirmed purity at the protein level it was demonstrated fairly early that it showed charge microheterogeneity [2] which could be ascribed to glycosylation differences. It has a high carbohydrate content (40%) with 5 N-linked oligosaccharide chains. The structure of these chains includes di-, tri- and tetraantennary types with and without fucosc residues (Fig. 1, 2A) [3-71. These chains are extensively sialylated, accounting for the acidic PI of this molecule and also give rise to the extensive charge heterogeneity. During the acute-phase immunological reaction the forms with complex glycans arc produced preferentially over the diantennary forms. Glycosylation variations of AGP have been found in various disease states. AGP from ascitic fluid of patients with stomach cancer was separated by chromatography on Con A-Sepharose into Con A-nonbound (AGP-21, 43, 5kDa, 70%) and Con A-bound (AGP-2, 41.5kDa, 24% and AGP-3, 40.0 kDa, 5%) forms, differing in monosaccharide composition [8]. Comparative study of N-glycan structures by HPLC of fluorescence-labelled oligosaccharides showed that the molecular forms differ in the ratio of the di-, tri-, and tetraantennary complex type carbohydrate N-chains. The molecular forms of AGP differ from nAGP in amounts of Lewis X (Lex)-fragments and agalacto-oligosaccharides. Plasma protein glycoforms rich in diantennary complex type N-glycans (type I) increase in acute infections, while in some diseases with chronic inflammatory changes, more branched N-glycans (type 11) increase. In sera from 109 human immunodeficiency

175

virus (H1V)-infected persons, 38 rheumatoid arthritis patients, and 44 healthy subjects, the composition of AGP glycoforms was studied using crossed immunoaffinity electrophoresis with concanavalin A (Con A) as a ligand [9]. In patients with chronic inflammatory changes classifications I, 11, and 111, the distribution of AGP glycoforms was analogous to that in normal subjects. Type I alterations were observed in patients in group IV who had no signs of arthritis. Type I1 changes, analogous to those found in rheumatoid arthritis, were seen in group IV patients who developed arthritis. Most significant type I changes were associated with Pneumocystis carinii pneumonia. Rat plasma glycoproteins, but not human, contain terminal 0-acetyl, N-glycolylneuraminic acids in the carbohydrate side-chains of “acute-phase” glycoproteins (predominantly AGP) [ 101. The microheterogeneity of AGP has been studied in 43 patients with early rheumatoid arthritis (RA) without clinical features of intercurrent infection [ 111. In contrast to previous reports, suggesting a decrease in Con A reactivity in patients with RA, high values of AGP reactivity with Con A were found in patients with a disease of short duration, similar to those in patients with acute bacterial infections. Conversely, normal or decreased values of AGP reactivity coefficients were found in patients with a disease of longer duration. Together with previous findings suggesting that cytokines control the glycosylation of acute phase proteins, these results indicate that differences in the microheterogeneity of AGP in early and longstanding RA reflect differences in cytokine action at different stages of the disease. Acute inflammation as well as cirrhosis induces increases in a-1,3-fucosylated AGP molecules, detected first of all by decreased binding to Con A [12,13] and later by the reactivity of AGP, presumably containing three or more fucosylated N-acetyllactosamine units, towards the hcose-binding Aleuria aurantia lectin in crossed affino-immunoelectrophoresis of human sera [14]. In at least part of these Lewis Xtype glycans (Gal~I-[Fuca1-3]GlcNAc-R) appeared to be substituted also with an a-2,3linked sialic acid residue. The structures of a fucosylated tetrasialyl oligosaccharide from cirrhotic AGP [ 12,131 is shown in Fig. 2B and should be compared with the structure from healthy donors in Fig. 2A. Acute inflammation induces a strong increase in sialyl Lewis Xsubstituted AGP molecules that persist at a high level throughout the inflammatory period. These changes may represent a physiological feedback response to the interaction between leukocytes and inflamed endothelium, mediated via sialylated Lewis X structures and the selectin endothelial-leukocyte adhesion molecule. The reducing oligosaccharides released from AGP by conventional hydrazinolysis have been analyzed [ 151. At least 13 different asialo-N-glycans were detected. The carbohydrate structures were assigned by comparison with the known AGP carbohydrate structures and known N-glycan structures by comparison of retention times. In addition to the hitherto known AGP carbohydrate structures, a number of sulfated N-glycans were also tentatively identified. Serum and synovial fluid have been obtained at the same time from 22 patients with rheumatoid arthntis and analyzed for microheterogeneity of a 1-acid-glycoprotein [ 161. In most samples the glycosylation pattern was similar, a nonreactive variant and two Con A reactive variants (the first and the second). In seven samples of synovial fluid an extra third peak was observed representative of the fraction strongly reactive with Con A.

176

NeuSAc(a2-3)Gal(~l4)GlcNAc(~l 4)

\

NeuSAc(a24)Gal(fi14)GlcNAc(fi1-2)Man(a1-6) \

\

Fuc(a1-3)

Man(filM)GIcNAc(fi14)GlcNAc(fil-N)Asn

I

\ -

Neu5Ac(a2-3)Gal(fi14)GlcNAc(fil-2)Man(a 1-3)

(A)

I

I

NeuSAc(a2-6)Gal(fi14)GlcNAc(fil-6) Fuc(a 1-3)

\

Neu5Ac(a2-3)Gal(fi14)GlcNAc(fil4)

\

NeuSAc(a2-6)Gal(fi14)GlcNAc(fiI-2)Man(a 1-6)

\ Fuc(a 1-3)

Man(PI4)GlcNAc(fil4)GlcNAc(fiI-N)Asn (B) \

NeuSAc(a2-3)Gal(P 14)GlcNAc(fi1-2)Man(a 1-3) NeuSAc(a2-6)Gal(fi14)GlcNAc(fi 1-6)

I

I

Fig. 2. Highly fucosylated tetraantennary glycans from a 1 -acid glycoprotein from a cirrhotic patient.

AGP contains complex di-, tri- and tetraantennary glycan chains and can be fractionated into three molecular variants using Con A-lectin chromatography based on variations in these structures. Standard HPLC profiles have been developed to analyze the percentage and distribution of the glycoforms present at each glycosylation site in AGP and its molecular variants [17]. The proportions of di-, tri- and tetraantennary glycans differ at each site for the three molecular variants. The most strongly retained variant from Con A has diantennary glycans at all five sites, whereas the unretained variant is completely devoid of diantennary structures. Only glycosylation site I1 of the five present is completely diantennary in the retained and weakly retained variants. In addition, the two gene products of AGP were glycosylated differently. The “site-directed” model of processing offers the most consistent explanation for the structures seen at the individual glycosylation sites of AGP.

2.2. Antithrombin III Antithrombin 111 (ATIII) is an a2-globulin with a mass of about 58kDa made up of about 425 amino acids. It has a high degree of sequence similarity with al-antitrypsin suggesting a common origin. Two isoforms occur naturally in human plasma. The a-AT111 isoform has four N-glycans attached to Asn 96, 135, 155, and 192. The p-ATIII isoform lacks carbohydrate on Asn 135 (N135), which is near the heparin binding site, and binds heparin with higher affinity than does a-ATIII. Two isoforms (a’ and are also produced when the normal human ATIII cDNA sequence is expressed in baculovirus-infected insect cells, and the recombinant fi-isoform also binds heparin with higher affinity than the recombinant a-isoform. Consensus sequences (CSs) of the ATIII N-glycosylation sites are N-X-S for 135 and N-X-T for 96, 155, and 192. Database and in oitro glycosylation studies suggest that N-X-S CSs are used less efficiently than N-X-T CSs. The 0-ATIII isoform might result from inefficient core glycosylation of the

B’)

177

N135 N-X-S CS due to the presence of a serine, rather than a threonine, in the third position. ATIIIs with N-X-S, N-X-T, and N-X-A consensus sequences were expressed in baculovirus-infected insect cells. In contrast to the N-X-S sequence, which expressed a mixture of a’ and p’ molecules, the N-X-T variant produced a’ exclusively, while the N-X-A variant produced fi’exclusively. Serine in the third position of the N135 CS is responsible for “partial” glycosylation and leads to production of p-ATIII [ 181.

2.3. al-Antitrypsin

a 1-Antitrypsin (Mr 54 kDa, 394 amino acids) is a major plasma component (290 mg/ml) with an important protease inhibitor activity. The carbohydrate content is about 12% [ 191. This glycoprotein is initially synthesized as a 49 kDa single polypeptide chain with the high mannose core oligosaccharide structure GlqMang GlcNAc linked to Asn at 46, 83, and 247. The mature glycoprotein is 54kDa with complex type glycans. Two types of carbohydrate chains are present in normal individuals, diantennary and triantennary glycans (Fig. 3). Gal(fi1-4)GlcNAc(fi1-2)Man(a1-6)

\ [GlcNAc(fiI

I Man@ 14)GlcNAc(fi 14)GlcNAc

Gal(fi14)GlcNAc(fil-Z)Man(a 1-3)

(A)

I

Gal@ 1-4)GlcNAc(P 1-2) \

Man(a1-3) Gal@ 14)GlcNAc(fi1-4)I

\ Man(fi14)GlcNAc(fi 1-4)GlcNAc

(B)

I

Gal(fi1-4)GlcNAc(~1-2)Man(a1-6)’ Fig. 3. (A) Di- and (B) triantennary glycans from antitrypsin.

A diantennary chain with a bisecting N-acetylglucosamine residue was also detected. The chains are present in 80% diantennary, 14% diantennary chain with a bisecting N-acetylglucosamine residue and 6% triantennary, respectively [20]. In hepatocellular carcinoma a diantennary carbohydrate chain with a fucose residue at the innermost N-acetylglucosamine residue was also detected which had a higher affinity for Lens culinaris agglutinin and appeared characteristic for this disease. Alveolar epithelial cells were shown to produce a 1-antitrypsin (AAT) with a modified glycosylation compared to serum AAT [21]. A monoclonal antibody against a specific epitope on al-antitrypsin was shown to be a valuable diagnostic marker for autoimmune conditions [22]. Evidence suggests that this epitope is influenced by the glycosylation of the molecule implying that this is altered in these disorders. In patients with cystic fibrosis[23] significant glycosylation changes in serum al-AT were only seen with free Con A and WGA; this probably results from a reduced synthesis of the diantennary side-chains or by their increased catabolism. Changes in meconium AAT glycosylation were more pronounced with free Con A and LCA. These differences may be useful in the diagnosis of cystic

178

fibrosis. a 1-Antitrypsin was among those hepatic plasma glycoproteins affected in chronic alcoholics, compared to controls, by differences in glycosylation that could be removed by treatment with sialidase [24,25].

2.4. Apolipoproteins Plasma contains several apolipoproteins, involved in lipid and cholesterol transport. Several have been reported to be glycosylated in low amounts including apoB, apoE and apoC-I11[26]. ApoB contains about 2-2.5% carbohydrate as complex-type N-glycan whereas apoE appears to contain O-linked glycan [27]. ApoE is synthesized in sialylated form but in plasma it is 80% desialylated. Site-specific structural characterization of the glycosylation of human 1ecithin:cholesterol acyltransferase (LCAT) has been carried out using microbore reversed-phase high performance liquid chromatography coupled with electrospray ionization mass spectrometry (HPLUESIMS) [28]. Monitoring of carbohydrate-specific fragment ions during HPLCESIMS located eight different groups of glycopeptides in a human LCAT protein digest. In addition to the four expected N-linked glycopeptides of LCAT, a di-O-linked glycopeptide was detected, as well as three additional glycopeptides. Structural information on the oligosaccharides from all eight glycopeptides was obtained by sequential glycosidase digestion of the glycopeptides followed by HPLC/ESIMS. All four potential N-linked glycosylation sites (Asn 20, Asn 84, Asn 272, and Asn 384) of LCAT contained sialylated triantennary andor diantennary complex structures. Two unexpected O-linked glycosylation sites at Thr 407 and Ser 409 of the LCAT O-linked glycopeptide were identified, each with sialylated Gal(@1-3)N-GalNAc structures. The three additional glycopeptides were from a copurifying protein, apolipoprotein D, which contains potential N-linked glycosylation sites at Asn 45 and Asn 78. These glycopeptides bear sialylated triantennary oligosaccharides or fucosylated sialylated diantennary oligosaccharides. Previous studies of LCAT indicated that removal of the glycosylation site at Asn 272 converts this protein into a phospholipase [29]. Most apolipoprotein B (apoB) in rat hepatocyte membranes is membrane-bound. Various data based on glycosylation specific monoclonal antibodies and carbohydratespecific labelling suggest that membrane-bound apoB is more glycosylated than plasma lipoprotein apoB [30]. N-glycans of apolipoprotein (apo) B- 100 in Watanabe heritable hyperlipidemic (WHHL) rabbit and fasting Japanese White rabbits are heterogeneous [3 I]. The N-glycans of apo B-100 consist of one neutral (N) and two acidic fractions (A1 and A2). N contained a high mannose type oligosaccharide consisting of MansGlcNAq to Man9 GlcNAcz, while A 1 and A2 contained monosialylated and disialylated complex type oligosaccharides, respectively. The molar ratio varied among the five WHHL rabbits. There was an inverse correlation between the ratio of acidic oligosaccharide fractions (A1 +A2) and serum cholesterol levels (r=-0.971, P less than 0.01) in the five WHHL rabbits. The N-glycosylation of apo B-1 00 is closely related to cholesterol metabolism in WHHL rabbits. ApoD consists of 169 amino acid residues, including five cysteines. Cys41 and Cysl6 are joined by a disulfide bridge. There is an intramolecular disulfide bridge between Cys8

179

and Cysll4 and an intermolecular bridge between Cysll6 of apoD and Cys6 of apoA-I1 [32]. N-glycosylation sites were found at Asn 45 and Asn 78. Apolipoprotein F has a molecular mass of 29 kDa and is composed of 162 amino acids. The cDNA sequence predicts that apolipoprotein F is a proteolytic product of a larger protein [33]. 2.5. Ceruloplasmin Ceruloplasmin is an important copper-containing glycoprotein enzyme (ferroxidase) involved in oxidation of Fe2+ to Fe3+ so that it can be transported by serotransferrin. Some mutations in the ceruloplasmin gene are associated with severe iron metabolism and distribution disorders leading to massive iron accumulation in the liver, brain and pancreas as well as to retinal problems and diabetes. Ceruloplasmin also has an important anti-oxidant function in protecting tissue against damage caused by free oxygen radicals. It is secreted as a holoprotein with six atoms of copper incorporated and has a molecular mass of 132 kDa and 8% carbohydrate content. Seven potential N-glycosylation sites exist but only four of these are occupied by oligosaccharides of the di- or triantennary type on which fucose may be present (Fig. 4) [34,35]. Differences in the oligosaccharide structure are responsible for the microheterogeneity on isoelectric focussing. NeuSAc(a2-6)Gal(fi14)GlcNAc(fi1-2)Man(a 1-3)

\

Fuc(al-6),,-,

Man(fiI4)GlcNAc(fi14)GlcNAc(fil -N)Asn

NeuSAc(a2-6)Gal(fiI 4)GlcNAc(flI-2)Man(a 1-6) Fuc(al-3)

,

(A)

o-l

\

Neu5A~(a2-3/6),_~Gal(fi14)GlcNAc((ll4)

\

NeuSAc(a24)Gal(fi14)GlcNAc(fil-2)Man(a 1-3)

\ Man(fi14)GlcNAc(fi 14)GlcNAc@-N)Asn

(B)

Neu5Ac(a2-6)GaI(fl14)GlcNAc(fiI-2)Man(a14~ Fig. 4.(A) Di- and (B) tnantennary glycans from cemloplasmin

2.6. CI inhibitor

Carbohydrate chains of Cl-inhibitor were identified by a binding assay using different lectins [36]. Lectins from Sumbucus nigru (SNA) and Muuckiu umurensis (MAA), that are specific for sialic acids, were shown to bind to C1-inhibitor. Lectin from Duturu stramonium (DSA) also reacted with the inhibitor indicating complex and hybrid sugar structures. C l-inhibitor was enzymatically desialylated and reexamined for lectin binding. SNA and MAA did not react anymore, but in addition to DSA, peanut agglutinin, which can bind to carbohydrate chains after sialic acids are removed, bound to desialylated C1-inhibitor. C l-inhibitor contains about 30 sialic acid residues per molecule. SDS-

180

NeuSAc(a24)Gal@ 14)GlcNAc@1-2)Man(a 1-3)

\ Fuc(a1-6), 0.3 Man(@14)GlcNAc(P14)GlcNAc(P1-N)Asn

(A)

I

NeuSAc(a2-3)Gal(fi 14)GlcNAc(P1-2)Man(a 1-3)

\

Fuc(a1-6),

0.3

Man(~I4)GlcNAc(@I4)GlcNAc(~I-N)Asn(9) Neu5Ac(a2-3)Gal(f1 14)GlcNAc(@1-2)Man(a 1-6f Neu5Ac(a2-3)Gal(f1-3)GalNAcol

(C)

Fig. 5. (A,B) Diantennary glycans and (C) 0-glycan from CI inhibitor

polyacrylamide gel electrophoresis showed that desialylated C 1-inhibitor had a higher mobility than native C 1-inhibitor. The N-terminal sequence of desialylated C 1-inhibitor was the same as that of native C1-inhibitor and no change in the inhibition of human plasma kallikrein was observed. The structure of the N- and 0-linked glycans of C1 inhibitor have been established by NMR spectroscopy and are shown in Fig. 5 [37]. 2.7. Complement 2.7.I . Complement C3 Of the 30 distinct complement proteins recognized to date, C3 is probably the most versatile and multifunctional molecule known, interacting with at least 20 different proteins. It plays a critical role in both pathways of complement activation and participates in phagocytic and immunoregulatory processes. Structural and functional analysis of C3 from different species, in addition to phylogenetic information, provides insights into the structural elements mediating the various functions. cDNA for human C3 has been cloned and its functional sites analyzed [38]. C3 consists of an a- (992 amino acids) and a fi-chain (645 amino acids) with a calculated molecular mass of about 181.5 kDa. The sequence contains three potential N-glycosylation sites: two on the a- and one on the b-chain of C3. One site on each chain is known to be glycosylated. 2.8. Factor J complement inhibitor Factor J (FJ), a new inhibitor of the complement system is a cationic molecule (PI 3 9.6 under native conditions, or PI = 8.1 under denaturing conditions) with a high carbohydrate content (40%) that is able to interact with different lectins, suggesting complex glycosylation [39]. Digestions with different proteinases did not affect activity. After b-glucuronidase digestion, FJ lost 80% of its initial activity. Consequently, glycosylation plays an important role in the inhibitory activity of FJ. 2.9. Factor V Coagulation factor V is a glycoprotein essential for haemostasis by accelerating the activation of prothrombin. Factor V is secreted as a single-chain polypeptide of 330 kDa

181

and is cleaved by the serine proteases thrombin or factor Xa to release B domain peptides of 71 and 150kDa to give the active species, factor Va. Factor Va contains the aminoterminal 94 kDa fragment associated with the 74 kDa carboxy-terminal fragment held together by divalent cations. The plasma concentration is about 8pg/ml and it is also found in platelet a-granules. The B-domain of factor removed during activation, contains 25 potential N-glycosylation sites. Inhibition of N-glycan addition with tunicamycin prevented secretion of factor V but an inhibitor of complex oligosaccharide addition, deoxymannojirimycin, did not affect secretion, although the specific activity of factor V was slightly increased. Thus, complex oligosaccharide addition was not required for secretion or functional activity of factor V. A23187 treatment inhibited addition of serinehhreonine 0-linked glycans to factor V [40].

2.10. Factor VII Factor VII is a single-chain, multidomain, vitamin K-dependent, plasma glycoprotein (50kDa) involved in the extrinsic pathway of blood coagulation. The plasma concentration is about 2pglml. Single-chain factor VII is converted into the two-chain serine protease factor VIIa by cleavage at Arg 152-Ile 153 by several coagulation proteases such as XIIa, IXa, Xa or thrombin. Factor VII is N-glycosylated at Asn 145 and Asn 322 [41] and 0-glycosylated at Ser 52 and Ser 60 in an EGF-like domain. The Ser 52 linked oligosaccharide was shown to be Xyll-zGlc-. A single fucose residue was found linked to Ser 60. Bovine factor VII is similarly glycosylated [42].

2.11. Factor VIII Factor VIII is a complex, plasma glycoprotein involved in blood coagulation and is processed intracellularly to yield a metal-ion-associated heterodimer of three chains, 85, 89 and 93 kDa, respectively, stabilized through association with von Willebrand factor. The plasma concentration is about 0.2 pg/ml. Deficiency leads to haemophilia A. Factor VIII is also deficient in certain types of von Willebrand’s disease due to the lack of the carrier molecule and may be responsible for some of the symptoms in that disorder. The asparagine-linked carbohydrate chains of blood coagulation factor VIII preparations were released as oligosaccharides by hydrazinolysis and purified from human plasma of blood group A donors [43]. Structural study of the oligosaccharides by sequential exo- and endoglycosidase digestion and by methylation analysis revealed that factor VIII preparations contain mainly high mannose type and di-, tri-, and tetraantennary complextype glycans (Fig. 6). Some of the diantennary complex-type glycans from human plasma factor VIII contain blood group A and/or H determinants. A small number of the triantennary complex-type glycans contain the Gal@ 14)[Fuc(a 1-3)]GlcNAc(fll4)[Gal(~l-4)GlcNAc(fll-2)]Man(a1-3) group. The Asn-linked glycans of factor VIII purified from porcine plasma were released as oligosaccharides by hydrazinolysis [44]. Structural study of each oligosaccharide by sequential exoglycosidase digestion and by methylation analysis revealed that porcine factor VIII, like the human, contains high mannose-type and di-, tri- and tetraantennary complex-type glycans. Sixty-seven percent of the complex-type glycans contained the

182

I

Gal(fil4)GlcNAc(~l-2)Man(al-3)

\

Fuc(al-6)

Man(~l4)GlcNAc(~14)GlcNAc(BI-N)Asn(A)

Neu5Ac(a2-3/6)0-2

/

Fuc(a 1-2)Gal(fi14)GlcNAc(fiI -2)Man(a 1-6) Gal@ I4)GlcNAc(fi I -2)Man(a 1-3)

\

Fuc(a1-6)

\

0-1

Man(BI4)GlcNAc(fiI4)GlcNAc(fiI-N)Asn (B) Gal@ 14)GlcNAc(fil-2)Man(al4)

I

Gal(~l4)GlcNAc(~l-2)Man(al-3) Gal(a 1-3)l-z

\

Neu5Ac(a2-3/6)0-1 Gal@ I4)GlcNAc(B I -2)Man(a 1-6)

I

Man(al-6) Man(a1-3) / Man(al-3)

I Gal@ l4)GlcNAc(fi 1-3)o-1 Neu5Ac(a2-316)&3

\

0-1

(C)

/

\ Man(Bl4)GlcNAc(~l4)GlcNAc(fiI-N)Asn(D)

Man(al-2)o~

1

Fuc(al-6)

Man(fil4)GlcNAc(fi I 4)GlcNAc(BI -N)Asn

/

Gal(fi14)GlcNAc(fi1-2)Man(a I -3/6) Gal@ I 4)GlcNAc(B 1-6)

\ Gal(~l4)GlcNAc(fil-2)i Gal(Bl4)GlcNAc(Bl 4),

\

Fuc(al-6)

1 . 0

\

Man(Bl4)GlcNAc(BI4)GlcNAc(BI-N)Asn

(G)

I

Man(a I -3/6)'

I Gal@ I4)GlcNAc(fi1-2)' Gal((3I4)GlcNAc(fi I-2)Man(al-6/3)

\ Ga1(~14)GlcNAc(~14) \

Man(al-316)

Fuc(al-6)

\

0-1

Man(B14)GlcNAc(B14)GlcNAc(fiI-N)Asn (H) /

I

Gal(B14)GlcNAc(f31-2)

Gal(fil4)GlcNAc(fi1-2)Man(al-6/3)

\

Fuc(al-6) o-l \

Man(fi I 4)GlcNAc(B I 4)GlcNAc(B I -N)Asn Gal(fiI4)GlcNAc(fiI4) / \ / Fuc(a1-3) Man(al-3/6)

(I)

/

Gal@14)GlcNAc@I -2)

Fig. 6 . (A-C,E) Di-, (F,H,I) tn- and (C) tetraantennary complex type and (D) high mannose type triantennary glycans from factor VIII.

183

Gal(a1-3)Gal group and 23% of the diantennary complex-type glycans contained the bisecting GlcNAc residue. These structures were not detected in the glycans of human plasma factor VIII. In uitro competition of von Willebrand factor and anti-Gal antibody for binding to factor VIII revealed that von Willebrand factor prevented antibody binding to Gal(a 13)Gal groups in porcine factor VIII glycans. This suggests that anti-Gal antibody present in human plasma may not interact with the glycans of porcine factor VIII used in treatment of haemophilia. Studies of pharmacokinetic parameters of recombinant factor VIII infused into baboons revealed that its half-life in blood circulation is similar to that of plasma-derived factor VIII, suggesting that the oligosaccharide structural differences between them do not affect the fate of factor VIII in uiuo [43]. In transfected mammalian cells inhibition of N-glycan addition by treatment with tunicamycin prevented secretion of factor VIII, whereas treatment with an inhibitor of complex oligosaccharide biosynthesis, deoxymannojirimycin, did not affect secretion, although the specific activity of factor VIII was slightly increased [45]. Thus, the presence of complex oligosaccharide was not required for secretion or functional activity of factor VIII. A23 187 treatment inhibited addition of serinehhreonine (0)-linked glycans to factor VIII. Factor VIII has been expressed in Spodopterufrugiperdu insect cells [46]. The construct retained the native signal sequence to allow secretion of recombinant protein into the culture medium. Initial studies revealed the production of secreted factor VIII with coagulation activity. The presence of N-glycans was demonstrated since the glycosylated molecule is similar in size to that expressed in mammalian cells.

2.12. Factor IX Factor IX (Christmas factor) is a plasma glycoprotein with a critical role in blood coagulation. Deficiency causes haemophilia B or Christmas disease, a severe bleeding disorder. Factor IX has a mass of 55 400 Da and activated IXa 46 500 Da. The normal concentration in plasma is 3 4 y g / m l . 0-Glycans have been identified in the activation peptide of human blood coagulation factor IX [47]. Gal-GalNAc-Thr, NeuNAc-(Gal-)GalNAc-Thr, and NeuNAc-Gal-GalNAc-Thr structures were 0-glycosidically linked to Thr 159 and 169 present on 35% of the total amount of activation peptide in circulating blood. 0and N-linked glycans were also released with hydrazine and analyzed [48]. Glycans were identified as mono- and disialyl Gal@ 1-3)GalNAc by two-dimensional HPLC mapping and as NeuSAc(a2-6)Gal(P 14)GlcNAc@I-3)Fuc by exoglycosidase digestion, methylation analysis, and Smith degradation. Carbohydrate composition and mass spectrometric analyses of tryptic and thermolytic peptides containing Ser 61 in the first EGF domain of human factor IX indicated the presence of a tetrasaccharide containing one residue each of sialic acid, galactose, N-acetylglucosamine, and fucose [49,50]. This structure attached to Ser 61 was also shown by hydrazinolysis of a peptide from this region followed by pyridylamination indicating that the reducing end was PA-Fuc. The results indicated that human factor IX has a novel tetrasaccharide linked to Ser 61 through the Fuc residue. Mass spectrometric analysis indicated that fucose was the attachment sugar

184

residue. The Ser 6 1 tetrasaccharide was not susceptible to a-fucosidase digestion. The complete structure of the tetrasaccharide was obtained by methylation and NMR analysis as NeuAc(a24)Gal(~14)GlcNAc(fll-3)Fuca( 1-0)-Ser6 1. Unusual glycans Xyll-2-Glc were shown to be linked to a serine residue (Ser 53) in the epidermal growth factor (EGF)-like domains of both human and bovine factor IX.

2.13. Factor X Factor X is a plasma protein involved in both the intrinsic and extrinsic pathways of blood coagulation. Factor X has a mass of 55 kDa and the activated Xa of 40kDa. The normal concentration in plasma is 6-8 pg/ml. Post-translational modifications of the protein involve y-carboxylation of specific glutamic acid residues, fl-hydroxylation of one aspartic acid residue, and N- and 0-linked glycosylation. Human blood coagulation factor X has two N-linked oligosaccharides at Asn 39 and Asn 49 and two 0-linked oligosaccharides at Thr 17 and Thr 29 in the region of the factor X activation peptide which is cleaved off during its activation by factor IXa. The structure of the oligosaccharides in the activation peptide region of human factor X has been determined[51]. The content of the neutral oligosaccharides at Asn 39 and Asn 49 residues were 32.5% and 30.0%, respectively. Six neutral and twelve monosialyl oligosaccharides isolated from both N-linked glycosylation sites showed similar elution profiles composed of di-, tri- and tetraantennary complex type oligosaccharides. The predominant component in neutral oligosaccharides was diantennary without a fucose residue. Two major monosialyl oligosaccharides were also diantennary without fucose and with a Neu5Ac(a2-6) residue. In addition, the structures of 0-linked oligosaccharides at Thr 17 and Thr 29 were suggested to be disialylated Gal(fl1-3)GalNAc sequences by component analysis. The effect of deglycosylation of bovine factor XI has been investigated with factor-Xactivating enzyme from Russell’s viper venom or extrinsic Xase (factor VIIdtissue factor/phospholipid) by examining the activation rates of derivatives of factor X prepared using 0-glycanase, sialidase, andor N-glycanase [52]. The removal of 0-linked carbohydrate resulted in a decrease in the rate of activation. Lectin binding and glycosidase treatment were also used [53] to investigate the functional role of carbohydrates on the activation peptide of factor X. Sumbucus nigru agglutinin, a lectin that binds to sialic acid a(24)-linked to galactose or N-acetylgalactosamine inhibits activation of human factor X in a dose-dependent manner. Inhibition of activation was observed for both intrinsic (factor IXa/VIIIa) and extrinsic (factor VIIa/tissue factor) pathway complexes. In accordance with this, removal of sialic acid residues from the activation peptide of factor X by sialidase also drastically reduces activation of the zymogen by these complexes. Parallel reduction of activity in classical clotting assays (activated partial thromboplastin time and prothrombin time) corresponds with this observation. These results also suggest a possible role of N-glycans in the activation of factor X. Thus, carbohydrate residues in factor X may play an important role in the activation of the zymogen.

185

2.14. Factor XI

Human factor XI is a glycoprotein composed of two identical chains linked by disulfide bonds with a molecular mass of 124 kDa. Normal plasma contains 4 yglml. Activation of factor XI by factor XIIa involves cleavage of both chains to 35 kDa and 25 kDa fragments. The active site serines are in the 25 kDa chains. Factor XIa thus consists of four subunits linked by disulfide bridges.

2.15. Factor XI1 Factor XI1 is a single polypeptide chain with a mass of 80 kDa, fully fucosylated at Thr 90 of the heavy chain in the N-terminal EGF domain [54]. The plasma content is 29 yg/ml.

NeuSAc(a2-6)Gal@ 1 -4)GlcNAc(B 1-2)Man(a 1-6)

\ Man(f3I4)GlcNAc(f314)GlcNAc-Asn (A) NeuSAc(a24)Gal(f314)GlcNAc(f31-2)Man(a 1-3)

1

Neu5Ac(a24)Gal(f314)GlcNAc(f3I-2)Man(a 1-6)

\ GlcNAc(f314)-Man(p 14)GlcNAc(f31 -4)GlcNAc-Asn I Neu5Ac(a24)Gal(f314)GlcNAc(f3l-2)Man(a 1-3)

(B)

Fig. 7. (A) Concanavalin A binding and (B) non-binding di- and triantennary glycans from a-fetoprotein.

2.16. a-Fetoprotein a-Fetoprotein is found in relatively high concentrations in the fetal and neonatal sera and in amniotic fluid of many species but it drops to low levels soon after birth. Distinct molecular variants of rat a-fetoprotein have been detected using Con A affinity chromatography [ 5 5 ] . Each a-fetoprotein variant contains two identical glycans which differ between the Con A reactive (Fig. 7A) and non-reactive forms (Fig. 7B).

2.17. Fibrinogen Fibrinogen is a three-chain protein synthesized in the liver that is essential for haemostasis both in linking together activated platelets in aggregation and in the formation of the insoluble polymer fibrin necessary for the consolidation of the haemostatic plug. Fibrinogen is glycosylated on two sites, one in the b- and one in the y-chain, both on Asn. Both glycans were determined to be diantennary structures (Fig. 8) [56]. The primary structures of two Asn-linked glycans from bovine fibrinogen have been determined by methylation analysis and NMR spectroscopy, and were also shown to be diantennary [57].

186

Gal(fil-4)GlcNAc(~l-2)Man(al-6)

\ Man(fil4)GlcNAc(fi 14)GlcNAc-Asn I

Gal(fil4)GlcNAc(fil-2)Man(a1-3) Fig. 8. Diantennary glycan from fibrinogen.

2.18. Fibronectin The fibronectins consist of isohomodimers of two nearly identical cysteine-linked 225 kDa subunits. They are adhesive proteins and can bind to a range of molecules including denatured collagen, fibrin and DNA. They are also important for cell adhesion. The carbohydrate content of fibronectins varies considerably depending on the source and is thought to be important for certain physical properties [58]. The extracellular matrix adhesion molecule fibronectin exhibits different isoforms caused by alternative splicing as well as by variation in 0-glycosylation, as recently demonstrated. Although fibronectin is widely distributed in normal tissues, the individual isoforms have been found to show restricted tissue distribution and association with malignancies. The monoclonal antibody FDC-6 defines a cancer-associated de nouo glycosylation of a specific Thr residue in the C-terminal region of the fibronectin molecule termed oncofetal fibronectin. Oral squamous cell carcinomas, premalignant lesions, and normal oral mucosa have been studied immunohistologically using the FDC-6 antibody [59]. Selective expression of the oncofetal fibronectin epitope was closely related to the invading carcinoma. Previous studies indicated that the de nouo glycosylation is induced by a novel transferase activity only found in fetal and carcinoma cell lines, placenta and hepatoma tissues. The gelatin-binding region of fibronectin contains three Asn-linked carbohydrate moieties, one on the second type I1 module and two on the eighth type I “finger” module. Carbohydrate groups were enzymatically removed from two non-overlapping gelatin-binding fragments (GBFs), 2 1 kDa GBF (modular composition 18-19) and, with much greater difficulty, 30 kDa GBF (modular composition 16-111-112-17) [60]. Fluorescence and calorimetric analyses indicated that module I8 was strongly destabilized by deglycosylation so that the apoform melts near physiological temperatures. A similar effect was caused by decreasing the pH of the holoform to 6.0, suggesting that one or more histidines are important for stability of module 18. The 21 kDa fragment exhibited an acid-induced change in fluorescence that occurred at higher pH in the deglycosylated derivative, providing further evidence of a stabilizing role for one or both carbohydrate moieties. In contrast, the stability of module I12 was unaffected by removal of its single carbohydrate. To determine if differential glycosylation of fibronectin (Fn) in inflammatory synovial fluid (SF) included expression of an oncofetal epitope (Onf Fn) previously detected only on Fn derived from embryonal or neoplastic tissue [61], Fn was purified from plasma, SF and synoviocyte conditioned medium by affinity chromatography and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blotting using a monoclonal antibody (FDC-6) specific for the Onf Fn. The Onf Fn was not expressed on

187

Fn isolated from normal or rheumatoid arthritis (RA) plasma but was strongly expressed on Fn from RA SF and to a lesser extent osteoarthritis SF. Onf Fn was also detected on Fn secreted by cultured RA synoviocytes. 2.19. Hemopexin Hemopexin has a molecular mass of 57kDa and is a glycoprotein involved in haem disposal present in plasma at levels between 8 and lOrng/lOOml. It contains about 22% carbohydrate. The human glycoprotein contains five N-linked and one 0-linked oligosaccharide chain, the glycans are of the di- and triantennary type. The 0-linked glycan blocks the N-terminal threonine [62]. Porcine hemopexin has been isolated in >99% purity as tested by crossed immunoelectrophoresis. Porcine hemopexin has a molecular mass of 62kDa. Based on carbohydrate and sialic acid analyses, it was proposed that hemopexin contains two diantennary (similar to Fig. 7A) and one triantennary glycan chains [63]. The structure of several rat hemopexin glycans were determined and the main component was shown to be a trisialyl diantennary oligosaccharide (Fig. 9) [64]. NeuSAc(a24) \

NeuSAc(a2-6)Gal(fi 14)GlcNAc(fi1-2)Man(a 1-6)

\ Man(fi 1-4)GlcNAc(fil4)GlcNAc-Asn

NeuSAc(a2-6)Gal(fi14)GlcNAc(fi 1-2)Man(a 1-3)

I

Fig. 9. Trisialyl diantennary glycan from hemopexin.

2.20. a2-HS-glycoprotein a2-HS-glycoprotein is the human equivalent of fetuin in animals. Complex type N-glycosides of bovine fetuin contain comparable amounts of triantennary (Fig. 1OA) and tri’-antennary (Fig. 1OB) as major asialo-structures. A specific galactosyltransferase may exist for the biosynthesis of the tri‘-antennary structure. The N-glycosides of ovine fetuin also have both triantennary and tri’-antennary structures in a ratio similar to that of bovine fetuin. However, the major N-glycoside of porcine fetuin has a fucosyl diantennary complex type structure (Fig. 1OC) and human a2-HS-glycoprotein has an N-glycoside which is almost exclusively a nonfucosylated diantennary structure (Fig. 1OD). This species-specific presence of N-glycosides of fetuins and comparison with N-glycosides of other glycoproteins suggest that the polypeptide sequence of a glycoprotein may affect its N-glycan structure by regulating the activity of specific glycosyltransferases [65]. The A-chain of human plasma a2HS-glycoprotein contains two diantennary N-glycans linked to Asn 138 and 158 (Fig. 10E) and two 0-linked trisaccharides of the types shown in Fig. 10F linked to Thr 238 and 252 [66]. The B-chain has one 0-linked trisaccharide of the same type. The N-glycans are sialylated to about 90%.

188

Gal(@14)GlcNAc(fil-2)Man(a 1 4 )

\ Man@1-4)GlcNAc(fi 14)GlcNAc (A) I

Gal(fi14)GlcNAc(fi 1 -2)Man(al-3) Gal(BI-4)GlcNAc(fi1-4)1 Gal((ll4)GlcNAc(fi 1-6)

\

Gal(fil4)GlcNAc(fiI-2)Man(a1-6)

\ Gal@1-4)GlcNAc(fi 1-2)Man(a 1-3)

Man(Bl4)GlcNAc(fiI 4)GlcNAc (B) I

Gal@ 14)GlcNAc(fi1-2)Man(a 1 4 )

\

Fuc(a 1-6)

\

Man(fi14)GlcNAc(@I4)GlcNAc (C)

Gal(fi14)GlcNAc(fi1-2)Man(a 1-3)

I

Gal(fi14)GlcNAc(fi1-2)Man(a 1-6)

\ Gal(fi14)GlcNAc(B 1 -2)Man(a 1-3)

Man(fil-I)GlcNAc(pl 4)GlcNAc (D) I

NeuSAc(a24)Gal(B 14)GlcNAc(fi1 -2)Man(a 14)

\ Man@ 14)GlcNAc(@1 4)GlcNAc(@I-N)Asn I

(E)

NeuSAc(a2-3)Gal(fi 1-4)Gal(@1-3)GalNAc(a 1-0)

(F)

NeuSAc(a2-6)Gal(~14)GlcNAc(fi1-2)Man(a 1-3)

Fig. 10. (C-E) Di- and (A,B) triantennary glycans and (F) 0-linked tetrasaccharide from a l - H S glycoprotein,

2.21. a2-Leucine-rich glycoprotein a2-Leucine-rich glycoprotein is the prototype of the leucine-rich domain family containing 13 repeats of a 24 amino acid sequence [67]. Plasma contains 2.1 mg per 100 ml; the function is unknown. The consensus sequence is LXXLXLXXNXLXXLPXXLLXXXXX, the first part of which forms an @-fold and the second part an a-helix. Based on the crystal structure of porcine ribonuclease inhibitor it can be predicted that these repeats will pack together in an arc-like form. Four N-glycans and one 0-glycan are present. The 0-glycan is close to the N-terminus. One potential N-site is not glycosylated.

2.22. a2-Macroglobulin a2-Macroglobulin is a major plasma component with nearly 300mg/100ml and a high molecular mass (820 kDa) containing about 10% carbohydrate in 3 1 glycans [68]. It has an important role in controlling protease activity in the blood. No complete oligosaccharide

189

structures have been reported yet. Abnormal glycosylation of a2-macroglobulin, a nonacute-phase protein has been detected in various pathological conditions including autoimmune diseases. The protein was purified from serum samples and cerebrospinal fluids (CSF) from patients with autoimmune diseases: systemic lupus erythematosus (SLE), rheumatoid arthritis (RA), mixed connective tissue disease (MCTD), scleroderma (SCL), Sjogren’s syndrome (SS), and polymyositis (PM); diseases of probable autoimmune origin: hepatopathies (HP); diseases of suspected autoimmune origin: schizophrenia and Alzheimer’s disease (AZ); and conditions not related to autoimmunity: pregnancy and elevation of the carcinoembryonic antigen (CEA) as well as from normal donors. Con A-reactive fragments on Western blots increased specifically and significantly in samples derived from patients of SLE, SCL, MCTD, and RA [69]. Linkage-specific lectins revealed the presence of Neu5Ac(a2-6)Gal but not Neu~Ac(a2-3)Gal in porcine a2-macroglobulin glycan [70].a2-Macroglobulin was found to be microheterogeneous in plasma samples from alcoholic patients but microheterogeneity disappeared after treatment with sialidase [24]. 2.23. Plusminogen

Plasminogen is the precursor of the serine protease plasmin which is involved in dissolution of fibrin clots and, hence, in removal of thrombi. The plasma concentration is about 12 mg/100 ml and the molecular mass 81 kDa. It contains 17.1% carbohydrate on two sites, one N-linked and one 0-linked oligosaccharide. A variant is also known that is only 0-glycosylated. The structures of the oligosaccharides have been determined [7 I731, and are shown in Fig. l l . NeuSNAc(a24)Gal(fi1-4)GlcNAc(fi1-2)Man(a1-6/3)

\ Man@ 1-4)GlcNAc(fi 14)GlcNAcAsn288 (A) NeuSNAc(a2-6)GaI(fi 1-4)GlcNAc(fi 1-2)Man(a 1-316)

NeuSAc(a2-3)Gal(fi1-3)GalNAcThr345 (B) I NeuSNA~(a2-6)~_,

Fig. 1 1. (A) Diantennary glycan and (B) 0-linked branched tetrasaccharide from plasminogen.

2.24. al-Proteinuse inhibitor al-Proteinase inhibitor is a secreted, monomeric protein glycosylated at Asn 46, 83, and 247. Single, double, and triple mutants have been made by changing the codons specifying these Asn residues to encode Gln. The mutant proteins were transiently expressed in COS-I cells. All variants with altered glycosylation sites are secreted at reduced rates, are partially degraded, accumulate intracellularly, and some form Nonidet P-40-insoluble aggregates. The carbohydrate attached at Asn 83 seems to be of particular importance to the export of both al-proteinase inhibitors M and S from the endoplasmic reticulum. All

I90

mutations affecting glycosylation of a 1-proteinase inhibitor S notably reduce secretion, cause formation of insoluble aggregates and influence degradation of the altered proteins. The variant of al-proteinase inhibitor S lacking all three glycosylation sites is poorly secreted, is incompletely degraded, and accumulates in unusual perinuclear vesicles [74]. Thus, N-linked oligosaccharides in a l -proteinase inhibitor S are vital to its efficient export from the endoplasmic reticulum and subsequent processing. 2.25. Protein C Protein C is a glycoprotein with a molecular mass of 62kDa present in plasma at a concentration of 4 pg/ml. Thr 68 of Protein C in the N-terminal EGF domain is not fucosylated but the Asp 71 is fully 6-hydroxylated [54]. Protein C exists in a- (70%) and 6-forms (30%) in plasma. The f3-form is smaller than the a-form. This difference appears to be due to the presence of four N-glycans on the a-form but only three on the f3-form. The fourth site at Asn 329 (Asn-Ala-Cys) is different from the usual Asn-X-Ser/Thr but corresponds to N-glycosylation sites in von Willebrand factor and bovine Protein C. The likelihood of glycosylation at this site may be related to the speed of formation of the Cys33 1-Cys345 disulfide bond [75].

2.26. Protein S Protein S contains three potential N-glycosylation sites at Asn 458,468 and 489. Mutation studies indicate that each site is glycosylated to the same extent and that the blockage of glycosylation generally enhances the cofactor activity [76]. 2.27. Prothrombin

Prothrombin is the precursor of the serine protease thrombin, essential for normal haemostasis in activation of platelets and in catalysing the polymerisation of fibrinogen to fibrin. It is synthesized in the liver and contains about 8% total carbohydrate in four N-glycans. Three of these lie on the Pro region and only one is present on the mature a-thrombin molecule. The structure of the thrombin glycan has been determined by mass spectrometry and is the same as that shown in Fig. 5A with fucose absent or present [77].

2.28. Transferrin The function and structure of the oligosaccharides on the iron-binding transferrin family of proteins is dealt with in detail in Chapter 10. 2.29. Vitronectin

Vitronectin is a glycoprotein with a mass of 70kDa involved in cell adhesion and spreading. Plasma contains about 200-300 pglnil. Vitronectin contains three potential N-glycosylation sites. The structures of the N-linked oligosaccharides present on human plasma vitronectin have been determined [78]. Oligosaccharides were released from the

191

vitronectin by PNGase F digestion and tagged with 2-aminopyridine. The pyridylaminooligosaccharides were then fractionated by anion-exchange and reverse-phase HPLC. Ten major pyridylamino-oligosaccharides were isolated. The linkages and locations of sialic acid residues were determined by desialylation with Salmonella sialidase in combination with acid. The asialo forms were then analyzed by two-dimensional carbohydrate mapping, component monosaccharide analysis and 400 MHz H-NMR spectroscopy. The major oligosaccharides of human vitronectin were of the diantennary N-acetyllactosamine type, with less of the tri- and a small amount of the mono-antennary type, to which one to three sialic acid residues were linked, mostly a(2-6)-linked, although a(2-3) linkages were also present. Several binding activities of vitronectin may be related to its glycan moiety based on the specific features of the N-glycans. The structures of N-linked oligosaccharides on porcine plasma vitronectin have been similarly determined [79]. Nine major pyridyl-amino-oligosaccharides were isolated. After desialylation, the asialo-forms were analyzed by two-dimensional sugar mapping, component sugar analysis and 400 MHz H-NMR spectroscopy. The major oligosaccharides of porcine vitronectin were of the fucosylated diantennary type, with a small amount of the triantennary N-acetyllactosamine type, to which one to three sialic acids residues were linked. Sialic acids were predominantly a(2-6)-linked, although a(2-3) linkages were also present, and fucose was linked to the innermost N-acetylglucosamine through an a ( 1-6) linkage. Each pyridylamino-oligosaccharide population contained NeusGc and NeusAc in a molar ratio of 1 :2-9, and NeuSGc were located predominantly on the Man(a1-6) antenna.

'

'

2.30. Von Willebrand factor Von Willebrand factor is a very large (1-1 0 MDa) multimeric protein synthesized in endothelial cells and in megakaryocytes. It is constitutively secreted from endothelial cells and is also stored in granules in these cells and in platelets and released when they are activated, for example, by thrombin. Von Willebrand factor (vWf) plays an essential role in haemostasis by acting as the carrier molecule for factor VIII. It also binds to collagen on vascular subendothelium and provides the attachment site for circulating platelets in the first step of primary haemostasis. The receptor for von Willebrand factor on platelets is the GPIb-V-IX complex. Several domains on vWf have been identified as involved in binding to GPIb, collagen and heparin. Six 0-glycans flank the A1 domain cystine loop containing the GPIb binding domain. Removal of sialic acid produces asialo-vWf which binds spontaneously to platelet GPIb. These results indicate that the active structure of the GPIb binding domain (the A1 domain) of vWf is dependent on glycosylation [go]. The structure of some of these oligosaccharides has been determined. A major N-glycan representing about 45% of the carbohydrate chains is a monosialylated monofucosylated diantennary glycan of the N-acetyllactosaminic type (Fig. 12A) [8 11. A tetraantennary glycan of the N-acetyllactosaminic type has also been isolated and its structure determined by methylation analysis and 500 MHz 'H-NMR (Fig. 12B) [82]. The major 0-glycan of human vWf was shown to be a diantennary tetrasaccharide (Fig. 12C) [83].

192

NeuSAc(a24)Gal(fi I 4)GlcNAc(fi I -2)Man(a 1-6)

\ Man(fiI4)GlcNAc(fi14)GlcNAc (A) I

Gal(fiI-rl)GlcNAc(fil-2)Man(aI-2)

I Fuc(a 14)

Gal@ l 4 ) G l c N A c ( f i l 4 ) \ Man(al-3)

\

Neu5Ac(a2-6)Gal(fi 14)GlcNAc(!3 1 -2f Gal(fil4)GlcNAc(fi1-2)~

Gal@ l4)GlcNAc(fi1-6)

Fuc(a 1-6)

\

Man(fil4)GlcNAc(fi14)GlcNAc (B)

Man(a 1-6f I Neu5Ac(a24)

\ NeuSAc(a2-3)Gal(b 1-3)

GalNAcol (C) I

Fig. 12. (A,B) Di- and tetraantennary glycans and (C) 0-linked branched tetrasaccharide from von Willebrand factor.

3. Platelets The basic role of platelets is haemostasis, or stopping bleeding, when tissue is damaged. The platelets adhere to exposed subendothelium on vessel walls, are activated and aggregate to form a thrombus that covers the damaged area or blocks off complete vessels when they are involved. These functions involve a number of receptors, all of which are membrane glycoproteins. Platelets are very rich in glycoproteins and contain some 11 times more sialic acid than erythrocytes.

3.1. Glycoprotein Ib- FIX complex This complex is responsible for the primary adhesion of platelets to von Willebrand factor as well as modulating the platelet response to thrombin. There are four separate subunits, GPIba and f~covalently linked by a disulfide bond, GPIX tightly bound non-covalently, and GPV more loosely bound non-covalently. GPIba contains the primary binding sites but the other chains seem to be necessary for physiological function. Absence of GPIbV-IX for genetic reasons leads to the bleeding disorder Bernard-Soulier syndrome. 3.1.1. Glycoprotein Iba/glycocalicin

Glycoprotein Iba is one of the major glycoproteins on the platelet surface containing a large part of the sialic acid present on the platelet surface. There are at least 25 000 copies per platelet but there may be up to 50 000 if molecules present in the surface-connected canalicular system are included. This chain contains the primary von Willebrand factor binding site and a thrombin-binding site. Glycocalicin is the extracellular domain which is readily split off by several proteases, particularly endogenous calpain and contains 40% carbohydrate by mass, all the glycosylation of GPIba. Glycoprotein Iba contains four

193 Neu5Ac(a2-3/6)Gal(fi 14)GlcNAc(fiI-2)Man(a1-6)

\

Fuc(al-6)

\

Man(fiI4)GlcNAc(fi14)GlcNAc(A) Neu5Ac(a2-3/6)Gal(fi14)GlcNAc(fil-2)Man(al-3)’ NeuSAc(a2-3/6)Gal(fi 14)GlcNAc(fi I-2)Man(a 1-6)

\

Fuc(al-6)

\

Man(/314)GlcNAc(fi14)GlcNAc (B)

Neu5Ac(a2-3/6)Gal(fiI 4)GlcNAc(fiI 4) \

Man(a 1-3)

I

NeuSAc(a2-3,6)Gal(fi 14)GlcNAc(fiI -2; NeuSAc(a2-3)Gal(fi 1 4)GlcNAc(/31-6)

\

Man(a 1-6)

Neu5Ac(a2-3)Gal(fiI4)GlcNAc(fi1-2)/ NeuSAc(a2-3)Gal(fi14)GlcNAc(f114),

\

Fuc(a I”),

\

Man(fi14)GlcNAc(fi14)GlcNAc(C)

NeuSAc(a2-3)Gal(fil4)GlcNAc(fi1-6)GalNAc (D) NeuSAc(a2-3)Gal(fiI -3)

I

Neu5NAc(a2-3)Gal(fiI-3)GalNAc (E)

NeuSAc(a2-3)Gal(fil-3)GalNAc (F) Neu5Ac(a24)

I

Fuc(a 1-2)Gal(fi 1-4)GlcNAc(fi 1-6)GalNAc I Gal(fi1-3)

(G)

Fig. 13. (A) Diantennary, (B) triantennary and (C) tetraantennary glycans and 0-linked (D) hexa-, (E) tri-, (F) tetra-, and (G) pentasaccharides from glycoprotein Iba.

putative N-glycosylation sites two of which are known to be glycosylated from primary sequencing data. The N-glycosylation structures were released by hydrazinolysis, isolated and structure determined by methylation analysis, glycosidase treatment and Smith degradation [84]. Typical di- and trisialylated complex-type structures were found and are illustrated in Figs. 13A and B. A tetrasialylated monofucosylated tetraantennary chain was later detected and its structure determined using NMR spectroscopy (Fig. 13C) [85]. 0-linked oligosaccharide alditols were obtained by alkaline borohydride treatment of glycocalicin and the structure of the major chain was determined by methylation and glycosidase treatments[86] or by NMR[87,88] and found to be a hexasaccharide as shown in Fig. 13D. Minor tri-, tetra- and pentasaccharides with the structures shown in Figs. 13E, F and G were also obtained. Glycocalicin contains a domain that is particularly rich in 0-glycosylation and includes five mucin-like repeats. Within this region is a size polymorphism consisting of a 13 amino-acid sequence [89] that is found either once or repeated twice, three or four times, termed D, C, B and A phenotypes, respectively. Since each repeat contains five putative 0-glycosylation sites the difference in molecular

194

mass of 2kDa between these is explicable. Although these differences are referred to as polymorphism, it remains unclear if there exist differences in function resulting from them. Thus, the function of this highly 0-glycosylated domain, which acts as a semi-stiff rod, appears to be to hold the receptor domain out from the platelet surface so that it is available for interaction with von Willebrand factor and thrombin. The fact that this may be extended more or less into the extracellular space may not be without consequences. The high glycosylation of this domain also functions to protect it against proteolysis which is restricted to small unprotected regions between domains. 3.1.2. Glycoprotein IbB The small subunit of GPIb with a molecular mass of 27 kDa and linked to the a-subunit by a disulfide bond has one leucine-rich repeat containing a single N-glycosylation site. Monosaccharide analysis showed that this was likely to be a diantennary glycan of the N-acetyllactosamine type. The lack of N-GalNAc argued against any 0-glycosylation [90]. 3.1.3. Glycoprotein IX Glycoprotein IX has overall a very similar structure to glycoprotein Ibp and is only slightly smaller with a mass of 22 kDa. It also has one leucine-rich repeat containing a single N-glycosylation site which also appears to be a diantennary glycan of the N-acetyllactosamine type. The lack of GalNAc also indicates that there is no 0-glycosylation [90]. 3.I . 4. Glycoprotein V Glycoprotein V is loosely non-covalently associated with the rest of the GPIb-IX complex, probably in a 1:2 ratio. It has a mass of 82kDa and contains 16 leucine-rich repeats, as well as disulfide-bridged loops that appear to be conserved among all the members of this family. An important feature is a thrombin-cleavage site just under these domains which separates the molecule into a 69 kDa soluble fragment and a 20 kDa membrane bound fragment. There are eight putative N-glycosylation sites of which seven are in the leucine-rich domain and one in the disulfide-bridged loops. Based on the molecular mass estimated from the protein backbone (59 m a ) , it seems likely that most, if not all of these sites are glycosylated. There is also evidence, based on primary sequence data, that some 0-glycosylation sites in the 20 kDa fragment near to the membrane are glycosylated [91]. However, the mucin-like repeats found in GPIba are not present in GPV: 3.1.4.1. Leucine-rich domains. The leucine-rich domains consist of a sequence of 24 amino acids forming a loop consisting of an ag-fold and an a-helix. These loops pack together to form arc-like structures or, in some cases with large numbers of repeats, horse-shoe-like or even circular, spiral structures. As mentioned above, several of the N-glycosylation sites in GPIb-V-IX are in the leucine-rich domains and lie without exception on the outside of the loops, generally on the outside face of the arc-like structure.

195

3.2. GIycoprotein IIb-IIIa

(aI1bp3)

Glycoprotein IIb-IIIa (allt$3), also known as CD41/61, is the major member of the integrin family on platelets and was, indeed, important for working out many of the structural and functional aspects of integrins. Integrins typically consist of two subunits an a and a B and form the link between adhesive proteins, such as are present in the extracellular matrix, and the cell cytoskeleton. Thus, they are often found to play an important role in cell adhesion and locomotion. In the case of platelets, GPIIb-IIIa is involved in secondary adhesion of activated platelets to vascular subendothelium. It is particularly important in platelet-platelet adhesion called aggregation, essential for formation of a stable thrombus in haemostasis, and for clot retraction, necessary to hold together the sides of a wound to allow repair to start. The a-subunits are characterized by the presence of four or more cation binding sites, critical for the conformation of the subunit and to allow interactions with the B-subunit and form the binding site. The b-subunit contains a domain rich in cysteine bridged loops but its function is not yet clear. GPIIb-IIIa belongs to the category of integrins that require activation to bind their ligand. In resting platelets it is to a large extent passive although a minor amount may always be in the active state. Activation occurs via changes in the platelet interior caused by signal transduction from other classes of receptor and seems to involve kinase/phosphatase cascades although phosphorylation of the integrin itself appears not to be necessary. There are about 50 000 copies of GPIIb-IIIa on the platelet surface but many are also present on the surface of a-granules and are exposed after platelet activation. Absence of this integrin for genetic reasons leads to the bleeding disorder, Glanzmann’s thrombasthenia.

3.2.1. Glycoprotein IIb

Glycoprotein IIb has a mass of 120kDa non-reduced and of 110 kDa reduced due to the separation of a large and small subunit held together by a disulfide bond. There are four potential N-glycosylation sites in the heavy chain and one in the light chain [92]. GPIIb contains mainly complex type glycans in contrast to GPIIIa accounting for the difference in binding to lectins [93]. The major types present are shown in Fig. 14.

I Neu5Ac(a2-3/6),

-3

I

I Neu5Ac(a2-3/6)

Gal(fiI4)GlcNAc(B14) \ Man(a1-6) Gal@ 14)GlcNAc(B 1-2) I

\

Fuc(a1-6)

\ -

I

Man(b14)GlcNAc@ 1 -4)GlcNAc

Gal(fi 14)GlcNAc(fi 1-2)Man(a 1-3)

(A)

t

Gal@ 14)GlcNAc(b 1-2)Man(a 1-6)

\

Fuc(a1-6)

\ -

,

M a n ( f i I 4 ) G l c N A c ( ~ l 4 ) G l c N A c (B)

I -2

Cial(fi14)GlcNAc((lI-2)Man(a 1-3)

I

196

Man(a 1 4 )

\

Man(a1-2),_,

Man(a 1-6) Man(a1-3) I \ Man(Bl4)GlcNAc(fi 14)GlcNAc I Man(al-3)

Fig. 15. High mannose type triantennary glycan from glycoprotein 1Ila.

3.2.2. Glycoprotein IIIa Glycoprotein IIIa has an apparent molecular mass on gel electrophoresis of 90 kDa nonreduced and 100 kDa reduced due to the large number of intramolecular disulfide bridges. Before reduction the molecule is very compact and after reduction much more stretched out. There are six potential N-glycosylation sites [94]. N-glycosylation on GPIIIa is predominantly of the high mannose type [93]. A typical structure is shown in Fig. 15. 3.3. D l Integrin family Little is known about the glycosylation on this family which is represented on platelets by GPIIa (CD29). From the lectin binding pattern it seems likely that it also contains principally high mannose-type glycans. The a-chains in this family are GPIa, GPIc and GPIc’ (CD49b, e, f). The structures of N-glycans obtained from human placenta integrin a5g1, corresponding to platelet GPIc-Ha, have recently been investigated [95]. A total of 35 different oligosaccharide structures were identified, 10 neutral, 6 monosialyl, 10 disialyl, 7 trisialyl and 2 tetrasialyl. High-mannose type glycans were only 1.5% of the total, the rest being all complex type and representatives of nearly all structures were found. The most prominent structure was the diantennary di-a-(2,3)-sialyl fucosyl corresponding to that shown in Fig. 5B. The major sialic acid linkage was a-(2,3) and 50% of all oligosaccharides were fucosylated at the reducing end GlcNAc. Tetraantennary structures of the same type as shown in Fig. 13C with and without fucose on the reducing terminus were also common.

3.4. CD36 (GPIIIb, GPIV) CD36 is a family of glycoproteins found on the surface of a wide range of different cells including platelets, endothelial cells and monocytes as well as mammary epithelial cells. The molecular mass varies considerably between tissues and species although the DNA coding for this protein appears very similar, implying large differences in glycosylation[96,97]. In platelets CD36 has been implicated in a large number of receptor functions including those for collagen, thrombospondin, oxidised LDL and for Plasmodium falciparum infected erythrocytes. The role of glycosylation in modulation of these receptor functions remains unknown. Glycosylation of CD36 from bovine mammary epithelial cells has been studied in the greatest detail [98]. Based on hydrazinolysis studies it was estimated that each CD36 molecule has on average six N-glycans. From the DNA sequence ten potential N-glycosylation sites were present [99]. Several types of oligosaccharide were found including several hybrid type chains. High mannose

197

Man(a1-2)Man(a1-3)Man(al-6)

\

Fuc(al-6)

\ -

,

Man(@I4)GlcNAc(fi1-4)GlcNAc (A) Gal(fi14)GlcNAc(~I-2)Man(a 1-3)

I

Man(al-6) \

Man(al-3)

Man(a 1-6) I \

Fuc(al-6)

\ -

,

Man(fi14)GlcNAc(fiI4)GlcNAc (B)

Gal(fiI 4)GlcNAc(fi1-2)Man(a 1-3)

I

Man(a 1-6)

\

Man(a 1-2)

Man(a 1-6) !Man(a 1-3/

Gal(fi14)GlcNAc(fi1-2)Man(a 1-3)

\

Man(a 1-2)Man(a 1-3)Man(a 1-6)

\ GalNAc(fiI-4)GIcNAc(fil -2)Man(al-3)

i

Fuc(a1-6)

(C)

\ -

Man(fi I -l)GlcNAc(fi 1-4)GlcNAc (D) I

GalNAc(fi14)GlcNAc(fiI -2)Man(a 1-6)

\ GalNAc(fi1-4)GlcNAc(fiI -2)Man(al-3)

Fuc(a1-6)> o-l

Man@ 14)GlcNAc(fi1-4)GlcNAc I

Fuc(a1-6)

\ -

I

Man@ 1 4)GlcNAc(fi1-4)GlcNAc I

(E)

Fig. 16. Mixed type mannoselcomplex type (A,D) diantennary and (B,C) tnantennary glycans and (E) diantennary complex type glycans from CD36.

type, hybrid-type, and di-, tri-, and tetraantennary complex type glycans were all found (Fig. 16). Based upon the patterns of N-glycosylation observed here and in comparison with other cell types, the conclusion was drawn that the trimming pathway of high mannose type oligosaccharides to hybrid-type ones may be distinct in different cell types. A proportion (28% of all glycans) of the hybrid and complex type which bound to Wistaria floribundu lectin contained the GalNAc(/314)GlcNAc groups instead of Gal(B14)GlcNAc in the outer chain. Some of these contained NeuSAc(a24)GalNAc. Most of the hybrid-type glycans with the Gal/GalNAc(fi 14)GlcNAc outer chain on the Man(a1-3) antenna contained an unusual Man(a I-2)Man(a 1-3) group on their Man(a 16) antenna (Fig. 16). Less is known about glycosylation of CD36 in human platelets although there has been one study [IOO] indicating the presence of both N- and 0-glycans in overall 28% carbohydrate content and both branched tetra- and disaccharides were found, probably with one or two terminal sialic acid residues.

198

3.5. PECAM-I (CD31) Platelet endothelial cell adhesion molecule- 1 (PECAM-1, CD3 1) is a member of the IgG superfamily found on various cells and may bind to glycosaminoglycans. There are nine potential N-glycosylation sites [ 10 11 but no detailed investigations have been done. 3.6. Thrombospondin Thrombospondin is a high molecular mass (420 kDa) glycoprotein found in the a-granules of platelets but it is also secreted by a range of other cells such as fibroblasts. It seems to play an important role in cellular adhesion and contains fibrinogen, heparin and collagen binding domains [ 1021. There are six potential N-glycosylation sites [ 1031.

Acknowledgements Work carried out at the Theodor Kocher Institute was supported by grants from the Swiss National Science Foundation (3 1-42336.94) and Hoffmann-La Roche Ltd. The supply of buffy coats for the isolation of platelets by the Central Laboratory of the Blood Transfusion Service of the Swiss Red Cross is gratefully acknowledged.

References [ l ] Bowman, B.H. (1993) Hepatic Plasma Proteins: Mechanisms of Function and Regulation. Academic Press, San Diego, New York. [2] Schmid, K., Binette, J.P., Kamiyama, S., Pfister, V. and Takahashi, S. (1962) Biochemistry I , 959-966. [3] Van den Eijnden, D.H., Joziasse, D.M., Dorland, L., Van Halbeek, H., Vliegenthart, J.F.G. and Schmid, K. (1980) Biochem. Biophys. Res. Commun. 92, 839-845. [4] Fournet, B., Montreuil, J., Strecker, G., Dorland, L., Haverkamp, J.F.G., Vliegenthart, J.F.G., Binette, J.P. and Schmid, K. (1978) Biochemistry 17, 5206-52 14. [5] Schmid, K., Binette, J.P., Dorland, L., Vliegenthart, J.F.G., Fournet, B. and Montreuil, J. (1979) Biochim. Biophys. Acta 581, 356-359. [6] Van Halbeek, H., Dorland, L., Vliegenthart, J.F.G., Montreuil, J., Fournet, B. and Schmid, K. (1981) J. Biol. Chem. 256, 5588-5590. [7] Cardon, P., Parente, J.P., Leroy, Y., Montreuil, J. and Fournet, B. (1986) J. Chromatogr. 356, 135-146. [8] Shiian, S.D., Nasonov, V.V., Bovin, N.V., Aleshkin, V.A., Novikova, L.I. and Liutov, A.G. (1994) Bioorg. Khim. 20, 1125-1 131. [9] Mackiewicz, A., Khan, M.A., Gorny, A,, Kapcinska, M., Juszczyk, J., Calabrese, L.H. and Espinosa, L.R. (1994) J. Infect. Dis. 169, 1360-1363. [lo] Grootveld, M., Claxson, A.W., Chander, C.L., Haycock, I?, Blake, D.R. and Hawkes, G.E. (1993) FEBS Lett. 322, 26tL276. [ I I] Hrycaj, I?, Sobieska, M., Mackiewicz, S. and Muller, W. (1993) Ann. Rheum. Dis. 52, 138-141. [I21 Biou, D., Konan, D., FCger, J., Agneray, J., Leroy, Y., Cardon, P., Fournet, B. and Durand, G. (1987) Biochim. Biophys. Acta 913, 308-312. 1131 Wieruszeski, J.M., Fournet, B., Konan, D., Biou, D. and Durand, G. (1988) FEBS Lett. 238, 390-394. [I41 De Graaf, T.W., Van der Stelt, M.E., Anbergen, M.G. and Van Dijk, W. (1993) J. Exp. Med. 177, 657-666.

199 [I51 Hermentin, P., Witzel, R., Doenges, R., Bauer, R., Haupt, H., Patel, T., Parekh, R.B. and Brazel, D. (1992) Anal. Biochem. 206, 419429. [I61 Lacki, J.K., Klama, K., Samborski, W., Mackiewicz, S.H. and Muller, W. (1994) Clin. Rheumatol. 13, 598604. [I71 Treuheit, M.J, Costello, C.E. and Halsall, H.B. (1992) Biochem. J. 283, 105-1 12. [18] Picard, V, Ersdal-Badju, E. and Bock, S.C. (1995) Biochemistry 34, 8433-8440. 1191 Schultze, H.E., Heide, K. and Haupt, H. (1962) Klin. Wochenschr. 40, 427429. [20] Saitoh, A., Aoyagi, Y. and Asakura, H. (1993) Arch. Biochem. Biophys. 303, 281-287. [21] Venembre, P., Boutten, A,, Seta, N., Dehoux, M.S., Crestani, B., Aubier, M. and Durand, G. (1994) FEBS Lett. 346, 171-174. [22] Saso, L., Silvestnni, B., Lahita, R. and Cheng, C.Y. (1993) Inflammation 17, 383400. [23] Duthel, S. and Revol, A. (1993) Clin. Chim. Acta 215, 173-187. [24] Wang, J.-S., Tsutsumi, M., Ueshima, Y., Takase, S., Matsuda, Y. and Takada, A. (1993) Alcohol Suppl. IA, 21-28. [25] Kawahara, H., Matsuda, Y., Tsuchishima, M., Wang, X.E. and Takada, A. (1993) Alcohol Suppl. IA, 29-3 5. [26] Suita-Mangano, P., Howard, S.C., Lennarz, W. and Lane, M.D. (1982) J. Biol. Chem. 257, 42924300. [27] Reardon, C.A., Driscoll, D.M., Davis, R.A., Borchardt, R.A. and Getz, G.S. (1986) J. Biol. Chem. 261, 46384645. [28] Schindler, P.A., Settinen, C.A., Collet, X., Fielding, C.J. and Burlingame, A.L. (1995) Protein Sci. 4, 79 1-803 1291 Francone, O.L., Evangelista, L. and Fielding, C.J. (1993) Biochim. Biophys. Acta 1166, 301-304 [30] Wong, L. and Torbati, A. (1994) Biochemistry 33, 1923-1929 [3 I ] Tsunemitsu, M., Ishikawa,Y., Taniguchi, T., Fukuzaki, H., Yokoyama, M. (1992) Atherosclerosis 93, 229-235. [32] Yang, C.Y, Gu, Z.W., Blanco-Vaca, F., Gaskell, S.J., Yang, M., Massey, J.B., Gotto Jr., A.M. and Pownall, H.J. (1994) Biochemistry 33, 12451-12455. [33] Day, J.R., Albers, J.J., Gilbert, T.L., Whitmore, T.E., McConathy, W.J. and Wolfbauer, G. (1994) Biochem. Biophys. Res. Commun. 203, 1 146-1 15 1. [34] Yamashita K., Liang, C.G., Funakoshi, S. and Putnam, F.W. (1981) J. Biol. Chem. 256, 1283-1289. [35] Endo M., Suzuki, K., Schmid, K., Fournet, B., Karamanos, Y., Montreuil, J., Dorland, L., Van Halbeek, H. and Vliegenthart, J.F.G. (1982) J. Biol. Chem. 257, 8755-8760. [36] Schoenberger, O.L. (1992) FEBS. Lett. 314, 430434. [37] Strecker, G., Ollier-Hartmann, M.P., Van Halbeek, H., Vliegenthart, J.F.G., Montreuil, J. and Hartmann, L. (1985) C. R. Acad. Sci. Pans 301, 571-576. [38] De Bruijn, M.H.L. and Fey, G.H. (1985) Proc. Natl. Acad. Sci. USA 82, 708-712. [39] Jimenez-Clavero, M.A., Gonzalez-Rubio, C., Fontan, G. and Lopez-Trascasa, M. (1994) Immunol. Lett. 42, 185-190. [40] Pittman, D.D., Tomkinson, K.N. and Kaufman, R.J. (1994) J. Biol. Chem. 269, 17329-17337. [41] Thim, L., Bjoem, S., Christensen, M., Nicolaisen, E.M., Lund-Hansen, T., Pedersen, A.H. and Hedner, U. (1988) Biochemistry 27, 7785-7793. [42] Nishimura, H., Kawabata, S., Kisiel, W., Hase, S., Ikenaka, T., Takao, T., Shimonishi, Y. and Iwanaga, S. (1989) J. Biol. Chem. 264, 20320-20325. [43] Hironaka, T., Furukawa, K., Esmon, P.C., Fournel, M.A., Sawada, S., Kato, M., Minaga, T. and Kobata, A. (1992) J. Biol. Chem. 267, 8012-8020 [44] Hironaka, T, Furukawa, K, Esmon, P.C., Yokota, T., Brown, J.E., Sawada, S., Fournel, M.A., Kato, M., Minaga, T. and Kobata, A. (1993) Arch. Biochem. Biophys. 307, 316-330. [45] Pittman, D.D., Tomkinson, K.N. and Kaufman, R.J. (1994) J. Biol. Chem. 269, 17329-17337. [46] Webb, E., Tkalcevic, J., Edwards, S., Hocking, D. and Nisbet, I . (1993) Biochem. Biophys. Res. Commun. 190, 536-543. [47] Agarwala, K.L., Kawabata, S., Takao, T., Murata, H., Shimonishi, Y., Nishimura, H. and Iwanaga, S. (1994) Biochemistry 33, 5167-5171.

200 [48] Kuraya, N., Omichi, K., Nishimura, H., Iwanaga, S. and Hase, S. (1993) J. Biochem (Tokyo) 114, 763-765. [49] Harris, R.J., Van Halbeek, H., Glushka, J., Basa, L.J., Ling, VT, Smith, K.J. and Spellman, M.W. (1993) Biochemistry 32, 6539-6547. [50] Nishimura, H., Takao, T., Hase, S., Shimonishi, Y. and Iwanaga, S. (1992) J. Biol. Chem. 267, 1752017525. [51] Nakagawa, H., Takahashi, N., Fujikawa, K., Kawamura, Y., Iino, M., Takeya, H., Ogawa, H. and Suzuki, K. (1995) Glycoconjugate J. 12, 173-181. [52] Inoue, K. and Morita, T.(1993) Eur. J. Biochem. 218, 153-163. [53] Sinha, U. and Wolf, D.L. (1993) J. Biol. Chem. 268, 3048-3051. [54] Harris, R.J., Ling, VT. and Spellman, M.W. (1992) J. Biol. Chem. 267, 5102-5107. [55] Bayard, B., Kerkaert, J . 2 , Strecker, G., Dorland, L., Van Halbeek, H. and Vliegenthart, J.F.G. (1983) Eur. J. Biochem. 137, 319-323. [56] Topfer-Petersen, E. (1980) Thromb. Haemostas. 42, 4 3 4 5 . [57] Debeire, P., Montreuil, J., Moczar, E., Van Halbeek, H. and Vliegenthart, J.F.G. (1985) Eur. J. Biochem. 151, 607411. [58] Yamada, K.M. (1983) Annu. Rev. Biochem. 52, 761-799. [59] Mandel, U., Hamilton-Therkildsen, M., Reibel, J., Sweeney, B., Matsuura, H., Hakomori, S., Dabelsteen, E. and Clausen, H. (1992) APMIS 100, 817-826. [60] Ingham, K.C., Brew, S.A. and Novokhatny, VV (1995) Arch. Biochem. Biophys. 316, 235-240. [61] Carsons, S., Clausen, H . and Wolf, J. (1994) J. Rheumatol. 21, 1888-1891. [62] Takahashi, N., Takahashi, Y. and Putnam, F.W. (1985) Proc. Natl. Acad. Sci. USA 82, 73-77. [63] Van Gelder, W., Huijskes-Heins, M.I., Hukshorn, C.J, De Jeu-Jaspars, C.M., Van Noort, W.L. and Van Eijk, H.G. (1995) Comp. Biochem. Physiol. Biochem. Mol. Biol. 1 I I , 171-179. [64] Bernard, N., Engler, R., Strecker, G., Montreuil, J., Van Halbeek, H. and Vliegenthart, J.F.G. (1984) Glycoconjugate J. 1, 123-140. [65] Hayase, T., Rice, K.-G., Dziegielewska, K.-M., Kuhlenschmidt, M., Reilly, T. and Lee, Y.-C. (1992) Biochemistry 3 1, 491 5 4 9 2 1. [66] Watzlawick, H., Walsh, M.-T., Yoshioka, Y., Schmid, K. and Brossmer, R. (1992) Biochemistry 31, 12198-12203. [67] Takahashi, N., Takahashi, Y. and Putnam, F.W. (1985) Proc. Natl. Acad. Sci. USA 82, 1906-1910. [68] Dunn, J.T. and Spiro, R.G. (1967) J. Biol. Chem. 242, 5549-5556. [69] Saso, L., Silvestrini, B., Guglielmotti, A,, Lahita, R. and Cheng, C-Y. (1993) Inflammation 17,465479. [70] Ryan-Poirier, K.-A. and Kawaoka, Y. (1993) Virology 193, 974-976. [71] Hayes, M.L. and Castellino, F.J. (1979) J. Biol. Chem. 254, 8768-8771. [72] Hayes, M.L. and Castellino, EJ. (1979) J. Biol. Chem. 254, 8772-8776. [73] Hayes, M.L. and Castellino, EJ. (1979) J. Biol. Chem. 254, 8777-8780. [74] Samandari, T. and Brown, J.-L. (1993) Protein Sci. 1400-1410. [75] Miletich, J.P. and Broze, G.J. (1992) J. Biol. Chem. 265, 11397-11404. [76] Lu, D., Rydzewski, A. and Long, G.L. (1995) Blood 86, 1489 (abstr.). [77] Nilsson, B., Home 111, McD.K. and Gralnick, H.R. (1983) Arch. Biochem. Biophys. 224, 127-133. [78] Ogawa, H., Yoneda, A,, Seno, N., Hayashi, M., Ishizuka, I., Hase, S. and Matsumoto, 1. (1995) Eur. J. Biochem. 230, 994-1000. [79] Yoneda, A., Ogawa, H., Matsumoto, I., Ishizuka, I., Hase, S. and Seno, N. (1993) Eur. J. Biochem. 218, 797-806. [80] Goudemand, J., Mazurier, C., Samor, B., Bouquelet, S., Montreuil, J. and Goudemand, M. (1985) Thromb. Haemostas. 53, 390-395. [81] Debeire, ,.'F Montreuil, J., Samor, B., Mazurier, C., Goudemand, M., Van Halbeek, H. and Vliegenthart, J.F.G. (1 983) FEBS Lett. 151, 22-26. [82] Samor, B., Michalski, J.C., Debray, H., Mazurier, C., Goudemand, M., Van Halbeek, H., Vliegenthart, J.F.G. and Montreuil, J. (1986) Eur. J. Biochem. 158, 295-298. [83] Samor, B., Michalski, J.C., Mazurier, C., Goudemand, M., De Waard, P., Vliegenthart, J.F.G., Strecker, G. and Montreuil, J. (1989) Glycoconjugate J. 6, 263-270.

20 1 [84] Tsuji, T. and Osawa, T. (1987) J. Biochem. 101, 241-249. [85] Korrel, S.A.M., Clemetson, K.J., Van Halbeek, H., Kamerling, J.P., Sixma, J.J. and Vliegenthart, J.F.G. (1988) FEBS Lett. 228, 321-326. [86] Tsuji, T., Tsunehisa, S., Watanabe, Y., Yamamoto, K., Tohyama, H. and Osawa, T. (1983) J. Biol. Chem. 258, 63354339. [87] Korrel, S.A.M., Clemetson, K.J., Van Halbeek, H., Kamerling, J.P., Sixma, J.J. and Vliegenthart, J.F.G. (1984) Eur. J. Biochem. 140, 571-576. [88] Korrel, S.A.M., Clemetson, K.J., Van Halbeek, H., Kamerling, J.P., Sixma, J.J. and Vliegenthart, J.F.G. (1985) Glycoconjugate J. 2, 229-234. [89] Lbpez, J.A., Ludwig, E.H. and McCarthy, B.J. (1992) J. Biol. Chem. 267, 10055-10061. [90] Wicki, A.N. and Clemetson, K.J. (1987) Eur. J. Biochem. 163, 43-50. [91] Lanza, F., Morales, M., De La Salk, C., Cazenave, J.-P., Clemetson, K.J., Shimomura, T. and Phillips, D.R. (1993) J. Biol. Chem. 268, 20801-20807. [92] Poncz, M., Eisman, R., Heidenreich, R., Silver, S.M., Vilaire, G., Surrey, S., Schwanz, E. and Bennett, J.S. (1987) J. Biol. Chem. 262, 8476-8482. [93] Tsuji, T. and Osawa, T. (1985) J. Biochem. 100, 138771398, [94] Fitzgerald, L., Steiner, B., Rall, S.C., Lo, S. and Phillips, D.R. (1987) J. Biol. Chem. 262, 3936-3939. [95] Nakagawa, H., Zheng, M., Hakomori, S . , Tsukamoto, Y., Kawamura, Y. and Takahashi, N. (1996) Eur. J. Biochem. 237, 76-85. [96] Greenwalt, D.E., Watt, K.W.K., Hasler, T., Howard, R.J. and Patel, S. (1990) J. Biol. Chem. 265, 1629616299. [97] Greenwalt, D.E., Lipsky, R.H., Ockenhouse, C.F., Ikeda, H., Tandon, N.N. and Jamieson, G.A. (1992) Blood 80, 1105-1115. [98] Nakata, N., Furukawa, K., Greenwalt, D.E., Sato, T. and Kobata, A. (1993) Biochemistry 32, 43694383. [99] Oquendo, P., Hundt, E., Lawler, J. and Seed, B. (1989) Cell 58, 95-101. [IOO] Tsuji, T. and Osawa, T. (1986) J. Biochem. 100, 1077-1085. [I011 Newman, P.J., Berndt, M.C., Gorski, J., White 11, G.C., Lyman, S., Paddock, C. and Muller, W.A. (1990) Science 247, 1219-1222. [I021 Lahav, J. (1993) Thrombospondin. CRC Press, West Palm Beach, FL. [I031 Lawler, J. and Hynes, R.O. (1986) J. Cell Biol. 103, 1635-1648.

This Page Intentionally Left Blank

J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins II 0 1997 Elsevier Science B.V. All rights reserved

CHAPTER 10

Transferrin superfamily An outstanding model for studying biochemical evolution Jean Montreuil, Genevikve Spik and Joel Mazurier Uniuersiti des Sciences et Technologies de Lille, Laboratoire de Chimie Biologique (UMR no. 111 du CNRS), 59655 Villeneuve dilscq Cedex, France

1. Introduction I . I . DeJinition of transferrin superfamily The superfamily of transferrins (for reviews, see refs. [ 1-10]), serotransferrins from blood plasma, lactotransferrins (also called lactoferrins) from mammalian milk, external secretions and polymorphonuclear leukocytes, ovotransferrins from avian and reptile egg-white and melanotransferrins (also called p97 protein) from cell membranes, constitutes a group of well conserved glycoproteins which present the following common physicochemical and biological properties: (i) their molecular mass is around 80 kDa; (ii) they are constituted of a single polypeptide chain organized in two lobes originating from a gene duplication; (iii) each lobe binds reversibly one Fe3+ ion and iron-saturated transferrins develop a salmon-pink color; (iv) the protein moiety of transferrins presents a high degree of homology (about 50%); (v) all transferrins are glycosylated (2.2-1 1.2% total sugars), except some fish transferrins; (vi) they play a key role in iron transport and are recognized by specific membrane receptors; (vii) they are inhibitors of the growth of microorganisms by a mechanism of ferrideprivation. Consequently, the transferrins represent a remarkable model for answering the question posed by Hughes and Butters [ 111: “Glycosylation pattern in cells: an evolutionary marker?”. This explains why we undertook, 30 years ago, a systematic and comparative study on glycan primary structures of transferrin from tissues and biological fluids of different species. In addition, our second aim was to address another important question: “Why are transferrins glycosylated?’. But, we must confess that, on the basis of the determination of the glycan primary and three-dimensional structure of transferrins, we are still unable to establish relationships between primary structure and function of the glycans, on the one hand and, on the other hand, to answer the two aforementioned questions as well as the following one: “How to explain the great microheterogenicity of their glycan moieties?”. At the moment, the mystery thickens, since, as observed in our laboratory and in others [12], the partial or complete deglycosylation of transferrins does not affect the reversible iron fixation, the recognition and binding to the reticulocyte membrane, or the iron transfer into the cell. In addition, as shown by Van Eijk et al. [13], serotransferrin molecules with the same iron content, but with variable antennae number and sialic acid content, i.e. from pentasialo- to asialotransferrins, show the same iron delivery to rat reticulocytes. In the same way, Mason et al. [ 141 have demonstrated that a 203

204

non-glycosylated mutant of serotransferrin expressed in BHK cells, after mutation of the two asparagine carbohydrate linkage sites to aspartic acid residues, binds to HeLa S3 cells with the same avidity and to the same extent as the glycosylated protein, so proving that the glycans have no role in this function. Consequently, the present chapter does not deliver any information about the structurefunction relationships of transferrink glycans. However, it is devoted to an interesting study of ‘horizontal evolution’ through the superfamily of transferrins. In this respect, the present chapter is a useful addition to chapters 6 1 2 of volume 29A [15] and to chapters 1-8 of the present volume [16]. In fact, these chapters are devoted to the description of the primary structure of glycoprotein glycans from bacteria to man and are thus relevant to ‘vertical evolution’.

1.2. Biological importance of iron In Aphorisms, Hippocrates asserted that “diseases which are not healed by cures, are healed by iron”. We know now that iron is distributed in all of the living organisms and that all of the cells need iron to survive and to grow. For this reason, Nature has coined the transferrins of which the mission is to furnish this invaluable metal to all living cells (for a review, see ref. [7]). The total iron content of the normal human adult is approximately 3-5g and may be subdivided into four compartments: red blood cell iron, tissue iron, storage iron and transport iron. The largest iron fraction, 1.5-3 g, consists of hemoglobin of the red blood cells. The tissue iron-porphyrin proteins such as myoglobin, cytochromes, as well as ironsulfur proteins, peroxidases and catalases, account for only 0.1-0.3 g of iron. The storage iron fraction (1.0-1.5g) is widely deposited in tissues of the body, the largest fraction being in the liver. Its main form is ferritin which can store 23% of its weight of iron (5000 Fe3+ ions per mole) and has a molecular mass of approximately 480 kDa in its iron-free form of apoferritin. The link between the various forms of tissue and storage iron is serotransferrin. Only 3 4 m g of iron is present as serum iron which is equivalent to about 30% of iron-saturated serotransferrin.

2. The transferrin superfamily In the present chapter, the term ‘transferrin’, introduced forty-nine years ago by Laurel1 and Ingelman [ 171, will concern any protein possessing the characteristics above listed in section 1.1. Consequently, all of the compounds we describe possess, as a trade-mark, this common root in their name, according to Williams [ 181 who wrote on the chaotic nomenclature of the group: “All members of the group should carry the same family name. Such names as conalbumin, lactoferrin, p97 and sciatin are not acceptable”. 2.1. The saga of transferrins

The first transferrin was isolated in 1900 from hen egg-white and the definition of its properties, chiefly its inhibitory power of bacterial growth due to its affinity for

205

iron, contributed to the discovery of the other transferrins. This explains the order of presentation of the transferrins we have adopted since it respects the chronology of the events.

2. I . I . Ouotrunsferrins (conalbumins) Although conalbumin, as it was called, was isolated by Osborne and Campbell in 1900 [19] from hen egg-white, almost half a century was to pass before the protein was identified in 1944 as the antimicrobial agent of egg-white whose properties were suppressed by the addition of iron as demonstrated by Schade and Caroline[20]. As related by Schade in a review full of humour[21], this discovery was the issue of a contract signed by Schade with the Medical Corps of the U.S. Army for producing a bacteriophage preparation effective against Shigellu dysenteriae. In order to protect the phages from the effects of lyophilisation, Schade used some additives, including eggwhite and observed that the growth inhibition of Shigellu was proportional to the eggwhite concentration of the medium. On the basis of this observation, Schade obtained by fractional ammonium sulfate precipitation a pure and very active compound which bound ferric iron to produce a salmon-pink color and to which he gave the Greek-rooted name of ‘siderophilin’ [22]. A few months later, Alderton et al. [23] isolated the protein they identified as conalbumin. In this way the bacteriostatic property of egg-white, which was well known for a long time, was explained. In this regard, as Schade wrote in an excellent review on non-heme metalloproteins [24]: “One may wonder whether Shakespeare was aware of this property, for in King Leur, Act 111, Scene VII, just following the excision of the eyes of the Duke of Gloucester and a cut face by the Duke of Cornwall, one servant speaks to a second: ‘I’ll fetch some flax and white of eggs to apply to his bleeding face’ ”. Schade opened the way to the discovery of serotransferrins. He suggested that blood might have a protein component similar to that of conalbumin and capable of so lowering the availability of iron to microbial pathogens that it might serve as a protective factor against infection of the host. Using Cohn’s method of plasma fractionation, Schade confirmed this view by characterizing a fraction which exhibited both the bacterial growth inhibitory and color-producing properties shown by conalbumin [25,26]. Conalbumin comprises 3-16% of the egg-white from various species of birds. It is structurally related to plasma serotransferrin, the polypeptide chain being coded by the same structural gene [27], but the carbohydrate moiety being different (see section 3.1.3). Because of the close relationship of these two proteins, the name ouotrunsferrin is now used in place of conalbumin. 2.1.2. Serotrunsferrins The existence in human plasma of non-heme iron was first demonstrated by Fontes and Thivolle in 1925 [28]. The association of bound iron with the globulin fraction was established by Starkenstein and Harvalik in 1933 [29] and, with the advent of electrophoresis, iron-binding was localized to the 61-globulin fraction by Surgenor et al. in 1949 [30]. In the same year, the same authors, on the basis of the studies of Holmberg and Laurel in 1947 [31] and of the aforementioned works of Schade and Caroline in 1946 [25], isolated and characterized the (3, -iron-binding globulin from Cohn’s fraction IV-3,4 to which Schade et al. [22] gave the name siderophilin. The crystallisation of the

206

serotransferrin was achieved by Koechlin in 1952 [32]. The presence of carbohydrate in serotransferrin was first reported by Surgenor et al. in 1949 [30]. In the meantime, in 1947, Laurel1 and Ingelman [ 171 had independently purified the ‘red protein’ from pig plasma and in the same year proposed the name ‘transferrin’ which has since been adopted as the generic name of the proteins of this family: serotransferrin (instead of siderophilin) present in blood and some external secretions, ovotransferrin (instead of conalbumin) in avian egg-white, lactotransferrin (also called lactoferrin) from milk, external secretions and leukocytes and melanotransferrin (instead of p97) in melanocyte and normal cell plasma membrane. A dozen mammalian and some frog, fish and insect serotransferrins were later isolated and characterized. In humans, the plasma concentration of serotransferrin, which is biosynthesized by the liver, varies from 3 to 3.5 g per liter and it is worthwhile to mention that the protein is only 20-30% iron-saturated, an important fact considering the inhibitory power of bacterial growth provided by aposerotransferrin acting via ferrideprivation (see above). Important variations of the plasma serotransferrin concentration are observed in physiological and pathological situations. They are often accompanied by dramatic modifications of glycan primary structure (see section 3.2).

2.1.3. Lactotransferrins (lactoferrins) Lactotransferrins (also called lactoferrins) were isolated for the first time from human and cow milk in 1960. Isolation of human lactotransferrin was described almost simultaneously by Montreuil et al. [33,34] and by Johansson [35]. By applying a fractionation procedure which associated an ammonium sulfate concentration gradient and a pH gradient, Montreuil et al. obtained a red-salmon coloured glycoprotein. The authors described in detail the physicochemical properties of the compound, demonstrated that it was glycosylated, bound reversibly ferric ions and possessed all the characteristics of a transferrin. However, it did not immunologically crossreact with human serotransferrin and had a 200-fold higher affinity for iron. These specific characteristics led Montreuil et al. to call this compound: lactosiderophilin or lactotransferrin [34]. For unknown reasons, the lactotransferrin is now commonly called lactoferrin. In 1960, Johansson described the isolation of a ‘red protein’ binding iron, devoid of Soret’s band and losing the metal at low pH [35]. In 1961, Blanc and Isliker [36] isolated from human milk by rivanol precipitation, a protein that they called lactoferrin and that they further identified as lactotransferrin. The term lactoferrin was reintroduced by Masson et al. in 1965 [37]. Bovine milk lactotransferrin first reported in 1939 by Serrensen and Serrensen [38], was isolated by Groves [39] and described under the name of bovine milk ‘red protein’. Since these first discoveries, lactotransferrin has been isolated from milk of numerous mammals: goat [40], mare [41], monkey [42], mouse [43,44], rabbit [45] and sow [46,47]. These findings caught the attention of many milk companies since, rightly or wrongly, a nutritional role was suggested for these proteins in iron transport for the mammalian newborn. The discovery of lactotransferrin specific receptors in the brush-border of enterocytes [48] supports this concept. In addition, as suggested by Montreuil et al. [34], human lactotransferrin could intervene in the intestinal defence of breast-fed infants by a

207

mechanism of iron-deprivation identical to that previously demonstrated by Schade and Caroline for ovo- and serotransferrin. Indeed, this biological role of lactotransferrin is now well established and extended to all mucous secretions (see section 2.3.2.1.2). In 1963, using immunoelectrophoresis, Biserte et al. [49] demonstrated that lactotransferrin was present in bronchial secretions. This result was confirmed in 1965 by Masson et al. [37] who, in addition, localized the lactotransferrin in human bronchial glands by immuno-histochemistry. Lactotransferrin has been identified, mainly due to the work of Masson, in most mucous secretions, i.e., bronchial and intestinal mucus and in various biological fluids such as saliva, tears, synovial fluid, seminal plasma, pancreatic juice, bile and, in very small amount, in blood plasma. In milk and mucous secretions, lactotransferrin is associated with secretory IgA (sIgA) and lysozyme and we now know that this association represents a powerful system of defence of mucosae (see section 2.3.2.1.2). The discovery by Masson et al. in 1969 [50,5 13 of the presence of lactotransferrin in the specific granules of neutrophilic leukocytes suggested its participation in the mechanisms of cell-mediated defence and opened the way to active research on this fascinating protein. 2.1.4. Melanotransferrin (human melanoma-associated antigen p97) Melanotransferrin (also called p97) (for a review see ref. [52]), discovered in 1980 by Brown et al. [53,54] and by Dippold et al. [55], is a cell-surface sialoglycoprotein that is present in most human melanomas, in certain foetal tissues, but only in trace amounts in normal adult tissues. This tumor-associated antigen was first identified in human melanoma by using monoclonal antibodies. Analysis of somatic cell hybrids and in situ hybridization have shown that the p97 gene, like the genes for transferrin and transferrin receptor, is located on chromosomal region 3q2 1-3q29 [56,57]. This observation suggested that p97 plays a role in iron metabolism and led Rose et al. [58] to propose “that p97 could be renamed melanotransferrin to denote its original identification in melanoma cells and its evolutionary relationship to serotransferrin and lactotransferrin, the other members of the transferrin superfamily”. In addition to being detected at various levels in other tumours such as lymphomas, melanotransferrin has subsequently been found in a wide range of cultured normal cell types including liver cells, intestinal epithelial cells, foetal intestinal cells, umbilical cord, placenta, sweat gland ducts, capillary endothelium of human brain and reactive microglia of Alzheimer’s disease patients. In contrast to other molecules of the transferrin family, melanotransferrin is the only one so far shown to be connected to the cell membrane by a glycosylphosphatidylinositol (GPI) anchor as demontrated by Alemany et al. [59] and Food et al. [60]. In addition, examination of melanotransferrin peptide sequence has led to the conclusion that the protein possesses only one iron binding site at its N-terminal half [61].

2.2. Comparative study of transferrin peptide chains 2.2.I . Primary and three-dimensional structure Fourteen complete transferrin and transferrin-related amino acid sequences have now

208

been determined, either directly at the protein level, or by translation from the nucleotide sequence of the corresponding mRNA or gene: human [57,62-64], pig [65], horse [66], rat [67], rabbit [68], chicken [69], frog Xenopus Zueuis [70], cockroach [71], flesh-fly Surcophugu peregrinu [72] and tobacco hornworm Munducu sextu [73] serotrunsferrins, human [74-761, bovine [77-791, pig [47,80], goat [8 13 and mouse [82] lactotrunsferrins and human melanotrunsferrin [58] (for reviews, see refs. [ 10,831). On the basis of the following data, it is clear that transferrin evolution was highly conservative. (1) All proteins consist of a peptide chain of 680-700 amino acid residues which can be divided into two homologous halves. The insect transferrin isolated from hemolymph of Surcophugu peregrinu fly is distinct since its molecular mass is 65 kDa only and, unlike the other transferrins, the similarity between its N-terminal and C-terminal halves is only 19% [72]. (2) As shown in Table 1, the sequence homology between eight different transferrins, as well as between N- and C-terminal halves of the molecules, is very extensive. These results favour the hypothesis of Williams [84] according to which transferrins originated by gene duplication. This hypothesis has been verified at the gene level by the study of exon distribution within the genes of hen ovotransferrin and human serotransferrin [85,86]. Melanotransferrin, a cell membrane transferrin-like glycosylphosphatidylinositol-anchored protein, is considered to be an intermediate between transferrins and receptors for transferrins [60]. Finally, the confirmation of the gene duplication hypothesis came from the isolation from the primitive urochordate F‘yuru stolonifera of a transferrin-like molecule of molecular mass 41 kDa with only one high affinity iron-binding site [87]. In experiments described by Bowman et al. [88], hybridization of human serotransferrin cDNA with fragments of Pyuru DNA showed hybridization signals. It was concluded that the intragenic duplication that produced the amplified transferrin gene and enabled transferrin to gain an additional iron-binding site would have occurred at some point in time which has been estimated as some 400 to 500 million years ago, at the evolutionary period wherein prochordates and humans separated. However, according to Escriva et al. [67] who have constructed the phylogenetic tree of the transferrin family, at least three gene duplications have occurred during the evolution of the transferrin sequences. An initial duplication occurred before the separation of arthropods and chordates, as both insect transferrins have duplicated N- and C-termini. A second duplication occurred in the branch leading to vertebrates before the emergence of land animals. This duplication gave rise to human melanotransferrin which can be observed to be re ancient than any other vertebrate transferrin. The third duplication took place b ore the appearance of mammals, and lactotransferrins are the resulting products. (3) The position of disulfide bridges within the two halves of the peptide chain are highly conserved in transferrins. (4) When considering the structure of the iron-binding sites, the amazing conservative evolution of transferrins bursts upon view. In fact, (i) in all transferrins, four ligands are provided by the surrounding protein structure in both iron-binding sites of N- and C-lobes: one carboxylate from aspartic acid, two phenolate oxygens

7

209

Table 1 Percentage homology between N- and C-terminal halves of various transferrins [ 101 Alignment

Percentage homology a

score (SD)a Human

Human N C N

x

C 34.0

pig N

C

Horse N C

Rabbit N C

Rat C

Hen N

C

Frog N C

Hornworm N C

48.6 71.9 45.1 75.7 44.8 81.9 43.5 39.3

53.8

45.8 71.3 45.1 70.6 48.6 75.9 66.9

45.9

54.3 39.7 51.2 30.3 33.2 45.1 43.9 44.7 28.0 32.6

x

x

46.2 42.7 42.4 28.4 31.4

Pig

N 40.5 25.6

45.5 74.9 43.6 73.4 43.0 42.9

54.6

C 22.3 44.3 22.1 x 42.6 73.4 45.1 73.4 60.5 N 67.7 29.3 76.9 22.5 x 43.4 74.2 43.2 41.9 C 24.5 76.8 27.8 61.7 24.7 x 43.8 70.9 65.2

46.2

54.1 38.6 52.3 28.8 30.1

Horse

55.2

44.5 46.9 42.1 29.9 31.5

Rabbit

N 60.8 28.3 75.1 28.9 61.0 15.8

Rat Hen

25.3

x

N 25.3 18.8 28.3 17.8 26.5 18.1 34.3 15.3

0.6

29.9

22.5

C 25.6 35.7 23.4 34.4 26.9 35.9 21.6 32.5 18.0

22.5

53.0 20.7

0.7

12.8

9.2 12.6

0.8

17.7

9.0 12.1 10.1

43.0

52.6 39.0 51.7 27.9 28.6

42.9 39.9

55.8

43.8 44.0 42.4 28.0 29.7

66.3

44.3

53.0 35.9 50.5 30.0 31.3

C 8.6 21.5 6.9 19.9 6.5 26.2 5.0 26.0 x N 30.6 29.7 29.5 28.0 37.1 21.4 34.0 22.8 3.7 C 19.5 31.7 25.2 36.4 21.7 62.4 21.6 47.3 24.8

33.3

48.0 31.3 51.4 25.3 27.8

x

41.4 42.9 44.2 28.2 32.9

x

C 23.6 55.2 28.5 44.9 28.9 59.9 24.5

Frog

Hornworm N 13.0 C a

9.9 10.2

x

7.5 10.2 10.8 14.6 11.5

13.8 11.0 15.8 11.7 14.5

8.5 15.0 10.2

37.2 50.7 27.5 28.0 x

37.1 28.9 26.6 x

8.4

27.9 29.0 x

30.3

8.5

x

N,C: N- and C-terminal part, respectively.

Fig. 1. Schematic representation of the iron- and anion-binding site in lactotransferrin (Anderson et al. [89]).

from tyrosine and one imidazole nitrogen from histidine (Fig. 1) [89]. In addition, one arginine intervenes in the prior fixation of a carbonic acid anion which is a prerequisite for the binding of iron; (ii) the position of these amino-acids along the peptide chain of all of the transferrins is highly conserved as demonstrated in Table 2 [90].

210 Table 2 Structural comparison of the iron binding sites of N-lobe of rabbit serotransferrin. human lactotransferrin and hen ovotransferrin [90] Rabbit serotransferrin

Human lactotransferrin

Hen ovotransferrin

Arg-124

Arg-121

Arg-121

Arp-63

ASP-60

His-249

ASP-60 His-253

Tyr-95,I 88

Tyr-92,192

Tyr-92,191

His-250

The conservative evolution of the primary structure of transferrin peptide chains leads, as a consequence, to a conservative evolution of transferrin three-dimensional structure (for a review see ref. [91]). In fact, X-ray crystallographic studies of human lactotransferrin [89,92-971, rabbit serotransferrin [98] and hen ovotransferrin[99] show (see Fig. 20) that the single peptide chain of these transferrins is folded into two lobes corresponding to the N-terminal and C-terminal halves of the molecules: in human lactotransferrin for example, amino acids 1 to 333 and 345 to 691, respectively. Both lobes are associated by non-covalent hydrophobic amino acid interactions, salt bridges and water molecules as demonstrated by Legrand et al. [loo] for human lactotransferrin and by Ikeda et al. [ l o l l for hen ovotransferrin. Both lobes are joined by a short connecting peptide of about twelve amino acids: eleven in human lactotransferrin (amino acids 334-344), sometimes in an extended conformation, as in rabbit serotransferrin [98] and hen ovotransferrin [99] and sometimes in a three-turn a-helix conformation as in human lactotransferrin [92] (see Fig. 2 1C). The three-dimensional pictures of human seroand lactotransferrin peptide chains are perfectly superimposable [9 11 with very few differences, so demonstrating conservative evolution of transferrins led to a conservative evolution of their three-dimensional structures. 2.2.2. Location of glycosylation sites 2.2.2.1. Serotransferrins. All known serotransferrins contain one or two glycans of the N-acetyllactosaminic type which are located in the C-terminal lobe of the polypeptide chain. Hen [102], rabbit [103,104], pig [I051 and rat [67] serotransferrins contain a single glycan located in a very similar position which does not correspond to Asn-4 13 in human serotransferrin (Fig. 2). As shown in Fig. 3 the two glycosylation sites (Am-413 and 611) of human serotransferrin may be occupied either by di-, tri- or tetraantennary glycans of the N-acetyllactosaminie type [ 106,1071. Concerning the triantennary glycans two positions of the third antenna exist: p-1,4-linked to the a-1,3-mannose residue of the inner-core, namely 2,4-variant, or b-1,6-linked to the a-176-mannoseresidue of the inner-core, namely 2,6-variant. This leads to several human transferrin glycovariants which are separable by Con A-Sepharose chromatography into glycovariants Tf-I, Tf-11 and Tf-111 described in section 3.1.1 [ 1081. The location of 2 4 - and 2,6-variants can be determined by using serial lectin affinity chromatography, fast atom bombardment-mass spectrometry and H-NMR

'

21 1

Human

679

Rat

695 490

Pig

727

497

Rabbit

676 485

Hen

686 473

618

Horse 1

687

Horse 2

687

729

Melano Tf 38

135

515

Fig. 2. Location of glycans in human [106,107], rat [67], pig [105], rabbit [103,104], hen [102], horse (1,2: vanants 1 and 2, respectively [I 14,1151) serotransfernns and human melanotransferrin [52].

spectroscopy. The results obtained show that the ratio of 2,4-variant versus 2,6-variant at the glycosylation site Asn-413 is approximately 5:1, whereas this ratio is 1:l for the glycosylation site Asn-61 1 [ 1091. In addition, on the basis of their different content in sialic acids, the glycovariants have been separated by isoelectric focusing [ 1101 or by high performance pellicular anion-exchange chromatography [ 1 111. The most accurate method for detecting the presence of human serotransferrin glycovariants is electro-spray ionization mass-spectrometry. On the basis of a theoretical mass of 75 143 daltons for the polypeptide chain of human serotransferrin [ 141, the mass values of the different serotransferrin glycovariants which could be found in normal and pathological sera are given in Fig. 3 [ 1121. Horse serotransferrin presents two glycovariants with one (Asn-496) [ 1 131 or two (Asn-496 and -6 19) glycans [ 1141 due to a mutation from Asn-638 to Ser-640 [1 151. Potential glycosylation sites are present in j s h [116], tobacco hornworm [73] and insects [72] serotransferrins, but the position of glycans in their peptide chain has not yet been determined. Figure 2 synthesizes the present data concerning the location of glycans on the peptide chain of different serotransferrins. It is remarkable that, except for melanotransferrin, all of the glycans or potential glycosylation sites are located in the C-domain and in very conserved positions.

212 1

75143

473

1

'?'

,679

N

79260

3SA

79551

4SA 1

2

1

2

1

79916

80207

80863

81519

4SA

1 5SA

'

1

2

1

2

' 1

1

6SA

I

7SA

2

Fig. 3. Location and molecular masses from 79 260 to 81 519 of human serotransferrin glycovariants. Open circle, NeuAc; solid square, Gal; solid circle, GlcNAc; solid diamond, Man; 1 to 679, amino acids of the peptide chain; x SA, number of sialic acid residues [ 1 121.

2.2.2.2. Ouotransferrins. The peptide chain of hen ovotransferrin is identical to that of hen serotransferrin. However, both glycoproteins differ only by the structure of their glycans (see Figs. 16B,D). Two potential glycosylation sites have been identified

213

Human 1

Human 2

138

479

624

138

479

624

Mouse

692

692

689

cow

689 233281

368

476 545

(a) Mare

692

Goat

692

(a) Swine

667 366

472

571

Fig. 4. Location of the glycosylation sites on the peptide chain of human variants 1 [74] and 2 [117], mouse [121], cow [I 181, mare, goat [81] and swine[80,120] lactotransferrins: (a), mapping not determined; solid circles on bars, glycosylated sites; heavy bars, non-glycosylated sites. In cow lactotransferrin, glycans are of the oligomannosidic type at positions 233 and 545, of the oligomannosidic or N-acetyllactosaminic type at positions 368 and 476. Only the glycans of the N-acetyllactosaminic type at position 476 contain one N-acetylgalactosamine residue.

in the C-terminal lobe (Asn-473 and -618) (Fig. 2) but only residue Asn-473 is glycosylated [ 1021. N-glycosylation site mapping of lactotransferrins from different species is described in Fig. 4. The number of N-glycosylation sites present in lactotransferrin polypeptide chains is quite variable and two types of lactotransferrins can be distinguished. Lactotransferrins with N-acetyllactosamine type glycans possess one or two glycosylation sites, while lactotransferrins with both oligomannosidic and N-acetyllactosamine type glycans contain five potential glycosylation sites. Contrary to serotransferrins in which glycans are located in the C-domain only, glycans of lactotransferrins are located in both N- and C-domains. H u m a n lactotransferrin possesses three potential glycosylation sites (Asn- 137, 478 and 624) and only two glycans in positions Asn-137 and Asn-478 [74]. However, Van Berkel et al. [ 1171 have identified a human lactotransferrin glycovariant glycosylated on Asn-624. Bovine lactotransferrin possesses five potential glycosylation sites, but the presence of only four glycans has been demonstrated and Asn-281 has not been glycosy-

2.2.2.3. Lactotransferrins.

214

lated [ 1181. The glycans of bovine lactotransferrin present a high degree of structural heterogeneity (see Fig. 14). The location of the four glycosylation sites is given in Fig. 4. The glycan of the N-acetyllactosamine type containing the GalNAc(fi1-4)GlcNAc unit is linked to Asn-476. This result suggests that a recognition signal for the N-acetylgalactosaminyltransferase should exist in bovine lactotransferrin and that the gene for this enzyme is expressed in bovine mammary gland. During lactation, the relative ratios of the different types of glycans linked to the four glycosylation sites vary significantly. Only oligomannosidic structures are present in cow lactotransferrin, up to one month before calving, whereas three glycans of the oligomannose type and one of the N-acetyllactosamine type have been detected in colostrum [119]. This result suggests either an effect of the hormonal status or of the rate of lactotransferrin biosynthesis on the type of glycosylation. Caprine lactotransferrin has, like bovine lactotransferrin, five potential glycosylation sites (Fig. 4), but the location of glycans has not yet been defined. Porcine lactotransferrin possesses three potential N-glycosylation sites which have been located in the C-terminal lobe (Asn-366, 472 and 571) (Fig. 4) [80] but only one site is glycosylated as recently demonstrated by Coddeville et al. [120]. Murine lactotransferrin polypeptide chain carries two glycans of the N-acetyllactosaminic type (Fig. 4) [121]. 2.2.2.4. Melanotransferrin. Melanotransferrin is a sialoglycoprotein [56] with three potential glycosylation sites located in the N-terminal and C-terminal lobes (Fig. 2) [52]. However, structure and location of glycans have not yet been defined. 2.2.2.5. Conclusion. In conclusion, considering in Figs. 2 and 4 the location of the glycans on the sero- and lactotransferrin polypeptide chains from different species, it is remarkable to note that glycans are located in highly conserved positions of the C-terminal lobe: Asn-490 to 496 and Asn-611 to 619 for serotransferrins and Asn-472 to 479 for lactotransferrins. This observation suggests that these particular glycans should play a crucial role in the biological functions of transferrins probably by interacting with an important peptide sequence. 2.3. Role of transferrins and of their receptors 2.3.1. Serotransferrin 2.3.1.1. Role of serotransferrin. Serotransferrin is now recognized as the protein transporting iron throughout the vascular as well as the lymphatic system (for a review, see ref. [7]). First, from the iron entering through the intestinal mucosa to the storage organs, and, secondly, from these storage organs to all cells requiring iron for growth and maintenance, and, in greatest quantity, for maturation of the developing erythrocytes. The capture of iron by cells is mediated by a specific membrane receptor of which a schematic structure proposed by Trowbridge [122] is given in Fig. 5. Serotransferrin, two thirds of which is iron-free in normal plasma, may intervene in the defence of organisms against bacteria due to its ability to bind iron so tighly under physiological conditions that the concentration of free ionic iron in plasma and lymph

215

671 AMINO ACID

EXTRACELLULARDOWIN

Fig. 5 . Schematic representation of the human serotransferrin receptor. C, cysteine residues; positions 89 and 98, disulfide bridges; position 62, fatty acid chain. From Trowbridge et al. [122].

is very low. Thus, the creation of a nutritional iron lack establishes an inhospitable environment for many pathogens leading to an inhibition of their growth, as demonstrated by Schade et al. [26,123]. This constitutes a defense mechanism against disease which reinforces the immunity system and serves as protection against free-iron toxicity, 2.3.1.2. Serotransferrin receptors. In many cell types, the uptake of iron is mediated by a specific receptor first identified as a placenta brush border membrane glycoprotein in 1979, by Sussman et al. [124] (for reviews, see refs. [125-1281). The number of receptors per cell varies from several ten thousands to almost a million. Investigations in a large number of tissues including tumors have revealed significant increases of serotransferrin receptors in all dividing cells so that the measurement of receptor expression has become a standard procedure to determine the growth potential of in uiuo tumors (for a review, see refs. [129,130]). Serotransferrin receptor genes are encoded on the same chromosome (chromosomal region 3q21-3q29) in humans [13 1,1321 as those for sero- and melanotransferrins (for a review, see ref. [SS]). Receptors of numerous cells have been identified by the use of monoclonal antibodies and appear to have similar structures to that represented in Fig. 5. They are all disulfidelinked dimers consisting of two identical transmembrane glycosylated units (MM: 95 kDa) of 760 amino acids with the N-terminus facing the cytoplasm. The cytoplasmic domain corresponds to the first 61 amino acid residues and is followed by a single hydrophobic

216

transmembrane region of 28 residues. Ser-24 is phosphorylated by a protein kinase C. The serotransferrin receptor, like other type I1 membrane proteins, does not have a cleavable signal sequence and it has been suggested that such transmembrane proteins have appeared early in the course of biochemical evolution [133]. The two subunits are linked by two disulfide bonds. Cysteine at position 62 is the site of acylation by a fatty acid residue. Serotransferrin receptor is a glycoprotein containing both N- and 0-linkages to the peptide chain[134-138]. The location in the extracellular domain of the three N-glycosylation sites Asn-25 1, 3 17 and 727 was deduced from the cDNA sequence of the receptor [131,132]. The sites of N-glycosylation of serotransferrin receptors are highly conserved in evolution. In fact, all of the receptors cloned up to date, including human, chicken, Chinese hamster, mouse and rat, possess three sites of N-glycosylation [ 13 1,132,1391. The first two sites are equivalent to human sites Asn-25 1 and Asn-3 17 [ 139,1401. Rodents have two glycosylation sites [ 1391 and chicken one glycosylation site [ 1411. The presence of one 0-glycan linked through an N-acetylgalactosamine residue to threonine-I04 near the transmembrane domain has been established by Do and Cummings [ 1371, The type of glycan is not randomly distributed along the peptide chain. Asn-251 site contains glycans of the N-acetyllactosaminic type while Asn-727 is entirely of the oligomannosidic type [142,143]. Glycans linked to Asn-317 are a mixture of glycans of the N-acetyllactosaminic, oligomannosidic and hybrid type [ 1443. Serotransferrin receptor N-glycans play a key role in the folding and transport of the receptor to the cell surface. Inhibition of N-glycosylation by treatment of cells with tunicamycin blocks translocation of the receptor to the plasma membrane [ 145,1461 interferes with formation of active dimers [145,146] and leads to the retention of the receptor in the endoplasmic reticulum [ 147,1481. Site-directed mutagenesis led to the same results [ 147-1501. Although all three sites contribute to the correct folding, transport and functioning of the serotransferrin receptor, the glycan conjugated to Asn-727 is the most critical to the structure and function of the receptor [ 143,lSO]. Glycans of sites 25 1 and 317 are the least critical for the folding and transport of the receptor. The serotransferrin-to-cell cycle, simultaneously discovered in 1983 by Dautry-Varsat et al. [151] and Klausner et al. [152], is a shuttle involving a complex pathway of endocytosis and recycling of the serotransferrin-serotransferrin receptor. Endocytosis is initiated by the binding of diferric serotransferrin to the receptor and is followed by the accumulation of the complex in clathnn-coated vesicles. Vesicles fuse with endosomes and progressive acidification to pH 5.5 leads to the release of iron, leaving the apotransferrin molecule still firmly bound to its receptor. Finally, the apotransferrinreceptor complex is recycled back to the plasma membrane and, encountering a neutral pH, the apo-serotransferrin dissociates from its receptor, is released in the circulation and goes off in a renewed search for iron at the level of storage tissues. 2.3.2. Lactotransferrin 2.3.2. I . Role of lactotransferrin. When the first lactotransferrins were isolated from mammalian milks, two hypotheses were proposed about the role they could play [34]: iron nutrition of the new-born and antibacterial defence of the gut due to growth inhibition

217

of microorganisms by iron deprivation (see sections 2.1.1 and 2.1.2). However, the fundamental discovery by Masson et al. in 1969 [50,5 I] of the presence of lactotransferrin in the granules of neutrophilic leukocytes, suggesting the participation of this protein in cell-mediated defence, gave rise to active research in this field. Two events attest to the development of knowledge on the biological role of lactotransferrin. First, the proliferation of literature in this domain in the past five years [7,129,153-1631. Second, the organization every two years of an international symposium on “Lactoferrin, structure and function” in the USA in 1993 and 1995 [159,163] and in France in 1997. We know now that lactotransferrin plays roles in iron transport, in local defence of epithelia, in cell-mediated defence of organisms and as a growth factor. 2.3.2.1.1. Iron transport. Experiments in oitro carried out in 1979 by Cox et al. [164] and using human intestinal biopsies demonstrated that human lactotransferrin can donate iron to intestinal mucosal cells. The characterization of a specific intestinal lactotransferrin receptor in rabbit [48], mouse [ 165,1661, Rhesus monkey [ 1671 and human foetal intestinal brush border membranes [ 1681 reinforces the concept of the role of lactotransferrin in intestinal iron absorption. However, despite these findings, the nutritional activity of lactotransferrin is still a subject of controversy (for a review, see ref. [7]). 2.3.2.1.2. Bacteriostasis and bactericidal effects. The antibacterial effect of lactotransferrin was first demonstrated by Masson and Heremans [169] and by Oram and Reiter [ 1701. This effect could be explained by iron-deprivation, iron being essential for the growth of numerous microorganisms. In fact, only apolactotransferrin is active. However, the mechanism of antibacterial activity is more complex and it is clear that lactotransferrin protects the mucosae in association with, at least, secretory IgA (sIgA) and lysozyme as first demonstrated by Bullen et al. [171] and Spik et al. [172,173]. The association of human sIgA with human (or bovine) lactotransferrin strongly inhibits the growth of pathological bacteria, in contrast to sIgA or lactotransferrin alone. The inhibitory power of the mixture is reinforced by the lysozyme which is present in human milk in a association with lactotransferrin in a molar ratio of 2: 1 [ 1741. The role of this association has been elucidated by Perraudin and Preels [ 1751 who showed that bacteria submitted to lysozyme action were agglutinated by lactotransferrin due to charge-charge interactions. Lactotransferrins also possess a bactericidal effect observed in 1977 by Arnold et al. [ 1761; according to Tomita et al. [ 177,1781 this is due to the liberation of short peptides like lactoferricin B (fragment 1 7 4 1 of bovine lactotransferrin) by partial proteolysis. Lactoferricin B derived from the peptidic segment binding to lactotransferrin receptors. The industrial interest in lactoferricin is evident. 2.3.2.1.3. Cell mediated defence of organism. A concise description of the different mechanisms of cell mediated defence of the organism by lactotransferrin would be an “impossible mission”. Consequently, the reader is referred to a series of excellent books and reviews [7,158-1621. Briefly, lactotransferrin has been shown to be involved in numerous inflammatory events (for reviews, see refs. [179,180]) and in immune response functions such as regulation of granulocyte monocyte colony stimulating factor synthesis with suppression of myelopoiesis by inhibition of the production of IL-1 (for a review, see ref. [lSl]),

218

regulation of interleukin synthesis [ 181,1821, natural killer cell activation and antitumor effects [129,183,184] and maturation of T- and B-cells [lS5-187]. 2.3.2.1.4. Growthfactor actiuity. Recent studies have demonstrated that lactotransferrin may promote growth of intestinal epithelial cells, suggesting that it might play a part in maturation of the intestine in the newborn (for reviews, see refs. [161,162]), of B-lymphocytes [1881 and of PHA-activated peripheral blood lymphocytes [ 189,1901.

2.3.2.2. Lactotransferrin receptors. The existence of a lactotransferrin receptor was first demonstrated by Van Snick and Masson in 1976 [191] at the surface of mouse peritoneal macrophages and lymphocytes. Since this discovery, the presence of lactotransferrin receptors has been demonstrated at the surface of various cells (for reviews, see refs. [156,158,192,193]): rabbit [48], mouse [165,166], monkey [167] and human [168] enterocytes; human HT29 and Caco-2 enterocyte cell lines [ 1941; human monocytes (reviewed in ref. [ 195]), human alveolar macrophages [ 1961, human neutrophils [ 195,1971, human resting lymphocytes [ 1971, human activated lymphocytes [ 1891, human Jurkat T cell line [190], human epithelial mammary cell line [198], human platelets [199,200] and megakaryocytes [201], hepatocytes [202,203] and in bacteria (for a review see refs. [204,205]). Contrary to serotransferrin receptors, little is known about the structure, physicochemical properties and gene expression of lactotransferrin receptors. Their main characteristics, which are totally different from those of serotransferrin receptors, can be summarized as follows: (i) Lactotransferrin receptors constitute a single peptide chain of about 1 10 kDa (100 to 13OkDa for intestinal receptor, 105kDa for lymphocyte and platelet receptors) [48,192,201]. A soluble fraction of 95 kDa is liberated by limited proteolysis [206]. (ii) Receptors are glycosylated but nothing is known about the structure and role of glycans [165,166,199]. (iii) Lactotransferrin receptors do not bind serotransferrin [48]. However, due to the high homology between the peptide chains of lactotransferrins, a given receptor is able to bind lactotransferrins originating from different species [ 1921. This explains the interest of milk companies in bovine lactotransferrin which is prepared in ton amounts and added to the food of young infants. (iv) The receptor binding site of lactotransferrins from different origins is located in the N-terminal part of the molecule covering amino acid residues 4 to 52. The conformation of this peptide segment of human and bovine lactotransferrin is very similar, explaining that bovine lactotransferrin is able to recognize human lactotransferrin receptor [ 1921. (v) Contrary to serotransferrin receptor which binds iron-saturated serotransferrin only, lactotransferrin receptor binds apo- and ferri-lactotransferrin 1481. 2.3.3. Ouotransferrin At the moment, we do not know if ovotransferrin can serve as an antibiotic protecting the egg and embryo development. In fact, if bacteria are present in the oviduct and appear in the egg-white, the bacteriostatic effect of ovotransferrin, which is only in the

219

apoprotein form in egg-white, could inhibit their growth by ferrideprivation. Additionally, it is possible that ovotransferrin plays a role in iron nutrition and in differentiation of the developing chick embryo. 2.3.4. Melanotransferrin The role of melanotransferrin has been recently elucidated by Kennard et al. [207] who demonstrated that this membrane bound iron binding protein is involved in the transferrin-independent uptake of iron in mammals but from iron-citrate and not from iron-transferrin complexes. This alternative iron uptake pathway may not function in the normal recirculation of iron within the body but might play a role during iron overload. On the other hand, rapidly proliferative tumor cells like melanocytes could use the alternative pathway to increase iron uptake. This independent system could also participate in the absorption of iron by intestinal cells that have no transferrin receptor on their lumenal surfaces [208], but express a transferrin-like GPI-linked iron-binding protein at the apical surface of fetal intestinal epithelial cells [209].

3. Comparative study of transferrin glycan primary structures The comparative study of sero-, lacto-, ovo- and melanotransferrins from different species or tissues shows (Table 3) that all are N-glycosylproteins which differ in their carbohydrate content (2.2 to 11.2% total sugars), in the number of glycans (1 to 4), except in serotransferrins of some fishes which are not glycosylated, and of antennae (2 to 4) as well as in the fucosylation rate. All of the glycans are of the N-acetyllactosaminic type, except in bovine, sheep and goat lactotransferrins which contain additional oligomannosidic structures (for reviews see refs. [ 119,2101). In addition, significant modifications of serotransferrin glycan primary structure have been observed in different physiological and pathological situations (see section 3.2.2). 3. I . Normal transferrin glycans

3. I . 1. Serotransferrin glycans As shown in Table 3 and in Figs. 6 to 10, serotransferrins of analyzed species possess 0 to 2 diantennary glycans of the N-acetyllactosaminic type, except human and fish serotransferrins which contain tri- and tetraantennary glycans, but in lower amounts. For example, in human serotransferrin (Fig. 6A-D), the relative proportions of serotransferrin having two triantennary glycans (glycovariant Tf-I), one diantennary and one triantennary glycan (glycovariant Tf-11) and two diantennary glycans (glycovariant Tf-111) (Fig. 6C,D) are 1, 17 and 82%, respectively [lOS]. In this regard, it is interesting to note, as described in section 3.2, that in some physiological or pathological situations, the amount of triand tetraantennary glycans increases significantly. Serotransferrin glycans are generally non-fucosylated except in human cerebrospinal fluid (trace amounts) [246,247], rat (20-30% of the molecules) [221], pig (100% of the molecules) [220] and the serotransferrin-like glycoprotein from mouse milk [2 18,2191 in which this protein co-exists with a lactotransferrin. None of the serotransferrin glycans

w 0 N

Table 3 Characteristics of sero-, ovo- and lactotransferrin glycans from different species Transferrins

Carbohydrate

Number of

Types and number of glycan

content ("h)

glycans

structuresa L

M

Fucosylation a-l,3b

Bisecting

GalNAc

Figure

a-1,6"

Serotransferrins Human

5.80

2

bi + tri +tetra

6A-D

Bovine

2.90 2.85 2.85-5.20 2.90 2.90 3.00 2.90 2.20 3.50

1

bi

6A,B

1

bi

16D 9 6A 7 6A 8 6A

Hen Horse Marsupial Mouse Rabbit Rat Sheep Fishe

1-2

bi

1

bi

1

bi + tetra

1

bi

1

bi

1

bi

1

tri tetra

+ +

+

10

Ovotransferrins Hen Turkey

2.65 3.70

tri bi

+ +

16B 16A continued on next page

Table 3, continued Transferrins

Carbohydrate content (“h)

Types and number of glycan structuresa

Number of glycans

Fucosylation ~

~

M

a-1,3b

a-1,6‘

+

+

Bisecting GlcNAc

GalNAc

Figure

L

Lactotransferrins

Human (milk) Human (1eukocyte)g Mouse Bovine Goat Pig

6.40

2

bi

2

bi

3.50

1

1 1.20

11.00

4 4

3.40

1

bi (4-2) bi(2)

+ + +

bi

+

bi 2 4 2

M, oligomannosidic type; L, N-acetyllactosaminic type; bi, tri, tetra: bi, tri and tetraantennary glycans, respectively. a-1,3, fbcose residue linked in C-3 position to GlcNAc of N-acetyllactosamine residue. a-1,6, fucose residue linked in C-6 position to the GlcNAc residue conjugated to the peptide chain. a

g

11B-D 11A 1 lB,C

+ + +

13-1 5 11B,C; 15B-D

Kangaroo, opossum and wallaby serotransferrin Carp big-head. Agalactoglycan. Polymorphonuclear leukocytes.

N

tl

222

NeuSAc(a2-6)Gal(fi14)GlcNAc(fi I -2)Man(a 14)

\ NeuSAc(a2-6)Gal(fi 14)GlcNAc(fi 1-2)Man(a 1-3)

Man(fi14)-R I

(A)

Gal(fi14)GlcNAc(fi 1-2)Man(a 1-6) \

NeuSAc(a2-6)Gal(fi 14)GlcNAc(fi 1 -2)Man(a 1-3)

han(fi1-4)-~ (B) I

NeuSAc(a2-3)Gal(fi14)GlcNAc(fi14)

\

Man(a1-6)

\

NeuSAc(a2d)Gal(fi 14)GlcNAc(fi 1-2)’

Man(fi14)-R

(C)

I

NeuSAc(a2-6)Gal(B 14)GlcNAc(fi1-2)Man(a 1-3) NeuSAc(a2-6)GaI(fi 14)GlcNAc(fi 1-2)Man(a 1-6)

\ NeuSAc(a2-6)Gal(fi 14)GlcNAc(fi 1-2)

NeuSAc(a2-3)Gal(fiI 4)GlcNAc(P 1 4 )

\

Man(fi14)-R

(D)

I

Man(a1-3) I

Fig. 6. Primary structure of serotransferrin diantennary glycans from human (A,B) [211-213], cow (A,B) [ I 19, 2141, rabbit (A) [103], sheep (A) [ I 19,2151, marsupial (kangaroo, opossum, wallaby) (A) [ I 19,2161. Primary structure of human serotransferrin triantennary glycans (C,D) [I 19-21 71. R, GlcNAc(fi14)GlcNAc(fil-N)Asn.

contains a bisecting residue of N-acetylglucosamine residue, except fish serotransferrin and human cerebrospinal fluid transferrin (Figs. 10A and 17A). The only sialic acid found to date in serotransferrin glycans is N-acetylneuraminic acid, except in horse serotransferrin glycans which contain 4-O-acety1, N-acetylneuraminic acid in addition to N-acetylneuraminic acid (Fig. 9) and in mouse serotransferrin in which only N-glycolylneuraminic acid is present (Fig. 7).

3.1.2. Lactotransferrin glycans Lactotransferrins from human, bovine, porcine, caprine and murine milk are N-glycosylproteins with diantennary glycans of the N-acetyllactosaminic type, a- 1,6-fucosyIated on the N-acetylglucosamine residue linked to the peptide chain. Only human lactotransferrin has a-l,3-fucosylated N-acetyllactosamine residues (Fig. 11). Cow, sheep and goat lactotransferrins contain additional glycans of the oligomannosidic type (Fig. 15). Only human lactotransferrin possesses poly-N-acetyllactosaminic glycans (Fig. 12). Bovine lactotransferrin glycans present a great microheterogeneity (Fig. 13). They are characterized a-1,3-linked galactose residues in terminal position (Fig. 14) and by N-acetylgalactosamine residues replacing galactose residues (Fig. 14) as in caprine lactotransferrin glycans. Interestingly, the glycan primary structure of lactotransferrin extracted from human polymorphonuclear leukocytes is identical to that of the non-fucosylated diantennary

223

Neu5Gc(a2-6,3)Gal(fi14)GlcNAc(fi1-2)Man(a 1-6)

\ Man(fil+)-R I

Neu5Gc(a24,3)Gal(fi14)GlcNAc(fi 1-2)Man(a 1-3) NeuSGc(a2-6)Gal(fi14)GlcNAc(fi1-2)Man(a I d )

\ Man(fi14)-R I

NeuSGc(a2-3)Gal(fi1-3)GlcNAc(fi1-2)Man(al-3)

I (a241

NeuSGc

NeuSGc(a2-3)Gal(fi14)GlcNAc(fi14)

\

Man(a1-6)

\

Neu5Gc(a2-3,6)Gal(fi14)GlcNAc(fi 1-2)’ NeuSGc(a2-6)Gal(fil 4)GlcNAc(fil -2)Man(a 1-3)

Man@ I 4 ) - R I

NeuSGc(a2-3)Gal(fi14)GlcNAc(fil4)

\

Man(al-6)

\

NeuSGc(a2-6)Gal(fi 1 4)GlcNAc(fi 1-2)’ NeuSGc(a2-6)Gal(fi 1 4)GlcNAc(fi1-2)

Man(fil4)-R \

Man(a1-3)

I

I

NeuSGc(a2-3)Gal(fi14)GlcNAc(fi 14)

Fig. 7. Primary structure of the glycans from mouse serotransferrin [ 1 19,2 18,2 191 R, GlcNAc(fil+)[Fuc(a 1-6)10-, GlcNAc(kN)Asn.

glycan of human serotransferrin (Fig. I IA). This data should be taken into consideration when human milk fucosylated lactotransferrin is used in binding experiments with various cells, mainly with cells of the immune system. 3.1.3. Ouotransferrin glycans In contrast to sero- and lactotransferrins, glycans of ovotransferrins from avian egg-white contain a bisecting N-acetylglucosamine residue, like other glycoproteins from oviducts, such as ovomucoid and ovalbumin for instance (Fig. 16A-C). Like all of the avian egg glycoproteins, ovotransferrins are not fucosylated. Concerning the hen sero- and ovotransferrin glycans, it is worthwhile noticing that they present completely different primary structure, whereas both transferrins which derived from the same gene [27]possess exactly the same polypeptide chain [69].In fact, hen serotransferrin glycan has a diantennary sialylated structure (Fig. 16D). This result does not favor the hypothesis that the protein moiety controls the structure of its own glycans, but rather that the glycan primary structure depends only upon the enzyme equipment of secreting cells, serotransferrin being synthesized by the liver, and ovotransferrin by the oviduct.

224 NeuSAc(a2-6)GaI((3 I -4)GlcNAc(p 1-2)Man(a 1-6)

\ Man(p14)-R I

(A)

NeuSAc(a2-3)Gal(fiI -3)GlcNAc(fi 1-2)Man(a 1-3)

I~

4

)

NeuSAc

Neu5Ac(a2-3)Ga1((31-3)GlcNAc(~l-2)Man(a 1-6)

\ Man(fi1-4-R I

(B)

NeuSAc(a2-3)Gal(fiI -3)GlcNAc(p1-2)Man(a 1-3)

I

032-6)

NeuSAc NeuSAc(a2-6)Gal(fi 14)GlcNAc(fi 1-2)Man(a 14)

\ NeuSAc(a2-6)Gal(p 14)GlcNAc@1-2)Man(a 1-3)

Man(fi14)-R I

(C)

[NeuSA~(a2-6)]~-, Gal@14)GlcNAc(fi1-2)Man(a 1 4 )

\ Man(fil4)-R

(D)

I

Neu5Ac(a2-6)Gal(fl1-4)GlcNAc(~l-2)Man(a 1-3) Fig. 8. Primary structures of the glycans from rat serotransferrin (A,B,C) [221] and from rat mammary gland GlcNAc(fi-N)Asn. transferrin (D) [67]. R, GlcNAc(~l-4)[Fu~(a1-6)]~-~

NeuSAc(a2-6)Gal(f314)GlcNAc(~I-2)Man(a 1-6)

\ NeuSAc(a2-6)Gal(fi 1-4)GlcNAc(p 1-2)Man(a 1-3)

Man(p14)-R I

NeuSAc(a2-6)Gal(f~I4)GlcNAc(fi1 -2)Man(a 1 4 )

\ Neu4,5Ac2(a2-6)Gal(fi1 4)GlcNAc(p 1-2)Man(a 1-3)

Man(b14)-R I

Neu4,5Ac2(a2-6)Ga1(@I-4)GlcNAc(fi 1-2)Man(a 1-6)

\ NeuSAc(a2d)Gal(fi14)GlcNAc(~1-2)Man(a1-3)

Man(p1 4 ) - R I

Neu4,5Ac2(a2-6)Gal(fi1 4)GlcNAc(fi1-2)Man(a 1 4 )

\ Neu4,5Ac2 (a2-6)Gal(D 14)GlcNAc@1-2)Man(a 1-3)

Man(fi1-4)-R I

Fig. 9. Primary structure of the four glycans identified in the three horse serotransferrin variants [113,114,119, 222,2231. R, GlcNAc(fi14)GlcNAc(flI-N)Asn.

225 [Gal(fiI 4)GlcNAc(fi 1-6)]o-l

\

[Neu5A~(a2-3)]~_~ Gal(@14)GlcNAc(fiI-2)Man(a 14)

\ (A)

GlcNAc(fi14)-Man(fil-4-R I

[Neu5Ac(a2-3)lo_, Gal(fiIL4)GlcNAc(@1-2)Man(a 1-3) [Gal(fiI 4)GlcNAc(B 14)Io-l NeuSAc(a2-3)Gal(fi 14)GlcNAc(fi1 -2)Man(a 1-6)

\ Man(@I 4 ) - R

NeuSAc(a2-3)Gal(fi14)GlcNAc(fiI-2)Man(a1-3)

(B)

I

Fig. 10. Primary structure of glycans from fish serotransferrins. A, Carp big-head (Aristichlhys nobilis) [224]; B, pike (Esox lucius) [225]. R, GlcNAc(fiI4)GlcNAc(fil-N)Asn.

NeuSAc(a2-6)Gal(fi I4)GlcNAc(@1-2)Man(al-6)

\ Man@14)GlcNAc(fi 1-rl)GlcNAc(@1-N)Asn NeuSAc(a2-6)Gal(fiI 4)GlcNAc(fiI -2)Man(a 1-3)

(A)

I

Gal(fi1-4)GlcNAc(fi1-2)Man(al-6)

\ Man(fi14)GlcNAc(fi 14)GlcNAc(fiI-N)Asn I I @l+) Neu5Ac(a24)Gal(fi14)GlcNAc(fil-2)Man(al-3) Fuc

6'

5'

(B)

4'

Neu5Ac(a2-6)Gal(fil-4)GlcNAc(fi1-2)Man(a1-6)

2 1 \3 Man@I 4)GlcNAc(fi14)GlcNAc(fil-N)Asn (C) I I W+) NeuSAc(a2-6)Gal(fi14)GlaNAc(fiI-2)Man(a1-3) Fuc 6 5 4 1' Fuc

I (at-3)

Gal@14)GlcNAc(@1-2)Man(a1-6)

\ Man@1 -4)GlcNAc(fi 14)GlcNAc(fiI-N)Asn I

NeuSAc(a2-6)Gal(fil4)GlcNAc(fiI-2)Man(al-3)

(D)

I (aid) Fuc

[Ne~5Ac(a2-3)]~_~ Gal(fiI4)GlcNAc(fi1-2)Man(a1-6)

\ Man(fi14)GlcNAc(fiI4)GlcNAc(fiI -N)Asn

Gal(fiI4)GlcNAc(fiI-2)Man(a1-3) [Neu5A~(a2-3)]~_~

I

I

(E)

(U1-h)

[Fuclo-1

Fig. 1 I . Primary structure of the glycans from human leukocyte lactotransferrin (A) [226] and human (B,C,D) [211,227-2301, Rhesus monkey (A) [231], sheep (A) [232], goat (B,C) [I 19,228,2321, and mouse (B,C) [ I 19, 218,2331 milk lactotransferrins. E, human recombinant lactotransferrin expressed in BHK cells [234].

226

Fuc

I (al-3)

Gal@ I--I)GlcNAc(fiI-3)Gal(fiI -4)GlcNAc(fiI -2)Man(a 1-6)

\ Man@ I--I)-R

Gal(fi14)GlcNAc(fl1-2)Man(a 1-3; Fuc

I

(Ul-3)

Gal(fiI4)GlcNAc(fiI-2)Man(a1 4 )

\ Man(fil4)-R NeuSAc(a2-6)Gal(fi 14)GlcNAc(fi 1-3)Gal(fi 14)GlcNAc(fiI-2)Man(a 1-3; Fig. 12. Primary structure of the poly-N-acetyllactosaminic glycans from human lactotransferrin [235]. R, GlcNAc(fl1-4)[Fuc(a 1 -6)]o-l GlcNAc(kN)Asn.

Gal(a 1-3)Gal(fi14)GlcNAc(fi 1 -2)Man(a 1-6)

\ Man@ I --I)GlcNAc(fi 14)GlcNAc(fi 1-N)Asn /

NeuSAc(a24)Gal(fi14)GlcNAc(fiI-Z)Man(a1-3) [NeuSAc(a2-6)]GalNAc(fi14)GlcNAc(fi 1-2)Man(a 1-6)

\ Man(fil--I)GlcNAc(fil4)GkNAc(fiI-N)Asn NeuSAc(a2-6)Gal(fi14)GlcNAc(fiI-2)Man(a 1-3)

I

GlcNAc((3-N)Asn. Fig. 13. Microheterogeneity of cow lactotransferrin [236]. R, GlcNAc(fil~)[Fuc(a14)],~, NeuSAc(a24)Gal( p 1-4) Gal(ald)Gal@ 1 4 ) Gal(P1-4)

Gd(p14)i

NeuSAc(a2-6)Gal(Pl-4) , 1 NeuSAc(a24)GalNAc(Pl-4) GalNAc( p 1 4 )

GlcNAc(pl-2)Man(al-6)

>

GlcNAc(pl-2)Man(al-3)

\

Man(pl-4tR

/

/

Fig. 14. Primary structure of diantennary glycans from cow lactotransferrin with an a-1,3-Gal residue in the terminal position [I 19,2361 and with a GalNAc residue replacing a Gal residue.

3.1.4. Concluding remarks

In conclusion, the characteristics of the carbohydrate moieties of transferrins are specific to each of them on the basis of the primary structure of glycans and of their number and position in the peptide chain as demonstrated in Figs. 2 to 4. In addition, each transferrin presents unexplained microheterogeneity.

221

Man(fi14)-R Man(al-2)Man(a 1-2)Man(a 1-3)

(A)

I

Man(a1-6)

\

Man(a1-6)

\ Man(bI4)-R Man(a 1-3)

(B)

I

Man(a 1-6)

\

Man(a1-6)

\

Man(a1-3)’

Man(fil4)-R

(C)

I

Man(a 1-3) Man(a1-2)Man(a 1-6)

\

Man(a 1-6) Man(alL2)Man(a 1-3)’

\ Man(fil4)-R

[Man(a1-2)J0_,Man(aIL3)Man(a 1-3)

(D)

I

Fig. 15. Primary structure of the glycans of the oligomannosidic type from cow [236,237] (A-D), goat (BD) [I 19,2321, sheep (B-D) [232] and Rhesus monkey (D) [231] milk lactotransferrins. R, GlcNAc(b14)GlcNAc(fi1-N)Asn.

3.2. Physiopathological modijications of transferrin glycan primary structure Human serotransferrin presents a microheterogeneity based on the co-existence of biand triantennary glycans of the N-acetyllactosaminic type leading to three glycovariants. The characterization and quantitation of these glycoforms is easily carried out by immuno-affinity electrophoresis as well as their isolation by affinity chromatography on concanavalin A-Sepharose columns. In this way, serotransferrin (TO isolated from healthy donors is resolved into three glycovariants: Tf-I (less than 1%), Tf-I1 (17 f2%) and Tf-I11 (82 3%), containing, two triantennary glycans, one tri- and one diantennary glycan and two diantennary glycans [ 1081, respectively. There are two “isomers” of the triantennary glycans (ratio 1 : 1) in which the third antenna is either E-l,4-linked to the a - 1,3-mannose residue or fi-1,6-linked to the a-1,6-mannose residue. Variations in the structure and location of serotransferrin glycans were detected in physiological and pathological cases [242].

*

3.2.1. Physiological modijications Serotransferrin is involved in the process of iron transfer to the foetus by a placental receptor-mediated mechanism. The three glycovariants of serotransferrin and their binding to human syncytiotrophoblast microvillar membranes have been studied by Ltger et al. [108]. The results obtained by these authors and others[243,244] showed that in the serum of pregnant women, especially in the last 3.2.1.1. Pregnancy.

228

three months of pregnancy, the serum concentration of serotransferrin reached 4.5-5 g per liter and the relative proportions of the glycovariants Tf-I and Tf-I1 increased from 1 to 6 f 1% and from 17 2 to 26 3~3%, respectively, while that of Tf-I11 decreased from 82 3 to 67 3~3%. In addition, an increase of sialylation has been observed [243]. The binding of the three serotransferrin glycovariants to the receptor of the syncytiotrophoblast plasma membranes has been studied and no difference in the binding affinity has been observed [108].

*

*

3.2.1.2. Embryogenesis. Changes in glycosylation of chicken serotransferrin synthesized during embryogenesis and by primary cultures of chicken embryo hepatocytes have been observed by Jacquinot et al. [241]. In the three transferrins analysed, the glycans were of the diantennary N-acetyllactosaminic type, having several prominent features. In particular, the embryo serotransferrin glycan (Fig. 16C) differed from that of chicken serotransferrin (Fig. 16D) by the presence of a bisecting N-acetylglucosamine residue, suggesting a developmental change in glycosylation. The glycan structure of the transferrin secreted by the embryo hepatocytes in primary culture was marked by the presence of fucose a-1,6-linked to the core N-acetylglucosamine, suggesting that expression of the fucosyltransferase activity is dependent on cell culture conditions. Moreover, comparative analysis of chicken serotransferrin (Fig. 16D) and ovotransferrin (Fig. 16B) glycans reinforces the idea that the glycosylation of two identical polypeptide chains is organ specific. [Gal(fi 14)10-, GlcNAc(fi1-2)Man(a 1-6)

\ GlcNAc(fi14)-Man(fi 14)-R I [Neu5Ac(a2-6)lo_, Gal(fi1-4)GlcNAc(f3 1-2)Man(a 1-3)

(A)

GIcNAc@-2)Man(a 1-6)

\ GlcNAc(fiI--l)-Man(fiI4)-R(B) GlcNAc(fi1-2)

\ Man(a1-3)

I

I

GlcNAc(fl 1 4 )

Gal(pl4)GlcNAc(flI-2)Man(a 1-6)

\ GlcNAc(fiIH)-Man(fil4)-R I Neu5Ac(a2-6)Gal(fi14)GlcNAc(fi1-2)Man(a 1-3)

(C)

[Neu5Ac(a2-6)],-, Gal@ 14)GlcNAc(fi I-2)Man(a 1 4 )

\ Man(fil-4)-R

(D)

Fig. 16. Primary structure of glycans from turkey ovotransferrin (A) [ I 19,2381, hen ovotransferrin (B) [239, 2401, chicken embryo serum (C), and chicken serotransferrin (D) [241]. R, GlcNAc(fiI-4)GlcNAc(fiI-N)Asn. Glycans of transferrin from embryo hepatocytes secreted into culture medium are a-I ,6-fucosylated.

229 GlcNAc(fi1 -2)Man(a 1-6)

\ GlcNAc@4)-Man(fiI 4)GlcNAc(fi14)GIcNAc(fil -N)Asn GlcNAc(fi1-2)Man(a 1-3)

I

I

(A)

(Ul-6)

Fuc

Gal(fl14)GlcNAc(fiI-2)Man(a 1-6)

\ Man(fil4)GlcNAc(fi 1 -4)GlcNAc(fil-N)Asn Gal@ 1 -4)GlcNAc(fi1-2)Man(a1-3)

(B)

I

Fig. 17. Major glycan structures of human cerebrospinal fluid transferrin [246,247]. Part of glycan B is sialylated.

3.2.1.3. Tissue-dependent glycosy lation. 3.2.1.3. I. Human cerebrospinal jluid transferrin. Transferrin is an important protein constituent of human cerebrospinal fluid, some 30% being present in an unsialylated form [245]. The glycan primary structure of the so-called fiz-transferrin or T-globulin, has been recently determined by Hoffmann et al. [246]. The major structure turned out to be an N-acetyllactosaminic type agalacto-diantennary oligosaccharide with bisecting N-acetylglucosamine and proximal fucose (Fig. 17A). Analysis of a second transferrin preparation containing both asialo- and sialotransferrin revealed another major glycan species derived from the sialylated transferrin variant which is galactosylated and lacks fucose and bisecting N-acetylglucosamine (Fig. 17B). The first result supports the hypothesis of a de nouo “brain-type’’ glycosylation of intrathecally synthesized cerebrospinal fluid proteins and the second one is in favor of an enzymatic desialylation of serotransferrin. 3.2.1.3.2. Human seminal transferrin. Human seminal transferrin (MM: 80 m a ) contains 6.1% sugars. The primary structure of the major N-linked glycan is identical to that of the serotransferrin diantennary glycan of Fig. 6A [247]. 3.2.1.3.3. Rat mammary-gland transferrin. Rat milk transferrin contains four glycovariants that differ only in their sialic acid content. The primary structure of the two major variants has been determined by Escriva et al. [67]. As shown in Fig. XD, the glycoforms contain either one or two N-acetylneuraminic acid residues a-2,6-linked to galactose in a conventional diantennary glycan of the N-acetyllactosaminic type. Most contain fucose a-1,6-linked to the proximal N-acetylglucosamine residue. 3.2.1.3.4. Recombinant transferrins. In view of the potential importance of human lactotransferrin for the production of artificial milk and for the pharmaceutical industry, numerous attemps have been made to produce recombinant human lactotransferrin using different cells: BHK cells [234,248], yeast, Aspergillus orizae [249], Saccharomyes cereuisiae [250] and tobacco cells [25 11; or transgenic animals [252]. But production of recombinant glycoproteins represents a formidable challenge. In fact, the yields are often low and, in addition, the glycan primary structures do not conform to the native ones (for reviews, see refs. [253,254]). Legrand et al. [234] have shown that expression in BHK cells of a full-length cDNA coding for human lactotransferrin led to diantennary glycans of the N-acetyllactosamine type, a-2,3-disialylated

230

(80%) and a-2,3-monosialylated (20%) forms. In addition, 70% of total glycans were a-1,6-fucosylated at the proximal GlcNAc residue, as in the native human milk lactotransferrin. The absence of N-acetylneuraminic acid a-2,6-linked to the terminal galactose residue is explained by the absence of a-2,6-sialyltransferase in BHK cells. 3.2.2. Serotransferrin and disease Serotransferrin glycans are highly sensitive to numerous pathological modifications of protein glycosylation. For this reason, serotransferrin is used as a tool to probe changes in N-glycosylation in pathological situations, particularly in liver diseases (for reviews, see refs. [242,255,256]). 3.2.2.1. Liver diseases. By applying the crossed affino-electrophoresis method, Spik et al. [255,256] have demonstrated that the human serotransferrin glycans are profoundly modified in liver diseases such as viral hepatitis and alcoholic cirrhosis with a marked increase of triantennary glycans (see also refs. [257-2591). 3.2.2.2. Cancer. The same kind of alterations have been observed by Yamashita et al. [260] in the serotransferrin of patients with hepatocellular carcinoma: dramatic increase of tri- and tetraantennary structures, a- 1,3-fucosylation of peripheral N-acetyllactosamine residues and, in small amounts, presence of bisected diantennary glycans. Identical results have been obtained from HepG2 cell line transferrin. In fact, profound modifications of glycan structure of serotransferrin synthesized by the human hepatocarcinoma cell line HepG2 have been shown by Campion et al. [261]. A comparative study of normal serotransferrin and of HepG2 transferrin shows the presence of tri-, tetra- and pentaantennary glycans of the N-acetyllactosaminic type, with fucose residues a-1,3-1inked to peripheral N-acetylglucosamine residues (Fig. 18). These results indicate that the increase Gal(@14)GlcNAc(@ 1 4 ) [Fuc(a 1-3)]

\

Man(a 1-6)

I

\

Gal(@14)GlcNAc(@1-2)/ [NeuSAc(a2-3)I2 Gal(fiI4)GlcNAc(@1-2) \

[NeuSAc(a24)], Gal@14)GlcNAc(@1 4 )

Man(al-3) I

Man(@14)-R I

Gal(@14)GlcNAc(@1 4 )

\

[Fuc(a

I

[NeuSAc(a2-3)I2

Gal(@I4)GlcNAc(@14)-Man(a 1-6) Gal(@14)GlcNAc(@1-2)l Gal(@1-4)GlcNAc(/3-2) \

Man(a1-3)

[NeuSAc(a24)I2 Gal(@14)GlcNAc(@1-4)

\

Man(@1 4 ) - R I

I

Fig. 18. Primary structure of tetra- and pentaantennary glycans from human serotransferrin secreted into culture medium of human hepatocarcinoma cell line Hep G2. The a-l,3-linked fucose residue is conjugated to the GlcNAc of one of the antennae [261]. R, Gl~NAc(~14)[Fuc(a1-6)]~~~GlcNAc(~-N)Asn.

23 1

in the number of antennae in transferrin glycans synthesized by the hepatocarcinoma cell line is much more pronounced than in liver diseases such as alcoholic cirrhosis and that, in addition, the malignant transformation of human liver induces fucosylation. These results might be due, at least in part, to the regulation of N-acetylglucosaminyltransferase V activity as has been observed in numerous cancer cells [262]. 3.2.2.3. CDG syndromes. The carbohydrate-deficient glycoprotein (CDG) syndromes are a family of genetic multisystemic diseases with severe nervous system involvement, growth retardation and hepatopathy during infancy which were first reported by Jaeken et al. [263] (for recent revicws, see refs. [264,265]). CDG-type I syndrome [264,266,267] is due to a deficiency in the oligosaccharidyltransferase which transfers “en bloc” onto the nascent protein the oligosaccharide linked to dolichol diphosphate. Later, van Schaftingen and Jaeken [ 1995, FEBS Lett. 377, 3 18-3201 demonstrated that the syndrome was due in fact to a phosphomannomutase deficiency, an enzyme which provides the mannose- 1-phosphate required for the initial steps of protein glycosylation. This leads to four transferrin isoforms: non-glycosylated, glycosylated in Asn-4 13 or in Asn-6 1 1 and in both Asn-4 13 and 6 1 1. CDG-type I1 syndrome [265,268] is a separate variant since it is characterized by a severe decrease in the activity of N-acetylglucosaminyltransferase I1 (UDPGlcNAc: a6-D-mannoside fi- 1,2-N-acetylglucosaminyltransferase).As a consequence, the serotransferrin isoforms contain two truncated monoantennary glycans of which the primary structures are described in Fig. 19. Man(a 1-6)

\ NeuSAc(a2-6)Gal(~14)GlcNAc(P1-2)Man(a1-3)

Man(pl-4)-R i

Fig. 19. Primary structure of the glycan from human serotransferrin isolated from a patient with carbohydratedeficient syndrome (CDG) type I1 [265,268]. R, GlcNAc(PI-4)GlcNAc((ll-N)Asn.

3.2.2.4. Serotransferrin in HEMPAS. Congenital dyserythropoietic anaemia type I1 or HEMPAS (hereditary erythroblastic multinuclearity with positive acidified serum lysis test) is a genetic disease caused by membrane disorganisation of erythroid cells (for a review, see refs. [269,270]). A defect in N-acetylglucosaminyltransferase 11 or in a-mannosidase I1 has been suspected. As a result, the serum glycoproteins, serotransferrin in particular, are taken up by the hepatocytes and the Kuppfer cells. Analysis carried out by Fukuda et al. [269,270] showed the presence of oligomannosidic (M6 to M9 structures) and hybrid type glycans in the HEMPAS transferrin.

4. Three-dimensional structure of transferrins As mentioned above (see section 2.2. l), the three-dimensional structure of the peptide chains of rabbit serotransferrin, of human and bovine lactotransferrins and of ovotrans-

232

ferrin have been defined by X-ray diffraction. However, the X-ray diffraction patterns provided little information on the 3D-structure of the glycans and on their relationships and interactions with the peptide chain. For example, in the case of human lactotransferrin only the disaccharide Fuc(a I-6)GlcNAc linked to asparagine is visible [97]. 4.1. Three-dimensional structure of glycans

The only view we had in the 1970s of the 3D-structure of glycans was mainly speculative. In fact, it resulted from molecular building of the diantennary glycan from human serotransferrin and lactotransferrin whose glycan primary structure had just been determined [211,212]. This approach led to the description of the Y, T, bird and broken-wing conformation (for reviews, see refs. [210,227,253,27 1-2741). This situation was due to the difficulties encountered in the crystallization of glycans and glycoproteins. However, little by little, the application of sophisticated techniques like NMR, EPR and neutron scattering led to results which favoured the emerging concept of conformation interconversions due to the mobility of antennae [227,271-2741. This concept was recently verified on the basis of molecular modelling [275,276], molecular dynamics simulations [277] and X-ray diffraction [278-2831 data concerning human sero- and lactotransferrin. Results obtained from X-ray diffraction are unambiguously demonstrative. In fact, Cambillau et al. (for a review, see ref. [278]) have devised an elegant method of co-crystallisation of the iso-lectins I and I1 of Lathyrus ochrus with glycans and glycoproteins. A free glycan [279] and a so-called N2-peptide [280,283] from human lactotransferrin were co-crystallized with the lectin and X-ray diffraction data from single crystals were collected at 2.3 A resolution. The results are illustrated in Fig. 20 which shows that the 3D-structure of the free glycan (bird-conformation) and of the N2-glycopeptide (broken wing-conformation) are different. Thus, the concept

Fig. 20. Spatial conformation (a) of a free asialoglycan and (b) of the N2-asialoglycan from human lactotransfcrrin, determined by X-ray diffraction [283]. Numbers correspond to the numbering used in structure C of Fig. 1 1.

233

of flexibility of antennae proposed about 20 years ago[227,272] (for reviews, see refs. [2 10,272-2741) is now firmly established by experimental data. .4.2. Three-dimensional structure of transferrins as glycoproteins As mentioned above, X-ray diffraction of transferrin furnishes little information on the 3D-structure of the glycans and the images we have today remain largely speculative since they result from molecular modelling studies. We have represented in Fig. 21 the 3D-structure, determined by molecular modelling on the basis of X-ray diffraction data of rabbit serotransferrin [276] and of human lactotransferrin [89,92]. In rabbit serotransferrin, the single glycan linked to the peptide chain is immobilized into only

Fig. 21, Molecular modelling (A,B) of rabbit serotransferrin glycan and (C) of human lactotransferrin [192, 210,2751: (A) 3D structure of rabbit serotransferrin; (B) interaction of rabbit serotransferrin glycan in a broken-wing conformation with a peptide segment (amino acids 254 to 271) in an a-helix conformation, 7,7', N-acetylneuraminic acid residues (see Fig. 6A). (C) 3D structure of human lactotransferrin. Arrows indicate the position of glycans.

234

one conformation, the broken wing-conformation. In addition, the glycan is conjugated to the Asn-485 residue of the C-lobe while its two antennae interact with a peptide portion (amino acid residues 254 to 271 in an a-helix conformation) of the N-lobe (Figs. 21A and B). This result strongly supports the hypothesis that, in the case of rabbit serotransferrin, the glycan might reinforce the association of the two lobes and contribute to maintain the protein moiety in a biologically active 3D-structure. On the contrary, in human lactotransferrin, the glycans do not seem to interact with the peptide chain (Fig. 21C). They are free in space and could thus play a role as recognition signals. In addition, lactotransferrin glycans are in opposite positions in the N-and C-lobes while serotransferrin glycans (not shown) are both located in the C-domain of the protein and are very near in space.

Acknowledgements The authors are very grateful and want to thank Laurence Onraet and Khadija Khalfa for their skilful assistance in typing this manuscript.

Refeven ces [ I ] Sutton, H.E. and Jamieson, G.A. (1972), In: A. Gottschalk (Ed.), Glycoproteins. Their Composition, Structure and Function. Elsevier, Amsterdam, Part A, pp. 6 5 3 4 9 8 . [2] Aisen, P. and Listowsky, I. (1980). Annu. Rev. Biochem. 49, 357-393. [3] Spik, G., Montreuil, J., Crichton, R.R. and Mazurier, J. (1985), Proteins of Iron Storage and Transport, Proc. 7th Intern. Conference on Proteins of Iron Metabolism, Villeneuve d'Ascq (France), 30 June-5 July 1985. Elsevier, Amsterdam. [4] Baker, E.N., Rumball, S.V. and Anderson, B.F. (1987) Trends Biochem. Sci. 12, 350-353. [S] Arosio, F'., Cairon, G., Levi, S. (1989) In: M. de Sousa and Brock J.H. (Eds.), Iron in Immunity, Cancer and Inflammation. Wiley, New York, pp. 55-79. [6] De Jong, G., Van Dijk, J.P. and Van Eijk, H.G. (1990) Clin. Chim. Acta 190, 1-46, [7] Crichton, R.R. (1 991) Inorganic Biochemistry of Iron Metabolism. Ellis Honvood, Chichester. [8] Crichton, R.R. and Ward, R.J. (1992) Biochemistry 31, 11255-11264. [9] Crichton, K.R. and Ward, R.J. (1992) In: R.B. Lauffer (Ed.), Iron and Human Disease. CRC Press, London, pp. 23-75. [lo] Baldwin, G.S. (1993) Comp. Biochem. Physiol. 106B. 203-218. [ l l ] Hughes, R.C. and Butters, T.D. (1981) Trends Biol. Sci. 228-230. [I21 Van Eijk, H.G., Van Noort, W.L., Kroos, M.J. and Van der Heul, C. (1982) Clin. Chim. Acta 121, 209-216. [I31 Van Eijk, H.G., Van Noort, W.L., De Jong, G. and Koster, J.F. (1987) Clin. Chim. Acta 165, 141-145. [I41 Mason, A.B., Miller, M.K., Funk, W.D., Banfield, D.K., Savage, K.J., Oliver, R.W.A., Green, B.N., MacGillivray, R.T.A. and Woodworth, R.C. (1 993) Biochemistry 32, 5472-5479. [ 151 Montreuil, J., Vliegenthart, J.F.G. and Schachter, 11. (Eds.) ( 1 995), Glycoproteins, New Comprehensive Biochemistry, Vol. 29a. Elsevier, Amsterdam. [ I 61 Montreuil, J., Vliegenthart, J.F.G. and Schachter, H. (Eds.) (1997), Glycoproteins 11, New Comprehensive Biochemistry, Vol. 29b. Elsevier, Amsterdam. [I71 Laurel], C.B. and Ingelman, B. (1947) Acta Chem. Scand. I , 770-776. [I81 Williams, J. (1985) In: G. Spik, J. Montreuil, R.R. Crichton, J. Mazurier (Eds.), Proteins of' Iron Storage and Transport. Elsevier, Amsterdam, pp. 13-23. [19] Osborne, T.B. and Campbell, G.F. (1900) J. Am. Chem. Soc. 22, 4 2 2 4 2 6 .

235 [20] Schade, A.L. and Caroline, L. (1944) Science 100, 1 4 1 5 . [21] Schade, A.L. (1985) In: G. Spik, J. Montreuil and J. Mazurier (Eds.), Proteins of Iron Storage and Transport. Elsevier, Amsterdam, pp. 3-1 2. [22] Schade, A.L., Reinhart, R.W. and Levy, H. (1949) Arch. Biochem. 20, 170-172. [23] Alderton, G., Ward, W.H. and Fevold, H.L. (1946) Arch. Biochem. 11, 9-13. [24] Schade, A.L. (1966) Protides in the Biological Fluids, pp. 13-23. [25] Schade, A.L. and Caroline, L. (1 946) Science 104, 340-341, [26] Schade, A.L. (1963) Biochem. Z. 338, 140. [27] Williams, J. (1962) Biochem. J. 83, 355. [28] Fontes, G. and Thivolle, L. (1925) C.R. SOC.Biol. Pans 93, 687. [29] Starkenstein, S. and Harvalik, Z. (1933) Arch. Exptl. Pathol. Pharmakol. 172, 75. [30] Surgenor, D.M., Koechlin, B.A. and Strong, L.E. (1949) J. Clin. Invest. 28, 73-96. [31] Holmberg, C.G. and Laurel, C.B. (1947) Acta Chem. Scand. 1, 944-950. [32] Koechlin, B.A. (1952) J. Am. Chem. SOC74, 2649. [33] Montreuil, J. and Mullet, S. (1960) C.R. Acad. Sci. Pans 250, 1736-1737. [34] Montreuil, J., Tonnelat, J. and Mullet, S. (1960) Biochim. Biophys. Acta 45, 413421. [35] Johansson, B. (1960) Acta Chem. Scand. 14, 5 10-5 12. [36] Blanc, B. and Isliker, H. (1961) Bull. SOC.Chim. Biol. 43, 929-943. [37] Masson, P.L., Heremans, J.F. and Prignot, J. (1965) Experientia 21, 604-605. [38] Ssrensen, M. and Ssrensen, S.P.L. (1939) C.R. Lab. Carlsberg 2B, 55-99. [39] Groves, M. (1960) J. Am. Chem. SOC.82, 3345-3350. [40] Oram, J.D. and Reiter, B. (1968) Biochim. Biophys. Acta 170, 351-365. [41] Jolles, J., Donda, A,, Amiguet, P. and Jollks, P. (1984) FEBS Lett. 176, 185-188. [42] Davidson, L.A. and Lonnerdal, B. (1986) Pediatr. Res. 20, 197-201. [43] Kinkade, J.M., Kendall-Miller, W.K. and Segars, S.F.M. (1976), Biochim. Biophys. Acta 446,40741 8. [44] Sawatzki, G., Hoffmann, F.A. and Kubanek, B. (1983), Infect. Immun. 39, 659465. [45] Godovac-Zimmermann, J. (1988) Biol. Chem. Hoppe Seyler 369, 93-96. [46] Roberts, T.K. and Boursnell, J.C. (1975) J. Reprod. Fertil. 42, 579-582. [47] Hutchens, W., Magnuson, J.S. and Yip, T.T. (1989) Biochim. Biophys. Acta 999, 323-329. [48] Mazurier, J., Montreuil, J. and Spik, G. (1985) Biochim. Biophys. Acta 821, 453460. [49] Biserte, G., Havez, R. and Cuvelier, R. (1963) ExposCs Annuels de Biochimie MCdicale, Vol. 24. Masson, Paris, pp. 85-1 20. [SO] Masson, P.L., Heremans, J.F. and Schonne, E. (1969) J. Exp. Med. 130, 643458. [51] Baggiolini, M., De Duve, C., Masson, P.L. and Heremans, J.F. (1970) J. Exp. Med. 131, 559-570. [52] Brown, J.P., Rose, T.M. and Plowman, G.D (1985) In: G. Spik, J. Montreuil, R.R. Crichton and J. Mazurier (Eds.), Proteins of Iron Storage and Transport. Elsevier, Amsterdam, pp. 3 9 4 6 . [53] Brown, J.P., Wright, P.W., Hart, C.E., Woodbury, R.G., Hellstrom, K.E. and Hellstrom, I. (1980) J. Biol. Chem. 255, 49804983. [54] Woodbury, R.G., Brown, J.P., Yeh, M.Y., Hellstrom, I . and Hellstrom, K.E. (1980) Proc. Natl. Acad. Sci. USA 77, 2183-2187. [55] Dippold, W.G., Lloyd, K.O., Li, L.T., Ikeda, H., Oettgen, H.F. and Old, L.J. (1980) Proc. Natl. Acad. Sci. USA 77, 61 14-61 18. [56] Plowman, G.D., Brown, J.P., Enns, C.A., Schroder, J., Nikinmaa, B., Sussman, H.H., Hellstrom, K.E. and Hellstrom, I. (1983) Nature 303, 70-72. [57] Yang, F., Lum, J.B., McGill, J.R., Moore, C.M., Naylor, S.L., Van Bragt, P.H., Baldwin, W.D. and Bowman, B.H. (1984) Proc. Natl. Acad. Sci. USA 81, 2752-2756. [58] Rose, T.M., Plowman, G.D., Teplow, D.B., Dreyer, W.J., Hellstrom, K.E. and Brown, J.P. (1986) Proc. Natl. Acad. Sci. USA 83, 1261-1265. [59] Alemany, R., Rosa Vila, M., Franci, C., Egea, G., Real, EX. and Thompson, J.M. (1993) J. Cell Sci. 104, 1155-1 162. [60] Food, M.R., Rothenberger, S., Gabathuler, R., Haidl, I.D., Reid, G. and Jefferies, W.A. (1994) J. Biol. Chem. 269, 3034-3040.

236 [61] Baker, E.N., Baker, H.M., Smith, C.A., Stebbins, M.R., Kahn, M., Hellstrom, K.E. and Hellstrom, 1. (1992) FEBS Lett. 298, 215-218. [62] McGillivray, R.T.A., Mendez, E., Sinha, S.K., Sutton, M.R., Lineback-Zins, J. and Brew, K. (1982) Proc. Natl. Acad. Sci. USA 79, 2504-2508. [63] McGillivray, R.T.A., Mendez, E., Shewall, J.G., Sinha, S.K., Lineback-Zins, J. and Brew, K.(1983) J. Biol. Chem. 258, 3543-3553. [64] Uzan, G., Frain, M., Park, I., Besmond, C., Maessen, G., Salatrepat, J., Zakin, M.M. and Kahn, A. (1984) Biochem. Biophys. Res. Commun. 119, 273-281 [65] Baldwin, G.S. and Weinstock, J. (1988) Nucleic Acids Res. 16, 8720. [66] Carpenter, M.A. and Broad, T.E. (1993) Biochim. Biophys. Acta 1173, 230-232. [67] Escriva, H., Pierce, A,, Coddeville, B., Gonzales, F., Benaissa, M., LCger, D., Wieruszeski, J.M., Spik, G. and Pamblanco, M. (1995) Biochim. J. 307, 47-55. [68] Banfield, D.K., Chow, B.K.C., Funk, W.D., Robertson, K.A., Umelas, T.M., Woodworth, R.C. and McGillivray, R.T.A. (1991) Biochem. Biophys. Acta 1089, 262-265. [69] Jeltsch, J.M. and Chambon, F! (1982) Eur. J. Biochem. 122, 291-295. [70] Moskaitis, J.E., Pastori, R.L. and Schoenberg, O.R. (1990) Nucleic Acids Res. 18, 6135. [71] Jamroz, R.C., Gasdaska, J.R., Bradfield, J.Y. and Law, J.H. (1993) Proc. Natl. Acad. Sci. USA 90, 1320-1324. [72] Kurama, T., Kurata, S. and Natori, S. (1995) Eur. J. Biochem. 228, 229-235. [73] Bartfeld, N.S. and Law, J.H.( 1990) J. Biol. Chem. 265, 21684-21691. [74] Metz-Boutigue, M.H., JollBs, J., Mazurier, J., Schoentgen, F., Legrand, D., Spik, G., Montreuil, J. and JollCs, F! (1984) Eur. J. Biochem. 145, 659476. [75] Powell, M.J. and Ogden, J.E. (1990) Nucleic Acids Res. 18, 4013. [76] Rey, M.W., Woloshuk, S.L., de Boer, H.A. and Pieper, F.R. (1990) Nucleic Acids Res. 18, 5288. [77] Mead, F!E. and Tweedie, J.W. (1990) Nucleic Acids Res. 18, 7167. [78] Pierce, A,, Colavizza, D., Benaissa, M., Maes, F,! Tartar, A,, Montreuil, J. and Spik, G. (1991) Eur. J. Biochem. 196, 177-1 84. [79] Goodman, R.E. and Schanbacher, EL. (1991) Biochem. Biophys. Res. Commun 180, 75-84, [80] Lydon, J.P., O’Malley, B.R., Saucedo, O., Lee, T., Headon, D.R. and Conneely, O.M. (1992) Biochim. Biophys. Acta 1132, 97-99. [81] Le Provost, F., Nocart, M., GuCrin, G. and Martin, P. (1994) Biochem. Biophys. Res. Commun. 203, 1324-1332. [82] Pentecost, B.T. and Teng, C.T. (1987) J. Biol. Chem. 262, 10134-10139. [83] Welch, S. (1990) Comp. Biochem. Physiol. 97B, 417427. [84] Williams, J. (1982) Trends Biochem. Sci. 8, 272-275. [85] Cochet, M., Gannon, F., Hen, R., Maroteaux, L., Perrin, F. and Chambon, F! (1979) Nature 282, 567574. [86] Schaeffer, E., Lucers, M.A., Jeltsch, J.M., Py, M.C., Levin, M.J., Chambon, P., Cohen, G.N. and Zakin, M.M. (1987) Gene 56, 109-1 16. [87] Martin, A.W., Huebers, H., Huebers, J., Webb, J. and Finch, C.A. (1984) Blood 64, 1048-1052. [88] Bowman, B.H., Yang, F. and Adrian, G.S. (1988) Adv. Genet. 25, 1-38. [89] Anderson, B.F., Baker, H.M., Norris, G.E., Rice, D.W. and Baker, E.N. (1989) J. Mol. Biol. 209, 711734. [90] Louie, G. (1993) Curr. Opin. Struct. Biol. 3, 401408. [91] Baker, E.N. and Lindley, P.F. (1992) J. Inorg. Biochem. 47, 147-160. [92] Anderson, B.F., Baker, H.M., Dodson, E.J., Norris, G.E., Rumball, S.V, Waters, J.M., Baker, E.N. (1987) Proc. Natl. Acad. Sci. USA 84, 1769-1773. [93] Anderson, B., Baker, H., Norris, G., Rumball, S. and Baker, T. (1990) Nature 344, 784-785. [94] Baker, E.N., Anderson, B.F., Baker, H.M., Haridas, M., Jameson, G.B., Norris, G.E., Rumbal, S.V and Smith, C.A. (1991) Int. J. Biol. Macromol. 13, 122-130. [95] Baker, E.N., Anderson, B.F., Baker, H.M., Day, C.L., Haridas, M., Norris, G.E., Rumball, V, Smith, C.A. and Thomas, D.H. (1994) Adv. Exp. Med. Biol. 357, 1-12.

237 [96] Haridas, M., Anderson, B.F., Baker, H.M., Norris, G.E. and Baker, E.N. (1994) Adv. Exp. Med. Biol. 357, 235-238. [97] Haridas, M., Anderson, B.F. and Baker, E.N. (1995) Acta Cryst. D51, 629-646. [98] Bailey, S., Evans, R.E., Garrat, R.C., Gorinsky, B., Hasnain, S., Thuti, H., Lindley, PF., Mydin, A,, Sarra, R. and Watson, J.L. (1988) Biochemistry 27, 58045812. [99] Kurokawa, H., Mikami, B. and Hirose, M. (1996) J. Mol. Biol. 254, 196-207. [loo] Legrand, D., Mazurier, J., Colavizza, D., Montreuil, J. and Spik, G. (1990) Biochem. J. 266, 575-581. [I011 Ikeda, H., Nabuchi, Y., Nakazato, K., Tanaka, Y. and Satake, K. (1985) FEBS Lett. 182, 305-309. [lo21 Williams, J. (1968) Biochem. J. 108, 57-61. [I031 Lkger, D., Tordera, V, Spik, G., Dorland, L., Haverkamp, J. and Vliegenthart, J.F.G. (1978) FEBS Lett. 93, 255-260. [I041 Evans, R.W., Aitken, A. and Patel, K.J. (1988) FEBS Lett. 238, 3 9 4 2 . [I051 Sharma, N.D., Evans, R.W., Patel, K.J., Gorinsky, B., Mallet, A.I. and Aitken, A. (1994) Biochim. Biophys. Acta 1206, 286-288. [I061 Hatton, M.W.C., Marz, L.., Berry, L.R., Debanne, M.T. and Regoeczi, E. (1979) Biochem. J. 181, 633638. [lo71 Marz, L., Hatton, M.W.C., Berry, L.R. and Regoeczi, E. (1982) Can. J. Biochem. 60, 624430. [lo81 Lkger, D., Campion, B., Decottignies, J.P., Montreuil, J. and Spik, G. (1989) Biochem. J. 257, 231-238. [I091 Fu, D. and Van Halbeek, H. (1992) Anal. Biochem. 206, 53-53. [I101 De Jong, G. and Van Eijk, H.G. (1988) Electrophoresis 9, 589-598. [I 1 I ] Rohrer, J.S. and Avdalovic, N. (1966) Protein Exp. Purif. 7, 3 9 4 4 . [ I 121 Coddeville, B., Carchon, H., Jaeken, J., Briand, G. and Spik, G. (1997) to be published. [ I 131 Stratil, A,, Glasnak, V,Tomasek, K., Williams, J. and Clamp, J.R. (1984) Anim. Blood Groups Biochem. Genet. 15, 285-297. [I 141 Coddeville, B., Stratil, A,, Wieruszeski, J.M., Strecker, G., Montreuil, J. and Spik, G. (1989) Eur. J. Biochem. 186, 583-590. [I151 Carpenter, M.A. and Broad, T.E. (1993) Genome 37, 157-165. [116] Stratil, A,, Tomasek, V, Clamp, J.R. and Williams, J. (1985) Comp. Biochem. Physiol. 80B, 909-91 1. [117] Van Berkel, P.H.C., Van Veen, H.A., Geerts, M.E.J., De Boer, H.A. and Nuijens, J. (1996) Biochem. J. 319, 117-122. [I 181 Colavizza, D., Decottignies, J.P. and Spik, G., unpublished results. 11191 Spik, G., Coddeville, B. and Montreuil, J. (1988) Biochimie 70, 1459-1469. [120] Coddeville, B., Newport, M.J. and Spik, G., unpublished results. [I211 Decottignies, J.P., Coddeville, B. and Spik, G., unpublished results. [122] Trowbridge, I.S. (1988) Prog. Allergy 45, 121-146. [I231 Rossen, R., Schade, A.L., Butler, W.T. and Kasel, J.A. (1966) J. Clin. Invest. 45, 768. [I241 Hamilton, T.A., Wada, H.G. and Sussman, H.H. (1979) J. Supramol. Struc. 11, 503-515. [I251 Testa, V (1985) Curr. Top. Hematol. 5, 127-161. [I261 Brock, J.H. (1985) in: P.M. Harrison (ed.) Metalloproteins, Part 2. Verlag-Chemie, Weinheim, pp. 183262. [I271 Crichton, R.R. (1991) In: R.B. Crichton (Ed.) Inorganic Biochemistry of Iron Metabolism. Ellis Honvood, Chichester, pp. 120-130. [I281 Qian, Z.M. and Tang, P.L. (1995) Biochim. Biophys. Acta 1269, 205-214. [I291 De Sousa, M. and Brock, J.H. (1989) Iron in Immunity, Cancer and Inflammation. Wiley, Chichester. [I301 Trowbridge, I.S. and Shakelford, D.E. (1986) Biochem. SOC.Symp. 51, 117-129. [131] Mc Clelland, A,, Kuhn, L.C. and Ruddle, F.H. (1984) Cell 39, 267-274. [132] Schneider, C., Owen, M.J., Banville, D. and Williams, J.G. (1984) Nature 311, 675-678. [133] Sabatini, D.D., Kreibich, G., Morinoto, T. and Adensick, M. (1982) J. Cell. Biol. 92, 1-22. [I341 Omary, M.B. and Trowbridge, I S . (1981) J. Biol. Chem. 256, 12888-12892. [I351 Schneider, C., Sutherland, R., Newman, R. and Greaves, M. (1982) J. Biol. Chem. 257, 8516-8522. [I361 Do, S.I., Enns, C.A. and Cummings, R.D. (1990) J. Biol. Chem. 265, 114-125. [I371 Do, S.I. and Cummings, R.D.(I992) Glycobiology 2, 345-353. [I381 Hayes, G.R., Enns, C.A. and Lucas, J.J. (1992) Glycobiology 2, 355-359.

238 [I391 Roberts, K.P. and Griswold, M.D. (1990) Mol. Endocrinol. 4, 531-542. [I401 Collawn, J.F., Lai, A., Domingo, D., Fitch, M., Hatton, S. and Trowbridge, I S . (1993) J. Biol. Chem. 268, 21 686-21 692. [I411 Gerhardt, E.M., Chan, L.L., Jing, S., Meiying, Q. and Trowbridge, IS.(1991) Gene 102, 249-254. [I421 Newman, R., Schneider, C., Sutherland, R., Vodinelich, L. and Greaves, M.G. (1982) Trends Biochem. Sci. 7, 397400. [I431 Hayes, G.R., Williams, A., Costello, C.E., Enns, C.A. and Lucas, J.J. (1995) Glycobiology 5, 227-232. [144] Orberger, G., Geyer, R., Stirm, S. and Tauber, R. (1992) Eur. J. Biochem. 205, 257-267. [I451 Reckhow, C.L. and Enns, C.A. (1988) J. Biol. Chem. 263, 7297-7301. [I461 Hunt, R.C., Riegler, R. and Davis, A.A. (1989) J. Biol. Chem. 264, 9643-9648. [I471 Williams, A.M. and Enns, C.A. (1991) J. Biol. Chem. 266, 17648-17654. [I481 Hoe, M.H. and Hunt, R.C. (1992) J. Biol. Chem. 267, 49164929. [149] Yang, B., Hoe, M.H., Black, P., Hunt, R.C.(1993) J. Biol. Chem., 268, 7435-7441. [I501 Williams, A.M. and Enns, C.A. (1993) J. Biol. Chem., 268, 12780-12786. [I511 Dautry-Varsat, A,, Ciechansver, A. and Lodish, H.F. (1983) Proc. Natl. Acad. Sci. USA 80,2258-2262. [I521 Klausner, R.D., Ashwell, G., Van Renswoude, J., Harford, J.B. and Bridges, K.R. (1983) Proc. Natl. Acad. Sci. USA 80, 2263-2266. [ 1531 Masson, P. (1 970) La Lactoferrine, ProtCine des Secretions Externes et des Leucocytes Neutrophiles. Arscia, Bruxelles. [I541 Machnicki, M. (1991) Biological Properties of Lactoferrin, Folia Biologica 37, 65-76. [I551 Sanchez, L., Calvo, M. and Brock, J.H. (1992) Arch. Dis. Child. 67, 657-661. [I561 Baker, N.E. and Lindley, P.F. (1992) J. Inorg. Biochem. 47, 147-160. [I571 Baker, N.E. (1993) Perspectives Bioinorg. Chem. 2, 161-205. [I581 Renner, B. and Sawatzki, G. (1993) New Perspectives in Infant Nutrition. Georg Thieme Verlag, Stuttgart. [I591 Hutchens, T.W., Rumball, S.V. and Lonnerdal, B. (1994) Lactoferrin: Structure and Function. Plenum Press, New York. [I601 Lonnerdal, B. and Iyer, S. (1995) Annu. Rev. Nutr. 15, 93-1 10. [I611 Levay, P.F. and Viljoen, M. (1995) Haematologica 80, 252-267. [I621 Nuijens, J.H., Van Berkel, P.H.C. and Schanbacher, EL. (1996) J. Mammary Gland Biol. Neoplasia I , 283-293. [ 1631 Hutchens, T.W. and Lonnerdal, B. (1997) Lactoferrin: Interactions and Biological Functions. Humana Press, Totowa, NJ. [I641 Cox, T.M., Mazurier, J., Spik, G., Montreuil, J. and Peters, T.J. (1979) Biochim. Biophys. Acta 588, 120-128. [I651 Hu, W.L., Mazurier, J., Sawatzki, G., Montreuil, J. and Spik, G. (1988) Biochem. J. 249, 435441. [I661 Hu, W.L., Mazurier, J., Montreuil, J. and Spik, G. (1990) Biochemistry 29, 535-541. [I671 Davidson, L.A. and Lonnerdal, B. (1988) Am. J. Physiol. 254, G580-GS85. [I681 Kawakami, H. and Lonnerdal, B. (1991) Am. J. Physiol. 261, 841-846. [I691 Masson, P.L. and Heremans, J.F. (1966) Prot. Biol. Fluids 14, 115-124. [I701 Oram, J.D. and Reiter, B. (1968) Biochim. Biophys. Acta 170, 351-365. [I711 Bullen, J.J., Rogers, H.J. and Leigh, L. (1972) Br. Med. J. I , 69-75. [I721 Spik, G., ChCron, A,, Montreuil, J. and Dolby, J.M. (1978) Immunology 35, 663-671. [I731 Spik, G., Jorieux, S., Mazurier, J., Navarro, J., Romond, C. and Montreuil, J. (1984) In: A.F. Williams and J.D. Baum (Eds.), Human Milk Banking, Nestle Nutrition Workshop Series, Vol. 5. Raven Press, New York, pp. 133-143 [174] Jorieux, S., Mazurier, J., Montreuil, J. and Spik, G. (1985) Prot. Biol. Fluids 32, 1 15-1 18. [I751 Perraudin, J.P. and Prieels, J.P. (1982) Biochim. Biophys. Acta 718, 4 2 4 8 . [I761 Arnold, R.R., Cole, M.F. and Mc Ghee, J.R. (1977) Science 197, 263-265. [I771 Tomita, M., Takase, M., Wakabayashi, H. and Bellamy, W. (1994) In: T.W. Hutchens, S.V. Rumball and B. Lonnerdal (Eds.), Lactoferrin, Structure and Function. Plenum Press, New York, pp. 209-21 8. [I781 Saito, H., Takase, M., Tamura, Y., Shimamura, S. and Tomita, M. (1994) In: T.W. Hutchens, S.V. Rumball and B. Lonnerdal (Eds.), Lactoferrin, Structure and Function. Plenum Press, New York, pp. 219-230.

239 [I791 Baynes, R.D. and Bezwoda, W.R. (1994)Adv. Exp. Med. Biol. 357, 133-142. [I801 Britigan, B.E., Serody, J.S.and Cohen, M.S. (1994)Adv. Exp. Med. Biol. 357, 143-156. [I811 Mantel, C., Miyazawa, K. and Broxmeyer, H.E. (1994)Adv. Exp. Med. Biol. 357, 121-132. [I821 Machnicki, M., Zimecki, M. and Zagulski, T. (1993)Int. J. Exp. Pathol. 74,433439. [183] Shau, H., Kim, A. and Golub, S.H. (1992)J. Leukocyte Biol. 51, 343-349. [I841 Bezault, J., Bhimani, R., Wiprovnick, J. and Furmanski, P. (1994)Cancer Res. 54,2310-2312. [I851 Brock, J.H. and Mainou-Fowler, T. (1983)Immunol. Today 4, 347-351. [I861 Zimecki, M., Mazurier, J., Machnicki, M., Wieczorek, Z., Montreuil, J. and Spik, G. (1991) Immunol. Lett. 30, 119-124. [I871 Zimecki, M., Mazurier, J., Spik, G. and Kapp, J.A. (1995)Immunology 86, 122-127. [I881 Hashizume, S.,Kuroda, K. and Murakami, H. (1983)Biochim. Biophys. Acta 763, 377-382. [189] Mazurier, J., Legrand, D., Hu, W.L., Montreuil, J. and Spik, G. (1989)Eur. J. Biochem. 179,481487. [I901 Mazurier, J., Legrand, D., Leveugle, B., Rochard, E., Montreuil, J. and Spik, G. (1994)In: T.W. Hutchens, S.V. Rumball and B. Lonnerdal (Eds.), Lactoferrin, Structure and Function. Plenum Press, New York, pp.lI1-119. [191] Van Snick, J.L. and Masson P.L. (1976)J. Exp. Med. 144, 1568-1580. [I921 Spik, G., Legrand, D., Leveugle, B., Mazurier, J., Mikogami, T., Montreuil, J., Pierce, A. and Rochard, E. (1993) In: B. Renner and G. Sawatzki (Eds.), New Perspectives in Infant Nutrition. Thieme Verlag, Stuttgart, pp. 77-83. [I931 Spik, G., Legrand, D., Leveugle, B., Mazurier, J., Mikogami, T., Montreuil, J., Pierce, A. and Rochard, E. (1994)In: T.W. Hutchens, S.V. Rumball and B. Lonnerdal (Eds.), Lactoferrin, Structure and Function. Plenum Press, New York, pp. 13-19. [I941 Mikogami, T., Heyman, M., Spik, G. and Desjeux, J.F. (1994)Am. J. Physiol. 267, G308-G315. [I951 Birgens, H.S. (1994)In: T.W. Hutchens, S.V. Rumball and B. Lonnerdal (Eds.), Lactoferrin, Structure and Function. Plenum Press, New York, pp. 99-1 09. [I961 Goavec, M., Mazurier, J., Montreuil, J. and Spik, G.(1985)C.R. Acad. Sci. Paris 301,689-694. [197]Birgens, H.S. (1984)Scand. J. Haematol. 33, 275-280. [198]Rochard, E., Legrand, D., Lecocq, M., Hamelin, R., Montreuil, J. and Spik, G. (1992)Anticancer Res.

12, 2047-2052. [I991 Leveugle, M., Mazurier, J., Legrand, D., Mazurier, C., Montreuil, J. and Spik, G. (1993)Eur. J. Biochem. 213, 1205-1211. [200] Maneva, A,, Taleva, B., Manev, V. and Sitakov, L. (1993)Int. J. Biochem. 25, 707-712. [201] Nillesse, N., Pierce, A,, Lecocq, M., Benaissa, A. and Spik, G. (1995)Biol. Cell 82, 149-159. [202] Mc Abee, D. and Esbensen, K. (1991)J. Biol. Chem. 266,23624-23631. [203] Ziere, G.J., Van Dijk, M.C.M., Bijsterboch, M.K. and Van Berkel, T.J.C. (1992)J. Biol. Chem. 267, 11229-1 1235.. [204]Naidu,S.S., Erdei, J., Czirok, E., Kalfas, S., Gado, I . , Thoren, A,, Forsgren, A. and Naidu, A.S. (1991) APMIS 99, 1142-1 150.. [205] Gray-Owen, S.C. and Schryvers, A.B. (1996)Trends Microbiol. 4, 185.. [206] Bi, B.Y., Leveugle, B., Liu, J.L., Collard, A,, Coppe, P, Roche, A.C., Nillesse, N., Capron, M., Spik, G. and Mazurier, J.(1994) Eur. J. Cell. Biol. 65, 164-171. [207] Kennard, M.L., Richardson, D.R., Gabathuler, R., Ponka, P. and Jefferies, W.A. (1995)EMBO J. 14, 417841 86. [208] Conrad, M.E. and Umbreit, J.M. (1993)Am. J. Hematol. 42,67-73. [209] Danielsen, E.M. and Van Deurs, B. (1995)J. Cell Biol. 131, 939-950. [210] Spik, G., Coddeville, B., Mazurier, J., Bourne, Y., Cambillau, C. and Montreuil, J. (1994), in: T.W. Hutchens, B. Lonnerdal and S. Rumball (Eds.), Lactoferrin Structure and Function. Plenum Press, New York, pp. 21-32. [21I] Spik, G., Vandersyppe, B., Fournet, B., Bayard, B., Charet, P., Bouquelet, S., Strecker, G. and Montreuil, J. (1974)In: J. Montreuil (Ed.), Proc. 2nd Int. Symp. on Glycoconjugates, Villeneuve d’Ascq, 1973.Editions du CNRS, Paris, Vol. I , pp. 483499. [212] Spik, G., Bayard, B., Fournet, B., Strecker, G., Bouquelet, S. and Montreuil, J. (1975)FEBS Lett. 50, 296-299.

240 [213] Dorland, L., Haverkamp, J., Schut, B.L., Vliegenthart, J.F.G., Spik, G., Strecker, G., Fournet, B. and Montreuil, J. (1977) FEBS Lett. 77, 15-20. [214] Coddeville, B., Brock, J. and Spik, G., unpublished results. [215] Coddeville, B., Gukrin, G. and Spik, G., unpublished results. [216] Coddeville, B., Morgan, E.H. and Spik, G., unpublished results. [217] Spik, G., Debruyne, V., Montreuil, J., Van Halbeek, H. and Vliegenthart, J.F.G. (1985) FEBS Lett. 183, 65-69. [218] Leclercq, Y., Sawatzki, G., Wieruszeski, J.M., Montreuil, J. and Spik, G. (1987) Biochem. J. 247, 571578. [219] Coddeville, B., Regoeczi, E., Strecker, G., Plancke, Y. and Spik, G., Biochem. J. (in press). I2201 Coddeville, B., Stratil, A. and Spik, G., unpublished results. [221] Spik, G., Coddeville, B., Strecker, G., Montreuil, J., Regoeczi, E., Chindemi, PA. and Rudolph, J.R. (1991) Eur. J. Biochem. 195, 397-405. [222] Stratil, A. and Glasnik, V. (1981) Anim. Blood Groups Biochem. Genet. 12, 113-122. [223] Stratil, A,, TomBsek, V., Bobak, P. and Glasnak, V. (1984) Anim. Blood Groups Biochem. Genet. 15, 89-1 01. [224] Coddeville, B., Stratil, A,, Strecker, G., Plancke, Y. and Spik, G., unpublished results. [225] Stratil, A,, TomLek, V., Clamp, J.R. and Williams, J. (1985) Comp. Biochem. Physiol. 80B, 909-91 I . [226] Derisbourg, P., Wieruszeski, J.M., Montreuil, J. and Spik, G. (1990) Biochem. J. 269, 821-825. [227] Montreuil, J. (1974) Pure Appl. Chem. 42, 431-477. [228] Spik, G., Coddeville, B., Legrand, D., Mazurier, J., Leger, D., Goavec, M. and Montreuil, J. (1985) In: G. Spik, J. Montreuil, R.R. Crichton and J. Mazurier (Eds.), Proteins of Iron Storage and Transport. Elsevier, Amsterdam, pp. 47-51, [229] Spik, G. and Mazurier, J. (1977) in: E.B. Brown, P. Aisen, J. Fielding and R.R. Crichton (Eds.), Proteins of Iron Metabolism. Grune and Stratton, New York, pp.143-151. [230] Spik, G., Strecker, G., Fournet, B., Bouquelet, S., Montreuil, J., Dorland, L., Van Halbeek, H. and Vliegenthart, J.F.G. (1982) Eur. J. Biochem. 121, 413419. [23 I ] Coddeville, B., Lonnerdal, B. and Spik, G., unpublished results. [232] Coddeville, B., Zagulski, T. and Spik, G., unpublished results. [233] Decottignies, J.P., Coddeville, B., Sawatzki, G. and Spik, G., unpublished results. [234] Legrand, D., Salmon, V., Coddeville, B., Benaissa, M., Plancke, Y. and Spik, G. (1995) FEBS Lett. 365, 57-60. [235] Matsumoto, A,, Yoshima, H., Takasaki, S. and Kobata, A. (1982) J. Biochem. (Japan) 91, 143-152. [236] Coddeville, B., Strecker, G., Wieruszeski, J.M., Vliegenthart, J.F.G., Van Halbeek, H., Peter-Katalinic, J., Egge, H. and Spik, G. (1992) Carbohydr. Res. 236, 145-164. [237] Van Halbeek, H., Dorland, L., Vliegenthart, J.F.G., Spik, G., Cheron, A. and Montreuil, J. (1981) Biochim. Biophys. Acta 675, 293-296. [238] Coddeville, B. and Spik, G., unpublished results. [239] Spik, G., Fournet, B., Montreuil, J. (1979) C.R. Acad. Sci. Paris 288, 967-970. [240] Dorland, L., Haverkamp, J., Vliegenthart, J.F.G., Spik, G., Fournet, B. and Montreuil, J. (1979) Eur J. Biochem. 100, 569-574. [241] Jacquinot, P.M., LCger, D., Wieruszeski, J.M., Coddeville, B., Montreuil, J. and Spik, G. (1994) Glycobiology 4, 617-624. [242] De Jong, G., Feelders, R., Van Noort, W.L. and Van Eijk, H.G. (1995) Glycoconjugate J. 12,219-226. [243] Van Eijk, H.G., Van Noort, W.L., De Jong, G. and Koster, J.F. (1987) Clin. Chim. Acta 165, 141-145. [244] De Jong, G., Van Noort, W.L., Feelders, R.A., Jeu-Jaspars, C.M.H. and Van Eijk, H.G. (1992) Clin. Chim. Acta 212, 2 7 4 5 . [245] Keir, G., Zeman, A., Brookes, G., Porter, M. and Thompson, E.J. (1992) Ann. Clin. Biochem. 29, 2 10-2 13. [246] Hoffmann, A,, Nimtz, M., Getzlaff, R. and Conradt, H.S. (1995) FEBS Lett. 359, 164-168. [247] D’Andrea, G., D’Alessandro, A.M., Salucci, M.L. and Oratore, A. (1994) J. Prot. Chem. 13, 31-36. [248] Stowell, K.M., Rado, T.A., Funk, W.D. and Tweedie, J.W. (1991) Biochem. J. 276, 349-355

24 1 [249] Ward, P.P., Lo, J.Y., Duke, M., May, G.S., Headon, D.R. and Conneely, O.M. (1992) Biotechnology 10, 784-789. [250] Liang, Q. and Richardson, T. (1993) J. Agr. Food Chem. 41, 180&-1806. [251] Mitra, A. and Zhang, Z. (1994) Plant Physiol. 106, 977-981. [252] Platenburg, G.J., Kootwijk, E.P.A., Kooiman, RM., Woloshuk, S.L., Nuijens, J.H., Krimpenfort, P.J.A., Pieper, F.R., De Boer, H.A. and Strijker, R. (1994) Transgenic Res. 3, 99-108. [253] Montreuil, J. (1995) In: J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.) Glycoproteins, New Comprehensive Biochemistry, Vol. 29a. Elsevier, Amsterdam, pp. 1-1 2. [254] Montreuil, J. (1993) In: C. Rivat and J.F. Stoltz (Eds.). Colloque INSERM no. 227. John Libley Eurotext, Montrouge, France, pp. 283-292. [255] Spik, G., Debruyne, V. and Montreuil, J. (1983), In: H. Popper, W. Reutter, E. Kottgen and F. Gudat (Eds.), Structural Carbohydrates in Liver. MTP Press, Boston, pp. 477483. [256] Debruyne, V., Montreuil, J. and Spik, G. (1984) Prot. Biol. Fluids 31, 6 3 4 8 . [257] Stibler, H. (1991) Clin. Chem. 37, 2029-2037. [258] Stibler, H. (1993) Acta Neurol. Scand. 88, 279-283. [259] Xin, Y., Lasker, J.M. and Lieber, C.S. (1993) Hepatology 22, 1462-1468. [260] Yamashita, K., Koide, K., Endo, T., Iwaki, Y. and Kobata, A. (1989) J. Biol. Chem. 264, 2415-2423. [261] Campion, B., Ltger, D., Wieruszeski, J.M., Montreuil, J. and Spik, G. (1989) Eur. J. Biochem. 184, 405413. [262] Hahn, T.J. and Goochee, C.F. (1992) J. Biol. Chem. 267, 23982-23987. [263] Jaeken, J., Van Eijk, H.G., Van der Heul, C., Corbeal, L., Eeckels, R. and Eggermont, E. (1984) Clin. Chim. Acta 144, 245-247. [264] Yamashita, K. and Ohno, K. (1996) In: J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins and Disease, New Comprehensive Biochemistry, Vol. 30. Elsevier, Amsterdam, ch. 1 6a. [265] Jaeken, J., Spik, G. and Schachter, H. (1996) In: J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins and Disease, New Comprehensive Biochemistry, Vol. 30. Elsevier, Amsterdam, ch. 16b. [266] Ohno, K., Yuasa, I., Akaboshi, S., Itoh, M., Yoshida, K., Ehara, H., Ochiai, Y. and Tokeshita, K. (1992) Brain Dev. 14, 30-35. [267] Yamashita, K., Ideo, H., Okhura, T., Fukushima, K., Yuasa, I., Ohno, K. and Takeshita, K. (1993) J. Biol. Chem. 268, 5783-5789. [268] Jaeken, J., Schachter, H., Carchon, H., De Cock, P., Coddeville, B. and Spik, G. (1994) Arch. Dis. Child. 71, 123-127. [269] Fukuda, M.N. (1996) In: J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins and Disease, New Comprehensive Biochemistry, Vol. 30. Elsevier, Amsterdam, ch. 8. [270] Fukuda, M.N., Gaetani, G.F., Izzo, P., Scartezzini, P. and Dell, A. (1992) Br. J. Haematol. 82, 745-752. [271] Montreuil, J., Foumet, B., Spik, G. and Strecker, G.(1978) C.R. Acad. Sci. Paris 287, 837-840. [272] Montreuil, J. (1980) Adv. Carbohydr. Chem. Biochem. 37, 157-223. [273] Montreuil, J. (1984) Biol. Cell 51, 115-132. [274] Montreuil, J. (1984) Pure Appl. Chem. 56, 859-877. 12751 Mazurier, J., Dauchez, M., Vergoten, G., Montreuil, J. and Spik, G. (1991) C.R. Acad. Sci. Paris 313, 7-14. [276] Mazurier, J., Dauchez, M., Vergoten, G., Montreuil, J. and Spik, G. (1991) Glycoconjugate J. 8, 390399. [277] Dauchez, M., Mazurier, J., Montreuil, J., Spik, G. and Vergoten, G. (1992) Biochimie 74, 63-74. [278] Cambillau, C. (1995) In: J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins, New Comprehensive Biochemistry, Vol. 29a. Elsevier, Amsterdam, 29a, pp. 29-65. [279] RougC, P.,Bourne, Y. and Cambillau, P. (1992) J. Biol. Chem. 267, 197-203. [280] Bourne, Y., Nesa, M.P., Rouge, P., Mazurier, J., Legrand, D., Spik, G., Montreuil, J. and Cambillau, C. (1992) J. Mol. Biol. 227, 938-941. [281] Bourne, Y., Mazurier, J., Legrand, D., Rougt, P., Montreuil, J., Spik, G. and Cambillau, C. (1994) Curr. Biol. 2, 209-219.

242 12821 Bourne, Y., Mazurier, J., Legrand, D., Rouge, P., Montreuil, J., Spik, G. and Cambillau, C. (1994) Structure 2, 209-219. [283] Bourne, Y., Van Tilbeurgh, H. and Cambillau, C. (1993) Curr. Opin. Struct. Biol. 3, 681-686. [284] Bailey, S., Evans, R.E., Garratt, R.C., Gorinsky, B., Hasnain, S., Horsburgh, C., Thuti, H., Lindley, P.F., Mydin, A,, Sarra, R. and Watson, J.L. (1988) Biochemistry 27, 5804-5812.

J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins II

0 1997 Elsevier Science B.Y All rights reserved CHAPTER 11

Chemistry, biochemistry and biology of sialic acids Roland Schauer' and Johannis P. Kamerling2 'Biochemisches Institut, Christian-AIbrechts-LIniuersilatzu Kiel,Germany, 'Bijuoet Center, Department of Bio-Organic Chemistry, Utrecht Vniversiy, The Netherlands

List of Abbreviations Abbreviations used for sialic acids (Sia) are included in Tables 1 and 13. Alt altrose

GlcA

D-glucuronic acid

GlcNAc

N-acetyl-o-glucosamine

GlcN

o-glucosamine

Ara

L-arabinose

GD3

disialoganglioside

Asn

L-asparagine

GM 1

monosialoganglioside

ASP CDP

L-aspartate cytidine diphosphate

Cer CI

ceramide chemical ionization

CMP

cytidine monophosphate

CoA

coenzyme A

CTP

cytidine triphosphate

Gul

gulose

HeP HPLC

heptose

HPTLC

high-performance thin-layer chromatography

high-performance liquid chromatography

IgG

immunoglobulin G

C7-NeuSAc NeuSAc missing the C8-C9 part C8-NeuSAc NeuSAc missing the C9 part

IgM Leu

immunoglobulin M

Da DFP

Dalton diisopropylfluorophosphate

Man

o-mannose

ManNAc

N-acetyl-D-mannosamine

DIG DMB

digoxigenin 1,2-diamino-4,5methylenedioxybenzene

MS

mass spectrometry

MU

4-methylumbelliferyl

NMK

nucleoside monophosphate kinase

DNA

desoxyribonucleic acid

NMR

nuclear magnetic resonance

EI

electron impact

NOESY

EPR

electron paramagnetic resonance

nuclear Overhauser enhancement spectroscopy

FAB Fuc

fast atom bombardment L-fucose

P, Pi

phosphate pulsed amperometric detection pyruvate kinase

L-leucine

FucN

L-fucosamine

PAD PEP

Gal

D-galactose

PK

GalNAc-ol

N-acetylgalactosaminitol

PPase

inorganic phosphatase

GalANGro

N-galacturonyl-2-aminoglycerol

PPi

pyrophosphate

GalNAc

N-acetyl-D-galactosamine

PYR

pyruvate

GDP Glc GLC

guanosine diphosphate D-glucose gas-liquid chromatography

ROESY

rotating-frame nuclear Overhauser enhancement spectroscopy

QuiNAc

N-acetyl-o-quinovose

243

phosp hoenolpyruvate

244 sulfate SDS/PAGE sodium dodecyl sulfate/polyacrylamide gel electrophoresis Ser L-serine TADH Thermoanaerobium brockii alcohol dehydrogenase trifluoroacetic acid TFA thin-layer chromatography TLC S

Thr TOCSY

L-threonine total correlation spectroscopy

UDP

uv

uridine diphosphate ultraviolet

xYl

D-XylOSe

2D

two-dimensional

I . Introduction Since the discovery of N-acetylneuraminic acid, the most universal sialic acid, at the end of the 1930s [1,2] as well as the structural and stereochemical elucidation of its free and bound forms at the end of the 1960s (reviewed in ref. [3]), there has been a continual increase in the number of sialic acid types (1994: more than 40) recognized to occur in a variety of living organisms. It is now generally accepted that naturally occurring sialic acids are monosaccharides which influence many important biological and pathological phenomena. In previous articles [3-51, the former literature has been extensively reviewed with respect to a number of chemical, biological, metabolic, functional as well as historical aspects. Since 1982 [4,5], the proliferation of the literature on the chemistry, biochemistry and (molecular) biology of sialic acids has accelerated dramatically, and several short reviews dealing with specific aspects of sialic acids have appeared [6-121. The aim of this chapter is to collate a mixture of, what is in our opinion, relevant information published before 1982 and new data appeared since then. Because of the profusion of biochemical and biological data published in the last few years, it was necessary to select those reports which, we believe, reflect potential trends in the future development of sialobiologyr

2. General characteristics of sialic acid In Table 1 a survey of the 43 naturally occurring members of the sialic acid family [ 135 13, together with their abbreviations and typical biological sources, is presented. The general name “sialic acid” is derived from the Greek “sialos”, meaning saliva. Taking into account the Rules for Carbohydrate Nomenclature, as recommended by the IUPAC-IUBMB Joint Commission on Biochemical Nomenclature (1999, the mother-molecule neuraminic acid, which does not occur in free form in nature, due to its immediate cyclization to form an internal Schiff base, is systematically named 5-amino-3,5-dideoxy-~-gbcero-~galacto-non-2-ulosonic acid (Fig. 1), and abbreviated as Neu, whereby the D-notation is implied in the trivial name. Chemically, this nine-carbon-containing monosaccharide is a 2-keto-carboxylic acid, a deoxysugar, and an aminosugar. The amino group is generally N-acetylated (5-acetam~do-3,5-d~deoxy-~-g~cero-~-galacto-non-2-ulopyranoson~c acid;

Table 1 Survey of established structures of naturally occurring members of the sialic acid familya Name

Abbreviation

Typical biological sources

Neuraminic acid

Neu

does not occur in free form; only found in gangliosides

N-Acetylneuraminic acid

Neu5Ac

5-N-Acetyl-4-0-acetyl-neuraminicacid

Neu4,5Ac2

5-N-Acetyl-7-0-acetyl-neuraminicacid

Neu5,7Ac2

5-N-Acetyl-8-0-acetyl-neuraminicacid

Neu5,8Ac2

5-N-Acetyl-9-0-acetyl-neuraminicacid

Neu5,9Ac2

oligosaccharides, polysaccharides and glycoconjugates from man, higher animals, and microorganisms; free form in body fluids ungulate, monotreme and guinea pig oligosaccharides and glycoconjugates animal glycoconjugates; bacterial polysaccharides bovine glycoproteins; bacterial polysaccharides human and animal glycoconjugates; bacterial (1ipo)polysaccharides

5-N-Acetyl-4,9-di-O-acetyl-neuraminic acid 5-N-Acety1-7,9-di-O-acetyl-neuraminic acid

Neu4,5,9Ac3

equine glycoproteins

~151

Neu5,7,9Ac3

bovine glycoproteins and bacterial lipopolysaccharides

[ 15J6,I 8,22,23,

5-N-Acetyl-8,9-di-O-acetyl-neuraminic acid 5-N-AcetyI-7,8,9-tri-O-acetyl-neuraminic acid 5-N-Acetyl-9-O-~-lactyl-neuraminic acid

Neu5,8,9Ac3

bovine glycoproteins

Neu5,7,8,9Ac4

bovine glycoproteins

NeuSAc9Lt

5-N-Acetyl-4-0-acetyl-9-0-lactyl-neuraminic acid

Neu4,5Ac29Lt

human and animal glycoproteins; free form in body fluids equine glycoproteins

Reference(s)

[13,15-181

[I 5,16,18-21]

[15,16,18,22-241 [ 15,22-241

[15,16,18,22-261

26,271 [ 15,16,18,22,23] [15,221 [15,28,29] [15,301 continued on next page

N

N m P

Table 1, continued Name

Abbreviation

Typical biological sources

5-N-Acetyl-8-0-methyl-neuraminic acid 5-N-Acetyl-9-0-acetyl-8-0-methyl-neuraminic acid 5-N-Acetyl-8-0-sulpho-neuraminic acid 5-N-Acetyl-9-0-phosphoro-neuraminic acid

NeuSAc8Me

starfish glycoconjugates starfish glycoconjugates

5-N-Acetyl-2-deoxy-2,3-didehydro-neuraminic acid 5-N-Acetyl-9-0-acetyl-2-deoxy-2,3-didehydro-neuraminic acid 5-N-Acetyl-2-deoxy-2,3-didehydro-9-0-lac~l-neuraminic acid 5-N-Acetyl-2,7-anhydro-neuraminic acid

Neu2en5Ac Neu2en5,9Ac2 Neu2enSAc9Lt Neu2,7anSAc

N-Glycolylneuraminic acid

NeuSGc

glycoconjugates from most higher animals

4-0-Acetyl-5-N-glycolyl-neuraminicacid 7-0-Acetyl-5-N-glycolyl-neuraminic acid

Neu4AcSGc Neu7AcSGc

8-0-Acetyl-5-N-glycolyl-neuraminicacid 9-0-Acetyl-5-N-glycolyl-neuraminicacid

Neu8AcSGc Neu9AcSGc

ungulate glycoconjugates glycoconjugates from most higher animals bovine glycoproteins

7,9-Di-O-acetyl-5-N-glycolyl-neuraminic acid 8,9-Di-0-acetyl-5-N-glycolyl-neuraminicacid 7,8,9-Tri-0-acetyl-5-N-glycolyl-neuraminic acid 5-N-glycolyl-9-0-lactyl-neuraminic acid

Neu7,9Ac25Gc Neu8,9Ac25Gc

5-N-glycolyl-8-0-methyl-neuraminic acid 9-0-Acetyl-5-N-glycolyl-8-0-methyl-neuraminic acid 7,9-Di-0-acetyl-5-N-glycolyl-8-O-methyl-neuraminic acid

NeuSGc8Me

Neu5,9Ac28Me NeuSAc8S NeuSAc9P

Neu7,8,9Ac35Gc NeuSGc9Lt

Reference(s)

sea urchin glycolipids biosynthetic intermediate to NeuSAc body fluids and tissues urine and tissues urine and tissues urine, wet cerumen, leech

glycoconjugates from most higher animals bovine glycoproteins bovine glycoproteins bovine glycoproteins porcine glycoproteins

Neu9AcSGc8Me

starfish glycoconjugates starfish glycoconjugates

Neu7,9Ac25Gc8Me

starfish glycoconjugates

[ 15-17,34]

[351 [351 [36-391

[I 3,15,16,18,25, 401 [ 15,16,18] [15,221 [361 [ 15,16,18,22]

Table 1, continued Name

Abbreviation

Typical biological sources

5-N-Glycolyl-8-0-sulpho-neuraminic acid N-(0-Acetyl)glycolylneuraminic acid N-(0-Methy1)glycolylneuraminic acid 2-Deoxy-2,3-didehydro-5-N-glycolyl-neuraminic acid 9-0-Acetyl-2-deoxy-2,3-didehydro-5-~-glycolyl-neuraminic acid 2-Deoxy-2,3-didehydro-S-N-glycolyl-9-0-lactyl-neuraminic acid 2-Deoxy-2,3-didehydro-5-N-glycolyl-8-0-methyl-neuraminic acid 2,7-Anhydro-5-N-glycolyl-neuraminic acid 2,7-Anhydro-S-N-glycolyl-8-~-methyl-neuraminic acid

NeuSGc8S

sea urchin glycolipids

NeuSGcAc

rat thrombocytes

NeuSGcMe

Neu2enSGc9Lt

starfish glycolipids body fluids and tissues urine urine and tissues

Neu2enSGc8Me Neu2,7anSGc

starfish rat urine

Neu2,7anSGc8Me

starfish

2-Keto-3-deoxynononic acid

Kdn

fish, amphibian and mammalian glycoconjugates; bacterial polysaccharides

[46-501

9-0-Acetyl-2-keto-3-deoxynononic acid

Kdn9Ac

fish glycoconjugates

[511

Neu2enSGc Neu2en9AcSGc

Reference(s)

a If possible, typical biological sources have been indicated. In the case of a sialic acid structure being proven by mass spectrometry andor NMR spectroscopy, reference numbers refer in general to such studies or to review articles including this sialic acid.

N

m P

E

COOH

C

I

O

H

I

c=o I p z

H-C-OH-

I

H,N-C-H

I

HO-C-H I

c - OH'

t

H

H-C-OH A ~ H N - &I H

I

I

H-

-

AcH( D-galacto

rH-C > -OH H

0-L-H reference --c H- I - OH

atom

H

I

H-C-OH

I

H- C - OH

I

D-glycero

CH,OH

I

OH

CH,OH

'w 1

OH

HO

H

AcHN

5

OH

qOOH

I

HOH,C

CH,OH

OH

3

?H

1

H

o

H

H

~

~

C

O

O

H

HO OH

Fig. 1. Chemical structures for simple sialic acids in different views. (a) 5-arnino-3,5-dideoxy-o-glycero-o-galacfo-non2-ulosonic acid (Neu, open chain, Fischer projection formula); (b) 5-acetamido-3,5-dideoxy-~-g~ceroa-o-g~~acfonon-2-ulopyranosonic acid (a-NeuSAc, Fischer projection formula, note that C7 is the anomeric reference atom); (c) a-NeuSAc (Haworth projection formula); (d) a-NeuSAc ('C, chair conformation); (e) 3-deoxy-o-glycero-~~-galacto-non-2-ulopyranosonic acid (P-Kdn, 'C, chair conformation). Note that the o-notation is part of the trivial name.

249

0

II

H0 H,

c

>kOH

O-P-O-

HO

AcHN

OH Fig. 2. Chemical structure of cytidine 5’-(5-acetam~do-3,5-d~deoxy-D-g~cero-~-D-g~~~cfo-non-2-ulopyranosylonate monophosphate (CMP-P-Neu5Ac).

N-acetylneuraminic acid; Neu5Ac) (Fig. 1) or N-glycolylated (5-hydroxyacetamido3,5-d~deoxy-~-glycero-~-galacto-non-2-ulopyranoson~c acid; N-glycolylneuraminic acid; Neu5Gc). As is evident from Table 1, the hydroxyl groups may be free, esterified (acetylated, lactylated, sulfated, phosphorylated) or etherified (methylated). In the case where a sialic acid bears a hydroxy group instead of an amino group at C5, so-called deaminated neuraminic acid, it is systematically named 3-deoxy-D-glycero-D-galacto-non2-ulopyranosonic acid (Fig. l), and is abbreviated as Kdn (2-keto-3-deoxynononic acid; also in this case the D-notation is part of the trivial name). Sialic acids are relatively strong acids, e.g. Neu5Ac has a pK, value found in the range of 2.2-3.0 in various studies with an average of 2.6. This strong acidity is responsible for processes such as autohydrolysis of sialic-acid-containing carbohydrate chains. Conformationally, sialic acids adopt the 2Cs chair conformation, having the glycerol side chain in an equatorial orientation [52]. Due to hydrogen bonding of HO7 and H08, leading to a trans-orientation of these groups, the glycerol side chain is not as flexible as one might expect. This has consequences for the course of the mild periodate oxidative degradation of this moiety in the case of substitution at C9[53]. Free sialic acids have mainly the b-anomeric ring structure (>93%), whereas glycoconjugate-bound sialic acids occur specifically in the a-anomeric form [54]. In nucleotide-bound sialic acids, i.e. CMP-Neu5Ac (Fig. 2), a P-configuration for the glycosidic bond is present [55]. Crystalline Neu5Ac occurs specifically in the 0-anomeric form [56]. With respect to 0-acetylation patterns in sialic acids, it is worthwhile to mention that spontaneous migration of 0-acetyl groups can occur between C7, C8 and C9. At pH values at which no significant de-0-acetylation is observed (e.g. physiological pH values), Neu5,7Ac2 can readily transform into Neu5,9Ac2, whereas Neu5,7,9Ac3 yields an equilibrium of Neu5,7,9Ac3 and Neu5,8,9Ac3 in a molar ratio of approximately 1:l; Neu4,5Ac2 does not give rise to 0-acetyl migrations[23]. Also starting from a-Neu5,8,9Ac3 4-aminophenylthio-glycosidea 1:1 equilibrium between the 8,9- and 7,9-di-O-acetylated derivatives is established [57]. 9-0-Acetylated N-acylneuraminic acids have been found in both glycoconjugates and oligolpolysaccharides of different

250 AcHN

OH

I

1

AcHN

y3rcoo

1

OH

HO

H

-

CH,OH

COOH

N ----+-OH

CH OH H

o H

e

AcHN o

O

W

/ o

o

0

H

HO

Fig. 3. Chemical structures of (a) 5-N-acetyl-2-deoxy-2,3-didehydro-neuraminic acid (Ne3en5Ac); (b) 5-Nacetyl-2,7-anhydro-neuraminic acid (Ne3,7an5Ac); and (c) the tautomers of S-N-acetyl-4,8-anhydroneuraminic acid (Neu4,SanSAc).

biological origin (Table 1). The same is true for 7-0-acetylated N-acylneuraminic acids. It should be emphasized that Neu5,9Ac2 and other side-chain-0-acylated sialic acids occur in man[%] (see also sections 8.4.2 and 8.4.3), e.g. in human colon. Ungulates form the major source for 4-0-acetylated N-acylneuraminic acids [ 131, whereas minor sources are monotremes [19], guinea pigs [21] and humans [59]. In an evaluation of the naturally occurring 0-acylation patterns, it is evident that 0-acyl groups are most frequently found at C9 of both NeuSAc and NeuSGc. 0-Methylated and 0-sulfated N-acylneuraminic acids have only been found in lower animals, e.g. in echinoderms [13,31,42]. 5-N-Acyl-2-deoxy-2,3-didehydro-neuraminic acids like Neu2enSAc (S-acetamido-2,6anhydro-3,5-d~deoxy-~-g~cero-~-galacto-non-2-enon~c acid, Fig. 3) occur in free form in nature. Moreover, they can be generated from corresponding CMP-N-acylneuraminic acids in a non-enzymatic elimination reaction, occurring under physiological and, much faster, under alkaline conditions [34,35]. Small amounts of Neu2enSAc are formed by a water elimination side reaction from NeuSAc during influenza-B-virus-sialidase-catalyzed desialylations of sialoglycoconjugates [60]. S-N-Acety1-2,7-anhydro-neuraminic acid (Neu2,7anSAc, Fig. 3), which occurs in free form in nature [36,37], can be generated from NeuSAc(a2-3)Gal(fi 1- containing glycoconjugates using a sialidase isolated from the leech Macrobdella decara [38,6 13.

25 1

It is clear that such a sialidase has a highly unusual specificity (see sections 5.1, 8.4.6 and 9.2). The methyl ester of Neu2,7anSAc is formed as a very minor by-product during methanolysis [62]. 5-N-Acetyl-4,8-anhydro-neuraminicacid (Neu4,8anSAc, Fig. 3), having CH2-COCOOH as a side chain of a pyranose ring with the 7C4 conformation, does not occur as such in nature. Initially, it has been isolated from an acid hydrolysate of edible bird’s nest substance [63]. Furthermore, Neu4,8an5Ac was detected in a sialic acid mixture obtained after acid hydrolysis of equine serum, whereas incubation of Neu4,5Ac2 under alkaline conditions showed, in addition to de-0-acetylation, a partial conversion into Neu4,8an5Ac [41]. Heating of solid sodium Neu5Ac (3 h, 140°C) or a solution of Neu5Ac at pH > 11 or at pH 2.0 (30 min, 80°C) also yielded Neu4,8anSAc (sodium salt) as a major degradation product, probably in two tautomeric forms [64]. Hyperexcretion of free Neu5Ac in urine, defined as sialuria, has been observed in several mentally retarded patients [ 6 5 4 8 ] (see section 10.5).

3. Occurrence of sialic acids in biornolecules Members of the sialic acid family occur mainly in bound form in higher animals, from the echinoderms onwards in evolution, but also in some viruses, various bacteria, protozoa, and pathogenic fungi [8,13,69]. Generally, they are constituents of glycoproteins [20,21,25,40,46-48,5 1,70-I071 (Table 2), glycolipids [13,3 1,44,108-1241 (Table 3), and oligosaccharides [ 16,19,25,76,125-1551 (Table 4), usually occurring as terminal monosaccharide units, and of homo- and heteropolysaccharides [24,26,27,49,156I851 (Table 5). However, glycosylphosphatidylinositol membrane anchors have also been shown to contain sialic acid [ 1861. Neu5Ac- and Neu5Gc-containing glycans frequently occur together, whereby ratio differences reflect species specificity, tissue specificity or physiological fluid specificity. Also Kdn has been found together with NeuSAc/SGc in specific glycoconjugates (Table 2). It should be noted that the presence of Neu5Gc in normal human tissue and soluble glycoproteins has not been established conclusively [ 13,187,1881. As is evident from Table 2, NeuSAc/SGc- and Kdn-containing elements in Nand O-glycoprotein glycans do occur in many different microenvironments, however, with a restricted glycosidic linkage specificity. In general, N-acylneuraminic acids are a-2,3- or a-2,6-linked with D-galactose (Gal), a-2,3- or a-2,6-linked with N-acetyl-Dgalactosamine (GalNAc), a-2,6-linked with N-acetyl-D-glucosamine (GlcNAc), or a-2,8linked with other N-acylneuraminic acids. These types of glycosidic linkages have been firmly established by different analytical methods, including NMR spectroscopy [76,77]. The sialic-acid-containing N-linked carbohydrate chains form part of the complex (Nacetyllactosamine and N,N’-diacetyllactosediamine subtypes; mono-, di-, tri-, tri’-, and tetraantennae) or the hybrid type of structures [ 1891. For the mucin-type 0-glycans, the N-acylneuraminic-acid-containingsequences are extensions of most of the wellestablished core types [ 1901. In addition, terminal NeuSAc(a24)Gal[75], terminal NeuSAc(a24)GlcNAc [96], and internal Fuc( 14)NeuSGc [97] sequences have been reported, and only one example of a terminal NeuSAc(a2-9)NeuSAc(a2- dimer has

252 Table 2 Sialic-acid-containing elements in N- and 0-glycoproteins a Partial structure

N

0 Ref(s).

Neu5Ac(a2-3)Gal(fil-3)GlcNAc(filNeu5Ac(a2-3)Gal(fi1-3)~eu5Ac(a2-6)]GlcNAc(fil-

+

[70,71]

+

NeuSAc(a2-3)Gal(fi 1-3)[Fuc(a 14)]GlcNAc(fiI-

[70,72,73] +

[741 [751 [70,73]

Neu5Ac(a24)Gal(fil-3)[Neu5Ac(a2-6)]GlcNAc(filGal(fi1-3)[Neu5Ac(a24)]GlcNAc(fi 1Neu5Ac(a2-3)Gal(fil4)GlcNAc(fi 1NeuSAc(a2-3)Gal(fi 14)[Fuc(a 1-3)]GlcNAc(fi 1-

+

+ +

[77,78]

NeuSAc(a2-3)[Gal(fiI 4)]Gal(fi 14)GlcNAc(filNeuSAc(a2-3)[GalNAc(fiI 4)]Gal(fi l4)GlcNAc(fi1-

+ + +

[79,80] [77,81]

Neu5Ac(a2-3)[6S]GaI(fil 4)GlcNAc(fil-

+

[821

Neu4,5Ac2(a2-3)Gal(fi1 4)GlcNAc(filNeu4,5Ac2(a2-3)[Ga1((3 14)]Gal(fil4)GlcNAc(fil-

+

Neu5,9Ac2(a2-3)Gal(fi14)GlcNAc(filNeu5Gc(aZ-3)Gal(fil4)GlcNAc(filNeuSAc(a24)Gal(fiI4)GlcNAc(fiI Neu5Ac(a2-6)Gal(fi14)GlcNAc(fiI-

+ +

~ 3 1 [791 [25,84]

Neu5Ac(a24)Gal(fiI -4)[6S]GlcNAc(filNeu5Ac(a24)[GalNAc(a 1-3)]Gal(fil4)GlcNAc(fil-

+

Neu4,5Ac2(a2-6)Gal(fi 14)GlcNAc(fiI-

+

Neu5Gc(a2-6)Gal(~14)GlcNAc(fil-

+

Neu5Ac(a2-3){ Gal(fiI -4)) ,_2Gal(fi14)GlcNAc(fil -

+

fNeu5Ac(a2-3)Gal(fi1-4)[tNeu5Ac(a2-3)]Gal(fil4)[fFuc(aI-3)]GlcNAc(fiI -

+

+ + +

+

[76,77]

[25,84]

+

+ + +

[751 [76,77]

~ 5 1 [861 [20,21,87,88] [401 [80,89]

NeuSAc(a2-3)Gal(fiI -3)GalNAc(a 1Neu5Ac(a2-3)Gal(fi1-3)GalNAc(a 1-O)Ser/Thr

+

~901 [771

+

1771

Neu5Ac(a2-8)Neu5Ac(a2-3)Gal(fi1-3)GalNAc(a l-O)Ser/Thr

+

[911

NeuSAc(a2-3)[GalNAc(fi I4)]Gal(fi 1-3)GalNAc(a 1-O)Ser/Thr

+ +

1771 [771

NeuSGc(a2-3)Gal(fil-3)GalNAc(a I-O)Ser/Thr Neu5Ac(a2-3)GalNAc(fi14)GlcNAc(filNeu5Ac(a2-6)GaINAc(fil 4)GlcNAc(fiINeu5Ac(a2-3)[GalNAc(fi14)]GalNAc(fi 1-3)Gal(fi14)Gal(fiINeu5Gc(a2-3)[GalNAc(fil4)]GalNAc(fil-3)Ga1(fil 4)Gal(fiI -

+

1921 [93,94]

+ +

[951

+

[771

NeuSAc(a2-6)GalNAc(a I-O)Ser/Thr NeuSGc(a2-6)GalNAc(a I-O)Ser/Thr

+

[771 [771

Neu5Ac(a24)GlcNAc(fi1-3)[Neu5Ac(a2-6)]Gal(fil-

+

-Fuc( 14)Neu5Gc(2-

+

+

Neu5Ac(a2-9)Neu5Ac(a2-3/6)Gal(fi 14)GlcNAc(fil-

+

{Neu5Ac(a2-8)},Neu5Ac(a2-3)Ga1(fi14)GlcNAc(~1{Neu5Ac(a2-8)},Neu5Ac(a2-6)GalNAc(al -O)Ser/Thr

+

+

[961 1971 [981 [991 I771

continued on next page

253

Table 2, continued ~

Partial structure

N

0 Ref(s).

{NeuSGc(a2-8)},NeuSGc(a24)GalNAc(a 1-O)Ser/Thr {NeuSAc(a2-8)},- with 4Ac, 7Ac, 9Ac or 9Lt

+

+

[771 [loo]

(NeuSGc(a2-8)},- with 4Ac, 7Ac or 9Ac

+

[51,100]

{Neu5Ac/SGc(a2-8)},- with 4Ac, 7Ac or 9Ac

+

[loo]

{NeuSGc(a2-05)},,-

+

[loll [lo21

Kdn(a2-3)Gal(fiI 4)GlcNAc(fi1Kdn(a2-3)Gal(o1-3)GalNAc(a 1-3)GalNAc(a 1-O)Ser/Thr

+ +

[471

Kdn(a2-3)[CalNAc(fi 14)]GalNAc(fi1-3)Gal(fi14)Gal(~lKdn(a2-6)GalNAc(a I-O)Ser/Thr

+

[771 [48,77,103]

{ Kdn(a2-8)}, Kdn(a24)GalNAc(a I -O)Ser/Thr Fuc(a 1-4)[Fuc(a 1-S)]Kdn(a24)GalNAc(a 1-O)Ser/Thr

+ [471

-Fuc(a14)Kdn(aZ-6)GalNAc(al-O)SeriThr Kdn(a2-6) [Kdn(a2-3)] G alNAc(a 1-0)Thr Kdn(a2-3)GalNAc(al-O)Thr Fuc(al4)[{ Fuc(a 1-5)}0-, ]Kdn(a2-3)Gal(fiI -

+

+ + +

[I041

[I051

+

[lo61 [106a] [I041

(Kdn(a2-8)},- with 4Ac, 7Ac or 9Ac Kdn(a2-8){Ne~SAc(a2-8)}~- with fAc

+ +

[I071 [lo71

Kdn(a2-8){Ne~SGc(a2-8)}~-with f A c Kdn(a2-8){Ne~SAc/SGc(a2-8)}~-with fAc

+

~461 [I071

Kdn9Ac(a2-8) { NeuSGc(a2-8)}, -

+ ~511

+

+

a In the case of a specific fragment being established by ' H NMR spectroscopy, the reference refers to such a study or to a review that includes the fragment. S means sulfate.

been described [98]. In a number of structures not only sialic acid is responsible for the acidic character of the carbohydrate chain, but also sulfate. Here, an unusual example is the Neu5Ac(a2-3)[6S]GaI@ 14)GlcNAc element [82]. The list of sialic-acid-containing sequences, in which N-acylneuraminic acid is replaced by Kdn is continuously growing (Table 2). Interestingly, also structural elements occur which have not been found so far for N-acylneuraminic acids, e.g. Kdn(a2-6)[Kdn(a2-3)]GalNAc(al-O)Thr [ 1061. General structural information with respect to glycoprotein glycans is presented in volume 29a of the New Comprehensive Biochemistry series, and detailed information with respect to poly-N-acylneuraminyl-containing glycoproteins and Kdn-containing glycoproteins in the present volume 29b. Several of the sialic-acid-containing sequences present in N- and 0-glycoprotein glycans, also occur in glycolipids, milk and urinary oligosaccharides, and in (1ipo)polysaccharides of different biological origin. An impression of this overlap in structures can be obtained from an inspection of the structural data in Tables 3-5, summarizing sialic-acid-containing elements of glycolipids, structures of milk and urinary sialo-oligosaccharides, and structures or elements of sialic-acid-containing (lipo)polysaccharides, respectively. It is interesting to note that

v, N P

Table 3 Compilation of sialic-acid-containingfragments in glycolipidsa Structure

References

Neu5Ac(a2-3)Gal(P 1- I’)Cer {Neu5Ac(a2-8)}o~,Neu5Ac(a2-6)Glc(~l-l’)Cer Neu5Gc/5Gc8S(a2-6)G1c(~l-1’)Cer Neu5Ac(a24)Glc(~l-8)Neu5Ac(a24)Glc(~I-l’)Cer Neu5Ac(a2-8)Neu5Ac(a24)Glc(P14)Glc(~l-I’)Cer Neu5Ac/5Gc/4AcSGc/5Gc8Me/5GcMe(a2-3)Ga1(~14)Glc(~1-1 ’)Cer Neu5Ac/5,7Ac2/5,9Ac2/5,7,9Ac3/5Ac8S/5Gc(a2-8)Neu5Ac(a2-3)Gal(~1 -4)Glc(PI -1’)Cer Neu5Ac/5Gc(a2-8)Neu5Gc(a2-3)Gal(P14)Glc(P1-1 ‘)Cer Neu5Ac(a2-8)Neu5Ac(a2-8)Neu5Ac(a2-3)Gal(~l4)Glc(~l-l ‘)Cer Neu5,9Ac2(a2-8)Neu5Ac(a2-8)Neu5Ac(a2-3)Gal(~1 -4)Glc(~l-I’)Cer GalNAc(~14)[Neu5Ac/5Ac8S/5Gc/5Gc8S(a2-3)]Gal(~14)Glc(~ l-l’)Cer l-l’)Cer Gal(PI-3)GalNAc(P 14)[Neu(a2-3)]Gal(~l4)Glc(~ [R-3]GalNAc(P14)[Neu5Ac/5Gc(a2-3)]Gal(P 14)Glc(P 1-1 ’)Cer [R-3]GalNAc(fi4 I)[{Neu5Ac(a2-8)} l-2Neu5Ac(a2-3)]Gal( 1~ 4)Glc(~l-l’)Cerc [R-3]GalNAc(f 1~ 4)[Neu5,9Ac2 (a2-8){ Neu5A~(a2-8)}~-~ Neu5Ac(a2-3)]Gal(P 14)Glc(PI -1 ’)Cer‘ [R-3]GalNAc(PI4)[{Neu5Gc(a2-8)} l_zNeu5Gc(a2-3)]Gal(P 14)Glc(P1-1’)Cerb l-l’)Cer [R-3]GalNAc(fi-4)[Neu5Ac(a2-8)Neu5Gc(a2-3)]Gal(~l4)Glc(~ I Ara( 14)Gal(P 1 4 ) [{Gal@ 1-8))o-1 ]Neu5Gc/5Gc8Me(a2-3)Gal(P 14)Glc(P 1-1 ’)Cer Neu5Gc8Me(2-3)GalNAc( 1-3)Gal( 14)Glc( 1-1’)Cer Neu5Gc8Me(a2-3)[NeuSGc8Me(a24)]GalNAc(~1-3)Gal(~I 4)Glc(Pl-l’)Cer Neu5Ac/5,9Ac2/5Gc(a2-3)Gal(~l-3)GalNAc(~1 {Neu5Ac(a2-8)} ,-2Neu5Ac(a2-3)Ga1(~1-3)GalNAc((3 1Neu5Gc(a2-8)Neu5Gc(a2-3)Gal(flI-3)GalNAc(P 1Neu5Ac(a2-3)[Neu5Ac(a24)]Gal(~1-3)GalNAc(~ 1continued on next page

Table 3, continued Structure

References

Neu5Ac(a2-3)[[R-3]GalNAc(fil4)]Gal(fil-3)GalNAc(filNeu5Gc(a2-3)[GalNAc(!314)]Gal(fil-3)GalNAc(fil{Neu5Ac(a2-8)}o~,Neu5Ac(a2-3)Gal(fi1-3)[{Neu5Ac(a2-8)}~~lNeu5Ac(a2~)]GalNAc(~1 NeuSAc(a2-3)Gal(fil-3)~eu5Ac(a2-6)]GalNAc(fi14)[{Neu5Ac(a2-8)}o~l Neu5Ac(a2-3)]Gal(fi1-4)Glc(fil-l’)Cer NeuSAc(a24)Gal(fil4)GlcNAc(filNeu5Ac/5Gc(a2-3)Gal(fil4)GlcNAc(fi 1{Neu5Ac(a2-8)}1-2Neu5Ac(a2-3)Ga1(fi 14)GlcNAc(fi1 Neu5Ac/5Gc(a2-8)Neu5Gc(a2-3)Gal(fl1 4)GlcNAc(P1 Neu5Ac(a2-3)[GalNAc(~14)]Gal(fil4)GlcNAc(filNeuSAc(a2-3)GaI(fi1 -4)[Fuc(a1 -3)]GlcNAc(fi1Neu5Ac(a2-3)Gal(fi1-3)GlcNAc(filNeu5Ac(a2-3)[GalNAc(fiI4)]Gal(fi1-3)GlcNAc(fllNeu5Ac(a2-3)Gal(fil-3)[Fuc(al4)]GlcNAc(~lNeu5Ac(a2-3)Gal(~1-3)~eu5Ac(a2-6)][R-4]GlcNAc(fil-c Neu5Ac(a2-3)GalNAc(fi1-3)Gal(a 1 NeuSAc(a2-3)GalNAc(fi1-3)Gal(fi1 (NeuSGc8Me(a2-05)},Kdn(a2-3)Gal(fil4)Glc(fi 1-1 ’)Cer

[R-3]GalNAc(pI 4)[Kdn(a2-3)]Gal(fl 14)Glc(fil-l’)Cer Kdn/Kdn9Ac(a2-3)Gal(fi1 -3)GalNAc(fi1Kdn(a2-3)Gal(fi1 -3)[Neu5Ac/Kdn(a2-6)]GalNAc(fi1 a

S, sulfate; Cer, ceramide. R, saccharide.

R, H or saccharide

Table 4 Sialic-acid-containing human milk oligosaccharides Structures

References

Neu5Ac(a2-3)Gal(fl 1-4)Glc a-e Neu5Ac(a24)Gal(~l-4)Glca,b

[ 16,125-1271

Neu5Ac(a2-3)Gal(fi1-3)GlcNAc(~l-3)Gal(fil4)Glc Neu5Ac(a2-3)Gal(~l-3)[NeuSAc(a2-6)]GlcNAc(fil-3)Gal(~14)Glc qf Neu5Ac(a2-3)Gal(~l-3)[Fuc(a 14)]GlcNAc(fll-3)Gal(~l4)Glc

[ 12S,l26,128,129]

Neu5Ac(a2-3)Gal(P 1-3)[Fuc(a 1 4 ) ][NeuSAc(a24)]GIcNAc(PI-3)Gal(fi 1-4)Glc NeuSAc(a2-3)[Fuc(a 1-2)]Gal(P 1-3)GlcNAc(fi 1-3)Gal(P 14 ) G l c

[ 129,1301

Gal(~1-3)[NeuSAc(a24)]GlcNAc(fil-3)Gal(flI4)G1~ a Gal@1-3) [NeuSAc(a24)]GlcNAc(fi1-3)Gal(fi14)Glc

[ 16,125-1 271

[12S,128-130] [ 1 2 83~1,1321

[1331 [12S,126,128]

Fuc(a 1-2)Gal(fi 1-3)[NeuSAc(a24)]GlcNAc(P 1-3)Gal(P 14)Glc NeuSAc(a24)Gal(~14)GlcNAc(~1-3)Gal(fil 4)Glca

[1341 [128,131,13S] [ 125,126,1281

NeuSAc(a2-3)Gal(P 1-3)GlcNAc(@1-3) { Gal(fi14)[Fuc(a 1-3)]GlcNAc(fi 14)}Gal(P 14)Glc Neu5Ac(a2-3)Gal(~1-3)GlcNAc(fil-3)[NeuSAc(a2-6)Gal(fll-4)GlcNAc(fiI4)]Ga1((314)Glc

~361 ~291

NeuSAc(a2-3)Gal(~1-3)[NeuSAc(a2-6)]GlcNAc(~l-3)[Gal(~l4)GlcNAc(~14)]Gal(~I4 ) G l c NeuSAc(a2-3)Gal(~1-3)[NeuSAc(a2-6)]GlcNAc(P l-3)[Fuc(a 1-3)Gal(~l-4)GlcNAc(~l4)]Gal(fi14)Glc NeuSAc(a2-3)Gal(fi1-3)[NeuSAc(a2-6)]GlcNAc(fi l-3){ Gal(fi14)[Fuc(al-3)]GlcNAc(fi14)}Gal(PI -4)Glc NeuSAc(a2-3)Gal(fi1-3)[Neu5Ac(a2-6)]GlcNAc(P 1-3)[Neu5Ac(a24)Gal(~l4)GlcNAc(~l-6)]Gal(~l-4)Glc NeuSAc(a2-3)Gal(fi1-3)[Fuc(a 1-4)]GlcNAc(~l-3)[Gal(fiI4)GlcNAc(fil4)]Gal(~I 4)Glc NeuSAc(a2-3)Gal(P 1-3)[Fuc(a 14)]GlcNAc(P 1-3){ Gal(P14)[Fuc(a 1-3)]GlcNAc(P 1-6)}Gal(@ 14)Glc Neu5Ac(a2-3)Gal(fil-3)[Fuc(a 14)]GlcNAc(fi 1-3)[Neu5Ac(a24)Gal(PI4)GlcNAc(Pl4)]Gal(~14)Glc

~291 [1371 [ 129,130,137]

[1381 [1391 [1391 [1391

Neu5Ac(a2-3)Gal(PI-3)[Fuc(a

IH)]GlcNAc(fl 1-3)[Gal(~I-4)GlcNAc(~ld)Gal(~ 1-4)GlcNAc(~1-6)]Gal(PI4)Glc Neu5Ac(a2-3)Gal(P 1-3)[Fuc(a 14)]GlcNAc(fi 1-3) {Gal@14)[Fuc(a 1-3)]GlcNAc(fl 1-3)Gal(fl14)GlcNAc@ 1-6)}Gal(~14)Glc

[ 1401

Neu5Ac(a2-3)Gal(P 1-3)[Fuc(a 1-4)]GlcNAc(P 1-3)[Gal(B 1-3)GlcNAc(P 1-3)Gal(fi 1-4)GlcNAc(P 1-6)]Gal(P 1 4 ) G l c NeuSAc(a2-3)Gal(P 1-3)[Fuc(a 1-4)]GlcNAc(fi 1-3) { Gal(fi1-3)[Fuc(a 1-4)]GlcNAc(fiI-3)Gal(fl 14)GlcNAc(j314)}Gal(fi1 4 ) G l c

~411

[1411 [1411 continued on next page

Table 4, continued Structures

References

Neu5Ac(a2-3)Gal(fi 1-3)[Fuc(a 14)]GlcNAc(!31-3){ Gal@1-3)GlcNAc(fi 1-3)Gal(P 14)[Fuc(a 1-3)]GlcNAc(P 1 4 ) } Gal@ 1-4)Glc Neu5Ac(a2-3)Gal(P 1-3)[Fuc(a 14)]GlcNAc(fi1-3){ Fuc(a 1-2)Gal(P 1-3)GlcNAc(fi 1-3)Gal(fi 1-4)[Fuc(a 1-3)]GlcNAc(fi 14)}Gal(fi 14)Glc Gal(fi1-3)GlcNAc(fi 1-3)[Neu5Ac(a24)Gal(fi 14)GlcNAc(fi1-6)]Gal(B 1 4 ) G l c

Gal(fiI-3)[NeuSAc(a24)]GlcNAc(fi1-3) { Gal(fiI4)[Fuc(a 1-3)]GlcNAc(P 14)}Gal(fi 14 ) G l c Gal(~1-3)~eu5Ac(a24)]GlcNAc(~l-3)[Gal(~1-3)GlcNAc(~14)]Gal(~l4)Glc Gal(~1-3)[Neu5Ac(a24)]GlcNAc(~l-3)[Fuc(al-2)Gal(~l-3)GlcNAc(~ 14)]Gal@l4)Glc Gal@1-3)[Fuc(a 1-4)]GlcNAc(fi 1-3)[Neu5Ac(a24)Gal(fil-4)GlcNAc(fi1-6)]Gal(PI 4 ) G l c Gal(P1-3)[Fuc(a 14)]GlcNAc(P 1-3){Neu5Ac(a2-3)Gal(fi 14)[Fuc(a 1-3)]GlcNAc(P 1-6)}Gal(P 14 ) G l c Fuc(al-Z)Gal(fi 1-3)GlcNAc(fil-3)~eu5Ac(a24)Gal(fil 4)GlcNAc(fiI4)]Gal(fi14)Glc Fuc(a 1-2)Gal(fi 1-3)GlcNAc(fi 1-3) {NeuSAc(a2-3)Gal(fi 14)[Fuc(a 1-3)]GlcNAc(fi 14)}Gal(fi 1 4 ) G l c Fuc(a 1-2)Gal(fi 1-3)[Fuc(a 14)]GlcNAc(@1-3)[Neu5Ac(a2-6)Gal(P 14)GlcNAc(fi1d)]Gal(B 1-4)Glc Gal@1 4)GlcNAc(P 1-3)[NeuSAc(a24)Gal(fi 1-4)GlcNAc(fiI-6)]Gal(P 1 4 ) G l c

Neu5Ac(a24)Gal(fil4)GlcNAc(~l-3)[Gal(fi 14)GlcNAc(~l-6)]Gal(~14)Glc Neu5Ac(a24)Gal(fi14)GlcNAc(fi1-3){ Gal@ 14)[Fuc(a1-3)]GlcNAc(fi 14)}Gal(P 14)Glc Neu5Ac(a24)Gal(fi1-4)GlcNAc(fi1-3){ Fuc(a 1-2)Gal(fiI 4)[Fuc(a 1-3)]GlcNAc(fi 14)}Gal(fi 1 4 ) G l c Neu5Ac(a24)Gal(fil4)GlcNAc(fil -3)[NeuSAc(a24)Gal(P 14)GlcNAc(fil4)]Gal(fil4)Glc Neu5Ac(a2-6/3)Ga1(fi1-4)[Fuc(a1-3)]GlcNAc(fi1-3)~eu5Ac(a2-3/6)Gal(fil-4)GlcNAc(fi14)]Gal(P14)Glc

Neu5Ac(a2-3)GaI(fi14)[Fuc(al-3)]Glc NeuSAc(a24)Gal(!3 14)GlcNAc(P1-3)Gal(fi 14)[Fuc(a 1-3)IGlc

Neu5Ac(a2-3)Gal(~1-3)[Neu5Ac(a2-6)]GlcNAc(fiI-3)Gal(~14)[Fuc(a 1-3)]Glc NeuSAc(a2-3)Gal(P 1-3) {Gal@14)[Fuc(a 1-3)]GlcNAc(P 14)}Gal(P 14)Glc

Neu5Ac(a2-3)GaI(fi1-3)[Fuc(a l4)]GlcNAc Neu5Ac(a2-3)Ga1((31-3)[Fuc(a 14)]GlcNAc(fi 1-3)Gal

continued on next page

Table 4, notes Also occurring in feces of preterm infants fed on breast milk [147]. Also occurring in bovine colostrum and human (pregnancy) urine [148]. Other sialyloligosaccharides in bovine colostrum are: NeuSAc(a2-3)GaL Neu5Gc(a2-3)Gal(P14)Glc, Neu5Ac(a2-8)Neu5Ac(a2-3)Gal(~1-4)Glc,NeuSAc(a26)Gal(P 14)GlcNAc, Neu5Ac(a2-6)Gal(~14)GlcNAc(a1-P, NeuSAc(a2-6)Gal(P 1-4)[6P]GlcNAc. Other sialyloligosaccharides in human (pregnancy) urine are: Neu5Ac(a2-3)Gal(~14)GlcNAc, Neu5Ac(a24)Gal(Pl4)GlcNAc, NeuSAc(a2-3)GaI(~l4)GlcNAc(al-P, Neu5Ac(a24)Gal(Pl4)GlcNAc(al-F', Neu5Ac(a2-3)Gal(P1-3)[Neu5Ac(a24)]GalNAc, Neu5Ac(a2-3)Gal(~l-3)[Neu5Ac(a24)]GalNAc(alO)Ser, Neu5Ac(a2-3)Gal(PI-3)GalNAc(al -P, Neu5Ac(a2-3)Gal(~1-1 L)-myo-inositol, NeuSAc(a2-3)Gal(PI -)scyllo-inositol. The urinary carbohydrates, except the inositol derivatives and the tetrasaccharide, do also occur in hemofiltrates of patients with end-stage renal disease; in addition NeuSAc(a2-8)NeuSAc, Neu5Ac(a2-3)Gal(fll-3)GalNAc(aI-O)Ser-Leu and Neu5Ac(a2-3)Gal(PlIt)Xyl(Pl-O)Ser have been detected [ 149,1501. For NMR data of Neu5Ac(a2-3)Gal(~l4)Glc, Neu5Ac(a2-3)Gal(~l4)GlcNAc,and Neu5Ac(a24)Gal(PIIt)GlcNAc, see also [ 1511. Urine of patients with aspartylglycosaminuria contains Neu5Ac(a2-3)Gal(~l4)GlcNAc(Pl -N)Asn, Neu5Ac(a24)Gal(~14)GlcNAc(P1-N)Asn, and NeuSAc(a23)Gal(P 14)GlcNAc(~l-3)Gal(~I 4)GlcNAc(PI -N)Asn [76,152]. Urine of patients with P-mannosidosis contains NeuSAc(a24)Man((3l4)GlcNAc [ 1531. ' Neu5Ac(a2-3)[6S]Gal(P14)Glc occurs in rat milk and mammary gland [154]. Neu5,9Ac2(a2-3)Gal(P14)Glc occurs in rat urine (251. Neu4,5Ac2(a2-3)Gal(~14)Glcoccurs in monotreme milk [I 91. Occurs also in human pregnancy urine [155]. a

v h 1 ,

m

259

for both glycoprotein and glycolipid glycans a sequence has been found comprising the oligomerization of Neu5Gc residues through their anomeric centers and N-glycolyl moieties, Neu5Gc(a2-05)Neu5Gc(a2-05)Neu5Gc(a2- [3 1,lO 11; in the case of the glycolipid material the NeuSGc residues are also 8-U-methylated [3 I]. In microbial polysaccharides, besides internal 8- and 9-substituted NeuSAc residues, also internal 4- and 7-substituted Neu5Ac units have been frequently found (Table 5). It should be noted that some of the sialic-acid-containing glycan fragments and polysialic acids are specifically found in lipopolysaccharides and capsular polysaccharides of pathogenic bacteria, leading to severe problems in the development of suitable vaccines. A typical source for sialo-oligosaccharides generated from N- and O-linked glycans is the urine of sialidosis and I-cell disease patients [191-1991, though this is not discussed in detail here (see section 10.5, and volume 30 of the New Comprehensive Biochemistry series). Patients with other inborn errors of metabolism, like aspartylglycosaminuria [76,152] or P-mannosidosis [ 1531 excrete small amounts of structurally unusual sialooligosaccharides, of which the formulae have been included in the footnotes of Table 4. Finally, sequence information of the already mentioned glycosylphosphatidylinositol anchor is available, showing that the glycan core consists of Mana-Mana-Man-[NeuSAcGal-Ga1NAc)Man-GlcN-inositol [ 1861. The phenomenon of intramolecular lactone formation, often reported for polysialic acid [200-2021 and for gangliosides [203-2061, has not been detected so far in glycoprotein sialoglycans. In the case of a NeuSAc(a2-8)NeuSAc sequence, lactonization affords a NeuSAc(a2-8,1-9)Neu5Ac element (Fig. 4), whereby the COOH group of one residue reacts with H 0 9 of an adjacent residue, to give a six-membered ring. Similarly NeuSAc(a2-9)NeuSAc can be converted into Neu5Ac(a2-9,1-8)NeuSAc, and NeuSAc(a2-3)Gal into Neu5Ac(a2-3,1-2)Gal or Neu5Ac(a2-3,1-4)Gal. In a polysialic acid chain of a-2,8-linked NeuSAc residues the a-2,8/ 1,9-lactonization can be effected under relatively mild conditions, like mild acid treatment [200], yielding a water-insoluble polymer. The NeuSAc(a2-9,1-8)Neu5Ac formation in a polysialic acid chain can only be realized by carbodiimide treatment [20 11, illustrating a more difficult condensation with the secondary H 0 8 group. The difference in reactivity between the primary and the secondary OH group has been nicely illustrated for the alternating a-2,8/a-2,9polysialic acid of E. coli K92 (Table 5) [202]. For gangliosides, it has been stated that ganglioside lactones occur also as such in nature[206], and that in this way the negative charge of a ganglioside under physiological conditions can be modulated. Lactonization makes the oligosaccharide chain also more rigid, which may have important biological implications. Treatment with carbodiimide or glacial acetic acid can even convert a Neu5Ac(a2-8)NeuSAc(a2-3)Gal(~ 1- into a NeuSAc(a2-8,1-9)NeuSAc(a23,1-2)Gal(Pl- sequence [204,205]. Over the years NeuSAc has been prepared by a variety of methods. Several biological sources have been explored to isolate this sialic acid in high amounts. Among them are edible birds nest substance [207], urine of sialuria patients [65], colominic acid {Neu5Ac(a2-8)}, produced by E. coli strains [208], and hen’s egg chalaza, egg-yolk membranes and delipidated egg yolk [209,210]. The large scale organic synthesis of Neu5Ac is still complicated (see section 6.1). However, recent biotechnological routes, using sialate-pyruvate lyase (aldolase; see section 9.5), readily yield large amounts of

Table 5 Survey of structures or elements of sialic-acid-containingmicrobial polysaccharides Structwdelement

Species

Reference(s)

Neisseria meningitidis B Escherichia coli K235

~41 ~561

Capsular polysacharides (CPS)

{ -8)Neu5Ac(a2-8)Neu5Ac(a2-),

{ -8)Ne~SAc(a2-8)NeuSAc(a2-}~ (with7Ac or 9Ac) {-9)Ne~5Ac(a2-9)Neu5Ac(a2-}~ (with7Ac or 8Ac) {-8)Ne~5Ac(a2-9)Neu5Ac(a2-}~ {-4)Neu5Ac(a24)Gal(aI -},, (-4)Neu5Ac(a24)Glc(al-},, (withAc) {-3)GalNAc(fi14)Gal(a14)Neu5,9Ac2(a2-3)Gal(fil -}" { -4)Glc(fi14)[NeuSAc(a2-3)Gal(fi1-4)GlcNAc(fi1 -3)]Gal(fi1 - } n { -4)Glc(fiI 4)meu5Ac(a2-3)Gal(fi 1-3)GlcNAc(fi1 -3)]Gal(fi1-},, { -4)GlcNAc(fil-3)[Gal(fil4)]Gal(fil4)Glc(fil-3)Glc(fil-2)[Neu5Ac(a2-3)]Gal(fi 1 -},, {-4)Glc(fil-6)[Neu5Ac(a2-3)Gal(fil4)]GlcNAc(fi 1-3)Gal(fiI-}" { -4)Glc(a 1 4 ) [NeuSAc(a2-3)Gal(fi14)GlcNAc(fi14)]Gal(fi 1-4)Glc(fi1 -}" {-4)Glc(fi14)[Neu5Ac(a2-3)Gal(fil4)GlcNAc(fil4)]Glc(a1-4)[Glc(fil-3)]Gal(fi1-}" { -4)Glc(fil4)[Neu5Ac(a2-3)Gal(~l-3)]Glc(fil-3)Gal(fi1-}n { -4)Glc(fi1-4)[Neu5Ac(a2-3)Gal(fi 1-4)GlcNAc(fi14)]Glc(a 14)Gal(fi1-}n { -3)Glc(fi1-2)[Kdn(a24)][3Ac]GlcA(a1-3)Man(a1-3)Glc(a 1-},,

Pasteurella haemolytica A-2

[1571

Morarella nonliquefaciens Escherichia coli KI

[1581 [24,159]

Neisseria meningitidis C

~41

Escherichia coli K92

[i601

Neisseria meningitidis W 135

~41 ~41 [26,161]

Neisseria meningitidis Y Escherichia coli K9

Group B Streptococcus la Group B Streptococcus Ib Group B Streptococcus 11 Group B Streptococcus I11 Group B Streptococcus IV Group B Streptococcus V Group B Streptococcus VI Group B Streptococcus VII Klebsiella ozaenae K4 (2211)

Lipopolysaccharides (LPS)

{ -3)GalNAc(fi1-7)Neu5Ac(a2-3)[Glc(a1-2)]Glc(fiI-}" { -3)GlcNAc(fi1-7)Neu5Ac(a2-3)[Gal(al-2)]Glc(fiI -}" {-3)GalNAc(fi14)Gal(a14)Neu5,7/8,9Ac3(a2-3)Gal(fil -},, (also with 9Ac)

Escherichia coli 024

[ 166,1671

Escherichia coli 056

[ 166,1671

Escherichia coli 0104

[26,271 continued on next page

Table 5, continued ~

~~

Structurelelement

Species

Reference(s)

{-3)[+6Ac]GlcNAc(fi 1-7)Neu5Ac(a2-3)[CH3(NH)C-2]FucN(a1 -},,

Salmonella arizonae 0 2 1

El681

{ -4GalANGro(al-4)Neu5Ac(a2-3)GalANGro(fi1-3)QuiNAc(fi 1- } n { -3)[Glc(al4)]GalNAc(@14)Neu5Ac(a24)Glc(alM)[Glc(a 1-4)Gal(fi1L6)Gl~((3l-3)]Gal((31-}~

Vibrio cholerae H11 (non-01) Hafnia aluei 0 2

~ 9 t1701

Neu5Ac(a2-3)Gal(@14)GlcNAc(fi 1-3)Gal(@I4)Glc((314)Hep-

Neisseria meningitidis

[ 171-1731

Neu5Ac(a2-3/6)Gal(@ l4)GlcNAc(@l-3)Gal(fi1-4)Hep-

Haemophilus ducreyi Campylobacter jejuni 0 1

[ 1741

NeuSAc(a2-3)[GalNAc(@1-4)]Gal(@ 1-3)[Gal(al-Z)]Gal(fi1-3)Hep-

1

Core regions of lipopolysaccharides (LOS) a

[I751 ~761

NeuSAc(a2-3)Gal(fi 1-3)[Gal(a 1-2)]Gal(fi 1-3)Hep-

Campylobacter jejuni 0 2

Neu5Ac(a2-3)Gal(fil-3)GalNAc(fil-4)pJeuSAc(a2-3)]Gal(fi 1-3)HepNeu5Ac(a2-3)[GalNAc(fil-4)]Gal(fil4)Glc((31-2)Hep-

Campylobacter jejuni 0 4 Campylobacter jejuni 0 2 3 and 0 3 6

[I751 [I751

Gal(a 14)Glc((3l-7)Neu5Ac

Rhodobacter capsulatus 37b4

[I771

a Some other species for which it has been demonstrated that their LPS or LOS contains sialic acid are: Rhizobium meliloti MI IS (LPS, Neu5Ac/Neu5,9Ac2)[178]; Salmonella toucra (LPS, terminal and 4-linked NeuSAc) [179]; Salmonella djakarta (LPS, terminal NeuSAc) [I 801; Salmonella isaszeg (LPS, terminal and 4-linked Neu5Ac) [ 1801; Citrobacter freundii 0 3 7 (LPS, terminal and 7-linked NeuSAc) [ 1791; Pasteurella haemolytica 1 and 5 (LPS, Neu5Ac) [181]; Haemophilus injruenzae (LOS, terminal NeuSAc) [182,183]; Neisseria gonorrhoeae (LOS, terminal Neu5Ac) [184,185]. Note that for E. coli 0 2 4 and 0 5 6 and H. aluei 0 2 also terminal Neu5Ac has been found [179]. GalANGro, N-galacturonyl-2-aminoglycerol.

262

C //O /I

0 AcHN

Fig. 4. Lactonization of a NeuSAc(a2-8)NeuSAc(a2sequence, affording a Neu5Ac(a2-8,1-9)Neu5Ac(a2fragment.

this sialic acid (see section 6.1). Efficient procedures to prepare Neu5Gc from colominic acid via de-N-acetylation, N-acryloylation and reductive ozonolysis, followed by acid or enzymatic hydrolysis of the formed (NeuSGc(a2-X)}, has appeared in refs. [211,212]. Also porcine submandibular gland is a good source for the large scale preparation of NeuSGc, whereas Neu5,9Ac2 can be prepared from bovine submandibular gland [2 13). Besides the occurrence of sialic acids [(0-acetylated) Neu5Ac and Kdn] in microbial polysaccharides (Table 5), some lipopolysaccharides have shown to contain sialic-acidlike monosaccharides. They differ from sialic acids in the presence of an additional amino function at C7, a deoxy function at C9, and in the configuration at the chiral centers: 5,7-diamino-3,5,7,9-tetradeoxy-~-glycero-~-manno-non-2-ulosonic acid and 5,7-diamino3,5,7,9-tetradeoxy-~-g~cero-~-ga~acto-non-2-u~osonic acid [2 14,214al. So far, the two amino groups have been found to be substituted in different combinations, yielding acetamido, formamido, (R)-3-hydroxybutyramido,4-hydroxybutyramido, or acetamidino functions. Also 0-acetylation can occur, whereas the (R)-3-hydroxybutyryl group at C7 can be used to link monosaccharides in a polysaccharide chain. Typical species are Pseudomonas aeruginosa strains [214], Shigella boydii type 7 [214], Vibrio salmonicida [214a], Vibrio cholerae 0 2 [215], Vibrio alginolyticus strain 945-80 [216], Salmonella arizonae 0 6 1 [2 171, Yersinia ruckeri 0 1 [2 181, Legionella pneumophila strain 1 [219,220], and PseudomonasJluorescens ATCC 49271 [220a]. For a review, see ref. [220], but it should be noted that the absolute configuration of 5,7-diamino-3,5,7,9tetradeoxy-~-glycero-~-ga~acto-non-2-u~osonic acid was earlier assigned as D-glycero-Lgalacto- [2 14,2 14a,216-2 19,220al.

4. Screening of biological materials for the presence of sialic acid For the staining of sialic acids in tissues and cells, a great variety of techniques is available. Classical histochemistry of sialic-acid-containing glycoconjugates makes use of either binding of cationized dyes (e.g. Alcian blue, cationized ferritin, ruthenium red) or selective periodate oxidation (derivatization of generated aldehyde groups of sialic acids with e.g. p-dimethylaminobenzaldehyde,dansylhydrazine, rhodamine, biotidferritin-conjugated avidin). The second approach is strongly dependent on sidechain modifications [53]. Comprehensive reviews on this subject have appeared [6,22 11.

263

Nowadays, specific lectins are frequently used to detect the presence of glycosidically bound sialic acids in complex carbohydrates. However, the specificity of lectins is generally broad, and positive information has always to be checked in control experiments using e.g. sialidases in the presence and absence of inhibitors. For the histochemical analysis, lectins may be conjugated e.g. with gold particles, peroxidase, rhodamine or fluorescein isothiocyanate [6]. A large series of lectins which recognize sialic acid have been demonstrated to occur in nature, and most of their biological sources have been summarized in refs. [6,11,222]; see also refs. [223-2271. For updated reviews focusing on lectins, see the present volume 29b of the New Comprehensive Biochemistry series. Generally, lectins have been isolated from lower animals such as prawns, snails, crabs, spiders, scorpions, lobsters, slugs, oysters, but also from plants, rat brain and B cells. While some of these lectins bind to both NeuSAc and NeuSGc, others are specific for NeuSAc. 0-Acetylation may also influence the binding strength of the lectin, both in a positive and in a negative manner (see section 10.3). For introductory glycoprotein analysis, in answering questions like “what monosaccharides are in the glycoprotein glycans?”, two sialic-acid-specific plant lectins, having also a glycosidic linkage specificity, have been included in commercially available kits. These lectins are the agglutinins from Muuckia umurensis and Sumbucus nigra, being diagnostic for NeuSAc(a2-3)Gal and NeuSAc(a2-6)Gal/GalNAc elements, respectively. For screening purposes, the lectins are labelled with digoxigenin-succinyl-E-amidocaproic acid hydrazide (DIG, a spacer-linked steroid hapten digoxigenin). After SDS/PAGE of the (g1yco)protein mixture and Western blotting, sialoglycoprotein bands with a-2,3and/or a-2,6-linked NeuSAc can be labelled by one or both of these DIG-labelled lectins, whereby the detection is carried out in an enzyme immuno-assay using a digoxigeninspecific antibody conjugated to alkaline phosphatase, followed by color development with 5-bromo-4-chloro-3-indolyl-phosphate/4-nitroblue tetrazolium chloride [228,229]. Viruses and antibodies can also be used for the detection of sialic acid in complex carbohydrate systems. The sialic-acid-binding properties of a number of viruses have been established, and it has been shown that influenza C virus, bovine corona virus and encephalomyelitis virus recognize 9-0-acetylated sialic acids (see also sections 8.4.2, 9.1 and 10.3). Microtiter-plate and nitrocellulose-membrane assays have been developed that use the hemagglutinin (receptor) and the receptor destroying activities (sialate 9-0acetylesterase) of the influenza C virus to specifically detect bound 9-0-acetylated sialic acids in glycoproteins [230,23 I]. Although the recognition site for Neu5,9Ac2 and the esterase activity are located on the same viral glycoprotein, these activities can be separated using different temperatures for the binding (4°C) and the enzyme reaction (4-methylumbelliferyl acetate or a-naphtyl acetate as substrates; 20-37°C). Other approaches are based on binding of the virus to immobilized ligands, and detection of the virus with monoclonal antibodies, whereby the esterase was selectively inactivated by the use of diisopropylfluorophosphate (DFP) [232]. Virus particles have also been labelled with radioactive isotopes [232] or biotin [233]. The application of the assay with virus particles for the staining of 0-acetylated sialic acids in tissue sections will be described in section 8.4.2. Recently, a new technique using a soluble chimera of the hemagglutininesterase portion of the hemagglutinin-esterase-fusion-glycoprotein from influenza C virus and the Fc portion of human IgG, has been reported [234,235]. Such a chimera retains

264

both its recognition and enzymatic functions, and also has the binding properties of the Fc portion of IgG. The probing can be carried out on blots and TLC plates taking into account precautions for the recognition and esterase activities (this has to be inhibited by DFP in the test) as discussed above. An interesting electrochemical method for the determination of bound sialic acid has been developed, making use of a potentiometric four-channel thick-film sensor [236]. The sialidase sensor consists of a bilayer of a membrane containing Clostridium perfringens sialidase immobilized in a poly(viny1 acetate)-polyethylene copolymer, which is placed on top of an H+-selective poly(viny1 chloride)-poly(viny1 acetate) indicator membrane. The enzyme-induced release of bound sialic acid leads to a concomitant decrease in pK, of the carboxyl function of sialic acid. This decrease affords a local pH change inside the sialidase-containing sensor membrane, which is monitored by the H+-selective indicator membrane. The pH optimum of the sialidase sensor was pH 4 for sialyllactose, mucin and colominic acid. Finally, TLC analysis may also be of great help in the screening of biological materials for the presence of sialic acids. This item will be discussed in section 5.3.1.

5. Isolation and analysis of sialic acids 5.1. Liberation

The characterization of the type of sialic acid in sialoglycoconjugates is frequently carried out after release and (partial) isolation. The cleavage of sialic acid from sialic-acidcontaining material is mainly performed by two methods, namely, acid hydrolysis and enzymatic hydrolysis. Both approaches have advantages and disadvantages. In the case of acid hydrolysis, problems arise with respect to de-0-esterification, which complicates quantitative analysis procedures. With respect to the enzymatic hydrolysis with sialidases, linkage specificity as well as a reduced or complete lack of susceptibility have to be taken into account. Moreover, in most cases much lower amounts of sialic acids are released by sialidases than by acid hydrolysis, which may be due to the different accessibility of the sialic acids in the biomolecules to be analyzed. Additionally, in work-up procedures and analyses, pH values below 4 and over 6 should be avoided to prevent migration of the 0-acetyl group at C7/C8 and hydrolysis of 0-acetyl groups as much as possible [23,57]. Several approaches have been reported for the effective acid hydrolysis of the labile a-2,3-, a-2,6- and a-2,8-linkages. All these procedures suffer from being not optimal in giving the real spectrum of sialic acids originally present in the sialoglycoconjugate under study, especially in the case of a mixture of (0-acylated) N-acylneuraminic acids. Terminal sialic acids are released in high yield (and low destruction) using a two-step acid hydrolysis procedure comprising treatment with formic acid (pH 2, 1 h, 70°C), followed by HCl (pH 1, 1 h, SOOC). After each step, the liberated sialic acids must be recovered by centrifugation, ultrafiltration or dialysis [I 1,12,213,237]. It should be noted that in the case of a spectrum of (0-acylated) N-acylneuraminic acids, the supernatant, ultrafiltrate or diffusate of the formic acid hydrolysis contains the majority of the 0-acylated N-acylneuraminic acid, whereas that of the HCl hydrolysis contains

265 Table 6 Substrate specificity of commercially available sialidases a Source of sialidase

PH optimum

a-2,3

Arthrobacter ureafaciens Vibrio cholerae

4.4-5.5' 5.0-6.5g

100

100

Clostridium perfingens Salmonella typhimurium

5.5-7.2'

100

5.0-7.01

100

Newcastle disease virus

5.0-6.0

100

Relative rates of cleavage are indicated (1 00 =full activity). Neu5Ac(a2-3)lactose. NeuSAc(a2-6)lactose. a

Neu5Ac(a2-8)Neu5Ac(a2-3)lactose. NeuSGc(a2-3)lactose. Sialyllactose, pH 5.0-5.5, and colominic acid, pH 4.34.5. g Sialyllactose, pH 5.0-5.5.

'

a-2,6C

a-2,Sd

a-2,3e

167 53

53 31

12 25

44

44

20

PI PI

0.4

n.d.k.l

n.d.

[247]

0.2

78

11

Reference

151

r51

High-molecular-mass isoenzyme (P. Roggentin, personal communication) (see also section 9.2). ' Sialyllactose, pH 5.0-5.1 (acetate buffer) and pH 5.8-6.0 (phosphate buffer). Depending on the buffer system used. n.d., not determined. { -8)Ne~5Ac(a2-8)NeuSAc(a2-}~, 0.1% and {-9)Neu5Ac(a2-9)Neu5Ac(a2-),, 0.08%.

J

' '

mostly Neu5Ac and Neu5Gc. In the case of low-molecular-mass substances, isolations can be carried out by gel-permeation chromatography. Although these conditions do not lead to significant de-N-acylation, de-0-acylation has been shown to occur to an extent of about 30-50%. One has to consider that milder acidic conditions result in incomplete release of sialic acids. Acid hydrolysis with acetic acid (2M, 3 h, SOOC) as suggested in ref. [238] did not improve the yield of 0-acylated N-acylneuraminic acids [ l 11. When focusing on sialic acid analysis, the use of H2S04 (0.05 M) is not recommended because of work-up problems. In connection with the HPLC analysis of Neu5Ac and NeuSGc, microwave hydrolysis in 2 M acetic acid has shown to be an interesting alternative [239]. In the methanolysis procedure, as used for the standard quantitative monosaccharide GLC analysis of glycoconjugates [240,241], 1 M methanolic HCl (24 h, 85OC) is applied. However, under these conditions released sialic acids are completely de-N, 0-acylated, which makes this approach unsuitable for the characterization of different types of sialic acid. It is, however, a reliable approach for the determination of the total amount of a mixture of (0-acetylated) N-acylneuraminic acids. When using a milder methanolysis procedure (0.05 M methanolic HCl, 1 h, 80°C) the de-N-acylation but not the de-0-acylation is strongly reduced [242]. In the quantitative determination of N-acylneuraminic acids in poly-N-acylneuraminicacid-containing glycoproteins, the release of free sialic acid was shown to be optimal using a combined mild acid hydrolysis (0.1 M TFA, 3 h, 8O0C)/subsequent mild methanolysis (0.05 M methanolic HCl, 1 h, SOOC) [243]. This method is also advised for the analysis of the Kdn content in Kdn-containing glycoproteins. In contrast, optimal release of Kdn from poly-Kdn-containing glycoproteins is obtained by mild methanolysis with a longer incubation time [243]. In the latter case the standard conditions of methanolysis also give good results.

266

The enzymatic release of N-acylneuraminic acids can be carried out under such mild conditions (low temperature, pH 5 4 ) , that destruction, migration, and de-0-acylation is kept at a minimum. In Table 6 the substrate specificity of commercially available sialidases from Arthrobacter ureafaciens, Yibrio cholerae, Clostridium perfringens, Salmonella typhimurium, and Newcastle disease virus, using simple N-acylneuraminyllactose substrates is compared (for comprehensive reviews on sialidases, see refs. [5,33,244-2461). It is evident that the sialidases show different ratios of cleavage rates for the a-2,3-, a-2,6- and a-2,8-linkages. The finding of a strong preference of the Newcastle disease virus sialidase for a-2,3-linkages holds also for other viral sialidases, such as those from fowl plague virus and influenza A2 virus. The latter enzyme also has a low specificity for a-2,g-linkages [5]. Among the bacterial sialidases, A. ureafaciens sialidase has a certain preference for a-2,6linkages. The S. typhimurium sialidase is the only bacterial sialidase with a viral sialidase-like pronounced preference for a-2,3-linkages [247]. Recently, two sialidases from Bacteroides fragilis having a higher preference for the cleavage of a-2,8linkages, when compared with a-2,3- and a-2,6-linkages, were isolated [248]. The sialidase, recently isolated from Macrobdella leech, cleaves only NeuSAc(a2-3)Gal linkages [61]. In general, Neu5Ac residues are released faster than NeuSGc residues. In a study using different N- and 0-glycoproteins with a-2,3- and/or a-2,6-linked N-acylneuraminic acids as substrates [antifreeze glycoprotein, ovine submandibular gland glycoprotein, a I-acid glycoprotein; Neu5Ac(a2-6)GalNAc(a l-O)Thr/Ser, NeuSAc/SGc(a26)[Gal(b 1-3)]GalNAc(a 1-O)Thr, NeuSAc/SGc(a2-3)Gal(fi 1-3)GalNAc(a 1-O)Thr, Neu5Ac(a2-3)Gal(B 14)GlcNAc(b 1-0)R; Neu5Ac/5Gc(a2-6)Gal(~l-4)GlcNAc(fi 1-O)R] and sialidases from A. ureafaciens, V cholerae, C. perfringens, Newcastle disease virus, fowl plague virus, and influenza A2 virus, roughly similar patterns of substrate specificity as for sialyllactoses were found. However, it was demonstrated that the core oligosaccharide andor the protein structure may also influence the rate of release for different glycosidic linkages[249]. In the case of S. typhimurium sialidase, also N-glycoprotein a-2,3-sialoglycans were susceptible to efficient cleavage, but not mucin 0-glycoprotein sialoglycans [247]. The most recently discovered sialic acid, NeuSGcAc, could not be released with V cholerae sialidaseE431. A comparison of the different commercially available sialidases shows that the C. perfringens sialidase iso-enzyme with a molecular mass of about 63 kDa has the broadest specificity. It should be noted that C. perfringens in fact produces two sialidases, the larger of which (63kDa) is commercially available (P. Roggentin, personal communication) (see also section 9.2). As described for the acid hydrolysis procedure, the work-up of enzymatically released sialic acids can be achieved employing various methods, depending on the starting sialoglycoconjugate material. 4-0-Acetylated neuraminic acids in any glycoconjugate are resistant to most sialidases tested so far; only viral sialidases show a low but significant activity. With the exception of Streptococcus sanguis sialidase [250], the presence of 0-acetyl substituents at C7-C9 leads to a reduced rate of cleavage by all sialidases, so that prolonged incubations are necessary for an efficient release [251]. More information has been collected in a detailed study with bacterial and viral sialidases and 4-methylumbelliferyl (MU) a-glycosides of 4-, 7-, and 9-0-acetylated N-acetylneuraminic acids as substrates [252]. In contrast to

267

the other sialidases tested, the fowl plague virus sialidase catalyzes a slow release of Neu4,5Ac2 from a-Neu4,5Ac2-MU. The recent finding of {Neu5Gc(a2-05)}, elements in a glycoprotein [ l o l l initiated a kinetic study of the enzymatic and non-enzymatic hydrolysis of Neu5Gc(a2-05)Neu5Gc and Neu5Gc(a2-8)NeuSGc [253]. It turned out that at pH < 3.8 the rate of acid hydrolysis of the unusual a-2,05-linkage was greater than that of the normal a-2,8-linkage. However, at pH > 3.8 reverse results were obtained; NeuSGc(a2-8)NeuSGc released a small but detectable amount of Neu5Gc even at pH 6. The a-2,05-linkage was only partially cleaved by C. perfringens and! l cholerae sialidases, and was essentially resistant to A . ureafaciens sialidase. The detection of sialidases that can release Kdn is so far highly limited. The liver of the loach Misgurnus fossilis was found to contain a sialidase capable of releasing both Neu5Ac and Kdn from sialoglycoconjugates [39]. While the sialidases investigated so far in detail require an NH-acyl group at C5 for full activity, the loach enzyme can handle both NH-acyl and OH functions at C5. The rainbow trout also turned out to be a useful source for the isolation of a sialidase, active in releasing both Neu5Ac and Kdn [254]. Recently, a sialidase was isolated from Sphingobacterium multiuorum that specifically released a-2,3-, a-2,6- and a-2,8-linked Kdn; Neu5Ac and Neu5Gc were not liberated [255]. Before fractionation, pools of free sialic acids can be freed from contaminants by several methods, including ion-exchange chromatography, cellulose chromatography, reversed-phase chromatography, preparative TLC [6,11,121. Generally, one of the purification procedures for the pool of sialic acids comprises Dowex ion-exchange chromatography at low temperature. After passage through a cation-exchange resin (Dowex SOW-XS, H+-form), the pool of sialic acids is adsorbed to an anion-exchange resin (Dowex 2-X8 or 1-X8, HCOO--form). Elution from the anion-exchange resin is generally carried out with &2 M formic acid. The ion-exchange chromatography should be carried out rapidly, as prolonged contact of acylneuraminic acids with the resin or the solvent systems used may result in degradation, 0-acetyl migration and/or de-0-acetylation. After rotary evaporation or lyophilization, the sialic acid pools are stored at -20°C. In general, mixtures of sialo-oligosaccharides from {Neu5Ac(a2-8)},, (NeuSAc(a2{Neu5Gc(a2-8)},, and (Kdn(a2-8)},, are 9)},, {-8)Neu5Ac(a2-9)Neu5Ac(a2-),, generated by limited acid hydrolysis [211,256,257]. Depending on the polysialic acid, different conditions have been applied. Also attention has been paid to the intramolecular self-cleavage of polysialic acids such as {Neu5Ac(a2-8)}, [258]. Adjacent COOH groups with a high pK, (3.9-5.5) act as proton donors for general acid catalysis. The lability is seen under mild acidic conditions, that can be encountered in various physiological situations. {Neu5Ac(a2-8)},3 is substantially more unstable than {Neu5Ac(a2-8)}2-3. A highly useful enzyme for the depolymerization of polysialyl carbohydrate chains, yielding oligosialyl compounds, is endo-sialidase (endo-N) produced by infection of E. coli K1 with a lytic bacteriophage [9,259,260]. The enzyme is specific in cleaving a-2,8-linkages, and requires at least five Neu5Ac or Neu5Gc residues for activity. A limited digestion of {NeuSAc(a2-8)},-R affords mainly {Neu5Ac(a2-8)}4, with some {Neu5Ac(a2-8)} 1-3. Alternating a-2,8/a-2,9-polysialyl chains, as present in some

268

bacterial polysaccharides, are also cleaved, but a-2,9-linked polysialyl chains are resistant. Poly-Kdn and {Neu5Gc(a2-05)}, are not substrates for endo-N. A similar endo-sialidase associated with phage particles, namely, E. coli bacteriophage $92, has been reviewed in ref. [261]. 5.2. Colorimetry One of the oldest methods used to detect and to quantitate sialic acids is colorimetry [3,6,12,237,262]. When carried out on non-purified samples, the influence of contaminants interfering with the assays has to be taken into account. Greatest problems are encountered when using cells or tissue extracts, as the level of contamination is inevitably high. Moreover, factors such as non-identical reactions of different sialic acids in the same assay and the non-specificity of the reactions for the sialic acids are important. Although several colorimetric methods have been developed in the past, only two main procedures are currently routinely applied, namely the orcinol/Fe3+/HC1assay, known as the “Bial” reaction, and the periodic acidhhiobarbituric acid assay. For microadaptations of these two different tests, see ref. [l 11. In the first assay, the sample is mixed with orcinol, FeC13 and concentrated HCl and heated at 96°C. The formed purple to red-violet chromophore is extracted with isoamyl alcohol and its absorbance measured at 572 nm. Because of the use of HCl, the method can be used to quantitate the total amount of both free and glycosidically bound sialic acids. Due to the strongly acidic conditions, ester groups are released. As the assay does not discriminate between bound and free sialic acids, it is widely used to monitor the presence of sialic acids in either form during fractionation of biological material. It should be noted that other monosaccharides, especially pentoses, hexoses and uronic acids interfere with the assay, which is of importance when small amounts of sialic acid are present. In the second assay, only suitable to quantitate free sialic acids, sialic acid is oxidized by periodate at 37°C under strongly acidic conditions. The oxidation leads to the formation of a prechromogen, a six carbon aldehyde, which then yields the chromogen fi-formyl pyruvic acid by aldol cleavage between C4 and C5. The chromogen reacts with thiobarbituric acid to give a red chromophore, the absorbance of which is measured at 549 and 532nm. In principle two approaches can be followed, called the Warren method and the Aminoff method. Differences between the methods lie in the acidity of the initial periodate oxidation and in the solvent used for the extraction of the pigment (cyclohexanone, Warren; acidic 1-butanol, Aminoff). It should be noted that substituents in the glycerol side chain severely influence the periodate oxidation. Therefore, in the case of ester substituents, a prior saponification is necessary (0.1 M NaOH, 37°C). Types of free sialic acid which do not yield the chromogen are negative in this test. Several compounds have been shown to interfere with the periodate/thiobarbituric acid assay, most especially 2-deoxyribose, 2-keto-3-deoxyaldonic acids other than Kdn, disaccharides such as lactose and maltose, and unsaturated fatty acids. The greatest errors arise in the quantitation of sialic acids from cellular extracts or homogenates containing membrane and nucleic acid material. Therefore, prior ion-exchange column chromatography and removal of lipids

269

by ether extraction are of advantage. Special attention to the periodic acidthiobarbituric acid assay of Kdn has been paid in ref. [243]. When both tests are used in combination, a differentiation between total and free sialic acid is possible, allowing the calculation of the amount of glycosidically bound sialic acid. In a new approach for the direct determination of free and bound sialic acid, an acidic ninhydrin assay has been proposed [263]. Heating of solutions of sialic-acid-containing material with ninhydridacetic acid/37% HCl at 100°C yields a stable chromophore, the absorbance of which can be measured at 470nm. In view of the comments made in section 5.1 with respect to the release of sialic acid by sialidases, a quantification procedure for bound sialic acid based on the enzymatic analysis of pyruvate, formed after sialidase/aldolase treatment [6,264], should be handled with care. Of the various fluorimetric assays available for sialic acid analysis, the method which allows the discrimination between sialic acids with or without 0-acyl groups at C8 and/or C9 may be of interest [6,265]. After mild periodate oxidation, formaldehyde, derived from C9 in the case of non-substituted H 0 9 and H 0 8 is derivatized with acetylacetone in the presence of ammonium acetate, leading to the fluorigen 3,Sdiacetyl1,4-dihydr0-2,6-dimethylpyridine (4 10 nm excitation, 5 10nm emission). It is evident that all contaminants producing formaldehyde under the influence of periodate will interfere with this sialic acid analysis. Finally, for the quantitative estimation of the 0-acyl content of sialic-acid-containing samples (Hestrin assay), also a colorimetric assay is routinely used. The method is based on the reaction with hydroxylamine in alkaline medium yielding hydroxamates, which form with Fe(C104)3 red chromophores, the absorbance of which is measured at 520 nm [262].

5.3. Chromatography 5.3.1. Thin-layer chromatography

From the beginning of free sialic acid analysis, TLC has played a major role in screening and tentative assignment procedures, and over the years several solvent systems for both cellulose and silicagel plates have been reported [6,237,262]. One of the most popular TLC methods for the analysis of (substituted) N-acylneuraminic acids comprises the use of plastic HPTLC plates precoated with cellulose and 1-propanol/l-butanol/O. 1 M HCl (2:1:1, v/v/v) as solvent system. It shows the best and most reproducible separation of different sialic acids, and is less sensitive to interfering substances when compared with other systems. Generally, the visualization of the different sialic acids is carried out by spraying with the orcinol/Fe3+/HC1reagent [237], yielding typical purple bands. For quantitative purposes densitometry is also used. It should be noted that due to differences in cellulose quality (impurities), even after pre-washing of the plates, the reproducibility of Rf values is relatively low. Therefore, analyses should be carried out in the presence of an appropriate reference sialic acid mixture on a separate lane. To give an impression of the separation capacity of cellulose plates, Table 7 summarizes the Rf values of a series

270 Table 7 Thin-layer chromatographic migration rates (Rf) of sialic acids on 0.1 mm cellulose plates using 1-propanol/lbutanol/O.l M HCI (2:1:1, v/v/v) [6] Compound Neu5Ac Neu4,5Ac2 Neu5,7Ac2 Neu5,9Ac2 Neu5,7,9Ac3 Neu5,8,9Ac3 Neu5,7,8,9Ac4 Neu5Ac9Lt

Rf

0.45

Compound

0.60 0.54 0.63 0.70 0.75

NeuZen5Ac Neu5Gc Neu4Ac5Gc Neu9Ac5Gc Neu7,9Ac25Gc Neu5GcAc

0.80

NeuZenSGc

Rf 0.55 0.35 0.65 0.55 0.70 0.51 0.45

0.56

of N, 0-acyl-neuraminic acids from one experiment [6]. In the case of radio-labelled sialic acids, the bands can also be traced by radio-TLC-scanning. Application of the TLC method in a two-dimensional procedure with intermediate ammonia treatment gives information about the type of N-acylneuraminic acid of the constituting NO-acylneuraminic acids. In this way a differentiation is possible for example between co-migrating Neu9AcSGc and Neu5,7Ac2 [6]. In principle, de-0-acetylations can also be carried out by sialate 0-acetylesterases [266]. For the analysis of oligomers of a-2,g-linked NeuSAc, NeuSGc or Kdn, a TLC procedure on silicagel with the solvent system 1-propanol/25% ammonia/water (12:2:5, v/v/v) has shown good results [267]. In this way, mixtures of {Neu5Ac(a2-8)}2 to {Ne~SAc(a2-8)}~4,{Ne~5Gc(a2-8)}~to {Neu5G~(a2-8)}~~, or (Kdn(a2-8)}2 to {Kdn(a2-8)}7 are well separated. It should be noted that (NeuSAc(a2-X)}, and {Neu5Gc(a2-8)}, can be visualized also with a resorcinol spray reagent, whereas for { Kdn(a2-8)}, the orcinol spray reagent is needed. 5.3.2. High-performance liquid chromatography Initially, the separation of sialic acids was mainly carried out by cellulose chromatography at low temperature [6,237]. However, nowadays HPLC fractionations using different column materials, elution protocols and detection techniques have replaced this approach [6,11,268]. The application of HPLC has also introduced a rapid method for tentative assignments of sialic acids in complex mixtures, based on elution times of known standards, being more reliable when more than one HPLC procedure is followed. Moreover, a rapid method for quantification of released sialic acids has become available. Due to the relatively short HPLC runs, also fast transitions between members of the sialic acid family due to migration of substituents, introduction of substituents, cleavage of substituents, or other (enzymatic) modifications can easily be monitored. First detailed reports on the separation of non-derivatized sialic acids deal with the application of Aminex A-28 or A-29 anion-exchange chromatography using 0.75 mM [269, 2701 or 0.5 mM [268] Na2S04 as eluting system and UV monitoring at 200 nm (nanomole range). In a different approach, fluorigenic derivatives of sialic acids, prepared by reaction

27 1

with 1,2-diamin0-4,5-methylenedioxybenzene(DMB) in the presence of 2-mercaptoethanol and sodium hydrogensulfite, have been separated by C reversed-phase HPLC [27 1-2731, using acetonitrile/methanol/water (9:7:84, v/v/v) as solvent system and fluorescence monitoring at 373 nm excitation and 448 nm emission wavelengths. An appropriate cutoff filter may be used instead. The fluorescence labelling makes a relatively specific and highly sensitive (femto- to picomole range) detection possible. However, using radiolabelled sialic acids, it was found that the derivatization reaction is not quantitative [268]. For an adapted protocol, see ref. [l 11. Another interesting approach is the conversion of NeuSAc/SGc into chemiluminescent quinoxalinone derivatives using 4,s-diaminophthalhydrazide dihydrochloride (a-keto acid derivatization) [274]. These derivatives are analyzed by reversed-phase HPLC (femtomole range), whereby the chemiluminescence detection follows the reaction of the derivatives with hydrogen peroxide in the presence of potassium hexacyanoferrate(II1) in alkaline solution. Other conversions of NeuSAc, useful for HPLC separations, include the derivatization with 4,4'-dicarboxy-2,2'-biquinoline [275], 2-cyanoacetamide [276], periodatethiobarbituric acid [277], benzoic anhydride [278], 4'-hydrazino-2-stilbazole [279] and 1,2-diamino-4,5-dimethoxybenzene[280]. Taking advantage of the separation capacity of the anion-exchange resin CarboPac PA- l, fractionation of non-derivatized sialic acids at neutral pH, using sodium acetate as eluent and pulsed amperometric detection (PAD) following postcolumn addition of alkali, has shown excellent results in terms of the wide array of sialic acids that can be separated, the sensitivity of the detection method (picomole range), and the relative ease of use for preparative work (without PAD detection) [268]. The general problem of quantification in PAD analyses, due to differences in detector response attributed to differences in the number of free hydroxyl groups of the various components separated, holds also for sialic acids. So far, only for a limited number of sialic acids relative detector response factors have been calculated (e.g. NeuSAc, 30 500; Neu5,9Ac2, 14 500; NeuSGc, 35 400). It is important to note that pH values > 11, as usually applied in CarboPac-PAD analyses of oligosaccharides (for such a NeuSAc/NeuSGc separation, see ref. [28 l]), will lead to rapid de-U-acylation of the 0-acylated sialic acids [282]. This phenomenon can even occur between the point of postcolumn alkali addition and the entry into the PAD detector [268]. In an evaluation of five different HPLC methods, it turned out that no single method is adequate to completely separate and quantitate complex mixtures of sialic acids [268], and the use of multi-dimensional HPLC is advised. As a clear illustration of this statement, Table 8 is included. This evaluation also compares a series of essential features of the five HPLC methods, namely, sensitivity, specificity of detection, separation by number of hydroxyl groups or substituents, separation of isomers, preparative use, avoiding of ester migration during purification, and avoiding of ester loss during purification. In this comparison [268,270,271,283,284] the HPLC systems I (CarboPac PA-I) and 111 (TSK-ODS 120T, DMB derivatives) gave the highest averaged scores in terms of applicability. A major advantage of HPLC system V (Aminex A-29) is the short running time, only 5 4 min, which makes this approach highly attractive for studying enzymatic conversions. In order to obtain information about the structure of sialic acids, HPLC is a very useful

272 Table 8 High-performance liquid chromatographic elution times of sialic acids in five different HPLC systems I to V, relative to Neu5Ac [268] Compound

la

IIb

Neu2Me Neul ,2Me2

0.17

0.13

0.40

0.85

IV

IIIC

.oo '

*

Ve

0.23 0.46 1 .oo

1 .oo 1.41

1.98

I .22 0.95 1.30 1.17 1.62

1.39

2.04

Neu5Ac

1 .oo

1 .oo

1 .OO/l

Neu4,5Ac2

0.76

0.39

1.6811.71

'

Neu5,7Ac2

0.74 0.95

0.36 0.35

1.06 1.5711.62

'

Neu5,7(8),9Ac3 Neu5,7,8,9Ac4

0.74

0.23

1.90

Neu5AcZMe

0.73

I .02

Neu4,8an5Ac

2.15

0.58

Neu2en5Ac

2.21

0.66

1.54

1.68

Neu5Gc

1.17

1.50

0.84/0.78'

1.20

1.33

Neu4Ac5Gc

0.86 1.06

0.54 0.49

1.53 1.30'

1.59 1.57

1.69

0.82 0.67 0.59

0.75 0.36 0.29

1.03/0.98 1.70 2.20

0.94 1.21

0.89

Neu5,9Ac2

Neu9AcSGc Neu5Gc8Me Neu9AcSGc8Me Neu7,9Ac25Gc8Me

0.62

0.73

Kdn Acetic acid

0.34

0.08

0.87 1.47 1.76 2.05 0.73

1.70

1.47

' I .70

System 1: Anion-exchange chromatography on CarboPac PA-I with PAD detection at room temperature; 5 mM sodium acetate for 5 min, then a linear gradient to 50% 5 mM sodium acetate/50% 5 mM acetic acid in 30 min; mixing of the column effluent with 300mM NaOH; running time 18-24 min [268]. System 11: Amine adsorptiodion suppression chromatography on Micropak AX-5 with UV monitoring (200 nm) at room temperature; acetonitrile/water/0.25 M sodium dihydrogenphosphate (72: I8:6, v/v/v); running time 21-24 min [283]. System 111: Reversed-phase chromatography on TSK-ODS 120T and fluorescence monitoring (373 nm excitation and 448 nm emission) at room temperature; DMB derivatives; acetonitrile/methanoI/water (9:7:84, v/v/v); running time 12-13 min [271]. System I V Anion-exchange chromatography on Aminex HPX-72s with UV monitoring (200 nm) at 40°C; 0.1 M sodium sulfate; running time 15-1 7 min [284]. System V Anion-exchange chromatography on Aminex A-29 with UV monitoring (200nm) at room temperature; 0.75 mM sodium sulfate; running time 5 4 min [36,270]. Taken from [ 1 I], RP-18 column. a

'

technique to be applied in combination with mild chemical or enzymatic degradation methods. For instance, HPLC before and after alkaline treatment of a mixture of 0-acylated sialic acids can give information about the de-0-acylated sialic acids present, e.g. in terms of their N-acyl substituents. Also the linkage specificity of sialidases in releasing sialic acids from sialoglycoconjugates (see section 5.1) can be monitored by HPLC. Furthermore, the specificity of enzymes involved in sialic acid metabolism can be studied in this way. In this respect, interesting results have been obtained

213

with aldolase, cleaving sialic acids to N-acylmannosamine derivatives and pyruvate, and with sialate 9-O-acetylesterase, hydrolyzing 0-acetyl groups from C9 of sialic acids. The aldolase degrades Neu5Ac faster than Neu5Gc; a slow degradation has been observed for 0-acylated sialic acids, not affecting 4-0-acetylated sialic acids at all [5]. A typical example of the HPLC analysis of enzyme reactions, in which esterase and aldolase are involved, and including the non-enzymatic conversion of an 0-acetyl group from C7 to C9, is presented in Fig. 5 [6,7,36]. Other examples are the determination of sialidase activity (sialyllactose as substrate, Neu5Ac as product, and Neu2en5Ac as inhibitor), CMP-Neu5Ac synthase activity (disappearance of NeuSAc, appearance of CMP-NeuSAc), CMP-Neu5Ac phosphodiesterase activity (appearance of NeuSAc, disappearance of CMP-Neu5Ac) [36]. Recently, 5-N-acetyl-9-0-acetyl2-(N-dansyl-4-aminophenylthio)-a-neuraminicacid has been proposed as a highly sensitive fluorescent substrate for the HPLC measurement of sialate 9-0-acetylesterase (334 nm excitation, 564 nm emission) [285]. As a thioglycoside, the compound is very stable in acidic aqueous solution and towards enzymatic hydrolysis by sialidases. In this context, it is also worthwhile mentioning that a sensitive HPLC assay has been developed for the tracing of sialyltransferase activity, making use of the synthetic fluorigenic acceptor lactose 2-[(2-pyridyl)amino]ethyl glycoside [286]. Details for a HPLC separation of CMP-NeuSAc, CMP-NeuSGc and CMP-Kdn on a DC-613 cation-exchange column are reported in ref. [ 1071. In addition to the fractionation procedures described for free sialic acids, several approaches have been reported for the separation of sialyl-oligomers. These compounds with a degree of polymerization up to 16 have been fractionated with varying results using conventional gel-filtration, TLC, DEAE-Sephadex A-25, and HPLC methods (see section 5.1 for preparation; see section 5.3.1 for TLC). A survey of literature has been included in ref. [287]. In general, mixtures of sialo-oligomers from (NeuSAc(a2-8)},, {NeuSAc(a2-9)},, {Neu5Gc(a2-8)},, or { Kdn(a2-8)}, can be isolated on a preparative scale via convential DEAE-Sephadex A-25 [256,257] or DEAEToyopearl 650M [267] anion-exchange chromatography. HPLC procedures comprise anion-exchange and adsorption-partition chromatography. A mixture of { NeuSAc(a28)}2-16 has been efficiently separated on a Zorbax SAX anion-exchange column using 0.2-1 M NaCl in lOmM phosphate buffer pH 3.5 [287]. Also adsorptionpartition chromatography on polystyrene DC-6 13 using mixtures of 0.02-0.025 M sodium phosphate buffer pH 7.4 and acetonitrile as solvent system, has shown good results [267]. On Mono Q anion-exchange columns, excellent results were obtained in the separation of { N e u S A ~ ( a 2 - 8 ) ) ~ -{Ne~5Gc(a2-8)}~_lo, ~~, or {Kdn(a2-8)}~_~after conversion into alditols with NaBH4 (or NaBT4), and using a NaCl gradient in Tris-HC1 buffer pH 8 as elution system [267,288]. In this context, several studies have focused on the determination of the chain lengths of sialo-oligomers and -polymers (for a review of methods currently employed in the analysis of polysialic acids, see ref. [289]), and recently a highly detailed adapted methodology for the analysis of a-2,8-linked sialooligomers and -polymers has appeared [290]. Using three variable assay procedures, providing overlapping information, details could be provided with respect to the degree of polymerization, the simultaneous identification of NeuSAc, Neu5Gc and Kdn when present in a single preparation, and the ability to distinguish qualitatively between

N

-4 P

0.03

lyase

-

__c

(C)

0.02

(i

-

0.01 -

z

8

16

8

16

8

16

Fig. 5. HPLC profiling (Aminex A-28, 0.75 mM NazSO,) of the chemo-enzymatic conversion of Neu5,7Acz into pyruvate and N-acetylD-mannosamine (ManNAc). (a) Intramolecular migration of the 0-acetyl group from C7 to C9 under slightly alkaline conditions, yielding Neu5,9Acz, accompanied by some de-0-acetylation; (b) enzymic release of the 0-acetyl group at C9 with the aid of sialate-9-0-acetylesterase, yielding Neu5Ac; (c) aldolase (1yase)-catalyzed degradation of NeuSAc, yielding pyruvate and ManNAc.

275

reducing and non-reducing polymers. The developed approach may include mild periodate oxidation (degradation of non-reducing terminal unit) in combination with reduction (degraded glycerol side chain yielding C7-sialic acid; reducing unit if present affording the corresponding stereoisomeric alditols), whereas monomer analysis is carried out after sialidase or acid hydrolysis on CarboPac PA-1 with pulsed amperometric detection. In the structural analysis of glycoprotein-derived N- and 0-linked sialic-acid-containing carbohydrate chains, fractionation procedures based on HPLC play a major role. As this aspect is outside the scope of this chapter, no details are included. Typical examples, making use of anion-exchange chromatography (Mono Q, CarboPac), and normal (e.g. Lichrosorb-NH2) or reversed-phase chromatography, are found in refs. [25,81,84,133,291-3001. In this context, it is also worth noting the recent use of highperformance capillary electrophoresis for the separation of glycoprotein-derived N-glycan chains [301] and 0-glycan chains [302].

5.3.3. Gas-liquid Chromatography combined with muss spectrometry As discussed in section 5.1, methanolysis of free and glycosidically bound sialic acids gives rise to the formation of methyl ester b- and a-methyl glycosides. Using the conditions of the standard quantitative monosaccharide analysis of glycoconjugates, de-N-acylatiodde-esterification takes place, which means that NO-acylneuraminic acid residues are converted into neuraminic acid methyl ester methyl glycoside (8, -96%; a, -4%). For the characterization by GLC the sialic acid methyl ester methyl glycoside is derivatized via N-acetylatiodtrimethylsilylation or pertrifluoroacetylation [241]. It should be noted that during the N-acetylation step H 0 9 of NeuSAc, when not substituted, is partially 0-acetylated (-4%). Using milder methanolysis conditions [242], an N-acetylation step is not necessary, yielding a method to determine both Neu5Ac and Neu5Gc by GLC. In principle, the latter approach is also suitable to determine sialic acids bearing only 0-alkyl groups. GLC analysis is generally carried out on SE-30 type column materials. Starting from free sialic acids (purified or as a pool), mainly present in their 8-anomeric forms, volatile sialic acid derivatives are generated using mild derivatization procedures such as esterification with diazomethane followed by trimethylsilylation [ 15,3031 or pertrimethylsilylation [ 11,3041. With respect to silylation cocktails, it should be noted that N-methyl-N-trimethylsilyl-2,2,2-trifluoroacetamide/pyridine leads to the formation of N-trimethylsilyl derivatives, yielding two different peaks for each sialic acid [36,305]. Subsequent GLC analysis is generally carried out on SE-30 or OV-17 type column materials. Both types of derivatives are highly suitable for MS analysis, and GLC coupled with electron impact (EI) MS formed the basis for the development of a highly reliable mass spectrometric method for the identification of sialic acids. Originally set up for the GLC-EI MS analysis of mixtures of NO-acylneuraminic acids [303], the method has also proved to be useful for the analysis of other naturally occurring sialic acids, of (partially) 0-methylated sialic acid methyl ester methyl glycosides as obtained in methylation analyses, and of synthetic sialic acid(s) (derivatives) [6,11,15]. In the following the principles of the EI MS identification procedure will be explained. Typical derivatives are trimethylsilylated methyl ester or pertrimethylsilylated derivatives of N, 0-acylneuraminic acids or of N-acyl-0-alkylneuraminic acids, acetylated N-acyl-

216

I

CHORg

I

CH2ORs

J cH=6Rs

I

CHZORg

I

R40

- CHOR~CHOR~CH~OR~ - NHRgRs' H

I

H

- RzOH - R40H

ICI

Fig. 6 . Survey of the selected fragment ions A-H worked out for the following derivatives: trimethylsilylated methyl ester or pertrimethylsilylated derivatives of NO-acylneuraminic acids or of N-acyl-0-alkylneuraminic acids, and trimethylsilylated/methylated N,N-acy1,methyl-neuraminic acid methyl ester methyl glycosides [ 151.

0-alkylneuraminic acid methyl ester methyl glycosides, trimethylsilylatedrnethylated N,N-acy1,methyl-neuraminic acid methyl ester methyl glycosides (methylation analysis), and acetylatedmethylated N,N-acy1,methyLneuraminic acid methyl ester methyl glycosides (methylation analysis). The determination of the type, number, and position of the 0-acyl or 0-alkyl groups as well as the type of the N-acyl group in neuraminic acids is facilitated by the highly specific EI mass spectra of the derivatized compounds. In Fig. 6, a schematic survey is depicted showing the selected fragment ions A-H, which furnish the information (abundances and mlz values of the ions) necessary to deduce the complete structure of the sialic acids. Fragments A and B indicate the molecular mass of the sialic acid derivatives and thereby the type and the number of substituents. Fragments C-H can be used for the determination of the positions of the different substituents. Fragment A is formed from the molecular ion by the elimination of a methyl group originating from a trimethylsilyl substituent in trimethylsilylated (0-acylated0-alkylated) N-acylneuraminic acid derivatives. When RSI = CH3 (methylation analysis), the eliminated methyl group can

277

also originate from the NN-acy1,methyl group. Fragment B is formed by elimination of the C l part of the molecule. Eliminations of OCOCH3 in 0-acylated sialic acid derivatives and of NH2COCH3 in N-acetylneuraminic acid derivatives, which in principle give rise to the same mlz value as fragment B in the case of R1 =CH3, can be neglected. For 0-trimethylsilylated N 0-acylneuraminic acids (B-anomers) it holds that, when compared to their methyl esters, in their trimethylsilyl esters the intensity of fragment A decreases relative to B. Fragment C is formed by elimination of C8-C9, with localization of the charge on position 7. In general, cleavage occurs between two alkoxylated carbon atoms, or between an acetoxylated and an alkoxylated carbon atom, rather than between two acetoxylated carbon atoms. In accordance with the fragmentation rules for partially methylated alditol acetates [306], the charge is preferentially located on an ether oxygen instead of on an ester oxygen. Therefore, fragment C has only significant abundance if C7 bears an ether group. When an ester group is present at C7, this fragment ion is absent or hardly observable. Fragment D is formed from fragment C by consecutive eliminations of R20H and R40H. It is evident that the occurrence of this fragment ion is dependent on the presence of fragment C. Fragment E is formed by elimination of the whole side-chain C7-C8-C9 and the substituent at C5. This fragment ion is not observed if an 0-acyl group is attached to C4, illustrating that the transition state in the McLafferty rearrangement is more favored when the substituent at C4 is an ether group rather than an ester group. For 0-trimethylsilylated N, 0-acylneuraminic acids (B-anomers) it holds that, when compared to their methyl esters, in their trimethylsilyl esters the intensity of fragment E is much reduced but still present; instead, an additional fragment derived from fragment E by loss of Me3SiOH is clearly present. Fragment F contains C8-C9. Based on the same fragmentation rules as mentioned above for fragments C and D, this ion can only readily be formed if an ether group is attached to C8. Fragment G consists of the C4-CS part of the molecule. Fragment H, necessary to use for derivatives containing only 0-alkyl substituents, is formed by elimination of the C9 part, followed by elimination of R4OH and R70H. For instance, this fragment is useful to discriminate between an OSiMe3 group at C8 or C9 in trimethylsilylated partially methylated N-acylneuraminic acids. Finally, for quadrupole analyzers, in the high mass range the fragment ions A, B and C often are of low intensity, especially when only small amounts of material are available. In Table 9 a survey is presented of GLC retention times and of characteristic EI MS fragment ions for a series of naturally occurring sialic acids, analyzed as their trimethylsilylated methyl ester or as their pertrimethylsilylated derivatives [6,11,15,3 1,38,43,307]. Although the sialic acids predominantly occur in the B-anomeric form, the a-anomer could occasionally be detected separately from the B-anomer, in most cases as a small shoulder. As a typical example, in Figs. 7a,b the EI mass spectra of the trimethylsilylated methyl ester of B-NeuSAc and of the pertrimethylsilylated derivative of P-NeuSAc, respectively, are depicted. Additional spectra have been published in refs. [ 15,43,304,307]. For a detailed survey of fragment ions of other derivatives mentioned, including mass spectra, and those obtained from periodate-oxidized sialic acids (C7-NeuSAc, C8-NeuSAc, C7-NeuSGc and C8-NeuSAc), see refs. [15,308]; for EI MS data of Neu4,8anSAc, see ref. [63]; for EI MS data of permethylated Kdn, see ref. [46]. In additional studies, the suitability of chemical ionization (CI) for the GLC-MS analysis of pertrimethylsilylated N, 0-acylneuraminic acids has been investigated. Isobu-

Table 9 GLC and characteristic EI MS fragment ions (70 eV) of (i) trimethylsilylated methyl ester (TM) and (ii) pertrimethylsilylated (PT) derivatives of naturally occurring sialic acids (o-anomers) a Sialic acid

RN~~sA~ TM

Neu5Ac

1.00

Neu4,5Ac2

1.18 1.04

Neu5,7Ac2 Neu5,8Ac2

1.05

Neu5,9Ac2 Neu4,5,9Ac3

1.13 1.31

Neu5,7,9Ac3 Neu5,8,9Ac3 Neu5,7,8,9Ac4

1.14 1.19 1.15

Neu5Ac9Lt

2.55

Neu4,5Ac29Lt Neu5Ac8Me

3.01

Neu5,9Ac28Me Neu2en5 Ac Neu2,7an5Ac Neu5Gc Neu4Ac5Gc Neu7Ac5Gc Neu9Ac5Gc Neu7,9Ac25Gc Neu8,9Ac25Gc

1.09

PT 1.00 1.05

Fragment (mlz) A

1.04

0.98 1 .oo 1.01

1.81 2.02 1.83

1.19

2.04 2.01 I .99

1.21

Reference(s)

D

E

G

F

TM

PT

TM

PT

TM

PT

TM

PT

TM

PT

TM

PT

TM

PT

668 638 638

726 696

624 594 594

624 594

478 448

536 506

298 298

356 356

317

375

173 143 173

[11,15]

-

205 205

173

-

143

[11,15]

638 1.02

C

B

594

638 608

696

594 564

594

608

666

564

608 578

666

740 710

798

-

317

478

298

317

-

478 448

536

564

-

564 534

564

478

696 666

696

668 638 578

-

205 205 205

298 298

356

-

-

-

536

298

356

478 448

536 536

566 536

636 564

-

756 726 726

814 784

712 682 682

712 682

726 696 696

784

682 652 652

682

-

298 298

388

462 566 536 566

298

624 594

386 386

356

173 143

173

317

375

175

175

173

317 317

375

-

-

173 173

173 173

[151 [6,15] [11,15]

317

375

277

173 143

173

[11,151 [15,307]

356 444 444

227

-

-

566

386

285

205

317

375

-

-

317 317 317

375

205 205

-

205 205 205

205 205

261 231 261

175 175

175

261 261 261

-

-

~ 5 1 [11,31] [11,31] [11,15]

173 261

[6,38] [11,15]

231

[6,15]

261

~ 5 1 [11,15]

173 173

147 117

-

317 444

277 277

315 375

-

386

-

-

-

624

175

356 356

446 374

-

175 175

317 -

536

375

~ 5 1 ~ 5 1 [11,15]

173

~ 5 1 u51 continued on next page

4 N W

Table 9, continued Sialic acid

Fragment ( m / z )

RNeuSAc

TM

PT

Neu7,8,9Ac35Gc Neu5Gc8Me Neu9Ac5Gc8Me Neu5GcAc Neu2en5Gc

1.93

B

A TM

PT

666

TM

D

C PT

TM

PT

-

622

Reference@)

TM

E PT

-

TM

F PT

TM

G PT

-

317

TM

PT

1.14 1.17

756 726

654 624

624 624

444 444

375 375

147 117

26 1 261 261

~ 5 1 [11,31] [11,31]

1.21

784

682

594

414

375

205

231

[43]

724

-

534

444

285

205

-

[61

The RNeuSAc values of the TM derivatives on 3.8% SE-30 at 215°C are given relative to the TM derivative of P-NeuSAc. The RNeuSAc values of the PT derivatives on CP-Sil 5 (capillary column), using the program 5 midl40"C; 2Wmin up to 220°C; 15 mid220"C, are given relative to the PT derivative of P-NeuSAc. For the preparation of TM derivatives, see 1151; for the preparation of PT derivatives, see [304]. For an explanation of the minus signs, see text. a

-4 N

W

280

100-

(A)

%

I

E

317

50-

n b 00

L.h 200

100

h .

3 00

LOO

A 668

B 62L

I

L

L

L

L

t I,

J

L

I

G 173

A 726

+--%+--

Fig. 7. (A) EI mass spectrum (70 eV) of the trimethylsilylated methyl ester of b-NeuSAc; (B) EI mass spectrum (70 eV) of the pertrimethylsilylated derivative of P-NeuSAc.

28 1

tane[304], as well as methane and ammonia[ll] were used as reactant gases. The CI mass spectra are characterized in the high mass range by [M+H]+ pseudomolecular ions, and typical major fragment ions derived from [M+H]+ by loss of R20H (fragment I), &OH (fragment I'), and RzOH+&OH (fragment 11). It was found that methane in particular gave CI spectra that also include several of the typical fragment ions observed in the EI spectra. In addition to GLC-MS, HPLC-CI MS with Aminex A-29 as column material and ammonium formate in water/acetonitrile as solvent system has been explored for the analysis of underivatized N , 0-acylneuraminic acids [6,309]. Although the positive-ion mass spectra allow the discrimination between different N-acylneuraminic acids (NeuSAc, NeuSGc) and the determination of the degree of 0-acetylation (Neu5,9Ac2, Neu5,7,9Ac3, Neu5,7,8,9Ac4), the position of the 0-acetyl groups (Neu4,5Ac2, Neu5,7Ac2, Neu5,9Ac2) could not be established. For the latter assignment, the combination with specific elution positions of standards on the HPLC column is advised. In the interpretation of the various fragment ions, the open-chain structure of the sialic acids has been generally used. NeuSAc and NeuSGc have also been converted into phosphatidylethanolamine dipalmitoyl derivatives, and after separation by HPTLC and subsequent isolation, the sialic acid derivatives were analyzed by liquid secondary ion MS [281]. In both cases intense [M - HI- ions together with sodium attachment ions were detected. For the detection of Neu5Ac on human tumor mucin, after liberation with sialidase, electrospray MS has been used [310]. 5.3.4. Fast atom bombardment mass spectrometry Free sialic acids, isolated after cleavage from glycoconjugate starting material, have been investigated, without derivatization, by FAB MS using 5% aqueous acetic acid solutions for loading into glycerol on the FAB target [4 I]. The positive and negative FAB mass spectra of each sialic acid showed clear [M+H]+ and [M-HI- pseudomolecular ions, respectively. Sialic acid mixture analysis (pg range) made the recognition of subgroups of sialic acids with the same molecular mass possible (e.g. NeuSAc(Ac)l, NeuSAc(Ac)z, NeuSAc(Ac)3, NeuSGc(Ac)3). However, a differentiation between positional isomers was not possible. Sialic acids were also studied after derivatization, which improves the sensitivity [4 I]. Direct peracylation failed to produce suitable derivatives, but reduction under acidic conditions followed by peracylation (perdeuteroacetylation or perpropionylation) gave good results. Generally, the sialic acids give rise to two major pseudomolecular ions, corresponding to the peracylated open chain form and an open-chain-derived lactone form, and a minor pseudomolecular ion corresponding to an open-chain-derived anhydrofonn (2,6 and/or 4,8). As the lactone peak is markedly reduced in the spectrum of Neu4,5Ac2, H 0 4 seems to be mainly involved in the lactonization. In the case of sialic acid mixtures, a fast sialic acid subgroup analysis based on molecular masses is possible; again, a differentiation between positional isomers cannot be achieved. Careful analysis of the negative FAB spectra of reduced and perpropionylated sialic acids in mixtures demonstrated that these spectra could also be used for quantitative purposes. As worked out for mixtures of NeuSAc, Neu5,9Ac2 and NeuSGc, an estimate of the relative amounts of these sialic acids can be given with an error of 1&I 5%, when the sum of the intensities

282

of the [M + HIf ions of the linear and the lactone forms of each component is compared, taking into account that the molar response of Neu5Gc is approximately 50% of that of Neu5Ac. In order to generate sialic-acid-derived compounds, which can be used to differentiate between positional isomers, use has been made of the rather difficult periodate oxidation under mild or more rigorous conditions [53,308]. Under both conditions, the resulting aldehyde groups were derivatized with p-amino-benzoic acid ethyl ester, reductively introduced at acidic pH without loss of the native 0-acetyl functions [41]. Sialic acids treated in this way were additionally reduced and peracylated, and then analyzed by FAB MS. Mixtures of products with different ring sizes (original, lactonized, anhydro) and/or open chain forms, depending on the substitution pattern, are often obtained. Of the mono-0-acetylated N-acylneuraminic acids Neu4,5Ac2, Neu5,7Ac2, Neu5,9Ac2, Neu4Ac5Gc and Neu9Ac5Gc were investigated, however, no attention was paid to the behavior of Neu5,8Ac2. Neu5,7(8),9Ac3 and Neu7,9Ac25Gc were also included in these investigations. Although not discussed in this chapter, FAB MS is widely used in the characterization of glycoprotein-derived N- and 0-linked sialic-acid-containing carbohydrate chains. Typical information can be found in refs. [311,312].

5.3.5. ' H NMR spectroscopy Since the introduction of high-resolution H NMR spectroscopy for the structural analysis of glycoprotein-derived glycans, a huge amount of NMR data have been generated, and highly detailed reviews on N-linked [76] and 0-linked [77] carbohydrate chains have appeared. The continuous expansion in the amount of data has made it necessary to develop computerized search programs, and, in connection with the still growing Complex Carbohydrate Structural Database (CARBBANK), attention has been paid to the development of a NMR-spectroscopic data base of carbohydrate structures, called SUGABASE [3 131. Sialic-acid-containing oligosaccharides/glycopeptides constitute a considerable majority of the glycoprotein glycans. In addition to the two reviews mentioned above, a specific review focusing on the NMR spectroscopy of sialic acids has also been published [ 161. Free as well as glycosidically bound sialic acid give rise to highly characteristic 'H NMR parameters. The 'H NMR spectra are generally recorded in D20, and because of the pH dependency of the proton chemical shifts, the spectral data are standardized at pD 6-7. The choice of the pH is also of importance in view of the earlier discussed de-0-acylation, 0-acyl migration, and autohydrolysis. As a typical example of a free sialic acid, in Fig. 8 the 500MHz 'H NMR spectrum of Neu5Ac in D20 at pD 7 is depicted. The spectrum shows a minor and a major set of protons, reflecting the subspectra of the a- and p-anomer of NeuSAc, respectively (a$ = 7:93), and especially the H3e,3a signals, resonating outside the bulk signal, can be used for the differentiation between both anomers. The effect of pH on the proton chemical shifts is clearly illustrated by the positions of the Neu5Ac H3e,3a resonances. At pD 1.4, the H3e and H3a signals of B-Neu5Ac resonate at S 2.313 and 6 1.880, respectively, whereas these values are 6 2.208 and 6 1.827 at pD 7. In the case of a-NeuSAc, at pD 1.4, the resonances are found at 6 2.718 and S 1.705,

'

SAC

H5

rSAc

HO

H7

I!

,

4.1

4.0

3.9

3.8

3.6

n30

3.5

2.2 ppm

2.1

2.0

Fig. 8. Resolutionenhanced 500MHz 'H NMR spectrum of Neu5Ac dissolved in D20recorded at pD 7 and 27°C.

1.9

1 .I

1.7

1.6

284

respectively, and at pD 7.0 at 6 2.730 and 6 1.621, respectively. Over the years, a large number of (naturally occurring) sialic acids and related derivatives have been analyzed, and in Table 10 a survey of the chemical shift values of a selected group of compounds, including naturally occurring sialic acids and CMP-sialic acids, is presented [16,31,38,51,64,107,308,314,315]. Inspection of these data reveals the typical influences of the substituents on the chemical shift values of the skeleton H-atoms. Comparison of the H3a,3e resonances of NeuSAc and NeuSGc indicate more downfield positions for those of NeuSGc (A6 M + 0.02). Additional data for NeuSAc, Neu5,7Ac2, Neu5,9Ac2, Neu5,7,9Ac3, and Neu5,8,9Aq in 0.1 M sodium phosphate/D20 at 37°C and pD 7-7.5 are reported in ref. [23]. The 'H NMR studies of free NO-acylneuraminic acids have shown that the anomeric equilibrium of a 7-0-acetylated sialic acid differs strongly from that of the sialic acids not substituted at C7. In contrast to the normal equilibrium values of a$ M 7:93, both Neu5,7Ac2 and Neu5,7,9Ac3 have an equilibrium ratio of about 23:77. Additional ' H NMR data of methyl glycosides and methyl ester methyl glycosides of N,0-acylneuraminic acids and some sialyllactoses can be found in ref. [16]. The 'H NMR data of the (methyl ester) methyl glycoside of (3-Kdn have been reported in ref. [46], and those of CMP-9amino-NeuSAc and CMP-9NAc-NeuSAc in ref. [316]. ' H NMR spectroscopy has shown to be an excellent method to monitor chemical and biochemical conversions of sialic acids, directly in the NMR-tube or by analysis of isolated reaction products. A typical example is the demonstration of the release of a-NeuSAc as the primary product of bacterial and viral sialidase action on Neu5Ac(a2-glycosides and oligosaccharides [3 17-3201. The initial formation of a-NeuSAc, as traced by 'H NMR spectroscopy, formed an excellent probe to investigate the kinetics of the mutarotation of NeuSAc by means of 'H NMR analysis in dependency of the pH [321]. At pD 5.4 the establishment of the equilibrium of mutarotation turned out to be rather slow, but at higher and lower pD values a more rapid establishment was observed, so that at pD 1.3 and pD 11.7 mutarotation was too fast to be measured. With the ability to generate a-NeuSAc in situ, the aldolase-catalyzed degradation of NeuSAc to pyruvate and N-acetylmannosamine (ManNAc) could also be investigated in more detail using ' H NMR spectroscopy [322]. Using sialidase (pH optimum 5.4) and aldolase acid (pH optimum 7.2) from C. perfringens and N-acetyl-2-azido-2-deoxy-a-neuraminic as substrate at pH 5.4, only released a-NeuSAc was found to be consumed by the aldolase, yielding specifically a-ManNAc followed by a fast mutarotation to a,P-ManNAc. These findings confirmed earlier work using Neu5Ac(a2-3)lactose as a-Neu5Ac-generating system and crystalline b-NeuSAc [319]. In the reversed reaction a-ManNAc is the substrate [322]. For more details about the aldolase-catalyzed degradation, see section 9.3. In connection with these studies, it has to be noted that under comparable conditions the activities of sialidase and aldolase in D20 are only about 50% of those in H20. With respect to 0-acetyl migrations, also the earlier mentioned (see section 2) spontaneous conversion at physiological pH of Neu5,7Ac2 into Neu5,9Ac2 and of Neu5,7,9Ac3 into Neu5,8,9Ac3 has been monitored by H NMR spectroscopy [23]. Furthermore, the NMR approach has shown its value in the determination of substrate specificities of various sialidases using substrates with differently linked sialic acid residues [3233251.

'

Table 10 'H Chemical shifts for different types of sialic acids. Chemical shifts are given in ppm relative to internal acetone in D 2 0 (6 2.225) at 300K, unless indicated otherwise Sialic acid

uD

Chemical shift H3a

H3e

H4

H5

H6

H7

H8

H9

H9'

Ref 4Ac

SAC

7Ac

8Ac, 8Me

B-NeuSAc

2

1.880

2.313

4.067

3.93 1

4.056

3.556

3.750

3.841

3.619

2.053

-

-

a-NeuSAc

2

1.705 2.718

n.d. a

3.85

3.684

3.53

3.75

3.85

3.62

2.036

-

-

B-NeuSAc

7

1.827 2.208

4.024

3.899

3.984

3.514

3.753

3.835

3.608

2.050

-

-

a-NeuSAc

7

1.621 2.730

n.d.

n.d.

n.d.

n.d.

n.d.

n.d.

n.d.

2.030

-

-

P-Neu4,5Ac2

7

1.951

2.249

5.274

4.15

4.15

3.570

3.775

3.844

3.619

1.992

-

-

~-Neu5,7Ac2

4

1.905

2.236

3.950

3.767

4.246

5.045

3.91 1

3.629 3.444

a-Neu5,7Ac2

4

1.649 2.757

n.d.

n.d.

n.d.

n.d.

n.d.

n.d.

n.d.

P-Neu5,9Ac2

7

1.833 2.221

4.024

3.913

3.99 1

3.571

3.977

4.365

a-Neu5,9Ac2

7

1.624 2.720

n.d.

n.d.

n.d.

n.d.

n.d.

fi-Neu5,7,9Ac3

2

1.924 2.303

3.978

3.775

4.293

5.162

a-Neu5,7,9Ac3

2

1.686

n.d.

n.d.

n.d.

n.d.

B-Neu5,8,9Ac3

7

1.838 2.189

3.978

3.903

3.780

3.838

0-NeuS, 8,9Ac3

2

1.862 2.250

Neu2enSAc

6

-

Neu2,7an5AcC

7

Neu4,8an5Acd

7

2.751

1.976b 2.144b

-

1.947b 2.12gb

-

4.187

2.057

-

-

n.d.

n.d.

n.d.

-

-

4.140

4.106

4.106

1.981b 2.134b

-

n.d.

n.d.

n.d.

1.956b n.d.b

-

5.114 4.528

4.287

2.057

-

2.089

4.006

3.912

3.830

3.866

5.1 15 4.545

4.287

2.059

-

2.091

5.690 4.470

4.051

4.213

3.601

3.936

3.885

3.646

2.068

-

-

2.167

2.007

3.953

3.919

4.543

4.434

3.537

3.592

3.755

2.035

-

-

2.983

2.844

4.188

4.333

3.852

3.496

3.363

3.814

3.734

2.041

-

-

2.143

1.723

n.d.

4.249

n.d.

3.439

3.268

3.690

n.d.

2.041

--

-

9Ac

5Gc

Table 10, continued Sialic acid

PD

Chemical shift H3a

P-Neu5Ac2P

7

H3e

1.548 2.403

CMP-P-Neu5Ac

8

CMP-P-NeuSAc9Ac

7

1.64

C8-P-Neu5Ac

6

1.814 2.203

1.639 2.484 2.48

Ref

H4

H5

H6

H7

H8

H9

H9'

4.093

3.888

4.239

3.386

4.028

3.883

3.581

4.066

3.92

4.141

3.456

3.92

3.90

3.622 n.d.

4Ac

5Ac

7Ac

8Ac, 8Me

9Ac

2.045

-

2.054

-

5Gc

-

[I61

-

[I61

n.d.

3.96

n.d.

3.51

4.07

n.d.

2.05

2.08

-

[314]

3.998

3.892

3.772

3.722

3.655'

-

2.055

-

-

[308]

-

-

2.044

-

-

I3081

[308]

3.586' C7-P-NeuSAc

6

1.821 2.238

3.992

3.710

3.800

3.671g 3.6349

C7-a-Neu5Ac

6

1.599 2.630

n.d.

n.d.

n.d.

n.d.

-

-

-

2.026

-

-

P-Neu5Gc

7

1.840 2.243

4.127

4.002

4.106

3.549

3.777

3.821

3.613

-

-

4.143

[16]

n.d.

n.d.

n.d.

n.d.

n.d.

-

4.12

[16]

a-Neu5Gc

7

1.644 2.749

n.d.

n.d.

P-Neu9AcSGc

7

1.842 2.234

4.14

4.006

4.109

3.570

3.970

4.365

4.183

-

2.115

4.144

[I61

a-Neu9Ac5Gc

7

1.649 2.751

n.d.

n.d.

n.d.

n.d.

n.d.

n.d.

n.d.

-

n.d.

4.123

[I61

P-NeuSGc8Me

7

1.863 2.219

4.017

3.559

3.432

3.932

3.652

3.425

-

4.133

[31]

a-Neu5GcSMe

7

1.643 2.55 1

n.d.

n.d.

n.d.

n.d.

n.d.

n.d.

-

4.125

[31]

3.60

-

-

4.11

[314]

-

4.144

[308]

4.1 10 3.983 n.d.

n.d.

CMP-P-Neu5Gc

7

1.64

2.48

3.92

4.03

n.d.

3.42

3.60

3.87

C8-P-Neu5Gc

6

1.832 2.218

4.097

3.976

3.887

3.709

3.657'

-

3.589' continued on next page

Table 10, continued ~

Sialic acid

Chemical shift

PD H3a

C8-a-NeuSGc C7-P-Neu5Gc

6 6

Ref.

H3e

H4

H5

H6

H7

H8

H9

H9'

4Ac

5Ac

7Ac

8Ac, 8Me

9Ac

5Gc

1.622 2.665

n.d.

n.d.

n.d.

n.d.

n.d.

-

-

-

-

-

-

-

4.129

[308]

-

-

-

-

-

-

-

-

4.133

[308]

-

-

-

-

-

-

-

-

4.116

I3081

-

-

-

1.840 2.253

4.096

3.798

3.906

3.6749 3.6329

C7-a-NeuSGc

6

n.d.

2.650

n.d.

n.d.

n.d.

n.d.

6-Kdn

7

1.837 2.265

4.015

3.603

3.995

3.881

3.772

3.883

3.675

-

-

-

[I061

o-Kdn9Ac

7

1.768 2.158

3.94

3.575

3.94

3.91

3.91

4.376

4.238

-

-

-

-

2.124

-

[51]

CMP-P-Kdnh

7

1.568 2.379

3.968

3.538

4.046

3.863

3.896

3.731

3.618

-

-

-

-

-

-

[I071

n.d., not determined. On one line values may have to be interchanged. ' Sialic acid occurs in 5C2 conformation. Sialic acid occurs in 'C4 conformation and is present in two tautomeric forms; chemical shifts are relative to HOD at 6 4.750; H3a= H3 and H3e = H3'. Values are assigned relative to the HOD signal at 6 4.81 at 296 K.

a

H8 and H8'. H7 and H7'. Personal communication Y. Inoue; values are assigned relative to 2,2-dimethyl-2-silapentane-5-sulphonatein D 2 0 (set to 0 ppm). Note that in CMP-6-Kdn the H8 and H9 signals have been interchanged when compared with ref. [3 151.

g

N m 4

288 Table 11 'H-Chemical shifts for the H3e and H3a signals of sialic acids as part of N- and 0-linked glycoprotein glycans. Chemical shifts are given in ppm relative to internal acetone in D 2 0 (6 2.225) at 300 K, or relative to internal 2,2-dimethyl-2-silapentane-5-sulphonate in D 2 0 set to 0 ppm (marked ') Structural element

H3a

H3e

NeuSAc(a2-3)Gal(fiI -

1.78-1.81

2.75-2.78

Neu4,5Ac2(a2-3)Gal(P 1-

1.93

2.77

Neu5,9Ac2(a2-3)Gal(fi1 Neu5Gc(a2-3)Ga1((31-

1.80

2.76

1.81-1.82

2.77-2.79

Kdn(a2-3)Gal(fi14)GlcNAc(flIKdn(a2-3)Gal(fi1-3)GalNAc(al-3)GalNAc-ol Neu5Ac(a2-3)[Gal(fil-4)]Gal(fi1NeuSAc(a2-3)[GalNAc(fi 14)]Gal(fl1 -

1.75-1 .76a

2.11-2.72

1.72a

2.69a

1.82a 1.92-1.94 1.70-1.72

2.73 a 2.66-2.68 2.67

Gal(~l-3)[Neu5Ac(a24)]GlcNAc(fil-2)Man(a1-

1.85 1.71 1.73-1.74 1.71-1.72

2.68 2.67 2.69-2.70 2.73-2.74

Gal(fi1-3)[Neu5Ac(a24)]GlcNAc(fi 1-4)Man(a 1NeuSAc(a24)GalNAc(fi 1 -

1.76-1.77 1.70-1.72

2.72-2.73 2.66

Neu5Gc(a2-3)[GalNAc(fil-4)]GalNAc(fi1-

1.85-1.86

Kdn(a2-3)[GalNAc(fiI 4)]GalNAc(Bl-

2.55-2.56 2.48 a

~~~

~

Neu5Ac(a2-6)Gal(flINeu4,5Ac2(a2-6)Gal(fi 1Neu5,9Ac2(a2-6)Gal(fi1 NeuSGc(a24)GalbI -

Neu5Ac(a24)GalNAc-ol

1.79a 1.69-1.71

Neu5Gc(a24)GalNAc-oI

1.71-1.72

2.74-2.75

Kdn(a2-6)GalNAc-01

1.64-1.66

2.66-2.68

a

2.72-2.74

Another interesting pH phenomenon, for the first time demonstrated by 'H NMR spectroscopy, is the complete replacement of H3a by a D-atom, when free Neu5Ac is kept in an alkaline D20 solution at pD 9.0 [326]. In fact, H3a can be exchanged in the pH range 6.5-9.0, and the H-D exchange is reversible. In the 'H NMR spectrum the H3a signal disappears, and the coupling patterns of H3e and H4 alter. At pD 12.4, H3e can also be replaced by a D-atom [327]. On the basis of this finding, the exchange experiments were also carried out in T20, yielding T-labelled NeuSAc, which was converted enzymatically into T-labelled CMP-Neu5Ac [326]. In glycosidically linked NeuSAc the H3 atoms are not exchangeable, rendering this specific labelling technique suitable for sialyltransferase experiments. In the H NMR structural-reporter-group concept, developed for the structural analysis of glycoprotein N- and 0-glycans, advantage is taken of the fact that a number of the H-atoms of the constituting monosaccharides, pccurring in special microenvironments, resonate outside the bulk-signal region [76,77,328]. In the case of glycosidically linked a-sialic acids the structural reporters are the H3a- and H3e-atoms, and the N-acetyl or N-glycolyl groups. The positions of the H3a,3e signals reflect

289

not only structural information with respect to the type of sialic acid present, but also with respect to the coupled monosaccharide in terms of type of linkage and type of monosaccharide. Furthermore, 0-acyl substituents induce additional shifts for other H-atoms. In Table 11 a survey is presented of chemical shifts of H3e and H3a signals of sialic acid residues occurring in different linkage types as part of N- and 0-glycans. For a further fine-tuning of the chemical shifts within the presented ranges, influenced by the different microenvironments wherein the sialic acid residues occur, see refs. [76,77] and the references cited in Table 2. It should be noted that the presence of a certain sialic acid in a certain linkage also influences the structural-reportergroup signals of other monosaccharide residues [76,77]. More detailed information with respect to the H3e,3a chemical shifts of sialic acids in sialo-oligomers and -polymers can be obtained from refs. [31,47,77,329,330]. A series of H3a,3e signals of sialic acids in milk- and glycolipid-derived oligosaccharides have been included in ref. [77]. More general NMR data of glycosidically linked sialic acids in glycolipids, milk and urinary oligosaccharides and (lipo)polysaccharides, if available, can be found in the references cited in Tables 3-5 (see also ref. [331]). For a series of sialocarbohydrates, it has been shown that H6 of NeuSAc, easily traced from the TOCSY H3a,H6 correlation, also has potential value for discriminating between a-2,3- (6 3.63 & 0.01 1) and a-2,6- (6 3.70 0.017) linked NeuSAc in NeuSAc(a2-3/6)Gal(~l-O)R/NeuSAc(a26)GalNAc(al-O)Ser/Thr [33 13. However, for branched oligosaccharides, this rule is not valid if the monosaccharide in the branching position is in the alditol form [332]. The H3a,3e/NAc structural reporters of Neu5Ac have proved to be suitable in studying sialylation reactions in terms of positional specificity and branching (N-linked carbohydrate chains) specificity, using different sialyltransferases and CMP-Neu5Ac as donor [333-3351 (see also section 6.3). In more biophysical studies, several aspects of sialic acids have been investigated by NMR spectroscopy. Although so far mainly a-forms of bound sialic acid have been detected, good differentiation systems for a- and p-forms are essential, and a number of empirical rules have been reported [336]. In a heteronuclear 2D-approach it could be demonstrated that the determination of the geminal C,H coupling constant 2J(C2,H3a) offers a unique criterion for the anomeric assignment in sialic acid glycosides (a, -8 Hz; p, -3 to -4Hz) [337]. Also the values of the vicinal C,H coupling constants 3J(C1,H3a) can be applied for this differentiation (a, -6 Hz; 0, -1 Hz) [55,336,338]. More details with respect to anomeric determinations have been reviewed in ref. [339]. 13C NMR data of sialic acids and glycoprotein-derived sialocarbohydrates have been reviewed in ref. [ 161. In several sialic-acid-related investigations, e.g. synthetic studies, I3C NMR spectroscopy forms part of the analysis techniques, and will not be reviewed in detail. For some additional data, more directly related with the glycoprotein glycan character of this chapter, see refs. [48,63,64,103,105,152,33 1,332,337,340-3431,

*

6. Chemo-enzymatic highlights in sialic acid chemistvy During the last ten years, activity in sialic acid chemistry has grown exponentially. Synthetic as well as biosynthetic routes for the preparation of sialic acids, sialic acid

290

derivatives, analogues, glycosides and sialoglycoconjugates have been explored. The main reason for this considerable interest in preparing sialic acids and sialic-acidcontaining compounds lies in the fact that sialic acids were found to be among the most biologically important carbohydrate units in glycoconjugates. The progress in organic synthetic protocols and the availability of relevant enzymes in suitable amounts made it realistic to develop the sialic acid field from a preparative synthetic side. Initially, the main tools were to prepare suitable derivatives to study the properties of sialic acids, or to prepare substrates and inhibitors for sialidases, sialyltransferases or for sialic acid converting enzymes. Although these tools are still highly relevant, the preparation of sialo-oligosaccharides using strictly organic synthetic or enzymatic methods, or a mixture of both, are also receiving considerable attention. For relevant reviews on preparative (bio)synthetic aspects, see refs. [339,344-3531. 6.1. Free sialic acids

Several protocols have been followed for the organic chemical synthesis of NeuSAc. At first, the approaches were based on condensation reactions of (derivatives of) N-acetylD-mannosamine (ManNAc) or N-acetyl-D-glucosamine (GlcNAc) with (derivatives of) oxaloacetic acid (for a review, see ref. [3]), however, the yields were very low. In this context, a procedure was worked out, that allowed modifications at Cl-C3 [354]. A total synthesis of NeuSAc from non-carbohydrate precursors has been reported in ref. [355]. Using a protocol for indium-mediated allylations of aldehydes, NeuSAc was prepared in good yields from a ManNAc precursor [356]. Furthermore, synthetic routes for NeuSAc have been proposed, based upon 1-deoxy- 1-nitro-sugar chemistry, that should also allow the preparation of NeuSAc analogues, modified at several carbon atoms of the skeleton [357-3591. A separate route also yielded NeuSAc [360]. A synthesis starting with the aldol condensation of D-glucose (Glc) and oxaloacetic acid, followed by adaptation of the substituent at C5 has been described in ref. [361]. More recently, another approach for the organic synthesis of NeuSAc and NeuSAc derivatives, based on the cis-selective Wittig reaction of benzoyl 2,3-O-isopropylidene-a-~-lyxo-pentodialdo1,4-furanoside with [(3S)-3,4-(isopropylidenedioxy)butyl]-triphenylphosphonium iodide as a first step, has been reported in ref. [362]. Comments on the acetylation of NeuSAc and its methyl ester have been published in ref. [363]. Starting from NeuSAc, in some derivatization reactions 1,4- as well as 1,7-lactone formation has been observed [349]. For the organic synthesis of Kdn, several routes were employed [364-3681, among them procedures starting from NeuSAc or from D-mannOSe (Man). NeuSAc-aldolase-catalyzed condensations of ManNAc and pyruvate [3,33], initially only investigated to understand sialic acid metabolism, have been optimized for preparative purposes. In principle, ManNAc can be generated from the cheaper GlcNAc in an alkaline epimerization process, yielding an epimeric mixture of which only the monosaccharide with the D-manno-configuration is recognized by the aldolase [369,370]. However, ManNAc can also be generated from GlcNAc in a GlcNAc-epimerasecatalyzed isomerization [37 I]. A multigram-scale enzymatic synthesis based on the aldol condensation of ManNAc and pyruvate in the presence of phosphate, catalyzed by immobilized microbial NeuSAc-aldolase, has been reported [372]. A similar approach,

29 1

using the aldolase enclosed in a dialysis membrane instead of being immobilized, has also been described [373]. Although in uiuo, the conversion of Neu5Ac into Neu5Gc occurs exclusively on the level of activated sialic acid (see section 8.4.1), Neu5Gc can be prepared in uitro by incubating a mixture of N-glycolyl-D-mannosaminelN-glycolylD-glucosamine and pyruvate with immobilized aldolase [374]. Interestingly, a series of other sugars also turned out to be accepted by the aldolase, and Man and 2-deoxyD-glucose in particular are excellent substrates [372,375-3771. In the case of Man as starting material, relatively moderate amounts of Kdn have been prepared [254,376]. A series of free 0-acetylated sialic acids, i.e., Neu4,5Ac2, Neu5,9Ac2, Neu4,5,9Ac3, and Neu5,7,8,9Ac4, together with the benzyl ester a-glycosides of Neu5,7Ac2 and Neu5,7,9Ac3, have been synthesized by organic synthetic routes using protecting group techniques [378,379]. Partially 0-acetylated sialic acid derivatives have also been prepared using more simple synthetic routes. @-Neu5,9Ac21,2Me2, fi-Neu4,5,9Ac31,2Me2, and @-Neu4,5,8,9Ac41,2Me2 were obtained from @-NeuSAcl,2Me2 by using N-acetylimidazole [53]. To realize 9-O-acetylations, also other acetylating reagents were applied, such as trimethyl orthoacetate [3 80,38 I], acetyl chloride [378], and dimethylacetamide dimethyl acetal[382]. In a recent comprehensive study, in particularly the use of trimethyl orthoacetate and dimethylacetamide dimethyl acetal was explored using the 4-aminophenylthio, 4-nitrophenylthio, and 4-nitrophenyl glycosides of COOH-esterified a-Neu5Ac as acceptors, and depending on the acetylating reagent a range of partially 0-acetylated derivatives could be generated [57]. One of the naturally occurring 0-acetylated sialic acids, Neu5,9Ac~,has also been synthesized in an enzymatic way [383,384] on a gram-scale [372]. After the enzymatic acetylation of 0 6 of ManNAc, using isopropenyl acetate and protease N as a catalyst, 2-N-acetyl-6-O-acetyI-~-mannosamine was condensed with pyruvate as catalyzed by the aldolase. These two enzymatic steps turned out to be highly regio- and stereoselective. Following another route, Neu5,9Ac2 has been synthesized enzymatically by incubating Neu5Ac with trichloroethyl acetate in pyridine using porcine pancreas lipase as a catalyst [385]. An enzymatic synthesis of Neu5Ac9Lt has also been worked out [370,384]. For the study of biochemical pathways, several isotopically labelled sialic acids and sialic acid derivatives have been prepared, both by enzyme-catalyzed synthesis and by organic synthesis. A survey of labelled sialic acids is presented in Table 12 [107,344,386]. In enzymatic procedures, use is generally made of the aldolase-catalyzed condensation of N-acybmannosamines and (phosphoenol)pyruvate, suitably labelled in one or both of the two synthons. For the preparation of N-[l-'4C]acetyl- and N - [ I-'4C]glycolylneuraminic acid, as well as O-['4C]acetylated sialic acids, surviving slices of submaxillary salivary glands incubated with [ 1-I4C]acetate, followed by isolation of the glycoprotein fraction and mild acid hydrolysis, have been used. A number of these labelled sialic acids have been converted into their CMP-glycosides, and subsequently incorporated into glycoconjugates (see section 6.3). Of course, labelling of glycoconjugates can also be carried out by periodate oxidatiodtritiated borohydride reduction, thereby converting sialic acids, if chemically possible, into their radiolabelled C7 and C8 analogues. Using the latter approach, fluorescent probes (dansylhydrazine, dansylethylenediamine, fluoresceinamine) and EPR spin labels can also be incorporated (see references cited in ref. [344]). The same holds for glycine [387].

292 Table 12 Survey of radiolabelled sialic acidsa Sialic acid

Reference

N-[3H]Acetylneuraminic acid N - [ 1-'4C]Acetylneuraminic acid N-A~etyl-[3-~H]neuraminicacid N-Acetyl-[9-) Hlneuraminic acid N-Acetyl-[ l-14C]neuraminicacid

N-A~etyl-[4-'~C]neurarninic acid N-A~etyl-[2-~~C,9-~H]neuraminic acid

S-N-AcetyI-9-azido-9-deoxy-[ 1-14C]neuraminicacid 5-N-A~etyl-4-O-methyI-[3-~ Hlneuraminic acid 5-N-[1-'4C]Acetyl-2-deoxy-2,3-didehydro-neuraminicacid

N-[l-14C]Glycolylneuraminic acid N-Glycolyl-[ 1-14C]neuraminicacid

N-Gly~olyl-[2-~~C,9-~H]neuraminic acid [14C]-2-Keto-3-deoxynononic acid a

For specific references, see [344].

Several interesting sialic acid variants and sialic acid derivatives have been synthesized, and a list is presented in Table 13 [3 16,344,347,349,357-359,368,370,376-380,384,3884461. A number of these compounds have been surveyed in ref. [3SO]. Both organic synthetic and aldolase-catalyzed routes have been followed. The major part of these compounds were prepared to study sialic acid metabolism (aldolase, CMP-NeuSAc synthase), sialic acid transfer (sialyltransferases), sialic acid release (sialidases), inhibition phenomena, or hemagglutinin-sialic acid interactions, and biological details are presented in sections 8-10. Compounds reported up to 1982 have been reviewed earlier [344]. Of special interest are the fluorescent and photoactivatable sialic acid derivatives [390,4 191, which can be applied, after conversion into their corresponding CMP-glycosides, to detect enzyme activities or to follow biological processes (see sections 6.2 and 8.2). In the context of sialic acid variants, the following compounds are also of interest. In view of the similarity in acidity of a tetrazoie group and a carboxyl function, a variant of Neu5Ac has been prepared containing a CN4H instead of a COOH group [447]. Also the synthesis of a series of Neu5Ac derivatives with specifically introduced tert-butyldimethylsilyl groups have been reported [400,448]. Furthermore, variants of 2d-2Ha-NeuSAc and 2d-2He,NeuSAc, in which the carboxyl function has been replaced by a phosphono (PO3H2) group [449], and a phosphonic acid analogue of Neu2en5Ac [450], have been synthesized. In addition to 6-amino-2,6-dideoxy-sialic acids, as mentioned in Table 13, the preparation of 2-C-hydroxymethyl derivatives [45 I], and C6 and C7 analogues [452] have also been reported.

Table 13 List of sialic acids, prepared along organic chemical or aldolase-catalyzed routes for use in biochemical studiesa Compound

Abbreviation

N-Acetylneuraminic acid

NeuSAc

5-N-Acetyl-2-deoxy-2,3-didehydro-neuraminic acid

NedenSAc

2-Deoxy-2,3-didehydro-neuraminicacid

Neu2en

N-Glycolylneuraminic acid

Neu5Gc

2-Deoxy-2,3-didehydro-5-N-glycolyl-neuraminic acid

Ne3enSGc

5-Azido-neuraminic acid

NeuSN,

5-Azido-2-deoxy-2,3-didehydro-neuraminic acid

NeuZenSN,

2-Keto-3-deoxynononic acid

Kdn

2,3-Didehydro-2,3-d~deoxy-~-g~cero-~-ga~acfo-non-2-ulopyr~oson~c acid

Kdn2en

N-Acetyl-[ 1-’3C]neuraminicacid

[ l-”C]NeuSAc

N-A~etyl-[3-~H]neuraminic acid

[3-*H]NeuSAc

N-A~etyl-[6-~H]neuraminic acid

[6-2H]Neu5Ac

Reference@)

N-Aminoacetyl-neuraminic acid

NeuSAcNH2

N-Thioacetyl-neuraminic acid

NeuSAcSH

5-N-Acetyl-2,7-anhydro-neuraminic acid

Neu2,7an5Ac

5-N-Acetyl-2-deoxy-2-Hax-neuraminic acid

2d-2Ha,-Neu5Ac

[392-3941

5-N-Acetyl-2-deoxy-2-HCq -neuraminic acid 5-N-Acetyl-2-deoxy-2-Heq-4-oxo-neuraminic acid

2d-2HCq-Neu5Ac

[393-3951

2d-2Heq-40x0-Neu5Ac

W I

5-N-Acetyl-2-deoxy-4-epi-neuraminicacid

2d-4epi-NeuSAc

5-N-AcetyI-2-deoxy-7-epi-2-H,,-neuraminicacid

2d-7epi-2HCq-Neu5Ac

[3971 [394,398]

5-N-Acetyl-2-deoxy-8-epi-2-Heq-neuraminic acid 5-N-Acetyl-2-deoxy-7,8-diepi-2-H,,-neuraminic acid

2d-8epi-2Heq-Neu5Ac

[394,398]

2d-7,8epi2-2Heq-Neu5Ac

[394,398] continued on next page

N

rg

w

N

W P

Table 13, continued Compound

Abbreviation

Reference(s)

5-N-Acetyl-3-fluoro-neuraminic acid

3F,, -Neu5Ac

[344,392]

5-N-Acetyl-3-hydroxy-neuraminic acid

30Heq-Neu5Ac

[344,399]

5-N-Acetyl-4-0-acetyl-neuraminic acid

Neu4,5AcZ

[378,380]

5-N-Acetyl-4-O-acetyl-4-epi-P-neuraminic acid methyl ester methyl glycoside

4epi-PNeu4,5Ac21 ,2Me2

5-N-Acetyl-4-deoxy-neuraminicacid

4d-Neu5Ac

[4001 [358,4014031

5-N-Acetyl-4-deoxy-4-iodo-neuraminic acid

41-Neu5Ac

[4041

5-N-Acetyl-4-deoxy-4-(R)-C-methyl-neuraminic acid

[4051

5-N-Acetyl-4-deoxy-4-(S)-C-methyl-neuraminic acid

[d051 [357,4061

5-N-Acetyl-4-epi-neuraminicacid

4epi-Neu5Ac

5-N-Acetyl-4-epi-4-0-methyl-neuraminic acid ethyl ester

4epi-Neu5Ac 1Et4Me

5-N-Acetyl-4-0-methyl-~-neuraminic acid (ethyl esterkthyl glycoside)

Neu5Ac4Me

[3441 [344,407]

5-N-Acetyl-4-0x0-neuraminic acid (methyl P-glycoside or ethyl a-glycoside)

40x0-Neu5Ac

[408,409]

5-N-Acetyl-7-deoxy-neuraminic acid

7d-Neu5Ac

[377,410]

5-N-Acetyl-7-epi-neuraminic acid 5-N-Acetyl-7-0-methyl-neuraminic acid

7epi-Neu5Ac

[411,412]

Neu5Ac7Me

P771

5-N-Acetyl-7-0x0-fi-neuraminic acid methyl glycoside

7oxo-fiNeu5Ac2Me

[4081

5-N-Acetyl-8-deoxy-neuraminicacid

8d-Neu5Ac

5-N-Acetyl-8-epi-neuraminicacid

8epi-Neu5Ac

[4 101 [411,412]

5-N-Acetyl-8-0-methyl-neuraminic acid

Neu5AcBMe

5-N-Acetyl-8-0x0-neuraminicacid methyl a- and P-glycoside

8oxo-Neu5Ac2Me

W41 [394,408,413]

5-N-Acetyl-9-0-acetyl-neuraminicacid

Neu5,9Acz

see text

5-N-Acetyl-9-S-acetyl-9-thio-neuraminic acid

[4141 continued on next page

Table 13, continued Compound

Abbreviation

Reference(s) ~~

S-N-Acetyl-9-amino-9-deoxy-neuraminic acid

9amino-NeuSAc

acid (methyl a-glycoside) S-N-Acetyl-9-azido-9-deoxy-neuraminic

9azido-NeuS Ac

S-N-Acetyl-9-(4-azidobenzamido)-9-deoxy-ne~uaminic acid S-N-Acetyl-9-(4-azidosalicylamido)-9-deoxy-neuraminic acid 5-N-Acetyl-9-benzamido-9-deoxy-neuraminic acid

9NBz-NeuSAc

5-N-Acetyl-9-(4-benzoylbenzamido)-9-deoxy-neuraminic acid

5-N-Acetyl-9-cyano-9-deoxy-a-neuraminic acid benzyl glycoside S-N-Acetyl-9-O-(N-dansylglycyl)-neuraminicacid 5-N-Acetyl-9-deoxy-neuraminic acid (methyla-glycoside)

9cyano-aNeuSAcZBn

5-N-Acetyl-9-deoxy-9-(3-fluoresceinylthioureido)-neuraminic acid

9d-NeuSAc 9fluoresceinyl-NeuSAc

S-N-Acetyl-9-deoxy-9-fluoro-neuraminic acid

9F-NeuSAc

S-N-Acetyl-9-deoxy-9-iodo-neuraminic acid

91-NeuSAc

S-N-Acetyl-9-deoxy-9-thioacetamido-neuraminic acid 5-N-Acetyl-9-O-(dimethylphosphinyl)-neuraminicacid 5-N-Acetyl-9-O-glycy1-neuraminic acid methyl ester

NeuSAc9GlylMe

S-N-Acetyl-9-hexanoylamido-9-deoxy-neuraminic acid

5-N-Acetyl-9-0-lactyl-neuraminicacid 5-N-Acetyl-9-0-methyl-neuraminic acid

9NHx-NeuSAc NeuSAc9Lt NeuSAc9Me

5-N-Acetyl-9-0-phosphoro-neuraminic acid

NeuSAc9P

~~

~

[316,390, 415 4 171 [316,384,390, 417,4181 [4191 [4191 [316,390,415, 4171 [4191 13941 [4201 [394,410,421] [390,415,422, 4231 [384,4244261 [344,427] [390,3911 [3W 13441 [316,390,4171 [370,384] [3471 [3441 continued on next page

W N

N m W

Table 13, continued Compound

Abbreviation

5-N-Acetyl-9-thio-neuraminic acid

5-N-Acetyl-4,9-di-U-acetyl-neuraminic acid

Neu4,5,9Ac3

5-N-Acetyl-2,7-dideoxy-2-Heq-neuraminic acid

2,7dz-2Heq-Neu5Ac 2,8dz-2Heq-Neu5Ac 4,7dz-Neu5Ac 7,9dz-Neu5Ac 7,8epi2-Neu5Ac

5-N-Acetyl-2,b-dideoxy-2-Heq-neuraminic acid

5-N-Acetyl-4,7-dideoxy-neuraminic acid 5-N-Acetyl-7,9-dideoxy-neuraminic acid 5-N-Acetyl-7,b-diepi-neuraminic acid (methyl a-glycoside)

5-N-Acetyl-7,7-dimethoxy-P-neurarninic acid methyl glycoside

5-N-Acetyl-4,7,9-trideoxy-neuraminic acid

5-N-Acetyl-4-azido-2,3-didehydro-2,4-dideoxy-neuraminic acid 5-N-Acetyl-9-S-(4-azido-2-nitrophenyl)-2,3-didehydro-2,9-dideoxy-9-thio-neuraminic acid 5-N-Acetyl-2-deoxy-2,3 -didehydro-4-epi-neuraminicacid (methyl ester) 5-N-Acetyl-2-deoxy-2,3-didehydro-4-oxo-neuraminic acid 5-N-Acetyl-2-deoxy-2,3 -didehydro-6-thio-neuraminicacid 5-N-Acetyl-2-deoxy-2,3-didehydro-7-epi-neuraminic acid 5-N-Acetyl-2-deoxy-2,3-didehydro-8-epi-neuraminic acid

5-N-Acetyl-2-deoxy-2,3-didehydro-7,8-diepi-neura~nic acid

[4101

[4211 [394,411,412]

Neu5,7,8,9Ac4 4,7,9d3-Neu5Ac

[3791 [4211

4amino-Neu2en5Ac 4azido-Neu2enSAc

~4281 [428430] [430,431]

5-N-Acetyl-4-allylamino-2,3-didehydro-2,4-dideoxy-neuraminicacid 5-N-Acetyl-4-amino-2,3-didehydro-2,4-dideoxy-neuraminic acid

[4141 P781 [394,398] [394,398]

[4081 [394,408,413]

5-N-Acetyl-8,8-dimethoxy-neuraminic acid methyl a-and P-glycoside 5-N-Acetyl-7,8,9-tri-U-acetyl-neuraminic acid

Reference(s)

4epi-NeuZenSAc 4oxo-Ne3en5Ac

14321 [344,433,434] [409,433,435] [4361

7epi-Neu2enSAc bepi-Neu2enSAc 7,8epi2-Neu2en5Ac

[3501 [4371 [3501

5-N-Acetyl-2-deoxy-2,3-didehydro-8,8-dimethoxy-neuraminic acid continued on next page

Table 13,continued Compound

Abbreviation

Reference(s)

5-N-Acctyl-2,3-didehydro-2,4-didcoxy-4-dimcthylamino-neuram~nic acid 5-N-Acetyl-2,3didehydro-2,4-dideoxy-4-gu~idinyl-neuraminic acid

4guanidino-Neu2enjAc

[428,429]

[4281

5-N-Acetyl-4-(N-hydroxy-N-allylamino)-2,3-didehydro-2,4-dideoxy-neuraminic acid

[4281

5-N-Acctyl-4-( 2-hydroxycthylamino)-2,3-dide~~dro-2,4-dideoxy-neuraminic acid

[4281 [4381 P501

5-N-acetyl-2,3-didchydro-2,4-didcoxy-ncuraminic acid 5-N-acetyl-2,3-didehydro-2,8-dideoxy-neuraminic acid 5-N-acetyl-2,3 didehydro-2,9-dideoxy-neuraminic acid

4d-NeuZenSAc 7d-Neu2enSAc 8d-NedenSAc 9d-Neu2en5Ac

5-N-acetyl-2,3didehydro-2,4,7-trideoxy-neuraminic acid

4,7d2-Neu2en5Ac

5-N-acetyl-2,3-didehydro-2,7-dideoxy-neuraminic acid

4-Acetamido-5-N-acetyl-2,3-didehydro-2,4-dideoxy-neuraminic acid

13501 [350]

13501 14281

5-N-Acetyl-6-amino-2-dcoxy-2-H,,-neuraminicacid 5-N-Ace~yl-6-amino-2-deoxy-2-H,,-4-epi-neuraminiti atiid 5-N-Acetyl-6-amin0-2-deoxy-2-H,~-4-epi-neuraminic acid 5-N-Acetyl-6-thio-neuraminic acid 5-N-Acetyl-5-epi-6-thio-neuraminic acid 5-N-Acetyl-4,5-diepi-6-thio-neurarninic acid 9-U-Acetyl-5-N-glycolyl-neuraminic acid

Neu9Ac5Gc

4-Acetamido-5-N-aetyl-4-deoxy-neuraminic acid

4NAc-Neu5Ac

4-Acelamido-3,4-~deoxy-n-g~~ero-~-~~~u~f~-non-2-ulopyranosonic acid

iso-Neu4Ac 2d-2F,-3Fq-Neu5Ac

5-Acetamido-2,6-anhydro-3,5-dideoxy-2,3-dIfluoro-~-urubino-~-~u~o-nononic acid 5-Acetamido-2,6-anhydro-3,5-dideoxy-4-C-met~yl-~-e~~~~o-~-u~~~o-nononic acid 5-Acetamido-2,6-anhydro-3,5-dideoxy-4-C-methyl-~-erythro-~-munno-nononic acid

conrinued on next page

W N 4

N W

Table 13,continued Compound

Abbreviation

Reference(s)

5-Acetamido-2,6-anhydro-4-C-methyl-3,4,5-tndeoxy-D-eryfhro-L-u~fro-nononic acid

W I

5-Acetamido-2,6-anhydro-4-C-methyl-3,4,5-trideoxy-~-eryfhro-~-mu~~o-nononic acid

t3961 [3961 [442,443]

5-Acetam~do-2,6-anhydro-l-C-methylene-3,4,5-tndeoxy-~-glycero-~-gulacfo-nonon~c acid 5-Acetamido-2,5-dideoxy-2,3-difluoro-~-eryfhro-~-gZuco-non-2-ulopyranosonic acid 5-C-Acetamidomethyl-5-deamino-neuraminic acid

9-Acetamido-5-N-acetyl-9-deoxy-neuraminic acid

9NAc-NeuS Ac

[4441 [3 16,390,415, 4171

5-N-Glycolyl-9-0-phosphoro-neuraminic acid

Neu5Gc9P

[W

5-Bromo-3,5-d~deoxy-~-glycero-~-gulucfo-non-2-ulopyranoson~c acid

5Br-Kdn 5epi-Kdn 4epi-Kdn 5epi-5F-Kdn 5d-Kdn 7d-Kdn 7epi-7,9F2-Kdn

3-Deoxy-D-g~cero-D-gu~o-non-2-u~opyranosonic acid 3-Deoxy-D-g~cero-D-fa~o-non-2-u~opyranosonic acid 3,5-Dideoxy-5-~uoro-D-g~cero-D-gu1o-non-2-u~opyra~0sonic acid 3,5-Dideoxy-~-g~cero-~-ga~acfo-non-~-u~opyranosonic acid

3,7-Dideoxy-D-g~cero-D-gu~acfo-non-2-ulopyranosonic acid 3,7,9-Tndeoxy-7,9-difluoro-D-glycero-~-alfro-non-2-ulopyranosonic acid a

For literature references before 1982 and additional lists of sialic acid(?.)

(derivatives), see ref.[344].

[376,384,445]

299

6.2. Glycosides of sialic acids The organic synthesis of a long series of alkyl and aryl a-glycosides of N-acylneuraminic acids has been previously reported (for reviews, see refs. [339,344]). One of the famous condensation reactions (classical Koenigs-Knorr method) comprised the silver carbonate-promoted condensation of 5-N-acetyl-4,7,8,9-tetra-O-acetyl-2-chloro-2-deoxy(3-neuraminic acid or the corresponding methyl ester (Fig. 9, structure A) with the appropriate alcohol, followed by removal of the protecting groups. In order to improve the yields, much attention has been paid to better catalysts. In these glycosidation reactions typical side reactions are the formation of unsaturated sialic acid derivatives (elimination of HCl) and of (3-glycosides(see section 6.3). For the preparation of simple (3-glycosides, N-acylneuraminic acids are often heated with the appropriate alcohol in the presence of an acid catalyst, followed by saponification of the formed ester. However, complex alcohols give rise to problems (for a review, see ref. [344]). A mild and efficient Raney nickel-catalyzed deuteration procedure has been reported for Neu5Ac glycosides, with a rate of exchange at C8 > C9 > C7 >> C4 [453]. Attention has also been paid to the synthesis of N-, S- and Se-glycosides, which are sialidase stable [339,454,455]. Specific S-glycosides are used as sialic acid donors in sialoglycoconjugate organic synthesis (see section 6.3). Early examples are the syntheses of the 4-nitrophenyl N- and S-glycosides of a-Neu5Ac [456]. Of special interest are the syntheses of 5-N-acetyl-2-azido-2-deoxy-a- and p-neuraminic acids [322,457,458]. The azides can readily be converted into the corresponding 2-amino derivatives, and used in e.g. N-acylation reactions [459]. For a preparation of the 6-thioanalogue of 2azido-a-NeuSAc, see ref. [436]. In further investigations, a series of S-glycosides of a-Neu5Ac was synthesized (thiophenyl, 4-nitrothiophenyl, 4-aminothiophenyl, 2-mercaptopyridyl), starting from 5-N-acetyl-4,7,8,9-tetra-O-acetyl2-chloro-2-deoxy-~-neuraminic acid methyl ester and using triethylbenzylammonium chloride as a phase transfer catalyst 114541. These compounds turned out to be effective sialidase inhibitors. For the detection of sialidase activity both (naturally occurring) oligosaccharides and simple a-glycosides are used. In these assays, two approaches can be followed, namely, determination of the released (modified) sialic acid or identification of the released aglycon. In the case of a-glycosides, in which the released aglycon concentration is measured spectrophotometrically or detected on solid supports, substrates with synthetically introduced aglycons having specific chromogenic properties, are used. Among such substrates, released aglycons can be detected directly or after condensation with specific reagents. One of the oldest substrates is the 4-nitrophenyl glycoside of a-NeuSAc [456], whereby released 4-nitrophenol is estimated by absorption at 400nm. In an adapted synthetic version, the compound has been prepared by coupling of 5-N-acetyl-4,7,8,9-tetra-O-acetyl-2-chloro-2-deoxy-(3-neuraminic acid methyl ester with sodium-nitrophenoxide in N,N-dimethylformamide, and subsequent deprotection [460]. Another suitable substrate is the 3-methoxyphenyl glycoside of a-NeuSAc, synthesized by coupling of 5-N-acetyl-4,7,8,9-tetra-O-acetyl-2-chloro-2deoxy-(3-neuraminic acid with 3-methoxyphenol in the presence of silver carbonate, followed by de-0-acetylation. Liberated 3-methoxyphenol is determined after coupling

300 X

OAc

I

(A)

1

AcOH z AcHN OAc X = OAc. F, CI, Br Y = H, Me, Bn. All

ACOH 2CW AGO C

O

O

M

e

AcoH 2 A ccd

w

c

o

AcHN

AcHN OAc

o

M

e

OH OAc

X=F; Y=SePh X=Br; Y=SPh X=CI: Y = S P h X = F ; Y=SPh Z = Me or Bn I

0-P(

COOMe

OAc

('3

1

Acd AcHN

AcHN

OAc

OAc

X Et, Bn Y = Me. Bn

Y = Me, Ph, C-OEt

II

S

COOMe

OAc

I

1

COOMe Acd AcHN

SPh

OAc AcHN OAc

Fig. 9. Frequently used NeuSAc donors in the organic synthesis of sialo-oligosaccharides.

ox

ox

301

with the diazonium salt of 4-amino-2,5-dimethoxy-4'-nitroazobenzene(red colored product) [461] or with 4-aminoantipyrine in the presence of the oxidizing agent potassium ferricyanide (colored quinone) [462]. The most popular fluorigenic substrate is the 4-methylumbelliferyl glycoside of a-NeuSAc, which is prepared by different methods [344]. A convenient synthesis is the condensation of 5-N-acetyl-4,7,8,9-tetraO-acetyl-2-chloro-2-deoxy-~-neuraminic acid methyl ester with the sodium salt of 4-methylumbelliferone in N,N-dimethylformamide, followed by deprotection [463]. Released 4-methylumbelliferone is measured at 360 nm (excitation)/440 nm (emission). Although the 4-methylumbelliferyl glycoside of a-Kdn has been synthesized starting from Neu5Ac [445], also a direct route using the glycosyl chloride of peracetylated b-Kdn methyl ester and the sodium salt of 4-methylumbelliferone has been explored [254]. In addition, several 4-methylumbelliferyl a-glycosides of sialic acid variants and substituted sialic acids, including partial 0-acetylated ones, have been synthesized (e.g. refs. [252,350,399,445,464]. To develop a sensitive assay for the analysis of the linkage specificity of bacterial and viral sialidases, Neu5Ac(a2-3)- and NeuSAc(a2-6)Gal(P 1O)C6H4N02 were synthesized enzymatically by using a-2,3- and a-2,6-sialyltransferase, respectively, CMP-NeuSAc (see section 6.3), and p-nitrophenyl-P-Galp [465]; after cleavage of NeuSAc, p-nitrophenol can be released by additional treatment with b-galactosidase. For the localization of sialidase on electropherograms or for histochemistry, the chromogenic 5-bromo-indol-3-yl glycoside of a-NeuSAc has been synthesized by coupling of 5-N-acetyl-4,7,8,9-tetra-O-acetyl-2-chloro-2-deoxy-~-neuraminic acid methyl ester with 1-acetyl-5-bromo-3-hydroxyindole,and subsequent deprotection [466]. The unstable intermediate 5-bromo-indoxyl, released by sialidase, is readily transformed into insoluble blue-green 5,Si-dibromo-indigo, which marks the sites of enzyme activity. To facilitate the screening of bacterial colonies or plaques for sialidase activity, the 5-bromo-4-chloro-indo1-3-yl glycoside variant has also been synthesized [467]. Using the same sialic acid synthon as starting product, the 4-azido-2-nitrophenyl S-glycoside of a-Neu5Ac has been prepared, which is a potential photoaffinity probe reagent for the screening of sialidases in tissues and the purification of sialic-acid-binding proteins [468]. The sialidase-resistant thioglycosyl linkage also makes the incorporation of 35 S possible. In order to detect sialate 9-0-acetylesterase activity, a highly sensitive fluorescent substrate, 5-N-acetyl-9-O-acetyl-2-[4-(dansylamino)phenylthio]-a-neuraminic acid, has been synthesized (see also sections 5.3.2 and 9.1) [285]. The regioselective acetylation at 0 9 of the dansylated S-glycoside was carried out with trimethyl orthoacetate. Other useful fluorescent substrates for sialate 0-acetylesterase assays comprise 5-N-acetyl-7,8,9-tri-O-acetyl-2-[4-(dansylamino)phenylthio]-a-neuraminic acid and 5-N-acetyl-4-O-acetyl-2-[4-(dansylamino)phenylthio]-a-neuraminic acid [469], and the 4-[3-(fluoresceinyl)thioureido]phenyl S-glycoside of a-Neu5,9Ac2 [57]. In all cases the fluorescent groups have been coupled to the glycosidic 4-aminophenylthio group of 0-acetylated NeuSAc derivatives. In connection with the generation of a monoclonal antibody to free Neu5Ac for the purpose of establishing a simple and specific assay of NeuSAc in serum and urine, a broad series of sialic acid a- and P-glycosides have been synthesized using substituted glycerol, substituted sphinganine and cholesterol as aglycons [470].

3 02 NaOH/H20

GlcNAc

pyruvate

ManNAc

Neu5Ac

C Neu5Ac aldolase

CTP

pyruvate

CMP-P-Neu5Ac

PEP

Fig. 10. One-pot synthesis of CMP-6-NeuSAc from ManNAc and CMP [473]. PEP, phosphoenolpyruvate; PPi, pyrophosphate.

The synthesis of CMP-sialic acids is generally carried out enzymatically using CTP and CMP-sialic acid synthase as a catalyst [33,3 14,370,47 1,4721. A multigram-scale one-pot synthesis of CMP-@-NeuSAchas been reported in ref. [473]. ManNAc, prepared by basecatalyzed epimerization of GlcNAc, was reacted with sodium pyruvate in the presence of NeuSAc-aldolase to yield NeuSAc (see section 6.1). For the formation of CMPNeuSAc, CTP was generated in situ from CMP by using adenylate kinase, pyruvate kinase, and phosphoenolpyruvate, and reacted with NeuSAc in the presence of CMPNeuSAc synthase (Fig. 10). Instead of a one-pot synthesis, for practical reasons it is easier to generate and store crude solutions of NeuSAc and CTP. For the use of GlcNAc in combination with GlcNAc-epimerase, see ref. [474]. Experiments with cloned CMPNeuSAc synthases from E. coli systems with NeuSAc and Kdn showed a high specificity for NeuSAc, thereby suggesting that in this case the 5-acetamido group is critical [384]. Chemical syntheses of CMP-NeuSAc, applying the phosphoramidite method [475] or using sialyl phosphites [476], have also been described. Furthermore, a synthetic approach for the preparation of CMP-NeuSGc based on the phosphite method has appeared [477]. In addition to CMP-p-NeuSAc, CMP-@-Neu5,9Ac2,CMP-b-NeuSGc, and CMP-8-Kdn, a large series of artificial CMP-sialic acids have been prepared biochemically on microscale starting from the corresponding sialic acid (see references cited in Table 13) and CTP. Among them are CMP-9azido-NeuSAc, CMP-9amino-NeuSAc, CMP9NAc-NeuSAc and other C9-modified CMP-sialic acids, CMP-NeuSAcNH2, CMPNeuSAc4Me, and CMP-4d-NeuSAc [33,3 14,350,390,407,419,4781. The CMP-sialic acids have found a broad application in enzymatic sialylations using different sialyltransferases (see sections 6.3 and 8.3). Several of the artificial CMP-sialic acids turned out to be suitable donors for asialo-a, -acid glycoprotein as acceptor with Gal(@lH)GlcNAc a-2,6-sialyltransferase from rat liver as a biocatalyst [3 16,402,4171. The transfer of CMP-9amino-NeuSAc is of considerable interest, as a-linked 9amino-NeuSAc in sialoglycoconjugates is not a substrate for bacterial, viral or mammalian sialidases tested so far. CMP-9amino-NeuSAc and CMP-NeuSAcNHz have also been used as synthons

303

to prepare fluorescent and photoactivatable analogues [419]. Because of the defined acceptor specificity, sialyltransferases in combination with fluorescent or photoactivatable donor CMP-sialic acids are excellent tools for selective introduction of a fluorescent or photoactivatable substituent to a distinct glycoconjugate. The latter reference [4 191 also includes kinetic data and information concerning the fluorimetric sialyltransferase assay. Typical fluorescent products comprise the CMP-sialic acids of S-N-acetyl-9-deoxy-9(3-fluoresceinylthioureido)-neuraminicacid (CMP-9fluoresceinyl-Neu5Ac), 5-N-acetyl9-(7-amino-4-methylcoumarinyl)acetamido-9-deoxy-neuraminic acid (CMP-9AMCANeuSAc), 5-N-acetyl-9-deoxy-9-(fluoresceinylaminomonochlorotriazinyl)amino-neuraminic acid (CMP-~MTAF-N~USAC), and N-(3-fluoresceinylthioureido-acetyl)neuraminic acid (CMP-NeuSfluoresceinyl). In the preparation of photoactivatable derivatives the NH29 group of CMP-9amino-NeuSAc has been substituted with a 4-azidobenzoyl, a 4-azidosalicyl, a 4-benzoylbenzoyl or a 4-azido[T]benzoyl group. In a similar way, CMP-NeuSAcNH2 has been labeled with a 4-azidobenzoyl group. Of special interest is the recently reported chemical synthesis of CMP-{NeuSAc(a2-8)NeuSAc} [479]; an attempt to prepare this compound biosynthetically with NeuSAc(a2-8)NeuSAc and CMP-sialic acid synthase failed so far [472]. In addition to the preparation of regular CMP-sialic acids, synthetic approaches have been worked out for the organic synthesis of a S-(N-acetylneuraminy1)nucleoside analogue [480,48 11 and other CMP-sialic acid variants [482]. For the immobilization of sialic acids on Sepharose solid supports, which provides potentially useful affinity materials, see the references cited in ref. [344]. The preparation of an affinity adsorbent with immobilized sialic acid through a thioglycosidic linkage has been described in ref. [483]. Synthetic sialidase-stable a-Neu5,9Ac2 p-aminophenylthio glycoside has been immobilized directly or by a six-carbon long spacer group to agarose for lectin isolations [382]. The allyl glycoside of a-NeuSAc has been applied as a starting material for the synthesis of NeuSAc-neoglycoproteins and pseudopolysaccharides. These polymers containing multivalent sialic acid are in principle useful for various applications related with recognitionhindinglinhibition processes. Reductive ozonolysis of the allyl then MezS), followed by coupling of the formed aldehyde to protein carriers group (03, (E-aminogroup of lysine) by sodium cyanoborohydride-mediated reductive amination, yielded neoglycoproteins with varying amounts of NeuSAc [484,485]. Copolymerization of the allyl glycoside with acrylamide generated a water-soluble pseudopolysaccharide [484]. In order to create a longer spacer arm for copolymerization with acrylamide, the allyl glycoside was converted into a 3-(2-aminoethylthio)propyl glycoside by reaction with cysteamine hydrochloride, after which the amino function was N-acryloylated [486]. The same principle of conjugation or copolymerization via N-acryloyl groups was also used for the preparation of sialo-oligosaccharide-neoglycoproteins and copolymers of sialo-oligosaccharides and acrylamide [487]. Using the strategy of reductive amination, p-formylphenyl glycoside of a-NeuSAc was also conjugated with proteins [488], and starting from the p-nitrophenyl 0- and S-glycosides, p-N-acryloylamino analogues were synthesized, which could be copolymerized with acrylamide, yielding water-soluble pseudopolysaccharides with Neu5Ac and acrylamide in different ratios [489], or directly coupled with polylysine [490]. Using trimethyl orthoacetate, the NeuSAc residue of the

3 04

S-glycoside-containing polymer was converted into Neu5,9Ac2 [489]. Finally, a series of interesting NeuSAc and Neu5,9Ac2-based dendrimers have been synthesized [49 11.

6.3. Sialo-oligosaccharides The organic synthesis of oligosaccharides having terminal a-linked sialic acid has proved to be highly complex. The specific difficulties arise from three factors inherent in the sialic acid molecule. First, the carboxylic acid function at the anomeric center (C2) electronically disfavors oxonium ion formation. Secondly, from a steric point of view, the carboxyl function restricts the glycoside formation. Thirdly, the presence of a neighboring methylene group in the ring (C3), instead of a substituted carbon atom, eliminates the possible assisting andor directing effect of an adjacent substituent [346]. This means that side reactions can be relatively important, mainly the thermodynamically favored fi-glycoside and 2,3-dehydro-derivative formation, and low yields are quite often obtained. Initially, the synthesized glycosidic linkages comprised mainly NeuSAc(a26)Gal(fi 1-, NeuSAc(a24)GlcNAc((31-, NeuSAc(a2-6)Glc(fll-, NeuSAc(a2-3)Gal((3 1-, NeuSAc(a2-3)GlcNAc(fl1- [344], and over the years these glycosidic linkages, together with NeuSAc(a24)GalNAc(al-, still receive most of the attention. In view of the desire to prepare biologically relevant carbohydrate chains, this is understandable. In Fig. 9 a series of typical NeuSAc donors, introduced by different research groups with the aim of increasing both the glycosidation yield and the a-stereoselectivity, is depicted [339,34&348]. For each class of donors, some further information is presented in the following paragraphs. The oldest approach of synthesizing sialo-oligosaccharides is the one starting from 2-deoxy-2-halo-fi-NeuSAc derivatives (Fig. 9, structure A). Methyl S-acetamido-4,7,8,9tetra-O-acetyl-2-ch~oro-2,3,5-t~deoxy-~-g~cero-fi-~-ga~acto-non-2-u~opyranosonate, X= C1, Y = Me, turned out to be a particularly useful donor [492], and typical promoters are silver and mercury salts. Due to poor stereoselectivity, a$-glycoside mixtures are generally obtained, and HC1-elimination from the donor is a major side reaction. The reaction with secondary hydroxy groups in particular gave rise to problems. In the case of the aim to prepare NeuSGc-containing oligosaccharides, also the N-glycolyl group in the donor analogue is 0-acetylated [493]. For the synthesis of NeuSAc(a2-9)NeuSAq see ref. [494]. In another approach, a series of 3-substituted NeuSAc donors was prepared, starting from peracetylated [495] or perbenzylated [496,497] Neu2enSAc methyl ester, thereby making use of the highly reactive 2,3-double bond to form adducts (Fig. 9, structures B-F) (see also ref. [498]). In the case of structure B as donor with silver triflate as a promoter, only fi-glycosidic linkages were created, and among several products, the NeuSAc(P2-8)NeuSAc linkage was synthesized [495]. Structure C yielded mainly (3-glycosidic linkages. From structure D, only the bromo variant is effective, although a$-glycoside mixtures are still formed. The bromo variant of structure D with silver triflate as a promoter has been applied in the synthesis of NeuSAc(a2-8)NeuSAc and NeuSAc(a2-9)NeuSAc linkages [499,500]. Structures E and F form another series of donors, and E with X = Br and Y = SPh (mercury salts as promoter) has been shown to give particularly high glycosidation yields and a-stereoselectivity.

305

A third type of donor involves the use of S-methyl or S-phenyl a-glysosides (Fig. 9, structure G) [498,501-5031. Initially developed to synthesize S-glycosides, making use of sodium salts of the peracetylated Neu5Ac methyl ester a- or 8-thioglycosides and suitable protected bromides [504-5061 (see also refs. [507,508]), this type of donors has shown to be highly attractive in 0-glycosidation reactions. In these couplings, frequently used promoters are dimethyl(methy1thio)sulfonium triflate or N-iodosuccinimideltriflic acid [347]. The choice of the solvent system is very important, as it greatly influences the stereoselectivity; e.g. acetonitrile gives mainly a-glycosidation. In addition to reports dealing with the synthesis of many monosialo-oligosaccharides, including those with a NeuSAc(a2-2)Glc, a NeuSAc(a2-3)GlcNAc, and a NeuSAc(a2-3)GalNAc sequence [509], typical examples are the creation of NeuSAc(2-9)NeuSAc [5 10,5 1 11 and NeuSAc(a2-8)NeuSAc [512] linkages, as well as sialyl LeX sequences (native and variants) (ref. [513] and references cited therein). In this context, it is also interesting to note that several syntheses of sialo-oligosaccharides include the use of a separately prepared disaccharide donor with a terminal a-linked sialic acid [514]. Although in general the donors contain a N-acetyl group at C5, other examples have been reported with the phthaloyl (benzeneselenenyl triflate as a promoter [5 151) or the tert-butoxycarbonyl function as N-protecting group, e.g. in the case of the synthesis of Neu-containing glycoconjugates [5 16,5171. Following similar routes, Kdn-containing oligosaccharides have also been synthesized [5 181. As a variation on this theme, the application of S-sialyl xanthates (SCSOEt) as donors in sialo-oligosaccharide synthesis has led to interesting results, including a high a-stereoselectivity [5 19-5221. Here, also the use of 0-benzoyl protection instead of 0-acetyl protection has been proposed [523,524]. Furthermore, the preparation of sialic acid S-glycosyl donors employing S,S’-bis( 1-phenyl- 1H-tetrazol-5y1)dithiocarbonate should be mentioned [525]. Another efficient donor combines parts of the structures F and G, yielding an a-thioglycoside with a SPh substituent at C3 (Fig. 9, structure H) [526]. The high stereoselective a-sialylation was obtained using either methyl sulfenyl bromide/silver triflate or N-iodosuccinimide/triflic acid as promoters. Sialyl phosphites with trimethylsilyl triflate as a promoter have additionally been shown to be of practical use (Fig. 9, structure I), affording good yields and a-stereoselectivity [527-5291, and examples include the synthesis of sialyl LeX sequences [528]. For a detailed study on the evaluation of different sialyl phosphites, see ref. [530]. For the organic synthesis of sialo-oligosaccharides with di- or trimeric Neu5Ac elements, specific glycosyl donors have been prepared directly from NeuSAc(a28)NeuSAc or Neu5Ac(a2-8)Neu5Ac(a2-8)Neu5Ac [53 1-5331. Treatment of the free oligosaccharides with H+-resin in methanol, followed by 0-acetylation and subsequent replacement of the anomeric acetoxy group by a phenylthio function yielded the corresponding peracetylated methyl ester phenyl 2-thioglycosides, in which Neu5Ac residues are linked via a (a2-8,l-9) lactone ring (Fig. 9, structure J for a disialosyl donor). For additional data with respect to the preparation of dimeric donors with structure A at the reducing site, see ref. [534]. In terms of preparative chemistry, the use of CMP-8-Neu5Ac as a glycosyl donor and (cloned) a-2,3/6-sialyltransferasesas biocatalysts have achieved a permanent position in the planning of synthetic routes for sialo-oligosaccharide chains; see e.g. refs. [3 14,342, 345,351,472,473,535-5441. Especially in the field of the preparation of sialyl LeX frag-

306

PEP Neu5Ac(a2-3)Gal(P1-4)[Fuc(al-3)]GlcNAc(P1-O)R GDP-Man

GDP-FUC

NADP

Neu5Ac(a2-3)Gal(PI -4)GlcNAc(pI-0)R

lranslerase

GlcNAc(p1-0)R

Gal(pl-4)GlcNAc(p1-O)R

x

4-epimerase UDP-Gal

(

c%pEp PYR

CMP-Neu5Ac

XCTP

"yKkPEpI

p1,4-Gal translerase

UDP-Gal

IM PPi

Neu5Ac

P Pase

PYR

UDP-Glc

2 Pi

ManNAc

XUTP UDP-Glc pyrophosphorylase

2 Pi-

P Pase

PPi

Glc(a1-O)P

Fig. 1 1. Chemo-enzymic synthesis of the sialyl Le" sequence [541]. NMK, nucleoside-monophosphate kinase; PEP, phosphoenolpyruvate; PK, pyruvate kinase; PPase, inorganic phosphatase; Pi, phosphate; PPi, pyrophosphate; PYR, pyruvate; TADH, Thermoanaerobium brockii alcohol dehydrogenase.

ments, excellent results have been obtained. To illustrate the enzymatic conversions, a series of typical examples will be presented. A first comprehensive example comprises the enzymatic sialylation on a microscale of oligosaccharides containing Gal(fil-3)GlcNAc, Gal(P14)GlcNAc, Gal(pl-4)Glc, and Gal(P1-3)GalNAc sequences by using different purified mammalian sialyltransferases [342], creating NeuSAc(a24)Gal, NeuSAc(a23)Gal, or NeuSAc(a2-6)GlcNAc linkages, as verified by 'H and 13C NMR spectroscopy (see section 5.3 3. The generation of the Neu5Ac(a24)GaINAc(a 1-0) sequence on a microscale has been described in ref. [536]. A second example (Fig. 11) is the chemo-enzymatic synthesis on a preparative scale of NeuSAc(a2-3)Gal(p 14)[Fuc(ctl3)]GlcNAc(P l-O)CH2CH=CH2 (and analogues) using P- 1,4-galactosyltransferase and recombinant a-2,3-sialyltransferase and a-1,3-fucosyltransferase with in situ regeneration of UDP-Gal, CMP-Neu5Ac and GDP-Fuc [541]. A third example is the one-pot enzymatic synthesis of Neu5Ac(a24)Gal(~l-4)GlcNAc (and analogues) based on a P-galactosidase-catalyzed galactosylation, using lactose as a donor and GlcNAc as

307

an acceptor, and a pig liver a-2,6-sialyltransferase-catalyzed sialylation with in situ regeneration of CMP-NeuSAc [542]. A fourth example is the enzymatic synthesis of NeuSAc(a24)Gal(fi 14)GlcNAc(fi l-O)pent-4-ene, a precursor for the organic chemical synthesis of higher oligosaccharides [472]. The trisaccharide was synthesized starting from GlcNAc(fi l-O)pent-4-ene, UDP-Gal (in situ generated from UDP-Glc catalyzed by UDP-Gal 4-epimerase), and NeuSAc in a one-pot reaction employing fi- 1,4-galactosyltransferaseand a-2,6-sialyltransferase in a complete cofactor regeneration system. The availability of specific sialyltransferases will certainly contribute to a further expansion of this area. In this context, the recent finding of a novel sialyltransferase which catalyzes the transfer of Kdn from CMP-Kdn to the non-reducing termini of oligo/polysialyl chains, thereby capping a further elongation of a (NeuSGc(a2-8)}, chain, is of interest [545]. In the framework of the finding that two NeuSAc(a2-6)Gal(fil4)GlcNAc units are the receptor determinants for the influenza virus hemagglutinin, these elements have been systematically anchored on a Gal residue in order to design structures capable of bimodal viral binding, and along chemo-enzymatic routes heptasaccharides with the general formula NeuSAc(a2-6)Gal(fi 14)GlcNAc(fil-x)[NeuSAc(a2-6)Gal(fi 14)GlcNAc(fil-y)]Gal(fi1-O)(CH~)~COOMe, where x and y are 2 and 3, 2 and 4, 2 and 6, 3 and 6, and 4 and 6, respectively, have been synthesized[546]. The concept of preparing compounds with a multivalent presentation of NeuSAc(a2-6)Gal(fi 14)GlcNAc((31- fragments on a linear or branched (via lysine) peptide backbone has been nicely worked out in ref. [547]. After the organic synthesis of a large series of peptide backbones in which GlcNAc((31-N)Asn units were incorporated, the oligosaccharide extensions were performed enzymatically by using fi- 1,4-galactosyltransferase and a-2,6-sialyltransferase. Neu5,9Ac2-containing oligosaccharides have been prepared along both organic chemical and enzymatic routes. In a reaction with trimethyl orthoacetate, NeuSAc(a2-6)Gal(fi14)Glc could be readily converted into Neu5,9Ac2(a2-6)Gal(filL4)Glc [380]. Employing CMP-Neu5,9Ac2 and immobilized porcine liver Gal((3lL4)GlcNAc a-2,6-sialyltransferase, Neu5,9Ac2(a2-6)Gal(fi lL4)GlcNAc has been synthesized [478]. The NeuSAc4Me-, Neu5Ac9Me-, and 8epi-NeuSAc-thioglycoside donors have been used to synthesize the corresponding sialoglycoconjugates [548], whereas also thioglycoside donors of 4d-, 7d-, 8d-, and 9d-NeuSAc have been prepared [549]. In other enzymatic approaches, use has been made of a fi-D-gaiactoside a-2,3-transsialidase from Trypanosoma cruzi as biocatalyst (see also section 9.2.3). This enzyme catalyzes the reversible transfer of NeuSAc from a donor-substrate of the sequence NeuSAc(a2-3)Gal(fi 1-O)Rl to virtually any Gal@ l-O)R2 acceptor substrate, affording a product NeuSAc(a2-3)Gal(fil-O)R2 [550-5521 (Fig. 12). A frequently used donor substrate is Neu5Ac(a2-3)Gal(fi1-4)Glc. In the case of Gal(fil-3/4)GlcNAc(fil-O)R sequences, the GlcNAc unit should not be substituted with a Fuc residue, as in LeX or Lea determinants [551]. The trans-sialidase reaction has been used in the chemo-enzymatic preparation of a water-soluble polyacrylamide, bearing multivalent NeuSAc(a2-3)Gal(filL4)GlcNAc elements [553]. To solve the problem of the poor enzymatic a-2,3-sialylation of Gal 2-(trimethylsily1)ethyl (3-glycoside using all known a-2,3-sialyltransferasesand CMP-NeuSAc, attention has been paid to the development of a sequence of enzymatic reactions, including cloned a-2,3-sialyltransferase and

308

b

z

I 0

r

0

8

I

0

-5

0

I

P

'

P

z

g

23 x

309

CMP-NeuSAc synthase, yielding an alternative active sialyl-donor-substrate in situ (e.g. NeuSAc(a2-3)1acto-N-tetraose), which can be used by trans-sialidase [554]. In this way it was possible to convert Gal 2-(trimethylsi1yl)ethyl P-glycoside into NeuSAc(a2-3)Gal 2-(trimethylsily1)ethyl 13-glycoside, a sialodisaccharide that can readily be transformed into a disaccharide donor, of interest for additional organic syntheses. In another study 4-MU-NeuSAc was tested as a donor with lactose as acceptor[555]. Starting from periodate treatedreductive aminated 4-MU-NeuSAc derivatives, interesting possibilities for the inclusion of fluorescent or photolyzable groups were demonstrated. Bacterial sialidases have also been explored in synthetic approaches. In a reverseenzyme reaction with A . ureafaciens sialidase, incubation of a concentrated solution of NeuSAc and lactose yielded NeuSAc(a24)Gal(P 1 4 ) G l c and Gal@14)[NeuSAc(a26)IGlc [556]. Similar experiments were carried out with immobilized K cholerae sialidase, using NeuSAc p-nitrophenyl a-glycoside as a donor. In this way NeuSAc(a2-x)Gal and NeuSAc(a2-x)Glc linkages could be produced, in which the a-2,6-linkage dominated over the a-2,3-linkage [557]. Transglycosylation of a NeuSAc unit using NeuSAc(a28)NeuSAc as a donor to Gal(P14)GlcNAc and Gal(b14)Glc was performed using sialidases of various origin [558]. Although the yields were low, a high regioselectivity was observed. The C. perfringens, A . ureufaciens and c! cholerae sialidases generated a-2,6-linkages, and the Newcastle disease virus sialidase a-2,3-linkages with the terminal Gal residue. For detailed information with respect to the synthesis of C-glycosides of sialic acids, see refs. [559-5631. As an example, the synthesis of a multivalent material that consists of the C-glycoside of NeuSAc, which is resistant to viral sialidase hydrolysis, should be mentioned [562].

7. Conformational aspects of sialic acids Earlier studies have appeared on the X-ray crystallography of both crystalline P-NeuSAc. H20 and B-NeuSAclMe*lHzO[56,564]. In an additional study, the crystal and molecular structure of a-NeuSAc1,2Me2 was also analyzed[565]. The C=O bond of the COOH function is approximately coplanar with the ring C-0 bond in a-NeuSAcl,2Me2, whereas in both P-NeuSAc and P-NeuSAclMe the C=O bond is found to be nearly eclipsed with the anomeric C-0 bond. In all three derivatives, the N-acetyl group is essentially planar, adopting the Z-conformation of a peptide bond. For a-NeuSAc 1,2Me2 and P-NeuSAclMe a hydrogen bond between the H-atom of H 0 7 and the carbonyl 0-atom of AcNHS was observed. The overall conformation of the glycerol side chain is the same for all three derivatives, as far as non-H atoms are concerned. In a-NeuSAc1,2Me2 a hydrogen bond between the H-atom of H 0 8 and the carbonyl 0-atom of the COOMe group is detectable. One of the oldest NMR studies on the conformation of NeuSAc is that focused on the spatial structure of aNeuSAc2Me in D20 [566,567]. On the basis of 'H-'H coupling constants in combination with 13C spinlattice relaxation times ( T I ) , a model could be constructed in which the amide H-atom of AcNHS is hydrogen-bonded to 0 7 , and the H-atom of H 0 8 is hydrogen-bonded to the ring-oxygen. A third hydrogen bond between the carbonyl 0-atom of AcNHS

310

and the H-atom of H 0 4 was suggested on the basis of molecular model building. In this model, apparently, the anomeric center is not involved in any hydrogen bonding, leading to the same conformation for a- and p-anomers. Independent of the models discussed above, the results fit the observation made by 'H NMR spectroscopy that H 0 7 and H 0 8 usually occur in a tuuns-orientation[53]. An NMR (H2O-suppressed 1D TOCSY, ROESY, NOESY) study, carried out on NeuSAc(a2-3)Gal(~14)GlcNAc and NeuSAc(a2-6)Gal(B 14)GlcNAc in 85% H20/15% (CD3)2CO, and aimed to detect hydroxyl and amido protons, indicated that in both compounds, thus irrespective of the type of linkage, the H-atom of H 0 8 of NeuSAc is involved in a strong intramolecular hydrogen bond [568]. In view of the fact that 7epi-NeuSAc and 7Jepi2-NeuSAc are substrates for CMP-sialic acid synthase (see section 6.1), but not 8epi-Neu5Ac [411], a conformational study on the side chain conformation of these sialic acids and NeuSAc itself (all
31 1

8. Biosynthesis of sialic acids 8.1. General Cells from higher animals and various microorganisms produce sialic acids in a long pathway starting from glucose [5,8,33]. Only mammalian erythrocytes, which have lost their nucleus during maturation, can no longer synthesize sialic acids, although they are still sialylated on their surface. The presence of sialic acids on the outer cell membrane is considered to be a prerequisite for the viability of mammalian cells, at least in tissues or fluids of organisms. After the loss of cell surface sialic acids by sialidase, these monosaccharides are rapidly restored [576]. Since the last reviews on sialic acid metabolism [5,33],little progress has been made on the enzymatic and regulatory mechanisms as well as on the molecular biology involved in the biosynthesis of N-acetylneuraminic acid (NeuSAc). However, further insight has been gained into the biosynthesis, degradation and role of modified sialic acids, especially N-acetyl hydroxylation, 0-acetylation and 0-methylation. An overview on the reactions involved, including catabolic enzymes, is shown in Fig. 13. The reactions, subcellular site, regulation, pathobiochemical role and molecular genetics of the enzymes activating sialic acids with CTP and transferring them onto nascent glycoconjugate molecules have also been intensively studied in animal and bacterial cells and much progress has been made. These new aspects of sialic acid metabolism will be discussed below. The regulation of sialic acid biosynthesis and consequently the sialylation of glycoproteins and gangliosides of cells seem to be governed in a rather variable manner, resulting in characteristic qualitative and quantitative differences during cell differentiation, growth, functional changes, ageing and malignant transformation of cells. Details will be given in section 10 on the biological role of sialic acids. In order to influence these processes, inhibitors of sialic acid biosynthesis or breakdown e.g. by sialidases are required. Various metal ions such as Zn2+, Cu2+ and selenite were shown to inhibit enzymes of sialic acid biosynthesis in rat liver homogenates [578]. They impair the activity of the key regulatory enzyme UDP-Nacetylglucosamine-2’-epimerase. In the same tissue, 3-O-methyl-N-acetyl-~-glucosamine efficiently inhibits N-acetylglucosamine and N-acetylmannosamine kinases involved in NeuSAc formation [579]. Consequently, the incorporation of N-acetylhexosamines into sialic acids in human hepatoma cells was reduced. After intraperitoneal injection of N-propanoybhexosamines, N-propanoylneuraminic acid could be isolated from tissues at different relative amounts, showing that the enzymes for sialic acid biosynthesis and transfer can tolerate changes in the N-acetyl moiety of precursor acylhexosamines [580]. Other inhibitors specific for sialic acid enzymes will be mentioned as appropriate in the text. 8.2. Biosynthesis of CMP-sialic acids The enzyme CMP-NeuSAc synthase, which is also known as NeuSAc cytidylyltransferase (EC 2.7.7.43), is ubiquitous in pro- and eukaryotic cells synthesizing sialic acids from glucose [5,33]. In contrast to other nucleotide-sugar-synthesizing enzymes, it is located

Neu5Gc8Me8

Hexose

CMP-Neu5Gc

4

eNeusGc I 6 Golgi

'

8

0

Neu5Gc

lysosomes plasma membranes etc.

Neu5Ac

CMP-Neu5Ac

Golgi

8

Neu4,5Ac2

8 Neu5Ac

pyruvate

ManNGc

cytosor (?)

Neu5Ac

ManNAc

d

Neu5,7(9)Ac2

Fig. 13. Metabolism scheme of sialic acids. Anabolic (solid arrow) and catabolic (dashed arrow) reactions are indicated. For literature see the text. Enzymes: 1, CMP-sialate synthase (EC 2.7.7.43); 2, sialyltransferases (EC 2.4.99.1. . . . ); 3, CMP-Neu5Ac hydroxylase (EC 1.14.99.18); 4, acetylCoA:sialate 4-0-acetyltransferase (EC 2.3.1.44); 5, acetyl-CoA:sialate 7(9)-0-acetyltransferase (EC 2.3.1.45); 6 , S-adenosyl-L-methionine:sialate 8-0-methyltransferase (proposed EC 2.1.1.78); 7, sialate 4-or 9-0-acetylesterases (EC 3.1.1 S3); 8, sialidase (EC 3.2.1.1 8); 9, sialate-pyruvate lyase (aldolase; EC 4.1.3.3). Both NeuSAc and Neu5Gc can be 0-acetylated by the two 0-acetyltransferases. There may also exist a sulfotransferase, since sulfated sialic acids have been found in e.g. echinoderms [ 13,5771. ( a ) , Sialic-acid-accepting nascent glycoconjugate.

313

in the nucleus of mammalian cells, although it has not yet unequivocally been shown that CMP-NeuSAc is really produced in this compartment [%I]. It has been isolated from both animal and bacterial sources [5,33,582], and particularly large activities exist in frog [583] and trout liver [582], as well as in Escherichia coli [584,585]. The latter enzyme, which is expressed in E. coli K1 and K92 only at temperatures higher than 20°C, has been cloned and sequenced [586] and its molecular properties have been described. It exhibits 59% homology with the corresponding, cloned gene from Neisseria meningitidis [587, 5881. The availability of this enzyme is increasingly important, since CMP-NeuSAc is required not only for sialylation of natural substances, but also for the combined chemical and enzymatic synthesis of sialylated complex carbohydrates in research and glycotechnology (see section 6.2 for details and ref. [S89]). With the synthases from different sources it is possible to activate not only NeuSAc, but also NeuSGc, 0-acetylated sialic acids and a variety of chemically prepared sialic acid derivatives (section 6.2). The transfer of non-natural sialic acids onto glycoconjugates may be useful for the study of sialic-acid-degrading enzymes and for various aspects of cell biology like cellular interactions and fluorescent labelling of cell surface glycoconjugates by CMP9-fluoresceinyl-NeuSAc [390,4 191. The enzyme from trout also activates Kdn occurring as oligomers in this animal [ 107,1221. With regard to the reaction mechanism it has been shown with the E. coli synthase and 13C-labelledNeuSAc in NMR studies that CMP-fi-NeuSAc is probably formed by a direct transfer of the anomeric oxygen of fi-NeuSAc to the a-phosphate of CTP [584]. Activation of NeuSAc is only possible, if the hydroxyl groups at C4 and C8 are in the correct, natural conformation [350]. Correspondingly, 4epi- and 8epi-NeuSAc cannot be linked to CMP by the rat liver cytidylyltransferase. Furthermore, the enzyme requires the presence of a hydroxyl group at C8, since 8d-NeuSAc is inactive. This is in contrast to 4d-NeuSAc and 9d-NeuSAc. Substituents at C4, such as in NeuSAc4Me or Neu4,SAc2, are tolerated. The strong competitive inhibition effect of 6-NeuSAc methyl glycoside shows that the H 0 2 group interacts with the enzyme via the lone electron pair of the oxygen atom and not by a hydrogen bridge. Little information is available on the active site of this enzyme. In the recombinant enzyme from E. coli two cysteine residues were recognized at the positions 129 and 329, one of which seems to be at or close to the active center, as was shown by inhibition with thiol-specific reagents partly in the presence of CTP. However, exchange of these cysteine residues for other amino acids by site-directed mutagenesis did not change the enzyme activity significantly [590]. Several cytidylyltransferases were shown to be inhibited by CMP and CDP [S82]. The observation that 2'-CMP, 3'-CMP and 2',3'-cyclic CMP do not inhibit shows that the phosphate residue at the 5' position of ribose is required for inhibition. These studies also demonstrate that the ribose moiety must be intact for inhibition and that chemical modifications of the cytidine residue tested do not lead to inhibitors. An inhibitory effect of the methyl fi-glycoside and the 4-0x0-derivative of NeuSAc on the cytidylyltransferase from rat liver was described [408].

314

8.3. Transfer of sialic acids Specific patterns of oligosaccharides as components of glycoproteins and glycolipids cover the surface of individual cell types and may change during variation of the function or differentiation of these cells. This “make up” of cells, which is formed by the expression of a set of glycosyltransferases [335,59 1-5941 gains enormous diversity by the transfer of sialic acids into mostly terminal position of the oligosaccharide chains and in different linkages (see Tables 2-6). This great variety of structures may be even more enlarged by the modification of sialic acids, i.e. N-acetyl-hydroxylation, 0-acetylation, 0-lactylation, 0-methylation or 0-sulfation (see Table 1 for natural sialic acids). Many studies have been carried out in the past few years on the transfer of sialic acids onto oligosaccharides, polysaccharides, glycoproteins and glycolipids mediated by sialyltransferases of animal cells and bacteria. These have given new insight into the wide occurrence and subcellular distribution of these enzymes comprising a family of over 10 members, their substrate specificities yielding a-2,3-, a-2,6-, a-2,8- and a-2,9linkages, regulation of gene expression, and especially primary structures. Since these topics have been summarized in refs. [593,595] and are mentioned in several chapters of volume 29a of New Comprehensive Biochemistry series, here only a selection of some aspects of these enzymes mainly from the most recent literature will be reported and a list of cloned sialyltransferases (Table 14) is included. The isolation of CMPNeu5Ac:Gal(f314)GlcNAc a-2,6- and Ga1((31-3/4)GlcNAc a-2,3-sialyltransferasesfrom rat liver (ST6Gal I and ST3Gal 111, respectively) by affinity chromatography has been described [624]. The sialyltransferase assay is usually carried out with radioactive CMP-Neu5Ac and the sialic acid acceptor, followed by identification of the radioactive product. As this product is often membrane-bound and the radioactive cosubstrate is expensive, a non-radioactive assay was developed [625], in which CMP released from CMP-NeuSAc by the transfer of Neu5Ac is determined with HPLC (see section 5.3.2). Since CMP-Neu5Ac can be analyzed in the same run, this test is also suited for the activity of the CMP-NeuSAc synthase described above. With regard to acceptor specificity, the oligosaccharide sequences accepting sialic acids by the action of sialyltransferases are summarized in Tables 2-6. Although sialyltransferases exhibit strong linkage specificity, in some cases they seem to be involved in the sialylation of not one but several glycoconjugates [595]. In mitochondria1 outer membranes, there exist two sialyltransferases with galactoside a-2,3- and a-2,6specificities, which preferably sialylate 0- and N-glycans, respectively, of glycoproteins [626]. CMP-Neu5Ac:Gal(f31~)GlcNAc-R-a-2,6-sialyltransferase(ST6Gal) was found to sialylate the disaccharide unit GalNAc(f3I4)GlcNAc-R N-glycosidically linked to a variety of milk glycoproteins [627]. It is concluded from these studies that the same oligosaccharide moiety of a number of human glycoproteins is a-2,6-sialylated in a corresponding way by one sialyltransferase. With regard to tissue specificity, a-2,3- and a-2,6-sialyltransferase activities specific for Gal@14)GlcNAc-R residues occur in both human liver and placenta [628]. In contrast to the acceptor specificity, the donor specificity of sialyltransferases is less pronounced, although differences exist. They can transfer not only Neu5Ac but

315

also other sialic acids from their CMP-glycosides, e.g. both the natural Neu5Gc and 0-acetylated species as well as synthetic Neu5Ac derivatives, which have been mentioned in sections 6.2 and 8.2. This is of great importance for cell biological studies. In order to gain more insight into the structural requirements of the acceptor for rat liver sialyltransferases, the residues in the acceptors Gal@1-3)GlcNAc and Gal((314)GlcNAc were modified mainly by deoxygenation of the pyranose rings to identify the key polar groups required for transfer[629]. Both the H 0 6 group of the P-Gal residue and the N-acetyl group of the GlcNAc unit are required for the activity of a-2,6-sialyltransferase (ST6Gal I). The a-2,3-sialyltransferase (ST3Gal 111) requires the H03, H 0 4 and H 0 6 groups of terminal B-Gal, and some influence from the subterminal sugars was noticed. Structural variants of the Gal residue can also be recognized by rat liver Gal(B14)GlcNAc a-2,6-sialyltransferase. Besides Gal@14)GlcNAc(P 10)Me and GalNAc((314)GlcNAc(Pl-O)Me, also Man((314)GlcNAc((31-0)Me, Glc(P14)GlcNAc(Pl-O)Me and GlcNAc((314)GlcNAc((3l-O)Me turned out to be acceptors, as proven by 'H NMR spectroscopy[630]. In a more detailed systematic study the substrate specificity of a-2,3- and a-2,6-sialyltransferases was explored using a series of synthetic Hex( 14)GlcNAc((31-2)Man(a l-O)(CH2)7CH3 acceptors with Hex = P-D-Gal, 4-deoxy-P-~-Gal, 4-O-methyl-(3-~-Gal,4-deoxy-4-fluoro-P-~-Gal, P-D-GlC, 3-0-methyl(3-D-Gu1, P-D-Gal, 3-deoxy-P-~-Gal,3-deoxy-3-fluoro-(3-~-Gal,3-amino-3-deoxy-P-~-Gal, a-L-Alt and P-L-Gal [630a]. These reactions are relevant to the finding of unusual sialicacid-containing structures, namely NeuSAc(a2-6)Man(P 14)GlcNAc in the urine of patients with 8-mannosidosis [ 1531. They now enable the synthesis of unusual sialic-acidcontaining carbohydrate sequences. Sialyltransferases may be glycoproteins and the carbohydrate chain can modify enzyme activity. Thus, deglycosylation of rat liver Gal((314)GlcNAc a-2,6-sialyltransferase protein markedly decreases enzyme activity, showing that the trimannosyl-N,N'-diacetylchitobiose core with GlcNAc attached is particularly important for the expression of catalytic activity [63 11. Glycosylation of a-2,6-sialyltransferase from rat hepatoma cells with N-acetyllactosamine- and oligomannose-type sugar chains was also reported [632]. The polypeptide part of an acceptor glycoprotein may also influence the activity of sialyltransferase, as was shown for desialylated human chorionic gonadotropin and bovine colostrum CMP-Neu5Ac:Gal((314)GlcNAc-R a-2,6-sialyltransferase [633]. When compared with the heterodimeric form of this hormone, the glycans of the a-subunit exhibited a higher kinetic efficiency (V/KM). In contrast, the branch specificity of this enzyme, i.e. the preference for the Ga1((314)GlcNAc((31-2)Man(al-3)Man branch for attachment of the first sialic acid molecule into the diantennary glycans of desialylated gonadotropin was not influenced. Interestingly, evidence was obtained that a-2,3-sialyltransferase from cultured C6 glioma cells [634] must be phosphorylated for activity and it is postulated that enzyme activity is regulated by the action of Ca2+/calmodulinand phosphatase, respectively. We are only beginning to understand the regulation of the expression of sialyltransferase activity. The higher level of sialylation of many tumor cells and increased sialyltransferase activities are well known (see chapter 3 of Vol. 30 and section 10.5 below]. Transformation of FR3T3 cells with the c-Ha-ras oncogene resulted in a marked increase of the expression of (3-galactoside a-2,6-sialyltransferase activity and,

Table 14 Sialyltransferases with known gene structuresa Abbreviations

Structures synthesized

Tissue

Ref.

Comments

Mammalian enzymes

ST3Gal I

porcine submand. gland mouse brain

ST3Gal I1

chick embryo mouse and rat brain

ST3Gal 111

rat liver

acts on 0-linked chains and glycolipids

similar to ST3Gal I, but seems to act better on glycolipids than ST3Gal I; kinetics see ref. [600] acts preferentially on Gal@-3)GlcNAc

human placenta mouse brain ST3Gal IV

Sia(a2-3)Gal(fil-3)GalNAc-Rand Sia(a2-3)Gal(fi1-4)GlcNAc-R

ST6Gal I

human melanoma, placenta, mouse brain rat liver human placenta

also called STZ or SAT-3, acts preferentially on 0-glycans [Gal(pl-3)GalNAc] and on N-glycans [Gal(fi14)GlcNAc], also on glycolipids acts on N-glycans, possibly also on lacto-neo-glycolipids

mouse liver chick embryo ST6GalNAc 1

Sia(aZ4)GalNAc-R

chick embryo mouse submand. gland chick testis human epithelial cells

ST6GalNAc I1

requires Gal((31-3)GalNAc as substrate, does not act on GalNAc-R

mouse ST8Sia I

Sia(aZ4)Sia-R (glycolipid)

human melanoma mouse brain

synthesizes GD3 >> GT 1a and GQ 1b continued on next page

Table 14, continued Abbreviations

Structures synthesized

Tissue

STSSia I1

Sia(a24)Sia-R

rat brain

also known as STX, specificity see ref. [618]

human fetal brain

human homologue to STX [617], partial cDNA-clone murine homologue to STX [617], acts on N-glycans

mouse brain ST8Sia I11

Sia(a24)Sia-R

mouse brain

ST8Sia IV

Sia(a24)Sia-R (NCAM)

Chinese hamster fibroblasts

STlSia V

Sia(a24)Sia-R

mouse brain

PST

Sia(a24)Sia-R on Sia-polymers

Neisseria rneningitidis

PST

Sia(a24)Sia-R on Sia-polymers

Escherichia coli K1

PST

Sia(a24)Sia-R and Sia(a2-9)Sia-R on Sia-polymers

Escherichia coli K92

mouse lung

Ref.

Comments

acts on N-glycans, synthesis of GD3 and GT3 synthesis of polysialic acid on NCAM acts on a-2,3-linked Sia of N-glycans acts on glycoprotein

Bacterial enzymes

group B synthesis of alternating a-2,8- and a-2,9-linkages

With regard to nomenclature, ST stands for sialyltransferase, 3, 6 and 8 for the linkage formed, Gal or Sia for the monosaccharide carrying the acceptor hydroxyl group, and Roman numbers (I-V) for the chronological order of published primary structures. The species designation should be given as a single letter at the beginning (e.g. pSTGal 1 for the porcine, mSTGal I for the murine and cSTGal I for the chicken enzyme). This nomenclature follows suggestions by Tsuji et al. [623a]. PST stands for bacterial polysialyltransferases. In the structures synthesized, Sia is used, since the sialyltransferases may transfer not only NeuSAc, but also other natural as well as synthetic sialic acids (see the text). As an example, the systematic name of ST3Gal I is CMP-NeuSAc:GaI(~l-3)GalNAc a-2,3-sialyltransferase (EC 2.4.99.4). However, this name would also be used for ST3Gal I1 and ST3Gal IV

a

318 Table 15 Natural and synthetic inhibitors of sialyltransferases"

Ki values (mM)

Inhibitor ST6Gall

ST3GalIC

5'-CMP

0.090

0.064

5'-CDP'

0.050

Cytidine

0.050 0.046 0.13

2-Thiocytidine

0.15

CMP-dialdehyde

3.3

5.7

UDP-dialdehyde

n.t.

n.t.

5'-CTP

6'-deoxy-lactosaminide

'

Ref(s)

Lactosyl-ceramide a-2,3-sialyltransferase

0.060 22 n.i.g

0.76

6'-thio-lactosaminide

3.78

6'-disulfide-lactosaminide

2.00

Protein (20 kDa; calf brain)

8O%J

[642,645]

Protein (14.8 and 22.4 kDa; rat brain)

50%

[642,646]

For sialyltransferase nomenclature see Table 14. From rat liver. From porcine submandibular gland. From embryonic chicken brain. n.t., not tested. This compound inhibits b-galactoside a-2,6-sialyltransferase from bovine colostrum with a Ki of 0.025 mM [647]. g mi., no significant inhibition. a

Obtained by periodate oxidation of the ribose moiety. i Lactosaminide stands for the methyl glycoside of N-acetyllactosamine. J This inhibition was observed at 20 pg/ml[642]; ovine submandibular gland CMP-Neu5Ac: GalNAc a-protein a-2,6-sialyltransferase was inhibited to a similar extent [645]. Inhibitor concentration 20 pg/ml.

'

correspondingly, of the a-2,6-sialylation of cell surface glycoconjugates [635]. Sialylation may also be increased by hormones, e.g. interferon [636], although in this publication the activity of sialyltransferase was not described. However, in the rat intestine hydrocortisone leads to an increase of sialyltransferase activity during tissue maturation [637]. Similar observations were made with rat fibroblast FR3T3 cells [638]. Cytokines were found to induce the expression of 6-galactoside a-2,6-sialyltransferase in human endothelial cells, which mediates sialylation of adhesion molecules and CD22 ligands [639]. Based on mRNA measurements using in situ hybridization histochemistry, it was found that the expression of (3-galactoside a-2,6-sialyltransferase in rat retina increases in the early post-natal days but to different degrees in the different retinal cell types[640]. Five different sialyltransferases were described to be involved in the synthesis of specific terminal sialoside structures of human fetal and adult tissues, and each gene of these enzymes seems to be regulated independently [619]. During the development of trout eggs a-2,6-, a-2,8-sialyl- and a-2,8-polysialyltransferase activities required for the synthesis of polysialoglycoproteins were expressed at various levels [64 I].

319

Little is known about inhibitors of sialyltransferases, although such compounds are very interesting for the modulation of sialylation, e.g. to study tumor biology and cancer therapy. Various sialic acids and their derivatives, as well as other synthetic or natural substances were tested as inhibitors (refs. [642,642a], Table 15). The Ki values of periodatetreated UDP and CMP, resulting in two aldehyde groups from the ribose ring each, which inhibit CMP-Neu5Ac:lactosylceramide a-2,3-sialyltransferase in microsomes of embryonic chicken brain in an irreversible, competitive manner [643], are also presented. While the natural 5’-CMP nucleotide is a potent inhibitor of rat liver and porcine submandibular gland sialyltransferases, the 3’-CMP isomer has no inhibitory effect. Strikingly, cytidine itself and its 2-thio-derivative also markedly inhibit the a-2,6- and a-2,3-sialyltransferases tested. Other substituents on cytidine both on the ribose and base moiety abolish this inhibitory effect. Oxidative opening of the ribose group of CMP significantly reduces the inhibitory potency of CMP on the liver and submandibulary gland enzymes, but yields a good inhibitor for the chicken brain sialyltransferase. The interesting compound 5-fluoro-2‘,3’-isopropylidene-5‘-O-(4-N-acetyl-2,4-dideoxy-3,6,7,8-tetraO-acetyl- 1-methoxycarbonyl-D-glycero-cr-D-galucto-octapyranosyl)uridine (KI-8 1 10), a CMP-Neu5Ac analogue, has been synthesized and inhibits the sialylation of cells in culture and pulmonary metastasis of mouse colon adenocarcinoma [648,649]. Although this compound does not inhibit the pure sialyltransferases from rat liver and porcine submandibular gland, it seems to inhibit the translocation of CMP-NeuSAc through the Golgi membrane and thus sialylation of nascent glycoconjugates, as was observed with mouse and trout microsomes [642] as well as with Golgi vesicles from human liver and colorectal cancer cells [650]. As consequence of reduced sialylation, the growth of Sindbis viruses in cell culture was much hindered (M. Odenthal-Schnittler et al., unpublished observations). While these inhibitors are on the donor site of sialyltransferase substrates, the effect of inhibiting substances on the acceptor site of NeuSAc has also been reported [644]. For example, the methyl glycosides of 6‘-deoxy- and 6’-thio-N-acetyllactosamine, which inhibit rat liver a-2,6-sialyltransferase in the millimolar range, are shown in Table 15. In all these cases, the inhibition does not follow pure Michaelis-Menten kinetics, but is of mixed competitivehon-competitive type. Natural sialyltransferase inhibitors have also been found, which are considered to play a role in the regulation of the synthesis of sialoglycoproteins and gangliosides. Whereas heparin and other polyanions are weak inhibitors, proteins from rat and calf brain have been isolated which inhibit a-2,6-sialyltransferases from ovine submandibular gland and rat liver by about 80% at a concentration of 20pgimol [641,645]. Two proteins of molecular mass 14.8 and 22.4 kDa isolated from rat brain inhibit both CMP-Neu5Ac:lactosylceramide a-2,6sialyltransferase [646] and rat liver a-2,6-sialyltransferase [642]. This inhibitor also occurs in chicken and bovine brain as well as in other tissues of these animals and of rat [646]. Many enzymatic and histochemical (see refs. [651,652] and chapter 5 of Vol. 29a, “Glycoproteins”) studies have shown that sialylation occurs in the trans-Golgi network [653] as a terminal glycosylation step after the transfer of CMP-sialic acid from the cytosol through the Golgi membrane by a specific translocator (see also Fig. 15 below). This protein was shown to transfer both CMP-Neu5Ac and CMP-NeuSGc equally well [654]. A detailed study on the glycosylation including sialylation of oligosaccharides

320

in rat liver Golgi preparations was reported in refs. [655,656]. In rat liver hepatoma cells, the glycosylation of a-2,6-sialyltransferase is involved in the maturation of the enzyme to a higher molecular mass form [632]. The enzyme follows the secretory pathway as a membrane protein and is retained at a late Golgi stage. Due to rapid degradation in a post-Golgi compartment, the enzyme, exhibiting a half-life time of 3 h, is secreted from the cell in only small amounts. For the retention of the sialyltransferase in the trans-Golgi network, signals in the enzyme’s polypeptide chain were suggested to be responsible for the specific binding to Golgi membranes. It was shown with mutant and chimeric a-2,6-sialyltransferase that the cytoplasmatic tail and signal anchor region alone are not sufficient for Golgi retention [657]. However, when two lysine residues were placed next to the signal anchor on the luminal side of the enzyme, more efficient Golgi retention was observed. This suggests that the signal anchor region is not sufficient for Golgi retention and that it can be replaced by any transmembrane region which allows correct spacing and folding of sequences flanking the membrane. So far, the acceptor specificity of nearly 20 eu- and prokaryotic sialyltransferases has been elucidated and 15 different cDNA clones of these enzymes have been obtained (Table 14 and references therein). This list demonstrates the rapid progress in this field and also includes the first cloning of mammalian polysialyltransferase (polysialyltransferase- 1 from hamster ovary cells) [620]. The same group has recently published the molecular analysis of the biosynthetic pathway of the a-2,8-polysialic acid capsule by Neisseria meningitidis serogroup B [658]. The sialyltransferase clones contain a stretch of about 50 amino acids with significant homology and this so-called sialyl motif is possibly involved in substrate binding [601]. Using polymerase chain reactions with degenerate primers deduced from this motif, additional cDNA clones have been obtained [6 171, for which, however, no sialyltransferase activity can yet be assigned. Sialyltransferases, like other glycosyltransferases, are structurally highly related, consisting of a short NH2-terminal cytoplasmic domain, a signal-membrane anchor domain, a proteolytically sensitive stem region, and a large COOH-terminal catalytic domain. For example, the deduced amino acid sequence of the mouse brain Gal(@1-3)GalNAca-2,3-sialyltransferase shows 80% identity with that of the porcine submandibular gland Gal(@1-3)GalNAca-2,3-sialyltransferase [597]. 8.4. Enzymatic modijication of sialic acid 8.4.1. Biosynthesis and functions of N-glycolylneuraminic acid NeuSGc is present in essentially all animal groups of the deuterostomate lineage, from the echinodermata up to the mammals, as component of all types of glycoconjugates [5,8,33, 659,6601. Remarkably, in healthy human tissues it is missing, but it is expressed in small quantities in some tumors [661]. Although NeuSGc has never been found in bacteria, it is a component of membrane glycoconjugates of Trypanosoma cruzi [662]. The wide occurrence of this sialic acid and its tissue-specific and developmentally regulated expression suggest that it has important biological functions. However, very few established functions have yet been allocated to NeuSGc. Our knowledge of this field is summarized in Table 16. It shows that in addition to structural aspects, the presence of NeuSGc modulates the general functions of the main, precursor sialic acid, NeuSAc.

32 1 Table 16 Biological significance of N-glycolylneuraminic acid Phenomenon

Reference(s)

Influence on physicochemical properties of glycoconjugates, e.g., increasing hydrophilicity

[51

Slower hydrolysis rate by sialidases when compared with NeuSAc, thus possibly hindering spreading and virulence of bacteria

[5,33,245]

Slower degradation rate by sialate-pyruvate lyase

[5,33,2451

Slower rate of demasking of subterminal galactose residues, thus prolonging the biological effect of sialic acids

~6631

Modulation of sialic acid-receptor interactions, e.g., increase of the binding of murine B lymphocyte CD22 adhesin and of Escherichiu coli K99 lectin; decrease of the binding of mouse macrophage sialoadhesin and Myelin-Associated Glycoprotein (MAG)

[664-6671

Specific antigenic epitopes, such as Hanganutziu-Deicher antigens

[668-6701

Dog and cat blood group determinants

[671473]

Tumor-associated antigen in man and chicken

[661,670]

Differentiation marker

[307,6741

The N-glycolyl group can be an acceptor site of 0-acetylation and O-methylation, yielding Neu5GcAc and Neu5GcMe (Table 1) in rat thrombocytes and starfish, respectively

[43,441

The N-glycolyl group can be a site of glycosylation, leading to (a) Neu5Gc8Me(a2-05)Neu5Gc8Me- chains in the starfish Asferius rubens (b) Branching points of glycans in echinoderm glycoconjugates

[6751 [5,13,3 1,6761

This is best illustrated with sialidases, the action of which is suppressed by the presence of a N-glycolyl group in sialic acids. This may afford some protection for host cells from attack by pathogen-derived sialidases, since the removal of sialic acids facilitates the further degradation of cells by other glycosidases and proteases [245]. Neu5Gc may therefore effectively act against the virulence of certain infectious bacteria. It is also conceivable that Neu5Gc may mask subterminal galactose residues of oligosaccharide chains thus preventing their recognition and phagocytosis by macrophages more potently than does Neu5Ac 116631. Other cell recognition phenomena, such as binding of sialic acids to sialoadhesin of mouse macrophages [664] or to the enterotoxic E. coli strain K99, which infects young pigs [665], is modulated by Neu5Gc. Since the Neu5Gc epitope on glycoconjugates is antigenic to man and leads to Hanganutziu-Deicher antibodies [668,677], the clinical application of recombinant glycoproteins, e.g. erythropoietin (EPO), containing small amounts of Neu5Gc due to

322 N A D H x ~ ~ ~ ~ Z ~ ~ ' L ~ ; xCytochrome

NAD+

Cytochrome b5 reductase (red)

b5 ( r e d x Hy droxy lase (ox)

Cytochrome b5 (ox)

Hydroxylase (red)

CMP-Neu5Gc + H 2 0

CMP- NeuSAc + O2

Fig. 14. Redox-protein compounds involved in the formation of NeuSGc by CMP-NeuSAc hydroxylase [ 188, 680,6811.

their production in cells such as CHO cells, requires special attention [669]. Formation of such antibodies would lead to rapid removal of the drug from the blood stream and the immuno-complex accumulated in renal glomeruli may cause nephritis or other symptoms known as serum sickness. To the authors' knowledge, however, no such problems with recombinant glycoproteins have been reported. While it was already detected in 1968/69 that NeuSGc is derived from NeuSAc by hydroxylation of the N-acetyl moiety [678,679], it was not until 1988 that CMPNeuSAc was recognized as the substrate for this hydroxylase [ 1881. This fact, together with the solubility of the enzyme following tissue homogenization stimulated purification and characterization of the enzyme from porcine submandibular gland [680] and mouse liver [68 1,6821. The enzyme from both sources is a monomer with a molecular mass of 6SkDa. The CMP-NeuSAc hydroxylase was also studied in the gonads of the starfish Asterias rubens, which belongs to the phylogenetically oldest animals synthesizing NeuSGc [683]. It has many characteristics in common with the mammalian enzymes. All hydroxylases studied so far show high affinity for CMP-NeuSAc [660], i.e. apparent KM values in the low micromolar range (e.g. 1.3 pM for the mouse liver enzyme [684]). They require a number of cofactors for activity including molecular oxygen, reduced pyridine nucleotides, and, as was shown for the mammalian enzymes, cytochrome b5, and cytochrome b5-reductase. These form an electron transport chain resulting in the incorporation of one oxygen into the methyl residue of the N-acetyl group, as shown in Fig. 14. Some experimental evidence was presented[685] that at the beginning of this concerted action the binding of CMP-NeuSAc to the hydroxylase changes the conformation of the enzyme in a way leading to the recognition and binding of the enzyme by cytochrome b5 and the formation of a ternary complex. After the transfer of electrons from NAD(P)H to the enzyme through cytochrome b5, CMP-NeuSAc is converted to CMP-NeuSGc which was shown not to bind to the enzyme and is therefore released with concomitant dissociation of the ternary complex. The observation that the addition of iron salts can stimulate enzyme activity and iron-binding substances are potent inhibitors, together with electron-spin-resonance studies of the purified hydroxylase, strongly point to the presence of an iron-sulfur center in the active site, which is possibly of the Rieske type [686]. The CMP-NeuSAc hydroxylases from mouse liver [687] and pig submandibular gland [688] have both been cloned. The gene structures show high homology. In addition to the CMP-NeuSAc-specific hydroxylase, pig submandibular glands were reported to contain a second hydroxylase, which is specific for free NeuSAc and also

323

\

Gal

SIALYLTRANSFERASE

Gal

I

NeuSGc

CMP

CMP

CMP-Neu5Gc

CMP-NeuSGc CMP-Neu5Ac HYDROXYLASE

t

02, NADH, Fe ion

CMP-NeuSAc

Fig. 15. Biosynthesis of CMP-N-glycolylneuraminic acid with the aid of CMP-N-acetylneuraminic acid monooxygenase in the cytosol, translocation of CMP-Neu5Gc into the Golgi vesicle and transfer of the Neu5Gc moiety onto nascent glycoconjugates [ 188,654,6841. From ref. [690] by permission of Oxford University Press, Oxford.

differs from the former hydroxylase with respect to its sensitivity to certain inhibitors [689]. The activity of CMP-NeuSAc hydroxylase and accordingly the concentration of CMPNeu5Gc in the cytosol probably play the most important role in regulating the level of sialylation with Neu5Gc. This is assumed because neither the Golgi CMP-sialic acid antiporter nor the sialyltransferases exhibit a pronounced preference for CMP-Neu5Ac or CMP-NeuSGc (Fig. 15). The activity of the hydroxylase may thus be tuned so that the ratio of Neu5Gc and Neu5Ac required in the resulting glycoconjugates is generated in the form of CMP-glycosides in the cytosol. The multiplicity of functions and constitutive expression of the cytochrome b5 system, together with the lack of effect of any metabolites

324

on the hydroxylase activity [68 13, suggest that the rate of production of CMP-NeuSAc is regulated at the level of the expression of the monooxygenase. The tissue-specific and developmental factors affecting the production of this protein on the gene level as well as a possible influence of the oxygen pressure still remain to be elucidated. No experimental explanation has so far been obtained for the formation of NeuSGc in human tumors, which may be caused by the anomalous expression of a CMP-NeuSAc hydroxylase gene, which seems to be dormant or repressed in normal tissues. Significantly, in T-cell lymphomas of baboons the level of NeuSGc, as component of gangliosides, was much higher than in those from the corresponding normal tissues [691]. The fusion of human B lymphocytes with mouse myeloma cells leads to high expression of NeuSGc, showing activation of the enzymes involved in NeuSGc biosynthesis due to the somatic cell fusion process [692]. It was furthermore reported that the insertion of retroviruses into Chinese hamster ovary cells almost completely replaces NeuSAc by NeuSGc [693].

8.4.2. Biosynthesis and functions of 0-acetylated sialic acids 0-Acetylated sialic acids are found in all types of glycoconjugates and in oligo- and polysaccharides of many animal species from the echinoderms onwards and also in some bacterial species (refs. [5,7,8,11,13,27,690,694]; see also sections 2 and 3, and Table 1). This modification is either at position 4 of the pyranose ring of both NeuSAc and NeuSGc, as was found in horse, donkey, guinea pig and echidna (Tachyglossus aculeatus), or in the glycerol side chain of sialic acids. In the latter case, the 0-acetyl group is most frequently located at C9, but it may be accompanied by acetylation at C8 and/or C7, leading to di- and tri-0-acetylated species. This side-chain 0-acetylation is very frequent in the animal kingdom and is also found in tissues and fluids of man. However, in the expression of 0-acetylated sialic acids ethnic differences do exist in colonic mucin of Sino-Japanese and non-Sino-Japanese races. 0-Acetylation was more frequent in the latter ethnic groups studied (British, Icelanders, South African blacks and Bahrainis) than in Chinese and Japanese. It is speculated that loss of 0-acetylation may be due to a single mutation of the gene regulating 0-acetylation and that this may be related to selection pressure resulting from differential enteric colonization by bacterial flora and from resulting infectious diseases [695]. The occurrence of differently 0-acetylated sialic acids suggests an important biological role of these modifications. We are only just beginning to understand these effects, which influence many biological and pathobiochemical systems, as listed in Table 17. As with NeuSGc, 0-acetylation may influence the physicochemical properties of sialylated glycoconjugates, which is especially important on cell surfaces. Correspondingly, an increased protective effect of the mucin containing a large quantity of 0-acetylated sialic acids and lining the endothelia of human colon is discussed [245]. This protection against an aqueous environment containing a large quantity of different bacteria may be due firstly to the greater hydrophobicity imparted by the 0-acetyl groups and secondly to a greater resistance of 0-acetylated sialic acids to microbial sialidases when compared with NeuSAc. Sialic acid 0-acetylation not only prevents (in the case of 4-0-acetylation) or much reduces (in the case of side-chain 0-acetylation) the activity of viral, bacterial and animal (including trypanosomal) sialidases, but also impairs the activity of acylneuraminate-

325 Table 17 Effects of 0-acetyl groups in sialic acids Phenomenon

Reference(s)

Influence on physicochemical properties of glycoconjugates, e.g. increasing hydrophobicity

PI

Hindering biologically active sialoglycoconjugates from degradation, by inhibition of sialidase, trans-sialidase, endoglycosidase and lyase action

[5,33,660,696]

Hindering cell degradation, e.g., of erythrocytes

[5,690,697]

Hindering binding of influenza A and B viruses, reovirus, Plasmodium falciparum, mouse macrophage sialoadhesin, B cell adhesion molecule CD22

[664,698-7001

Hindering activation of the alternate complement pathway, probably by preventing sialic acid binding to factor H

[701,702]

Hindering infection of cattle by Trypanomma brucei

t7031

Decreasing antigenicity

[5,668,704]

Providing epitopes for the binding of influenza C, Corona and encephalomyelitis viruses

[705-7071

Increase of bacterial virulence

[7081

Providing epitopes for recognition by antibodies

[5,7091

Providing epitopes for recognition by lectins

[709-71 I]

Representing differentiation antigens and influencing morphogenesis

[5,668,712-7 161

Representing tumor-associated antigens

[5,668,717-7231

Regulating environmental adaptation

[7241

pyruvate lyase, the next enzyme involved in the catabolism of sialic acids [5,33] (see Fig. 13). This effect may facilitate recycling of sialic acids released from glycoconjugates. 0-Acetyl groups have a further hindering effect on the binding by several proteins, as was observed with the hemagglutinins of influenza A and B viruses, which only bind to non-0-acetylated NeuSAc and Neu5Gc [699], with the malaria parasite Plasmodium falciparum, where the attachment of 0-acetylated mouse erythrocytes was weaker than to non-0-acetylated cells [700], and with sialoadhesin from mouse macrophages [664]. In contrast to this masking of recognition sites, 9-0-acetylated sialic acids are specifically recognized by influenza C virus [705,725], as well as encephalomyelitis and corona viruses [706], leading to binding of these viruses to cells with concomitant infection. In the case of influenza C virus, this may occur in human nasal epithelia, which contain 0-acetylated gangliosides and glycoproteins [J. Tack et al., unpublished]. In human nasal mucin 5-10% of the Neu5Ac residues were found to be 0-acetylated [726].

326

A polyacrylamide glycopolymer containing Neu5,9Ac2 was synthesized (see section 6.2) as inhibitor of influenza C virus hemagglutination, which is useful as a ligand for binding studies and as a potential anti-influenza C virus drug [489]. Influenza C viruses can be used for the detection of glycoproteins and gangliosides with 9-0-acetylated sialic acids on thin-layer chromatograms, and blots of SDS-polyacrylamide gels as well as in histochemical sections (see ref. [727] and unpublished data). This staining procedure, described in section 4, is possible with the aid of the hemagglutinin and the sialate-9-0-acetylesterase activities (section 9.1) of influenza C virus, which is part of the viral spike hemagglutinin-esterase-fusion (HEF) glycoprotein [234,728,729]. In Fig. 16 an example for the localization of Neu5,9Ac2 in the mucin-producing Goblet cells of the mucosa of human colon is shown. With these techniques the frequent occurrence of Neu5,9Ac2-containing glycoconjugates especially in epithelia and in endothelia of blood vessels of rat, cow and man was demonstrated [234,235,727,730,73Oa]. Sialic acids 0-acetylated at C4 cannot be stained by influenza C virus. In addition to these functions, 0-acetylated sialic acids are involved in differentiation and tumorigenesis as evidenced by their differential expression during these processes (for some examples see Table 17). In the adenoma-carcinoma sequence in colorectum, analysis of sialic acid 0-acetylation may assist in diagnosis of tumor disease[723]. 9-0-Acetyl-GD3 was found to be a surface marker for basal cell carcinomas, which may lead to novel, immunological therapeutic interventions [7 181, as is also expected for Neu5,9Ac2-containing glycoconjugates of melanoma cells. The great importance of 0-acetylation during ontogenesis was demonstrated by transfection of sialate-9-0-acetylesterase into mouse embryos, which led to developmental abnormalities in these transgenic mice [714]. There is also evidence that the expression of 0-acetyltransferase activity in human T lymphocytes is regulated by maturation events taking place in the periphery [716]. Although both the sialate 4- and 9(7)-U-acetyltransferases (EC 2.3.1.44 and 45) were discovered 27 years ago in bovine and equine submandibular glands, respectively [73 1,7321, the enzymes have not yet been purified and characterized in detail. Furthermore, nothing is known about the regulation of their expression at the gene level. In contrast to the CMP-Neu5Ac hydroxylase, the 0-acetyltransferases are bound to subcellular membranes, as was studied with bovine and equine submandibular glands [733] and with rat liver [284,734]. Using detergents, the 0-acetyltransferase from bovine submandibular glands was partially solubilized and enriched about 600-fold [7]. This labile enzyme exhibits a pH optimum of 7.2 and transfers acetyl groups from acetyl-CoA only onto a-glycosidically bound sialic acids of glycoconjugates. This corresponds with the finding that 0-acetylation of the side chain at C7 and C9 of Golgibound sialic acids occurs after the translocation of acetyl-CoA into this compartment [8]. The complete pattern of sialic acid side-chain 0-acetylation may require several sialate 0-acetyltransferases each with a specificity for only one position. On the other hand, the primary insertion place for the 0-acetyl function may be the H 0 7 group alone from where the ester group migrates even under physiological conditions to the 9 position, presumably via C8 [7,23], leaving H 0 7 ready for a new transfer. Thus, for complete side-

321

Fig. 16. Staining with influenza C virus of Neu5,9Ac2-containing mucin in goblet cells (arrow) of the mucosa of human colon tissue. Thin-sections were overlayed with virus suspension and virus bound to 0-acetylated sialic acids visualized using FITC-conjugated anti-influenza antibodies 17301. (top) Goblet cells containing native much are stained yellow-green. (bottom) After removal of Neu5,9Ac2 by sialidase, 110 staining occurs and the cells appear dark. This effect can also be obtained by saponification of the ester groups. Magnification about 400-fold.

328

chain 0-acetylation only one enzyme, acetyl-CoA:sialate 7-O-acetyltransferase, may be necessary together with non-enzymatic migration of this substituent. It is also possible that several sialic acid 0-acetyltransferases preferring glycoproteins or gangliosides or at least specific oligosaccharide chains exist. Enzymes were reported to be specific for NeuSAc in N-glycans of rat liver[735] and for gangliosides in human melanoma cells [694,736], respectively. An 0-acetyltransferase incorporating acetyl residues into C7 and C9 of polysialic acid was demonstrated to occur in K1-positive Escherichia coli [ 1591. From the results reported above, an 0-acetyltransferase specific for the 0-glycan of mucin in bovine submandibular glands could also be postulated.

8.4.3. Biosynthesis of 9-0-lactylated sialic acid 5-N-Acetyl-9-0-lactylneuraminicacid (NeuSAc9Lt) has been found in man, cow and trout (see refs. [5,8,13,582,737], see also Table 1). Not much is known with regard to the origin of this compound. However, investigation with a particulate fraction from horse liver suggested that this modification occurs enzymatically, although the exact mechanism and type of lactyl donor is not known [29]. 8.4.4. Biosynthesis of 8-0-methylated sialic acids Sialic acid 0-methylation is a modification that seems to be restricted to the echinodermata, especially in certain starfish [5,8,13]. 0-Methyl groups have only been found in the 8 position of NeuSGc and in much lower quantities on NeuSAc [675]. 9-0-Acetylated derivatives of 8-0-methylated sialic acids were also detected in small amounts (Table 1). The existence of a sialate 8-0-methyltransferase required for the biosynthesis of this sialic acid was first demonstrated to occur in the starfish Asterias rubens [675,738]. The enzyme has now been further characterized after solubilization by detergents from a crude preparation of starfish gonads as a membrane-associated protein that transfers methyl groups from S-adenosyl-L-methionine preferably onto glycosidically linked NeuSGc residues. Neither free NeuSGc nor CMP-Neu5Gc are substrates for this enzyme. Horse erythrocytes or glycopeptides from pig submandibular gland mucin immobilized on Sepharose, containing almost exclusively Neu5Gc residues, are excellent substrates for the sialate 0-methyltransferase [739]. The systematic name S-adenosyl~-methionine:sialate-8-O-methyltransferase (EC 2.1.1.78) has been proposed.

8.4.5. Biosynthesis of 5-N-acetyl-2-deoxy-2,3-didehydro-neuraminic acid This Neu5Ac derivative lacking a glycosidic hydroxyl group at C2 has been found as free sialic acid in small quantities in blood serum, saliva and urine of man [5,8,13] and in tissue extracts of starfish [36] (Table 1). It is unknown, whether this compound, which is a potent competitive inhibitor of sialidases (see below), has a biological function. Its formation under physiological conditions has been described in section 2. 8.4.6. Biosynthesis of 5-N-acetyl-2,7-anhydro-neuraminic acid The occurrence of this neuraminic acid derivative (Table 1) has first been described in cerumen of men [37], although its origin in this material remained unknown. Some hints to the possible enzymatic formation (by microorganisms?) come from the discovery

329

that Neu2,7anSAc is a reaction product of a leech (Macrobdella decura) sialidase ("sialidase I,") with sialoglycoconjugates [61] (see also section 2).

8.4.7. Occurrence of 2-keto-3-deoxynononic acid (Kdn) This substance (for the chemistry see section 3, Tables 1 and 3) was found in the eggs of rainbow trout [740], and the egg jelly coat from the amphibians Pleurodeles waltlii [48], Axolotl mexicanum [105], and Xenopus laeuis [741]. In trout, it is part of gangliosides [ 1221 and glycoproteins, where it can terminate long sialic acid chains or form poly-Kdn chains [267]. To our knowledge, nothing is known about the biosynthesis of this sugar, including the possibility of deamination of neuraminic acid at C5. However, the linkage of Kdn to CMP by the cytidylyltransferase from the testis of rainbow trout has been reported [107,122]. This enzyme can also activate Neu5Ac and Neu5Gc. A Golgi-membrane-bound Kdn-transferase from the ovaries of the rainbow trout Oncorhynchus mykiss has been characterized [545]. This enzyme terminates the growth of a-2,8-polysialyl chains in a series of actions of four glycosyltransferases involved in the glycosylation of ovary glycoproteins. Three of these transferases are sialyltransferases, the first transferring Neu5Gc to C6 of serine- or threonine-bound GalNAc, the second linking another Neu5Gc residue to C8 of this first Neu5Gc and the third enzyme being an a-2,8-polysialyltransferase for chain elongation, before termination of its growth by the 2,s-Kdn-transferase [545]. Remarkably, a developmentally regulated a-2,S-poly Kdn was also discovered in various tissues of mammals [J. Roth, personal communkation].

9. Catabolism of sialic acids A well-balanced system of enzymes involved in the degradation of sialoglycoconjugates is required for the normal functions of cells and tissues. A lack of catabolic enzymes, often due to genetic errors, may lead to severe diseases, as may a surplus of such enzymes, e.g. due to infections by microorganisms. These problems will be addressed in section 10.5. Three enzyme systems are required for the catabolism of sialic acids: 0-acetylesterases, sialidases and lyases. Furthermore, a sialic acid transporter has been described (section l0.-5), which carries liberated sialic acids from lysosomes into the cytosol, where they are either degraded by the lyase or recycled after activation with CTP (section 8.3). Sialic acid permeases provide bacteria with sialic acids for nutritional purposes (section 9.4). 9.1. Sialate-0-acetylesterases

One of the main functions of sialic acid U-acetyl groups is their inhibitory effect on the action of both sialidases and sialic acid-specific lyases [5,33]. While a 4-U-acetyl group completely hinders the action of these enzymes (with the exception of a slow release by viral sialidases [252]), such ester groups at the sialic acid side chain appreciably hinder hydrolysis of the glycosidic bond of these sugars and their further breakdown by lyases. The existence of esterases acting on 0-acetylated sialic acids prior to sialidase is therefore a prerequisite or at least supports the rapid turnover of 0-acetylated sialoglycoconjugates. The possibility of the existence of such enzymes was raised by the observation that

330

horse glycoproteins, which are heavily 4-U-acetylated, seem to have a normal turnover, and sialidase studied in horse liver cannot act on 4-U-acetylated sialic acids [742]. Consequently, the existence of such esterases in horse liver was assumed and two enzymes were found, one hydrolyzing only 9-U-acetyl groups and the other mainly 4-U-acetyl residues from sialic acids (ref. [266], and unpublished work). Furthermore, sialate-9-U-acetylesterases (EC 3.1.1.53) have been isolated and characterized from influenza C virus [705,743], bovine brain [744] and rat liver [745,746]. They also occur in human erythrocytes [747] as well as in enteric bacteria [748]. These enzymes specifically release 9-U-acetyl groups from free and glycosidically bound sialic acids. 9-U-Lactyl groups, U-acetyl groups of Neu5,7Ac2 and methyl-esterified carboxyl groups of sialic acid are not hydrolyzed. Although they also attack U-acetyl groups of non-physiological substances, e.g. naphthyl acetate [266,749], they belong to the few esterases which have been recognized to have a physiological function [266]. Amide groups including the synthetically added 9-N-acetyl moiety of 9NAc-Neu5Ac (Table 13) are resistant to the action of these esterases. However, the latter Neu5Ac derivative binds to influenza C virus hemagglutinin and it does this in a stable manner, as it cannot be removed by the viral esterase [750]. The staining of glycoconjugates containing U-acetylated sialic acids with influenza C virus and the problems involved due to esterase activity were discussed in sections 4 and 8.4.2. The mammalian sialate esterases are considered to be involved in the turnover processes of glycoconjugates in lysosomes, thus facilitating the action of sialidases as was discussed for the rat liver enzyme [735,745]. Tissue esterases may also participate in the recycling of sialoglycoconjugates, as the U-acetyl groups of the ganglioside GD3 in human melanoma cells turn over faster than the underlying sialic acids of this molecule [75 11. It should be noted that the de-N-acetylated sialic acid (Table l), as a component of GD3 ganglioside, can also occur in the same malignant cells [14], as was shown by immunological means. Thus, the unstable neuraminic acid molecule can exist in nature, however, only in glycosidic linkage (section 1). The viral enzyme may be of pathophysiological significance and be involved in the binding to mucins and to endothelial cell surfaces prior to endocytosis of the virus particles [266,728]. With regard to the epidemiology of viral infections, the viral esterase might contribute to the unmasking of new receptor sites to facilitate superinfection by type A and B influenza viruses [752]. Sialate esterase from human enteric bacteria acts on U-acetylated sialic acids of colon mucin and thus facilitates the further degradation of the mucin barrier by bacterial sialidase and other glycosidases [748]. The activity of sialate-U-acetylesterase can easily be followed using free Neu5,9Acz either isolated from bovine submandibular gland much [5,13] or U-acetylated chemically (see section 6) and HPLC determination (see section 5.3.2). Alternatively, 5-N-acetyl9-U-acetyl-2-[4-(dansylamino)phenylthio]-a-neuraminic acid [285] acts as a specific and highly sensitive, fluorescent esterase substrate (see sections 5.3.2 and 6.2). With the corresponding 7,8,9-tri-U-acetylated derivative it was shown that influenza C virus esterase can remove all three U-acetyl groups [469]. Since this esterase acts only on the U-acetyl residue at C9, de-0-acetylation of the whole sialic acid side chain can occur only after migration of the other U-acetyl groups to the primary alcohol group at C9. These experiments provide further evidence for the existence of the cascade of

33 1

alternative enzymatic de-0-acetylation and non-enzymatic 0-acetyl migration proposed earlier (refs. [5,7,33], and sections 2 and 5.3.5).

9.2. Sialidases These enzymes (EC 3.2.1.18) are essential tools in sialic acid metabolism, mainly in catabolic reactions, usually hydrolyzing the 0-glycosidic linkages between the terminal sialic acids and the subterminal monosaccharides of free and glycoconjugatebound oligosaccharides. Sialidase, as well as its substrate, is common in metazoan animals of the deuterostomate lineage from echinoderms to mammals [33,244,246,753]. Diverse viruses and microorganisms, such as fungi, protozoa and bacteria also produce sialidases [33,244,246,753-7561, although they mostly lack sialic acids. The most recent review on the biochemistry and function of the large sialidase family discusses great variety in cellular location, molecular mass and substrate specificity [246]. Therefore, only a few, more general aspects of sialidases will be summarized, such as the type of sialidase reactions, including trans-sialidases, the molecular biology, the pathophysiological significance and inhibitors. 9.2.1. ljpes of sialidases Sialidases are exo- or endo-sialidases, hydrolyzing either terminal sialic acid residues of complex carbohydrates or internal sialic acid glycosidic bonds of oligo- or polysialyl chains. Most microbial and animal sialidases belong to the first group [33,244,246,7537551, in the following simply referred to as “sialidases”, while an endo-sialidase has been found in the coliphage E hydrolyzing the a-2,g-linkages of colominic acid [757]. Based on the nature of the chemical reaction, sialidases can tentatively be grouped into eight types, irrespective of the nature of the complex carbohydrate attacked and the subcellular location of the enzyme. (1) Most microbial and animal sialidases readily hydrolyze terminal a-2,3-, a-2,6and a-2,8-glycosidic linkages, although at different rates. In the majority of cases, the a-2,3-linkage is cleaved most rapidly [244,246,753-7551. Due to steric hindrance the side-positioned (branched) terminal Neu5Ac as in GM 1, however, is resistant to the action of most sialidases with the exception of the Arthrobacter ureafaciens sialidase. The sialidases from C. perfingens and V cholerae cleave the glycosidic bond of their substrates between C2 of Neu5Ac and the glycosidic oxygen, as demonstrated by release experiments carried out in H2[I80] buffer solutions and monitoring by GLC [758]. Sialic acids are released in the a-anomeric form after which they slowly mutarotate to the more stable b-form (see section 5.3.5). Although Neu2en5Ac is a competitive inhibitor for most members of this sialidase group, it has not been detected to be an intermediate of enzyme reaction, as was the case for some viral sialidases (see below). (2) With regard to the primary reaction product, Salmonella typhimurium sialidase behaves quite differently, since it was reported not to retain the configuration of sialic acid during the hydrolysis of Neu5Ac p-nitrophenyl a-glycoside. As primary reaction product the b-isomer was detected by optical rotation measurements [759,760]. (3) Close investigation of the reaction mechanism of influenza virus sialidases [60,320] has shown that Neu2en5Ac appears as a transition intermediate, resulting in the

332

appearance of small amounts of this substance in free form during prolonged incubation of sialic acids with influenza B virus sialidase and sialyllactose, as was identified by GLC/MS analysis [60]. Thus, this sialidase can “synthesize its own inhibitor”. During the reaction, the chair conformation of NeuSAc is distorted to the boat conformation by the concerted action of mainly aspartate, arginine, glutamate and tyrosine residues of the active center. The aspartate serves as proton donor for the release of the glycosidic partner of sialic acid. After the formation of the oxocarbonium intermediate, the addition of a hydroxyl group leads to the formation of a-NeuSAc, or, in a side reaction by deprotonation, to Neu2enSAc. The site of enzyme catalysis was shown by X-ray crystallographic studies of the inhibitor-influenza A NPsialidase complex [76 I]. Here, twelve amino acid residues directly interact with Neu2enSAc. Additionally, a further six conserved amino acids exist lining the active site pocket. On the basis of these studies, a mechanism of influenza virus sialidase action was suggested [762]. (4) As mentioned in sections 2 and 8.4.5, a sialidase from the leech Macrobdella decora yields Neu2,7an5Ac as reaction product [6 I]. The purified enzyme interacts only with the a-2,3-glycosidically bound sialic acid of sialyllactose and various sialoglycoconjugates. This marked linkage specificity is unique among the sialidases. The origin of Neu2,7anSAc in cerumen [37] is unknown, but it is hypothesized that it results from the action of a similar, probably microbial sialidase. The small amounts of this sialic acid found in human urine [763] may have the same origin. (5) The bacteriophage E which infects strains of E. coli displaying the a-2,g-linked polysialic acid colominic acid possesses an eado-sialidase, which was first described in ref. [764]. The reports on such endo-sialidases associated with K1 coli phages have been summarized [765]. These enzymes recognizing a-2,g-linkages of both NeuSAc and NeuSGc are believed to be involved in the attachment of the phages to the bacteria. (6) Shortly after the discovery of Kdn, now also included in the sialic acid family (sections 1-3, Tables 1-3), an enzyme activity was discovered in the loach Misgurnus fossilis hydrolyzing the terminal a-glycosidic linkage of this monosaccharide [39,766]. Although conventional sialidases cannot hydrolyze Kdn bonds, the Misgurnus Kdn-ase can also act on Neu5Ac bound to 4-methylumbelliferone (MU) or GM3 ganglioside. Kdn-ase found in several tissues, particularly in the ovary of rainbow trout, also has sialidase activity [254]. In contrast, a Kdn-ase from Sphingobacterium multivorum only hydrolyzes a-2,3-, a-2,6- and a-2,8-bonds of Kdn glycosides, but does not cleave Neu5Ac or NeuSGc from a variety of sialoglycoconjugates tested [255]. The use of these hydrolases for analytical purposes is mentioned in section 5. I . (7) All these rather different sialidase types specifically recognize a-glycosidic bonds. However, a hydrolase acting on the fi-glycosidic bond of CMP-Neu5Ac is CMP-sialate hydrolase (EC 3.1.4.40) which occurs in plasma membranes [767]. It is assumed to be involved in the regulation of the cellular level of CMP-sialic acids. Evidence for the existence of a Golgi-associated CMP-sialate hydrolase was also obtained in rat and mouse liver [654]. (8) While all these enzymes hydrolytically cleave a-glycosidic bonds of sialic acids, another type of sialidase has been discovered, which can also hydrolyze these linkages but preferably forms sialic acid linkages under physiological conditions in the presence of suitable acceptors. These are the trans-sialidases, occurring in parasites, and a

333

special section (9.2.3) is devoted to these enzymes. Since in principle all enzymatic reactions are reversible, transglycosidation of a Neu5Ac unit is also possible by the use of conventional, bacterial sialidases under appropriate conditions. For examples of corresponding glycoside syntheses see section 6.3. 9.2.2. Primary structures of sialidases Comparison of the enzymatic properties of sialidases from a variety of microbial and animal sources revealed that these enzymes are highly diverse with respect to subcellular location, molecular mass, number of subunits, isoelectric point, temperature optimum, influence of Ca2+on activity, substrate specificity and specific activity [244,246,753,754]. The only common properties are the a-anomeric configuration of the sialic acid glycosidic linkage and the acidic pH optimum (pH 5.0-6.1), though the trypanosomal trunssialidases exhibit maximum activity at neutral pH [755]. As these properties apparently do not reveal much relationship between the different microbial sialidases, the primary structures of the enzymes were further investigated at the DNA level, from which it became clear that, although their enzymatic properties show many differences, the proand eukaryotic sialidases so far investigated belong to one superfamily [660,768]. In the first primary structures of microbial sialidases, obtained by cloning and sequencing of the respective genes from Clostridium perfringens [769], Vibrio cholerae [770], Clostridium sordellii [771] and Salmonellu typhimurium [772], an amino acid sequence motif was detected, which is repeated four-fold in each protein: S-X-D-X-GX-T-W [773]. This motif, named the Asp-box, was found in all 16 sialidases of animals, trypanosomes, and bacteria, which have so far been sequenced (see refs. [660,768] and Table 18). In viral sialidases, however, the motif was rarely detectable (e.g. only the sialidase from N9 influenza A virus strain exhibits the complete motif[786] and has probably undergone mutational alterations). Further identical motifs, or single amino acids at certain positions, became evident by an alignment of the pro- and eukaryotic sialidase sequences known [660]. Gaps had to be introduced as a consequence of the differences in protein size. The central regions of these proteins are especially homologous and exhibit most of the conserved amino acids. A further motif, the F-R-I-P region, which is located N-terminally from the first Asp-box, is highly conserved in clostridial sialidases, but was found to be degenerated to X-R-X-P, when further bacterial and animal sialidase sequences were included in the alignment. Altogether, sixteen amino acids are found to be conserved. The function of the conserved and repeated Asp-box is not yet fully understood. In the N9 influenza A virus sialidase [786], the single corresponding motif is located as part of a b-pleated sheet polypeptide at the connections between the four protein subunits. Immunological studies revealed that the Asp-box I of Trypanosoma cruzi trans-sialidase (see below) is inaccessible to antibodies, which has been taken to indicate that it is not part of the external catalytic domain [788]. Site-directed mutagenesis experiments of the “sma11”(43 kDa) sialidase isoenzyme of C. perfringens resulted in only small alterations of enzyme activity by changing some of the Asp-box amino acids [789], while the exchange of other highly conserved amino acids drastically reduced enzyme activity and increased the K M value, e.g. by replacement of the N-terminally located, conserved arginine-37, probably belonging to the active center, with lysine.

334 Table 18 Cloned sialidases and trans-sialidases with Asp-boxesa Organism

Reference(s)

Rattus rattus (rat)

Cricetulus griseus (hamster) Macrobdella decora (leech) Trypanosoma cruzi Trypanosoma rangelic Actinomyces viscosus Bacteroides fragilis Clostridium perfingens (“small” isoenzyme) Clostridium perfringens (“large” isoenzyme) Clostridium septicum Clostridium sordellii Micromonospora uiridifaciens Salmonella typhimurium Streptococcus pneumoniae Vibrio cholerae Influenza virus (A/whale/Maine/l/84; H I3n9) Bacteriophage E‘ a

The gene structures can be found in the EMBL Data Library. The reaction product is Neu2,7anSAc. Sialidase-like protein. This enzyme has also been cloned [769a]. No Asp-boxes were found so far in the other viral sialidase sequences of the EMBL Data Library. Endo-sialidase.

More insight into the possible role and spatial location within the protein molecule of the Asp-boxes and the other conserved amino acids came from the crystal structure of the sialidase from S. typhimurium [790]. The enzyme has the shape of a six-bladed propeller, each blade consisting of four-stranded antiparallel fi-sheets (Fig. 17). The axis of this propeller passes through the active site. The Asp-boxes are located at topologically equivalent positions on the outside of the blades which is not in favor of a role of these sequence motifs in enzyme catalysis. Cocrystallization of the enzyme with the inhibitor Neu2enSAc enabled the identification of the amino acids involved in sialic acid binding and catalysis. Accordingly, the active site consists of an arginine triad, a hydrophobic pocket and key tyrosine and glutamic acid residues. The interaction of these residues with different parts of the NeuSAc molecule is shown in Fig. 18 for three bacterial sialidases. Identical amino acids are assumed to be involved in catalysis by the low-molecular-mass sialidase from C. perfringens, as was shown by the finding of conserved amino acids at positions identical or similar to the S. typhimurium enzyme, site-directed mutagenesis

Fig. 17. Three-dimensional structure of Salmonella typhimurium LT2 sialidase with bound inhibitor NedenSAc obtained from X-ray crystallography. From ref. [790] by permission of National Academy of Sciences, New York.

and hydropathy studies of the amino acid sequence [Kleineidam et al., unpublished]. The same is valid for the K cholerae sialidase, although the positions of the catalytic amino acids are different, due to the large size of this enzyme [791]. Since the tertiary structure of bacterial sialidases is similar to that found in viral sialidases (e.g. refs. [756,792-7941, for a review see ref. [246]) and residues of the Asp-box have also been found, viral and bacterial sialidases can be considered to belong to one enzyme family. The few mammalian sialidases sequenced so far, i.e. from rat muscle [774] or hamster ovary cells [775], as well as trypanosomal sialidases (see in section 9.2.3), also show many structural features in common with viral and bacterial sialidases and are correspondingly members of this family. A DNA sequence of human origin showing features similar to microbial sialidase genes is also known [795]. However,

336

Asn 318"

128" T ~ 1P 2 4 ~ ~

I e

.OH

342 Tyr 347'P I 740"

1 231" Glu 23OCp 619"

246 Arg 24!jCp 1 635" I I I

I

.*-

CO'o:O' I '

309 erg 312cp 712"

.

'.

*. -.... Fig. 18. Amino acids essential for bacterial sialidase action. The positions of the amino acid residues interacting with different parts of the NeuSAc molecule or forming a hydrophobic pocket (by Leu, Trp and Met; indicated by a dotted line at the left side of the sialic acid molecule) are shown. Vc, Vibrio cholerue sialidase 17911; St, Sulmonellu typhimurium sialidase [790]; and Cp, Clostridium perfringens sialidase [R.G. Kleineidam, personal communication].

the expression of an active enzyme from this nucleotide sequence has not yet been reported. The human sialidase gene is located on chromosome 10 [796]. Interestingly, the leech sialidase resulting in Neu2,7anSAc also belongs to this family, since the consensus repeat S-X-D-X-G-X-T-W was found in tryptic peptides of this enzyme [61]. The following are considerations with regard to the evolutionary origin of sialidases. Those enzymes whose gene structures are known, have been found in mammals or in microorganisms frequently living in a symbiotic or pathogenic association. This fact and the similarity of the structures and properties of mammalian and microbial sialidases together with the observation that the acquisition of sialidase activity may be of benefit for the microorganisms (see below) has led to the assumption that the sialidase gene may have been acquired by microorganisms from the host [797]. This may have occurred during infections by the uptake of DNA from decomposed host cells or during the life of microorganisms in intestine. It is therefore necessary to investigate the sialidase genes from lower animals, particularly of echinodermata, which are considered to be the "inventors" of both the sialidase and sialic acids. The sialidase from the starfish Asterias rubens was isolated to apparent homogeneity [798] and experiments for the elucidation of its gene structure are intended. Evidence for a horizontal sialidase gene transfer between bacteria has been obtained by comparison of the similarities of bacterial sialidases so far sequenced [246,660,768,799]. It was found that some of the sialidases are related in accordance with the phylogenetic distances of their producers, e.g. Micromonospora viridifaciens and Actinomyces viscosus

337

sialidase, or the “large” (72kDa) isoenzyme of C. perfringens and the C. septicum sialidase. On the other hand, the three “small” sialidases ( 4 2 4 4 m a ) produced by S. typhimurium, C. perfringens and C. sordellii, exhibit a higher similarity than is expected from the phylogenetic relationship of the bacterial species. These Gram-positive (Clostridia) or Gram-negative (Salmonella) bacteria are quite distinct from an evolutionary point of view, but are found at the same ecological location, e.g. in wounds or intestine of vertebrates. Here, an exchange of genes even between unrelated microorganisms might be possible via phages, transposons or plasmids by mechanisms of transfection or conjugation. The participation of phages is indicated by typical sequence motifs up- and downstream from sialidase genes of S. typhimurium and M. uiridfaciens [772,783], and from the observation that the “small” sialidase gene is located near a phage attachment site on the chromosome of C. perfringens [800]. Further evidence for such a horizontal gene transfer comes from the investigation of the percentages of the bases G + C, which in the sialidase genes are frequently atypical for the chromosomal DNA of the respective bacterial species. From these data the hypothesis of a gene transfer, from mammals to microorganisms, is supported, since on basis of the mol% G + C , the origin of bacterial sialidase is expected in an organism exhibiting genes which contain about 45 mol% G + C, a value not uncommon in higher animals [660]. 9.2.3. Trans-sialidases As all chemical reactions, the sialidase reaction is theoretically reversible. This has been exploited, by choosing suitable conditions and bacterial or viral sialidases, for the a-2,3- or a-2,6-sialylation of oligosaccharides as final step in the course of chemical synthesis (see section 6.3). During the last years, several sialidases have been discovered in some protozoa which behave like normal sialidases if only water is present, but preferably transfer sialyl residues from one glycan chain to the terminal galactose of another non-sialylated oligosaccharide or glycoconjugate forming a-2,3-linkages. The first hint for the existence of an unusual sialic acid transfer reaction was given in 1983, when sialic acids (Neu5Ac and NeuSGc) were detected in Trypanosoma cruzi [662,801], which were found not to be synthesized by the parasites themselves. Since the molar ratio of these trypanosomal sialic acids corresponded to that of the incubation medium, their acquisition from this source, or in the case of an infection, from the host glycoconjugates, was assumed. Furthermore, an involvement of the sialidase activity found in these trypanosomes [802] and later localized on their cell surface, was suspected. These observations finally led to the discovery of enzymes called “trans-sialidases”, first in the American species 1: cruzi [803-8051 and later in the African species T. brucei [755] and I: congolense [806]. A closer investigation revealed the presence of this enzyme in the whole 1: brucei group, i.e. several strains of 7: brucei brucei, T. brucei rhodesiense and I: brucei gambiense [806]. It also occurs in Endotrypanum promastigotes, an intraerythrocytic flagellate distantly related to trypanosomes [807]. In contrast, many other members of the kinetoplastida lack both sialidase and trans-sialidase activities, or express only sialidase, such as 1: uiuax [808] and T. rangeli [809]. This rapidly expanding field has been reviewed in refs. [246,810]. It is possible to discriminate between morphologically indistinguishable trypanosomatids by measurement of trans-sialidase and sialidase activities [S 111. Furthermore, the occurrence

338

of sialic acids on major cell-surface epitopes correlates with the expression of transsialidase [812]. With regard to substrate specificity, trans-sialidases exhibit a preference for sialyl a-2,3linkages in common with most of the usual sialidases. While for the 7: cruzi enzyme only a-2,3- but not a-2,6-linked sialic acids are donors in the transfer reaction [550], the 7: brucei trans-sialidase can also transfer a-2,6-linked sialic acids from sialyllactose to other glycans, however at an eight times lower rate than from sialyl(a2-3)lactose [755]. While both Neu5Ac and NeuSGc are transferred equally well, Neu5,9Ac2 from the corresponding sialyllactose derivative serves as sialyl donor only at a much reduced rate[696]. This may be the reason why N’dama cattle exhibiting a high degree of 0-acetylation of erythrocyte sialic acids are more trypanotolerant than Zebu cattle, whose erythrocyte sialic acids are much less 0-acetylated [703]. Only P-galactose residues of a variety of oligosaccharides and glycoconjugates are acceptors of these transferases. Sialyl(a2-3)lactose is the best donor for the trans-sialidase from 7: brucei and lactose the best acceptor when compared with serum glycoproteins, mucins and gangliosides [755]. MUNeu5Ac also serves as a sialyl donor. A novel substrate for the assay of trans-sialidase activities is MUB-D-galactopyranoside as sialic acid acceptor which was tested with the T brucei enzyme [806]. The main advantage of this accurate and sensitive assay is the avoidance of radioactive lactose as substrate. Due to its wide substrate specificity but very limited linkage specificity, the transsialidase from 7: cruzi has been used for the sialylation of oligosaccharides, yielding NeuSAc(a2-3)Gal sequences, in glycotechnological experiments [554], discussed in detail in section 6.3. Remarkably, the trans-sialidases from T. cruzi and T brucei are not inhibited by Neu2en5Ac [550,809,813]. As this is in contrast to most other sialidases described in the groups 1-3 of the previous section, Neu2en5Ac is probably not an intermediate of the trans-sialidase reaction. Furthermore, N-(4-nitrophenyl)oxamic acid and related N-acylanilines do not inhibit 7: cruzi or 7: brucei trans-sialidases, which is also in contrast to the influence of this substance on the activity of sialidases from Y cholerae and other bacteria [813]. The term “trans-sialidase” used for the transglycosidase described in this section is not an adequate denomination of this enzyme, which also can react as a conventional sialidase, thus representing a hybrid glycosidase/glycosyltransferase. The reactions are shown in Fig. 12 of section 6.3. The product of the transsialidase reaction corresponds to that formed by sialyltransferases, however, instead of CMP-NeuSAc, a-sialyl residues bound to P-galactose of glycan chains or to aglycones such as 4-methylumbelliferone (MU) are the sialyl donors. Therefore, the systematic name for the glycosyltransferase function of this enzyme, acylneuraminosyl(or sialyl-):~-~-galactoside-a-2,3-sialyltransferase (EC 2.4.99.7?), has been proposed. This name takes into account that both NeuSAc and Neu5Gc can be transferred. Presently, several laboratories are intensively studying the gene structures of trypanosomal trans-sialidases, in order to understand the differences between this unique enzyme and conventional sialidases. In accordance with the latter sialidases, the transsialidase genes of 7: cruzi [768,776,814-8181 also contain several Asp-boxes, first found in bacterial sialidases (see section 9.2.2 and Table 18). In addition, the T cruzi trans-

339

sialidase contains conserved amino acids, found to be involved in the catalysis of S. typhimurium sialidase [790] at similar positions [768,819]. It was shown at least for one

of these amino acids, namely tyrosine-342, that it is involved in enzymatic catalysis [819]. Enzyme activity was lost after exchange of this amino acid by a histidine residue. Remarkably, the latter, inactive mutant is a member of a group of proteins with similar structures, expressed by several homologous genes, of which only some exhibit transsialidase activity [815,8 161. The others are enzymatically inactive and their function is as yet unknown. However, these studies show that all animal, microbial and probably viral exo-sialidases and trans-sialidases studied belong to one gene family [660,768]. As first observed by Pereira et al. [776], the T cruzi trans-sialidase consists of an N-terminal half, containing the enzymatic function and a C-terminal portion, composed mainly of a tandem series of twelve amino acid repeats. Four Asp-boxes exist within the N-terminus of the enzyme. However, only three of these conserved motifs occur in positions of the amino acid sequence similar to those found in bacterial sialidases, while the distance between the third and the fourth box is much larger in the parasite enzyme than in the bacterial proteins. It may be assumed that this extended sequence together with some other structural features is a domain functioning in the trans-sialidase mechanism [768,8 151. The carboxy-terminal repetitive motif is most probably not involved in catalytic activity [815,8201 and a recombinant T cruzi trans-sialidase lacking the amino acid repeats was shown to retain the enzymatic activity [821]. Remarkably, the American parasite 7: rangeli secretes a conventional sialidase [809] lacking transsialidase activity [822], the primary structure of which is related (70% identity of the amino acid sequences) to that of the trans-sialidase of T cruzi [777]. Sequences encoding the tandemly repeated motif characteristic of the members of the trans-sialidase family were not detected in the T rangeli genomic DNA. 7: cruzi trans-sialidase first named SAPA (“Shed-Acute-Phase-Antigen”) and found in human serum in Chagas disease, is also unique from an immunological aspect, since it contains two immunologically distinct domains and seems to be a naturally chimeric protein having functionally independent enzymatic and antigenic domains [823]. (It is thus reminiscent of c! cholerae sialidase which consists of a catalytic domain and a non-catalytic sequence similar to those of legume lectins [768,791].) Antibodies against both regions have been raised, and the one recognizing the catalytic region inhibits enzyme activity [788,823,824]. The SAPA epitopes were shown to be located on the tandemly repeated 12-amino-acid units and to induce an early and strong antibody response in acute and congenital human infections, as well as in mouse infections. In patients, antibodies against the immunodominant, non-enzymatic domain are formed much earlier (8 days) after infection than against the enzymatic domain (about 30 days) [823,824]. The availability of antibodies may be suitable for diagnosis and therapy of Chagas disease which affects several million people in Central and South America. The structural and functional properties of TYypanosoma trans-sialidases have been reviewed [825,826]. 9.2.4. Pathophysiological signijicance of sialidases and trans-sialidases 9.2.4.1. Eukaryotic (trans-)sialidases. It is well known that sialidases are involved in the breakdown of sialoglycoconjugates, which may be of physiological or pathological

340

significance. These enzymes are physiologically important in the lysosomes, their main localization in animal cells, where they release sialic acids from glycoconjugates, thus initiating the degradation of these molecules by other glycosidases or proteases [5,8,33, 2441. The significance of mammalian sialidases is indicated by diseases, called sialidosis [827]. In these genetic disorders, sialidase activity is reduced or lacking, resulting in reduced turnover of sialylated glycoconjugates, excretion of sialo-oligosaccharides in the urine as well as mental and physical impairments (see below). Since sialidases not only occur in lysosomes, but also in other compartments such as the cytosol or plasma membranes [33,244,246,660], other cellular functions of these enzymes are assumed. This is easily understandable, since sialic acids are not only structural but also functional constituents of many biologically active molecules of the cell, for example receptors, ion channels and growth-regulating molecules (refs. [246,660] and section 10). Sialidases may regulate the functional activity of these and other molecules. Plasmamembrane-bound sialidase for instance is involved in the degradation of cell membrane gangliosides, which may not only represent a catabolic function, but also a regulative effect on the biological functioning of these molecules [828]. A striking example of such a role is the myelin-associated sialidase of rat oligodendroglial cells which was found to adhere to immobilized GMl and to be inhibited by this ganglioside [829]. This interaction can also be inhibited by the sialidase inhibitor Neu2enSAc. It is concluded from these experiments that the sialidase-GM1 interaction plays a role in the formation and stabilization of the multilamellar structure of the myelin sheath. Another system of eukaryotic sialidases, whose significance we are just beginning to understand, is the trypanosomal trans-sialidase described above. It is evident that the main function of this enzyme is the acquisition of sialic acids by the parasites at expense of the host which is presumably of benefit for the trypanosomes. In the case of 7: brucei the accepting glycoprotein is the GPI-anchored “procyclic acidic repetitive protein” (PARP/procyclin) [755] and in the case of 7: cruzi the Ssp-3 epitope of trypomastigotes [823]. In epimastigote forms glycoproteins with molecular masses of 3 8 4 3 kDa were described to be the main sialic acid acceptor, and the structures of the oligosaccharides of these glycoproteins, linked 0-glycosidically via GlcNAc to serine or threonine, were elucidated [830]. Some understanding of this chemical modification of cell surface molecules may come from consideration of the period of sialylation during the life cycles of the African and American trypanosomes. In 7: brucei, sialylation only occurs in the insect stage, where sialic acids are probably derived from mammalian blood sucked in by the tsetse fly [806]. In contrast, in T cruzi sialic acid is acquired by the parasite not in the insect (triatomine bugs) stage, but only during its life in the mammalian host. Since sialylation may be of advantage to the trypanosomes, it is hypothesized that the dense, charged glycocalix formed by sialylation of 7: brucei may protect 7: brucei from the attack of proteases, glycosidases, antibodies or complement from the fly’s digestive tract or from the ingested blood meal [806]. It is also conceivable that the parasites require sialic acids for maturation and adhesion to the fly’s salivary glands before transfer to the mammalian hosts. Furthermore, after infection, sialylation may be of benefit for the survival of the parasites in the primary infection site (“chancre”) of the host’s skin, before the non-sialylated variable surface antigenic glycoproteins (VSG) are expressed. These

341

antigens then represent another potential site for the immunological protection of the parasites when sialic acids are no longer available. Since 7: cruzi does not express such variable antigens on the cell surface, it is assumed that it starts to express trans-sialidase in the mammalian host to capture sialic acids for defense against the host’s immune and complement systems and against phagocytosis [8 18,83 1,8321. The role of sialic acid in the resistance of 7: cruzi trypomastigotes to complement has clearly been shown [833]. Other experiments demonstrated that the acquisition of sialic acids increases the infectivity of 7: cruzi and is involved in invasion of host cells by these parasites [823,83 1,8341, and accordingly the action of sialidase reduces their virulence. In mice the virulence-promoting activity of trans-sialidase was achieved with tiny doses of 1-2 pg enzymelkg animal [835]. (An increase of pathogenicity, i.e. evasion of host defence systems, is also achieved by sialylation of some pathogenic bacteria, such as group B Streptococcus [836] and Neisseria gonorrhea [837] (see also Table 3 and sections 3 and 10.2). Correspondingly, monoclonal antibodies against the Ssp-3 epitope, the target of trans-sialidase activity and required for invasion after sialylation inhibit infection of the host cell [823,838]. Such antibodies or other transsialidase inhibitors (see below), as well as prevention of the expression of this enzyme on the gene level, may become potent therapeutic tools in Chagas and Nagana (sleeping) diseases. The role of conventional sialidases secreted by non-sialylated trypanosomes lacking an insect vector (see above), such as 7: rungeli [809,822] or 7: v i v a 118061, is unknown. It is likely that they hydrolyze sialic acids from host glycoconjugates for nutritional purposes as do bacteria (see below). It is, however, also possible that these trypanosomes demask recognition sites on host cells, e.g. subterminal galactose residues, for their attachment mediated by lectins or other receptors on the parasite’s surface. This hypothesis clearly warrants further investigation. Infection of host cells by T. cruzi parasites is a complicated mechanism which is not yet fully understood. However, sialylation of both cell types seems to be necessary for this process. For the attachment and penetration of 7: cruzi into host cells the latter must be sialylated, as was shown with CHO cell mutants expressing no or varying amounts of sialic acids. O-Linked sialylated glycans of host cells were shown to be particularly important for infection [838]. Various trypanosomal surface glycoproteins are considered to be involved in adhesion, i.e. the developmentally regulated MTS-gp82 [839], a small mucin-like sialoglycoprotein [840], or enzymatically inactive members of the (trans-)sialidase family, which may be capable of binding sialic acid without cleaving it, and in this way may function as mammalian-stage-specific surface receptors [84 I]. On the parasite surface, the relative balance between exposed sialic acid and galactoselN-acetylgalactosamine residues may determine the parasite’s capacity to invade host cells including macrophages [842]. A stage-specific 82 kDa adhesion molecule of 7: cruzi metacyclic trypomastigotes may play a role in host cell invasion [843]. After binding, 7: cruzi enters host cells via an acidic vacuole, which fuses with lysosomes [8 18,844-8461. The parasitophorous vacuole formed in this way is later disrupted in order to release the trypanosomes into the cytosol. Truns-sialidase activity seems to be required for the latter step, as it desialylates the highly sialylated lysosomal

342

membrane, which then becomes more susceptible to a parasite hemolysin [844,846]. The crucial role of sialic acid in host-cell-parasite interaction and the reversible sialylation of cell surface molecules mediated by sialidase and trans-sialidase activities, which in some respects resembles the infection of mammalian cells by influenza viruses, has also been pointed out in other studies [826,847]. In a species of Acanthamoeba, another protozoon, the sialidase of the trophozoites and cysts also seems to play a pathophysiological role, as it was shown to be involved in colonization and damage of the sialic-acid-rich corneal epithelium [848]. Spreading of the microorganisms may have been facilitated by a decrease in the viscosity of the protective mucus layer occurring after desialylation. 9.2.4.2. Bacterial sialidases. In bacteria, sialidases also seem to be of great significance in pathogenesis, although only few well-defined effects are known, in contrast to the wide occurrence of this enzyme. Strikingly, bacterial species pathogenic for higher animals and man often produce sialidases so that it is suspected that they enhance virulence. It was also discussed above that bacteria may have acquired the sialidase gene from the mammalian host, which may have been of evolutionary advantage. Certainly, not only one function can be attributed to bacterial sialidases, since their cellular localization varies much with the species. They may be intracellular, periplasmic, membrane-bound or excreted, and they vary in their properties such as complexity, molecular mass and substrate specificity (refs. [33,244,246,660,768], and section 9.2). However, they belong to one gene family. One of the main functions of bacterial sialidases is nutritive, providing an energy source for the microorganisms [245,246]. This has long been assumed and a model first proposed for the C. perfringens enzyme shows that sialic acids released from host tissues by secreted sialidase are taken up by the bacteria with the aid of a permease, intracellularly cleaved by acylneuraminate-pyruvate lyase to pyruvate and N-acylmannosamines (section 9.3) and then further metabolized [5,33,245]. After the loss of sialic acids, host cell membrane glycoconjugates are more vulnerable to the action of other glycosidases and proteases secreted by bacteria, since the presence of terminal sialic acids on glycoconjugates is known to hinder the action of these enzymes [5]. Thus, sialidase can be considered as pacemaker in the breakdown of host cells by a lytic cocktail secreted by many infectious bacteria. Suitable models are required to verify this assumption. For example, the advantage of the availability of free sialic acid for bacterial survival has been demonstrated in two model systems [849]. The growth of two isogenic strains of Bacteroides fragilis in CHO cell culture monolayers in oitro or in rat granuloma pouches in oioo was compared. While one strain expressed sialidase, the activity of this enzyme was deleted in the other group. At the beginning of the culture, both strains grew equally well, however, after 48 to 72 h, the sialidase-producing strain was enriched. The latter bacteria grew more rapidly than those lacking sialidase, because after the consumption of glucose some 16 to 24h after infection, other carbon sources, including liberated sialic acid, were required for further bacterial growth. This experiment supports the assumption that sialidase may facilitate spreading and thus virulence of bacteria in infected tissues. Similar conclusions were drawn from studies on arteritis caused by Evysipelothrix

343

rhusiopathiae [850], in which a close correlation between bacterial invasion, desialylation and cell infiltration in the common iliac artery was observed. Several clostridial species secrete large amounts of this enzyme into their environment, which, together with other lytic enzymes, is believed to facilitate their rapid and lifethreatening spreading and desialylation of vital structures in the whole organism [85 1, 8521. The sialidase production of C. perfringens is so great that it can be detected not only in wound fluids, but also in blood serum. Antibodies were raised against sialidases from C. perfringens, C. septicum and C. sordellii, the main causative agents for clostridial myonecrosis or gas oedema, which allow an early, sensitive and specific diagnosis of this infection in wound fluids or blood serum. This enables a prompt, specific and therefore life-saving treatment of this severe disease. The action of sialidase also contributes to the liquification of mucin barriers lining endothelia, as mentioned above for Acanthamoeba infection of the eye. This property is most relevant in tracheobronchial, intestinal and urogenital tracts [245,853]. Often other hydrolases, “mucinases”, support the viscosity-decreasing effect initiated by mere desialylation of mucins, as was observed with intestinal [854,855], vaginal [856] and other[857] bacteria, as well as with the flagellates Trichomonas vaginalis and 7:foetus [858]. Microbial sialidase is considered a risk factor for intra-uterine infection and preterm birth [858]. 0-Acetylation of mucin-bound and membrane-bound sialic acids makes them more resistant towards sialidase action [8,33,245,853,854]. This may be one of the reasons why intestinal mucins, especially of the colon, are often 0-acetylated (ref. [245], and section 8.4.2). The high level of 0-acetylation of sialic acids observed in the endothelia of blood vessels, e.g. in liver, detected by histochemical methods using influenza C virus hemagglutinin, is assumed to have a similar function [234,235,730]. A pathogenic role facilitating colonization and invasion of the host during the development of pneumonia is also attributed to the sialidase from Streptococcus pneumoniae [859]. This enzyme has also been sequenced and found to contain four copies of the sequence S-X-D-X-G-X-T-W typical for sialidases [784]. The antigenic potential of this enzyme as well as other pneumococcal proteins makes them potentially suitable candidates for the vaccination against lung inflammation (refs. [784,860], and [K. Fischer, personal communication]). During the inflammation of bovine lung, sialidase from Pasteurella haemolytica may play a supportive role [861]. In all these cases, the production of sialidase may be physiologically important for the microorganisms causing an enhancement of their virulence. A further, well-defined example for such a role is the Vibrio cholerae sialidase, which is an accessory virulence factor, causing a partial desialylation of higher sialylated gangliosides in intestinal endothelia, resulting in GM 1 [770]. Cholera toxin subsequently binds to this ganglioside and is taken up by intestinal cells, which are stimulated to secrete cell fluid required for growth of the vibrions in the intestinal lumen. The efficiency of the sialidase in this process is increased by the capability of the enzyme to attach to cell surfaces via lectinlike domains [768,791]. Since sialic acids are involved in so many important physiological processes in animals and man (section lo), the flooding of a tissue or the whole organism with bacterial sialidases is potentially disastrous, leading to acute and chronic diseases.

344

As a consequence of sialidase production by microorganisms and viruses, chronic inflammatory or immunological diseases such as glomerulonephritis or asthma may result [245,795]. Sialidase-producing bacteria in disease and phenomena resulting from the action of sialidases in bacterial infections have been summarized [245]. 9.2.4.3. Viral sialidases. Sialidase is a surface-bound enzyme of ortho- and paramyxoviruses as reviewed[246,862]. The enzymes of influenza A and B viruses have been intensively studied with regard to their structure, substrate specificity, immunological properties and function during viral infection. These sialidases play a crucial role during penetration of the virus into cells and probably also in their release from infected cells, especially of respiratory endothelia, after the attachment to these cells via the sialicacid-specific receptor hemagglutinin [246,863,864]. The first step of this process is the binding of viruses to cell surfaces, which is followed by loosening of these linkages as a prerequisite for endocytosis of virus particles. Thus, the host sialic acid is exploited by the infecting virus. Sialidase may also be required by the virus to reach the endothelial surface and not to be bound to the highly sialylated, sticky, protective mucus layer, which would lead to elimination of the virus from the respiratory tract. 9.2.5. Sialidase and trans-siulidase inhibitors The involvement of sialidases in the severe pathogenesis of many viruses and other microorganisms and the aggravation of these infectious diseases including late and often irreversible, e.g. immunological effects prompted pharmaceutical strategies for the inhibition of these enzymes. Some sialidase inhibitors will therefore be described in the following. For the terminology and abbreviations of natural and synthetic sialic acids see Tables 1 and 13 and section 6.1. A survey of sialidase inhibitors was given earlier [5,33,865] and some recent inhibitors of higher efficiency ( K , < 10"' M) are listed in Table 19. Neu2enSAc is a competitive inhibitor which is still widely used, since it inhibits most viral, bacterial and mammalian sialidases with a Ki of around 5pM. This substance is therefore very useful to prove the specificity of sialidase action when studying the role of sialic acids in biological experiments. The replacement of the N-acetyl moiety in Neu2en5Ac by an N-trifluoroacetyl residue leads to an about 2-fold better inhibitor of K cholerue sialidase M with a K ,of 1.3-1.8 x 10-6M [389,866]. In the case of 9d-NeuSAc a Ki of 1 x was measured [869]. Using K cholerue sialidase, it was shown, with various Neu2enSAc derivatives structurally varied at C4, that the highest inhibitory effect was obtained with compounds having an axial hydroxyl residue, as in the parent substance Neu2enSAc [430], whereas the 4epi- and 4d-derivatives were less inhibitory. These studies also revealed that 4amino-Neu2enSAc inhibits K cholerue sialidase only slightly, which is in contrast to influenza virus sialidases. The transition-state analogue 4d-Neu2enSAc inhibits influenza virus sialidase with a K ,of 8 x M) [441]. Most interest is currently being focussed on the inhibition of viral sialidases, since their role in the infection mechanism has long been known. The fact that Neu2enSAc is a transition-state analogue for the sialosyl residue during the viral enzyme reaction described in section 9.2.1 [60,320], served as a point of reference in the design of new inhibitors. The observation that the 4-0-acetyl group in the ring of Neu4,5Ac2

Table 19 Sialidase inhibitorsa Name

Inhibition (M)

Sialidases tested

Reference( s)

Neu2enSAc

Ki = 1 x 10-6-2 x 1 0-5

viral, bacterial, mammalian

[428,865]

5-Trifluoroacetyl-Neu2en

Ki = 8 x 10-7-2x Ki = 4 x 10-R-7 x 10-4 K, = 1 x 10-9-1 x lo-* K, = 2 x 1 0-6-8 x

viral, bacterial

[389,866]

viral, bacterial, mammalian

[428,867,868] [428,867,868]

4amino-Neu2enSAc 4guanidino-Neu2enSAc 3Feq-Neu5Ac 2d-6amino-2Ha,-Neu5Ac

2d-4epi-6amino-2Ha,-Neu5Ac Phosphonate analogue of Neu5Ac 7d-Neu2enSAc 8d-Neu2enSAc 9d-Neu2enSAc

Ki = 5 . 4 ~ Ki = 2 . 9 ~ Ki =5.5 x K~= 9.ox 10-5 K,=5.0~10-~ Ki = 1 .I x 10-5 Ki = 3.8 x

viral, bacterial, mammalian viral, bacterial, mammalian bacterial bacterial

14491

bacterial bacterial

18691

Neu5Aca(2-S-6)Gal(fi14)Glc(fil-I )ceramide

Ki = 2.8 x 10-6-4.0x 10-4 ~,=1.5x10-5

Siastatin A and B N - ( 1,2Dihydroxypropyl)-siastatin B

Ki = 1.7x 10-5%3 x 1O-’ IC50=1 . 0 ~ 1 0 - ~ - 2 . 7 x I O - ~

bacterial bacterial

N - ( 1,2Dihydroxypropyl)-4-deoxy-siastatinB 3epi-Siastatin B 5,7,8,4’-Tetrahydroxyflavone(Isoscutellarein) 5,7,4’-Trihydroxy-8-methoxy-flavone

I 0-(-5 10-5 IC5,= 1 ~ 1 0 - ~ - 7 . 4 x l O - ~ ~~=4.1x10-5 IC,, = 5 s X 10-5

bacterial viral viral

a

IC,,

=5

[438,452] [438,452]

bacterial bacterial

bacterial viral, bacterial viral

8epi-Neu2enSAc Neu5Aca(2-S-6)GIc(fil-1 )ceramide

14421

viral

18691 18691 18701 18711 18711 [865,872]

18731 18731 W41

18751 18751

For the abbreviation of natural and synthetic sialic acids see Tables 1 and 13, respectively

w L P n

346

completely hinders the action of mammalian and bacterial sialidases [5,33] but allows slow hydrolysis of the glycosidic bond by viral enzymes [252,876], as well as the finding by X-ray crystallography studies that in the active center of viral sialidases only acidic amino acids (i.e. aspartic acid) interact with H 0 4 of sialic acid, led to the synthesis of much more potent inhibitors of influenza viruses. These are 4amino-Neu2enSAc and 4guanidino-Neu2enSAc [428,429,867,868,877], which exhibit K,values in the nanomolar range, corresponding to 2 4 orders of magnitude lower than Neu2enSAc. The guanidino derivative inhibits more effectively than the amino compound. 4-Guanidino-Neu2enSAc was shown to inhibit influenza virus A sialidase in a slow-binding fashion, and a Ki value of 3 x lo-'' M was observed for the tightly bound form under steady-state conditions [878]. This strong binding to viral sialidases and not to those from bacteria or mammals, is due to the existence of a pocket in the binding site of viral sialidases near to the H 0 4 group of sialic acid [867,879,880]. This pocket is considered to be the reason why Neu4,5Ac2 [252] and NeuSAc4Me [407] can be released by viral sialidase, but not by the bacterial enzyme. Since the other sialidases do not seem to have this pocket (for bacterial sialidases see refs. [790,881]) and are therefore inhibited to a much lesser extent, the new Neu2enSAc derivatives appear suitable for the prophylaxis and/or therapy of influenza A and B virus infections [867]. It has been shown for 4guanidino-Neu2enSAc that it inhibits the growth of a wide range of influenza A and B viruses in in vitro systems, including human respiratory epithelium, at lower concentrations than classical inhibitors of virus replication such as amantadine or ribavirin [868,882-8841. The same effect was observed in mice and ferrets when the new compounds were administered as a nasal spray [867,880,882], thus raising hopes that these non-toxic substances can be used therapeutically during epidemics of influenza in man. These viral sialidase inhibitors are an excellent example of how drugs may be designed on the basis of a crystal structure of a target protein and on other biochemical parameters [867]. Potent inhibitors of bacterial sialidases or trypanosomal trans-sialidases are also desirable as pharmaceuticals to assist the therapy of inflammatory diseases caused by these microorganisms. Some natural substances exist which inhibit viral sialidases in the 10 pM range, such as siastatin A and B and panosialin [865] as well as flavonoids [875] (Table 19). Two epimers of siastatin B, 3-epi-siastatin B and 3,4-diepi-siastatin B, have been synthesized which are good inhibitors of sialidases from various influenza virus strains, as well as from Sendai and Newcastle disease viruses [874]. They effectively inhibit the growth of these viruses in cell cultures. Sialidases from Streptococcus sp. and C. perfringens are inhibited to an extent comparable to Neu2en5Ac by the N - ( 1,2-dihydroxypropyl) derivative of siastatin B and its 4-deoxy analogue [873]. Isoscutellarein (5,7,8,4'-tetrahydroxyilavone) isolated from the leafs of Scutellaria baicalensis is a good inhibitor of influenza virus sialidases (K,= 4 x M). It is not toxic to mice and inhibits replication of influenza viruses in Madin-Darby bovine kidney cells. It was shown to be a better inhibitor than amantadine. The 8-methoxy derivative of isoscutellarein is also a potent inhibitor [ 8 8 5 ] . Thioglycoside-analogues of gangliosides [87 13 (for structures see Table 19) are good inhibitors of viral sialidases, but hinder the bacterial enzymes to a considerably lower degree (Ki= 0.1 mM). Various sialyl(a2-6)galactosides in which one of the hydrogen atoms at C6 of Gal has been replaced by a methyl group, yielding R and S isomers, and in some of which thioglycosidic bonds occur, act on both viral and bacterial

341

sialidases [886]. Although their inhibition constants are only in the millimolar range, these compounds are expected to have important implications in the rational design of sialidase inhibitors. The finding that these glycosides are the first common inhibitors of both hemagglutinin and sialidase of influenza viruses may prove to be of great advantage. The o-(difluoromethy1)phenyLglycoside of a-Neu5Ac irreversibly inhibits C. perfingens sialidase [887]. This mechanism-based inhibitor acts by a fluoromethylene-quinone formed spontaneously from the liberated difluoromethylphenol. Analogues of free NeuSAc, such as H 0 4 epimers of 6-amino-6-deoxy-sialic acid are competitive inhibitors of medium potency for both viral and bacterial sialidases [438]. 3F-Neu5Ac strongly and competitively inhibits bacterial, viral and mouse spleen sialidases [442]. Other strategies for the production of sialidase inhibitors are the Neu5Ac isomer with a 6-acetylamino group occurring in the furanose ring form, which could be interpreted as a 6-acetylamino derivative of Kdn or 4epi-Kdn [888]. Promising inhibitors also represent 4azido-NeuSAc-, 4amino-Neu5Ac- and 4NAc-NeuSAc-containing sialosides and thiosialosides [889]. Another principle for tailor-made sialidase inhibitors is the design of low-molecularmass compounds that mimic the binding function of a macromolecular antibody. Such studies are based on the crystal structure of an influenza virus N9-sialidase (antigen)-NC4 1 (antibody) complex, showing the direct contact of four amino acid residues on the antibody binding surface with the active-site loop 368-370 of the antigen [890]. Correspondingly, a constrained cyclic peptide composed of 5 amino acids was synthesized, which mimics the receptor-bound conformation of these amino acids and inhibits the sialidase with a Ki of 0.1 mM. Although this is about 3 orders of magnitude less than the parent protein, because the antibody forms contacts with about 17 amino acid residues of the antigen sialidase, it nevertheless opens a possible new approach for the design of more potent inhibitors of this important viral enzyme. The rapid development of research on trypanosomal trans-sialidases, their involvement in the pathogenic mechanism and the morbidity of men and agricultural animals caused by these parasites, has evoked considerable interest in inhibitors of these enzymes. One potential hope is the availability of antibodies, mentioned in sections 9.2.3 and 9.2.4.1. There it was also outlined that classical sialidase inhibitors like Neu2en5Ac and N-(4-nitrophenyl)oxamic acid are not effective on the trans-sialidases studied. A nonimmunological, non-competitive inhibitor of T cruzi sialidase activity, named cruzin, was first found in blood plasma from patients with Chagas’ disease and later identified as high density lipoprotein (HDL) [834,891]. This lipoprotein does not inhibit sialidases from 7: rungeli, various bacteria and influenza viruses. Remarkably, in high concentrations it enhances infection of culture cells by 1: cruzi trypomastigotes, while K cholerue sialidase reduces it, which points to the significance of sialylated structures on the parasites during infection. All these effects could be explained by the fact that HDL is not an inhibitor of 7: cruzi sialidase, but as a glycoprotein provided sialic acids for the trans-sialidase action of this enzyme, thus mimicking an inhibition of enzymatic sialic acid release in the test system used [8 151.

9.3. Sialate-pyruuate base As shown in Fig. 13, sialic acids liberated by sialidase are degraded by the aldolase systematically named sialate (acy1neuraminate)-pyruvate lyase (EC 4.1.3.3), resulting in

348 H07 and H08 of moderate importance for binding and activity

1

COO-

HO9 not important for binding and activRy

COO- essential for binding and activity

H 02 essential for binding and activity hydrophobic interactions preferred; larger substituents accepted for binding and activity

OH 0

3

equatorial H03 of minor influence on binding; activity is abolished

hydrophilic interactions important at C4; axial or equatorial methyl groups diminish binding and prevent activity; equatorial OH essential for activRy

Fig. 19. Model of sialic acid structural features required for the interaction with sialate lyase and binding of inhibitors. Based on data from ref. [892].

acylmannosamines and pyruvate. This reaction and the occurrence of the soluble enzyme in both bacteria and higher animals as well as the influence of sialic acid substituents and enzyme inhibitors has been reviewed [5,8,33]. The biological role of this lyase in bacteria is believed to be nutritive, by splitting the sialic acid 9-carbon chain as prerequisite for further consumption in the energy-producing metabolism. In animals, the main function of the cytosolic enzyme seems to be in the degradation of sialic acids in order to regulate sialic acid metabolism by prevention of recycling of this sugar. With regard to substrate specificity, the lyase from pig kidney, the first mammalian sialic-acid-specific aldolase purified [89 1a], shows similar behavior to the enzyme from Clostridium perfringens studied earlier [5,33]. The relative cleavage rates for NeuSAc, NeuSGc and Neu5,9Ac2 are loo%, 70% and 33%, respectively. Neu4,SAq and Neu2enSAc are inactive. An investigation into the kinetic behavior of various epi- and deoxy-analogues of NeuSAc with the lyase from C. perfringens revealed that modifications at C8 and C4 strongly reduce cleavage rates and affinity of the substrates to the enzyme, or are even inhibitory [409,892,893]. 4Epi- and 4d-NeuSAc were most inhibitory, with Ki values in the millimolar range. Computer-assisted drawings of these sialic acids (CPK models) indicate that the region most important for the binding of sialic acids to the enzyme is an equatorial zone stretching from C8 via the ring oxygen atom to C4 of the monosaccharide. The substituents at CS and C9 may be varied to a greater extent without significantly disturbing enzyme action. The structural features of sialic acid required for the interaction with the lyase [892] are depicted in Fig. 19. With regard to enzyme catalysis, for the C. perfringens lyase a Schiff-base mechanism involving a lysine residue was elucidated to be part of the catalysis of this enzyme [S, 331. The participation of a histidine residue in the cleavage reaction was also shown to be likely. The same amino acids probably are also parts of the active center of the aldolase purified from porcine kidney [891a], as was shown by borohydride reduction in the presence of NeuSAc and the influence of reagents such as Rose Bengal which interacts with histidine. Heavy metal ions and other substances were also inhibitory, pointing to the essential role of cysteine residues.

349

'

H NMR spectroscopic investigations have shown that only the a-anomer of NeuSAc is consumed by the C. perfingens lyase, yielding a-ManNAc, as described in detail in section 5.3.5. From this observation, the chemical modification experiments, and models drawn in refs. [33,892], a reaction scheme for the heterolytic fragmentation of cyclic sialic acids into pyruvic acid and N-acylmannosamines is delineated (Fig. 20, overleaf). Since the enzyme reaction is reversible, conditions for the synthesis of NeuSAc and natural or synthetic derivatives in high yield as well as of Kdn with the aid of bacterial lyase were elaborated (see section 6.1). The recombinant and overexpressed sialatepyruvate lyase from E. coli is now in wide use as a specific chiral catalyst which mediates highly enantioselective aldol condensation reactions leading to a variety of sialic acids. The enzyme first purified from E. coli[894] was cloned, the nucleotide sequence of its gene elucidated [895-8971 and later crystallized [898]. The lyase gene encodes a polypeptide of 297 amino acids [897]. The three-dimensional structure of this aldolase from E. coli has been investigated by X-ray crystallography and shown to be a tetramer, each subunit representing an eight-stranded a/B-barrel[899] (Fig. 2 1). The active center was tentatively identified as a pocket located at the carboxy-terminal end of this barrel. Lysine-165 lies within this pocket and is probably the reactive residue forming the Schiffbase intermediate with the substrate described above (see Fig. 20). Several additional amino acids were recognized to line this pocket. It is tempting to speculate that, as with sialidases, an exchange of the gene for the sialate lyase between mammals and microorganisms may have occurred. This assumption is based on the similarity of many properties of both bacterial and mammalian lyases studied so far. It has been mentioned in section 5.3.2 and refs. [11,12] that the commercially available sialate lyase is an excellent tool for the analysis, e.g. by HPLC or GLC, of a substance suspected to be a sialic acid. The enzyme has also proved to be useful in clinical determinations of these monosaccharides [900,90 11. For the analysis of glycoconjugatebound sialic acids, e.g. in the bloodstream, a combination with sialidase treatment is necessary. 9.4. Sialic acid permease

Although only few experiments have been carried out, sialic acids do not seem to be able to penetrate bacterial and animal cell membranes at significant amounts. Only minimal amounts of radioactivity from NeuSAc are taken up by surviving tissue slices from porcine submandibular gland [679]. Furthermore, Neu5Gc and Neu2en5Ac administered orally and intravenously into mouse and rat are excreted rapidly, mainly in the urine [902]. While the unsaturated sialic acid was excreted completely, less than 10% of the radioactivity from NeuSGc was found in the tissues. It is assumed that most of this portion was cleaved by the sialate-pyruvate lyase before its uptake by cells. Since sialic acids released from host glycoconjugates by sialidases are of advantage for the growth of bacteria, as described above, it is conceivable that these microorganisms developed a mechanism for the uptake of these anionic sugars [245]. First hints for the existence of a sialate permease were obtained with C. perfingens when studying the induction of both sialidase and sialate lyase with free or glycopeptide-bound

B

A

D

C

HOV AcHN

E

F

G

H

I

ManNAc + Pyruvate

Fig. 20. Proposed reaction scheme of sialate-pyruvate lyase from Clostridium perfringens. Based on data from refs. [33,892].

w 0 VI

351

Fig. L1. 1 hree-dimensional structure of the sialate-pyruvate lyase from Escherichia coli. (top) Viewed down the p-barrel axis from the carboxy-terminal end of this aldolase. The putative catalytic residue Lys-165 is shown in ball-and-stick representation. (bottom) Putative active site of NeuSAc lyase showing the side chains of nine of the residues forming the surface of the pocket. Carbon atoms are white, oxygens black and nitrogens grey. From ref. [899] by permission of Current Biology Ltd., London.

352

Neu5Ac [903]. A K M value of 0.3 mM for Neu5Ac was measured for this hypothetical transporter protein. This question has not yet been solved in C. perfringens, although recent experiments have also shown that sialic acid uptake is not due to mere diffusion (C. Traving et al., unpublished). However, in E. coli evidence for a sialic acid permease was obtained, based on molecular biological studies [585,904].

10. Physiological and pathobiochemical signijicance of sialic acids The growing understanding of the involvement of sialic acids in many biological and pathological processes has considerably enhanced interest and research in this field. It started in 1946 when McFarlane Burnet and colleagues [905] recognized the reversible binding of influenza viruses to mammalian erythrocytes via cell surface sialic acids. Several reviews on the biological aspects of sialic acid biochemistry have appeared [5,8,245,660,663,690,853,90&909]. From these articles it is clear that there are general functions of sialic acids resulting from their physico-chemical properties such as electronegativity, hydrophilicity and relatively large size, coupled with their exposed position on cells and secreted macromolecules. However, there is a growing number of examples where sialic acids exert specific functions, mostly in molecular and cellular adhesion phenomena. The influence of natural and artificial modifications of the structure of Neu5Ac on enzymatic and some biological phenomena has been mentioned at relevant places throughout this article, showing that the configuration of NeuSAc and related molecules is of great importance for the metabolism and biology of these monosaccharides, including those of the whole glycoconjugate molecules, cells and tissues, to which sialic acids are bound. When trying to summarize the numerous roles of sialic acids, three groups of main effects can be described. Although there is considerable overlap, this classification has proved to be very useful and has been supported by many recent examples. Due to lack of space, only a few characteristic examples will be given here and new trends will be emphasized. 10.I . General physico-chemical effects

Firstly, due to their high acidity (pK value around pH 2), sialic acids are fully dissociated under physiological conditions and act as anions for the binding of inorganic and organic cations, e.g. on cell surfaces, thus facilitating the cellular uptake of substances with physiological or pharmacological significance. Cell aggregation may be prevented by the repulsive effects of the negatively charged sialic acids, as was observed for example when studying the adhesion of cultured cells to their substratum [910]. These cells adhered better to a layer of desialylated gangliosides than to the native, negatively charged, species, or after desialylation of the cells. The prevention of erythrocyte aggregation by surface sialic acids is a further, excellent example of their repulsive effect [911,912]. After the removal of sialic acids by sialidase, erythrocytes were found to aggregate readily, either without the influence of other compounds or under the influence of immunoglobulins and fibrinogen, respectively. Correspondingly, erythrocytes which are undersialylated in

353

diabetes, as was analyzed with their glycophorin components, tend to increase aggregation and adhesion to blood vessel walls. This may be of importance in the development of vascular disease during diabetes [913]. The high degree of sialylation existing in the epithelial sialomucin “episialin”, which is expressed not only on several secretory epithelial cell types but also in many carcinoma cell lines, is also believed to have antiadhesion function. This effect may facilitate metastasis of tumor cells [914]. Such a role can also be attributed to the sialic acids found on the surface of Streptococci of the serological group B. In contrast to the wild-type strain, an asialo mutant obtained by the transposon method showed stronger cell adherence, probably due to the more hydrophobic surface resulting from the lack of sialic acid [915]. In contrast to this repulsive effect of sialic acids, cell adhesion may be facilitated via positively charged substances or Ca2+ bridges. Furthermore, binding of Ca2+ to ganglioside sialic acids is of great importance in the function of nervous tissues [916]. These glycoconjugates are localized in clusters on neuronal and especially synaptic membranes in the vicinity of a membrane-bound calcium pump, thus facilitating the supply of these ions for neuronal cells. These calcium-ganglioside interactions may modulate neuronal functions, not only for the short-term process of synaptic transmission of information, but also for long-term events of neuronal adaptations including storage of information. The highly charged anionic sialic acid residues, a-2,3- and a-2,6glycosidically bound to the trabecular meshwork of the human eye, are considered to be involved in the regulation of the aqueous outflow and the control of intraocular pressure [917]. Secondly, sialic acids may modulate the conformation of glycoproteins and gangliosides, thus influencing their arrangement, physical properties and biological functions in cell membranes, as well as the solubility, thermal stability and other properties of soluble glycoproteins including sialylated enzymes and hormones [5,8,907]. This is, indeed, a very large field and we still are only beginning to understand the influences of glycoconjugate conformations in cell biology. To obtain such information on individual molecules existing on living cells is particularly difficult to obtain with the techniques available. However, with a-dodecyl-Neu5Ac anchored to the surface of a phospholipidbased membrane fragment, it was shown by ‘H and 13C NMR studies that the sialic acid pyranose ring is extended into the aqueous phase with the carboxyl group remaining close to the membrane surface, where it may interact with a hydrogen acceptor [918]. For glycoproteins it has been postulated that the carbohydrate moiety mimics the effect of a molecular chaperon [919]. The anti-proteolytic effect, long known [906] to be exerted by glycan chains on glycoproteins in general and sialic acids in particular, should also be mentioned here. This may be due to conformational changes mediated by sialic acids, to their negative charge, or to mere shielding of e.g. peptide bonds from proteases by the bulky carbohydrate moiety. The high grade of sialylation of lysosomal cell membranes[920] may be a protective barrier against proteases in this cellular compartment. Glycosylation of influenza virus hemagglutinin governs hydrolysis of certain peptide bonds by host proteases in order to enhance the virulence of these viruses [863,921]. It is suggested that the removal of sialic acid from the carbohydrate moiety of the hemagglutinin of influenza B virus by sialidase is essential for the cleavage of this glycoprotein

354

by cellular protease [922]. Mucins, which are highly glycosylated and often heavily sialylated [5,13,659,906] are excellent anti-proteolytic substances of high viscosity, protecting endothelia particularly in the intestine or respiratory tract from the attack by proteases and pathogenic substances [5,8,245,660,853]. The high viscosity of mucins is due, to an extent, to their sialic acid and often also sulfate content, or, in lower animals, to the presence of other anionic components such as uronic acids, sulfate or phosphate residues [5,853,906,923]. This property enables mucins to act as very efficient biological lubricants, thus for example facilitating the transport of foodstuffs through the intestine, assisting in the movement of the eye bulbus, facilitating the mechanical processes of reproduction, or serving as protection for fish and many lower animals, the body surfaces of which must be lubricated for movement in aqueous or muddy environments. The structure, biosynthesis and molecular biology of mucins expressing cell- and tissue specificity have been reviewed [722,924]. The next two large groups of sialic acid functions are two-fold: sialic acids are either masks of biological recognition systems or they are directly recognized during the interaction of molecules and cells [663,925]. Since the masking functions of these sugars were recognized earlier, they will be described first. 10.2, Sialic acids masking biological recognition sites It is well established that the main reason why sialic acids occur chiefly on membrane surfaces as well as on circulating and secreted molecules is because of the protective role these highly acidic and relatively bulky sugar molecules play. In this way, cells and molecules are protected against damaging influences from the environment, such as proteases and immunological or phagocytotic agents. This may be the reason why so many microorganisms such as trypanosomes (see section 9.2.4. l), bacteria (see below) and even pathogenic fungi [92&928] also developed mechanisms to sialylate cell surface components. Remarkably, in the fungi Sporothrix schenkii [926] and Fonsecaea pedrosi [928], both of which are pathogenic to man, NeuSAc and NeuSGc were found. The protective role of sialic acids was shown in the case of S. schenkii, whose phagocytosis by macrophages was increased after sialidase treatment [929]. In addition to this more general function, sialylation can mask specific biological recognition sites and, by reversible sialylation, can control biological events such as the action of hormones, ion channels, growth and ageing, some examples of which will be discussed in the following. One of the most frequent biological recognition sites of glycoconjugates, which are often masked by sialic acids, are galactose residues. These monosaccharides, after their exposure by sialidase or reduced sialyltransferase activity, interact with a variety of galactose-specific receptors and are thus involved in many physiological and pathological events (for reviews see refs. [5,8,663,925,93&934]). The galactose-specific lectin of animals was found to be a phylogenetically very old protein, with homologues occurring in organisms ranging from the sponge Geodia cydonium to man [935,936]. One of the best known galactose-specific receptors is the hepatic asialoglycoprotein receptor, which was the first well-characterized animal lectin [933,937]. The properties, molecular biology and biosynthesis of this Ca2+-dependent receptor have been intensively studied (e.g.

355

ref. [938]). It is proposed to function as a “vacuum cleaner” for the removal of serum glycoconjugates, which had become “non-self’ and have exposed galactose residues, resulting from the loss of sialic acids. However, this lectin also mediates the hepatic binding and uptake of hepatitis B virus particles by the same mechanism[939]. This clearance of asialoglycoconjugates is also a problem during the therapeutic administration of recombinant glycoproteins, including those with hormonal or enzymatic activity, or the follicle-stimulating hormone isolated from urine [94&945]. Only the fully sialylated glycoproteins exhibit a long life in serum and thus extended biological activity, in contrast to those possessing terminal mannose or galactose residues. In the latter cases, rapid removal mediated by carbohydrate-specific receptors takes place. In contrast to this positive effect of sialic acids on the circulation time and the biological activity of these glycoprotein hormones, the same sugar negatively influences the hormonal effect of e.g. thyrotropin [940], erythropoietin [944] and prolactin [946] in uitro. It was shown in these studies that asialo-erythropoietin binds better to the hormone receptor than the sialylated form. Deglycosylated human lutropin and chorionic gonadotropin also bind better to the receptors than the native form, but their hormonal activities, such as CAMP formation and steroidogenesis are lost [947]. The presence of sialic acid residues is a functional requirement of the hormone molecules and may modulate the efficiency of signal transduction. It was shown in these studies that a sialic acid residue from a chorionic gonadotropin glycan either binds to the hormone receptor directly or to a neighboring lectin-like molecule. This interaction is essential for signal transduction mediated by the hormone. A comparable receptor which is responsible for the sequestration of sialic-acid-depleted blood cells, such as erythrocytes [948], lymphocytes [949] and thrombocytes [43] exists on the surface of liver and spleen macrophages, and also on the peritoneal phagocytes of some animals [925,950,95 I]. This receptor was isolated from rat peritoneal macrophages [952]. It mediates binding of desialylated blood cells via demasked galactose and N-acetylgalactosamine residues by macrophages and requires Ca2+ for this function. Immunoglobulin or complement factors do not influence this process. Although sialidase-treated erythrocytes and thrombocytes are phagocytozed after adhesion [43,948], lymphocytes are released from the macrophages within 24 h after binding [949], probably due to the resynthesis of their sialic acid moieties. The binding of the three cell types to phagocytes is shown in Fig. 22. This lectinophagocytosis not only acts on blood cells damaged by sialidase, which may have resulted from viral or bacterial infections (see above), but also seems to be responsible for the removal of aged human erythrocytes [953]. It was shown with lectins and by resialylation with 9-fluoresceinyl-Neu5Ac that old blood cells expose more terminal galactose residues compared with young erythrocytes. Loss of sialic acid residues from glycophorin during the life-time of human erythrocytes seems to be responsible for the clearance of senescent red blood cells from circulation [954]. Injection of sialidase into mice induces lectin-mediated thrombocytopenia, which stimulates thrombopoiesis [955]. As a consequence of programmed cell death (“apoptosis”), changes of surface carbohydrates including a loss of sialic acids and exposure of galactose residues occurs leading to phagocytosis of apoptotic cells [956]. In neonatal rat liver cell cultures, the asialoglycoprotein receptor has been implicated in the ingestion by healthy hepatocytes

356

Fig. 22. Binding of sialidase-treated rat erythrocytes, lymphocytes and thrombocytes to homologous peritoneal macrophages by the galactose-specific receptor. (a) Scanning electron microscopy of erythrocyte-macrophage interaction (from ref. [925] by permission of Kodansha Publishers, Tokyo). After prolonged incubation, the erythrocytes were ingested. (b) Scanning electron microscopy of cultured, sialidase-treated lymphocytes (L) bound by two macrophages (M). Lymphocytes were not phagocytozed but separated from the macrophages within one day [949]. (c) Scanning electron microscopy of sialidase-treated thrombocytes to a macrophage adherent to a Petri dish. From ref. [925] by permission of Kodansha Publishers, Tokyo.

351

Fig. 22c.

of neighboring cells undergoing apoptosis [957]. After desialylation of rooster sperm glycocalyx, the fertility of the spermatozoa was reduced due to an increased rate of sequestration in the hen’s reproductive tract when compared with untreated cells [958]. This effect is believed to be due to demasking of antigenic or other recognition sites on the spermatozoa. However, treatment of human motile spermatozoa with sialidase results in better attachment of these cells to the zona pellucida of oocytes [959]. It is concluded from this observation that the release of sialic acid from the sperm plasma membrane could be one of the capacitation events necessary for unmasking certain sperm surface antigens or carbohydrate ligands implicated in zona pellucida recognition. The surface of human spermatozoa is highly sialylated, containing predominantly Neu5Ac(a2-6)Gal/GalNAcglycoconjugate sequences [959,960]. It was also shown with a sialylated wild-type and a mutant Streptococcus strain that sialic acids hinder phagocytosis of the bacteria and thus increase their virulence [915]. The galactose receptor also enables rat peritoneal macrophages to take up either free [96 I ] or gold-particle-bound asialoglycoproteins [925,950,962]. Interestingly, for activity this receptor requires the presence of sialic acids on the macrophages to which it is bound, as was shown by sialidase treatment [925]. No explanation for this phenomenon is available so far. It may be due to the requirement of sialic acid for stabilization of the conformation of the receptor, to prevent self-aggregation, if it is a glycoprotein, or to hinder interaction with neighboring desialylated molecules in the macrophage membrane. The interaction of galactose-exposing particles with rat Kupffer cells has also been characterized [963]. From the reversible binding of lymphocytes mentioned above, as well as from the inhibition of binding of erythrocytes by enzymatic resialylation [964], a model has been proposed for the reversible binding of cells via sialic acid and galactose, which is

358

Fig. 23. Model showing how sialidase and sialyltransferase may regulate cellular interactions, i.e. association and dissociation mediated by a galactose-recognizing system. Symbols: solid square, galactose; open oval, sialic acid; open castle, galactose-recognizing receptor; dashed sausage, hypothetical recognition site (cluster of three galactose residues) by the lectin. From ref. [965]by permission of Gustav Fischer Verlag, Jena.

regulated by the activity of sialidase and sialyltransferase [925,950] (Fig. 23). This may generally operate in cell communication, especially in growth, differentiation, ageing, malignant transformation and metastasis. The protection of host cells by sialic acids from the interaction resulting in colonization and possibly infection with bacteria is illustrated by the adhesion of Bacteroides intermedius to sialidase-treated human buccal epithelial cells and erythrocytes [966]. A corresponding phenomenon was described above in the sialidase section 9.2.4 for another oral bacterium, Actinomyces uiscosus, which is also capable, by sialidase secretion, to demask its own attachment sites on host cells. Pseudomonas aeruginosa, a dangerous, infectious agent in patients suffering from cystic fibrosis or other diseases, employs pili to adhere to asialo-GM1 [967,968]. As sialic acid interferes with binding, GM1 itself is not a ligand for this bacterium. Sialic acid residues on cell surfaces are important as modulators of immune and complement reactivity. Most importantly, animal cells and some microorganisms are shielded from the influence of these defence systems, since sialic acids can mask corresponding recognition sites. This has many physiological and pathological implications, as it may prevent autoimmune diseases, but can also increase the virulence of microbial infections or of cancer.

359

A classical example for this protective effect is the interaction of the non-immune channel fish lctalurus punctutus [969] with various Gram-negative bacteria. Very little bactericidal activity is exerted by the alternative complement pathway against the fish pathogens containing sialic acid, in contrast to a very strong bactericidal response against the non-pathogens lacking sialic acid. To show that only the sialic acid coat protected the bacteria against the complement system and thus rendered them virulent, the bacteria were treated with sialidase. This destroyed their virulence, due to an enhancement of the bactericidal response of the fish. The virulence of Gonococci (Neisseria gonorrhea) is also much increased by sialylation of the bacteria [ 184,185,970]. Terminal Gal(fi14)GlcNAc structures of the bacterial 4.5 kDa lipopolysaccharide are sialylated during the life-time of the bacteria in the host organism. In this way, Gonococci in patients become resistant to complement-mediated killing by serum and to phagocytosis [971]. This resistance is lost in subcultures of the bacteria, but can be acquired again in the host. It is also absent in mutants lacking the necessary lipopolysaccharides. However, after infection with such Gonococci, variants appeared in the patient again expressing the acceptor for sialic acids, which shows a fascinating capability of the bacteria to adapt to the life in the host [970]. Closer investigation has revealed that sialic acids also mask porin epitopes on the Gonococci, which results in the reduction of the binding and bactericidal activity of antibodies recognizing this peptide moiety [972]. Furthermore, binding to human neutrophils and killing of these bacteria is prevented after sialylation [973]. Curiously, this chemical modification of Gonococci living in the host is possible by a sialyltransferase expressed by the bacterium itself and the sialic acid donor substrate, CMP-NeuSAc, produced by the host. The properties of the sialyltransferase extracted from the bacteria have been reported [974]. It would be very interesting to compare the primary structure of this enzyme with those of host sialyltransferases. This mechanism is different from that used by trypanosomes for sialylation, although in both cases sialic acids are acquired from the host. One of the major functions of sialic acids on mammalian cells is their masking effect on immunological recognition sites, as has been reviewed [5,8,663,795]. Loss of sialic acids, which may occur by exogenous sialidases from infections by microorganisms and viruses (see section 9.2.4), can thus lead to autoimmune diseases. One example is glomerulonephritis of human kidney, probably resulting from streptococcal infections. We are only at the beginning to understand such pathobiochemical mechanisms. It was also shown that cultured epithelial cells from guinea pig after treatment with bacterial sialidase bind more IgM than the controls [795,975]. These cells and those from lung epithelium taken from influenza-virus-infected animals were found to be hyposialylated. In serum, the level of IgG autoantibodies binding to the sialidase-treated cells increased [975]. It is assumed that these autoantibodies participate in cellular disfunctions and modify bronchoreactivity occurring during the infection of the respiratory tract. Asthma and chronic pulmonary disease may be the consequence of such desialylation. In these processes, the parasympathetic tone of the airways and the affinity of muscarinic agonists is impaired and the interaction of acetylcholine and histamine with their receptors also seems to be disturbed [795,976].

3 60

10.3. Sialic acids representing biological recognition sites

Whereas for a long time the main function of terminal sialic acid residues was considered as a mask for recognition sites on cell surfaces, e.g. galactose residues or antigens as outlined above, it is only in the last few years that receptors have been described which connect the structural diversity of sialylated glycoconjugates to specific functions in cellular interactions. 10.3.1. Sialic acid receptors of microorganisms, plants and lower animals Many microorganisms, plants and animals express proteins which bind to sialic acids as components of glycoconjugates. Such sialic-acid-recognizing proteins or lectins from microorganisms, plants and invertebrates (mainly snails, scorpions and crabs) have been summarized in Tables 20 and 21, and have been discussed elsewhere in this series. They show more or less specific binding to different sialic acids or even specific sialic acid linkages and oligosaccharide sequences, as reviewed [709]. A careful investigation of the binding determinants of the sialic-acid-specific lectin from e.g. the slug L i m a jauus has revealed that the a-anomer, the N-acetyl group, and an intact side chain of sialic acid are necessary for high affinity interaction. An axial position of the hydroxyl group at C4, hydroxylation of the N-acetyl group or substitution by a formyl residue, as well as 9-0-acetylation or periodate cleavage of the glycerol chain of NeuSAc considerably reduced binding to the lectin[224]. Since these organisms do not express sialic acids themselves, it is unlikely that these lectins play a role as sialic-acid-binding proteins in cell recognition. However, they may be involved in defence against sialic-acid-containing microorganisms and animals. Some of the sialic-acid-binding lectins have proved to be useful tools in the analysis and histochemistry of glycoconjugates. Most frequently used are the agglutinins from wheat germ (WGA), LimafIauus (LFA), Sambucus nigra (SNA) and Maackia amurensis (MAA) (see section 4). Initial insight into the putative sialic-acidbinding sites of these unique lectins can be gained from a comparison of the primary structure of MAA with other plant lectins [lOOl]. A conserved asparagine residue was found to be exchanged by aspartate (Asp-135), which is proposed to be involved in the interaction between MAA and sialo-oligosaccharides. It has long been known that pathogens, i.e. viruses, mycoplasma, bacteria and protozoa, take advantage of cell surface sialic acids to adhere to their respective host cells (for summaries see refs. [660,1002,1003] and chapter 13 of the present volume). Of greatest importance in this connection is the role of sialic acids in the attachment to and infection of mammalian cells by influenza viruses (see section 9.2.4.3). From 500 MHz ‘H NMR spectroscopic [I0041 and from X-ray crystallographic studies [1005,1006], models for the position of sialic acid in the binding pocket and the interaction with certain amino acids of influenza A virus hemagglutinins were delineated (Fig. 24) [394]. The use of sialic acid analogues modified in the glycerol side chain and in the N-acetyl and in the carboxyl groups of NeuSAc, gave support to these models and revealed that the hydroxyl group at C9 does not interact with the hemagglutinin [394]. The N-acetyl moiety is critical for the interaction of sialic acids via hydrophobic bonds. These studies were extended by investigation of the crystal structures of this influenza virus hemagglutinin with different high-affinity analogues

361 Table 20 Pathogenic microorganisms and toxins binding to sialic acids on host cells Pathogen

Viruses Influenza A and B

Specificity

Ref(s).

Neu5Ac (some strains prefer NeuSAc(a2J)Gal or NeuSAc(a2-6)Gal, dependent on host specificity)

Influenza C

Neu5,9Ac2

Corona virus

Neu5,9Ac2

Sendai virus

NeuSAc

Polyoma virus

NeuSAc(a2-3)Gal(~I-3)GalNAc

Rotavirus group C

NeuSAc

Mycoplasma Mycoplasma pneumoniae

NeuSAc(a2-3)Gal on polylactosamine chains

Bacteria

Streptococcus sanguis Escherichia coli K99 Escherichia coli, S-fimbriae (newborn human meningitis) Bordetella bronchiseptica Pseudomonas aeruginosa

0-linked sialylated tetrasaccharides NeuSGc-containing glycolipids NeuSAc(a2-3)Gal(/31-3)GalNAc NeuSGc-GM3, GD3, GDlb NeuSAc NeuSAc

Helicobacter (Campylobacter) pylori

NeuSAc(a2-3)Lactose > NeuSAc(a2-6)Lactose

Streptococcus suis

NeuSAc(a2-3)Gal(/31-4)GlcNAc(~1 -3)Gal

Protozoa

Plasmodium falcQarum (Malaria, MSA-I)

NeuSAc

Trypanosoma cruzi (Chagas disease) Tritrichomonas species

NeuSAc NeuSAc (a-2,3- and a-2,6-linkages)

Toxins

Vibrio cholerae toxin

GM 1

Petusis toxin

Neu5Ac

Tetanus toxin

sialoglycolipids

of sialic acid in order to design possibly therapeutic inhibitors of viral attachment to host cells, which may prevent membrane fusion and circumvent evasion of inhibition by antigenic variation of the viruses [ 10061. A series of poly(acry1ic acid-co-acrylamides) and dextrans bearing pendant glycine-4-amidobenzyl a-NeuSAc groups were synthesized for anti-influenza chemotherapy. Some of these compounds suppress virus replication in embryonated eggs [ 10071. X-ray crystallographic studies of influenza virus sialidaseantibody complexes gave interesting insight into the structure of such complexes between two cross-reacting antibodies both recognizing sialidase [ 10081.

362 Table 21 Sialic-acid-binding lectins from plants and invertebrates (see ref. [709] and references therein) Source

Specificity

A: Plants

Wheat germ Trificum oulgare

NeuSAc < GlcNAc

Elderberry Sambucus nigra

NeuSAc(a24)Gal/GalNAc

Maackia amurensis

NeuSAc(a2-3)Gal(fi14)GlcNAc

B: Invertebrates Snail Dolabella

NeuSAc

Slug Limaxflaous Snail Cepaea hortensis

NeuSAc > NeuSGc

Snail Achafinafulica

NeuSAc(a2-3)Gal> NeuSAc(a2-6)Gal

Snail Pila globosa

NeuSGc

Oyster Crassosfrea gigas

NeuSAc

Horseshoe crab Limulus polyphemus

NeuSAc

Lobster Homarus americanus

NeuSAc, NeuSGc

Horseshoe crab Tachypleus tridentatus

NeuSAc, NeuSGc

NeuSAc > NeuSGc

Scorpion Androcfonus ausfralis

sialyllactose

Horseshoe crab Carcinoscorpius rotunda

NeuSAc(a24)Gal> NeuSAc(a2J)Gal

Scorpion Centruroides sculpturatus

NeuSAc, NeuSGc

Prawn Macrobrachium rosenbergii

NeuSAc

Scorpion Masticoprocfus giganteus

NeuSAc

Spider Aphonopelma cepaeahortensis

sialoglycoproteins

Scorpion Heterometrus granulomanus

NeuSAc(a2-3)Lactose

Prawn Peneaus monodon

NeuSAc

Scorpion Parurocfonus mesaenis

sialoglycoproteins

The interest in sialic-acid-specific adhesion of bacteria is increasing, since it often is a critical step in infectious diseases. Examples are the inflammation of gastric mucosa by Helicobacter pylori after adhesion to sialoglycoproteins of the cell surface [993,1009] and meningitis of infants as well as urinary tract infections by E. coli [985-9891. The binding of the latter bacteria via their S fimbriae to epithelial, e.g. buccal cells can be inhibited by sialylated mucins from human milk. Correspondingly, in human newborn babies sialylated substances have an anti-infective effect. Colostrum was shown to be more effective in this respect than mature milk [1010]. The S fimbrial adhesin of E. coli represents a protein complex and the genes encoding its subunits have been sequenced [988]. The introduction of specific mutations into the subunit gene sfaS revealed that a region of six amino acids of the adhesin which includes two lysine and one arginine residues is involved in the interaction of S fimbriae with sialic acid [990]. It should also be mentioned that several bacterial toxins are known which bind to gangliosides in a sialic-acid-dependent manner, e.g. cholera, pertussis and tetanus toxins

363

Fig. 24. Model for the position of sialic acid in the binding pocket of the hemagglutinin of influenza virus, based on X-ray crystallographic studies. Some of the hydrogen bonds proposed in this model are shown by dashed lines. From ref. [394] by permission of European Journal of Biochemistry, Zurich.

(ref. [660] and Table 20). The role of host sialic acids in the binding of protozoa is also of growing interest, and the possible involvement of this monosaccharide in the attachment of trypanosomes to host cells during infection has already been mentioned (section 9.2.4). The role of sialic acid in the invasion of erythrocytes by malaria parasites has been demonstrated and a 175 kDa protein of Plasmodium falciparum has been identified as an erythrocyte-binding protein with sialic acid specificity [ 10111. It preferentially binds to NeuSAc(a2-3)Gal sequences on the 0-linked tetrasaccharide of glycophorin. Tritrichomonads (Tritrichomonas suis, foetus and mobiliensis) also exhibit a sialic-aciddependent adhesion to host cells and mucus 11226,9961. A corresponding lectin consisting of three subunits was isolated from Tritrichomonas mobiliensis, which recognizes free as well as a-2,3- and a-2,6-linked Neu5Ac [226]. A model of binding was delineated from the observation showing that optimal interaction with the receptor requires a free

364

carboxylic group, the N-acetyl moiety and a free H 0 8 group of sialic acid. Remarkably, 0-acetylation at C9 or C4 does not influence adhesion. 10.3.2. Sialic acid receptors of vertebrates Sialic-acid-recognizing receptors have been discovered in a variety of vertebrates, mostly mammalian tissues, summarized in Tables 22-24. This line of research was much stimulated by the discovery of a family of cell adhesion molecules now generally called selectins[l033]. Due to its great importance, this topic has been reviewed frequently [660,828,1034-10391. It will, therefore, not be discussed in detail here. The selectin family consists of three members, L-selectin, found on lymphocytes and other white blood cells, E-selectin expressed on endothelial cells activated by cytokines, and P-selectin occurring on activated blood platelets and endothelia (Table 23). They possess a similar primary structure, representing type I trans-membrane glycoproteins containing an amino-terminal carbohydrate recognition domain, a single epidermal growth-factor-like domain, a variable number of short consensus repeats and a relatively short carboxy-terminal cytoplasmic domain. The selectins also share common aspects in their function. They all play crucial roles in the initial event of white blood cell adhesion to specific endothelia, so-called rolling. Before firm adhesion, cells floating in the blood stream begin to slow down by rolling along the endothelial lining of the vessel. This is mediated by selectins interacting with sialic-acid-containing ligands. The specificity of this interaction is accomplished by the expression pattern of the receptors and their appropriate ligands. The selectins bind with low affinity to many small, sialylated, fucosylated and in some cases sulfated carbohydrates, of which the prototypes are the tetrasaccharides sialyl-Lewis’ {sLeX;Neu5Ac(a2-3)Gal(fil4)[Fuc(a1-3)]GlcNAcfi-R} and the isomeric sialyl-Lewisa {sLea;Neu5Ac(a2-3)Gal(fil-3)[Fuc(a14)]GlcNAc~-R}. At least two of the selectins (Table 23) recognize sLe’ andor sLea, which are expressed at high levels on leukocytes [ 10401 and some tumor cells, particularly of the colon [ 104 110441 as components of glycoproteins or gangliosides. Sialyl-Lewis’ has also been shown to be a marker of dysplasia in the colonic adenoma-carcinoma sequence [1045]. It has therefore been proposed that selectins are involved in malignant growth and metastatic events. Remarkably, 6’-sulfated sLeX(Table 23) has been identified as a major ligand (GlyCAM-1) involved in the initial attachment of lymphocytes via L-selectin to high endothelial venules of lymph nodes [ 10461. The synthesis of this “addressin”, as L-selectin ligands are called now, 0-sulfated at C6 of galactose, has been reported [1047]. Using an L-selectin chimera, it could be shown that ligands for L-selectin also occur on the myelin sheaths of neurons of the central nervous system [1048]. Progress has been made in understanding how these molecules function at the atomic level, by determination of the three-dimensional structure of a portion of the E-selectin molecule [ 1049,10501. A spin label study has shown that the interaction of selectin on lymphocytes with sialic acid severely restricts the rotational mobility of the cell surface proteins and lipids. Additionally, the cytoplasmic viscosity increases appreciably [ 10511. Presently, much interest is focussed on the role of the selectins during inflammation and in transplantation medicine, since it was recognized that selectins are involved in the rolling and adhesion of granulocytes to endothelia under the influence of inflammatory cytokines [ 1034,1037,1039,1052]. This is a multistep process in which other adhesion

365 Table 22 Vertebrate sialic-acid-binding proteins a Source

Specificity

Reference(s)

Frog egg

sialylated glycoproteins

[I0121

Rat uterus

NeuSAc

[I0131

Rat brain

NeuSAc, NeuSGc

[I0141

Rat brain myelin

gangliosides, preferentially GTI b, GQI b, GDlb

[IOIS]

Human endometrium

NeuSAc, NeuSGc

[I0161

Human placenta (IgG)

0-acetylated sialic acids

[lo171

Blood (factor H of alternate complement pathway)

sialylated glycoconjugates, other polyanionic molecules

[I0181

Murine macrophages (sialoadhesin)

NeuSAc(a2-3)Gal(fi 1-3)GalNAc on glycoproteins and glycolipids

[lo191

Bovine heart (calcyclin)

NeuSAc, NeuSGc

[ 10201

Human placenta (sarcolectin)

NeuSAc, NeuSGc

[I0211

B lymphocytes (CD22)

NeuSAc(a2-6)Gal(fl 14)GlcNAc

[ 1022,10231

a

For selectins and sialoadhesins see Tables 23 and 24. Table 23 Distribution and binding specificity of selectins (for references see the text)

Selectin

Cell type

Ligand determinant

E-selectin

activated endothelia

sialyl-Le', NeuSAc(a2-3)Gal(fiI4)[Fuc(a 1-3)IGlcNAc sialyl-Lea, NeuSAc(a2-3)Gal(fiI-3)[Fuc(a I4)]GlcNAc

L-selectin

leukocytes

sialylated, sulfated and fucosylated 0-glycans like

NeuSAc(a2-3)(SO4-6)GaI(fll4)[Fuc(al -3)IGlcNAc P-selectin

activated platelets and endothelia sialyl-LeXand sialyl-Lea

molecules of the integrin superfamily [I0531 are also involved. Such reactions may occur under the influence of microbial infections [ 10541 or other inflammatory events, such as in skin and neuronal tissue [I0551 or in lung injury [1056]. The application of sLeXoligosaccharide has thus been shown to have a protective effect in such inflammatory reactions, e.g. lung inflammation due to oxygen radical formation by cobra venom. From this research, a therapeutic effect of sLeX analogues in the treatment of many diseases including reperfusion injury observed after organ transplantation, heart attack and stroke is expected, as shown in animal models (ref. [ 10571 and J.C. Paulson, personal communication). Consequently, much work is dedicated to the synthesis of selectin ligands on a commercial scale (ref. [lo581 and section 6.3). Another strategy to prevent

Table 24 Sialoadhesins, a family of immunoglobulin-like adhesion molecules binding to sialylated glycans Adhesion molecule

Occurrence

Sialoadhesin

macrophage subpopulations

Myelin-associated glycoprotein (MAG)

Number of Ig-domains

Ligand determinant

Target cells

References

17

Sia(a2-3)Gal(f31-3)GalNAc Sia(a2-3)Gal(fl1-3/4)GlcNAc

myeloid cells

[I0191 [664,1024,1025]

myelin of oligodendrocytes and Schwann cells

5

Sia(a2-3)Gal@-3)GalNAc

neurons, oligodendrocytes

[664,1026]

CD22

B cells

7

Sia(a24)Gal(f314)GlcNAc

lymphocytes

[ 1023,1027-10301

CD33

myelomonocytic cells

2

Sia(a2-3)Gal(p1-3)GalNAc Sia(a2-3)Gal(P1-3/4)GlcNAc

myelomonocytic cells

[I0311 [ 10321

367

selectin-mediated adhesion phenomena in pathological states is the neutralization of the biological activity of these receptors by antibodies [ 1059,10601. Other well defined sialic-acid-dependent adhesion receptors are sialoadhesin [660,664, 1019,1024] found on specific subsets of macrophages in bone marrow and lymphatic tissues, such as lymph nodes and spleen, as well as CD22 [660,1025,1061,1062], a B-cell-specific protein, CD33, occurring on myelomonocytic cells [ 1031,10321 and MAG, the myelin-associated glycoprotein of neuronal tissues [ 1025,10631 (Table 24). All these receptors belong to the immunoglobulin superfamily [ 10251. The murine macrophage sialoadhesin has a molecular mass of 185 kDa as well as 17 immunoglobulin-like domains [ 10241 and binds to the sequence NeuSAc(a2-3)Gal(fl 1-3)GalNAc or NeuSAc(a23)Gal(@l-3/4)GlcNAc, i.e. both N- and 0-glycans on glycoproteins and glycolipids of cell surfaces, preferentially of the granulocytic lineage [ 10641. Remarkably, Neu5Gc and Neu5,9Ac2 are not recognized by the receptor[664]. With regard to its function, it may play a role in hematopoiesis, since it was found to be enriched on bone marrow macrophages at the contact sites with developing myeloid cells. A role in the trafficking of leucocytes in lymphatic organs is also assumed. The receptor preferentially binds to inflammatory and circulating neutrophils. CD22 occurs on B cells, where it mediates their binding to T cells and to neutrophils, monocytes and erythrocytes [1065]. The interaction with T cells is assumed to be involved in early B cell activation and may also modulate signalling through the surface IgMcell receptor complex. In contrast to sialoadhesin, CD22 binds specifically to sialyl(a26)Gal(b14)GlcNAc structures of N-glycans [ 1023,10251. Again, Neu5Gc and Neu5,9Ac2 modulate binding of this ligand. Whereas Neu5Gc increases binding to murine CD22, 9-0-acetylation of NeuSAc is inhibitory[664]. It can only be assumed so far that these sialic acid modifications, both occurring in mice, have a biological role in the activity of the two receptors. Recognition of Neu5Ac by human CD22 is also masked by 9-0-acetylation [235]. A further regulation of CD22 function is a-2,6-sialylation of the glycan chains of the receptor molecule itself [ 10661, which inhibits receptor activity. In this way, the ligand binding sites of CD22 may be blocked and the interaction of the receptor with ligands on adjacent cells prevented. In resting B cells, which only express low levels of a-2,6-sialyltransferase activity, CD22 binding sites may be free, thus enabling interaction with neighboring cells. Lymphocyte activation, however, results in an increase in the activity of this sialyltransferase, causing sialylation of the receptor and a decrease in cellular interaction. Regulation of the sialylation of CD22 may therefore mediate adhesion and trafficking at discrete stages of B cell differentiation. In this regulatory system the expression of sialylated ligands in non-lymphoid cells may be important. It was shown that inflammatory cytokines stimulate the expression of both ligands for CD22 and 6-galactoside a-2,6-sialyltransferase in endothelial cells of human umbilical vein [ 1028,10291. Thus, CD22 could direct interactions between mature B cells and endothelial cells during inflammatory stages. These events can be influenced, i.e. lymphocyte binding inhibited, by a-2,6-sialylated glycoproteins of the blood serum. Interestingly, the two serum molecules much involved in immune and inflammatory responses, haptoglobin and IgM, are selectively recognized by lymphocyte CD22 [ 10291. Using a soluble CD22-immunoglobulin fusion product or expression of CD22 on the surface of Chinese hamster ovary cells, it was shown that this lectin has a higher

368

apparent affinity for multiply sialylated substances over monosialylated, unbranched glycans [ 10301. This observation provides a mechanism for strong CD22-dependent cell adhesion despite the relatively low Kd for protein-carbohydrate binding. The myelin-associated glycoprotein (MAG), involved in the proper myelination of axons [ 10631, has recently been shown to be a sialic-acid-dependent receptor, recognizing a-2,3-glycosidically bound NeuSAc, preferably on 0-glycans [ 10251. It shares sequence similarity with both sialoadhesin and CD22, showing that it belongs to the same immunoglobulin superfamily. The term “sialoadhesin family” was proposed for this group of related sialic-acid-binding proteins [ 10251. Since MAG can activate the MAG-associated tyrosine kinase [ 10671, it is possible that sialylated glycoconjugate ligands play a role in signal transduction during myelination. Furthermore, myelinassociated sialidase adhering to the ganglioside GMl may be involved in the formation and stabilization of the multilamellar structure of the myelin sheath [829]. Other sialic-acid-binding activities have been found in mammalian tissues (Table 22), these being reviewed in ref. [660] and in chapter 14 of the present volume. Recently, a sialic-acid-binding protein was purified from human endometrium, which can bind to human spermatozoa, unless they have been sialidase-treated [ 10161. Glycophorin A of mouse erythrocytes is recognized by a homologous peritoneal macrophage receptor specific for sialylated carbohydrates [ 10681. It should, furthermore, be mentioned that various antibodies have been described, which recognize epitopes containing mono- or oligomeric sialic acids (glycoproteins, gangliosides and bacterial capsular polysaccharides) on a variety of cells (refs. [289,709,1011,1069-10711, and various chapters of this series). Monoclonal antibodies specific for a-2,8-linked Kdn sequences suitable for histochemistry are available now [1072]. The specificity of this interaction is higher than is generally the case for lectins, since the total carbohydrate structures bearing the sialic acids have a greater influence on the binding of antibodies. Interest is, therefore, increasing in the expression of sialyltransferase activity, this being a critical regulatory step in the formation of cell-surface differentiation antigens, which may be necessary for the function of these cells, so far most dramatically shown in lymphocytes [1070,1073,1074]. In the immune-reactivity of sialic acid epitopes, 0-acetylation and N-acetyl hydroxylation have a strong modulatory role [668,669,1075].

10.4. Do sialic acids have “specijic”functions? The question arises as to whether these monosaccharides indeed have specific functions. Certainly, sialic acids are involved in numerous specific biological events, and many of them have been discussed in this and other articles. However, on closer inspection, it seems that sialylation exerts its influence in the diverse biological systems by the factors discussed, i.e. charge, conformation, masking and interaction with receptors. With regard to conformation, sialic acids may contribute to the specific structure of a given glycoconjugate molecule and thus may influence its biological activity. The highest degree of specificity can be observed in sialic-acid-receptor interactions, which require structurally well-defined features on both sites, not only related to the sialyl group but also to the penultimate sugars, thus resembling the formation of an enzyme-substrate complex. The interaction of influenza C virus with 9-0-acetylated sialic acid may be such an

369

example. The specificity in these phenomena can be related mostly to the non-sialylated part of the molecule or the cell, for example to the biological specificity of a glycoprotein hormone, a bacterial toxin, a neuronal reaction, the activity of an ion channel, or the nature of the receptor and the kind of signal it transfers into the cell after ligand binding. In these systems, the sialic acid moiety plays an essential role, since its presence is a prerequisite for the receipt and transfer of biological messages. Furthermore, sialic acid can control and modulate these biological events. The exposed position of this highly negatively charged, hydrophilic molecule on extracellular glycoconjugates makes it suitable for this purpose. The physiological role of sialic acids could therefore be grouped on the basis of the long list of biological phenomena influenced by these carbohydrates. This will not be made here, since another grouping, into three main topics, was done in sections 10.110.3. However, a few interesting and unusual systems, influenced by sialic acids, as well as open questions for future research, will be mentioned. In reproductive biology, sialic acid polymers are receiving more and more attention. Polysialic acid with a(2-8)-linkages was found to be expressed on mouse embryos before and after implantation and the neuronal cell adhesion molecule (NCAM), which bears a polysialic acid chain, seems to be involved in cellular interactions in the early mammalian embryo [ 101,1076]. A similar function is attributed to polysialic acid observed during a short period of the larval stage of Drosophila melanogaster [1077]. It has to be noted that this is the first report on the occurrence of sialic acid in insects. The participation of sialic acids in different linkages and grade of polymerization during differentiation of cells and tissues was excellently demonstrated in a histochemical study [1078]. It was also reported there that polysialic acids typical for developing tissues are expressed in some malignant tumors. Sialic acid is involved in changes of the glycoconjugate patterns during the development of murine molar tooth germs [ 10791. Corresponding to the variable expression of sialic acids during ontogenesis, the expression of sialyltransferases varies with the state of development, e.g. during the maturation of human myeloid cells [ 10801 or of the epithelium of small intestine upon weaning [ 108 11. Sialic acid modification is also involved in differentiation. In several rat and cow organs, the expression of NeuSGc was found to be developmentally regulated [307,674]. Similar observations have been made for the appearance of Neu5,9Ac2 in chicken erythrocytes (ref. [712]; see also section 8.4.2). Many studies have revealed that sialic acids are of significance in the growth and functioning of neuronal cells. This can be mainly attributed to gangliosides, the most important carriers of sialic acids in these tissues [1082,1083]. How these molecules function is the object of intensive research, however, they are mainly involved in transmembrane signalling and the regulation of mediators. Interestingly, the synthetic glycoconjugate sialyl cholesterol also promotes neurite outgrowth in a mouse neuroblastoma cell line [1084]. Since this substance is transported to the nucleus, its localization may play an as yet unknown key role in neuritogenesis. Polysialic acid of NCAM affects cellular interactions not only during the development of the nervous system and skeletal muscle but also during the development of the ventricular conduction system in embryonic hearts of man and various animals[1085]. In adult rat brain, polysialic acid was found to be associated with sodium channels, and was shown to be

370

required for the conductivity of the channel and thus for the function of this nervous tissue [ 10861. Interestingly, polysialic acid is a rather labile substance, which easily degrades spontaneously even under physiological conditions, due to an anomalously high pK value of the internal sialic acid residues [258]. This intramolecular self-cleavage may influence neuronal adhesion, embryogenesis and bacterial virulence. 10.5. Medical signijcance of sialic acids

Since sialic acids are involved in so many biological phenomena, it is easily understandable that disturbances in the metabolism of these sugars, either due to genetic errors or at the post-translational level, may impair the physiological functions of an organism and thus lead to disease. Such pathophysiological influences of sialic acids and sialidases as e.g. virulence factor of microbial diseases have been mentioned throughout this chapter. Four diseases in which metabolic errors of sialic acids have been clearly recognized will be briefly discussed: galactosialidosis, Salla disease, Alzheimer disease and malignant tumors. (Sialidosis and sialuria are described by Michalski in chapter l b of Glycoproteins and Disease, Vol. 30 of this series). Galactosialidosis, first described as sialidosis [ 10871, is a rare genetic disorder and occurs in at least three clinical phenotypes: an early infantile form where abnormalities (sialidase deficiency and an increased amount of sialo-oligosaccharides) can be detected in utero; a late infantile form characterized by skeletal abnormalities, macular cherry-red spot and mild mental retardation; and a juvenile-adult form with variable age of onset of skeletal abnormalities, corneal clouding, macular cherry-red spot, neurological manifestations and mental retardation [ 1088-1 09 11. All three forms are characterized by accumulation of sialo- and galactoglycoconjugates and the excretion of sialyloligosaccharides. The primary defect of this genetically well-studied disease is in the capability of a protective protein, a serine carboxypeptidase which has also deamidase activity (see chapter l a by Jourdian of Glycoproteins and Disease, Vol. 30 of this series), to associate with p-galactosidase and sialidase to protect them from intralysosomal proteolysis. Fibroblasts from these patients show a marked reduction in B-galactosidase and virtually no sialidase activity. The genes encoding P-galactosidase, sialidase and the protective protein, are located on the human chromosomes 3, 10 and 20, respectively [ 1092,10931. With regard to the function of the protease normally stabilizing the ternary enzyme complex, various mutations have been found which affect the stability of the carboxypeptidase/deamidase central sheet [ 10941. The stoichiometry of this complex from human lysosomes has been described [1095]. Lysosomal accumulation of free sialic acid occurs in two phenotypically distinct inherited metabolic disorders, Salla disease and infantile sialic acid storage disease [ 10961. Salla disease is an autosomal recessive lysosomal storage disorder and was first observed in patients of Finnish ancestry, but also occurs outside Finland. The clinical symptoms are a slow progressive psychomotor retardation, impaired speech, ataxia and a prolonged course. Sialic acid accumulates in the lysosomes due to a defective efflux into the cytosol. The genetic defect affects the function of the specific transport protein for sialic acid and other acidic monosaccharides in the lysosomal membrane [ 10971. The Salla disease locus

37 1

was found on the long arm of chromosome 6, and the critical region for this locus is in the range of 190 kb [1098]. Both in Alzheimer’s disease and older Down’s syndrome subjects, a decrease in serum sialyltransferase activity was observed [1099,1100]. This was found to affect only the a-2,3-sialyltransferase,also leading to a decrease of a-2,3-linked sialic acid in serum glycoproteins. It can only be assumed at present that this observation mirrors reduced activity of sialyltransferase and decreased sialoglycoconjugate biosynthesis in neuronal tissue subjected to these degenerative diseases. Serum sialyltransferase may thus be an early biochemical marker of neurodegeneration. Cell surface carbohydrates are recognized to play crucial roles in malignant growth and metastasis, as is reported in detail in volume 30 of this series, “Glycoproteins and Disease”. Here, some recent observations on the special influence of sialic acids in these processes will be presented to complete the main aspects of sialic acid biology. There are several reports in the literature showing that the amount and type of sialylation of membrane components influence the metastatic potential of tumor cells. A higher amount of membrane-bound sialic acid as component of glycoproteins and gangliosides promotes invasion and interferes with intercellular adhesion. Sialic acid is also involved in tumor cell attachment to endothelial cells via sLex- or sLea-binding selectins (section 10.3.2), and in adhesion of tumor cells to substances of the extracellular matrix [ 1044,1101]. These oligosaccharides are therefore considered as markers of some tumors, e.g. of human colon [1102] or thyroid gland carcinomas [1103]. In the latter case, clinicopathological studies indicate that sLea antigens are related to biologically aggressive thyroid tumors. In human colon, sLeXwas found to be expressed only in cultivated cells from premalignant polyposis, but not in intermediate or carcinoma cells [ 11021. The higher degree of sialylation in human colorectal tumor tissue seems to be due mainly to an increase of a-2,6sialyltransferase activity sialylating N-glycan and N-acetyllactosamine sequences [ 110411061. When studying the activity of several glycosyltransferases in human colorectal adenoma cells during progression to cancer, a-2,3-sialyltransferase activity exhibited a peak in the intermediate premalignant stage, while a-2,6-sialyltransferase appeared to be turned off in the final stage, i.e. the adenomacarcinoma cells [ 11021. An increase of serum sialic acids was shown to be relevant for the diagnosis and prognosis of tumor diseases and leukemia[1101,1104,1107]. The role of sialic acids in the formation of metastases was also shown with nude mice and human colorectal cancer cell lines [ 11081. The better differentiated cells had higher levels of sialyltransferase activity and sialic acids which correlated with their enhanced ability to form liver metastases compared to the poorly differentiated, less sialylated cell lines. The role of sialic acid in metastasis was further demonstrated by inhibition of enzymatic sialyltransfer with a CMP-sialic acid derivative (refs. [648,649] and section 8.3), which resulted in reduced liver metastasis. An increase of a-2,6sialyltransferase activity was also observed in cultured Ehrlich ascites tumor cells [ 1 1091 and in human gastric epithelium carcinoma [ 1 1101. The increase of sialylation of an antigenic epitope of the latter tumor cells enhanced their metastatic potential. The expression of ras oncogenes in NIH 3T3 fibroblasts increases the sialic acid density and concomitantly the invasiveness of these cells and reduces cell surface galactosylation [ 11 111. Correspondingly, a-l,3-galactosyltransferase activity was decreased and

372

those of a-2,3- and a-2,6-sialyltransferase activities were increased. Furthermore, the number and/or length of polylactosaminoglycan chains and the branching of N-glycans was also enhanced. This is another example, in which multiple changes in the expression of glycoconjugates occur on malignant transformation. It was shown with rat 3Y1 cell lines that oncogenesis can be accompanied not only by increased sialylation but simultaneously by decreased sialidase activity, thus supporting hypersialylation and metastatic potential of the cells [ 11121. There was an inverse correlation between lysosomal-type sialidase activity and invasion. How the oncogenes mentioned operate in these processes is presently being studied. All these studies and experiments with murine lymphoid tumor cells [ 1 1131 suggest that changes either in the amount, the type or linkage of sialic acid in tumor cell glycoconjugates can affect tumor growth and metastasis.

11. Conclusion In the field of the primary structural analysis and synthesis of sialic acids and sialic-acidcontaining glycans, the methods developed have reached a high level of sophistication. The conformational analysis of sialylated carbohydrate chains of glycoconjugates in their natural environment needs specific attention in order to better understand the physical and biological functions of these glycosylated molecules. Although deep insight into the metabolism of the different sialic acids has been obtained, it is challenging to further unravel the secrets of expression and transfer of these monosaccharides on the enzyme and gene levels, as well as receiving information on their regulation including hormonal control. This is particularly valid for the new group in the sialic acid family, Kdn and its derivatives. The biological roles of sialic acids are least understood, although solving these open questions seems to gain momentum. Thus, sialobiology, sialopathology and also sialopharmacology will remain most fascinating topics of future research on sialic acids.

References [ I ] Blix, G. (1936) Hoppe-Seyler’s Z. Physiol. Chem. 240, 43-54. [2] Klenk, E. (1941) Hoppe-Seyler’s Z. Physiol. Chem. 268, 50-58. [3] Tuppy, H. and Gottschalk, A. (1972) In: A. Gottschalk (Ed.), Glycoproteins - Their Composition, Structure and Function, Part A. Elsevier, Amsterdam, pp. 403449. [4] Schauer, R. (Ed.) (1982) Sialic acids - Chemistry, Metabolism and Function. Springer, Wien. [5] Schauer, R. (1982) Adv. Carbohydr. Chem. Biochem. 40, 131-234. [6] Schauer, R. (1987) Methods Enzyrnol. 138, 132-161. [7] Schauer, R. (1987) Methods Enzymol. 138, 61 1-626. [8] Varki, A. (1992) Glycobiology 2, 2 5 4 0 . [9] Troy, F.A. (1992) Glycobiology 2, 5-23. [lo] Manzi, A.E. and Varki, A. (1993) In: M. Fukuda and A. Kobata (Eds.), Glycobiology A Practical Approach. IRL Press, Oxford, pp. 27-77. [I I ] Reuter, G. and Schauer, R. (1994) Methods Enzymol. 230, 168-199. [I21 Reuter, G. and Schauer, R. (1994) Methods Carbohydr. Chem. 10, 29-39. [13] Corfield, A.P. and Schauer, R. (1982) In: R. Schauer (Ed.), Sialic acids - Chemistry, Metabolism and Function. Springer, Wien, pp. 5-50. -

373 [I41 Sjoberg, E.R., Chammas, R., Ozawa, H., Kawashima, I., Khoo, K.-H., Morris, H.R., Dell, A,, Tai, T. and Varki, A. (1995) J. Biol. Chem. 270, 2921-2930. [I51 Kamerling, J.P. and Vliegenthart, J.F.G. (1982) In: R. Schauer (Ed.), Sialic acids - Chemistry, Metabolism and Function. Springer, Wien, pp. 95-1 25. [I61 Vliegenthart, J.F.G., Dorland, L., van Halbeek, H. and Haverkamp, J. (1982) In: R. Schauer (Ed.), Sialic acids - Chemistry, Metabolism and Function. Springer, Wien, pp. 127-172. [I71 Kamerling, J.P., Vliegenthart, J.F.G., Schauer, R., Strecker, G. and Montreuil, J. (1975) Eur. J. Biochem. 56, 253-258. [I81 Haverkamp, J., van Halbeek, H., Dorland, L., Vliegenthart, J.F.G., Pfeil, R. and Schauer, R. (1982) Eur. J. Biochem. 122, 305-3 1 1. [I91 Kamerling, J.P., Dorland, L., van Halbeek, H., Vliegenthart, J.F.G., Messer, M. and Schauer, R. (1982) Carbohydr. Res. 100, 331-340. [20] Damm, J.B.L., Voshol, H., Hird, K., Kamerling, J.P. and Vliegenthart, J.F.G. (1989) Eur. J. Biochem. 180, 101-110. [21] Hanaoka, K., Pritchett, T.J., Takasaki, S., Kochibe, N., Sabesan, S., Paulson, J.C. and Kobata, A. (1989) J. Biol. Chem. 264, 9842-9849. [22] Reuter, G., Pfeil, R., Stoll, S., Schauer, R., Kamerling, J.P., Versluis, C. and Vliegenthart, J.F.G. (1983) Eur. J. Biochem. 134, 139-143. [23] Kamerling, J.P., Schauer, R., Shukla, A.K., van Halbeek, H. and Vliegenthart, J.F.G. (1987) Eur. J. Biochem. 162, 601L607. [24] Jennings, H.J. (1983) Adv. Carbohydr. Chem. Biochem. 41, 155-208. [25] Bergwerff, A.A., van Oostrum, J., Asselbergs, F.A.M., Burgi, R., Hokke, C.H., Kamerling, J.P. and Vliegenthart, J.F.G. (1993) Eur. J. Biochem. 212, 639-656. [26] Kogan, G., Jann, B. and Jann, K. (1992) FEMS Microbiol. Lett. 91, 135-140. [27] Gamian, A., Romanowska, E., Ulrich, J. and Defaye, J. (1992) Carbohydr. Res. 236, 195-208. [28] Schauer, R., Haverkamp, J., Wember, M., Vliegenthart, J.F.G. and Kamerling, J.P. (1976) Eur. J. Biochem. 62, 237-242. [29] Kleineidam, R.G., Hofmann, O., Reuter, G. and Schauer, R. (1993) Glycoconjugate J. 10, 116-119. [30] Reuter, G., Pfeil, R., Kamerling, J.P., Vliegenthart, J.F.G. and Schauer, R. (1980) Biochim. Biophys. Acta 630, 306-3 10. [31] Bergwerff, A.A., Hulleman, S.H.D., Kamerling, J.P., Vliegenthart, J.F.G., Shaw, L., Reuter, G. and Schauer, R. (1992) Biochimie 74, 25-38. [32] Slomiany, A,, Kojima, K., Banas-Gruszka, Z. and Slomiany, B.L. (1981) Biochem. Biophys. Res. Commun. 100, 778-784. [33] Corfield, A.P. and Schauer, R. (1982) In: R. Schauer (Ed.), Sialic acids - Chemistry, Metabolism and Function. Springer, Wien, pp. 195-261, [34] Beau, J.-M., Schauer, R., Haverkamp, J., Kamerling, J.P., Dorland, L. and Vliegenthart, J.F.G. (1984) Eur. J. Biochem. 140, 203-208. [35] Shukla, A.K., Schroder, C., Nohle, U. and Schauer, R. (1987) Carbohydr. Res. 168, 199-209. [36] Schauer, R., Schroder, C. and Shukla, A.K. (1984) Adv. Exp. Med. Biol. 174, 75-86. [37] Suzuki, M., Suzuki, A,, Yamakawa, T. and Matsunaga, E. (1985) J. Biochem. (Tokyo) 97, 509-515. [38] Li, Y.-T., Nakagawa, H., Ross, S.A., Hansson, G. and Li, S.-C. (1990) J. Biol. Chem. 265, 2162921633. [39] Li, Y.-T., Yuziuk, J.A., Li, S.-C., Nematalla, A,, Hasegawa, A,, Kimura, M. and Nakagawa, H. (1994) Arch. Biochem. Biophys. 3 10, 243-246. [40] Damm, J.B.L., Bergwerff, A.A., Hird, K., Kamerling, J.P. and Vliegenhart, J.F.G. (1989) Recl. Trav. Chim. Pays-Bas 108, 35 1-359. [41] Manzi, A.E., Dell, A,, Parastoo, A. and Varki, A. (1990) J. Biol. Chem. 265, 8094-8107. [42] Kochetkov, N.K., Smirnova, G.P. and Chekareva, N.V. (1976) Biochim. Biophys. Acta 424,274-283. (431 Kluge, A,, Reuter, G., Lee, H., Ruch-Heeger, B. and Schauer, R. (1992) Eur. J. Cell Biol. 59, 12-20. [44] Higuchi, R., Inukai, K., Jhou, J.X., Honda, M., Komori, T., Tsuji, S. and Nagai, Y. (1993) Liebigs Ann. Chem. 359-366.

374 [45] Nohle, U., Shukla, A.K., Schroder, C., Reuter, G., Schauer, R., Kamerling, J.P. and Vliegenthart, J.F.G. (1985) Eur. J. Biochem. 152, 459463. [46] Nadano, D., Iwasaki, M., Endo, S., Kitajima, K., Inoue, S. and Inoue, Y. (1986) J. Biol. Chem. 261, 11550-1 1557. [47] Kanamori, A,, Inoue, S., Iwasaki, M., Kitajima, K., Kawai, G., Yokoyama, S. and Inoue, Y. (1990) J. Biol. Chem. 265, 21811-21819. [48] Strecker, G., Wieruszeski, J.-M., Michalski, J.-C., Alonso, C., Boilly, B. and Montreuil, J. (1992) FEBS Lett. 298, 3 9 4 3 . [49] Knirel, Y.A., Kocharova, N.A., Shashkov, A.S., Kochetkov, N.K., Mamontova, VA. and Solov’eva, T.F. (1989) Carbohydr. Res. 188, 145-155. [50] Ziak, M., Yu, B.X., Zuo, X.L., Zuber, C., Kanamon, A,, Kitajima, K., lnoue, S., Inoue, Y. and Roth, J. (1996) Proc. Natl. Acad. Sci. USA 93, 2759-2763. [51] Iwasaki, M., Inoue, S. and Troy, F.A. (1990) J. Biol. Chem. 265, 2596-2602. [52] Lutz, F,! Lochinger, W. and Taigel, G. (1968) Chem. Ber. 101, 1089-1094. [53] Haverkamp, J., Schauer, R., Wember, M., Kamerling, J.P. and Vliegenthart, J.F.G. (1975) Hoppe-Seyler’s Z. Physiol. Chem. 356, 1575-1583. [54] Yu, R.K. and Ledeen, R. (1969) J. Biol. Chem. 244, 1306-1313. [55] Haverkamp, J., Spoormaker, T., Dorland, L., Vliegenthart, J.F.G. and Schauer, R. (1979) J. Am. Chem. SOC.101, 48514853. [56] Flippen, J.L. (1973) Acta Cryst. B 29, 1881-1886. [57] Reinhard, B. and Faillard, H. (1994) Liebigs Ann. Chem. 193-203. [58] Haverkamp, J., Schauer, R., Wember, M., Farriaux, J.-P., Kamerling, J.P., Versluis, C. and Vliegenthart, J.F.G. (1976) Hoppe-Seyler’s 2. Physiol. Chem. 357, 1699-1705. [59] Miyoshi, I., Higashi, H., Hirabayashi, Y., Kato, S. and Naiki, M. (1986) Mol. lmmunol. 23, 631-638. [60] Burmeister, W.P., Henrissat, B., Bosso, C., Cusack, S. and Ruigrok, R.W.H. (1993) Structure 1, 19-26. [61] Chou, M.-Y., Li, S.-C., Kiso, M., Hasegawa, A. and Li, Y.-T. (1994) J. Biol. Chem. 269, 18821-18826. [62] Lifely, M.R. and Cottee, F.H. (1982) Carbohydr. Res. 107, 187-197. [63] Pozsgay, V., Jennings, H. and Kasper, D.L. (1987) Eur. J. Biochem. 162, 445450. [64] Saito, K., Sugai, K., Fujikura, K., Yamada, N., Goto, M., Ban, C., Hayasaka, E., Sugiyama, N. and Tomita, K. (1 989) Carbohydr. Res. 185, 307-3 14. [65] Montreuil, J., Biserte, G., Strecker, G., Spik, G., Fontaine, G. and Farriaux, J.-F! (1968) Clin. Chim. Acta 21, 61-69. [66] Renlund, M., Chester, M.A., Lundblad, A., Ada, P., Raivio, K.O., Autio, S. and Koskela, S.-L. (1979) Eur. J. Biochem. 101, 245-250. [67] Stevenson, R.E., Lubinsky, M., Taylor, H.A., Wenger, D.A., Schroer, R.J. and Olmstead, P.M. (1983) Pediatrics 72, 441449. [68] Palo, J., Rauvala, H., Finne, J., Haltia, M. and Palmgren, K. (1985) Clin. Chim. Acta 145, 237-242. [69] Krauss, J.H., Reuter, G., Schauer, R., Weckesser, J. and Mayer, H. (1988) Arch. Microbiol. 150, 584589. [70] Green, E.D., Adelt, G., Baenziger, J.U., Wilson, S. and van Halbeek, H. (1988) J. Biol. Chem. 263, 18253-1 8268. [71] Bendiak, B., Harris-Brandts, M., Michnick, S.W., Carver, J.P. and Cumming, D.A. (1989) Biochemistry 28, 64916499. [72] Bernard, N., Engler, R., Strecker, G., Montreuil, J., van Halbeek, H. and Vliegenthart, J.F.G. (1984) Glycoconjugate J. 1, 123-140. [73] Cumming, D.A., Hellerqvist, C.G., Harris-Brandts, M., Michnick, S.W., Carver, J.F! and Bendiak, B. (1989) Biochemistry 28, 6500-6512. [74] Hanisch, F.G., Uhlenbruck, G., Peter-Katalinic, J. and Egge, H. (1988) Carbohydr. Res. 178, 2 9 4 7 . [75] Takasaki, S., Yamashita, K., Suzuki, K., Iwanaga, S. and Kobata, A. (1979) J. Biol. Chem. 254, 85488553. [76] Vliegenthart, J.F.G., Dorland, L. and van Halbeek, H. (1983) Adv. Carbohydr. Chem. Biochem. 41, 209-374. [77] Kamerling, J.P. and Vliegenthart, J.F.G. (1992) Biol. Magn. Reson. 10, 1-194.

375 [78] [79] [SO] [81]

Rivera-Sagedro, A., Bergwefl, A.A., Kamerling, J.P. and Vliegenthart, J.F.G. (1997) in preparation. Inoue, S., Iwasaki, M., Ishii, K., Kitajima, K. and Inoue, Y. (1989) J. Biol. Chem. 264, 18520-18526. Iwasaki, M., Seko, A., Kitajima, K., Inoue, Y.and Inoue, S. (1992) J. Biol. Chem. 267, 24287-24296. HBrd, K., van Zadelhoff, G., Moonen, P., Kamerling, J.P. and Vliegenthart, J.F.G. (1992) Eur. J. Biochem. 209 (1 992) 895-9 15. [82] Pfeiffer, G., Stirm, S., Geyer, R., Strube, K.-H., BergwertT, A.A., Kamerling, J.P. and Vliegenthart, J.F.G. (1992) Glycobiology 2, 41 1 4 1 8 . [83] Van Kuik, J.A., HBrd, K. and Vliegenthart, J.F.G. (1992) Carbohydr. Res. 235, 53-68. [84] Hokke, C.H., Bergwerff, A.A., van Dedem, G.W.K., Kamerling, J.P. and Vliegenthart, J.F.G. (1995) Eur. J. Biochem. 228, 981-1008. [85] De Waard, P., Koorevaar, A., Kamerling, J.P. and Vliegenthart, J.F.G. (1991) J. Biol. Chem. 266, 42374243. [86] Slomiany, B.L., Murty, VL.N. and Slomiany, A. (1980) J. Biol. Chem. 255, 9719-9723. [87] D a m , J.B.L., HBrd, K., Kamerling, J.P., van Dedem, G.W.K. and Vliegenthart, J.F.G. (1990) Eur. J. Biochem. 189, 175-183. [88] Coddeville, B., Stratil, A,, Wieruszeski, .I.-M., Strecker, G., Montreuil, J. and Spik, G. (1989) Eur. J. Biochem. 186, 583-590. [89] Taguchi, T., Seko, A,, Kitajima, K., Inoue, S., Iwamatsu, T., Khoo, K.-H., Morris, H.R., Dell, A. and Inoue, Y.(1993) J. Biol. Chem. 268, 2353-2362. [90] Taguchi, T., Seko, A,, Kitajima, K., Muto, Y., Inoue, S., Khoo, K.-H., Morris, H.R., Dell, A. and Inoue, Y. (1994) J. Biol. Chem. 269, 8762-8771. [91] Fukuda, M., Lauffenburger, M., Sasaki, H., Rogers, M.E. and Dell, A. (1987) J. Biol. Chem. 262, 11952-1 1957. [92] Pfeiffer, G., Dabrowski, U., Dabrowski, J., Stirm, S., Strube, K.-H. and Geyer, R. (1992) Eur. J. Biochem. 205, 961-978. [93] Weisshaar, G., Hiyama, J., Renwick, A.G.C. and Nimtz, M. (1991) Eur. J. Biochem. 195, 257-268. [94] Bergwerff, A.A., van Oostrum, J., Kamerling, J.P. and Vliegenthart, J.F.G. (1995) Eur. J. Biochem. 228, 1009-1 019. [95] Iwasaki, M. and Inoue, S. (1985) Glycoconjugate J. 2, 209-228. [96] Slomiany, A. and Slomiany, B.L. (1978) J. Biol. Chem. 253, 7301-7306. [97] Van der Meer, A,, Kamerling, J.P., Vliegenthart, J.F.G., Schmid, K. and Schauer, R. (1983) Biochim. Biophys. Acta 757, 371-376. [98] Fukuda, M.N., Dell, A,, Oates, J.E. and Fukuda, M. (1985) J. Biol. Chem. 260, 6623-6631. [99] Finne, J. (1985) Trends Biochem. Sci. 10, 129-132. [loo] Sato, C., Kitajima, K., Tazawa, I., Inoue, Y., Inoue, S. and Troy, F.A. (1993) J. Biol. Chem. 268, 23675-23684. [I011 Kitazume, S., Kitajima, K., Inoue, S., Troy, F.A., Cho, J.-W., Lennarz, W.J. and Inoue, Y. (1994) J. Biol. Chem. 269, 22712-22718. [I021 Tezuka, T., Taguchi, T., Kanamari, A,, Muto, Y., Kitajima, K., Inoue, Y. and Inoue, S. (1994) Biochemistry 33, 64954502. I1031 Strecker, G., Wieruszeski, J.-M., Michalski, J.-C. and Montreuil, J. (1992) Biochem. J. 287,905-909. [lo41 Maes, E., Wieruszeski, J.-M., Plancke, Y. and Strecker, G. (1995) FEBS Lett. 358, 205-210. [lo51 Strecker, G., Wieruszeski, J.-M., Michalski, J.-C., Alonso, C., Leroy, Y., Boilly, B. and Montreuil, J. (1992) Eur. J. Biochem. 207, 995-1002. [I061 Kimura, M., Hama, Y., Sumi, T., Asakawa, M., Rao, B.N.N., Home, A.P., Li, S.-C., Li, Y.-T. and Nakagawa, H. (1994) J. Biol. Chem. 269, 32138-32143. [106a] Nakagawa, H., Sumi, T., Kimura, M., Hama, Y., Li, S.-C. and Li, Y.-T. (1995) Glycoconjugate J. 12, 516. [I071 Terada, T., Kitazume, S., Kitajima, K., Inoue, S., Ito, F., Troy, F.A. and Inoue, Y. (1993) J. Biol. Chem. 268, 2640-2648. [lo81 Yu, R.K. and Saito, M. (1989) In: R.U. Margolis and R.K. Margolis (Eds.), Neurobiology of Glycoconjugates. Plenum Publishing Corporation, New York, pp. 1 4 2 . [lo91 Kubo, H., Irie, A,, Inagaki, F. and Hoshi, M. (1990) J. Biochem. (Tokyo) 108, 185-192.

376

[I 101 Ren, S.-L., Scarsdale, J.N., Ariga, T., Zhang, Y.-J., Klein, R.A., Hartmann, R., Kushi, Y., Egge, H. and Yu, R.K. (1992) J. Biol. Chem. 267, 12632-12638. [ l l I] Waki, H., Murata, A,, Kon, K., Maruyama, K., Kimura, S., Ogura, H. and Ando, S. (1993) J. Biochem. (Tokyo) 113, 502-507. [112] Kazuya, I.-P., Hidari, J., Irie, F., Suzuki, M., Kon, K., Ando, S. and Hirabayashi, Y. (1993) Biochem. J. 296, 259-263. [I 131 Waki, H., Masuzawa, A,, Kon, K. and Ando, S. (1993) J. Biochem. (Tokyo) 114, 459462. [114] Nakao, T., Kon, K., Ando, S. and Hirabayashi, Y. (1991) Biochim. Biophys. Acta 1086, 305-309. [I 151 Muralikrishna, G., Reuter, G., Peter-Katalinic, J., Egge, H., Hanisch, F.-G., Siebert, H.-C. and Schauer, R. (1992) Carbohydr. Res. 236, 321-326. [116] Miller-Podraza, H., Miinsson, J.-E. and Svennerholm, L. (1991) FEBS Lett. 288, 212-214. [117] Nakamura, K., Suzuki, M., Taya, C., Inagaki, F., Yamakawa, T. and Suzuki, A. (1991) J. Biochem. (Tokyo) 110, 832-841. [118] Nakamura, K., Inagaki, F. and Tamai, Y. (1988) J. Biol. Chem. 263, 9896-9900. [119] Hirabayashi, Y., Nakao, T., hie, F., Whittaker, W , Kon, K. and Ando, S. (1992) J. Biol. Chem. 267, 12973-1 2978. [I201 Furukawa, K., Yamaguchi, H., Oettgen, H.F., Old, L.J. and Lloyd, K.O. (1988) J. Biol. Chem. 263, 18507-1 85 12. [I211 Fredman, P., Miinsson, J.-E., Wikstrand, C.J., Vrionis, ED., Rynmark, B.-M., Bigner, D.D. and Svennerholm, L. (1989) J. Biol. Chem. 264, 12122-12125. [122] Song, Y., Kitajima, K., Inoue, S. and Inoue, Y. (1991) J. Biol. Chem. 266, 21929-21935. [I231 Song, Y., Kitajima, K., Inoue, S., Muto, Y., Kasama, T., Handa, S. and Inoue, Y. (1993) Biochemistry 32, 9221-9229. [I241 Song, Y., Kitajima, K., Inoue, S., Khoo, K.-H., Morns, H.R., Dell, A. and Inoue, Y. (1995) Glycobiology 5, 207-218. [I251 Kobata, A. (1977) In: M.I. Horowitz and W. Pigman (Eds.), The Glycoconjugates, Vol. 1. Academic Press, New York, pp. 423440. [126] Gronberg, G., Lipniunas, P., Lundgren, T., Erlansson, K., Lindh, F. and Nilsson, B. (1989) Carbohydr. Res. 191, 261-278. [127] Platzer, N., Davoust, D., Lhermitte, M., Bauvy, C., Meyer, D.M. and Derappe, C. (1989) Carbohydr. Res. 191, 191-207. [I281 Strecker, G., Wieruszeski, J.-M., Michalski, J.-C. and Montreuil, J. (1989) Glycoconjugate J. 6, 67-83. [I291 Kitagawa, H., Takaoka, M., Nakada, H., Fukui, S., Funakoshi, I., Kawasaki, T., Tate, S., Inagaki, F. and Yamashina, I. (1991) J. Biochem. (Tokyo) 110, 598404. [I301 Gronberg, G., Lipniunas, P., Lundgren, T., Lindh, F. and Nilsson, B. (1990) Arch. Biochem. Biophys. 278, 297-3 1 1. [131] Wieruszeski, J.-M., Chekkor, A,, Bouquelet, S., Montreuil, J., Strecker, G., Peter-Katalinic, J. and Egge, H. (1985) Carbohydr. Res. 137, 127-138. [132] Kitagawa, H., Nakada, H., Numata, Y., Korosaka, A., Fukui, S., Funakoshi, l., Kawasaki, T., Shimada, K., Inagaki, F, and Yamashina, I. (1988) J. Biochem. (Tokyo) 104, 591-594. [133] Wang, W.T., Erlansson, K., Lindh, F., Lundgren, T. and Zopf, D. (1990) Anal. Biochem. 190, 182-1 87. [134] Huang, R.T.C. (1971) Hoppe-Seyler’s Z. Physiol. Chem. 352, 1645-1652. [135] Prieto, PA. and Smith, D.F. (1984) Arch. Biochem. Biophys. 229, 650456. [136] Gronberg, G., Lipniunas, P., Lundgren, T., Lindh, E and Nilsson, B. (1992) Arch. Biochem. Biophys. 296, 597-61 0. [I371 Yamashita, K., Tachibana, Y. and Kobata, A. (1976) Arch. Biochem. Biophys. 174, 582-591. [I381 Fievre, S., Wieruszeski, J.-M., Michalski, J.-C., Lemoine, J., Montreuil, J. and Strecker, G. (1991) Biochem. Biophys. Res. Commun. 177, 720-725. [ 1391 Kitagawa, H., Nakada, H., Kurosaka, A,, Hiraiwa, N., Numata, Y., Fukui, S., Funakoshi, I., Kawasaki, T., Yamashina, I., Shimada, I. and Inagaki, F. (1989) Biochemistry 28, 8891-8897. [140] Kitagawa, H., Nakada, H., Fukui, S., Funakoshi, I., Kawasaki, T., Yamashina, I., Tate, S. and Inagaki, F, (1993) J. Biochem. (Tokyo) 114, 504-508.

377 [I411 Kitagawa, H., Nakada, H., Fukui, S., Funakoshi, I., Kawasaki, T., Yamashina, I., Tate, S. and Inagaki, F. (1991) Biochemistry 30, 2869-2876. [I421 Wang, W.T., Lundgren, T., Lindh, F., Nilsson, B., Gronberg, G., Brown, J.P., Mentzer-Dibert, H. and Zopf, D. (1992) Arch. Biochem. Biophys. 292, 433441. [I431 Yamashita, K., Tachibana, Y. and Kobata, A. (1977) Arch. Biochem. Biophys. 182, 546-555. [I441 Tarrago, M.T., Tucker, K.H., van Halbeek, H. and Smith, D.F. (1988) Arch. Biochem. Biophys. 267, 353-362. [I451 Smith, D.F., Prieto, PA., McCrumb, D.K. and Wang, W.-C. (1987) J. Biol. Chem. 262, 12040-12047. [I461 Kitagawa, H., Nakada, H., Numata, Y., Kurosaka, A., Fukui, S., Funakoshi, I., Kawasaki, T., Shimada, I., Inagaki, F. and Yamashina, I. (1990) J. Biol. Chem. 265, 48594862. [I471 Sabhanval, H., Nilsson, B., Gronberg, G., Chester, M.A., Dakour, J., Sjoblad, S. and Lundblad, A. (1988) Arch. Biochem. Biophys. 265, 390406. [I481 Parkkinen, J. and Finne, J. (1987) Methods Enzymol. 138, 289-300. [I491 Weisshaar, G., Baumann, W., Friebolin, H., Brunner, H., Mann, H., Sieberth, H.-G. and Opferkuch, H.-J. (1987) Biol. Chem. Hoppe-Seyler 368, 647-658. [I501 Weisshaar, G., Baumann, W., Friebolin, H., Brunner, H., Mann, H., Sieberth, H.-G. and Opferkuch, H.-J. (1987) Biol. Chem. Hoppe-Seyler 368, 1545-1558. [I511 Berman, E. (1983) Carbohydr. Res. 118, 9-20. [I521 Breg, J., Kroon-Batenburg, L.M.J., Strecker, G., Montreuil, J. and Vliegenthart, J.F.G. (1989) Eur. J. Biochem. 178, 727-739. [I531 Van Pelt, J., Dorland, L., Duran, M., Hokke, C.H., Kamerling, J.P. and Vliegenthart, J.F.G. (1990) J. Biol. Chem. 265, 19685-19689. [I541 Ryan, L.C., Carubelli, R., Caputto, R. and Trucco, R.E. (1965) Biochim. Biophys. Acta 101,252-258. [I551 Lemonnier, M. and Bourrillon, R. (1976) Carbohydr. Res. 51, 99-106. [I561 Rodriguez-Aparicio, L.B., Reglero, A,, Ortiz, A.I. and Luengo, J.M. (1 988) Appl. Microbiol. Biotechnol. 27, 474483. [I571 Adlam, C., Knight, J., Mugridge, A,, Williams, J. and Lindon, J. (1987) FEMS Microbiol. Lett. 42, 23-25. [I581 Devi, S.J.N., Schneerson, R., Egan, W., Vann, W.F., Robbins, J.B. and Shiloach, J. (1991) Infect. Immun. 59, 732-736. [159] Higa, H.H. and Varki, A. (1988) J. Biol. Chem. 263, 8872-8878. [ 1601 Gonzalez-Clemente, C., Luengo, J.M., Rodriguez-Aparicio, L.B., Ferrero, M.A. and Reglero, A. (1990) Biol. Chem. Hoppe-Seyler 371, 1101-1 106. [161] Dutton, G.G.S., Parolis, H. and Parolis, L.A.S. (1987) Carbohydr. Res. 170, 193-206. [I621 Wessels, M.R., Pozsgay, V, Kasper, D.L. and Jennings, H.J. (1987) J. Biol. Chem. 262, 8262-8267. [I631 DiFabio, J.L., Michon, E, Brisson, J.-R., Jennings, H.J., Wessels, M.R., Benedi, V-J. and Kasper, D.L. (1989) Can. J. Chem. 67, 877-882. [I641 Wessels, M.R., DiFabio, J.L., Benedi, V-J., Kasper, D.L., Michon, F., Brisson, J.-R., Jelinkovi, J. and Jennings, H.J. (1991) J. Biol. Chem. 266, 6714-6719. [165] Kogan, G., Uhnn, D., Brisson, J.-R., Paoletti, L.C., Kasper, D.L., von Hunolstein, C., Orefici, G. and Jennings, H.J. (1994) J. Carbohydr. Chem. 13, 1071-1078. [ 165a] Kogan, G., Brisson, J.-R., Kasper, D.L., von Hunolstein, C., Orefici, G. and Jennings, H.J. (1 995) Carbohydr. Res. 277, 1-9. [166] Gamian, A,, Kenne, L., Mieszala, M., Ulrich, J. and Defaye, J. (1994) Eur. J. Biochem. 225, 12111220. [167] Torgov, VI., Shashkov, AS., Jann, B. and Jann, K. (1995) Carbohydr. Res. 272, 73-90. [I681 Vinogradov, E.V, Knirel, Y.A., Shashkov, AS., Paramonov, N.A., Kochetkov, N.K., Stanislavsky, E.S. and Kholodkova, E.V (1994) Carbohydr. Res. 259, 59-65. 11691 Vinogradov, E.V, Holst, O., Thomas-Oates, J.E., Broady, K.W. and Brade, H. (1992) Eur. J. Biochem. 210,491498. [170] Gamian, A., Romanowska, E., Dabrowski, U. and Dabrowski, J. (1991) Biochemistry 30, 5032-5038. [171] Pavliak, V, Brisson, J.-R., Michon, F., Uhrin, D. and Jennings, H.J. (1993) J. Biol. Chem. 268, 1414614152.

378 [I721 Yamasaki, R., Griffiss, J.M., Quinn, K.P. and Mandrell, R.E. (1993) J. Bacteriol. 175 (1993) 45654568. [173] Kim, J.J., Phillips, N.J., Gibson, B.W., Griffiss, J.M. and Yamasaki, R. (1994) Infect. Immun. 62, 15661575. [174] Melaugh, W., Phillips, N.J., Campagnari, A.A., Tullius, M.V and Gibson, B.W. (1994) Biochemistry 33, 13070-13078. [I751 Aspinall, G.O., McDonald, A.G., Raju, T.S., Pang, H., Moran, A.P. and Penner, J.L. (1993) Eur. J. Biochem. 213, 1017-1027. [176] Aspinall, G.O., McDonald, A.G., Raju, T.S., Pang, H., Kurjanczyk, L.A., Penner, J.L. and Moran, A.P. (1993) Eur. J. Biochem. 213, 1029-1037. [I771 Krauss, J.H., Himmelspach, K., Reuter, G., Schauer, R. and Mayer, H. (1992) Eur. J. Biochem. 204, 2 17-223. [I781 Defives, C., Bouslamti, R., Derieux, J.C., Kol, 0. and Fournet, B. (1989) FEMS Microbiol. Lett. 57, 203-208. [I791 Gamian, A. and Kenne, L. (1993) J. Bacteriol. 175, 1508-1513. [180] Basu, S., Schlecht, S., Wagner, M. and Mayer, H.L. (1994) FEMS Immunol. Med. Microbiol. 9, 189198. [I811 Utley, S.R., Bhat, U.R., Byrd, W. and Kadis, S. (1992) FEMS Microbiol. Lett. 92, 211-216. [I821 Gibson, B.W., Melaugh, W., Phillips, N.J., Apicella, M.A., Campagnan, A.A. and McLeod Griffiss, J. (1993) J. Bacteriol. 175, 2702-2712. [183] Mandrell, R.E., McLaughlin, R., Kwaik, Y.A., Lesse, A,, Yamasaki, R., Gibson, B., Spinola, S.M. and Apicella, M.A. (1992) Infect. Immun. 60, 1322-1328. [184] Smith, H., Cole, J.A. and Parsons, N.J. (1992) FEMS Microbiol. Lett. 100, 287-292. [185] Wetzler, L.M., Barry, K., Blake, M.S. and Gotschlich, E.C. (1992) Infect. Immun. 60, 3 9 4 3 . [186] Stahl, N., Baldwin, M.A., Hecker, R., Pan, K.-M., Burlingame, A.L. and Prusiner, S.B. (1992) Biochemistry 31, 5043-5053. [187] Nohle, U. and Schauer, R. (1984) Hoppe-Seyler’s Z. Physiol. Chem. 365, 1457-1467. [188] Shaw, L. and Schauer, R. (1988) Biol. Chem. Hoppe-Seyler 369, 477486. [189] Kamerling, J.P. (1994) Pure Appl. Chem. 66, 2235-2238. [I901 Schachter, H. and Brockhausen, 1. (1992) In: H.J. Allen and E.C. Kisailus (Eds.), Glycoconjugates Composition, Structure, and Function. Marcel Dekker, New York, pp. 263-332. [191] Dorland, L., Haverkamp, J., Vliegenthart, J.F.G., Strecker, G., Michalski, J.-C., Foumet, B., Spik, G. and Montreuil, J. (1978) Eur. J. Biochem. 87, 323-329. [I921 Lecat, D., Lemonnier, M., Derappe, C., Lhermitte, M., van Halbeek, H., Dorland, L. and Vliegenthart, J.F.G. (1984) Eur. J. Biochem. 140, 415420. [193] Kuriyama, M., Ariga, T., Ando, S., Suzuki, M., Yamada, T., Miyatake, T. and Igata, A. (1985) J. Biochem. (Tokyo) 98, 1049-1054. [194] Van Pelt, J., Kamerling, J.P., Vliegenthart, J.F.G., Verheijen, F.W. and Galjaard, H. (1988) Biochim. Biophys. Acta 965, 3 6 4 5 . [I951 Van Pelt, J., Kamerling, J.P., Vliegenthart, J.F.G., Hoogeveen, A.T. and Galjaard, H. (1988) Clin. Chim. Acta 174, 325-336. [196] Van Pelt, J., van Kuik, J.A., Kamerling, J.P., Vliegenthart, J.F.G., van Diggelen, O.P. and Galjaard, H. (1988) Eur. J. Biochem. 177, 327-338. [197] Van Pelt, J., van Bilsen, D.G.J.L., Kamerling, J.P. and Vliegenthart, J.F.G. (1988) Eur. J. Biochem. 174, 183-187. [198] Van Pelt, J., Hhrd, K., Kamerling, J.P., Vliegenthart, J.F.G., Reuser, A and Galjaard, H. (1989) Biol. Chem. Hoppe-Seyler 370, 191-203. [199] Van Pelt, J., Kamerling, J.P., Bakker, H.D. and Vliegenthart, J.F.G. (1991) J. lnher. Metab. Dis. 14, 730-740. [200] Lifely, M.R., Gilbert, A.S. and Moreno, C. (1981) Carbohydr. Res. 94, 193-203. [201] Lifely, M.R., Gilbert, A S . and Moreno, C. (1984) Carbohydr. Res. 134, 229-243. [202] Lifely, M.R., Lindon, J.C., Williams, J.M. and Moreno, C. (1985) Carbohydr. Res. 143, 191-205. [203] Bassi, R., Riboni, L., Sonnino, S. and Tettamanti, G. (1989) Carbohydr. Res. 141-146.

379 [204] Fronza, G., Kirschner, G., Acquotti, D. and Sonnino, S. (1989) Carbohydr. Res. 195, 51-58. [205] Ando, S., Yu, R.K., Scarsdale, J.N., Kusunoki, S. and Prestegard, J.H. (1989) J. Biol. Chem. 264, 3478-3483. [206] Terabayashi, T.,Ogawa, T. and Kawanishi, Y. (1990) J. Biochem. (Tokyo) 107, 868-871. [207] Zilliken, F, and O’Brien, P.J. (1960) Biochem. Prep. 7, 1-5. [208] Uchida, Y., Tsukada, Y. and Sugimon, T. (1973) Agric. Biol. Chem. 37, 2105-2110. [209] Juneja, L.R., Koketsu, M., Nishimoto, K., Kim, M., Yamamoto, T. and Itoh, T. (1991) Carbohydr. Res. 214, 179-186. [210] Koketsu, M., Juneja, L.R., Kawanami, H., Kim, M. and Yamamoto, T. (1992) Glycoconjugate J. 9 (1 992) 70-74. [211] Roy, R. and Pon, R.A. (1990) Glycoconjugate J. 7, 3-12. [212] Roy, R. and Boratynski, J. (1990) Biotechnol. Lett. 12, 347-352. [213] Schauer, R. and Corfield, A.P. (1982) In: R. Schauer (Ed.), Sialic acids - Chemistry, Metabolism and Function. Springer, Wien, pp. 51-57. [214] Knirel, Y.A. and Kochetkov, N.K. (1987) FEMS Microbiol. Rev. 46, 381-385. [214a] Edebrink, P., Jansson, P.-E., Bsgwald, J. and Hoffmann, J. (1996) Carbohydr. Res. 287, 225-246. [215] Kenne, L., Lindberg, B., Schweda, E., Gustafsson, B. and Holme, T. (1988) Carbohydr. Res. 180, 285-294. [216] Nazarenko, E.L., Shashkov, AS., Knirel, Y.A., lvanova, E.F! and Ovodov, Y.S. (1990) Bioorg. Khim. 16, 1426-1429. [217] Vinogradov, E.V., Shashkov, AS., Knirel, Y.A., Kochetkov, N.K., Dabrowski, J., Grosskurth, H., Stanislavsky, E.S. and Kholodkova, E.V. (1992) Carbohydr. Res. 231, 1-1 1. [218] Beynon, L.M., Richards, J.C. and Perry, M.B. (1994) Carbohydr. Res. 256, 303-317. [219] Knirel, Y.A., Rietschel, E.T., Marre, R. and Zahringer, U. (1994) Eur. J. Biochem. 221, 239-245. [220] Zahringer, U., Knirel, Y.A., Lindner, B., Helbig, J.H., Sonesson, A,, Marre, R. and Rietschel, E.T. (1995) In: J. Levin, C.R. Alving, R.S. Munford and H. Redl (Eds.), Progress in Clinical and Biological Research, Vol. 392, Bacterial endotoxins - Lipopolysaccharides from genes to therapy. Wiley and Liss, New York, pp. 113-139. [220a] Knirel, Y.A., Helbig, J.H. and Zahringer, U. (1996) Carbohydr. Res. 283, 129-139. [221] Culling, C.F.A. and Reid, P.E. (1982) In: R. Schauer (Ed.), Sialic acids - Chemistry, Metabolism and Function. Springer, Wien, pp. 173-193. [222] Mandal, C. and Mandal, C. (1990) Experientia 46, 433441. [223] Mercy, S.P.D. and Ravindranath, M.H. (1993) Eur. J. Biochem. 215, 697-704. [224] Knibbs, R.N., Osborne, S.E., Glick, G.D. and Goldstein, I.J. (1993) J. Biol. Chem. 268, 18524-18531. [225] Tsuboi, I . and Yanagi, K. (1993) Comp. Biochem. Physiol. 1048, 19-26. [226] Babil, I?, Pindak, F.F., Wells, D.J. and Gardner, W.A. (1994) Biochem. J. 299, 341-346. [227] Powell, L.D. and Varki, A. (1994) J. Biol. Chem. 269, 10628-10636. [228] Haselbeck, A. and Hosel, W. (1990) Glycoconjugate J. 7, 63-74. [229] Haselbeck, A,, Schickaneder, E., von der Eltz, H. and Hosel, W. (1990) Anal. Biochem. 191, 25-30. [230] Zimmer, G., Reuter, G. and Schauer, R. (1992) Eur. J. Biochem. 204, 209-215. [231] Zimmer, G., Suguri, T., Reuter, G., Yu, R.K., Schauer, R. and Herrler, G. (1994) Glycobiology 4, 343-349. [232] Manuguerra, J.-C., DuBois, C. and Hannoun, C. (1991) Anal. Biochem. 194, 425432. [233] Muchmore, E.A. and Varki, A. (1987) Science 236, 1293-1295. [234] Klein, A,, Krishna, M., Varki, N.M. and Varki, A. (1994) Proc. Natl. Acad. Sci. USA 91, 7782-7786. [235] Sjoberg, E.R., Powell, L.D., Klein, A. and Varki, A. (1994) J. Cell Biol. 126, 549-562. [236] Aubeck, R., Eppelsheim, C., Brauchle, C. and Hampp, N. (1993) Analyst 118, 1389-1392. [237] Schauer, R. (1978) Methods Enzymol. 50, 64-89. [238] Varki, A. and Diaz, S. (1984) Anal. Biochem. 137, 236-247. I2391 Lagana, A,, Marino, A,, Fago, G. and Martinez, B.P. (1993) Anal. Biochem. 215, 266-272. [240] Kamerling, J.P., Gerwig, G.J., Vliegenthart, J.F.G. and Clamp, J.R. (1975) Biochem. J. 151, 491495. [241] Kamerling, J.P. and Vliegenthart, J.F.G. (1989) In: A.M. Lawson (Ed.), Clinical Biochemistry -

380 Principles, Methods, Applications, Vol. I , Mass spectrometry. Walter de Gruyter, Berlin, pp. 175263. [242] Yu, R.K. and Ledeen, R.W. (1970) J. Lipid Res. 11, 506-516. [243] Kitajima, K., Inoue, S., Kitazume, S. and Inoue, Y. (1992) Anal. Biochem. 205, 244-250. [244] Corfield, A.P., Michalski, J.-C. and Schauer, R. (1981) In: G. Tettamanti, P. Durand and S. Di Donato (Eds.), Sialidases and Sialidosis, Perspectives in Inherited Metabolic Diseases, Vol. 4. Edi Ermes, Milan, pp. 3-70. [245] Corfield, T. (1992) Glycobiology 2, 509-521. [246] Saito, M. and Yu, R.K. (1995) In: A. Rosenberg (Ed.), Biology of the Sialic Acids. Plenum Press, New York, pp. 261-313. [247] Hoyer, L.L., Roggentin, P., Schauer, R. and Vimr, E.R. (1991) J. Biochem. (Tokyo) 110, 462467. [248] Tanaka, H., Ito, F. and Iwasaki, T. (1994) J. Biochem. (Tokyo) 115, 318-321. [249] Corfield, A.P., Higa, H., Paulson, J.C. and Schauer, R. (1983) Biochim. Biophys. Acta 744, 121-126. [250] Varki, A. and Diaz, S. (1983) J. Biol. Chem. 258, 12465-12471. [251] Corfield, A.P., Sander-Wewer, M., Veh, R.W., Wember, M. and Schauer, R. (1986) Biol. Chem. HoppeSeyler 367, 433439. [252] Kleineidam, R.G., Furuhata, K., Ogura, H. and Schauer, R. (1990) Biol. Chem. Hoppe-Seyler 371, 71 5-719. [253] Kitazume, S., Kitajima, K., Inoue, S., Troy, EA., Lennarz, W.J. and Inoue, Y. (1994) Biochem. Biophys. Res. Commun. 205, 893-898. [254] Angata, T., Kitajima, K., Inoue, S., Chang, J., Warner, T.G., Troy, EA. and Inoue, Y. (1994) Glycobiology 4, 5 17-523. [255] Kitajima, K., Kuroyanagi, H., Inoue, S., Ye, J., Troy, F.A. and Inoue, Y. (1994) J. Biol. Chem. 269, 2 1415-2 14 19. [256] Jennings, H.J., Roy, R. and Michon, E (1985) J. Immunol. 134, 2651-2657. [257] Lifely, M.R., Nowicka, U.T. and Moreno, C. (1986) Carbohydr. Res. 156, 123-135. [258] Manzi, A.E., Higa, H.H., Diaz, S. and Varki, A. (1994) J. Biol. Chem. 269, 23617-23624. [259] Hallenbeck, P.C., Vimr, E.R., Yu, E, Bassler, B. andTroy, F.A. (1987) J. Biol. Chem. 262,3553-3561. [260] Troy, EA., Hallenbeck, P.C., McCoy, R.D. and Vim, E.R. (1987) Methods Enzymol. 138, 169-1 85. [261] Kwiatkowski, B. and Stirm, S. (1987) Methods Enzymol. 138, 786-792. [262] Schauer, R. and Corfield, A.P. (1982) In: R. Schauer (Ed.), Sialic acids - Chemistry, Metabolism and Function. Springer, Wien, pp. 77-94. [263] Yao, K., Ubuka, T., Masuoka, N., Kinuta, M. and Ikeda, T. (1989) Anal. Biochem. 179, 332-335. [264] Sugahara, K., Sugimoto, K., Nomura, 0. and Usui, T. (1980) Clin. Chim. Acta 108, 493498. [265] Shukla, A.K. and Schauer, R. (1982) Hoppe-Seyler’s Z. Physiol. Chem. 363, 255-262. [266] Schauer, R., Reuter, G. and Stoll, S. (1988) Biochimie 70, 1511-1519. [267] Kitazume, S., Kitajima, K., Inoue, S. and Inoue, Y. (1992) Anal. Biochem. 202, 25-34. [268] Manzi, A.E., Diaz, S. and Varki, A. (1990) Anal. Biochem. 188, 20-32. [269] Shukla, A.K. and Schauer, R. (1982) J. Chromatogr. 244, 81-89. I2701 Shukla, A.K. and Schauer, R. (1986) Anal. Biochem. 158, 158-164. [271] Hara, S., Yamaguchi, M., Takemori, Y., Furahata, K., Ogura, H. and Nakamura, M. (1989) Anal. Biochem. 179, 162-1 66. [272] Hokke, C.H., Bergwerff, A.A., van Dedem, G.W.K., van Oostrum, J., Kamerling, J.P. and Vliegenthart, J.F.G. (1990) FEBS Lett. 275, 9-14. [273] Hayakawa, K., De Felice, C., Watanabe, T., Tanaka, T., linuma, K., Nihei, K., Higuchi, S., Ezoe, T., Hibi, I. and Kurosawa, K. (1993) J. Chromatogr. 620, 25-3 1. [274] Ishida, J., Nakahara, T. and Yamaguchi, M. (1992) Biomed. Chromatogr. 6, 135-140. [275] Shukla, A.K., Scholz, N., Reimerdes, E.H. and Schauer, R. (1982) Anal. Biochem. 123, 78-82. [276] Honda, S. and Suzuki, S. (1984) Anal. Biochem. 142, 167-174. [277] Powell, L.D. and Hart, G.W. (1986) Anal. Biochem. 157, 179-185. [278] Unland, F. and Miithing, J. (1992) Biomed. Chromatogr. 6, 155-159. [279] Kobayashi, K., Akiyama, Y., Kawaguchi, K., Tanabe, S. and Imanari, T. (1985) Anal. Sci. 1, 81.

38 1 [280] Hara, S., Yamaguchi, M., Takemori, Y., Nakamura, M. and Ohkura, Y. (1986) J. Chromatogr. 377, 11 1-1 19. [281] Budd, T.J., Dolman, C.D., Lawson, A.M., Chai, W., Saxton, J. and Hemming, F.W. (1992) Glycoconjugate J. 9, 274-278. [282] Shukla, A.K. and Schauer, R. (1982) Fresenius Z. Anal. Chem. 31 1, 377. [283] Diaz, S. and Varki, A. (1985) Anal. Biochem. 150, 3 2 4 6 . [284] Diaz, S., Higa, H.H., Hayes, B.K. and Varki, A. (1989) J. Biol. Chem. 264, 19416-19426. [285] Reinhard, B., Becker, A,, Rothermel, J. and Faillard, H. (1992) Biol. Chem. Hoppe-Seyler 373,63-68. [286] Sato, T., Omichi, K. and Ikenaka, T. (1988) J. Biochem. (Tokyo) 104, 18-21. [287] Ferrero, M.A., Luengo, J.M. and Reglero, A. (1991) Biochem. J. 280, 575-579. [288] Hallenbeck, P.C., Yu, F. and Troy, F.A. (1987) Anal. Biochem. 161, 181-186. [289] Ye, J., Kitajima, K., Inoue, Y., Inoue, S. and Troy, F.A. (1994) Methods Enzymol. 230, 460484. [290] Ashwell, G., Berlin. W.K. and Gabriel, 0. (1994) Anal. Biochem. 222, 495-502. [291] Townsend, R.R., Hardy, M.R., Hindsgaul, 0. and Lee, Y.C. (1988) Anal. Biochem. 174,459470. [292] Townsend, R.R., Hardy, M.R., Cumming, D.A., Carver, J.P. and Bendiak, B. (1989) Anal. Biochem. 182, 1-8. [293] Hermentin, P., Witzel, R., Vliegenthart, J.F.G., Kamerling, J.P., Nimtz, M. and Conradt, H.S. (1992) Anal. Biochem. 203, 281-289. [294] Hayase, T., Sheykhanazari, M., Bhavanandan, V.P., Savage, A.V and Lee, Y.C. (1993) Anal. Biochem. 21 1, 72-80. [295] Lee, Y.C. and Rice, K.G. (1993) In: M. Fukuda and A. Kobata (Eds.), Glycobiology - A Practical Approach. IRL Press, Oxford, pp. 127-163. [296] Piller, F, and Piller, V (1993) In: M. Fukuda and A. Kobata (Eds.), Glycobiology - A Practical Approach. IRL Press, Oxford, pp. 291-328. [297] Hounsell, E.F. (Ed.) (1993) Glycoprotein Analysis in Biomedicine. Humana Press, Towota. [298] Toomre, D.K. and Varki, A. (1994) Glycobiology 4, 653-663. [299] Hase, S. (1994) Methods Enzymol. 230, 225-237. [300] Ruan, S. and Lloyd, K.O. (1994) Glycoconjugate J. 1 1 , 249-256. [301] Hermentin, P., Doenges, R., Witzel, R., Hokke, C.H., Vliegenthart, J.F.G., Kamerling, J.P., Conradt, H.S., Nimtz, M. and Brazel, D. (1994) Anal. Biochem. 221, 2 9 4 1 . [302] Kakehi, K., Susami, A,, Taga, A,, Suzuki, S. and Honda, S. (1994) J. Chromatogr. 680, 209-215. [303] Kamerling, J.P., Vliegenthart, J.F.G., Versluis, C. and Schauer, R. (1975) Carbohydr. Res. 41, 7-17. [304] Reuter, G. and Schauer, R. (1986) Anal. Biochem. 157, 3 9 4 6 . [305] Schroder, C. and Schauer, R. (1982) Fresenius Z. Anal. Chem. 311, 385. [306] Lonngren, J. and Svensson, S. (1974) Adv. Carbohydr. Chem. Biochem. 29, 41-106. [307] Schauer, R., Stoll, S. and Reuter, G. (1991) Carbohydr. Res. 213, 353-359. [308] Reuter, G., Schauer, R., Szeiki, C., Kamerling, J.P. and Vliegenthart, J.F.G. (1989) Glycoconjugate J. 6, 3 5 4 4 . [309] Shukla, A.K., Schauer, R., Schade, U., Moll, H. and Rietschel, E. Th. (1985) J. Chromatogr. 337, 231-238. [310] Devine, P.L., Clark, B.A., Birrell, G.W., Layton, G.T., Ward, B.G., Alewood, P.F. and McKenzie, 1.F.C. (1991) Cancer Res. 51, 5826-5836. [31 I ] Dell, A. (1987) Adv. Carbohydr. Chem. Biochem. 45, 19-72. [312] Dell, A,, Khoo, K.-H., Panico, M., McDowell, R.A., Etienne, A.T., Reason, A.J. and Morris, H.R. (1993) In: M. Fukuda and A. Kobata (Eds.), Glycobiology - A Practical Approach. IRL Press, Oxford, pp, 187-222. [313] Van Kuik, J.A. and Vliegenthart, J.F.G. (1994) Carbohydr. Europe 10, 31-32. [314] Higa, H.H. and Paulson, J.C. (1985) J. Biol. Chem. 260, 8838-8849. [315] Auge, C. and Gautheron, C. (1988) Tetrahedron Lett. 29, 789-790. [3 161 Gross, H.J., Bunsch, A., Paulson, J.C. and Brossmer, R. (1987) Eur. J. Biochem. 168, 595-602. [317] Friebolin, H., Brossmer, R., Keilich, G., Ziegler, D. and Supp, M. (1980) Hoppe-Seyler’s Z. Physiol. Chem. 361, 697-702.

3 82 [318] Friebolin, H., Baumann, W., Brossmer, R., Keilich, G., Supp, M., Ziegler, D. and Von Nicolai, H. (1981) Biochem. Int. 4, 321-326. [319] Deijl, C.M. and Vliegenthart, J.F.G. (1983) Biochem. Biophys. Res. Commun. 1 1 I , 668-674. [320] Chong, A.K.J., Pegg, M.S., Taylor, N.R., von Itzstein, M. (1992) Eur. J. Biochem. 207, 335-343. [321] Friebolin, H., Kunzelmann, P., Supp, M., Brossmer, R., Keilich, G. and Ziegler, D. (1981) Tetrahedron Lett. 22, 1383-1 386. [322] Baumann, W., Freidenreich, J., Weisshaar, G., Brossmer, R. and Friebolin, H. (1989) Biol. Chem. Hoppe-Seyler 370, 141-149. [323] Friebolin, H., Baumann, W., Keilich, G., Ziegler, D., Brossmer, R. and Von Nicolai, H. (1981) HoppeSeyler’s Z. Physiol. Chem. 362, 1455-1463. [324] Paulson, J.C., Weinstein, J., Dorland, L., van Halbeek, H. and Vliegenthart, J.F.G. (1982) J. Biol. Chem. 257, 12734-12738. [325] Friebolin, H., Baumann, W., Hauck, M., Kurz, D., Wajda, R., Weisshaar, G., Keilich, G., Ziegler, D., Brossmer, R. and Von Nicolai, H. (1984) Hoppe-Seyler’s Z. Physiol. Chem. 365, 1309-1321. [326] Dorland, L., Haverkamp, J., Schauer, R., Veldink, G.A. and Vliegenthart, J.F.G. (1 982) Biochem. Biophys. Res. Commun. 104, 1 1 14-1 119. [327] Friebolin, H., Schmidt, H. and Supp, M. (1981) Tetrahedron Lett. 22, 5171-5174. [328] Hird, K. and Vliegenthart, J.F.G. (1993) In: M. Fukuda and A. Kobata (Eds.), Glycobiology - A Practical Approach. IRL Press, Oxford, pp. 223-242. [329] Nomoto, H., Iwasaki, M., Endo, T., Inoue, S., Inoue, Y. and Matsumura, G. (1982) Arch. Biochem. Biophys. 218, 335-341. [330] Kitajima, K., Nomoto, H., Inoue, Y., Iwasaki, M. and Inoue, S. (1984) Biochemistry 23, 310-316. [331] Machytka, D., Klein, R.A. and Egge, H. (1994) Carbohydr. Res. 254, 289-294. [332] Strecker, G., Wieruszeski, J.-M., Cuvillier, O., Michalski, J.-C. and Montreuil, J. (1992) Biochimie 74, 39-52. [333] Joziasse, D.H., Schiphorst, W.E.C.M., van den Eijnden, D.H., van Kuik, J.A., van Halbeek, H. and Vliegenthart, J.F.G. (1985) J. Biol. Chem. 260, 714-719. [334] Nemansky, M. and van den Eijnden, D.H. (1993) Glycoconjugate J. 10, 99-108. [335] Van den Eijnden, D.H. and Joziasse, D.H. (1993) Curr. Opinion Struct. Biol. 3, 71 1-721. [336] Hori, H., Nakajima, T., Nishida, Y., Ohrui, H. and Meguro, H. (1988) Tetrahedron Lett. 29, 63 17-6320. [337] Prytulla, S., Lambert, J., Lautenvein, J., Klessinger, M. and Thiem, J. (1990) Magn. Reson. Chem. 28, 888-901. [338] Prytulla, S., Lautenvein, J., Klessinger, M. and Thiem, J. (1991) Carbohydr. Res. 215, 345-349. [339] Okamoto, K. and Goto, T. (1990) Tetrahedron 46, 5835-5857. [340] Berman, E. (1986) Carbohydr. Res. 152, 3 3 4 6 . [341] Berman, E. and Bendel, P. (1986) FEBS Lett. 204, 257-260. [342] Sabesan, S. and Paulson, J.C. (1986) J. Am. Chem. SOC.108, 2068-2080. [343] Berman, E., Dabrowski, U. and Dabrowski, J. (1988) Carbohydr. Res. 176, 1-15. [344] Vliegenthart, J.F.G. and Kamerling, J.P. (1982) In: R. Schauer (Ed.), Sialic acids - Chemistry, Metabolism and Function. Springer, Wien, pp. 59-76. [345] Toone, E.J., Simon, E.S., Bednarski, M.D. and Whitesides, G.M. (1989) Tetrahedron 45, 5365-5422. [346] DeNinno, M.P. (1991) Synthesis 583-593. [347] Hasegawa, A. and Kiso, M. (1992) In: H. Ogura, A. Hasegawa and T. Suami (Eds.), Carbohydrates Synthetic Methods and Applications in Medicinal Chemistry. Kodansha, Tokyo, pp. 243-266. [348] Ikeda, K. and Achiwa, K. (1992) In: H. Ogura, A. Hasegawa and T. Suami (Eds.), Carbohydrates Synthetic Methods and Applications in Medicinal Chemistry. Kodansha, Tokyo, pp. 267-28 1 . [349] Ogura, H. (1992) In: H. Ogura, A. Hasegawa and T. Suami (Eds.), Carbohydrates - Synthetic Methods and Applications in Medicinal Chemistry. Kodansha, Tokyo, pp. 282-303, [350] Zbiral, E. (1992) In: H. Ogura, A. Hasegawa and T. Suami (Eds.), Carbohydrates - Synthetic Methods and Applications in Medicinal Chemistry. Kodansha, Tokyo, pp. 304339. [351] Ichikawa, Y., Look, G.C. and Wong, C.-H. (1992) Anal. Biochem. 202, 215-238. [352] Ogura, H. (1994) Yakugaku Zasshi 114, 277-303.

383 [353] Wong, C.-H. and Whitesides, G.M. (1994) Enzymes in Synthetic Organic Chemistry, Tetrahedron Organic Chemistry, Vol. 12. Elsevier Science, Oxford. [354] Benzing-Nguyen, L. and Perry, M.B. (1978) J. Org. Chem. 43, 551-554. [355] Danishefsky, S.J., DeNinno, M.P. and Chen, S. (1988) J. Am. Chem. SOC.110, 3929-3940. [356] Gordon, D.M. and Whitesides, G.M. (1993) J. Org. Chem. 58, 7937-7938. [357] Baumberger, F. and Vasella, A. (1986) Helv. Chim. Acta 69, 1205-1215. [358] Baumberger, F. and Vasella, A. (1986) Helv. Chim. Acta 69, 1535-1541. [359] Julina, R., Muller, I., Vasella, A. and Wyler, R. (1987) Carbohydr. Res. 164, 415432. [360] Csuk, R., Hugener, M. and Vasella, A. (1988) Helv. Chim. Acta 71, 609418. [361] Yamamoto, T., Teshima, T., Inami, K. and Shiba, T. (1992) Tetrahedron Lett. 33, 325-328. [362] Bodenmiiller, A. and Schmidt, R.R. (1994) Liebigs Ann. Chem. 541-548. [363] Marra, A. and Sinay, P. (1989) Carbohydr. Res. 190, 317-322. [364] Shirai, R., Nakamura, M., Hara, S., Takayanagi, H. and Ogura, H. (1988) Tetrahedron Lett. 29, 44494452. [365] Shirai, R. and Ogura, H. (1989) Tetrahedron Lett. 30, 2263-2264. [366] Chan, T.-H. and Li, C.-J. (1992) J. Chem. SOC.Chem. Commun. 747-748. [367] Dondoni, A. and Marra, A. (1994) Carbohydr. Lett. I, 4 3 4 6 . [368] Dondoni, A,, Marra, A. and Merino, P. (1994) J. Am. Chem. SOC.116, 3324-3336. [369] Auge, C., David, S. and Gautheron, C. (1984) Tetrahedron Lett. 25, 46634664. [370] David, S. and Auge, C. (1987) Pure Appl. Chem. 59, 1501-1508. [371] Kragl, U., Gygax, D., Ghisalba, 0. and Wandrey, C. (1991) Angew. Chem. Int. Ed. Engl. 30, 827-828. [372] Kim, M.-J., Hennen, W.J., Sweers, H.M. and Wong, C.-H. (1988) J. Am. Chem. SOC.110,6481-6486. [373] Bednarski, M.D., Chenault, H.K., Simon, E.S. and Whitesides, G.M. (1987) J. Am. Chem. SOC. 109, 1283-1285. [374] Lubineau, A,, Augt, C., Gautheron-Le Narvor, C. and Ginet, J.-C. (1994) BioMed. Chem. 2, 669-674. [375] AugC., C. and Gautheron, C. (1987) J. Chem. SOC.Chem. Commun. 859-860. [376] Auge, C., Gautheron, C., David, S., Malleron, A., Cavaye, B. and Bouxom, B. (1990) Tetrahedron 46, 201-214. [377] Halcomb, R.L., Fitz, W. and Wong, C.-H. (1994) Tetrahedron: Asymmetry 5, 2437-2442. [378] Hasegawa, A,, Murase, T., Ogawa, M., Ishida, H. and Kiso, M. (1990) J. Carbohydr. Chem. 9,415428. [379] Hasegawa, A,, Murase, T., Ogawa, M., Ishida, H. and Kiso, M. (1990) J. Carbohydr. Chem. 9,429439. [380] Ogura, H., Furuhata, K., Sato, S., Anazawa, K., Itoh, M. and Shitori, Y. (1987) Carbohydr. Res. 167, 77-86. [381] Furuhata, K. and Ogura, H. (1989) Chem. Pharm. Bull. 37, 2037-2040. [382] Reinhard, B., Gotz, C. and Faillard, H. (1992) Biol. Chem. Hoppe-Seyler 373, 1243-1248. I3831 Augt, C., David, S., Gautheron, C. and Veyritres, A. (1985) Tetrahedron Lett. 26, 2439-2440. [384] Liu, J.L.-C., Shen, G.-J., Ichikawa, Y., Rutan, J.F., Zapata, G., Vann, W.F. and Wong, C.-H. (1992) J. Am. Chem. SOC.114, 3901-3910. [385] Forstner, M., Freytag, K. and Paschke, E. (1989) Carbohydr. Res. 193, 294-295. [386] Shigeta, S., Winter, H.C. and Goldstein, I.J. (1994) Carbohydr. Res. 264, 111-121. [387] Murray, M.C., Bhavanandan, V.P., Davidson, E.A. and Reinhold, V. (1989) Carbohydr. Res. 186, 255265. [388] Kragl, U., Godde, A,, Wandrey, C., Kinzy, W., Cappon, J.J. and Lugtenburg, J. (1993) Tetrahedron: Asymmetry 4, 1 193-1 202. [389] Schreiner, E., Zbiral, E., Kleineidam, R.G. and Schauer, R. (1991) Carbohydr. Res. 216, 61-66. [390] Brossmer, R. and Gross, H.J. (1994) Methods Enzymol. 247, 153-176. [391] Isecke, R. and Brossmer, R. (1994) Tetrahedron SO, 7445-7460. [392] Petrie, C.R., Sharma, M., Simmons, O.D. and Korytnyk, W. (1989) Carbohydr. Res. 186, 326-334. [393] Schmid, W., Christian, R. and Zbiral, E. (1988) Tetrahedron Lett. 29, 3643-3646. [394] Kelm, S., Paulson, J.C., Rose, U., Brossmer, R., Schmid, W., Bandgar, B.P., Schreiner, E., Hartmann, M. and Zbiral, E. (1992) Eur. J. Biochem. 205, 147-153. [395] Hanessian, S. and Girard, C. (1994) Synlett 863-864. [396] Hartmann, M. and Zbiral, E. (1991) Liebigs Ann. Chem. 795-801.

384 [397] Kessler, J., Heck, J., Tanenbaum, S.W. and Flashner, M. (1982) J. Biol. Chem. 257, 505&5060. [398] Bandgar, B.P., Hartmann, M., Schmid, W. and Zbiral, E. (1990) Liebigs Ann. Chem. 1185-1 195. [399] Okamoto, K., Hasegawa, T., Toyomaki, Y., Yamakawa, M. and Okukado, N. (1992) Chem. Pharm. Bull. 40, 2728-2734. [400] Brandstetter, H.H. and Zbiral, E. (1983) Liebigs Ann. Chem. 2055-2065. [401] Hagedom, H.W. and Brossmer, R. (1986) Helv. Chim. Acta 69, 2127-2132. [402] Gross, H.J. and Brossmer, R. (1987) Glycoconjugate J. 4, 145-156. [403] Estenne, G., Saroli, A. and Doutheau, A. (1991) J. Carbohydr. Chem. 10, 181-195. [404] Luger, P., Zaki, C., Hagedorn, H.-W. and Brossmer, R. (1987) Carbohydr. Res. 164, 49-58. [405] Hartmann, M. and Zbiral, E. (1 990) Tetrahedron Lett. 3 1, 2875-2878. [406] Gross, H.J., Kovac, A,, Rose, U., Watzlawick, H. and Brossmer, R. (1988) Biochemistry 27, 42794283. [407] Beau, J.-M. and Schauer, R. (1980) Eur. J. Biochem. 106, 531-540. [408] Hartmann, M., Christian, R. and Zbiral, E. (1990) Liebigs Ann. Chem. 83-91. [409] Gross, H.J. and Brossmer, R. (1988) FEBS Lett. 232, 145-147. [410] Zbiral, E., Brandstetter, H.H. and Schreiner, E.P. (1988) Monatsh. Chem. 119, 127-141. [41 I] Zbiral, E. and Brandstetter, H.H. (1985) Monatsh. Chem. 116, 87-98. [412] Salunkhe, M., Hartmann, M., Schmid, W. and Zbiral, E. (1988) Liebigs Ann. Chem. 187-189. [413] Hartmann, M. and Zbiral, E. (1991) Monatsh. Chem. 122, 995-1003. [414] Isecke, R. and Brossmer, R. (1995) Carbohydr. Res. 274, 303-31 1. [415] Sticher, U., Gross, H.J. and Brossmer, R. (1991) Glycoconjugate J. 8, 45-54. [416] Gross, H.J. and Brossmer, R. (1988) Glycoconjugate J. 5, 41 1 4 1 7 . [417] Gross, H.J., Rose, U., Krause, J.M., Paulson, J.C., Schmid, K., Feeney, R.E. and Brossmer, R. (1989) Biochemistry 28, 7386-7392. [418] Brossmer, R., Rose, U., Kasper, D., Smith, T.L., Grasmuk, H., Unger, EM. (1980) Biochem. Biophys. Res. Commun. 96, 1282-1289. [419] Brossmer, R. and Gross, H.J. (1994) Methods Enzymol. 247, 177-193. [420] Fitz, W. and Wong, C.-H. (1994) J. Org. Chem. 59, 8279-8280. [421] Schreiner, E., Christian, R. and Zbiral, E. (1990) Liebigs Ann. Chem. 93-97. [422] Gross, H.J. and Brossmer, R. (1988) Eur. J. Biochem. 177, 583-589. [423] Gross, H.J., Sticher, U. and Brossmer, R. (1990) Anal. Biochem. 186, 127-134. [424] Conradt, H.S., Bunsch, A. and Brossmer, R. (1984) FEBS Lett. 170, 295-300. [425] Sharma, M., Petrie, C.R. and Korytnyk, W. (1988) Carbohydr. Res. 175, 25-34. [426] Gross, H.J., Bunsch, A. and Brossmer, R. (1984) Hoppe-Seyler’s Z. Physiol. Chem. 365, 994-995. [427] Petrie, C.R. and Korytnyk, W. (1981) Anal. Biochem. 131, 153-159. [428] Holzer, C.T., von ltzstein, M., Jin, B., Pegg, M.S., Stewart, W.P. and Wu, W.-Y. (1993) Glycoconjugate J. 10, 4 0 4 4 . [429] Von Itzstein, M., Wu, W.-Y. and Jin, B. (1994) Carbohydr. Res. 259, 301-305. [430] Schreiner, E., Zbiral, E., Kleineidam, R.G. and Schauer, R. (1991) Liebigs Ann. Chem. 129-134. [431] Von Itzstein, M., Jin, B., Wu, W.-Y. and Chandler, M. (1993) Carbohydr. Res. 244, 181-185. [432] Warner, T.G. (1987) Biochem. Biophys. Res. Commun. 148, 1323-1329. [433] Kumar, V, Tanenbaum, S.W. and Flashner, M. (1982) Carbohydr. Res. 103, 281-285. [434] Kumar, V., Kessler, J., Scott, M.E., Patwardhan, B.H., Tanenbaum, S.W. and Flashner, M. (1981) Carbohydr. Res. 94, 123-130. [435] Flashner, M., Kessler, J. and Tanenbaum, S.W. (1983) Arch. Biochem. Biophys. 221, 188-196. [436] Mack, H. and Brossmer, R. (1987) Tetrahedron Lett. 28, 191-194. [437] Zbiral, E., Brandstetter, H.H., Christian, R. and Schauer, R. (1987) Liebigs Ann. Chem. 781-786. [438] Driguez, P.-A., Barrere, B., Quash, G. and Doutheau, A . (1994) Carbohydr. Res. 262, 297-310. [439] Baumberger, F., Vasella, A. and Schauer, R. (1988) Helv. Chim. Acta 71, 429445. [440] Zbiral, E., Schreiner, E. and Christian, R. (1989) Carbohydr. Res. 194, C15-Cl8. [441] Dondoni, A,, Boscarato, A. and Marra, A. (1994) Tetrahedron: Asymmetry 5, 2209-2212. [442] Hagiwara, T., Kijima-Suda, I., Ido, T., Ohrui, H. and Tomita, K. (1994) Carbohydr. Res. 263, 167-172. [443] Nakajima, T., Hori, H., Ohrui, H., Meguro, H. and Ido, T. (1988) Agric. Biol. Chem. 52, 1209-1215.

385 [444] Koppert, K. and Brossmer, R. (1992) Tetrahedron Lett. 33, 8031-8034. [445] Schreiner, E. and Zbiral, E. (1990) Liebigs Ann. Chem. 581-586. [446] David, S., Malleron, A. and Cavaye, B. (1994) Carbohydr. Res. 260, 233-241. [447] Schmid, W., Zbiral, E., Christian, R. and Schulz, G. (1986) Liebigs Ann. Chem. 2104-21 11, [448] Brandstetter, H.H., Zbiral, E. and Schulz, G. (1982) Liebigs Ann. Chem. 1-13. [449] Wallimann, K. and Vasella, A. (1990) Helv. Chim. Acta 73, 1359-1372. [450] Vasella, A. and Wyler, R. (1991) Helv. Chim. Acta 74, 451463. [451] Bernet, B., Murty, A.R.C.B. and Vasella, A. (1990) Helv. Chim. Acta 73, 940-958. [452] Glanzer, B.I., Gyorgydeak, Z., Bernet, B. and Vasella, A. (1991) Helv. Chim. Acta 74, 343-369. [453] Thomson, D.S. and Prestegard, J.H. (1990) Carbohydr. Res. 196, 206-210. 14541 Rothermel, J. and Faillard, H. (1989) Biol. Chem. Hoppe-Seyler 370, 1077-1084. [455] Rothermel, J. and Faillard, H. (1990) Carbohydr. Res. 208, 251-254. [456] Privalova, I.M. and Khorlin, A.Ya. (1969) Izv. Akad. Nauk. SSSR, Ser. Khim. 12, 2785-2792. [457] Supp, M., Rose, U. and Brossmer, R. (1980) Hoppe-Seyler’s Z. Physiol. Chem. 361, 338. [458] Tropper, ED., Andersson, F.O., Braun, S. and Roy, R. (1992) Synthesis 618-620. [459] Rothermel, J., Weber, B. and Faillard, H . (1992) Liebigs Ann. Chem. 799-802. [460] Eschenfelder, V. and Brossmer, R. (1 987) Carbohydr. Res. 162, 294-297. [461] Tuppy, H. and Palese, P. (1969) FEBS Lett. 3, 72-75. [462] Santer, U.V., Yee-Foon, J. and Glick, M.C. (1978) Biochim. Biophys. Acta 523, 435442. [463] Myers, R.W., Lee, R.T., Lee, Y.C., Thomas, G.H., Reynolds, L.W. anduchida, Y. (1980)Anal. Biochem. 101, 166-174. [464] Baumberger, F., Vasella, A. and Schauer, R. (1986) Helv. Chim. Acta 69, 1927-1935. [465] Kodama, H., Baum, L.G. and Paulson, J.C. (1991) Carbohydr. Res. 218, 1 1 1-1 19. [466] Eschenfelder, V. and Brossmer, R. (1987) Glycoconjugate J. 4, 171-178. [467] Fujii, I., Iwabuchi, Y., Teshima, T., Shiba, T. and Kikuchi, M. (1993) BioMed. Chem. 1, 147-149. [468] Warner, T.G. and Lee, L.A. (1988) Carbohydr. Res. 176, 21 1-218. [469] Roth, A. and Faillard, H. (1993) Liebigs Ann. Chem. 485489. [470] Kawamura, A,, Kijima-Suda, I., Sugimoto, M., Itoh, M., Takada, K., Tomita, K., Ogawa, T. and Nagai, Y. (1990) Biochim. Biophys. Acta 1033, 201-206. [471] Heidlas, J.E., Williams, K.W. and Whitesides, G.M. (1992) Ace. Chem. Res. 25, 307-314. [472] Stangier, P., Treder, W. and Thiem, J. (1993) Glycoconjugate J. 10, 26-33. [473] Simon, E.S., Bednarski, M.D. and Whitesides, G.M. (1988) J. Am. Chem. Soc. 110, 7159-7163. [474] Kittelmann, M., Klein, T., Kragl, U., Wandrey, C. and Ghisalba, 0. (1992) Ann. NY Acad. Sci. 672, 444450. [475] Makino, S., Ueno, Y., Ishikawa, M., Hayakawa, Y. and Hata, T. (1993) Tetrahedron Lett. 34, 27752778. [476] Martin, T.J. and Schmidt, R.R. (1993) Tetrahedron Lett. 34, 1765-1768. [477] Yoshino, T. and Schmidt, R.R. (1995) Carbohydr. Lett. 1, 329-334. [478] Auge, C., Fernandez-Fernandez, R. and Gautheron, C. (1990) Carbohydr. Res. 200, 257-268. [479] Kajihara, Y., Koseki, K., Ebata, T., Kodama, H., Matsushita, H. and Hashimoto, H. (1994) Carbohydr. Res. 264, CI-C5. [480] Kanie, O., Nakamura, J., Kiso, M. and Hasegawa, A. (1987) J. Carbohydr. Chem. 6, 105-1 15. [481] Smalec, B. and von Itzstein, M. (1995) Carbohydr. Res. 266, 269-272. [482] Ikeda, K., Nagao, Y. and Achiwa, K. (1992) Carbohydr. Res. 224, 123-131. [483] Suzuki, J., Murakami, K. and Nishimura, Y . (1993) J. Carbohydr. Chem. 12, 201-208. [484] Roy, R., Laferriire, C.A., Gamian, A. and Jennings, H.J. (1987) J. Carbohydr. Chem. 6, 161-165. [485] Roy, R. and Laferriere, C.A. (1990) Can. J. Chem. 68, 2045-2054. [486] Roy, R. and Laferritre, C.A. (1988) Carbohydr. Res. 177, CI-C4. [487] Roy, R. and Laferritre, C.A. (1990) J. Chem. SOC.Chem. Commun. 1709-171 1. [488] Roy, R., Tropper, ED., Romanowska, A,, Letellier, M., Cousineau, L., Meunier, S.J. and Boratynski, J. (1 991) Glycoconjugate J. 8, 75-8 1. [489] Roy, R., Andersson, F.O., Harms, G., Kelm, S. and Schauer, R. (1992) Angew. Chem. Int. Ed. Engl. 31, 1478-1481.

386 [490]Roy, R., Pon, R.A., Tropper, ED. and Anderson, F.O. (1993)J. Chem. SOC.Chem. Commun. 264-265. [491] Roy, R., Zanini, D., Meunier, S.J. and Romanowska, A. (1994)ACS Symp. Ser. 560,104119. [492] Kuhn, R., Lutz, F? and MacDonald, D.L. (1966)Chem. Ber. 99,611417. [493]Numata, M., Sugimoto, M. Shibayama, S. and Ogawa, T. (1988)Carbohydr. Res. 174,73-85. [494] Shimizu, C. and Achiwa, K. (1 987) Carbohydr. Res. 166,314-316. [495] Okamoto, K., Kondo, T. and Goto, T. (1986)Chem. Lett. 1449-1452. [496] Ito, Y.and Ogawa, T. (1987)Tetrahedron Lett. 28,62214224. [497] Ito, Y.and Ogawa, T. (1988)Tetrahedron Lett. 29,3987-3990. [498]Kirchner, E., Thiem, F., Dernick, R., Heukeshoven, J. and Thiem, J. (1988) J. Carbohydr. Chem. 7, 453486. [499]Okamoto, K., Kondo, T. and Goto, T. (1986)Tetrahedron Lett. 27,5229-5232. [500]Okamoto, K., Kondo, T. and Goto, T. (1988)Tetrahedron 44,1291-1298. [501]Kanie, O., Kiso, M. and Hasegawa, A. (1988)J. Carbohydr. Chem. 7,501-506. [502] Murase, T., Ishida, H., Kiso, M. and Hasegawa, A. (1988)Carbohydr. Res. 184,Cl-C4. [503] Ito, Y.and Ogawa, T. (1988)Tetrahedron Lett. 29,1061-1064. [504] Hasegawa, A,, Nakamura, J. and Kiso, M. (1986) J. Carbohydr. Chem. 5, 11-19. [505] Hasegawa, A., Nakamura, J. and Kiso, M. (1986) J. Carbohydr. Chem. 5, 21-31. [506]Hasegawa, A., Morita, M., Ito, Y., Ishida, H. and Kiso, M. (1990)J. Carbohydr. Chem. 9,369-392. [507]Bennett, S., von Itzstein, M. and Kiefel, M.J. (1994)Carbohydr. Res. 259,293-299. [508]Hasegawa, A,, Ogawa, H., Ishida, H. and Kiso, M. (1992)Carbohydr. Res. 224,175-184. [509]Hasegawa, A,, Ogawa, M., Kojima, Y. and Kiso, M. (1992)J. Carbohydr. Chem. 11, 333-341. [510]Ito, Y.,Numata, M., Sugimoto, M. and Ogawa, T. (1989)J. Am. Chem. SOC.1 1 1, 8508-8510. [511] Hasegawa, A,, Ogawa, M., Ishida, H. and Kiso, M. (1990)J. Carbohydr. Chem. 9,393414. [512] Ishida, H., Ohta, Y., Tsukada, Y., Kiso, M. and Hasegawa, A. (1993)Carbohydr. Res. 246,75-88. [513] Hasegawa, A., Fushimi, K., Ishida, H. and Kiso, M. (1993)J. Carbohydr. Chem. 12,1203-1216. [514] Hotta, K., Ishida, H., Kiso, M. and Hasegawa, A. (1994)J. Carbohydr. Chem. 13, 175-191. [515] Ito, Y.,Ogawa, T., Numata, M. and Sugimoto, M. (1990)Carbohydr. Res. 202,165-175. [516]Fujita, S., Numata, M., Sugimoto, M., Tomita, K. and Ogawa, T. (1992)Carbohydr. Res. 228.347-370. [517]Fujita, S., Numata, M., Sugimoto, M., Tomita, K. and Ogawa, T. (1994)Carbohydr. Res. 263,181-196. [518] Terada, T., Kiso, M. and Hasegawa, A. (1993)J. Carbohydr. Chem. 12,425440. [519] Marra, A. and Sinay, P (1989) Carbohydr. Res. 187,3542. [520] Marra, A. and Sinay, P. (1990)Carbohydr. Res. 195,303-308. [521] Lonn, H. and Stenvall, K. (1992)Tetrahedron Lett. 33, 115-116. [522] Tropper, ED., Anderson, F.O., Cao, S. and Roy, R. (1992)J. Carbohydr. Chem. 11, 741-750. [523] Birberg, W. and Lonn, H. (1991)Tetrahedron Lett. 32, 7453-7456. [524]Birberg, W. and Lonn, H. (1991)Tetrahedron Lett. 32,7457-7458. [525]Takeda, K., Tsuboyama, K., Toni, K., Furuhata, K., Sato, N. and Ogura, H. (1990)Carbohydr. Res. 203,57-63. [526]Ercegovic, T. and Magnusson, G. (1994)J. Chem. SOC.Chem. Commun. 831-832. [527]Martin, T.J. and Schmidt, R.R. (1992)Tetrahedron Lett. 33, 6123-6126. [528]Kondo, H., Ichikawa, Y. and Wong, C.-H. (1992)J. Am. Chem. SOC.114,8748-8750. [529]Kondo, H., Aoki, S., Ichikawa, Y., Halcomb, R.L., Ritzen, H. and Wong, C.-H. (1994)J. Org. Chem. 59,864-877. (5301 Martin, T.J., Brescello, R., Toepfer, A. and Schmidt, R.R. (1993)Glycoconjugate J. 10, 16-25. [531] Matsuzaki, Y., Nunomura, S., Ito, Y., Sugimoto, M., Nakahara, Y. and Ogawa, T. (1993) Carbohydr. Res. 242,Cl-C6. [532] Ishida, H.-K., Ishida, H., Kiso, M. and i-iasegawa, A. (1994)Tetrahedron: Asymmetry 5,2493-2512. [533] Ishida, H., Ishida, H., Kiso, M. and Hasegawa, A. (1994)J. Carbohydr. Chem. 13, 655464. [534] Abbas, S.Z., Sugiyama, S., Diakur, J., Pon, R.A. and Roy, R. (1990)J. Carbohydr. Chem. 9,891-901. [535] Thiem, J. and Treder, W. (1986)Angew. Chem. Int. Ed. Engl. 25,1096. [536] De Heij, H.T., Kloosterman, M., Koppen, PL., van Boom, J.H. and van den Eijnden, D.H. (1988) J. Carbohydr. Chem. 7,209-219. [537] AugC, C., Gautheron, C. and Pora, H. (1989)Carbohydr. Res. 193,288-293.

387 [538] I5391 [540] [541]

Unverzagt, C., Kunz, H. and Paulson, J.C. (1990) J. Am. Chem. SOC.112, 9308-9309. Lubineau, A., Augtt, C. and Francois, P. (1992) Carbohydr. Res. 228, 137-144. Ichikawa, Y., Liu, J., Shen, G.-J. and Wong, C.-H. (1991) J. Am. Chem. SOC.113, 6300-6302. Ichikawa, Y., Lin, Y.-C., Dumas, D.P., Shen, G.-J., Garcia-Junceda, E., Williams, M.A., Bayer, R., Ketcham, C., Walker, L.E., Paulson, J.C. and Wong, C.-H. (1992) J. Am. Chem. SOC. 114, 92839298. [542] Herrmann, G., Ichikawa, Y., Wandrey, C., Gaeta, F.C.A., Paulson, J.C. and Wong, C.-H. (1993) Tetrahedron Lett. 34, 3091-3094. [543] Kashem, M.A., Wlasichuk, K.B., Gregson, J.M. and Venot, A.P. (1993) Carbohydr. Res. 250, 129-144. [544] Pohlentz, G., Trimborn, M. and Egge, H. (1 994) Glycobiology 4, 6 2 5 6 3 I . [545] Angata, T., Kitazume, S., Terada, T., Kitajima, K., Inoue, S., Troy, F.A. and Inoue, Y. (1994) Glycoconjugate J. 1 I , 493499. [546] Sabesan, S., Duus, J.Q., Neira, S., Domaille, P., Kelm, S., Paulson, J.C. and Bock, K. (1992) J. Am. Chem. SOC.114, 8363-8375. [547] Unverzagt, C., Kelm, S. and Paulson, J.C. (1994) Carbohydr. Res. 285-301. [548] Hasegawa, A,, Adachi, K., Yoshida, M. and Kiso, M. (1992) J. Carbohydr. Chem. 11, 95-1 16. [549] Hasegawa, A., Adachi, K., Yoshida, M. and Kiso, M. (1992) Carbohydr Res. 230, 257-272. [550] Vandekerckhove, F., Schenkman, S., Pontes de Carvalho, L., Tomlinson, S., Kiso, M., Yoshida, M., Hasegawa, A. and Nussenzweig, V (1992) Glycobiology 2, 541-548. [551] Scudder, P., Doom, J.P., Chuenkova, M., Manger, I.D. and Pereira, M.E.A. (1993) J. Biol. Chem. 268, 9886-9891. [552] Ferrero-Garcia, M.A., Trombetta, S.E., Sanchez, D.O., Reglero, A,, Frasch, A.C.C. and Parodi, A.J. (1993) Eur. J. Biochem. 213, 765-771. [553] Nishimura, %-I., Lee, K.B., Matsuoka, K. and Lee, Y.C. (1994) Biochem. Biophys. Res. Commun. 199, 249-254. [554] Ito, Y. and Paulson, J.C. (1993) J. Am. Chem. SOC.115, 7862-7863. [555] Lee, K.B. and Lee, Y.C. (1994) Anal. Biochem. 216, 356-364. [556] Maru, I., Ohta, Y., Okamoto, K., Suzuki, S., Kakehi, K. and Tsukada, Y. (1992) Biosci. Biotech. Biochem. 56, 1557-1561. I5571 Thiem, J. and Sauerbrei, B. (1991) Angew. Chem. Int. Ed. Engl. 30, 1503-1505. [558] Ajisaka, K., Fujimoto, H. and Isomura, M. (1994) Carbohydr. Res. 259, 103-1 15. [559] Nagy, J.O. and Bednarski, M.D. (1991) Tetrahedron Lett. 32, 3953-3955. [560] Paulsen, H. and Matschulat, P. (1991) Liebigs Ann. Chem. 487495. [561] Wallimann, K. and Vasella, A. (1991) Helv. Chim. Acta 74, 1520-1532. [562] Nagy, J.O., Wang, P., Gilbert, J.H., Schaefer, M.E., Hill, T.G., Callstrom, M.R. and Bednarski, M.D. (1992) J. Med. Chem. 35, 45014502. [563] Spevak, W., Nagy, J.O., Charych, D.H., Schaefer, M.E., Gilbert, J.H. and Bednarski, M.D. (1993) J. Am. Chem. SOC.115, 1146-1 147. [564] O’Connell, A.M. (1973) Acta Cryst. B 29, 2320-2328. [565] Kooijman, H., Kroon-Batenburg, L.M.J., Kroon, J., Breg, J.N. and de Boer, J.L. (1990) Acta Cryst. C 46, 4 0 7 4 1 0. [566] Czarniecki, M.F. and Thornton, E.R. (1976) J. Am. Chem. SOC.98, 1023-1025. [567] Czarniecki, M.F. and Thornton, E.R. (1977) J. Am. Chem. SOC.99, 8273-8279. [568] Poppe, L. and van Halbeek, H. (1991) J. Am. Chem. SOC.113, 363-365. [569] Christian, R., Schulz, G., Brandstetter, H.H. and Zbiral, E. (1987) Carbohydr. Res. 162, 1-1 I . [570] Christian, R., Schreiner, E., Zbiral, E. and Schulz, G. (1989) Carbohydr. Res. 194, 4 9 6 1 . [571] Lindon, J.C., Vinter, J.G., Lifely, M.R. and Moreno, C. (1984) Carbohydr. Res. 133, 59-74. [572] Michon, F., Brisson, J.-R. and Jennings, H.J. (1987) Biochemistry 26, 8399-8405. [573] Yamasaki, R. and Bacon, B. (1991) Biochemistry 30, 851-857. 15741 Brisson, J.-R., Baumann, H., Imberty, A,, Perez, S. and Jennings, H.J. (1992) Biochemistry 3 I , 49965004. [575] Baumann, H., Brisson, J.-R., Michon, F., Pon, R. and Jennings, H.J. (1993) Biochemistry 32, 40074013.

388 [576] Hughes, R.C., Sanford, B. and Jeanloz, R.W. (1972) Proc. Natl. Acad. Sci. USA 69, 942-945. [577] Kitazume, S., Kitajima, K., Inoue, S., Haslam, S.M., Morris, H.R., Dell, A,, Lennan, W.J. and Inoue, Y. (1996) J. Biol. Chem. 271, 6694-6701. [578] Zeitler, R., Banzer, J.-P., Bauer, C. and Reutter, W. (1992) BioMetals 5, 103-109. [579] Zeitler, R., Giannis, A,, Danneschewski, S., Henk, E., Henk, T., Bauer, C., Reutter, W. and Sandhoff, K. (1992) Eur. J. Biochem. 204, 1165-1 168. [ S O ] Kayser, H., Zeitler, R., Kannicht, C., Grunow, D., Nuck, R. and Reutter, W. (1992) J. Biol. Chem. 267, 16934-1 6938. [581] Pels Rijcken, W.R., Overdijk, B., van den Eijnden, D.H. and Fenverda, W. (1993) Biochem. J. 293, 207-213. [582] Schmelter, T., Ivanov, S., Wember, M., Stangier, P., Thiem, J. and Schauer, R. (1993) Biol. Chem. Hoppe-Seyler 374, 337-342. [583] Schauer, R., Haverkamp, J. and Ehrlich, K. (1980) Hoppe-Seyler’s Z. Physiol. Chem. 361, 641-648. [584] Ambrose, M.G., Freese, S.J., Reinhold, M.S., Warner, T.G. and Vann, W.F. (1992) Biochemistry 31, 775-780. [585] Vimr, E.R. (1992) J. Bactenol. 174, 6191-6197. [586] Vann, W.F., Silver, R.P., Abeijon, C., Chang, K., Aaronson, W., Sutton, A,, Finn, C.W., Lindner, W. and Kotsatos, M. (1987) J. Biol. Chem. 262, 17556-17562. I5871 Edwards, U. and Frosch, M. (1992) FEMS Microbiol. Lett. 96, 161-166. [588] Ganguli, S., Zapata, G., Wallis, T., Reid, C., Boulnois, G., Vann, W.F. and Roberts, 1,s. (1994) J. Bactenol. 176, 45834589. [589] Ichikawa, Y., Wang, R. and Wong, C.-H. (1994) Methods Enzymol. 247, 107-127. [590] Zapata, G., Roller, P.P., Crowley, J. and Vann, W.F. (1993) Biochem. J. 295, 485491. [591] Paulson, J.C. and Colley, K.J. (1989) J. Biol. Chem. 264, 17615-17618. [592] Kleene, R. and Berger, E.G. (1993) Biochim. Biophys. Acta 1154, 283-325. [593] Basu, S.C. (1991) Glycobiology I , 469475. [594] Clausen, H., Bennett, E.P. and Dabelsteen, E. (1992) APMIS Suppl. 27, 100, 9-17. [595] Basu, S., Basu, M. and Basu, S.S. (1995) In: A. Rosenberg (Ed.), Biology of the Sialic Acids. Plenum Publ. Corp., New York, pp. 69-93. [596] Gillespie, W.M., Kelm, S. and Paulson, J.C. (1992) J. Biol. Chem. 267, 21004-21010. [597] Lee, Y.-C., Kurosawa, N., Hamamoto, T., Nakaoka, T. and Tsuji, S. (1993) Eur. J. Biochem. 216, 377-3 85. [598] Kurosawa, N., Hamamoto, T., Inoue, M. and Tsuji, S. (1995) Biochim. Biophys. Acta 1244,216222, [599] Lee, Y.-C., Kojima, N., Wada, E., Kurosawa, N., Nakaoka, T., Hamamoto, T. and Tsuji, S. (1994) J. Biol. Chem. 269, 10028-10033. [600] Kojima, N., Lee, Y.-C., Hamamoto,T., Kurosawa, N. and Tsuji, S. (1994) Biochemistry 33, 5772-5776. [601] Wen, D.X., Livingston, B.D., Medzihradszky, K.F., Kelm, S., Burlingame, A.L. and Paulson, J.C. (1992) J. Biol. Chem. 267, 2101 1-21019. [602] Kitagawa, H. and Paulson, J.C. (1993) Biochem. Biophys. Res. Commun. 194, 375-382. [603] Kono, M., Ohyama, Y., Lee, Y.-C., Hamamoto, T., Kojima, N. and Tsuji, S. (1997) Glycobiology 7, 469479. [604] Sasaki, K., Watanabe, E., Kawashima, K., Sekine, S., Dohi, T., Oshima, M., Hanai, N., Nishi, T. and Hasegawa, M. (1993) J. Biol. Chem. 268, 22782-22787. [605] Kitagawa, H. and Paulson, J.C. (1994) J. Biol. Chem. 269, 1394-1401. [606] Weinstein, J., Lee, E.U., McEntee, K., Lai, P.-H., Paulson, J.C. (1987) J. Biol. Chem. 262, 1773517743. [607] Grundmann, U., Nerlich, C., Rein, T., and Zettlmeissl, G. (1990) Nucl. Acid Res. 18, 667. [608] Hamamoto, T., Kawasaki, M., Kurosawa, N., Nakaoka, T., Lee, Y.-C. and Tsuji, S. (1993) Bioorg. Med. Chem. 1, 141-145. [609] Kurosawa, N., Kawasaki, M., Hamamoto, T., Nakaoka, T., Lee, Y.-C., Arita, M. and Tsuji, S. (1994) Eur. J. Biochem. 219, 375-381. [610] Kurosawa, N., Hamamoto, T., Lee, Y.-C., Nakaoka, T., Kojima, N. and Tsuji, S. (1994) J. Biol. Chem. 269, 1402-1409.

389 [61 I ] Kurosawa, N., Kojima, N., Inoue, M., Hamamoto, T. and Tsuji, S. (1994) J. Biol. Chem. 269, 1904819053. [612] Soutiropoulou, G., Anisowicz, A. and Sager, R. (1 994) GenBank(TM)/EMBL Data Library, accession number U14550. [612a] Kurosawa, N., Inoue, M., Yoshida, Y. and Tsuji, S. (1996) J. Biol. Chem. 271, 15109-15116. [613] Sasaki, K., Kurata, K., Kojima, N., Kurosawa, N., Ohta, S., Hanai, N., Tsuji, S. and Nishi, T. (1994) J. Biol. Chem. 269, 15950-15956. [614] Nara, K., Watanabe, Y., Maruyama, K., Kasahara, K., Nagai, Y. and Sanai, Y. (1994) Proc. Natl. Acad. Sci. USA 91, 7952-7956. [615] Haraguchi, M., Yamashiro, S., Yamamoto, A,, Furukawa, K., Takamiya, K., Lloyd, K.O., Shiku, H. and Furukawa, K. (1994) Proc. Natl. Acad. Sci. USA 91, 10455-10459. [616] Yoshida, Y., Kojima, N., Kurosawa, N., Hamamoto, T. and Tsuji, S. (1995) J. Biol. Chem. 270, 1462814633. [617] Livingston, B.D. and Paulson, J.C. (1993) J. Biol. Chem. 268, 11504-1 1507. [618] Kojima, N., Yoshida, Y., Kurosawa, N., Lee, Y.-C. and Tsuji, S. (1995) FEBS Lett. 360, 1 4 . [619] Kitagawa, H . and Paulson, J.C. (1994) J. Biol. Chem. 269, 17872-17878. [620] Eckhardt, M., Muhlenhoff, M., Bethe, A,, Koopman, J., Frosch, M. and Gerardy-Schahn, R. (1995) Nature 373, 715-718. [621] Yoshida, Y., Kojima, N. and Tsuji, S. (1995) J. Biochem. 118, 658-664. [621a] Kono, M., Yoshida, Y., Kojima, N. and Tsuji, S. (1996) J. Biol. Chem. 271, 29366-29371. [622] Weisgerber, C., Hansen, A. and Frosch, M. (1991) Glycobiology I , 357-365. [623] Steenbergen, S.M., Wrona, T.J. and Vimr, E.R. (1992) J. Bacteriol. 174, 1099-1 108. [623a] Tsuji, S., Datta, A.K. and Paulson, J.C. (1996) Glycobiology 6, V-VII. [624] Sabesan, S., Paulson, J.C. and Weinstein, J. (1994) Methods Enzymol. 247, 237-243. I6251 Spiegel, L.B., Hadjimichael, J. and Rossomando, E.F. (1992) J. Chromatogr. 573, 23-27. [626] Gamier, F., Baubichon-Cortay, H., Louisot, P. and Gateau-Roesch, 0. (1991) J. Biochem. 110, 702707. [627] Nemansky, M. and van den Eijnden, D.H. (1992) Biochem. J. 287, 31 1-316. [628] Nemansky, M., Schiphorst, W.E.C.M., Koeleman, C.A.M. and van den Eijnden, D.H. (1992) FEBS Lett. 3 12, 3 1-36, [629] Wlasichuk, K.B., Kashem, M.A., Nikrad, P.Y, Bird, P., Jiang, C. and Venot, A.P. (1993) J. Biol. Chem. 268, 13971-13977. [630] Hokke, C.H., van der Ven, J.G.M., Kamerling, J.P. and Vliegenthart, J.F.G. (1993) Glycoconjugate J. 10, 82-90. [630a] Van Dorst, J.A.L.M., Tikkanen, J.M., Krezdorn, C.H., Streift, M.B., Berger, E.G., van Kuik, J.A., Kamerling, J.P. and Vliegenthart, J.F.G. (1996) Eur. J. Biochem. 242, 674-681. [631] Fast, D.G., Jamieson, J.C. and McCaffrey, G. (1993) Biochim. Biophys. Acta 1202, 325-330. [632] Bosshart, H. and Berger, E.G. (1992) Eur. J. Biochem. 208, 341-349. [633] Nemansky, M., Edzes, H.T., Wijnands, R.A. and van den Eijnden, D.H. (1992) Glycobiology 2, 109117. [634] Reboul, P., George, P., Geoffroy, J., Louisot, P. and Broquet, P. (1992) Biochem. Biophys. Res. Commun. 186, 1575-1581. [635] Le Marer, N., Laudet, V., Svensson, E.C., Cazlaris, H., Van Hille, B., Lagrou, C., Stehelin, D., Montreuil, J., Verbert, A. and Delannoy, P. (1992) Glycobiology 2, 49-56. [636] Yogeeswaran, G., Fujinami, R., Kiessling, R. and Welsh, R.M. (1982) Virology 121, 363-371. [637] Kolinski, J., Ivanov, S. and Chelibonova-Lorer, H. (1988) FEBS Lett. 242, 57-60. [638] Vandamme, V., Pierce, A., Verbert, A. and Delannoy, P. (1993) Eur. J. Biochem. 211, 135-140. [639] Hanasaki, K., Varki, A,, Stamenkovic, I. and Bevilacqua, M.P. (1994) J. Biol. Chem. 269, 1063710643. [640] Uehara, F., Ohba, N., Nakashima, Y., Yanagita, T., Ozawa, M. and Muramatsu, T. (1993) Exp. Eye Res. 56, 89-93. [641] Kitazume, S., Kitajima, K., Inoue, S., Inoue, Y. and Troy 11, EA. (1994) J. Biol. Chem. 269, 1033010340.

390 [642] [642a] [643] [644] [645] [646] [647] [648] [649] [650] [651] [652] [653] [654] [655] [656] [657] [658] [659] [660] [661] [662] [663] [664] [665] [666] [667] [668] [669] [670] [671] [672] [673] [674] [675]

[676]

Schmelter, T. (1991) Doctoral thesis, University of Kiel. Kleineidam, R.G., Schmelter, T., Schwarz, R.T. and Schauer, R. (1997) Glycoconjugate J. 14, 5 7 4 6 . Cambron, L.D. and Leskawa, K.C. (1993) Biochem. Biophys. Res. Commun. 193, 585-590. Kajihara, Y., Hashimoto, H., Kodama, H., Wakabayashi, T. and Sato, K.4. (1993) J. Carbohydr. Chem. 12, 991-995. Van den Eijnden, D.H. and Schiphorst, W.E.C.M. (1988) In: R. Schauer and T. Yamakawa (Eds.), Sialic Acids - Proc. Jap.-German Symp. Verlag Wissenschaft und Bildung, Kiel, pp. 128-129. Albarracin, I., Lassaga, F.E. and Caputto, R. (1988) Biochem. J. 254, 559-565. Paulson, J.C., Reanck, J.I. and Hill, R.L. (1977) J. Biol. Chem. 252, 2363-2371. Kijima-Suda, I., Miyamoto, Y., Toyoshima, S., Itoh, M. and Osawa, T. (1986) Cancer Res. 46, 858862. Kijima-Suda, I., Miyazawa, T., Itoh, M., Toyoshima, S. and Osawa, T. (1988) Cancer Res. 48, 37283732. Harvey, B.E. and Thomas, P. (1993) Biochem. Biophys. Res. Commun. 190, 571-575. Berger, E.G., Thurnher, M. and Muller, U. (1987) Exp. Cell Res. 173, 267-273. Roth, J. (1987) Biochim. Biophys. Acta 906, 405436. Gleeson, P.A., Teasdale, R.D. and Burke, J. (1994) Glycoconjugate J. 1 I , 381-394. Lepers, A,, Shaw, L., Schneckenburger, P., Cacan, R., Verbert, A. and Schauer, R. (1990) Eur. J. Biochem. 193, 715-723. Hayes, B.K., Freeze, H.H. and Varki, A. (1993) J. Biol. Chem. 268, 16139-16154. Hayes, B.K. and Varki, A. (1993) J. Biol. Chem. 268, 16155-16169. Dahdal, R.Y. and Colley, K.J. (1993) J. Biol. Chem. 268, 26310-26319. Edwards, U., Muller, A,, Hammerschmidt, S., Gerardy-Schahn, R. and Frosch, M. (1994) Mol. Microbiol. 14, 141-149. Gottschalk, A. (Ed.) (1960) The Chemistry and Biology of Sialic Acids and Related Substances. University Press, Cambridge. Schauer, R., Kelm, S., Reuter, G., Roggentin, P. and Shaw, L. (1995) In: A. Rosenberg (Ed.), Biology of the Sialic Acids. Plenum Publ. Corp. New York, pp. 7-67. Kawai, T., Kato, A,, Higashi, A., Kato, S. and Naiki, M. (1991) Cancer Res. 51, 1242-1246. Schauer, R., Reuter, G., Muhlpfordt, H., Andrade, A.F.B. and Pereira, M.E.A. (1983) Hoppe-Seyler’s Z. Physiol. Chem. 364, 1053-1057. Schauer, R. (1985) Trends Biochem. Sci. 10, 357-360. Kelm, S., Schauer, R., Manuguerra, J.-C., Gross, H.-J. and Crocker, P.R. (1994) Glycoconjugate J. 11, 576-585. Kyogashima, M., Ginsburg, V. and Krivan, H.C. (1989) Arch. Biochem. Biophys. 270, 391-397. Maget-Dana, R., Veh, R.W., Sander, M., Roche, A.-C., Schauer, R. and Monsigny, M. (1981) Eur. J. Biochem. 114, 11-16, Ouadia, A,, Karamanos, Y. and Julien, R. (1992) Glycoconjugate J. 9, 21-26. Schauer, R. (1988) In: A.M. Wu (Ed.), The Molecular Immunology of Complex Carbohydrates, Adv. Exp. Med. Biol. Vol. 228. Plenum Press, New York, pp. 47-72. Noguchi, A,, Mukuria, C.J., Suzuki, E. and Naiki, M. (1995) J. Biochem. 117, 5 9 4 2 . Asaoka, H., Nishinaka, S., Wakamiya, N., Matsuda, H. and Murata, M. (1992) Immunol. Lett. 32, 9 1-96. Gtiot-Wenk, M., Pahlsson, P., Chisholm-Chait, A,, Spitalnik, P.F., Spitalnik, S.L. and Giger, U. (1993) Animal Genet. 24, 401407. Hashimoto, Y., Yamakawa, T. and Tanabe, Y. (1984) J. Biochem. 96, 177771782, Andrews, G.A., Chavey, P.S., Smith, J.E. and Rich, L. (1992) Blood 79, 2485-2491. Muchmore, E.A. (1992) Glycobiology 2, 337-343. Bergwerff, A.A., Hulleman, S.H.D., Kamerling, J.P., Vliegenthart, J.F.G., Shaw, L., Reuter, G. and Schauer, R. (1993) In: J. Roth, U. Rutishauser and F.A. Troy I1 (Eds.), Polysialic Acid. Birkhauser Verlag, Basel, pp. 201-212. Sugita, M. (1979) J. Biochem. 86, 765-772.

39 1 [677] Watarai, S., Kushi, Y., Shigeto, R., Misawa, N., Eishi, Y., Handa, S. and Yasuda, T. (1995) J. Biochem. 117, 1062-1069. [678] Schauer, R., Schoop, H.J. and Faillard, H. (1968) Hoppe-Seyler’s Z. Physiol. Chem. 349, 645452. [679] Schoop, H.J., Schauer, R. and Faillard, H. (1969) Hoppe-Seyler’s Z. Physiol. Chem. 350, 155-162. [680] Schlenzka, W., Shaw, L. and Schauer, R. (1994) Glycobiology 4, 675-683. [681] Schneckenburger, P., Shaw, L. and Schauer, R. (1994) Glycoconjugate J. 11, 194-203. [682] Kawano, T., Kozutsumi, Y., Kawasaki, T. and Suzuki, A. (1994) J. Biol. Chem. 269, 9024-9029. [683] Schlenzka, W., Shaw, L. and Schauer, R. (1993) Biochim. Biophys. Acta 1161, 131-138. [684] Shaw, L. and Schauer, R. (1989) Biochem. J. 263, 355-363. [685] Takematsu, H., Kawano, T., Koyama, S., Kozutsumi, Y., Suzuki, A. and Kawasaki, T. (1994) J. Biochem. 115, 381-386. [686] Schlenzka, W., Shaw, L., Schauer, R., Schmidt, C.L., Bill, E. and Trautwein, A.X. (1994) J. Inorg. Biochem. 56, 29. [687] Kawano, T., Koyama, S., Takematsu, H., Kozutsumi, Y., Kawasaki, H., Kawashima, S., Kawasaki, T. and Suzuki, A. (1995) J. Biol. Chem. 270, 16458-16463. I6881 Schlenzka, W., Shaw, L., Kelm, S., Schmidt, C.L., Bill, E., Trautwein, A.X., Lottspeich, F. and Schauer, R. (1996) FEBS Lett. 385, 197-200. [689] Mukuna, C.J., Mwangi, W.D., Noguchi, A,, Waiyaki, G.P., Asano, T. and Naiki, M. (1995) Biochem. J. 305, 459464. [690] Schauer, R. (1991) Glycobiology 1, 449452. [691] Malykh, Ya.N., Dyatlovitskaya, E.V., Lemenovskaya, A.F., Sorokin, M.M., Indzhiya, L.V., Yakovleva, L.A. and Bergel’son, L.D. (1991) Biochemistry 56, 723-728. [692] Muthing, J., Steuer, H., Peter-KataliniC, J., Marx, U., Bethke, U., Neumann, U. and Lehmann, J. (1994) J. Biochem. 116, 64-73. [693] Hubbard, S.C., Walls, L., Ruley, H.E. and Muchmore, E.A. (1994) J. Biol. Chem. 269, 3717-3724. I6941 Sjoberg, E.R., Manzi, A.E., Khoo, K.-H., Dell, A. and Varki, A. (1992) J. Biol. Chem. 267, 1620016211. [695] Campbell, F,, Appleton, M.A.C., Fuller, C.E., Greeff, M.P., Hallgrimsson, J., Katoh, R., Ng, O.L.I., Satir, A,, Williams, G.T. and Williams, E.D. (1994) J. Pathol. 174, 169-174. [696] Engstler, M., Hubl, U. and Schauer, R. (1995) Glycoconjugate J. 12, 520. [697] Kiehne, K. and Schauer, R. (1992) Biol. Chem. Hoppe-Seyler 373, 11 17-1 123. [698] Pacitti, A.F. and Gentsch, J.R. (1987) J. Virol. 61, 1407-1415. [699] Higa, H.H., Rogers, G.N. and Paulson, J.C. (1985) Virology 144, 279-282. Schauer, R., Howard, R.J., Palese, P. [700] Klotz, EW., Orlandi, P.A., Reuter, G., Cohen, S.J., Haynes, J.D., and Miller, L.H. (1992) Mol. Biochem. Parasitol. 51, 49-54. [701] Varki, A. and Kornfeld, S. (1980) J. Exp. Med. 152, 532-544. [702] Koistinen, V. (1992) Biochem. J. 283, 317-319. [703] Shugaba, A,, Umar, I., Omage, J., Ibrahim, N.D.G., Andrews, J., Ukoha, A.I., Saror, D.I. and Esievo, K.A.N. (1994) J. Comp. Pathol. 110, 91-95. [704] Larkin, M., Knapp, W., Stoll, M.S., Mehmet, H. and Feizi, T. (1991) Clin. Exp. Immunol. 85,536-541. [705] Herrler, G., Rott, R., Klenk, H.-D., Muller, H.-P., Shukla, A.K. and Schauer, R. (1985) EMBO J. 4, 1503-1 506. [706] Schultze, B., Gross, H.-J., Brossmer, R. and Herrler, G. (1991) J. Virol. 65, 6232-6237. [707] Schultze, B., Gross, H.-J., Brossmer, R., Klenk, H.-D. and Herrler, G. (1990) Virus Res. 16, 185-194. [708] Frasa, H., Procee, J., Torensma, R., Verbruggen, A,, Algra, A,, Rozenberg-Arska, M., Kraaijeveld, K. and Verhoef, J. (1993) J. Clin. Microbiol. 3 1, 3 174-3 178. [709] Zeng, F.Y. and Gabius, H.-J. (1992) Z. Naturforsch. 47c, 641-653. [710] Ravindranath, M.H. and Paulson, J.C. (1987) Methods Enzymol. 138, 520-527. [711] Sen, G. and Mandal, C. (1995) Carbohydrate Res. 268, 115-125. [712] Herrler, G., Reuter, G., Rott, R., Klenk, H.-D. and Schauer, R. (1987) Biol. Chem. Hoppe-Seyler 368, 45 1 4 5 4 . [713] Chou, D.K.H., Flores, S., Jungalwala, F.B. (1990) J. Neurochem. 54, 1598-1607. [714] Varki, A,, Hooshmand, F., Diaz, S., Varki, N.M. and Hednck, M. (1991) Cell 65, 65-74.

392 Sen, G., Chowdhury, M. and Mandal, C. (1994) Mol. Cell. Biochem. 136, 65-70. Reivinen, J., Holthofer, H. and Miettinen, A. (1994) Int. Immunol. 6, 1409-1416. Cheresh, D.A., Reisfeld, R.A. and Varki, A.P. (1984) Science 225, 844-846. Paller, AS., Arnsmeier, S.L., Robinson, J.K. and Bremer, E.G. (1992) J. Invest. Dermatol. 98, 226232. I7191 Hutchins, J.T. and Reading, C.L. (1988) J. Cell. Biochem. 37, 3 7 4 8 . [720] Huang, C.-B., Xu, J., Wu, K.-R. and Wang, M.-W. (1988) Anticancer Res. 8, 829-832. [721] Blum, A S . and Barnstable, C.J. (1987) Proc. Natl. Acad. Sci. USA 84, 8716-8720. [722] Jass, J.R. and Roberton, A.M. (1994) Pathol. Intern. 44, 487-504. I7231 Agawa, S. and Jass, J.R. (1990) J. Clin. Pathol. 43, 527-532. 17241 Rahmann, H., Hilbig, R., Marx, J., Beitinger, H. and Mehlfeld, R. (1987) J. Therm. Biol. 12, 81-85. 17251 Suzuki, Y., Nakao, T., Ito, T., Watanabe, N., Toda, Y., Guiyun, X., Suzuki, T., Kobayashi, T., Kimura, Y., Yamada, A., Sugawara, K., Nishimura, H., Kitame, F., Nakamura, K., Deya, E., Kiso, M. and Hasegawa, A. (1992) Virology 189, 121-131. [726] Reuter, G., Schauer, R. and Bumm, I? (1988) In: R. Schauer and T. Yamakawa (Eds.), Sialic Acids, Proc. Jap.-German Symp. Verlag Wissenschaft und Bildung, Kiel, pp. 258-259. [727] Harms, G., Corfield, A.P., Schauer, R. and Reuter, G. (1993) Biol. Chem. Hoppe-Seyler 374,946-947. [728] Herrler, G. and Klenk, H.-D. (1991) Adv. Virus Res. 40, 213-234. [729] Vlasak, R., Krystal, M., Nacht, M. and Palese, I? (1987) Virology 160, 4 1 9 4 2 5 . [730] Harms, G., Reuter, G., Corfield, A.P. and Schauer, R. (1996) Glycoconjugate J. 13, 621-630. [730a] Klein, A,, Krishna, M., Varki, N.M. and Varki, A. (1994) Proc. Natl. Acad. Sci. USA 91, 7782-7786. [731] Schauer, R. (1970) Hoppe-Seyler’s Z. Physiol. Chem. 351, 5 9 5 4 0 2 . [732] Schauer, R. (1970) Hoppe-Seyler’s Z. Physiol. Chem. 351, 749-758. [733] Schauer, R., Casals-Stenzel, J., Corfield, A.P and Veh, R.W. (1988) Glycoconjugate J. 5 , 257-270. [734] Higa, H.H., Butor, C., Diaz, S. and Varki, A. (1989) J. Biol. Chem. 264, 19427-19434. [735] Butor, C., Diaz, S. and Varki, A. (1993) J. Biol. Chem. 268, 10197-10206. [736] Sjoberg, E.R. and Varki, A. (1993) J. Biol. Chem. 268, 10185-10196. [737] Corfield, A.P., Wagner, S.A., Safe, A,, Mountford, R.A., Clamp, J.R., Kamerling, J.P., Vliegenthart, J I G . and Schauer, R. (1993) Clin. Sci. 84, 573-579. [738] Schauer, R. and Wember, M. (1985) In: E.A. Davidson, J.C. Williams and N.M. Di Ferrante (Eds.), Glycoconjugates, Proc. 8th Int. Symp., Vol. I . Praeger, New York, pp. 266-267. [739] De Freese, A,, Shaw, L., Reuter, G. and Schauer, R. (1993) Glycoconjugate J. 10, 330. [740] Inoue, Y.(1993) In: J. Roth, U. Rutishauser and F.A. Troy I1 (Eds.), Polysialic Acid. Birkhauser Verlag, Basel, pp. 171-181. [741] Strecker, G., Wieruszeski, J.-M., Plancke, Y. and Michalski, J.-C. (1993) Glycoconjugate J. 10, 345. [742] Sander, M., Veh, R.W. and Schauer, R. (1979) In: R. Schauer, I? Boer, E. Buddecke, M.F. Kramer, J.F.G. Vliegenthart and H. Wiegandt (Eds.), Glycoconjugates, Proc. 5th Int. Symp. on Glycoconjugates. Thieme, Stuttgart, pp. 358-359. [743] Schauer, R., Reuter, G., Stoll, S., Posadas del Rio, F., Herrler, G. and Klenk, H.-D. (1988) Biol. Chem. Hoppe-Seyler 369, 1121-1 130. [744] Schauer, R., Reuter, G., Stoll, S. and Shukla, A.K. (1989) J. Biochem. 106, 143-150. [745] Butor, C., Higa, H.H. and Varki, A. (1993) J. Biol. Chem. 268, 10207-10213. [746] Higa, H.H., Manzi, A. and Varki, A. (1989) J. Biol. Chem. 264, 19435-19442. [747] Varki, A., Muchmore, E. and Diaz, S. (1986) Proc. Natl. Acad. Sci. USA 83, 882-886. 17481 Corfield, A.P., Wagner, S.A., O’Donnell, L.J.D., Durdey, P., Mountford, R.A. and Clamp, J.R. (1993) Glycoconjugate J. 10, 72-8 1. [749] Garcia-Sastre, A., Villar, E., Manuguerra, J.C., Hannoun, C. and Cabezas, J.A. (1991) Biochem. J. 273, 435441. [750] Herrler, G., Gross, H.J., Imhof, A., Brossmer, R., Milks, G. and Paulson, J.C. (1992) J. Biol. Chem. 267, 12501-12505. [751] Manzi, A.E., Sjoberg, E.R., Diaz, S. and Varki, A. (1990) J. Biol. Chem. 265, 13091-13103. [752] MuAoz-Barroso, I., Garcia-Sastre, A., Villar, E., Manuguerra, J.-C., Hannoun, C. and Cabezas, J.A. (1992) Virus Res. 25, 145-153. [715] [716] [717] [718]

393 [753] Rosenberg, A. and Schengrund, C.-L. ( I 976) In: A. Rosenberg and C.-L. Schengrund (Eds.), Biological Roles of Sialic Acid. Plenum Press, New York, pp. 295-359. [754] Miiller, H.E. (1974) Behring Inst. Mitt. 55, 34-56. [755] Engstler, M., Reuter, G. and Schauer, R. (1993) Mol. Biochem. Parasitol. 61, 1-14. [756] Colman, P.M. (1994) Prot. Sci. 3, 1687-1696. [757] Long, G.S., Taylor, PW. and Luzio, J.P. (1993) In: J. Roth, U. Rutishauser and F.A. Troy II (Eds.), Polysialic Acid. Birkhauser Verlag, Basel, pp. 137-144. [758] Deijl, C.M., Kamerling, J.P. and Vliegenthart, J.F.G. (1984) Carbohydr. Res. 126, 338-342. [759] Guo, X. and Sinnott, M.L. (1993) Biochem. J. 296, 291-292. [760] Guo, X., Laver, W.G., Vimr, E. and Sinnott, M.L. (1994) J. Am. Chem. SOC.116, 5572-5578. [761] Bossart-Whitaker, P., Carson, M., Babu, Y.S., Smith, C.D., Laver, W.G. and Air, G.M. (1993) J. Mol. Biol. 232, 1069-1083. [762] Janakiraman, M.N., White, C.L., Laver, W.G., Air, G.M. and Luo, M. (1994) Biochemistry 33, 81728179. [763] Schauer, R., Shukla, A.K., Schroder, C. and Miiller, E. (1984) Pure Appl. Chem. 56, 907-921. [764] Kwiatkowski, B., Boscheck, B., Thiele, H. and Stirm, S. (1982) J. Virol. 43, 697-704. [765] Rougon, G. (1993) Eur. J. Cell Biol. 61, 197-207. [766] Li, Y.-T., Yuziuk, J.A., Li, S.-C., Nematalla, A,, Hasegawa, A,, Tsutsumi, M. and Nakagawa, H. (1993) Glycobiology 10, 525. [767] Kean, E.L. and Bighouse, K.J. (1974) J. Biol. Chem. 249, 7813-7823. [768] Vimr, E.R. (1994) Trends Microbiol. 2, 271-277. [769] Roggentin, P., Rothe, B., Lottspeich, F. and Schauer, R. (1988) FEBS Lett. 238, 31-34. [769a] Chien, C.H., Shann, Y.J. and Shen, S.Y. (1996) Enzyme Microbiol. Technol. 19, 267-276. [770] Galen, J.E., Ketley, J.M, Fasano, A,, Richardson, S.H., Wasserman, S.S. and Kaper, J.B. (1992) Infect. Immun. 60, 4 0 6 4 1 5 . [771] Rothe, B., Roggentin, P., Frank, R., Blocker, H. and Schauer, R. (1989) J. Genet. Microbiol. 135, 3087-3096. [772] Hoyer, L.L., Hamilton, A.C., Steenbergen, S.M. and Vimr, E.R. (1992) Mol. Microbiol. 6, 873-884. [773] Roggentin, P., Rothe, B., Kaper, J.B., Galen, J., Lawrisuk, L., Vimr, E.R. and Schauer, R. (1989) Glycoconjugate J. 6, 349-353. [774] Miyagi, T., Konno, K., Emori, Y., Kawasaki, H., Suzuki, K., Yasui, A. and Tsuiki, S. (1993) J. Biol. Chem. 268, 26435-26440. [775] Ferrari, J., Harris, R. and Warner, T.G. (1994) Glycobiology 4, 367-373. [776] Pereira, M.E.A., Mejia, J.S., Ortega-Barria, E., Matzilevitch, D. and Prioli, R.P (1991) J. Exp. Med. 174, 179-191. [777] Buschiazzo, A,, Cremona, M.L., Campetella, O., Frasch, A.C.C. and Sanchez, D.D. (1993) Mol. Biochem. Parasitol. 62, 115-1 16. [778] Henningsen, M., Roggentin, P. and Schauer, R. (1991) Biol. Chem. Hoppe-Seyler 372, 1065-1072. [779] Yeung, M.K. (1993) Infect. Immun. 61, 109-116. [780] Akimoto, S., Ono, T., Tsutsui, H., Kinouchi, T., Kataoka, K. and Ohnishi, Y. (1994) Biochem. Biophys. Res. Commun. 203, 914-921. [781] Rothe, B., Rothe, B., Roggentin, P and Schauer, R. (1991) Mol. Genet. 226, 190-197. [782] Traving, C., Schauer, R. and Roggentin, P. (1994) Glycoconjugate J. I I , 141-151. 17831 Sakurada, K., Ohta, T. and Hasegawa, M. (1992) J. Bacteriol. 174, 6896-6903. [784] Camara, M., Boulnois, G.J., Andrew, P.W. and Mitchell, T.J. (1994) Infect. Immun. 62, 3688-3695. [785] Galen, J.E., Vimr, E.R., Lawrisuk, L. and Kaper, J.B. (1990) In: R.B. Sack and Y. Zinnaka (Eds.), Advances in Research on Cholera and Related Diarrheas. Vol. 7. KTK Scientific Publishers, Tokyo, pp. 143-153. [786] Air, G.M., Webster, R.G., Colman, P.M. and Laver, W.G. (1987) Virology 160, 346-354. [787] Bryant, J.M., Luzio, J.P. and Taylor, P. (1995) Glycoconjugate J. 12, 509. [788] Prioli, R.P., Ortega-Barria, E., Mejia, J.S. and Pereira, M.E.A. (1992) Mol. Biochem. Parasitol. 52, 85-96. [789] Roggentin, T., Kleineidam, R.G., Schauer, R. and Roggentin, P. (1992) Glycoconjugate J. 9, 235-240.

394 [790] Crennell, S.J., Garman, E.F., Laver, W.G., Vimr, E.R. and Taylor, G.L. (1993) Proc. Natl. Acad. Sci. USA 90, 9852-9856. [791] Crennell, S., Garman, E., Laver, G., Vimr, E. and Taylor, G. (1994) Structure 2, 535-544. [792] Varghese, J.N., Laver, W.G. and Colman, P.M. (1983) Nature 303, 3 5 4 0 . 17931 Colman, P.M., Varghese, J.N. and Laver, W.G. (1983) Nature 303, 41-44. [794] Varghese, J.N., McKimm-Breschkin, J.L., Caldwell, J.B., Kortt, A.A. and Colman, P.M. (1992) Proteins 14, 327-332. [795] Corfield, A.P., Lambrt, C.R., Michalski, J.-C. and Schauer, R. (1992) Conferences Philippe Laudat 1991. INSERM, Paris, pp. 11 1-175. [796] Simpson, N.E. and Cann, H.M. (1991) Cytogenet. Cell Genet. 58, 428458. [797] Schauer, R. and Vliegenthart, J.F.G. (1982) In: R. Schauer (Ed.), Sialic Acids - Chemistry, Metabolism and Function. Springer Verlag, Wien, New York, pp. 1-3. [798] Schauer, R. and Wember, M. (1989) Biol. Chem. Hoppe-Seyler 370, 183-190. [799] Roggentin, P., Schauer, R., Hoyer, L.L. and Vimr, E.R. (1993) Mol. Microbiol. 9, 915-921. [800] Canard, B., and Cole, S.T. (1990) FEMS Lett. 66, 323-326. [801] Pereira, M.E.A. (1983) Science 219, 144441446, [802] Prioli, R.P., Mejia, J.S., Aji, T., Aikawa, M. and Pereira, M.E.A. (1991) Trop. Med. Parasitol. 42, 146-1 50. 18031 Previato, J.O., Andrade, A.F.B., Pessolani, M.C.V. and MendonCa-Previato, L. (1985) Mol. Biochem. Parasitol. 16, 85-96. [804] Zingales, B., Carniol, C., de Lederkremer, R. and Colli, W. (1987) Mol. Biochem. Parasitol. 26, 135144. [805] Schenkman, S., Jiang, M.-S., Hart, G.W. and Nussenmeig, V. (1991) Cell 65, 11 17-1 125. [806] Engstler, M., Schauer, R. and Brun, R. (1995) Acta Tropica 59, 117-129. [807] Medina-Acosta, E., Paul, S., Tomlinson, S. and Pontes-de-Carvalho, L.C. (1994) Mol. Biochem. Parasitol. 64, 273-282. [808] Esievo, K.A.N. (1979) Proc. 16th Meeting of OAU/STRC International Council for Trypanosomiasis Research and Control, Yaunde, Cameroun, pp. 205-21 0. [SO91 Reuter, G., Schauer, R., Prioli, R. and Pereira, M.E.A. (1987) Glycoconjugate J. 4, 339-348. [810] Engstler, M. and Schauer, R. (1993) Parasitol. Today 9, 222-225. [81 I] Medina-Acosta, E., Franco, A.M.R., Jansen, A.M., Sampol, M., Nevts, N., Pontes-de-Carvalho, L., Grimaldi Jr., G . and Nussenzweig, V. (1994) Eur. J. Biochem. 225, 333-339. [8 121 Pontes de Carvalho, L.C., Tomlinson, S., Vandekerckhove, E, Bienen, E.J., Clarkson, A.B., Jiang, M.-S., Hart, G.W. and Nussenzweig, V. (1993) J. Exp. Med. 177, 4 6 5 4 7 4 . [813] Engstler, M., Ferrero-Garcia, M.A., Parodi, A.J., Schauer, R., Storz-Eckerlin, T., Vasella, A,, Witzig, C. and Zhu, X. (1994) Helv. Chim. Acta 77, 1166-1 174. [814] Parodi, A.J., Pollevick, G.D., Mautner, M., Buschiazzo, A., Sanchez, D.O. and Frasch, A.C.C. (1992) EMBO J. 11, 1705-1710. [815] Cross, G.A.M. and Takle, G.B. (1993) Annu. Rev. Microbiol. 47, 3 8 5 4 1 I . [816] Uemura, H., Schenkman, S., Nussenzweig, V. and Eichinger, D. (1992) EMBO J. 11, 3837-3844. [817] Ruef, B.J., Dawson, B.D., Tewari, D., Fouts, D.L. and Manning, J.E. (1994) Mol. Biochem. Parasitol. 63, 109-120. [818] Colli, W. (1993) FASEB J. 7, 125771264, [819] Cremona, M.L., Sanchez, D.O., Frasch, A.C.C. and Campetella, 0. (1995) Gene 160, 123-128. [820] Schenkman, S., Chaves, L.B., Pontes de Carvalho, L.C. and Eichinger, D. (1994) J. Biol. Chem. 269, 7970-7975. [821] Campetella, O.E., Uttaro, A.D., Parodi, A.J. and Frasch, A.C.C. (1994) Mol. Biochem. Parasitol. 64, 337-340. I8221 Pontes-de-Carvalho, L.C., Tomlinson, S. and Nussenzweig, V. (1993) Mol. Biochem. Parasitol. 62, 19-26. [823] Leguizamon, M.S., Campetella, O.E., Cappa, S.M.G. and Frasch, A.C.C. (1994) Infect. Immun. 62, 344 1-3446.

395 [824] Leguizamon, M.S., Campetella, O., Russomando, G., Almiron, M., Guillen, I., Cappa, S.M.G. and Frasch, A.C.C. (1994) J. Infect. Dis. 170, 1570-1574. [825] Schenkman, S., Eichinger, D., Pereira, M.E.A. and Nussenzweig, V. (1994) Annu. Rev. Microbiol. 48, 499-523. [826] Frasch, A.C.C. (1994) Parasitology 108, S37-S44. [827] Shimmoto, M., Fukuhara, Y., Itoh, K., Oshima, A,, Sakuraba, H. and Suzuki, Y. (1993) J. Clin. Invest. 91, 2393-2398. [828] Hakomori, S.-I. (1994) In: K. Bock and H. Clausen (Eds.), Complex Carbohydrates in Drug Research. Munksgaard, Copenhagen, pp. 337-349. [829] Saito, M. and Yu, R.K. (1993) J. Neurosci. Res. 36, 127-132. [830] Previato, J.O., Jones, C., Goncalves, L.P.B., Wait, R., Travassos, L.R. and MendonGa-Previato, L. (1994) Biochem. J. 301, 151-159. [831] Schenkman, S. and Eichinger, D. (1993) Parasitol. Today 9, 218-222. [832] Do Carmo Ciavaglia, M., de Carvalho, T.U. and de Souza, W. (1993) Biochem. Biophys. Res. Commun. 193, 718-721. [833] Tomlinson, S., Pontes de Carvalho, L.C., Vandekerckhove, F. and Nussenzweig, V (1994) J. Immunol. 153, 3141-3147. [834] Prioli, R.P., Mejia, J.S. and Pereira, M.E.A. (1991) Eur. J. Epidemiol. 7, 344-348. [835] Chuenkova, M. and Pereira, M.E.A. (1995) J. Exp. Med. 181, 1693-1703. [836] Wessels, M.R., Rubens, C.E., Benedi, V-J. and Kasper, D.L. (1989) Proc. Natl. Acad. Sci. USA 86, 8983-8987. [837] Frangipane, J.V. and Rest, R.F. (1993) Infect. Immun. 61, 1657-1666. [838] Schenkman, S., Kurosaki, T., Ravetch, J.V and Nussenzweig, V (1992) J. Exp. Med. 175, 1635-1641. [839] Araya, J.E., Cano, M.I., Yoshida, N. and da Silveira, J.F. (1994) Mol. Biochem. Parasitol. 65, 161-169. [840] Reyes, M.B., Pollevick, G.D. and Frasch, A.C.C. (1994) Gene 140, 139-140. [841] Kahn, S., Kahn, M., Van Voorhis, W.C., Goshom, A,, Strand, A,, Hoagland, N., Eisen, H. and Pennathur, S. (1993) Mol. Biochem. Parasitol. 60, 149-152. [842] De Araujo-Jorge, T.C., Barbosa, H.S. and Meirelles, M.N.L. (1992) Mem. Inst. Oswaldo Cruz, Rio de Janeiro 87, Suppl. V, 43-56. [843] Ramirez, M.I., de Cassia Ruiz, R., Araya, J.E., da Silveira, J.F. and Yoshida, N. (1993) Infect. Immun. 61, 3636-3641. [844] Fenton Hall, B., Webster, P., Ma, A.K., Joiner, K.A. and Andrews, N.W. (1992) J. Exp. Med. 176, 3 13-325. [845] Fenton Hall, B. and Joiner, K.A. (1993) J. Euk. Microbiol. 40, 207-213. [846] Andrews, N.W. (1994) Braz. J. Med. Biol. Res. 27, 471475. [847] Vermelho, A.B. and Meirelles, M.N.L. (1994) Mem. Inst. Oswaldo Cruz, Rio de Janeiro 89, 69--79. [848] Pellegrin, J.-L.J., Ortega-Barria, E., Barza, M., Baum, J. and Pereira, M.E.A. (1991) Invest. Ophthalmol. Visual Sci. 32, 3061-3066. [849] Godoy, V.G., Miller Dallas, M., Russo, T.A. and Malamy, M.H. (1993) Infect. Immun. 61,441 5 4 4 2 6 . [850] Nakato, H., Shinomiya, K. and Mikawa, H. (1986) Pathol. Res. Pract. 181, 31 1-319. [851] Roggentin, T., Kleineidam, R.G., Majewski, D.M., Tirpitz, D., Roggentin, P. and Schauer, R. (1993) J. Immunol. Methods 157, 125-133. [852] Roggentin, T., Roggentin, P., Kleineidam, R.G., Majewski, D.M. and Schauer, R. (1992) Trends Glycosci. Glycotechnol. 4, 263-268. [853] Schauer, R. (1992) Nachr. Chem. Tech. Lab. 40, 1227-1231. [854] Corfield, A.P., Wagner, S.A., Clamp, J.R., Kriaris, M.S. and Hoskins, L.C. (1992) Infect. Immun. 60, 3971-3978. [855] Boureau, H., Decre, D., Carlier, J.P., Guichet, C. and Bourlioux, P. (1993) Res. Microbiol. 144, 405410. 18561 McGregor, J.A., French, J.I., Jones, W., Milligan, K., McKinney, P.J., Patterson, E. and Parker, R. (1994) Am. J. Obstet. Gynecol. 170, 1048-1060. [857] Homer, K.A., Whiley, R.A. and Beighton, D. (1994) J. Med. Microbiol. 41, 184-190.

396 [858] Silva Filho, EC., Breier-Saraiva, E.M., Tosta, M.X. and de Souza, W. (1989) Mol. Biochem. Parasitol. 35, 73-78. [859] Paton, J.C., Andrew, P.W., Boulnois, G.J. and Mitchell, T.J. (1993) Annu. Rev. Microbiol. 47, 89-1 15. [860] Boulnois, G.J. (1992) J. Genet. Microbiol. 138, 249-259. [861] Straus, D.C., Unbehagen, P.J. and Purdy, C.W. (1993) Infect. Immun. 61, 253-259. [862] Nuss, J.M. and Air, G.M. (1994) J. Mol. Biol. 235, 747-759. [863] Herrler, G., Hausmann, J. and Klenk, H.-D. (1995) In: A. Rosenberg (Ed.), Biology of the Sialic Acids. Plenum Publ. Corp., New York, pp. 315-336. [864] Hu, X., Ray, R. and Compans, R.W. (1992) J. Virol. 66, 1528-1534. [865] Schauer, R. and Corfield, A.P. (1981) In: F.G. de las Heras and S. Vega (Eds.), Medicinal Chemistry Advances. Pergamon Press, Oxford, New York, pp. 423434. [866] Meindl, P., Bodo, G., Palese, P., Schulman, J. and Tuppy, H . (1974) Virology 58, 4 5 7 4 6 3 . [867] Von Itzstein, M., Wu, W.-Y., Kok, G.B., Pegg, M.S., Dyason, J.C., Jin, B., Van Phan, T., Smythe, M.L., White, H.F., Oliver, S.W., Colman, P.M., Varghese, J.N., Ryan, D.M., Woods, J.M., Bethell, R.C., Hotham, VJ., Cameron, J.M. and Penn, C.R. (1993) Nature 363, 4 1 8 4 2 3 . [868] Woods, J.M., Bethell, R.C., Coates, J.A.V, Healy, N., Hiscox, S.A., Pearson, B.A., Ryan, D.M., Ticehurst, J., Tilling, J., Walcott, S.M. and Penn, C.R. (1993) Antimicrob. Agents Chemother. 37, 1473-1479. [869] Zbiral, E., Schreiner, E., Christian, R., Kleineidam, R.G. and Schauer, R. (1989) Liebigs Ann. Chem. 159-165. [870] Zbiral, E., Schreiner, E., Salunkhe, M.M., Schulz, G., Kleineidam, R.G. and Schauer, R. (1989) Liebigs Ann. Chem. 519-526. [871] Suzuki, Y., Sato, K., Kiso, M. and Hasegawa, A. (1990) Glycoconjugate J. 7, 349-356. [872] Umezawa, H., Aoyagi, T., Komiyama, H., Morishima, H., Hamada, M. and Takeuchi, T. (1974) J. Antibiot. 27, 963-969. [873] Kudo, T., Nishimura, Y., Kondo, S. and Takeuchi, T. (1993) J. Antibiot. 46, 300-309. [874] Nishimura, Y., Umezawa, Y., Kondo, S., Takeuchi, T., Mori, K., Kijima-Suda, I., Tomita, K., Sugawara, K. and Nakamura, K. (1993) J. Antibiot. 46, 1883-1889. [875] Nagai, T., Miyaichi, Y., Tomimori, T., Suzuki, Y. and Yamada, H. (1992) Antiviral Res. 19, 207-217. [876] Messer, M. (1974) Biochem. J. 139, 415420. [877] Taylor, N.R. and von Itzstein, M. (1994) J. Med. Chem. 37, 6 1 6 4 2 4 . [878] Pegg, M.S. and von Itzstein, M. (1994) Biochem. Mol. Biol. Int. 32, 851-858. [879] Corfield, T. (1993) Glycobiology 3, 4 1 3 4 2 2 . [880] Taylor, G. (1993) Nature 363, 401402. [881] Taylor, G., Vimr, E., Garman, E. and Laver, W.G. (1992) J. Mol. Biol. 226, 1287-1290. [882] Ryan, D.M., Ticehurst, J., Dempsey, M.H. and Penn, C.R. (1994) Antimicrob. Agents Chemother. 38, 2270-2275. [883] Hayden, F.G., Rollins, B.S. and Madren, L.K. (1994) Antiviral Res. 25, 123-131. [884] Thomas, G.P., Forsyth, M., Penn, C.R. and McCauley, J.W. (1994) Antiviral Res. 24, 351-356. [885] Nagai, T., Miyaichi, Y., Tomimori, T., Suzuki, Y. and Yamada, H. (1990) Chem. Pharm. Bull. 38, 1329-1332. [886] Sabesan, S., Neira, S., Davidson, F., Duus, J . 0 . and Bock, K. (1994) J. Am. Chem. SOC. 116, 16161634. [887] Driguez, P.-A,, Barrere, B., Chantegrel, B., Deshayes, C., Doutheau, A. and Quash, G. (1992) Bioorg. Med. Chem. Lett. 2, 1361-1366. [888] Yamamoto, T., Kumazawa, H., Inami, K., Teshima, T. and Shiba, T. (1992) Tetrahedron Lett. 33, 57915794. [889] Sabesan, S. (1994) Bioorg. Med. Chem. Lett. 4, 2457-2460. [890] Smythe, M.L. and von Itzstein, M. (1994) J. Am. Chem. SOC.116, 2725-2733. [891] Prioli, R.P., Rosenberg, I. and Pereira, M.E.A. (1990) Mol. Biochem. Parasitol. 38, 191-198. [891a] Schauer, R. and Wember, M. (1996) Biol. Chem. Hoppe-Seyler 377, 293-299. [892] Zbiral, E., Kleineidam, R.G., Schreiner, E., Hartmann, M., Christian, R. and Schauer, R. (1992) Biochem. J. 282, 51 1-516.

397 [893] Schauer, R., Stoll, S., Zbiral, E., Schreiner, E., Brandstetter, H.H., Vasella, A. and Baumberger, F, (1987) Glycoconjugate J. 4, 361-369. [894] Uchida, Y., Tsukada, Y. and Sugimori, T. (1 984) J. Biochem. 96, 507-522. [895] Aisaka, K. and Uwajima, T. (1986) Appl. Environ. Microbiol. 51, 562-565. [896] Kawakami, B., Kudo, T., Narahashi, Y.and Horikoshi, K. (1986) Agric. Biol. Chem. 50, 21 55-2 158. [897] Ohta, Y., Shimosaka, M., Murata, K., Tsukada, Y. and Kimura, A. (1986) Appl. Microbiol. Biotechnol. 24, 386-391. [898] Aisaka, K., Igarashi, A,, Yamaguchi, K. and Uwajima, T. (1991) Biochem. J. 276, 541-546. [899] Izard, T., Lawrence, M.C., Malby, R.L., Lilley, G.G. and Colman, P.M. (1994) Structure 2, 361-369. [900] Horiuchi, T. and Kurokawa, T. (1989) Clin. Chim. Acta 182, 117-122. [901] Hommes, F.A. (1991) In: EA. Hommes (Ed.), Techniques in Diagnostic Human Biochemica Genetics: A Laboratory Manual. Wiley-Liss, New York, pp. 233-237. [902] Nohle, U., Beau, J.-M. and Schauer, R. (1982) Eur. J. Biochem. 126, 543-548. [903] Nees, S. and Schauer, R. (1974) Behring Inst. Mitt. 55, 68-78. [904] Vimr, E.R. and Troy, F.A. (1985) J. Bacteriol. 164, 845-853. [905] Burnet, EM., McCrea, J.F. and Stone, J.D. (1946) In: A. Gottschalk (Ed.), The Chemistry and Biology of Sialic Acids and Related Substances. University Press, Cambridge, pp. 98-1 05. I9061 Faillard, H. and Schauer, R. (1972) In: A. Gottschalk (Ed.), Glycoproteins, Their Composition, Structure and Function. Elsevier, Amsterdam, pp. 1246-1267. [907] Varki, A. (1993) Glycobiology 3, 97-130. [908] Reutter, W., Kottgen, E., Bauer, C. and Gerok, W. (1982) In: R. Schauer (Ed.), Sialic Acids - Chemistry, Metabolism and Function. Springer, Wien, pp. 263-305. [909] Rosenberg, A. and Schengrund, C.-L. (Eds.) (1976) Biological Roles of Sialic Acid, Plenum Press, New York. I9101 Shimamura, M., Shibuya, N., Ito, M. and Yamagata, T. (1994) Biochem. Mol. Biol. Int. 33, 871-878. [91 I ] Bohler, T. and Linderkamp, 0. (1993) Clin. Hemorheol. 13, 775-778. [912] Izumida, Y., Seiyama, A. and Maeda, N. (1991) Biochim. Biophys. Acta 1067, 221-226. [913] Rogers, M.E., Williams, D.T., Niththyananthan, R., Rampling, M.W., Heslop, K.E. and Johnston, D.G. (1992) Clin. Sci. 82, 309-3 13. [914] Litvinov, S.V. and Hilkens, J. (1993) J. Biol. Chem. 268, 21364-21371. [915] Wibawan. I.W.T. and Lammler, C. (1991) J. Genet. Microbiol. 137, 2721-2725. [916] Rahmann, H. (1992) In: N.N. Osborne (Ed.), Current Aspects of the Neurosciences. The Macmillan Press Ltd., pp. 87-125. [9 171 Chapman, S.A., Bonshek, R.E., Stoddart, R.W., Mackenzie, K.R. and McLeod, D. ( 1 994) Br. J. Ophthalmol. 78, 632-637. [918] Aubin, Y. and Prestegard, J.H. (1993) Biochemistry 32, 3422-3428. [919] Jaenicke, R. (1993) Philos. Trans. R. SOC.London B 339, 287-295. [920] Fukuda, M. (1991) J. Biol. Chem. 266, 21327-21330. [921] Klenk, H.-D. and Garten, W. (1994) Trends Microbiol. 2, 3 9 4 3 . [922] Yamamotogoshima, F. and Maeno, K. (1994) Microbiol. Immunol. 38, 819-822. [923] Corfield, A. (1993) In: B. Renner and G. Sawatzki (Eds.), New Perspectives in Infant Nutrition. Georg Thieme Verlag, Stuttgart, New York, pp. 57-65. [924] Verma, M. and Davidson, E.A. (1994) Glycoconjugate J. 1 1 , 172-179. [925] Schauer, R. (1992) In: H. Ogura, A . Hasegawa and T. Suami (Eds.), Carbohydrates - Synthetic Methods and Applications in Medicinal Chemistry. Kodansha, Tokyo, pp. 340-354. [926] Alviano, C.S., Pereira, M.E.A., Souza, W., Oda, L.M. and Travassos, L.R. (1982) FEMS Microbiol. Lett. 15, 223-227. [927] Soares, R.M.A., Alviano, C.S., Angluster, J. and Travassos, L.R. (1993) FEMS Microbiol. Lett. 108, 3 1-34. [928] Souza, E.T., Silva-Filho, F.C., de Souza, W., Alviano, C.S., Angluster, J. and Travassos, L.R. (1986) J. Med. Vet. Mycol. 24, 145-153. [929] Oda, L.M., Kubelka, C.F., Alviano, C.S. and Travassos, L.R. (1983) Infect. Immun. 39, 497-504. [930] Sharon, N. (1994) Struct. Biol. I , 843-845.

398 [931] Hughes, R.C. (1994) Glycobiology 4, 5-12. [932] Drickamer, K. (1994) In: M. Fukuda and 0. Hindsgaul (Eds.), Molecular Glycobiology. IRL Press, Oxford, pp. 53-87. [933] Geffen, I. and Spiess, M. (1992) Int. Rev. Cytol. 137B, 181-219. [934] Liener, I.E., Sharon, N., Goldstein, I.J. (Eds.) (1986) The Lectins. Academic Press, Orlando. [935] Miiller, W.E.G., Schroder, H.C., Schacke, H., Miiller, I.M. and Gamulin, V. (1993/94) Endocytobiosis and Cell Res. 10, 185-204. 19361 Gamulin, V, Rinkevich, B., Schacke, H., Kruse, M., Miiller, I.M. and Miiller, W.E.G. (1994) Biol. Chem. Hoppe-Seyler 375, 583-588. [937] Ashwell, G. and Morell, A.G. (1974) Adv. Enzymol. 41, 99-128. [938] Lodish, H.F., Kong, N. and Wikstrom, L. (1992) J. Biol. Chem. 267, 12753-12760. [939] Treichel, U., Meyer zum Biischenfelde, K.-H., Stockert, R.J., Poralla, T. and Gerken, G. (1994) J. Genet. Virol. 75, 3021-3029. [940] Thotakura, N.R., Szkudlinski, M.W. and Weintraub, B.D. (1994) Glycobiology 4, 525-533. [941] Martinuk, S.D., Manning, A.W., Black, W.D. and Murphy, B.D. (1991) Biol. Reproduct. 45, 598-604. [942] Jenkins, N. and Curling, E.M.A. (1994) Enzyme Microbiol. Technol. 16, 354-364. [943] Henkin, J., Dudlak, D., Beebe, D.P. and Sennello, L. (1991) Thrombosis Res. 63, 215-225. [944] Dong, Y.J., Kung, C. and Goldwasser, E. (1992) J. Cell. Biochem. 48, 269-276. [945] Matteri, R.L., Warikoo, P.K. and Bavister, B.D. (1992) Am. J. Primatol. 27, 205-213. [946] Lascols, O., Cherqui, G., Munier, A,, Picard, J. and Capeau, J. (1994) Cell. Mol. Biol. 40, 359-371. [947] Amano, J. and Kobata, A. (1993) Arch. Biochem. Biophys. 305, 618421. [948] Miiller, E., Schroder, C., Sharon, N. and Schauer, R. (1983) Hoppe-Seyler’s Z. Physiol. Chem. 364, 14 10-1 420. [949] Fischer, C., Kelm, S., Ruch, B. and Schauer, R. (1991) Carbohydr. Res. 213, 263-273. [950] Schauer, R. (1992) In: E. Windler and H. Greten (Eds.), Hepatic Endocytosis of Lipids and Proteins. W. Zuckschwerdt Verlag, Munchen, pp. 127-136. [951] Sato, S. and Hughes, R.C. (1994) J. Biol. Chem. 269, 44244430. [952] Kelm, S. and Schauer, R. (1988) Biol. Chem. Hoppe-Seyler 369, 693-704. [953] Bratosin, D., Mazurier, J., Debray, H., Lecocq, M., Boilly, B., Alonso, C., Moisei, M., Motas, C. and Montreuil, J. (1995) Glycoconjugate J. 12, 258-267. [954] Rolfes-Curl, A,, Ogden, L.L., Omann, G.M. and Aminoff, D. (1991) Exp. Gerontol. 26, 327-345. [955] Stenberg, P.E., Levin, J., Baker, G., Mok, Y. and Corash, L. (1991) J. Cell Physiol. 147, 7-16. [956] Savill, J., Fadok, V, Henson, P. and Haslett, C. (1993) Immunol. Today 14, 131-136. [957] Dini, L., Autuori, F., Lentini, A,, Oliverio, S. and Piacentini, M. (1992) FEBS Lett. 296, 174-178. [958] Froman, D.P. and Thursam, K.A. (1994) Biol. Reproduct. 50, 1094-1099. I9591 Lassalle, B. and Testart, J. (1994) J. Reproduct. Fertil. 101, 703-711. [960] Calzada, L., Salazar, E.L. and Pedron, N. (1994) Arch. Androl. 33, 87-92. [961] Kelm, S. and Schauer, R. (1986) Biol. Chem. Hoppe-Seyler 367, 989-998. [962] Ruch, B., Kelm, S. and Schauer, R. (1988) In: R. Schauer and T. Yamakawa (Eds.), Sialic Acids 1988, Proc. Jap.-German Symp. Sialic Acids. Verlag Wissenschaft und Bildung, Kiel, pp, 268-269. [963] Kuiper, J., Bakkeren, H.F., Biessen, E.A.L. and van Berkel, T.J.C. (1994) Biochem. J. 299, 285-290. [964] Kelm, S., Shukla, A.K., Paulson, J.C. and Schauer, R. (1986) Carbohydr. Res. 149, 59-64. [965] Reuter, G., Kelm, S. and Schauer, R. (1988) Acta Histochem. Suppl. XXXVI, 51-79. [966] Okuda, K., Ono, M. and Kato, T.(1989) Infect. Immun. 57, 1635-1637. [967] Sheth, H.B., Lee, K.K., Wong, W.Y., Srivastava, G., Hindsgaul, O., Hodges, R.S., Paranchych, W. and Irvin, R.T. (1994) Mol. Microbiol. 11, 715-723. [968] Devaraj, N., Sheykhnazari, M., Warren, W.S. and Bhavanandan, V.P. (1994) Glycobiology 4,307-3 16. [969] Ourth, D.D. and Bachinski, L.M. (1987) Dev. Comp. Immunol. 1 1, 55 1-564. [970] Smith, H. (1991) Proc. R. SOC.London B 246, 97-105. [971] Jarvis, G.A. (1995) Trends Microbiol. 3, 198-201. [972] Elkins, C., Carbonetti, N.H., Varela, VA., Stirewalt, D., Klapper, D.G. and Sparling, P.F. (1992) Mol. Microbiol. 6, 2617-2628. [973] Rest, R.F. and Frangipane, J.V (1992) Infect. Immun. 60, 989-997.

399 [974] Mandrell, R.E., Smith, H., Jarvis, G.A., Griffiss, J.M. and Cole, J.A. (1993) Microb. Pathogen. 14, 307-313. [975] Nahori, M.-A,, Etievant, M., Terzidis-Trabelsi, H., David, B. and Lambre, C.R. (1994) Immunol. Lett. 42, 67-73. [976] Haddad, E.-B. and Gies, J.-P. (1992) Eur. J. Pharmacol. 21 1, 273-276. [977] Paulson, J.C. (1985) In: M. Conn (Ed.), The Receptors, Vol. 2. Academic Press, New York, pp. 131219. [978] Rogers, G.N., Herrler, G., Paulson, J.C., and Klenk, H.-D. (1986) J. Biol. Chem. 261, 5947-5951. [979] Vlasak, R., Luytjes, W., Spaan, W. and Palese, P. (1988) Proc. Natl. Acad. Sci. USA 85, 45264529. [980] Markwell, M.A.K. and Paulson, J.C. (1980) Proc. Natl. Acad. Sci. USA 77, 5693-5697. [981] Cahan, L.D. and Paulson, J.C. (1980) Virology 103, 505-509. [982] Svensson, L. (1992) J. Virol. 66, 5582-5585. [983] Loomes, L.M., Uemura, K . 4 , Childs, R.A., Paulson, G.N., Rogers, G.N., Scudder, P.R., Michalski, J.-C., Hounsell, E.F., Taylor-Robinson, D. and Feizi, T. (1984) Nature 307, 560-563. [984] Murray, P.A., Levine, M.J., Tabak, L.A. and Reddy, M.S. (1982) Biochem. Biophys. Res. Commun. 106, 390-396. [985] Teneberg, S., Willemsen, P.T.J., de Graaf, F.K., Stenhagen, G., Pimlott, W., Jovall, P.-A,, Angstrom, J. and Karlsson, K.-A. (1994) J. Biochem. 116, 560-574. [986] Parkkinen, A,, Rogers, G.N., Korhonen, T., Dahr, W. and Finne, J. (1986) Infect. Immun. 54, 3 7 4 2 . [987] Hanisch, F,-G., Hacker, J. and Schroten, H. (1993) Infect. Immun. 61, 2108-21 15. [988] Hacker, J., Kestler, H., Hoschiitzky, H., Jam, K., Lottspeich, F. and Korhonen, T.K. (1993) Infect. Immun. 61, 544-550. [989] Hacker, J. and Morschhauser, J. (1 994) In: P. Klemm (Ed.), Fimbriae. Adhesion, Genetics, Biogenesis, and Vaccines. CRC Press, Boca Raton, Ann Arbor, London, Tokyo, pp. 27-36. [990] Morschhauser, J., Vetter, V, Korhonen, T., Uhlin, B.E. and Hacker, J. (1993) Zbl. Bakt. 278, 165-176. [991] Ishikawa, H. and Isayama, Y. (1987) Infect. Immun. 55, 1607-1609. [992] KO, H.L., Beuth, J., Solter, J., Schroten, H., Uhlenbruck, G. and Pulverer, G. (1987) Infection 15, 237-240. [993] Hirmo, S., Kelm, S., Schauer, R., Nilsson, B. and Wadstrom, T. (1996) Glycoconjugate J. 13, 10051011. [994] Liukkonen, J., Haataja, S., Tikkanen, K., Kelm, S. and Finne, J. (1992) J. Biol. Chem. 267, 21 10521111. [995] Perkins, M.E. and Rocco, L.J. (1988) J. Immunol. 141, 3190-3196. [996] Pakandl, M. and Grubhoffer, L. (1994) Comp. Biochem. Physiol. 108B, 529-536. [997] Schengrund, C.-L. and Ringler, N.J. (1989) J. Biol. Chem. 264, 13233-13237. [998] Brennan, M.J., David, J.L., Kenimer, J.G. and Manclark, C.R. (1988) J. Biol. Chem. 263,48954899, [999] Armstrong, G.D., Clark, C.G. and Heerze, L.D. (1994) Infect. Immun. 62, 2236-2243. [IOOO] Schiavo, G., Demel, R. and Montecucco, C. (1991) Eur. J. Biochem. 199, 705-711. [IOOI] Konami, Y., Ishida, C., Yamamoto, K., Osawa, T. and Irimura, T. (1994) J. Biochem. 115, 767-777. [I0021 Goldhar, J. (1994) Methods Enzymol. 236, 21 1-23 1. [I0031 Mandrell, R.E., Apicella, M.A., Lindstedt, R. and Leffler, H. (1994) Methods Enzymol. 236,231-254. [I0041 Sauter, N.K., Bednarski, M.D., Wurzburg, B.A., Hanson, J.E., Whitesides, G.M., Skehel, J.J. and Wiley, D.C. (1989) Biochemistry 28, 8388-8396. [I0051 Weis, W., Brown, J.H., Cusack, S . , Paulson, J.C., Skehel, J.J. and Wiley, D.C. (1988) Nature 333, 426-43 1. [I0061 Watowich, S.J., Skehel, J.J. and Wiley, D.C. (1994) Structure 2, 719-731. [I0071 Mochalova, L.V, Tuzikov, A.B., Marinina, VP.,Gambaryan, AS., Byramova, N.E., Bovin, N.V and Matrosovich, M.N. (1994) Antiviral Res. 23, 179-190. [I0081 Malby, R.L., Tulip, W.R., Harley, V.R., McKimm-Breschkin, J.L., Laver, W.G., Webster, R.G. and Colman, P.M. (1994) Structure 2, 733-746. [I0091 Lelwala-Guruge, J., Ascencio, F., Kreger, AS., Ljungh, A. and Wadstrom, T. (1993) Zbl. Bakt. 280, 93-106.

400 [ I O I O ] Schroten, H., Plogmann, R., Hanisch, F.-G., Hacker, J., Nobis-Bosch, R. and Wahn, V. (1993) Acta Paediatr. 82, 6-1 1. [ I O I I ] Holder, A.A., Blackman, M.J., Borre, M., Burghaus, P.A., Chappel, J.A., Keen, J.K., Ling, I.T., Ogun, S.A., Owen, C.A. and Sinha, K.A. (1994) Biochem. SOC.Trans. 22, 291-295. [I0121 Titani, K., Takio, K., Kuwada, M., Nitta, K., Sakakibara, E, Kawauchi, H., Takayanagi, G. and Hakomori, S.4. (1987) Biochemistry 26, 2189-2194. [I0131 Chakraborty, I., Mandal, C. and Chowdhury, M. (1993) Mol. Cell. Biochem. 126, 77-86. [lo141 Popoli, M. and Mengano, A. (1988) Neurochem. Res. 13, 63-67. [I0151 Tiemeyer, M., Yasuda, Y. and Schnaar, R.L. (1989) J. Biol. Chem. 264, 1671-1681. [I0161 Banejee, M. and Chowdhury, M. (1994) Human Reproduct. 9, 1497-1504. [I0171 Ahmed, H. and Gabius, H.-J. (1989) J. Biol. Chem. 264, 18673-18678. [I0181 Meri, S. and Pangburn, M.K. (1990) Proc. Natl. Acad. Sci. USA 87, 3982-3986. [lo191 Crocker, P.R., Kelm, S., Dubois, C., Martin, B., McWilliam, B.S., Shotten, D.M., Paulson, J.C. and Gordon, S. (1991) EMBO J. 10, 1661-1669. [I0201 Zeng, F.Y. and Gabius, H.J. (1991) Arch. Biochem. Biophys. 289, 137-144. [I0211 Zeng, F.Y. and Gabius, H.J. (1992) Biochem. Int. 26, 17-24. [I0221 Sgroi, D., Varki, A,, Braesch-Andersen, S. and Stamenkovic, I. (1993) J. Biol. Chem. 268,701 1-7018. [I0231 Powell, L.D., Sgroi, D., Sjoberg, E.R., Stamenkovic, I. and Varki, A. (1993) J. Biol. Chem. 268, 70197027. [I0241 Crocker, P.R., Mucklow, S., Bouckson, V, McWilliam, A,, Willis, A.C., Gordon, S., Milon, G., Kelm, S. and Bradfield, I? (1994) EMBO J. 13, 44904503. [I0251 Kelm, S., Pelz, A,, Schauer, R., Filbin, M.T., Tang, S., de Bellard, M.-E., Schnaar, R.L., Mahoney, J.A., Hartnell, A,, Bradfield, P. and Crocker, P.R. (1994) Curr. Biol. 4, 965-972. [I0261 Fujita, N., Sato, S., Kurihara, T.,Kuwano, R., Sakimura, K., Inusuka, T., Takahashi, Y. and Miyatake, T. (1 989) Biochem. Biophys. Res. Commun. 165, 1 162-1 169. [I0271 Stamenkovic, I. and Seed, B. (1990) Nature 345, 74-77. [I0281 Hanasaki, K., Varki, A. and Powell, L.D. (1995) J. Biol. Chem. 270, 7533-7542. [ 10291 Hanasaki, K., Powell, L.D. and Varki, A. ( I 995) J. Biol. Chem. 270, 7543-7550. [I0301 Powell, L.D., Jain, R.K., Matta, K.L., Sabesan, S. andVarki, A. (1995) J. Biol. Chem. 270,7523-7532. [I0311 Simmons, D. and Seed, B. (1988) J. Immunol. 141, 2797-2800. [I0321 Freeman, S.D., Kelm, S., Barber, E.K. and Crocker, P.R. (1995) Blood 85, 2005-2012. [I0331 Bevilacqua, M., Butcher, E., Furie, B., Furie, B., Gallatin, M., Gimbrone, M., Harlan, J., Kishimoto, K., Lasky, L., McEver, R., Paulson, J., Rosen, S., Seed, B., Siegelman, M., Springer, T., Stoolman, L., Tedder, T., Varki, A,, Wagner, D., Weissman, I. and Zimmerman, G. (1991) Cell 67, 233. [I0341 Varki, A . (1994) Proc. Natl. Acad. Sci. USA 91, 7390-7397. [I0351 Hughes, R.C. (1992) Curr. Opin. Struct. Biol. 2, 687-692. [I0361 Stad, R.K. and Buurman, W.A. (1994) Cell Adhes. Commun. 2, 261-268. [I0371 Lowe, J.B. (1994) In: M. Fukuda and 0. Hindsgaul (Eds.), Molecular Glycobiology. IRL Press, Oxford, pp. 163-205. [I0381 Feizi, T. (1994) In: F. Wieland and W. Reutter (Eds.), 44th Colloquium Mosbach 1993, Glyco- and Cellbiology. Springer-Verlag, Berlin, Heidelberg, pp. 145-1 60. [I0391 Bischoff, J. (1994) Trends Glycosci. Glycotechnol. 6, 351-365. [I0401 Ito, K., Handa, K. and Hakomori, S.-I. (1994) Glycoconjugate J. 1 1 , 232-237. [I0411 Hanski, C., Drechsler, K., Hanisch, F.-G., Sheehan, J., Manske, M., Ogorek, D., Klussmann, E., Hanski, M.-L., Blank, M., Xing, P-X., McKenzie, I.F.C., Devine, P.L. and Riecken, E.-0. (1993) Cancer Res. 53, 40824088. [I0421 Hakomori, S.4. (1991) Curr. Opin. Immunol. 3, 646453. [I0431 Basu, M., Basu, S.S., Li, Z., Tang, H. and Basu, S. (1993) Indian J. Biochem. Biophys. 30, 324-332. [I0441 Fukuda, M. (1994) In: K. Bock and H. Clausen (Eds.), Complex Carbohydrates in Drug Research. Munksgaard, Copenhagen, pp. 353-365. [I0451 Hanisch, F.-G., Hanski, C. and Hasegawa, A. (1992) Cancer Res. 52, 3138-3144. [I0461 Hemmerich, S. and Rosen, S.D. (1994) Biochemistry 33, 48304835.

40 1 [I0471 Scudder, P.R., Shailubhai, K., Duffin, K.L., Streeter, P.R. and Jacob, G.S. (1994) Glycobiology 4, 929-933. [I0481 Huang, K., Kikuta, A. and Rosen, S.D. (1994) J. Neuroimmunol. 53, 133-141. [I0491 Graves, B.J., Crowther, R.L., Chandran, C., Rumberger, J.M., Li, S., Huang, K.-S., Presky, D.H., Familletti, P.C., Wolitzky, B.A. and Burns, D.K. (1994) Nature 367, 532-538. [ 10501 Drickamer, K. ( I 994) Glycobiology 4, 245-248. [I0511 Hajela, K., Kayestha, R. and Kumar, G.P. (1994) FEBS Lett. 354, 217-219. [I0521 Kotovuori, P., Tontti, E., Pigott, R., Shepherd, M., Kiso, M., Hasegawa, A,, Renkonen, R., Noramo, P., Altieri, C.D. and Gahmberg, C.G. (1993) Glycobiology 3, 131-136. [I0531 Diamond, M.S. and Springer, T.A. (1994) Curr. Biol. 4, 506-517. [I0541 Boggemeyer, E., Stehle, T., Schaible, U.E., Hahne, M., Vestweber. D. and Simon, M.M. (1994) Cell Adhes. Commun. 2, 145-157. [I0551 Simmons, R.D. and Cattle, B.A. (1992) J. Neuroimmunol. 41, 123-130. [lo561 Mulligan, M.S., Paulson, J.C., De Frees, S., Zheng, 2.-L., Lowe, J.B. and Ward, P.A. (1993) Nature 364, 149-151. [I0571 Buerke, M., Weyrich, A S . , Zheng, Z.-L., Gaeta, F.C.A., Forrest, M.J. and Lefer, A.M. (1994) J. Clin. Invest. 93, 1140-1 148. [I0581 Ichikawa, Y. and Wong, C.43. (1994) In: K . Bock and H. Clausen (Eds.), Complex Carbohydrates in Drug Research. Munksgaard, Copenhagen, pp. 1 18-1 30. [I0591 Chen, L.Y., Nichols, W.W., Hendricks, J.B., Yang, B.C. and Mehta, J.L. (1994) Cardiovasc. Res. 28, 14 14-1 422. [I0601 Winn, R.K., Liggitt, D., Vedder, N.B., Paulson, J.C. and Harlan, J.M. (1993) J. Clin. Invest. 92, 20422047. [I0611 Stamenkovic, I., Sgroi, D. and Aruffo, A. (1992) Cell 68, 1003-1004. [I0621 Van den Berg, T.K., Brev6, J.J.P., Damoiseaux, J.G.M.C., Dopp, E.A., Kelm, S., Crocker, P.R., Dijkstra, C.D. and Kraal, G. (1992) J. Exp. Med. 176, 647-655. [I0631 Meyer-Franke, A. and Barres, B. (1994) Curr. Biol. 4, 847-850. [I0641 Crocker, P.R., Freeman, S., Gordon, S. and Kelm, S. (1995) J. Clin. Invest. 95, 635-643. [I0651 Engel, P., Nojima, Y., Rothstein, D., Zhou, L.-J., Wilson, G.L., Kehrl, J.H. and Tedder, T.F. (1993) J. Immunol. 150, 4719-4732. [I0661 Braesch-Andersen, S. and Stamenkovic, I. (1994) J. Biol. Chem. 269, 11783-1 1786. [I0671 Umemori, H., Sato, S., Yagi, T., Aizawa, S. and Yamamoto, T. (1994) Nature 367, 572-576. [I0681 Beppu, M., Takahashi, T., Hayashi, T. and Kikugawa, K. (1994) Biochim. Biophys. Acta 1223,47-56. [I0691 Devi, S.J.N., Robbins, J.B. and Schneerson, R. (1991) Proc. Natl. Acad. Sci. USA 88, 7175-7179. [lo701 Duk, M., Sticher, U., Brossmer, R. and Lisowska, E. (1994) Glycobiology 4, 175-181. [I0711 Younes-Chennoufi, A.B., Leger, J.-M., Hauw, J.-J., Preud’homme, J.-L., Bouche, P., Aucouturier, P., Ratinahirana, H., Lubetzki, C., Lyon-Caen, 0. and Baumann, N. (1992) Ann. Neurol. 32, 18-23. [I0721 Kanamori, A,, Inoue, S., Xulei, Z., Zuber, C., Roth, J., Kitajima, K., Ye, J., Troy 11, F.A. and Inoue, Y. (1994) Histochemistry 101, 333-340. [I0731 Bast, B.J.E.G., Zhou, L.-J., Freeman, G.J., Colley, K.J., Ernst, T.J., Munro, J.M. and Tedder, T.F. (1992) J. Cell Biol. 116, 4 2 3 4 3 5 . [I0741 Portner, A,, Peter-KataliniC, J., Brade, H., Unland, E, Buntemeyer, H . and Muthing, J. (1993) Biochemistry 32, 12685-12693. [I0751 Mukuria, C.J., Noguchi, A,, Suzuki, E. and Naiki, M. (1994) Jpn. J. Med. Sci. Biol. 47, 253-264. [I0761 Kimber, S.J., Bentley, J., Ciemerych, M., Moller, C.J. and Bock, E. (1994) Eur. J. Cell Biol. 63, 102113. [I0771 Roth, J., Kempf, A,, Reuter, G., Schauer, R. and Gehring, W.J. (1992) Science 256, 673-675. [I0781 Roth, J. (1993) Histochem. J. 25, 687-710. [ 10791 Jowett, A.K., Kimber, S.J. and Ferguson, M.W.J. ( I 994) J. Anat. 185, 89-94. [I0801 Skacel, P.O., Edwards, A.J., Harrison, C.T. and Watkins, W.M. (1991) Blood 78, 1452-1460. [I0811 Vertino-Bell, A,, Ren, J., Black, J.D. and Lau, J.T.Y. (1994) Dev. Biol. 165, 126-136. [I0821 Nagai, Y. and Tsuji, S. (1994) Prog. Brain Res. 101, 119-126.

402 [I0831 Rahmann, H., Rosner, H., Kortje, K.-H., Beitinger, H. and Seybold, V. (1994) Prog. Brain Res. 101, 127-145. [I0841 Yamashita, T., Tsuji, S. and Nagai, Y. (1991) Glycobiology 1, 149-154. [I0851 Watanabe, M., Timm, M. and Fallah-Najmabadi, H. (1992) Dev. Dynam. 194, 128-141. [I0861 Zuber, C., Lackie, P.M., Catterall, W.A. and Roth, J. (1992) J. Biol. Chem. 267, 9965-9971. [I0871 Cantz, M. (1982) In: R. Schauer (Ed.), Sialic Acids - Chemistry, Metabolism and Function. Springer Verlag Wien, New York, pp. 307-320. [lo881 D’Agrosa, R.M., Hubbes, M., Zhang, S., Shankaran, R. and Callahan, J.W. (1992) Biochem. J. 285, 833-838. [lo891 Okamura-Oho, Y., Zhang, S. and Callahan, J.W. (1994) Biochim. Biophys. Acta 1225, 244-254. [lo901 Sasagasako, N., Miyahara, S., Saito, N., Shinnoh, N., Kobayashi, T. and Goto, I. (1993) Clin. Genet. 44, 8-1 1. [I0911 Schmid, B., Paton, B.C., Sandhoff, K. and Harzer, K. (1992) Hum. Genet. 89, 513-518. [lo921 Mueller, O.T., Henry, W.M., Haley, L.L., Byers, M.G., Eddy, R.L. and Shows, T.B. (1986) Proc. Natl. Acad. Sci. USA 83, 18 17-1 82 1. [I0931 Grzeschik, K.H. and Skolnick, M.H. (1991) Cytogenet. Cell Genet. 58, 785-799. [lo941 Elsliger, M.-A. and Potier, M. (1994) Prot. Struct. Funct. Genet. 18, 81-93. [I0951 Pshezhetsky, A.V. and Potier, M. (1993) Biochem. Biophys. Res. Commun. 195, 354-362. [I0961 Mancini, G.M.S., Hu, P., Verheijen, EW., van Diggelen, O.P., Janse, H.C., Kleijer, W.J., Beemer, EA. and Jennekens, F.G.I. (1992) Eur. J. Pediatr. 151, 590-595. [lo971 Pisoni, R.L. and Thoene, J.G. (1991) Biochim. Biophys. Acta 1071, 351-373. [I0981 Haataja, L., Schleutker, J., Laine, A.-P., Renlund, M., Savontaus, M.-L., Dib, C., Weissenbach, J., Peltonen, L. and Aula, P. (1994) Am. J. Hum. Genet. 54, 1042-1049. [I0991 Maguire, T.M., Gillian, A.M., O’Mahony, D., Coughlan, C.M., Dennihan, A. and Breen, K.C. (1994) Neurobiol. Aging 15, 99-102. [I 1001 Maguire, T.M., O’Mahony, D., Gillian, A.M., Dennihan, A. and Breen, K.C. (1994) Neurodegeneration 3, 129-133. [IIOI] Narayanan, S. (1994) Ann. Clin. Lab. Sci. 24, 376-384. [I 1021 Vavasseur, F., Dole, K., Yang, J., Matta, K., Myerscough, N., Corfield, A,, Paraskeva, C. and Brockhausen, I. (1994) Eur. J. Biochem. 222, 415424. [1103] Vierbuchen, M., Schroder, S., Larena, A., Uhlenbruck, G. and Fischer, R. (1994) Virchows Arch. Int. J. Pathol. 424, 205-21 1. [1104] Kemmner, W., Kriick, D. and Schlag, P. (1994) Clin. Exp. Metastasis 12, 245-254. [1105] Sata, T., Roth, J., Zuber, C., Stamm, B. and Heitz, P.U. (1991) Am. J. Pathol. 139, 1435-1448. [1106] Dall’Olio, F., Malagolini, N., Guerrini, S., Serafini-Cessi, F. (1993) Biochem. Biophys. Res. Commun. 196, 714-720. [1107] Patel, P.S., Adhvaryu, S.G., Balar, D.B., Parikh, B.J. and Shah, P.M. (1994) Anticancer Res. 14, 747751. [I1081 Harvey, B.E., Toth, C.A., Wagner, H.E., Steele Jr., G.D. and Thomas, P. (1992) Cancer Res. 52, 17751779. [1109] Mostafapour, M.K. and Goldstein, I.J. (1993) Arch. Biochem. Biophys. 303, 255-259. [1110] David, L., Nesland, J.M., Funderud, S. and Sobrinho-Simdes (1993) Cancer 72, 1522-1527. [I 11 11 Easton, E.W., Bolscher, J.G.M. and van den Eijnden, D.H. (1991) J. Biol. Chem. 266, 21674-21680. I11121 Miyagi, T., Sato, K., Hata, K. and Taniguchi, S. (1994) FEBS Lett. 349, 255-259. [1113] Takano, R., Muchmore, E. and Dennis, J.W. (1994) Glycobiology 4, 665-674.

J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), GIycoproteins 11

0 1997 Elsevier Science B.V. All rights reserved CHAPTER 12

Plant lectins: tools for the study of complex carbohydrates Irwin J. Goldstein, Ph.D., and Harry C. Winter, Ph.D. Department of Biological Chemistry, Uniuersib of Michigan Medical School, Ann Arboc MI 48109-0624, USA

Ronald D. Poretz, Ph.D. Department of Biochemistry and Microbiology, Rutgers Uniuersity, New Brunswick, NJ 08903, USA

1. Introduction The literature on plant lectin chemistry and biochemistry has increased exponentially during the past decade and continues at a rapid pace. New lectins of unique specificity have been isolated, characterized, and employed in biological studies; numerous lectins have been crystallized and their X-ray structures solved at high resolution, both in their native form and complexed with mono- and oligosaccharides; the genes encoding scores of lectins have been cloned and expressed in heterologous systems and in a few cases, mutations generated, shedding light on the molecular basis of carbohydrate-lectin interaction. This chapter will describe the properties of a selection of lectins present in the tissues of plants (seeds, bulbs, tubers, bark and leaves), in order to provide the reader with an understanding of the general nature of these carbohydrate-specific proteins. The emphasis of the discussion will be directed to the application of lectins as reagents for the detection, isolation and characterization of glycoconjugates, and the contemporary methodologies used in these applications. Lectins were defined in 1980 as carbohydrate-binding proteins of non-immune origin that agglutinate cells or precipitate polysaccharides or glycoconjugates [ 11. This definition implies that lectins are multivalent. It is of course possible that carbohydrate-binding proteins will be found or generated (by synthesis, chemical modification or molecular biological means) that contain a single binding site. An alternate definition was coined by Barondes [2] who defined lectins as carbohydrate-binding proteins that are not enzymes or antibodies. Interestingly, carbohydrate transport proteins were not included in this definition. Numerous reviews have appeared since the last review by the authors [3]. These include several monographs [4] and a series of Aduances in Lectin Research [5]. For a review of agglutinins present in higher fungi, see the paper by Pemberton [6].

2. Isolation, puriJication and characterization of plant lectins Detailed discussions of the general procedures for the isolation and purification of plant lectins are found in former reviews [3,4,7]. Numerous purification protocols also 403

404

may be found in Methods in Enzymology [S]. Although the classical techniques of protein chemistry have and continue to be employed for the purification of plant lectins, most present-day schemes involve a combination of classical approaches and affinity chromatography. Virtually all contemporary schemes for lectin purification exploit the specific sugar binding specificity of lectins in which carbohydrate ligands are immobilized and employed as matrices for their adsorption. (See a chart of immobilized ligands employed for lectin purification, Table 1.) More recent examples of affinity purification schemes include the isolation of the Maackia arnurensis leukoagglutinin on murine-laminin-Sepharose [9], the Alocasia indica lectin on asialofetuin silica [ 101 and the Coccinea indica lectin on denatured soybean-agglutinin-Sepharose [ 1 I]. Characterization of lectins as proteins is similar to the characterization of any protein; descriptions may be found in published reviews [3,4,7].

3. Structure and carbohydrate-binding specijicity of lectins Plant lectins comprise a large heterogenous group of (g1yco)proteins. As cell agglutinating molecules, they are multivalent containing two or more carbohydrate binding sites. They range in size from 8500Da (stinging nettle lectin) to 125000Da, for example the lima bean lectin (1 24 000 Da) and the Phaseolus uulgaris L4 isolectin (126 000 Da). Classified according to their taxonomic families, lectins exhibit common physical chemical properties. Legume lectins are largely (but not exclusively) glycoproteins containing 2-4 subunits and divalent metal ions (e.g. Ca2' and Mn2') which are required for expression of their carbohydrate-binding and hence biological activity. Members of the family Solanaceae (e.g. tomato, potato, Datura strarnoniurn) consist of two polypeptide domains: an unsubstituted pure peptide and a highly glycosylated portion (approx. 40%) containing mainly L-arabinose' (covalently bound to hydroxylysine) but also D-galactose. Lectins of the family Gramineae are homodimers containing many disulfide bonds which confer great physical stability; they are devoid of both carbohydrate and metal ions, examples include wheat germ agglutinin, rice and barley lectins. Secondary structures of all lectins examined by the techniques of high resolution X-ray diffraction and circular dichroism have been shown to consist primarily of &sheets and random coils with little or no a-helices [12]. The carbohydrate portions of several lectins have been investigated. To summarize: except for lectins of Solanaceae which contain L-arabinose and galactose (see above), the glycoportion of lectins are composed of GlcNAc and Man with the occasional L-FUCand/or D-Xyl. Examples are soybean agglutinin with N-linked Man9 GlcNAc2 [ 131, Erythrinu cristugulli [ 141, and Phaseolus uulgaris [ 151 which contain N-linked Man9 GlcNAqXyl Fuc. The Gramineae and other monocotyledonous lectins are devoid of carbohydrate units. In cases where the carbohydrate has been removed, or is lacking in recombinant lectins, full activity is retained [ 16,171. The amino acid sequence homologies that exist among various families of lectins is of interest and of fundamental importance. It has been determined that such homologies

'

All sugars are of the D-COnfigUratlOn except where noted otherwise, e.g. L-fucose

Table 1 Solid-phase adsorbants employed for the isolation of lectins Sephadex Untreated Stroma-derivatized Agarose/Sepharose Untreated Acid-treated Mannose-derivatized Mannan-derivatized 3-0-Methylglucosamine-derivatized N-Acetylglucosaminyl-derivatized N,W-Diacetylchitobiosyl-derivatized N,W ,W’-Triacetylchitotriosyl-derivatized N -Acetylgalactosamine-derivatized N-Acetylgalactosaminyl-derivatized Galactosyl-derivatized Galactose-derivatized Lactose-derivatized L-Fucose(sy1)-derivatized Glycoprotein-derivatized

Glycopeptide-derivatized

Canaualia ensiformis. Lens culinaris, Viciafaba, Pisum satiuum, T cracca, !l satiua. Lathyrus satiuus, L. tingitanus, Onobiyehis uiciifolia, Dioclea grandipora, Cratylia mollis (1 & 2 ) Phaseolus acutifolius var. latifolius Wistaria floribunda agglutinin, Sophora japonica, Arachis hypogaea. Eiythrina spp., Hura crepitans, Ricinus communis, Abrus precatorius. Adenia digitata. Viscum album, Momordica charantia. Bauhinia purpurea Crotalaria juncea. H. crepitans, A. digitata. M. charantia, B. purpurea. Euphorbia marginata, Iris germanica CIiuia miniata, Cymbidium hyb., Epipactis helleborine, Galanthus niualis, Hippeastrum hyb., Leucojum spp., Narcissus spp., Tulipa gesneriana I, Vicia eruilia. 0.vicifolia Tulipa gesneriana, Allium ascalonicum !b faba Triticum uulgaris, Hordeum uulgare, Secale cereale, Solanum tuberosum. Grzffonia simplicifolia 11, Codium tomentosum Datura stramonium S. tuberosum, Ulex europaeus 11 Psophocarpus tetragonolobus, Vicia cracca, Maclura pomifera. Falcata japonica Caragana arborescens, Cytisus scoparius, Clerodendron trichotomum Vicia uillosa. Glycine mar, Maclura pomifera, Arachis hypogaea. Griffonia simplicifolia, Clerodendron trichotomum Eiythrina spp., Crotalaria juncea, Moluccella laeuis, Galactia tashiroi. Phaseolus mungo, Sclerotinia spp. Arachis hypogaea, Eiythrina spp.. Sarothamnus welwitschii. Sphenostylus stenocarpus, Telfairia occidentalis Lotus tetragonolobus. Ulex europaeus I Triticum uulgaris, Oiyza satiua, Solanum tuberosum. Phaseolus uulgaris, Adenia digitata. Vicia spp., Aleuria aurantia, Bauhinia purpurea. Trichosanthes japonica, Pleurotus ostreatus, Amaranthus cruentus, Biyonia dioica, Colchicum autumnale, Sambucus nigra I , Eranthis hyemalis, Artocarpus atilis, Irisx hollandica. Ficus cunia, Maackia amurensis. Pinellia terneata, Tulipa gesneriana 11, Robinia pseudoacacia, Coccinea indica Phytolacca americana, Euonymus europaeus, Vicia graminea, Luffa acutangula, Maackia amurensis continued on next page

P

8 Table 1, continued Polyacrylamide (or BioGel P) Guaran entrapped Erythrocyte stroma entrapped Lactobionate-derivatized 1-Amino-I -deoxy-lactitol-derivatized Melibionate-derivatized Galactosylderivatized N-Acetylgalactosaminyl-derivatized Lactosyl-derivatized L-Fucosyl-derivatized Maleylated much copolymer Chitin, untreated Starch N,N,N'-Triacetylchitotriose-derivatized Lactose-derivatized L-Fucose-derivatized Polyleucyl-derivatizedrnucin

Arachis hypogaea Lactarius spp., Aegopodium podagraria, Amaranthus leucocarpus, Phaseolus coccineus var. aluba Psophocarpus tetragonolobus Arachis hypogaea Maclura pornifera, Griffonia simplicifolia Arachis hypogaea, GriFonia simplicifolia I , Erythrina spp., Momordica charantia Maclura pomifera Ricinus communis Ulex europaeus I Ulex europaeus 11 Triticum uulgaris. Secale cereale, Gnffonia simplicifolia 11, Chelidonium majus, Cyphomandra betacea, Psathyrella uelurina, Urtica dioica Cytisus sessilifolius, Ulex europaeus 11 Arachis hypogaea U l a europaeus I

Illex europaeus 11, Phaseolus lunatus, Dolichos bifIorus, Macrofyloma axillare, Lycopersicon esculentum, Vicia uillosa. mstaria floribunda, Sophora japonica, Maclura pomifera, Euonymus europaeus, Grtffonia simplicifolia I, Lotus tetragonolobus

Cross-linked Guaran

Gbcine m a , Ricinus communis, Echinocystis lobata, Psophocarpus tetragonolobus

Cross-linked Arabinogalactan

Arachis hypogaea, Crotalaria juncea, Ricinus communis, Momordica charantia

Silica (Synsorb, Toyopearl, etc.) Oligosacc haride-derivatized Glycoprotein-derivatized

Phaseolus lunatus. Amphicarpa bracteata, Griffonia simplicifolia IV, Amaranthus caudatus. Irisx hollandica Amaranthus spp., Alocasia indica, Hericium erinaceum

Cellulose, mucin-derivatized

Vicia cracca

407

alone do not define carbohydrate specificity, i.e. a group of lectins that exhibit extensive homology may display quite different carbohydrate-binding specificities. The most extensive studies of homology among plant lectins have been carried out with the legume seed lectins, many of which have been sequenced at the amino acid or DNA level [ 181. Overall interspecific homology (taking into account the circular permutation in Con A and similar lectins from members of the tribe Diocleae) among the legume lectins ranges from 35 to 70%, with some regions of the proteins being highly conserved irrespective of the carbohydrate-binding specificity, and other regions being more variable. Young and Oomen [I91 compared sequences of 11 lectins from various tribes of the Papilionoideae subfamily which exhibited different major carbohydratebinding specificities (Glc/Man, Gal/GalNAc, L-FUC,GlcNAc and complex). They found that a ring of hypervariability occurs around the carbohydrate binding site, as identified by X-ray crystal structures of Con A and pea lectin. An additional area of variability is in the C-terminal region, but the core regions of all these lectins are highly homologous. This situation is analogous to the hypervariable and conserved regions of immunoglobulins. Pairs of lectins having similar binding specificity appeared no more homologous than pairs having different specificities but of similar taxonomic relationship. Only the GalNAc-binding lectins appeared to have greater overall homology, but all were from the same tribe (Phaseoleae). The recently recognized mannose-specific lectins of monocot storage tissue have also been extensively sequenced, largely by cDNA cloning techniques. These lectins sequences show considerable homology among the families involved [20-221 (Amaryllidaceae, Liliaceae, and Orchidaceae), but not to the leguminous lectins, whether of MadGlc, or of other binding specificity (Fig. 1). Typically, more than one sequence and as many as 7 [23] are detected in any species, indicating the presence of multigene families of common origin. Sequences of four distinct clones from Clivia miniata [21] showed a high degree of homology (>90%), as did four cDNA clones of orchid lectins (3 from Listera ooata and one from Epipactis helleborine) [22]. Comparison between sequences representing the mature proteins shows interspecific homology of about 50% or greater. The cDNA sequences reveal the presence of an N-terminal signal peptide of 20-30 residues, and a C-terminal region of the precursor polypeptide that are cleaved to give the mature protein. Between species, there is little or no homology of these cleaved peptides. Lectins which bind to the T-antigen (Gal@1-3)GalNAc) are found in Artocarpus integrifolia (Jacalin) and Maclura pomifera, two species of the family Moraceae. They are tetramers of a heterodimer containing a small peptide of about 2kDa, and a large polypeptide of 14.7kDa. The isoforms of the 2kDa (20 amino acid) light peptide from each species show 8&90% identity; homology between the species is about 50% [24]. Comparison of the heavy chain sequences, including some unresolved microheterogeneity, shows 85% overall homology [25]. Again, little or no homology is evident between these and either the legume or monocot lectins (Fig. 1). Amaranthin, a further T-antigen-binding lectin from Amaranthus caudatus seeds, is a homodimer [26]. These observations suggest that carbohydrate-binding and agglutinating activities were not conserved via evolution from a common ancestral lectin gene, but that these properties have been acquired by many different archetypal proteins throughout the course

100 NPA

110

140

I50

180

170

I80

I33i

120 YSPONKAlWASNTDGEN~GHFVClLOKDRNVVlYGTDRWATG YTGAVGIPESPASEKYPT~SGKITPT~SEKYPTTGKIKLVTGK

-- -L(---L-SLP

-

I ----NRYGS-YVV-- -RNCTVG I--AEONKV-- IVR IVDVTGSA------NIHGA--VGV-C-APON-.L~EMIKLV.R--L-T-

LoA

- N N R - ~ - Y - - - D - V R - - I -NV I L - -

HHA

-A-R-OV------O---.-NY-----------------

ConA

DKRLSAVVSYPN~ADSATV~SYDVDLDNVLPEWVRVGLSASTGLYK~~~ETNTILSWSFTSKLKSNSTH

LBL

T-L-V-SLV--SGST-YI~-EK-EK-~-KS------Ni-F~-~-FN-GNV-ADDY--~--A---SDG1PCEDLSLANlVLNKlL

190

200

210

220

230

240

250

260

UEAII T-S-1VSL---SDCT-Ni-TASS---KA1-----S--F-GGV-NAA~.KFDHDY---Y---N-EA-OsOT

Fig. I . Comparison of representative amino acid sequences from lectins of various families and binding specificities. Monocot mannose-specific lectins: NPA, Narcissus pseudonarcissus agglutinin; LOA, Listera ouaia agglutinin, one of seven genes sequenced; HHA, amaryllis lectin. Gal/GalNAc-specific lectins from Moraceae: AIA, Ariocarpus infegrifolia agglutinin (Jacalin) large subunit; MPA, Maclura pornifera agglutinin, large subunit. Legume lectins: Con A, Canaoalia ensiformis lectin (MadGlc-specific); LBL, lima bean (Phaseolus lunafus) lectin (GaVGalNAc-specific); UEA 11, Ulex europaeus agglutinin I1 (GlcNAc-specific). The monocot sequences show the entire gene sequence, indicating by arrows post-translational cleavage sites at the N- and C-terminal regions. Con A is shown as its cyclic permutation to align with the other legume lectins, the C-terminus indicated by the asterisk. Homology between members of each taxonomically related group is shown as a dash (-) denoting identity, and underlined superscripted letter showing conservative substitutions. Dots indicate gaps introduced to maximize alignment of homologous sequences.

of evolution. The constraints of primary structure required to exhibit lectin activity are minimal; hence homology by convergent evolution generally does not occur. One exception to this general concept may be found in the chitin-binding lectins (GlcNAcspecific) of the Gramineae, as represented by wheat germ agglutinin. The relationship of WGA to other chitin-binding proteins has been well-reviewed recently [27,28]. Wheat germ and other Gramineae lectins are comprised of single polypeptide chains having four GlcNAc-binding domains, each with multiple intradomain disulfide linkages and highly conserved amino acid residues involved in GlcNAc binding. This domain is termed the hevein domain because of its occurrence in hevein, an abundant protein in latex of Heoea brasiliensis (rubber tree); it is conserved in highly homologous forms in a wide variety of proteins in the plant kingdom, as well as a few examples in animals

409

(snake venom disintegrins), generally in single copies or in 2 or 4 repeats within a polypeptide strand. It occurs, for example, in lectins from plants of the Solanaceae (e.g. potato [29]) and Urticaceae families, in class I chitinases (although the chitinase activity itself occurs in an unrelated domain), in wound-induced proteins, and in the small antimicrobial peptide from Amaranthus caudatus. This last peptide appears to be unrelated to the Amaranthus lectins which have been the subject of recent studies [26]. It is uncertain whether this domain has developed independently several times, or whether it is a very old gene that has been conserved through a great deal of evolutionary divergence. The latter explanation appears to be the more plausible, since other GlcNAc-binding lectins, such as Ulex europaeus lectin I1 (see Table Al, appendix) do not have a sequence homologous to the hevein domain [ 191. Originally, it was believed that the carbohydrate-binding specificity of lectins was directed solely against monosaccharides. In fact, Makela [30] suggested that lectinreactive monosaccharides could be divided into four classes based on their configuration at C-3 and C-4 of the pyranose ring. Utilizing immunochemical (precipitin and hapten inhibition studies) and modern biophysical techniques (equilibrium dialysis, direct or difference ultraviolet and fluorescence spectroscopy, NMR spectroscopy, X-ray crystallography, microcalorimetry) it has now been established that many (if not most) plant lectins have combining sites that recognize sequences of sugars ranging from two to five monosaccharides. In addition to a sequence of monosaccharides, the binding specificity of lectins often extend to specific linkages. As examples, lectins of the family Solanaceae (tomato, potato, Datura stramonium) specifically bind chitodextrins composed of three to five fl- 1,4-1inked GlcNAc units [3 11; several lectins recognize the T-antigen (Gal@1-3)GalNAc): Amaranthus caudatus [26] and Artocarpus integrifolia [32]; Griffonia simplicifolia IV binds to the Lewis b tetrasaccharide [32-341, and the Maackia amurensis lectin recognizes the trisaccharide sequence NeuSAc(a2-3)Gal(fl 1-4)GalNAc [ 3 5 ] . Among specificity groups are lectins which recognize mannose/glucose, solely mannose, galactoselN-acetylgalactosamine, (or principally one of these sugars), N-acetylglucosamine, L-fucose, and N-acetylneuraminic-acid-containing oligosaccharides. Interestingly, thus far, no xylose-, arabinose- or ribose-recognizing lectins have been discovered, nor has a lectin which binds B-glucans been reported. The most elegant and visually convincing argument for the binding of oligosaccharides to plant lectins is presented by the X-ray crystallographic studies of oligosaccharidelectin complexes. Numerous such studies have been published during the past ten years (see Quiocho [36] and Rini [37] for reviews). A listing of the X-ray crystal structure (arranged in chronological order) of plant lectins is presented in Table 2 commencing with the structure of concanavalin A at 3 resolution [38,39]. Interestingly, it was not until 1989 that a definitive high resolution X-ray crystal structure of the concanavalin A-methyl a-D-mannopyranoside complex was elucidated [40], and the most recently reported is that of peanut agglutinin [41]. Reviews on carbohydrate-protein interactions are also worthy of note [4245]. The structure of Griffonia simplicifolia IV (GS IV) and its complex with the Lewis b human blood group determinant at 2.0 resolution has been solved [51]. The molecular recognition of Leb-OMe by GS IV involved both polar and extensive non-polar

a

410

Table 2 X-Ray crystal structures of lectins Lectin

Monosaccharide specificity

Year

MadGlc GalNAc

1972-1989

Yiciafaba (Favin) Pisum sativum (Pea) Eythrina corallodendron (ECorL) Grrffonia simplicifolia (GS IV) Lathyrus ochrus (LOL-I) Lathyrus ochrus (LOL-11) Lens culinaris (Lentil) Psophocarpus tetragonolobus

Man/Glc Man/Glc

1986

GalNAc GalNAc

1989

Arachis hypogaea (Peanut)

Reference(s)

Legumes Canaoalia ensiformis (Con A) Glycine m a (SBA)

Non-legumes Triticum aestiuum (WGA) Ricinus communis (rich) Galanthus nivalis (GNA) Urtica dioica (Stinging nettle)

Man/Glc

1984 1986, 1993 1991, 1993 1991, 1992

MadGlc MadGlc

1992

Gal/NAc

1993

Gal

1994

GlcNAc/NANA

1974-1992

Gal

1981, 1987

[571 [58,58a]

Man GlcNAc

1990, 1995

[59,59a]

1993

[601

1993

interactions. Key polar interactions include side chains of Asp 89 and Asn 135 and the peptide NH of Gly 107; non-polar interactions involve Tyr 105, Phe 108 and Tyr 223 of GS IV. The crystal structure of wheat germ agglutinin isolectin 1 (WGA 1) complexed with NeuAc(a2-3)Gal(~1-3)[NeuAc(a2-6]GalNAc(a 1-O)Ser/Thr was determined at 2.0 resolution [57]. A preference for a-2,6-NeuAc over a-2,3-NeuAc-linked oligosaccharides was observed in the aromatic-rich sites of domains A and B of WGA. The X-ray structure of a diantennary octasaccharide of the N-acetyllactosamine type complexed to isolectin I from Lathyrus ochrus at 2.3 A resolution was solved [53]. The complex is stabilized by numerous hydrogen bonds, many also involving water molecules. Van der Waals interactions, including some with aromatic residues are also involved. It is apparent that GlcNAc((31-2)Man residues play an important role in the oligosaccharidelectin interaction. The X-ray structure of the Erythrinu corullodendron lectin complexed with lactose was solved at 2.6 A resolution. It was determined that only the galactose moiety of the lactose ligand was situated within the binding site; it is stabilized by seven hydrogen bonds and numerous hydrophobic interactions [50]. The X-ray crystal structure of the pea lectin bound to the branched trimannosyl core oligosaccharide found in N-linked glycoproteins was solved at 2.6 resolution [49]. It

A

A

41 1

was reported that “the trisaccharide binds primarily through one of the terminal a-linked mannose residues” [49]. Stabilization of the complex is by hydrogen bonds involving amino acid residues Asp-8 1, Gly-99, Asn- 125, Ala-2 17 and Glu-2 18 and oxygen atoms at C-3, C-4, C-5 and C-6 of the carbohydrate [49]. NMR has provided important information on the mechanism of binding of carbohydrate ligands to lectins[61-661. In a series of papers, Kronis and Carver investigated the binding of sialylated oligosaccharides to wheat germ agglutinin; line broadening of proton signals of the Neu5Ac residues was observed [62,63]. Neurohs and colleagues [64] obtained kinetic parameters for the interaction of 13C-enriched methyl 0-lactoside with peanut agglutinin by using line broadening. The interaction of oligosaccharides to concanavalin A was studied by Brewer and Brown using a nuclear magnetic relaxation dispersion technique [65]. Recently, Takesada et al. [66] investigated the interaction of the Japanese elderberry bark lectin (Sambucus sieboldiana) with oligosaccharides by NMR. They observed that the signal of H-4 of galactose in methyl 0-galactoside and methyl 6-lactoside was broadened (indicative of a specific interaction), but not those of H- 1, OCH3 or H-2. A fascinating new dimension of specificity in carbohydrate-protein interactions was discovered by Brewer and his colleagues [67-69]. They observed that interaction of lectins with di-, tri- and tetraantennary oligosaccharides and glycopeptides results in “the formation of homogeneous (homopolymeric) aggregates” [67-701 which give the appearance of well defined lattice structures under the electron microscope [68]. Additionally, mixtures of multivalent glycopeptides react with lectins to give “individual cross-linked homogeneous precipitates” [67]. Conversely, “binary mixtures of lectins with branched oligosaccharides form separate cross-linked lattice structures between each lectin and the carbohydrate” [70]. The X-ray crystal structure of soybean lectin cross-linked with a diantennary analog of the blood group I carbohydrate antigen has been solved (70 A). The number of purified lectins is much too great to elucidate individually their carbohydrate-binding properties. For the purpose of this review, lectins are grouped and discussed in terms of their carbohydrate-binding specificity. Furthermore, the decision was made to concentrate on those lectins which have proved to be most useful in isolating and characterizing glycoconjugates present on cell surfaces and as they occur in solution. 3. I . Mannose/glucose-binding lectins The mannoselglucose-binding lectins comprise a large group of agglutinins present in the family Leguminosae, primarily in seeds. Prominent among these lectins are those from the jack bean (Canaualia ensiformis; concanavalin A), the lentil (Lens culinaris), the pea (Pisum satiuum), the fava bean (Vicia faba) and the common vetch (Vicia cracca). All these agglutinins are metalloproteins requiring metal ions e.g. Ca2+and/or Mn2+ for their carbohydrate-binding activity. In an important discovery, Foriers and coworkers [71], and others [72,73], observed an intriguing relationship between concanavalin A, a homotetramer, and the pea, lentil, Vicia faba (all composed of two light a-chains and two heavy p-chains), as well as other leguminous plant lectins. Termed a circular permutation of amino acid sequences,

412

Fig. 2. Diagrammatic representation of sequence homology relationships of leguminous lectins. (Adapted from Strosberg et al. [3]). Positions of the post-translational cleavages (PTC) and post-translational ligations (PTL) are indicated. Sector-shaped boxes indicate highly homologous regions used to deduce the alignment of the circularly permuted Diocleae (e.g. Con A) lectins and the heterodimeric lectins with the unprocessed singlechain lectins. Sequences of Con A within the boxes are shown; all lectins named have the same sequence or an homologous one with no more than two conservative substitutions at these points. The additional C-terminal sequence of the outermost group is highly homologous within the group, but not to any sequence in the other groups.

it was determined that maximum homology is observed when lectins from the lentil, the pea and Vicia faba are aligned such that their a-chains correspond to residues 70119 of concanavalin A and their @-chainscommencing at residue 120 of concanavalin A (Fig. 2). In similar fashion the primary amino acid sequence of the legume lectins such as the peanut and soybean agglutinins can be related to Con A in which positions 122 to 237 and 1 to 69 are aligned with positions 1 to 185 of the soybean and peanut proteins, and positions 70 to 121 (of Con A) aligned with positions 186 to 238 of the soybean and peanut agglutinin. Genetic analysis has demonstrated that the permutation is caused by post-translational cleavage and re-ligation in the case of Con A, rather than a rearrangement of genetic elements. These relationships reveal the extensive evolutionary (taxonomic) relationships which exist among the legume lectins. This extends to the

413

amino acid sequences of their metal-binding sites, glycosylation sites if the lectins are glycoproteins, hydrophobic binding sites, etc. 3.1. I. Concanavalin A

Without doubt, concanavalin A (Con A) is the most celebrated and has proven to be one of the most useful of the plant lectins. Its physical chemical properties and carbohydratebinding properties are well documented in previous reviews [3,8]. Suffice it to note that it was first isolated and crystallized by Sumner and Howell in 1936, who showed it to require metal ions for its activity [74]. By virtue of its interaction with branched a-D-glucans, it is readily prepared by affinity chromatography on crossed-linked dextran (Sephadex) [75, 761. A homotetramer at pH 7 (subunit M , = 26 500 Da) of Con A has been sequenced [77] and its crystal structure determined both in its native form [38,39] and complexed with methyl a-D-mannopyranoside [40] and Man(a l-3)[Man(a 1-6)IMan [49]. The carbohydrate-binding specificity of Con A has been studied in great detail by every conceivable technique. It binds D-glucose, D-fructose, D-mannose, N-acetylD-glucosamine and related monosaccharides [3,8]. The a-anomer of D-mannose is the monosaccharide most complementary to the Con A sugar binding site. The hydroxyl groups most critical for binding to the lectin are those at positions C-3, C-4 and C-6 of the D-pyranose ring system. The manno-configuration, with an axial hydroxyl group at C-2 binds to Con A with an affinity about four times greater than the D-ghco-configuration. Certain sugars, e.g. the aldopentose D-arabinose and the ketohexose D-fructose also bind to Con A [3,8]. These sugars in their five-membered furanoid ring form, are similar in their orientation of their hydroxyl groups at C-2, C-3 and C-5 (arabinofuranose) and C-3, C-4 and C-6 (fiuctofuranose) as the orientation of the hydroxyl groups at C-3, C-4 and C-6 of the mannopyranosyl ring [3,8]. Examination of a large number of oligosaccharides indicated only those containing non-reducing, terminal a-D-glucosyl, mannosyl, or N-acetylglucosaminyl end groups bind to the lectin. A notable exception was sophorose (Glc b1-2 Glc) [78], whose binding is attributed to the free hydroxyl groups at the C-3, C-4 and C-6 positions of the reducing glucose residue group of the disaccharide [78]. a-Mannosyl residues substituted at their C-2 position also bind to Con A [79]. These units are common constituents of glycoproteins. A recent and most important discovery was that the branched trisaccharide Man(a13)[Man(a l-6)]Man, as occurs in all N-asparagine-linked glycoproteins, binds to Con A with high affinity [80-821. Similarly, the high affinity binding of Con A to branched, high mannose oligosaccharides and glycopeptides, is via the Man(a 14)[Man(a 13)lMan trimannosyl moiety of the a-1,6-antenna of these carbohydrate chains [83]. This observation forms the basis of many of the procedures for the isolation of glycoproteins on columns of immobilized Con A. Several classes of non-polar binding sites also are present on the Con A molecule. Two of these, a site adjacent to the carbohydrate-specific site that interacts with phenyl b-glycosides of mannose, glucose and N-acetylglucosamine [84], and 4 low affinity sites (one per subunit) that bind various non-polar molecules such as tryptophan and indoleacetic acid [85], are of special interest.

414

3.1.2. Pea lectin The pea lectin (Pisum satiuum) is a tetramer composed of two light a-chains and two heavy &chains bound by non-covalent forces [86]. Actually, the pea lectin occurs as two very closely related isolectins with common 0-chains and different but homologous a-polypeptide chains. The X-ray crystal structure of the pea lectin has been determined at 2 A [48,49]. The molecule is peanut-shaped and strikingly similar to the Con A dimer with a high content of 0-structure. The heterotetramer contains two non-interacting binding sites per molecule ~71. Carbohydrate-binding studies indicate a specificity similar to Con A: the pea lectin binds mannose, glucose, fructose and L-sorbose with methyl a-mannopyranoside being the most potent inhibitor of phosphomannan-lectin precipitation [88]. In contrast to Con A, which has a strict requirement for an unmodified C-3 hydroxyl group, the pea, lentil, and Viciafaba lectins bind 3-0-methyl and 3-0-benzyl derivatives of glucose 10 times more avidly than the free sugar [89]. Of even greater interest, methyl 2J-di-O-methyl a-D-glucopyranoside is 18 times more potent as an inhibitor than the parent glucoside, suggesting the presence in these lectin proteins of a non-polar region in proximity to these hydroxyl groups [89]. A fundamental finding revealed that the high affinity binding of N-linked glycopeptides to the pea (and lentil) lectins requires the presence of an a-L-fucosyl residue attached to the asparagine-linked N-acetylglucosamine residue [90,95].

3.1.3. Lentil lectin The lentil lectin (Lens culinaris) occurs as two isolectins termed LcH-A and -B [91] (or I and I1 [92]). The two proteins consist of two a- and two fbchains, differ in electrophoretic mobility, and have an apparent M , = 46 000 Da. They are both pure metalloproteins requiring Ca2+ and Mn+' for activity. Stein et al. [93] determined two binding sites per molecule LcH-A with K , = (100f24) M-' for methyl a-glucoside, a rather low binding constant. As with other lectins in this group, the lentil lectins bind primarily to the a-anomer of mannose. The hydroxyl groups at C-3, C-4 and C-6 are important for lectin binding, although some latitude at the C-2 position is tolerated, e.g. 2-deoxy-~-arubino hexose also binds to the lectin. Similar to the pea lectin, Allen et al. [89] noted that 3-0-methyl and 3-0-benzyl glucose were better inhibitors than glucose. These studies have been extended to include p-nitrobenzyl derivatives [94]. The presence of an a-L-fucosyl group attached to the asparagine-linked N-acetylglucosamine residue is an absolute requirement for the binding of glycopeptides and oligosaccharides of the poly-N-acetyllactosamine type to the lentil lectin [90,95]. 3.1.4. Fauin Favin, the lectin present in the fava bean (Viciu fuba), is a tetrameric glycoprotein composed of two a-chains ( M , = 5571 Da) and two 0-chains ( M , = 20 700 Da) united by non-covalent forces [96,97]. Amino acid analyses reveal high proportions of acidic and hydroxylic amino acids, substantial amounts of hydrophobic amino acids, and the absence of cysteine and methionine [96,97]. Both the a- and p-chains have been

415

sequenced [72,73]. The X-ray crystallographic structure of favin has been solved at 2.0A [47]. The amino acid sequences of the a- and @-chains of ! F faba lectin are homologous with Con A as represented by the circular permutation of amino acid sequences as alluded to for the pea and lentil lectins [71-731. This includes amino acid residues involved in metal and carbohydrate binding [72,73]. The fava bean lectin exhibits a carbohydrate-binding specificity similar to that of the pea and lentil: it shows a specificity toward mannose (4 times more potent than glucose) with methyl a-mannopyranoside and a,a’-trehalose being especially strong inhibitors [98,99]. Although similar to Con A in its carbohydrate-binding specificity, favin differs from the jack bean lectin in that it recognizes 3-0-methyl, benzyl and p-nitrobenzyl a-methyl glucoside and mannoside [97,100]. Additionally, it was noted by Debray et al. [95] that the presence of an a-1,6-linked L-fucosyl group on the N-acetylglucosamine residue linked to asparagine of N-linked carbohydrate chains enhances the binding of favin to its carbohydrate receptors. 3.2. Monocotyledonous mannose-binding lectins

A novel class of monocotyledonous bulb lectins has been isolated and characterized from the families Amaryllidaceae: snow drop (Galanthus niualis; GNA), daffodil (Narcissus pseudonarcissus; NPA), amaryllis (Hippeastrum hyb.; HHA) and Liliaceae: garlic (Allium satiuum; ASA), ramsons (Allium ursinum; AUA), shallot (Allium ascalonicum; AAA), and tulip (Tulipa gesneriana). Also included in this group are agglutinins from the Orchidaceae family (Listera ooata, LOA; Epipactis helleborine, EHA; and Cymbidium hyb.) [23,109]. This group of lectins is unique in that they all bind D-mannose and its derivatives but are inactive toward D-glucose and N-acetyl-D-glucosamine. This property distinguishes these lectins from the legume mannose/glucose-binding lectins (e.g. Con A and lectins from the pea, lentil, Viciafaba). They are readily isolated on columns of immobilized mannose and occur as dimeric and tetrameric proteins [23,101-1111. Additionally, they all occur in multiple forms (isolectins). Not surprisingly, there is extensive amino acid sequence homology among the bulb lectins; this is also expressed in cross-reactivity reactions using polyclonal antibodies raised against the lectins. Table 3 presents the carbohydrate-binding specificity of several monocotyledonous bulb lectins. In all cases, specificity is directed toward mannose oligosaccharides. All the bulb lectins form specific precipitation curves with a-mannans (e.g. yeast mannans), but unlike the legume mannose/glucose-binding lectins, do not precipitate a-glucans such as glycogens, amylopectins, and dextrans. Hence, the bulb lectins can serve as valuable tools for the isolation and characterization of a-linked mannooligosaccharides, a-mannans, and high mannose glycoproteins. Examples include the single step purification of murine IgM, human a-macroglobulin, and yeast mannan [ 11I]. Furthermore, yeast a-mannan has been shown to be resolved from glycogen on immobilized Listera ouata agglutinin [ 1091. A further solely mannose-binding lectin was isolated from a Nigerian legume Bowringia mildbraedii Harms [ 1 121. Its specificity is for “oligosaccharides bearing the sequence Man(al-2)Man(a 1-6)Man(a1-6)Man(a1-6), for example Mang GlcNAc and

416 Table 3 Inhibition by various sugars of six mannose-specific lectin-yeast mannan precipitation systems" Relative inhibitory potencyb

Sugar

Man Man a-OMe

ASAC

AUAC

GNAC

NPAC

HHAC

1 .o

1.5

AAACzd

1 .o

I .o

1 .o

1 .o

1 .o

1.3

1.6

1.2

1.5

2.9

0.3 2. I

0.2

0.5

1.7

3.3

3.2

12.1 14.2

2.8 3.1

5.9

Man B-OMe Man a 2 Man

K0.4 <1.4

<3.6

Man a 3 Man

11.5 11.9 11.5

<7.2

Man a 4 Man a-OMe

4.4

K3.2

I .9

<< 1.4

5.1 6.9

4.3 1.9

<0.7 5.1 0.7

<2.0

Man a 6 Man a-OMe Man a 3 Man a6 Man

8.3 20.0

10.7

<<3.6

11.3

28.3

3.8

13.8

so

Man a 3 Man a-OMe Man a 3 Man a-0-ally1

Man a 6

10

10.5

\ Man a-OMe

I

20.0

<7.2

Man a 3 a

After ref. [ 1061. No inhibition by l00mM Glc, Glc a 4 Glc, and Gal. Abbreviations: ASA, garlic; AUA, ramsons; GNA, snowdrop; NPA, daffodil; HHA, amyrillis; AAA, shallot Asialofetuin precipitation system [ 1071.

Man7 GlcNAc" [ 1121. Other mannose-specific lectins recently have been reported [ 1131. Interestingly, several of these mannose-specific lectins exhibit inhibitory activity toward HIV infectivity [ 113-1 151. It will be noted that the snowdrop lectin (GNA) is specific for Man(a1-3)Man linkages and exhibits high affinity for the branched trisaccharide Man(al-3)[Man(al6)lMan [ 104,l lo]; the daffodil lectin (NPA) for a-1,3- and a-1,6-linked mannose disaccharides [ 1051; the amaryllis lectin (HHA) for a - 1,6-linked mannose oligosaccharides [105]; and the garlic (ASA) and ramsons (AUA) lectins for Man(a1-3)Man units [106]. Similar to the snowdrop lectin, the orchid lectin (Listera ouata) is highly specific for a-1,3-mannose oligomers [ 1091. Of considerable interest was the finding that several of these mannose-specific bulb lectins could generate typical precipitation curves with a linear mannopentaose indicating their ability to interact with internal a-mannosyl groups [ 1081 (c.f. Con A). In fact, the amaryllis lectin gave a precipitation reaction with the reduced mannopentaose, thus indicating its ability to interact with two sites on this small oligosaccharide [108]. 3.3. N-acetyl-D-glucosarnine-bindinglectins

The N-acetyl-D-glucosamine-binding lectins comprise a diverse group of lectins which, except in a few cases, recognize sequences of B-1,4-linked GlcNAc. The exceptions include the MadGlc-binding lectins which exhibit a secondary, low binding affinity

417

for terminal, non-reducing a-GlcNAc units, and Grzfonia simplicifolia I1 which also recognizes terminal GlcNAc units, but containing both a- or 13-anomeric glycosidic linkages. Included in this group are lectins from several plant families: Solanaceae (jimson weed, potato, tomato), Gramineae (barley, rice, rye, wheat), Urticaceae (stinging nettle) and Leguminosae as mentioned above and also including those from Cytisus sessilifolius and Ulex europaeus 131. 3.3.1. Wheat germ agglutinin Wheat germ agglutinin is the most studied of these lectins and also one of the most useful in its biomedical application. The lectin is a homodimer composed of subunits of M , = 23 600 Da which dissociates into monomers at acid pH. It is a pure, metal-free protein devoid of carbohydrate residues, which is isolated as a mixture of four isolectins differing in electrophoretic mobility [ 116,1171. The lectin has been sequenced (1 71 amino acids per polypeptide chain) and its X-ray crystallographic structure determined at a 2.0 A resolution [ 1 18,1191. Each protomeric unit of wheat germ agglutinin consists of 4 structurally homologous domains with a high degree of amino acid sequence homology among the 4 domains. Four interlocking disulfide bonds, each within one of the 4 domains, result in a highly compact stable protein [ 1 181. The carbohydrate-binding specificity of wheat germ agglutinin is directed against sequences of fi-1,4-GlcNAc-linked residues - the chitodextrins [ 1 16,1201. Each monomer contains two identical, non-interacting binding sites which are complementary to 3 or 4 fi-1,4-GlcNAc units [ 1211. Of the monosaccharides examined, only GlcNAc binds to WGA; ManNAc does not bind and GalNAc binds only weakly. Interestingly, it was observed that sialic-acid-containing glycoconjugates and oligosaccharides appeared to interact with WGA. This was first reported by investigators who found that sialidase-treated animal cells had reduced or abolished binding with the lectin [ 120,1221. Hapten inhibition studies indicated that N-acetylneuraminic acid and its derivatives were weak inhibitors of the lectin. Similarity in configuration of sialic acid to GlcNAc at their N-acetamido groups (positions C-5 and C-2, respectively) and adjacent hydroxyl groups (positions C-4 and C-3) account for this recognition by WGA [123,124]. Complexes of WGA with sialic-acid-containing oligosaccharides and glycopeptides have been analyzed by Wright [ 125,1261 using X-ray crystallographic analysis. She showed that only two molecules of sialyllactose bind to the WGA dimer. In an important study, Wright also determined the X-ray crystallographic structure of WGA bound to a tryptic sialoglycopeptide from its erythrocyte glycophorin A receptor [ 1271.

3.3.2. Tomato lectin Tomato lectin (Lycopersicon esculentum), also a chitin-binding agglutinin is a glycoprotein composed of approximately equal amounts of protein and carbohydrate (85% L-arabinose, 15% galactose) [ 128,1291. The lectin, a single glycosylated polypeptide chain, has M , = 71 000 Da, is rich in hydroxyproline (1 6%) to which are glycosidicallylinked chains of L-arabinose, and lacks leucine and valine [ 128,1291. Tomato lectin binds 3 or 4 p-1,4-linked GlcNAc units [128,129].

418

3.3.3. Potato lectin Potato lectin (Solunum tuberosum) is a dimeric glycoprotein composed of two identical subunits ( M , = 50000 Da) which are linked by non-covalent forces to give an aggregate M , = 100 000 Da [ 130,1311. Physical chemical studies indicate the lectin to be a long cylindrical molecule. The lectin may be considered to be comprised of two very dissimilar domains: one domain contains all of the hydroxyproline and all of the carbohydrate and the remainder of the molecule all of the cysteine as well as the carbohydratebinding sites [ 1311. Structural studies of the carbohydrate residues indicate chains of 3 or 4 L-arabinose units linked to hydroxyproline. The molecule also contains single a-galactosyl groups bound to serine. It was determined that each potato lectin monomer contains a single binding site for N-acetylated chitotriose and chitotetraose [ 1321. Carbohydrate-binding specificity studies indicate N-acetylated chitotetraose and chitopentaose to be the best inhibitors: these oligosaccharides are 20- and 50-fold more potent, respectively, than NN‘-diacetylchitobiose [ 130,1331. 3.3.4. Cytisus sessilifolius Extracts of Cytisus sessilifolius exhibit anti-H(0) and anti-A2 activity [ 1341. The lectin was purified on immobilized N,N”N”-triacetylchitotriose [161]; it has M , = 1 l0000Da. N,N‘-diacetylchitobiose was the most potent inhibitor tested although cellobiose and laminaribiose Glc@1-3)Glc were also good inhibitors, whereas gentiobiose and sophorose (Glc(b1-2)Glc) were poor, as were all a-linked disaccharides [ 1351. 3.3.5. Datura strumonium lectin Datum stramonium seeds contain a chitin-binding lectin that agglutinates human erythrocytes non-specifically. It is composed of two non-identical subunits ( M , = 40 000 and 46000Da) joined by disulfide bonds [ 136,1371. Molecular mass estimates vary considerably due to its high carbohydrate content (37%) consisting of galactose (7.5%) and L-arabinose (93%). The lectin also contains high amounts of hydroxyproline (1 l%), half-cystine (16%), and glycine (12%). Carbohydrate-binding studies indicate the Datura lectins binding site is most complementary to the per-N-acetylated chitodextrins with the triose and tetraose being 560 and 1000 times, respectively, more inhibitory than the biose N,N‘-diacetylchitobiose [ 138,1391. Most dramatically, the branched pentasacwas almost three orders charide Gal(~I-4)GlcNAc(~1-6)[Gal(~l-4)GlcNAc(~l-2)]Man of magnitude more active than N,N’-diacetylchitobiose [ 1391. The lectin also recognizes Gal@1-4)GlcNAc sequences; hence polylactosamine-containing glycopeptides, and oligosaccharides may be isolated on the immobilized lectin [ 140,1411. 3.3.6. Grifoniu (Bandeirueu) simplicifolia II lectin The seeds of Griffoniu (Bundeiruea) simplicifolia contain a second lectin - GS 11. It is a tetramer composed of four apparently identical subunits ( M , = 30 000) with an aggregate, M , = 113 000 Da by gel filtration [142]. The lectin is a glycoprotein with three cysteine residues per subunit; a disulfide bond links two subunits [142]. Each subunit contains a binding site for GlcNAc with K , = 1.3x lo5 M-’ at 37°C [ 1431. GS I1 does not agglutinate A, B or 0 erythrocytes but will agglutinate acquired-b, T-activated and Tk polyagglutinable cells [ 142,1441.

419

Carbohydrate-binding specificity studies indicate a strict specificity for GlcNAc and its derivatives: a preference exists for the a-anomer and GlcNAc a-2,6-linked to a second saccharide which is more active than the a-2,3-linked isomer. In this regard, GS I1 resembles the lectins from Laburnum alpinurn and Cytisus sessilifolius [145]. Interestingly, taking advantage of its slight reactivity with Glc, GS I1 has been used as a histochemical stain for glycogen [146]. 3.4. N -acetylgalactosamine/galactose-binding lectins

This group of lectins, among the earliest to be investigated among plant agglutinins, vary in their specificity from those which are primarily GalNAc-specific, to those which exhibit little specificity toward either sugar, to those which are most specific for Gal. 3.4.1. Dolichos bijlorus lectin The lectin present in horse gram (Dolichos biforus) seeds is a blood-group-A-specific lectin. Bird [ 1471 first reported seed extracts agglutinated human type A, erythrocytes 1O-fold more strongly than A2 cells. Purified by affinity chromatography on immobilized polyleucyl hog A + H substance [ 1481, the Dolichos bijlorus lectin consists of a mixture of isolectins. A tetrameric glycoprotein, the D. biflorus lectin is composed of two similar but distinct subunits: I ( M , = 27 700 Da) and I1 ( M , = 27 300 Da) with an aggregate M , = 1 l0000Da [149]. The D. bzjlorus lectin consists of two isolectins, A and B, separable by fractionation on concanavalin A-Sepharose [ 1501. Each isolectin consists of distinct subunits: IA and IIA in form A and IB and IIB in form B. The first 30 NHz-terminal amino acids of IA and IIA are identical but differ at their carboxyl ends [151]. The D. bzjlorus lectin has been cloned, thus establishing its primary structure [ 1521. By equilibrium dialysis, it was established that the D. biflorus lectin has two carbohydrate binding sites per molecule for N-acetylgalactosamine. Similar to other legume lectins, the D. biforus lectin is a metalloprotein, requiring Ca2+for its activity. Carbohydrate-binding studies reveal that the lectin is specific for the a-anomeric form of GalNAc [146]. Interestingly, many years later it was shown that the Forssman disaccharide GalNAc(a1-3)GalNAc and the corresponding Forssman pentasaccharide, were 36- and 63-fold more potent inhibitors than GalNAc [153]. The Dolichos bgorus lectin has been used extensively to probe for a-GalNAc-terminated residues in animal cells and tissue. 3.4.2. Lima bean lectin The lima bean lectin (Phaseolus lunatus) was the first agglutinin shown to possess blood group specificity [ 1541. The purified lectin, a glycoprotein, was prepared by salt fractionation and gel filtration [ 1551, by specific adsorption to insolubilized type A blood group substance, followed by elution with GalNAc [ 1561 and by affinity chromatography on the immobilized type A trisaccharide [ 1571. The lectin occurs in varying states of aggregation. It is composed of 30 000 Da subunits, two of which are joined by a disulfide bond; two such dimers constitute the lectin of M , = 124 000 Da whereas 4 dimers lead to the 247000Da form. The lectin is a metalloprotein requiring Ca2+ and Mn2+ for its

420

carbohydrate-binding activity. Each subunit contains a cysteine residue which is required for both its carbohydrate- and metal-ion-binding activities. A peptide containing the essential sulfiydryl group was isolated by Roberts et al. [ 1581, and shown to be homologous with corresponding peptides present in Con A, the pea, lentil, and Vicia faba lectins. The primary structure of the lectin was established by cloning the gene encoding the lima bean lectin [159]. Each subunit contains a single GalNAc binding site [160]. Additionally, the lima bean lectin contains two classes of hydrophobic binding sites: each subunit has a low affinity site which binds fluorescence probes such as 1,8-aniIinonaphthalenesulfonicacid [ 1571; and a single high affinity site per lectin tetramer, which binds adenine and its N6 derivatives related to cytokinins [ 1601. The carbohydrate-binding specificity of the lima bean lectin studied by inhibition of precipitation, and a sulfiydryl group protection assay [161], revealed the type A trisaccharide GalNAc(a 1-3)[~-Fuc(al-2)]Gal to be the best inhibitor, -40 times more potent than GalNAc. The a-glycosides of GalNAc are -8 times more potent than the corresponding 6-anomers. 3.4.3. Soybean agglutinin Soybean meal contains a lectin which agglutinates protease-treated or transformed animal erythrocytes but not normal ones [ 162,1631. First prepared by conventional protein purification techniques, the soybean agglutinin is now most commonly purified by affinity chromatography on immobilized GalNAc columns [ 164,1651. Comprehensive biophysical characterization revealed the soybean agglutinin to be a tetrameric glycoprotein [ 1661 with each subunit consisting of 153 amino acids [ 1671 having M , = 30000 Da. Similar to other legume lectins, soybean agglutinin is rich in acidic and hydroxylic amino acids, is devoid of cysteine [ 1661, and possesses a high content of 6-pleated sheet; it also exhibits extensive amino acid homology with other legume lectins. Carbohydrate-binding studies indicate a preference for N-acetylgalactosamine in its a-anomeric form; GalNAc a-methyl glycoside is 30 times more potent than the corresponding a-galactoside [ 1681. The most potent inhibitor described thus far is the type A disaccharide GalNAc(a13)Gal. However, the addition of an a-L-fucosyl group to the Gal, as occurs in the type A trisaccharide, greatly diminished its activity [ 1681. A hydrophobic site on soybean agglutinin was identified by its interaction with N-dansylgalactosamine [ 1691. The soybean lectin possesses four carbohydrate binding sites per molecule, one for each subunit 61691. 3.4.4. Erythrina lectins Seed extracts of a large number of Erythrina species contain proteins of the Gal/GalNAcspecific group of lectins. A large number of these lectins have been purified and characterized. They are glycoproteins composed of two non-covalently linked subunits and have aggregate molecular masses in the range of 5 6 0 0 0 4 8 000 Da. Not surprisingly, these leguminous seed lectins are also metalloproteins requiring Ca2+ and Mn+* for their carbohydrate-binding activity, contain no cysteine and low methionine and

42 1

exhibit extensive N-terminal amino acid homology. Equilibrium dialysis studies [ 170,17 I] indicate two binding sites per molecule. Carbohydrate-binding studies showed a somewhat greater specificity of GalNAc over Gal at the monosaccharide level. Of much greater interest however, was the finding that N-acetyllactosamine was 2&50-fold more active for many of the Erythrina species studies [ 172,1731. A pentasaccharide with two terminal LacNAc end groups, Gal@4)GlcNAc(fil-2)[Gal(fi I-4)GlcNAc(fi 1-6)]Man, had three times greater activity than the disaccharide. 3.4.5. Peanut lectin The peanut (Arachis hypogaea) contains a lectin with anti-T (Gal(fi1-3)GalNAc) activity [ 1741. This antigen appears on human erythrocytes following treatment with sialidase and leads to the phenomenon known as polyagglutinability as monitored by the peanut lectin [ 1751. Peanut agglutinin, purified by numerous affinity purification schemes, is a tetrameric protein composed of four carbohydrate-free subunits, M , = 27000 Da [176]. The lectin is a metalloprotein rich in acidic and hydroxylic amino acids and devoid of cysteine [ 1761. Uhlenbruck et al. [177] were the first to establish the disaccharide Gal(fi1-3)GalNAc to be a very active inhibitor of hemagglutination. GalNAc was twice as good an inhibitor as Gal [178]. In view of the fact that peanut agglutinin recognizes the disaccharide Gal(fil-3)GalNAc, it is not surprising that the antifreeze glycoprotein from antarctic fish which. contains Gal@ 1-3)GalNAc(a 1-0)ThrAlaAla repeating units was the best precipitating agent of the lectin [ 1781. 3.4.6. Maclura pomifera lectin The Maclura pomifera lectin, present in the seeds of the Osage orange fruit, occurs as 5 isolectins; each is a tetrameric protein composed of two different subunits ( M , = 10000 and 12 000 Da) [ 179,l SO]. The isolectin proteins are devoid of carbohydrate and metal ions; they are rich in aspartic acid, glycine and hydroxylic amino acid, but lack cysteine and contain only a single methionine residue per polypeptide chain. More recent work indicates that, like the Artocarpus integrifolia lectin from the same family (Moraceae), the lectin also contains a small subunit of 20 amino acids [24]. The Maclura lectin was first shown to be inhibited by N-acetylgalactosamine, galactose and melibiose by Cawley et al. [ 18 11. It exhibits high specificity toward Gal(fi1-3)GalNAc and glycosides of GalNAc, whereas GalNAc(a 1-3)Gal and GalNAc(a 1-3)GalNAc are poor inhibitors of the lectin [182]. 3.4.7. Winged bean lectin The winged bean (Psophocarpus tetragonolobus) contains two lectins: a basic lectin (WBA I) with p l 10 and an acidic lectin (WBA 11) with p l 5.5 [183-1851. The basic agglutinin (WBA I) consists of two identical subunits M , = 29 000 Da and strongly recognizes human A, blood groups substance and moderately A2 and B substances [ 183,1861. The primary structure of the lectin has been determined [ 1871. Carbohydrate-binding studies reveal WBA I to display a specificity for a-GalNAc and secondarily for a-Gal. The A-trisaccharide (GalNAc(aI--3)[~-Fuc(a1-2)lGal) and its extended derivatives were

422

shown to be most complementary to the lectins carbohydrate binding site [188,189]. The 2-dansyl derivative of GalN was strongly bound to WBA I ( K , = 4.17 x 1O5 M-' ) due to a positive entropic contribution to binding [ 1881.

3.4.8. Jack fruit lectin The jack fruit, Artocarpus integrifolia, contains two isolectins which specifically recognize the T-antigen Gal(f31-3)GalNAc [190-1921. It is a glycoprotein composed of two different types of polypeptide chains with M , = 16200i~1200and 2090f300Da [193,194]. The primary sequence of the two chains has been determined and the heavy chain has been shown to possess an internal repeat spanning residues 7 4 4 and 76130 [ 1931. The Artocarpus lectin, known as jacalin, binds a-D-galactosyl groups as well as the Thomsen-Friedenreich T-antigen Gal(f31-3)GalNAc [ 19 1,1921. It is especially noteworthy that the A. integrifolia lectin binds to and precipitates both serum and secretory immunoglobulin A [ 195,1961. 3.4.9. Castor bean lectins The castor bean (Ricinus communis) contains two lectins: a highly cytotoxic, weak agglutinating protein, and a strong erythroagglutinin [ 197,1981. Although structurally related, the two lectins are quite distinct in their biological properties. The cytotoxic component (termed ricin, ricin D, E and RCA 11) is a heterodimer ( M , = 63 000 Da) [ 1991 composed of a B chain which binds 2 moles of GalIGalNAc and an A chain that catalyzes the cleavage of an adenine residue from ribosomal RNA resulting in inhibition of protein synthesis [200-2021. The two chains are linked by a single disulfide bond. RCA I, the erythroagglutinin, has proved to be a useful reagent for the detection of terminal non-reducing f3-galactosyl groups of polysaccharides, glycoproteins and glycolipids. The lectin, a tetramer ( M , = 120000Da), is also composed of A and B chains. The tetramer binds two moles of saccharide. Carbohydrate-binding studies show that lactose is bound 6 to 7 times more effectively than either galactose or melibiose [203]. The snail lectin, Helix pomatia, although of animal origin, is a useful reagent for the determination of a-N-acetylgalactosaminyl end groups [204]. 3.4.10. Griffonia (Bandeiraea) simplicifolia I lectin Griffonia (Bandeiraea) simplicifolia lectin I (GS I) was first reported to exhibit anti-blood group B activity [205]. The lectin was isolated on a-galactosyl-substituted matrices [206,207] but later was shown to consist of five tetrameric isolectins termed Ad, A3B, A2B2, AB3, and B4. The B subunit (M,=33000Da) was shown to be highly specific for a-galactosyl groups and as GS I-B4 to agglutinate type B erythrocytes, whereas the A subunit ( M , = 32000 Da) recognizes N-acetylgalactosaminyl end groups and agglutinates type A erythrocytes [208]. The GS I-B4 isolectin is especially useful, being the only truly a-galactosyl-specific lectin. It was first utilized to demonstrate the presence and characterization of a-galactosyl-containing glycoproteins on the surface of Ehrlich ascites tumor cells [209]. The occurrence of a-galactosyl end groups in murine laminin [210] and thyroglobulin [211] followed.

42 3

3.5. L-Fucose-binding lectins Interestingly, there are fewer L-fucose-binding lectins than any other group of plant agglutinins. They vary in specificity from recognition of only a portion of the methyl pentose ring structure - the eel hemagglutinin, not discussed here - to Griffonia simplicifolia IV which binds Leb blood group tetrasaccharide and related structures.

3.5.1. Asparagus pea lectin The Lotus tetragonolobus (Asparagus pea) lectin was discovered by Renkonnen in 1948 [212] as a type 0 human erythrocyte agglutinin. Morgan and Watkins [213] showed that L - h o s e inhibited the agglutination of type 0 cells by the Lotus lectin; blood group H, but not A- or B substance also inhibited the agglutination reaction. The lectin has been purified by affinity precipitation and chromatography [2 14-2 161. The L. tetragonolobus lectin occurs as three isolectins (A, B and C). Lectins A and C are tetramers, whereas B is a homodimer. All three isolectins lack cysteine, contain two residues of methionine per subunit, and are rich in aspartic acid and hydroxylic amino acid[217]. They also are metalloproteins requiring metal ions for their carbohydrate-binding activity. Several useful studies compare the carbohydrate-binding specificity of L-fucose-binding lectins. They include a table comparing the relative inhibitory potency of saccharides toward several L-fucose-binding lectins (Lotus, Ulex I, Aleuria aurantia (orange peel fungus), as well as the eel lectin), in a monograph on lectins [218], a report by Petryniak and Goldstein [219], and one by Konami et al. [220]. The a-methyl glycoside of L-fucose is ten times more potent an inhibitor of the Lotus lectin than the 6-methyl glycoside, or the free sugar. Interestingly, the Lotus lectin tolerates methoxyl groups at C-2 and C-3, but not at C-4 or C-5 [221]. Structures similar to the H-active type 2 chains of blood group substance such as ~-Fuc(al-2)Gal(614)Glc are good inhibitors [2 161. However, additional substitution on the galactosyl residue as occurs in the type-A- and -B-active trisaccharides drastically reduces their inhibitory potency [2 161. Lotus lectin will bind ~ - F u c ( 1-6)GlcNAc a but binds very poorly to GlcNAc(fi14)[Fuc(a l-6)]GlcNAc indicating that the lectin has limited ability to recognize fucose linked to subterminal sugars in an oligosaccharide [218]. 3.5.2. Ulex europaeus lectins Ulex europaeus seeds contain two agglutinins: an L-fucose-binding lectin (c!europaeus I, UEA I) and c! europaeus I1 (UEA 11), which is most potently inhibited by N-acetylated chitodextrins and L-Fuc(a 1-2)Gal(fi1-4)GlcNAc. UEA I has been isolated by affinity chromatography on immobilized L-fucose columns [215]. Although originally believed to be a dimer of two dissimilar polypeptide chains, it has been shown to be composed of two identical subunits containing 243 amino acid residues ( M , = 26 669 Da) held together by non-covalent interactions. The lectin is a metalloprotein requiring Ca2+ and Zn2+ or Mn2+ for activity. UEA I has been sequenced and shown to exhibit extensive homology with many other legume lectins, but not with Lotus tetragonolobus, another L-fucose-binding lectin. Carbohydrate-binding studies indicate strong preference for 2-0-linked a-L-fucosyl units, the best trisaccharide inhibitor of the lectin being ~ - F u c ( 1-2)Gal(61-4)Glc. a

424

The ~-Fuc(al-3), L-Fuc(a1-6)Glc and ~-Fuc(al-2)Gal(fi14)Glc are non-inhibitors. In contrast to the Lotus lectin which exhibits increased reactivity toward difucosylated oligosaccharides, Ulex I shows reduced activity toward these substances.

3.5.3. Trichosanthes juponica lectin The root tuber of Trichosanthes japonica contains two lectins: a sialic-acid-binding lectin (TJA I) and a lectin (TJA 11) which recognizes L-fucose-terminated oligosaccharides [222]. TJA 11, purified by affinity chromatography on immobilized porcine stomach mucin, consists of two different subunits ( M , = 29 and 33 kDa) held together by disulfide bond(s) [222]. The lectin is a glycoprotein with two binding sites toward Fuc(a 1-2)Gal(B 1-3)GlcNAc(fi l-3)Gal(@ 14)Glc OT. The carbohydrate-binding activity of the TJA I1 lectin, studied by inhibition of agglutination and passage of tritiumlabeled oligosaccharides through a column of the immobilized lectin, indicated that the lectin “fundamentally recognizes a fi-galactosyl residue and the binding strength increases on substitution at the C-2 position with a fucosyl or acetamido group” [222]. The fundamental receptor for TJA I1 is Fuc(al-2)Gal(fil-3/4)Glc/GlcNAc-. Fuc(a 12)Gal(fi14)Glc is bound 260 times better than Gal; L-FUCdoes not bind to the lectin. 3.5.4. Griffonia simplicifolia IV lectin The Griffonia simplicifolia IV (GS IV) lectin is a heterodimer composed of subunits of M , = 27 000 and 29 000 Da [223]. The lectin, a glycoprotein, reacted most strongly with H and Leb blood group substance [223,224]. GS IV has been crystallized and its X-ray structure determined [5 11. Although of animal origin, the lectin from the eel (Anguillu anguilla) is an interesting and useful reagent with anti-O(H) activity [225]. It is specific for a-L-fucosyl groups. 3.6. Lectins with complex binding sites 3.6.1. Red kidney bean lectin The red kidney bean Phaseolus vulgaris, contains five isolectins composed of two similar and different subunits, E and L [226,227]. These isolectins are designated E4, E3L, E2L2, EL, and L4 [226,227]. The E subunits are erythroagglutinating whereas the L subunits bind to lymphocytes and are mitogenic. Hence E4 is the strongest hemagglutinin, whereas L4 exhibits strong leukoagglutinating and mitogenic activity. Carbohydratebinding specificity studies indicate complex binding sites for both homotetramers. Employing glycopeptides and synthetic oligosaccharides it was shown that both the erythroagglutinin and leukoagglutinin bind to Asn-linked, complex-type oligosaccharides and that the galactose residues are important determinants of binding [228,229]. It was also shown that terminal sialic acid residues did not interfere with interaction with these lectins [229,230]. Using immobilized E4, Irimura et al. [229] and Cummings and Kornfeld [230] showed that diantennary complex-type Asn-linked oligosaccharides interact with high affinity with E4. These studies indicate that PHA-Ed binds to diantennary N-linked oligosaccharides containing two galactosyl residues, Furthermore, a bisecting GlcNAc residue as present on glycophorin molecules of human erythrocytes, does not interfere, but rather enhances binding.

425

In similar fashion, it was shown that PHA-L4 has a similar, but distinct carbohydratebinding specificity [230]: it fails to interact with (bisected) diantennary glycopeptides, but binds to tri- and tetraantennary glycopeptides “containing outer galactosyl residues and an a-linked mannosyl residue substituted at positions C-2 and C-6. Triantennary glycopeptides substituted at C-2 and C-4 are not reactive with immobilized PHA-L4agarose” [230]. Using synthetic oligosaccharides and a hapten inhibition of the precipitation assay, Hammarstrom et al. [23 11 reported that a pentasaccharide contains two N-acetyllactosamine disaccharide units linked through GlcNAc residues to a Man residue at position C-2 and C-6 was most complementary to the PHA-L4 binding site. Interestingly, Mirkov and Chrispeels[232] showed that Asn 128 was essential for the carbohydrate-binding activity of PHA-L inasmuch as site directed mutagenesis gave a mutant lectin (N 128D) which failed to bind carbohydrate, the leucoagglutinating and mitogenic activities of PHA-L were also abolished. Also employing immobilized lectin columns, and homogenous oligosaccharides released from Asn with N-glycanase, Green and Baenziger [233] observed that “in virtually all cases, L- and E-PHA yielded identical results. Both lectins retarded oligosaccharides bearing a-2,3- but not a-2,6-linked sialic acid.” Desialylated, di-, tri-or tetraantennary oligosaccharides were retarded to varying extends by both lectins, but subsequent removal of galactose decreased or eliminated binding. Furthermore, they as well as others [233,234] also reported that a bisecting GlcNAc residue attached to the b-linked core Man residues of Asn-linked oligosaccharides enhanced interaction with both lectins. Bierhuizen et al. [235] observed that it is essential to desialylate and to defucosylate glycans prior to application to L-PHA-agarose. 3.7. Sialic-acid-binding lectins Sialic acids are of widespread occurrence as components of the oligosaccharide chains of glycoproteins and glycolipids present in animal cells and tissues. Lectins which recognize sialic acid are valuable reagents for the detection, isolation and characterization of glycoconjugates containing this sugar (see 236-238 for reviews). Most such proteins occur in invertebrates such as crabs (e.g. limulin, the lectin from the horseshoe crab Limulus polyphemus), lobsters and slugs [236-2381. However, a few sialic-acid-binding lectins of plant origin have proved to be exceptionally useful. They are the elderberry bark lectin (Sambucus nigra) [240], Trichosanthes japonica tuber lectin I [240] and the Maackia amurensis leukogglutinin [35,241]. Wheat germ agglutinin (WGA) has also been shown to recognize NeuSAc-terminated oligosaccharide chains (see under N-acetylglucosamine-binding lectins for a discussion).

3.7.1. Sambucus nigra I lectin The Sambucus nigra I lectin (SNA), a tetrameric protein (M,= 140000Da), was first shown to bind primarily to NeuSAc(a2-6)Gal/GalNAc disaccharide sequences by Shibuya et al. [239]. This lectin has found important application in the distinction and separation of oligosaccharides and glycopeptides containing Neu5Ac(a2-6)Gal/GalNAc from the isomeric a-2,3-saccharide [242] e.g. oligosaccharides from milk, porcine thyroglobulin, human transferrin and bovine fetuin [242], human colorectal cancer

426

tissue [243] and Ehrlich ascites tumor cells [244,245]. Both a free carboxyl group at C-1 and the glyceryl side chain (C7-C9) of Neu5Ac are required for interaction with SNA I [239,242]. 3.7.2. Maackia amurensis leukagglutinin The Maackia amurensis leukagglutinin (MAL), a tetrameric protein composed of two different subunits (38 000 and 50 000 Da), binds the trisaccharide sequence NeuSAc(a23)Gal(~14)GlcNAc/Glcas found in murine laminin [246]. The lectin requires three intact sugars and also recognizes this sugar sequence if the N-acetyl group of Neu5Ac is replaced by an N-glycolyl group. 3.7.3. Trichosanthes japonica lectin The Trichosanthes japonica lectin recognizes Neu5Ac(a2-6)Gal(~14)GlcNAc trisaccharide sequences [247]. Stable, monovalent subunits of SNA and MAL have been prepared by reduction of disulfide bonds and alkylation with 4-vinylpyridine. These derivatives have proved to be useful reagents for the study of cell surface glycoconjugates by flow cytometry [248,249]. Both SNA and MAL have found wide application in the immunohistochemical localization of glycoconjugates containing Neu5Ac in a-2,3- and a-2,6-linkage to Gal and GalNAc. The elegant studies of Roth and his colleagues are especially noteworthy [25& 2521.

4. Lectin-based approaches for the isolation and characterization of

glycoconjugates The strong influence of the immunological/serological parentage of the lectin field is most evident in the applications of these proteins as reagents to detect and characterize polysaccharides and glycoconjugates. In considering lectins as reagents to study carbohydrates, they can be envisioned simply as (g1yco)proteins with one or more regions capable of reversibly binding a subset of saccharides [ 1,3]. In practice they may be viewed as being similar to IgG, though lacking an effector activity. Accordingly, many of the immunochemical techniques employing IgG as a specific reagent can be borrowed, almost in toto, for applications utilizing lectins. One important difference, however, must be emphasized. The structure of IgG is quite different from that of lectins, and unlike immunoglobulins, different lectins are structurally distinct. In practice this means that approaches and applications successful with one lectin may require modifications to optimize the use of techniques with another lectin. With the exception of chemical modification of the lectin, in most instances such variations are relatively minor. For example, though the pH optima of the lectins could differ, such as below pH 7 for Con A [253] and above pH 8 for S. japonica agglutinin [254], virtually all lectins exhibit significant carbohydrate-binding activity at pH values near neutrality. On the other hand, where succinylation of Con A results in conversion of the multimeric form of the lectin

427

to a divalent protein retaining carbohydrate-binding activity [255], similar treatment of other lectins, such as wheat germ agglutinin [256],does not have similar effects and may result in altered reactivity of the protein. Approaches for the application of lectins to study glycoconjugates can be subdivided on the basis of the physical state of the carbohydrate. Specifically, whether the glycancontaining substance is soluble or particulate (cell surface or solid phase) bound. In turn, different techniques may be applied for the detection, quantification or structural elucidation of the glycan. The following description of select methods utilizing lectins in glycobiology is designed to acquaint the reader with potential approaches and to direct the reader to relevant literature references. 4.1. Lectin-based reagents

The rapid expansion of lectin-based applications for the detection and quantification of glycoconjugates has been led by the development of commercially available, purified and chemically derivatized lectins, and in some cases, anti-lectin antibodies. Over 50 purified plant lectins are sold commercially by a number of producers and vendors, with this number growing annually. Equally important is the ease by which investigators can obtain lectins labeled with various fluorescent dyes, haptenic moieties, biotin, and radioactive atoms, as well as conjugated to enzymes and solid-phase supports. These derivatized lectins are useful for either direct or indirect detection and quantification techniques, or for the physical separation of particulate-bound or soluble glycoconjugates. Table 4 lists many of the commercially available lectin reagents and sources. 4.2. Soluble glycoconjugates

The techniques using lectin-based reagents to isolate or characterize water soluble glycoconjugates have followed two fundamental approaches: (1) to employ methods designed for soluble analytes; or alternatively, ( 2 ) to convert the soluble analyte to a solid-phase-bound glycoconjugate and employ techniques appropriate for such media. Quantitative precipitation analysis with soluble lectin reagents and affinity adsorptiodchromatography utilizing solid-phase-bound lectin are examples of the former method, and blot analysis, where the glycoconjugate is fixed to a solid matrix, is representative of the latter techniques. By far, affinity adsorptiodchromatography has become the most common method to isolate and/or characterize soluble glycans and glycoconjugates. However, the commercial availability of the appropriate reagents and supplies in kit formats, and convenience of use is certainly popularizing the blot methodology as a stand alone procedure to study pure glycoconjugates, or as a second dimension of analysis in conjunction with gel electrophoresis and HPLC to explore individual glycoproteins within complex mixtures of biopolymers. It is noteworthy to emphasize that quantitative precipitation analysis, lectin affinity chromatography, or lectin-based blot analysis is capable of supplying valuable structural information. Such information, however, is only inferred and must be weighed in light of structural evidence developed by alternative, preferably chemical, means. Furthermore, glycoproteins often contain multiple sites of glycosylation, each exhibiting microheterogeneity. Accordingly,

428 Table 4 Some commercially availabile plant lectin reagents Lectin

Purified lectin

Fluorochrome labeled

Biotin or hapten labeled

Protein conjugated

Gold conjugated

Immobilized

Abrus precatorius

AB a

AB

AB

B

B

AB

A egopodium podagraria

B

B

B

B

B

B

Agaricus bisporus

A

B

B

B

B

B

Aleuria aurantia

CD

D

CD

Amaranthus caudatus

BCD

BC

BCD

B

Arachis hypogaea

ABCD

ABC

ABCD

AB

AB

ABC

Artocarpus integrifolia

ABC

BC

ABC

AB

B

ABC

Bauhinia purpurea

ABC

ABC

ABC

B

B

B

Bryonia dioica Caragana arborescens

B

B

B

B

B

AB

AB

AB

B

B

Cicer arietinum Colchicum auiumnale

AB

B

B

B

B

B

B

B

B B B B

B

B

Canaoalia ensiformis

ABCD

ABC

ABCD

AB

AB

ABC B

C B

Cytisus scoparius

AB

B

B

B

B

Datura stramonium

ABCD

BC

BCD

B

B

BC

Dolichos biJorus

ABC

A BC

ABC

AB

AB

ABC

Erythrina corallodendron

A

A

A

Erythrina cristagalli

ABC

ABC

ABC

AB

B

BC

Euonymus europaeus

ABC

BC

BC

B

Galanthus nioalis Glycine inax Griffonia simplicifolia I

ABCD

BC

BCD

B

B

ABC

ABCD

ABC

ABC

AB

AB

ABC

ABC

ABC

ABC

AB

B

BC

Grvonia simplicifolia I A,

ABC

AC

C

B

C

Griffonia simplicifolia I B, Griffonia simplicifolia I 1

ABC

ABC

ABC

AB

B

B

ABC

ABC

BC

AB

B

ABC

Laburnum alpinum

B

B

B

B

Laburnum anagyroides

B

B

B

B

Lathyrus odoratus Lens culinaris Lotus tetragonolobus Lycopersicon esculentum

A

A

ABC

ABC

ABC

B

AB

ABC

ABC

ABC

ABC

AB

B

ABC

ABCD

ABC

ABC

B

B

BC

Maackia arnurensis

ABCD

BC

BCD

B

B

B

Maclura pomifera Narcissus pseudonarcissus

ABC

ABC

ABC

AB

B

B

ABC

B

BC

B

B

B

BC

B

Oryza sativa

B

B

B

B

B

B

Perseau americana Phaseolus coccineus

ABD

AB

B

B

AB

A

A

AB A

continued on next page

429

Table 4, continued Lectin

Purified lectin

Fluorochrome labeled

Phaseolus lunalus Phaseolus vulgaris-E

AB ABC

Phaseolus vulgaris-L Pisum sativum

ABCD

ABC

ABCD

ABC

ABC

ABC

Psophocarpus tetragonolobus I

BC

B

BC

B

Biotin or hapten labeled

Protein conjugated

AB

B

B

B

B

ABC

AB

B

B

ABC

B

B

ABC

B

B

ABC

Gold conjugated

Immobilized

Psophocarpus tetragonolobus I1

BC

B

BC

B

Ricinus communis I

ABCD

ABC

ABCD

AB

B

ABC

Ricinus communis I1 Robinia pseudoacacia

ABC

ABC

ABC

AB

B

ABC

AB

B

B

B

Sambucus nigra

ABCD

BC

BCD

B

B

BC

Solanum tuberosum

ABC

ABC

BC

B

B

BC

Sophora japonica

ABC

ABC

ABC

B

B

BC

Trichosanthes kirilowii

B

B

B

B

B

B

Triticum vulgaris Tulipa sp.

ABCD B

ABC

ABCD

ABC

AB

ABC

B

B

B

B

B

Ulex europaeus

ABCD

ABC

ABC

AB

AB

ABC

Urtica dioica Viciafaba

B

B

B

B

B

B

AB

AB

B

B

B

B

B

ABC

Vicia villosa

ABC

ABC

ABC

B

Vicia villosa B4

A

A

A

A

Vigna radiata

AB

B

B

Viscum album

AB

B

B

B

B

B

ABC

AB

B

BC

WistariaJoribunda

ABC

ABC

Letter denotes vendor. A, Sigma Chemical Co. (St. Louis, MO); B, EY Laboratories, Inc. (San Mateo, CA); C, Vector Laboratories, Inc. (Burlingame, CA); D, Boehringer Mannheim Corp. (Indianapolis, IN).

a

such glycoproteins will exhibit patterns of reactivity which reflect a composite of all the different glycan units and their isoforms found on the glycoprotein. Confirmatory experiments are advised where possible, when structural information is acquired. This may be accomplished by determining the reactivity of the glycoconjugate following treatment with appropriate exo- and/or endoglycosidases with each lectin that had exhibited a positive reaction with the intact biopolymer. Judicious selection of the enzymes will help demonstrate the glycan-related specificity of the interaction, and help define additional structural characteristics, or resolve complex lectin reactivity patterns caused by multiple glycan structures on a single polypeptide. 4.2.1. Lectin precipitation analysis of glycans The quantitative precipitin analysis of antibody-antigen interactions as introduced

430

by Heidelberger and colleagues [257] and extended by Kabat’s laboratory [258] for the structural characterization of polysaccharides and carbohydrate-rich biopolymers established the basis for the use of lectins in a similar manner to study these glycans. The characteristics of the precipitation of polysaccharides by lectin parallel those of the antibody-antigen interaction in solution [259,260]. The method is based upon the quantitative measurement of a given amount of lectin precipitated by increasing concentrations of the glycan. A plot of the amount of lectin precipitated with respect to the quantity of glycan added to the reaction produces curves characteristic of the properties of the lectin and glycan. The formation of the precipitate is dependent upon the multi-valency of the lectin, and the number of available lectin-binding saccharide units, often reflected by the degree of branching or sites of glycosylation of the biopolymer [3,253,259,260]. This technique, preceded only by the method of inhibition of agglutination analysis, was one of the earliest applications of lectins for the structural analysis of polysaccharides. It has proved quite effective for the study of blood group substances [26 1-2651, glycoproteins [209,266-2681, and homo- and heteropolysaccharides [269-2721. Recent use of contemporary biophysical approaches in conjunction with quantitative precipitation analysis has shed new light on the nature of the interactions which result in the formation of the lectin-glycan precipitate, thereby allowing for further interpretation of data from such analyses for the characterization of glycans [68,70,273]. Though an uncommon procedure today, structural information concerning sequence, linkage and branching patterns [3] is readily discerned by its use. The application of quantitative precipitation analysis is exemplified by a study of the branched pneumococcal S-14 heteropolysaccharide with RCA, WGA and G. simplicifolia agglutinin [274]. This glycan readily gave precipitation curves with both WGA and RCA, but not with the a-Gal-binding lectin, GS I. These results established the presence of a terminal fi-Gal end group and a b-GlcNAc residue linked at its C-4 position. Smith degradation of the S- 14 polysaccharide (limited periodate oxidation, followed by reduction with NaBH4 and mild acid hydrolysis) resulted in a product which precipitated with WGA, but no longer interacted with the fi-Gal-binding RCA. These and related results suggested that the original and Smith degraded polysaccharide were composed of the repeating oligosaccharide units, -6)GlcNAc(fil-3)[Gal(fil4)]Gal(~14)Glc and -6)GlcNAc(Bl3)Gal(fi14)Glc, respectively. Interestingly, this is the first report demonstrating that WGA is capable of binding internal C-6-substituted GlcNAc residues. The related technique of agar gel double diffusion analysis originally developed by Ouchterlony [275] for structural studies involving antibody-antigen reactions is very easy to use and was employed most effectively by Goldstein and So [276] for the structural analysis of glucans and fructans with Con A. The method relies on the formation of a specific precipitate in the agar gel caused by the interaction of concentration gradients emerging from opposing reservoirs (wells cut into the gel bed) containing the lectin or the reactive glycan preparation. Precipitation patterns caused by the interaction of different glycans andor lectins can yield important information regarding the presence of common or distinguishing features among different glycopolymers, as well as glycan heterogeneity of the preparation. In particular, this method is capable of determining whether the reactivity of a glycan preparation with more than one lectin is due to different reactive glycan moieties on a single biopolymer, or on separate polysaccharides

43 1

or glycoconjugates. This information is complementary to that obtained by the lectinbased blot analysis (see below), which is incapable of discriminating between such possibilities. The reader is referred to Oudin and Williams [277] for a review of the method and concepts of double diffusion analysis, as applied to the antibody-antigen system. Though still employed [278,279], surprisingly, lectin-based agar gel double diffusion for the analysis of glycan-containing preparations never gained popularity. This technique, requiring only small amounts of the reactive preparation, offers the user much potentially valuable information for only a small investment in effort and supplies. 4.2.2. Lectin afinity adsorption/chromatography Affinity adsorption of antigens to solid-phase-bound antibody, as an early application of affinity chromatography [280] is the historic keystone for the lectin affinity separation of glycans and glycoconjugates. The approach is based upon the concept that the solid-phase-bound lectin is capable of specifically retarding or tightly binding soluble glycoconjugates flowing past the adsorbant contained within the flow path. The initial applications of solid-phase lectin adsorption was limited primarily to the separation of bound (lectin-reactive) from unbound (lectin-unreactive) glycoprotein utilizing the only commercially available lectin at that time, Con A [281,282]. The protocols followed closely those employed for the isolation of antigens involving antibody adsorbants. Specifically, the sample containing the glycoprotein of interest is applied by gravity flow to a glass column containing the Con A adsorbant. This is followed by extensive washing with a neutral buffer to elute all unbound material. The specifically bound glycoproteins are eluted separately by the addition of the competitive inhibitory saccharide, methyl a-lo-mannopyranoside. More sophisticated approaches evolved, at first related to the sequential separation of numerous bound glycoproteins by use of either step [282] or gradient [242] elution techniques with buffers containing increasing concentrations of inhibitory saccharide, or by change of temperature, which affects the affinity of E-PHA for saccharide [283]. Serial lectin affinity chromatography [230,284], a logical, though innovational advance in the application of solid-phase lectin adsorbants for the purification and characterization of glycan-containing biopolymers, set the ground work for the contemporary technology of lectin affinity chromatography. The method utilizes a number of different lectin affinity chromatography steps, employing lectins with different carbohydrate specificities, in a sequential manner to resolve individual glycans or glycoconjugates contained within a complex mixture. Today, lectin affinity chromatography is commonly used to study glycoproteins, and has been applied with limited success to the separation and analysis of glycolipids using detergentcontaining aqueous buffers[285] or with a mixed solvent system composed of up to 50% tetrahydrofuran in aqueous buffers [286,287]. The binding constant of the glycan-lectin complex, under the conditions employed, probably represents the most important criterion for determining the potential interaction of a ligand with the adsorbant. Nevertheless, related factors must be considered. These are concerned with the density of lectin binding sites per volume of adsorbant, and the effective valency (number of lectin-reactive glycan units per glycoconjugate molecule) of the biopolymer. Obviously, the concentration of lectin binding sites will influence both the rate and capacity of ligand binding by the adsorbant, and increasing

432

Con A

$ 7 Mixture of oligosaccharides I-IX

a

Sugar inhibitors: 1. 10 mM methyl a-D-glucopyranoside 2.100 mM methyl a-wnannopyranoside

7

Lentil lectin

I

v, VII

Fraction Number

Fraction number

Fig. 3a. Flow diagram for the separation of the components of a complex mixture of oligosaccharides by serial lectin affinity chromatography. Depending upon the lectin adsorbant, specific oligosaccharides are either unbound (not retarded by the adsorbant), retarded (and eluted without the need of a saccharide inhibitor), or are tightly bound and then require either lOmM methyl a-o-glucopyranoside or 100 mM methyl a-o-mannopyranoside for elution. Where appropriate, each eluted peak is concentrated and the saccharide inhibitor is removed prior to application to the second affinity column. The structures of the individual oligosaccharides are shown in Fig. 3b. (Adapted from ref. 288.)

valency of the glycoconjugate will result in an increased probability of a single binding phenomenon causing retardation of the whole glycoconjugate. Accordingly, determination of the appropriate lectin adsorbants and the order of their sequential use in serial affinity chromatography depends upon the complexity and nature of the mixture of glycoconjugates to be separated or analyzed. In general for complex mixtures, it is advised that the first lectin adsorbant be capable of differentiating broad categories of potential ligands. Often Con A is employed for the separation of oligomannose and hybrid glycans from diantennary complex oligosaccharides, both of which bind, but to different degrees, to the lectin. These in turn, are separated from the unbound 0-linked and multi-antennary glycans. Further steps in the serial affinity separation procedure employ lectins capable of differentiating glycans within each category, based upon unique structural characteristics of the individual biopolymers. A detailed discussion of these and other practical considerations for serial lectin affinity chromatography is presented in a recent publication by Cummings [288]. Different approaches emphasizing specific problems, however, are presented by others [283,289]. A general description of the logic associated with serial affinity chromatography to separate individual components of a complex mixture of glycans is given in Figs. 3a and 3b. As shown, the flow diagram illustrates a procedure capable of resolving a number of different oligosaccharide structures. The strength of the technique however, lies in its versatility. It is important to realize that each step in serial lectin affinity chromatography can be tailored, in regard

43 3

Man(al-2)Man(a 1-6)

\ Man(a1-6) Man(al-2)Man(al-3) I

\ Man(bI4)GlcNAc(B 14)GlcNAc(P 1-N)Asn

Man(al-2)Man(a 1-2)Man(a 1-3)

i

NeuAc(a2-6)Gal(fi14)GlcNAc(fi1-2)Man(a 1-6)

\ Man(fil4)GlcNAc(fil4)GIcNAc(~l-N)Asn NeuAc(a2-6)Gal(fil4)GlcNAc(BI-Z)Man(a 1-3)

i

Gal(a 1-3)Gal(b I4)GlcNAc(B I -2)Man(a 1-5)

\ Man@ I4)GlcNAc(fi 14)GlcNAc(fi 1-N)Asn i Gal(a I-3)Gal(~l4)GlcNAc(fil-2)Man(al-3) NeuAc(a2-3)Ga1((3 14)GlcNAc(B I-2)Man(a 1-6)

\ NeuAc(a2-3)Gal(f3 I4)GlcNAc(B 1-2)

Man(gl-4)GlcNAc(BI-4)GlcNAc(Bl-N)Asn

\

Man(a1-3)

I (a1-6)

i

Fuc

i

NeuAc(a2-6)Gal(P I-I)GlcNAc(B 1-4) NeuAc(a2-3)Gal(B 14)GlcNAc(fi 1-6)

\

Man(al-6) NeuAc(a2-3)Gal(fi 14)GIcNAc(B 1-2) i

\

Man(Bl4)GlcNAc(fi 1-4)GlcNAc(B I -N)Asn i (UI-6) NeuAc(a2-3)Gal(B 14)GlcNAc(BI -2)Man(a 1-3) Fuc

I

Gal@ 14)GlcNAc(fi1-6)

\

Man(al-6)

\

Gal@ 14)GlcNAc(fi 1-2) I

Man(~l4)GlcNAc(f3l4)GlcNAc(fil-N)Asn

I (d 6 )

Gal(@1-4)GlcNAc(fi 1-2)Man(a 1-3)

Fuc

GlcNAc(f3 1-6)

\

Man(a 1-6) \ GlcNAc(B 1-2) I Man@ I4)GlcNAc(fi I-4)GlcNAc(B I-N)Asn I (a) 6 ) GlcNAc(B I-Z)Man(a 1-3) Fuc

I

NeuAc(a2-3)Ga1((3 14)GlcNAc(B1-6)

\

NeuAc(a2-3)Gal(l) 14)GlcNAc(B 1-2)Man(a 1-5)

\ Man(fil4)GlcNAc(fil4)GlcNAc(~l-N)Asn (VIII)

I

I

NeuAc(a2-3)Gal(fi 14)GlcNAc(B1-2)Man(a 1-3)

(01-6)

Fuc

I

NeuAc(a2-3)Gal(fi 14)GlcNAc(fi1 4 ) NeuAc(a2-6)

\

NeuAc(a2-3)Gal(B I J ) G a l N A c ( a I-N)SeriThr Fig. 3b. S t r u c t u r e s of the individual oligosaccharides resolved in Fig. 3a.

434

to the lectin employed and the conditions required to resolve bound or retarded glycans, to meet the specific needs of the problem under investigation. 4.2.3. Lectin-based blot analysis The union of a technique employing lectin recognition of specific glycan structures as a second dimension with another procedure for the separation of biopolymers was introduced, independently, by a number of laboratories in the mid 1970s. Both enzymecomplexed lectins [290] and radiolabeled lectins 1291-2941 were employed to probe gels directly for components separated by denaturing SDS-PAGE. Guzman et al. [295], on the other hand, examined native prolyl hydroxylase, separated on non-denaturing discontinuous gels to demonstrate the glycoprotein nature of the enzyme. The enzymic activity and lectin-related radioactivity were co-localized in the gel. The approach of in situ lectin binding or co-precipitation employed in these methods proved quite useful [296]. The most significant drawback to this technology was the time consuming steps involving the diffusion regulated introduction of the lectin reagent and washout of excess or unwanted reagents. Fortunately, shortly afterwards, the technique of employing antibody identification of antigens separated by PAGE and electroblotted onto a non-specific, solid-phase adsorbant sheet was introduced by Towbin et al. [297]. This led to the application of lectins in a similar manner employing both indirect (anti-lectin antibody) or direct (radiolabeled or enzyme-complexed lectin) detection of the transferred glycoconjugates [298-30 13. Both one-dimensional dot blot analysis and two-dimensional analysis involving passive or electroblotting of glycoconjugates separated first by PAGE related procedures are routinely performed today. The commercial introduction by Boehringer Mannheim Company, of digoxigenin-labeled lectins, in conjunction with a highly specific anti-digoxigenin antibody for the indirect localization of the hapten-containing lectin, serves as a convenient and sensitive approach for the detection of glycoconjugates transferred or adsorbed onto a membrane support. Haselbech et al. [ 1 151 present a useful description of this technique. Two-dimensional analysis of glycolipids was introduced independently employing an overlay technique of thin layer plates utilizing either radiolabeled anti-glycolipid antibodies [302,303] or lectins [285]. Smith’s laboratory [286,287,304] further extended the use of lectin-based glycolipid identification methods for both TLC and affinity chromatography. As a further evolution of the methodology, the application of lectin identification as a second dimension in association with HPLC has been used to define the elution position of glycoproteins and glycopeptides. Aliquots of each fraction, eluted from the HPLC column can be applied to a solid-phase filter support in order to fix the eluting (g1yco)proteins or (g1yco)peptides by non-specific reversible binding. Subsequent probing of the blocked filter with lectin reagents in a direct or indirect blot analysis allows for detection of glycan-containing fractions. To overcome the potential difficulties of poor fixation to nitrocellulose filters by some (g1yco)proteins and (glyco)peptides, Canas et al. [305] employed Immobilon-AV derivatized PVDF filters, which are capable of forming covalent bonds with nucleophilic groups, such as aminocontaining glycoconjugates. Following blocking of the filter with Tris/Tween 20 and ethanolamine, the application of biotinylated RCA and visualization with the Vectastain ABC kit permitted detection of glycopeptides down to 1-2 pmoles. The HPLC-lectin

435

blot analysis approach is convenient to use and may well gain popularity for the micropreparation of glycopeptides. 4.2.4. Lectin affinity electrophoresis The two-dimensional detection of antigens by immunoelectrophoresis incorporates the resolving power of gel electrophoresis and the discrimination and detectability attained by immunodiffusion reactions, occurring in the second dimension perpendicular to that of the first dimension of electrophoresis [306]. This method has been modified further whereby the second dimension of analysis involves electrophoresis of those components separated by the first dimension into a gel homogeneously infused with antibody [307]. The resulting precipitin pattern from this procedure reflects the resolving power of the electrophoretic separation in the first dimension, and the immuno-detectable properties and concentrations of each antibody-reactive substance in the second dimension. The lectin-based detection of glycoconjugates derived from this procedure has been termed crossed affinity immunoelectrophoresis, crossed affino-immunoelectrophoresis, or crossed immuno-affinoelectrophoresis and includes the incorporation of an appropriate lectin into the electrophoretic gel of the first dimension [308,309]. Accordingly, the final precipitin pattern evident in the antibody-containing second dimension gel also reflects the ability of the immuno-detectable glycoconjugate(s) to be retarded by the lectin. Conditions for the separation in each dimension must be such that little or no electrophoretic migration occurs by the lectin in the first dimension, and the antibody in the second dimension. Furthermore, to limit the effects of lectin on the migration and immunoreactivity of the glycoconjugate in the second dimension, a neutral saccharide inhibitor of the lectin is incorporated into the antibody-containing gel. A complete discussion of the theory and some applications of this technique is given by Heegard and Berg-Hansen [3 lo]. As shown in Fig. 4, crossed affinity immunoelectrophoresis is easily capable of demonstrating the glycan heterogeneity of each of the two electrophoretically distinct forms of human ceruloplasmin, detected by crossed immunoelectrophoresis [3 111. The procedure is quite routine and has gained popularity for the detection of glycosylation variants of clinically important glycoproteins [3 121, as well as for the structural analysis of the glycan moieties of glycoproteins [313-3151. In particular, much work has been done for the analysis of E-PHA, LCA and Con A separable variants of a-fetoprotein by crossed affinity immunoelectrophoresis and their relationship to malignancy [3 15,3 161. In recent years the method of crossed affinity immunoelectrophoresis has been modified whereby Western blot analysis, employing the antibody of choice or a second lectin, is substituted for the second dimension of electrophoresis in an antibody-containing gel. This approach is demonstrated by the procedure used for the convenient detection of multiple a-fetoprotein variants in liver diseases [3 17,3181. High density polyacrylamide gel electrophoresis of fluorescent end-labeled glycans, originally introduced in 1981 [3 191 exhibits high resolving power. Though separation of the different glycans is based primarily upon size, sequence information is obtainable when used in conjunction with specific glycosidase treatment of the glycan [3 19-3241. Recently, lectin affinity adsorption has been used as a second dimension to obtain structural information [324]. Batch adsorption, on an analytical scale, employing 50 pl of Con A- or RCA I-Sepharose particles allowed for the separate size analysis by PAGE of

436

Fig. 4. Crossed immunoelectrophoresis and crossed affinity immunoelectrophoresis patterns of human ceruloplasmin. c, Crossed immunoelectrophoresis with anti-ceruloplasmin; c-wga, crossed affinity immunoelectrophoresis with WGA (L oulguris agglutinin); c-lca, crossed affinity immunoelectrophoresis with LCA (L.culinuris agglutinin); y-wga, crossed affinity immunoelectrophoresis with WGA of component y from c; y-lca, crossed affinity immunoelectrophoresis with LCA of component y from c; z-wga, crossed affinity immunoelectrophoresis with WGA of component z from c; z-lca, crossed affinity immunoelectrophoresis with LCA of component z from c. All crossed affinity immunoelectrophoreses were performed with the lectin in the first dimension and anti-ceruloplasmin and appropriate lectin saccharide inhibitor incorporated into the second dimension gel. (Taken from ref. 3 1 1 .)

lectin-reactive and -unreactive fluorescent labeled glycans from ovalbumin. It was shown that this heterogeneous mixture contained Con A-reactive and -unreactive derivatized glycans of identical size and electrophoretic mobility, but of quite different structures. 4.2.5. Integration of lectin-based methodologies for soluble glycoconjugates Two companion publications from R. Kornfeld’s laboratory [325,326] together, serve as an excellent example of the integration of a number of different lectin-based techniques for the purification, characterization and intracellular localization of a specific glycoprotein. The authors utilized lectin affinity chromatography, lectin-blot analysis and lectin fluorescence microscopy to help define the properties of the Golgi enzyme, N-acetylglucosamine- 1-phosphodiester a-N-acetylglucosaminidase (PGase). To accomplish this, Con A and WGA were employed for adsorption purification steps,

437 Table 5 Summary of lectin blot analysis of bovine liver N-acetylglucosamine- 1 -phosphodiester a-N-acetylglucosaminidase (PGase) a Treatment of PGase

None N-glycanase Neuraminidase

Agglutinin A . hypogaea Gal(fil-3)GalNAc

S.nigra

M. amurensis

G. nivalis terminal Man

D. slramonium Gal((314)GlcNAc

+

-

+

f -

-

-

-

-I-

Neu(a2d)Gal Neu(a2-3)Gal (GalNAc)

-b -

+

+

-

-I-

Adapted from ref. [326]. Positive (+) and negative (-) symbols denote reactivity or lack of reactivity, respectively, of lectin reagent with the region of an SDS-PAGEkransfer blot corresponding to PGase.

a

digoxigenin-labeled G. nivalis agglutinin, S. nigra agglutinin, M. amurensis agglutinin, PNA and D. stramonium agglutinin were used in a commercially available kit format for enzyme amplified blot analysis of the purified PGase, and Texas-Red-labeled wheat germ agglutinin served as a reagent for fluorescence microscopy to denote the location of the trans-Golgi and trans-Golgi network of tissue cultured cells. Serial lectin chromatography of partially purified PGase with Con A- and WGA-Sepharose adsorbants resulted in a six-fold purification of the product. It is noteworthy, that as stated by the authors, purification of glycoproteins employing lectin-based techniques may be expected to result in selection of a subpopulation of the target glycoprotein. Since it is not unusual for a single glycosylated protein to exhibit glycan microheterogeneity, lectins capable of distinguishing among the glycan isoforms would segregate subpopulations of the total microheterogeneous population. For instance, pilot studies on the purification of PGase with WGA-Sepharose chromatography demonstrated that at least four different subpopulations of the enzyme bound to the adsorbant, and were eluted separately with increasing concentrations of N-acetylglucosamine. A fifth (possibly mixed) population of the PGase, representing 2 W O % of the total enzyme activity applied to the column, did not bind to the adsorbant and therefore, was not included in the final purified product. Evidently, the use of lectin chromatography in this instance, allowed for purification of the PGase, but also produced a product which was not representative, in regard to glycosylation, of the microheterogeneous enzyme in the original tissue extract. The carbohydrate structures of the lectin-based purified PGase was partially characterized by two-dimensional SDS-PAGEAectin blot analysis employing specific exoand endoglycosidases. As shown in Table 5, different combinations of glycosidase treatments and reactivity to the panel of lectin probes allowed the authors to conclude the probable presence of both 0- and N-linked glycans falling within the following structural groups, respectively: 0-Linked structure, NeuAc(a2-3)Gal(~l-3)[NeuAc(a24)]GalNac(a 1-O)Ser/Thr, N-Linked structure, [NeuAc(a2-3)Gal-GlcNAc-Man]~-Man-GlcNAc-GlcNAc-Asn. Lastly, these authors performed immunofluorescence microscopy (see section 4.3.4, Lectin histochemistry) to localize the PGase in cultured Vero cells. To allow for

438

the delineation of different Golgi regions of the cell, fluorochrome-labeled WGA was employed to preferentially stain the trans-Golgi and trans-Golgi network, thereby distinguishing these organelles from the medial- and cis-Golgi regions. By double, indirect fluorescence microscopy the authors were able to confirm that, although the PGase was terminally processed by sialyl transferase, the enzyme did not exhibit a transGolgi or trans-Golgi network location, but rather was present in the cis-Golgi andor medial-Golgi as suggested by others. The combination of different lectin-based approaches, as described above, served the researchers well both for the purification and the characterization of the glycoprotein, as well as for studies concerned with the biological function of the enzyme. 4.3. Cell-bound glycoconjugates

Often in glycobiology, the initial observations of functionally important glycans are related to those found on membrane-bound glycoconjugates. Numerous lectin-based methods, derived in great part from those developed for serology and cellular immunology, are available to the glycobiologist for the detection and characterization of cell-bound glycans, as well as for the enrichment of cell populations possessing the glycoconjugates of interest. The most common approaches employed are those which result in lectininduced aggregation or binding of the target cells, or detection of lectin-marked cells or subcellular compartments possessing reactive membrane-bound glycans. 4.3.1. Agglutination analysis Cellular agglutination is a convenient and sensitive initial approach for the detection of cell-surface-bound glycoconjugates of potential interest. The utilization of lectins with this technique has a long and consistent history of reliability [3,327]. It is based upon the ability of di- or multivalent lectins to cause cells in suspension to clump by specifically binding to appropriate glycoconjugates on the surface of adjacent cells. The size of the cell aggregates may be quite large and are often visible through a low power microscope or magnifying glass. Accordingly, cells which express exposed, surface target glycan(s) will be agglutinated. The degree of agglutination depends upon the relative reactivity of the glycan, the number of reactive structures on the surface of the cell, and the concentration of the lectin. It is clear, from the model of agglutination presented here, that an increase in lectin concentration will lead to an increase in the degree of agglutination of potentially reactive cells, up to a limit where the lectin becomes saturating. Under such a condition, all reactive cells will be covered maximally by the lectin, thereby causing an inhibition of the lectin-induced heterophilic cross-links between cells. In such a case, reduced or no agglutination is observed. Therefore, it is advisable that cell agglutination reactions be performed at varying concentrations of the lectin so as to determine the optimum concentration resulting in agglutination of target cells. One must keep in mind that the potential exists for the cells to represent a mixed population, resulting in the optimum concentration of lectin causing less than complete agglutination of all cells. The strength of lectin-induced agglutination however, is that it allows for a relatively rapid determination of the cell types which exhibit surface-bound glycoconjugates of interest.

439

Agglutination of animal cells dates back to the earliest studies performed with lectins [328]. The method proved most effective in determining the carbohydrate nature of the erythrocyte ABO blood group antigens [ 147,213,261,329-3321, identification of distinctive hemopoietic cell populations [333], and led to a host of studies associating differences in surface glycan structures with tumorigenicity and metastasis [334-3371. This latter area of study has matured to the point where specific tumor-associated glycan moieties have been structurally characterized, and their appearance related to the expression of specific glycosyltransferases [338-3411. The large number of commercially available lectins makes it convenient to employ lectin agglutination to detect and characterize distinctive groups of related microorganisms. In the late 1970s and early 1980s literature appeared on the use of Con A and other lectins to characterize amoeba [342], entameba [343,344] and bacteria [345,346]. Periera et al. [347] and Katzin and Colli [348] employed lectins to distinguish developmental stages of the protozoal parasite, Typanosoma cruzi. A similar approach continues today for the characterization of distinctive cell-surface glycoconjugates at each stage of development of these microorganisms [349]. Contemporary efforts utilizing lectin agglutination have broadened into all aspects of microbiology. These include the application of lectin panels to pseudo-serotype and subdivide bacterial [350-3551, and protozoal [356,357] strains, as well as to differentiate drug-resistant and -sensitive Trichornonas uaginalis [358]. It is noteworthy that established serogroups are further subdivided on the basis of distinctive agglutination profiles with the use of large numbers of lectins. Such results, however, may in part reflect nutritional or other factors which affect the metabolic condition of the organism, as expressed in the nature of surface sugar moieties, and not fundamental genetic differences. As with all approaches involving lectins, proper controls are necessary. This is even more of a concern when studying microorganisms. It is now well appreciated that bacteria and other microbes as well as animal cells may display cell surface lectins capable of interacting with glycan moieties, such as those found on plant lectin glycoproteins [359]. Consequently, care must be taken to differentiate agglutination of the target cells caused by the carbohydrate-binding property of the lectin reagent, from that caused by the cell-derived lectin interacting with the glycan units of the lectin reagent. The use of controls incorporating saccharides capable of specifically and competitively inhibiting the lectin reagent is strongly advised. Potential interference due to cell-bound lectins may be determined in a separate investigation employing neoglycoproteins possessing glycan moieties structurally related to those found on the plant lectin glycoproteins. Cottin et al. [360] demonstrated that specific neoglycoproteins are capable of interacting with the surface bacterial lectins resulting in agglutination. Such studies will help eliminate the possibility of target cell-derived lectins as a source of potential confksion. More recently, mammalian cell agglutination has been superseded by more sophisticated methods such as fluorescence flow cytometry (see later). The ease and relative inexpense of the former, however, when weighed against the information gained still makes agglutination analysis an attractive, effective analytical tool. A recent report by Knibbs et al. [245] serves as an excellent example of the methodology and logic of this approach, when used to explore the structure-function relationships of surface glycans from mammalian cells. These researchers employed a panel of 19 lectins for comparative

440 Table 6 Lectin reactivities of Ehrlich ascites cell variants' Lectin

G. simplicifolia I (GS I) G. simplicifolia I1 (GS 11) R. comrnunis (RCA) T vulgaris (WGA) C. ensiformis (Con A)

I! oulgaris E (PHA-E) I! oulgaris L (PHA-L) D. biflorus (DBA) L.fravus (LFA) S. nigra (SNA) M. amurensis L (MAL) M. amurensis E (MAH) A . caudafus (ACA) A . hypogaea (PNA) G. nioalis (GNA)

Y faba (VFA) D. sframonium (DSA) E. variegata (EVA) (1. europaeus (UEA)

EAT-wtb

EAT-c

0.23

0.47

2330 0.32

>500 0.32

0.98 < I .27

7.8

7.8

3.9

I .27

7.8

1.95

>340 0.5 >335

>I70 3.9 <0.65

0.48 6.3 >496 <0.49 >250 4.1 1.96 7.8 >500

0.97 1.58 0.48

>500 >250 3.8 0.49 >500 >500

Taken from ref. [245]. The numbers represent the concentration of lectin (pg/ml) required for 25% agglutination of the cells. EAT-wt, wild type, non-adherent cells; EAT-c, adherent Ehrlich cells grown in culture. a

agglutination analysis of Ehrlich ascites tumor cells grown as the ascitic form (EAT-wt), and an anchorage-dependent variant (EAT-c) selected to grow in tissue culture as a monolayer. As shown in Table 6, select lectins are capable of distinguishing between these two cell types. In particular, the sialic-acid-specific lectins from S. nigru, M. umurensis, and L. flauus help define the nature of the sialic acid linkage. More specifically, the authors describe how the information gained from the agglutination patterns (and the known specificities of the lectins) in conjunction with chemical and enzymic analyses of cell surface carbohydrate, allowed them to conclude that: (1) both types of cells contain a-2,3-linked sialic acid, but only the EAT-c cells possess a-2,6-Neu5Ac terminal structures; (2) the cell surface of the EAT-c variant contains a greater percentage of more highly glycosylated glycoproteins than that of the EAT-wt cells; (3) EAT-c cells exhibit an elevated number of terminal residues as compared to the EAT-wt cells; (4) the EAT-c cells possess fewer @-Gal-and B-GlcNAc-terminating chains than do the EAT-wt cells and consequently, the EAT-c variant exhibits surface glycoproteins which have oligosaccharide chains more completely modified by chain-terminating a-linked Gal and sialic acid groups than does the EAT-wt form. This study demonstrates that the careful and thoughtful application

441

of lectins in cell agglutination analysis can be a productive, preliminary approach to define more specific directions for the elucidation of structure-function relationships of cell-bound glycoconjugates. A derivative method of cellular agglutination, which has been quite effective for the isolation of restricted populations of suspension cells from heterogeneous mixtures, allows for the physical separation of preparative quantities of reactive cells. Consequently, the biological and biochemical characterization of unique cell populations, differentiated by the nature of surface glycans, can be studied. Li and Osgood[361] first demonstrated the practicality of this approach by employing I? vulgaris agglutinin for the preparative scale separation of erythrocytes from blood leukocytes. Reisner, Sharon and colleagues [333,362-3641 established the groundwork and developed largescale methods for the preparation of unique human and murine thymocyte populations which exhibit differential reactivity to peanut agglutinin. Subsequent studies described these as immature (PNA+) and mature (PNA-) thymocytes. Recently, Gillespie et al. [365] suggested that regulation of a specific Gal@1-3)GalNAc(a2-3) sialyltransferase is responsible for masking specific PNA+ sites during the differentiation of the immature T-lymphocytes into mature cells. In a similar manner, WGA, SBA and D. biJIorus agglutinin have been employed for the isolation of murine pluripotent stem cells [366], the preparation of human stem cells from bone marrow cells [362,363,367], or the separation of activated murine splenocytes exhibiting different levels of anti-tumor activity [368], respectively. A number of other lectins have been used to prepare distinct lectin-reactive and -unreactive populations of human and murine leukocytes, though in many cases, a correlation of the individual separated populations with unique functional properties is not yet apparent or is ambiguous [359,369]. In some cases, as with SBA, these methods have allowed, with limited success, the preparation of therapeutic amounts of thymocyte populations for transplantation between histo-incompatible donor and recipient [370,37 13. Recently, with the advent of monoclonal antibodies directed toward epitopes found on discrete leukocyte populations, efforts to obtain purified preparations of specific hemopoietic cell types for experimental and clinical applications have employed lectins in conjunction with such Mab reagents [372-3751. A significant advantage of lectins over other reagents is the convenience in removal of any bound and extraneous lectin. The addition of inhibitor saccharide to the cell washes is adequate for the effective release of bound lectin. The original methodology of Reisner et al. [362] for the use of PNA and SBA to separate T- and B-lymphocyte subpopulations, as modified by Schwartz et al. [376], remains the clearest example of this technology. Fig. 5 describes the application of SBA to the separation of murine T- and B-lymphocytes from a mixed population of cells derived from either bone marrow or spleen. In order to more completely separate T- and B-cells obtained from human bone marrow, so as to yield a T-cell-depleted population adequate for human allogeneic transplantation, Reisner et al. [367] introduced a preliminary E-rosetting step. This procedure, whereby sheep red blood cells (SRBC) are allowed to react with the bone marrow leukocyte fraction at 0°C for 2 h, results in the formation of T-cell-SRBC aggregates. These aggregates can be separated by centrifugation through a ficoll-hypaque solution. The non-rosetting mononuclear cells can be utilized for preparation of a T-cell-depleted hemopoietic precursor population suitable for human transplantation procedures.

442

1) Resuspend and wash in

2M D-galactose in PBS 2) Resuspend and wash in PBS

Fig. 5. Flow diagram for the fractionation of murine spleen and bone marrow cells with soybean agglutinin (G. m a agglutinin). (Adapted from ref. 376.)

4.3.2, Leetin-coated magnetic beads and $asks Alternative approaches related to the general concept described above have given the casual user additional convenient techniques for the preparative isolation of lectindifferentiated cell populations. Among the more popular are the use of lectins bound to solid phases in the form of tissue culture flasks or magnetic beads. The former technique treats the separation of those cells bound to the lectin adsorbant from unbound cells in a manner similar to the procedures used to separate non-adhering cells from those which adhere to a growth surface. Following addition of a mixed cell population to the lectin-coated flask and allowing for an appropriate incubation time, the unbound cells are removed by gentle panning of the suspension. The bound, “adhering” cells may be removed with the simple addition of a solution containing the saccharide inhibitor of the lectin or by agitation of the flask to physically dislodge the bound cells into the suspension medium. The procedure for activation of polystyrene flasks and binding of the lectin is described by Lebkowski et al. [375], and lectin- and antibody-coated flasks for this method are commercially available from Applied Immune Sciences (Menlo Park, CA). Wagner [377] describes the combined use of elutriation centrifugation, followed by selection of separated cell populations sequentially using an SBA-polystyrene flask and an anti-CD34-monoclonal-antibody-coatedpolystyrene flask for selective adsorption to obtain a highly enriched human hemopoietic progenitorhtem cell population. The utilization of lectin- and antibody-coated magnetic beads to separate different cell populations had been described originally using iron-oxide-containing polyacrylamideagarose [378] or methacrylate-based beads [379]. Today the technique employs commercially available protein-binding polystyrene-coated ferromagnetic particles, and a magnetic source used to attract the particles. The monodispersed microbeads and a

443

magnetic concentrator for this method is distributed by Dynal (Lake Success, NY). The mixed cell population is suspended in a vessel containing lectin-bound magnetic particles to allow interaction of the reactive cells with the adsorbant. Placement of a magnetic source alongside the reaction container results in immobilization of the magnetic lectin adsorbant-bound cell aggregates, allowing for the convenient removal of the suspension medium containing unbound cells. Addition of a competitive saccharide inhibitor of the lectin to a suspension of the lectin positive cells bound to the adsorbant results in the dissociation of the cells and the magnetic particles. Again, use of the external magnetic source permits the isolation of the lectin-positive cell population. Such an approach, using PNA- and anti-CD 19-monoclonal-antibody-coated magnetic particles, is described as a potential method to purge myeloma contaminated bone marrow cells for use in autologous bone marrow transplantation [373]. It is evident that numerous approaches to the application of solid-phase-bound lectins for the separation of cell populations is open to many clever technical innovations. 4.3.3. Fluorescence activated lectin-based flow cytometv and cell sorting The growing availability of relatively expensive fluorescence activated cell sorting (FACS) instruments has resulted in the popularization of this approach for the analysis and preparative isolation of lectin differentiated cell populations. The utility of the instrument is derived from its ability to quantitatively analyze the light scattering and fluorescent property of individual cells presented in a flow of liquid, in a manner similar to conventional fluorescence flow cytometers, and within prearranged detection limits, divert the cell to alternative flow paths. Accordingly, a mixed cell population, differentially stained by a surface-bound fluorescent-labeled lectin, may be fractionated based upon the amount of lectin (in terms of fluorescence) bound by each cell in the population. An approach for use of the analytical and preparative properties of the FACS is exemplified by a report by Chambers et al. [374] on the functional properties of rat natural killer cells. A number of studies [38&383] have demonstrated that lectins are capable of distinguishing subpopulations of large granular lymphocytes, some of which exhibit natural killer (NK) cell activity. Flow cytometric studies by McCoy et al. [384] showed that NK cells appear heterogeneous in regard to their interactions with the GSI-B4 lectin, and the agglutinins from !L villosa and L. esculentum, and that the binding patterns changed following interleukine stimulation of the cells. Subsequently, Chambers et al. [374] employed the analytical and cell sorting capabilities of FACS to isolate two NK cell populations (those able to bind an NK-cell-selective monoclonal antibody, anti-NKR-P 1) which differed in their interactions with L. esculentum agglutinin. Interestingly, the two lectin differentiated NK cell populations exhibit differences in their ability to perform a number of NK-cell-related activities. Apparently, phenotypic changes in cell surface glycoconjugate expression is associated with cellular differentiation regarding acquisition of specific NK cell activity by large granular lymphocytes. Highly enriched, relatively pure minor populations of cells expressing distinctive cell surface glycans, such as the NKR-P1+ and L. esculentum agglutinin' cell population, may be obtained by employing two independent fluorescent markers (two lectins andor antibodies) in a two-step FACS procedure based upon fluorescence and light scatter

444

properties. This is described in detail by McCoy et al. [385] employing L. esculentum and anti-NKR-P 1 monoclonal antibody. Advantage can be taken of the quantitative aspects of fluorescence flow cytometry to study the interaction of labeled lectins with populations of erythrocytes. To further investigate the hypothesis that surface sialic acid is lost during senescence of mammalian erythrocytes, thereby facilitating their clearance from circulation, Aminoff and colleagues [386] employed quantitative fluorescence flow cytometry using both labeled lectins and anti-IgG and -1gM. This investigation was extended and further refined recently [387] by the application of flow cytometry to the determination of both Kd and binding sites per erythrocyte for a number of lectins capable of recognizing terminal sialic acid, galactose or N-acetylgalactosamine residues. Bratosin et al. [387] demonstrated that “old” erythrocytes, separated from “young” erythrocytes by Percoll gradient centrifugation, exhibited approximately 40% fewer M. amurensis agglutinin (a-2,3-sialic acid groups) binding sites than did “young” cells. This contrasted with only a 27% loss of S. nigru agglutinin (a-2,6-sialic acid groups) binding sites per cell upon aging. As expected, care must be taken in these procedures to minimize lectinmediated agglutination, usually by utilizing a low concentration of lectin reagent. The development of non-agglutinating lectin derivatives will be beneficial for such studies. Kaku and collaborators describe the preparation of fluorescent, monovalent derivatives of the lectins from S. sieboldiana bark [248] and M. amurensis [249] and their use as probes for the fluorescence flow cytometric analysis of cells expressing specific sialylated glycoconjugates. 4.3.4. Lectin histochemistry Cell biologists have long made effective use of lectins for the study of cellular metabolism and differentiation through both light and electron microscopic histochemical techniques. Conceptually, lectin histochemical techniques parallel those developed for immunocytochemistry and depend upon the specific interaction of the cell-bound glycoconjugate target with the lectin reagent linked to an appropriate visual reporter system. Opaque agents, such as lectin-bound gold particles or lectin-conjugates of enzymes which produce opaque reaction products, yield high quality histochemical slides for light microscopy. Both direct fluorescent labeled lectins or indirect fluorescence microscopy, whereby a second fluorochrome-labeled reagent such as an anti-lectin antibody is employed as the visualizing agent, have been used successfully. Similarly, lectin conjugates (or second reagents able to detect the bound lectin) containing an electron dense agent, such as ferritin or gold particles, or an enzyme, such as horseradish peroxidase, which enzymatically produces an electron dense reaction product, have yielded excellent results for the ultracytochemical localization of specific glycoconjugates. The work of Nicolson and Singer [388,389] laid the foundation for lectin-electron microscopy. Today these techniques are employed regularly for the characterization of different cell types in tissue sections, as well as for the association and characterization of specific or unique glycoconjugates with cellular transformation or development. In a recent review of histochemical methods for glycobiology, with a particular focus on lectin histochemistry, Danguy et al. [390] describe a variety of select histological studies employing lectins to elucidate the structure and function of

445

glycoconjugates. The reader is directed to this review to gain an insight into the potential of the application of lectin histochemistry to the functional and structural analysis of cell-bound glycans. Glycobiologists interested in delving into both technical aspects as well as applications of both electron and light lectin histochemistry may find the publication by Li et al. [391] to be of interest. This paper, though focusing solely on the use of digoxigenin-labeled lectins in indirect microscopic methods, presents useful technical comments and methodology. The authors also describe the integration of lectin blot techniques with lectin histochemistry to further characterize the structure of the glycoconjugates which had been identified by microscopy. A word of caution to the uninitiated; as with all methods utilizing lectins, proper controls and verification procedures should be employed. The use of competitive saccharide inhibitors of the lectin will help confirm the specificity of the lectin-mediated binding. Similarly with indirect identification approaches, care should be taken to verify that the other reagents in the staining process do not cause false positive reactions. A number of lectins and other plant glycoproteins possess an a- I ,3-linked core fucose which is immunogenic in rabbits [ 14,15,320,321,323,392]. Anti-lectin [320,321,323], or in fact anti-horseradish peroxidase [39 11 antibody may contain a population of anti-carbohydrate antibody which is capable of binding directly to glycan epitopes in the absence of the lectin. An increasing number of lectins being employed as reagents contain such an epitope, which also appears to be present in some insect tissues [393-3961. Lectin histochemistry is a relatively specialized skill not often found in glycobiology laboratories. The commercial availability of many of the necessary lectin and related histochemical reagents, however, has made this approach quite popular with cell biologists, thereby further stimulating interest in structure-function relationships of glycoconjugates within that community.

References [ I ] Goldstein, I.J., Hughes, R.C., Monsigny, M., Osawa, T. and Sharon, N. (1980) Nature (London) 285, 66. [2] Barondes, S.H. (1988) Trends Biochem. Sci. 13, 480482. [3] Goldstein, I.J. and Poretz, R.D. (1986) Isolation, physicochemical characterization and carbohydratebinding specificity of lectins. In: I.E. Liener, N. Sharon and I.J. Goldstein (Eds.), The Lectins: Properties, Functions, and Applications in Biology and Medicine. Academic Press, New York. [4] Sharon, N. and Lis, H. (1989) Lectins. Chapman and Hall, London. Pusztai, A. (1991) Plant Lectins. Cambridge University Press, Cambridge, pp. 1-263. Doyle, PD. and Slifkin, M. (Eds.) (1994) Lectin-Microorganism Interactions. Marcel Dekker, New York. Pusztai, A. and Bardocz, S. (Eds.) (1995) Lectins: Biomedical Perspectives. Taylor & Francis, London. [5] Franz, H. (Ed.) (1988-1991) Advances in Lectin Research, Vol. 1-3. VEB Verlag Volk und Gesundheit, Berlin. [6] Pemberton, R.T. (1994) Mycol. Res. 98 (3), 277-290. [7] Goldstein, I.J. and Hayes, C.E. (1978) Adv. Carbohydr. Chem. Biochem. 35, 127-334. [8] Methods in Enzymology (1972) 28, 313-368; (1978) 50, 323-373; (1982) 83, 357-377; (1987) 138, 536-564; (1989) 179, 327-331; 342-347; (1994) 230, 66-86. [9] Knibbs, R.N., Goldstein, I.J., Ratcliffe, R.M. and Shibuya N. (1991) .I.Biol. Chem. 266, 83-88. [lo] Singh, J., Kamboj, S.S., Sandhu, R.S., Shangary, S. and Kamboj, K. (1993) Phytochemistry 33, 979983.

446

[I I ] Sanadi, A.R. and Surolia, A. (1994) J. Biol. Chem. 269, 5072-5077. [I21 Young, N.M., Johnston, R.A.Z., Szabo, A.G. and Watson, D.C. (1989) Arch. Biochem. Biophys. 270, 596-603. [13] Dorland, L., van Halbeek, H., Vliegenthart, J.F.G., Lis, H. and Sharron, N. (1980) J. Biol. Chem. 256, 7708-771 1. [14] Ashford, D., Dwek, R.A., Welpy, J.K., Amatayakul, S., Homans, S., Lis, H., Taylor, G.N., Sharon, N. and Rademacher, T.W. (1987) Eur. J. Biochem. 166, 31 1-320. [ 151 Sturm, A,, Bergwerff, A.A. and Vliegenthart, F.G. (1992) Eur. J. Biochem. 204, 3 13-3 16. [ 161 Hoffman, L.M. and Donaldson, D.D. (1987) BioTechnology 5, 157-1 60. [ 171 Jordan, E.T. and Goldstein, I.J. (1994) J. Biol. Chem. 269, 7674768 1. [18] Sharon, N. and Lis, H. (1990) FASEB J. 4, 3198-3208. [I91 Young, N.M. and Oomen, R.P. (1992) J. Mol. Biol. 228, 924-934. [20] Van Damme, E.J.M., Goldstein, I.J., Vercammen, G., Vuylsteke, J. and Peumans, W.J. (1992) Physiol. Plant 86, 245-252. [21] Van Damme, E.J.M., Smeets, K., Van Leuven, F. and Peumans, W.J. (1994) Plant Mol. Biol. 24, 825830. [22] Van Damme, E.J.M., Balzarini, J., Smeets, K., Van Leuven, F. and Peumans, W.J. (1994) Glycoconj. J. 11, 321-332. [23] Van Damme, E.J.M., Smeets, K., Torrekens, S., Van Leuven, F. and Peumans, W.J. (1994) Eur. J. Biochem. 221, 769-777. [24] Young, N.M., Johnston, R.A.Z., Szabo, A.G. and Watson, D.C. (1989) Arch. Biochem. Biophys. 270, 596-603. [25] Young, N.M., Johnston, R.A.Z. and Watson, D.C. (1991) FEBS Lett. 282, 382-384. [26] Rinderle, S.J., Goldstein, I.J., Matta, K.L. and Ratcliffe, R.M. (1989) J. Biol. Chem. 264, 16123-16131. [27] Raikhel, N.V. and Lee, H.-I. (1993) Annu. Rev. Plant Physiol. Plant Mol. Biol. 44, 591-615. [28] Bientema, J.J. (1994) FEBS Lett. 350, 159-163. [29] Kieliszewski, M.J., Showalter, A.M. and Leykam, J.F. (1994) Plant J. 5, 849-861. [30] Makela, 0. (1957) Studies on Hemagglutinins of Leguminous Seeds. Walin & Goos, Helsinki. [31] Poretz, R.D. and Goldstein, I.J. (1986) Studies on Hemagglutinins of Leguminous Seeds. Walin & Goos, Helsinki, pp. 85-1 19. [32] Sastry, M.V.K., Banarjee, P., Patanjal, S.R., Swamy, M.S., Swarnalatha, G.V. and Surolia, A. (1986) J. Biol. Chem. 261, 11726-11733. [33] Shibata, S., Goldstein, I.J. and Baker, D.A. (1982) J. Biol. Chem. 257, 9324-9329. [34] Spohr, U., Hindsgaul, 0. and Lemieux, R.U. (1985) Can. J. Chem. 63, 2644-2652. [35] Knibbs, R.N., Goldstein, I.J., Ratcliffe, R.M. and Shibuya, N. (1991) J. Biol. Chem. 266, 83-88. [36] Quiocho, F.A. (1986) Annu. Rev. Biochem. 55, 287-315. [37] Rini, J.M. (1995) Annu. Rev. Biophys. Biomol. Struct. 24, 551-557. [38] Edelman, G.M., Cunningham, B.A., Reeke Jr, G.N., Becker, J.W., Waxdal, M.J. and Wang, J.L. (1972) Proc. Natl. Acad. Sci. USA 69, 2580-2584. [39] Hardman, K.D. and Ainsworth, C.F. (1972) Biochemistry 1 1 , 49104919. [40] Derewenda, Z., Yariv, J., Helliwell, J.R., Kalb (Gilboa), A.J., Dodson, E.J., Papiz, M.Z., Wan, T. and Campbell, J. (1989) EMBO J. 8, 2189-2193. [41] Banerjee, R., Mane, S.C., Ganesh, V., Das, K., Dhanaraj, V., Mahanta, S.K., Suguna, K., Surolia, A. and Vijayan, M. (1994) Proc. Natl. Acad. Sci. USA 91, 227-231. [42] Vyas, N.K. (1991) Curr. Opin. Struct. Biol. 1, 741-749. [43] Lis, H. and Sharon N. (1991) Curr. Opin. Struct. Biol. 1, 750-759. [44] Sharon N. (1993) Trends Biochem. Sci. 18, 221-226. [45] Einsphar, H. (1991) Trans. Am. Crystallogr. Assoc. 25, 1-22. [46] Shaman, B., Shoham, M., Yonath, A,, Lis, H. and Sharon, N. (1984) J. Mol. Biol. 174, 723-725. [47] Reeke Jr, G.N. and Becker, J.W. (1986) Science 234, 1 108-1 1 11. [48] Einspahr, H., Parks, E.H., Suguna, K., Subramanian, E. and Suddath, EL. (1986) J. Biol. Chem. 261, 16518-1 6527.

447 [49] Rini, J.M., Hardman, K.D., Einspahr, H., Suddath, EL. and Carver, J.P. (1993) J. Biol. Chem. 268, 10126-1 10132. [50] Shaanan, B., Lis, H. and Sharon, N. (1989) Science 254, 862-866. [51] Delbaere, L.T.J., Vandonselaar, M., Prasad, L., Quail, J.W., Wilson, K.S. and Dauter, Z. (1993) J. Mol. Biol. 230, 950-965. [52] Bourne, Y., Rouge, P. and Cambillau, C. (1990) J. Biol. Chem. 265, 18161-18165. [53] Bourne, Y., Rouge, P. and Cambillau, C. (1992) J. Biol. Chem. 267, 197-203. [54] Bourne, Y., NCsa, M.-P., RougB, I?, Mazurier, J., Legrand, D., Spik, G., Montreuil, J. and Cambillau, C. (1992) J. Mol. Biol. 227, 938-941. [55] Loris, R., Steyaert, J., Maes, D., Lisgarten, J., Pickersgill, R. and Wyns, L. (1993) Biochemistry 32, 8772-8781. [56] Sankaranarayanan, R., Pun, K.D., Ganesh, V, Banerjee, R., Surolia, A. and Vijayan, M. (1993) J. Mol. Biol. 229, 558-560. [57] Wright, C.S. (1992) J. Biol. Chem. 267, 14345-14352. [58] Villafranca, J.E. and Robertus, J.D. (1981) J. Biol. Chem. 256, 554-556. [58a] Montfort, W., Villafranca, J.E., Monzingo, A.F., Ernst, S.R., Katzin, B., Rutenber, E., Xuong, N.H., Hamlin, R. and Robertus, J.D. (1987) J. Biol. Chem. 262, 5398-5403. [59] Wright, C.S., Kaku, H. and Goldstein, I.J. (1990) J. Biol. Chem. 265, 1676-1677. [59a] Hester, G., Kaku, H., Goldstein, I.J. and Wright, C. (1995) Nature Struct. Biol. 2, 472479. [60] Loris, R., Thi, M.H., Lisgarten, J. and Wyns, L. (1993) Proteins 15, 205-208. [61] Craik, D.J. and Higgins, K.A. (1989) Annu. Rep. NMR Spectrosc. 22, 61-138. [62] Kronis, K.A. and Carver, J.P. (1982) Biochemistry 21, 3050-3057. [63] Kronis, K.A. and Carver, J.P. (1985) Biochemistry 24, 826-833; 834-840. [64] Neurohs, K.J., Mantsch, H.H., Young, N.M. and Bundle, D.R. (1982) Biochemistry 21, 498-503. [65] Brewer, C.F. and Brown, R.D. (1979) Biochemistry 18, 2555-2562. [66] Takesada, H., Shibuya, N. and Nagashima, N. (1992) J. Biochem. 112, 143-146. [67] Bhattacharyya, L., Khan, I.M. and Brewer, C.F. (1988) Biochemistry 27, 8762-8767. [68] Bhattacharyya, L., Fant, J., Lonn, H. and Brewer, C.F. (1990) Biochemistry 29, 7523-7530. [69] Bhattacharyya, L., Khan, M.I., Fant, J. and Brewer, C.F. (1989) J. Biol. Chem. 264, 11543-1 1545. [70] Bhattacharyya, L., Haraldsson, M. and Brewer, C.F. (1988) Biochemistry 27, 1034-1041. [70a] Dessen, A., Gupta, D., Sabesan, S., Brewer, C.F. and Sacchettini, J.C. (1995) Biochemistry 34, 49334942. [71] Foriers, A,, De Neve, R., Kanarek, L. and Strosberg, A.D. (1978) Proc. Natl. Acad. Sci. USA 75, 1136-1139. [72] Cunningham, B.A., Hemperly, J.J., Hopp, T.P. and Edelman, G.M. (1979) Proc. Natl. Acad. Sci. USA 76, 321 8-3232. [73] Hopp, T.P.,Hemperly, J.J. and Cunningham, B.A. (1982) J. Biol. Chem. 257, 44734483. [74j Sumner, J.B. and Howell, S.F. (1936) J. Bacteriol 32, 227-237. [75] Agrawal, B.B.L. and Goldstein, I.J. (1965) Biochem. J. 96, 23C-25C. [76] Olson, M.O.J. and Liener, I.E. (1967) Biochemistry 6, 105-1 11. [77] Edelman, G.M., Cunningham, B.A., Reeke Jr, G.N., Becker, J.W., Waxdal, M.J. and Wang, J.L. (1972) Proc. Natl. Acad. Sci. USA 69, 2580-2585. [78] Goldstein, I.J., Iyer, R.N., Smith, E.E. and So, L.L. (1967) Biochemistry 6, 2372-2377. [79] So, L.L. and Goldstein, I.J. (1968) J. Biol. Chem. 243, 2003-2007. [80] Carver, J.P., MacKenzie, A.E. and Hardman, K.D. (1985) Biopolymers 24, 49-63. [81] Brewer, C.F., Bhattacharyya, L., Brown 111, R.D., Koenig, S.H. (1985) Biochem. Biophys. Res. Commun. 127, 1066-1071. [82] Kaku, H., Goldstein, LJ., Oscarson, S. (1991) Carbohydr. Res. 213, 109-116. [83] Brewer, C.F. and Bhattacharyya, L. (1988) Glycoconj. J. 5, 159-173. [84] Poretz, R.D. and Goldstein, I.J. (1971) Biochem. Pharmacol. 20, 2727-2739. [85] Edelman, G.M. and Wang, J.L. (1978) J. Biol. Chem. 253, 3016-3022. [86] Marik, T., Entlicher, G. and Kocourek, J. (1974) Biochim. Biophys. Acta 336, 53-61. [87] Trowbridge, I S . (1974) J. Biol. Chem. 249, 6004-6012.

448 [88] [89] [90] [91]

Van Wauwe, J.-P., Loontiens, F.G. and DeBruyne, C.K. (1975) Biochim. Biophys. Acta 3 7 9 , 4 5 6 4 6 1 , Allen, A.K., Desai, N.N. and Neuberger, A. (1976) Biochem. J. 155, 127-135. Kornfeld, K., Reitman, M.L. and Kornfeld, R. (1981) J. Biol. Chem. 256, 6633-6640. Howard, I.K., Sage, H.J., Stein, M.D., Young, N.M., Leon, M.A. and Dyckes, D.F. (1971) J. Biol. Chem. 246, 1590-1 595. [92] Ticha, M., Entilicher, G., Kostir, J.V. and Kocourek, J. (1970) Biochim. Biophys. Acta 221, 282-289. [93] Stein, M.D., Howard, I.K. and Sage, H.J. (1971) Arch. Biochem. Biophys. 146, 353-355. [94] Loganathan, D., Osborne, S.E., Glick, G.D. and Goldstein, I.J. (1992) Arch. Biochem. Biophys. 299, 268-274. [95] Debray, H., Decout, D., Strecker, G., Spik, G. and Montreuil, J. (1981) Eur. J. Biochem. 117, 41-45. [96] Horstmann, C., Rudolph, A. and Schmidt, P. (1978) Biochem. Physiol. Pflanz. 173, 31 1-321. [97] Allen, H.J. and Johnson, E.A.Z. (1976) Biochim. Biophys. Acta 444, 374-385. [98] Perera, C.B. and Frumin, A.M. (1966) Science 151, 821. [99] Wang, J.L., Becker, J.W., Reeke Jr, G.N. and Edelman, G.M. (1974) J. Mol. Biol. 88, 259-262. [IOO] Loganathan, D., Osborne, S.E., Glick, G.D. and Goldstein, I.J. (1992) Arch. Biochem. Biophys. 299, 268-274. [ l o l l Van Damme, E.J.M., Allen, A.K. and Peuman, W.J. (1987) FEBS Lett. 215, 140-144. [I021 Van Damme, E.J.M., Allen, A.K. and Peuman, W.J. (1988) Physiol. Plant 73, 52-57. [I031 Van Damme, E.J.M., Goldstein, I.J. and Peuman, W.J. (1991) Phytochemistry 30, 509-514. [I041 Shibuya, N., Goldstein, I.J., Van Damme, E.J.M. and Peumans, W.J. (1988) J. Biol. Chem. 263,728-734. [I051 Kaku, H., Van Damme, E.J.M., Peumans, W.J. and Goldstein, I.J. (1990) Arch. Biochem Biophys. 279, 298-3 04. [I061 Kaku, H., Van Damme, E.J.M., Peumans, W.J. and Goldstein, I.J. (1992) Carbohydr. Res. 229, 347-353. [I071 Mo, H., Van Damme, E.J.M., Peumans, W.J. and Goldstein, I.J. (1993) Arch. Biochem. Biophys. 306, 431438. [I081 Kaku, H. and Goldstein, I.J. (1992) Carbohydr. Res. 229, 337-346. [I091 Saito, K., Komae, A,, Kakuta, M., Van Damme, E.J.M., Peumans, W.J., Goldstein, I.J. and Misaki, A. (1993) Eur. J. Biochem, 217, 677-681. [IlO] Kaku, H. and Goldstein, I.J. (1991) Carbohydr. Res. 213, 109-116. [l 1 I] Shibuya, N., Berry, J.E. and Goldstein, I.J. (1988) Arch. Biochem. Biophys. 267, 676-680. [112] Animashaun, T. and Hughes, R.C. (1989) J. Biol. Chem. 264, 46574663. [I131 Animashaun, T., Mahmood, N., Hay, A.J. and Hughes, R.C. (1993) Antiviral Chem. Chemother. 4, 145-153. [ I 141 Balzarini, J., Schols, D., Neyts, J., Van Damme, E.J.M., Peumans, W.J. and De Clercq, E. (1991) Antimicrob. Agents 35, 410-416. [ I 151 Haselbeck, A,, Scheckander, E., von der Eltz, H. and Hosel, W. (1990) Anal. Biochem. 191, 25-30. [I161 Allen, A.K., Neuberger, A. and Sharon, N. (1973) Biochem. J. 131, 155-162. [ I 171 Rice, R.H. and Etzler, M.E. (1975) Biochemistry 14, 40934099. [118] Wright, C.S. (1984) J. Mol. Biol. 178, 91-104. [119] Wright, C.S. (1989) J. Mol. Biol. 209, 4 7 5 4 8 7 . [I201 Burger, M.M. and Goldberg, A.R. (1967) Proc. Natl. Acad. Sci. USA 57, 359-366. [I211 Nagata, Y. and Burger, M.M. (1974) J. Biol. Chem. 249, 3116-3122. [122] Cuatrecasas, P. (1973) Biochemistry 12, 1312-1323. [I231 Bhavanandan, V.P. and Katlic, A.W. (1979) J. Biol. Chem. 254, 4 0 0 M 0 0 8 . [I241 Peters, B.P., Ebisu, S., Goldstein, I.J. and Flashner, M. (1979) Biochemistry 18, 5505-551 1. [I251 Wright, C.S. (1980a) J. Mol. Biol. 139, 53-60. [I261 Wright, C.S. (1980b) J. Mol. Biol. 141, 267-291. [I271 Wright, C.S. (1992) J. Biol. Chem. 267, 14345-14352. [128] Nachbar, M.S., Oppenheim, J.D. and Thomas, J.O. (1980) J. Biol. Chem. 255, 2056-2061. [I291 Kilpatrick, D.C. (1980) Biochem. J. 185, 169-172. [I301 Allen, A.K. and Neuberger, A. (1973) Biochem. J. 135, 307-314. [I311 Owens, R.J. and Northcote, D.H. (1980) Phytochemistry 19, 1861-1862. [I321 Ashford, D., Allen, A.K. and Neuberger, A. (1982) Abstr. 12th Int. Congr. Biochem. p. 345.

449 [I331 Pardoe, G.I., Bird, G.W.G. and Uhlenbruck, G. (1969) 2. Immunitaetsforsch. Allerg. Klin. Immunol. 137, 4 4 2 4 5 7 . [I341 Renkonen, K.O. (1948) Ann. Med. Exp. Biol. Fenn. 26, 66-72. [I351 Matsumoto, 1. and Osawa, T. (1974) Biochemistry 13, 582-588. [I361 Kilpatrick, D.C., Yeoman, M.M. and Kilpatrick, S.P. (1978) Plant Sci. Lett. 13, 3 5 4 0 . [I371 Crowley, J.F. and Goldstein, I.J. (1981) FEBS Lett. 130, 149-152. [I381 Kilpatrick, D.C. and Yeoman, M.M. (1978) Biochem. J. 175, 1151-1153. [I391 Crowley, J.F., Goldstein, I.J., Arnap, J. and Lonngren, I. (1984) Arch. Biochem. Biophys. 231, 524-533. [I401 Cummings, R.D. and Kornfeld, S. (1984) J. Biol. Chem. 259, 6253-6260. [I411 Yamashita, K., Totani, K., Ohkura, T., Takasaki, S., Goldstein, 1.J. and Kobata, A. (1987) J. Biol. Chem. 262, 1602-1 607. [I421 Shanker Iyer, P.N., Wilkinson, K.D. and Goldstein, I.J. (1976) Arch. Biochem. Biophys. 177, 330-333. [I431 Ebisu, S., Shankar Iyer, PN. and Goldstein, 1.J. (1978) Carbohydr. Res. 61, 129-138. [I441 Judd, W.J., Beck, M.L., Hicklin, B.L., Shankar lyer, P.N. and Goldstein, I.J. (1977) Vox Sang. 33, 246251. [I451 Osawa, T. (1966) Biochim. Biophys. Acta 115, 507-510. [I461 Henniger, R.A., Shulte, B.A. and Spicer, S.S. (1986) Histochem. J. 78, 589-596. [I471 Bird, G.W.G. (1951) Curr. Sci. 20, 298-299. [I481 Etzler, M.E. and Kabat, E.A. (1970) Biochemistry 9, 869-877. [I491 Carter, W.G. and Etzler, M.E. (1975) Biochemistry 14, 2685-2689. [I501 Carter, W.G. and Etzler, M.E. (1975) J. Biol. Chem. 250, 2756-2762. [I511 Roberts, D.M., Walker, J. and Etzler, M.E. (1982) Arch. Biochem. Biophys. 218, 213-219. [I521 Harada J.J., Spadoro-Tank, J., Maxwell, J.C., Schnell, D.J. and Etzler, M.E. (1990) J. Biol. Chem. 265, 4997-5001. [I531 Baker, D.A., Sugii, S., Kabat, E.A., Ratcliffe, R.M., Hermentin, P. and Lemieux, R.U. (1983) Biochemistry 22, 2741-2750. [I541 Boyd, W.C. and Reguera, R.M (1949) J. Immunol. 62, 333-339. [I551 Could, N.R. and Scheinberg, S.L. (1970) Arch. Biochem. Biophys. 137, 1-1 I . [I561 Galbraith, W. and Goldstein, I.J. (1972) Biochemistry 11, 3976-3984. [I571 Roberts, D.D. and Goldstein, I.J. (1982) J. Biol. Chem. 257, 11274-1 1277. [I581 Roberts, D.D. and Goldstein, I.J. (1984) J. Biol. Chem. 259, 909-914. [I591 Jordan, E. and Goldstein, 1.J. (1994) J. Biol. Chem. 269, 7674-7681. [I601 Roberts, D.D. and Goldstein, I.J. (1983) J. Biol. Chem. 258, 13820-13824. [I611 Roberts, D.D. and Goldstein, 1.J. (1984) J. Biol. Chem. 259, 903-908. [I621 Liener, 1.E. (1953) J. Nutr. 49, 527-539. [ I631 Sela, B.-A,, Lis, H., Sharon, N. and Sachs, L. ( I 970) J. Biomol. Struct. Dynam. 2, 41 -54. [I641 Gordon, J.A., Blumberg, S., Lis H . and Sharon N. (1972) FEBS Lett. 24, 193-196. [I651 Allen, A.K. and Neuberger, A. (1975) FEBS Lett. 50, 362-364. [I661 Lotan, R., Siegelman, H.W., Lis, H . and Sharon N. (1974) J. Biol. Chem. 249, 1219-1224. [I671 Vodkin, L.O., Rhodes, P.R. and Goldberg, R.B. (1983) Cell 34, 1023-1031. [I681 Pereira, M.E.A., Kabat, E.A. and Sharon, N. (1974) Carbohydr. Res. 37, 89-102. [I691 de Boeck, H., Lis, H. van Tilbeurgh, H., Sharon, N. and Loontiens, EG. (1984) J. Biol. Chem. 259, 7067-7074. [I701 Bhattacharyya, L., Das, P.K. and Sen, A. (1981) Arch. Biochem. Biophys. 211, 4 5 9 4 7 0 . [I711 de Boeck, H., Loontiens, EG., Lis, H. and Sharon N. (1984) Arch. Biochem. Biophys. 234,297-304. [I721 Iglesias, J.L., Lis, H. and Sharon, N. (1982) Eur. J. Biochem. 123, 247-252. [I731 Kalades, EM., Kabat, E.A., Iglesias, J.L., Lis, H. and Sharon N. (1982) Arch. Biochem. Biophys. 217, 624-637. [I741 Bird, G.W.G. (1964) Vox Sang. 9, 748-759. [I751 Bird, G.W.G. and Wingham, J. (1971) Scand. J. Haematol. 8, 307-308. [I761 Lotan, R., Skutelsky, E., Danon, D. and Sharon, N. (1975) J. Biol. Chem. 250, 8518-8523. [I771 Uhlenbruck, G., Pardoe, I . and Bird, G.W.G. (1969) Z . Immunitaetsforsch. Allerg. Klin. Immunol. 138, 423433.

450 [I781 [179] [I801 [I811 [182] [I831 [184] [ 1851 [I861 [I871 [188] [I891 [190] [I911 [I921 [I931 [194] [195] [196] [197] [I981 [I991 [200] [201] [202] [203] [204] [205] [206] [207] [208] [209] [210] [21 I] [212] [213] [214] [215] [216] [217] [218] [219] [220] [221] [222] [223] [224]

Pereira, M.E.A., Kabat, E.A., Lotan, R. and Sharon, N. (1976) Carbohydr. Res. 51, 107-118. Bausch, J.N. and Poretz, R.D. (1977) Biochemistry 16, 5790-5794. Bausch, J.N., Richey, J. and Poretz, R.D. (1981) Biochemistry 20, 2618-2620. Cawley, L.P., Jones, J.M. and Teresa, G.W. (1967) Transfusion (Philadelphia) 7, 343-346. Sarkar, M., Wu, A.M. and Kabat, E.A. (1981) Arch. Biochem. Biophys. 209, 204-218. Appukuttan, P.S. and Basu, D. (1981) Anal. Biochem. 113, 253-255. Kortt, A.A. (1984) Eur. J. Biochem. 138, 519-525. Kortt, A.A. (1985) Arch. Biochem. Biophys. 236, 544-554. Khan, M.I., Sastry, M.V.K. and Surolia, A. (1986) J. Biol. Chern. 261, 3013-3019. Pun, K.D. and Surolia, A. (1994) J. Biol. Chem. 269, 30917-30926. Matsuda, T., Kabat, E.A. and Surolia, A. (1989) Mol. Immunol. 26, 189-195. Pun, K.D. and Surolia, A. (1994) Pure Appl. Chem. 66, 497-502. Chatterjee, B.P., Vaith, P., Chatterjee, S., Karduck, D. and Uhlenbruck, G. (1979) Int. J. Biochem. 10, 321-327. Kumar, G.S., Appukuttan, P.S. and Basu, D. (1982) J. Biosci. 4, 257-261. Sastry, M.V.K., Banarjee, P., Patanjali, S.R., Swamy, M.J., Swarnalatha, G.V. and Surolia, A. (1986) J. Biol. Chem. 261, 11726-1 1733. Mahanta, S.K., Sanker, S., Rao, N.V.S.A.V.P., Swamy, M.J. and Surolia, A. (1992) Biochem. J. 2854, 95-101. Moreira, R.J. and Ainouz, I.L. (1981) Biol. Plant. 23, 186-192. Rogue-Barreira, M.C. and Compos-Neto, A. (1985) J. Immunol. 134, 257-261. Kondoh, H., Kobayashi, K, Hagiwara, K. and Kajii, T. (1986) J. Immunol. Methods 88, 171-173. Takahashi, T., Funatsu, G. and Funatsu, M. (1962) J. Biochem. (Tokyo) 52, 50-53. Olsnes, S., Refsnes, K. and Pihl, A. (1974) Nature (London) 249, 627-631. Gurtler, L.G. and Horstmann, H.J. (1973) Biochim. Biophys. Acta 295, 582-594. Olsnes, S. and Pihl, A. (1973) Biochemistry 12, 3121-3125. Endo, Y., Mitsui, K., Motiguki, M. and Tsurugi, K. (1987) J. Biol. Chem. 262, 5908-5912. Lord, J.M., Roberts, L.M. and Robertus, J.D. (1994) FASEB J. 8 (2), 201-208. Nicolson, G.L. (1974) Int. Rev. Cytol. 39, 89-190. Hammarstrom, S. (1973) Methods Enzymol. 28 (B), 368-383. Makela, 0. and Makela, I? (1956) Ann. Med. Exp. Biol. Fenn. 34, 4 0 2 4 0 4 . Horejsi, V. and Kocourek, J. (1973) Biochim. Biophys. Acta 297, 346351. Hayes, C.E. and Goldstein, I.J. (1974) J. Biol. Chem. 249, 1904-1914. Murphy, L.A. and Goldstein, I.J. (1977) J. Biol. Chem. 252, 47394742. Eckhardt, A.E. and Goldstein, I.J. (1983) Biochemistry 22, 5290-5297. Shibata, S., Peters, B.P., Roberts, D.D., Goldstein, I.J. and Liotla, L.A. (1982) FEBS Lett. 142, 194-198. Spiro, R.G. and Bhoyroo, V.D. (1984) J. Biol Chem. 259, 9858-9866. Renkonen, K.G. (1948) Ann. Med. Exp. Biol. Fenn. 26, 66-72. Morgan, W.T.J. and Watkins, W.M. (1953) Br. J. Exp. Pathol. 34, 94-103. Yanv, J., Kalb, A.J. and Katchalski, E. (1967) Nature (London) 215, 890-891. Blumberg, S., Hildesheim, J., Yariv, J. and Wilson, K.J. (1972) Biochim. Biophys. Acta 264, 171-176. Pereira, M.E.A. and Kabat, E.A. (1974) Biochemistry 13, 3184-3192. Kalb, A.J. (1968) Biochim. Biophys. Acta 168, 532-536. Goldstein, I.J. and Poretz, R.D. (1986) In: I.E. Liener, N. Sharon and I.J. Goldstein (Eds.), The Lectins: Properties, Functions and Applications in Biology and Medicine. Academic Press, New York, p. 199. Petryniak, J. and Goldstein, I.J. (1986) Biochemistry 25, 2829-2838. Konami, Y., Tsuji, T., Toyoshima, S., Matsumoto, I. and Osawa, T. (1989) Chem. Pharm. Bull. 37, 729-73 1. Springer, G.F. and Williamson, I? (1962) Biochem. J. 85, 282-291. Yamashita, K., Ohkura, T., Umetsu, K. and Suzuki, T. (1992) J. Biol. Chem. 267, 25414-25422. Shibata, S., Goldstein, I.J. and Baker, D.A. (1982) J. Biol. Chem. 257, 932449329, Kaladas, P.M., Kabat, E.A., Shibata, S. and Goldstein, I.J. (1983) Arch. Biochem. Biophys. 223, 309318.

45 1 [225] Shugishita, S. (1935) Juzenkwai-Zasshi 40 ( S ) , 1938. [226] Miller, J.B., Noyes, C., Heinrikson, R., Kingdon, H.S. and Yachnin, S. (1973) J. Exp. Med. 138, 939951. [227] Yachnin, S. and Svenson, R.H. (1972) Immunology 22, 871-883. [228] Kornfeld, R. and Kornfeld, S. (1970) J. Biol. Chem. 245, 2536-2545. 12291 Inmura, T., Tsuji, T., Tagami, S., Yamamoto, K. and Osawa, T. (1981) Biochemistry 20, 560-566. [230] Cummings, R.D. and Kornfeld, S. (1982) J. Biol. Chem. 257, 11230-1 1234. [231] Hammerstrom, S., Hammarstrom, M.L., Sundblad, G., Arnap, J. and Lonngren, J. (1982) Proc. Natl. Acad. Sci. USA 79, 1611-1615. [232] Mirkov, T.E. and Chrispeels, M.J. (19 ) Glycobiology 3, 581-587. [233] Green, E.D. and Baenziger, J.U. (1987) J. Biol. Chem. 262, 12018-012029. [234] Narasimhan, S., Feed, J.C. and Schachter, H. (1986) Carbohydr. Res. 149, 65-83. [235] Bierhuizen, M.F.A., Edzes, H.T., Schiphorst, W.E.C.M., van den Eijnden, D.H. and van Dijk, W. (1988) Glycoconj. J. 5, 85-97. [236] Zeng, F.Y. and Gabius, H.-J. (1992) Z. Naturforsch., p. 47c. [237] Mandal, C. and Mandal, C. (1990) Experientia 46, 433441. [238] Goldstein, I.J. and Poretz, R.D. (1986) In: I.E. Liener, N. Sharon and I.J. Goldstein (Eds.), The Lectins: Properties, Functions and Applications in Biology and Medicine. Academic Press, New York, pp. 209214. [239] Shibuya, N., Goldstein, I.J., Broekaert, W.F., Nsimba-Lubaki, M., Peeters, B. and Peuman, W.J. (1987) J. Biol. Chem. 262, 1596-1601. [240] Yamashita, K., Umetsu, K., Suzuki, T. and Ohkuro, T. (1992) Biochemistry, 31, 11647-11650. [241] Wang, W.C. and Cummings, R.D. (1988) J. Biol. Chem. 263, 45764585. [242] Shibuya, N., Goldstein, I.J., Broekaert, U.F., Nsimba-Lubaki, M., Peeters, B. and Peuman, W.J. (1987) Arch. Biochem. Biophys. 254, 1-8. [243] Brasalier, S.R., Rockwell, R.W., Dahiya, R., Duh, Q.Y. and Kim, Y.S. (1990) Cancer Res. 50, 12991307. [244] Mostafapour, M.K. and Goldstein, I.J. (1993) Arch. Biochem. Biophys. 303, 255-259. [245] Knibbs, R.N., MacCallum, D.K., Lillie, J.H. and Goldstein, I.J. (1994) Glycobiology 4, 4 1 9 4 2 8 . [246] Knibbs, R.N., Perini, F. and Goldstein, I.J. (1980) Biochemistry 28, 63794392. [247] Yamashita, K., Umetsu, K., Suzuki, T. and Ohkura, T. (1992) Biochemistry 31, 11647-1 1650. [248] Kaku, H. and Shibuya, N. (1992) FEBS Lett. 306, 176-180 [249] Kaku, H., Mori, Y., Goldstein, I.J. and Shibuya, N. (1993) J. Biol. Chem. 268, 13237-13241. [250] Taatjes, D.J., Roth, W., Peumans, W. and Goldstein, I.J. (1988) Histochem. J. 20, 4 7 8 4 9 0 . [251] Sata, T., Lackie, P.M., Taatjes, D.J., Peumans, W. and Roth, J. (1989) J. Histochem. Cytochem. 37, 1577-1 588. [252] Sata, T., Roth, J., Zuber, C., Stamm, B. and Heitz, P.U. (1991) Am. J. Pathol. 139, 1435-1448. [253] So, L. and Goldstein, I.J. (1967) J. Biol. Chem. 242, 1617-1622. [254] Chien, S.-M., Singla, S. and Poretz, R.D. (1975) J. Immunol. Methods 169-174. [255] Gunther, G.R., Wang, J.L., Yahara, I, Cunningham, B.A. and Edelman, G.M. (1973) Proc. Natl. Acad. Sci. USA 70, 1012-1016. [256] Monsigny, M., Sene, C., Obrenovitch, A,, Roche, A.-C., Delmotte, EM. and Boschetti, E. (1979) Eur. J. Biochem. 98, 3 9 4 5 . [257] Heidelberger, M. (1956) Annu. Rev. Biochem. 25, 6 4 1 4 5 8 . [258] Kabat, E.A. and Mayer, M.M. (1971) In: C.C. Thomas (Ed.), Experimental Immunochemistry. Springfield IL, pp. 4 0 5 4 3 3 . [259] Goldstein, I.J., Hollerman, C.E. and Smith, E.E. (1965) Biochemistry 4, 876-883. [260] Poretz, R.D. and Goldstein, I.J. (1968) Immunology 14, 165-174 12611 Boyd, W.C., Shapleigh, E. and McMaster, M. (1955) Arch. Biochem. Biophys. 55, 226-234. [262] Petryniak, J., Pereira, M.E.A. and Kabat, E.A. (1977) Arch. Biochem. Biophys. 178, 118-134. [263] Wood, C., Kabat, E.A., Murphy, L.A. and Goldstein, I.J. (1979) Arch. Biochem. Biophys. 198, 1-1 1. [264] Chien, S.-M., Lemanski, T. and Poretz, R.D. (1974) Immunochemistry 1 I , 501-506. [265] Poretz, R.D. and Watkins, W.M. (1972) Eur. J. Biochem. 25, 455462.

452 [266] [267] [268] [269] [270] [271] [272] [273] [274] [275] [276] [277] [278] [279] [280] [281] [282] [283] [284] [285] [286] [287] [288] [289] [290] [291] [292] [293] [294] [295] [296] [297] [298] [299] [300] [301] [302] [303] [304] [305] [306] [307] [308] [309] [310] [31 I] [312] [313]

Edge, A.S.B. and Spiro, R.G. (1985) J. Biol. Chem. 260, 15332-15338. Wu, A.M., Wu, J.H. and Shen, F. (1994) Biochem. Biophys. Res. Commun. 198, 251-256. Wold, A.E., Motas, C., Svanborg, C. and Mestecky, J. (1994) Scand. J. Immunol. 39, 195-201. Cifonelli, J.A. and Smith, E (1955) Anal. Chem. 27, 1639-1649. Goldstein, I.J., Poretz, R.D., So, L.L. and Yang, Y. (1968) Arch. Biochem. Biophys. 127, 787-794. Smith, E.E., Guja Smith, Z.H. and Goldstein, 1.J. (1968) Biochem. J. 107, 715-724. Goldstein, I.J., Cifonelli, J.A. and Duke, J. (1974) Biochemistry 13, 867-870. Mandal, D.K. and Brewer C.F. (1992) Biochemistry 31, 12602-12609. Ebisu, S., Lonngren, J. and Goldstein, I.J. (1977) Carbohydr. Res. 58, 187-191. Ouchterlony, 0. (1958) Prog. Allergy 5, 1-78. Goldstein, I.J. and So, L.L. (1965) Arch. Biochem. Biophys. I 1 I , 407414. Oudin, J. and Williams, C.A. (1971) In: C.A. Williams and M.W. Chase (Eds.), Methods in Immunology and Immunochemistry. Academic Press, New York, pp. 103-234. Sijtsma, L., Wouter, J.T.M. and Hellingwerf, K.J. (1990) J. Bacteriol. 172, 7126-7130. Iraqui, F., Zalazar, F.E., Chiabrando, G.A., Romero, 0. and Vides, M. (1992) J. Immunol. Methods 156, 199-204. Weetall, H.H. and Weliky, N. (1965) Biochim. Biophys. Acta 107, 150-152. Ogata, S., Muramatsu, T. and Kobata, A. (1975) J. Biochem. 78, 687-696. Krusius, T., Finne, J. and Rauvala, H. (1976) FEBS Lett. 72, 117-120. Kobata, A. and Yamashita, K. (1989) Methods Enzymol. 179, 46-54. Merkle, R.K. and Cummings, R.D. (1987) Methods Enzymol. 138, 232-259. Momoi, T., Tokunaga, T. and Nagai, Y. (1982) FEBS Lett. 141, 6-10. Torres, B.V. and Smith, D.F. (1988) Anal. Biochem. 170, 209-219. Smith, D.F. and Torres, B.V. (1989) Methods Enzymol. 179, 3 0 4 5 . Cummings, R.D. (1994) Methods Enzymol. 230, 66-86. Kobata, A. and Endo, T. (1992) J. Chromatogr. 597, 11 1-122. Wood, J.G. and Sarinana, F.O. (1975) Anal. Biochem. 69, 320-322. Tanner, M.J.A. and Anstee, D.J. (1976) Biochem. J. 153, 265-270. Robinson, P.J., Bull, F.G., Anderton, B.H. and Roitt, I.M. (1975) FEBS Lett. 58, 330-333. Gurd, J.W. and Evans, W.H. (1976) Can. J. Biochem. 54, 477480. Burridge, K. (1976) Proc. Natl. Acad. Sci. USA 73, 44574461. Guzman, N., Berg, R.A. and Prockop, D.J. (1976) Biochem. Biophys. Res. Commun. 73, 279-285. Burridge, K. (1978) Methods Enzymol. 50, 54-64. Towbin, H., Staehelin, T. and Gordon, J. (1979) Proc. Natl. Acad. Sci. USA 76, 43504354. Glass, W.F., Briggs, R.C. and Hnilica, L. (1981) Anal. Biochem. 115, 219-224. Hawkes, R. (1982) Anal. Biochem. 123, 143-146. Gershoni, J.M. and Palade, G.E. (1982) Anal. Biochem. 124, 396405. Gershoni, J.M. and Palade, G.E. (1983) Anal. Biochem. 131, 1-15. Magnani, J.L., Brockhaus, M., Smith, D.F., Ginsburg, V., Blaszczyk, M., Mitchell, K.F., Steplewski, Z. and Koprowski, H . (1981) Science 212, 55-56. Magnani, J.L., Brockhaus, M., Smith, D.F. and Ginsburg, V. (1982) Methods Enzymol. 83, 235-241. Smith, D.F. (1983) Biochem. Biophys. Res. Commun. 115, 360-367. Canas, B., Dai, Z., Lackland, H., Poretz, R. and Stein, S. (1993) Anal. Biochem. 21 I , 179-182. Grabar, P. and Williams, C.A. (1953) Biochim. Biophys. Acta 10, 193-194. Laurell, C.-B. (1965) Anal. Biochem. 10, 358-361. Bug-Hansen, T.C. (1973) Anal. Biochem. 56, 480488. Berg-Hansen, T.C. and Takeo, K. (1989) Electrophoresis 10, 81 1-870. Heegard, N.H. and Berg-Hansen, T.C. (1990) Appl. Theor. Electrophor. 1, 249-259. Hansen, J.-E., Heegard, P.M.H., Ssren, P.J., Bent, N.-P. and Bug-Hansen, T.C. (1988) Electrophoresis 9, 273-278. Taketa, K. (1991) J. Chromatogr. 569, 229-241. De Graaf, T.W., Van der Stelt, M.E., Anbergen, M.G. and van Dijk, W. (1993) J. Exp. Med. 177, 657-666.

453 Wojczyk, B.S., Hoja, D.A. and Litynska, A.M. (1992) Int. J. Biochem. 24, 1561-1571. Taketa, K., Fujii, Y. and Taga, H. (1993) Electrophoresis 14, 1333-1337. Magna, D., Seta, N., Lebrun, D., Durand, G. and Durand, D. (1992) Clin. Chem. 38, 1418-1424. Shimazu, K., Taniichi, T., Satomura, S., Matsuura, S., Taga, H. and Taketa, K. (1993) Clin. Chim. Acta 214, 3-12. [318] Taketa, K., Endo, Y.,Sekiya, C., Tanikawa, K., Koji, T., Taga, H., Satomuura, S., Kawai, T. and Hirai, H. (193) Cancer Res. 53, 5419-5423 [319] Poretz, R.D. and Pieczenik,G. (1981) Anal. Biochem. 115, 170-176. 13201 Kaladas, P.M., Goldberg, R. and Poretz, R.D. (1983) Mol. Immunol. 20, 727-735. [321] Fournet, B., Leroy, Y., Wieruszeski, J.-M., Montreuil, J. and Poretz, R.D. (1987) Eur. J. Biochem. 166, 321-324. [322] Lee, S.-O., Connolly, J.M., Ramirez-Soto, D. and Poretz, R.D. (1990) J. Biol. Chem. 265, 5833-5839. [323] Ramirez-Soto, D. and Poretz, R.D. (1991) Carbohydr. Res. 213, 27-36. [324] Rago, R., Ramirez-Soto, D. and Poretz, R.D. (1992) Carbohydr. Res. 236, 1-8. [325] Mullis, K.G., Huynh, M. and Kornfeld, R.H. (1994) J. Biol. Chem. 269, 1718-1726. [326] Mullis, K.G. and Kornfeld, R.H. (1994) J. Biol. Chem. 269, 1727-1733. [327] Gold, E.R. and Balding, P. (1975) Receptor Specific Proteins. Plant and Animal Lectins. Elsevier, New York, pp. 151-236. [328] Stillmark, H. (1889) Uber Rizin, ein giftiges Ferment aus den Samen von Ricinus communis L. und einigen anderen Euphorbiaceen. Inaugural Dissertation, Dorpat. [329] Boyd, W.C. (1963) Vox Sang. 8, 1-32. [330] Watkins, W.M. and Morgan, W.T.J. (1952) Nature (London) 169, 825-826. [331] Makela, 0. (1957) Ann. Med. Exp. Biol. Fenn. 35 (Suppl. I I ) , 1-133. [332] Pereira, M.E. and Kabat, E.A. (1979) CRC Crit. Rev. Immunol. 1, 33-78. [333] Reisner, Y., Linker-Israeli, M. and Sharon, N. (1976) Cell. Immunol. 25, 129-134. [334] Aub, J.C., Tieslau, C. and Lankester, A. (1963) Proc. Natl. Acad. Sci. USA 50, 613-619. [335] Aub, J.C., Sanford, B.H. and Cote, M.N. (1965) Proc. Natl. Acad. Sci. USA 54, 396-399. [336] Burger, M.M. and Martin, G.S. (1972) Nature New Biol. 237, 9-12. [337] Rapin, A.M. and Burger, M.M. (1974) Adv. Cancer Res. 20, 1-91. [338] Yamishita, K., Tachibana, Y.,Okhura, T. and Kobata, A. (1985) J. Biol. Chem. 260, 3963-3969. [339] Holmes, E.H., Hakomori, S. and Ostrander, G.K. (1987) J. Biol. Chem. 262, 15649-15658. [340] Yousefi, S., Higgens, E., Daoling, Z., Pollex-Kruger, A,, Hindsgaul, 0. and Dennis, J.W. (1991) J. Biol. Chem. 266, 1772-1782. [341] Easton, E.W., Bolscher, J.G.M. and van den Eijnden, D.H. (1991) J. Biol. Chem. 266, 21674-21680. [342] Josephson, S.L., Weik, R.R. and John, D.T. (1977) Am. J. Trop. Med. Hyg. 26, 856-858. [343] Martinez-Palomero, A,, Gonzales-Robles, A. and de la Torre, M. (1973) Nature (London) 245, 186187. [344] Tissl, D., Martinez-Palomero, A. Arguello, C., de la Torre, M. and de la Hoz, R. (1977) J. Exp. Med. 145, 652-665. [345] Raychowdhury, M.K., Goswami, R. and Chakrabarti, P. (1982) Ind. J. Exp. Biol. 20, 748-750. [346] Davidson, S.K., Keller, K.F. and Doyle, R.J. (1982) J. Clin. Microbiol. 15, 547-553. [347] Pereira, M.E.A., Loures, M.A., Villalta, F. andde Andrade, A.F.B. (1980) J. Exp. Med. 152, 1375-1392. [348] Katzen, A.M. and Colli, W. (1983) Biochem. Biophys. Acta 7 2 7 , 4 0 3 4 1 I . [349] de Andrade, A.F., Estaves, M.J., Angluster, J., Gonzales-Perdomo, A. and Goldenberg, S. (1991) J. Genet. Microbiol. 137, 2845-2849. [350] Kellens, J.T., Jacobs, J.A., Peumans, W.J. and Stobberingh, E.E. (1994) J. Med. Microbiol. 41, 14-19. [351] Kellens, J.T., Jacobs, J.A., Peumans, W.J. and Stobberingh, E.E. (1993) J. Med. Microbiol. 3 9 , 4 4 0 4 4 5 . [352] Moyes, A. and Young, H. (1992) J. Med. Microbiol. 37, 51-55. [353] Jones, J.W. (1993) J. Clin. Pathol. 46, 761-763. [354] Corbel, M.J. and Abraham, A.M. (1992) FEMS Lett. 69, 245-248. [355] Moyes, A. and Young, H. (1993) Brit. J. Biomed. Sci. 50, 295-300. [356] Zajicek, I? and Lukes, J. (1992) Folia Parasitol. (Praha) 39, 195-200. [357] Haido, R.M., Estaves, M.J. and Barreto-Bergter, E. (1992) J. Protozool. 39, 609-612. [314] [315] [316] [317]

454 [358] Dias Filho, B.P., de Andrade, A.F., de Souza, W., Estaves, M.J. and Angluster, J. (1992) Microbios 71, 55-64. [359] Sharon, N. (1968) In: I.E. Liener, N. Sharon and I.J. Goldstein (Eds.), The Lectins. Properties, Functions and Applications in Biology and Medicine. Academic Press, New York. pp. 494-526. [360] Cottin, J., Loiseau, D., Robert, R., Mahoza, C., Carbonnelle, B. and Senet, J.M. (1990) FEMS Lett. 56, 3 0 1-305. [361] Li, J.G. and Osgood, E.E. (1949) Blood 4, 670475. [362] Reisner, Y., Ravid, A. and Sharon, N. (1976) Biochem. Biophys. Res. Commun. 72, 1588-1591. [363] Reisner, Y., Itzicovitch, L., Meshorer, A. and Sharon, N. (1978) Proc. Natl. Acad. Sci. USA 75, 29332936. [364] Reisner, Y., Biniaminov, M., Rosenthal, E., Sharon, N. and Ramot, B. (1979) Proc. Natl. Acad. Sci. USA 76, 4 4 7 4 5 1. [365] Gillespie, W., Paulsen, J.C., Kelm, S., Pang, M. andBaum, L.G. (1993) J. Biol. Chem. 268, 3801-3804. [366] Bauman, J.G.J., Wagemaker, G. and Visser, J.W.M. (1986) J. Cell. Physiol. 128, 133-142. I3671 Reisner, Y., Kapoor, N., O’Reilly, R.J. and Good, R.A. (1980) Lancet Dec. 20/27, 1320-1324. [368] Okada, T., Higuchi, M., Maruyama, T., Imai, Y. and Osawa, T. (1990) Br. J. Cancer 61, 241-249. [369] Poretz, R.D., Tang, M. and Vucenik, I. (1986) Immunol. Invest. 15, 521-529. [370] Reisner, Y., Kapoor, N., Kirkpatrick, D., Pollack, M.S., Dupont, B., Good, R.A. andO’Reilly, R.J. (1981) Lancet 2, 327-33 I . [371] Reisner, Y., Kapoor, N., Kirkpatrick, D., Cunningham-Rundles, S., Dupont, B., Hodes, M.Z., Good, R.A. and O’Reilly, R.J. (1983) Blood 61, 341-363. [372] Visser, J.W.M., Bauman, J.G.J., Mudler, A.H., Eliason, J.F., de Leeuw, A.M. (1984) J. Exp. Med. 59, 1576-1 590. [373] Rhodes, E.G.H., Baker, I?, Rhodes, J.M., Davies, J.M. and Cawley, J.C. (1991) Exp. Hematol. 19, 833837. [374] Chambers, W.H., Brumfield, A.M., Hanky-Yanez, K., Lakomy, R., Herberman, R.B., McCaslin, D.C., Olszowy, M.W. and McCoy, J.I? (1992) J. Immunol. 148, 3658-3665. [375] Lebkowski, J.S., Schain, L.R., Okrongly, D., Levinsky, R., Harvey, M.J. and Okarma, T.B. (1992) Transplantation 53, 101 1-1019. [376] Schwartz, E., Lapidot, T., Gozes, D., Singer, T.S. and Reisner, Y. (1987) J. Immunol. 138, 460465. [377] Wagner, J.E. (1992) Semin. Hematol. 29, 6-9. [378] Molday, R.S. (1977) Nature (London) 268, 437438. [379] Antoine, J.-C., Ternynck, T., Rodngot, M., Avrameas, S. (1978) lmmunochemistry 15, 443452. [380] Haller, O., Gidlund, M., Hellstrom, U., Hammarstrom, S. and Wigzell, H. (1978) Eur. J. Immunol. 8, 765-771. [381] Vose, B.M., Blackledge, G., Crowther, D. and Gallaher, J. (1982) Immunology 46, 619-627. [382] Harris, D.T., Iglesias, J.L., Argov, S., Toomey, J. and Koren, H.S. (1987) J. Leukocyte Biol. 42, 163-170. [383] Frydecka, I., Slesak, B., Benczur, M. and Harlozinska-Szmyrka, A. (1987) J. Natl. Cancer Inst. 78, 1145-1 148. [384] McCoy, J.P., Hanley-Yanez, K., Brumfield, A.M., Herberman, R.B. and Chambers, W.H. (1990) Cell. Immunol. 127, 275-283. [385] McCoy, J.I?, Chambers, W.H., Lakomy, R., Campbell, J.A. and Stewart, C.C. (1991) Cytometry 12, 268-274. [386] Gutowski, K.A., Hudson, J.L. and Aminoff, D. (1991) Exp. Gerontol. 26, 315-326. [387] Bratosin, D., Mazurier, J., Debray, H., Lecocq, M., Boilly, B., Alonso, C., Moisei, M. Motas, C. and Montreuil, J. (1995) Glycoconj. J. 12, 258-267. [388] Nicolson, G.L. and Singer, S.J. (1971) Proc. Natl. Acad. Sci. USA 68, 942-945. [389] Nicolson, G.L. and Singer, S.J. (1974) J. Cell Biol. 60, 236-248. [390] Danguy, A,, Akif, F., Pajak, B. and Gabius, H.-J. (1994) Histol. Histopathol. 9, 155-171. [391] Li, W.-P., Zuber, C. and Roth, J. (1993) Histochemistry 100, 347-356. [392] Ashford, D., Dwek, R.A., Rademacher, T.W., Lis, H. and Sharon, N. (1991) Carbohydr. Res. 213, 215227.

455 [393] Kurosaka, A., Yano, A. Itoh, N., Kuroda, Y., Nakagawa, T.and Kawasaki, T. (1991) J. Biol. Chem. 266, 4 1 6 8 4 1 72. 13941 Katz, F., Moats, W. and Jan, Y.N. (1988) EMBO J. 7, 3471-3477. [395] Prenner, C., Mach, L., Glossl, J. and Man, L. (1992) Biochem. J. 284, 377-380. [396] Staudacher, E., Altman, F., Man, L., Hard, K., Kamerling, J.P. and Vliegenthart, J.F. (1992) Glycoconj. J. 9, 82-85.

Appendix This appendix consists of three tables and the list of references for these tables. Table A1 lists the properties of purified plant lectins, Table A2 summarizes representative applications of lectins to biomedical research, and Table A3 provides a classification of the species listed in Table A l .

Table A l Properties of purified plant lectins Lectin

Molecular weight Native KDa

Subunit KDa

11

11

Subunit

Specificitya

Cells agglutinatedb

Mana6[Mana3]Man > Mana6(3)Man

Rab

Reference(s)

Mannose-specific group

Allium ascalonicum A. satiuum A. ursinum Cymbidium hyb. Epipactis helleborine

monomer

Mana3ManOR > Mana6[Mana3]ManOR Mana6[Mana3]ManOR > Mana3ManOR 25

12.5

a2

Man, ASiaFetuin= Thyroglobulin

Rab

25

12

a2

Man, ASiaFetuin = Thyroglobulin

Rab

a4

Mana6[Mana3]Man > Mana3Man > (Mana6)3

Rab,

(Ma11a6)~> Mana6[Mana3]ManOR

Galanthus niualis Hippeastrum hyb.

50 56

14.3

a4

Leucojum aestiuum L. uernum Listera ouata

25

13

a2

Narcissus lobularis N pseudonarcissus Tulipa gesneriana I

13

RabT

25

13

25

12.5

25

13

a2

(Mana6), > Mana6[Mana3]ManOR

67

17

a4

(Mar1a6),,~ > (ManaZ),,,

47

a13.3; p11.9

azp2

Man,_,GlcNAc

106

26.5 a13; PI4

a4

MeaMan > aMan > Glc > GlcNAc

ABO

a4

aMan > aGlc

Rab

azpz

aMan > aGlc > GlcNAc

ABO

a2

Thyroglobulin

W T ,

a2

(Mana3),_, > Mana3ManOR > (Mar1a3)~

Rab

(Mana6 > 2)2

Rab, Yeast

Mannose/glucose group

Bowringia mildbraedi Canaualia ensformis Cliuia miniata Cratylia mollis (1 & 2) Dioclea grandiflora

100

Lathyrus odoratus

52

25

ab Rab>AO>>B

31(1); 60(2) 25 a5.8; g20

continued on next page

Table A 1, continued ~

Lectin

Molecular weight

Subunit

Cells agglutinatedb

Reference(s)

aMan > aGlc > GlcNAc aMan > aGlc MeaMan >Man > aGlc, GlcNAc

ABO

[71 [71

aMan > aGlc 3-OMeGlc > 2.3-diOMeGlc > MeaMan > Man > Glc

Cat >> Rab

structure

Native KDa

Subunit KDa

L. sativus L. tingitanus Lens culinaris

49 50 46

a4.4; 819 a5.0; 820 (15.7; 817.5

Onobrychis viciifolia Pisum satiuum

53 50

26.5 a5.7; 817

Vicia cracca (Man) Y ervilia

a5.8; 817.5 a4.7; 821

Yfaba

44 53 52.5

a5.6; 820.7

a262

Man,Glc,Fruc,MeaMan,Mal,aa’Tre 2,3-diOMeMeaGlc > 3-OMeGlc > 3-OBzGlc > Man

T sativa

40

a 6 ; 814

a282

Melz > MeaMan > Man > Glc

Brachypodium syluaticum Chelidonium majus

36 23

18

a2

(GlnNAc(34), > GlcNAc

a9.5; 81 1.5

aB

(GlcNAc(34)4,3,, >> GIcNAc

Coccinea indica Codium fragile C. tomentosum

64 >60

32

a2

Cyphomandra betacea Cytisus sessilifolius

50

a282

a282 a282 a2

a282 a2B2

a282

ABO ABO ABO ABO ABO ABO (Sase) Rab > A > BO

N-Acetylglucosamine group

ABO ABO

GalNAc > GlcNAc Rab, H,

16.3; 15.6 (GlcNAc(34)4 > (GlcNAc(34)3 GlcNAc(34GlcNAc > ~-Fuca2Gal~4GlcNAc

Rab 0

monomer

(GlcNAc[34)3 > (GlcNAc(34)z > GIcNAc (Gl~NAc(34),,~= LacNAc > GlcNAc

a4

GlcNAca3GallGlc > GlcNAc~3Gal,GlcNAca6Gal

ABO ABO Tk

25

a2

a8

110

Datura stramonium

86

a40; (346

Ficus cunia‘ Griffonia simplicifolia I1

3.5 1 I3

3 .5 30

Hordeum uulgare Lufa acutangula

36

18

48

24

~~~~~~

Specificitya

a2

(GlcNAcfM), > GlcNAc

a2

(G1cNAc(34)5>4,3>2

ABO Rab

[71 [71 [71 [71 [71 [71 ~71

Table A l , continued Lectin

Molecular weight Native KDa

Subunit

Specificitya

monomer

Lycopersicon esculentum

71

71

Oryza satiua

38 19

18-19 -

monomer

40

monomer

Secale cereale

40 36

(G1cNAc(341)4>3>2 (GlcNAc(34)"> GlcNAc (GlcNAc(34)" GlcNacfi3Gal(34GlcNAc > GlcNAc!36GalppNC~

a2

(GlcNAc(34)">> GlcNAc

Solanum tuberosum

100

Triticum uulgaris

43.2

Wlex europaeus I1

105 8.5

Phytolacca americana Psathyrella uelutina

Urtica dioica

GalactoselN-acetylgalactosarninegroup 135

Abrus precatorius (Agglutinin)

Cells agglutinated

Subunit KDa

18 50 21.6 23-25 8.5

a33; 036; p'37.5

a2

a2 a2

(GlcNAC(34),_5 (GlcNAc(34), > GlcNAcP > NeuSAc

monomer

(GicNAc(34)4>3,2 GlcNAc oligomers

a2w

fiGal> GalNAc

a4

O>AB ABO 0

ABO ABO ABO 0 >AB Rab, ABO

A . precatorius (toxin A)

62.5

a31; a'33; 836

A. precatorius (toxin C) Adenia digiiata A . digitata (6B)

67 5743

a29; 636 a28; p35 a27; 031

Gal > Me1 > Lac,Raf,o-Fuc >> L-FUC

Rab > ABO

Aegopodium podagraria

480 58.5

60(a,a')

GalNAc >Lac > Fuc > Gal > Me1 Galf33GalNAc-T,S

ABO ABO

Amaranthus caudatus

62

A . cruentus

66

33-36 35

(NeuSAc)Galfl3GalNAc> GalNAca3Gal Fetuin >> GalNAc; GalOR = 0

ABO Rab,ABO

A . leucocarpus

66

3346

ASiaFetuin > Fetuin > Human IgG > >GalNAc

H~ > H ~

65.8

37

ASiaFetuin > Fetuin >> GalNAc

Rab, ABO

Agaricus bisporus

A. spinosus

57

pGal> GalNAc

Reference(s)

Table A l , continued Lectin

Molecular weight Native KDa

Subunit KDa

Amphicarpa bracteata

135

Arachis hypogaea

110

a28.5; 836; y32 27

A. integrifolia Bauhinia purpurea

54 195

a12; PI5 44

Boletus satanas Botiytis cinerea

63 34 61

63 17

Subunit

a4

18; 14.5

Artocarpus altilis

Biyonia dioica Caragana arborescens Ceratobasidium cornigerum

103

a32; 830 26

Clerodendron irichotomum

56

28

Codium fragile (see under GlcNAc group) Colchicum autumnale

100

a15; 810 31

Cratylia mollis (3)

Specificitya

Cells agglutinatedb

GalNAca3GalNAc > GalNAca3Gal Galb3GalNAc > a,fiGal; Lac > PGal

A1 T cells

Galp3GalNAc

a2

Gal83GalNAc; MeaGal >> MePGal Galfi3GalNAc > a,BGalNAc > a,PGal Gal MeaGal > GalNAc, Gal, Me1

aP

GalNAc

P

a3 a4

monomer

a4

GalNAc >Gal, eomplex GalNAc

H, Rat Rab, ABO 0 >AB

GalNAcOR (a@= !3@ > Me > OH) > Lac > Gal

0

a484

Lac > GalN > GalNAc >Gal (Ho only) Gal

ABO, Rab ABO

a4?

Lac,Mel> GalNAc >Gal

ABO (Sase)

[591

A,

>> A2

[71

120 75-90?

31 31

Dolichos biflorus

110

27.3; 27.7

a282

GalNAca3GalNAc >> GalNAc

>50 5648

a30; 828 26-33 30.9

aP

GalNAc

dimer dimer

Galfi4GlcNAc > GalNAc > Gal Lac,Gen,MeOGal> MeaGal,Gal> GalNAc

Erythrina spp. g Euphorbia marginata

ABO Rab > ABO

a2

Crotalaria juncea Cyiisus scoparius Echinocystis lobata Eranthis hyemalis

[581 [81 171

a4

ABO

52.5

Reference(s)

Structure

ABO ABO, Rab

[601 [61,621

[71 ~ 3 1 continued on next page

$, W

P m 0

Table A l , continued Lectin

Molecular weight Native KDa

Subunit KDa

Falcata japonica

125

Galactia tashiroi

90 120 114 I20

34 24

Glycine max Grifonia simplicifolia I d Hura crepitans Iris germanica

30 a32; 033 31 35; 30 a21; 634

Iris x hollandica

60

Lactarius deliciosus

37 37

a18; B19

4046 108

a10; p12

L. deterrimus Maclura pomifera Macrotyloma axillare’ Moluccella laeuis Momordicn charantia Mucuna deeringiana Phaseolus lunatus’ Phaseolus mungo Pleurotus ostreatus

130 129

110 124 137

18

a,1327 26-46 a29; 836

Subunit

Specificitya

Cells agglutinatedb

Reference(s)

GalNAc(a@)O@>> GalNAc > Gal >> GlcNAc,GalN

A, > A 2 > A ,

[64,651

GalNAc >> Gal,MeGal,Lac GalNAc > Gal a4,GalNAc >Gal; @,Gal

ABO > Rab A>>B

[661 [71

A(a4), B(B4) Shp > ABO AB>B

[71 [71 ~671

R4T) ABO ABO O>A>B

[681 ~ 9 1 [701 17~71,721

Structure

GalNAc > Ga1,MepGlc GalNAc >Lac > Gal GalNAca3Gal> GalNAca6Gal> GalNAcOMe GalB3GalNAc > GalB3GlcNAc, GalNAc-1 P Galfi3GalNAc pN@aGalNAc > pN@aGal >> GalNAc > Gal GalNAc GalNAcaOR > GalNAcBOR >> GalOR Lac >> MeflGal> a,BGal> GalNAc

A, [71 BONN> AMM> OMM [73,74] ABO

[71

RabT, Shp

[751

A , >A2 RabT

[71 [761

GalNAc >Lac > Gal, Me(a,P)Gal> Raf > Me1

ABO>Rab

[771

29(a,B,v) a30; 634

Gal > GalNAc; Lac > Me1 BGal> aGal >> GalNAc a,pGal> GalNAc GalNAc, human a-glycoproteins

ABO (t. as) ABO ABO

[71 [71 [71 [78-811

a,a’,P,3 1 66 a44; @ l

GalNAca3GalNAc > GalNAc >Gal GalNAca3[~-Fuca2]Gal>GalNAc MeaGal> pN@aGal> Me1 > Gal > GalNAc

54

Psophocarpus tetragonolobus

87 52

Ricinus communis I

120

R. communis 11 (Ricin) Robinia pseudoacacia I11 Sambucus nigra I1

63 105

a30; 034

51

30

29; 30.5

GalNAca3/6Gal> GalNAc > GalNAcfi6Gal

Rab > ABO

[34,821 continued on next page

Table A l , continued ~~

~

Lectin

Molecular weight Native KDa

Subunit KDa

Sarothamnus welwitschii

64

Sclerotinia miyabaena

34

21.5 17-18

132.8

32.5(a,b)

Sophora japonica

Specificitya

Cells agglutinatedb

a3

GalNAc > MeaGal> Lac,Mel

A,>A>BO Rab,

a2B2

MeaGal> GalNAc > Lac > Gal; AS-Much > M u c h GalNAc > Gal

Subunit Structure

26

Sphenostylus stenocarpus

GlcNAc > Lac > Raf,Gal MeBGal >Lac > GalNAc,MeaGal,Mel

Telfairia occidentalis

180

Tetracarpidium conophorum 1

70 34

a,(330 34 34

64

a33; (329

aB

Fuca2Galp4GlcNAc > (Galp4GlcNAc)4,3,2

56 114, 125

a26; (330 33

aB

Lac > BGal

105

T conophorum I1 Trichosanthes japonica (11) L kirilowii Vicia cracca I , I1

(a(3)3 as-Sa monomer

GalNAc > Me(3Gal > Lac > Ga1,MeaGal Lac > Me(3Gal> Gal,MeaGal,GalNAc

B > A >> O(H) 0 >AB ABO A A O>>B > A A, >A2

a4

a4

(R = T or S peptide) (G~IB~G~INAc-OR),

A B O ~> A B O ~

an(34-n

GalNAc >> Gal

At(a4); Tn((34)

a29; (334

a(3

Gal >> GalNAc

ABO, animal

60

30

a2

GalNAca6Gal> aGalNAc > BGalNAc

ABO

Grifonia simplicifolia IV

72 56.6

31 a27; (329

aB

Mea-L-Fuc > Lea,Leb ~-Fuca2Gal(33[~-Fuca4]GlcNAc

ABO Leb

Lotus telragonolobus A

120

a4

Fuca6GlcNAc > Fuca2Ga1~4[Fuca3]GlcNAc

O(H)

Lotus tetragonolobus B

58

27.8 27

Ulex europaeus I

60

a29; 831

aB

!l graminea F uillosa

110-

26 a34; (336

Viscum album

120 60

Wistaria floribunda

Reference(s)

L-Fucose group

Aleuria aurantia

a2

OW)

a2

Fuca2Gal@GlcNAc~> FucaZGal(34Glc> MeaFuc

O(H) continued on next page

P

0'

tR N Table A l , continued Lectin

Molecular weight

Subunit

Specificitya

Cells agglutinated

Reference(s)

a2P2

NeuSGc >NeuSAc > NeuSAca3'Lac > GalUA

Pig > A,O > B

~ 9 1

a4

NeuSAca3Galp4GlcNAc

[ 12,90-941

34.5; 37.5

a2P2

Neu5Aca6Gal/GalNAc >> Lac

[12,93-971

a38; 832

a6

Neu5Aca6Galp4GlcNAc, S03-6Gal@GlcNAc

[981

Native KDa

Subunit KDa

Hericium erinaceum

54

a1S; 616

Maackia amurensis

130

33

Sambucus nigra

140

Trichosanthesjaponica (I)

70

Sialic acid group

Complex group (not inhibited by monosaccharides)

Alocasia indica Boodlea coacta (A-D)

55 14-20

13 14-20

a4

ASiafetuin (3 Galfi3GalNAc?)

Rab, GP, Rat

monomers

High mannose glyconjugates, mannan

Rab, ABO(P, BO>A2,H,

43

26

a2

Human IgM, bov. fetuin

126.7

3s

a4

Gala3[Fuca2]LacNAc > Fuca2LacNAc

Marchartia polymorpha

20

18

monome1

Subs. A, B; IgG, lactalbumin

Phaseolus coccineus (I, 11)

120

f! coccineus

112

Cicer arietinum Euonymus europaeus

'

0 > Rab, Bov. ABO ABO

a4

Phaseolus vulgaris Erythroagglutinin

11s

30

a4

Tetraantennary octasaccharide

0

Leucoagglutinin

115

30

a4

(Galp4Gl~NAc)~P2,6Man

Leucocytes

145,175

34.5, 37, 39

?

ASiaFetuin > Thyroglob. > fetuin > orosomucoid

Rab, ABO

Pinellia terneata

40

10-1 1

a4

Mannan, thyroglobulin

Rab

Robinia pseudoacacia I

59

34

a2

Ovomucoid

40

a26; 614

k? vulgaris'

Sclerotium rolfsii Tulipa gesneriana I1

45

Hog gastric mucin > bov. submax. ASiamucin Thyroglobulin > mucin, orosomucoid

E. coli Mouse, rat

continued on next page

Table A l , notes Abbreviations for oligosaccharides and substituents: Gen, gentibiose; Lac, lactose; Mal, maltose; Melz, melezitose; Raf, raffinose; Tre, trehalose; Asia, asialo-; Bz, benzyl; Me, methyl; NCP, nitrophenyl; CP, phenyl. Abbreviations for blood cells: Rab, rabbit; RabT, trypsinized rabbit cells; H, human, type unspecified; ABO, human, type specified; Bov. bovine; Shp, sheep; Sase, sialidase-treated. Ficus semicordata (J.E. Smith) is preferred over E: cunia (Hamilton ex Roxburgh). Griffonia is accepted over Bandeiraea. Synonymous with Amaranthus hypochondriacus. Synonymous with Artocarpus heterophyllus. g The following species of Erythrina possess nearly identical lectins: arborescens, caffra, corallodendron, cristagalli, Jadelliformis, latissima, litosperma, lysistemon, perrieri, stricta, suberra, zeyheri. Falcata japonica (Oliver) V Komarov = Amphicarpa edgeworthii Benth. i Macroiyloma axillare (E. Mey) Verde, formerly Dolichos axillaris E. Mey. Synonymous with I? limensis. Nearly identical lectins are found in S. minor. S. trifoliorirm, and 3 isolates of S. sclerotiorum. Euonymus has been incorrectly spelled and indexed as “Euonymus” in some literature and data bases; europaeus (not europaea) is the correct form of the specific epithet. Variety ’Xlubia”. ” Variety Red Kidney Bean. O Variety “Great Northern”. a

*

J

’

464 Table A2 Representative applications of lectins to biomedical research Lectin

Application

Reference(s)

Mannose/glucose group Bowringia mildbraedi

Effect on infectivity of HIV

Cymbidium hyb.

Inhibitor of HIV and CMV replication in uitro

Epipactis helleborine

Inhibitor of HIV and CMV replication in vitro

Galanthus nivalis

Effect on infectivity of HIV Purification of HIV and SIV envelope glycoproteins Glycosylation of rat spermatids Purification of mouse and human glycoproteins Typing 6-hemolytic Streptococci

Lens culinaris

Analysis of human a-fetoprotein

Leucojum aestivum

Effect on infectivity of HIV

Listera ouata

Separation of a-o-mannans

Characterize strains of Neisseria gonorrhoeae

Inhibitor of HIV and CMV replication in uitro Narcissus lobularis

Effect on infectivity of HIV

N-Acetylglucosamine group Codium tomentosum

Interaction with intestinal parasites

C. fragile

See C. tomentosum

Datura stramonium

Purification of poly-LacNAc glycopeptides

Griffonia simplicifolia I1

Detect blood-group related poly-LacNAc in tissues Characterize strains of Neisseria gonorrhoeae Detect human colonic carcinoma glycoproteins

Lycopersicon esculen turn

Agglutinate type B streptococci type Staphylococcus aureus strains Identify pathogenic trypanosomes

Psathyrella velutina Solanurn tuberosum

Study of glycosylation abnormalities of rheumatoid IgG Agglutinate type B streptococci Characterize strains of Neisseria gonorrhoeae

Triticum uulgaris

Numerous

Wlex europaeus I1

Trace maturation of rat kidney cells

Wrtica dioica

Induction of human IFN-g in lymphocytes Activate specific T-cells Inhibit HIV and CMV replication in vitro

GalactoselN-acetylgalactosamine group Aegopodium podagraria

Typing P-hemolytic Streptococci

Arachis hypogaea

Numerous

[341 [120,125-1331 continued on next page

465

Table A2, continued Lectin

Application

Artocarpus integrifolia

Separation of 0-linked glycoproteins

Bauhinia purpurea

Isolation of specific B- and T-cell populations

Reference(s)

Identify pathogenic trypanosomes Boletus satanas

Release of IL-la and IL-2 Protein synthesis inhibitor

Biyonia dioica

Typing B-hemolytic Streptococci

Canavalia ensiformis

Numerous

Ceratobasidium cornigerum

Typing 0-hemolytic Streptococci

Clerodendron trichotomum

Screening of hematopoietic cells

Dolichos bgorus

Numerous

Eranthis hyemalis

Protein synthesis inhibitor

Erythrina cristagalli

Distinguish embryonic & mature rat kidney cells

Glycine m a

Remove T cells from bone marrow

Griffonia simplicifolia I

Detect Tn antigen in cancer cells

Differentiate types of alveolar macrophages Identify alveolar macrophages in respiratory disease Maclura pornifera

Rat T cell studies

Moluccella laeois

Typing of NM blood groups

Ricinus communis I

Trace development of chick thymus Assay for sialidase

Robinia pseudoacacia 111

Study of cyclic nucleotide phosphodiesterase in peripheral blood lymphocytes

Sambucus nigra 1

Typing fi-hemolytic Streptococci

Sclerotinia spp.

Typing 6-hemolytic Streptococci

Sophora japonica

Identify alveolar macrophages in respiratory disease Identify pathogenic trypanosomes

Sphenostylus stenocarpus

HIV infectivity studies

Tetracarpidium conophorum

Effect on HIV infectivity

Vicia graminea

Identification of polyagglutinating erythrocytes in hemoglobin variant

!I oillosa

Separate T-cell types Stain cancer cells Identify pathogenic trypanosomes

Viscum album

Specific staining of microglial cells Stain amyloid plaques in Alzheimer’s

L-Fucose group

AIeuria aurantia

Characterize fucosylated glycoproteins in brain

[1511 continued on next page

466

Table A2, continued Lectin Lotus tetragonolobus

Application

Reference(s)

Purify fucose-containing glycoconjugates Identify pathogenic trypanosomes Development of chick thymus

Ulex europaeus 1

Identify secretors Identify alveolar macrophages in respiratory disease Determine epithelial cell differentiation Characterize strains of Neisseria gonorrhoeae Identify pathogenic trypanosomes

Sialic acid group Maackia amurensis

Glycosylation of rat spermatids Characterization of sialyltransferases in Ehrlich ascites cells Characterization of sialidases Typing 6-hemolytic Streptococci

Sambucus nigra I1

Glycosylation of rat spermatids Characterization of sialyltransferases in Ehrlich ascites cells Characterization of sialidases Sialylation of neoplastic colon tissue Isolation of sialylated glycoconjugates Typing &hemolytic Streptococci

Complex group Euonymus europaeus

Purify human complement fractions

Phaseolus vulgaris

Trace development of optic tectum

Pinellia terneata

Motor nerve terminal studies Typing 6-hemolytic Streptococci

Robinia pseudoacacia I

~1541 [ 155,1561 [ 108,109]

[341

467 Table A3 Classification of species listed in Table Al Order and Family

Genus and species

Tissue used

Agaricaceae

Agaricus bisporus

Fruiting body

Boletaceae

Boletus satanas

Fruiting body

Coprinaceae

Fruiting body

Polyporaceae

Psathyrella veluiina Pleurotus ostreatus

Russelaceae

Lactarius spp.

Carpophore

Hericium erinaceum

Fruiting body

Botrytis cinerea

Mycelia, sclerotia

Algae & fungi (Thallophyta)

Agricales

Fruiting body

Aphyllophorales Hydnaceae Helotiales Sclerotinaceae

Sclerotinia spp.

Mycelia, sclerotia

Sclerotium roljsii

Culture fluid

Aleuria aurantia

Fruiting body

Pezizales Pezizaceae Siphonales Codiaceae

Codium spp.

Siphonocladiales Valoniaceae

Boodlea coacta

Tulasnellales Ceratobasidiaceae

Ceralobasidium cornigerum

Mosses & liverworts (Bryophyta)

Marchartiales Marchartiaceae

Marchartia polymorpha

Gametophyte

Brachypodium sylvaiicum Oiyza saliva Secale cereale

Germ Seed Germ

Triiicum vulgaris

Germ

Tulipa gesneriana

Bulb

Alliuma spp. Colchicum autumnale

Tuber

Seed plants (Spermatophyta) Monocotyledoneae

Glumiflorae Gramineae

Liliiflorae Liliaceae

Bulb

continued on next page

468

Table A3, continued Order and Family

Genus and species

Tissue used

Clioia miniata

Leaf

Galanthus nioalis

Bulb

Liliiflorae (cont a) Amaryllidaceae

lridaceae

Leucojum spp.

Bulb

Narcissus spp.

Bulb

Hippeastrum hyb.

Bulb

Iris germanica

Rhizome

Iris x hollandica

Bulb

Listera ovata

All

Microspermae (Orchidales) Orchidaceae

Cymbidium hyb.

All except root

Epipactis helleborine

All

Alocasia indica

Tuber

Pinellia terneata

Rhizome

Spathiflorae Araceae

Dicotyledoneae Urticales Moraceae

Artocarpus spp.

Seed

Ficus cunia

Seed

Maclura pomifera

Seed

Urtica dioica

Rhizome

t7iscum album

Vegetative

Amaranthaceae

Amaranthus spp.

Seed

Phytolaccaceae

Phytolacca americana

Root (all)

Eranthis hyemalis

Tuber

Chelidonium majus

Seed

Abrus precatorius

Seed

Urticaceae Santalales Loranthaceae Centrospermae

Ranales Ranunculaceae Rhoeadales Papaveraceae Rosa1es Leguminoseae (Fabaceae)

Amphicarpa bracteata

Seed

Arachis hypogaea Bauhinia purpurea

Seed Seed continued on next page

469

Table A3, continued Order and Family

Genus and species

Rosales

Bowringia mildbraedii

Seed

Canavalia ensifarmis Caragana arborescens

Seed

Cicer arietinum

Seed

Cratylia mollis Crotalaria juncea

Seed

Cytisus spp.

Seed

Dioclea grandiflora Dolichos biforus

Seed

-

Leguminoseae

(cant 3)

Seed

Seed

Seed

Erythrina spp. Falcata (=Amphicarpa) Galactia tashiroi

Seed

Glycine max

Seed

Seed Seed

Griffonia simplicifolia

Seed

Lathyrus spp. Lens culinaris Lotus tetragonolobus

Seed Seed

Maackia amurensis

Seed

Macrotyloma axillare

Seed

Seed

Mucuna deeringiana

Seed

Onobrychis uiciifolia Phaseolus spp.

Seed Seed

Pisum satiuum

Seed

Psophocarpus tetragonolobus

Tuber, seed

Robinia pseudoacacia

Seed

Sarothamnus welwitschii

Seed

Sophora japonica

Seed

Sphenostylus stenocarpus Ulex europaeus

Seed Seed

Vicia spp.

Seed

Kstaria floribunda

Seed

Geraniales Euphorbiaceae

Euphorbia marginata

Latex

Hura crepitans

Seed

Ricinus communis Tetracarpidium conophorum

Seed Seed

Euonymus europaeus

Seed

Sapindales Celastraceae

continued on next page

470

Table A3, continued Order and Family

Genus and species

Tissue used

Parietales Passifloraceae

Adenia digitata

Root

Umbelliflorae Umbelliferae (Apiaceae)

Aegopodium podagraria

Rhizome

Clerodendron frichotomum

Fruit Seed Fruit

Tubiflorae Verbenaceae Labiatiae Solanaceae

Moluccella laeois Cyphomandra betacea

Solanum tuberosum

Seed Fruit Tuber

Rubiales Caprifoliaceae

Sambucus nigra

Bark

Cucurbitales Cucurbitaceae

Bryonia dioica

Root Fruit (phloem) Seed Fruit (phloem) Seed Tuber Root

Datura stramonium Lycopersicon esculenturn

Coccineu indica Echinocystis lobata Luffa acutangula Momordica charantiu Trichosanthesjaponica

T kirilowii a

Allium is placed in Amaryllidaceae by some taxonomists. Some classifications elevate the three subfamilies of Leguminoseae to families in the Order Leguminales

References [I] Mo, H., van Damme, E.J.M., Peumans, W.J. and Goldstein, I.J. (1993) Arch. Biochem. Biophys. 306, 431438. [2] van Damme, E.J.M., Goldstein, I.J. and Peumans, W.J. (1991) Phytochemistry 30, 509-514. [3] Kaku, H., Goldstein, I.J., van D a m e , E.J.M. and Peumans, W.J. (1992) Carbohydr. Res. 229,347-353. [4] Animashaun, T. and Hughes, R.C. (1989) J. Biol. Chem. 264,46574663. 151 Chawla, D., Animashaun, T., Hughes, R.C., Harris, A. and Aitken, A. (1993) Biochim. Biophys. Acta 1202, 3 8 4 6 . [6] Animashaun, T., Mahmood, N., Hay, A.J. and Hughes, R.C. (1993) Antiviral Chem. Chemother. 4, 145-153. [7] Goldstein, I.J. and Poretz, R.D. (1986) in: 1.E. Liener, N. Sharon and I.J. Goldstein (Eds.), The Lectins: Properties, Functions, and Applicalions in Biology and Medicine. Academic Press, New York, pp. 33247. [8] Van Damme, E.J., Smeets, K., Van Leuven, F. and Peumans, W.J. (1994) Plant Mol. Biol. 24, 825-830. [9] Paiva, P.M.G. and Coelho, L.C.B.B. (1992) Appl. Biochem. Biotech. 36, 113-1 18.

47 1 [ 101 Van Damme, E.J., Smeets, K., Torrekens, S., Van Leuven, F. and Peumans, W.J. (1 994) Eur. J. Biochem. 221, 769-777. [I I] Shibuya, N., Goldstein, I.J., van Damme, E.J.M. and Peumans, W.J. (1988) J. Biol. Chem. 263,728-734. [12] Martinez-Menarguez, J.A., Aviles, M., Madrid, J.F., Castells, M.T. and Ballesta, J. (1993) Eur. J. Cell Biol. 61, 21-33. [I31 Gilljam, G. (1993) AIDS Res. Human Retrovir. 9, 431438. [I41 van Damme, E.J.M., Allen, A.K. and Peumans, W.J. (1988) Physiol. Plant. 73, 52-57. [I51 Kaku, H., van Damme, E.J.M., Peumans, W.J. and Goldstein, I.J. (1990) Arch. Biochem. Biophys. 279, 298-304. [I61 Wood, S.D., Reynolds, C.D., Lambert, S. and McMichael, R.A.D. (1994) Acta Crystallogr. D50, 110111. [I71 van Damme, E.J.M., Allen, A.K. and Peumans, W.J. (1987) Plant Physiol. 85, 566-569. [I 81 Saito, K., Komae, A,, Kakuta, M., van Damme, E.J.M., Peumans, W.J., Goldstein, I.J. and Misaki, A. (1993) Eur. J. Biochem. 217, 677-681. [I91 Kaku, H. and Goldstein, I.J. (1991) Carbohydr. Res. 213, 109-116. [20] Kaku, H. and Goldstein, I.J. (1992) Carbohydr. Res. 229, 337-346. [21] Oda, Y. and Minami, K. (1986) Eur. J. Biochem. 159, 239-245. [22] Peumans, W.J., De Ley, M., Stinissen, H.M. and Broekaert, W.F. (1985) Plant Physiol. 78, 379-383. [23] Sanadi, A.R. and Surolia, A. (1994) J. Biol. Chem. 269, 5072-5077. [24] Llovo, J., Lopez, A,, Fabregas, J. and Munoz, A. (1993) J. Infect. Dis. 167, 1477-1480. [25] Rogers, D.J., Loveless, R.W. and Balding, P. (1986) in: T.C. Bog-Hansel1 and E. Van Dnessche (Eds.), Lectins, Vol V Walter de Gruyter, Berlin, pp. 155-160. [26] Fabregas, J., Munoz, A., Llove, J. and Carracedo, A. (1988) J. Exp. Marine Biol. Ecol. 124, 21-30. [27] Xu, C., Moore, C.H., Fountain, D.W. and Yu, P.-L. (1992) Plant Sci. 81, 183-189. [28] Ray, S., Ahmed, H., Basu, S. and Chatterjee, B.P. (1993) Carbohydr. Res. 242, 247-263. [29] Ananthuram, V, Patanjali, S.R., Swamy, M.J., Sanadi, A.R., Goldstein, I.J. and Surolia, A. (1986) J. Biol. Chem. 261, 14621-14627. [30] Kochibe, N. and Matta, K.L. (1989) J. Biol. Chem. 264, 173-177. [31] Tsuchiya, N., Endo, T., Matsuta, K., Yoshinoya, S., Takeuchi, F., Nagano, Y., Shiotu, M., Furukawa, K., Kochibe, N., Ito, K. and Kobata, A. (1993) J. Immunol. 151, 1137-1146. [32] Peumans, W.J., DeLey, M. and Broekaert, W.F. (1984) FFBS Lett. 177, 99-103. [33] Huesing, J.E., Murdock, L.L. and Shade, R.E. (1991) Phytochemistry 30, 3565-3568. [34] Kellens, J.T.C., Jacobs, J.A., Peumans, W.J. and Stobberingh, E.E. (1993) J. Med. Microb. 39,440445. [35] Peumans, W.J., Nsimba-Labaki, M., Peeters, B. and Broekaert, W.F. (1985) Planta 164, 75-82. [36] Presant, C.A. and Komfield, S. (1972) J. Biol. Chem. 247, 6937-6945. [37] Rinderle, S.J., Goldstein, I.J., Matta, K.L. and Ratcliffe, R.M. (1989) J. Biol. Chem. 264, 16123-16131. [38] Rinderle, S.J., Goldstein, I.J. and Remsen, E.E. (1990) Biochemistry 29, 10555-10561. [39] Koeppe, S.J. and Rupnow, J.H. (1988) J. Food Sci. 53, 1412-1417. [40] Zenteno, E. and Ochoa, J.-L. (1988) Phytochemistry 27, 3 13-3 17. [41] Zenteno, E., Lascurain, R., Montano, L.F., Vazquez, L., Debray, H. and Montreuil, J. (1992) Glycoconj. J. 9, 204-208. [42] Singh, J., Kamboj, K.K., Kamboj, S.S., Sandnu, R.S. and Shangary, S. (1993) Plant Sci. 94, 47-53. [43] Pekelharing, J.M. and Animashaun, T. (1989) Biochem. SOC.Trans. 17, 131-132. [44] Aucouturier, P., Mihaesco, E., Mihaesco, C. and Preud’homme, J.-L. (1987) Mol. Immunol. 24, 503511. [45] Hortin, G.L. and Trimpe, B.L. (1990) Anal. Biochem. 188, 271-277. [46] Sastry, M.V.K., Banarjee, I?, Patanjal, S.R., Swamy, M.J., Swarnalatha, G.V. and Surolia, A. (1986) J. Biol. Chem. 261, 11726-1 1733. [47] Sastry, M.YK. and Surolia, A. (1986) Biosci. Rep. 6, 853-860. [48] Kretz, O., Creppy, E.E., Boulanger, Y. and Dirheimer, G. (1989) Arch. Toxic. Suppl. 13, 422427. [49] Kretz, O., Creppy, E.E. and Dirheimer, G. (1991) Toxicology 66, 213-224. [50] Kretz, O., Reinholt, J., Creppy, E.E. and Dirheimer, G. (1992) Tox. Lett. 65, 763-766.

[65] [66] [67] [68] [69]

Licastro, F., Morini, M.C., Kretz, O., Dirheimer, G., Creppy, E.E. and Stirpe, F. (1993) Int. J. Biochem. 25, 789-792. Kellens, J.T.C., Goldstein, I.J. and Peumans, W.J. ( 1 992) Mycol. Res. 96, 495-502. Peumans, W.J., Nsimba-Lubaki, M., Carlier, A.R. and Van Driessche, E. (1984) Planta 160, 222-228. Bloch, R., Jenkins, J., Roth, J. and Burger, M.M. (1976) J. Biol. Chem. 251, 5929-5935. Kellens, J.T.C. and Peumans, W.J. (1990) in: D.J.L. Freed (Ed.), Lectins, Vol. 7. Sigma, St. Louis, MO., pp. 57-62. Kitagaki, H., Seno, N., Yamaguchi, H. and Matsumoto, I. (1985) J. Biochem. (Tokyo) 97, 791-799. Shichijo, S., Shibata, H., Tsunosue, R., Shiotsuki, K., Hara, A,, Ito, K., Shiraishi, M. and Yokoyama, M.M. (1987) Leuk. Res. 11, 603-608. Peumans, W.J., Allen, A.K. and Cammue, B.P. (1986) Plant Physiol. 82, 1036-1039. Young, N.M., Watson, D.C. and Williams, D.E. (1984) Biochem. J. 222, 4 1 4 8 . Lonngren, J., Goldstein, I.J., and Bywater, R. (1976) FEBS Lett. 68, 31-35. Cammue, B.P., Peeters, B. and Peumans, W.J. (1985) Biochem. J. 227, 949-955. Kumar, M.A., Timm, D.E., Neet, K.E., Owen, W.G., Peumans, W.J. and Rao, A.G. (1993) J. Biol. Chem. 268, 251 76-25 183. Stirpe, F., Licastro, F., Morini, M.C., Parente, A., Savino, G., Abbondanza, A,, Bolognesi. A., Falasca, A.I. and Rossi, C.A. (1993) Biochim. Biophys. Acta 1158, 33-39. Nakajima, T., Kogure, T. and Furukawa, K. (1986) Exp. Clin. Immunogenet. 3, 187-194. Nakajima, T., Yazawa, S., Kogure, T. and Furukawa, K. (1988) Biochim. Biophys. Acta 964,207-212. Fukuda, N., Yoshimaru, A,, Hidaka, T., Ohta, H. and Yamamoto, K. (1994) Biosci. Biotech. Biochem. (Toyko) 58, 4 2 3 4 2 4 . Ferens-Sieczkowska, M. and Morawiecka, B. (1993) Acta Biochim. Pol. 40, 123-124. Mo, H., van Damme, E.J.M., Peumans, W.J. and Goldstein, I.J. (1994) J. Biol. Chem. 269, 7666-7673. Giollant, J., Giollant, M., Damez, M. and Dusser, M. (1991) J. Biochem. 109, 840-845. Giollant, M., Guillot, J., Damez, M., Dusser, M., Didier, P. and Didier, E. (1993) Plant Physiol. 101, 513-522. Bausch, J.N. and Poretz, R.D. (1 977) Biochemistry 16, 5790-5794. Bausch, J.N., Richey, J. and Poretz, R.D. (1981) Biochemistry 20, 2618-2620. Lis, H., Latter, H., Adar, R. and Sharon, N (1988) FEBS Lett. 233, 191-195. Lis, H. and Sharon, N. (1994) Trends Glycosci. Glycotechnol. 6, 65-74. Mo, H. and Goldstein, I.J. (1994) Glycoconjugate J., in press. Sharma, S. and Salahuddini, A. (1993) J. Agr. Food Chem. 41, 700-703. Conrad, F. and Rudiger, H. (1994) Phytochemistry 36, 277-283. Wantyghem, J., Goulut, C., Frenoy, J.P., Torpin, E. and Goussault, Y. (1986) Biochem. J. 237,483489. Fleischmann, G. and Rudiger, H. (1986) Biol. Chem. Hoppe-Seyler 367, 27-32. Wantyghem, J., Baron, M.H., Picquart, M. and Lavialle, F. (1990) Biochemistry 29, 6600-6609. Sabeur, G., Wantyghem, J. and Schuller, E. (1986) Biochimie 68, 581-585. Kaku, H., Peumans, W.J. and Goldstein, I.J. (1990) Arch. Biochem. Biophys. 277, 255-262. Sampietro, A.R. and Vattuone, M. (1994) Phytochemistry 35, 841-845. Animashaun, T. (1983) in: T.C. Bog-Hansen and G.A. Spengler (Eds.), Lectins: Biology, Biochemistry, Clinical Biochemistry 3. De Gruyter, Berlin, pp. 639-664. Togun, R.A., Animashaun, T., Kay, J.E. and Aboderin, A.A. (1994) Phytochemistry 35, 1125-1 130. Animashaun, T., Togun, R.A. and Hughes, R.C. (1994) Glycoconjugate J. 11, 299-303. Yamashita, K., Ohkura, T., Umetsu, K. and Suzuki, T. (1992) J. Biol. Chem. 267, 25414-25422. Yeung, H.W., Ng, T.B., Wong, D.M., Wong, C.M. and Li, W.W. (1986) Int. J. Peptide Protein Res. 27, 208-220. Yeung, H.W., Wong, D.M., Ng, T.B. and Li, W.W. (1986) Int. J. Peptide Protein Res. 27, 325-333. Kawagishi, H., Mori, H., Uno, A. and Kimura, A. (1994) FEBS Lett. 340, 56-58. Kawaguchi, T., Matsumoto, I. and Osawa, T. (1974) J. Biol. Chem. 249, 2786-2792. Knibbs, R.N., Goldstein, I.J., Ratcliffe, R.M. and Shibuya, N. (1991) J. Biol. Chem. 266, 83-88. Wang, W.-C. and Cummings, R.D. (1988) J. Biol. Chem. 263,45764585. Moustafapour, M.K. and Goldstein, I.J. (1993) Arch. Biochem. Biophys. 303, 255-259.

473 [94] Rogerieux, F., Belaise, M., Terzidis-Trabelsi, H., Greffard, A,, Pilatte, Y. and Lambre, C.R. (1993) Anal, Biochem. 21 1, 200-204. [95] Shibuya, N., Goldstein, I.J., Broekaert, W.F., Nsimba-Lubaki, M., Peeters, B. and Peumans, W.J. (1987) Arch. Biochem. Biophys. 254, 1-8. [96] Dall’Olio, I.F. and Trere, D. (1993) Eur. J. Histochem. 37, 257-265. [97] Shibuya, N., Goldstein, I.J., Broeckaert, W.F., Nsimba-Lubaki, M., Peeters, B. and Peumans, W.J. (1987) J. Biol. Chem. 262, 1596-1601. [98] Yamashita, K., Umetsu, K., Suzuki, T. and Ohkura, T. ( I 992) Biochemistry 3 1, 1 1647-1 1650. [99] Singh, J., Kamboj, S.S., Sandhu, R.S., Shangary, S. and Kamboj, K. (1993) Phytochemistry 33, 979983. [IOO] Hori, K., Miyazawa, K. and Ito, K. (1988) Bot. Mar. 29, 323-328. [I011 Kolberg, J., Michaelson, T.E. and Sletten, K. (1983) Z. Physiol. Chem. 364, 655-664. [I021 Adam, K.P. and Becker, H. (1993) Experientia 49, 1098-1 100. [I031 Nowakova, N. and Kocourek, J. (1974) Biochim. Biophys. Acta 359, 320-333. [I041 Ochoa, J.L. and Kristiansen, T. (1982) Biochim. Biophys. Acta 705, 396-404. [I051 Shi, W.-X., Shen, Z.-M., Sun, C. and Yang, J.T. (1993) 3. Prot. Chem. 12, 153-157. [lo61 Kamemura, K., Furuichi, Y., Umekawa, H. and Takahashi, T. (1993) Biochim. Biophys. Acta 1158, 181-188. [I071 Wang, K.-Y., Tao, Z.-J., Wu, K.-Z., Mo, H., Lee, S. and Sun, C. (1981) Acta Biochim. Biophys. Sin. 13, 4 2 3 4 2 6 . [I081 Shi, Y.-L., Xu, Y.-F. and Zhang, H. (1994) Sci. China B 37, 448453. [lo91 Zhang, H. and Shi, Y.-L. (1994) Sci. China B 37, 547-556. [IIO] Inbar, J. and Chet, 1. (1994) Microbiology 140, 651-657. [I 1 I] Oda, Y., Minami, K., Ichida, S. and Aonuma, S. (1987) Eur. J. Biochem. 165, 297-302. [I 121 Balzarini, J., Neyts, J., Schols, D., Hosoya, M., Van Damme, E., Peumans, W. and De Clercq, E. (1992) Antiviral Res. 18, 191-207. [ I 131 Taketa, K., Izumi, M. and Ichikawa, E. (1983) Ann. NY Acad. Sci. 417, 61-68. [ I 141 Taketa, K. and Izumi, M. (1984) Protides Biol. Fluids 31, 387-390. [ I 151 Schalla, W.O., Whittington, W.L., Rice, R.J. and Larsen, S.A. (1985) J. Clin. Micro. 22, 379-382. [ I 161 Cummings, R.D. and Kornfield, S. (1984) J. Biol. Chem. 259, 6253-6260. [ I 171 Ito, N., Kawahara, S., Hirano, Y., Morimura, Y., Nakajima, K., Uchida, K. and Hirata, T. (1994) Histochem. J. 26, 327-336. [ I 181 Nakayama, J., Okano, A,, Maeda, H., Miyachi, M., Ota, H., Katsuyama, T. and Kanai, M. (1990) Jap. J. Cancer Res. 8 1, 388-395. [ I 191 Slifkin, M. and Cumbie, R. (1987) J. Clin. Microbiol. 25, 1172-1 175. [I201 Jarlov, J.O., Rosdahl, V.T., Yeo, M. and Marples, R.R. (1993) J. Med. Microbiol. 39, 305-309. [121] Miranda-Santos, I.K.F. and Pereira, M.E.A. (1985) Am. J. Trop. Med. Hyg. 33, 839-844. [I221 Fernandez, J.G., Sanchez, A.J., Melcor, C., Chamorro, C.A., Garcia, C. and Paz, P. (1994) J. Anat. 184, I 37- 1 45. [I231 Imamura, H., Akimoto, Y., Chino, I . and Hirano, H . (1993) Acta Histochem. Cytochem. 26, 349-358. [I241 Galelli, A. and Truffa-Bachi, P. (1993) J. Immunol. 151, 1821-1831. [I251 Bird, G.N. and Winghall, J. (1971) Scand. J. Hematol. 8, 307-308. [I261 Moller, I? (1982) Virchows Arch. 396, 313-317. [I271 Vierbuchen, M. and Klein, P.J. (1983) Lab. Invest. 48, 181-186. [I281 Ree, H.J. and Hsu, S.-M. (1983) Cancer 51, 1631-1638. [I291 DeMaio, A,, Lis, H . , Gershoni, J. and Sharon, N. (1986) Cell Immunol. 99, 345-353. [I301 Shue, G.L., Kana, S., Kato, M., Oguchi, H., Kobayashi, T., Koiwai, T., Takoo, M., Furuta, S., Kanai, M. and Homma, T. (1993) Scand. J. Gastroenterol. 28, 599-604. [I311 Meyer, K.C., Powers, C., Rosenthal, N. and Auerbach, R. (1993) Am. Rev. Resp. Dis. 148, 1325-1334. [I321 Ota, T., Kohno, H., Maeda, M., Tanino, M. and Odashima, S. (1993) Oncology Res. 5, 235-243. [I331 Muramatsu, T., Muramatsu, H., Kasai, M., Habu, S. and Okumura, K. (1980) Biochem. Biophys. Res. Comm. 96, 1547-1553. [I341 Imai, Y. and Osawa, T. (1983) Scand. J. Immunol. 18, 217-224.

474 [135] Partin, K.M., Patneau, D.K., Winters, C.A., Mayer, M.L. and Buonanno, A. (1993) Neuron 1 1 , 10691082. [I361 Jacobs, L.R. and Huber, P.W. (1985) J. Clin. Invest. 75, 112-118. [I371 Watanabe, M., Muramatsu, T., Shirane, H. and Ugai, K. (1981) J. Histochem. Cytochem. 29, 779-790. [138] Farr, A.G., Anderson, S.K., Braddy, S.C. and Mejino Jr, J.L.V., (1988) J. Immunol. 140, 1014-1021. [139] Kasper, M., Haroske, G. and Mueller, M. (1994) Acta Histochem. 96, 63-73. [140] O’Reilly, R.J., Collins, N.H., Kernan, N., Brochstein, J., Dinsmoor, R., Kirkpatrick, D., Siena, S., Keever, C., Jordan, B., Shank, B., Wolf, L., Dupont, B. and Reisner, Y. (1985) Transplant. Proc. 17, 455-459. [141] Chen, Y.-F., Boland, C.R., Kraus, E.R. and Goldstein, I.J. (1994) Int. J. Cancer 57, 561-567. [142] Jones, J.M. and Feldman, J.D. (1973) J. Immunol. 111, 1765-1770. [143] Sasaki, K., Watanabe, E., Kawashima, K., Sekine, S., Dohi, T., Oshima, M., Hanai, N., Nishi, T. and Hasegawa, M. (1994) J. Cell. Biochem. Suppl. 18D, 272. [I411 Onodera, S. (1994) Biol. Pharm. Bull. 17, 29-33. [I451 King, M.J., Liew, Y.W., Moores, P.P. and Bird, G.W.G. (1988) Transfusion 28, 549-555. [I461 Brines, R. and Lehner, T. (1988) Immunology 63, 247-253. [I471 Kitamura, K., Kiyono, H., Fujihashi, K., Eldrige, J.H., Beagley, K.W. and McGhee, J.R. (1988) J. Immunol. 140, 1385-1392. [148] Schoenbeck, S., Hanner, M.J. and Kognoff, M.F. (1989) J. Exp. Med. 169, 1491-1496. [I491 Schumacher, U., Kretzscgmar, H., Pfueller, U. (1994) Acta Neuropathol. 87, 422424. [150] Susuki, H., Franz, H., Yamamoto, T., Inasaki, Y. and Konno, H. (1988) Neuropath. Appl. Neurobiol. 14, 221-227. [I511 Ohlon, C. and Karlsson, J.O. (1983) Brain Res. 264, 99-104. [I521 Leu, R.W., Whitley, S.B., Herriot, J. and Huddleston, D.J. (1980) Cell. Immunol. 52, 414428. [153] Neogrady, S., Galfi, P., Veresegyhazy, T., Bardocz, S. and Pusztai, A. (1994) Histochem. J. 26, 197-206. [154] Wouterlood, F.G. and Groenewegen, H.J. (1991) Prog. Histochem. Cytochem. 22, 1-78. [I551 Ngo, T.D., Davies, D.C., Egedi, G.Y. and Tombol, T. (1994) J. Anat. 184, 129-136. [I561 Schultz, D.R. and Arnold, P.I. (1984) Acta Pathol. Microbiol. Immun. Scand. C Suppl. 284, 59-66.

J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins 11 0 1997 Elsevier Science B.V. All rights reserved

CHAPTER 13

Microbial lectins and their glycoprotein receptors Nathan Sharon and Halina Lis Department of Membrane Research and Biophysics, The Weizmann Institute of Science, Rehovot 76100, Israel

1. Introduction Numerous microorganisms appear to have selected carbohydrates as the preferred attachment sites on their host cells and tissues [l]. This is not surprising, in view of the abundance of carbohydrates on the surfaces of eukaryotic cells, in the form of glycoproteins and glycolipids. Binding to the carbohydrates is mediated by microbial surface lectins, members of a large group of microbial adhesins, that play a major role in the initiation of infection and in non-opsonic phagocytosis [2-51. Host range, tissue tropism and target cell specificity demonstrated by a particular microbe are determined, at least in part, by a stereochemical fit between microbial lectins and complementary carbohydrate receptors on host cell surfaces [6]. Adhesion to insoluble carbohydrates may also play an important role in the biofouling of marine surfaces and in ecological phenomena such as biodegradation, as well as in the monitoring by bacteria of the nutrient status of the environment [7]. Although a considerable number of these lectins are well defined, the existence of many more has been inferred from experiments on the effect of carbohydrates on the interaction of intact microorganisms with different target cells. The oldest, and perhaps most thoroughly studied system of this type is the interaction between the hemagglutinin of influenza virus A and N-acetylneuraminic acid' on cell surfaces. It is responsible not only for the adhesion of the virus to the cells, but also for fusion of the viral membrane with the host cell membrane and is also the viral component to which protective antibody is directed. Virtually all bacterial species and genera express lectins or lectin-like activities, frequently of more than one type and with different specificities. However, it is usually not known whether individual cells coexpress multiple types of lectin or each lectin is confined to a distinct cell population. Many Gram negative bacteria (for example, Escherichia coli and Salmonellae spp.) and a few Gram positive ones (e.g. certain actinomyces), produce surface lectins that are often in the form of submicroscopic hairlike appendages known as fimbriae (pili) that protrude from the surface of the cells. The best characterized bacterial surface lectins with respect to their molecular properties, carbohydrate specificity and genetics are the type 1 fimbriae specific for mannose and the type P fimbriae specific for galabiose, [Gal(al+Gal], produced by many strains of E. coli. Other examples are S fimbriae of E. coli, specific for NeuAc(a2-3)Gal, and type 2

'

All sugars are of the D-configuration, except for fucose which is L475

476

fimbriae of oral actinomyces, specific for Gal(fi1-3)GalNAc or GalNAc fi- that are not so well characterized. Non-fimbrial lectins that are components of the cell surface have also been described. The case of Pseudomonas aeruginosa is unusual in that certain strains of the organism produce intracellular lectins, at least one of which appears also on the bacterial surface. Among the fungal and protozoal lectins only a few have been studied in detail. One of these is the galactose-specific lectin of the protozoa Entamoeba histolytica. It mediates adhesion of the parasite to human colonic mucin glycoproteins and has a central role in the contact-dependent cytolysis or histolysis for which the parasite is named. A sialicacid-specific lectin has been isolated from merozoites of the human malarial parasite, Plasmodium falciparum. An unusual lectin is that of the protozoan Giardia lamblia, specific for mannose-6-phosphate, which is activated by trypsinization. The carbohydrates to which the microbial lectins bind are in the form of glycoproteins and glycolipids (Table 1). Many lectins interact with both classes of glycoconjugate; this is not unexpected, since these compounds frequently contain identical oligosaccharides. Identification of the receptors for microbial lectins on animal cells (or in secretions) is based on methods such as binding of intact bacteria or of the isolated lectin either to blots of electrophoretically. separated cell membrane glycoproteins [8] or to thin layer chromatograms of glycolipids extracted from the cells. Glycoproteins can also be analyzed by affinity chromatography of cell membrane extracts on immobilized microbial lectins. The finding that a glycoprotein or glycolipid interacts with the bacteria (or isolated lectin) in such tests does not prove, however, that it serves as a functional receptor in uiuo. Most importantly, it should be obtained from cells susceptible to infection by the organism studied. Also, antibodies to the presumed receptor should inhibit binding of the organism to the cells and cells devoid of the receptor (as seen sometimes in mutants) should lack the ability to bind the microorganism. Glycolipids that serve as native receptors for microbial lectins have been identified in a large number of cases, but information on glycoproteins that play such a role is still scarce. In the following we shall discuss mainly those lectins of viruses, bacteria, protozoa and fungi that are known to interact with either glycoproteins alone or with glycoproteins as well as glycolipids. We shall also deal with their role in infection and lectinophagocytosis. For bacterial lectins that appear to bind exclusively to glycolipids, see recent reviews on the subject [9-1 I].

2. Viruses 2. I . Sialic-acid-specijic 2.1.1. Influenza uirus The ability of the virus to agglutinate erythrocytes has been first reported in 1941. It took more than a decade before it was shown that influenza virus binds to erythrocytes and other cells via N-acetylneuraminic acid residues present on the cell surface and that this binding is a prerequisite for initiation of infection [24,25]. Other viruses, such as Sendai, Newcastle disease, polyoma and rotavirus also exhibit an affinity for sialic

477 Table 1 Carbohydrates as attachment sites for infectious agents a Organism

Target tissue

Reference( s)

Carbohydrate Structure

Form’

Respiratory tract

NeuAc(a2-6)Gal

GP

[I21

Respiratory tract

NeuAc(a2-3)Gal

GP

1121

type c Parvovirus B 19

Respiratory tract

9-0-AcNeuAc(a2-3)Gal

GP

[I21

Erythroid cells

GalNAc(@1-3)Gal(a 1-4)Gal(@1 4 )

GSL

[I31

Polyoma virus

Epithelial cells

NeuAc(a2-3)Gal

GP

[I41

Urinary tract

Man(a 1-3)[Man(a 1-3)[Man(al4)]

Urinary tract

Gal(a 1-4)Gal

GP GSL

[4,81 [9]

GP GSL

PI

GP GP

[81 [I61

Viruses Influenza type A type B

Bacteria

E. coli type 1 type p type

s

Neural

NeuAc(a2-3)Gal(fi 1 -3)GalNAc NeuGc(a2-3)Gal

type CFAI/II

Intestine

NeuAc(a2-3)Gal

type KI

Endothelial cells

GlcNAc(fi14)GlcNAc

type K88ac

Intestine

Gal(fi1-3)GalNAc

type K99

Intestine

NeuGc(a2-3)Gal(fi14)Glc

Oral

Gal(@14)Glc

Actinomyces naeslundi

GalNAc(B 1-3)Galfi

Neisseria gonorrhoea

Genital

Gal@14)Glcfi NeuAc(a2-3)Gal(fi 1-4)GlcNAc

Streptococcus sanguis

Oral

NeuAc(a2-3)Gal(fi1-3)GalNAc

GP GSL GSL GP

[I51

[17] [I81 [I91 [20]

GSL GP GP

[9]

GSL

[22]

GP

[23]

[21]

Fungi

Candida albicans

Skin and mucosa Gal(@14)Glc Fuc(a I-2)Gal

a

For protozoa, see Table 3.

’ Predominant; GP, glycoproteins; GSL, glycolipids. acid [ 121. The hemagglutinin (lectin) of the influenza virus, responsible for its attachment to cells, was purified, crystallized, and studied in detail, culminating in the elucidation of the three-dimensional structure of its complex with N-acetylneuraminic-acid-containing oligosaccharides at the atomic level [26-281. The subunit of the lectin is composed of two polypeptides, HA1 and HA2 (with molecular masses of 36 and 26kDa, respectively), covalently linked by a single disulfide bond, and it associates non-covalently to form trimers that are located on the surface of the viral membrane [27]. It is a glycoprotein, with six N-linked oligosaccharide chains attached to HA1 and one to HA2. Except for one oligomannose unit, all are complex, bi- or triantennary structures, with three of them containing sulfated galactose [29]. The carbohydrate binding site is located in a pocket of the HA1 polypeptide chain, in a domain of the lectin protruding from the membrane,

478

and is composed of amino acids that are largely conserved in the numerous strains of the virus [27]. Other conserved residues are found behind the pocket and seem to stabilize the architecture of the site without interacting with the carbohydrate. Over 100 strains of influenza virus, mostly of the A- and B-types, were examined for their ability to bind to enzymatically modified erythrocytes carrying terminal N-acetylneuraminic acid attached to galactose either by an a2-3 or a2-6 linkage [ 12,261. Differences in their specificity with respect to this linkage were correlated with the species origin of the virus. Thus, human isolates preferentially agglutinated resialylated erythrocytes containing the NeuAc(a24)Gal sequence, while the avian and equine isolates exhibited preference for NeuAc(a2-3)Gal. Strains of influenza C virus (as well as coronaviruses) do not bind N-acetylneuraminic acid at all, but only recognize its derivative, 9-0-acetyl-N-acetylneuraminic acid; the 9-0-acetyl group is critical for mediating cellular attachment [30]. Comparison of the primary sequences of hemagglutinins of the human virus with those of mutants showing decreased affinity for NeuAc(a2-6)Gal and markedly increased affinity for NeuAc(a2-3)Gal, revealed that they differ in a single amino acid substitution, Leu226 in the parental strains being replaced by glutamine in the mutants. Similar studies with avian isolates and their variants showing the reverse change in specificity (from a23 to a 2 4 linked N-acetylneuraminic acid), again revealed a substitution only at position 226 - from glutamine to leucine. This illustrates that a single amino acid substitution can alter the sugar specificity of a lectin. Although residue 226 is located in the carbohydrate binding site of the hemagglutinin, it is not in direct contact with the bound sugar, as shown by crystallographic studies of wild-type influenza virus hemagglutinin complexed with NeuAc(a2-6)Gal(fi14)Glc [sialyl(a24)lactose] and of a mutant hemagglutinin complexed with NeuAc(a2-3)Gal(fil-4)Glc [sialyl(a2-3)lactose]. The suggestion has therefore been made that the change in specificity is due to conformational differences between the mutant and wild-type proteins. X-ray crystallographic analysis of the complex of the hemagglutinin with bound sialyl(a2-6)lactose, [NeuAc(a24)Gal(fi 1-4)Glc], placed the sialic acid in the binding pocket with one side of the pyranose ring in tight contact to the protein and the other side facing the solvent [28] (Fig. 1). It also permitted to predict potential hydrogen bonds and van der Waals contacts of sialic acid atoms with amino acids in the binding site based on their proximity to each other. The validity of these predictions was tested by evaluating the affinity for the hemagglutinin of a series of synthetic analogs of N-acetylneuraminic acid with modified functional groups, by determining: (i) the ability of the analogs to inhibit viral attachment to cells [3 11 and (ii) the equilibrium dissociation constants for the binding of these analogs to the hemagglutinin [32]. It was thus confirmed that the two carboxylate oxygens of N-acetylneuraminic acid receive hydrogen bonds from the hydroxyl group of Szr-136 and from the main chain amide group at Asn-137, respectively, and that both bonds are necessary for binding. These studies also provided additional evidence on the critical importance of the hydrophobic contact between the acetamido group at C-5 of N-acetylneuraminic acid with the indole ring of Trp- 153. On the other hand, the hydroxyl at C-9 does not seem to participate in ligand binding, contrary to what has been proposed from the crystal structure. The structures of complexes of the hemagglutinin with four sialic acid analogs, with

479

\7%

Thr 155

I, Trp 153

n

Fig. 1. Model for the position of sialic acid in the binding pocket of influenza virus hemagglutinin. This model for the best fit has been deduced from the difference electron density maps of X-ray crystallographic studies. Some of the hydrogen bonds proposed in this model are shown by dashed lines. (Taken from ref. [31]; with permission from the authors.)

affinities 10- to 100-fold higher than that of N-acetylneuraminic acid, were determined by high-resolution X-ray crystallography [33]. In these analogs, the sialic acid core was substituted at the 4 or 6 positions with spaced naphthyl or dansyl groups. In each of the complexes, the sialic acid moiety was equivalently positioned in the binding site of the hemagglutinin, while the substituent groups that differentiate the high-affinity analogs frclm each other interacted with hydrophobic patches and polar residues adjacent to the binding site. In addition to the binding site discussed above, the hemagglutinin possesses a secondary site, at the interface between HA1 and HA2 polypeptide chains [34]. However, crystallographic studies have shown that sialyl(a24)lactose does not bind to the secondary site at all, while sialyl(a2-3)lactose binds to this site with at least 4 times lower affinity than to the first one. Another virus specific for sialic acid is murine polyoma virus. Two types of strains are known that differ in their specificity for sialic acid oligosaccharides: those that form large plaques bind to oligosaccharides terminating in NeuAc(a2-3)Gal, whereas the small plaque strains also tolerate branched structures with a second, a2-6 linked, sialic acid, e.g. NeuAc(a2-3)Galp3 ~ e u A c ( a 2 4 ) ] G a l N A c These . strains also differ in their ability to form tumors in mice - the small plaque strains produce few, if any, tumors, while

480

the large plaque strains are highly tumorigenic. The critical difference in the structure of the viral protein of these strains is in residue 91, which is glycine in the small plaque strains and glutamic acid in the large plaque strains. Crystallographic studies at 3.65 A resolution of the viral protein from small plaque strains in complex with sialyl-3-lactose have shown that the combining site is in the form of a shallow groove and that both the sialic acid and the galactose form contacts with the protein [14]. This is in contrast to influenza virus hemagglutinin, where only sialic acid makes such contacts. 2.2. Other specijicities Information on the existence of viral lectins with specificities other than for sialic acid is scant. Thus, parvovirus B 19 binds GalNAc(P 1-3)Gal(a14)Gal(fi1-4) [ 131, Simplex virus binds heparan sulfate [35] and HIV recombinant envelope glycoproteins interact specifically with certain N-linked carbohydrate units of glycoproteins, e.g. of the oligomannose type, with mannose-6-phosphate and sulfated polysaccharides, such as heparin and dextran sulfate [36,37].

3. Bacteria The largest number of microbial lectins characterized to date are from bacterial sources (Table 1).

3.1. Mannose-specijic (type I jimbriae) 3.1.1. Enterobacteria Type 1 fimbriae are expressed by most strains of E. coli, as well as by other enterobacteriaceae, such as Klebsiella pneumoniae [38,39] (Fig. 2). They are heteropolymeric organelles, about 7 nm in diameter and 100 to 200 nm in length, consisting of helically arranged subunits (pilins) of several different types, assembled in a well defined order[5,40]. The bulk of the fimbrial filament (shaft) is made up of polymers of a major subunit, FimA, mw 17 kDa. In addition, fimbriae contain a cassette of three minor ancillary subunits, FimF, FimG and FimH. The latter is the only subunit that possesses a

Fig 2. Type I fimbriated Eschenchia coh; magnified 24 OOOx (courtesy Dr. A. Gbarah)

48 1

carbohydrate binding site and is thus responsible for the sugar specificity of the finibriae. Although FimH is present both at the distal tip and at intervals along the length of the filament, only the subunit at the tip appears to be able to mediate mannose-sensitive adhesive interactions; the subunits at the other positions are inaccessible to the ligand. In other types of fimbriae (e.g. type P) the carbohydrate binding subunit is exclusively located to the tip. Isolated FimH binds mannose-containing glycoproteins and adheres in a mannosespecific manner to human neutrophils [41]. Moreover, it triggers an oxidative burst in a manner that mimics the activity of type 1 fimbriae. In addition, inert microspheres coated with FimH, but not with bovine serum albumin, are phagocytosed by neutrophils. Mutants lacking t h e j m H gene, but not genes encoding other fimbrial subunits, fail to bind to eukaryotic cells. Two proteins, FimC and FimB, are involved in the biogenesis of type 1 fimbriae, without being part of the final structure. The former acts as a periplasmic chaperon that stabilizes fimbria subunits in the periplasm through the formation of distinct complexes. The subunit-chaperon complexes are targeted to FimB, an outer membrane protein, which organizes their ordered secretion into an extracellular polymer [42]. Chaperons are also required for the assembly of other types of fimbriae, as well as of non-fimbrial lectins [43451. The expression of fimbriae is phase variable, i.e. bacteria shift periodically between a fimbriated and non-fimbriated state [38]. As a result, a given bacterial population will always contain cells of both phenotypes. The on- and off-phase variation is controlled at the transcriptional level and involves the inversion of a 314-base pair DNA segment harboring the promoter of t h e j m A gene [46]. Remarkable differences are found in the size and antigenic properties of the structural subunit of type 1 fimbriae (FimA) among different species of enterobacteria[47]. In contrast, a high degree of conservation is found between FimH proteins from type 1 fimbriae expressed by various species of the Enterobacteriaceae family, although they differ in their fine sugar specificity [48]. Thus, the combining sites of type 1 fimbriae of E. coli and K. pneumoniae correspond to the size of a trisaccharide and are in the form of a depression or pocket on the surface of the lectin [49]. In the case of the E. coli lectin, there are probably three adjacent subsites, each of which accommodates a monosaccharide residue. In the proximity of the combining site there is a hydrophobic binding region, as indicated by the finding that aromatic a-mannosides are significantly more powerful inhibitors than methyl a-mannoside. In contrast, several Salmonella species examined bind aromatic a-mannosides, as well as the trisaccharide Man(a I-3)Man(fi I-4)GlcNAc, weaker than methyl a-mannoside, indicating that the combining site of the Salmonella lectin is probably smaller than that of E. coli and K. pneumoniae, and is devoid of a hydrophobic region. Although very similar in their carbohydrate specificity, the lectins from E. coli and K. pneumoniae differ in their affinity for aromatic mannosides. Thus, 4-methylumbelliferyl a-mannoside is about 10-fold more effective than p-nitrophenyl a-mannoside in inhibiting yeast aggregation by type 1 fimbriated E. coli, while it is only 4 times more effective than p-nitrophenyl a-mannoside as inhibitor of K. pneumoniae [50]. Since the FimH subunits of the two organisms exhibit 88% homology, the possibility has been

482 Table 2 Glycoprotein receptors for microbial lectins Organism

Source of glycoprotein

Designation

Human granulocytes, peritoneal macrophages

CD1 l/CD18 integrin

Human granulocytes

NCA-50

Colonic mucosa Urine

IgA Tamm-Horsfall glycoprotein

Sialic acid specific E. coli CFA/I I? fakiparum

Human epithelial cells

26 kDa gp

Human erythrocytes

Glycophorin

P aeruginosa, PAK

Human oral epithelium

82 kDa, 40-50 kDa gp’s

Mannose specific E. coli type 1

Reference( s)

Mouse corneal cells S. sanguis

Human saliva

400 kDa gp

Human oral epithelium

160 kDa gp

Human saliva

180 kDa gp

GaVGalNAc specific

A . naeslundii type 2

B. pertussis toxin

CHO cells

165 kDa gp

Human T cells, Jurkatt

43 kDa gp; 70 kda gp

considered that the differences in specificity are due to differences in the presentation of the combining sites of these subunits in the fimbriae. To test this hypothesis, two types of hybrid fimbriae were genetically generated: in one of these the E. coli FimH was presented on a filament of K. pneumoniae structural subunits (EcFimH-KpFimA); in the other K. pneumoniae FimH was presented on a shaft of E. coli FimA (KpFimH-EcFimA) [50]. It was found that the specificity of the EcFimH-KpFimA hybrid with respect to aromatic a-mannosides was similar to that of native K. pneumoniae lectin, whereas that of KpFimH-EcFimA was like that of the E. coli lectin. These results indicate that the shaft on which FimH is presented plays a role in modulating the specificity of type 1 fimbriae lectins, probably by imposing conformational constraints on the carbohydrate binding subunit. The interspecies heterogeneity of the FimA subunit of enterobacteria thus ensures significant diversity in their sugar specificity and as a result, in the function of their lectins, as reflected by their ability to mediate adhesion to a particular type of animal cell. The notion that the fimbrial filament can influence the specificity of the carbohydrate binding moiety is novel and contrasts with the P fimbriae system, in which PapG appears to be the sole determinant of binding specificity. Recently a form of type 1 fimbriae has been described which, in addition to binding carbohydrates, interacts also with non-glycosylated regions of proteins in a mannoseinhibitable manner [51]. It is not clear whether this interaction occurs via the carbohydrate binding site proper and how it is inhibited by mannose. The difference between the two

483

functional forms of the fimbriae may be due to subtle variations in FimH or to quantitative or qualitative differences in the assembly of one or more of the subunits. In this context it should be noted that concanavalin A, a mannose/glucose-specific plant lectin, has also been shown to bind peptides in a carbohydrate-inhibitable manner [52,53]. Type 1 fimbriated E. coli or the isolated fimbriae bind to glycoproteins from diverse sources [8]. These include a 65 kDa glycoprotein from guinea pig erythrocytes [54], the carcinoembryonic antigen, normally localized at the apical border of epithelial cells of the large intestine, secretory IgA and IgA myeloma proteins, especially those of the IgA2 subclass [55] Tamm-Horsfall glycoprotein (often referred to as uromucoid) [56], as well as several constituents of mucous layers [57] (see also section 6.1.2). The fimbriae bind to three glycoproteins derived from the cell membrane of human granulocytes (or neutrophils)[58] (Table 2). Two of them have been identified as components of the integrin superfamily CD1 l/CD18 (also known as leukocyte adhesion molecules). The fimbriated bacteria bound in a mannose-specific, dose-dependent and saturable manner to the isolated integrin in wells of microtiter plates and on SDS-PAGE gels; this binding was inhibited by monoclonal antibodies to integrin 1591. Monoclonal antibodies to CDI 1/CD18, but not to other granulocyte surface antigens, inhibited binding of the bacteria to the granulocytes. The same molecules serve also as receptors for type 1 fimbriae on human peritoneal macrophages [60]. In addition, a glycoprotein called NCA-50, which belongs to a family of non-specific cross-reacting antigens associated with the granulocyte membrane, was reported to specifically bind type 1 fimbriae [61]. 3.2. Sialic-acid-specijic

3.2.I . Escherichia coli Certain strains of E. coli isolated from humans and farm animals express fimbrial lectins specific for glycoconjugates containing sialic acids [3-51. This conclusion is based primarily on the observation that hemagglutination caused by these organisms is decreased or completely abolished by treatment of the erythrocytes with sialidase. Examples of such lectins are type S fimbriae of bacterial strains causing sepsis and meningitis in newborn infants, CFA (Colonization Factor Antigen) I and I1 of human enterotoxigenic E. coli isolates, as well as E. coli K99 fimbriae of enterotoxigenic strains isolated from piglets, calves and lambs suffering from diarrhea. The structure of S fimbriae is very similar to that of type 1 fimbriae. They are composed of a major subunit and several minor components, of which only one binds sialic acid. In contrast, both in the CFAs and K99 fimbriae the major subunit also contains the carbohydrate binding site [72,73]. Comparison of the amino acid sequence of the carbohydrate binding subunit of type S fimbriae with those of the major CFA 1 and K99 subunits revealed the presence of a common motif, rich in basic amino acids (Fig. 3). Site-specific mutagenesis experiments showed that a lysine and an arginine residue in this region play a part in ligand binding [74]. S-fimbriated E. coli combine with a2-3 linked sialic acid residues on integral membrane glycoproteins. Limited trypsinization of human erythrocytes completely abolished binding of such bacteria, indicating that glycophorin A is their sole receptor on the erythrocytes[75]. S-fimbriated E. coli also bind to sialic acid on gangliosides,

484

Organism

Amino acid residues

Reference

Escherichia coli

SfaS

-

K99

LYS”~ -

-

LysSh

-

CFAiI Helicobacter pyIori

Lys”‘

-

Lys134 -

Ala

~~

-

Arg

Ala

&I

Ser

-

Lys

[74]

LYS -

Asp

-

Asp

Arg

[74]

Val

Lys

[74]

&

Glu

Lys -

[89]

Lys -

Val

Arg -

Thr

Fig. 3. Sequence homology in the binding region of sialic-acid-specific bacterial lectins. Identical or functionally identical residues are underscored. Gaps have been introduced for optimal alignment.

preferentially NeuGc(a2-3)Gal and NeuAc(a2-8)NeuAc [ 151, as well as to sulfated glycolipids [76]. The latter binding apparently occurs through a different fimbrial subunit than that which interacts with sialic acid. The specificity of CFA I is not well defined, beyond the fact that it recognizes sialic acid [77,78]. A sialoglycoprotein with an apparent molecular weight of 26,000 is the only glycoprotein from the human erythrocyte membrane that binds CFA I fimbriated E. coli [621. K99-bearing E. coli bind to sialylated mucus glycoproteins [79]. The binding is not inhibited by sialic acid nor by other simple sugars, but by glycopeptides isolated from glycoproteins of bovine plasma, suggesting that the lectin recognizes complex carbohydrate structures. Glycopeptides bearing the terminal NeuGc(a2-3)Gal sequence strongly inhibited hemagglutination caused by E. coli K99, demonstrating the specificity of these bacteria for N-glycoloylneuraminic acid [SO]. The physiological receptor for K99 on intestinal epithelial cells of pig [Sl] and horse [82] appears, however, to be the glycolipid N-glycoloylneuraminyllactosyl ceramide [NeuGc(a2-3)Gal(P14)Glc~lCer]. 3.2.2. Streptococcus suis The Gram positive bacterium S. suis, a common cause of sepsis, meningitis and other serious infxtions in young piglets and also of meningitis in humans, agglutinates human erythrocytes, but not after treatment with sialidase. Resialylation of the desialylated erythrocytes with a2-3 sialyltransferase resulted in strong agglutination of the cells by the bacteria, whereas resialylation with sialyltransferases having different specificities gave cells that were poorly agglutinated [83]. Blotting experiments revealed binding to band 3, band 4.5 and glycophorin (as well as to polyglycosyl ceramides) of human erythrocyte membranes. The involvement of glycophorin as a ligand for the bacteria on intact cells is however excluded by the finding that trypsinization of the cells does not affect their agglutination by the bacteria and by the agglutinability of En(a-) erythrocytes which are defective in glycophorin A. The ligands for S. suis are thus the sialylated polyN-acetyllactosamine glycans carried by band 3 and band 4.5.

485

3.2.3. Streptococcus sanguis S. sanguis, an oral microorganism, adheres to saliva-coated tooth surfaces by binding

to salivary glycoproteins. The binding to these glycoproteins on SDS-PAGE blots was abolished by their treatment with sialidase, as well as with hydrofluoric acid, but was not affected by peptide-N-glycanase F (PNGaseF), indicating that S. sanguis binds to sialic acid on 0-linked chains of the glycoprotein(s) [84]. A 23 kDa membrane glycoprotein from human buccal epithelial cells bearing 0-linked NeuAc(a2-3)Gal(@ 1-3)GalNAc chains was implicated as receptor for S. sanguis OMZ9 on these cells [21]. The cDNA of the sialic-acid-specific lectin of S. sanguis codes for a polypeptide of 1435 residues (calculated mw of 158.4kDa) with three unique domains, two of which consist of repetitive amino acid sequences [66]. The third, which resides near the carboxy terminus, contains 48% proline. The lectin bound to a single salivary glycoprotein of mw 400 kDa [66]. Binding was inhibited by sialic acid and was abolished by desialylation of the glycoprotein; the best inhibitor was N-acetylneuraminyllactose.

3.2.4. Helicobacter pylori H. pylori is a pathogen which colonizes the mucus layer of human gastric tissues and is associated with gastritis and peptic ulcers (and possibly also with gastric carcinoma). It exhibits several specificities, one of which is for sialic acid. H. pylori agglutinates erythrocytes [85] and binds to mouse adrenal gland cells [86] in a sialic-acid dependent fashion. Furthermore, specific binding of H. pylori to acid glycosphingolipids extracted from human gastric mucosa, such as the ganglioside GM3, has been reported[87]. In addition, H. pylori has an affinity for fucose (see section 3.4.1) and interacts with glycolipids such as cerebroside and sulfated lactosylceramide [88], which lack both sialic acid and fucose. The inhibitory activity of lactosylceramide sulfate and GM3 ganglioside on hemagglutination induced by H. pylori was additive, consistent with the possibility that two distinct lectins are involved in the binding to sialic acid and to sulfated glycolipids, respectively. The sialic-acid-specific lectin is a fibrillar surface protein with a mw of 20 kDa. Its cDNA has been cloned, sequenced and expressed in E. coli [89]. A sequence of the lectin (residues 134-139) was found to be homologous to a region that forms part of the carbohydrate binding sites of the sialic-acid-specific lectins of S-fimbriae, K99 and CFA I (cf. Fig. 3). An antibody against a synthetic peptide containing the above sequence blocked hemagglutination of human erythrocytes by H. pylori, suggesting that in this lectin, too, it is part of the carbohydrate binding site. 3.2.5. Mycoplasma pneumoniae M. pneumoniae, a well-established pathogen of the human respiratory tract, is another organism specific for sialic acid. It adheres to animal cells primarily through a lectin, known as P1 protein (mw 170kDa), which is densely clustered at the tip of the organism [90]. The lectin is specific for N-acetylneuraminic acid linked a2-3 to terminal galactose residues of the poly-N-acetyllactosamine sequence of blood type I/i antigen, as shown by binding experiments with sialidase-treated human erythrocytes that have been resialylated by specific sialyltransferases [9 11. The preference for sialic acid a2-3 linked, rather than a2-6 linked, was confirmed by the finding that the oligosaccharides and glycoproteins containing the former linkage were more inhibitory than those containing

486

the latter one. The most potent inhibitors were glycopeptides derived from bands 3 and 4.5 of human erythrocytes, and the bovine erythrocyte glycoprotein GP-2, all rich in poly-N-acetyllactosamine chains. Further evidence for the importance of such sequences in binding the M. pneumoniae lectin was provided by experiments with human blood type i erythrocytes, whose linear poly-N-acetyllactosamine chains are susceptible to digestion with endo 6-galactosidase. Following treatment with the enzyme, the binding of M. pneumoniae to the erythrocytes decreased by 85% [92]. 3.3. Gal and GulNAc-specijk

3.3.1. Escherichia coli CS3, a subcomponent of CFA I1 of enterotoxigenic E. coli binds specifically to GalNAc(6 1-4)Gal[93]. This was demonstrated by inhibition studies, using well-characterized antibodies and glycoconjugates of defined structures. Support for these findings was provided by electron microscopic experiments showing that the disaccharide, 0-linked to bovine serum albumin via a spacer, localized around bacteria expressing CS3 but not around CS3-negative mutants. Enterotoxigenic E. coli producing K88 fimbriae occur in three serological variants: ab, ac or ad, that differ in their fine carbohydrate specificity. All are galactose-specific, but whereas K88ab fimbriae recognize the sequence Gal(al-3)Gal[94], K88ac fimbriae appear to bind preferentially to Gal@-3)GalNAc and Fuc(al-2)Gal(~1-3/4)GlcNAc[ 171. These carbohydrate structures were shown to be present in two porcine brush border glycoproteins of 210 and 240kDa that bind K88ac fimbriae, but not K88ab and K88ad fimbriae. They were not detected in glycoproteins from brush borders of piglets that do not bind K88 fimbriated E. coli. 3.3.2. Pseudomonas aeruginosu I! aeruginosa, an opportunistic pathogen, capable of causing infections of eye, lung, skin and other parts of the body, produces two well-Characterized intracellular lectins; one is specific for galactose (PA-I), the other for fucose (PA-11, see section 3.4.2) [95]. PA-I exhibits a preference for a-galactosides, especially those with a hydrophobic aglycone[96]. The cDNA of PA-I was isolated and shown to encode a chain of 121 amino acids (mw 12.7kDa) with a predominant central hydrophilic core between two hydrophobic domains [97]. PA-I agglutinates papain-treated human erythrocytes independent of blood type. However, it exhibits preferential affinity for the branched oligosaccharides bearing both A and B blood group determinants that are present in the saliva of AB secretors [98]. Although PA-I is located mainly intracellularly, evidence has been obtained that it is also exposed on the cell surface [99]. This could explain the finding that injection of the purified lectin into mice protected the animal against lethal doses of the live bacteria [951. I? aeruginosa appears to bind largely to glycolipids [l], but binding to human respiratory mucins has also been reported. The binding is sensitive to periodate oxidation of the mucins, suggesting the involvement of the carbohydrate chains of the mucins and a (putative) surface lectin(s) on the bacteria [ 1001. Inhibition studies have shown that the

487

organism recognizes Gal(@1-3)GlcNAc and Gal(fl14)GlcNAc, but has no affinity for sulfated glycopeptides [loll. This lectin(s) has, however, not yet been isolated and its relation to PA-I is not known. 3.3.3. Actinomyces species Actinomyces naeslundii and Actinomyces viscosus are prominent oral bacteria that colonize tooth and mucosal surfaces by binding to epithelial cells or other bacteria, such as S. sanguis. These interactions are mediated by-galactose/@-N-acetylgalactosaminespecific lectins, associated with type 2 fimbriae [ 19,20,102]. Although isolated type 2 fimbriae alone do not agglutinate either S. sanguis or sialidase-treated erythrocytes, lactosespecific agglutination occurred when the cells were incubated with multivalent complexes, formed by crosslinking the fimbriae with small amounts of specific antibody. The actinomyces lectin has not yet been purified. A fimbrial subunit gene,jmA, has been cloned, but the protein it encodes is apparently not involved in the interaction with S. sanguis. The carbohydrate specificities of the lectins of A . viscosus T14V and A. naeslundii WVU45 were defined using galactose-containing oligosaccharides as inhibitors of coaggregation with S. sanguis 34. The most effective disaccharide inhibitor was Gal(fi1-3)Gal, which was more than 10 times as active as lactose and also more active than any galactose disaccharide tested. Receptors for the actinomyces lectins have been isolated and extensively characterized from four streptococcal strains. All are linear cell wall polysaccharides, composed of repeating hexa- (or hepta-) saccharide units linked by phosphodiester bonds to the 6-carbon of the non-reducing terminal sugar of the repeating unit and all contain N-acetylgalactosamine. Two of them contain the sequence Galf((31-6)GalNAc(fl1-3)Gal(al-),in the other two it is Galf(P1-6)Ga1((3 13)GalNAc(al-) [102a]. This region is considered to be important in determining the recognition of the streptococci by the actinomyces lectins. The binding of A . naeslundii WVU45 to sialidase-treated monolayers of epithelial cells was inhibited by pretreatment of the latter with peanut agglutinin and Bauhinia purpurea lectin [103]. Although these two lectins differ in their fine specificity, both react well with Gal@ 1-3)GalNAc, and similar to the actinomyces lectins, their binding to epithelial cells is enhanced by treatment of the cells with sialidase. In contrast, Erythrina cristagalli lectin, specific for Gal@ 14)GlcNAc, failed to inhibit bacterial binding. These and other experiments with lectins led to the conclusion that the receptor for the actinomyces lectin on epithelial cells is most likely 0-linked GaI((31-3)GalNAc. A. naeslundii binds to a glycoprotein of mw 160 kDa extracted from oral epithelial cells [67]. A. naeslundii WVU45, as well as some other strains of actinomyces expressing type 2 fimbriae, bind also to sialidase-treated polymorphonuclear leukocytes, resulting in the activation of the latter cells, phagocytosis and destruction of the bacteria. The interaction of A . naeslundii WVU45 with the leukocytes was inhibited by the same lectins that inhibited the binding of these bacteria to oral epithelial cells. The receptors for the organism on the surface of sialidase-treated polymorphonuclear cells were identified as a 130 kDa glycoprotein as well as asialoganglosides with Gal((31-3)GalNAc termini [ 1041. Certain actinomyces species, such as A. naeslundii 12104 and A. viscosus 19246 and 147, have an affinity for GalNAcfi-terminating oligosaccharides. These strains exhibit heterogeneous receptor specificities and bind to different salivary and submaxillary

488

glycoproteins on blots [20]. A 180 kDa salivary glycoprotein that binds A. naeslundii has been isolated and characterized [68]. A special type of coaggregation of oral bacteria is that resulting from the bridging between one cell type and its partner by a third organism. This happens with Prevotella loescheii PK1295, which can serve as a bridge between Streptococcus oralis 34 and Actinomyces israeliz PK14, two Gram positive oral bacteria that are otherwise unable to coaggregate. Coaggregation of P loescheii PK1295 with S. oralis 34 is inhibited by lactose, while that with A. israelii PK14 is not, indicating that two different kinds of adhesins are involved, at least one of which is a lectin [105]. The latter has been isolated and purified to electrophoretic homogeneity, the first from oral bacteria. It is a fimbriaassociated protein with a molecular weight of 450kDa and consists of six identical subunits. In its oligomeric form it agglutinates S. oralis 34 and a variety of sialidasetreated erythrocytes in a lactose-sensitive manner, while the individual protomers blocked coaggregation between P loescheii PIC1295 and S. oralis 34 [ 1061. 3.3.4. Rhizobia A lectin specific for galactose has been isolated from Bradyrhizobium japonicum [ 107, 1081. It is a protein of 38kDa that binds lactose about 15 times better that galactose and does not recognize N-acetylgalactosamine. The lectin is localized at one pole of the bacterial cell surface, which is coincident with the site of cell-cell contact in homotypic aggregation of the bacteria and in their adhesion to the cultured soybean cell line SB-1. Such topological distribution is consistent with a role for the lectin in the polar binding of the organism to soybean roots.

3.3.5. Myxobacteria Myxobacteria differ from other bacteria in being social organisms. They tend to maintain close contact with each other and to aggregate into swarms. Aggregation is a developmentally regulated process that occurs as the cells differentiate from the vegetative form into mature spores. In Myxococcus xanthus, high hemagglutinating activity was found in extracts of the mature cells, but not in early vegetative cells. A lectin was purified from extracts of the aggregated stage of this organism. In solution it exists as a monomer with an apparent molecular weight of 28 kDa. The hemagglutinating activity of the lectin was not inhibited by simple sugars, only by glycoproteins such as fetuin, glycophorin and rabbit IgG, all of which contain the 0-linked tetrasaccharide NeuAc(a23)Gal(fl1-3)[NeuAc(a2-6)]GalNAc.The penultimate galactose was directly implicated in the affinity of the lectin for the saccharide, since inhibition by asialofetuin was diminished to one-fifteenth by fl-galactosidase treatment. The lectin was detected on the surface of developmental, but not vegetative cells, localized in distinct patches at one or both of the cell poles. This localization suggests that the lectin may function in end-to-end cellular interactions during aggregation. 3.4. Fucose-specific 3.4.1. Helicobacter pylori A specificity exhibited by H. pylori (in addition to that for sialic acid) is for fucose.

489

The putative lectin discriminates between closely related fucosylated epitopes as well as carbohydrate core chains, and only binds H and Lewisb antigens expressed on lacto-series type 1 but not type 2 chains [109]. To test the hypothesis that Leb antigen functions as receptor for H. pylori and mediates its attachment to gastric pit/mucous cells [ 1 101, mice that normally do not synthesize this carbohydrate structure were genetically engineered to produce it by transfection with human ~ ~ 1 , 3 1fucosyltransferase 4 [I 1 I]. Expression of Leb in the transgenic mice was associated with acquisition of the ability to bind clinical isolates of H. pylori. Binding was blocked by pretreatment of the bacteria with soluble Leb-serum human albumin conjugates.

3.4.2. Pseudomonas aeruginosa PA-11, an intracellular lectin of P aeruginosa, is specific for fucose with an unusually high affinity ( K , = 1.5x lo6 M-') and interacts weakly also with mannose ( K , = 3.1 x lo2 M-2) [112]. 3.4.3. Vibrio cholerae Agglutination of human group 0 erythrocytes by K cholerae and adhesion of the organism to brush borders are specifically inhibited by fucose, and to a lesser extent by niannose. It has been suggested that structures on eukaryotic cell surfaces containing fucose may function as receptors for a vibrio lectin and may therefore be an important determinant of host susceptibility to these bacteria. A soluble lectin specific for fucose, produced by ci cholevae strain CA401 was purified to apparent homogeneity and was found to focus at three different PI: 6.3, 5.3 and 4.7 [I 131. Thus, there are apparently three distinct PI isotypes of the lectin that exist as non-covalently associated polymers of 32kDa subunits. The lectin possessed proteolytic activity, which likewise focused at pH values 6.3, 5.3 and 4.7. It was therefore concluded that the soluble lectin is a bifunctional molecule, capable of mediating hemagglutination and proteolysis. 3.4.4. Others Recently, a fucose-specific lectin associated with the bacterial surface of Rhizobium lupinii has been purified, but its role in the interactions between the bacteria and the lupin root has not been established [ I 141. A fucose-specific lectin has also been isolated from the cell walls of Agrobacterium tumefaciens, a bacterium belonging to the Rhizobiaceaea family which infects dicotyledonous plants and forms crown gall [ 1 151. 3.5. Multiple specijcities 3.5.1. Bordetella pertussis Pertussis toxin, produced by virulent strains of B. pertussis, the etiological agent of whooping cough, is a classical A-B type toxin comprised of an A subunit that possesses ADP-ribosyltransferase activity and is responsible for most of the biological effects of the toxin, and a B subunit with affinity for carbohydrates. The B subunit of the pertussis toxin is a pentamer composed of four different subunits (S2-S5). The toxin acts as a hemagglutinin and exhibits dual carbohydrate specificity, due to

490

Selectins

Toxin subunits S 2 P l 9 Y G R c A N K T R A L T = S3

419

Y G R

P N C; T R A

T

E 1.

R'6

E J ' Y I . I I H V T I < C W ~ '

E

Rz6

T"

I.

u I<

Q I T I' <; W

si'

Fig. 4.Sequence homology between subunits S2 and S3 of Bordetellupertussis toxin and the selectins. Identical residues are underscored. Gaps have been introduced for optimal alignment. (Modified from ref. [123].)

the presence of a subunit (S2) that binds galactose-containing glycoconjugates such as lactosylceramide, and a subunit (S3) that recognizes sialylated glycoconjugates (e.g. gangliosides). Competitive inhibition studies using recombinant subunits indicated that both S2 and S3 mediate, in an additive manner, the toxin dependent attachment of B. pertussis to human macrophages [ 1 161. It was suggested, therefore, that each subunit interacts with a different receptor(s). The two subunits exhibit 80% homology. Their amino terminal regions contain a hexapeptide (residues 18-23) homologous to a sequence (residues 62-67) found in the carbohydrate binding pocket of wheat germ agglutinin (WGA); mutations in this hexapeptide of S2 abolished carbohydrate recognition [ 1171. The WGA-like region is followed by a segment (residues 30-53) which shows sequence similarity to selectins [I 181, a family of C-type eukaryotic lectins [I 191 (Fig. 4). The amino termini of S2 and S3 are thus a composite of a plant and an animal lectin arranged in tandem. This region appears to be important in determining the carbohydrate specificity of the subunits. Chimeric recombinant subunits S2 and S3, obtained by mutational interchange of amino acid residues 37-52, exhibited switched specificities, i.e. S2 became specific for gangliosides and S3 for lactosylceramide [ 1201. Concomitantly, a switch of target cell specificity occurred with respect to macrophages and ciliated cells. Mutations at Asn93 or Asn"' of S2 [121] or at Tyr8* or Ly.sl0' of S3 [122] abolished carbohydrate binding. The sequence similarity with selectins extends to shared function, in that S2 and S3 inhibit competitively the interaction of neutrophils with surfaces coated with P-selectin and E-selectin, respectively [ 1231. The recent solution of the crystal structure of a dimeric form of pertussis toxin provided support for the composite nature of the molecule[l24]. Thus, the WGAlike region of S2 superimposes on the homologous part of the three-dimensional structure of WGA (residues 62-67), while the segment homologous to the C-type lectins superimposes on the core of rat mannose binding protein, a C-type lectin. The crystal structure of the dimeric form of the toxin in complex with a diantennary undecasaccharide isolated from human serum transferrin was determined at 3.5 A [ 1251.

49 1

The carbohydrate is bound to equivalent sites on S2 and S3 of the B oligomer, occupying three of the four combining sites present in the dimeric unit. Despite the relatively low resolution, NeuAc(a2-6)Gal (which forms the terminus of each branch of the undecasaccharide) was clearly identified in the binding sites with the aid of energy calculations and NMR spectroscopy. The sialic acid forms both hydrogen bonds and hydrophobic interactions with the protein, while no contacts with the galactose are seen. Surprisingly, in the crystal structure, S2 and S3 interact with sialic acid in the same way, although the two subunits appear to exhibit different sugar specificities in solution. In oitro, the toxin binds to glycoproteins as well as glycolipids. The ability of fetuin to inhibit hemagglutination of goose erythrocytes by the toxin was abolished by treatment of the glycoprotein with sialidase and was restored by resialylation with the aid of a2-6 sialyltransferase [126]. On blots of separated proteins from Chinese hamster ovary (CHO) cells, the intact toxin, as well as the purified B subunit, bound a single band of 165 kDa; sialidase treatment of the blots abolished binding [69]. With lectin-resistant CHO mutants deficient in sialic acid, no toxin-binding component was revealed [ 1271. The toxin did not bind to neutral glycolipids or gangliosides from CHO cells, nor to sialoparagloboside, although the latter terminates in NeuAc-N-acetyllactosamine. It appears thus that the 165 kDa glycoprotein is the receptor for pertussis toxin on CHO cells. Toxin-binding proteins on human peripheral T lymphocytes, as well as on T-cellderived lines (HPB-ALL and Jurkatt), all of which respond to pertussis toxin by rapid second messenger production, were detected by labeling the cells with the toxin bound to a crosslinking agent, followed by gel electrophoresis, and by binding of the toxin to blots of SDS-PAGE separated membrane proteins [70]. In all cases, a single band of 43 kDa was revealed, which was absent from non-responsive cells. The 43 kDa protein may thus be a functional receptor for the toxin. In a different report, also using peripheral T lymphocytes and Jurkatt cells and a similar labeling technique as above, a protein of 70 kDa was detected [71]. A band of the same mobility was found among several components isolated from surface-iodinated T lymphocytes by affinity chromatography on the immobilized toxin. However, it cannot be concluded that binding to this (g1yco)protein was carbohydrate-mediated, since no controls with inhibitory sugars were carried out, and elution from the affinity column was with high pH and not with carbohydrate. In addition to the toxin, virulent strains of B. pertussis express another lectin, known as filamentous hemagglutinin. It is found in a cell-associated form outside the outer membrane as a microcapsule, or in a soluble form secreted into the culture medium during growth. It is a large molecule of 220kDa with an affinity for lactosylceramide when assayed by direct binding with extracts of ciliated cells or standard glycolipids and has also an affinity for sulfated saccharides, such as Gal-6-sulfate, dextran sulfate and heparin. The carbohydrate binding region of the lectin was mapped to amino acids 11411279 with the aid of a monoclonal antibody that blocked bacterial binding to ciliated cells and to lactosylceramide [128]. Mutant strains of B. pertussis with deletions of this region did not bind to ciliated cells or macrophages and reacted poorly with the antibody. The region contains a stretch exhibiting sequence similarity with 20 amino acids (30-51) of the lectin domain of S2 of the toxin.

492

4. Fungi Lectins have been found in some soilborne plant pathogenic fungi, such as Rhizoctonia solani and related species [ 129,1301 and in different members of the Sclerotiniaceae [ 13 I , 1321. Most of these lectins are specific for galactose1N-acetylgalactosamine although some (e.g. that from Sclerotium rolfsii [ 1321) recognize only animal glycoproteins such as fetuin and mucin. A lectin specific for N-acetylgalactosamine has also been isolated from the nematode-trapping fungus Arthrobotrys oligospora [ 1331. From the same source a protein has been obtained which, although immunologically identical with that mentioned above, is not inhibited by N-acetylgalactosamine, only by glycoproteins [ 1341.

5. Protozoa Only a few protozoan lectins are known (Table 3). Table 3 Carbohydrate specificity of protozoal lectins Organism

Specificity

Reference

Entamoeba histolytica

GaliCalNAc

[I351

Plasmodium falciparum

(GlcNAc), NeuAc(a2-3)Gal

[ 1521 [I451

GlcNAc

Giardia lamblia

Man-6-P

[I551

5.1. Gal and GalNAc-specijk 5.1.1. Entamoeba histolytica The pathogenic protozoan E. histolytica causes disease in humans by disruption and invasion of the colonic mucosa. It produces two lectins, specific for Gal1GalNAc [ 1351 and for chitooligosaccharides, respectively (for the latter lectin, see section 5.3. I ) . The Gal/GalNAc-specific lectin has a mw of 260kDa and is composed of two subunits (170 and 31135kDa) linked by disulfide bonds[136]. The heavy subunit is an integral membrane protein with a large cysteine-rich extracellular domain and a short cytoplasmic tail [137,138]. The two types of light subunit are synthesized from the same cDNA and the difference in molecular weight is probably the result of post-translational modifications, possibly of glycosylation [139]. The 31 kDa (but not the 35 kDa) isoform of the light subunit contains a phosphatidylinositol-glycan anchor; the 17013 1 kDa isolectin is thus an unusual protein, in that it is composed of a transmembrane and a phosphatidylinositolglycan anchored component. The purified lectin partially inhibits binding of amoebic trophozoites to cultured human intestinal epithelial and CHO cells. In addition, antibodies directed to the 170 kDa subunit

493

inhibit binding of the amoeba to CHO cells, suggesting that this subunit is primarily responsible for mediating adhesion. Interestingly, certain monoclonal anti-heavy chain antibodies caused an enhancement of galactose-inhibitable adhesion of the parasite to CHO cells [ 1401. One of the antibodies also enhanced binding of the purified lectin to such cells, indicating that enhanced adhesion of the parasite may be due to direct activation of the galactose-binding activity of the lectin. It is thus possible that amoebae regulate their in oiuo adhesion to galactose-containing substrates via changes in lectin activity. The epitopes for both the adhesion-inhibiting and adhesion-enhancing antibodies are located in the cystein-rich domain of the 170 kDa subunit, suggesting that the carbohydrate binding site may reside in this region [141]. The extent of binding of E. histolytica to wild-type CHO cells and three lectinresistant CHO mutants with altered glycosylation patterns was directly correlated with the presence of terminal galactose residues on these cells: very poor binding was observed to CHO cell mutants deficient in galactose [ 1421. Moreover, galactose derivatives, in particular N-acetyllactosamine, efficiently blocked binding of the amoebae to CHO cells, while no inhibition was seen with N-acetylglucosamine or N, N'-diacetylchitobiose. The adhesion of the parasite to CHO cells was also completely inhibited by very low concentrations (1 pg/ml) of purified human colonic mucin and this inhibition was galactose-specific [ 1431; direct binding experiments have shown that the amoebic lectin binds the mucin with a very high affinity (Kd = X X lo-" M-'). Therefore, the galactosespecific, rather than the N-acetylglucosamine-specific, lectin is most probably responsible for the recognition of mammalian cells by E. histolytica. Specificity studies with a series of multivalent glycoconjugates with non-reducing terminal galactose or N-acetylgalactosamine residues confirmed the preference of the lectin for the latter monosaccharide and revealed a dramatic increase in the affinity of the lectin for its ligands when presented in multivalent form [144]. A neoglycoprotein having an average of 40 galactose residues linked to bovine serum albumin was about 17000 times more potent as inhibitor of hemagglutination by the lectin than the monovalent sugar, and a neoglycoprotein with an average of 39 N-acetylgalactosamine residues was 140 000-fold more potent than the monosaccharide. The latter polymer bound to membranes of E. histolytica with a dissociation constant of about IOnM. These findings are in line with the above mentioned high affinity of the lectin for human colonic mucin and support the possibility that this class of glycoprotein is a potential target for E. histolytica binding in uiuo.

5.2. Sialic-acid-spec@ 5.2.I . Plasmodium fakiparum The invasion of human erythrocytes by plasmoidal merozoites is a key event during malaria infection. Sialidase treatment of the erythrocytes renders the cells resistant to invasion by different isolates of I? fakiparum, suggesting that the interaction between the parasite and the erythrocytes is mediated by a plasmodia1 lectin specific for sialic acid [ 1451. Different mezoroite surface components have been proposed as the putative lectins, among them the well-defined surface antigen Pf200 [ 1461. Binding of the isolated

494

antigen to erythrocytes was inhibited by sialidase treatment of the cells and by monoclonal antibodies against the glycosylated domain of glycophorin. A protein of 175 kDa (designated EBA-175), which is released from merozoites into the culture medium, was purified by binding to, and elution from, human red blood cells [147]. Its interaction with the erythrocytes was inhibited best by NeuAc(a2-3)Gal and less by NeuAc(a24)Gal; free N-acetylneuraminic acid was not inhibitory. The structure of the sialic acid is critical, because removal of the 9-0-acetyl group from mouse erythrocytes, which converted the mouse sialic acid to the human form, enhanced binding of EBA-175 to the mouse cells [148]. The cDNA of the lectin was cloned; it codes for a polypeptide chain of 1435 amino acids, including a 19 residue leader sequence [149]. Antibodies against a synthetic peptide encompassing amino acid residues 1062-1 103 inhibited merozoite invasion in uitro, as well as binding of the purified lectin to erythrocytes, indicating that this region may include the carbohydrate combining site of the lectin. In line with this possibility was the finding that the nucleotide sequence of this region is conserved among several plasmodium strains from various regions of the world. The inhibitory activities of the antipeptide antibodies were shown to be primarily, if not exclusively, directed against amino acids 1069-1087 [150]. Antibodies against a peptide corresponding to the above amino acid stretch inhibited the growth of the parasite similarly to antibodies against the larger peptide and, moreover, blocked the interactions of the latter antibodies with the purified lectin and with the parasite. Surprisingly however, using truncated portions of the lectin expressed in COS cells, the binding region was mapped to a cystein-rich domain closer to the N-terminal of the protein [I511 The fragment (amino acids 487-673, designated as region F2) bound erythrocytes with the same pattern as native EBA-175. The receptors for the plasmodia1 sialic-acid-specific lectin on human erythrocytes are the 0-linked chains of glycophorin. Cells devoid of glycophorin (MKMK erythrocytes), or having glycophorin with modified 0-chains (Tn and Cad) are, to a large extent, resistant to invasion [147]. Interestingly, this is also the case with En(a-) erythrocytes that contain glycophorin B, but not glycophorin A. These cells also failed to bind to EBA-175 and fragment F2 [151]. The two glycophorins contain the same 26 amino acids and eleven clustered 0-linked chains at the N-terminus, but differ in the rest of the molecule. This was taken to indicate that an amino acid sequence specific for glycophorin A is necessary for binding, possibly by contributing an unique conformation to the sialic acid residues recognized by the lectin. This is supported by the finding that glycopeptide 1-64 of glycophorin A, but not a mixture of glycopeptides 1-34 and 35-64, inhibited the binding of erythrocytes to EBA-175 [151].

5.3. N-Acetylglucosamine- and chitooligosaccharide-speciJic 5.3.I. Entamoeba histolytica The E. histolytica lectin specific for chitooligosaccharides is a 220 kDa plasma membrane glycoprotein that inhibits the attachment of amoebic tropozoites to layers of cultured MDCK cells [152,153].

495

5.3.2. Plasmodium falciparum Besides the sialic-acid-specific lectin (see section 5.2. l), isolates of P falciparum from different geographical regions express a lectin specific for N-acetylglucosamine [ 1541. The merozoite surface of such isolates was specifically labeled with a neoglycoprotein bearing multiple residues of this monosaccharide, as evidenced by fluorescence and electron microscopy. Affinity chromatography of merozoite lysates on immobilized N-acetylglucosamine and elution with the free sugar yielded three protein bands of 45, 83 and 120kDa. The same bands were revealed when blots of merozoite proteins were probed with the N-acetylglucosamine containing neoglycoprotein. 5.4. Mannose-6-phosphate-spec& 5.4.1. Giardia lamblia

Giardia lamblia, a protozoan parasite that causes widespread diarrheal disease, expresses a surface membrane associated lectin, named taglin, which is specifically activated by limited proteolysis with trypsin, a protease present in abundance at the site of infection [ 1551. Trypsin-treated lysates of G. lamblia trophozoites agglutinate rabbit erythrocytes as well as enterocytes, which are the target cells of the parasite in uioo. The agglutination was best inhibited by mannose 6-phosphate; mannose 1-phosphate and other phosphorylated sugars were not inhibitory. A monoclonal antibody which inhibited hemagglutination by the activated lysate, and is thus presumably directed against the lectin, recognized a protein of 28/30kDa on immunoblots of trophozoite lysates. The lectin activity of this protein was demonstrated by binding of erythrocytes to bands in the same molecular weight range as those recognized by the monoclonal antibody. Immunoprecipitation of G. lamblia lysates with the monoclonal antibody also yielded a protein of 28/30 m a , which has however not been further characterized.

6. Biological roles 6.I . Infection A major role of the microbial lectins is to mediate the adhesion of the organisms to host cells, an initial stage of infection. This has been extensively demonstrated both in uitro, in studies with isolated cells and cell cultures, and in uiuo in experimental animals.

6.I . 1. Viruses The binding of the influenza virus lectin (hemagglutinin) to carbohydrates containing sialic acid on the surface of the target cells leads to the attachment of the virus to the cells. This is followed by fusion of the viral and cellular membranes, allowing release of the viral genome into the cytoplasm and subsequent replication. Removal of sialic acid from the cell membranes by sialidase abolishes binding and prevents infection, while enzymatic reattachment of sialic acid or insertion of sialic-acid-containing oligosaccharides (for example, in the form of glycolipids) into the membranes of sialidase-treated cells restores the ability of the cells to bind the virus and to be infected by it [26,27]. Detailed

496

knowledge of the sialic-acid-hemagglutinin interaction (cf. section 2.1.1) provides a basis for the attempts to design antiviral drugs that would block viral attachment to cells. An inhibitor, targeted to the conserved amino acids of the combining site or to the receptor carbohydrate, might be effective against influenza viruses of different subtypes. It would be independent of the antigenic changes that accompany the recurrent epidemics for which these viruses are renowned.

6.1.2. Bacteria In the case of bacteria that possess the ability to produce fimbriae, the infectivity of the fimbriated phenotype is, as a rule, higher than that of the non-fimbriated [2,4]. Thus, when mice were infected intravesicularly with two isogenic E. coli mutants, both of which were type 1 fimbriated, but only one of which possessed mannose-binding activity, only the latter induced urinary tract infections. Also, infection by enterobacteria carrying fimbrial lectins was prevented by passive or active pre-immunization of the animals against the fimbriae. Although the anti-fimbrial antibodies inhibited bacterial adhesion in uitro, there is no compelling evidence that they are directed against the carbohydrate binding sites of the fimbrial lectin. Table 4 Inhibitors of sugar-specific adhesion prevent infection in animals and humans Organism

Animal

Site of infection

Inhibitor

Ref.

Type 1 fimbriated Escherichia coli

Mice

Urinary tract

MeaMan

[41

Mice

Gastrointestinal tract

Mannose

[41

Mice

Urinary tract

Anti-Man antibody

[41

Klebsiella pneumoniae

Rats

Urinary tract

MeaMan

[41

Shigella Jexnerii

Guinea pigs

Eye

Mannose

[41

K99 fimbriated E. coli

Calves

Gastrointestinal tract

Glycopeptides of bovine plasma glycoproteins

[ 1581

Bordetella pertussis (toxin)

Rabbit

Lung

Lactose, anti-Lea antibody

[I661

Pseudomonas aeruginosa

Man

External auditory canal

Gal + Man +NeuAc

[I591

More significantly, in three separate systems, each employing different type 1 fimbriated enterobacterial species, mannose and methyl a-mannoside inhibited infection of the urinary tract of mice and rats (Table 4). In each of these glucose (or methyl a-glucoside) which is not an inhibitor of type 1 fimbriae, did not affect the infectivity of the injected bacteria. In addition, antibodies against the carbohydrate structures recognized by type 1 fimbriae on epithelial cells also prevented urinary tract infection in mice by type 1 fimbriated E. coli. Similarly, N-acetylneuraminic acid considerably reduced colonization of lung, liver and kidney by I? aeruginosa injected intravenously

497

None

PT-/FHA-

Lactose

AntiLea

Sialic acid

AntiA

Anti-

CR3

r H *

PT

RAD 1098 Anti-CR3 Sialic acid Lactose Anti-A

* -

__

--t7

Anti-Le a VIR+

I

0 20 40 60 80 100 Lung weight (% of VIR+ over control) Fig. 5. Effect of adhesion inhibitors (A) on the colonization of the respiratory tract by Bordete//a periussis and (B) on pulmonary edema in animals challenged with the virulent BP536 strain. Anti-Lea and anti-A are antibodies against blood group determinants Lea and A, respectively; anti-CR3 is the antibody against the receptor for complement fragment C3bi. VIR+, control without inhibitor. (Modified from ref. [I 571.)

to mice [156]. It was also demonstrated that introduction into the trachea of rabbits infected with B. pertussis of lactose or monoclonal anti-Lea antibodies resulted in the abortion of colonization of the respiratory tract by the bacteria and blocking of pulmonary edema [157] (Fig. 5 ) . These findings provide some of the most convincing evidence for

498

the central role of bacterial lectins in infection, in particular in mucosal colonization. They also illustrate the great potential of simple carbohydrates in the prevention of infections caused by bacteria that express surface lectins and raise hopes for the development of anti-adhesive drugs for human use. Other reports also support a role for bacterial lectins in natural infections. Sialylated glycoproteins, administered orally, protected colostrum-deprived, newborn calves against lethal doses of enterotoxigenic E. coli K99 [158]. In a clinical trial in humans, patients with otitis externa (a painful swelling with secretion from the external auditory canal) caused by P aeruginosa were treated at the site of infection with a solution of galactose, mannose and N-acetylneuraminic acid. The results were fully comparable to those obtained with conventional antibiotic treatment [ 1591. Expression of the bacterial lectins during the infection process is not always beneficial to the survival of the bacteria. Thus, type 1 fimbriated bacteria may adhere to phagocytic cells, leading to ingestion and killing of the bacteria (see section 6.2) or combine with soluble glycoproteins such as IgA. Since colonic mucosa is known to be rich in IgA2 secreting cells, it has been suggested that one important function of secretory IgA could be to bind and agglutinate intestinal bacteria and thus prevent their attachment to host cells [55]. Tamm-Horsfall glycoprotein binds mannose-specific E. coli via its single oligomannose unit [56] and may thus serve as a vehicle for the clearance of bacteria from the urinary tract. A novel mechanism for lectin-dependent adhesion has been proposed in the binding of P aeruginosa to animal cells, based on the observation that when lysates of this organism were mixed with the intact bacteria, the latter acquired the ability to bind to corneal epithelial culture cells and that the binding was specifically inhibited by galactose and mannose [ 1601. It was therefore proposed that the intracellular lectins of P aeruginosa (PA-1 and PA-2, cf. sections 3.3.2 and 3.4.2), released from the bacteria in uiuo, bind to surviving intact cells and mediate adhesion of the organisms to galactose- or mannosecontaining glycoconjugates on epithelial cells. The involvement of mannose residues as receptors for P aeruginosa is supported by studies showing that concanavalin A, a mannose-specific lectin, inhibited binding of the bacteria to corneal epithelia [ 1611. The b-galactoside-specific lectins produced by oral actinomyces, such as A . naeslundii and A . uiscosus, facilitate initial colonization of epithelial surfaces of the mouth and teeth by mediating the attachment of the bacteria to galactose residues either on the surface of the epithelial cells, or on the surface of other bacteria (for example, S. sanguis) which are adsorbed to the enamel of the teeth [162]. 6.1.3. Fungi

The lectins of pathogenic fungi have been implicated in the specific interactions between plant pathogens and Trichoderma, a natural antagonist to other fungi and a well-known biocontrol agent, leading to the destruction of the pathogen [ 1631. Direct evidence for this hypothesis was obtained with the aid of a biomimetic system, in which nylon fibers coated with a fungal lectin simulate hyphae of the fungus from which the lectin is derived. When Trichoderma harzianum was allowed to grow on fibers coated with the purified lectin from Sclerotium rolfsii, it coiled around the nylon fibers and produced hooks in a pattern similar

499

Fig. 6 . Scanning electron micrographs of the different stages of the interaction between Trichoderma harzianum and nylon fibers coated with purified Sclerotium rolfsii lectin. (A) A typical branching of the Trichoderma towards the fibers and contact of the branch tip with the fiber surface (bar= 10pm); (B) subsequent elongation of the firmly attached tip along the fiber surface (bar= 1 pm); (C) Trichoderma hyphae coiled fibre, producing additional branches. The hyphal coils and branches adhere tightly to the fiber surface (bar= 10pm). (From ref. [132], by permission from publisher.)

to that observed with the real host hyphae [ 1321 (Fig. 6). The incidence of interaction was significantly (six times) higher with lectin-treated fibers than with untreated ones. 6.1.4. Protozoa The inhibitory effect of the purified GalIGalNAc-specific lectin of E. histolytica and of anti-lectin antibodies on the binding of the parasite in oitro to animal cells and on the contact-dependent killing of the latter prompted attempts to use the lectin as a protective agent in uiuo. Pre-immunization of gerbils with the purified lectin provided complete protection from liver abscesses in a majority (67%) of animals injected intrahepatically with E. histolytica trophozoites [ 1641. However, the sera of the immunized animals, while completely blocking adhesion at 1 :10 dilution, increased adhesion when tested at 1:1000 dilution. This finding, which is analogous to the enhancement of adhesion by certain

500

monoclonal anti-lectin antibodies mentioned above, complicates attempts to understand the role of adhesion-blocking antibodies in protection against disease. The ability to elicit protective immunity has now been mapped to a cysteine-rich fragment of the lectin, encompassing amino acids 758-1 134 [ 1651. Facilitating the adhesion of amoebae to intestinal epithelial cells of the host is only one of the ways in which the lectins may enhance the pathogenicity of the parasite. For instance, the Gal/GalNAc-specific lectin increases the resistance of E. histolytica trophozoites to complement-dependent killing [ 1661. Also, once invasion has taken place and the amoebae have spread through the host, the lectins mediate binding of the parasite to other cells and tissues, in particular to hepatocytes, initiating the killing of these cells. In addition, the lectins enable the amoebae to bind bacteria carrying the appropriate sugars. The bound bacteria are subsequently ingested and serve as a source of nutrition for the parasite, increasing its virulence [167].

6.2. Non-opsonic phagocytosis Different microorganisms in serum-free media attach readily via their surface lectins to phagocytic cells. Such attachment in the absence of opsonins is often followed by activation of the phagocytes and ingestion of the bacteria; sometimes, killing of the bacteria is observed. The process has been named lectinophagocytosis, in analogy to opsonophagocytosis, in which recognition between the microorganisms and the phagocytic cells is mediated by serum constituents termed opsonins (mainly IgG antibodies and the C3b and C3bi fragments of the C3 component of complement) [168,169]. The best characterized system of lectinophagocytosis is that of bacteria carrying mannose-specific lectins, mainly E. coli, in the form of type 1 fimbriae. Shigellaflexneri M90T, a nonfimbriated organism that, unless opsonized, does not bind to or activate phagocytic cells, when transfected with the cluster of genes encoding type 1 fimbriae of E. coli, bound to human granulocytes and mouse peritoneal macrophages, and activated the latter [ 1701. As already mentioned (section 3.1.l), the adhesion molecules CDI l/CD18 serve as the major leukocyte receptors for type 1 fimbriae [59,60]. Because CDI lb,c/CD18 constitute the receptor CR3 for complement fragment C3bi, the findings indicate that the receptors that participate in opsonophagocytosis function in lectinophagocytosis as well. Lectinophagocytosis mediated by bacterial surface lectins also occurs with type 2 fimbriated Actinomyces species, specific for Gal(f3-3)GalNAc and Gal(fi14)Glc [ 17 I ] (cf. section 3.3.3). Phagocytosis of mutants deficient in such fimbriae is negligible. Binding and phagocytosis are markedly enhanced by pretreating the granulocytes with sialidase. B. pertussis too adheres to, and is internalized by cultured macrophages in oitro, and alveolar macrophages in oioo, in the absence of opsonins. It is, however, not killed, but persists within the cells. Either one of the lectins (the filamentous hemagglutinin or the toxin) can mediate the adhesion of the bacteria to the macrophages. In this case, however, the hemagglutinin does not act as a lectin, since it binds through its Arg-GlyAsp sequence rather than through the carbohydrate combining site; this binding seems to be a prerequisite for the subsequent uptake of the bacteria. The adhesion of B. pertussis to macrophages proved to be a cooperative process, involving both the toxin and the

501

hemagglutinin. Pretreatment of the macrophages with the whole toxin, the B subunit or recombinant S2 or S3 enhanced the adhesion of hemagglutinin-producing bacterial strains, but had no effect on adhesion of hemagglutinin-deficient bacteria [I 161. The enhanced binding was CDI IWCD18 dependent, as demonstrated by increased binding of erythrocytes coated with C3bi (known to bind to the above antigens) to macrophages treated as above. These results suggest a communication between the receptor(s) for the toxin and the integrin [ 1 161. It has been argued that lectinophagocytosis may function as a defence mechanism against microbial infections in uiuo at sites, such as the renal medulla and peritoneal cavity, especially during dialysis, or in situations (e.g. in patients infected by microorganisms prior to the development of an immune response) where opsonic activity is poor [168]. Indirect evidence for such a role came from experimental infections with mixed bacterial phenotypes (isogens) of E. coli, one of which was type 1 fimbriated and the other of which was non-fimbriated. Whenever the organisms reached phagocyte-rich sites, those with the non-fimbriated phenotype survived, while at phagocyte-poor sites those with the fimbriated phenotype survived, irrespective of the bacterial species employed or the route or site of infection. The selective survival of the non-fimbriated phenotype in phagocyterich sites was attributed to elimination of the fimbriated phenotype by phagocytes. It was suggested that phase variation, a random on-off process that allows the bacteria to alternate between fimbriated and non-fimbriated states, and thus to choose the phenotype conducive for their survival, is an important virulence trait of type 1 fimbriated bacteria. Recent experiments provide direct evidence for the possibility that lectinophagocytosis occurs in uiuo. Injection of type 1 fimbriated, but not of non-fimbriated, E. coli cells into the peritoneal cavity of mice led to the activation of peritoneal macrophages, as measured by the release of the lysosomal enzyme, N-acetyl-fi-glucosaminidase [ 1721. Methyl coinjected into the peritoneum with the bacteria, specifically inhibited the release of the enzyme. Furthermore, no release was observed following injection of fimbriated bacteria into a macrophage-depleted peritoneum. If indeed lectinophagocytesis will prove to be an important defense mechanism against infections, it should be taken into account in attempts to develop anti-adhesive drugs, especially those based on blocking carbohydrate mediated recognition by different infectious agents.

References [ I ] Karlsson, K.-A. (1995) Curr. Opin. Struct. Biol. 5, 622-635. [2] Sharon, N. (1986) In: I.E. Liener, N. Sharon and 1.J. Goldstein (Eds.), The Lectins: Properties, Functions and Applications in Biology and Medicine. Academic Press, Orlando, ch. 9, pp. 493-526. [3] Mirelman, D. (Ed.) (1986) Microbial Lectins and Agglutinins. John Wiley and Sons, New York. [4] Ofek, I. and Sharon, N. (1990) Curr. Top. Microbiol. Immunol. 152, 91-113. [S] Ofek, I . and Doyle, R.J. (1994) Bacterial Adhesion to Cells and Tissues. Chapman and Hall, New York, London. [6] Hultgren, S.J., Abraham, S., Caparon, M., Falk, P.,St. Geme 111, J.W. and Normark, S. (1993) Cell 73, 887-901. [7] Yu, C., Lee, A.M., Bassler, B.L. and Roseman, S. (1991) J. Biol. Chem. 266, 24260-24267. [8] Sharon, N. and Ofek, I. (1995) Methods Enzymol. 253, 91-98. [9] Karlsson, K.-A. (1989) Annu. Rev. Biochem. 58, 309-350.

502

[lo] Karlsson, K.-A,, Angstrom, J., Bergstrom, J. and Lanne, B. (1992) APMS (Suppl. 27) 100, 71-83. [ I I ] Karlsson, K.-A, Milh, M.A., Angstrom, J., Bergstrom, J., Dezfoolian, H., Lanne, B., Leonardsson, 1. and Teneberg, S. (1992) In: T.K. Korhonen et al. (Eds.), Molecular Recognition in Host-Parasite Interactions. Plenum Press, New York, pp. 1 15-1 32. [I21 Lenz, T.L. (1990) J. Genet. Virol. 71, 751-766. [I31 Brown, K.E., Anderson, S.M. and Young, N.S. (1993) Science 262, 114-1 17. [I41 Stehle, T., Yan, Y., Benjamin, T.L. and Harrison S.C. (1994) Nature 369, 160-163. [15] Hanisch, F.-G., Hacker, J. and Schroten, H. (1993) Infect. Immun. 61, 2108-2115. [I61 Prasadarao, N.V., Wass, C.A. and Kim, K.S. (1996) Infect. Immun. 64, 154-160. [I71 Seignole, D., Grange, P., Duval-Iflah, Y. and Mouricourt, M. (1994) Microbiology 140, 2467-2473. [I81 Seignole, D., Mouricout, M., Duval-Iflah, Y., Quintard, B. and Julien, R. (1991) J. Genet. Microbiol. 137, 1591-1601. [I91 Stromberg, N. and Karlsson, K.-A. (1990) J. Biol. Chem. 265, 11251-11258. [20] Stromberg, N., Boren, T., Carlen, A. and Olsson, J. (1992) Infect. Immun. 60, 3278-3286. [21] Neeser, J.-R., Grafstrom, R.C., Woltz, A,, Brassart, D., Fryder, V. and Guggenheim, B. (1995) Glycobiology 5, 97-104. [22] Jimenez-Luch, V., Ginsburg, V. and Krivan, H.C. (1990) Infect. Immun. 58, 2085-2090. [23] Brassart, D., Woltz, A,, Golliard, M. and Neeser, J.-R. (1991) Infect. Immun. 59, 1605-1613, [24] Gottschalk, A. (1960) The Chemistry and Biology of Sialic Acids and Related Substances. Cambridge University Press, 1 15 pp. [25] Schauer, R. (1991) Glycobiology 1, 449452. [26] Paulson, J.C. (1985) In: P.M. Conn (Ed.), The Receptors, Vol. 2. Academic Press, New York, pp, 131219. [27] Wiley, D.C. and Skehel, J.J. (1987) Annu. Rev. Biochem. 56, 365-394. [28] Weis, W., Brown, J.H., Cusack, S., Paulson, J.C., Skehel, J.J. and Wiley, D.C. (1988) Nature 333, 4 2 6 4 3 1. [29] Nieman, H., Dabrowski, J., Dabrowski, U., Geyer, R., Keil, W., Klenk, H.D. and Stirm, S. (1985) Eur. J. Biochem. 146, 523-532. [30] Vlasak, R., Luytjes, W., Spaan, W. and Palese, P. (1988) Proc. Natl. Acad. Sci. USA 85, 4526. [31] Kelm, S., Paulson, J.C., Rose, U., Brossmer, R., Schmid, W., Bandgar, B.P., Schreiner, E., Hartmann, M. and Zbiral, E. (1992) Eur. J. Biochem. 205, 147-153. [32] Sauter, N.K., Hanson, J.E., Glick, G.D., Brown, J.H., Crowther, R.L., Park, S.-J., Skehel, J.J. and Wiley, D.C., 1992, Biochemistry 3 1, 9609-9621. [33] Watowich, S.J., Skehel, J.J. and Wiley, D.C. (1994) Structure 2, 719-731. [34] Sauter, N.K., Glick, G.D., Crowther, R.L., Park, S.-J., Eisen, M.B., Skehel, J.J., Knowles, J.R. and Wiley, D.C. (1992) Proc. Natl. Acad. Sci. 89, 324-327. [35] Shieh, M.T., WuDunn, D., Montgomery, R.I., Esko, J.D. and Spear, P.G. (1992) J. Cell Biol. 116, 1273-1281. [36] Haidar, M., Gluckman, J.C. and Gattegno, L. (1992) Glycobiology 2, 429435. [37] Mbemba, E., Gluckman, J.C. and Gattegno, L. (1994) Glycobiology 4, 13-21. [38] Duguid, J.P. and Old, D.C. (1980) In: E.H. Beachey (Ed.), Receptors and Recognition, Vol. 6. Chapman and Hall, London, pp. 185-217. [39] Klemm, P. (1985) Rev. Infect. Dis. 7, 321-340. [40] Clegg, S. and Gerlach, G. (1987) J. Bacteriol. 169, 934-938. [41] Tewari, R., MacGregor, J.I., Ikeda, T., Little, J.R., Hultgren, S.J. and Abraham, S.N. (1993) J. Biol. Chem. 268, 3009-3015. [42] Jones, C.H., Pinkner, J.S., Nicholes, A.V, Slonim, L.N., Abraham, S.N. and Hultgren, S.J. (1993) Proc. Natl. Acad. Sci. USA 90, 8397-8401. [43] Hultgren, S.J., Normark, S. and Abraham, S. (1991) Annu. Rev. Microbiol. 45, 383415. [44] Jacob-Dubuisson, F., Kuehn, M. and Hultgren, S.J. (1993) Trends Microbiol. 1, 50-55. [45] Labigne-Roussel, A., Schmidt, M.A., Walz, W. and Falkow, S. (1985) J. Bacteriol. 162, 1285-1292. [46] Abraham, S.N., Freitag, C.S., Gander, R.M., Clements, J.R., Thomas, V.L. and Eisenstein, B.I. (1986) Mol. Biol. Med. 3, 495-508.

503 [47] Sharon, N. and Ofek, I. (1986) In: D. Mirelman (Ed.), Microbial Lectins and Agglutinins. Wiley, New York, pp. 55-8 1 . [48] Abraham, S.N., Sun, D., Dale, J.B. and Beachey, E.H. (1988) Nature, 336, 6 8 2 4 8 4 . [49] Firon, N., Ofek, 1. and Sharon, N. (1984) Infect. Immun. 43, 1088-1090. [50] Madison, B., Ofek, I., Clegg, S. and Abraham, N.S. (1994) Infect. Immun. 62, 843-848. [51] Sokurenko, E.V, Courtney, H.S., Abraham, S.N., Klemm, P. and Hasty, D.L. (1992) Infect. Immun. 60, 47094719. [52] Oldenburg, K.R., Loganathan, D., Goldstein, I.J., Schultz, P.G. and Gallop, M.A. (1992) Proc. Natl. Acad. Sci. USA 89, 5393-5397. [53] Scott, J.K., Loganathan, D., Easley, R.B., Gong, X. and Goldstein, I.J. (1992) Proc. Natl. Acad. Sci. 89, 5398. [54] Giampapa, C.S., Abraham, S.N., Chiang, T.M. and Beachey, E.H. (1988) J. Biol. Chem. 263, 53625367. [55] Wold, A.E., Mestecky, J., Tomana, M., Kobata, A., Ohbayashi, H, Endo, T. and Svanborg-Eden, C. (1990) Infect. Immun. 58, 3073-3077. [56] Kumar, S. and Muchmore, A.V (1990) Kidney Int. 37, 1395-1401. [57] Wadolkowski, E.A., Laux, D.C. and Cohen, P.S. (1988) Infect. Immun. 56, 1036-1043. [58] Rodriguez-Ortega, M., Ofek, I. and Sharon, N. (1997) Infect. Immun. 55, 968-973. [59] Gbarah, A., Gahmberg, C.G., Ofek, I., Jacobi, I . and Sharon, N. (1991) Infect. Immun. 59, 45244530. [60] Gbarah, A,, Gahmberg, C.G., Bonner, G. and Sharon, N. (1993) J. Leukoc. Biol. 54, 1 1 1-1 13. [61] Sauter, S.L., Rutherfurd, S.M., Wagener, C., Shively, J.E. and Hefta, S.A. (1991). Infect. Immun. 59, 2485-2493. [62] Pieroni, P., Worobec, E.A., Paranchych, W. and Armstrong, G.D. (1988) Infect. Immun. 56, 1334-1340. [63] Hermentin, P. (1987) Parasitol. Today 3, 52-55. [64] Doig, P., Paranchych, W., Sastry, P.A. and Irvin, R.T. (1989) Can. J. Microbiol. 35, 1141-1 145. [65] Rudner, X.L., Zheng, Z., Berk, R.S., Irvin, R.T. and Hazlett, L.D. (1991) Invest. Ophtalmol. Visual Sci. 33, 2185. [66] Demuth, D.R., Golub, E.E. and Malamud, D. (1990) J. Biol. Chem. 265, 7120-7126. [67] Brennan, M.J., Cisar, J.O. and Sandberg, A.L. (1986) Infect. Immun. 52, 840-845. [68] Babu, J.P., Dabbous, M.K. and Abraham, S.N. (1991) J. Periodont. Res. 26, 97-1 10. [69] Brennan, M.J., David, J.L., Kenimer, J.G. and Manclark, C.R. (1988) J. Biol. Chem. 263,48954899. [70] Rogers, T.S., Corey, S.J. and Rosoff, P.M. (1990) J. Immunol. 145, 678-683. [71] Clark, C.G. and Armstrong, G.D. (1990) Infect. Immun. 58, 3840-3846. [72] Buhler, T., Hoschutzky, H. and Jann, K. (1991) Infect. Immun. 59, 3876-3882. [73] DeGraaf, F.K. (1 990) Curr. Top. Microbiol. Immunol. 15 1, 29-53. [74] Morschhauser, J., Hoschutzky, H., Jann, K. and Hacker, J. (1990) Infect. Immun. 58, 2133-2138. [75] Parkkinen, J., Rogers, G.N., Korhonen, T., Dahr, W. and Finne, J. (1986) Infect. Immun. 54, 3 7 4 2 . [76] Prasadarao, N.V, Wass, C.A., Hacker, J., Jann, K. and Kim, K.S. (1993) J. Biol. Chem. 268, 1035610363. [77] Neeser, J.-R., Chambaz, A,, Hoang, K.Y.and Link-Amster, H. (1988) FEMS Microbiol. Lett. 49, 301-307. [78) Sjoberg, PO., Lindahl, M., Porath, J. and Wadstrom, T. (1988) Biochem. J. 255, 105-1 1 I . [79] Mouricout, M.A. and Julien, R.A. (1987) Infect. Immun. 55, 1216-1223. [80] Lindahl, M., Brossmer, R. and Wadstrom, T. (1987) Glycoconj. J. 4, 51-58. [81] Teneberg, S., Willemsen, P., DeGraaf, F.K. and Karlsson, K.A. (1990) FEBS Lett. 263, 10-14. [82] Ono, E., Abe, K., Nakazawa, M. and Naiki, M. (1989) Infect. Immun. 57, 907-91 I . [83] Liukkonen, J., Haataja, S., Tikkanen, K., Kelm, S. and Finne, J. (1992) J. Biol. Chem. 267, 21 10521111. [84] Murray, P.A., Prakobphol, A., Lee, T., Hoover, C.I. and Fisher, S.J. (1992) Infect. Immun. 60, 31-38. [85] Evans, D.G., Evans Jr., D.J., Moulds, J.J. and Grahm, D.Y. (1988) Infect. Immun. 56, 2896-2906. [86] Evans, D.G., Evans Jr., D.J., and Graham, D.Y. (1989) Infect. Immun. 57, 2272-2278. [87] Saitoh, T., Natomi, H., Zhao, W.L., Okuzumi, K., Sugano, K., Iwamori, M. and Nagai, Y. (1991) FEBS Lett. 282, 385-387.

504 [88] Slomiany, B.L., Piotrkowski, K., Samanta, A., VanHorn, K., Murty, VL.N. and Slomiany, A. (1989). Biochem. Int. 19, 929-936. [89] Evans, D.G., Karjalainen, T.K., Evans Jr., D.J., Graham, D. and Lee, C.-H. (1993) J. Bacteriol. 175, 674-683. [90] Razin, S. and Jacobs, E. (1993) J. Genet. Microbiol. 138, 407-4022. [91] Loomes, L.M., Uemura, K., Childs, R.A., Paulson, J.C., Rogers, G.N., Scudder Jr., P.R., Michalski, J.-C., Hounsell, E.F., Taylor-Robinson, D. and Feizi, T. (1984) Nature 307, 560-563. [92] Loomes, L.M., Uemura, K. and Feizi, T. (1985) Infect. Immun. 47, 15-20. [93] Wenneris, C., Neeser, J.-R. and Svennerholm, A.M. (1995) Infect. Immun. 63, 640-646. [94] Willemsen, P.T. and de Graaf, F.K. (1992) Microbiol. Pathogenet. 12, 367-375. [95] Gilboa-Garber, N. and Garber, N. (1992) In: H.J. Allen and E.C. Kisailus (Eds.), Glycoconjugates: Composition, Structure and Function. Marcel Dekker, New York, pp. 541-591. [96] Garber, N., Guempel, U., Belz, A., Gilboa-Garber, N. and Doyle, R.J. (1992) Biochim. Biophys. Acta 1 116, 331-333. [97] Avichezer, D., Katcoff, D.J., Garber, N.C. and Gilboa-Garber, N. (1992) J. Biol. Chem. 267, 2302323027. [98] Gilboa-Garber, N. and Mizrahi, L. (1985) Experientia 41, 681-682. [99] Glick, J. and Garber, N. (1983) J. Genet. Microbiol. 129, 3085-3090. [loo] Lamblin, G., Lhermitte, M., Klein, A,, Houdre, N., Scharfman, A,, Ramphal, R. and Roussel, P. (1991) Am. Rev. Respirat. Dis. 144, S19-S24. [ I O I ] Ramphal, R., Carnoy, C., Fievre, S., Michalski, J.-C., Houdret, N., Lamblin, G., Strecker, G. and Roussel, P. (1991) Infect. Immun. 59, 700-704. [I021 Cisar, J.O. (1986) In: D. Mirelman (Ed.), Microbial Lectins and Agglutinins. Wiley, New York, pp. 183196. [102a] Kolenbrander, P.E. (1991) In: M. Dworkin (Ed.), Microbial Cell-Cell Interactions. Am. Soc. Microbiol., Washington, pp. 303-329. [I031 Brennan, M.J., Cisar, J.O., Vatter, A.E. and Sandberg, A.L. (1984) Infect. Immun. 46, 459-464. [I041 Sandberg, A.L., Ruhl, S., Joralmon, R.A., Brennan, M.J., Sutphin, M.J. and Cisar, J.O. (1995) Infect. Immun. 63, 2625-263 1. [I051 London, J., Hand, A.R., Weiss, E.I. and Allen, J. (1989) Infect. Immun. 57, 3940-3944. [I061 London, J. and Allen, J. (1990) J. Bacteriol. 172, 2527-2534. [lo71 Ho, S.C., Schindler, M. and Wang, J.L. (1990) J. Cell Biol. 111, 1639-1643. [I081 Loh, J.T., Ho, S.C., De Freijter, A.W., Wang, J.L. and Schindler, M. (1993) Proc. Natl. Acad. Sci. USA 90, 3033-3037. [lo91 Boren, T., Falk, P., Roth, K.A., Larson, G. and Normark, S. (1993) Science 262, 1892-1895. [I 101 Falk, P., Roth, K.A., Bortn, T., Westhlom, T.U., Gordon, J.1. and Normark, S. (1993) Proc. Natl. Acad. Sci. USA 90, 2035-2039. [ I l l ] Falk, P.G., Bry, L., Holgersson, Gordon J.I. (1995) Proc. Natl. Acad. Sci. USA 92, 1515-1519. [I 121 Garber, N., Guempel, U., Gilboa-Garber, N. and Doyle, R.J. (1987) FEMS Microbiol. Lett. 48, 331334. [I 131 Finkelstein, R.A. and Hanne, L.F. (1982) Infect. Immun. 36, 1199-1208. [I 141 Wisniewski, J.P., Monsigny, M. and Delmotte, EM. (1994) Biochimie 76, 121-128. [I 151 Depierreux, C., Kang, H.-C., Guerin, B., Monsigny, M. and Delmotte, F. (1990) Glycobiology I , 643649. [I 161 van’t Wout, J., Burnette, W.N., Mar, V, Rozdzinski, E., Wright, S.D. and Tuomanen, E.I. (1992) Infect. Immun. 60, 3303-3308. [I171 Heerze, L., Chong, P. and Armstrong, G. (1993) J. Biol. Chem. 267, 25810-25815. [I 181 Springer, T.A. (1990) Nature 346, 425. [I 191 Drickamer, K. and Taylor, M.E. (1993) Annu. Rev. Cell Biol. 9, 237-264. [I201 Saukkonen, K., Burnette, W.N., Mar, V, Masure, H.R. and Tuomanen, E. (1992) Proc. Natl. Acad. Sci. USA 89, 118-122. [I211 Lobet, Y., Feron, C., Dequesne, G., Simoen, E., Hauser, P. and Locht, C. (1993) J. Exp. Med. 177, 79-87.

505

[I221 Loosmore, S., Zealey, G., Cockle, S., Boux, H., Chong, F’., Jacoob, R. and Klein, M. (1993) Infect. Immun. 61, 23 16-2324. [I231 Rozdzinski, E., Burnette, W.N., Jones, T., Mar, V. and Tuomanen, E. (1993) J. Exp. Med. 178,917-924. [I241 Stein, P., Boodhoo, A., Amstrong, G., Cockle, S., Klein, M. and Read, R. (1994) Structure 2, 45-57. [I251 Stein, P.E., Boodhoo, A., Armstrong, G.D., Heerze, L.D., Cockle, S.A., Klein, M.H. and Read, R.J. (1994) Nature Struct. Biol. I , 591-596. [I261 Armstrong, G.D., Howard, L.A. and Peppler, M.S. (1988) J. Biol. Chem. 263, 8677-8684. [I271 Witvliet, M.H., Burns, D.L., Brennan, M.J., Poolman, J.T. and Manclark, C.R. (1989) Infect. Immun. 57, 3324-3330. [I281 Prasad, S.M., Yin, Y., Rodzinski, E., Tuomanen, E.I. and Masure, H.R. (1993) Infect. Immun. 61, 2780-2785. [I291 Barak, R., Elad, Y. and Chet, I. (1986) Arch. Microbiol. 144, 346-349. [I301 Kellens, J.T.C. and Peumans, W.J. (1991) Mycol. Res. 95, 1235-1241. [I311 Kellens, J.T.C., Goldstein, I.J. and Peumans, W.J. (1992) Mycol. Res. 96, 495-502. [I321 Inbar, J. and Chet, 1. (1994) Microbiology 140, 651-657. [ 1331 Borrebaeck, C.A.K., Mattiasson, B. and Nordbrink-Hertz, B. (1984) J. Bacteriol. 159, 53-56. [I341 Rosen, S., Ek, B., Rask, L. and Tunlid, A. (1992) J. Genet. Microbiol. 138, 2663-2672. [I351 Ravdin, J.I. and Guerrant, R.L. (1981) J. Clin. Invest. 68, 1305-1313. [I361 Petri Jr., W.A. (1991) ASM News 57, 299-306. [I371 Tannich, E., Ebert, F. and Horstmann, R.D. (1991) Proc. Natl. Acad. Sci. USA 88, 1849-1853. [I381 Mann, B.J., Vedvick, T., Torian, B. and Petri Jr., W.A. (1991) Proc. Natl. Acad. Sci. USA 88, 32483252. [139] McCoy, J.J., Mann, B.J., Vedvick, T.S., Pak, Y., Heimark, D.B. and Petri Jr., W.A. (1993) J. Biol. Chem. 268, 24223-2423 1. [I401 Petri Jr., W.A., Snodgrass, T.L., Jackson, T.F.H.G., Gathiram, V., Simjee, A.E., Chadee, K. and Chapman, M.D. (1990) J. Immunol. 144, 4803-4809. [I411 Mann, B.J., Chung, C.Y., Ashley, J.M., Braga, L.L. and Snodgrass, T.T. (1993) Infect. Immun. 61, 1772-1778. [I421 Li, E., Becker, A. and Stanley Jr., S.L. (1998) J. Exp. Med. 167, 1725-1730. [I431 Chadee, K., Petri Jr., W.A., Innes, D.J. and Ravdin, J.I. (1987) J. Clin. Invest. 80, 1245-1254. [I441 Adler, P.,Wood, S.J., Lee, Y.C., Lee, R.T., Petri Jr., W.A. and Schnaar, R.L. (1995) J. Biol. Chem. 270, 5 164-5 17 1. [I451 Breuer, W.V., Ginsburg, H. and Cabantchik, Z.1. (1983) Biochim. Biophys. Acta 755, 263-271. [I461 Perkins, M.E. and Rocco, L.J. (1988) J. Immunol. 141, 3190-3196. [I471 Orlandi, PA., Klotz, E.W. and Haynes, J.D. (1992) J. Cell Biol. 116, 901-909. [I481 Klotz, F.W., Orlandi, P.A., Reuter, G., Cohen, S.J., Haynes, J.D., Schauer, R., Howard, R.J., Pelese, P. and Miller, R.H. (1992) Mol. Biochem. Parasitol. 51, 49-54. [I491 Sim, B.K.L., Orlandi, P.A., Haynes, J.D., Klotz, E.W., Carter, J.M., Camus, D., Zegans, M.E. and Chulay, J.D. (1990) J. Cell Biol. 11 1, 1877-1884. [I501 Sim, B.K.L., Carter, J.M., Deal, C.D., Holland, C., Haynes, J.D. and Gross, M. (1994) Exp. Parasitol. 78, 259-268. [I511 Sim, B.K.L., Chitnis, C.E., Wasniowska, K., Hadley, T.J. and Miller, L.H. (1994) Science 264, 19411944. [I521 Kobiler, D. and Mirelman, D. (1980) Infect. Immun. 29, 221-225. [153] Rosales-Encina, J.L., Meza, I., Lopez-De-Leon, A,, Talamas-Rohana, P. and Rojkind, M. (1987) J. Infect. Dis. 156, 790-797. [I541 El Moudni, B., Philippe, M., Monsigny, M. and Schrevel, J. (1993) Glycobiology 3, 305-312. [I551 Ward, H.D., Keusch, G.T. and Pereira. M.E.A. (1990) BioEssays 12, 211-215. [I561 KO, H.L., Beuth, J., Solter, J., Schroten, H., Uhlenbruck, G. and Pulverer, G. (1987) Infection 15, 237-240. [I571 Saukkonen, K., Cabellos, C., Burroughs, M., Prasad, S. and Tuomanen, E. (1991) J. Exp. Med. 173, 1143-1 149. [I581 Mouricout, M., Petit, J.M., Carias, J.R. and Julien, R. (1990) Infect. Immun. 58, 98-106.

506 [I591 Steuer, M.K., Herbst, H., Beuth, J., Steuer, M., Pulverer, G. and Matthias, R. (1993) Otorhinolaryngol. Nova 3, 18-25. [I601 Wentworth, J.S., Austin, F.E., Garber, N., Gilboa-Garber, N., Paterson, C.A. and Doyle, R.J. (1991) Biofouling 4, 99-104. [161] Blaylock, W.K., Yue, B.Y. and Robin, J.B. (1990) CLAO J. 16, 223-227. [I621 Kolenbrander, P.E., Ganeshkumar, N., Cassels, F.J. and Hughes, C.V. (1993) FASEB J. 7, 4 0 6 4 1 3 . [I631 Manocha, M.S. and Chet, Y. (1990) Can. J. Microbiol. 36, 69-76. [164] Petri Jr., W.A. and Ravdin, J.I. (1991) Infect. Immun. 59, 97-101. [I651 Soong, C.J.G., Kain, K.C., Abd-Alla, M., Jackson, F.H.G. and Ravdin, J.I. (1995) J. Infect. Dis. 171, 645-65 1. [166] Braga, L.L., Ninimiya, H., McCoy, J.J., Eacker, S., Wiedmer, T., Pham, C., Wood, S., Sims, P.J. and Petri Jr., W.A. (1992) J. Clin Invest. 90, 1131-1 137. [I671 Mirelrnan, D. (1987). Microbiol. Rev. 51, 272-284. [I581 Ofek, I. and Sharon, N. (1988) Infect. Immun. 56, 539-547. [I691 Ofek, I., Rest, R.F. and Sharon, N. (1992) ASM News 58, 429435. [I701 Gbarah, A,, Mirelman, D., Sansonetti, P.J., Verdon, R., Bernhard, W. and Sharon, N. (1993) Infect. Immun. 61, 1687-1693. [I711 Sandberg, A.L., Mudrick, L.L., Cisar, J.O., Metcalf, J.A. and Malech, H.L. (1988) Infect. Immun. 56, 267-269. [I721 Bernhard, W., Gbarah, A. and Sharon, N. (1992) J. Leukoc. Biol. 52, 343-348.

J. Montreuil, J.F.G. Vliegenthart and H . Schachter (Eds.), Glycoproteins II

0 1997 Elsevier Science B.V. All rights reserved CHAPTER 14

Adhesive glycoproteins and receptors R. Colin Hughes National Institute f o r Medical Research, The Ridgeway, Mill Hill, London NW7 I A A , U K .

1. Introduction The adhesive interactions that attract different types of cells to each other and to the extracellular matrix play critical roles in a wide variety of developmental and pathological events. It has been known for many years that cells dispersed from different embryonic tissues such as liver and heart can reform tissue-like structures when incubated in an appropriate medium or can sort out when cells from different tissues are mixed (Fig. la). These properties were first established in primitive multi-cellular organisms such as marine sponges by Wilson [ I ] and in mammalian embryonic tissues by Holfreter [2] and have been amply confirmed since. In developed multi-cellular organisms the structure and organization of the various tissues depends on the maintenance of adhesions between neighbouring cells and between cells and the extracellular matrix [3]. For example, in a relatively simple tissue such as the blood vessel wall an endothelial cell layer is firmly attached to a basement membrane, and the underlying stromal tissue forms a medium for migratory cells such as fibroblasts (Fig. 1b). This structure involves a variety of cell-cell and cell-matrix adhesive events. In simple epithelia, individual cells are firmly adherent to the basement membrane and to each other. There is considerable evidence that the surface of each epithelial cell is differentiated into domains with distinct functional and chemical composition, and distinguished by unique adhesion structures (Fig. Ic). The free, apical surface domain is primarily involved in regulation of nutrient uptake, and hence is enriched in uptake systems such as ion channels, and in protection of mucosal surfaces. This domain is also relatively enriched in glycolipids and in glycosyl-phosphatidyl-inositol-anchoredmembrane glycoproteins. Indeed, the glycosylphosphatidyl inositol (GPI) anchor has been suggested to be a contributing signal for the secretion of proteins to the apical domain of polarized epithelial cells [4]. The apical surface domain of polarized epithelial cells is considered to be non-adhesive, although it is an important site for the initial adhesive interactions of invasive micro-organisms during infection. The lateral surface domain of polarized epithelial cells is the site of cell-cell contact and communication and is distinguished by several types of junctional complexes. Tight junctions are the most apically disposed and functionally are crucial in maintaining a polarized state by prevention of the movement of proteins from the apical domain into lateral or basal domains. They also act as a permeability barrier between the functionally distinct apical and baso-lateral domains. Gap junctions are almost ubiquitously expressed in cells and their main function is to facilitate chemical communication between adjacent cells by forming channels for the exchange of low molecular mass metabolites. The two remaining junctional types (Fig. 1c) are the desmosomes and the intermediate or adherens junctions. These junctions are primarily responsible for initiating cell-cell contacts with 507

Tissues

Dissociation

Aggregation

f Basement membrane

Apical membrane

c) Xght junctions

+ Desmosomes + Gap junctions +

Lateral membrane

Adherens junctions

Basal membrane

f Basement membrane laminins wllagen’t e IV, proteoglyXns

Migration

t

Attachment

Spreading

Matrix glycoproteins fohaaens, lib onectin. am1 n Fig. 1. Cell-cell and cell-matrix interactions.

the specificity expected from cell sorting phenomena (Fig. la) and reinforcing these specific adhesive events, and are therefore of special relevance to the subject of this chapter. The basal membrane of polarized cells (Fig. lc) contains the molecules involved in binding to components of basement membranes. These adhesive contacts are mediated

509

by interactions between specific receptors and the extracellular matrix components characteristic of basement membranes, namely collagens especially type IV collagen, the glycoproteins laminin and fibronectin and the proteoglycans especially heparan sulfate proteoglycans. The adhesive contacts of the basal surface such as hemidesmosomes (not to be confused with desmosomes, as shown by a distinctly different chemical composition [5,6]) and focal contacts, are morphologically distinct features where the space between the cell surface and the underlying substratum is decreased compared with less adhesive areas. Basal cell contacts with the substratum is a key feature for the maintenance not only of epithelial organization but for the growth and motility of many cell types. In the context of understanding the organization of epithelia there is a lively debate on the ordering of specific adhesive events during the establishment of the polarized structure. Some argue that the initial contacts of individual cells with a substratum is crucial in stabilizing the cell monolayer, after which other types of adhesions and surface regionalization takes place to produce a fully polarized epithelium. Others suggest that initial cell-cell contacts precede firm attachment of adherent cells to the substratum. Probably, both mechanisms occur in different systems. In any event, the adhesive contacts between cells and a substratum are not confined to those discussed above for polarized epithelial cells, and even highly motile cells such as fibroblasts are in intimate association with matrix elements as shown in Fig. 1b. The widespread ability of various cell types to form cell-substratum adhesions is readily demonstrated in culture. When cells are plated onto an appropriate substratum formed of extracellular matrix components such as fibronectin or laminin the cells attach very rapidly by interactions of matrix receptors on the cell surface and eventually reorganize to adopt a characteristic spread morphology by recruitment of the specific cell surface receptors on the underside, ventral surface of the spread cell to form close and focal adhesive contacts (Fig. Id). Such contacts are crucial for the normal growth of many cell types. Gene expression in particular is often very sensitive to extracellular signals mediated through cellular interactions with a defined matrix acting as an adhesive substratum [7,8]. Locomotion of cells also depends critically on cell-substratum interactions and is involved in many important events during normal development, as in cell movements during gastrulation and in the segregation and dissemination of neural crest cells, in wound healing and in bone remodelling [9]. It is to be noted that the motile cell, approximated to by culture models of cell-substratum adhesion (Fig. Id), has a polarity every bit as complex as that of a fully polarized cell in an established epithelium. The most conspicuous feature of the polarized motile cell is the leading edge, a thin vein-like structure free of cytoplasmic organelles, that extends from the cell body in the direction of locomotion. The ventral surface in contact with the substratum is distinct structurally and morphologically from the upper, dorsal surface that becomes denuded of adhesive molecules. In the last ten years or so there have been major advances in understanding many of the adhesive interactions described above and in identifying the cell adhesion molecules involved. Much of this progress has been due to the production of specific antibodies able to perturb cell-cell or cell-substratum contacts in functional adhesion assays. The antibodies, prepared against membrane fractions conventionally or by hybridoma technologies, proved to be invaluable reagents for the isolation of functional adhesion

510

molecules and importantly for their cloning in expression libraries. More recently, with the increasing evidence for homologous families of cell adhesion molecules, cross-hybridization and polymerase chain reaction (PCR) methods using degenerate oligonucleotide primers have revealed many more members. These advances are leading to a satisfying synthesis of the mechanisms and molecules involved in specific adhesive events and their regulation. Common features are emerging and a rational classification of cell adhesion molecules, responsible for mediating different types of adhesive interactions, is now possible. The cadherins are a family of calcium-dependent cell-cell adhesion molecules that are widely expressed in epithelia of higher vertebrates. Analogous members of the cadherin family are also being discovered in other organisms such as Drosophila and C. elegans which offer significant advantages for the genetic dissection of cell adhesion mechanisms. Another major class of cell adhesion molecules belongs to the immunoglobulin superfamily and is typified by N-CAM, the first thoroughly characterized of cell adhesion molecules. The N-CAM family mediates calcium-independent cell-cell adhesive interactions. Finally, the integrins are cell surface receptors that mainly but not exclusively mediate cell interactions with extracellular matrix components in a cationdependent manner. In parallel with progress in characterization of these cell surface adhesion molecules, the biochemistry of the extracellular matrix has also advanced in understanding over the last few years. In particular, the isolation, cloning and structural analysis of large, modular glycoproteins of the extracellular matrix has progressed rapidly and has obvious relevance to the present discussion. Other major components of the extracellular matrix, including the collagens and the proteoglycans are described elsewhere [ 10,111. In this chapter I shall focus on the cell adhesion molecules and matrix components as glycoproteins. In particular the glycosylation of these components and its significance for biological function will be considered. Evidence that carbohydrate recognition does play an important role in adhesive events including leukocyte adhesion and in microbial invasion is summarized elsewhere [12-141. In the specific context of the cell adhesion mechanisms to be discussed here, however, there is little evidence to suggest a direct role of carbohydrate. Nonetheless, glycosylation of the cell adhesion molecules under discussion has been shown in several cases to be an important modulator of function.

2. Cadherin family of adhesive glycoproteins 2. I. Classical cadherins

The classical cadherins are integral membrane glycoproteins that mediate calciumdependent adhesion of many cell types [15,16]. The first member of this family to be recognized was E-cadherin in mouse teratocarcinoma cells and other epithelial cells and also called uvomorulin or cell CAM 120/80. Monovalent antibody fragments from a polyclonal serum raised against F9 teratocarcinoma cells were shown to block the calcium-dependent aggregation of these cells in uitro. The antibody was then used to identify a -120 kDa glycoprotein, which subsequently was cloned, sequenced

-

51 1

HAV \site,

N-cadherin

NH2

E-cadherin

NH2

f-cadherin

NH2

T????

???

I E2 I E3 I E4 I E5

''

O 'H

COOH

Fig. 2 . Classical cadherins. The extracellular domain comprises a propeptide (solid) which is removed in the active protein, four segments showing internal homologies ( E l E4), an extra domain E5, a transmembrane domain (dashed) and a cytoplasmic domain (CD). Putative glycosylation sites are indicated by open circles. The primary adhesion determinant HAV is indicated.

-+-COOH

and established as E-cadherin. Similar experiments were performed earlier by Kemler who found that an antibody raised against F9 cells blocked compaction of early mouse embryos. Further experiments showed that the antibody recognized a -120 kDa glycoprotein which was called uvomorulin. The chicken equivalent of E-cadherin, called L-CAM, was identified in chicken hepatocytes by an antibody capable of blocking the calcium-dependent aggregation of these cells [ 171. N-cadherin was originally identified in mouse and chicken brain and proved to be identical to adhesion molecules of neural and mesodermal origin called A-CAM and N-CAL CAM. P-cadherin was identified in placenta of mouse or human origin. The early work indicated that these cadherins are tissue-specific and are immunologically distinct [ 18,191. Subsequent cloning and sequencing studies showed however, that all of these glycoproteins share significant structural similarities and sequence homologies as indicated in Fig. 2 . Comparison of the primary structure of E-cadherin with that of P-cadherin for example shows 58% aminoacid homology. Similarly the homology between mouse E-cadherin and chicken L-CAM is about 65%. The structure of the classical cadherins typified by E-, P- and N-cadherins (Fig. 2) comprises a pro-peptide domain that is removed by limited proteolysis during activation, four extracellular domains that show significant internal homologies especially between the first two domains El and E2, a connecting segment E5, a transmembrane domain responsible for integration of the glycoproteins into cellular membranes and a cytoplasmic tail. Each of the homologous domains E L E 4 contains two putative calcium-binding sites. A synthetic peptide based on one of these sites has been shown directly to bind calcium, a property that is presumed to be important in maintaining the adhesive activity of the cadherins [20]. Site-directed mutagenesis in just one of the eight putative cation binding sites of E-cadherin induces a sensitivity to proteolysis and inactivation of adhesive activity [20]. The dominant site of adhesive activity has been located to domain E l , in which an amino-acid sequence His-Ala-Val (HAV) appears to be crucial. Monoclonal antibodies that block cadherin-mediated adhesions recognize epitopes that map to this or adjacent sequences and site-directed mutagenesis in or around this sequence also abolish adhesive activity [21]. Interestingly, although the HAV motif is found in E-, P- and N-cadherin, the immediately adjacent sequences do differ, suggesting that an extended recognition sequence in this region of each molecule is responsible for the specificity of adhesive

512

Extracellular space

Cell B

Cell A

I

y-catenin

AdhesJve interactions

3B Adhesjve inferacttons

Plasma membrane

Fig. 3. Models of cadherin-mediated cell adhesion

interactions of individual cadherins as discussed later. The role of the extra E5 domain is less clear but some adhesion blocking monoclonal antibodies recognize epitopes that map in this region [22]. The cytoplasmic domain of the classical cadherins show the highest degree of sequence homology, suggesting that this domain plays a vital role in mediating cadherin function. Intercellular adhesion mediated by the cadherins is homotypic, that is to say apposing cells expressing the same cadherin interact through interactions between the extracellular domains of interacting cadherins. There are two possible mechanisms (Fig. 3); the El domains of cadherin molecules on apposing cells may act directly with one another or each El domain may contact an internal E5 domain. Whatever the mechanism, it is clear that these initial contacts precipitate a cascade of events which lead to a fully established junctional complex. The highly conserved cytoplasmic tail of the cadherins associate with the intracellular proteins a-, 8- and y-catenins [23-251. In pulse-chase experiments it has been shown that the precursor cadherin molecule is already complexed with one or possibly two (3-catenin molecules which may be important for targeting of the cadherin-catenin complex to potential junctional sites. After the initial contacts the cytoplasmic tails of the cadherins become associated with a-catenin and y-catenin and initiate assembly of actin filaments, which stabilize the junction. a-Catenin has significant sequence homology with vinculin [24], a protein that is localized in other adhesive specializations such as focal contacts involved in the coupling of extracellular matrix receptors to actin-based cytoskeletal elements. It is reasonable to assume therefore that association of a-catenin with the cytoplasmic tail of contacting cadherin molecules is a key event in actin polymerization and formation of the mature junctional complex. Cadherin-mediated interactions appear to be the dominant factors in the cell-specific sorting shown in Fig. la. The clearest evidence for such a role comes from transfection experiments [26]. Mouse L fibroblasts lack cadherins. Transfection of L cells with fulllength cDNA clones of E-cadherin induce an efficient, calcium-dependent adhesion in the cells, which morphologically become epithelioid. Furthermore, when L cells are transfected separately with cDNA clones of E-cadherin or P-cadherin and mixed, the

513

El

E2

E3

E4

E5

cells sort out into aggregates expressing one but not both of the cadherins. Interestingly, cells transfected with chicken L-CAM cDNA, as well as showing increased cell-cell adhesivity, also increased their content of desmosomal and gap junctions, indicating that cadherin-mediated contacts are the crucial initial events in establishing a fully polarized and communicating epithelium [27]. As indicated in Fig. 2 the cadherins are glycoproteins containing several potential N-glycosylation sites. There are four putative glycosylation sites in E- and P-cadherin and seven in N-cadherin, only a few of which are conserved. Tunicamycin treatment of cells resulted in expression of P- or E-cadherin molecules reduced in size by 6-8kDa, indicating that 3-4 of the N-glycosylation sites are utilized. Tunicamycin-treated F9 cells expressing unglycosylated forms of E-cadherin aggregated in a calcium-dependent manner, indicating that surface expression and function of the cadherin do not require N-glycosylation [28]. Similarly, chick myoblasts treated with tunicamycin were found to aggregate in a calcium-dependent manner [29]. Interestingly, however, the tunicamycintreated F9 cells appeared to accumulate a high molecular mass (-130kDa) form of E-cadherin which did not appear at the cell surface and may represent the precursor cadherin which had not undergone proteolytic processing [28]. Overall surface expression of E-cadherin was reduced in the tunicamycin-treated cells suggesting that glycosylation may be important in the efficient processing of cadherin precursors and transport to the cell surface, perhaps by modulation of interactions with (3-catenin. Other evidence that specific glycosylation may be important in cadherin function is available. In the chick retina N-cadherin appears to be associated with a cell surface N-acetylglucosaminyl phosphotransferase [30,3 I]. Antibodies directed against the transferase blocked completely N-cadherin-mediated adhesion of retinal cells and pancreatic islet cells indicating that this association, and presumably a special glycosylation event involving N-cadherin, was important in adhesive function [32,33]. The transferase is reported to be present in many tissues in association with N-cadherin, especially at neuromuscular junctions [34]. Characterization of the transferase and its action on N-cadherin will be of great interest in considering the role of glycosylation in cadherin function. Recently, several novel members of the cadherin family have been discovered; M-cadherin in mouse embryonic skeletal muscle [35,36], R-cadherin in neuronal cells [37] and T-cadherin in myoblasts [38]. The structure of T-cadherin (Fig. 4) is of special interest. This cadherin in its mature form lacks the transmembrane and cytoplasmic domains of classical cadherins and is integrated into the membrane by a GPI anchor. Although T-cadherin is a severely truncated form and additionally lacks an HAV sequence in the El domain it appears to be functional. Transfection of cDNA encoding T-cadherin into a cell devoid of cadherin expression confers calcium-dependent aggregation on the cells [38]. The mechanism of T-cadherin adhesion is intriguing and presumably is independent of interactions with catenins and direct coupling with actin filaments.

514

2.2. Desmosomal glycoproteins Desmosomes are punctate intercellular junctions that occur between the apposed lateral membranes of polarized epithelial cells (Fig. lc). Desmosomes together with the actincoupled adherence junctions discussed in the previous section play an important role in binding cells together [5,6,39-41]. However, the link between epithelial polarity and desmosomal junctions is not as well established as for the adherence junctions. Possibly the formation of desmosomal junctions, although an important event, is secondary to formation of adherence junctions involving classical cadherins. However, reduction of desmosomes has been described in some epithelial tumors and correlated with transformation and tumorigenicity of the cells [42,43]. Perhaps the best way of looking at desmosomal junctions is as a form of “spot-welding’’ of the contacting membranes between epithelial cells that reinforce initial contacts formed through junctions involving classical cadherins, and helping in formation and stabilization of the polarized structure. The desmosomal junction is a morphological entity that is very distinct from adherence junctions involving classical cadherins [44]. The plasma membrane domains of apposing cells are separated by a 20-30nm thick layer, which in cross section in electron micrographs reveals a midline structure and electron-dense threads stretching laterally from the midline back to the plasma membrane. The cytoplasmic face of each membrane domain is covered with an electron-dense plaque to which bundles of intermediate filaments attach, rather than the actin filaments present in adherence junctions containing classical cadherins. In most epithelia the intermediate filaments are based on keratins but rarer examples of desmosomal junctions coupled into intermediate filaments of the vimentin or desmin type are known [5]. The structure of desmosomal junctions has been intensively studied in stratified epithelial tissues such as bovine muzzle epithelium and human epidermal tissue [5]. Biochemical analyses have identified two major families of glycoproteins, the desmoCollins and the desmogleins. In addition, there are four major non-glycosylated proteins: desmoplakins I and 11, plakoglobin which has homology with p-catenin of actin-coupled adherence junctions [45], and a 75 kDa protein which is commonly found in desmosomes of stratified epithelia but not in simple, single layered epithelia. The arrangement of these components in the desmosomal structure has been determined using specific antibodies and immunolocalization particularly by electron microscopy (Fig. 5). Recent cloning and sequencing data [39,4 11 have revealed first that the desmosomal glycoproteins have extensive homologies with the classical cadherins. Second, they have revealed an unexpected heterogeneity in that there appears to be at least three desmocollin genes DSC 1, 2 and 3 which each encode two glycoprotein products arising from alternative RNA splicing and differing by about 6 kDa. Homology identities between the desmocollins of different genes are considerably less than the homologies shown between the classical cadherin family, showing that these desmocollins represent separate gene families. However, all desmocollins share a similar structural organization with each other and with the classical cadherins. The structures of desmocollins Dsc3a and DsQb, the products of alternatively spliced transcript of the DSC3 gene are shown in Fig. 6. The extracellular portions contain four internal repeats (El-E4) that carry calcium-binding sites as in the classical cadherins, an extra domain (E5) adjacent to

515 Plasma membrane

Plaque

Intermediate lilaments

Fig. 5 . Organization of structural components of the desmosomal junction.

the plasma membrane-spanning domain and a cytoplasmic domain. Dsc3b differs from Dsc3a in the cytoplasmic portion, due to the inclusion in the transcript of an extra 46 base pair exon containing an in-frame stop codon, resulting in a truncated (-14 m a ) cytoplasmic domain, rather than a -20 kDa domain for Dsc3a. The homologies between the cytoplasmic domains of the desmocollins and the classical cadherins are relatively low, consistent with their presumptive coupling into different cytoskeletal elements. Additionally, sequence differences in this region generated through alternative transcript splicing of the desmocollins, which is not found in classical cadherins, indicates RAL site

0-glycans

n

FAT site

Dsc 3a (desmocollin) 1 I

Fig. 6 . Structures of desmosomal cadherins. Nomenclature from Buxton et al. [48]. For explanation of symbols see Fig. 2. Dsc3a and Dsc3b are the products of alternative RNA splicing of the DSC3 gene: see text for details.

516

the existence of other regulatory mechanisms in the attachment of the desmosomal glycoproteins to the cytoskeleton, the function of which remains to be elucidated. The desmocollins carry six putative N-glycosylation sites (Fig. 6) and the available evidence suggests that at least some of these are utilized. Desmocollins Dsc3a and Dsc3b isolated from cells treated with tunicamycin have apparent molecular sizes reduced by about 5-lOkDa from about 130 and 115kDa for the two products. This indicates the presence of two to four N-glycans of average structure [46,47]. Pulse-chase studies in cultured MDCK cells showed a time-dependent shift of the N-glycans to endo-H-resistance, indicating their conversion into complex-type structures. The presence of 0-glycans was not indicated since tunicamycin-treated cells did not incorporate any radioactivity from UDP-GlcNAc into either desmocollin. The pro-peptide domain of desmocollins Dsc3a and Dsc3b carry one putative N-glycosylation site and another is present in the E l domain adjacent to the site of partial proteolysis required for removal of the pro-peptide and formation of the mature form of the desmosomal glycoproteins. Proteolytic processing appears to be a late event in integration of the desmocollins into the desmosomal structures [47]. There is some suggestive evidence (see Fig. 6 in ref. [46]) to correlate glycosylation with the regulation of this proteolytic event since in tunicamycin-treated cells the amounts of mature Dsc3a and Dsc3b products appeared to be reduced compared with control cells supporting full N-glycosylation, and additional products of lower molecular mass but still reactive with specific antibodies were detected. The second family of desmosomal glycoproteins, the desmogleins, is also composed of multiple genes, DSG1, DSG2 and DSG3 [48]. The structure of the product, Dsgl, of one of these genes, DSG1, is shown in Fig. 6. Two unique features of this glycoprotein require comment. First, the cytoplasmic segment is much more extended than the cytoplasmic domains of either the desmocollin or classical cadherin families, containing at its most carboxyl-terminal end a 29 amino-acid residue sequence repeated six times followed by a glycine-rich sequence. A more membrane proximal domain rich in cysteine residues may play a crucial role in molecular interactions within the desmosomal plaque (see Fig. 5 ) . Secondly, the extra domain E5 of desmoglein Dsgl contains a serine-threonine enriched sequence which may be a preferred site for 0-glycosylation [49] as indicated by reactivity with the Aruchis hypogeu (peanut) lectin and by a shift in molecular size by about 5 kDa resulting in a loss of peanut lectin but not of concanavalin A staining after an alkaline (3-elimination reaction [46]. The identification of galactosamine in a desmoglein fraction [50] also suggests 0-glycosylation. The adhesive interactions involved in formation of desmosomal junctions presumably are mediated through the extracellular domains of the desmosomal glycoproteins, both the desmocollins and the desmogleins. Antibody fragments against the desmocollins are able to inhibit desmosome formation [Sl]. However, the details of these interactions are unknown. The E l domain of the desmocollins and desmogleins are most closely homologous to each other and to the classical cadherins, indicating that the desmosomal junction is formed by homotypic or possibly heterotypic, i.e. desmocollin to desmoglein interactions between these domains, as in the case of cadherin-type junctional interactions (Fig. 5 ) . The HAV sequence which forms part of the interactive sequence of classical cadherins is missing in desmosomal glycoproteins which contain instead the analogous sequence

517

Arg-Ala-Val (RAL) in desmoglein Dsgl and Phe-Ala-Thr (FAT) in desmocollins Dsc3a and Dsc3b. However, confirmatory evidence showing a role in adhesive interactions of these sequences will require adhesion blocking monoclonal antibodies recognizing these epitopes, and the study of transfectants expressing mutant desmosomal glycoproteins lacking these peptide determinants. Key questions for future study include the distribution and tissue specificity of expression of the so far identified desmosomal proteins and glycoproteins. Secondly, the distribution of individual desmosomal glycoproteins and proteins within a single desmosoma1 structure is unknown. Lastly, the function of glycosylation of the desmosomal glycoproteins in junction formation remains to be clarified. Some evidence for a role of correct glycosylation of desmosomal glycoproteins in desmosome formation comes from one study showing that the frequency of desmosomes in aggregating chick corneal cells was reduced by moderate amounts of tunicamycin that inhibited incorporation of mannose but not of leucine into the glycoproteins [52].In this case, protease inhibitors prevented the reduction in desmosome formation indicating that the function of glycosylation may be in accurate proteolytic processing and activation of the desmosomal glycoproteins. Clearly, more direct experiments, for example with non-adherent cells transfected with selected desmosomal glycoproteins under conditions where glycosylation patterns can be controlled, will be necessary to address these important questions.

3. Immunoglobulin superfamily of adhesive glycoproteins The Ig-superfamily contains many proteins involved in immune recognition such as products of the MHC complex and accessory molecules [53]. In addition there are ten or more members associated mainly with nervous tissues in mature animals and several others in non-nervous tissue that are important factors in cell-cell and cellsubstratum adhesion in non-immune cells. See [54] and [55] for detailed discussion of other aspects of Ig-superfamily glycoproteins. All of the cell adhesion glycoproteins in the family contain a variable number of Ig-like domains of about one hundred aminoacid residues, usually but not always defined within a pair of disulfide-bonded cysteine residues, and of the C2 type fold. In many cases the molecules contain variable numbers of another type of modular sequence known as the fibronectin type 111 repeat, since it was discovered in fibronectin. In the following discussion, some principles of the structure and functions of this large family of cell adhesion molecules will be considered with particular emphasis on the interplay between different members in adhesion and modulation of adhesive interactions by carbohydrates.

3. I . Neroe cells 3. I . I . N-CAM: the prototype adhesion molecule of the immunoglobulin superfamily N-CAM was first identified in chick embryo nervous tissue using an antibody that specifically inhibited the re-aggregation of dissociated nerve cells [56,57]. Subsequently, the antibodies were shown to alter the normal patterns of differentiation in the chick embryo, to inhibit the bundling of neurites into nerve fascicules and to prevent the

518 Exon

1,2 3,4

5,6

7,8 g,lO I 1 12 13 14

16 17

18

19

Fig. 7. Simplified structures of N-CAM isoforms. (a) The 180 kDa form of N-CAM encoded by exons 1-14 and 16-1 9 of the N-CAM gene. The extracellular portion contains five (I-V) C2 type immunoglobulin domains and two fibronectin type-III-like repeats. The transmembrane peptide sequence is encoded within exon 16 and the cytoplasmic tail continues with sequences from exons 17-19. Putative N-glycosylation sites are indicated and may be of conventional complex-type (solid circle) or bear polysialic acid glycans (openkolid circle). Two sites of alternative RNA splicing between exons 7,s and 12,13 are indicated (triangle). A hepann-binding site in the second Ig-like domain is indicated by the filled segment. (b) The I4OkDa form of N-CAM. The extracellular portion (dashed line) is similar to that described in (a) but the cytoplasmic domain lacks the coding sequence of exon 18 which is excised by RNA splicing. (c) The 120 kDa form of N-CAM. The extracellular portion (dashed line) with similar structural properties as isoforms shown in (a) and (b) utilizes a coding sequence in exon 15 which signals post-translational addition of a phospha1;dylinositol glycan for insertion into the membrane (wavy line). Other isoforms of N-CAM are discussed in the text.

enervation of muscle. Although originally found in nervous tissue, hence the name neural cell adhesion molecule (N-CAM), N-CAM is now known to be expressed in many tissues and cell types, especially in the embryo and appears early in development on derivatives of all three germ layers [58]. It is believed to play important roles in controlling cell aggregation and movement during embryogenesis. Inactivation of the N-CAM gene in mice by gene targeting offers further support for these conclusions [59]. Although somewhat surprisingly the homozygous mutant animals were found to be fertile and generally normal up to four months of age despite a complete loss of N-CAM immunoreactivity, the animals were smaller than controls, showed significant reduction in total brain size, particularly the olfactory bulb, and displayed deficiencies in spatial learning in maze-exploring tests. The olfactory bulb shows a high degree of synaptic modelling even in adults and the major effect of N-CAM inactivation on this structure is consistent with the view that N-CAM is critically involved in the formation and regeneration of neural connections. N-CAM is encoded by a single gene that gives rise to a large number of isoforms by alternative RNA splicing and by post-translation modifications including unique patterns of glycosylation[60]. The three major forms (Fig. 7) are similar in their extracellular

519

portion which contains five Ig-like domains and two fibronectin type 111 repeats. The largest 180kDa N-CAM polypeptide (the Id form, Fig. 7a) contains a conventional transmembrane segment which crosses the membrane, and a long cytoplasmic extension. The 140 kDa isoform (sd, Fig. 7b) lacks the cytoplasmic sequence encoded by exon 18 which is deleted by RNA splicing, and the smallest 120kDa isoform (ssd, Fig. 7c) is the product of an alternatively spliced transcript containing a new exon 15, the use of which eliminates exons 16-19 and introduces into the polypeptide a signal for post-translational modification by attachment of a GPI anchor for insertion into the membrane. Additional isoforms can be formed by insertion of a short exon n coding for ten amino-acid residues in the fourth Ig-like domain and more complex splicing events involving four short exons a-d expressed singly or in various combinations in the membrane proximal stem. The latter insertion leading to an additional 37 aminoacid sequence was originally thought to be muscle-specific [61] but is now believed to be more widely distributed. The regulation of these slicing choices is currently unknown although there is abundant evidence to show that expression of particular N-CAM isoforms is developmentally controlled. For example, the 180 kDa isoform is expressed preferentially in post-mitotic neurones in the central nervous system [62] whereas the exon-n-containing isoform predominates in postnatal brain development [63] and the GPI-anchored forms containing some or all of exons a-d express maximally during myogenesis [64,65]. All isoforms of N-CAM appear to be heavily glycosylated (Fig. 7a) as shown by significantly reduced sizes of the polypeptides obtained from tunicamycin-treated cells or after endo-glycosidase treatment. The presence of 0-glycans has not been described so far but it is intriguing that the extra sequence of 37 amino-acid residues introduced by utilization of exons a-d contains eleven residues each of proline and serinelthreonine [6 I] and hence is a good candidate for modification by 0-glycosylation. The most striking feature of N-CAM glycosylation, however, is the presence of polysialic acid on the N-glycosylation sites in or adjacent to the fifth Ig-like domain (Fig. 7a). Detailed discussion of polysialic acid is given in ref. [66]. In homozygous transgenic mice lacking N-CAM an 85% loss of total polysialic acid content was found in various tissues including brain [59] indicating that this type of glycosylation is largely unique for the N-CAM polypeptide although the polysialylation of minor glycoproteins cannot be ruled out. Unfortunately, little is known of the biosynthesis of polysialic acid except it is a late Golgi or post-Golgi event [67]. A single polysialic acid chain in N-CAM may contain up to 30 or more sialic acid residues but the exact chain length varies according to source of N-CAM [68,69]. Most strikingly, N-CAM in fetal or immature tissues is more heavily decorated with polysialic acid than in adult tissues, a difference of presumed functional significance in N-CAM-mediated adhesive interaction [70,7 11. Interestingly, polysialic acid appears to bind rather specifically to a homeobox sequence of the protein product (PAntP) of the antennapedia gene of Drosophila [72]. Structural comparisons between polysialic acid and double-stranded DNA suggest that a sequence of eight a2-8 linked sialic acid residues resembles one large groove of DNA. Confirmation of these intriguing findings will be of great interest, particularly since some homeobox genes are involved in specifying neuronal lineages or regulating neuronal development [73]. The possibility that a homeobox function, namely binding to promoter or enhancer

520

regions of target genes, may be affected in such a way by glycosylation of neuronally expressed proteins is intriguing. N-CAM mediates cell-cell adhesion in a calcium-independent manner. Early evidence suggested a homophilic mechanism in which N-CAM molecules on apposing cells interact directly. Thus, N-CAM incorporated into liposomes aggregated in a manner that was inhibited by N-CAM-specific antibodies. Similarly, mouse L cells that contain no endogenous N-CAM were found to aggregate after transfection with full-length N-CAM cDNA clones [57]. Direct binding studies with N-CAM protein in solid phase assays are consistent with this conclusion and indicate a relatively high homophilic binding affinity ( k , =6.9x lo-' M) for the adult form of N-CAM and a significantly lower value ( k , = 1.23x M) for newborn rat N-CAM [74]. Earlier studies had also shown that chick embryo cells expressing the fetal form of N-CAM aggregated more strongly after treatment to remove polysialic acid [75,77]. However, localization of the extracellular domain(s) involved in homophilic interactions of N-CAM has proved to be unexpectedly difficult. The original data [57,75] showed that a proteolytic fragment of N-CAM containing the first four Ig-like domains retained homophilic binding activity and more recently the epitope for a monoclonal antibody blocking homophilic binding has been mapped to the third Ig-like domain [78]. In a detailed study [79] fragments of N-CAM produced in a bacterial expression system and hence lacking carbohydrate, were tested for their capacity to support adhesion of neuronal cells. The product of constructs containing the amino-terminus and first two Ig-like domains were found to be best in supporting neuronal cell adhesion whereas the migration of neuronal cells on a laminin substratum required the Ig-like domains I, I11 and IV. Finally, an epitope in domain V recognized by a blocking monoclonal antibody appears to be required for an N-CAM function in adhesive interactions of chick retinal cells [SO]. Taken together these results indicate that the N-CAM molecules carry in the extracellular domain several sites that can mediate different types of adhesive interactions. Since such a complexity of adhesive interactions is unlikely to be explained solely by homophilic binding the possibility of other ligands is being actively considered. One heterophilic interaction between N-CAM and the L1 neuronal glycoprotein will be discussed in the next section and other evidence for heterophilic adhesion of embryonic retinal cells involving N-CAM has recently been demonstrated [S 1,821. However, the binding partners in these heterophilic interactions are unknown. One potential heterophilic binding site has been mapped to the second Ig-like domain, which binds to heparin, raising the possibility that interactions with heparin-related molecules, such as heparan sulfate proteoglycans expressed on cell surfaces or in the extracellular matrix, may mediate adhesive interactions involving N-CAM [83]. Peptides based on the heparin-binding sequence inhibit retinal cell adhesion to N-CAM [84]. It is not clear at present however whether the heparin-binding site on N-CAM truly mediates heterophilic adhesion, perhaps in synergy with homophilic interactions formed at other domains, or if heparin binding to N-CAM somehow modulates the homophilic interactions, for example by an induced conformational change in the molecule. In this latter model, the binding of heparan sulfate with N-CAM, expressed either on the same cell or on apposing cells, might induce a conformational charge to enhance homophilic interactions between contacting N-CAM molecules.

52 1

The mechanism by which polysialic acid affects N-CAM activity is also unknown. It is possible that the large hydrodynamic space occupied by these glycans in Ig-like domain V (Fig. 7a) blocks an adhesive site in this or other domains. Secondly, the presence of polysialic acid may prevent kinking of the N-CAM polypeptide at the fifth Ig-like domain observed in electron micrographs [85,86]. Polysialic acid substitution of N-CAM can regulate adhesions between cells mediated by other molecules, for example N-cadherin [87], suggesting that polysialylation of N-CAM may affect many N-CAM-dependent and -independent adhesive events [70,7 13. One way of explaining such a general effect takes into account that the bulky polysialic acid chain represents a major factor in regulating the width of the pericellular matrix surrounding cells and the formation of cell-cell contacts. Direct measurements of the intercellular space between gently pelleted cells expressing polysialylated N-CAM show that the average intercellular space is about 25% greater than in control cells expressing non-sialylated N-CAM [88]. Thus, treatment of cells carrying polysialylated N-CAM with endo-sialidase allowed significantly closer contacts to be made between apposing cells. Interestingly, removal of heparan- or chondroitin-proteoglycans, which also are major components of the pericellular space had no effect on the intercellular separation of the pelleted cells [88]. These findings could explain at least in part the broad biological effects of polysialic acid and are supported by other in uiuo evidence. During muscle development several isoforms of N-CAM appear in a developmentally regulated manner. Interestingly, while N-CAM was found to be distributed on all myotube surfaces, the polysialylated isoforms were restricted to regions of the myotube surface that had recently separated [87]. Other studies show that a very highly polysialylated N-CAM isoform is specifically expressed in the adult rat in a subset of neuronal precursors in the dentate granular layer of the hippocampus, an area in which synaptogenesis proceeds postnatally [89] and during motoneuron outgrowth and pathway guidance [90]. Another carbohydrate-mediated modulation of N-CAM adhesive activity has been described [91]. The adhesion of neuronally derived PC12 cells to monolayers of mouse 3T3 fibroblasts expressing transfected human or chick N-CAM was shown to be enhanced by pretreatment of the neuronal cells with ganglioside GMl. This effect appears to be secondary to the initial homophilic contacts between N-CAM molecules on apposing cells. 3.1.2. LI glycoprotein

The L1 glycoprotein was first described in the central nervous system of the mouse [92,93] where it is expressed on a subset of post-mitotic neurons whose exons form bundles (fasciculate) with other L1-positive axons [94]. In the peripheral nervous system L1 is also involved in neuronal cell interactions with glia and in neurite outgrowth. In the chick the L1 homologue is the neuronal-glial-cell adhesion molecule (Ng) [95]. L1 mediates calcium-independent adhesive interactions. The extracellular portions of L1 contains six Jg-like domains and five fibronectin type 111 repeats (Fig. 8a). The glycoprotein is inserted into the membrane by a hydrophobic spanning peptide segment and the short cytoplasmic segment shows sequence variation due to alternative RNA splicing [96]. The extracellular portion of L1 contains 20 potential N-glycosylation sites, many of which

522

Fig. 8. Structures of LI, MAG and PO glycoproteins. The sequences of LI (a), MAG (b) and PO (c) glycoproteins predicted from analysis of cDNA clones as well as protein in the case of PO are shown to variably include in the extracellular portions Ig-like domains, fibronectin type 111 repeats (hatched), a transmembrane segment and a cytoplasmic domain. Potential N-glycosylation sites are indicated (open circle). Splicing variants in the cytoplasmic domains of LI and MAG are shown (open triangle).

must be utilized to account for a high content (approximately 30%) of carbohydrate in the intact L1 glycoprotein. In earlier studies Ng-glycoprotein [95] was shown to mediate homophilic interactions. More recently, these interactions have been shown to be enhanced by N-CAM and the carbohydrate moieties of L1 appear to be involved in Ll-N-CAM interactions [97,98]. All neuronal cell types containing L1 also express N-CAM; L1 is predominantly expressed on axons whereas N-CAM is more uniformly distributed. In oitro, beads coated with L1 glycoprotein aggregated very inefficiently whereas beads coated with L1 and N-CAM aggregated well, either when incubated alone or in mixture with L1-coated beads. Interestingly, beads coated with L1 and N-CAM aggregated preferentially with beads containing only L1 rather than the L1- and N-CAM-coated beads [97,98]. Similar findings were obtained using cell aggregation assays with cell populations containing L1 only or both L1 and N-CAM [97,98]. These results suggest that although L1 can interact with itself to mediate cell-cell contacts, an Ll-N-CAM complex on one cell interacts far better as a counter receptor for L1 expressed on an apposing cell (Fig. 9). When the cells expressing L1 and N-CAM were treated with castanospermine to block early glucosidase processing of N-glycans, aggregation was severely reduced [98]. Swainsonine, a specific inhibitor of processing mannosidase 11, had a similar but less potent effect [98] suggesting that oligomannosidic but not hybrid- or complex-type glycans were most important in mediating the effect. Treatment of the L1-containing cells with castanospermine had no inhibitory effect on the aggregation of these cells when mixed with untreated cells containing L1 and N-CAM showing that the glycosylation of the adhesion molecules was important for complex formation between them in the

523 Lower affinity binding

Cell A

Higher affinity binding

e-

Membrane

domain

Cell B

Fig. 9. Adhesive interactions of the L1 glycoprotein. Homophilic interactions between L1 glycoproteins on apposing cells A and B form cell-cell contacts. Binding of Ll by N-CAM in the same membrane enhances the affinity of homophilic L1 interactions. L1-N-CAM binding may involve a glycan moiety (oligomannosidic-type) of L1 and a lectin activity within the fourth Ig-like domain of N-CAM. Adapted from Kadmon et al. [97,98].

same cell membrane. Inspection of the N-CAM sequence pointed to a weak but possibly significant homology in the fourth Ig-like domain to the carbohydrate recognition domain of C-type animal lectins and to some plant lectins [99]. In other assays, oligomannosidic glycans or glycopeptides purified from ribonuclease B blocked neurite outgrowth from transplants of early postnatal mouse cerebellum. Surprisingly, a synthetic peptide based on the putative C-type lectins-like domain also had significant inhibitory effects. These results, when confirmed by other studies are intriguing and suggest that carbohydrate-mediated interactions between L1 and N-CAM molecules can modulate their adhesive functions (Fig. 9). Some caution is needed however before concluding that the putative carbohydrate-lectin interaction affects directly hornophilic Ll-Ll contacts. A monoclonal antibody against one of the carbohydrate moieties of L1 blocks calcium channels activated by L1-mediated contacts [ 1001, which effectively modulates neurite extension [ 1011 and conceivably other adhesive interactions. 3.1.3. Myelin glycoproteins MAG and PO Myelin is a multi-layered membrane that surrounds conducting axons and acts as an insulation sheath. The two major myelin glycoproteins are MAG and PO, both of which are unique to the nervous system [ 1021. MAG glycoprotein mediates neuron-oligodendrocyte and oligodendrocyte-oligodendrocyte interactions, is produced by oligodendrocyte and Schwann cells and is induced after initial contacts of axons. Both glycoproteins, but especially PO which accounts for over half of the protein in peripheral myelin, stabilize the myelin structure. MAG has a molecular mass of -100kDa and contains about 30% carbohydrate. The extracellular portion (Fig. 8b) consists of five Ig-like domains [103,104] with about eight N-glycans almost entirely of the complex-type and mainly containing a bisecting N-acetylglucosamine residue and a core a l - 6 fucose residue [ 105,106]. The MAG glycoprotein exists in two forms generated by alternative splicing in the coding region of the cytoplasmic tail of the polypeptide [107]. The smaller S-form containing a truncated cytoplasmic tail is the major form in mature animals whereas the larger L-form appears earlier in development. When incorporated into liposomes MAG shows binding to cultured neuronal cells [ 1081 and specific antibodies to MAG inhibit adhesion of oligodendrocytes [109]. Adhesions involving MAG appear to be heterotypic but the

524

counter-receptors have not been identified. However, binding of MAG to heparin and various collagen isotypes has been detected [ 1101 and MAG has been implicated in collagen fibril formation [ 1 111. The PO glycoprotein is a small (28 kDa) glycoprotein unique in the Ig-superfamily in having a single variable (V) Ig-like domain (Fig. 8c) which carries one N-glycan [ 1 12, 1131. The structure of this glycan varies as a function of Schwann cell maturation [114]. In newborn rat nerve the glycan is predominantly of the endo-H-resistant complex-type whereas in adult animals it is mostly endo-H-sensitive indicating a developmentally regulated change to hybrid- or oligomannosidic-type structures. Podulso et al. [ 1151 have shown that when axonal contacts decrease through axonal degeneration after nerve transection, the glycosylation of PO changes from predominantly complex-type to oligomannosidic-type. Structural predictions suggest that the glycan fits into a cavity at the interface between the extracellular domain and the membrane, perhaps maintaining the orientation of PO at the membrane surface [116]. Interestingly, the glycan moiety of PO is required [ 117,1181 for the homophilic interactions shown to occur between PO molecules. Transfected CHO cells carrying PO mutants lacking the glycosylation site fail to aggregate although the mutant PO protein is at the cell surface [ 1 171. Similarly, PO-transfected CHO cells deficient in GlcNAc transferase 1 and hence unable to convert oligomannosidic glycans into hybrid- or complex-type chains fail to aggregate although transfected control cells do aggregate [I 171. It is not known if this requirement for hybrid- or complex-type oligosaccharides in PO adhesive function implies a direct involvement in the binding event or a function in maintaining an active conformation of the molecule. However, the glycopeptide of PO does block aggregation of PO-transfected cells [ 1 191 suggesting a direct effect. It is possible that a glycosylation switch in PO is a mechanism by which homophilic interactions are modulated and provides a mechanism by which the Schwann cell ensures that compact myelin is formed, maintained or degraded. 3.1.4. The L2/HNK-I carbohydrate

Two monoclonal antibodies L2 and HNK-1 were found to react with a common epitope on certain neuronal adhesive glycoproteins including N-CAM, L1, PO and MAG [ 109,1201231. Expression of the epitope is uneven so that only part of each glycoprotein in any tissue sample reacts. Glycolipids from human peripheral nerve or fetal brain are also recognized by the antibodies and have been characterized as sulfo-3-glucuronyl paragloboside and sulfo-3-glucuronyl neolactohexaosyl ceramide [ 124-1 261, the sulfo-3-glucuronyl group being the crucial determinant. The exact structure of the immunoreactive epitope on the neuronal glycoproteins is unknown but presumably contains the sulfated uronic acid residue. A glucuronyl transferase capable of transferring glucuronic acid from UDP-glucuronic acid to the terminal galactose fi 1-4 N-acetyl glucosamine units of asialoorosomucoid has been identified [ 1271 which is a good candidate for assembly of the L2/HNK-l epitope on neuronal glycoproteins. In MAG and PO glycoproteins the epitope is present exclusively on hybrid- or complex-type glycans that bind to pea lectin affinity columns indicating core al-6 fucose substitution [ 105,1061. Monovalent fragments of the L2 or HNK-1 monoclonal antibody block neuronal adhesions [128] and neurite outgrowth [ 1291. Furthermore, glycolipids or oligosaccharides carrying the L2/HNK- 1 epitope block neurite outgrowth [128]. The mechanism of these effects is unknown. One

525

possibility is interaction with domains of the neuronal glycoproteins that have affinity to other sulfated carbohydrates such as heparin. However, a recent study [ 1301 suggests that the sulfated glycolipid and heparin may bind separately to different domains.

3.2. Non-neuronal cells Other members of the Ig-superfamily function in conjunction with counter-receptors from separate families of adhesion molecules in various events such as leukocyte trafficking and the movement of immune cells and probably tumor cells across cell barriers in inflammation and metastasis. These glycoproteins (Fig. 10) are the immune cell adhesion molecules ICAM-1, -2 and -3, the vascular cell adhesion molecule VCAM and the platelet-endothelial cell adhesion molecule PECAM [ 13 1-1341.

3.2.1. /CAM Intercellular cell adhesion molecule 1 (ICAM-1) is widely expressed on cells of hematopoietic and non-hematopoietic tissues including lymphocytes, monocytes, fibroblasts, endothelial and epithelial cells. ICAM-I expression is low in normal tissues but inflammatory stimuli such as tumor necrosis factor a (TNF-a), interleukin 1p (IL-lp), interferon y and bacterial endotoxin dramatically up-regulate gene expression and cell surface localization [ 13 11. Up-regulation of ICAM-1 on vascular endothelium is an important step in initiating leukocyte adhesion and represents an early stage in migration of these cells into tissue sites of inflammation by extravasation [135]. Antibodies to ICAM- 1 inhibit leukocyte adhesion to endothelial cells and granulocyte migration through the endothelium [ 136,1371, binding of T-lymphocytes to fibroblasts and endothelial cells, and the interactions between cytotoxic T-lymphocytes and certain target cells [ 1381. The ICAM-1 polypeptide of 55 kDa is heavily glycosylated and the mature glycoprotein is 76 to 1 14 kDa on different cell types [ 1381 indicating tissue-specific glycosylation. The primary structure of ICAM-1 deduced from cDNA cloning [139-1411 contains an extracellular portion of five Ig-like domains, a transmembrane segment and a very short cytoplasmic tail of 28 amino-acid residues. Eight N-glycosylation sites are predicted, the majority of which must be utilized to account for the size difference between the unglycosylated and mature glycoproteins [ 138,1391. The N-glycans appear to be of the complex-type resistant to endo-H and must be highly branched or contain polylactosamine from the average size of the glycan at each site (about 5 m a ) . ICAM-I exhibits several binding activities, all of the heterophilic type involving counter-receptors. The first counter-receptor to be identified was the LFA-1 integrin [142], a heterodimer sharing a common p 2 chain with integrins Mac-1 and P150/95 as discussed later. Cells expressing LFA- 1, but not negative controls, bound to ICAM- 1 incorporated into liposomes and the binding was specifically inhibited by ICAM- 1 antibodies. Site-directed mutagenesis has shown that the first two Ig-like domains of ICAM-1 are involved in LFA-1 binding [ 1431. ICAM- 1 also binds the Mac- 1 integrin and the determinant for binding resides in the third Ig-like domain [144]. Finally, ICAM-1 is a cell surface receptor for the major serotype of rhinoviruses [ 1451 and the human malaria parasite [ 1461. ICAM-I-mediated adhesion of cells is sensitive to cytoskeleton-disrupting agents such as cytochalasin, suggesting that

526

a)

NH2

COOH

Fig. 10. Structures of ICAM, VCAM and PCAM adhesive glycoproteins. The sequence of ICAM-I (a), ICAM-2 (b), ICAM-3 (c), VCAM (d) and PECAM (e) predicted from cDNA sequencing. See Figs. 7 and 8 for other details.

the small cytoplasmic tail of ICAMs forms associations with the cytoskeleton, especially a-actinin [ 1471. In early studies it was found that anti-ICAM-1 monoclonal antibodies did not inhibit the LFA- 1-dependent adhesion of some cell types suggesting the presence of other LFA- 1-binding molecules. The molecules have now been identified, cloned and sequenced and shown to be ICAM homologues. ICAM-2 (Fig. lob) is found on endothelial cells even in the unstimulated state and may play a key role in the circulation of normal lymphocytes [ 1481. ICAM-2 is an integral membrane glycoprotein with two Ig-like domains that have closest sequence homologies with the two most N-terminal domains of ICAM-1 [149]. There are six N-glycosylation sites and the size difference

527

between the 5 5 4 5 kDa mature glycoprotein and the unglycosylated 28 kDa polypeptide shows that most of these must be occupied by high molecular mass oligosaccharides. ICAM-3 [ 1501 has recently been cloned and sequenced [ 151-1 531 (Fig. 1Oc). The mature 124 kDa glycoprotein contains five Ig-like domains in the extracellular portion which are 52% identical in sequence to the corresponding regions of ICAM-1, with highest identity contained within domain 2 and the first half of domain 3. ICAM-3 is the most highly glycosylated LFA-1 ligand with fifteen sites for N-glycosylation in a predicted -57 kDa polypeptide. N-glycanase treatment yielded an 87 kDa band which retained carbohydrate suggesting the presence of substitutions other than N-glycans or partial resistance to N-glycanase. However, consensus sites for 0-glycosylation are not present in ICAM-3 and the unusually high frequency of N-glycans (one in every 30 aminoacid residues) may well account for resistance to endo-glycosidases especially as with ICAM-1 and ICAM-2 these appear to be large, highly branched or extended complex-type structures. Although the homologies between the ICAMs are striking especially when compared on a domain to domain basis, their individually distinct structures and different patterns of expression suggest specialized roles. Part of this specialization may result from the relatively low homologies between the cytoplasmic tails of the three ICAM molecules which may be relevant to the activation of separate secondary events after LFA-1 binding. In any case, the ability of a combination of monoclonal antibodies to ICAM-1, -2 and -3 to inhibit completely the LFA- 1-dependent activation of mixed leukocyte populations suggests strongly that all of the major factors in LFA-1 adhesive interactions have now been identified [ 1541.

3.2.2. VCAM The vascular cell adhesion molecule VCAM was first identified as a cytokine-inducible adhesion molecule on endothelium, mediating the interactions of endothelial cells with tumor cells and leukocytes [ 155-1 571. Subsequently VCAM has been found to be present on macrophages, dendritic cells and muscle cells. The structure (Fig. 10d) shows the presence in the largest -1lOkDa forms of seven Ig-like domains, a transmembrane sequence and a short cytoplasmic tail of nineteen amino-acid residues. A splicing variant lacking the fourth Ig-like domain has been described [I581 and is expressed as a minor isoform in some tissues. In another splicing variation a new exon coding for 36 amino-acids is introduced behind the exon coding for the third Ig-like domain. This new sequence contains a signal for the addition of a GPI anchor [ 159,1601. The largest VCAM isoform contains six N-glycosylation sites, all except one of which are deleted from the GPI-anchored forms. Interestingly, in the mouse two GPI-anchored isoforms of VCAM have been described of 44 and 45 kDa and these forms appear to differ in glycosylation [161]. VCAM interacts specifically with the 0 1 integrin VLA4 El561 and binding involves the Ig-like domains 1 and 4 which show high sequence homologies [ 162,1631. Interactions between VCAM and VLA4 have been shown to be important in several immune functions involving many leukocyte classes in line with the widespread presence of VLA4 on all leukocytes except neutrophils [ 156,1641.

528

In muscle VCAM-I is present on myoblasts and myotubes whereas VLA4 is induced as the primary tubes form, and interactions between these molecules are believed to play an important role in myogenesis [ 1651.

3.2.3. PECAM PECAM- 1 is expressed on platelets, granulocytes, monocytes, lymphocytes, macrophages and endothelial cells as well as certain tumor cells [166-1721. PECAM-1 is a 120130kDa transmembrane glycoprotein carrying about 40% carbohydrate. There are nine N-glycosylation sites in the extracellular portion which consists of six Ig-like domains (Fig. 10e). Murine L cell transfectants expressing PECAM-1 aggregate in a PECAM- 1-dependent manner indicating that this glycoprotein is capable of homotypic interactions [ 1711. However, PECAM- 1 binds sulfated proteoglycans and the aggregation of transfected L cells is effectively inhibited by heparin, heparan sulfate or after treatment of the cells with heparinase [173]. These results suggest either a modulation of homophilic binding between PECAM- 1 molecules by binding of sulfated polysaccharides or a heterophilic interaction between PECAM- 1 and surface bound proteoglycans. The precise roles for PECAM-1 in uiuo are not clear. The molecule is concentrated at cell-cell contacts along the lateral domains of endothelial cells [ 166,1711 and in culture specific antibodies block lateral adhesions in confluent endothelial cell monolayers [ 1711. Cells transfected with PECAM- 1 transcripts show enhanced cell-cell adhesion and become less motile [ 1741. Exposure to PECAM- 1 antibodies of vascular endothelial cells, cultured in a three-dimensional collagen gel which supports angiogenesis, was found to severely limit normal multi-cellular tubule formation [ 1751. Other studies suggest that PECAM-1 is essential for late events in leukocyte migration and extravasation [ 1761 and is expressed on the surface of human tumor cells where it mediates adhesion to microvascular endothelial cells [ 1771.

4. Matrix glycoproteins The extracellular matrix surrounding cells is morphologically and biochemically diverse, reflecting a diversity in function and the cell types laying down the matrix. Apart from roles in maintenance of tissue architecture and cell migration the extracellular matrix is an important reservoir of growth factors and cytokines, many of which are now known to bind specifically to matrix components. Broadly speaking, as shown in Fig. 1b, the extracellular matrix can be categorized into two morphologically and biochemically distinct structures. Basement membranes provide a support for epithelia and separate them from underlying mesenchyme. In adult tissues basement membranes contain mainly collagen type IV, non-collagenous glycoproteins such as laminin and entactin and heparan sulfate proteoglycans. The interstitial matrix surrounding mesenchymal cells contains a variety of collagen types (I, 111, V, VI, XI1 according to location), non-collagenous glycoproteins such as fibronectin, and proteoglycans. In this section the major non-collagenous glycoproteins of these matrices will be described to emphasize common structural and functional features and the potential roles of glycosylation. The non-collagenous components of extracellular matrices are mosaic proteins of high

Gangliosides Bacteria Fibronectin Fibrin

-

529

%PI

as$,

Gelatin

n

COOH

t

U Fibronectin

ED-A

Subunit 1

t

ED-B

U

t

ss

L_ _ _ - -

NH2

Subunit 2

Fig. 1 1. Structure of a fibronectin subunit. Fihronectin contains three types of sequence modules, type 1 (open rectangles), type 11 (large open circles) and type 111 (open squares). Two type 11 nodules (solid squares) can be present or absent due to alternative RNA splicing (ED-A and ED-B). Additional splicing variants produce sequence variables in the type 111 modules of the V region. Subunits are joined in an anti-parallel manner by two cysteine pairs at the carboxyl terminals. The two free sulfhydrals are shown (SH). The sites of N-glycosylation (small open circles) and 0-glycosylation (small solid circles) are indicated. Binding domains as indicated include two each of heparin- and fibrin-binding domains, a gelatin-collagen binding site and sites for binding of certain bacteria and for fibronectin-fibronectin interactions. The major cell binding site recognized by integrin asO, contains the RGD recognition sequence and a second cell binding site occurs in the V region and is recognized by integrin a&.

molecular mass containing multiple functional domains capable of interacting with each other or with other matrix components, an important requirement for formation of the matrix. Additionally, these proteins contain cell-binding domains responsible for their diverse effects on the adhesion, activation and motility of cells. 4. I. Fibronectin

Fibronectin is a dimer consisting of -220 kDa subunits [I 78-1 801. Although encoded by a single gene, the fibronectin protein shows great structural variety due to alternative RNA splicing. Two splicing regions are located in the central part of the protein and are called extra domains A and B (ED-A and ED-B). In the mature protein these domains can either be included totally or totally excluded. The third splicing region called the variable (V) or connecting sequence (CS) is more complex and the sequence can be completely absent, present in part or present in full as a result of intra-exon splicing. Up to twenty fibronectin variants result from various combinations of splicing patterns and subunit dimerization. Splicing patterns appear to be tissue-specific and are regulated in development [ 1811. For example, the plasma form of fibronectin, produced by hepatocytes, lacks the ED-A and ED-B encoded sequences and contains distinct polypeptide subunits resulting from splicing variants in the V region [ 1821. Fibronectin is a mosaic protein made up of a number of modular repeats (Fig. 1 I). The type I and I1 modules are disulfide bonded loops encircling 45-50 amino-acid residues whereas the type 111 module contains on average 90 amino-acid residues and is not circumscribed by a disulfide bond. These modules appear to be basic structural units and homologous sequences are widely found in structural and catalytic proteins [ 1781. Their distinct folding characteristics have been studied in solution by NMR and other physical methods [ 183-1 851 but in the intact protein presumably several modules interact to form larger structural and functional domains. Electron microscopy using rotary shadowing

530

reveals fibronectin as a V-shaped structure which is formed by interchain disulfide bonding between two pairs of cysteine residues in the carboxyl-terminal region [ 1861. The two polypeptide subunits are joined in anti-parallel fashion as shown by the cysteinecysteine pairing and by NMR analysis of carboxyl-terminal fragments [ 187,1881. Partial proteolysis of fibronectin produces fragments that have distinct binding properties (Fig. 11). The first to be described was a gelatin- or collagen-binding domain situated (Fig. 11) at the N-terminus and involving the type I1 repeats [189]. A 14 residue amino-acid sequence located between the second type I1 repeat and the adjacent type I repeat appears to be essential for collagen [ 1901 binding but additional putative collagen-binding sites that include the first type 111 repeat have recently been described [191]. Similarly fibrin binding appears to utilize at least two separate sites on fibronectin (Fig. 11). The interactions are weak but are stabilized by cross-linking with blood factor XI11 transglutaminase in coagulation [ 1921. Interestingly, cellular forms of fibronectin which, unlike plasma fibronectin, include ED sequences do not become incorporated in cross-linked fibrin [193]. A number of pathogenic bacteria bind to fibronectin, an important determinant in adherence to and colonization of host tissues [14]. The binding site for S.aweus appears to be close to, but distal to the fibrin binding site at the N-terminus of fibronectin (Fig. 11). Fibronectin also binds to sulfated polysaccharides such as heparin (Fig. 11). The significance of these interactions are believed to relate to the binding of heparan sulfate proteoglycans and other proteoglycans which may occur in extracellular matrices [194]. As far as heparin binding is concerned, binding appears to be of highest affinity at the most carboxyl-terminal site under physiological conditions [ 195,1961. An interesting question is the specificity of fibronectin interactions with heparin-like molecules. Recent data [ 1971 suggest that specific structural features within the heparin molecule can confer preferential binding as shown by the ability to fractionate commercial heparin into components with distinctly different avidities for fibronectin. Identification of the structures of the high affinity heparins for fibronectin, and similar analysis of heparan sulfates will be of great interest. Defined structural features of heparin are known to be involved in binding to other proteins such as antithrombin and growth factors [ 1981 raising the possibility that interactions between matrix components such as fibronectin and heparan sulfate proteoglycans may be regulated at the level of polysaccharide chain modifications in the latter. Fibronectin binds to other sulfated polymers such as chondroitin sulfate proteoglycans [199]. In this case however the free polysaccharide chains do not interact [200] suggesting that the core protein may be important in binding. The sites on fibronectin responsible for binding these proteoglycans have not been identified. At least two cell-binding sites have been localized in fibronectin which mediate the attachment and spreading (Fig. Id) of various cells though interactions with integrins a5pl or a& (Fig. 11). These interactions are peptidebased [201] and peptide fragments derived from the cell-binding domains of fibronectin can substitute to some extent for the complete molecule in cell attachment assays, including the RGD sequence in the tenth type 111 domain recognized by a5pl integrin, and an RGD-independent sequence present in the V site of alternative splicing, recognized by a& integrin as discussed later. Another property of fibronectin is its ability to assemble at nucleation sites on the cell surface into disulfide cross-linked fibrillar aggregates independently of interactions

53 1 Galpl + 4GlcNAcp1+ 2Manal

L6 Manpl + 4GlcNAcpl + 4GlcNAc NeuAcuZ + 6Galp1 + 4GlcNAcpl*

2Manal

NeuAca2 -+ 6Galp1+ 4GlcNAcPl+ 2Manal

’

k g Manpl NeuAca2 + 6Galp1+ 4GlcNAcP1+ 2Manal

-+

4GlcNAcp

-+

4GlcNAc

Fig. 12. Structures of the major asparagine-linked glycans of human plasma fibronectin. Data from Takasaki et al. [215].

with other matrix components [ 1851. Several regions of fibronectin are involved in self-assembly. The first five amino-terminal type I modules and the first type I11 module (Fig. 1 1) are especially important [202-2041 but the carboxyl-terminal region [205] and the RGD-containing type I11 repeat [206] may be required although conflicting results have been reported [205,207]. Fibrillogenesis of fibronectin also appears to involve gangliosides. Ganglioside-deficient mutant fibroblasts were found to produce and secrete fibronectin but not form a pencellular fibronectin matrix unless treated with mixed ganglioside preparations. Some of these apparently become associated with the cell surface and perhaps served as nucleation sites for fibronectin assembly [20821 11. The ganglioside binding site on fibronectin has been localized [212] to a region, already associated with fibrillogenesis, at the amino-terminus of fibronectin (Fig. 1 1). In general, highly sialylated glycolipids such as GD2 gangliosides bind preferentially. Interestingly, a cryptic lectin-like site that binds oligosaccharides terminated with sialic acid has been detected in plasma fibronectin after partial proteolysis [2 131. Lectin activity could be demonstrated in a 23kDa proteolytic fragment originating from the major RGD cell-binding domain but not in a larger 125kDa fragment containing this and flanking domains. It was suggested that in the larger fragment and in fibronectin itself the lectin site is occupied by a sialic acid moiety of an oligosaccharide substituted elsewhere on the same or an interacting polypeptide [213]. However, the relevance of these findings to the role of gangliosides in fibrillogenesis, for example as a bridging molecule to the cell surface or between fibronectin molecules, is unknown. The fibronectin subunit contains seven N-glycosylation sites (Fig. 11) as confirmed by protein sequencing of bovine plasma fibronectin [214] and inferred from sequences of other fibronectins deduced from cDNA clones. As is the case with all glycoproteins the glycosylation of fibronectin is species-, tissue- and cell-specific and within a single species shows interesting developmentally regulated patterns. These points are best demonstrated by comparison of the N-glycans of human plasma fibronectin [2 151 representing the adult form, and the fibronectins of human amniotic fluid [216] and human placenta [217], which are fetally derived forms synthesized by epithelial cells of the amniotic membrane and endothelial cells of fetal blood vessels, respectively. The N-glycans of human plasma fibronectin are sialylated diantennary complex glycans containing three or four terminal sialic acid residues (Fig. 12). In human amniotic fluid fibronectin the dominant glycan is again the diantennary complex-type but unlike plasma

532

fibronectin, these glycans contain a substantial proportion of core Fuca 1-6 substituents and in addition about one third carry bisecting GlcNAcfi1-4 residues. Small proportions of triantennary glycans with or without core fucose and bisecting N-acetylglucosamine residues were also found (Fig. 13). About half of these glycans carry NeuAc a23 and a2-6 linked, the remainder being neutral oligosaccharides. By contrast plasma fibronectin contains only NeuAc a 2 4 substituents. In another study of human amniotic fluid fibronectin [2 181 no bisected diantennary-type complex glycans were detected and triantennary isoforms appeared to be the major glycans. These differences may relate to the use of mid-term amniotic fluid in the latter work compared with term fluid used by Takamato et al. [216]. Changes in fibronectin splicing variants in amniotic fluid during pregnancy have been reported [2 19,2201 which could account for fine differences in glycosylation patterns. The glycans of human placental fibronectin (Fig. 14) are characterized by the predominance of tetra- and triantennary glycans which may or may not be core fucosylated and sialylated[217]. In addition these glycans may carry polylactosamine (PLA) extensions. The diantennary glycan fraction is more minor than the highly branched and PLA extended chains and is bisected with GlcNAcB 1-4 residues. About half of the glycans are sialylated, exclusively by NeuAca2-3 residues, a common feature of placental glycoproteins [2 171. The glycan structures of bovine plasma fibronectin have also been described [221] and show interesting species variation compared with human fibronectins. Of special interest is the presence in the diantennary complex-type chains of type I Galfil-3GlcNAc units, NeuAca2-4 substituents on the terminal galactose units and disialylated chains containing this sialic acid residue and an internal sialic acid linked a2-6 to the penultimate N-acetylglucosamine residues. The glycosylation patterns of cell- or matrix-associated fibronectins have been less studied than soluble forms. Human and hamster fibroblast fibronectins have predominantly diantennary complex chains, in the case of the former bearing core fucose substituents and NeuACa2-3 terminal residues [222,223]. Undifferentiated mouse embryonal carcinoma cells appear to synthesize an unusual form of fibronectin containing in addition to PLA structures a covalently attached heparan sulfate [224]. The site of attachment of the heparan sulfate is unknown but could involve the 0-glycosylation site identified[225,226] in some fibronectin splice variants (Fig. 11). In human amniotic fluid fibronectin [218] the 0-glycans have been identified as Galfi1-3GalNAc, NeuAca2-3Galfi 1-3GlcNAc and NeuAca2-3GalB 1-3(NeuAc a2-6)GlcNAc, typical of many 0-glycosylated glycoproteins. It is possible that glycosylation at this site could change according to cell type or developmental stage to introduce other types of-glycans, even glycosaminoglycans. An interesting developmental change in fibronectin glycosylation was reported in chondrocytes [227]. It appears that the glycans surrounding the collagen binding domain of fibronectin isolated from these cells are endo-Hi-sensitive (oligomannosidic-type or hybrid-type) and become processed to endo-H-resistant complex-type glycans after treatment of the cells with vitamin A. The functional significance in glycosylation differences between various fibronectins remains to be determined. Fibronectins with greatly altered glycosylation patterns due to inhibition of processing in cells by monensin [228,229], swainsonine [230] or in

533 GlcNAcDl Galpl + 4GlcNAcPl -+2Manal

+

-f. Fucal

GlcNAcp1

t Fucal

+

6 L6 Manpl + 4GlcNAcDl -P 4GlcNAc Galpl + 4GlcNAcPl +2Manal ,a

+

Gal01 * 4GlcNAcPl *2Manal

+ 6

k gManpl + 4GlcNAcPl + 4GlcNAc GlcNAcPl +2Manal Y 3

Galpl

-P

GlcNAcPl 4GlcNAcPl +2Manal

Galpl + 4GlcNAcpl

c

,

:Manal

t Fucal

c

6 k gManpl + 4GlcNAcPl + 4GlcNAc

f3

Galpl + 4GlcNAcP1

t Fucal Gal(31 + 4GlcNAcPl

+2Manal

GalPl + 4GlcNAcPl

-+ 2Manal

+

6

L6 Manpl + 4GlcNAcP1 + 4GlcNAc f3

-f Fucal

GalPl + 4GlcNAcPl +2Manal GalPl + 4GlcNAcp1

+

6 k gManpl + 4GlcNAcPl + 4GlcNAc

f3 Manal

Galpl + 4GlcNAcpl f 2

Galpl

* 4GlcNAcpl

t Fucal L6Manal 2

fl

c

6 L6 Manpl * 4GlcNAcp1 * 4GlcNAc

Galpl * 4GlcNAcpl Y3 Galpl + 4GlcNAcpl+ 2fvlanal

Fig. 13. Structures of the major asparagine-linked oligosaccharides of human term amniotic fluid fibronectin. Data from Takamoto et al. [216].

534 GalpI * 4GlcNAcpl

t Fucal

x6Manal 2 x6Manpl f Galpl * 4GlcNAcDl d3

+

6 * 4GlcNAcpl * 4GlcNAc

Galpl + 4GIcNAcpl

x6Manal

f

2

t Fucal

+

x

GalpI* 4GlcNAcpl Galpl * 4GIcNAcPlx,

3

6 Manpl .+4GlcNAcpl * 4GlcNAc

#

Galpl + 4GlcNAcpl f

Galpl* (Galpi * 4GlcNAcp1 * 3) -n

f Manal

'

t Fucal

4GlcNAcpl

4

x6 Manal

x:

4GlcNAcpI f12

Galpl

-t

Galpl

+ 4GlcNAcp1 + 2Manai

i

6 Manpl -t 4GlcNAcpl * 4GlcNAc

Galpl * 4GlcNAcpI

(Galpl -* 4GIcNAcpl

-*

3)1-n

.f Fucal

+

h6 Manul

* 4G1cNAcp1f 2 Galpl * 4GlcNAcpI

Galpl

Galpl * 4GlcNAcp1 f

2

Manal

6

'*6 ManpI

f13

GlcNAcpl

Galpl -c 4GlcNAcpl Galpl * 4GlcNAcpI

* 2Manal * PManal

+

4 kg Manpl

-*

t Fucal

+

6

* 4GlcNAcDl + 4GlcNAc

f3

GlcNAcpl Galpl + 4GlcNAcp1

* 4GIcNAcpl -t 4GlcNAc

+

2Manai

x 6 Mtnpl

t Fucal

+ 6

* 4GlcNAcpI * 4GlcNAc

GlcNAcpl + 2Manal f 3

t Fucal Galpi * 4GlcNAcfil

+

* 2Manal

6

x 6 Manpl

Galpl * 4GlcNAcp1

* PManal

* 4GlcNAcpl -c 4GlcNAc

f3

Fig. 14. Structures of the major asparagine-linked oligosaccharides of human placental fibronectin. Data from Takamoto et al. [217].

535

glycosylation-defective ricin-resistant cells [23 1,2321 appear to be secreted normally from cells and are functional in cell adhesion activity and other properties. However, the presence of carbohydrate appears to stabilize fibronectin against proteolysis [233,234] and PLA glycosylation of human fetal fibronectin is reported to weaken binding affinity to gelatin [234,235]. In the latter study, binding to heparin was unaffected, suggesting that the relatively high degree of glycosylation in the collagen binding domain (Fig. 11) may modulate binding activity preferentially. 4.2. Laminin

Laminin, a large >800 kDa multi-subunit glycoprotein of basement membranes, was first isolated from the murine Engelbroth-Holm-Swarm (EHS) tumor and from the murine parietal yolk sac (PYS) carcinoma [236,237]. Laminin is a key component of basement membranes in mature organisms and in very early embryogenesis, for example in the 2 - 4 cell stage of mouse embryos where it is the major basement membrane component to be expressed. Structural analysis of this complex molecule has used a combination of biochemical studies of the intact glycoprotein, its fragments obtained by partial proteolysis with for example elastase (fragment E3 and E8) and pepsin (fragment P l), electron microscopy of the intact glycoprotein and fragments alone or in combination with ligands such as type IV collagen, and cDNA cloning and sequencing [238-2411. These studies have produced a detailed model of the laminin molecule which is shown in simplified form in Fig. 15. To date little is known about the structure of laminin in normal tissues, in particular glycosylation status, due to the difficulties inherent in characterization of such large and complex multi-subunit glycoproteins and availability of sufficient amounts of material for structural analysis. However, the available data indicate that the EHS laminin represents only one of a family of related molecules which differ in subunit composition and post-translational modifications including glycosylation. Laminin extracted from the EHS tumor basement membrane is a disulfide-bonded complex of three distinct subunits B 1, B2 and A of 220, 2 10 and -400 kDa, respectively. The three subunits are arranged (Fig. 15) into a cross-shaped molecule comprising one long arm terminated by a complex multi-globular structure and three short arms with larger globular structures at each end and within the short B 1 arm. The complete aminoacid sequences of the B1, B2 and A chains of EHS laminin have been determined by cDNA cloning and sequencing [242-2441 and show that the non-globular regions of the short arms contain numerous EGF-like repeats of about fifty amino-acid residues with eight cysteines. In addition, the short arms of the A and B2 chains have two and one, respectively, of a modified EGF-like domain which contains an insert of 18&200 aminoacid residues. The carboxyl-terminal segments of each chain combine to form a long rod-like structure characteristic of interacting triple a-helices. The interactions between the subunits of laminin are stabilized by disulfide bonding at the centre of the cross and at the end of the long arm rod. In keeping with its large and complex structure, laminin has been shown to participate in a number of binding interactions with diverse cell surface receptors as discussed later. In addition, laminin contains binding sites for other matrix components such

536 A

la)

-

-

Collagen? Self-assembly 7€4

-Chain

nidogen binding

1

-Chain YlGSR 67 kDa receptor

1 €6

cell binding

neurite promotion a6pl

1 1I Heparin

-7 c 82

5

A

laminin

M merosin

5

w A

S-laminin

c M S-merosin

Fig. 15. Structure of laminin isoforms. (a) Models of murine EHS tumor laminin deduced from cDNA sequencing, secondary structure predictions, and analysis of proteolytic fragments. EGF-like domains are indicated by small open circles and modified EGF-like domains containing -200 amino-acid inserts are shown as small solid oblongs. Globular domains are shown by large solid circles and a large a-helical domain involving all three subunits by long, open columns. Putative interchain disulfide bonds are indicated. Some proteolytic fragments mentioned in the text (E3, E4, E8) and ligand binding regions are shown. (b) Laminin isoforms containing A chain (M) or B1 chain (S) variants. (c, facing page) Glycosylation sites in EHS laminin (open triangles).

537

A Chsln

Fig. 15c.

as collagen, heparin and heparan sulfate proteoglycans and the matrix glycoprotein nidogedentactin, interactions that play important roles in assembly and maintenance of extracellular matrices. The location of some of these binding domains have been determined, principally by study of the binding properties of proteolytic fragments. A strong binding domain for heparin is located in a proteolytic fragment E3 deriving from

538

the carboxyl-terminal region of the long arm of laminin. Recently, the carboxyl-terminal region of the A chain containing the five small globular domains (Fig. 15a) has been produced in a Baculovirus expression system and shown to have an enhanced heparin binding activity relative to intact laminin [245,246]. There appear to be two separate heparin-binding sites, one each in the innermost (Gl) and sub-terminal (G4) globular structures in this region. Presumably, the latter site accounts for the heparin binding activity of fragment E3 (Fig. 15a). Interestingly, a larger protein constructed from the five globular domains of the A chain with part of the distal rod-like structure of this chain showed a reduced heparin binding activity, more similar to that of intact laminin. This result indicates that the binding activities of one or more of the individual globular domains are modulated in some way by the overall conformations of the sites in the intact mc lecule compared with the individual domains. Laminin also binds to other sulfated sugars such as a monogalactosyl sulfatide, galactosyl ceramide-13-sulfate [247]. However, binding appears to be independent of the major heparin binding sites and requires the globular termini of the short arms of one or both of the B chains, which have been shown to have weak heparin binding activities [248]. The biological significance of sulfatide binding is unknown. As for heparin binding activity, presumably this reflects an interaction with heparan sulfate proteoglycans which co-exist with laminin in basement membranes. Laminin also binds to collagen IV, another major component of basement membranes [249,250], probably through the globular domains at the end of the B1 chain short arm. Two other interactions that are probably important for integration of laminin into basement membranes involve binding to the glycoprotein, nidogedentactin and self-assembly of laminin. The binding of nidogedentactin to laminin is especially important since this appears to increase the affinity of laminin for other basement membrane components, in particular collagen type IV and proteoglycans [25 1,2521. Nidogedentactin binds to the short arm of the B2 chain (Fig. 15a) and a single EGF-like repeat has been identified as the dominant binding structure [253]. Laminin can aggregate in solution in a calcium-dependent manner which may also be an important property in matrix assembly. The globular domain at the end of the B1 chain is particularly important for self-assembly [254,255]. In addition to interactions with other matrix components, several sites are involved in interactions with cell surface receptors, including integrin and non-integrin components. Many cells can attach and spread (Fig. Id) on laminin-coated surfaces and at least six integrins bind laminin namely alB1, a z p I , a3P1,a&, avPl, and a 6 B 4 [256]. Many of the binding domains of laminin recognized by these integrins remain to be determined. However, an elastase fragment (E8) retains binding capacity for a& integrin [2572611 as indicated in Fig. 15a. Recently a synthetic peptide containing 21 aminoacid residues based on a sequence within the last globular domain at the carboxylterminus of the A chain was found to bind specifically to a 3 B l integrinE2621. An additional integrin binding site that contains the RGD sequence has been located in proteolytic fragment P1 [259]. This site is apparently not active in mediating cell attachment in the intact laminin molecule and its significance is unknown. There is evidence that possibly alBl[258,263] may bind at this or adjacent sites. The binding sites of a 2 B 1 and asp4 integrins are unknown. A non-integrin laminin binding protein,

539

the 67 kDa receptor [264] recognizes a sequence YIGSR in one of the EGF-like domains in the short arm of the B1 chain (Fig. 15). The EHS tumor laminin represents one member of a family of related molecules. Variant A and B chains have been described and these, in conjunction with the A or B chains already described, assemble into unique complexes that are tissue- and cell-type-specific (Fig. 15b). Merosin is a homologue of the EHS laminin A chain [265] present in human placenta and heart muscle. The complete primary sequence from cDNA cloning shows a very similar domain structure to the EHS A chain [266]. S-laminin is a B1 chain homologue abundant in laminin at skeletal neuromuscular junctions. The sequence[267] contains a tripeptide Leu-Arg-Glu (LRE) motif that appears to be a recognition unit required for interaction with an as yet unidentified cell receptor involved in neuronal cell adhesion and neurite outgrowth. Other variants of the EHS A chain for example the K chain [268], the B1 chain, an avian eye laminin subunit [269], and a truncated B2 chain in a human tumor [270] are known. At the protein level there is good evidence for laminin isoforms containing various combinations of the A, B 1, B2, S and M chains as shown in Fig. 15b. Expression of these various forms is developmentally and tissue regulated [271-2731. Laminin is a highly glycosylated molecule. In EHS laminin there are 46 potential N-glycosylation sites in the A chain and fourteen each in the B1 and B2 chains (Fig. 15c). Studies of laminin biosynthesis in cell cultures in the presence of tunicamycin showed non-glycosylated chains of -300, 205 and 185kDa for the A, B1 and B2 subunits, respectively [274-2761, in reasonable agreement with the expected values (336, 194 and 174 kDa) from sequences deduced from cDNA analysis. These results indicate carbohydrate contents of about 25%, 10% and 15% for the A, B1 and B2 chains of murine tumor laminin. Structural analysis of the glycans present in each laminin subunit is a formidable task given the large number of potential glycans and so far only some general structures are available. However, at least forty N-glycan structures are present and distributed among all chains, especially along the rod-like structure in the long arm. Some of the earlier data was obtained by plant lectin reactivities. Griffonia simplicifoliu B4 isolectin was first shown to stain basement membranes [277] and subsequently EHS laminin [278,279] indicating the presence of terminal a-galactosyl residues. Reactivity was strongest in laminin fragments containing the terminal globular domains of the short arms and least in the rod-like structure of the long arm. In addition EHS laminin and fragments reacted with Con A, indicating the presence of high mannose, hybrid or diantennary complex-type chains, and with Phaseolus oulguris and Datura strumonium lectins indicating more highly branched structures. Ricinus communis agglutinin (RCA-I) recognized A and B subunits of EHS laminin and the merosin M chain of murine muscle [280] indicating the presence of b-galactosyl residues. EHS laminin A and B chains also bind potato lectin [28 I] diagnostic of polylactosamine glycans. More detailed analysis of glycopeptides, obtained from EHS laminin and separated into various fractions by Con A-Sepharose chromatography [282], showed the presence of di-, tri- and tetraantennary complex-type chains, some of which contained terminal sialic acid or a-galactose residues. Others carried polylactosamine extensions. In addition, high mannose-type glycans binding tightly to Con A-Sepharose were identified. In EHS laminin all of the sialic acid is a2-3 linked [283]. Evidence for small amounts

540

Fragment Structure

El

E3

P1

Gal-GlcNAc-Man\ Gal-GlcNAc-Man-R'

+

+

+

Gal-GIcNAc-Gal-GlcNAc-Man\ Gal-GlcNAc-Man-R'

+

Gal-GlcNAc-Man\ Gal-GlcNAc-Man-R'

+

+

+

Gal-GlcNAc'

(Gal-GlcNAc)-Gal-GlcNAc-Man \ -Gal-GlcNAc-Man-R' -Gal-GlcNAJ

+

Gal-GlcNAc Gal-GlcNAc-Man-R' Gal-GlcNAc-Man/ GaT-GlcNAc'

+

,

Fig. 16. Branching patterns and presence of polylactosamine in complex-type N-glycans of laminin fragments. Data from Fujiwara et al. [289]. Each structure is variably terminated with sialic acid and/or a-galactosyl core residues. residues. R' is Man~14GlcNAc~l4GlcNAcfFuca14

of unusual diantennary glycans containing fucosyl substituents on terminal penultimate N-acetylglucosamine residues has also been reported. Only one study to date has reported on the question as to how these diverse N-glycan structures are distributed in the laminin molecule [284]. Asparagine-linked oligosaccharides were released from well-characterized proteolytic fragments including the E3 heparin-binding domain and the E4 and P1 fragments (Fig. IS), the latter originating from segments of all subunits in the centre of the cross of the intact molecule. Fragment E4 contains about 3-4 N-glycans, fragment E3 contains N-glycans in the last two globular domains originating from the extreme carboxyl-terminus of the A chain and fragment PI contains up to ten N-glycans. Interestingly, the branching patterns and polylactosamine content of complex-type oligosaccharides from these three fragments were strikingly different, indicating heterogeneous glycosylation in various regions of the laminin subunits (Fig. 16). In general, the globular domain fragments E3 and E4 appear to be enriched in simple diantennary complex-type glycans, whereas the centre-cross fragment P 1 carries more highly branched glycans, many carrying polylactosamine extensions. It will be interesting to extend these observations to other domains of EHS laminin, in view of the functions suggested for glycosylation in laminin as described later. In addition, the carbohydrate structural analysis of laminin from sources other than the murine EHS tumor is needed to generalize the findings discussed above. The likelihood that laminin glycosylation does vary according to species, source and development stage is already apparent from a simple compositional

54 1

comparison between laminins from EHS and another murine source, parietal endodermal cells [236], which indicates a relatively high content of N-acetylgalactosamine in the latter. The EHS laminin by contrast contains little N-acetylgalactosamine and hence is not significantly 0-glycosylated. 4.3. Nidogedentactin

Nidogen was identified as a component co-purified with the EHS tumor laminin [285]. Subsequent cDNA cloning [286] showed identity with another glycoprotein, entactin, detected in murine parietal endodermal and other cell cultures [276,287]. The glycoprotein of -150 kDa consists of three globular domains G1, G2 at the N-terminus and G3 at the C-terminus separated by more extended linking regions (Fig. 17a). The size (-140kDa) of the unglycosylated protein isolated from tunicamycin-treated cells indicates that each of the two N-glycosylation sites in the first and second globular domains (Fig. 17a) are substituted with asparagine-linked oligosaccharides. The predicted glycan size at each glycosylation site (-5 kDa) suggests highly branched, complex-type structures, consistent with a -2% glucosamine content of the glycoprotein [288] and structural analysis [289]. The purified glycoprotein also contains about seven 0-glycans [289]. Nidogedentactin is believed to play a key role in formation and organization of basement membranes since it binds separately to laminin, collagen IV and heparan sulfate proteoglycan and promotes the assembly of ternary complexes of these components [25 1,252,2901. Laminin binding to nidogedentactin involves a single EGF-like domain in the B2 short arm of laminin as discussed previously and the C-terminal G3 domain of nidogedentactin [29 11. The separate binding of collagen type IV and heparan sulfate proteoglycan is mediated by sites in the G2 globular domain [291]. Interestingly, unlike the major heparan sulfate proteoglycan-binding sites of laminin and fibronectin, heparin is not recognized and the interaction of nidogedentactin appears to involve the protein core of the proteoglycan [252,29 11. Nidogedentactin has been shown to mediate cell attachment of epithelial cells [292] and non-epithelial cells 12931. The central globular domain of nidogedentactin contains an RGD recognition domain which in certain cases may be functional in cell attachment [294]. However, the a3P, integrin which mediates adhesion of human carcinoma cells to nidogedentactin [295], appears not to utilize this sequence and its binding site on nidogedentactin is unknown. 4.4. Tenascin

Tenascin was discovered independently on several occasions and given different names. Electron microscopy of chick fibronectin preparations showed in addition to the typical hinged two-armed fibronectin molecule a large six-armed structure, named hexabrachion [296] which was shown to account entirely for the hemagglutinating activity previously suggested to be an inherent property of chick fibronectin [297]. Later, molecules called myotendinous antigen, cytotactin, J 1 glycoprotein and gliomamesenchymal extracellular matrix antigen GMEM [298,299] were shown by cloning and sequencing data to be the same entity. However, further studies have shown that there

542

(a>

proteog CollaYen ycan m

Laminin

NHz

COOH

-

RGD

Collagen proteoglycan

Heparin

4 % OOH

Domain

I

NHZ

\

Calcium . binding . . /

Fig. 17. Model structures of nidogedentactin (a), tenascin (b) and SPARC/osteonectin (c). (a) Nidogedentactin, a -150 kDa glycoprotein, contains three globular domains (Gl, G2 and G 3 ) separated by a short flexible link region and a rod-like element consisting of several EGF-like motifs (solid oblongs) one of which includes an RGD sequence. Two putative N-glycosylation sites (solid circles) are present which apparently are glycosylated. In addition the presence of N-acetylgalactosamine indicates 0-glycosylation, the location of which is unknown. (b) The subunit of tenascin contains fourteen EGF-like motifs (solid ovals), invariably eight fibronectin type I11 repeats (1-8) (open squares) and a variable number of additional type 111 repeats (A1 etc.) according to source, and a C-terminal fibrinogen-like domain (). An RGD sequence and potential N-glycosylation site (solid circle) are shown. Heparin-binding domains are indicated. (c) SPARC/osteonectin, a -43 kDa glycoprotein consists of an acidic domain (I) including several cation binding sites, a domain (11) with homologies to follistatin an inhibitor of OTGF-like growth factors, a-helical domain (111) and a domain containing (IV) an EF-hand cation binding site. (d) The single N-glycan (solid circle) is substituted in domain I1 in an exposed loop as predicted from secondary structure modelling.

543

are many variants of the basic structure, derived principally by alternative RNA splicing of the gene transcripts (Fig. 17b). Three major splice variants have molecular masses of -230, 200 and 190 kDa [300,301] which differ in the inclusion of three, one or no extra fibronectin type 111 repeats, respectively, in the C-terminal half of the protein (Fig. 17b). Additional splicing variants have subsequently been described containing up to seven extra type 111 repeats [302] indicating that tenascin is present in extracellular matrices in a large variety of forms, that presumably have functional but as yet unknown significance. The amino-terminal domain of tenascin consists of four repetitive sequences homologous to motifs in the C-terminal rod-like extension of laminin. This domain is followed by fourteen EGF-like domains, a variable number of fibronectin type I11 repeats according to the splicing pattern, and a fibrinogen-like globular domain. In the intact molecule, three subunits assemble at the N-terminal domain and two trimeric species are linked covalently at the N-terminus by a disulfide bond to produce the six-armed mature structure. The tenascin subunits are extensively glycosylated since their molecular masses predicted from sequence data and migration of non-glycosylated polypeptides on SDS-polyacrylamide gel electrophoresis indicate values of 200, 180 and 170 kDa, respectively [3011. The higher carbohydrate content of the larger variants is expected from the numerous putative glycosylation sites in the alternatively spliced fibronectin type 111 repeats (Fig. 17b). The average size predicted for the N-glycans of a fully glycosylated 230 kDa tenascin subunit (-2-3 kDa) could accommodate universal glycosylation with simple glycans or a more site-restricted glycosylation with branched, complex-type glycans. Tenascin is especially prominent in embryonic tissues in areas adjacent to rapidly proliferating epithelia where extensive cell migration or branching occurs, in the developing central nervous system along the pathways of migrating cells, in developing connective tissue and at uterine implantation sites [298,299,303]. In the adult, tenascin is predominant at the tips of intestinal villi where cell shedding occurs [304,305], and at the corticomedullary junction of thymic lobules at sites of thymocyte maturation and migration [306]. Tenascin binds to extracellular matrix glycoproteins such as fibronectin [307-3091, heparin [3 lo], heparan sulfate [3 1I] and chondroitin sulfate [3 12-3 141 proteoglycans. Some of these interactions appear to be regulated by the inclusion of extra fibronectin type I11 repeats, and hence higher glycosylation, in tenascin splicing variants. For example, the small tenascin variant has a higher affinity for fibronectin than tenascin isoforms containing variable domains [309]. Co-regulation of tenascin and chondroitin sulfate proteoglycan expression has recently been reported in limb development [3 151 suggesting functional significance in the binding affinity between these two molecules. The binding sites on tenascin for heparin involve separately a region covering the fourth and fifth fibronectin type 111 repeats and the fibrinogen-like globular C-terminal domain [3 161. Interestingly, the latter domain produced in a bacterial expression system was shown to support the weak attachment of primary rat fibroblasts [3 161. The adhesion was specific since heparin or heparan sulfate proteoglycan blocked attachment as did treatment of the cells with heparinase or with inhibitors of sulfation reactions such as chlorate or growth in sulfate-depleted media. The role of the second heparin-binding domain of tenascin, which appears to be of higher affinity [3 161, is unknown as is the significance of the RGD sequence in the third fibronectin type 111 repeat of tenascin.

544

In simple adhesion assays in culture it is generally agreed that surfaces coated with tenascin are non-adhesive or weakly adhesive [298,317]. In some cases it blocks adhesion of cells mediated by other matrix glycoproteins such as fibronectin [3 18-3201, consistent with the association of tenascin with sites of epithelial expansion and cell migration. However, a bacterially expressed fragment of tenascin containing the type 111 repeats 1-6 does appear to support the attachment and spreading of fibroblasts and other cells [62], suggesting a strong adhesive effect which is cryptic in the intact molecule. This fragment includes the RGD sequence of the third type 111 repeat (Fig. 17b). Interestingly, a human RGD-dependent tenascin binding integrin has been obtained from glioma cells by affinity chromatography of cell extracts on a tenascin column [321] suggesting that tenascin fragments produced in oioo may have adhesion promoting rather than adhesion inhibiting effects, at least on some cells. 4.5. SPARUosteonectin

This 43 kDa glycoprotein was originally described as osteonectin, a non-collagenous major component of calcifying tissue capable of binding calcium, collagen and hydroxyapatite and believed to play a crucial role in bone mineralization [322,323] and a marker of differentiated bone cells [324]. However, the glycoprotein is more widely expressed in developing tissues and is produced and secreted from mouse parietal endodermal cells or proliferating endothelial cells, where it was called SPARC for Secreted protein, acidic and rich in cysteine [325,326]. The glycoprotein is also a major constituent of basement membranes, for example of the EHS tumor as an acidic, EDTA-extractable component BM40 [327]. The glycoprotein (Fig. 17c,d) has a multi-domain structure [328] which contains in the N-terminal domain five acidic calcium-binding motifs. Cations enhance binding of the glycoprotein to collagen type IV and presumably are essential for integration of the glycoprotein into basement membranes [329]. A relatively less helical domain (Fig. 17d) following the major calcium binding domain contains the single N-glycosylation site which is conserved across species. This domain bears significant homology to follistatin, an inhibitor of growth factors such as activin, and to serpin-type protease inhibitors. The third domain is largely a-helical and the fourth C-terminal domain contains an additional calcium binding site of the EF-hand type (Fig. 17c,d). The unglycosylated polypeptide has a predicted molecular mass of -33 kDa. Structural analysis of the glycoprotein secreted from parietal endodermal cells showed that the N-glycans comprised predominantly (86%) diantennary complex-type oligosaccharides with terminal sialic acid and core fucose residues [330]. In addition smaller amounts of hybrid-type (5%) and oligomannosidic-type (9%) glycans were detected. As with most glycoproteins, the glycosylation of SPARC/osteonectin appears to be cell-type-specific. For example, the glycoprotein secreted from thrombin activated platelets does not bind to concanavalin A affinity columns, contrary to what would be expected for glycoproteins carrying diantennary glycans. However, it did bind to Lens culinaris lectin columns indicating the presence of triantennary core fucosylated complex-type glycans [33 13. Strikingly, the glycoprotein extracted by EDTA from cortical bone bound tightly to Con A columns and the N-glycans were exclusively endo-H-sensitive, indicating an oligomannosidic- or hybrid-type structure [33 11. Solid state binding assays showed that

545

while the bone-extracted glycoprotein exhibited high affinity binding to collagens I, 111 and V, platelet-derived glycoprotein failed to bind [33 11. However, the product secreted from parietal endodermal cells binds various collagens [332] suggesting that if indeed different glycosylation is involved in modulating collagen-binding activities of SPARC/osteonectin, the effect must be rather subtle and controlled by the degree of branching of complex-type N-glycans. The function of SPARC/osteonectin is believed to be to mediate the mineralization of collagen by affinities of the glycoprotein for calcium and phosphate ions and hydroxyapatite as well as collagen. The role in non-ossified tissues is unknown. The glycoprotein has not been shown to be a cell attachment protein but inhibits the spreading on collagen substrata of endothelial cells and fibroblasts [332] and disrupts polarized endothelial cell layers [328]. It remains to be determined if these effects act directly on adhesive interactions or are indirect since the glycoprotein, when added exogenously to cultured synovial fibroblasts is reported to induce the secretion of several metalloproteinases involved in collagenous matrix turnover [333]. Interestingly, a peptide sequence in the a-helical third domain (Fig. 17c) retains this inducing activity. 4.6. Osteopontin Osteopontin was described as a non-collagenous glycoprotein in the mineralized bone matrix [334,335]. Sequencing showed identity with a component secreted from normal rat kidney (NRK) cells and other cells stimulated by tumor promoters[336] and a transformation-associated antigen [337]. The polypeptide has a predicted molecular mass of -32 kDa but behaves anomalously on SDS-polyacrylamide gel electrophoresis where it runs variously as a -43-75 kDa component [336], due in part to the presence of a single N-glycan and several 0-glycans, protein phosphorylation and a high content of acid amino-acids [334,336,338]. The glycoprotein when coated onto a substratum mediates the attachment and spreading of cells through an RGD sequence in osteopontin and an integrin, probably avfJ3 [339], on the responding cells. In bone the primary function of osteopontin is believed to be the adhesion of osteoclasts to sites of bone integrin [341]. resorption [340] where it has been shown to co-localize with the Osteopontin also binds to fibronectin at an unknown site. Interestingly, both protein phosphorylation and glycosylation appear to modulate this interaction. Binding to the pericellular fibronectin-rich matrix of NRK cells was found only for a phosphorylated and non-glycosylated isoform whereas binding to soluble plasma fibronectin was shown by the non-phosphorylated and N-glycosylated isoform [338]. Furthermore, the latter binding was abolished by removal of the N-glycan from osteopontin showing either a direct involvement of the glycan or an indirect effect on maintaining an active conformation for binding to plasma fibronectin.

5. Integrins 5.1. Structure and specijicity The name integrin was introduced to describe an integral membrane glycoprotein receptor for cell adhesion to fibronectin identified by a monoclonal antibody which was shown to

546

a 4

I

aI , I

P 7

D 4

P 2

Fig. 18. Subunits and assembly of integrins. Seven p subunits and fourteen a subunits are shown which can form in various combinations to produce dimeric, non-covalently bonded complexes with various ligand binding properties, some of which are indicated in boxes. Abbreviations: Fn, fibronectin; Lm, laminin; coll, collagen; Fb, fibrinogen; Op, osteopontin; ICAM, ICAM family I or 2; BM, unidentified basement membrane component(s). a, is also called p150 or CDI Ic; aL and aM are CDI l a and CDllb, respectively; p2 is CD18 in the immunological literature.

block cell-fibronectin interactions [342]. The antibody was used to purify by expression cloning a cDNA encoding what later became known as the @ I subunit of the avian receptor. At about the same time a mammalian fibronectin receptor was identified by affinity chromatography of cell extracts on immobilized fibronectin fragments [343]. Very quickly thereafter the sequences of both the a and fi subunits of the human fibronectin receptor were reported [344,345]. Subsequent studies have shown the existence of a rather large family of related integrin receptors which has been extensively reviewed [346-35 11. At present at least fourteen a subunits and eight fi subunits have been identified. Although the a and fi subunits are dissimilar, the various subunits within either the a or fi group show 40-50% homologies at the protein level and often considerably more. The integrins are heterodimers consisting of a and fi subunits that associate in various combinations non-covalently. Consequently, at the protein level the number of functional integrins is rather large as indicated in Fig. 18. The a subunits vary in size between 120 and 180kDa and the @ subunits (except the unique 6 4 subunit) between 90 and 110 kDa. Additional variation is generated by the alternative splicing of several subunits in their cytoplasmic domains including @ I [352,353], p3 [354], (34 [355-3571, a3 [358], a6 [359,360] and allb [361]. The significance of these splicing variants is unknown at present but is believed to involve modulation of integrin affinity for their respective

547

ligands, transduction of extracellular signals and interaction with the cytoskeleton in maturing adhesive complexes [362-3651. Typically, the integrin a and f3 subunits are glycosylated polypeptides that contain a short cytoplasmic segment of about thirty amino-acid residues, a transmembrane segment of similar length and a large extracellular domain of unique structure (Fig. 19a). The a subunits contain in the extracellular portion several, usually four, aspartic-acid-rich sequences that are responsible for cation binding which is essential for ligand binding of the fully assembled integrin (Fig. 19b). Several a subunits (a3, as, a6, a l l b and a v ) also are proteolytically cleaved within a sequence enclosed by an intra-subunit disulfide bond. Proteolytic cleavage gives a light 25-30 kDa chain which contains the cytoplasmic and transmembrane domains, and a heavy chain containing most of the extracellular portion of the subunit. The reason why some a subunits undergo post-translational proteolytic processing is unknown. Other a subunits contain an insert of 180-200 aminoacid residues, the I-domain, that is related to sequence repeats found in certain collagens. The significance of the I-domain is unknown but since at least two of the integrins (alp1 and a$I) containing an I-domain bind collagen it is possible that the I-domain provides a homophilic ligand binding activity. However, other integrins with I-domain, such as the aL-,a M - and axbzintegrins, have not been shown as yet to bind collagens. The f3 subunits contain within the extracellular portion four cysteine-rich sequences believed to be folded by extensive internal disulfide bonds and which together with the N-terminal domain of the a subunit form the ligand binding site(s) (Fig. 19b). The (34 subunit is so far the only exception to the domain structure described, in that it contains a large -1000 residue cytoplasmic extension (Fig. 19a). The 8 4 subunit is commonly found in combination with an a6 subunit at hemidesmosomes in epithelial cells [36&368] suggesting that its cytoplasmic extension is involved in interactions with intermediate filaments which accumulate at such sites. By contrast the shorter cytoplasmic domains of other fi subunits may, at least in certain af3combinations, associate with actin-based filaments at points of cell-substratum or other contacts [369]. The a and p subunits are extensively glycosylated. For example, the fibronectin-binding integrin a& contains 13-14 potential N-glycosylation sites. Graded digestion of the a5 and subunits with N-glycanase indicate that at least 7-8 of these sites are utilized [370]. This is in agreement with metabolic labelling studies which show extensive glycosylation and maturation of N-glycans from endo-H-sensitive to endo-H-resistant, complex-type structures [346]. Functional integrins are formed by non-covalent associations of a and f3 subunits. Electron micrographs of purified integrins al1b/b3 [371] and a5bI [372] show that the dimer forms a globular head (approximately 80 A x 12G-150A), supported by two clearly separated stalks extending to the membrane integration and cytoplasmic domains (Fig. 19b). Electron microscopy [373], chemical cross-linking data and site-directed mutagenesis studies locate the ligand-binding site in the globular head at the a,b subunit contact area [347,349]. These contacts include the cation-binding domains of the a subunit, which presumably contribute to maintaining the binding conformation, and several of the N-glycosylation sites. However, the role of the latter in assembly or function of the binding site is not clear at present. There is increasing evidence however that assembly of a@dimers is not always sufficient for high affinity binding to ligands. The

548 Plasma membrane Proleolytic

I

p ,,‘

4

1COOH

(b) 80 8,

l4

Ligand binding

280 A

Stalk

Proteolysis

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

U

f-

Membrane integration

f--.

Cytoplasmic domain

Fig. 19. Structure of integrins. (a) Generalized domain structure as illustrated by the fibronectin receptor,

asplintegrin. The structure consists of two polypeptides, that form a non-covalently bonded dimeric complex. The a, subunit contains a short cytoplasmic domain, a transmembrane segment and a large extracellular domain which typically includes four cation-binding sites (thick lines). Post-translational proteolytic processing of the

a3,as, a6, allb and a, subunits occurs at a conserved site indicated to produce a heavy chain (-140 kDa) and a light chain (-22 kDa) held together by intrachain disulfide bonding. Certain a subunits including a , , a2,aL,ax and a,,,, also contain a 200 amino-acid inserted segment, the I-domain (shaded rectangle) near the N-terminus. The putative N-glycosylation sites (solid circles) number fourteen in the a, subunit not all of which are conserved in other subunits. The bI subunit contains a short cytoplasmic domain, a transmembrane segment and a larger extracellular domain which includes several (at least four) cystine-rich domains (hatched squares) and thirteen putative N-glycosylation sites. Uniquely (so far) the fi4 subunit contains a large cytoplasmic extension. (b) Dimensions of the asp, integrin deduced from electron micrographs of the purified receptor. The proteolytic site of the as subunit is shown. The cation-binding sites of the a , subunit appear to be located in the globular head (shaded area) in association with the PI subunit and are believed to be crucial for subunit interactions and ligand binding.

549

clearest example is integrin aIIb/fi3 which is present on platelets and binds fibrinogen (Fig. 18a) with high affinity only after induction of the cells by agonists such as thrombin [374,375]. Conversion to high affinity binding states have also been reported for aM(32 [376] and a5fil [377,378] integrins. The mechanisms involved in integrin activation are not known at present but in certain cases appear to involve conformational changes as revealed by interactions with monoclonal antibodies recognizing conformational-sensitive epitopes. Some of these antibodies can actually induce an activated state in target integrins [379,380], one of which binds to the I-domain of a$:! 113811, possibly by binding to the relevant epitope and stabilization of an active conformation. Other mechanisms for integrin activation undoubtedly exist however, including phosphorylation at cytoplasmic domains of fi subunits and the action of lipid modulators [382]. One of the activating factors appears to be gangliosides which interact specifically with integrins. It was first shown that GD2 and GD3 gangliosides are associated with the avo3 integrin in human melanoma cells [383]. Previous studies had shown that ganglioside GM3 was localized with fibronectin and presumably the a5P1 integrin at focal contacts between fibroblasts and a substratum [384]. More recently it was found that a ganglioside-deficient cell line interacted very weakly with a fibronectincoated substratum, whereas a variant of the cells selected for a high ganglioside GM3 content interacted strongly 13851. The two cell types expressed similar amounts of both major fibronectin-binding integrins, a& and a&, strongly supporting an effect of the ganglioside on integrin activation. Liposomes incorporating a partially active a5PI integrin was stimulated in adhesion to a fibronectin-coated substratum by inclusion into the liposomes of GM3 ganglioside. Interestingly, a threshold effect was observed in that the enhancement induced by GM3 was abolished at high concentrations of the ganglioside. Somewhat similar findings were reported for another cell line expressing integrin avB3[386]. The site(s) in the integrin dimer which interact with ganglioside needs to be identified. The ligand binding specificity of integrins depends on the specific combinations of a and f3 subunits as indicated in Fig. 18. The ligands fall into three major classes, namely: extracellular matrix components such as laminin, collagen and fibronectin; plasma proteins such as fibrinogen or complement factors; and cell surface glycoproteins such as ICAM, VCAM and osteopontin. The binding specificity of particular af3combinations is well exemplified by the properties of the leukocyte integrins a ~ f 3 2and a& in that only the latter binds to both ICAM-1 and ICAM-2, whereas binding of a~f32is confined to ICAM-1 (Fig. 18). This pair of integrins also illustrates another important feature of binding specificity. Although the same ligand, in this case ICAM-1, may bind multiple integrins the recognition sites are completely different. Thus the binding site of ICAM-1 for a~(32 lies in the third Ig-like domain whereas aLf& binds to the first Ig-like domain [143,144]. Similarly, alf3,and a2Plare collagen receptors on a wide variety of cell types but bind at distinct regions of the collagen al subunit [387,388]. Other examples are integrins alp1,aspl and which bind to distinct sites on laminin (Fig. 15) and integrins a5f31 and a& which bind (Fig. 11) respectively to an RGD-containing sequence in a type 111 repeat of fibronectin and within an alternatively spliced sequence near the C-terminus of the fibronectin subunit. Another general feature of the integrins is the ability of a specific combination to bind several, apparently completely different,

550

ligands. For example, the a3p1 integrin binds laminin, fibronectin, collagen and also as recently shown entactidnidogen [295] and a novel laminin-related matrix component of keratinocytes called epilgrin [3891. Finally, the binding specificity of a complex containing the same a subunit may be controlled by the partner p subunit as indicated by the significantly different properties of the avPl and avo3 integrins (Fig. 18; see also refs. [347,390,391]. The determinants of these differences in binding specificities are unknown but could involve different oligosaccharide structures, in as much as the pattern of N-glycan processing of a particular subunit can be modulated by its binding partner as discussed in the next section. Recently, the belief that the (31 integrins act predominantly to promote cell interactions with extracellular matrix components has been questioned by increasing evidence for a role of these integrins in cell-cell interactions. The a3B1integrin has been identified at cell-cell contact sites 1392,3931 and several other integrins have been located to intercellular junctions both in cell cultures and in uiuo [394]. It has been shown that the integrin mediates intercellular adhesion of keratinocytes through aspl to azpl interactions [395]. The a361 integrin also appears to interact homophilically [396]. An affinity matrix containing purified a3p1 showed specific binding of a3PIand such binding was cation requiring, indicating an involvement of the ligand binding globular domain. The a3pl matrix showed no affinity for a561 or avP3indicating the specificity of the interaction. It remains to be determined how important such homotypic or heterotypic integrin interactions are in cell-cell adhesion and how they coordinate with intercellular adhesions mediated by other homophilic cell adhesion molecules such as cadherins. 5.2. Biosynthesis The biosynthesis and assembly of integrins must accommodate several intriguing aspects of their expression and function in different cell types. First the pattern of integrin expression in a particular cell type is usually complex and often involves the association of different a subunits with the same 0 subunit. The cells of many tissues for example express high levels of alp1,azpl and a& integrins involved in cell adhesions to collagens and laminin but relatively low amounts of the fibronectin receptor a& integrin. Similarly most leukocytes express the aLfiz integrin whereas aMb2and axB2expression is more cell-specific. Secondly, the relative expression of integrins sharing a common 6 subunit is developmentally regulated or altered drastically in response to changed states such as inflammation or wound healing, the latter often being associated with integrins, for example ref. [397]. up-regulation of a50l integrin compared with other Differentiation of monocytes into tissue macrophages is characterized by down-regulation of aLPzexpression and a massive up-regulation of axP2integrin [382,398]. These changes in integrin expression are tightly controlled in response to changing requirements for specific adhesive interactions. One of the best models to illustrate this point is the epidermis, a stratified epithelium. Keratinocytes migrating from the proliferative basal layer show decreased adhesiveness to basement membrane components such as laminin and collagen type IV and to fibronectin which is reflected by reduced expression of aZfi1, a3p1and a501 integrins [394].

55 1

Little is known at present regarding the mechanisms regulating the surface expression of individual integrins and control of the integrin repertoire shown by different cell types under different circumstances. It is clear however that the mechanisms are diverse and can occur over very different time scales, ranging from the very rapid up-regulation of surface expression of specific integrins, for example in inflammatory cells, to much slower events. Some of the rapid responses in integrin surface expression are explained by significant intracellular pools and induced transport from these pools to the cell surface. For example, in circulating granulocytes and monocytes pre-assembled aMp2and are present in intracellular granules which upon activation fuse with the plasma membrane [399,400]. Interestingly, the other major leukocyte integrin a& is not held in intracellular granules suggesting that the targeting signal for transport of the integrins to intracellular storage sites resides in the a subunits, or to differences in post-translational modifications, for example subunit glycosylation. The a and (3 integrin subunits are synthesized in the endoplasmic reticulum from independent precursors in all classes so far investigated including the (32-associated integrins [401,402], the aIIb/(33integrin [403-405] and the (31 -associated integrins [406,407]. Subunit biosynthesis is under transcriptional and translational control that can be modulated by, for example transforming growth factor (3, (TGFP1) [406-408]. Interestingly, TGF(31 up-regulates synthesis of the aLbut not of the (32 subunit [408] again confirming the independence of a and (3 subunit biosynthesis within the same cell. Another indication of the uncoupled biosynthesis of a and (3 subunits is the presence intracellularly of a large excess pool of (3 subunit, most clearly shown in the case of the (31 subunit and assembly of the fibronectin integrin a5(3l [409-411]. At this stage both a5 and (31 subunits carry oligomannosidic N-glycans sensitive to endo-H (Fig. 20). As judged by the differences in apparent molecular size of the subunits estimated from SDS-polyacrylamide gel electrophoresis, the predicted polypeptide sizes and assuming a medial N-glycan size of about 2.5 kDa, it appears that approximately five to six N-glycans may be present on each subunit out of the full complement of potential N-glycosylation sites present in these subunits (Fig. 19a) in agreement with results obtained by N-glycanase treatment of the mature integrin[370]. Assembly of a(3 dimers takes place in the endoplasmic reticulum before oligosaccharide processing takes place to generate endo H-resistant N-glycans (Fig. 20), analogous to the assembly of other multimeric complexes such as the T-cell receptor and class I or I1 histocompatibility antigens. Similarly, association of a(3 complexes appears to be critical for further oligosaccharide processing and intracellular transport of the (32 integrins [401]. There is evidence that the kinetics of integrin assembly is rate-limited by the availability of appropriate a-subunits and that unutilized (3-subunits undergo degradation, apparently in a non-lysosomal compartment (Fig. 20) [412]. Assembly of the a5(3l dimer appears to require some maturation event in the PI subunit. Pulse chase experiments have identified a (31 intermediate which differs from the (31 -subunit in fully assembled aspl complexes in electrophoretic mobility [406]. The difference between the two forms can be abolished by reduction, indicating that the putative maturation step involves progressive formation of disulfide bonds presumably within the cysteine-rich domains of the PI -subunit. The factors controlling (3-subunit maturation are unknown but presumably act in step with provision of a-subunits. Since the latter pool is regulated by factors such as TGF(3l it is possible that the maturation

552 Endoplasrnic

reticulum

-*

Golgi Cations (1: Ca"

Glycolipids GD 2

-

Plasma membrane

TGF-P

I

-

integrin

I I

Processing reactions glucosidases mannosidase

v Maturation compartment

Degradation Compartment

Mannosidase I

GlcNAcTI Mannosidase II

+ 29":'

Fig. 20. Assembly of an integrin. The a and 0 subunits are N-glycosylated with immature oligosaccharides post-translationally in the endoplasmic reticulum. Subunit transcription can be regulated by various effectors including transforming growth factor 0 (TGF-p). The 0 subunit is usually in large excess relative to expressed a subunits and undergoes polypeptide maturation which also is regulated. Assembly of a0 dimers occurs before extensive processing of the N-glycans and undimerized 0 subunits may be degraded. Subsequent transport of afl subunits through the Golgi apparatus to the cell surface is accompanied by terminal N-glycan processing, association of dimeric a0 structures with cations and with ganglioside in certain cases. The steps illustrated are based on limited data obtained for very few integrins, hence are speculative and may not be followed by all integrins. Furthermore, full activation of surface-expressed integrins may require additional steps as discussed in the text.

of (3-subunits or translocation of active (3-subunits to sites containing the a-subunits are coordinately regulated by effector molecules such as T G F ~ as I indicated in Fig. 20. An early event in integrin assembly appears to be binding of cations [413] which occurs before extensive processing of N-glycans in either subunit. The importance of cation binding for stable a(3 dimer formation is shown by site-directed mutagenesis experiments to inactivate one of the cation binding domains of the a subunit [413]. Although the mutant a subunit formed an arlb(33 complex the integrin failed to exit the endoplasmic reticulum suggesting that additional folding events in integrin maturation were prevented with a consequent block in intracellular transport. Mutations in the cation binding domains of the a I l b subunit are associated with Glanzmann thrombasthemia, a rare bleeding disorder characterized by a failure of platelets to express al&3 integrin at the cell surface and a defect in an ability to bind fibrinogen and participate in formation of the hemostatic plug [4 131. Later events in integrin assembly involve transport of the complex through the Golgi apparatus where oligosaccharide processing and proteolytic cleavage of certain a subunits [4 141 takes place. At a late stage in intracellular transport, association of certain integrins,

553

for example the av63 integrin, with gangliosides must take place, in as much as formation of polysialylated gangliosides, such as GDI and GD2 which preferentially combine with integrins, is known to occur in trans-Golgi compartments [415]. Finally, the assembled integrin is transported to the cell surface where in most cases it exists in a resting state and becomes further activated by diverse mechanisms as discussed previously. Evidence for the scheme of integrin assembly shown in Fig. 20 is fragmentary and the model may not be applicable to all integrins. Nevertheless, additional evidence on the glycosylation of integrin subunits is compatible with the scheme. Assembly of the fibronectin receptor a& has been studied in Ricl4 cells, a BHK cell mutant lacking N-acetylglucosaminyl transferase I and hence unable to modify N-glycans beyond the MansGlcNAc2Asn stage [370]. In these cells the precursor (31 subunit was found to carry N-glycans larger than the fully processed MansGlcNAc2 glycan of the mature subunit indicating that the bulk precursor pool had not been translocated into the cis-Golgi compartment containing mannosidase I. This result clearly shows that the putative maturation compartment holding 61 subunit precursors (Fig. 20) resides proximally to cis-Golgi compartments. The proposal that subunit association is an early event taking place in a pre-Golgi compartment (Fig. 20) is given strong support by the fact that (32 [4 161 and 6 3 [414] subunits are glycosylated differently when present in integrins containing distinct a subunits. This is presumably due to oligosaccharide processing constraints imposed by the structurally different a subunits on the (32 or 6 3 subunit glycosylation sites. Thus structural analysis of the oligosaccharides at specific N-glycosylation sites in the (32 subunit of the aLb2and aMb2integrins showed that the glycosylation patterns at four of the sites differed significantly according to the nature of the a subunit [416]. Since both integrins were synthesized in the same macrophage cell line these differences presumably cannot be due to different glycosylation potentials of the cell producing the individual integrins. They were interpreted [4 161 to reflect the influence of quaternary structure on the processing of the 6 2 subunit as a consequence of association with different a subunits and terminal oligosaccharide processing of already assembled integrin complexes. Although this interpretation is reasonable, some caution is warranted since the intracellular fate of the (32 integrins differ as discussed previously. It is possible that the different glycosylation of fi subunits complexed with various a subunits could result from separate trafficking pathways of different integrins, after subunit assembly in the endoplasmic reticulum and before completion of terminal oligosaccharide processing. Interestingly, four out of five glycosylation sites examined [416] in the (j2 subunit of aMP2integrin carry larger oligomannosidic glycans than the p 2 subunit of aLp2.The glycans of the latter are more sialylated and enriched in highly branched glycans. It is possible that these differences could reflect an early exit from Golgi compartments of integrins destined for transport into intracellular storage compartments. It would be interesting to determine the glycosylation of (32 subunits of integrins in tissue macrophages which do not contain intracellular pools of these receptors [399,400]. Information on the detailed structures of the N-glycans of fully assembled mature integrins is sparse, especially for a particular integrin expressed, unlike the 6 2 integrins, in different cell types where significant differences in glycosylation would be expected to occur. So far the most complete analysis of N-glycan structure in the integrins is available for the human leukocyte (32 integrins [417]. Oligosaccharides released by

554

hydrazinolysis from a mixed fraction of human lymphoid tissue aM-, a L - and integrins were characterized (Fig. 2 1). Several notable features are apparent. The major oligosaccharide fractions were of the conventional Man9-5GlcNAc2 type (3 8% of total) and diantennary complex-type glycans (17% of total) terminated with one or two NeuAc a 2 + 3 or 6 residues. However, a major fraction (18% of total) was shown to be a tetraantennary complex-type glycan variously containing outer-branch polylactosamine extensions terminated in NeuAc a 2 3 or 6 residues and with a Fuc a1 -i 3 GlcNAc modification on penultimate N-acetylglucosamine residues (Fig. 2 I A). Triantennary isomers containing similar terminal modifications accounted for 1 1% of total glycans. Most interestingly, a significant fraction (9% of total) contained a polylactosamine extension terminated by a type I Galfil + 3GlcNAc disaccharide (Fig. 21B). The polylactosamine extension terminated by the type I Galfil -+ 3GlcNAc disaccharide was not modified by the Fucal -i 3GlcNAc substituent. These structural features containing type I substituents are quite unique and determination of their distribution among the subunits of the various a subunits and fi2 subunit of the leukocyte integrins may provide intriguing insights into a potential function of carbohydrates in integrin trafficking and function. The presence of sulfated glycans in the fi2 integrins was suggested by metabolic labelling of mouse lymphoid cells with [35S]sulfate[418]. Although sulfated oligosaccharides are not major components of fi2 integrins [417] the sulfated structure recognized by the L2/HNK-l antibody has been detected in immunoblots of f i ~integrins, including the a5fil fibronectin receptor[419]. Both the a and fi subunits reacted with the antibody after SDS-polyacrylamide gel electrophoresis, but interestingly the epitope was cryptic in the native undenatured integrins. Since the L2/HNK-1 epitope has been implicated in several adhesive events mediated by other molecules expressing this antigenic determinant, its presence in the integrin family is an intriguing finding of possible functional significance, especially in heterotypic interactions between integrins and other families of adhesion molecules. The question of 0-glycosylation of integrins has not been examined closely. However, one interesting study [420] has shown unequivocally the presence of a single 0-glycosylation site in a mutant a l l b subunit. This mutation is an Ileu + Ser replacement in the heavy chain of the a l I b subunit which is recognized immunologically by expression of the Baka (NPA-3a) allo-antigen. The Baka determinant requires post-translational processing of the a I I b subunit for expression, and immunoreactivity is sensitive to sialidase treatment indicating an involvement of carbohydrate. Presumably, since a normal phenotype persists in the patients the mutated a I I b subunit carrying the 0-glycan interacts normally to form an integrin functional in fibrinogen binding. ---f

5.3. Roles of carbohydrate The evidence is increasing that glycosylation of integrins plays important modulatory roles in the function of these molecules. In other studies it appears that the glycosylation of at least one extracellular matrix glycoprotein, namely laminin, is important in cellular adhesion mediated by integrins. So far most information on the role of carbohydrates in integrin-mediated adhesive interactions has been obtained for the fibronectin receptors

555

a& and a&. As discussed earlier, terminal oligosaccharide processing of a3- and pl-subunits appears not to be essential for a5b1 assembly, transport to the cell surface nor fibronectin binding since a fully active integrin was synthesized and expressed by Ricl4 mutant cells [370]. Interestingly, Ricl4 cells were shown to be grossly defective in adhesion to fibronectin-coated substrata [23 1,2321. However, the defect in N-acetylglucosaminyl transferase I in Ricl4 cells leads to pleiotropic effects, including synthesis and surface expression of under-sulfated heparan sulfate proteoglycans [42 11. This altered proteoglycan was shown to have reduced affinity for fibronectin and to turn over more rapidly than the fully sulfated proteoglycan of BHK cells. A relation between lowered sulfation of heparan sulfate proteoglycans and reduced affinity for fibronectin has been reported in other systems [422]. This may well account for the observed decreased adhesion to fibronectin by Ricl4 cells since high efficiency adhesion requires both the cell binding domain, recognized by asplintegrin, and at least one heparin binding domain of fibronectin as discussed earlier. Additional evidence that terminal oligosaccharide processing is not required for functional a3filintegrin comes from the isolation of the receptor from extracts of cells treated with Swainsonine [423], an inhibitor which induces an increase in hybrid-type N-glycans in glycoproteins. In mouse fibroblasts however it appears that a&, integrin in cells treated with 1-deoxymannojirimycin, an a-mannosidase IAAB inhibitor, fails to bind to fibronectin [423]. The inhibitor did not prevent association of a and p subunits, transport of the complex to the cell surface nor surface receptor density. These interesting findings could suggest that a limited amount of mannosidase processing, perhaps down to MansGlcNAc:! oligosaccharides, is required for assembly and maturation of a potentially functional aspl integrin. In BHK cells, 1-deoxymannojirimycin does not block formation of functional aspl integrin [370]. However, this inhibitor may not be equally effective in preventing N-glycan processing in all cells. In BHK cells treated with the inhibitor it was found that extensive mannosidase processing still occurred, presumably through alternative pathways involving inhibitor-insensitive processing mannosidases [424]. Other work indicates that terminal sialylation of integrin subunits can enhance ligand binding affinity. Desialylation of cell surface glycoproteins leads to decreased fibronectin binding of melanoma cells [425,426] and the hematopoietic K562 cell line [427]. Treatment of K562 cells [427] and other leukaemic cell lines [428] with phorbol esters also leads to reduced adhesion to fibronectin and decreased sialylation of integrin a5bi subunits. In human fibroblasts it is reported [423] that only the as subunit, and not the subunit is sensitive to sialidase suggesting that the effects of phorbol ester may be to initiate sialylation of the fi1 subunit as well as over-sialylation of the a subunit. However, no direct structural data are available on the integrin subunits from untreated or treated cells to support this possibility. A close relationship between sialylation of integrin subunits and fibronectin binding has also been reported in metastatic and non-metastatic melanoma cells [429]. A weakly metastatic, wheat-germagglutinin-resistant mutant Wa4-b 1 was found to be less adherent to endothelial cell extracellular matrix than the highly metastatic parent B- 16 melanoma cells, presumably an important determining factor in metastatic potential. The integrins from both cell types contained similar amounts of oligomannosidic-type glycans and complex-type glycans [430]. However, the tri- and tetraantennary glycans in the mutant integrins were

556

relatively poorly sialylated and contained instead terminal Gal(3I-4[Fuca l-3]GlcNAc structures. Thus, the under-sialylation of mutant integrins may be due to expression of a fucosyl transferase which modifies N-acetyllactosamine termini and blocks subsequent sialylation, resulting in resistance to the selective lectin. The role of N-glycans appears to differ in other integrins. As discussed earlier a 4 B 1 integrin binds to an alternatively spliced V-domain of fibronectin (Fig. 11) and its binding is blocked by a synthetic peptide based on a region CS-I within this domain [201,347]. Functional a& integrin could be isolated from cells inhibited in N-glycan processing with either Swainsonine or 1-deoxymannojirimycin and affinity for CS-I peptide was similar to a4(31 integrin from control cells[431]. Monensin treatment of cells, which also prevents N-glycan processing, similarly had no effect on expression of a 4 r j l integrin with binding affinity for CS-I peptide. Further, the (31 subunit of a& integrin in immature mouse thymocytes is reported to contain under-processed N-glycans and less sialic acid than from mature thymocytes although the former cells adhere more efficiently to fibronectin [432]. Therefore it appears that early Golgi processing modifications of N-glycans are not required for efficient interactions of integrin a& with the V-domain, in contrast to data obtained for aspl integrin binding to the RGD segment. Similarly, the RGD-dependent interaction of a& integrin is also apparently not dependent on N-glycan processing [433]. On the other hand, glycosylation of the (31 subunit, possibly sialylation, appears to be important for the function of an integrin, in myelination [434]. As discussed previously, the (32 integrins in human peripheral lymphocytes (mainly T-cells) contain significant amounts of the non-sialylated Gal@14[FucaI-3]GlcNAc terminal sequences (Fig. 2 1). This structure may be distributed preferentially on a M ( 3 2 integrins, as shown by the increased relative amounts of terminal galactose on the (32 subunit of a M B 2 compared with aLP2integrin[418]. Recently, a clear role of glycosylation in ICAM-I interactions with (32 integrins has been demonstrated, specifically the nature of the N-glycans in the third Ig-like domain of ICAM-I (Fig. 10a) which contains the binding site for a ~ f 3 2 . Deletion of N-glycosylation sites in this domain enhanced binding to purified a ~ ( 3 2and mouse fibroblasts transfectants expressing ICAM-I showed enhanced adhesion to purified a M P 2 after treatment of the cells with sialidase or with 1 -deoxymannojirimycin [ 1441. Similarly, ICAM-1 produced in the Baculovirus insect cell expression system and hence carrying only small oligomannosidic glycans, bound with enhanced affinity to neutrophils expressing aMc32 [ 1441. Taken together, these results show that the size of the N-glycans in the third Ig-like domain of ICAM-1 can modulate the interactions with a ~ ( 3 2 integrin at this site. The interactions with aLo2are not affected as expected since aLP2binds to the first Ig-like domain in ICAM-I. These findings may have functional significance in uiuo, in cells expressing several (j2 integrins or in mixed cell populations differentially expressing separate 8 2 integrins. Thus neutrophil adhesion to recombinant ICAM- 1 may depend primarily on a M B 2 if ICAM-I has smaller N-glycans but on aLP2if ICAM-I carries more complex-type glycans [ 1441. Similarly, neutrophil adhesion to endothelial cells proceeds mainly through ~ ~ ( interactions 3 2 after treatment of the endothelial cells with 1-deoxymannojirimycin [435]. It is interesting that ICAM- 1 isolated from different cell types differs in the extent of N-glycan processing [138]. This suggests that in uiuo

557

Galpl + 4GliNAq31 t Fucul \6

? Fucal (NeuSAca2 --t 3 or 6)0-2

Galpl

f

t Fucal f Galpl

k 4 Manal

* 4GlcNAcpl

f Fucal f

Fig. 21. Unique structures of the N-glycans of 4GlcNAc.

Manal

+ 4GlcNAcp1 f 2

6(3) Manpl + 413 3(6)

f2

3

fiz integnns. Data from Asada et al. [417]. R=GlcNAc(3-

ICAM- 1 glycosylation may regulate interactions with other cells carrying (32 integrins. For example, myeloid cells expressing a& integrins may be selected if ICAM-1 on the binding partner cells is under-glycosylated or contains poorly processed N-glycans, whereas lymphocytes may interact with other cells bearing fully glycosylated ICAM-I with highly processed N-glycans through aLP2integrin. In particular, it is possible that selectivity in leukocyte adhesion to sites of activated endothelium may be regulated at least in part by differential glycosylation patterns on ICAM-I during its up-regulation by different inflammatory mediators. Recently, some interesting experiments have suggested a role for carbohydrates in the integrin-mediated interactions of cells with laminin (Fig. 22). Initially it was found that concanavalin A inhibited the spreading but not attachment of mouse melanoma cells on mouse laminin [436,437]. Similarly, concanavalin A blocked neurite outgrowth but not attachment of a neuronal cell line, rat pheochromocytoma PC12. When these cells were plated out on laminin obtained from tunicamycin-treated cells they attached comparably to fully glycosylated laminin, but spreading and neurite outgrowth was inhibited [437]. Similar results were obtained with EHS laminin, partially deglycosylated by periodate oxidation [43 81. The predominant laminin-binding integrin appeared to be a& integrin [438] which binds to the E8 fragment of EHS laminin and as described earlier is known to promote neurite outgrowth in neuronal cells (Fig. IS). Laminin obtained from mouse tumor cells treated with castanospermine, an inhibitor of processing glucosidases, or kifunensine, a potent inhibitor of mannosidase IAIIB, when tested as a substratum for melanoma cell adhesion was found to be as effective (castanospermine)

558 Glycopeptide Lectin

J

Melanoma

H lntegrin

fi

4--

Mannose

(b)

Attachment

Glycosylated laminin

Spreading

Fig. 22. Speculative model of laminin mediated adhesion and spreading of mouse melanoma cells. Adapted from Chandrasekaran et al. [440,441]. Mouse melanoma cells attach to both unglycosylated (a) or fully glycosylated (b) laminin but spread only on glycosylated laminin. Spreading on unglycosylated laminin is triggered by binding of soluble oligomannosidic ligands, e.g. glycopeptides derived from laminin, to a melanoma cell surface lectin. An activation signal induced by occupation of the lectin site may be involved in the initiation of cell spreading, in this model by activation of integrin and enhancement of integrin-laminin interactions.

or more effective (kifunensine) than laminin obtained from untreated tumor cells. These results suggested that the essential oligosaccharides in laminin promoting cell spreading and neurite outgrowth were of the oligomannosidic-type [43944 11. Remarkably, when polyvalent oligomannosides were added, together with melanoma cells, to a substratum of unglycosylated laminin the cells attached and spread as well or almost as well as on fully glycosylated laminin. Oligomannosidic glycopeptides obtained from mouse tumor laminin or yeast mannan were found to be similarly active whereas complex-type glycopeptides were inactive [439441]. However, a minimum active structure appeared to be required since a Man3 GlcNAc2 oligosaccharide was inactive whereas Man6GlcNAc2 and larger oligosaccharides were equally effective [441]. These results suggest a duel receptor model for melanoma cell adhesion to laminin (Fig. 22). Cell attachment initially is formed by interactions of a& integrin with a binding site in the E8 fragment of laminin. However, spreading requires secondary events which appear to require ligation of the melanoma surface by an oligomannosidic glycan which can be presented in solution (Fig. 22a) or as a component of the laminin substratum (Fig. 22b). Since melanoma cells cannot attach to a substratum of polyvalent mannose BSA [440], it seems clear that ligation of the putative oligomannosidic-binding site on the melanoma cell surface is not in itself an adhesive event. Rather, it in some way potentiates interactions involving integrin-laminin binding. A putative mannose-binding lectin was detected on the surface of melanoma cells attached to unglycosylated laminin, which became unreactive to a

559

fluorescence mannan probe as the cells spread and then reappeared [440,441]. These results suggest strongly that the ligation of the melanoma lectin site by oligomannoside need only be transient to promote the signal required to trigger spreading. During spreading the fl1 integrin became recruited to the ventral surface of the cell in contact with the laminin substratum, as expected since presumably spreading is accompanied by multiplication of integrin-laminin interactions. Although many mechanisms are obviously possible it seems reasonable to propose (Fig. 22) that ligation of the lectin site by oligomannoside may induce some signal that promotes integrin affinity for laminin to mediate the spreading process. The precise location within the laminin molecule of the active oligomannosidic glycans is unknown. As described earlier the oligomannosidic glycans of EHS tumor contain six to eight mannose residues. There is indirect evidence to suggest [441] that the active fraction may reside in the long arm of laminin, although apparently not in that portion of the long arm containing the E8 fragment. The model (Fig. 22) may apply to cells other than mouse melanoma cells in that somewhat similar results were obtained for laminininduced neurite outgrowth of PC12 neuronal cells. However, many other cells do not respond to oligomannosidic glycans including fibroblasts, rhabdomyosarcoma cells and melanocytes [440]. Other plant lectins, unlike concanavalin A, block the initial attachment of melanoma cells or neuronal PC12 cells to fully glycosylated laminin. These include wheat germ agglutinin and Griflonia simplicijolia lectin [437,442]. The mechanism of this inhibition is unknown but appears to involve lectin binding to laminin oligosaccharides near to integrin reactive sites and steric interference of integrins with laminin polypeptide [436,437]. However, some preliminary evidence for the presence of a-galactosyl residues, which are recognized by Griflonia simplicijolia lectin, on the a6 integrin subunit in melanoma cells [443] may need a revision of this proposal. Interestingly, two endogenous mammalian lectins, galectins-1 and -3, also have been reported to block integrin-mediated cell adhesion to substratum, in particular laminin [281,4441. Detailed studies of their binding specificities [281,445] show that the minimum structure required for high affinity binding appears to be polylactosamine glycans. The lectins bind laminin from murine EHS tumor and fetal fibronectins which contain large amounts of polylactosamine glycans [28 11. The simplest interpretation of these results is that the endogenous mammalian lectins bind to polylactosamine glycans of laminin close to cell binding sites and block, by steric interference, interactions mediated by cell surface integrins [445-447]. Direct evidence that a mammalian lectin inhibits the interactions between EHS laminin and the major laminin-binding integrin a& of skeletal myoblasts has recently been reported [448].

References [ I ] Wilson, H.V. (1907) J. Exp. Zool. 5, 245-258. [2] Holtfreter, J. (1948) Ann. N.Y. Acad. Sci. 49, 709-760. [3] McNeill, H. and Nelson, W.J. (1992) In: T.P. Fleming (Ed.), Epithelial Organization and Development. Chapman and Hall, London, pp. 53-88. [4] Rodriguez-Boulan, E. and Zurzola, C. (1993) J. Cell Sci. 17, 9-12. [5] Schwarz, M.A., Owaribe, K., Kartenbeck, J. and Franke, W.W. (1990) Annu. Rev. Cell Biol. 6 , 4 6 1 4 9 1 .

560 [6] [7] [8] [9] [lo] [Ill [I21 [13] [I41 [15] [16] [I71 [I81 [I91 [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49]

Legan, P.K., Collins, J.E. and Garrod, D.R. (1992) BioEssays 14, 385-393. Adams, J.C. and Watt, EM. (1993) Development 17, 1183-1 198. Stamatoglou, S. and Hughes, R.C. (1994) FASEB J. 8, 420427. Luna, E.J. and Hill, A.L. (1992) Science 258, 955-964. Noelken, M.E., and Hudson, B. (1995) In: J. Montreuil, J I G . Vliegenthart and H. Schachter (Eds.), Glycoproteins, New Comprehensive Biochemistry, Vol. 29A. Elsevier, Amsterdam, ch. 12. Silbert, J.E., Bernfield, M. and Kokenyesi, R. (1997) ch. I , this volume. Feizi, T. (1997) ch. 15, this volume. Varki, A. (1994) Proc. Natl. Acad. Sci. USA 91, 7390-7397. Sharon, N. and Lis, H. (1997) ch. 13, this volume. Takeichi, M. (1991) Science 251, 1451-1455. Kemler, R. (1992) Sem. Cell Biol. 3, 149-155. Gallin, W.J., Sorkin, B.C. and Edelman, G. and Cunningham, B.A. (1987) Proc. Natl. Acad. Sci. USA 84, 2808-2812. Shirayoshi, Y., Hatta, K., Hosoda, M., Tsunasawa, S., Sakiyama, F. and Takeichi, M. (1986) EMBO J. 5, 2485-2488. Nose, A,, Nagafuchi, A. and Takeichi, M. (1987) EMBO J. 6, 3655-3661. Ozawa, M., Engel, J. and Kemler, R. (1990) Cell 63, 1033-1038. Nose, A,, Tsuji, K. and Takeichi, M. (1990) Cell 61, 147-155. Ozawa, M., Hoschutzky, H., Kerrenknecht, K. and Kemler, R. (1991) Mech. Dev. 33, 49-56. Ozawa, M., Baribault, H. and Kemler, R. (1989) EMBO J. 8, 1711-1717. Nagafuchi, A,, Takeichi, M. and Tsukita, S. (1991) Cell 65, 849-857. Ozawa, M. and Kemler, R. (1992) J. Cell Biol. 116, 989-996. Nagafuchi, A,, Shirayoshi, Y., Okazaki, K., Yasuda, K. and Takeichi, M. (1987) Nature 329, 341-343. Mege, R.M., Matsuzaki, F., Gallin, W.J., Goldberg, J.I., Cunningham, B.A. and Edelman, G.M. (1988) Proc. Natl. Acad. Sci. USA 85, 7274-7278. Shirayoshi, Y., Nose, A,, Iwasaki, K. and Takeichi, M. (1986) Cell Struct. Funct. 1 1 , 245-252. Knudsen, K. (1985) J. Cell Biol. 101, 891-897. Balsamo, J., Pratt, R.S., Emmerling, M.R., Grunwald, G.B. and Lilien, J. (1986) J. Cell Biol. 32, 125141. Balsamo, J. and Lilien, J. (1990) J. Biol. Chem. 265, 2923-2928. Balsamo, J., Thiboldeaux, R., Swaminathan, N. and Lilien, J. (1991) J. Cell Biol. 113, 4 2 9 4 3 6 . Bauer, G.E., Balsami, J. and Lilien, J. (1992) J. Cell Sci. 103, 1235-1241. Scott, L.J.C., Balsamo, J., Sanes, J.R. and Lilien, J. (1990) J. Neurosci. LO, 346-350. Donalies, M., Cramer, M., Ringwald, M. and Starzinski-Powitz, A. (1991) Proc. Natl. Acad. Sci. USA 88, 8024-8028. Moore, R. and Walsh, F.J. (1993) Development 117, 1409-1420. Inuzuka, H., Miyatani, S. and Takeichi, M. (1991) Neuron 7, 69-79. Rantsch, B. and Doors-Zimmerman, M.T. (1991) Neuron 7, 3 9 1 4 0 2 . Buxton, R.S. and Magee, A.I. (1992) Sem. Cell Biol. 3, 157-167. Garrod, D.R. (1993) Curr. Opin. Cell Biol. 5, 3 0 4 0 . Magee, A.1. and Buxton, R.S. (1991) Curr. Opin. Cell Biol. 3, 854-861. Ren, J., Hamada, J., Takeichi, N., Fujikawa, S. and Kobayashi, H. (1990) Cancer Res. 50, 358-362. Green, K.J., Stappenbeck, T.H., Noguchi, S., Oyasu, R. and Niles, N.A. (1991) Exp. Cell Res. 193, 134-143. Farquhar, M.G. and Palade, G.E. (1963) J. Cell Biol. 17, 375412. Butz, S., Stappert, J., Weissig, H. and Kemler, R. (1992) Science 257, 1142-1 144. Penn, E.J., Hobson, C., Rees, D.A. and Magee, A.I. (1987) J. Cell Biol. 105, 57-68. Penn, E.J., Burdett, I.D.J., Hobson, C., Magee, A.I. and Rees, D.A. (1987) J. Cell Biol. 103,2327-2334. Buxton, R.S., Cowin, P., Franke, W.W., Garrod, D.R., Green, K.J., King, LA., Koch, P.J., Magee, A.I., Rees, D.A., Stanley, J.R. and Steinberg, M.J. (1993) J. Cell Biol. 121, 4 8 1 4 8 3 . Wheeler, G.N., Parker, A.E., Thomas, C.L., Ataliotis, P., Poynter, D., Arnemann, J., Rutman, A.J.,

56 1 Pidsley, S.C., Watt, EM., Rees, D.A., Buxton, R.S. and Magee, A.I. (1991) Proc. Natl. Acad. Sci. USA 88, 4796-4800. [SO] Kapprell, H.-P., Cowin, P., Franke, W.W., Ponsting, H. and Opferkuch, H.J. (1985) Eur. J. Cell Biol. 36, 217-229. [51] Cowin, P., Mattey, P. and Garrod, D.R. (1984) J. Cell Sci. 70, 41-60. [52] Overton, J. (1982) Dev. Biol. 92, 66-72. [S3] Williams, A.F. and Barclay, A.N. (1988) Annu. Rev. Immunol. 6, 381406. [54] Yamashita, K. and Kobata, A. (1996) In: J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins and Disease, New Comprehensive Biochemistry, Vol. 30. Elsevier, Amsterdam, ch. 3d. [55] Finne, J. (1997) ch. 3, this volume. [56] Cunningham, B.A. (1986) Trends Biochem. Sci. 1 1 , 4 2 3 4 2 6 . [57] Cunningham, B.A., Hemperley, J.J., Murray, B.A., Prediger, E.A., Brackenbury, R. and Edelman, G.M. (1987) Science 236, 799-804. [58] Thiery, J.P., Duband, J.L., Rutishauser, V. and Edelman, G.M. (1982) Proc. Natl. Acad. Sci. USA 79, 6737-6741. [59] Cremer, H., Lange, R., Christoph, A., Plomann, M . , Vopper, G., Roes, J., Brown, R., Baldwin, S., Kraemer, P., Scheff, S., Bartheis, D., Rajewsky, K. and Wille, W. (1994) Nature 367, 4 5 5 4 5 9 . [60] Goridis, C. and Brunet, J.-F. (1992) Sem. Cell Biol. 3, 189-197. [61] Dickson, G., Gower, H.J., Barton, H.C., Prentice, H.M., Elsom, VL., Moore, S.E., Cox, R.D., Quinn, C., Putt, W. and Walsh, F.S. (1987) Cell 50, 1 1 19-1 130. [62] Prieto, A.L., Crossin, K.L., Cunningham, B.A. and Edelman, G.M. (1989) Proc. Natl. Acad. Sci. USA 86, 9579-9583. [63] Small, S.J. and Akeson, R. (1990) J. Cell Biol. 1 1 1, 2089-2096. [64] Moore, S.E., Thompson, J., Kirkness, V., Dickson, J.G. and Walsh, FS. (1987) J. Cell Biol. 105, 13771386. [65] Hamshere, M., Dickson, G. and Eperon, 1. (1991) Nucl. Acid Res. 19, 4709-4716. [66] Roth, J. (1996) In: J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins and Disease, New Comprehensive Biochemistry, Vol. 30. Elsevier, Amsterdam, ch. 14. [67] Alcaraz, G. and Goridis, C. (1991) Eur. J. Cell Biol. 55, 165-173. [68] Finne, J., Bitter-Suerman, D., Goridis, C. and Finne, U. (1987) J. Immunol. 138, 44024407. [69] Livingstone, B.D., Jacobs, J.L., Glick, M.C. and Troy, F.A. (1988) J. Biol. Chem. 263, 9443-9448. [70] Rutishauser, U. and Landmesser, M. (1991) Trends Neurosci. 14, 528-532. [71] Rutishauser, U. (1992) Development (Suppl.), 99-104. [72] Joliot, A.H., Triller, A,, Volovich, M., Pernelle, C. and Prochiantz, A. (1991) New Biologist 3, 11211134. [73] Edelman, G.M. and Jones, F.S. (1993) J. Biol. Chem. 268, 20683-20686. [74] Moran, N. and Bock, E. (1988) FEBS Lett. 242, 121-124. [75] Cunningham, B.A., Hoffman, S . , Rutishauser, U., Hemperley, J.J. and Edelman, G.M. (1983) Proc. Natl. Acad. Sci. USA 80, 31 16-3120. [76] Ruishauser, U., Watanabe, M., Silver, J., Troy, F.A. and Vim, E.R. (1985) J. Cell Biol. 101, 1842-1849. [77] Rutishauser, U., Acheson, A,, Hall, A.K., Mann, D.M. and Sunshine, J. (1988) Science 240, 53-57. [78] Rao, Y., Wu, X.-E, Gariepy, J., Rutishauser, U. and Siu, C.-H. (1992) J. Cell Biol. 188, 937-949. [79] Frei, T., von Bohlen, E, Wille, W. and Schachner, M. (1992) J. Cell Biol. 118, 177-194. [80] Pollenberg, G.E. and Beck-Sickinger, A. (1993) Dev. Biol. 156, 324-340. [81] Murray, B.A. and Jensen, J.J. (1992) J. Cell Biol. 117, 1311-1320. [82] Woo, M.K. and Murray, B.A. (1993) Exp. Cell Res. 204, 336-345. [83] Cole, G.J. and Glaser, L. (1986) J. Cell Biol. 102, 4 0 3 4 1 2 . [84] Reyes, A.A., Akeson, R., Brezina, L. and Cole, G.J. (1990) Cell Regulation I , 567-576. [ 8 5 ] Hall, A.K. and Rutishauser, U. (1987) J. Cell Biol. 104, 1579-1586. [86] Becker, J.W., Erickson, H.B., Hoffman, S., Cunningham, B.A. and Edelman, G.M. (1989) Proc. Natl. Acad. Sci. USA 86, 1088-1092. [87] Fredette, B., Rutishauser, U. and Landmesser, L. (1993) J. Cell Biol. 123, 1867-1888. [88] Yang, P., Yin, X. and Rutishauser, U. (1992) J. Cell Biol. 116, 1487-1496.

562 [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [99]

Seki, T. and Arai, Y. (1993) J. Neurosci. 13, 2351-2358. Tang, J., Landmesser, L. and Rutishauser, U. (1992) Neuron 8, 1031-1044. Doherty, P., Ashton, S.V., Skaper, S.D., Leon, A. and Walsh, FS. (1992) J. Cell Biol. 117, 1093-1099. Rathjen, F.G. and Schachner, M. (1984) EMBO J. 3, 1-10, Moos, M., Tacke, R., Scherer, H., Teplow, D., Frith, K. and Schachner, M. (1988) Nature 334,701-703. Persohn, E. and Schachner, M. (1987) J. Cell Biol. 105, 569-576. Grumet, M. and Edelman, G.M. (1988) J. Cell Biol. 106, 487-503. Reid, R.A. and Hemperley, J.J. (1992) J. Mol. Neurosci. 3, 127-135. Kadmon, G., Kowitz, A,, Altevogt, I? and Schachner, M. (1990) J. Cell Biol. 110, 193-208. Kadmon, G., Kowitz, A,, Altevogt, I? and Schachner, M. (1990) J. Cell Biol. 110, 209-218. Horstkorte, R., Schachner, M., Magyar, J.P., Vorherr, T. and Schmitz, 9. (1993) J. Cell Biol. 121, 14091421. [IOO] Asou, H. (1992) J. Cell Physiol. 153, 318-320. [ l o l l Williams, E.J., Doherty, P., Turner, G., Reid, R.A., Hemperley, J.J. and Walsh, FS. (1992) J. Cell Biol. 119, 883-892. [I021 Brady, R.O. and Quarles, R.H. (1988) Cell. Mol. Neurobiol. 8, 139-148. [I031 Sato, S., Fugita, N., Kurihara, T., Kuwano, R., Sakimura, K., Takahashi, Y. and Miyatake, T. (1989) Biochem. Biophys. Res. Commun. 163, 1473-1480. [I041 Spagnol, G., Williams, M., Srinivasan, J., Golier, J., Bauer, P., Lebo, L.V. and Latov, N. (1989) J. Neurosci. Res. 24, 137-142. [I051 Burger, D., Perruisseau, G., Simon, M. and Steck, A.J. (1992a) J. Neurochem. 58, 845-853. [I061 Burger, D., Perruisseau, G., Simon, M. and Steck, A.J. (1992b) J. Neurochem. 52, 854-861. [I071 Salzer, J.L., Holmes, W.P. and Colman, D.R. (1987) J. Cell Biol. 104, 957-965. [I081 Sadoul, R. (1990) J. Neurosci. Res. 25, 1-13. [lo91 Poltorak, M., Sadoul, R., Keilhauer, G., Landa, C., Fahrig, T. and Schachner, M. (1987) J. Cell Biol. 105, 1893-1899. [IIO] Fahrig, T., Landa, C., Pesheva, P.,Kuhn, K. and Schachner, M. (1987) EMBO J. 6, 2875-2883. [I 1 I] Probstmeier, R., Fahrig, T., Spies, E. and Schachner, M. (1992) J. Cell Biol. 116, 1063-1070. [I 121 Lemke, G. and Axel, R. (1985) Cell 40, 501-508. [ I 1 31 Sakamoto, Y., Kitamura, K., Yoshimura, K., Nishijima, T. and Uyemura, K. (1 987) J. Biol. Chem. 252, 42084214. [I 141 Brunden, K.R. (1992) J. Neurochem. 58, 1659-1666. [I 151 Poduslo, J.F., Dyck, PJ and Berg, C.T. (1985) J. Neurochem. 44, 388400. [I 161 Wells, C.A., Saavedra, R.A., Inouye, H. and Kirschner, D.A. (1993) J. Neurochem. 61, 198771995, [117] Filbin, M.T. and Tennekoon, (3.1. (1991) Neuron 7, 845-855. [I 181 Filbin, M.T. and Tennekoon, G.I. (1993) J. Cell Biol. 122, 451459. [I 191 Yazaki, T., Miura, M., Asou, H., Kitamura, K., Toya, S. and Uyemur, K. (1992) FEBS Lett. 307, 361366. [I201 Rutishauser, U. and Goridis, C. (1986) Trends Genet. 2, 72-76. [121] Schachner, M., Faussner, A,, Kruie, J., Lindner, J., Meier, F.G., Raylhen, F.G. and Wemecke, H. (1983) Cold Spring Harbor Symp. Quantum Biol. 48, 557-568. [I221 McGarry, R.C., Nelfand, S.L., Quarles, R.H. and Roder, J.C. (1985) Nature 306, 376-378. [ 1231 Bollensen, E. and Schachner, M. (1 987) Neurosci. Lett. 82, 77-82. [ 1241 Ariga, T., Kohriyama, T., Freddo, I., Latov, N., Saito, M., Kon, K., Ando, S., Suzuki, M., Hemling, M.E., Rinehart, K.L., Kusunoki, S. and Yu, R.K. (1987) J. Biol. Chem. 262, 848-853. [ 1251 Chou, D.K.H., Ilyas, A.A., Evans, J.E., Quarles, R.H. and Jungalwala, F.B. ( I 985) Biochem. Biophys. Res. Commun. 128, 383-388. [I261 Chou, D.K.H., Schwarting, G.A., Evans, J.E. and Jungawala, F.B. (1987) J. Neurochem. 49, 865-873. [I271 Oka, S., Terayama, K., Kawashima, C. and Kawasaki, T. (1992) J. Biol. Chem. 267, 22711-22714. [I281 Kunemunde, V., Jungawala, F.B., Fischer, G., Chou, D.K.H., Keichauer, G. and Schachner, M. (1988) J. Cell Biol. 106, 213-223. [I291 Riopelle, R.J., McGarry, R.C. and Roper, J.C. (1986) Brain Res. 367, 20-25. [I301 Hall, H., Liu, L., Schachner, M. and Schmitz, 9 . (1993) Eur. J. Neurosci. 5, 3 4 4 2 .

563 [I311 Springer, T.A. (1990) Nature 346, 425434. [I321 Wawryk, S.O., Novotny, J.R., Wicks, 1 2 , Wilkinson, D., Haner, R.D., Salvaris, G., Welch, K., Fecondo, J. and Boyd, A.W. (1989) Immunol. Rev. 108, 135-161. [I331 Albelda, S.M. and Buck, C.A. (1990) FASEB J. 4, 2868-2880. [I341 Buck, C.A. (1992) Sem. Cell Biol. 3, 179-188. [I351 Harlan, J.M. (1985) Blood 65, 513-525. [136] Smith, C.W., Marlin, S.D., Rothlein, R., Tuman, C. and Anderson, D.C. (1989) J. Clin. Invest. 83, 2008-2017. [I371 Smith, C.W., Rothlein, R., Hughes, B.J., Mariscalco, M.M., Rudfoff, I.J., Schmalsteig, F.C. and Anderson, D.C. (1988) J. Clin. Invest. 82, 17461756. [138] Dustin, M.L., Rothlein, R., Bhan, A.K., Dinarello, C.A. and Springer, T.A. (1986) J. Immunol. 137, 245-254. [I391 Staunton, D.E., Marlin, S.D., Stratowa, C., Dustin, M.L. and Springer, T.A. (1988) Cell 52, 925-933. [I401 Simmons, D., Makgoba, M.W. and Seed, B. (1988) Nature 331, 624427. [I411 Horley, K.J., Carpenito, C., Baker, B. and Takei, F. (1989) EMBO J. 8, 2889-2896. [I421 Marlin, S.D. and Springer, T.A. (1987) Cell 51, 813-819. [I431 Staunton, D.E., Dustin, M.L., Erickson, H.P. and Springer, T.A. (1990) Cell 61, 243-249. [I441 Diamond, M.S., Staunton, D.E., Marlin, S.D. and Springer, T.A. (1991) Cell 65, 83 961-971. [I451 Staunton, D.E., Merluzzi, VJ., Ruthlein, R., Barton, R., Marlin, S.D. and Springer, T.A. (1989b) Cell 56, 849-853. [146] Berendt, A.R., Simmons, D.L., Tansey, J., Newbold, C.I. and Marsh, K. (1989) Nature 341, 57-59. [I471 Carpen, O., Pallai, P., Staunton, D.E. and Springer, T.A. (1992) J. Cell Biol. 118, 1223-1234. [I481 de Fougerolle, A,, Stacker, S.A., Schwarting, R. and Springer, T.A. (1991) J. Exp. Med. 174,253-260. [I491 Staunton, D.E., Dustin, M.L. and Springer, T.A. (1989) Nature 339, 61-64. [I501 de Fougerolle, A.R. and Springer, T.A. (1992) J. Exp. Med. 175, 185-190. [I511 Vazuez, R., Hoffman, PA., Tomita, J.K., Dickinson, E.S., Jasman, R.L., St. John, T. and Gallatin, W.N. (1992) Nature 360, 485488. [I521 Fawcett, J., Holness, C.L.L., Needham, L.A., Turley, H., Gatter, K.C., Mason, D.Y. and Simmons, P.L. (1992) Nature 360, 481483. [I531 de Fougerolle, A.R., Klickstein, L.B. and Springer, T.A. (1993) J. Exp. Med. 177, 1187-1 192. [I541 de Fougerolle, A.R., Qin, X. and Springer, T.A. (1994) J. Exp. Med. 179, 619429. [I551 Osborn, L., Ijession, C., Tizard, R., Vassallo, C., Luhowsky, S., Chi-Rosen, G. and Lobb, R. (1989) Cell 59, 1203-1211. [I561 E k e s , M.L., Osborn, L., Takada, Y., Crouse, C., Luhowsky, S. and Lobb, R.R. (1990) Cell 60,577-584. [I571 Rice, G.E., Munro, J.M., Corless, C. and Bevilacqva, M.P. (1991) Am. J. Pathol. 138, 385-393. [I581 Cybulsky, M.I., Fries, J.W.U., Williams, A.J., Sultan, P., Eddy, R., Byers, M., Shows, T., Gimbone, M.A. and Collins, T. (1991) Proc. Natl. Acad. Sci. USA 88, 7859-7863. [I591 Moy, P., Lobb, R., Tizard, R., Olson, D. and Hession, C. (1993) J. Biol. Chem. 268, 8835. [I601 Terry, R.W., Kwee, L., Levine, J.F. and Labow, M.A. (1993) Proc. Natl. Acad. Sci. USA 90,5919-5923. [I611 Hahne, M., Lentner, M., Jager, U. and Vestweber, P. (1994) Eur. J. Immunol. 24, 421428. [I621 Osborn, L., Vassallo, C. and Benjamin, C.D. (1992) J. Exp. Med. 176, 99-106. [I631 Vonderheide, R.H. and Springer, T.A. (1992) J. Exp. Med. 175, 1433-1437. [I641 Freedman, AS., Munro, J.M., Rice, G.E., Bevilacqua, M.R., Morimoto, C., McIntyre, B.W., Rhynhart, K., Pober, J.S. and Nadler, L.M. (1990) Science 249, 1030-1037. [I651 Rosen, G.D., Sanes, J.R., Le Chance, R., Cunningham, J.M., Roman, J. and Dean, D.C. (1992) Cell 69, I107-1I19. [I661 Muller, W.A., Ratti, C.M., McDonell, S.L. and Cohn, Z.A. (1989) J. Exp. Med. 170, 399414. [I671 Knapp, W., Dorken, B., Rieber, P., Schmidt, R.H., Stein, H. and van dem Borne, A.E. (1989) Blood 74, 1448-1 450. [I681 Newman, P., Bernot, M.C., Gorski, J., White, J.C., Lyman, S., Paddock, C. and Muller, W.A. (1990) Science 247, 1212-1222. [I691 Stockinger, H., Gadd, S.J., Eher, R., Majdic, 0.. Schreiber, W., Kasinrerk, W., Strass, B., Schnable, E. and Knapp, W. (1990) J. Immunol. 145, 3889-3897.

564 [I701 Simmons, D., Walker, C., Power, C. and Pigott, R. (1990) J. Exp. Med. 171, 2147-2152. [I711 Albelda, S.M., Muller, W.A., Buck, C.A. and Newman, P.J. (1991) J. Cell Biol. 114, 1059-1068. [I721 Zehnder, J.L., Hirai, K., Shatzky, M., McGregor, Y.L., Leuitt, L.J. and Leung, L.L.K. (1992) J. Biol. Chem. 267, 5243-5249. [I731 Delisser, H.M., Muller, W.A., Newman, P.J. and Albelda, S.M. (1991) J. Cell Biol. 115, 70a. [174] Schimmenti, L.A., Yan, H.-C., Madri, J.A. and Albelda, S.M. (1992) J. Cell Physiol. 153, 4 1 7 4 2 8 . [I751 Menvin, J.R., Tucker, A,, Albelda, S.M. and Mardi, J.A. (1990) J. Cell Biol. 1 1 1 , 157. [I761 Muller, W.A., Weigl, S., Deng, X. and Phillips, P.M. (1994) J. Exp. Med. 179, 449460. [I771 Tang, D.G., Chen, Y.Q., Newman, P.J., Shi, L., Gao, X., Diglio, C.A. and Honn, K.V. (1993) J. Biol. Chem. 268, 22883-22894. [I781 Hynes, R.O. (1985) Annu. Rev. Cell Biol. I , 67-90. [I791 Schwarzbauer, J.E. (1991a) Curr. Opin. Cell Biol. 3, 786-791. [I801 Mosher, D.F. (1989) Fibronectins. Academic Press, New York, pp. 53-69. [181] Hynes, R.O., George, E.L., Georges, E.N., Guan, J.L., Rayburn, H. and Yang, J.T. (1992) Cold Spring Harbor Symp. Quantum Biol. 57, 249-258. [182] Tressel, T., McCarthy, J.B., Calaycay, J., Lee, T.D., Legesse, K., Shively, J.E. and Pande, H. (1991) Biochem. J. 274, 73 1-739. [I831 Main, A.L., Harvey, T.S., Baron, M . , Boyd, J. and Cambell, I.D. (1992) Cell 71, 671-678. [184] Novokhatny, V., Schwarz, E, Atha, D. and Ingham, K. (1992) J. Mol. Biol. 227, 1182-1 191. [I851 Mosher, D.F. (1993) Curr. Opin. Struct. Biol. 3, 214-222. [I861 Engel, J. (1991) Curr. Opin. Cell Biol. 3, 779-785. [ 1871 An, S.S.A., Jimenez-Barbero, J., Petersen, T.E. and Llinas, M. (1992) Biochemistry 3 1, 9927-9933. [I881 Kar, L., Lai, C.-S., Wolff, C.E., Nettesheim, D., Sherman, S. and Johnson, M.E. (1993) J. Biol. Chem. 268, 8580-8589. [I891 Engvall, E., Ruoslahti, E. and Miller, E.J. (1978) J. Exp. Med. 147, 1584-1595. [I901 Owens, R.J. and Baralle, F.E. (1986) EMBO J. 5, 2825-2830. [I911 Garcia-Pardo, A. and Gold, L.I. (1993) Arch. Biochem. Biophys. 304, 181-188. [192] Mosher, D. (1975) J. Biol. Chem. 250, 6614-6621. [193] Wilson, C.L. and Schwarzbauer, J.E. (1992) J. Cell Biol. 119, 923-934. [I941 Heremans, A,, De Cock, B., Cassiman, J.J., van dem Berghe, H. and David, G. (1990) J. Biol. Chem. 265, 8716-8724. [I951 Hayashi, M. and Yamada, K.M. (1982) J. Biol. Chem. 257, 5263-5267. [I961 Hayashi, M. and Yamada, K.M. (1983) J. Biol. Chem. 258, 3332-3340. [197] San Antonio, J.D., Slover, J., Lawler, J., Karnovsky, M.J. and Lander, A.D. (1993) Biochemistry 32, 47464755. [198] Kjellen, L. and Lindahl, U. (1991) Annu. Rev. Biochem. 60, 4 4 3 4 7 5 . [199] Olberg, A. and Ruoslahti, E. (1982) J. Biol. Chem. 257, 48594863. [200] Yamada, K.M., Kennedy, P.W., Groten Dorst, G.R. and Momoi, T. (1981) J. Cell Physiol. 109,343-351. [201] Yamada, K.M. (1991) J. Biol. Chem. 266, 12809-12812. [202] Limper, A.H., Quade, B.J., La Chance, R.M., Birdenmeier, T.M., Rangwala, T.S. and McDonald, J.A. (1991) J. Biol. Chem. 266, 9697-9702. [203] Morla, A. and Ruoslahti, E. (1992) J. Cell Biol. 118, 421429. [204] Morla, A,, Zhang, Z. and Ruoslahti, E. (1994) Nature 367, 193-196. [205] Tanaka, K.I., Maeda, T., Titani, K. and Sekiguchi, K. (1992) FEBS Lett. 299, 155-158. [206] McDonald, J.A., Quade, B.J., Broekelmann, T.J., La Chance, R., Forsman, K., Hasegawa, E. and Akiyama, S. (1987) J. Biol. Chem. 262, 2957-2967. [207] Schwarzbauer, J.E. (1991) J. Cell Biol. 113, 1463-1473. [208] Perkins, R.M., Kellie, S., Patel, B. and Critchley, D.R. (1982) Exp. Cell Res. 141, 231-243. [209] Yamada, K.M., Critchley, D.R., Fishman, P.H. and Moss, J. (1983) Exp. Cell Res. 143, 295-302. [210] Matyas, G.R., Evers, D.L., Radinsky, R. and Morre, D.J. (1986) Exp. Cell Res. 162, 296-318. [21 I ] Spiegel, S., Yamada, K.M., Hom, B.H., Moss, Y. and Fishman, P.H. (1986) J. Cell Biol. 102, 1898-1906. [212] Thompson, L.K., Horowitz, P.M., Bentley, K.L., Thomas, D.D., Alderette, J.F. and Klebe, R.J. (1986) J. Biol. Chem. 261, 5209-5214.

565 [213] Hormann, H., Richter, H. and Jelinic, V. (1984) Hoppe-Seyler’s Z. Physiol. Chem. 365, 517-524. [214] Skortstengaard, K., Jensen, M.S., Sahl, P., Peterson, T.E.J. and Magnusson, S. (1986) Eur. J. Biochem. 161, 4 4 1 4 5 3 . [215] Takasaki, S., Yamashita, K., Suzuki, K. and Kohata, A. (1980) J. Biochem. 88, 1587-1 594. [216] Takamoto, M., Endo, T., Isemura, M., Yamaguchi, Y., Okamora, K., Kochibe, N. and Kobata, A. (1989) J. Biochem. 106, 228-235. [217] Takamoto, M., Endo, T., Isemura, M., Kochihe, N. and Kobata, A. (1989) J. Biochem. 105, 742-750. [218] Krusius, T., Fukuda, M., Dell, A. and Ruoslahti, E. (1985) J. Biol. Chem. 260, 41104116. [219] Linnala, R., von Koskull, H. and Virtanen, 1. (1994) FEBS Lett. 337, 167-170. [220] Mosher, D. (1989) Fihronectins. Academic Press, New York, pp, 20-23. [221] Takasaki, S., Yamashita, K., Suzuki, K., Iwanaga, S. and Kobata, A. (1979) J. Biol. Chem. 254, 85488553. [222] Fukuda, M. and Hakomori, S. (1979) J. Biol. Chem. 254, 5442-5450. [223] Fukuda, M., Leuery, S.R. and Hakomon, S. (1982) J. Biol. Chem. 257, 6856-6860. [224] Cossu, G. and Warren, L. (1983) J. Biol. Chem. 258, 5603-5607. [225] Matsuura, H., Takio, K., Titani, K., Greene, T., Levery, S.B., Salyan, M.E.K. and Hakomeri, S.I. (1988) J. Biol. Chem. 263, 3314-3322. [226] Matsuura, H., Greene, T. and Hakomori, S.I. (1989) J. Biol. Chem. 264, 10472-10476. [227] Bernard, B.A., DeLuca, L.M., Hassell, J.R., Yamada, K.M. and Olden, K. (1984) J. Biol. Chem. 259, 53 10-53 15. [228] Ledger, P.W., Nishimoto, S.K., Hayashi, S. and Tanzer, M.L. (1983) J. Biol. Chem. 258, 547-554. [229] Sai, S. and Tanzer, M.L. (1984) Exp. Cell Res. 154, 4 0 2 4 1 1. [230] Arumugham, R.G. and Tanzer, N.L. (1983) J. Biol. Chem. 258, 11883-1 1889. [231] Pena, S.D.J. and Hughes, R.C. (1978) Nature 276, 80-83. [232] Hughes, R.C. and Mills, G. (1986) J. Cell Physiol. 128, 402412. [233] Olden, K., Pratt, R.M. and Yamada, K.M. (1979) Proc. Natl. Acad. Sci. USA 76, 3343-3347. [234] Jones, G.E., Arumugham, R.G. and Tanzer, M.L. (1986) J. Cell Biol. 103, 1663-1670. [235] Zhu, B.C. and Laine, R.A. (1985) J. Biol. Chem. 260, 40414045. [236] Chung, A.E., Jaffe, R., Freeman, I.L., Vergnes, J.P., Braginski, J.E. and Carlin, B. (1979) Cell 16, 277287. [237] Timpl, R., Rohde, H., Gehron, R.P., Rennard, S.I., Foidart, J.M. and Martin, G.R. (1979) J. Biol. Chem. 254, 9933-9937. [238] Martin, G.R. (1987) Annu. Rev. Cell Biol. 3, 57-85. [239] Timpl, R. (1989) Eur. J. Biochem. 180, 487-502. [240] Beck, K., Hunter, I. and Engel, J. (1990) FASEB J. 4, 148-160. [24 11 Engel, J. ( I 992) Biochemistry 3 I , 10644-1 065 1. [242] Sasaki, M., Kohno, K., Kato, S., Martin, G.R. and Yamada, Y. (1987) Proc. Natl. Acad. Sci. USA 84, 935-939. [243] Sasaki, M., Kleinman, H.K., Huher, H., Deutzmann, R. and Yamada, Y. (1988) J. Biol. Chem. 263, 16536-1 6544. [244] Sasaki, M. and Yamada, Y. (1988) J. Biol. Chem. 262, 171 11-171 17. [245] Sung, U., O’Rear, Y.R. and Yurchenko, P.D. (1993) J. Cell Biol. 123, 1255-1268. [246] Yurchenko, P.D., Sung, U., Ward, M.D., Yamda, Y. and O’Rear, J.J. (1993) J. Biol. Chem. 268, 83568365. [247] Roberts, D.D., Rao, C.N., Magnan, J.L., Spitalnik, S.L., Liotta, L.A. and Ginsburg, V. (1985) Proc. Natl. Acad. Sci. USA 82, 1306-1310. [248] Ott, U., Odermatt, E., Engel, J., Furthmay, R.M. and Timpl, R. (1982) Eur. J. Biochem. 123, 63-72. [249] Charonis, A S . , Tsilihary, E.C., Yurchenko, ED. and Furthmayer, H. (1985) J. Cell Biol. 100, 18481853. [250] Laurie, G.W., Bing, J.T., Kleinman, W.K., Hassell, J.R., Aumailley, M., Martin, G.R. and Feldmann, R.J. (1986) J. Mol. Biol. 189, 205-216. [251] Fox, J.W., Mayer, U., Nischt, R., Aumailley, M., Reinhardt, D., Wiedemann, H., Mann, K., Timpl, R., Krieg, T., Engel, J. and Chu, M.-L. (1991) EMBO J. 10, 3137-3146.

566 [252] Battaglia, C., Mayer, U., Aumailley, M. and Timpl, R. (1992) Eur. J. Biochem. 208, 359-366. [253] Mayer, U., Nischt, R., Poschl, E., Mann, K., Fukuda, K., Gerl, M., Yamada, Y. and Timpl, R. (1993) EMBO J. 12, 1879-1885. [254] Bruch, M., Landwehr, R. and Engel, J. (1989) Eur. J. Biochem. 185, 271-279. [255] Schnittny, J.C. and Yurchenko, P.D. (1990) J. Cell Biol. 110, 825-832. [256] Mecham, R.P. (1991) FASEB J. 5, 2538-2546. [257] Gehlsen, K.R., Dickerson, K., Argraves, W.S., Engvall, E. and Ruoslahti, E. (1989) J. Biol. Chem. 264, 19034-1 9038. [258] Tomaselli, K.J., Hall, D.E., Flier, L.A., Gehlsen, K.R., Turner, D.C., Carbonetto, S. and Reichardt, L.F. (1990) Neuron 5, 651462. [259] Aumailley, M., Gerl, M., Sonnenberg, A,, Deutzmann, R. and Timpl, R. (1990) FEBS Lett. 262, 82-86. [260] Hall, D.E., Reichardt, L.F., Crowley, E., Holley, B., Moezzi, H., Sonnenberg, A. and Damsky, C.H. (1990) J. Cell Biol. 110, 2175-2184. [261J Sonnenberg, A,, Linders, C.J.T., Modderman, P.W., Damsky, C.H., Aumailley, M. and Timpl, R. (1990) J. Cell Biol. 110, 2145-2155. [262] Gehlsen, K.R., Sriramarao, P, Furcht, L.T. and Skubitz, A.P.N. (1992) J. Cell Biol. 117, 449459. [263] Lallier, T., Deutzmann, R., Perris, R. and Bonner-Frazer, M. (1994) Dev. Biol. 162, 451464. [264] Graf, J., Iwamoto, Y., Sasaki, M., Martin, G.R., Kleinman, H.K., Robey, F.A. and Yamada, Y . (1987) Cell 48, 989-996. [265] Ehrig, D., Leivo, I., Agraves, W.S., Ruoslahti, E. and Engvall, E. (1990) Proc. Natl. Acad. Sci. USA 87, 3264-3268. [266] Vuolteenato, R., Nissinen, M., Sainio, K., Byers, M., Eddry, R., Hirvonen, H., Shows, T.B., Sariola, H., Engvall, E. and Tryggvason, K. (1994) J. Cell Biol. 124, 381-394. [267] Hunter, D.D., Shah, V, Merlie, J.P. and Sanes, J.R. (1989) Nature 338, 229-233. [268] Marinovich, M.P., Lunstrum, G.P., Keene, D.R. and Burgeson, R.E. (1992) J. Cell Biol. 119, 695-703. [269] O’Rear, J.J. (1992) J. Biol. Chem. 267, 20555-20557. [270] Kallunki, P., Saino, K., Eddry, R., Byers, M., Kallunki, T., Sariola, H., Beck, K., Hirvonen, H., Shows, T.B. and Tryggvason, K. (1992) J. Cell Biol. 119, 679493. [271] Engvall, E., Eanvicker, D., Haaparanta, T., Ruoslahti, E. and Sanes, J.R. (1990) Cell Regulat. 1, 731740. [272] Glukhova, M., Koteliansky, V, Fondacci, C., Marotte, F. and Rappaport, L. (1993) Exp. Cell Res. 157, 437447. [273] Brown, J.C., Wiedemann, H. and Timpl, R. (1994) J. Cell Sci. 107, 329-338. [274] Cooper, A.R., Kurkinen, M., Taylor, A. and Hogan, B.L.M. (1981) Eur. J. Biochem. 119, 189-197. [275] Howe, C.C. and Dietzschold, B. (1983) Dev. Biol. 98, 385-391. [276] Kurkinen, M., Barlow, D.P., Jenkins, J.R. and Hogan, B.L.M. (1983) J. Biol. Chem. 258, 6543-6548. [277] Peters, B.P. and Goldstein, I.J. (1979) Exp. Cell Res. 120, 321-334. [278] Shibata, S., Peters, B.P., Roberts, D.D., Goldstein, I.J. and Liotta, L.A. (1982) FEBS Lett. 142, 194-197. [279] Rao, G.N., Goldstein, I.J. and Liotta, L.A. (1983) Arch. Biochem. Biophys. 227, 118-124. [280] Wadsworth, J.D.F., Okuno, A. and Strong, PN. (1993) Biochem. J. 295, 537-541. [281] Sato, S. and Hughes, R.C. (1992) J. Biol. Chem. 267, 69834990. [282] Arumugham, R.G., Hsieh, T.C.-Y. and Laine, R.A. (1986) Biochim. Biophys. Acta 883, 112-126. [283] Knibbs, R.N., Perini, F. and Goldstein, I.J. (1989) Biochemistry 28, 63794392. [284] Fujiwara, S., Shinkai, H., Deutzmann, R., Paulsson, M. and Timpl, R. (1988) Biochem. J. 232,453461. [285] Paulsson, M., Aumailley, M., Deutzmann, R., Timpl, R., Beck, K. and Engel, J. (1987) Eur. J. Biochem. 166, 11-19. [286] Mann, K., Deutzmann, R., Aumailley, M., Timpl, R., Raimondi, L., Yamada, Y., Pan, T.-C., Conway, D. and Chu, M.-L. (1989) EMBO J. 8, 65-72. [287] Carlin, B., Jaffe, R., Bender, B. and Chung, A.E. (1981) J. Biol. Chem. 256, 5209-5214. [288] Paulsson, M., Deutzmann, R., Dziadek, M., Nowack, H., Timpl, R., Wber, S. and Engel, J. (1986) Eur. 3. Biochem. 156, 467478. [289] Fujiwara, S., Shinkai, H., Mann, H. and Timpl, R. (1993) Matrix 13, 215-222.

567 [290] Aumailley, M., Rattaglia, C., Mayer, U., Reinhardt, D., Nischt, R., Timpl, R. and Fox, J.W. (1993) Kidney Int. 43, 7-12. [291] Reinhardt, D., Mann, K., Nischt, R., Fox, J.W., Chu, M.-L., Krieg, T. and Timpl, R. (1993) J. Biol. Chem. 268, 10881-10887. [292] Chakravati, S., Tam, M.F. and Chung, A.E. (1990a) J. Biol. Chem. 265, 10597-10603. [293] Tsao, T., Hsieh, J.C., Durkin, M.E., Wu, C., Chakravati, S., Dong, L.J., Lewis, M. and Chung, A.E. (1990) J. Biol. Chem. 265, 5188-5191. [294] Chakravati, S., Ming, T.F. and Chung, A.E. (1990b) J. Biol. Chem. 265, 10597-10603. [295] Dedhar, S., Jewell, K., Rojiani, M. and Gray, V (1992) J. Biol. Chem. 267, 18908-18914. [296] Erickson, H.P. and Iglesias, J.L. (1984) Nature 31 I , 267-269. [297] Yamada, K.M., Yamada, S.S. and Pastani, I. (1975) Proc. Natl. Acad. Sci. USA 72, 3158-3162. [298] Chiquet-Ehrismann, R. (1990) FASEB J. 4, 2598-2604. [299] Erickson, H.P. and Bourdon, M.A. (1989) Annu. Rev. Cell Biol. 5, 71-92. [300] Jones, F.S., Hoffman, S., Cunningham, B.A. and Edelman, G.M. (1989) Proc. Natl. Acad. Sci. USA 86, 1905-1 909. [301] Spring, J., Beck, K. and Chiquet-Ehrismann, R. (1989) Cell 59, 325-334. [302] Tucker, R.P., Spring, J., Baumgartner, S., Martin, D., Hagios, C., Poss, P.M. and Chiquet-Ehrismann, R. (1994) Development 120, 637447. [303] Julian, J., Chiquet-Ehrismann, R., Erickson, H.P. and Carson, D.D. (1994) Development 120, 661-671, [304] Aufderheide, E. and Ekblom, I? (1988) J. Cell Biol. 107, 2341-2349. [305] Probstmeier, R., Martini, M. and Schachner, M. (1990) Development 109, 313-321. [306] Hemesath, T.J. and Stefansson, J. (1994) J. Immunol. 152, 422428. [307] Lightner, VA, and Erickson, H.P. (1990) J. Cell Sci. 95, 263-277. [308] Faissner, A,, Kruse, J., Kuhn, K. and Schachner, M. (1990) J. Neurochem. 54, 1004-1015. [309] Chiquet-Ehrismann, R., Matsuoka, Y., Hofer, U., Spring, J., Bernasconi, C. and Chiquet, M. (1991) Cell Regulat. 2, 927-938. [310] Marton, L.S., Gulcher, J.R. and Stefansson, K. (1989) J. Biol. Chem. 264, 13145-13149. [31 I ] Salmivirta, M., Elenius, K., Vainio, S., Hofer, U., Chiquet-Ehrismann, R., Thesleff, 1. and Jalkanen, M. (1991) J. Biol. Chem. 266, 7733-7739. [312] Chiquet, M. and Fambrough, D.M. (1984) J. Cell Biol. 98, 1937-1946. [313] Vaughan, L., Huber, S., Chiquet, M. and Winterhalter, K.H. (1987) EMBO J. 6, 349-353. [314] Hoffman, B. and Edelman, G.M. (1987) Proc. Natl. Acad. Sci. USA 84, 2523-2527. [315] Klatt, K.P., Yang, E.V. and Tassava, R.A. (1992) Differentiation 50, 133-140. [316] Aukhil, I., Joshi, P., Yan, Y.and Erickson, H.P. (1993) J. Biol. Chem. 268, 2542-2553. [317] Chiquet-Ehrismann, R. (1991) Curr. Opin. Cell Biol. 3, 80G804. [318] Lotz, M.M., Burdsal, C.A., Erickson, H.I? and McClay, D.R. (1989) J. Cell Biol. 109, 1795-1806. [319] Hafrar, W., Chiquet-Ehrismann, R. and Tucker, R.P. (1989) Dev. Biol. 132, 14-25. [320] Chiquet-Ehrismann, R., Kalla, I?, Pearson, C.A., Beck, K. and Chiquet, M. (1988) Cell 53, 383-390. [321] Bourdon, M.A. and Ruoslahti, E. (1989) J. Cell Biol. 108, 1149-1 155. [322] Termine, J.D., Kleinnman, H.K., Whitson, S.W., Conn, K.M., McGarvey, M.L. and Martin, G.R. (1981) Cell 26, 99-105. [323] Termine, J.D., Robey, P.G., Fisher, L.W., Shimokawa, H., Drum, M.A., Conn, K.M., Hawkins, G.R., Cruz, J.B. and Thompson, K.G. (1984) Proc. Natl. Acad. Sci. USA 81, 2213-2217. [324] Jundt, G., Berghauser, K.-H., Termine, J.D. and Schuiz, A. (1987) Cell Tissue Res. 248, 409415. [325] Sage, H., Johnson, C. and Bornstein, I? (1984) J. Biol. Chem. 259, 39934007. [326] Mason, I.J., Taylor, A,, Williams, J.G., Sage, H. and Hogan, B.L.M. (1980) EMBO J. 5, 1465-1472. [327] Dziadek, M., Paulsson, M., Aumailley, M. and Timpl, R. (1986) Eur. J. Biochem. 161, 455464. [328] Lane, T.F. and Sage, H. (1994) FASEB J. 8, 163-173. [329] Mayer, U., Aumailley, M., Mann, K., Timpl, R. and Engel, J. (1991) Eur. J. Biochem. 198, 141-150. [330] Hughes, R.C., Taylor, A,, Sage, H. and Hogan, B.L.M. (1987) Eur. J. Biochem. 163, 57-65. [331] Kelm, R.J. and Mann, K.G. (1991) J. Biol. Chem. 266, 9632-9639. [332] Sage, N., Vernon, R.B., Funk, S.E., Everitt, E.A. and Angello, J. (1989) J. Cell Biol. 109, 341-356. [333] Tremble, PM., Lane, T.F., Sage, H. and Werb, Z. (1993) J. Cell Biol. 121, 1433-1444.

568 [334] Prince, C.W., Oosawa, T., Butler, W.T., Tomana, M., Bhown, A.S., Bhown, M. and Schrohenloher, R.E. (1987) J. Biol. Chem. 262, 2900-2907. 13351 Franzen, A. and Heinegard, D. (1985) Biochem. J. 232, 715-724. [336] Craig, A.M., Smith, J.H. and Denhardt, D.T. (1989) J. Biol. Chem. 264, 9682-9689. [337] Senger, D.R. and Perruzi, C.A. (1986) Cancer Res. 45, 5818-5823. [338] Singh, K., DeVouge, M.W. and Mukherjee, B.B. (1990) J. Biol. Chem. 265, 18696-18701. [339] Olberg, A,, Franzen, A,, Heinegard, D., Pierschbacher, M. and Ruoslahti, E. (1988) J. Biol. Chem. 263, 19433-19436. [340] Denhardt, D.T. and Guo, X. (1993) FASEB J. 7, 1475-1482. [341] Flores, M.E., Norgard, M., Heinegard, D., Reinholt, F.P. and Andersson, G. (1992) Exp. Cell Res. 201, 526-530. [342] Tamkun, J.W., DeSimone, D.W., Fonda, D.A., Patel, R.S., Buck, C., Honvitz, A.F. and Hynes, R.O. (1986) Cell 46, 271-282. [343] Pytela, R., Pierschbacher, M.D. and Ruoslahti, E. (1985) Cell 40, 191-198. [344] Argraves, W.S., Suzuki, S., Arai, H., Thompson, K., Pierschbacher, M.D. and Ruoslahti, E. (1987) J. Cell Biol. 105, 1183-1 190. [345] Fitzgerald, L.A., Poncz, M., Steiner, B., Rall, S.C., Bennett, J.S. and Phillips, D.B. (1987) Biochemistry 26, 8158-8165. [346] Akiyama, S.K., Nagata, K. and Yamada, K.M. (1990) Biochim. Biophys. Acta 1031, 91-110. [347] Humphries, M.J. (1990) J. Cell Sci. 97, 585-592. [348] Hynes, R.O. (1987) Cell 48, 549-554. [349] Hynes, R.O. (1992) Cell 69, 11-25. [350] Albelda, S.M. (1993) Lab. Invest. 68, 4-17. [35 I] Juliano, R.L. (1987) Biochim. Biophys. Acta 907, 261-278. [352] Altruda, F., Cervella, P., Tarone, G., Botta, C., Balzac, F., Stefanuto, G. and Silengo, L. (1990) Gene 95, 26 1-266. [353] Languino, L.R. and Ruoslahti, E. (1992) J. Biol. Chem. 267, 71 16-7120. [354] van Kuppevelt, T.H., Languino, L.R., Gailit, J.O., Suzuki, S. and Ruoslahti, E. (1989) Proc. Natl. Acad. Sci. USA 86, 5415-5418. [355] Suzuki, S. and Naitoh, Y. (1990) EMBO J. 9, 757-763. [356] Hogervorst, F., Kuikman, I., von dem Borne, A.E. and Sonnenberg, A. (1990) EMBO J. 9, 765-770. [357] Tamura, R.N., Ruzzo, C., Starr, L., Chambers, J., Reichardt, L.F., Cooper, H.M. and Quaranta, V (1990) J. Cell Biol. 1 1 1, 1593-1604. [358] Tamura, R.N., Cooper, H.M., Collo, G. and Quaranta, V (1991) Proc. Natl. Acad. Sci. USA 88, 1013810187. [359] Hogervorst, E, Kuikman, I., van Kessel, A.G. and Sonnenberg, A. (1991) Eur. J. Biochem. 199, 425433. [360] Cooper, H.M., Tamura, R.N. and Quaranta, V (1991a) J. Cell Biol. 115, 843-850. [361] Bray, P.F., Leung, C.S.I. and Shuman, M.A. (1990) J. Biol. Chem. 265, 9587-9590. [362] Hemler, M.E., Kassner, P.D. and Chan, B.M.C. (1992) Cold Spring Harbor Symp. Quantum Biol. 57, 21 3-220. [363] de Curtis, I. and Reichardt, L.F. (1993) Development 118, 377-388. [364] Balzac, F., Belkin, A.M., Koteliansky, V.E., Balabanov, Y.V., Altruda, F,, Silengo, L. and Tarone, G. (1993) J. Cell Biol. 121. [365] Hogervorst, F., Admiraal, L.G., Niessen, C., Kuikman, I., Janissen, H., Daams, H. and Sunnenberg, A. (1993) J. Cell Biol. 121. [366] Stepp, M.A., Spurr-Michaud, S., Tisdale, A., Elwell, J. and Gipson, I.K. (1990) Proc. Natl. Acad. Sci. USA 87, 8970-8974. [367] Sonnenberg, A,, Calafat, J., Janssen, H., Daams, H., van der Raaij-Helmer, L.M., Falcioni, R., Kennel, S.J., Aplin, J.D., Baker, J., Loizidou, M. and Garrod, D. (1991) J. Cell Biol. 113, 907-917. [368] Kurpakus, M.A., Quaranta, V and Jones, J.C.R. (1991) J. Cell Biol. 115, 1737-1750. [369] Burridge, K., Fath, K., Kelly, T., Nuckolls, G. and Turner, G. (1988) Annu. Rev. Cell Biol. 4, 487-525. [370] Koyama, T. and Hughes, R.C. (1992) J. Biol. Chem. 267, 25939-25944.

569 [371] Carrell, N.A., Fitzgerald, L.A., Steiner, B., Erickson, H.P. and Phillips, D.R. (1985) J. Biol. Chern. 260, 1743-1 749. [372] Nermut, M.V., Green, N.M., Eason, P., Yamada, S.S. and Yamada, K.M. (1988) EMBO J. 7,40934099. [373] Weise, J.W., Nagaswami, C., Vilaire, G. and Bennett, J.S. (1992) J. Biol. Chem. 267, 16637-16643. [374] Plow, E.F. and Ginsberg, M.H. (1981) J. Biol. Chem. 256, 947779482, [375] Plow, E.F., Strouji, A.H., Meyer, D., Marguerie, G. and Ginsberg, M.H. (1984) J. Biol. Chem. 259, 5388-5391. [376] Altieri, D.C., Bader, R., Mannucci, P.M. and Edgington, T.S. (1988) J. Cell Biol. 107, 1893-1900. [377] Arroyo, A.G., Garcia-Pardo, A. and Sanchez-Madrid, F. (1993) J. Biol. Chem. 268, 9863-9868. [378] Faull, R.J., Kovack, N.L., Harlan, J.M. and Ginsberg, M.H. (1993) J. Cell Biol. 121, 155-162. [379] Keizer, G.D., Visser, W., Vliem, M. and Figdor, C.G. (1988) J. Immunol. 140, 1393-1400, [380] van Kooyk, Y., Weder, P., Hogervorst, F., Verhoeven, A.J., van Seventer, G., Velde, A.A., Borst, P. Keizer, G.D. and Figdor, C.G. (1991) J. Cell Biol. 112, 345-354. [381] Landis, R.C., Bennett, R.I. and Hogg, N. (1983) J. Cell Biol. 120, 1519-1527. (1993) [382] Smyth, S.S., Jonecki, C.C. and Panse, L.V. (1993) Blood 81, 2827-2843. [383] Cheresh, D.A., Pytela, R., Pierschbacher, M.D., Klier, F.G., Ruoslahti, E. and Reisfeld, R.A. (1987) J. Cell Biol. 105, 1163-1 173. [384] Okada, Y., Mugnai, G., Bremer, E.G. and Hakomori, S. (1984) Exp. Cell Res. 155, 4 4 8 4 5 6 . [385] Zheng, M., Fang, H., Tsuruoka, T., Tsuji, T., Sasaki, T and Hakornori, S.4. (1993) J. Biol. Chem. 268, 221 7-2222. [386] Barletta, E., Mugnai, G . and Ruggieri, S. (1993) Biochem. Biophys. Res. Cornmun. 192, 214-222. I3871 Gulberg, D., Gehlsen, K.R., Turner, D.C., Ahlen, K., Zijenah, L.S., Barnes, M.J. and Rubin, K. (1992) EMBO J. I I , 3865-3873. [388] Morton, L.F., Peachey, A.R., Zijenah, L.S., Goodall, A.H., Humphries, M.J. and Barnes, M.J. (1994) Biochem. J. 299, 791-797. [389] Carter, W.G., Ryan, M.C. and Gahy, P.J. (1991) Cell 65, 599-610. [390] Miyauchi, A,, Alvarez, J., Greenfield, E.M., Teti, A,, Grano, H., Colucci, S., Zambonn-Zallone, A., Teitelbaum, S.L., Cheresh, D.T. and Hruska, K.A. (1991) J. Biol. Chem. 266, 20369-20374. [391] Zhang, Z., Morla, A.O., Vuori, K., Bauer, J.S., Juliano, R.L. and Ruoslahti, E. (1993) J. Cell Biol. 122, 235-242. [392] Kaufmann, R., Frosch, D., Westphal, C., Webber, L. and Klein, C.E. (1989) J. Cell Biol. 109, 18071815. [393] Takeuchi, K.-I., Tsuji, T., Hakomori, S.-I. and Irimura, T. (1994) Exp. Cell Res. 21 1, 133-141. [394] Adams, J.C. and Watt, EM. (1990) Cell 63, 4 2 5 4 3 5 . [395] Symington, B.E., Takada, Y. and Carter, W.G. (1993) J. Cell Biol. 120, 523-535. [396] Sriramarao, P., Steffner, P. and Gehlsen, K.P. (1993a) J. Biol. Chem. 268, 22036-22041. [397] Grinnell, F. (1992) J. Cell Sci. 101, 1-5. [398] Arnaout, M.A. (1990) Blood 75, 1037-1050. [399] Miller, L.J., Bainton, D.F., Borregard, N. and Springer, T.A. (1987) J. Clin. Invest. 80, 535-543. [400] Fryer, D.R., Morganroth, M.L., Rogers, C.E., Arnaout, M.A. and Tudd, R.F. (1988) J. Clin. Immunol. Immunopathol. 46, 272-28 1. [401] Miller, L.J. and Springer, T.A. (1987) J. Imrnunol. 139, 842-852. [402] Remold-O’Donnell, E. and Savage, B. (1988) Biochemistry 27, 3 9 4 8 . [403] Bray, P.F., Rosa, J.P., Lingappa, V.R., Kan, Y.W., McEver, R.P. and Shurnan, M.A. (1986) Proc. Natl. Acad. Sci. USA 83, 1480-1484. [404] Duperray, A., Troesch, A., Berthier, R., Chagnon, E., Frachet, P., Jzan, G. and Margueneg, G. (1989) Blood 74, 1603-1611. [405] O’Toole, T.E., Luftus, J.C., Plow, E X , Glass, A.A., Harper, J.R. and Ginsberg, M.H. (1989) Blood 74, 14-18. [406] Heino, Y., Ignotz, R.A., Hernler, M.E., Crouse, C. and Massague, Y. (1989) J. Biol. Chem. 864, 380388. [407] Roberts, C.J., Birkenmeier, T.M., McQuillan, J.J., Akiyama, S.K., Yamada, S.S., Chew, W.T., Yamada, K.M. and McDonald, J.S. (1988) J. Biol. Chem. 263, 45864592.

570 [408] Ignotz, R.A., Heino, J. and Massague, J. (1989) J. Biol. Chem. 264, 389-392. [409] Akiyama, S.K. and Yamada, K. (1987) J. Biol. Chem. 262, 17536-17542. [410] Jaspers, M., de Strooper, B., Spaepen, M., van Leuven, F., David, G., van der Bergne, H. and Cassiman, J.-J. (I 988) FEBS Lett. 23 1, 402406. [41 I ] de Strooper, B., van Leuven, F., Carmeliet, G., van den Bergne, H. and Cassiman, J.-J. (1991) Eur. J. Biochem. 199, 25-33. [412] Vekeman, S., Jaspers, M. and Cassiman, J.-J. (1993) FEBS Lett. 327, 207-212. [413] Wilcox, D.A., Wautier, J.L., Pidard, D. and Newman, P.J. (1994) J. Biol. Chem. 269, 44504457. [414] Troesch, A,, Duperray, A,, Polack, B. and Marguerie, G. (1990) Biochem. J. 268, 129-133. [415] Young, W.W., Lutz., M.S., Mills, S.E. and Lechler-Osbern, S. (1990) Proc. Natl. Acad. Sci. USA 87, 6838-6842. [416] Dahms, N.M. and Hart, G.W. (1986) J. Bioi. Chem. 261, 13186-13196. [417] Asada, M., Furukawa, K., Kantor, C., Gahmberg, C. and Kobata, A. (1991) Biochemistry 30, 15611571. [418] Dahms, N.M. and Hart, G.W. (1985) J. Immunol. 134, 3978-3956. [419] Pesheva, P., Honvitz, A.F. and Schachner, M. (1987) Neurosci. Lett. 83, 303-306. [420] Calvete, J.J. and Muniz-Dim, E. (1993) FEBS Lett. 328, 30-34. [421] Couchman, J.R., Austria, R., Woods, A. and Hughes, R.C. (1988) J. Cell Physiol. 136, 228-236. [422] Stamatoglou, S. and Keller, J.M. (1983) J. Cell Biol. 96, 1820-1823. [423] Akiyama, S.K., Yamada, S.S. and Yamada, K.M. (1989) J. Biol. Chem. 264, 18011-18018. [424] Hughes, R.C., Foddy, L. and Bause, E. (1987) Biochem. J. 2, 537-545. [425] Cheresh, D.A., Pierschbacher, M.D., Nerzig, M.A. and Mujoo, K. (1986) J. Cell Biol. 102, 688-696. [426] Bums, G.F., Lucas, C.M., Krissanen, G.W., Werkmeister, Y., Scanlon, D.B., Simpson, R.J. and Vadas, N.A. (1988) J. Cell Biol. 107, 1225-1230. [427] Symington, B.E., Symington, F.W. and Rohrschneider, L.R. (1989) J. Biol. Chem. 264, 13258-13266. [428] van den Water, L., Aronson, D. and Braman, V. (1988) Cancer Res. 48, 5730-5737. [429] Oz, O.K., Campbell, A. and Tao, T.-W. (1989) Int. J. Cancer 44, 343-347. [430] Kawano, T., Takasaki, S., Tao, T.-W. and Kobata, A. (1993) Int. J. Cancer 53, 91-96. [431] Bednarczyk. J.L., Szabo, M.C. and McIntyre, B.W. (1992) J. Biol. Chem. 267, 25274-25281. [432] Wadsworth, J.D.F., Haluorson, M.J., Chang, A.C. and Coligan, J.E. (1993b) J. Immunol. 150, 847-857. [433] Cheresh, D.A. and Harper, J.R. (1987) J. Biol. Chem. 262, 1434-1437. [434] Malek-Hedayat, S. and Rome, L.H. (1994) J. Cell Biol. 124, 1039-1046. [435] Sriramarao, P., Berger, E., Chambers, J.D., Arfors, K.-E. and Gehlsen, K.R. (1993b) J. Cell. Biochem. 51, 360-368. [436] Dean, J.W., Chandrasekaran, S. and Tanzer, M.L. (1988) Biochem. Biophys. Res. Commun. 156, 41 1416. [437] Dean, J.W., Chandrasekaran, S. and Tanzer, M.L. (1990) J. Biol. Chem. 265, 12553-12562. [438] Chammas, R., Veiga, S.S., Line, S., Potoenjak, I? and Brentani, R.R. (1991) J. Biol. Chem. 266, 33493355. [439] Chandrasekaran, S., Dean, J.W., Giniger, M.S. and Tanzer, M.L. (1991) J. Cell. Biochem. 46, 115-124. [440] Chandrasekaran, Tanzer, M.L. and Giniger, M.S. (1994) J. Biol. Chem. 269, 3356-3366. [441] Chandrasekaran, Tanzer, M.L. and Giniger, M.S. (1994) J. Biol. Chem. 269, 3367-3373. [442] Codogno, P., Doyennette-Moyne, M.A., Botti, J. and Aubrey, M. (1988) J. Cell Physiol. 136,463470. [443] Chammas, R., Veiga, S.S. Travassos, L.R. and Brentani, R.R. (1993) Proc. Natl. Acad. Sci. USA 90, 1795-1 799. I4441 Cooper, D.N.W., Massa, S.M. and Barondes, S.H. (1991b) J. Cell Biol. 115, 1437-1448. [445] Hughes, R.C. (1992a) Curr. Opin. Struct. Biol. 2, 687492. [446] Hughes, R.C. (1992b) Biochem. SOC.Trans. 20, 279-284. [447] Hughes, R.C. (1994) Glycobiology 4, 5-12. [448] Gu, M., Wang, W., Song, W.K., Cooper, D.N.W. and Kaufman, S.J. (1994) J. Cell Sci. 107, 175-1 81.

J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins II 0 1997 Elsevier Science B.V. All rights reserved CHAPTER 15

Carbohydrate differentiation antigens Ii, SSEA-1 (Le”) and related structures Prototype mammalian carbohydrate antigens that serve as ligands in molecular recognition Ten Feizi The Glycosciences Laboratory, Northwick Park Hospital, Walford Road, Harrow, Middx. HA1 3UJ U K .

Abbreviations S E A - 1 stage specific embryonic antigen I

1. Introduction Understanding the biological roles of the diverse oligosaccharides of glycoproteins, proteoglycans and glycolipids at the surface of cells has been one of the challenges in cell biology, a challenge traditionally shunned by cell biologists because of the heterogeneities of these substances and the lack of non-laborious methods for their detailed analysis and structure/function assignments. Repeatedly, however, developments in immunological frontiers have drawn attention to oligosaccharides as prominent cell surface antigens of animal cells. Notable among these are the major blood group antigens A, B and H(0) [1,2] around which revolves the practice of safe blood transfusion, and a diverse array of differentiation antigens (surface antigens whose expression changes during embryonic development, cell differentiation and maturation) the existence of which was particularly highlighted by work with naturally occurring monoclonal antibodies to the I and i antigens, and hybridoma-derived monoclonal antibodies [3]. What is now regarded as “differentiation antigen” status of the blood group A, B, H and Lewis antigens was already apparent from results of histochemical studies which showed changes in the expression of these antigens in various organs during the development of the human fetus [4]. There were also observations from the bloodbank arena that levels of the I and i antigens on human red cells change in the course of the first year of life (i antigen diminishes and I increases concomitantly) [ 5 ] . When the i and I antigens recognized by monoclonal autoantibodies were characterized as oligosaccharide backbones (linear and branched chains, respectively) of the major blood group antigens [ 6 ] , the serological observations could be interpreted as reflecting an increase in the proportions of the branched oligosaccharide backbone structures in the course of the first year of life. Thus, here were clues to the occurrence of programmed changes in glycosyltransferase activities during the developmental processes. i

Gal(~1-4)GlcNAc(~1-3)Gal(~1-4)GlcNAc(~1-3)Gal(~I-4)GlcNAc(~1-3)Gal(~1-

57 1

572

Gal(fiI4)GlcNAc(fi 1-6)

\ Gal(fil4)GlcNAc(fil-3)Gal(fi 1-

I Gal@ I4)GlcNAc(fi1-3)

I

The sequence-specific, natural monoclonal antibodies proved to be powerful reagents, and through their use, remarkable stage-specific changes were revealed in the branching patterns of the oligosaccharide backbones during murine embryonic development [7]. There followed the finding that an antigen transiently expressed on murine embryonic cells is the Forssman carbohydrate antigen [8,9], and that the 8-cell stage-specific embryonic antigen of the mouse, SSEA-I, consists of the a-1,3-fucosylated blood group chain, Le” [lo], and that many other antigenic markers of developmental stage, differentiation fate and maturation state are blood-group-related carbohydrate structures [3,11]. This was thanks to the advent of the hybridoma technology [I21 and the surge of research activity to generate monoclonal antibodies in the search for cellular components with important functions. SSEA-l(Le’) Gal@ 1 -4)GlcNAc(fi 1 -

I

Fuc(a1-3)

The finding that oligosaccharides are among these much sought antigens led the author to suggest that oligosaccharides, among them the blood group and related antigens, may have roles as “area codes” which determine cell migration pathways and serve as ligands in macromolecular interactions in health and disease [3,13,14]. In that case, endogenous lectins would be candidates for roles as “postmen”, “policemen” and “traffic signs” involved in the obedient interpretation of the area codes! [ 131. Moreover, we proposed [3,15] that the several tumor-associated carbohydrate antigens identified using hybridoma antibodies [3,11], although of little promise as targets for cancer therapy [3], may be involved in the disordered behaviour of the tumor cells. These developments were naturally welcomed by carbohydrate chemists and biochemists, but may have done little to enthuse the population of biologists at large. However, attitudes have been changing following the discoveries by the molecular cloning route that several proteins with important effector functions in the mechanisms of inflammation and host defence contain motifs with predicted carbohydrate-binding activities (reviewed in ref. [ 161). Notable among such proteins are the leukocyte-to-endothelium adhesion molecules, the selectins [ 17-20]. Not only is information accumulating on the saccharide ligands for these proteins [ 19,21-23], but new principles are emerging that carbohydrate-protein interactions are essential elements of effector functions in immunity. These principles have caught the imagination of biotechnology and of the pharmaceutical industry [24]; and there is optimism that carbohydrate-based compounds will find a place as drugs in the management of disorders of immunity. This chapter recounts aspects of two antigen systems: Ii and SSEA-1 (Le”), that are prototype carbohydrate differentiation antigens which have served to establish some

513

principles at the heart of glycobiology, namely, insights into biological roles of specific oligosaccharide chains of glycoproteins and glycolipids. Whereas the bioactivities of some of these oligosaccharide sequences are not yet established with certainty, those of others have been defined in terms of their roles as determinants of the tropism of a microbial pathogen, or in terms of their recognition by defined endogenous carbohydrate-binding proteins that are intimately involved in the marshalling of the body's defence mechanisms through cell adhesion events that they mediate.

2. The I and i antigens and their sialyl forms 2.1. Background The I and i antigens are best known as antigens expressed on human red cells, but they are also expressed on leukocytes, macrophages and on certain epithelial glycoproteins (reviewed in ref. [6]). They are recognized by autoantibodies that occur in high titres persistently in a chronic hemolytic disorder (cold hemagglutinin disease), and transiently following infections such as those caused by Mycoplasma pneumoniae (anti-I) and Epstein-Barr virus (anti-i). These autoantibodies have low affinities, and agglutinate red cells only at temperatures below 37"C, hence their description as cold agglutinins. The anti-I and -i cold agglutinins in chronic hemolytic disorders are usually monoclonal; many of the transiently occurring cold agglutinins are also of restricted clonality [6]. On the one hand, the triggering of the anti-I autoantibodies with considerable frequency following M. pneumoniae infection, and on the other, their monoclonal nature, has made this antigen system appealing for in depth investigations of the host-parasite interactions, the pathogenesis of autoantibody production, and the elucidation of the structures and functions of the antigens. The story that has unfolded has given some clear insights into the developmental regulation of glycosylation, and into the biological roles of long oligosaccharide sequences of poly-N-acetyllactosamine type [25-281 as ligands in molecular interactions. Salient developments are highlighted in the sections that follow. 2.2. Biochemical nature of Ii antigens

Exploratory studies in the author's laboratory on the transient autoimmune disorder elicited by M. pneumoniae infection had indicated that there is little if any I antigen in the cultured micro-organism recognized by the elicited anti-I in the infected patient (reviewed in ref. [29]). Therefore the mycoplasma was considered unlikely to be a useful source of I-active substance for characterization. Work in the field of blood transfusion had established, first, that the I antigen is strongly expressed on the red cells of the vast majority of adults, second, that the i antigen is strongly expressed on those of the fetus and neonate and of rare adults that lack I antigen [5,30], and that the level of I antigen increases during the first year of life and i decreases concomitantly. The possibility that I antigen is a carbohydrate structure was raised by a finding that I antigen of red cells could be degraded with a bacterial filtrate containing glycosidases [31]. As

574

in the case of the major blood group antigens, it was the finding of I antigen activity in water soluble mucin-type substances that greatly facilitated characterization. Thus, when a water soluble I substance in human milk [32] was found to be rich in galactose and N-acetylglucosamine and low in fucose[33], a relationship of the I antigen to the backbones of the blood group antigens (then referred to as precursor chains) was suggested. This was readily corroborated by studies of ovarian cyst glycoproteins with and without blood group A, B and H activities, and with the ABH-active glycoproteins from which terminal monosaccharides were removed by Smith degradation [33]. One of the anti-I sera, that of patient Ma with chronic cold hemagglutinin disease, (anti-I Ma), was shown to recognize the Gal(Pl4)GlcNAc(P1-6) sequence known to constitute one branch of the backbone sequence, and the masking effect of the ABH monosaccharides on the I Ma epitope was demonstrated [33]. Subsequent work with erythrocyte glycolipids [34-371 in conjunction with chemically synthesized oligosaccharides [38-40] has resulted in elucidation of the biochemical nature of both the I and the i antigenic determinants. Anti-i was shown to recognize antigenic determinants on a linear oligosaccharide (octasaccharide) sequence consisting of the repeating type 2: G a l ( ~ 1 4 ) G l c N A c1-3)(~ sequence, and anti-I, determinants on the branched sequence formed by the addition of an N-acetyllactosamine branch joined by a 1,6 linkage (i.e. the I Ma determinant) to the inner galactose of the i antigen structure. It was already known that carbohydrate chains of fetal red cells are predominantly simple, unbranched structures, while those of adult red cells are more complex. and branched [41]. Thus a biochemical explanation was found for the change of human red cells from i to I antigen status in the first year of life. This was also an explanation for the predominance of simple oligosaccharides on erythrocytes of rare adults with i-type red cells [42]. No two anti-I and -i antibodies were found to be absolutely identical in their reaction patterns with various branched and linear oligosaccharide analogues (such differences were later observed also with hybridoma antibodies to defined carbohydrate sequences [43]) and it was clear that there are three main types of anti-I. The first (exemplified by anti-I Ma) recognizes the “( 1,4)-, (1,6)-” branch; and the second reacts best with the “(1,4)-, (1,3)-” sequence in the presence of branching, and the third requires both branches to be available [36]. Moreover, individual antibodies were found to differ in their ability to accommodate their antigenic determinants in the presence of additional glycosylations of the backbone sequence. Substitution of I and i active structures with a-l,2-Fuc was found to completely mask their reactivity, whereas substitution with a-2,3-NeuAc or a- 1,3-Gal had a different effect with individual antibodies: with certain antibodies no masking, with others partial masking, and with others still, complete masking of the antigenic determinants (reviewed in ref. [44]). These observations provided an explanation for the marked differences in the reactivities of these monoclonal anti-I with various complex glycoproteins [33,45] and with various panels of erythrocyte variants [46]. These principles which are now known to apply to hybridoma-derived monoclonal antibodies, on the one hand highlight the need to precisely characterize the fine specificities of the antibodies, and on the other, they demonstrate the power of well characterized antibodies as reagents in glycosylation studies.

515

2.3. Anti-I and -i antibodies as reagents in studies of cell dierentiation Anti4 and -i antibodies have served as valuable reagents in studies of antigenic changes during early embryogenesis [7]. They revealed that the earliest embryo cells of the mouse express I antigen without i. Only at the start of differentiation, when primary endoderm cells can be discerned, is i antigen expressed. Thus the linear backbone sequence is a stage-specific marker of murine embryonal epithelial differentiation. Epithelial cells at later stages of differentiation were studied both using mouse teratocarcinoma cells that undergo differentiation during culture in oitro, and mouse embryos at later stages of development. Anti-I and -i antibodies, in conjunction with other hybridoma-derived antibodies provided a wealth of information on changes in the poly-N-acetyllactosamine family of oligosaccharides during stages of the developmental processes. Some of the epithelia were lacking in i and I but they reacted strongly with anti-blood group H, suggesting that fucose al-2 linked to galactose that constitutes bloodgroup H masks the expression of the I and i antigens on the backbones. H Fuc( a 1 -2)Gal( fi1-4,3)GlcNAc(B 1-

These and other observations during embryonic development [47,48] together with those on the changing antigenicities of human red cells during the first year of life, served to confer the status of differentiation antigens to oligosaccharides, a concept amply expanded with the advent of the hybridoma antibodies. Moreover, they extended the initial observations by Muramatsu [49-521 that in embryos and teratocarcinoma cells there occur glycoproteins with long carbohydrate chains (poly-N-acetyllactosamine type) susceptible to endo-fi-glycosidase. Remarkable spatio-temporal patterns of Ii antigen expression were also observed in the developing chick embryo [53]. 2.4. Anti-I and -i antibodies as immunosequencing reagents for glycoprotein oligosaccharides

Because of their exquisite specificities for linear and branched oligosaccharides of poly-N-acetyllactosamine type which occur typically as isomeric sequences that are difficult to separate from one another, anti-i and -I antibodies served as invaluable reagents providing insights into the remarkable differences among individual cells in expression of these antigens (and by inference the differences in details of glycosylation) in various established cell lines [54]. Thus it is no wonder that glycoproteins, more often than not, have heterogeneous oligosaccharides (microheterogeneity at individual glycosylation sites) even when isolated from apparently homogeneous cell lines. Anti-I and -i antibodies were also shown to be powerful as immunosequencing reagents for the oligosaccharides of defined cellular glycoproteins and glycolipids. For example, immunoprecipitation of radioiodinated human red cell membranes using anti-I, revealed band 3 protein as the main immunoreactive component, thus providing the first evidence for the presence of poly-N-acetyllactosamine sequences on this glycoprotein [55]. A further new principle established using these antibodies as reagents was that different membrane glycoconjugates may be revealed on cells depending on

576

whether the protein or oligosaccharide moieties are labelled. Thus, when the carbohydrate moieties of the red cells were labelled using galactose oxidase/sodium borotritide, the immunoprecipitation with anti-I revealed the carbohydrate-rich polyglycosyl ceramides rather than band 3 [55]. The knowledge that there is a lack of expression of genetically predicted A, B and H antigens in certain epithelial tumors [56,57] stimulated us to use monoclonal anti-I and anti-i antibodies to look for inappropriately expressed I and i antigen activities on glycoproteins extracted from gastrointestinal tumors. Glycoprotein extracts from metastatic (but not primary) colon tumors were found to express these antigens [58]. Later studies showed that the I Ma determinant behaves as a tumor-associated antigen on glycoprotein extracts of gastric carcinomas derived from secretors [59]. We extended the studies of mucin-type gastrointestinal glycoproteins to include several other carbohydrate epitopes recognized by hybridoma antibodies and found different repertoires among secretors and non-secretors and differences in their tumor derived glycoproteins [60]. The author has proposed that the immunosequencing approach, fine tuned to incorporate the neoglycolipid approach [6 1,621, would greatly facilitate mapping the repertoire of oligosaccharides in normal and neoplastic epithelia. For example, antibody binding experiments may be performed with neoglycolipids generated from mucin glycopeptides or from particular glycosylation sites on the glycopeptides, and chromatographed on silica gel plates. Other applications of anti-I and -i antibodies for immunosequencing included some investigations of membrane glycoproteins of human lymphocytes and promyelocytic cells. Immune precipitation experiments with galactose oxidasehodium borotritide labelled lymphocytes revealed glycosylation differences: major differences in the expression of the I and i antigens on cells of T and B origins. The “T200” family of glycoproteins (also termed leukocyte common antigen or CD45) on lymphocytes of B rather than T origins were shown to strongly express I and i activities, and by inference to have unsubstituted poly-N-acetyllactosamine sequences [63]. Further experiments with the purified leukocyte common antigen species from the two cell types led to the first demonstration of polydisperse 0-linked oligosaccharides on these glycoproteins, of which those of B rather than T cell types were shown to be susceptible to endo-(3-galactosidase [64]. Unfractionated thymocytes and T-lymphoblastoid cell lines were found to differ from cells of B origins in expressing the I Ma determinant on the major sialoglycoprotein [63] (now termed sialophorin or CD43). This was therefore the first indication of the presence of the core 2 (ref. [65]) sequence on the 0-linked chains of CD43 of activated T lymphocytes [66]. On the promyelocytic cell line HL60, abundant, diffusely migrating cell surface glycoconjugates were revealed with apparent molecular weights greater than 100kDa[15]. These were also bound by anti SSEA-1, and probably contained the glycoproteins now known to serve as counter-receptors for the selectins which are discussed in section 3.3. Similarly, polydisperse high apparent molecular weight components, not visualisable by conventional protein staining were demonstrated on murine embryonal carcinoma cells [67]. Such glycoproteins, rather than glycolipids, were the carriers also of the stage specific embryonic antigen SSEA- 1. The immunosequencing approach has also been applied to the glycolipids of blood cells [68,69]. For example, the use of an anti-i antibody in conjunction with sequence-specific

hybridoma antibodies and endo-fi-galactosidase resulted in the demonstration of an i active ceramide octasaccharide as well as several species of neutral glycolipids with six or more than ten monosaccharides carrying the Le" antigen and others the Ley antigen [69]. 2.5. Roles of Ii and related sequences as ligands

From their very structures and because of the different carbohydrate substituents that they may support, the Ii-type oligosaccharides were regarded good candidates for roles as ligands in molecular recognition [44,70]. With improved techniques, it should be possible to rigorously investigate such roles for oligosaccharides of this family. Here I highlight some evidence which has accumulated in the course of immunopathobiological investigations. 2.5.1. Sialyl Ii as host-cell ligands for a pathogen Mycoplasma pneumoniae M. pneumoniae causes a spectrum of acute respiratory tract disease in humans ranging from mild upper respiratory tract infection to full blown pneumonia. Transient immunological disorders are well known to occur following this infection, the most common being the production of high titre autoantibodies to the I antigen (reviewed in ref. [71]). As in chronic idiopathic autoimmune hemolytic anaemia of cold agglutinin type, the anti-I elicited in M. pneumoniae infection are low affinity autoantibodies that show red cell binding, agglutination and complement fixation at temperatures below 37°C. The precise mechanism of autoimmunization in this excellent model of autoimmunity in humans awaits elucidation. However, the molecular characterization of the I autoantigen and of the host-cell receptors for the infective agent (as described below) have indicated collectively that the disorder has its origin in the interaction of the infective agent with I-type saccharides on host cells. Thus this post-infective disorder may be regarded as a prototype of carbohydrate-triggered pathology. M. pneumoniae does not produce I antigen; and although there is evidence for the uptake of trace amounts of I antigen-active substances by the infective agent from the medium in which it is grown [29], it is not known whether this is a passive adsorption or whether it is a receptor-mediated uptake. Any passively adsorbed microbe-associated I antigen is not thought to be the immunogen that triggers the autoantibody production as this is unlikely to be the sole host antigen adsorbed by the infective agent (reviewed in ref. [71]). Moreover, in experimental immunizations of rabbits, anti-I was not elicited when the mycoplasma or human red cells alone were tested as the immunogens, but were elicited only when red cells with specifically bound mycoplasma were used [72]. This suggested that the microbe-host cell complex or interaction products trigger the production of anti-I, a concept corroborated by observations on the nature of the hostcell ligands for this agent, as discussed below. M.pneumoniae adheres to human erythrocytes and to a variety of other cell types, as well as to the ciliated epithelium which is the primary site of infection. The attachment sites on erythrocytes have been shown to be members of the Ii antigen family with sialic acid a-2,3 linked to galactose [73,74]: NeuAc(a2-3)Gal(fil-4)GlcNAc(fiI-3)Gal(fi I-4)GlcNAc(fiI-3)Gal(fiI4)GlcNAc(fi 1-

578

Gal@ 1 4 ) G l c N A c ( f i 1 4 )

\ Gal@ 14)GlcNAc(fi1-3)Gal(B 14)GlcNAc(fiII

NeuAc(a2-3)Gal(B 14)GlcNAc(fiI-3)

Using anti-I and -i antibodies as reagents in studies of the primary site of infection, i.e. the bronchial epithelium, sialyloligosaccharides of this family have been found to be richly displayed and highly polarized at the apicalAumina1 aspects of the epithelium as well as at the microvilli to which the mycoplasma is know to adhere [75,76]. These various observations have led us to propose that in the inflamed areas of the respiratory tissues, receptor-mediated complexes formed between the host oligosaccharides and the lipid-rich mycoplasma (functioning as an adjuvant) may serve as a trigger for autoimmunization, and that predominance of anti-I rather than anti-i after infection may reflect a greater abundance of branched carbohydrate ligands of I-type at the surface of host cells with which the mycoplasma forms complexes[75,76]. In a proportion of patients there is evidence for the occurrence of autoantibodies that recognize the sialyl-I antigen [77]. Thus these transiently occurring autoantibodies are anti-receptor antibodies directed either against the backbone domains or against the entire host-cell ligand structures. In another study [78], M. pneumoniae was found to bind to sulfated glycoconjugates. Knowing that sulfated oligosaccharide sequences of the neolacto family are produced by the bronchial epithelium [79], it will be interesting to apply state of the art technologies to isolate and characterize precisely the ligands among the glycans polarized at the surface of the bronchial epithelium. 2.5.2. Ii-type sequences as ligands for endogenous carbohydrate-binding proteins? As mentioned above, from the outset of her awareness of the biochemical nature of the Ii antigens, and the developmentally regulated changes in their expression, the author has considered most likely that they and their variously substituted forms constitute ligands for one or more endogenous carbohydrate-binding protein(s). Thus, upon becoming aware of the work from the Kornfeld Laboratory [80] on b-galactoside binding lectin (now termed galectin-1 [Sl]), isolated from bovine heart, and its binding to glycoprotein oligosaccharides having Gal@ 14)GlcNAc sequences at their non-reducing termini, for example, glycopeptides derived from bovine thymocytes containing Gal@ 14)GcNAc(B1-3 or 6) sequences [82], we investigated the activities of Ii-type sequences as ligands for this lectin [83]. We observed that Ii-active (blood group ABH-inactive) mucins are more potent inhibitors of the lectin than ABH-active mucins. Furthermore, by affinity chromatography, we isolated from the bovine heart tissue Ii-active substances which were potent inhibitors of the lectin [83], and raised the possibility that Ii-active oligosaccharide sequences may serve as surface receptors whose perturbation may affect the metabolism of cells [54]. However, in exploratory cell culture experiments in which Ii positive lymphoblastoid cell lines (human) were cultured in the presence of the purified lectin, or freshly isolated human or bovine lymphocytes were cultured in the presence of the bovine lectin or stimulated with phytohaemagglutinin in the presence of the bovine lectin, no effect was observed in the presence of the added lectin [Y. Katagiri and the author, 1980, unpublished observations].

579

In the meantime considerable information has accumulated on glycoproteins and glycolipids with oligosaccharides of poly-N-acetyllactosamine type that are bound by galectin- 1 and related proteins [84-881. And observations have been made consistent with the participation of galectins-1 and -3 in the adhesion of tumor cells to endothelial cells [89]. Also there have been descriptions made of nuclear [90-941 and cytoplasmic as well as extracellular[95] locations of the galectins which lack the signal sequence for secretion, and of the secretion of galectin-1 by a novel mechanism shared by certain growth regulatory molecules [96]. Based on the findings on the epidermal growth factor (EGF) receptor poly-N-acetyllactosamine sequences capped with LeXand Lea sequences, as well as blood group A [97] (the latter is a ligand for galectin-3 [98]), we proposed that growth regulatory units, and even networks may exist at the cell surface linked by oligosaccharide chains (of glycoproteins and glycolipids) and complementary lectins, and that this would be dependent on correct glycosylation and availability of desired lectins [15,99]. This was put forward in the hope of stimulating research on the possible regulation (fine tuning) of receptor activities through interactions of their oligosaccharides with endogenous lectins such as the galectins. Networks and lattices can indeed be formed in oitro by galectin- 1 [ 100,lO11, and there has been some support for the concept of galectin-mediated growth regulation from research on the EGF receptor in mutant cells lacking blood group A antigen and the finding that they have altered responses to EGF [102]. In addition, the activity of the low affinity IgE receptor (FcERI) on mast cells has been shown to be modulated by galectin-3 [103]. The finding that the Ii-type sequences capped with the blood group B, as well the A determinant are the preferred ligands for galectin-3 [87] lead us to propose that here may lie one clue to genetic susceptibilities to certain allergic diseases, or more precisely a clue to genetic determinants of the severity of allergic responses. This is something that may be worthwhile to investigate in view of some suggestive evidence of an over-representation of blood group B in a statistical study of pollen allergy in the human [104], and the evidence for a greater susceptibility of blood group A+ than blood group A- rabbits to experimental allergic encephalitis [ 1051. More recently evidence has been provided indicating that in activated (but not resting T lymphocytes) galectin-1 may trigger cell death by the apoptosis cascade through a carbohydrate-mediated interaction with the T200/CD45 molecule on the surface of these cells [ 1061. Of course important functional clues are much awaited from experimental knock-out of galectin genes, most likely of multiple galectin genes, in view of the lack of a defective phenotype with the murine galectin- 1 knock-out [ 1071.

3. Stage-spec@ embyyonic antigen-1, SSEA-1 (CDI 5/LeX/L5) and the sialyl and sulfated analogues 3.1. Background

The discovery of the status of the LeX oligosaccharide sequence as a differentiation antigen came following the activities of developmental biologists who were among the first to use hybridomas to generate “tailor-made” antibodies to surface antigens

580

of embryonic cells, in the search for molecules with important functions during the developmental processes [ 1081. Initially, the 8-cell stage embryonic antigen, SSEA-1, seemed elusive, for it could not be identified readily by the usual techniques of immune precipitation and polyacrylamide gel electrophoresis [ 1081. Following observations on the changing expression of the I and i antigens in the early mouse embryo, P.W. Andrews and the author (unpublished) considered the possibility that the anti-SSEA- 1 may be directed against a carbohydrate structure related to the Ii antigen system. Indeed this proved to be the case, but the pattern of expression of the antigen in the human was found to be very different though not any the less interesting from the biological point of view as recounted below. 3.2. Biochemical nature of SSEA-I and related antigens in mouse and human

We examined the reactivity of anti-SSEA-1 against a variety of glycoproteins known to express the I and i antigens; and found that the antibody binds to the intestinal glycoproteins (meconium) of the human fetus [lo]. Using a series of structurally defined oligosaccharides to inhibit the binding of the anti-embryo antibody, measured by radioimmunoassay, we established that anti-SSEA- 1 recognizes the 3-fucosylated N-acetyllactosamine and deduced that during mouse development, backbone structures of Ii-type can be converted into SSEA-I-active structures by a-l,3-fucosylation [ 101. A second fucose joined by al-2 linkage to the galactose residue, as in the Ley structure was found to mask completely the expression of the SSEA-I determinant, thus corroborating the concept introduced as a result of the Ii experience that the sequential addition (or deletion) of monosaccharides (including sialic acid) might be one mechanism for the appearance, disappearance and reappearance of antigens during stages of embryogenesis and differentiation [7,10,13]. Ley

Gal(~1-4)GlcNac(~1-

I

Fuc(a 1-2)

I

Fuc(a1-3)

The specificity of anti-SSEA- 1 was also investigated by examining its reactivities with glycolipids isolated from various human cells and tissues [ 109,110]; initial conflicting conclusions between results in the two laboratories were shortly resolved, and it was confirmed [ 1 1 1,1121 that the determinant consists of the 3-fucosyl-N-acetyllactosamine sequence, a sequence first identified on the blood group chains of human epithelial glycoproteins [113,114], and later found on glycolipids of human adenocarcinomas [ I IS]. The sequence was also found on serum glycoproteins, and among the oligosaccharides of human milk and urine, and was termed LeX to distinguish it from the isomeric Lea oligosaccharide sequence (reviewed in refs. [ 10,116,117]). In contrast to its relatively wide distribution in the human, the SSEA1/Lex antigen was found to occur only at restricted sites in the mouse [ 1081. Among these sites are certain parts of the brain [ 1 181201, and interestingly the developmentally regulated neural antigen, L5, of the mouse and other animal species has now been identified as the Le" sequence [121]. The pattern of expression of Le" antigen is a striking example of the species differences that occur in the expression of carbohydrate differentiation antigens. SSEA- 1

581

was found to be expressed by human granulocytes but not on those of the mouse, rat, rabbit or the pig [ 1221. Moreover, hybridoma antibodies, VEP8 and VEP9 selected for binding exclusively to granulocytes among human peripheral blood cells were shown to recognize the Le"/SSEA- 1 sequence [ 1 17,1231. There followed reports of many granulocyte-specific monoclonal antibodies (now collectively designated anti-CD 15) that recognize the Le" determinant on human granulocytes (ref. [124]; see ref. [125] for a review). Moreover, by immunochemical means evidence was obtained [ 1221 that among cells of the human peripheral blood, granulocytes and monocytes can be distinguished from the other cell types by their expression of the Le" and sialyl-Le" sequences. Biochemical evidence corroborated the existence of the Le" and sialyl-Le" sequences on glycoproteins and glycolipids of human cells of myeloid type [126-1281. Variants of the fucosyl, sialyl sequence based on poly-N-acetyllactosamine backbones were demonstrated by immunochemical[69,129] and biochemical means [ 126,130,13 11 on cells of the myelomonocytic series.

3.3. LeX, sialyl-LeX and related sequences as ligands for endogenous carbohydratebinding proteins The assignment of the Le" sequence to the stage-specific antigen, SSEA-1 [lo], which appears on the early mouse embryo coincident with the first cell-cell adhesion event, compaction [ 1081, raised the interesting possibility that this oligosaccharide sequence may be a ligand for an embryonic adhesion protein [ 101. Some evidence for the involvement of oligosaccharides of this family in the embryonic adhesion has been obtained [ 1321341. However, a cognate recognition protein (a receptor) in this system has not yet been identified. One proposal [ 1351 is that an entirely carbohydrate-mediated mechanism, LeXLe" interaction, may be operating here. The Le" oligosaccharide sequence has been shown to be recognized and bound by more than one protein of the innate immune system. These include the human macrophage endocytosis receptor [ 1361 and serum mannan-binding protein [ 1371. The precise biological significance of this specificity is uncertain, although it is likely that it comes to play in the body's defense mechanisms (endocytosis and opsonization, respectively) against microbial pathogens whose surface polysaccharides contain the Le" sequence. It is in the area of mechanisms of leukocyte trafficking in the inflammatory response that dramatic developments have occurred in our understanding of the functions of the Le" and related oligosaccharides as ligands. The finding of lectin-like motifs on the leukocyte-to-endothelium adhesion molecules, E-, L- and P-selectins [ 17-20], which have crucial roles at the initial stages of leukocyte recruitment in inflammation, stimulated intense research activities to identify their oligosaccharide ligands, knowing that inhibitors of the selectin adhesion will inhibit the subsequent cell extravasation cascades. Knowledge that the E- and P-selectins bind granulocytes and monocytes, not surprisingly, focussed research on the Le" and sialyl-Le" oligosaccharides as candidate ligands. And it is now well established from work in several laboratories that the 3'-sialyl-LeX sequence is a recognition structure for all three selectins, and that the asialo-Le" is also bound, though less strongly than the sialyl analogue, by P- and E-selectins. Details of the binding specificities of the three selectins towards naturally

582

occurring oligosaccharide sequences on glycoproteins and glycolipids have been reviewed elsewhere [19,21-23,1381, suffice it to say that the 3'-sulfated analogue of sialyl-Le", as well as the isomeric sequences, 3'-sialyl-Lea and 3'-sulfated Lea (the latter two are typically found on epithelial cells) are also ligands for the three selectins, and information is accumulating on the types of carrier proteins and carrier sphingolipids that can present these ligands in forms that serve as their biological counter-receptors in the leukocyte tethering and rolling phenomena that serve to recruit the leukocytes that are destined to extravasate. Expression of the Le" and Lea series oligosaccharides changes markedly on the glycoproteins and glycolipids of malignant cells, as reviewed by Hakomori in another volume in this series [139]. There are many observations and correlations that are consistent with the concept that malignant cells with strong expression of oligosaccharide chains in this series may highjack the leukocyte extravasation route through for example sialyl-LeX and sialyl-Lea-mediated binding to the endothelial selectins, and thereby give rise to metastases [140-1421.

4. Perspectives The Ii and SSEA-l/LeXantigen systems are just a few of the large array of developmentally regulated oligosaccharide sequences that are known on glycoproteins and glycQlipids. The uncovering of biological roles of these oligosaccharide sequences has thus far been, largely, the cumulative result of what could be described as incidental observations. With the increased awareness of examples of functionally important oligosaccharides sequences, and advances in techniques applicable in the glycosciences [ 1431, it can be envisaged that, increasingly, there will be more focussed efforts to unravel the roles of specific oligosaccharides as ligands in molecular recognition. Hopefully, technologies such as the neoglycolipid technology [61,62] which is applicable for the immobilization of any desired oligosaccharide sequence (for binding experiments), in conjunction with protein expression-cloning, will enable the pin-pointing and characterization of the cognate proteins in various tissues of the body. But, as is clear from the galectin system, discoveries of carbohydrate-binding proteins (the receptors) and of oligosaccharide ligands are only beginnings, and then begins the task of identifying the physiological counter-receptors (i.e. the macromolecular carriers of the oligosaccharide ligands) on the way to unravelling the cellular signals that can be generated following the receptor-counter-receptor interactions. For the design of oligosaccharide-based therapeutic substances that will be non-teratogenic and non-toxic, it will be important to establish a database of carbohydrate recognition systems in different tissues of the body.

A cknowledgements The author is supported by program grant No. E400/622 from the Medical Research Council, and thanks Mrs. Margaret Runnicles for her help in assembling the manuscript.

583

References I

I

Watkins, W.M. (1980) Adv. Human Genet. 10, 1-136. Kabat, E.A. (1982) Am. J. Clin. Pathol. 78, 281-292. Feizi, T. (1985) Nature 314, 53-57. Szulman, A.E. (1966) Annu. Rev. Med. 17, 307-322. Marsh, W.L. (1961) Br. J. Haematol. 7, 200-209. Feizi, T. (1981) Immunol. Commun. 10, 127-156. Kapadia, A., Feizi, T.and Evans, M.J. (1981) Exp. Cell Res. 131, 185-195. Willison, K.R. and Stern, P.L. (1978) Cell 14, 785-793. Willison, K.R., Karol, R.A., Suzuki, A., Kundu, S.K. and Marcus, D.M. (1982) J. Immunol. 129, 603609. Gooi, H.C., Feizi, T., Kapadia, A., Knowles, B.B., Solter, D. and Evans, M.J. (1981) Nature 292, 156158. Hakomori, S. (1985) Cancer Res. 45, 2405-2414. Kohler, G. and Milstein, C. (1975) Nature (London) 256, 495497. Feizi, T. (1981) Trends Biochem. Sci. 6, 333-335. Feizi, T.(1989) In: G. Bock and S. Harnett (Eds.), Carbohydrate Recognition in Cellular Function, Ciba Foundation Symposium 145. Wiley, Chichester, pp. 62-79. Feizi, T.and Childs, R.A. (1987) Biochem. J. 245, 1-1 1. Drickamer, K. (1993) Curr. Opin. Struct. Biol. 3, 393400. Brandley, B.K., Swiedler, S.J. and Robbins, P.W. (1990) Cell 63, 861-863. McEver, R.P. (1991) J. Cell. Biochem. 45, 156-161. Bevilacqua, M.P. and Nelson, R.M. (1993) J. Clin. Invest. 91, 379-387. Rosen, S.D. (1993) Histochemistry 100, 185-191, Feizi, T. ( I 991) Trends Biochem. Sci. 16, 84-86. Feizi, T. (1991) Curr. Opin. Struct. Biol. 1, 766-770. Feizi, T. (1993) Curr. Opin. Struct. Biol. 3, 701-710. Edgington, S.M. (1992) Bio/Technology 10, 383-389. Gardas, A. and Koscielak, J. (1974) FEBS Lett. 42, 101-104. Jarnefelt, J., Rush, J., Li, Y.T. and Laine, R.A. (1978) J. Biol. Chem. 253, 8006-8009. Krusius, T., Finne, J. and Rauvala, H. (1978) Eur. J. Biochem. 92, 289-300. Hanfland, P., Dabrowski, J. and Roelcke, D. (1979) In: R. Schauer, P. Boer, E. Buddecke, M.F. Kramer, J.F.G. Vliegenthart and H. Wiegandt (Eds.), Glycoconjugates. Georg Thieme Publishers, Stuttgart, 1pp. 520-521. Uemura, K., Loomes, L.M., Childs, R.A. and Feizi, T. (1988) Infect. Immun. 56, 3015-3020. Wiener, AS., Moor-Janokoowski, J., Gordon Eve, B. and Davis, J. (1965) Am. J. Phys. Antropol. 23, 389-396. 1 Marcus, D.M., Kabat, E.A. and Rosenfeld, R.E. (1963) J. Exp. Med. 118, 175-194. Feizi, T. and Marsh, W.L. (1970) Vox Sang. 18, 379-382. Feizi, T., Kabat, E.A., Vicari, G., Anderson, B. and Marsh, W.L. (1 97 1) J. Immunol. 106, 1578-1 592. Feizi, T., Childs, R.A., Hakomori, S. and Powell, M.E. (1978) Biochem. J. 173, 245-254. Niemann, H., Watanabe, K., Hakomori, S., Childs, R.A. and Feizi, T.(1978) Biochem. Biophys. Res. Commun. 81, 1286-1293. Feizi, T., Childs, R.A., Watanabe, K. and Hakomori, S.I. (1979) J. Exp. Med. 149, 975-980. Watanabe, K., Hakomori, S., Childs, R.A. and Feizi, T. (1979) J. Biol. Chem. 254, 3221-3228. Feizi, T., Wood, E., Auge, C., David, S. and VeyriBres, A. (1978) lmmunochemistry 15, 733-736. Gooi, H.C., Veyrieres, A,, Alais, J., Scudder, P, Hounsell, E.F. and Feizi, T. (1984) Mol. Immunol. 21, 1099-1 104. Feizi, T., Hounsell, E.F., Alais, J., VeyriBres, A. and David, S. (1992) Carbohydr. Res. 228, 289-297. Watanabe, K. and Hakomori, S. (1976) J. Exp. Med. 144, 644653. Fukuda, M.N., Fukuda, M. and Hakomori, S. (1979) J. Biol. Chem. 254, 5458-5465 Gooi, H.C., Hounsell, E.F., Edwards, A,, Majdic, O., Knapp, W. and Feizi, T. (1985) Clin. Exp. Immunol. 60, 151-158.

584 [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] 1591 [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81]

[82] [83] [84] [85]

Feizi, T. (1980) Blood Trans. Immunohaematol. 23, 563-577. Feizi, T. and Kabat, E.A. (1972) J. Exp. Med. 135, 1247-1258. Dzierzkowa-Borodej, W., Seyfried, H. and Lisowska, E. (1975) Vox Sang. 28, 110-122. Pennington, J.E., Rastan, S., Roelcke, D. and Feizi, T. (1985) J. Embryol. Exp. Morphol. 90, 335-361. Fenderson, B.A., Eddy, E.M. and Hakomori, S. (1988) Differentiation 38, 124-133. Muramastu, T., Gachelin, G., Damonneveille, M., Delarbre, C. and Jacob, F, (1979) Cell 18, 183-191. Muramastu, T., Gachelin, G. and Jacob, F. (1979) Biochim. Biophys. Acta 587, 392406. Muramastu, T., Condamine, H., Gachelin, G. and Jacob, F. (1980) J. Embryol. Exp. Morphol. 57,25-36. Muramastu, T., Gachelin, G., Nicolas, J.F., Condamine, H., Jocob, F. and Jokob, H. (1978) Proc. Natl. Acad. Sci. USA 75, 2315-2319. Thorpe, S.J., Bellairs, R. and Feizi, T. (1988) Development 102, 193-210. Childs, R.A., Kapadia, A. and Feizi, T. (1980) Eur. J. Immunol. 10, 379-384. Childs, R.A., Feizi, T. and Tonegawa, Y. (1979) Biochem. J. 181, 533-538. Masamune, H., Kawasaki, H., Abe, S., Oyama, K. and Yamaguchi, Y. (1958) J. Exp. Med. 68, 81-91. Davidsohn, I., Kovarik, S. and Lee, C.L. (1966) Arch. Pathol. 81, 381-390. Feizi, T., Turberville, C. and Westwood, J.H. (1975) Lancet ii, 391-393. Picard, J.K., Waldron Edward, D. and Feizi, T. (1978) J. Clin. Lab. Immunol. 1 , 119-128. Feizi, T., Gooi, H.C., Childs, R.A., Picard, J.K., Uemura, K., Loomes, L.M., Thorpe, S.J. and Hounsell, E.F. (1984) Biochem. SOC.Trans. 12, 591-596. Tang, P.W., Gooi, H.C., Hardy, M., Lee, Y.C. and Feizi, T. (1985) Biochem. Biophys. Res. Commun. 132, 4 7 4 4 8 0 . Feizi, T., Stoll, M.S., Yuen, C.-T., Chai, W. and Lawson, A.M. (1994) Methods Enzymol. 230,484-519. Childs, R.A. and Feizi, T. (1981) Biochem. Biophys. Res. Commun. 102, 1158-1 164. Childs, R.A., Dalchau, R., Scudder, P., Hounsell, E.F., Fahre, J.W. and Feizi, T. (1983) Biochem. Biophys. Res. Commun. 1 1 0 , 4 2 4 4 3 1. Hounsell, E.F. and Feizi, T. (1982) Med. Biol. 60, 227-236. Fukuda, M. (1991) Glycohiology 1, 347-356. Childs, R.A., Pennington, J., Uemura, K., Scudder, P., Goodfellow, P.N., Evans, M.J. and Feizi, T. (1983) Biochem. J. 215, 491-503. Uemura, K., Childs, R.A., Hanfland, P. and Feizi, T. (1983) Biosci. Rep. 3, 577-588. Knapp, W. and Feizi, T. (1985) Uemura, K., Macher, B.A., DeGregorio, M., Scudder, P, Buehler, .I., Biochim. Biophys. Acta 846, 26-36. Feizi, T., Kapadia, A. and Yount, W.J. (1980) Proc. Natl. Acad. Sci. USA 77, 376-380. Feizi, T. and Loveless, R.W. (1996) Am. J. Respirat. Crit. Care Med. In press. Feizi, T., Taylor-Robinson, D., Shields, M.D. and Carter, R.A. (1969) Nature 222, 1253-1256. Loomes, L.M., Uemura, K., Childs, R.A., Paulson, J.C., Rogers, G.N., Scudder, P., Michalski, J., Hounsell, E.F., Taylor-Robinson, D. and Feizi, T. (1984) Nature 307, 560-563. Loomes, L.M., Uemura, K. and Feizi, T. (1985) Infect. Immun. 47, 15-20. Loveless, R.W., Griffiths, S., Fryer, P.R., Blauth, C. and Feizi, T. (1992) Infect. Immun. 60,40154023. Loveless, R.W. and Feizi, T. (1989) Infect. Immun. 57, 1285-1289. Roelcke, D., Kreft, H., Northoff, H. and Gallasch, E. (1991) Transfusion 31, 627-630. Krivan, H.C., Olson, L.D., Barile, M.F. and Ginsburg, V. (1989) J. Biol. Chem. 264, 9283-9288. Lo-Guidice, J.M., Perini, J.M., Lafitte, J.-J., Ducourouble, M.-P., Roussel, P. and Lamblin, G. (1995) J. Biol. Chem. 270, 27544-27550. de Waard, A., Hickman, S. and Kornfeld, S. (1976) J. Biol. Chem. 251, 7581-7587. Barondes, S.H., Castronovo, V, Cooper, D.N.W., Cummings, R.D., Drickamer, K., Feizi, T., Gitt, M.A., Hirabayashi, J., Hughes, C., Kasai, K., Lefffer, H., Liu, E-T., Lotan, R., Mercurio, A.M., Monsigny, M., Pillai, S., Poirier, F., Raz, A., Rigby, P.W.J., Rini, J.M. and Wang, J.L. (1994) Cell 76, 597 Kornfeld, R. (1978) Biochemistry 17, 1415-1423. Childs, R.A. and Feizi, T. (1979) FEBS Lett. 99, 175-179. Zhou, Q. and Cummings, R.D. (1993) Arch. Biochem. Biophys. 300, 6-17. Solomon, J.C., Stoll, M.S., Penfold, P., Abbott, W.M., Childs, R.A., Hanfland, P and Feizi, T. (1991) Carbohydr. Res. 213, 293-307.

585 [86] Knibbs, R.N., Agnval, N., Wang, J.L. and Goldstein, 1.J. (1993) J. Biol. Chem. 268, 14940-14947. [87] Feizi, T., Solomon, J.C., Yuen C.-T., Jeng, K.C.G., Frigeri, L.G., Hsu, D.K. and Liu F,-T. (1994) Biochemistry 33, 6342-6349. [88] Sato, S. and Hughes, R.C. (1995) J. Biol. Chem. 267, 6983-6990. [89] Lotan, R., Belloni, P.N., Tressler, R.J., Lotan, D., Xu, X.-C. and Nicolson, G.L. (1994) Glycoconj. J. 1 1 , 462468. [90] Childs, R.A. and Feizi, T. (1980) Cell Biol. Int. Rep. 4, 775 [91] Beyer, E.C. and Barondes, S.H. (1980) J. Supramol. Struct. Cell. Biochem. 13, 219-227. [92] Moutsatsos, I.K., Wade, M., Schindler, M. and Wang, J.L. (1987) Proc. Natl. Acad. Sci. 84, 6452-6456. [93] Carding, S.R., Thorpe, S.J., Thorpe, R. and Feizi, T. (1985) Biochem. Biophys. Res. Commun. 127, 680-686. [94] Wang, J.L., Werner, E.A., Lamg, J.G. and Patterson, R.J. (1992) Biochem. SOC.Trans. 20, 269-274. [95] Barondes, S.H., Cooper, D.N.W., Gitt, M.A. and LefRer, H. (1994) J. Biol. Chern. 269, 20807-20810. [96] Cooper, D.N.W. and Barondes, S.H. (1990) J. Cell Biol. 110, 1681-1691. [97] Childs, R.A., Gregoriou, M., Scudder, P., Thorpe, S.J., Rees, A.R. and Feizi, T. (1984) EMBO J. 3, 2227-2233. [980] Leffler, H. and Barondes, S.H. (1986) J. BIoL Chem. 261, 10119-10126. [991] Feizi, T. and Childs, R.A. (1985) Trends Biochem. Sci. 10, 24-29. [IOO] Bourne, Y., Bolgiano, B., Liao, D.-I., Strecker, G., Cantau, P, Herzberg, O., Feizi, T. and Cambillau, C. (1994) Struct. Biol. I , 863-870. [I011 Mandal, D.K. and Brewer, C.F. (1992) Biochemistry 31, 8465-8472. [I021 Defize, L.H.K., Arndt-Jovin, D.J., Jovin, T.M., Boonstra, J., Meisenhelder, J., Hunter, T., de Hey, H.T. and de Laat, S.W. (1988) J. Cell Biol. 107, 939-949. [I031 Frigeri, L.G., Zuberi, R.I. and Liu, F.-T. (1993) Biochemistry 32, 7644-7649. [I041 Koers, W.J., Houben, G.F. and Berrens, L. (1989) Allerg. Immunol. 35, 167-172. [I051 Villarroya, H., Dalix, A.M., Paraut, M. and Oriol, R. (1990) Autoimmunity 6, 47-60. [I061 Perillo, N.L., Pace, K.E., Seilhamer, J.J. and Baum, L.G. (1995) Nature 378, 736-739. [ 1071 Poirier, E and Robertson, E.J. (1 993) Development I 19, 1229-1236. [I081 Solter, D. and Knowles, B.B. (1978) Proc. Natl. Acad. Sci. USA 75, 5565-5569. [I091 Nudelman, E., Hakomori, S., Knowles, B.B., Solter, D., Nowinski, R.C., Tam, M.R. and Young, W.W. ( I 980) Biochem. Biophys. Res. Commun. 97, 4 4 3 4 5 I . [I101 Hakomori, S., Nudelman, E., Levery, S., Solter, D. and Knowles, B.B. (1981) Biochem. Biophys. Res. Commun. 100, 1578-1586. [ I l l ] Hounsell, E.F., Gooi, H.C. and Feizi, T. (1981) FEBS Lett. 131, 279-282. [ I 121 Kannagi, R., Nudelrnan, E., Levery, S.B. and Hakomori, S. (1982) J. Biol. Chem. 257, 14865-14874. [ I 131 Lloyd, K.O., Kabat, E.A. and Licerio, E. (1968) Biochemistry 7, 2976-2990. [ I 141 Marr, A.M.S., Donald, A.S.R. and Morgan, W.T.J. (1968) Biochem. J. 110, 789-791. [I151 Yang, H.J. and Hakomori, S. (1971) J. Biochem. (Japan) 246, 1192-1200. [I 161 Hakomori, S. and Kobata, A. (1974) In: M. Sela (Ed.), The Antigens, Vol. 11. Academic Press, San Diego, CA, pp. 79-140. [ 1171 Gooi, H.C., Thorpe, S.J., Hounsell, E.F., Rumpold, H., Kraft, D., Forster, 0. and Feizi, T. ( I 983) Eur. J. Immunol. 13, 306-312. [ I 181 Fox, N., Damjanov, I., Knowles, B.B. and Solter, D. (1983) Cancer Res. 43, 669-678. [ I 191 Yarnamoto, M., Boyer, A.M. and Schwarting. G.A. (1985) Proc. Natl. Acad. Sci. USA 82, 3045-3049. [I201 Marani, E. and Mai, J.K. (1992) Histochem. J. 24, 852-868. [121] Streit, A,, Yuen, C.-T., Loveless, R.W., Lawson, A.M., Finne, J., Schmitz, B., Feizi, T. and Stern, C.D. (1996) J. Neurochem. 66, 834-844. [I221 Thorpe, S.J. and Feizi, T. (1984) Biosci. Rep. 4, 673-685. [I231 Feizi, T. (1983) In: D. Evered and J. Whelan (Eds.), Fetal Antigens and Cancer, Ciba Foundation Symposium 96 (1982). Pitrnan Books, London, pp. 2 16-22 1. [I241 Urdal, D.L., Brentnall, T.A., Bernstein, I.D. and Hakornori, S. (1983) Blood 62, 1022-1026. [I251 Magnani, J.L. (1984) Biochem. SOC.Trans. 12, 543-545.

586 [I261 Fukuda, M.N., Dell, A,, Oates, J.E., Wu, P., Klock, J.C. and Fukuda, M. (1985) J. Biol. Chem. 260, 1067-1 082. [I271 Spooncer, E., Fukuda, M., Klock, J.C., Oates, J.E. and Dell, A. (1984) J. Biol. Chem. 259,47924801. [I281 Mizugochi, A,, Takasaki, S., Maeda, S. and Kobata, A. (1984) J. Biol. Chem. 259, 11949-1 1957. [I291 Macher, B.A., Buehler, J.,Scudder, P., Knapp, W. andFeizi,T. (1988) J. Biol. Chem. 263, 10186-10191. [I301 Fukuda, M., Bothner, B., Ramsamooj, P., Dell, A,, Tiller, P.R., Varki, A. and Klock, J.C. (1985) J. Biol. Chem. 260, 12957-12967. [I311 Fukuda, M.N., Dell, A,, Tiller, P.R., Varki, A., Klock, J.C. and Fukuda, M. (1986) J. Biol. Chem. 261, 2376-2382. [I321 Bird, J.M. and Kimber, S.J. (1984) Dev. Biol. 104, 449460. [I331 Fenderson, B.A., Zehavi, U. and Hakomori, S. (1984) J. Exp. Med. 160, 1591-1596. [I341 Rastan, S., Thorpe, S.J., Scudder, P., Brown, S., Gooi, H.C. and Feizi, T. (1985) J. Embryol. Exp. Morphol. 87, 115-128. [I351 Eggens, I., Fenderson, B., Toyokuni, T., Dean, B., Stroud, M. and Hakomori, S. (1989) J. Biol. Chem. 264, 9476-9484. [I361 Shepherd, V.L., Lee, Y.C., Schlesinger, P.H. and Stahl, P.D. (1981) Proc. Natl. Acad. Sci. USA 78, 1019-1 022. [I371 Childs, R.A., Drickamer, K., Kawasaki, T., Thiel, S., Mizuochi, T. and Feizi, T. (1989) Biochem. J. 262, 131-138. [I381 Crocker, P.R. and Feizi, T. (1996) Curr. Opin. Struct. Biol. 6, 679-691. [I391 Hakomori, S. (1996) In: J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins and Disease, Comprehensive Biochemistry, Vol. 30. Elsevier, Amsterdam, pp. 243-276. [I401 Irimura, T., Nakamori, S., Matsushita, Y., Taniuchi, Y., Todoroki, N., Tsuji, T., Izumi, Y., Kawamura, Y., Hoff, S.D., Cleary, K.R. and Ota, D.M. (1993) Cancer Biol. 4,319-324. [I411 Kageshita, T., Hirai, S., Kimura, T., Hanai, N., Ohta, S. and Ono, T. (1995) Cancer Res. 55, 1748-1751. [I421 Hakomori, S. (1995) In: J.P. Cartron and P. Rouger (Eds.), Molecular Basis of Major Human Blood Group Antigens. Plenum Press, New York, pp. 421443. [I431 Lennarz, W.J. and Hart, G.W. (1994) In: W.J. Lennarz and G.W. Hart (Eds.), Guide to Techniques in Glycobiology. Academic Press, San Diego, CA.

J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins II Elsevier Science B.V! CHAPTER 16

Cell adhesion and recognition mechanisms in the nervous tissue Jean-Pierre Zanetta Centre National de la Recherche Scientifque. Center of Neurochemistiy, 5, rue Blaise Pascal, 67000 Strasbourg, France

Abbreviations CNS

central nervous system

MBP

CSL

cerebellar soluble lectin

MOG

myelin basic protein myelin-oligodendrocyte glycoprotein

EAE

experimental allergic encephalomyelitis

PNS

peripheral nervous system

MAG

myelin-associated glycoprotein

I . Introduction The formation of a structure as complex, structurally and functionally, as the nervous tissue necessarily involves specific cell contacts, adhesion and recognition processes. The formation of the specialized contacts between neurons in the synapses is only one aspect of this complexity. As a matter of fact, the formation of this tissue can be schematized in several steps as follows. After the initial phase of neuronal cell proliferation, the cell bodies of immature neurons undergo migration on matrixes or along pre-existing astrocytic fibers to reach their adult location. Later on, or simultaneously, these neuronal cells grow axons, sometimes for very long distances, which have to reach the proper targets (generally neurons but also muscles and sensitive targets). What are the molecular mechanisms involved in directing axons to grow in the proper direction, following defined ways and passing through various structures without establishing contacts? How is the true contact established, systematically or after specific cell recognition mechanisms? Why, after a plethoric formation of synapses, are some of them degenerating and sometimes their cell body as well? At this stage, the nervous system is still not perfectly functional and later stages of wrapping by glial cells are taking place: myelination of some axons (and why only some axons?), and wrapping of neuronal cell bodies or clusters of neurons by astrocytic veils. These steps represent a crude scheme defining the major problems to be solved before understanding the formation of the nervous tissue. Besides biology, the molecular mechanisms sustaining the specific formation of such a complex structure are of crucial importance, and hence a fundamental role of glycobiological processes is expected.

2. Mechanisms of neuronal migration Two types of neuronal migration have been discriminated depending upon the stage of differentiation of the nervous tissue. In the early stages of migration of neural 587

588

cells from the neural crest, the guideline has been suggested to be constituents of the extracellular matrix: fibronectin, laminin and proteoglycans [ 1-41. Evidence was obtained through the inhibitory effect of anti-fibronectin Fab fragments. However, immunocytochemical studies indicate that this matrix [5,6] is also rich in HNK-1 epitope (glucuronic acid-3-sulfate) which has been involved in cell adhesion mechanisms, as well as N-CAM and N- and T-cadherins [7,8]. The importance of fi- 1,4-gaIactosyI-transferase has been documented [9]. In fact, due to difficulties in performing immunohistochemical studies at the ultrastructural level in this embryonic tissue, it is not well established what are the precise localizations of these antigens, and if they are actually in contact with neural cells. Recent experiments of gene inactivation represent a break-through in the understanding of these mechanisms. A special role was attributed to the embryonic form of N-CAM during this period. Besides potential adhesive properties due to homophilic interactions, this N-CAM shares the HNK-1 carbohydrate epitope (which has been shown to be involved in adhesion mechanisms) and the polysialosyl a-2,s- sequences (which were suggested to play anti-adhesive roles). When the N-CAM gene was inactivated [lo], the embryonic development, including migration of neural crest cells, is normal, although N-CAM and its polysialosyl a-2,8 sequences are not expressed at all. This indicates that N-CAM and its associated glycans do not have a significant role in this type of migration. In contrast, the inactivation of the GlcNAc-transferase I gene [ 1 I , 121, provokes abnormal formation of the neural tube and death of the embryo at mid-gestation. These results indicate an essential role of hybrid-type and complex N-glycans in these mechanisms. Interestingly, since the GlcNAc-transferase I is expressed in earlier stages in embryo as an active enzyme, this suggests that a switch-over in glycobiological interactions is occurring at this period, which makes these types of glycans indispensable. In later stages, the migration of neurons is clearly along astrocytic processes (Fig. 1) through a mechanism called “contact guidance” [ 131. The surface of migrating neurons is in close apposition to the guide, the radial glia. This mechanism has been extensively studied in the developing rat cerebellum either in oioo and in oitro. In the latter case, the cultures of cerebellar explants allow to follow after short time labelling of proliferating neurons, their migration from the external germinal layer to the internal granular layer by autoradiography. The molecular mechanism of migration has been studied by a large number of groups working on different hypotheses. Attempts to inhibit migration using neuraminidase treatments or ganglioside-specific toxins were unsuccessful [ 141. No inhibition was obtained using either chondroitinase ABC treatments or incubation of the explants with glycosaminoglycans [ 151. Consequently, it was concluded that neither gangliosides nor proteoglycans were involved in the contact guidance mechanism, and more generally in migration. A large variety of Fab fragments produced from antibodies to various glycoproteins were also tested with variable results. The neural cell adhesion molecule, N-CAM, the N-cadherin, the cytotactin, do not appear to be involved in this mechanism [14]. Significant inhibition of migration was obtained using anti-Ll/Ng-CAM [ 15-17] and anti-astrotactin antibodies [ 181 and using HNK-1 and anti-13 antibodies [ 17,19,20], specific for glucuronic-3-sulfate and MansGlcNAc2 epitopes, respectively. An almost complete inhibition of migration was obtained with small amounts of Fab fragments directed against the cerebellar soluble lectin (CSL)

589 P I A HATER

Fig. I . Left: Comprehensive scheme of the mechanism of contact migration in the cerebellum. Neurons (N, granule cells) proliferating in the external germinal layer (EGL) close to the pia mater constitute densely packed layers. In area a, they are detaching the one from the other and make intimate contacts with pre-existing processes of astrocytes (Ast). Migration occurs (area b). When neurons arrive close to the cell body of astrocytes (area c), they leave their guide to reach the internal granular layer (IGL). Right: Suggested molecular mechanism of adhesion between granule cells and astrocyte fibers during contact guidance of cell migration [21]. The lectin CSL synthesized by astrocytes is externalized and makes bridges between N-glycans (solid circles) of ligands at the neuronal (Neu) surface (one major being the P3l) and ligands at the surface of the astrocyte fiber. The fact that the P3 1 is a highly N-glycosylated GPI-anchored glycoprotein may allow a strong transverse interaction with astrocytes and a free mobility in the plane of the membrane (without interactions with the neuronal cytoskeleton).

and against the P3 1 glycoprotein, a major ligand of CSL on the neuronal surface [21]. The latter is structurally and immunologically related to the B-cell antigen CD24, the heat stable antigen and to nectadrin [22-251. This glycoprotein, also important for the homotypic adhesion between neurons [23], comprises only 3 1 amino-acids, has a glycosyl-phosphatidyl inositol (GPI) anchor and shares 4 N-glycans [22]. It is transiently expressed at the neuronal cell surface [26]. Although the structure of its N-glycans is unknown, they are binding to concanavalin A and a small part of the N-glycans transiently expresses the HNK- 1 epitope [26]. Therefore, the most probable mechanism of contact between astrocytes and migrating neurons is the formation of bridges between glycoprotein N-glycans of the surface of both partners by the polyvalent lectin CSL (Fig. 1). Adhesion is not the only mechanism involved in migration (Fig. 1). First, neurons have to detach, individually or by groups, from the external germinal layer, where neurons are highly compacted and have close metabolic relationships. Thus, the first step involves proteases as demonstrated by the action of protease inhibitors on neuron migration [27]. After migration, the neurons are leaving their guide by an unknown mechanism. But, due to the involvement in guidance of the P31, a highly N-glycosylated GPI-anchored glycoprotein, it may be assumed that besides proteases [27], exo- or endo-glycosidases and phosphatidylinositol-specificphospholipases C and D may be important.

590

3. Mechanisms of axonal growth The hypotheses concerning the molecular mechanisms involved in axonal growth are founded essentially on two views. The first is to consider that axonal growth is directed by neurotrophic substances, produced by target cells and recognized by the growth cones of axons, attracting and guiding the axons to the target. The second is the hypothesis of a preformed way along which axons are growing [28]. The situation probably involves both kinds of mechanisms. In uitro studies indicate that a neurotrophic effect takes place, although, there is no evidence for it in uiuo. Nevertheless, it can be assumed that it occurs actually, as for other cells in the field of biology. The nature of these postulated neurotrophic substances is unknown, but evidence was provided that, during growth of incoming axons, groups of target neurons are biochemically different and consequently are potentially capable of producing specific neurotrophic substances [29]. Pictures of the preformed way have been observed, particularly using anti-carbohydrate antibodies like CC1 specific for the 0-GalNAc determinant [30,3 I] or for the CD15 antigen [32,33]. Interactions between glycoconjugates and lectins have been suggested in the dorsal root ganglia [34-391. Based on developmental immunohistochemical studies at the ultrastructural level (immunoprecipitate in-between growing axons [40,4 l]), it was suggested that galectins could be involved in neurite fasciculation (axons do not grow as isolated axons but make bundles following a pioneer axon). This hypothesis found experimental support recently [42]. Besides their role in neurite fasciculation (Fig. 2), it might be postulated that the externalized galectins inhibit the interaction of axons with the surrounding extracellular matrix containing laminin. Indeed, it has been demonstrated that the L-14 galectin externalized by myoblasts binds to glycans of laminin in the extracellular matrix but not to cell surface oligosaccharides. This binding inhibits the laminindependent adhesion of myoblasts on the matrix, allowing myoblasts to migrate and fuse[43]. In fact, there is no in oioo biological evidence for this second role. For example, in the peripheral nervous system (PNS), axons are growing along immature Schwann cells, which produce a laminin-containing basal lamina, but morphological AXON

Galectin

Ga lect in

Fig. 2. Left: Scheme showing how axons are growing as fascicles in a matrix. Right: Hypothetical scheme of a molecular mechanism of neurite fasciculation involving galactose residues (Gal) mostly associated with glycolipids [Zanetta, unpublished observations] at the surface of growing axons and extracellular galectins [40, 41 1. The cellular origin of the extracellular galectins (neuronal or astrocytic) remains unknown. These galectins are no more detectable around axons in areas where they are contacting their target cells in order to make synapses. A role of laminin as a matrix for axonal growth in uiuo is not likely.

59 1

studies indicate that growing axons are never in contact with this matrix. Similarly, central nervous system (CNS) neurons are growing axons through an extracellular matrix where laminin is absent. Furthermore, the neurite outgrowth observed in uitro on a laminin substratum is not modified by anti-laminin antibodies, suggesting that the effect was due to another unidentified factor [44]. Other molecules have been involved in neurite fasciculation. The initial view of a role of N-CAM was retracted [39]. It is worth mentioning that some growing axons are rich in the embryonic form of N-CAM having polysialosyl a-2,8 sequences, considered as inhibitory of cell adhesion mechanisms [45]. But virtually normal axonal growth is occurring in mutant mice with an inactivated N-CAM gene [ 101 and the embryonic form of N-CAM is not re-expressed during axonal regrowth after lesion in the adult CNS [46]. Arguments for the involvement of Ll/Ng-CAM were obtained through the effects of Fab fragments (at high concentrations) which are capable of significantly reducing neurite fasciculation [l 11 and the positive effect of LI/Ng-CAM on the neurite outgrowth in uitro [47]. Due to different results obtained using different experimental models and in uiuo and in oitro experiments, the fine biochemical mechanisms of the neurite fasciculation and axonal growth remain largely speculative. The cultures of undissociated dorsal root ganglia grown in the presence of nerve growth factor (NGF) may provide an easy experimental model for testing the different molecular hypotheses.

4. Mechanisms of synaptogenesis Since the specificity of the formation of nerve connections is the most important functional event in the formation of the nervous tissue, due to the complexity of the networks, it was suggested that highly specific cell recognition mechanisms were taking place. The initial chemio-affinity hypothesis “one couple of neurons, one couple of complementary recognition molecules” [48], is not supported because of the insufficient information in the human or animal genome. This theory was inflected to that of “group to group recognition”. However, the concept of recognition itself was challenged because of the results of in uitro experiments: neurons in cultures are capable of forming contact, with presynaptic differentiations, on positively charged beads. Mixtures of neurons in cultures can make synapses the one with the other, not respecting the specificity observed in uiuo. However, the results of these experiments cannot be extrapolated to the in uiuo situation for several reasons: (i) neurons in cultures loose the major characteristic of the neuron: its polarity (in uitro, the molecules distribute uniformly on the cell body, dendrite and axon, in contrast with the in oivo situation); (ii) the surface molecules expressed in uitro are not necessarily the same as in uiuo; (iii) the timing and modalities of expression of these molecules could be affected by the in uitro situation. The model of the cerebellar mutant mice stuggerer is very interesting in this respect. Although afferent fibers contact their natural target, they cannot make synapses. Thus, contact or proximity alone is not sufficient for the formation of synapses: recognition is a preliminary step in uiuo. The molecular mechanism of neuron-neuron recognition has been poorly studied due to its complexity: for determining if an enzyme or an antibody or other substances are

592

es

.li

Parallel Fiber

Purkinje Cell

y t +!

I

Lectine R1 Double walled Coated Vesicles

Fig. 3. Left: Schematic drawing showing the geometry of the major contingent of synapses in the cerebellum. These synapses are formed by the axons (called the parallel fibers (PF)) of the granule neurons (G) with the dendritic spines of the terminal dendrites (PCD) of Purkinje cells. A single parallel fiber is establishing around 400 synapses (termed “en passant”) with the dendrites of about 20 different Purkinje cells. Right: Before synaptogenesis, the surface of the parallel fibers is covered by a large amount of low Mr glycoproteins, endowed with N-glycans (solid circles) binding to concanavalin A, including the P3 1 glycoprotein. At the period just before synaptogenesis, the membrane-bound dimeric lectin RI is expressed at high levels and is in part externalized at the surface of some Purkinje cell dendrites. The contact between RI and its axonal ligand is followed by a phagocytosis of parallel fiber membrane concentrating glycoprotein ligands of R I . This results in the production in the Purkinje cell dendrites of double walled coated vesicles containing RI and its ligands. The distribution of these coated vesicles in Purkinje cell dendrites is limited to those showing an increased R1 expression. Morphological evidence suggests that these structures fuse with the endosomal compartment in multivesicular bodies where the interaction lectin RI-ligand is dissociated at acidic pH, the lectin being recycled in other portions of the dendrite. The use of chloroquine, which increases the pH of the endosomal and lysosomal compartments, provokes a huge accumulation of membranes containing the two complementary molecules in non-functional lysosomes [52]. This mechanism is associated with a strong increase in a-mannosidase activity, but, surprisingly, this increase is only associated with that of a cytosolic a-mannosidase working only at neutral pH [93]. When parallel fibers are sectioned in adult animals (R1 is expressed at low level and the P31 glycoprotein is no more detectable), there is an immediate high level expression of the P31 glycoprotein in granule cell bodies then on parallel fibers, with a simultaneous increased expression of R1 in the Purkinje cell dendrites, restricted to the area of the lesion. Ultrastructural studies using specific antibodies indicate that the mechanisms occurring during normal synaptogenesis take place during re-innervation [46].

able to inhibit synapse formation, it is necessary to count synapses at the ultrastructural level, and in uitro studies could induce their own artefacts. N-CAM was proposed as a fundamental molecule for synaptogenesis, but recent experiments of inactivation of its gene [lo] are not compatible with this hypothesis. Developmental studies of the localization of galectins (using polyclonal antibodies which recognize all galectins) at the optical microscope indicated a strong transient accumulation in the neuronal dendrites (and particularly the dendritic spines, the structures on which synapses are formed) at the period of synaptogenesis in different brain areas [40,41]. However, this assumption was not verified at the ultrastructural level, since the immunoprecipitate was never observed extracellularly [40,4 11, but actually accumulated within the dendritic spines. The absence of externalization of galectins observed in uiuo in these important structures for synaptogenesis is incompatible with a role of galectins as inter-neuronal

593

recognition molecules. An alternative molecular mechanism proposed for neuronal recognition is concerned with a mannose-binding lectin-glycoprotein interaction. It has been documented for the formation of synapses between parallel fibres (incoming axons) and Purkinje cells (target neurons) in the rat cerebellum (Fig. 3). The calciumindependent, Mans GlcNAcz-specific, divalent and membrane-bound lectin R1 [49,50], intracellularly localized outside the period of synaptogenesis, is externalized at the surface of the Purkinje cell dendrites at the period of synaptogenesis [50], where it contacts specific glycoproteins transiently expressed at the surface of parallel fibers [5 11. These glycoproteins are the P31 glycoprotein (see above) and two major low Mr glycoproteins binding to concanavalin A, one of them being probably Thy-1 . This contact is followed by the specific receptor R1-mediated internalization of the glycoproteins into Purkinje cells, where glycoproteins are degraded [52]. Experiments designed for impairing the normal cerebellar development (absence of thyroid hormone, treatment with tunicamycin [53,54]) suggest that R1 and its glycoprotein ligands participate in the recognition mechanism associated with the first step of synaptogenesis. In the case of tunicamycin, the late degeneration of neurons not having succeeded in forming synapses with their targets resembles the “en cascade” degeneration observed in the stuggever mutant [55]. The hypothesis of the role of lectin R1 in neuronal recognition has been substantiated by the finding that the two complementary molecules (lectin R1 at the surface of the dendrites of Purkinje cells and the P31 glycoprotein at the surface of parallel fibers) are re-expressed in adult rat cerebella during repair 2 4 4 8 h after sections of parallel fibers [46]. Lectin R1-mediated internalization occurs thereafter as a first step of repair in the adult animal. This mechanism may provide an explanation for the increased number of “double walled coated vesicles” observed at the ultrastructural level during the period preceding synaptogenesis in various parts of the CNS [56-601. But, the stabilization of the specific contact in the synapse is certainly supported by another mechanism, since the lectin is not present in mature synapses. Several studies indicated that synaptic junctions are rich in glycoconjugates and specially glycoproteins [61,62] and that the mechanisms involved in this complex contact are essentially calcium-dependent [63].

5. Mechanisms of glial wrapping The final stage of maturation of the nervous tissue involves two important steps, the astrocyte wrapping of neurons or groups of neurons and the myelination of axons. Although functionally important, the former mechanism has not been studied in contrast with myelination. Myelin is a compact structure made of the rolling up of the processes of oligodendrocytes in the CNS (Fig. 4a) and Schwann cells in the PNS with one major difference: oligodendrocytes can myelinate up to forty different axons, whereas Schwann cells are specific for one axon. This wrapping is separated from the axonal surface in most areas with the exception of the lateral loops of the nodes of Ranvier (the nodes of Ranvier are the spaces between two myelin sheaths), where specific junctions between axons and myelinating cells are present. Before starting the myelination process, the myelinating cells have to recognize the axonal surface. This specific contact is demonstrated by ultrastructural studies and by the evidence that axonal membranes can stimulate the

594

Fig. 4. (a) Comprehensive scheme of the structure of the CNS myelin (produced by oligodendrocytes). When observed at the electron microscope, myelin shows an alternation of dense lines (corresponding to the apposition of the cytoplasmic faces of the membrane of the processes of oligodendrocytes; in black) and the intraperiodic lines (corresponding to the apposition of the extracellular surfaces; in white). Note the presence of a penaxonal space (OUT), where the adult form of MAG is located, which does not allow contact between myelin and the axonal membrane. The regions of contact are restricted to the lateral loop of the nodes of Ranvier, where special junctions are found. (b) Hypothesis on the molecular mechanism involved in the first contact between axons and myelinating cells and in the formation of the junction in the lateral loop of the nodes of Ranvier. CSL produced and externalized by myelinating cells makes bridges between the N-glycans (solid circles) of its ligands at the surface of the two cells. The “early” MAG is the major ligand at the surface of the myelinating cell and P31 the major ligand at the surface of the axon. The fact that this major axonal ligand of CSL is a GPI-anchored glycoprotein (with a free mobility in the plane of the membrane and not interacting with the cytoskeleton) may explain why the axons (and their neurofilaments and microtubules) are not twisted during myelination. Such junctions may behave as ball bearings. (c) Hypothetical mechanisms involved in the compaction of PNS myelin. Myelin basic protein (MBP) concentrated close to the surface is supposed to play a role in the association between the cytoplasmic faces, possibly in association with sulfatides. For the apposition of the extracellular faces, the lectin CSL makes punctual bridges between a specific N-glycans (solid circles) of the PO glycoprotein, whereas other N-glycans (solid circles and open squares) are not involved in the interaction.

proliferation of immature myelinating cells at the period preceding myelination. It was also assumed that myelination takes place only for axons having a sufficiently large diameter. This view is contradicted by the observation that small diameter axons can be myelinated whereas some intermediate size or large axons are not. All compounds involved in the contact between axon and myelinating cells are glycoproteins. Most of them are members of the superfamily of immunoglobulins: the myelin-associated glycoprotein, MAG, on myelinating cells, the L l/Ng-CAM and the N-CAM antigens on axons or a splicing variant of Ll/Ng-CAM on Schwann cells. Based on their localization, the majority of MAG molecules do not seem to be involved in the

595

contact with axons, since most of it is found in the no-man’s land between these structures. A specific role of MAG in keeping the space between the two surfaces (anti-adhesive role) has been proposed[64,65]. But one part of MAG is localized in the areas of contacts as demonstrated by immunogold labelling [66,67]. Its role in the interaction with axons has been documented through the inhibitory action of anti-MAG Fab fragments [68]. Ll/Ng-CAM was also involved based on the effect of specific Fab fragments [69,70]. Homophilic interactions are supposed to involve the homophilic binding sites of MAG and the variant of Ll/Ng-CAM on myelinating cells and Ng-CAM on the axonal surface, but experiments in favour of such a hypothesis are lacking. In contrast, evidence was provided that HNK-1 epitopes shared by the multi-sulfated N-glycans of MAG [71] are involved in contacts [68]. The endogenous lectin CSL [72] is also present in the areas of contacts as well as its neuronal ligand, the P3 1 glycoprotein [26] and one minor form of MAG [73751. Two different alternative splicing variants of MAG (the “early” and “late” forms) have been identified [76], their protein sequences differing only by the intracytoplasmic domains. They differ from their solubility properties, the “late” form being solubilized by lithium di-iodosalicylate, whereas the “early” form remains insoluble. The two forms have N-glycans containing the HNK-1 epitope, but only the “early” form is interacting with CSL [73], indicating that the two forms of the rat MAG have different N-glycans. In fact, Schwann cells from newborn rats (which have the “early” MAG and CSL at their surface) do not adhere to a substratum of immobilized “late” MAG (which is not CSL-binding) [77]. But these cells are adhering to the PO glycoprotein which shares CSLbinding N-glycans [77]. Since this binding persists when PO is denatured (boiling in SDS in the presence of reducing agents), it is suggested that the important factors for adhesion are the N-glycans and not the polypeptide chain. Consequently, two different mechanisms are proposed for explaining the contact between axons and myelinating cells, one involving homophilic interactions and the other lectin-carbohydrate interactions. For the latter, it is proposed (Fig. 4b) that the lectin CSL produced by myelinating cells is externalized and makes bridges between surface glycans shared by the “early” MAG polypeptide chains [73] and, possibly, the variant of Ll/Ng-CAM on myelinating cells [75] and the P3 1 glycoprotein [74] and, possibly, Ll/Ng-CAM and N-CAM on the axonal surface. This glycobiological hypothesis can also explain the mitogenic effect of axonal membranes on Schwann cells [78,79]. Based of the ability of heparin to solubilize the mitogenic activity from axonal membranes, it was assumed that the mitogen was a molecule bound to a heparan sulfate proteoglycan of the axonal membrane. In fact, recent studies [75] indicate that the mitogen is the lectin CSL bound to its axonal glycoprotein ligands. The mitogenic activity is inhibited by anti-CSL monovalent antibodies and can be solubilized by low concentration of Man6GlcNAcz N-glycans [75]. The fact that the mitogen is also solubilized by heparin is explained by the heparin-binding properties of CSL (heparin inhibits the binding of CSL to ligands containing only N-glycans on blots). A role for homophilic interactions in the stabilization of compact myelin has been suggested. This hypothesis is sustained by the adhesive role of PO glycoprotein, the major glycoprotein of the PNS. This glycoprotein is a short member of the superfamily of immunoglobulins which has only one occupied N-glycosylation site. The structures of the five N-glycans of the bovine PO have been determined[80]. These hybrid-type N-glycans are all fucosylated on the first GlcNAc and sulfated on the second GlcNAc

596

of the core. One of them is sharing the HNK-1 epitope on the lactosaminic branch. The role of PO in adhesion was evidenced by transfection experiments with the gene of this molecule [81]. However, transfection of this gene in CHO cell mutants lacking GlcNActransferase I is not followed by increased adhesion [82] although PO is expressed at the cell surface as in the transfected wild-type CHO cells, indicating an important role of the N-glycans. The demonstration that increased adhesion after transfection in C6 cells is completely inhibited by the glycopeptide fraction prepared from endogenous PO [83] indicates that N-glycans are directly involved in the adhesion process. It is noteworthy that the lectin CSL, binding to PO N-glycans, is present in both CHO and C6 cells, where it plays an important role in the aggregation process of these cells [84,85]. But, besides PO, other glycoproteins may play a role in myelin stabilization. For example, PO is absent from the CNS myelin. Low Mr not yet identified glycoproteins were detected using labelled CSL, which are largely reduced or absent in dysmyelinating mutants. The demonstration that anti-CSL Fab fragments are capable of destroying the structure of myelin produced in uitro [86] suggests that these minor low Mr glycoproteins could play a role similar to PO, in relationship with CSL.

6. Conclusions Although the precise mechanism of involvement of nervous glycoproteins in adhesion and recognition phenomena occurring during the ontogenesis of the nervous tissue are not known, there is a large body of evidence that glycoconjugates are playing a predominant role. A special interest is focused on HNK-I epitope found on N-glycans, but there is still no direct evidence that HNK-1 is itself involved in adhesion processes. The inhibitory properties of antibodies to HNK-1 may be due to a steric hindrance impeding the accessibility to important carbohydrate determinants of N-glycans, particularly sulfation and fucosylation of the core GlcNAc residues [71,80]. The important anti-adhesive role of the N-glycans of the PO molecule[83] and of antibodies to lectin CSL, binding the N-glycans of PO and MandGlcNAc2 N-glycans, on the one hand, the binding of lectin R1 to MangGlcNAc2 [87], on the other hand, suggests that adhesion and recognition mechanisms are mostly supported by carbohydrate-lectin interactions. The inhibition of essential steps of brain ontogenesis using castanospermine and deoxynojirimycin [88] but not using swainsonine or deoxymannojirimycin may be indicative of a very important role of oligomannosidic and hybrid-type N-glycans in the development of the nervous tissue. Unfortunately, the isolation of CNS glycoproteins is impeded by the complexity of the tissue, the low amounts of accessible material and the “strange” structure of glycoproteins of interest (like the P3 1 glycoprotein) for determining the structures of their glycans. Furthermore, it appears that the important glycoproteins (often alternative splicing variants) are transiently expressed as well as special glycans necessary for a function. The latter may be dependent on minor or special [7 1,801 or non-orthodox glycan structures [89,90]. Besides glycoproteins of the superfamily of immunoglobulins, there is increasing evidence that GPI-anchored glycoproteins may have a fundamental role in adhesion and recognition mechanisms, specially those which are heavily N-glycosylated

597

on a short polypeptide chain (P3 1 and Thy-1). For Thy-1, the 1 11 amino-acid long GPIanchored form containing three N-glycans [911 has adhesive properties different from the transmembrane form [92]. It could be of interest to have more information on the structure of the glycans of such molecules, which could be easily manipulated when detached from the membranes by a phosphatidylinositol-specific phospholipase C.

References [ I ] Perris, R., Paulsson, M. and Bronnerfraser, M. (1989) Dev. Biol. 136, 222-238. [2] Poelmann, R.E., Giittenberger-de Groot, A.C., Mentink, M.M.T., Delpech, B., Girard, N. and Christ, B. (1990) Anat. Embryol. 182, 29-39. [3] Tan, S.S., Crossin, K.L., Hoffman, S. and Edelman, G.M. (1987) Proc. Natl. Acad. Sci. USA 84, 79777981, [4] Perris, R., Krotoski, D., Lallier, T., Domingo, C., Sorrell, J.M. and Bronnerfraser, M. (1991) Development 1 1 1, 583-599. [S] Layer, P.G. and Kaulich, S. (1991) Cell Tissue Res. 265, 393407. [6] Sadaghiani, B. and Vielkind, J.R. (1990) Development 110, 197-209. [7] Bronnerfraser, M., Wolf, J.J. and Murray, B.A. (1992) Dev. Biol. 153, 291-301. [8] Ranscht, B. and Bronnerfraser, M. (1991) Development I 1 1, 15-22. [9] Hathaway, H.J. and Shur, B.D. (1992) J. Cell Biol. 117, 369-382. [lo] Cremer, H., Lange, R., Christoph, A,, Plomann, M., Vopper, G., Roes, J., Brown, R., Baldwin, S., Kraemer, P., Scheff, S., Barthels, D., Rajewsky, K. and Wille, W. (1994) Nature 367, 4 5 5 4 5 9 . [ I 11 Ioffe, E. and Stanley, P. (1994) Proc. Natl. Acad. Sci. USA 91, 728-732. [I21 Metzler, M., Gertz, A,, Sarkar, M., Schachter, H., Schrader, J.W. and Marth, J.D. (1994) EMBO J. 13, 2056-2065. [I31 Sidman, R.L. and Rakic, P. (1973) Brain Res. 62, 1-35. [I41 Fischer, G., Kunemund, V and Schachner, M. (1986) J. Neurosci. 6, 605-612. [I51 Miura, M., Asou, H., Kobayashi, M. and Uyemura, K. (1992) J. Biol. Chem. 267, 10752-10758. [I61 Asou, H., Miura, M., Kobayashi, M., Uyemura, K. and Itoh, K. (1992) Neurosci. Lett. 144, 221-224. [I71 Fahrig, T., Schmitz, B., Weber, D., Kucherer-Ehret, A,, Faissner, A. and Schachner, M. (1990) Eur. J. Neurosci. 2, 153-1 61. [I81 Edmondson, J.C., Liem, R.K.H., Kuster, J.E. and Hatten, M.E. (1988) J. Cell Biol. 105, 505-517. [I91 Kadmon, G., Kowitz, A,, Altevogt, P and Schachner, M. (1990) J. Cell Biol. 110, 209-218. [20] Kadmon, G., Kowitz, A,, Altevogt, P and Schachner, M. (1990) J. Cell Biol. 110, 193-208. [21] Lehmann, S., Kuchler, S., Theveniau, M., Vincendon, G. and Zanetta, J.-P. (1990) Proc. Natl. Acad. Sci. USA 87, 6455-6459. [22] Nedelec, J., Pierres, M., Moreau, H., Barbet, J., Naquet, P., Faivre-Sarrailh, C. and Rougon, G. (1992) Eur. J. Biochem. 2 0 3 , 4 3 3 4 4 2 , [23] Kadmon, G., Eckert, M., Sammar, M., Schachner, M. and Altevogt, P (1992) J. Cell Biol. 118, 12451258. [24] Kadmon, G., Halbach, EWJ., Schachner, M. and Altevogt, P (1994) Biochem. Biophys. Res. Commun. 198, 1209-1215. [25] Kay, R., Rosten, PM. and Humphries, R.K. (1991) J. Immunol. 147, 1412-1416. [26] Kuchler, S., Rougon, G., Marschal, P., Lehmann, S., Reeber, A,, Vincendon, G. and Zanetta, J.-P (1989) Neuroscience 33, 1 11-124. [27] Moonen, G., Grau-Wagemans, M.P. and Selak, I. (1982) Nature 298, 753-755. [28] Goodman, C.S. and Shatz, C.J. (1993) Cell 72, 77-98. [29] Wassef, M., Zanetta, J.-P., Brehier, A. and Sotelo, C. (1985) Dev. Biol. 1 1 I , 129-137. [30] Schwarting, G.A. and Crandall, J.E. (1991) Brain Res. 547, 239-248. [31] Schwarting, G.A., Deutsch, G., Gattey, D.M. and Crandall, J.E. (1992) J. Neurobiol. 23, 130-142. [32] Oudega, M., Marani, E. and Thomeer, R.T.W.M. (1992) Histochem. J. 24, 869-877.

598 [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [SO] [SI] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72]

Reifenberger, G., Sieth, P., Niederhaus, M. and Wechsler, W. (1992) Histochem. J. 24, 890-901. Dodd, J. and Jessel, T.M. (1986) Exp. Biol. 124, 225-238. Dodd, J. and Jessell, T.M. (1988) Science 242, 692-699. Dodd, J., Morton, S.B., Karagogeos, D., Yamamoto, M. and Jessell, T.M. (,1988) Neuron I , 105-1 16. Hynes, R.O., Patel, R. and Miller, R.H. (1986) J. Neurosci. 6, 867-876. Hynes, M.A., Buck, L.B., Gitt, M., Barondes, S.H., Dodd, J. and Jessel, T.M. (1989) 189-210. Karagogeos, D., Morton, S.B., Casano, F., Dodd, J. and Jessell, T.M. (1991) Development 112, 5 1 4 7 . Joubert, R., Kuchler, S., Zanetta, J.-P., Bladier, D., Avellana-Adalid, V., Caron, M., Doinel, C. and Vincendon, G. (1989) Dev. Neurosci. 11, 397413. Kuchler, S., Joubert, R., Avellana-Adalid, V., Caron, M., Bladier, D., Vincendon, G. and Zanetta, J.-P. (1989) Dev. Neurosci. 11, 414-427. Mahanthappa, N.K., Cooper, D.N.W., Barondes, S.H. and Schwarting, G.A. (1994) Development 120, 1373-1 384. Cooper, D.N.W., Massa, S.M. and Barondes, S.H. (1991) J. Cell Biol. 115, 1437-1448. David, S. (1988) J. Neurocytol. 17, 131-144. Rutishauser, U., Acheson, A., Hall, A.K., Mann, D.N. and Sunshine, J. (1988) Science 240, 53-57. Lehmann, S., Kuchler, S., Gobaille, S., Marschal, P., Badache, A., Reeber, A,, Vincendon, G. and Zanetta, J.-P. (1993) Brain Res. Bull. 30, 515-521. Hankin, M.H. and Lagenaur, C.F. (1994) J. Neurobiol. 25, 472487. Sperry, R.W. (1965) In: R.L. De Han and U. Urspring (Eds.), Organogenesis. Holt, Rinchart and Winston, New York, pp. 161-186. Zanetta, J.-P., Dontenwill, M., Meyer, A. and Roussel, G. (1985) Dev. Brain Res. 17, 233-243. Dontenwill, M., Roussel, G. and Zanetta, J.-P. (1985) Dev. Brain Res. 17, 245-252. Zanetta, J.-P., Roussel, G., Ghandour, M. S., Vincendon, G. and Gombos, G. (1978) Brain Res. 142, 301-3 19. Dontenwill, M., Devilliers, G., Langley, O.K., Roussel, G., Hubert, P, Reeber, A,, Vincendon, G. and Zanetta, J.-P. (1983) Dev. Brain Res. 10, 287-299. Zanetta, J.-P., Dontenwill, M., Reeber, A. and Vincendon, G. (1987) NATO AS1 Series H 2, 92-104. Zanetta, J.-P., Dontenwill, M., Reeber, A,, Vincendon, G., Legrand, C., Clos, J. and Legrand, J. (1985) Dev. Brain Res. 21, 1 4 . Sotelo, C. and Changeux, J.-P. (1974) Brain Res. 67, 519-526. Altman, J. (1971) Brain Res. 30, 311-322. Chen, S. and Hillman, E. (1982) Brain Res. 240, 205-220. Eckenhoff, M.F. and Pysh, 5.3. (1979) J. Neurocytol. 8, 623438. Privat, A. (1974) Brain Res. 69, 125-129. Palacios-Prii, E.L., Palacios, L. and Mendoza, R.V. (1981) J. Submicrosc. Cytol. 13, 145-167. Bezamahouta, C., Zanetta, J.-P., Clos, J., Meyer, A. and Vincendon, G. (1988) Dev. Brain Res. 40, 193200. Gurd, J.W., Brown, I.R., Bissoon, N., Cudmore, S., Ni, B. and Soullitre, J. (1992) Biochem. SOC.Trans. 20, 379-382. Sealock, R. and Froehner, S.C. (1994) Cell 77, 617-619. Trapp, B.D., Quarles, R.H. and Griffin, J.W. (1984) J. Cell Biol. 98, 1272-1278. Trapp, B.D., Quarles, R.H. and Suzuki, K. (1984) J. Cell Biol. 99, 594406. Martini, R. and Schachner, M. (1986) J. Cell Biol. 103, 2439-2448. Martini, R. and Schachner, M. (1988) J. Cell Biol. 106, 1735-1746. Poltorak, M., Sadoul, R., Keilhauer, G., Landa, C., Fahrig, T. and Schachner, M. (1987) J. Cell Biol. 105, 1893-1 899. Wood, M.W., Schachner, M. and Bunge, R.P. (1990) J. Neurosci. 10, 3635-3645. Seilheimer, B., Persohn, E. and Schachner, M. (1989) J. Cell Biol. 109, 3095-3103. Field, M.C., Wing, D.R., Dwek, R.A., Rademacher, T.W., Schmitz, B., Bollensen, E. and Schachner, M. (1992) J. Neurochem. 58, 993-1000. Zanetta, J.-P., Meyer, A., Kuchler, S. and Vincendon, G. (1987) J. Neurochem. 49, 1250-1257.

599 [73] Badache, A., Burger, D., Villarroya, H., Robert, Y., Kuchler, S., Steck, A. J. and Zanetta, J.-P. (1992) Dev. Neurosci. 14, 342-350. [74] Kuchler, S., Herbein, G., Sarlieve, L.L., Vincendon, G. and Zanetta, J.-P. (1989) Cell. Mol. Biol. 35, 581-596. [75] Badache, A., Lehmann, S., Kuchler-Bopp, S., Hand, N. and Zanetta, J.-P. (1993) Glycobiology 5, 371383. [76] Lai, C., Brow, M.A., Nave, K.-A., Noronha, A.B., Quarles, R.H., Bloom, F.E., Milner, R.J. and Sutcliffe, J.G. (1987) Proc. Natl. Acad. Sci. USA 84, 43374341. [77] Zanetta, J.-P., Badache, A., Maschke, S., Marschal, P. and Kuchler, S. (1994) Histol. Histopathol. 9, 385-412. [78] Wood, P.M. and Bunge, R.P. (1975) Nature 256, 662-664. [79] DeCoster, M.A. and DeVries, G.H. (1989) J. Neurosci. Res. 22, 283-288. [80] Uyemura, K. and Kitamura, K. (1991) Comp. Biochem. Physiol. 98, 63-72. [81] Filbin, M.T., Walsh, F.S., Trapp, B.D., Pizzey, J.A. and Tennekoon, (3.1. (1990) Nature 344, 871-872. [82] Filbin, M.T. and Tennekoon, ‘3.1. (1991) Neuron 7, 845-855. [83] Yazaki, T., Miura, M., Asou, H., Kitamura, K., Toya, S. and Uyemura, K. (1992) FEBS Lett. 307, 361366. [84] Lehmann, S., Kuchler, S., Badache, A,, Zaepfel, M., Meyer, A. and Zanetta, J.-P (1991) Eur. J. Cell Biol. 56, 433442. [85] Maschke, S., Kuchler-Bopp, S. and Zanetta, .I.-P. (1995) Eur. J. Cell Biol. (submitted). [85] Kuchler, S., Fressinaud, C., Sarlieve, L. L., Vincendon, G. and Zanetta, J.-P. (1988) Dev. Neurosci. 10, 199-21 2. [87] Marschal, P., Reeber, A,, Neeser, J.-R., Vincendon, G. and Zanetta, J.-P. (1989) Biochimie 71,645453. [88] Bhat, N.R. and Zhang, P. (1994) J. Neurosci. Res. 39, 1-10. [89] Kurosaka, A,, Yano, A,, Itoh, N., Kuroda, Y., Nakagawa, T. and Kawasaki, T. (1991) J. Biol. Chem. 266, 41 684172. [90] Kurosaka, A.,Tsuyuki, Y. and Tahara, S. (1994) Biochem. Biophys. Res. Commun. 204, 342-347. [91] Rademacher, T.W. and Dwek, R.A. (1989) In: G. Bock and S. Harnett (Eds.), Carbohydrate Recognition in Cellular Function. Wiley, Chichester, pp. 241-256. [92] Tiveron, M.C., Nosten-Bertrand, M., Jani, H., Garnett, D., Hirst, E.M.A., Grosveld, E, Morns, R.J. (1994) J. Cell Sci. 107, 1783-1796. [93] Zanetta, J.-P., Roussel, G., Dontenwill, M. and Vincendon, G. (1983) J. Neurochem. 40, 202-208.

This Page Intentionally Left Blank

J. Montreuil, J.F.G. Vliegenthart and H. Schachter (Eds.), Glycoproteins I1 0 1997 Elsevier Science B.V. All rights reserved CHAPTER 17

Neoglycoproteins Reiko T. Lee and Yuan C. Lee Department of Biology,Johns Hopkins University, Baltimore, MD 21218, USA

List of Abbreviations BSA

bovine serum albumin

MU

CMP

cytidine monophosphate

PFP

pentafluorophenyl

Con A

concanavalin A

PNP

p-nitrophenyl

GDP

guanine diphosphate

IgG LeX

immunoglobulin G

RP-HPLC reverse phase high performance liquid chromatography

4-methylumbelliferyl

Lewis x antigen

1. History and deJinition Glycoconjugates are those hybrid biochemicals which contain carbohydrate chemically bonded to some other component such as peptide (glycopeptide), protein (glycoprotein), lipid (glycolipid), and others. About twenty years ago, we prepared a series of proteins chemically modified with carbohydrates of defined structure to overcome the problem of heterogeneity of carbohydrate groups inherent in natural glycoproteins [ 11. The term “neoglycoproteins” was coined to describe such carbohydrate-modified proteins. Subsequently, the usage of “neoglycoproteins” has become quite popular, and the prefix “neo” has spread to carbohydrate-modified lipids (neoglycolipids), carbohydrate-modified polymers (neoglycopolymers) etc. Eventually a term “neoglycoconjugates” was invented to include all of these chemically or enzymatically modified carbohydrate hybrids [2]. There have been several reviews on the topics of neoglycoproteins [ 1,3,4] and neoglycoconjugates [2,591.

2. Prepa ra tion of neoglycoproteins Preparation of neoglycoproteins typically starts with natural proteins or glycoproteins [ 101. Proteins most suitable for modification with carbohydrate derivatives are those of good water solubility and those with many different types of side chains. Bovine serum albumin (BSA) is a popular protein for this purpose because it is devoid of carbohydrate and can be obtained inexpensively in a reasonable purity. Neoglycoproteins can also be prepared from bioactive proteins such as enzymes, immunoglobulins, hormones and growth factors. As will be shown in later sections, such neoglycoproteins are useful for targeting and analysis. 60 1

602

COOH

HS

Fig. 1. Side chains of proteins amenable for modification with carbohydrate derivatives.

Figure 1 illustrates the side chains of proteins useful for the formation of neoglycoproteins. Among them, the E-amino group of lysyl side chains is the most frequently used for modification of proteins, because the lysyl group is often abundant, and its side chain is flexible and solvent accessible. In the next four sections, we will describe the sugar attachment via these four functional groups.

2.1. ModiJication of primaly amino groups 2.1.1. Reductive amination Preparation of neoglycoproteins by reductive amination was pioneered more than 20 years ago [ 1 13. Many subsequently refined methods emanate from this procedure. In the original procedure, a reducing di- or oligosaccharide (having a masked carbonyl group as hemiacetal) is mixed with the protein to be modified together with a reducing agent such as NaCNBH3. The reaction is typically carried out at near neutrality and at room temperature. The reaction between the carbonyl group of a sugar and an amino group of a protein initially generates a Schiff’s base which is preferentially reduced to produce a permanent covalent bond (Fig. 2). The choice of reducing agent is quite important. NaCNBH3 was originally chosen because it reduces Schiff’s base much faster than aldehyde at neutral to slightly basic pH [ 111. Although NaBH4 can reduce Schiff’s base as

g

* ou Ho

i

NHAc

Q NH

NHAc

Fig. 2. Reductive amination with oligosaccharides. P denotes protein.

603

NaOMe

ACO

HO

8

OMe

2

EO

Fig. 3. Modification of proteins with glycosides with terminal aldehydic or imidate groups. P denotes protein.

well, it also reduces the aldehyde (reducing sugar), and therefore is not a suitable reagent for this purpose. More recently, pyridine borane and other amino borane complexes [ 121 are becoming more attractive alternatives because of their longer shelf life and ease of handling. The modified amino groups (becoming secondary or tertiary amines) remain positively charged under physiological conditions, so that little perturbation of the tertiary structure of proteins is observed [13]. This is a very important factor to consider when bioactive proteins are to be modified. The reductive amination is still one of the gentlest ways to attach the reducing oligosaccharides directly to proteins or other amino compounds [ 6 ] .It is widely used by biochemists and biologists because of the simplicity of its operation. Preparation of reducing oligosaccharides from N-glycosides of glycoproteins is becoming very practical, thanks to ready availability of endo-P-hexosaminidases and glycoamidases [ 141. Moreover, availability of instrumentation for automated hydrazinolysis makes preparation of reducing oligosaccharides from both N- and 0-glycosides of glycoproteins less hazardous, so that attainment of pure reducing oligosaccharides is now limited only by the efficiency of subsequent purification. An inevitable shortcoming of the reductive amination is that it converts the reducing terminus into an acyclic form. If such a conversion is known a priori to be detrimental to the biological functions of the oligosaccharide, the reductive amination method should be avoided. Another drawback of this method as applied to reducing oligosaccharides is its low reaction rate, often requiring several days. This is because of extremely low concentrations of the acyclic form of reducing sugars which is the reacting species. The problem can be circumvented by the use of glycosides possessing an w-aldehydic group in the aglycon, such as shown in eq. (2) of Fig. 3 [13]. The synthetic process for these sugar derivatives is quite simple for mono- and disaccharides. As expected, reductive amination of a protein with this unmasked aldehyde group proceeded very much faster, the reaction being complete within 24 h. Alternatively, naturally obtained reducing oligosaccharides can be reduced to sugar

604

0 HzNNHA

C HZ)!

1 ) HONO

,OMe CONHCHzCH, OMe

1

2) R'N& (Glycopcptide)

,OMe

0 RNHA(cH*)4/cmHcH2cH\

1

OMe

Mild acid

CONHCH&H=O RNH Bm-Pyridine

jk

RNH

1

NHz-Prolcia

CONHCHzCH~-NH-protein H214/

Fig. 4. Conjugation of glycopeptide to protein using a heterobifunctional reagent.

alcohols (alditols), which are then oxidized by periodate to generate a new aldehydic group. The condition of periodate oxidation must be controlled in such a way (e.g. 1 mM periodate at room temperature for 10 min) so that only vicinal glycols of the reduced sugar (acyclic) are oxidized to generate aldehydic group(s) without affecting vicinal glycols in the sugar rings. The newly generated aldehydic group remains mostly acyclic because the oxidized terminal sugar chain is usually too short for cyclization. This approach is especially useful for the preparation of neoglycoproteins from 0-glycosides of glycoproteins which are usually liberated by (3-elimination concurrent with reduction to yield oligosaccharide alditols. Often a glycopeptide with structurally well-defined oligosaccharide chain can be obtained in pure form easier than the corresponding oligosaccharide. If this is the case, a direct conjugation of such a glycopeptide can alleviate the effort of releasing and further modifying the reducing oligosaccharide for conjugation. Moreover, either the amino or the carboxyl terminus of the glycopeptide can be utilized for conjugation. A heterobifunctional reagent containing an acyl hydrazide at one end for coupling to the amino group of a glycopeptide and an aldehydic group (masked as acetal) on the other end to be coupled to protein by reductive amination was developed to facilitate conjugation of glycopeptides to proteins [15] (Fig. 4). A very high efficiency of conjugation under mild conditions was attained even at submicromol levels.

605

2.1.2. Amidination Amidination aims for attaining the same goal as reductive amination described above, i.e. retention of the positive charge of amino groups to be modified. In its original version, an imidoester was generated from a nitrile-containing thioglycoside by treating it with sodium methoxide (eq. (1) of Fig. 3). The imidoester thus generated reacts with amino groups of protein to form amidino linkages, which are still positively charged under physiological conditions. This method is suitable for mono- or disaccharides, but the preparation of cyanomethyl thioglycosides becomes more cumbersome for larger oligosaccharides. The advantages of the thioglycosidic linkage are: (i) its resistance to glycosidases which allows neoglycoprotein probes to survive in many biological systems; and (ii) specific chemical hydrolysis by mercuric ion which allows easy analysis [ 161. The thioglycosidic linkage was also employed in the earlier example of the o-aldehydic glycosides (Fig. 3). 2.1.3. Acylation Although acylation of amino groups of protein is an easy reaction, it suffers from the change of physical properties (changing a positively charged group into a neutral group) after modification, which may also alter biological properties of the proteins. A carboxyl function can be created on a sugar moiety by oxidizing the reducing end of di- or oligosaccharides with bromine or other mild oxidizing agents to form aldonic acids (or their lactones). This approach has the same disadvantage as the reductive amination in that the reaction converts the reducing terminal sugar residue into an acyclic form. An alternative approach is to transform an oligosaccharide into a glycoside containing a carboxyl group [1,17]. In either case, conjugation of the carboxyl groups to amino groups can be mediated by water-soluble carbodiimide or other coupling agents [ 181. Alternatively, carboxyl groups can also be converted into their active esters [6,18] or into the corresponding azides for coupling to the protein amino groups [17]. The latter two methods avoid cross-linking of proteins, which may accompany the carbodiimide mediated coupling. 2.1.4. p-Zsothiocyanato-phenyl glycosides p-Aminophenyl glycosides can be derived from commercially available p-nitrophenyl glycosides (commonly used as glycosidase substrates), and can be used in the same way as o-aminoalkyl glycosides in conjugation to carboxyl groups of proteins. However, they can also be utilized after conversion into p-isothiocyanato-phenyl glycosides by treatment with thiophosgene. p-Isothiocyanato-phenyl glycosides react quite readily with &-aminogroups [ 191. Alternatively, p-aminophenyl glycosides can be diazotized to react with tyrosyl groups in proteins as will be discussed in section 2.3.

2.2. Modijication of carboxyl groups To conjugate oligosaccharides to carboxyl groups of a protein, the oligosaccharides must be converted into amines. Oligosaccharides can be directly converted to glycosylamines, or made into glycosides containing a terminal amino group in the aglycon. The

606

conjugation of the resulting amines to proteins uses the same procedures as described in section 2.1.3.

2.2.1. ModiJication with glycosylamine derivatives The use of glycosylamines for conjugation to carboxyl groups of protein is partially based on the desire to chemically synthesize “natural” glycopeptides - i.e. sugar residues linked to peptides via P-amide of asparagine (Glyc-Asn) [20,21]. To prepare an Asn-linked oligosaccharide, the original method converts a glycosyl halide into a glycosyl azide, which in turn is hydrogenolyzed to yield a glycosylamine. This is then coupled to p-COOH of a suitably protected aspartic acid derivative to yield Glyc-Asn. When applied to a protein, both Asp and Glu will be amidated. In a simpler method for preparation of glycosylamines [22,23], reducing oligosaccharides are treated with saturated ammonium hydrogencarbonate for several days with daily replenishing of solid ammonium hydrogencarbonate. After removal of excess ammonium hydrogencarbonate, the glycosylamine formed is purified on a column of cation exchange resin. Although glycosylamines are unstable as such in aqueous solution, acylation (i.e. with Asp-derivatives to form Asn-oligosaccharides) will stabilize them.

- &R

R*NH2

OH

PH

OH

Fig. 5. Formation of glycosylamine and its derivatives. R, oligosaccharide.

607

In a recent variation of this method [24], a glycosylamine is first chloroacetylated for stabilization, which is then reacted further. In another recent scheme [25], glycosylamines from an oligosaccharide mixture are tyrosinated so that the products are more easily separated by RP-HPLC, and permit radio-iodination as well (Fig. 5).

2.2.2. Glycamine derivatives A glycamine is formed when a reducing sugar is reductively aminated, and the sugar structure becomes acyclic. This is in contrast to a glycosylamine which retains the cyclic sugar structure. While a primary amino group can be formed by direct reaction of reducing oligosaccharide and ammonia under reductive conditions, our recent scheme [26] of preparing glycamine derivatives is to use benzylamine to form N-benzyl glycamine first, because it gives much higher yields than ammonia or aliphatic amines under comparable conditions. The product is stable for storage, yet unmasking of the amino group is readily accomplished by hydrogenolysis of the benzyl group, and the resulting amine is used for formation of neoglycoproteins. This method allows efficient formation of glycamine even from minute quantities of reducing oligosaccharide. Kallin et al. 123,271 took advantage of the same principle, but used 4-trifluoroacetamido-aniline as the amino component in reductive amination (Fig. 6). The trifluoroacetyl group in the aminated product is easily removed to expose a new aryl amino group, which can then be diazotized to be coupled to tyrosine in a protein (see section 2.3). The efficiency of sugar derivatization with this compound is quite high, though the introduction of an aromatic ring is not desirable for some experiments. As

Fig. 6. Formation of glycamine from oligosaccharide

608

described in section 2.1.4, the anilino group can also be treated with thiophosgene to yield an isothiocyanate group which reacts readily with &-aminogroups [ 191. This method neutralizes the positively charged amino groups in addition to introducing aromatic rings, and thus dramatically increases the hydrophobicity of protein and the propensity for alteration in protein conformation. The advantage of glycamine over glycosylamine is that the former is more stable during the amidation reaction to form neoglycoproteins, typically using water soluble carbodiimide at pH 4-5. Glycamines have also been used to modify poly-carboxylic compounds, such as poly-Asp or poly-Glu [26,28]. 2.2.3. wdminoalkyl glycosides Oligosaccharides can be transformed into glycosides containing a terminal amino group in the aglycon that can be used to react directly with carboxyl groups of proteins. We have prepared 6-aminohexyl glycosides for a number of applications [29,30]. To synthesize o-amino glycosides, several o-amino alkanols are commercially available for reaction with activated glycosyl derivatives. Poly-ethyleneglycol glycosides with terminal amino group have been reported [31]. (See Fig. 7 for structures.)

R A

-0

B

--CHrNH2

C

-CH2CmH

Fig. 7. Polyethyleneglycol glycosides with terminal functional group of: (A) cyclic acetal group [86]; (B) amino group [3 11; (C) carboxyl group [87].

2.3. Modification of tyrosyl group Coupling of diazotized glycosides [ 191 or other carbohydrate derivatives containing an anilino group to proteins is mainly via the tyrosyl group (Fig. 6). The diazo-coupled proteins are usually colored, and the additional hydrophobicity can result in detrimental physical properties such as low solubility. Usually there are not as many tyrosyl groups in a protein compared with lysyl or aspartyl/glutamyl residues. 2.4. Modification of cysteinyl group The number of cysteinyl residues in a protein is usually limited, but it can be used to advantage in a limited and specific attachment of carbohydrate. Some proteins have a

609

Fig. 8. Modification of a protein SH group with carbohydrate derivatives.

readily accessible cysteinyl residue which can be modified. Serum albumins and IgG are examples of this type. In certain cases some disulfide bonds (cystine) can be reduced to provide additional cysteinyl residues without affecting the structural stability or biological activity of the protein. If a protein is being produced by genetic engineering, a cysteinyl residue can be introduced at a desired site in the sequence. For such modifications, the carbohydrate must have a functional group reactive to the SH function. Examples are haloacetamido-, maleimido-, and organomercurial groups (Fig. 8). Most conveniently, these groups should be placed at the terminal position of the aglycon. For example, a 6-aminohexyl glycoside can be readily transformed into a corresponding 6-chloroacetamido derivative. 2.5. Conjugation of polysaccharides to proteins Conjugation of polysaccharides to proteins is a popular method for preparing efficacious vaccines. However, attachment of polysaccharides to proteins provide a set of problems different from those of small, structurally defined oligosaccharides. Typically, polysaccharides used are chemically polydisperse, and the preparation of glycosides from polysaccharides is impractically difficult. Among the modification methods for carbohydrates mentioned above (sections 2.12.4), there are a few which can be applied to polysaccharides for their attachment to a protein. For instance, the method of selectively oxidizing the reducing terminal sugar with periodate to generate an aldehydic function has been applied successfully to polysaccharides, such as hyaluronic acid. The reduced and selectively oxidized hyaluronic acid was first conjugated to 1,6-diaminohexane by reductive amination, and the resulting conjugate was then attached to proteins [32]. The advantage of this method is that the point of attachment is always at the originally “reducing” terminus, and there is only one site of attachment per chain of polysaccharide. However, since this method relies on selective periodate oxidation of acyclic polyols as compared to uic. glycols in cyclic sugars, the oxidation conditions must be stringently controlled. The oxidation obviously cannot be selective for the reduced terminus if the remainder of the reduced polysaccharide contains exocyclic glycols, such as in a-2,9-linked polysialic acids.

610

Incorporating aldehydic functions in a polysaccharide can be accomplished by reacting the polysaccharide with chloroacetaldehyde dimethyl acetal. After unmasking of the acetal, the modified polysaccharide is reacted with proteins by reductive amination [33]. Since the initial reaction of chloroacetaldehyde dimethyl acetal with polysaccharide lacks site-specificity, this method suffers from randomness as well as multiplicity of the attachment site on the polysaccharide. The selective periodate oxidation can be actually used to fragment a polysaccharide as well as to provide aldehydic groups. A mild periodate oxidation of a-2,9-linked polysialic acids can result in fragmentation yielding dialdehydic fragments. Similarly, 1,5-phosphorylated ribitol in the Haemophilus injluenzae polysaccharide can be oxidized with periodate to provide dialdehydic products. These dialdehydic products actually can be used to cross-link between modified proteins. Free or potentially free amino groups exist in certain polysaccharides (e.g. GlcN in heparin and chitosan). Treatment of such polysaccharides with nitrous acid causes GlcN to be deaminated to yield 2,5-anhydro-~-mannoseby ring contraction with inversion accompanied by cleavage of the affected glycosidic bonds. Poly- or oligosaccharides thus produced contain an active aldehydic group at the reducing terminus, which can be used for conjugation to proteins or amino compounds by reductive amination. This was the approach used to prepare low-mass heparin derivatives possessing anticoagulant activities [34]. Usually smaller oligosaccharides have higher reactivity than their higher analogs with respect to the reducing terminus. It is thus advantageous to fragment a polysaccharide if the reducing terminus is to be used directly for conjugation and the polysaccharide in its entirety is not needed for biological activity. Partial hydrolysis has been effectively used to fragment polysaccharides from certain microorganisms, and the resulting fragments used for vaccine preparation by reductive amination with appropriate protein carriers [35].

2.6. Enzymatic methods Although there are only few enzymatic methods available to directly attach carbohydrate moieties including glycopeptides to proteins, enzymatic methods are very powerful in constructing oligosaccharides to be used for chemical modification of proteins or for modification of existing carbohydrate units in (neo)glycoproteins. The strength of enzymes in constructing oligosaccharides is their exquisite specificity. For example, when a glycosyl-transferase is used for glycosylation, in addition to the structure of donor sugar, the nature of acceptor sugar, the position of attachment site, and the anomeric configuration are quite specific. There are also enzymes which allow transfer of the whole preexisting glycan moieties. In this chapter, the topics of construction of oligosaccharides per se will not be presented. Instead enzymatic methods for modification of proteins or glycoproteins to create new oligosaccharide groups will be discussed.

2.6. I. Use of glycosyl-transferases Glycosyl transferases are quite powerful for stepwise addition of glycosyl units to existing (neo)glycoproteins. Paulson et al. [36] used a-2,3- and a-2,6-specific sialyl transferases to resialylate red blood cells that had been extensively desialylated enzymatically. Since

61 1

\

FUC-Tf-1 Gav1-4GkNAcp 3

I

Fucal

Fuc-Tf-2

Fuca1-2Galf3 1-4GlcNAcp

1

Fig. 9. Modification of neoglycoproteins with glycosyl-transferases. Different Fuc-transferases add L-FUCon different sugars via different linkages.

these remodeled cell surface glycoconjugates contain only a single specific type of sialyl linkage, this enabled them to probe critically the specificities of viral binding. In another example of remodeling of glycoconjugates, extant Gal residues on hen ovalbumin were replaced with I3C-labeled Gal by de-galactosylation followed by attachment of I3C-labeled Gal to the exposed GlcNAc by the action of a galactosyl-transferase [37391. Such derivatives allowed them to study lectin (Erythrinu crystugulli agglutinin and soy bean agglutinin) binding by I3C-NMR. Hill and coworkers modified GlcNAc-BSA [40] by attaching Gal and L-FUCsequentially with the aid of a galactosyl-transferase and one of two different L-Fuc-transferases. The different L-Fuc-containing neoglycoproteins were used to study L-Fuc-binding proteins [4 1,421 (Fig. 9). Similarly, a successive galactosylation and sialylation by glycosyl transferases has also been applied to polyacrylamide polymers containing GlcNAc pendants [43]. Other groups have also employed this type of chemo-enzymatic approaches [44,45]. Modification of the sugar residues on neoglycoproteins or natural glycoproteins needs not be with natural sugar residues. Many glycosyl-transferases can be enticed to use sugar nucleotides containing unnatural sugars as donor substrates. In a notable example, an L-Fuc-transferase was used to transfer a derivative of fucose having the blood group B trisaccharide structure linked via a spacer arm [46] (Fig. 10). A change in serological properties of the cells resulted from the modification with such a fucosederivative (“sneaky-B”). N-Acetylneuraminic acid (Neu5Ac) modified with a fluorescent or photoactivatable group at C-9 can be activated into the corresponding CMP-derivative (with CMP-Neu5Ac synthase) and transferred onto appropriate oligosaccharide acceptors by several of the sialyl-transferases [47].

612

L-FUC

Fuc-Transfercase

Fig. 10. Transfer of “sneaky-B” (a GDP-derivative of fucose carrying “B-trisaccharide”) by fucosyl transferase

+ AcNH

5, Trans-sialylase

COOH ACN

R I 4-MdJrnbelllleron0, p-NOz-phenol, G8I. etc.

Fig. 1 1. Transfer of NeuSAc and derivatized NeuSAc by Trypanosoma cruzi trans-sialylation

2.6.2. Use of glycosidases in transglycosylation Glycohydrolases can often perform transglycosylation [48], especially when the acceptor concentration is high. The main advantages of using glycosidases in comparison to using glycosyl-transferases are the lower costs for enzymes as well as substrates (oligosaccharides or simple glycosides rather than sugar nucleotides are used as donors). Glycosidase-catalyzed oligosaccharide formations are usually specific with respect to the anomeric configuration but lack stringent positional specificity. A good example is the trans-sialylase from Trypanosoma cruzi, which is capable of transferring an a-2,3-linked NeuSAc residue to galactosyl residues (Fig. 1 I). This enzyme is interesting in that the natural donor substrate and the product are both sialyl groups a-2,3-linked to a Gal residue. Its ability to use a ketoside of NeuSAc rather than CMPNeuSAc (which is more laborious to prepare) as donor gives a significant advantage. Interestingly, common chromogenic substrates for sialidase, such as 4-methylumbelliferyl

613

orp-nitrophenyl ketoside of Neu5Ac (4-MU-Neu5Ac and PNP-NeuSAc, respectively) can also serve as glycosyl donors. We have shown that 4-MU-Neu5Ac derivatives modified at the exocyclic side chain can serve as donor substrates as well [49] (Fig. 11). Some endo-p-hexosaminidases show a great promise in the en bloc transfer of oligosaccharides rather than one single monosaccharide at a time done by glycosyltransferases. For example, endo-A from Arthrobacter [50,5 I] and endo-M from MuCOY [52,53] can transfer high-mannose type and complex type oligosaccharides, respectively. The acceptor can be a free mono- or oligosaccharide or a peptide containing terminal GlcNAc residue. The yield of transglycosylation catalyzed by endo-A can be raised to more than 90% by inclusion of some organic solvent in the reaction medium [50].

2.6.3. Use of transglutaminase Although most natural N-glycosides contain the GlcNAc-Asn linkage, Yan and Wold [54] developed an ingenious method of preparing neoglycoproteins containing analogous sugar-Gln linkage by attaching glycopeptides or other amino-terminated glycosides to Gln by the action of transglutaminase. For example, the y-CONH2 groups of Gln in p-casein can be modified with the a-amino group of glycan-Asn by this enzyme. Of course, all the amino groups in f3-casein must be masked first to prevent cross-linking of the proteins, but this is easily done by a transient masking of amino groups. In the above example, conjugation of four Gln sites with glycan-Asn was achieved. 2.7. Glycoproteins of non-cooalent attachment

The interaction between biotin and avididstreptavidin is so strong that it can almost be considered as a covalent bond. Wold et al. [55] allowed biotinylated glycopeptides or 6-aminoalkyl glycosides to be bound to avidin or streptavidin. Although avidin or streptavidin has four biotin-binding sites, only three biotinylated glycosides or glycopeptides can be bound readily, the fourth insertion being quite difficult. This type of neoglycoproteins has an advantage of the site of attachment being geometrically well defined, although the number of oligosaccharide chains attached is limited [56]. An oligosaccharide mixture derivatized with 2,6-diaminopyridine-modifiedbiotin can be efficiently separated by reverse-phase HPLC [57] and the separated oligosaccharide derivatives can be used for construction of neoglycoproteins via the biotin-avidin interaction.

3. Synthetic glycopeptides Glycopeptides are more difficult to synthesize than the conventional peptides, because common protection-deprotection reactions used in the peptide synthesis can cause serious problems to the protective groups used for carbohydrates. Many of these problems have been solved by recent technical innovations [2 11. The combination of 9-fluorenylmethyloxycarbonyl (Fmoc) group for N-protection and pentafluorophenyl (PFP) group as the activating group for the carboxylic acid allows GlcNAc-Asn

614

or GalNAc-Ser/Thr as a building block to be incorporated into the peptide sequence in the solid-phase synthetic scheme. A sugar chain is then built onto the monosaccharide stub by the aid of enzymes. The endo-A and endo-M enzymes mentioned earlier offer promise in this area. Synthetic glycopeptides can be used directly in conjugation with proteins via either amino terminus or carboxyl terminus. The transglutaminase method (section 2.6.4) can also be used for the construction of neoglycoproteins.

4. Applications of neoglycoproteins Neoglycoproteins are used in many different areas of glycobiology. Only a few important examples will be presented in this chapter.

4. I . Probing carbohydrate-protein interactions In the first application of neoglycoproteins, bovine serum albumin (BSA) modified with different monosaccharide derivatives was used to demonstrate that the carbohydrate receptor of mammalian liver cells bind Gal/GalNAc-BSA specifically [ 10,401. Similarly, chicken hepatocytes were found to specifically bind GlcNAc-BSA [ 5 8 ] and mammalian lung macrophages were found to bind Man-BSA and L-FUC-BSA[59-61]. Thus, neoglycoproteins proved to be extremely valuable in determination of sugar-binding specificities of animal lectins/receptors, and in the cell biological studies of their receptor functions. Beyond deciphering sugar-binding specificities, these experiments also clearly showed two important aspects of these lectins: (i) only the terminal sugar residue is recognized; and (ii) near logarithmic increase in the binding potency is brought about by a linear increase in the sugar density. The latter phenomenon (glycoside cluster effect) turns out to be a very important aspect of many carbohydrate-protein interactions. 4.2. Use in isolation of carbohydrate-binding proteins In contrast to naturally derived oligosaccharides of a single specific structure or a unique glycoform of a glycoprotein, neoglycoproteins of mono- and disaccharides can be easily prepared in relatively large quantities. An attractive application of neoglycoproteins is for isolation of carbohydrate-binding proteins. For example, the hepatic lectins mentioned above as well as that from alligator liver [62] were isolated by affinity chromatography using Sepharose bearing appropriate neoglycoproteins. 4.3. Cytochemical markers The presence of carbohydrate-binding proteins in tissue sections can be readily detected with neoglycoproteins marked with fluorescent probes. Neoglycoproteins, in comparison to simple oligosaccharides, have a high survival rate during cytochemical procedures because of their large molecular size and stronger affinity resulting from the “glycoside

615

cluster effect” (section 4.1). The use of neoglycoenzymes (see below) in similar applications can further increase the detection sensitivity. Changes in carbohydratebinding activities occurring during malignancy have been probed and documented by such methods [63]. The use of neoglycoproteins also greatly benefits flow cytometric assay [64]. Here again, the possibility of fluorescent labeling and potential for strong binding affinity by neoglycoproteins are important virtues. 4.4. Neoglycoenzymes

Enzymes modified with carbohydrates (neoglycoenzymes) can be used in cytochemistry as described above or in biochemical detection of lectins in solid-phase assays to gain greater sensitivity in analysis. For example, bacterial fi-galactosidase modified with paminophenyl a-D-mannopyranoside via amide linkage was useful in determination of Con A immobilized on plastic microtiter plates, and lactose-modified fi-galactosidase was effective in histochemical detection of galactoside-specific lectins [63]. Other enzymes frequently used for these applications are alkaline phosphatase and horse radish peroxidase. There are a number of colorimetric, fluorometric, and chemiluminescent substrates available for these enzymes. 4.5. Biomedical applications

Neoglycoproteins have been valuable in tumor diagnosis [65]. Gal-modified human serum albumin, further modified with technetium, has been useful for the monitoring of liver receptor function, which can be correlated to the disease status [66]. A more intriguing application is to achieve targeted delivery of DNA via a carbohydrate-mediated entry into specific cells [67]. A high-affinity neoglycopeptide having three terminal GalNAc residues has been used for the purpose of delivering DNA [68] as well as anti-sense methyl phosphonate nucleotide analogs to the liver [69]. Carbohydrate-mediated targeting of drugs in general has been reviewed recently [70]. The outer coat of bacteria often contains a large amount of polysaccharides. Many vaccines against microorganisms utilize the antigenicity of these polysaccharides [35]. Although polysaccharides alone can be used as vaccines, conjugation to suitable proteins greatly enhances their immunogenicity. Conjugation to proteins is especially important in vaccines for infants, since oligo- and polysaccharides alone are poor immunogens in general, and especially for infants.

5. Advantages of neoglycoproteins Neoglycoproteins were deviced to overcome complications arising from heterogeneity of oligosaccharides in natural glycoproteins. Since the carbohydrate moieties in neoglycoproteins are structurally well-defined, they can provide unambiguous answers in the studies of carbohydrate-protein interaction.

616

Bioactive oligosaccharides in glycoconjugates are extremely difficult to isolate, and obtain in a significant quantity in pure form. Fortunately, it has become increasingly clear that only a short segment, often only the terminal sugar, of bioactive glycoconjugates is involved in biological interactions. Thus the total synthesis of the oligosaccharide structures is usually unnecessary for the construction of effective neoglycoconjugates. Good examples for the last case are the hepatic carbohydrate-binding proteins mentioned above which can be isolated in pure form by simple affinity chromatography using Sepharose modified with appropriate neoglycoprotein containing only a specific monosaccharide derivative (see ref. [62]). Binding of a carbohydrate ligand involves both hydrogen bonding and hydrophobic interactions. Usually not all the functional groups of a carbohydrate ligand are needed for binding. Neoglycoproteins containing an analog of natural ligand with a specific structural alteration (e.g. removal of a single hydroxyl group) can be valuable in determining the binding mode of the receptor/lectin. In the case of mammalian hepatic lectins, BSA containing 6-0-methylated Gal residues was found to be bound as well as the Gal counterparts, thus establishing the non-essentiality of the 6-OH group of Gal in binding [71]. This information was invaluable in the later design of affinity labeling reagents for these receptors [72,73] as well as for the attachment of fluorescent probes for conformational analysis of a triantennary glycopeptide [74]. A very important aspect of using neoglycoproteins is to take advantage of the strong glycoside cluster effects they can produce. One of the hallmarks of neoglycoproteins is the ease with which a large number of sugar residues can be attached. For example, BSA can easily be modified with up to 40 sugar residues per molecule. For many carbohydrateanimal lectin interactions, the affinity increases exponentially up to the sugar density of -20 mol per mol of BSA. Thus highly sugar-substituted neoglycoproteins often exhibit affinities as strong as or even stronger than the best natural ligands [58,75]. In one of the recent examples, sialyl LeX attached to BSA showed much enhanced affinity toward L-selectin compared to the monovalent sialyl LeXderivative [76]. Actually the glycoside cluster effect does not depend solely on the sugar density. Unlike peptides or nucleotides, carbohydrates, especially those of glycoconjugates have branched structures, and the recognition of such carbohydrates often involves recognition of multiple terminal sugars as well as branching patterns. In the case of hepatic lectin on rat hepatocytes, a Gal-terminated triantennary oligosaccharide of N-glycoside appears to possess structure or conformation complementary to the lectin-binding sites, in that its affinity is approximately the geometrical product of the monovalent ligand (i.e. the maximal glycoside cluster effect) [74,77,78]. Most other carbohydrate-animal lectin interactions seem to give lesser degrees of glycoside cluster effects. The basis for such glycoside cluster effects is discussed in detail in a recent review [49]. In some instances, neoglycoconjugates are prepared for the purpose of improving physical properties. If the hydrophobic nature of bioactive glycolipids is undesirable, they can be transformed into neoglycoproteins to provide greater aqueous solubility [79,80]. This approach was instrumental in the discovery of ganglioside receptors in the central nervous system [8 11. Similarly, neoglycoproteins containing active oligosaccharide fragments (antigenic determinants) of a glycolipid from Mycobacteriurn leprae were extremely valuable for diagnosis of the diseases caused by Mycobacteria [82,83].

617

cw-

Fig. 12. Transglycosylation with ceramide glycanase from American leech

To transform glycolipids to neoglycoprotein, a ceramide glycanase from the American leech is very effective. This glycohydrolase possesses a high transglycosylation activity (Fig. 12), so that the entire oligosaccharide structure of a glycosphingolipid can be transferred to a potential linker compound, which then can be used for conjugation to proteins [84,85].

6. Other neoglycoconjugates In addition to neoglycoproteins, there are now neoglycolipids, neoproteoglycans, etc.,

618

which are called collectively “neoglycoconjugates” [2]. The advantages of other neoglycoconjugates are similar to those of neoglycoproteins. Details on neoglycoconjugates other than neoglycoproteins are covered in a number of recent review articles [5,7-91.

7. Conclusions It should be clear from the foregoing presentation that neoglycoproteins and neoglycoconjugates in general have a wide range of applicability. As the synthetic methodology advances, more sophisticated neoglycoconjugates will be prepared and used. In such an endeavor, chemo-enzymatic methods are predicted to be instrumental. Progress in molecular biology of glycosyl-transferases accelerates the usage of this class of enzymes in the construction of neoglycoconjugates.

Acknowledgment The support of National Institutes of Health Research Grants, DK09970 and GM4994 is gratefully acknowledged.

References 111 Stowell, C.P. and Lee, Y.C.

(1980) Adv. Carbohydr. Chem. Biochem. 37, 225-281. [2] Magnusson, G. (1 986) In: D.A. Lark (Ed.), Protein-Carbohydrate Interactions in Biological Systems. Academic Press, London, pp. 215-228. [3] Aplin, J.D. and Wriston, J.C. (1981) CRC Crit. Rev. Biochem. 10, 259-306. 141 Lee, Y.C. and Lee, R.T. (1982) In: M.I. Horowitz (Ed.), The Glycoconjugates, Vol. IV Academic Press, New York, pp. 57-83. [ 5 ] Lee, Y.C. and Lee, R.T. (1991) In: H.-J. Gabius and S. Gabius (Eds.), Lectins and Cancer. Springer, Berlin, pp. 53-70. [6] Lee, Y.C. and Lee, R.T. (1992) In: H.J. Allen and E.C. Kisailus (Eds.), Glycoconjugates. Marcel Dekker, Inc., New York, pp. 121-165. [7] Lee, Y.C. and Lee, R.T. (1994) Neoglycoconjugates: Preparation and Applications. Academic Press, San Diego, CA, p. 549. [8] Lee, Y.C. and Lee, R.T. (1994) In: J.N. Abelson and M.I. Simon (Eds.), Methods in Enzymology, Vol. 242. Academic Press, San Diego, CA, p. 450. [9] Lee, Y.C. and Lee, R.T. (1994) In: J.N. Abelson and M.I. Simon (Eds.), Methods in Enzymology, Vol. 247. Academic Press, San Diego, CA, p. 328. [lo] Krantz, M.J., Holtzman, N.A., Stowell, C.P. and Lee, Y.C. (1976) Biochemistry 15, 39633-3968. [I I] Gray, G.R. (1974) Arch. Biochem. Biophys. 163, 426428. [12] Cabacungan, J.C., Ahmed, A.I. and Feeney, R.E. (1982) Anal. Biochem. 124, 272-278. [I31 Lee, R.T. and Lee, Y.C. (1980) Biochemistry 19, 156-163. [14] Takahashi, N. and Muramatsu, T. (1992) Handbook of Endoglycosidases and Glycoamidases. CRC Press, Boca Raton, FL, p. 362. [15] Lee, R.T., Wong, T.-C., Lee, R., Yue, L. and Lee, Y.C. (1989) Biochemistry 28, 1856-1861. [16] Krantz, M.J. and Lee, Y.C. (1976) Anal. Biochem. 71, 318-321. [I71 Lemieux, R.U., Bundle, D.R. and Baker, D.A. (1975) J. Am. Chem. SOC.97, 40764083. [I81 Andersson, M. and Oscarson, S. (1993) Bioconjugate Chem. 4, 246-249.

619 [I91 McBroom, C.R., Samanen, C.H. and Goldstein, I.J. (1972) Methods Enzymol. 28, 212-219. [20] Garg, H.G. and Jeanloz, R.W. (1985) Adv. Carbohydr. Chem. Biochem. 43, 135-201. [21] Meldal, M. (1994) In: Y.C. Lee and R.T., Lee (Eds.), Neoglycoconjugates: Preparation and Applications. Academic Press, San Diego, CA, pp. 144-196. [22] Likhosherstov, L.M., Piskarev, VE,, Derevitsakaya, VA. and Kochetkov, N.K. (1990) Bioorg. Chem. 16, 124-1 26. [23] Kallin, E., Loenn, H. and Norberg, T. (1986) Glycoconj. J. 3, 311-319. [24] Manger, I.D., Wong, S.Y.C., Rademacher, T.W. and Dwek, R.A. (1992) Biochemistry 31, 10733-10740. [25] Tamura, T., Wadhwa, M.S. and Rice, K.G. (1994) Anal. Biochem. 216, 335-344. [26] Yoshida, T. and Lee, Y.C. (1994) Carbohydr. Res. 251, 249-254. [27] Kallin, E. (1994) In: Y.C. Lee and R.T. Lee (Eds.), Neoglycoconjugates: Preparation and Applications. Academic Press, San Diego, CA, pp. 199-223. [28] Yoshida, T., Lasky, L. and Lee, Y.C. (1992) Glycobiology 2, 489489. [29] Chipowsky, S. and Lee, Y.C. (1973) Carbohydr. Res. 31, 339-346. [30] Weigel, P.H., Naoi, M., Roseman, S. and Lee, Y.C. (1979) Carbohydr. Res. 70, 83-91. [31] Bertozzi, C.R. and Bednarski, M.D. (1991) J. Org. Chem. 56, 43264329. [32] Raja, R.H., LeBoeuf, R.D., Stone, G.W. and Weigel, P.H. (1984) Anal. Biochem. 139, 168-177. [33] Bogwald, J., Seljelid, R. and Hoffman, J. (1986) Carbohydr. Res. 148, 101-107. [34] Malsch, R., Guerrini, M., Torn, G., Loehr, G., Casu, B. and Harenberg, J. (1994) Anal. Biochem. 217, 255-264. [35] Jennings, H.J. and Sood, R.K. (1994) In: Y.C. Lee and R.T. Lee (Eds.), Neoglycoconjugates: Preparation and Applications. Academic Press, San Diego, CA, pp. 324-371. [36] Paulson, J.C., Rogers, G.N., Carroll, S.M., Higa, H.H., Pritchett, T., Milks, G. and Sabesan, S. (1984) Pure Appl. Chem. 56, 797-805. [37] Berman, E., Brown, J., Lis, H. and Sharon, N. (1985) Eur. J. Biochem. 152, 447451. [38] Berman, E., Lis, H. and James, T.L. (1986) Eur. J. Biochem. 161, 589-594. [39] Berman, E. and Lis, H. (1987) Biochim. Biophys. Acta 924, 403407. [40] Lee, Y.C., Stowell, C.P. and Krantz, M.J. (1976) Biochemistry 15, 3956-3963. [41] Lehrman, M.A., Haltiwanger, R.S. and Hill, R.L. (1986) J. Biol. Chem. 261, 7426-7432. [42] Lehrman, R., Pizzo, S.V., Imber, M.J. and Hill, R.L. (1986) J. Biol. Chem. 261, 7412-7418. [43] Nishimura, S.-I., Lee, K.B., Matsuoka, K. and Lee, Y.C. (1994) Biochem. Biophys. Res. Commun. 199, 249-252. [44] Sabesan, S. and Paulson, J.C. (1986) Biochemistry 108, 2068-2080. [45] Sabesan, S., Duus, J., Neirs, S., Domaille, F'., Kelm, S., Paulson, J.C. and Bock, K. (1992) J. Am. Chem. Soc. 114, 8363-8375. [46] Srivastava, G., Kaur, K.J., Hindsgaul, 0. and Palcic, M.M. (1992) J. Biol. Chem. 267, 22356-22361. [47] Gross, H.J., Rose, U., Karuse, J.M., Paulson, J.C., Schmid, K., Feeney, R.E. and Brossmer, R. (1989) Biochemistry 28, 7386-7392. [48] Ichikawa, Y., Look, G.C. and Wong, C.-H. (1992) Anal. Biochem. 202, 215-238. [49] Lee, K.B. and Lee, Y.C. (1994) Anal. Biochem. 216, 358-364. [50] Fan, J.Q., Takegawa, K., Iwahara, S., Kondo, A,, Kato, I.,Abeygunawardana, C. and Lee, Y.C. (1995) J. Biol. Chem. 270, 17723-17729. [51] Takegawa, K., Tabuchi, M., Yamaguchi, S., Kondo, A., Kato, I. and Iwahara, S. (1995) J. Biol. Chem. 270, 3094-3099. [52] Yamamoto, K., Watanabe, J., Kadowaki, S. and Kumagai, H. (1993) Seikagaku 65, 1014. [53] Kadowaki, S., Yamamoto, K., Fujisaki, M. and Tochikura, T. (1991) J. Biochem. 110, 17-21. I541 Yan, S.-C.B. and Wold, F. (1984) Biochemistry 23, 3759-3765. [55] Chen, VJ. and Wold, E (1984) Biochemistry 23, 3306-331 I. [56] Shao, M.-C., Sokolik, C.W. and Wold, F. (1994) In: Y.C. Lee and R.T. Lee (Eds.), Neoglycoconjugates: Preparation and Applications. Academic Press, San Diego, CA, pp. 224-249. [57] Rothenberg, B.E., Hayes, B.K., Toomre, D., Manzi, A.E. and Varki, A. (1993) Proc. Natl. Acad. Sci. USA 90, 11939-1 1943. [58] Lee, Y.C. (1992) FASEB J. 6, 3193-3200.

620 [59] Stahl, P., Schlesinger, P.H., Sigardson, E., Rodman, J. and Lee, Y.C. (1980) Cell 19, 207-215. [60] Shepherd, VL., Lee, Y.C., Schlesinger, P.H. and Stahl, P.D. (1981) Proc. Natl. Acad. Sci. USA 78, 10191022. [61] Stahl, F. and Gordon, S. (1982) J. Cell Biol. 93, 49-56. [62] Lee, R.T., Yang, G.C., Kiang, J., Bingham, J.B., Golgher, D. and Lee, Y.C. (1994) J. Biol. Chem. 269, I961 7-1 9625. [63] Gabius, S., Hellmann, K.-P., Hellmann, T., Brinck, U. and Gabius, H.-J. (1989) Anal. Biochem. 182, 4 4 7 4 5 1. [64] Midoux, P., Roche, A.-C. and Monsigny, M. (1987) Cytometry 8, 327-334. [65] Gabius, H.-J., Brinck, U., Kayser, K., Schauer, A,, Stiller, D. and Gabius, S. (1994) In: Y.C. Lee and R.T. Lee (Eds.), Neoglycoconjugates: Preparation and Applications. Academic Press, San Diego, CA, pp. 404424. [66] Kudo, M., Vera, D.R. and Stadalnik, R.C. (1994) In: Y.C. Lee and R.T. Lee (Eds.), Neoglycoconjugates: Preparation and Applications. Academic Press, San Diego, CA, pp. 373402. [67] Lee-Young, A.W., Wu, G.Y. and Wu, C.H. (1994) In: Y.C. Lee and R.T. Lee (Eds.), Neoglycoconjugates: Preparation and Applications. Academic Press, San Diego, CA, pp. 51 1-537. [68] Menvin, J.R., Noell, G.S., Thomas, W.L., Chiou, H.C., DeRome, M.E., McKee, T.D., Spitalny, G.L. and Findeis, M.A. (1994) Bioconjugate Chem. 5, 612-620. [69] Hangeland, J.J., Levis, J.T., Lee, Y.C. and Ts’o, P.O.P. (1995) Bioconjugate J. 6, 695-701. [70] Ouchi, T. and Ohya, Y. (1994) In: Y.C. Lee and R.T. Lee (Eds.), Neoglycoconjugates: Preparation and Applications. Academic Press, San Diego, CA, pp. 4 6 4 4 9 8 . [71] Lee, R.T., Myers, R.W. and Lee, Y.C. (1982) Biochemistry 24, 6292-6293. [72] Lee, R.T. and Lee, Y.C. (1986) Biochemistry 25, 6835-6841. [73] Rice, K.G. and Lee, Y.C. (1990) J. Biol. Chem. 265, 18423-18428. [74] Rice, K.G. and Lee, Y.C. (1993) Adv. Enzymol. Related Areas Mol. Biol. 66, 41-83. [75] Lee, Y.C. and Lee, R.T. (1995) Acc. Chem. Res. 28, 321-327. [76] Welply, J.K., Abbas, S.Z., Scudder, P., Keene, J.L., Broschat, K., Casnocha, S., Gorka, C., Steininger, C., Howard, S.C., Schmuke, J.J., Graneto, M., Rotsaert, J.M., Manger, I.D. and Jacobs, G.S. (1994) Glycobiology 4, 259-265. [77] Lee, Y.C., Townsend, R.R., Hardy, M.R., Lonngren, J., Arnarp, J., Haraldsson, M. and Lonn, H. (1983) J. Biol. Chem. 258, 199-202. [78] Lee, Y.C., Townsend, R.R., Hardy, M.R., Lonngren, J. and Bock, K. (1984) In: T.F. Lo, T.Y. Liu and C.H. Li (Eds.), Biochemical and Biophysical Studies of Proteins and Nucleic Acids. Elsevier, New York, pp. 349-360. [79] Chatterjee, D., Douglas, J.T., Cho, S.-N., Rea, T.H., Gelber, R.H., Aspinall, G.O. and Brennan, P.J. (1985) Glycoconjugate J. 2, 1877208. [80] Tiemeyer, M., Yasuda, Y. and Schnaar, R.L. (1989) J. Biol. Chem. 264, 1671-1681. [81] Tiemeyer, M., Swank-Hill, P. and Schnaar, R.L. (1990) J. Biol. Chem. 265, 11990-1 1999. [82] Gaylord, H. and Brennan, P.J. (1987) Ann. Rev. Microbiol. 41, 645-675. [83] Aspinall, G.O., Chatterjee, D. and Brennan, P.J. (1995) Adv. Carbohydr. Chem. Biochem. 51, 169-242. [84] Li, Y.-T., Carter, B.Z., Rao, B.N., Schweingruber, H. and Li, S.-C. (1991) J. Biol. Chem. 266, 1072310726. [85] Li, Y.-T. and Li, S.-C. (1994) In: Y.C. Lee and R.T. Lee (Eds.), Neoglycoconjugates: Preparation and Applications. Academic Press, San Diego, CA, pp. 250-260. [86] Verez-Bencomo, V., Campos-Valdes, M., Marino-Albernas, J.R., Fernandez-Santana, V, HernandezRensoli, M. and Perez-Martinez, C.S. (1991) Carbohydr. Res. 271, 263-267. [87] Anderson, M., Oscarson, S. and Oberg, L. (1993) Glycoconjugate J. 10, 197-201.

Subject Index AB secretors 486 Abrus precatorius 405,428,458,468 Acanthamoeba castelloni 94 Acer pseudoplatanus 138 5-Acetamido-2,6-anhydro-3,5-dideoxy-~-glyceroD-galacto-non-2-enonic acid 250

p-I ,2N-Acetylglucosaminyltransferase I p-I ,2N-Acetylglucosaminyltransferase I1

138 138

N-Acetyllactosamine 410,414,421,493 N-Acetyl-D-mannosamine 274 Acetyl migration 328 0-Acetyl migration 249 5-Acetamido-3,5-d~deoxy-~-glycero-a-~-galacto- N-Acetylneuraminic acid (Neu5Ac) 244,249, non-2-ulopyranosonic acid 248 409,417,476478,494,61 1 9-0-Acetyl-N-acetylneuraminicacid 478 N-Acetylneuraminyllactose 485 5-N-Acetyl-2,7-anhydro-neuraminicacid 250 0-Acetyltransferases 326 biological origin 328 Escherichia coli 328 Macrobdella decora 329 melanoma 328 sialidase L 329 al -Acid glycoprotein 174-176 5-N-Acetyl-4,8-anhydro-neuraminicacid 250, a-Actinin 526 25 1 Actinomyces 476,487,488,500 0-Acetylated sialic acids Actinomyces israelii 488 biosynthesis 324-328 Actinomyces naeslundii 487,488 corona virus 325 Actinomyces uiscosus 487 encephalomyelitis virus 325 Activin 544 functions 324-328 Acute inflammation 175 influenza A virus 325 Acute-phase reaction I74 influenza B virus 325 5-N-Acyl-2-deoxy-2,3-didehydro-neuraminic acid influenza C virus 325 250 occurrence 324 Acylation 605 Plasmodium falciparum 325 aldonic acids 605 Tachyglossus aculeatus 324 reductive amination 605 9-0-Acetylated sialic acids, recognition by 263 water-soluble carbodiimide 605 bovine corona virus 263 Acylneuraminate-pyruvate lyase 325,342 encephalomyelitis virus 263 Adenia digifata 405,458,470 influenza C virus 263 Adenoma-carcinoma sequence, colorectum 326 Acetyl-CoA, translocation 326 S-Adenosylmethionine 94 5-N-Acetyl-2-deoxy-2,3-didehydro-neuraminic acid S-Adenosyl-L-methionine 328 250 S-Adenosyl-L-methionine:sialatebiological origin 328 8-0-methyltransferase 328 N-Acetylgalactosamine 406,409,444,458,464, Adenovirus, fiber proteins 42 487,488,493 Adherence junctions 507,514 N-Acetylgalactosamine-bindinglectins 419422 Adhesion 475,482,493,495,500 N-Acetylglucosamine 74,405,409,415,457, Actinomyces uiscosus 358 464,493,494 Racteroides intermedius 358 N-Acetyl-o-glucosamine 413,415 Pseudomonas aeruginosa 358 N-Acetyl-D-glucosamine-binding lectins 416sialic-acid-specific 362

419

CD22 367 CD33 367 colostrum 362

N-Acetylglucosamine-1 -phosphodiester a-Nacetylglucosaminidase 436,437 N-Acetyl-P-glucosaminidase 501 N-AcetylglucosaminyI phosphotransferase 5 13 N-Acetylglucosaminyl transferan 93 N-Acetylglucosaminyltransferase 75, 78,79

Escherichia coli 362 Helicobacter pylori 362 myelin-associated glycoprotein (MAG) tyrosine kinase 368

621

367

622 Adhesion - sialic-acid-specific (cont 8 ) sialoadhesin 367 sialylated mucins, milk 362 Tritrichomonas foetus 363 Tritrichomonas mobiliensis 363 Tritrichomonas suis 363 Adhesion molecules, leukocyte-to-endothelium

572,581 Adhesive receptors 173 Adjuvant 578 ADP-nbosyltransferase 489 Aegopodium podagraria 406,428,458,464,470 Affinity adsorption 435 Affinity adsorption/chromatography 427,431-

434 56,404,413,419,420, 423,424,436,476,491 Affinity electrophoresis 435,436 Affinity precipitation 423 Agaricus bisporus 428,458,467 Agarose 405 Agglutination 484,488 Agglutination analysis 430,438441 Agglutinins 403470 LimaxJlavus (LFA) 360 Maackia amurensis (MAA) 360 Ricinus communis (RCA) 539 Sambucus nigra (SNA) 360 wheat germ (WGA) 360 Aggrecan 7,16,17 aggregate formation with hyaluronan 16 Aggregation 488 Agonist receptors 173 AGP in ascitic fluid 174 Agrobacterium tumefaciens 489 Affinity chromatography

Albumen gland 140 Aldolase 273,274,284,290-292,347 Aleuria aurantia 405,423,428,461,465,467 Alkaline borohydride treatment 56 Alkylation 426 Allergic responses 579 Allium 467 Allium ascalonicum 405,415,416,456 Allium saiivum 415,416,456 Allium ursinum 415,416,456 Allogeneic transplantation 441 Alocasia indica 404,406,462,468 a-1,3-bound fucose 56 Alveolar epithelial cells 177 Alveolar macrophages 465,466 Alzheimer’s disease 370,371,465 Amaranthin 407 Amaranthus 406,468

Amaranthus caudaius 406,407,409,428,440,

458 Amaranthus cruentus 405,458 Amaranthus leucocarpus 406,458 Amaranthus spinosus 458 Amaryllidaceae 407,415, 416 Amaryllis, see Hippeastrum Ambystoma maculaium 167 Ambystoma mexicanum 166 Amidination 605 cyanomethyl thioglycosides 605 imidoester 605 thioglycosidic linkage 605 Amino acid sequence 422 Amino acids 41 1 w-Aminoalkyl glycosides 608 1 -Amino-I-deoxy-lactitol 406 5-Amino-3,5-dideoxy-~-g~cero-~-galacfo-non-2ulosonic acid 244,248 Amoeba 439,493 CAMP 107 Amphibian glycoproteins 163-170 Amphibians N-acetyl-fi-D-glucosaminidase 165 Ambystoma 165 Ambystoma maculaium I66 Ambystoma tigrinurn 168 Bufo bufo 165 Bufo japonicus japonicus 165,169 eggs 164 fucosyltransferase 169 glycoproteins 163-170 glycosyltransferases I69 integumentary mucins I65 jelly 164 lectin 165 LeX 166 Ley 166 Pleurodeles 165 polyspermy 165 Rana temporaria 168 Rana utricularia I68 sperm binding 164 Xenopus laevis 169 integumentary mucins 165 Amphicarpa bracteata 406,459,468 Amphiglycan 19 Amphomycin 75 Amyloid plaques 465 Amylopectin 415 Androctonus australis 126 Anemia 78 Angiogenesis 528 Anguilla anguilla 424

623 antennapedia gene 5 19 Anti-adhesive drugs 501 Anti-blood group H 575 Anti-I 571-582 Anti-i 571-582 Antibacterial defence 216 Antibodies 63, 64, 403, 427, 429431, 435, 444, 476, 487, 493 Anticoagulant 61 0 Antigen GMEM 541 Antigenic determinants LeX 163, 571-582 sialyl 579 sulfated 579 Ley 163 Mycobacterium leprae 61 6 Antigens 41 1, 4 2 9 4 3 1, 435 Antigens, I and i 571-582 allergic diseases 579 allergic encephalitis 579 antibodies as reagents in glycosylation studies 574 antigen masking 575 band 3 576 band 3 protein 575 counter-receptors S76 embryogenesis 575 endo-b-galactosidase 577 gastrointestinal tumors 576 I Ma determinant 576 LeX 577 Ley 577 membrane labelling 575 neoglycolipid technology 576 poly-N-acetyllactosamine 575 polyglycosyl ceramides 576 secretors 576 selectins 576 teratocarcinoma cells 575 tumor-associated antigen 576 antigen masking 574 apoptosis 579 autoantibodies 573, 577 Epstein-Barr virus 573 Mycoplasrna pneumoniae 573 autoimmune disorder 573 autoimmune hemolytic anaemia 577 blood group A 579 blood group antigens 574 blood group B 579 carbohydrate-binding proteins 578 galectin-1 578 lectins 578 CD43 576

CD45 576 cold agglutinins 573, 577 counter-receptors 582 epidermal growth factor (EGF) receptor epithelia 578 epithelial glycoproteins 573 galectin-I 579 galectin-3 579 glycoproteins 576 HL60 576 host-parasite interactions 573 human adenocarcinomas 580 I Ma epitope 574 immunosequencing 576 lectins, cytoplasmic 579 lectins, extracellular 579 lectins, nuclear 579 leukocyte common antigen 576 leukocytes 573 Ley 580 ligands 578 lymphoblastoid cell lines 578 lymphocytes 576 macrophages 573 microbial attachment sites 577 milk 574 mouse embryo 580 mucin-type substances 574 ovarian cyst glycoproteins 574 poly-N-acetyllactosamine 573, 576 promyelocytes 576 receptor-mediated complexes 578 red cells 573 sialophorin 576 sialyl li immunological disorders 577 Mycoplasrna pneumoniae 577 pneumonia 577 respiratory tract infection 577 sialyl-I antigen 578 T200 576 T and B cells 576 T200/CD45 579 Antigens, inappropriate expression 576 Antithrombin 111 176, 177 a 1-Antitrypsin 176-1 78 Apolipoprotein B 178 Apolipoprotein D 178 Apolipoprotein F 179 Apolipoproteins 178, 179 Apoptosis 355 Aposerotransferrin 206 Apotransferrin receptor complex 21 6 Arabinogalactan 406

579

624 Arabinose 409 o-Arabinose 4 13 L-Arabinose 404, 4 I 7, 4 I8 Arachis hypogaea 405, 406, 410412, 421, 428, 437, 440, 441, 443, 459, 464, 468 Arg-Gly-Asp 500 Arthrobotrys oligospora 492 Arthropod hemocyanins 124-127 Artocarpus 468 Artocarpus atilis 405, 459 Artocarpus integrifolia 407409, 421, 422, 428, 459, 465 Asialofetuin 404 Asialoganglosides 487 Asialoglycoprotein receptor 354 Asn-oligosaccharide 606 Asparagus pea, see Lotus tetragonolobus Aspartate-boxes 334 Aspartylglycosaminuria 259 Asthma 344, 359 Astrocyte 587, 589 Astrotactin 588 Autoantibodies 577, 578 Autoantibody production 573 Autoantigen 577 Autoimmune conditions 177 Autoimmune diseases 359 Autoimmunization 577, 578 Avididstreptavidin 613 B cells 576 Bacteria 173, 439, 480491, 496 N-acetylneuraminic acid 496, 498 Actinomyces naeslundii 498 Actinomyces oiscosus 498 anti-adhesive drugs 498 antibodies 496 Bordetella pertussis 497 colonic mucosa 498 concanavalin A (Con A) 498 corneal epithelia 498 edema 497 enterobacteria 496 enterobacteriaceae 480 epithelial cells 498 Escherichia coli 480, 481, 498 Escherichia coli K99 498 fimbriae 496 type 1 496, 498 galactose 498 glucose 496 IgA 498 immunization 496 infections 498

killing by phagocytes 500 Klebsiella pneumoniae 480, 481 lectin dependent adhesion 498 mannose 496, 498 mannose-specific 480 methyl a-glucoside 496 methyl a-mannoside 496 mouth 498 mucosal 498 otitis externa 498 PA-I 49'8 PA-2 498 phagocytes cells 498 Pseudomonas aeruginosa 496 respiratory tract 497 Streptococcus sanguis 498 Tamn-Horsfall glycoprotein 498 teeth 498 trachea 497 urinary tract 496, 498 Bacterial polysaccharides 63 Baculovirus 42, 176, 177 gp41 42 overexpression system 42 Bak" determinant 554 Band 3 484, 486 Band 4.5 484, 486 Bandeiraea, see Griffonia Barley, see Hordeum ouigare Basal cell carcinomas 326 Basement membrane 507-509, 528, 535, 538, 539, 544 Basic phosphoprotein (BPP, UL32) 42 Basophils 21, 173 Bauhinia purpurea 405, 428, 459, 465, 468 3-0-Benzyl glucose 414 &pleated sheet 420 Biglycan 17, 18 Binding site 477, 478, 491, 493 Biochemical evolution 203 BioGel P, see Polyacrylamide Biological lubricants 354 Biotechnology 572 Biotin 613 Bird-conformation 232 Blood 173-198 Blood factor XI11 transglutaminase 530 Blood group antigens A, B and H 574, 576, 578 Blood group substances 430 Lewis b 423,424 type A 419, 423, 439 type A , 421 type A, 421 type B 439

625 type B, 421 type H 419, 423, 424 type NM 465 type 0 439 Blood groups 486 A 579 B 579, 611 blood group chain, a-I ,3-fucosylated 572 H type 4 in hemocyanins 132 Blot analysis 427, 43 I , 4 3 4 4 3 7 Boletus satanas 459, 465, 467 Bone marrow cells 441 Boodlea coacta 462, 467 Bordetella pertussis 489491, 500 Botrytis cinerea 459, 467 Bovine lactotransferrin 213, 214, 218 Bovine serum albumin (BSA) 481 Bowringia mildbraedii 41 5, 456, 464, 469 Brachypodium sylvaticum 457, 467 Bradyrhizobium japonicum 488 Brain glycoproteins 55-66 Brevican 16, 17 Broken wing-conformation 232, 234 Bromelain 137 Bronchial secretions 207 Brush border 486, 489 Blyonia dioica 405, 428, 459, 465, 470 Bufo bufo 170 C1 inhibitor 179, 180 Cadherins 5 10-5 17, 550 E-Cadherin 5 10-5 13 M-Cadherin 5 13 N-Cadherin 51 I , 513, 521, 588 P-Cadherin 5 1 1-5 13 R-Cadherin 5 13 T-Cadherin 5 13 CAM, see Cell adhesion molecules Canaoalia ensformis 405, 410, 411, 428, 440, 456, 465, 469 Cancer 40 Capsular polysaccharides 62, 63 Capture of iron 2 14 Caragana arborescens 405, 428, 459, 469 Carbohydrate antigens 571-582 Carbohydrate-binding proteins 65, 572, 573, 578 Carbohydrate binding site 407 Carbohydrate chains diantennary 177 triantennary 177 Carbohydrate content of hemocyanins 130 (Carbohydrate) differentiation antigens 57 1-582 antigens, 1 and i 571 blood group antigens

A 571 B 571 changes in expression 571 H ( 0 ) 571 Lewis 571 cell adhesion 573 cell surface antigens 571 embryonic antigen, stage-specific 572 Forssman carbohydrate antigen 572 glycosyltransferase activities 57 I hybridoma technology 572 Le" 572 oligosaccharides as area codes 572 selectins 572 SSEA-I (stage specific embryonic antigen 1) 572 tropisms of microbes 573 tumor-associated carbohydrate antigens 572 Carbohydrate transport proteins 403 Carbohydrate-triggered pathology 577 Carbohydrate-protein interactions 572 Carboxyl groups 605 Castanospemine 522, 557, 596 Castor bean, see Ricinus communis Catenins 512-514 CD15 590 CD24 589 CD29 196 CD31 198 CD36 196, 197 collagen 196 di-, tri-, and tetraantennary glycans 197 erythrocytes 196 high-mannose glycan 196 hybrid glycans 196 mammary epithelial cells 196 oxidised LDL 196 Plasmodium falciparum 196 thrombospondin I96 CD43 576 CDlI/CDI8 500 CDllb/CD18 501 CD I I b,c/CDI 8 500 Cell adhesion 65, 353 Cell adhesion molecules (CAM) 364, 510 A-CAM 511 ICAM, see ICAM L-CAM 51 1, 513 N-CAL CAM 51 1 N-CAM (neural cell adhesion molecule) 58, 63, 65, 5 10, 5 17-524, 588, 59 I , 592,594,595 Ng-CAM (neuron-glia cell adhesion molecule) 62

626 Cell adhesion molecules (CAM) (cont if) PECAM (platelet endothelial cell adhesion molecule) 525, 528 PECAM-1 (platelet endotheliaf cell adhesion molecule-1) 198, 528 VCAM (vascular cell adhesion molecule) 525, 527, 528, 549 Cell agglutination 577 Cell aggregation 352 Cell CAM 120180 510 Cell free systems mammalian cells 73 toxoplasma 73 trypanosomes 72 yeast 73 Cell lines 575 8-Cell stage embryo 580 Cell surface 475, 476 Cell surface carbohydrates 55 Cell surface labelling 61 Cell wall polysaccharides 487 Cell-cell adhesion 507-509, 520 Cell-substratum adhesion 509 Cellobiose 41 8 Cellular agglutination 441 Cellular immunology 438 Cellulose 90, 1 1 I , 406 Ceratobasidium cornigerum 459, 465, 467 Cerebellum 588, 589, 592, 593 Cerebroglycan 20 Cerehroside 485 Ceruloplasmin 179, 436 free oxygen radicals I79 CFA, see Colonization Factor Antigen Chagas disease 339, 341 Chaperons 353, 481 Charge microheterogeneity 174 Chelidonium majus 406, 457, 468 Chemical ionization (CI) MS 277 electrospray MS 28 1 liquid secondary ion MS 281 Chemo-enzymatic methods 61 1, 618 Chemo-enzymic synthesis 306 Chemotherapeutic strategies 77 Chitin 408, 417 Chitinase 409 Chitodextrins 409, 417, 418, 423 Chitooligosaccharide-specific 494 Chitooligosaccharides 492 Chitopentaose 41 8 Chitotetraose 418 Chitotriose 418 CHO cells 491493, 596 Cholera toxin 343

Cholesterol 178 Cholesterol transport 178 Chondroitin, biosynthesis 10 Chondroitin sulfate 2, 6, 16-18, 20, 21, 46, 47, 521, 530, 543 degradation by hyaluronidase 13 proteoglycan 59, 60 structure 5 Chondroitinase 588 Chromatin proteins 39, 40 Chronic alcoholics 178 Chronic inflammation 175 Cicer arietinum 428, 462, 469 Ciliated cells 490, 491 Circular permutation 41 1, 415 Cirrhosis I75 Class I histocompatibility antigens 55 I Class 11 histocompatibility antigens 55 1 Clathim coated pits 82 Clerodendron trichotomum 405, 459, 465, 470 Clivia miniata 405, 407, 456, 468 Clostridial myonecrosis 343 CMP-9-fluoresceinyl-Neu5Ac 3 13 CMP-glycosides 292 CMP-Neu5Ac 289 hydroxylase Asterias rubens 322 iron-sulfur center, Rieske type 322 redox proteins 322 phosphodiesterase activity 273 synthase 292, 31 1 active site 3 I3 activity 273 inhibition 3 I3 CMP-P-NeuSAc (cytidine 5’-(5-acetamido3,5-dideoxy-o-glycero-~-~-galacto-non-2ulopyranosylonate monophosphate) 249 CMP-NeuSGc (CMP-N-glycolylneuraminic acid) 322 hiosynthesis 323 CMP-sialate hydrolase 332 CMP-sialic acids antiporter 323 biosynthesis 3 1 1-3 13 C-myc 40 Coaggregation 488 Coagulation factor V 180, 18 1 deoxymannojirimycin 181 factor Va 181 factor Xa 181 haemostasis 180 plasma concentration 181 platelet a-granules 18 1 prothrombin 180

627 thrombin 181 tunicamycin 181 Coagulation factor VII 18 1 blood coagulation 181 epidermal growth factor (EGF) I81 extrinsic pathway 18 1 factor VIIa 181 plasma 181 plasma concentration I8I vitamin K 181 Coagulation factor VIII I81 -1 83 baboons 183 deoxymannojinmycin 183 di-, tri-, and tetraantennary glycans 181 Gal(a 1-3)Gal group 183 haemophilia A 181 high mannose type 181 plasma 183 plasma concentration I 8 1 porcine factor VIII 183 Spodoptera frugiperda 183 tunicamycin 183 von Willebrand factor I 8 I , 183 von Willebrand's disease 181 Coccinea indica 405,457,470 Coccinea indica leukoagglutinin 404 Codium 467 Codium fragile 457,459,464 Codium tomentosum 405,457,464 Colchicum autumnale 405,428,459,467 Collagens 509,510,524,528,537,544,547,549,

550 type I 545 type II 545 type IV 509,528,535,538,541,544,550 type V 545 type IX 19 Colonic carcinoma 464 Colonic mucin 493 Colonic mucosa 492 Colonization Factor Antigen (CFA) 483,486 Colorectal cancer 425 Combining site 478 Complement 466,500 Complement C3 180 immunoregulatory processes 180 Complement fixation 577 Complex binding site lectins 424,425 Complex Carbohydrate Structural Database (CARBBANK) 282 Complex oligosaccharides 432 Complex type oligosaccharides 613 Complex unit 477 Con A, see Concanavalin A

Conalbumin, see Ovotransferrin Concanavalin A (Con A ) 56,57,409,41 1416,

419,426,430432,435437,439,440,483, 539,615 Conformational epitope 63 Connective tissue of gastropods 137 Contact guidance 588,589 Contact site A (csA) 109 Contact site B (csB) 108 Convergent evolution 408 Copper-containing glycoprotein 179 Core 2 sequence 576 Core structures 55-57 Cortical alveolus glycoproteins 144-154 Corticomedullary junction 543 Counter-receptors 582 COWlactotransferrin 214 Craylia mollis 405,456,459,469 CROSREL (cross relaxation) 137 Cross-reacting determinant (CRD) 72 Cross relaxation (CROSREL) 137 Crotalaria juncea 405,406,459,469 Crown gall 489 C-type lectins 490,523 Cymbidium 405,415,416,456,464,468 Cyphomandra betacea 406,457,470 Cysteine proteinase 97 Cysteinyl residues 608 Cystic fibrosis I77 Cytidine 5'-(5-acetamido-3,5-dideoxy-~-glycerofl-D-galacto-non-2-ulopyranosylonate monophosphate (CMP-fl-Neu5Ac) 249 Cytisus 469 Cytisus scoparius 405,428,459 Cytisus sessili$olius 406,417419,457 Cytochalasin 525 Cytochemical markers 614 Cytochrome b, 322 Cytochrome b,-reductase 322 Cytokeratins 41,42 Cytokines 364,367 Cytomegalovirus 464 Cytoplasmic glycosylation 102 Cytotactin 541,588 Cytotoxicity 422 Daffodil, see Narcissus pseudonarcissus N-Dansylgalactosamine 420 Datura stramonium 404,405,409,417,418,428,

437,440,457,464,470 Deacetylation 75,78 Deaminated neuraminic acid Decorin 7,17 binding 18

249

628 Decorin (cont a) binding to collagen fibrils 18 binding to transforming growth factor P (TGFP)

common antigen 1 (CAI) 104 conditioned medium factors (CMFs) 107 contact site A (csA) 101, 107, 11 I 18 contact site B (csB) 109 De-U-esterification 264 contact site C (csC) I09 2-Deoxy-~-arabinohexose 4 14 cysteine proteinases 98, 100 3-Deoxy-D-g&cero-o-ga~acto-nononic acid I45 development 89-9 1, I09 3-Deoxy-D-g&cero-o-ga~acto-non-2-u~opyranosonic ecmA 112 acid 249 ecmB 112 3-Deoxy-o-g&cero-~-D-ga~aclo-non-2FP21 102 ulopyranosonic acid 248 fucosyl transferase 102 acid 3-Deoxy-D-g&cero-~-ga~acto-nonu~osonic gp24 107, 108 I63 gp80 96, 101, 105, 106, 109 Deoxymannojirimycin 596 gp130 109 I-Deoxymannojirimycin 92, 555, 556 gp138A 106 Deoxynojirimycin 596 gp138B 107, 109 Dermatan, biosynthesis 10 gp150 107, 1 1 1 Dermatan sulfate 2, 6, 18 lectins cofactor activation 25 concanavalin A (C 1 A) 106, 114 structure 5 LCA 106 Desmocollins 5 14-5 17 RCA-I 114, 115 Desmogleins 514, 516, 517 SBA 114, 115 Desmoplakins 5 14 WGA 105, 106, 109, 113, 114 Desmosomal glycoproteins 5 14-5 17 life cycle 89-91, 106-1 15 Desmosomal junction 5 14, 5 I6 mAb 103 Desmosomes 507, 509, 513, 514, 516, 517 Man-630, 96 Detergent 43 1 mating types 106 Detergent partitioning, Triton X-I14 72 mod4 104, 114 Developmentally regulated oligosaccharides 582 modB 100, 104, 105, 109, 110, 114 Dextran 413, 415 modB-dependent 97 Dextran sulfate 491 modC 105 N,N’-Diacetylchitobiose 405, 418, 493 modD 105 N.N’-Diacetyllactosediaminetype 132 modE 105 5,7-Diamino-3,5,7,9-tetradeoxy-~-g&cero-~morphogenesis 89 galacto-non-2-ulosonic acid 262 phosphoglycosylation 97 5,7-Diamino-3,5,7,9-tetradeoxy-~-g&cero-~-manno- polysaccharide 90 non-2-ulosonic acid 262 ponticulin 101, 106, 107 Diantennary glycans 56, 196 pre-spore vesicle (PSV) 90, 112-1 14 Diazotization 608 PsA 100, 101, 105, 106 Dictyostelium discoideum 89-1 17 PsB 115 adhesion molecules 108-1 1 I slug 90 EDTA-resistant I09 SP29 101 EDTA-sensitive 108 SP75 115 gpl50 110 SP80 115 gp24 110 SP96 115 post-aggregation I 10 spore coat 90, 100, 112, 113, 115 aggregation 90 ST310 112 anterior-like cells (ALC) 90 ST430 112 CA2 104 stalk cells 105 CA3 104 stalk tube 90 CAMP 90 surface sheath 90, I I 1 cell adhesion 1 I 1 Differentiation 61, 65, 575 cellulase 91 Differentiation antigens 57 1-582 cellulose 90, I15 Dioclea grandijora 405, 456, 469

629 Diocleae 407, 412 Disaccharide type A 420 Disintegrins 409 Disulfide bonds 404, 408, 417419, 422, 424, 426 Dithiothreitol (DTT) 75 cDNA cloning 407 Dolichos biforus 406, 419, 428, 440, 441, 459, 465, 469 Dol-P-Man 75 Double walled coated vesicles 592, 593 Down’s syndrome 37 1 DTT (dithiothreitol) 75

EBA-I75 494 Echinocystis lobata 406, 459, 470 Edible bird’s nest substance 25 1 Effector functions in immunity 572 EF-hand 544 EGF 579 EGF-like repeats 535, 538, 539, 541, 543 Egg-white 205 Ehrlich ascites tumor cells 422, 426, 440, 466 Elderberry, see Sambucus nigra Japanese, see Sambucus sieboldiana Electron impact (El) MS identification 275, 278 N-acyl-0-alkylneuraminicacid 276 N.0-acylneuraminic acids 275 Electron microscope 41 I Elutriation centrifugation 442 Embryogenesis 535 Embryonal carcinoma cells 576 Embryonic cells 580 Embryonic development 575 Embryos 575 chick 575 murine 575 En(a-) erythrocytes 484 Endo-A 614 Endo-F 93 Endo-H 92 Endo-H-resistant 92, 109 Endo-M 614 Endo-N (endosialidase) 63, 64, 267, 331, 332, 52 1 Escherichia coli bacteriophage 092 268 Escherichia coli KI 267 Endo-b-galactosidase 61, 486 Endo-0-glycosidase 575 Endo-0-hexosaminidases 603 Endocytosis 58 1 Endoplasmic reticulum I90 Endosialidase, see Endo-N

Endothelial cells 173, 579 Engelbroth-Holm-Swarm (EHS) tumor 535, 539, 541,544 Entactin 528 Entactinhidogen 550 Entamoeba 439 Entamoeba histolytica 493, 494 chitooligosaccharides 494 Enterobacteria 48 1 Enterotoxigenic 483 Enzymatic methods 61 0-613 Enzymes 403 Eosinophils I73 Epidermis 550 Epilgrin 550 Epipactis helleborine 405, 407, 415, 416, 456, 464, 468 Episialin 353 Epithelial 487 Epithelial cell differentiation 466 Epithelial cells 483, 485, 492, 575 EPR spin labels 291 Equilibrium dialysis 409, 421, Eranthis hyemalis 405, 459, 465, 468 E-rosettes 441 Erythrina 405, 406, 412, 420, 421. 459, 469 Erythrina corallodendron 41 0, 428 Erythrina crislagalli 404, 428, 465 Erylhrina crystagalli agglutinin 61 1 Erythrina uariegafa 440 Erythroagglutinin 422, 424 Erythrocyte band 4. I 41 Erythrocytes 60, 173, 355, 418, 441, 444, 476, 478, 483489, 491, 494, 501 animal 420 human 421 type A , 419 type A, 419 type 0 423 type A 422 type B 422 Erythropoietin 321 Escherichia coli 475, 483, 484, 486, 500, 501 Escherichia coli KI 63 Estrogen receptor 40 N-Ethylmaleimide (NEM) 75 Eukaryotic initiation factor 2 (eIF-2) 41 Euonymus europaeus 405, 406, 428, 462, 466, 469 Euphorbia rnarginata 405, 459, 469 Extracellular matrix 507, 509, 5 10, 5 12, 520, 528, 529, 537, 543, 554 Eye infections 486

630 Factor IX (Christmas factor) 183, 184 Christmas disease 183 EGF (epidermal growth factor) domain 183 epidermal growth factor (EGF) 184 haemophilia B 183 plasma concentration 183 Factor J complement inhibitor 180 Factor X 184 di-, tri- and tetraantennary oligosaccharides 184 extrinsic pathway 184 extrinsic Xase (factor VIIa/tissue factor/ phospholipid) 184 factor VIIdtissue factor complex 184 factor-X-activating enzyme 184 glutamic acid y-carboxylation 184 intrinsic pathway 184 plasma concentration 184 prothrombin 184 thromboplastin 184 Factor XI 185 plasma concentration I85 Factor XI1 185 epidermal growth factor (EGF) 185 Falcata (=Amphicarpa) 469 Falcata japonica 405, 460 Fatty acids alkyl-acylglycerol 70 ceramide 70 diacylglycerol 70 human erythrocyte acetylcholinesterase 70 palmitate 70 slime mold and yeast GPIs 70 stearoyl-lysoglycerol 70 Fava bean, see Viciafaba Favin 414, 415 Ferrideprivation 2 19 (apo)Ferritin 204 Ferroxidase 179 a-Fetoprotein 185, 435, 464 Fetuin 425, 462, 488, 491, 492 FGF (Fibroblast Growth Factor) 24 Fibrillogenesis 53 1 Fibrinogen 185, 549, 552 Fibroblast Growth Factor (FGF) 24 Fibroglycan 19 Fibromodulin 17, 18 Fibronectin 186, 187, 509, 517, 528-535, 541, 543-545, 549, 550, 555, 556 carcinoma cell lines 186 cell-binding sites 530 collagen 186 collagen binding 530, 532, 535 connecting sequence (CS) 529

embryonal tissue 186 extra domains ED-A 529 extra domains ED-B 529 extracellular matrix 186 fetal cells 186 fibrin 186 fibrin binding 530 glycosylation 53 1-535 hepatoma 186 monoclonal antibody FDC-6 186 oral mucosa 186 oral squamous cell carcinomas 186 osteoarthritis 187 pathogenic bacteria 530 placenta 186 rheumatoid arthritis 187 splicing 529 synovial fluid 186 synoviocytes 186, I87 type I repeat 529, 530 type I1 repeat 529, 530 type 111 repeat 517, 519, 521, 530, 531, 543, 549 variable connecting sequence 556 variable (V) 529 Fibronectin receptor 546 Ficus cunia 405, 457, 468 Filamentous hemagglutinin 491 FimA 482 JimA 487 JimA gene 481 FimB 481 Fimbriae (pili) 475, 482, 487, 488, 501 type I 480483, 500, 501 CDIl/CDI8 483 integrin 483 mannose 475 type 1 macrophages 483 type INCA-50 483 type 2 487, 500 Gal(P13)GalNAc 476 type K88 486 type K99 483, 484 Escherichia coli K99, 484 type P 48 1, 482 galabiose 475 type S 483, 485 NeuAc(a2-3)Gal 475 FimC 481 FimH 4 8 1 4 8 3 Fish glycoproteins 143-1 59 teleost eggs 144 Flow cytometry 426, 443, 444

63 1 Fluorescence activated cell sorting (FACS) 443, 444 Fluorescence flow cytometry 439 Fluorescence microscopy 436, 438 Fluorescence spectroscopy 409 Fluorescent glycans 435, 436 Fluorescent group 61 I Focal contacts 509, 512, 549 Follistatin 544 Forssman disaccharide 4 19 Forssman pentasaccharide 41 9 FP2l 45, 46 Fractionation of glycans 56 Fractionation of glycopeptides 56 Fructans 430 Fructose 414 D-Fructose 41 3 Fuc(aI-2)Gal(fiI-3/4)GlcNAc 486 Fucose 56, 174, 179, 485, 486, 489 L-Fucose 405, 406, 465 a-L-Fucose 414 L-Fucose-binding lectins 423, 424 Fucose-specific 488 a-L-Fucosyl 420 3-Fucosyl-N-acetyllactosamine 580 3-Fucosylated N-acetyllactosamine 580 a-1,3-Fucosylation 580 Fucosyltransferase 489 cytosol 45, 46 Fungi 403, 492, 498 Sclerotium rolfsii 498 Trichoderma 498 Trichoderma harzianum 498 Gal 486 Galactia tashiroi 405, 460, 469 Galactose 405, 409, 41 1, 417, 424, 444, 458, 464, 478, 480, 486-488, 490, 493 a-Galactose 59, 65 D-Galactose 404 Galactose-binding lectins 56, 419-422 Galactoselhr-acetylgalactosamine 492 Galactose oxidase 576 Galactose-recognizing receptor 358 Galactose-specific lectin, sponge Geodia cydoniurn 354 Galactose-specific receptor 354, 356 Galactosialidosis 370 p-galactosidase 370 serine carboxypeptidase 370 sialidase 370 sialo-oligosaccharides 370 excretion 370 p-Galactosidase 488

6-Galactoside binding lectin 578 a-Galactosides 486 Galactosyl 406 a-Galactosyl 41 8 Galactosyltransferase 34, 38, 61, 93, 588 Galanthus nivalis 405, 410, 415, 416, 428, 437, 440, 456, 464,468 Galectin 579, 582, 590, 592 Galectin-1 559, 579 Galectin-3 559, 579 Gal(a 1-3)Gal 486 Gal(a14)Gal 475 Gal(aI-3)GalNAc in adrenal medulla 59 in brain 59 in muscle 59 in teratocarcinoma cells 59 Gal(fiI-3)GalNAc 486, 487, 500 Gal@-3)GlcNAc 56, 487 Gal(fiI4)Glc 500 Gal(fiI4)GlcNAc 487 GalJ(fil-6)Gal(filL3)GalNAc(al-) 487 Galf(fil-6)GalNAc(~l-3)Gal(al-) 487 GalNAc-specific 486 GalNAc(p14)Gal 486 Gal-6-sulfate 491 Ganglioside 259, 483, 490, 491, 521, 531, 549, 553, 588 Ganglioside GD3, 0-acetylated 330 Ganglioside GM, 485 Gap junctions 507, 513 Garlic, see Allium sativum Gastric carcinomas 576 Gastric cells 489 Gastritis 485 GDP-Fuc 105 GDP-Fuc:Gal(p1-3)GalNAc (Fuc to Gal) a-I ,2-fucosyltransferase 140 GDP-Fuc:Gal(fi1-4)GlcNAc (Fuc to GlcNAc) a-I ,3-fucosyltransferase 140 GDP-Man 105 Gel electrophoresis 427 Gel filtration 56, 60, 62, 419 Gene expression 403 Gentiobiose 41 8 Giardia lamblia 495 enterocytes 495 erythrocytes 495 mannose-I-phosphate 495 parasite 495 protozoan 495 taglin 495 GIPLs (glycoinositolphospholipids) 71 GLC 278

632 GLC-EI MS analysis 275 GlcNAc GlcNAcl-P 97 GlcNAc- 1 -P phosphodiester a-N-acetylglucosaminidase 94 GlcNAc-P-Ser 97 GlcNAc-6-SOd 104 intersecting 92 0-GIcNAc 3 3 4 2 GlcNAc-epimerase 290 GlcNAc transferase I 588, 596 0-GlcNAc transferase 35-38 0-GlcNAcase 37 cytosolic active site 37 deglycosylase 42 inhibitors 38 subunit structure 37 GlcNAc(fi1-3)Man-01 59 GlcNAc(fi1-3)Man-ol in brain 59 GlcNAc-PI synthesis, deficiency 75 Glc-phosphotransferase 43 Glial wrapping 593-596 a2-Globulin 176 t-Globulin 229 Glomerulonephritis 344, 359 Glucans 430 a-Glucans 415 a-D-Ghcans 4 13 P-Glucans 409 fl-Glucosaminidase 35 Glucosaminoglycans, heparan sulfate 20 Glucose 409, 414, 456, 464 D-Glucose 413, 415 Glucose-I -phosphate phosphodiesterase 43 Glucose polymer 65 al,3-Glucosidase 104 a-Glucoside 496 Glucosyltransferase 44 0-Glucuronidase 180 Glycamine 607, 608 E-amino groups 608 diazotization 607 isothiocyanate group 608 reductive amination 607 4-trifluoroacetamido-aniline 607 water-soluble carbodiimide 608 Glycan cores 57 N-Glycanase 425, 437 Glycans 438, 531, 532 Gbcine m a 405, 406, 410412, 420, 428, 441, 442, 460, 465, 469, 488 Glycoamidases 603 Glycocalicin 192-194 A, B, C and D phenotypes 193

calpain 192 size polymorphism 193 surface-connected canalicular system I92 tetraantennary glycan 193 thrombin 192, 194 von Willebrand factor 192, 194 Glycoconjugate conformations 353 Glycoconjugates 41 1 Glycogen 44, 45, 415,419 Glycogen synthase 44, 45 Glycogenin 44, 45 glucosylation of Tyr-194 44 Glycoinositolglycerolipids 70 Glycoinositolphospholipids (GIPLs) 7 I Glycolipids 254, 425, 431, 434, 475, 476, 484486, 491, 495, 575, 576 Glycolipids of erythrocytes 574 N-Glycoloylneuraminic acid 484 N-Glycoloylneuraminyllactosyl ceramide 484 N-Glycolylneuraminic acid 249 biological significance 321 biosynthesis 320-324 functions 320-324 occurrence 320 Glycopeptides 56, 60, 62, 65, 405, 486, 61 3 Glycophorin 355, 424, 483, 484, 494 Glycophorin A 368, 417 Glycophospholipid 101, 107 Glycophosphosphingolipids 71 BM40 Glycoprotein 544 a2-HS-Glycoprotein 187 ovine fetuin I87 tn’-antennary glycans 187 triantennary glycans 187 N-Glycoprotein glycans 25 1 0-Glycoprotein glycans 25 1 Glycoprotein hormones 355 Glycoprotein Ib-V-IX 192 thrombin 192 von Willebrand factor 192 Glycoprotein Iba 192-194 A, B, C and D phenotypes 193 calpain 192 size polymorphism 193 surface-connected canalicular system 192 tetraantennary glycan 193 thrombin 192, 194 von Willebrand factor 192, 194 Glycoprotein Ibfi 194 diantennary glycan 194 leucine-rich repeat 194 Glycoprotein Ilb 195 Glycoprotein IIb-IIIa 195 CD41/61 195

633 cell adhesion 195 clot retraction I95 extracellular matrix I95 Glanzmann’s thrombasthenia 195 haemostasis 195 signal transduction 195 subendothelium 195 thrombus 195 wound healing 195 Glycoprotein llla 196 high mannose glycan 196 Glycoprotein IX 194 diantennary glycan 194 glycoprotein It$ I94 leucine-rich repeat 194 Glycoprotein N-glycans 288 Glycoprotein 0-glycans 288 Glycoprotein receptors 475 Glycoprotein Thy-I rat 69 Glycoprotein V 194 leucine-rich repeat I94 thrombin 194 Glycoproteins 575, 576 adhesive 507-559 amphibian 163-1 70 antifreeze 42 I preparation 55 Glycosaminoglycans 1-25, 46, 532, 588 biosynthesis 7-1 2 chondroitin sulfate I0 dermatan sulfate 10 Golgi 8, 9 heparan sulfate 10 heparin I0 hyaluronan 10 keratan sulfate 10 precursor activation 8 chondroitin sulfate 5, 6, 10, 16-21 degradation by bacterial lyases 14 degradation by endoglycosidases 12 degradation by hyaluronidase 12 degradation in lysosomes 14 dermatan sulfate 5, 6, 10, 18 heparan sulfate 5, 6, 10, 18, 19, 21 heparin 5, 6, 10, 21 hyaluronan 4, 10 interactions with proteins 22-25 keratan sulfate 7, 10, 16, 18 linkage region 6, 7 nucleus 48 polymerization 10 structure 3-7 structure/function relations 22-25

sulfation 10 Glycosciences, techniques 582 N-Glycosides 603, 616 0-Glycosides 603, 604 Glycosides of sialic acids 299-304 CMP-sialic acid synthesis 302 dendrimers 304 fluonmetric sialyltransferase assay 303 immobilization of sialic acids 303 Koenigs-Knorr method 299 N-glycosides 299 9-0-acetylesterase 301 Se-glycosides 299 S-glycosides 299 sialate U-acetylesterase 301 sialidase 299 sialidase inhibitors 299 Glycosphingolipids 485 Glycosyl-phosphatidylinositol 69-84 Glycosylamines 605-607 U-Glycosylation 5 16, 5 19, 532, 554 Glycosylation sites 21 0, 575 Glycosylation specific monoclonal antibodies 178 Glycosyltransferases 439, 610, 61 1, 618 L-Fuc-transferases 6 I I galactosyl-transferase 6 I 1 in snail connective tissue 137-140 ovalbumin 61 1 sialyltransferases 61 0, 61 1 Glypican 20 GM, 485 GM 1 -gangliosidosis 60 Goblet cells 327 Golgi 77, 438 in glycosaminoglycan biosynthesis 8, 9 Golgi enzyme 436 Gp80 I10 GPI anchors 69-84, 507, 513, 519, 527, 589, 597 attachment 77 endoplasmic reticulum 77 transamidation 77 biosynthesis 74-79 Dol-P-Man (dolichyl-phosphoryl-mannose) 75 phosphoethanolamine 75, 76 topology 78 endoplasmic reticulum 78 yeast mutants 78 cleavage pancreatic granule membrane protein, GP2 80 phospholipases 80

634 GPI anchors - cleavage (contbr) secretion 80 functions 79-83 mammalian 75, 76 remodeling 77 fatty acids 77 myristate 77 structural diversity a-linked galactoses 70 a-linked mannose 70 sialic acids 70 structure 70-72 glucosamine 70 inositol 70 mannose 70 phosphoethanolamine 70 trypanosome 74-77 GPI-anchored proteins caveolae 81 chemical cleavage 72 hydrofluoric acid 72 mild alkali treatment 72 nitrous acid 72 endocytosis 81, 82 enzymatic cleavage 72 phospholipase C 72 lateral diffusion 80 metabolic labeling 72 potocytosis 81, 82 sorting 80, 81 targetting epithelial cells 80, 81 signal 80, 81 GPIIIb 196, 197 GPI-linked Thy-1, deficiency 75 GPIV 196, 197 Gramineae 404, 408, 4 17 Granulocytes 483, 500 Grijlbnia simplicifolia 405, 406, 409, 410, 417419, 422, 424, 428, 430, 440, 443, 457, 460, 461, 464, 465, 469 Group B Streptococcus 341 Growth factor activity 218 Growth regulation 579 Guaran 406 H antigen 489 Hanganutziu-Deicher antibodies 321 Hapten inhibition 409, 417 HAV sequence 5 1 1, 5 13 Heat stable antigen 589 a-Helices 404 Helicobacter pylori 485, 488, 489 Helix pomatia 422

Hemagglutinating activity 488 Hemagglutination 438, 483, 485, 489, 491, 493 Hemagglutinin 477, 480, 489, 495, 500, 501 see also under Virus - influenza galactose 491 hydrogen bonds 478, 491 hydrophobic interactions 491 influenza virus N-acetylneuraminic acid 475 NeuAc(a2-6)Gal 491 sialic acid 491 sialic acid interaction, influenza A virus 360 undecasaccharide 491 van der Waals contacts 478 Hemidesmosomes 509, 547 Hemocyanin glycans Androctonus australis I26 Astacus leptodactylus 127 Helix pomatia 132 Lymnaea stagnalis 129, 137-140 mollusc 129-1 37 Panulirus interruptus 127 Hemocyanins 123-140 Acila castrensis 129, 130 allosteric behavior 124 Androctonus australis 124, 126, 127 arthropods 124-1 27 Astacus leptodactylus 124, 126 Astacus leptodactylus oligomannose type 128 binuclear Cu(I) site 123 blood group H type 4 132 Buccinum undatum 130 Busycon canaliculatum I29 Busycon carica 130 carbohydrate content 124-1 36 Colus gracilus 130 conformational analysis 136 Cupiennius salei 126 Eutypelma califoornicum 124, 126 Eutypelma helluo 124 Helix pomatia 124, 129, 130 Helix pomatia N-linked carbohydrate chains 133 hemolymph 123 Homarus americanus I26 Lewisx 140 Limulus polyphemus 126 Loligo forbesi 130 Lymnaea stagnalis 124, 129, 130 Lymnaea stagnalis N-linked carbohydrate chains 131 Megathura crenulata 130 3-0-methylgalactose 129 3-0-methylmannose 129

635 Molecular Dynamics 137 molluscs 124, 125, 129-136 Mopalia muscosa 130 Neptunica antiqua 130 NuttallinaJIuna 130 Octopus vulgaris 129, 130 oligomannose type 124 Panulirus interruptus 124, 126, I29 quaternary structure 126 Scutigera coleoptrata 126 Sepia oficinalis 130 Stenoplen conspicua 129, 130 xylose 129 xylose-containing elements 136 Hemolymph 124 Hemopexin 187 di- and triantennary glycans 187 Hemopoietic cells 439, 441 Hemostasis 173 Hemostatic plug 185 Hen ovotransferrin 2 10, 2 12 Hen serotransferrin 2 I2 Heparan, biosynthesis 10 Heparan sulfate 2, 6, 18-22, 65, 509, 520, 52 1, 528, 530, 532, 537, 538, 541, 543, 555 anticoagulant capability 24 binding to proteins 22-24 degradation by endouronidase I3 degradation in lysosomes 14 nuclear 22 nucleus 48 structure 5, 6 Heparin 2, 6, 21, 22, 176, 491, 520, 524, 525, 528, 530, 535, 537, 538, 541, 543, 555, 595 anticoagulant capability 24 binding to proteins 22-24 biosynthesis 10 degradation by endouronidase 13 degradation in lysosomes 14 structure 5, 6 Heparinase 543 Hepatic lectins 614, 616 Hepatocellular carcinoma 177 Hepatocytes 178 Hepatoma 3 1 I , 320 Hericium erinaceum 406, 462, 467 Heuea brasiliensis 408 Hevein domain 408, 409 Hexabrachion 541 (3-Hexosaminidase 94 Hexosaminidases 37, 38 High-mannose glycans 56, 58, 196 High-mannose glycoproteins 41 5

High-mannose type oligosaccharides 177, 178, 613 High-mobility group (HMG) proteins 48 Hippeasirurn 405, 415, 416, 456, 468 Histochemical stain 419 Histochemistry 437, 444, 445 HMG proteins 49 HNK-I 60, 64, 65, 588, 589, 595, 596 Homeobox genes 5 19 Homeobox sequence 5 19 Homology 490 amino acid sequence 404409, 41 1, 415, 417, 423 interspecific 407 Hordeum vulgare 405, 417, 457 Horse gram, see Dolichos biJlorus Horseradish peroxidase 92, 104 Horseshoe crab, see Limulus polyphemus HPLC 272, 274, 427, 434 HPLUESIMS 178 HPLC-CI MS 281 Human cytomegalovirus (HCMV) 42 Human foetal intestinal brush border 2 17 Human immunodeficiency virus (HIV) 175, 416, 464, 465 Human lactotransferrin 210, 213, 222 Human melanoma-associated antigen p97 207 Human melanotransferrin 208 Human meningitis 484 Human serotransferrin 21 0 Hura crepifans 405, 460, 469 Hyaluronan (hyaluronic acid) 2 aggregate formation 16, 17 biosynthesis I0 rat cerebellum 47 structure 4 Hybridoma antibodies 574-577 Hybridoma-derived antibodies 575 Hydrazinolysis 603 Hydrogen bonds 410, 41 1 Hydrophobic aglycone 486 Hydrophobic binding site 413 Hydrophobic interactions 4 10 Hydroxyproline 417, 418 Hyosophorins 144-154 biological function 151-154 calcium ion binding 153 Cfupeopallasii I 5 I Cyprinus carpio 1 5 1 Fundulus heteroclitus 151 Paralichthys olivaceus I5 I pentaantennary glycan 150 Plecoglossus altivelis 151 poly-N-acetyllactosamine 150

636 Hyosophorins (conta) sperm agglutination 152, 153 tetraantennary glycan 15 1 Tribolodon hakonensis 151 Hypervariability 407 I/i antigen 485 I Ma 574 I Ma determinant 576 ICAM (immune cell adhesion molecule) 525-527, 549 ICAM-I 525-527, 549, 556, 557 ICAM-2 526, 527, 549 ICAM-3 527 I-cell disease 259 Ictalurus punctatus 359 I-domain 549 eIF-2 kinases 41, 42 IgE receptor (FcERI) 579 IgG 488, 500 Immune precipitation 580 lmmunoelectrophoresis 435 crossed-affinity 435 lmmunofluorescence microscopy 437 Immunogens 577 Immunoglobulin 407, 426 Immunoglobulin A 422 Immunoglobulin (Ig) superfamily 367, 5 10, 517-528 Immunohistochemical localization 426 Immunological defence 65 Immunoprecipitation 575, 576 Immunosequencing 575, 576 Indoleacetic acid 4 I3 Infantile sialic acid storage disease 370 Infection 476, 495, 577 Inflammation 173, 525, 572 Influenza virus 477, 495 type C 368 N-9-0-acetylneuraminic acid staining 327 Ingestion 500 Insect transferrin 208 PI Integrin family 196 Integrin receptors 546 Integrin superfamily 365 Integrin XIIbb3 195 Integrins 501, 510, 538, 544-559 alPl 538, 549, 550 a2Dl 538, 549, 550 a3PI 538, 541, 549, 550 cell-cell adhesion 550 a4BI 530, 549, 555, 556 asBl 530, 547, 549-551, 553-556 a& 538, 549, 550, 557

ad34

538 559 allbiP3 547, 551 allbf’3 552 aLD2 549-551, 556, 557 aMP2 549-551, 553, 556 avP1 550 avD3 545, 549, 550, 553, 556 axP2 550, 551 activation 547-549 assembly 547, 551 biosynthesis 550-554 carbohydrate 554-559 cation binding 547, 552 cell-cell adhesion 550 glycosylation 547, 553 I-domain 547 LFA-I 525-527 ligand binding 549, 550 Mac-I 525 P150/95 525 a-subunit 546, 547 P-subunit 546, 547 phosphorylation 549 VLA4 527, 528 Intercellular junctions 550 Interleukin 443, 465 Intermediate filaments 41 Interstitial matrix 528 Intestinal glycoproteins 580 Intestinal lactotransferrin receptor 21 7 Intramolecular migration 274 Iris germanica 405, 460, 468 Irisxhollandica 405, 406, 460, 468 Iron 204 Iron accumulation brain 179 diabetes 179 liver 179 pancreas 179 retinal problems 179 Iron-binding sites 208 Iron metabolism 179 Iron nutrition 216, 219 Iron transport 2 17 lsolectins 414, 415, 417, 419, 4 2 1 4 2 4 p-lsothiocyanato-phenyl glycosides 605 a7PI

J1 glycoprotein 541 Jacalin 407, 422 Jack bean, see Canaualia ensiformis Jimson weed, see Datura stramonium

K562 cell line 555 K chain 539

63 7 Kdn (2-Keto-3-deoxynononic acid) 163, 249 Kdn 7,9-0-Ac 165 Kdn 7-0-Ac 165 Kdn 9-0-Ac 165 Kdn-ase Misgurnus fossilis 332 Sphingobacterium multiuorum 332 Kdn-gp 0-glycans 154-1 56 Kdn-transferase 329 Oncorhynchus mykiss 329 2,8-Kdn-transferase 329 metabolism 329 occurrence 329 Axolotl mexicanum 329 Pleurodeles waltlii 329 Xenopus lueuis 329 Keratan, biosynthesis I0 Keratan sulfate 2, 7, 16, 18, 59 degradation in lysosomes 14 structure 6 Keratinocytes 550 Keratins 514 Kidney bean, see Phaseolus uulgaris Kifunensine 557 Kinases 35, 43 Klebsiella pneumoniue 482 Kupffer cells 357 LI, cell adhesion glycoprotein 58, 62, 520-524 L2 60, 64, 65 L2 monoclonal antibody 64 L2/HNK-I 554 L2/HNK-I carbohydrate 524, 525 L3 588 L3 monoclonal antibody 58 L5 monoclonal antibody 65 LIINg-CAM 588, 591, 594, 595 Laburnum alpinum 419, 428 Laburnum anugyroides 428 Lactalbumin 462 Luctarius 406, 467 Lactarius deliciosus 460 Lacturius deterrimus 460 Lactobionate 406 Lactoferrin 206 Lactonization 262 Lactose 405, 406, 487, 488 sialyl-3 480 sialyl(a2-3) 479 sialyl(a2-6) 479 Lactosiderophilin, see Lactoferrin Lactosylceramide 485, 490, 491 Lactotransferrin 206, 207, 213, 216, 217, 232 antitumor effects 2 I8

aposerotransferrin 206 bovine 231 goat 206 human 231, 233 human milk 230 human polymorphonuclear leukocytes 222 human serotransferrin 223 immune response 217 i n biological fluids bile 207 pancreatic juice 207 saliva 207 seminal plasma 207 synovial 207 tears 207 inflammatory events 2 17 interleukin synthesis 21 8 mare 206 maturation of T- and B-cells 21 8 monkey 206 mouse 206 natural killer cell activation 218 neutrophilic leukocytes 207 rabbit 206 role 216 sow 206 Lactotransferrin glycans 222 Lactotransferrin receptors 21 8 bacteria 218 Caco-2 enterocyte cell 218 enterocytes 2 18 epithal mammary cell 2 18 hepatocytes 2 18 human 218 human alveolar macrophages 2 18 human HT29 218 human monocytes 2 I8 human neutrophils 218 human platelets 21 8 Jurkat T cell 2 18 megakaryocytes 21 8 monkey 218 mouse 218 rabbit 218 9-0-Lactylated sialic acid, biosynthesis 328 Laminaribiose 418 Laminin 426, 509, 520, 528, 535-541, 543, 549, 550, 554, 557-559, 590, 591 glycosylation ' 539-541 heparin binding 538 S-Laminin 539 Lathyrus 4 12, 469 Lathyrus ochrus 232, 410 Lathyrus odorutus 428, 456

638 Lathyrus satiuus 405, 457 Lathyrus tingitanus 405, 457 Lattice structures 4 1 1 Lecithin:cholesterol acyltransferase (LCAT) 178 Lectin affinity chromatography 21 0 Lectin-based reagents 427 Lectin-like motifs 58 1 Lectin-resistant 49 1 Lectinophagocytosis 355, 476, 500, 501 Lectins 56, 174-177, 179, 180, 184, 185, 189, 197, 578, 579 affinity adsorption 435 affinity adsorptiodchromatography 427, 43 1434 affinity chromatography 404, 413, 419, 420, 423, 424, 436, 476, 491 affinity electrophoresis 435, 436 affinity precipitation 423 agglutination analysis 430, 438441 animal 490 Bauhinia purpurea 487 biotin-labeled 428, 429, 434 blot analysis 427, 43 I , 4 3 4 4 3 7 carbohydrate-binding specificity 407, 409, 41 1, 415, 417-419 cerebellar soluble lectin (CSL) 588, 589, 594, 595 concanavalin A (Con A), see Concanavalin A Datura stramonium 539 digoxigenin-labeled 434, 437 Entamoeba histolytica 476 enzyme-complexed 428, 429, 434 fluorochrome-labeled 428, 429 fungal 476 galactose 476 galactose-binding 56 Giardia lamblia 476 glucose 483 gold-conjugated 428, 429 Grrffonia simplicifolia 559 Grrffonia simplicijolia B4 isolectin 539 immobilized 428, 429 intracellular 486 lectin RI 592, 593 Limaxflavus 360 mannose 483 mannose-6-phosphate 476 microbial 475-501 non-fimbrial 476 Phaseolus uulgaris 539 plant 403470, 483, 490 isolation 403, 404 purification 403, 404 potato lectin 539

protozoal 476 Salmonella 48 1 sialic-acid 476 sialic-acid-binding invertebrates 362 plants 362 sialic-acid-specific Limaxflavus 360 use in microscopy 444 Legumes 404,407, 408 Leguminosae 411, 417 Lens culinaris 405, 410412, 414, 428, 432, 436, 457, 464, 469 Lentil, see Lens culinaris Leucine-rich domain 194 a2-Leucine-rich glycoprotein 188 leucine-rich repeat 188 plasma concentration 188 porcine ribonuclease inhibitor 188 Leucine-rich repeats 194 Leucine-rich motifs 17 Leucojum 405, 468 Leucojum aesiiuum 456, 464 Leucojum uernum 456 Leukemia 371 Leukoagglutinin 404, 424 Leukocyte adhesion 557 Leukocyte extravasation 582 Leukocyte tethering and rolling 582 Leukocytes 173, 175, 441,487 Lewis b 409 Lewis x in hemocyanins 140 Lewis x-type glycans 175 Lewisb antigen 489 Ligands 404, 571-573, 577, 578, 581, 582 Ligands, polarized 578 Ligands, saccharide 572 Liliaceae 407, 415, 416 Lima bean, see Phaseolus lunatus Limaxflaous 440 Limulin 425 Limulus polyphemus 126, 425 Lipid-linked oligosaccharide 93 Lipid transport I78 Lipopeptidophosphoglygans (LPPG) 7 1 Lipophosphoglycans (LPG) 70 Lipopolysaccharide 359 Listera ovata 407, 408, 41 5, 41 6, 456, 464, 468 Locomotion of cells 509 Lotus tetragonolobus 405, 406, 423, 428, 461, 466, 469 LPG (lipophosphoglycans) 70 LPPG (lipopeptidophosphoglygans) 71 Luffa acutangula 405, 457, 470

639 Lumican 17, 18 Lung infections 486 Lyase 347 active center 349 catalysis 348 cloning 349 Clostridium perfringens 348 Escherichia coli 349 inhibitors 348 ‘H NMR spectroscopy 349 pig kidney 348 Schiff-base mechanism 348 sialic acid analysis 349 substrate specificity 348 X-ray crystallography 349 Lycopersicon esculenium 406, 409, 417, 428, 443, 444, 458, 464, 470 Lymnaea siagnalis hemocyanin glycans biosynthesis 137-140 Lymphocytes 173, 355, 424, 441, 443 B-Lymphocytes 441, 465 T-Lymphocytes 441,465, 491, 576, 579 Lymphomas 207 Lysosomal enzymes 93, 94, 104, 106 Lysosomes 96, 113 Lysozyme 207 Maackia amurensis 404, 405, 409, 425, 426, 428, 432, 437, 440, 444, 462, 466, 469 Maclura pomifera 405-408, 421, 428, 460, 465, 468 a-Macroglobulin 41 5 a2-Macroglobulin 188, I89 alcoholism I89 Alzheimer’s disease 189 autoimmune diseases 189 autoimmunity 189 carcinoembryonic antigen 189 cerebrospinal fluid 189 non-acute-phase protein 189 plasma concentration 188 polymyositis 189 rheumatoid arthritis 189 schizophrenia 189 sclerodenna 189 serum 189 Sjogren’s syndrome 189 systemic lupus erythematosus 189 Macrophages 355, 356, 490, 491, 500, 501 Macrotyloma axillare 406, 460, 469 MAG 523, 524, 594, 595 Magnetic beads 442 Malaria 476, 525 Malignant cells 582

Malignant transformation 372 Malignant tumors 370, 371 colon 371 human colorectal tumor 371 thyroid gland carcinomas 371 Man-O-Ser/Thr 60 Man(aI-3)Man(f314)GlcNAc 481 Mannan 405, 462 a-Mannan 415 a-o-Mannan 464 Mannitol 59 Mannose 405, 408, 409, 414416, 456, 464, 489, 500 adhesion 481 o-Mannose 413, 415 0-Mannose 59 0-Mannose linked glycans 59, 60 0-Mannose linked oligosaccharides 59, 60, 65 Mannose-binding lectins 415, 416 Mannose/glucose-binding lectins 41 I Mannose-inhibitable 482 Mannose-6-phosphate (Man-6-P)93, 495 cation-dependent Man-6-P receptor 94 cation-independent Man-6-P receptor 93 methyl transferase 94 M-6-P phosphate methyltransferase 95 a-Mannosidases 92, 104 a-Mannosides aromatic 48 1 f3-Mannosidosis 259, 315 Man-640, 104 Marchariia polymorpha 462, 467 Mass spectrometry 178 Mast cells 21 Matrix glycoproteins 528-545 MDCK cells 516 Meconium 177, 580 Melanoma 207, 326 Melanoma cells 555, 557, 559 human 549 Melanotransferrin 207, 208, 214, 219 aposerotransferrin 206 in cells capillary endothelium 207 foetal intestinal cells 207 humain brain 207 intestinal epithelial cells 207 liver cells 207 microglia of Alzheimer’s disease 207 placenta 207 sweat gland ducts 207 umbilical cord 207 uptake of iron 2 19 Melibionate 406 Melibiose 421

640 Meningitis 62, 66, 484 Meningococci, group B 63 Merosin 539 Merozoites 494 Metal-binding site 41 3 Metalloproteinases, degradation of proteoglycans cathepsins I2 matrix proteases 12 trypsin-like enzymes 12 Metalloproteins 41 1, 414, 4 1 9 4 2 1 Metastasis 353, 371, 439, 582 Metastatic potential 555 Methanolysis 265 8-0-Methylated sialic acids, biosynthesis 328 Asterias rubens 328 echinodermata 328 Methyl 2,3-di-O-methyl a-D-ghcopyranoside 414 3-0-Methylgalactose 123 4-0-Methylgalactose 123 4-0-Methyl-D-galactose 132 Methyl fi-galactoside 41 1 Methyl a-D-glucopyranoside 432 3-0-Methylglucosamine 405 3-0-Methyl glucose 414 Methyl a-glucoside 414 Methyl p-lactoside 41 I Methyl a-mannopyranoside 414, 415 Methyl a-D-mannopyranoside 41 3, 43 I , 432 3-0-Methylmannose 123 Methyl a-mannoside 48 I , 501 Methyltransferase 140 4-Methylumbelliferyl a-mannoside 48 I MHC complex 517 Microbial adhesins 475 Microbial pathogens 58 I Microcalorimetry 409 Microglial cells 465 Microheterogeneity 407 Microorganisms 439 Microvilli 578 Milk 218 Mimicry 65 Mitogen 424 Mollusc hemocyanin 125, 129-1 36 Mollusc hemocyanin glycans, synthetic oligosaccharides I37 Moluccella laeuis 405, 460, 465, 470 Momordica charantia 405, 406, 460, 470 Monensin 532, 556 Monoclonal antibodics 571, 576 hybridoma-derived 571, 574 sequence-specific 572 Monocot storage tissue 407

Monocytes 173, 581 Monooxygenase 324 Monosaccharides 409 Moraceae 407, 421 Mouse 217 Mucin glycopeptides 576 Mucin-type glycoproteins 576 Mucin-type 0-glycans 25 1 Mucins 327, 354, 406, 462, 476, 486, 492, 493, 578 hog gastric 424 intestinal 343 porcine stomach 424 viscosity 354 Mucopolysaccharides 2 Mucopolysaccharidoses 2, 14 Mucosa 487 Mucuna deeringiana 460, 469 Mucus 484, 485, 489 Multi-antennary glycans 56 Multigene families 407 Multivalent 493 Murine IgM 415 Mycoplasma pneumoniae 485, 486, 573, 577 Myelin 587, 593-596 Myelin basic protein (MBP) 594 Myelin glycoproteins 523, 524 MAG 523, 524 PO 523, 524 Myelination 556 Myeloma 443 Myogenesis 528 Myotendinous antigen 541 Myxobacteria 488 Myxococcus xanthus 488 N-asparagine-linked glycopeptide 414 N-asparagine-linked glycoprotein 413 Nagana disease 341 Narcissus 405, 468 Narcissus lobularis 456, 464 Narcissus pseudonarcissus 408, 4 15, 4 16, 428, 456 Natural killer cells 21, 443 NIE-I 15 neuroblastoma cells 60 Nectadrin 589 Neisseria gonorrhoeae 341, 464, 466 sialyltransferase 359 virulence 359 NEM (N-ethylmaleimide) 75 Nematode 492 Neoglycoconjugates 601, 616 antigenic determinants 6 I6 ceramide glycanase, American leech 617

64 1 diagnosis, Mycobacteria 616 ganglioside receptors 61 6 glycohydrolase 617 glycolipids 616, 617 glycosphingolipid 617 transglycosylation 61 7 Neoglycoenzymes 615 alkaline phosphatase 615 chemiluminescent substrates 615 cytochemistry 6 I5 horse radish peroxidase 615 lectins 615 solid-phase assays 615 Neoglycolipid technology 582 Neoglycolipids 576, 601 Neoglycopolymers 601 N-CAM (neural cell adhesion molecule), see under Cell adhesion molecules Neoglycoproteins 439, 493, 601-618 biomedical applications 61 5 antisense nucleotide 615 DNA, targeted delivery 615 immunogenicity 61 5 technetium 6 15 tumor diagnosis 615 vaccines 615 bovine serum albumin (BSA) 601, 614, 616 carbohydrate-binding proteins (lectins) 6 I4 carbohydrate-protein interaction 61 4, 61 5 hydrogen bonding 6 I6 hydrophobic interactions 61 6 chicken hepatocytes 6 I4 cytochemical procedures 6 14 flow cytometry 61 5 fluorescent labeling 61 5 L-Fuc-BSA 61 4 Gal/GalNAc-BSA 614 GlcNAc-BSA 6 I4 glycoside cluster effect 61 4-6 I6 hepatic lectins 616 lung macrophages 6 14 malignancy 61 5 mammalian liver cells 614 Man-BSA 614 neoglycoenzymes 61 5 receptor 61 6 L-selectin 6 I6 sialyl LeX 616 Nerve growth factor 61, 62, 591 Nervous tissue cell adhesion 587-597 Nervous tissue glycoproteins 55-66 Networks and lattices, lectin-mediated 579 NeuSAc 612 conformation 309

donors, organic synthesis 300 mutarotation 284 synthesis 349 NeuAc-N-acetyllactosamine 491 NeuAc(a2-3)Gal 478, 494 NeuAc(a2-6)Gal 478 NeuAc(a3-6)Gal 494 NeuAc(a2-3)Gal(fl1-3)GalNAcchains 485 NeuAc(a2-3)Gal(P 1-3)[NeuAc(a2-6)]GalNAc 488 NeuAc(a2-3)Gal~3[NeuAc(a2-6)]GalNAc 479 NeuAc(a2-8)NeuAc 484 Neu2en5Ac, transition-state analogue 344 NeuSGc, presence in normal human tissue 251 NeuCc(a2-3)Gal 484 NeuCc(a2-3)Gal(~I4)GlcfiI-Cer 484 Neural cell adhesion 58, 63 Neural crest 588 Neuraminic acid 244 Neuraminidase 437 Neurite outgrowth 523, 524, 539, 557, 559, 591 Neuroblastoma cells, NIE-I 15 60 Neurocan 7, 16, 17, 60 Neurofilament assembly 42 Neurofilaments 41 Neuromuscular junctions 513, 539 Neuron migration 587, 589 Neuronal adhesions 524 Neuronal cell adhesion 539 Neuronal differentiation 6 1 Neutrophilic leukocytes 217 Neutrophils 173, 490 granulocyte polyrnorphonuclear leucocyte 48 1 NG2 21 Ng glycoprotein 52 I , 522 Nidogenientactin 537, 538, 541 laminin binding 541 NILE (nerve growth factor-inducible large external glycoprotein) 62 P-Nitrobenzyl a-methyl glucoside 4 15 p-Nitrophenyl glycosides 605 P-Nitrophenyl a-mannoside 48 I N-linked carbohydrate chains 25 I N-linked GlcNAc 47, 48 N-linked glycans 56-58, 62 N-linked glycoproteins 41 0 nucleus 49, 50 NMR spectroscopy I80 I3C-NMR 611 'H-NMR spectroscopy 252 H NMR structural-reporter-group concept 288 Nodes of Ranvier 593 Non-fimbrial 481 Non-heme iron 205

'

642 Non-opsonic phagocytosis 500 Non-secretors 576 Non-polar binding site 413 Non-polar interactions 410 N-terminal signal peptide 407 Nuclear magnetic resonance 409, 41 1 Nuclear pore assembly 39 Nuclear pore proteins (nucleoporins) 38, 39 Nuclear proteins 38 Nucleoporins, see Nuclear pore proteins a-2,05-linkage 267 9-0-Ac-NeuAc 494 Oligodendrocytes 523, 593, 594 Oligomannose 432 complex unit 477 Oligosaccharide backbones 571, 572 Oligosaccharide-based therapeutic substances 582 Oligosaccharide microheterogeneity 575 Oligosaccharide sequences 571-582 Oligosaccharides 477, 486, 571, 575, 578 biological roles 571, 573, 582 chemically synthesized 574 complex type 424 di- or triantennary type 179 milk 256 N-asparagine-linked 424 0-linked N-acetylglucosamine 3 3 4 2 0-linked fucose, 0-Fuc 46 0-linked GlcNAc 47, 49 0-linked glycans 56, 58, 59, 432 0-linked glycoproteins 465 0-linked mannose 47, 56 0-man 43 0-linked oligosaccharides 58, 59, 62, 576 Oncogenes 40, 372 Onobtychis uiciifolia 405, 457, 469 Oocytes 357 Opsonic activity 501 Opsonins 500 Opsonization 58 I Opsonophagocytosis 500 Optic tectum 466 Oral bacteria 488 Orange peel fungus, see Aleuria auraniia Orchidaceae 407, 415, 416 Orosomucoid 174, 462 Oryza saiiva 405, 417, 428, 458, 467 Osage orange, see Maclura pornifera Osteopontin 545, 549 Ouchterlony double diffusion 430 Ovalbumin 223, 436 Oviducal secretions in amphibians 164

Ovomucoid 223, 462 Ovotransferrin 205, 208, 2 18, 232 aposerotransferrin 206 Ovotransferrin glycans 223 Oxidative burst 481 Oxygen-transporting proteins 123 PO 523, 524, 595, 596 P31 589, 592-597 p43 autoantigen 40 p62 39 Ph7 41, 42 p97 aposerotransferrin 206 PA-I 486, 487 PA-I1 486, 489 Panning 442 Papilionoideae 407 PAPS (3’-Phosphoadenylyl 5’-phosphosulfate) 2, 8 Parafusin 43, 44 Parasites 173, 493 Paroxysmal nocturnal hemoglobinuria 78, 79 CD59 1 9 decay-accelerating factor (DAF) 78 Pathogenesis 65 PC12 pheochromocytoma cells 61 Peanut, see Arachis hypogaea Peanut agglutinin 487 Peptic ulcers 485 Peptide:N-glycanase (PNGase) in fish I57 Peptide:N-glycanase F 485 Pericellular matrix 521 Perinuclear vesicles 190 Periodate oxidation 486 Peritoneum 501 Peritoneum cavity 501 Perivitelline fluid 152 osmoregulation 152 Perlecan 8, 18, 19 Perseau americana 428 Pertussis toxin 489, 490 Phago-lysosomes 97 Phagocyte 501 Phagocytosis 108, 180, 475, 487, 500 blood cells 355 Pharmaceutical industry 572 Phase variation 501 Phaseoleae 407 Phaseolus 469 Phaseolus acutifolius var. latifolius 405 Phaseolus coccineus 428, 462 Phaseolus coccineus var. aluba 406 Phaseolus lunaius 406, 408, 412, 419, 420, 429, 460

643 Phaseolus mungo 405, 460 Phaseolus oulgaris 138, 404, 405, 412, 424, 425, 429, 43 1, 432, 440, 441, 462, 466 Phenyl (3-glycoside 4 I3 Phosphacan 60 Phosphatases 35, 43 Phosphatidylethanolamine 76 Phosphatidylinositol 19, 20, 22 Phosphatidylinositol-glycan 492 Phosphatidylinositol-specific phospholipase C (PI-PLC) 101 3’-Phosphoadenylyl 5’-phosphosulfate (PAPS) 2, 8, 95 Phosphoethanolamine transferase inhibitors 76 Phosphoglucomutase 43, 44 Phosphoglycosylation 97 Phospholipases 178 C 589, 597 D 589 Phosphomannan 4 I4 Phosphomethyldiester 94 Phosphoproteins 34 Phosphorylation 34, 35, 40-42 Photoactivatable group 61 1 Phototaxis 90 Physiological modifications 227 Phytolacca americana 405, 458, 468 PI-PLC (phosphatidyl-inositol-specific phospholipase C) 101, 110 Piglets 486 meningitis 484 Pilins 480 PineNia terneutu 405, 462, 466, 468 Pisum satioum 405, 407, 4 1 W I 2 , 414, 415, 429, 457, 469 Plakoglobin 5 14 Plant 492 Plasma 173-191 Plasma kallikrein 180 Plasminogen 189 fibrin 189 plasma concentration 189 plasmin 189 thrombus 189 Plasmodium 494 Plasmodium falcipurum 476, 4 9 3 4 9 5 N-acetylglucosamine 495 affinity chromatography 495 erythrocytes 493, 494 glycophorin 494 malaria 493 merozoites 493 neoglycoprotein 495 sialic acid 495

sialidase 493, 494 surface antigen Pf200 493 Platelet endothelial cell adhesion molecule-I, see PECAM-I under Cell adhesion molecules Platelets 173, 192-1 98 erythrocytes 192 haemostasis 192 subendothelium 192 thrombus 192 Pleurodeles waltl 166 Pleurotus oslreatus 405, 460, 467 Pneumococcal polysaccharide 430 Pneumocystis carinii 175 Polar interactions 410 Poly-N-acetyllactosamine 485, 486, 575, 579 Poly-N-acetyllactosamine glycans 60-62, 65, 484 in erythrocytes 60 in teratocarcinoma cells 60 Polyacrylamide 406 Polyacrylamide gel electrophoresis 435 Poly-N-acy lneuraminic-acid-containing glycoproteins 265 Polyagglutinability 41 8, 42 1 Polyagglutinating erythrocytes 465 Polyclonal antibodies 41 5 Polyglycosyl ceramides 484 a-2,S-Poly Kdn 329 Polylactosamine 6 0 4 2 , 418, 525, 532, 535, 539, 554, 559 Poly-a-2,S-linked N-acetylneuraminic acid 62, 63 antibody specificity 63 Drosophila 63 elasmobranchs 63 Escherichia coli KI 62 group B meningococci 62 sahnonid fish eggs 62 sodium channel glycoprotein 63 Polymorphonuclear leukocytes 487 Polyoma, see Virus Polysaccharides 58 I , 609, 61 0 aldehyde 610 Haemophilus injuenzae 610 hyaluronic acid 609 periodate oxidation 609, 6 10 reductive amination 609 vaccine 610 Polysialic acid 62, 63, 65, 259, 267, 369, 370, 519-52 1 Drosophila melanogaster 369 Polysialoglycoprotein (PSGP) 144, 145 a-2,6-ST 148 a-2.8-ST 148

644 Polysialoglycoprotein (PSGP) (contii) a-2,8-p0lyST 148 Kdn-transferase 149 rainbow trout 144 Salmonidue 144 Oncorhynchus 144 Sulmo 144 Sulvelinus 144 sialyltransferase I48 Polysialosyl 588 Polysialyltransferase 320 Polyvinylidene difluoride membranes 434 Porcine pancreas lipase 29 1 Post-translational cleavage 41 2 Post-translational ligation 41 2 Potato, see Solunum tuberosum Precipitation analysis 427, 429-43 1 Precipitation curves 41 6 Precipitin 409, 435 Prevotellu luescheii 488 Primary structure 408 Prolyl hydroxylase 434 Promyelocytic cells 576 N-Propanoylneuraminic acid 3 I 1 Protease inhibitor 177 Protease N 291 Protein kinases 35 Proteinase I 97 al -Proteinase inhibitor 189, 190 Proteins PI protein 485 protein C 190 plasma concentration 190 von Willebrand factor 190 protein S 190 sialic acid-binding, in vertebrates 365 synthesis inhibition 422 inhibitor 465 translocation 38 Proteoglycans 1-25. 46, 65, 509, 5 10, 520, 52 I , 528, 530, 537, 538, 541, 543, 555 biosynthesis 7-12 glucuronosyl transferase 10 cell surface 19-2 I core protein 2 acylation 7 aggrecan 7, 16, 17 amphiglycan 19 betaglycan 2 1 biglycan 17, 18 biosynthesis 7 brevican 16, 17 cerebroglycan 20

decorin 7, 17, 18 fibroglycan 19 fibromodulin 17, 18 glypican 20 lumican 17, 18 neurocan 7, 16, 17 NG2 21 perlecan 18, 19 phosphorylation 7 ryudocan 19 serglycin 7, 21 syndecan 7, 19, 20 versican 16, 17 core protein proteolytic trimming aggrecab 7 decorin 7 neurocan 7 serglycin 7 core protein structure 3, 15-22 core protein xylosylation I0 core protein-glycosaminoglycan linkage region 2 degradation and turnover 12-1 4 degradation by metalloproteases I2 function 15-22 intracellular 2 I , 22 linkage oligosaccharide 3 linkage region 6, 7 mast cells, basophils 21 matrix 16 N- and/or 0-linked oligosaccharides 8 part-time 17, 19, 21, 22 phosphatidylinositol 19, 20, 22 structure 3-7 Prothrombin 190 fibrin 190 fibrinogen 190 liver 190 thrombin 190 Protozoa 439, 4 9 2 4 9 5 , 499 adhesion 499, 500 amoebae 500 antibodies 500 complement 500 Enfamueba hisfulyfica 499, 500 epithelial cells 500 GaliGalNAc 499, 500 gerbils 499 hepatocytes 500 immunization 499 pathogenicity 500 protection 499, 500 trophozoites 499 virulence 500

645 Psathyrella velutina 406, 458, 464, 467 Pseudomonas aeruginosa 476, 486, 489 Psophocarpus tetragonolobus 405, 406, 4 10, 42 1. 422, 429, 460, 469 Pulmonary disease 359 Purkinje cells 592, 593 Pyruvate 274 Rabbit 178, 217 Rabbit serotransferrin 2 10 Ramsons, see Allium ursinum Rana temporaria 17 I Random coil 404 ras oncogenes 37 I Rat a-fetoprotein 185 Rat nuclear pore proteins 39 Ricinus communis agglutinin (RCA-I) 539 67 kDa Receptor 539 Receptor activities, regulation (fine tuning) 579 Receptors 66, 581, 582 adhesive 507-559 Red cells 575, 576 Red protein, see Lactoferrin Reductive amination 602-604, 61 0 o-aldehydic group 603 amino borane complexes 603 (3-elimination 604 heterobifunctional reagent 604 periodate oxidation 604 pyridine borane 603 sodium cyanoborohydride (NaCNBH3) 602 sugar alcohol (alditols) 604 Regulation of glycosylation 573 Regulatory proteins 34 REMI (restriction enzyme mediated integration) I I6 Renal medulla 501 RER in glycosaminoglycan biosynthesis 9 Respiratory tract 485 Restriction enzyme mediated integration, see REMI RGD sequence 530, 531, 538, 541, 543, 544, 549, 556 Rhesus monkey 2 17 Rheumatoid arthritis 175 Rhinovirus 525 Rhizobia 488 Rhizobiaceaea 489 Rhizobium lupinii 489 Rhizoctonia solani 492 Ribose 409 Ribosomal RNA 422 Rice, see Oryza satiua Ricin 422, 460

Ricin-resistant cells 535, 553, 555 Ricinus communis 405, 406, 41 0, 422, 429, 430, 434, 435, 440, 460, 465, 469 RNA polymerase I1 39, 40 Robinia pseudoacacia 405, 429, 460, 462, 465, 466, 469 ROESY (rotating-frame nuclear Overhauser enhancement spectroscopy) I37 Rye, see Secale cereale Ryudocan 19 Sainfoin, see Onobrychis viciifolia Saliva 485, 486 Salivary glycoprotein 487, 488 Salla disease 370 Salmonella 475 Salt fractionation 41 9 Sambucus nigra 405, 425, 426, 429, 437, 440, 444, 460, 462, 465, 466, 470 Sambucus sieboldiana 41 I , 444 Sarothamnus welwitschii 405, 461, 469 Schwann cells 523, 524, 593-595 Sclevotinia 405, 465, 467 Sclerotinia miyabaena 46 I Sclerotiniaceae 492 Sclerotium rolfii 462, 467, 492 SDS-PAGE 434, 437 Secak cereale 405, 406, 41 7, 458, 467 Secretors 576 Secretory IgA 207 Selectins 56, 65, 175, 364, 365, 371, 490, 582 addressin 364 binding specificity 364, 365 distribution 365 E-selectin 364, 490 L-selectin 364 P-selectin 364, 490 roles 364 Selenate 97 Senescence 444 Sephadex 405 Sepharose 405, 435 Sepsis 484 Serglycin 3, 6, 7, 2 1, 22 Serial lectin affinity chromatography 43 I , 432 Serine 418 Serology 438 Serotransferrin 179, 206, 208, 214, 232 aposerotransferrin 206 chicken embryogenesis 228 disease 230 glycans 219 glycovariants 21 1

646 Serotransferrin - glycovariants (contii) Tf-I 227 Tf-11 227 Tf-111 227 human 227, 230 alcoholic cirrhosis 230 CDG syndromes 231 CDG-type I syndrome 231 CDG-type I1 syndromes 23 1 in HEMPAS 231 liver diseases 230 microheterogeneity 227 viral hepatitis 230 placental receptor 227 pregnancy 227 rabbit 231, 233 receptor 216, 218 N-glycans 21 6 role in iron transport 214 Serotransferrin-to-cell cycle 2 16 Serpin-type protease inhibitors 544 Serum 175 Shallot, see Allium ascalonicum Shed-Acute-Phase-Antigen (SAPA) 339 fi-Sheet 404 Shigellaflexneri 500 Sia-gp 154 Sialate 0-acetylesterase 329-33 1 occurrence bovine brain 330 enteric bacteria 330 human erythrocytes 330 influenza A virus 330 influenza B virus 330 influenza C virus 330 rat liver 330 substrate specificity 330 Sialate 9-0-acetylesterase 263, 273, 274 Sialate 4-0-acetyltransferase 326 Sialate 9(7)-O-acetyltransferase 326 Sialate 0-methyltransferase 328 Sialate-pyruvate lyase 347-349 Sialic-acid-binding lectins 425, 426 Sialic-acid-binding protein, endometrium 368 Sialic acid chemistry 289-309 free sialic acids 290-292 Kdn, organic synthesis 290 NeuSAc, organic synthesis 290 organic chemical synthesis 290 Sialic-acid-containing elements in N-glycoproteins 252 in 0-glycoproteins 252 Sialic-acid-containing microbial polysaccharides 260

Sialic acid permease 349-352 Sialic-acid-recognizing proteins 360 Sialic-acid-recognizing receptor 364 Sialic-acid-specific lectin, Limaxflauus 360 Sialic acids 56, 58, 59, 62, 175, 179, 243-372, 417, 424, 440, 444, 462, 466, 476480, 483436, 488, 493-495 see also N-Acetylneuraminic acid 0-acetyl 146 0-acetyl migration 264, 284 U-acetylated 327, 343 biological significance 325 9-0-acetylation 367 0-acetylation patterns 249 acid hydrolysis 264 0-acyl migration 282 aldolase 259 aldolase-catalyzed 293 anti-proteolytic effect 353 aspartylglycosaminuria 258 autohydrolysis 282 bacterial toxins, binding 362 biological recognition sites 360-368 masking 354-359 biosynthesis 3 1 1-329 biosynthesis inhibitors 3 I 1 bovine colostrum 258 bovine submandibular gland 262 catabolism 329-352 0-acetylesterases 329 lyases 329 sialic acid permeases 329 sialidases 329 C-glycosides 309 colominic acid 259, 262 colorimetry 268, 269 Aminoff method 268 “Bial” reaction 268 fluonmetric assays 269 Hestrin assay 269 ninhydrin assay 269 orcinol/Fe3+/HCIassay 268 periodic acid/thiobarbituric acid assay 268 Warren method 268 complement reactivity 358 conformational aspects 309, 3 10 de-N-acetylated 330 de-0-acylation 282 edible birds nest substance 259 egg yolk 259 egg-yolk membranes 259 enzymatic hydrolysis 264 enzymatic modification 320-329

647 erythrocyte invasion 363 Plasmodium falciparum 363 fast atom bombardment mass spectrometry 28 1, 282 feces 258 Fonsecaea pedrosi 354 galactose residues, masking 354 gas-liquid chromatography mass spectrometry 275-281 general physico-chemical effects 352-354 glycolipids 251, 253, 289 glycoproteins 25 1 glycosylphosphatidylinositol 259 glycosylphosphatidylinositol membrane anchors 25 1 hemagglutinin binding 363 hen's egg chalaza 259 high-performance capillary electrophoresis 275 high-performance liquid chromatography 270275 0-acylated sialic acids 272 anion exchange 271 anion-exchange chromatography 270 cellulose chromatography 270 fluorescence labelling 271 reversed-phase HPLC 27L histochemistry 262 homo- and heteropolysaccharides 25 1 human (pregnancy) urine 258 immune reactivity 358, 368 interaction with hemagglutinin 496 isolation and analysis 264-289 Kdn-containing glycoproteins 253 lactonization 259 lectins specific for 263 Legionella pneumophila 262 (1ipo)polysaccharides 253, 289 0-mannosidosis 258 mass spectroscopy 245 medical significance 370-372 Alzheimer disease 370 galactosialidosis 370 leukemia 371 malignant tumors 370 Salla disease 370 sialidosis 370 sialuria 370 metabolism scheme 3 12 microbial polysaccharides 259 microwave hydrolysis 265 milk 253 milk oligosaccharides 289 naturally occurring 245

NMR spectroscopy 245, 251 ' H NMR 282-289 0-lactyl 146 oligosaccharides 25 1 organic chemical preparation 293 organic synthesis 259 pathobiochemical significance 352-372 pathogenic microorganisms and toxins, binding 361 permease 342 physiological significance 352-372 pK value 352 plant lectins specific for 263 Maackia amurensis 263 Sambucus nigra 263 poly-N-acylneuraminy I-containing glycoproteins 253 porcine submandibular gland 262 Pseudomonas aeruginosa 262 Pseudomonas juorescens ATCC 4927 1 262 radiolabelled 292 receptor functions 360-368 Salmonella arizonae 0 6 I 262 screening 262-264 Shigella boydii 262 sialate-pyruvate lyase 259, 348 reaction scheme 350 three-dimensional structure 35 1 sialic-acid-like monosaccharides 262 Sporothrix schenkii 354 sulfate 253 thin-layer chromatography 269, 270 urinary oligosaccharides 253 urine of sialuria patients 259 Vibrio alginolyticus 262 Vibrio cholerae 0 2 262 Vibrio salmonicida 262 Yersinia ruckeri 0 I 262 Sialidase 178, 265, 292, 321, 324, 331-347, 417, 421, 466,485, 487, 488, 491, 495, 500 Acanthamoeba 342 Aclinomyces viscosus 336 active sites 334 activity 273 amino acid sequence motif 333 antibodies 343 Arthrobacter ureafaciens 266, 267, 33 1 Asterias rubens 336 bacterial 284, 342-344 nutritive effect 342 virulence enhancing 342 Bacteroides fragilis 266, 342 cellular interactions, regulating 358 cloning 334

648 Sialidase (conrid) Clostridium perfringens 264, 266, 267, 284, 331, 333, 336, 337, 343 Clostridium septicum 337 Clostridium sordellii 333, 337 crystal structure 334 deficiency 370 Erysipelothrix rhusiopathiae 343 essential amino acids 336 eukaryotic 339-342 evolutionary origin 336 fowl plague virus 266 horizontal gene transfer 336 influenza A virus 333, 346 influenza A, virus 266 influenza B virus 250 inhibitor constants 344 inhibitors 344-347 tailor-made 347 Macrobdella deeora leech 250, 266, 332 Micromonospora oiridifaciens 336 Misgurnus fossilis 267 Newcastle disease virus 266 Pasteurella haemolytica 343 pathogenesis 344 pathophysiological significance 339-344 primary structures 333-337 Aspartate box 333 protozoa 337 rainbow trout 267 reaction mechanism 33 1 Salmonella typhimurium 266, 33 I , 333, 334, 336, 337 three-dimensional structure 335 sensor 264 Sphingobacterium multioorum 267 Streptococcus pneumoniae 343 Streptococcus sanguis 266 tertiary structure 335 Trichomonas foetus 343 Trichomonas uaginalis 343 Vibrio cholerae 266, 267, 33 I , 333, 336 viral 284, 344 influenza A virus 344 influenza B virus 344 orthomyxoviruses 344 paramyxoviruses 344 X-ray crystallography 346 trans-Sialidases 308, 332, 337-339 acceptor, procyclic acidic repetitive protein 340 antibodies 339 cloning 334 Endotrypanum 337

gene family 339 gene structures 338 inhibition 338 inhibitors 344-347 cruzin 347 pathophysiological significance 339-344 substrate specificity 338 Trypanosoma brucei 337 Trypanosoma congolense 337 Trypanosoma cruzi 333, 337 Trypanosoma rangeli 337 Trypanosoma oioar 337 Sialidase-treated erythrocytes, binding 356 Sialidase-treated lymphocytes, binding 356 Sialidase-treated thrombocytes, binding 356 Sialidosis 259, 370 Sialo compounds, autohydrolysis 249 Sialo-oligosaccharides 304-309 CMP-(l-Neu5Ac 305 CMP-Neu5,9Ac2 307 2-deoxy-2-halo-fi-Neu5Ac derivatives 304 dimeric Neu5Ac elements 305 a-I ,3-fucosyltransferase 306 P-galactosidase-catalyzed galactosylation 306 P-D-galactoside a-2,3-trans-sialidase Trypanosoma cruzi 307 (3- I ,4-galactosyltransferase 306 LeX 305 S-methyl a-glysosides 305 Neu5Ac donors 304 Neu2en5Ac methyl ester peracetylated 304 perbenzylated 304 S-phenyl a-glysosides 305 sialidase Arthrobacter ureafaciens 309 bacterial 309 Clostridium pevfingens 309 Vibrio cholerae 309 trans-sialidase 307 sialyl LeX 305 sialyl phosphitcs 305 a-2,3/6-sialyltransferases 305 a-2,3-sialyltransferase 306 a-2,6-sialyltransferase 307 trimeric Neu5Ac elements 305 Sialoadhesins 321, 325, 365 ligand specificity 367 sialylated glycan specificity 366 Sialobiology 244 Sialoglycoproteins 484 ' H NMR spectroscopy 288 Sialuria 251, 370 SialyI cholesterol, neurite outgrowth 369

649 Sialyl Ii antigens 577 Sialyl Led 364 Sialyl LeX 56, 60, 65, 175, 364, 579, 581 3’-Sialyl-LeX 581 Sialyl-oligomers 273 Sialylation 490 Sialyloligosaccharides 578 a-2,3-Sialyltransferase,glioma cells 3 15 a-2,6-Sialyltransferase Ehrlich ascites tumor cells 371 hepatoma 315 human colorectal tumor 371 human gastric epithelium carcinoma 371 Sialyltransferases 93, 289, 292, 3 14-320, 438, 466, 484, 485, 49 I acceptor specificity 3 14, 3 15 activity 273 assay 314 cDNA clones 320 cellular interactions, regulating 358 cloned 314 CMP-sialic acid 3 19 donor specificity 3 I4 expression 3 15 Golgi 319 inhibitors 3 18, 3 I9 nomenclature 3 I7 a-2,8-polysialyltransferase 329 sialyl motif 320 translocator 3 19 Siderophilin 205 Siderophilin, see Serotransferrin Signal sequences 77, 78 Signal transduction 82, 83, 355 diacylglycerol 83 inositol phosphate glycan (IPG) 83 phospholipases 83 Silica 406 Silica gel chromatography 576 Simplex virus, see under Viruses Site-directed mutagenesis 425 Skin infections 486 Small intestine 56 Smith degradation 574 Snail albumen gland 140 Snow drop, see Galanthus nioalis Sodium borotritide 576 Solanaceae 404, 409, 4 I7 Solanum tuberosum 405, 409, 417, 418, 429, 458, 464, 470 Solid-phase adsorbants 405, 406 Soluble glycoconjugates 427 Sophoru japonica 405, 406, 426, 429, 461, 465, 469

Sophorose 413, 418 L-Sorbose 4 I4 Soy bean agglutinin 61 I Soybean, see Gbcine ma\SPARC/osteonectin 544, 545 collagen binding 545 Spermatids 464, 466 Spermatozoa, fertility 357 Sphenostylus sfenocarpus 405, 461, 465, 469 Splenocytes 441 Spreading factor I2 SSEA (stage specific embryonic antigen) adhesion proteins 58 I macrophage endocytosis receptor 581 serum mannan-binding protein 58 1 anti-CD15 581 cell-cell adhesion 581 compaction 58 1 granulocytes 58 I hybridoma antibodies 581 leukocyte trafficking 58 I Lex-LeX interaction 581 myelomonocytic series 58 I poly-N-acetyllactosamiiie 58 I E-selectins 581 L-selectins 581 P-selectins 581 VEP8 and VEP9 581 SSEA-I (stage specific embryonic antigen I ) 571-582 Stage specific embryonic antigen, see SSEA Staggerer 593 Staphylococcus auieus 464 Stem cells 441 Stinging nettle, see Urfica dioica Streptococci 464466, 487 Sfreptococcus oralis 488 Strepfococcus sanguis 485, 487 Sfreptococcus suis 484 Stroma 405, 406 Structure diantennary 174, 176, 178 tetraantennary 174, I76 triantennary 174, 176, I78 Submaxillary glycoprotein 487 Submaxillary salivary glands 291 Substratum 509, 545, 549, 558 f l ~Subunit 546 0 4 Subunit, epithelial cells 547 SUGABASE 282 Sulfate esters 96 Sukdted LeX 579 3’-Sulfated LeX 582 epithelial cells 582

650 Sulfated N-glycans I75 Sulfatide 538 Sulfo-3-glucuronyl neolactohexaosyl ceramide 524 Sulfo-3-glucuronyl paragloboside 524 Swainsonine 92, 522, 532, 555, 556, 596 Synaptogenesis 587, 591-593 Syndecan 7, 19, 20 Synovial fluid I75 Synthetic glycopeptides 613, 614 9-fluorenylmethyloxycarbonyl(Fmoc) group 61 3 GalNAc-Ser/Thr 614 GlcNAc-Am linkage 613 pentafluorophenyl (PFP) group 61 3 solid-phase synthesis 614

T-antigen 407, 409 T cell 576 T cell activation 82, 83 T cell mutants 75, 76, 79 T cell receptor 551 Talin 41 Tam-Horsfall glycoprotein 483 Targets for cancer therapy 572 Terfiria occidentalis 405, 461 Tenascin 541-544 Teratocarcinoma cells 60, 510, 575 Tetraantennary glycans 56, 62, 196 Tetracarpidium conophorum 46 I , 465, 469 TGF(3, 551, 552 Thin-layer chromatography 270, 434 Thomsen-Friedenreich antigen 422 see also T-antigen Three-dimensional structure 477 glycans 232 transferrins 23 1 Thrombasthemia 552 Thrombin 549 Thrombocytes 355 Thrombospondin 198 cell adhesion 198 collagen 198 fibrinogen 198 heparin 198 platelets a-granules 198 Thy-1 glycoprotein 65, 593, 597 Thymocytes 441 Thymus 466 Thyroglobulin 425, 462 Tight junctions 507 Tissue-specific glycosylation 65

T-labelled CMP-NeuSAc 288 TLC analysis 264 Tn antigen 465 Tomato, see Lycopevsicon escirlentum Tooth, colonization by Actinomyces uiscosus 487 Toxin 491, 500, 501 Transcription factor Spl 39 Transcription factors 39, 40 Transferrin 190, 203-234, 425, 490 asialo serotransferrin 203 cancer 230 hepatocellular carcinoma 230 HepG2 cell line 230 evolution 208 gene duplication 208 glycan primary structures 2 I9 glycovariants 2 10 human cerebrospinal fluid 229 human seminal 229 lactoferrin 203 lactotransferrin 203, 219 see also Lactotransferrin antibacterial effect 217 bacterial effect 217 lactoferricin B 2 17 bovine 214 caprine 214 cow 222 goat 222 lysozyme 2 17 murine 214 porcine 214 secretory IgA 217 sheep 222 melanotransferrins 203 ovotransferrins 203 pentasialo serotransferrin 203 rat mammary-gland 229 rat milk 229 recombinant 229 sequence homology 208 serotransferrin 203 see also Serotransferrin fish 211, 219 hen 223 horse 21 I human 219 human cerebrospinal fluid 219 insects 21 1 mouse milk 219 pig 219 rat 219 serotransferrin receptor 21 5 tobacco hornworm 21 1

65 1 fi,-Transferrin 229 Transferrin glycan, physiopathological modifications 227 Transferrin-like molecule 208 Transglutaminase 6 13 fl-casein 613 GlcNAc-Asn linkage 613 N-glycosides 61 3 Transglycosylation 612, 61 3 endo-(l-hexosaminidases 6 13 endo-A Arthrobacter 6 I3 endo-M Mucor 6 13 glycohydrolases 612 glycosidases 612 glycosyl-transferases 6 12 trans-sialylase Trypanosoma cruzi 6 I2 a,a’-Trehalose 4 I5 N,N!N”-Triacetylchitotriose 405 Triantennary glycans 56, 62 Triantennary glycopeptide 61 6 Triantennary oligosaccharides 61 6 Trichomonas uaginalis 439 Trichosanthes japonica 405, 424426, 461, 462, 470 Trichosanthes kirilowii 429, 461, 470 Trisaccharide type A 420, 421 Triticum aestiuum 4 I0 Trilicum oulgaris 405, 406, 408, 410, 411, 417, 425, 427, 429, 430, 436438, 440, 441, 458, 464, 467 Trophozoites 492, 500 Trypanosoma cruzi 439 infectivity 341 Trypanosomes 464466 Tsetse fly 340 Tulip, see Tulipa gesneriana Tulipa 429 Tulipa gesneriana 405, 415, 416, 456, 462, 467 Tumor cells 579 Tumor-derived glycoproteins 576 Tumorigenicity 439 Tumors, metastatic 576 Tumors, primary 576 Tunicamycin 107, 513, 516, 517, 519, 539, 541, 557 Turnover rate 82 Type 1 fimbriae, see under Fimbriae Type A saccharide 420 Type A trisaccharide 421 Type P fimbriae. see under Fimbriae Type S fimbriae, see under Fimbriae Tyrosine kinases 83

UDP-N-acetylglucosamine-2’-epimerase3 1 1 UDP-Gal:GalNAc(fiI 4)GlcNAc(PI-R fi-1,3-galactosyltransferase 138 UDP-GalNAc:GlcNAc(fil-R fi- 1,4-N-acetylgalactosaminyltransferase 138 UDP-Xyl:GlcNAc(fi1-2)Man(a 1-3)Man(fil -R (Xyl to Manfi) fi-I ,2-xylosyltransferase I38 Ulex europaeus 405, 406, 408, 409, 412, 417, 423, 424, 429, 440, 458,461, 464, 466, 469 Ultraviolet spectroscopy 409 Uncovering enzyme 94 Undecasaccharide 490 Urochordate Pyura stolonifera 208 Urtica dioica 406, 410, 417, 429, 458, 464, 468 Urticaceae 409, 4 17 Uvomorulin 5 10, 5 I 1 Vaccines 259 Van der Waals interactions 410 Variable surface antigenic glycoproteins (VSG) 340 V-Erb-a 40 Vero cells 437 Versican 16, 17 aggregate formation with hyaluronan 17 Vetch, see Vicia cracca Vibrio cholerae 489 Vicia 405, 469 Vicia album 429 Vicia cracca 405, 406, 41 I , 457, 461 Vicia eruilia 405, 457 Vicia faba 405, 410412, 414, 415, 429, 440, 457 Vicia graminea 405, 461, 465 Vicia saliva 405, 457 Vicia uillosa 405, 406, 429, 443, 46 I , 465 Vigna radiata 429 Vinculin 41, 512 Viral proteins 42 Virus hemagglutinins, sialic acid interaction ‘ H NMR spectroscopy 360 X-ray crystallography 360 Viruses 173, 476-480, 495 corona 478 influenza 476480 hemagglutinins 478 NeuAc(a2-6)Gal(PI -4)Glc [sialyl(a2-6)lactose] 478 Newcastle disease 476 polyoma 476, 479 rota 476 Sendai 476 Simplex 480

652 Viscum album 405, 461, 465, 468 Vitamin A 532 Vitellogenesis 156, I57 Vitronectin 190, I9 I diantennary N-glycans 191 mono-antennary glycans 191 plasma concentration 190 sialic acids 191 triantennary glycans 19 1 Von Willebrand factor 19 1 as carrier molecule for factor VIIl collagen 191 endothelial cells 191 GPlb-V-IX complex 191 haemostasis I9 I heparin 191 megakaryocytes 19 I NMR spectroscopy I9 I platelets 191 a-granules I91 tetraantennary glycan 19 1 thrombin 191

Western blot analysis 435 WGA, see Wheat germ agglutinin (WGA) Wheat, see Tritium uulguris Wheat germ agglutinin (WGA) 38, 39, 490, 559 Wheat-germ-agglutinin-resistant cells 555 Winged bean, see Psophocarpus retrugonolobus Msiaria Joribunda 405, 406, 429, 461, 469 Wound-induced proteins 409

191

Xenopus oocytes 38, 39 X-ray crystal structure 403, 407, 414, 415, 417 X-ray crystallography 309, 4 0 9 4 1 1, 4 7 8 4 8 0 , 490 X-ray diffraction molecular dynamics simulations 232 molecular modelling 232 Xylose 92, 123, 409 Xylosyl transferase 9 Y, T, bird and broken-wing conformation Yeast mannan 415 Yeast mutants 75

232