glycobiologyand medicine and...proceedings of the 7th jenner glycobiology and medicine symposium...

GLYCOBIOLOGY AND MEDICINEProceedings of the 7th Jenner Glycobiology

and Medicine Symposium

ADVANCES IN EXPERIMENTAL MEDICINE AND BIOLOGY

Editorial Board:

NATHAN BACK, State University of New York at Buffalo

IRUN R. COHEN, The Weizmann Institute of Science

DAVID KRITCHEVSKY, Wistar Institute

ABEL LAJTHA, N. S. Kline Institute for Psychiatric Research

RODOLFO PAOLETTI, University of Milan

Recent Volumes in this Series

Volume 561

CHEMISTRY AND SAFETY OF ACRYLAMIDE IN FOOD

Edited by Mendel Friedman and Don Mottram

Volume 562

CHOLINGERGIC MECHANISMS

Edited by José Gonzalez-Ros

Volume 563

UPDATES IN PATHOLOGY

Volume 564

GLYCOBIOLOGY AND MEDICINE

Edited by John S. Axford

Volume 565

SLIDING FILAMENT MECHANISM IN MUSCLE CONTRACTION: FIFTY YEARS OF

RESEARCH

Edited by Haruo Sugi

Volume 566

OXYGEN TRANSPORT TO TISSUE XXVI

Edited by Paul Okunieff, Jacqueline Williams, and Yuhchyau Chen

Volume 567

THE GROWTH HORMONE-INSULIN-LIKE GROWTH FACTOR AXIS DURING

DEVELOPMENT

Edited by Isabel Varela-Nieto and Julie A. Chowen

Volume 568

HOT TOPICS IN INFECTION AND IMMUNITY IN CHILDREN II

Edited by Andrew J. Pollard and Adam Finn

Volume 569

EARLY NUTRITION AND ITS LATER CONSEQUENCES: NEW OPPORTUNITIES

Edited by Berthold Koletzko, Peter Dodds, Hans Akerbloom, and Margaret Ashwell

A Continuation Order Plan is available for this series. A continuation order will bring delivery of each

new volume immediately upon publication. Volumes are billed only upon actual shipment. For fur-

ther information please contact the publisher.

Edited by David C. Chieng and Gene P. Siegal

GLYCOBIOLOGY ANDMEDICINEProceedings of the 7th Jenner Glycobiology and Medicine Symposium

Edited by

John S. Axford

St George’s, University of London, UK t

ISBN-10 0-387-25514-1 (HB)ISBN-10 0-387-25515-X (e-book)ISBN-13 978-0-387-25514-9 (9 HB)ISBN-13 978-0-387-2551 (e-book)15-6

A C.I.P. Catalogue record for this book is available from the Library of Congress.

P.O. Box 17, 3300 AA Dordrecht, The Netherlands.

Printed on acid-free paper

All Rights Reserved

© 2005 Springer

No part of this work may be reproduced, stored in a retrieval system, or transmitted

in any form or by any means, electronic, mechanical, photocopying, microfilming, recording

or otherwise, without written permission from the Publisher, with the exception

of any material supplied specifically for the purpose of being entered

and executed on a computer system, for exclusive use by the purchaser of the work.

Printed in the Netherlands.

www.springeronline.com

Published by Springer,

INTRODUCTION

7th Jenner Glycobiology and Medicine SymposiumSunday 5 – Wednesday 8 September 2004WW

John S. Axford

The potential for glycobiology to improve the practice of medicine has been wellrecognised, which is why biannual meetings concerning the association have beentaking place for the last 14 years. The science of glycobiology has matured rapidly,and with it the far reaching clinical implications are becoming understood. The nextdecade is going to see this final frontier of science conquered. The impact thisunderstanding of glycobiology will have upon our practice of medicine is going tobe exciting. The 7th Jenner Glycobiology and Medicine Symposium was designedto reflect these advances. All the major clinical areas were involved, with contributionsfrom pivotal players in science and medicine.As with our previous meetings, junior scientists were involved as we recognise

that at the end of the next decade they will be in the driving seat. This introductionserves as a taster to whet your appetite.From embryogenesis to pathogenesis, glycosylation plays a pivotal role.

Complex and hybrid N-glyans and O-fucose glycans are critical in oocyte develop-ment and function. This area must surely be a fertile ground for glycosylationresearch.The pathogenesis of viral infections involves sugars at every turn. Hepatitis C

virus and Bovine viral diarrhoea virus are diseases that are opening themselves toscrutiny. The BVDV has proven very useful in the evaluation of the antiviral activityof molecules that inhibit morphogenesis and/or viral entry. Infection by humanimmunodeficiency virus type-1 is characterised by low levels of neutralising antibod-ies. One broadly neutralising human monoclonal antibody is 2G12.This has 3 pos-sible combining sites and recognises a cluster of oligomannose residues on the‘‘immunologically silent’’ face. This recognition provides exciting challenges forimmunogen design.

vJohn S. Axford (ed.), Glycobiology and Medicine, v-viii.

© 2005 Springer. Printed in the Netherlands.

St George’s, University of London, UK

vi J. S. Axford

N- and O-linked glycosylation of enveloped glycoproteins permits Ebola virusbinding to host cells. It is thought that an alternative pathway to the calnexin-calreticulin folding and quality control pathway is being used by the viralglycoproteins.

Immune mechanisms will be a major focus for clinical intervention over thenext decade. This will involve modulation of both the innate and adaptive immunedefences.

The innate immune system provides the first line of defence to invading patho-gens, and recognition of pathogens governs the induction, and type of pathogen-specific adaptive responses.

Schistosomiasis is a major tropical parasitic disease. Recently several antigenpresenting cell-associated lectins, that show interaction with egg glycoproteins of S.mansoni, such as the dendritic cell-specific DC-SIGN, L-SIGN on liver sinusoidendothelial cells, MGL (macrophage galactose-type lectin) and galectin 3, have beenidentified. Since their glycan ligands occur on many parasitic helminths, that theymay constitute parasite patterns for lectin-mediated immune recognition. Forexample, Lex interacts with DC-SIGN on dendritic cells and it is thought that thisinteraction may play a role in triggering dendritic cells s to mount to a Th2 response.

Mannan binding lectin (MBL) is an oligomeric protein designed to recognizepathogen-associated molecular patterns. The biological importance of MBL wasindicated when opsonin-deficient children with recurrent infections were found to begenetically deficient in MBL. Further interest in this molecule was sparked by theobservation of complement activation upon binding to carbohydrates. Glycosylationis now known to play a central role in the MBL pathway of complement activationand the glycosylation of the mannose receptor determines its functional specialisation.

Glycan structures that can act as potential ligands for MBL have been identifiedon all the immunoglobulins. In human serum only IgG-G0 and polymeric anddimeric IgA have been shown to bind MBL and initiate the lectin pathway ofcomplement. This is thought to occur through GlcNAc-terminating glycan structures.

Disease associations with sugar changes are plentiful, when the adaptive immunesystem is considered. This may involve fundamental processes, for example glycosyl-ation related molecular mechanisms are thought to involve the function of the T cellco-receptor CD8; which will have far reaching implications if abnormal.

Sugar associations with cancer have been recognised for some time. Therecontinues to be new data concerning ovarian cancer and arthritis, but research isexpanding into new areas. Sugars have now been shown to be associated with thepathological mechanism associated with the GPI anchorage of the prion protein,pigeon fanciers’ lung and muscular dystrophy.

At least six different forms of muscular dystrophy are caused by genes thatffffencode glycosyltransferases, and when malfunctioning result in a secondary deficiencyin the glycosylation of dystroglycan.

Autoimmune arthritis has been associated with the generation of remnant glyco-epitopes by gelatinase B.gelatinase B/matrix metalloproteinase-9, which is aninflammatory mediator and effector. Considerable amounts of gelatinase B areffffreleased by neutrophils in the synovial cavity of patients with rheumatoid arthritis.This is thought to be linked to the pathogenesis of arthritis as gelatinase B-deficientmice are resistant to antibody-induced arthritis. Determination of T-cell reactivity

Introduction vii

against the gelatinase B-cleaved fragments of collagen II indicates that, there aremany glyco-epitopes present in collagen II, which reinforces the role of glycopeptideantigens in autoimmunity.

It is however always exciting when clinical anecdotes translate into therapeuticand diagnostic possibilities. Glycobiology is at that transition.

Once disease mechanisms have been understood, the next step is to determinewhether this information can be used to devise therapeutic options. Predictably,therapeutic hypotheses are plentiful. For example, there is a possibility that poly-saccharide may be used to block skin inflammation.There have been new developments in treating glycosphingolipid storage dis-

eases. This may not be a common group of diseases, but for those that have it thisopens the door to improved quality of life.

The glycosphingolipid lysosomal storage diseases result from defects in glyco-sphingolipid catabolism. They are progressive disorders, the majority of whichinvolve storage and pathology in the central nervous system. A new approach totreatment is substrate reduction therapy (SRT), using small molecule inhibitors toreduce the rate of glycosphingolipid biosynthesis, to offset the catabolic defect. Oneffffof these drugs, NB-DNJ has recently been approved for clinical use in type 1 Gaucherdisease, following a successful clinical trial. There is also the potential of combiningSRT with drugs that target the downstream consequences of storage.

RA is a common disorder where the available diagnostic tests eg rheumatoidfactor, anti-citrulinated cyclic peptide, lack sensitivity. The diagnostic potential ofIgG glycosylation has been previously discussed and we await the results fromprospective trials.

Indeed, abnormal galactosylation of polyclonal IgG in ANCA associated sys-temic vasculitis patients has now been reported and the diagnostic potential of thistechnology for other autoimmune rheumatic diseases is significant.

Experiments looking at the cause behind these sugar changes indicate bothquantitative and qualitative changes in the RA serum GTase isoform profile. This islikely to be due to a greater proportion of hypersialylated isoforms, which have thepotential to adversely affect the catalytic activity of the enzyme, thus providing affffpossible mechanism for post-translational regulation of GTase activity in RA. It alsoprovides further evidence that RA glycosylation changes may be more general thanpreviously indicated and encompass proteins other than IgG. These observationscan only strengthen the potential of sugars as RA disease biomarkers.

The selectin family of adhesion molecules mediates the initial attachment ofleukocytes to venular endothelial cells at sites of tissue injury and inflammation. Forexample in staphylococcal arthritis. Fucoidin, a sulfated polysaccharide from sea-weed, binds to and blocks the function of L-and P-selectins thereby inhibitingleukocyte rolling and adhesion to endothelial surface. Treatment with fucoidin hasbeen shown to reduce the severity of septic arthritis within the first three daysfollowing bacterial infection. It is suggested that the efficient treatment of septicarthritis should encompass a combination of antibiotics and immunomodulation.Gastroenterologists want to know more about normal and abnormal bacteria

that inhabit our bowels. O-acetyl sialic acid expression in colorectal mucosa hasbeen shown to be regulated by enteric microflora; as demonstrated by the loss ofsialic acid oligo-O acetylation after elimination of the faecal flow. There is therefore

viii J. S. Axford

potential to use this observation to quantitate bacterial colonisation and perhapsinterfere with disease associated pathology.Biological therapies will be the new treatments of the next decade. The impact

of glycosylation on the structure and function of natural and recombinant (thera-peutic) IgG antibody is therefore important to get to grips with.It has been shown that the in vivo micro-environment can have a profound

influence on the glycosylation profile of IgG-Fc. This may reflect the unique structuralrelationship between the oligosaccharide and the protein. The ‘‘core’’ heptasaccharideis essential for FccRI, FccRII, FccRIII and C1 activation whilst outer arm sugarresidues can influence these and other functions, e.g. FccRIII, FcRn, MBL, MR.Thus, fidelity of glycosylation is essential to the effector function profile of antibodiesffffand in the future the oligosaccharide will be used to function as a structural ‘‘rheostat’’to generate specific glycoforms exhibiting optimal effector activities for a particularffffdisease target.Cellular glycoengineering for fully human glycosylation and optimised sialyl-

ation of proteins is therefore going to be important if these molecules are going tobe fully and specifically active. Most pharmaceutical proteins are expressed in bac-teria, yeast or mammalian cells resulting in proteins lacking glycosylation or carryingglycans which largely differ from human carbohydrate chains in various aspectsffffincluding sialylation. However, relationships between the N-glycan structures andbiological activities of, for example, recombinant human erythropoietins producedusing different culture conditions and purification procedures are now better under-ffffstood. It is nevertheless apparent that novel glycoprotein expression technology willneed to be developed to address this problem and data is now available to demon-strate how this can be done.The above introduction adds up to the fact that Glycobiology is an extremely

exciting science to be involved in. Additionally, if you are a clinician, it is even moregripping as you will be at the forefront of important clinical developments.I hope these proceedings stimulate you as much as they have me and I look

forward to seeing you at Jenner 8!

REFERENCES

Axford J Keida C Van Dijk WV, Rudd PM. 6th Jenner Glycobiology and Medicine. CPD Bulletin.

Immunology & Allergy 2004; 3(3): 84–87.

Alavi A, Axford J. Glycobiology of the Rheumatic Diseases: an update. Adv Exp Med Biol. 2003; 535:

271–80.

Axford J, Keida C & Dikk WV. Meeting Report 5th Jenner Glycobiology and Medicine. Glycobiology

2001; 11(2): 5G-7G.

Axford JS. 4th Jenner International Glycoimmunology Meeting: A Review. Immunol Today 1997; 18(11):

511–513.

Axford JS. 3rd Jenner International Glycoimmunology Meeting. Immunol Today 1995; 16(5):213–215.

Axford JS. 2nd Jenner International Glycoimmunology Meeting. Immunol Today 1993; 14(3): 104–106.

CONTENTS

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v

1. Glycosylation: Disease Targets and Therapy . . . . . . . . . . . . . . . . . . . . . . . 1Nicole Zitzmann, Timothy Block, Anund Methta, Pauline Rudd,Dennis Burton, Ian Wilson, Frances Platt, Terry Butters, andRaymond A. Dwek

2. Long Alkylchain Iminosugars Block the HCV P7 Ion Channel . . . . . . . 3D. Pavlovic, W. Fischer, M. Hussey, D. Durantel, S. Durantel,N. Branza-Nichita, S. Woodhouse, R. A. Dwek, and N. Zitzmann

3. The Bovine Viral Diarrhoea Virus: A Model for the Study of AntiviralMolecules Interfering with N-Glycosylation and Folding of EnvelopeGlycoprotein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5D. Durantel, N. Branza-Nichita, S. Durantel, R.A. Dweek, andN. Zitzmann

4. Antibody Recognition of a Carbohydrate Epitope: A Template for HIVVaccine Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7Chris Scanlan, Daniel Calarese, Hing-Ken Lee, Ola Blixt,Chi-Huey Wong, Ian Wilson, Dennis Burton, Raymond Dwek,and Pauline Rudd

5. Interaction of Schistosome Glycans with the Host Immune System . . . . 9Irma van Die, Ellis van Liempt, Christine M. C. Bank, and WietskeE. C. M. Schiphorst

6. The Mannan-Binding Lectin (MBL) of Complement Activation:Biochemistry, Biology and Clinical Implications . . . . . . . . . . . . . . . . . . . . 21Jens Christian Jensenius

7. Killer Cell Lectin-like Receptors and the Natural Killer Cell GeneComplex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Ø. Nylenna, L. M. Flornes, I. H. Westgaard, P. Y. Woon, C. Naper,J. T. Vaage, D. Gauguier, J. C. Ryan, E. Dissen, and S. Fossum

8. Glycosylation Influences the Ligand Binding Activities of MannoseReceptor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25Yunpeng Roc Su, Clarence Tsang, Talitha Bakker, James Harris,Siamon Gordon, Raymond A. Dwek, Luisa Martinez-Pomares andPauline M. Rudd

ix

Introduction

x

9. Human Immunoglobulin Glycosylation and the Lectin Pathway ofComplement Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27James N. Arnold, Louise Royle, Raymond A. Dwek,Pauline M. Rudd, and Robert B. Sim

10. Gelatinase B Participates in Collagen II Degradation and ReleasesGlycosylated Remnant Epitopes in Rheumatoid Arthritis . . . . . . . . . . . 45P. E. Van den Steen, B. Grillet, and G. Opdenakker

11. Hyaluronan in Immune Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57Alan J. Wright and Anthony J. Day

12. Glycosylation and the Function of the T Cell Co-receptor CD8 . . . . . . 71David A. Shore, Ian A. Wilson, Raymond A. Dwek, andPauline M. Rudd

13. Immunogenecity of Calreticulin-bound Murine Leukemia VirusGlycoprotein gp90 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85Yusuke Mimura, Denise Golgher, Yuka Mimura-Kimura,Raymond A. Dwek, Pauline M. Rudd, Tim Elliott

14. Glycosylation and GPI Anchorage of the Prion Protein . . . . . . . . . . . . 95N. M. Hooper

15. Glycosylation Defects and Muscular Dystrophy . . . . . . . . . . . . . . . . . . . 97Derek J. Blake, Christopher T. Esapa, Enca Martin-Rendon, andR. A. Jeffrey McIlhinneyffff

16. Roles of Complex and Hybrid N-Glycans and O-Fucose Glycans inOocyte Development and Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99S. Shi, S. A. Williams, H. Kurniawan, L. Lu, L., and P. Stanley

17. Mucin Oligosaccharides and Pigeon Fanciers’ Lung . . . . . . . . . . . . . . . 101C. I. Baldwin, A. Allen, S. Bourke, E. Hounsell, and J. E. Calvert

18. Differential Glycosylation of Gelatinase B from Neutrophils andffffBreast Cancer Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103Simon A. Fry, Philippe E. Van den Steen, Louise Royle,Mark R. Wormald, Anthony J. Leathem, Ghislain Opdenakker,Pauline M. Rudd, and Raymond A. Dwek

19. Detection of Glycosylation Changes in Serum and Tissue Proteins inCancer by Lectin Blotting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113R. E. Ferguson, D. H. Jackson, R. Hutson, N. Wilkinson,P. Harnden, P. Selby, and R. E. Banks

20. Carbohydrates and Biology of Staphylococcal Infections . . . . . . . . . . . . 115Andrej Tarkowski, Margareta Verdrengh, Ing-Marie Jonsson,Mattias Magnusson, Simon J Foster, and Zai-Quing Liu

21. New Developments in Treating Glycosphingolipid Storage Diseases . . 117Frances M. Platt, Mylvaganam Jeyakumar, Ulrika Andersson,Raymond A. Dwek and Terry D. Butters

22. Fucosylated Glycans in Innate and Adaptive Immunity . . . . . . . . . . . . 127J. B. Lowe

xi

23. New Insights into Rheumatoid Arthritis Associated GlycosylationChanges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129Azita Alavi, Andrew J. Pool, and John S. Axford

24. Production of Complex Human Glycoproteins in Yeast . . . . . . . . . . . . 139Tillman Gerngross

25. Relationship Between the N-Glycan Structures and BiologicalActivitities of Recombinant Human Erythropoietins Produced UsingDifferent Culture Conditions and Purification Proceduresffff . . . . . . . . . . . 141C-T. Yuen, P. L. Storring, R. J. Tiplady, M. Izquierdo, R. Wait,C. K. Gee, P. Gerson, P. Lloyd, and J. A. Cremata

26. Glycosylation of Natural and Recombinant Antibody Molecules . . . . . 143Roy Jefferisffff

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

1

GLYCOSYLATION: DISEASE TARGETSAND THERAPY

Nicole Zitzmann1, Timothy Block2, Anund Methta2, Pauline Rudd1,Dennis Burton2, Ian Wilson3, Frances Platt1, Terry ButtersTT 1, andRaymond A. Dwek1

1Oxford UniversityDepartment of BiochemistryGlycobiology Institute, Oxford1Jefferson CentreffffDoylestown, USA3The Scripps Research InstituteLa Jolla, USA

Four different glycosylation approaches illustrate strategies for providing therapy inffffdisease targets. Two of these are to develop antiviral therapies using iminosugarderivatives. The first approach involves targeting virus encoded protein(s), the secondtargets host cell encoded protein(s) necessary for virus survival. The latter couldpotentially prove more resistant to the development of viral escape mutants, aproblem plaguing most conventional drug therapies.

In the case of hepatitis C virus (HCV), which affects about 3% of the worldffffpopulation, it is possible to employ both strategies. Using bovine viral diarrhoeavirus (BVDV) as a model organism for HCV we showed that inhibition of the hostcell ER alpha-glucosidases I and II using the glucose analogue deoxynojirimycin(BuDNJ) leads to an antiviral activity caused by a reduction in viral secretion dueto the interference with viral envelope glycoprotein folding and subsequent impair-ment of viral morphogenesis. However, it is also possible to target the virally encodedHCV protein p7 which can form ion channels, using long alkylchain derivativesof either DNJ or the galactose analogue deoxygalactonojirimycin (DGJ). N7-oxanonyl-6deoxy-DGJ is currently in phase I clinical studies.Worldwide, more than 350 million people are chronically infected with hepatitis

B virus (HBV). Glucosidase inhibitors have been shown to be antiviral against HBVin tissue culture and in the woodchuck model of chronic HBV infection. The Msurface antigen glycoprotein of HBV folds via the calnexin pathway. Glucosidase

1

John S. Axford (ed.), Glycobiology and Medicine, 1-2.


2 N. Zitzmann et al.

inhibitors which prevent this interaction prevent the formation and secretion ofHBV. The misfolded M surface antigen is retained within the cell and may itself actas a long lived ‘‘drug’’ which prevents virus formation. As in the case of HCV, asecond class of long alkylchain iminosugars which do not inhibit glycan processingare also potent antiviral agents and may act at a stage before viral envelopment, butthe mechanism is still unknown. Toxicology studies are underway to identify acompound from these classes of orally available drugs for clinical trials for thetreatment of chronic HBV infection.The humoral immune response to infection by human immunodeficiency virus

type-1 (HIV-1) is characterised by low levels of neutralising antibodies, particularlythose which have a broad specificity against many different isolates. One broadlyffffneutralising human monoclonal antibody is 2G12. This has a novel antibody structurewith 3 possible combining sites and recognises a cluster of oligomannose residueson the ‘‘immunologically silent’’ face. This recognition provides exciting challengesfor immunogen design. The use of imino sugars as with other viruses may alsoprovide additional possibilities for antiviral therapy.T he glucosphingolipid (GSL) storage diseases are a family of progressive disorders

in which GSL species are stored in the lysosome, as a result of defects in the enzymesof the GSL degradation pathway. Specific diseases include Gaucher, Tay-Sachs,Fabry, Sandhoff and GM1 gangliosidosis. GSL storage diseases occur at a collectivefrequency of 1 in 18,000 live births and are one of the most common cause ofneurodegenerative disease in infants and children. Our drug based strategy formanagement of these diseases is to inhibit partiallyGSL synthesis using imino sugars.Slowing the rate of synthesis of GSLs will lead to fewer entering the lysosome forcatabolism, reducing the rate of storage. This substrate reduction therapy (SRT) haslead to a world wide approved oral drug (NB-DNJ) for Gaucher type-1 disease.

2

LONG ALKYLCHAIN IMINOSUGARS BLOCK THEHCV P7 ION CHANNEL

D. Pavlovic1, W. Fischer1, M. Hussey1, D. Durantel2, S. Durantel2,N. Branza-Nichita3, S. Woodhouse1, R. A. Dwek1, and N. Zitzmann1

1Oxford Glycobiology InstituteDepartment of BiochemistryUniversity of Oxford, Oxford, UK2Virus des hepatites et pathologies associeesINSERM U271, Lyon, France3Institute of BiochemistrySplaiul IndependenteiBucharest, Romania

The small p7 protein of the hepatitis C virus (HCV) and the closely related bovineviral diarrhea virus (BVDV) can form ion channels in artificial membranes (see

3



Figure 1. Channel recordings of synthetic HCV p7 inserted into a black lipid membrane (BLM). Channel

activity is shown for +/_

50 mV and +/_ _

100 mV. The clos_

ed state is shown as a solid line. Channel openings

are deviations from this line. Solutions are the same in the cis and trans chamber: 0.5 M KCl, 5 mM Hepes,

1 mM CaCl2, pH 7.4. HCV p7 is added on the trans side to a final concentration of approximately 50 microM.

Scale bars are 10 s and 100 pA.

4 D. Pavlovic et al.

Fig. 1). Ion channel activity can be suppressed by long alkylchain iminosugar derivat-ives, which have been shown to be antiviral in BVDV infectivity assays. Treatmentwith these inhibitors does not affect viral RNA replication. However, the infectivityffffof virions secreted in the presence of the inhibitors is impaired. The physiologicalrole of the p7 ion channel during the viral life cycle is unknown and is beinginvestigated using inhibitory iminosugars as well as a BVDV construct from whichp7 has been deleted.

REFERENCES

1. Study of the mechanism of the antiviral action of iminosugar derivatives against Bovine Viral Diarrhea

Virus. D. Durantel, N. Branza-Nichita, S. Carrouee-Durantel, T. D. Butters, R. A. Dwek and

N. Zitzmann (2001), Journal of Virology 75 (19), 8987–8998

2. The hepatitis C virus p7 protein forms an ion channel which is inhibited by long alkylchain iminosugar

derivatives. D. Pavlovic, D. C. A. Neville, O. Argaud, B. Blumberg, R. A. Dwek, W. B. Fischer and N.

Zitzmann (2003), PNAS 100 (10), 6104–610

3. Effect of interferon, ribavirin and iminosugar derivatives on viral infection in cells persistently infectedffff

with non-cytopathic BVDV: A comparative study. D. Durantel, S. Carrouee-Durantel, N. Branza-

Nichita, R. A. Dwek and N. Zitzmann (2004), Antimicrobial Agents and Chemotherapy 48 (2), 497–504

3

THE BOVINE VIRAL DIARRHOEA VIRUS:A MODEL FOR THE STUDY OF ANTIVIRALMOLECULES INTERFERING WITHN-GLYCOSYLATION AND FOLDING OFENVELOPPE GLYCOPROTEIN

D. Durantel1, N. Branza-Nichita2, S. Durantel1, R. A. Dweek3,and N. Zitzmann3

1Laboratoire des virus hepatiques et pathologies associeesINSERM U271, Lyon, France2Institute of BiochemistrySector 6, Bucharest, Romania3Glycobiology InstituteDepartment of BiochemistryUniversity of Oxford, Oxford, U.K.

The current treatment of chronic hepatitis C combines interferon alpha and ribavirinand is effective in only half of the patients treated. Considerable effff fforts are beingffffmade to develop novel anti-HVC molecules with a better efficacy particularly forrefractory patients. Molecules targeting specifically viral activities are the moststudied. However, an antiviral strategy based uniquely on the utilisation of this typeof molecules is expected to encounter problems caused by the emergence of viralescape mutants, as already widely described for HIV and HBV. Alternativeapproaches and molecules are needed to complement antiviral strategies based oninhibitors of viral enzyme. Ideally, new molecules should target steps of the viralcycle that are potentially less likely to give rise to resistance. The assembly andmorphogenesis of HCV belong to these yet untargeted steps of the life cycle.As no cellular system able to support the morphogenesis and secretion of HCV

particles is yet available, the bovine viral diarrhoea virus (BVDV), which is phyllo-genetically close and shares biological features with HCV, has been used as asurrogate model for the study of antiviral molecules interfering with theN-glycosylation and folding of viral envelope glycoprotein.We have demonstrated that some analogues of glucose (deoxynojirimycin), also

5John S. Axford (ed.), Glycobiology and Medicine, 5-6.


6 D. Durantel et al.

generically called iminosugars, are good inhibitors of morphogenesis and preventviral re-entry by reducing the infectivity of released virions. The mechanism of actionhas been studied at molecular level for these iminosugars presenting antiviral activityagainst BVDV. Two different mechanism of action have been defined to explain theffffwhole antiviral effect. DNJ derivatives inhibit host ER a-glucosidases, thus preventingffffthe trimming of 2 glucoses from triglycosylated N-glycans and the subsequentinteraction with lectin chaperone. This inhibition results in the misfolding of viralglycoprotein and the subsequent defect in assembly, budding and viral secretion.Moreover, DNJ derivatives induce a diminution of viral infectivity and thereforeprevent re-infection of cells by neo-formed particles. This is likely due to the incorp-oration of non functional envelope glycoprotein complexes.In conclusion, the BVDV has proven very useful to evaluate the antiviral activity

of molecules that inhibit morphogenesis and/or viral entry. The BVDV will remainan interesting model for HCV while waiting for the development of a cell culturesystem able to fully propagate the latter.

4

ANTIBODY RECOGNITION OF A CARBOHYDRATEEPITOPE: A TEMPLATE FOR HIV VACCINEDESIGN

Chris Scanlan1,2, Daniel Calarese2, Hing-Ken Lee2, Ola Blixt2,Chi-Huey Wong2, Ian Wilson2, Dennis Burton2, Raymond Dwek1,and Pauline Rudd1

1Glycobiology InstituteUniversity of OxfordSouth Parks Rd, Oxford 0X1 3QU2The Scripps Research Institute10550 N.Torrey PinesLa Jola, CA 92037

The humoral response to HIV-1 infection is typically characterized by low levels ofneutralizing antibodies, especially antibodies which can provide sterilizing immunityagainst a wide range of HIV isolates. However, a small number of antibodies, isolatedfrom infected individuals, have been shown to protect against HIV challenge in

Figure 1. Most of the antigenic surface of HIV-1 gp120 is glycosylated.

7



8 C. Scanlan et al.

animal models. As such, these antibodies are potential templates for HIV vaccinedesign. One such antibody is the broadly neutralizing antibody 2G12. Alaninescanning mutagenesis, glycosidase digests and competition experiments demonstratedthat 2G12 binds to a cluster of alpha1�2-linked mannose residues on the outer faceof gp120. Cyrstallographic studies showed that IgG 2G12 exhibits a unique domain-exchanged Fab configuration. Mutagenesis of 2G12 Fab, combined with carbohyd-rate inhibition experiments explained how the unusual structure of 2G12 is able torecognize its neutralization epitope on gp120. Synthetic mimics of the 2G12 epitopeare currently under evaluation as potential immunogens.

REFERENCES

1. Angew Chem Int Ed Engl. 2004 Feb 13;43(8):1000.

2. Science. 2003 Jun 27;300(5628):2065–71.

3. J Virol. 2002 Jul;76(14):7306–21.

5

INTERACTION OF SCHISTOSOME GLYCANS WITHTHE HOST IMMUNE SYSTEM

Irma van Die, Ellis van Liempt, Christine M. C. Bank, and WietskeE. C. M. Schiphorst

Department of Molecular Cell Biology and ImmunologyVU University Medical CenterVan der Boechorststraat 7, 1081 BT Amsterdamthe Netherlands

1. INTRODUCTION

Schistosomiasis is a parasitic disease caused by trematodes that affects moreffffthan 200 million people worldwide, mostly children in developing countries.Annually, 200 000 deaths are estimated to be associated with schistosomiasis (van derWerf et al., 2002). Until now, attempts to control infection and disease have mostlyfailed but the disease can be effectively treated by chemotherapy (Praziquantel ). Oneffffof the most striking features of schistosomiasis is that the worms are experts inmodulation and evasion of the host immune response, to enable their survival,migration and development in different host tissues. It is becoming increasingly clearffffthat schistosome glycoconjugates play a crucial role in the evasion mechanisms thatare exploited by the parasites. Here we will summarize our studies that aim toincrease our molecular understanding of the role of specific schistosome glycanantigens in immune modulation. We will focus on the interactions of schistosomeglycans with the host immune system that result in the mounting of T helper cell 2(Th2) responses.

2. THE LIFE CYCLE OF SCHISTOSOMES

Schistosomes have a complicated life cycle, requiring two hosts. Male and femaleworms live in the veins of the abdomal cavity of their vertebrate host, sometimes formore than 20 years. Here they mate and produce eggs. Schistosome eggs that become

9



10 I. van Die et al.

lodged within host tissues are the major cause of pathology. Part of the eggs escapefrom the body and eventually reach the water with faeces or urine. In the watermiracidia hatch from the eggs and infect a suitable snail host. After asexual reproduc-tion in the snail, cercariae leave the snail and can penetrate the skin of their vertebratehost when they come into contact with the water. While penetrating the skin,cercariae loose their tails and become schistosomula that subsequently migrate viathe lungs to the veins of the abdomal cavity and develop to sexually mature adultworms. Three major schistosome species are discriminated that infect humans.Although their life cycles basically are similar, each Schistosome has a specific snailintermediate host that is essential for their development. Schistosoma mansoni thatoccurs mainly in Africa, the Middle East and South America requires snails of thegenus Biomphalaria as intermediate host. S. haematobium, found in Africa as well asin parts of the Middle East and Asia, is transmitted by snails of the genus Bulinus,whereas S. japonicum occurs mainly in South-east Asia and China and is transmittedby snails of the genus Oncomelania. All three species can also infect rodents, whichare often used as model systems and are required, in combination with the specificsnail intermediate hosts, to maintain the cycle in the laboratory for scientific research.

3. SCHISTOSOME GLYCANS AND THE SYNTHESIS OFNEOGLYCOPROTEINS

Glycans antigens are abundantly present on the surface of the different parasiteffffstages and within their excretory/secretory products. Several reviews have summar-ized the structures of glycan antigens found, as well as data that demonstrate thatthese glycans are the major focus of the host immune response (Cummings andNyame, 1996; Cummings and Nyame, 1999; Hokke and Deelder, 2001). Remarkableis the absence of sialylation in the schistosome glycans, and the high degree offucosylation in structural compositions that are not found in humans. The glycanscan be very large, consisting of many different monosaccharides that diffff ffer in sequenceffffand anomeric linkage. Therefore, one glycan molecule can encompass differentffffantigenic determinants (glycan antigens, see Fig. 1A). To elucidate their functionalrole in the host immune response, it is essential that individual glycan antigens canbe studied separately. Such glycan antigens have been synthesized in vitro usingenzymatic methods and subsequently coupled to BSA or other suitable carriers toyield neoglycoconjugates (van Remoortere et al., 2000) (Fig. 1B).

Basically, a chemical, enzymatic, or a combined chemo-enzymatic approach canbe applied for in vitro glycan synthesis. A drawback of chemical synthesis is that itrequires a complete control of the stereoselectivity. For enzymatic synthesis glycosid-ases and glycosyltransferases (GTs), Golgi enzymes that are involved in the in vivobiosynthesis of carbohydrate moieties on glycoproteins, are used. GTs catalyze thetransfer of sugar moieties from activated nucleotide-sugars to specific acceptor molec-ules, and act sequentially to built an oligosaccharide on a carrier molecule. The useof GTs offers significant advantages because it is fast and combines a high regio-ffffand stereospecificity with the potential availability of many different glycosidicfffflinkages. However, although many mammalian glycosyltransferases are cloned andavailable in recombinant form for glycan synthesis, the number of glycosyltransferases

Interaction of Schistosome Glycans 11

Figure 1. Schistosome glycan antigens and neoglycoconjugates.

A. Structure of a schistosome N-glycan (Srivatsan et al., 1992a), with different glycan antigens, LDN,ffff

LDNF and core-fucose, indicated.

B. Construction of a neoglycoconjugate carrying LDNF. The glycan antigen is synthesized using

different glycosyltransferases and coupled to BSA (van Remoortereffff et al., 2000).

Table 1. Glycosyltransferases derived from invertebrate and plant sources that may be useful for the

synthesis of typical helminth glycan antigens

Glycosyltransferase Source Recombinant Reference

b4-GalNAcT (�GlcNAcb) L . stagnalis no (Mulder et al., 1995)

b4-GalNAcT (�GlcNAcb) C. elegans yes (Kawar et al., 2002)

Core a3-FucT A. thaliana yes (Bakker et al., 2001; Fabini et al., 2001)

Drosophila yes

Core b2-XylT A. thaliana yes (Strasser et al., 2001)

a2 FucT (�Xylb) C. elegans yes (Zheng et al., 2002)

a2 FucT (�Fuca) T . ocellata no (Hokke et al., 1998)

b4 GlcNAcT (�GlcNAcb) L . stagnalis yes (Bakker et al., 1994)

b4 GlcT (�GlcNAcb) L . stagnalis no (Van Die et al., 2000)

that can be applied to synthesize typical schistosome or other helminth glycanstructures is limited. Since schistosomes share glycan antigens with plants and otherinvertebrates such as snails or the free-living nematode Caenorhabditis elegans, glyco-syltransferases derived from these sources are useful for synthesis purposes. In severalstudies, recombinant GTs cloned from C. elegans or Arabidopsis thaliana, as well asextracts from schistosomes and snails have been exploited to synthesize schistosomeglycan antigens such as LDN-DF or core-xylose/ core-fucose. In Table 1 a numberof non-human glycosyltransferases that have been applied (van Remoortere et al.,2003b; Remoortere et al., 2000; Nyame et al., 2004), or may be useful for the synthesisof Schistosome glycan antigen, are summarized.


4. SCHISTOSOME GLYCANS ARE THE MAJOR FOCUS OF THEHOST IMMUNE RESPONSE

Schistosome glycans play an important role in the hosts humoral and cellularimmune responses carbohydrate antigens

within Schistosome egg antigens (SEA) that are referred to in this review, are shown.A strong humoral response has been found against the fucosylated glycan epitopesGalNAcb1-4(Fuca1-3)GlcNAc (LDNF), GalNAcb1-4(Fuca1-2Fuca1-3)GlcNAc(LDN-DF) and Fuca1-3GalNAcb1-4GlcNAc (FLDN) in both infected animals andhumans (Nyame et al., 2000; van Remoortere et al., 2001; Eberl et al., 2002; vanRemoortere et al., 2003a; Naus et al., 2003). Sera of infected hosts also contain lowamounts of antibodies against Galb1-4(Fuca1-3)GlcNAc (Lex, CD15), a glycanepitope shared by humans and schistosomes (Nyame et al., 1998; van Remoortereet al., 2001; Eberl et al., 2002). These Lex antibodies may induce autoimmunereactions, as was shown by their ability to mediate complement-dependent cytolysisof myeloid cells and granulocytes (Nyame et al., 1996; Nyame et al., 1997; Van Damet al., 1996). Interestingly, Lex occurs on all parasite stages, i.e. cercariae, eggs,schistosomula and adult worms (Srivatsan et al., 1992b; Cummings and Nyame,1996; van Remoortere et al., 2000). Lex determinants have been found as repeatingtrisaccharides in tri- and tetra-antennary N-glycans of membrane glycoproteins(Srivatsan et al., 1992b), as core-2-based O-glycans on the secreted circulating cath-odic antigen (Van Dam et al., 1994) and on cercarial glycolipids (Wuhrer et al.,2000). Interestingly, the immune system discriminates between Lex expressed inmonomeric and polymeric form as shown by different types of antibody responsesffffin mice towards these glycan structures (Van Roon et al., 2004). Lex containingglycoconjugates also induce proliferation of B-cells from infected animals, which

to infection. In Fig. 1, some of the major

secrete IL-10 and PGE2 (Velupillai and Harn, 1994), and induce the production ofVVIL-10 by peripheral blood mononuclear cells from schistosome-infected individuals(VelupillaiVV et al., 2000). In a murine schistosome model, Lex is an effective adjuvantfffffor induction of a Th2 response (Okano et al., 2001), and it has been demonstratedthat sensitization with Lex results in an increased cellular response towards SEA-coupled beads implanted in the liver and to the formation of granulomas (Jacobset al., 1999).

5. DENDRITIC CELLS RECOGNIZE PARASITE-DERIVEDGLYCANS

Dendritic cells (DCs) form a link between innate and adaptive immunity andare therefore crucial in the defense against pathogens, such as schistosomes (Paluckaand Banchereau, 2002; Janeway, Jr. and Medzhitov, 2002). They are localized inperipheral tissues throughout the body and recognize invading pathogens usingpattern recognition receptors including Toll-like receptors and lectins. Lectins areproteins that contain carbohydrate recognition domains (CRDs) that specifically


mechanism of binding, and function. Many C-type lectins (showing Ca2+-dependentbinding of their carbohydrate ligands) function in the capture of glycoconjugatesand subsequent presentation of the antigen to the immune system (Weis et al., 1998;Figdor et al., 2002; Engering et al., 2002). Current views are that the principalfunction of at least some C-type lectins is the recognition and clearance of glycosyl-ated self-antigens, in order to induce tolerance (‘t Hart and Kooyk, 2004). Thisfunction may then be exploited by pathogens to escape immune attack. It is becomingevident that a pathogen-DC interaction is mediated by multiple sets of ligand-receptor interactions to generate a pathogen-specific response, and several distinctDC subsets have been identified to express different receptors. By targeting diffff fferentffffDC subsets, an invading pathogen can trigger mixed responses, and it is thoughtthat interaction of glycans with their receptor-lectin can either enhance or opposeTLR signalling thereby modulating the DC phenotype and outcome of the inducedimmune response (Gantner et al., 2003; Geijtenbeek T.B. et al., 2002). Some C-typelectins contain ITIM or ITAM sequence motifs in their cytoplasmic tails that indeedsuggest a potential role in either immunosuppression or immunoactivation (Figdoret al., 2002). After internalization of bound components to allow antigen-processingand presentation, DCs migrate to secondary lymphoid organs, where they presentthe captured antigens to resting T cells and, dependent on the received stimuli, inducetolerance or initiate adaptive immune responses.It should be noted that in addition to DC associated C-type lectins also members

of other lectin classes, such as galectins (galactose-binding lectins) or siglecs (sialicacid binding lectins) present on DC or on other antigen-presenting cells, have beenimplicated in host-pathogen interactions (van den Berg et al., 2004; Sato andNieminen, 2004; Jones et al., 2003). For example, we recently showed that macro-phage-derived galectin-3 is highly expressed in liver granuloma’s of schistosomeinfected hamsters and binds to LDN glycan antigens on schistosome egg antigens.

Interestingly, in vitro studies demonstrated that galectin-3 mediates phagocytosis ofLDN containing neoglycoconjugates by activated macrophages indicating a role forgalectin-3 in innate immunity to schistosomes (van den Berg et al., 2004).DCs are central in directing Th1-Th2 responses and molecular patterns on the

pathogen that are recognized by DC that capture pathogen determinants, are crucialfor biasing the Th immune response (Jankovic et al., 2002). In mouse models, DCpulsed with SEA potently stimulate Th2 responses both in vivo and in vitro whilefailing to undergo a conventional maturation process (MacDonald et al., 2001). Inaddition, in immunization experiments using a Lex neoglycoprotein, a strong Th2response was mounted that was dependent on the presense of Lex on the protein.Surprisingly, the antibodies that were generated appeared protein-specific, indicatingthat Lex merely had a function as an efficient Th2-stimulating adjuvans (Okanoet al., 2001). We therefore hypothesized that Lewis-x might be recognized by aspecific lectin on DC, and that this interaction would be (one of ) the signal(s) thatcould trigger DCs to acquire a Th2 inducing phenotype.

different types of lectins are discriminated that diffff ffer in carbohydrate specificity,ffffbind to a variety of sugars present on cell-wall or secreted glycoconjugates. Several


6. IDENTIFICATION OF LECTINS ON DENDRITIC CELLS THATBIND SCHISTOSOME EGG ANTIGENS

Because DC are central in directing Th1-Th2 responses, we searched for a cell-surface receptor expressed on human immature DC that interacts with S. mansoniegg glycoproteins (SEA). SEA is a mixture of glycoproteins, containing manyimmunogenic glycan antigens that potentially could interact with lectins on the iDC.To detect binding of SEA to human immature DC, a fluorescent-bead adhesionassay was developed. Fluorescent beads were precoated with streptavidin and thenused to capture biotinylated SEA. The conjugated beads were allowed to interactwith DC, similar to the approach described by (Geijtenbeek et al., 1999). We showedthat the immature DCs strongly bound SEA, and that this binding could be blockedcompletely by EDTA, suggesting that one or more C-type lectin(s) on DC mediatethe binding to SEA. Identification of the C-type lectin(s) involved showed that asubstantial part of the binding was mediated via the C-type lectin DC-SIGN (vanDie et al., 2003; Van Liempt et al., 2004). In addition, the C-type lectin MGL(macrophage galactose lectin) was shown to be responsible for part of the bindingof iDC to SEA (SJ van Vliet, et al, manuscript submitted). These data show thatdifferent C-type lectins on iDCs participate in the binding to Schistosome eggffffantigens, which implies that each interaction may contribute to the induction of thefinal DC function.

7. BINDING OF DC-SIGN TO SCHISTOSOME EGG ANTIGENS ISMEDIATED THROUGH INTERACTION WITH THE GLYCANSLEXIS-X AND LDNF

As DC-SIGN has been reported to display affinity to both mannose and fucosewe explored whether the binding of DC-SIGN to SEA is fucose mediated. Using acompetitive ELISA it was demonstrated that this binding indeed can be blocked bymonoclonal antibodies (mAbs) specific for the fucosylated glycans Lex and LDNF,respectively. A combination of anti-Lex and anti-LDNF mAbs strongly blocked thebinding of DC-SIGN to SEA. These binding properties have been established bydirect binding studies of DC-SIGN to Lex and LDNF containing oligosaccharides/neoglycoconjugates. Lex and LDNF glycans are both major glycan antigens withinSEA and resemble each other by containing a terminal fucose (1–3 linked to aGlcNAc residue. Our data have demonstrated that other fucosylated glycan antigenswithin SEA, such as core-fucose and LDN-DF, do not constitute ligands forDC-SIGN (van Die et al., 2003) (E. van Liempt ea, unpublished). This indicates thatbinding of DC-SIGN to Lex and LDNF glycan antigens is specific and does notonly depends on the presence of a fucose. Based on our binding studies and thecrystal structure of DC-SIGN (Feinberg et al., 2001) we recently proposed a molecu-lar model of the binding of DC-SIGN to Lex (van Liempt et al., 2004) that was


to the binding (van Liempt et al., 2004; Guo et al., 2004). The grey sphere indicates a calcium ion.

essentially similar to a model based on the crystal structure of DC-SIGN in complexwith Lex (Guo et al., 2004). Both studies showed that the fucose of Lex stronglyinteracts with Val351 in DC-SIGN, and that both the galactose and GlcNAc showadditional contacts with the CRD of DC-SIGN and contribute to the binding. Fromthe molecular an N-acetylgroup linkage to the

C2 of galactose within the Lex trisaccharide, as is found in LDNF, will not interactwith the CRD of DC-SIGN, which is in agreement with the observed binding ofDC-SIGN to LDNF. By contrast, there is no place in the binding pocket for an

Figure 2. Binding of Lewis-x to the CRD of DC-SIGN. The three monosaccharides of Lex all participate

model (Fig. 2) it can be deduced that

additional fucose a2-linked to the Fuca1, 3GlcNAc moiety of LDNF, such as found

in LDN-DF, which explains the lack of binding of DC-SIGN to LDN-DF. Ourdata indicate that DC-SIGN shows a strongly increased binding to multivalentlypresented Lex and LDNF (Van Liempt et al, unpublished). The optimal bindingmost likely is dependent on the spacing and presentation of the antigens within theglycan, and further studies are underway to identify the actual schistosome ligandsthat bind DC-SIGN.

8. CONCLUDING REMARKS

Our knowledge of the remarkable role that Schistosome glycoconjugates playin the immunobiology of schistosomiasis is rapidly growing. We now are getting


insight in the receptors on dendritic cells and macrophages that interact withSchistosome glycan antigens, although the functional consequences of these inter-actions mostly remain to be understood. Our data indicate that Lex interacts withDC-SIGN on DC, which in combination with the data from Okano et al., (2001),suggest that this interaction may play a role in triggering DCs to direct the mountingof a Th2 response. This hypothesis is further strengthened by recent data showingthat inHelicobacter pylori infection, Lex positive variants block Th1 cell developmentthrough interaction with DC-SIGN whereas Lex negative variants induce a strongTh1 cell response (Bergman et al., 2004). However, the functional role of DC-SIGN-Lex interaction needs to be further investigated in schistosome infections and thisinteraction clearly cannot explain the general potential of helminth to induce Th2responses. Although DC-SIGN has a broad carbohydrate recognition potential(Appelmelk et al., 2003), several helminths that we tested are not recognized byDC-SIGN (unpublished results). In addition, also other glycan antigens, such ascore-fucose and core-xylose that are not recognized by DC-SIGN (Van Liempt et al.,manuscript in preparation), can strongly induce Th2 responses (Faveeuw et al., 2003).It is expected that other factors, such as the presence or absence of maturationsignals, the glycan-carriers and cross-talk between lectins and Toll-like receptorswithin DCs, will contribute to define DC-function. The availability of otherTh2-inducing helminth-type glycoconjugates will be important to identify their recep-tors, and investigate the immunomodulatory properties of these glycoconjugates.Thus the molecular understanding how parasite glycans trigger DCs to induce Th2responses remains an issue of high priority for the next future.

ACKNOWLEDGEMENTS

We thank Anne Imberty for composing the molecular model of DC-SIGN withLewis-x (Fig. 3)

REFERENCES

Bakker, H., Agterberg, M., Van Tetering, A., Koeleman, C.A.M., Van den Eijnden, D.H., and Van Die, I.,

1994, A Lymnaea stagnalis gene, with sequence similarity to that of mammalian b1-

4-galactosyltransferases, encodes a novel UDP- GlcNAc:GlcNAc b-R b1-4-N-acetylglucosaminyl-

transferase. J. Biol. Chem. 269: 30326–30333.

Bakker, H., Schijlen, E., De Vries, T., Schiphorst, W.E., Jordi, W., Lommen, A., Bosch, D., and Van, D., I.,

2001, Plant members of the alpha1-3/4-fucosyltransferase gene family encode an alpha1-

4-fucosyltransferase, potentially involved in Lewis(a) biosynthesis, and two core alpha1-

3-fucosyltransferases. FEBS L ett. 507: 307–312.

Bergman, M.P., Engering, A., Smits, H.H., Van Vliet, S.J., Van Bodegraven, A.A., Wirth, H-P., Kapsenberg,

M.L. Vandenbroucke-Grauls, M.J.E., Van Kooyk, Y. and Appelmelk, B.J., 2004, Helicobacter pylori

modulates the T helper cell 1/T helper cell 2 balance through phase-variable interaction between

lipopolysaccharide and DC-SIGN. J. Exp. Med. 200: 979–990.

Cummings, R.D. and Nyame, A.K., 1996, Glycobiology of schistosomiasis. FASEB Journal 10: 838–848.

Cummings, R.D. and Nyame, A.K., 1999, Schistosome glycoconjugates. Biochim. Biophys. Acta 1455:

363–374.


Eberl, M., Langermans, J.A.M., Vervenne, R.A., Nyame, K.A., Cummings, R.D., Thomas, A.W., Coulson,

P.S., and Wilson, R.A., 2002, Antibodies to glycans dominate the host response to schistosome larvae

and eggs: is their role protective or subversive? J. Infect. Dis.ff 183: 1238–1247.

Engering, A., Geijtenbeek, T. B. H., and van Kooyk, Y., 2002, Immune escape through C-type lectins on

dendritic cells. T rends in Immunol. 23: 480–485.

Fabini, G., Freilinger, A., Altmann, F., andWilson, I.B.H., 2001, Identification of core alpha 1, 3-fucosylated

glycans and cloning of the requisite fucosyltransferase cDNA from Drosophila melanogaster.

Potential basis of the neural anti-horseradish peroxidase epitope. J. Biol. Chem. 276: 28058–28067.

Faveeuw, C., Mallevaey, T., Paschinger, K., et al., 2003, Schistosome N-glycans containing core alpha

3-fucose and core beta 2-xylose epitopes are strong inducers of Th2 responses in mice. Eur J Immunol

33: 1271–1281

Feinberg, H., Mitchell, D.A., Drickamer, K., and Weis, W.I., 2001, Structural basis for selective recognition

of oligosaccharides by DC-SIGN and DC-SIGNR. Science 294: 2163–2166.

Figdor, C.G., van Kooyk, Y., and Adema, G.J., 2002, C-type lectin receptors on dendritic cells and

langerhans cells. Nature Rev. Immunol. 2: 77–84.

Gantner, B.N., Simmons, R.M., Canavera, S.J., Akira, S., and Underhill, D.M., 2003, Collaborative induc-

tion of inflammatory responses by dectin-1 and Toll-like receptor 2. J. Exp. Med. 197: 1107–1117.

Geijtenbeek T.B., van Vliet, Sandra J., Koppel E.A., Sanchez-Hernandez M., Vandenbroucke-Grauls C.M.,

Appelmelk, B., and van Kooyk, Y., 2002, Mycobacteria target DC-SIGN to suppress dendritic cell

function. J Exp Med. 197: 7–17.

Geijtenbeek, T.B.H., van Kooyk, Y., van Vliet, S.J., Renes, M.H., Raymakers, R.A.P., and Figdor, C.G.,

1999, High frequency of adhesion defects in B-Lineage acute lymphoblastic leukemia. Blood 94:

754–764.

Guo, Y., Feinberg, H., Conroy, E., Mitchell, D.A., Alvarez, R., Blixt, O., Taylor, M.E., Weis, W.I., and

Drickamer, K., 2004. Structural basis for distinct ligand-binding and targeting properties of the

receptors DC-SIGN and DC-SIGNR. Nat. Struct. Mol. Biol. 11: 591–598.

’t Hart, B.A. and Van Kooyk, Y., 2004, Yin-Yang regulation of autoimmunity by DCs. T rends in Immunol.

25: 353–359.

Hokke, C. H. and Deelder, A. M., 2001, Schistosome glycoconjugates in host-parasite interplay. Glycoconj.

J. 18: 573–587.

Hokke, C.H., Neeleman, A.P., Koeleman, C.A.M., and Van den Eijnden, D.H., 1998, Identification of an

a3-fucosyltransferase and a novel a2-fucosyltransferase activity in cercariae of the schistosome

T richobilharzia ocellata. Glycobiology in press.

Jacobs, W., Deelder, A.M., and Van Marck, E., 1999, Schistosomal granuloma modulation. II. Specific

immunogenic carbohydrates can modulate schistosome-egg-antigen-induced hepatic granuloma

formation. Parasitol. Res. 85: 14–18.

Janeway, C.A., Jr. and Medzhitov, R., 2002, Innate immune recognition. Ann. Rev. Immunol. 20: 197–216.

Jankovic, D., Liu, Z., and Gause, W.C., 2002, Th1-and Th2-cell commitment during infectious disease:

asymmetry in divergent pathways. T rends in Immunol. 22: 450–457.

Jones, C., Virji, M., and Crocker, P.R., 2003, Recognition of sialylated meningococcal lipopolysaccharide

by siglecs expressed on myeloid cells leads to enhanced bacterial uptake. Mol. Microbiol. 49:

1213–1225.

Kawar, Z.S., van Die, I., and Cummings, R.D., 2002, Molecular cloning and enzymatic characterization of

a UDP-GalNAc:GlcNAcbeta -R b1, 4-N-acetylgalactosaminyltransferase from Caenorhabditis ele-

gans. J. Biol. Chem. 277: 34924–34932.MacDonald, A.S., Straw, A.D., Bauman, B., and Pearce, E.J., 2001, CD8– dendritic cell activation status

plays an integral role in influencing Th2 response development. J. Immunol. 167: 1982–1988.

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, Identification of a novel UDP-GalNAc:GlcNAcb-R

b1-4 N-acetylgalactosaminyltransferase from the albumen gland and connective tissue of the snail

Lymnaea stagnalis. Eur. J. Biochem. 227: 175–185.

Naus, C.W.A., van Remoortere, A., Ouma, J.H., Kimani, G., Dunne, D.W., Kamerling, J.P., Deelder, A.M.,

and Hokke, C.H., 2003, Specific antibody responses to three Schistosome-related carbohydrate

structures in recently exposed immigrants and established residents in an area of Schistosoma

mansoni endemicity. Infect. Immun.ff 71: 5676–5681.


Nyame, A.K., Debose-Boyd, R., Long, T.D., Tsang, V.C.W., and Cummings, R.D., 1998, Expression of Lexantigen in Schistosoma japonicum and S. haematobium and immune responses to Lex in infectedanimals: lack of Lex expression in other trematodes and nematodes. Glycobiology 8: 615–624.

Nyame, A. K., Kawar, Ziad S., and Cummings, R.D., 2004, Antigenic glycans in parasitic infections:

implications for vaccines and diagnosis. Arch. Biochem. Biophys. 426: 182–200.

Nyame, A.K., Leppanen, A., Bogitsh, B.J., and Cummings, R.D., 2000, Antibody responses to the fucosyl-

ated LacdiNAc glycan antigen in Schistosoma mansoni-infected mice and expression of the glycan

among schistosomes. Exp. Parasitol. 96: 202–212.

Nyame, A.K., Pilcher, J.B., Tsang, V.C.W., and Cummings, R.D., 1996, Schistosoma mansoni infection in

humans and primates induces cytolytic antibodies to surface Lex determinants on myeloid cells. Exp.Parasitology 82: 191–200.

Nyame, A.K., Pilcher, J.B., Tsang, V.C.W., and Cummings, R.D., 1997, Rodents infected with Schistosoma

mansoni produce cytolytic IgG and IgM antibodies to the Lewis x antigen. Glycobiology 7: 207–215.

Okano, M., Satoskar, A.R., Nishizaki, K., and Harn, D.A., Jr., 2001, Lacto-N-fucopentaose III found on

Schistosomamansoni egg antigens functions as adjuvant for proteins by inducing Th2-Type response.

J. Immunol. 167: 442–450.

Palucka, K. and Banchereau, J., 2002, How dendritic cells and microbes interact to elicit or subvert

protective immune responses. Curr. Opin. Immunol. 14: 420–431.

Sato, Sachiko and Nieminen, J., 2004, Seeing strangers or announcing ‘‘danger’’: Galectin-3 in two models

of innate immunity. Glycoconjugate J. 19: 583–591.

Srivatsan, J., Smith, D.F., and Cummings, R.D., 1992a, Schistosoma mansoni synthesizes novel biantennary

Asn-linked oligosaccharides containing terminal b-linked N-acetylgalactosamine. Glycobiology 2:

445–452.

Srivatsan, J., Smith, D.F., and Cummings, R.D., 1992b, The human blood fluke Schistosoma mansoni

synthesizes glycoproteins containing the Lewis X antigen. J. Biol. Chem. 267: 20196–20203.

Strasser, R., Mucha, J., Mach, L., Altmann, F., Wilson, I.B.H., Glxssl, J., and Steinkellner, H., 2001,

Molecular cloning and functional expression of b1,2-xylosyltransferase cDNA from Arabidopsis

thaliana. FEBS L etters 472: 105–108.

Van Dam, G.J., Bergwerff, A.A., Thomas-Oates, J.E., Rotmans, J.P., Kamerling, J.P., Vliegenthart, J.F.G.,ffff

and Deelder, A.M., 1994, The immunologically reactive O-linked polysaccharide chains derived from

circulating cathodic antigen isolated from the human blood fluke Schistosoma mansoni have Lewis x

as repeating unit. Eur. J. Biochem. 225: 467–482.

Van den Berg, T. K., Honing, H., Franke, N., van Remoortere, A., Schiphorst, W. E. C. M., Liu, F-T,

Deelder, A. M., Cummings, R. D., Hokke, C. H., and Van Die, I., 2004, LacdiNAc-glycans constitute

a parasite pattern for galectin-3-mediated immune recognition. J. Immunol. 173: 1902–1907.

Van Die, I, Cummings, R. D., van Tetering, A., Hokke, C. H., Koeleman, C. A. M., and van den Eijnden,

D. H., 2000, Identification of a novel UDP-Glc:GlcNAcb1, 4-glucosyltransferase in Lymnaea stag-

nalis that may be involved in the synthesis of complex-type oligosaccharide chains. Glycobiology

10: 263–271.

Van Die, I., Van Vliet, S.J., Nyame, A.K., Cummings, R.D., Bank, C.M.C., Appelmelk, B., Geijtenbeek,

T.B.H., and van Kooyk, Y., 2003, The dendritic cell-specific C-type lectin DC-SIGN is a receptor for

Schistosoma mansoni egg antigens and recognizes the glycan antigen Lewis x. Glycobiology 13:

471–478.

Van Liempt, E., Imberty, A., Bank, C.M., Van Vliet, S.J., Van Kooyk, Y., Geijtenbeek, T.B., and Van Die, I.

(2004). Molecular basis of the differences in binding properties of the highly related C-type lectinsffff

DC-SIGN and L-SIGN to Lewis X trisaccharide and Schistosoma mansoni egg antigens. J. Biol.Chem. 279: 33161–33167.

Van Remoortere, A., Hokke, C., Van Dam, G.J., Van Die, I., Deelder, A., and Van den Eijnden, D.H., 2000,

Various stages of Schistosoma express Lewisx, LacdiNAc, GalNAcb1-4 (Fuca1-3)GlcNAc andGalNAcb1-4(Fuca1-2Fuca1-3)GlcNAc carbohydrate epitopes: detection with monoclonal antibod-

ies that are characterized by enzymatically synthesized neoglycoproteins. Glycobiology 10: 601–609.

Van Remoortere, A., Vermeer, H.J., Van Roon, A.M., Langermans, J.A., Thomas, A.W., Wilson, R.A., Van

Die, I., Van den Eijnden, D.H., Agoston, K., Kerekgyarto, J., Vliegenthart, J.F., Kamerling, J.P., Van

Dam, G.J., Hokke, C.H., and Deelder, A.M., 2003a, Dominant antibody responses to

Fuca1-3GalNAc and Fuca1-2Fuca1-3GlcNAc containing carbohydrate epitopes in Pan troglodytes

vaccinated and infected with Schistosoma mansoni. Exp. Parasitol. 105, 219–225.


Van Remoortere, A., Bank, C.M.C., Nyame, A.K., Cummings, R.D., Deelder, A.M., and Van Die, I., 2003b,

Schistosoma mansoni-infected mice produce antibodies that cross-react with plant, insect, and mam-

malian glycoproteins and recognize the truncated biantennaryN-glycan Man3GlcNAc2-R.Glycobiology 13: 217–225.

Van Remoortere, A., Hokke, C.H., Van Dam, G.J., Van Die, I., Deelder, A.M., and Van den Eijnden, D.H.,

2000, Various stages of Schistosoma express Lewisx, LacdiNAc, GalNAcb1-4(Fuca1-3)GlcNAc, andGalNAcb1-4(Fuca1-2Fuca1-3)GlcNAc carbohydrate epitopes: detection with monoclonal antibod-

ies that are characterized by enzymatically synthesized neoglycoproteins. Glycobiology 10: 601–609.

Van Remoortere, A., Van Dam, G.J., Hokke, C.H., Van den Eijnden, D.H., Van Die, I., and Deelder, A.M.,

2001, Profiles of immunoglobulinM (IgM) and IgG antibodies against defined carbohydrate epitopes

in sera of Schistosoma-infected individuals determined by Surface Plasmon Resonance. Infect. Immun.ff

69: 2396–2401.

Van Roon, A.M., Van de Vijver, K.K., Jacobs, W., Van Marck, E.A., van Dam, G.J., Hokke, C.H., and

Deelder, A.M., 2004, Discrimination between the anti-monomeric and the anti-multimeric Lewis X

response in murine schistosomiasis.Microbes and Infection 6: 1125–1132.

Van Dam, G.J., Claas, F.H.J., Yazdanbakhsh, M., Kruize, Y.C.M., Van Keulen, A.C.I., Ferreira, S.T.M.F.,

Rotmans, J.P., and Deelder, A.M., 1996, Schistosoma mansoni excretory circulating cathodic antigen

shares Lewis-x epitopes with a human granulocyte surface antigen and evokes host antibodies

mediating complement-dependent lysis of granulocytes. Blood 88: 4246–4251.

Velupillai, P., Dos Reis, E.A., Dos Reis, M.G., and Harn, D.A., 2000, Lewisx-containing oligosaccharideattenuates schistosome egg antigen-induced immune depression in human schistosomiasis. Human

Immunology 61: 225–232.

Velupillai, P. and Harn, D.A., 1994, Oligosaccharide-specific induction of interleukin 10 production by

B220+ cells from schistosome-infected mice: a mechanism for regulation of CD4+ T-cell subsets.Proc. Natl. Acad. Sci. USA 91: 18–22.

Weis, W. I., Taylor, Maureen E., and Drickamer, K., 1998, The C-type lectin superfamily in the immune

system. Immunol. Rev. 163: 19–34.

Wuhrer, M., Dennis, R.D., Doenhoff, M.J., Lochnit, G., and Geyer, R., 2000, Schistosoma mansoni cercarialffff

glycolipids are dominated by Lewis X and pseudo-Lewis Y structures. Glycobiology 10: 89–101.

Zheng, Q., Van Die, I., and Cummings, R.D., 2002, Molecular cloning and characterization of a novel

a1,2-fucosyltransferase (CE2FT-1) from Caenorhabditis elegans. J. Biol. Chem. 277: 39823–39832.

6

THE MANNAN-BINDING LECTIN (MBL) PATHWAYOF COMPLEMENT ACTIVATION: BIOCHEMISTRY,BIOLOGY AND CLINICAL IMPLICATIONS

Jens Christian Jensenius

Department of Medical Microbiology and ImmunologyUniversity of Aarhus, Denmark

MBL is an oligomeric protein designed to recognize pathogen-associated molecularpatterns (PAMPs). It belongs to the family of collagen-like defence molecules charac-terized by being comprised of several subunits each composed of three polypeptides.The polypeptide of about 30 kD presents a collagen region attached to a globularhead containing the recognition structure. Each subunit thus presents three bindingsites. Hence the oligomer, typically comprising four subunits presents a substantialnumber of binding sites, each of low affinity, with the avidity and selectivity of themolecules being determined through multiple interactions. The collectins have sugar-binding, C-type-lectin globular domains, the ficolins have fibrinogen-like domains of

A B

A. The overall structure of the human collectins, SP-A and MBL, and L-ficolin. The schematic structures

represent interpretations of the electron micrographs. Two MBL oligomers are shown. The structures are

referred to as sertiform (sertula=small umbel )

B. Overview of the complement system with focus on the MBL pathway. Different MBL-MASP complexesffff

are involved in the formation of the C3 convertase, C4bC2b, and in direct activation of C3. Complexes

of ficolin and MASP also activate complement.

21



22 J. C. Jensenius

less defined specificity, complement C1q has domains recognizing structures onimmunoglobulins.

The biological importance of MBL was indicated when opsonin-deficient chil-dren with recurrent infections were found to be genetically deficient in MBL. Furtherinterest in this molecule was sparked by the observation of complement activationupon binding to carbohydrates. This activation is mediated by MBL-associatedserine proteases, MASPs, which have now been found associated also with ficolins.The MASPs present domain structures identical to those of C1r and C1s of theclassical complement pathway. MBL has affinity for terminal, non-reducing sugarspresenting horizontal 3- and 4-OH groups, e.g., glucose, mannose and fucose. Ficolinsare also cited as being lectins, but we have recently shown L-ficolin to be selectivefor acetyl groups on both sugars and on other molecules. The ligands for H- andM-ficolin remain undefined. Thus the widely adopted term ‘‘lectin complementpathway’’ appears inappropriate for MASP-mediated complement activation.

REFERENCE

Holmskov U, Thiel S, Jensenius JC. Collections and ficolins: humoral lectins of the innate immune defence.

Annu Rev Immunol. 2003;21:547–78.

7

KILLER CELL LECTIN-LIKE RECEPTORS AND THENATURAL KILLER CELL GENE COMPLEX

Ø. Nylenna1, L. M. Flornes1, I. H. WestgaardWW 1, P. Y. Woon2,C. Naper1, J. T. Vaage,VV 1, D. Gauguier2, J. C. Ryan3, E. Dissen1,and S. Fossum1

1Institute of Basic Medical SciencesUniversity of Oslo, Norway2Wellcome Trust Centre for Human GeneticsUniversity of Oxford, U.K.3Department of Arthritis and ImmunologyUniversity of California, USA

The natural killer cell gene complex (NKC), which maps to the distal parts of mousechromosome 6 and rat chromosome 4, and in the human to the short arm ofchromosome 12, encodes type 2 membrane receptors belonging to the group VC-type lectin superfamily (CLSF), lacking the evolutionary conserved calcium/saccharide binding amino acid residues found in other CLSF receptors. It containsall group V CLSF genes currently known, except Klrg1 which in rodents lies 6–7 Mbproximal to the NKC (see Fig. 1), and nothing but such genes, except Gabarapl1.Due to expansion of the Nkrp1 (Klrb) and in particular the Ly49 (Klra) multigenefamilies the complex is particularly large in rodents, where it can be divided intothree parts: a proximal part encoding Nkrp1 and Clr receptors, a middle partencoding a variety of group V CLSF receptors, and a large distal part encodingLy49 receptors. In the rat the NKC spans 3.3 Mb and is predicted to contain 67CLSF loci (including some pseudogenes), distributed as indicated in the figure. Tothe extent ligands are known, the NKC encoded receptors do not bind saccharides(with the exception of Dectin-11), but rather MHC class I and related ligands (rev.in2). Recently, mouse Nkrp1d and -f were shown to bind Clr molecules, providingthe first example of CLSF receptor/ligand pairs3. Functionally the NKC encodedreceptors have opposing regulatory roles on leukocyte activation, the activatingmediating their effects via protein tyrosine kinases and the inhibitory via proteinfffftyrosine phosphatases. Close to, but distinct from the NKC lies a smaller genecomplex (called APLEC4) encoding opposing regulatory leukocyte receptors, but

23



24 Ø. Nylenna et al.

Figure 1. The chromosal regions in man, mouse and rat containing the NKC (lower box) and the adjacent

APLEC plus the Klrg1 gene. In the rat the NKC is predicted to contain 67 CLSF loci (based on the rat

genome sequence from the BN strain. Numbers to the left indicate distances in Mb, numbers to the right

(bold font) indcate number of loci.

preferentially expressed by professional antigen presenting cells and neutrophils andwith the calcium/saccharide binding amino acid residues conserved (hence classifiedas group II CLSF receptors). The presentation will concentrate on the rat, wherewe now have cloned close to all of the genes, and in particular on the highly dynamicLy49 gene region, to which the functional alloreactivity gene Nka previously wasmapped5,6.

REFERENCES

1. Brown, G.D. et al. J Exp Med 196, 407–412 (2002).

2. Yokoyama,W.M. & Plougastel,B.F.M. Nature Reviews Immunology 3, 304–316 (2003).

3. Iizuka,K. et al. Nature Immunol 4, 801–807 (2003).

4. Flornes, L.M. et al. Immunogenetics, in press (2004)

5. Dissen,E. et al. J. Exp. Med. 183, 2197–2207 (1996).

6. Nylenna, Ø. et al. Submitted.

8

GLYCOSYLATION INFLUENCES THE LIGANDBINDING ACTIVITIES OF MANNOSE RECEPTOR

Yunpeng Roc Su1,2, Clarence Tsang1, Talitha Bakker2,James Harris2, Siamon Gordon2, Raymond A. Dwek1,Luisa Martinez-Pomares2 and Pauline M Rudd1

1Glycobiology Institute2Sir William’s Dunn School of PathologyOxford University, OX1, 3QU, Oxford, UK

Murine mannose receptor (MR) contains seven N-linked and three O-linked oligosac-charides and differential binding properties have been described for MR isolatedfffffrom the liver and the lung. We hypothesised that these different binding activitiesffffcould be controlled by glycosylation. In this study the relationship between MRglycosylation and its function has been investigated using MR transductants gener-ated in both wild type CHO cells and glycosylation-deficient LEC cells. The investi-gation shows that glycosylation does not affect the subcellular distribution,ffffproteolytic processing and endocytic capacity of the receptor, but has a major effectffffin its binding capacity. Cells bearing MR modified with Man5GlcNAc2 sugars(Man-5 MR) completely lost its mannose-internalisation activity, which is associatedwith CRD4–5 of MR. In agreement with this observation purified soluble Man-5MR lost the capability to bind mannan in vitro. The desialylation modification ofMR also results in a 70% reduction of cellular internalisation activity and a lowefficient mannan binding activity in vitro. However, cells bearing MR modified withMan5GlcNAc2 sugars or desialylated glycans do retain their sulphated sugarinternalisation activity, which is associated to the cysteine-rich (CR) domain.Interestingly, in vitro SO4-3-galactore-PAA binding study indicated desialylated MRhas better affinity than wild-type MR. Subsequent gel filtration and BIAcore studiesshowed that desialylated MR tend to form self-associated structure and multiplepresentation of CR domain could enhance its affinity to sulphated sugars dramatic-ally. These results, for the first time, suggest a role for glycosylation, especiallyterminal sialylation of MR, in manipulating its dual ligand binding activities in vivo.



26 Yunpeng Roc Su et al.

REFERENCES

Fiete, D. J. et al., 1998, A cysteine-rich domain of the ‘‘mannose’’ receptor mediates GalNAc-4-SO4 binding.

Proc Natl Acad Sci U S A. 2089–2093.

Feinberg, H. et al. (2000). Structure of a C-type carbohydrate recognition domain from the macrophage

mannose receptor. J Biol Chem. 275, 21539–21548.

Rudd, P. M. et al. (1997). Oligosaccharide sequencing technology. Nature. 205–207

Stanley, P. (1989). Chinese hamster ovary cell mutants with multiple glycosylation defects for production

of glycoproteins with minimal carbohydrate heterogeneity. Mol Cell Biol. 377–383.

9

HUMAN IMMUNOGLOBULIN GLYCOSYLATIONAND THE LECTIN PATHWAY OF COMPLEMENTACTIVATION

James N. Arnold1, Louise Royle2, Raymond A. Dwek2,Pauline M. Rudd2, and Robert B. Sim1

1MRC Immunochemistry Unit2Oxford Glycobiology InstituteDepartment of BiochemistryUniversity of OxfordSouth Parks Road, Oxford OX1 3QU, UK

1. INTRODUCTION

Immunoglobulins are the major secretory products of the adaptive immunesystem. They are glycoproteins which are found in all higher vertebrates (mammals,birds, reptiles, amphibians, bony and cartilaginous fish, but not in jawless fish(agnatha)) (Litman et al., 1999). In humans there are five classes IgG, IgM, IgA, IgEand IgD. The immunoglobulins share similar structures (Fig. 1). Each immunoglob-ulin molecule is composed of two identical disulphide bridged class-specific heavychains, each disulphide bridged to a light chain of which there are two isoformsnamed k and l. Both heavy and light chains are composed of regions called immuno-globulin domains. The immunoglobulin fold/domain is about 105–120 amino acidslong and is composed of b-sheet secondary structure (Amzel and Poljak, 1979).The role of immunoglobulins is to bind to antigens via their N-terminal (variable

amino acid sequence) domains and to mediate effector functions, such as activationffffof complement (Malhotra et al., 1995; Roos et al., 2001) or binding to receptors viatheir constant (invariable sequence) domains (Mimura et al., 2000; Shields et al.,2001). During immunoglobulin synthesis, rearrangement of gene segments andsomatic mutation creates variation in amino acid sequence in the N-terminal domains(named VH and VL domains for Variable Heavy and Light chains respectively). Thelight chains have one V domain and one constant sequence domain (CL). Thesequence of all l chain C domains is the same, and the sequence is homologous to

27



28 J. N. Arnold et al.

Figure 1. Immunoglobulin Structure

a) The structure of IgG showing the Variable Heavy (VH) and Constant Heavy (CH), Variable Light

(VL) and Constant Light (CL) domains. The diagram identifies the Fab, Fc and flexible hinge regions of

the molecule. This hinge varies in length between the different immunoglobulin classes and is replaced byffff

additional CH domain in IgE and IgM. The approximate positioning of the Asn-297 N-linkage site for

glycans is marked. b) Diagrammatic representation of IgG1, IgD, IgA1, IgE and IgM showing N- and O-

linked glycan positions, and inter-chain disulphide bridges. The domains themselves contain intra-domain

disulphide bridges, although these are not marked. IgM circulates in the serum in both pentameric and

hexameric forms, in which the monomeric units are disulphide bridged together. Pentameric IgM contains

a single J chain but hexameric IgM does not (Weirsma et al., 1998).

Human Immunoglobulin Glycosylation 29

the C domain shared by all k chains. Heavy chains have 3 or 4 C domains. Thesequences of the C domains are class or subclass specific, i.e. all IgGs have identicalconstant regions, as do all IgMs. Each clone of B lymphocytes secretes only oneimmunoglobulin molecule, which has V regions unique to that particular B cellclone. Total IgG isolated from human serum therefore contains 4 subclasses, eachwith similar but distinct constant regions, and with 105–106 different V regionffffsequences.IgM and IgD occur both as soluble forms (in serum) and membrane-bound

forms on B lymphocytes (Van Boxel et al., 1972). The membrane-bound forms havean additional trans-membrane segment, C-terminal to the constant regions. IgA,IgG, IgE are all soluble molecules: IgG is the most abundant in serum (10–15 mg/ml),while IgA is the most abundant immunoglobulin overall. Most IgA is secretedthrough epithelia into the mucous lining of the gastrointestinal and respiratory tract,and into tears, saliva and milk (Norderhaug et al., 1999). The secreted form isgenerally dimeric and contains an extra glycosylated polypeptide chain, SC (SecretoryComponent) and glycosylated 16KDa J chain (Johansen et al., 2001; Royle et al.,2003), which is also found in pentameric forms of IgM (Wiersma et al., 1998). Thesingle J chain is disulphide bridged to two C-termini of both IgM and IgA molecules(Wiersma et al., 1998; Royle et al., 2003). IgA in serum is predominantly monomericbut also forms dimers and higher polymers (Delacroix et al., 1982; Roos et al., 2001).IgE is the lowest abundance immunoglobulin, occurring as a monomer at<1 mg/ml.IgD also occurs as a low abundance monomer in serum at <30 mg/ml, while IgMis at high concentrations (~2.5 mg/ml). IgM occurs predominantly as pentamersand hexamers, although a small amount of monomer also circulates (Sørensenet al., 1999).

The different classes of immunoglobulin are distinct in their major effff ffectorfffffunctions. IgM is principally associated with complement classical pathway activationvia binding of C1q (Wiersma et al., 1998). IgG also activates complement via classical(Duncan and Winter, 1988) and alternative pathways (Anton et al., 1989) andmediates ADCC (Antibody Dependent Cell Cytotoxicity) (Sarmay et al., 1992). IgEis associated with mast cell and basophil stimulation in allergic conditions. IgA insecretions may act mainly to agglutinate (immobilise) or neutralise micro-organisms(Lamm, 1997). No effector functions have been identified for IgD.ffff

In addition to their enormous diversity of amino acid sequences and antigen-binding specificity, immunoglobulins display considerable diversity in the locationand number of glycosylation sites (both N- and O-linked) and great diversity inglycan structure. The glycans attached to the immunoglobulins are important forimmunoglobulin solubility (Tarentino et al., 1974), subcellular transport and secre-tion (Gala and Morrison, 2002), conformation (Mimura et al., 2000), binding to Fcreceptors (Mimura et al., 2000), normal plasma clearance (Skockert, 1995) andcomplement activation (Malhotra et al., 1995). This chapter discusses both the glycanstructures that are attached to the normal human serum immunoglobulins and theirpotential roles in complement activation through the binding of the serum ‘recogni-tion’ lectin, Mannan Binding Lectin (MBL), and the subsequent activation of thelectin pathway of the complement system.


2. GLYCOSYLATION OF THE IMMUNOGLOBULINS

2.1. IgG

There are four subclasses of IgG, named IgG1–4, that differ in their heavy chainffffconstant region sequence and disulphide bridging. The subclasses have distinctiveglycan pools (Jefferisffff et al., 1990).

All IgGs have a single N-linked glycosylation site on each heavy chain in theCH2 domain at Asn-297 (Fig. 1). There are no conserved glycosylation sites in thelight chain or variable regions of the heavy chain. The glycan population attachedat Asn-297 contains three sets of glycoforms termed IgG-G0, -G1 and -G2 (Fig. 2b).The IgG-G2 biantennary glycans occupying Asn-297 have two arms that bothterminate in galactose residues. This set of glycoforms accounts for approximately16% of total IgG glycans. Approximately 35% are IgG-G1, which lack a terminalgalactose residue on one biantennary arm, exposing a GlcNAc residue. IgG-G0glycans make up 35% and neither biantennary arm contains a galactose residue.The final 14% of serum IgG glycans consist of IgG-G2 or -G1 glycoforms whichare sialylated. Within the glycans of IgG there is a diversity of structures caused bythe presence of bisecting GlcNAc residues (B in Fig. 2a) (approximately 30% of totalIgG1 glycan pool), core fucose (Fc in Fig. 2a) (approximately 70% of the total IgG1glycan pool) and sialylation of the terminal 1,3 arm galactose residues (S in Fig. 2a)(14% of total IgG1 glycan pool) (Butler et al., 2003)).

IgG1 is the most abundant subclass of IgG in the serum. IgG2 and IgG3 havea preferred linkage of the galactose residues to the a1,3 arm mannose, whereas IgG1has preferential linkage of galactose to the a1,6 arm mannose. IgG4 is reported tocontain predominantly fully galactosylated structures (Jefferisffff et al., 1990).There is considerable amino acid sequence diversity in the variable regions, and

N-linked glycosylation sites can occur in the variable regions. These are relativelyrare. A recent survey of heavy chain variable region cDNA sequences showed thatonly 7 out of 75 (9.3%) had a potential N-linked glycosylation site in the variableregion (Zhu et al., 2002). The glycans that occupy these sites are predominantlysialylated structures, with a high incidence of bisecting GlcNAc residues (Youingset al., 1996.; Wormald et al., 1997).

2.2. IgM

IgM is found predominantly in the serum as a pentameric structure disulphidebridged at the CH3 domains and at the tail piece (a flexible region following theCH4 domain) and believed to form a ring structure. Pentameric IgM also has a Jchain that contains a single N-linked glycosylation site. IgM can also adopt ahexameric structure that contains no J chain (Wiersma et al., 1998). IgM heavychain (m chain) has five N-linked glycosylation sites at Asn-171, Asn-332, Asn-395,Asn-402, and Asn-563. Asn-402 and Asn-563 have been shown to be occupied byoligomannose structures (Chapman and Kornfeld, 1979; Wormald et al., 1991). Theother N-linked glycosylation sites on each m chain in normal human serum IgM areoccupied predominantly by complex biantennary glycans. The most predominantglycan is FcGlcNAc2A2G2S1 (26% of total glycan pool). Sialylated structures


Figure 2. Glycan Structure and IgG Glycoforms.

a) Shows the general nomenclature used to describe sugar residues, bond angles and sugar linkages of the

different glycan structures that occupy glycoproteins. b) Shows the predominant glycan structures thatffff

occupy the Asn-297 site in IgG. The glycans shown may also vary by the presence of absence of a core

Fucose and/or bisecting GlcNAc.


Figure 3. IgA Glycosylation Types.

IgA has two subclasses, IgA1 and IgA2, and both have N-linked glycosylation at Asn-263 and Asn-459.

IgA1 contains nine potential O-linked sites in the hinge region, of which five or six have been shown to

be occupied. *The sixth O-linked site occupies one or more of Ser-224, Thr-233, Ser-238 or Ser-240 (Tarelli

et al., 2004). IgA2 has no potential O-linked sites in its hinge region. IgA2 is subdivided into IgA2m(1)

which has two additionalN-linked glycosylation sites in the CH1 domain and CH2 domain, and IgA2m(2)

which has these additional N-linked sites but also a third additional CH1 domain N-linked glycosyl-

ation site.

(61.8%), core fucosylated (65%) and bisected structures (38%) are present in thetotal glycan pool (J.Arnold unpublished data).

2.3. IgA

IgA has two conserved N-linked glycosylation sites, at Asn-263 in the CH2domain and Asn-459 located in the 18 amino acid tail piece on each a chain. Thereare two subclasses of IgA designated IgA1 and IgA2. IgA2 has two forms thatcontain two (IgA2m(1)) or three (IgA2m(2)) extra conserved N-linked glycosylationsites respectively (Fig. 3).IgA occurs in several different oligomeric forms, and is present both in serumffff

and in secretions. Serum IgA and Secretory IgA (SIgA), have distinct populationsof glycan structures.The 23 amino acid hinge region in IgA1 contains nine potential O-linked sites

of which five have been shown to be occupied (Mattu et al., 1998: Baenziger andKornfeld, 1974b). These sites are Thr-228, Ser-230, Ser-232, with Thr-225 and Thr-236


Figure 4. Core I Structures, Neutral, Mono-, and Di-Sialylated.

The neutral, mono- and di-sialylated Core I O-linked glycan structures, that have been identified on serum

IgA1 and also IgD hinge regions. The nomenclature is explained in Fig. 2.

partially occupied (Mattu et al., 1998). Recently a sixth occupied O-linked site atone or more of Ser-224, Thr-233, Ser-238 or Ser-240 (Tarelli et al., 2004) has beenidentified. No O-linked glycans have been identified on IgA2.

2.3.1. Serum IgA

Serum IgA consists mainly of IgA1. IgA1 and IgA2 contain similar N-linkedglycan structures (Endo et al., 1994; Royle et al., 2003). Over 80% of the glycansare di-galactosylated bi-antennary complex glycans. Less than 10% are tri- andtetra-antennary structures (Mattu et al., 1998). Sixty four percent of the glycanstructures are sialylated and 95% of these are linked a2–6 to galactose. The predom-inant glycan is GlcNAc2A2G2S2 (24%). The glycan pool has 36% of the glycanscontaining a core fucose residue and 25% containing a bisecting GlcNAc residue.The oligosaccharides attached to the Fab in IgA2 differ from those that occupy theffffFc in for example, the presence of triantennary structures and outer-arm fucoseresidues such as GlcNAc2A3G3FS3 which accounts for 3.7% of total Fab glycanpool (Mattu et al., 1998).The O-linked glycans on the heavy chain of IgA1 have been identified (Mattu

et al., 1998, Field et al., 1994 and Rudd et al., 1994)). The hinge is predominantlyoccupied by mono-sialylated core I structures (37%) and neutral core I structures(31%) (Mattu et al., 1998) (Fig. 4).

2.3.2. Secretory IgA

SIgA is a dimer, held together with a J chain (which has one N-linked site) andSecretory Component (SC). The SC is the extracellular portion of the epithelialpolymeric Ig receptor (pIgR), and is required for transcytosis of the IgA across theepithelium to the mucosal surface. The SC has seven N-linked glycosylation sites.SIgA contains both IgA1 and IgA2 populations.SIgA is present in mucosal secretions such as colostrum and milk and can bind

to microorganisms, their metabolic products and toxins, preventing their attachment


to the epithelium and facilitating their excretion. This process is known as immuneexclusion (reviewed by Lamm, 1997).TheN-linked glycans from the heavy chain of colostrum SIgA consist of approxi-

mately 15% sialylated structures (solely a2–6 linked sialic acids), with over 75% ofstructures containing a bisecting GlcNAc and 50% being core fucosylated.Oligomannose structures account for 12% of the N-linked glycan pool. There is alack of glycan processing of the N-linked glycans on the a-chain of SIgA, as only20% of structures are fully galactosylated, and 66% have an exposed terminalGlcNAc residue. The major structures occupying SIgA heavy chain areFcGlcNAc2A2B (30%), GlcNAc2A2B (21%) and FcGlcNAc2A2BG1 (8%) (Royleet al., 2003).There is a large diversity of O-linked glycan structures on the heavy chain of

SIgA1, which contains over 50 different structures of up to 15 residues in size (Royleffffet al., 2003), in contrast to the restricted pool of structures present on serum IgA1.The glycans occupying the J chain single N-linked site are predominantly

sialylated biantennary structures (75%). Fifty percent of all structures are corefucosylated, and 50% of the neutral structures contain a bisecting GlcNAc residue(Royle et al., 2003). Interestingly no bisecting GlcNAc is present on the sialylatedstructures (Royle et al., 2003).The sevenN-linked sites of the SC are occupied by a large diversity of structures,

many of which are not found on the immunoglobulin heavy chains, for example,outer-arm fucosylated glycans. The presence of these structures may be explainedpartially by the fact that epithelial cell glycosylation machinery glycosylates the SC,whereas the plasma cell glycosylates the immunoglobulin. The SC N-linked glycansare predominantly bi-antennary structures. Tri-antennary (11.7%) and tetra-anten-nary (<1%) structures are also present (Royle et al., 2003). Over 70% of the glycansare sialylated, predominantly mono-sialylated structures, and over 65% of the glycanscontain a core fucose (Royle et al., 2003).

2.4. IgD

IgD has three N-linked glycosylation sites in the Fc at Asn-354, Asn-445,Asn-496 (Takahashi et al., 1982). Asn-354 in the CH2 domain is occupied solely byoligomannose structures (GlcNAc2Man5-9 ) (Mellis and Baenziger, 1983a; Arnoldet al., 2004) which represent 34% of the total glycans. The predominant oligomannosestructure is GlcNAc2Man8 . Glucosylated mannose structures (GlcNAc2Man9Glc1 ,GlcNAc2Man8Glc1 and GlcNAc2Man7Glc1 ) are also present (Arnold et al., 2004;Mellis and Baenziger, 1983a). The other 66% of glycan structures have been shownto terminate in galactose or sialic acid. These glycans occupy the two CH3 domainN-linked glycosylation sites Asn-445 and Asn-496. At these two CH3 N-linked sites71% of the oligosaccharides are sialylated; both mono- (53%) and di-sialylated(47%) glycans have been identified. Twenty nine percent of glycans terminate ingalactose residues, 50% contain core fucosylation and 50% of the glycans containa bisecting GlcNAc (Arnold et al., 2004) at these two sites.The hinge region of IgD contains several potential O-linked glycosylation sites.

In an IgDmyeloma protein IgD:WAH,O-linked glycans occupy Ser-106 and Thr-126,-127, -131 and -132, although it is uncertain if Thr-131 and –132 are both occupied


(Mellis and Baenziger, 1983b; Takahashi et al., 1982). Another myeloma IgD:NIG-65contains seven O-linked glycosylation sites; the five identified in IgD:WAH and alsoSer-110 and Thr-113 (Takayasu et al., 1982). The O-linked glycans present on thehinge region are solely Core I structures: di-, mono-sialylated and neutral structures(Fig. 4) (Arnold et al., 2004; Mellis and Baenziger, 1983b).

2.5. IgE

IgE has seven N-linked glycosylation sites in the e chain at Asn-140, Asn-168,Asn-218, Asn-265, Asn-371, Asn-383, Asn-394 (Dorrington and Bennich, 1978). TheAsn-394 N-linked glycosylation site is occupied solely by oligomannose structures(Dorrington and Bennich, 1978; Baenziger and Kornfeld, 1974b). The predominantoligomannose structure is GlcNAc2Man5 (8.3% of the total glycan pool). The othersix exposed glycosylation sites on each e chain are occupied predominantly withsialylated glycan structures (46%mono- 42% di-sialylated structures), 12% galactoseterminating structures, 68% core fucosylated and 14% bisected structures (Arnoldet al., 2004).

3. MANNOSE BINDING LECTIN (MBL) AND THE LECTINPATHWAY OF COMPLEMENT ACTIVATIONAA

3.1. MBL

MBL (Fig. 5) is a glycoprotein, also known as Mannan/Mannose BindingProtein and is member of the collectin family of proteins (Malhotra et al., 1994).Collectins are large oligomeric proteins with multiple lectin domains and collagenousregions. MBL is synthesized in the liver and secreted into the blood stream. MBLis an important component of the innate immune system, which binds calcium-dependently to sugars that have hydroxyl groups on the carbon-3 and carbon-4orientated in the equatorial plane of the pyranose ring (Weis et al., 1992). This givesMBL affinity for mannose, fucose and N-acetyl glucosamine (GlcNAc) (Turner et al.,1996). This specificity allows MBL to bind to sugar arrays on the surfaces ofmicroorganisms, including bacteria, viruses and fungi (Holmskov et al., 1994), butnot to human glycoprotein glycans, the structures of which generally terminate ingalactose or sialic acid. MBL has a structure and function similar to that of C1q,the recognition molecule that initiates the classical pathway of complement. MBLbinds to sugar residues via the Carbohydrate Recognition Domain (CRD) ( lectin)heads. The affinity of a single CRD for carbohydrate is very weak (10−3M) (Iobstet al., 1994). Multiple CRD binding leads to a much greater avidity. Levels of MBLin human serum vary greatly between individuals (Turner, 1996), from below 50ng/mlto above 10ug/ml. The variation of MBL levels is caused by several identifiedpolymorphisms in the coding sequence and promoter regions of the MBL gene(Madsen et al., 1995). The coding sequence polymorphisms disrupt the Gly-X-Yrepeat that is found in the collagenous region destablilising the collagen triple helixformation (Sumiya et al., 1991: Lipscombe et al., 1992), and consequently heterozy-gotes have low levels of MBL in the blood. Low levels of MBL have been linked tosevere and recurrent infections in children (Summerfield et al., 1997).


Figure 5. Structure of MBL.

MBL is composed of identical 25kDa polypeptides that form a trimer through the formation of a triple

helix of the collagen-like regions that is the basis of the MBL subunit (or monomer). This subunit can

then disulphide bridge at its N-terminus to form higher order structures. MBL circulates in the serum

mainly as a hexameric molecule (i.e. six subunits, 18 polypeptide chains). The collagen-like region is

attached to a Carbohydrate Recognition Domain (CRD) which binds to sugar arrays that have hydroxyl

groups on the carbon-3 and carbon-4 orientated in the equatorial plane of the pyranose ring (Weis

et al., 1992).

MBL participates in the host defense response through two major pathways.Firstly, it acts directly as an opsonin, promoting phagocytosis of foreign material towhich it has bound. There are several candidate receptors through which this processmay be mediated. The main candidate receptor is cell surface calreticulin (Sim et al.,1998; Ogden et al., 2001), but there is also evidence for the participation of comple-ment receptor 1 (CR1: CD35) (Ghiran et al., 2000). The second pathway throughwhich MBL functions is by triggering the lectin pathway of complement activationvia MBL associated serine protease-2 (MASP-2) (Vorup-Jensen et al., 2000; Hajelaet al., 2002).

3.2. MASPs

MBL circulates in the serum bound to the serine protease pro-enzymes, MASPs,of which three have been identified to date; MASP-1, MASP-2 (Matsushita et al.,1992; Thiel et al., 1997) and MASP-3 (Dahl et al., 2001). The MBL-MASP complexwas shown to be capable of consuming the complement components C2 and C4(Ikeda et al., 1987). It is now generally accepted from recombinant protein workthat MASP-2 is solely responsibly for the cleavage of C2 and C4 to produce C4b2a(Vorup-Jensen et al., 2000). This provides MASP-2 with a function similar to thatof C1s in the C1 complex. The biological roles for MASP-1 and MASP-3 arecurrently unknown. MASP-1 has been shown to cleave ‘dead’ C3 (C3 in which thethiolester bond has hydrolyzed) at a slow rate. Cleavage of physiological ‘live’ C3(C3 in which the thiolester bond is intact) occurs at a very slow rate, suggested to


Figure 6. The Complement System.

The lectin and classical pathways rely on cleavage of complement protein C4, forming C4b, to which C2

binds and is cleaved, that leads to the formation of C4b2a, a C3 convertase that activates C3. C3 is cleaved

into C3a and C3b, which is further cleaved to form the iC3b opsonin. Activation of C3 leads to the

formation of the membrane attack complex which causes cell lysis. The alternative pathway relies on

preformed C3b, or C3(H2O) which forms spontaneously at a slow rate. C3b binds factor B, which iscleaved by Factor D to form another C3 convertase, C3bBb. The C3 convertases are inactivated by decay

accelerating factor, Factor H, C4b-binding protein and complement receptor I, which speed up the dissoci-

ation of the convertase. C3b and C4b when bound by cofactors such as Factor H are cleaved by Factor

I and inactivated. The C3 convertases have naturally short half lives in the circulation.

be too slow to be physiologically important (Hajela et al., 2002). MASP-1 alsocleaves Factor XIII (plasma transglutaminase) and fibrinogen, two substrates ofthrombin, potentially implicating MASP-1 in localized coagulation (Hajela et al.,2002).

3.3. Complement and the lectin pathway of complement activation

The complement system (Fig. 6) is a major part of the innate immune responsethat eliminates foreign and altered-self cells by opsonisation and lysis. It is the body’sfirst line of defense against infectious agents. The complement system recognizesforeign matter through proteins with specific binding affinities to potential PathogenAssociated Molecular Patterns (PAMPs) including lipopolysaccharide, lipoproteins,peptidoglycan, lipoarabinomannan and oligosaccharide and charge arrays. The bind-ing of ‘recognition’ proteins MBL and C1q leads to the activation of the complementcascade which is controlled and propagated through serine proteases and regulateddirectly by a serpin, C1-inhibitor (Cooper, 1985) that binds and inactivates thesecascade triggering proteases. There are three routes of complement activation; theclassical, alternative and lectin pathways. The classical pathway is triggered by theC1 complex. The C1 complex is composed of C1q and 2 each of the serine proteasesC1r and C1s (Arlaud et al., 1987).


Figure 7. MBL and the Immunoglobulins

A summary of the interaction of MBL with the immunoglobulins.

MBL is the recognition molecule of the lectin pathways of complement activa-tion. Binding of MBL to a target activates MASPs. Activated MASP-2 cleaves thecomplement protein C4, forming C4b, to which C2 binds and is also cleaved byMASP-2, leading to the formation of C4b2a, a C3 convertase that activates C3. C3is cleaved into C3a and C3b, which is further cleaved to form the iC3b opsonin.Activation of C3 leads on to the formation of the membrane attack complex (MAC)that causes cell lysis.

4. THE INTERACTION OF MBL WITH THE IMMUNOGLOBULINS

The immunoglobulins contain populations of glycans, some of which terminatein mannose or GlcNAc which are potential binding ligands for lectin-like recognitionproteins of the innate immune system, such as MBL, macrophage Mannose Receptorand the surfactant proteins SP-A and SP-D. The known interactions of MBL withimmunoglobulins are summarised in Fig. 7.The glycans of IgG have restricted motion because of the terminal galactose

residues attached to the glycan structures. The IgG CH2 domain has a hydrophobicarea on the peptide surface of each heavy chain. Galactoses attached to the a1,6arm of the glycan interact with this region, and this together with >80 otherinteractions such as hydrogen bonding and van der Waals interactions holds theglycan in contact with the protein surface, which also prevents further processing toattach terminal sialic acid (Wormald et al., 1997). The glycans therefore have limited


mobility. In IgG-G0, the glycans do not have terminal galactose residues, butterminal GlcNAc residues. These glycans are more mobile, as the glycan-proteininteractions are not sufficient to hold the glycan anchored to the protein surface(Wormald et al., 1997). MBL has been shown to bind to the terminal GlcNAcresidues of the IgG-G0 glycans (Malhotra et al., 1995). IgG-G0 glycoforms havebeen shown to increase dramatically in Rheumatoid Arthritis (RA) (Parekh et al.,1985). This increase has been shown to correlate with disease activity (Rook et al.,1991). Garred et al. (2000) correlated MBL levels with disease onset and progressionin RA patients. This was consistent with the suggestion by Malhotra et al. (1995)that activating the lectin pathway of the complement system could be a potentialroute to additional inflammation in RA.

IgD has the same domain structure as IgG, however the glycans found at theN-linked site homologous to that in IgG (Asn-297 in IgG and Asn-354 in IgD) aresolely oligomannose structures. Although these are potential ligands for MBL, MBLdoes not bind IgD (Arnold et al., 2004). The oligomannose glycans at Asn-354 areinaccessible to MBL because the complex glycans occupying Asn-445 on the CH3domain block the access to the oligomannose glycans (Arnold et al., 2004).MBL has been shown to interact with certain polymeric types of IgA but not

SIgA (Roos et al., 2001; Royle et al., 2003). MBL binds to polymeric and dimericforms of IgA with the highest avidity, but MBL does not bind to monomeric serumIgA (Roos et al., 2001). The glycans in IgA with which MBL is interacting have notbeen identified although it has been inferred from models that all the glycans onIgA (but not SIgA) are exposed and could potentially bind (Mattu et al., 1998).SIgA contains a large array of glycans terminating in GlcNAc residues (Royle

et al., 2003). However these structures are masked from lectin binding by the SCwhich wraps around the IgA heavy chains. The SC itself contains predominantlysialylated complex glycans (Royle et al., 2003). The SC structure blocks access ofMBL to the IgA glycans, although it has been suggested that these may be revealedwhen SC binds to pathogens (Royle et al., 2003).

IgE has a different domain structure from IgG, IgD and IgA. The hinge peptidesffffare replaced by immunoglobulin domains which form a rigid dimer. The crystalstructure of the Fc and CH2 hinge domain showed an asymmetrically bent quatern-ary structure, where the CH2 domain bends over one side of the Fc (Wan et al.,2002). Oligomannose structures occupy Asn-394 (homologous site to Asn-297 inIgG and Asn-354 in IgD) (Dorrington and Bennich, 1978; Arnold et al., 2004). MBL,however, does not bind IgE (Arnold et al., 2004). The access to these oligomannoseglycans is prevented because of the CH2 hinge domain which completely blocksaccess to the oligomannose glycans from one side. The CH2 hinge domain is proposedto ‘flip’ between two bent quaternary conformations with the CH2 hinge domainson either side of the Fc domain, preventing access to the oligomannose glycans fromboth sides of the Fc (Arnold et al., 2004).IgM is found in the serum as a pentamer and a hexamer. The IgM monomer

unit has a very similar structure to that of IgE, with an Ig domain replacing thehinge region. IgM, however, contains oligomannose glycans at twoN-linked glycosyl-ation sites, located at Asn-402, homologous to the Asn-394 in IgE (and Asn-297 inIgG and Asn-354 in IgD) and at Asn-563 at the C-terminus (Wormald et al., 1991).It has been shown that immobilised human IgM does not bind MBL on microtitre


plates (Roos et al., 2003). The oligomannose glycans at Asn-402 are predicted to beinaccessible on the basis that it is similar in structure to IgE, where the CH2 domain‘flips,’ between two bent quaternary conformations (J. Arnold and M. Wormald,unpublished data). The accessibility of the tail piece oligomannose glycans is currentlyunknown and under investigation. It may be the case that the structural change thatoccurs upon IgM binding to antigen (referred to as the staple form of IgM), maypresent the oligomannose sugars to MBL for binding. There have been reports (seeFig. 7) of human IgM binding to rat MBL (Koppel and Solomon, 2001) and human,bovine and murine IgM binding to rabbit MBL (Nevens et al., 1992) (Fig. 7). In thelatter case it appears that MBL may be binding only a small subpopulation ofhuman IgM (J.Arnold, unpublished data).

5. CONCLUSIONS

The glycans attached to the immunoglobulins have a great diversity in structure,location and number. The predominant complex glycan structures are biantennary,which are variably galactosylated and sialylated. There is also a high proportion ofstructures that contain either or both a bisecting GlcNAc and/or core fucose residue,in different percentages between the immunoglobulins.ffffGlycan structures that could act as potential ligands for MBL have been

identified on all the immunoglobulins. In human serum only IgG-G0 and polymericand dimeric IgA have been shown to bind MBL and initiate the lectin pathway ofcomplement (Malhotra et al., 1995; Roos et al., 2001). In other immunoglobulinssmall quantities of GlcNAc-terminating glycan structures have been identified in theglycan pool. These structures may define small subpopulations of the immuno-globulins to which MBL could bind.

REFERENCES

Amzel, L. M., and Poljak, R. J. (1979). Three-dimensional structure of immunoglobulins. Annu Rev

Biochem 48, 961–997.

Anton, L. C., Alcolea, J. M., Sanchez-Corral, P., Marques, G., Sanchez, A., and Vivanco, F. (1989). C3

binds covalently to the C gamma 3 domain of IgG immune aggregates during complement activation

by the alternative pathway. Biochem J 257, 831–838.

Arlaud, G. J., and Colomb, M. G. (1987). Modelling of C1, the first component of human complement:

towards a consensus? Mol Immunol 24, 317.

Arnold, J. N., Radcliffe, C. M., Wormald, M. R., Royle, L., Harvey, D. J., Crispin, M., Dwek, R. A., Sim, R.ffff

B., and Rudd, P. M. (2004) In Press J Immunol.

Baenziger, J., and Kornfeld, S. (1974). Structure of the carbohydrate units of IgA1 immunoglobulin. I.

Composition, glycopeptide isolation, and structure of the asparagines linked oligosaccharide units. J

Biol Chem 249, 7260–7269.

Baenziger, J., Kornfeld, S., and Kochwa, S. (1974). Structure of the carbohydrate units of IgE immunoglob-

ulin. I. Over-all composition, glycopeptide isolation, and structure of the high mannose oligosacchar-

ide unit. J Biol Chem 249, 1889–1896.

Baenziger, J., Kornfeld, S., and Kochwa, S. (1974). Structure of the carbohydrate units of IgE immunoglob-

ulin. II. Sequence of the sialic acid-containing glycopeptides. J Biol Chem 249, 1897–1903.

Bos, N. A., Bun, J. C., Popma, S. H., Cebra, E. R., Deenen, G. J., van der Cammen, M. J., Kroese, F. G., and

Cebra, J. J. (1996). Monoclonal immunoglobulin A derived from peritoneal B cells is encoded by


both germ line and somatically mutated VH genes and is reactive with commensal bacteria. Infect

Immun 64, 616–623.

Butler, M., Quelhas, D., Critchley, A. J., Carchon, H., Hebestreit, H. F., Hibbert, R. G., Vilarinho, L., Teles,

E., Matthijs, G., Schollen, E., et al. (2003). Detailed glycan analysis of serum glycoproteins of patients

with congenital disorders of glycosylation indicates the specific defective glycan processing step and

provides an insight into pathogenesis. Glycobiology 13, 601–622.

Chapman, A., and Kornfeld, R. (1979). Structure of the high mannose oligosaccharides of a human IgM

myeloma protein. II. The minor oligosaccharides of high mannose glycopeptide. J Biol Chem 254,

824–828.

Cooper, N. R. (1985). The classical complement pathway: activation and regulation of the first complement

component. Adv Immunol 37, 151–216.

Crispin, M. D., Ritchie, G. E., Critchley, A. J., Morgan, B. P., Wilson, I. A., Dwek, R. A., Sim, R. B., and

Rudd, P. M. (2004). Monoglucosylated glycans in the secreted human complement component C3:

implications for protein biosynthesis and structure. FEBS Lett 566, 270–274.

Dahl, M. R., Thiel, S., Matsushita, M., Fujita, T., Willis, A. C., Christensen, T., Vorup-Jensen, T., and

Jensenius, J. C. (2001). MASP-3 and its association with distinct complexes of the mannan-binding

lectin complement activation pathway. Immunity 15, 127–135.

Delacroix, D. L., Dive, C., Rambaud, J. C., and Vaerman, J. P. (1982). IgA subclasses in various secretions

and in serum. Immunology 47, 383–385.

Dorrington, K. J., and Bennich, H. H. (1978). Structure-function relationships in human immunoglobulin

E. Immunol Rev 41, 3–25.

Duncan, A. R., and Winter, G. (1988). The binding site for C1q on IgG. Nature 332, 738–740.

Endo, T., Mestecky, J., Kulhavy, R., and Kobata, A. (1994). Carbohydrate heterogeneity of human myel-

oma proteins of the IgA1 and IgA2 subclasses. Mol Immunol 31, 1415–1422.

Field, M. C., Amatayakul-Chantler, S., Rademacher, T. W., Rudd, P. M., and Dwek, R. A. (1994). Structural

analysis of the N-glycans from human immunoglobulin A1: comparison of normal human serum

immunoglobulin A1 with that isolated from patients with rheumatoid arthritis. Biochem J 299 (Pt

1), 261–275.

Gala, F. A., and Morrison, S. L. (2002). The role of constant region carbohydrate in the assembly and

secretion of human IgD and IgA1. J Biol Chem 277, 29005–29011.

Garred, P., Madsen, H. O., Marquart, H., Hansen, T. M., Sorensen, S. F., Petersen, J., Volck, B., Svejgaard,

A., Graudal, N. A., Rudd, P. M., et al. (2000). Two edged role of mannose binding lectin in rheumat-

oid arthritis: a cross sectional study. J Rheumatol 27, 26–34.

Ghiran, I., Barbashov, S. F., Klickstein, L. B., Tas, S. W., Jensenius, J. C., and Nicholson-Weller, A. (2000).

Complement receptor 1/CD35 is a receptor for mannan-binding lectin. J Exp Med 192, 1797–1808.

Hajela, K., Kojima, M., Ambrus, G., Wong, K. H., Moffatt, B. E., Ferluga, J., Hajela, S., Gal, P., and Sim,ffff

R. B. (2002). The biological functions of MBL-associated serine proteases (MASPs). Immunobiology

205, 467–475.

Holmskov, U., Malhotra, R., Sim, R. B., and Jensenius, J. C. (1994). Collectins: collagenous C- type lectins

of the innate immune defense system. Immunol Today 15, 67–74.

Ikeda, K., Sannoh, T., Kawasaki, N., Kawasaki, T., and Yamashina, I. (1987). Serum lectin with known

structure activates complement through the classical pathway. J Biol Chem 262, 7451–7454.

Iobst, S. T., and Drickamer, K. (1994). Binding of sugar ligands to Ca(2+)-dependent animal lectins. II.Generation of high-affinity galactose binding by site-directed mutagenesis. J Biol Chem 269,

15512–15519.

Jefferis, R., Lund, J., Mizutani, H., Nakagawa, H., Kawazoe, Y., Arata, Y., and Takahashi, N. (1990). Affff

comparative study of the N-linked oligosaccharide structures of human IgG subclass proteins.

Biochem J 268, 529–537.

Johansen, F. E., Braathen, R., and Brandtzaeg, P. (2001). The J chain is essential for polymeric Ig receptor-

mediated epithelial transport of IgA. J Immunol 167, 5185–5192.

Koppel, R., and Solomon, B. (2001). IgM detection via selective recognition by mannose-binding protein.

J Biochem Biophys Methods 49, 641–647.

Lamm, M. E. (1997). Interaction of antigens and antibodies at mucosal surfaces. Annu Rev Microbiol

51, 311–340.

Lipscombe, R. J., Sumiya, M., Hill, A. V., Lau, Y. L., Levinsky, R. J., Summerfield, J. A., and Turner, M. W.

(1992). High frequencies in African and non-African populations of independent mutations in the

mannose binding protein gene. HumMol Genet 1, 709–715.


Litman, G. W., Anderson, M. K., and Rast, J. P. (1999). Evolution of antigen binding receptors. Annu Rev

Immunol 17, 109–147.

Madsen, H. O., Satz, M. L., Hogh, B., Svejgaard, A., and Garred, P. (1998). Different molecular eventsffff

result in low protein levels of mannan-binding lectin in populations from southeast Africa and South

America. J Immunol 161, 3169–3175.

Malhotra, R., Lu, J., Holmskov, U., and Sim, R. B. (1994). Collectins, collectin receptors and the lectin

pathway of complement activation. Clin Exp Immunol 97 Suppl 2, 4–9.

Malhotra, R., Wormald, M. R., Rudd, P. M., Fischer, P. B., Dwek, R. A., and Sim, R. B. (1995).

Glycosylation changes of IgG associated with rheumatoid arthritis can activate complement via the

mannose-binding protein. Nat Med 1, 237–243.

Matsushita, M., and Fujita, T. (1992). Activation of the classical complement pathway by mannose-binding

protein in association with a novel C1s-like serine protease. J Exp Med 176, 1497–1502.

Mattu, T. S., Pleass, R. J., Willis, A. C., Kilian, M., Wormald, M. R., Lellouch, A. C., Rudd, P.

M., Woof, J. M., and Dwek, R. A. (1998). The glycosylation and structure of human serum IgA1, Fab, and

Fc regions and the role of N-glycosylation on Fc alpha receptor interactions. J Biol Chem 273,

2260–2272.

Mellis, S. J., and Baenziger, J. U. (1983). Structures of the O-glycosidically linked oligosaccharides of

human IgD. J Biol Chem 258, 11557–11563.

Mellis, S. J., and Baenziger, J. U. (1983). Structures of the oligosaccharides present at the three asparagine-

linked glycosylation sites of human IgD. J Biol Chem 258, 11546–11556.

Mimura, Y., Church, S., Ghirlando, R., Ashton, P. R., Dong, S., Goodall, M., Lund, J., and Jefferis, R.ffff

(2000). The influence of glycosylation on the thermal stability and effector function expression offfff

human IgG1-Fc: properties of a series of truncated glycoforms. Mol Immunol 37, 697–706.

Nevens, J. R., Mallia, A. K., Wendt, M. W., and Smith, P. K. (1992). Affinity chromatographic purification

of immunoglobulin M antibodies utilizing immobilized mannan binding protein. J Chromatogr

597, 247–256.

Norderhaug, I. N., Johansen, F. E., Schjerven, H., and Brandtzaeg, P. (1999). Regulation of the formation

and external transport of secretory immunoglobulins. Crit Rev Immunol 19, 481– 508.

Ogden, C. A., deCathelineau, A., Hoffmann, P. R., Bratton, D., Ghebrehiwet, B., Fadok, V. A., and Henson,ffff

P. M. (2001). C1q and mannose binding lectin engagement of cell surface calreticulin and CD91

initiates macropinocytosis and uptake of apoptotic cells. J Exp Med 194, 781–795.

Parekh, R. B., Dwek, R. A., Sutton, B. J., Fernandes, D. L., Leung, A., Stanworth, D., Rademacher, T. W.,

Mizuochi, T., Taniguchi, T., Matsuta, K., and et al. (1985). Association of rheumatoid arthritis and

primary osteoarthritis with changes in the glycosylation pattern of total serum IgG. Nature 316,

452–457.

Rook, G. A., Steele, J., Brealey, R., Whyte, A., Isenberg, D., Sumar, N., Nelson, J. L., Bodman, K. B., Young,

A., Roitt, I. M., and et al. (1991). Changes in IgG glycoform levels are associated with remission of

arthritis during pregnancy. J Autoimmun 4, 779–794.

Roos, A., Bouwman, L. H., Munoz, J., Zuiverloon, T., Faber-Krol, M. C., Fallaux-van den Houten, F. C.,

Klar-Mohamad, N., Hack, C. E., Tilanus, M. G., and Daha, M. R. (2003).

Functional characterization of the lectin pathway of complement in human serum. Mol Immunol 39,

655–668.

Roos, A., Bouwman, L. H., van Gijlswijk-Janssen, D. J., Faber-Krol, M. C., Stahl, G. L., and Daha, M. R.

(2001). Human IgA activates the complement system via the mannan-binding lectin pathway. J

Immunol 167, 2861–2868.

Royle, L., Roos, A., Harvey, D. J., Wormald, M. R., van Gijlswijk-Janssen, D., Redwan el, R. M., Wilson, I.

A., Daha, M. R., Dwek, R. A., and Rudd, P. M. (2003). Secretory IgA N- and O- glycans provide a

link between the innate and adaptive immune systems. J Biol Chem 278, 20140–20153.

Rudd, P. M., Fortune, F., Patel, T., Parekh, R. B., Dwek, R. A., and Lehner, T. (1994). A human T-cell

receptor recognizes ‘O’-linked sugars from the hinge region of human IgA1 and IgD. Immunology

83, 99–106.

Sarmay, G., Lund, J., Rozsnyay, Z., Gergely, J., and Jefferis, R. (1992). Mapping and comparison of theffff

interaction sites on the Fc region of IgG responsible for triggering antibody dependent cellular

cytotoxicity (ADCC) through different types of human Fc gamma receptor. Mol Immunol 29,ffff

633–639.

Shields, R. L., Namenuk, A. K., Hong, K., Meng, Y. G., Rae, J., Briggs, J., Xie, D., Lai, J., Stadlen, A., Li, B.,


et al. (2001). High resolution mapping of the binding site on human IgG1 for Fc gamma RI, Fc

gamma RII, Fc gamma RIII, and FcRn and design of IgG1 variants with improved binding to the Fc

gamma R. J Biol Chem 276, 6591–6604.

Sim, R. B., Moestrup, S. K., Stuart, G. R., Lynch, N. J., Lu, J., Schwaeble, W. J., and Malhotra, R. (1998).

Interaction of C1q and the collectins with the potential receptors calreticulin (cC1qR/collectin recep-

tor) and megalin. Immunobiology 199, 208–224.

Sorensen, V., Sundvold, V., Michaelsen, T. E., and Sandlie, I. (1999). Polymerization of IgA and IgM: roles

of Cys309/Cys414 and the secretory tailpiece. J Immunol 162, 3448–3455.

Stockert, R. J. (1995). The asialoglycoprotein receptor: relationships between structure, function, and

expression. Physiol Rev 75, 591–609.

Sumiya, M., Super, M., Tabona, P., Levinsky, R. J., Arai, T., Turner, M. W., and Summerfield, J. A. (1991).

Molecular basis of opsonic defect in immunodeficient children. Lancet 337, 1569– 1570.

Summerfield, J. A., Sumiya, M., Levin, M., and Turner, M. W. (1997). Association of mutations in mannose

binding protein gene with childhood infection in consecutive hospital series. Bmj 314, 1229–1232.

Takahashi, N., Tetaert, D., Debuire, B., Lin, L. C., and Putnam, F. W. (1982). Complete amino acid

sequence of the delta heavy chain of human immunoglobulin D. Proc Natl Acad Sci U S A 79,

2850–2854.

Takayasu, T., Suzuki, S., Kametani, F., Takahashi, N., Shinoda, T., Okuyama, T., and Munekata, E. (1982).

Amino acid sequence of galactosamine-containing glycopeptides in the hinge region of a human

immunoglobulin D. Biochem Biophys Res Commun 105, 1066–1071.

Tarelli, E., Smith, A. C., Hendry, B. M., Challacombe, S. J., and Pouria, S. (2004). Human serum IgA1 is

substituted with up to six O-glycans as shown by matrix assisted laser desorption ionisation time-of-

flight mass spectrometry. Carbohydr Res 339, 2329–2335.

Tarentino, A. L., Plummer, T. H., Jr., and Maley, F. (1974). The release of intact oligosaccharides from

specific glycoproteins by endo-beta-N-acetylglucosaminidase H. J Biol Chem 249, 818– 824.

Thiel, S., Vorup-Jensen, T., Stover, C. M., Schwaeble, W., Laursen, S. B., Poulsen, K., Willis, A. C., Eggleton,

P., Hansen, S., Holmskov, U., et al. (1997). A second serine protease associated with mannan-binding

lectin that activates complement. Nature 386, 506–510.

Turner, M. W. (1996). Mannose-binding lectin: the pluripotent molecule of the innate immune system.

Immunol Today 17, 532–540.

Valdimarsson, H., Stefansson, M., Vikingsdottir, T., Arason, G. J., Koch, C., Thiel, S., and Jensenius, J. C.

(1998). Reconstitution of opsonizing activity by infusion of mannan-binding lectin (MBL) to MBL-

deficient humans. Scand J Immunol 48, 116–123.

Van Boxel, J. A., Paul, W. E., Terry, W. D., and Green, I. (1972). Communications. IgD-bearing human

lymphocytes. J Immunol 109, 648–651.

Vorup-Jensen, T., Petersen, S. V., Hansen, A. G., Poulsen, K., Schwaeble, W., Sim, R. B., Reid, K. B., Davis,

S. J., Thiel, S., and Jensenius, J. C. (2000). Distinct pathways of mannan- binding lectin (MBL)- and

C1-complex autoactivation revealed by reconstitution of MBL with recombinant MBL-associated

serine protease-2. J Immunol 165, 2093–2100.

Wan, T., Beavil, R. L., Fabiane, S. M., Beavil, A. J., Sohi, M. K., Keown, M., Young, R. J., Henry, A. J.,

Owens, R. J., Gould, H. J., and Sutton, B. J. (2002). The crystal structure of IgE Fc reveals an

asymmetrically bent conformation. Nat Immunol 3, 681–686.

Weis, W. I., Drickamer, K., and Hendrickson, W. A. (1992). Structure of a C-type mannose- binding protein

complexed with an oligosaccharide. Nature 360, 127–134.

Wiersma, E. J., Collins, C., Fazel, S., and Shulman, M. J. (1998). Structural and functional analysis of J

chain-deficient IgM. J Immunol 160, 5979–5989.

Wormald, M. R., Rudd, P. M., Harvey, D. J., Chang, S. C., Scragg, I. G., and Dwek, R. A. (1997). Variations

in oligosaccharide-protein interactions in immunoglobulin G determine the site-specific glycosylation

profiles and modulate the dynamic motion of the Fc oligosaccharides. Biochemistry 36, 1370–1380.

Wormald, M. R., Wooten, E. W., Bazzo, R., Edge, C. J., Feinstein, A., Rademacher, T. W., and Dwek, R. A.

(1991). The conformational effects of N-glycosylation on the tailpiece from serum IgM. Eur Jffff

Biochem 198, 131–139.

Youings, A., Chang, S. C., Dwek, R. A., and Scragg, I. G. (1996). Site-specific glycosylation of human

immunoglobulin G is altered in four rheumatoid arthritis patients. Biochem J 314 (Pt 2), 621–630.

Zhu, D., McCarthy, H., Ottensmeier, C. H., Johnson, P., Hamblin, T. J., and Stevenson, F. K. (2002).

Acquisition of potential N-glycosylation sites in the immunoglobulin variable region by somatic

mutation is a distinctive feature of follicular lymphoma. Blood 99, 2562–2568.

10

GELATINASE B PARTICIPATES IN COLLAGEN IIDEGRADATION AND RELEASES GLYCOSYLATEDREMNANT EPITOPES IN RHEUMATOIDARTHRITIS

P. E. Van den Steen, B. Grillet, and G. Opdenakker

Laboratory of ImmunobiologyRega Institute for Medical ResearchUniversity of LeuvenMinderbroedersstraat 10, 3000 Leuven, Belgium

1. INTRODUCTION

Rheumatoid arthritis is an autoimmune disease characterized by chronicinflammation of the joints. It is associated with the activation of autoreactive T-cellsand with production of autoantibodies. The main auto-antigen is collagen type II,which is a major constituent of the cartilage in the joint. The inflammation causescartilage degradation, hyperplasia of synovial membranes, accumulation of excessivesynovial fluid, and, in the worst cases, bone erosion. The exact aetiology is notknown, but it is clear that inflammatory reactions and auto-antibodies, which activatethe complement cascade, are main causes of the cartilage degradation. Inhibition ofinflammation by interference with some of the main pro-inflammatory cytokines,interleukin-1 (IL-1) and tumor necrosis factor-a (TNF-a) has proven to be beneficialand constitutes the basis of currently approved treatments.

Matrix metalloproteinases are a family of neutral Zn2+-dependent endoprote-ases, which share a number of homologous domains (Nagase and Woessner, 1999).These domains are the Zn2+-containing active site, kept inactive by a propeptide inthe pro-enzyme form, and a hemopexin domain (Fig. 1). The latter domain is presentin most MMPs except for matrilysins (MMP-7 and -26) and cysteine-array MMP(CA-MMP or MMP-23). Several MMPs contain additional domains, such as acarboxyterminal membrane anchor in membrane-type MMPs (MT-MMPs), afibronectin-like domain in gelatinases A and B (MMP-2 and -9), and a uniquemucin-type domain in gelatinase B. The latter domain is often named collagen type

45



46 P. E. Van den Steen et al.

Figure 1. Domain structure of MMPs

MMPs are a family of neutral endoproteinases with a number of conserved domains, including the

prodomain and the active enzyme and Zn2+-binding domains, which together form the active site of theenzymes. Most MMPs have also a carboxyterminal hemopexin domain, except for the matrilysins (MMP-7

and –26) and cysteine-array MMP (MMP-23). MT-MMPs additionally contain a GPI membrane anchor

(for MT4-MMP and MT6-MMP) or a transmembrane domain with a short cytoplasmic domain (the

other MT-MMPs). Gelatinases contain a fibronectin domain, and gelatinase B contains also a unique

mucin-like domain, named the collagen V domain, which is an ideal attachment site for clustered O-linked

sugars. The theoretical attachment sites for N-linked sugars are indicated with a Y symbol, one of which

is conserved among most MMPs (indicated by a filled diamond). On top, the conserved histidines are

shown which interact with the catalytic Zn2+, and the conserved cysteine in the propeptide, also interactingwith the catalytic Zn2+ to keep the enzyme latent.

Gelatinase B Participation in Collagen II Degradation 47

V domain because it contains a large number of proline residues, conferring somehomology to collagen type V. However, it is not a true collagen-like domain sinceit does not contain a glycine at every three amino acids, which is a main characteristicof collagen. Instead, this domain in gelatinase B contains repeats of the sequenceS/T-X-X-P, and is therefore an ideal site for the clustered attachment of O-linkedglycans, which are abundantly present on gelatinase B (Van den Steen et al., 1998;Mattu et al., 2000; Van den Steen et al., 2001).MMPs are well-known for their ability to cleave most extracellular matrix

components. In particular, the three classical collagenases, interstitial collagenase(MMP-1), neutrophil collagenase (MMP-8) and collagenase-3 (MMP-13), are ableto cleave triple-helical collagen, which is otherwise highly resistant to most proteases(Jeffrey, 1998). Furthermore, stromelysins and matrilysins degrade proteoglycans,ffffand gelatinases degrade a variety of other ECM components such as gelatin, elastin(Senior et al., 1991), link protein (Nguyen et al., 1993) and collagen type V (Hibbset al., 1987). Besides cleaving ECM proteins, MMPs also cleave a variety of regu-latory molecules, such as serine protease inhibitors, cytokines and chemokines(Opdenakker et al., 2001).

2. ROLE OF GELATINASE B IN RHEUMATOID ARTHRITIS

2.1. Expression of Gelatinase B in Rheumatoid Arthritis

The expression of gelatinase B in synovial fluids of patients with rheumatoidarthritis has been documented more than 10 years ago (Opdenakker et al., 1991)and was confirmed in several other studies (Sopata et al., 1995; Ahrens et al., 1996;Gruber et al., 1996). The amount of gelatinase B, as measured by gelatin zymography,correlates with inflammatory markers such as IL-6 and IL-8 (Opdenakker et al.,1991; Van den Steen et al., 2002b). Furthermore, gelatinase B is present underdifferent forms in synovial fluids: monomers of around 92 kDa, homodimers offfff200 kDa and a complex of gelatinase B with neutrophil gelatinase B-associatedlipocalin (NGAL) (Fig. 2). The latter complex is only synthesized by neutrophils(Kjeldsen et al., 1993; Triebel et al., 1992), indicating that neutrophils are the mainproducers of gelatinase B in the synovial fluid. In fact, the chemokine IL-8, which isalso upregulated in the synovial fluids (Peichl et al., 1991; Rampart et al., 1992; Seitzet al., 1992), attracts neutrophils from the blood vessels to the synovial fluid(Akahoshi et al., 1994) and stimulates these cells to degranulate. Since gelatinase Bis present in the granules in the proform, this results in the release of progelatinaseB (Masure et al., 1991). The propeptide may be removed in different ways, includingffffproteolysis by several proteases and oxidation of a conserved cysteine in the propep-tide, which normally is bound to the Zn2+ ion (Van den Steen et al., 2002a). Whichactivation pathway is active in the synovial fluid of RA patients is unknown, but itis likely that stromelysin-1 contributes to the activation, since it is also increased inarthritic synovial fluids (Ribbens et al., 2000) and it is an efficient activator ofgelatinase B (Ogata et al., 1992). However, it is also possible that reactive oxygenspecies, produced by the activated neutrophils, are contributing to the activation(Peppin and Weiss, 1986). The active form of the protease is seen in a limited number


Figure 2. Gelatin zymography of synovial fluids

Synovial fluids from arthritis patients were analysed by gelatin zymography. The positions of monomeric

pro- and active gelatinase B, homodimeric gelatinase B, gelatinase B complexed with NGAL and gelatinase

A are indicated. In some synovial fluids ( lanes 1 and 6), the activated forms of gelatinase B are present,

whereas other synovial fluids only contain gelatinase A (lane 3).

of samples. However, the presence of active gelatinase B on zymography does notprove its activity in the original synovial fluid, since it may still be complexed withthe tissue inhibitor of metalloproteinases (TIMP)-1 (Murphy andWillenbrock, 1995).As this complex dissociates during the electrophoresis, TIMP-1 cannot influence theactivity on gelatin-zymography. Therefore, it was analysed whether net activity waspresent in the synovial fluids, using an activity assay with fluorescently labelledgelatin coated onto microspheres. A high activity was measured in some patients,and in serial samples from the same patients the activity varied greatly with time(Van den SteenVV et al., 2002b).

2.2. Role of Gelatinase B in Cartilage Breakdown

The presence of gelatinolytic activity in the synovial fluids and synovial tissuesof RA patients suggests a role for this enzyme in the disease. Gelatinase B knock-out mice are resistant to anti-collagen II antibody-induced arthritis (Itoh et al.,2002). This is in sharp contrast to gelatinase A knock-out mice, which are moresusceptible for arthritis development. The latter finding may be explained by anti-inflammatory activities of gelatinase A, in particular the cleavage of monocytechemotactic protein-3. This cleavage of MCP-3 aborts its potential to induce signaltransduction through its receptor, but does not impair receptor binding and thereforeconverts the chemokine in an antagonist (McQuibban et al., 2000). In contrast,gelatinase B processes the neutrophil chemoattractant interleukin-8 into a more than10-fold more potent variant, inducing a positive feedback loop which fuels inflamma-tion (Van den Steen et al., 2000). These studies show the importance of the cleavageof regulatory molecules by MMPs in acute and chronic inflammation (Opdenakkeret al., 2001).


Figure 3. Combined action of neutrophil collagenase and gelatinase B on native collagen II

Collagenases, and in particular neutrophil collagenase (MMP-8), efficiently cleave native collagen II at a

single site, resulting in the formation of 34and 14fragments. When both collagenase and gelatinase B are

added to native collagen II, the collagen is completely degraded into small fragments, as is visualised on

this SDS-PAGE analysis. The positions of uncleaved native collagen II and the 34fragment of collagen II

are indicated.

However, the most efficient substrate of gelatinase B is gelatin, which is heat-denatured collagen. Triple helical collagen II cannot be cleaved by gelatinase B,because the peptide bonds are at the inner side of the triple helix and not accessiblefor most proteases, including gelatinase B. However, collagenases can unwind thetriple helix locally and subsequently cleave the peptide chain at a single site, generat-ing the so-called 3

4and 14fragments. Neutrophil collagenase, which is usually secreted

together with gelatinase B from the granules of neutrophils, is particularly efficientin the cleavage of collagen type II (Jeffrey, 1998). Although this cleavage does notffffresult in complete denaturation of the triple helix (Marini et al., 2000), the 3

4and 14

fragments are susceptible to cleavage by gelatinase B (Fig. 3). The two fragmentsare degraded by gelatinase B into small peptides, and these were analysed afterseparation of the fragments by RP-HPLC (Van den Steen et al., 2002b; Van denSteen et al., 2004). To identify each peptide in each fraction, all fractions weresubjected to Edman degradation and mass spectrometry. Edman degradation yieldsthe aminoterminal sequence, and mass spectrometry defines the exact size and thusthe carboxyterminus of the peptides (figure 4). In this way, 25 and 30 differentffffcleavage sites were identified in bovine and human denatured collagen II, respectively.The finding and definition of the reaction products confirm that gelatinase B particip-ates in the complete degradation of collagen II into small peptides (Table 1).

Two main immunodominant T-cell epitopes occur in collagen II (Brand et al.,1994; Myers et al., 1995; Rosloniec et al., 1996; Andersson et al., 1998). The positionof the cleavage sites, relative to the known immunodominant autoreactive T-cellepitopes, suggests that the degradation of collagen II by the combined action ofcollagenase and gelatinase B not only leads to the degradation of the cartilage, butalso releases intact immunodominant epitopes. These remnant epitopes may evenlay at the basis of autoimmunity, since they may be loaded onto MHC-II of antigen


Table 1. Comparison of the cleavage sites by gelatinase B in natural human type II collagen

P6 P5 P4 P3 P2 P1 P1∞ P2∞ P3∞ P4∞ P5∞ P6∞

A POH G P Q G F42 Q G N POH G

P P* G P Q G A93 R G F POH G

P M G P R G L147LL P(OH) G E R G

R T G P A G A159 A G A R G

P A G A A G A162 R G N D G

A R G P E G A207 Q G P R G

S POH G P A G A225 S G N P* G

P L G P K G Q273 T G E POH G

A POH G P A G E306 E G K R G

P I G P POH G E327 R G A POH G

E R G P S G L357LL A G P K(OHex) G

L A G P K(OHex) G A363 N G D P(OH) G

L POH G A R G L381LL T G R POH G

P P(OH) G P Q G A414 R G Q POH G

L P(OH) G A P(OH) G L447LL R G L POH G

A POH G P S G F483 Q G L POH G

L V G P R G E519 R G F POH G

S POH G A Q G L534LL Q G P R G

L Q G P R G L540LL P* G T P G

A Q G P P* G L567LL Q G M POH G

P P* G P A G A618 N G E K(OH) G

E T G P POH G T651 S G F A G

P POH G T S G F654 A G P POH G

P Q G P T G V699VV T G P K(OHex) G

A Q G P POH G A714 T G F POH G

K D G P K G A750 R G D S G

P P* G P Q G L795LL A G Q R G

L A G Q R G I801 V G L POH G

P V G P P* G L846LL T G P A G

P A G A R G I927 Q G P Q G

L K G H R G F954 T G L Q G

Residues showing considerable consensus are shown in bold. At P1∞, hydrophobic residues areindicated in italic and the position in the sequence is indicated in subscript. POH, hydroxypro-line; KOHex, glycosylated hydroxylysine, P*, Pro with probable but uncertain hydroxylation.

presenting cells. Loading of the remnant epitopes can occur either after internalisationand further processing, or at the outside of the antigen presenting cell, since emptyMHC-II is expressed on the cell surface, together with HLA-DM, which is knownto catalyze loading of peptides in MHC-II (Santambrogio et al., 1999; Arndt et al.,2000). Inflammatory stimuli in the joint will therefore induce degradation of collagenII by collagenases and gelatinase B, and additionally activate the antigen presentingcells to upregulate costimulatory molecules, increasing their capacity to activateT-cells (Inaba et al., 2000). Furthermore, latent autoreactive T-cells are present inmany healthy individuals (Sun et al., 1991). This may mean that, with a susceptiblegenetic background and with the presence of autoreactive T-cells, an inflammationin the joint may induce rheumatoid arthritis. This theory was originally formulatedin the context of multiple sclerosis and was named ‘‘remnant epitopes generateautoimmunity’’ or REGA model (Opdenakker and Van Damme, 1994). The REGA


Figure 4. Use of Edman degradation and mass spectrometry to analyse the gelatinase B cleavage sites

and posttranslational modifications in collagen II

Denatured bovine and human collagen type II were digested with natural human gelatinase B. The

resulting fragments were separated by RP-HPLC, and the resulting fractions were analysed by Edman

degradation and mass spectrometry. The combination of these technologies enabled to determine 30

different gelatinase B cleavage sites and to localize a large number of posttranslational modifications inffff

collagen type II.

model is further supported by the study of Pu et al., who showed the existence ofT-cells which react only with epitopes that are destroyed after intracellular processingof the antigen, suggesting that extracellular processing and loading on MHC-II mustoccur (Pu et al., 2002).

3. POSTTRANSLATIONAL MODIFICATIONS OF HUMANCOLLAGEN II

The use of both mass spectrometry and Edman degradation for the determina-tion of the gelatinase B cleavage sites in type II collagen also allowed the determina-tion of a large number of posttranslational modifications in the sequence (Fig. 4).These include proline-hydroxylation, lysine-hydroxylation and glycosylation of


hydroxylysine. With the use of the outlined procedure, 9 different glycosylation sitesffffwere determined, of which one is in the major immunodominant epitope and one inits close vicinity (Van den Steen et al., 2004; Van den Steen et al., 2002b). This hasmajor implications for the immunogenicity of this immunodominant epitope, as itis conceivable that sugars on a peptide will modify its recognition by the immunesystem. This was shown by several groups. Glycosylation of an antigenic peptidecan influence its binding in the groove of MHC-I or MHC-II, as such a sugar maymake extensive contacts within the groove or also provide steric hindrance to thebinding (Haurum et al., 1995). Another possibility is that the sugar points out ofthe groove, consequently modifying the interaction with the T-cell receptor and evenleading to activation of different T-cell clones (Haurumffff et al., 1994; Corthay et al.,1998). In this context, it is interesting to note that glycosylated collagen II is morepotent to elicit arthritis than a chemically deglycosylated variant (Michaelsson et al.,1994) or a recombinant underglycosylated form of collagen II (Myers et al., 2004).Furthermore, autoreactive T-cells from patients with rheumatoid arthritis reactpreferentially with a synthetic glycosylated form of the immunodominant epitopethan with the unglycosylated peptide (Backlund et al., 2002). Therefore, the formalproof that a glycan is present on the immunodominant epitope of natural humancollagen II (Van den Steen et al., 2004) is significant for the further understandingof the autoimmune process in rheumatoid arthritis patients. Furthermore, the knownimmunodominant epitopes in collagen II have been detected by the technique ofepitope scanning with unmodified synthetic peptides. It is therefore conceivable thatother, modified immunodominant epitopes are present in collagen II. This notion isfurther exemplified by the finding that, after RP-HPLC separation of the degradationproducts of collagen II after gelatinase B cleavage, some fractions without knownimmunodominant epitopes induce T-cell reactivity (Van den Steen et al., 2004).

Figure 5. Generation of glycosylated remnant epitopes from collagen II by neutrophil collagenase and

gelatinase B

Triple helical collagen is highly resistant to proteolysis. Only collagenases, e.g. neutrophil

collagenase/MMP-8, are able to locally unwind the triple helix and to perform a single cleavage. Thereafter,

the collagen becomes susceptible to degradation by gelatinase B. The resulting peptides, some of which

contain oligosaccharides, can subsequently be loaded onto MHC-II of antigen presenting cells and be

presented to autoreactive T-cells.


REFERENCES

Ahrens, D., Koch, A.E., Pope, R.M., Stein-Picarella, M., and Niedbala, M.J. (1996). Expression of matrix

metalloproteinase 9 (96-kd gelatinase B) in human rheumatoid arthritis. Arthritis Rheum. 39,

1576–1587.

Akahoshi, T., Endo, H., Kondo, H., Kashiwazaki, S., Kasahara, T., Mukaida, N., Harada, A., and

Matsushima, K. (1994). Essential involvement of interleukin-8 in neutrophil recruitment in rabbits

with acute experimental arthritis induced by lipopolysaccharide and interleukin-1. Lymphokine

Cytokine. Res. 13, 113–116.

Andersson, E.C., Hansen, B.E., Jacobsen, H., Madsen, L.S., Andersen, C.B., Engberg, J., Rothbard, J.B.,

McDevitt, G.S., Malmstrom, V., Holmdahl, R., Svejgaard, A., and Fugger, L. (1998). Definition of

MHC and T cell receptor contacts in the HLA-DR4 restricted immunodominant epitope in type II

collagen and characterization of collagen-induced arthritis in HLA-DR4 and human CD4 transgenic

mice. Proc. Natl. Acad. Sci. U.S.A. 95, 7574–7579.

Arndt, S.O., Vogt, A.B., Markovic-Plese, S., Martin, R., Moldenhauer, G., Wolpl, A., Sun, Y., Schadendorf,

D., Hammerling, G.J., and Kropshofer, H. (2000). Functional HLA-DM on the surface of B cells and

immature dendritic cells. EMBO J. 19, 1241–1251.

Backlund, J., Carlsen, S., Hoger, T., Holm, B., Fugger, L., Kihlberg, J., Burkhardt, H., and Holmdahl, R.

(2002). Predominant selection of T cells specific for the glycosylated collagen type II epitope

(263–270) in humanized transgenic mice and in rheumatoid arthritis. Proc. Natl. Acad. Sci. U.S.A.

99, 9960–9965.

Brand, D.D., Myers, L.K., Terato, K., Whittington, K.B., Stuart, J.M., Kang, A.H., and Rosloniec, E.F.

(1994). Characterization of the T cell determinants in the induction of autoimmune arthritis by

bovine a1(II)-CB11 in H-2q mice. J. Immunol. 152, 3088–3097.

Corthay, A., Backlund, J., Broddefalk, J., Michaelsson, E., Goldschmidt, T.J., Kihlberg, J., and Holmdahl,

R. (1998). Epitope glycosylation plays a critical role for T cell recognition of type II collagen in

collagen-induced arthritis. Eur. J. Immunol. 28, 2580–2590.

Gruber, B.L., Sorbi, D., French, D.L., Marchese, M.J., Nuovo, G.J., Kew, R.R., and Arbeit, L.A. (1996).

Markedly elevated serum MMP-9 (gelatinase B) levels in rheumatoid arthritis: a potentially useful

laboratory marker. Clin. Immunol. Immunopathol. 78, 161–171.

becomes sensitive for cleavage by gelatinase B and is fragmented into small peptidesthat still contain the immunodominant epitopes. These auto-antigenic peptides,named remnant peptides, may be loaded onto MHC-II and may be presented byantigen presenting cells to autoreactive T-cells (Fig. 5). This leads to the activationof these T-cells, resulting in the formation of autoantibodies, which can activate thecomplement cascade, further enhancing the inflammation. Furthermore, some of theremnant epitopes are modified by hydroxylation and glycosylation, which increasestheir immunogenicity.

ACKNOWLEDGEMENTS

P.E. Van den Steen is a postdoctoral fellow of the Belgian National Fund forScientific Research (F.W.O.-Vlaanderen). Supported by the GeconcerteerdeOnderzoeksActies (GOA 2002–2006) and by the F.W.O.-Vlaanderen.

4. CONCLUSIONS

Inflammation of the joint, as seen in rheumatoid arthritis, is associated with theexpression of MMPs, in particular collagenases and gelatinase B. Collagenases suchas MMP-8 cleave the triple helical collagen type II, which is one of the maincomponents of joint cartilage, at a single position. As a consequence, collagen II


Haurum, J.S., Tan, L., Arsequell, G., Frodsham, P., Lellouch, A.C., Moss, P.A., Dwek, R.A., McMichael,

A.J., and Elliott, T. (1995). Peptide anchor residue glycosylation: effect on class I major histocompat-ffff

ibility complex binding and cytotoxic T lymphocyte recognition. Eur. J. Immunol. 25, 3270–3276.

Hibbs, M.S., Hoidal, J.R., and Kang, A.H. (1987). Expression of a metalloproteinase that degrades native

type V collagen and denatured collagens by cultured human alveolar macrophages. J. Clin. Invest.

80, 1644–1650.

Inaba, K., Turley, S., Iyoda, T., Yamaide, F., Shimoyama, S., Reis, Germain, R.N., Mellman, I., and

Steinman, R.M. (2000). The formation of immunogenic major histocompatibility complex class

II-peptide ligands in lysosomal compartments of dendritic cells is regulated by inflammatory stimuli.

J. Exp. Med. 191, 927–936.

Itoh, T., Matsuda, H., Tanioka, M., Kuwabara, K., Itohara, S., and Suzuki, R. (2002). The role of matrix

metalloproteinase-2 and matrix metalloproteinase-9 in antibody-induced arthritis. J. Immunol. 169,

2643–2647.

Jeffrey, J.J. (1998). Interstitial collagenases. In Matrix Metalloproteinases, W.C.Parks and R.P.Mecham,ffff

eds. Academic Press, pp. 15–42.

Kjeldsen, L., Johnsen, A.H., Sengelov, H., and Borregaard, N. (1993). Isolation and primary structure of

NGAL, a novel protein associated with human neutrophil gelatinase. J. Biol. Chem. 268,

10425–10432.

Marini, S., Fasciglione, G.F., De Sanctis, G., D’Alessio, S., Politi, V., and Coletta, M. (2000). Cleavage of

bovine collagen I by neutrophil collagenase MMP-8. Effect of pH on the catalytic properties asffff

compared to synthetic substrates. J. Biol. Chem. 275, 18657–18663.

Masure, S., Proost, P., Van Damme, J., and Opdenakker, G. (1991). Purification and identification of

91-kDa neutrophil gelatinase. Release by the activating peptide interleukin-8. Eur. J. Biochem.

198, 391–398.

Mattu, T.S., Royle, L., Langridge, J., Wormald, M.R., Van den Steen, P.E., Van Damme, J., Opdenakker,

G., Harvey, D.J., Dwek, R.A., and Rudd, P.M. (2000). O-Glycan analysis of natural human neutrophil

gelatinase B using a combination of normal phase- HPLC and online tandem mass spectrometry:

implications for the domain organization of the enzyme. Biochemistry 39, 15695–15704.

McQuibban, G.A., Gong, J.H., Tam, E.M., McCulloch, C.A., Clark-Lewis, I., and Overall, C.M. (2000).

Inflammation dampened by gelatinase A cleavage of monocyte chemoattractant protein-3. Science

289, 1202–1206.

Michaelsson, E., Malmstrom, V., Reis, S., Engstrom, A., Burkhardt, H., and Holmdahl, R. (1994). T cell

recognition of carbohydrates on type II collagen. J. Exp. Med. 180, 745–749.

Murphy, G. and Willenbrock, F. (1995). Tissue inhibitors of matrix metalloendopeptidases. Methods

Enzymol. 248, 496–510.

Myers, L.K., Miyahara, H., Terato, K., Seyer, J.M., Stuart, J.M., and Kang, A.H. (1995). Collagen-induced

arthritis in B10.RIII mice (H-2r): identification of an arthritogenic T-cell determinant. Immunology

84, 509–513.

Myers, L.K., Myllyharju, J., Nokelainen, M., Brand, D.D., Cremer, M.A., Stuart, J.M., Bodo, M., Kivirikko,

K.I., and Kang, A.H. (2004). Relevance of posttranslational modifications for the arthritogenicity of

type II collagen. J. Immunol. 172, 2970–2975.

Nagase, H. and Woessner, J.F. (1999). Matrix metalloproteinases. J. Biol. Chem. 274, 21491–21494.

Nguyen, Q., Murphy, G., Hughes, C.E., Mort, J.S., and Roughley, P.J. (1993). Matrix metalloproteinases

cleave at two distinct sites on human cartilage link protein. Biochem. J. 295, 595–598.

Ogata, Y., Enghild, J.J., and Nagase, H. (1992). Matrix metalloproteinase 3 (stromelysin) activates the

precursor for the human matrix metalloproteinase 9. J. Biol. Chem. 267, 3581–3584.

Opdenakker, G., Masure, S., Grillet, B., and Van Damme J. (1991). Cytokine-mediated regulation of

human leukocyte gelatinases and role in arthritis. Lymphokine Cytokine. Res. 10, 317–324.

Opdenakker, G. and Van Damme, J. (1994). Cytokine-regulated proteases in autoimmune diseases.

Immunol. Today 15, 103–107.Opdenakker, G., Van den Steen, P.E., and Van Damme, J. (2001). Gelatinase B: a tuner and amplifier of

immune functions. Trends. Immunol. 22, 571–579.Peichl, P., Ceska, M., Effenberger, F., Haberhauer, G., Broell, H., and Lindley, I.J. (1991). Presence offfff

NAP-1/IL-8 in synovial fluids indicates a possible pathogenic role in rheumatoid arthritis. Scand. J.

Immunol. 34, 333–339.

Haurum, J.S., Arsequell, G., Lellouch, A.C., Wong, S.Y., Dwek, R.A., McMichael, A.J., and Elliott, T.

(1994). Recognition of carbohydrate by major histocompatibility complex class I-restricted, glyco-

peptide-specific cytotoxic T lymphocytes. J. Exp. Med. 180, 739–744.


Pu, Z., Carrero, J.A., and Unanue, E.R. (2002). Distinct recognition by two subsets of T cells of an MHC

class II-peptide complex. Proc. Natl. Acad. Sci. U.S.A. 99, 8844–8849.

Rampart, M., Herman, A.G., Grillet, B., Opdenakker, G., and Van Damme, J. (1992). Development and

application of a radioimmunoassay for interleukin-8: detection of interleukin-8 in synovial fluids

from patients with inflammatory joint disease. Lab. Invest. 66, 512–518.

Ribbens, C., Andre, B., Jaspar, J.M., Kaye, O., Kaiser, M.J., De Groote, D., and Malaise, M.G. (2000).

Matrix metalloproteinase-3 serum levels are correlated with disease activity and predict clinical

response in rheumatoid arthritis. J. Rheumatol. 27, 888–893.

Rosloniec, E.F., Whittington, K.B., Brand, D.D., Myers, L.K., and Stuart, J.M. (1996). Identification of

MHC class II and TCR binding residues in the type II collagen immunodominant determinant

mediating collagen-induced arthritis. Cell Immunol. 172, 21–28.

Santambrogio, L., Sato, A.K., Carven, G.J., Belyanskaya, S.L., Strominger, J.L., and Stern, L.J. (1999).

Extracellular antigen processing and presentation by immature dendritic cells. Proc. Natl. Acad. Sci.

U.S.A. 96, 15056–15061.

Seitz, M., Dewald, B., Ceska, M., Gerber, N., and Baggiolini, M. (1992). Interleukin-8 in inflammatory

rheumatic diseases: synovial fluid levels, relation to rheumatoid factors, production by mononuclear

cells, and effects of gold sodium thiomalate and methotrexate. Rheumatol. Int.ffff 12, 159–164.

Senior, R.M., Griffin, G.L., Fliszar, C.J., Shapiro, S.D., Goldberg, G.I., and Welgus, H.G. (1991). Human

92– and 72-kilodalton type IV collagenases are elastases. J. Biol. Chem. 266, 7870–7875.

Sopata, I., Wize, J., Filipowicz-Sosnowska, A., Stanislawska-Biernat, E., Brzezinska, B., and Maslinski, S.

(1995). Neutrophil gelatinase levels in plasma and synovial fluid of patients with rheumatic diseases.

Rheumatol. Int. 15, 9–14.

Sun, J.B., Olsson, T., Wang, W.Z., Xiao, B.G., Kostulas, V., Fredrikson, S., Ekre, H.P., and Link, H. (1991).

Autoreactive T and B cells responding to myelin proteolipid protein in multiple sclerosis and con-

trols. Eur. J. Immunol. 21, 1461–1468.

Triebel, S., Blaser, J., Reinke, H., and Tschesche, H. (1992). A 25 kDa alpha 2-microglobulin-related protein

is a component of the 125 kDa form of human gelatinase. FEBS Lett. 314, 386–388.

Van den Steen, P., Rudd, P.M., Dwek, R.A., and Opdenakker, G. (1998). Concepts and principles of

O-linked glycosylation. Crit. Rev. Biochem. Mol. Biol. 33, 151–208.

Van den Steen, P.E., Dubois, B., Nelissen, I., Rudd, P.M., Dwek, R.A., and Opdenakker, G. (2002a).

Biochemistry and molecular biology of gelatinase B or matrix metalloproteinase-9 (MMP-9). Crit

Rev. Biochem. Mol. Biol. 37, 375–536.

Van den Steen, P.E., Opdenakker, G., Wormald, M.R., Dwek, R.A., and Rudd, P.M. (2001). Matrix

remodelling enzymes, the protease cascade and glycosylation. Biochim. Biophys. Acta 1528, 61–73.

Van den Steen, P.E., Proost, P., Brand, D.D., Kang, A.H., Van Damme, J., and Opdenakker, G. (2004).

Generation of Glycosylated Remnant Epitopes from Human Collagen Type II by Gelatinase B.

Biochemistry 43, 10809–10816.

Van den Steen, P.E., Proost, P., Grillet, B., Brand, D.D., Kang, A.H., Van Damme, J., and Opdenakker, G.

(2002b). Cleavage of denatured natural collagen type II by neutrophil gelatinase B reveals enzyme

specificity, post-translational modifications in the substrate, and the formation of remnant epitopes

in rheumatoid arthritis. FASEB J. 16, 379–389.

Van den Steen, P.E., Proost, P., Wuyts, A., Van Damme, J., and Opdenakker, G. (2000). Neutrophil

gelatinase B potentiates interleukin-8 tenfold by aminoterminal processing, whereas it degrades

CTAP-III, PF-4, and GRO-alpha and leaves RANTES and MCP-2 intact. Blood 96, 2673–2681.

Peppin, G.J. and Weiss, S.J. (1986). Activation of the endogenous metalloproteinase, gelatinase, by trig-

gered human neutrophils. Proc. Natl. Acad. Sci. U.S.A. 83, 4322–4326.

11

HYALURONAN IN IMMUNE PROCESSES

Alan J. Wright and Anthony J. Day

MRC Immunochemistry UnitDepartment of BiochemistryUniversity of OxfordSouth Parks Road, Oxford OX1 3QU, UK

1. INTRODUCTION

The linear polysaccharide hyaluronan (HA) is a glycosaminoglycan consistingentirely of a repeating disaccharide of glucuronic acid and N-acetyl glucosamine (seeFig. 1), which unlike other glycosaminoglycans (e.g., heparin and chondroitinsulphate) is neither sulphated, nor attached to a protein core (Day and Sheehan,2001; Fraser et al., 1997; Tammi et al., 2002). While HA molecules are usually ofvery high molecular weight (i.e., 105 to 107 Da), smaller fragments and oligosacchar-ides of HA have also been detected under certain physiological or pathologicalconditions. HA is highly polar existing as a polyanion and, in aqueous solutions(such as synovial fluid), its large hydrodynamic volume and viscoelastic propertiesgive rise to its important space filling, filtering and lubricating functions(Hardingham, 2004; Laurent et al., 1996). It is also a vital structural component ofextracellular matrix, where it has diverse roles in development, wound healing,

Figure 1. The structure of a trisaccharide of hyaluronan showing the alternating b(1–4) and b(1–3)

linkages.

57



58 A. J. Wright and A. J. Day

ovulation and the immune response (Blundell et al., 2004b; Tammi et al., 2002;). Forexample, at sites of inflammation HA is involved in the preliminary stages ofleukocyte adhesion and transendothelial migration.

Given the simplicity of HA’s chemical structure it might seem, at first glance,surprising that it is implicated in so many different processes. However, there isffffincreasing evidence that this diversity results from the interaction of HA with specificHA-binding proteins (known as hyaladherins) to form HA-protein complexes withdistinct architectures and functional activities (see Blundell et al., 2004b). In thisregard, it has been proposed that individual hyaladherins are able to capture andpropagate particular conformations of the polysaccharide, leading to the formationof different higher order structures (Day and Sheehan, 2001); in the absence offfffproteins, solution HA is thought to correspond to a stiffened random coil, existingffffas a highly dynamic ensemble of chaotically interchanging, semi-ordered states(reviewed in Blundell et al., 2004a). The repeating nature of HA (and its chemicalfidelity), coupled with its conformational repertoire, makes it a perfect scaffold toffffform periodic protein arrays, especially where cooperative interactions lead to clus-tering of hyaladherins along the HA chain (Day and Sheehan, 2001); complexes ofthis type are well known components of cartilage matrix. Furthermore, the cross-linking of multiple HA strands, via protein-protein interactions, can result in hugecable-like structures, which have been shown to form during inflammation and havedistinctive leukocyte-binding properties (de la Motte et al., 2003, 2004; Majorset al., 2003).

The majority of hyaladherins are members of the Link module superfamily(Blundell et al., 2004b; Day and Prestwich, 2002) containing a common proteindomain of ~100 amino acids that mediates the interaction with HA. To date three-dimensional structures have only been determined for the HA-binding domains oftwo members of this superfamily; i.e., TSG-6 (tumor necrosis factor-stimulatedgene-6) and CD44. The interaction domain of TSG-6, an inflammation-associatedprotein (see Milner and Day, 2003), consists of a single independently folded Linkmodule that alone can support high affinity HA binding (Blundell et al., 2003; Kohdaet al., 1996). In the case of CD44, a cell surface receptor for HA, additional N- andC-terminal sequences flanking the Link module are necessary to form a stable fold,giving rise to an HA-binding domain of ~150 amino acids (Teriete et al., 2004).This review will focus on CD44 and TSG-6 describing their involvement in immuneprocesses, in particular their roles in the HA-mediated interactions of leukocyteswith the vascular endothelium at sites of inflammation.

2. CD44

CD44 is ubiquitously expressed on leukocytes, and many other cell types (includ-ing tumour cells), where it represents the major cell surface receptor for hyaluronan(Arrufo et al., 1990; Lesley et al., 1997; Pure and Cuff, 2001; Terieteffff et al., 2004;Toole, 2004). Although CD44 constitutes a family of hyaluronan receptors varyingin size from ~80 to 250 kDa arising from alternate splicing (Pure and Cuff, 2001),ffffthe major species on leukocytes is the standard or hematopoetic form (CD44H) thatdoes not utilise variant exons. As can be seen in Fig. 2, CD44 consists of anN-terminal extracellular HA-binding domain (highly conserved in mammals with

59

Figure 2. The domain structure of CD44.

~85% amino acid identity; Isacke and Yarwood, 2002), a membrane proximalextracellular domain, a transmembrane domain and a C-terminal cytoplasmic tail.

As part of the immune response, circulating leukocytes become localised at sitesof inflammation via their adhesion to, and rolling along, the vascular endothelium;this allows sampling of the local environment potentially leading to tight adhesionand transendothelial migration (Butcher, 1991; Kubes and Kerfoot, 2001; Siegelmanet al., 2000). In the case of lymphocytes and monocytes the initial tethering androlling can be mediated by CD44 (on their cell surfaces) interacting with HA immo-bilised on the blood vessel wall (DeGrendele et al., 1996; Mohamadzadeh et al.,1998). Importantly, CD44 on circulating leukocytes does not bind HA constitutively,but can be induced to do so in response to inflammatory signals (Lesley et al., 1993;1997), which also lead to increased expression of HA by endothelial cells; tightregulation of these processes is essential to avoid host tissue damage due to inappro-priate extravasation (e.g., as seen in autoimmune disease; Siegelman et al., 1999).The induction of CD44’s HA-binding activity can be mediated in a number ofdifferent ways. For instance, activated leukocytes change their cell surface CD44 toffffa HA-binding form after stimulation by cytokines, such as tumour necrosis factorand interferon-c, or through T cell receptor triggering (DeGrendele et al., 1997; Pureand Cuff, 2001). Alternatively, CD44 can be induced to bind HA by factors thatffffinteract directly with the receptor; e.g., there are certain monoclonal antibodiesagainst the CD44_HABD that can induce it to bind HA (Lesley et al., 1993; Zhenget al., 1995). As discussed in detail below our recent studies on CD44 and TSG-6have provided novel insights into the molecular basis of CD44 regulation.

3. ANALYSIS OF THE HA-BINDING SURFACE OF CD44

The N-terminal HA-binding domain of CD44 (CD44_HABD) contains a Linkmodule, together with flanking N- and C- terminal sequences that are necessary to

Hyaluronan in Immune Processes


Figure 3. Leukocytes can adhere to and roll along the blood vessel endothelium via CD44-HA interactions.

Figure 4. Cartoons showing the backbone folds of (A) the Link module of TSG-6 (Link_TSG6) and (B)

the~150 amino acid CD44 HA-binding domain (CD44_HABD). In CD44 the N- and C-terminal flankingsequences (coloured black) extend the b sheet structure of the Link module.

form a stable fold (Peach et al., 1993; Banerji et al., 1998). The recent determinationof the 3D structure of the CD44_HABD, by both X-ray crystallography and NMR(TerieteTT et al., 2004), has revealed that the N- and C- terminal extensions, which arelinked by a disulphide bond, form an extra lobe of structure in intimate contact withone side of the Link module (Fig. 4). This structure has given new information onhow CD44 may bind HA and how this interaction may be regulated by N-linkedglycosylation (Teriete et al., 2004).

Site-directed mutagenesis of CD44 (Peach et al., 1993; Bajorath et al. 1998)

61

revealed that four residues in the Link module were essential for HA binding (i.e.,Arg41, Tyr42, Arg78 and Tyr79), where mutation of any one of these caused completeloss of functional activity. In addition, a further ten ‘important’ residues were implic-ated in HA recognition, with five located in the Link module (Lys38, Lys68, Asn100,Asn101 and Tyr105) and five in the extra lobe formed from the N- (Arg29) andC-terminal extensions (Arg150, Arg154, Lys158 and Arg162). Mapping these aminoacids onto the CD44_HABD structure shows that the four key binding residuesform a continuous patch on one face of the molecule (which also includes Lys38),and that Asn100, Asn101 and Tyr105 are clustered close by, extending the bindingsite to one side (Teriete et al., 2004). The basic amino acids in the C-terminalextension are also on this face of the CD44_HABD, but are widely spaced, andArg29 and Lys68 are on the opposite side of the structure (see Fig. 5).

NMR studies on CD44 have determined that almost an entire face of the HABDis highly perturbed on its interaction with HA; the chemical shift changes are centredround Arg41 and Tyr42, and overlap appreciably with the ligand-binding residuesdetermined experimentally (Teriete et al., 2004). However, these perturbations aretoo widespread to all be caused solely by the direct interaction of a single HAmolecule and a more likely explanation is that the protein undergoes a conforma-tional change on binding, which is consistent with the observed perturbation of thehydrogen bond network (Teriete et al., 2004). The widespread distribution of aminoacids implicated by mutagenesis is also difficult to reconcile with the recognition ofa single HA molecule. Therefore, it has been suggested (Teriete et al., 2004) thatCD44 could accommodate HA in two different, mutually exclusive, binding positionsffffor modes (see Fig. 5), where mode 1 is likely to represent the HA-binding siteconserved across the Link module superfamily (Blundell et al., 2003). While thishypothesis does not explain the involvement of Arg29 or Lys68 in the interactionwith HA, it seems unlikely that, given their locations, they could play a direct rolein binding. In this regard, it cannot be ruled out that the mutation of these, or otherresidues implicated as functional, might have a deleterious effect on the formationffffof an optimal binding conformation.

4. REGULATION OF CD44 FUNCTION BY N-GLYCOSYLATIONNN

As mentioned above, CD44, although present on the surface of circulatingleukocytes, does not bind HA constitutively. However, during inflammation, activeCD44 molecules are formed that can mediate cell attachment and rolling (reviewedin Lesley et al., 2004; Teriete et al., 2004). While there are several possible ways inwhich this could occur, one of the major mechanisms of receptor activation has beenshown to involve remodelling of N-linked glycans on CD44. Importantly, the levelof cell surface N-glycosylation dictates the activation state of CD44, and the enzym-atic removal of sialic acid (e.g., by an endogenous sialidase; Gee et al., 2003; Katohet al., 1999), or inhibition of N-glycan biosynthesis, induces HA binding (Englishet al., 1998; Katoh et al., 1995; Lesley et al., 1995). In this regard, there are five N-linked glycosylation sites in the HABD of human and mouse CD44 (residues 25, 57,100, 110, 120 in the former), and the individual mutation of two of these (i.e., Asn25



Figure 5. A) A surface representation of the CD44_HABD (in two orientations; one 180° around a verticalaxis from the other) showing the relative positions of the essential binding residues (dark grey) and those

implicated as functionally important ( lighter grey). B) The widespread nature of the amino acids implicated

from mutagenesis could be explained by two different modes of HA binding, which are represented byffff

arrows; the HA is shown such that the arrowheads are at its non-reducing termini, which is the polarity

determined previously for HA binding to Link_TSG6 (Blundell et al., 2003). The CD44_HABD in (B) is

shown in the same orientation as the left-hand molecule in (A).

or Asn120) converts CD44 from an inducible to constitutively active state on lympho-cyte cell lines (English et al., 1998). The recent elucidation of the tertiary structureof the HA-binding domain has allowed the position of these key regulatory glycosyl-ation sites to be determined (Teriete et al., 2004). As can be seen from Fig. 5, Asn25is present on the HA-binding face of CD44 in close proximity to the mode 1interaction site. It is not difficult to imagine, therefore, how the presence of glycosyl-ation at this location could sterically block ligand recognition (Fig. 6). In the caseof Asn120, it is less obvious how the attachment of an N-glycan could inhibit bindingsince it is located on the opposite face of the protein. The most likely explanation isthat glycosylation at Asn120 has the potential to interfere with CD44 self-associationand so prevent receptor clustering (see Fig. 6), a process that is known to be

63

Figure 6. A schematic diagram showing the effects offfff N-linked glycosylation on HA-CD44 interactions.

The presence of certain N-glycans at Asn25 and Asn120 may have an inhibitory effect on HA binding byffff

CD44+ cells and their removal (or remodelling) could either unlock the binding groove or facilitatereceptor clustering, respectively.

important in CD44 regulation. For example, it is well established that cross-linkingof CD44 by certain antibodies (such as IRAWB14) can trigger large increases in itsHA-binding activity on leukocyte (Lesley et al., 1993; 2000). Interestingly, the epitopefor IRAWB14 (Zheng et al., 1995) lies immediately adjacent to Asn120 in the CD44structure and is centred on Lys68, one of the two residues on the rear face of theHADB whose mutation affects HA binding (Bajorathffff et al., 1998). It seems likely,therefore, that this face of CD44 plays an important role in the interaction with HA,but further work is required to determine its contribution to CD44 oligomerisationand exactly how this is modulated by changes in glycosylation. Conflicting reportson the effect of glycosylation on CD44-HA interactions (Bartolaziffff et al., 1996; Englishet al., 1998; Zheng et al. 1997) indicates there is considerable complexity to beunderstood, much of which is likely to result from the tissue and cell specific natureof CD44 biology.

5. TSG-6

TSG-6 is a 35 kDa HA-binding protein that is not generally present in healthyadult tissues, but is expressed in response to inflammatory cytokines and certain



growth factors (Lee et al., 1992; Milner and Day, 2003; Wisniewski and Vilcek,2004). Although TSG-6 has been found to be associated with diseases such asarthritis, it is likely to have an anti-inflammatory role, acting as part of a negativefeedback loop to resolve inflammation (Getting et al., 2002). For example, TSG-6 isa potent down regulator of cytokine-induced neutrophil migration, via its inhibitionof polymorphonuclear-endothelial cell interactions (Cao et al., 2004). In this regard,deletion of the TSG-6 gene leads to increased extravasation of neutrophils intosynovial tissues in murine models of arthritis causing extensive joint destruction,whereas overexpression of TSG-6 has a chondroprotective effect (see Szaffff nto et al.,2004). However, it should be noted that the inhibition of neutrophil migration byTSG-6 is likely to be independent of its HA-binding function (Getting et al., 2002).TSG-6 is also expressed in the context of inflammation-like processes such asovulation, where it serves to stabilise the HA-rich extracellular matrix that formsaround the oocyte during cumulus matrix expansion via its interaction with inter-a-inhibitor and PTX3 (Day et al., 2004; Fulop et al., 2003; Salustri et al., 2004). HAcross-linking of this type is likely to occur at inflammatory sites (e.g., in articularjoint disease), where the HA/protein complexes formed could potentially have alteredleukocyte-binding properties (see Day et al., 2004).

The HA-binding domain of TSG-6 consists of a single Link module (denotedLink_TSG6) that has been extensively characterized by NMR spectroscopy and site-directed mutagenesis (Blundell et al., 2003; Getting et al., 2002; Kahmann et al.,2000; Kohda et al., 1996; Mahoney et al., 2001). For instance, we have recentlydetermined the solution structures of its free and HA-bound forms, revealing that aconformational change occurs in the Link module on its interaction with HA(Blundell et al., 2003). This exposes a HA-binding groove, lined with the key func-tional amino acids implicated by mutagenesis studies (Getting et al., 2002; Mahoneyet al., 2001), which is in a similar position to the proposed mode 1 interaction sitein CD44 (Teriete et al., 2004). Although the locations of the HA-interaction surfacesare likely to be conserved in TSG-6 and CD44, there appear to be some majordifferences in the details of the residues and sequence positions involved in mediatingffffHA binding in the two proteins (Mahoney et al., 2001). It is likely therefore, thatthe interaction networks will be distinct in these HA-protein complexes, potentiallyinvolving the stabilisation of HA in different bound conformations.ffff

6. TSG-6 ENHANCES HYALURONAN BINDING TO CELLSURFACE CD44

Given that TSG-6 is a HA-binding protein, which is expressed by a wide rangeof cell types in response to inflammatory stimuli (e.g., monocytes, macrophages,neutrophils, dendritic cells, microvascular endothelium, vascular smooth muscle cellsand fibroblasts (Lesley et al., 2004; Milner and Day, 2003)), it was thought possiblethat it may act as a competitive inhibitor of CD44-mediated leukocyte migration(Lee et al., 1992). This hypothesis has recently been tested (Lesley et al., 2004) andit was found that, rather than being an inhibitor of CD44-HA adhesion, TSG-6 isable to enhance or induce HA binding to CD44 on constitutive or inducible lymphoidcell backgrounds, respectively. In the latter case, EL4 cells (derived from a T cell

65

Figure 7. The interaction of TSG-6 with HA may lead to the formation of cross-linked HA fibres.

TSG-6/HA complexes of this type could induce receptor clustering, leading to an increase in HA binding.

lymphoma), which express CD44 on their cell surface but do not interact constitu-tively with HA (i.e., they can be considered a model for circulating lymphocytes),could be induced to bind HA in a manner reminiscent of the superagonist monoclonalantibodies (e.g., IRAWB14) discussed above (Lesley et al., 2004). Importantly,enhancement by TSG-6 of the CD44-mediated interaction of lymphoid cell lineswith HA was seen under conditions of flow at shear forces comparable to those thatoccur in post-capillary venules. In order for TSG-6 to have its modulatory effect itffffwas found necessary to preincubate the HA with TSG-6 at concentrations thatsaturate the majority of the protein-binding sites on the polysaccharide. Thesemixtures retained their activity over time, even after dilution, suggesting the formationof stable TSG-6/HA complexes. In this regard, the recombinant Link module domain(i.e., Link_TSG6) was also able to enhance/induce CD44’s interaction with HA(albeit with less potency than the full length protein), where mutation of HA-bindingresidues impaired its modulatory effect. This, and the lack of a detectable interactionffffof TSG-6 with CD44+ cells in the absence of HA, support the conclusion that theenhancing/inducing activity depends solely on the interaction of TSG-6 with HA.The demonstration that amino acids outside the HA-binding site are also involvedin this activity gave rise to the hypothesis that TSG-6 may self-associate leading tothe cross-linking of multiple HA chains to form extended fibres (Lesley et al., 2004).As shown in Fig. 7 engagement of TSG-6/HA fibres with cell surface CD44 couldpromote receptor clustering leading to an increase/activation of HA binding due toavidity effects or, potentially, changes in CD44 conformation. A similar mechanismffff



is likely to be involved in the binding of HA cable-like structures (produced forexample by mucosal smooth muscle cells in inflamed human colon) to CD44 onnon-activated monocytes, which are not able to interact constitutively with free HA(see de la Motte et al., 2004).The finding that CD44-expressing EL4 cells (that do not constitutively bind

HA) are able to interact with TSG-6/HA complexes and exhibit enhanced rollingon TSG-6/HA substrates, provides a mechanism whereby circulating lymphocytesmight become adhesive in an inflammatory milieu where such complexes could bepresent (Lesley et al., 2004); both HA and TSG-6 are upregulated in blood vesselsduring inflammation. For example, it is possible that retention of TSG-6/HA com-plexes on endothelial cells could facilitate the CD44-mediated recruitment of leuko-cytes and thus be proinflammatory. Alternatively, release of TSG-6/HA complexesinto the local circulation and their binding to CD44 on leukocytes could inhibit theadhesive interactions of circulating cells with the vascular endothelium, and therebyhave an anti-inflammatory effect. In this regard, TSG-6ffff /HA complexes were foundto be effective competitors of CD44ffff + cell attachment and rolling on immobilisedHA (Lesley et al., 2004). However, further research is necessary to address the exactrole of TSG-6 in the regulation of CD44-mediated leukocyte migration.

7. CONCLUSIONS

The polysaccharide HA is important in mediating the initial attachment androlling of mononuclear leukocytes on the vascular endothelium at sites of inflamma-tion through its interaction with CD44, a process that necessitates tight regulation.Recent structural studies have provided novel insights into how changes in the N-linked glycosylation of CD44 could either unblock the HA-binding site or allowreceptor clustering, thus switching CD44 into an active state. The inflammation-associated hyaladherin TSG-6 has also been implicated in the positive modulationof CD44-HA interactions on lymphocytes via a mechanism that is likely to involvethe formation of cross-linked HA fibres which can engage and cluster multiplereceptor molecules on the leukocyte surface. Further work is now required toinvestigate the formation and structure of these and other multi-molecularHA/protein complexes, to allow a better understanding of their roles in the regulationof immune cell adhesion.

8. ACKNOWLEDGEMENTS

We would like to thank Dr. Peter Teriete for providing Figs. 2 and 6, and Dr.Caroline M. Milner for review of the manuscript. AJD acknowledges the support ofthe ARC and MRC, and AJW was the recipient of a MRC Studentship.

REFERENCES

Aruffo, A., Stamenkovic, I., Melnick, M., Underhill, C.B., and Seed, B., 1990, CD44 is the principal cellffff

surface receptor for hyaluronate., Cell, 61:1303–1313.

67

Banerji, S., Day, A.J., Kahmann, J.D. and Jackson, D.G., 1998, Characterization of a functional hyalu-

ronan-binding domain from the human CD44 molecule expressed in Escherichia coli. Protein Expr.

Purif., 14:371–381.

Bajorath, J., Greenfield, B., Munro, S.B., Day, A.J. and Aruffo, A., 1998, Identification of CD44 residuesffff

important for hyaluronan binding and delineation of the binding site. J. Biol. Chem. 273:338–343.

Bartolazzi, A., Nocks, A., Aruffo, A., Spring, F. and Stamenkovic I., 1996, Glycosylation of CD44 isffff

implicated in CD44-mediated cell adhesion to hyaluronan. J. Cell Biol., 132:1199–1208.

Blundell, C.D., DeAngelis, P.L., Day, A.J. and Almond, A., 2004a, Use of 15N-NMR to resolve moleculardetails in isotopically-enriched carbohydrates: sequence-specific observations in hyaluronan oligo-

mers up to decasaccharides. Glycobiology, 14:999–1009.

Blundell, C.D., Mahoney, D.J., Almond, A., DeAngelis, P.L., Kahmann, J.D., Teriete, P., Pickford, A.R.,

Campbell, I.D. and Day, A.J., 2003, The link module from ovulation- and inflammation-associated

protein TSG-6 changes conformation on hyaluronan binding. J. Biol. Chem., 278:49261–49270.

Blundell, C.D., Seyfried, N.T. and Day, A.J., 2004b, Structural and functional diversity of hyaluronan-

binding proteins. In: Chemistry and Biology of Hyaluronan (H.G. Garg and C. A. Hales eds),

Elservier, pp189–204.

Butcher, E.C., 1991, Leukocyte-endothelial cell recognition: three (or more) steps to specificity and diversity.

Cell, 67:1033–1036.

Cao, T.V., La M., Getting, S.J., Day, A.J. and Perretti, M., 2004, Inhibitory effects of TSG-6 Link moduleffff

on leukocyte-endothelial cell interactions in vitro and in vivo. Micocirculation, 11:615–624.

Day, A.J. and Prestwich, G.D., 2002, Hyaluronan-binding proteins: tying up the giant. J. Biol. Chem.,

277:4575–4579.

Day, A.J., Rugg, M.S., Mahoney, D.J. and Milner, C.M., 2004, The role of hyaluronan-binding proteins in

ovulation. In: HA2003 Proceedings at http://www.matrixbiologyinstitute.org/ha03/toc.htm.

Day, A.J. and Sheehan, J.K., 2001, Hyaluronan: polysaccharide chaos to protein organisation. Curr. Opin.

Struc. Biol., 11:617–622.

de la Motte, C.A., Hascall, V.C., Drazba, J., Bandyopadhyay, S.K. and Strong, S.A., 2003, Mononuclear

leukocytes bind to specific hyaluronan structures on colon mucosal smooth muscle cells treated with

polyinosinic acid:polycytidylic acid: inter-alpha-trypsin inhibitor is crucial to structure and function.

Am. J. Pathol., 163:121–133.

de la Motte, C.A., Drazba, J., Bandyopadhyay, S., Majors, A., Hascall, V.C. and Strong, S.A., 2004, Viral

stimuli induce novel hyaluronan cable structures on colon smooth muscle cells that bind leukocytes

externally and nuclei internally. In: HA2003 Proceedings at http://www.matrixbiologyinstitute.org/

ha03/toc.htm.

DeGrendele, H.C., Estess, P., Picker, L.J. and Siegelman, M.H., 1996, CD44 and its ligand hyaluronate

mediate rolling under physiological flow: A novel lymphocyte-endothelial cell primary adhesion

pathway. J. Exp. Med., 183:1119–1130.

DeGrendele, H.C., Kosfiszer, M., Estess, P. and Siegelman, M.H., 1997, CD44 activation and associated

primary adhesion is inducible via T cell receptor stimulation. J. Immunol., 159:2549–2553.

English, N.M., Lesley, J.F. and Hyman, R., 1998, Site-specific de-N-glycosylation of CD44 can activate

hyaluronan binding, and CD44 activation states show distinct threshold densities for hyaluronan

binding. Cancer Res., 58:3736–3742.

Fraser, J.R., Laurent, T.C. and Laurent, U.B., 1997, Hyaluronan: its nature, distribution, functions and

turnover. J. Intern. Med., 242:27–33.

Fulop, C., Szanto, S., Mukhopadhyay, D., Bardos, T., Kamath, R.V., Rugg, M.S., Day, A.J., Salustri, A.,

Hascall, T.T. and Mikecz K., 2003, Impaired cumulus mucification and female sterility in tumor

necrosis factor-induced protein-6 deficient mice. Development, 130:2253–2261.

Getting, S.J., Mahoney, D.J., Cao, T., Rugg, M.S., Fries, E., Milner, C.M., Perretti, M. and Day, A.J., 2002,

The link module from human TSG-6 inhibits neutrophil migration in a hyaluronan- and inter-a-

inhibitor-independent manner. J. Biol. Chem., 277:51068–51076.

Gee, K., Kozlowski, M. and Kumar, A., 2003, TNFa induces functionally active hyaluronan-adhesive

CD44 by activating sialidase through p38 mitogen-activated protein kinase in lipopolysaccharide-

stimulated human monocytic cells. J. Biol. Chem., 278:37275–37287.

Hardingham, T.E., 2004, Solution properties of HA. In: Chemistry and Biology of Hyaluronan (H.G. Garg

and C. A. Hales eds), Elservier, pp1–19.

Isacke, C.M. and Yarwood, H., 2002, The hyaluronan receptor, CD44. Int. J. Biochem. Cell Biol.,

34:718–721.



Kahmann, J.D., O’Brien, R., Werner, J.M., Heinegard, D., Ladbury, J.E., Campbell, I.D. and Day A.J.,

2000, Localization and characterization of the hyaluronan-binding site on the link module from

human TSG-6. Structure Fold Des., 8:763–774.

Katoh, S., Miyagi, T., Taniguchi, H., Matsubara, Y., Kadota, J., Tominaga, A., Kincade, P.W., Matsukura,

S. and Kohno, S., 1999, An inducible sialidase regulates the hyaluronic acid binding ability of

CD44-bearing monocytes. J. Immunol., 162:5058–5061.

Katoh S., Zheng Z., Oritani K., Shimozato T. and Kincade P.W. (1995). Glycosylation of CD44 negatively

regulates its recognition of hyaluronan. J. Exp. Med., 182:419–429.

Kohda, D., Morton, C.J., Parkar, A.A., Hatanaka, H., Inagaki, F.M., Campbell, I.D. and Day A.J., 1996,

Solution structure of the link module: a hyaluronan-binding domain involved in extracellular matrix

stability and cell migration. Cell, 86:767–775.

Kubes, P. and Kerfoot, S.M., 2001, Leukocyte recruitment in the microcirculation: the rolling paradigm

revisited. News Physiol. Sci., 16:76–80.

Laurent, T.C, Laurent, U.B and Fraser, J.R., 1996, The structure and function of hyaluronan: An overview.

Immunol. Cell Biol., 74:A1–7.

Lee, H.L., Wisniewski, H.G. and Vilcek, J., 1992, A novel secretory tumor necrosis factor-inducible protein

(TSG-6) is a member of the family of hyaluronan binding proteins, closely related to the adhesion

receptor CD44. J. Cell Biol., 116:545–557.

Lesley, J., English, N., Perschl, A., Gregoroff, J.ffff and Hyman, R., 1995, Variant cell lines selected for

alterations in the function of the hyaluronan receptor CD44 show differences in glycosylation. J. Exp.ffff

Med., 182:431–437.

Lesley, J., Gal, I., Mahoney, D.J., Cordell, M.R., Rugg, M.S., Hyman, R., Day, A.J. and Mikecz, K., 2004,

TSG-6 modulates the interaction between hyaluronan and cell surface CD44. J. Biol. Chem.,

279:25745–25754.

Lesley, J., Hascall, V.C., Tammi, M. and Hyman, R., 2000, Hyaluronan binding by cell surface CD44. J.

Biol. Chem., 275:26967–26975.

Lesley, J., Hyman, R., English, N., Catterall, J.B. and Turner, G.A., 1997, CD44 in inflammation and

metastasis. Glycoconj. J., 14:611–622.

Lesley, J., Kincade, P.W. and Hyman, R., 1993, Antibody-induced activation of the hyaluronan receptor

function of CD44 requires multivalent binding by antibody. Eur. J. Immunol., 8:1902–1909.

Mahoney, D.J, Blundell, C.D and Day, A.J., 2001, Mapping the hyaluronan-binding site on the link module

from human tumor necrosis factor-stimulated gene-6 by site-directed mutagenesis. J. Biol. Chem.,

276:22764–22771.

Majors, A.K., Austin, R.C., de la Motte, C.A., Pyeritz, R.E., Hascall, V.C., Kessler, S.P., Sen, G. and Strong,

S.A., 2003, Endoplasmic reticulum stress induces hyaluronan deposition and leukocyte adhesion. J.

Biol. Chem., 278:47223–47231.

Milner, C.M. and Day, A.J., 2003, TSG-6: a multifunctional protein associated with inflammation. J. Cell.

Sci., 116:1863–1873.

Mohamadzadeh, M., DeGrendele, H., Arizpe, H., Estess, P. and Siegelman, M., 1998, Proinflammatory

stimuli regulate endothelial hyaluronan expression and CD44/HA-dependent primary adhesion. J.

Clin. Invest., 101:97–108.

Peach, R.J., Hollenbaugh, D., Stamenkovic, I. and Aruffo, A., 1993, Identification of hyaluronic acidffff

binding sites in the extracellular domain of CD44. J. Cell Biol., 122:257–264.

Pure, E. and Cuff, C.A., 2001, A crucial role for CD44 in inflammation. Trends. Mol. Med., 7:213–221.ffff

Salustri, A., Garlanda, C., Hirsch, E., De Acetis, M., Maccagno, A., Bottazzi, B., Doni, A., Bastone, A.,

Mantovani, G., Beck Peccoz, P., Salvatori, G., Mahoney, D.J., Day, A.J., Siracusa, G., Romani, L.

and Mantovani A., 2004, PTX3 plays a key role in the organization of the cumulus oophorus

extracellular matrix and in in vivo fertilization. Development, 131:1577–1586.

Siegelman, M.H., DeGrendele, H.C. and Estess, P., 1999, Activation and interaction of CD44 and hyalu-

ronan in immunological systems. J. Leukocyte Biol., 66:315–321.

Siegelman, M.H., Stanescu, D. and Estess, P., 2000, The CD44-initiated pathway of T-cell extravasation

uses VLA-4 but not LFA-1 for firm adhesion. J. Clin. Invest., 105:683–691.

Szanto, S., Bardos, T., Gal, I., Glant, T.T. and Mikecz, K., 2004, Enhanced neutrophil extravasation and

rapid progression of proteoglycan-induced arthritis in TSG-6-knockout mice. Arthritis Rheum.,

50:3012–3022.

Tammi, M.I., Day, A.J. and Turley, E.A., 2002, Hyaluronan and homeostasis: a balancing act. J. Biol.

Chem., 277:4575–4579.

69

Teriete, P., Banerji, S., Noble, M., Blundell, C.D., Wright, A.J., Pickford, A.R., Lowe, E., Mahoney, D.J.,

Tammi, M.I., Kahmann, J.D., Campbell, I.D., Day, A.J. and Jackson, D.G., 2004, Structure of the

regulatory hyaluronan binding domain in the inflammatory leukocyte homing receptor CD44. Mol.

Cell, 13:483–496.

Toole, B.P., 2004, Hyaluronan: from extracellular glue to pericellular cue. Nat. Rev. Cancer, 4:528–539

Wisniewski, H.G. and Vilcek, J., 2004, Cytokine-induced gene expression at the crossroad of innate

immunity, inflammation and fertility: TSG-6 and PTX3/TSG-14. Cytokine Growth Factor Rev.,

15:129–146.

Zheng, Z., Cummings, R.D., Pummill, P.E. and Kincade, P.W., 1997, Growth as a solid tumor or reduced

glucose concentrations in culture reversibly induce CD44-mediated hyaluronan recognition by

Chinese hamster ovary cells. J. Clin. Invest., 100:1217–1229

Zheng, Z., Katoh, S., He, Q., Oritani, K., Miyake, K., Lesley, J., Hyman, R., Hamik, A., Parkhouse, R.M.

and Farr, A.G., 1995, Monoclonal antibodies to CD44 and their influence on hyaluronan recognition.

J. Cell. Biol., 130:485–95.


12

GLYCOSYLATION AND THE FUNCTION OF THET CELL CO-RECEPTOR CD8

David A. Shore1,2, Ian A. Wilson2,3, Raymond A. Dwek1 andPauline M. Rudd1

1The Glycobiology InstituteDepartment of BiochemistryUniversity of OxfordOxford, OX1 3QU, UK2Department of Molecular Biology3Skaggs Institute for Chemical BiologyThe Scripps Research Institute10550 North Torrey Pines RoadLa Jolla, CA 92037, USA.

1. INTRODUCTION

The CD8 glycoprotein functions at the surface of cytotoxic T-lymphocytes(CTLs) as an essential co-receptor in T cell activation in response to peptide antigencomplexed with the class I major histocompatibility complex (pMHC). CD8 interactswith the class I pMHC in a bidentate manner with the T cell receptor (TCR), toinitiate and augment T-cell signalling through the CD3 subunits of the TCR complex(reviewed in Devine, 1999; Garcia, 1999; Gao, 2000; Wang, 2000; van der Merwe,2003).In the absence of CD8 at the cell surface, the CTL response to peptide antigen

is significantly impaired (Luescher, 1995). Likewise, the CD8/pMHC interactiongreatly enhances TCR-mediated T cell sensitivity to antigen (Purbhoo, 2001). TheCD8 glycoprotein comprises two distinct subunits, alpha(a) and beta(b), and isexpressed at the cell surface as a mixture of covalently associated aa homodimersand ab heterodimers. Activation of the genes encoding the a and b subunits of CD8is under the control of separate promoters, and the expression of CD8aa and CD8abis highly cell-type specific (Gangadharan, 2004 and references therein).

The ab heterodimeric form of CD8 is the abundant form of the co-receptor atthe surface thymocytes and mature conventional class I restricted T cells (reviewed



72 D. A. Shore et al.

in Zamoyska, 1994). CD8ab is the primary co-receptor for abTCR-mediated signal-ling in response to antigen, and is 100 fold more effective than CD8ffff aa (Arcaro, 2000;Arcaro, 2001; Cawthon, 2001). Mice deficient in CD8b exhibit a substantially reducednumber of peripheral CD8+ T cells, implicating CD8b in the development anddifferentiation of such cells (Crooks, 1994, Fung-Leung, 1994).ffff

The aa and ab isoforms of CD8 are functionally distinct (reviewed inGangadharan, 2004). In contrast to CD8ab, expression of CD8aa is not limited toclass I-restricted T cells. The aa receptor has been identified on the surface of anumber of different cell types, including CD4ffff + T cells, lymphoid-related dendriticcells, cdTCR cells, NK cells and intraepithelial lymphocytes of the gut. It is thoughtthat CD8aa acts in a class I pMHC-independent manner to elicit a regulatory rolein T cell function. Specifically, a strong interaction between CD8aa and the thymicleukaemia antigen (TL) has been demonstrated (Devine, 2002; Liu, 2003). In thisinstance, the TCR-independent interaction between CD8aa and the non-antigenpresenting TL is thought to modify TCR activation signals received by antigen-stimulated cells (Cheroutre, 1995; Leishman, 2001).

2. STRUCTURE OF THE CD8 GLYCOPROTEIN

Despite a sequence identity of only 20%, the a and b subunits are predicted tohave a similar topology (Norment, 1988). Each subunit comprises an ectodomainof 150–170 amino acids, a type-2 single pass transmembrane domain and a shortcytoplasmic region (Fig. 1a). The CD8 ectodomain consists of an amino terminalimmunoglobulin V-set domain, tethered to the membrane by a 30–50 residue flexiblestalk. The intracellular region of the a subunit of CD8 interacts with the proteintyrosine kinase LckP56 through a conserved CxCP motif (Fung-Leung, 1993; Arcaro,2001; Kim, 2003).The stalk region of both CD8 subunits is rich in proline, serine and threonine

residues, and is highly O-glycosylated in all species studied. These characteristicslikely confer an extended but relatively rigid conformation to the stalk region, as inleukosialin and mucins (reviewed in Rudd, 1999). Amino acid sequencing of themembrane-distal region of rat CD8a has determined that four threonine residues,T122, T126, T132 and T134 are post-translationally modified with O-linked carbo-hydrates (Classon, 1992). The distribution of potential N-linked glycosylation sitesbetween the a and b subunits is highly variable across species. The presence of atleast one N-linked glycosylation site on the b subunit is common to all species,whereasN-glycosylation of the a subunit appears to be species specific (Merry, 2003).

3. INTERACTION OF CD8 WITH CLASS I MHC

Binding studies involving soluble forms of the murine and human CD8 andclass I MHC have shown that the affinities for class I pMHC of both the aa and abdimers in solution are equivalent (Sun, 1997, Kern, 1999). Furthermore, CD8 inter-actions are generally an order of magnitude weaker than those involving the TCR,although TCR/pMHC complex affinity is highly dependent upon TCR haplotype

Glycosylation and the Function of the T Cell Co-receptor CD8 73

Figure 1. The overall topology and function of the CD8ab co-receptor.

(a) The a and b subunits of CD8 are covalently associated via disulphide bonds within the stalk region

of the receptor. The a and b Ig-like regions dimerise to form the class I pMHC binding site. The a subunit

of CD8 is associated with the intracellular phospho-tyrosine kinase p56LcK. The b subunit stalk is shorterthan that of the a subunit, potentially producing a bowed effect in theffff ab dimer. (b) The IgSF head group

of CD8 spans the intercellular gap to bind the membrane proximal region of the class I MHC. The

flexibility inherent within the CD8 stalk region likely enables the head group to adopt the optimal

orientation for pMHC interaction. Interaction between CD8 and class I pMHC brings the associated

p56LcK tyrosine kinase into close association with the ITAMs of the TCR/CD3 complex, such that anintracellular signaling cascade is initiated/augmented and maintained for the duration of the CD8/MHC

interaction.


and antigen (i.e. strong or weak agonist) and varies substantially (Kern, 1999; Garcia,1996; 2003 van der Merwe). Hence, although recruitment of CD8 to thenascent pMHC/TCR may somewhat contribute to the overall affinity of the TCRfor pMHC, it is unlikely that the primary co-receptor function of CD8 is the additionof structural support. It seems more probable that CD8 functions predominantly tobring the protein kinase LcKp56 into close association with the intracellular regionof the CD3 subunits, giving rise to TCR/CD3 signal transduction when boundto pMHC (Fig. 1b) (Arcaro, 2001; Purbhoo, 2001; Doucey, 2003; Palacios, 2004).

The crystal structure of a murine CD8aa immunoglobulin super family (IgSF)domain dimer in complex with the class I MHC H-2Kb has been determined (Kern,1998). This structure describes an interaction between the membrane proximal regionof the MHC and the six CDR-equivalent loops of the CD8 dimer in an antibody-like manner, perpendicular to the long axis of the MHC. In this model, the contribu-tion of the individual CD8 subunits to the binding interface is asymmetric, with oneof the CD8a subunits in the ‘‘upper’’ a1 position accounting for~70% of the bindingsite (Fig. 2). Both CD8 subunits contact the protruding acidic loop of the non-polymorphic MHC a3 domain which is considered the primary binding site; however,the a1 subunit of the CD8 dimer makes additional contacts with the MHC a2domain, as well as with b2-microglobulin (Kern, 1998; Li, 1998). This mode ofinteraction was also observed in the crystal structure of the human form of CD8aain complex with the human class I MHC molecule HLA-A2 (Gao, 1997). Currentlyno structural data are available to determine the orientation of the CD8b subunitin complex with class I MHC, although given the predicted similarities between IgSfdomains of the two subunits, it is likely that binding of CD8ab to class I MHC isequivalent to that of CD8aa. Mutations in the CDR-like regions of the CD8asubunit have a dramatic effect upon the ability of CD8ffff ab to interact with pMHC,whereas similar mutations in the b subunit have little effect upon complex formationffff(Devine, 1999a). Furthermore, transfection studies to express a chimeric human bbhomodimer at the cell surface suggest that this form of the co-receptor is incapableof interacting with conventional class I MHC, indicating a critical role for the asubunit in co-receptor binding (Devine, 2000). These findings suggest that the asubunit dominates the binding site with MHC, whereas the b subunit occupies the‘‘lower’’ a2 position in the interface. However, the shorter b subunit stalk wouldappear to favour a model in which the b subunit occupies the ‘‘upper’’ a1 positionin the CD8/MHC complex (as in Fig. 1). The precise position of the CD8ab dimerrelative to the MHC, therefore, remains a contentious issue which must await aCD8ab/pMHC complex crystal structure.

4. FUNCTION OF THE CD8b SUBUNIT

It is apparent that the b subunit contributes significantly to the function of CD8as a co-receptor to TCR mediated signalling (Irie, 1998; Witte, 1999; Bosselut, 2000).However, the mechanism underlying the enhanced effector function of CD8ffff b overthat of CD8a is unclear. It has been suggested that binding of the b subunit to classI pMHC induces a conformational change within the class I receptor, which in turnenhances the interaction between the pMHC and TCR (Garcia, 1996). Given the


Figure 2. The interaction between the murine CD8aa head group and the class I pMHC H-2Kb.The domain architecture of the CD8aa head group dimer is similar to that of an immunoglobulin Fab

variable region. The 6 CDR-like loops of the CD8aa dimer clamp around an extended loop between the

C and D strand of the class I pMHC a3 domain. Binding of the CD8a subunits to class I pMHC is

asymmetric; the ‘‘upper’’ (CD8a1) subunit accounts for ~70% of the solvent exposed area within theCD8 binding pocket. Interaction between the CD8a1 subunit and the pMHC are strengthened by addi-

tional interactions with the MHC a2 domain, as well as with the CD8a DE loop and the b2M domain.The interaction between CD8 and pMHC is of relatively low affinity (~60 mM) and as such is likely tobe reversible and sensitive to avidity effectsffff in vivo.

current structural information for the class I pMHC, both in complex with the TCRand in complex with the CD8, it is not clear how such a change may be instigatedat the CD8 binding site and transmitted through the MHC to the peptide bindingregion.

The cytoplasmic tail of the CD8b subunit is palmitoylated (Arcaro, 2000), whichis thought to enable CD8 to interact with cell membrane rafts more readily, therebyenhancing interaction between CD8 and raft-associated receptor molecules, such asTCR/CD3 (Arcaro, 2001). However, other work has shown that the enhancedactivity of the b subunit over the a subunit of CD8 resides primarily within thestalk-like region of the ectodomain (Witte, 1999). Mutant CD8a subunits, comprisingthe a IgSF domain and the b stalk region are capable of restoring CD8b(−) CTLactivity in the presence of class I pMHC to nearly that of wildtype, as assessed byinterleukin production and cellular proliferation (Witte, 1999.) Furthermore, thepresence of the b stalk region enhances the co-receptor activity, and produces adistinct topology in the resulting CD8/MHC complex (Wong, 2003).


In all species studied, the CD8b stalk region is shorter than that of the a subunitby 10–13 amino acids. This differentiation in stalk length would potentially produceffffa bowed effect within the stalk region of theffff ab heterodimer, in contrast with thatof the symmetric CD8aa homodimer. Although the purpose of this structural bifurca-tion is unknown, it may well contribute to the enhanced effecter function of theffff bsubunit by orientating the ab heterodimer at the cell surface such that class I MHCinteraction is favoured.

5. CD8/MHC INTERACTION IS MODULATED BY O-LINKEDGLYCOSYLATION

The glycosylation state of T cell surface markers is associated with the T cellmaturation level and activation state (Piller, 1988; Wu, 1996). Specifically, immaturethymocytes and activated, mature T cells exhibit lower levels of cell-surface sialylationthan mature resting T cells. Several lines of evidence indicate developmental regula-tion of CD8/class I MHC interaction that is independent of TCR specificity (reviewedin Daniels, 2002; Baum, 2002; Gascoigne, 2002). Work from the laboratories of EllisReinhertz and Steve Jameson has shown that changes in O-glycosylation associatedwith T cell maturation have a direct effect upon the ligand binding properties offfffCD8 in vivo (Moody, 2001; Daniels, 2000; Daniels, 2001).Daniels et al. demonstrated differences in the affff ffinity of CD8 for soluble tetra-

mers of non-cognate pMHC by comparing CD8’s ability to interact withsoluble pMHC before and after T cell maturation. A marked reduction in class-IMHC/CD8 interaction is associated with the positive selection of CD8+ T cells fromCD4+/CD8+ double positive (DP) thymocytes in this assay. To determine the effectffffof sialylation of cell-surface glycoproteins upon MHC/CD8 interaction, thymocytesand mature T cells extracted from the lymph nodes of mice were treated withneuraminidase. Strikingly, a significant increase in MHC/CD8 interaction wasobserved following neuraminidase treatment of mature CD8+ T cells. Hence, Danielsand co-workers propose a mechanism whereby CD8 interaction with pMHC isdiminished following T cell maturation in a glycosylation-dependent manner; spe-cifically, changes in the surface sialylation of T cells that occur in thymocyte develop-ment have a dramatic impact upon the ability of CD8 to bind class I MHC(Daniels, 2001).Moody et al. have observed a similar TCR-independent variation in the affinity

of CD8 for pMHC tetramers in DP thymocytes as compared to resting, matureCD8+ T cells. Furthermore, Moody and et al. report that CD8ab heterodimersaccount for the overwhelming majority of CD8 at the thymocyte cell surface, anddemonstrate that glycosylation of the b subunit is critical to the variable activity ofCD8. Immunoprecipitation studies indicate that CD8b is subject to the most substan-tial modification in O-linked glycosylation during thymocyte maturation.Variation in T cell sialylation is brought about in part by developmentally-

regulated expression of the enzyme ST3 Gal-1, which catalyses addition of sialic acidto core 1 O-linked glycans (Priatel, 2000). Moody et al. demonstrated that ST3Gal-1 is largely responsible for the increase in sialylation of CD8 at the surface ofT cells following maturation (Moody, 2001).


Further work implicated addition of sialic acid moieties to O-linked carbohyd-rates at specific locations of the CD8b stalk region as responsible for the effects offfffreduced CD8ab/MHC interaction, resulting from an increase in ST3 Gal-1 expressionwhich accompanies the thymic development of CD8+ T cells from the CD4+/CD8+double-positive to CD8+ single-positive stage (Moody, 2003). The sialylation of O-linked carbohydrates attached to peptide fragments derived from tryptic digestionof the murine CD8b subunit was also analysed by nanospray ES-MS. Five of fourteenpotential O-glycosylation sites in the stalk region of CD8b were found to be post-translationally modified, of which three are conserved in all species so far studied.The modified threonines (T120, T121, T124, T127 and T128) are clustered in themembrane distal portion of the CD8b stalk, in close association to the IgSF domain.Surprisingly, core-1 O-glycans of the CD8b subunits of both thymocytes and matureCD8+ T cells were found to be mono-sialylated. Additional sialylation of a singleO-glycan at position T120-T124 was observed solely in CD8+ SP thymocytes andis, therefore, likely responsible for the enhanced CD8/MHC binding activity of thesecells over immature CD4+/CD8+ DP thymocytes.

6. THE INFLUENCE OF O-GLYCOSYLATION UPON THEEXTENSION OF THE CD8 STALK

The finding that a decrease in CD8/pMHC binding results from a development-ally mediated increase in CD8b O-glycan sialylation upon T cell maturation accountsfor the predominance of CD8ab over CD8aa in T cell selection and development.However, it is not immediately clear how changes to O-glycans in the stalk regionof CD8 can have such a dramatic effect upon the CD8ffff /MHC binding site, which issituated some 30 A away. It has also been proposed that the presence of sialic acidin the stalk region gives rise to an electrostatic repulsion effect between CD8 andffffclass I pMHC. However, thermodynamic considerations make this mechanismextremely unlikely.Moody et al. explain their results by suggesting that the addition of sialic acids

to O-glycans in the stalk region stabilizes or re-orientates the CD8 head group insuch a way that interaction with class I pMHC is no longer favoured. There isevidence that O-glycans can have an influence upon the structural characteristics ofpolypeptides to which they are attached (Gerken, 1989, Shogren, 1989, Rudd, 1999a)and it is possible that changes to the structure of O-glycans in the membrane distalregion of the CD8 stalk influence the extent of the extension and conformation ofthe stalk polypeptide.To investigate whether the structure of the CD8 stalk is sensitive to modification

by the addition of sialic-acid adducts to O-glycans in the membrane distal region, astructure-function analysis of CD8 carbohydrates was carried out in our laboratory.To this end, soluble forms of the CD8a and b subunit ectodomains were designed(as described elsewhere) and expressed in the Chinese hamster ovary (CHO) cell linesK1 and Lec3.2.8.1 (Merry, 2003). The CHO Lec3.2.8.1 cell line lacks the enzymeGlcNAc transferase T1, necessary for the processing of complex-form oligosacharidesin the golgi apparatus, and so produces only oligomannose N-linked glycans (seeChen, 2003). The profile of O-linked glycans of proteins produced in the Lec3.2.8.1


cell was thus expected to be significantly different to that of proteins expressed inffffthe K1 cell line.

The use of a glycosylation-deficient expression system was chosen as a preferredmeans of producing differentially-glycosylated protein over the alternatives of geneticffffmodification and chemical inhibition of glycosylation due to the potential problemsinherent in these systems, such as protein mis-folding (Butters, 1998). An unglycosyl-ated form of the CD8aa IgSF domain dimer lacking the stalk region, produced inthe E.coli expression system, was used as a control for these experiments. Given thesignificant functional activity of O-glycosylation specific to the b subunit of CD8 asopposed to that of the a subunit, it was reasonable to expect some differencesffffbetween the N-linked glycosylation profiles of the two subunits. However, analysisof N-glycans removed from separated CD8a and b constructs produced in CHO K1cells detected no significant differences between the two subunits, despite the veryfffflow level of protein sequence conservation.A comprehensive O-glycan analysis of CD8 proteins produced in both the CHO

K1 and CHO LecR systems was carried out by HPLC analysis. O-glycans in Thestalk region of CD8 produced in the CHO K1 system contained O-glycans of thetype 1 core Gal-b1,3 GalNAc structure, the majority of which were mono anddi-sialylated, although a small amount (15%) of unsialylated Gal-b1,3 GalNAc wasalso observed. In contrast, sialylation of O-glycans attached to CD8 produced in theLecR system was extremely limited, and was restricted to the infrequent addition ofa single sialic acid to the core-1 disaccharide.An analysis of O-glycan occupancy in CD8 constructs was carried out by

electrospray mass spectrometry and subsequent MS/MS fragmentation of trypticpeptide fragments of the rat CD8 stalk region. Three threonine residues (T126, T132and T134) within the stalk-like region of CD8 constructs produced in both K1 andLecR systems were occupied, although the addition of hexose residues to the GalNAccore was extremely limited in proteins derived in LecR cells (see Fig. 3).To examine the extent of polypeptide chain extension induced by O-glycans in

the stalk region of CD8 constructs produced in the CHO K1, CHO LecR and E.coliexpression systems, sedimentation (s) and Perrin (P) values were determined for eachconstruct by analytical ultracentrifugation (AUC). The (s) values determined for eachsample varied significantly, due to the different sizes of the variably glycosylatedffffforms of the protein. Critically, the experimentally-derived P value, which givesinformation regarding the shape of the molecule, was roughly equivalent for boththe CHO K1 and CHO LecR derived constructs. However, a substantial differenceffffwas observed between the E.coli derived form of soluble IgSF CD8 and thoseexpressed in the mammalian systems, which confirms that the system is indeedsensitive to both molecular mass and shape. Taken together, these data suggest thatthe presence of additional hexose residues and/or sialic acids attached to the N-acetylgalactosamine core of O-glycans in the stalk of CD8 have little or no overtaffect upon the length of the polypeptide stalk region (Fig. 4).ffff

7. CONCLUSIONS AND FUTURE PERSPECTIVES

The majority of cell surface glycoproteins are modified by terminal sialylation(Rudd, 1999a; 2001; 2004). The sialylation of T cell surface glycoproteins is directly


Figure 3. O-glycan analysis of CHO K1 and CHO Lec 3.2.8.1-derived CD8 constructs.

O-glycans released from the CHO K1 derived protein consisted of the type 1 core Gal-b1,3GalNAc

disaccharide, and its sialylated tri- and tetrasaccharide derivatives. Positive ion MALDI mass spectra of

the desialylated tryptic peptides from CHO Lec 3.2.8.1 derived protein indicated that ~70% of the O-glycans present consisted of only a single GalNAc residue, whilst the remaining 30% where extended

b1,3-galactose to form the type 1 core Gal-b1,3GalNAc. The O-glycosylation profile of CHO Lec-derived

protein was characterized by the absence of the di-sialylated structure as well as a substantial decrease in

the relative amounts of the mono-sialylated structure; 82% of the CHO Lec 3.2.8.1-derived type 1 core

Gal-b1,3GalNAc disaccharides were non-sialylated.

associated with the developmental state and activation state of the T cell (Piller,1988; Harrington, 2000; Starr, 2003; Pappu, 2004). The increase in T cell surfacesialylation that accompanies T cell maturation from the immature double-positivestage to mature CD8+ T cells has a negative effect upon the ability of CD8 toffffrecognize and interact with class I MHC (Moody, 2001; Daniels, 2001). This variationin CD8/MHC interaction is independent of MHC haplotype and TCR specificity.The stalk polypeptide of CD8 extends the IgSF head group of the co-receptor

across the intercellular junction to contact its binding site on the membrane proximalregion of the class I MHC. O-glycans present in the membrane distal region of thestalk polypeptide of both subunits likely act to rigidify this region of the polypeptideby reducing the overall extension of the stalk from the theoretical maximumof ~3.7 A/residue to ~2.6 A/residue as observed in mucins, as compared to~1.5 A/residue in the a-helical conformation. Data from our laboratory suggest thatthe addition of a single GalNAc moiety to threonine residues in the membrane distalstalk polypeptide is sufficient to induce this rigidification in the CD8 stalk (Merry,2003). The presence of additional carbohydrate residues attached to the terminalGalNAc moiety has no further effect upon the conformation of the CD8 stalkffffpolypeptide. However, the addition of sialic acids to these O-glycans provides a levelof post-translational control over the function of CD8+ T cells with regards to classI pMHC interaction (reviewed in Daniels, 2002). Given the apparent inability ofsialic acids in the CD8 stalk to directly influence the conformation of this polypeptide,the upregulation of sialylation during thymocyte development must have alternativeoutcomes for CD8 activity.


Figure 4. Molecular bead modeling of the soluble forms of CD8aa produced in (a) E.coli, (b) CHO

Lec3.2.8.1 and (c) CHO K1. In this simulation, the structure of the CD8aa head group is based on the

previously solved crystal structure of murine CD8aa in complex with classI p MHC (Kern, 1998). The

stalk region of mammalian CD8aa was modeled according to the amino acid sequence described in Merry,

2003, assuming an extension of 2.6 A per residue, as observed for mucins.N-linked andO-linked oligosacch-

arides have been modeled according to the glycan analysis detailed in Merry, 2003.

The S and P values given for each protein species are the calculated sedimentation (s) coefficients and

Perrin (P) functions applicable to each bead model, generated using a computational simulation program,

assuming a stalk region extension of 2.6 A/residue. Sexp and Pexp values are those determined experimentallyby AUC.

Critically, the calculated Pexp values, which give information regarding the shape of eth molecule, aresimilar for both CHO Lec and CHO K1 derived forms of the protein. These data suggest that, although

the presence or absence of the N-terminal stalk region per se is significant to the overall shape and

extended conformation of the molecule in solution, the presence of additional sialic acids in the CHO K1

derived protein has little influence upon the extension of the stalk region as compared with that of the

CHO Lec R derived protein.

It has been reported that clusters of O-linked carbohydrates attached to mucin-like polypeptides have the effect of restricting the conformational plasticity withinffffsuch polypeptides (Gerken, 1989; Shogren, 1989). It is tempting to speculate thatthe addition of sialic acids to O-glycans in the highly extended CD8 stalk regionmay function to further stiffen the polypeptide and, thereby, limit the flexibility offfffthe co-receptor head group. The effects of subtle changes to the rigidity of theffffpolypeptide stalk in the b subunit may be enhanced by the difference in lengthffffbetween the a and b subunits, such that the orientation of the IgSF dimer is restricted,resulting in decreased interaction with the binding site of pMHC. It is also possiblethat an increase in sialic acid content within the stalk region acts to reduce CD8co-localisation at the cell surface, which would account for the reduced pMHCtetramer binding observed by Moody et al and Daniels et al. Given the low affinitybinding between CD8 and pMHC, it is likely that these interactions are under the


influence of avidity effects, such as density-dependent binding, which are independentffffof any structural changes within the CD8 or pMHC. Further investigation of theeffects of increased CD8ffff b sialylation upon individual CD8ab/MHC complex forma-tion is required in order to establish the structural effects of sialylation upon thisffffinteraction.

In view of the functional significance of the stalk region of the CD8b subunit,it is interesting to speculate on the function of the equivalent stalk region of theCD8a subunit, which is similarly rich in threonine, serine and proline residues. Todate, however, there have been no reports of an analogous glycosylation dependentfunctional variation in the homodimeric CD8aa isoform of the co-receptor. Likewise,there has been no evidence to suggest that the CD4 co-receptor is subject toglycosylation-dependent classII MHC binding, suggesting that activation and mat-uration state-dependent control is specific to the CD8ab/pMHC interaction. TheCD4 and CD8 co-receptors are structurally distinct, despite their apparently similarbiological functions as a co-receptor for TCR activation (see Gao, 2002). The CD4co-receptor has evolved to address the problem of spanning the inter-cellular junctionby association of four IgSF domains, whereas the CD8 receptor utilises an extendedlinker domain to serve this purpose. The reasons underlying the structural differencesffffbetween CD4 and CD8 have yet to be determined. It is possible that the presenceof O-glycans in the CD8 stalk amenable to post-translational modification may beconstitutive to CD8+ T cell function and, therefore, fundamental to the structuraldifferentiation of the CD4 and CD8 co-receptors.ffff

Although it is widely accepted that sialylation of O-linked carbohydrates at theT cell surface has a profound effect upon T cell activities, the significance of variationffffin the sialylation of individual glycoproteins in vivo is difficult to determine. Recentwork examined the effects of the global de-sialylation of T cell surface markers onffffthe ability of mature T cells to react to peptide antigen (Starr, 2004). T cells exhibita specific loss in sensitivity to low affinity ligands following maturity, while sensitivityto higher affinity ligands is maintained. Starr and colleagues report that the level ofthe response by mature T cells to low-affinity peptide antigen is restored to the levelof immature thymocytes in the presence of neuraminidase, suggesting a role for sialicacids in the fine tuning of the TCR activation threshold. Although these results maybe explained in part by an enhanced CD8 activity following desialylation, it isapparent that the variable sialylation of other T cell surface glycoproteins is equallycritical to T cell function in response to peptide antigen.

ACKNOWLEDGEMENTS

IAW is supported by the National Institutes of Health Grants AI42266 andCA58896.

REFERENCES

Arcaro, A., Gregoire, C., Bakker, T. R., Baldi, L., Jordan, M., Goffin, L., Boucheron, N., Wurm, F., van der

Merwe, P. A., Malissen, B., and Luescher, I. F. (2001). CD8b endows CD8 with efficient coreceptor


function by coupling T cell receptor/CD3 to raft-associated CD8/p56( lck) complexes. J Exp Med

194, 1485–95.

Arcaro, A., Gregoire, C., Boucheron, N., Stotz, S., Palmer, E., Malissen, B., and Luescher, I. F. (2000).

Essential role of CD8 palmitoylation in CD8 coreceptor function. J Immunol 165, 2068–76.

Baum, L. G. (2002). Developing a taste for sweets. Immunity 16, 5–8.

Bosselut, R., Kubo, S., Guinter, T., Kopacz, J. L., Altman, J. D., Feigenbaum, L., and Singer, A. (2000).

Role of CD8b domains in CD8 coreceptor function: importance for MHC I binding, signaling, and

positive selection of CD8+ T cells in the thymus. Immunity 12, 409–18.Butters, T. D., Sparks, L. M., Harlos, K., Ikemizu, S., Stuart, D. I., Jones, E. Y., and Davis, S. J. (1999).

Effects of N-butyldeoxynojirimycin and the Lec3.2.8.1 mutant phenotype on N-glycan processing inffff

Chinese hamster ovary cells: application to glycoprotein crystallization. Protein Sci 8, 1696–701.

Cawthon, A. G., Lu, H., and Alexander-Miller, M. A. (2001). Peptide requirement for CTL activation

reflects the sensitivity to CD3 engagement: correlation with CD8ab versus CD8aa expression. J

Immunol 167, 2577–84.

Chen, W., and Stanley, P. (2003). Five Lec1 CHO cell mutants have distinct Mgat1 gene mutations that

encode truncated N-acetylglucosaminyltransferase I. Glycobiology 13, 43–50.

Cheroutre, H., Holcombe, H. R., Tangri, S., Castano, A. R., Teitell, M., Miller, J. E., Cardell, S., Benoist, C.,

Mathis, D., and Huse, W. D. (1995). Antigen-presenting function of the TL antigen and mouse CD1

molecules. Immunol Rev 147, 31–52.

Classon, B. J., Brown, M. H., Garnett, D., Somoza, C., Barclay, A. N., Willis, A. C., and Williams, A. F.

(1992). The hinge region of the CD8a chain: structure, antigenicity, and utility in expression of

immunoglobulin superfamily domains. Int Immunol 4, 215–25.

Crooks, M. E., and Littman, D. R. (1994). Disruption of T lymphocyte positive and negative selection in

mice lacking the CD8 beta chain. Immunity 1, 277–85.

Daniels, M. A., Devine, L., Miller, J. D., Moser, J. M., Lukacher, A. E., Altman, J. D., Kavathas, P.,

Hogquist, K. A., and Jameson, S. C. (2001). CD8 binding to MHC class I molecules is influenced by

T cell maturation and glycosylation. Immunity 15, 1051–61.

Daniels, M. A., Hogquist, K. A., and Jameson, S. C. (2002). Sweet ‘n’ sour: the impact of differentialffff

glycosylation on T cell responses. Nat Immunol 3, 903–10.

Daniels, M. A., and Jameson, S. C. (2000). Critical role for CD8 in T cell receptor binding and activation

by peptide/major histocompatibility complex multimers. J Exp Med 191, 335–46.

Devine, L., and Kavathas, P. B. (1999a). Molecular analysis of protein interactions mediating the function

of the cell surface protein CD8. Immunol Res 19, 201–10.

Devine, L., Kieffer, L. J., Aitken, V., and Kavathas, P. B. (2000). Human CD8ffff b, but not mouse CD8 beta,

can be expressed in the absence of CD8 alpha as a beta beta homodimer. J Immunol 164, 833–8.

Devine, L., Rogozinski, L., Naidenko, O. V., Cheroutre, H., and Kavathas, P. B. (2002). The complementar-

ity-determining region-like loops of CD8a interact differently with beta 2-microglobulin of the classffff

I molecules H-2Kb and thymic leukemia antigen, while similarly with their alpha 3 domains. JImmunol 168, 3881–6.

Devine, L., Sun, J., Barr, M. R., and Kavathas, P. B. (1999a). Orientation of the Ig domains of CD8a beta

relative to MHC class I. J Immunol 162, 846–51.

Doucey, M. A., Goffin, L., Naeher, D., Michielin, O., Baumgartner, P., Guillaume, P., Palmer, E., and

Luescher, I. F. (2003). CD3 delta establishes a functional link between the T cell receptor and CD8.

J Biol Chem 278, 3257–64.

Fung-Leung, W. P., Kundig, T. M., Ngo, K., Panakos, J., De Sousa-Hitzler, J., Wang, E., Ohashi, P. S.,

Mak, T. W., and Lau, C. Y. (1994). Reduced thymic maturation but normal effector function offfff

CD8+ T cells in CD8b gene-targeted mice. J Exp Med 180, 959–67.Fung-Leung, W. P., Louie, M. C., Limmer, A., Ohashi, P. S., Ngo, K., Chen, L., Kawai, K., Lacy, E., Loh,

D. Y., and Mak, T. W. (1993). The lack of CD8a cytoplasmic domain resulted in a dramatic decrease

in efficiency in thymic maturation but only a moderate reduction in cytotoxic function of CD8+ Tlymphocytes. Eur J Immunol 23, 2834–40.

Gangadharan, D., and Cheroutre, H. (2004). The CD8 isoform CD8aa is not a functional homologue of

the TCR co-receptor CD8ab. Curr Opin Immunol 16, 264–70.Gao, G. F., and Jakobsen, B. K. (2000). Molecular interactions of coreceptor CD8 and MHC class I: the

molecular basis for functional coordination with the T-cell receptor. Immunol T oday 21, 630–6.Gao, G. F., Rao, Z., and Bell, J. I. (2002). Molecular coordination of ab T-cell receptors and coreceptors

CD8 and CD4 in their recognition of peptide-MHC ligands. T rends Immunol 23, 408–13.


Gao, G. F., Tormo, J., Gerth, U. C., Wyer, J. R., McMichael, A. J., Stuart, D. I., Bell, J. I., Jones, E. Y., and

Jakobsen, B. K. (1997). Crystal structure of the complex between human CD8a(a) and HLA-A2.

Nature 387, 630–4.

Garcia, K. C., Scott, C. A., Brunmark, A., Carbone, F. R., Peterson, P. A., Wilson, I. A., and Teyton, L.

(1996). CD8 enhances formation of stable T-cell receptor/MHC class I molecule complexes. Nature

384, 577–81.

Garcia, K. C., Teyton, L., and Wilson, I. A. (1999). Structural basis of T cell recognition. Annu Rev Immunol

17, 369–97.

Gascoigne, N. R. (2002). T-cell differentiation: MHC class I’s sweet tooth lost on maturity.ffff Curr Biol

12, R99-R101.

Gerken, T. A., Butenhof, K. J., and Shogren, R. (1989). Effects of glycosylation on the conformation andffff

dynamics of O-linked glycoproteins: carbon-13 NMR studies of ovine submaxillary mucin.


Harrington, L. E., Galvan, M., Baum, L. G., Altman, J. D., and Ahmed, R. (2000). Differentiating betweenffff

memory and effector CD8 T cells by altered expression of cell surfaceffff O-glycans. J Exp Med 191,

1241–6.

Irie, H. Y., Mong, M. S., Itano, A., Crooks, M. E., Littman, D. R., Burakoff, S.ffff J., and Robey, E. (1998). The

cytoplasmic domain of CD8 beta regulates Lck kinase activation and CD8 T cell development. J

Immunol 161, 183–91.

Kern, P., Hussey, R. E., Spoerl, R., Reinherz, E. L., and Chang, H. C. (1999). Expression, purification, and

functional analysis of murine ectodomain fragments of CD8aa and CD8ab dimers. J Biol Chem

274, 27237–43.

Kern, P. S., Teng, M. K., Smolyar, A., Liu, J. H., Liu, J., Hussey, R. E., Spoerl, R., Chang, H. C., Reinherz,

E. L., and Wang, J. H. (1998). Structural basis of CD8 coreceptor function revealed by crystallo-

graphic analysis of a murine CD8aa ectodomain fragment in complex with H-2Kb. Immunity 9,519–30.

Kim, P. W., Sun, Z. Y., Blacklow, S. C., Wagner, G., and Eck, M. J. (2003). A zinc clasp structure tethers

Lck to T cell coreceptors CD4 and CD8. Science 301, 1725–8.

Leishman, A. J., Naidenko, O. V., Attinger, A., Koning, F., Lena, C. J., Xiong, Y., Chang, H. C., Reinherz,

E., Kronenberg, M., and Cheroutre, H. (2001). T cell responses modulated through interaction

between CD8aa and the nonclassical MHC class I molecule, TL. Science 294, 1936–9.

Li, S., Choksi, S., Shan, S., Hu, X., Gao, J., Korngold, R., and Huang, Z. (1998). Identification of the CD8

DE loop as a surface functional epitope. Implications for major histocompatibility complex class I

binding and CD8 inhibitor design. J Biol Chem 273, 16442–5.

Liu, Y., Xiong, Y., Naidenko, O. V., Liu, J. H., Zhang, R., Joachimiak, A., Kronenberg, M., Cheroutre, H.,

Reinherz, E. L., and Wang, J. H. (2003). The crystal structure of a TL/CD8aa complex at 2.1 A

resolution: implications for modulation of T cell activation and memory. Immunity 18, 205-15.

Luescher, I. F., Vivier, E., Layer, A., Mahiou, J., Godeau, F., Malissen, B., and Romero, P. (1995). CD8

modulation of T-cell antigen receptor-ligand interactions on living cytotoxic T lymphocytes. Nature

373, 353–6.

Merry, A. H., Gilbert, R. J., Shore, D. A., Royle, L., Miroshnychenko, O., Vuong, M., Wormald, M. R.,

Harvey, D. J., Dwek, R. A., Classon, B. J., Rudd, P. M., and Davis, S. J. (2003). O-glycan sialylation

and the structure of the stalk-like region of the T cell co-receptor CD8. J Biol Chem 278, 27119–28.

Moody, A. M., Chui, D., Reche, P. A., Priatel, J. J., Marth, J. D., and Reinherz, E. L. (2001).

Developmentally regulated glycosylation of the CD8ab coreceptor stalk modulates ligand binding.

Cell 107, 501–12.

Moody, A. M., North, S. J., Reinhold, B., Van Dyken, S. J., Rogers, M. E., Panico, M., Dell, A., Morris,

H. R., Marth, J. D., and Reinherz, E. L. (2003). Sialic acid capping of CD8b core 1-O-glycans controls

thymocyte-major histocompatibility complex class I interaction. J Biol Chem 278, 7240–6.Norment, A. M., and Littman, D. R. (1988). A second subunit of CD8 is expressed in human T cells. Embo

J 7, 3433–9.Palacios, E. H., and Weiss, A. (2004). Function of the Src-family kinases, Lck and Fyn, in T-cell develop-

ment and activation. Oncogene 23, 7990–8000.Pappu, B. P., and Shrikant, P. A. (2004). Alteration of cell surface sialylation regulates antigen-induced

naive CD8+ T cell responses. J Immunol 173, 275–84.Piller, F., Piller, V., Fox, R. I., and Fukuda, M. (1988). Human T-lymphocyte activation is associated with

changes in O-glycan biosynthesis. J Biol Chem 263, 15146–50.


Priatel, J. J., Chui, D., Hiraoka, N., Simmons, C. J., Richardson, K. B., Page, D. M., Fukuda, M., Varki, N.

M., and Marth, J. D. (2000). The ST3Gal-I sialyltransferase controls CD8+ T lymphocyte homeost-asis by modulating O-glycan biosynthesis. Immunity 12, 273–83.

Purbhoo, M. A., Boulter, J. M., Price, D. A., Vuidepot, A. L., Hourigan, C. S., Dunbar, P. R., Olson, K.,

Dawson, S. J., Phillips, R. E., Jakobsen, B. K., Bell, J. I., and Sewell, A. K. (2001). The human CD8

coreceptor effects cytotoxic T cell activation and antigen sensitivity primarily by mediating completeffff

phosphorylation of the T cell receptor zeta chain. J Biol Chem 276, 32786–92.

Rudd, P. M., Elliott, T., Cresswell, P., Wilson, I. A., and Dwek, R. A. (2001). Glycosylation and the immune

system. Science 291, 2370–6.

Rudd, P. M., Wormald, M. R., and Dwek, R. A. (2004). Sugar-mediated ligand-receptor interactions in the

immune system. T rends Biotechnol 22, 524–30.

Rudd, P. M., Wormald, M. R., Harvey, D. J., Devasahayam, M., McAlister, M. S., Brown, M. H., Davis,

S. J., Barclay, A. N., and Dwek, R. A. (1999). Oligosaccharide analysis and molecular modeling of

soluble forms of glycoproteins belonging to the Ly-6, scavenger receptor, and immunoglobulin

superfamilies expressed in Chinese hamster ovary cells. Glycobiology 9, 443–58.

Rudd, P. M., Wormald, M. R., Stanfield, R. L., Huang, M., Mattsson, N., Speir, J. A., DiGennaro, J. A.,

Fetrow, J. S., Dwek, R. A., and Wilson, I. A. (1999). Roles for glycosylation of cell surface receptors

involved in cellular immune recognition. J Mol BiolMM 293, 351–66.

Shogren, R., Gerken, T. A., and Jentoft, N. (1989). Role of glycosylation on the conformation and chain

dimensions of O-linked glycoproteins: light-scattering studies of ovine submaxillary mucin.


Starr, T. K., Daniels, M. A., Lucido, M. M., Jameson, S. C., and Hogquist, K. A. (2003). Thymocyte

sensitivity and supramolecular activation cluster formation are developmentally regulated: a partial

role for sialylation. J Immunol 171, 4512–20.

Sun, J., and Kavathas, P. B. (1997). Comparison of the roles of CD8 aa and CD8 ab in interaction with

MHC class I. J Immunol 159, 6077–82.

van der Merwe, P. A., and Davis, S. J. (2003). Molecular interactions mediating T cell antigen recognition.

Annu Rev Immunol 21, 659–84.

Wang, J., and Reinherz, E. L. (2000). Structural basis of cell-cell interactions in the immune system. Curr

Opin Struct Biol 10, 656–61.

Witte, T., Spoerl, R., and Chang, H. C. (1999). The CD8b ectodomain contributes to the augmented

coreceptor function of CD8ab heterodimers relative to CD8aa homodimers. Cell Immunol 191, 90–6.

Wong, J. S., Wang, X., Witte, T., Nie, L., Carvou, N., Kern, P., and Chang, H. C. (2003). Stalk region of

b-chain enhances the coreceptor function of CD8. J Immunol 171, 867–74.

Wu, W., Harley, P. H., Punt, J. A., Sharrow, S. O., and Kearse, K. P. (1996). Identification of CD8 as a

peanut agglutinin (PNA) receptor molecule on immature thymocytes. J Exp Med 184, 759–64.

Zamoyska, R. (1994). The CD8 coreceptor revisited: one chain good, two chains better. Immunity 1, 243–6.

13

IMMUNOGENICITY OF CALRETICULIN-BOUNDMURINE LEUKEMIA VIRUS GLYCOPROTEIN gp90

Yusuke Mimura1,2, Denise Golgher2, Yuka Mimura-KimuraYY 1,Raymond A. Dwek1, Pauline M. Rudd1, and Tim Elliott2

1Glycobiology InstituteDepartment of BiochemistryUniversity of OxfordSouth Parks Road, Oxford, OX1 3QU, UK2Cancer Sciences DivisionSchool of MedicineUniversity of SouthamptonTremona Road, Southampton, SO16 6YD, UK

1. INTRODUCTION

Class II molecules of the major histocompatibility complex (MHC class IImolecules) bind peptides derived from protein antigens delivered into endocyticcompartments and present these peptides to CD4+ T cells (Cresswell, 1994; Germain,1994). In the endocytic pathway antigens are unfolded and cleaved into fragmentsduring transport through the increasingly acidic endosomal network, from earlyendosomes to lysosomes. Newly synthesized MHC class II molecules associated withthe invariant chains (Ii) are transported to a late endocytic compartment where theIi is removed by proteolytic cleavage. MHC class II molecules bind polypeptideantigens by removal of the Ii-derived peptide CLIP (class II invariant chain-derivedpeptide) through a peptide-exchange process that is catalyzed by MHC-encodedDM molecules (Fig. 1) (Denzin and Cresswell, 1995; Sherman et al., 1995; Sloanet al., 1995; Wolf and Ploegh, 1995). The initial form of antigen that binds to classII molecules may be short peptides 10-20 amino acids in length, generated by acidendopeptidases (Watts, 2001). This pathway could be termed ‘‘cut/trim first, bindlater’’ model and became the paradigm for the binding of peptides to MHC class IImolecules. However, there is an alternative pathway independent of DM moleculesthat involves early capture of unfolded, extended sequence byMHC class II molecules(Pinet et al., 1994; Pinet et al., 1995; Sercarz et al., 1993). Subsequent trimming of

85



86 Y. Mimura et al.

Figure 1. Overview of the MHC class II processing and presentation pathway.

Figure 2. Two postulated pathways for antigen presentation by MHC class II molecules.

the MHC-bound peptide would give rise to a short peptide epitope that roughlycorresponded to the ‘‘footprint’’ of the MHC binding site. The ability of MHC classII molecules to bind extended polypeptide sequences is facilitated by the fact thatits peptide binding groove is open at each end, unlike the MHC class I bindinggroove which terminates with deep pockets at either end. This could be termed ‘‘bindfirst, cut/trim later’’ model (Fig. 2) (Sercarz and Maverakis, 2003). Although thislatter pathway is an attractive proposition with respect to epitope selection in anaggressive proteolytic environment, whether this model is the canonical mode ofoperation remains unclear (Watts, 2004) because this model has been investigatedmostly using well-defined hen egg lysozyme (HEL) system whose immunodominantdeterminant HEL(52-61) can be presented by the DM-dependent classical pathway(Castellino et al., 1998; Lindner and Unanue, 1996; Pinet et al., 1995). This chapterwill be concerned with presentation by MHC class II molecules of unique tumor

Immunogenicity of Leukemia Virus Glycoprotein gp90 87

antigen gp90 (Huang et al., 1996) whose antigenicity is dependent on the conforma-tion and calreticulin (CRT) binding (Golgher et al., 2001). We will review evidencefor the ‘‘bind first’’ model and describe how this notion is related to the conforma-tional dependence of gp90 antigenicity. As CD4+ T cells play an important role inanti-tumor immune response (Pardoll and Topalian, 1998), our findings may forma basis for the development of cancer vaccine design by manipulating the conforma-tion of tumor antigens.An exogenous antigen is taken up by an antigen presenting cell (APC) throughendocytosis, unfolded and cleaved during transport from early endosomes to lyso-somes by acid endoproteases including cathepsin D and cathepsin S (Cat). NascentMHC class II molecules are associated with Ii trimers in the endoplasmic reticulum(ER). This complex with CLIP in the class II binding site is competent for exit fromthe ER and moves through the Golgi apparatus to the trans-Golgi network (TGN).Most of MHC class II - Ii complexes traffic directly to later endocytic locations suchas MHC class II loading compartments (MIIC) and lysosomes from the TGN(Pathway (1)), others traffic to lysosomes after internalization from the plasmamembrane (Pathway (2)) and from the TGN to early endosomes before their trans-port to lysosomes (Pathway (3)), with the relative contribution of each pathwayvarying from APC to APC (Hiltbold and Roche, 2002). The exchange of CLIP formore stable antigenic peptides is catalyzed by the DMmolecules. The bound peptidesundergo proteolytic trimming to form a final size of 15 to 20 residues. The peptide-loaded MHC class II molecules are transported to the cell surface to be recognizedby CD4+ T cells. Pre-existing surface peptide - MHC class II complexes can alsointernalize and recycle through endosomes, where peptides can be exchanged in aDM-independent manner (Pathway (4)).

2. INFLUENCE OF ANTIGEN CONFORMATION ONPRESENTATION BY MHC CLASS II MOLECULES

Since the original studies of T cell recognition of protein antigen where recogni-tion of denatured antigen was shown to be as efficient as native antigen (Chesnutet al., 1980), it has long been assumed that antigen conformation is not an importantfactor in determining antigenicity because T cells generally recognize protein antigensthat have been processed and subsequently bound to MHC class II molecules. Theprocessing requirement could reflect the incapacity of native proteins to interactwith MHC molecules and form MHC-antigenic peptide complexes. However, manystudies have shown that MHC class II molecules can engage peptides of muchgreater length than those typically eluted from MHC class II molecules and thatthis can occur during normal processing. Sette et al. (1989) detected positive bindingof urea-denatured and reduced versions of ovalbumin, bovine serum albumin, HEL,and transferrin to appropriate MHC class II molecules but not for their native forms,indicating that an unfolding step by exposure to low pH or proteolysis in endocyticcompartments is a prerequisite for MHC class II binding to antigens. Although suchtightly folded globular molecules do not bind to MHC class II molecules, a nativeprotein molecule as large as fibrinogen (340 kDa) has been shown to bind to pre-fixed APCs in the absence of processing. The epitope recognized was localized to a

88 Y. Mimura et al.

region of the molecule with conformational flexibility that gained access to MHCclass II peptide-binding grooves (Lee et al., 1988). Furthermore, some intact proteinssuch as HEL, beef insulin and pigeon cytochrome c have been shown to be presentedto specific T cells by MHC class II molecules bound at low pH in the presence orabsence of reducing agent without a requirement of proteolytic cleavage or irrevers-ible denaturation (Jensen, 1993). These results indicate the ability of MHC class IImolecules to bind and present intact or even native proteins to CD4+ T cells,depending on determinant accessibility. As a noticeable consequence of class IIcapture of unfolding antigens, the immunogenic T cell determinants could becomeproteolytically inaccessible. This protection hypothesis has been tested using a 34amino acid HEL peptide containing the high-affinity I-Ak binding site HEL(52-61)in the core with extension of 12 unnatural D-amino acids at both sides so that thispeptide would be resistant to proteolysis except in the core region (Donermeyer andAllen, 1989). The authors showed that the determinant became resistant to chymo-trypsin digestion if HEL(40-73) was allowed to first bind to I-Ak, otherwise it wastotally destroyed. The protection hypothesis suggests that MHC class II binding toa large antigen fragment can increase the possibility to present T cell determinants.

3. ANTIGEN PRESENTATION BY TWO DISTINCT POPULATIONSOF MHC CLASS II MOLECULES

In addition to the classical DM-dependent pathway, recycling of cell-surfaceMHC class II molecules has been suggested as a pathway for presentation of someantigens including RNase A, influenza virus hemagglutinin and myelin basic protein(Adorini et al., 1989; Harding and Unanue, 1989; Nadimi et al., 1991; Pinet et al.,1994; Pinet et al., 1995). Cell surface MHC class II molecules have been shown tointernalize and return to the cell surface rapidly. Pinet et al. (1995) have reportedthat truncation of either one of the _ or _ cytoplasmic tails of HLA-DR moleculeseliminates internalization of HLA-DR and presentation of influenza hemagglutininwhile Ii-dependent presentation of matrix antigen from the same virus particles isunaffected by the truncations. Thus, the compartment for peptide loading on recyclingffffclass II molecules seemed likely to be distinct from the one where newly synthesizedclass II molecules are transported. This alternative processing pathway was alsodescribed with the finding that intact partially-folded HEL forms stable complexeswith mature I-Ak molecules in low pH compartments independently of proteasesand DM (Lindner and Unanue, 1996). By subcellular fractionation it has beenshown that early endosomes generate RNase(42-56)-I-Ak complexes by theDM-independent processing pathway (Griffin et al., 1997). Zhong et al. (1997)addressed the involvement of newly synthesized (Ii-associated) versus mature classII (Ii-free) in effective presentation of distinct determinants in HEL by truncationsffffof leucine-based cytoplasmic tails in Ii and MHC class II molecules, demonstratingthat a requirement for MHC class II internalization is inversely correlated with arequirement for Ii expression. These results suggest that the classical pathway requiresIi and DM expression for peptide loading onto newly synthesized MHC class IImolecules in late endocytic compartments whereas the alternative pathway utilizesrecycling, mature MHC class II molecules independently of protein synthesis, Ii and


DM for peptide loading in earlier compartments than the classical pathway. Inaddition to recycling mature MHC class II molecules, Ii-free MHC class II moleculeshave been shown to be generated by removal of Ii in an early endosomal compart-ment without involvement of cathepsin S or DM, allowing the direct binding to high

molecular-weight polypeptides (Villadangos et al., 2000). The existence of multipleendocytic compartments for peptide loading would be advantageous since eachcompartment would provide a unique environment to accommodate the processingof a wide variety of class II-restricted epitopes. Some epitopes may be revealed in

early endosomes and destroyed before reaching later compartments, whereas othersmay require the more active proteolytic environment of late endosomes and lyso-somes for release. The simultaneous presence of these mechanisms increases the

possiblilties of displaying foreign epitopes that will be recognized by CD4+ T cells.

4. MHC-GUIDED PROCESSING AND IMMUNODOMINANCE

‘‘MHC-guided processing’’ was originally hypothesized by Sercarz et al. (1993)and a consequence of ‘‘bind first, trim later’’ model. The notion is closely related to

immunodominance and stresses the importance of intramolecular competition

between the multiple determinants on a single long peptide for binding to an MHCmolecule (Deng et al., 1993). Protein antigens typically contain multiple epitopescapable of binding to MHC class II molecules, yet T cell responses are limited toonly a small number of these determinants. The ability of the immune system to

regulate and focus T cell responses to a select number of epitopes is termed immuno-dominance. The hierarchy of T cell responses observed in vivo, namely dominant,subdominant, and cryptic (silent), reflects in part this selective presentation of epi-topes by MHC class II molecules, as well as the influences of T cell responsivenessand repertoire. Biochemical and functional studies of MHC class II molecules haverevealed the preferential display of immunodominant epitopes, with conversely lowerlevels of MHC-restricted presentation of subdominant or cryptic peptides derivedfrom the same Ag (Ma et al., 1999; Nelson et al., 1992; Viner et al., 1995). Thespecific reactions within APC which influence epitope selection remain poorlydefined, with both Ag processing and MHC binding potentially playing key roles.Early capture of unfolded, extended sequence by an MHC class II molecule outlinedby the notion of MHC-guided processing is an attractive proposition for the epitopeselection in an aggressive proteolytic environment. Castellino et al. (1998) showed,using the HEL model, that large complexes of 120 kDa could be found in theendocytic pathway that comprised a single HEL polypeptide chain of about 70

amino acids, bound to two different MHC class II isotypes, I-Affff k and I-Ek. Thisclearly shows that binding of MHC class II molecules can occur before excessive

antigen processing. HIV envelope glycoprotein (gp140) epitopes recognized by HIV-specific CD4+ T cells have been found to be located in exposed, non-helical loopsor strands on one face of the molecule (Surman et al., 2001). Presumably, MHC-guided processing may play a role in the selection of immunodominant gp140epitopes.

90 Y. Mimura et al.

5. INCOMPLETE FOLDING IN THE ER CAN ALTERANTIGENICITY OF A GLYCOPROTEIN

Golgher et al. (2001) have reported an interesting MHC class II-restricted tumorantigen, endogenous murine leukemia virus envelope protein gp90 (Fass et al., 1997;Lenz et al., 1982; Pinter and Fleissner, 1977), that is expressed in many mouse tumorlines while antigenicity varies among the cell lines. gp90 is synthesized in the endo-plasmic reticulum (ER) as a precursor protein, where its amino-terminal signalsequence is cleaved. The polypeptide then undergoes disulfide bonding and is modi-fied by the addition and processing of multiple asparagine-linked high mannose-typesugars, and oligomerizes, prior to transport to the Golgi apparatus. In the Golgiapparatus, the high mannose-type sugars are modified to complex-type, and theprecursor protein polypeptide backbone is cleaved to form an extracellular glycopro-tein gp70 and a transmembrane protein p15E. p15E anchors gp70 to the plasmamembrane by way of noncovalent interactions and in some instances, a disulfidebond (Gliniak et al., 1991; Pinter et al., 1978). CD4+ T cell hybridoma clonesspecific for gp90 were obtained from mice immunized with colon adenocarcinomacell line CT26 genetically engineered to secrete granulocyte/macrophage colony-stimulating factor. Importantly, the recognition of gp90 by the CD4+ T cellhybridomas is strictly dependent on the conformation of gp90 (Golgher et al., 2001).Unfolding of the protein by disulfide bond reduction and alkylation markedly reducesantigenicity, and thermal denaturation abrogates T cell recognition, which is differentfffffrom presentation of other model antigens such as HEL and ovalbumin describedabove. Furthermore, the T cell hybridomas do not respond to its mature form, gp70.Interestingly, the gp90 from other tumor lines of the same histological origin is notantigenic although the nucleotide sequence of their gp90 cDNAs was identical(unpublished data). In CT26 cells gp90 is found to be retained in the ER, associatedmostly with the ER chaperone calreticulin (CRT) that recognizes monoglucosylatedhigh mannose-type oligosaccharides (Glc1Man5-9GlcNAc2 ) on nascent glycopro-teins (Helenius and Aebi, 2001; Parodi, 2000; Sitia and Braakman, 2003; Trombetta,2003). It was noted that the CRT-bound gp90 is highly antigenic (Golgher et al.,2001). Incomplete folding of antigenic gp90 is evidenced by the presence of the CRT-ligand, Glc1Man9GlcNAc2 , in the antigenic gp90 of CT26 (YM, manuscript inpreparation).

Why is gp90 presentation dependent on the conformation? Generally, denatura-tion and unfolding have a minimum effect on antigen presentation by MHC classffffII molecules (Allen and Unanue, 1984; Streicher et al., 1984) because an intactprotein antigen eventually is always recognized as a denatured protein or fragment,as a result of intracellular processing by an APC. Therefore, it is very likely that thegp90 epitope is susceptible to proteolysis, i.e., destructive processing. Trypsin diges-tion abrogated gp90 presentation by pre-fixed APCs, in contrast to ovalbumin whereprior tryptic digestion liberates the immunodominant epitope. However, if nativegp90 is incubated with pre-fixed APCs, the epitope became resistant to trypsin andis successfully presented to the T cell hybridoma (YM, manuscript in preparation).Therefore, the gp90 epitope needs to be bound and protected by MHC class IImolecules before processing. Furthermore, gp90 presentation was not affected byffffleupeptin or chloroquine treatment of APCs, contrary to ovalbumin (YM, manuscript


in preparation), suggesting that gp90 binds to recycling mature MHC class IImolecules in a DM-independent manner. Taken together, gp90 presentation is con-sistent with the alternative pathway involving MHC-guided processing.It should be noted that differential processing by APCs of the same glycoproteinffff

can occur, depending on the folding status of the glycoprotein and cell lines thatsynthesize it. This phenomenon cannot be under-evaluated because MHC-guidedprocessing can convert a cryptic epitope into an immunodominant one by protectingthe determinant from destructive processing. Such a situation may permit inductionof an anti-tumor immune response as well as an autoimmune reaction by thepresentation of ‘‘self ’’ epitopes to which tolerance has not been established. Themechanism by which incomplete folding of a glycoprotein is promoted in CT26 cellsremains unclear although oxidoreductases (e.g., ERp72, ERp57, PDI, Ero-1), ERmolecular chaperones (e.g., Bip, CRT, CNX) or differential N-glycosylation siteffffoccupancy may be involved. If conformational changes of a glycoprotein antigencould be induced by modulating the activities of oxidoreductases in the ER, it wouldallow for manipulation of hierarchy of T cell response to the antigen, which mayform a basis for the improvement of cancer vaccine design.

6. THE ROLE OF GLYCOSYLATION IN INFLUENCING MHCCLASS II-RESTRICTED ANTIGEN PROCESSING

The extent to which MHC class II epitopes are generated via a bind/trim versusa trim/bind pathway will depend on a number of factors including competitionbetween MHC class II and destructive proteases for a relatively short stretch ofpolypeptide within a larger precursor protein. Accessibility of this site to the twomutually exclusive activities may in turn depend upon the route of antigen uptakeand the timing of its subsequent encounters with processing factors such as oxidored-uctases and hydrolases. The presence of post-translational modifications carried byprotein antigens could clearly affect any of these events. Although formal proof hasffffyet to be provided for the direct influence of glycosylation on the outcome of antigenprocessing, compelling indirect evidence exists linking N-linked glycosylation to theantigenicity of glycoproteins for T cell responses directed towards non-glycopeptideepitopes. Thus Sjolander et al. (1996) demonstrated that HIV-1 gp160 lacking threeN-linked glycans in its C-terminal CD4-binding region were unable to induce inmice T cell responses to non-glycosylated epitopes in this region of the molecule;whereas the fully glycosylated gp160 could. One explanation for this observation isthat the physical location of the peptide epitope within the native protein leads todifferential processing and consequent epitope selection and furthermore that prox-ffffimal glycosylation may influence this process – perhaps by protecting regions of themolecule from destructive processing. Consistent with this idea is the observationthat extensive T cell epitope mapping of the gp160 glycoprotein has shown thatepitope clustering occurs in four hotspots that comprise relatively short sequencesthat were bordered by regions of heavy glycosylation on exposed strands (Surmanet al., 2001). In other studies, the generation of MHC class II-restricted epitopesfrom tumor antigens (tyrosinase and gp90 described above) has been shown to bedependent on these antigens being glycosylated (Golgher et al., 2001; Housseau et al.,

92 Y. Mimura et al.

2001). Both these studies provide indirect evidence that the T cell epitope is notglycosylated. In the former case, site-directed mutagenesis of each of seven potentialN-glycosylation sequons showed that four sites were required to generate ‘‘immuno-genic’’ tyrosinase. In the latter example, it was shown that other structure perturba-tions to gp90 such as reduction and thermal denaturation (which did not affectffffglycosylation) also led to loss of antigenicity – consistent with the notion thatdenaturation renders the protein susceptible to ‘‘overprocessing’’.

REFERENCES

Adorini, L., Appella, E., Doria, G., Cardinaux, F. and Nagy, Z.A. (1989) Competition for antigen presenta-

tion in living cells involves exchange of peptides bound by class II MHC molecules. Nature, 342,

800–803.

Allen, P.M. and Unanue, E.R. (1984) Differential requirements for antigen processing by macrophages forffff

lysozyme-specific T cell hybridomas. J Immunol, 132, 1077–1079.

Castellino, F., Zappacosta, F., Coligan, J.E. and Germain, R.N. (1998) Large protein fragments as sub-

strates for endocytic antigen capture by MHC class II molecules. J Immunol, 161, 4048–4057.

Chesnut, R.W., Endres, R.O. and Grey, H.M. (1980) Antigen recognition by T cells and B cells: recognition

of cross-reactivity between native and denatured forms of globular antigens. Clin Immunol

Immunopathol, 15, 397–408.

Cresswell, P. (1994) Assembly, transport, and function of MHC class II molecules. Annu Rev Immunol,

12, 259–293.

Deng, H., Apple, R., Clare-Salzler, M., Trembleau, S., Mathis, D., Adorini, L. and Sercarz, E. (1993)

Determinant capture as a possible mechanism of protection afforded by major histocompatibilityffff

complex class II molecules in autoimmune disease. J Exp Med, 178, 1675–1680.

Denzin, L.K. and Cresswell, P. (1995) HLA-DM induces CLIP dissociation from MHC class II alpha beta

dimers and facilitates peptide loading. Cell, 82, 155–165.

Donermeyer, D.L. and Allen, P.M. (1989) Binding to Ia protects an immunogenic peptide from proteolytic

degradation. J Immunol, 142, 1063–1068.

Fass, D., Davey, R.A., Hamson, C.A., Kim, P.S., Cunningham, J.M. and Berger, J.M. (1997) Structure of a

murine leukemia virus receptor-binding glycoprotein at 2.0 angstrom resolution. Science, 277,

1662–1666.

Germain, R.N. (1994) MHC-dependent antigen processing and peptide presentation: providing ligands for

T lymphocyte activation. Cell, 76, 287–299.

Gliniak, B.C., Kozak, S.L., Jones, R.T. and Kabat, D. (1991) Disulfide bonding controls the processing of

retroviral envelope glycoproteins. J Biol Chem, 266, 22991–22997.

Golgher, D., Korangy, F., Gao, B., Gorski, K., Jaffee, E., Edidin, M., Pardoll, D.M. and Elliott, T. (2001)ffff

An immunodominant MHC class II-restricted tumor antigen is conformation dependent and binds

to the endoplasmic reticulum chaperone, calreticulin. J Immunol, 167, 147–155.

Griffin, J.P., Chu, R. and Harding, C.V. (1997) Early endosomes and a late endocytic compartment generate

different peptide-class II MHC complexes via distinct processing mechanisms.ffff J Immunol, 158,

1523–1532.

Harding, C.V. and Unanue, E.R. (1989) Antigen processing and intracellular Ia. Possible roles of endo-

cytosis and protein synthesis in Ia function. J Immunol, 142, 12–19.

Helenius, A. and Aebi, M. (2001) Intracellular functions of N-linked glycans. Science, 291, 2364–2369.

Hiltbold, E.M. and Roche, P.A. (2002) Trafficking of MHC class II molecules in the late secretory pathway.

Curr Opin Immunol, 14, 30–35.

Housseau, F., Moorthy, A., Langer, D.A., Robbins, P.F., Gonzales, M.I. and Topalian, S.L. (2001) N-linked

carbohydrates in tyrosinase are required for its recognition by human MHC class II-restricted

CD4(+) T cells. Eur J Immunol, 31, 2690–2701.Huang, A.Y., Gulden, P.H., Woods, A.S., Thomas, M.C., Tong, C.D., Wang, W., Engelhard, V.H.,

Pasternack, G., Cotter, R., Hunt, D., Pardoll, D.M. and Jaffee, E.M. (1996) The immunodominantffff


major histocompatibility complex class I-restricted antigen of a murine colon tumor derives from an

endogenous retroviral gene product. Proc Natl Acad Sci USA, 93, 9730–9735.

Jensen, P.E. (1993) Acidification and disulfide reduction can be sufficient to allow intact proteins to bind

class II MHC. J Immunol, 150, 3347–3356.

Lee, P., Matsueda, G.R. and Allen, P.M. (1988) T cell recognition of fibrinogen. A determinant on the A

alpha-chain does not require processing. J Immunol, 140, 1063–1068.

Lenz, J., Crowther, R., Straceski, A. and Haseltine, W. (1982) Nucleotide sequence of the Akv env gene. J

V irol, 42, 519–529.

Lindner, R. and Unanue, E.R. (1996) Distinct antigen MHC class II complexes generated by separate

processing pathways. Embo J, 15, 6910–6920.

Ma, C., Whiteley, P.E., Cameron, P.M., Freed, D.C., Pressey, A., Chen, S.L., Garni-Wagner, B., Fang, C.,

Zaller, D.M., Wicker, L.S. and Blum, J.S. (1999) Role of APC in the selection of immunodominant

T cell epitopes. J Immunol, 163, 6413–6423.

Nadimi, F., Moreno, J., Momburg, F., Heuser, A., Fuchs, S., Adorini, L. and Hammerling, G.J. (1991)

Antigen presentation of hen egg-white lysozyme but not of ribonuclease A is augmented by the major

histocompatibility complex class II-associated invariant chain. Eur J Immunol, 21, 1255–1263.

Nelson, C.A., Roof, R.W., McCourt, D.W. and Unanue, E.R. (1992) Identification of the naturally pro-

cessed form of hen egg white lysozyme bound to the murine major histocompatibility complex class

II molecule I-Ak. Proc Natl Acad Sci USA, 89, 7380–7383.

Pardoll, D.M. and Topalian, S.L. (1998) The role of CD4+ T cell responses in antitumor immunity. CurrOpin Immunol, 10, 588–594.

Parodi, A.J. (2000) Protein glucosylation and its role in protein folding. Annu Rev Biochem, 69, 69–93.

Pinet, V., Malnati, M.S. and Long, E.O. (1994) Two processing pathways for the MHC class II-restricted

presentation of exogenous influenza virus antigen. J Immunol, 152, 4852–4860.

Pinet, V., Vergelli, M., Martin, R., Bakke, O. and Long, E.O. (1995) Antigen presentation mediated by

recycling of surface HLA-DR molecules. Nature, 375, 603–606.

Pinter, A. and Fleissner, E. (1977) The presence of disulfide-linked gp70-p15(E) complexes in AKR murine

leukemia virus. V irology, 83, 417–422.

Pinter, A., Lieman-Hurwitz, J. and Fleissner, E. (1978) The nature of the association between the murine

leukemia virus envelope proteins. V irology, 91, 345–351.

Sercarz, E.E., Lehmann, P.V., Ametani, A., Benichou, G., Miller, A. and Moudgil, K. (1993) Dominance

and crypticity of T cell antigenic determinants. Annu Rev Immunol, 11, 729–766.

Sercarz, E.E. and Maverakis, E. (2003) Mhc-guided processing: binding of large antigen fragments. Nat

Rev Immunol, 3, 621–629.

Sette, A., Adorini, L., Colon, S.M., Buus, S. and Grey, H.M. (1989) Capacity of intact proteins to bind to

MHC class II molecules. J Immunol, 143, 1265–1267.

Sherman, M.A., Weber, D.A. and Jensen, P.E. (1995) DM enhances peptide binding to class II MHC by

release of invariant chain-derived peptide. Immunity, 3, 197–205.

Sitia, R. and Braakman, I. (2003) Quality control in the endoplasmic reticulum protein factory. Nature,

426, 891–894.

Sjolander, S., Bolmstedt, A., Akerblom, L., Horal, P., Olofsson, S., Morein, B. and Sjolander, A. (1996)

N-linked glycans in the CD4-binding domain of human immunodeficiency virus type 1 envelope

glycoprotein gp160 are essential for the in vivo priming of T cells recognizing an epitope located in

their vicinity. V irology, 215, 124–133.

Sloan, V.S., Cameron, P., Porter, G., Gammon, M., Amaya, M., Mellins, E. and Zaller, D.M. (1995)

Mediation by HLA-DM of dissociation of peptides from HLA-DR. Nature, 375, 802–806.

Streicher, H.Z., Berkower, I.J., Busch, M., Gurd, F.R. and Berzofsky, J.A. (1984) Antigen conformation

determines processing requirements for T-cell activation. Proc Natl Acad Sci USA, 81, 6831–6835.

Surman, S., Lockey, T.D., Slobod, K.S., Jones, B., Riberdy, J.M., White, S.W., Doherty, P.C. and Hurwitz,

J.L. (2001) Localization of CD4+ T cell epitope hotspots to exposed strands of HIV envelopeglycoprotein suggests structural influences on antigen processing. Proc Natl Acad Sci USA, 98,4587–4592.

Trombetta, E.S. (2003) The contribution of N-glycans and their processing in the endoplasmic reticulum

to glycoprotein biosynthesis. Glycobiology, 13, 77R-91R.Villadangos, J.A., Driessen, C., Shi, G.P., Chapman, H.A. and Ploegh, H.L. (2000) Early endosomal

maturation of MHC class II molecules independently of cysteine proteases and H-2DM. Embo J,19, 882–891.

94 Y. Mimura et al.

Viner, N.J., Nelson, C.A. and Unanue, E.R. (1995) Identification of a major I-Ek-restricted determinant ofhen egg lysozyme: limitations of lymph node proliferation studies in defining immunodominance and

crypticity. Proc Natl Acad Sci USA, 92, 2214–2218.

Watts, C. (2001) Antigen processing in the endocytic compartment. Curr Opin Immunol, 13, 26–31.

Watts, C. (2004) The exogenous pathway for antigen presentation on major histocompatibility complex

class II and CD1 molecules. Nat Immunol, 5, 685–692.

Wolf, P.R. and Ploegh, H.L. (1995) How MHC class II molecules acquire peptide cargo: biosynthesis and

trafficking through the endocytic pathway. Annu Rev Cell Dev Biol, 11, 267–306.

Zhong, G., Romagnoli, P. and Germain, R. N. (1997) Related leucine-based cytoplasmic targeting signals

in invariant chain and major histocompatibility complex class II molecules control endocytic pre-

sentation of distinct determinants in a single protein. J Exp Med, 185, 429–438.

14

GLYCOSYLATION AND GPI ANCHORAGE OF THEPRION PROTEIN

N. M. Hooper

School of Biochemistry and MicrobiologyLeeds Institute for GeneticsHealth and TherapeuticsUniversity of Leeds, Leeds, UK

Prion diseases or transmissible spongiform encephalopathies are a group of neurod-egenerative disorders including scrapie in sheep, bovine spongiform encephalopathyin cattle, Creutzfeldt–Jakob disease and Gerstmann–Straussler–Scheinker disease inhumans. In prion diseases the normal cellular form of the prion protein (PrPC)undergoes a conformational conversion to the b-sheet-rich scrapie isoform (PrPSc).Although PrPC is critical for the development of prion disease through its conversioninto PrPSc, the physiological role of PrPC is less clear. PrPC undergoes a variety ofpost-translational processing events, including glycosylation, GPI anchorage and

Figure 1.



96 N. M. Hooper

proteolysis. PrP contains two N-glycosylation sequons at Asn180 and Asn196, bothof which can be glycosylated. However, either one or both of the acceptor Asn mayremain unglycosylated, giving rise to unglycosylated, mono-glycosylated anddi-glycosylated proteins. The ratio of these three glycoforms appears to be character-istic for particular ‘strains’ of PrP. N-glycosylation of PrP is dramatically influencedby its membrane topology and by the distance of the Asn sequons to the C-terminusof the protein. The role of these post-translational modifications in the life cycle ofPrPC will be discussed. The C-terminal signal peptide directs the addition of aglycosyl-phosphatidylinositol (GPI) anchor to the protein within the lumen of theER. This GPI anchor, along with a determinant in the N-terminal region of theprotein, promotes the association of PrPC with cholesterol-rich lipid rafts that areinvolved in the trafficking of the protein and its conversion to PrPSc.

REFERENCES

1. Walmsley, A. R., Zeng, F. and Hooper, N. M. (2001). Membrane topology influences N-glycosylation of

the prion protein. EMBO J. 20, 703–712.

2. Walmsley, A. R. and Hooper, N. M. (2003). Distance of sequons to the C-terminus influences the cellular

N-glycosylation of the prion protein. Biochem. J. 370, 351–355.

3. Walmsley, A. R., Zeng, F. and Hooper, N. M. (2003). The N-terminal region of the prion protein

ectodomain contains a lipid raft targeting determinant. J. Biol. Chem. 278, 37241–37248.

15

GLYCOSYLATION DEFECTS AND MUSCULARDYSTROPHY

Derek J. Blake1, Christopher T. Esapa1, Enca Martin-Rendon2, andR. A. Jeffrey McIlhinneyffff 3

1Department of Pharmacology, University of Oxford. MansfieldRoad, Oxford, UK2Stem Cell Research Laboratory, National Blood Service, OxfordCentre, Oxford, UK3MRC Anatomical Neuropharmacology Unit, University of Oxford,Oxford, UK

Glycosylation is an important post-translational modification of many proteins inthe secretory pathway. Mutations in several genes involved in glycan metabolismare known to cause different types of congenital disorders of glycosylation; a genetic-ffffally heterogeneous group of diseases that affect multiple organs and are frequentlyffffassociated with developmental delay, haematological and immunological anomalies.In addition to these disorders, it is now apparent that at least six different forms offfffmuscular dystrophy are caused by genes that encode actual or putative glycosyltransf-erases. Although these diseases are clinically distinct, they are all associated with asecondary deficiency in the glycosylation of a-dystroglycan. Hypo-glycosylation ofa-dystroglycan disrupts a link between the membrane and proteins in the extracellularmatrix such as laminin, resulting in muscle disease and in several cases a neuronalmigration disorder. At least three allelic disorders are caused by mutations in thegene encoding fukutin-related protein (FKRP). These are; congenital muscular dys-trophy type 1C (MDC1C), limb girdle muscular dystrophy 2I (LGMD2I) andcongenital muscular dystrophy (CMD) with brain malformations and mental retard-ation. These diseases result from any one of 36 different missense mutations in theffffgene encoding FKRP. FKRP is a DxD motif-containing type II membrane proteinthat is targeted to the medial Golgi-apparatus by an N-terminal signal anchorsequence. We have found that FKRP mutations associated with the most severedisease phenotypes result in retention of the mutant protein in the endoplasmicreticulum (Figure 1). The ER-retained mutants have a prolonged association withcalenxin and are preferentially degraded by the proteasome. These data suggest that



98 D. J. Blake et al.

Figure 1.

impaired intracellular trafficking and proteasomal degradation contribute to thecellular pathology of CMD caused by mutations in the FKRP gene.

REFERENCES

1. Brockington M et al., (2001) Am J. Hum. Genet. 69:1198–209.

2. Esapa CT et al., (2002) Hum. Mol. Genet. 11: 3319–3331.

3. Martin-Rendon E and Blake DJ (2003) T rends Pharm. Sci. 24: 178–183.

16

ROLES OF COMPLEX AND HYBRID N-GLYCANSAND O-FUCOSE GLYCANS IN OOCYTEDEVELOPMENT AND FUNCTION

S. Shi, S. A. Williams, H. Kurniawan, L. Lu, and P. Stanley

Department of Cell BiologyAlbert Einstein College MedicineNew York, NY 10461, USA

Roles for complex or hybrid N-glycans in oocyte maturation and function wereinvestigated using female mice with a floxed Mgat1 gene (Mgat1F) carrying a Cre-recombinase transgene under the control of the zona pellucida protein 3 (ZP3)promoter (1). Inactivation of theMgat1 gene responsible for the synthesis of complexand hybrid N-glycans is embryonic lethal, but homozygous mutant blastocysts arerescued by maternal Mgat1 gene transcripts. Following deletion of the Mgat1 genein oocytes, females with mutant oocytes had reduced fertility and fewer oocytes aftersuperovulation. All mutant oocytes had a zona pellucida (ZP) that contained ZP1,ZP2 and ZP3 glycoproteins but they did not contain complex N-glycans and mutantZP were thin, deformed and loosely attached (Fig. 1). Nevertheless, mutant oocyteswere efficiently fertilized, all embryos implanted and ~70% developed to E9.5(Mgat1−/−) or birth (Mgat1+/− or Mgat1+/+). However, ~40%–50% embryos were

Figure 1. Mgat1+/+ oocyte ( left); Mgat1−/− oocyte with ZP (right).



100 S. Shi et al.

retarded in development to various degrees at E3.5. Surprisingly, heterozygousblastocysts that synthesized complex N-glycans were both retarded and normal inmorphology at E3.5 indicating that a proportion of Mgat1−/− oocytes are unableto support development of a zygote, even when they synthesize complex and hybridN-glycans. Additionally unexpected, was the finding that Mgat1−/− zygotes coulddevelop normally during pre-implantation, implant and progress to E9.5 in theabsence of complex and hybrid N-glycans. Therefore sugars on complex or hybridN-glycans are required at some stage of oogenesis for the generation of a develop-mentally competent oocyte, but fertilization, pre-implantation development andimplantation may proceed in their absence. When the Pofut1ff gene was deleted inoocytes using the ZP3Cre transgene, a similar result was obtained. The Pofut1ff geneis responsible for initiating the synthesis of O-fucose glycans found on certain EGFrepeats of Notch receptors and their ligands (2). Deletion of Pofut1ff gives embryoniclethality and a phenotype characteristic of embryos with severe Notch signalingdefects (3). After deletion in oocytes, while ZP and egg morphology were normal,about 50% females had a reduced litter size. These females were found to have ahigh proportion of malformed and under-developed embryos at E3.5, although thetotal number of pre-implantation embryos was normal. Therefore a loss of Pofut1in oocytes did not appear to inhibit oocyte maturation, ovulation or fertilization butgenerated a majority of oocytes that were unable to facilitate development to normalblastocysts following fertilization. The implications of these findings for the roles ofNotch signaling in oocyte development will be discussed.

REFERENCES

(1) Ye, Z. and Marth J. D. (2004) Glycobiology, 14, 547–558.

(2) Wang, Y., Shao, L., Shi, S., Harris, R. J., Spellman, M. W., Stanley, P., Haltiwanger, R. S. (2001)WW J. Biol

Chem. 276, 40338–40345.

(3) Shi, S. and Stanley, P. (2003) Proc. Natl. Acad. Sci. USA., 100, 5234–5239.

17

MUCIN OLIGOSACCHARIDES AND PIGEONFANCIERS’ LUNG

C. I. Baldwin1, A. Allen2, S. Bourke3, E. Hounsell4, and J. E. Calvert2

1School of Applied SciencesNorthumbria UniversityNewcastle upon Tyne, UK2School of Cell and Molecular BiosciencesNewcastle UniversityNewcastle upon Tyne, UK3Department of Respiratory MedicineRoyal Victoria InfirmaryNewcastle upon Tyne, UK4School of Biological and Chemical SciencesBirkbeck, University of London, UK

Pigeon fanciers’ lung (PFL), a form of extrinsic allergic alveolitis, is an immunolo-gically mediated lung disease that occurs in susceptible individuals after repeatedinhalation of pigeon antigens. The disease is characterised by hypersensitivity reac-tions that occur in the distil bronchioles and alveoli which may lead to irreversiblepulmonary fibrosis. Previous studies have shown that IgG responses to carbohydratedeterminants on pigeon intestinal mucin (PIM) are key to the development ofdisease. To specifically identify carbohydrate epitopes on PIMO-linked oligosacchar-ides were released by hydrazinolysis, separated by reverse phase-HPLC and thereleased free reducing oligosaccharides were then coupled to poly-L-lysine (PLL).The antigenicity of the resultant polyvalent conjugates was tested with sera frompigeon fanciers by both dot blot and ELISA. Further to this two dimensional 1HNMR spectroscopy and GC-MS was used to determine the structure of one of themajor antigenic oligosaccharides of PIM. The sera reacted with a large number ofPLL-oligosaccharide conjugates confirming that anti-PIM IgG1 and IgG2 recogniseO-linked oligosaccahrides. ELISA showed that IgG1 responses were mainly toglycoforms in the earlier fractions (2–20) whilst IgG2 responses were directed againstsome of the very early fractions (2–10) and the later fractions (21–35) suggestingthat there are at least two major epitopes on PIM. Furthermore IgG1 responses

101



102 C. I. Baldwin et al.

were significantly higher in symptomatic individuals against oligosaccharides foundin fractions 15–17. As anti-mucin IgG1 responses correlate with disease it may bethat these fractions contain disease associated epitopes. Structural analysis revealedthe presence of two novel forms of LeX substituted at C-3 of the b-Gal with eitherb-GalNAC or b-GlcNAc. These glycoforms have not been described in mammalianglycosystems and one may expect both to be highly antigenic. These observationssupport the concept that PIM is a key antigen in the development of PFL.

18

DIFFERENTIAL GLYCOSYLATION OFGELATINASE B FROM NEUTROPHILS ANDBREAST CANCER CELLS

Simon A. Fry1, Philippe E. Van den Steen2, Louise Royle1,Mark R. Wormald1, Anthony J. Leathem3, Ghislain Opdenakker2,Pauline M. Rudd1, and Raymond A. Dwek1

1Glycobiology InstituteDepartment of BiochemistryUniversity of OxfordSouth Parks Road, Oxford OX1 3QU, UK2Rega Institute for Medical ResearchLaboratory of Molecular ImmunologyUniversity of LeuvenMinderbroedersstraat 10, B-3000 Leuven, Belgium3Department of SurgeryRoyal Free and University College London Medical School67–73 Riding House Street, London W1W 1EJ, UK

1. INTRODUCTION

The matrix metalloproteases (MMPs) are a family of zinc-dependant endopep-tidases (Nagase and Woessner 1999) that are involved in extracellular matrix (ECM)remodelling in a variety of physiological and pathological processes. MMP degrada-tion of ECM proteins is associated with many aspects of cancer progression, includingcancer cell growth, differentiation, apoptosis, migration and invasion, as well asffffregulation of angiogenesis and immune surveillance (Egeblad and Werb 2002).Gelatinase B (MMP-9) is structurally one of the most complex MMPs

(Opdenakker, Van den Steen et al. 2001; Van den Steen, Dubois et al. 2002). AllMMPs have a prodomain (proteolytically removed to yield active enzyme), an activedomain and a zinc-binding domain. Gelatinase B also has a gelatin-binding fibronec-tin domain, a collagen type V-like domain and a carboxyterminal hemopexin domain.Human natural gelatinase B from neutrophils is heavily glycosylated (Rudd, Mattu

103



104 S. A. Fry et al.

et al. 1999; Mattu, Royle et al. 2000). 3 consensus sequences for the attachment ofN-linked glycans are present in gelatinase B, 1 in the prodomain and 2 in the activedomain. N-glycan site analysis revealed that only 2 of the 3 sites are occupied (Vanden Steen, Opdenakker et al. 2001; Kotra, Zhang et al. 2002), one each in theprodomain (Asn38-Leu-Thr) and active domain (Asn120-Tyr-Ser). The O-linked gly-cans are probably located in the Thr, Ser and Pro rich collagen V- like domain,since this sequence contains ideal clustered attachment sites for O-linked oli-gosaccharides.

Gelatinase B is typically stored in large amounts in neutrophil granules to allowrapid release in innate immune responses (Masure, Proost et al. 1991). Expressionis increased in many cancer cell types, including breast carcinomas (Zucker, Lysiket al. 1993), and experimental metastasis assays demonstrate that it is required foreffective cell invasion and metastasis (Hua and Muschel 1996; Itoh, Taniokaffff et al.1999). Gelatinase B has these biological properties because some of its substratesare ECM proteins. As well as cleaving gelatins (denatured collagens), gelatinase Bcan cleave the 3

4fragment of collagen type II (produced following collagenase cleavage

at a single site) (Van den Steen, Proost et al. 2004), collagen type V (Hibbs, Hoidalet al. 1987) and perhaps (Mackay, Hartzler et al. 1990) collagen type IV (a majorcomponent of basement membranes) (Wilhelm, Collier et al. 1989; Okada, Gonojiet al. 1992). Other gelatinase B ECM substrates include elastin (Senior, Griffin et al.1991), aggregan (Fosang, Neame et al. 1992), link protein (Nguyen, Murphy et al.1993) and galectin-3. (Ochieng, Fridman et al. 1994)

Aberrant protein glycosylation is a common feature of cancer (Feizi 1985; Saitoh,Wang et al. 1992; Lloyd, Burchell et al. 1996; Kim and Varki 1997; Granovsky, Fataet al. 2000). This chapter discusses the known glycosylation of natural neutrophilgelatinase B compared to that of gelatinase B secreted from MCF-7 breast cancercells.

2. MATRIX METALLOPROTEASE GLYCOSYLATION

MMPs are synthesised as latent proenzymes that are activated by proteolysis.As such, MMPs act sequentially in a protease cascade that produces active proteaseswhose combined effects mediate ECM remodelling (Fig. 1). Alterations in the glycos-ffffylation of certain MMPs can alter their enzymatic activity. This has consequencesin terms of individual MMP properties and more widely in terms of the proteasecascade.

Plasminogen has 3 N-linked glycosylation sites and occurs naturally in 2 popula-tions that differ only by the absence of anffff N-glycan at N288. The presence of an N-glycan at this site down regulates plasminogen activation by tissue-type plasminogenactivator (Rudd, Woods et al. 1995). Although the observed effect is only a 2-foldffffreduction, these enzymes initiate the protease cascade which terminates in gelatinaseB activation. A similar glycosylation-induced alteration in activation in other MMPswould be amplified through the cascade providing a potentially powerful means ofregulation.

Increased b1–6 GlcNAc branching of N-linked glycans has long been associatedwith tumour metastasis (Dennis, Laferte et al. 1987; Fernandes, Sagman et al. 1991).

Differential Glycosylation of Gelatinase Bffff 105

Figure 1. The protease cascade. Each enzyme has another enzyme as a substrate, arrows indicate the

activation cascade of latent proenzymes. u-PA, urokinase-type plasminogen activator; t-PA, tissue-type

plasminogen activator.

Increased b1–6 GlcNAc branching (mediated by UDP-GlcNAc a-mannosideb1–6–N-acetylglucosaminyltransferase) leads to increased resistance of matriptase todegradation (Ihara, Miyoshi et al. 2002). This in turn promotes matrix metallo-proteinase activation through enhanced activation of urokinase-type plasminogenactivator (Fig. 1).

MMPs, including gelatinase B, that have been secreted and activated can stillbe regulated by Tissue Inhibitor of Metalloprotease (TIMP) inhibition (Murphy andDocherty 1992). TIMP-1 exerts its inhibitory effect by binding with high affff ffinity toboth progelatinase B and activated gelatinase B (Murphy, Houbrechts et al. 1991).However, enzymatic desialylation of gelatinase B has been demonstrated to reduceTIMP-1 inhibition by ~50% (Van den Steen, Opdenakker et al. 2001). Together,these findings highlight how intimately glycosylation is associated with gelatinase Bactivity and activation.

3. PRODUCTION OF MCF-7 GELATINASE B ANDGLYCOSYLATION ANALYSIS

In order to perform glycan analysis, 100 mg of gelatinase B was required. Phorbol-myristate-acetate (PMA) is known to upregulate expression levels of gelatinase B in


many cell types by directly activating protein kinase C (PKC) (Masure, Proost et al.1991; Opdenakker, Masure et al. 1991; Houde, de Bruyne et al. 1993). ActivatedPKC phosphorylates a range of substrates leading to enhanced transcription of thegelatinase B gene (Houde, de Bruyne et al. 1993). Optimal gelatinase B expressionlevels were achieved by incubating MCF-7 cells in serum free media for 48 hours inthe presence of 10 ng/ml PMA (data not shown). Cell supernatants were harvested,filtered and purified by affinity chromatography on gelatin-sepharose as previouslydescribed (Masure, Proost et al. 1991).In order to liberate N-glycans from purified gelatinase B, samples were reduced

and alkylated prior to SDS-PAGE. Gel bands were cut out and washed alternativelywith 20mM sodium bicarbonate and acetonitrile. N-glycans were removed fromprotein in the gel bands by overnight incubation with Peptide N-glycosidase F at37°C. N-linked glycans were recovered from the gel pieces with water and sonicationbefore being dried and 2-AB labeled. As O-glycans can not be removed enzymatically,manual hydrazinolysis was performed as previously described (Royle, Mattu et al.2002). Briefly, gelatinase B samples were dialysed against 0.1% trifluoroacetic acid,lyophilised and then cryogenically dried. O-linked glycans were released by incuba-tion with anhydrous hydrazine for 6hr at 60°C, then re-N-acetylated and desaltedbefore peptides were removed. Released N- and O-glycans were fluorescently labelledwith 2-AB according to the method of Bigge et al. (Bigge, Patel et al. 1995)

2-AB labelled N- and O-linked glycans of gelatinase B were resolved byNP-HPLC. By comparison with a standard dextran hydolysate ladder, oligosacchar-ide elution positions could be expressed as glucose units (GU). Oligosaccharidestructures were assigned by reference to the GU values of a database of standardsugars (Royle, Mattu et al. 2002), and confirmed by using a series of parallelexoglycosidase digestions.

4. GELATINASE B GLYCOSYLATION

MCF-7 gelatinase B is post-translationally modified differently to natural neu-fffftrophil gelatinase B. The N- and O-linked glycans of neutrophil gelatinase B havebeen sequenced (Rudd, Mattu et al. 1999; Mattu, Royle et al. 2000). More than 95%of the N-linked glycans are partially sialylated, core fucosylated bianntennary struc-tures with and without outer arm a1–3 linked fucose (Fig. 2a). The O-linked glycans(ranging from 2–10 monosaccharides) mainly consisted of type II cores with lactosam-ine extensions, with or without sialic acid or outer arm fucose (Fig. 2b).The N-linked glycans of MCF-7 gelatinase B are all core-fucosylated bianten-

nary structures (Fig. 3a). There is one mono-sialylated and one disialylatedN-glycan,and all outer arm fucosylation occurs through a mixture of a1–2 and a1–3 linkages.The O-linked glycans of MCF-7 gelatinase B differ from those of neutrophil gelatinaseffffB more dramatically than the N-linked glycans. MCF-7 gelatinase B O-glycans arerelatively small (in the range of 2–6 monosaccharides) and heavily sialylated withno fucosylation (Fig 3b). The most abundant O-glycans have type I cores(Galb1–3GalNAc) with fewer galactosylated core II structures (Galb1–4GlcNAcb1–6[Galb1–3]GalNAc) present.


Figure 2. N- and O-Linked glycans of neutrophil gelatinase B. a) Diagrammatic representation of a

neutrophil gelatinase B bianntennary N-linked glycan. Glycosylation analysis showed this glycan is pre-

dominantly present in a core-fucosylated form, with varying amounts of outer arm a1–3 fucosylation or

sialylation. b) A typical neutrophil gelatinase B O-linked glycan. Core II O-glycans containing 1–3 lactosa-

mine repeats predominate. Fucosylation (a1–3linked) and/or sialylation (a2–3 and a2–6 linked) are also

present. The scheme for glycan representation is as follows: #, mannose; 2, N-acetylgalactosamine; 1,galactosamine; &, N-acetylglucosamine; k, fucose; 0, sialic acid; dashed line, a-linkage; full line,b-linkage; , linkage position.

5. DISCUSSION

The N-linked glycans of human neutrophil and MCF-7 gelatinase B are verysimilar. The main difference is the presence offfff a1–2 linked outer arm fucose onMCF-7 gelatinase B N-glycans (absent from neutrophil gelatinase B N-glycans). Afunction is yet to be ascribed to gelatinase B N-glycans, although 1 of the 2 N-glycans is known to reside on the propeptide, so could conceivably influence proteo-lytic activation. The minor differences in neutrophil and MCF-7ffff N-glycosylation areunlikely to impact on the biological activity of gelatinase B.In contrast to the N-linked glycans, there are dramatic differences in theffff O-

glycans. In comparison to neutrophil gelatinase B O-glycans, the O-glycans of MCF-7gelatinase B are truncated by the addition of sialic acid to core I structures (Fig. 4).This finding is consistent with that of truncation of MUC1 O-glycans in breastcancer (Lloyd, Burchell et al. 1996; Dalziel, Whitehouse et al. 2001). The majorityof neutrophil gelatinase B core I O-glycans are N-acetylglucosylated to form core IIstructures. These are further elongated with N-acetyllactosamine repeats to produceextended O-linked glycans that are completely absent from MCF-7 gelatinase B.Interestingly, these differences inffff O-glycosylation will alter the nature of interactionwith the gelatinase B substrate, galectin-3.


Figure 3. N- and O-linked glycans of MCF-7 gelatinase B. a) Diagrammatic representation of an MCF-7

gelatinase B bianntennary N-linked glycan. Glycosylation analysis shows this glycan is always core-

fucosylated, with varying amounts of outer arm a1–3 and/or a1–2 linked fucose or a2–3 linked sialic

acid. b) A typical MCF-7 gelatinase B O-linked glycan. Mono- and disialylated core I O-glycans predomin-

ate. There is an absence of fucosylation. The scheme for glycan representation is as follows: #, mannose;2, N-acetylgalactosamine; 1, galactosamine; &, N-acetylglucosamine; k, fucose; 0, sialic acid; dashed

line, a-linkage; full line, b-linkage; , linkage position.

Galectin-3 is a 30kDa b-galactoside binding lectin that has a carboxy-terminalcarbohydrate recognition domain (CRD) (Seetharaman, Kanigsberg et al. 1998) anda flexible Pro, Tyr and Gly-rich amino-terminal domain that contains a gelatinaseB cleavage site. This amino-terminal domain is required for the non-covalent self-association of between 2 and 5 galectin-3 molecules (Ochieng, Platt et al. 1993;Barboni, Bawumia et al. 1999; Ahmad, Gabius et al. 2004), allowing cross-linkingof suitable glycans. Breast cancer galectin-3 expression has been reported to correlatepositively with disease progression (Honjo, Nangia-Makker et al. 2001). Also,galectin-3 induced expression of cell surface adhesion molecules and the directinteraction of galectin-3 with suitably glycosylated proteins, such as laminin andintegrins, may favour tumorigenesis (Ochieng, Leite-Browning et al. 1998; Matarrese,Fusco et al. 2000).A property of galectin-3 is a dramatic increase in affinity in response to increases

in repeating lactosamine units in the substrate; ie, LN (Kd=26 mM), LN2 (1.3 mM),LN3 (0.35 mM) (Hirabayashi, Hashidate et al. 2002). The glycan analysis performedhere shows that neutrophil gelatinase B has a higher proportion of O-glycans thatare galectin-3 ligands, some of which contain lactosamine repeats, than MCF-7gelatinase B (Table 1). Indeed, changes in galectin-3 binding to neutrophil andMCF-7 gelatinase B have been demonstrated (Fry et al., in press). Truncation ofgelatinase B O-glycans reduces galectin-3 binding and may therefore facilitate cancercell invasion.

Gelatinase B is the only MMP to contain a collagen type V-like domain. This


Figure 4. Neutrophil and MCF-7 gelatinase B O-glycans differ in their core structures. Core structures offfff

O-linked glycans are represented as a % of the total O-glycan pool. a) 70.1% of neutrophil gelatinase B

O-glycans have a type II core. b) 90.4% of MCF-7 gelatinase B O-glycans have a type I core.

Table 1. Gelatinase B O-linked glycans that are galectin-3 ligands. O-glycans that are

galectin-3 ligands are represented as a % of the total O-glycan pool. LNxNN , number (x) of

lactosamine repeats.

% O-glycans that are galectin-3 ligands

Gelatinase B source LN1 LN2 LN3 Total

Neutrophil 36.3 6.1 3.8 46.2

MCF-7 8.6 – – 8.6

56 amino acid sequence contains repeating Ser, Thr and Pro residues and thus isthought to be the site of clustered O-linked glycans (18 potential O-glycosylationsites as predicted by the NetOGlyc 3.1 server, www.cbs.dtu.dk/services/NetOGlyc/).The function of this domain is as yet unknown, but similar peptide sequences thatare heavily O-glycosylated are extended and possess an increased rigidity of thepeptide chain (Lukacik, Roversi et al. 2004). Hence, a role for the repeating Ser, Thrand Pro residues may be to extend the polypeptide chain to maximise separation


between the hemopexin and the other domains. The O-glycans of MCF-7 gelatinaseB are all short and carry negatively charged sialic acid. Charge repulsion mayfacilitate peptide backbone extension and alter gelatinase B properties accordingly.Gelatinase B secreted from MCF-7 breast cancer cells is aberrantly O-glycosyl-

ated when compared to natural neutrophil gelatinase B. Although N-glycosylationprofiles are similar, MCF-7 O-glycans are truncated and heavily sialylated comparedto those of neutrophil gelatinase B. This aberrant glycosylation results in a reductionin the proportion of gelatinase B O-glycans that are galectin-3 ligands, and a completeabsence of multiple N-acetyllactosamine repeats that are bound with high affinityby galectin-3. These cancer associated alterations in glycosylation may facilitatecancer cell metastasis by altering the peptide conformation of gelatinase B and/orits interaction with galectin-3.

REFERENCES

Ahmad, N., H. J. Gabius, et al. (2004). ‘‘Galectin-3 precipitates as a pentamer with synthetic multivalent

carbohydrates and forms heterogeneous cross-linked complexes.’’ J Biol Chem 279(12): 10841–7.

Barboni, E. A., S. Bawumia, et al. (1999). ‘‘Kinetic measurements of binding of galectin 3 to a laminin

substratum.’’ Glycoconj J 16 (7): 365–73.

Bigge, J. C., T. P. Patel, et al. (1995). ‘‘Nonselective and efficient fluorescent labeling of glycans using

2-amino benzamide and anthranilic acid.’’ Anal Biochem 230 (2): 229–38.

Dalziel, M., C. Whitehouse, et al. (2001). ‘‘The relative activities of the C2GnT1 and ST3Gal-I glycosyl-

transferases determine O-glycan structure and expression of a tumor-associated epitope on MUC1.’’

J Biol Chem 276 (14): 11007–15.

Dennis, J. W., S. Laferte, et al. (1987). ‘‘Beta 1–6 branching of Asn-linked oligosaccharides is directly

associated with metastasis.’’ Science 236 (4801): 582–5.

Egeblad, M. and Z. Werb (2002). ‘‘New functions for the matrix metalloproteinases in cancer progression.’’

Nat Rev Cancer 2(3): 161–74.

Feizi, T. (1985). ‘‘Demonstration by monoclonal antibodies that carbohydrate structures of glycoproteins

and glycolipids are onco-developmental antigens.’’ Nature 314(6006): 53–7.

Fernandes, B., U. Sagman, et al. (1991). ‘‘Beta 1–6 branched oligosaccharides as a marker of tumor

progression in human breast and colon neoplasia.’’ Cancer Res 51(2): 718–23.

Fosang, A. J., P. J. Neame, et al. (1992). ‘‘The interglobular domain of cartilage aggrecan is cleaved by

PUMP, gelatinases, and cathepsin B.’’ J Biol Chem 267(27): 19470–4.

Granovsky, M., J. Fata, et al. (2000). ‘‘Suppression of tumor growth and metastasis in Mgat5-deficient

mice.’’ Nat Med 6(3): 306–12.

Hibbs, M. S., J. R. Hoidal, et al. (1987). ‘‘Expression of a metalloproteinase that degrades native type V

collagen and denatured collagens by cultured human alveolar macrophages.’’ J Clin Invest 80(6):

1644–50.

Hirabayashi, J., T. Hashidate, et al. (2002). ‘‘Oligosaccharide specificity of galectins: a search by frontal

affinity chromatography.’’ Biochim Biophys Acta 1572 (2–3): 232–54.

Honjo, Y., P. Nangia-Makker, et al. (2001). ‘‘Down-regulation of galectin-3 suppresses tumorigenicity of

human breast carcinoma cells.’’ Clin Cancer Res 7(3): 661–8.

Houde, M., G. de Bruyne, et al. (1993). ‘‘Differential regulation of gelatinase B and tissue-type plasminogenffff

activator expression in human Bowes melanoma cells.’’ Int J Cancer 53 (3): 395–400.

Hua, J. and R. J. Muschel (1996). ‘‘Inhibition of matrix metalloproteinase 9 expression by a ribozyme

blocks metastasis in a rat sarcoma model system.’’ Cancer Res 56(22): 5279–84.

Ihara, S., E. Miyoshi, et al. (2002). ‘‘Prometastatic effect of N-acetylglucosaminyltransferase V is due toffff

modification and stabilization of active matriptase by adding beta 1–6 GlcNAc branching.’’ J Biol

Chem 277(19): 16960–7.

Itoh, T., M. Tanioka, et al. (1999). ‘‘Experimental metastasis is suppressed in MMP-9-deficient mice.’’ Clin

Exp Metastasis 17(2): 177–81.


Kim, Y. J. and A. Varki (1997). ‘‘Perspectives on the significance of altered glycosylation of glycoproteins

in cancer.’’ Glycoconj J 14 (5): 569–76.

Kotra, L. P., L. Zhang, et al. (2002). ‘‘N-Glycosylation pattern of the zymogenic form of human matrix

metalloproteinase-9.’’ Bioorg Chem 30 (5): 356–70.

Lloyd, K. O., J. Burchell, et al. (1996). ‘‘Comparison of O-linked carbohydrate chains in MUC-1 mucin

from normal breast epithelial cell lines and breast carcinoma cell lines. Demonstration of simpler and

fewer glycan chains in tumor cells.’’ J Biol Chem 271 (52): 33325–34.

Lukacik, P., P. Roversi, et al. (2004). ‘‘Complement regulation at the molecular level: the structure of decay-

accelerating factor.’’ Proc Natl Acad Sci U S A 101 (5): 1279–84.

Mackay, A. R., J. L. Hartzler, et al. (1990). ‘‘Studies on the ability of 65-kDa and 92-kDa tumor cell

gelatinases to degrade type IV collagen.’’ J Biol Chem 265 (35): 21929–34.

Masure, S., P. Proost, et al. (1991). ‘‘Purification and identification of 91-kDa neutrophil gelatinase. Release

by the activating peptide interleukin-8.’’ Eur J Biochem 198 (2): 391–8.

Matarrese, P., O. Fusco, et al. (2000). ‘‘Galectin-3 overexpression protects from apoptosis by improving

cell adhesion properties.’’ Int J Cancer 85 (4): 545–54.

Mattu, T. S., L. Royle, et al. (2000). ‘‘O-glycan analysis of natural human neutrophil gelatinase B using a

combination of normal phase-HPLC and online tandem mass spectrometry: implications for the

domain organization of the enzyme.’’ Biochemistry 39(51): 15695–704.

Murphy, G. and A. J. Docherty (1992). ‘‘The matrix metalloproteinases and their inhibitors.’’ Am J Respir

Cell Mol Biol 7(2): 120–5.

Murphy, G., A. Houbrechts, et al. (1991). ‘‘The N-terminal domain of tissue inhibitor of metalloproteinases

retains metalloproteinase inhibitory activity.’’ Biochemistry 30(33): 8097–102.

Nagase, H. and J. F. Woessner, Jr. (1999). ‘‘Matrix metalloproteinases.’’ J Biol Chem 274 (31): 21491–4.

Nguyen, Q., G. Murphy, et al. (1993). ‘‘Matrix metalloproteinases cleave at two distinct sites on human

cartilage link protein.’’ Biochem J 295 (Pt 2): 595–8.

Ochieng, J., R. Fridman, et al. (1994). ‘‘Galectin-3 is a novel substrate for human matrix

metalloproteinases-2 and –9.’’ Biochemistry 33 (47): 14109–14.

Ochieng, J., M. L. Leite-Browning, et al. (1998). ‘‘Regulation of cellular adhesion to extracellular matrix

proteins by galectin-3.’’ Biochem Biophys Res Commun 246 (3): 788–91.

Ochieng, J., D. Platt, et al. (1993). ‘‘Structure-function relationship of a recombinant human galactoside-

binding protein.’’ Biochemistry 32 (16): 4455–60.

Okada, Y., Y. Gonoji, et al. (1992). ‘‘Matrix metalloproteinase 9 (92-kDa gelatinase/type IV collagenase)

from HT 1080 human fibrosarcoma cells. Purification and activation of the precursor and enzymic

properties.’’ J Biol Chem 267(30): 21712–9.

Opdenakker, G., S. Masure, et al. (1991). ‘‘Cytokine-mediated regulation of human leukocyte gelatinases

and role in arthritis.’’ Lymphokine Cytokine Res 10 (4): 317–24.

Opdenakker, G., P. E. Van den Steen, et al. (2001). ‘‘Gelatinase B: a tuner and amplifier of immune

functions.’’ T rends Immunol 22(10): 571–9.

Royle, L., T. S. Mattu, et al. (2002). ‘‘An analytical and structural database provides a strategy for

sequencing O-glycans from microgram quantities of glycoproteins.’’ Anal Biochem 304(1): 70–90.

Rudd, P. M., T. S. Mattu, et al. (1999). ‘‘Glycosylation of natural human neutrophil gelatinase B and

neutrophil gelatinase B-associated lipocalin.’’ Biochemistry 38(42): 13937–50.

Rudd, P. M., R. J. Woods, et al. (1995). ‘‘The effects of variable glycosylation on the functional activities offfff

ribonuclease, plasminogen and tissue plasminogen activator.’’ Biochim Biophys Acta 1248 (1): 1–10.

Saitoh, O., W. C. Wang, et al. (1992). ‘‘Differential glycosylation and cell surface expression of lysosomalffff

membrane glycoproteins in sublines of a human colon cancer exhibiting distinct metastatic poten-

tials.’’ J Biol Chem 267 (8): 5700–11.

Seetharaman, J., A. Kanigsberg, et al. (1998). ‘‘X-ray crystal structure of the human galectin-3 carbohydrate

recognition domain at 2.1-A resolution.’’ J Biol Chem 273 (21): 13047–52.Senior, R. M., G. L. Griffin, et al. (1991). ‘‘Human 92– and 72-kilodalton type IV collagenases are elastases.’’

J Biol Chem 266(12): 7870–5.Van den Steen, P. E., B. Dubois, et al. (2002). ‘‘Biochemistry and molecular biology of gelatinase B or

matrix metalloproteinase-9 (MMP-9).’’ Crit Rev Biochem Mol Biol 37 (6): 375–536.Van den Steen, P. E., G. Opdenakker, et al. (2001). ‘‘Matrix remodelling enzymes, the protease cascade and

glycosylation.’’ Biochim Biophys Acta 1528 (2–3): 61–73.Van den Steen, P. E., P. Proost, et al. (2004). ‘‘Generation of glycosylated remnant epitopes from human

collagen type II by gelatinase B.’’ Biochemistry 43 (33): 10809–16.


Wilhelm, S. M., I. E. Collier, et al. (1989). ‘‘SV40-transformed human lung fibroblasts secrete a 92-kDa type

IV collagenase which is identical to that secreted by normal human macrophages.’’ J Biol Chem

264 (29): 17213–21.

Zucker, S., R. M. Lysik, et al. (1993). ‘‘M(r) 92,000 type IV collagenase is increased in plasma of patients

with colon cancer and breast cancer.’’ Cancer Res 53 (1): 140–6.

19

DETECTION OF GLYCOSYLATION CHANGES INSERUM AND TISSUE PROTEINS IN CANCER BYLECTIN BLOTTING

R. E. Ferguson, D. H. Jackson, R. Hutson, N. Wilkinson,P. Harnden, P. Selby, and R. E. Banks

Cancer Research UK Clinical CentreSt. James’s University HospitalLeeds, L59 7TF, UK

Renal cancer is the tenth most common cancer but its incidence is increasing, withthe 22% increase in rate in the last 10 years being the largest change for any cancerin females [1]. Surgery is the main therapy for organ-confined disease, however,over 50% of cases present with locally advanced or metastatic disease which isresistant to chemotherapy. Ovarian cancer is the fourth most common malignancyin females and due to the relatively asymptomatic progression, over 70% patients

Figure 1. 1D PAGE blot of 4 patient matched normal and malignant renal tissues probed with one of a

panel of lectins. Boxed region highlights specific differences between the matched normal and renal cancerffff

tissue samples.

113



114 R. E. Ferguson et al.

present with advanced disease [2]. In both cases there is a clear need for tumourmarkers to enable earlier diagnosis of disease as well as the identification of noveltherapeutic targets.

It is estimated that over 60% of all proteins, and virtually all secreted andtransmembrane proteins are glycosylated [3] and there is abundant evidence foralterations of glycosylation in cancer with glycan structures playing a role in cancercell homing and metastasis. Lectins are proteins which bind specific glycans and aneffective method to study cancer-associated glycosylation changes is by lectin-basedffffprofiling [4]. We have studied changes in glycosylation of serum proteins and normaland tumour tissue samples using lectin-based profiling, with the aim that such cancer-associated glycoforms will form the basis of either tumour marker assays or thera-peutic targets. 1D- and 2D-PAGE blots probed with a panel of lectins has demon-strated several differences in glycoprotein profile for both malignancies (see Fig. 1)ffffand has identified a potential novel ovarian cancer serum marker which is currentlyundergoing downstream validation to determine its diagnostic utility.

REFERENCES

1. Cancer statistics 2002 – Cancer Research UK.

2. Ries, L. A. G. et al. (2001) SEER Cancer Statistics Review 1973–1998, Section 20: Ovarian Cancer.

Bethesda: National Cancer Institute.

3. Apweiler, R. et al. (1999) Biochim Biophys Acta, 1473: 2 1–34.

4. Dwek M. V., et al. (2001) Proteomics, 1:756–62.

20

CARBOHYDRATES AND BIOLOGY OFSTAPHYLOCOCCAL INFECTIONS

Andrej Tarkowski1, Margareta Verdrengh1, Ing-Marie Jonsson1,Mattias Magnusson1, Simon J Foster2, and Zai-Quing Liu1

1Department of Rheumatology and Inflammation ResearchGoteborg University, Sweden2Department of Molecular Biology and BiotechnologyUniversity of Sheffield, UK

The genus Staphylococcus includes more than 30 species but only three are of majorclinical importance: S. aureus, S. epidermidis, and S. saprophyticus. S. aureus, thatthis review deals with, is a commensal found on the skin of 1/3 of the entire humanpopulation. However, in many healthy and diseased subjects S. aureus colonizationreaches much higher numbers. The increasing number of immunocompromisedsubjects on one hand and the appearence of methicillin-resistant staphylococci onthe other should increase the attempts to better understand the biology of host-bacterium interaction in staphylococcal diseases. In this abstract we focus on theimpact of certain carbohydrate structures within staphylococci on their virulenceand ability to cause inflammatory responses. Furthermore, we describe action ofcarbohydrates stemming from seaweed, being able to downregulate inflammatoryresponses caused by staphylococcal infection by interacting with rolling propertiesof endogenous leukocytes.

Carbohydrate constituents of the staphylococcal cell wall include polysaccharidemicrocapsule and peptidoglycans. We have shown that staphylococci, defective withrespect to the expression of the type 5 capsular polysaccharides triggered in murinerecipients lower frequency and severity of arthritis as well as infection related mortal-ity than the congeneic wild-type strain. Further in vitro studies suggested that thisoutcome was due to the enhanced phagocytosis and intracellular killing of bacteriathat lacked capsular polysaccharides. This finding was employed in a recent studywhere vaccination with staphylococcal polysaccharides (coupled to protein carrierto increase their immunogenicity) was used in a clinical setting. It was shown thathemodialysis patients receiving such a conjugate vaccine were partially protectedagainst staphylococcal bacteremia. Peptidoglycan component of the staphylococcal



116 A. Tarkowski et al.

cell wall proved to be an extremely potent pro-inflammatory substance, even ifadminstered alone in a single dose intra-articularly. Inflammation triggered by pepti-doglycans was mediated by combined action of lymphocytes and monocytes asproved in a recent study. Peptidoglycans may very well prove to be important causeof staphylococcal sepsis/septic shock along with actions of superantigens. Our recentand still unpublished data (Magnusson et al.) indicate that the peptidoglycan com-ponent that probably is responsable for its inflammatory action is Pam-Cys. Indeed,Pam-Cys alone is capable to trigger severe synovitis, when injected intra-articularly.This finding opens the perspective to interact with pro-inflammatory properties ofstaphylococcal peptidoglycans by silencing Toll Like receptor 2, mediating the actionof Pam-Cys.

The selectin family of adhesion molecules mediates the initial attachment ofleukocytes to venular endothelial cells at sites of tissue injury and inflammation. Forthis reason expression of selectins mediates inflammation also in case of staphylococ-cal arthritis. Fucoidin, a sulfated polysaccharide from seaweed, binds to and blocksthe function of L-and P-selectins thereby inhibiting leukocyte rolling and adhesionto endothelial surface. Treatment with fucoidin was used to assess whether signs ofearly joint inflammation triggered by invading staphylococci could be blocked.Indeed, severity of septic arthritis was significantly decreased within the first threedays following bacterial infection. Similar result was obtained using P-selectin defi-cient mice indicating that this selectin interaction is of crucial importance for extra-vasation of leukocytes during staphylococcal arthritis. Since P-selectin deficiencydecreases phagocytic activity of neutrophils it should be kept in mind that the effcientfffftreatment of staphylococcal arthritis should encompass combination of antibioticsand immunomodulation. Indeed, recent experimental and clinical studies indicatethat this principle is clearly superior to antibiotic treatment alone.

21

NEW DEVELOPMENTS IN TREATINGGLYCOSPHINGOLIPID STORAGE DISEASES

Frances M. Platt, Mylvaganam Jeyakumar, Ulrika Andersson,Raymond A. Dwek and Terry D. Butters

Department of BiochemistryUniversity of OxfordSouth Parks Road, Oxford OX1 3QU, UK

1. GLYCOSPHINGOLIPIDS

Eukaryotic cells have complex membranes that consist of a number of lipid andprotein species. The majority of the lipids present are phospholipids, with glyco-sphingolipids being present at much lower levels, representing only a few percent oftotal cellular lipid content. Glycosphingolipids are not required for membrane integ-rity (Ichikawa, Nakajo et al. 1994) but function in conjunction with cholesterol inthe formation of membrane microdomains or rafts that facilitate certain signallingevents in cells (Kobayashi and Hirabayashi 2000; Galbiati, Razani et al. 2001; Munro2003). They have been implicated to play a role in a number of biological processes(Bektas and Spiegel 2004; Zhang and Kiechle 2004) and are required for embryogen-esis (Yamashita, Wada et al. 1999). In this review we will focus on the sub-family ofglycosphingolipids that have glucosylceramide (Ichikawa and Hirabayashi 1998) astheir core structure, the glucosphingolipids (GSLs).

2. GSL BIOSYNTHESIS AND CATABOLISM

GSLs are synthesised within the Golgi apparatus by the sequential addition ofmonosaccharides to ceramide (Sandhoff and Kolter 2003). The first step in thispathway is the transfer of glucose to ceramide to generate glucosylceramide(Ichikawa, Sakiyama et al. 1996; Chujor, Feingold et al. 1998; Ichikawa, Ozawa et al.1998). This reaction is catalyzed by ceramide glucosyltransferase, an enzyme thathas its catalytic domain on the cytosolic side of an early Golgi compartment

117



118 F. M. Platt et al.

(Futerman and Pagano 1991; Trinchera, Fabbri et al. 1991). The GlcCer is theneither transported directly to other sites within the cell (Warnock, Lutz et al. 1994)or flipped into the Golgi lumen and further modified by the sequential addition ofmonosaccharides via the action of glycosyltransferases (Sandhoff and Kolter 2003).This further processing results in two families of GSLs, the neutral GSLs and thegangliosides (Maccioni, Daniotti et al. 1999). The gangliosides contain one or moresialic acid residue and are present at high levels in the CNS (Walkley, Zervaset al. 2000).

There are no human disease states associated with the early steps in GSLbiosynthesis, implying this pathway may be essential for embryonic development.Consistant with this, knocking out the ceramide glucosyltransferase in the mouse isembryonically lethal (Yamashita, Wada et al. 1999). However, there has been arecent report of a severe human infantile onset epilepsy syndrome in an Amishpedigree resulting from a proven defect in the GM3 synthase gene, preventing theformation of complex gangliosides (Simpson, Cross et al. 2004). This raises thepossibility that other human diseases resulting from defects in ganglioside biosyn-thesis exist, but have yet to be identified.

GSLs typically re-cycle from and to the cell surface via the Golgi apparatus(Pagano, Puri et al. 2000) but can also be routed to the lysosome for degradationwhere they are subjected to the sequential action of specific glycohydrolases (Sandhoffand Kolter 2003). Disease states are known to result from defects in many of thesteps in the GSL catabolic pathway. The substrate for the defective enzyme accumu-lates in the lysosome and leads to pathology. The resulting diseases are termed theGSL lysosomal storage disorders (Jeyakumar, Butters et al. 2002; Platt andWalkley 2004).

3. GSL LYSOSOMAL STORAGE DISEASES

The GSL storage diseases are individually rare but collectively affect approxi-ffffmately 1:18,000 live births (Meikle, Hopwood et al. 1999). They result from theinherited defects in the genes that encode the lysosomal hydrolases or their co-factorsrequired for GSL catabolism in the lysosome. The substrate(s) for the defectiveenzyme accumulates in the endosomal/lysosomal system leading to disease. The ageof onset of clinical signs is highly variable depending on how the specific mutationaffects enzyme function andffff /or stability (Wraith 2004). Relatively modest levels ofresidual enzyme activity can slow the rate of disease progression where as individualsalmost null for the enzyme activity in question will develop disease in utero or duringthe early post-natal period and have the most severely attenuated life span(Winchester 2004). The diseases fall into two main categories, those in which neutralGSLs are stored (Gaucher and Fabry disease) and those in which gangliosides arestored (the GM1 and GM2 gangliosidoses).

4. SECONDARY STORAGE OF GSLs

Storage of GSLs also occurs in a number of storage diseases in which theprimary defect is independent of GSL catabolism (Walkley 2004). The reason for

New Developments in Glycosphingolipid Storage Diseases 119

the accumulation of these secondary species is unknown, but there is emergingevidence implicating the secondary storage of GSLs contributing to the diseaseprocess (Zervas, Somers et al. 2001). Reducing GSL storage in these diseases maytherefore constitute a rational approach to their potential therapy (Platt andButters 2004).

5. THERAPEUTIC APPROACHES FOR GSL STORAGE DISEASES

As GSL storage diseases are the result of defects in the genes encoding thecatabolic enzymes of the lysosome, introducing the functional wild type enzyme or

gene should correct the problem. As the majority of GSL storage disorders involveGSL storage in the CNS, correction of disease in the brain is essential if these

therapies are going to deal with more than the visceral manifestation of these diseases(Schiffmann and Brady 2002).ffff

Currently, enzyme replacement therapy is in clinical use for the non-CNSdiseases, type 1 Gaucher and Fabry (Brady 2003; Neufeld 2004). However, wecurrently have no therapies for treating storage diseases of the brain. Gene therapy

remains elusive (Cabrera-Salazar, Novelli et al. 2002), although eventual efficacy ofthis approach in these diseases is anticipated as the hurdle to cross to achievetherapeutic benefit in these diseases is a very low one. Even a small increase inresidual enzyme activity can greatly reduce disease severity (Winchester 2004).

There are also cell-based therapies such as bone marrow transplantation (BMT)(Erikson, Groth et al. 1990; Ringden, Groth et al. 1995; Dobrenis 2004). The amountof brain reconstitution with BM-derived microglial cells is low and this limits thelevel of functional enzyme achievable in the CNS (Krivit, Sung et al. 1995). Therisks associated with the transplantation procedure and the need for HLA-matcheddonors limits the clinical application of BMT. When it has been used it generallytends to arrest the disease process but does not reverse pre-existing clinical symptoms.

New emerging experimental cell-based therapies involve the introduction of

neuronal stem cells directly into the brain to serve as a source of wild type enzymeand also to potentially replace dead or dying cell. This approach has the potentialto treat all neurodegenerative diseases, including the GSL storage diseases (Snyder,Daley et al. 2004).The other therapeutic option is to decrease the synthesis of the stored substrate

using small molecule enzyme inhibitors. This approach was suggested first by Radin(Inokuchi and Radin 1987; Radin 1996) and has been termed substrate reductiontherapy (SRT) (Platt and Butters 2004). As GSL species cannot be completelydegraded in the lysosome, as a result of the inherited enzyme deficiency, the biosyn-

thesis of fewer GSL molecules reduces the influx of GSLs into the lysosome allowingmore of the molecules synthesised by the cell to be catabolised. The aim is to balance

synthesis with the impaired rate of degradation. There are several advantages to thisapproach. These include use of an orally acting drug, use of a drug which penetratesthe CNS and by targeting an early step in the GSL biosynthetic pathway one drugcould potentially be used to treat multiple GSL storage diseases (Fig. 1).


6. SUBSTRATE REDUCTION THERAPY (SRT)

Small molecule inhibitors of the first step in GSL biosynthesis have beenidentified that have the potential to be used for SRT. To date there are two distinctchemical classes of these inhibitors that have been characterised, the PDMP seriesof compounds (Abe, Inokuchi et al. 1992) and the imino sugars (Butters, Dwek et al.2003; Butters, Mellor et al. 2003). In this article we will be focusing on the iminosugars as these are now in clinical use (Lachmann 2003).The imino sugars are stereochemical monosaccharide mimetics that have a ring

nitrogen atom instead of the oxygen (Winchester and Fleet 1992; Butters, van denBroek et al. 2000). The inhibition of ceramide glucosyltransferase by these iminosugars (Fig. 1) is critically dependant upon N-alkyation of the ring nitrogen with atleast a 4 carbon alkyl chain (Butters, Mellor et al. 2003). The prototypic compoundis N-butyldeoxynojirimycin (NB-DNJ) (Platt, Neises et al. 1994a). The galactoseanalogue N-butyldeoxygalactonojirimycin (NB-DGJ) also inhibits the ceramide glu-cosyltransferase (Platt, Neises et al. 1994b). The mechanism for their inhibitoryproperty is not fully understood but may in part be due to their structural similarityto ceramide (Butters, Mellor et al. 2003).

7. SRT IN THE SANDHOFF DISEASE MOUSE MODEL

Proof of principle studies using NB-DNJ were conducted in a mouse model ofSandhoff disease. The mouse lacked the b-subunit of b-hexosaminidase resulting inloss of the HexA (ab) and Hex B (bb) isoenzymes, with only very low level of residualenzyme activity conferred by Hex S (aa) (Sango, Yamanaka et al. 1995). GM2ganglioside storage results, along with storage of GA2 due to the action of a

Figure 1. Schematic representation of glycosphingolipid biosynthesis. The step in the pathway inhibited by

the imino sugars NB-DNJ and NB-DGJ is the ceramide glycosyltransferase catalysed biosynthesis of

glycosylceramide from ceramide.


lysosomal sialidase that converts GM2 to its asialo derivative GA2. This mousemodel has a clinical phenotype similar to human Tay-Sachs and Sandhoff disease.Sandhoff mice were fed on a diet containing NB-DNJ and were monitored for therate of disease progression, using behavioural tests that measure motor coordinationand muscle strength (Jeyakumar, Butters et al. 1999). It was found that the pre-symptomatic period was extended in response to SRT, the rate of clinical declinewas slowed and life expectancy increased by approximately forty percent. Thisdemonstrated that NB-DNJ mediated SRT could potentially be of therapeutic valuein managing storage diseases, including those involving pathology in the CNS.A clinical trial was initiated in type 1 Gaucher disease, a disorder with well-

defined clinical endpoints and no CNS involvement.

8. CLINICAL TRIALS OF NB-DNJNN

Type 1 Gaucher disease is a macrophage disorder characterised by hepatospleno-megaly, anaemia and bone disease (Beutler and Grabowski 2001). Patients wererecruited (Cambridge, Amsterdam, Prague and Jerusalem) into a 1-year open-labelclinical trial of NB-DNJ (also termed OGT-918) (Cox, Lachmann et al. 2000). Allpatients were unable or unwilling to receive ERT, the standard of care in this disease(Neufeld 2004). Liver and spleen volumes and haematological parameters weremeasured. Biochemical markers were also assessed including chitotriosidase (Aertsand Hollak 1997), cell surface leukocyte GM1 and plasma levels of GlcCer, thestorage lipid. Oral dosing was typically 100 mg OGT-918 three times daily.Pharmacokinetic profiling in a subgroup of the 28 patients showed that the

drug reached maximum plasma concentrations at 2.5 hours with a plasma half-lifeof 6.3 hours. Steady state concentrations of OGT-918 were achieved after 15 daysof dosing. The mean peak level of OGT-918 over the 12 month study was 6.8 mMwith trough values of 3.9 mM (Cox, Lachmann et al. 2000; Moyses 2003).

9. CLINICAL TRIALS

Spleen and liver volumes showed a significant reduction ((15%, 11.8–18.4,p<0.001) and (7%, 3.4–10.5, p<0.001) respectively) after six months of therapy. At12 months the decrease from base line was 19% (14.3–23.7, p<0.001) and 12%(7.8–16.4, p<0.001) respectively (Cox, Lachmann et al. 2000). This was comparableto the response observed in patients of the same disease severity at baseline receivingERT (Lachmann and Platt 2001). Chitotriosidase activity showed a time dependentreduction consistent with a reduction in the number of Gaucher cells. (Cox,Lachmann et al. 2000). Haemoglobin and platelet counts showed trends towardsimprovement, with a greater improvement in haemoglobin noted in patients whowere anaemic at baseline. A statistically significant improvement in platelet countswas achieved following 12 months of treatment.

Longer-term efficacy and safety were evaluated in patients that had completed12 months of therapy (Elstein, Hollak et al. 2004). Eighteen of the 22 patients thatwere eligible entered the extension phase and were followed for a further two years.


Continued and increasing efficacy was observed in all clinical parameters assessed.There was a marked improvement in platelet counts and haemoglobin levels relativeto the 12-month time point. No serious adverse events were reported and no furthercases of peripheral neuropathy emerged in the extension phase (two cases werereported and the patients withdrew in the initial 12 month study). GI tract sideeffects (due to intestinal sucraseffff /isomaltase inhibition by NB-DNJ) persisted in thesepatients, but to a lesser extent than in the first 12 months. Bone marrow fat fractionmeasurements were made in two patients and improvements were noted at 12months, with further improvement found at 36 months. This parameter reflects areduction in the number of Gaucher cells in the bone marrow in response to therapy.Consistent with this finding, chitotriosidase levels continued to decline in the exten-sion phase.

10. REGULATORY APPROVAL

In 2002 the European regulatory authority (EMEA) approvedNB-DNJ (miglus-tat, Zavesca≤) for the treatment of type 1 Gaucher disease (mild to moderate disease,unwilling or unable to receive ERT) (Cox, Aerts et al. 2003; Lachmann 2003). Inthe same year approval was granted in Israel. The FDA approved miglustat in theUSA in 2003, under the same label.

11. CURRENT CLINICAL TRIALS

Clinical trials with miglustat are currently in progress in type 3 Gaucher disease,late onset Tay-Sachs disease and in Niemann-Pick type C. Results from these trialsare anticipated in 2005.

12. SECOND-GENERATION COMPOUNDS

The galactose analogue NB-DGJ is equivalent to NB-DNJ in terms of potencyagainst the ceramide glucosyltransferase (Platt, Neises et al. 1994) but lacks manyof the additional enzyme inhibitory properties associated with NB-DNJ (Andersson,Butters et al. 2000). Significantly, it does not inhibit the gut disaccharidasessucrase/isomaltase, the property of NB-DNJ that causes osmotic diarrhoea(Andersson, Butters et al. 2000). Also, NB-DGJ does not cause weight loss. NB-DGJwas evaluated in the mouse model of Sandhoff disease and dose escalation waspossible with increasing benefit in terms of extended survival and improved function(Andersson, Smith et al. 2004). This compound is now in phase 1 in healthyvolunteers.

13. INFLAMMATION AS AN ADDITIONAL TARGET FORTREATING GSL STORAGE DISEASES

To date two main downstream consequences of storage have been reported thatare potentially amenable to pharmacological intervention, altered calcium homeost-asis (Ginzburg, Kacher et al. 2004) and macrophage/microglial cell mediated


inflammation in the brain (Wada, Tifftffff et al. 2000; Jeyakumar, Thomas et al. 2003;Wu and Proia 2004). The CNS inflammation has been extensively characterised andmouse models of the GM2 gangliosidoses (Tay-Sachs, Late Onset Tay-Sachs (LOTS),Sandhoff ) and GM1 gangliosidosis have been studied to determine whether there isa common neuro-inflammatory component to these disorders (Jeyakumar, Thomaset al. 2003). During the disease course, the expression of a number of inflammatorymarkers have been studied in the central nervous system (CNS), including MHCclass II, CD68, CD11b (CR3), 7/4, F4/80, nitrotyrosine, CD4 and CD8. Cytokineproduction was also profiled (TNFa, TGFb1 and IL1b) and blood-brain barrier(BBB) integrity determined. The kinetics of apoptosis and the expression of Fas andTNF-R1 were also assessed. In all symptomatic mouse models, a progressive increasein local microglial activation/expansion and infiltration of inflammatory cells wasnoted. Altered blood-brain barrier permeability was detected in Sandhoff and GM1mice, but not in the more mildly affected LOTS mice. Progressive CNS inflammationffffwas coincident with the onset of clinical signs in these mouse models. These datasuggested that inflammation might play an important role in the pathogenesis of thegangliosidoses (Jeyakumar, Thomas et al. 2003).

Recently, anti-inflammatory drugs and anti-oxidants (vitamin E and vitamin C)have been evaluated in the mouse model of Sandhoff disease and found to showefficacy as monotherapies and to synergise with SRT (Jeyakumar, Smith et al. 2004).As this approach utilises drugs already in common usage, the translation of thesefindings into clinical studies could potentially be quite rapid. It is likely that combina-tion therapy will provide the most benefit for the infantile onset GSL storage diseases,which remain the most challenging group of patients to treat, due to the lack ofsignificant levels of residual enzyme activity.

14. PERSPECTIVE

Despite the relative rarity of the GSL storage diseases there has been animpressive amount of progress made in both understanding the underlying thepathobiology of these diseases and in the development of multiple therapeuticapproaches of which BMT, ERT and SRT are currently in clinical use.

In common with the more common neurodegenerative diseases such asAlzheimer’s, the GSL storage diseases are neuro-inflammatory diseases with inflam-mation in the brain contributing to disease progression. How inflammation is trig-gered by the storage of GSLs is not yet known. However, the efficacy of non-steroidalanti-inflammatory drugs and anti-oxidant therapies in the Sandhoff mouse offers theffffprospect that targeting inflammation as an adjunctive therapy in these and relatedstorage disorders may be of clinical benefit. It is to be anticipated that over the nextfew years further new targets for storage disease therapy will emerge and be evaluatedin clinical studies either alone or in combination with other proven or experimentaltherapies.

REFERENCES

Abe, A., J. Inokuchi, et al. (1992). ‘‘Improved inhibitors of glucosylceramide synthase.’’ J Biochem Tokyo

111(2): 191–6.


Aerts, J. M. and C. E. Hollak (1997). ‘‘Plasma and metabolic abnormalities in Gaucher’s disease.’’ Baillieres

Clin Haematol 10(4): 691–709.

Andersson, U., T. D. Butters, et al. (2000). ‘‘N-butyldeoxygalactonojirimycin: a more selective inhibitor of

glycosphingolipid biosynthesis than N-butyldeoxynojirimycin, in vitro and in vivo.’’ Biochem

Pharmacol 59(7): 821–9.

Andersson, U., D. Smith, et al. (2004). ‘‘Improved outcome of N-butyldeoxygalactonojirimycin-mediated

substrate reduction therapy in a mouse model of Sandhoff disease.’’ Neurobiol Dis 16(3): 506–15.

Bektas, M. and S. Spiegel (2004). ‘‘Glycosphingolipids and cell death.’’ Glycoconj J 20(1): 39–47.

Beutler, E. and G. Grabowski (2001). Gaucher disease. The metabolic and molecular bases of inherited

diseases. C. R. Scriver, A. L. Beadet, D. Valle and W. S. Sly. New York, McGraw Hill. 3: 3636–3668.

Brady, R. O. (2003). ‘‘Enzyme replacement therapy: conception, chaos and culmination.’’ Philos Trans R

Soc Lond B Biol Sci 358(1433): 915–9.

Butters, T. D., R. A. Dwek, et al. (2003). ‘‘Therapeutic applications of imino sugars in lysosomal storage

disorders.’’ Curr Top Med Chem 3(5): 561–74.

Butters, T. D., H. R. Mellor, et al. (2003). ‘‘Small-molecule therapeutics for the treatment of glycolipid

lysosomal storage disorders.’’ Philos Trans R Soc Lond B Biol Sci 358(1433): 927–45.

Butters, T. D., L. A. G. M. van den Broek, et al. (2000). ‘‘Molecular requirements of imino sugars for the

selective control of N-linked glycosylation and glycospingolipid biosynthesis.’’ Tetrahedron

Assymetry 11: 113–124.

Cabrera-Salazar, M. A., E. Novelli, et al. (2002). ‘‘Gene therapy for the lysosomal storage disorders.’’ Curr

Opin Mol Ther 4(4): 349–58.

Chujor, C. S. N., K. R. Feingold, et al. (1998). ‘‘Glucosylceramide synthase activity in murine epidermis:

quantitation, localization, regulation, and requirement for barrier homeostasis.’’ J Lipid Res 39(2):

277–285.

Cox, T., R. Lachmann, et al. (2000). ‘‘Novel oral treatment of Gaucher’s disease with

N-butyldeoxynojirimycin (OGT 918) to decrease substrate biosynthesis.’’ Lancet 355(9214): 1481–5.

Cox, T. M., J. M. Aerts, et al. (2003). ‘‘The role of the iminosugar N-butyldeoxynojirimycin (miglustat) in

the management of type I (non-neuronopathic) Gaucher disease: a position statement.’’ J Inherit

Metab Dis 26(6): 513–26.

Dobrenis, K. (2004). Cell-mediated delivery systems. Lysosomal Disorders of the Brain. F. M. Platt and S.

U. Walkley. Oxford, Oxford University Press: 339–380.

Elstein, D., C. Hollak, et al. (2004). ‘‘Sustained therapeutic effects of oral miglustat (Zavesca,ffff

N-butyldeoxynojirimycin, OGT-918) in type 1 Gaucher disease.’’ J Inherit Metab Dis In Press.

Erikson, A., C. G. Groth, et al. (1990). ‘‘Clinical and biochemical outcome of marrow transplantation for

Gaucher disease of the Norrbottnian type.’’ Acta-Paediatr-Scand 79(6–7): 680–5.

Futerman, A. H. and R. E. Pagano (1991). ‘‘Determination of the intracellular sites and topology of

glucosylceramide synthesis in rat liver.’’ Biochem J 280(Pt 2): 295–302.

Galbiati, F., B. Razani, et al. (2001). ‘‘Emerging themes in lipid rafts and caveolae.’’ Cell 106(4): 403–11.

Ginzburg, L., Y. Kacher, et al. (2004). ‘‘The pathogenesis of glycosphingolipid storage disorders.’’ Semin

Cell Dev Biol 15(4): 417–31.

Ichikawa, S. and Y. Hirabayashi (1998). ‘‘Glucosylceramide synthase and glycosphingolipid synthesis.’’ Tr

Cell Biol 8(5): 198–202.

Ichikawa, S., N. Nakajo, et al. (1994). ‘‘A mouse B16 melanoma mutant deficient in glycolipids.’’ Proc Natl

Acad Sci U S A 91(7): 2703–7.

Ichikawa, S., K. Ozawa, et al. (1998). ‘‘Molecular cloning and characterization of the mouse ceramide

glucosyltransferase gene.’’ Biochem Biophys Res Commun 253(3): 707–11.

Ichikawa, S., H. Sakiyama, et al. (1996). ‘‘Expression cloning of a cDNA for human ceramide glucosyl-

transferase that catalyzes the first glycosylation step of glycosphingolipid synthesis.’’ Proc Natl Acad

Sci USA 93(10): 4638–43.

Inokuchi, J. and N. S. Radin (1987). ‘‘Preparation of the active isomer of 1-phenyl-2-decanoylamino-

3-morpholino-1-propanol, inhibitor of murine glucocerebroside synthetase.’’ J Lipid Res 28(5):

565–71.

Jeyakumar, M., T. D. Butters, et al. (1999). ‘‘Delayed symptom onset and increased life expectancy inSandhoff disease mice treated with N-butyldeoxynojirimycin.’’ Proc. Natl.Acad. Sci. U.S.A. 96(11):6388–6393.

Jeyakumar, M., T. D. Butters, et al. (2002). ‘‘Glycosphingolipid lysosomal storage diseases: therapy andpathogenesis.’’ Neuropathol Appl Neurobiol 28(5): 343–57.


Jeyakumar, M., D. Smith, et al. (2004). ‘‘NSAIDS incraese survival in the Sandhoff disease mouse: synergy

with NB-DNJ.’’ Annals of Neurology In Press.

Jeyakumar, M., R. Thomas, et al. (2003). ‘‘Central nervous system inflammation is a hallmark of pathogen-

esis in mouse models of GM1 and GM2 gangliosidosis.’’ Brain 126(Pt 4): 974–87.

Kobayashi, T. and Y. Hirabayashi (2000). ‘‘Lipid membrane domains in cell surface and vacuolar systems.’’

Glycoconj J 17(3 –4): 163–71.

Krivit, W., J. H. Sung, et al. (1995). ‘‘Microglia: the effector cell for reconstitution of the central nervousffff

system following bone marrow transplantation for lysosomal and peroxisomal storage diseases.’’

Cell-Transplant 4(4): 385–92.

Lachmann, R. H. (2003). ‘‘Miglustat. Oxford GlycoSciences/Actelion.’’ Curr Opin Investig Drugs 4(4):

472–9.

Lachmann, R. H. and F. M. Platt (2001). ‘‘Substrate reduction therapy for glycosphingolipid storage

disorders.’’ Exp. Opin. Invest. Drugs 10: 455–466.

Maccioni, H. J., J. L. Daniotti, et al. (1999). ‘‘Organization of ganglioside synthesis in the Golgi apparatus.’’

Biochim Biophys Acta 1437(2): 101–18.

Meikle, P. J., J. J. Hopwood, et al. (1999). ‘‘Prevalence of lysosomal storage disorders.’’ JAMA 281(3):

249–54.

Moyses, C. (2003). ‘‘Substrate reduction therapy: clinical evaluation in type 1 Gaucher disease.’’ Philos

Trans R Soc Lond B Biol Sci 358(1433): 955–60.

Munro, S. (2003). ‘‘Lipid rafts: elusive or illusive?’’ Cell 115(4): 377–88.

Neufeld, E. F. (2004). Enzyme replacement therapy. Lysosomal disorders of the brain. F. M. Platt and S.

U. Walkley. Oxford, Oxford University Press: 327–338.

Pagano, R. E., V. Puri, et al. (2000). ‘‘Membrane traffic in sphingolipid storage diseases.’’ Traffic 1(11):

807–15.

Platt, F. M. and T. D. Butters (2004). Inhibition of substrate synthesis: a pharmacological approach for

glycosphingolipid storage disease therapy. Lysosomal Disorders of the Brain. F. M. Platt and S. U.

Walkley. Oxford, Oxford University Press: 381–408.

Platt, F. M., G. R. Neises, et al. (1994a). ‘‘N-butyldeoxynojirimycin is a novel inhibitor of glycolipid

biosynthesis.’’ J Biol Chem 269(11): 8362–5.

Platt, F. M., G. R. Neises, et al. (1994b). ‘‘N-butyldeoxygalactonojirimycin inhibits glycolipid biosynthesis

but does not affectffff N-linked oligosaccharide processing.’’ J Biol Chem 269(43): 27108–14.

Platt, F. M. and S. U. Walkley (2004). Lysosomal defects and storage. Lysosomal Disorders of the Brain.

F. M. Platt and S. U. Walkley. Oxford, Oxford University Press: 32–49.

Radin, N. S. (1996). ‘‘Treatment of Gaucher disease with an enzyme inhibitor.’’ Glycoconj J 13(2): 153–7.

Ringden, O., C. G. Groth, et al. (1995). ‘‘Ten years’ experience of bone marrow transplantation for Gaucher

disease.’’ Transplantation 59(6): 864–70.

Sandhoff, K.ffff and T. Kolter (2003). ‘‘Biosynthesis and degradation of mammalian glycosphingolipids.’’

Philos Trans R Soc Lond B Biol Sci 358(1433): 847–61.

Sango, K., S. Yamanaka, et al. (1995). ‘‘Mouse models of Tay-Sachs and Sandhoff diseases differ inffff

neurologic phenotype and ganglioside metabolism.’’ Nat Genet 11(2): 170–6.

Schiffmann, R. and R. O. Brady (2002). ‘‘New prospects for the treatment of lysosomal storage diseases.’’ffff

Drugs 62(5): 733–42.

Simpson, M. A., H. Cross, et al. (2004). ‘‘Infantile-onset symptomatic epilepsy syndrome caused by a

homozygous loss-of-function mutation of GM3 synthase.’’ Nature Genetics In Press.

Snyder, E. Y., G. Q. Daley, et al. (2004). ‘‘Taking stock and planning for the next decade: realistic prospects

for stem cell therapies for the nervous system.’’ J Neurosci Res 76(2): 157–68.

Trinchera, M., M. Fabbri, et al. (1991). ‘‘Topography of glycosyltransferases involved in the initial glycosyl-

ations of gangliosides.’’ J-Biol-Chem 266(31): 20907–12 issn: 0021–9258.

Wada, R., C. J. Tifft,ffff et al. (2000). ‘‘Microglial activation precedes acute neurodegeneration in Sandhoffdisease and is supressed by bone marrow transplantation.’’ Proc Natl Acad Sci USA 97(20):

10954–10959.

Walkley, S. U. (2004). ‘‘Secondary accumulation of gangliosides in lysosomal storage disorders.’’ Semin

Cell Dev Biol 15(4): 433–44.

Walkley, S. U., M. Zervas, et al. (2000). ‘‘Gangliosides as modulators of dendritogenesis in normal andstorage disease-affected pyramidal neurons.’’ Cereb Cortex 10(10): 1028–37.ffff

Warnock, D. E., M. S. Lutz, et al. (1994). ‘‘Transport of newly synthesized glucosylceramide to the plasmamembrane by a non-Golgi pathway.’’ Proc Natl Acad Sci U S A 91(7): 2708–12.


Winchester, B. (2004). Primary defects in lysosomal enzymes. Lysosomal Disorders of the Brain. F. M.

Platt and S. U. Walkley, Oxford University Press: 81–130.

Winchester, B. and G. W. Fleet (1992). ‘‘Amino-sugar glycosidase inhibitors: versatile tools for glycobiolog-

ists.’’ Glycobiology 2(3): 199–210.

Wraith, J. E. (2004). Clinical aspects and diagnosis. Lysosomal disorders of the brain. F. M. Platt and S. U.

Walkley. Oxford, Oxford University Press: 50–77.

Wu, Y. P. and R. L. Proia (2004). ‘‘Deletion of macrophage-inflammatory protein 1 alpha retards neurodeg-

eneration in Sandhoff disease mice.’’ Proc Natl Acad Sci U S A 101(22): 8425–30.

Yamashita, T., R. Wada, et al. (1999). ‘‘A vital role for glycosphingolipid synthesis during development and

differentiation.’’ Proc. Natl. Acad. Sci. USA 96: 9142–9147.ffff

Zervas, M., K. L. Somers, et al. (2001). ‘‘Critical role for glycosphingolipids in Niemann-Pick disease type

C.’’ Curr Biol 11(16): 1283–7.

Zhang, X. and F. L. Kiechle (2004). ‘‘Review: Glycosphingolipids in health and disease.’’ Ann Clin Lab Sci

34(1): 3–13.

22

FUCOSYLATED GLYCANS IN INNATEAND ADAPTIVE IMMMUNITY

J. B. Lowe

Department of Pathology, the Life Sciences Institute, and theHoward Hughes Medical InstituteUniversity of MichiganAnn Arbor, MI, USA

Mammalian glycans are characterized by fucose modifications. Some of these areconstitutively expressed, whereas others are under temporal, developmental, andlineage-specific control. These modifications include fucose in alpha1,3-linkage toaspragine-linked, lipid-linked and serine/threonine-linked glycans. Others includefucose in direct linkage to serines and threonines of EGF-like domains, and throm-bospondin repeat domains, in several of proteins. To address the functions of thesefucose modifications in vivo, we have created and characterized mice with targetednull mutations in genes that control glycan and protein fucosylation. Mice withdeletions on a pair of alpha1,3fucosytransferases exhibit deficits in selectin-dependentneutrophil and T lymphocyte recruitment in acute inflammation, with an accompany-ing faulty innate immunity. Adaptive immunity in these mice is also faulty, due todefective L-selectin counter-receptor activity in HEVs of peripheral nodes, and acorresponding deficit in homing of naive T lymphocytes. In mice that are homozygousfor a null allele at the FX locus, GDP-fucose synthesis is conditionally active onlywhen fucose is supplied to the mice in their chow or water. In the absence ofexogenous fucose, these mice exhibit selectin counter-receptor defects akin to thoseobserved in the alpha1,3fucosytransferase deficient mice. Fucose-deficient FX nullmice also display numerous other abnormalities, including a reversible thymicatrophy phenotype that is apparently consequent to loss of Notch signaling activityin the pathway the controls thymocyte development from multipotent thymic progen-itors. These observations provide mechanistic insight into important functional rolesfor fucose-modifications in mammals.

127



128 J. B. Lowe

REFERENCES

1. Haltiwanger RS and Lowe JB. Role of glycosylation in development. Annu Rev Biochem. 73:491–537,

2004.

2. Haines N and Irvine KD. Glycosylation regulates Notch signalling. Nat Rev Mol Cell Biol. 4:786–797,

2003.

3. Smith PL, Myers JT, Rogers CE, Zhou L, Petryniak B, Becker DJ, Homeister JW, and Lowe JB.

Conditional control of selectin ligand expression and global fucosylation events in mice with a targeted

mutation at the FX locus. J Cell Biol. 158:801–815, 2002.

23

NEW INSIGHTS INTO RHEUMATOID ARTHRITISASSOCIATED GLYCOSYLATION CHANGES

Azita Alavi, Andrew J. Pool and John S. Axford

Biochemistry and ImmunologyAcademic Unit for Musculoskeletal DiseaseSt Georges Hospital Med SchoolCranmer Terrace, London, SW17 0RE, UK

1. INTRODUCTION

The link between RA, reduced b1,4-Galactosyltransferase (GTase) enzyme activ-ity (1–7), and immunoglobulin G (IgG) hypo-galactosylation is a well documentedphenomenon that may be linked to the pathology of rheumatoid arthritis (RA;3, 8–11).The IgG molecule represents one of the most powerful effector components offfff

the immune system and as such its structural diversification via changes in its,conserved, oligosaccharide constituents is of fundamental importance in our under-standing of pathological mechanisms associated with diseases such as RA.

Glycosylation changes have been shown to have profound effects on the stability,ffffconformation, antigenicity and, consequently, the overall function of IgG. It istherefore, not surprising that physiological IgG exists as a population of glycoforms,each conveying different physical and or biochemical properties that may result infffffunctional diversity.The concise informational package of the complex glycans and the significant

degree of heterogeneity that theses highly variable groups of branched ring structuresconfer to the protein backbone is governed by a finely tuned mechanism that relieson the action and interplay of a group of enzymes known as the glycosidases andglycosyltransferases.

1.1. b1,4-Galactosyltransferase Subfamily

b1,4-Galactosyltransferase is a subfamily of the glycosyltransferase super-family,which comprise of at least seven members (12–15). The most widely distributed,

129



130 A. Alavi et al.

principle, member of this family, and the most extensively studied in terms of bothRA and the control of galactosylation, is the classical GTase which has recently beendesignated as GTase-I (13, 15). The other, more recent, additions (GTase-II – GTase-VII) identified by primary sequence similarities (15), are differentially expressed,ffffoften present at comparatively low levels, and are still under investigation withregard to their biological significance, if any, in terms of the control of galactosylation(14, 16–18). GTase has a wide range of biological functions, and has been extensivelystudied in relation to RA and its role in the galactosylation of the terminal N-acetylglucosamine on the complex N-linked biantennary oligosaccharides located inthe CH2 domain of IgG (1–6, 12, 19). These sugars are an integral feature of IgGand are known to affect various, Fc-mediated, effff ffector functions (20).ffff

1.2. Reduced GTase Activity in RA

Reduced B cell GTase activity and the corresponding decreased IgG galactosyl-ation (IgG-G0) in RA, and various animal models (3, 5, 7–8, 10), appear to bedirectly linked to the pathogenic features associated with RA (9, 11). For example,agalactosylation of IgG can trigger the inappropriate activation of complement (21),is an important component of rheumatoid factor-IgG complexes in RA (22, 23) andhas been shown to be pathogenic in animal models of this disease (8, 11). In addition,IgG-G0 has been found to be a significant diagnostic and prognostic feature of RA,which together with rheumatoid factor status predicts a more severe disease (24–26).GTase, in particular GTase-I, has been shown to be the principal regulator of

IgG galactosylation as demonstrated by the specific alteration of b1-4GTase-I expres-sion in a human IgG-secreting cell line, via transfection with sense / antisense humanb1-4GTase-I cDNA (12). GTase galactosylation of IgG is a pre-secretory event thatoccurs, principally, within the trans-Golgi apparatus (27, 28). In addition to itsbiosynthetic function within the cell, GTase is also involved in a wide variety ofother complex biological processes, including cell-cell and cell-matrix interactions,and is therefore widely distributed and important in cell migration, matrix formationand signal transduction (28, 29). GTase is, therefore, present in the Golgi as well ason the cell surface and in a soluble form in various body fluids e.g. serum (30).Structurally, GTase consists of a short positively charged region attached to a

tightly folded globular (catalytic) domain which is attached to the membrane by aheavily glycosylated stem region (27, 28). The enzyme is synthesised in the roughendoplasmic reticulum and has a half-life of#20 hr, as a membrane bound glycopro-tein in the Golgi apparatus, after which it is released (via the proteolytic cleavage ofthe stem region) in a catalytically active soluble form (31). Serum GTase compositionis, therefore, reflective of GTase at the cellular level (27, 30, 31).

1.3. Regulation of GTase Activity in RA

Regulation of GTase activity is complex (32) and controlled in part by transcrip-tion, translation and post-synthetic modifications e.g. the degree of phosphoryl-ation (33).

However, to date, no evidence has been found to explain the reduction in GTase

Serum Galactosyltransferase Isoenzyme Changes In Rheumatoid Arthritis 131

activity in RA. Studies aimed at transcriptional and translational control have foundno evidence of reduced GTase mRNA expression (19) or reduced amount of GTaseprotein in RA B Cells (6). Furthermore, there is also no evidence of unique B cellpolymorphisms of the GTase gene or the gene controlling phosphorylation in RApatients (34), or any evidence for an intracellular inhibitor of GTase in RA (5).Together, these findings would suggest that the reduction in GTase activity in RA,is unlikely to be due to either a genetic abnormality or reduced expression of theenzyme, and points to the possibility of post-synthetic regulatory modifications.Potential mechanisms for post-translational regulation of GTase are evident

(35), and could give rise to various isoforms of GTase that may exhibit alteredkinetics (35, 36).In this respect it is interesting to note that a number of serum GTase isoforms

exist. At least 12 isoforms with broad and similar acceptor specificities have beendemonstrated in both health and disease (36–39). This heterogeneity may be due tothe fact that GTase is a sialoglycoprotein. It contains #10% carbohydrates, withone N- (40) and a variety of O- linked glycan chains (mucin-type sugar chains: ABO& Lewis blood group determinants) that are expressed in accord with the bloodgroup (41) and thus highly heterogeneous (42).

Sialic acid content, however, appears to be the principal determinant of theobserved charge heterogeneity of GTase isoforms, e.g. in cancer (38) and althoughits function is unknown, it may influence the enzymatic activity of GTase, which hasbeen shown to be influenced by its pI value (35–36).

The possibility that RA may be associated with the differential expression offfffGTase isoforms was suggested by pilot studies in which we have previously demon-strated (a) the differential incorporation of galactose onto a variety of diffff fferentffffacceptor molecules, and (b) significant changes in the gross isoelectric focusing (IEF)profiles (pH 3–10) of i) soluble serum and ii) peripheral B cell (CD19), GTase derivedfrom a group of patients with RA and healthy individuals (43).

1.4. IEF Profiling of Serum GTase

In a recent study (44) we extended our preliminary pilot IEF investigations andused a liquid phase IEF system to determine whether there were GTase isoformsspecific to, or elevated in RA, and investigated whether these changes could beattributed to possible sialylation differences.ffffFor this purpose we used serum samples from patients with 1) RA (n=9), 2) a

disease control group (n=9) and a group of healthy individuals (n=10).We used Rotofor IEF to separate serum GTase by charge and used a 2%

ampholyte solution; pH specificity of 4–6. The fractions were harvested (20 in total ),neutralised and assayed using a well characterised GTase assay that uses Uridinediphospho-d-[6-3H] galactose as the donor sugar and ovalbumin as the acceptormolecule. To standardise the data from each assay, the enzyme activity (cpm×103 )of each sample was expressed as a percentage of the sum of the activity of all thefractions (% Total GTase activity). In order to ascertain whether sialylation had arole to play, both serum and purified GTase (bovine) were desialylated withClostridium Perfringens sialidase and then assayed using an ampholyte with a pHrange of 5–7.

132 A. Alavi et al.

Figure 1. Percentage difference in the IEF profile of the RA group compared to the 2 control groups; DCffff

(shaded) & HI (black).

1.5. GTase Isoforms

GTase isoforms have previously been identified in both HI and malignancy,where a large degree of the observed charge heterogeneity has been attributed tochanges in the degree of sialylation (35, 36, 38). Our aim was to ascertain whetherthere were RA associated serum changes in the GTase isoform composition andwhether these changes could be attributed to possible sialylation differences. Wholeffffserum which represents secreted GTase from a number of different cell types, includ-ffffing lymphocytes, was used in preference to purified GTase, as the purificationprocedure may have selectively removed certain GTase isoforms in preference toothers (37).

Solution phase IEF demonstrated highly significant ( p<0.0001) differencesffffbetween the groups studied with the RA IEF profile, being significantly differentfffffrom that of the disease control group as well as the healthy controls (Fig. 1).Analysis of variance applied to each pair of groups demonstrated that the differenceffffin the shape of the IEF profile was highly significant for RA vs. DC or HI ( p<0.0001).There was no significant difference between the IEF profiles of the DC and HI groups.ffff

1.6. Differential Expression of GTase Isoforms in RAffff

The IEF differences were found to be predominantly the result of changes inffffthe pI of two distinct, albeit broad, peaks of GTase activity (Fig. 2) and pointtowards the possibility that the RA patients are synthesising larger quantities of therelatively more acidic GTase isoforms.


Figure 2. Representative IEF profiles: (a) a RA patient; (b) a disease control patient; and (c) a healthy

individual.

1.7. RA is Associated with a More Acidic GTase IEF Profile

Further analysis of the two main peaks of GTase activity revealed that the RAIEF differences were due to acidic shifts in the pI of the two main peak of activityffff(Fig. 3).

134 A. Alavi et al.

Figure 3. Percentage total GTase activity and pH value of (a) Peak 1 and (b) Peak 2 for RA patients

(black) and HI (white). Bars represent mean and standard error. The RA peak 2 was significantly more

acidic (p<0.01) and constituted a greater proportion of the total activity (p<0.01), when compared tothe HI peak.

Peak one

Although, the peak in the RA group was more acidic, there were no significantdifferences in the mean pI of this peak between the three groups examined 4.49ffff[range 4.30–4.65], 4.63 [range 4.38–4.86] and 4.60 [range 4.40–5.00] for the RA,DC and HI respectively (Fig. 3a). There were also, no significant differences in theffffactivity associated with this peak, which constituted 19.7% [range 17.2–26.0], 16.9%[range 12.1–22.8] and 17.7% [range 9.3–27.3], of the total GTase activity in theRA, DC and HI respectively.

Peak two

The RA peak was significantly more acidic when compared to the peak in theDC ( p<0.05) and the HI group ( p<0.01; Fig. 3b) and constituted a significantlygreater proportion of the total GTase activity (RA: 16.1% [range 13.3–18.8]), whencompared to the second peak in the DC (13.5% [range 10.2–18.1; p<0.05]) and HI(12.6% [range 8.9–17.7]; p<0.01) group.

1.8. Are Theses GTase Isoforms Unique to RA?

The acidic isoforms seen in RA are unlikely to be unique to this disease, as asmall percentage of both DC and HI show a second peak in a similar position. Thisis not surprising since a previous study examining GTase isoforms in HI andmalignancy, has shown that HI express most isoforms of the enzyme (30). The resultsof our studies and those of others would, in fact, indicate that the difference betweenffffdisease states lies in the relative quantities of each individual isoform.

On comparing the IEF profile of RA in relation to the disease control group,it is apparent that the association between the acidic GTase isoform and RA isespecially strong. Signifying that the RA associated changes in the GTase isoformprofile are specific and not due to inflammation per se as all the DC patientsinvestigated had active disease at the time of sampling.

Interestingly, when comparing serum RA GTase profiles with those found inmalignant disease, there appears to be a number of similarities, in particular, thepresence of a large and prominent peak at pH 4.75 and the absence of a peak


Figure 4. The effect of sialidase treatment on the IEF profile of serum GTase in 3 patients with RAffff

(inactive; shaded circle, active RA1; black squares, and active RA2; black triangle) and 2 HI (white

symbols): (a) Pre-desialylation and (b) Post-desialylation (bracket bars represent the range of activity for

each peak). Sialidase treatment resulted in similar IEF profiles of GTase in both RA and HI.

normally found at pH 5.10 (45), suggesting that isoforms with greater overall negativecharge predominate in both cancer and RA patients.

1.9. What is the Principal Cause of the Observed Charge Heterogeneity?

Sialidase treatment of sera from both RA patients and HI with markedlydiffering GTase isoform profiles, lead to three peaks of activity with comparable pHffffvalues (Fig. 4), which would suggest that the acid shift in the RA profile, is due inpart, to the presence of isoforms with higher sialic acid content.

The fact, that the differences in GTase isoform profiles, in both RA and malignantffffdisease (30, 36), appear to be primarily due to the presence of increased quantitiesof hypersialylated GTase, suggests that there may be a common mechanism by whichthis occurs. The most likely explanation is the high levels of sialyltransferase foundin both RA and malignancy (46, 47).

1.10. Do These Hypersialylated Isoform(s) Exhibit Altered Kinetics?

The enzymatic activity of purified GTase increased by an average of 75%following desialylation (Table 1).

The finding that GTase activity can be influenced by its degree of sialylationsupports earlier observations, demonstrating that the enzymatic activity of GTasemay, in part, be influenced by its pI value; isoforms with low pI values havingdecreased, and those with higher pI values having increased enzyme activity (36).

136 A. Alavi et al.

Table 1. The effect of desialylation on the activity of purified GTase. GTase activity is expressed as Meanffff

counts per minute (cpm) of duplicate assays

GTase Activity GTase Activity

Sialylated Desialylated

GTase (m Units) (cpm×103 ) (cpm×103) % Increase in Activity

3.10 7.94 14.38 81

6.25 16.12 25.79 60

12.50 33.97 63.24 86

25.00 79.15 138.46 75

50.00 190.61 326.30 71

CONCLUSION

In conclusion, the data from our recent studies demonstrate quantitative andqualitative changes in the RA serum GTase isoform profile. These changes appearto be, predominantly, due to increased synthesis of a greater proportion of hypersialy-lated isoforms, which have the potential to adversely affect the catalytic activity offfffthe enzyme, thus providing a possible mechanism for post-translational regulationof GTase activity in RA. It also provides further evidence that RA glycosylationchanges may be more general than previously indicated and encompass proteinsother than IgG.

REFERENCES

1. Axford JS, Mackenzie L, Lydyard PM, Hay FC, Isenberg DA, Roitt IM. Reduced B-cell galactosyl-

transferase activity in rheumatoid arthritis. Lancet 1987;2(8574):1486–8.

2. Furukawa K, Matsuta K, Takeuchi F, Kosuge E, Miyamoto T, Kobata A. Kinetic study of a galactos-

yltransferase in the B cells of patients with rheumatoid arthritis. Int Immunol 1990;2(1):105–12.

3. Axford JS, Sumar N, Alavi A, Isenberg DA, Young A, Bodman KB, et al. Changes in normal glycosyl-

ation mechanisms in autoimmune rheumatic disease. J Clin Invest 1992;89(3):1021–31.

4. Wilson IB, Platt FM, Isenberg DA, Rademacher TW. Aberrant control of galactosyltransferase in

peripheral B lymphocytes and Epstein-Barr virus transformed B lymphoblasts from patients with

rheumatoid arthritis. J Rheumatol 1993;20(8):1282–7.

5. Alavi A, Axford J. Evaluation of beta 1,4-galactosyltransferase in rheumatoid arthritis and its role in

the glycosylation network associated with this disease. Glycoconj J 1995;12(3):206–10.

6. Keusch J, Lydyard PM, Berger EG, Delves PJ. B lymphocyte galactosyltransferase protein levels in

normal individuals and in patients with rheumatoid arthritis. Glycoconj J 1998;15(11):1093–7.

7. Alavi A, Axford JS, Hay FC, Jones MG. Tissue-specific galactosyltransferase abnormalities in an

experimental model of rheumatoid arthritis. Ann Med Interne (Paris) 1998;149(5):251–60.

8. Rademacher TW, Williams P, Dwek RA. Agalactosyl glycoforms of IgG autoantibodies are patho-

genic. Proc Natl Acad Sci U S A 1994;91(13):6123–7.

9. Young A, Sumar N, Bodman K, Goyal S, Sinclair H, Roitt I, et al. Agalactosyl IgG: an aid to

differential diagnosis in early synovitis. Arthritis Rheum 1991;34(11):1425–9.ffff

10. Alavi A, Arden N, Spector TD, Axford JS. Immunoglobulin G glycosylation and clinical outcome in

rheumatoid arthritis during pregnancy. J Rheumatol 2000;27(6):1379–85.

11. Kuroda Y, Nakata M, Hirose S, Shirai T, Iwamoto M, Izui S, et al. Abnormal IgG galactosylation in

MRL-lpr/lpr mice: pathogenic role in the development of arthritis. Pathol Int 2001;51(12):909–15.

12. Keusch J, Lydyard PM, Delves PJ. The effect on IgG glycosylation of altering beta1, 4–ffff

galactosyltransferase-1 activity in B cells. Glycobiology 1998;8(12):1215–20.


13. Lo NW, Shaper JH, Pevsner J, Shaper NL. The expanding beta 4-galactosyltransferase gene family:

messages from the databanks. Glycobiology 1998;8(5):517–26.

14. Guo S, Sato T, Shirane K, Furukawa K. Galactosylation of N-linked oligosaccharides by human

beta-1,4-galactosyltransferases I, II, III, IV, V, and VI expressed in Sf-9 cells. Glycobiology

2001;11(10):813–20.

15. Amado M, Almeida R, Schwientek T, Clausen H. Identification and characterization of large galactos-

yltransferase gene families: galactosyltransferases for all functions. Biochim Biophys Acta

1999;1473(1):35–53.

16. Kido M, Asano M, Iwakura Y, Ichinose M, Miki K, Furukawa K. Normal levels of serum glycopro-

teins maintained in beta-1, 4-galactosyltransferase I-knockout mice. FEBS Lett 1999;464(1– 2):75–9.

17. Kotani N, Asano M, Iwakura Y, Takasaki S. Knockout of mouse beta 1,4-galactosyltransferase- 1 gene

results in a dramatic shift of outer chain moieties of N-glycans from type 2 to type 1 chains in hepatic

membrane and plasma glycoproteins. Biochem J 2001;357(Pt 3):827–34.

18. Kotani N, Asano M, Iwakura Y, Takasaki S. Impaired galactosylation of core 2 O-glycans in erythro-

cytes of beta1,4-galactosyltransferase knockout mice. Biochem Biophys Res Commun

1999;260(1):94–8.

19. Jeddi PA, Bodman-Smith KB, Lund T, Lydyard PM, Mengle-Gaw L, Isenberg DA, et al. Agalactosyl

IgG and beta-1,4-galactosyltransferase gene expression in rheumatoid arthritis patients and in the

arthritis-prone MRL lpr/lpr mouse. Immunology 1996;87(4):654–9.

20. Jefferis R, Lund J. Interaction sites on human IgG-Fc for FcgammaR: current models. Immunol Lettffff

2002;82(1–2):57–65.

21. Malhotra R, Wormald MR, Rudd PM, Fischer PB, Dwek RA, Sim RB. Glycosylation changes of IgG

associated with rheumatoid arthritis can activate complement via the mannose-binding protein. Nat

Med 1995;1(3):237–43.

22. Soltys AJ, Hay FC, Bond A, Axford JS, Jones MG, Randen I, et al. The binding of synovial tissue-

derived human monoclonal immunoglobulin M rheumatoid factor to immunoglobulin G preparations

of differing galactose content. Scand J Immunol 1994;4ffff 0(2):135–43.

23. Bond A, Kerr MA, Hay FC. Distinct oligosaccharide content of rheumatoid arthritis-derived immune

complexes. Arthritis Rheum 1995;38(6):744–9.

24. Isenberg DA, Lydyard PM, Axford JS, Hay FC, Rook GW, Roitt IM, et al. Galactosylation of IgG

associated oligosaccharides. Lancet 1988;2(8605):288.

25. van Zeben D, Rook GA, Hazes JM, Zwinderman AH, Zhang Y, Ghelani S, et al. Early agalactosylation

of IgG is associated with a more progressive disease course in patients with rheumatoid arthritis:

results of a follow-up study. Br J Rheumatol 1994;33(1):36–43.

26. Axford J, Gunnane G, FitzGerald O, Martin Bland J, Bresnihan B, Frears ER. Rheumatic disease

differentiation using immunoglobulin G sugar printing by high density electrophoresis. J Rheumatolffff

2003; 30(12):2540–6.

27. Alavi A. The Glycosyltransferases. In Abnormalities of IgG glycosylation and immunological dis-

orders. 1996:149–169.

28. Kleene R, Berger EG. The molecular and cell biology of glycosyltransferases. Biochim Biophys Acta

1993;1154(3–4):283–325.

29. Shur BD, Evans S, Lu Q. Cell surface galactosyltransferase: current issues. Glycoconj J

1998;15(6):537–48.

30. Davey RA, Harvie RM, Cahill EJ, Levi JA. Serum galactosyltransferase isoenzymes as markers for

solid tumours in humans. Eur J Cancer Clin Oncol 1984;20(1):75–9.

31. Strous GJ. Golgi and secreted galactosyltransferase. CRC Crit Rev Biochem 1986;21(2):119– 51.

32. Dinter A, Berger EG. The regulation of cell- and tissue-specific expression of glycans by glycosyltransf-

erases. Adv Exp Med Biol 1995;376:53–82.

33. Zhang SW, Xu SL, Cai MM, Yan J, Zhu XY, Hu Y, et al. Effect of p58GTA on beta-1,4– galactosyl-ffff

transferase 1 activity and cell-cycle in human hepatocarcinoma cells. Mol Cell Biochem

2001;221(1–2):161–8.

34. Delves PJ, Lund T, Axford JS, Alavi-Sadrieh A, Lydyard PM, MacKenzie L, et al. Polymorphism and

expression of the galactosyltransferase-associated protein kinase gene in normal individuals and galac-

tosylation-defective rheumatoid arthritis patients. Arthritis Rheum 1990;33(11):1655–64.

35. Furukawa K, Roth S. Embryonic and adult forms of two galactosyltransferases differ in their degreesffff

of sialylation. Eur J Biochem 1985;150(1):175–80.

138 A. Alavi et al.

36. Gerber AC, Kozdrowski I, Wyss SR, Berger EG. The charge heterogeneity of soluble human galactosyl-

transferases isolated frommilk, amniotic fluid and malignant ascites. Eur J Biochem 1979;93(3):453–60.

37. Davey R, Bowen R, Cahill J. The analysis of soluble galactosyltransferase isoenzyme patterns using

high resolution agarose isoelectricfocusing. Biochem Int 1983;6(5):643–51.

38. Davey R, Harvie R, Cahill J, Levi J. Serum galactosyltransferase isoenzyme patterns of cancer patients

with liver involvement. Br J Cancer 1986;53(2):211–5.

39. Uemura M, Sakaguchi T, Uejima T, Nozawa S, Narimatsu H. Mouse monoclonal antibodies which

recognize a human (beta 1–4)galactosyl-transferase associated with tumor in body fluids. Cancer Res

1992;52(22):6153–7.

40. Endo T, Amano J, Berger EG, Kobata A. Structure identification of the complex-type, asparagine-

linked sugar chains of beta-D-galactosyl-transferase purified from human milk. Carbohydr Res

1986;150:241–63.

41. Amano J, Straehl P, Berger EG, Kochibe N, Kobata A. Structures of mucin-type sugar chains of the

galactosyltransferase purified from human milk. Occurrence of the ABO and Lewis blood group

determinants. J Biol Chem 1991;266(18):11461–77.

42. Malissard M, Berger EG. Improving solubility of catalytic domain of human beta-1,4– galactosyl-

transferase 1 through rationally designed amino acid replacements. Eur J Biochem

2001;268(15):4352–8.

43. Soltys A, Alavi A, Dalziel M, Axford J. Galactosyltransferase isoenzymes in rheumatoid arthritis and

healthy individuals [abstract]. Glycosylation & Disease 1994;1:204.

44. Alavi A, Axford JS, Pool A. Serum galactosyltransferase isoform changes in rheumatoid arthritis. J

Rheumatol. 2004; 31(8):1513–20.

45. Qian GX, Liu CK, Waxman S. Abnormal isoelectric focusing patterns of serum galactosyltransferase

activity in patients with liver neoplasia. Proc Soc Exp Biol Med 1984;175(1):21–4.

46. Kessel D, Allen J. Elevated plasma sialyltransferase in the cancer patient. Cancer Res 1975;35(3):670–2.

47. Basset C, Durand V, Mimassi N, Pennec YL, Youinou P, Dueymes M. Enhanced sialyltransferase

activity in B lymphocytes from patients with primary Sjogren’s syndrome. Scand J Immunol

2000;51(3):307–11.

24

PRODUCTION OF COMPLEX HUMANGLYCOPROTEINS IN YEAST

Tillman Gerngross

Engineering and the Department of Biological SciencesDartmouth College, Hanover, USA

Recent advances in the Glycobiology field have helped to establish a relationshipbetween therapeutic protein function and glycosylation structures. Most of thesestudies rely on the comparison of mixed glycoforms, which complicate the clearinterpretation of distinct structure activity relationships. We describe the use ofcombinatorial genetic libraries to engineer yeast cells that perform entirely human-like glycosylation with exceptional fidelity and uniformity. The use of these librariesto elucidate structure function relationships of glycoproteins and the ability tomanufacture complex glycoproteins with unprecedented control over glycosylationwill be discussed.

139John S. Axford (ed.), Glycobiology and Medicine, 139.


25

RELATIONSHIPS BETWEEN THE N-GLYCANSTRUCTURES AND BIOLOGICAL ACTIVITIES OFRECOMBINANT HUMAN ERYTHROPOIETINSPRODUCED USING DIFFERENT CULTURECONDITIONS AND PURIFICATION PROCEDURES

C-T. Yuen1, P. L. Storring1, R. J. Tiplady1, M. Izquierdo2, R. Wait3,C. K. Gee1, P. Gerson1, P. Lloyd1, and J. A. Cremata2

1National Institute for Biological Standards and ControlPotters Bar, Herts., UK2Centre for Genetic Engineering and BiotechnologyHavana, Cuba3Kennedy Institute of RheumatologyHammersmith, London, UK

Eight preparations of recombinant human erythropoietin (rhEPO) with differingffffisoform compositions were produced by using different culture conditions and puri-fffffication procedures. The N-glycan structures of these rhEPOs were analyzed usingan HPLC based, with fluorescent detection profiling procedure (Yuen et al, 2002)and identified using matrix-assisted laser desorption ionization mass spectrometry.The specific activities of these rhEPOs were estimated by in vivo and in vitro mousebioassays.The eight rhEPOs were found to differ in their isoform compositions, as judgedffff

by isoelectric focusing, their N-glycan profiles, and in their in vivo and in vitrobioactivities. N-glycan analyses identified at least 23 different structures among theseffffrhEPOs, including bi-, tri- and tetra-antennary N-glycans, with or without fucosyl-ation or N-acetyllactosamine extensions, and sialylated to varying degrees. Massspectrometry also indicated the presence of N-glycans with incomplete outer chainsterminating in N-acetylglucosamine residues, and of molecular masses consistentwith phosphorylated or sulphated oligomannoside structures. The tetrasialylatedtetraantennary N-glycan contents of the eight rhEPOs were found to be significantlyand positively correlated with their specific activities as estimated by mouse in vivobioassay, and significantly and negatively correlated with their specific activities as



142 Yuen et al.

estimated by mouse in vitro bioassay. It is concluded that the tetrasialylated tetraan-tennary N-glycan content of rhEPO is a major determinant for its in vivo biologicalactivity in the mouse.

REFERENCE

Yuen, C-T., Gee, C.K. and Jones, C (2002). Biomedical Chromatography, 16, 247–254.

C-T.

26

GLYCOSYLATION OF NATURAL ANDRECOMBINANT ANTIBODY MOLECULES

Roy Jefferisffff

Immunity & InfectionUniversity of Birmingham, B15 2TT UK

Antibodies are often referred to as adaptor molecules that link humoral and cellularimmunity. Antibody/antigen interactions generate immune complexes that trigger acascade of inflammatory mechanisms resulting in the elimination of pathogens andresolution of infection. These inflammatory reactions are generally protective, how-ever, unless effectively regulated they can also cause ‘‘bystander’’ damage.ffffInflammatory cascades are triggered by interactions of complexed IgG-Fc with oneor more effector ligands, e.g. cellular receptors (Fcffff cRI, FccRIII, FccRIII), the C1qcomponent of complement, mannan binding lectin (MBL), the neonatal receptor(FcRn), the mannose receptor (MR) etc [1–4]. The inflammatory cascade can includeantibody-dependent-cellular-cytotoxicity (ADCC), complement-dependent-cellular-cytotoxicity (CDCC), phagocytosis, the oxidative burst, release of inflammatorymediators etc.The profile of inflammatory (protective) reactions mediated by the human

antibody isotypes has been determined through research programmes spanningdecades. However, these studies have been conducted, mostly, in vitro and employedheterologous systems, e.g. the use of guinea pig complement. It is not possible toreadily extrapolate from these studies to antibody-mediated mechanisms activatedin the intact animal since the response will be comprised of multiple antibodyisotypes, differing epitope specificities, affff ffinity etc. The advent of recombinant anti-body therapeutics challenges us to anticipate the effector functions optimal for affffdisease indication and to select the antibody isotype accordingly. Further, we maygenerate new antibody constructs optimized to activate a given inflammatory cas-cade(s). A majority of antibody drugs approved for clinical use have been based onthe human IgG molecule. The effector functions activated by chimeric or humanizedffffantibodies can be selected, in part, through the choice of the human IgG subclassemployed; it is unequivocally established that appropriate glycosylation of the IgG-Fcis essential.

143



144 R. Jefferisffff

Figure 1. The alpha backbone structure of human IgG showing functional regions – see text; light chain:

grey; heavy chains: black.

The basic structure of the IgG molecule is of two light and two heavy chainsin covalent and non-covalent association to form three independent protein moietiesconnected thorough a flexible linker or hinge region, Fig. 1. Two of these moietiesare of identical structure and each expresses an antigen specific binding site, the Fabregions; the third, the Fc, expresses interaction sites for ligands that activate clearancemechanisms. N-linked glycosylation of the IgG-Fc through Asn297 is a definingfeature and is essential to the expression of multiple effector activities [1–4]. Forffffpolyclonal human IgG ca. 10 – 20% of Fab moieties bear N-linked oligosaccharides.The glycosylation motifs are present in the variable regions of the light or heavychains and may be germline encoded or introduced through somatic mutation [4,5]. Effector mechanisms mediated through Fcffff cRI, FccRII, FccRIII, C1q, MBL andMR are ablated or severely compromised for aglycosylated IgG-Fc. The influenceof glycosylation on FcRn binding and activation appears to be minimal.

1. IgG-Fc GLYCOSYLATION

The oligosaccharide moiety is of the complex diantennary type and exhibitsheterogeneity with respect to terminal sugars, Fig. 2. The minimal structure observedfor normal human IgG is a heptasaccharide having terminal N-acetylglucosamineresidues (full ‘‘bond’’ lines). The possible addition of fucose, galactose, bisectingN-acetylglucosamine and sialic acid (dotted ‘‘bond’’ lines), generates the multipleglycoforms present in polyclonal and monoclonal IgG preparations.

There are unique features associated with IgG-Fc glycosylation. The site ofattachment, Asn 297, is proximal to the N-terminal region of the CH2 domain, andthe oligosaccharide ‘‘runs forward’’ with terminal sugar residues being exposed atthe CH2/CH3 domain interface. The oligosaccharide appears to be ‘‘enclosed’’within the protein structure and has defined secondary/tertiary structure [6]. This

Glycosylation of Natural and Recombinant Antibody Molecules 145

Figure 2. The structures of the possible diantennary oligosaccharide structure attached to IgG-Fc at

asparagine 297 (Asn297 ). The core ‘‘core’’ heptasaccharide present in normal human IgG is shown inupright type; this generates the G0 structure. Additional sugar residues that may be attached to the ‘‘core’’

are shown in italic type; thus G0F represents a fucosylated GO (G0F) oligosaccharide.

is due to multiple non-covalent interactions between the oligosaccharide and theprotein surface of the CH2 domains. In other glycoproteins oligosaccharides areattached to asparagine residues exposed on the surface of the protein, are highlymobile and interact with the aqueous phase.Analysis of neutral oligosaccharides released from normal polyclonal IgG reveals

twelve principle structures [7]. The possible addition of sialic acid provides for atotal of 21 structures. Since each heavy chain may bear a unique oligosaccharide atotal of>400 glycoforms can be anticipated. It remains to determine the significanceof this heterogeneity for polyclonal antibody populations and its impact on theactivation of inflammatory cascade(s) for antibodies to a given antigen (pathogen).

Insights into the significance of IgG antibody glycoforms for functional activityis being obtained from experiences with recombinant antibody molecules Since post-translational modifications of proteins are species, tissue and site specific the glycosyl-ation of recombinant antibody molecules varies between production vehicles, e.g.CHO, NSO cells etc. Analysis of currently licensed therapeutic antibodies, producedin CHO, NSO or Sp2/0 reveals simple glycoform profiles in which the G0F glycoformpredominates. The terminal sugar residue of G0F glycoforms is N-acetylglucosamine;there is evidence that immune complexes incorporating this glycoform may bind andactivate mannan-binding protein, with the activation of a pseudo-classical comple-ment pathway, and cellular mannose receptor, facilitating uptake by dendritic cells[8 – 10]. By contrast there is evidence that the fully galactosylated glycoform (G2F)facilitates placental transfer and complement activation [11, 12]. An additionalheterogeneity may be present in the form of the presence or absence of the C-terminallysine residue has been removed.

Cell engineering has been employed to generate glycoforms of IgG present innormal polyclonal IgG that are absent or present in small yield from CHO cells.

146 R. Jefferisffff

Transfection of CHO cells with the N-acetylglucosamine III transferase gene (GTIII)generated an anti-neuroblastoma IgG antibody with high level of bisectingN-acetylglucosamine that had 10–20 fold increased FccRIII mediated ADCC efficacy[13]; a similar improvement was reported for the therapeutic anti-CD20 antibody(Rituximab) [14]. Dramatic increases in FccRIII mediated ADCC efficacy havebeen reported for glycoforms of antibodies lacking core fucose [15–17]. Interestingly,the increased efficacy was restricted to FccRIII mediated activities. It has also beenshown that addition of sialic acid in an a(2�6) configuration, as opposed to a(2�3),results in improved FccRI and C1 activities [18]. The finding of increased FccRIIImediated ADCC for IgG antibodies: (i) with bisecting N-acetylglucosamine and (ii)the absence of fucose suggests that the most efficient glycoform would be an antibodycombining both these structural attributes. It is interesting to note that such glyco-forms are essentially absent from normal polyclonal human IgG.

The contribution of oligosaccharides to structure and function of IgG has beenprobed for a series of homogeneous antibody glycoforms and revealed that a simpletrisaccharide may confer both stability and function to the IgG-Fc [19 – 21]. Thesestudies show that whilst the influence on function is considerable it results from verysubtle conformational differences. It was concluded that the lower hinge region offfffthe IgG molecule exhibits ‘‘plasticity’’ and may exist as an equilibrium of conformersthat allows for multiple ligand binding specificities. Functional activities and glyco-form profiles may also be manipulated by alanine replacement of amino acid residuesthat make non-covalent interactions with the oligosaccharide [22].

Interaction sites for the FccRI, FccRII, FccRIII and C1q ligands have been‘‘mapped’’ to the amino acid residues in the lower hinge region and the hingeproximal region of the CH2 domain. Whilst the isolated IgG-Fc fragment expressesFccRI, FccRII, FccRIII and C1q binding properties x-ray crystallographic analysisshows this hinge proximal region of protein structure to be disordered [6]. Previouslypredicted interaction sites for FccRIII have been validated through x-ray crystallo-graphic analysis of a complex of IgG1-Fc and a soluble form of FccRIII. It showsasymmetric binding of the receptor to each of the lower hinge heavy chain regionssuch that they become ordered. Interestingly, there is no significant contact with theoligosaccharides although their presence is essential to the formation of complex[23, 24].

2. IgG-Fab GLYCOSYLATION

It is established that 15 – 20% of polyclonal human IgG molecules bear N-linkedoligosaccharides within the IgG-Fab region, in addition to the conserved glycosyl-ation site at Asn 297 in the IgG-Fc (4, 25). There are no consensus sequences forN-linked oligosaccharide within the constant domains of either the kappa or lambdalight chains or the CH1 domain of heavy chains, therefore, when present they areattached in the variable regions of the kappa (Vk), lambda (Vl) or heavy (VH )chains; sometimes both. In the immunoglobulin sequence database ~20% of IgGV regions have N-linked glycosylation consensus sequences (Asn-X-Thr/Ser; whereX can be any amino acid except proline). Interestingly, these consensus sequences

Glycosylation of Natural and Recombinant Antibody Molecules 147

are mostly not germline encoded but result from somatic mutation – suggestingpositive selection for improved antigen binding (4, 5).A monoclonal human IgG, isolated the serum of patient with multiple myeloma,

was shown to bear oligosaccharide in the VL region, in addition to the normalIgG-Fc. The IgG-Fc oligosaccharide could be removed with PNGase F but the VLoligosaccharide was resistant; in contrast the VL oligosaccharide but not the IgG-Fcoligosaccharide was removed on exposure to endoglycosidase F. The therapeuticErbitux (Cetuximab) also exhibits IgG-Fab glycosylation, at Asn 88 in the VH region.Only the IgG-Fc oligosaccharide could be removed with PNGase F. In each caseoligosaccharide analysis of the IgG-Fc and IgG-Fab shows that whilst the IgG-Fcbears mainly G0F oligosaccharides the IgG-Fab bears complex diantennary sialyl-ated structures. Thus, there is distinct site specificity to the glycoforms generated.

Hypogalactosylation of polyclonal IgG-Fc has been reported for a number ofinflammatory autoimmune diseases. An extreme example is in Wegener’s granulom-atosis and microscopic polyangiitis [26]. Oligosaccharide analysis of the IgG-Fcand IgG-Fab shows that whilst the IgG-Fc bears mainly G0F oligosaccharides theIgG-Fab bears complex diantennary sialylated structures. This suggests that whilstthe galactosylation machinery is intact environmental factors are exerting an influ-ence on IgG-Fc galactosylation. The functional significance for IgG-Fab glycosyl-ation of polyclonal IgG has not been fully evaluated but data emerging formonoclonal antibodies suggests a positive, neutral or negative influence on antigenbinding.

These studies demonstrate that the micro-environment, in vivo, can have aprofound influence on the glycosylation profile of the IgG-Fc. This may reflect theunique structural relationship between the oligosaccharide and the protein. The‘‘core’’ heptasaccharide is essential for FccRI, FccRII, FccRIII and C1 activationwhilst outer arm sugar residues can influence these and other functions, e.g. FccRIII,FcRn, MBL, MR. Thus, fidelity of glycosylation is essential to the effector functionffffprofile of antibodies. However, the oligosaccharide can function as a structural‘‘rheostat’’ to generate specific glycoforms exhibiting optimal effector activities for affffparticular disease target.

REFERENCES

1. Jefferis, R., Lund, J. & Pound, J. (1998) IgG-Fc mediated effff ffector functions: molecular definition offfff

interaction sites for effector ligands and the role of glycosylation. Immunol.Rev. 163:50–76.ffff

2. Jefferis, R. (2001) Glycosylation of human IgG antibodies: relevance to therapeutic applications.ffff

Biopharm. 14:19–26.

3. Jefferis, R. and Lund, J. (2002) Interaction sites on human IgG-Fc for Fcffff cR: current models.

Immunol.Lett. 82:57–65.

4. Jefferis, R. (2005) Glycosylation of Recombinant Antibody Therapeutics. Biotechnology Progress.ffff

In press.

5. Dunn-Walters D.; Boursier L.; Spencer J.(2000) Effect of somatic hypermutation on potentialffff

N-glycosylation sites in human immunoglobulin heavy chain variable regions. Molecular Immunology.

37:107–13.

6. Deisenhofer, J., (1981) Crystallographic refinement and atomic models of a human Fc fragment and its

complex with fragment B of protein A from Staphylococcus aureus at 2.9- and 2.8-A resolution.

Biochemistry 20:2361–2370.

148 R. Jefferisffff

7. Jefferis, R, Lund, J, Mizitani, H, Nakagawa, H, Kawazoe, Y, Arata, Y & Takahashi, N. (1990) Affff

comparative study of the N-linked Oligosaccharide structures of human IgG subclass proteins.

Biochem. J. 268, 529–537.

8. Malhotra. R.; Wormald, M. R.; Rudd, P. M.; Fischer, P. B.; Dwek, R. A.; Sim R. B. (1995) Glycosylation

changes of IgG associated with rheumatoid arthritis can activate complement via the mannose-binding

protein. Nature Medicine. 1, 237–243.

9. Abadeh, S., Church, S., Dong, S., Lund, J., Goodall, M. and Jefferis R. (1997) Remodelling theffff

oligosaccharide of human IgG antibodies: effects on biological activities. Biochem.Soc.Trans. 25:S661ffff

10. Dong, X.; Storkus, W. J.; Salter, R. D. (1999) Binding and uptake of agalactosyl IgG by mannose

receptor on macrophages and dendritic cells. J.Immunol. 163, 5427–5434.

11. http://www.fda.gov/cder/biologics/review/ritugen112697-r2.pdf

12. Kibe, T., Fujimoto, S., Ishida, C., Togari, H., Okada, S., Nalagawa, H., Tsukamoto, Y Takahashi, N.

(1996) Glycosylation and placental transport of IgG. J.Clin.Biochem.Nutrition. 21:57–63.

13. Umana, P, Jean-Mairet J, Moudry R, Amstutz H, Bailey JE. (1998) Engineered glycoforms of an anti-

neuroblastoma IgG1 with optimized antibody-dependent cellular cytotoxic activity. Nat Biotechnol.

17:176–80.

14. Davies, J., Jiang, L., LaBarre, MJ., Anderson, D., Reff, M. (ffff 2001) Expression of GTIII in a recombinant

anti-CD20 CHO production cell line: Expression of antibodies of altered glycoforms leads to an

increase in ADCC thro’ higher affinity for FcRIII. Biotech.Bioeng. 74:288–294.

15. Shields RL, Lai J, Keck R, O’Connell LY, Hong K, Meng YG, Weikert SH, Presta LG. (2002) Lack of

Fucose on Human IgG1 N-Linked Oligosaccharide Improves Binding to Human Fcgamma RIII and

Antibody-dependent Cellular Toxicity. J. Biol Chem. 277:26733–40.

16. Shinkawa T, Nakamura K, Yamane N, Shoji-Hosaka E, Kanda Y, Sakurada M, Uchida K, Anazawa

H, Satoh M, Yamasaki M, Hanai N, Shitara K. (2003) The absence of fucose but not the presence of

galactose or bisecting N-acetylglucosamine of human IgG1 complex-type oligosaccharides shows the

critical role of enhancing antibody-dependent cellular cytotoxicity. J Biol Chem. 278:3466–73.

17. Okazaki A, Shoji-Hosaka E, Nakamura K, Wakitani M, Uchida K, Kakita S, Tsumoto K, Kumagai I,

Shitara K. (2004) Fucose depletion from human IgG1 oligosaccharide enhances binding enthalpy and

association rate between IgG1 and FcgammaRIIIa. J Mol Biol. 336:1239–49.

18. Jassal, R., Jenkins, N., Charlwood, J., Camilleri, P., Jefferis, R. and Lund, J. (2001) Sialylation of humanffff

IgG-Fc carbohydrate by transfected rat a(2 – 6) sialyltransferase. Biochem. Biophys.Res.Comm.

286:243–249.

19. Mimura Y., Church S., Ghirlando R,. Dong S., Goodall M., Lund J. and Jefferis R. (2000) The influenceffff

of glycosylation on the thermal stability and effector function expression of human IgG1-Fc: propertiesffff

of a series of truncated glycoforms. Molecular Immunology. 37:697 – 706.

20. Mimura, Y., Sondermann, P., Ghirlando, R., Lund, J., Young, S.P., Goodall, M. and Jefferis, R. (2001).ffff

The role of oligosaccharide residues of IgG1-Fc in FccIIb binding. J.Biol.Chem. 276:45539–45547.

21. Krapp, S., Mimura, Y., Jefferis, R., Huber, R. and Sondermann, P. (2003) Structural analysis of humanffff

IgG glycoforms reveals a correlation between oligosaccharide content, structural integrity and Fcc-

receptor affinity. J.Mol.Biol. 325:979–989.

22. Lund, J., Winter, G., Jones, P.T., Pound, J., Tanaka, T., Walker, M.R., Artymiuk, P.J., Arata, Y., Burton,

D.R., Jefferis, R. and Woof, J.M. (1991) Human Fcffff cRI and FccRII interact with distinct but overlap-

ping sites on human IgG. J.Immunol. 147 2657 – 2662.

23. Sondermann P. Huber R. Oosthuizen V. Jacob U. (2000). The 3.2-A crystal structure of the human

IgG1 Fc fragment-Fc gammaRIII complex. Nature. 406(6793):267–273.

24. Radaev S, Motyka S, Fridman WH, Sautes-Fridman C, Sun PD. (2001) The structure of a human type

III Fcgamma receptor in complex with Fc. J Biol Chem. 276:16469–77.

25. Youings, A.; Chang, S. C.; Dwek, R. A.; Scragg, I. G. (1996) Site-specific glycosylation of human

immunoglobulin G is altered in four rheumatoid arthritis patients. Biochem J. 314, 621–30.

26. Holland, M., Takada, K., Okomoto, T., Takahashi, N., Kato, K., Adu, D., Ben-Smith, A., Harper, L.,

Savage, C.O.S. and Jefferis, R. (2002) Hypogalactosylation of serum IgG in patients with ANCA-ffff

associated systemic vasculitis. Clin.exp.Immunol. 129:183–190.

INDEX

a-dystroglycan 97 bovine viral diarrhoea virus (BVDV) vantiviral molecules interfering withadaptive immune system vi

fucosylated glycans 127–8 N-glycosylation, model forstudy 5–6alanine scanning mutagenesis 8

allergic alveolitis see pigeon fanciers’ lung model organism for HCV 1bowels, bacteria, normal/abnormal viianti-citrulinated cyclic peptide vii

anti-inflammatory drugs 123C1-complex 37–8anti-oxidants 123C1-inhibitor 37antibiotic treatment, staphylococcalC1g component of complement 143infection 116C-type lectin superfamily (CLSF) 23–4antibody molecules (natural andCaenorhabditis elegans 11recombinant), glycosylation of 143–7calnexin, pathway 1antibody-dependent-cellular-cytotoxicitycalreticulin-bound murine leukaemia virus(ADCC) 29, 143

glycoprotein, gp90,antiviral therapies 1immunogenecity 85–92APLEC 23–4

cancer, associated glycosylation 113–14Arabidopsis thaliana 11carbohydrate epitope, antibody recognition,Asn-100 61

HIV vaccine design 7–8Asn-101 61carbohydrate recognition domainAsn-120 62, 63(fig.)

(CRD) 35Asn-25 62, 63(fig.)cartilage breakdown, gelatinase B 48–51autoimmune arthritis viCD44 58–63autoimmune rheumatoid arthritis seeHA-binding surface, analysis 59–61rheumatoid arthritisN-glycosylation, regulation by 61–3structure 59(fig.)bacterial colonisation vii

b1,4-Galactosyltransferase (GTas) CD8 glycoprotein, glycosylation 71–81CD8 b subunit, function 74–6rheumatoid arthritis 129–36

activity, reduced 130 CD8 stalk, influence of O-glycosylationupon the extention 77–8activity, regulation 130–1

IEF profiling 131–2 CD8/MHC interaction, O-linkedglycosylation, modulated by 76–7isoforms 132–5

subfamily 129–30 class I MHC, interactions with 72–4structure 72, 73(fig.)biological therapies viii

Biomphalaria 10 cell-based therapies 119cellular glycoengineering viiiBip 92

bone marrow transplantation 119 cellular receptors (Fc c RI,Fc c RII,Fc cRIII) 143bovine spongiform encephalopathy in

cattle 95 ceramide 120(fig.)

149

150 Index

ceramide glycosyltransferase 117–18 2-G12 monoclonal antibody v, 2, 8cercariae 10 galactose-binding lectins 13Cetuximab 147 galactosylation, abnormal viichondroitin 57 galectin-3 108CLIP (class II invariant chain-derived galectins 13

peptide) 85, 87 ganglioside biosynthesis, defects 118CNX 91 Gaucher disease 2, 118collagen type-II vi–vii type-1 viirheumatoid arthritis 45–53 gelatin 49posttranslocational gelatin zymography 47, 48(fig.)

modifications 51–2 gelatinase A 45collagen type-V 47 gelatinase B vicollagenase-3 (MMP-13) 47 differential glycosylation, from neutrophilsffffcollectins 21–2, 35 and breast cancer cells 103–10complement classical pathway rheumatoid arthritis, role in 45–53

activation 27, 29 cartilage breakdown 48–51complement system 37(fig.) gelatinase B/matrix metalloproteinase-9 viinnate immune response 37–8 Gerstmann–Straussler–Scheinkercomplement-dependent-cellular cytotoxicity disease 95

(CDCC) 143 glucosidase inhibitors 1–2congenital muscular dystrophy (CMD) 97 glucosphingolipid (GSL) storage disease 2type 1C (MDC1C) 97 glycan(s)‘‘core’’ heptasaccharide viii antigens 10, 11–12(figs)Creutzfeldt–Jakob disease 95 structures viCRT 91 glycosaminoglycans 57

glycosphingolipid storage diseases viiDC-SIGN vi, 14–16glycosphingolipid(s) 117–18Dectin-1 23biosynthesis and catabolism 117–18dendritic cellslysosomal storage disease 118parasite-derived glycans,secondary storage 118–19reconditionglycosphingolipid(s) storage disease, newTh1/Th2 responses 13–14

developments 117–23deoxygalactonojirimycin (DGJ) 1, 5–6inflammation as an additionaldeoxynojirimycin (BuDNJ) 1

target 122–3Ebola virus v NB-DNJ clinical trials 121–2Edman degradation 49, 51(fig.) substrate reduction therapy 120–1enteric microflora vii therapeutic approaches 119, 120(fig.)enzyme replacement therapy 119 glycosylation, disease targets andErbitux 147 therapy 1–2Ero-1 91 glycosylceramide 117, 120(fig.)ERp57 91 glycosyltransferases (GT)ERp72 91 GM1 gangliosidosis 2eukaryotic cells 117 GM3 synthase gene, defect 118

gp90, immunogenicity 85–92Fabry disease 2, 118factor VIII (plasma transglutaminase) 37

heparin 57fibrinogen 37hepatitis C virus (HCV) vficolins 21–2disease targets and therapies 1–2fucoidin vii, 116P7 ion channel, long alkylchainfucosylated glycans, innate and adaptive

iminosugars, blocking role 3–4immunity 127–8fukutin-related protein (FKRP) 97 HIV vaccine design (a template), antibody

12–13

10–11, 97

Index 151

recognition of a carbohydrate L-selectin viiepitope 7–8 L-SIGN vi, 14–16

human immunodeficiency virus type-1 lactoceramide 120(fig.)(HIV-1) v laminin, hypo-glycosylation 97

gp120 antigenic surface, Late Onset Tay–Sachs disease 123glycosylated 7(fig.) LDN 11–12(figs)

humoral immune response to infection 2 LDNF 11–12(figs)human immunoglobulin G (IgG) 143 lectin blotting, glycosylation changes inalpha backbone structure 144(fig.) serum and tissue proteins inhuman immunoglobulin glycosylation, lectin cancer 113–14

pathway of complement lectin complement pathway 22activation 27–40 mannose binding lectin (MBL) 35–6

hyaladherins 58 MBL-associated serine proteaseshyaloronan 57–66 (MASPs) 36–8structure 57(fig.) lectinshypo-glycosylation 97 C-type 13

cell-associated viiminosugar(s) 120Lewis-xderivatives 1binding to the CRD of DC-SIGN 14–16long alkylchain, HCVp7 ion channel

Lex 3blocking 3–4antibodies, autoimmune reactions 12–13morphogenesis inhibitors, viral re-entrylimb girdle muscular dystrophy 2Iprevention 5–6

(LGMD2I) 97immunoglobulin A (IgA) 27Link module superfamily 58, 60–1glycosylation 32–4lipoarabinomannan 37immunoglobulin D (IgD) 27lipopolysaccharide 37glycosylation 34–5lipoproteins 37immunoglobulin E (IgE) 27Ly49 (Klra) 23–4glycosylation 35

immunoglobulin G (IgG) 27glycosylation 30immunoglobulin G macrophage galactose-type lectin

(IgG)-Fab glycosylation 146–7 (MGL) viimmunoglobulin G (IgG)-Fc major histocompatibility complex (MHC)

glycosylation 144–6 class II molecules 85–9immunoglobulin M (IgM) 27 antigen presentation by two distinctglycosylation 30–2 populations 88–9immunoglobulins 27–9 overview and presentationmannose binding lectin (MBL), pathway 86(figs)

interactions with 38–40 presentation, influence of antigenrole 27 conformation 87–8solubility 29 class II-restricted tumour antigen 90–2structure 28(fig.) processing, role of glycosylation insynthesis 27–9

influencing 91–2immunomodulation 116

guided processing andinnate immune system vi

immunodominance 89complement system 37–8

Man5-GlcNAc2 sugars 25fucosylated glycans 127–8

mannan binding lectin (MBL) 3, 29, 143interferon a 5

pathway of complement activation,interleukin-1 45

biochemistry, biology and clinicalinterstitial collagenase (MMP-1) 47

implications 21–2isoelectric focusing (IEF) 131

mannan/mannose Binding protein 35structure 36(fig.)killer cell lectin-like receptors 23–4

12–13

13

152 Index

mannose binding lectin (MBL) O-glycans 76–81analysis 79(fig.)immunoglobulins 38–40

lectin pathway of complement O-linked glycosylation vO-linked glycosylationactivation 35–8

structure 36(fig.) CD8 stalk extension, influenceupon 77–8mannose receptor (MR) 38, 143

mass spectrometry 49, 51(fig.) CD8/MHC interaction 76–7oligosaccharide viii, 37mast cell, IgE 29

matrilysins 47 Oncomelania 10ovarian cancer vimatrix metalloproteinases (MMP) 45,

103–5 asymptomatic progression, need fortumour markers 114structure 46(fig.)

MBL-associated serine proteases oxidative burst 143oxidoreductases 91(MASPs) 21–2, 36–7

lectin pathway of complementactivation 36–8 p7 protein 3–4

MCF-7 gelatinase B, production and ion channels formation/activity 3–4glycosylation analysis 105–10 P-selectin vii

methicillin-resistant staphylococci 115 1D/2D-PAGE, glycoprotein profile,minacidia 10 malignancies 113–14mucin oligosaccharides 101–2 Pam-Cys 116murine mannose receptor (MR), ligand pathogen-associated molecular patterns

binding, glycosylation influence 25–6 (PAMPs) 21–2, 37–8PPmuscular dystrophy vi PDI 91glycosylation defects 97–8 peptidoglycan(s) 37

staphylococcal cell wall 115–16N7-oxanonyl-6deoxy-DGJ 1

phagocytosis 143N-butyldeoxynojirimycin (NB-DNJ) vii, 2,

phorbol-myristate-acetate (PMA) 105–6120–1

pigeon fanciers’ lung viclinical trials 121–2

mucin oligosaccharides 101–2N-glycans v

pigeon intestinal mucin (PIM) 101oocyte development and function, role

polysaccharide skin inflammation viiin 99–100

posttranslocational modifications, humanN-linked glycosylation v

collagen-II 51–2CD44 function, regulation 61–3

Praziquantel 9natural killer cell gene complex

prion disease 95(NKC) 23–4

prion protein vineoglycoproteins, synthesis of 10–12

glycosylation and GPI anchorage 95–6neonatal receptor (FcRn) 143

protease, cascade 104, 105(fig.)neuronal migration disorder 97

proteoglycans 47neuronal stem cells 119neutrophil collagenase (MMP-8) 47, 49,

recombinant human erythropoietins viii53(fig.)recombinant human erythropoietins,neutrophil gelatinase B-associated lipocalin

N-glycan structures and biological(NGAL) 47, 48(fig.)activities 141–2Nkrp1 (Klrb) 23–4

recombinant IgG antibody viiinovel anti-HVC molecules 5recombinant therapy viiiREGA model 50–1O-acetyl sialic acid expression viirenal cancerO-fucose glycans vincidence 113oocyte development and function, role

in 99–100 metastatic disease, chemotherapy 113

Index 153

rheumatoid arthritis vi stromelysins 47substrate reduction therapy (SRT) vii, 2,associated glycosylation changes, new

insights 129–36 119–20Sahdhoff disease, mouse model 120–1gelatinase B, expression in 45–53

rheumatoid factor vii sugarscancer, association with viribavirin 5viral infections, pathogenesis vsynovial fluids, gelatinase B expression 47,Sandhoff disease 2, 120–3

Schistosoma haematobium 10 48(fig.)cartilage breakdown 48–51Schistosoma japonicum 10

Schistosoma mansoni 10 systemic vasculitis viischistosome egg antigens (SEA) 14major glycan antigens within 12 T helper cells-2 (Th-2) responses,

schistosome glycans schistosomiasis 9–16host immune response, major Tay–Sachs disease 2, 123

focus 12–13 thrombin 37interaction with the host immune tissue inhibitor of metalloproteinases-1

system 9–16 (TIMP-1) 48, 105Lewis-x, binding to the CRD of the toll-like receptors

CD-sign transmissible spongiformneoglycoproteins, synthesis 10–12 encephalopathies 95–6parasite-derived glycans, dendritic cells trematodes 9

recognition tropical parasitic diseases vischistosome egg antigens, dendritic cells tumour markers, ovarian cancer serum

binding, lectins identification 14 marker 114schistosomes, life cycle 9–10 tumour necrosis factor a 45schistosomiasis vi, 9 tumour necrosis factor-stimulatedschistosomula 10 gene-6 58, 60(fig.), 63–6scrapie in sheep 95 hyaluronan binding to cell surfaceselectin family of adhesion molecules vii CD44 64–6septic arthritis vii Tyr-105 61serpin 37sialic acid 131

viral infections, pathogenesis vsialic acid-binding lectins 13

vitamin C, 123sialidase treatment 135

vitamin E 123siglecs 13

van der Waals interactions 38staphylococcal arthritis viistaphylococcal infection, carbohydrates and

Wegener’s granulomatosis 147biology 115–16Staphylococcus aureus 115Staphylococcus epidermidis 115 yeast, production of complex human

glycoproteins 139Staphylococcus saprophyticus 115

13–16

12–14

12

glycobiologyand medicine and...proceedings of the 7th jenner glycobiology and medicine symposium...

Documents