ADVANCES IN PROTEIN CHEMISTRY EDITED BY FREDERIC M. RICHARDS DAVID E. EISENBERG Department of Molecular Biophysics Department of Chemistry and Biochemistry and Biochemistry University of California, Los Angeles Yale University Los Angeles, California New Haven, Connecticut PETER S. KIM Department of Biology Massachusetts Institute of Technology Whitehead lnstitute for Biomedical Research Howard Hughes Medical lnstitute Research Laboratories Cambridge, Massachusetts VOLUME 49 Antigen Binding Molecules: Antibodies and T-cell Receptors EDITED BY EDGARHABER Department of Biological Sciences Harvard School of Public Health Harvard Medical School Boston, Massachusetts ACADEMIC PRESS San Diego London Boston New York Sydney Tokyo Toronto @ This book is printed on acid-free paper. Copyright 0 1996 by ACADEMIC PRESS All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Academic Press, Inc. 525 B Street, Suite 1900, San Diego, California 92101-4495, USA http:lfwww.apnet.com Academic Press Limited 24-28 Oval Road, London NWl 7DX, UK http://www.hbuk.co.uk/ap/ International Standard Serial Number: 0065-3233 International Standard Book Number: 0-1 2-034249-9 PRINTED IN THE UNITED STATES OF AMERICA 96 97 9899 00 01BC 9 8 7 6 5 4 3 2 1 PREFACE Immunology began at the end of the nineteenth century when Paul Ehrlich became fascinated by precipitation reactions in the serum of immunized animals. His work was followed by a systematic analysis by Karl Landsteiner of the exquisite specificity of what we now know to be the antigen-antibody interaction. Yet the precipitin reaction remained in the realm of phenomenology until the 1950s, during which Rodney Porter and Gerald Edelman defined the modular structure of the antibody molecule and demonstrated that antigen binding was the property of only one domain of this multifunctional molecule. Renaturation experiments from my laboratory and those of Michael Sela and Charles Tanford soon implied that there had to be a great many different antibody molecules to account for the wide range of antigen binding specificities, since the amino acid sequence of an antibody determined its specificity. After the earliest amino acid sequences of myeloma proteins (homogeneous surrogates for the very heterogeneous serum antibodies) became available in 1965 from Norbert Hilschmann and Lyman Craig, it was apparent that a hypervari- able region accounted for the antigen binding domain. The ability to clone examples of real antibodies by the hybridoma method of George Kohler and Cksar Milstein then led to a surge of activity in the 1970s that defined many aspects of the structure and function of the antigen binding site. Soon it was clear that the same principles of binding applied not only to humoral antibodies but to T-cell receptors and many members of what we now call the immunoglobulin superfamily. Interest in antigen binding waned for a decade while the detailed mechanisms of the cellular immune response were explored and defined. Now in the 1990s we are witnessing a revival of interest in antigen bindings as its broad relevance to many immunologic processes has become appar- ent and a variety of new analytic tools have allowed for deeper insights into the nature of antigen binding. Advances in molecular engineering permit the recapitulation of clonal selection and affinity maturation of antibodies in vitro. It is possible to produce antibodies to selected antigens without the mediation of an animal host, and it is even possible to turn an antigen binding site into a highly selective enzyme. This volume brings the reader up to date on the covalent and three- dimensional structures of the antibody molecule’s antigen binding domain and the synthesis and use of this domain as a separate small molecule. The reader will find a full account of antibody three-dimensional structure (as xi xii PREFACE revealed by x-ray crystallography and computational biochemistry) as well as an analysis of how antigens bind. The nature and the structure of an antigen are defined, and the affinity maturation of antibodies is examined in relation to gene structure and diversification. The exciting field of cata- lytic antibodies has advanced to include a range of enzymatic functions not even contemplated when the first examples were described a few years ago. The T-cell receptor, which has some elements in common with the anti- body molecule, is far more complex in its antigen recognition function. The T-cell receptor is analyzed here in the context of its binding to anti- gen and to the essential major histocompatibility complex. The role of the T-cell receptor’s accessory proteins in binding and activation is also defined. Not only immunologists but also biologists and chemists should profit from reading this volume: it reveals a mature yet evolving field of broad interest to other areas of science. EDGAR HABER ANTIGEN-SPECIFIC T-CELL RECEPTORS AND THEIR REACTIONS WITH COMPLEXES FORMED BY PEPTIDES WITH MAJOR HISTOCOMPATIBILITY COMPLEX PROTEINS By HERMAN N. EISEN, YURl SYKULEV, and THEODORE J. TSOMIDES Center for Cancer Research and Department of Biology Massachusetts Institute of Technology Cambridge, Massachusetts0 2139 I. Overview ............................................... 1 11. History a ound .......... 3 111. T-cell Receptor Genes ............................................ 8 IV. T-cell Receptor Proteins 11 V. T-cell Receptor Ligands: Peptide-MHC Complexes. ................... 15 VI. T-cell Responses to PepMHC 22 A. Affinity: Intrinsic Equilibri 23 B. Kinetics ................. 26 C. Time Required to Approach .......................... 29 D. Epitope (PepMHC) Density 30 E. TCR-PepMHC Engagement: .................... 32 F. Specificity, Degenerac 35 VII. T-cell Receptor Accessory Proteins. ................................. 39 A. CD3andc.. ......... 39 B. CD4andCD8.. ...... 40 VIII. Altered Peptide Ligands: Partial Agonists and Antagonist 41 IX. MHC Restriction by Self and Nonself MHC: The Paradox 43 X. Concluding Remarks. ............. 47 References ...................................................... 48 I. OVERVIEW The most distinctive feature of the vertebrate immune system is its ability to recognize an enormous number of organic molecules and molec- ular complexes, termed antigens, distinguishing broadly between those that are foreign to the responding animal (nonself) and those that are indigenous (self). This property is due to antigen-specific receptors on lymphocytes, small cells that comprise -5% of all cells in the body (esti- mated at 101'-lO1z out of about loL3ce lls in an adult human). The recep- tors on the two major classes of lymphocytes, B and T cells, are similar structurally but profoundly different functionally. On B cells the receptors 1 ADVANCES IN Copyright 63 1996 by Academic Press, Inc. PROTEIN CHEMISTRY W1. 49 All rights of reproduction in any form reselved. 2 HERMAN N. ElSEN ET AL. are immunoglobulins (Ig) embedded in the cell surface as integral mem- brane proteins; in response to recognition of antigens, B cells produce large amounts of the receptors and secrete them as soluble antibody mole- cules. The antigen-specific receptors on T cells (T-cell receptors or TCR) are also Ig-like cell surface integral membrane proteins; their recognition of antigens triggers T cells to exercise a great variety of functions but not to secrete the receptors. Both Ig and TCR molecules are heterodimers, each subunit consisting of two or more domains, with each domain having a characteristic three- dimensional shape called the Ig fold. The N-terminal domains (termed variable or V domains) differ in amino acid sequence from one lymphocyte clone to another. The variable domains of each heterodimer pair to form a single antigen binding site that determines the unique ability of each clone to recognize and respond to only very few of the millions of different antigens to which an individual animal can respond. The enormous diver- sity of B- and T-cell receptors arises from the many germline gene seg- ments that encode them; as each lymphocyte matures, different combina- tions of these segments are joined (combinatorial diversity) and additional variations in sequence are introduced at the junctures (junctional diver- sity), leading to an immense number of variable domain sequences. Despite extensive similarities in amino acid sequence between Ig and TCR and in the organization and recombination of the gene segments that encode them, these receptors differ remarkably in the universe of antigens they recognize. The antigens recognized by antibodies (or their membrane- bound form on B cells) vary enormously: Physically they may be soluble, colloidal, particulate, or parts of virions or microbial or eukaryotic cells, and chemically they may be proteins, peptides, carbohydrates, lipids, nucleic acids, or any of a limitless number of diverse small organic molecules. In contrast, the TCR reviewed here normally recognize and respond only to complexes formed between small peptides and a specialized set of pro- teins encoded by the major histocompatibility complex (MHC). Because MHC molecules (sometimes called histocompatibility antigens) are integral membrane proteins, these complexes (termed pepMHC) are confined to cell surfaces, and the TCR of a T cell is therefore normally able to recog- nize antigenic complexes only on the surfaces of other cells, called antigen presenting cells or target cells. Inasmuch as both TCR and their natural pepMHC ligands are embedded in cell surface membranes, analysis of their interaction at the molecular level poses a major challenge. Later in this review we focus on how this challenge is being met and emphasize recent results that illuminate the way in which TCR react with (or “recognize”) their natural pepMHC ligands, particularly the TCR on those T cells (called cytotoxic T lymphocytes or CTL) that destroy other cells (termed target cells). ANTIGEN-SPECIFIC T-CELL RECEPTORS 3 11. HISTORAYND BACKGROUND The immune system has been under study for about 100 years, but only in the past 30 years have T and B cells been distinguished and only in the past 12 years have TCR molecules and the genes encoding them been identified. Nevertheless, for decades before T cells and TCR emerged as recognized entities, their existence was foreshadowed by certain antigen- specific inflammatory responses produced by injecting antigens into the skin of individuals with previous exposure to such antigens, either through natural infection of deliberate inoculation (immunization). Around 1890, Robert Koch showed that the injection of tubercle bacilli (or a mixture of proteins called “tuberculin” from supernatants of tubercle bacilli cultures) elicited an intense inflammatory response in guinea pigs if they had been previously infected with these microorganisms. The responses appeared 1248h r following antigen injection, and similar delayed-type hypersensi- tivity (DTH) responses, always specific for the original inciting antigen, were subsequently demonstrated with crude protein mixtures from many other microbes (bacteria, fungi, and later, viruses). Responses sometimes occurred following deliberate immunization with purified proteins (e.g., ovalbumin) or even with small organic molecules applied to the skin, pro- viding they reacted in situ to form covalent derivatives of skin proteins (as with 2,4-dinitrochlorobenzene,o r a catechol in the case of poison ivy). In contrast to the late appearance of DTH responses, many other antigen-specific skin responses appear almost immediately, e.g., within 1 min or sometimes 2-3 hr. Because the transfer of serum antibodies from an immunized to a nonimmunized (naive) individual confers on the recipient the same prompt antigen-specific responses, it was clear that these rapid or immediate-type responses were mediated by antibodies. But serum failed to transfer the delayed-type responses (e.g., to tuberculin). Although efforts were made to reconcile these failures with a role for special antibodies, it came to be widely believed that antigen-specific cells rather than soluble antibody molecules were the direct mediators of DTH responses. This belief was supported by the finding that DTH responses could be transferred to naive recipients with inflammatory cells from im- munized donors, although the transferred cells were complex mixtures of leukocytes that probably included some antibody-forming cells. The reso- lution of all doubt came after T cells were distinguished from B cells, largely through studies involving extirpation of the thymus from newborn mice. These studies led to the establishment of a clear dichotomy between those lymphocytes that develop from immature precursors in the thymus (T cells) and those that develop to maturity in the bone marrow (B cells). B cells were shown to be the source of Ig and antibodies, and T cells were 4 HERMAN N. EISEN ET AL shown to be required for the optimal production of antibodies. In addi- tion, highly purified populations of T cells could transfer DTH responses, suggesting that T cells bear receptors that recognize antigens. Early studies on the nature of the antigen-recognizing receptors on T cells were marked by intense disagreements arising largely, it appears in retrospect, because the T-cell populations studied were often contami- nated by small numbers of B cells and because reliance was placed almost exclusively on serological analyses using anti-Ig antisera. Antigen-specific molecules on T cells were variously suggested to be Ig molecules firmly attached to T cells, to be some other (non-Ig) type of cell-associated pro- tein, or to be some other type of nonprotein informational macromole- cule. The debate was resolved by the development of monoclonal antibody (MAb) technology and the ability to grow T-cell clones in culture. Using MAb raised against T-cell clones, one a malignant lymphoma and the others normal T cells, three independent studies succeeded in immuno- precipitating a T-cell surface protein that was unique to cells of the immu- nizing clone (Allison et al., 1982; Meuer et al., 1983; Haskins et al., 1983). The clonally specific (or clonotypic) protein was in each case a disulfide- linked heterodimer (-90 kDa) consisting of a relatively acidic membrane- bound a chain and a more basic membrane-bound p chain. The idea that these clone-specific heterodimers were antigen-specific receptors was strengthened by two findings. First, antibodies against them could block antigen-driven responses of the corresponding T-cell clones (Lancki et al., 1983). Second, amino acid sequences of proteolytic fragments from isolated a and p subunits suggested that, like Ig heavy and light chains, they had some regions where amino acid sequences varied from clone to clone and others where these sequences were invariant (Kappler et al., 1983; Meuer et al., 1984; McIntyre and Allison, 1983). The isolation and sequencing of cDNA clones for the b subunit (Yanagi et al., 1984; Hedrick et al., 1984; Saito et al., 1984a) and then for the a subunit (Saito et al., 1984b; Chien et al., 1984) finally demonstrated unambiguously that the clonally diverse heterodimers greatly resembled Ig and had all the characteristics expected of cell surface integral membrane proteins serv- ing as antigen-specific receptors. Moreover, introduction of genes for both the a and /3 subunits into a T-cell clone having unrelated specificity trans- ferred the antigen-specific responsiveness of the donor cell to the recipient cell (Dembic et al., 1986; Saito et al., 1987). In the course of searching a T-cell cDNA library for clones for the a and p subunits, cDNA for a third related subunit was found (Saito et al., 1984a). Termed y, the third gene turned out to encode an Ig-like chain that paired with the Ig-like product of a fourth gene, 6, to form a y6 heterodimer (Chien et al., 1984). Closely similar to the ab TCR, y6 heterodimers are ANTIGEN-SPECIFIC T-CELL RECEPTORS 5 another type of antigen-specific receptor found on a subset of T cells + located primarily in epithelia. yd cells constitute about 1-5% of periph- eral T cells in mice and humans; these TCR will not be considered further here, as their natural ligands are not as well-defined as those recognized by aB TCR and the hnction of yd T cells is still not clear. Ten years before the molecular identity of the ab TCR began to take shape, several observations pointed to the special character of its natural ligands. By transferring T cells to athymic (nude) mice, Kindred and Shreffler (1972) saw that these cells reacted only when the T cells and the recipients had the same MHC type. The ineffectiveness of MHC-dispartate T cells could not be attributed to their immune elimination in the recipi- ent because athymic mice do not reject allografts (see Alloaggression, Section IX). It was therefore concluded that the MHC “must play an active role in ensuring cooperation between B and T cells.” Shortly thereafter, in vitro studies of T-cell enhancement of the produc- tion of antibodies by B cells (Katz et al., 1973) and of T-cell responses to antigens presented by macrophages (Rosenthal and Shevach, 1973) clearly indicated that successful T-cell responses required the responding T and B cells, or T cells and macrophages, to have the same MHC type. Then, in a seminal publication, Zinkernagel and Doherty (19 74) reported that virus- infected target cells were lysed by T cells from a mouse infected by that virus only if the target cells expressed the same MHC products as the infected animal. Work with different antigens revealed similarly that antigen- specific lysis of antigen-bearing target cells by cytotoxic T cells depended on target cell expression of a proper MHC protein (Shearer, 1974; Bevan, 1975; Gordon et al., 1975). This dual requirement of antigen recognition, referred to as “MHC restriction” (Zinkernagel and Doherty, 1979), has since been found to be characteristic of all a/?T CR-mediated reactions. To explain MHC restriction, two models were proposed. In the two- receptor model, T cells had one receptor for an MHC product and an- other for antigen, and both had to be occupied for a successful T-cell response (Cohn and Epstein, 1978). According to the alternative one- receptor model, each T cell had a single type of receptor that recognized an antigen-MHC complex. The one-receptor model was shown to be cor- rect by various approaches. In one, a hybridoma resulting from the fusion of two cells recognizing antigen A and MHC X or antigen B and MHC Y + + responded specifically to cells expressing either A X or B Y but not to + + those expressing A Y or B X, indicating that the T cell did not see the antigen and the MHC product separately (Kappler et al., 1981). The antigeneMHC complex was sometimes referred to as “altered self’ be- cause in this context the restricting MHC is indigenous (self) with respect to the responding T cells and is therefore nonimmunogenic, while the 6 HERMAN N. EISEN ET AL antigen is foreign (nonself) and somehow alters the character of self MHC, imposing “foreignness” and ability to elicit a reaction. Mature T cells bearing ap TCR fall into two groups, each marked by one of two cell surface glycoproteins. Termed CD4 and CD8 (formerly called Lyt-1 and Lyt-2,3 in the mouse), both are present on,immature (double positive) T cells in the thymus, and one of them is lost by what appears to be a stochastic process (Corbella et al., 1994) to yield single positive (CD4+ or CD8+) mature T cells. Some mature CD4+ T cells, termed helper (Th2) cells, are required for optimal B-cell responses, while other helper cells (Thl) produce cytokines that cause inflammation, as in DTH responses. In contrast, CD8+T cells, which also produce some cyto- kines, behave primarily as CTL, lysing target cells that bear appropriate antigenic MHC complexes. The CD4/CD8 T-cell dichotomy also extends to the types of MHC pro- tein that restrict antigen recognition. Class I MHC proteins (MHC-I) restrict antigen recognition by CD8+ T cells, and class I1 MHC proteins (MHC-11) restrict antigen recognition by CD4+ T cells (Table I). Virtually all cells express MHC-I and can present peptides to CD8+ cells, whereas only specialized cells (principally macrophages, dendritic cells, and B cells) ex- press MHC-I1 for interactions with CD4+ cells. In common usage, cells that present pepMHC-I1 complexes to CD4+ cells are called antigen pre- senting cells (APC), whereas those that present pepMHC-I complexes to CD8+ CTL are termed target cells. Once it became established that a single heterodimeric receptor on T cells reacts with an antigen displayed on the surface of another cell and TABLE I Peptide Binding to MHC-I and MHC-II Molecules Parameter MHC-I MHC-I1 Distribution All nucleated cells Specialized APC Domain structure az, a3 + Pzm al, a2 + PI,P Z Accessory molecule CD8 CD4 Typical T-cell response Cytolytic activity B-cell help, DTH Origin of most bound Cytosolic (endogenous) Extracellular and membranous peptides proteins - Length of bound peptides Usually 8-9 residues Variable, 12-25 residues Pockets in MHC groove Yes Yes Peptide N and C termini Buried in groove May extend outside groove Many critical contacts Peptide side chains Peptide backbone - Equilibrium constant, Ka -104-109 M-I measured 106-1 OH h.T’ measured