Artigo Acesso aberto Revisado por pares

The three-dimensional structure of a T-cell antigen receptor Valpha Vbeta heterodimer reveals a novel arrangement of the Vbeta domain

1997; Springer Nature; Volume: 16; Issue: 14 Linguagem: Inglês

10.1093/emboj/16.14.4205

ISSN

1460-2075

Autores

Dominique Housset, Gilbert Mazza, Claude Grégoire, C. Piras, Bernard Malissen, Juan Carlos Fontecilla‐Camps,

Tópico(s)

Toxin Mechanisms and Immunotoxins

Resumo

Article15 July 1997free access The three-dimensional structure of a T-cell antigen receptor VαVβ heterodimer reveals a novel arrangement of the Vβ domain Dominique Housset Dominique Housset Laboratoire de Cristallographie et Cristallogénèse des Protéines, Institut de Biologie Structurale 'Jean-Pierre Ebel' CEA-CNRS, 41 avenue des Martyrs, 38027 Grenoble, cedex 1, FranceD.Housset and G.Mazza contributed equally to this work Search for more papers by this author Gilbert Mazza Gilbert Mazza Laboratoire de Cristallographie et Cristallogénèse des Protéines, Institut de Biologie Structurale 'Jean-Pierre Ebel' CEA-CNRS, 41 avenue des Martyrs, 38027 Grenoble, cedex 1, France Centre d'Immunologie INSERM-CNRS de Marseille-Luminy, Case 906, 13288 Marseille, cedex 9, FranceD.Housset and G.Mazza contributed equally to this work Search for more papers by this author Claude Grégoire Claude Grégoire Centre d'Immunologie INSERM-CNRS de Marseille-Luminy, Case 906, 13288 Marseille, cedex 9, France Search for more papers by this author Claudine Piras Claudine Piras Laboratoire de Cristallographie et Cristallogénèse des Protéines, Institut de Biologie Structurale 'Jean-Pierre Ebel' CEA-CNRS, 41 avenue des Martyrs, 38027 Grenoble, cedex 1, France Search for more papers by this author Bernard Malissen Corresponding Author Bernard Malissen Centre d'Immunologie INSERM-CNRS de Marseille-Luminy, Case 906, 13288 Marseille, cedex 9, France Search for more papers by this author Juan Carlos Fontecilla-Camps Corresponding Author Juan Carlos Fontecilla-Camps Laboratoire de Cristallographie et Cristallogénèse des Protéines, Institut de Biologie Structurale 'Jean-Pierre Ebel' CEA-CNRS, 41 avenue des Martyrs, 38027 Grenoble, cedex 1, France Search for more papers by this author Dominique Housset Dominique Housset Laboratoire de Cristallographie et Cristallogénèse des Protéines, Institut de Biologie Structurale 'Jean-Pierre Ebel' CEA-CNRS, 41 avenue des Martyrs, 38027 Grenoble, cedex 1, FranceD.Housset and G.Mazza contributed equally to this work Search for more papers by this author Gilbert Mazza Gilbert Mazza Laboratoire de Cristallographie et Cristallogénèse des Protéines, Institut de Biologie Structurale 'Jean-Pierre Ebel' CEA-CNRS, 41 avenue des Martyrs, 38027 Grenoble, cedex 1, France Centre d'Immunologie INSERM-CNRS de Marseille-Luminy, Case 906, 13288 Marseille, cedex 9, FranceD.Housset and G.Mazza contributed equally to this work Search for more papers by this author Claude Grégoire Claude Grégoire Centre d'Immunologie INSERM-CNRS de Marseille-Luminy, Case 906, 13288 Marseille, cedex 9, France Search for more papers by this author Claudine Piras Claudine Piras Laboratoire de Cristallographie et Cristallogénèse des Protéines, Institut de Biologie Structurale 'Jean-Pierre Ebel' CEA-CNRS, 41 avenue des Martyrs, 38027 Grenoble, cedex 1, France Search for more papers by this author Bernard Malissen Corresponding Author Bernard Malissen Centre d'Immunologie INSERM-CNRS de Marseille-Luminy, Case 906, 13288 Marseille, cedex 9, France Search for more papers by this author Juan Carlos Fontecilla-Camps Corresponding Author Juan Carlos Fontecilla-Camps Laboratoire de Cristallographie et Cristallogénèse des Protéines, Institut de Biologie Structurale 'Jean-Pierre Ebel' CEA-CNRS, 41 avenue des Martyrs, 38027 Grenoble, cedex 1, France Search for more papers by this author Author Information Dominique Housset1, Gilbert Mazza1,2, Claude Grégoire2, Claudine Piras1, Bernard Malissen 2 and Juan Carlos Fontecilla-Camps 1 1Laboratoire de Cristallographie et Cristallogénèse des Protéines, Institut de Biologie Structurale 'Jean-Pierre Ebel' CEA-CNRS, 41 avenue des Martyrs, 38027 Grenoble, cedex 1, France 2Centre d'Immunologie INSERM-CNRS de Marseille-Luminy, Case 906, 13288 Marseille, cedex 9, France *Corresponding authors. E-mail: [email protected] or E-mail: [email protected] The EMBO Journal (1997)16:4205-4216https://doi.org/10.1093/emboj/16.14.4205 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info The crystal structure of a mouse T-cell antigen receptor (TCR) Fv fragment complexed to the Fab fragment of a specific anti-clonotypic antibody has been determined to 2.6 Å resolution. The polypeptide backbone of the TCR Vα domain is very similar to those of other crystallographically determined Vαs, whereas the Vβ structure is so far unique among TCR Vβ domains in that it displays a switch of the c″ strand from the inner to the outer β-sheet. The β chain variable region of this TCR antigen-binding site is characterized by a rather elongated third complementarity-determining region (CDR3β) that packs tightly against the CDR3 loop of the α chain, without leaving any intervening hydrophobic pocket. Thus, the conformation of the CDR loops with the highest potential diversity distinguishes the structure of this TCR antigen-binding site from those for which crystallographic data are available. On the basis of all these results, we infer that a significant conformational change of the CDR3β loop found in our TCR is required for binding to its cognate peptide-MHC ligand. Introduction The specific recognition of antigen by T cells and its ensuing transduction into intracellular signals are accomplished by a multi-subunit complex denoted as the T-cell receptor (TCR)-CD3 complex. The TCR α and TCR β polypeptides, which confer antigen-binding capacity to these membrane-bound complexes, consist of an amino-terminal variable (V) region and a carboxy-terminal constant (C) region. The determination of the three-dimensional structure of an isolated TCR β chain (Bentley et al., 1995) and of a homodimer of Vα domains (Fields et al., 1995) indicated that both Vα and Vβ regions are structurally related to the V domains of the heavy (H) and light (L) chains of immunoglobulins (Igs), and showed that peptide loops homologous to Ig complementarity-determining regions (CDRs) protrude at the membrane-distal ends of both TCR Vα and Vβ domains where they collectively form the antigen-binding site. As previously observed for Ig V domains, the first and second CDR equivalents are encoded within V gene segments, whereas the third CDR equivalent is formed during somatic DNA recombination events involving the juxtaposition of Vα and Jα gene segments in TCR α chain genes, and of Vβ, Dβ and Jβ gene segments in TCR β chain genes. During V(D)J recombination, the coding ends of the TCR gene segments are subjected to various degrees of base deletion, addition or both. Due to this extensive junctional diversity, the Vα and Vβ CDR3s are responsible for most of the diversity observed in αβ TCRs (Davis and Bjorkman 1988; Jores et al., 1990). The αβ TCRs recognize peptides that are bound to either class I or class II products of the major histocompatibility complex (MHC). X-ray structures of MHC class I molecules complexed with different peptides have shown that the latter are buried in the groove formed between the α1 and α2 helices, leaving only a few of their side chains available for direct TCR contact (Fremont et al., 1992, 1995; Zhang et al., 1992; Madden et al., 1993; Young et al., 1994). These data have led to the view that TCR VαVβ regions most likely recognize a composite surface made of residues belonging to a given antigenic peptide and to both MHC α1 and α2 helices (Sun et al., 1995). Here we report the 2.6 Å resolution structure of a TCR single chain Fv fragment (scFv) derived from the KB5-C20 mouse cytotoxic T cell clone complexed to a Fab fragment of the specific monoclonal anti-clonotypic antibody Désiré-1 (Albert et al., 1982; Hua et al., 1985; Hue et al., 1990; Grégoire et al., 1991, 1996). While this manuscript was in preparation, two reports on TCR crystal structures have appeared. The first one describes the structure of a complete αβ TCR ectodomain derived from the 2C mouse alloreactive cytotoxic T-cell clone (Garcia et al., 1996a). Interestingly, the antigen-binding sites of the KB5-C20 and 2C TCRs use distinct Vα-Vβ gene segment combinations and exhibit TCR β-chain CDR3 region lengths which lie at opposite tails of the TCR β CDR3 size distribution (Candéias et al., 1991; Pannetier et al., 1993). Despite these marked differences, both TCRs recognize the same MHC class I molecule (H-2Kb) when complexed to different peptides. Comparison of their combining sites thus provides a unique opportunity to examine the range of TCR conformational variability allowed for the recognition of a given class I MHC molecule. The second report corresponds to the structure of a complex containing a human αβ TCR bound to an HLA-A2 molecule loaded with a nonapeptide derived from the HTLV-1 virus (A6 TCR-Tax-HLA-A2 complex, Garboczi et al., 1996). Using the atomic coordinates of this ternary complex, we have carried out rigid-body rotations of the KB5-C20 TCR scFv and H-2Kb structures onto the positions occupied by their respective human counterparts, and found that a significant conformational change of the KB5-C20 CDR3β loop is required for binding to its cognate peptide-MHC ligand. Moreover, this hypothetical model allowed us to define a minimum set of plausible interactions between the KB5-C20 TCR and H-2Kb molecule, and to compare them with the ones reported in the A6 TCR-Tax-HLA-A2 complex. Results and discussion Overall structure The crystal structure of the complex of the KB5-C20 TCR scFv [Vα2.3(AV2S3)-JαA10/Vβ2(BV2S1)-Dβ2-Jβ2.3] and Désiré-1 Fab has been determined to 2.6 Å resolution (see Materials and methods and Table I). The model is well defined except for the extra eight amino acids unique to the N-terminus of Vα2 polypeptides (Gahéry-Ségard et al., 1996; Grégoire et al., 1996) and the 24 residue linker used to connect the Vα and Vβ domains (Grégoire et al., 1996). All these residues were omitted from the final model. Structure solution and refinement statistics are given in Table II. Table 1. Data collection statistics Resolution (Å) No. of measurements No. of reflections [F>3σ(F)] Rsym (last shell) Completeness (%) (last shell) Redundancy (last shell) (last shell) Form 1 2.9 147 342 29 732 (27 746) 0.091 (0.400) 86.7 (67.4) 5.0 (3.1) 7.2 (1.8) Form 2 2.6 75 774 24 517 (18 474) 0.086 (0.300) 86.7 (75.6) 3.1 (2.5) - Table 2. Molecular replacement and refinement statistics Molecular replacement statistics Resolution range (Å) R-factor (%) Correlation coefficient (%) One Cκ-CH1 15.0-3.5 58.6 13.9 One VL-VH 15.0-3.5 59.0 11.1 Two Cκ-CH1 15.0-3.5 53.4 35.8 Two VL-VH 15.0-3.5 54.1 32.9 Two Fabs 15.0-3.5 51.3 40.7 Two Fabs + two Vα-Vβ 15.0-3.5 48.8 48.3 10.0-2.9 48.7 48.1 Resolution (Å) No. of reflections No. of atoms R-factor Rfree Stereochemistry (SD from ideality) Form 1 F >3σ(F) 26 296 24.9 31.6 bonds 0.014 Å 8.0-2.9 10 280 angles 0.039 Å all 28 266 27.0 33.9 Form 2 F >3σ(F) 18 105 21.0 30.4 bonds 0.013 Å 10.0-2.6 5438 angles 0.034 Å all 20 232 22.1 31.5 correlation coefficient = <(Fobs− )(Fcalc−Fcalc>)>/ . Rfree is an R-factor calculated on a subset of reflections (5%), not used in the refinement. Non-crystallographic symmetry restraints have been used during refinement of crystal form 1. Refinement statistics Stereochemistry (SD from ideality) Figure 1 shows the backbone ribbon representation of the complex. All six Fab CDRs interact with the TCR scFv (CDR1L, six residues; CDR2L, one residue; CDR3L, three residues; CDR1H, three residues; CDR2H, five residues; CDR3H, five residues). The antibody-accessible surface area buried by the interaction with the TCR Fv is 1343 Å2 as calculated with X-PLOR (Brünger, 1990) using a 1.5 Å radius probe. Conversely, the areas of the buried surfaces on the Vα and Vβ domains correspond to 545 Å2 (22 residues) and 833 Å2 (26 residues), respectively. Consistent with the anti-clonotypic nature of the Désiré-1 antibody, the segments of the KB5-C20 TCR scFv that interact with Désiré-1 encompass both Vα and Vβ CDRs. As shown in Figure 1, the Vα CDR3 interacts with Fab residues contributed by the CDR2L, CDR1H and CDR3H, whereas the Vβ CDR2 interacts with CDR3L and CDR2H residues. It should be emphasized that the only contact between the Vβ CDR3 and the Désiré-1 Fab is the perpendicular interaction of the aromatic rings of W100β and Y32CDR1L. The limited number of TCR CDRs involved in the interaction with the Désiré-1 Fab fragment is reflected by the fact that the geometrical centre of the Fab-TCR interface lies near the c′ strand of the TCR Vβ domain. This is in contrast with the structure of an Fab-anti-idiotype Fab complex (Bentley et al., 1990) where the two Ig Fab fragments are roughly aligned along their longest dimension and interact mostly through their respective CDRs. Therefore, bacterial superantigens (Fields et al., 1996), peptide-MHC complexes (Garboczi et al., 1996) and the Désiré-1 antibody (this paper) differ markedly in the way they bind to TCR V domains. Nevertheless, these TCR ligands are all capable of efficiently activating T cells. Figure 1.Overall stereoscopic view of the complex between the KB5-C20 TCR scFv and Désiré-1 Fab fragment. The TCR scFv is at the top of the figure. The β strands were determined with the program PROCHECK (Laskowski et al., 1993), and are represented as arrows, the α helices are depicted in green, and the CDRs as black coils.The β strands are labelled according to Bork et al. (1994). The linker connecting the C-terminus of Vα to the N-terminus of Vβ is not seen in the electron density map and is depicted as a dotted line. Its path appears to influence neither the Vα-Vβ association, nor the CDR conformation. This figure was produced with MOLSCRIPT (Kraulis, 1991). Abbreviations are as follows: a, TCR Vα domain; b, TCR Vβ domain; H, Ig heavy chain; L, Ig light chain; N-ter, NH2-terminus; C-ter, COOH-terminus. Download figure Download PowerPoint The topology of Vα and Vβdomains The amino acid sequence of the Vα2.3 segment used by the KB5-C20 TCR is 26% identical to that of the Vα4.2 segment used by the 1934.4 TCR (Fields et al., 1995). Despite their identical names, note that the Vα2.3 segment used by the KB5-C20 TCR does not constitute the mouse homologue of the Vα2.3 segment found in the human A6 TCR (Clark et al., 1995; Garboczi et al., 1996). Accordingly, these two Vα segments share only 48% identity at the protein level. Based on the analysis of the three-dimensional structures of the 1934.4 Vα4.2 and 2C Vα3 segments, TCR Vα domains have been shown to be unique among Ig-related V domains in that their fifth β-strand (also known as c″; Bork et al., 1994) forms hydrogen bonds with the d strand of the outer β-sheet (Fields et al., 1995; Garcia et al., 1996a). Such a switch of the c″ strand from the inner to the outer β-sheet was found to remove a surface protrusion from the Vα domain and, based on the packing found in a Vα crystal (Fields et al., 1995), was hypothesized to permit the initiation of T-cell activation via αβ TCR dimerization (Fields et al., 1995). In the KB5-C20 Vα domain, the c″ strand is also switched (Figure 2A). Moreover, consistent with their relative levels of primary sequence identities, the KB5-C20 Vα2.3 polypeptide backbone superposes better with A6 Vα2.3 than with Vα4.2 [root mean square (r.m.s.) differences of 0.77 and 0.92 Å are obtained for the positions of 98 pairs and 86 pairs of structurally equivalent α carbons, respectively; Figure 2A]. Figure 2.Stereoscopic view of the α-carbon backbone of TCR Vα and Vβ domains. V domains were optimally superimposed with the program ALIGN (Satow et al., 1986). The β strands are represented as thick lines and labelled according to Bork et al. (1994). (A) Diagram of the KB5-C20 Vα2.3 domain (black) superposed onto the A6 Vα2.3 (dark grey) and 1934.4 Vα4.2 (light grey) domains. Note that the CDR loops found in the KB5-C20 Vα2.3, 1934.4 Vα4.2 and A6 Vα2.3 domains display closely related conformations. (B) Diagram of the KB5-C20 Vβ2 domain (black) superposed onto the A6 Vβ12.3 (dark grey) and 14.3.d Vβ8.2 (light grey) domains. The c″ strand of the KB5-C20 Vβ domain is switched and hydrogen bonded to strand d, as previously observed in the A6 Vα2.3, 2C Vα3, 1934.4 Vα4.2 and KB5-C20 Vα2.3 domains. In the KB5-C20 Vβ2 domain, both the c″ and d strands are three residues longer than those found in the 14.3.d Vβ8.2 domain (see Table III). The 14.3.d CDR3β is three residues shorter than the KB5-C20 CDR3β, and adopts a different conformation, presumably due to the lack of Vα partner in the 14.3.d Vβ structure. The A6 CDR3β is two residues shorter than the KB5-C20 CDR3β and its tip folds away from CDR3α, opening a cavity at the CDR3α-CDR3β interface that accommodates the side chains of the central residues of the HTLV-1 peptide (Garboczi et al., 1996). Download figure Download PowerPoint The amino acid sequence of the Vβ2 segment used by the KB5-C20 TCR is 27% identical to that of the Vβ8.2 domain used by both the 14.3.d and 2C TCRs (Bentley et al., 1995; Garcia et al., 1996a), and 28% identical to the human Vβ12.3 segment used by the A6 TCR (Garboczi et al., 1996). When the KB5-C20 Vβ domain is superimposed onto its 14.3.d and A6 counterparts, r.m.s. differences of 0.98 and 1.04 Å are obtained for 78 and 80 pairs of equivalent α carbons, respectively. The major differences are restricted to the CDR3 region and c″ strand. As shown in Figure 2B, the four residue long c″ strand found in the KB5-C20 Vβ domain is hydrogen-bonded to the d strand of the outer β-sheet. This folding topology is different from those of the Vβ8.2 and Vβ12.3 domains (Table III, Bentley et al., 1995; Garboczi et al., 1996; Garcia et al., 1996a), and reminiscent of the ones observed for the strand-switched Vα domains found in the 1934.4, 2C, A6 and KB5-C20 TCRs (Fields et al., 1995; Garboczi et al., 1996; Garcia et al., 1996a; this paper). Therefore, the switch of strand c″ from the inner to the outer β-sheet does not constitute an exclusive attribute of TCR Vα domains. Inspection of the crystal structure of the KB5-C20 TCR scFv (this paper), and of available full-length TCR β chain structures (Bentley et al., 1995; Garboczi et al., 1996), further indicates that the switch of the c″ strand observed in the KB5-C20 Vβ2 domain results neither from artefactual contacts with the Désiré-1 Fab fragment, nor from the absence of the TCR Cβ domain [in the structures solved by Bentley et al. (1995) and Garboczi et al. (1996), the Cβ domain does not contact this part of the Vβ domain]. Table 3. List of the β-sheet hydrogen bonds found between the c″-d and c′-c″ strands in the KB5-C20 Vα domain and in the KB5-C20, 14.3.d and A6 Vβ domains KB5-C20 Vα2.3 KB5-C20 Vβ2 14.3.d Vβ8.2 and A6 hVβ12.3 c″ d c″ d c′ c″ 54O N66 54O N69 56N O64 56N O67 48O N56 56O N64 56O N67 48N O56 58N O62 58N O65 46O N58 58O N62 58O N65 60N O63 60O N63 The Vα-Vβ interface In the KB5-C20 TCR scFv, the Vα and Vβ domains are related by a 175° rotation axis. The accessible surface areas buried at the Vα-Vβ interface are 1029 Å2 for Vα and 1067 Å2 for Vβ. Thus, a total surface area of 2096 Å2 is buried at the KB5-C20 interface, as opposed to values of 1160 and 1575 Å2 in the cases of the 2C and A6 Vα-Vβ interfaces, respectively (Garboczi et al., 1996; Garcia et al., 1996a). Such scattered values are mostly accounted for by the differential contribution of residues belonging to CDRs (mainly CDR3s), which is higher in the KB5-C20 and A6 TCRs than in the 2C TCR. Many of the contacts between the KB5-C20 Vα and Vβ domains are conserved in the 2C and A6 TCRs, as well as in Ig V domains (Chothia et al., 1985; Garboczi et al., 1996; Garcia et al., 1996a). Figure 3 shows the Vα-Vβ contacts found in the KB5-C20 TCR scFv. They include a pair of side chain-side chain hydrogen bonds between Q37α and Q37β, a symmetric pair of side chain-main chain hydrogen bonds involving Y35α-L106β and L104α-Y35β, and a hydrophobic core formed by Y35α, P43α, F89α, F106α, Y35β, L43β, L106β and F108β. Other interdomain contacts occur between residues I105α and W45β, as well as between the main chain oxygen of F106α and Q42β. Figure 3.Stereoscopic view of the KB5-C20 Vα-Vβ interface. The Vα and Vβ domains are on the left and right sides, respectively. Only side chains of residues involved in interdomain interactions are shown. Residues are colour-coded according to their chemical nature. Acidic residues are red, basic residues dark blue, polar residues light blue, hydrophobic residues yellow and aromatic residues purple. Download figure Download PowerPoint As discussed below, in the absence of peptide-MHC ligand, the KB5-C20 Vα and Vβ CDR3s protrude from the plane formed by the remaining CDRs and show a rather extensive contact surface area. A major feature of this CDR3α-CDR3β interface is the presence of an ionic interaction between R93α and E105β (Figure 3). Additional stabilization of this composite protrusion comes from several van der Waals contacts and two hydrogen bonds (Q95αNϵ2-W100βO and G69αN-S103βOγ). Composed of 13 residues, the KB5-C20 CDR3β loop contrasts with the shorter CDR3β loops found in the 2C and A6 Vβ domains (Figure 4), and interacts with CDR1α through a hydrogen bond between N30α and G101β. On the other hand, the KB5-C20 CDR3α loop interacts with CDR2β through a hydrogen bond between the main chain oxygen of R101α and the side chain of T48β. Comparison of our data with those reported by Garcia et al. (1996a) and Garboczi et al. (1996) indicates that the KB5-C20, 2C and A6 TCRs adopt very similar interdomain β-sheet packings, and suggests that neither the binding of the Désiré-1 Fab nor the absence of TCR C domains have significantly modified the native association of the V domains present in the KB5-C20 TCR scFv. Considering that some of the V segments used by these three TCRs share 80% identity in their amino acid sequences (Gahéry-Ségard et al., 1996). Comparison of their sequences shows the presence of conservative and non-conservative replacements within the CDRs (CDR1, positions 26, 28, 30; CDR2, positions 48, 50, 51, 54), as well as in the framework (positions 7, 16, 18, 19, 21, 36, 40, 42, 64, 72, 73, 78). Based on the three-dimensional structure of the Vα2.3 segment reported here, most of the residues varying within the mouse Vα2 subfamily could be assigned to lateral, solvent-exposed regions, and the few which are buried (positions 18, 36, 42, 64, 73), or located at the Vα-Vβ interface (position 48), should not modify the overall three-dimensional structure. Thus, the pattern of Vα folding is likely to be closely conserved within a given Vα subfamily. The TCR antigen-binding site As shown in Figure 5, the six CDRs cluster together at the membrane-distal tip of the KB5-C20 TCR scFv, the α and β HV4 loops lying at opposite ends of the binding site. Visual inspection of the recently published structure of the 2C TCR combining site (Garcia et al., 1996a) and rigid-body rotation of the KB5-C20 TCR scFv onto the A6 TCR structure reported by Garboczi et al. (1996) indicate that the relative positions of the CDRs, with the exception of CDR3β, are rather conserved in these three TCR antigen-binding sites (compare Figures 2 and 5 with Figure 7B in Garcia et al., 1996a). However, as expected from the fact that their respective CDR3β sequences lie at three scattered points of the size distribution observed for TCR β chain CDR3s (Candéias et al., 1991; Pannetier et al., 1993), the major differences between the structures of these TCR binding sites are confined to the CDR3β loops. As shown in Figure 5, the elongated KB5-C20 CDR3β loop packs tightly against the CDR3α loop without leaving an intervening hydrophobic pocket as found in both the A6 and 2C TCR binding sites (Garboczi et al., 1996; Garcia et al., 1996a). Moreover, in the absence of peptide-MHC ligand, the KB5-C20 CDR3α and CDR3β protrude from a plane formed by the remaining CDRs, thereby occupying a central position in the binding site. Figure 5.Depiction of the α-carbon backbone of the KB5-C20 Vα and Vβ domains. The following colour codes have been used: Vα in light brown, Vβ in light blue, CDR1s in green, CDR2s in red, CDR3s in blue and HV4s in violet. In this orientation, the antigen-binding site is viewed from above, as in Figure 7B in Garcia et al. (1996a). Download figure Download PowerPoint Figure 6.Stereo pairs depicting plausible contacts between the KB5-C20 TCR CDR loops (dark grey), a bound octapeptide (medium grey) and the H-2Kb α1 and α2 helices (light grey). In this orientation, the TCR-peptide-MHC complex is viewed from the side, so that the H-2Kbα2 helix is in the foreground and the α1 helix is behind the peptide. The interactions between the KB5-C20 TCR and its cognate peptide-MHC complex were modelled by rotating these molecules onto the positions occupied by their human counterparts found in the A6 TCR-Tax-HLA-A2 complex (Garboczi et al., 1996). These plausible interactions are depicted as dashed lines connecting the corresponding pairs of α-carbons. The interactions also present in the A6 TCR-Tax-HLA-A2 complex are shown as bold dashed lines, whereas those specific to the KB5-C20 TCR-peptide-H-2Kb complex are depicted as thin dashed lines. The configuration adopted by the peptide-MHC-unliganded CDR3β is shown as a dashed loop. Note that the interaction of residue Q95α with the second carbonyl group of the

Referência(s)