1 Introduction. Rajiv Khanna 1, Sharon L. Silins 1, Zhiping Weng 2, David Gatchell 2, Scott R. Burrows 1 and Leanne Cooper 1

Transcription

1 Eur. J. Immunol : MHC-peptide-TCR interactions in HLA cross-restriction 1587 Cytotoxic T cell recognition of allelic variants of HLA B35 bound to an Epstein-Barr virus epitope: influence of peptide conformation and TCR-peptide interaction Rajiv Khanna 1, Sharon L. Silins 1, Zhiping Weng 2, David Gatchell 2, Scott R. Burrows 1 and Leanne Cooper 1 1 Tumour Immunology Laboratory, EBV Unit, Queensland Institute of Medical Research and Joint Oncology Program, University of Queensland, Brisbane, Australia 2 Department of Biomedical Engineering, Boston University College of Engineering, Boston, USA Fine specificity analysis of HLA B35-restricted Epstein-Barr virus (EBV)-specific cytotoxic T lymphocyte (CTL) clones revealed a unique heterogeneity whereby one group of these clones cross-recognized an EBV epitope (YPLHEQHGM) on virus-infected cells expressing either HLA B*3501 or HLA B*3503, while another group cross-recognized this epitope in association with either HLA B*3502 or HLA B*3503. Peptide binding and titration studies ruled out the possibility that these differences were due to variation in the efficiency of peptide presentation by the HLA B35 alleles. Sequence analysis of the TCR genetic elements showed that these clonotypes either expressed BV12/AV3 or BV14/ADV17S1 heterodimers. Interestingly, CTL analysis with monosubstituted alanine mutants of the YPLHEQHGM epitope indicated that the BV12/AV3 + clones preferentially recognized residues towards the C terminus of the peptide, while the BV14/ADV17S1 + clones interacted with residues towards N terminus of the peptide. Molecular modelling of the MHC-peptide complexes suggests that the differences in two floor positions (114 and 116) of the HLA B35 alleles dictate different conformations of the peptide residues L3 and/or H7 and directly contribute in the discerning allele-specific immune recognition by the CTL clonotypes. These results provide evidence for a critical role for the selective interaction of the TCR with specific residues within the peptide epitope in the fine specificity of CTL recognition of allelic variants of an HLA molecule. Key words: Virus / Cytotoxic T lymphocyte / HLA allele / Peptide / TCR Received 16/12/98 Revised 25/1/99 Accepted 26/1/99 1 Introduction The interaction of T cells with the MHC-peptide complex is a critical step towards the initiation and propagation of specific immune responses. The specificity of this interaction is determined by two distinct components, namely MHC-restricted presentation of a peptide epitope, and a heterodimeric g cell surface protein called the TCR. T cells recognize a wide range of self and nonself antigenic determinants by rearrangement of variable (TCRAV and TCRBV), diversity (TCRBD) and joining (TCRAJ and TCRBJ) TCR gene segments, as well as N [I 19123] Abbreviations: CDR: Complementarity-determining region EBNA:EBV nuclear antigen LCL: Lymphoblastoidcell line(s) region diversity at the junctional regions. The molecular details of the interactions between the TCR and the MHC-peptide complex have been the focus of considerable interest over the last decade [1]. Earlier studies have used point mutations to explore the role of individual variable regions and complementarity determining regions (CDR) in recognition by T cells [2 3]. After considerable effort, Garcia et al. [4] and Garboczi et al. [5] have resolved the crystal structures of g TCR complexed with MHC-peptide ligands. In these structures, V CDR1 lies over the N terminus of the peptide, while V g CDR3 is aligned over the middle and C-terminal regions of the peptide. In addition, the CDR3 region of V and the CDR1 region of V g contact the middle and C-terminal region of the peptide, respectively. The CDR from both V and V g also contact the MHC helices. These results are consistent with the studies of peptide immunization and mutation experiments which predicted a diagonal WILEY-VCH Verlag GmbH, D Weinheim, /99/ $ /0

2 1588 R. Khanna et al. Eur. J. Immunol : configuration of the TCR with respect to the MHCpeptide complex [1, 3]. Immune recognition of peptide epitopes by specific T lymphocytes is also constrained by the polymorphism within MHC class I alleles expressed by the target cells. These constraints result from either: (1) variation in the surface of MHC class I molecules that directly interact with the TCR; (2) changes in the set of peptides that can be presented to the TCR; or (3) changes in conformation of bound peptide. Considerable work has been carried out to address the first two forms of constraints [6, 7], and studies using MHC or peptide variants have implied that conformational variation is also an important mechanism for T cell reactivity [8]. In the present study we have exploited a novel pattern of allelic heterogeneity in the immune recognition of virus-infected cells by a panel of CTL clones to reveal the structural mechanism of TCR interaction with a peptide epitope bound to allelic variants of an HLA molecule. Molecular analysis of these clones, expressing two distinct types of TCRBV and TCRAV gene rearrangements, revealed that immune recognition of target cells expressing different HLA B35 allelic variants was dependent upon the specific interactions of these TCR with individual residues within the peptide epitope. Moreover, modelling of the HLA-peptide complexes indicated that minor amino acid differences in the peptide binding groove of the HLA B35 allelic variants directly influence the conformation of the peptide epitope, which may contribute to the allele-specific immune recognition by the two T cell clonotypes. These results allow us to propose a hypothesis for the TCR-peptide interaction and its influence on allelic heterogeneity in T cell recognition. We believe that this hypothesis accommodates the emerging trends in TCR-ligand interaction and represents a general mechanism of T cell recognition of allelic variants of HLA molecules. 2 Results 2.1 YPLHEQHGM-specific CTL clones display heterogeneity in recognition of target cells expressing different allelic variants of HLA B35 Earlier studies from our laboratory have shown that EBVspecific CTL from HLA B35 + individuals consistently show a strong response to the YPLHEQHGM epitope encoded within the EBV nuclear antigen 3 (EBNA3) [9]. Fine specificity analysis of YPLHEQHGM-specific CTL clones revealed an interesting heterogeneity in recognition of virus-infected cells expressing natural HLA B35 variants. Representative data from five such CTL clones are shown in Fig. 1A E. CTL clones Cl.37, Cl.46 and Cl.51 strongly recognized HLA B* and HLA B* lymphoblastoid cell lines (LCL), while HLA B* LCL were poorly recognized. On the other hand, clones Cl.56 and Cl.18 lysed HLA B* and HLA B LCL with insignificant lytic activity against HLA Figure 1. HLA B35-restricted, EBV-specific CTL clones display heterogeneity in their recognition of virus-infected cells expressing allelic variants of HLA B Cr-labeled HLA B*3501 +, B* or B* LCL were exposed to CTL effectors at different E/T ratios. CTL clones used in this assay were Cl.56, Cl.46, Cl.37, Cl.51 and Cl.18 as indicated. The results are expressed as percent specific lysis. E/T ratios of 1:1, 2:1 and 4:1 were used in the assay.

3 Eur. J. Immunol : MHC-peptide-TCR interactions in HLA cross-restriction 1589 B* LCL. To further confirm this heterogeneity, these CTL clones were also tested against synthetic peptidesensitized PHA blasts from HLA B*3501 +, B* or B* individuals. Data presented in Fig. 2A E clearly show a similar pattern of heterogeneity in CTL lysis of peptide-sensitized PHA blasts. CTL clones Cl.37, Cl.46 and Cl.51 recognized YPLHEQHGM-coated PHA blasts from HLA B* and HLA B* donors very efficiently, while CTL clones Cl.56 and Cl.18 lysed peptidesensitized HLA B* and HLA B PHA blasts, with very low levels of lytic activity against HLA B* PHA blasts coated with the peptide. 2.2 Variable peptide binding does not account for heterogeneity in CTL recognition One possible explanation for the data presented above is that the HLA B35 alleles bind the YPLHEQHGM peptide with different efficiencies. To explore this possibility, we assayed MHC stabilization on T2 transfectants expressing HLA B*3501, B*3502 or B*3503 cdna using serially diluted YPLHEQHGM peptide. T2 transfectants were incubated with the peptide at 26 C and the levels of stable MHC expression were assessed with a conformation-sensitive mab [10]. Data from one such experiment are presented in Fig. 3. T2 transfectants expressing three different subtypes of the HLA B35 allele showed almost identical patterns of MHC stabilization over a range of peptide concentrations. Moreover, similar results were also obtained using a modified stabilization assay that took into account the peptide off rates (by thoroughly washing the peptide away before a 4-h incubation at 37 C; data not shown). These results strongly suggest that heterogeneity in CTL recognition is unlikely due to variable binding of the YPLHEQHGM peptide by these HLA B35 alleles. 2.3 HLA B35 subtype discrimination by CTL clones is due to distinct TCR rearrangement To investigate the TCR repertoire used by the YPLHEQHGM-specific T cells, a panel of CD8 + clones displaying heterogeneity in CTL recognition was analyzed for TCRA (V-J-C) and TCRB (V-D-J-C) rearrangements. These rearrangements were identified using the single-strand ligation to single-stranded cdna technique (SLIC) [11], followed by direct sequencing of SLICgenerated PCR products. Alignment of junctional TCRA and TCRB nucleotide and polypeptide sequences (Fig. 4) revealed that each clone expressed one of two distinct TCR g heterodimers. CTL clones recognizing HLA B* and B* targets shared common TCRA and TCRB genetic rearrangements (BV12/BJ2S1 and AV3/ AJ9S7), while all CTL clones recognizing HLA B* and B* targets expressed BV14/BJ1S6 and ADV17S1/AJ17S8 heterodimers. Apart from a few con- Figure 2. Discerning cytotoxic T cell reactivity by YPLHEQHGM-specific CTL clones against peptide sensitized PHA blasts expressing either the HLA B*3501, HLA B*3502 or HLA B*3503 allele. 51 Cr-labeled PHA blasts were presensitized with varying concentrations of peptides (shown on the x-axis) and then exposed to CTL effectors. CTL clones used in this assay were Cl.56, Cl.46, Cl.37, Cl.51 and Cl.18 as indicated. An E/T ratio of 4:1 was used in the assay.

4 1590 R. Khanna et al. Eur. J. Immunol : CTL analysis with monosubstituted alanine mutants of the YPLHEQHGM peptide reveals heterogeneity in TCR interactions with each MHC-peptide complex Figure 3. MHC stabilization analysis on the T2 cell line lacking the transporter associated with antigen presentation (TAP), transfected with cdna encoding either HLA B*3501, HLA B*3502 or HLA B*3503. T2 transfectants were initially incubated with 200? l of serially diluted YPLHEQHGM peptide (shown on the x-axis) for h at 26 C followed by incubation at 37 C for 2 3 h. HLA B35 expression on these cells was analyzed by FCM using SFR8-Bw6 antibody. Data are expressed as relative fluorescence intensity. served amino acids staggered between the non-germline-encoded regions of the CDR3 loops in both chains of these two TCR, there was no apparent conservation of CDR3 amino acid composition. To investigate the molecular interactions of different MHC-peptide complexes with these TCR, CTL clones were screened for lysis of target cells sensitized with monosubstituted alanine mutants of the YPLHEQHGM peptide. Our initial experiments with alanine mutants and the wild-type peptide indicated that a peptide concentration of 0.001? g/ml was the most appropriate for delineating subtle effects on the interaction of each TCR with the peptide bound to the HLA B35 alleles (data not shown). PHA blasts from HLA B*3501 +, B* or B* individuals were presensitized with either YPL- HEQHGM or the alanine mutant peptides (0.001? g/ml for 1 h). Following incubation, excess peptide was washed off and the cells were exposed to either the BV12/AV3 + or the BV14/ADV17S1 + CTLclonesinastandard 51 Cr-release assay. Representative data from one of these assays are presented in Fig. 5A and B. The level of CTL lysis of target cells coated with the alanine mutants was compared with that of the wild-type YPLHEQHGM peptide. Both BV12/AV3 + CTL clones (which recognize HLA B* and B* targets) and BV14/ADV17S1 + CTL clones (which recognize HLA B* and B* targets) showed significant loss of lytic activity following replacement of residues H4, E5, Q6 or M9, while much less pronounced effects were noted when Y1, P2 or G8 residues were substituted with alanine. However, a clear Figure 4. TCR and TCR g chain sequences of YPLHEQHGM-specific CTL clones that recognize target cells expressing B*3502/B*3503 (A) or B*3501/B*3503 (B). The deduced amino acid sequence of the CDR3 loop, defined according to Chothia et al. [29], is shown putatively supported by two framework branches (FW), TCRV gene segments are assigned according to Arden et al. [25] and TCRAJ and TCRBJ gene segments are assigned according to Moss et al. [30] and Toyonaga et al. [31], respectively. The TCRAJ and TCRBJ germ-line sequences are underlined.

5 Eur. J. Immunol : MHC-peptide-TCR interactions in HLA cross-restriction 1591 Figure 5. CTL analysis of YPLHEQHGM-specific clones expressing distinct TCR genetic elements, using monosubstituted alanine mutants of the YPLHEQHGM peptide. PHA blasts from HLA B*3501 +, B* or B* donors were presensitized with either the YPLHEQHGM peptide or individual monosubstituted alanine mutants (shown on the x-axis; peptide concentration: 0.001? g/ml) for 1 h at 37 C. Following incubation, excess peptide was washed off and the cells were exposed to either the BV12/AV3 + (Cl.56; Fig. 5A) or the BV14/ADV17S1 + (Cl.46; Fig. 5B) CTL clones. An E/T ratio of 4:1 was used in the assay. The results are expressed as relative percent specific lysis by comparison with the level of lysis with the wild-type peptide. distinction was noticed between these two sets of clones when alanine was substituted at positions L3 and H7. The BV12/AV3 + CTL clones showed negligible lysis of target cells coated with the YPLHEQAGM mutant, while substitution of the L3 residue with alanine had a much less dramatic effect on CTL lysis (Fig. 5A). On the other hand, CTL activity of the BV14/ADV17S1 + CTL clones was significantly reduced by replacing the L3 residue with alanine, and very little effect was seen with the H7 1 A mutation (Fig. 5B). These results strongly suggest that immune recognition by the BV12/AV3 + clones depends on interaction with residues towards the C terminus of the peptide, while the BV14/ADV17S1 + clones interact more closely with residues towards the N terminus of the peptide. 2.5 Molecular modelling of the HLA B35 allelic variants complexed with the YPLHEQHGM peptide To provide more insight into the B35-peptide-TCR interactions, theoretical models of HLA B*3501, B*3502 and B*3503 complexed with the YPLHEQHGM peptide were constructed. Sequence comparison of the HLA B35 alleles revealed that HLA B*3502 differs from B*3501 at three positions: 109 (L 1 F), 114 (D 1 N) and 116 (S 1 Y), while B*3503 differs from B*3501 only at position 116 (S 1 F). Overlapping models for these allelic variants are shown in Fig. 6A. It is clear that positions 114 and 116 lie on the floor of the peptide binding groove and thus can directly affect peptide binding. Position 109 lies at the

6 1592 R. Khanna et al. Eur. J. Immunol : Figure 6. Molecular modelling of the HLA B35 allelic variants and the YPLHEQHGM peptide conformations bound to individual HLA B35 alleles. Side chain conformational searches for HLA B35 subtypes (B*3501, B*3502 and B*3503) and the backbone and side chain conformational searches for the YPLHEQHGM peptide were performed by using CONGEN. (A) shows the top view of overlapping models of HLA B*3501 (blue), B*3502 (red) and B*3503 (yellow). Amino acid differences at positions 109 (L for B*3501 and B*3503 and F for B*3502), 114 (D for B*3501 & B*3503 and N for B*3502) and 116 (S for B*3501, Y for B*3502 and F for B*3503) amongst these allelic variants are also shown. Side chain conformation of position 97 (R for all B35 allele) is shown in this figure since it is adjacent to positions 114 and 116 and its conformation is highly variant among different MHC molecules. Overlapping backbone conformations of the YPLHEQHGM peptide bound to B*3501 (blue), B*3502 (green) and B*3503 (yellow) are shown in (B). (C, E and G) show the side view of peptide conformations bound to B*3501, B*3502 and B*3503 respectively, while the top view of peptide YPLHEQHGM bound to these alleles is shown in (D, F and H).

7 Eur. J. Immunol : MHC-peptide-TCR interactions in HLA cross-restriction 1593 edge of the beta floor and beneath the 2 helix (Fig. 6A). It is in direct contact with neither the peptide nor the TCR. It is possible that alterations at this position can displace the MHC 2-helix and affect TCR binding indirectly. However, comparison of crystal structures of different MHC alleles ruled out this possibility. HLA A*0201 and A*6801 have amino acid F at position 109, and B*0801, B*3501 and B*5301 have L. The superposition of the crystal structures of these MHC molecules revealed that the alpha2-helix occupies the same position regardless which residue is at position 109. We therefore conclude that the presence of F (for B*3502) or L (for B*3501 and B*3503) at position 109 is unlikely to influence TCR binding. Our modelling strategy for peptide conformations was to place the terminal two positions (1 and 9) of the peptide first and carry out backbone and side chain conformational searches for the remaining peptide positions. For each HLA B35 subtype, the peptide conformations, with the lowest molecular energy calculated using CONGEN, were considered. These potential structures were subsequently compared with the functional data and finally three peptide conformations, one each of HLA B*3501, B*3502 and B*3503, were selected. These are shown in Fig. 6B H. Overlapping models of the backbone of the YPLHEQHGM peptide bound to the allelic variants showed significant differences in conformations (Fig. 6B). Differences were further revealed by analysis of side chain conformations of the peptide. YPLHEQHGM peptide bound to B*3501 (side view shown in Fig. 6C; top view: Fig. 6D) indicates that residues H4 and H7 point upwards, while E5 and Q6 point sideways and are thus accessible to the BV12/AV3 TCR. In contrast to this conformation, the YPLHEQHGM peptide bound to B*3502 (side view shown in Fig. 6E, top view Fig. 6F) clearly shows that in addition to E5 and Q6, the L3 residue is also pointing upwards, while the H7 residue is not accessible for TCR interaction. These models are in accordance with data presentd in Fig. 5, which clearly showed that the BV12/AV3 TCR recognizes H7 as a critical residue, while residue L3 is important for the BV14/ ADV17S1 TCR interaction. Thus differential orientation of the H7 and L3 residues in the B*3501 and B*3502- peptide complexes may control the heterogeneity in CTL recognition pattern. The conformation of the YPL- HEQHGM peptide bound to B*3503 (side view shown in Fig. 6G, top view Fig. 6H) shows both the L3 and H7 residues pointing upwards, thus CTL clones expressing either BV12/AV3 or BV14/ADV17S1 can efficiently recognize target cells presenting the HLA B*3503- YPLHEQHGM complex. 3 Discussion By exploiting a novel pattern of allelic heterogeneity in the immune recognition of virus-infected cells by a panel of CTL clones, we have studied molecular principles of TCR interaction with a peptide epitope bound to naturally occurring variants of a single HLA molecule. Molecular characterization of HLA B35-restricted CTL clones specific for an EBV epitope (YPLHEQHGM) from the EBNA3 protein revealed an interesting pattern of heterogeneity with respect to their recognition of virus-infected cells expressing allelic variants of the HLA B35 molecule. These clones recognized virus-infected cells expressing either HLA B*3501/HLA B*3503 or HLA B*3502/HLA B*3503 alleles. This heterogeneity was also noted when peptide YPLHEQHGM was added exogenously on EBVnegative target cells. One of the possible explanations for these observations was that individual allelic variants of the HLA B35 molecule differed in their efficiency in binding the peptide. However, peptide binding assays with T2 transfectants expressing the different subtypes of HLA B35 showed very little difference in their binding efficiency. Moreover, we also ruled out the possibility that peptide off rates may contribute to this heterogeneity. A number of studies have used mutagenesis to illustrate the critical role of the variable regions, CDR or individual TCR residues in recognition by T cells [5, 6, 12]. Moreover, crystal structures of g TCR have shown that the CDR which span the junctional regions interact directly with peptide epitope bound to MHC molecules [4, 5]. Thus it is possible that both TCR rearrangement and peptide conformation may contribute to the discriminatory patterns of CTL reactivity. Indeed, TCR sequence analysis of CTL clones revealed two distinct g rearrangements that showed no apparent conservation of either the TCRBV, TCRAV, TCRDJ, TCRAJ or CDR3 regions. Such dramatic differences in TCR rearrangements for CTL clones recognizing the same peptide epitope are intriguing, especially in the context of the earlier mutagenesis studies which showed a limited alteration in TCR usage following immunization with variant peptides [1, 3]. In our study, it is highly unlikely that these differences can be attributed to the interaction of the TCR with MHC residues on the surface of the three allelic variants of HLA B35 molecules. Since most of the amino acid variations within these B35 alleles are concentrated in the peptide binding groove, it is possible that these structurally distinct TCR rearrangements have been selected to accommodate different peptide conformations. Having established that the CTL clones displaying allelic heterogeneity in fine specificity expressed distinct TCR

8 1594 R. Khanna et al. Eur. J. Immunol : genetic elements, we designed our next set of experiments to delineate the precise mode of interaction of the TCR with the YPLHEQHGM peptide bound to individual HLA B35 subtypes. Another important aim of these experiments was to explore the potential role of peptide conformational differences in the heterogeneity of CTL specificity. To address these issues we adopted an approach which involved testing CTL activity against HLA B*3501 +, B* or B* target cells sensitized with monosubstituted alanine mutants of the YPL- HEQHGM peptide. A clear distinction was noticed when alanine mutants for peptide positions 3 and 7 were compared. Replacement of H7 with alanine completely abolished CTL reactivity for the BV12/AV3 + CTL clones, while lytic activity of the BV14/ADV17S1 + clones was significantly reduced following substitution of the L3 residue with alanine. Based on these results we concluded that the two groups of CTL clones under analysis differed considerably in their mode of interaction with individual residues within the peptide epitope. One of the important questions that still remained unanswered was to determine why these CTL clones show selective recognition of some allelic variants of HLA B35, while other variants bound to the same peptide are not recognized. To address this question we constructed theoretical models of the YPLHEQHGM peptide bound to the HLA B35 allelic variants using homology models. In accordance with our functional data with the CTL clones, each HLA B35 allelic variant showed unique conformations for the YPLHEQHGM peptide. One critical difference in peptide conformation with respect to the HLA B*3501 and B*3502 alleles was the orientation of L3 and H7. Residue H7 points upwards and away from the peptide binding groove when YPLHEQHGM associates with B*3501, while the L3 residue is not accessible for TCR interaction. In contrast, orientation of these residues is completely reversed when this peptide associates with the B*3502 allele. These models strongly support the conclusions drawn from the CTL analysis with monosubstituted alanine mutant peptides. Interestingly, the conformation of peptide YPLHEQHGM bound to B*3503 shows both L3 and H7 residues pointing upwards and away from the peptide binding groove, thus allowing efficient interaction with either the BV12/ AV3 or the BV14/ADV17S1 TCR. Based on these theoretical models and recent crystallographic analyses of the MHC-peptide-TCR complexes [4, 5], it is tempting to speculate that the diagonal orientation of the TCR with respect to the MHC-peptide complex directly contributes towards the heterogeneity in T cell recognition described in this report. Thus, the BV14/ADV17S1 and BV12/AV3 TCR appear to be oriented to interact with either N terminus or C terminus peptide residues, respectively, allowing discrimination between the B*3501 and B*3502 alleles. On the other hand, the conformation of YPLHEQHGM peptide bound to the B*3053 allele permits unbiased CTL recognition by both CTL clonotypes. Taken together, it appears that the heterogeneity in CTL recognition of peptide bound to HLA B35 allelic variants is dependent upon (1) the TCR genetic elements expressed by the clones and (2) the selective interaction of the TCR heterodimers with specific residues within the peptide epitope. 4 Materials and methods 4.1 Establishment and maintenance of cell lines LCL were established by exogenous transformation of peripheral B cells [13]. HLA B*3501 +, B* and B* LCL were transformed with EBV derived from the B95.8 cell line which encodes the epitope YPLHEQHGM [9, 14]. HLA B*3501 +, B* and B* PHA-stimulated blasts were prepared as described earlier [13, 14]. In addition, the mutant LCL T-lymphoblastoid hybrid cell line, 174 CEM.T2 (referred to as T2 cells) [15] expressing either HLA B*3501 (T2.B*3501), B*3502 (T2.B*3502) or B*3503 (T2.B*3503) were also used in the study. The T2.B*3501 cell line has been described elsewhere [16]. T2.B*3502 and T2.B*3503 cell lines were established as follows. B*3503 cdna was excised from puc18 plasmid [17] (a kind gift from Dr. Leo Satz; Laboratorio de Immunogenetica, Argentina) using SalI and HindIII restriction enzymes. HLA B*3502 cdna was constructed using Quick Change site-directed mutagenesis (Stratagene) according to the manufacturer s instructions. Essentially, HLA B*3503 cdna was mutated at amino acid positions 109 (L 1 F), 114 (D 1 M) and 116 (F 1 Y) and then confirmed by automated sequencing (ABI 377 DNA Sequenator). Both HLA B*3502 and HLA B*3503 cdna were cloned into the selectable mammalian expression vector pcdna3.1 ( ) (Invitrogen). These expression vectors were transfected into T2 cells as previously described [18]. T2 cells expressing these genes were selected by flow cytometry using SFR8-Bw6, anti-hla Bw6 mab [19]. YPLHEQHGM-specific CTL clones were generated from donor NB (HLA A2, A24, B7, B*3503) by agar cloning as previously described [13] following initial stimulation with either + -irradiated (8000 rad) autologous LCL or autologous PBMC precoated with peptide YPLHEQHGM (1? Mfor1h)usinga responder:stimulator ratio of 50:1 or 5:1, respectively. Antigen specific of these CTL clones was confirmed by standard 51 Cr-release assay using recombinant vaccinia virusinfected and YPLHEQHGM-sensitized target cells [20]. 4.2 MHC stabilization assays To assess MHC binding by the HLA B35-restricted epitope, YPLHEQHGM and alanine mutants, T2.B*3501, T2.B*3502

9 Eur. J. Immunol : MHC-peptide-TCR interactions in HLA cross-restriction 1595 and T2.B*3503 cells ( ) were incubated with 200? lof twofold dilutions of each of the peptides (0.0002? g/ ml 20? g/ml) at 26 C for h, followed by incubation at 37 C for 2 3 h. After the incubations, HLA B35 expression was measured by FCM using a HLA Bw6 mab (SFR8 Bw6). 4.3 Cytotoxicity assay CTL clones were tested in duplicate for cytotoxicity in the standard 5-h 51 Cr-release assay. Where synthetic peptide was involved, it was added directly to 51 Cr-labeled targets and incubated for 1 h before excess unbound peptide was washed off. Peptides were synthesized by Chiron Mimotopes (Chiron Corp., Emeryville, CA) on a 1-mg scale using Pin-Technology [21]. 4.4 RNA isolation and cdna synthesis Total RNA was extracted from CTL using Total RNA Isolation Reagent (Advanced Biotechnologies, London, GB). Antisense TCRBC (C b1 ) and TCRAC (C a1 )primers were used to generate first-strand cdna from 2 5? gof total RNA, and this was followed by RNA hydrolysis, the removal of excess primer, and ligation of an anchor oligonucleotide (5 P-CACGAATTCACTATCGATTCTGGAACCTTCA- GAGG-NH 3 3 ) to the 3 end of the cdna as described previously [11]. 4.5 Amplification of rearrangement TCRA and TCRB sequences TCRA and TCRB sequences were amplified by PCR using a primer complementary to the anchor and nested TCRAC or TCRBC genes. Primer sequences and amplification conditions employed were same as described previously [11]. The clonality of the CTL was further confirmed by TCRAV and TCRBV family-specific PCR. TCRBV sequences were amplified as previously described [22]. TCRAV sequences were amplified with one of 32 5 TCRAV family-specific oligonucleotides (V 1-32) and a 3 TCRAC (C ) constant primer. The oligonucleotides V 1-12/V 17-18/C and V 13-16/V 22-29/V 32 were synthesized according to Davies et al. [23] and Steinle et al. [24], respectively. The V 15, 16 and 17 primers reported in Davies et al. [23] were reassigned to V 19, 20 and 21, respectively, in accordance with the new nomenclature or Arden et al. [25]. TCRAV families 30 and 31 were amplified with the following oligonucleotides: V 30: 5 CTTCACCCTGTATTCAGCTGGG 3 ; and V 31: 5 CTG- CAGCTTCTTCAGAGAGAGACAATGG 3. Amplification of the ADV17S1 sequence was performed using a 5 primer (5 - GGATTCCAATTATAAACTGTGCTT-3 ) and the 3 TCRAC (C ) primer. All amplification conditions were the same as those described above. 4.6 Nucleotide sequencing QiAEX-recovered PCR products were sequenced in both directions with a PRISM Ready Reaction DyeDeoxy Terminator Cycle Sequencing Kit and a 373A DNA sequencer (Applied Biosystems Inc., Foster City, CA) as described previously [11]. 4.7 Computer assisted modelling of the MHC-peptide complexes Coordinates for the crystal structures of HLA B*3501 (PDB code 1A1N) and B*5301 (PDB code 1A1O) were kindly provided by Dr. Yevonne Jones, Oxford University. In the crystal structure 1A1N, B*3501 was bound to an eight-residue-long peptide [26, 27]. However, the binding modes of eightresidue-long peptides are very different from those nine residues long. Since B*3501 and B*5301 have very similar sequences, and both bind peptides with proline at the anchor position 2, we used the N-terminal coordinates of the nine-residue-long LS6 peptide in the B*3501 crystal structure (1A1O) as the N-terminal position of the EBV peptide for the three B35 allelic variants. The C-terminal positions of the crystal structures 1A1O, 1VAD (H-2K b ) and 1HOC (H-2D b ), which have either M or F, were tried for the placement of the YPLHEQHGM peptide into the three B35 grooves, and the position providing the lowest binding energy was adopted. After the N- and C-terminal coordinates of the peptide had been determined, the side chain conformations of three positions (97, 114, 116) of the HLA B35 subtypes (B*3501, B*3502, and B*3503) and the backbone and side chain conformations of YPLHEQHGM were searched using CONGEN [28]. This program is designed to carry out exhaustive searches for the specified degrees of freedom and rank the conformations according to their molecular mechanics energy, including van der Waals, electrostatics, bond and angles according to CHARMM19 force field. The van der Waals cutoff was set to 15 kcal/mol and any conformations with higher van der Waals energy were discarded. For each HLA B35 subtype, 20 MHC-peptide complex structures with the lowest molecular mechanics energy calculated using CONGEN were retained and energy minimization was carried out to remove any van der Waals clashes. Finally, peptide structures which showed the strongest agreement with the CTL functional analysis were selected for each B35 subtype. Acknowledgments: This work was supported by grants from National Health and Medical Research Council (NHMRC), Australia. RK is supported by an R. Douglas Wright Fellowship from the NHMRC, Australia. We would like to thank Prof. D. J. Moss and Prof. Charles Delisi for valuable advice and continuing support.

10 1596 R. Khanna et al. Eur. J. Immunol : References 1 Davis, M. M., Boniface, J. J., Reich, Z., Lyons, D., Hampl,J.,Arden,B.andChien,Y.,Ligand recognition by alpha beta T cell receptors. Annu. Rev. Immunol : Patten, P. A., Rock, E. P., Sonoda, T., Fazekas de St. Groth, B., Jorgensen, J. L. and Davis, M. M., Transfer of putative complementarity-determining region loops of T cell receptor V domains confers toxin reactivity but not peptide/mhc specificity. J. Immunol : Sant Angelo, D. B., Waterbury, G., Preston-Hurlburt, P.,Yoon,S.T.,Medzhitov,R.,Hong,S.C.andJaneway, C. A., Jr., The specificity and orientation of a TCR to its peptide-mhc class II ligands. Immunity : Garcia, K. C., Degano, M., Stanfield, R. L., Brunmark, A., Jackson, M. R., Peterson, P. A., Teyton, L. and Wilson, I. A., An alpha-beta T cell receptor structure at 2.5 angstrom and its orientation in the TCR-MHC complex. Science : Garboczi, D. N., Ghosh, P., Utz, U., Fan, Q. R., Biddison, W. E. and Wiley, D. C., Structure of the complex between human T-cell receptor, viral peptide and HLA- A2. Nature : Ajitkumar, P., Geier, S.S., Kesari, K. V., Borriello, F., Nakagawa,M.,Bluestone,J.A.,Saper,M.A.,Wiley, D. C. and Nathenson, S. G., Evidence that multiple residues on both the alpha-helices of the class I MHC moleculearesimultaneouslyrecognizedbythet-cellreceptor. Cell : Grandea, A. G. and Bevan, M. J., Single-residue changes in class I major histocompatibility complex molecules stimulate responses to self peptides. Proc. Natl. Acad. Sci. USA : Dyall, R., Fremont, D. H., Jameson, S. C. and Nikolic- Zugic, J., T cell receptor (TCR) recognition of MHC class I variants: intermolecular second-site reversion provides evidence for peptide/mhc conformational variation. J. Exp. Med : Burrows, S.R., Gardner, J., Khanna, R., Steward, T., Moss, D. J., Rodda, S. and Suhrbier, A., Five new cytotoxic T cell epitopes identified within Epstein-Barr virus nuclear antigen 3. J. Gen. Virol : Burrows,J.M.,Burrows,S.R.,Poulsen,L.M.,Sculley, T. B., Moss, D. J. and Khanna, R., Unusually high frequency of Epstein-Barr virus genetic variants in Papua New Guinea that can escape cytotoxic T-cell recognition: implications for virus evolution. J. Virol : Argaet, V. P., Schmidt, C. W., Burrows, S. R., Silins, S. L., Kurilla, M. G., Doolan, D. L., Suhrbier, A., Moss, D. J., Kieff, E., Sculley, T. B. and Misko, I. S., Dominant selection of an invariant T-cell antigen receptor in response to persistent infection by Epstein-Barr virus. J. Exp. Med : Manning, T. C., Schlueter, C. J., Brodnicki, T. C., Parke, E. A., Speir, J. A., Garcia, K. C., Teyton, L., Wilson, I. A. and Kranz, D. M., Alanine scanning mutagenesis of an alpha beta T cell receptor: mapping the energy of antigen recognition. Immunity : Moss, D. J., Misko, I. S., Burrows, S. R., Burman, K., McCarthy, R. and Sculley, T. B., Cytotoxic T-cell clones discriminate between A- and B-type Epstein-Barr virus transformants. Nature : Burrows, S. R., Khanna, R., Burrows, J. M. and Moss, D. J., An alloresponse in humans is dominated by cytotoxic T lymphocytes (CTL) cross-reactive with a single Epstein-Barr virus CTL epitope: implications for graftversus-host disease. J. Exp. Med : Salter,R.D.andCresswell,P.,Impaired assembly and transport of HLA-A and -B antigens in a mutant TxB cell hybrid. EMBO J : Takiguchi, M., Kawaguchi, G., Sekimata, M., Hiraiwa, M., Kariyone, A. and Takamiya, Y., The role of the conserved residue in pocket A and the polymorphic residue in pocket E of HLA-B*3501 in presentation of human minor histocompatibility peptides to T cells. Int. Immunol : Satz, M. L., Fernandez-Vina, M., Theiler, G. C., Marcos, Y. C., Lindel, N., Capucchio, M., Gorodezky, C., Fainboim, L. and Stastny, P., Allelic heterogeneity of HLA-B35 subtypes in different populations as assessed by DNA typing. Tissue Antigens : Khanna, R., Burrows, S. R., Argaet, V. and Moss, D. J., Endoplasmic reticulum signal sequence facilitated transport of peptide epitopes restores immunogenicity of an antigen processing defective tumour cell line. Int. Immunol : Radka,S.F.,Kostyu,D.D.andAmos,D.B.,A monoclonal antibody directed against the HLA-Bw6 epitope. J. Immunol : Khanna, R., Burrows, S. R., Kurilla, M. G., Jacob, C. A., Misko, I. S., Sculley, T. B., Kieff, E. and Moss, D. J., Localization of Epstein-Barr virus cytotoxic T-cell epitopes using recombinant vaccinia: implications for vaccine development. J. Exp. Med :

11 Eur. J. Immunol : MHC-peptide-TCR interactions in HLA cross-restriction Valerio, R. M., Benstead, M., Bray, A. M., Campbell, R.A.andMaeji,N.J.,Synthesis of peptide analogues using the multipin peptides synthesis method. Anal. Biochem : Silins, S. L., Cross, S. M., Elliott, S. L., Pye, S. J., Burrows,J.M.,Moss,D.J.andMisko,I.S.,Selection of a diverse TCR repertoire in response to an Epstein-Barr virus-encoded transactivator protein BZLF1 by CD8 + cytotoxic T lymphocytes during primary and persistent infection. Int. Immunol : Davies, T. F., Martin, A., Concepcion, E. S., Graves, P., Cohen, L. and Ben-Nun, A., Evidence of limited variability of antigen receptors on intrathyroidal T cells in autoimmune thyroid disease. N. Engl. J. Med : Steinle, A., Reinhardt, C., Jantzer, P. and Schendel, D. J., In vivo expansion of HLA-B35 alloreactive T cells sharing homologous T cell receptors: evidence for maintenance of an oligoclonally dominated allospecificity by persistent stimulation with an autologous MHC/ peptide complex. J. Exp. Med : Arden, B., Clark, S. P., Kabelitz, D. and Mak, T. W., Mouse T-cell receptor variable gene segment families. Immunogenetics : Smith, K. J., Reid, S. W., Harlos, K., McMichael, A. J., Stuart,D.I.,Bell,J.I.andJones,E.Y.,Bound water structure and polymorphic amino acids act together to allow the binding of different peptides to MHC class I HLA-B53. Immunity : Smith, K. J., Reid, S. W., Stuart, D. I., McMichael, A. J.,Jones,E.Y.andBell,J.I.,An altered position of the alpha 2 helix of MHC class I is revealed by the crystal structure of HLA-B*3501. Immunity : Bruccoleri, R. E. and Karplus, M., Prediction of the folding of short polypeptide segments by uniform conformational sampling. Biopolymers : Chothia, C., Boswell, D. R. and Lesk, A. M., The outline structure of the T-cell alpha beta receptor. EMBO J : Moss, P. A. H., Rosenberg, W. M. C., Zintzaras, E. and Bell, J. I., Characterization of the human TCR alpha-chain repertoire and demonstration of a genetic influence on Valpha usage. Eur. J. Immunol : Toyonaga, B., Yoshikai, Y., Vadasz, V., Chin, B. and Mak, T. W., Organization and sequences of the diversity, joining, and constant region genes of the human T- cell receptor beta chain. Proc. Natl. Acad. Sci. USA : Correspondence: Rajiv Khanna, Queensland Institute of Medical Research, Bancroft Centre, 300 Herston Rd, Brisbane, Australia 4029 Fax: rajivk qimr.edu.au