The Cys-67 Residue of HLA-B27 Influences Cell Surface Stability, Peptide Specificity, and T-cell Antigen Presentation*

Cys-67 of HLA-B27 is located in the B pocket, which determines peptide-binding specificity. We analyzed effects of the Cys-67 3 Ser mutation on cell surface ex- pression, peptide specificity, and T-cell recognition of HLA-B*2705. Surface expression was assessed with antibodies recognizing either native or unfolded HLA proteins. Whereas native B*2705 molecules predominated over unfolded ones, this ratio was reversed in the mutant, suggesting lower stability. Comparison of B*2705-and Cys-67 3 Ser-bound peptides revealed that the mu- tant failed to bind (cid:1) 15% of the B*2705 ligands, while binding as many novel ones. Two peptides with Gln-2 found in both B*2705 and Cys-67 3 Ser are the first demonstration of natural B*2705 ligands lacking Arg-2. Other effects of the mutation on peptide specificity were: 1) average molecular mass of natural ligands higher than for B*2705, 2) bias against small residues at peptide position (P) 1, and 3) increased P2 permissiveness. The results suggest that the Cys-67 3 Ser mutation weakens B pocket interactions, leading to decreased stability of the mutant-peptide complexes. This may be partially below. C1R cell lines were cultured in Dulbecco’s modified Eagle’s medium supplemented with 7.5% FBS (both from Life Technologies).RMA-Sisa TAP-deficient murine cell line (24, 25). RMA-S transfec- tant cells expressing B*2705 and human (cid:2) 2m have been described and 50- (cid:1) l fractions were collected. Mass Spectrometry (MS) Analysis and Sequencing— peptide composition of HPLC fractions MALDI-TOF MS Peptides— Peptides HPLC. correct molecular

The molecular basis for the very strong association of HLA-B27 with ankylosing spondylitis (AS) 1 (1), reactive arthritis (2), and other spondyloarthropathies remains unknown. Among the pathogenetic hypotheses proposed, a classical one assumes that cytotoxic T lymphocytes (CTL) activated in response to external antigen would eventually cross-react with a self-peptide constitutively presented by HLA-B27 (3). This hypothesis is supported by much indirect evidence (4 -8). Alternatively, HLA-B27 heavy chain homodimers, whose formation in vitro is dependent on an unpaired Cys-67 residue (9), were suggested to act as noncanonical peptide presenting molecules, perhaps stimulating abnormal T-cell responses (10,11). It has also been proposed that HLA-B27 association to AS might be unrelated to peptide presentation, but rather to B27 heavy chain misfolding and triggering of endoplasmic reticulum stress responses (12)(13)(14).
The Cys-67 residue of HLA-B27 is an integral part of its B pocket, which determines the specificity of the molecule for peptides with Arg-2 (15). This residue plays a critical role in controlling the thermodynamic stability of soluble HLA-B27peptide complexes (16). This same study claimed that mutation of Cys-67 did not significantly affect the peptide binding properties of HLA-B27, but this conclusion was based on the affinity of a single HLA-B27 ligand and a few peptide analogs.
The pathogenetic role of Cys-67 was tested on HLA-B27 transgenic rats. In this system, HLA-B27 triggers a spontaneous inflammatory disease with many features of human spondyloarthropathies (17). Rats transgenic for the Cys-67 3 Ser (C67S) mutant showed a disease phenotype similar to wild type B*2705 transgenic rats of the same strain, suggesting that Cys-67 has little influence on disease (18).
In this study we have addressed the role of Cys-67 on cell surface expression, peptide specificity, and T-cell recognition of HLA-B27. We demonstrate that the C67S mutation decreases the stability of HLA-B27-peptide complexes at the cell surface, induces distinct alterations in the endogenous HLA-B27-bound peptide repertoire, and influences T-cell allorecognition.

MATERIALS AND METHODS
Cell Lines and DNA-mediated Gene Transfer-HMy2.C1R (C1R) is a human lymphoid cell line with low expression of its endogenous class I antigens (19,20). B*2705-C1R transfectant cells have been described elsewhere (21).
A genomic construct containing the C67S mutant of the B*2705 gene cloned into pUC19 (22,23) was a kind gift of Dr. Joel D. Taurog (University of Texas Southwestern Medical Center, Dallas, TX). This construct was designated as JHG1.1b. Exons 1 and 2 of the mutant were resequenced in our laboratory to confirm its correct structure. To obtain a C67S-C1R transfectant cell line, approximately 10 7 C1R cells were washed twice, and resuspended in 0.8 ml of PBS. Ten g of linearized JHG1.1b plasmid and 1 g of linearized pSV2.neo were co-transfected by electroporation at 960 V, 300 microfarads in a Bio-Rad Gene Pulser. After 10 min at room temperature, cells were seeded in a culture flask in Dulbecco's modified Eagle's medium supplemented with 10% FBS. Twenty-four hours after electroporation, 1 mg/ml G418 (Life Technologies, Paisley, United Kingdom) was added to the cell culture. Surviving cells were tested for expression of the C67S mutant by flow cytometry as described below. C1R cell lines were cultured in Dulbecco's modified Eagle's medium supplemented with 7.5% FBS (both from Life Technologies).
For flow cytometry analysis, approximately 2 ϫ 10 5 C1R or RMA-S transfectant cells were washed twice in 200 l of PBS, and resuspended in 50 l of undiluted mAb supernatant. After incubating for 20 min, cells were washed three times in 200 l of PBS and resuspended in 50 l of fluorescein isothiocyanate-conjugated anti-mouse IgG rabbit antiserum (Calbiochem-Novabiochem GmbH, Schwalbach, Germany), incubated for 20 min, and washed three times in 200 l of PBS. All operations were done at 4°C. Flow cytometry was carried out in an Epics Profile XL instrument (Coulter Electronics Inc., Hialeah, FL).
Isolation of B*2705-and C67S-B27-bound Peptides-This was carried from 10 10 C1R transfectant cells lysed in 1% Nonidet P-40 in the presence of a mixture of protease inhibitors, after immunopurification of HLA-B27 with the W6/32 mAb and acid extraction, exactly as described elsewhere (32). HLA-B27-bound peptide pools were fractionated by HPLC at a flow rate of 100 l/min as described previously (33), and 50-l fractions were collected.
Mass Spectrometry (MS) Analysis and Sequencing-The peptide composition of HPLC fractions was analyzed by matrix-assisted desorption ionization time-of-flight (MALDI-TOF) MS using a calibrated Kompact Probe instrument (Kratos-Shimadzu) operating in the positive linear mode, as described previously (33). Alternatively, a Bruker Re-flex™ III MALDI-TOF mass spectrometer (Bruker-Franzen Analytic GmbH, Bremen, Germany) equipped with the SCOUT™ source in positive ion reflector mode was also used, as described previously (34).
Peptide sequencing was carried out by quadrupole ion trap nanoelectrospray MS/MS in an LCQ instrument (Finnigan ThermoQuest, San Jose, CA), exactly as detailed elsewhere (35,36). In some cases, peptide sequencing was also done by post-source decay (PSD) MALDI-TOF MS, as described previously (34).
In all cases peptide-containing HPLC fractions were dried and resuspended in 5 l of methanol/water (1:1) containing 0.1% formic acid. Aliquots of 0.5 or 1 l were used for MALDI-TOF or nanoelectrospray MS analyses, respectively.
Synthetic Peptides-Peptides were synthesized using the standard solid-phase Fmoc (N-(9-fluorenyl)methoxycarbonyl) chemistry, and were purified by HPLC. The correct composition and molecular mass of purified peptides were confirmed by amino acid analysis using a model 6300 amino acid analyzer (Beckman Coulter, Palo Alto, CA), which also allowed their quantification.
Epitope Stabilization Assay-The epitope stabilization assay used to measure peptide binding was performed as described (37). Briefly, B*2705 RMA-S transfectants were incubated at 26°C for 22 h in RPMI 1640 medium supplemented with 10% heat-inactivated FBS. They were then washed three times in serum-free medium, incubated for 1 h at 26°C with various peptide concentrations without FBS, transferred to 37°C, and collected for flow cytometry after 4 h. B*2705 expression was measured using 50 l of hybridoma culture supernatant containing the mAb ME1. Binding of the RRYQKSTEL peptide, used as positive control, was expressed as C 50 , which is the molar concentration of the peptide at 50% of the maximum fluorescence obtained at the concentration range used (10 Ϫ4 to 10 Ϫ9 M). Binding of other peptides was assessed as the concentration of peptide required to obtain the fluorescence value at the C 50 of the control peptide. This was designated as EC 50 .
T-cell Clones and Cytotoxicity Assay-A set of allospecific CTL clones raised against B*2705 were used. The generation, culture conditions, and fine specificity of these CTL clones has been described previously (38). Recognition of B*2705 and C67S by these CTL was analyzed using a standard 51 Cr release cytotoxicity assay, essentially as described elsewhere (39).

RESULTS
The C67S Mutation Alters Cell Surface Expression and Stability of HLA-B27-Cell surface expression of B*2705 on C1R transfectants was measured by flow cytometry with the ME1, W6/32, BBM.1, and HC10 mAbs (Fig. 1). W6/32 and ME1 recognize distinct epitopes on the heavy chain of the native HLA-B27 molecule (3,22,40). BBM.1 recognizes an epitope on human ␤2m (30), and HC10 recognizes an epitope on unfolded, but not native, HLA class I heavy chains (31). Whereas surface expression of C67S in our transfectants was lower than expression of B*2705 when measured with W6/32, ME1, or BBM.1, it was higher when measured with HC10. This result indicates that the ratio between unfolded and native HLA-B27 molecules on the cell surface was higher for C67S than for B*2705, suggesting lower stability of the mutant molecules.
High Overlap between the B*2705-and C67S-bound Peptide Repertoires-A systematic comparison of B*2705-and C67Sbound peptide repertoires was carried out to determine the effect of the mutation on peptide specificity. The corresponding peptide pools were isolated by acid extraction, after immunoprecipitation of the B27 molecules with the W6/32 mAb, and fractionated by HPLC under identical conditions. The peptide composition of individual HPLC fractions was analyzed by MALDI-TOF MS. The MS spectrum of each HPLC fraction from one of the molecules was compared with the MS spectra of the correlative, previous, and following HPLC fractions from the other molecule. This was done to account for slight shifts in retention times between consecutive chromatographic runs. Ion peaks with the same (Ϯ1) mass/charge (m/z) among the HPLC fractions compared were considered to reflect identical peptides shared by both molecules. Ion peaks in one HPLC fraction not found in the counterparts from the other molecule were considered to be peptides differentially bound to one molecule.
One example of such comparison is shown in Fig. 2. HPLC fraction 161 from B*2705 was compared with HPLC fractions 160 -162 from C67S. Of 14 ion peaks compared from the B*2705 fraction, 12 had counterparts in one or more of the HPLC fractions from C67S, and 2 were not detected in the mutant.
A total of 1193 peptides from B*2705 and 1175 peptides from C67S were compared in this analysis (Table I). Of the B*2705 ligands, 1005 (84%) were shared by the C67S mutant, and 188 (16%) were not found in the mutant. Conversely, 171 peptides from C67S (15%) were not found in B*2705. These results indicate that the C67S mutation is compatible with binding of a large majority, but not all, of the natural B*2705 ligands, and influences peptide specificity also by allowing binding of new ligands absent from the B*2705-bound pool.
Although MALDI-TOF MS is not a quantitative technique because ion peak signal intensity is affected by multiple fac- were stained with the indicated mAb, followed by fluorescein isothiocyanate-conjugated rabbit-anti mouse IgG Ab. The background staining of untransfected C1R cells with W6/32 and BBM.1 is due to endogenous class I molecules of these cells (20). Peak heights were normalized in each panel to the peak height of untransfected CIR cells.
tors, an attempt was made to identify the extent in which peptide amount could be affected by the C67S mutation. Thus, to identify peptides common to both molecules but much more abundant in one of them, we selected ion peaks in each HPLC fraction whose intensity was higher than 50% of the maximum signal intensity in that fraction. Their amount was measured as the total number of millivolts (mV) corresponding to each ion peak in all the HPLC fractions in which it was detected. When the total intensity of a given ion peak was more than 10 times higher in one molecule than in the other, the corresponding peptide was assigned as a quantitative difference. One example (Fig. 2), is the ion peak at m/z 1088.1, in HPLC fraction 161 from B*2705 (SRTPYHVNL). Of 252 peptides compared by these criteria (Table I), 25 (10%) were much more prominent in B*2705, and 31 (12%) predominated in C67S. These results suggest that, in addition to determining differential binding of some peptides, the C67S mutation also influences the amount of bound peptide in at least an additional 22% of the shared ligands.
The C67S Mutant Favors Binding of Bigger Peptides-The size distribution of natural ligands bound to B*2705 or C67S showed a gaussian pattern in both cases, with mean peptide sizes ([M ϩ H] ϩ ) of 1154 and 1185 Da, respectively (Fig. 3A). This subtle size difference, which was reproducible in a second independent comparison (data not shown), became more obvious when the size distribution of peptides differentially bound to either B*2705 or C67S was compared (Fig. 3B). Whereas the mean size of B*2705 peptide differences was 1089 Da, that of C67S differences was 1295 Da. Thus, peptides differentially bound to the mutant had an mean size 206 Da bigger than B*2705 differences. This is compatible with a slightly bigger length (1-2 residues), a bias toward bigger residues at some positions, or both.
The C67S Mutant Selects against Peptides with Small Residues at Position 1-A total of 36 natural ligands bound to B*2705 or to the C67S mutant were sequenced by quadrupole ion trap nanoelectrospray MS/MS (Fig. 4). Of these, 29 peptides, including 20 nonamers, 8 decamers, and 1 undecamer, were shared by both molecules. Among these shared ligands, nine showed quantitative differences, six of them predominant in B*2705, and three of them predominant in C67S. In addi-   tion, six peptides, including five nonamers and one decamer, were differentially bound to B*2705, and one nonamer was differentially bound to C67S.
In general there were no obvious differences between peptides common to B*2705 and C67S, and peptides found only or predominantly in B*2705, except at position (P) 1. Of the 12 peptides B*2705-specific or predominant in B*2705 (Fig. 4), 11 (92%) had a small P1 residue: Gly, Ala, or Ser. Together, these three amino acids accounted only for 25% of the other shared ligands. In contrast, whereas basic P1 residues (Arg, Lys, and His) were found in 12 (60%) of the shared ligands, and in 2 of 3 peptides predominant in the mutant, these residues were absent from peptides differentially or predominantly bound to B*2705. These results strongly suggest that the C67S mutation disfavors binding of peptides with small P1 residues, while allowing basic residues at this position.
There were more peptides with Pro-4 among those differentially or predominantly bound to B*2705 (58%) than among shared ones (15%). However, the significance of this is dubious because there was Pro at positions 4, 5, or 6 in 40% of the shared ligands (Fig. 4).
Natural B*2705 Ligands Lacking Arg-2-Two peptides with Gln-2 were sequenced from the C67S mutant, both of which had counterparts in the B*2705-bound peptide pool (Fig. 4). The MALDI-TOF MS spectra of HPLC fraction 143 from C67S and B*2705 showed an ion peak at m/z 1056.4. The ion peak from the mutant was fragmented by quadrupole ion trap nanoelectrospray MS/MS, which allowed us to determine the sequence of the corresponding peptide as RQTGIVLNR (M ϭ 1055.6 Da) (Fig. 5A). Assignment of this sequence and, in particular, distinction of Gln-2 from the isobaric Lys residue was based on: (a) match of this sequence, but not of that with Lys-2, with a human protein in the data base (Fig. 4), (b) detection of the y 8 * fragment ion (m/z 883.3), which is consistent with neutral loss of ammonia caused by cyclation of the N-terminal Gln residue in the y 8 Љ ion, and (c) the MS/MS spectrum of the synthetic RQTGIVLNR peptide was identical to the peptide from the mutant (data not shown). The corresponding ion peak from B*2705 was sequenced by PSD-MALDI-TOF MS. Its spectrum (Fig. 5B) was essentially identical to that of the synthetic RQTGIVLNR peptide (Fig. 5C). This is the first demonstration that B*2705 binds peptides with Gln-2 in vivo.
Similarly, the MALDI-TOF MS spectra of HPLC fractions 173 from C67S and B*2705 revealed an ion peak with m/z 1023.4. The sequence of this peptide obtained from the C67S mutant was RQVIPIIGK (Fig. 4). In this case, PSD-MALDI-TOF or electrospray MS/MS spectra of the B*2705 counterpart could not be obtained, so that assignment of this peptide as a natural B*2705 ligand relies only on identity of molecular mass and retention time with the peptide from C67S.
Extended B Pocket Specificity of the C67S Mutant-A third peptide lacking Arg-2 was found in HPLC fraction 155 from the mutant. In the MALDI-TOF MS spectrum of this fraction, there was an ion peak at m/z 1105.4, which was not detected in the corresponding HPLC fractions from B*2705 (data not shown), so that this peak was assigned as a peptide difference from the mutant. The nanoelectrospray MS/MS spectrum of this peptide was consistent with the sequence YKFSGFTQK (Fig. 6A). Distinction between the isobaric Gln/Lys residues was based on: 1) unambiguous matching with a human sequence in the data base (Fig. 4), 2) lack of a prominent fragment ion corresponding to neutral loss of ammonia from the y 8 Љ ϩ or y 8 Љ 2ϩ fragment ions, which would be expected if the peptide would have Gln-2, because of cyclation of this residue when it becomes N-terminal in the y 8 Љ ions, and 3) identity of the MS/MS spectrum with that of synthetic YKFSGFTQK (Fig.  6B), but not with the synthetic analog with Gln-2 (Fig. 6C). In this latter spectrum, prominent y 8 * fragment ions corresponding to neutral loss of ammonia from y 8 Љ ϩ and y 8 Љ 2ϩ fragment ions were a clear differential feature distinguishing this spectrum from those of the C67S ligand and the synthetic YKFS-GFTQK peptide.
This result demonstrates that the C67S mutation relaxes the specificity of the B pocket, allowing for residues other than Arg-2 or Gln-2.
Peptides Containing Gln-2, but Not Lys-2, Bind Efficiently B*2705 in Vitro-The three peptide ligands lacking Arg-2 found in the previous analyses were tested for cell surface binding to B*2705 in an epitope stabilization assay (Fig. 7). The two peptides with Gln-2 bound B*2705 with high efficiency (EC 50  High Allospecific T-cell Epitope Sharing between B*2705 and C67S-Because most allospecific CTL are peptide-dependent, another way to look at the effect of C67S on peptide specificity is to examine the cross-reactivity of allospecific CTL raised against B*2705 with the C67S mutant. Thus, 11 anti-B*2705 CTL clones were tested for recognition of C67S-C1R transfectant target cells in a classical cytotoxicity assay (Table II). Seven CTL clones (64%) showed significant cross-reaction (Ͼ50% relative lysis), whereas four CTL clones (36%) crossreacted more weakly (Ͻ40% relative lysis). These results indicate that many allospecific T-cell epitopes are conserved in the mutant. However, epitope sharing was lower than the peptide sharing determined by comparison of the B*2705-and C67Sbound peptide repertoires (Table I). This result suggests that some shared ligands may bind B*2705 and the mutant in different conformations, so that the corresponding T-cell epitopes are altered.

DISCUSSION
The Cys-67 residue is a structural feature of HLA-B27 shared by few other allotypes such as subtypes of HLA-B14, B15, B38, B39, and B73 (41). However, only HLA-B27 and the rare allotype HLA-B73 (42,43) also have Lys-70, which confers to Cys-67 enhanced chemical reactivity (44). Although this residue might be modified by homocysteine in vivo upon bacterial invasion (45), it seems that the overwhelming majority of HLA-B27 molecules in normal cells have a free Cys-67. Because this residue is an integral part of the B pocket (15), it was expected to influence peptide binding and antigen presentation. Thus, in this study we investigated the role of Cys-67 in: 1) cell surface expression and stability, 2) peptide specificity, and 3) T-cell allorecognition.
Early studies with HLA-B27 mutants bearing different sub-stitutions at position 67 demonstrated that these mutations did not impair cell surface expression, but destroyed some mAb epitopes. These results were interpreted as a limited conformational effect of the mutations on the antigenic structure of HLA-B27 (22,23). Recent studies have shown that Cys-67 controls the formation of HLA-B27 heavy chain homodimers retaining secondary structure, some peptide binding capacity, and reactivity with the conformation-dependent W6/32 mAb (9). More relevant to the present study, Reinelt et al. (16) demonstrated that Cys-67 have a significant influence on the thermodynamic stability of B*2705-peptide complexes in solution. The C67S mutation decreased the thermal stability of the molecule, albeit less than other pocket B mutants, and promoted faster dissociation. These results are consistent with the inverted ratio of HLA-B27 associated fluorescence at the cell surface measured with W6/32, ME1, and BBM.1, relative to the fluorescence measured with HC10, between B*2705-and C67S-C1R transfectants. The increased HC10 reactivity observed with the mutant strongly suggests that a higher fraction of C67S-peptide complexes are dissociated at the cell surface relative to the wild type. Although the molecular determinants of this decreased stability cannot be established from our experiment, a simple explanation of the result is that the strength of the interaction of many HLA-B27-peptide complexes is reduced in the mutant because of its altered B pocket structure, leading to decreased stability.
Assessment of the role of Cys-67 in determining the peptide specificity of HLA-B27 required a systematic comparison of the endogenous peptide repertoires of the wild type and mutant, which was undertaken in this study. Our results demonstrated that, although the C67S mutant still binds many of the natural B*2705 ligands, its peptide specificity was distinct from the wild type in some significant aspects: 1) the mutant failed to bind, within the detection limits of the MS techniques used, approximately 15% of the B*2705-bound peptides, while gaining the capability to bind a similar percentage of new ligands, 2) the mutant modulated, either positively or negatively, the amount of bound peptide for at least an additional 22% of natural B*2705 ligands (because conservative criteria were used to assess quantitative peptide differences among shared ligands, it is conceivable that this percentage may be significantly higher); 3) the mutant modulated residue preferences at P1; and 4) it increased permissiveness at the B pocket, allowing for P2 residues not found among B*2705 ligands.
The effect on P1 residues, consisting of a certain bias against small residues and toward basic ones, is presumably an indirect consequence of weaker pocket B interactions in the mutant. Basic P1 residues may provide additional strength to the interaction of peptides with HLA-B27, as a result of hydrogen bonding of the P1 residue with Glu-63 in HLA-B27, as well as additional van der Waals interactions of the aliphatic portions of the P1 side chains with Trp-167, in the A pocket. This was observed, for instance, in H-2K b complexed with an Arg-1containing peptide (46), or in HLA-B53 complexed with a Lys-1-containing peptide (47). These interactions are not possible, or are weaker, with Gly and other small P1 residues. Stronger anchoring of the P1 residue could partially compensate for weaker pocket B interactions. van der Waals interactions with Trp-167 are also important with bulky aromatic P1 residues, such as the Tyr-1 found in the only sequenced C67S-specific ligand. Significantly, an even stronger bias toward basic P1 residues than that observed in C67S was reported for B*2703 (48,49). This allotype has an unusual mutation affecting an otherwise conserved residue in the A pocket, which weakens anchoring of the peptidic N terminus (50,51). This suggests that weakening of canonical interactions in peptide-binding pockets make P1 side chain interactions involving basic residues particularly relevant, consistent with our results in C67S.
An interesting observation concerning B pocket specificity in this study was the finding of natural B*2705 ligands with Gln-2. This is the first demonstration of natural B*2705 ligands lacking Arg-2. A previous publication that described a Gln-2-containing ligand of B*2705 (52) was later retracted (53), and the description of this ligand is not believed to be valid. 2 The finding was not totally unexpected, because previous studies from ourselves (54) and others (55) showed that some Gln-2-containing peptides bind B*2705 in vitro with significant efficiency. Because the overwhelming majority of natural B*2705 ligands contain Arg-2, identification of Gln-2-containing peptides in the B*2705-bound pool was facilitated by detection of these peptides as more prominent ion peaks in correlative HPLC fractions from the mutant, perhaps reflecting increased allowance for this residue in C67S. For this reason, it cannot be inferred from our results that finding of 2 Gln-2containing peptides of 35 sequenced ligands from B*2705 reflects the true proportion of peptides carrying the Gln-2 motif in the B*2705-bound pool. That HLA-B27 presents peptides lacking Arg-2 in vivo has obvious implications in studies aimed at predicting and identifying B27-restricted antigens such as, for instance, those from arthritogenic bacteria (56).
The effect of the C67S mutation on relaxing B pocket specificity was also demonstrated by sequencing of one ligand of this mutant not found in B*2705, which contained Lys-2. This residue has never been found among natural HLA-B27 ligands.
The observed effect of the C67S mutation on allorecognition by B*2705-specific CTL shows that Cys-67 plays a role in antigen presentation and is consistent with the observed alterations of the peptide repertoire. The limited panel of CTL clones available for our study imposes a cautious interpretation of the apparently lower T-cell cross-reactivity with the mutant, 2 J. D. Taurog, personal communication.  Functional Role of Cys-67 in HLA-B27 relative to the homology of peptide repertoires. However, this result strongly suggests that the 15% lack of overlap between the peptide repertoires of B*2705 and C67S is not simply caused by lack of sensitivity of our MS techniques for peptides with low abundance, because CTL would presumably detect peptide amounts undetected in our MS analysis (57). Rather, our CTL data are consistent with the possibility that some shared ligands are not antigenically identical when presented by B*2705 and the mutant. This would imply that, in addition to its effect on peptide binding, the C67S mutation may alter the conformation and antigenicity of some bound peptides.
In conclusion, our results indicate that Cys-67 plays a role in determining the cell surface stability, peptide binding specificity, and T-cell recognition of HLA-B27. They also suggest that increased unfolding of the mutant at the cell surface, relative to the wild type, may be the result of decreased stability of C67Speptide complexes, mainly as a consequence of weakened pocket B interactions.
Studies in transgenic rats indicate that the phenotype of HLA-B27-associated disease is at most modestly affected by the C67S mutation (18). This suggests that the effect of this mutation on peptide specificity has not much relevance for disease. It has been suggested that HLA-B27-associated disease may be a consequence of HLA-B27 misfolding (13). Failure of the C67S mutation to exacerbate disease in transgenic rats (18), despite decreased thermodynamic (16) and cell surface stability of the mutant, does not seem to support this model. However, both our study and that of Reinelt et al. (16) deal with HLA-B27-peptide complexes. Therefore, they do not necessarily imply an influence of the C67S mutation on intracellular folding of the HLA-B27 heavy chain. This should be addressed in further studies.