Voltar
Artigo Acesso aberto Revisado por pares
BIBTEX RIS

Major Histocompatibility Complex Class I Molecules Bind Natural Peptide Ligands Lacking the Amino-terminal Binding Residue in Vivo

2001; Elsevier BV; Volume: 276; Issue: 47 Linguagem: Inglês

10.1074/jbc.m105981200

ISSN

1083-351X

Autores

Jesús Yagüe, Anabel Marina, Jesús Vázquez, José A. Łópez de Castro,

Tópico(s)

Peptidase Inhibition and Analysis

Resumo

Major histocompatibility complex (MHC) class I-peptide complexes are stabilized by multiple interactions, including those of the peptidic NH2-terminal group in the A pocket of the MHC molecule. In this study, the characterization of four natural HLA-B39 ligands lacking the amino-terminal binding residue is reported. These peptides were found in the endogenous peptide pool of one or more of the B*3901, B*3905, and B*3909 allotypes and sequenced by nanoelectrospray mass spectrometry. Control experiments ruled out that they resulted from exopeptidase trimming of their NH2-terminally extended counterparts: NAc-SHVAVENAL, EHGPNPIL, IHEPEPHIL, and EHAGVISVL, also present in the same peptide pools, during purification. HAGVISVL and HVAVENAL behaved similarly to the corresponding NH2-terminally extended peptides in their binding to B*3901 and B*3909 at the cell surface in vitro, and in cell surface stabilization of B*3901. This is, to our knowledge, the first demonstration that peptides lacking the amino-terminal binding residue bind in vivo to classical MHC class I molecules. The results indicate that canonical MHC-peptide interactions in the A pocket are not always necessary for endogenous peptide presentation. Major histocompatibility complex (MHC) class I-peptide complexes are stabilized by multiple interactions, including those of the peptidic NH2-terminal group in the A pocket of the MHC molecule. In this study, the characterization of four natural HLA-B39 ligands lacking the amino-terminal binding residue is reported. These peptides were found in the endogenous peptide pool of one or more of the B*3901, B*3905, and B*3909 allotypes and sequenced by nanoelectrospray mass spectrometry. Control experiments ruled out that they resulted from exopeptidase trimming of their NH2-terminally extended counterparts: NAc-SHVAVENAL, EHGPNPIL, IHEPEPHIL, and EHAGVISVL, also present in the same peptide pools, during purification. HAGVISVL and HVAVENAL behaved similarly to the corresponding NH2-terminally extended peptides in their binding to B*3901 and B*3909 at the cell surface in vitro, and in cell surface stabilization of B*3901. This is, to our knowledge, the first demonstration that peptides lacking the amino-terminal binding residue bind in vivo to classical MHC class I molecules. The results indicate that canonical MHC-peptide interactions in the A pocket are not always necessary for endogenous peptide presentation. major histocompatibility complex monoclonal antibody β2-microglobulin cytotoxic T lymphocyte transporter associated with antigen processing human T-cell lymphotropic virus fetal bovine serum trifluoroacetic acid high performance liquid chromatography mass spectrometry matrix-assisted desorption-ionization time of flight flow microfluorometry MHC1 class I molecules constitutively bind endogenous peptides, usually 8–11 residues long, and present them at the cell surface for recognition by cytotoxic T lymphocyte (CTL). Peptides bind to the MHC molecule through a complex array of interactions. Some of these are sequence-independent, involving the NH2 and carboxyl termini and the peptide main chain. Other interactions are sequence-dependent and involve some of the peptide side chains (1Madden D.R. Annu. Rev. Immunol. 1995; 13: 587-622Crossref PubMed Scopus (708) Google Scholar). A variable number of water molecules are involved in hydrogen bonding with the peptide and the MHC molecule and play an important stabilizing role. The peptide NH2 and COOH termini are anchored in the A and F pocket, respectively, of the peptide binding site (2Garrett T.P. Saper M.A. Bjorkman P.J. Strominger J.L. Wiley D.C. Nature. 1989; 342: 692-696Crossref PubMed Scopus (591) Google Scholar, 3Saper M.A. Bjorkman P.J. Wiley D.C. J. Mol. Biol. 1991; 219: 277-319Crossref PubMed Scopus (964) Google Scholar), establishing hydrogen bonds with conserved MHC residues in these pockets. These interactions have a significant contribution to the stability of MHC-peptide complexes (4Madden D.R. Gorga J.C. Strominger J.L. Wiley D.C. Cell. 1992; 70: 1035-1048Abstract Full Text PDF PubMed Scopus (610) Google Scholar, 5Bouvier M. Wiley D.C. Science. 1994; 265: 398-402Crossref PubMed Scopus (244) Google Scholar). Peptidic anchor side chains interact in other pockets of the peptide binding site. In HLA class I molecules the position (P)2 and PΩ side chains, which bind in pockets B and F, respectively, are the main anchor residues of class I-bound ligands. Other residues, most notably at P3, are important auxiliary anchors (6Ruppert J. Sidney J. Celis E. Kubo R.T. Grey H.M. Sette A. Cell. 1993; 74: 929-937Abstract Full Text PDF PubMed Scopus (614) Google Scholar). Natural MHC-peptide complexes are very stable with long half-lives (7Cerundolo V. Elliott T. Elvin J. Bastin J. Rammensee H.G. Townsend A. Eur. J. Immunol. 1991; 21: 2069-2075Crossref PubMed Scopus (158) Google Scholar, 8Ojcius D.M. Abastado J.P. Casrouge A. Mottez E. Cabanie L. Kourilsky P. J. Immunol. 1993; 151: 6020-6026PubMed Google Scholar, 9Burshtyn D.N. Barber B.H. J. Immunol. 1993; 151: 3082-3093PubMed Google Scholar, 10van der Burg S.H. Visseren M.J. Brandt R.M. Kast W.M. Melief C.J. J. Immunol. 1996; 156: 3308-3314PubMed Google Scholar, 11Levitsky V. Zhang Q.J. Levitskaya J. Masucci M.G. J. Exp. Med. 1996; 183: 915-926Crossref PubMed Scopus (114) Google Scholar). In the absence of appropriate peptides, such as in cells lacking the transporter associated with antigen processing (TAP), the stability of the class I molecule is drastically reduced and its expression at the cell surface largely impaired (12Ljunggren H.G. Karre K. J. Exp. Med. 1985; 162: 1745-1759Crossref PubMed Scopus (642) Google Scholar, 13Ljunggren H.G. Stam N.J. Ohlen C. Neefjes J.J. Hoglund P. Heemels M.T. Bastin J. Schumacher T.N. Townsend A. Karre K. Ploegh H.L. Nature. 1990; 346: 476-480Crossref PubMed Scopus (784) Google Scholar, 14Townsend A. Ohlen C. Bastin J. Ljunggren H.G. Foster L. Karre K. Nature. 1989; 340: 443-448Crossref PubMed Scopus (875) Google Scholar, 15Cerundolo V. Alexander J. Anderson K. Lamb C. Cresswell P. McMichael A. Gotch F. Townsend A. Nature. 1990; 345: 449-452Crossref PubMed Scopus (326) Google Scholar). Because of its significant contribution to peptide stability (5Bouvier M. Wiley D.C. Science. 1994; 265: 398-402Crossref PubMed Scopus (244) Google Scholar), a free NH2 terminus is found in the overwhelming majority of natural class I-bound peptides. However, finding of an N α-acetylated natural ligand of HLA-B39 (16Yague J. Alvarez I. Rognan D. Ramos M. Vazquez J. Lopez de Castro J.A. J. Exp. Med. 2000; 191: 2083-2092Crossref PubMed Scopus (22) Google Scholar) demonstrated that a blocked NH2 terminus does not necessarily impair peptide binding in vivo. Nevertheless, this peptide bound less efficiently than a nonacetylated analog. Recently, the crystal structure of HLA-A*0201 in complex with a human T-cell lymphotropic virus (HTLV)-1-derived Tax8 peptide lacking the amino-terminal binding residue was reported (17Khan A.R. Baker B.M. Ghosh P. Biddison W.E. Wiley D.C. J. Immunol. 2000; 164: 6398-6405Crossref PubMed Scopus (145) Google Scholar). This study demonstrated that one such peptide may bind in vitro to class I molecules, albeit with reduced stability relative to its canonical counterpart. Binding was possible because the A pocket was filled with hydrogen-bonded water molecules that partially compensated for the loss of canonical interactions involving the NH2terminus of the P1 residue. In that study the Tax8 peptide had a negligible ability to sensitize target cells for lysis by one Tax9-specific CTL clone. However, in an earlier report (18Utz U. Koenig S. Coligan J.E. Biddison W.E. J. Immunol. 1992; 149: 214-221PubMed Google Scholar), Tax8 sensitized targets for lysis by Tax-specific CTL generated from HTLV-1-infected individuals against this Tax peptide. That study did not rule out the possibility that this reflects cross-reactivity from some CTL stimulated in vivo against the dominant Tax9 epitope and, therefore, did not show that Tax8 was a natural ligand of HLA-A2. In this study, we report, for the first time to our knowledge, the identification of natural class I ligands lacking the amino-terminal binding residue, from endogenous peptide pools. These peptides were isolated from three HLA-B39 allotypes: B*3901, B*3905, and B*3909. These molecules bind peptides with either Arg2 or His2, but the preference of each allotype for either motif is variable; B*3905 has a higher preference for His2 than B*3901, whereas B*3909 has a marked preference for Arg2(19Falk K. Rotzschke O. Takiguchi M. Gnau V. Stevanovic S. Jung G. Rammensee H.G. Immunogenetics. 1995; 41: 162-164Crossref PubMed Scopus (42) Google Scholar, 20Yague J. Ramos M. Vazquez J. Marina A. Albar J.P. Lopez de Castro J.A. Tissue Antigens. 1999; 53: 227-236Crossref PubMed Scopus (16) Google Scholar, 21Yague J. Ramos M. Ogueta S. Vazquez J. Lopez de Castro J.A. Tissue Antigens. 2000; 56: 385-391Crossref PubMed Scopus (7) Google Scholar). All the peptides lacking the P1 residue found had His as the B pocket-binding motif. RPMI 1640 medium containing 25 mmHepes buffer, AIM V medium, fetal bovine serum (FBS), streptomycin sulfate, and penicillin G were purchased from Life Technologies, Inc. (Paisley, United Kingdom). Leupeptin, pepstatin, and aprotinin were purchased from Roche Molecular Biochemicals (Mannheim, Germany). Trifluoroacetic acid (TFA), iodoacetamide, phenylmethanesulfonyl fluoride, brefeldin A, human β2-microglobulin (β2m), and hygromycin B were purchased from Sigma.l-Glutamine was purchased from Merck (Darmstadt, Germany). NaN3, NaCl, EDTA, and Tris-HCl were purchased from Fluka (Buchs, Switzerland). Durapore membrane filter type HVLP was purchased from Millipore Corp. (Bedford, MA). The CNBr-activated Sepharose™ 4B was purchased from Amersham Pharmacia Biotech AB (Uppsala, Sweden). Centricon C-3 was purchased from Amicon (Beverly, MA). Deltapak C18 and Sep-Pak t-C18 column were purchased from Waters (Milford, MA). All synthetic peptides were made in our Protein Chemistry facility, purified by HPLC, and stored as stock solutions in water, without dimethyl sulfoxide, before use. HMy2.C1R (C1R) is a human lymphoid cell line with low expression of its endogenous class I antigens (22Storkus W.J. Howell D.N. Salter R.D. Dawson J.R. Cresswell P. J. Immunol. 1987; 138: 1657-1659PubMed Google Scholar, 23Zemmour J. Little A.M. Schendel D.J. Parham P. J. Immunol. 1992; 148: 1941-1948PubMed Google Scholar). C1R transfectants expressing B*3901, B*3905, or B*3909 have been described previously (20Yague J. Ramos M. Vazquez J. Marina A. Albar J.P. Lopez de Castro J.A. Tissue Antigens. 1999; 53: 227-236Crossref PubMed Scopus (16) Google Scholar, 21Yague J. Ramos M. Ogueta S. Vazquez J. Lopez de Castro J.A. Tissue Antigens. 2000; 56: 385-391Crossref PubMed Scopus (7) Google Scholar). RMA-S is a TAP-deficient murine cell line (12Ljunggren H.G. Karre K. J. Exp. Med. 1985; 162: 1745-1759Crossref PubMed Scopus (642) Google Scholar). B*3901-RMA-S transfectant cells were kindly supplied by Dr. Masafumi Takiguchi (Division of Viral Immunology, Center for AIDS Research, Kumamoto University, Kumamoto, Japan). B*3909-RMA-S has been described previously (20Yague J. Ramos M. Vazquez J. Marina A. Albar J.P. Lopez de Castro J.A. Tissue Antigens. 1999; 53: 227-236Crossref PubMed Scopus (16) Google Scholar). Both RMA-S transfectants express human β2m. C1R cells were grown in Dulbecco's modified Eagle's medium, pH 7.4, containing 7.5% heat-inactivated FBS, 100 μg/ml streptomycin sulfate, and 100 units/ml penicillin G. RMA-S transfectant cell lines were grown in RPMI 1640 medium containing 25 mm Hepes buffer and 7.5% heat-inactivated FBS, without antibiotics but with 0.3 mg/ml hygromycin B for the B*3901 transfectant. Approximately 1010 HLA-B*3901-, B*3905-, or B*3909-C1R transfectant cells were lysed in 20 mm Tris-HCl, 150 mm NaCl, 1% Nonidet P-40, pH 7.4, containing the following protease inhibitors: 10 μg/ml leupeptin, 2 μg/ml pepstatin A, 2 μg/ml aprotinin, 18.5 μg/ml iodoacetamide, 1 mm EDTA, 348 μg/ml phenylmethanesulfonyl fluoride, and 0.02% NaN3. Lysates were centrifuged at 4 °C for 10 min at 1,500 × g, and then for 1 h at 38,000 × g. The supernatant was filtered through a 0.45-μm Durapore membrane filter, pre-cleared with Sepharose-ethanolamine beads, and subjected to affinity chromatography with W6/32-Sepharose. The murine W6/32 mAb is an IgG2a, specific for a monomorphic HLA-A,B,C determinant (24Barnstable C.J. Bodmer W.F. Brown G. Galfre G. Milstein C. Williams A.F. Ziegler A. Cell. 1978; 14: 9-20Abstract Full Text PDF PubMed Scopus (1583) Google Scholar). The column was washed with: (a) 250 ml of NET (50 mm Tris, 150 mm NaCl, 5 mm EDTA, 0.1% NaN3, pH 7.4) containing 10% saturated NaCl and 0.5% Nonidet P-40, (b) 250 ml of NET containing 5% of saturated NaCl and 0.5% Nonidet P-40, and (c) 500 ml of NET. Peptide elution was done with 0.1% TFA in water at room temperature. Collected fractions were filtered through Centricon C-3. Material with Mr <3000 Da was subjected to reverse phase HPLC on a Deltapak C18 column maintained at 30 °C, using a flow rate of 100 μl/min and the following linear gradient: 0–20 min, 100% A; 100 min, 56% A; 140 min, 100% B; 141 min, 100% acetonitrile; 145 min, 100% acetonitrile, 200 μl/min. Buffers A and B were 0.1% TFA in water and 80% acetonitrile (v/v), 0.1% TFA in water, respectively. Fractions (50 μl) were collected at 30-s intervals. Peptide composition analysis, zoomscan, and sequencing by electrospray ion trap MS was carried out with an LCQ instrument (Thermo Finnigan, San Jose, CA), using the "nanospray" interface as detailed elsewhere (25Marina A. Garcia M.A. Albar J.P. Yague J. Lopez de Castro J.A. Vazquez J. J. Mass Spectrom. 1999; 34: 17-27Crossref PubMed Scopus (58) Google Scholar). In some cases, sequence assignments were confirmed by refragmenting some fragment ions (MS3) arising from the parental one, and by MS/MS sequencing of the corresponding synthetic peptides. Zoomscan is a high resolution method for determining accurate peptide mass and charge of ionic species, in which a narrow precursor ion window is selected to incorporate several isotopomers. The charge states of individual product ions were determined at enhanced resolution by scanning across a limited mass/charge (m/z) range. In some experiments, 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, Manchester, United Kingdom) operating in the positive linear mode as described previously (26Paradela A. Alvarez I. Garcia-Peydro M. Sesma L. Ramos M. Vazquez J. Lopez de Castro J.A. J. Immunol. 2000; 164: 329-337Crossref PubMed Scopus (36) Google Scholar), using 1 μl of the HPLC fractions. The epitope stabilization assay used to measure peptide binding was performed as described (27Galocha B. Lamas J.R. Villadangos J.A. Albar J.P. Lopez de Castro J.A. Tissue Antigens. 1996; 48: 509-518Crossref PubMed Scopus (37) Google Scholar). Briefly, B*3901-RMA-S or B*3909-RMA-S transfectants were incubated at 26 °C for 24 h. They were then incubated 1 h at 26 °C with 10−4 to 10−9m peptide without FBS, transferred to 37 °C, and collected for FMF after 3 h (B*3901) or 4 h (B*3909). B*3901 or B*3909 expression was measured using 50 μl of hybridoma culture supernatant containing the mAb W6/32, as described previously (20Yague J. Ramos M. Vazquez J. Marina A. Albar J.P. Lopez de Castro J.A. Tissue Antigens. 1999; 53: 227-236Crossref PubMed Scopus (16) Google Scholar). Binding was expressed as the C50, which is the molar concentration of a given peptide at 50% of the maximum fluorescence obtained with that peptide within the concentration range used. Approximately 2 × 105 B*3901-RMA-S transfectant cells/well were incubated at 26 °C for 18 h in RPMI 1640 medium supplemented with 10% heat-inactivated FBS, washed twice with serum-free medium, and further incubated with 100 μm peptide and 100 nmhuman β2m for 1 h at 26 °C in the presence of 5 μg/ml brefeldin A, to avoid appearance of newly synthesized class I MHC molecules, in a total volume of 100 μl. Cells were then washed with RPMI 1640 medium, resuspended in brefeldin A-containing AIM V medium, and transferred to 37 °C. Cells were removed at various times and subjected to FMF with the W6/32 mAb as in the previous paragraph. Cell surface stability of the B*3901-peptide complexes was measured as DT50, which is the time corresponding to 50% of the maximum fluorescence, measured at time 0 after transfer to 37 °C. In a first assay, three synthetic peptides: EHAGVISVL (0.78 nmol), IHEPEPHIL (0.40 nmol), and NAc-SHVAVENAL (0.99 nmol) were dissolved in 400 μl of NET buffer, pH 7.4, and incubated for 90 min at room temperature in a NET-washed W6/32-Sepharose column, like those used for immunoaffinity purification of HLA-B39. Elution of the peptides was carried out with five column volumes of 0.1% TFA in water at 0.3 ml/min. The 15-ml eluate was concentrated to 5 ml, passed through Centricon C-3, and then through a Sep-Pak t-C18 column equilibrated with 0.1% TFA in water. The sample was loaded into the column and rinsed with 15 ml of 0.1% TFA. Peptides were eluted with 10 ml of 20% water in CH3CN, brought to a final volume of 100 μl of 0.1% aqueous TFA, and subjected to reverse phase HPLC in the same conditions used for HLA-B39-bound peptide pools. A separate blank for the W6/32-Sepharose and Sep-Pak t-C18 columns was run in parallel. In a second assay, 8.2 × 108 untransfected C1R cells were lysed and further processed exactly as done for isolation of HLA-B39-bound peptides. Immediately after washing the lysate supernatant from the W6/32-Sepharose immunoaffinity column, a mixture of three synthetic peptides, EHAGVISVL (0.39 nmol), IHEPEPHIL (0.17 nmol), and NAc-SHVAVENAL (0.41 nmol) dissolved in 400 μl of NET buffer, pH 7.4, was loaded into the column and eluted with 20 ml of 0.1% TFA in water at 0.3 ml/min at room temperature. Peptide-containing fractions were filtered through Centricon C-3 and subjected to reverse phase HPLC as in the previous assay. Composition of HPLC fractions eluting in the 70–90-min range was analyzed by MALDI-TOF MS. The HLA-B*3905-bound peptide pool was isolated from B*3905-C1R transfectant cells after immunopurification of the class I molecule and acid extraction. Peptides were fractionated by HPLC (Fig.1). Full scan analysis of HPLC fraction N.158 (elution time: 78 min) by nanoelectrospray quadrupole ion trap MS revealed a prominent ion peak at m/z 486.1. Zoomscan analysis was consistent with the [M + 2H]2+ ion of a peptide with a molecular mass (M) of 970.4 Da (Fig.2 A). The MS/MS fragmentation spectrum of this ion peak was consistent with the sequence HEPEPHIL (Fig. 2 B). This was confirmed by MS/MS fragmentation of the corresponding synthetic peptide (data not shown). An NH2-terminally extended peptide, IHEPEPHIL, found in the same peptide pool (Fig. 1), had been previously reported as a natural B*3905 ligand (21Yague J. Ramos M. Ogueta S. Vazquez J. Lopez de Castro J.A. Tissue Antigens. 2000; 56: 385-391Crossref PubMed Scopus (7) Google Scholar). Thus, these results indicate that a peptide in the B*3905-bound pool lacks the P1 residue of another natural ligand from the same peptide pool.Figure 2Panel A, full scan MS spectrum of HPLC fraction N.158 from the B*3905-bound peptide pool. Zoomscan analysis (inset) of the major ion peak (m/z486.1) revealed that it corresponded to the [M+2H]2+ ion of a peptide with M = 970.4 Da. Panel B, MS/MS fragmentation spectrum of the ion peak at m/z 486.1 in Panel A; identified fragment ions of the b and y" series, and prominent ion peaks corresponding to secondary series, are indicated. Ion peaks at m/z 235.1 and 574.1 were identified as internal sequence ions PH and PEPHI upon refragmentation (MS3) of the latter one (data not shown). The peptide sequence assigned is shown. It was confirmed by MS/MS fragmentation of the corresponding synthetic peptide. The nomenclature of Roepsstorff and Fohlman (35Roepstorff P. Fohlman J. Biomed. Mass. Spectrom. 1984; 11: 601Crossref PubMed Scopus (2371) Google Scholar) was used for the main and secondary series of fragment ions, and that of Biemann (36Biemann K. Methods Enzymol. 1990; 193: 886-887Crossref PubMed Scopus (419) Google Scholar) for internal sequence ions.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Three additional peptides related to previously described B*3905 ligands with His2 by lack of their P1 residue were found in the same peptide pool. In HPLC fractions N.155 (elution time: 76.5 min) and N.175 (elution time: 86.5 min), zoomscan analysis revealed the presence of peptides with molecular mass of 746.4 and 794.3, respectively (Fig. 3), whose sequence by nanoelectrospray MS/MS (Fig. 4) demonstrated that they were peptides with His1: HGPNPIL and HAGVISVL, respectively. The sequence of the latter peptide was confirmed by MS/MS fragmentation of the corresponding synthetic peptide. NH2-terminally extended counterparts of these peptides (EHGPNPIL and EHAGVISVL) were known natural B*3905 ligands (21Yague J. Ramos M. Ogueta S. Vazquez J. Lopez de Castro J.A. Tissue Antigens. 2000; 56: 385-391Crossref PubMed Scopus (7) Google Scholar). A third peptide, with M = 851.3 was detected in HPLC fraction N.145 (elution time: 71.5 min) from B*3905. Although this peptide was not revealed initially by full scan or zoomscan analyses, its presence was confirmed by direct MS/MS fragmentation analysis focused at the corresponding m/z value, after its finding in B*3909 (Fig. 3). The sequence of this peptide, as determined by nanoelectrospray MS/MS (Fig. 4), was HVAVENAL. This peptide was the (−P1) counterpart of a previously reported N α-acetylated HLA-B39 ligand NAc-SHVAVENAL (16Yague J. Alvarez I. Rognan D. Ramos M. Vazquez J. Lopez de Castro J.A. J. Exp. Med. 2000; 191: 2083-2092Crossref PubMed Scopus (22) Google Scholar).Figure 4MS/MS fragmentation spectra of the parent ions in Fig. 3. Identified fragment ions of the b and y" series, and prominent ion peaks corresponding to secondary series, are indicated. Internal sequence ion GPNPI and related ones were assigned upon re-fragmentation (MS3) of the y6"+fragment ion. The GVISV and related internal sequence ions were assigned by MS3 refragmentation of the ion at m/z 456.0 (data not shown). The peptide sequences assigned are shown. The HVAVENAL and HAGVISVL sequences were confirmed by MS/MS fragmentation of the corresponding synthetic peptides. Nomenclature is as in Fig. 2.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Besides being B*3905 ligands, NAc-SHVAVENAL and EHAGVISVL were also isolated from B*3901 and B*3909 (16Yague J. Alvarez I. Rognan D. Ramos M. Vazquez J. Lopez de Castro J.A. J. Exp. Med. 2000; 191: 2083-2092Crossref PubMed Scopus (22) Google Scholar, 20Yague J. Ramos M. Vazquez J. Marina A. Albar J.P. Lopez de Castro J.A. Tissue Antigens. 1999; 53: 227-236Crossref PubMed Scopus (16) Google Scholar). In addition, IHEPEPHIL was also found from B*3901 in this study (TableI). Thus, the presence of the corresponding (−P1) peptides was searched in the peptide pools from these subtypes, using the same approach as in B*3905. As summarized in Table I, HVAVENAL, HEPEPHIL, and HAGVISVL were found in the B*3901-bound peptide pool. The former and latter peptides were also detected in B*3909. The identity of these peptides was established by MS/MS sequencing in all cases (Fig. 4 and data not shown).Table IIdentification of natural ligands of B*3901, B*3905, and B*3909 lacking the amino-terminal binding residue(+P1)/(−P1)B*3901B*3905B*3909MFractionMFractionMFractionNAc-SHVAVENAL/HVAVENAL980.4 /851.2155 /145980.4 /851.2155 /145980.5 /851.3155 /145EHGPNPIL/HGPNPIL875.4 /746.4156 /155IHEPEPHIL/HEPEPHIL1083.6 /970.4166 /1581083.6 /970.4166 /158EHAGVISVL/HAGVISVL923.3 /794.2178 /176923.5 /794.3177 /175923.4 /794.3177 /175Peptides lacking an amino-terminal binding residue (−P1) and their NH2-terminally extended counterparts (+P1) are indicated. The molecular mass (M) of each peptide, as determined by nanoelectrospray MS, and HPLC fraction number in which they eluted are indicated for each HLA-B39 allotype. Open table in a new tab Peptides lacking an amino-terminal binding residue (−P1) and their NH2-terminally extended counterparts (+P1) are indicated. The molecular mass (M) of each peptide, as determined by nanoelectrospray MS, and HPLC fraction number in which they eluted are indicated for each HLA-B39 allotype. Two control experiments were done to rule out the possibility that the (−P1) peptides found might result from residual exopeptidase activity during the isolation procedure, despite using protease inhibitors. The first experiment addressed the possibility of murine exopeptidase contamination in the W6/32-Sepharose immunoaffinity column used for purification of HLA-B39 from cell lysates. Known amounts of the synthetic NAc-SHVAVENAL, IHEPEPHIL, and EHAGVISVL peptides were incubated for 90 min into the column at pH 7.4, eluted with 0.1% aqueous TFA, and fractionated by HPLC in the same conditions used for fractionation of B39-bound peptide pools. (Fig.5 A). The three synthetic peptides were recovered with high yield. HPLC fractions around the retention times of the corresponding (−P1) peptides were analyzed by MALDI-TOF MS, and showed no evidence for these peptides (not shown). This result indicates that the (−P1) peptides found in the B39-bound peptide pools do not result from contamination of the immunoaffinity column by murine exopeptidases. The second experiment addressed the possibility that residual exopeptidase activity in the cell lysate could have contaminated the immunoaffinity column. Untransfected C1R cells were lysed exactly like the B39 transfectants, in the presence of protease inhibitors. Cell lysates were loaded into, incubated, and washed out from the W6/32 immunoaffinity column at neutral pH, using the same procedure as for purification of HLA-B39. The synthetic NAc-SHVAVENAL, IHEPEPHIL, and EHAGVISVL peptides were then loaded into the lysate-treated column, incubated for 10 min at neutral pH, and eluted with 0.1% aqueous TFA at lower rate than that used for elution of B39-bound peptide pools, to increase any putative exopeptidase action at this stage, and fractionated by HPLC (Fig. 5 B). Again, the synthetic precursors were recovered with high yield, and no evidence for the presence of the corresponding (−P1) peptides could be detected by MALDI-TOF MS analysis of the corresponding HPLC fractions (Fig.6). This result indicates that the (−P1) peptides in the B39-bound peptide pool do not result from exopeptidase activity in the cell lysates. For simplicity, ∼10-fold fewer cells than for peptide isolation were used in this control experiment, but the whole process, including treatment of the lysate with protease inhibitors, was correspondingly scaled down. Because exopeptidases are highly active enzymes, it is very unlikely that the lower number of cells used may have impaired our ability to detect any significant peptidase activity in the cell lysate, particularly with the incubation conditions used. Two pairs of synthetic peptides, EHAGVISVL/HAGVISVL and NAc-SHVAVENAL/HVAVENAL, were tested for binding to B*3901 and B*3909 at the cell surface, using an epitope stabilization assay with RMA-S transfectant cells (Fig. 7). For B*3901, binding of HAGVISVL was very efficient (C50: 3 ± 0.2 μm) and similar or slightly better than binding of its NH2-terminally extended counterpart EHAGVISVL (6 ± 0.5 μm). Binding of NAc-SHVAVENAL and HVAVENAL to B*3901 was also very similar to each other (C50: 19 ± 2 and 16 ± 2 μm, respectively) and significantly lower than binding of the other peptide pair. Similar results were obtained for B*3909 (Fig. 7); binding of each (−P1) peptide was very similar to binding of its NH2-terminally extended counterpart, and binding of the EHAGVISVL/HAGVISVL pair (C50: 10 ± 2 and 9 ± 2 μm, respectively) was significantly better than binding of the NAc-SHVAVENAL/HVAVENAL pair (C50: 28 ± 2 and 33 ± 2 μm, respectively). The EHAGVISVL/HAGVISVL and NAc-SHVAVENAL/HVAVENAL peptide pairs were also tested for their capacity to stabilize HLA-B*3901 on the surface of brefeldin A-treated RMA-S transfectant cells (Fig.8). Again, similar or slightly higher stability of HAGVISVL (DT50: 3.5 ± 0.5 h) in complex with B*3901 was observed, relative to EHAGVISVL (DT50: 2.5 ± 0.1 h). The stability of NAc-SHVAVENAL (DT50: 1.5 ± 0.2 h) was moderate, and lower than for the previous peptide pair, but similar to that of HVAVENAL (DT50: 1.62 ± 0.04 h). These results indicate that, for two pairs of natural HLA-B39 ligands, lack of the P1 residue have little or no effect on cell surface binding and stability to HLA-B39 in vitro. This study demonstrated that peptides lacking the amino-terminal binding residue are found in HLA-B39-bound peptide pools together with their NH2-terminally extended canonical counterparts. Four examples were identified using an MS-based approach. The trivial explanation that (−P1) peptides might have resulted from exopeptidase activity during the isolation procedure was ruled out by the following findings. (a) Cells were lysed in the presence of a mixture of protease inhibitors, following a standard method known to protect MHC class I ligands from degradation during purification; (b) the possibility of exopeptidase contamination of the immunoaffinity column arising either from the murine mAb or from the cell lysate was ruled out by control experiments showing absence of degradation of synthetic precursors. The experimental conditions in these control experiments would have favored the action of any putative exopeptidase, because the three synthetic peptides were incubated in the column at neutral pH, which favors the action of these enzymes. During isolation of class I-bound peptides, these are protected from enzymatic degradation in their bound state, and peptide dissociation was carried out at pH 2, which strongly disfavors most exopeptidase activity. Binding of the peptidic NH2 terminus in the A pocket makes a significant contribution to stability of MHC-peptide complexes. A network of H-bonds is established between the peptidic NH2-terminal group and A pocket residues from the MHC class I molecule (1Madden D.R. Annu. Rev. Immunol. 1995; 13: 587-622Crossref PubMed Scopus (708) Google Scholar, 4Madden D.R. Gorga J.C. Strominger J.L. Wiley D.C. Cell. 1992; 70: 1035-1048Abstract Full Text PDF PubMed Scopus (610) Google Scholar). Substitution of the peptidic NH2terminus by a methyl group decreases the Tm of the MHC-peptide complex by 22 °C (5Bouvier M. Wiley D.C. Science. 1994; 265: 398-402Crossref PubMed Scopus (244) Google Scholar). This significant decrease in stability could be explained by loss of three H-bonds in the A pocket, as a consequence of the fact that the CH3-terminal group was located away from the A pocket and the position naturally occupied by the NH2-terminal group was filled by a water molecule (28Bouvier M. Guo H.C. Smith K.J. Wiley D.C. Proteins. 1998; 33: 97-106Crossref PubMed Scopus (37) Google Scholar). This loss in stability contributes to explain that the overwhelming majority of classical MHC class I ligands have a free NH2 terminus. However, that this feature may occasionally be nonessential was recently demonstrated by our report of a natural N α-acetylated ligand of HLA-B39 (16Yague J. Alvarez I. Rognan D. Ramos M. Vazquez J. Lopez de Castro J.A. J. Exp. Med. 2000; 191: 2083-2092Crossref PubMed Scopus (22) Google Scholar). Although the mode of binding of this peptide to the class I molecule remains unknown, it was suggested that N α -acetylated P1 residue might not occupy the A pocket and this might be occupied by water molecules that would partially compensate for the absence of a free NH2-terminal group. A recent report described the crystal structure of an HTLV-1-derived Tax8 peptide lacking the amino-terminal binding residue, refolded in vitro with HLA-A*0201 (17Khan A.R. Baker B.M. Ghosh P. Biddison W.E. Wiley D.C. J. Immunol. 2000; 164: 6398-6405Crossref PubMed Scopus (145) Google Scholar). In this complex, the conformation of the bound peptide was very similar to that of its NH2-terminally extended Tax9 counterpart, except at the A pocket, which was occupied only by two water molecules. This mode of binding was less stable than binding of the canonical counterpart (ΔTm = −16 °C), as also reported for H-2 K d (29Fahnestock M.L. Johnson J.L. Feldman R.M. Tsomides T.J. Mayer J. Narhi L.O. Bjorkman P.J. Biochemistry. 1994; 33: 8149-8158Crossref PubMed Scopus (45) Google Scholar), but did not prevent formation of a "closed conformation" of the peptide binding site. In an earlier report (18Utz U. Koenig S. Coligan J.E. Biddison W.E. J. Immunol. 1992; 149: 214-221PubMed Google Scholar), Tax8 sensitized HLA-A2-positive targets for lysis by Tax-specific CTL from HTLV-1-infected individuals expanded in vitro with this peptide. Although these results are compatible with Tax8 being a natural HLA-A2 ligand, this seems unlikely on the basis of the limited stability of Tax8/HLA-A2 complexes (17Khan A.R. Baker B.M. Ghosh P. Biddison W.E. Wiley D.C. J. Immunol. 2000; 164: 6398-6405Crossref PubMed Scopus (145) Google Scholar). Furthermore, Utz et al. (18Utz U. Koenig S. Coligan J.E. Biddison W.E. J. Immunol. 1992; 149: 214-221PubMed Google Scholar) did not rule out the possibility that the CTL reactivity observed could be the result of cross-reaction of some CTL activated in vivo against the dominant Tax9 epitope. That such cross-reaction did not occur with a Tax9-specific CTL clone (17Khan A.R. Baker B.M. Ghosh P. Biddison W.E. Wiley D.C. J. Immunol. 2000; 164: 6398-6405Crossref PubMed Scopus (145) Google Scholar) does not rule out this possibility in a polyclonal T-cell response. For instance, Tax-specific CTL clones critically depend on interactions with a cluster of 3 residues (Arg65, Lys66, Ala69) in the HLA-A2 α1-helix, but reactivity of individual clones is affected differently by mutations at each of these positions (30Baker B.M. Turner R.V. Gagnon S.J. Wiley D.C. Biddison W.E. J. Exp. Med. 2001; 193: 551-562Crossref PubMed Scopus (76) Google Scholar), which underlines clonal diversity. The similar binding of HVAVENAL relative to NAc-SHVAVENAL could be explained on the basis that the latter peptide lacks a free NH2 terminus, so that the NAc-Ser group may not bind in the A pocket, and the binding mode of the acetylated ligand might actually be similar to HVAVENAL. However, this explanation cannot apply to the similar behavior of the EHAGVISVL/HAGVISVL peptide pair in our in vitro assays. It should be noted that loss of a P1 residue with free NH2terminus is reflected in our epitope stabilization assay. For instance, binding of HVAVENAL or NAc-SHAVAVENAL was ∼10-fold lower in this assay than binding of SHVAVENAL (16Yague J. Alvarez I. Rognan D. Ramos M. Vazquez J. Lopez de Castro J.A. J. Exp. Med. 2000; 191: 2083-2092Crossref PubMed Scopus (22) Google Scholar). Two alternative explanations can be proposed for the similar binding of EHAGVISVL/HAGVISVL. First, stabilization of the (−P1) peptide by sequence-dependent anchors might be high enough as to make the relative contribution of the canonical NH2 terminus to the overall binding affinity and stability sufficiently small to go unnoticed in our assays. This is unlikely to be a general rule and would explain why only few (−P1) peptides have been found. Related to this, it is possible that removal of P1 residues, whose side chains make a negative contribution to the overall binding energy in peptides well stabilized by other anchors, might favor presence of the corresponding (−P1) ligands. A second possibility is that stabilization of the A pocket by H-bonded water molecules might be higher for HLA-B39 ligands with His as the B pocket-binding residue, than for ligands of HLA-A2 or other class I molecules. Structural differences among class I molecules in the A pocket, and the polarity of the peptidic His, might provide a possible basis for higher water-mediated stabilization of the MHC/(−P1) peptide complex. For instance, although the three HLA-B39 allotypes in our study bind peptides with either Arg2 or His2, the four (−P1) ligands found had His as the B pocket-binding motif. Moreover, after extensive studies in our laboratory with HLA-B27-bound peptides, which have Arg2, we never found evidence for (−P1) peptides from this class I protein. Ultimately, x-ray diffraction studies would be required to establish the binding mode of natural His-containing (−P1) ligands in complex with HLA-B39. The apparently similar binding properties of the (−P1) peptides described in this study relative to their NH2-terminally extended counterparts, their structural similarities, notably the NH2-terminal His residue, and the well established contribution of the canonical NH2 terminus to peptide stability (5Bouvier M. Wiley D.C. Science. 1994; 265: 398-402Crossref PubMed Scopus (244) Google Scholar), suggest that occurrence of (−P1) ligands in vivo is rare among class I MHC proteins, although it may be somewhat more frequent in particular allotypes, such as HLA-B39. (−P1) ligands arising from internal protein sequences may be directly produced by the proteasome, along with their NH2-terminally extended counterparts, or result from aminopeptidase trimming further down in the antigen-processing pathway. However, finding of the HVAVENAL ligand was surprising because its NH2-terminally extended counterpart was N α-acetylated, and the sequence corresponded to the amino-terminal portion of a helicase (31You L.R. Chen C.M. Yeh T.S. Tsai T.Y. Mai R.T. Lin C.H. Lee Y.H. J. Virol. 1999; 73: 2841-2853Crossref PubMed Google Scholar). In vitro digestion of an N α-acetylated synthetic precursor with 20 S proteasome resulted in direct generation of NAc-SHVAVENAL and absence of cleavage after NAc-Ser (16Yague J. Alvarez I. Rognan D. Ramos M. Vazquez J. Lopez de Castro J.A. J. Exp. Med. 2000; 191: 2083-2092Crossref PubMed Scopus (22) Google Scholar). Several possibilities may be considered to explain the presence of HVAVENAL as a natural HLA-B39 ligand. First, it could arise from an internal protein region, independent from the DBX helicase from which NAc-SHVAVENAL most likely arose (16Yague J. Alvarez I. Rognan D. Ramos M. Vazquez J. Lopez de Castro J.A. J. Exp. Med. 2000; 191: 2083-2092Crossref PubMed Scopus (22) Google Scholar). We cannot rule out this possibility, but consider it unlikely because HVAVENAL did not match any sequence in the protein data base other than that of the DBX helicase. Second, HVAVENAL might be directly generated by the proteasome before NH2-terminal acetylation, for instance, during degradation of incorrect DBX polypeptides. This possibility would also appear unlikely because the proteasome cleaves inefficiently near the free NH2 termini of protein substrates. A third possibility, which we favor, is that HVAVENAL is generated after proteasomal generation of a peptide precursor by amino-peptidase trimming. Such precursor might arise from nonacetylated or acetylated DBX polypeptides. The latter situation would imply a role of cytosolic acyl-amino acid-releasing enzymes (32Tsunasawa S. Narita K. Ogata K. J. Biochem. (Tokyo). 1975; 77: 89-102PubMed Google Scholar, 33Mitta M. Ohnogi H. Mizutani S. Sakiyama F. Kato I. Tsunasawa S. DNA Res. 1996; 3: 31-35Crossref PubMed Scopus (14) Google Scholar, 34Scaloni A. Ingallinella P. Andolfo A. Jones W. Marino G. Manning J.M. J. Protein Chem. 1999; 18: 349-360Crossref PubMed Scopus (9) Google Scholar) in the generation of a class I ligand. In conclusion, our results demonstrate, for the first time to our knowledge, that classical MHC class I molecules bind in vivopeptides lacking the canonical P1 residue. Together with our previous report of an N α-acetylated ligand (16Yague J. Alvarez I. Rognan D. Ramos M. Vazquez J. Lopez de Castro J.A. J. Exp. Med. 2000; 191: 2083-2092Crossref PubMed Scopus (22) Google Scholar), this is a second example of natural class I ligands lacking a canonical structure at their NH2 termini. The functional significance of (−P1) ligands is unknown, but their presence should be taken into account for epitope prediction and mapping, both in defensive T-cell responses and autoimmunity. We thank Samuel Ogueta, Rosana Rogado, and Alberto Monteagudo for help in MS and Fernando Barahona for peptide synthesis.

Referência(s)