Of several thousand peptides presented by the major histocompatibility molecule HLA-A2.1, at least nine are recognized by melanoma-specific cytotoxic T lymphocytes (CTLs). Tandem mass spectrometry was used to identify and to sequence one of these peptide epitopes. Melanoma-specific CTLs had an exceptionally high affinity for this nine-residue peptide, which reconstituted an epitope for CTL lines from each of five different melanoma patients tested. Recognition by multiple CTL lines suggests that this may be a promising candidate for use in peptide-based melanoma vaccines.

SummaryT lymphocytes recognize antigens consisting ofpeptides presented by class I and II major histocompatibility complex (MHC) molecules. The peptides identified so far have been predictable from the amino acid sequences of proteins. We have identified the natural peptide target of a CTL clone that recognizes the tyrosinase gene product on melanoma cells. The peptide results from posttranslational conversion ofasparagine to aspartic acid. This change is of central importance for peptide recognition by melanoma-specific T cells, but has no impact on peptide binding to the MHC molecule. This posttranslational modification has not been previously described for any MHC-associated peptide and represents the first demonstration of posttranslational modification of a naturally processed class I-associated peptide. This observation is relevant to the identification and prediction of potential peptide antigens. The most likely mechanism for production of this peptide leads to the suggestion that antigenic peptides can be derived from proteins that are translated into the endoplasmic reticulum.C lass I molecules of MHC bind to peptides derived from intracellular pathogens or from proteins expressed in tumor cells, and present them on the cell surface to the host immune system (1-3). Identification of the specific peptides that constitute T cell epitopes has been difficult without prior knowledge of the source protein. However, peptides recognized by human melanoma--specific T cells have recently been identified from five proteins using two alternative strategies. One approach has been to generate genomic or cDNA libraries from tumor cells followed by transfection of progressively smaller subsets of these molecular clones into cells that express the appropriate MHC molecule, but not the tumor-specific epitope (4--14). Molecular clones that encode T cell epitopes are identified by their ability to reconstitute tumor-specific T cell recognition of the transfected cells. The exact T cell epitope is then identified by a combination of molecular subcloningThe contributions of the first two authors were equivalent and their order should be considered arbitrary. and the use of synthetic peptides based on the predicted amino acid sequence. This approach led to the identification of antigens encoded by genes whose expression is specific for tumors, such as MAGE and of other antigens related to melanocyte differentiation such as tyrosinase (4, 13). In the second approach, naturally occurring peptides associated with MHC molecules on the tumor cells are directly extracted, fractionated by HPLC, and used to reconstitute recognition by melanoma-specific CTL of a nonmelanoma cell expressing appropriate MHC molecules (15). The peptide epitope within a reconstituting peptide fraction is identified and sequenced by tandem mass spectrometry (16,17). Using this approach, a peptide, YLEPG-PVTA, from the protein Pmel-17/gp100, was identified as an epitope for HLA-A2.1-restricted, melanoma-specific CTL from multiple individuals (18).

Minor histocompatibility antigen disparities between human leukocyte antigen (HLA)-matched bone marrow donors and recipients are a major risk factor for graft versus host disease (GVHD). An HLA-A2.1-restricted cytotoxic T cell clone that recognized the minor histocompatibility antigen HA-2 was previously isolated from a patient with severe GVHD after HLA-identical bone marrow transplantation. The HLA-A2.1-bound peptide representing HA-2 has now been identified. This peptide appears to originate from a member of the non-filament-forming class I myosin family. Because HA-2 has a phenotype frequency of 95 percent in the HLA-A2.1-positive population, it is a candidate for immunotherapeutic intervention in bone marrow transplantation.