USE OF COMPLETE ELUTED PEPTIDE SEQUENCE DATA FROM HLA-DR AND HLA-DQ MOLECULES TO PREDICT T-CELL EPITOPES, AND THE INFLUENCE OF THE NONBINDING TERMINAL REGIONS OF LIGANDS IN EPITOPE SELECTION

In diseases with a strong association with an HLA haplotype, identific ation of relevant T cell epitopes may allow alteration of the patholog ic process. In this report we use a reverse immunogenetic approach to predict possible HLA class II-restricted T cell epitopes by using comp lete pool sequencing data. Data from HLA-DR2(B11501), -DR3(B1*0301), -DQ2(A10501, B1*0201), and -DQ8(A1*0301, B1*0302) alleles were used b y a computer program that searches a candidate protein to predict liga nds with a relatively high probability of being processed and presente d, This approach successfully identified both known T cell epitopes an d eluted single peptides from the parent protein. Furthermore, the pro gram identified ligands from proteins in which the binding motif of th e HLA molecule was unable to do so. When the information from the nonb inding N- and C-terminal regions in the pool sequence was removed, the ability to predict several ligands was markedly reduced, particularly for the HLA-DQ alleles, This suggests a possible role for these regio ns in determining ligands for HLA class II molecules, Thus, the use of complete eluted peptide sequence data offers a powerful approach to t he prediction of HLA-DQ and -DR peptide ligands and T cell epitopes.