GROUPING HLA-B LOCUS SEROLOGIC SPECIFICITIES ACCORDING TO SHARED STRUCTURAL MOTIFS SUGGESTS THAT DIFFERENT PEPTIDE-ANCHORING POCKETS MAY HAVE CONTRASTING INFLUENCES ON THE COURSE OF HIV-1 INFECTION

Two different groups of HLA-B specificities were associated with two contrasting outcomes of HIV-1 infection. HLA-B45, -B49, and -B50 were each found at a moderately increased frequency among individuals responding to HIV-1 infection with a marked circulating and infiltrative CD8 T-cell lymphocytosis, a slow rate of CD4 T-cell decline, very low frequency of opportunistic infections, and low viral strain heterogeneity. In contrast, among HIV-infected individuals with more rapid progression to opportunistic infections, HLA-B35 was found to be increased in frequency and to act as a dominant marker for this adverse outcome. HLA-B45, -B49, and -B50 contain identical peptide-anchoring ''B'' and ''C-terminal'' pocket structures, which differ greatly from those present in HLA-B35, implying that different immunogenic peptides are likely to be bound by these two groups of alleles. Placing HLA-B45, -B49, and -B50 into one structurally defined group revealed a much stronger and statistically significant association with the CD8 lymphocytosis syndrome (OR = 5.3, p = 0.0005). The B pocket structure in these alleles contains an easily accessible lysine residue at position 45, suggesting that the P2 or P3 anchor residue of a bound peptide is negatively charged. Additionally, by observing the effect on the ORs of adding structures containing amino acid substitutions in the C-terminal pocket- of HLA-B45, -B49, and -B50, this region was also shown to influence susceptibility to this host response. These results suggest that, when analyzing HLA disease susceptibility and resistance associations, grouping HLA specificities according to regions of shared structure and resistance associations, grouping HLA specificities according to regions of shared structure and peptide-binding function may provide insight into disease mechanisms involving binding of immunogenic peptides. Moreover, they suggest the possibility that various peptide-anchoring structures present in MHC class I molecules may contribute ro different outcomes of HIV-1 infection, possibly by initiating qualitatively different host immune responses.