Because T cell responses are critical for defense against viral infections, a series of synthetic peptides derived from the predicted sequence for HIV-1 proteins gp41, pg120, gag, and viral polymerase were used to test the T cell proliferative response of HIV-1 seropositive individuals. Of HIV-1-infected donors from various clinical categories 90% (27/30) had sensitized cells that proliferated in response to at least one of 21 HIV peptides tested. Cells from HIV seronegative controls did not proliferate (0/9) in response to these HIV peptides. Individuals with fewer clinical manifestations of HIV-1 disease responded to a greater number of peptides (average for asymptomatic seropositives = 8.1 peptides; AIDS patients averaged 2.0). The number of peptides recognized also correlated with absolute number of CD4+ cells, but not with delayed cutaneous hypersensitivity to a (non-HIV) battery of Ag. However, clinical stage at no time correlated with the response to any particular peptide. Response patterns differed considerably among individuals, and some peptides stimulated proliferation in many (48%) HIV-infected donors (peptides gp41-2 and pol-3), whereas another peptide elicited no T cell response in any donor tested (peptide gp120-8). We have also begun to investigate the basis for individual heterogeneity of T lymphocyte proliferative responses of HIV-infected donors to the 21 HIV synthetic peptides. Peptide structure and HLA class II determinants both influenced patterns of lymphocyte responses. Reactivity correlated with peptide size, the presence of alpha and beta secondary structure and lack of reverse turn potential. Hydropathy and charge had no predictive value. Peptides derived from HIV sequences that vary highly among strains tended to be recognized less frequently. HIV-infected lymphocyte donors were HLA typed to examine the influence of the MHC on T lymphocyte proliferation. Analysis of the frequencies of individuals reacting to specific peptides, when compared to the allele frequencies in the population at large, indicated association of some responses to DR alleles. More DR association was observed with peptides that showed "moderate" reactivity than with those that were "highly" reactive. We suggest that highly reactive peptides are capable of forming a structure closer to an "ideal" T cell epitope that can associate with many DR alleles. In contrast, "moderately" reactive determinants have less favorable structures for interaction, are more limited in their ability to interact and therefore show more restriction to specific class II alleles.