In H-2d mice, the immunodominant determinant of the HIV-1-IIIB gp160 envelope glycoprotein recognized by CD8+ CTL is represented by a 15-residue synthetic peptide (315-329: RIQRGPGRAFVTIGK). This peptide is seen in association with the Dd class I MHC molecule expressed on H-2k L cell fibroblast targets. We explored the structural requirements for CTL recognition of this peptide at the levels of both the peptide molecule and the class I MHC molecule. Using several transfectants expressing recombinant Dd/Ld molecules, we found that presentation of this epitope required both the alpha 1 and alpha 2 domains of the Dd molecule, in contrast to certain instances of allorecognition for which alpha 1 of Dd was sufficient in association with alpha 2 of Ld. Because this peptide derives from a hypervariable segment of the HIV envelope, substituted peptides could be used to define not only the structures affecting interaction of peptide with class I MHC molecule and with the TCR, but also the structural basis for the effect of naturally occurring viral variation on CTL recognition. The CTL-LINE specific for this HIV-1-IIIB-derived sequence could not recognize the HIV-1-RF variant-derived sequence from exactly the same site (315-329:--HIGPGRVIYATGQ). Peptides with single amino acid substitutions from the HIV-1-IIIB sequence toward the HIV-1-RF sequence were made to test the effect of each residue significantly affected recognition, and only one, 324(F), was obligatory. Moreover, both 322(R) and 324(F) substituted peptides failed to inhibit the binding of the wild type peptide to the MHC molecule. Therefore, the amino-acids 322(R) and 324(F) seem to be involved in regulating peptide interaction with the Dd class I MHC molecule. In contrast, 325(V) appeared to affect interaction with the TCR. We suggest that sequence variations among known HIV-1 isolates that affect peptide binding to MHC such as those described here, if occurring during the course of infection of an individual, could result in failure of the MHC molecules of that individual to present the peptide. If the number of dominant HIV CTL epitopes is indeed very limited, such a blind spot could allow the virus to escape immune control, proliferate rapidly, and cause AIDS.