Gag-specific CTLs are known to have stronger ability to control HIV-1 replication than others that are protein-specific. Therefore, the analysis of Gag escape mutants is expected to clarify the mechanisms of immune control in HIV-1-infected donors. However, only a limited number of Gag escape mutants have been identified so far. A previous study suggested the possibility that Gag28-3R (KW9-3R) is an escape mutant from HLA-A*2402-restricted KW9-specific CTLs but did not show any evidence of it. Here we sought to demonstrate that KW9-3R is selected as escape mutant by KW9-specific CTLs. KW9-specific CTLs showed a remarkable reduction in recognition of target cells infected with the KW9-3R mutant. The sequence analysis of HIV-1 from 58 HIV-1-infected individuals showed that the frequency of the KW9-3R mutant was significantly higher in HLA-A*2402(+) individuals than in HLA-A*2402(-) individuals. Longitudinal analysis of an HLA-A*2402(+) individual with HIV-1 early infection showed that this escape mutant was selected over an approximately 2-year period. These results together indicate that Gag28-3R is an escape mutant selected by HLA-A*2402-restricted KW9-specific CTLs. Further analysis of this epitope will clarify the role of HIV-1-specific CTLs in the control of HIV-1 among the Japanese population, since 70% of them carry this allele.