Significant amounts of different tRNA molecules are present in retroviral particles, but one specific tRNA species functions as primer in reverse transcription. It is generally believed that the HIV-1 virus uses the tRNA(Lys,3) molecule as primer. This is based on sequence complementarity between the 3' end of tRNA(Lys,3) and the primer-binding site (PBS) on HIV-1 genomic RNA. Recent biochemical analyses indicated that tRNA(LYs,3) is indeed incorporated into viral particles. Interestingly, tRNA(Lys,3) could not be detected in virions produced by HeLa-CD4+ cells [(1992) Biochem. Biophys. Res. Commun. 185, 1105-1115]. In order to test whether alternative tRNA molecules can function as primer in HIV replication, we performed a series of experiments based on the observation that tRNA primer sequences are inherited by the viral progeny. We cultured HIV-1 for prolonged periods of time in HeLa-CD4+ cells, but did not detect sequence changes in the PBS region. Furthermore, we found PBS-mutants to be replication-incompetent, again suggesting that HIV-1 solely uses tRNA(Lys,3) as primer. Most importantly, we obtained revertants of one such PBS-mutant, which had restored a wild-type PBS sequence. This tRNA(Lys,3)-mediated repair demonstrates a general requirement for this primer in HIV-1 reverse transcription.