How to recover from a bad start: adaptation of HIV-1 transcription start site mutants during serial passaging in culture

HIV-1 uses neighboring sequences as transcription start sites and generates multiple unspliced transcripts, including two major transcripts with three guanosines (3G) or one guanosine (1G) at the 5' end. Although only differing by 2-nt, 3G RNA and 1G RNA are functionally distinct. We have previously generated two mutants: the TTG virus mainly expresses 1G RNA, whereas the plusAC virus predominantly expresses 3G RNA. Both mutant viruses are replication-competent but exhibit fitness defects. Here, we passaged the plusAC virus in T cells and characterized the changes near the transcription start sites. We observed the rapid loss of the plusAC virus genotype and the emergence of multiple revertants. All major revertants that dominated the cultures had a 1- to 3-nt deletion that compensated for the dinucleotide insertion in the plusAC virus. These major revertants express more than one major transcript, preferentially package 1G RNA, and have improved replication kinetics compared with the plusAC virus. Most major revertants likely arose through errors during reverse transcription, including misalignments during minus-strand DNA transfer, nucleotide deletion in a homopolymer run, or deletion of a short direct repeat. Additionally, we have determined that a T-to-G substitution near transcription start sites occurs at ~5% per replication cycle by copying the guanosine cap. Converting a base to guanosine through cap copying has been observed in multiple positions, but always directly upstream of a major transcription start site. Taken together, our findings demonstrate the selection pressure for expressing functionally distinct unspliced RNA species to optimize replication fitness.IMPORTANCEHIV-1 unspliced RNA serves as the mRNA to translate Gag/Gag-Pol polyproteins and as the virion genome. HIV-1 produces two major RNA species: 1G RNA is preferentially packaged and 3G RNA is favorably translated, although each transcript can perform both functions. We have previously generated a replication-competent mutant virus that mainly expresses 3G RNA and observed that this mutant has replication fitness defects. We found that the mutant virus improved its replication kinetics after passaging, indicating adaptation. Our analyses showed that, through mutations occurring during DNA synthesis, multiple revertants arose rapidly to replace the input mutant virus. The major revertants regained the ability to generate more than one major transcript and preferentially package 1G RNA. These results highlight the importance of expressing HIV-1 RNA species that serve distinct functions and the ability of HIV-1 to adapt through mutations in the genome.