HIV-1 sequences in the epidemic suggest an alternative pathway for the generation of the Long Terminal Repeats

To generate the long-terminal repeats (LTR) that border the integrated viral genome, two-strand transfer steps must occur during reverse transcription. Analysis of the genetic polymorphisms that are present in the LTR of HIV-1 heterozygous virions in single infection cycle studies has revealed which of the two copies of genomic RNAs is used for each transfer event. Thus, the first event of strand transfer has been described to be either intra- or intermolecular, while the second event is generally intramolecular. Here, we repeated these analyses using sequences from HIV databases and extended the study to the regions surrounding the LTR. We observed a striking correlation between the pattern of recombination in the LTR and the phylogenetic origin of the surrounding sequences. This correlation suggests that the second-strand transfer can be either intra- or intermolecular and, interestingly, could reflect an effect of proximity between nucleic acids that would guide this transfer. This factor could be particularly relevant for heterozygous viruses containing highly divergent genomic RNAs, such as those considered in the present study.