Effects of varying sequence similarity on the frequency of repeat deletion during reverse transcription of a human immunodeficiency virus type 1 vector

Abstract

Genetic recombination contributes to human immunodeficiency virus type 1 (HIV-1) diversity, with homologous recombination being more frequent than nonhomologous recombination. In this study, HIV-1-based vectors were used to assay the effects of various extents of sequence divergence on the frequency of the recombination-related property of repeat deletion. Sequence variation, similar in degree to that which differentiates natural HIV-1 isolates, was introduced by synonymous substitutions into a gene segment. Repeated copies of this segment were then introduced into assay vectors. With the use of a phenotypic screen, the deletion frequency of identical repeats was compared to the frequencies of repeats that differed in sequence by various extents. During HIV-1 reverse transcription, the deletion frequency observed with repeats that differed by 5% was 65% of that observed with identical repeats. The deletion frequency decreased to 26% for repeats that differed by 9%, and when repeats differed by 18%, the deletion frequency was about 5% of the identical repeat value. Deletion frequencies fell to less than 0.3% of identical repeat values when genetic distances of 27% or more were examined. These data argue that genetic variation is not as inhibitory to HIV-1 repeat deletion as it is to the corresponding cellular process and suggest that, for sequences that differ by about 25% or more, HIV-1 recombination directed by sequence homology may be no more frequent than that which is homology independent.

FIG. 1.

HIV-1-transducing vectors. (A) Structure of pHIVlacWT. LTR, long terminal repeat; Ψ, packaging signal; RRE, Rev response element; CMV, cytomegalovirus immediate-early promoter; SV40, simian virus 40 promoter; puro, puromycin resistance gene. (B) Nucleotide sequence alignment of 156-nt EcoRI-XhoI lacZ fragments from pHIVlacWT, pHIVlac95%, pHIVlac91%, pHIVlac82%, pHIVlac73%, pHIVlac63%, and pHIVlac58%. Nucleotides of the variants that differed from wild type sequence are highlighted (white type on black background). (C) Structure of pHIVlaac95%-WT, a representative repeat vector. Its 5′ repeat was derived from the 156-nt lacZ fragment of pHIVlac95%, and its 3′ repeat was derived from pHIVlacWT. Abbreviations are as defined for panel A.

FIG. 2.

Repeat deletion rates for vectors with identical lacZ variant repeats. The repeat vector structures are illustrated with boxed percentages (e.g., the first repeat vector structure listed represents pHIVlaac95%-95%). The homology between repeats was 100% for each vector in this series. The repeat deletion rates were calculated as described in the text (means [bars] minus standard errors of the means [error bars] are shown).

FIG. 3.

Effects of various extents of sequence identity on repeat deletion frequencies. (A) Repeat deletion rates for vectors with wild- type lacZ sequence in the 3′ repeat. The repeat vector structures are illustrated with boxed percentages (the 5′ and 3′ repeats are as indicated in the left and right boxes, respectively). For example, the second repeat vector structure listed represents pHIVlaac95%-WT, where the homology between repeats was 95%. Deletion rates were calculated as described in the text (means [bars] minus standard errors of the means [error bars] are shown). (B) Repeat deletion rates for vectors with wild-type lacZ sequence in the 5′ repeat. The methods for illustrating repeat vector structures and calculation of deletion rates are as described for panel A.