Background: Previous studies suggest that heterosexually transmitted subtype A and subtype C HIV, but not non-heterosexually transmitted subtype B HIV, possess signature genotypes of compact and less glycosylated envelope variable loops. To understand whether other heterosexually transmitted HIV also displays similar or other signature sequence characteristics, we examined envelope sequences from 13 donor-recipient pairs in Uganda.

Methods: From a trial of sexually transmitted diseases treatment to prevent HIV-1 acquisition in the Rakai district of Uganda, newly infected subjects with documented HIV-1 seroconversion over prospective follow-up and their epidemiologically identified monogamous heterosexual partner were examined in this study. A minimum of 8 full-length envelope sequences were isolated and examined from the earliest available serum sample after HIV-1 acquisition from the newly infected subject and the corresponding collection day matching plasma sample from the transmitting partner.

Results: Samples from the newly infected partner were collected a median of 189 days after estimated HIV-1 seroconversion (range 142 to 359 days). Each recipient's sequences clustered with the corresponding donor's sequences in neighbor joining phylogenetic analysis, which confirmed the epidemiological linkage. Of the 13 couples, 10 were infected with subtype D HIV-1 and the remaining 3 pairs had subtype A HIV-1. Sequences in the recipients showed lower diversity (p <0.001) and divergence (p <0.001) compared to the donor sequences. Recipient envelope V1-V2 (p = 0.002) and V1-V4 (p <0.001) sequences were significantly shorter compared to donor sequences (Wilcoxon rank-sum test stratified by couple). There was no significant difference in the number of potential N-linked glycosylation sites between recipient and donor sequences within the V1-V2 (p = 0.91) or V1-V4 (p = 0.40) segments. Within each pair, signature amino acids differentiated the donor and recipient sequences although these differences were not consistent among all couples.

Conclusions: Our results suggest that a small subset of HIV closely related to ancestral viruses are selected during heterosexual transmission of predominantly subtype D HIV-1 in Uganda. Similar to heterosexually transmitted subtype A and subtype C HIV, viruses in newly infected Ugandan subjects also have shorter envelope variable loops but they do not have a significantly lower number of potential N-linked glycosylation sites.