Background:  The implications of specific mutations in HIV-1 pol for antiretroviral resistance are well known, but HIV-1 pol evolution over the course of infection without drug pressure is less clear. A better understanding of these evolutionary dynamics may assist molecular epidemiology studies that use widely available pol genotypic data. This study aimed to determine the velocity of HIV-1 pol evolution, the rate of nucleotide substitutions between pol sequences sampled longitudinally (distance over time), in a large primary HIV infection cohort as antiretroviral-naïve individuals progress from primary to chronic infection.

Methods:  Participants infected with HIV-1 subtype B had >1 population-based pol sequence (Viroseq v2.0, 1300 bp), remained antiretroviral-naïve (>6 months) and were not found to be dually infected during the study period. Analysis included: HIV-1 pol evolution, resistance-associated mutations, CD4 count, viral load, and HLA haplotype. Using a Geneious Pro package, HIV-1 pol velocity was calculated by comparing an individual’s baseline sequence to their latest available sequence and observed nucleotide substitutions were divided by sequence length and time from baseline. In addition, for individuals with ≥3 population-based pol sequences, we compared HIV-1 pol velocity derived from intervals between baseline to earliest time point and baseline to latest time point using the Wilcoxon matched pairs test.

Results:  Ninety-five individuals enrolled in a primary infection cohort (1 July 1996, through 30 May 2007) were included with a mean period of observation of 2.74 years (min = 0.49, max = 8.97) since primary infection. The mean nucleotide substitution per site per year in HIV-1 pol was 4.181x10^–3 (SD = 2.570x10^–3). Even after 7 years of follow up, no participant had greater than 2% divergence between any 2 sequences. This velocity of pol was not associated with HLA, CD4, or viral load. HIV-1 pol velocity was higher at earlier time points versus later time points (p <0.0001).

Conclusions:  HIV-1 pol evolution was time dependent with the rate of nucleotide substitutions occurring at a faster pace during the earliest stages of infection. However, at any stage, the velocity of HIV-1 subtype B pol velocity was rather slow, thereby supporting the use of these types of data to characterize transmission networks.