Sequence analysis of the polymerase domain of HIV-1 reverse transcriptase in naive and zidovudine-treated individuals reveals a higher polymorphism in alpha-helices as compared with beta-strands.

We report a statistical analysis of genetic heterogeneity of the reverse transcriptase (RT)-coding region of human immunodeficiency virus type 1. Both newly determined sequences and sequences contained in the data banks have been examined. For the calculations, the viral samples and the regions within the RT molecule were divided in two groups. The viral samples were split into those from patients not subjected to antiretroviral therapy and those from patients treated with zidovudine (AZT, 3'-azido-3'-deoxythymidine) alone or in combination with other RT inhibitors. The RT-coding region was divided into segments encoding beta-strands and segments encoding alpha-helices. A significantly lower heterogeneity was observed in beta-strands relative to the alpha-helix coding segments. Application of the D test of Tajima has provided evidence of operation of negative (or purifying) selection in sequences from viruses of patients not subjected to antiretroviral treatment as well as in treated patients. In the group of untreated individuals, regions encoding beta-strands are subjected to stronger negative selection than those encoding alpha-helices. It is likely that the observed differences reflect stronger functional constraints in beta-strands than in alpha-helices of RT.