Associations between Amino Acids in the Evolution of HIV Type 1 Protease Sequences under Indinavir Therapy

Significant diversity exists in amino acid sequences encoding HIV-1 protease in individuals naive for protease inhibitors, which could influence the rate of evolution of resistance. High-level resistance to indinavir requires multiple substitutions among at least 11 amino acid sites, and no single substitution was observed in all of 29 resistant isolates obtained from patients on long-term indinavir monotherapy. We have analyzed the evolution of PR in these sequences. The divergence from the baseline amino acid sequence by week 24 was 4%, increasing more than 7% by week 60. The mean difference between sequences from different patients at baseline was 6% (3-9%), rising to 10% after 40 weeks (3-16%), although at all time points nonsynonymous substitutions were less frequent than synonymous nucleotide changes. Analysis of associations between variants at different amino acid sites using a mutual information statistic revealed four pairs of sites to be significantly associated. In three cases these associations included residue 82. Clusters of baseline and week 24 amino acid sequences identified by maximum parsimony did not correlate significantly with the IC95 to indinavir, although a weak correlation of baseline clusters with phenotype at the week 24 time point was suggested.