Naturally occurring polymorphisms in the protease of human immunodeficiency virus type 1 (HIV-1)subtype C would be expected to lead to adaptive (compensatory) changes in protease cleavage sites. To test this hypothesis, we examined the prevalences and patterns of cleavage site polymorphisms in the Gag, Gag-Pol, and Nef cleavage sites of C compared to those in non-C subtypes. Codon-based maximum-likelihood methods were used to assess the natural selection and evolutionary history of individual cleavage sites. Seven cleavage sites
(p17/p24, p24/p2, NC/p1, NC/TFP, PR/RT, RT/p66, and p66/IN) were well conserved over time and in all HIV-1 subtypes. One site (p1/p6gag) exhibited moderate variation, and four sites (p2/NC, TFP/p6pol, p6pol/PR, and Nef)
were highly variable, both within and between subtypes. Three of the variable sites are known to be major determinants of polyprotein processing and virion production. P2/NC controls the rate and order of cleavage, p6gag is an important phosphoprotein required for virion release, and TFP/p6pol, a novel cleavage site in the transframe domain, influences the specificity of Gag-Pol processing and the activation of protease. Overall, 58.3% of the 12 HIV-1 cleavage sites were significantly more diverse in C than in B viruses. When analyzed as
a single concatenated fragment of 360 bp, 96.0% of group M cleavage site sequences fell into subtype-specific phylogenetic clusters, suggesting that they coevolved with the virus. Natural variation at C cleavage sites may play an important role, not only in regulation of the viral cycle but also in disease progression and response to therapy.