HIV-1 Gag, Pol, and Env diversified with limited adaptation since the 1980s.

Knowledge of HIV-1 global sequence diversity is critical for developing an effective prophylactic against HIV-1 infection. We developed the Hervé platform to analyze and visualize trends in HIV-1 diversification. Using Hervé, we analyzed 4,830 Env, 4,407 Gag, and 3,002 Pol publicly available independent sequences corresponding to subtypes A1, A6, B, C, D, F1, and G and circulating recombinant forms (CRFs) 01_AE, 02_AG, and 07_BC; sequences were sampled between 1980 and 2020 from 82 countries. HIV-1 diversified with a median of 1.82 amino acid substitutions per year in Env, 0.297 in Gag, and 0.779 in Pol. Yet, Env subtype B diversification plateaued post-2000. Pairwise diversity within subtypes and CRFs increased by 41.82% (range = 24.85%-54.41%) in Env, 56.93% (15.38%-89.16%) in Gag, and 46.12% (11.70%-70.57%) in Pol. Consensus sequences based on sequences sampled in each decade remained relatively stable over time. Similarly, at antibody epitope sites, only 0-8 residues that were minority variants became consensus over time in any subtype/CRF and only one known drug resistance mutation site differed from the reference (subtype G). The apparent contradiction between the fast diversification of HIV-1 and its limited adaptation illustrates that HIV-1 evolution is not directional and its consensus is at the intersection of millions of within-host selective processes occurring in a star-like manner. While a consensus sequence is a better representation of HIV-1 diversity than any individual sequence, consensus sequences have progressively become more distant from the circulating sequences they represent.IMPORTANCEGlobal surveillance of HIV-1 sequences is critical for designing relevant prophylactic and therapeutic interventions to infection. We designed an open-source platform, Hervé, for analyzing and visualizing the diversification dynamics of HIV-1 protein sequences. We characterized the evolution of over 12,000 HIV-1 Env, Gag, and Pol protein sequences from 1980-2020 and found that, despite a steady increase in intra-subtype and circulating recombinant form diversity, the most frequent residue at each site, i.e., the consensus, has varied only moderately.