Figure 3

Selection Map of HIV-1 gp120. Positively selected variants are at those codons where one or more mutations from the consensus confer selective advantage on the virus (capitalized and listed below the consensus). Lowercase letters indicate mutations where the neutral drift hypothesis is not ruled out. Negatively selected variants are not shown. Numbering is according to the HXBc2 sequence, where residue one is the amino-terminal methionine of the signal sequence. The asterisk indicates a 310Q-311R insert in HXBc2 that is not common enough to include in the profile (the alignment gap is the consensus sequence, and 2^H+σ < 1.5). The hyphens in the numbering indicate codons absent from HXBc2. Many of the positively selected gp120 variants lie outside the variable regions, V1-V5, and are instead found in the constant regions proposed by Starcich et al. [61]. This is not troubling in that the V1-V5 designations were based on an analysis of only five gp120 sequences, so the "variable" and "constant" designations are not truly indicative of variability (B. Foley, Los Alamos, pers. commun.; this paper). For more appropriate positional references, we indicate secondary structure motifs [44]. The four numbered regions are repetitive and gap-filled and for which we could not generate a reasonable alignment; because we are unable to align these subregions reliably and because the alignment is a prerequisite of our analysis, we cannot reasonably comment further on these regions. The CD4BS epitopes are yellow; CD4i epitopes are blue; positions that belong to both CD4BS and CD4i are green [45].