Figure 6

Compartmentalization “Hot spots” across full-length HIV-1 env and the network of correlated mutations. Results of analyses identifying compartmentalization hot spots (A). All positions listed are significant in ≥ 2 individuals. Subjects with statistically non-compartmentalized (blue) and compartmentalized (yellow) CSF quasispecies are shown. For each hot spot, corresponding HXB2 gp160 (white) and specific gene-region (C1-gp41) numbering (red) is shown (regions not drawn to scale). Amino acids at hot spot positions in CSF (C) alignments meeting the threshold for statistical significance are shaded in purple. Corresponding plasma (P) amino acids are shown directly below. (B) An adjacency matrix demonstrating hot spot positions with amino acids sharing statistically significant mutual information (MI) in ≥ 2 subjects. Lines connect positions sharing MI, with the thickness of the line corresponding to the number of studied subjects in which the correlation was identified. Surface representation of the hot spots on a gp120 trimer based on the SOSIP crystal structure in a side (C) and top (D) view. The three gp120s in the trimer were rendered as surfaces and colored light grey, grey and blue, respectively. For simplicity, the hot spot residues were colored individually and only displayed in one of the gp120s (light grey). A CD4 molecule (with N-terminal D1D2 domains; cyan ribbon) was placed onto this gp120 by superimposition of its complex with a gp120 core to indicate the location of the CD4 binding site (light red). The entire V1V2 region and V3 region are colored light green and light orange, respectively, while the hot spots are colored more intensely. Note that the hot spot residues form several spatial clusters on the gp120 surface, including the V2 tip region (dark green), V3 region (dark orange), the CD4 binding site proximal region (red), the outer domain cluster (blue) and the inner domain cluster (pink).