Background:  Entry of HIV-1 requires binding of the gp120 Env protein to CD4, and to a co-receptor (CCR5 or CXCR4). The V3 region of Env is the major determinant of co-receptor binding. Subtype B viruses frequently switch late in infection from using CCR5 to using CXCR4. However, in subtype C virus, the use of CXCR4 is less common.

Method:  We aligned ~350 V3 sequences each from subtype B and C HIV-1 and calculated Shannon entropy at each position to determine the difference in the number and frequency of amino acid substitutions in the V3 region between subtype B and C. Full-length Env and V3 peptides were used in ELISA with a panel of anti-V3 monoclonal antibodies to determine binding differences between subtype B and C Env V3. Chimeric V3 peptides and Envs were generated where the V3 amino acids that differ between subtype B and subtype C were individually mutated. These Env were also pseudotyped and used to infect U373 CCR5-expressing MAGI cells to determine functionality of the mutant Env proteins.

Results: The V3 region of subtype B and C Env displayed different patterns of entropy at sites flanking the V3 turn. Both had a “saw-toothed” pattern of high/low entropy, but these positions were displaced between the 2 subtypes. This suggests that V3 has a β hairpin structure flanking the turn where the residues on one face of the β sheet have higher entropy, and that this face differs between subtype B and subtype C. ELISA data revealed that in chimeric Env proteins, both the backbone of Env and the V3 sequence itself are important in defining V3 as an epitope. Furthermore, using full-length Env and V3 peptides where 3 amino acid differences between consensus subtype B and C V3 sequences were individually mutated to the other subtype, we showed that the anti-V3 antibodies consistently fell into 3 distinct classes of binding to all Env and V3 peptides. Finally, when HIV-1 was pseudotyped with these chimeric Env, virion infectivity was reduced, suggesting that other differences further distinguish the Env conformation of the 2 subtypes.

Conclusions:  These results suggest that the 3 classes of binding by anti-V3 antibodies to full-length Env and V3 peptides reflect 3 distinct structural elements of the V3 region. The data also imply that the structure of V3 differs between subtype B V3 and subtype C V3 in ways that could affect V3 immunogen design.