Positive Selection at Key Residues in the HIV Envelope Distinguishes Broad and Strain-Specific Plasma Neutralizing Antibodies

The development of HIV broadly neutralizing antibodies (bNAbs) has previously been shown to be associated with viral evolution and high levels of genetic diversity in the HIV envelope (Env) glycoprotein. However, few studies have examined Env evolution in those who fail to develop neutralization breadth in order to assess whether bNAbs result from distinct evolutionary pathways. We compared Env evolution in eight HIV-1-infected participants who developed bNAbs to six donors with similar viral loads who did not develop bNAbs over three years of infection. We focused on Env V1V2 and C3V4, as these are major targets for both strain-specific neutralizing antibodies (nAbs) and bNAbs. Overall evolutionary rates (ranging from 9.92 x 10(-3) to 4.1 x 10(-2) substitutions/site/year) and viral diversity (from 1.1% to 6.5%) across Env, and within targeted epitopes, did not distinguish bNAb donors from non-bNAb donors. However, bNAb participants had more positively selected residues within epitopes than those without bNAbs, and several of these were common among bNAb donors. A comparison of the kinetics of strain-specific nAbs and bNAbs indicated that selection pressure at these residues increased with the onset of breadth. These data suggest that highly targeted viral evolution rather than overall envelope diversity is associated with neutralization breadth. The association of shared positively selected sites with the onset of breadth highlights the importance of diversity at specific positions in these epitopes for bNAb development, with implications for the development of sequential and cocktail immunization strategies. IMPORTANCE Millions of people are still being infected with HIV decades after the first recognition of the virus. Currently, no vaccine is able to elicit bNAbs that will prevent infection by global HIV strains. Several studies have implicated HIV Env diversity in the development of breadth. However, Env evolution in individuals who fail to develop breadth despite mounting potent strain-specific neutralizing responses has not been well defined. Using longitudinal neutralization, epitope mapping, and sequence data from 14 participants, we found that overall measures of viral diversity were similar in all donors. However, the number of positively selected sites within Env epitopes was higher in bNAb participants than in strain-specific donors. We further identified common sites that were positively selected as bNAbs developed. These data indicate that while viral diversity is required for breadth, this should be highly targeted to specific residues to shape the elicitation of bNAbs by vaccination.