Variable Loop Glycan Dependency of the Broad and Potent HIV-1-Neutralizing Antibodies PG9 and PG16

The HIV-1-specific antibodies PG9 and PG16 show marked cross-isolate neutralization breadth and potency. Antibody neutralization has been shown to be dependent on the presence of N-linked glycosylation at position 160 in gp120. We show here that (i) the loss of several key glycosylation sites in the V1, V2, and V3 loops; (ii) the generation of pseudoviruses in the presence of various glycosidase inhibitors; and (iii) the growth of pseudoviruses in a mutant cell line (GnT1^−/−) that alters envelope glycosylation patterns all have significant effects on the sensitivity of virus to neutralization by PG9 and PG16. However, the interaction of antibody is not inhibited by sugar monosaccharides corresponding to those found in glycans on the HIV surface. We show that some of the glycosylation effects described are isolate dependent and others are universal and can be used as diagnostic for the presence of PG9 and PG16-like antibodies in the sera of HIV-1-infected patients. The results suggest that PG9 and PG16 recognize a conformational epitope that is dependent on glycosylation at specific variable loop N-linked sites. This information may be valuable for the design of immunogens to elicit PG9 and PG16-like antibodies, as well as constructs for cocrystallization studies.It is argued that an effective HIV vaccine should include a component that induces a broadly neutralizing antibody response (2, 3, 21, 25, 32, 37, 39, 54). The key target for broadly neutralizing HIV antibodies is the envelope spike, which consists of a compact, metastable heterodimeric trimer of the glycoproteins gp120 and gp41 (43, 62).gp120 is one of the most heavily glycosylated proteins known, with up to 50% of its mass arising from carbohydrates attached to roughly 25 N-linked glycosylation sites (31) determined by the NXT/S consensus sequence (where X can be any amino acid except Pro) (1). Glycosylation significantly impacts the folding and conformation of envelope spikes, thus affecting antigenicity and immunogenicity (30, 35). Carbohydrates are generally poorly immunogenic, and the dense covering of glycans is often referred to as the “silent face” or “glycan shield” (58). The glycans have also been suggested to have an important role in viral transmission through interaction with lectins, in particular the C-type lectin DC-SIGN, which is found on the surfaces of dendritic cells and is thought to aid the transport of virus to anatomical sites rich in CD4⁺ T cells, such as lymph nodes (8, 16).Although the positioning of N-linked protein glycosylation is encoded by the protein sequence (1), the type of glycan displayed (high mannose, hybrid, or complex) is not under direct genetic control but is determined by the three-dimensional structure of a protein and its interaction with the biosynthetic cellular environment, including accessibility to glycan-processing enzymes (50). For example, highly clustered glycans prevent access of the processing enzymes, leading to high-mannose-type glycans being displayed (6, 23). Therefore, the glycosylation of recombinant HIV envelope proteins can vary significantly depending on the protein sequence, structure, and the cell in which they are expressed (50). Although the positions of many glycans are relatively conserved between isolates and clades (60), there can be variation in the occupancy and precise nature of the glycans displayed at these positions on recombinant envelope (7, 17-19, 61). However, we have recently observed major differences between the glycosylation of recombinant envelope proteins and envelope expressed on the virion surface, with the latter being dominated by Man_5-9GlcNAc₂ oligomannose glycans (9). Nevertheless, significant glycan heterogeneity remains on the virion surface.Recently, two new neutralizing antibodies, PG9 and PG16, were isolated from an African clade A-infected donor and shown to be both broad and potent (56). From a panel of 162 viruses, PG9 neutralized 127 and PG16 neutralized 119 viruses at a median potency that exceeded that of the broadly neutralizing antibodies—2G12, b12, 2F5, and 4E10—by about an order of magnitude. In a TZM-bl neutralization assay, PG9 has been shown to neutralize 87% of a panel of 82 viruses (M. Seaman, unpublished data). Both PG9 and PG16 show preferential trimer binding and interact with an epitope formed from conserved regions of the V1/V2 and V3 variable loops. Mutation of N160, an N-linked glycosylation site in the V2 loop, completely abolishes PG9 and PG16 neutralization, suggesting the N160 glycan is important in forming the PG9 and PG16 epitope. Further, PG9 shows significant binding to monomeric gp120 DU422 and treatment of the glycoprotein with Endo H (removing high-mannose glycans) results in significant reduction in antibody binding. Occasionally, neutralization of some pseudoviruses by PG16 in particular has revealed an unusual neutralization profile with a shallow slope and plateaus at <100%. We hypothesized that this unusual neutralization profile may be related to antibody sensitivity to glycosylation and, more specifically, could be due to glycan profile or partial glycosylation at critical sites.We show here that loss of any one of several glycosylation sites in the V1, V2, and V3 loops has significant effects on the sensitivity of pseudovirus to neutralization by PG9 and PG16. Generating pseudovirus in the presence of various glycosidase inhibitors also has notable effects on antibody neutralization. We show that some of these effects are isolate dependent and others are universal and can be used to help identify the presence of PG9 and PG16-like antibodies in the serum of HIV-1-infected patients (57). For some isolates displaying aberrant neutralization profiles as described above, we found that changing the glycan profile on the HIV-1 trimer using glycosidase inhibitors or a mutant cell line resulted in higher neutralization plateaus and neutralization profiles with the more usual sigmoidal shape. Changes in sensitivity to neutralization were also observed for some but not all isolates. The antibody-gp120 interaction was not inhibited by sugar monosaccharides found in glycans on the HIV envelope. The results suggest PG9 and PG16 recognize a conformational epitope that is dependent on the glycosylation at specific variable loop N-linked glycosylation sites. This information may be valuable for the design of immunogens to elicit PG9 and PG16-like antibodies, as well as constructs for cocrystallization studies.