Structural Rearrangements Maintain the Glycan Shield of an HIV-1 Envelope Trimer After the Loss of a Glycan

Abstract

The HIV-1 envelope (Env) glycoprotein is the primary target of the humoral immune response and a critical vaccine candidate. However, Env is densely glycosylated and thereby substantially protected from neutralisation. Importantly, glycan N301 shields V3 loop and CD4 binding site epitopes from neutralising antibodies. Here, we use molecular dynamics techniques to evaluate the structural rearrangements that maintain the protective qualities of the glycan shield after the loss of glycan N301. We examined a naturally occurring subtype C isolate and its N301A mutant; the mutant not only remained protected against neutralising antibodies targeting underlying epitopes, but also exhibited an increased resistance to the VRC01 class of broadly neutralising antibodies. Analysis of this mutant revealed several glycans that were responsible, independently or through synergy, for the neutralisation resistance of the mutant. These data provide detailed insight into the glycan shield's ability to compensate for the loss of a glycan, as well as the cascade of glycan movements on a protomer, starting at the point mutation, that affects the integrity of an antibody epitope located at the edge of the diminishing effect. These results present key, previously overlooked, considerations for HIV-1 Env glycan research and related vaccine studies.

Figure 1

Distribution of potential N-linked glycosylation sites (PNGSs) for the CAP45.G3 strain (blue) relative to the PNGSs of the HIV-1 reference strain, HXB2 (orange). The gp160 conserved (C1-C4), variable (V1-V5) and gp41 sequence regions are labelled and shaded to indicate the borders of each region.

Figure 2

3D representation of the N-linked glycosylation sites of the (a) CAP45.G3 strain (computationally determined) and the clade A BG505 strain (crystal structure). The protein and glycan residues are shown as surfaces and the glycans are labelled according to HXB2 numbering. The depicted orientation is such that the lipid membrane is located at the top and the V1/V2-loop regions are at the bottom of the figure.

Figure 3

N301A mutant model with different orientations (a–c, 120° increments) to show each protomer (different shades of tan) and the residues with statistically significant increases in their average AASA ratios relative to the wild-type simulation (difference in average AASA <10% is orange (not important), ≥10% is blue (important)). A statistically significant increase was evaluated at a 5% significance level using a bootstrap approach (see Methods). The V3 (pink) and CD4 (green) regions are outlined. The depicted orientation is such that the lipid membrane is located at the top and the V1/V2-loop regions are at the bottom of the figure.

Figure 4

Neighbourhoods of the (a) wild-type and (b) N301A mutant glycans nearest to the CD4 sub-region for the three protomers (A, B, and C): glycans N197 (orange), N301 (blue, only present in the wild-type), and N386 (pink). Horizontal lines represent residues that are in different neighbourhoods when comparing the wild-type and N301A mutant. The arrows originate from the Cα atom of the Asn and end at the average centre of mass position for the selected glycans. The representations are cropped around the CD4 binding site of each protomer and the surface representation of residues on adjacent protomers are shown as opaque.

Figure 5

Neighbourhoods of the (a) wild-type and (b) N301A mutant glycans nearest to the V3 sub-region for the three protomers (A, B, C): glycans N156 (orange), N262 (yellow), N301 (blue, only present in the wild-type), N442 (pink), and N446 (green). Horizontal lines represent residues that are in different neighbourhoods when comparing the wild-type and N301A mutant. The arrows originate from the Cα atom of the Asn and end at the average centre of mass position for the selected glycans. The representations are cropped around the V3 region of each protomer and the surface representation of residues on adjacent protomers are shown as opaque.

Figure 6

Domino effect of glycan conformational changes on the three protomers. (a) Structures illustrating the location of glycans. The arrows originate from the Cα atom of the Asn and end at the average centre of mass position for the selected wild-type (orange) and N301A mutant (blue) glycans identified by the hydrogen bond analysis for protomers A, B and C. Curved arrows indicate the degree of change in directionality, e.g. the large movement for glycan N399 on protomer C. The representations are cropped around the V3 region of each protomer and the surface representation of residues on adjacent protomers are shown as opaque. (b) Each schematic - protomer A, B and C - shows the hydrogen bond network starting from glycan N442 (first node) on protomer A, B and C, respectively. Solid borders around glycan nodes and broad arrows represent glycans and interactions that form part of the domino effect considered, whereas dotted borders and fine arrows indicate any glycans and interactions outside the considered domino effect. The arrows connecting glycan nodes distinguish interactions that are either greater on the wild-type (WT; orange) or on the N301A mutant (M; blue).