HIV-1 escapes from N332-directed antibody neutralization in an elite neutralizer by envelope glycoprotein elongation and introduction of unusual disulfide bonds

Abstract

Background: Current HIV-1 immunogens are unable to induce antibodies that can neutralize a broad range of HIV-1 (broadly neutralizing antibodies; bNAbs). However, such antibodies are elicited in 10-30 % of HIV-1 infected individuals, and the co-evolution of the virus and the humoral immune responses in these individuals has attracted attention, because they can provide clues for vaccine design.

Results: Here we characterized the NAb responses and envelope glycoprotein evolution in an HIV-1 infected "elite neutralizer" of the Amsterdam Cohort Studies on HIV-1 infection and AIDS who developed an unusually potent bNAb response rapidly after infection. The NAb response was dependent on the N332-glycan and viral resistance against the N332-glycan dependent bNAb PGT135 developed over time but viral escape did not occur at or near this glycan. In contrast, the virus likely escaped by increasing V1 length, with up to 21 amino acids, accompanied by the introduction of 1-3 additional glycans, as well as 2-4 additional cysteine residues within V1.

Conclusions: In the individual studied here, HIV-1 escaped from N332-glycan directed NAb responses without changing the epitope itself, but by elongating a variable loop that shields this epitope.

Keywords: Broadly neutralizing antibodies; Cysteines; Envelope glycoprotein; Glycans; HIV-1; N332; Variable regions.

Fig. 1

Development of heterologous and autologous NAb responses in elite neutralizer D16916. a Neutralizing activity over the course of HIV-1 infection against the individual viruses of a 6-virus panel that is representative for HIV-1 variation worldwide [8, 32]. The ID₅₀ against each virus as well as the geometric mean midpoint titer for all 6 viruses combined are given for D16916 sera taken 8, 11, 14, 19, 22, 26, 30, 34 and 38 months post-SC. A color scale is used to indicate low (green) to high (red) ID₅₀ values. b Longitudinal neutralization of autologous viruses. The mean ID₅₀ values of longitudinal sera (x-axis) against the autologous viruses (3 for month 7 and 6 for month 11, all in grey) are plotted in green (month 7) and blue (month 11). The red line indicates the geometric mean midpoint titer against the heterologous 6-virus panel from a [14]

Fig. 2

Sensitivity of viral isolates from individual D16916 to bNAbs. Clonal HIV-1 variants from 7 and 11 months post-SC were tested for their neutralization sensitivity against bNAbs PGT135 (a), b12 (b), 12A21 (c), PG16 (d) and PGT145 (e) and grouped according to their target epitope OD-glycan, CD4-bs and V1V2 apex. The graphs show IC₅₀ values for each virus isolate, as determined by linear regression. Differences were considered statistically significant when p values were ≤ 0.05, represented by asterisks (*p ≤ 0.05; **p ≤ 0.005, ***p ≤ 0.001). The horizontal bars represent the median IC₅₀ value per time point

Fig. 3

Extension of V1 and insertion of unusual disulfide bonds. a Amino acid alignment of the D16916 V1 loop of clonal HIV-1 variants isolated over time. Cysteine residues are indicated in dark grey boxes and potential N-linked glycans in light grey. The frequencies of additional cysteines (0: blue, 2: red, 4: green) in the sequences from viruses isolated at 7, 11 and 19, 22 and 30 months post-SC combined are shown in the pie charts. HXB2 numbering has been used to annotate the positions of the conserved cysteine residues. Arrows indicate the positions of the introduced cysteine residues. b The length of V1 and c the number of PNGS in V1 of clonal HIV-1 variants over the course of infection are shown. In both panels, the horizontal bars represent the mean values per time point tested and differences were considered statistically significant when p values were ≤ 0.05, represented by asterisks (***p ≤ 0.001)

Fig. 4

Involvement of V1 length in resistance to bNAbs PGT135, b12 and 12A21. Correlation plots between the V1 length and neutralization sensitivity for mAbs PGT135 (a), PGT121 (b), b12 (c), 12A21 (d), and VRC01 (e). The r and p values for the linear regression are given. For correlations that were statistically significant, the regression line is shown. (f) Top view of the Env trimer with one protomer shown in dark blue and the other protomers in light blue. The V1 loops are indicated in red and the N332-glycan is shown on one protomer. g View of the trimer in the same orientations as f in complex with PGT135 (yellow) and PGT122 (gray). We used PGT122 in our figure instead of PGT121 since the HIV-1 Env trimer structures were solved in complex with PGT122, whereas for PGT121 the structure was only solved in complex with a glycan [42] or recently a complex of a PGT121 precursor in complex with the HIV-1 Env trimer [120]. It was shown that PGT121 and PGT122 bind to the Env protein with the same angle of approach in a very similar way. h Detailed view of the expected clash of PGT135 with the V1 loop indicated by an asterisk. i View of the trimer in the same orientations as f in complex with b12 (green) and 12A21 (dark green) and VRC01 (gray). j Detailed view of the expected clash of b12 with the V1 loop indicated by an asterisk. The figures were drawn using pymol (www.pymol.org) by aligning the gp120 structures of 4JM2.pdb (gp120 plus PGT135; PMC3823233 [60]), 2NY7.pdb (gp120 plus b12: PMC2584968 [121]), 4JPW.pdb (gp120 plus 12A21; PMC3792590 [122]), and 3NGB.pdb (gp120 plus VRC01; PMC2981354 [123]) with the dark blue protomer of the BG505 SOSIP.664 trimer in complex with PGT122 and 35O22 (4TVP.pdb [43])

Fig. 5

Accommodation of long V1, associated with viral fitness loss, by compensatory mutations. a Viral replication of clonal viral isolates from month 7 versus month 11 with the horizontal bars representing the median p24 values over the viruses isolated at that time point. b Linear regression between viral replication, expressed as p24 production per day, and V1 length for viruses from month 7 and 11. Virus isolates are colored based on number of additional cysteine residues in their V1 (0: blue, 2: red, 4: green). c V1 sequence alignment of representative virus isolates from individual D16916 with 0, 2 and 4 additional cysteines (clones D16916.7.2C9, D16916.7.1E3, D16916.11.2D1), as well as HIV-1_LAI, HIV-1_LAI mutants 1 and 2, and their revertants. d CA-p24 ELISA of HIV-1_LAI and mutant virus stocks, produced by transient transfection of HEK293T cells. e TZM-bl cells were infected with 500 pg CA-p24, and infectivity was measured after 48 h infection. f Schematic V1/V2 topology. β-strands are depicted as purple arrows and disulphide bonds as yellow lines. The V1 and V2 loop are indicated in green and blue lines, respectively. The substitutions designed to restore the epitopes for bNAbs PG9 and PG16 are underlined and the locations of the HIV-1_LAI reversions are indicated in bold. g Ribbon diagram of f with the position of the HIV-1_LAI reversions indicated as red spheres (except L → P at the fourth position of the insert) and labeled according to their position in the linear sequence. Note that the elongated V1 with additional cysteine residues is not depicted