Enhanced fusion and virion incorporation for HIV-1 subtype C envelope glycoproteins with compact V1/V2 domains

Abstract

In infected people, the HIV-1 envelope glycoprotein (Env) constantly evolves to escape the immune response while retaining the essential elements needed to mediate viral entry into target cells. The extensive genetic variation of Env is particularly striking in the V1/V2 hypervariable domains. In this study, we investigated the trade-off, in terms of fusion efficiency, for encoding V1/V2 domains of different lengths. We found that natural variations in V1/V2 length exert a profound impact on HIV-1 entry. Variants encoding compact V1/V2 domains mediated fusion with higher efficiencies than related Envs encoding longer V1/V2 domains. By exchanging the V1/V2 domains between Envs of the same infected person or between two persons linked by a transmission event, we further demonstrated that V1/V2 domains critically influence both Env incorporation into viral particles and fusion to primary CD4 T cells and monocyte-derived dendritic cells. Shortening the V1/V2 domains consistently increased Env incorporation and fusion, whereas lengthening the V1/V2 domains decreased Env incorporation and fusion. Given that in a new host transmitted founder viruses are distinguished by compact Envs with fewer glycosylation sites, our study points to fusion and possibly Env incorporation into virions as limiting steps for transmission of HIV-1 to a new host and suggests that the length and/or the N-glycosylation profile of the V1/V2 domain influences these early steps in the HIV life cycle.

FIG 1

Alignment of V1/V2 domains of WT Envs with natural variations in V1/V2. Sequences from the 37 WT Envs from the four transmission pairs of the Lusaka cohort were aligned using multiple-sequence alignments with hierarchical clustering (by use of the MultiAlin program [61]). The V1/V2 loop was identified by comparison with the sequence of HxB2. Each Env genotype is named with the number corresponding to the transmission pair, a letter corresponding to the person in the couple (F for the acutely infected female and M for the male transmitting partner), and a number corresponding to the Env clone. Note that although the Env genes from the female recipient encode identical V1 or V2 domains, they each encode a slightly different variant due to differences in other parts of the Env. The yellow areas highlight the part of the sequences with variation in the amino acid within the transmission pair.

FIG 2

Analysis of fusion with the virion-based fusion assay. (A) Gating strategy used to measure HIV-1 fusion to MDDCs or resting or activated PBLs. For measuring fusion to CD4 T cells, PBLs were infected, allowed to fuse to CD4 T cells, and immunostained. Resting T cells were gated for CD3⁺ CD4⁺, and activated T cells were gated for CD3⁺ CD4⁺ CD45RO⁺ β7⁺. The fusion gates were always set on the uninfected samples, where usually less than 0.01% of the cells showed a shift in fluorescence. Note the increase in the number of cells in the fusion gates for 81A. FSC, forward scatter; SSC, side scatter. (B) Representative fusion FACS plot obtained with WT Envs from transmission pair 109. Viral supernatants from 293T cells transfected with WT clones of pair 109 were normalized for p24^Gag levels and used to infect MDDCs and resting or activated CD4 T cells.

FIG 3

Reproducibility of fusion measurements. (A) Triplicate infections were performed with the same viral preparation with MDDCs or CD4 T cells. Error bars represent standard deviations. (B) Donor-to-donor target cell variability. Results of experiments in which the same viral preparations were used to infect the cells of two donors are presented. To account for changes in HIV-1 susceptibility, the fusion ratio corresponding to the percentage of cells supporting infection was divided by the fusion measurement for 81A. (C) Variability due to different plasmid preparation, virion production, quantification, and infection. For four genotypes (Env 1 to Env 4), two or three clones (represented by different symbols) were tested simultaneously using the same donor target cells. Note the similarity of the fusion phenotypes for identical genotypes.

FIG 4

Correlation between fusion and V1/V2 length of WT Envs. Fusion mediated by the 37 WT Envs corresponding to four transmission pairs was tested by the fusion assay with resting or activated CD4 T cells and MDDCs. The ratio between WT Env fusion and 81A fusion was first calculated. Then, results for replicate infections of the same target cells with the same viral preparation and replicate donors were successively averaged, as described in Materials and Methods. The median values of fusion relative to the value for 81A were used for statistical analysis. The correlation between change in length and fusion was analyzed using Spearman's rank correlation. CI, confidence interval. Best-fit lines were added to highlight the negative correlations.

FIG 5

Alignment of WT and mutant Envs. Mutations affecting the V1 or V2 domain were designed by swapping amino acids between Envs of a transmission pair and/or the same infected individual. (A to D) WT and mutant Envs of each of the four transmission pairs of the Lusaka cohort, respectively. The V1/V2 mutants were named with the Env WT backbone, followed in parentheses by the clone name from which V1 or V2 was copied. Mutations are highlighted in bold.

FIG 6

Correlation between changes in fusion and in V1/V2 length. The ratio between fusion of the mutant and that of the corresponding WT clone was calculated for the 19 V1/V2 mutants. Results for replicates of clones, replicate infections of the same target cells with the same viral preparation, and replicate donors were successively averaged, as described in Materials and Methods. The median values of Mut/WT were used for statistical analysis. The correlation between change in length and fusion was analyzed using Spearman's rank correlation.

FIG 7

Measurement of Env incorporation. Viral preparations of WT and V1/V2 mutants were lysed, and 100 ng of p24^Gag from the viral lysate was analyzed by Western blotting. The b13 antibody was used to measure the incorporation of Env into virions, while the p24^Gag antibody was used to control for the virion input. (A) Viral incorporation for the set of WT and V1 mutant Envs corresponding to pair 109. The ratio between the intensities of the Env and p24^Gag bands was first calculated for all the genotypes. Then, the fold change in Env incorporation between the mutant and its corresponding WT was calculated by dividing the ratio of Env/p24 obtained for the mutant by the one obtained for the corresponding WT. These fold changes are given for each mutant of transmission pair 109. (B) Relationship between changes in Env incorporation and V1/V2 length. The Mut/WT fold change in Env incorporation was then graphed against the changes in Env length measured by subtracting the number of amino acids of the WT sequence from the number of amino acids of the mutant (WT − Mut).

FIG 8

Relationship between changes in V1/V2 loop length, Env incorporation, and fusion to MDDCs and CD4 T cells. (A) Correlation between V1/V2 length and envelope incorporation. The ratio between the intensity of the Env and p24^Gag signal was measured for the entire set of WT and mutant genotypes. This quantification was performed twice using two sets of viral production. The change in envelope incorporation (Mut/WT) was calculated for each set, averaged, and plotted versus the changes in length (WT − Mut). (B) Link between envelope incorporation (Mut/WT) and fusion to resting or activated CD4 T cells and MDDCs (Mut/WT). (C) Effect of Env titration on fusion to MDDCs and CD4 T cells. Proviral DNA of the 109F4 WT envelope was cotransfected into 293T cells with NL4-3 ΔEnv at the indicated ratio and with BlaM-Vpr. After normalization by the use of the p24^Gag signal, MDDCs and CD4 T cells were infected with the equivalent of 500 ng of p24^Gag for 2 h. The level of fusion was reported relative to the level of fusion obtained with 100% WT 109F4. Envelope incorporation was measured by Western blotting with the b13 antibody, and p24^Gag was used to control for the amount of input virions.

FIG 9

Link between changes in the number of potential glycosylation sites in the V1/V2 domain, V1/V2 length, Env incorporation, and fusion to MDDCs and CD4 T cells. Correlations between changes in the number of potential glycosylation sites identified by NX(T/S) and changes in V1/V2 length (A), Env incorporation (B), and fusion to MDDCs or resting or activated CD4 T cells (C) were determined. (D) Summary of the correlation between the changes incurred by the mutations for the four variables. Note that the four parameters were significantly correlated (P < 0.05, Spearman's test) for all comparisons except the relationship between Env incorporation and MDDC fusion (the changes in Env incorporation levels did not correlate with fusion).