Different infectivity of HIV-1 strains is linked to number of envelope trimers required for entry

Abstract

HIV-1 enters target cells by virtue of envelope glycoprotein trimers that are incorporated at low density in the viral membrane. How many trimers are required to interact with target cell receptors to mediate virus entry, the HIV entry stoichiometry, still awaits clarification. Here, we provide estimates of the HIV entry stoichiometry utilizing a combined approach of experimental analyses and mathematical modeling. We demonstrate that divergent HIV strains differ in their stoichiometry of entry and require between 1 to 7 trimers, with most strains depending on 2 to 3 trimers to complete infection. Envelope modifications that perturb trimer structure lead to an increase in the entry stoichiometry, as did naturally occurring antibody or entry inhibitor escape mutations. Highlighting the physiological relevance of our findings, a high entry stoichiometry correlated with low virus infectivity and slow virus entry kinetics. The entry stoichiometry therefore directly influences HIV transmission, as trimer number requirements will dictate the infectivity of virus populations and efficacy of neutralizing antibodies. Thereby our results render consideration of stoichiometric concepts relevant for developing antibody-based vaccines and therapeutics against HIV.

Figure 1. HIV-1 strains differ in the number of trimers required for entry.

(A) Theoretical predictions of relative virus infectivity over the fraction of dominant-negative mutant env (f_m) according to our model. Curves for T ranging from 1 to 8 are shown assuming the trimer number distribution across virions to follow a discretized Beta distribution with constant mean 12.95 and variance 45 . (B) Relative infectivity of mixed trimer infection experiments with 11 HIV-1 strains using the R508S/R511S (left) and V513E (right) dominant-negative env mutants. Infectivity of pseudotyped virus stocks expressing the indicated ratios of wild type and dominant-negative mutant envs was measured on TZM-bl reporter cells. Infectivity of virus stocks containing solely wt envelope were set as 100%. Data depict mean and SD from 2 to 4 independent experiments. For each virus the individual curve fits resulting from the model to the obtained data were evaluated (S1 Fig.). (C) Mathematical estimates of T derived from the data shown in (B). The R508S/R511S (black circles) and V513E (open squares) mutations were analyzed individually. Bootstrap analyses demonstrating the robustness of the obtained estimates of T are shown in S1 Fig. (D and E) Mixed trimer virus stocks for strain JR-FL (D) and SF-162 (E) carrying the V513E mutation were assayed on healthy donor PBMC and compared to data obtained with TZM-bl target cells. Data depict mean and SD from 2 independent experiments.

Figure 2. The entry stoichiometry governs virus population infectivity.

(A) Scheme depicting the influence of the entry stoichiometry on virus population infectivity. Different Ts (exemplified here: T = 1 and T = 7) will determine the minimum number of trimers that a virion requires in order to be infectious. (B) Correlation analysis (Pearson) of virus strain infectivity (measured by infection of TZM-bl reporter cells and expressed in arbitrary relative light units (RLU) per µl of virus stock) and the estimated T (plotted as mean of the independent R508S/R511S and V513E estimates shown in Fig. 1C). Virus infectivities are depicted as mean values derived from 3 independent experiments. (C) Mathematical modeling to investigate the influence of entry stoichiometry on virion population infectivity. The data depict how T = 2 and T = 7 translate into different fractions of a virion population being potentially infectious, in dependence on the trimer number distribution across the virion population. As shown in (D), the overall infectivity of a virus population decreases with increasing T. For (C) and (D) we assumed the trimer number distribution across virions to follow a discretized Beta distribution with constant mean 12.95 and variance 45 .

Figure 3. V1V2 deletion impairs virus infectivity and is reflected by a high stoichiometry of entry.

(A) Comparison of infectivity of pseudoviruses expressing wt and V1V2-deleted envs upon infection of TZM-bl cells. Data points depict mean values of luciferase reporter activity per µl virus stock measured in 3 independent experiments. The p-value was calculated by a paired t-test. (B and C) Relative infectivity of mixed trimer infection experiments of 10 wt envs and their V1V2 deleted variants using the R508S/R511S (B) and the V513E (C) dominant-negative mutations are shown. Infectivity of pseudotyped virus stocks expressing the indicated ratios of dominant-negative mutant envs was measured on TZM-bl cells. Infectivity of virus stocks containing solely functional envelope were set as 100%. Data depict mean and SD from 2 to 4 independent experiments. (D) Estimates of T for the wt and V1V2-deleted envs derived from mixed trimer experiments. Data points are the mean of the individual estimates of T obtained with the R508S/R511S and V513E dominant-negative mutations. The p-value was calculated by a paired t-test.

Figure 4. Virus entry kinetics correlate with the stoichiometry of entry.

(A) Entry kinetic curves for JR-FL wt and JR-FL ΔV1V2. Synchronized pseudovirus infection of TZM-bl cells following spinoculation was terminated by addition of T-20 at the indicated timepoints. Infectivity reached after 120 minutes was set as 100% and all data were normalized relative to this value. Data are mean and SD from 3 independent experiments. (B) Half maximal entry time for wt and V1V2-deleted envs was calculated from kinetic profiles shown in Fig. 4A and S7B Fig. Time (in minutes) required to reach 50% entry into target cells is depicted. Data shown are means derived from 2 to 4 independent experiments. The p-value was calculated by a paired t-test. (C) Correlation analysis (Pearson) of wt (black symbols) and V1V2-deleted env (red symbols) half-maximal entry time and estimated T.

Figure 5. Loss of the N160 glycosylation site in gp120 steers the stoichiometry of entry and virus infectivity.

(A) Titration of CAP88 wt (black) and CAP88 K160N (red) pseudoviruses on TZM-bl reporter cells. Data shown are mean and SD of luciferase reporter activity upon infection measured in 2 independent experiments. Inset: Infectivity comparison of CAP88 wt and K160N depicted as relative light units (RLU) of luciferase reporter activity per µl of pseudovirus stock upon infection of TZM-bl cells. The fold infectivity difference is indicated. (B) Estimates of T for the CAP88 wt and K160N variant. Mean and range of the individual estimates using the R508S/R511S and V513E dominant-negative mutations are shown.

Figure 6. Point mutations in JR-FL dictate virus infection efficacy and entry stoichiometry.

(A) Infectivities of JR-FL wt and indicated point mutant envs were determined by titration of virus stocks on TZM-bl reporter cells and are shown normalized to JR-FL wt. Data depict mean and SD from 4 independent experiments. (B) Relative infectivity of mixed trimer infection experiments with the specified JR-FL variants using the R508S/R511S and V513E dominant-negative mutations are shown. Infectivity of pseudotyped virus stocks expressing the indicated ratios of dominant-negative mutant envs was measured on TZM-bl cells. Infectivity of virus stocks containing solely functional envelope were set as 100%. Data depict mean and SD from 2 independent experiments. (C) Mathematical analyses of the data shown in (B) yielded estimates of T, shown here as mean and range of the individual T estimates obtained with the R508S/R511S and V513E dominant-negative mutations. (D) Analysis of virus entry kinetics for the four JR-FL variants were performed as shown S7A Fig.