smFRET imaging of full-length Env on HIV-1 virions from the gp120–gp41 structural perspective.

(A) Domain organization of BG505 Env with click fluorophores conjugated to ncAA sites at S401 on gp120 and R542 on gp41. “ncAA” refers to unnatural amino acids incorporated at amber (TAG) sites via amber suppression, and “click” dyes are site-specifically conjugated through click chemistry (see Fig. S1 for details). (B – D) Functional validation of HIV-1Q23 EnvBG505 S401ncAA R542ncAA virions for smFRET studies. Immunoblotting (B), tomographic slices (C), and neutralization curves (D) confirm that the dual-ncAA incorporation does not compromise Env trimer functionality on virions used for smFRET. Viruses were generated from transfections using 100% of the indicated HIV-1 constructs, and the data shown represent the resulting particles. (E) HIV-1Q23 virions carrying fluorescently labeled full-length EnvBG505, imaged by prism-based TIRF microscopy. The Env trimer (PDB 4ZMJ) was fitted into the electron density map of the membrane-bound Env trimer (EMD-21412). Two click fluorophores (donor LD555-TTZ and acceptor LD655-TTZ) were site-specifically introduced at S401ncAA and R542ncAA of a single protomer (gp120 in cyan, gp41 in pink; wild-type protomers in grey). Labeling sites are shown as green and red spheres, respectively. (F) Fluorescence emission spectra of HIV-1Q23 virions carrying click-labeled EnvBG505 S401ncAA R542ncAA, showing LD555 and LD655 signals from dual-color virions (solid cyan) compared with spectra from single-color–labeled virions (dashed lines). (G) Representative donor (green) and acceptor (red) fluorescence traces and corresponding FRET efficiency trajectories (blue) with hidden Markov modeling (HMM) idealization (magenta), from two individual ligand-free HIV-1Q23 virions carrying EnvBG505 S401* R542* (left and right panels). The black arrow marks single-step photobleaching. Three FRET-populated states are indicated by color-coded bands.

smFRET observation of global conformational opening of Env trimer from gp120–gp41 and gp120 V1V4 structural perspectives.

(A, B) Schematic models of Env conformational states. (A) gp120 adopts three major conformations: “PT” (pre-triggered, structurally unknown), PC (prefusion closed), and CO (CD4-bound open), shown for a single protomer within the trimer, with the other two protomers remaining in the PT state. (B) Fully open Env trimer schematic showing all three gp120 subunits in the CD4-bound state. (C, D) Structural visualization of Env conformations from two smFRET probe perspectives. Env trimers in the prefusion closed (C, PDB #4TVP) and fully open (D, PDB #5VN3) states are shown with probe positions highlighted: gp120–gp41 (S401* R542*) and gp120 V1V4 (N136* S401*). (E, F) FRET histograms (left panels) and population contour plots (right panels) of EnvBG505 S401* R542* (gp120–gp41 perspective) on native virions without ligand (E) or with soluble CD4 (sCD4) plus the co-receptor–mimicking antibody 17b (F). Ligand-free Env samples three primary conformational states (as in panel A), whereas sCD4 and 17b stabilize the fully open conformation (as in panel B). Here and elsewhere, Nm indicates the number of FRET trajectories - molecules that were used to construct the histogram and were fit to the sum of three Gaussian distributions (see Table S1). State occupancies are reported as percentages. Contour plots represent 12-second cumulative FRET trajectories for all molecules, showing the frequency of sampled conformations over time. (G, H) Validation from the gp120 V1V4 perspective using EnvBG505 N136* S401*. FRET histograms confirm a conformational shift toward the open-dominated population (see Fig. S6 for contour plots).

Stabilization of HIV-1 Env in the prefusion closed state by gp120–gp41 interface or FP bNAbs

(A) Dose–response neutralization curves (left) of HIV-1Q23 EnvBG505 by the gp120–gp41 interface or FP bNAbs 8ANC195, VRC34, and PGT151, with binding epitopes mapped onto the membrane-bound Env trimer (EMD-21412, right). (B) FRET histogram (left) and population contour plot (right) of HIV-1Q23 EnvBG505 S401* R542* in the presence of 8ANC195, overlaid with ligand-free Env (dashed gray), showing modest conformational shifts toward downstream states. (C, D) FRET histograms (left) and contour plots (right) for VRC34 (C) and PGT151 (D), as in B, revealing redistribution of Env populations from pre-triggered (PT) to prefusion closed (PC) state dominance. (E) Bar graph (left, mean ± s.e.m.) and line graph (right) showing relative state occupancy of Env under different antibody conditions, demonstrating progressive enrichment of the PC state by 8ANC195, VRC34, and PGT151.

MPER-directed bNAbs enrich Env in the prefusion-closed (PC) state, while 10E8.4/iMab also enriches Env in the CD4-bound open (CO) state.

(A) HIV-1 neutralization by MPER-directed antibodies DH511.2_K3 and VRC42, and 10E8.4/iMab. Left: neutralization curves; right: MPER binding epitopes on Env. (B, C) FRET histograms of EnvBG505 S401* R542* in the presence of DH511.2_K3 (B), or VRC42 (C). These MPER-directed bNAbs shift Env toward dominance of the PC state. The FRET histogram for Env with sCD4 and 17b (dashed gray) was included in each plot for reference. (D, E) smFRET results for bi-valent 10E8.4/iMab alone (D), which shifts Env further downstream along the opening pathway with a broad range across PC and CO, and with sCD4 and 17b (E), which further stabilizes Env in the open CO state. (F, G) Bar graph (F, mean ± s.e.m.) and line graph (G) showing relative state occupancy of Env under different antibody-binding conditions, demonstrating shifts from PT to PC with MPER bNAbs, and progression toward CO with 10E8.4/iMab alone or combined with sCD4 and 17b. (H) Bar graph of changes in state occupancy of Env exerted by MPER antibodies relative to ligand-free. (I) A derived schematic summarizing smFRET results of preferential conformations of Env targeted by MPER-directed bNAbs and 10E8.4/iMab.

Kinetic analysis of Env transitions exerted by gp41-directed antibody and 10E8.4/iMab, and an integrative model depicting allosteric control across conformational states.

(A, B) gp41-directed antibody and the bispecific 10E8.4/iMab alter Env transition frequencies with slight effects on kinetic rates while preserving the opening pathway. (A) Transition density plots (TDPs) showing the order, directionality, and frequency of Env conformational transitions in the presence of representative bNAbs. Each TDP plots the initial versus final FRET value for every detected transition, with color intensity indicating the relative frequency of those transitions (transitions per second). Nt = total number of transitions from Nm (total number of molecules), indicated on each plot. These plots reveal antibody-dependent redistribution of state sampling among PT, PC, and CO conformations. (B) Schematic (top) and table (bottom) summarizing transition rates between allowable Env conformations in the presence of interface/FP, MPER, and 10E8.4/iMab antibodies. (C) Integrative working model of gp41 antibody-mediated allosteric control of Env conformational states. Native Env on virions transitions through three primary states (PT, PC, and CO) along the opening pathway toward fusion, depicted as sequential checkpoints along the opening/fusion road targeted by antibodies. Most gp41-directed bNAbs act preferentially at the PC checkpoint, whereas 10E8.4/iMab can act at both PC and CO. The MPER- and CD4-binding arms of 10E8.4/iMab may function as a mechanical brace, restraining Env from breaking free at PC, CO, and the secondary checkpoints between them, which could explain its superior potency.

Model-fitting statistics and parameters for FRET histograms.

FRET histograms acquired from two structural perspectives (top: S401 – R542; bottom: N136 – S401) of Ab-incubated Env on native virions were fitted independently using constrained three-state models. Each state was described by a Gaussian distribution N (μ, σ2) corresponding to PT, PC, and CO states. The mean (μ) and standard deviation (σ) parameters were determined separately for each perspective based on visual inspection of trajectories exhibiting state-to-state transitions and idealization of individual trajectories using multi-state Hidden Markov modeling. State probabilities are reported as mean ± s.e.m. Goodness-of-fit was assessed using R2 and RMSE (root-mean-square error), with R2 values approaching 1 and RMSE values approaching 0 indicating a high-quality fit.

Amber suppression and click-labeling strategy for preparing HIV-1Q23 EnvBG505 virions for smFRET.

(A) Genetic code expansion via amber suppression. Amber suppressor tRNA and tRNA synthetase (tRNAPyl/NESPylRSAF) incorporate the ncAA trans-cyclooct-2-en-L-lysine (TCO*A) at TAG codons engineered into Env (S401TAG in gp120 and R542TAG in gp41) on intact virions produced in mammalian cells. The supply of TCO*A to transfected cells enables its incorporation at the designated positions, providing reactive handles for subsequent fluorophore conjugation. (B) Bioorthogonal click labeling via SPIEDAC. Tetrazine-conjugated Cy3 and Cy5 derivatives (LD555-TTZ and LD655-TTZ) react with the strained alkene of TCO*A through strain-promoted inverse electron-demand Diels–Alder cycloaddition (SPIEDAC). The conjugated fluorophores (dyes) are depicted as red spheres. TCO*A Functional groups are shown in the modeled membrane-present Env trimers (right panel).

Dual-amber suppression efficiency of HIV-1Q23 EnvBG505 S401TAG R542TAG and particle size distribution of the resulting virions.

(A) Infectivity of dual-ncAA (dual-amber) virions was measured on TZM-bl cells. Virions were produced by transfecting HEK293T cells with plasmids amber-free HIV-1Q23 EnvBG505 S401TAG R542TAG. Infectivity (mean ± SD, n=3) was normalized to wild-type EnvBG505 (WT) under suppression conditions. With amber suppression (tRNAPyl/NESPylRSAF + TCO*A), the dual-amber construct retained ∼20% of WT infectivity, whereas non-suppression controls (no supplement) showed negligible infectivity. Mock denotes particles produced without Env plasmids. (B) Example negative-staining transmission electron microscopy images of the resulting HIV-1Q23 EnvBG505 S401ncAA R542ncAA virions produced under amber suppression conditions. (C) Particle size distributions of HIV-1 particles (as in panel B) measured by nanoparticle tracking analysis (NTA, ZetaView). Dual-ncAA Env virions exhibited size distributions comparable to WT Env virions.

Biophysical fluorescence properties of functionalized fluorophores and single–color–labeled EnvBG505 HIV-1Q23 S401ncAA R542ncAA virions.

(A) Excitation and emission spectra of free LD555-tetrazine (LD555-TTZ) and LD655-tetrazine (LD655-TTZ). (B) Excitation and emission spectra of single-color–labeled S401ncAA R542ncAA virions, in which both ncAA sites (S401 on gp120 and R542 on gp41) were labeled with the same fluorophore via SPIEDAC click chemistry. LD555-only denotes virions labeled with LD555-TTZ at both sites, and LD655-only denotes virions labeled with LD655-TTZ at both sites. These single-color–labeled virions serve as spectral controls.

Relative protomer movements in symmetric PT, PC, and CO Env trimers.

Cartoon representations of Env trimers in three major conformations: “PT” (pre-triggered, structurally unknown), PC (prefusion closed), and CO (CD4-bound open). This figure complements Figure 2A by showing the three conformations as symmetric trimers, whereas Figure 2A depicts a single protomer transitioning within an otherwise PT-state trimer.

Representative traces of Env predominantly in the open conformation.

Example fluorescence and FRET traces from an individual EnvBG505 S401* R542* virion under opening conditions with soluble CD4 (sCD4) and the co-receptor–mimicking antibody 17b.

smFRET population contour plots of EnvBG505 N136* S401* on native virions.

(A) Ligand-free condition. (B) With soluble CD4 (sCD4) and the co-receptor–mimicking antibody 17b. Each plot represents 12 seconds of cumulative trajectories, showing the frequency of conformations sampled over time. These data provide the contour-plot representation corresponding to the FRET histograms shown in Figure 2G, H.

Molecular dynamics simulation of dye-conjugated soluble Env trimer in prefusion closed state and CD4-bound state.

(A) Sampling space of dye conformations during a 10-ns molecular dynamics simulation of dye-conjugated soluble Env S401TCO*-A – LD555 R542TCO*A – LD655 in prefusion closed state (RCSB PDB ID: 4TVP). Dye positions were extracted from the equilibrated portion of the trajectory and superimposed after backbone alignment. LD555 is shown in green, LD655 in red, and the protein backbone in grey. (B) Sampling space of dye conformations during a 10-ns molecular dynamics simulation of dye-conjugated Env in CD4-bound open state (RCSB PDB ID: 5VN3). Dye positions were extracted and displayed as described in (A). (C) Inter-dye distance as a function of simulation time (25 ns) for dye-conjugated Env trimer in prefusion closed state and CD4-bound open state, calculated from the center-of-mass distance between the two dyes at each saved frame. Distance traces were smoothed using a centred moving average window of 40 ps for visualization. (D) Distribution of inter-dye distances. Distances were calculated as the center-of-mass separation between the two dyes over the trajectories shown in (C). Mean inter-dye distances are 48.68 ± 1.75 Å (prefusion closed state) and 73.75 ± 2.94 Å (CD4-bound open state), respectively (mean ± SD).

Representative smFRET traces of HIV-1Q23 EnvBG505 S401* R542* with gp120-gp41 interface or fusion peptide bNAbs.

(A – C) Example donor and acceptor fluorescence (top) and corresponding FRET efficiency traces (bottom) from an individual EnvBG505 S401* R542* in the presence of 8ANC195 (A), VRC34 (B), and PGT151 (C).

smFRET population contour plots of Env with MPER-directed and bi-valent antibodies.

(A, B) Contour plots of HIV-1 EnvBG505 S401* R542* in the presence of MPER-directed antibody DH511.2_K3 or VRC42 (A), and bi-valent 10E8.4/iMab (B) under the indicated conditions. Each plot represents 12 seconds of cumulative trajectories from all molecules included in the FRET histograms of Figure 4, showing the frequency of conformational populations sampled over time.

Survival probability plots used to estimate transition rates among Env conformations.

Survival probability plots showing the probability of occupying a given FRET-defined conformation as a function of dwell time. Data points represent pooled dwell times from individual molecules, and curves show exponential fits used to extract transition rate constants (see Methods). These rate estimates correspond to the transitions summarized in Figure 5.