Bioorthogonal click labeling of an amber-free HIV-1 provirus for in-virus single molecule imaging

Summary

Structural dynamics of HIV-1 envelope (Env) glycoprotein mediates cell entry and facilitates immune evasion. Single-molecule FRET using peptides for Env labeling revealed structural dynamics of Env, but peptide use risks potential effects on structural integrity/dynamics. While incorporating noncanonical amino acids (ncAAs) into Env by amber stop-codon suppression, followed by click chemistry, offers a minimally invasive mean, this approach proved technically challenging for HIV-1. Here, we developed an intact amber-free HIV-1 system that overcomes hurdles of preexisting viral amber codons. We achieved dual-ncAA incorporation into Env on amber-free virions, enabling smFRET studies of click-labeled Env that validated the previous peptide-based labeling approaches by confirming the intrinsic propensity of Env to dynamically sample multiple conformational states. Amber-free click-labeled Env also enables real-time tracking of single virion internalization and trafficking in cells. Our system thus permits in-virus bioorthogonal labeling of proteins, compatible with studies of virus entry, trafficking, and egress from cells.

In Brief

Ao et al. developed an intact amber-free HIV-1 system that enables minimally invasive Env tagging by genetic code expansion and in-virus bioorthogonal click labeling for single-molecule FRET structural dynamics and advanced microscopy studies of virus entry, trafficking, and egress in living cells.

Graphical Abstract:

Introduction

The sparsely distributed envelope (Env) glycoprotein on human immunodeficiency virus 1 (HIV-1) is responsible for the host cellular receptor binding and cell entry while escaping neutralizing antibodies. Consequently, Env is a main target for designing vaccines, antibodies, and antiviral drugs. Env interacts with the host cellular receptor CD4 and coreceptor CCR5/CXCR4 to mediate viral membrane fusion into target cells through a series of large-scale conformational changes.^1–10 Functional Env is a homotrimer-of-heterodimer consisting of the gp120 exterior subunits and gp41 transmembrane subunits.^11–16 The gp120 subunit binds the primary receptor CD4 and coreceptor CCR5/CXCR4 on the susceptible target cells.⁴ The binding of CD4 induces opening of gp120 and the formation of binding sites for coreceptors. After binding to CCR5/CXCR4, additional conformational changes in gp120 are relayed to the membrane-anchored gp41 subunit. Subsequent conformational refolding of gp41 induces fusion of the viral membrane with the host cell membrane.^17–22 Besides mediating viral entry, Env harbors virus-specific antigenic epitopes and thus serves as the main target for antibody responses during viral infection.^1,23–27 Nevertheless, Env has evolved to successfully evade the host immune responses through multiple mechanisms, including extensive glycan shielding, regions of hypermutation, and conformational flexibility.^28–31 Collectively, Env exhibits a high capacity to undergo conformational switching upon interacting with host cells, which is essential for virus entry and immune evasion during HIV-1 infection life cycle. Thus, a better understanding of conformational dynamics of native Env on infectious viruses will inform the development of HIV-1 prevention and treatment strategies, such as vaccines, antibodies, and small-molecule inhibitors.^1,7–10

Single-molecule Förster resonance energy transfer (smFRET) imaging has provided unique insights into the HIV-1 entry mechanism by visualizing real-time conformational changes of Env tagged with a donor/acceptor pair of fluorophores undergoing distance-sensitive energy transfer on intact virions (Figure 1A).^32–35 Attaching donor/acceptor fluorophores to two sites of a molecule within their Förster distance of a few nanometers is a prerequisite for smFRET.³⁶ We have used enzymatic labeling for smFRET imaging of HIV-1 Env,^32,37–39 in which a customized Cy3/Cy5 FRET pair was used to label gp120 variable loops V1 and V4, respectively. This approach uses enzymes that site-specifically recognize the introduced short peptides (Q3 and A1) and transfer dye-conjugated substrates to specific residues within these peptides.^40,41 This approach requires 6 and 12 amino acids insertion into the gp120 V1 and V4, respectively. These insertions have generated concerns regarding the potential impact on structural integrity and unpredictable interaction with antibodies.⁴² Although these concerns have been addressed by extensive functional validations,^32,33,43 there are still some unreconciled differences between virus-associated states of Env identified by smFRET^8–10 and the structures revealed by cryo-electron microscopy (cryo-EM).^7,39 The pretriggered conformational state of Env reported by smFRET using peptide labeling tags has not been structurally observed by cryoEM.^7,39 There was a concern that the existence of this “missing” state might be promoted by peptide tags used for labeling. Thus, a minimal tagging approach for attaching dyes to Env is critical to directly address this concern. Such a labeling approach holds promise for further advancing our knowledge of Env-mediated virus entry in future smFRET studies.

Figure 1. Experimental design for minimally invasive smFRET of HIV-1 Env on the intact virion using amber suppression and bioorthogonal click labeling.

(A) smFRET of Env trimer in the virus context. The Env structure is adapted from PDB 4ZMJ (dually labeled gp120, brown; non-labeled gp120, cyan; gp41, blue).

(B, C) Theoretical (B) and experimental (C) schematics of using two-step minimal genetically encoded amber (TAG) tags for fluorescence labeling - amber suppression (step 1) followed by click chemistry (step 2). Amber suppressor tRNA and tRNA synthetase (tRNA^Pyl/NESPylRS^AF) introduce ncAA trans-cyclooct-2-en – L – lysine (TCO*) into the desired TAG position in Env (POI) on an assembled intact virion generated in mammalian cells. The supply of TCO* to the transfected cells allows the incorporation of TCO* at the TAG insertion site in Env, to which clickable fluorophores can be attached.

(D) Bioorthogonal click labeling via SPIEDAC. Tetrazine-conjugated Cy3 and Cy5 derivatives (LD555-TTZ and LD655-TTZ) react with the strained alkene of the TCO* via strain-promoted inverse electron-demand Diels-Alder cycloaddition (SPIEDAC). Fluorophores/dyes are colored green and red, respectively.

Design

The recently developed amber (TAG) stop-codon suppression through genetic code expansion in mammalian cells and copper-free bioorthogonal click chemistry to label dyes on proteins by only altering a single amino acid^44–50 seems an ideal labeling strategy for Env. This approach includes suppressing amber stop codons with clickable noncanonical/unnatural amino acids (ncAAs) by an engineered suppressor tRNA and its cognate tRNA synthetase (tRNA/RS) and the subsequent click labeling of dyes onto ncAAs.^44,51 The use of amber suppression, advantageous over peptide tags, minimizes the impact of tagging on the structural integrity of proteins, thus offering a less invasive means for site-specific labeling of Env (Table S1). Despite its obvious advances, click labeling of HIV-1 proteins in their close-to-native states has been rare, except for the two published pioneering reports.^44,52 Sakin and colleagues have generated singly click-labeled Env expressed on mammalian cells for monitoring Env mobility at the plasma membrane.⁴⁴ However, this pioneered study did not achieve click labeling of Env in the viral context, due in part to low single ncAA incorporation into Env at each of the three optimized sites.⁴⁴ Single-site click labeling of capsids in the viral context has been recently achieved, enabling tracking the journey of click-labeled capsids from the plasma membrane to the nucleus.⁵² The drastic differences in the number of copies of capsid vs. Env (thousands vs. a few) in/on each virion is likely the reason singly click-labeled Env in the viral context has been so difficult. In addition, smFRET requires dually click-labeled Env at proximal sites, contingent upon successful incorporation of two ncAAs into a single gp120 protomer that does not interfere with efficient virus assembly and release. This stringency imposes additional technical barriers. Thus, coupling a single or a pair of dyes to Env on virions through genetic code expansion–click chemistry has been lacking due to lower copies of Env on HIV-1, in addition to the low amber-suppression efficiency in mammalian cells.

Here, we aimed to establish a two-step minimally invasive labeling strategy for smFRET imaging on the virus via genetically encoding ncAA tags in gp120 and click labeling ncAAs with Cy3/Cy5 donor/acceptor dyes for prism-based total internal reflection fluorescence (prism-TIRF) microscopy (Figure 1). Incorporation of ncAAs into Env on intact HIV-1 virions can be achieved by co-transfecting HEK293T cells with the amber suppression machinery plasmid encoding for tRNA^Pyl/NESPylRS^AF and an amber-tagged Env plasmid, in the presence of ncAAs.^44,46,53,54 We introduced ncAA trans-cyclooct-2-ene-L-lysine (the A isomer, hereafter abbreviated TCO*) at an amber (TAG) stop codon engineered into gp120 (Figure 1B and C). TCO* reacts with tetrazine-conjugated Cy3 or Cy5 fluorophores by click chemistry via strain-promoted inverse electron-demand Diels-Alder cycloaddition (SPIEDAC, Figure 1D).^44,48

Results

Construction of dually amber-suppressed click-labeled Env on intact HIV-1 virions

We screened for amber suppression (ncAA incorporation) efficiency of Env TAG variants in the presence of tRNA^Pyl/NESPylRS^AF in HEK293T by comparing the HIV-1_{Q23 BG505} ^55–57 infectivity on TZM-bl cells with the counterpart without tRNA^Pyl/NESPylRS^AF. To identify amber-tolerable positions in gp120, we generated four full-length virus Env TAG variants that each carried single TAG insertion at different sites within V1 loop, denoted as N133_TAG, N136_TAG, N137_TAG, and T139_TAG. The amber-inserted positions were at the same or adjacent to the insertion sites of peptide tags.^33,43 N136_TAG showed better amber tolerance than the other three, as manifested by a higher ratio of infectivity under amber-suppression vs. non-suppression conditions (Figure S1). We combined N136_TAG with six amber insertion sites in the gp120 V4 loop and generated six dually amber-tagged HIV-1_{Q23 BG505} clones. Among six clones, N136_TAG S401_TAG, and N136_TAG S413_TAG can better tolerate amber insertions in Env by showing slight increases in infectivity under suppression conditions compared to non-suppression conditions (Figure S1). Attempts to visualize truncated (at Amber stop) and suppressed full-length Env by western blotting failed, likely because of insufficient virus particles released and low suppression efficiency. Related, we noticed that HIV-1_{Q23 BG505} wild-type infectivity exhibited an approximate 50- to 100-fold reduction (Figures 2 and S1) in the presence of the amber suppressor system compared to the results without. This reduction had limited our ability to screen amber-tolerable sites but provided a possible clue for improving amber suppression efficiency and ncAA incorporation into Env.

Figure 2. Construction and validation of an amber-free primary HIV-1 system.

(A) Alternations of four amber TAG termination codons of pol, vif, vpu, rev open reading frames (ORFs) in full-length wildtype HIV-1_{Q23 BG505} to ochre TAA termination codons, generating an amber-free HIV-1 genome.

(B–D) Amber-free HIV-1_{Q23 BG505} yields an approximate 10-fold increase in released infectivity (B) and an enhanced Env expression (C) and retains similar susceptibility to potent dodecameric CD4 molecules (sCD4_D1D2–Igαtp, D) compared to the amber-abundant wildtype counterpart. (B) Release of virus infectivity (mean ± SD) of the mutants carrying a termination codon change at Pol from amber to ochre (TAG to TAA) (Left), and amber-free viral particles (Right) in comparison to that of the wildtype counterpart. The red arrow points to a 5- to 10-fold increase in infectivity caused by a single termination codon change at Pol. For comparisons, infectivity for the wildtype HIV-1 was measured separately and thus shown twice in graph (B). Where indicated as “Amber-Suppression” or “S,” a plasmid encoding tRNA^Pyl/NESPylRS^AF was co-transfected with the HIV-1 genome, and the TCO* at 250 μM was added to the media. In contrast, only HIV-1 genome was transfected under the “Non-Suppression” or “NS” condition (grey bar in B). WT or wildtype, the original amber-abundant HIV-1_{Q23 BG505}.

(C) Env processing and incorporation (Figure S2) of viral particles were assessed by SDS-PAGE of virus supernatants, followed by western blotting using the polyclonal anti-gp120 and anti-HIV-IG. Repeated twice.

(D) Neutralization curves (mean ± SD) of amber-free and amber-abundant viral particles by SCD4_D1D2-Igαtp molecules.

Development of an intact amber-free HIV-1 system for improving amber suppression efficiency in the target protein

We then explored the cause for 50 to 100-fold reduction of viral infectivity. Analysis of the HIV-1 Q23 genome revealed that the full-length HIV-1_{Q23 BG505} contains four pre-existing amber termination codons in the Pol, Vif, Vpu, and Rev open reading frames (Figure 2A and Table S2). These ambers might incorporate TCO* by amber suppressors against intrinsic cellular translation machinery. Consequently, the read-through of supposed-to-terminate-translation amber codons is problematic through the generation of additional amino acids at the C termini of the above proteins before termination that could exhibit dominant-negative effects (Table S2). To test our hypothesis, given the critical role of Pol in the HIV-1 infection lifecycle, our first attempt was made to change amber to ochre (TAG to TAA) in Pol. Strikingly, a single TAG-to-TAA (amber to ochre) alternation in Pol improved infectivity under the suppression condition by approximately 5 to 10-fold (Figure 2B - left), confirming significant issues arising from the preexisting ambers in the HIV-1 genome. Thus, we reasoned that the previously observed unsatisfying efficiency of amber suppression and the resulting undetectable ncAA incorporated Env on gels were likely in part due to problematic suppression of preexisting ambers in HIV-1.

We then constructed an amber-free HIV-1 system by modifying all four termination codons from amber to ochre presented in the Pol, Vif, Vpu, and Rev genes (Figure 2A). A more than 10-fold increase in infectivity was obtained with the amber-free HIV-1_{Q23 BG505} construct, compared with the wildtype – amber-abundant (Figure 2B - right). Env expression and incorporation into virions of amber-free and amber-abundant HIV-1 constructs were further verified by immunoblots of virus supernatants using the polyclonal anti-gp120 and anti-HIV IgG (Human serum). Consistently, gp120 and Gag/p24 expression levels were higher in amber-free than in amber-abundant virions under the same suppression conditions (Figures 2C and S2). In control experiments, amber-free virions were as sensitive to neutralization by a potent soluble dodecameric sCD4_D1D2–Igαtp as wild-type virions were (Figure 2D). Thus, the developed amber-free system provides an optimal backbone for inserting genetically encoded amber tags in Env.

Intact amber-free HIV-1 carrying clickable dual-ncAA Env retains sensitivities to trimer-specific neutralizing antibodies

We next reconstructed and reevaluated amber-suppression efficiencies of previously screened Env TAG variants in the new amber-free backbone. We identified thus far superior amber-tolerable positions: N136_TAG in the V1 loop, S401_TAG and S413_TAG in the V4 loop (Figure S3). We generated four dually tagged Env constructs in the context of an intact amber-free HIV-1, including our choice of two dual-amber N136_TAG S401_TAG and N136_TAG S413_TAG and two hybrid amber/peptide N136_TAG V4A1 and V1Q3 S401_TAG mutants as comparisons (Figure 3A). Single-amber Env variants N136_TAG, S401_TAG, and S413_TAG retained ~ 30% of infectivity (suppression efficiency) compared to the wildtype (Figure 3B), similar to the reported 5 ~ 20% infectivity of three different NL4–3 Env modified with single amber introduced into gp120.⁴⁴ Hybrid amber/peptide N136_TAG V4A1 and V1Q3 S401_TAG showed ~20% suppression efficiency (Figure 3B, left and middle panels). Comparable to infectivity of hybrid Env, dual-amber Env (N136_TAG S401_TAG and N136_TAG S413_TAG) were successfully suppressed with efficiencies of 10 to 20% relative to the wild-type virus (Figure 3B, right panel). In contrast to bacteria, the efficiency of amber suppression or ncAA incorporation in mammalian cells is known to be challenging,^44,46,58, and for Env it might be unlikely to reach 100%. Truncated Env proteins due to translation termination following incomplete suppression might exhibit dominant-negative effects on Env trimer folding and reduce infectivity. Therefore, it is not unexpected that the infectivity of virus carrying amber-tagged virus Env can be somewhat lower than peptide-tagged virus Env (V1Q3 V4 A1 tags on BG505 did not affect viral infectivity in our previous study⁴³). Western blotting of post-transfected cell supernatants revealed that all four dually tagged Env variants were successfully suppressed, proteolytically processed (gp120 signal), and incorporated onto amber-free virions, in contrast to no traceable Env expression under non-suppression conditions (Figure 3C, top). By comparison, both gp120 and gp160 expressions were observed in post-transfected cell lysates (Figure 3C, bottom). The presence of gp160 in cell lysate and its absence in cell supernatant implies that precursor gp160 Env is largely (if not entirely) proteolytically cleaved to produce mature gp120/gp41 on the released virions. The morphology of packaged virus particles is unaffected in 2D EM images acquired by cryo-electron tomography (Figure 3D). Nanoparticle Tracking Analysis (NTA) showed the viral particles with a mode diameter size of 109.2 ± 15.9, 125.7 ± 3.4, and 121.8 ± 4.2 nm for amber-free virions carrying no-amber or two dual-amber Envs, respectively (Figure S4). These mode diameters are close to the diameter of 120 nm reported for native HIV particles⁵⁹, suggesting the normal assembly of ncAA-incorporated virions.

Figure 3. Efficient ncAA incorporation into dually amber-tagged Env on the amber-free HIV-1 virus that retains sensitivities to trimer-specific neutralizing antibodies.

(A) Scheme showing tag insertion sites and denotations of labeling tags in BG505 gp120 subunit.

(B) Release of infectivity (mean ± SD) on TZM-bl cells from HEK293T cells transfected with plasmids encoding singly amber-tagged, hybrid peptide/amber-tagged, dually amber-tagged Env_BG505 in the context of an amber-free Q23 backbone (HIV-1_{Q23 BG505}).

(C) Env processing and incorporation of amber-free, dually amber-tagged Env (N136_TAG S401_TAG and N136_TAG S413_TAG), or hybrid peptide/amber-tagged Env (N136_TA V4A1 and V1Q3 S401_TAG) viral particles. Proteins were analyzed by immunoblotting using a polyclonal antiserum against HIV-1 gp120.

(D) A representative cryoET tomographic slice of amber-free and two dually amber-tagged HIV-1 viral particles, respectively. Scale bar: 50 nm.

(E) Neutralization curves (mean ± SD) of amber-free and dually amber-tagged HIV-1 viral particles by trimer-specific antibodies, including PG9, PG16, and PGT151. Dose-dependent relative infectivity was normalized to the mean value determined in the absence of the corresponding ligand. All viral particles generated in this study were prepared using 100% of indicated HIV-1 constructs during transfection.

We next tested the sensitivities of full-length dual-ncAA Env virions to trimer-specific neutralizing antibodies, including Env apex-targeting PG9 and PG16 and gp120/gp41 interface-directed PGT151.^33,60,61 Neutralization curves of dual-amber Env viral particles resembled that of amber-free wildtype ones (Figure 3E), indicating the preservation of neutralization sensitivities of dual-ncAA incorporated Env to trimer-specific antibodies. Thus, we demonstrated that our constructed dual-amber Env variants were successfully expressed with efficient ncAA insertions, processed, and incorporated into intact amber-free virions that retained neutralization sensitivities to trimeric Env-directed antibodies. The consistent immunoblots and neutralization results indicate efficient trimerization and maturation of full-length dual-ncAA incorporated Env proteins, which present properly on intact HIV-1 virions. These functional validations pave the way for minimally invasive smFRET of Env.

smFRET imaging of dually click-labeled Env on the intact amber-free HIV-1 virion

We next performed smFRET imaging of dually click-labeled Env on amber-free intact virions (Figures 1A and 4). Three different types of donor/acceptor-dually labeled Env were investigated by smFRET imaging, including 1) dually click-labeled Env (N136_TAG S401_TAG and N136_TAG S413_TAG), 2) hybrid click/peptide Env (N136_TAG V4A1 and V1Q3 S401_TAG), and 3) dual-peptide Env (V1Q3 V4A1). Dually click-labeled Env was generated by randomly coupling tetrazine-conjugated Cy3/Cy5 derivatives LD555-TTZ and LD655-TTZ to TCO* ncAAs that are incorporated into Env V1V4 loops (exampled in Figure 4A or flipped sites of labeling). Two hybrid click/peptide Env were used as comparisons, in which one label was attached by click chemistry and the other by enzymatic modification of a peptide tag. Two labels of dual-peptide Env, LD555-CD and LD655-CoA were attached by enzymatic modification of two peptide tags as previously described.^32,33,43,62

Figure 4. Representative fluorescence and FRET trajectories of unliganded dually click-labeled Env on amber-free virions at the single-molecule level.

(A) A donor/acceptor-paired labeled protomer within an Env trimer at indicated sites on an intact virion was immobilized on a quartz slide and imaged by prism-TIRF microscope. Donor dye was excited with an incident laser and transferred its energy to nearby acceptor dye depending on its proximity. The structure of the Env trimer (PDB 4ZMJ)¹⁵ was fitted into in-situ electron density map of the ligand-free trimer (EMD-21412)³⁴. Dually labeled gp120 monomer, brown; unlabeled gp120 monomer, cyan; gp41, blue; donor fluorophore, green; acceptor fluorophore, red.

(B, C) Two exampled fluorescence (Top panel) and derived FRET trajectories (Bottom panel) of individual dually click-labeled amber-free virions. LD555-TTZ and LD655-TTZ were randomly attached to TCO* at sites N136_TAG and S413_TAG (Top panel). Their corresponding FRET trajectory (blue) with overlaid idealization was generated using Hidden Markov Modeling (HMM) (purple) (Bottom panel). The monitored individual gp120 monomers undergo transitions between three FRET-identified conformational states (State 1, low FRET; State 2, high-FRET; State 3, intermediate-FRET). Photobleaching points (B, the acceptor bleaches first; C, the donor bleaches first) indicate time-course fluorescence/FRET monitoring at the single-molecule level.

As smFRET monitors a single FRET-labelled gp120 protomer within an Env trimer in otherwise wild-type trimers on a single virion, we could achieve 1 labeled-protomer:1 trimer:1 virion by adjusting the co-transfecting ratio of tagged vs. wild-type Env plasmids in HEK293T cells.^32,33,63 Based on the double-amber suppression efficiency, we used an approximate 4:1 ratio of amber-free wildtype and dual-amber virus Env constructs to achieve ideal conditions for smFRET: single FRET-labelled gp120 protomer embedded in an otherwise wildtype environment. Fluorescently labeled virions were immobilized on a streptavidin-coated quartz plate-coverslip sample chamber through the incorporation of biotinylated lipids into the virus membrane. The LD555-TTZ label on immobilized virions was excited, and fluorescence intensities of LD555-TTZ and LD655-TTZ were recorded simultaneously by our customized prism-TIRF microscope. As in smFRET studies of enveloped viruses³⁸, single-color virions were automatically filtered based on the cross-correlation between two fluorescence trajectories, whereas multi-labeled protomers in a trimer were manually filtered based on stepwise photobleaching events. We extracted hundreds of fluorescence trajectories with anti-correlated features between donor and acceptor fluorescence intensities (Figure 4B and C; top). Anti-correlation is the revealing signature of energy transfer between a donor and an acceptor, reflecting conformational changes of Env on individual virions. The donor or acceptor photobleaching point, either acceptor bleaching first (Figure 4B) or donor bleaching first (Figure 4C), indicates the detection of a single FRET-labeled Env on an immobilized intact virion. We further derived the FRET efficiency (FRET traces/trajectories, exampled in Figures 4B and 4C; bottom), documenting real-time conformational changes over time of individual Env on virions.

Consolidation of Env conformational profiles by minimally invasive amber-click smFRET systems in the context of intact amber-free virions

Qualitative and quantitative smFRET analysis of dually click-labeled Env in the amber-free virus context revealed consistent three-state conformational profiles of Env on the viruses with results obtained from previous dual-peptide Env and two hybrid click/peptide Env (Figures 5, S5, S6, and Table S3). A pair of donor/acceptor dyes attached to different Env constructs include LD555-TTZ/LD655-TTZ (R₀ ~61.0 Å) for labeling both N136_TAG S401_TAG and N136_TAG S413_TAG, LD555-TTZ/LD655-CoA (R₀ ~60.0 Å) for N136_TAG V4A1, LD555-CD/LD655-TTZ (R₀ ~61.0 Å) for V1Q3 S401_TAG, and LD555-CD/LD655-CoA (R₀ ~60.0 Å) for V1Q3 V4A1. In bulk photophysical measurements, no noticeable differences were found in the spectra of different LD555 or LD655 derivatives (Figure S5A) or shifts in excitation or emission spectra for all dyes used after attaching to Env constructs. For example, the spectra of single-color N136_TAG S401_TAG that was singly labeled by LD555-TTZ or LD655-TTZ were almost identical to the spectra of free dyes (Figure S5B). As expected, the fluorescence spectra of dual-color N136_TAG S401_TAG labeled by LD555-TTZ and LD655 showed emission peaks of both dyes (Figure S5C).

Figure 5. Consolidation of conformational landscapes of Env by smFRET imaging of click-labeled Env on amber-free virions.

(A–C) FRET histograms of dually click-labeled Env at sites N136_TAG S401_TAG in the context of intact virions under different conditions: ligand-free (A), PGT151-bound (B), and sCD4_D1D2–Igαtp-bound (C). The number (N_m) of individual trajectories was compiled into a conformation-population FRET histogram and fitted into a 3-state Gaussian distribution (solid gray line) centered at ~0.1-FRET (dashed cyan), ~0.3-FRET (dashed red), ~0.6-FRET (dashed green). The fitted distributions (solid gray line) were overlaid with the corresponding referenced distributions of dual-peptide Env V1Q3 V4A1, where ref stands for the referenced distribution.

(D–F) Experiments as in A–C, revealing conformational landscapes for the ligand-free (D) and upon binding of PGT151 (E) or sCD4_D1D2–Igαtp (F) to dually click-labeled Env N136TAG S401TAG.

(G) Bar graph of relative conformation-occupancy of virus Env using different FRET probes. FRET histograms represent mean ± SEM, determined from three randomly assigned populations of FRET traces under indicated experimental conditions. Relative occupancies of each conformational state and fitting parameters were summarized in Table S3.

In single-molecule experiments, previous smFRET using peptide tags showed that native Env on the virus is intrinsically dynamic, transiting from the pre-triggered conformation (State 1) through a partially open intermediate (State 2) to the fully activated open conformation (State 3) upon binding to CD4 molecules.^32,33,38,43 FRET trajectories report time-correlated relative donor-acceptor distance changes associated with conformational changes between V1 and V4 regions within the Env monomer. Compiling hundreds of FRET traces obtained under different conditions led to respective merged histograms fitted with Gaussian functions, indicating conformational distributions with state occupancies (Figures 4 and 5). FRET histograms of Env revealed by dually click-labeled N136_TAG S401_TAG (Figure 5A–C), N136_TAG S413_TAG (Figure 5D–F), as well as hybrid click/peptide N136_TAG V4A1 or V1Q3 S401_TAG (Figure S6), were in good agreement with previous results from dual-peptide Env V1Q3 V4A1 (overlaid lines). Similarly, the three-state conformational ensemble contained State 1 (low-FRET with mean 0.11 ± 0.02), State 2 (high-FRET with mean 0.63 ± 0.05), and State 3 (intermediate-FRET with mean 0.28 ± 0.04). Transitions between low- and high-FRET states and high- and intermediate-FRET states occurred frequently, but the low- and intermediate-FRET states were rarely observed, as shown in the representative traces (Figures 4B, 4C, and S6A). The unliganded N136_TAG S401_TAG, N136_TAG S413_TAG, N136_TAG V4A1, and V1Q3 S401_TAG tagged Env in the amber-free virus context all predominantly resided in a low-FRET State 1 with the mean occupations of 41%, 50%, 42%, and 52%, respectively, similar to the previous studies (Figures 5, S6, and Table S3).^33,43 Potent sCD4_D1D2–Igαtp shifted the conformational landscape of Env towards intermediate-FRET State 3-dominant distributions with the occupations of respective 38%, 47%, 45%, and 50% from the above-mentioned four smFRET systems (Figures 5, S6, and Table S3). Adding the PGT151 antibody that asymmetrically targets the interface between gp120 and gp41 resulted in the preferential conformation to be high-FRET State 2 with the mean occupations of 54%, 58%, 57%, and 50% (Figures 5, S6, and Table S3). Donor/acceptor dyes coupled to distinct pairs of sites on Env may contribute to the observed variations in the relative occupancy of three conformations in different FRET systems.

The minimal tags used in dually click-labeled Env did not appear to narrow the FRET histograms compared to the relatively longer peptide tags used in dual-peptide Env (Figure 5). The similarity of FRET histograms using different tags could be explained by the high degrees of flexibility and fast motions of Env variable loops V1V4 in which tags are inserted. The sampling space of tags is likely within that of V1V4 loops; thus, the impact of using different tags is likely to be buried by the intrinsic flexibility of V1V4 loops. The motions of V1V4 loops (including tags) are at a faster time scale (nanoseconds to microseconds) and are beyond our temporal resolution of 40 milliseconds, which are also averaged out during global milliseconds-to-seconds conformational motions of Env. In summary, the FRET histogram similarities could be due in part to the fast-sampling flexibility of V1V4 loops, freely rotated dyes, and similar R₀ values (~60.0 vs. 61.0 Å) of different donor/acceptor pairs. All factors mentioned above likely contributed to the consistency in conformational propensities of Env probed using all established FRET systems (Figure 5G). These results reinforce our previous finding of Env sampling three primary large-scale milliseconds-to-seconds global conformations. Altogether, minimally invasive amber-click smFRET systems of Env in the context of intact amber-free virions consolidated conformational landscapes of Env and ruled out the possibility that Env conformational states are affected by peptide tagging.

Application of amber-free HIV-1 Env labeling for single virus tracking in cells

We further explored the potential of the amber-free HIV-1 system for adopting click labeling in live cell imaging settings. For the purposes of prolonged single virus tracking, virions containing 100% amber N136_TAG-Cy5 mutant Env were produced and labeled with LD655-tetrazine. Live cell confocal microscopy of single-labeled amber-free HIV-1 N136_TAG-Cy5 revealed virus internalization into target TZM-bl cells labeled with SPY555-Actin. Viruses were pre-bound to cells in the cold, unbound particles washed away, and their entry was initiated by shifting to 37 °C. The trajectory and the Cy5 and SPY555 intensity profiles of a representative single virion are shown in Figure 6A, B, and Movie 1. The virion lingered at the cell periphery for about 20 min and then was rapidly internalized, as evidenced by initial colocalization with the peripheral actin cortex followed by net displacement to the center of the cell (Fig. 6B, C). Consistent with endosomal trafficking, the virus’ speed reached 0.17 μm/s. (Figure 6D). The super-linear mean square displacement (MSD) profile for the trajectory further supports a net directional movement of the virus away from the cell periphery. (Figure 6E). These findings support the notion that HIV-1 is internalized and transported in endosomes.^64,65 Our results thus demonstrate the versatility of the amber-free virus labeling approach that can be applied to live-cell tracking of HIV-1 entry into cells.

Figure 6. Trafficking of amber-free HIV-1_{Q23 BG505} virions carrying 100% click-labeled Env N136_TAG-Cy5 in TZM-bl HeLa cells.

(A) Live cell imaging of SPY555-Actin (magenta) labeled TZM-bl cells with an internalization event of a single virion with click-labeled Env N136_TAG-Cy5 (green). The HIV-1 particle of interest initially associated with the cell periphery is marked with a dashed yellow circle.

(B) Representative single particle trajectory of the click-labeled HIV-1 internalization event from panel A. The single HIV-1 trajectory is overlayed on the source image to demonstrate the depth of virus internalization into the cell.

(C) Representative single particle intensity trace for the particle of interest from panel A. The internalization event is indicated with the black arrowhead on the panel.

(D) Instantaneous particle speed demonstrating an increase in virus trafficking speed during internalization into the cell, consistent with virus transport in endosomes. The internalization event is signified with an arrow.

(E) Mean squared displacement (MSD) plot of the particle from panel A demonstrating that the particle of interest undergoes directional motion for a fraction of its trajectory coinciding with the internalization even.

Discussion

Here, we developed and optimized an intact amber-free HIV-1 provirus that significantly improves the incorporation of ncAAs into Env as a minimally invasive method to label viral proteins at single-residue precision by genetic code expansion. The developed provirus further enabled bio-orthogonal click labeling of Env and the subsequent establishment of the minimal invasive amber-click smFRET systems for Env. The amber-free HIV-1 system is versatile and valuable for in-virus protein biorthogonal click labeling, particularly as lentiviral cores are often used to study glycoproteins from other viruses by pseudotying.^{62,63,66–68}

Demands for minimally invasive labeling strategy in smFRET imaging of Env

To achieve smFRET of Env, we had previously deployed an alternative approach of introducing short peptide tags ^40,41 into variable loops of gp120.^32,33,43 Nevertheless, the introduction of a few amino acids still holds the possibility of affecting Env conformations in unpredictable ways.⁴² There was also a concern, or controversy, that the use of peptide tags for FRET labeling could be the cause for the observation of an additional conformational state at the smFRET level that has not been structurally characterized. The amber-click strategy for site-specific incorporation of ncAAs and labeling facilitates the identification of additional, more permissive labeling sites that lower the risk of functional impact.^51,69,70 This approach is also promising for accessing the sterically restricted gp41 subunit that does not tolerate peptide insertions in our hands.

Challenges in amber-click labeling the sparse Env on the virus

The challenges in achieving dually amber-click labeling of Env on the virus were initially underestimated and impeded our long-sought-after minimally invasive smFRET imaging of Env. No success of singly click-labeled Env on intact HIV-1 virions was reported. On the positive side, progress has been made with click-labeling of pseudotyped influenza A hemagglutinin⁶³ and Ebolavirus glycoprotein.⁶⁶ However, the number of Env copies on HIV-1 is much lower than influenza A hemagglutinin or Ebolavirus glycoproteins. Labeling of HIV-1 capsids has recently been reported⁵² and in a new preprint⁷¹ posted at the time of this submission. Again, a few copies of Env on each virion are in sharp contrast to thousands of copies of capsid protein (p24).^13,72–74 Apparently, major challenges, such as low copies of Env on virions, incomplete suppression in mammalian cells, and high Env structural complexity, have impeded the application of ncAAs to Env.

Pinpointing a limiting factor of amber suppression in the HIV-1 genome

We realized that many HIV-1 proviruses carry an abundant number of amber codons in their genomes. These preexisting ambers could result in aberrant extensions of viral proteins, resulting in the synthesis of dominant-negative proteins. We thus constructed an intact amber-free HIV-1 system by genetically modifying naturally occurring amber codons (TAG) to ochre codons (TAA) in critical protein-coding genes, warranting their correct translating terminations. Altering the amber codon of Pol largely rescued virus infectivity (Figure 2B), implying that the aberrant extension of Pol is likely the major reason for the poor production/infectivity of naturally amber-abundant HIV-1_Q23 (Table S2). Changing the amber of Vpu to ochre causes a point mutation in Env, but it is within the signal peptide and does not appear to interfere with Env secretion (Figure 2C). Modifying the amber of Rev to ochre results in a mutation in the C-tail of Env gp41, but it does not seem to affect Env conformational dynamics (Figures 5 and S6).

While we could not biochemically document problematic suppression of amber stop codons within the HIV-1 genome under suppression conditions, our reasoning that it is likely the main hurdle of applying genetic code expansion to HIV-1 is based on the following observations: 1) all amber stop codons are expected to be indiscriminately read through by suppressor tRNAs inside the cytosol of producer cells including ambers in the HIV-1 genome; 2) indeed, we observed a strong rescue of infectivity by a single change from amber to ochre in Pol and further rescue by modifying all ambers to ochres in intact HIV-1 genome (Figure 2B); 3) our constructed amber-free system enabled the first-ever visualization of singly and dually amber-suppressed virus Env that can be detected by western blotting (Figure 3C), in contrast to unsuccessful attempts made on the amber-abundant virions by us and others.

The amber-free HIV-1 system is compatible with click-labeling of dually amber-suppressed Env, enabling minimally invasive smFRET imaging of Env on the virus

Built upon our newly developed amber-free HIV-1 provirus, we further optimized two candidate clones encoding dually TAG-tagged Env. In contrast to the pioneering study⁴⁴, our infectivity and immunoblot results provided evidence of enhanced suppression efficiency and resulting effective ncAA incorporation in Env (Figure 3). smFRET results of dually click-labeled Env in parallel with hybrid click/peptide Env were in global agreement with our previous observations (Figures 4 and 5).^32,33,43. These results further rule out the scenarios that sCD4_D1D2–Igαtp and PGT151 interact with peptide tags/dyes, and that peptide tags affect the presence of smFRET-identified conformational states of Env. Our results directly prove that the Env states observed at the level of smFRET are an intrinsic dynamic property of Env and not due to a labeling artifact using peptide tags.

The amber-free context also holds great promise for smFRET imaging of the highly constrained transmembrane subunit gp41 of Env, for which the fusion-inducing conformational events remain elusive. The amber-free HIV-1 system is expected to significantly increase our success rate for site-specifically introducing a pair of clickable donor/acceptor dyes into conformational switching regions of gp41 by only altering a single amino acid per site. Therefore, the potential of amber-click Env in the amber-free context will be exploited in our future smFRET imaging of conformational dynamics and allostery during virus entry.

Application of an amber-free HIV-1 system for in-virus protein bioorthogonal labeling

The amber-free clinically relevant infectious virus system is highly versatile for screening amber-tolerable sites in any HIV-1 proteins, offering a high chance of efficient labeling of proteins in/on HIV-1 virions. It, therefore, permits visualizing/tracking under-investigated viral proteins within the context of infectious virus particles close to their native states during the trafficking of viruses in cells using live-cell microscopy, correlative imaging, and super-resolution microscopy. For example, the identified amber-tolerable N136_TAG, S401_TAG, and S413_TAG sites on Env in the amber-free Q23 virus context can be used for virus tracking in living cells using confocal, TIRF, super-resolution microscopy. Searching for bioorthogonal clickable sites in the capsid in the context of our amber-free clinically relevant HIV-1_Q23 infectious system, which is more relevant than a lab-adapted HIV-1 isolate,^52,71 is promising for tracking capsid uncoating and other imaging applications.

In-virus protein labeling commonly involves the use of peptide tags, fluorescent proteins, or antibody staining. Peptide tags attached to viral proteins or fluorescent proteins packaged into virions could affect the trafficking speed or trajectories of virions. Antibodies used in antibody staining could trigger unpredictable cellular responses, which further complicates the understanding of virus-host interactions. Click-labeled virions have the potential to minimize these effects. The broad application of click labeling will open gates for visualizing complicated processes in the virus lifecycle, as demonstrated by the recent inspiring HIV-1 studies.^44,52,71 Importantly, live cell tracking of single HIV-1 entry and fusion has been commonly performed using pseudoviruses over-incorporating Env (see, for example ^64,75). The use of full-length HIV-1 proviruses carrying Env trimers at natural levels is likely more presentative of bona fide virions. Our study directly labels and tracks native HIV-1 particles carrying a proper number of copies of a nearly fully proteolytically processed Env precursor.

To demonstrate the potential applications of our amber-free HIV-1 system, we tracked single viruses in live cells (Figure 6 and Movie 1). These experiments demonstrate the feasibility of long-term tracking of single virions with only a few Env trimers incorporated. These studies are significant in the light of evidence supporting productive HIV-1 entry through an endocytic route in HeLa-derived and primary CD4⁺ T-cells.^64,75 Further studies are required to deeply understand HIV-1 internalization dynamics and its relevance to productive infection.

Limitations of the study

The amber-suppressing efficiencies of Env are still not perfectly satisfactory. Two factors likely limit amber suppression in mammalian cells. Firstly, the efficiency of incorporating ncAAs is still limited by unpredictable competition with cellular amber codons translations, e.g., release factors.^44,76–80 Secondly, the tRNA^pyl/PylRS^AF and its advanced NES systems from Methanosarcina species are not exploited fully in the eukaryotic cells,⁴⁴ leading to the decrease of amber-suppression efficiency, with the expectation of the exponential decrease efficiency for a duall amber-suppression in the same protein.⁸¹ Future efforts to develop promiscuously amber-suppressed systems with higher suppression efficiency are thus required. This may include an improved pylRS system with enhanced translations⁴⁶ and the design of an orthogonal tRNA/pylRS pair permitting simultaneous incorporation of two distinct unnatural amino acids against different stop codons.⁴⁵ Due to the instrumental limitations, we have not yet explored their full potential in advanced microscopic studies of click-labeled viral proteins beyond smFRET imaging of Env. Nevertheless, the success of dually click labeling Env at the single-molecule level in this study holds great promise for bioorthogonal click labeling other viral proteins with high copies at the ensemble or single-molecule levels.

STAR ★ METHODS

RESOURCE AVAILABILITY

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Maolin Lu (maolin.lu@uthct.edu). This study generates new recombinant DNA (plasmids).

Materials Availability

All primers used for constructing amber-free HIV-1 and amber-tagged Env plasmids are shown in Table S4. All template and constructed recombinant DNA are listed in the key resources table. The replication-incompetent amber-free HIV-1 provirus plasmid is available from Addgene (plasmid ID: 213007) or from the lead contact upon request. All other unique reagents generated in this study are available from the lead contact upon request.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
PG9	NIH HIV Reagent Program	Cat # ARP-12149
PG16	NIH HIV Reagent Program	Cat # ARP-12150
PGT151	Peter D. Kwong, NIH/NIAID	N/A
Anti-HIV-1-gp120	NIH HIV Reagent Program	Cat # ARP-288
Anti-HIV-1-p24	NIH HIV Reagent Program	Cat # ARP-3537
HRP-conjugated rabbit anti-sheep IgG secondary antibodies	Invitrogen	Cat # 31240
HRP-conjugated rabbit anti-goat IgG secondary antibodies	Invitrogen	Cat # 31402
Chemicals, peptides, and recombinant proteins
Dulbecco’s Modified Eagle Medium	Gibco	Cat # 2509022
Fetal bovine serum	Atlanta Biologicals	Cat # S11550
L-glutamine	Gibco	Cat # 25030-081
Penicillin-streptomycin	Gibco	Cat # 15140-122
0.05% Trypsin-EDTA, phenol red	Gibco	Cat # 25300054
Dulbecco’s Phosphate Buffered Saline	Gibco	Cat # 14190144
Opti-MEM reduced serum medium	Gibco	Cat # 31985070
Polyethyleneimine	Polyscience	Cat # NC1014320
Fugene 6	Promega	Cat # E2311
Unnatural amino acid transcyclooct-2-ene lysine (TCO*)	Sichem	Cat # SC8008
4–12% Bis Tris polyacrylamide	Bio-Rad	Cat # NP0322BOX
MOPS SDS running Buffer	Invitrogen	Cat # NP0001
Trans-Blot Turbo Trans Buffer	Bio-Rad	Cat # 10026938
Precision plus protein dual color standards	Invitrogen	Cat # 1610374
Tween20	Invitrogen	Cat # BP337-500
Nonfat-Dried Milk Bovine	Sigma	Cat # SLBZ5377
Clarity western ECL substrate	Bio-Rad	Cat # 170-5060
Sucrose	Sigma	Cat # S0389
Transglutaminase	Sigma	Cat # 80146-85-6
AcpS (acyl-carrier-protein synthase)	Maolin Lu Lab, University of Texas at Tyler	N/A
PEG-passivated, streptavidin-coated quartz slides	This paper	N/A
Trolox	Sigma	Cat # 238813
Cyclooctatetraene	Sigma	Cat # 138924
Nitrobenzyl alcohol	Sigma	Cat # N12821
Protocatechuic acid	Sigma	Cat # 37580
Protoatechuate 3,4-deoxygenase	Sigma	Cat # P8279
LD555-Tetrazine	Lumidyne Technologies	N/A
LD655-Tetrazine	Lumidyne Technologies	N/A
LD555-COA	Lumidyne Technologies	N/A
LD6555-CD	Lumidyne Technologies	N/A
BCN-OH quencher	Sigma	Cat # 742678
PEG2000-biotin	Avanti Polar Lipids	Cat # 880129P
Streptavidin	Invitrogen	Cat # S888
60% (w/v) Opti-prep	Sigma	Cat # D1556
Nitrobenzyl alcohol	Sigma	Cat # N12821
Protocatechuic acid	Sigma	Cat # 37580
Protocatechuic-3,4-dioxygenase	Sigma	Cat # P8279
Acetone, EM-Grade, Glass-Distilled	Electron Microscopy Sciences	Cat # 10015
SPY555-Actin	Cytoskeleton	CAT # CY-SC202
Critical commercial assays
KAPA SYBR FAST qPCR Master Mix (2X) Kit	KAPA Biosystems	Cat # KK4600
Pierce™ Gaussia Luciferase Glow Assay Kit	Thermo Fisher Scientific	Cat # 16158
Experimental models: cell lines
HEK293T	ATCC	Cat # CRL-3216
TZM-bl	NIH HIV Reagent Program	Cat # ARP-8129
Experimental models: organisms/strains
HIV-1 lentiviral particles carrying envelope	This paper	N/A
Oligonucleotides
Primers for constructing plasmids, see Table S4	This paper	N/A
Recombinant DNA
Wildtype HIV-1Q23 BG505	Julie Overbaugh Lab, Fred Hutch Cancer Center	N/A
Amber-free HIV-1Q23 BG505	This paper	N/A
tRNAPyl/NESPylRSAF	Edward Lemke Lab, Johannes Gutenberg –University Mainz	N/A
Amber-free HIV-1Q23 BG505 N136TAG V4A1	This paper	N/A
Amber-free HIV-1Q23 BG505 V1Q3 S401TAG	This paper	N/A
Amber-free HIV-1Q23 BG505 N136TAG S401TAG	This paper	N/A
Amber-free HIV-1Q23 BG505 N136TAG S413TAG	This paper	N/A
HIV-1-lnGluc	Walther Mothes Lab, Yale University	N/A
Deposited data
Replication-incompetent amber-free HIV-1Q23 BG505 deltaRT	Addgene	Cat # 213007
Replication-competent amber-free HIV-1Q23 BG505	This paper	Mendeley Data (https://doi.org/10.17632/d9ww5mh46d.1)
Software and algorithms
GraphPad Prism v8.4.3	GraphPad	https://www.graphpad.com/
MATLAB	Mathworks	https://www.mathworks.com/
PyMOL	Schrödinger	https://pymol.org/2/
Chimera	University of California, San Francisco	https://bio3d.colorado.edu/SerialEM/
SerialEM software package	David N. Mastronarde, University of Colorado Boulder	https://bio3d.colorado.edu/SerialEM/
IMOD software package	David N. Mastronarde, University of Colorado Boulder	https://bio3d.colorado.edu/imod/
Other
Trans-Blot Turbo system	Bio-Rad	N/A
ProFlex PCR System	Applied biosystems	N/A
NanoSight instrument	Malvern Panalytical	N/A
Bio-Rad ChemiDoc™ XRS+ System	Bio-Rad	N/A
BioTek Synergy H1 Microplate Reader	BioTek	N/A
Prism-based TIRF Microscope	Maolin Lu Lab, University of Texas at Tyler	N/A
Gradient master 108	Biocomp	N/A
FluoroMax - Spectrofluorometer	Horiba Scientific	N/A
Gravity-driven plunger apparatus	Walther Mothes Lab, Yale University	N/A
QUANTIFOIL® holey carbon grids	Electron Microscopy Sciences	Cat # Q250-CR1
Titan Krios G3 microscope	Thermo Fisher Scientific	N/A
Zeiss LSM880 laser scanning confocal microscope	Gregory Melikian Lab, Emory University	N/A
96-well white plates for luciferase assays	Costar	Cat # 3917
ThermalGrid Rigid Strip PCR tubes	Denville Scientific Inc	Cat # C18064
Acrodisc 32 mm Syringe Filter w/0.45 μm Supor Memberane Non-Pyrogenic.	PALL Life Sciences	Cat # 4654
Immun-Blot PVDF membrane	Bio-Rad	Cat # 1620174
Mini Trans-Blot Filter Paper	Bio-Rad	Cat # 1703932

Data and Code Availability

• Sequence and map of replication-incompetent amber-free HIV-1 provirus plasmid have been deposited into Addgene under plasmid number 213007 listed in the key resources table. Sequences and maps of both replication-competent and incompetent amber-free HIV-1 provirus have been deposited at Mendeley. The DOI is listed in the key resources table. Other data reported in this paper will be shared by the lead contact upon request.

• This paper does not report original code.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Cell Lines and Culture Conditions

HEK293T cells (human embryonic kidney, ATCC) were used for producing replication-complement and replication-incompetent HIV-1 viruses. TZM-bl cells (generated from HeLa cells) that stably express receptor CD4 and co-receptors CCR5 and CXCR4 were used as target cells for quantifying the infection abilities of produced HIV-1 viruses⁸². These cell lines were cultured in high glucose Dulbecco’s Modified Eagle Medium supplemented with 10% (v/v) fetal bovine serum, 100 μg/mL penicillin-streptomycin, and 6 mM L-glutamine, and maintained in a 37 °C incubator supplied with 5% CO₂.

METHOD DETAILS

Plasmid construction

The wildtype HIV-1_{Q23 BG505}, an infectious clade A clone, contains full-length Env_BG505 with Q23 as the backbone^55–57 (a gift from Julie Overbaugh, Fred Hutchinson Cancer Center, Seattle). BG505 is a clinical Clade A transmitted/founder Env strain. HIV-1_{Q23 BG505} ΔRT is a replication-incompetent clone that lacks the reverse transcriptase gene^32,33,43. Amber-free HIV-1_{Q23 BG505} plasmid was generated from HIV-1_{Q23 BG505} plasmid in which TAG stop codons presented in the Pol, Vif, Vpu and Rev genes were altered to TAA stop codons. Tagged-Env_BG505 plasmids with amber-free Q23 as backbone include amber-free HIV-1Q23 BG505 N136TAGV4A1, V1Q3S401TAG, N136TAGS401TAG and N136TAGS413TAG. Amber-free N136_TAGV4A1 plasmid was tagged with a single amber codon TAG at the position of codon 136 (Asn) in the variable V1 loop and an A1 peptide tag (A1, GDSLDMLEWSLM) in the variable V4 loop within gp120 of Env, while plasmid amber-free V1Q3 S401_TAG contained the insertions of Q3 peptide tag (Q3, GQQQLG) in V1 and a single amber codon at the position of 401 (Ser) in the V4 of gp120. Similarly, N136_TAGS401_TAG contains two ambers inserted at positions 136 (Asn) in V1 and 401 (Ser) in V4, respectively. N136_TAGS413_TAG was tagged by two ambers introduced at positions 136 (Asn) in V1 and 413 (Ser) in V4, respectively. These plasmids were constructed from amber-free HIV-1_Q23 BG505 and HIV-1_{Q23 BG505} ΔRT for generating amber-click labeled Env in the context of intact virions. During the initial screening of amber-tolerance sites on Env_BG505 V1 and V4 loops, single amber- or dual amber-tagged Env were generated from wildtype amber-abundant HIV-1_{Q23 BG505}. We generated different versions of amber-free HIV-1_{Q23 BG505} clones, including ΔRT, ΔEnv (a new virus packing system with BG505 Env deletion), and a ΔRT ΔEnv. All the plasmids were sequenced before being used.

Viral infectivity measurements

Viral infectivity of HIV-1_{Q23 BG505} was determined by using a Gaussia luciferase reporter^83,84 and quantified using a Gaussia luciferase flash assay kit, according to the manufacturer’s instructions. Briefly, amber-abundant, or amber-free HIV-1Q23 BG505 viruses, or Env-tagged ones, were produced, respectively, in HEK293T cells that were co-transfected with the corresponding HIV-1_{Q23 BG505} plasmid, and intron-regulated luciferase plasmid (HIV-1-InGluc) using transfection reagent polyethyleneimine (PEI Max^™) or Fugene 6. Under amber suppression conditions, in addition to the HIV-1_Q23 BG505 and HIV-1-InGluc plasmids, the transfection solution also includes a tRNA^Pyl/NESPylRS^AF plasmid (1:3 HIV-1_{Q23 BG505} plasmid) supplemented with 250 μM unnatural amino acid (ncAA) transcyclooct-2-ene lysine (TCO*). The plasmid tRNA^Pyl/NESPylRS^AF is composed of a pyrrolysine tRNA and its cognate aminoacyl tRNA synthetase pair. PylRS^AF is the double mutant (Y306A, Y384F) of the M. mazei pyrrolysine tRNA synthetase, accommodating bulky side-chain moieties in ncAAs. tRNA^Pyl/NESPylRS^AF plasmid can be expressed in mammalian cells, such as HEK293T cells.^45,46 Viruses in the cell supernatants were harvested at 40 hour (h) post-transfection, filtered through 0.2 μm filters, and concentrated by ultracentrifugation at 25,000 rpm for 2 h. Viruses diluted in the medium were added onto TZM-bl cells pre-seeded in a 48-well plate. At 40 to 48 h post-infection, 100 μl cell supernatant was transferred to 96-well plates for the measurement of gaussian luciferase activity using a Gaussia luciferase flash assay kit. Relative viral infectivity was determined by normalizing relative light units (RLU) to the control group in each assay.

Immunoblotting

Viral particles of amber-abundant, amber-free HIV-1_{Q23 BG505}, and amber-free Env-tagged N136_TAG S401_TAG, N136_TAG S413_TAG, V1Q3 S401_TAG or N136_TAG V4A1 produced in HEK293T cells were verified by anti-gp120 western blotting under both suppression and non-suppression experimental conditions. Viral particles of amber-abundant and amber-free HIV-1_{Q23 BG505} were also analyzed using anti-Gag immunoblotting. Viral particles were produced similarly as that used in the infectivity assay, except lack of HIV-1-InGluc. For immunoblot assays, both the supernatants and cells were harvested after 40–48 h post-transfection of HEK293T cells. Virus supernatants were followed by ultracentrifugation of 40,000 rpm for 1.5 h, and virus pellets were resuspended in PBS buffer. Protein samples from virus supernatants and cell lysates were loaded on a 4–12% BisTris gel and transferred onto PVDF membrane by semi-dry blotting (Bio-Rad trans-blot turbo transfer system, 1.4 A constant and 20V for 7 min) with precision plus protein dual color standards PVDF membranes were blocked with 5% (w/v) skimmed milk powder in PBS for 1 h at room temperature and then incubated with primary antibodies of anti-HIV-1-gp120, and anti-HIV-1-p24⁸⁵, diluted in PBST (PBS with 0.05% Tween 20) buffer overnight at 4 °C. Membranes were followed by incubation with HRP-conjugated rabbit anti-sheep/-goat IgG secondary antibodies for 1 h at room temperature. Detection and quantitation of bound antibodies were conducted using clarity western ECL substrate and Bio-Rad ChemiDoc^™ XRS⁺ System with Image Lab^™ Software.

Neutralization assays

HEK293T cells were seeded and co-transfected with a mix of plasmids encoding either amber-abundant wildtype or amber-free HIV-1_{Q23 BG505}, amber-free N136_TAG S401_TAG, or amber-free N136_TAG S413_TAG, together with HIV-1-InGluc and tRNA^Pyl/NESPylRS^AF. 250 μM TCO* was added to the culture medium. At 40 h post-transfection, produced viruses in cell supernatants were collected, filtered, and concentrated by ultracentrifugation. Prior to adding onto the target TZM-bl cells, respective HIV-1 viruses as indicated in figures were incubated with sCD4_D1D2–lgαtp, PG9⁶⁰, PG16⁶⁰ or PGT151 at the 10-fold serial dilutions (10² to 10⁻² μg/mL) for 60 min at room temperature, as previously described³³. Viruses in the presence of different concentrations of ligands were then added on pre-seeded TZM-bl cells in the 48-well plate and were further incubated for another 40–48 h. Approximately two days later, cell supernatants were transferred to 96-well plates for the measurement of Gaussia luciferase activity, as described earlier in the viral infectivity measurements. Relative infectivity (mean ± SD) was normalized to the mean value determined from the control counterpart in each neutralization group.

2D cryo-electron microscopy

After 40 h post-transfection of 293T cells, 120 mL viral supernatants of amber-free HIV-1_{Q23 BG505}, amber-free N136_TAG S401_TAG, and amber-free N136_TAG S413_TAG were harvested, filtered and centrifugated on a 15% sucrose cushion by ultracentrifugation at 25,000 rpm for 2 h. Virus pellets were resuspended with 30 μL DPBS. Viral particles were mixed with a 6 nm gold tracer at a 1:3 ratio. 5 μL of the mixture was placed onto freshly glow discharged holey carbon grids (Quantifoil^™ R 2/1 on 200 copper mesh) and incubated for 1 min. Grids were blotted with filter paper and plunge-frozen into liquid ethane using a homemade gravity-driven plunger apparatus. Samples were screened with a Titan Krios G3 microscope (Titan Krios, Thermo Fisher Scientific) operating at 300 kV on a Gatan K3 camera in counting mode, at a magnification of 64,000×, with a calibrated pixel size of 1.346 Å.

Diameter size measurement of produced virus particles

HEK293T cells were seeded in 10 cm dishes with 12 mL complete growth medium. On the next day, cells with 80% confluency were transfected with 12 μg DNA plasmid and 1 mg/mL PEI Max^™ at a molar ratio of 1:3 (w: w). Plasmid DNA and PEI were separately diluted in Opti-MEM reduced serum medium (Thermal Fisher) for 5 min at room temperature (RT). PEI was then mixed well with plasmid DNA and incubated for 15 min at room temperature. The final mixture was added dropwise to the cells. For the amber suppression, HEK293T cells transfected with HIV-1 plasmid DNA, tRNA^Pyl/NESPylRS^AF and PEI at a ratio of 3: 1: 12 (w: w: w) were maintained in complete growth medium supplemented with a final concentration of 250 μM TCO*. The exchange of fresh complete growth medium (supplemented with 250 μM TCO* for amber suppression) was carried out at 6 h post-transfection. At 16 h post-transfection, an additional 250 μM TCO* was added to the amber suppression transfection systems. After 40 h post-transfection, the supernatants containing virions were harvested and centrifugated at a low speed of 200 g for 10 min. The supernatants were purified on a 15% sucrose cushion (w/v phosphate-buffered saline, PBS) by ultracentrifugation at 40,000 rpm for 1.5 h, using Bechman SW-41 Ti Swing Rotor. Virus pellets were resuspended with 200 μL borate buffered saline (BBS) buffer. The diameter size distribution of virus particles diluted in borate buffered saline (BBS) buffer was measured by Nanoparticle Tracking Analysis (NTA) using NanoSight instrument. Three measurements were carried out with 30 seconds duration for each measurement and the data were analyzed by NTA 3.3 Dev Build 3.3.301. Detected tracks were translated into a size distribution using maximum likelihood estimation with an assumed distribution (FTLA method).

Preparation of intact HIV-1 virions for smFRET imaging

HIV-1_{Q23 BG505} virions incorporated with tagged Env_BG505 (amber-tagged, amber/peptide-tagged, or peptide-tagged) in V1 and V4 loops were prepared in a similar way as previously described.³³ HIV-1 virions used for smFRET imaging were prepared as replication-incompetent viral particles that lack reverse transcriptase (ΔRT), meaning that all the full-length HIV-1Q23 BG505 plasmids used for transfection are ΔRT versions. In this study, we generated three different types of viruses with peptide-, hybrid amber/peptide-, or amber-tagged Env: 1) dual-peptide V1Q3 V4A1; 2) hybrid click/peptide V1Q3 S401_TAG and N136_TAG V4A1 in the amber-free virus context; 3) dual-amber N136_TAG S401_TAG and N136_TAG S413_TAG in the amber-free virus context. V1Q3 V4A1 Env carries a Q3 tag in V1 and an A1 tag in V4. As previously described,^33,43 V1Q3 V4A1 virions were produced by co-transfecting HEK293T cells with a 40:1 plasmid: plasmid ratio of wild-type tag-free Env virion: V1Q3 V4A1, to ensure that on average only one dually peptide-tagged protomer within a trimer was available for labeling on a single virion. The plasmid: plasmid ratio of tag-free: tagged Env virions was adjusted according to the reduction in amber suppression efficiency of amber-tagged or amber/peptide-tagged Env. N136_TAG S401_TAG, N136_TAG S413_TAG, V1Q3 S401_TAG, N136_TAG V4A1 on the amber-free virions were made similarly, by co-transfecting HEK293T cells with a 4:1 ratio of amber-free tag-free Env to Env-tagged plasmids, along with an amber suppressor plasmid supplemented with unnatural amino acids. For instance, amber-free N136_TAG S413_TAG viruses, carrying two clickable sites at N136_TAG in V1 and S413_TAG in V4, were produced by co-transfecting HEK293T cells with a tag-free ΔRT plasmid, an Env-tagged variant N136_TAG S413_TAG ΔRT plasmid, and a bi-cistronic plasmid tRNA^Pyl/NESPylRS^AF encoding the amber suppressor tRNA and PylRS^AF. The amount of tRNA^Pyl/NESPylRS^AF used is 1/3 of the total amount of plasmids that encode Env (including untagged and tagged). The unnatural amino acid TCO* (250 μM) was supplemented to transfected HEK293T cells during the transfection and re-supplemented 12-h post-transfection.

Fluorescent labeling Env on intact HIV-1 virions

After 40 h post-transfection, viruses in the supernatants were filtered, harvested, and concentrated on the 15% sucrose cushion by ultracentrifugation at 25,000 rpm (SW28, Beckman Coulter) for 2 h. Virus pellets were resuspended in the labeling buffer containing 50 mM HEPES, 10 mM MgCl₂ and 10 mM CaCl₂.^32,33,43 Fluorescently labeling, including enzymatic labeling and click labeling, of virus-associated Env has been previously described^32,33,43. For enzymatic labeling, the resuspended V1Q3 V4A1 viruses in labeling buffer were supplied with cadaverine-conjugated Cy3 (LD555-CD, 0.5 μM) and coenzyme A (CoA)-conjugated Cy5 (LD655-CoA, 0.65 μM)⁴¹, and AcpS (5 μM)⁴⁰, and then incubated overnight at room temperature. For click labeling, resuspended amber-free N136_TAG S401_TAG or N136_TAG S413_TAG viruses that carry two click-chemistry-reactive TCO* were fluorescently labeled in a reaction mixture of 0.1 μM tetrazine-conjugated Cy3 and Cy5 derivatives (LD555-TTZ and LD655-TTZ). For enzymatic/click labeling, resuspended amber-free N136_TAG V4A1 or V1Q3 S401_TAG viruses that carry one peptide and one clickable TCO* were fluorescently labeled by the corresponding enzymatically conjugated dye along with the specific enzyme and the FRET-paired click dye. For the click-labeled or enzymatic/click-labeled viruses, a quenching step of 10 min incubation with 1μM BCN-OH quencher followed the labeling step. PEG2000-biotin was subsequently added to the labeling reaction at a final concentration of 0.1 mg/ml and followed by incubation at room temperature for 30 min with rotation. Excessive label and lipid were removed by ultracentrifugation over a 6–18% Optiprep gradient at 40,000 rpm for 1 h. The labeling virions were stored at −80 °C until further study.

smFRET imaging data acquisition

All smFRET experiments of Env on intact HIV-1 virions were performed on a customized prism-based total internal reflection fluorescence (prism-TIRF) microscope, as previously described.^33,62,68,86 Fluorescently labeled HIV-1 viruses were incubated in the absence or presence of 0.1 mg/ml sCD4_D1D2–lgαtp or 0.1 mg/ml PGT151 in the imaging buffer containing 50 mM Tris (pH 7.4), 50 mM NaCl, a cocktail of triplet-state quenchers, 2 mM protocatechuic acid (PCA), and 8 nM protocatechuic-3,4-dioxygenase (PCD) at room temperature for 30 min before smFRET imaging, as previously described.^32,33,43,87 The ligand and antibody concentrations were approximately 5-fold above the 95% inhibitory concentration. HIV-1 viruses carrying fluorescently labeled Env were immobilized on a PEG-passivated biotinylated quartz-coverslip imaging chamber coated with streptavidin. Based on the refractive indexes at the interface between quartz and sample solution, the evanescent field was generated by the total internal reflection of 532-nm single-wavelength laser excitation (Ventus, Laser Quantum) directed to a prism. The donor fluorophore labeled on Env on intact HIV-1 virions is further directly excited by the generated TIRF field. Fluorescence from both donor and acceptor fluorophores labeled on Env was collected through a water-immersion Nikon objective (1.27-NA 60x) and then optically separated by a MultiCam LS image splitter (Cairn Research) with a dichroic filter (Chroma). The separated fluorescence signals from the donor and acceptor passed through two emission filters (ET590/50, ET690/50, Chroma) mounted on the image splitter, respectively. Fluorescence signals were recorded simultaneously on two synchronized sCMOS cameras (Hamamatsu ORCA-Flash4.0 V3) at 25 Hz for 80 seconds. Where indicated, viruses were pre-incubated with the corresponding antibody/ligand for 30 mins at room temperature before imaging and that antibody/ligand was continuously present during smFRET imaging.

smFRET data analysis

smFRET data were viewed, processed, and analyzed by a customized SPARTAN software package that was generously shared by Scott Blanchard laboratory (https://www.scottcblanchardlab.com/software)⁸⁶ and customized MATLAB-based scripts. Image stacks of smFRET data were extracted as individual fluorescence time series trajectories (fluorescence traces) of donor and acceptor labeled on HIV-1 virions. At the single-molecule level, the background signal was evaluated and then subtracted from recorded signals according to the level of signals at the single-step fluorophore bleaching points. Fluorescence traces of the donor and acceptor over 80 seconds were then corrected for the donor to acceptor crosstalk. The time-correlated donor-to-acceptor energy transfer efficiency (FRET values or FRET in graphs) was determined based on FRET= I_A/(γI_D+I_A). I_D and I_A are the fluorescence of the donor and acceptor, respectively. The correlation coefficient γ compensates for detection efficiencies variations between donor and acceptor channels, as the response of optics and cameras to different wavelengths varies slightly. FRET traces, real-time FRET values over time, were further derived from corresponding fluorescence traces of paired donor and acceptor. FRET traces indicate the relative distance changes between donor and acceptor over time, which are ultimately translated to real-time global conformational dynamics of the donor/acceptor labeled Env in the context of an intact virion. We use stringent criteria to filter out noisy signals. Fluorescence or its corresponding FRET Traces were automatically excluded from further analysis if it lacked signals from either donor or acceptor or both, and if it contained multiple labeled donors/ acceptors. After the initial automatic filtering, the remaining traces with sufficient signal-to-noise ratio need to display anti-correlation between donor and acceptor fluorescence before being manually included for further data analysis. The time-correlated anti-correlated feature is a signature of donor-to-acceptor energy transfer in response to real-time conformational changes of the host molecule Env, thus is an indicator of fully active/dynamic Env on the virion. The use of automatic and manual filters ensures that only one dually FRET-labeled protomer within one Env trimer on one virion is included for further analysis. FRET traces that pass our filters were included and compiled into FRET histograms/distributions, reflecting ensembles of multiple conformational states – conformational ensembles. FRET histograms - conformational ensembles represent mean ± SEM, determined from three randomly assigned populations of FRET traces under indicated experimental conditions. FRET histograms were further fitted into the sum of three Gaussian distributions using the least-squares fitting algorithm in MATLAB, based on visual inspection of traces that show clear state-to-state transitions and the idealization of individual traces using 3-state hidden Markov modeling.^32,33,43,88 Each Gaussian distribution represents one distinct conformational state of Env on the virus. The area under each Gaussian estimates the relative probability of Env occupying that conformational state, displayed as relative state occupancy.

Calculation of R₀ values

The Förster distance, $R_0$, is calculated by the following equation:⁸⁹

$$R_0={(\frac{9\mathrm{l}\mathrm{n}\left(10\right)}{128{\pi }^5{N}_A}\mathrm{*}\frac{k^2{\mathrm{\Phi }}_D}{n^4}\mathrm{*}{10}^{17}\mathrm{*}{J}_{\mathrm{\lambda }})}^{\frac{1}{6}}(\text{in}\text{Å})$$

where $n$ is the refractive index of the medium, $N_A$ is Avogadro’s constant, ${\mathrm{\Phi }}_D$ is the quantum yield of the donor in the absence of the acceptor, $k^2$ is the orientation factor, $J_{\mathrm{\lambda }}$ is the spectral overlap integral. For biomolecules in water, the medium used in these experiments, $n$ is assumed to be 1.33⁹⁰, within the dynamic isotropic averaging assumption $k^2$ is 2/3 (freely rotated dyes with flexible linkers)⁹¹. $J_{\mathrm{\lambda }}-\mathrm{J}\left(\lambda \right)$ is the spectral overlap integral between the steady-state donor emission and the steady-state acceptor absorption. $\mathrm{J}\left(\lambda \right)$ can itself be defined using the equation: ⁸⁹

$$\mathrm{J}\left(\lambda \right)={\int }_0^{\mathrm{\infty }} {\mathrm{f}}_{\mathrm{D}}\left(\lambda \right){\mathrm{\epsilon }}_{\mathrm{A}}\left(\lambda \right){\lambda }^{4}d\lambda$$

where ${\mathrm{f}}_{\mathrm{D}}\left(\lambda \right)$ is the normalized emission spectrum of the donor and ${\mathrm{\epsilon }}_{\mathrm{A}}\left(\lambda \right)$ is the extinction of the acceptor at wavelength $\lambda$. To integrate $\mathrm{J}\left(\lambda \right)$, the trapezoid rule was used. The steady-state emission spectra of the donor and the excitation spectra of the acceptor were measured on a Horiba FluoroMax spectrofluorometer (model FluoroMax Plus-C-SP). No spectra shifts were observed for free dyes vs. dyes labeled on viruses. $R_0$ values were similar for different donor/acceptor pairs used in our study, in which $R_0$ values were calculated to be 60.0 Å, 61.0 Å, 61.0 Å, and 60.0 Å for donor/acceptor-paired LD555-CD/LD655-CoA, LD555-tetrazine/LD655-tetrazine, LD555-CD/LD655-tetrazine, and LD555-tetrazine/LD655-CoA, respectively.

Live cell imaging of click-labeled virions

Amber-free HIV-1_{Q23 BG505} N136_TAG -TCO* viral particles made using 100% mutant Env construct were click-labeled by incubating for 6 hours at room temperature with 0.1 μM LD655-TTZ (Cy5-TTZ). Unreacted dye was removed by washing with PBS buffer, followed by centrifugation (25,000 rpm, 2 hr). Trafficking of amber-free HIV-1_{Q23 BG505} N136_TAG virions labeled with Cy5-TTZ in HeLa-derived TZM-bl cells ^92–94 was visualized with live cell confocal microscopy. Briefly, 50,000 TZM-bl cells were plated onto collagen-coated imaging dishes with #1.5 cover glass inserts (Matek, CAT# P35G-1.5–10-C). Labeled virions were prebound to TZM-bl cells via centrifugation for 20 min at 4°C at 1,550×g in DPBS with Ca²⁺ and Mg²⁺. After virus-cell binding, cells were washed with cold DPBS, and transferred to pre-warmed Fluorobrite DMEM medium supplemented with 10% fetal bovine serum and 20 mM HEPES, pH 7.2. The imaging medium contained the SPY555-Actin (Cytoskeleton, CAT# CY-SC202) live cell actin probe, and imaging was performed in the continued presence of the probe to visualize the actin cortex, as per the manufacturer’s instructions. 3D time-lapse live cell imaging was carried out on a Zeiss LSM880 laser scanning confocal microscope equipped with a humidity and temperature-controlled environmental chamber and using a C-Apo 63x/1.4 NA oil immersion objective. All live cell imaging experiments were performed using the highly attenuated 561 nm (for SPY555) & 633 nm (for Cy5-TTZ) laser lines, and the emitted light was detected by gallium-arsenide spectral detectors with 0.22 × 0.22 μm pixel size. Live cell micrographs were collected using 10–12 Z-sections (0.7 μm spacing), with each frame being imaged every 15–17 seconds.

QUANTIFICATION AND STATISTICAL ANALYSIS

Quantification and statistical analysis are detailed in the figure legends and the STAR Methods. As indicated, analysis was performed using GraphPad Prism 9, MATLAB, or EXCEL software. Data (mean ± SD or SEM) were derived from at least three independent observations.

Supplementary Material

1

Movie 1. Internalization of an amber-free click-labeled HIV-1_{Q23 BG505} Env N136_TAG-Cy5 virion into a live cell, related to Figure 6.

Representative movie of a single Env N136_TAG-Cy5 (green) virion entering into a TZM-bl cell labeled with SPY555-Actin (magenta). The virion initially binds to the cell periphery and later undergoes internalization into the peripheral actin cortex, where high degrees of mobility and displacement are observed within the cell.

Highlights.

• We constructed an intact amber-free HIV-1 provirus for in-virus click labeling

• System enables dual-ncAA insertion for biorthogonal click labeling on intact virions

• smFRET of Env reinforced its structural dynamics observed via conventional labeling

• Live cell imaging tracked amber-free click-labeled Env virions in cells

SIGNIFICANCE.

The minimally invasive bioorthogonal click labeling approach based on genetic code expansion combined with click chemistry has been rarely applied to fluorescently labeling HIV-1 proteins partially due to intrinsic technical limitations. Conformational dynamics of sparsely distributed envelope (Env) glycoprotein on HIV-1 mediate cell entry and facilitate antibody evasion. Dually fluorescent labeling Env is a prerequisite for enabling a unique “spectroscopic ruler” or “molecular ruler” approach to access spatiotemporal information on conformations sampled by Env on the virus, so least-invasive click labeling has been highly desirable. Here, we overcame a technical bottleneck of genetic code expansion application for HIV-1 protein labeling by establishing new full-length clinically relevant HIV-1 strains, which enabled, for the first time, singly and dually click labeling of Env in the context of virions. More specifically, we successfully incorporated two adjacent clickable unnatural amino acids into Env on intact HIV-1 virions. Using click-labeled amber-free virions, real-time single-molecule/virus spectroscopic imaging of Env on virions consolidated the presence of multiple conformations as an intrinsic property of Env dynamics that is not due to labeling artifacts and tracked the internalization route of HIV-1 into host cells. The amber-free system could open new opportunities for click labeling of functional proteins in HIV-1 and of surface glycoproteins of other viruses.