Skip to main content
PMC home page
eLife logoLink to eLife
. 2018 Mar 21;7:e34271. doi: 10.7554/eLife.34271

HIV-1 Env trimer opens through an asymmetric intermediate in which individual protomers adopt distinct conformations

Xiaochu Ma 1, Maolin Lu 1, Jason Gorman 2, Daniel S Terry 3, Xinyu Hong 1, Zhou Zhou 3, Hong Zhao 3, Roger B Altman 3, James Arthos 4, Scott C Blanchard 3, Peter D Kwong 2, James B Munro 5,✉, Walther Mothes 1,✉
Editor: Pamela J Bjorkman6
PMCID: PMC5896952  PMID: 29561264

Abstract

HIV-1 entry into cells requires binding of the viral envelope glycoprotein (Env) to receptor CD4 and coreceptor. Imaging of individual Env molecules on native virions shows Env trimers to be dynamic, spontaneously transitioning between three distinct well-populated conformational states: a pre-triggered Env (State 1), a default intermediate (State 2) and a three-CD4-bound conformation (State 3), which can be stabilized by binding of CD4 and coreceptor-surrogate antibody 17b. Here, using single-molecule Fluorescence Resonance Energy Transfer (smFRET), we show the default intermediate configuration to be asymmetric, with individual protomers adopting distinct conformations. During entry, this asymmetric intermediate forms when a single CD4 molecule engages the trimer. The trimer can then transition to State 3 by binding additional CD4 molecules and coreceptor.

Research organism: Virus

Introduction

The human immunodeficiency virus 1 (HIV-1) enters CD4+ T cells through the interaction of its envelope glycoprotein (Env) with cell-surface receptor CD4 and coreceptor CCR5 or CXCR4 (Wyatt and Sodroski, 1998; Doms and Moore, 2000). Env is a trimer with each protomer consisting of two subunits: the surface subunit gp120 that binds to receptor and coreceptor, and the transmembrane subunit gp41 that mediates fusion of viral and cellular membranes. CD4 binding induces conformational changes in Env that expose structural elements required for coreceptor binding including the V3 loop and the bridging sheet (Trkola et al., 1996; Kwong et al., 1998; Huang et al., 2007; Liu et al., 2008; Pancera et al., 2010; Wang et al., 2016; Herschhorn et al., 2017; Ozorowski et al., 2017). Binding of coreceptor triggers additional conformational changes in gp41, including the formation of an extended gp41 structure that subsequently collapses into a stable six-helix bundle, which is thought to drive viral and cellular membranes together for fusion (Pancera et al., 2010; Blumenthal et al., 2012; Harrison, 2015). Correspondingly, Env’s capacity to undergo extensive conformational changes is critical for virus entry.

At the same time, Env evades immune surveillance through ‘conformational masking’ (Kwong et al., 2002), which protects key functional elements within the trimer from being recognized by antibodies. This feature renders most Env-targeting antibodies non-neutralizing. However, a portion of patients develop potent broadly neutralizing antibodies that can prevent immunodeficiency virus infections in animal models, lower the viral load when administered to HIV-1-infected patients and can restore immunological control in the absence of antiretroviral therapy (ART) in non-human primates (Wu et al., 2010; Walker et al., 2011; Klein et al., 2012; Caskey et al., 2015; Gautam et al., 2016; Lu et al., 2016; Schoofs et al., 2016; Nishimura et al., 2017). Antibodies that are broadly neutralizing tend to recognize closed Env conformations (Munro et al., 2014; Guttman et al., 2015). Substantial efforts have been made to characterize structurally closed Env trimers using stabilized soluble ectodomains as well as detergent-solubilized Env proteins (Julien et al., 2013; Lyumkis et al., 2013; Pancera et al., 2014; Lee et al., 2016). However, how the Env trimer opens, and through what structural intermediates it transitions, is poorly understood.

We have previously applied single-molecule Fluorescence Resonance Energy Transfer (smFRET) imaging to visualize the dynamics of individual Env molecules on the surface of native virions of two HIV-1 strains, NL4-3 and JR-FL (Munro et al., 2014). Donor and acceptor fluorophores were introduced into the variable loops V1, and V4 or V5 of a single gp120 subunit in an otherwise unlabeled virus. smFRET analysis revealed that single Env protomers spontaneously transit between three distinct conformational states exhibiting low-, intermediate- and high-FRET values (Munro et al., 2014). The unliganded Env prefers a low-FRET pre-triggered conformation. In the presence of soluble CD4 (D1D2 domain, sCD4), and the additional presence of the coreceptor-surrogate antibody 17b, some Env trimers could be stabilized in high-FRET and intermediate-FRET conformations, respectively (Munro et al., 2014).

The FRET-indicated Env conformational states initially observed require further structural assignments. Various lines of evidence suggest that the low-FRET state corresponds to the pre-triggered conformation: (1) it is the most populated conformation of the unliganded Env; (2) it is more populated in a clinical isolate such as HIV-1JR-FL that is more neutralization-resistant than the laboratory-adapted HIV-1NL4-3; and (3) it is stabilized by broadly neutralizing antibodies, and the small-molecule conformational blocker BMS-626529 (Munro et al., 2014; Pancera et al., 2014; Kwon et al., 2015; Herschhorn et al., 2017; Pancera et al., 2017). In contrast, the structural nature of the intermediate- and high-FRET states has been unclear. The stabilization of the high-FRET State two in HIV-1NL4-3 by sCD4 and of intermediate-FRET State three by sCD4/17b suggested that they might represent CD4 or coreceptor-bound conformations, respectively. However, this conflicts with the finding that the CD4 mimetic JRC-II-191 stabilizes intermediate-FRET configurations (State 3) (Munro et al., 2014), as we would expect CD4 mimetics to reproduce the conformational impact of CD4. Also, there are no substantial Env-structural differences between the gp120 bound to CD4 or to CD4 and 17b (Kwong et al., 1998; Ozorowski et al., 2017). Finally, HIV-1JR-FL does not respond to sCD4 in the same way as HIV-1NL4-3.

Here we provide smFRET analysis that clarify the nature of the intermediate-FRET and high-FRET Env conformations. Through smFRET measurements, we demonstrated that dodecameric CD4 (sCD4D1D2-Igαtp), which is a potent neutralizer of HIV-1, stabilizes the intermediate-FRET State 3 of the Env in three strains HIV-1NL4-3, HIV-1JR-FL and HIV-1BG505. We observed no difference in predominant conformational state of Env bound to CD4/17b, to JRC-II-191 (a CD4 small molecule mimetic), or to sCD4D1D2-Igαtp, all of which corresponded to the open three-CD4-bound conformation of the trimer (State 3). Moreover, by using a mixed trimer assay, we identified the intermediate configuration of State two to be an asymmetric Env trimer. In this configuration, a single CD4 can engage the trimer such that individual protomers adopt distinct conformations. An asymmetric trimer is thus a prevalent functional intermediate, which can open further either spontaneously or through binding to additional CD4 molecules or to coreceptor.

Results

State 3 corresponds to the gp120 conformation of the three-CD4-bound HIV-1 Env trimer

smFRET imaging of HIV-1 Env trimers carrying a single pair of donor and acceptor fluorophores in the V1 and V4 loops of gp120, previously established for HIV-1 isolates NL4-3 and JR-FL (Munro et al., 2014), was extended to include the mother-to-child transmitted founder (T/F) virus BG505 widely used for structural studies (Figure 1—figure supplement 1) (Julien et al., 2013; Lyumkis et al., 2013; Sanders et al., 2013; Pancera et al., 2014; Kwon et al., 2015; Scharf et al., 2015; Wang et al., 2016; Ozorowski et al., 2017). The insertion of Q3 and A1 tags into the variable loops V1 and V4 of gp120 for enzymatic labeling was validated by infectivity assays, protein expression and antibody neutralizations, indicating that the Env functions remained minimally affected even in the 100% dually tagged virus (Figure 1—figure supplements 2 and 3).

Surface-bound HIV-1NL4-3, HIV-1JR-FL and HIV-1BG505 viruses carrying a single dually-tagged Env were imaged by total internal reflection fluorescence (TIRF) microscopy. In agreement with previous results (Munro et al., 2014), FRET trajectories and histograms showed that the gp120 protomers in unliganded HIV-1 Env predominantly resided in a low-FRET conformation, but had inherent access to both intermediate- and high-FRET conformations (Figure 1A and B, left). Consistent with the features of Tier two viruses, the HIV-1 isolates JR-FL and BG505 (Koyanagi et al., 1987; Wu et al., 2006) exhibited a higher occupancy of the pre-triggered low-FRET state as compared to the lab-adapted NL4-3 (compare Figures 1E and H, B).

Figure 1. State 3 corresponds to the gp120 conformation of the three-CD4-bound HIV-1 Env trimer.

(A) Representative smFRET trace for unliganded HIV-1NL4-3 Env. (Top) The donor fluorophore (green) was attached to the V1 loop and the acceptor fluorophore (red) was attached to the V4 loop. (Bottom) Corresponding FRET trajectory (blue) with overlaid idealization generated by Hidden Markov Modeling (HMM) (red). (B) (Left) Probability distribution of FRET values compiled from all the individual HIV-1NL4-3 Env molecules (N = number of FRET traces analyzed). The histogram was fitted to a sum of three Gaussian distributions, with means of 0.13, 0.3 and 0.63, which are corresponding to States 1, 3 and 2, as indicated. The percentage indicates the occupancy of each FRET state. Error bars represent standard errors calculated from histograms from three independent sets of FRET traces. (Right) Transition Density Plot (TDP) of all the observed transitions in unliganded HIV-1NL4-3 Env. Color bar shows the scale used to indicate the frequency of each transition. (C, D) sCD4 (0.1 mg/ml) (C) or sCD4D1D2-Igαtp (0.1 mg/ml) (D) was incubated with the virus for 30 min prior to imaging. FRET histogram and TDP are as in (B). (E–G) Probability distributions (left) and TDPs (right) for HIV-1JR-FL Env for the unliganded (E) (Note: FRET histogram and TDP were from previous data set for direct comparison [Herschhorn et al., 2016]), sCD4-bound (0.1 mg/ml) (F) and sCD4D1D2-Igαtp-bound (0.1 mg/ml) (G). (H–J) Probability distributions (left) and TDPs (right) for HIV-1BG505 Env for the unliganded (H), sCD4-bound (0.1 mg/ml) (I) and sCD4D1D2-Igαtp-bound (0.1 mg/ml) (J) viruses are displayed as in HIV-1NL4-3. (K) Schematic illustration of the closed and open conformations of the Env trimer. The unliganded conformation is in blue and CD4-bound conformation is in pink. Green and red starts represent donor and acceptor fluorophores, respectively. Sizes of the stars represent relative change of fluorescence between donor and acceptor dyes and dotted line indicated changes of inter-dye distances.

Figure 1.

Figure 1—figure supplement 1. Peptide insertion sites into the V1 and V4 loops of gp120 of three HIV-1 isolates.

Figure 1—figure supplement 1.

Q3 and A1 peptides were inserted into the V1 and V4 loops of gp120 in HIV-1NL4-3, HIV-1JR-FL, HIV-1BG505, respectively (Munro et al., 2014)(HIV-1BG505, this report).
Figure 1—figure supplement 2. Infectivity and Env incorporation of single or dually tagged HIV-1BG505 viruses.

Figure 1—figure supplement 2.

Q3 (GQQQLG) or A1 (GDSLDMLLEWSLM) tags were inserted separately or together into the V1 and V4 loops of full-length Q23_BG505 virus. (A) Infectivity was measured from three independent experiments by Gaussia Luciferase assay and normalized to WT (%). (B) Env incorporation into virions was detected by Western blotting using an antiserum to HIV-1 gp120 (NIH AIDS Reagent Program) well as an anti-p24 antibody.
Figure 1—figure supplement 3. Dually tagged HIV-1BG505 antibodies are neutralized by trimer specific antibodies.

Figure 1—figure supplement 3.

Neutralization of HIV-1 BG505_WT (black) and dually labeled BG505_V1Q3_V4A1 (purple) by the broadly neutralizing antibodies (A) PG9 and (B) PG16. X-axis depicts increasing concentration of antibodies (μg/ml) and the y-axis shows the relative infectivity compared to control in the absence of ligands. Infectivity was measured using the Gaussia Luciferase assay.
Figure 1—figure supplement 4. smFRET histogram for HIV-1BG505 Env bound to sCD4 and 17b.

Figure 1—figure supplement 4.

HIV-1BG505 Env carrying virions were incubated with sCD4 and 17b (0.1 mg/ml each) at room temperature for 30 min prior to TIR-FM imaging. FRET trajectories were compiled into histogram and fitted into 3-state Gaussian curves.
Open in a new tab

We applied Hidden Markov Modeling (HMM) using a three-state model to analyze the sequence of transitions between the conformational states and displayed them in a transition density plot (TDP) (Figure 1B, right). The unliganded HIV-1NL4-3 Env frequently transitioned between the low-FRET and high-FRET states, and between the high- and intermediate-FRET states; whereas transitions between low- and intermediate-FRET were rarely observed. Given that the low-FRET state corresponds to a mature pre-triggered conformation of the trimer, these data suggest that the high-FRET state represents a default conformational intermediate during opening of the Env trimer. Based on this sequence of FRET transitions, we refer to the low-FRET state as State 1, the high-FRET state as State two and the intermediate-FRET state as State 3 (Herschhorn et al., 2016).

Consistent with previous observations, addition of sCD4 (0.1 mg/ml) stabilized HIV-1NL4-3 Env in State 2 (Figure 1C, left). The predominant FRET transitions observed were between States 2 and 3. The observed stabilization of State two by sCD4 was initially difficult to reconcile with the observed State three stabilization by the CD4-mimetic small molecule, JRC-II-191, as well as the combination of sCD4 and the coreceptor-surrogate antibody 17b (Munro et al., 2014). To clarify the response of HIV-1NL4-3 Env to receptor CD4, a dodecameric CD4 oligomer (sCD4D1D2-Igαtp, 0.1 mg/ml), which is 300 to 1000-fold more potent in neutralizing HIV-1 than sCD4, was used to provide a high local CD4 density (Arthos et al., 2002; Kwong et al., 2002). Strikingly, sCD4D1D2-Igαtp stabilized HIV-1NL4-3 Env in State 3 (Figure 1D), indicating that both sCD4/17b (Munro et al., 2014) and an oligomerized CD4, both stabilize HIV-1NL4-3 Env in the three-CD4-bound trimer configuration (State 3).

In contrast to the NL4-3 isolate, both HIV-1JR-FL and HIV-1BG505 were less responsive to sCD4, with only small increases in the occupancies of States 2 and 3 (Figure 1F and I) although sCD4 and 17b stabilized both in State 3 (Munro et al., 2014) (Figure 1—figure supplement 4). Importantly, as was observed for HIV-1NL4-3 Env, addition of the dodecameric sCD4D1D2-Igαtp stabilized both HIV-1JR-FL and HIV-1BG505 in State 3 (Figure 1G and J). This suggests that for all three HIV-1 isolates the intermediate-FRET State 3 corresponds to the three-CD4-bound conformation (Figure 1K).

State 2 arises from an asymmetric trimer, in which a single CD4 molecule engages HIV-1 Env

The assignment of the three-CD4-bound conformation as State 3 leaves the nature of the structural intermediate State 2 unresolved. We used the stabilization of HIV-1NL4-3 State 2 by sCD4 as an assay to identify the origin of the high-FRET signal. We hypothesized that the different conformational effects of sCD4 and sCD4D1D2-Igαtp on HIV-1NL4-3 Env are linked to the differences in CD4 occupancy. The dodecameric sCD4D1D2-Igαtp, with its higher avidity and local CD4 density, may engage two or three protomers of the trimer (Arthos et al., 2002; Bennett et al., 2007) while monomeric sCD4 binds with less avidity and may only interact with one protomer. The distinct high-FRET conformation of State 2 may then arise from either a distinct conformation in the first CD4-bound protomer, or from neighboring gp120 subunits.

To distinguish between these possibilities, we took advantage of the D368R mutation in the CD4-binding pocket of Env that reduces affinity for CD4 (Olshevsky et al., 1990). The D368R mutant of HIV-1NL4-3 Env was resistant to the neutralization of both sCD4 and sCD4D1D2-Igαtp (Figure 2A, Table 2). The mutant Env was expressed similarly to that of WT Env (Figure 2—figure supplement 1). Their infectivity was reduced by ~100 fold, but the signal to noise ratio in our experiments was orders of magnitude above background permitting a neutralization assay and demonstrating the resistance of D368R to sCD4 (Figure 2—figure supplement 2). Trimers that uniformly contained the D368R mutation exhibited similar conformational landscapes to that of wild-type HIV-1, and did not respond to sCD4 and sCD4D1D2-Igαtp as WT HIV-1NL4-3 (Figure 2—figure supplement 3A). Thus, unlike the double mutant D368R/E370R that was stabilized in State 1 (Munro et al., 2014) (likely due to E370R interfering with the adoption of the bridging sheet), the single D368R mutation exhibited little or no conformational effects on the trimer. This allowed us to engineer a ‘mixed trimer 1’ where the two unlabeled protomers carried the D368R mutation to prevent binding of CD4 and the single protomer carrying the fluorophores remained competent for CD4 binding (Figure 2B). smFRET analysis for the mixed HIV-1NL4-3 trimer 1 revealed a clear State 3 stabilization by sCD4 (Figure 2D). The occupancy of State 2 was similar to the native HIV-1NL4-3 trimer bound by sCD4D1D2-Igαtp (for comparison, see Figure 1D). Thus, gp120 bound to CD4 adopts a State 3 conformation regardless of the CD4 occupancy elsewhere within the trimer, even when only a single CD4 molecule engages the trimer.

Figure 2. State 2 corresponds to an asymmetric trimer, in which a single CD4 molecule engages HIV-1 Env.

(A) Neutralization curves of WT and D368R HIVNL4-3 viruses by sCD4 and sCD4D1D2-Igαtp. Data represent three independent experiments ± standard deviation. (B) Scheme to illustrate generation of mixed HIV-1 Env trimer 1, in which the two unlabeled protomers contained the D368R mutation to prevent CD4 binding, and the CD4-binding competent WT protomer carried the donor and acceptor fluorophores (green, red stars in scheme above). (C) Scheme to illustrate generation of mixed HIV-1 trimer 2, in which sCD4 can only engage gp120 domains adjacent to the labeled domain. Given the co-transfection protocol of indicated HIV-1 plasmids, only 50% of all trimers are expected to exhibit this configuration. 25% of trimers are expected to carry D368R mutation in all three protomers and the remaining 25% would carry two CD4-binding competent protomers next to the labeled mutant gp120. (D) FRET histogram as in Figure 1 for the mixed HIV-1NL4-3 Env trimer 1. (E) FRET histogram for the mixed HIV-1NL4-3 Env trimer 2. Neutralization curves for HIV-1JR-FL (F) or HIV-1BG505 (I) were shown as in HIV-1NL4-3. FRET histograms for mixed HIV-1JR-FL (G–H) and HIV-1BG505 (J–K) Env trimer 1 and 2 are shown as HIV-1NL4-3. Viruses were incubated with sCD4 (0.1 mg/ml) or sCD4D1D2-Igαtp (0.01 mg/ml) for 30 min prior to imaging as indicated. (L) Schematic illustration of the asymmetric opening of the Env trimer. The CD4-bound conformation is in pink, and the conformational intermediate in the asymmetric trimer is in purple. The purple x indicates the D368R mutation. Green and red stars represent donor and acceptor fluorophores, respectively. Sizes of the stars represent relative change of fluorescence between donor and acceptor dyes and dotted line indicated changes of inter-dye distances.

Figure 2.

Figure 2—figure supplement 1. D368R carrying Envs are expressed and incorporated into virions similar to wild-type.

Figure 2—figure supplement 1.

HIV-1 containing WT or D368R Env were produced in HEK293 cells and viral supernatants collected 40 hr post-transfection. Virus was concentrated by centrifugation at 20,000 g and analyzed by SDS-PAGE and Western blot analyses using the anti-Env antibody 2G12 or patient serum (NIH AIDS Reagent Program) to detect Env of HIV-1NL4-3 and HIV-1JR-FL w, or HIV-1BG505, respectively, and anti-p24 to detect capsid.
Figure 2—figure supplement 2. Infectivity of HIV-1D368R viruses.

Figure 2—figure supplement 2.

Infectivity of WT and D368R containing HIV-1 isolates was determined from three independent measurements with ±standard deviation (SD); presented normalized to WT infectivity (%).
Figure 2—figure supplement 3. D368R inhibits sCD4 and sCD4D1D2-Igαtp binding of all three HIV-1 Envs.

Figure 2—figure supplement 3.

(A) FRET histograms as in Figure 1 for the unliganded (left), sCD4-bound (middle) and sCD4D1D2-Igαtp-bound (right) HIV-1NL4-3 Envs carrying the D368R mutation in all protomers. (B) FRET histograms as in (A) for the HIV-1JR-FL Env carrying the D368R mutation in all protomers. (C) FRET histograms as in (A) for the HIV-1BG505 Env carrying the D368R mutation in all protomers. sCD4 (0.1 mg/ml) and sCD4D1D2-Igαtp (0.01 mg/ml) were incubated for 30 min prior to imaging.
Open in a new tab

We next examined the conformation of the CD4-binding incompetent protomers (carrying a D368R mutation) next to the single-CD4-bound protomer, which we designated ‘mixed trimer 2’. To this end, HEK293 cells were co-transfected with a 1:1 ratio of plasmid encoding HIV-1 D368R mutant and wild-type HIV-1NL4-3 Env in excess over plasmid encoding HIV-1D368R with dually tagged Env. As a result,~50% of the trimers were expected to carry fluorophores in the CD4-binding incompetent D368R protomer adjacent to a single protomer that can bind CD4 (Figure 2C). Under these conditions,~25% of trimers were expected to carry the D368R mutation in all three protomers. This subpopulation cannot bind CD4. The remaining 25% carry two CD4-binding competent protomers next to the labeled mutant gp120 (Figure 2C).

Strikingly, despite the heterogeneity in the population of trimers, imaging of the HIV-1NL4-3 mixed trimer two restored the high occupancy of the State two conformation (Figure 2E). Hence, the observed high-FRET state likely originates from an unbound gp120 protomer adjacent to a CD4-bound protomer. Since the State three and State two stabilizations were observed for the same single-CD4 condition, but with dyes placed either into the protomer that binds a single CD4 or adjacent to it, this trimer must be asymmetric with protomers adopting distinct conformations (Figure 2L).

We next similarly generated mixed Env trimers for the Tier 2 HIV-1 isolates HIV-1JR-FL and HIV-1BG505. Since these Tier two viruses were less responsive to sCD4, we used the more potent ligand dodecameric sCD4D1D2-Igαtp. The D368R mutants for both HIV-1 isolates made all Envs largely resistant to sCD4 and sCD4D1D2-Igαtp (Figure 2F and I and Figure 2—figure supplement 3B and C). Intriguingly, the engineering of mixed trimers that can only bind a single CD4 allowed the asymmetric trimer to be stabilized, confirming that the high-FRET of the default structural intermediate originates from protomers adjacent to a single bound CD4 molecule (Figure 2G,H,J,K). The broad high-FRET peaks in sCD4D1D2-Igαtp-bound mixed trimer two suggest a potential continuum of various conformational states. These data support the hypothesis that the trimer spontaneously opens and closes through an asymmetric trimer configuration. Moreover, during virus entry, a single CD4 may initially engage a closed trimer (Kwon et al., 2015; Hu et al., 2017), and a very early version of this intermediate may be presented in the single CD4-bound structure of the Env trimer stabilized by the DS-SOSIP (Liu et al., 2017) when not fixed by DS mutation, CD4 binding to the trimer would induce the CD4-bound conformation in the bound protomer (Figure 2L). Binding of a single CD4 may loosen interaction of the V1/V2 loops in the trimer association domain so that neighboring protomers can adopt a conformation in which the V1 and V4 loops are closer to each other.

Binding of additional CD4 molecules or coreceptor surrogate antibody 17b completely opens the Env trimer

To further investigate how the conformation of Env changes when an additional CD4 binds to the trimer, we designed ‘mixed trimer 3’ for all three isolates, where two protomers are competent for binding to CD4, but the donor and acceptor fluorophores reside in the single protomer carrying a D368R mutation (Figure 3A). This allowed us to detect the conformation of the ligand free protomer when two CD4 molecules bind to the other two protomers within the Env trimer. Interestingly, when two CD4 molecules are bound to the Env trimer, the ligand-free protomer adopted the State three conformation, indicating that the trimer had fully opened (Figure 3B–D). Binding of two CD4 molecules is enough to flip the third protomer open indicating cooperativity between the three protomers. Our data are consistent with previous reports on the necessity of binding of multiple CD4 molecules to Env, as well as CD4 clustering (Yang et al., 2006; Salzwedel and Berger, 2009).

Figure 3. Binding of additional CD4 molecules or coreceptors completely opens the Env trimer.

(A) Scheme to illustrate generation of mixed HIV-1 Env trimer 3, in which two unlabeled protomers are CD4-binding competent and the one protomer carrying the donor and acceptor fluorophores (green, red stars in scheme above) is CD4-binding defective because it has the D368R mutation. (B–D) FRET histograms as in Figure 1 for the mixed HIV-1 Env trimer 3. sCD4D1D2-Igαtp (0.01 mg/ml) was incubated with the virus for 30 min prior to imaging. (E) Scheme to illustrate generation of mixed HIV-1 Env trimer 2, as in Figure 2. (F–H) FRET histograms as in Figure 1 for the mixed HIV-1 Env trimer 2. sCD4 (0.1 mg/ml) and 17b (0.1 mg/ml) were incubated with the virus for 30 min prior to imaging. (I) Schematic illustration of the further activation of the Env trimer. from the asymmetric intermediate. The CD4-bound conformation is in pink, and the conformational intermediate in the asymmetric trimer is in purple. The purple x indicates the D368R mutation. Green and red starts represent donor and acceptor fluorophores, respectively. Sizes of the stars represent relative change of fluorescence between donor and acceptor dyes and dotted line indicated changes of inter-dye distances.

Figure 3.

Figure 3—figure supplement 1. Model for the activation of the HIV-1 Env trimer through asymmetric intermediates.

Figure 3—figure supplement 1.

The Env configuration of the structural trimer intermediates depicts the conformational state of each protomer (State 1 = blue, State 2 = purple, State 3 = pink) within the trimer. See text for details.
Open in a new tab

We then asked if the asymmetric trimer intermediate, containing a single CD4-bound protomer (State two predominant configuration), is competent for coreceptor binding. To this end, we imaged the mixed trimer 2 of all three HIV-1 isolates and incubated them with both sCD4 (0.1 mg/ml) and coreceptor surrogate antibody 17b (0.1 mg/ml) (Figure 3E). We observed that the mixed trimer two bound by both a single sCD4 molecule and 17b also adopted the State three conformation (Figure 3F–H). This suggests that coreceptor binding would also be sufficient to trigger the complete opening of the single-CD4 bound asymmetric trimer (Figure 3I, Figure 3—figure supplement 1).

Kinetic and thermodynamic analysis of smFRET data

We performed a full kinetic and thermodynamic analysis of the FRET trajectories to quantify how ligand binding remodels the energy landscape governing Env dynamics. This analysis was permitted by organic fluorophores with enhanced photostability (Zheng et al., 2014), as well as the application of sCMOS cameras, which offered access to greater numbers of FRET trajectories (Juette et al., 2016). The occupancies in the FRET states were determined from the FRET histograms, and used to calculate the differences in free energies between states i and j according to ΔG°ij=-kBTln(Pi/Pj), where Pi and Pj are the occupancies of the ith and jth state in the histogram, respectively, and kB is the Boltzmann constant.

In HIV-1NL4-3, the predominant effect of sCD4 binding was to profoundly stabilize State 2. Before binding of sCD4, the relative free energy of HIV-1NL4-3 State 2 was 0.37 kBT higher than State 1 (Figure 4A, left). Following binding of sCD4, the free energy of State 2 was 0.54 kBT lower than State 1, resulting in a net change of −0.91 kBT. With respect to activation energies, sCD4 binding decreased the activation energy for the transitions from State one and State three into State 2, and increased the activation energy for transitions out of State 2. Along with the thermodynamic stabilization of State 2, these data offer an energetic explanation for the accumulation of molecules in State 2 (Figure 1C). In contrast, the predominant effect of sCD4D1D2-Igαtp binding was to stabilize State 3 (Figure 4A, right, Figure 1D). Before sCD4D1D2-Igαtp binding, the free energy of State 3 was 0.11 kBT lower than State 2. Following sCD4D1D2-Igαtp binding the free energy of State 3 was 0.69 kBT lower than the liganded State 2. With the liganded State 2 being 0.28 kBT lower than unliganded, this results in a net change of −0.86 kBT.

Figure 4. Kinetic analysis of HIV-1 Env binding to CD4.

Differences in free energies (ΔG°ij) between the FRET states, and the changes in activation energies (ΔΔG‡ij) arising from sCD4 and sCD4D1D2-Igαtp binding to (A) HIV-1NL4-3 (B) HIV-1JR-FL and (C) HIV-1BG505. ΔG°ij values were calculated from the occupancies of each FRET state in the FRET histograms. ΔΔG‡ij values were calculated from the rate constants for the observed transitions for the unliganded (black), sCD4-bound (green) and sCD4D1D2-Igαtp-bound (blue) HIV-1 Env. The dotted arrow lines indicate the actual ΔΔG‡ij of the forward and backward transitions while the solid line in between indicate the averaged value of the two. Energies are displayed in units of kBT, where kB is Boltzmann constant, and T is temperature.

Figure 4.

Figure 4—figure supplement 1. Survival probability distributions for transitions observed in HIV-1NL4-3 Env.

Figure 4—figure supplement 1.

(A–C) Survival probability distributions (gray) for each transition for the unliganded (A), sCD4-bound (B), and sCD4D1D2-Igαtp-bound (C) HIV-1NL4-3 Env. Each distribution was fit to the sum of two exponential distributions (y = A1 exp –k1t + A2 exp –k2t) (red). R2 of curve fitting was indicated in each panel.
Figure 4—figure supplement 2. Survival probability distributions for transitions observed in HIV-1JR-FL Env.

Figure 4—figure supplement 2.

Survival probability distributions as in Figure 4—figure supplement 1 for HIV-1JR-FL Env.
Figure 4—figure supplement 3. Survival probability distributions for transitions observed in HIV-1BG505 Env.

Figure 4—figure supplement 3.

Survival probability distributions as in Figure 4—figure supplement 1 for HIV-1BG505 Env.
Open in a new tab

Next, the dwell times in each FRET state identified by the application of HMM were compiled into histograms and fit to exponential distributions (Figure 4—figure supplement 1), revealing the rates of transition between the FRET states (Table 1). The rate constants obtained from this analysis were used to calculate the change in activation energies between two states ΔΔG‡ij = -kBT ln (kijliganded/kijunliganded), where kij is the rate of transition from the ith to jth FRET state. According to this analysis, sCD4 binding decreased the activation energy for the transitions from State one and State three into State 2, and increased the activation energy for transitions out of the State 2. Along with the thermodynamic stabilization of State 2, these data offer an energetic explanation for the accumulation of molecules in State 2. In contrast, the major effect of sCD4D1D2-Igαtp binding was the thermodynamic stabilization of State 3, with only minor effects on the observed kinetics (Figure 4A, right).

Table 1. Rates of transition between all observed FRET states for all three HIV-1 Envs.

The distribution of dwell times in each FRET state, determined through Hidden Markov Modeling (HMM), were fit to the sum of two exponential distributions (y = A1 exp –k1t + A2 exp –k2t) (Figure 4—figure supplements 1–3). The weighted average of the two rate constants from each fit are presented. Error bars represent 95% confidence intervals propagated from the kinetics analysis.

HIV-1NL4-3 k1→2 (s−1) k2→1 (s−1) k2→3 (s−1) k3→2 (s−1)
 Unliganded 1.50 ± 0.04 3.05 ± 0.04 1.89 ± 0.05 1.32 ± 0.04
 sCD4 2.10 ± 0.04 1.52 ± 0.04 1.21 ± 0.03 1.85 ± 0.03
 sCD4D1D2-Igαtp 1.60 ± 0.02 2.84 ± 0.04 1.87 ± 0.04 1.13 ± 0.02
HIV-1JR-FL
 Unliganded 1.30 ± 0.03 2.21 ± 0.03 1.70 ± 0.02 1.25 ± 0.02
 sCD4 1.52 ± 0.03 1.76 ± 0.04 2.25 ± 0.03 1.37 ± 0.05
 sCD4D1D2-Igαtp 2.02 ± 0.06 2.06 ± 0.05 1.81 ± 0.06 1.06 ± 0.03
HIV-1BG505
 Unliganded 1.10 ± 0.04 4.37 ± 0.08 2.86 ± 0.05 1.77 ± 0.03
 sCD4 1.21 ± 0.03 4.32 ± 0.08 1.81 ± 0.09 2.79 ± 0.04
 sCD4D1D2-Igαtp 1.62 ± 0.06 2.82 ± 0.05 2.92 ± 0.04 2.31 ± 0.05
Open in a new tab

Table 2. IC50 of sCD4 or sCD4D1D2-Igαtp of all three HIV-1 isolates.

Neutralization data were analyzed by nonlinear regression analysis and the Ab concentrations (μg/ml) at which 50% of virus infectivity was inhibited, were calculated.

sCD4 sCD4D1D2-Igαtp
WT D368R WT D368R
HIV-1NL4-3 0.56 >50 0.03 >50
HIV-1JR-FL 3.43 >50 0.06 >50
HIV-1BG505 1.26 >50 0.03 >50
Open in a new tab

We then repeated these experiments for the clinical HIV-1 isolate JR-FL and BG505. The unliganded HIV-1JR-FL and HIV-1BG505 Envs more stably adopted State one as compared to the HIV-1NL4-3 Env. This finding indicates that, relative to the other states, the closed state has lower energy in HIV-1JR-FL compared with the lab-adapted HIV-1NL4-3 isolate (Figure 4B and C, Figure 1E and H). Specifically, the energy of State 1 of HIV-1JR-FL and HIV-1BG505 was lower than that of State 2 by 1.05 kBT and 1.16 kBT, respectively, as compared to 0.37 kBT for NL4-3 (Figure 4B and C). We also found a modest decrease in the activation energy for the sCD4-induced transition out of State 1, and decreased activation energy for exchanges between States 2 and 3 (Figure 4B and C, left). Importantly, as was observed for HIV-1NL4-3 Env, addition of the dodecameric sCD4D1D2-Igαtp stabilized State 3 (Figure 4B and C, right), decreasing the energy to 0.6 kBT and 0.79 kBT lower than that of State 2 of HIV-1JR-FL and HIV-1BG505, respectively. (Figure 4B and C, right).

Interestingly, HIV-1JR-FL and HIVBG505 Envs were not as dramatically stabilized in State two by sCD4 binding as was observed for HIV-1NL4-3 Env. Rather, binding of sCD4 and sCD4D1D2-Igαtp to HIV-1JR-FL and HIV-1BG505 Env exhibited similar tendencies in the stabilization of State two as well as State 3. sCD4D1D2-Igαtp was clearly more efficient in promoting the accumulation of molecules in the State three conformation. Thus, while there are distinct differences in the ability of sCD4 to activate Env for all three strains, the potent dodecameric sCD4D1D2-Igαtp stabilizes Envs from all three isolates in the open conformation (State 3).

Discussion

Here we used smFRET to identify the nature of the three conformational states sampled by HIV-1 Env (Munro et al., 2014). The most populated FRET State one corresponds to the gp120 conformation of the mature pre-triggered Env trimer. The State three stabilization by oligomerized sCD4D1D2-Igαtp, sCD4 and 17b, and the CD4 mimetic small molecule JRC-II-191 (Munro et al., 2014), indicated that State three corresponds to the gp120 conformation of the open, three-CD4-bound trimer. By using a mixed trimer assay, we identified the most prevalent default intermediate State two to be an asymmetric Env trimer configuration engaging a single CD4 molecule. The individual protomers within the asymmetric trimer adopted different conformations, and State two population originated from the two unbound protomers adjacent to the single CD4-bound protomer.

Based on these single-molecule analyses and previous results, we propose a model of HIV-1 Env activation by receptor CD4 and coreceptor (Figure 3—figure supplement 1). This description of the molecular events during Env activation requires a nomenclature that specifies the conformational state of each protomer within the trimer. The unliganded trimer resides predominantly in a conformation in which all three protomers are in the pre-triggered State one configuration, which can be represented as the (1,1,1) configuration (where each of the three numbers defines the FRET state of each protomer within the trimer). Binding of a single CD4 molecule then induces conformational changes in the CD4-bound protomer, such that the trimer adopts a (3,2,2) configuration. In the (3,2,2) configuration, the CD4-bound protomer exhibits intermediate-FRET (State 3), while the two neighboring protomers exhibit high-FRET (State 2). The near absence of direct transitions between State one to State three suggests that the trimer opening likely occurs through a conformational intermediate with a FRET value similar to State two such as a (2,1,1) or (2,2,2) configuration.

The complete activation of HIV-1 Env then likely requires either a local clustering of CD4 or coreceptor binding. This would induce of the Env trimer to reach more open configurations (either (3,3,2) or (3,3,3)) (Yang et al., 2006; Salzwedel and Berger, 2009; Ozorowski et al., 2017). Coreceptor binding then leads to the activation of gp41 and the induction of additional conformational changes necessary for membrane fusion (Figure 3—figure supplement 1). During the spontaneous opening and closing of the HIV-1 Env trimer between States 1 and 3, it is possible that the opening protomer does not fully adopt the three-CD4-bound conformation. Rather it may only partially open, just far enough for the trimer apex to be disrupted and the adjacent protomers adopting the State two high-FRET conformation.

In this paper, we used smFRET to reveal asymmetric trimer configurations in the opening of the HIV-1 Env trimer. Asymmetric structural intermediates during the activation of viral fusion machines may be more common than previously thought. Single receptor engagement and sequential shedding has been observed for the Murine Leukemia Virus (MLV) Env (Riedel et al., 2017; Sjöberg et al., 2017). And a recent study on the activation of pre-fusion HIV-1 gp41 to downstream intermediate states suggested that gp41 also opens through asymmetric intermediates (Khasnis et al., 2016). Our work illustrates the power of smFRET to define prevalent functional intermediates in the opening of the HIV-1 Env trimer.

Materials and methods

Cell lines

Name and Source: HEK293, ATCC, catalog # CRL- CRL-1573

Use and Authentication: The HEK293 cell line is used due to its high transfectability and gives high titers for produced retroviruses. The cell line is obtained as an early passage from ATCC which carries out Cell Authentication using Short Tandem Repeat Profiling before distribution. We are maintaining a big batch of frozen stocks from early passage. The cell line was tested and confirmed to mycoplasma free.

Name and Source: TZMBL, AIDS RRP # 8129

Use and Authentication: Used as indicator cell line for determining infectious units of produced HIV-1 viruses. We have tested the surface expression of CCR5 and CXCR4 using specific antibodies and challenging with R5 and X4-tropic viruses. We have also carried out PCR-based analyses to ascertain that the cell line is not infected with XMRV that was shown to increase infectivity of Nef-deficient viruses due to expression of glycoGag. The cell line was tested and confirmed to mycoplasma free.

Preparation of labeled virions

The plasmid pNL4-3_V1Q3_V4A1 ΔRT encoding for HIV-1NL4-3 Env and plasmid Q23_BG505_V1Q3-V4A1 ΔRT encoding for HIV-1BG505 Env carrying the Q3 (GQQQLG) and A1 (GDSLDMLEWSLM) labeling peptides (Wu et al., 2006; Yin et al., 2006; Zhou et al., 2007) in V1 and V4 loops of full-length pNL4-3 ΔRT construct and full-length Q23_BG505 ΔRT construct, respectively; and the Env-expressing plasmid pCAGGS_JR-FL_V1Q3_V4A1 containing the peptides in analogous positions were previously described (Munro et al., 2014). The D368R point mutation (Olshevsky et al., 1990) was introduced into untagged constructs and V1Q3_V4A1 tagged constructs of full-length NL4-3, BG505 and Env-expressing JR-FL by overlap-extension PCR. All viruses were produced by co-transfecting HEK293 cells using the following ratios of plasmids:

HIV-1NL4-3: 40:1 of pNL4-3 ΔRT to pNL4-3_V1Q3_V4A1 ΔRT;

HIV-1BG505: 40:1 of Q23_BG505 ΔRT to Q23_BG505_V1Q3_V4A1 ΔRT;

HIV-1JR-FL: 1:1 of the pNL4-3 ΔEnvΔRT and the Env-expressing constructs, which were at 40:1 for pCAGGS_JR-FL to pCAGGS_JR-FL_V1Q3_V4A1.

HIV-1NL4-3_D368R: 40:1 of pNL4-3_D368R ΔRT to pNL4-3_V1Q3_V4A1_D368R ΔRT.

HIV-1BG505_D368R: 40:1 of Q23_BG505_D368R ΔRT to Q23_BG505_V1Q3_V4A1_D368R ΔRT.

HIV-1JR-FL_D368R: 1:1 of the pNL4-3 ΔEnvΔRT and the Env-expressing constructs, which were at 40:1 for pCAGGS_JR-FL_D368R to pCAGGS_JR-FL_V1Q3_V4A1_D368R.

HIV-1NL4-3 mixed trimer 1: 40:1 of pNL4-3_D368R ΔRT to pNL4-3_V1Q3_V4A1 ΔRT.

HIV-1BG505 mixed trimer 1: 40:1 of Q23_BG505_D368R ΔRT to Q23_BG505_V1Q3_V4A1 ΔRT.

HIV-1JR-FL mixed trimer 1: 1:1 of the pNL4-3 ΔEnvΔRT and the Env-expressing constructs, which were at 40:1 for pCAGGS_JR-FL_D368R to pCAGGS_JR-FL_V1Q3_V4A1.

HIV-1NL4-3 mixed trimer 2: 20:20:1 of pNL4-3 ΔRT to pNL4-3_D368R ΔRT to pNL4-3_V1Q3_V4A1_D368R ΔRT.

HIV-1BG505 mixed trimer 2: 20:20:1 of Q23_BG505 ΔRT to Q23_BG505_D368R ΔRT to Q23_BG505_V1Q3_V4A1_D368R ΔRT.

HIV-1JR-FL mixed trimer 2: 1:1 of the pNL4-3 ΔEnvΔRT and the Env-expressing constructs, which were at 20:20:1 for pCAGGS_JR-FL to pCAGGS_D368R to pCAGGS_JR-FL_V1Q3_V4A1_D368R.

Statistically, the 20:20:1 ratio for mixed trimer 2 will yield 50% trimers with the desired mixed trimer two shown in Figure 2B.

HIV-1NL4-3 mixed trimer 3: 40:1 of pNL4-3 ΔRT to pNL4-3_V1Q3_V4A1_D368R ΔRT.

HIV-1BG505 mixed trimer 3: 40:1 of Q23_BG505 ΔRT to Q23_BG505_V1Q3_V4A1_D368R ΔRT.

HIV-1JR-FL mixed trimer 3: 1:1 of the pNL4-3 ΔEnvΔRT and the Env-expressing constructs, which were at 40:1 for pCAGGS_JR-FL to pCAGGS_JR-FL_V1Q3_V4A1_D368R.

Virus was harvested 40 hr post transfection and concentrated by centrifugation over a 15% sucrose cushion at 20,000 x g for 2 hr. Virus pellets were resuspended in the labeling buffer containing 50 mM HEPES, 10 mM MgCl2, 10 mM CaCl2. Resuspended virus was enzymatically labeled with fluorophores in a reaction mixture containing Cy3B(3S)-cadaverine (0.5 μM), transglutaminase (0.65 μM; Sigma Aldrich), LD650-CoA (0.5 μM) (Lumidyne Technologies), and AcpS (5 μM) (Zhou et al., 2007). The labeling reaction was incubated at room temperature overnight. 0.1 mg/ml PEG2000-biotin (Avanti Polar Lipids) was added to the labeling reaction and incubated for 30 min at room temperature. Under these conditions, the labeling efficiencies of the Q3 and A1 tags are 40% and 55%, respectively (Munro et al., 2014). The virus was then purified away from excess fluorophore and lipid by ultracentrifugation (1 hr at 150,000 x g) over a 6–18% Optiprep (Sigma Aldrich) gradient containing 50 mM Tris pH 7.4 and 50 mM NaCl. The fractions containing viral particles were identified by p24 western blot and stored at −80°C.

smFRET imaging

Labeled viruses were immobilized on a passivated quartz microscope slide coated with streptavidin. smFRET imaging was conducted at room temperature on a prism-based TIRF microscope with a 60 × 1.27 NA water-immersion objective (Nikon). Donor fluorophores were excited by a 532 nm laser (Laser Quantum) at ~0.1 kW per cm2. Data were recorded at 25 frames per second for 2000 frames by two synchronized ORCA-Flash4.0v2 sCMOS cameras (Hamamatsu, 2048 × 2048), as described (Juette et al., 2016). The imaging buffer contained 50 mM Tris pH 7.4, 50 mM NaCl, and a cocktail of triplet-state quenchers. 2 mM protocatechuic acid (PCA) and 8 nM protocatechuic 3,4-dioxygenase (PCD) were also included to remove molecular oxygen (Aitken et al., 2008). For experiments performed in the presence of ligands, 0.1 mg/ml sCD4 or sCD4D1D2-Igαtp were incubated with the immobilized virus at room temperature for 30 min prior to imaging.

smFRET data analysis

The smFRET trajectories were extracted from the movies and processed using SPARTAN (Juette et al., 2016). FRET values were calculated according to FRET = IA/(γID + IA), where IA and ID are the fluorescence intensities of acceptor and donor dyes, respectively, and γ is the coefficient correcting for the difference in detection efficiencies of the donor and acceptor channels. The FRET histograms were fit to the sum of 3 Gaussian distributions. The occupancy in each FRET state was determined by the area under each Gaussian curve. Error bars were generated by propagating the uncertainties of the fits through the occupancy calculation. The occupancies in the FRET states were used to calculate the differences in free energies between states i and j according to ΔG°ij=-kBTln(Pi/Pj), where Pi and Pj are the occupancies of the ith and jth state in the histogram, respectively, and kB is the Boltzmann constant.

smFRET traces were idealized using a segmental k-means algorithm with a 3-state model (Qin, 2004). The frequencies of transitions were displayed in transition density plots (TDP). Dwell times in each states were compiled into histograms and fit to the sum of two exponential distributions (y = A1 exp –k1t + A2 exp –k2t). The reported rates were determined by averaging the two rate constants, weighted by their respective amplitudes. The apparent change in activation energy barriers ΔΔG‡ij was calculated according to ΔΔG‡ij = -kBT(kijliganded/kijunliganded), where kij is the rate constants determined from the exponential curves, kB is the Boltzmann constant, T is the temperature in kelvin. The uncertainties of ΔG°ij and ΔΔG‡ij were determined by propagating the uncertainties from the curve fitting.

Infectivity measurements

Viruses carrying wild-type, D368R mutant HIV-1 Env were generated by transfection of HEK293 cells with full-length WT or mutant pNL4-3 or Q23_BG505 constructs, or at a 1:1 ratio of pNL4-3 ΔEnv and WT or mutant pCAAGS_JR-FL, along with 1:6 ratio of HIV-1-InGluc to viral constructs, using Fugene 6 (Promega, Madison, WI). Supernatants from 24 and 48 hr post-transfection were combined, filtered through a 0.45 μM filter (Pall Corporation) and titered on TZMbl cells. Titers were determined 48 hr post-infection either by measuring luciferase activity using the Gaussia luciferase assay kit (NEB). For testing the effects of ligands on infectivity, viruses were incubated with ligands for 30 min at room temperature prior to addition to TZMbl cells.

The V1V4-tagged BG505 was validated in form of the 100% tagged virus, not as the dye-labeled virus since the incomplete labeling efficiencies would leave enough unmodified trimers on the surface of a virus to account for nearly undiminished infectivity. While we cannot exclude additional effects of the dyes, we have not observed anisotropy for singly labeled virus, the dyes are highly hydrophilic (Zheng et al., 2014), do not associate with the viral membrane and no dye associated with viruses in the absence of labeling tags (Munro et al., 2014).

Acknowledgements

We thank Alon Herschhorn and Joseph Sodroski for discussions. Patent applications pertaining to this work are U.S. Patent Application 13/202,351, Methods and Compositions for Altering Photophysical Properties of Fluorophores via Proximal Quenching (SCB, ZZ); U.S. Patent Application 14/373,402 Dye Compositions, Methods of Preparation, Conjugates Thereof, and Methods of Use (SCB, ZZ); and International and US Patent Application PCT/US13/42249 Reagents and Methods for Identifying Anti-HIV Compounds (SCB, JBM, WM). SCB is a co-founder of Lumidyne Corporation.

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Contributor Information

James B Munro, Email: james.munro@tufts.edu.

Walther Mothes, Email: walther.mothes@yale.edu.

Pamela J Bjorkman, California Institute of Technology, United States.

Funding Information

This paper was supported by the following grants:

  • National Institutes of Health GM116654 to Walther Mothes.

  • National Institutes of Health AI116262 to James B Munro.

  • National Institutes of Health GM098859 to Scott C Blanchard.

  • National Institutes of Health GM056550 to Scott C Blanchard, Walther Mothes.

  • Cancer Research Institute Irvington Fellows Program to James B Munro.

  • National Institutes of Health AI042853 to James B Munro.

  • China Scholarship Council Yale World Scholars to Xiaochu Ma.

  • National Institutes of Health AI005023-17 to Peter D Kwong.

Additional information

Competing interests

No competing interests declared.

Patent applications pertaining to this work are U.S. Patent Application 13/202,351, Methods and Compositions for Altering Photophysical Properties of Fluorophores via Proximal Quenching (S.C.B., Z.Z.); U.S. Patent Application 14/373,402 Dye Compositions, Methods of Preparation, Conjugates Thereof, and Methods of Use (S.C.B., Z.Z.); and International and US Patent Application PCT/US13/42249 Reagents and Methods for Identifying Anti-HIV Compounds (S.C.B., J.B.M., W.M.). S.C.B. is a co-founder of Lumidyne Corporation.

Author contributions

Conceptualization, Data curation, Formal analysis, Visualization, Writing—original draft, Writing—review and editing.

Data curation, Formal analysis, Writing—review and editing.

Conceptualization, Data curation, Methodology, Writing—review and editing.

Data curation, Software.

Software.

Methodology.

Methodology.

Methodology.

Resources.

Resources, Methodology, Writing—review and editing.

Conceptualization, Resources, Supervision, Funding acquisition, Writing—review and editing.

Conceptualization, Formal analysis, Supervision, Funding acquisition, Methodology, Writing—original draft, Writing—review and editing.

Conceptualization, Formal analysis, Supervision, Funding acquisition, Investigation, Writing—original draft, Project administration, Writing—review and editing.

Additional files

Transparent reporting form
elife-34271-transrepform.docx (244.7KB, docx)
DOI: 10.7554/eLife.34271.019

References

  1. Aitken CE, Marshall RA, Puglisi JD. An oxygen scavenging system for improvement of dye stability in single-molecule fluorescence experiments. Biophysical Journal. 2008;94:1826–1835. doi: 10.1529/biophysj.107.117689. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Arthos J, Cicala C, Steenbeke TD, Chun TW, Dela Cruz C, Hanback DB, Khazanie P, Nam D, Schuck P, Selig SM, Van Ryk D, Chaikin MA, Fauci AS. Biochemical and biological characterization of a dodecameric CD4-Ig fusion protein: implications for therapeutic and vaccine strategies. The Journal of Biological Chemistry. 2002;277:11456–11464. doi: 10.1074/jbc.M111191200. [DOI] [PubMed] [Google Scholar]
  3. Bennett A, Liu J, Van Ryk D, Bliss D, Arthos J, Henderson RM, Subramaniam S. Cryoelectron tomographic analysis of an HIV-neutralizing protein and its complex with native viral gp120. Journal of Biological Chemistry. 2007;282:27754–27759. doi: 10.1074/jbc.M702025200. [DOI] [PubMed] [Google Scholar]
  4. Blumenthal R, Durell S, Viard M. HIV entry and envelope glycoprotein-mediated fusion. Journal of Biological Chemistry. 2012;287:40841–40849. doi: 10.1074/jbc.R112.406272. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Caskey M, Klein F, Lorenzi JC, Seaman MS, West AP, Buckley N, Kremer G, Nogueira L, Braunschweig M, Scheid JF, Horwitz JA, Shimeliovich I, Ben-Avraham S, Witmer-Pack M, Platten M, Lehmann C, Burke LA, Hawthorne T, Gorelick RJ, Walker BD, Keler T, Gulick RM, Fätkenheuer G, Schlesinger SJ, Nussenzweig MC. Viraemia suppressed in HIV-1-infected humans by broadly neutralizing antibody 3BNC117. Nature. 2015;522:487–491. doi: 10.1038/nature14411. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Doms RW, Moore JP. HIV-1 membrane fusion: targets of opportunity. The Journal of Cell Biology. 2000;151:F9–14. doi: 10.1083/jcb.151.2.f9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Gautam R, Nishimura Y, Pegu A, Nason MC, Klein F, Gazumyan A, Golijanin J, Buckler-White A, Sadjadpour R, Wang K, Mankoff Z, Schmidt SD, Lifson JD, Mascola JR, Nussenzweig MC, Martin MA. A single injection of anti-HIV-1 antibodies protects against repeated SHIV challenges. Nature. 2016;533:105–109. doi: 10.1038/nature17677. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Guttman M, Cupo A, Julien JP, Sanders RW, Wilson IA, Moore JP, Lee KK. Antibody potency relates to the ability to recognize the closed, pre-fusion form of HIV Env. Nature Communications. 2015;6:6144. doi: 10.1038/ncomms7144. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Harrison SC. Viral membrane fusion. Virology. 2015;479-480:498–507. doi: 10.1016/j.virol.2015.03.043. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Herschhorn A, Gu C, Moraca F, Ma X, Farrell M, Smith AB, Pancera M, Kwong PD, Schön A, Freire E, Abrams C, Blanchard SC, Mothes W, Sodroski JG. The β20-β21 of gp120 is a regulatory switch for HIV-1 Env conformational transitions. Nature Communications. 2017;8:1049. doi: 10.1038/s41467-017-01119-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Herschhorn A, Ma X, Gu C, Ventura JD, Castillo-Menendez L, Melillo B, Terry DS, Smith AB, Blanchard SC, Munro JB, Mothes W, Finzi A, Sodroski J. Release of gp120 restraints leads to an entry-competent intermediate state of the HIV-1 envelope glycoproteins. mBio. 2016;7:e01598-16. doi: 10.1128/mBio.01598-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Hu G, Liu J, Roux KH, Taylor KA. Structure of Simian Immunodeficiency Virus Envelope Spikes Bound with CD4 and Monoclonal Antibody 36D5. Journal of Virology. 2017;91:e00134-17. doi: 10.1128/JVI.00134-17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Huang CC, Lam SN, Acharya P, Tang M, Xiang SH, Hussan SS, Stanfield RL, Robinson J, Sodroski J, Wilson IA, Wyatt R, Bewley CA, Kwong PD. Structures of the CCR5 N terminus and of a tyrosine-sulfated antibody with HIV-1 gp120 and CD4. Science. 2007;317:1930–1934. doi: 10.1126/science.1145373. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Juette MF, Terry DS, Wasserman MR, Altman RB, Zhou Z, Zhao H, Blanchard SC. Single-molecule imaging of non-equilibrium molecular ensembles on the millisecond timescale. Nature Methods. 2016;13:341–344. doi: 10.1038/nmeth.3769. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Julien JP, Cupo A, Sok D, Stanfield RL, Lyumkis D, Deller MC, Klasse PJ, Burton DR, Sanders RW, Moore JP, Ward AB, Wilson IA. Crystal structure of a soluble cleaved HIV-1 envelope trimer. Science. 2013;342:1477–1483. doi: 10.1126/science.1245625. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Khasnis MD, Halkidis K, Bhardwaj A, Root MJ. Receptor activation of HIV-1 env leads to asymmetric exposure of the gp41 Trimer. PLoS Pathogens. 2016;12:e1006098. doi: 10.1371/journal.ppat.1006098. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Klein F, Halper-Stromberg A, Horwitz JA, Gruell H, Scheid JF, Bournazos S, Mouquet H, Spatz LA, Diskin R, Abadir A, Zang T, Dorner M, Billerbeck E, Labitt RN, Gaebler C, Marcovecchio P, Incesu RB, Eisenreich TR, Bieniasz PD, Seaman MS, Bjorkman PJ, Ravetch JV, Ploss A, Nussenzweig MC. HIV therapy by a combination of broadly neutralizing antibodies in humanized mice. Nature. 2012;492:118–122. doi: 10.1038/nature11604. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Koyanagi Y, Miles S, Mitsuyasu RT, Merrill JE, Vinters HV, Chen IS. Dual infection of the central nervous system by AIDS viruses with distinct cellular tropisms. Science. 1987;236:819–822. doi: 10.1126/science.3646751. [DOI] [PubMed] [Google Scholar]
  19. Kwon YD, Pancera M, Acharya P, Georgiev IS, Crooks ET, Gorman J, Joyce MG, Guttman M, Ma X, Narpala S, Soto C, Terry DS, Yang Y, Zhou T, Ahlsen G, Bailer RT, Chambers M, Chuang GY, Doria-Rose NA, Druz A, Hallen MA, Harned A, Kirys T, Louder MK, O'Dell S, Ofek G, Osawa K, Prabhakaran M, Sastry M, Stewart-Jones GB, Stuckey J, Thomas PV, Tittley T, Williams C, Zhang B, Zhao H, Zhou Z, Donald BR, Lee LK, Zolla-Pazner S, Baxa U, Schön A, Freire E, Shapiro L, Lee KK, Arthos J, Munro JB, Blanchard SC, Mothes W, Binley JM, McDermott AB, Mascola JR, Kwong PD. Crystal structure, conformational fixation and entry-related interactions of mature ligand-free HIV-1 Env. Nature Structural & Molecular Biology. 2015;22:522–531. doi: 10.1038/nsmb.3051. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kwong PD, Doyle ML, Casper DJ, Cicala C, Leavitt SA, Majeed S, Steenbeke TD, Venturi M, Chaiken I, Fung M, Katinger H, Parren PW, Robinson J, Van Ryk D, Wang L, Burton DR, Freire E, Wyatt R, Sodroski J, Hendrickson WA, Arthos J. HIV-1 evades antibody-mediated neutralization through conformational masking of receptor-binding sites. Nature. 2002;420:678–682. doi: 10.1038/nature01188. [DOI] [PubMed] [Google Scholar]
  21. Kwong PD, Wyatt R, Robinson J, Sweet RW, Sodroski J, Hendrickson WA. Structure of an HIV gp120 envelope glycoprotein in complex with the CD4 receptor and a neutralizing human antibody. Nature. 1998;393:648–659. doi: 10.1038/31405. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Lee JH, Ozorowski G, Ward AB. Cryo-EM structure of a native, fully glycosylated, cleaved HIV-1 envelope trimer. Science. 2016;351:1043–1048. doi: 10.1126/science.aad2450. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Liu J, Bartesaghi A, Borgnia MJ, Sapiro G, Subramaniam S. Molecular architecture of native HIV-1 gp120 trimers. Nature. 2008;455:109–113. doi: 10.1038/nature07159. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Liu Q, Acharya P, Dolan MA, Zhang P, Guzzo C, Lu J, Kwon A, Gururani D, Miao H, Bylund T, Chuang GY, Druz A, Zhou T, Rice WJ, Wigge C, Carragher B, Potter CS, Kwong PD, Lusso P. Quaternary contact in the initial interaction of CD4 with the HIV-1 envelope trimer. Nature Structural & Molecular Biology. 2017;24:370–378. doi: 10.1038/nsmb.3382. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lu CL, Murakowski DK, Bournazos S, Schoofs T, Sarkar D, Halper-Stromberg A, Horwitz JA, Nogueira L, Golijanin J, Gazumyan A, Ravetch JV, Caskey M, Chakraborty AK, Nussenzweig MC. Enhanced clearance of HIV-1-infected cells by broadly neutralizing antibodies against HIV-1 in vivo. Science. 2016;352:1001–1004. doi: 10.1126/science.aaf1279. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Lyumkis D, Julien JP, de Val N, Cupo A, Potter CS, Klasse PJ, Burton DR, Sanders RW, Moore JP, Carragher B, Wilson IA, Ward AB. Cryo-EM structure of a fully glycosylated soluble cleaved HIV-1 envelope trimer. Science. 2013;342:1484–1490. doi: 10.1126/science.1245627. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Munro JB, Gorman J, Ma X, Zhou Z, Arthos J, Burton DR, Koff WC, Courter JR, Smith AB, Kwong PD, Blanchard SC, Mothes W. Conformational dynamics of single HIV-1 envelope trimers on the surface of native virions. Science. 2014;346:759–763. doi: 10.1126/science.1254426. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Nishimura Y, Gautam R, Chun TW, Sadjadpour R, Foulds KE, Shingai M, Klein F, Gazumyan A, Golijanin J, Donaldson M, Donau OK, Plishka RJ, Buckler-White A, Seaman MS, Lifson JD, Koup RA, Fauci AS, Nussenzweig MC, Martin MA. Early antibody therapy can induce long-lasting immunity to SHIV. Nature. 2017;543:559–563. doi: 10.1038/nature21435. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Olshevsky U, Helseth E, Furman C, Li J, Haseltine W, Sodroski J. Identification of individual human immunodeficiency virus type 1 gp120 amino acids important for CD4 receptor binding. Journal of Virology. 1990;64:5701–5707. doi: 10.1128/jvi.64.12.5701-5707.1990. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Ozorowski G, Pallesen J, de Val N, Lyumkis D, Cottrell CA, Torres JL, Copps J, Stanfield RL, Cupo A, Pugach P, Moore JP, Wilson IA, Ward AB. Open and closed structures reveal allostery and pliability in the HIV-1 envelope spike. Nature. 2017;547:360–363. doi: 10.1038/nature23010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Pancera M, Lai YT, Bylund T, Druz A, Narpala S, O'Dell S, Schön A, Bailer RT, Chuang GY, Geng H, Louder MK, Rawi R, Soumana DI, Finzi A, Herschhorn A, Madani N, Sodroski J, Freire E, Langley DR, Mascola JR, McDermott AB, Kwong PD. Crystal structures of trimeric HIV envelope with entry inhibitors BMS-378806 and BMS-626529. Nature Chemical Biology. 2017;13:1115–1122. doi: 10.1038/nchembio.2460. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Pancera M, Majeed S, Ban YE, Chen L, Huang CC, Kong L, Kwon YD, Stuckey J, Zhou T, Robinson JE, Schief WR, Sodroski J, Wyatt R, Kwong PD. Structure of HIV-1 gp120 with gp41-interactive region reveals layered envelope architecture and basis of conformational mobility. PNAS. 2010;107:1166–1171. doi: 10.1073/pnas.0911004107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Pancera M, Zhou T, Druz A, Georgiev IS, Soto C, Gorman J, Huang J, Acharya P, Chuang GY, Ofek G, Stewart-Jones GB, Stuckey J, Bailer RT, Joyce MG, Louder MK, Tumba N, Yang Y, Zhang B, Cohen MS, Haynes BF, Mascola JR, Morris L, Munro JB, Blanchard SC, Mothes W, Connors M, Kwong PD. Structure and immune recognition of trimeric pre-fusion HIV-1 Env. Nature. 2014;514:455–461. doi: 10.1038/nature13808. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Qin F. Restoration of single-channel currents using the segmental k-means method based on hidden Markov modeling. Biophysical Journal. 2004;86:1488–1501. doi: 10.1016/S0006-3495(04)74217-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Riedel C, Vasishtan D, Siebert CA, Whittle C, Lehmann MJ, Mothes W, Grünewald K. Native structure of a retroviral envelope protein and its conformational change upon interaction with the target cell. Journal of Structural Biology. 2017;197:172–180. doi: 10.1016/j.jsb.2016.06.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Salzwedel K, Berger EA. Complementation of diverse HIV-1 Env defects through cooperative subunit interactions: a general property of the functional trimer. Retrovirology. 2009;6:75. doi: 10.1186/1742-4690-6-75. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Sanders RW, Derking R, Cupo A, Julien JP, Yasmeen A, de Val N, Kim HJ, Blattner C, de la Peña AT, Korzun J, Golabek M, de Los Reyes K, Ketas TJ, van Gils MJ, King CR, Wilson IA, Ward AB, Klasse PJ, Moore JP. A next-generation cleaved, soluble HIV-1 Env trimer, BG505 SOSIP.664 gp140, expresses multiple epitopes for broadly neutralizing but not non-neutralizing antibodies. PLoS Pathogens. 2013;9:e1003618. doi: 10.1371/journal.ppat.1003618. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Scharf L, Wang H, Gao H, Chen S, McDowall AW, Bjorkman PJ. Broadly neutralizing antibody 8ANC195 recognizes closed and open states of HIV-1 Env. Cell. 2015;162:1379–1390. doi: 10.1016/j.cell.2015.08.035. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Schoofs T, Klein F, Braunschweig M, Kreider EF, Feldmann A, Nogueira L, Oliveira T, Lorenzi JC, Parrish EH, Learn GH, West AP, Bjorkman PJ, Schlesinger SJ, Seaman MS, Czartoski J, McElrath MJ, Pfeifer N, Hahn BH, Caskey M, Nussenzweig MC. HIV-1 therapy with monoclonal antibody 3BNC117 elicits host immune responses against HIV-1. Science. 2016;352:997–1001. doi: 10.1126/science.aaf0972. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Sjöberg M, Löving R, Lindqvist B, Garoff H. Sequential activation of the three protomers in the Moloney murine leukemia virus Env. PNAS. 2017;114:2723–2728. doi: 10.1073/pnas.1617264114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Trkola A, Dragic T, Arthos J, Binley JM, Olson WC, Allaway GP, Cheng-Mayer C, Robinson J, Maddon PJ, Moore JP. CD4-dependent, antibody-sensitive interactions between HIV-1 and its co-receptor CCR-5. Nature. 1996;384:184–187. doi: 10.1038/384184a0. [DOI] [PubMed] [Google Scholar]
  42. Walker LM, Huber M, Doores KJ, Falkowska E, Pejchal R, Julien JP, Wang SK, Ramos A, Chan-Hui PY, Moyle M, Mitcham JL, Hammond PW, Olsen OA, Phung P, Fling S, Wong CH, Phogat S, Wrin T, Simek MD, Koff WC, Wilson IA, Burton DR, Poignard P, Protocol G Principal Investigators Broad neutralization coverage of HIV by multiple highly potent antibodies. Nature. 2011;477:466–470. doi: 10.1038/nature10373. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Wang H, Cohen AA, Galimidi RP, Gristick HB, Jensen GJ, Bjorkman PJ. Cryo-EM structure of a CD4-bound open HIV-1 envelope trimer reveals structural rearrangements of the gp120 V1V2 loop. PNAS. 2016;113:E7151–E7158. doi: 10.1073/pnas.1615939113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Wu X, Parast AB, Richardson BA, Nduati R, John-Stewart G, Mbori-Ngacha D, Rainwater SM, Overbaugh J. Neutralization escape variants of human immunodeficiency virus type 1 are transmitted from mother to infant. Journal of Virology. 2006;80:835–844. doi: 10.1128/JVI.80.2.835-844.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Wu X, Yang ZY, Li Y, Hogerkorp CM, Schief WR, Seaman MS, Zhou T, Schmidt SD, Wu L, Xu L, Longo NS, McKee K, O'Dell S, Louder MK, Wycuff DL, Feng Y, Nason M, Doria-Rose N, Connors M, Kwong PD, Roederer M, Wyatt RT, Nabel GJ, Mascola JR. Rational design of envelope identifies broadly neutralizing human monoclonal antibodies to HIV-1. Science. 2010;329:856–861. doi: 10.1126/science.1187659. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Wyatt R, Sodroski J. The HIV-1 envelope glycoproteins: fusogens, antigens, and immunogens. Science. 1998;280:1884–1888. doi: 10.1126/science.280.5371.1884. [DOI] [PubMed] [Google Scholar]
  47. Yang X, Kurteva S, Ren X, Lee S, Sodroski J. Subunit stoichiometry of human immunodeficiency virus type 1 envelope glycoprotein trimers during virus entry into host cells. Journal of Virology. 2006;80:4388–4395. doi: 10.1128/JVI.80.9.4388-4395.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Yin J, Lin AJ, Golan DE, Walsh CT. Site-specific protein labeling by Sfp phosphopantetheinyl transferase. Nature Protocols. 2006;1:280–285. doi: 10.1038/nprot.2006.43. [DOI] [PubMed] [Google Scholar]
  49. Zheng Q, Juette MF, Jockusch S, Wasserman MR, Zhou Z, Altman RB, Blanchard SC. Ultra-stable organic fluorophores for single-molecule research. Chem. Soc. Rev. 2014;43:1044–1056. doi: 10.1039/C3CS60237K. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Zhou Z, Cironi P, Lin AJ, Xu Y, Hrvatin S, Golan DE, Silver PA, Walsh CT, Yin J. Genetically encoded short peptide tags for orthogonal protein labeling by Sfp and AcpS phosphopantetheinyl transferases. ACS Chemical Biology. 2007;2:337–346. doi: 10.1021/cb700054k. [DOI] [PubMed] [Google Scholar]

Decision letter

Editor: Pamela J Bjorkman1

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "HIV-1 Env trimer opens through an asymmetric intermediate in which individual protomers adopt distinct conformations" for consideration by eLife. Your article has been favorably evaluated by Wenhui Li (Senior Editor) and three reviewers, one of whom is a member of our Board of Reviewing Editors. The following individual involved in review of your submission has agreed to reveal his identity: Joseph D Puglisi (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

The manuscript of Ma et al. addresses the longstanding debate about the conformational state of the HIV-1 Env, and the relationship of conformation to viral entry. Clearly, this is a conformationally flexible system involving challenging protein chemistry; as such, traditional structural approaches have led to conflicting models for the pathway of Env recognition by CD4 and co-receptors, and subsequent conformational steps along the viral entry pathway. Here the authors continue their elegant single-molecule FRET experiments applied to the Env protein on intact virions. They build on prior FRET measurements to assign the nature of the three distinct FRET states observed previously. They use clever choices of mutants, and ligand complexes to illuminate the conformational states of Env, and provide a low resolution cryo-EM structure that they claim supports their key observation – that Env opens through a key asymmetric intermediate involving changes in a single protomer.

The biophysical studies are an elegant mixture of single-molecule biophysics and challenging biochemistry/virology that deserves publication in eLife after changes that would greatly improve the presentation. However, our recommendation is that the EM structure should either be removed from this paper or extensively revised, which would require additional refinement and analyses of the structure.

We have the following suggestions for revising the smFRET part of the paper.

1) The authors use a challenging biochemical approach using orthogonal peptide tagging to allow double labeling of the individual protomers, combined with careful titration of unlabeled Env through co-expression. They have presented this approach previously, but it would have been useful for readers to know the extent of labeling efficiency for the virions. Also as described in the first paragraph of the subsection “State 3 corresponds to the gp120 conformation of the three-CD4-bound HIV-1 Env trimer”, the authors validate that insertion of the Q3 and A1 tags for fluorophore labeling did not affect Env functions using various assays, but it is not clear whether they tested labeled viruses in these assays, or viruses with just the peptide insertions. The peptide tag is just one component, and perhaps not the largest one, that distinguishes an unlabeled Env from a labeled Env. If it is not possible to do the validation assays using labeled viruses, then the authors should discuss caveats associated with possible steric occlusion from adding the dyes and whatever other components must be added to attach dyes to the peptide tags.

2) The use of a dodecameric CD4 oligomer (sCD4D1D2-Igαtp) for capturing state 3 in NL4-3 is interesting and promotes different effects than sCD4 in the 3 trimers analyzed by smFRET. It is unclear, however, if the "high local CD4 density" resulting from the architecture of this dodecameric CD4 oligomer would geometrically allow multiple copies of CD4 to engage a single Env? A cartoon showing the structure of the construct used in these experiments demonstrating this is feasible would address this and support the results presented. Figure 1D indicates that sCD4D1D2-Igαtp stabilizes state 3 for HIV-1(NL4-3), which corresponds to the open conformation with 3 CD4 bound, whereas the authors also claim that a single CD4 bound to the mixed trimer 1 leads to the same result (Figure 2D). How can these results be reconciled? And were smFRET studies done using sCD4D1D2-Igαtp on mixed trimer 1 (labeled protomer that can bind CD4 with two protomers that can't bind CD4)? This would allow a direct comparison of sCD4D1D2-Igαtp effects on native HIV-1NL4-3 trimer and mixed HIV-1NL4-3 trimer 1. As shown in Figure 2—figure supplements 3-5, the effects of sCD4D1D2-Igαtp were evaluated for JR-FL and BG505 viruses but not for NL4-3. Minor suggestion: it would be easier to compare if Figure 2—figure supplements 3-5 were combined into one figure.

3) The schematics used in Figure 1, 2 and 3 are confusing. It is hard to see the donor and acceptor, and it's not clear why they should have different FRET values in the different conformational states. It's also hard to see the difference between closed and open on such a small figure. Maybe using axial lines for the 3 protomers, and making the dyes bigger (and their positions different in the states) would help. These are the key figures for understanding all the single-molecule data, so the authors should spend more time on presentation. In particular, in Figure 1K, the donor and acceptor dyes are shown as being the same distance apart whether they are on the closed or the open trimer. This is confusing, since the dyes would be separated by long distances in the low FRET State 1 but by shorter distances in the intermediate and high FRET states 3 and 2. There is little discussion in this paper as to whether it makes structural sense for the dyes to become closer together upon CD4 binding, which is especially important if the authors want to include the cryo-EM structure of Env in the same paper. In particular, this sentence, "Binding of a single CD4 may loosen interaction of the V1/V2 loops in the trimer association domain so that neighboring protomers can adopt a conformation in which the V1 and V4 loops are closer to each other." Is this consistent with what is known about the relative locations of V1 and V4 in closed and open Env structures?

4) Figure 3I. Why is 17b shown as binding to the protomers in a mixed trimer that can't bind CD4? 17b won't bind to most Envs in the absence of CD4. It seems reasonable that 17b could bind to the protomer(s) that can bind CD4 in a mixed trimer, but what evidence is there that 17b can also bind to the CD4-incompetent protomers in a mixed trimer?

5) The authors should present prior measurements on CD4 affinity/kinetics wherever possible.

6) Figure 1 legend. Why are the authors using standard errors here instead of standard deviations as they used for other error analyses?

7) For FRET traces, it would be nice to show with lines the high, medium, and low states in each trace.

8) Materials and methods subsection “Preparation of labeled virions”. The protocol for use of the enzyme AcpS for labeling of one of the peptide tags should be cited.

9)Discussion, fifth paragraph. The authors say they have compared smFRET values for labeled BG505 virus and labeled BG505 SOSIP in a paper cited as Lu, Ma et al., 2017, under review. As the current submitted paper includes smFRET data for BG505 virus and an EM structure of a BG505 SOSIP, the paper under review is directly relevant to evaluating the submitted paper and should be given to the reviewers.

Cryo-EM structure

The structure itself (both results and methods) is problematic, and there is little effort to correlate the EM and smFRET results. We can't figure out why the authors think that the EM structure validates the smFRET results because the EM methods are inadequately described and/or possibly done incorrectly, the figures are poor quality, and the description of the structure is vague. We think this paper would stand alone as an important contribution without the EM structure, but if the authors want to include the EM, the following issues must be addressed before publication in any journal:

1) The Materials and methods section describing the EM procedures lacks a lot of information. How many particles were collected for the untilted and for the tilted data sets? How many were kept from each dataset after each step of classification? How many frames were collected in each movie? How many classes were used during 3D classification? How was CTF correction achieved for the two datasets; in particular, how was this done for the tilted data? Was dose-weighting used with motioncor2? What was the total exposure time and how many sub-frames was the total dosage divided into? The authors should use the ab initio model as a reference to do 3D classification, and detailed information about 3D classes should be included in the Methods and also shown in a supplementary figure. Additionally, since 2D/3D classification during cryoEM processing often leads to focusing on the "best" class(es), the authors should address whether there are other conformation(s) present in the dataset.

2) In Figure 4—figure supplement 1C. The authors only show 4 representative 2D classes (all of which look bad and appear to exhibit orientation bias, but this may be due to low resolution of the figure). Did they only use the particles belonging to these 4 classes as input for ab initio model generation? The authors should describe more details about the ab initio model generation input parameters for cryoSPARC in the Materials and methods section, they should use the ab initio model as a reference to do 3D classification, and detailed information about 3D classes should be included in the Materials and methods and also shown in a supplementary figure. Figure 4—figure supplement 4F should show the entirety of gp41 and gp120 rigid-body docked into the EM density, rather than just representative segments. Figure 4—figure supplement 1E is labeled as Map-to-model FSC but the legends states it is the gold standard FSC from refinement. It can only be one of the two.

3) What model was used as a reference for 3D refinement and was the reference model low pass filtered? The authors should verify that they are using the gold-standard FSC resolution estimation that was generated after post processing because this step would generate a more reliable resolution estimation. This information is especially important to know in the case of this structure because the map density and fitting in Figure 4—figure supplement 1F look worse than an 8Å resolution structure. Indeed, the resolution claim of 8Å seems optimistic since the gp41 inner helices are likely the best resolved part of the map and the one the authors chose to show. As a result, the reader cannot judge how confident the authors can be about the different CD4 orientation detected compared to the higher resolution 3 CD4-bound structure. This is further confounded by the strong preferred orientation of the sample in vitreous ice where each 2D class appears to be of the same orientation.

4) Materials and methods. We don't understand this sentence: "Atomic models of gp120 in the CD4-bound conformation, and 4-domain CD4, were fitted into the cryo-EM density by rigid body docking…" This implies that CD4-bound gp120 coordinates were fit into all three protomers in the cryo-EM map, but only one protomer is bound to CD4 in this structure. Also, the authors should discuss whether all four domains of CD4 were ordered in the density. It seems unlikely that D3 and D4 were as well-ordered as D1-D2 since CD4 exhibits some flexibility at the D2-D3 interface. Indeed, D2 of CD4 was not well-resolved in the cryo-EM structures of CD4-bound Env cited by the authors. How CD4 was fit is important because in Figure 4C, the authors used the fitted model to describe a CD4 D1-D2 domain orientation change. However, since full-length sCD4 is presumably flexible, this calculation would be unreliable if based on fitting all four domains. The authors should try lowering the contour level of the map and fitting only D1D2 or maybe only D1 and then do the comparison. Finally PDB codes should be cited for all coordinates used.

5) In order to publish this structure, the authors must show a figure with local resolutions plotted onto the structure. Without this, it is impossible to interpret sentences like this: "…the two adjacent ligand-free gp120 protomers are much more disordered than the CD4-bound protomer. At this resolution, the precise conformational state of the gp120 protomers could not be resolved." First, it's not clear what the resolutions of the ligand-free gp120 protomers actually are (need to see the requested figure to assess). Second and most importantly, if the precise conformational state of the ligand-free gp120 protomers cannot be resolved, then how is the cryo-EM structure relevant to interpreting the smFRET results?

6) The quality of Figure 4 is poor and it was difficult/impossible for all three reviewers to confirm the claimed asymmetry of the Env trimer. Additionally, it is also very difficult to see any of the proposed conformational changes on panels C (right) and D.

7) If the authors decide to retain the cryo-EM structure in this paper, they should provide the map mrc file as well as fitted coordinates file for the reviewers to better judge the cryo-EM map and the model building.

8) "The model of a gradual opening through asymmetric intermediates is supported by cryo-EM that depicts a single CD4 molecule bound to the DS-SOSIP.664 that cannot open, and the current structure with SOSIP.664 that can open in response to CD4…" We don't follow this logic. The existence of the previously published DS-SOSIP/single CD4 cryo-EM structure that is closed is not relevant to this model because the DS-SOSIP is prevented from opening due to introduction of a S-S bond, therefore what is the evidence that an Env without the added DS disulfide bond would remain closed?

9) Related to the point 8 above, Figure 4B shows a hypothetical progression of Env conformations from closed, to a single CD4-bound closed Env (the DS-SOSIP/single CD4 structure), to the current structure (described in the figure as single CD4-bound open structure), to the published 3 CD4-bound, open Env structure. This figure is misleading because it implies that the closed DS-SOSIP/single CD4 structure is a relevant intermediate (see above) and that all three protomers in the current structure are open. The current structure is said to "open in response to CD4" – is that what this EM structure shows? We can't tell from the figures. Figure 4B describes the current structure as a "single CD4-bound open structure" which might be interpreted as meaning that all three protomers adopt the open conformation. But in the Results section, the structure is referred to as an "asymmetric, open structure" which might mean that only the CD4-bound gp120 adopts the open conformation. "…the CD4 molecule is bound solely to one protomer in an open trimer configuration…" and later "…the current structure where the SOSIP.664 partially opens in response to binding of a single CD4…" These examples show that the text is confusing and possibly contradictory. Is the Env trimer in this structure "open" or is it "partially open"? Is the structure asymmetric or symmetric? The figures can't be used to address these questions, and since we don't know how disordered the non CD4-bound protomers are (point 9 above), the EM structure detracts from this paper.

eLife. 2018 Mar 21;7:e34271. doi: 10.7554/eLife.34271.022

Author response

The biophysical studies are an elegant mixture of single-molecule biophysics and challenging biochemistry/virology that deserves publication in eLife after changes that would greatly improve the presentation. However, our recommendation is that the EM structure should either be removed from this paper or extensively revised, which would require additional refinement and analyses of the structure.

We like to note up front that Peter Kwong and Priyamvada Acharya have agreed to remove the EM structure from the manuscript. Peter Kwong and Jason Gorman have supported our project on other aspects, and thus will remain authors. However, all authors involved in the structural work have agreed to be removed from the authors list, including Priyamvada Acharya.

We have the following suggestions for revising the smFRET part of the paper.

1) The authors use a challenging biochemical approach using orthogonal peptide tagging to allow double labeling of the individual protomers, combined with careful titration of unlabeled Env through co-expression. They have presented this approach previously, but it would have been useful for readers to know the extent of labeling efficiency for the virions. Also as described in the first paragraph of the subsection “State 3 corresponds to the gp120 conformation of the three-CD4-bound HIV-1 Env trimer”, the authors validate that insertion of the Q3 and A1 tags for fluorophore labeling did not affect Env functions using various assays, but it is not clear whether they tested labeled viruses in these assays, or viruses with just the peptide insertions. The peptide tag is just one component, and perhaps not the largest one, that distinguishes an unlabeled Env from a labeled Env. If it is not possible to do the validation assays using labeled viruses, then the authors should discuss caveats associated with possible steric occlusion from adding the dyes and whatever other components must be added to attach dyes to the peptide tags.

Each new HIV-1 isolate that is being introduced for smFRET imaging was carefully validated as shown in this manuscript for the BG505. This is done with the 100% tagged Env.

With respect to labeling efficiencies, we have previously determined the labeling efficiencies and found them to be 40% for the Q3 tag and 55% for the A1 tag (Munro et al., 2014). We are still using the same protocols and have not observed a drop in the number of labeled particles in each preparation. Given that labeling efficiencies are not 100%, it’s not valid to test the infectivity of “100% labeled” virus since there are several trimers on the surface of the virus and incomplete labeling leaves enough unlabeled Env to maintain similar infectivity as WT. Thus, the test of Q3 and A1 peptide insertion on virus infectivity was done solely on viruses carrying 100% labeling peptides, but not on fluorophore-labeled viruses. We now mention this in the last paragraph of the Materials and methods subsection “Preparation of labeled virions”. While dyes could affect the outcome, which we acknowledge in the last paragraph of the Materials and methods subsection “Infectivity measurements”, we have not observed anisotropy for singly labeled virions suggesting no trapping of dyes in specific environments. These dyes are highly hydrophilic (Zheng et al., 2014). They also do not interact with the viral membrane. The association of dyes with virions is entirely dependent on the presence of labeling tag (Munro et al., 2014).

For us most important have been the biological controls. The smFRET signal has been responsive to ligands, mutations in a manner that correlates with virological data (Munro et al., 2014; this paper; Herschhorn et al., 2016). Moreover, the Tier 2 viruses JR-FL and BG505 are more closed and less responsive to sCD4 than the Tier 1 lab-adapted HIV-1 isolate NL4-3 (Figures 1B, 1E and 1H).

2) The use of a dodecameric CD4 oligomer (sCD4D1D2-Igαtp) for capturing state 3 in NL4-3 is interesting and promotes different effects than sCD4 in the 3 trimers analyzed by smFRET. It is unclear, however, if the "high local CD4 density" resulting from the architecture of this dodecameric CD4 oligomer would geometrically allow multiple copies of CD4 to engage a single Env? A cartoon showing the structure of the construct used in these experiments demonstrating this is feasible would address this and support the results presented. Figure 1D indicates that sCD4D1D2-Igαtp stabilizes state 3 for HIV-1(NL4-3), which corresponds to the open conformation with 3 CD4 bound, whereas the authors also claim that a single CD4 bound to the mixed trimer 1 leads to the same result (Figure 2D). How can these results be reconciled? And were smFRET studies done using sCD4D1D2-Igαtp on mixed trimer 1 (labeled protomer that can bind CD4 with two protomers that can't bind CD4)? This would allow a direct comparison of sCD4D1D2-Igαtp effects on native HIV-1NL4-3 trimer and mixed HIV-1NL4-3 trimer 1. As shown in Figure 2—figure supplements 3-5, the effects of sCD4D1D2-Igαtp were evaluated for JR-FL and BG505 viruses but not for NL4-3. Minor suggestion: it would be easier to compare if Figure 2—figure supplements 3-5 were combined into one figure.

The dodecameric CD4 oligomer was generated and extensively studied by the group of James Arthos (Bennet et al., 2007; Arthos et al., 2002). The interaction of the sCD4D1D2-Igαtp with native virions has been directly visualized by Sriram Subramaniam using cryo-electron tomography. 12xCD4 was found to cover several gp120 subunits of the same trimer as well as neighboring trimers highlighting the strong avidity (Bennet et al., 2007). The off-rate for sCD4D1D2-Igαtp from viruses is basically zero (Arthos et al., 2002).

In Figure 2D, however, only one protomer in a trimer is capable of binding to CD4. This therefore leads to a single-CD4-bound asymmetric trimer, regardless of whether the ligand is a single D1D2 CD4 molecule or sCD4D1D2-Igαtp. Viruses carrying the D368R mutation are also resistant to sCD4D1D2-Igαtp (Figure 2—figure supplement 3). Thus, upon CD4 binding, this single gp120 adopts the conformation of State 3 that corresponds to the CD4-bound conformation.

The reason why we use either sCD4 or sCD4D1D2-Igαtp is because the Tier 1 lab-adapted HIV-1 isolate is highly responsive and sensitive to sCD4, whereas the Tier 2 viruses JR-FL and BG505 are not. To trigger the CD4 bound conformation, we have to use the more potent ligand 12xCD4 for both Tier 2 isolates. If we use sCD4D1D2-Igαtp on NL4-3, it would be more than 1000x above the inhibitory concentration. We are trying to image all ligands at ~10x above the IC90. It wouldn’t be wise to test one ligand at ~10x above an IC90 and another at 1000x above the IC90.

Upon request, we have combined Figure 2—figure supplements 3-5 into one single Figure 2—figure supplement 3 for better comparison. And have added the figure of 100% D368R NL4-3 in Figure 2—figure supplement 3A.

3) The schematics used in Figure 1, 2 and 3 are confusing. It is hard to see the donor and acceptor, and it's not clear why they should have different FRET values in the different conformational states. It's also hard to see the difference between closed and open on such a small figure. Maybe using axial lines for the 3 protomers, and making the dyes bigger (and their positions different in the states) would help. These are the key figures for understanding all the single-molecule data, so the authors should spend more time on presentation. In particular, in Figure 1K, the donor and acceptor dyes are shown as being the same distance apart whether they are on the closed or the open trimer. This is confusing, since the dyes would be separated by long distances in the low FRET State 1 but by shorter distances in the intermediate and high FRET states 3 and 2. There is little discussion in this paper as to whether it makes structural sense for the dyes to become closer together upon CD4 binding, which is especially important if the authors want to include the cryo-EM structure of Env in the same paper. In particular, this sentence, "Binding of a single CD4 may loosen interaction of the V1/V2 loops in the trimer association domain so that neighboring protomers can adopt a conformation in which the V1 and V4 loops are closer to each other." Is this consistent with what is known about the relative locations of V1 and V4 in closed and open Env structures?

We have modified the figures to schematically illustrate the changes in the dye distance. We can, however, not relate the FRET data to current structural models as it is not trivial and requires the determination of the FRET values observed in Env protein complexes characterized structurally, which requires a separate study. For this reason, we have also removed the EM structure from this manuscript.

4) Figure 3I. Why is 17b shown as binding to the protomers in a mixed trimer that can't bind CD4? 17b won't bind to most Envs in the absence of CD4. It seems reasonable that 17b could bind to the protomer(s) that can bind CD4 in a mixed trimer, but what evidence is there that 17b can also bind to the CD4-incompetent protomers in a mixed trimer?

Because the single protomer that engages CD4 in these asymmetric trimers is already in State 3 and there is no difference between the CD4-bound and the CD4/17b bound gp120 conformation, neither in our smFRET assay nor structurally in gp120 or in the SOSIP trimer in Ozorowski et al., 2017). Thus, 17b likely binds in trans to the neighboring protomers. Mechanistically 17b likely binds by capturing a preexisting State 3 conformation that is more frequently sampled in State 2 as compared to the more closed State 1. However, this is rather speculative. It could in principal be tested experimentally in mixed trimers by combining the D368R mutation with mutations that prevent 17b binding, but introducing several mutations into Env can lead to non-linear effects and phenotypes are increasingly difficult to interpret. We have therefore decided to stay away from this speculation.

Additional evidence can be found in previous publications such as (Herschhorn et al., 2016) demonstrating that Env mutants residing in State 2 more readily engage coreceptors, leading toward downstream conformations.

We have previously shown in the 2014 Munro paper that the frequently opening Tier 1 lab-adapted NL4-3 can open in response to 17b alone, while the more closed Tier 2 JR-FL needs CD4.

5) The authors should present prior measurements on CD4 affinity/kinetics wherever possible.

We are working with native virions and measure infectivity, not affinities for recombinant proteins. We present the neutralization curves for CD4 and sCD4D1D2-Igαtp for all three HIV-1 isolates (Figures 2A, 2F and 2I; Figure 1—figure supplement 3). We now include a table with the calculated IC50 (Table 2), but we cannot include affinities.

6) Figure 1 legend. Why are the authors using standard errors here instead of standard deviations as they used for other error analyses?

Standard error= standard deviation/ (square root of sample size). We chose standard error for the estimation of data quality of histograms because it is more important for such a big number of data points that both reflect the mean and the accuracy of mean, which standard error takes into account. While for neutralization curves, there are only about 10 data points for each sample mean, therefore it is more important to present how each individual data point is different from the mean, which is reflected by standard deviation.

7) For FRET traces, it would be nice to show with lines the high, medium, and low states in each trace.

We have added lines indicating Low-, Intermediate- and High-FRET states in the FRET trace with Hidden Markov Modeling idealization (Figure 1A).

8) Materials and methods subsection “Preparation of labeled virions”. The protocol for use of the enzyme AcpS for labeling of one of the peptide tags should be cited.

We have added lines indicating Low-, Intermediate- and High-FRET states in the FRET trace with Hidden Markov Modeling idealization (Figure 1A).

9)Discussion, fifth paragraph. The authors say they have compared smFRET values for labeled BG505 virus and labeled BG505 SOSIP in a paper cited as Lu, Ma et al., 2017, under review. As the current submitted paper includes smFRET data for BG505 virus and an EM structure of a BG505 SOSIP, the paper under review is directly relevant to evaluating the submitted paper and should be given to the reviewers.

We have removed any reference to the Lu manuscript since the EM structure has been removed from this manuscript.

Cryo-EM structure

The structure itself (both results and methods) is problematic, and there is little effort to correlate the EM and smFRET results. We can't figure out why the authors think that the EM structure validates the smFRET results because the EM methods are inadequately described and/or possibly done incorrectly, the figures are poor quality, and the description of the structure is vague. We think this paper would stand alone as an important contribution without the EM structure, but if the authors want to include the EM, the following issues must be addressed before publication in any journal.

We have removed the EM structure from the manuscript.

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    Transparent reporting form
    elife-34271-transrepform.docx (244.7KB, docx)
    DOI: 10.7554/eLife.34271.019

    Articles from eLife are provided here courtesy of eLife Sciences Publications, Ltd

    RESOURCES