Field-Based Affinity Optimization of a Novel Azabicyclohexane Scaffold HIV-1 Entry Inhibitor

Small-molecule HIV-1 entry inhibitors are an extremely attractive therapeutic modality. We have previously demonstrated that the entry inhibitor class can be optimized by using computational means to identify and extend the chemotypes available. Here we demonstrate unique and differential effects of previously published antiviral compounds on the gross structure of the HIV-1 Env complex, with an azabicyclohexane scaffolded inhibitor having a positive effect on glycoprotein thermostability. We demonstrate that modification of the methyltriazole-azaindole headgroup of these entry inhibitors directly effects the potency of the compounds, and substitution of the methyltriazole with an amine-oxadiazole increases the affinity of the compound 1000-fold over parental by improving the on-rate kinetic parameter. These findings support the continuing exploration of compounds that shift the conformational equilibrium of HIV-1 Env as a novel strategy to improve future inhibitor and vaccine design efforts.


Introduction
The HIV-1 Env complex, the sole viral protein on the outer surface of the virion, is the main focus of antigen design for antibody-based vaccines and small molecule entry inhibitors [1]. The Env complex itself is a trimer of heterodimeric subunits, gp120 and gp41, and orchestrates the series of events that allow deposition of the viral contents into the host cell, making it a primary determinant of viral infectivity. Over recent years, several structures of recombinant Env trimers have been determined by cryo-electron microscopy and x-ray crystallography, including membrane-extracted trimers and engineered soluble trimer mimics [2][3][4][5][6][7][8][9][10][11]. These structures have revealed many insights about the structural features that govern sensitivity and resistance to neutralization of HIV-1 by antibodies and small molecules.
A number of academic groups and pharmaceutical companies are investing time and effort into the development of small molecule inhibitors of the HIV-1 entry process. Encouraging results have been obtained in one entry inhibitor group, originally developed by Bristol-Myers Squibb, and currently being further developed by ViiV Healthcare [12]. Phase III trials of BMS-663068 (or fostemsavir as it has been dubbed) are ongoing. However, all results from these studies on BMS-663068 prodrug entry inhibitor indicate that it would only be of utility as salvage therapy for highly treatment-experienced patients.
Given the enormous potential of HIV-1 entry inhibitors as both pre-and post-exposure prophylactic regimens, our group has been exploring the modification of piperazine scaffold entry inhibitors [12][13][14][15][16][17][18][19]. We have successfully expanded, designed and tested a number of scaffolds other than piperazine, developing compounds with nanomolar potency and specificity to HIV-1 Env [16,20,21]. Moreover, we have recently demonstrated by using a novel surface plasmon resonance (SPR) interaction assay, that the off-rate of the compounds for a soluble recombinant Env trimer is strongly correlated with the potency of the compounds in the single round infection assay [20]. The results presented herein further extends our work developing novel entry inhibitors by demonstrating the conformational effects of previously disclosed entry inhibitors on the Env complex. Moreover, we demonstrate the affinity optimization of one of these scaffolds, the azabicyclohexane scaffold, which has the most pronounced effect on the conformation of the Env target. The results from this work have broad implications for future HIV-1 entry inhibitor designs that capitalize on the malleability and metastability of the Env protein complex.

Results and Discussion
To gain insight into optimization strategies and mechanisms of action of a selection of our entry inhibitor compounds with differing core scaffolds, we wished to perform molecular docking against the published structure of the HIV-1 SOSIP.664 gp140 trimer. First, however, we had to confirm that the compounds actually do directly interact with the Env complex. We have previously demonstrated the interaction of four of our HIV-1 entry inhibitors with the B41 SOSIP.664 gp140 trimer using surface plasmon resonance [20]. Therefore, as a precursor to molecular modeling, we performed SPR interaction analysis upon SC11 (dipyrrolidine scaffold), SC15 (azetidine scaffold), SC28 (azabicyclo-hexane scaffold), and SC45 (tetrahydropyridine scaffold) ( Figure 1). Figure 2 shows the representative sensograms for each compound interacting with the B41 SOSIP.664 gp140 trimer and Table 1 shows the kinetic parameters obtained after analysis. Given the enormous potential of HIV-1 entry inhibitors as both pre-and post-exposure prophylactic regimens, our group has been exploring the modification of piperazine scaffold entry inhibitors [12][13][14][15][16][17][18][19]. We have successfully expanded, designed and tested a number of scaffolds other than piperazine, developing compounds with nanomolar potency and specificity to HIV-1 Env [16,20,21]. Moreover, we have recently demonstrated by using a novel surface plasmon resonance (SPR) interaction assay, that the off-rate of the compounds for a soluble recombinant Env trimer is strongly correlated with the potency of the compounds in the single round infection assay [20]. The results presented herein further extends our work developing novel entry inhibitors by demonstrating the conformational effects of previously disclosed entry inhibitors on the Env complex. Moreover, we demonstrate the affinity optimization of one of these scaffolds, the azabicyclohexane scaffold, which has the most pronounced effect on the conformation of the Env target. The results from this work have broad implications for future HIV-1 entry inhibitor designs that capitalize on the malleability and metastability of the Env protein complex.

Results and Discussion
To gain insight into optimization strategies and mechanisms of action of a selection of our entry inhibitor compounds with differing core scaffolds, we wished to perform molecular docking against the published structure of the HIV-1 SOSIP.664 gp140 trimer. First, however, we had to confirm that the compounds actually do directly interact with the Env complex. We have previously demonstrated the interaction of four of our HIV-1 entry inhibitors with the B41 SOSIP.664 gp140 trimer using surface plasmon resonance. [20] Therefore, as a precursor to molecular modeling, we performed SPR interaction analysis upon SC11 (dipyrrolidine scaffold), SC15 (azetidine scaffold), SC28 (azabicyclohexane scaffold), and SC45 (tetrahydropyridine scaffold) ( Figure 1). Figure 2 shows the representative sensograms for each compound interacting with the B41 SOSIP.664 gp140 trimer and Table 1 shows the kinetic parameters obtained after analysis.  . Colored lines represent actual data collected from the dilution series, whereas black lines signify the fits to a 1:1 binding model. Interaction parameters derived from 5 sets of data are given in Table 1. 3.83 ± 1.12 × 10 3 5.02 ± 2.67 × 10 −4 0.131 µM SC15 3.01 ± 0.188 × 10 5 5.44 ± 0.677 × 10 −3 0.0181 µM SC28 1.39 ± 0.14 × 10 4 6.99 ± 0.43 × 10 −3 0.511 µM SC45 1.38 ± 0.045 × 10 4 1.52 ± 0.0294 × 10 −2 1.10 µM As predicted, each of the compounds interacted robustly with the B41 SOSIP.664 gp140 trimer, albeit with differing kinetics, providing a clear basis for computational docking of the inhibitors into the structure of soluble Env trimer. Therefore, we docked all of the compounds onto the recently published B41 SOSIP Env structure in complex with BMS-386150, using the binding pocket of the cocrystalized ligand (PDB code: 6MUG) [22] to examine potential differences in binding orientation. Interestingly, compounds SC11, SC15 and SC45 were easily docked onto the binding site of B41.SOSIP structure, similar to the co-crystalized ligand BMS-386150 (Figure 3), without any need to perform flexible docking. However, the azabicyclo-hexane compound SC28 did not successfully dock using the previous rigid docking protocol, and instead, we had to use flexible protein-ligand docking in order to achieve a plausible model of SC28 at the binding site. This need for flexibility in the docking protocol was subsequently found to be due to the fact that the presence of the azabicyclohexane moiety, in order to dock to the binding pocket, appears to induce changes in the orientations of certain pocket side chains, especially in the β20/21 loop residue W427, and the α1 helix residue W112 (Figures 3 and 4). It is well documented that changes or interactions at one site can have profound . Colored lines represent actual data collected from the dilution series, whereas black lines signify the fits to a 1:1 binding model. Interaction parameters derived from 5 sets of data are given in Table 1. 3.01 ± 0.188 × 10 5 5.44 ± 0.677 × 10 −3 0.0181 µM SC28 1.39 ± 0.14 × 10 4 6.99 ± 0.43 × 10 −3 0.511 µM SC45 1.38 ± 0.045 × 10 4 1.52 ± 0.0294 × 10 −2 1.10 µM As predicted, each of the compounds interacted robustly with the B41 SOSIP.664 gp140 trimer, albeit with differing kinetics, providing a clear basis for computational docking of the inhibitors into the structure of soluble Env trimer. Therefore, we docked all of the compounds onto the recently published B41 SOSIP Env structure in complex with BMS-386150, using the binding pocket of the co-crystalized ligand (PDB code: 6MUG) [22] to examine potential differences in binding orientation. Interestingly, compounds SC11, SC15 and SC45 were easily docked onto the binding site of B41.SOSIP structure, similar to the co-crystalized ligand BMS-386150 (Figure 3), without any need to perform flexible docking. However, the azabicyclo-hexane compound SC28 did not successfully dock using the previous rigid docking protocol, and instead, we had to use flexible protein-ligand docking in order to achieve a plausible model of SC28 at the binding site. This need for flexibility in the docking protocol was subsequently found to be due to the fact that the presence of the azabicyclo-hexane moiety, in order to dock to the binding pocket, appears to induce changes in the orientations of certain pocket side chains, especially in the β 20/21 loop residue W427, and the α1 helix residue W112 (Figures 3 and 4). It is well documented that changes or interactions at one site can have profound effects upon the overall structure and conformation of the Env complex. Moreover, it has been demonstrated recently that the β 20/21 loop is a regulator of conformational transitions in HIV-1 Env and it is possible that modulation of this region by SC28 could induce gross changes in structure [23,24].
Molecules 2018, 23, x 4 of 30 effects upon the overall structure and conformation of the Env complex. Moreover, it has been demonstrated recently that the β20/21 loop is a regulator of conformational transitions in HIV-1 Env and it is possible that modulation of this region by SC28 could induce gross changes in structure [23,24].  Based on the SPR and docking results, and specifically the participation of the β20/21 loop in the binding of SC28, we decided to test the effects of four of the compounds on the overall structure of B41 SOSIP.664 using negative-stain electron microscopy (NS-EM). B41 SOSIP.664 has been reported to be a strong immunogen like the "gold standard" BG505 SOSIP.664, with most broadly neutralizing epitopes presented on its surface, yet its appearance by NS-EM suggests a much greater degree of movement of gp120 subunits relative to the three-fold symmetry axis, a behavior sometimes referred to as trimer "breathing" [25][26][27]. When no compound is included, a small population (about 20-25%) of imaged B41 trimers adopted what we infer to be a more open conformation, while the remaining trimers had a more tightly-packed, closed phenotype ( Figure 5A). The inclusion of a 10-fold molar excess of any of the four compounds appears to drive the equilibrium from open to closed trimers, at least qualitatively. Statistics of NS-EM analysis are summarized in Table S1. To make a better effects upon the overall structure and conformation of the Env complex. Moreover, it has been demonstrated recently that the β20/21 loop is a regulator of conformational transitions in HIV-1 Env and it is possible that modulation of this region by SC28 could induce gross changes in structure [23,24].  Based on the SPR and docking results, and specifically the participation of the β20/21 loop in the binding of SC28, we decided to test the effects of four of the compounds on the overall structure of B41 SOSIP.664 using negative-stain electron microscopy (NS-EM). B41 SOSIP.664 has been reported to be a strong immunogen like the "gold standard" BG505 SOSIP.664, with most broadly neutralizing epitopes presented on its surface, yet its appearance by NS-EM suggests a much greater degree of movement of gp120 subunits relative to the three-fold symmetry axis, a behavior sometimes referred to as trimer "breathing" [25][26][27]. When no compound is included, a small population (about 20-25%) of imaged B41 trimers adopted what we infer to be a more open conformation, while the remaining trimers had a more tightly-packed, closed phenotype ( Figure 5A). The inclusion of a 10-fold molar excess of any of the four compounds appears to drive the equilibrium from open to closed trimers, at least qualitatively. Statistics of NS-EM analysis are summarized in Table S1. To make a better Based on the SPR and docking results, and specifically the participation of the β 20/21 loop in the binding of SC28, we decided to test the effects of four of the compounds on the overall structure of B41 SOSIP.664 using negative-stain electron microscopy (NS-EM). B41 SOSIP.664 has been reported to be a strong immunogen like the "gold standard" BG505 SOSIP.664, with most broadly neutralizing epitopes presented on its surface, yet its appearance by NS-EM suggests a much greater degree of movement of gp120 subunits relative to the three-fold symmetry axis, a behavior sometimes referred to as trimer "breathing" [25][26][27]. When no compound is included, a small population (about 20-25%) of imaged B41 trimers adopted what we infer to be a more open conformation, while the remaining trimers had a more tightly-packed, closed phenotype ( Figure 5A). The inclusion of a 10-fold molar excess of any of the four compounds appears to drive the equilibrium from open to closed trimers, at least qualitatively. Statistics of NS-EM analysis are summarized in Table S1. To make a better assessment on the stabilizing effects of the small molecules on the trimer, we used differential scanning calorimetry (DSC) to measure changes in thermostability upon the addition of small molecules ( Figure 5B). All four compounds had a positive impact on B41 SOSIP.664 trimer thermostability, with SC11 having the largest effect ( Figure 5B). The combined observations of a measured increase in thermostability (DSC) and an inferred shift towards a more compact state (EM) suggested that the compounds are able to halt the natural dynamics of an Env trimer (that are critical for the virus during receptor recognition and host cell fusion), and are in agreement with the role of the β 20/21 loop in stabilization of the ground state of Env [23]. Because SC28 had both experimentally-determined stabilizing effects and in silico suggestions of inducing conformational changes, the results prompted us to search for and design analogues of the azabicyclohexane scaffold with improved affinity.
Molecules 2018, 23, x 5 of 30 assessment on the stabilizing effects of the small molecules on the trimer, we used differential scanning calorimetry (DSC) to measure changes in thermostability upon the addition of small molecules ( Figure 5B). All four compounds had a positive impact on B41 SOSIP.664 trimer thermostability, with SC11 having the largest effect ( Figure 5B). The combined observations of a measured increase in thermostability (DSC) and an inferred shift towards a more compact state (EM) suggested that the compounds are able to halt the natural dynamics of an Env trimer (that are critical for the virus during receptor recognition and host cell fusion), and are in agreement with the role of the β20/21 loop in stabilization of the ground state of Env [23]. Because SC28 had both experimentallydetermined stabilizing effects and in silico suggestions of inducing conformational changes, the results prompted us to search for and design analogues of the azabicyclohexane scaffold with improved affinity. Averaged particles with phenotypes of "breathing" trimers are highlighted with yellow boxes. All samples were stained with 2% (w/v) uranyl formate. (B) Differential scanning calorimetry curves (left) and summary of melting temperatures (Tm) (right). Primary and secondary peaks are those with the highest and second highest intensity (Cp, specific heat capacity), respectively. ΔTm is the relative change from the inhibitor-free control.
In the absence of any experimentally-derived structural information on the bioactive conformation of our inhibitors, we used the docking model of SC28 as described above as an input into Spark (Cresset, UK) to identify nonclassical bioisosteres of the selected region. From the literature available on the piperazine-based entry inhibitors, it is clear that a primary determinant of potency is the methyltriazole-azaindole head group of the compounds. [28][29][30][31] We, therefore, focused on this region, looking for changes suggested by Spark that may modulate either drug-like parameters or potency, and were significantly different from the methyltriazole-azaindole head group of SC28. [16,20,21,32] From this analysis, we docked five compounds onto the B41 SOSIP Env structure (PDB code: 6MUG) [22] using Glide (XP-mode), and looked at the overall ligand-protein interaction energies. Table 2 shows the calculated ligand-protein interaction energies for these compounds in comparison to the parental compound SC28.
Interestingly, all of the azabicyclohexane analogues produced broadly similar overall poses with the distortion of the β20/21 loop region but a range of calculated interaction energies. The two compounds with the greatest interaction energies in this analysis were SC28 and SC56, which differ only in the replacement of the methyltriazole with an amine-oxadiazole. This change only appeared Averaged particles with phenotypes of "breathing" trimers are highlighted with yellow boxes. All samples were stained with 2% (w/v) uranyl formate. (B) Differential scanning calorimetry curves (left) and summary of melting temperatures (T m ) (right). Primary and secondary peaks are those with the highest and second highest intensity (C p , specific heat capacity), respectively. ∆T m is the relative change from the inhibitor-free control.
In the absence of any experimentally-derived structural information on the bioactive conformation of our inhibitors, we used the docking model of SC28 as described above as an input into Spark (Cresset, UK) to identify nonclassical bioisosteres of the selected region. From the literature available on the piperazine-based entry inhibitors, it is clear that a primary determinant of potency is the methyltriazole-azaindole head group of the compounds [28][29][30][31]. We, therefore, focused on this region, looking for changes suggested by Spark that may modulate either drug-like parameters or potency, and were significantly different from the methyltriazole-azaindole head group of SC28. [16,20,21,32] From this analysis, we docked five compounds onto the B41 SOSIP Env structure (PDB code: 6MUG) [22] using Glide (XP-mode), and looked at the overall ligand-protein interaction energies. Table 2 shows the calculated ligand-protein interaction energies for these compounds in comparison to the parental compound SC28. Interestingly, all of the azabicyclohexane analogues produced broadly similar overall poses with the distortion of the β 20/21 loop region but a range of calculated interaction energies. The two compounds with the greatest interaction energies in this analysis were SC28 and SC56, which differ only in the replacement of the methyltriazole with an amine-oxadiazole. This change only appeared to have one large effect upon the binding pocket, with the orientation of the W427 side chain being completely different between the two binding models ( Figure 4). This analysis suggests that SC56 should have greater affinity for the HIV-1 Env than the parental compound, whereas the other modifications run a range of having potentially lower or equal affinities to SC28. We therefore chose to synthesize these compounds in order to test the predictions from the design and docking analyses.
The docking of the new compounds (designated SC49, SC50, SC52, SC55, and SC56) onto the B41 SOSIP Env structure implies a rank order of affinities. Therefore, after synthesis, and prior to antiviral testing, we sought to establish that the new compounds retained the target specificity of the parental SC28 compound and whether this rank order held true. We again chose to demonstrate this via SPR. Figure 6 shows the representative sensorgrams for each compound interacting with the B41 SOSIP.664 gp140 trimer and Table 3 shows the kinetic parameters obtained after analysis.
As can be seen in Table 3, all analogues retained target specificity and interacted with the B41 SOSIP.664 gp140 trimer. A range of affinities were observed between the analogues with a general agreement with the docking results, i.e., SC49 and SC52 having lower affinities relative to SC28. Rather satisfyingly, however, was the finding that SC56 indeed had a greater affinity for the B41 SOSIP.664 gp140 trimer than SC28 with a 1000-fold difference improvement in the K D of SC56 (0.5 nM) in comparison to the parental SC28 (0.5 µM) and similar overall binding responses to SC28 but over a concentration series 1000-fold more dilute. The dissociation rates of SC56 and SC28 are nearly identical indicating that the contributing factor in the kinetics to this increase in affinity is the association rate with the ka parameter of SC56 increasing by a factor of 1000 in comparison to SC28.
After demonstrating that each of the five SC28 derivatives and SC28 interact with the B41 SOSIP.664 gp140 trimer, and that we had greatly improved the affinity in one of the analogues, SC56, we then tested them for potency against HIV-1 using the HIV-1 single round infection assay. In this system, 293T cells are co-transfected with the envelope-deficient HIV-1 NL4-3 vector (pNL4-3-LucR + E − ; a gift of N. R. Landau, New York University), [33] which carries the luciferase-reporter gene; and the envelope expressing vector from the B41 [26], HxBc2 [34], JR-CSF [35,36], JR-FL [37], or YU-2 [38,39] HIV-1 isolates. This co-transfection yields recombinant single-round infectious envelope-pseudotyped luciferase-reporter HIV-1 viruses. The B41, JR-CSF, JR-FL, and YU-2 envelopes were all originally isolated by directly cloning samples from HIV-1 infected patients and therefore were never subjected to the potential selection imposed by passage of the virus in tissue culture. B41, JR-CSF, JR-FL, and YU-2 are all relatively resistant to neutralization by soluble CD4 and antibodies directed against the HIV-1 envelope glycoproteins and are classified as tier 2 isolates that utilize the CCR5-receptor for entry. Therefore they are representative of the clinically most abundant viruses. The use of the HxBc2 reference strain Env, along with the Envs from four primary R5-tropic viruses allows an assessment of the generality of results obtained. The pseudotyped viruses are then used to infect U87.CD4.CCR5 (B41, JR-CSF, JR-FL, and YU-2) or U87.CD4.CXCR4 (HxBc2) target cells in the presence and absence of compounds and infectivity is quantified by measuring luciferase levels in cell lysates (Luciferase Assay System, Promega, Fitchburg, WI, USA) using a microplate luminometer (GloMax, Promega). The toxicity of the compounds was also assessed in parallel as outlined in Materials & Methods. The results of this analysis are shown in Figure S1 and S2 and the values are summarized in Table 4.  Table 2.  . Colored lines represent actual data, whereas black lines indicate the fits to a 1:1 interaction model. Interaction parameters derived from five sets of data are given in Table 2. system, 293T cells are co-transfected with the envelope-deficient HIV-1 NL4-3 vector (pNL4-3-LucR + E − ; a gift of N. R. Landau, New York University), [33] which carries the luciferase-reporter gene; and the envelope expressing vector from the B41 [26], HxBc2 [34], JR-CSF [35,36], JR-FL [37], or YU-2 [38,39] HIV-1 isolates. This co-transfection yields recombinant single-round infectious envelopepseudotyped luciferase-reporter HIV-1 viruses. The B41, JR-CSF, JR-FL, and YU-2 envelopes were all originally isolated by directly cloning samples from HIV-1 infected patients and therefore were never subjected to the potential selection imposed by passage of the virus in tissue culture. B41, JR-CSF, JR-FL, and YU-2 are all relatively resistant to neutralization by soluble CD4 and antibodies directed against the HIV-1 envelope glycoproteins and are classified as tier 2 isolates that utilize the CCR5receptor for entry. Therefore they are representative of the clinically most abundant viruses. The use of the HxBc2 reference strain Env, along with the Envs from four primary R5-tropic viruses allows an assessment of the generality of results obtained. The pseudotyped viruses are then used to infect U87.CD4.CCR5 (B41, JR-CSF, JR-FL, and YU-2) or U87.CD4.CXCR4 (HxBc2) target cells in the presence and absence of compounds and infectivity is quantified by measuring luciferase levels in cell lysates (Luciferase Assay System, Promega, Fitchburg, WI, USA) using a microplate luminometer (GloMax, Promega). The toxicity of the compounds was also assessed in parallel as outlined in Materials & Methods. The results of this analysis are shown in Figure S1 and S2 and the values are summarized in Table 4.
As summarized in Table 4, SC28 and each of the analogues tested exhibited antiviral effects in the single-round infection assay utilizing viruses pseudotyped with HIV-1 Env from either the B41, HxBc2, JR-CSF, JR-FL, or YU-2 isolates. The analogues exhibited a range of potencies, largely in correlation with the docking and the SPR results. However, we found that SC56 despite having a much higher affinity for the B41 SOSIP.664 gp140 trimer than SC28, had a median IC50 only two-fold over that of SC28. We have previously demonstrated that the potency of our entry inhibitors has a direct correlation with the kinetic off-rate parameter. [20] Comparison of the kinetics of SC28 and SC56 shows that the off-rates for their interaction with the B41 SOSIP.664 gp140 trimer are almost identical, providing a rationale for their similar potencies, and further corroborating our previous study.

SC49
we then tested them for potency against HIV-1 using the HIV-1 single round infection assay. In this system, 293T cells are co-transfected with the envelope-deficient HIV-1 NL4-3 vector (pNL4-3-LucR + E − ; a gift of N. R. Landau, New York University), [33] which carries the luciferase-reporter gene; and the envelope expressing vector from the B41 [26], HxBc2 [34], JR-CSF [35,36], JR-FL [37], or YU-2 [38,39] HIV-1 isolates. This co-transfection yields recombinant single-round infectious envelopepseudotyped luciferase-reporter HIV-1 viruses. The B41, JR-CSF, JR-FL, and YU-2 envelopes were all originally isolated by directly cloning samples from HIV-1 infected patients and therefore were never subjected to the potential selection imposed by passage of the virus in tissue culture. B41, JR-CSF, JR-FL, and YU-2 are all relatively resistant to neutralization by soluble CD4 and antibodies directed against the HIV-1 envelope glycoproteins and are classified as tier 2 isolates that utilize the CCR5receptor for entry. Therefore they are representative of the clinically most abundant viruses. The use of the HxBc2 reference strain Env, along with the Envs from four primary R5-tropic viruses allows an assessment of the generality of results obtained. The pseudotyped viruses are then used to infect U87.CD4.CCR5 (B41, JR-CSF, JR-FL, and YU-2) or U87.CD4.CXCR4 (HxBc2) target cells in the presence and absence of compounds and infectivity is quantified by measuring luciferase levels in cell lysates (Luciferase Assay System, Promega, Fitchburg, WI, USA) using a microplate luminometer (GloMax, Promega). The toxicity of the compounds was also assessed in parallel as outlined in Materials & Methods. The results of this analysis are shown in Figure S1 and S2 and the values are summarized in Table 4. As summarized in Table 4, SC28 and each of the analogues tested exhibited antiviral effects in the single-round infection assay utilizing viruses pseudotyped with HIV-1 Env from either the B41, HxBc2, JR-CSF, JR-FL, or YU-2 isolates. The analogues exhibited a range of potencies, largely in correlation with the docking and the SPR results. However, we found that SC56 despite having a much higher affinity for the B41 SOSIP.664 gp140 trimer than SC28, had a median IC50 only two-fold over that of SC28. We have previously demonstrated that the potency of our entry inhibitors has a direct correlation with the kinetic off-rate parameter. [20] Comparison of the kinetics of SC28 and SC56 shows that the off-rates for their interaction with the B41 SOSIP.664 gp140 trimer are almost identical, providing a rationale for their similar potencies, and further corroborating our previous study.       As summarized in Table 4, SC28 and each of the analogues tested exhibited antiviral effects in the single-round infection assay utilizing viruses pseudotyped with HIV-1 Env from either the B41, HxBc2, JR-CSF, JR-FL, or YU-2 isolates. The analogues exhibited a range of potencies, largely in correlation with the docking and the SPR results. However, we found that SC56 despite having a much higher affinity for the B41 SOSIP.664 gp140 trimer than SC28, had a median IC 50 only two-fold over that of

SC28.
We have previously demonstrated that the potency of our entry inhibitors has a direct correlation with the kinetic off-rate parameter [20]. Comparison of the kinetics of SC28 and SC56 shows that the off-rates for their interaction with the B41 SOSIP.664 gp140 trimer are almost identical, providing a rationale for their similar potencies, and further corroborating our previous study.  [21]. SC15, SC28, and SC45 were synthesized as outlined in Tuyishime et al. [16].

Synthesis of SC49
General Procedure for the Preparation of 2 . SC11 was synthesized as outlined in Tuyishime et al. [21]. SC15, SC28, and SC45 were synthesized as outlined in Tuyishime et al. [16].

Synthesis of SC49
General procedure for the preparation of 2 Vinylmagnesium bromide 1a (1 M, 543.04 mL) was cooled below −60 °C with vigorous stirring under N2. A solution of 1 (30 g, 136 mmol) in THF (100 mL) was added dropwise slowly that the temperature was kept below −60 °C. The reaction mixture was warmed to −40 to −50 °C and stirred for an additional 1 h. TLC (petroleum ether: ethyl acetate = 2:1, Rf = 0.47) show that the reaction was complete. Saturated aqueous NH4Cl (200 mL) was added slowly. The layers were separated, and the aqueous layer was extracted with EtOAc (3 × 200 mL). The organic extracts were washed with brine (200 mL), dried over Na2SO4, filtered, and concentrated. To the residue CH2Cl2 (100 mL) was added and the solid formed was collected by filtration and washed with CH2Cl2 (50 mL) to give 2 (8 g, 27% yield) as a brown solid. 1  complete. Saturated aqueous NH4Cl (200 mL) was added slowly. The layers were separated, and the aqueous layer was extracted with EtOAc (3 × 200 mL). The organic extracts were washed with brine (200 mL), dried over Na2SO4, filtered, and concentrated. To the residue CH2Cl2 (100 mL) was added and the solid formed was collected by filtration and washed with CH2Cl2 (50 mL) to give 2 (8 g, 27% yield) as a brown solid. 1  General Procedure for the Preparation of 4 A mixture of 3 (800 mg, 4.96 mmol) in MeOH (10 mL) and conc. HCl (10 mL) was stirred at 90 °C for 16 h. TLC (petroleum ether: ethyl acetate = 2:1, Rf = 0.01) showed that the reaction was complete. The organic solvent was evaporated, and the precipitate was filtered off and dried to give A mixture of 2 (4 g, 18.6 mMol) and CuCN (3.3 g, 37.2 mMol) in DMF (30 mL) was stirred at 150 • C for 1 h, after which TLC (petroleum ether: ethyl acetate = 2: 1, Rf = 0.51) showed reaction completion. The mixture was diluted with EtOAc (100 mL), washed with water (100 mL), brine (100 mL), and concentrated. The residue was purified by column chromatography on silica gel and eluted with petroleum ether: ethyl acetate = 4: 1 to give 3 (900 mg, 30% yield) as a yellow solid. complete. Saturated aqueous NH4Cl (200 mL) was added slowly. The layers were separated, and the aqueous layer was extracted with EtOAc (3 × 200 mL). The organic extracts were washed with brine (200 mL), dried over Na2SO4, filtered, and concentrated. To the residue CH2Cl2 (100 mL) was added and the solid formed was collected by filtration and washed with CH2Cl2 (50 mL) to give 2 (8 g, 27% yield) as a brown solid. 1  General Procedure for the Preparation of 4 A mixture of 3 (800 mg, 4.96 mmol) in MeOH (10 mL) and conc. HCl (10 mL) was stirred at 90 °C for 16 h. TLC (petroleum ether: ethyl acetate = 2:1, Rf = 0.01) showed that the reaction was complete. The organic solvent was evaporated, and the precipitate was filtered off and dried to give A mixture of 3 (800 mg, 4.96 mMol) in MeOH (10 mL) and conc. HCl (10 mL) was stirred at 90 • C for 16 h. TLC (petroleum ether: ethyl acetate = 2:1, Rf = 0.01) showed that the reaction was complete. The organic solvent was evaporated, and the precipitate was filtered off and dried to give 4 (500 mg, 56% yield) as a brown solid. 1  General Procedure for the Preparation of 6 Compound 5 (370 mg, 1.92 mmol) was added to a mixture of AlCl3 (1.53 g, 11.5 mmol) and 1ethyl-3-methylimidazol-3-ium chloride (566 mg, 3.83 mmol). Then, 5a (523 mg, 3.83 mmol) was added slowly to the solution, and the mixture was stirred at 25 °C for 15 h. TLC (ethyl acetate: petroleum ether = 2:1, Rf = 0.01) showed that the conversion was more than 50% and LCMS showed the desired A solution of 4 (450 mg, 2.50 mMol), DIEA (967 mg, 7.49 mMol), and HATU (1.04 g, 2.75 mMol) in THF (10 mL) was stirred at 25 • C for 0.5 h. Then methylamine (675 mg, 9.99 mMol, HCl salt) was added, and the mixture was stirred at 25 • C for 16 h. TLC (dichloromethane: methanol = 20: 1, Rf = 0.65) showed that the reaction was complete. The mixture was diluted with EtOAc (20 mL), washed with water (20 mL), brine (20 mL) and dried over Na 2 SO 4 , and concentrated. The residue was purified by column chromatography on silica gel and eluted with petroleum ether: EtOAc = 5:1 to give 5 (400 mg, 83% yield) as a white solid. 1  General Procedure for the Preparation of 6 Compound 5 (370 mg, 1.92 mmol) was added to a mixture of AlCl3 (1.53 g, 11.5 mmol) and 1ethyl-3-methylimidazol-3-ium chloride (566 mg, 3.83 mmol). Then, 5a (523 mg, 3.83 mmol) was added slowly to the solution, and the mixture was stirred at 25 °C for 15 h. TLC (ethyl acetate: petroleum ether = 2:1, Rf = 0.01) showed that the conversion was more than 50% and LCMS showed the desired mass of the product. Water was added (20 mL) slowly to the mixture at 0 °C. The precipitate was Compound 5 (370 mg, 1.92 mMol) was added to a mixture of AlCl 3 (1.53 g, 11.5 mMol) and 1-ethyl-3-methylimidazol-3-ium chloride (566 mg, 3.83 mMol). Then, 5a (523 mg, 3.83 mMol) was added slowly to the solution, and the mixture was stirred at 25 • C for 15 h. TLC (ethyl acetate: petroleum ether = 2:1, Rf = 0.01) showed that the conversion was more than 50% and LCMS showed the desired mass of the product. Water was added (20 mL) slowly to the mixture at 0 • C. The precipitate was filtered off, and dried to give 6 (200 mg, crude) as a yellow solid.
General Procedure for the Preparation of 7 Compound 5 (370 mg, 1.92 mmol) was added to a mixture of AlCl3 (1.53 g, 11.5 mmol) and 1ethyl-3-methylimidazol-3-ium chloride (566 mg, 3.83 mmol). Then, 5a (523 mg, 3.83 mmol) was added slowly to the solution, and the mixture was stirred at 25 °C for 15 h. TLC (ethyl acetate: petroleum ether = 2:1, Rf = 0.01) showed that the conversion was more than 50% and LCMS showed the desired mass of the product. Water was added (20 mL) slowly to the mixture at 0 °C. The precipitate was filtered off, and dried to give 6 (200 mg, crude) as a yellow solid.

The Synthesis of SC50
General Procedure for the Preparation of 9 8a (163 mg, 1.16 mmol). Then the mixture was stirred at 25 °C for 16 h. LCMS showed that the desired product was produced. The mixture was concentrated. The residue was purified by neutral prep-HPLC to give SC49 (22 mg, 8% yield) as a light yellow solid. 1

The Synthesis of SC50
General Procedure for the Preparation of 9 To a mixture of 1-ethyl-3-methylimidazol-3-ium chloride (4.1 g, 27.9 mmol) was added AlCl3 (11.2 g, 83.7 mmol) in portions at 0 °C and stirred for 30 min, then 2 (3.0 g, 13.9 mmol) was added in portions at 0 °C under N2. The mixture was stirred at 0 °C for 0.5 hour, then ethyl 2-chloro-2-oxoacetate (3.8 g, 27.9 mmol) was added dropwise, then the mixture was stirred at 25 °C for 2 h. TLC (petroleum ether: ethyl acetate = 3: 1, Rf = 0.2) showed the reaction was occurred and one new spot was generated. The mixture was poured onto water (50 mL) and stirred for 10 min. The aqueous phase was extracted with ethyl acetate (50 mL × 3 mL). The combined organic phase was washed with brine (30 mL × 2mL), dried over Na2SO4, filtered and concentrated in vacuum to give 9 (3 g, 68% yield) as a brown solid. 1

General Procedure for the Preparation of 10
To a mixture of 1-ethyl-3-methylimidazol-3-ium chloride (4.1 g, 27.9 mMol) was added AlCl 3 (11.2 g, 83.7 mMol) in portions at 0 • C and stirred for 30 min, then 2 (3.0 g, 13.9 mMol) was added in portions at 0 • C under N 2 . The mixture was stirred at 0 • C for 0.5 h, then ethyl 2-chloro-2-oxo-acetate (3.8 g, 27.9 mMol) was added dropwise, then the mixture was stirred at 25 • C for 2 h. TLC (petroleum ether: ethyl acetate = 3: 1, R f = 0.2) showed the reaction was occurred and one new spot was generated. The mixture was poured onto water (50 mL) and stirred for 10 min. The aqueous phase was extracted with ethyl acetate (50 mL × 3 mL). The combined organic phase was washed with brine (30 mL × 2mL), dried over Na 2 SO 4 , filtered and concentrated in vacuum to give 9 (3 g, 68% yield) as a brown solid. 1  To a mixture of 9 (3 g, 9.5 mmol) in MeOH (30 mL) and H2O (10 mL) was added K2CO3 (2.6 g, 19 mmol) in one portion at 25 °C under N2. The mixture was stirred at 25 °C for 5 h. TLC (petroleum ether:ethyl acetate = 1: 1, Rf = 0.03) showed the reaction was completed. The mixture was concentrated in reduced pressure at 45 °C. The residue was poured onto water (50 mL) and stirred for 10 min. The mixture was acidified with HCl (1 M) to pH ~4, the aqueous phase was extracted with ethyl acetate (50 mL × 3mL). The combined organic phase was washed with brine (30 mL × 2mL), dried with anhydrous Na2SO4, filtered and concentrated in vacuum to give 10 (2.0 g, 73% yield) as a yellow solid.

General Procedure for the Preparation of 11
To a mixture of 10 (900 mg, 3.1 mmol) in DMF (30 mL) was added HATU (1.55 g, 4.08 mmol) and DIEA (1.22 g, 9.4 mmol) in one portion at 25 °C under N2. The mixture was stirred at 25 °C for 30 min, then 6a (745 mg, 3.7 mmol) was added in portions and the mixture was stirred for 5 h. TLC (petroleum ether:ethyl acetate = 1:1, Rf = 0.6.) showed one main spot was generated. The mixture was poured onto water (50 mL) and stirred for 10 min. The aqueous phase was extracted with ethyl acetate (30 mL × 2). The combined organic phase was washed with brine (20 mL × 2), dried with anhydrous Na2SO4, filtered and concentrated in vacuum. The residue was purified by silica gel chromatography (column height: 250 mm, diameter: 100 mm, 100-200 mesh silica gel, petroleum ether:ethyl acetate = 1:1) to give 11 (2.0 g, 68% yield) as a yellow solid.
To a mixture of 9 (3 g, 9.5 mMol) in MeOH (30 mL) and H 2 O (10 mL) was added K 2 CO 3 (2.6 g, 19 mMol) in one portion at 25 • C under N 2 . The mixture was stirred at 25 • C for 5 h. TLC (petroleum ether:ethyl acetate = 1: 1, R f = 0.03) showed the reaction was completed. The mixture was concentrated in reduced pressure at 45 • C. The residue was poured onto water (50 mL) and stirred for 10 min. The mixture was acidified with HCl (1 M) to pH~4, the aqueous phase was extracted with ethyl acetate (50 mL × 3mL). The combined organic phase was washed with brine (30 mL × 2mL), dried with anhydrous Na 2 SO 4 , filtered and concentrated in vacuum to give 10 (2.0 g, 73% yield) as a yellow solid. 1  To a mixture of 9 (3 g, 9.5 mmol) in MeOH (30 mL) and H2O (10 mL) was added K2CO3 (2.6 g, 19 mmol) in one portion at 25 °C under N2. The mixture was stirred at 25 °C for 5 h. TLC (petroleum ether:ethyl acetate = 1: 1, Rf = 0.03) showed the reaction was completed. The mixture was concentrated in reduced pressure at 45 °C. The residue was poured onto water (50 mL) and stirred for 10 min. The mixture was acidified with HCl (1 M) to pH ~4, the aqueous phase was extracted with ethyl acetate (50 mL × 3mL). The combined organic phase was washed with brine (30 mL × 2mL), dried with anhydrous Na2SO4, filtered and concentrated in vacuum to give 10 (2.0 g, 73% yield) as a yellow solid.

General Procedure for the Preparation of 11
To a mixture of 10 (900 mg, 3.1 mmol) in DMF (30 mL) was added HATU (1.55 g, 4.08 mmol) and DIEA (1.22 g, 9.4 mmol) in one portion at 25 °C under N2. The mixture was stirred at 25 °C for 30 min, then 6a (745 mg, 3.7 mmol) was added in portions and the mixture was stirred for 5 h. TLC (petroleum ether:ethyl acetate = 1:1, Rf = 0.6.) showed one main spot was generated. The mixture was poured onto water (50 mL) and stirred for 10 min. The aqueous phase was extracted with ethyl acetate (30 mL × 2). The combined organic phase was washed with brine (20 mL × 2), dried with anhydrous Na2SO4, filtered and concentrated in vacuum. The residue was purified by silica gel chromatography (column height: 250 mm, diameter: 100 mm, 100-200 mesh silica gel, petroleum ether:ethyl acetate = 1:1) to give 11 (2.0 g, 68% yield) as a yellow solid.

General Procedure for the Preparation of 12
To a mixture of 10 (900 mg, 3.1 mmol) in DMF (30 mL) was added HATU (1.55 g, 4.08 mmol) and DIEA (1.22 g, 9.4 mmol) in one portion at 25 °C under N2. The mixture was stirred at 25 °C for 30 min, then 6a (745 mg, 3.7 mmol) was added in portions and the mixture was stirred for 5 h. TLC (petroleum ether:ethyl acetate = 1:1, Rf = 0.6.) showed one main spot was generated. The mixture was poured onto water (50 mL) and stirred for 10 min. The aqueous phase was extracted with ethyl acetate (30 mL × 2). The combined organic phase was washed with brine (20 mL × 2), dried with anhydrous Na2SO4, filtered and concentrated in vacuum. The residue was purified by silica gel chromatography (column height: 250 mm, diameter: 100 mm, 100-200 mesh silica gel, petroleum ether:ethyl acetate = 1:1) to give 11 (2.0 g, 68% yield) as a yellow solid.

General Procedure for the Preparation of 12
A mixture of 11 (2.0 g, 4.28 mmol) in HCl/EtOAc (20 mL) was stirred at 25 °C for 5 h. LCMS showed the reaction was completed, and the desired product was generated. The mixture was filtered and washed with EtOAc, and concentrated in vacuum to give 12 (1.2 g, 76% yield) as a yellow solid.

General Procedure for the Preparation of 13
A mixture of 11 (2.0 g, 4.28 mMol) in HCl/EtOAc (20 mL) was stirred at 25 • C for 5 h. LCMS showed the reaction was completed, and the desired product was generated. The mixture was filtered and washed with EtOAc, and concentrated in vacuum to give 12 (1.2 g, 76% yield) as a yellow solid.  To a solution of 13 (100 mg, 0.212 mMol), 3-pyridylboronic acid (39 mg, 0.318 mMol) in dioxane (5 mL) and H 2 O (1 mL) was added Pd(dppf)Cl 2 (15.5 mg, 21 mMol) and K 2 CO 3 (87.9 mg, 637 mMol) in one portion at 25 • C then heated at 110 • C for 12 h. TLC (dichloromethane:methanol = 10:1, R f = 0.4) showed reaction completion, LCMS showed the desired product was formed. After cooling to 25 • C, ethyl acetate (30 mL) was added and combined and concentrated under reduce pressure. The residue was purified by prep-TLC (dichloromethane:methanol = 10:1, R f = 0.4) to give SC50 (90 mg, 91% yield) as an off-white solid. 1

Preparation of N-(3-Benzoyl-3-azabicyclo[3.1.0]hexan-6-yl)-2-(7-cyano-4-methoxy-1H-indol-3-yl)-2oxoacetamide (38)
At room temperature, to a stirred solution of 37 (500 mg, 2.05 mmol) in DCM (30 mL) were added (6-amino-3-azabicyclo[3.1.0]hexan-3-yl)(phenyl)methanone HCl salt (540 mg, 2.26 mmol) and DIPEA (1 mL, 6.15 mmol), followed by HATU (935 mg, 2.46 mmol) in portions. The resulting mixture was stirred at room temperature for 30 min, and then diluted with DCM (10 mL) and water (20 mL). The organic layer was washed with brine, dried over Na2SO4, filtered and concentrated under reduced pressure. At room temperature, to a solution of 38 (300 mg, 0.7 mmol) in EtOH (10 mL) was added NH2OH (3 mL, 50 wt.% in water). After stirred at room temperature for 30 min, the reaction mixture was partitioned between DCM and water. The organic layer was separated, and the aqueous layer was At room temperature, to a solution of 38 (300 mg, 0.7 mMol) in EtOH (10 mL) was added NH 2 OH (3 mL, 50 wt.% in water). After stirred at room temperature for 30 min, the reaction mixture was partitioned between DCM and water. The organic layer was separated, and the aqueous layer was extracted with DCM. The combined organic layers were dried over Na 2 SO 4 , filtered and concentrated under reduced pressure to give the crude title product 39, which was used in the next step without At room temperature, to a solution of 39 (323.0 mg, 0.7 mmol) in anhydrous THF (10 mL) was added 2,2,2-trichloroacetic anhydride (648.0 mg, 2.1 mmol) dropwise. The resulting mixture was stirred at room temperature overnight. After completion of the reaction as indicated by LC-MS, the reaction was quenched with ice water. The resulting solution was extracted with DCM. The combined organic layers were washed with water, dried over Na2SO4, filtered and concentrated under reduced pressure to give the crude title product 40, which was used in the next step without further At room temperature, to a solution of 39 (323.0 mg, 0.7 mMol) in anhydrous THF (10 mL) was added 2,2,2-trichloroacetic anhydride (648.0 mg, 2.1 mMol) dropwise. The resulting mixture was stirred at room temperature overnight. After completion of the reaction as indicated by LC-MS, the reaction was quenched with ice water. The resulting solution was extracted with DCM. The combined organic layers were washed with water, dried over Na 2 SO 4 , filtered and concentrated under reduced pressure to give the crude title product 40, which was used in the next step without further purification. LC-MS (ESI): m/z [M/M + 1] + = 588.00/590.00.
grid. The drop was blotted using Whatman #1 filter paper, a 3 µL drop of 2% (w/v) uranyl formate was added to the sample-treated surface of the grid, and the sample stained for 45 s before blotting. Samples were imaged using an FEI Talos Arctica electron microscope (Thermo Fisher) operating at 200 kV and an FEI Ceta 16M CMOS camera (Thermo Fisher). Single frame exposures were collected with a total dose of~25 e − /Å 2 at a magnification of 73,000, resulting in a pixel size of 1.98 Å at the specimen plane. Data processing and analysis has been described previously [26,50].

Differential Scanning Aalorimetry (DSC)
Experiments were performed using a MicroCal VP-Capillary differential scanning calorimeter (Malvern Panalytical, Westborough, MA, USA). Before the experiments were carried out, B41 SOSIP.664 was dialyzed against phosphate-buffered saline (PBS). The protein concentration was subsequently adjusted to 0.25 mg/mL. Similar to the EM methods above, 0.5 mL of B41 SOSIP.664 were added to tubes containing 2 µg lyophilized compound SC11, SC15, SC28 or SC45, mixed by pipetting, and allowed to incubate for 1 h at room temperature. The samples, along with a control containing protein alone, were into the instrument cell and thermal denaturation was probed at a scan rate of 90 • C/h, with PBS in the reference cell. Buffer correction, normalization, and baseline subtraction procedures were performed using the Automated Origin 7.0 software (Malvern Panalytical, Westborough, MA, USA). The data were fitted using a non-two-state model.

Docking of Compounds SC11, SC15 and SC45
The Env protein (pdb code: 6MUG) was prepared by the Protein Preparation Wizard implemented with Maestro (Schrödinger Maestro Version 11.5.011, New York, NY, USA, mM share Version 4.1.011, New York, NY, USA, Release 2018-1, Platform Darwin-x86_64). The Grid box was centered on the co-crystalized BMS-386150. For validating the glide docking protocol, the original ligand BMS-386150 was built using the LigPrep tool (Schrödinger Maestro Version 11.5.011, New York, NY, USA) and docked using glide-XP mode. The predicted docked pose matched the original co-ordinates of the co-crystalized BMS-386150 with an RMS value of 0.1 Å. SC11, SC15 and SC45 was docked using the same glide-XP mode and the top ranked pose was selected. The docked protein-ligand complexes were then refined using Prime (VSGB solvation model and OPLS3e forcefield, entire protein refinement).
3.13.2. Docking of Compounds SC28, SC49, SC50, SC52, SC55 and SC56 Glide induced-Fit (extended sampling settings) was used to flexibly sample the ligands and the protein pocket. The top ranked poses were then refined using Prime (VSGB solvation model and OPLS3e forcefield, entire protein refinement).

Conclusions
In this study, we demonstrated that four entry inhibitors with different core scaffolds still interact with the B41 SOSIP.664 gp140 trimer via SPR but with different kinetics. After examining the effects of these four compounds on the overall structure of B41 SOSIP.664 trimer using NS-EM and DSC, we discovered that all of them, including SC28 (azabicyclohexane core), were able to stabilize the SOSIP and the EM results suggested a shift in the conformational equilibrium of the SOSIP from a heterogeneous population of open and closed trimers to a more homogenous pool of closed trimers. The findings suggest that a molecule like SC28 is able to stabilize the SOSIP and halt or severly impair the ability of receptor-mediated conformational dynamics of Env, making it a potential allosteric fusion inhibitor. Using computational field-and structure-based design methods, we chose five analogues of SC28 that had a modified methyltriazole-azaindole head group and retained the azabicyclo-hexane core region. These five SC28 derivatives all retained target specificity to B41 SOSIP.664 gp140 trimers and exhibited antiviral activity. Most striking was the difference between SC28 and SC56, both compounds had nearly identical potencies, but had a 1000-fold difference in affinity. We have previously demonstrated for this class of compounds that the dissociation rate is correlated to the potency and when looking at the contributing factor to this increase in affinity, we observed that both compounds had similar dissociation rates, but that the association rates also differed by a factor of 1000. This further confirms that by decreasing the dissociation rate we can improve the potency for future analogues of our entry inhibitors to help bring this class of inhibitors closer towards clinical utility. By continuing to explore compounds that shift the conformational equilibrium of Env trimers, this class of entry inhibitors will have profound implications for future inhibitor and vaccine design efforts.
Supplementary Materials: The following are available online at, Figure S1: Single round infection assay graphs, Figure S2: Toxicity of compounds, Figure S3: Sensorgram of SC56, Table S1: NS-EM statistics.