|Home | About | Journals | Submit | Contact Us | Français|
Cellular infection by HIV-1 is initiated with a binding event between the viral envelope glycoprotein gp120 and the cellular receptor protein CD4. The CD4:gp120 interface is dominated by two hotspots: a hydrophobic gp120 cavity capped by Phe43CD4 and an electrostatic interaction between residues Arg59CD4 and Asp368gp120. The CD4 mimetic small-molecule NBD-556 (1) binds within the gp120 cavity; however, 1 and related congeners demonstrate limited viral neutralization breadth. Herein, we report the design, synthesis, characterization and structures of gp120 in complex with small-molecules that simultaneously engage both binding hotspots of gp120. The compounds specifically inhibit viral infection of 42 Tier 2 clade B and C viruses and are shown to be antagonists of entry into CD4-negative cells. Dual hotspot design thus provides both a means to enhance neutralization potency of HIV-1 entry inhibitors and a novel structural paradigm for inhibiting the CD4:gp120 protein-protein interaction.
Infection by the HIV-1 virus and the subsequent progression to AIDS1, 2 begins with protein-protein binding events between the trimeric envelope glycoprotein spike (Env) and host cell receptors.3–7 Each trimeric Env spike is composed of three gp120 envelope glycoproteins and three gp41 transmembrane proteins.8–10 The HIV-1 viral entry process begins with two consecutive gp120-protein binding events, each associated with changes in Env conformation.7, 11–13 Attachment of HIV occurs when gp120 binds to the T-cell CD4 receptor3, 14 anchored to the cell membrane. Binding to CD4 then induces a gp120 conformational change,11, 12 which exposes the binding site for the transmembrane chemokine co-receptor (CCR5 or CXCR4).15–17 Once co-receptor binding occurs, gp41 rearranges to form the six-helix bundle which inserts into the host cell membrane, culminating in fusion and viral entry.6, 18–20
Inhibition of the initial entry event of the HIV-1 virus into host cells remains a compelling, yet elusive means to prevent or treat HIV-1 infection and AIDS. Currently, only two therapeutic agents that target the viral entry process, are approved in the U.S. for the treatment of individuals infected with HIV-1: maraviroc,21 a small molecule CCR5-antagonist; and enfuvirtide,22 a 36-amino acid peptide that binds gp41 and prevents formation of the fusion peptide. Indeed, to date, limited progress has been made on developing inhibitors that target the initial step in viral entry, CD4:gp120 binding.23–26 Hence, disruption of the earliest CD4:gp120 protein–protein interactions at the host cell-viral interface remains an important, yet difficult strategy for blocking HIV-1 infection.
A number of X-ray crystal structures of the gp120 core bound to the D1D2 domains of CD4 and the Fab of the neutralizing antibody 17b, a surrogate for the co-receptor, have been solved.27–34 The structure of the unbound form of the simian immunodeficiency virus (SIV) gp120, which has a 35% sequence identity with HIV-1 gp120, indicates an invariant outer domain, with conformational differences in both the bridging sheet and inner domain.35 In the ternary HIV-1 CD4:gp120:17b complex a large internal cavity is formed at the interface of the inner, outer, and bridging sheet domains of gp120 (Figure 1A).31, 32 Furthermore, a recent crystallographic study of the unliganded HIV-1 gp120 monomeric core demonstrates the propensity of gp120 to assume the CD4-bound conformation when not restrained by the presence of variable loops and interactions with gp41 in the trimer spike.36 At the CD4:gp120 interface, residue Phe43CD4 is located on the CDR2-like loop and binds within the hydrophobic cavity of gp120, termed the ‘Phe43 cavity’, while Arg59CD4 is located on a neighboring β-strand and forms an electrostatic interaction with Asp368gp120 (Figure 1B).31, 32 Importantly, the amino acids lining Phe43 cavity, as well as Asp368gp120, are well-conserved among primary HIV-1 isolates.37 Upon mutation of Phe43CD4 to Ala43CD4 a 550-fold reduction of binding affinity for gp120 was observed.38 Equally significant, the protein-protein interaction is reduced 9-fold when Arg59CD4 is mutated to Ala. Thus, the mutagenesis data and high degree of residue conservation highlight the functionally critical role that both the Phe43 cavity and Asp368gp120 binding hotspots play in mediating the HIV-1 viral entry process. Although inhibition of protein-protein interactions continues to be a significant challenge,39–41 recent examples indicate that efficient inhibition can be achieved by targeting surface hotspots.39, 42 The large CD4:gp120 contact surface, the orthogonal arrangement between the Phe43 cavity and the Asp368gp120 hotspots, and the therapeutic potential of targeting this interface comprise a challenging opportunity to combine structure based design and synthesis to develop novel small molecule antagonists of the viral attachment and entry process.
Previous screening of a small-molecule library for inhibitors of viral fusion led Debnath and coworkers to identify two inhibitors of CD4:gp120 binding, NBD-556 (1) and NBD-557 (2) (Table 1).43 Subsequent studies in our laboratories revealed that while 1 and 2 inhibit HIV-1 viral entry in CD4-positive, CCR5-expressing T-cells, 1 and 2 actually activate viral infection in CD4-negative cells (Table 1, column 3).44 Thus, in the context of CD4-negative cells, these small molecules both function as surrogates of the CD4-receptor and serve as agonists by promoting HIV-1 entry. Promotion of HIV-1 entry by NBD compounds may be possible in CD4-independent HIV-1 variants;45, 46 therefore, the agonistic properties of 1 and 2 must be eliminated for this chemotype. The thermodynamic signature of 1 binding to gp120 provides further evidence of the CD4-mimetic properties. For example, soluble CD4 (sCD4) binding to gp120 exhibits a highly favorable binding enthalpy balanced with an unfavorable entropy associated with molecular ordering.47, 48 Binding of 1 to the gp120 core is also characterized by both a favorable change in enthalpy (Δ and a large, unfavorable entropic component (−TΔS) (Table 1).44, 49 The similarity in thermodynamic signature for sCD4 and 1 suggests that the small molecules both induce and stabilize gp120 in a CD4-bound like conformation.
SAR studies identified three pharmacophoric elements that define the NBD chemotype: Region I, a para-halogenated phenyl ring; Region II, the oxalamide linker; and Region III, a tetramethylpiperidine ring.44, 50–52 Initial optimization of Region I and II44, 51 led to the synthesis of JRC-I-191 (3) with improved binding affinity.53 Mutagenesis and modeling of 3 revealed that these small molecules bind to the highly conserved gp120 cavity and compete with CD4 binding.44, 49, 51, 54 Unexpectedly, we found that 3 inactivates HIV-1 by inducing irreversible conformational changes which result in an active, yet transient, intermediate state of the Env trimer. If the metastable trimer does not encounter a target cell within the lifetime of the activated intermediate, it decays to a state that is no longer competent for viral entry.55 Further optimization of 3 led to the synthesis and biological evaluation of TS-II-224 (4), the most potent inhibitor of CD4:gp120 binding possessing the NBD chemotype.44, 51 While analogues 3 and 4 have improved affinities for gp120, as measured by Isothermal Titration Calorimetry (ITC), they are also more effective agonists of viral entry (relative to 1) in CD4-negative cells, clearly an undesired property (Table 1).44, 51 The recent cocrystal structure of 1 bound to gp120 (Clade C1086), at a resolution of 3.0 Å, confirms that 1 binds within the conserved Phe43 cavity.36 However, 1 does not form specific interactions with Asp368gp120 at the second binding hotspot. Thus, the essential Arg59CD4-Asp368gp120 electrostatic interaction on the CD4:gp120 surface has yet to be integrated in NBD small-molecule design. Hence, we hypothesized that engaging the Asp368gp120 binding hotspot would significantly improve viral inhibition properties, while also targeting a residue that is well-conserved across many HIV-1 subtypes.37 Herein, we describe a computational approach for the design of Region III analogues that target both the Phe43 cavity and the Asp368gp120 hotspots.
Recently, Kwon et al. determined four structures of the unliganded gp120 “extended core” (coree) from clade B (YU2 strain), clade C (C1086 and ZM109 strains), and clade A/E (93TH057 strain) primary HIV-1 isolates.36 The gp120 coree includes the N-terminus but excludes the variable loops and facilitates crystallization of the unliganded gp120. The clade A/E93TH057 construct of gp120 coree produced the highest resolution structure (1.9 Å).36 This protein also produced well diffracting crystals in complex with VRC01-like antibodies.34, 56 Therefore, we employed the same clade A/E gp120e as a template for small molecule cocrystallizations, with the exception that we mutated His375 to Ser within the Phe43 cavity to accommodate ligand binding. The crystal structure of 4 bound to clade A/E gp120(H375S) coree was determined at 2.0 Å resolution by molecular replacement (Figures 1C–E and Table 3). This structure reveals that 4 binds similarly to 1 in the ligand:gp120 complex,36 with Region I bound deep within the Phe43 cavity and forming aromatic stacking interactions with Phe382gp120 and Trp427gp120, as well as hydrophobic contacts with Val255gp120 and Ile424gp120. Both amide nitrogens of Region II form hydrogen bonds with the main-chain carbonyls on opposite sides of the Phe43 cavity (Gly473gp120 from the outer domain and with Asn425gp120 from the bridging sheet domain). In the cavity vestibule, one Region III gem-dimethyl moiety forms van der Waals contacts with the bridging sheet domain residues Trp427gp120, Asn428gp120 and Glu429gp120, while the second geminal dimethyl group contacts the outer domain residues Gln473gp120 and Asn474gp120. This cocrystal structure was employed as the starting point for our virtual screening and inhibitor design efforts.
Based on the critical role of the Arg59CD4-Asp368gp120 electrostatic interaction in CD4:gp120 recognition and binding,38 we hypothesized that addition of a similar chemotype to Region III, directed towards the Asp368gp120 hotspot, would improve the viral inhibition properties of 4. The cocrystal structure of 4 bound to the clade A/E gp120(H375S) coree indicated the close proximity of C4 of the tetramethylpiperidine ring (Region III) and the carboxylate side-chain of Asp368gp120 (Figure 1E). Rather than performing systematic synthetic modifications of Region III congeners, we chose a virtual screening strategy to identify alternative scaffolds that would contain a basic amine oriented towards Asp368gp120. Hence, an analogue possessing a primary amine attached to C4 of the 4-amino-tetramethylpiperidine was constructed in silico to afford query structure 5 (Figure 2). While geminal diamine 5 is not a chemically stable entity, we exploited this archetype to replicate the desired interactions between the small molecule and gp120. The prototype was assessed with the docking program GOLD57, 58 to provide a three-dimensional model that incorporated the desired trajectory of the amino group. Following our previously reported virtual screening paradigm51 employing the ROCS shape-based similarity algorithm,59–61 the amine prototype was used to search the Zinc Database of commercially available compounds.62, 63 Virtual screening identified several bicyclic primary amines, such as amino-bicyclo-nonanols, indanols, and diaminoindanes, that displayed both shape and chemotype similarity to prototype 5 and directed a hydrogen bond donor towards Asp368gp120(Table S1 in Supporting Information). In the end, we chose the synthetically versatile indane scaffold and docked the 1,2 and 1,3-diaminoindane enantiomers with GOLD.57, 58 The trans-1,2-diaminoindane isomers (−)-6 and (+)-6 were predicted to form weak polar interactions with Asp368gp120 and thus were selected for synthesis (Figure 2).
To assess the suitability of the 1,2-diamines as Region III platforms, we began with the synthetic conjugation of the commercially available cis-1-amino-indan-2-ol (−)-8 to ethyl oxalamide 7 (Figure 3). Amide bond formation was achieved under thermal conditions to furnish (+)-9. Enantiomer (−)-9 was also prepared employing the same procedure, beginning with (+)-8. These cis-indanol enantiomers (+)-9 and (−)-9 exhibited weak inhibition of viral entry (data not shown). Given that the trans-1,2-diaminoindanes exhibited the requisite spatial arrangement in the docking studies, we next converted (+)-9 to the corresponding trans-amine (+)-6. Tosylation of alcohol (+)-9, followed by SN2 displacement with sodium azide furnished (+)-10. Reduction of the azide employing Lindlar’s catalyst then provided the desired amine AWS-I-45 [(−)-6]. An identical synthetic sequence, beginning with (+)-9, was employed to construct the enantiomer, AWS-I-50 [(+)-6]. Although (−)-6 and (+)-6 exhibited non-specificity in our standard viral-entry inhibition assays (vide infra), conversion of the alcohol to the amines demonstrated an increase in binding affinity [KD (−)-6 = 1.2 μM; KD (+)-6 = 1.9 μM] compared to 1 (KD = 3.7 μM). Encouraged by this result, we sought to mimic further the Arg59CD4 side chain. Assessment of guanidine analogues by docking predicted favorable interactions between the ligand and Asp368gp120. These analogs were thus selected for synthesis. Here we employed 1H-pyrazol-1-carboxamidine monohydrochloride (11) to convert amine (+)-6, as well as the corresponding enantiomer (−)-6, to the desired guanidinium functionality, affording AWS-I-169 [(−)-12] and DMJ-I-228 [(+)-12], respectively. Note that following purification by high performance liquid chromatography (HPLC), the guanidinium formate salts of (−)-12 and (+)-12 were isolated and employed in all biological assays. The synthesis of (+)-12 is illustrated in Figure 3.
Amines (−)-6 and (+)-6 and the corresponding guanidinium salts (−)-12 and (+)-12 were first tested against mono-tropic and dual-tropic HIV-1 strains in single-round infection of recombinant HIV-1, encoding firefly luciferase.64 The recombinant viruses employed were pseudotyped with HIV-1 envelope glycoproteins derived from either a CXCR4, laboratory-adapted HXBc2 isolate, or the CCR5, primary YU2 isolate. To evaluate compound specificity for HIV-1, the viruses were pseudotyped with the envelope glycoproteins of the amphotropic murine leukemia virus (A-MLV), an unrelated retrovirus.53 In both mono-tropic (HXBc2, YU2, ADA, JRFL) and dual-tropic (89.6 and KB9) viruses, (−)-6 and (+)-6 inhibited entry on cells co-expressing CD4 and CCR5 or CXCR4 with IC50 in the range of 50 to 90 μM, but also exhibited non-specificity, by inhibiting entry of the A-MLV control virus with a similar IC50 value (Table 2). The guanidinium salts (−)-12 and (+)-12, on the other hand, demonstrated improved inhibition of viral entry (22.9 and 21.3 μM, respectively). Moreover, the inhibition proved to be completely specific to HIV-1, as neither compound was found to inhibit entry of A-MLV (Table 2 and Figure 4). Assessment of the in vitro cytotoxicity of (+)-12 in Cf2Th-CD4-CCR5 cells did not demonstrate measurable inhibition of cell growth (Figure S1 in Supporting Information). Thus, analogues (−)-12 and (+)-12 posses significantly improved antiviral activities relative to the starting compounds 1–4.
To characterize further the antiviral breadth, (−)-12 and (+)-12 were tested against 42 diverse HIV-1 strains (clades B and C). While the parent compound 1 was found to neutralize HIV-1 with a geometric mean titer (GMT) IC50 = 29.3 μM, only 12% of isolates were found to have an IC50 < 10 μM (Figure 5 and Table S3 in Supporting Information). In contrast, both (−)-12 and (+)-12 and were found to display 100% neutralization breadth among the viral isolates tested. Compound (−)-12 possessed a GMT IC50 of 8.9 μM with 52% of the viral isolates having an IC50 < 10 μM. Similarly, (+)-12 revealed a mean IC50 of 7.9 μM with 57% of the isolates having an IC50 < 10 μM. Compounds (−)-12, (+)-12 and 1 all exhibited less antiviral activity against clade C viruses compared to clade B strains. Given the high degree of sequence conservation in the Phe43 cavity and that the gp120 trimer exhibits a high degree of conformational diversity in solution48, 65, 66 the decreased potency in clade C is not easily assignable.
Given the broadly neutralizing capabilities of the indane-based analogues, we asked if (−)-12 and (+)-12 could pose a therapeutic disadvantage by functionally replacing CD4 and enhancing cellular infection? We have previously measured the enhancement of HIV-1 entry in CD4-independent viruses demonstrating that 1–4 enhance the entry of HIV-1 YU2 viruses in CD4-negative cells (vide supra).44, 51 We found that, both amines (−)-6 and (+)-6 enhanced viral entry in CD4-negative cells at levels comparable to those previously measured for 1–4 (Table 2). Remarkably, the guanidinum salts (−)-12 and (+)-12 displayed no appreciable entry enhancement of YU2 viruses into CD4-negative, CCR5-expressing cells (Table 2). We further evaluated whether the lack of viral enhancement might be related to the unproductive binding of (−)-12:gp120 or (+)-12:gp120 complexes to the CCR5 receptor. Surface Plasmon Resonance (SPR) assessment of the CCR5 antibody surrogate 17b binding to the small molecule:gp120 complexes indicated that both 4 and the indane congeners [(−)-6, (+)-6, (−)-12 and (+)-12], enhance binding of 17b to both the monomeric core and full-length gp120 (Figure S2 in Supporting Information). Therefore, the enhancement of viral entry into CD4-negative, CCR5-expressing cells is likely unrelated to the formation of the 17b binding epitope. Hence, the addition of the guanidinium functionality reduces the undesired enhancement of viral infection observed for 1–4, (−)-6, and (+)-6. We therefore conclude that (−)-12 and (+)-12 function as viral entry antagonists in the context of CD4-negative cells expressing the CCR5 coreceptor.
The binding of the novel analogues to full-length gp120 from the YU2 strain was next characterized by ITC. The measured binding affinities of (−)-12 (KD = 0.30 μM), and (+)-12 (KD = 0.25 μM), demonstrate that incorporation of the guanidinium functionality restores potency comparable to parental compound 4 (KD = 0.30 μM) (Table 2). We also note that the observed discrepancies between measured IC50 and KD values (Table 2) reflect the differences in small molecule binding to the full Env trimer and the gp120 monomer in viral and ITC binding assays, respectively. The binding efficiency index (B.E.I.) is defined as the binding affinity (KD) divided by molecular weight while the surface efficiency index (S.E.I.) is defined as the binding affinity (KD) divided by the molecular polar surface area.67 Both (−)-12, and (+)-12, with identical affinity and structural properties yield the same calculated B.E.I. and S.E.I., 16.8 and 6.4, respectively. When compared to 22 known inhibitors of protein-protein interactions (B.E.I. = 11.7 ± 2.4 and S.E.I. = 7.2 ± 3.0)41, 68 and 92 marketed drugs, (B.E.I. = 25.8 ± 7.9 and S.E.I. = 14.5 ± 8.7), the values for (−)-12, and (+)-12 fall within the lower bounds of binding efficiency and surface efficiency space for marketed drugs.67
We next assessed the enthalpic and entropic contributions to binding as compared to sCD4 binding (Table 2 and Figure 6A). The binding of sCD4 to gp120 at 25 °C is associated with a large enthalpy change (ΔH = −34.5 kcal/mol), that is partially negated by a large unfavorable entropy change (−TΔS = 23.6 kcal/mol) and is associated with a large negative heat capacity (ΔCp) of –1,800 cal/(K × mol).49 Such a binding event suggests a large molecular reordering of gp120 upon CD4 binding.48, 49 Compounds 1–4 also exhibit a large unfavorable entropy upon binding to gp120.49 Thus, these small molecules bind and stabilize a gp120 conformation similar to that observed in the CD4 bound-state. Characterization of analogues (−)-6 and (−)-12 (Table 2) indicated a thermodynamic signature resembling that of 4 (Figure 6B and Table 2). Conversely, the antipodes (+)-6 and (+)-12 exhibited decreased unfavorable entropy contributions (−TΔS = 7.6 and 5.9 kcal/mol, respectively) compared to 1–4, (−)-6 and (−)-12. Therefore, the (+)-6 and (+)-12 enantiomers, as shown in Figure 6, have a distinct and preferable thermodynamic signature, with gp120 incurring a lower entropic penalty upon binding. The associated changes in heat capacity for the binding of 4, (−)-12 and (+)-12, calculated from the temperature dependence of the binding enthalpies, are −738 ± 36, −817 ± 15 and −398 ± 5 cal/(K × mol), respectively (Figure 6B). The change in the heat capacity for (+)-12 approaches the expected value for the burial of a small hydrophobic molecule.69, 70 Thus, enantiomer (+)-12 displays chemical properties and a thermodynamic signature resembling small molecule drug binding.
To determine if the antiviral and thermodynamic properties of (−)-12 and (+)-12 were exerted through direct binding to the Phe43 cavity hotspot, we measured viral inhibition in the Phe43 cavity filling mutant S375Wgp120 YU2 virus, wherein the tryptophan side chain restricts access to the Phe43 cavity.71 Unlike the wild-type virus, the YU2 S375Wgp120 was completely resistant to (−)-12 and (+)-12, (Table S2 in Supporting Information) supporting the claim that these new compounds exhibit their antiviral effects through binding within the conserved Phe43 cavity in context of the functional trimeric Env spike. Cocrystallization experiments with compounds (+)-6, (−)-12 and (+)-12 bound to the monomeric A/E gp120(H375S) coree were employed to define further the binding interactions of these small molecule ligands with gp120.
Amine (+)-6 and guaninidium salts (−)-12 and (+)-12, were cocrystallized in complex with the clade A/E gp120(H375S) coree and yielded high resolution X-ray cocrystal structures at 1.8 Å, 1.8 Å, and 1.9 Å, respectively (Figures 7, ,8,8, ,99 and Table 3).72 In the three crystal structures, Regions I and II maintain similar interactions, as observed with 4, binding deep within the Phe43 cavity of gp120 (Figures 7, ,88 and and9).9). The mean RMSD values of the backbone and side-chain atoms for residues lining the Phe43 cavity are 0.39 Å and 0.68 Å, respectively, indicating structural conservation in the Phe43 hotspot among the ligand:gp120 complexes. Moreover, comparison of (+)-6, (−)-12 or (+)-12:gp120 complexes with the unliganded gp120 structure36 or the CD4:gp120 structure (PDB: 1G9M),31 shows that the Asp368gp120 side-chain has the same orientation upon small molecule binding. Inspection of Region III in the (+)-6:gp120 complex (Figure 7) reveals that the 2-amino group does not interact with the Asp368gp120 hotspot. Nonetheless, the indane scaffold of (+)-6 occupies a similar position to the tetramethylpiperidine ring of 4 (Figure S3A in Supporting Information). The five-membered ring of (+)-6 overlaps with the piperidine ring while the aromatic ring overlaps with one of the gem dimethyl groups of 4, (Figure S3A in Supporting Information). The aromatic ring of (+)-6 also forms extensive contacts with outer domain residues Gly473gp120 and Asn474gp120. Superposition of the (+)-6:gp120 structure with a CD4:gp120 structure indicates that the indane arene ring overlaps with Asn40CD4 and Gly41CD4 main-chain atoms, while the five-membered ring overlaps with the main-chain atoms of Ser42CD4 and Phe43CD4 (Figure S3A and S3B in Supporting Information). Therefore, the indane ring effectively mimics the CD4 β-turn spanning the interface of the outer domain and bridging sheet of gp120.
Addition of the guanidinium group to afford (−)-12 and (+)-12, was envisioned to strengthen the ligand interactions with the Asp368gp120 hotspot to resemble more closely the native protein-protein interaction. The 1.8 Å cocrystal structure of the (−)-12:gp120 complex reveals that the indane ring is rotated 90° in the cavity vestibule (relative to Region II), leaving the plane of the ring tilted away from the gp120 surface (Figure 8). Hence, the arene ring of (−)-12 does not form contacts with the outer domain as observed in the (+)-6:gp120 complex (Figure 8). Furthermore, the altered orientation imposed by the stereochemical configuration of (−)-12 dictates that the guanidinium functionality approach Asp368gp120 near the outer domain rather than near the bridging sheet as observed with (+)-6. Interestingly, the guanidinium moiety of (−)-12 does not form a salt bridge with Asp368gp120. Instead, the guandinium group forms hydrogen bonds with a network of ordered water molecules (Wat601-Wat603) and forms a single ionic interaction with Asp368gp120 (Figure 8). Therefore, in the (−)-12:gp120 complex, the guanidinum moiety resides at the center of a hydrogen-bonded network between ordered watered and the outer domain residues Gly472gp120 and Asp368gp120.
Further analysis of the (+)-12:gp120 cocrystal structure indicates that the guanidinium group of (+)-12 approaches Asp368gp120 between the bridging sheet and outer domains (Figure 9). The guanidinium moiety of (+)-12 also forms a single ionic interaction with Asp368gp120 and in addition forms an intra-molecular hydrogen bond to the Region II oxalamide and a hydrogen bond to crystallographic water Wat501. Interestingly, Wat501 occupies a position equivalent to that of Arg59CD4 observed in the CD4:gp120 complex (Figure S3B in Supporting Information).31 Wat501 is at the center of a water mediated hydrogen-bonding network that connects (+)-12 with both the outer domain (Asp368gp120) and the bridging sheet (Met426gp120). Thus, between the two complexes [(+)-12 versus (−)-12], a unique pattern of hydrogen bonding exists between the guanidinium moiety, ordered water molecules and the gp120 outer and bridging sheet domains. As judged by the dissociation constant of (+)-12 (KD = 0.25 μM) and (−)-12 (KD = 0.30 μM) compared to (+)-6 (KD = 1.9 μM), the guanidinium-Asp368gp120 interaction and water mediated hydrogen bonding contributes favorably to the binding affinity of both (+)-12 and (−)-12 to gp120. Furthermore, the distinct thermodynamic signatures of gp120 binding between the (+)-12 and (−)-12 enantiomers are consistent with the different binding modes revealed in the crystal structures of these enantiomeric compounds bound to gp120
We have targeted the highly conserved Phe43 cavity and Asp368gp120 hotspots at the protein-protein interface of the CD4:gp120 complex to improve the antiviral potency and breadth of 1–4. In this work a modified ROCS “scaffold hopping”,60, 61 virtual screening strategy was employed to develop inhibitors that target both the Phe43gp120 and Asp368gp120 hotspots. A hypothetical 1,1-diamine molecule (5), encompassing the desired chemotype and spatial features was employed to search chemical space to identify a lead compound possessing a basic amine directed towards the Asp368 hotspot. This strategy, in conjunction with molecular design and synthesis, has led to the identification of two novel analogues that employ the trans-1,2-disubstituted indane scaffold [e.g. (−)-12 and (+)-12] to direct a guanidinium group towards the Asp368gp120 hotspot.
We subsequently demonstrated that (−)-12 and (+)-12 specifically inhibit HIV-1 entry through interaction with the Env spike by virtue of the incorporation of a guandinium group. Both (−)-12 and (+)-12 have improved and broad antiviral potency, and, importantly, do not enhance viral infection in CD4-negative cells. The wide neutralization breath against diverse strains of clade B and C viral isolates validates the importance of considering strain variability when evaluating binding hotspots appropriate for inhibitor design. Furthermore, in the context of CD4-negative cells, amines (−)-6 and (+)-6 function as agonists, promoting the CD4-independent viral entry process. Incorporation of the guanidinium functionality to afford (−)-12 and (+)-12 converts these agonists to antagonists of viral entry and eliminates the undesired property of promoting CD4-independent viral entry. The ability to suppress the undesired agonist properties demonstrates that the Phe43 cavity of gp120 and the NBD compound class remains an attractive target for the development of HIV-1 entry inhibitors.
The crystal structures described herein further confirm our design criteria and represent the highest resolution structures, to date, of a small molecule entry antagonist bound to the gp120 core. Importantly, the guanidinium groups of (−)-12 and (+)-12 form an electrostatic interaction with Asp368gp120 and reveal two distinct water-mediated hydrogen-bonding networks between the antagonist guanidinium group and Asp368gp120. Interestingly, (−)-12 and (+)-12 also display distinct thermodynamic signatures. The smaller unfavorable entropy for (+)-12 binding to gp120, along with the favorable binding efficiency and surface efficiency indices, may make this enantiomer a more suitable candidate for continued optimization. The modest 20 μM IC50 exhibited (−)-12 and (+)-12 indicates additional optimization will be required to make this class of molecules more effective. In Region I, the surface complementarity of the halogenated phenyl ring could be improved and in Region III, the guaninidium-Asp368gp120 interactions could be strengthened and the adjacent bound solvent incorporated in compound design. Therefore, the high-resolution crystal structures of these antagonists, with demonstrated improvement in anti-viral potency and breadth, provide a novel structural paradigm to facilitate future cycles of design, synthesis and biological evaluation to develop this class of small molecule inhibitors of HIV-1 entry.
Molecules were constructed in MOE73 ionized using MOE’s WashMDB function, and hydrogens were added. The small molecule conformation was minimized to a gradient of 0.01 with the MMFF94x74, 75 using a distance-dependent dielectric constant of 1.
Using the X-ray crystal structure of 4 bound to HIV-1 clade A/E gp120(H375S) coree, hydrogen atoms were added and tautomeric states and orientations of Asn, Gln and His residues were determined with Molprobity.76, 77 Hydrogens were added to crystallographic waters using MOE (2010).73 The OPSLAA78 force field in MOE was used and all hydrogen atoms were minimized to an rms gradient of 0.01, holding the remaining heavy atoms fixed. A stepwise minimization followed for all atoms, using a quadratic force constant (100) to tether the atoms to their starting geometries; for each subsequent minimization, the force constant was reduced by a half until 0.25. This was followed by a final cycle of unrestrained minimization.
GOLD (version 4.0.1)57, 58 was used for docking and the binding site was defined by using the crystallographic position of 4. One hundred genetic algorithm (GA) docking runs were performed with the following parameters: initial_virtual_pt_match_max=3.5, diverse_solutions=1, divsol_cluster_size=1, and divsol_rmsd=1.5. All other parameters were set as defaults.
Flipper from Open Eye was used to expand compounds with unspecified chirality prior to generation of conformers. Using Omega (version 2.2.1)79 from Open Eye with default parameters, a maximum of 50 low energy conformers for all compounds in the Zinc Database (version 7)62, were generated and stored in sd files of approximately 10,000 molecules. ROCS60, 61 searches were run using 3D coordinates from the docked binding mode of the amine containing tetramethylpiperidine prototype. The Implicit Mills Dean80 force field was used to match chemotypes as well as shape. Hits were ranked by a combination of Tanimoto and the scaled Color Score (ComboScore). Primary amines with a ComboScore > 1.0 were selected from the set of hits, filtered for currently commercial availability, conjugated in silico and were docked with GOLD57, 58 and scored with a mass-corrected Goldscore.
All reactions were conducted in oven-dried glassware under an inert atmosphere of nitrogen or argon, unless otherwise stated. All solvents were reagent or high performance liquid chromatography (HPLC) grade. Anhydrous CH2Cl2 and THF were obtained from the Pure Solve™ PS-400 system under an argon atmosphere. All reagents were purchased from commercially available sources and used as received. Reactions were magnetically stirred under a nitrogen atmosphere and monitored by either thin layer chromatography (TLC) with 0.25 mm E. Merck pre-coated silica gel plates or analytical HPLC. Yields refer to chromatographically and spectroscopically pure compounds. Optical rotations were measured on a JASCO P-2000 polarimeter within the specified solvent. Proton and carbon-13 NMR spectra were recorded on a Bruker AM-500 at 305 K, unless otherwise noted. Chemical shifts are reported relative to chloroform (δ 7.26), methanol (δ 3.31), or dimethyl sulfoxide (δ 2.50) for 1H NMR and either chloroform (δ 77.0), methanol (δ 49.2), or dimethyl sulfoxide (δ 39.4) for 13C NMR. High-resolution mass spectra (HRMS) were recorded at the University of Pennsylvania Mass Spectroscopy Service Center on either a VG Micromass 70/70H or VG ZAB-E spectrometer. Analytical HPLC was preformed with a Waters HPLC-MS system, consisting of a 515 pump and Sunfire C18 reverse phase column (20 μL injection volume, 5 μm packing material, 4.5 × 50 mm column dimensions) with detection accomplished by a Micromass ZQ mass spectrometer and 2996 Photodiode Array detector. Preparative HPLC was preformed with a Gilson 333/334 preparative pump system equipped with a 5 mL injection loop, Sunfire C18 OBD column (5 μm packing material, 19 × 100 mm column dimensions) equipped with a UV-Vis dual wavelength detector (210 and 254 nm) and 215 liquid handling module. Solvent systems employed were based on the following buffers: Buffer A: H2O containing 0.05% formic acid; Buffer B: MeCN containing 0.05% formic acid. The purity of new compounds was judged by NMR and HPLC-MS to be in excess of 95%.
To a solution containing Ethyl 2-(4-chloro-3-fluorophenylamino)-2-oxoacetate (7) (1.87 g, 7.62 mmol) in 20 mL of EtOH was added (1S,2R)-1-amino-2,3 dihydro-1H-inden-2-ol [(−)-8, 1.19 g, 7.98 mmol] in one portion. The reaction mixture was stirred and heated to 150 °C in a sealed tube for 1 hour, after which the suspension was allowed to cool to room temperature. The precipitate was collected by vacuum filtration and washed with small portions of dichloromethane to provide 2.22 g (83%) of analytically pure TK-II-52 (+)-9 as a white semi-crystalline solid; [α]29D = +103.5° (c = 0.56, DMSO); 1H NMR (500 MHz, DMSO-d6) δ 11.19 (s, 1H), 8.36 (d, J = 8.7 Hz, 1H), 7.97, (dd, J = 2, 11.7 Hz, 1H), 7.77, (d, J = 8.8 Hz, 1H), 7.59, (t, J = 8.7 Hz, 1H), 7.28-7.18, (m, 4H), 5.46, (d, J = 4.9 Hz, 1H), 5.25, (dd, J = 5.2, 8.6 Hz, 1H), 4.52, (dd, J = 4.6, 8.6 Hz, 1H), 3.14, (dd, J = 4.9, 16.2 Hz, 1H), 2.88, (d, J = 16.1 Hz, 1H); 13C NMR (125 MHz, DMSO-d6) δ 159.3, 158.8, 156.8, (d, JCF = 242.5 Hz), 141.0, 140.8, 138.2, (d, JCF = 10 Hz), 131.6, 127.7, 126.5, 125.0, 124.1, 117.5 (d, JCF = 2.9 Hz), 114.5 (d, JCF = 17.5 Hz), 108.6, (d, JCF = 25 Hz), 71.6, 56.9. HRMS (ES+) m/z 371.0572 [(M+Na)+; calcd for C17H14ClFN2O3: 371.0575].
To a solution containing (+)-9 (2.20 g, 6.31 mmol) in a mixture of dichloroethane (50 mL) and THF (10 mL) was added p-toluenesulfonyl chloride (3.60 g, 18.88 mmol), followed by NEt3 (2.64 mL, 18.91 mmol), DMAP (0.77 g, 2.82 mmol), and a stir bar. A reflux condenser was attached and the solution was heated to 60 °C and stirred for 2 hours. After cooling, the reaction mixture was quenched with 100 mL of saturated NH4Cl solution, followed by extraction with DCM/EtOAc (3:1, 3 × 100 mL). The combined organic fractions were washed with brine and dried over Na2SO4, filtered, and concentrated in vacuo. The crude product was sonicated in 50 mL of DCM/hexanes (1:1) for 30 minutes and then filtered. The filtrate was then washed with DCM/Hexanes (1:1, 2 × 25 mL) to provide 2.21 g (70%) of the desired tosylate [(+)-SI-1] as a white flakey solid; [α]D25 = +10.4° (c = 0.29, CH2Cl2); 1H NMR (500 MHz, DMSO-d6) δ 10.97 (s), 8.84 (d, J = 9.0 Hz, 1H), 8.00 (dd, J = 2.5, 12.0 Hz, 1H), 7.79 (dd, J = 2.0, 9.0 Hz, 1H), 7.75 (d, J = 8 Hz, 2H), 7.62 (t, J = 8.5 Hz, 1H), 7.33-7.22 (m, 6H), 5.49 (dd, J = 5.0, 8.5 Hz, 1H), 5.24 (dt, J = 1.5, 5.0 Hz, 1H), 3.34 (m, overlap with water, 1H), 3.13 (d, J = 16.0 Hz, 1H), 2.33 (s, 3H); 13C NMR (125 MHz, DMSO-d6) δ 159.4, 157.9, 156.8 (d, JCF = 243.8 Hz), 144.8, 138.7, 138.6, 138.2, (d, JCF = 10 Hz), 132.7, 130.6, 129.9, 128.3, 127.5, 127.0, 124.8, 123.9, 117.3, (d, JCF = 2.7 Hz), 114.4, (d, JCF = 17.5 Hz), 108.4, (d, JCF = 25.0 Hz), 82.7, 55.6, 37.6, 20.9. HRMS (ES+) m/z 503.0862 [(M+H)+; calculated for C24H20ClFN2O5S: 503.0844].
To a solution of (+)-SI-1, prepared above, (2.20 g, 4.38 mmol) in 8 mL DMSO was added NaN3 (2.86 g, 43.99 mmol) and a stir bar. The solution was heated to 50 °C and stirred for 2.5 hours. After cooling to room temperature, the reaction was quenched with H2O (50 mL) followed by extraction with DCM/EtOAc (3:1, 3 × 100 mL). The combined organic fractions were washed with brine (3 × 100 mL), dried over Na2SO4, filtered, and concentrated in vacuo. The crude product was then dissolved in a minimal amount of hot EtOAc, followed by addition of an equal portion of hot hexane. The obtained precipitate was collected, the mother liquor concentrated, and the residual solid was recrystallized via the same procedure. The isolated solids were combined to provide 1.30 g (79%) of the azide (+)-10 as a white flakey solid; [α]D25 = +40.7° (c = 0.51, EtOAc); 1H NMR (500 MHz, DMSO-d6) δ 11.14 (s, 1H), 9.66 (d, J = 9.0 Hz, 1H), 7.98 (dd, J = 2.0, 11.5 Hz, 1H), 7.77 (dd, J = 1.5, 8.5 Hz, 1H), 7.60 (t, J = 8.5 Hz, 1H), 7.28-7.22 (m, 3H), 7.15 (d, J = 7.0 Hz, 1H), 5.34 (t, J = 8.0 Hz, 1H), 4.54 (q, J = 8.0 Hz, 1H), 3.34 (dd, J = 7.5, 15.5 Hz, 1H), 2.88 (dd, J = 8.5, 15.5 Hz, 1H); 13C NMR (125 MHz, DMSO-d6) δ 160.1, 158.5, 156.8, (d, JCF = 242.5 Hz), 139.9, 138.8, 138.2 (d, JCF = 10 Hz), 130.5, 128.2, 127.1, 124.7, 123.4, 117.4, (d, JCF = 3.0 Hz), 114.4, (d, JCF =17.8 Hz), 108.5, (d, JCF = 25.6 Hz), 65.7, 59.4, 35.3. HRMS (ES+) m/z 372.0652 [(M-H)−; calculated for C17H12ClFN5O2: 372.0664].
To a solution of (+)-10 (1.30 g, 3.48 mmol) in 35 mL MeOH was added Lindlar’s catalyst (5% Pd/CaCO3, poisoned with lead, 650 mg) and a stir bar. Hydrogen was bubbled through the solution, after which the reaction mixture was stirred for 2 hours under a hydrogen atmosphere at room temperature. The reaction mixture was then filtered through a plug of celite and the filter cake washed with additional methanol. The filtrate was concentrated in vacuo to obtain the crude product, which was then purified by silica gel chromatography utilizing a gradient solvent system (DCM/MeOH/NH4OH 100:1:0.1 to 50:1:0.1 to 20:1:0.1 to 10:1:0.1) to afford 1.06 g (87%) of amine (+)-6 as a white flakey solid; [α]D25 = +93.0° (c = 0.14, MeOH); 1H NMR (500 MHz, CDCl3): δ 11.07 (s, 1H), 9.22 (d, J = 9.0 Hz, 1H), 7.98 (dd, J = 2.4, 11.9 Hz, 1H), 7.77 (ddd, J = 1.0, 2.5, 8.9 Hz, 1H), 7.60 (t, J = 8.7 Hz, 1H), 7.20–7.14 (m, 3H), 7.06 (d, J = 7.2 Hz, 1H), 5.00 (t, J = 8.5 Hz, 1H), 3.69 (q, J = 8.5 Hz, 1H), 3.08 (dd, J = 7.5 Hz, 1H), 2.64 (dd, J = 9.2 Hz, 1H); 13C NMR (125 MHz, DMSO-d6) δ 160.3, 159.0, 156.8, (d, JCF = 242.5 Hz), 142.0, 140.6, 138.3 (d, JCF = 10 Hz), 130.6, 127.4, 126.4, 124.4, 123.3, 117.3, (d, JCF = 3.75 Hz), 114.2, (d, JCF = 17.5 Hz), 108.4 (d, JCF = 25 Hz), 62.6, 59.6, 48.6. HRMS (ES+) m/z 348.0911 [(M+H)+; calculated for C17H15ClFN3O2: 348.0915].
To a solution containing (+)-6 (1.06 g, 3.05 mmol) in 35 mL DMF was added 1H-pyrazole-1-carboxamidine hydrochloride (11, 879 mg, 6.00 mmol), N,N-diisopropylethylamine, (2.10 mL, 12.06 mmol) and a stir bar. The solution was heated to 110°C for 16 hours and then allowed to cool. The light-red reaction mixture was concentrated to approx. 5 mL and diluted with 20 mL MeCN/H2O (1:1) and the solution was subjected to purification via HPLC to give 350 mg (27%) of (+)-12 as a white fluffy solid after lyophilization; [α]D26 = +5.8° (c = 0.72, MeOH) 1H NMR (500 MHz, CD3OD): δ 8.3 (br s, 1Hformate), 7.84 (dd J = 2.5, 11.5 Hz, 1H), 7.50 (d, J =8.75 Hz, 1H), 7.43 (t, J = 8.5 Hz, 1H)), 7.34 – 7.29 (m, 4 H), 5.27 (d, J = 4.5 Hz, 1H), 4.30 (m, 1H), 3.52 (dd, J = 7.5 Hz,1H), 2.93 (dd, J = 5, 16.5 Hz, 1H); 13C NMR (125 MHz, CD3OD) δ 162.8, 159.4, 159.3, (d, JCF = 244.4 Hz), 159.0, 141.9, 139.7, 139.1 (d, JCF = 9.8 Hz), 131.9, 130.6, 129.0, 126.4, 126.3, 118.3 (d, JCF = 3.3 Hz), 117.4 (d, JCF = 17.8 Hz), 110.0 (d, JCF = 26.3 Hz), 62.3, 59.7, 37.8. HRMS (ES+) m/z 390.1132 [(M+H)+; calculated for C18H18ClFN5O2: 390.1133]. Note that the formate counterion was not observed under the HRMS conditions
The enantiomeric compounds (−)-6 and (−)-12 were prepared using the procedures described above, starting with commercially available amino indanol (+)-8. The intermediates and final compounds possess identical 1H and 13C NMR spectra, as well as HRMS data. The observed optical rotations were: (−)-9: [α]29D = −112.1° (c = 0.48, DMSO); (−)-10: [α]D25 = −34.9° (c = 0.30, EtOAc); (−)-6: [α]D25 = −83.2° (c = 0.16, MeOH); and (−)-12: [α]D25 = −3.2° (c = 0.72, MeOH).
Compounds were dissolved in dimethyl sulfoxide (DMSO), and stored at 10 mM concentrations at −20°C. The compounds were diluted in Dulbecco Modified Eagle Medium (DMEM, Invitrogen) to create 1 mM solutions before use. Soluble CD4 (sCD4) was purchased from ImmunoDiagnostics (Woburn, MA). Human 293T embryonic kidney and canine Cf2Th thymocytes (ATCC) were grown at 37°C and 5% CO2 in DMEM (Invitrogen) containing 10% fetal bovine serum (Sigma) and 100 μg/mL of penicillin-streptomycin (Meditech, Inc.). Cf2Th cells stably expressing human CD4 and either CCR5 or CXCR481, 82 were grown in medium supplemented with 0.4 mg/mL of G418 (Invitrogen) and 0.20 mg/mL of hygromycin B (Roche Diagnostics). Using the Effectene transfection reagent (Qiagen), 293T human embryonic kidney cells were cotransfected with plasmids expressing the pCMVΔP1ΔenvpA HIV-1 Gag-Pol packaging construct, the wild-type or mutant HIV-1YU2 envelope glycoproteins or the envelope glycoproteins of the control amphotropic murine leukemia virus (A-MLV), and the firefly luciferase-expressing vector at a DNA ratio of 1:1:3 μg. For the production of viruses pseudotyped with the A-MLV glycoprotein, a rev-expressing plasmid was added. The single-round, replication-defective viruses in the supernatants were harvested 24–30 hours after transfection, filtered (0.45 μm), aliquoted, and frozen at −80°C until further use. The reverse transcriptase (RT) activities of all viruses were measured as described previously83.
Cf2Th/CD4-CCR5 or Cf2Th/CD4-CXCR4 target cells were seeded at a density of 6 × 103 cells/well in 96-well luminometer-compatible tissue culture plates (Perkin Elmer) 24 h before infection. On the day of infection, (1 to 100 μM) was added to recombinant viruses (10,000 reverse transcriptase units) in a final volume of 50 μL and incubated at 37°C for 30 minutes. The medium was removed from the target cells, which were then incubated with the virus-drug mixture for 2–4 hours at 37°C. At the end of this time point, complete medium was added to a final volume of 150 μL and incubated for 48 hours at 37°C. The medium was removed from each well, and the cells were lysed with 30 μL of passive lysis buffer (Promega) by three freeze-thaw cycles. An EG&G Berthold Microplate Luminometer LB 96V was used to measure luciferase activity in each well after the addition of 100 μL of luciferin buffer (15 mM MgSO4, 15 mM potassium phosphate buffer [pH 7.8, 1 mM ATP, 1 mM dithiothreitol) and 50 μL of 1 mM D-luciferin potassium salt (BD Pharmingen).
Compounds were dissolved in dimethyl sulfoxide (DMSO), and stored at 10 mM concentrations at −20°C. The compounds were diluted in Dulbecco Modified Eagle Medium (DMEM, Invitrogen) to create 1 mM solutions before use. In addition, DMSO (SIGMA) was also diluted in DMEM and used as negative control. Cytotoxicity assay was done in parallel with the cell-based infection assay as described in method section of this article using Cf2Th cells stably expressing human CD4 and CCR5. Briefly, Cf2Th/CD4-CCR5 target cells were seeded at a density of 6 × 103 cells/well in 96-well luminometer-compatible tissue culture plates (Perkin Elmer) 24 h before infection. On the day of infection, media from target cells were removed and 1 to 100 μM of (+)-12 or DMSO to 2% (equivalent to concentration of DMSO in 100 μM (+)-12) was added to target cells at a final volume of 50 μL and incubated at 37°C for 2–4 hrs. At the end of this time point, complete medium was added to a final volume of 150 μL and incubated for 48 hours at 37°C. 48 hrs later, the plates were equilibrated to room temperature for 30 min and 30 μL of substrate reagent, CellTiter-Glo (Promega), was added to each well and incubated at room temperature for 5–10 min. An EG&G Berthold Microplate Luminometer LB 96V was used to measure luciferase activity in each well for 5 seconds.
Isothermal titration calorimetric experiments were performed using a high-precision VP-ITC titration calorimetric system from MicroCal LLC. (Northampton, MA). The calorimetric cell (~1.4 mL), containing gp120 at a concentration of about 2 μM dissolved in PBS, pH 7.4 (Roche Diagnostics GmbH), with 2 % DMSO, was titrated with the different compounds dissolved in the same buffer at concentrations of 80 – 130 μM. The compound solution was added in aliquots of 10 μL at pre-set intervals. All solutions were degassed to avoid any formation of bubbles in the calorimeter during stirring. All experiments were performed at 25 °C. The heat evolved upon injection of compound was obtained from the integral of the calorimetric signal. The heat associated with the binding reaction was obtained by subtracting the heat of dilution from the heat of reaction. The individual binding heats were plotted against the molar ratio, and the values for the enthalpy change (ΔH) and association constant, Ka (Kd = 1/Ka), were obtained by nonlinear regression of the data.
HIV-1 Env-pseudoviruses were prepared by transfecting 293T cells with 10 μg of rev/env expression plasmid and 30 μg of an env-deficient HIV-1 backbone vector (pSGΔ3env), using Fugene 6 transfection reagents (Invitrogen). Pseudovirus-containing culture supernatants were harvested two days after transfection, filtered (0.45 μm), and stored at −80°C or in the vapor phase of liquid nitrogen. Neutralization was measured using HIV-1 Env-pseudoviruses to infect TZM-bl cells as described previously84–87 with minor modifications. Briefly, the test reagent [(+)-12, (−)-12, 1 or CD4-Ig] were diluted in complete media containing 10% DMSO. Then 40 μl of virus was added to 10 μl of serial diluted test reagent in duplicate wells of a 96-well flat bottom culture plate, and the virus-reagent mix was incubated for 30 min at 37°C. To keep assay conditions constant, sham media containing 10% DMSO was used in place of test reagent in specified control wells. The virus input was set at a multiplicity of infection of approximately 0.01–0.1, which generally results in 100,000 to 400,000 relative light units (RLU) in a luciferase assay (Promega, Madison, WI). The test reagent concentrations were defined at the point of incubation with virus supernatant. Neutralization curves were fit by nonlinear regression using a 5-parameter hill slope equation as previously described.86 The 50% or 80% inhibitory concentrations (IC50 or IC80) were reported as the reagent concentrations required to inhibit infection by 50% or 80%.
The HIV-1 gp160 protein sequences of isolates used in the neutralization assays were aligned using MUSCLE, for multiple sequence comparison by log-expectation.88, 89 The protein distance matrix was calculated by “protdist” using the Jones-Taylor-Thornton model90 and the dendrogram was constructed using the neighbor-joining method91 by “Neighbor”. The analysis was performed at the NIAID Biocluster (https://niaid-biocluster.niaid.nih.gov/). The trees were displayed with Dendroscope.92
A point mutation was introduced to pVRC8400-HIV-1 clade A/E 93TH057 ΔV123 expression vector to generate clade A/E93TH057 gp120 coree (H375S)-expressing plasmid. The plasmid construct was verified by DNA sequencing.
Clade A/E93TH057 gp120(H375S) coree was purified as described (manuscript submitted). Small molecules in 100% DMSO were incorporated in the purified gp120 to make a final concentration of 100 μM. Then, the small molecule: gp120 complexes were set up for crystallization using vapor diffusion at 20 °C. Crystals grew in a mixture of 0.5 μl protein-small molecule complex and 0.5 μl of reservoir solution containing 8–10% (v/v) PEG 8000, 5% iso-propanol, 0.1 M HEPES (pH 7.5). Crystals were soaked in cryo-protection solution containing 30% Ethylene glycol, 12% PEG 8000, and 0.1 M HEPES (pH 7.5), and were flash frozen in liquid nitrogen. Data were collected on beamline SERCAT ID-22 at the Advanced Photon Source, and processed with HKL2000.93 The structures were solved by molecular replacement with PHENIX94 using the coordinates of unliganded clade A/E93TH057 gp120 coree (PDB ID 3TGT). The initial Fo-Fc map generated after a rigid body refinement clearly indicated the electron densities of (−)-12 and (+)-12 and allowed us to place them into the densities manually using COOT.95 The initial densities of 4 and (+)-6, however, were not as clear as those found in (−)-12 and (+)-12, specifically in the Region III. After several rounds of refinement using PHENIX94, the R and values converged to 18.1–20.3% and 20.4–23.7%, respectively (Table 3). The geometry of the refined model was checked with Molprobity.96 Figures 1, ,2,2, ,77–9 were generated by PyMOL.97 To facilitate analysis of protein-ligand interactions in the context of the crystal structures, hydrogen atoms were added to all atoms within a 4.5 Å radius of the small molecule ligand and were minimized in the Merck Molecular Force Field 74, 75.
We thank Irwin Chaiken and Wayne Hendrickson and all the members of the PO1 Consortium Structure-Based Antagonism of HIV-1 Envelope Function in Cell Entry. Funding was provided by NIH GM 56550 to JL, EF, ABS, and JS. JL thanks the Pittsburgh Supercomputing Center for an allocation for computing resources #MCB090108.
Coordinates and structure factors have been deposited in the Protein Data Bank with accession numbers: 4DKO, 4DKP, 4DKQ, and 4DKR.