|Home | About | Journals | Submit | Contact Us | Français|
Betulinic acid (BA) derivatives can inhibit human immunodeficiency virus type 1 (HIV-1) entry or maturation depending on side chain modifications. While BA derivatives with antimaturation activity have attracted considerable interest, the anti-HIV-1 profile and molecular mechanism of BA derivatives with anti-HIV-1 entry activity (termed BA entry inhibitors) have not been well defined. In this study, we have found that two BA entry inhibitors, IC9564 and A43D, exhibited a broad spectrum of anti-HIV-1 activity. Both compounds inhibited multiple strains of HIV-1 from clades A, B, and C at submicromolar concentrations. Clade C viruses were more sensitive to the compounds than clade A and B viruses. Interestingly, IC9564 at subinhibitory concentrations could alter the antifusion activities of other entry inhibitors. IC9564 was especially potent in increasing the sensitivity of HIV-1YU2 Env-mediated membrane fusion to the CCR5 inhibitor TAK-779. Results from this study suggest that the V3 loop of gp120 is a critical determinant for the anti-HIV-1 activity of IC9564. IC9564 escape viruses contained mutations near the tip of the V3 loop. Moreover, IC9564 could compete with the binding of V3 monoclonal antibodies 447-52D and 39F. IC9564 also competed with the binding of gp120/CD4 complexes to chemokine receptors. In summary, these results suggest that BA entry inhibitors can potently inhibit a broad spectrum of primary HIV-1 isolates by targeting the V3 loop of gp120.
Human immunodeficiency virus type 1 (HIV-1) is the causative agent of AIDS. Many antiretroviral drugs have been developed to treat HIV-1 infection. In particular, the introduction of highly active antiretroviral therapy has been shown to reduce plasma viral loads to undetectable levels in many patients (11, 23). Despite this potent therapeutic effect, it is clear that the virus is suppressed rather than eradicated by highly active antiretroviral therapy (7, 10, 36). In addition, drug-related toxicities and the emergence of resistant viruses often compromise the effectiveness of treatment. Thus, novel drugs with unique modes of action are needed to improve the current treatment options for HIV-1 infection.
While no treatment that completely eradicates the virus is currently available, many drugs can inhibit HIV-1 at different stages of replication. Initial events in HIV-1 infection involve virus attachment and entry into host cells. Multiple steps in the entry process could be exploited for therapeutic intervention. HIV-1 entry is initiated when the HIV-1 surface envelope glycoprotein gp120 engages its cellular receptor, CD4, and subsequently its coreceptor, usually either CCR5 or CXCR4. These initial events trigger conformational changes in HIV-1 envelope glycoproteins and expose the fusion peptide of gp41 to initiate membrane fusion.
Potential antiviral targets on gp120 include the CD4 binding site and the chemokine receptor interactive sites. These potential drug targets are functional motifs in gp120 that are involved in virus entry. Hence, many anti-HIV-1 agents that target these key functional motifs have been identified. For example, sulfated polysaccharides, including heparin sulfate, dextran sulfate, dextrin sulfate, and cyclodextrin sulfate, were shown to exhibit anti-HIV activity by interacting with the V3 region of gp120 (1, 3, 5). In addition, there has been tremendous interest and effort to develop drugs that can block CD4/gp120 interactions. Recombinant proteins, such as the soluble CD4-immunoglobulin fusion protein PRO 542 (33) and the bacterial polypeptide cyanovirin (4), interfere with CD4/gp120 binding. The small molecule BMS-378806 was previously reported to inhibit many HIV-1 isolates by blocking the binding of gp120 to CD4 (21). In a separate study, BMS-378806 was shown to inhibit the CD4-induced exposure of the N-terminal heptad repeat of gp41 with no significant effect on CD4/gp120 binding (30).
In an effort to identify HIV-1 entry inhibitors with novel mechanisms of action, we have been synthesizing small molecules using betulinic acid (BA) as a scaffold. The viral targets of BA derivatives can vary depending on where their side chains reside. For example, IC9564, with a side chain at position 28, is an entry inhibitor (12). On the other hand, 3-O-(3′,3′dimethylsuccinyl) BA (DSB) (also termed PA-457 or Bevirimat), with a side chain at position 3, is a potent maturation inhibitor (18, 38). During HIV-1 maturation, the viral protease cleaves specific sites in the p55 Gag protein to form the viral structural proteins MA, CA (p24), NC, and p6 as well as the Gag spacer peptides SP1 (p2) and SP2 (34). DSB interferes with the processing of the p24/p2 cleavage site. While the mechanism of antimaturation activity is relatively well characterized for DSB, the molecular details of how BA entry inhibitors block HIV-1 Env-mediated membrane fusion are unclear. Mapping of the drug target by using drug-resistant viruses has been difficult. Our previous results indicated that the virus conceals the putative drug binding site by altering the conformation of gp120 rather than by changing the amino acid residues in the putative drug binding site (37).
In the present study, we characterized the spectra of antiviral activity and the molecular mechanisms of action of two BA entry inhibitors, IC9564 and A43-D (Fig. (Fig.1).1). Our results indicated that both compounds could potently inhibit a collection of primary HIV-1 isolates from clades A, B, and C. At subinhibitory concentrations, IC9564 significantly altered the antifusion activity of other entry inhibitors. The ability of IC9564 to increase the antientry activity of the R5 inhibitor TAK-779 was especially strong (greater than 100-fold). The strong influence of IC9564 on the antientry activity of TAK-779 suggested that the IC9564 binding site is likely to be intimately associated with the chemokine receptor binding sites in gp120. Indeed, data from this study suggest that the V3 loop of gp120, a domain involved in chemokine receptor binding, is also a critical determinant for the antientry activity of IC9564.
A collection of pseudoviruses derived from clades B and C used in this study was described previously (19, 20). Clade A viruses were kindly provided by Julie Overbaugh, Fred Hutchinson Cancer Research Center, Seattle, WA. V3 monoclonal antibody 447-52D was obtained from the NIH AIDS Research and Reference Reagent Program. Monoclonal antibody 39F was provided by James Robinson, Tulane University. COS, an African green monkey kidney cell line, was propagated in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (vol/vol) fetal bovine serum and 1% (vol/vol) penicillin-streptomycin (10,000 units/ml). TZM-bl cells (35), obtained from the NIH AIDS Research and Reference Reagent Program, were grown in DMEM with 10% (vol/vol) heat-inactivated fetal bovine serum and gentamicin at a concentration of 50 μg/ml. HOS.CCR5 and HOS.CXCR4 cells (8) were also obtained from the NIH AIDS Research and Reference Reagent Program. HOS.CCR5 and HOS.CXCR4 cells were cultured in DMEM with 10% (vol/vol) heat-inactivated fetal bovine serum supplemented with 1.0 μg/ml puromycin.
Inhibition of HIV-1 infection was measured as reduction in luciferase gene expression after a single round of virus infection of TZM-bl cells as described previously (19, 20, 35). Briefly, 200 50% tissue culture infective doses of virus was used to infect TZM-bl cells in the presence of various concentrations of compounds. Two days after infection, the culture medium was removed from each well, and 100 μl of Bright Glo reagent (Promega, San Luis Obispo, CA) was added to the cells for measurement of luminescence using a Victor 2 luminometer. The 50% inhibitory concentration (IC50) was defined as the concentration that caused a 50% reduction in the luciferase activity (relative light units [RLU]) compared to virus control wells.
Selection of IC9564-resistant M2-NLDH variants was carried out using MT4 cells. The drug-resistant mutants were generated using escalating concentrations of IC9564 starting at 2 μM, a concentration that inhibits 95% of virus replication, as previously described (12, 37). In addition to IC9564, the DH012-neutralizing antiserum C1206 was added to the culture medium at 1,000-fold dilution to suppress mutations that lead to conformational changes in gp120 (39). Viral replication was monitored by measuring the p24 level in the culture supernatants using an enzyme-linked immunosorbent assay (ELISA) kit from Perkin-Elmer (Waltham, MA). The HIV-1 Env gene of the selected IC9564-resistant variants was cloned into an Invitrogen TA cloning vector before sequencing.
Construction of Env mutants was achieved by using a QuikChange site-directed mutagenesis kit from Stratagene (La Jolla, CA). Two pairs of primers, 5′-C AAT ACA AGA AAA CGG ATA ACT CTA GGA-3′ and 5′-TCC TAG AGT TAT CCG TTT TCT TGT ATT G-3′ and 5′-CATTGAAGGAAAGGACACTATCACAC-3′ and 5′-GTGTGATAGTGTCCTTTCCTTCAATG-3′, were used to introduce G306R and N411K mutations into the M2-NLDH virus, respectively. The detailed mutagenesis protocol is described in the brochure provided by the manufacturer (Stratagene). The sequences of the desired mutants were confirmed by DNA sequencing.
A previously described HIV-1 infectivity assay was used in the experiments (14). A 96-well microtiter plate was used to set up the HIV-1 replication assay. HIV-1 at a multiplicity of infection of 0.001 was used to infect MT4 cells. Culture supernatants were collected on day 4 postinfection for p24 assays using an ELISA kit from Perkin-Elmer.
The fusion assay used in this study was described previously (14). The fusion assays were performed by transfecting monkey kidney cells (COS cells) with the expression vector pSRHS, which contains HIV-1 env and tat genes. COS cells (1 × 106 cells/ml) were mixed with 5 μg of the Env-expressing vectors and incubated on ice for 10 min. Electroporation was performed using a gene pulser (Bio-Rad, Hercules, CA), with the capacitance set at 950 μF and voltage at 150 V. The transfected COS cells were cultured for 1 day before being mixed with TZM-bl cells. TZM-bl cells were incubated with the Env-expressing COS cells in 96-well flat-bottom plates (Costar) overnight. Fusion was measured by quantifying the luciferase activity in the fused cells using a Bright-Glo luciferase assay kit and instructions provided by the manufacturer (Promega, San Luis Obispo, CA). Inhibition of the Env-mediated membrane fusion was expressed as a percentage of control Env-mediated membrane fusion in the absence of inhibitors.
A simple ELISA was used to assess the ability of the BA derivatives to compete for the binding of V3 antibodies to either HIV-1IIIB gp120 or HIV-1Ba-L gp120. HIV-1IIIB gp120 was purchased from ImmunoDiagnostic Inc., Woburn, MA. HIV-1Ba-L gp120 was obtained from the NIH AIDS Research and Reference Reagent Program. HIV-1IIIB and HIV-1Ba-L are X4 and R5 viruses, respectively. A detailed protocol for the ELISA was described previously (6). Briefly, gp120 at 10 μg/ml was coated onto ELISA plates. The ELISA plates were blocked with 3% bovine serum albumin in Tris-buffered saline (500 mM NaCl, 20 mM Tris [pH 7.5]) before incubation with 39F or 447-52D. Antibody binding was carried out in the presence of various concentrations of the BA derivative IC9564. A horseradish peroxidase-conjugated goat anti-human antiserum was used to detect 39F or 447-52D binding. The ELISA plates were read with a microplate reader (Bio-Tek, Winooski, VT) at an optical density at 490 nm to determine the binding of the antibodies.
The binding of the gp120/CD4 complex to chemokine receptors was performed by using a modified method described previously (9). HOS.CCR5 or HOS.CXCR4 cells (100 μl at 1 × 106 cells/ml) were plated onto poly-d-lysine-coated plates (Becton Dickinson) and incubated at 37°C overnight. A 1:1 mix of soluble CD4 (4.5 nM in culture medium) and HIV-1 gp120 (Ba-L or IIIB) was incubated at room temperature for 15 min. HIV-1Ba-L gp120 and soluble CD4 were obtained from the NIH AIDS Research and Reference Reagent Program. The gp120/CD4 complex was added to the cells in the presence of 5 μg/ml of compounds and incubated at 37°C for 1 h. A fluorescein isothiocyanate-conjugated anti-gp120 antibody (Biocat, Heidelberg, Germany) was added to each well and incubated for 1 h. After washing three times with phosphate-buffered saline, the bound gp120 was detected with a fluorometer (Perkin-Elmer, Shelton, CT) at 490 nm/520 nm. Nonspecific binding was performed by replacing the gp120/CD4 complex with gp120. The specific binding of the gp120/CD4 complex in the presence of the compounds was expressed as a percentage of the specific binding in the absence of the testing compounds. Specific binding is derived by subtracting the nonspecific binding from the observed gp120/CD4 binding.
Docking of IC9564 to the V3 loop region of the gp120 protein was performed using the MOE package (version 2006.08; Molecular Computing Group, Inc.). The X-ray structure of an HIV-1 gp120 core complex (Protein Data Bank accession number 2B4C) from the Protein Data Bank (www.pdb.org) was edited by retaining its gp120 core sequence only. Hydrogen atoms were added to the remaining structure, and partial charges were calculated by using default parameter settings. IC9564 was subsequently docked to the V3 loop region (residues 296 to 330) using an alpha triangle placement method. The docking complexes were sorted by their affinity dG and ASE scores separately. dG and ASE are two scoring functions designed to measure the binding affinity between a ligand and its binding protein. The affinity dG score is an energy-based scoring function and measures favorable interactions between a ligand and the receptor binding site. The ASE score is based primarily on the shape fit between the two. Each scoring function has its own strength. Top-scored complexes were visually analyzed, and the one best-fitting experimental datum was further energy minimized.
In order to determine the spectrum of anti-HIV-1 activity, IC9564 and A43-D were tested against a panel of HIV-1 reference strains representing genetic subtypes A, B, and C (19, 20). Approximately 90% of all HIV-1 isolates belong to these three subtypes. Among them, clade C viruses account for approximately 50% of all HIV-1 isolates globally. The reference strains all possessed an R5 (CCR5-using) phenotype except for NL4-3 and HXB2, which are T-cell-line-adapted X4 viruses.
In general, A43-D was approximately two- to threefold more potent against the HIV-1 isolates than IC9564 and the fusion inhibitor T20 (enfuvirtide) (Table (Table1).1). A43-D was synthesized for the purpose of determining whether the carboxylic group on IC9564 is required for its anti-HIV-1 entry activity. The results clearly indicated that the uncharged A43-D was slightly more potent than IC9564. In other words, a negative charge was not required for BA derivatives to inhibit HIV-1 entry. The highest concentration of the compounds used for the assays was 15 μM, which did not affect TZM-bl cell growth in the assays. The IC50 of the compounds on TZM-bl cells was greater than 50 μM in a 2-day assay using trypan blue exclusion to enumerate viable cells.
Among the HIV-1 subtypes tested, clade C viruses were most sensitive to the inhibition of A43-D, which reduced clade C virus infection by 50% (IC50) at an average concentration of 0.206 μM. On the other hand, the average IC50 values for A43-D against clade A and clade B viruses were 0.91 and 0.53 μM, respectively. Thus, the clade C viruses were approximately three- to fourfold more sensitive to inhibition by A43-D than the clade A and B viruses. More viruses from different clades need to be tested to determine whether clade C viruses are indeed more sensitive to the compounds than other genetic subtypes.
Our previous studies suggested that gp120 was the target of IC9564 (12, 37). The differential sensitivity of these HIV-1 subtypes to the BA entry inhibitors might be related to the genetic diversity associated with gp120. In contrast to the BA derivatives, the average IC50 of T20 against clades A, B, and C ranges from 1.2 μM to 1.5 μM. Perhaps the relatively conserved target of T20, the N-terminal heptad repeat of gp41, contributed to the smaller variation in anti-HIV-1 activity.
As a step toward dissecting the molecular mechanism of action, IC9564 was used as a model compound for its ability to interact with other HIV-1 entry inhibitors. Our hypothesis was that IC9564 should affect the potency of other entry inhibitors if their targets are either physically in close proximity or functionally linked to IC9564. To test this hypothesis, one entry inhibitor that targets gp120 (BMS-378806) and two that target chemokine receptors (TAK-779 and AMD3100) were chosen for study. Antifusion activity was determined in the presence of subinhibitory concentrations of IC9564. The Envs derived from the R5 virus YU-2 and the X4 virus NL4-3 were used in this series of experiments. YU-2 Env-mediated membrane fusion was used to determine the effect of IC9564 on the antifusion activity of the CCR5 inhibitor TAK-779 and the gp120 inhibitor BMS-378806. NL4-3 Env-mediated membrane fusion was used to determine the effect of IC9564 on the antifusion activity of the CXCR4 inhibitor AMD3100. NL4-3 Env-mediated membrane fusion is approximately 30-fold more sensitive than YU-2 Env-mediated membrane fusion to IC9564 (Fig. (Fig.2A).2A). BMS-378806 inhibited the YU-2 Env-mediated membrane fusion by 50% (IC50) at 56 ng/ml (138 nM) in the absence of IC9564. On the other hand, the IC50 for BMS-378806 was 14 ng/ml (34.5 nM) in the presence of 2 μg/ml (2.6 μM) of IC9564. The subinhibitory concentration of IC9564, 2 μg/ml (2.6 μM), altered the antifusion activity of BMS-378806 by fourfold (Fig. (Fig.2B).2B). IC9564 also exhibited a similar impact on the antifusion activity of AMD3100 (Fig. (Fig.2C2C).
Although IC9564 could affect the anti-HIV-1 activity of BMS-378806 and AMD3100, the effect of IC9564 on the antifusion activity of TAK-779 was much more prominent. A subinhibitory concentration of IC9564, 2 μg/ml (2.6 μM), could significantly change the antifusion potency of TAK-779 (Fig. (Fig.2D).2D). The IC50 for TAK-779 against YU-2 Env-mediated membrane fusion was 48 ng/ml (90 nM). On the other hand, it took only 0.053 ng/ml (0.1 nM) of TAK-779 to achieve the same effect in the presence of 2 μg/ml (2.6 μM) of IC9564. Such a drastic influence of anti-HIV-1 potency by IC9664 raised the possibility that the targets of IC9564 and TAK-779 might be physically in close proximity or functionally linked during HIV-1 entry. The strong impact of IC9564 on the antifusion activity of TAK-779 was not shared by the gp120 inhibitor BMS-378806. At a subinhibitory concentration, 10 ng/ml (25 nM), BMS-378806 altered the antifusion activity of TAK-779 by less than twofold (data not shown).
We previously showed that mutations in gp120 resulted in a conformational change in gp120 and allowed the virus to evade the antientry activity of IC9564 (37). In many ensuing efforts to develop IC9564-resistant mutants, the virus opted to change the conformation of gp120 to conceal the drug binding site instead of undergoing mutations in the putative drug binding site to escape IC9564 (data not shown). One strategy to overcome this obstacle is to compromise the ability of HIV-1 to evolve mutations that result in changing the conformation of gp120 and concealing the drug target. This strategy involved the use of M2-NLDH virus (39) and the chimpanzee immune serum C1206 (40). M2-NLDH is a C1206-resistant, IC9564-sensitive virus derived from DH012 (39). IC9564-resistant mutants derived from M2-NLDH reversed to an envelope phenotype similar to that of wild-type DH012 virus (37), which was very sensitive to C1206 and resistant to IC9564. C1206 at a 1,000-fold dilution completely neutralized DH012 infection. Thus, the selection of IC9564-resistant mutants from M2-NLDH in the presence of C1206 is expected to suppress mutations that lead to the conformational change and to increase the chance of evolving mutations that are near or within the drug binding site.
Under this selection condition, three mutations were detected in the envelope glycoproteins of the IC9564 escape variant M2-IC5/CP (Fig. (Fig.3A).3A). Two mutations, G306R and N411K, were in gp120 (Fig. (Fig.3B).3B). M2-IC5/CP did not have a change at amino acid residue 198 under the selection condition (Fig. (Fig.3B).3B). This result suggested that the chimpanzee serum C1206 successfully suppressed a change at amino acid residue 198 that can lead to IC9564 resistance by changing the conformation of gp120 (37). Another mutation is an I37T mutation in gp41. The I37T mutation was implicated in drug resistance associated with T20/DP178 (24). The I37T mutation was reintroduced into M2-NLDH by using site-directed mutagenesis. This gp41 mutation did not significantly affect IC9564 sensitivity (data not shown). Amino acid residue 306 in the V3 loop was shown to play a key role in determining chemokine receptor binding (2). The other gp120 mutation, N411K, is located in the N terminus of the β-19 strand, which was implicated in chemokine receptor binding (16, 25). The introduction of the N411K mutation into M2-NLDH resulted in a virus, M2-NLDH-N411K, that did not significantly affect IC9564 sensitivity (Fig. (Fig.3A).3A). On the other hand, the introduction of the G306R mutation into M2-NLDH resulted in a virus, M2-NLDH-G306R, that exhibited a degree of resistance to IC9564 similar to that of the IC9564 escape variant M2-IC5/CP (Fig. (Fig.3A).3A). This result suggested that the V3 loop is a key determinant in the entry inhibition activity of IC9564.
Although drug-resistant mutants were selected under conditions that favor mutations at or near the drug binding site, it remained possible that the V3 mutation G306R arose to conceal a distal IC9564 binding site. To determine whether V3 is indeed the target for IC9564, experiments to test whether IC9564 could compete with the binding of the V3-specific monoclonal antibodies 447-52D and 39F were performed. 447-52D recognizes an epitope at the tip of V3 (41), whereas the amino acid residues N terminal to the tip (amino acids 304 and 305) are critical for 39F binding (22). Binding of these antibodies to HIV-1IIIB gp120 was performed in the presence of various concentrations of IC9564. IC9564 inhibited the binding of these monoclonal antibodies to gp120 in a dose-dependent manner (Fig. (Fig.4A).4A). In contrast, IC9564 did not affect the binding of monoclonal antibody 2G12 to gp120. 2G12 recognizes a cluster of mannose residues on the outer face of gp120 (26, 28). Since IC9564 can inhibit both R5 and X4 viruses, gp120 derived from the R5 virus HIV-1Ba-L was also used in the V3 antibody binding assay. IC9564 effectively competed with the binding of 447-52-D to HIV-1Ba-L gp120 (Fig. (Fig.4B).4B). In addition, IC9564 did not compete with the binding of 2G12 to HIV-1Ba-L gp120. The ability of IC9564 to compete with the V3-specific antibodies supports the notion that the V3 loop is a key determinant for the anti-HIV-1 activity of the BA derivatives. However, the results did not rule out the possibility that IC9564 might have interacted with a distal site and transformed the V3 loop into a conformation that was less reactive with the V3 monoclonal antibodies.
The results of this study suggested that the V3 loop of gp120 is the likely target for a class of BA derivatives that inhibit viral entry. However, how these compounds bind the V3 loop and inhibit entry was not clear. It is possible that the binding of BA derivatives interferes with interactions of gp120 and the chemokine receptor. To test this possibility, IC9564 was tested to determine whether it competed with gp120/CD4 complex binding to chemokine receptors on HOS.CCR5 or HOS.CXCR4 cells. The gp120 derived from HIV-1Ba-L and derived from HIV-1IIIB were used for binding to HOS.CCR5 and HOS.CXCR4 cells, respectively. IC9564 at 5 μg/ml strongly inhibited the binding of HIV-1Ba-L gp120/CD4 or HIV-1IIIB gp120/CD4 complexes to HOS.CCR5 or HOS.CXCR4 cells (Fig. (Fig.5).5). BA, which did not inhibit HIV-1 entry, was used as a negative control for the binding assay. BA did not compete with the binding of gp120/CD4 complexes to the chemokine receptors on the HOS cells. These results strongly suggested that IC9564 interfered with the binding of the gp120/CD4 complex to chemokine receptors.
Using the X-ray crystal structure of an HIV-1 gp120 core with the V3 loop as a template (13), IC9564 was docked into the V3 loop region. A total of 100 different docking conformations were generated. In the energy-minimized model, the A ring of the BA scaffold resided at the base of the V3 loop, while the C-28 side chain was lining with the charged amino acid residues in the N-terminal half of the loop near the V3 tip (Fig. (Fig.66).
One of the most interesting aspects of BA derivatives is that the locations of side chains can orchestrate their mechanisms of action. A side chain in the C-3 position is needed for anti-HIV-1 maturation activity (18, 38). On the other hand, a C-28 side chain on the BA molecule is required for anti-HIV-1 entry activity (12, 31). The most well-characterized anti-HIV-1 maturation BA derivative is DSB (PA-457). DSB does not inhibit HIV-1 entry, but it is a potent anti-HIV-1 maturation inhibitor (18, 38). DSB was shown to be active against many HIV-1 isolates and is currently in phase II clinical trials. In contrast to the BA maturation inhibitors, the molecular mechanism of BA entry inhibitors was elusive. Although BA entry inhibitors have been shown to potently inhibit T-cell-adapted viruses such as HIV-1IIIB or NL4-3, their spectra of anti-HIV-1 activity were poorly characterized. The data presented in this study clearly indicated that IC9564 and A43-D exhibited broad-spectrum anti-HIV activity at submicromolar concentrations. It is worth noting that clade C viruses, which account for approximately 50% of all HIV-1 isolates in the world, are relatively more sensitive to the BA entry inhibitors than are clade A and clade B viruses.
All the viruses listed in Table Table11 are R5 viruses, except the X4 viruses HXB2 and NL4-3. One BA entry inhibitor, RPR103611, was shown to have antientry activity against some X4 viruses, with little activity against R5 viruses (17). Since RPR103611 and IC9564 are steroisomers, their mechanisms of anti-HIV action are likely to be the same. However, it remains a possibility that a minor structural difference in BA derivatives could affect their mechanisms of action. For example, PA-457 is a BA derivative that inhibits HIV-1 maturation without anti-HIV-1 entry activity.
The purpose of studying the interaction between IC9564 and other HIV-1 entry inhibitors was to gain insights into where the drug binding site resides. The strong influence of IC9564 on the antifusion activity of TAK-779 suggests that the IC9564 binding site might be intimately related to chemokine receptor binding sites in gp120. The binding of IC9564 might have resulted in an HIV-1 Env structure that retains a minimal function for chemokine receptor binding but that has reached a threshold that is susceptible to the CCR5 inhibitor TAK-779. It will be interesting to determine whether IC9564 can also strongly affect the anti-HIV-1 entry activity of other classes of CCR5 inhibitors. Since the experiments were designed to study the mechanism of action of IC9564 rather than the synergy between IC9564 and other entry inhibitors, the degree of synergy between the BA derivatives and other HIV-1 entry inhibitors remains to be determined.
The ability of the BA derivatives to inhibit primary HIV-1 isolates across the most prevalent genetic subtypes (clades A, B, and C) suggests that these compounds target a conserved functional motif in gp120. However, results obtained with drug-resistant viruses suggested that the V3 loop of gp120 is the main target. The strong competition of IC9564 with V3 monoclonal antibody binding to gp120 also supports the notion that V3 is the likely target of the compounds. The V3 region, by definition, contains a highly variable sequence. Thus, the broad anti-HIV activity of uncharged, HIV-1-specific A43-D was unexpected. It is likely that the drugs interact with a conserved functional motif in V3. Despite considerable sequence variation, a conserved V3 structural motif is likely needed to interact with chemokine receptors for virus-cell fusion to take place. Such a conserved V3 feature was implicated in previous studies on the structure of the V3 loop (13, 29, 32). Some V3 monoclonal antibodies, such as 447-52D, were shown to neutralize many primary isolates (41). Although IC9564 can compete with 447-52D for V3 binding, their mechanisms of action might not be identical. For example, all the clade B viruses used in this study are sensitive to the BA entry inhibitors, whereas only two of the clade B viruses tested, SS1196.1 and 6535.3, were sensitive to 447-52D (Table (Table1).1). SS1196.1 and another clade B virus, WITO4160.33, exhibited similar sensitivities to the BA derivatives. On the other hand, WITO4160.33 was not sensitive to 447-52D. Therefore, the BA entry inhibitors might have targeted a structural motif different from that of the 447-52D epitope. A similar argument against the notion that the G306R mutation in the IC9564-resistant virus is where the drug binding site resides could also be made. For example, HXB2 virus, which possesses an Arg (R) at amino acid residue 306, is very sensitive to the BA derivatives (Table (Table1).1). The G306R mutation in gp120 of M2-NLDH, which occurred at a site known to be a key determinant for coreceptor tropism, might result in a V3 conformation that is unfavorable for interactions with the BA entry inhibitors.
The V3 loop is one of the key structural motifs that are critical for chemokine receptor binding. It appears that the binding of BA derivatives to the V3 loop interfered with interactions between gp120 and chemokine receptors. The interaction of gp120 with chemokine receptors is thought to trigger subsequent conformational changes in gp41. This is in agreement with our previous observations that IC9564 induced an aberrant conformational change in gp120 and inhibited the conformational changes in gp41 (15). Thus, the BA derivatives could inhibit HIV-1 entry processes by inducing a fusion-incompetent gp120 conformation that is no longer capable of interacting with chemokine receptors and triggering the subsequent conformational changes in gp41 needed for fusion.
The computer docking model for the interaction between the BA derivatives and the V3 loop (Fig. (Fig.6)6) appears to agree with our previous structure-activity relationship study. For example, a bulky C-3 side chain of the A ring in the BA scaffold compromised the antientry activity of BA derivatives (14). It is possible that a bulky C-3 side chain negatively impacts the ability of the A ring to fit into the ligand binding pocket in the V3 base. Furthermore, the C-28 side chain is where the anti-HIV-1-entry activity resides. Alignment of the C-28 side chain with the V3 stem toward the N-terminal half of the V3 loop might impact the ability of V3 to interact with other structural motifs in gp120 and chemokine receptors.
In summary, the BA derivatives studied here were shown to possess potent broad-spectrum anti-HIV-1 activity. The most prevalent clade C viruses were relatively sensitive to this activity. In addition to broad anti-HIV-1 activity, IC9564 was able to affect the antifusion activity of other entry inhibitors. The greater-than-2-log10 alteration in the anti-HIV-1 entry activity of the R5 inhibitor TAK-779 is remarkable. Although the ability of IC9564 to sensitize HIV-1 to TAK-779 was unexpected, the most surprising result was that the V3 region of gp120 is a critical determinant for the antientry activity of the BA derivatives. The inhibition of a variety of primary HIV-1 isolates by targeting V3 supports the notion that a conserved structural motif exists in the V3 region. Such a conserved V3 structure could be an interesting target for both drug and vaccine development against HIV-1 infection.
This study is supported by NIH grants AI52022 and AI65310.
Published ahead of print on 22 October 2007.