Derivatives of Mesoxalic Acid Block Translocation of HIV-1 Reverse Transcriptase*

Background: The 4-chlorophenylhydrazone of mesoxalic acid (CPHM) is a known inhibitor of HIV-1 reverse transcriptase (RT). Results: We demonstrate that CPHM traps the pre-translocational conformation of the RT-DNA complex. Conclusion: The data validate this complex as a possible drug target. Significance: This work can therefore contribute to the development of novel classes of antiretroviral agents. The pyrophosphate mimic and broad spectrum antiviral phosphonoformic acid (PFA, foscarnet) was shown to freeze the pre-translocational state of the reverse transcriptase (RT) complex of the human immunodeficiency virus type 1 (HIV-1). However, PFA lacks a specificity domain, which is seen as a major reason for toxic side effects associated with the clinical use of this drug. Here, we studied the mechanism of inhibition of HIV-1 RT by the 4-chlorophenylhydrazone of mesoxalic acid (CPHM) and demonstrate that this compound also blocks RT translocation. Hot spots for inhibition with PFA or CPHM occur at template positions with a bias toward pre-translocation. Mutations at active site residue Asp-185 compromise binding of both compounds. Moreover, divalent metal ions are required for the formation of ternary complexes with either of the two compounds. However, CPHM contains both an anchor domain that likely interacts with the catalytic metal ions and a specificity domain. Thus, although the inhibitor binding sites may partly overlap, they are not identical. The K65R mutation in HIV-1 RT, which reduces affinity to PFA, increases affinity to CPHM. Details with respect to the binding sites of the two inhibitors are provided on the basis of mutagenesis studies, structure-activity relationship analyses with newly designed CPHM derivatives, and in silico docking experiments. Together, these findings validate the pre-translocated complex of HIV-1 RT as a specific target for the development of novel classes of RT inhibitors.

The reverse transcriptase (RT) of the human immunodeficiency virus type 1 (HIV-1) is a virally encoded RNA-and DNAdependent polymerase that represents an important target for antiretroviral therapeutics (1)(2)(3)(4). RT also possesses a ribonuclease H (RNase H) activity that degrades the RNA of RNA/ DNA replication intermediates. Two different classes of approved antiviral drugs target the polymerase active site (5) as follows: nucleoside analog RT inhibitors (NRTIs) 3 that compete with natural deoxynucleoside triphosphate (dNTP) pools for binding and incorporation (3,6) and the chemically diverse class of non-nucleoside analog RT inhibitors that bind to a hydrophobic pocket 10 Å away from the active site (7,8).
Phosphonoformic acid (PFA, foscarnet) (compound 1) ( Fig.  1) targets HIV-1 RT and other viral DNA polymerases and shows a broad spectrum of antiviral activities against HIV and several members of the herpesviridae, including herpes simplex virus types 1 and 2 (HSV1 and HSV2) and the human cytomegalovirus (HCMV) (9). However, toxic side effects limit its clinical utility. PFA is currently only approved as a component in second line regimens to treat infection with HCMV (10). Despite such unfavorable properties as a drug, the mechanism of action is unique and could potentially be exploited in efforts to design novel classes of RT inhibitors. PFA interferes with the translocation of RT between two distinct cycles of nucleotide incorporation (11)(12)(13). Catalytic incorporation of nucleotides by RT occurs stepwise (14,15). Following binding of the primer-template (P/T), the incoming deoxynucleoside triphosphate (dNTP) binds to the nucleotide-binding site (N-site). The chemical step is preceded by a conformational change in the flexible "fingers" subdomain that traps the nucleotide substrate in a closed conformation. Inorganic pyrophosphate (PP i ) is generated concomitantly with formation of a new phosphodiester bond (16,17). Following the release of PP i , RT has to translocate to shift the newly synthesized primer terminus to the priming site (P-site). Several lines of evidence suggest that the polymerase establishes a dynamic equilibrium between preand post-translocational conformations. Site-specific footprinting experiments revealed that the sequence of the nucleic acid substrate can influence this equilibrium (18). Certain sequences show an equal distribution, although other sequences show a bias toward pre-or post-translocation. Increasing concentrations of the next nucleotide substrate can convert a population of predominantly pre-translocated complexes into a population of predominantly post-translocated complexes (11,19). Conversely, PFA traps the pre-translocated complex of RT, which prevents binding of the next dNTP substrate (11).
It is currently unknown whether this type of inhibition is limited to small molecules that are structurally highly related to PFA or whether more complex compounds that provide specificity to HIV-1 RT are also capable of blocking RT translocation through a similar mechanism of action. The 4-chlorophenylhydrazone of mesoxalic acid (CPHM) (compound 2) (Fig. 1) is a potential candidate in this context. This compound was originally identified as a strand transfer inhibitor of HIV-1 RT (20,21). CPHM was shown to inhibit the RT-associated RNase H activity under assay conditions that allow multiple turnovers. Moreover, at high concentrations, CPHM delays the second of two consecutive nucleotide incorporation events. These findings pointed to a defect in RT translocation; however, the underlying mechanism remained elusive. Here, we demonstrate that CPHM, like PFA, binds to the polymerase active site and traps the pre-translocated complex. These findings demonstrate that the pre-translocated RT complex can be targeted by structurally diverse compounds, which provide novel opportunities in the development of antiretroviral drugs.

EXPERIMENTAL PROCEDURES
Enzymes, Nucleic Acids, and Small Molecules-Heterodimeric p66/p51 reverse transcriptase was expressed and purified as described previously (22). Wild-type HCMV polymerase (UL54) was derived from recombinant viruses generated by overlapping cosmids (23). The UL54 coding sequence was kindly provided by Dr. Guy Boivin (Laval University). The UL54 coding sequence was cloned into pPR-IBA1 (IBA) using the BsaI site to generate pPR-IBA1/UL54. This construct facilitates protein purification through Strep-tag affinity chromatography (IBA). D542A substitution was introduced to remove the 3Ј-5Јexonuclease activity. The amino acid substitution was introduced with PfuUltra DNA polymerase (Stratagene) according to the manufacturer's recommendations. The HCMV polymerase UL54 was expressed in insect cells using the Bac-to-Bac baculovirus expression system (Invitrogen) according to the manufacturer's recommendations. UL54 was purified using Strep-tag affinity chromatography (IBA) according to the manufacturer's recommendations. Oligo(deoxy)ribonucleotides were synthesized and purchased from Integrated DNA Technologies. Sequences are provided in the figures. Phosphonoformic acid was purchased from Sigma. CPHM and derivatives were provided by Dr. Sidney Hecht (Arizona State University) and Merck (West Point).
DNA Synthesis-To monitor the incorporation of multiple nucleotides by only HIV-1 RT, a 3-fold molar excess of PPT-57 DNA template was hybridized to 50 nM 5Ј-radiolabeled PPT DNA primer. The hybrid was heat-annealed in a buffer containing 50 mM Tris-HCl, pH 7.8, and 50 mM NaCl and then incubated with 500 nM RT in a buffer containing 50 mM Tris-HCl, pH 7.8, 50 mM NaCl, 0.3 mM EDTA, and 2 M each of dATP, dTTP, dGTP, and dCTP. The samples were treated with either no inhibitor, 20 M PFA, or 200 M CPHM. DNA synthesis was initiated by the addition of 6 mM MgCl 2 at 37°C. The reaction was stopped with formamide-loading dye containing xylene cyanol and bromphenol blue. Samples were resolved on a 15% denaturing polyacrylamide gel followed by phosphorimaging (Amersham Biosciences). To monitor the incorporation of multiple nucleotides by both HIV-1 RT and HCMV UL54, 100 nM DNA/DNA template-primer hybrid T1/P1 was preincubated for 10 min at 37°C with a given DNA polymerase (30 nM HIV-1 RT or 4 nM HCMV UL54) in a buffer containing 25 mM Tris-HCl, pH 8, 50 mM NaCl, 0.5 mM dithiothreitol (DTT), 0.2 mg/ml bovine serum albumin, 5% glycerol, and 10 M dNTP mix in the presence or absence of increasing concentrations of PFA or CPHM. Nucleotide incorporation was initiated by the addition of MgCl 2 to a final concentration of 10 mM, and the reactions were allowed to proceed for 30 min. The reactions were stopped by the addition of 3 reaction volumes of formamide containing traces of bromphenol blue and xylene cyanol. Samples were then subjected to 15% denaturing PAGE followed by phosphorimaging.
Electrophoretic Mobility Shift Assay (EMSA)-The formation of ternary complexes was monitored with a 5Ј-radiolabeled primer. The labeled primer was annealed to a 5-fold molar excess of the template. The DNA hybrid (50 nM) was then incubated with 500 nM WT HIV-1 RT. Increasing concentrations of PFA and CPHM were added to each sample and incubated for 10 min at room temperature. The complexes were subsequently challenged with 3 g/l heparin, followed by incubation for 1 h at room temperature. The samples were resolved on 6% nondenaturing polyacrylamide gels, analyzed with a phosphorimager (Amersham Biosciences), and quantified with Quantity One and ImageQuant software.
Site-specific Footprinting-Site-specific footprints with Fe 2ϩ were monitored on 5Ј-end-labeled DNA templates. The template (100 nM) was hybridized with the complementary primer (300 nM) as described above. The hybridization was conducted in a buffer containing 20 mM sodium cacodylate, pH 7, and 20 mM NaCl. The duplex was incubated with 750 nM HIV-1 RT in a buffer containing 120 mM sodium cacodylate, pH 7, 20 mM NaCl, and 6 mM MgCl 2 , in a final volume of 50 l. Prior to the treatment with Fe 2ϩ , complexes were preincubated at 37°C for 10 min with increasing concentrations of PFA or CPHM. Treatment with Fe 2ϩ was performed as described previously (18).
HIV-1 Replication Assays-Multiple cycle replication assays were performed with HIV isolate R8 in MT-4 human T-lymphoid cells as described (24).
Time Course of RNase H Activity-For steady-state (multiple turnover) RNase H reactions, 3Ј-radiolabeled RNA template (PBS-52r) was heat-annealed to a 3-fold excess of DNA primer (PBS-22Dpol). The RNA/DNA hybrid (150 nM) was incubated with 25 nM RT in a buffer of 50 mM Tris-HCl, pH 7.8, 50 mM NaCl, 500 M EDTA, and 6 mM MgCl 2 , in the absence or pres-ence of CPHM (100 M) or PFA (100 M). Reactions in the absence of inhibitor contained 0.1% dimethyl sulfoxide (DMSO). The reaction was allowed to proceed at 37°C and stopped by the addition of 100% formamide containing traces of xylene cyanol and bromphenol blue. RNA fragments were resolved on a 12% denaturing polyacrylamide gel and visualized by phosphorimaging. Pre-steady-state (single turnover) RNase H reactions were conducted using a Kin-Tek RQF-3 rapid quench-flow apparatus. RNA/DNA hybrids were prepared as described above to a final concentration of 100 nM and incubated with 1000 nM RT in a buffer containing 50 mM Tris-HCl, pH 7.8, 50 mM NaCl, 500 M EDTA and 6 mM MgCl 2 , in the presence or absence of CPHM or PFA (50 M). Primary and subsequent secondary RNase H cleavage events were monitored under the same conditions with 5Ј-radiolabeled chimeric RNA-DNA (PBS14r8d), heat-annealed to a 3-fold excess of DNA primer (PBS-22D). Increasing concentrations of CPHM or the RNase H inhibitor ␤-thujaplicinol (0, 0.8, 1.6, 3.2, 6.4, 12.5, 25, 50, and 100 M) were added to the reaction mixture to assess RNase H inhibition.
Molecular Modeling-Molecular models were generated using the crystal structure of the post-translocated ternary complex HIV-1 RT/chain-terminated hybrid/dTTP (Protein Data Bank code 1RTD) (25). The p66 subunit of HIV-1 RT, the chain-terminated hybrid, the two magnesium ions located at the polymerase active site, and dTTP were retained, and all other components were removed from the Protein Data Bank file. The ␤and ␥-phosphates of dTTP, as well as one of the oxygen atoms of the ␣-phosphate, were removed to generate a pseudo-pre-translocated complex. The structures of PFA and CPHM were downloaded from PubChem and were energyminimized for 100 cycles using Chimera (26). Protein and small molecular structures were prepared using AutoDockTools 1.5.4 (27); all hydrogens were added, and Gasteiger partial charges were assigned to the protein and the small molecules. Nonpolar hydrogens were merged with carbon atoms, and AD4 atom types were assigned. Protonation states for CPHM and PFA were assigned using AutoDock Tools 1.5.4.. Both nitrogens in the hydrazone moiety were partially protonated during the simulation. The protein and the inhibitors were saved as PDBqt files. A charge of ϩ2 was assigned to the magnesium ions. The grid box, of size 63 ϫ 63 ϫ 63, was centered at x ϭ 48.666, y ϭ 24.911, and z ϭ 41.55. The spacing in the grid box was 0.375 Å. PFA and CPHM were docked to the polymerase active site using a Lamarckian genetic algorithm. The docking algorithm parameters included 50 genetic algorithm runs, a randomized initial population of 150 inhibitor conformations, a maximum number of energy evaluations of 2,500,000, a maximum number of generations of 27,000, a rate of gene mutation of 0.02, and a rate of crossover of 0.8. Docking poses were scored using the AutoDock (27) empirical free energy function. Results that displayed a position root-mean-squared deviation of less than 0.5 Å were grouped into clusters; the lowest energy pose for each inhibitor was selected from the collection of docked poses of the highest ranked cluster.

RESULTS
Assessment of Antiviral Activity of CPHM-Although CPHM was shown to affect strand transfer reactions of HIV-1 RT, its potential as an antiviral agent remained to be assessed. We found that CPHM did not inhibit HIV replication in multiple cycle infectivity assays at a concentration up to 50 M nor did it exhibit any cytotoxicity. The control compounds efavirenz and indinavir exhibited IC 95 values of 23 and 34 nM when tested side-by-side with CPHM. The absence of antiviral effects could be due to a number of reasons related to restricted cellular uptake, compound stability, or limitations inherent to the particular mechanism of action. Here, we focused on detailed mechanistic studies with the goal to establish CPHM as a specific inhibitor of HIV-1 RT.
CPHM and PFA Induce Identical Pausing Patterns in DNA Synthesis by HIV-1 RT-To investigate the mechanism by which CPHM inhibits RT-mediated DNA synthesis, we initially examined the pattern of inhibition of CPHM and PFA during multiple nucleotide incorporation events (Fig. 1). The P/T system used in this experiment is derived from the PPT of the HIV-1 genome (28). The pausing patterns associated with CPHM and PFA, respectively, are identical in that inhibition of DNA synthesis occurs predominantly at positions ϩ3 and ϩ16. The two "hot spots" disappear with time, which shows that inhibition is reversible. Previous footprinting experiments revealed that P/T substrates corresponding to positions ϩ3 and ϩ16 shift the translocational equilibrium toward pretranslocation (11). Thus, these findings provide evidence to suggest that the mechanisms of action associated with the two inhibitors are highly related.
CPHM and PFA Trap the Pre-translocated Complex of HIV-1 RT-We subsequently used site-specific footprinting experiments to confirm that CPHM and PFA can stabilize the pretranslocated state of RT, thereby shifting the translocational equilibrium of RT at a given position to favor the pre-translocated state. Fe 2ϩ ions generate hydroxyl radicals in solution, which will selectively cleave the template at positions Ϫ17 or Ϫ18. A major cut at position Ϫ17 is indicative for the posttranslocated complex, whereas a cut at position Ϫ18 corresponds to the pre-translocated complex. Fe 2ϩ footprinting was performed using a primer-template system that biases RT to favor the post-translocated complex (Fig. 2). In the presence of increasing concentrations of either CPHM or PFA, the translocational equilibrium of WT RT is shifted to the pre-translocated complex, as shown by the change in the pattern of DNA cleavage products. These results are consistent with the RT pausing pattern and support the notion that CPHM and PFA act as inhibitors of the pre-translocated complex.
Binding of CPHM Requires Catalytic Metal Ions-Despite possible differences in binding of the two compounds, their negatively charged carboxylate and/or phosphonate moieties may interact with the bound divalent metal ions at the polymerase active site. Binary RT-P/T complexes are unstable when challenged with an enzyme trap, such as heparin. However, the complex can be stabilized in the presence of PFA, provided that the P/T sequence shows a bias toward pre-translocation (11). Here, we utilized a PPT-derived substrate that mimics inhibition at position ϩ3 and monitored complex formation through use of EMSA (Fig. 3). The EMSA was performed in the presence and absence of MgCl 2 to establish whether or not catalytic metal ions are required for inhibitor binding. The data show that complex formation increases in the presence of increasing concentrations of PFA or CPHM, provided that MgCl 2 was added to the reaction. Hence, binding of both CPHM and PFA depends on the presence of Mg 2ϩ , suggesting that CPHM, like PFA, also binds to the polymerase active site.  In an attempt to provide more detailed information on the binding site, we studied the effects of selected mutant enzymes on ternary complex formation monitored through EMSA (Table 1). In these studies we included mutants with changes of the three catalytic aspartic acid residues (D110N, D185N, and  D186N). We also included two mutant enzymes with amino acid changes K65R and E89K that are known to confer resistance to PFA (29 -31). Lys-65 is located in the vicinity of the phosphate moiety of an incoming nucleotide, whereas Glu-89 is found in close proximity to the template (25). Thus, both mutations contribute to the PFA-resistant phenotype through different mechanisms. K65R can potentially affect binding of PFA, whereas E89K was shown to disturb the translocational equilibrium of HIV-1 RT thereby diminishing access to the pretranslocated complex (11). Although D110N and D186N show subtle effects on complex formation, the D185N mutant shows   (33). Here, we included CPHM in these studies, and we measured RNase H activity under both multiple and single turnover conditions (Fig. 4). In the absence of inhibitor, two major cleavage fragments are produced by RNase H degradation of a 3Ј-radiolabeled template, corresponding to the post-translocated state of RT (Ϫ18) and the pre-translocated state of RT (Ϫ19) (33). Inhibition of RNase H activity with either PFA or CPHM is solely observed under multiple turnover conditions, i.e. steady-state conditions. However, no inhibitory effects are seen when CPHM or PFA are added to a preformed RT-P/T complex under single turnover conditions. The single cleavage product at position Ϫ19 is indicative for the pre-translocated state. In agreement with our previously published data on PFA, we propose that binding of either CPHM or PFA to the pre-translocated complex of RT diminishes the turnover. However, there is no evidence to suggest that these compounds can act as bona fide RNase H active site inhibitors.
To address this question in greater detail, we used a chimeric DNA-RNA/DNA primer-template substrate that allowed us to monitor primary and secondary RNase H cleavages. The latter were shown to be sensitive to inhibition by the RNase H active site inhibitor ␤-thujaplicinol (32). In this experiment, primary RNase H cleavage is seen 1 bp downstream of the DNA-RNA junction, followed by downstream secondary cleavages. The addition of CPHM in a dose-dependent manner did not affect RNase H degradation (Fig. 5), which provides strong evidence to show that this compound is indeed not a bona fide RNase H inhibitor. In contrast, secondary cleavages were efficiently inhibited in the presence of ␤-thujaplicinol, as described previously (32).
CPHM Does Not Inhibit the UL54 Polymerase of HCMV-To investigate the specificity of CPHM for HIV-1 RT, we performed DNA synthesis inhibition assays using both HIV-1 RT and the UL54 polymerase of HCMV (Fig. 6). Inhibition of DNA synthesis by PFA is evident with both polymerases; this lack of specificity is characteristic of PFA that inhibits several viral DNA polymerases. However, although CPHM inhibits DNA synthesis by HIV-1 RT, no inhibition of DNA synthesis by the HCMV UL54 enzyme is seen. This suggests that structural differences in the architecture of the polymerase active sites of these two enzymes can determine specificity for CPHM.

Structure-Activity Relationship Analysis of CPHM Derivatives Revealed Important Chemical Motifs for Inhibitor
Activity-In attempts to further characterize the specific binding site for the inhibitor, we synthesized a panel of derivatives of CPHM and tested their abilities to inhibit DNA synthesis by HIV-1 RT (Table 2 and Fig. 7). Representative CPHM derivatives were tested in the gel-based assay (compounds 3-6) (Fig.  7) to assess the importance of the dicarboxylate anchor motif and the aromatic ring of the putative specificity domain. To further probe the necessity of having a dicarboxylate moiety for inhibitory activity, we initially tested derivatives, which contained di-acetyl (compound 3) (Fig. 6), mono-ethyl ester, mono-carboxylate (compound 29) (Table 2), and di-ethyl ester (compound 30) ( Table 2) functionalities. None of these compounds displayed inhibition of DNA synthesis, confirming that the dicarboxylate moiety of CPHM is essential for inhibitor activity. In addition, we investigated the effect of replacing the chloro-group in the para-position of the phenyl ring (compounds 4 and 6 -15) ( Table 2) (compounds 4 and 6) (Fig. 6). Of the compounds tested from Table 2, only the p-bromo and p-methyl derivatives retained comparable activity to CPHM; the others were less active or inactive. Compounds 4 and 6 from Fig. 6 caused very weak pausing at position ϩ16 when compared with CPHM. Substitution at position R 1 (compound 28)  ( Table 2) abolishes inhibitor activity. These data demonstrate that the chlorophenylhydrazone moiety is a requirement for inhibitor activity and that even the removal of the chlorine from the inhibitor results in a dramatic decrease in compound activity. We subsequently derivatized the phenyl ring of CPHM with methyl groups to probe the binding site of the inhibitor. The m-methyl (compound 5) ( Table 2) (compound 5) (Fig. 7) and o-methyl (compound 24) (Table 2) derivatives of CPHM were tested for DNA synthesis inhibition. Although compound 24 showed weaker inhibitory activity than CPHM, compound 5 showed stronger activity than the parent compound, as evidenced by stronger pausing at positions ϩ3 and ϩ16 (compound 5) (Fig. 7B). This suggests that even subtle changes to the phenyl ring can influence the inhibitory activity of the compound. Finally, we examined p-cyano (compound 9) ( Table 2) and p-nitro (compound 6) ( Table 2) (compound 6) (Fig. 7) derivatives of CPHM to establish the requirement of a chlorogroup in the para-position for inhibitor activity. Both of these compounds had less activity in our assays, suggesting that although derivatization at the para-position is essential for compound activity, unfavorable interactions may arise when larger groups are introduced at this position.
In Silico Docking Revealed Differences in CPHM and PFA Binding-To gain structural insight into the binding modes of CPHM and PFA to HIV-1 RT, we constructed molecular models to illustrate stabilization of the pre-translocated complex of RT (Fig. 8). Using the crystal structure of the ternary complex composed of a chain-terminated primer-template and a bound dTTP (Protein Data Bank code 1RTD), we generated a pseudo pre-translocated complex by removing the ␤and ␥-phos- phates of the bound nucleotide. The fingers remain in the closed conformation as predicted for complexes with either PFA or CPHM. We proceeded to dock CPHM and PFA to the polymerase active site of RT using AutoDock 4.2, and the best pose from the highest ranked cluster was retained as the working model for inhibitor binding (Fig. 8). Immediately apparent is the proximity of PFA to the side chain of Lys-65. It is conceivable that upon mutation of this residue to arginine, the increased bulk of the amino acid side chain clashes with the phosphonate of PFA, resulting in diminished and/or unproductive binding of the inhibitor. CPHM, however, binds in a manner with its phenylhydrazone moiety pointing away from Lys-65. In addition, both compounds interact with the Mg 2ϩ ion distal from the primer terminus through their carboxylates. This may explain why Asp-185 is essential for binding of both CPHM and PFA.

DISCUSSION
RT translocation is an underexplored target. Like the natural dNTP substrate, NRTIs bind to the nucleotide-binding site that is accessible in the post-translocated state. The translocational equilibrium is re-established following NRTI incorporation and may show a bias toward post-or pre-translocational states, depending on the structure of the inhibitor. Recent studies have shown that the investigational NRTI 4Ј-ethynyl-2-fluoro-2Јdeoxyadenosine causes a shift toward pre-translocation (34). Although this compound contains a 3Ј-hydroxyl group, DNA synthesis is terminated because the next nucleotide substrate has no access to its designated binding site. 4Ј-Ethynyl-2fluoro-2Ј-deoxyadenosine is therefore dubbed a translocationdefective RT inhibitor (TDRTI) (35)(36)(37)(38). Nucleotide-competing RT inhibitors, such as INDOPY-1, can reversibly bind to and trap the post-translocated complex (39 -46). These compounds are likewise under pre-clinical investigation. PFA is the only known non-nucleosidic compound that was shown to trap the pre-translocated complex of HIV-1 RT. The structure of an HCMV-like polymerase in complex with DNA/DNA and PFA provided strong support for this mechanism of action (47).
Here, we demonstrate that the more complex mesoxalic acid derivative CPHM can specifically trap the pre-translocated complex of HIV-1 RT. Our data suggest that binding of the inhibitor is Mg 2ϩ -dependent. Mutations at the polymerase active site result in the loss of inhibitor binding. D185N shows the strongest effect in this regard. In addition, we have shown that the compound cannot stabilize a ternary complex when the enzyme is situated at sequence positions that show a bias toward post-translocation. The E89K mutation, which results in resistance to the pyrophosphate analog PFA by diminishing the population of pre-translocated complexes, is in our enzymatic assays also less sensitive to CPHM; chemical footprinting and RNase H mapping experiments confirm that both PFA and CPHM can trap the pre-translocated complex of HIV-1 RT.
It is also important to note that CPHM does not directly target the RNase H activity of HIV-1 RT. Rather, inhibition of RNase H-mediated RNA cleavage is exclusively seen under steady-state conditions, because binding of PFA (33) or CPHM (this study) to the polymerase active site facilitates formation of a stable ternary complex, which in turn diminishes the turnover of the reaction. Under single turnover conditions, this inhibitory effect is not observed, and the RNase H cleavage pattern demonstrates that the pre-translocated complex is trapped. Together, these data provide compelling evidence to suggest that the inhibitor binds to the polymerase active site.
In silico docking studies are in good agreement with the experimental data. An important observation is the differential effect of the K65R mutation on sensitivity to PFA and CPHM, respectively. Although sensitivity to PFA is reduced in this mutational context, CPHM shows hypersusceptibility. This difference can be rationalized if one considers different binding modes of the two compounds. In our model, binding of PFA appears to be stabilized by Lys-65 through an electrostatic interaction between the negatively charged phosphonate of PFA and the positively charged ⑀-amino group of the lysine side chain. The K65R mutation would result in a steric clash between the guanidinium of arginine and the phosphonate of PFA. Conversely, CPHM does not appear to be located in the vicinity of Lys-65. The guanidinium of K65R would be too far away from CPHM to clash with the compound, and it may even stabilize binding of CPHM to the polymerase active site.
Interestingly, mutations at active site residues Asp-110 and Asp-186 did not result in dramatic losses of binding for PFA and CPHM, although mutation of Asp-185 precluded inhibitor binding in both cases. This may be accounted for by the spatial orientation of Asp-185 with respect to the binding orientations of PFA and CPHM to the polymerase active site. The plane in which the divalent metal ions are coordinated may be essential for the proper binding of PFA and CPHM to the enzyme. The D185N mutation might cause a complete loss of metal stabilization in the active site, which would in turn result in the loss of PFA and CPHM binding.
This study opens the possibility for developing novel inhibitors of related chemical classes. Indeed, the selective inhibition of HIV-1 RT by CPHM points to structural differences in the polymerase active sites of HIV-1 RT and HCMV UL54. CPHM hypersusceptibility in the context of the K65R mutation suggests that pre-translocational inhibitors can be designed to escape resistance pathways that are associated with NRTIs. Although our SAR studies have shown a number of chemical constituents of CPHM, which are essential for inhibition of RT, we have demonstrated that derivatization of the phenyl ring may be an avenue to investigate for improved inhibitor potency. The absence of antiviral effects of CPHM could be due to the inability of the compound to permeate the cell membrane as it contains two negative charges, although this has not been directly evaluated in this study. A prodrug approach is likely required to address this problem. Further detailed SAR studies will be undertaken to dissect the chemical constituents required for inhibitor activity, as well as testing of activity-optimized compounds for toxicity and antiviral effects in cell culture.