Inhibition of HIV-1 Ribonuclease H by a Novel Diketo Acid, 4-[5-(Benzoylamino)thien-2-yl]-2,4-dioxobutanoic Acid*

Human immunodeficiency virus-type 1 (HIV-1) reverse transcriptase (RT) coordinates DNA polymerization and ribonuclease H (RNase H) activities using two discrete active sites embedded within a single heterodimeric polyprotein. We have identified a novel thiophene diketo acid, 4-[5-(benzoylamino)thien-2-yl]-2,4-dioxobutanoic acid, that selectively inhibits polymerase-independent RNase H cleavage (IC50 = 3.2 μm) but has no effect on DNA polymerization (IC50 > 50 μm). The activity profile of the diketo acid is shown to be distinct from previously described compounds, including the polymerase inhibitor foscarnet and the putative RNase H inhibitor 4-chlorophenylhydrazone. Both foscarnet and the hydrazone inhibit RNase H cleavage and DNA polymerization activities of RT, yet neither inhibits the RNase H activity of RT containing a mutation in the polymerase active site (D185N) or an isolated HIV-1 RNase H domain chimera containing the α-C helix from Escherichia coli RNase HI, suggesting these compounds affect RNase H indirectly. In contrast, the diketo acid inhibits the RNase H activity of the isolated RNase H domain as well as full-length RT, and inhibition is not affected by the polymerase active site mutation. In isothermal titration calorimetry studies using the isolated RNase H domain, binding of the diketo acid is independent of nucleic acid but strictly requires Mn2+implying a direct interaction between the inhibitor and the RNase H active site. These studies demonstrate that inhibition of HIV-1 RNase H may occur by either direct or indirect mechanisms, and they provide a framework for identifying novel agents such as 4-[5-(benzoylamino)thien- 2-yl]-2,4-dioxobutanoic acid that specifically targets RNase H.

HIV-1 reverse transcriptase (RT) 1 catalyzes the conversion of single-stranded RNA into double-stranded viral DNA. This heterodimeric protein is composed of two polypeptide subunits: p66 and p51. The two catalytic activities of reverse transcriptase are carried out within discrete active sites located in the p66 subunit. The N terminus of p66 catalyzes RNA-and DNA-dependent DNA polymerase activity, while the p15 domain at the C terminus of p66 catalyzes ribonuclease H (RNase H) activity. RNase H is required to cleave the RNA strand of the RNA:DNA heteroduplex intermediates in reverse transcription. This activity is essential for HIV-1 replication (1,2,3), thus RNase H is an attractive target for the development of new antiretroviral agents.
All RT inhibitors currently licensed for the treatment of HIV-1 infection inhibit the polymerase activity and bind to the polymerase domain of RT (reviewed in Ref. 4). Although there are several published examples of compounds that inhibit RNase H activity in vitro (5)(6)(7)(8)(9)(10)(11), it is still unclear whether their affect on RNase H function is mediated by direct binding to the RNase H domain. In some cases, these inhibitors have been shown to affect both the polymerase and RNase H activities of RT (8,9), which superficially suggests that these compounds may bind at both active sites. However, several studies have demonstrated that there is significant interdomain communication between the polymerase and RNase H domains of RT (e.g. Ref. 12), and potential allosteric effects on RNase H function are usually not ruled out.
In this report we have employed a series of RT polymerase and RNase H mutants to characterize the effects of several known inhibitors on each enzymatic function and to explore the behavior of a novel diketo acid inhibitor of HIV-1 RNase H: 4-[5-(benzoylamino)thien-2-yl]-2,4-dioxobutanoic acid (Compound I, Fig. 1). We demonstrate here that inhibitors that bind to the polymerase domain, such as foscarnet, can inhibit the RNase H activity in the absence of polymerization. The phenylhydrazone (Compound II) previously reported to be an RNase H inhibitor (10,11) is also shown to work by this mechanism. In contrast, 4-[5-(benzoylamino)thien-2-yl]-2,4-dioxobutanoic acid selectively binds to and inhibits the RNase H domain of HIV-1 RT. The metal-dependent behavior of this novel RNase H inhibitor is analogous to the mechanism of action proposed for HIV-1 integrase inhibitors of the same structural class (13,14). These studies demonstrate that inhibitors of RNase H function can be either direct or allosteric modulators of activity and suggest that prototype structures that have proven successful in developing integrase inhibitors with potent antiretroviral activity can be exploited to develop selective inhibitors of the RNase H activity of HIV-1 RT.
Protein Purification-HIV-1 RT mutants D185N (polymerase domain active site) and D443N (RNase H domain active site) were generated with the QuikChange site-directed mutagenesis kit (Stratagene). Full-length wild type and mutant RT proteins were expressed in E. coli BL21(DE3) cells and purified as described previously (19) with two variations: a 20-ml Hi-Prep 16/10 Heparin FF column (Amersham Biosciences, Inc.) replaced the phosphocellulose P-11 resin, and as a final step, an 8 ml poly(U)-Sepharose 4B column (Amersham Biosciences, Inc.) was added to eliminate residual E. coli RNase HI. The chimeric isolated HIV-1 RNase H domain protein containing an ␣-C helix substrate-binding loop derived from E. coli RNase HI (20, 21) (p15-EC) was overexpressed from a synthetic codon-optimized gene encoding the HIV-1 NL4 -3 sequence, and this protein was purified by methods described previously (21).
RNase H Polymerase-independent Cleavage Assays-RNase H activity was measured in 50 mM Tris-HCl, pH 7.8, 1 mM dithiothreitol, 6 mM MgCl 2 , 80 mM KCl, 0.2% polyethylene glycol 8000, 0.1 mM EGTA, and 5% dimethyl sulfoxide (Me 2 SO, Pierce Endogen, Inc.). Compounds were preincubated with 5 nM HIV-1 RT for 10 min at 37°C, and 40-l reactions were subsequently initiated with 50 nM nucleic acid substrate composed of the 32-mer 5Ј-[ 32 P]tC5U RNA template annealed to the complementary 12-mer DNA oligomer. Reactions proceeded for 20 min and were quenched with an equal volume of formamide gel-loading buffer. Substrate and product (25-mer) oligomers were resolved on a 20% denaturing polyacrylamide sequencing gel containing 8 M urea in Tris borate-EDTA buffer ( Fig. 2A), and results were quantified by PhosphorImager analysis (Amersham Biosciences, Inc.). Cleavage reactions with p15-EC were conducted in the same buffer with magnesium replaced by 2 mM MnCl 2. After the 10-min preincubation of 1 nM p15-EC with each compound of interest, reactions were initiated with 200 nM substrate and quenched after 10 min. The two product bands (25-mer and 28-mer) observed with p15-EC reflect the ability of this truncated protein to carry out multiple independent endonuclease cleavages (Fig. 2B); both bands were summed for quantification. To determine inhibitory potency (IC 50 ), compounds were titrated with 2-fold dilutions (Fig. 2C).
RNA-dependent DNA Polymerase Assays-The heterodimeric nucleic acid substrate used in the HIV-1 RT polymerase reactions was generated by annealing pD500, the DNA primer, to t500, the 500 nucleotide RNA template created by in vitro transcription (22). This primer-template substrate (100 nM final concentration) was combined with HIV-1 RT enzyme (5 nM) in assay buffer (50 mM Tris-HCl, pH 7.8, 1 mM dithiothreitol, 6 mM MgCl 2 , 80 mM KCl, 0.2% polyethylene glycol 8000) and immediately mixed with inhibitor or Me 2 SO (5%) prior to addition of dNTPs (5 M dNTPs, 5 M [␣-33 P]dATP) to initiate the reactions (50 l total volume). After 1-h incubation at 37°C, reactions were quenched by EDTA, spotted onto DE81 filter plates (Whatman), and washed (twice with 250 l of phosphate-buffered saline, once with 250 l of water, and once with 50 l of ethanol). After drying for 15 min, 50 l of scintillation fluid was added prior to quantification via Topcount (Packard).
Isothermal Titration Calorimetry-Isothermal titration calorimetry experiments were conducted in a MicroCal VP-ITC calorimeter. Protein samples were dialyzed in Buffer A (50 mM Tris-HCl, pH 7.0, 5% glycerol). The p15-EC protein was suspended to 18.6 M in Buffer A, and 200 M MnCl 2 was added when experiments were conducted with divalent cation. Ligand solutions were made in the same buffer as the protein. After p15-EC was added to the 1.4 ml of calorimeter sample cell and equilibrated at 30°C with a stirring speed of 490 rpm, a total of 25 aliquots (5 l) were injected into the cell. The heat released with each injection was measured, and heats of reaction were determined by integration of peak areas. Base-line correction, peak integration, and binding parameters (stoichiometry, K d and ⌬H°) were performed using the ORIGIN analysis software. Gibbs free energy (⌬G°ϭ ϪRT ln(1/K d )) and entropy of binding (T⌬S°ϭ ⌬H°Ϫ ⌬G°) were calculated from measured experimental values. Samples were run in triplicate and results were averaged to obtain binding parameters.

Inhibition of Polymerase-independent RNase H Cleavage-
There are striking structural and functional homologies between HIV-1 integrase and the RNase H domain of HIV-1 RT. Potent and specific compounds containing a diketo acid motif have been shown to inhibit HIV-1 integrase in vitro and in cell culture (13). As the purported mechanism of these inhibitors invokes sequestration of the active site divalent metals in integrase, it has been hypothesized that a similar mechanism could be exploited for other phosphotransferase enzymes. In screening a library of diketo acids we identified 4-[5-(benzoylamino)thien-2-yl]-2,4-dioxobutanoic acid (Compound I) as a novel inhibitor of the RNase H activity of HIV-1 RT.
In RNase H cleavage assays, the activity of Compound I against full-length wild type HIV-1 RT (IC 50 ϭ 3.2 M, Table I) was as potent as the most active RNase H inhibitors reported previously (8,23). Inhibition of RNase H by Compound I was independent of the order of addition to the assay; reactions initiated by either enzyme or substrate behaved identically. To further characterize this activity, Compound I was also evaluated in RNase H cleavage assays using an RT polymerase active site mutant (D185N) and a chimeric isolated HIV-1 RNase H domain, p15-EC. As expected for a bona fide RNase H inhibitor, the activity of the diketo acid against p15-EC (IC 50 ϭ 4.7 M) was comparable with wild type HIV-1 RT (Table I, Table I).
The HIV-1 RT inhibitor foscarnet and a previously described  RNase H activity. In particular, Compound III (data not shown) and nevirapine (12) increase RNase H cleavage rates and alter the position and number of cleavage products observed in vitro.
Characterization of Allosteric Inhibitors with HIV-1 RT Polymerase-To substantiate that the allosteric effect of foscarnet and Compound II on RNase H function is as a result of binding to the polymerase domain of RT, the compounds were evaluated in HIV-1 RT polymerization assays using wild type HIV-1 RT and an active site RNase H mutant (D443N). Wild type HIV-1 RT polymerase activity was inhibited by foscarnet and Compound II as well as AZT-TP and the NNRTI Compound III, but not by Compound I (Table II). The potencies listed are consistent with previously reported values for foscarnet, AZT-TP, and Compound III (17,24). In contrast to the effect of D185N mutation on RNase H inhibition, polymerase inhibition   (10), the reaction conditions were not optimal for evaluation; DNA synthesis is not rate-limiting when enzyme is equivalent or in excess of substrate, thus explaining the apparent discordance from results presented here.
Unlike foscarnet and the hydrazone, the diketo acid did not inhibit RNA-dependent DNA polymerization of wild type or D443N RT (IC 50 Ͼ 50 M). Although all three compounds had activity in RNase H cleavage assays using wild type RT, these results suggest that only Compound I is a selective inhibitor of HIV-1 RNase H catalysis.
Metal-dependent Interactions Observed by Isothermal Titration Calorimetry-Isothermal titration calorimetry experiments were conducted with the isolated RNase H domain hybrid protein (p15-EC) to characterize the interaction between the diketo acid (Compound I) and the RNase H domain of HIV-1 RT. Although no binding was observed either in the absence of metal or in the presence of Mg 2ϩ (Fig. 3B), a robust signal corresponding to the heat released was detected in the presence of MnCl 2 (Fig. 3, A and B). The divalent cation specificity observed for Compound I binding to p15-EC is consistent with the strict manganese requirement for p15-EC catalytic activity and supports the hypothesis that Compound I might interact with metal ions in the RNase H active site. As expected, no change in the heat of reaction was detected when either foscarnet or Compound II were titrated into buffer containing p15-EC (data not shown).
For Compound I, the fitted parameters (Fig. 3B, insert) derived from the ITC studies suggest that there is one binding site (n ϭ 0.9 Ϯ 0.2) on the RNase H domain for the diketo acid with an exothermic reaction upon association as shown by the negative Gibbs free energy and enthalpy changes. In addition, the dissociation constant calculated from the ITC analysis (8.9 Ϯ 3.5 M) is consistent with the IC 50 determined in enzymatic activity assays (3.2 M, Table I); the slight difference between the values may be due to a shift in solvent proton concentration (pH 7.0 versus pH 7.8, respectively). As the ITC studies were performed in the absence of nucleic acid substrate, the observation that binding is metal-dependent and that the calculated K d is comparable with the observed IC 50 suggests nucleic acid is not required for either binding of the inhibitor or the active site metals.
In summary, we have demonstrated that compounds from the diketo acid structural class can inhibit HIV-1 RNase H catalysis in vitro. Compound I is distinct from previously described RNase H inhibitors, and it selectively inhibits RNase H cleavage without affecting RNA-dependent DNA polymerization. Isothermal titration calorimetry studies confirm a direct interaction with the isolated RNase H domain and demonstrate a requirement for divalent cation. The suggestion that the thiophene diketo acid may interact with metals in the RNase H active site is consistent with the metal sequestration mecha-nism by which similar compounds affect HIV-1 integrase activity. Compound I also inhibits HIV-1 integrase in vitro (1.9 M) 2 ; however, the activity relationship for RNase H inhibition is clearly distinct from integrase, as previously described integrase inhibitors with the diketo acid pharmacophore have no effect on RNase H activity (Refs. 13 and 14 and data not shown). Although Compound I does not inhibit HIV-1 replication in cell culture, it provides an important proof of concept for direct inhibitors of the RNase H activity of HIV-1 RT. This work also outlines methods to distinguish indirect affects on RNase H function by compounds that bind to the polymerase domain and to identify selective RNase H inhibitors that may ultimately lead to the development of novel chemotherapeutics for HIV-1 infection.