The quinoline U-78036 is a potent inhibitor of HIV-1 reverse transcriptase.

The quinoline U-78036 represents a new class of non-nucleoside human immunodeficiency virus (HIV)-1 reverse transcriptase inhibitors. The agent possesses excellent antiviral activity at nontoxic doses in HIV-1-infected lymphocytes grown in tissue culture. Enzymatic kinetic studies of the HIV-1 reverse transcriptase (RT)-catalyzed RNA-directed DNA polymerase function were carried out in order to determine whether the inhibitor interacts with the template-primer or deoxyribonucleotide triphosphate (dNTP) binding sites of the polymerase. The data were analyzed using steady-state or Briggs-Haldane kinetics assuming that the template-primer binds to the enzyme first followed by the dNTP and that the polymerase functions processively. The calculated rate constants are in agreement with this model. The results show that the inhibitor acts as a mixed to noncompetitive inhibitor with respect to both the template-primer and the dNTP binding sites of the enzyme. Hence, U-78036 inhibits the RNA-directed DNA polymerase activity of RT by interacting with a site distinct from the template-primer and dNTP binding sites. Moreover, the potency of U-78036 is dependent on the base composition of the template-primer. The equilibrium constants for various enzyme-substrate-inhibitor complexes were at least seven times lower for the poly(rC).(dG)10-catalyzed system than the one catalyzed by poly(rA).(dT)10. In addition, the inhibitor does not impair the DNA-dependent DNA polymerase activity and the RNase H function of HIV-1 RT nor does it inhibit the RNA-directed DNA polymerase activity of the HIV-2, avian myoblastoma virus, and murine leukemia virus RT enzymes.


MATERIALS AND METHODS
The expression of HIV-1 RT and its purification have been described (24). For the polymerase assays, a partially purified RT preparation was used that was judged as 90-95% pure based on SDS polyacrylamide gel electrophoresis. This preparation was devoid of Escherichia coli RNase H activity and consisted of p51/p66 heterodimers of RT, with no evidence of monomeric RT in the form of p66 or p51 alone. For the RNase H assay, a highly purified preparation of the heterodimer p51/p66 RT was used. Its expression and purification have been described (25).
The standard reaction mixtures for the RNA-directed DNA polymerase assay contained 20 mM dithiothreitol, 60 mM NaCl, 0.05% Nonidet P-40 (Sigma), 10 mM MgC12, 50 mM Tris.HC1, pH 8.3,8 p~ of the cognate a-36S-labeled deoxyribonucleotide-5'-triphosphate (final specific activity 1 Ci/mmol), 10 pg/ml RNA template (poly(rA) or poly(rC)), 5 pg/ml of the appropriate primer (dT)lo or (dG),,, and 0.0274 p~ purified HIV-1 RT. The total volume of the reaction mixtures was 50 pl. The samples were incubated at 37 "C for 15 min. The reactions were terminated by the addition of equal volumes of 10% trichloroacetic acid. Incorporation of radiolabeled precursor was determined by collecting the precipitates on glass fiber filters, drying, and counting the samples.
The DNA-directed DNA polymerase activity of the RT enzyme was assessed as described above for the RNA-directed DNA polymerase assay. The synthetic template-primer used was poly(dC) . (dG)12-18 present at a concentration of 10 pg/ml.
The RNase H assay was conducted as described (26). In general, the assay follows the loss of trichloroacetic acid-precipitable radiolabeled RNA-DNA hybrid as a function of time. The specific assay mixtures contained 2.5 pg and 2 pCi/ml [3H]poly(rG) .poly(dC) (l:l), 50 mM Tris.HC1, pH 8.5, 5 mM MgCl*,O.O2% Nonidet P-40, and 3% glycerol. Incubation was for 10 min at 25 "C, and the reactions were terminated by the addition of equal volumes of 10% trichloroacetic acid. The loss of substrate was determined by collecting the precipitates on glass filters, drying, and counting the samples.
The infectivity assays in HIV-infected lymphocytes grown in culture were carried out by the syncytia reduction method (27) or by measuring the total amount of core p24 protein released into the culture medium and the total amount of viral RNA synthesized (6).
DNA polymerases a and 6 were purified from fetal calf thymus and assayed as described (28).
The avian myoblastoma virus and murine leukemia virus RT preparations were purchased from Life Technologies, Inc. These enzymes were assayed in the same standard reaction mixture as described above for the HIV RT. Sufficient amounts of the latter RT species were added per reaction mixture to incorporate approximately 0.04 nmol of dNTP in 15 min at 37 "C.

RESULTS
Antiviral Actiuity-The quinoline U-78036 is a potent inhibitor of HIV-1 RT and has shown excellent antiviral activity at nontoxic doses in HIV-1-infected lymphocytes grown in tissue culture. In the syncytia reduction assay, using MT-2 cells infected with HIV-1 (IIIb isolate), the IC, was <3 PM U-78036 and the ICso in terms of cytotoxicity to the host cells was >10 PM inhibitor (Table I). In the p24 release and total viral RNA synthesis assays carried out in HIV-1 (D34 isolate)infected peripheral blood mononuclear cells, the IC60 values in terms of p24 core protein released and total viral RNA synthesized were <0.1 PM U-78036 at day 3 and <1 PM at day 4 post-infection of the cells (Table 11). No apparent toxic effects on the host cells were observed at a concentration of 10 PM inhibitor. The IC, values for inhibition of the cellular

Antiviral activity of U-78036 in the syncytia reduction assay (in HIV-1 -infected MT-2 cells)
The cytotoxicity in MT-2 cells was >10 p~ U-78036 in 10 separate assays.

Sample
Number of syncytia/plate   SI, template-primer; SZ, dNTP. KO, K,, and Kz represent equilibrium constants between the inhibitor ( I ) , the enzyme, and its substrates. El', enzyme-product complex. polymerases a and 6 were 1300 and >3400 PM, respectively. These concentrations are significantly above those required to inhibit HIV-1 RT to the same extent as shown below.

Antiviral activity of U-78036 in HIV-1-infected peripheral blood mononuclear cells
Enzyme Kinetics-The kinetic data were first analyzed using Michaelis-Menten kinetics, which are based on a rapid equilibrium system wherein the enzyme, the substrate, and the enzyme-substrate complex are at equilibrium. This analysis yielded ambiguous results when either the template-primer or its cognate nucleotide were varied in the presence of a fixed amount of the other substrate. The rapid equilibrium treatment was thus abandoned, and the data were analyzed using steady-state Briggs-Haldane kinetics. In this latter case, the enzyme-substrate complex does not need to be in equilibrium with the enzyme and its substrate. However, shortly after initiation of the reaction, enzyme-substrate complex is formed at the same rate as it dissociates. A steady-state scheme, which includes all the reaction steps to be considered here, yields very complex velocity equations that are impractical to solve (see below). For these reasons, the general steady-state kinetic analysis used in this study was simplified as detailed in Fig. 2. The essential rate constants used in the following are defined in that figure. The steady-state kinetics were limited to the reactions occurring between the enzyme and the substrates, and rapid equilibrium kinetics were applied to the interactions between the inhibitor and the enzyme or the various enzyme-substrate complexes. Moreover, an ordered mechanism was assumed, whereby the template-primer complex binds first to the enzyme, followed by the addition of dNTP (29,30). The polymerase is a processive enzyme, and, after the addition of the first nucleotide, translocation occurs along the template, resulting in the incorporation of further nucleotides into the growing chain (29). Under these condi-tions, the formation of the phosphoester bond can be considered as irreversible, since the reverse reaction occurs at an extremely slow rate and the dissociation of the enzymeproduct complex into its components is also negligible during the initial reaction phase. Thus, the enzyme-product complex does not differ from the initial enzyme-template-primer complex in that the former shuttles back to the enzyme-templateprimer state where another nucleotide is added and this rate constant, designated as kw, is equal to bat representing the turnover number. The constant k+ represents the backward rate constant for the enzyme-template-primer-dNTP complex. The quaternary enzyme-inhibitor-template-primer-dNTP complex should be nonproductive, as no translocation to the enzyme-inhibitor-template-primer state should occur Hence, the HIV RT-catalyzed system considered here consists of two substrates, S1, S2, and one inhibitor, I. Therefore, the system contains eight enzyme species, i.e. E, ES1, ES2, ES1S2, EI, EIS1, EIS2, and EIS1S2. If conversion between any two of these enzyme species is possible, then the directed graph G representing the enzyme-catalyzed system is shown in Fig. 3 (34,35). This system is very complex, and its kinetic equations involve an excessive number of rate constants. To illustrate this, one can use Chou's graphic rule 1 of enzyme kinetics (34, 35) to estimate how many terms need to be considered in deriving the concentration for each of the enzyme species. The conversion between any two enzyme species can be expressed by a "zero-one" matrix A in which the element at the ith row and jth column is one if the enzyme species i can be converted to the enzyme species j ; otherwise, the element is zero. Such a zero-one matrix for the enzymecatalyzed system shown in Fig. 3 is given by Equation 1.
(thus, ku' + 0). where B1,l denotes the submatrix obtained by removing the first row and first column from the matrix B. Since G of Fig.   3 is a symmetric graph, i.e. a graph in which whenever there is an arc from the enzyme species i to j , there must be an arc from enzymes species j to i. Then the corresponding number of terms for each of the other seven enzyme species must be the same and also equal to 304 (34,35). The number of terms for the denominator is even larger and equal to 304 x 8 = 2432 rendering the system impractical to manageable solutions.
Since the system considered here is ordered in that the binding of S1 to E precedes the binding of S2, the system can be simplified significantly. Furthermore, the reactions between the inhibitor and the enzyme and various enzymesubstrate complexes are assumed to be diffusion-controlled (32,33), and the interconversion rates between E and EI, ES, and EIS,, ESISP and EIS1S2, respectively, occur much faster than those between the enzyme and its substrates. Thus, although the whole system is a steady-state one, there is an equilibrium between the low molecular weight inhibitor and the enzyme and the enzyme-substrate complexes (31). The whole system can be expressed as shown in Fig. 2 For such a simplified system, the rate of product formation is given by the velocity Equation 5 (see Ref.

22)
where RNA-directed DNA Polymerase-Enzymatic kinetic studies were performed with U-78036 and synthetic templateprimers to determine the type of inhibition pattern on the RNA-directed DNA polymerase function of HIV-1 RT with respect to the dNTP and template-primer binding sites.
In one set of experiments using the template-primer combination poly(rA) . (dT),, and dTTP as the two polymerase substrates, alternatively, one of the substrates was varied while the other one was kept constant. Three concentrations of inhibitor (25,50,and 100 PM) were studied in addition to appropriate controls containing no drug. The data were analyzed via computer using the steady-state kinetic model described above. The program simultaneously calculates the essential forward and backward rate constants as well as the inhibition constants of the reaction. The experimental results are listed in Fig. 4, and the calculated reaction rates are shown in Fig. 5. The rate constants derived by fitting the experimental data to Equation 5 for the association and dissociation constants kl and k l , respectively, representing the association and dissociation rate constants for the enzyme-poly(rA). The equilibrium constants between the inhibitor and the enzyme and the enzyme-substrate complexes were 4.6 pM for KO, 6.8 p~ for K,, and 8.1 pM U-78036 for Kz, respectively.
These constants are similar in magnitude and indicate that the inhibitor acts noncompetitively with respect to the poly(rC) (dG)lo and dGTP binding sites.
DNA-directed DNA Polymerase-U-78036 was also tested for its effect on the DNA-directed DNA polymerase function of RT using poly(dC) .oligo(dG) as the template-primer. No inhibition of this enzyme activity was observed with U-78036 concentrations up to 100 pM.
RNase H Assay-U-78036 did not inhibit this RT activity when tested at concentrations up to 100 p~.
Other Retroviral RT Species-The compound was also tested for its inhibitory activity against the RNA-directed DNA polymerase functions of HIV-2 RT, avian myoblastoma virus RT, and murine leukemia virus RT. None of these enzyme species was inhibited by U-78036 when tested at concentrations up to 100 pM.

DISCUSSION
The inhibition kinetics of U-78036 on the RNA-directed DNA polymerase domain of HIV-1 RT were investigated with respect to the template-primer and dNTP binding sites. The results were analyzed using a modified steady-state kinetic model. This model considers the following. (a) The reaction is ordered in that the template-primer binds to the enzyme first, followed by the binding of the dNTP species. (b) During the initial stages of the reaction, the enzyme-product complex does not dissociate into its components after the formation of a phosphoester bond and translocates back to the enzymetemplate-primer stage, and the polymerase is, therefore, processive. (c) The formation of the phosphoester bonds and the concomitant release of pyrophosphate are assumed to be irreversible, since the reverse reaction rate is negligible. (d) The binding of the low molecular inhibitor U-78036 to the free enzyme or to the various enzyme-substrate complexes follows rapid equilibrium kinetics. The results obtained are in agreement with this model. The reverse reaction rates ( k l ) for the dissociation of the enzyme-template-primer complexes in the absence of the inhibitor were approximately 0.3 s-l for both homopolymers studied. These values are very small and indicative for a processive polymerase. The translocation rate k@ was 1 s-' for the poly(rA). (dT)lo directed system and 0.6 s-' for the poly(rC). (dGho directed system in the absence of U-78036. The rate of dTMP incorporation is thus somewhat more efficient than the one for dGMP incorporation into their respective homopolymers. The pattern of inhibition exerted by U-78036 was mixed to noncompetitive with respect to the nucleic acid and dNTP binding sites of the RT enzyme and indicates that the inhibitor interacts with a site distinct from these two domains. The potency of the inhibitor depends on the base composition of the template-primers studied (poly(rA). (dT)lo uersus poly(rC). (dG)lo). While the KO values were equal in both systems, the Kl and K2 values were much higher for the poly(rA). (dT)lo-catalyzed system than for the poly(rC) . (dG)lo-catalyzed one. This indicates that the inhibitor binds much tighter to the free enzyme and the enzymepoly(rC) . (dG)lo containing complexes than to the ones containing poly(rA). (dT)lo.
Recently, other non-nucleoside classes of HIV-1 RT inhibitors have been described. These include the tetrahydroimidazo-[4,5,1-jk][1,4]-benzo-diazepin-2(1H)-one and -thione or TIBO compounds (1, 2), the dipyridodiazepinones (3,4), the l-[(2-hydroxyethoxy)-methyl]-6-(phenylthio)thymine or HEPT derivatives (5,6), the pyridinone derivatives (19), and the bisheteroarylpiperazines or BHAP compounds (20)(21)(22). The TIBO compound R82150 appears to be a specific inhibitor of HIV-1 RT-catalyzed RNA-directed DNA synthesis (3). Kinetic studies suggest that the inhibitor acts uncompetitively with respect to the nucleic acid binding site and noncompetitively with respect to the dNTP site. The IC, for DNAdirected DNA synthesis was 40 times higher than the one required for effective inhibition of RNA-directed DNA synthesis. Moreover, R82150 did not inhibit RNase H. The dipyrido-diazepinone nevirapine acts as a mixed inhibitor with respect to the poly(rA) .(dT)lo and poly(rC). (dG)lO binding sites and noncompetitively with respect to the dNTP binding sites during RNA-directed DNA synthesis by HIV-1 RT (36). Nevirapine also inhibits the DNA-catalyzed DNA polymerase function of HIV-1 RT. The pyridinone derivatives, like the TIBO compounds mentioned above, also seem to act as uncompetitive inhibitors with respect to the nucleic binding site and as noncompetitive inhibitors with respect to the dNTP binding site of the enzyme if the RT functions in the RNA-directed DNA mode (19). Moreover, the pyridinones inhibit the DNA-directed DNA polymerase of HIV-1 RT and showed a noncompetitive inhibition pattern with respect to the nucleic acid and dNTP binding sites in this case. The bisheteroarylpiperazine U-87201E is a noncompetitive inhibitor of both the nucleic acid and dNTP binding sites (22). It should be noted that all of these assessments, except for the arylpiperazine U-87201E, are based on Michaelis-Menten kinetics and not steady-state kinetics. Compared with these other classes of non-nucleoside RT inhibitors mentioned, the quinoline U-78036 appears to possess a somewhat unique inhibition pattern in that it specifically inhibits the RNAdirected DNA polymerase of HIV-1 RT without affecting the DNA-directed DNA polymerase or RNase H activities.