Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

The multifunctional human immunodeficiency virus (HIV) type 1 reverse transcriptase (RT) catalyzes the conversion of single-stranded viral genomic RNA into double-stranded DNA. HIV-1 RT has RNA-dependent DNA polymerase activity (RDDP) to synthesize a ()-strand DNA copy of the (+)-genomic RNA, RNase H activity to digest the RNA component of the resultant (+)RNA/()DNA duplex, and DNA-dependent DNA polymerase activity to synthesize (+)DNA using the ()DNA as template. Numerous inhibitors of HIV-1 RT have been identified, 10 of which are in current clinical use as treatment for HIV infection (see Sluis-Cremer et al., 2000a and Parniak and Sluis-Cremer, 2000 for recent reviews). However, the majority of these inhibitors are directed at the DNA polymerase activity of RT, and they all function primarily at preintegrational stages of virus replication. Unfortunately, HIV-1 readily develops resistance to these inhibitors, and therefore, the identification of other compounds that exhibit novel mechanisms of action is essential.

The gene for HIV-1 RT encodes a 66-kDa polypeptide; however, the biologically relevant form of the enzyme is considered to be a heterodimer consisting of two subunits of 66 and 51 kDa. The latter subunit is derived from the former by proteolytic processing during HIV-1 assembly and maturation. Although the precise kinetics of HIV-1 RT processing and dimerization are not well characterized, it is clear that the expression of RNA-dependent DNA polymerase and RNase H activities requires a dimeric enzyme (Restle et al., 1990, 1992; Divita et al., 1995a,b). Therefore, inhibition of RT dimerization or the disruption of RT dimer stability could provide alternative targets for HIV-1 inhibition (Divita et al., 1994; Morris et al., 1999a,b; Sluis-Cremer et al., 2000b). In regard to the latter strategy of RT inhibition, we have recently shown that 1-{spiro[4-amino-2,2-dioxo-1,2-oxathiole-5,3'-[2',5'-bis-O-(tert-butyldimethylsilyl)--D-ribofuranosyl]]}-3-ethylthymine (TSAOe³T) destabilizes the subunit interactions of both p66/p51 heterodimer and p66/p66 homodimer forms of HIV-1 RT (Sluis-Cremer et al., 2000b).

We previously identified N-(4-tert-butylbenzoyl)-2-hydroxy-1-naphthaldehyde hydrazone (BBNH; Fig. 1, compound 1) as a multitarget inhibitor of HIV-1 RT, active against both the DNA polymerase and RNase H activities of the enzyme (Borkow et al., 1997). We have proposed that BBNH exerts its multitarget inhibition by binding to two spatially separate sites in RT, one in the DNA polymerase domain and the other in the RNase H domain (Borkow et al., 1997; Arion et al., 2002). Furthermore, we have shown that treatment of HIV-1 chronically infected H9 cells with an appropriately formulated iron chelate of BBNH [Fe(III)BBNH as SP1093V] resulted in a nascent virus that was significantly attenuated in infectivity, thus suggesting a postintegrational target for this inhibitor. However, viral p24 levels and HIV-1 protein processing in the nascent virions were unaffected by Fe(III)BBNH (Parniak et al., 1998). Although this compound targets RT, its enzymatic activities are not necessary for the postintegrational stages of HIV replication. Because RT is an obligatory dimer, we considered that BBNH might act at the level of RT dimer stability.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Chemical structures of BBNH (1), BBSH (2), and DABNH (3).

Our results indicate that BBNH considerably decreases the dimeric stability of p66/p51 HIV-1 RT. To determine whether BBNH binding to both the DNA polymerase and RNase H domains in RT was an important criterion for dimer destabilization, we synthesized a variety of BBNH analogs and identified (4-t-butylbenzoyl)-2-hydroxy-1-salicylyl hydrazone (BBSH), which binds specifically to the polymerase domain, and (4,N,N-dimethylaminobenzoyl)-2-hydroxy-1-naphthyl hydrazone (DABNH), which binds only to the RNase H domain. The structures of BBNH, BBSH, and DABNH are illustrated in Fig. 1. Only BBNH and BBSH provided significant destabilization of the HIV-1 RT dimer, suggesting that the interaction with a site in the RT DNA polymerase domain is important for the dimer-destabilizing activity of the compound.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Recombinant HIV-1 RT p66/p51 heterodimer was expressed and purified essentially as described previously (Fletcher et al., 1996). [³H]dNTPs, -[³²P]ATP, and the homopolymeric template/primers (T/P) poly(rA)-oligo(dG)_12-18 and poly(rC)-oligo(dG)_12-18 were products of Amersham Biosciences (Montreal, QC, Canada). Heteropolymeric T/P was prepared using an RNA transcript derived from pHIV-PBS and a synthetic 18-mer oligonucleotide primer (Arts et al., 1994). The RNase H substrate poly([³H]rG)-poly(dC) was prepared as described previously (Starnes and Cheng, 1989). Nevirapine was obtained from Boehringer-Ingelheim (Canada) Ltd. (Laval, QC, Canada). BBNH and derivatives thereof were synthesized by standard methods (Edward et al., 1988). All other reagents were of the highest quality available and were used without further purification.

Assay of RT DNA Polymerase Activity. HIV-1 RT DNA polymerase activity was determined by a fixed time assay. Reaction mixtures (50 µl total volume) contained 50 mM Tris-HCl (pH 7.8, 37°C), 60 mM KCl, 10 mM MgCl₂, 5 µg/ml of either poly(rA)-oligo(dT)_12-18 or poly(rC)-oligo(dG)_12-18, and either 20 µM [³H]TTP or 10 µM [³H]dGTP. Reactions were initiated by the addition of 50 to 80 ng of RT (9-12 nM final concentration). Reaction mixtures were incubated at 37°C for 20 min and then quenched with 250 µl of ice-cold 10% trichloroacetic acid containing 20 mM sodium pyrophosphate. Quenched samples were left on ice for 20 min then filtered using 1.2-µm glass fiber Type C filter multiwell plates (Millipore Corp., Bedford, MA) and washed sequentially with 10% trichloroacetic acid containing 20 mM sodium pyrophosphate and with 100% ethanol. The extent of radionucleotide incorporation was determined by liquid scintillation spectrometry of the dried filters.

Assay of RT RNase H Activity. RT RNase H activity was assayed using either the homopolymeric poly([³H]rG)-poly(dC) substrate or a 5'-³²P-labeled heteropolymeric RNA template (5'-GAUCUGAGCCUGGGAGCU-3') annealed to a complementary DNA oligonucleotide (3'-CTAGACTCGGACCCTCGA-5'). Assays of RT RNase H activity using the homopolymeric substrate were carried out in reaction mixtures (50 µl total volume) that contained 50 mM Tris-HCl (pH 8.0, 37°C), 60 mM KCl, 10 mM MgCl₂, and 2 µg/ml poly([³H]rG)-poly(dC). Reactions were initiated by the addition of RT (generally, 50 ng of p66/p51 RT, to provide a 9 nM final concentration), followed by 20 min incubation at 37°C. Reactions were quenched by placing the tubes on ice followed by the addition of 100 µl of cold 7% perchloric acid. After 30 min on ice, the reaction mixtures were centrifuged at 12000g for 15 min. The supernatants (100 µl) were carefully removed, and the radioactivity was determined by liquid scintillation analysis.

PAGE Analysis of RT Processive DNA Synthesis in the Presence of Inhibitor. Assay conditions were identical to that described previously (Arion et al., 1996b). Briefly, HIV-1 RT was preincubated with ³²P-labeled heteropolymeric T/P at 37°C for 10 min. DNA synthesis was initiated by the simultaneous addition of dNTPs plus the polymerization trap heparin (0.2 mg/ml). Heparin competes with the binding of T/P for binding to free RT (Beard and Wilson, 1993). After appropriate incubation at 37°C in either the absence or presence of varying concentrations of RT inhibitors, reactions were quenched by addition of an equal volume of sequencing gel loading buffer. Samples were heated for 5 min at 100°C and then analyzed by denaturing PAGE using 16% polyacrylamide gels containing 7M urea. Products were visualized by autoradiography and quantified by densitometry.

Urea Denaturation of HIV-1 RT. HIV-1 RT (200 nM) was incubated with varying concentrations of urea (0-8M) in 50 mM Tris-HCl, pH 7.5, and 50 mM sodium sulfate for 1 h. Changes in protein structure and function were then assessed using a variety of probes. The intrinsic tryptophan fluorescence of the enzyme was monitored using a fluorescence excitation wavelength of 295 nm and fluorescence emission wavelengths of 335 (in the absence of urea) and 352 nm (in the presence of 8M urea) using a T-format fluorescence spectrometer. Data were plotted as a ratio of the fluorescence emission values at 335 and 352 nm [i.e., F₃₅₂(unfolded)/F₃₃₅(folded)]. Size-exclusion high-performance liquid chromatography (HPLC) was carried out essentially as described previously (Sluis-Cremer et al., 2000). Circular dichroism experiments were carried out using a Jasco J-750 spectrometer (Tokyo, Japan), and data are reported as mean ellipticity values at 220 nm. For RT activity analyses, an aliquot of the denatured RT was briefly assayed (4 min) to evaluate the residual RT activity. Four-minute assays were used to ensure that no other parameters (such as T/P) affected the equilibrium between heterodimer and monomeric forms. Similar strategies have been successfully employed previously (Divita et al., 1995).

Urea-Mediated Dissociation of the HIV-1 RT Heterodimer. Urea denaturation of p66/p51 HIV-1 RT, in the presence or absence of various concentrations of N-acylhydrazones, was carried out essentially as described previously (Sluis-Cremer et al., 2000b). Denaturation isotherms, obtained by following the loss of RT RDDP and RNase H activities with increasing concentrations of denaturant, can be approximated by a two-state transition between dimer and monomer (Divita et al., 1994). Denaturation curves were evaluated according to a linear extrapolation method (Pace, 1986). Briefly, an equilibrium constant, K_d, was calculated at each point in the transition region of the denaturation isotherm according to the following expression (Bowie and Sauer, 1989; Timm and Neet, 1992): K_d = P_t f_M²/(1  f_M), where P_t is the total protein concentration of RT and f_M is the fraction of monomeric protein. f_M was calculated from the urea dissociation isotherms using the algorithm: f_M = (y₁₀₀  y)/y₁₀₀), where y₁₀₀ and y correspond to RT activity in the absence (i.e., 100% active RT) and presence of varying concentrations of urea, respectively. A linear dependence of the Gibbs free energy of monomer formation (G = RT ln K_d) on the denaturation concentration is assumed (Schellman, 1978): G_D = G_D^H₂O  m[urea], where G_D^H₂O represents the difference in Gibbs free energy between the monomer and dimer transition in the absence of denaturant.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Urea Denaturation of the HIV-1 RT Heterodimer. Urea is a relatively strong denaturant, which at high concentrations favors the complete unfolding of polypeptide chains. To optimize the concentration of urea required to dissociate the RT subunits (but not promote significant unfolding of the polypeptide chains), we evaluated the changes in various structural and functional parameters of RT as a function of urea concentration (Fig. 2). In light of the structural complexity of the enzyme, a variety of probes were used to monitor changes occurring at different structural levels during the denaturation process. These included size-exclusion HPLC, tryptophan fluorescence (the 37 tryptophan residues in RT provide a "fingerprint" for tertiary structure), far-ultraviolet circular dichroism (CD; probe for protein secondary structure), and RT activity itself. The denaturation isotherms for each of these probes are illustrated in Fig. 2. The isotherms monitored by size-exclusion HPLC and RT activity are coincident, suggesting that loss of RT activity is a direct result of dissociation of the heterodimer and not a result of conformational change preceding dissociation. Analyses of the denaturation isotherms monitored by CD spectroscopy indicate that no significant changes in protein secondary are observed during the dimer-monomer transition. However, a small increase in the ratio of the tryptophan fluorescence emission at 335 and 352 nm is observed. This increase may result from the exposure of tryptophan residues that are buried in the RT dimer interface (five in the p66 subunit and two in the p51 subunit), as has been suggested previously (Divita et al., 1993). It was previously reported that the dissociation of RT subunits was accompanied by significant changes (increased fluorescence and large spectral shifts) in tryptophan fluorescence (Divita et al., 1993). However, in the present study, only small changes in fluorescence intensity and emission maximum wavelength were noted compared with that noted upon complete unfolding of the subunits (Fig. 2, inset). Our experiments suggest that the addition of a relatively low concentration of urea promotes dissociation of the subunits, and that this dissociation phenomenon can be monitored accurately by evaluating the loss of RT activity. Furthermore, it implies that no significant structural changes (or unfolding events) are associated with dissociation of the RT heterodimer subunits.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Denaturation isotherms of HIV-1 RT monitored by various structural [Trp fluorescence (), size-exclusion HPLC (), CD ()] and functional probes [RT activity ()]. RT (200 nM) was incubated in varying concentrations of urea (0-8M) for 1 h to allow for the establishment of a thermodynamic equilibrium. Trp fluorescence is reported as a ratio of the fluorescence emission value at 335 (folded protein) and 352 nm (8M urea denatured). Size-exclusion HPLC values are reported as percentage of heterodimeric RT as a function of urea concentration. CD values are reported as ellipticity values (°·cm2·dmol1). RT activity is plotted as the residual (%) DNA polymerase activity as a function of urea concentration. Inset, urea denaturation transition of HIV-1 RT monitored by Trp fluorescence (F352/F335). This inset highlights that the increase in Trp fluorescence from 0 to 3M urea is small relative to the overall transition from folded RT heterodimer to unfolded polypeptide chains.

Urea Dissociation of HIV-1 RT Heterodimer in the Absence and Presence of BBNH. Because RT activity is dependent on the quaternary structure of the enzyme, we used this parameter to evaluate the effect of BBNH binding on the intrinsic stability of the HIV-1 RT dimer. Figure 3 illustrates dissociation isotherms of heterodimeric p66/p51 RT. In the absence of inhibitor, the transition from RT dimer to monomer can be described by a sharp sigmoidal isotherm with a midpoint of denaturation ([urea]_1/2) at 2.30M urea. The Gibbs free energy of dimer dissociation in the absence of urea (G_D^H₂O) is estimated to be 10.1 kcal mol¹ for this transition, and the m-value (linear dependence of G_D on urea concentration) is 3.6 kcal mol¹ M¹. If HIV-1 RT is preincubated (20 min) with BBNH and then subjected to urea denaturation, the [urea]_1/2 value is shifted to lower urea concentrations. Values of 2.20, 1.85, and 1.76M were calculated for RT denaturation isotherms in the presence of 2, 10, and 25 µM of BBNH, respectively. Furthermore, the G_D^H₂O and m-value are decreased to 6.3 kcal mol¹ and 2.3 kcal mol¹ M¹ in the presence of 25 µM BBNH, respectively. The change in m-value in the presence of BBNH may reflect differences in the solvent exposure of RT hydrophobic residues in the dimeric and monomeric states and/or alterations in the pathway by which the subunits dissociate due to ligand-induced changes in the subunit interactions (Sluis-Cremer et al., 2000).

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Denaturation isotherms of HIV-1 RT in the absence () or presence of 2 (), 10 (), or 25 µM () BBNH. Denaturation isotherms are averaged data calculated from both RT RDDP and RT RNase H activity. Standard errors ranged from 0.3 to 3% over the range of the measurements in the experiments. Fraction monomeric RT was calculated according to equation described under Experimental Procedures. Values for GDH2O were calculated to be 10.1 ± 0.1 kcal mol1, 8.3± kcal mol1, 7.05 ± 0.24 kcal mol1, and 6.3 ± 0.33 kcal mol1 in the presence of 2, 10, and 25 µM BBNH, respectively.

Synthesis of BBNH Analogs and Characterization of Their RT-Inhibitory Properties. Previous biochemical data have suggested that BBNH binds to both the DNA polymerase and RNase H domains of HIV-1 RT (Borkow et al., 1997; Arion et al., 2002). To determine whether multisite binding is an important criterion for N-acylhydrazone-induced RT dimer destabilization, we used new BBNH analogs that were found to selectively inhibit either RT DNA polymerase or RNase H activity (Table 1), namely BBSH (Fig. 1, compound 2), which inhibits RT RDDP activity with a similar potency to BBNH but does not inhibit RT RNase H activity, and DABNH (Fig. 1, compound 3), which inhibits RT RNase H activity with a similar potency to BBNH (~4 µM) but is virtually inactive against RT RDDP activity (IC₅₀ 20 µM). Furthermore, we evaluated the ability of BBNH and BBSH to inhibit RT-containing mutations associated with resistance to nonnucleoside reverse transcriptase inhibitors (NNRTIs). NNRTI-resistant RT (including K103N, Y181C, Y188H, and E138K) remained sensitive to inhibition by BBNH and BBSH (data not shown), consistent with previously published data (Borkow et al., 1997).

View larger version (121K): [in this window] [in a new window]   Fig. 4.   Inhibition of RT RNase H activity by BBNH (lanes 1-5) and DABNH (lanes 6-10). A 5'-32P-labeled 18 nucleotide heteropolymeric RNA template (18-mer) was annealed to complementary 18 nucleotide DNA oligonucleotide. Reactions were carried out as described under Experimental Procedures using 2.5 pmol of RNA template and 10 nM HIV-1 RT. IC50 values of ~4 µM were calculated for both compounds.

Effect of BBNH Analogs on the Stability of RT Heterodimer Subunit Interactions. The curves for urea-induced dissociation of p66/p51 HIV-1 RT in the presence of varying concentrations of the N-acylhydrazones used in the present study showed that only BBNH and BBSH significantly destabilize the enzyme dimer in a concentration-dependent manner (Fig. 5; Table 2). The maximum change in G_D^H₂O (G) for p66/p51 HIV-1 RT is 3.80 kcal mol¹ and 3.72 kcal mol¹ in the presence of BBNH and BBSH, respectively (Table 2). This observed decrease in G_D^H₂O falls within the theoretical limit by which small-molecule ligand binding can contribute to the destabilization of macromolecular conformation (Freire, 1998). In contrast, the dissociation isotherms for the HIV-1 RT complex with DABNH are virtually identical to those obtained in the absence of inhibitor. It therefore seems that inhibitor binding to the DNA polymerase domain, but not to the RNase H domain, results in the observed destabilization of the RT dimer by N-acylhydrazones.

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Dependence of HIV-1 RT GDH2O on concentration of BBNH (), BBSH (), or DABNH (). The lines for the N-acylhydrazone concentration dependence are calculated as single-exponential decay curves (y = y0 + aebx). The asymptotic value of GDH2O is 6.3 kcal mol1 for BBNH [(GDH2O) = 3.55 kcal mol1], 6.68 kcal mol1 for BBSH [(GDH2O) = 3.42 kcal mol1], and 9.8 kcal mol1 for DABNH [(GDH2O) = 0.3 kcal mol1].

View this table: [in this window] [in a new window]   TABLE 2 Thermodynamic parameters calculated from HIV-1 RT dimer dissociation isotherms in the presence of different N-acylhydrazones Data were calculated as described under Experimental Procedures and are an average of at least two separate experiments, each carried out in duplicate. C1/2 is the concentration of urea at the midpoint of the denaturation isotherm. kD was calculated from the relationship GDH20 = RT ln kD.

Effect of Inhibitor Concentration on RT DNA Polymerase Activity under Processive Conditions. NNRTIs, such as nevirapine, block RT-catalyzed DNA synthesis in regions of the RNA or DNA template that are rich in secondary nucleic acid structure (Arion et al., 1996b; Götte et al., 1999). This results in RT exhibiting a mode of DNA polymerization that is more distributive in the presence of certain NNRTIs under single processive-cycle conditions (Kopp et al., 1991; Olsen et al., 1994; Arion et al., 1996b). We have previously shown significant differences in the processivity of DNA synthesis carried out by HIV-1 RT in the presence of nevirapine and TSAOe³T (Arion et al., 1996b). Because BBNH and BBSH destabilize RT subunit interactions, as does TSAOe³T (Sluis-Cremer et al., 2000b), we examined the effect of N-acylhydrazones on RT-processive DNA polymerization to ascertain whether there may be any similarities between these inhibitors and TSAOe³T.

View larger version (66K): [in this window] [in a new window]   Fig. 6.   Effect of inhibitor concentration on RT RDDP product distribution under single processive cycle conditions. A, autoradiogram of polymerization product profiles in the presence of BBNH (lanes 1-5), BBSH (lanes 6-10), and nevirapine (lanes 11-15). The concentration of inhibitor in each series is 0, 3, 6, 12, and 24 µM, respectively. The IC50 for BBNH, BBSH, and nevirapine on heteropolymeric T/P are 3, 2.5, and 1 µM, respectively. Experimental conditions are as described under Experimental Procedures. p + 173 represents the final DNA polymerization product. Arrows at sites p + 4 and p + 7 indicate regions at which increased dissociation is observed in the presence of nevirapine. B, graphical representation of the variation of selected polymerization products with increase in BBNH (), BBSH (), or nevirapine (). Percentage of RT polymerization product was determined by densitometric scanning of appropriate gel bands.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

To date, the majority of all compounds described that inhibit HIV-1 RT target the DNA polymerase activity of the enzyme (Parniak and Sluis-Cremer, 2000). BBNH is unique in that it inhibits both the DNA polymerase and RNase H activities of the enzyme (Borkow et al., 1997). We previously suggested that this multifunctional inhibition resulted from the binding of two molecules of BBNH per molecule of HIV-1 RT at two spatially separate sites. We proposed that one molecule bound close to the nonnucleoside RT inhibitor binding pocket in the polymerase domain of the enzyme, whereas the second molecule interacted with the spatially distinct RNase H active site (Borkow et al., 1997). Our recent molecular modeling and mutagenesis studies further support this hypothesis (Arion et al., 2002). BBNH binding to the RNase H domain of HIV-1 RT seems to require an aromatic stacking interaction between Tyr-501 and the naphthyl ring of BBNH. Mutation of Tyr-501 to Arg severely attenuates the ability of BBNH to inhibit the RNase H activity of HIV-1 RT. In contrast, the DNA polymerase activity of these mutant enzymes remained as sensitive to inhibition of BBNH as the wild-type enzyme.

We now show that BBNH binding to HIV-1 RT impacts on the protein-protein interactions of the heterodimeric enzyme by decreasing the Gibbs free energy of dimer destabilization by 3.8 kcal mol¹. To evaluate whether this observed effect was mediated by BBNH binding to one or both sites in RT, we synthesized more than 100 analogs of BBNH and evaluated their ability to inhibit RT RNase H and DNA polymerase activities. From these analogs, we identified two compounds that specifically inhibited either the RT DNA polymerase (BBSH) or RNase H (DABNH) activity of RT. As indicated above, we recently demonstrated that the binding of hydrazone molecules in the RT RNase H domain involves an important - stacking interaction between the phenol ring of Tyr-501 and the second ring of the naphthyl moiety of BBNH. The observation that BBSH, a BBNH analog with a reduced ring system (Fig. 1), is unable to inhibit RT RNase H is consistent with this model. At present, the structural features of DABNH that give rise to its unique inhibitory properties are not understood; however, crystallographic studies are in progress to evaluate its binding interaction with RT. Nonetheless, the identification of these BBNH analogs has unequivocally demonstrated that N-acylhydrazone compounds can be developed that target either activity, and furthermore, it has enabled us to identify that their binding to the RT DNA polymerase domain exerts the observed destabilization of heterodimeric HIV-1 RT. It should be noted that BBNH and BBNH do not dissociate the RT heterodimer subunits; instead, their binding seems to induce conformational changes that impact on the protein-protein interactions of the enzyme and consequently affect the DNA polymerase active site. We previously reported that TSAOe³T, a highly modified nucleoside based compound, inhibits RT DNA polymerase activity via an analogous mechanism (Sluis-Cremer et al., 2000b).

We previously showed that the binding of other NNRTIs, such as nevirapine and UC781, to RT do not destabilize RT subunit interactions (Sluis-Cremer et al., 2000b). Interestingly, Tachedjian et al. (2001) recently reported that several NNRTIs (including nevirapine and UC781) actually enhance the dimerization of HIV-1 RT. The mechanism by which this observed enhancement arises remains unclear. One hypothesis suggested that NNRTI binding stabilizes the RT heterodimer and thus shifts the dimerization binding equilibrium to heterodimer formation (Tachedjian et al., 2001). However, this hypothesis is not supported by the wealth of crystal structure data currently available, and evaluation of the buried surface area in the dimer interface of HIV-1 RT in the absence of inhibitors with that of the enzyme complexed with various NNRTIs does not correlate with the observed enhancement described by Tachedjian et al. (2001) (Menéndez-Arias et al., 2001). This is consistent with our observations that NNRTI binding to purified dimeric HIV-1 RT does not impact significantly on the dimeric structure of the enzyme (Sluis-Cremer et al., 2000b).

TSAOe³T also destabilizes RT subunit interactions, and our modeling studies suggested that TSAOe³T could bind to a site that is overlapping with, but distinct from, the NNRTI binding site (Sluis-Cremer et al., 2000b). X-ray diffraction studies show that BBNH is a flat planar molecule (Lanthier et al., 1997), and preliminary modeling studies suggest that it cannot be readily accommodated in the NNRTI binding pocket. We propose that N-acylhydrazones may bind to RT at a site close to, but distinct from, the NNRTI binding pocket in a manner analogous to TSAOe³T, and that this difference in binding relative to other NNRTIs may impact on RT subunit interaction stability. This different mode of binding is consistent with the differences in BBNH and TSAOe³T inhibition of RT RNA-dependent DNA synthesis under processive conditions compared with other NNRTIs, such as nevirapine. Neither TSAOe³T (Arion et al., 1996b) nor BBNH or BBSH promote dissociation and/or stalling of RT on a heteropolymeric RDDP template. The distribution of DNA synthesis products, including full-length product, do not differ significantly from that noted with uninhibited enzyme (except for the reduction in product levels). In contrast, nevirapine prevents the similar conditions (Arion et al., 1996a). On the other hand, nevirapine prevents the formation of full-length DNA product (p + 173), and there is a corresponding accumulation of smaller polymerization products in regions of the RNA template [that is, formation of full-length DNA product (p + 173) with a corresponding accumulation of smaller polymerization products in regions of the RNA template that are rich in secondary structure]. NNRTI binding alters the conformation of the "thumb domain" and displaces the orientation of the "primer grip" region in RT (Kohlstaedt et al., 1992; Ding et al., 1995a,b; Ren et al., 1995a,b). These regions are essential for the correct positioning of the primer terminus in the active site and also for RT DNA polymerization processivity and translocation (Jacques et al., 1994; Ghosh et al., 1996; Powell et al., 1997; Wohrl et al., 1997). Thus, the NNRTI-induced changes in RT structure can provide a plausible explanation for the increased enzyme dissociation and polymerization stalling during viral DNA synthesis in the presence of nevirapine under processive conditions. The exact molecular mechanism whereby N-acylhydrazones (and/or TSAOe3T) destabilize RT structure and how this destabilization process affects RT DNA polymerization remain unclear. It is of interest to note that neither TSAOe³T nor BBSH inhibit RT RNase H activity despite the inhibitor-induced weakening of the RT subunit interactions. This suggests that although the RT dimer form may be obligatory for the activity of RNase H (Restle et al., 1990, 1992), small changes in the stability of the subunit interactions of RT do not impact on this activity, even though they provide considerable alterations of the DNA synthetic capacity of the enzyme.

In summary, we have now identified two structurally diverse classes of small organic molecules, TSAOe³T (Sluis-Cremer et al., 2000b) and certain N-acylhydrazones, that act as "molecular crowbars" to destabilize the interactions of the RT dimer subunits. These molecules illustrate that small organic compounds can indeed impact on protein-protein interactions and may provide useful leads in the search for additional such compounds for use as antiviral agents.