NTPase / helicase of Flaviviridae : inhibitors and inhibition of the enzyme

Peter Borowski , Andreas Niebuhr, Herbert Schmitz, Ramachandra S. Hosmane, Maria Bretner, Maria A. Siwecka and Tadeusz Kulikowski Abteilung für Virologie, Bernhard-Nocht-Institut für Tropenmedizin, Hamburg, Germany; Department of Chemistry and Biochemistry, University of Maryland Baltimore County, Baltimore, U.S.A.; Institute of Biochemistry and Biophysics, Polish Academy of Sciences, Warszawa, Poland

The viral plus-strand RNA serves further as a template for the synthesis of several copies of minus-strand RNA.A membrane associated replicase complex consisting of at least two viral proteins carries out this synthesis: NS3 with its nucleoside-triphosphatase and helicase (NTPase/helicase) activities and NS5B with its RNA dependent RNA polymerase (RdRp) activity.The minus-strand RNA is transcribed into respective plus-strand RNA which in turn is assembled with the nucleocapsid (Westeway, 1987;Monath & Heinz, 1996;Rice, 1996;Meyers & Thiel, 1996;Leyssen et al., 2000).
Although the NS proteins are not constituents of the virus particle, their intact function, particularly of the components of the replication complex, is essential for virus replication (Bartenschlager, 1997;Ishido et al., 1998;Neddermann et al., 1999;Koch & Bartenschlager, 1999;Leyssen et al., 2000).In this context the NTPase and/or helicase activities of NS3 appear to be exceptionally attractive targets for termination of viral replication.

STRUCTURE AND FUNCTION OF NS3 NTPase/HELICASE
The first amino-acid sequence comparisons of DNA and RNA NTPase/helicases and other NTP consuming enzymes revealed a range of conserved motifs associated with NTP binding (Gorbalenya & Koonin, 1989;1993;Kadare & Haenni, 1997).These include Walker motif A, which binds the terminal phosphate groups of the NTP, and Walker motif B responsible for chelation of the Mg 2+ of the Mg-NTP complex (Walker et al., 1982).Based on the different sequences of Walker motif A the NTPase/helicases are arranged in three superfamilies (SF's): SF1 that is characterized by the classic Walker motif A (G-X-X-X-X-G-K-S/T) and SF2 and SF3 that reveal variations of the domain (A-X-X-G-X-G-K-S/T) and (G-X-G-X-G-K-S) respectively (Kadare & Haenni, 1997;Lüking et al., 1998;Fuller-Pace, 1994).In SF1 are classified alpha virus-like (nsP2-like) proteins; SF2 includes polypeptides similar to the NS3 protein (NS3-like proteins) encoded by potyviruses and the members of Flaviviridae family; SF3 includes picornavirus-like (2C-like) proteins (Gorbalenya et al., 1989;Lain et al., 1989).The SF2 NTPase/helicases are further divided, according to the sequence surrounding the conserved D-E residues (Walker motif B) in three different subgroups of proteins.The first is formed by the classic D-E-A-D box proteins, and the other are named D-E-A-H and D-E-X-H, based on their deviating Walker motif B sequences (Fuller-Pace, 1994;Kadare & Haenni, 1997;Lüking et al., 1998).The latter two subgroups are more heterogeneous than the D-E-A-D box proteins with respect to their sequence and biochemical function (Wassarman & Steitz, 1991;Lüking et al., 1998).The NTPase/helicases associated with the NS3 protein of the hepaciviruses and HGV possess D-E-C-H and those of pestiviruses the D-E-Y-H motif and are therefore members of the D-E-X-H box subgroup (Miller & Purcell, 1990;Gu et al., 2000).On the other hand the enzymes of the flaviviruses possess the D-E-A-H motif, thus belonging to the D-E-A-H box subgroup.The membership in the D-E-X-H or D-E-A-H box subgroups has no clear predictive value regarding the key properties of the enzyme.The insights into the dependency between the structure and function have come primarily from X-ray crystallography data of a representative family member, the HCV NTPase/helicase.The structure of the HCV enzyme in the absence of a nucleic acid has been solved to 2.1 Å (Yao et al., 1997) and 2.3 Å resolution (Cho et al., 1998) (Fig. 1).Further, the structure of the NTPase/helicase with a bound oligonucleotide has been solved to 2.2 Å resolution (Kim et al., 1998).The protein consists of three equally sized structural domains separated by a series of clefts.Domains 1 and 3 share with each other a more extensive interface than either of them shares with domain 2. As consequence, the clefts between domains 1 and 2 and domains 2 and 3 are the largest.Domain 2 is flexibly linked to the other two and may undergo a rigid-body movement relative to domains 1 and 3. Domains 1 and 2 contain all of the conserved helicase sequence motifs and, moreover, display structural similarities to the coresponding domains of PcrA helicase from Bacillus stearothermophilus and to Rep DNA helicase from Escherichia coli which are members of SF1 (Subramanya et al., 1996;Korolev et al., 1997).The structural comparisons made it possible to localize the mentioned Walker motifs A and B (motifs I and II of the SF2 helicases) on the surface of domain 1.In the absence of substrate residues of the Walker motifs bind to each other and additionally to the residues of the conserved T-A-T sequence (motif III).Motif III is a part of the flexible switch region, so called "hinge region", that connects the first and second domains of the enzyme (Yao et al., 1997;Kadare & Haenni, 1997).As demonstrated for the DEXX DNA helicase, the conformational changes of the molecule accompanying the NTP hydrolysis are transmitted by this switch sequence (Subramanya et al., 1996).
The role of the highly conserved arginine-rich motif (G-R-X-G-R-X-G-R; motif VI) localized on the surface of domain 2 is controversial.Yao et al. (1997) and Cho et al. (1998) proposed on the basis of their X-ray crystallographic data that motif VI is required for RNA binding and that a further cluster of positively charged amino acids may be involved in the interaction with the RNA substrate.This appears to be in accordance with the biochemical data obtained with the related HGV NTPase/helicase (Gwack et al., 1999) and with our own data obtained with isolated domain 1 of HCV NTPase/helicase (Borowski et al., 1999b).On the other hand, the structure of the HCV NTPase/helicase complexed with (dU) 8 , proposed by Kim (Kim et al., 1998), suggests rather that the arginine residues of these motifs are directly involved in the ATP binding.
Based on the structures of the enzymes and biochemical analyses, two alternative mechanisms of the unwinding reaction have been proposed.According to the first one, the so called "passive" mechanism, the NTPase/ helicase molecule binds single stranded regions of the substrate and does not actively participate in the separation of the strands forming the RNA or DNA duplexes (Matson & Kaiser-Rogers, 1990;Yao et al., 1997;  Lüking et al., 1998).Corroborating with this model is the observation that some proteins, like eukaryotic replication protein A (Georgaki et al., 1992), herpes simplex virus (HSV) type 1 ICP8 protein (Boehmer & Lehman, 1993), T4gp32 (Chase & Wiliams, 1986) and the protein vasa from Drosophila melanogaster (Hay et al., 1988) are capable of unwinding duplex DNA structures in an NTP-independent manner.Moreover, in a recent study Porter and Preugschat identified the strand separating activity of the HCV NTPase/helicase, that occurs in the absence of ATP, as being purely stoichiometric and not kinetic and confirmed its passive nature (Porter & Preugschat, 2000).Interestingly, our own investigations, performed on a group of closely related NTPase/helicases of flaviviruses demonstrated that the stoichiometry of the NTP-independent unwinding reaction (enzyme molecules : nucleotide bases of the substrate) vary significantly between the species of the genus (P.Borowski, A. Niebuhr, H. Schmitz, unpublished data).
The second, so called "active", mechanism of the unwinding, predicts an NTP-dependency of the reaction and at least two nucleic acid binding sites on the surface of the enzyme.According to this model the NTP-triggered conformational changes facilitate the binding of RNA or DNA substrate to the alternative sites (Yao et al., 1997).Thus, this modus of the unwinding requires either the presence of multiple substrate binding sites on a single polypeptide, or alternatively, as postulated for the majority of NTPase/helicases, the multiple substrate binding sites could result from the polymerization of the enzyme.Indeed, an oligomeric status of the HCV NTPase/helicase was demonstrated by Levin and Patel (1999) and such of the WNV enzyme was shown recently by us (Borowski et al., 2001a).
Although X-ray crystallography data supply some insights into the mechanism of the unwinding reaction, the basic information regarding the biochemical properties of NTPase/helicase has come from kinetic studies and mutational analyses.
The minimal amino-acid sequence of NS3 displaying helicase activity is approximately 400 residues in length and, for HCV, lies between residues 1209 and 1608 of its polyprotein (Kim et al., 1997).Nevertheless, most of the biochemical studies are performed with an enzyme consisting of the entire COOH-terminally localized NTPase/helicase domain of NS3 or with the full-length NS3 protein.However, studies comparing the biochemical properties of the full-length and truncated NS3 do not supply uniform results.Morgenstern et al. (1997) demonstrated significant differences between the proteins regarding the pH value and poly(U) concentration required for optimum ATPase activity of the HCV NTPase/helicase.Further, Kuo (Kuo et al., 1996) have shown that removal of the NH 2 -terminal 148 amino acids enhanced the NTPase activity of JEV NTPase/helicase.On the other hand, a comparative study with recombinant full-length NS3 and its isolated NH 2 -and COOH-terminal domains, performed by Gallinari et al. (1998), did not revealed significant differences in the enzymatic activities analyzed in independent in vitro assays.In this context it appears to be more appropriate to verify the inhibitory potential of NTPase/helicase inhibitors evaluated with the COOH-terminal NTPase/helicase domain, with full-length enzyme or even in the presence of other components of the replication complex.Up to date the NTPase activity could be demonstrated for a wide range of the Flaviviridae NS3 proteins.The common property of the enzymes, thus far tested, is their low selectivity towards the nucleobase of the NTP.The enzymes hydrolyzed all NTP's, dNTP's and even acyclovir triphosphate, acyclic ATP and tri-polyphosphate, which lack the ribosyl functionality of natural nucleosides (Tamura et al., 1993;Warrener & Collet, 1995;Preugschat et al., 1996) This property was documented for a wide spectrum of unrelated NTPase/helicases, like the DNA helicases from SV40 (Scheffner et al., 1989), bacteriophage T7 (Notarnicola et al., 1995) and E. coli (Moore & Lohman, 1994).The hydrolysis of the NTP's is stimulated by ribohomopolymers and 2¢-deoxyribohomopolymers.The response of the NTPase/helicases to the polymers varies among the species of the Flaviviridae enzymes.For example the ATPase activity of the HCV enzyme is stimulated with the following order of efficiency: poly(U)=poly(dU)>> poly(A)>poly(dT)=poly(C)>poly(dI)>poly(I)> poly(dA)>poly(dC)>poly(G), whereas the order of efficiency for the NTPase/helicase from YFV was the following: poly(A)>poly(C)> poly(U)>>poly(dU)>poly(I)>poly(dT)>poly(dI) >poly(dA)>poly(dC)>poly(G) (Suzich et al., 1993).As shown in numerous studies performed with full-length NS3 or the COOH-terminal NTPase/helicase domain, the extent of the activating effect corresponds closely to the rate of binding of the polynucleotides to the protein (Gwack et al., 1996;Bartenschlager 1997).Both, binding of the polynucleotide and polynucleotide mediated NTPase activation required a minimum length of the nucleic acid (15 nucleotides) (Preugschat et al., 1996).A similar minimum length was also reported for E. coli Rep helicase and eukaryotic nuclear DNA helicase (Chao & Lohman, 1991;Zhang et al., 1995).It has been demonstrated that this length of the polynucleotide may contribute to the binding energy needed for the complex formation (Preugschat et al., 1996).
The preference for the polymer that acts as activator of the NTPase reaction may be related to the presence of the homopolymeric motifs within the 3¢-UTR.Indeed, in a recent study a specific interaction of hepatitis C virus protease/helicase NS3 with the 3¢-terminal sequences corresponding to the viral 3¢-UTR was demonstrated (Banerjee & Dasgupta, 2001).The mechanistic significance of the activation mediated by the polynucleotide for virus replication remains unclear.
Recent genetic "knock out" experiments support the essentiality of the NS3 associated helicase activity for virus replication (Gu et al., 2000;Matusan et al., 2001).In the case of the members of Flaviviridae family the negative-stranded RNA must be synthesized using the parental positive-stranded RNA as a template.The resulting negative-stranded RNA is then used as a template for the synthesis of the positive-stranded progeny RNA, that is than assembled into viral particles.Since the negative and positive oriented RNA strands are complementary, the NS3 associated helicase activity appears to be necessary for strand separation.
Nevertheless, the helicase activity could be experimentally documented only in a few members of the Flaviviridae family (Fig. 2) (Suzich et al., 1993;Gwack et al., 1999;Li et al., 1999;Gu et al., 2000;Utama et al., 2000;Borowski et al., 2001a;2001b).All the enzymes exhibit a 3¢to 5¢-directionality with respect to the template strand (Kadare & Haenni, 1997) and, in contrast to the majority of the NTPase/helicases described, they are capable of unwinding DNA/DNA and RNA/RNA homoduplexes and RNA/DNA heteroduplexes (Tai et al., 1996 and own observations).The explanation for this lack of specificity is the fact that the interaction between the protein molecule and the DNA or RNA substrate is mediated by phosphate groups and not by the nucleotide base or sugar moieties (Yao et al., 1997;Kim et al., 1998).
Despite numerous kinetic analyses and solution of the structure of several NTPase/helicases it remains unclear how the NTP binding and hydrolysis are coupled to the unwinding of double-stranded substrate.The data suggest that the activities of the enzyme are not necessarily coupled: (i) The enzymes function as polynucleotide-stimulated and not as polynucleotide-dependent NTPases and the stimulation of ATPase activity of HCV or WNV NTPase/helicases by singlestranded nucleic acids is not related to the procession of the helicase along the RNA or DNA substrates (Tai et al., 1996;Hesson et al., 2000;Borowski et al., 2001a).(ii) A broad range of structurally unrelated compounds like 5-fluoro-2-selenocytosine or derivatives of O 6 -benzylguanine are able to inhibit or enhance the ATPase activity of the WNV NTPase/helicase without affecting the helicase activity (Borowski et al., 2001a).On the other hand, some chloroethylguanine derivatives stimulated the helicase activity of the enzyme with no apparent effect on the ATPase activity of the enzyme (Borowski et al., 2001a).(iii) The BVDV, HCV and WNV NTPase/helicases have different optimum conditions, such as pH-value, concentrations of salt or detergent, for the NTPase and helicase reactions (Preugschat et al., 1996;Morgenstern et al., 1997;Gallinari et al., 1998;Hesson et al., 2000;Borowski et al., 2001a).
These observations indicate that the inhibitors designed against the helicase should be evaluated by the helicase assay rather, than by testing of the NTPase activity or of RNA binding.

HELICASE ASSAY
The assay for the detection or measurement of helicase activity is based on a substrate (DNA/DNA or RNA/RNA homo-duplexes and RNA/DNA hetero-duplex) that is susceptible to unwinding.One of the strands, with the 3¢to 5¢-directionality, is referred to as template, and the other with the reverse orientation is referred to as released strain (Tai et al., 1996).The substrate is exposed to the enzyme in the presence of divalent ions (Mg 2+ , Mn 2+ ) and NTP.The single-stranded product of the reaction is separated from the substrate and quantified by a variety of methods (see below).Interestingly, the enzyme requires a short single-stranded structure of the substrate (on the template strand) to initiate the reaction (Tai et al., 1996;Gallinari et al., 1998).
For screening of the inhibitors of the helicase activity, a range of assays can be used: (i) The increasing signal assay developed by Kyono et al. (1998) using the scintillation proximity assay (SPA) system developed by Amersham.In this system, the unwinding of radiolabeled [ 3 H]DNA/RNA duplex by the NTPase/helicase is measured by hybridization of the released ss[ 3 H]DNA to a biotinylated complementary oligonucleotide.This is then bound to streptavidin-coated SPA beads, which results in a detectable scintillation signal.(ii) The assay developed by Kwong & Risano (1998), uses RNA/DNA hetero-duplex as a substrate.The radioactively labeled released strain is hybridized to a captured oligonucleotide, adsorbed to the wells coated with streptavidin (FlashPlate PLUS; NEN Life Science Products).The bound radioactivity is measured by scintillation counting.(iii) In the assay described by Hsu et al. (1998) an RNA/RNA homo-duplex is used as substrate.One of the strands (template) is coupled with biotin and immobilized on a streptavidincoated solid phase.The second (released strand) is labeled with digoxigenin.After terminating the unwinding reaction and removing the separated strand the remaining du-plex-bound digoxigenin-labeled strand is quantified with the use of tan ELISA assay employing anti-digoxigenin antibodies labeled with horseradish peroxidase.
In our studies we used as a substrate partly complementary RNA/RNA oligonucleotides and their deoxynucleotide versions of sequences reported previously (Gallinari et al., 1998).In this assay the 3¢-terminus of the released strand is labeled to a high specific activity with 32 P or 33 P. The product of the unwinding reaction is separated by Tris/borate/ EDTA/polyacrylamide gel electrophoresis and quantified by scintillation counting of the respective parts of the gel or alternatively by scanning of autoradiography films (Borowski et al., 2001a).The benefit of the method is the better quantifiability of the results due to the high reproducibility of the labeling of the released strand and the annealing procedure (Gallinari et al., 1998).By using this method we were able to test a broad range of substances for their inhibitory potential towards the helicase activity.

INHIBITION OF THE ENZYMATIC ACTIVITIES OF THE FLAVIVIRIDAE NTPase/HELICASES
The mentioned dissociation of the NTPase and helicase activities of the NTPase/helicases suggests that both activities of the en-zymes could be inhibited independently from each other.Indeed, the majority of the compounds discussed further inhibit selectively only one of the activities.

Inhibition of NTPase activity of Flaviviridae NTPase/helicases
According to the active mechanism of unwinding, NTP hydrolysis supplies the energy necessary for the strand displacing reaction.Thus, it could be expected that reduction of the accessibility of the NTP-binding site for NTP may lead to a decreased NTPase hydrolysis and, therefore, to a corresponding reduction of the unwinding rate.Consequently, compounds based on the structure of nucleoside-5¢-triphosphates seemed to be an effective tool for inhibition of the enzyme.
However, it was found that analogues of nucleoside-5¢-triphosphates, with a modified base such as ribavirin-TP (Borowski et al., 2000), IDA-TP or ITA-TP (Zhang et al., 2002), (Fig. 3.) when tested at ATP concentrations equal to the K m values determined for the ATPase reaction of each of the viral enzymes, inhibit the ATPase reaction mediated by HCV, JEV, DENV and WNV NTPase/helicases only weakly (Borowski et al., 2000;2001a;and unpublished data;Zhang et al., 2002) (Table 1).Moreover, at concentrations of ribavirin-TP, IDA-TP or ITA-TP higher than 500 mM and >10 ´Km for ATP an activation of the ATPase activity was observed.Higher concentrations of ATP in the reaction mixture caused further dramatic increase of the hydrolytic activity of the NTPase/helicases, for example in the presence of 1 mM ATP (corresponding to 100 ´Km value) and 0.5 mM ITA-TP an increase of 1000-1100% of the hydrolytic activity of WNV NTPase/helicase was measured.On the other hand, reduction of the ATP concentration to values that are below 1/10th of K m of the enzymes lead to an inhibition of the ATPase activity by ribavirin-TP, IDA-TP and ITA-TP.Successive lowering of ATP concentration down to values corresponding to 1 ´10 -5 of K m of each enzyme increases strongly the inhibitory effect of the compounds (Fig. 4).At very low ATP concentrations, all the nucleoside-5¢-triphosphate analogues act as classical competitive inhibitors of the ATPase activity of the HCV, JEV, DENV and WNV enzymes (Borowski et al., 2000;2001a;and unpublished data;Zhang et al., 2002).
Because of the mentioned lack of selectivity of the viral NTPase/helicases for the hydrolysis of NTP's, it could not be ruled out that the nucleoside-5¢-triphosphate analogues tested undergo partial hydrolysis to less active derivatives.This question was addressed by using the slowly hydrolysing adenosine-5¢-g-thiotriphosphate (ATP-g-S) and the non-hydrolysable ATP analogues 5¢-adenylimidodiphosphate (AMP-PNP) and b,g-methyleneadenosine-5¢-triphosphate (AMP-PCP).All the compounds were active towards the ATPase activity of the viral NTPase/helicases; either as an activator (at higher ATP concentrations) or as an inhibitor (at lowered ATP concentrations).The ATPase modulating effects mediated by ATP-g-S, AMP-PNP or AMP-PCP were found, however, to be considerably weaker when compared with those exerted by ribavirin-TP, IDA-TP or ITA-TP (Table 1).
According to the data, it appears that unaltered conformation of the terminal phosphate group is critical for theirs recognition by the Walker motifs of the NTPase/helicases tested.In corroboration of this hypothesis are our inhibition studies with a further non-hydrolysable ATP analogue, 5¢-O-(4-fluorosulfonylbenzoyl) adenosine (FSBA), in which the phosphate groups are substituted by the 4-fluorosulfonylbenzoyl moiety.The compound was completely devoid of inhibitory activity towards the ATPase activity of the enzymes.Noteworthy is that, in the absence of The ATPase assay was performed at indicated ATP concentrations (10 -5 ´Km , K m , 10 2 ´Km ) (Borowski et al., 2000;2001a).
The inhibition was expressed as the concentration of the compound at which the half-maximal inhibition was observed (IC 50 ).
The term ED 200 reflects the effective dose of compound at which 200% activity was measured.
ATP, FSBA binds covalently to the NTP-binding site of the NTPase/helicases and blocks it (Borowski et al., 1999b).As a consequence, significant reduction of the hydrolytic and unwinding activities of the enzymes was observed (Borowski et al., 2001a).
Surprisingly, the nucleoside-5¢-triphosphate analogues tested were not capable of significantly reducing the unwinding activity of HCV, WNV, JEV or DENV NTPase/helicase.Whereas ribavirin-TP modestly inhibited the helicase activity of the WNV enzyme (IC 50 = 120 mM (Borowski et al., 2001a)) and of the HCV enzyme (only under selected reaction conditions (IC 50 = 30 mM; Borowski et al., 2001b), it was devoid of any activity towards the JEV and DENV enzymes (unpublished data).Similarly, neither of the nucleoside-5¢-triphosphates influenced the helicase activity of the viral enzymes investigated up to concentrations as high as 500 mM.Interestingly, kinetic analyses and binding studies re-vealed that the reduction of the helicase activity mediated by ribavirin-TP did not result from the blockade of the NTP-binding site of the enzyme (Borowski et al., 2001a).
In the light of these observations the results of inhibitory studies obtained with paclitaxel, were not surprising.This compound, structurally not related to NTP, is known to interact with Walker motif A of the nucleotide binding pocket of efflux proteins, members of the ABC (ATP-binding cassette) superfamily of membrane proteins (Wu et al., 1998).Also in the case of the HCV NTPase/helicase, paclitaxel (Fig. 5A) was able to block the NTP-binding site and to inhibit the ATPase activity of the enzyme with a similar efficacy (the respective IC 50 values were 22 mM and 17 mM) in a competitive manner (Borowski et al., 1999b).Nevertheless, when tested as an inhibitor of the helicase activity of the enzyme, paclitaxel was not capable of inhibiting it, up to the concentration of 1 mM (unpublished data).It should be mentioned that paclitaxel does not influence the ATPase nor helicase activities of the WNV, JEV and DENV enzymes.A similar observation was made when we testedtrifluoperazine dihydrochloride, as inhibitor of the NTPase activity of the HCV enzyme (Fig. 5B) .The compound, a calmodulin antagonist (Ganapathi et al., 1991), is also chemically not related to nucleoside-5¢-triphosphate.It was capable of inhibiting NTP binding and the NTPase activity of HCV NTPase/helicase, but in contrast to paclitaxel by a non-competitive mechanism.The inhibition occurred with a similar efficacy: 98 mM and 105 mM, respectively, for half-maximal inhibition of ATP binding and the ATPase activity (Borowski et al., 1999b).Interestingly, trifluoperazine inhibits the helicase activity, however, at significantly higher concentrations (IC 50 = 600-700 mM).Whether the helicase activity of the HCV NTPase/helicase was inhibited as a consequence of the blockade of the NTP-binding site of the enzyme is unknown.In the case of WNV NTPase/helicase trifluoperazine inhibits poorly the  helicase activity (IC 50 = 1000-1100 mM) but does not influence its ATPase activity.

Inhibition of helicase activity of Flaviviridae NTPase/helicase
There is a range of mechanisms conceivable, by which the unwinding activity of NTPase/ helicases could be inhibited.Tai has demonstrated previously that polynucleotides increasing the NTPase activity of HCV NTPase/ helicase inhibited its unwinding activity.This inhibiting effect results from the competition of the polynucleotides with the RNA or DNA substrates for the nucleic acid binding site(s) (Tai et al., 1996).Several attempts to develop small-molecular inhibitors of the helicase activity of HCV NTPase/helicase acting at the level of the nucleic acid binding site have been described.Two series of chemically related compounds, reported hitherto only as patents by ViroPharma (Diana & Bailey, 1996;Diana et al., 1998), are composed of two benzimidazoles, or aminophenylbenzimidazoles attached to symmetrical linkers of variable lengths.These were reported to exhibit IC 50 values for inhibition of the HCV helicase in the low micromolar range, subsequently con-firmed and extended by an SAR study reported by Phoon (Phoon et al., 2001).
Our previous studies demonstrated that the lysine-rich histone H1 and the core histones H2B and H4 form stable complexes with HCV NTPase/helicase (Borowski et al., 1996).This protein-protein interaction leads to changing of the conformation of the histone molecules.As a consequence an alteration occurs of their properties as substrates for some serine/threonine protein kinases (Borowski et al., 1999a) and reduction of their DNA-binding capacity (Borowski et al., 1999c).On the other hand, the binding of the histones resulted in strong inhibition of the unwinding activity of the HCV NTPase/helicase with IC 50 in the nanomolar range.The NTPase/helicase binds histones by a short stretch of amino acids located within the "hinge region" connecting domains 1 and 2 (see above).Thus, one could speculate that the bound histone molecule might affect the mobility of domain 2 and, therefore, inhibit the "march" of the NTPase/helicase along double-stranded RNA or DNA.The currently running modelling studies should help to find out structurally similar small-molecular compounds that could mimic the action of histones.To our knowledge there are no reports addressing the influence of compounds that modulate the structure of the DNA or RNA substrate on the unwinding reaction mediated by any of the viral NTPase/helicases.This challenged us to test a range of established, commercially available DNA and RNA binding/intercalating agents as inhibitors of the NTPase and helicase activities of selected NTPase/helicases of the members of the Flaviviridae family.As demonstrated in Table 2, anthracycline antibiotics and mitoxantrone (Fig. 6) are very effective inhibitors of the helicase activity of the enzymes.Surprisingly, the closely related enzymes displayed significant differences in their response to the action of these compounds.Of particular interest are the parameters of the inhibition that is mediated by nogalamycin.This compound inhibited the helicase activity of the HCV enzyme with an IC 50 = 0.1 mM, whereas the unwinding activity of the WNV and DENV enzymes were reduced only marginally.Worthy of note is that, neither of the antibiotics examined reduced the NTPase activity of the enzymes up to milimolar concentrations.
Although the compounds in question are widely used in the clinic as immunosuppressant drugs, their cytotoxicity and weak penetration in to the cell limit their application as potential antivirals.On the other hand, the high inhibiting potential and selectivity of the compounds may make them attractive an-tiviral drugs.Nevertheless, a search for less toxic derivatives is necessary.
Recently, we have reported that the unwinding activity of WNV NTPase/helicase is significantly activated by N( 7)-chloroethylguanine and N(9)-chloroethylguanine (850% and 220% of the control at concentrations of 200 mM and 250 mM, respectively).This effect was not associated with enhanced consumption of ATP.The ATPase activity of the enzyme remained unchanged up to the concentrations of the compounds in the high milimolar range.On the other hand, the chemically related O 6 -benzyl-N(7)-chloroethylguanine activated the ATPase activity of the NTPase/helicase without affecting its unwinding activity (Borowski et al., 2001a).Similar modulating effects were observed with the further investigated HCV and JEV NTPase/helicases (unpublished data).The mechanism of the modulating effect remains unclear.The well-documented nucleotide binding studies of Porter (1998), together with our kinetic data (Borowski et al., 2000) suggest strongly the existence of a second nucleotide binding site within the NTPase/helicase of Flaviviridae.One could speculate that the second binding site that could be occupied by a nucleotide, nucleoside and even nucleotide base, probably  The helicase reaction was performed under standard conditions described previously (Borowski et al., 2001a) in the presence of increasing amounts of the compounds.The inhibition parameter was expressed as the concentration of the compound at which the half-maximal inhibition was observed (IC 50 ).
fulfils a regulatory function with respect to the NTPase and/or helicase activities of the enzyme.In agreement with the hypothesis are our ATP-binding studies performed with the isolated NTP-binding domain (domain 1) of the HCV NTPase/helicase demonstrating that the investigated chloroethylguanine derivatives do not influence ATP binding to the polypeptide.

CONCLUDING REMARKS
In view of the importance of modulation of RNA structures in diverse metabolic processes, RNA NTPase/helicases are probably of key importance in the life cycle of viruses whose genomes are composed of RNA.Thus, the compounds that reduce or modulate the activity of the enzymes could act as inhibitors of virus replication.Corroborating this hypothesis are the results of our in vitro and in vivo studies.An imidazo [4,5-d] H)-dione, that caused an inhibition of the helicase activity of WNV NTPase/helicase (IC 50 = 30 mM) displayed a similar inhibitory potency with respect to WNV replication in tissue cultures (Borowski et al., 2002;and unpublished results).Paradoxically, the activators of the helicase activity N(7)-chloroethylguanine or N(9)-chloroethylguanine, when applied in cell culture also caused a strong inhibition of WNV replication.The mechanism of the antiviral effect remains, however, to be elucidated.This observation is not without precedence.Recently, using high throughput screening, two independent groups identified the same class of aminothiazole derivatives, which inhibited the helicase activity of the herpes simplex virus (HSV) UL5/8/52 helicase/primase complex (Spector et al., 1998;Crute et al., 2002).The compounds inhibited HSV growth in cell culture and in an animal model (Crute et al., 2002).Although crystallization studies performed with HCV NTPase/helicase helped to determine the potential sites for inhibitor binding and the mode of their action, numer- ous molecular details of the mechanism of this class of enzymes, particularly how the enzyme recognizes (or not) the nucleotide base, the mechanism by which NTP hydrolysis is coupled to the unwinding reaction or which residues (motifs) are involved in RNA binding, remain unsolved.

Figure 2 .
Figure 2. Time course of the unwinding reaction mediated by HCV NTPase/helicase.The helicase assays were performed with 10 pmol of the enzyme and the unwinding reaction was terminated at the times indicated in the figure.The samples were separated on a 15% polyacrylamide gel containing 0.1% SDS for 14 h (reprinted from Borowski et al., (2001) Acta Biochim Polon.; 48: 739 with permission).