Modulation of the nucleoside triphosphatase/RNA helicase and 5'-RNA triphosphatase activities of Dengue virus type 2 nonstructural protein 3 (NS3) by interaction with NS5, the RNA-dependent RNA polymerase.

Dengue virus type 2 (DEN2), a member of the Flaviviridae family, is a re-emerging human pathogen of global significance. DEN2 nonstructural protein 3 (NS3) has a serine protease domain (NS3-pro) and requires the hydrophilic domain of NS2B (NS2BH) for activation. NS3 is also an RNA-stimulated nucleoside triphosphatase (NTPase)/RNA helicase and a 5'-RNA triphosphatase (RTPase). In this study the first biochemical and kinetic properties of full-length NS3 (NS3FL)-associated NTPase, RTPase, and RNA helicase are presented. The NS3FL showed an enhanced RNA helicase activity compared with the NS3-pro-minus NS3, which was further enhanced by the presence of the NS2BH (NS2BH-NS3FL). An active protease catalytic triad is not required for the stimulatory effect, suggesting that the overall folding of the N-terminal protease domain contributes to this enhancement. In DEN2-infected mammalian cells, NS3 and NS5, the viral 5'-RNA methyltransferase/polymerase, exist as a complex. Therefore, the effect of NS5 on the NS3 NTPase activity was examined. The results show that NS5 stimulated the NS3 NTPase and RTPase activities. The NS5 stimulation of NS3 NTPase was dose-dependent until an equimolar ratio was reached. Moreover, the conserved motif, 184RKRK, of NS3 played a crucial role in binding to RNA substrate and modulating the NTPase/RNA helicase and RTPase activities of NS3.

more severe forms, dengue hemorrhagic fever and dengue shock syndrome. These diseases pose a significant threat to humans living in DEN-infected Aedes aegypti mosquitoes endemic in areas that inhabit two-thirds of world population (5) DEN genome is a single-stranded RNA (10,723 nt in length for DEN2 New Guinea C strain used in this study (6)) of positive polarity. The viral RNA contains a long open reading frame coding for a polyprotein precursor. The polyprotein is processed into mature structural proteins, capsid (C), precursor membrane (prM), and envelope (E) and at least seven nonstructural proteins, NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5, by cellular signal peptidase and viral serine protease in the endoplasmic reticulum (for review, see Ref. 7). The 5Ј-end of the viral RNA is modified by a type I cap structure (m 7 GpppN; 2Ј-OH moiety of N is methylated).
DEN2 NS3 is a multifunctional protein of about 69 kDa. It includes a serine catalytic triad within the N-terminal 185 amino acid residues. The protease domain is activated by the hydrophobic protein NS2B which serves as a cofactor for the protease and forms a complex in the infected cells (8 -12). NS2B has three hydrophobic regions flanking a conserved hydrophilic domain. The hydrophilic domain of NS2B (NS2BH) alone is sufficient for protease activity in processing of the 2B/3 site in the precursor expressed in transfected cells or in an in vitro protease assay (13,14). NS2B is an endoplasmic reticulum resident integral membrane protein (14).
The region C-terminal to the protease domain of NS3 has conserved domains found in the DEXH family of NTPases/RNA helicases (15). The NTPase activity of NS3, involved in the hydrolysis of the ␥2␤␣ phosphoric anhydride bond (shown by an arrow) of NTP has been reported for several flaviviruses including DEN2 (16 -20). In addition, NS3 has the 5Ј-RNA triphosphatase activity (5Ј-RTPase), capable of hydrolyzing the ␥2␤␣ phosphoric anhydride bond of triphosphorylated RNA as shown for an N-terminal-truncated fragment of West Nile virus NS3 partially purified from infected cells (21) or full-length NS3 expressed and purified from Escherichia coli (20,22). The 5Ј-RTPase is the first of the three sequential enzymatic reactions that are involved in the addition of 5Ј-cap to RNA (for review, see Ref. 23). Based on mutation analysis and competition experiments using ATP substrate and analogs, it was concluded that the RNA-stimulated NTPase activity and the 5Ј-RNA triphosphatase activity involve a common active site determinant(s) (20,22).
The ATPase activity is intrinsic to viral and cellular RNA helicases and is required for unwinding of a double-stranded RNA substrate (for reviews, see Refs. 24 and 25). RNA helicase activities of NS3 of hepatitis C virus, a member of Hepacivirus genus (26 -35), and the p80 of bovine virus diarrhea virus, a member of the Pestivirus genus (36,37), have been well characterized. However, the RNA helicases of mosquito-borne flaviviruses have not been well studied. The N-terminal-truncated DEN2 (16,20) and Japanese encephalitis virus NS3 (38) were expressed in E. coli and purified to demonstrate the RNA helicase activity. The RNA helicase activity of the full-length NS3, its RNA binding properties, the kinetic parameters or the influence of the protease cofactor, NS2BH domain, on the RNA helicase activity has not been previously studied to date. In this study we report that the full-length NS3 with or without NS2B cofactor domain, expressed in E. coli and purified, exhibits a catalytically more efficient RNA helicase activity than the Nterminal-truncated NS3 helicase domain, suggesting that the protease domain enhances RNA helicase activity. A functionally active serine catalytic triad is not required for this enhancement of RNA helicase activity as shown by mutagenesis of the protease catalytic triad residue, H51A.
The multifunctional NS3 protein exists in a complex with NS5, which itself has two enzyme activities, the 5Ј-RNA Omethyl transferase involved in 5Ј-capping and the RNA-dependent RNA polymerase required for viral RNA replication in DEN2-infected cells (39) (for review, see Ref. 40). Therefore, we examined the role of NS5 interaction on the NTPase/RNA helicase and the RTPase activity of NS3. Our results show that the NS5 stimulates the NTPase activity of NS3 in a dose-dependent manner until an equimolar stoichiometry is attained, which supports the notion that the heterodimeric NS3⅐NS5 complex is the functional unit involved in unwinding doublestranded RNA during replication. Furthermore, RNA binding analyses by gel shift assays show that NS3 binds to RNA, and a positively charged motif, RKRK, which is conserved in Flavivirus NS3, is required for this RNA binding activity. Mutation of this motif significantly reduced the NTPase/RNA helicase activity of NS3, suggesting that this motif is an important determinant in NTPase/RNA helicase function of NS3.
Purification of Recombinant Proteins-E. coli strain, Top 10FЈ (Invitrogen), or BL21 was transformed with appropriate expression plasmid, and the cells were grown at 37°C in LB medium containing ampicillin (100 g/ml) and 0.5% w/v glucose to an optical density of 0.6 at 600 nm. Cells were centrifuged at 6000 ϫ g and resuspended in glucose-free LBϩ ampicillin (100 g/ml), and isopropyl-1-thio-␤-D-galactopyranoside (Sigma) was added to a final concentration of 1 mM. Cells were grown for 5 h at 30°C, harvested by centrifugation, and stored at Ϫ80°C until use. For purification recombinant proteins, the bacterial pellets were resuspended in native lysis buffer containing 100 mM sodium phosphate buffer, pH 7.5, 300 mM NaCl, and 40% glycerol. Cell lysates were loaded onto a cobalt-based immobilized metal affinity column (Talon TM , Clontech). After incubation and then washing, the recombinant proteins were eluted with 500 mM imidazole. The fractions containing the highest protein concentration were pooled, dialyzed against a buffer containing 50 mM sodium phosphate, pH 7.5, 300 mM NaCl, and 40% glycerol. In some cases the recombinant proteins were purified by using the fast protein liquid chromatography HiTrap Chelating HP 1-ml column) (AKTAprime system from Amersham Biosciences). The proteins were eluted with 20 mM Tris-Cl, pH 7.4, 500 mM NaCl, and 500 mM imidazole. Peak fractions were pooled and loaded again to the HiTrap SP FF (1 ml) column. The fractions with the highest protein concentrations were pooled and dialyzed against 20 mM Tris-HCl, pH 7.5, 75 mM NaCl, and 40% glycerol.
RNA Substrates-For the electrophoretic mobility shift assays and the 5Ј-RTPase assays, RNA of 100 nucleotides in length containing sequences from the 5Ј-end of DEN2 RNA (which includes the 5Ј-untranslated region of 96 nucleotides) was obtained by in vitro transcription of the PCR product. The template for PCR was the pSY2 plasmid (encoding the subgenomic RNA (43)). The concentration of the RNA was measured by using a spectrophotometer, and the integrity of RNA was verified by partially denaturing PAGE (3.5% containing 7 M urea) followed by ethidium bromide staining. For RNA helicase assays two oligomer RNAs, a 30-and 15-mer (44), were synthesized (Dharmacon). The sequences of the oligomers are 5Ј-CAUCAUGCAGGACAGUCG-GAUCGCAGUCAG-3Ј and 5Ј-GUAGUACGUCCUGUC-3. The 30-mer was labeled at the 5Ј-end using T4 polynucleotide kinase at 37°C for 2 h. Then the labeled oligomer was extracted by phenol/chloroform followed by ethanol precipitation. Unincorporated [␥-32 P]ATP was further removed by passing through two Sephadex G-25 spin columns (Bio-Rad). The purified radiolabeled oligomer was mixed with the complementary strand, heated to 95°C for 10 min, and cooled slowly to room temperature (Ͼ3 h).
NTPase and RTPase Assays-The NTPase and RTPase assays were based on the original colorimetric method described several decades ago to quantify the inorganic phosphate present in serum (45,46) and more recently in a study of HCV NS3-associated NTPase (47). The method used in our study has slight modifications. Briefly, the reaction mixture (50 l) consisted of 25 mM HEPES-K ϩ , pH 7.5, 1 mM MgCl 2 , 1 mM NTP (unless the NTP concentration was varied), and 140 nM purified NS3 protein. For RTPase assays, the NTP was substituted with triphosphorylated RNA (100 nucleotides in length). The reaction mixtures were incubated for 30 -60 min (unless otherwise indicated) at 37°C. Reactions were stopped by adding 10 l of trichloroacetic acid. A freshly prepared solution of 140 l containing a 6:1 ratio of 0.42% ammonium molybdate in 1 N H 2 SO 4 and 10% ascorbic acid was added and incubated for 20 -40 min at 42°C. The "molybdenum blue" formed was measured at A 780 nm using a spectrophotometer (SpectraMax, Molecular Devices).
RNA Helicase Assay-The reaction mixture (20 l) for RNA helicase consisted of 20 mM HEPES-KOH, pH 7.0, 2 mM dithiothreitol, 1.5 mM MgCl 2 , 2.5 mM ATP, 500 nM purified protein (unless the protein amounts were varied), and 5 nM substrate. The reaction was carried out at 37°C for various times and then terminated by adding 5 l of 5ϫ RNA loading dye (100 mM EDTA and 0.7% SDS). Samples were loaded onto a 12% native polyacrylamide gel. The gel was electrophoresed at 100 V for 1.5 h, dried, and subjected to autoradiography. The ratio of released single strands versus the total of single and double strands was quantified by PhosphorImager (Molecular Dynamics).

RESULTS
Expression and purification of DEN2 NS3 FL proteins. A previous study from our laboratory indicated that the N-terminaltruncated DEN2 NS3 (NS3del.2, referred to as NS3⌬160) protein had basal NTPase that was stimulated by the addition of poly(A). NS3⌬160 had the deletion of the serine protease domain but contained all the conserved domains attributed to NTPase/RNA helicases of DEXH family members. The kinetic constants of the NTPase activity of NS3⌬160 were close to those of other flaviviral NTPases. However, the RNA helicase activity of NS3del.2 required 2.7 M concentrations of purified NS3del.2 protein to unwind less than 5% of a 29-bp RNA duplex. On the other hand, the hepatitis C virus NS3 and the bovine diarrhea virus p80 RNA helicases required 0.1-1 pmol (ϳ10 -100 nM) of enzyme for unwinding similar substrates (27,36). We also reported that the RNA-stimulated NTPase activity of NS3del.2 was abolished by mutation of the positively charged motif, 184 RKRK 3 184 QNGN, of NS3 or by deletion of an additional 20-amino acid residues from NS3⌬160 (NS3⌬180) even though the latter mutant still contained the 184 RKRK motif very close to the N terminus. In either of these two mutants the basal NTPase activity was still retained (16). From these results we concluded that the 184 RKRK motif, although not required for basal NTPase activity, played an important role in the RNA-mediated stimulation of NTPase. The mechanism for requirement of the positively charged motif in RNA-stimulated NTPase was unknown. In this study we hypothesized that the weak RNA helicase activity associated with the NS3⌬160 protein might be because of suboptimal folding of the NS3⌬160 protein and/or binding to doublestranded RNA substrate in contrast to the RNA helicases associated with the full-length NS3 of hepatitis C virus or the p80 of bovine viral diarrhea virus. We also considered the alternate possibility that the DEN2 RNA helicase may be stimulated by interaction with other viral components such as NS5, the RNAdependent RNA polymerase, with which NS3 forms a stable complex in Flavivirus-infected cells (39,48). Moreover, none of the mosquito-borne flaviviral NS3 RNA helicases have been studied in detail to date.
In this study we launched detailed biochemical and kinetic analyses of NTPase/RNA helicase activities of full-length NS3 (NS3 FL ) containing both the serine protease domain at the N terminus and the conserved NTPase/RNA helicase domains at the C terminus. Because the protease domain of NS3 interacts with the hydrophilic domain of NS2B (NS2BH), we sought to examine the effect of this interaction on NTPase and RNA helicase activities. To this end we also constructed the expression plasmid, NS2BH(QR)-NS3 FL , encoding the precursor protein in which the NS2BH is linked to the protease domain of NS3 flanking the two amino acid residues, QR, occupying the P2 and P1 positions of the protease cleavage site. The recombinant protein expressed from this plasmid underwent cis cleavage at the protease-sensitive site QR2 due to interaction of NS2BH with the NS3 protease domain to produce the non-covalent binary complex, NS2BH/NS3 FL . We also constructed the NS2BH(QR)-NS3 FL (H51A) expression plasmid containing the mutant catalytic triad residue (H51A). The protein expressed from this plasmid did not undergo cis cleavage and was in the precursor form (see Fig. 2A) as expected. To analyze the role of 184 RKRK motif in NTPase and RNA helicase activities of NS3 FL , the NS3 FL containing the mutant motif 184 QNGN was constructed (Fig. 1). The transformed E. coli TOP10FЈ cells were grown, and the proteins were expressed and purified as described under "Materials and Methods." Fig. 2 shows the Coomassie Blue-stained SDS-PAGE of the purified proteins. To study the effect of NS5 on the NTPase and RNA helicase activities of NS3, NS5 protein was expressed from the pMHA77-3 plasmid as described previously (49) (Fig. 2C). The NS2BH(QR)-NS3 FL precursor underwent partial cleavage at the QR2 site during purification from the Talon TM column as shown in Fig. 2D (lane 2). We observed that if the pH of the buffer used for dialysis of the protein was at 9.0, the cis cleavage of the QR2 site was enhanced and was essentially complete (Fig. 2D, lane 1). The identities of the purified proteins were confirmed by immunoblotting using either anti-NS2B (Fig. 2E) or anti-NS3 antibodies ( Fig. 2F; data not shown). The NS3 FL and NS3 FL (4M) proteins were expressed and purified as described under "Materials and Methods" (Fig. 2B; data not shown).
NTPase Activities of NS3 FL, NS2BH(QR)/NS3 FL, and NS2BH(QR)-NS3 FL (H51A) Proteins-In our previous study of the biochemical characterization of the NTPase activity of the NS3⌬160 protein, a coupled system that measured the oxidation of NADH at 340 nm (16,18,50) was employed to assay the NTPase activity. In this study we used an assay adapted to microtiter plate format that is based on the conversion of the P i formed by ATP hydrolysis to intense blue color in the presence of ammonium molybdate and ascorbic acid (45,46) with slight modifications as described under "Materials and Methods." Under these conditions the colorimetric reaction obeyed Beer's law for at least up to 100 M P i . P i was the preferred substrate over PP i , and the rate of the NTP hydrolysis was linear with increasing concentrations of the enzyme (data not shown). The NTPase activity of NS3 at various ATP concentrations exhibited Michaelis-Menten kinetics (Fig. 3). The kinetic parameters of the NS3 FL proteins were then determined from the Lineweaver-Burk plots (Fig. 3, inset). The results from several assays indicated that the apparent K m of NS3 FL for ATP was 191 M, the apparent V max was 6.05 Ϯ 0.4 nmol s Ϫ1 /g of protein, and the K cat value was 0.92 Ϯ 0.35 s Ϫ1 (Table I).
Poly(A) Stimulation of the NTPase Activity of NS3 FL -Our previous study indicated that the NTPase activity of the NS3⌬160 deletion mutant was stimulated by poly(A), which required not only the region between the 161-180 amino acid FIG. 1. Plasmid constructs. The construction of NS3 FL was described previously (22) . The construction of the other three expression plasmids are described under "Materials and Methods." residues but also the 184 RKRK motif (16). This conclusion was reached based on our results that the poly(A)-stimulated NT-Pase activity was abolished with the NS3⌬160(4M) mutant protein containing the 184 RKRK 3 184 QNGN mutation or with the deletion of an additional 20 amino acid residues from NS3⌬160 protein (NS3⌬180 protein), although the basal NT-Pase activity of either protein was unaffected. We sought to examine whether the basal NTPase activity of the NS3 FL protein is also stimulated by poly(A) and whether the 184 RKRK 3 184 QNGN mutation has any effect on the poly(A)-stimulated NTPase activity. We expressed and purified the NS3 FL (4M) mutant protein under conditions similar to those for NS3 FL (data not shown) and assayed its NTPase activity. Fig. 4 shows that poly(A) stimulated the basal activity of NS3 FL protein but not the basal activity of the NS3 FL (4M) mutant. However, the degree of stimulation of the NTPase activity of NS3 FL was only about 2-fold compared with 5-7-fold for the NS3⌬160. This can be explained due to a higher basal NTPase activity of NS3 FL than the NS3⌬160. Moreover, the positively charged motif 184 RKRK was also important for RNA-stimulated NTPase activity of NS3 FL because mutation of this motif abolished this stimulation (Fig. 4). This result suggested that the 184 RKRK motif is involved in RNA binding, and the resultant conformational change is required for stimulation of NTPase activity.
Characterization of the RNA Helicase of NS3 FL Protein-Next, we examined the RNA helicase activities associated with NS3 FL and NS3 FL (4M) (Fig. 5). The helicase assay utilized synthetic double-stranded RNA substrate having a 3Ј-singlestranded terminus, formed by annealing of a synthetic 32 P-labeled 30-mer RNA to a 15-mer complementary RNA according to the method described previously (44). We used a 100-fold excess of the enzyme to substrate in all the experiments shown in Fig. 5, A-D. The results shown in Fig. 5 indicated that the presence of the protease domain in the NS3 FL had an enhancing effect on the RNA helicase activity compared with that of the protease domain-deleted NS3 (NS3⌬160) protein described previously (16). The NS3⌬160 protein was required at 2.7 M for RNA unwinding of less than 5% of the substrate RNA (16). Interestingly, the mutation of the 184 RKRK motif significantly reduced the RNA helicase activity of NS3 FL (Fig. 5D). This is the first report for the function of this positively charged motif in RNA helicase activity of any flaviviral NS3 protein.
We also examined whether the presence of the protease cofactor, NS2BH, in a binary complex with NS3 FL (NS2BH(QR)/NS3 FL ), which was formed after cleavage of the QR2 site, had any effect on the RNA unwinding activity catalyzed by NS3 FL . Our results showed that the presence of NS2BH cofactor domain at the N terminus of NS3 FL protease domain in the binary complex did not have a significant en-  Table I.  hancing effect on the RNA helicase activity of NS3 FL . We surmised that in NS2BH(QR)/NS3 FL , the activated protease in the binary complex had an opposing effect on the helicase activity as seen from the time course of RNA unwinding by NS3 FL versus NS2BH(QR)/NS3 FL . Therefore, we constructed the NS2BH(QR)-NS3 FL ( 51 His 3 Ala) plasmid in which the protease domain was inactivated by mutagenesis of the catalytic triad residue, 51 His 3 Ala and purified the mutant protein as an uncleaved precursor ( Fig. 2A). This protein had reproducibly, an enhanced RNA helicase activity over the NS3 FL (Fig. 5,  A versus B; see also E).
We sought to determine whether changing the ratio of the enzyme to substrate had any effect on the rate of unwinding. The unwinding of the substrate by the NS2BH(QR)/NS3 FL protein was carried out at an [E:S] ratio of 40 and 400. As shown in Fig. 5F, increasing the enzyme to substrate ratio significantly increased the activity of the NS3 RNA helicase. These results are consistent with a recent study which revealed that a large excess of the HCV NS3 helicase was required for optimal unwinding of the substrate (35).
The NTPase Activity of NS3 FL Is Stimulated by Interaction with NS5, the RNA-dependent RNA Polymerase-In DEN2infected cells, NS3 and NS5 exist as a complex and are thought to be important components of the viral replicase involved in RNA replication (39). However, the functional consequence of this interaction and which step this complex is required for viral replication have not yet been established. In this study we asked whether the NS5 protein has any influence on the NTPase activity of NS3. The NTPase activity of NS3 FL was assayed in the absence and presence of increasing amounts of NS5. The results shown in Fig. 6 indicate that NS5 stimulated the NTPase activity of NS3 FL in a dose-dependent manner until the molar ratio of 1:1 was reached. At this point the addition of further amounts of NS5 had no effect on the NTPase activity of NS3. From these results we conclude that a 1:1 stoichiometric complex of NS3⅐NS5 complex is the functional NTPase. Next, we examined whether the basal NTPase activity of NS3 FL (4M) mutant can also be stimulated by the addition of NS5. The results shown in Fig. 6 indicate that the NTPase activity of the NS3 FL (4M) mutant lost the ability to be stimulated by NS5, suggesting that stimulation of basal NTPase activity by RNA and NS5 is intimately linked.
The 5Ј-RTPase Activity of NS3 Is Also Stimulated by Interaction with NS5-We previously reported that the NS3 FL has 5Ј-RTPase, the cleavage of the ␥-␤ phosphodiester bond of a triphosphorylated RNA (22), and the same active site was FIG. 5. RNA helicase assays of the various NS3 FL proteins. The RNA helicase assays were performed using different NS3 FL proteins as indicated (A-D). Samples were removed, and the reactions were stopped at various time points and analyzed by polyacrylamide gel electrophoresis as described under "Materials and Methods." The slower migrating and faster migrating bands represent double-and single-stranded RNA, respectively. In D, lane 3, the enzyme NS2BHQR/NS3 FL was used as a positive control. The percent of double-stranded RNA substrate unwound was estimated by PhosphorImager (Molecular Dynamics) and plotted in each case (E). , NS2BH-NS3 FL (H51A); E, NS3 FL ; •, NS2BHQR/NS3 FL ; ƒ, NS3 FL (4M). These experiments were repeated four times with each of the proteins. F, the helicase assays of the NS2BHQR/NS3 FL protein were performed at two different substrate:enzyme ratios, 1:40 (E) and 1:400 (•). The data were plotted using the Sigma plot, version 8. involved in NTPase and 5Ј-RTPase activities of NS3 (20,22). The 5Ј-RTPase is the first enzymatic step required in the 5Ј-cap addition of viral RNA. The N-terminal region of NS5 has been reported to possess the 2Ј-O-methyltransferase, an activity involved in the formation of type I cap, and the crystal structure of this domain was reported previously (51). We surmised that the 5Ј-cap addition is carried out by an enzyme complex of NS3 and NS5 and that this interaction, which enhanced the NTPase activity of NS3, might also influence its 5Ј-RTPase activity. Fig.  7 shows that the addition of submolar amount of NS5 (50 nM) to 280 nM NS3 stimulated the RTPase activity of NS3 ϳ5-fold on a triphosphorylated RNA substrate (100 nucleotides in length). The apparent K m of DEN2 NS3 for triphosphorylated RNA substrate was 29.3 M, and the V max was 0.65 nmol of substrate hydrolyzed/s. The apparent K cat for the RTPase was 0.09/s, and the catalytic efficiency, K cat /K m , was 3185 M Ϫ1 s Ϫ1 . The turnover number of the RTPase activity was 10-fold lower than that of the NTPase activity of NS3 FL , although the catalytic efficiency for the two reactions was approximately similar (Table I).
184 RKRK Motif Is a Key Determinant for RNA Binding-Our results indicated that the 184 RKRK motif is an important determinant for both RNA-stimulated NTPase (Fig. 4) and NS5mediated stimulation of the NTPase (Fig. 6) and RTPase (data not shown) activities of NS3. These results suggested that this motif of NS3 FL might be involved in binding to RNA and modulating these activities. To test this notion, we performed electrophoretic mobility shift assays using a 32 P-labeled RNA fragment produced by T7 RNA polymerase-catalyzed in vitro transcription representing the sequence from the 5Ј-end of DEN2 genome (100 nucleotides in length) and purified NS3 FL and NS3 FL (4M) proteins. The results shown in Fig. 8 indicate that the wild type NS3 FL bound to the RNA probe and retarded the mobility of NS3 protein (lane 2). However, the mutant NS3 FL (4M) protein failed to bind to the probe (lane 3), which supported the notion that this motif is required for RNA binding and to modulate the RNA-stimulated NTPase and 5Ј-RTPase activities. DISCUSSION In this study we launched a detailed biochemical and kinetic analysis of the RNA-stimulated NTPase/RNA helicase and 5Ј-RNA triphosphatase activities of full-length DEN2 NS3 protein. We established the following key findings. 1) Full-length NS3 with the serine protease domain at the N terminus and the conserved motifs found in several RNA-stimulated NTPases/ RNA helicases at the C-terminal region of the protease domain is a more active RNA helicase than the N-terminal-truncated or the glutathione S-transferase-NS3 fusion protein that was reported previously (20,22,52). 2) Moreover, the interaction of NS3 FL with NS5 stimulated the NTPase in a dose-dependent manner up to the point when a 1:1 stoichiometry was reached, and beyond this point NS5 had no effect on the NTPase activity, suggesting that the NS3⅐NS5 complex is the functional unit for the NTPase activity of NS3, which is coupled to RNA unwinding (3). The positively charged motif 184 RKRK plays an important role in RNA binding, and this motif modulates both the RNA-stimulated NTPase, RNA helicase, and the 5Ј-RTPase activities, all requiring interaction with RNA. Recently, the structure of RNA helicase of yellow fever virus has been solved, which suggests that the positively charged side of the helicase structure might play a role in single-stranded nucleic acid binding. 2 However, the RKRK motif is not proximal to the NTP binding site nor to the presumed double-stranded RNA binding region. 2 Because our data clearly show that the poly(A)-stimulated NTPase activity of NS3⌬160 (16) as well as NS3 FL (this study) is abolished by mutation of RKRK 3 QNGN, it is possible that the poly(A) binds to the interdomain cleft 2 where single-stranded nucleic acid is bound as in other helicases including the HCV enzyme (53)(54)(55), and the resultant conformational change is involved in NTPase stimulation. The notion 2 J. Smith, Purdue University, personal communication. that a conformational change is required for NTPase stimulation is supported by our previous observation (16) that NS3⌬180 protein, which had a deletion of the N-terminal 180 amino acid residues, did not exhibit RNA-stimulated NTPase activity even though this protein has the positively charged motif situated close to the N terminus. Recent biochemical and kinetic analysis of HCV NS3 RNA helicase supports that multiple molecules of NS3 bound to the single-strand region or the junction of single-stranded and double-stranded regions of the substrate which seemed to be required for optimal unwinding (35). Blunt-end duplex is not an optimal substrate for unwinding by HCV NS3 helicase (44). Therefore, we propose that the positively charged motif in Flavivirus NS3 might play a role in initial binding of the enzyme to the single-stranded region of the substrate RNA before unwinding of the duplex region. Our observation that mutation of this positively charged motif significantly reduced the RNA helicase activity is also consistent with this notion.
We examined the conservation of this positively charged motif in NS3 protein of other flaviviruses. As shown in Fig. 9, this motif is conserved across the flaviviral NS3 NTPase/RNA helicases, suggesting an important role in viral life cycle. Interestingly, the amino acid sequences in the C-terminal side of this positively charged motif are also conserved, suggesting that for a conformational change induced by single-stranded nucleic acid binding, additional amino acid residues might also be involved. However, in contrast to our results, in another study using glutathione S-transferase fusion proteins of DEN2 NS3 and mutants within and outside the helicase motifs, the authors concluded that mutation of this positively charged motif enhanced the RNA helicase activity (52). The reason for this discrepancy is unclear and might be attributed to differences in the nature of the recombinant proteins and RNA substrates used in the two studies. Our observation that mutation of this positively charged motif abolished the NS5-mediated stimulation of NTPase and 5Ј-RTPase activities of NS3 when the two proteins are present in stoichiometric amounts suggested that this motif is also involved in stabilization of NS3/NS5 interaction to form a functional complex. Further work is needed to establish the role of this motif and the conserved amino acid residues juxtaposed with this motif in NS3/NS5 interaction in the presence and absence of RNA.