A Kinetic Analysis of the Oligonucleotide-modulated ATPase Activity of the Helicase Domain of the NS3 Protein from Hepatitis C Virus

Hepatitis C virus (HCV) helicase (E) formed spectrofluorometrically detectable complexes with a 16-mer and HF16 (a 16-mer with 5′-hexachlorofluoresceinyl moiety). The interaction of helicase with these effectors was investigated by kinetic techniques to determine if the complexes were kinetically competent for ATP hydrolysis. k cat values with the 16-mer and HF16 were 2.7 and 36 s−1, respectively. The maximal value of the rate constant for the approach of an intermediate to the steady-state level has to be at least 4-fold greater thank cat for it to be kinetically competent. This value was 1.2 s−1 with HF16 and “E·ATP” and was 1.82 s−1 with ATP and E·HF16. These values were too small for formation of these intermediates to be kinetically competent in ATP hydrolysis. Dissociation of “E·HF16·ATP” (0.34 s−1) was also too slow to contribute significantly to catalysis. Furthermore, theK m of E·HF16 for ATP (3 μm) was significantly less than theK m for ATP hydrolysis at a saturating concentration of HF16 (320 μm). HCV helicase has two nucleotide-binding sites per monomer. If the fluorescence changes observed were associated with structure changes preceding steady-state catalysis (isomerization), pre-steady-state data could be reconciled with the turnover data. Data for the 16-mer yielded similar conclusions.

Even though helicase activity was identified and the associated protein was purified over 20 years ago (1), the kinetic and chemical mechanisms for this class of enzymes have not been elucidated. In particular, the mechanism of coupling ATP hydrolysis to the unwinding of double-stranded DNA is incompletely understood. From extensive studies on the mechanism of action of Escherichia coli Rep helicase, Lohman and coworkers (2,(3)(4)(5)(6)(7)(8)(9)(10)(11)(12) have proposed that the catalytically active species of this enzyme is a DNA stabilized dimer. The two DNA-binding sites in the catalytic dimer unwind defined lengths of the DNA duplex by alternatively binding duplex and single-stranded DNA in a process coupled to ATP hydrolysis (2,5,6,12). Recently, it has been shown that E. coli helicase II unwinds approximately 4 base pairs during each catalytic cycle (13). Nucleotide binding also regulates the binding of singlestranded and double-stranded DNA to E. coli Rep helicase. In particular, ATP stimulates the rate of single-stranded DNA exchange by increasing the rate constant for dissociation of DNA (10). Because the active form of this enzyme is a dimer (2,3,11), there are minimally two potential sites for ATP binding. Bjornsen et al. (10) have provided kinetic evidence that suggested these two sites communicate. Recently Wong and Lohman (14) made heterodimers of Rep helicase containing ADP-AlF 4 tightly bound at the ATP site and covalently crosslinked single-stranded DNA at the DNA-binding site. Through the use of cis-(substrate analogues on the same subunit) and trans (substrate analogues on different subunits)-labeled dimers, they concluded that most of the observed ATPase activity was the result of the cis-labeled subunit. The ATPase activity of the subunit lacking single-stranded DNA was small because of a slow conformational change occurring prior to ADP release and not due to a diminished rate of ATP hydrolysis (14). Even though the kinetic mechanism for ATP hydrolysis by Rep helicase in the presence of single-stranded DNA has been clearly defined, the chemical mechanism for coupling ATP hydrolysis to double-stranded DNA unwinding is still unclear.
Hepatitis C virus (HCV) 1 genome encodes for an RNA helicase that presumably is essential for viral replication (15). Recently, a crystal structure of the HCV helicase has been reported (16). However, the protein was crystallized in the absence of a divalent metal cofactor such as Mg 2ϩ , which is required for ATP hydrolysis and nucleic acid unwinding, and in the absence of nucleic acid or nucleotide. Crystal structures of Rep helicase from E. coli (17) and a helicase from Bacillus stearothermophilus have also been reported. Co-crystals of Rep helicase with nucleic acid and ADP in the absence of Mg 2ϩ implicated motifs Ia, V, and III for single-stranded DNA binding, motifs I and IV for nucleotide binding, and motifs II and IV possibly functioned in the coupling of nucleotide and singlestranded DNA binding (17). These structures suggested a single nucleotide-binding site per monomer.
HCV helicase has a large intrinsic ATPase activity (k cat ϭ 3 s Ϫ1 (19)). Single-stranded DNA or RNA increases the k cat value up to 30-fold (19). In contrast, the k cat values for ATPase activities of bacterial helicases such as Rep helicase and heli-case II are increased over 1000-fold by single-stranded DNA (11). The turnover numbers for ATP hydrolysis by HCV helicase and Rep helicase in the presence of nucleic acid are similar. The differential DNA stimulation of these enzymes is the result of the large intrinsic ATPase activity of HCV helicase in the absence of single-stranded nucleic acid. This observation suggested that either HCV helicase ATPase activity was associated with a single site that was not tightly coupled to nucleic acid binding or HCV helicase may have ATPase activity associated with two separate sites on the enzyme. Further support for the latter suggestion was the finding that stoichiometry for ADP binding to the enzyme in the presence of excess Mg 2ϩ , F Ϫ , and poly(rU) was two nucleotides per nucleic acid-binding site (30).
HCV helicase forms spectrofluorometrically detectable intermediates with DNA (19) that react with ATP. The kinetic competence of the intermediates formed from the reaction of E⅐DNA with ATP or "E⅐ATP" with DNA was evaluated from pre-steady-state and steady-state data. In summary, the intermediates associated with these fluorescent changes were catalytically incompetent. These results suggested either the fluorescent changes were monitoring a catalytically inactive ATPbinding site with a fluorescently silent catalytically active site or the first cycle of interaction of enzyme with ATP was unique relative to subsequent cycles of ATP hydrolysis.

EXPERIMENTAL PROCEDURES
Materials-Rabbit muscle pyruvate kinase, rabbit muscle lactate dehydrogenase, NADH, ATP, phosphoenolpyruvate, and MOPS were from Sigma. HF16 and the 16-mer from Oligo Therapeutics, Inc., were purified by electrophoresis through 20% polyacrylamide gels in 6 M urea. Oligomer concentrations were calculated from the absorbance at 260 nm and the extinction coefficients provided by Oligo Therapeutics. The standard buffer was 0.5 M MOPS K ϩ and 3.5 mM MgCl 2 at pH 7.0. HCV helicase was purified as described previously (19). The concentration of monomer was calculated with ⑀ 280 ϭ 88 mM Ϫ1 cm Ϫ1 (19).
Assay of ATPase Activity-ATPase was assayed spectrophotometically by the absorbance change at 340 nm resulting from the oxidation of NADH that was coupled to the phosphorylation of ADP through pyruvate kinase and lactate dehydrogenase (20). Rates were calculated using a ⌬⑀ 340 ϭ 6.22 mM Ϫ1 cm Ϫ1 . A solution of the coupling components (40 units/ml pyruvate kinase, 40 units of lactate dehydrogenase, 8.0 mM phosphoenolpyruvate, and 800 M NADH) was prepared in the standard buffer at 5°C. This solution was diluted 4-fold with the standard buffer containing the selected concentrations of DNA and ATP. The reactions were initiated with HCV helicase. The t1 ⁄2 for phosphorylation of 5 M ADP in this assay was 3 s. The measured rate of ATP hydrolysis was not affected by decreasing the standard concentration of the coupling enzymes NADH and phosphoenolpyruvate by 50%.
Titration of HCV Helicase-Titration of helicase (E) with a ligand (L) was monitored by quenching of intrinsic protein fluorescence or by quenching of fluorescence of hexachlorofluoresceinyl-labeled DNA upon formation of the enzyme-ligand complex (E⅐L). It was assumed that this process could be described by the simplified mechanism of Equation 1, where K was the dissociation constant of E for L, [E t ] was the concentration of enzyme sites binding ligand, and [L t ] was the concentration of ligand.
Furthermore, it was assumed that the fractional fluorescence (F([L] t )) remaining was related to the total concentration of added ligand sites ([L t ]) by Equation 2 in which ⌬F ϱ was the fractional fluorescence decrease resulting from conversion of E to E⅐L.
Because the concentration of enzyme was in many cases comparable to the dissociation constant of the enzyme for L, the concentration of E⅐L for the simple scheme of Equation 1 was related to [L t ] by Equation 3.
During the titration of E with the 16-mer, which was monitored by the quenching of the intrinsic protein fluorescence ( ex ϭ 280 nm and em ϭ 340 nm), there was significant drift in the fluorescence signal. This was corrected for by a blank titration in which the experimental signal obtained after addition of the 16-mer to the enzyme was normalized to that observed after addition of an equivalent volume of buffer to the enzyme. The time between additions of sample to the protein was 30 s. Steady-state Kinetic Data Analysis-The dependence of the rate of ATP hydrolysis by HCV helicase on ATP, HF16, and the 16-mer concentrations was complicated. Equation 4 for a random rapid equilibrium scheme in which two molecules of ATP and one molecule of DNA bound to helicase described these data. The parameter of interest for these studies was k cat , which was relatively independent of the model fitted to the data. The substrate coefficients in Equation 4 were not used in the present studies.
Analysis of Time Courses-The time courses for fluorescence changes were described by exponential functions (Equation 5).
Where F(t) was the observed signal at time equal to t, F i was the respective amplitude of the phase associated with k i . F 0 was the final value of the fluorescence. The reaction of ligand (L) with enzyme was described by the two-step mechanism of Equation 6. For this mechanism, the dependence of the observed rate constant (k obs ) on ligand concentration ([L]) was given by Equation 7.
Where k obs ([L]) was the value of k obs for the indicated concentration of L; k Ϫ2 was the first-order rate constant at zero concentration of L; the sum of k 2 and k Ϫ2 was the first-order rate constant at infinite ligand concentration; and K was the concentration of L such that k obs ϭ k Ϫ2 ϩ 0.5k 2 . ⌬F([L]) was the amplitude of the fluorescence change as a function of [L] and ⌬F ϱ was the maximal signal change. Equations 7 and 8 were fitted to the amplitude data and the kinetic data globally to yield estimates for K, K Ϫ2 , k 2 , and ⌬F ϱ . If the value of K was much greater than the concentrations of L, Equation 7 was approximated by Equation 9.
If L were a substrate for the enzyme, Equation 6 was modified to include an additional step (k cat ) for the return of EL to E. In this modified scheme, the k Ϫ2 term in Equations 7-9 would be the sum of k Ϫ2 and k cat (k Ϫ2 , obs ). General Methods-Fluorescence spectra were recorded on a Kontron SFM 25 spectrofluorometer. Intrinsic protein fluorescence data ( ex ϭ 280 -300 nm, tryptophanyl and tyrosinyl residues, em ϭ 340 nm) were corrected for inner and outer filter effects. Fluorescence of HF16 was monitored with ex ϭ 510 nm and em ϭ 550 nm. Absorbance and conventional kinetic data were obtained from a UVIKON 860 spectrophotometer at 25°C or from a Spectramax 250 plate reader at 21°C. Rapid reactions were monitored on an Applied Photophysics SX.17MV Spectrophotometer (Leatherhead, UK). All concentrations of reactants are given after mixing on the stopped-flow spectrofluorometer. Entrance and exit slits were 10 mm in the absorbance mode and 2 mm in the fluorescence mode. The fluorescence of the hexachlorofluoresceinyl moiety was monitored on the stopped-flow spectrofluorometer with ex ϭ 500 nm and ex Ͼ530 nm. Intrinsic protein fluorescence was monitored with ex ϭ 280 -290 nm and em Ͼ305 nm. Stopped-flow time courses were an average of 4 -6 experiments. Fluorescence data from the stopped-flow spectrophotometer were presented as the voltage change from the beginning of the reaction. These changes were not normalized to the initial fluorescence of the solution because the solutions contained varying amounts of HF16 that were in excess of enzyme. For a group of experiments associated with a particular figure, the photomultiplier voltage was held constant. However, the magnitude of the voltage changes should not be compared between groups of experiments, only the voltage changes associated with a single group of experiments should be compared. The appropriate equations were fitted to the data by nonlinear least squares using SigmaPlot from Jandel Scientific (Corte Madera, CA).

RESULTS
k cat for ATP Hydrolysis by HCV Helicase with HF16 and the 16-mer as Single-stranded Nucleic Acid Effectors-The goal of the studies described herein was to determine if the formation of spectrally observed intermediates was a step in the steadystate catalytic cycle of ATP hydrolysis by HCV helicase. The criteria for kinetic competence of an intermediate in the catalytic cycle was that the value of the limiting first-order rate constant for the approach to the steady-state was at least 4-fold greater than the k cat for ATP hydrolysis. 2 Thus, the value of k cat was the number of importance determined from steadystate rate data. The value for k cat for ATP hydrolysis by HCV helicase is dependent on the single-stranded DNA effector. For example (dU) 18 increased the value of k cat 30-fold (19). With HF16 as the DNA effector, the ATPase activity of the enzyme was inhibited at small ATP concentrations and enhanced at large ATP concentrations (Fig. 1). The complex concentration dependence of the initial rate data was adequately described by Equation 4 with k cat ϭ 2.73 Ϯ 0.09 s Ϫ1 for HF16. The k cat for the 16-mer was calculated to be 36 Ϯ 7 s Ϫ1 from an analogous data set. The rate of hydrolysis of 1.6 mM ATP in the presence of 1.4 M HF16 (data not shown) was linearly dependent on enzyme concentrations between 1 and 60 nM. This result indicated that the value of k cat was independent of enzyme concentration so that it was valid to compare k cat values with kinetic parameters determined at different enzyme concentrations. The values for the K m of the enzyme for ATP at near-saturating concentrations of HF16 (1600 nM) and the 16-mer (6000 nM) were determined to be 320 Ϯ 50 and 410 Ϯ 20 M, respectively.
Affinity of HCV Helicase for HF16 and the 16-mer-The binding of HCV helicase to HF16 was monitored by the quenching of HF16 fluorescence upon formation of E⅐HF16. Titration of 22.4 nM HF16 with HCV helicase (Fig. 2) was described by Equations 2 and 3 (L ϭ HF16) with K ϭ 0.48 Ϯ 0.02 nM and ⌬F ϱ ϭ 0.457 Ϯ 0.006. Quenching of intrinsic protein fluorescence monitored the binding of HCV helicase to the 16-mer. Titration of 33.5 nM E (Fig. 2, inset) was described by Equations 2 and 3 (L ϭ 16-mer) with K ϭ 2.2 Ϯ 0.4 nM and ⌬F ϱ ϭ 0.377 Ϯ 0.006. Because the concentration of the fixed ligand was many fold greater than the K d for dissociation of E⅐DNA, the calculated values for K d were associated with considerable error. Nonetheless, these estimates for the values of K d were similar to the estimate values calculated from bimolecular association rate constants and dissociation rate constants (see below).
Kinetics for Equilibration of HCV Helicase with HF16 and the 16-mer-The time course for quenching of the fluorescence of HF16 by HCV helicase was a first-order process (Fig. 3,  inset). The pseudo first-order rate constant for this process (k obs ) was linearly dependent on the concentration of HF16 with [HF16] Ͼ [E] (Fig. 3). These data indicated that the K d for formation of the initial complex between E and HF16 (Equation 6) was much greater than the highest concentration of HF16 tested. Consequently, Equation 9 was fitted to these data to At saturating substrate concentration, the pseudo first-order rate constant for approach to the steady-state (k obs ) and k cat are given by Equations 11 and 12.
k obs ϭ k ϩ ␣k (Eq. 11) k cat ϭ ␣k ͑␣ϩ1͒ (Eq. 12) The expression of k obs in terms of ␣ and k cat is given by Equation 13.  The dependence of the pseudo first-order rate constant for the reaction of HF16 with HCV helicase appeared to extrapolate to a small value (Fig. 3, ordinate intercept) at zero concentration of HF16. This result suggested that the rate constant describing the dissociation of E⅐HF16 (k Ϫ2 in Equation 9) was very small. A direct measure of the dissociation of E⅐HF16 was made by monitoring the associated increase in fluorescence of HF16 ( ex ϭ 500 nm, em Ͼ530 nm) upon release from the enzyme in the presence of excess dextran sulfate that trapped free HCV helicase (Equation 14).
The fluorescence ( ex ϭ 500 nm, em Ͼ530 nm) of a solution of 13.2 nM E and 18.8 nM HF16 increased in a first-order process (k obs ) upon addition of 2.0 mM dextran sulfate. k obs had a value of 0.087 Ϯ 0.001 s Ϫ1 with 2.0 mM dextran sulfate and a similar value (0.109 Ϯ 0.001 s Ϫ1 ) with 0.4 mM dextran sulfate. Because k obs was similar for two dextran sulfate concentrations, this value was assigned to the dissociation rate constant (k Ϫ2 ) for E⅐HF16. Similar experiments with E⅐16 ( ex ϭ 280 nm, em Ͼ305 nm) yielded a value for k Ϫ2 of 0.096 Ϯ 0.009 s Ϫ1 .
The apparent dissociation constants (K) of E for HF16 and the 16-mer were calculated from the relationship K ϳ K d k Ϫ2 /k 2 , where k Ϫ2 was the effective rate constant for dissociation of E⅐HF16, and k 2 /K d was the effective association rate constant for reaction of HCV helicase and HF16 (Equation 6, L ϭ HF16). The values of K for HCV helicase and HF16 and the 16-mer were calculated from the kinetic data to be 0.14 and 1.0 nM, respectively, which were in reasonable agreement with the values estimated from titration data (0.48 and 2.2 nM, respectively).

Kinetics for the Approach to the Steady-state from E⅐ATP and HF16 or the 16-mer-
The time courses for the approach to the steady-state starting with 2.0 mM ATP, 13.2 nM E⅐ATP, and selected concentrations of HF16 were first-order processes (k obs ). The dependence of k obs on HF16 concentration was described by Equations 7 and 8 with k Ϫ2,obs ϭ 0.37 Ϯ 0.02 s Ϫ1 , k 2 ϭ 1.2 Ϯ 0.2 s Ϫ1 , K ϭ 240 Ϯ 70 nM, and ⌬F ϱ ϭ 0.40 Ϯ 0.03 V (Fig.  4). Similarly, the approach to the steady-state starting with 50 nM E⅐ATP and selected concentrations of the 16-mer was a first-order processes. Because the highest concentration of the 16-mer was significantly less than the K d for the reaction, Equation 9 was fitted to the data with k Ϫ2,obs ϭ 2.0 Ϯ 0.1 s Ϫ1 , Dissociation of E⅐HF16⅐ATP or E⅐16⅐ATP-The dissociation constant of the enzyme for HF16 was increased over 100-fold by ATP (compare data of Fig. 2 with that of Fig. 4, inset). Consequently, mixing ATP with a solution of E⅐HF16 ([HF16] Ͻ [K HF16 ]) should result in the partial dissociation of HF16 from E⅐ATP as depicted in Equation 15.
The release of HF16 from the enzyme was monitored by the associated increase in fluorescence of HF16 ( ex ϭ 500 nm, em Ͼ530 nm). The approach to the steady-state upon addition of 1.0 mM ATP to 9.3 nM E⅐HF16 that was generated from 9.3 nM helicase and 17.7 nM HF16 was a biphasic process that indicated that ATP bound to the E⅐HF16 complex to form an intermediate ("E⅐HF16⅐ATP") from which HF16 dissociated (Fig. 5). Equation 5 (i ϭ 2) was fitted to these data to give F 0 ϭ Ϫ0.494 Ϯ 0.003 V, F 1 ϭ Ϫ3.98 Ϯ 0.2 V, k 1 ϭ 1.01 Ϯ 0.03 s Ϫ1 , F 2 ϭ 4.3 Ϯ 0.2 V, and k 2 ϭ 0.403 Ϯ 0.008 s Ϫ1 . The early phase of the reaction was attributed to formation of E⅐HF16⅐ATP from E⅐HF16 and ATP, whereas the late phase of the reaction was attributed to dissociation of HF16 from E⅐HF16⅐ATP. When dextran sulfate was added to a steady-state mixture of ATP, HCV helicase, and HF16 to trap free E as it was cycled during catalysis, the time course for formation of free HF16 was monophasic (Fig. 5, inset). Equation 5 (i ϭ 1) was fitted to these data with F 0 ϭ Ϫ0.67 Ϯ 0.002 V, F 1 ϭ 0.067 Ϯ 0.002 V, and k 1 ϭ 0.341 Ϯ 0.002 s Ϫ1 . The value of the rate constant for the latter process was in good agreement with the value for the late phase of the reaction in the presence of ATP alone (0.403 s Ϫ1 ).
ATP also induced the dissociation of E⅐16. The release of the 16-mer from E⅐16 was monitored by the associated change in intrinsic protein fluorescence ( ex ϭ 280 nm, em Ͼ305 nm) upon formation of (E⅐ATP) from (E⅐16⅐ATP). The addition of 1.0 mM ATP to a solution of 50 nM E⅐16 generated from 50 nM E and 75 nM 16-mer resulted in a biphasic increase in fluorescence. These data (not shown) were described by Equation 5 (i ϭ 2) with F 0 ϭ Ϫ0.129 Ϯ 0.001 V, F 1 ϭ Ϫ0.83 Ϯ 0.07 V, k 1 ϭ 2.34 Ϯ 0.05 s Ϫ1 , F 2 ϭ 1.08 Ϯ 0.07 V, and k 2 ϭ 1.70 Ϯ 0.02 s Ϫ1 . The early phase of the reaction was attributed to formation of E⅐16⅐ATP from E⅐16 and ATP, whereas the late phase of the reaction was attributed to dissociation of the 16-mer from E⅐16⅐ATP.
Kinetics for the Approach to the Steady-state from E⅐HF16 or E⅐16 and ATP-The previous results indicated that the reaction of E⅐HF16 with ATP to form E⅐HF16⅐ATP was associated with a fluorescence decrease (Fig. 6A). Because the concentration of HF16 in these experiments was less than the dissociation constant of E⅐ATP for HF16, E⅐HF16⅐ATP partially dissociated to E⅐ATP and HF16. The dissociation phase of the reaction was eliminated by including high concentrations of HF16 in the reaction. Thus, the reaction of 2.0 mM ATP with 50 nM E⅐HF16, which was generated by premixing 50 nM E with 492 nM HF16, was a first-order process (Fig. 6A). Equation 5 (i ϭ 1) was fitted to these data with F 0 ϭ 0.262 Ϯ 0.01 V, F 1 ϭ Ϫ0.262 Ϯ 0.001 V, and k 1 ϭ 1.67 Ϯ 0.03 s Ϫ1 . This rate constant described the approach to the steady-state after mixing E⅐HF16 with ATP. The values of pseudo first-order rate constant (k obs ) and the amplitudes of the fluorescence change for this reaction were determined for selected ATP concentrations (Fig. 6B). Equations 7 and 8 were fitted to these data with K ϭ 121 Ϯ 7 M, k 2 ϭ 1.82 Ϯ 0.03 s Ϫ1 , k Ϫ2,obs ϭ 0.05 Ϯ 0.01 s Ϫ1 , and ⌬F ϱ ϭ 0.29 Ϯ 0.01 V. If E⅐ATP⅐HF16 were a catalytic intermediate in ATP steady-state ATP hydrolysis as depicted in Equation 15, the value of k Ϫ2,obs should have been the sum of k Ϫ2 and k cat .
A data set similar to that for E⅐ATP and HF16 was collected   for the reaction of E⅐ATP with the 16-mer. E⅐16 was generated by premixing 50 nM E with 12 M 16-mer. Addition of 1.0 mM ATP resulted in a first-order increase in intrinsic protein fluorescence (data not shown). Equation 5 (i ϭ 1) was fitted to these data with F 0 ϭ 0.109 Ϯ 0.002 V, F 1 ϭ Ϫ0.109 Ϯ 0.002 V, and k 1 ϭ 2.55 Ϯ 0.07 s Ϫ1 . The values of pseudo first-order rate constant (k obs ) and the amplitudes of the fluorescence change for this reaction were determined for selected ATP concentrations. Equations 7 and 8 were fitted to these data with K ϭ 170 Ϯ 20 M, k 2 ϭ 2.9 Ϯ 0.2 s Ϫ1 , k Ϫ2,obs ϭ 0.49 Ϯ 0.06 s Ϫ1 , and ⌬F ϱ ϭ 0.104 Ϯ 0.004 V. The kinetic parameters for these reactions are summarized in Table I. DISCUSSION Helicases catalyze the ATP-dependent unwinding of doublestranded DNA. The mechanisms for the unwinding reaction are being actively investigated in numerous laboratories (2,18,(21)(22)(23)(24). Recent mechanistic studies on E. coli Rep helicase are forming the basis for a detailed understanding of the kinetic mechanism for DNA unwinding by a helicase (2). Lohman (6) has classified the mechanism for helicase action as being either "active or "passive." In a passive type mechanism, the enzyme binds preferentially to single-stranded DNA and unwinds double-stranded DNA by binding to single-stranded DNA that is formed transiently from the duplex as the result of thermal fluctuations. In contrast, an active mechanism requires that the enzyme bind to both double-stranded DNA and singlestranded DNA, and the enzyme destabilizes multiple base pairs during each catalytic step. This mechanism requires that the enzyme have multiple DNA-binding sites either on the same subunit or as an oligomeric protein. During catalysis, one DNA-binding site alternatively binds double-stranded and single-stranded DNA as the enzyme rolls along the doublestranded DNA. Whichever mechanism is operative for unwinding of double-stranded DNA, the coupling of ATP hydrolysis to translocation is crucial for the unwinding process.
The coupling of ATP hydrolysis to conformational transitions of other domains in members of helicase superfamily I is being investigated by site-directed mutagenesis of conserved amino acids in each of the six motifs. For example, E. coli helicase II mutants D248N, K35M, and R284A greatly reduced the singlestranded DNA-stimulated ATPase activity (25)(26)(27). Recently, site-specific substitutions in each of the six motifs of HSV helicase (superfamily 1) produced large effects when in motifs I and II and small effects when in motifs III-VI on DNA-dependent ATPase activity (28). To supplement these studies, our goal was to characterize the single-stranded DNA-stimulated ATPase activity of HCV helicase by steady-state and presteady-state kinetic techniques. In particular, the kinetic competence of spectrofluorometrically observed intermediates was determined. HCV helicase quenched the fluorescence of HF16, whereas HF16 and the 16-mer quenched the intrinsic protein fluorescence. The kinetics for formation of these intermediates with modified fluorescence properties were compared with the values for k cat for ATP hydrolysis to decide whether or not the intermediates were kinetically competent for catalysis of ATP hydrolysis.
The dissociation of DNA from the enzyme in the presence of ATP was too slow for this process to be an obligatory step in the catalytic cycle. The rate constants for dissociation of E⅐HF16 and E⅐16 in the presence of 2.0 mM ATP were 0.341 s Ϫ1 and 1.70 s Ϫ1 , respectively, whereas the values of k cat were 2.73 and 36 s Ϫ1 , respectively. Because DNA dissociated from E⅐DNA infrequently during the catalytic cycle, the enzyme cycled primarily between E⅐DNA and E⅐DNA⅐ATP during ATP hydrolysis. If E⅐DNA⅐ATP were on the pathway for hydrolysis of ATP, the pseudo first-order rate constant for the approach to the steady-state starting with E⅐DNA and a saturating concentration of ATP must be at least 4-fold greater than the value of k cat . 2 Furthermore, the concentration of ATP that resulted in 50% of the maximal amount of this intermediate must be greater than or equal to the K m for ATP. These conditions were not met by the pre-steady-state intermediate formed from E⅐DNA and ATP. For instance, the maximal value for the first-order rate constant for the approach to the steady-state concentration of E⅐HF16⅐ATP starting with E⅐HF16 and ATP was 1.82 s Ϫ1 , whereas k cat was 2.73 s Ϫ1 . The minimal value for the rate constant for approach to the steady state at saturating concentration of ATP has to be 10.9 s Ϫ1 for this process to be on the catalytic cycle. Furthermore, the concentration of ATP that yielded 50% of the maximal amount of the intermediate for a saturating concentration of HF16 should be equal to the K m for the steady-state rate of ATP hydrolysis. The value of K m for the ATP at saturating concentrations of HF16 was 320 M, whereas the concentration of ATP that yield 50% of the intermediate was only 3 M. Thus, formation of this intermediate was not a competent step in the catalytic cycle of ATP hydrolysis.
These results suggested either that the formation of E⅐DNA⅐ATP occurred independently of and was unrelated to the catalytic process or that formation of this species introduced the active form of the enzyme into the catalytic cycle. If the latter were the case, an isomerization step could be included in the kinetic mechanism such that reaction of enzyme with the first molecule of ATP was different than its reaction with subsequent molecules of ATP (Scheme I). In this model, steady-state ATP hydrolysis was mediated principally through cycling of ‫ޅ‬ HF16 (Scheme I), whereas the reaction monitored on the stopped-flow spectrofluorometer was conversion of E HF16 to ‫ޅ‬ HF16 . These data have not addressed the fate of ATP in the conversion of E HF16 to ‫ޅ‬ HF16 . Possibly a structural isomerization dependent on phosphorylation of the enzyme was involved that transformed inactive enzyme into a catalytically active form. If this were the case, formation of ‫ޅ‬ HF16 and its reversion to E HF16 (Ͻ0.05 s Ϫ1 in Scheme I) could involve net hydrolysis of ATP.
The data presented herein for hydrolysis of ATP by E⅐HF31 suggested that the first cycle of interaction of ATP with E⅐HF31 was unique relative to subsequent cycles of interaction of the enzyme with ATP. This concept has been suggested for other DNA unwinding proteins such as the E. coli Rep helicase and the E. coli Rec A protein. For Rep helicase, the approach to the steady state for the reaction of 1.0 mM ATP with enzyme preloaded with 2-aminopurine-labeled dT(pT) 15 appeared to be too slow for the formation of ternary complex to be a kinetically competent step during ATP hydrolysis (10). Furthermore, Wong and Lohman (14) have suggested from studies on the K28I mutant of Rep helicase that a global conformation change of the protein occurs prior to the first turnover. Similar results have been described for a mutant E. coli Rec A protein (H163W) that catalyzed single-stranded DNA-dependent ATPase. The fluorescence of the H163W protein was used to monitor confor-SCHEME 1. Isomerization model for hydrolysis of ATP by E⅐HF16.
mational changes occurring during ATP hydrolysis (29). The first cycle of interaction of the mutant protein with ATP was too slow to be on the catalytic pathway. Consequently, Stole and Bryant (29) proposed an isomerization model similar to that described in Scheme I. The steady-state hydrolysis of ATP by these DNA unwinding proteins was preceded by a unique isomerization of the initial ternary enzyme-DNA-ATP complex that may be a general phenomenon for all DNA unwinding proteins.