A Partially Functional DNA Helicase II Mutant Defective in Forming Stable Binary Complexes with ATP and DNA A

To address the functional significance of motif III in Escherichia coli DNA helicase II, the conserved aspartic acid at position 248 was changed to asparagine. UvrDD248N failed to form stable binary complexes with either DNA or ATP. However, UvrDD248N was capable of forming an active ternary complex when both ATP and single-stranded DNA were present. The DNA-stimu- lated ATPase activity of UvrDD248N was reduced relative to that of wild-type UvrD with no significant change in the apparent K m for ATP. The mutant protein also demonstrated a reduced DNA unwinding activity. The requirement for high concentrations of UvrDD248N to achieve unwinding of long duplex substrates likely reflects the reduced stability of various binary and ter- nary complexes that must exist in the catalytic cycle of a helicase. The data suggest that motif III may act as an interface between the ATP binding and DNA binding domains of a helicase. The uvrDD248N allele was also characterized in genetic assays. The D248N protein complemented the UV- sensitive phenotype of a uvrD deletion strain to levels nearly equivalent to wild-type helicase II. In contrast, the mutant protein only partially complemented the mu-tator phenotype. A correlation between the level of ge- netic complementation and the helicase activity of UvrDD248N is discussed. The the helicase reaction using the 346-bp fully duplex substrate were resolved on a 6% nondenaturing polyacrylamide gel as described (George et al ., 1994). Polyacrylamide gels were imaged and quantified using phosphor storage technology and software (Molecular Dynamics). DNA-dependent ATPase Assays— The hydrolysis of ATP to ADP was measured as described previously (Matson and Richardson, 1983). ATPase reaction mixtures were identical with helicase reaction mixtures except that they contained 9 (cid:109) M M13mp7 ssDNA (nucleotide phosphate) and [ 3 H]ATP (22 cpm/pmol). For k cat determinations, the [ 3 H]ATP concentration was 540 (cid:109) M . For K m determinations, the [ 3 H]ATP concentration was varied between 25 and 500 (cid:109) M .

The unwinding of double-stranded DNA (dsDNA) 1 is essential for the complex processes of DNA replication, repair, and recombination (Lohman, 1992;Lohman, 1993;Matson et al., 1994;Matson and Kaiser-Rogers, 1990). The unwinding of duplex RNA and DNA-RNA hybrids is equally important in transcription (Brennan et al., 1987), RNA processing (Schmid and Linder, 1992;Wassarman and Steitz, 1991), and translation (Ray et al., 1985;Schmid and Linder, 1992;Wassarman and Steitz, 1991). Helicases catalyze these reactions by disrupting the hydrogen bonds between the complementary base pairs of duplex nucleic acid in an NTP hydrolysis-dependent reaction.
Helicases of prokaryotic, eukaryotic, and viral origin have been isolated and described (Lohman, 1992;Lohman, 1993;Matson, 1991;Matson et al., 1994;Matson and Kaiser-Rogers, 1990; Thommes and Hubscher, 1992). However, the mechanism of helicase-catalyzed nucleic acid unwinding is not fully understood. It has been proposed that most, perhaps all, helicases function as oligomers to provide for multiple DNA binding sites (Lohman, 1992;Lohman, 1993). In support of this notion, the assembly state of all DNA helicases examined thus far is a dimer or hexamer (Lohman, 1992;Lohman, 1993). A rolling model for helicase-catalyzed DNA unwinding has been proposed for Rep protein . Conformational changes in the protein, induced by ligand binding, allow the protein complex to alternate between states in which subunits interact with single-stranded DNA (ssDNA) or dsDNA. Evidence for such a mechanism has not yet been provided for other helicases.
Sequence analysis of helicases and putative helicases from a wide array of sources has revealed seven distinct regions of homology (Gorbalenya et al., 1988;Gorbalenya et al., 1989;Hodgman, 1988). These regions are referred to as motifs I, Ia, and II-VI. Presumably, these motifs partially define functional domains relevant to the known biochemical activities of these proteins. For example, motifs I and II are known to define a nucleotide binding/hydrolysis domain (Brosh and Matson, 1995;George et al., 1994;Fry et al., 1986;Jindal et al., 1994;Pause and Sonenberg, 1992;Walker et al., 1982;Washburn and Kushner, 1993;Zavitz and Marians, 1992). The functional significance of the remaining five motifs remains to be established but likely relates to activities required for polynucleotide unwinding.
To continue to address the structure-function relationship of a DNA helicase, we have replaced a highly conserved aspartic acid residue with asparagine in motif III of UvrDp (DNA helicase II). Biochemical characterization of the UvrDD248N protein has been performed to further our understanding of the functional significance of motif III in the mechanism of helicase-catalyzed DNA unwinding. Escherichia coli DNA helicase II is an excellent model enzyme for these studies because it has been extensively characterized in vitro. Helicase II is a DNAstimulated ATPase (Richet and Kohiyama, 1976) that prefers to unwind a dsDNA substrate on which a 3Ј-ssDNA tail is available for binding to initiate unwinding (Matson, 1986). At higher protein concentrations the enzyme can also initiate unwinding from a blunt end or nick (Runyon et al., 1990;Runyon and Lohman, 1989;. Helicase II unwinds DNA in a protein concentration-dependent manner (Abdel-Monem et al., 1977;Kuhn et al., 1979;Matson and George, 1987) and may act as a helix-destabilizing protein remaining bound to the ssDNA generated during an unwinding reaction (Runyon et al., 1990;Wessel et al., 1990).
Biochemical and genetic studies have demonstrated a role for helicase II in two DNA repair pathways, methyl-directed mismatch repair (Grilley et al., 1990;Lahue et al., 1989; and UvrABC-mediated nucleotide excision repair (Caron et al., 1985;Husain et al., 1985;Orren et al., 1991). The uvrDD248N allele was assessed for its ability to complement in DNA repair in an effort to characterize the functional significance of the conserved aspartic acid as it relates to the defined roles of helicase II in the cell.
DNA and Nucleotides-Bacteriophage M13mp18 and M13mp7 ss-DNAs and their derivatives were prepared as described (Lechner and Richardson, 1983). All unlabeled nucleotides were from U.S. Biochemical Corp. except ATP␥S, which was from Boehringer Mannheim. Plasmid pET81F1ϩ was kindly provided by Dr. P. J. Laipis (University of Florida). Plasmid pET9d was purchased from Novagen, Inc. Concentrations of DNA and nucleotides were determined by UV spectrophotometry using published extinction coefficients and are expressed as nucleotide equivalents.
Enzymes-Restriction endonucleases, DNA polymerase I (large fragment), phage T7 DNA polymerase, and phage T4 polynucleotide kinase were purchased from New England Biolabs, Inc., or U.S. Biochemical Corp. The reaction conditions used were those suggested by the supplier.
E. coli DNA helicase II was purified from BL21(DE3)/pLysS cells containing the pET9d-H2wt expression plasmid . The UvrDD248N protein was purified from BL21(DE3)⌬uvrD/pLysS cells containing the pET81-H2D248N expression plasmid. Ten liters of cells grown in LB media (Miller, 1972) plus ampicillin (200 g/ml) and chloramphenicol (30 g/ml) were induced for protein expression during log phase with isopropyl-␤-D-thiogalactopyranoside (0.5 mM) and harvested 4 h after induction. The procedure of  was used to purify both the wild-type and mutant proteins except for the following modifications in the purification of UvrDD248N. After pooling the peak fractions containing UvrDD248N from the heparin-agarose column, the pool was dialyzed against buffer A (20 mM KPO 4 (pH 7.2), 10% glycerol, 5 mM ␤-mercaptoethanol, and 0.5 mM EGTA) plus 150 mM KCl and loaded on to a DEAE-Sephadex A-50 column (4 ϫ 2.5-cm inner diameter) that had been equilibrated in buffer A plus 150 mM KCl. The DEAE-Sephadex column was washed to base line with buffer A and 150 mM KCl, and the protein eluted with a linear 10-column volume KCl gradient (150 mM-500 mM) in buffer A. The peak fractions containing UvrDD248N eluted in the range of 300 -440 mM KCl. These fractions were pooled, diluted to a protein concentration of 100 g/ml, and dialyzed against buffer G (20 mM Tris-HCl (pH 8.3), 20% glycerol, 1 mM EDTA, 0.5 mM EGTA, and 15 mM ␤-mercaptoethanol) plus 150 mM NaCl. The dialysate was loaded on to dsDNA-cellulose and further purified as described previously . Final protein concentration was determined using the helicase II extinction coefficient .

Methods
Site-directed Mutagenesis and DNA Constructions-For mutagenesis and expression, the uvrD gene was cloned into pET81F1ϩ, a T7 expression vector with a bacteriophage f1 origin of replication (Tanhauser et al., 1992). pET81F1ϩ was digested to completion with with BamHI, the 5Ј-overhang ends were filled in using DNA polymerase I (large fragment), and the DNA was extracted with phenol-chloroform followed by an ethanol precipitation. The DNA was resuspended in TE (10 mM Tris-HCl (pH 8.0), 1 mM EDTA) and digested with NcoI. pET9d-H2wt was digested with HindIII, the 5Ј-overhang ends were filled in using DNA polymerase I (large fragment), and the DNA was extracted with phenol-chloroform followed by an ethanol precipitation. The DNA was resuspended in TE and digested with NcoI, and the 2.5-kilobase pair DNA fragment containing the entire uvrD gene was isolated on an agarose gel and extracted using Geneclean (Bio 101, Inc.). The 2.5kilobase pair DNA fragment was ligated into pET81F1ϩ, prepared as described above, resulting in the construction pET81-H2wt. pET81-H2wt was the target for site-directed mutagenesis using published procedures (Kunkel et al., 1991;Zoller and Smith, 1991). Oligonucleotide 5Ј-CTGGTCGTCATTACCGACGAT-3Ј was used to alter codon 248 of uvrD from GAT (Asp) to AAT (Asn). The entire uvrD gene in pET81-H2D248N and pET81-H2wt was sequenced on a model 373A DNA Sequencer (Applied Biosystems) using the Taq DyeDeoxy TM Terminator Cycle Sequencing Kit (Applied Biosystems) to confirm the mutation at position 248. A 2.1-kilobase pair NdeI fragment, containing a portion of the uvrD gene containing the uvrDD248N mutation plus some 3Јflanking sequence was moved from pET81-H2D248N into pET9d-H2wt to yield pET9d-H2D248N.
Genetic Assays-The viability of bacterial strains exposed to ultraviolet (UV) light was measured as described previously (Brosh and Matson, 1995). The spontaneous mutation frequency for each cell strain was determined as described .
DNA Binding Assays-A nitrocellulose filter binding assay was used to measure binding of UvrD protein to DNA (Matson and Richardson, 1985). Binding reaction mixtures (20 l) contained 25 mM Tris-HCl (pH 7.5), 3 mM MgCl 2 , 20 mM NaCl, 5 mM 2-mercaptoethanol, the 92-base pair (bp) partial duplex helicase substrate (approximately 2 M nucleotide phosphate) (3.26 ϫ 10 8 cpm mol Ϫ1 ), and the indicated amount of helicase II. To determine the effect of nucleotide on DNA binding, reaction mixtures were altered to contain either 3 mM ATP␥S, 3 mM ADP, or no nucleotide as indicated in the appropriate figure legends. The reaction mixture was incubated at 37°C for 10 min, diluted with 1 ml of 37°C prewarmed reaction buffer containing 50 g/ml bovine serum albumin, and passed through a nitrocellulose filter (0.45 M; Whatman) at a flow rate of 4 ml/min. The filters were washed with an additional 2 ml of prewarmed reaction buffer. The dried filters were counted in a liquid scintillation counter. Background radioactivity bound in the absence of UvrD protein was subtracted from total radioactivity bound to the filter. Nitrocellulose filters were pretreated by boiling in deionized distilled water for 20 min and stored in reaction buffer containing 50 g/ml bovine serum albumin.
Data Analysis-The fraction of the 92-bp partial duplex DNA substrate specifically bound by UvrD protein was determined from the nitrocellulose filter binding assays. A Hill plot was used to analyze the data (Yong and Romano, 1995).
K d is the dissociation constant of the DNA-protein complex, [Pt] is the total concentration of UvrD protein present in the reaction, and f is the ratio of the amount of the bound DNA over the total amount of DNA present in the reaction. The logarithm of [Pt] was plotted against the logarithm of (f/(1 Ϫ f)), and the y intercept represented the logarithm of K d . ATP Binding Assays-Gel filtration was used to measure binding of ATP to UvrD proteins (Sung et al., 1988). Binding reaction mixtures (30 l) contained 10 mM Tris-HCl (pH 7.5), 4 mM MgCl 2 , 206 M [ 3 H]ATP (465 cpm/pmol), and either 2.2 M UvrD (monomer) or 2.2 M UvrDD248N (monomer). The reaction mixture was incubated at room temperature for 25 min and then applied, at room temperature, onto a 1.25-ml Sephadex G-50 column equilibrated in 10 mM Tris-HCl (pH 7.5), 100 mM NaCl, 4 mM MgCl 2 , and 50 g/ml bovine serum albumin. Four drop (110-l) fractions were collected. A 20-l aliquot of each fraction was mixed with 3 ml of Ecoscint A (National Diagnostics) and counted in a liquid scintillation counter. Background counts were determined by performing the same experiment in the absence of UvrD protein and subtracted from experimental data. Protein elution from the column was determined by immunoblotting.
Proteolytic Digestions of Helicase II Protein-UvrD and UvrDD248N proteins were lightly digested with ␣-chymotrypsin (Sigma) as described previously (Chao and Lohman, 1990). Reaction mixtures (15 l) contained either 1.8 M UvrD (monomer) or 1.5 M UvrDD248N (monomer). To determine the effect of ATP, Mg 2ϩ , or ssDNA on the proteolysis of helicase II, reaction mixtures contained 2 mM ATP␥S, 3 mM MgCl 2 , and/or 214 M M13mp7 ssDNA (nucleotide phosphate). One l of freshly prepared chymotrypsin (5 ng/l) was added to each reaction mixture, and the proteolysis reaction was incubated at 37°C for 2 min. The reactions were quenched by adding 15 l of 2 ϫ SDS gel loading buffer (Laemmli, 1970), and the samples were boiled for 2 min prior to electrophoresis on a 12% polyacrylamide gel run in the presence of SDS. The gel was stained with Coomassie Blue and destained.
Glutaraldehyde Cross-linking-Protein cross-linking reactions were performed at room temperature at a final concentration of UvrD or UvrDD248N of 2 M (monomer). Cross-linking reaction mixtures (20 l) contained 20 mM Tricine (pH 8.3), 50 mM NaCl, 20% glycerol, 5 mM MgCl 2 , and 5 mM 2-mercaptoethanol as described previously . In reactions containing ATP␥S or (dT) 10 , the concentrations were 2 mM and 4.2 M, respectively. Reactions were preincubated on ice for 15 min prior to the addition of glutaraldehyde (EM Grade, Electron Microscopy Sciences) to a final concentration of 0.01%. Reactions were quenched after 30 min by adding lysine acetate to a final concentration of 10 mM. Samples were resolved on 9% polyacrylamide gels containing SDS and stained with Coomassie Blue.
Helicase Assays-The partial duplex helicase substrates were constructed as described previously (Brosh and Matson, 1995). The 346-bp blunt-ended helicase substrate was constructed as described . Helicase assay reaction mixtures (20 l) contained 25 mM Tris-HCl (pH 7.5), 3 mM MgCl 2 , 20 mM NaCl, 5 mM 2-mercaptoethanol, 3 mM ATP, and the indicated amount of helicase II. The concentration of the partial duplex helicase substrate in the reaction was approxi-mately 2 M (nucleotide phosphate). The concentration of the 346-bp blunt end duplex substrate in the assay was approximately 0.5 M (nucleotide phosphate). Reactions were initiated by the addition of ATP and incubated at 37°C for 10 min with the exception of the 851-bp partial duplex DNA substrate, which was incubated for 20 min. Reactions were terminated by the addition of 10 l of 50 mM EDTA, 40% glycerol, 0.5% SDS, 0.1% bromphenol blue, 0.1% xylene cyanol. The products of helicase reactions were resolved on 6 or 8% nondenaturing polyacrylamide gels as described (Matson and George, 1987). The products of the helicase reaction using the 346-bp fully duplex substrate were resolved on a 6% nondenaturing polyacrylamide gel as described . Polyacrylamide gels were imaged and quantified using phosphor storage technology and software (Molecular Dynamics).
DNA-dependent ATPase Assays-The hydrolysis of ATP to ADP was measured as described previously (Matson and Richardson, 1983). ATPase reaction mixtures were identical with helicase reaction mixtures except that they contained 9 M M13mp7 ssDNA (nucleotide phosphate) and [ 3 H]ATP (22 cpm/pmol). For k cat determinations, the [ 3 H]ATP concentration was 540 M. For K m determinations, the [ 3 H]ATP concentration was varied between 25 and 500 M.

RESULTS
The conserved primary sequence and relative position of motif III in a number of E. coli and Saccharomyces cerevisiae DNA helicases ( Fig. 1) suggest that this region of the protein may have functional significance in the mechanism of helicasecatalyzed unwinding. To begin to address the functional role of motif III in DNA helicases the highly conserved aspartic acid in motif III of the well-characterized UvrD protein from E. coli was altered by site-directed mutagenesis. The negatively charged aspartic acid was replaced by a neutral asparagine (Fig. 1). The resulting mutant, designated UvrDD248N, was expressed, purified, and biochemically characterized.

Biochemical Characterization of UvrDD248N
The procedure used to purify UvrDD248N was altered slightly (see "Experimental Procedures") as compared with that used to purify the wild-type protein, since the mutant protein failed to bind to a ssDNA-cellulose column. The modified procedure resulted in purification of the mutant protein to apparent homogeneity as judged by the presence of a single band of protein on an SDS-polyacrylamide gel (see Fig. 3, lane F).
DNA Binding-The marked reduction in binding to ssDNAcellulose prompted us to examine the DNA binding properties FIG. 1. The D248N amino acid substitution in motif III of UvrD. The relative positions of the seven previously described helicase motifs are shown with the conserved sequence of motif III (Gorbalenya et al., 1988;Hodgman, 1988). The amino acid sequence of motif III from UvrD is compared with the corresponding segment found in S. cerevisiae and E. coli DNA helicases of superfamily 1. The motifs for the yeast helicases were described previously with the exception of scHelI (D. Bean and S. W. Matson, unpublished data): Dna2p (Budd and Campbell, 1995), Srs2p (Aboussekhra et al., 1989), and Pif1p (Foury and Lahaye, 1987). Alignments for the yeast proteins are as described previously with the exception of Dna2p (motif IV), and scHelI (motifs I-VI), which were determined in this laboratory. The motifs for E. coli helicases Rep, RecB, and UvrD were aligned as described previously (Gorbalenya et al., 1988;Hodgman, 1988). The sequences for E. coli proteins HelD (Wood and Matson, 1989) and TraI (Frost et al., 1994) were aligned with the other proteins in this study. Shown below the alignment is the sequence of UvrDD248N, in which the highly conserved aspartic acid is replaced with an asparagine.
FIG. 2. DNA binding by UvrDD248N and UvrD. Binding assays were performed as described under "Experimental Procedures" using the indicated amounts of UvrD (Ⅺ) or UvrDD248N (G). DNA binding incubations were performed in the presence of no nucleotide (A), 3 mM ADP (B), or 3 mM ATP␥S (C). Background values were typically less than 2% and have been subtracted from the reported data. These data represent the average of at least three independent determinations. of the mutant protein. Nitrocellulose filter binding assays using the 92-bp partial duplex DNA as a ligand were performed with both the UvrDD248N and the wild-type proteins (Fig. 2). The mutant enzyme was dramatically compromised in its ability to bind DNA in the absence of nucleotide or in the presence of ADP (Fig. 2, A and B). However, in the presence of the poorly hydrolyzed ATP analog ATP␥S, UvrDD248N demonstrated a binding isotherm similar to that measured for the wild-type enzyme (Fig. 2C). Moreover, the D248N mutant protein was able to bind ssDNA-cellulose when the resin was equilibrated in buffer containing 2 mM ATP (data not shown).
To determine the apparent dissociation constants (K d ) for the protein-DNA complexes, the data were analyzed by a Hill plot as described under "Experimental Procedures." In the absence of nucleotide, the apparent K d values for the wild-type and UvrDD248N proteins were 7 nM (monomer) and 165 nM (monomer), respectively (Table I). In the presence of ADP the apparent K d values for the wild-type and UvrDD248N proteins were 9 nM (monomer) and 340 nM (monomer), respectively. Thus, the stability of the UvrDD248N protein-DNA complex formed in the absence of nucleotide or in the presence of ADP was significantly reduced compared with the corresponding wild-type protein-DNA complex. In the presence of ATP␥S, the apparent K d values for the wild-type and mutant proteins were 3.8 nM (monomer) and 1.7 nM (monomer), respectively. In contrast to the results presented above, the wild-type and the mutant protein-DNA complexes, formed in the presence of ATP␥S, are nearly equal in stability.
Partial Proteolytic Cleavage-To probe further the interactions of UvrDD248N with ssDNA and ATP we examined the pattern of proteolytic cleavage by chymotrypsin in the presence of various ligands. Chao and Lohman (1990) have demonstrated that helicase II is protected from cleavage by chymotrypsin when ATP␥S and/or ssDNA is present and bound by the protein. We performed a set of limited proteolysis experiments with the wild-type and UvrDD248N proteins to test for the ability of these proteins to bind ATP␥S and/or M13 ssDNA (Fig.  3). In the absence of ATP␥S or ssDNA, UvrDD248N is cleaved by chymotrypsin into two fragments with approximate molecular masses of 53 and 29 kDa (lane G). The same cleavage pattern is obtained with wild-type UvrD (lane B). Thus, the mutation has not drastically altered the conformation of the protein. Wild-type UvrD protein is protected from cleavage when ssDNA is present as expected (lane D). In contrast, UvrDD248N is not protected from limited chymotrypsin cleavage when ssDNA is added (lane I). This result suggests that the mutant protein failed to bind stably to ssDNA in agreement with the nitrocellulose filter binding assay results. Binding of ATP␥S to wild-type helicase II also rendered the protein resistant to chymotrypsin cleavage (lane C). In contrast, UvrDD248N failed to be protected from cleavage by ATP␥S (lane H), suggesting that the mutant is also defective in its interaction with nucleoside triphosphate in the absence of DNA. When both ATP␥S and ssDNA are present, both UvrDD248N (lane J) and the wild-type protein (lane E) are protected from proteolytic cleavage. Thus, the mutant protein was capable of forming a ternary complex with ATP and ssDNA when both ligands were present. However, the above results indicate that UvrDD248N exhibited a defect in binding ssDNA in the absence of ATP as well as a defect in binding ATP in the absence of ssDNA. Similar results were obtained using trypsin as the source of protease, although the cleavage pattern was different (data not shown).
ATP Binding-The observation that UvrDD248N was susceptible to cleavage by chymotrypsin in the presence of ATP␥S led us to directly determine if the mutant protein was compromised in its ability to form a binary complex with ATP.
[ 3 H]ATP binding by UvrD and UvrDD248N was measured using a gel filtration assay (see "Experimental Procedures"). The results are shown in Table II. ATP binding by UvrDD248N was reduced to less than 15% that of wild-type helicase II as measured by this assay. This result is consistent with the results from the proteolytic cleavage protection experiments in which the presence of ATP failed to render UvrD248N resistant to chymotrypsin cleavage. Taken together, these data demonstrate that Asp 248 is required for stable binding of ATP to the protein in the absence of DNA.
ATPase and Helicase Activity-To further characterize the UvrDD248N mutant protein we measured the ssDNA-stimulated ATPase activity of the mutant protein and compared this activity with that of the wild-type protein. The ATP hydrolysis constants k cat and K m are shown in Table III. The k cat value of UvrDD248N was 5.0 s Ϫ1 , approximately 4% that of the wildtype protein. No significant change in the apparent K m for ATP was detected for the mutant protein compared with wild-type UvrD. These reactions are performed in the presence of excess ssDNA and thus must reflect the interaction of ATP with the protein in the ternary enzyme-DNA-ATP complex. The specificity constant (k cat /K m ) of UvrDD248N was reduced 19-fold  compared with wild-type UvrD, reflecting the decrease in k cat with little change in K m . Therefore, the replacement of Asp 248 with an asparagine negatively impacts the ATP hydrolysis reaction catalyzed by helicase II but does not significantly alter the interaction of ATP with the enzyme in the presence of ssDNA.
To determine the impact of the D248N mutation on helicase activity, we examined the unwinding activity of the mutant protein using both partial duplex and blunt duplex DNA substrates. Titrations of the UvrD and UvrDD248N proteins with partial duplex substrates are shown in Fig. 4. Wild-type helicase II, at a concentration of 0.68 nM (monomer), unwound 73% of the 20-bp partial duplex DNA substrate in a 10-min reaction (Fig. 4A). The UvrDD248N mutant protein achieved a comparable level of unwinding (72%) but at a considerably higher protein concentration (9.36 nM monomer). Wild-type helicase II unwound nearly 75% of the 92-bp partial duplex substrate at a concentration of 34 nM (monomer) (Fig. 4B). The D248N mutant required a concentration of 187 nM (monomer) to unwind approximately 70% of the 92-bp partial duplex. Thus the unwinding efficiency of the D248N mutant is moderately compromised on relatively short duplexes of 20 and 92 bp in length. However, the UvrDD248N mutant retained unwinding activity on short partial duplex substrates even at low concentrations of protein.
The unwinding reactions catalyzed by the wild-type and D248N mutant proteins were distinctly different as the length of the duplex region was increased to 343 bp (Fig. 4C). At 93.6 nM (monomer), UvrDD248N unwound only 2% of the 343-bp partial duplex, whereas wild-type enzyme unwound more than 75% of the partial duplex substrate at 90.5 nM (monomer). However, substantial levels of unwinding were attained using the UvrDD248N mutant at higher protein concentrations in a markedly biphasic titration curve. In contrast, the wild-type protein unwinds increasing amounts of the 343-mer at each incremental protein concentration. The discrepancy between the helicase activity of the wild-type and D248N mutant proteins is yet more dramatic on the 851-bp partial duplex DNA substrate (Fig. 4D). At a UvrDD248N concentration of 187 nM (monomer), 3% of the 851-mer is unwound. In contrast, the wild-type protein unwound more than 80% of the 851-bp substrate at 181 nM (monomer). The wild-type UvrD protein exhibited a typical protein concentration-dependent displacement of the 851-nucleotide fragment. The UvrDD248N mutant protein only achieved significant unwinding of the 851-bp partial duplex at the highest concentration of protein tested, again showing a pronounced biphasic protein titration in these unwinding reactions. Clearly, much higher concentrations of UvrDD248N were required to achieve unwinding of relatively long partial duplex substrates of 343 and 851 bp in length. However, at these high protein concentrations a significant fraction of the substrate was unwound.
Helicase II has also been shown to unwind blunt-ended duplex DNA substrates (Runyon et al., 1990;Runyon and Lohman, 1989). We tested the ability of the UvrDD248N mutant protein to unwind a 346-bp blunt duplex DNA substrate (Fig. 5). Wild-type helicase II, at a concentration of 136 nM (monomer), unwound about 50% of the 346-bp fragment in a 10-min reaction (Fig. 5A). Less than 0.5% of the 346-bp blunt duplex was unwound by the UvrDD248N mutant at a comparable concentration of protein (141 nM monomer). An increase in wild-type helicase II concentration to 543 nM (monomer) resulted in the unwinding of 95% of the 346-bp duplex. At a comparable mutant enzyme concentration (562 nM monomer) only 1.7% of the 346-bp blunt duplex was unwound. We also examined the blunt duplex unwinding reaction catalyzed by UvrD and UvrDD248N Reaction conditions for DNA-dependent ATP hydrolysis assays were as previously described . For k cat determinations, protein concentrations were 27.2 and 70.3 nM for UvrD and UvrDD248N, respectively. The production of [ 3 H]ADP from [ 3 H]ATP was measured in a 30-l reaction over a 5-min time course at 37°C. Aliquots (5 l) were taken at 1-min intervals, and the rate of ATP hydrolysis was determined from linear plots of ADP production versus time. In a typical reaction less than 20% of the ATP was hydrolyzed to minimize end product inhibition and substrate depletion effects. For K m determinations, protein concentrations were 2.71 and 33.7 nM for UvrD and UvrDD248N, respectively. Reaction mixtures (20 l) for K m determinations were incubated for 10 min at 37°C.  the UvrDD248N mutant protein at various time points during a 210-min incubation (Fig. 5B). At a concentration of 562 nM (monomer), UvrDD248N achieved a plateau with approximately 40% of the 346-bp blunt duplex unwound in 90 min. We conclude that at high enzyme concentrations UvrDD248N is capable of unwinding the 346-bp blunt duplex DNA substrate, albeit at a significantly lower rate than wild-type UvrD.
Dimerization Studies-Helicase II has been shown to selfassemble to form a dimer or higher order oligomer in solution . Upon binding DNA, the dimeric form of helicase II is stabilized. To test the ability of UvrDD248N to oligomerize, we examined protein dimer formation by treating the UvrDD248N mutant protein with glutaraldehyde and detecting cross-linked dimers on denaturing SDS-polyacrylamide gels as described . At a concentration of 2 M (monomer), both UvrDD248N and UvrD formed dimers and higher order oligomers in solution as previously reported for UvrD  (data not shown). In the presence of 2 mM ATP␥S and 4.2 mM (dT) 10 the extent of dimer formation was enhanced, as expected, for both proteins. These results demonstrate that UvrDD248N has the ability to form dimers and that these dimers are stabilized upon binding DNA and nucleotide, similar to results previously reported for wild-type helicase II.

Genetic Characterization of the uvrDD248N Allele
We have genetically characterized the uvrDD248N allele by assessing its ability to complement the loss of the wild-type protein in two DNA repair pathways. For these studies plasmids containing uvrD (pET9d-H2wt) or uvrDD248N (pET9d-H2D248N) were transformed into E. coli JH137 or JH137⌬uvrD. Immunoblot data have demonstrated that helicase II expression from the pET9d-H2wt plasmid (in the absence of induction) is slightly less than that produced from the chromosome of JH137 .
To assess the ability of UvrDD248N to function in UvrABCmediated nucleotide excision repair, we determined the relative UV sensitivity of strains expressing the mutant or wildtype proteins (Fig. 6). UvrDD248N, supplied on a plasmid, complemented the UV-sensitive phenotype of the uvrD deletion strain JH137⌬uvrD to nearly the same level as the wild-type uvrD allele. Thus, the UvrDD248N protein retains its ability to function in UvrABC-mediated nucleotide excision repair, although this protein is not as efficient as the wild-type protein.
The UvrDD248N mutant protein was also examined for its ability to complement the loss of helicase II in methyl-directed mismatch repair. The mutant allele was introduced into a uvrD deletion strain on a plasmid, and the spontaneous mutation frequency at the rpoB locus was measured and compared with relevant strains as shown in Table IV. The relative mutability values of JH137⌬uvrD, JH137⌬uvrD/pET9d-H2D248N, and JH137⌬uvrD/pET9d-H2wt were found to be 185, 16.7, and 0.94, respectively. Thus, uvrDD248N partially complemented JH137⌬uvrD in methyl-directed mismatch repair. The wildtype allele, expressed from the same plasmid, exhibited full complementation of the repair deficiency. The partial genetic complementation of UvrDD248N in mismatch repair suggests that the mutant protein can function in unwinding at least some of the hemimethylated duplexes containing a mismatch. The mutation frequencies of JH137 (with a wild-type copy of uvrD on the chromosome) transformed with pET9d-H2wt and pET9d-H2D248N were also measured and found to be equivalent to that of JH137. Thus, uvrDD248N is recessive to the wild-type allele in methyl-directed mismatch repair.

DISCUSSION
To explore the functional significance of motif III in superfamily 1 DNA helicases we have mutagenized a very highly conserved amino acid residue within this region of E. coli DNA helicase II (see Fig. 1). Motif III is separated by 16 -18 amino acids from motif II, which is known to be involved in nucleotide binding/hydrolysis, and is separated from motif IV, whose function is still unknown, by 24 -26 amino acid residues (Hodgman, 1988). Thus, in addition to amino acid sequence conservation, there is conservation of the spacing between motifs II and III FIG. 5. UvrDD248N catalyzes a slow unwinding reaction on a blunt duplex DNA substrate at high enzyme concentration. Helicase reactions using a 346-bp blunt duplex DNA substrate were as described under "Experimental Procedures." A, reactions (20 l) were initiated by the addition of the indicated amounts of UvrDD248N (G) or the wild type enzyme (Ⅺ) and incubated for 10 min at 37°C. B, the UvrDD248N reaction mixture was increased to 240 l, and a 20-l aliquot was removed at time zero. The reaction was initiated by the addition of UvrDD248N at a final concentration of 562 nM (monomer), and incubation was at 37°C. Aliquots were removed at the indicated times and analyzed by polyacrylamide gel electrophoresis as described under "Experimental Procedures." The data represent the average of at least three independent experiments.
There have been no biochemical studies to date regarding the function of motif III in superfamily 1 DNA helicases. However, mutation of the highly conserved glycine in motif III (see Fig. 1) of the herpes simplex virus UL5 protein to a serine resulted in the failure of the mutant protein to function in viral DNA replication (Zhu and Weller, 1992). Although the UL5 gene product is only one component of the three-protein helicaseprimase complex, this result demonstrates that motif III has a significant cellular function. To continue to define the functional significance of motif III we have mutated a highly conserved motif III residue in E. coli DNA helicase II, an enzyme that has been extensively characterized both biochemically and genetically. The specific mutant characterized here, UvrDD248N, lacks the negative charge on the conserved aspartic acid residue that was replaced with a neutral asparagine. Limited proteolysis studies with both trypsin and chymotrypsin have shown that this isosteric change does not grossly alter the structure of the UvrD protein.
The most significant defect we have noted in the UvrDD248N mutant protein is the inability of this protein to form stable binary complexes with either ATP or ssDNA. The failure of UvrDD248N to stably bind ssDNA was initially noted during purification of the mutant protein. Typically, helicase II is purified using a ssDNA-cellulose column. The UvrDD248N mutant failed to bind ssDNA-cellulose under the standard conditions used for purification. However, if the column was equilibrated with a buffer containing ATP, the protein bound with apparently normal affinity. Nitrocellulose filter binding assays and limited proteolysis protection experiments have confirmed that the UvrDD248N mutant does not stably bind DNA in the absence of ATP. However, the protein exhibits a normal interaction with ssDNA when ATP is present, as evidenced by a binding isotherm similar to that of the wild-type protein in nitrocellulose filter binding assays. Thus, a ternary complex consisting of enzyme-ATP-ssDNA can be formed and is enzymatically active as an ATPase and a DNA helicase. For this reason, motif III does not appear to be the ssDNA binding domain in DNA helicase II.
We also note that the mutant protein is not able to stably bind ATP in the absence of ssDNA as demonstrated using limited proteolysis protection experiments and gel filtration analysis of ATP-enzyme complexes. This result is surprising in view of the commonly held notion that motifs I and II of the DNA helicases are involved in ATP binding/hydrolysis. We suggest that the aspartic acid residue in motif III is involved in stabilizing the interaction between ATP and the enzyme and is not likely to be directly involved in ATP binding. Support for this interpretation derives from the fact that the K m for ATP in the ssDNA-stimulated ATPase reaction catalyzed by UvrDD248N is essentially the same as that of the wild-type enzyme. Moreover, there is evidence of a weak interaction with ATP in the gel filtration experiments (see Table II).
Taken together, the results presented here suggest that the aspartic acid residue in motif III of DNA helicase II is involved in stabilizing the interaction between ATP and enzyme in the absence of DNA and stabilizing the interaction between ssDNA and enzyme in the absence of ATP. Motif III is not likely to be the primary site of either ATP or ssDNA interaction with the protein, since a ternary complex involving the enzyme and both ligands can be formed and is active. We suggest that motif III may act as an interface between the DNA binding and ATP binding domains of this complex protein. It may help to stabilize the binding of either ligand to the protein while awaiting the binding of the second ligand. Binding of both ligands is a prerequisite to protein function as a helicase (i.e. both ATP and DNA must be bound for the helicase reaction to occur). It is also possible that mutations in motif III slightly alter the overall conformation of the protein. This motif may represent a specific fold responsible for bringing the ATP binding domain and the ssDNA binding domain into the correct juxtaposition to facilitate the ssDNA-stimulated ATPase reaction. These two possibilities are not mutually exclusive.
It is interesting, and perhaps unexpected, that this mutation in motif III has a significant effect on the k cat of the ssDNAstimulated ATPase reaction catalyzed by helicase II. If we assume that the reaction mechanism exhibited by helicase II is not ordered with respect to binding of either ligand, two possibilities for binding order can be envisioned (Fig. 7): (i) ATP binds followed by DNA (k 1 , k 2 ) or (ii) DNA binds followed by ATP (k 3 , k 4 ). Both pathways are significantly compromised in the UvrDD248N mutant with respect to the first step (k 1 and k 3 ). However, the second step in each pathway (k 2 and k 4 ) appears not to be seriously affected as evidenced by the wildtype K m for binding ATP in the presence of saturating ssDNA and the wild-type affinity for ssDNA in the presence of saturating ATP. 2 Thus, the reduced k cat for ATP hydrolysis seems to reflect an impairment of the actual hydrolysis event (k 5 ). This may be due to the fact that the ssDNA binding domain and the ATP binding domain are not precisely aligned for maximal catalytic efficiency. It is important to note that the specificity constant (k cat /K m ) is reduced 19-fold for this mutant as compared with reductions of Ͼ250-fold for mutations in motifs I and II, which directly impact the hydrolysis of ATP (Brosh and Matson, 1995;George et al., 1994).
Alternatively, if the measured k cat reflects the entire reaction pathway (i.e. k 1 , k 2 . . . k 4 plus the hydrolysis step, k 5 ), then the reduced stability of the binary complexes may be directly responsible for the reduced k cat . In this case, a reduced k 1 or k 3 or an increased k Ϫ1 and k Ϫ3 would substantially reduce the steady state concentration of each binary complex. If formation of a binary complex is the rate-limiting step, then overall conversion of ATP to ADP and P i would be reduced with no effect on the hydrolytic step (k 5 ) in the reaction pathway. At present, we cannot distinguish between these possibilities but suggest that the the D248N mutation may perturb the alignment between the ssDNA binding domain and the ATP binding do-2 R. M. Brosh and S. W. Matson, unpublished results. a Mutation frequency was determined by dividing the number of resistant colonies formed on selective agar by the total number of cells plated. Relative mutability values were obtained by dividing the mutation frequency of the cell strain in question by the frequency of the wild-type strain.
b Value previously determined in this lab . main, which in turn impairs the stability of the binary complexes. It is possible that both binary complex stability and catalytic efficiency are impaired in this mutant enzyme. The unwinding reaction catalyzed by the UvrDD248N mutant is also compromised. A decrease in helicase activity might be expected due to the decrease in k cat for ATP hydrolysis, since ATP hydrolysis is coupled to the unwinding reaction. However, it is not clear how tightly coupled these two reactions might be; therefore, it is possible that a decreased k cat for ATPase activity would result in only a small change in helicase activity. In the case of the UvrDD248N mutant, the helicase reaction is significantly reduced, and relatively high concentrations of mutant enzyme compared with wild-type helicase II are required to achieve substantial unwinding of partial duplex substrates. Moreover, the unwinding titration curves obtained using the mutant protein are markedly biphasic as compared with the relatively smooth hyperbolic curves obtained using the wildtype protein. The mutant enzyme exhibits a very low rate of unwinding at low protein concentrations and a rate of unwinding that approaches that exhibited by the wild-type protein at higher protein concentrations with each substrate tested. The transition point in this biphasic curve is at progressively higher protein concentrations as the length of partial duplex region increases. One explanation for this effect is that higher concentrations of UvrDD248N protein are required to coat the ssDNA generated during an unwinding reaction due to the reduced stability of the enzyme-DNA-ADP complex. Previous studies have shown that helicase II acts as its own helix destabilizing protein in an unwinding reaction by binding to the unwound strands (Runyon et al., 1990;Wessel et al., 1990). Both the UvrDD248N enzyme-ssDNA binary complex and the UvrDD248N enzyme-ssDNA-ADP ternary complex are significantly less stable than the corresponding complexes formed with the wild-type enzyme (see Table I). Thus, the probability of the UvrDD248N protein dissociating from partially unwound DNA is significantly increased. In this case, substantially more of the mutant enzyme may be required to unwind a long duplex region, since helicase II molecules constantly dissociate from the partially unwound substrate after they hydrolyze ATP. In support of this interpretation we have shown that the addition of SSB stimulated the unwinding reaction cata-lyzed by UvrDD248N. 3 Moreover, this interpretation is consistent with the notion that helicase II serves as its own helixdestabilizing protein during an unwinding reaction and underscores the importance of this property of helicase II.
We also note that the UvrDD248N mutant exhibits a reduced ability to unwind the 346-bp blunt duplex DNA substrate. Even at very high concentrations, the mutant enzyme performed poorly in a 10-min reaction. However, the mutant enzyme was able to catalyze significant unwinding of the blunt duplex substrate at high protein concentrations with long periods of incubation. The reduction in unwinding from a blunt duplex end may be the result of a defect in either initiation or propagation of the unwinding reaction. It is clear that the enzyme is compromised in propagation of the unwinding reaction, particularly on longer duplex regions. A defect in initiating an unwinding reaction from a blunt duplex end would further diminish overall unwinding of the substrate.  have provided evidence to suggest that the initiation of unwinding on blunt duplex DNA is a rate-limiting step. The substantial reduction in blunt duplex unwinding by UvrDD248N would suggest that the mutant enzyme is compromised in both the initiation and propagation steps. This would be consistent with the instability of the binary complex formed by the enzyme with DNA or ATP.
Despite the fact that the ssDNA-dependent ATPase and helicase reactions catalyzed by UvrDD248N are significantly reduced relative to the wild-type protein, this mutant is able to complement the absence of UvrDp in both excision repair and methyl-directed mismatch repair. Complementation provided by the UvrDD248N protein is at nearly wild-type levels in UvrABC-mediated excision repair of UV-induced damage. The fact that the UvrDD248N protein retained substantial activity with the 20-bp partial duplex substrate would indicate that the mutant enzyme should be capable of excising the 12-13-mer released in excision repair. A second requirement of UvrDp in this pathway is the turnover of UvrCp (Husain et al., 1985;Orren et al., 1991). The level of genetic complementation provided by the UvrDD248N mutant suggests that the mutant enzyme functions in this capacity.
Complementation of the helicase II deficit in methyl-directed mismatch repair is apparently not as robust as that observed for repair of UV-induced DNA damage, although it is difficult to directly compare results from the two different assays. It is possible that the complementation we observe is due to repair requiring short track length excision and resynthesis (i.e. the GATC is located close to the base pair mismatch), while repair requiring long track length excision and resynthesis is inefficient. This is consistent with the results of in vitro helicase assays indicating a more pronounced reduction in helicase activity on longer partial duplex substrates. We also note that the uvrDD248N allele is recessive to the wild-type allele in complementation of the defect in methyl-directed mismatch repair. Assuming that helicase II functions as a dimer and that the UvrDD248N protein can interact with the wild-type protein, this interaction must not dramatically interfere with the ability of the wild-type enzyme to function in the mismatch repair pathway. Alternatively, a population of wild-type helicase II dimers may exist, since expression of UvrDD248N using the pET9d expression vector is expected to be slightly less than expression of the wild-type protein from the chromosome .
Several studies that have begun to address the functional significance of motif III in members of a distantly related group of proteins (which include DNA and RNA helicases) designated 3 R. M. Brosh, unpublished results. superfamily 2 (Gorbalenya et al., 1988) are consistent with the results reported here. A motif III mutation in mammalian translation initiation factor (eukaryotic initiation factor 4A) eliminated the RNA-unwinding activity of the enzyme yet enhanced the ATPase activity 3-fold, suggesting that the ATPase and helicase activities of the mutant protein had become uncoupled (Pause and Sonenberg, 1992). A dominant negative mutation in motif III of the splicing factor PRP2 slightly reduced the RNA-stimulated ATPase activity of the mutant protein and resulted in the accumulation of stalled splicing complexes, leading the authors to propose that the dominant negative phenotype was due primarily to a defect in the putative RNA helicase activity of PRP2 protein (Plumpton et al., 1994). A mutation in motif III of E. coli RecG protein significantly reduced the ATPase activity of the mutant protein relative to wild-type protein and abolished the enzyme's ability to catalyze branch migration of Holliday junction intermediates (Sharples et al., 1994). Finally, mutations in motif III of herpesvirus protein UL9 decrease viral DNA replication (Martinez et al., 1992), suggesting that motif III of superfamily 2 is functionally significant.
We suggest that the aspartic acid residue in motif III of DNA helicase II, and perhaps in other helicases, is important in stabilizing the binary complex between protein and ssDNA or protein and ATP. Motif III is unlikely to be the primary binding site for either ligand. It is more likely that motif III serves as an interface between the ATP binding domain and the ssDNA binding domain of a DNA helicase and is important in maintaining the proper alignment between the two domains.