The DNA Helicase and Adenosine Triphosphatase Activities of Yeast Rad3 Protein Are Inhibited by DNA Damage A POTENTIAL MECHANISM FOR DAMAGE-SPECIFIC RECOGNITION*

Purified Rad3 protein from the yeast Saccharomyces cerevisiae is a single-stranded DNA-dependent ATPase and also acts as a DNA helicase on partially duplex DNA. In this study we show that the DNA helicase activity is inhibited when a partially duplex circular DNA substrate is exposed to ultraviolet (UV) radiation. Inhibition of DNA helicase activity is sensitive to the particular strand of the duplex region which carries the damage. Inhibition is retained if the single-stranded circle is irradiated prior to annealing to an unirradiated oligonucleotide, but not if a UV-irradi-ated oligonucleotide is annealed to unirradiated circu- lar single-stranded DNA. UV irradiation of single-stranded DNA or deoxyribonucleotide homopolymers also inhibits the ability of these polynucleotides to support the hydrolysis of ATP by Rad3 protein. UV radia- tion damage apparently blocks translocation of Rad3 protein and results in the formation of stable Rad3 protein-UV-irradiated DNA complexes. As a conse- quence, Rad3 protein remains sequestered on DNA, presumably at sites of base damage. The sensitivity of Rad3 protein to the presence of DNA damage on the strand along which it translocates provides a potential mechanism for damage recognition during nucleotide excision

11 To whom correspondence should be sent. 1988). A seventh gene (designated ERCC3") has been recently identified based on sequence homology with a gene (ERCC3) required for nucleotide excision repair in human cells (Weeda et al., 1990). This genetic complexity suggests that the specific recognition of base damage and/or the incision of DNA at such sites requires multiple proteins. However, at present the functional roles of the polypeptides encoded by the 7 genes identified above are unknown.
Purification and biochemical characterization of the RAD3 gene product has demonstrated that it is a single-stranded DNA-dependent ATPase and DNA helicase (Sung et al., 1987a(Sung et al., , 1987bHarosh et al., 1989). Based on the paradigm of nucleotide excision repair in the prokaryote Escherichia coli (Sancar and Sancar, 1988;Van Houten, 1990), it is tempting to consider that a DNA helicase might play a role in the excision of oligonucleotide fragments generated by damagespecific incision of DNA. Alternatively, a DNA helicase might be required for preincisional events and/or incision of DNA by specifically altering conformation a t or near sites of DNA damage, as has been suggested for the UvrAB complex of E. coli (Oh and Grossman, 1989;Seeley and Grossman, 1989;Selby and Sancar, 1990;Koo et al., 1991).
Previous studies demonstrated that the DNA helicase activity of Rad3 protein was partially inhibited when a circular DNA substrate containing a 30-base pair duplex region was exposed to ultraviolet (UV) radiation (Harosh et al., 1989). In the present studies we have explored the effect of UV radiation damage on the catalytic functions of Rad3 protein in greater detail. We show here that in addition to inhibition of DNA helicase activity, UV radiation of single-stranded DNA results in inhibition of the single-stranded DNA-dependent ATPase activity. This inhibition apparently results from the sequestration of Rad3 protein on the damaged polynucleotide, presumably at sites of base damage. Additionally, when partially duplex DNA carrying photoproducts exclusively on the oligonucleotide opposite to the strand to which Rad3 protein binds and translocates was utilized as a substrate, no inhibition of DNA helicase activity was detected. These observations suggest that Rad3 protein may play an important role in searching for and locating base damage in a strand-specific manner during nucleotide excision repair in yeast.
Purification of Rad3 Protein-The Rad3 overexpressing yeast strain BJ-YAR was grown and Rad3 protein was purified to homogeneity as described (Harosh et al., 1989) with the following modifications. The eluate from DEAE-Sepharose chromatography (fraction 11) was loaded directly onto a Blue Sepharose CL-GB column (10 x 2.5 cm) equilibrated with 0.28 M NaCl in buffer A (50 mM Tris-HC1, pH 7.5,5 mM EDTA, 10 mM Na2S205, 5 mM P-mercaptoethanol, 10% (v/v) glycerol). The column was washed with 3 bed volumes of 0.28 M NaCl in buffer A and then eluted with a linear gradient of NaCl (0.28-1.8 M) in 240 ml of buffer A. Fractions containing Rad3 protein were identified by immunoblotting (Naumovski, 1987), pooled, dialyzed against 3 changes of 50 mM NaCl in buffer A, and processed through DNA-cellulose and hydroxylapatite chromatography as described (Harosh et al., 1989). Rad3 protein was eluted from the hydroxylapatite column with 60 mM sodium phosphate and 0.1 M NaCl in buffer A. The purified enzyme migrated as a single band of 89 kDa on sodium dodecyl sulfate-polyacrylamide gels (Laemmli, 1970).
Preparation of Substrates for the DNA Helicase Assay-Partial duplex substrates for the DNA helicase assay consisted of M13mp18 single-stranded DNA annealed to either a 5' end-labeled 30-mer complementary oligonucleotide as described (Harosh et al., 1989), or to a ""P-labeled 206-mer complementary oligonucleotide produced by the single-stranded polymerase chain reaction (Bednarczuk et al., 1991;see below). For annealing, 0.5-1 pg of M13mp18 DNA and an equimolar amount of complementary oligonucleotide were mixed in 10 mM Tris-HC1, pH 7.0,l mM EDTA, 100 mM NaC1,lO mM MgC1, (50 pl final volume) and heated a t 98 "C for 10 min, followed by incubation at 72 "C (30-mer) or 82 "C (206-mer) for 20 min. Substrates were immediately purified by gel filtration through a Bio-Gel A-5m column equilibrated with 10 mM Tris-HC1, pH 7.0,l mM EDTA, 100 mM NaCl, at room temperature.
UV Irradiation of DNA-Aliquots (20 pl) of DNA in 10 mM Tris, pH 7.0, 1 mM EDTA, 100 mM NaCl were placed on ice in an open Petri dish and irradiated at a dose rate of 1 J.m-'.s-' using a germicidal lamp with a peak output at 254 nm.
Quantitation of Pyrimidine Dimers-The presence of pyrimidine dimers in the "'P-labeled strand of the partial duplex helicase substrate was demonstrated by using the pyrimidine dimer DNA-glycosylase/AP endonuclease from M. luteus. Reactions (10 pl) were performed according to the manufacturer's instructions and contained 50 mM KHPO,, pH 6.5, 40 mM NaC1, 1 mM EDTA, 1 ng of "Plabeled partial duplex DNA, 100 ng of unlabeled unirradiated Hind111 X DNA digest, and 500 units of enzyme. After incubation at 37 "C for 60 min, reaction products were resolved by denaturing polyacrylamide gel electrophoresis (Ogden and Adams, 1987) and visualized by autoradiography. Regions of the gel containing the full-length 206-mer fragment were removed and radioactivity was measured by counting Cerenkov radiations in a liquid scintillation counter. After exposure to a UV radiation dose of 3600 J/m', 56% of the DNA fragments were digested. The extent of degradation in control incubations containing unirradiated DNA was less than 4%. Assuming a Poisson distribution of pyrimidine dimers in the DNA, the calculated dimer density at saturation doses was 1 dimer/249 bases of single-stranded DNA.
ATPase and DNA Helicase Assay-Hydrolysis of ATP by Rad3 protein was measured as described (Harosh et al., 1989). DNA helicase activity was determined in reaction mixtures (20 pl) containing 30 mM potassium acetate buffer, pH 5.6, 5 mM MgCl,, 1 mM dithiothreitol, 70 pg/ml of bovine serum albumin, 1 mM ATP, and the indicated amounts of helicase substrate, competitor DNA, and Rad3 protein.
Reactions were stopped by the addition of 2 p1 of 40 mM EDTA, 1% sodium dodecyl sulfate, and 0.1% bromphenol blue in 40% (v/v) glycerol. "'P-Labeled oligonucleotides were resolved by electrophoresis on polyacrylamide gels and visualized by autoradiography as described (Harosh et al., 1989;Matson, 1989). The regions of the gels containing duplex DNA and displaced fragments were removed and assayed for Cerenkov radiations in a liquid scintillation counter. Helicase activity was expressed as the percentage of displaced fragments and all values were corrected for background by subtracting the release of fragments (typically <5%) obtained in control incubations without ATP.
DNA Binding Assay-The assay used to determine binding of Rad3 protein to single-stranded DNA was adapted from Matson and Richardson (1985). Binding reactions (20 p l ) contained 30 mM potassium acetate buffer, pH 5.6, 5 mM MgCl,, 1 mM dithiothreitol, 1 mM ATP, 1.5 nM (in nucleotides) linear 5' end-labeled M13 singlestranded DNA and varying amounts of Rad3 protein. After incubation for 10 min at 30 "C, reaction mixtures were diluted to 1.0 ml by the addition of ice-cold 30 mM potassium acetate buffer, pH 5.6, 5 mM MgCl,, 1 mM dithiothreitol, and passed through nitrocellulose filters. The filters were washed with 1.0 ml of buffer, dried, and assayed for Cerenkov radiations in a liquid scintillation counter. Protein binding was expressed as the percentage of input DNA. Background levels of DNA bound to the filters in the absence of added protein (typically <8%) were subtracted from all values. Nitrocellulose filters were pretreated by boiling in distilled water for 20 min, soaked in 0.2 M NaOH for 20 min, washed with several changes of distilled water, and stored at 4 "C in 10 mM Tris-HC1, pH 8.0, 1 mM EDTA.
Preparation of Linear Single-stranded M I 3 DNA-M13mpl8 DNA (10 pg) was incubated with 120 ng of pancreatic DNase I in 50 mM Tris-HC1, pH 7.5, 5 mM MgCI, for 1 min at 23 "C. The reaction was terminated by the addition of 10 mM EDTA. The linearized DNA was dephosphorylated by treatment with bacterial alkaline phosphatase, 5' end-labeled with T4 polynucleotide kinase, and purified as described (Runyon and Lohman, 1989).
Other Methods-DNA concentrations were determined by measuring absorbance at 260 nm and are expressed as nucleotide equivalents. The Coomassie Blue dye binding method of Bradford (1976) was used for determination of protein concentrations.

Inhibition of Rad3 D N A Helicase Activity by UV
Radiation of the Substrate-The DNA helicase activity of Rad3 protein was measured by the displacement of a "'P-labeled oligonucleotide fragment annealed to complementary single-stranded M13 DNA. A previous study demonstrated partial inhibition of the Rad3 DNA helicase activity when such a substrate containing a duplex region of 30 base pairs was exposed to UV radiation (Harosh et al., 1989). In the present studies we have extended these findings using a substrate with a duplex region of 206 base pairs. Quantitative analysis of DNA helicase activity demonstrated that Rad3 protein unwound 42% of the partial duplex substrate in the control reaction. Essentially no displacement of radiolabeled oligonucleotide was detected when the substrate was exposed to 3600 J/m2 of UV radiation (Fig. 1A). UV radiation-induced cross-links were not responsible for the observed inhibition since heat denaturation of the damaged substrate released >90% of the labeled oligonucleotide (data not shown). Inhibition of DNA helicase activity was UV radiation-dependent at doses between -120 and -900 J/mZ and plateaued at -1800 J/mZ (Fig.  1B).
Strand Specificity of Inhibition of Rad3 DNA Helicase Actiuity-The Rad3 DNA helicase translocates on the single- strand to which it is bound exclusively with a 5'+3' polarity. We therefore asked whether during the unwinding of partially duplex DNA, inhibition of the helicase activity was sensitive to base damage on the unbound strand, i.e. the complementary oligonucleotide. We constructed a partial duplex substrate containing photoproducts only in the DNA fragment to be displaced ( Fig. 2; substrate A). This was achieved by UV irradiating the complementary 206-mer at a dose of 3600 J/ m' prior to its annealing to M13 single-stranded DNA. Using the pyrimidine dimer-specific DNA glycosylase/AP endonuclease from M. luteus (Grafstrom et al., 1982), we determined that 56% of the partial duplex DNA molecules contained at least one pyrimidine dimer in the annealed oligonucleotide. Control substrates were constructed by annealing either unirradiated ( Fig. 2; substrate B ) or UV irradiated (3600 J/m'; substrate C ) M13 single-stranded DNA to the undamaged 206-mer. As shown in Fig. 2, in the presence of Rad3 protein the kinetics and extent of oligonucleotide displacement were indistinguishable with substrates A or B. As expected, the presence of UV radiation damage in substrate C inhibited Rad3 helicase activity to the same extent as that observed when the partial duplex was irradiated after annealing (Fig.  lA). These results suggest that Rad3 DNA helicase activity is uniquely sensitive to UV radiation damage in the DNA strand on which it is translocating.
age to Single-stranded DNA-The unwinding of duplex regions of DNA by Rad3 protein is accompanied by the hydrolysis of ATP to yield ADP and Pi (Sung et al., 1987a(Sung et al., , 1987bHarosh et al., 1989). The ATPase activity is strictly dependent on the presence of single-stranded DNA or deoxyribonucleotide homopolymers (Sung et al., 1987a(Sung et al., , 1987bHarosh et al., 1989). UV irradiation of single-stranded M13 DNA resulted in inhibition of the Rad3 ATPase activity. After incubation for 30 min, ATP hydrolysis in the presence of M13 singlestranded DNA UV irradiated at 3600 J/m' was 51% of that measured in the presence of an equal amount of unirradiated DNA (Fig. 3A).
To explore this inhibition in greater detail we utilized a variety of synthetic homodeoxyribopolymers as substrates. Poly(dT) supported ATP hydrolysis more efficiently than did single-stranded M13 DNA (data not shown). In contrast, poly(dA), poly(dC), and poly(dG) were much less effective activators of the ATPase activity of Rad3 protein (Fig. 3B). When poly(dT) was exposed to increasing levels of UV radiation a dose-dependent inhibition of ATP hydrolysis was observed (Fig. 4). The extent of the inhibition was greater with poly(dT) than with equimolar amounts of M13 singlestranded DNA at all doses tested and was linear with respect to the UV radiation dose between -120 and -600 J/m' (Fig.  4). On the other hand, UV irradiation of poly(dA) had very little effect on enzyme activity (Fig. 4). These results indicate that inhibition of Rad3 ATPase activity is a function of the number of lesions in the DNA, and are consistent with several studies demonstrating that pyrimidine bases are quantitatively major targets for photochemical alterations in DNA, whereas purines are altered to a lesser extent .
Inhibition of the Rad3 ATPase activity by UV-irradiated polymers was also examined in a series of competition experiments. Rad3 protein was incubated with unirradiated M13 single-stranded DNA in reaction mixtures supplemented with either unirradiated or UV irradiated (3600 J/m2) poly(dT). Addition of the unirradiated homopolymer stimulated the ATPase activity of Rad3 protein and ATP hydrolysis reached

FIG. 4. Inhibition of Rad3
ATPase activity by DNA damage as a function of UV dose. Rad3 protein (25 ng) was activated by poly(dA) (circles), M13 single-stranded DNA (squares), or poly(dT) (triangles) irradiated at the indicated UV doses. Final concentrations of the polynucleotides in the reaction were 7.6 p~ for poly(dT) and M13 DNA, and 76 PM for poly(dA). The incubation was stopped after 30 min and ATP hydrolysis was measured as described under "Experimental Procedures." ATPase activity (average of duplicate determinations) is expressed as a percentage of the unirradiated control. a maximum level of 352 pmol during a 30-min incubation (Fig. 5A ).
The same level of ATPase activity was obtained in a control reaction in the presence of unirradiated poly(dT) alone (Fig.  5B). Addition of increasing amounts of UV-irradiated poly(dT) to reactions with unirradiated single-stranded DNA resulted in reduced hydrolysis of ATP (Fig. 5A). At a poly(dT):single-stranded DNA molar ration of 1:1, the level of residual ATP hydrolysis (66 pmol) was close to that observed during incubation with UV-irradiated poly(dT) alone (40 pmol) (Fig. 5, A and B).
These results indicate that the inhibition of ATPase activity is not due to a reduced binding affinity of Rad3 protein for the UV-irradiated polymer relative to the unirradiated polymer or single-stranded DNA. On the contrary, they suggest that the protein might bind preferentially to UV-irradiated poly(dT), resulting in enzyme inhibition due to sequestration of Rad3 protein at or near sites of DNA damage.
Binding of Rad3 Protein to Single-stranded DNA-In order to directly examine the binding of Rad3 protein to UVirradiated single-stranded DNA, we incubated Rad3 protein with linear 5' end-labeled M13 DNA and analyzed the formation of protein-DNA complexes by retention of DNA on nitrocellulose filters. Addition of Rad3 protein to unirradiated single-stranded DNA led to maximal retention of -80% of the input DNA on nitrocellulose filters (Fig. 6). Half-maximal retention of input DNA was obtained at a molar ratio of Rad3 protein:Ml3 DNA of 58:l. This requirement for excess protein may reflect a low percentage of active enzyme molecules or, more likely, a low efficiency of binding of DNA-protein complexes to nitrocellulose filters (Matson and Richardson, 1985). Equilibrium binding was achieved after 10 min of incubation, since longer incubations did not significantly increase the fraction of bound DNA. ATP was not required for the formation of Rad3 protein-DNA complexes (data not shown). However, the triphosphate was routinely included in all reactions.  Binding of Rad3 protein to DNA was unaffected by the presence of UV radiation-induced damage (Fig. 6). Even at the highest UV dose used (3600 J/m2) we observed the same amount of DNA retention on filters. Concomitant measurement of ATP hydrolysis confirmed that UV irradiation of the linear substrate used in these experiments produced the same inhibitory effect on Rad3 ATPase activity as the circular DNA used in the experiments described above. These results provide direct confirmation that inhibition of the ATPase activity of Rad3 protein is not caused by reduced binding of protein to the UV-irradiated DNA substrate.

Dissociation Kinetics of Rad3 Protein-DNA Compkxes-
When challenged with an excess of cold competing singlestranded M13 DNA, Rad3 protein rapidly dissociated from radiolabeled M13 DNA (Fig. 7). After 30 min of incubation, 8.0% of the input DNA (corresponding to 11.3% of the initial amount of Rad3-DNA complexes) was retained on nitrocellulose filters. The majority of the complexes dissociated during the first 10 min after the addition of competing DNA (Fig.   7), yielding a calculated half-life of the complex of 3.9 min. When competitor DNA was added to incubations of Rad3 protein with UV-irradiated M13 DNA, Rad3 protein initially dissociated from the radiolabeled irradiated DNA at a similar rate (Fig. 7). However, 2 min after the addition of the competitor, DNA turnover of the protein was significantly reduced and a considerable fraction of the Rad3-UV-irradiated DNA complexes remained stable (Fig. 7). During the slower dissociation phase the half-life for dissociation of Rad3-UV-irradiated DNA complexes was calculated at 84.0 min. These results indicate that Rad3 protein remains tightly bound to the damaged substrate and, together with the observed inhibition of DNA helicase activity (Fig. l), suggest that sites of UV radiation damage block the translocation of Rad3 protein along single-stranded DNA. Sequestration of Rad3 Protein on UV-damaged Singkstranded DNA or Poly(dT)-To determine the fraction of Rad3 protein sequestered at sites of UV radiation damage in single-stranded DNA, we incubated purified Rad3 protein with varying amounts of unirradiated or UV-irradiated M13 single-stranded DNA for 10 min and then added a substrate for DNA helicase activity, consisting of circular singlestranded M13 DNA annealed to a complementary 5' endlabeled 30-mer primer. After incubation for a further 30 min, the reaction was stopped and helicase activity was measured. We concurrently determined that under standard assay conditions in the absence of a competing substrate, DNA helicase activity was linear with respect to Rad3 protein concentration between 5 and 100 ng of protein. This helicase competition assay therefore provided a convenient method for quantitating the fraction of active Rad3 protein that remained sequestered on the UV-irradiated single-stranded DNA.
A representative result is shown in Fig. 8A. As expected, U I lane: 1 2 3 4 5 6 FIG. 8. Sequestration of Rad3 protein on UV-irradiated M13 single-stranded DNA or UV-irradiated poly(dT). A, Rad3 protein (50 ng) was incubated for 10 min with the indicated concentrations of competitor M13mp19 single-stranded DNA. The M13 DNA was either unirradiated (-) or irradiated (+) at 3600 J/m2. DNA helicase substrate was then added to the reactions at a concentration of 1 p M (in nucleotides) and after a further 30 min of incubation samples were analyzed by polyacrylamide gel electrophoresis (see "Experimental Procedures" for details). The positions of partially duplex DNA and of the displaced oligonucleotide are indicated by the arrows. Lane 1, control reaction incubated in the absence of ATP; lane 2, DNA helicase activity obtained in the absence of competitor DNA (47.7% of primers displaced). B, Rad3 protein (50 ng) was incubated with unirradiated (-) or UV irradiated (+) poly(dT) or poly(dA), both at a concentration of 4 p~. After 10 min the reaction was supplemented with the DNA helicase substrate (0.4 pM) and samples were incubated for a further 30 min. Lane I , control; lane 2, helicase activity in the absence of competitor (67.7% primers displaced).
helicase activity was progressively reduced as a function of added competitor single-stranded DNA. At all concentrations tested, UV-irradiated (3600 J/m2) M13 DNA was a stronger competitor than the unirradiated DNA. For example, when 4 PM M13 DNA was used as the competitor (Fig. 8A, lane 7) helicase activity was 70.2% of that measured in its absence (Fig. &I, lune 2). On the other hand, 4 PM UV-irradiated DNA reduced helicase activity to 30.6% of the control reaction ( Fig. 8A, lune 8). Hence, an additional 39.6% of the active Rad3 enzyme molecules apparently remained sequestered on the UV-irradiated competitor DNA, and were not available for unwinding of the DNA helicase substrate. Assuming that all enzyme molecules were active, we calculated that 249 k 9.3 fmol of Rad3 protein were sequestered on UV-irradiated M13 DNA in reactions containing 4 p~ UV-irradiated M13 competitor DNA. This translates to an average of 1 molecule of Rad3 protein sequestered per 321.3 (f23.2) nucleotides of UV-irradiated M13 DNA.
This phenomenon was even more striking when poly(dT) was used for competition. Calculations from the data shown in Fig. 8B indicate that on the average 1 enzyme molecule of Rad3 protein was sequestered for every 151.9 (f17.9) nucleotides. This effect was not observed in the presence of UVirradiated poly(dA) (Fig. 8B), presumably reflecting the lower number of photochemical alterations that can be induced by UV irradiation of purines. The extent of the sequestration of Rad3 protein was essentially linear with respect to the UV radiation dose in the range 120-900 J/m2 and plateaued around 1800 J/m2 (Fig. 9).

DISCUSSION
Rad3 protein of 5' . cereuisiue is a DNA-dependent ATPase/ DNA helicase (Sung et al., 1987a(Sung et al., , 1987bHarosh et al., 1989). The ATPase activity utilizes exclusively single-stranded DNA; essentially no hydrolysis of ATP is detected in the presence of duplex DNA (Sung et al., 1987a, 198713;Harosh et al., 1989). The requirement for single-stranded DNA for nucleoside and deoxynucleoside triphosphate hydrolysis is central to the mechanism of unwinding of duplex DNA by all known DNA helicases (Matson and Kaiser-Rogers, 1990). It is generally assumed that the energy released during the hydrolysis of triphosphates is utilized in the helicase reaction, although little is known about the detailed mechanism of DNA unwinding (Matson and Kaiser-Rogers, 1990).
Studies on bacteriophage T7 G4 protein, which like Rad3  protein is a DNA helicase with a strict 5'+3' polarity with respect to the single strand to which it is bound, suggest that the hydrolysis of nucleoside triphosphates facilitates the processive unidirectional translocation of the protein along single-stranded DNA (Matson and Richardson, 1983;. Similar conclusions were reached by Brown and Romano (1989), who hypothesized that hydrolysis of nucleoside triphosphate by phage T7 G4 protein results in a conformational change in the protein when bound to single-stranded DNA and that this change is required for its translocation.
We previously reported that the displacement of a 30-mer complementary oligonucleotide from M13 single-stranded circular DNA by Rad3 protein was significantly reduced if the substrate was previously exposed to UV radiation (Harosh et al., 1989). Similar results were reported for the DNA helicase activity of E. coli UvrA/UvrB protein complexes (Oh andGrossman, 1987, 1989). In the present experiments we observed more complete inhibition of Rad3 DNA helicase activity with a UV-irradiated substrate containing a longer partially duplex region. The longer length of the oligonucleotide in this substrate is expected to increase the probability of generating photoproducts in the duplex region and this might explain the quantitative differences between the two studies.
A novel feature of the present experiments is the observation that inhibition of the Rad3 DNA helicase activity by UV radiation damage is strand specific and is effected only in the presence of damage on the strand on which the protein is translocating. These results are in contrast to those previously reported with the DNA helicase activity of the E. coli UvrAB protein, in which it was observed that irradiation of either the entire partially duplex DNA substrate or just the complementary oligonucleotide prior to annealing resulted in partial inhibition of helicase activity (Oh and Grossman, 1987). However, it should be noted that the biochemical properties of the yeast Rad3 and the E. coli UvrAB helicases are not identical. In particular, the E. coli UvrAB complex only unwinds short DNA duplex regions and acts in a distributive mode (Oh and Grossman, 1989).
The observation that UV radiation of single-stranded DNA or poly(dT) inhibits the hydrolysis of ATP by Rad3 protein is consistent with the inhibition of DNA helicase activity. A similar phenomenon was recently reported by Brown and Romano (1989), who observed inhibition of the dTTPase activity of phage T7 G4 protein in the presence of singlestranded DNA containing benzo[a]pyrene adducts. These investigators concluded that T7 G4 protein is sequestered at benzo[a]pyrene adducts. Our results support such a phenomenon. We demonstrated that in the presence of unirradiated single-stranded DNA, Rad3 protein forms short-lived DNAprotein complexes with a half-life of 3.9 min. However, in the presence of UV-irradiated DNA a significant fraction of Rad3-DNA complexes are considerably more stable, with a half-life of 84 min. Additionally, UV-irradiated single-stranded DNA competes more efficiently for DNA helicase activity than does unirradiated single-stranded DNA, suggesting that Rad3 protein remains sequestered on UV-irradiated DNA, presumably at sites of base damage.
The mechanism by which polynucleotide damage inhibits the nucleoside triphosphatase activity of DNA helicases such as the Rad3 and phage T7 G4 proteins is not clear. We have observed that single-stranded RNA is totally ineffective as a polynucleotide activator for ATP hydrolysis by Rad3 protein.' This observation, together with the observation that different deoxyribonucleotide homopolymers support ATP hydrolysis H. Naegeli, L. Bardwell, and E. C. Friedberg, manuscript in preparation.
by Rad3 protein with different efficiencies, suggests that the conformational change(s) in the protein that are required for ATP hydrolysis are exquisitely sensitive to the chemistry of the polynucleotide. Hence, damage to single-stranded DNA might inhibit the hydrolysis of nucleoside triphosphates, and RNA may be "read" by Rad3 protein as a form of damaged DNA. Additionally, if, as has been suggested by Brown and Romano (1989) for phage T7 G4 protein, the power stroke for translocation along single-stranded DNA is actually the exchange of nucleoside triphosphate for the diphosphate, continued translocation may become rate-limiting for adopting a conformation necessary for hydrolysis of the triphosphate.
Regardless of the precise mechanism(s) whereby DNA damage inhibits nucleotide hydrolysis and translocation of DNA helicases along single-stranded regions of DNA, this phenomenon potentially offers a highly specific mechanism for locating sites of base damage in DNA. Hence, Rad3 protein may play a fundamental role in damage-specific recognition during nucleotide excision repair. Following arrested translocation of Rad3 protein a t sites of damage, other proteins required for DNA incision may bind to Rad3-DNA complexes, eventually generating a conformational state required for endonucleolytic cleavage of DNA. Two attractive features of this model immediately come to mind. One is that the model readily accommodates the strand specificity associated with the excision of damaged nucleotides. Second, the model accommodates the observation that nucleotide excision repair operates on a wide variety of chemically distinct types of base damage. Indeed, recent studies in E. coli suggest that lesions as diverse as pyrimidine dimers and abasic sites are recognized by the UvrABC endonuclease (Lin and Sancar, 1989;Van Houten, 1990). The strand-specific effect of other types of base damage in DNA on the Rad3 DNA helicase and ATPase activities is currently under investigation.