Synergistic Interaction between Yeast Nucleotide Excision Repair Factors NEF2 and NEF4 in the Binding of Ultraviolet-damaged DNA*

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 274, No. 34, Issue of August 20, pp. 24257 24262, 1999 1999 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Synergistic Interaction between Yeast Nucleotide Excision Repair Factors NEF2 and NEF4 in the Binding of Ultraviolet-damaged DNA* (Received for publication, April 14, 1999, and in revised form, June 2, 1999) Sami N. Guzder, Patrick Sung, Louise Prakash, and Satya Prakash From the Sealy Center for Molecular Science, University of Texas Medical Branch, Galveston, Texas 77555-1061 and the Department of Molecular Medicine/Institute of Biotechnology, University of Texas Health Science Center at San Antonio, San Antonio, Texas 78245 Saccharomyces cerevisiae RAD4, RAD7, RAD16, and RAD23 genes function in the nucleotide excision repair (NER) of ultraviolet light (UV)-damaged DNA. Previous biochemical studies have shown that the Rad4 and Rad23 proteins are associated in a stoichiometric complex named NEF2, and the Rad7 and Rad16 proteins form another stoichiometric complex named NEF4. While NEF2 is indispensable for the incision of UV-damaged DNA in the in vitro reconstituted system, NEF4 stimulates the incision reaction. Both NEF2 and NEF4 bind UV-damaged DNA, which raises the intriguing possibility that these two complexes cooperate to achieve the high degree of specificity for DNA damage demarcation required for nucleotide excision repair in vivo. Consistent with this hypothesis, we find that NEF2 and NEF4 bind in a synergistic fashion to UV-damaged DNA in a reaction that is dependent on ATP. We also purify the Rad7 protein and show that it binds DNA but has no preference for UV-damaged DNA. Rad7 physically interacts with NEF2, suggesting a role for Rad7 in linking NEF2 with NEF4. In eukaryotes, nucleotide excision repair (NER) 1 of ultraviolet (UV)-damaged DNA is a highly intricate process which requires a large number of proteins. A defect in NER in yeast results in an extreme sensitivity to UV light and a hypermutational phenotype, and in humans, defective NER is the underlying cause of the skin cancer-prone syndrome xeroderma pigmentosum, observations that underscore the importance of this repair system in neutralizing the cytotoxicity and genotoxicity of UV light (1). During NER, the DNA lesion is first bound by damage recognition proteins, followed by the recruitment of additional NER factors and DNA unwinding by the two DNA helicases Rad3 and Rad25, present in TFIIH, to create a DNA bubble. Dual incision of the damage-containing strand in the unwound DNA, on the 5 -side by the Rad1-Rad10 nuclease and on the 3 -side by Rad2, results in the release of the lesion in the form * This work was supported by Grant CA41261 from the NCI, National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. To whom correspondence should be addressed: Sealy Center for Molecular Science, University of Texas Medical Branch, 6.104 Medical Research Bldg., 11th and Mechanic Sts., Galveston, TX 77555-1061. Tel.: 409-747-8602; Fax: 409-747-8608; E-mail: sprakash@scms.utmb. edu. 1 The abbreviations used are: NER, nucleotide excision repair; RPA, replication protein A; Q, ubiquinone; NTA, nitrilotriacetic acid; GST, glutathione S-transferase; ATP S, adenosine 5 -O-(3-thiotriphosphate). This paper is available on line at http://www.jbc.org 24257 of an oligonucleotide 30 nucleotides in length (for a discussion, see Ref. 2). A conserved set of proteins performs the same function in humans (3 5). Biochemical fractionation of Saccharomyces cerevisiae extract has revealed distinct subassemblies of the NER proteins, termed nucleotide excision repair factors or NEFs. Thus, Rad14 associates with the Rad1-Rad10 endonuclease to form NEF1 (6), Rad4 and Rad23 combine to form NEF2 (7), and Rad2 associates with the six subunit RNA polymerase II transcription factor TFIIH to form NEF3 (2). The combination of NEF1, NEF2, NEF3, and the heterotrimeric ssdna binding factor replication protein A (RPA) is sufficient to mediate dual incision of UV-damaged DNA in an in vitro reconstituted system (7), indicating that the basal NER machinery consists of these factors and also lends credence to the suggestion that NER is mediated by the stepwise incorporation of the aforementioned NEFs and RPA at the damage site (2). At the genomic level, another level of complexity exists in the repair of transcribed versus nontranscribed DNA. The repair of the transcribed strand requires the RAD26 gene, a homolog of the human CSB gene (8), whereas the repair of the nontranscribed strand requires the RAD7 and RAD16 genes (9, 10). Deletion of RAD7 or RAD16 causes an intermediate level of UV sensitivity, and the UV sensitivity of the rad7 rad16 double mutant is the same as that of the rad7 and rad16 single mutants. The Rad7 and Rad16 proteins exist together in a stable complex termed NEF4, and the purified Rad7-Rad16 heterodimer binds UV-damaged DNA in an ATP-dependent manner (11). NEF4 also has a DNA-dependent ATPase activity, and UV irradiation of DNA results in a marked inhibition of ATP hydrolysis (12). These observations have suggested a model in which NEF4 utilizes the free energy from ATP hydrolysis to translocate on DNA. Because of the attenuation of the ATPase activity, NEF4 stops translocating at the damage site and becomes stably bound to the damage. In this scenario, NEF4 would be the first protein complex to arrive at the damage site and would serve as the nucleation site for the subsequent assembly of the other repair factors (11, 12). Consistent with this hypothesis, the addition of NEF4 to the reconstituted NER system results in a marked stimulation of the proficiency of the incision reaction (11). In addition to NEF4, the Rad4-Rad23 complex (NEF2) also recognizes UV damage (13, 14). However, by contrast to the damage binding activity of NEF4, which is strongly dependent upon ATP, NEF2 damage binding shows no such dependence on a nucleotide co-factor (13). To begin delineating the intricacy of the damage recognition step in NER, we examine here whether the combination of NEF2 and NEF4 is more adept at binding UV-damaged DNA than either protein complex alone. Our results indicate a synergistic enhancement of damage binding by NEF2 and NEF4. We also purify the Rad7 protein

24258 Damage Binding by NEF2 and NEF4 and show that it binds DNA but has no specificity for UVdamaged DNA. Furthermore, we provide evidence for the physical interaction of Rad7 with NEF2, which suggests a specific function for Rad7 as a bridging factor between NEF2 and NEF4. MATERIALS AND METHODS Antibodies The anti-rad7, anti-rad4, anti-rad16, and anti-rad23 antibodies used in this study were all affinity purified from rabbit antisera raised against portions of the respective proteins expressed in and purified from Escherichia coli as described (11, 13, 15). Plasmids The construction of the Rad7 overexpression plasmid pr7.8 has been described previously (11). For overexpression of the glutathione S-transferase (GST)-Rad7 fusion protein, a 2.3-kilobase XbaI-EcoRI fragment containing the entire Rad7 coding sequence was cloned in plasmid pgex-3x to yield the GST-Rad7 expression plasmid pr7.4. Purification of Rad7 Protein The protease-deficient yeast strain LY2 harboring plasmid pr7.8 was grown overnight to the stationary phase in complete synthetic medium lacking leucine. The overnight starter culture was diluted ten times with fresh medium that also contained 2% galactose and was then incubated at 30 C for 10 h before the cells were harvested by centrifugation. We obtained 300 g of yeast paste from 80 liters of culture. Cell paste (300 g) of yeast strain LY2 containing pr7.8 was resuspended in cell breakage buffer containing protease inhibitors and disrupted by the French press. After clarification of the crude cell lysate by centrifugation (100,000 g, 90 min), the Rad7 protein was quantitatively precipitated by the addition of 0.22 g of ammonium sulfate per ml of lysate. The ammonium sulfate pellet was collected by centrifugation, and proteins were redissolved in K buffer (20 mm KH 2 PO 4, ph 7.5, 10% glycerol, 0.5 mm EDTA, 1 mm dithiothreitol) to give Fraction I with an ionic strength equivalent to 100 mm KCl and then loaded onto a Q-Sepharose column (2.5 5 cm; 25 ml). The Q-Sepharose flow-through (Fraction II) was applied onto a SP-Sepharose column (2.5 5 cm; 25 ml), which was developed with a 250-ml gradient of 100 500 mm KCl in K buffer. Fractions containing the peak of Rad7 protein (Fraction III) eluting at an ionic strength of 220 mm KCl were pooled, loaded onto a MacroHAP hydroxyapatite column (Bio-Rad; 2 ml), and equilibrated in K buffer with 100 mm KCl, which was developed with a 40-ml gradient of 20 300 mm potassium phosphate in the same buffer. Fractions containing Rad7 (Fraction IV; 7 ml) eluting at an ionic strength equivalent to 150 mm phosphate were dialyzed against 1 liter of buffer A (20 mm Tris-Cl, ph 7.0, 10% glycerol, 0.5 mm EDTA, 0.01% Nonidet P-40, 250 mm KOAc, 1 mm 2-mercaptoethanol) for5hat4 C.Thedialysate was mixed gently with 0.3 ml of Ni-NTAagarose matrix (Qiagen) for 3hat4 C. Thematrix was washed successively with buffer containing 10, 20, 30, 40, and 100 mm imidazole. Most ( 85%) of the Rad7 protein was recovered in the 100 mm imidazole eluate (Fraction V), which was loaded onto a Mono-S column (HR 5/5) equilibrated in buffer A containing 100 mm KCl. The Mono-S column was developed with a 12-ml gradient of 200 600 mm KCl, collecting 0.5-ml fractions. The fractions containing nearly homogeneous Rad7 protein were pooled (Fraction VI) and concentrated to a small volume using a Centricon 30 concentrator and stored in small aliquots at 70 C. Purification of Rad23 Protein To facilitate purification of Rad23 protein from yeast cells, the RAD23 gene was cloned downstream of the highly expressed yeast ADC1 promoter to yield the multicopy plasmid pjw112 (2 m, ADC1-RAD23). Rad23 protein overexpressed in the protease-deficient yeast strain LP2749 9B was purified to homogeneity as described previously (16). Purification of NEF2 The NEF2 complex, comprised of Rad4 and Rad23 proteins, was purified to homogeneity from yeast cells overexpressing the Rad4 protein alone because Rad23 protein is present in considerable molar excess over Rad4 protein. During the purification procedure, the Rad4 and Rad23 proteins were monitored by Western analysis using affinity purified antibodies as described previously (13). Other Protein Factors The Rad14 protein, Rad1-Rad10 complex, TFIIH, and RPA were purified as described (7). DNA Mobility Shift Assay A 130-base pair DNA fragment containing a stretch of thymine residues (17) was labeled at the 5 -end with 32 P and incubated with the indicated amounts of Rad7, NEF2, and NEF4 in 10 l of reaction buffer (30 mm potassium HEPES, ph 7.5, 30 mm KCl, 5mM MgCl 2,1mM dithiothreitol, 100 g/ml bovine serum albumin) and containing either 50 ng or the indicated amounts of HaeIII linearized X174 double-stranded DNA as competitor DNA. After 10 min at 30 C, samples were mixed with 2 l of gel loading buffer (0.1 M Tris-HCl, ph 7.0, 50% glycerol, and 0.05% Orange G) and electrophoresed in a 4.5% polyacrylamide gel in TAE buffer (40 mm Tris acetate, ph 7.4, 1 mm EDTA) at 4 C. The gels were dried and subjected to autoradiography using Kodak MR films, and autoradiograms were quantitated by a Bio-Rad GS670 densitometer. NER Reaction The NER reaction was carried out as described (7). Briefly, 75 ng of NEF2, 50 ng of RPA, 65 ng of NEF1, 65 ng of NEF3, and 60 ng of NEF4 or Rad7, where indicated, were incubated in reaction buffer (45 mm potassium HEPES, ph 7.9, 8 mm MgCl 2, 120 g/ml bovine serum albumin, 1.5 mm dithiothreitol, 2 mm ATP, an ATP regenerating system consisting of 30 mm creatine phosphate and 200 ng of creatine kinase) with 150 ng of M13mp18 DNA that had been irradiated with UV light emitting at 254 nm to a total dose of 30 J/m 2. Reaction mixtures were incubated at 30 C for varying times, deproteinized by treatment with SDS and proteinase K (7), and analyzed in a 0.8% agarose gel in TAE buffer (20 mm Tris acetate, ph 7.4, 0.5 mm EDTA). The DNA species were visualized by staining with ethidium bromide. Expression and Purification of GST-Rad7 Fusion Protein The GST- Rad7 fusion protein was overexpressed in the E. coli strain JM101, using the plasmid pr7.4. After induction with 1 mm isopropyl-1-thio-b- D-galactopyranoside, cells were broken and the GST-Rad7 fusion protein, which was localized in the inclusion bodies, was solubilized in 1.5% Sarkosyl buffer containing 0.1 M ethanolamine and 10 mm EDTA (ph 8.0). Triton X-100 was added to 2% to the supernatant containing the GST-Rad7 protein, and after clarification, the preparation was dialyzed extensively against phosphate-buffered saline at 4 C. The GST-Rad7 protein in this fraction was present at approximately 1.5 mg/ml. Affinity Binding to GST-Rad7 A Rad7 affinity matrix was generated by binding GST-Rad7 protein to glutathione-sepharose beads according to the instructions of the manufacturer (Amersham Pharmacia Biotech). The matrix was washed with 10 volumes of buffer (20 mm Tris acetate, 20% glycerol, 0.5 mm EDTA, and 0.01% Nonidet P-40) containing 50 mm KOAc, ph 7.0 and then resuspended in an equal volume of buffer containing 50 mm KOAc. An aliquot (2.5 l) of the beads was incubated with buffer containing 20 mm glutathione to check the amount of Rad7 bound to the matrix (1 g/ l beads). A similar protocol was followed to generate GST-Sepharose beads, using lysate containing GST (6.4 g/ l of beads). For binding experiments, purified Rad23 (0.5 g) or NEF2 (1.5 g) was mixed with 10 l of the GST- or GST-Rad7 affinity matrix in 150 l of buffer Q (20 mm Tris acetate, ph 7.0, 20% glycerol, 1.5 mm EDTA, 1 mm dithiothreitol, 0.01% Nonidet P-40) containing 50 mm KOAc at 4 C for 3 h. After removal of the unbound proteins, the matrix was washed with 0.2 ml of buffer Q containing 0.2, 0.4, and 0.8 M KOAc. Proteins were then eluted into 40 l of2%sdsby incubation of the affinity beads containing bound proteins at 37 C for 10 min. The supernatants (5 l), washes (5 l), and SDS-eluates (1.5 l) were analyzed for the presence of NEF2 or Rad23 protein by immunoblotting. RESULTS Synergistic Action of NEF2 and NEF4 in Damage Binding Binding of NEF4 to UV-damaged DNA is strongly stimulated by ATP or ATP S (11, 12), whereas NEF2 shows no dependence on a nucleotide in its binding to UV-irradiated DNA (13). To determine whether NEF2 and NEF4 interact synergistically in the binding of UV-damaged DNA, we used concentrations of NEF2 and NEF4 at which little binding of the UV-irradiated DNA fragment is seen with either factor alone and then investigated the effect of mixing the two NEFs in the presence of ATP. As shown in Fig. 1, A and B, whereas only 5% or less of the UV-damaged DNA fragment was bound by NEF2 or NEF4, combining the two NEFs resulted in a marked elevation in the level of binding ( 70%), an increase much higher than the simple sum of the binding activities seen with the two protein complexes (Fig. 1B). Concomitant with the increased binding of the UV-damaged DNA fragment, nucleoprotein complexes that have retarded mobilities (marked collectively as C* in Fig. 1A) compared with those formed between the DNA substrate and NEF2 or NEF4 alone (labeled C in Fig. 1A) were seen. Immunoblotting after transfer of the nucleoprotein complex from the polyacrylamide gel onto nitrocellulose has indicated the presence of both NEF2 and NEF4 in the C* complex, thus confirming that C* was a

Damage Binding by NEF2 and NEF4 24259 FIG. 1.Binding of UV-damaged DNA by NEF2 and NEF4. A, the 32 P-labeled UV-irradiated (10 kj/m 2 ) DNA (1 ng) was incubated in buffer (lane 1) and with NEF2 (10 ng), NEF4 (10 ng), or the combination of NEF2 and NEF4 (10 ng of each) in lanes 2, 3, and 4, respectively. ATP (2 mm) was included in these reactions. B, summary of the results in panel A. C, ATP promotes synergistic damage binding by NEF2 and NEF4. The 32 P-labeled DNA, either undamaged (lanes 1 9) or UV-irradiated (2 kj/m 2, lanes 10 18), was incubated with NEF2 (20 ng), NEF4 (20 ng), or the combination of NEF2 and NEF4 (20 ng of each), as indicated. Either ATP or ATP S,2mM final concentration, was added to the reactions, as indicated. D, the histogram summarizes the results in lanes 10 18 in panel C. Reaction mixtures were resolved in polyacrylamide gels, which were dried and then exposed to x-ray films to visualize the DNA and nucleoprotein species. Symbols in panels A and C: F, unbound DNA probe; C, complex formed between DNA probe and either NEF2 or NEF4; C*, ternary complexes formed among the DNA probe, NEF2, and NEF4. ternary complex consisting of the damaged DNA, NEF2, and NEF4. Under the conditions used, undamaged DNA is not bound by the combination of NEF2 and NEF4 (Fig. 1C). ATP Is Required for Ternary Complex Formation Because NEF4 binds and hydrolyzes ATP, and the presence of either ATP or ATP S is critical for optimal damage-specific DNA binding by NEF4 (11, 12), it was of considerable interest to examine whether the synergistic cooperation between NEF2 and NEF4 in damage-specific DNA binding is dependent on the presence of ATP. To do this, NEF2 alone, NEF4 alone, and the combination of NEF2 and NEF4 were incubated with undamaged DNA or with UV-irradiated DNA in the absence or the presence of ATP. As shown in Fig. 1, C and D, combining NEF2 and NEF4 results in a synergistic increase in the binding of UV-damaged DNA. Significantly, the formation of the NEF2- NEF4-UV-damaged DNA ternary complex was greatly diminished when ATP was omitted from the reaction mixture (Fig. 1C, compare lanes 16 and 17), indicating a requirement for ATP in the formation of the ternary nucleoprotein complex. As reported previously (11, 12) and reiterated here, ATP S, like ATP, was effective in promoting DNA damage binding by NEF4 (Fig. 1C, compare lanes 14 and 15 with lane 13). ATP S, however, did not stimulate the formation of the ternary complex of damaged DNA with NEF2 and NEF4 (Fig. 1C, compare lanes 16 and 18). This result stands in contrast to the damage binding by NEF4 alone, and it suggests that either only ATP is effective in inducing a conformation conducive for the interaction of NEF4 with NEF2 or that ATP hydrolysis is in fact indispensable for mediating the interaction of NEF2 with NEF4. No significant binding of the undamaged DNA fragment was seen even with the combination of NEF2 and NEF4, regardless of whether a nucleotide was present or not (Fig. 1C). Purification of Rad7 Protein The Rad7 protein does not contain any known conserved sequence motifs that might suggest a biochemical function for the protein. To gain insight into the biochemical role of Rad7, we overexpressed the protein in yeast cells (Fig. 2A) and purified it to near homogeneity (Fig. 2B). For the overexpression of Rad7 protein, the RAD7 gene tagged with a 6-histidine sequence at the amino terminus was placed under the control of the GAL-PGK promoter to yield plasmid pr7.8. For purifying Rad7 protein, extract from 300 g of the protease-deficient yeast strain LY2 harboring plasmid pr7.8 was subjected to ammonium sulfate precipitation, followed by chromatographic fractionation in columns of Q-Sepharose, SP-Sepharose, hydroxyapatite, nickel NTA-agarose, and finally in Mono-S, as described under Materials and Methods. The elution of Rad7 from the various columns was monitored by immunoblotting until the hydroxyapatite step, where Rad7 could be identified by Coomassie Blue staining of polyacrylamide gels in which the column fractions had been run. When 1 g of Rad7 protein eluting from the last step of purification in Mono-S column (Fraction VI Rad7 protein) was analyzed by denaturing polyacrylamide gel electrophoresis and Coomassie Blue staining, only the Rad7 protein was seen, indicating a high degree of purity of the preparation (Fig. 2B). The purified Rad7 preparation was devoid of Rad16 protein, as no Rad16 was detected when 1 g of the Rad7 preparation was subjected to immunoblot analysis with anti-rad16 antibodies

24260 Damage Binding by NEF2 and NEF4 FIG. 2.Overexpression and purification of Rad7 protein. A, a nitrocellulose blot containing extracts from strain LY2 harboring the cloning vector ppm231 (2 m GAL-PGK; lane 1), from strain LY2 harboring Rad7 overproducing plasmid pr7.8 (2 m GAL-PGK-RAD7; lane 2), and 15 ng of purified Rad7 protein (Fraction VI; lane 3) was probed with affinity purified anti-rad7 antibodies. B, purity analysis. Fraction VI Rad7 from the Mono-S step, 1 g, was run in an 8.5% denaturing polyacrylamide gel (lane 2) along with molecular mass markers (lane 1) and then stained with Coomassie Blue. under conditions wherein even 1 ng of Rad16 would have been seen. We obtained 60 g of Rad7 protein from the starting 300 g of yeast paste used, representing an overall recovery of 10%. Fraction VI Rad7 protein was used in the DNA binding and in vitro repair reactions described below. Rad7 Is a DNA Binding Protein Because NEF4 has a UV damage-specific DNA binding activity, we examined whether Rad7 protein alone can bind UV-damaged DNA. The results in Fig. 3A show that Rad7 has a DNA binding activity as indicated by the formation of a nucleoprotein complex with the 32 P-labeled DNA fragment. However, unlike the Rad7-Rad16 complex, Rad7 protein by itself does not discriminate between the UV-damaged and undamaged DNA, as essentially the same level of nucleoprotein complex was formed with both DNA species (Fig. 3A). The same results were obtained when ATP was omitted from the reaction buffer (data not shown). Furthermore, the binding of Rad7 to either the UV-irradiated or unirradiated DNA substrate can be effectively competed away by adding low amounts of a nonlabeled DNA to the preformed nucleoprotein complex (Fig. 3A), suggesting that the Rad7- DNA nucleoprotein complex is not very stable. In sharp contrast, NEF4 binds specifically and stably to UV-irradiated DNA even in the presence of a 100-fold excess of competitor DNA (Fig. 3B, lane 7). These results indicate that Rad7 is a DNA binding protein but it is devoid of the ability to recognize UV lesions. Rad7 Interacts with NEF2 In further characterizing the function of Rad7 protein, we considered the possibility that Rad7 may act as a physical link between NEF2 and NEF4. Because even with overexpression, only a small amount of Rad7 protein could be purified from yeast cells, to facilitate our study, we expressed Rad7 as a fusion polypeptide to GST and purified it from E. coli according to standard procedure entailing affinity chromatography on glutathione-sepharose. For examining whether there was a specific interaction between Rad7 and NEF2, either GST alone or purified GST-Rad7 immobilized on glutathione-sepharose was incubated with purified NEF2. The GST-Rad7 affinity matrix and the control GST matrix were washed with increasing concentrations of potassium acetate, and the bound NEF2 was eluted by treatment with SDS. The starting material, the supernatant containing unbound NEF2, the potassium acetate washes, and the SDS eluates were subjected to immunoblotting with anti-rad4 and anti- Rad23 antibodies to determine the amount of NEF2 that was FIG. 3.Rad7 is a DNA binding protein. A, DNA binding activity of Rad7 protein. The 32 P-labeled DNA fragment (1 ng) with or without prior treatment with UV light (10 kj/m 2 ) was incubated with 30 ng of Rad7 protein in the absence (lanes 2 and 7) or in the presence of 1, 2, and 5 ng of unirradiated HaeIII-digested X174 competitor DNA (lanes 3 5 and 8 10). Lanes 1 and 6 contain unirradiated and UV-irradiated DNA without Rad7, respectively. The reaction mixtures were resolved in a polyacrylamide gel followed by autoradiography to visualize the nucleoprotein complex (labeled as C) and free DNA probe (labeled as F). B, NEF4 nucleoprotein complex is unaffected by competitor DNA. Rad7 (30 ng, lanes 2 4) or NEF4 (70 ng lanes 5 7) was incubated with UV-damaged DNA (10 kj/m 2 ) in the presence of 30 (lanes 2 and 5), 60 (lanes 3 and 6), and 100 (lanes 4 and 7)-fold excess of unirradiated competitor DNA. Lane 1, DNA without protein. The reaction mixtures were resolved in a polyacrylamide gel, and the free probe (F) and nucleoprotein complexes (C) were visualized by autoradiography. specifically retained on the GST-Rad7 affinity matrix. Significantly, whereas greater than 90% of the input NEF2 was retained on the GST-Rad7 matrix, there was no binding of NEF2 to the GST matrix (Fig. 4A). Because we had purified Rad23 protein available (16), we examined whether there was a direct interaction between Rad23 and GST-Rad7. However, under conditions wherein NEF2 interacts strongly with GST-Rad7, we did not observe any binding of the Rad23 protein to the GST-Rad7 matrix (Fig. 4B). Taken together, the results demonstrate a specific and direct interaction of Rad7 with NEF2, but the interaction may be through the Rad4 protein or it may occur only with the Rad4-Rad23 complex. Rad7 Does Not Enhance the Incision Reaction Rad7 by itself shows no damage-specific binding (Fig. 3A), and it did not enhance the binding of UV-damaged DNA by NEF2 (data not shown). In a previous study (11), we showed that the addition of NEF4 to an in vitro NER reaction reconstituted with highly purified NEF1, NEF2, NEF3, and RPA markedly increased the efficiency of the damage-specific incision reaction. Because Rad7 interacts with NEF2, we examined whether the repair efficiency could be enhanced by the inclusion of the Rad7 protein in the reconstituted NER reaction. The addition of Rad7 to the basic incision reaction, however, did not result in significant stimulation or inhibition of the UV damage-specific incision (Fig. 5A, compare lanes 10 and 11 with lanes 6 and 7, respectively, and Fig. 5B), whereas, as reported previously, the inclusion of NEF4 enhances the incision reaction by 4 5-fold (Fig. 5A, compare lanes 8 and 9 with lanes 6 and 7, respectively,

Damage Binding by NEF2 and NEF4 24261 FIG. 4.Rad7 physically interacts with NEF2. A, NEF2 binds to GST-Rad7. Purified NEF2 (1.5 g) (lanes 1 and 7) was mixed with glutathione-sepharose containing either GST or GST-Rad7. The unbound fraction (lanes 2 and 8), 0.2 M (lanes 3 and 9), 0.4 M (lanes 4 and 10), 0.8 M (lanes 5 and 11) KOAc washes and SDS eluates (lanes 6 and 12) from the GST- and GST-Rad7 matrices were analyzed for their content of NEF2 by probing for Rad4 and Rad23 proteins using affinity purified antibodies. B, Rad23 does not interact with Rad7. Purified Rad23 protein (0.5 g) (lanes 1 and 7) was mixed with glutathione- Sepharose containing either GST and GST-Rad7. The unbound fraction (lanes 2 and 8), 0.2 M (lanes 3 and 9), 0.4 M (lanes 4 and 10), 0.8 M (lanes 5 and 11) KOAc washes and SDS eluates (lanes 6 and 12) from the matrices were analyzed for their content of Rad23 protein by immunoblotting. and Fig. 5B). These biochemical results and the results from previous genetic studies are consistent with the suggestion that both Rad7 and Rad16 proteins are required for the functional integrity of NEF4 in UV damage recognition and in NER enhancement. DISCUSSION Genetic studies have indicated that the RAD4 gene is indispensable for the NER of UV-damaged DNA in vivo, whereas the RAD23 gene affects the efficiency of the repair process (15). Previous studies from our laboratory have shown that the Rad4 and Rad23 proteins are tightly associated in a stoichiometric complex called NEF2. Consistent with the genetic results, NEF2 was shown to be indispensable for the dual incision of UV-damaged DNA in the in vitro NER reconstitution studies (7). Because the Rad4 and Rad23 proteins contain no identifiable sequence motifs that would predict an enzymatic activity, it had remained unclear as to whether NEF2 was important for NER by providing a distinct biochemical function or whether it primarily served as a scaffold in the assembly of the NER machinery. In fact, our previous studies have demonstrated that Rad23 interacts with Rad14 protein and with TFIIH (16). More recently, we showed that NEF2 possesses a DNA binding ability that is highly specific for UV-damaged DNA, but Rad23 by itself shows no DNA binding ability (13), indicating that either Rad4 is the damage recognition subunit of NEF2, or Rad4 and Rad23 proteins are both required for the DNA damage binding activity of NEF2. Similar to the DNA damage recognition ability of NEF2, the equivalent human NER complex consisting of the XPC and HHR23B proteins, also binds preferentially to UV-damaged DNA (18, 19). Taken together, the biochemical studies have provided tangible evidence that during NER, NEF2 serves as a damage recognition factor, and in addition, plays an important role in the assembly of the incision machinery at the damage site via specific proteinprotein interactions (16). The RAD7 and RAD16 genes affect the efficiency of NER in vivo with a particular requirement of these genes in the repair of nontranscribed DNA (9, 10). The addition of Rad7-Rad16 FIG. 5.NEF4, but not Rad7, enhances the incision of UV-damaged DNA. A, M13mp18 DNA unirradiated ( UV) (lanes 1 4) or irradiated with 30 J/m 2 of UV light ( UV) (lanes 5 11) was incubated with NEF1, NEF2, NEF3, and RPA (lanes 2, 6, and 7). Other lanes contained, in addition to these factors, either NEF4 (lanes 3, 8, and 9) or Rad7 (lanes 10 and 11), or both NEF4 and Rad7 (lane 4) at30 Cfor 8 or 15 min, as indicated. Lanes 1 and 5 contain unirradiated and UV-irradiated DNAs without any protein, respectively. The reaction mixtures were deproteinized, run in a 0.8% agarose gel, and stained with ethidium bromide to visualize the supercoiled form (SC) and open circular DNA (OC), generated as a result of damage-specific incision by the NER factors. B, results in lanes 6 through 11 of panel A are summarized in the histogram. complex, NEF4, to the reconstituted NER reaction results in a marked stimulation of the incision of UV-damaged DNA (11). Consistent with the presence in Rad16 of Walker-type sequence motifs suggestive of the ability to bind and hydrolyze a nucleoside triphosphate, our biochemical studies have indicated that NEF4 contains an intrinsic ATPase activity. The NEF4 ATPase activity requires DNA, and double-stranded DNA is more effective than single-stranded DNA for its activation (12). Like NEF2, NEF4 also binds specifically to UVdamaged DNA. However, unlike NEF2, which shows no dependence on ATP for damage binding, NEF4 requires ATP for maximal binding to UV-damaged DNA. ATP S is also effective, but to a lesser degree than ATP, in promoting DNA damage binding by NEF4 (12), which has led to the deduction that ATP binding alone is sufficient to induce a conformational change in NEF4 conducive for damage binding. The ATPase activity of NEF4 is markedly inhibited by the presence of UV damage in the DNA (12), suggesting that the free energy derived from ATP hydrolysis may fuel the translocation of NEF4 on DNA and that the movement of NEF4 is arrested upon encountering a DNA lesion. The lesion-bound NEF4 may then serve as the nucleation site for the loading of the remaining NER factors, including NEF1, NEF2, NEF3, and RPA. In addition to NEF2, Rad14, which is a component of NEF1, and RPA also bind UV-damaged DNA specifically (17, 20, 21); however, the sequence by which these damage recognition factors assemble at the site of lesion-bound NEF4 remains to be determined. To begin addressing the hypothesis that multiple damage

24262 Damage Binding by NEF2 and NEF4 recognition factors including NEF2 and NEF4 function together to achieve a high degree of specificity in damage demarcation during NER, we have now examined NEF2 and NEF4 for their possible cooperation in DNA damage recognition. Our results indicate that combining NEF2 and NEF4 results in synergistic binding of UV-irradiated DNA. We also find that the synergistic action of NEF2 and NEF4 in damage recognition is an ATP-dependent process, suggesting a role of ATP binding/hydrolysis in this reaction. To explore the role of Rad7 in NER, we overexpressed and purified this protein to near homogeneity. Our results show that Rad7 binds DNA but has no specificity for UV-damaged DNA template. Consistent with this result, addition of NEF4, but not of Rad7 alone, to the in vitro NER reaction results in marked stimulation of the incision reaction (Fig. 5). These observations suggest that either Rad16 protein by itself provides damage-specific binding in NEF4 or that the combination of Rad7 and Rad16 proteins is in fact necessary for the expression of damage-specific DNA binding. These scenarios have not yet been tested directly, since despite considerable efforts to purify the Rad16 protein we have thus far been unable to obtain a sufficient amount of highly purified Rad16 for these studies, which is because of the extremely low level of expression of Rad16 protein in the absence of Rad7 protein. Nevertheless, the fact that Rad16 contains two potential zinc binding, DNA binding motifs, a C 4 motif and a C 3 HC 4 ring finger motif (22), supports the suggestion that Rad16 by itself can discern between normal and damaged DNA species. If that was the case, then it is quite possible that the DNA binding ability of Rad7 has a role in enhancing the affinity of Rad16 for damaged DNA, or it increases the stability of the nucleoprotein complex formed between NEF4 and damaged DNA. In addition to the DNA binding activity, Rad7 also has the ability to physically interact with NEF2. Rad7 contains regions of marked hydrophobicity, including 12 tandemly repeated leucine-rich motifs (22), which may enable Rad7 to interact with NEF2. In summary, our results suggest that NEF4 functions as an ATPdependent DNA damage sensor and as an assembly factor during the incision phase of NER, and they provide support for the notion that the damage recognition reaction involves hierarchical interactions among different damage binding factors. Recent studies have suggested an involvement of the Rad7 and Rad16 proteins in the post-incision phase of NER (23). Thus, it appears that NEF4 plays a multifunctional role in NER, viz. in the very initial step of damage recognition in cooperation with NEF2 as shown in our studies here and also in the post-incision phase (23), perhaps in the turnover of the incision protein machinery, and in the recruitment of factors to perform the DNA repair synthesis reaction. REFERENCES 1. Friedberg, E. C., Walker, G. C., and Siede, W. (1995) DNA Repair and Mutagenesis, ASM Press, Washington, D. C. 2. Habraken, Y., Sung, P., Prakash, S., and Prakash, L. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 10718 10722 3. Mu, D., Park, C.-H., Matsunaga, T., Hsu, D. S., Reardon, J. T., and Sancar, A. (1995) J. Biol. Chem. 270, 2415 2418 4. Mu, D., Hsu, D. S., and Sancar, A. (1996) J. Biol. Chem. 271, 8285 8294 5. Mu, D., Wakasugi, M., Hsu, D. S., and Sancar, A. (1997) J. Biol. Chem. 272, 28971 28979 6. Guzder, S. N., Sung, P., Prakash, L., and Prakash, S. (1996) J. Biol. Chem. 271, 8903 8910 7. Guzder, S. N., Habraken, Y., Sung, P., Prakash, L., and Prakash, S. (1995) J. Biol. Chem. 270, 12973 12976 8. van Gool, A. J., Verhage, R., Swagemakers, S. M. A., van de Putte, P., Brouwer, J., Troelstra, C., Bootsma, D., and Hoeijmakers, J. H. J. (1994) EMBO J. 13, 5361 5369 9. Verhage, R., Zeeman, A.-M., de Groot, N., Gleig, F., Gang, D. D., van de Putte, P., and Brouwer, J. (1994) Mol. Cell. Biol. 14, 6135 6142 10. Mueller, J. P., and Smerdon, M. J. (1995) Nucleic Acids Res. 23, 3457 3464 11. Guzder, S. N., Sung, P., Prakash, L., and Prakash, S. (1997) J. Biol. Chem. 272, 21665 21668 12. Guzder, S. N., Sung, P., Prakash, L., and Prakash, S. (1998) J. Biol. Chem. 273, 6292 6296 13. Guzder, S. N., Sung, P., Prakash, L., and Prakash, S. (1998) J. Biol. Chem. 273, 31541 31546 14. Jansen, L. E. T., Verhage, R. A., and Brouwer, J. (1998) J. Biol. Chem. 273, 33111 33114 15. Watkins, J. F., Sung, P., Prakash, L., and Prakash, S. (1993) Mol. Cell. Biol. 13, 7757 7765 16. Guzder, S. N., Bailly, V., Sung, P., Prakash, L., and Prakash, S. (1995) J. Biol. Chem. 270, 8385 8388 17. Guzder, S. N., Sung, P., Prakash, L., and Prakash, S. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 5433 5437 18. Reardon, J. T., Mu, D., and Sancar, A. (1996) J. Biol. Chem. 271, 19451 19456 19. Sugasawa, K., Ng, J. M. Y., Masutani, C., Iwai, S., van der Spek, P. J., Eker, A. P. M., Hanaoka, F., Bootsma, D., and Hoeijmakers, J. H. J. (1998) Mol. Cell 2, 223 232 20. Clugston, C. K., McLaughlin, K., Kenny, M. K., and Brown, R. (1992) Cancer Res. 52, 6375 6379 21. Burns, J. L., Guzder, S. N., Sung, P., Prakash, S., and Prakash, L. (1996) J. Biol. Chem. 271, 11607 11610 22. Prakash, S., Sung, P., and Prakash, L. (1993) Annu. Rev. Genet. 27, 33 70 23. Reed, S. H., You, Z., and Friedberg, E. C. (1998) J. Biol. Chem. 273, 29481 29488