DNA Polymerase (cid:1) Protects Mouse Fibroblasts against Oxidative DNA Damage and Is Recruited to Sites of DNA Damage/Repair*

DNA polymerase (cid:1) (pol (cid:1) ) is a member of the X family of DNA polymerases that has been implicated in both base excision repair and non-homologous end joining through in vitro studies. However, to date, no phenotype has been associated with cells deficient in this DNA polymerase. Here we show that pol (cid:1) null mouse fibroblasts are hypersensitive to oxidative DNA damaging agents, suggesting a role of pol (cid:1) in protection of cells against the cytotoxic effects of oxidized DNA. Addition-ally, pol (cid:1) co-immunoprecipitates with an oxidized base DNA glycosylase, single-strand-selective monofunctional uracil-DNA glycosylase (SMUG1), and localizes to oxidative DNA lesions in situ . From these data, we conclude that pol (cid:1) and five thymidine residues were added as a termination signal. The resulting oligonucleotide duplex was cloned between the XhoI and HindIII sites of the pmH1P- pgkneo vector. Cell Sensitivity Assays— Cytotoxicity studies were performed as de- scribed previously (7). The day after plating the cells at a density of 40,000 cells per well in 6-well dishes, cells were exposed to increasing concentrations of 5-hydroxymethyl-2 (cid:1) -deoxyuridine (hmdUrd) (Sigma) for 24 h or H 2 O 2 (Sigma) for 1 h, washed with Hanks’ balanced salt solution, and ultimately allowed to grow for 4–5 days in growth media at 37 °C in a 10% CO 2 incubator. Triplicate experiments for each con- centration were counted using a cell lysis procedure (22), and the results are expressed as the “percent control growth” ((number of cells after treatment with hmdUrd)/(number of control cells) (cid:4) 100).

Mammalian cells express a large number of DNA polymerases that are considered to differ from one another according to specialized roles in DNA replication, repair, and other essential DNA transactions such as somatic hypermutation (1). Although these enzymes, numbering over two dozen, catalyze a similar chemical reaction, they differ in primary structure and in features of catalytic specificity as measured in vitro (2). In recent years, DNA polymerases have been grouped into categories or families based on primary structure similarities (3), and members of the X family have been assigned specialized roles in DNA repair (4).
DNA polymerase (pol ) 1 is a member of the X family of DNA polymerases, along with DNA polymerase ␤ (pol ␤) and several other enzymes (4). Pol ␤ has been shown to be important in the cellular defense against DNA base damage by virtue of its capacity to perform steps in base excision repair (BER) (5). For example, mouse embryonic fibroblast (MEF) cells that are deficient in pol ␤ are hypersensitive to the cytotoxic effects of DNA alkylating agents (6), and this appears to be due to the persistence of toxic intermediates of DNA repair that are normally removed by pol ␤ in wild-type cells (7). Biochemical characterizations of pol ␤ and steps in the BER pathway have focused attention on removal of the deoxyribose phosphate group from the BER intermediate after apurinic/apyrimidinic endonuclease (APE) cleavage of the abasic site (8), and this step is known to be important for alkylating agent-induced cytoxicity in MEFs (9).
DNA polymerase is a single polypeptide enzyme consisting of two overall domains, the N-terminal and the Cterminal (10,32). The latter is similar in size, structure, and enzymatic composition to pol ␤, including capacity for gapfilling synthesis and deoxyribose phosphate lyase activity on a BER substrate. It has also been shown that pol is able to conduct uracil-initiated BER in extracts of MEFs (11), as well as in assays where BER is reconstituted with purified enzymes (12). Nevertheless, there appears to be no requirement for pol in the MEF cellular response to DNA alkylating agents, since cells deficient in pol showed no hypersensitivity to these agents (13). Thus, the role of pol in the cellular BER response to DNA damage is still unclear. Another important feature of pol is that it has been implicated in double-strand break repair in mammalian cell extracts (14), whereas pol ␤ has not. Genetic studies of Saccharomyces cerevisiae pol IV first implicated a pol like enzyme in double-strand break repair (15,16). Yet, a similar role in living mammalian cells remains to be confirmed. Finally, the N-terminal domain of pol carries no known enzymatic activity; however, this domain contains a BRCT motif that may be involved in targeting the enzyme through protein-protein interactions (10,14).
In this study, we further examined the role of pol in cellular protection against oxidative DNA damage. DNA lesions such a 5-hydroxymethyluracil (hmUra) and 8-oxoguanine are produced by oxidative stress agents (17) and are repaired mainly by the BER pathway. We report here that MEFs with a pol gene deletion are hypersensitive to the oxidative stress agent H 2 O 2 and are even more hypersensitive to the introduction of hmUra into genomic DNA. DNA polymerase was found to localize to sites of oxidative damage by in situ analysis of cells, and the requirements for pol localization to damage/repair sites was examined.
DNA Constructs-The plasmid pEGFP-N1-dN-XN was modified from pEGFP-N1 (Clontech, Mountain View, CA) by deletion of the original NotI site and introduction of Xho-EcoRV-NotI site between the BglII and AgeI sites at the multiple cloning site. The plasmid pEGFP-N1-dN-XN, containing full-length pol (cloned between the XhoI and NotI sites of pEGFP-N1-dN-XN-pol ), was constructed similarly to the method previously described (20). The BRCT domain of pol , pol -BRCT 1-174 , encoding Met 1 to Arg 174 , was prepared by digesting the plasmid pEGFP-N1-dN-XN-pol with XhoI and StuI. The resulting fragment was purified and introduced into the plasmid pEGFP-N1-dN-XN between the XhoI and EcoRV sites. The polymerase domain of pol , pol -POL 242-575 encoding Ala 242 to Trp 575 , was amplified by the PCR from the plasmid pEGFP-N1-dN-XN-pol . The PCR primers, 5Ј-GTCCTCGAGGCACAGCCCTCAAGCCAGAAGGCG-3Ј and 5Ј-CTC-GAGCCAGTCCCGCTCAGCAGGTTCTCG-3Ј, were designed to incorporate XhoI sites at both termini. The resulting PCR fragment was introduced into the plasmid pEGFP-N1-dN-XN at the XhoI site. Each construct utilized the initiation codon (ATG) in Kozak sequences introduced in the host plasmid pEGFP-N1-dN-XN and was in-frame relative to EGFP at the C terminus of each protein. Single-strand selective monofunctional uracil-DNA glycosylase (SMUG1) was cloned between the BglI and SalI sites of the pEGFP-C1 vector (Clontech). The pol siRNA expression vector was constructed as described previously (21). The pol siRNA expression vector was constructed as previously described (21). Pairs of DNA oligonucleotides encoding hairpin RNAs were designed based on four different pol gene-specific targeted sequences (position on the pol gene is indicated in parentheses): 5Ј-ATTGA-GCAGACGGTCCGGA-3Ј (1259), 5Ј-GGGTTCCTCACAGATGACT-3Ј (1439), 5Ј-GCTCTGGATAAATGGGTCT-3Ј (779), and 5Ј-GGCAACTA-ACTACAATCTG-3Ј (820).
The 19-nucleotide target transcript sequence and its reverse complement were separated by a short spacer, and five thymidine residues were added as a termination signal. The resulting oligonucleotide duplex was cloned between the XhoI and HindIII sites of the pmH1Ppgkneo vector.
Cell Sensitivity Assays-Cytotoxicity studies were performed as described previously (7). The day after plating the cells at a density of 40,000 cells per well in 6-well dishes, cells were exposed to increasing concentrations of 5-hydroxymethyl-2Ј-deoxyuridine (hmdUrd) (Sigma) for 24 h or H 2 O 2 (Sigma) for 1 h, washed with Hanks' balanced salt solution, and ultimately allowed to grow for 4 -5 days in growth media at 37°C in a 10% CO 2 incubator. Triplicate experiments for each concentration were counted using a cell lysis procedure (22), and the results are expressed as the "percent control growth" ((number of cells after treatment with hmdUrd)/(number of control cells) ϫ 100).
siRNA Expression Vector Transfection and Clone Selection-The vectors for stable expression of siRNA against pol were transfected into pol ␤Ϫ/Ϫ MEFs and selected for antibiotic resistance as described previously (21). Cells transformed with siRNAs against the target sequence 5Ј-GGGTTCCTCACAGATGACT-3Ј demonstrated the lowest level of pol expression as determined by Reverse Transcriptase-PCR. From this pool of cells, three clones were isolated containing approximately a 60% reduction in pol expression that was confirmed by reverse transcriptase-PCR and Western blot analysis. In parallel, pol ␤Ϫ/Ϫ MEFs were also transfected with the empty pmH1P-pgkneo vector as a negative control.
Cell Extract Preparation-For Western blotting and in vitro BER assays, cell extracts were prepared as described previously (11). Briefly, cells were washed with phosphate-buffered saline, detached by scraping, pelleted, and resuspended in equal volumes of Buffer I, composed of 10 mM Tris-Cl, pH 7.8, 200 mM KCl, and 1 complete mini, EDTA-free protease inhibitor mixture tablet (Roche Applied Science, Mannhein, Germany) per 2.5 ml Buffer I solution, and Buffer II composed of 10 mM Tris-Cl, pH 7.8, 200 mM KCl, 2 mM EDTA, 40% (v/v) glycerol, 0.2% (v/v) Nonidet P-40, and 2 mM dithiothreitol. The suspension was rotated at 4°C for 1 h, and the resulting extracts were clarified by centrifugation. Supernatant fractions were recovered for use in Western blotting and in vitro BER assays after determining the protein concentration. For co-immunoprecipitation using cell extracts, pol ϩ/ϩ and pol Ϫ/Ϫ MEFs were washed and harvested as described above. Cell pellets were resuspended in lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 25 mM NaF, 0.1 mM sodium orthovanadate, 0.2% (v/v) Triton X-100, 0.3% (v/v) Nonidet P-40) containing protease inhibitors (0.1 mM phenylmethylsulfonyl fluoride, 1 mg/ml aprotinin, and 5 mg/ml leupeptin) and processed as described previously (23). Briefly, extracts were incubated on ice for 30 min and centrifuged at 20,800 ϫ g for 30 min at 4°C, and the supernatant fraction was used for co-immunoprecipitation assays.
Western Blotting -Cell extracts (30 -50 g) were suspended in SDS sample buffer. The proteins were heated at 95°C for 5 min, separated by 4 -12% NuPAGE Novex BisTris PAGE (Invitrogen), and transferred to nitrocellulose membranes. Initially the membranes were blocked with 5% (w/v) nonfat milk in Tris-buffered saline containing 0.05% (v/v) Tween 20 (TBST) and then incubated with primary antibodies against pol . The filters were then washed with TBST and secondary peroxidase-conjugated AffiniPure donkey anti-rabbit IgG antibody (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) was added. After three washes with TBST, enhanced chemiluminescence was used to detect the peroxidase conjugate by exposure to x-ray film.
In Vitro BER Assay-A 35-mer oligonucleotide containing a sitespecific hmUra modification at position 18 of the following sequence was purchased from Oligos Etc. (Wilsonville, OR): 5Ј-CTGCAGCTGAT-GCGCCGXACGGATCCCCGGGTA-3Ј, where "X" denotes the position of the hmUra. This oligonucleotide was annealed to a complementary strand that contained a G opposite the hmUra lesion. In vitro BER time course reactions were performed by incubating this 35-base pair oligonucleotide duplex (250 nM) with 15 g of MEF cell extracts at 37°C in a buffer containing 25 mM Tris, pH 7, 60 mM NaCl, 2 mM dithiothreitol, 0.2 mM EDTA, 1 mg/ml bovine serum albumin, and 10% (v/v) glycerol. At the indicated times, 3-l aliquots of the reaction mixture were removed for analysis and BER products were measured as described previously (11). In some cases, cell extracts were preincubated with either preimmune serum or a neutralizing antibody against pol for 30 min on ice and an additional 10 min at 25°C prior to performing in vitro BER reactions.
Co-immunoprecipitation Assays-Co-immunoprecipitation reactions with pol ϩ/ϩ or pol Ϫ/Ϫ cell extracts were performed by mixing 1 mg of total extract protein and 0.7 mg of rabbit non-immune IgG, anti-pol polyclonal, or anti-SMUG1 polyclonal antibody. The protein-antibody mixture was incubated at 4°C with rotation for 4 h and immunocomplexes were adsorbed onto TrueBlot anti-rabbit IgG IP beads (eBioscience, San Diego, CA) by incubating the mixture overnight at 4°C. The beads were washed four times with lysis buffer containing 0.1 mM phenylmethylsulfonyl fluoride, 1 mg/ml aprotinin, and 5 mg/ml leupeptin, resuspended in SDS sample buffer, and heated for 5 min, and the soluble proteins were separated by 4 -12% NuPAGE Novex BisTris PAGE. After transfer, proteins were detected as previously described using anti-SMUG1 antibody (1:1000 dilution) or anti-pol antibody (1:5000) as a primary probe and goat anti-rabbit IgG TrueBlot conjugated to horseradish peroxidase (1:5000 dilution) as a secondary antibody (23). Immobilized horseradish peroxidase activity was detected by enhanced chemiluminescence. The nitrocellulose filter was then stripped by incubation in a buffer containing 6.25 mM Tris-HCl, pH 6.8, 100 mM ␤-mercaptoethanol and 1% (v/v) SDS for 30 min at 50°C followed by two washes with TBST at room temperature. The presence of pol was confirmed by incubating the membrane with rabbit anti-pol polyclonal antibody. Similarly, the cell extract was immunoprecipitated with anti-SMUG1 polyclonal antibody to detect pol . After stripping the blot, the presence of SMUG1 was confirmed using anti-SMUG1 antibodies.
Co-immunoprecipitation of purified pol and SMUG1 proteins was performed in the presence of binding buffer (25 mM Tris, pH 8, 10% (v/v) glycerol, 100 mM NaCl, 0.01% (v/v) Nonidet P-40) containing protease inhibitors (0.1 mM phenylmethylsulfonyl fluoride, 1 mg/ml aprotinin, and 5 mg/ml leupeptin) as described previously (23). An equimolar mixture of 1.5 M pol or truncated pol and SMUG1 was combined with either anti-pol or anti-SMUG1 antibody and rotated for 4 h at 4°C in a final volume of 500 l. Antibody-containing protein complexes were adsorbed onto TrueBlot anti-rabbit IgG IP beads by incubating the mixture overnight at 4°C with agitation. The beads were then washed with binding buffer containing protease inhibitors, suspended in SDS sample buffer, and heated for 5 min at 95°C. Soluble proteins were separated by 4 -12% NuPAGE Novex BisTris PAGE and transferred to nitrocellulose membranes. The membranes were blocked in 5% (w/v) nonfat dry milk in TBST and immunoblotting was performed with the appropriate antibody as described above.
Microscopy and Laser-light Irradiation-HeLa, pol ϩ/ϩ, pol Ϫ/Ϫ, XRCC1ϩ/ϩ, or XRCC1Ϫ/Ϫ cells were plated on 60-mm glass-bottom dishes at a density of 1 ϫ 10 5 cells/dish and transiently transfected with variants of the pEGFP-N1-dN-XN mammalian expression vector using FuGENE 6 (Roche Diagnostics) as a transfection reagent. Forty-eight hours after transfection, cells expressing GFP-tagged proteins were identified and a line for the path of the laser was drawn that spanned the width of the cell's nucleus. A 365-or 405-nm laser was focused through a 40ϫ objective lens using a FV-500 confocal scanning laser microscopy system (Olympus, Tokyo) to irradiate the cell along this path and produce damage/repair sites. During irradiation, cells were covered with a chamber to prevent evaporation on a 37°C heating plate. The mean intensity of each focus was obtained after subtraction of the background intensity in the irradiated cell.

DNA Polymerase
Ϫ/Ϫ Cells Are Hypersensitive to Oxidative DNA Damage-causing Agents-MEFs that are pol -deficient by virtue of homozygous gene deletion were examined for sensitivity to DNA damaging agents. These pol null cells were not significantly more sensitive to methylmethane sulfonate than an isogenic wild-type line developed in parallel (data not shown). This is in contrast to the clear methylmethane sulfonate hypersensitivity observed with pol ␤ null cells (6). DNA polymerase null cells, however, exhibited slightly more hypersensitivity to H 2 O 2 than pol ␤ null cells (Fig. 1A). This observation encouraged us to examine the sensitivity of pol null and isogenic wild-type cells to hmdUrd. Use of this nucleoside allows introduction of a common oxidative base lesion into genomic DNA without exposing cells to an oxidative stress agent in the culture medium (24): hmdUrd is taken up, converted into the 5-hydroxymethyl-2Ј-deoxyuridine triphosphate, and incorporated into genomic DNA during replication. The hmUra lesion appears to be innocuous unless BER is initiated by SMUG1, which recognizes hmUra in genomic DNA and removes it, initiating the BER process (25,26). We found that pol null cells were strongly hypersensitive to hmdUrd (Fig.  1B), and this effect was somewhat stronger than that observed previously with pol ␤ null cells (7).
The hmdUrd hypersensitivity of the pol null cells could be complemented by recombinant expression of mouse pol to a level similar to that found in the isogenic wild-type cell line (Fig. 1, C and D). Thus, stable transfection of pol null cells with an expression vector for mouse pol and then selection of cells with resistance to hmdUrd allowed us to isolate two cell MEF cells were exposed to increasing concentrations of H 2 O 2 for 1 h. Percent control growth was plotted for each data point, representing the mean values of triplicate experiments. In some cases standard error bars are smaller than the data symbols. B, pol ϩ/ϩ (closed circles) and pol Ϫ/Ϫ (open circles) MEF cells were exposed to increasing concentrations of hmdUrd for 24 h and percent control growth was plotted. C, pol Ϫ/Ϫ MEFs were transfected either with an empty expression vector control (open circles) or a vector to express mouse pol (closed circles). These transfected cells were exposed to increasing concentrations of hmdUrd and surviving cells were counted and plotted as described above. These transfected cells were exposed to increasing concentrations of hmdUrd and surviving cells were counted and plotted as described above. F, cell extract protein (30 g) prepared from MEFs cells in E was separated by PAGE and immunoblotted. Photographs of enhanced chemiluminescencestained IBs are shown after reaction with antibodies specific to pol as indicated.
lines with an expression level of pol similar to that of wildtype cells. Analysis of these transfected cells revealed a protection by pol against hmdUrd-induced cytotoxicity, whereas cells transfected with an empty vector showed no such complementation (Fig. 1C). Additionally, in a series of experiments not shown, introduction of human pol into the pol null cells by expression vector transfection failed to produce any complementation of the hmdUrd hypersensitivity.
Next, we explored siRNA-mediated down-regulation, an alternate approach toward developing a pol deficiency in MEFs. Since both pol ␤ and pol appear to participate in repair of hmUra lesions in MEFs, we examined hmdUrd cytotoxicity in cells containing reduced expression of both these polymerases. For these experiments, pol ␤ null MEFs were stably transfected with a pol -siRNA-producing vector. Two clonal cell lines were eventually obtained with ϳ60% reduction in pol level (Fig.  1F), and the effect of this reduction in pol on hmdUrd sensitivity was examined. These pol knockdown cells were more sensitive to hmdUrd than were cells transformed in parallel with an empty vector and used as a negative control (Fig. 1E). We conclude from these results with MEFs that a deficiency in pol is associated with hypersensitivity to oxidative stress and to an oxidized base, hmUra, in genomic DNA.
DNA Polymerase Mediates Repair of hmUra in Vitro-Earlier, we had shown that pol is capable of mediating a back-up role to pol ␤ in MEF extract mediated BER of uracil-DNA (11). In view of the results on hmdUrd hypersensitivity described above, we wished to verify that pol is able to perform extract mediated BER of hmUra-DNA. Results of BER assays with cell extracts and an oligonucleotide DNA substrate carrying the hmUra lesion are shown in Fig. 2. DNA polymerase was capable of a role in extract mediated BER as revealed by pol null cell extracts when compared with wild-type cell extracts (Fig. 2, A and B). In addition, pol was able to mediate repair in an extract from pol ␤ null cells, as revealed by antibody inhibition (Fig. 2C). Finally, using a BER system reconstituted with purified human SMUG1, APE, pol or pol ␤, and DNA ligase I, we found robust repair of the hmUra containing substrate in reactions with either polymerase (Fig. 2D). This was reminiscent of earlier work by others (12) showing that purified pol is able to mediate BER of uracil-DNA.
DNA Polymerase Co-immunoprecipitates with SMUG1-In light of the dependence on pol for cellular resistance to hmdUrd, we wished to explore the possibility that pol and the DNA glycosylase for hmUra-DNA, SMUG1, may cooperate through a protein-protein interaction. Results from co-immunoprecipitation experiments with purified SMUG1 and pol are shown in Fig. 3. Initially, we examined an extract from pol ϩ/ϩ MEFs to see if co-immunoprecipitation of pol and SMUG1 could be observed. Both enzymes were co-immunoprecipitated with the respective antibody (Fig. 3, A and B). To determine whether pol and SMUG1 directly interact with each other, these precipitation reactions were repeated with purified pol and SMUG1 protein. SMUG1 was co-precipitated with antibody to pol (Fig. 3C) and conversely the antibody to SMUG1 co-precipitated pol (Fig. 3D). Negative controls in these experiments failed to reveal any nonspecific precipitation of either protein. Since the BRCT domain of pol has been hypothesized to participate in protein-protein interactions, we evaluated whether this domain was responsible for the interaction between SMUG1 and pol . As seen in Figs. 3E and F, a truncated version of pol lacking the BRCT domain co-precipitated with SMUG1 when an antibody against pol was used. The inverse precipitation reaction using SMUG1 antibody also co-precipitated these proteins. We conclude from these experiments that pol and SMUG1 directly interact allowing the  (lanes 1-4) or pol neutralizing antibody (lanes 5-8) as described. After preincubation, hmUra-containing 35-base pair duplex DNA was added, and in vitro BER was initiated. Repair reactions continued for 10 ( lanes 1 and 5), 15 (lanes 2 and 6), 30 (lanes  3 and 7), or 60 min (lanes 4 and 8). Photographs of the autoradiogram after denaturing PAGE are shown, and the position of the ligated BER product (35-mer) is indicated. D, results of reconstituted BER with purified enzymes, illustrating repair of hmUra. In vitro BER time course reactions were reconstituted with 20 nM SMUG1, 10 nM APE, 200 nM DNA ligase I, and 100 nM pol (lanes 1-3) or 10 nM pol ␤ (lanes 4 -6). BER reaction mixtures were incubated for 10 (lanes 1 and 4), 30 (lanes 2 and 5), or 60 min (lanes 3 and 6) and analyzed as described. Photographs of autoradiograms after denaturing PAGE are shown, and the position of the ligated BER product (35-mer) is indicated. enzymes to be co-immunoprecipitated, either from wild-type cell extract or a mixture of purified enzymes.
DNA Polymerase Is Recruited to Sites of Laser-induced Oxidative DNA Damage-The strategy for studying localization of enzymes involved in oxidative DNA damage repair in living mammalian cells has previously been established (20). Cells expressing GFP-tagged SMUG1 or pol were subjected to near-UV light irradiation in the nucleus to produce oxidative base damage, single-strand DNA breaks, and double-strand DNA breaks. The GFP-tagged enzymes were followed by confocal microscopy, as described previously (20). We used this system to examine localization of SMUG1 and pol to sites of damage/repair and obtained similar results with both HeLa cells and MEFs. Initially we observed that SMUG1 co-localized to sites of damage/repair (Fig. 4A). Also, SMUG1 accumulation at damage/repair sites was independent from pol expression, since SMUG1 localization was not impaired in a pol null background (Fig. 4B). However, the kinetics for SMUG1 foci formation were altered in the absence of pol . While SMUG1 foci rapidly dissipated after 20 s in wild-type cells, pol null cells retained SMUG1 foci for the 300 s duration of the study (Fig. 4B). We also examined the requirements for pol foci formation. DNA polymerase was also able to form foci at laser-irradiated sites (Fig. 4C), and its localization required the N-terminal domain of the protein (Fig. D). Although the laser treatment used in this study produces oxidative base damage, single-strand DNA breaks and double-strand DNA breaks, we hypothesize that pol accumulates at sites of oxidative base damage. In data not shown, we observe that the accumulation of pol is enhanced in the presence of RO-19-8022, a photosensitizer that increases production of oxidized bases after irradiation (20). Additionally, we examined accumulation at double-strand DNA breaks. HeLa cells were pretreated with bromodeoxyuridine to allow for its incorporation into genomic DNA. Upon laser irradiation of these cells, energy from the near-UV light is absorbed by bromine and released to break nearby phosphodiester bonds and produce double-strand DNA breaks (27). Using these conditions, we did not observe an increase in formation of pol foci (data not shown), suggesting that pol accumulation does not increase when the number of double-strand DNA breaks increases. However, we cannot exclude the possibility that small increases in recruitment of GFP-tagged pol or endogenous pol may not have been detectable in this study. The dependence upon XRCC1 and PARP activity was also evaluated for formation of pol foci. DNA polymerase recruitment was identical in a background of either wild-type or XRCC1Ϫ/Ϫ MEFs (Fig. 4, E and F). Similarly, treatment of HeLa cells with the inhibitor 1,5- dihydroxyisoquinoline under conditions that inhibited PARP failed to alter pol recruitment (data not shown). These results indicate that requirements for pol localization to sites of damage/repair were different from those observed for pol ␤ localization, which required both XRCC1 and PARP activity (20). DISCUSSION Mammalian cells are constantly exposed to endogenous and exogenous reactive oxygen species that are capable of modifying DNA. While cells have developed several mechanisms for protection against the cytotoxic and mutagenic effects of DNA damage, BER is considered the predominant DNA repair pathway for combating the detrimental effects of oxidized base lesions. Several subpathways of BER have been identified that are generally divided into two main categories: single nucleotide BER and long patch BER. While the goal of the BER subpathways is identical, each requires a unique repertoire of BER enzymes and accessory proteins to complete the repair process after its initiation, and there may be a need for specialized DNA polymerases in these processes. Typically, BER begins with a lesion-specific DNA glycosylase that cleaves the N-glycosidic bond between the damaged base and deoxyribose. In the case of some oxidative damage glycosylases, strand incision often accompanies base removal, whereas in other cases the glycosylase product is cleaved by APE, to produce a single-strand break containing BER intermediate. Gap tailoring and DNA synthesis ultimately generate the nicked DNA intermediate that is ligated to complete repair. DNA polymerase ␤ has been shown to have a role in protection of MEFs against DNA alkylating agents.
DNA polymerase ␤ also has a protective effect against DNA oxidizing agents, but in the absence of pol ␤, cells continue to show significant survival upon exposure to oxidizing compounds (6,28). Therefore, the existence of pol ␤-independent forms of oxidative DNA damage BER has been long suspected (29). Considering that pol is biochemically capable of BER (12) and shares key repair related characteristics with pol ␤, we decided to examine whether pol can protect cells against oxidative DNA damage. Initially, we found that MEFs contain-ing a deletion in the pol gene were hypersensitive to H 2 O 2 and were even more hypersensitive to incorporation of the oxidative lesion hmUra into genomic DNA. Next, we confirmed that pol either in cell extract or as a purified enzyme could conduct BER of an hmUra-containing substrate. We also found that pol could directly interact with SMUG1, a glycosylase that removes the hmUra lesion. Finally, we found that pol was recruited to sites in the nucleus of living cells that were repairing oxidative DNA lesions. These results are the first demonstration of a phenotype for pol -deficient mammalian cells and are also the first demonstration of the localization of pol to sites of oxidative repair in living cells. Previous studies with pol Ϫ/Ϫ mouse embryonic stem cells (13) and pol IV null S. cerevisiae (30) failed to reveal H 2 O 2 hypersensitivity in null cells. This may be due to the different cell types studied, thereby reflecting differential responses to oxidative stress. Additionally, these earlier studies generated oxidative DNA damage with ϳ10-fold higher concentrations of H 2 O 2 than used here. These higher concentrations are known to lower the efficiency of aldehydic DNA lesion production in mammalian cells, in contrast to the optimum concentrations of H 2 O 2 employed in our study to produce these lesions (31).
It is likely that pol exerts its protective effect against oxidative DNA damage by conducting steps in BER. The hypersensitivity of pol null cells to oxidative damaging agents indicates that the role of pol is not fully complemented by pol ␤ or the other DNA polymerases expressed in these cells. This suggests that there is a specialized role for pol in oxidized lesion repair. Yet, pol ␤ also appears to function in hmUra-DNA repair, since pol ␤ null cells are hypersensitive to hmdUrd (7), and pol ␤ is capable of mediating BER of hmUra-DNA in the cell extract. The mechanism underlying the dependence on pol , even in the presence of strong pol ␤ expression, remains to be elucidated. Similarly, although pol and pol ␤ share many features, the requirements for their respective recruitment to sites of damage/repair in living cells are different. Finally, we were interested in the observation that the Nterminal domain of pol was required for localization to sites of damage/repair. Thus, the recruitment to damage/repair sites appeared to be independent of the pol -SMUG1 interaction, which did not require the N-terminal domain of pol . It is possible that the protein-protein interaction between these two BER enzymes is important for coordination of BER steps, whereas the N-terminal domain of pol may target it to complexes assembled at sites of damage/repair.