Involvement of Vertebrate Polκ in Rad18-independent Postreplication Repair of UV Damage* 210

DNA damage, which is left unrepaired by excision repair pathways, often blocks replication, leading to lesions such as breaks and gaps on the sister chromatids. These lesions may be processed by either homologous recombination (HR) repair or translesion DNA synthesis (TLS). Vertebrate Polκ belongs to the DNA polymerase Y family, as do most TLS polymerases. However, the role for Polκ in vertebrate cells is unclear because of the lack of reverse genetic studies. Here, we generated cells deficient in Polκ (polκ cells) from the chicken B lymphocyte line DT40. Although purified Polκ is unable to bypass ultraviolet (UV) damage,polκ cells exhibited increased UV sensitivity, and the phenotype was suppressed by expression of human and chicken Polκ, suggesting that Polκ is involved in TLS of UV photoproduct. Defects in both Polκ and Rad18, which regulates TLS in yeast, in DT40 showed an additive effect on UV sensitivity. Interestingly, the level of sister chromatid exchange, which reflects HR-mediated repair, was elevated in normally cycling polκ cells. This implies functional redundancy between HR and Polκ in maintaining chromosomal DNA. In conclusion, vertebrate Polκ is involved in Rad18-independent TLS of UV damage and plays a role in maintaining genomic stability.

A wide range of potential insult to the genomic DNA is contributed not only by the environment, but also by cellular activities. DNA damages, which are left unrepaired by excision repair pathways, may arrest DNA replication, leading to breaks and gaps on the sister chromatids. These lesions are processed by two major postreplication repair pathways: homologous recombination (HR) 1 repair and translesion DNA synthesis (TLS) (1). TLS fills a daughter strand gap that encompasses a DNA damage on the template strand by employing a number of specialized DNA polymerases (Pols) (reviewed in Refs. [2][3][4]. On the other hand, HR functions in processing both gaps and breaks by facilitating recombination between damaged sister chromatids with the other intact ones. Thus, daughter-strand gaps can be processed by both TLS and HR in postreplicational repair. Human cells contain four Y family DNA polymerases, namely Pol (Rad30A), Pol (Rad30B), Pol (DinB1), and Rev1 (5)(6)(7). A number of biochemical studies have demonstrated that Y family DNA polymerases can bypass specific lesions on the template strand (8 -12), whereas in vivo function of these polymerases are as yet unclear except for Pol. Pol is mutated in a variant form of xeroderma pigmentosum (XP-V) (13,14), which is characterized by predisposition to skin cancer and elevated UV sensitivity.
Genetic studies of Saccharomyces cerevisiae suggest that all known components of TLS belong to the epistasis group of the RAD6-RAD18 genes (21,22), which are conserved from yeast to mammalian cells (23)(24)(25)(26)(27). The Rad6 protein is one of the ubiquitin-conjugating enzymes (E2s) and forms a tight complex with the Rad18 protein, which may have E3 ubiquitin ligase activity (28,29). Although all TLS is under Rad18 control, S. cerevisiae does not have a POLK gene. On the other hand, Schizosaccharimyces pombe does have a POLK ortholog, but no data about genetic relationship between Pol and Rad18 has been obtained. Similarly, it is not known whether Pol is regulated by Rad18 in vertebrate cells.
Here we present a phenotypic analysis of Pol-deficient cells (hereafter abbreviated as pol cells) derived from the chicken B lymphocyte line DT40 (30). Remarkably, pol cells showed elevated sensitivity to UV, although the data are not in agreement with previous biochemical studies. We also provide genetic data that Pol and Rad18 show an additive effect on UV sensitivity, suggesting that Pol is involved in Rad18-independent TLS of UV lesions.

EXPERIMENTAL PROCEDURES
Cloning of the Chicken POLK Gene-A chicken POL⌲ (GdPOL⌲) partial cDNA was amplified from chicken testis cDNA by RT-PCR with the primer pair, 5Ј-CCATAGTGCACATTGACATGG-3Ј and 5Ј-CAC-GAACACCAAATCTCCTTGC-3Ј, the design of which was based on the conserved polymerase motif in the human and mouse POL⌲ genes. An amplified 144-bp fragment was subcloned into a pCRII-TOPO vector (Invitrogen, Carlsbad, CA) and sequenced. A full-length Gd POLK cDNA was isolated by RACE reactions using the gene-specific primers, 5Ј-ACAACATACTCATTGATCCCACTGC-3Ј for 5Ј-RACE and 5Ј-AATC-CAGAGCTGAAGGAAAAGCC-3Ј for 3Ј-RACE. The sequences of five independently isolated cDNA clones were determined. The chicken POLK cDNA sequence has been deposited in the GenBank TM data base under the accession number AY118271.
Construction of Targeting and Expression Vectors-A 14-kb fragment containing part of the chicken genomic POLK locus was isolated from a DT40 genomic DNA library by hybridization with the 144-bp PCRamplified fragment. The positions of exons and introns were determined by base sequencing. Chicken POLK disruption constructs containing drug resistance genes, POLK-HIS and POLK-BSR were made by replacing 2-kb genomic sequences containing exons encoding amino acids 96 -209 with histidinol-D and blasticidin-S selection marker cassettes. Since DT40 cells have a single POLK gene, the genotype of POLK-depleted cells is described as POLK /Ϫ . POLK-BSR was used for generating rad18 pol (RAD18 Ϫ/Ϫ POLK /Ϫ ) cells from rad18 (RAD18 Ϫ/Ϫ ) cells containing the Histidinol-D-and Hygromycin-resistance genes (27). To construct an HA-tagged GdPOLK expression vector (designated as pcDB3), we amplified Gd POL⌲ cDNAs including the sequence encoding human hemaggulutinin at the 3Ј-end by high-fidelity PCR with primers 5Ј-GGATCCATGGACAACGTGAAAAATGTTGA-T-3Ј and 5Ј-GGATCCTCATGCGTAGTCAGGGACGTCGTACGGATAT-TTAA-AAAACATTTCAATTGTACGCT-3Ј and then inserted the PCR product into the BamHI site of pUHG10-3, an expression vector containing a tetracycline-repressible promoter (kindly provided from Dr. Bujard at ZMBH). To construct a human POLK expression vector (phPK-G), a full-length Hs POLK cDNA (16) was inserted into the EcoRI site of the multicloning sites of the expression vector, pCR3-loxPmulticloning sites/IRES-EGFP-loxP (designated as p176, in Ref. 31). The plasmids were linearized prior to transfection into DT40 cells; POLK-HIS, POLK-BSR, and pcDB3 were digested with XhoI, and p176 and phPK-G were digested with PvuI.
Transfection, Cell Culture, and Synchronization-The conditions for cell culture, selection, and DNA transfections have been described previously (32,33). Generation of RAD18 Ϫ/Ϫ cells is described in our earlier report (27). XPA-deficient cells were made so as to delete amino acids sequences Asp 125 -Gln 195 . DT40 cells have only a single XPA gene, which is located on a sex chromosome. The strategy for XPA gene disruption is described in detail on our web site (www.rg.med.kyoto-u. ac.jp/index-e.html). Mutant DT40 cells described in this article, pol (POLK /Ϫ ), rad18 (RAD18 Ϫ/Ϫ ), rad18 pol (RAD18 Ϫ/Ϫ POLK /Ϫ ), and xpa pol (XPA /Ϫ POLK /Ϫ ) cells, and stable transfectants of pol cells with the pcDB3, phPK-G, and p176 plasmids (designated as pol ϩcPol, pol ϩhPol, and pol ϩGFP cells, respectively) are deposited at the RIKEN Cell Bank (RIKEN, Wako, Japan). Cell synchronization at the G 1 /S phase transition was achieved by sequential nocodazole and mimosine blocks as described previously (34).
Flow Cytometric Analysis of Cell Cycle Progression and GFP Expression-To analyze the cell cycle, 1 ϫ 10 6 cells were exposed for 10 min to 20 M 5-bromo-2Ј-deoxyuridine (BrdUrd; Nacalai Tesque, Kyoto, Japan) and then harvested and fixed with 70% ethanol. Fixed cells were incubated with 2 M HCl containing Triton X-100 at 0.5%, treated with mouse anti-BrdUrd monoclonal antibody (BD PharMingen, San Diego, CA) and subsequently with FITC-conjugated anti-mouse IgG antibody (Southern Biotechnology Associates, Birmingham, AL). The cells were resuspended in phosphate-buffered saline containing propidium iodide (PI) at 5 g/ml for subsequent analysis with FACScaliber (BD Biosciences). To assess GFP expression, ϳ1 ϫ 10 6 cells were harvested and resuspended in 0.5 ml of phosphate-buffered saline for subsequent analysis with FACScaliber.
Colony Formation Assay-Serially diluted cells were plated in trip-licate onto 6-well plates with 5 ml/well of 1.5% (w/v) methylcellulose (Aldrich, Milwaukee, WI) containing Dulbecco's modified Eagle's medium/F-12 (Invitrogen Life Technologies, Inc.), 15% fetal calf serum (Equitech-Bio, Ingram, TX), 1.5% chicken serum (Sigma), and 10 M ␤-mercaptoethanol, and were incubated at 39.5°C for 6 -7 days. To analyze sensitivities to IR (ionizing radiation; ␥-ray), cells were plated into medium containing methylcellulose for a 1-h incubation at 39.5°C, and subsequently irradiated with a 137 Cs ␥-ray source (GammaCell 40E; Nordion International, Kanata, Ontario, Canada). For exposure of cells to UV, 3 ϫ 10 5 cells were suspended in 0.5 ml of phosphate-buffered saline containing 1% fetal calf serum, introduced onto 6-well cluster plates and irradiated with UVC (wavelength, 254 nm), followed by addition of 1 ml of complete medium. For exposure of cells to cisplatin (CDDP, cis-platinum (II) diaminodichloride; Nihon-Kayaku, Tokyo, Japan), 1 ϫ 10 5 cells were treated at 39.5°C in 1 ml of complete medium containing CDDP for 1 h. Exposure of cells to methyl methanesulfonate (MMS; Nacalai Tesque, Kyoto, Japan) was performed in medium containing no serum for 1 h at 39.5°C. Exposure of cells to AAF (Nacalai Tesque, Kyoto, Japan) or BaP (Nacalai Tesque, Kyoto, Japan) was performed at 1 ϫ 10 5 cells/ml in a 5:1 mixture of serum-free medium and S-9 mix liver extract (Oriental Yeast Co, Tokyo, Japan) for 1 h at 39.5°C. Analyses of Chromosome Aberrations and Sister Chromatid Exchange Events-Measurement of chromosome aberrations and sister chromatid exchanges (SCEs) was carried out as previously described (35,36) with the following modification for cell preparation. To analyze UV-induced chromosomal aberrations, 1 ϫ 10 6 asynchronous populations of cells were suspended in 0.5 ml of phosphate-buffered saline containing 1% fetal calf serum, spread onto 6-well plates, and exposed to 5 J/m 2 UVC, followed by addition of 10 ml of complete medium. Cells were harvested at every 3 h for 12 h. The cells were treated with 0.1 g/ml of colcemid (Invitrogen Life Technologies, Inc.) for the final 3 h of the incubation to enrich for mitotic cells. To measure SCEs induced by 4-nitroquinoline 1-oxide (4-NQO; Nacalai-Tesque, Kyoto, Japan), cells were treated with 0.2 ng/ml of 4-NQO for the last 8 h of incubation with 10 M BrdUrd for 18 h, which corresponds to two cell cycle periods of DT40 cells. The cells were also treated with 0.1 g/ml of colcemid for the final 2 h of the incubation to enrich for mitotic cells.

Isolation of Chicken POLK Gene and Generation of pol
Mutants-We isolated chicken POLK cDNA, which encodes a protein of 867 amino acids with a predicted molecular mass of 97 kDa (see Supplemental Data). The chicken Pol protein shows ϳ60% amino acid identity to the human and mouse Pol overall. Amino acid sequence conservation is significant particularly within the five conserved motifs (37) of the DNA polymerase Y family, with 90% identity to those of human and mouse Pol, and ϳ60% to those of E. coli DinB (38). Additionally, residues Asp 108 , Asp 199 , and Glu 200 in the polymerase motifs are perfectly conserved with strict intervals from bacterial DinB to the vertebrate Pol proteins. FISH analysis indicated that the chicken POLK gene is located on the sex chromosome Z (data not shown). DT40 cells carry the ZW sex chromosome (35) and carry only a single POLK gene.
Gene targeting constructs were generated to delete amino acids 96 -209 including all the DNA polymerase motifs. Targeting events were verified by the appearance of a 3.7-kb band and the disappearance of an 8.5-kb band in Southern blot analysis of EcoRI-digested genomic DNA (Fig. 1, A and B). We isolated three POLK-disrupted cell clones (D12, D32, and D33), which proliferated with the same kinetics as did wild-type cells (Fig. 1, C and D). These mutant clones are designated pol cells. Flow cytometric analysis showed a normal cell cycle distribution of an asynchronous population of pol cells (Fig. 1D).
Increased Sensitivity of pol Cells to UV in Early S Phase-To analyze the DNA repair capacity of pol cells, we examined viability of cells after various genotoxic treatments using colony survival assays. All three pol clones consistently showed a ϳ1.8-fold increase in UV sensitivity as judged by the evaluation of the doses that reduce survival to 1% (Fig. 2A).
The pol clones showed no significant increase in sensitivities to IR, cisplatin, or MMS in comparison to wild-type cells (Fig. 2,  B-D). Moreover, we were unable to detect significant increases in sensitivities of pol cells to AAF or BaP (Fig. 2, E and F).
To confirm that the elevated UV sensitivity is caused by defective Pol, the pol cells were reconstituted with the chicken or human Pol cDNA. We found that the reconstitution did restore the tolerance to UV to that of wild-type cells ( Fig.  2A). We next examined UV sensitivity of pol cells at each cell cycle phase, comparing with that of wild-type cells. To this end, the cells were synchronized at the G 1 /S boundary using sequential nocodazole and mimosine blocks (34). After the release from the G 1 /S block, cells were exposed to UV at a number of time points including early S phase (1.5 h after the release), late S phase (4.5 h after the release), and the G 2 /M phase (6 h after the release) (Fig. 3A). Wild-type cells showed increased resistance to UV as DNA replication progressed, reflecting lower frequencies of replication blockage caused by UV-induced damage at late S phase compared with early S phase. In contrast, only at early S phase, the UV sensitivity of pol cells was significantly greater than that of wild-type cells. The early S phase-specific increase in UV sensitivity of pol cells is in agreement with defective TLS in the mutant cells.
UV-induced damage is expected to block DNA replication, leading to gaps and double-strand breaks (DSBs) in sister chromatids. To evaluate UV-induced DSBs, we measured cytologically detectable chromosomal aberrations following UV irradiation in asynchronous populations of pol cells (Fig. 4). The cell cycle progression following UV irradiation is comparable between wild-type and pol cells (data not shown). Cells irradiated at late S to G 2 phase are expected to enter the M phase within 3 h following UV irradiation, whereas cells irradiated at G 1 to early S phase may need 3-6 h to enter the M phase (39). We found no increase in the level of spontaneous chromosomal aberrations in pol cells when compared with wild-type cells (Fig. 4), as expected from the normal growth property of pol cells (Fig. 1, C and D). The levels of UV-induced chromosomal aberrations in pol cells were up to 3-fold higher than those of wild-type cells in each 3-h period following UV irradiation (Fig.  4), indicating that the mutants are not able to deal with replication block as efficiently as are wild-type cells. It should be noted that in DT40 cells, UV irradiation induces both chromosome-type breaks, where a pair of sister chromatids are broken at the same site, and chromatid-type breaks, where only one sister chromatid is broken, although it is known that UV causes only chromatid-type breaks in other vertebrate cells (40). Thus, we may have to take into account DT40-specific mechanisms for the induction of chromosome-type breaks, such as efficient interactions between damaged chromatids and other intact ones through HR.
Defective Postreplication Repair Following UV Irradiation in pol Cells-There are three major pathways for repairing DNA damage induced by UV, i.e. the NER, HR, and TLS pathways (1). It is not entirely clear which DNA polymerases are involved in the NER or HR pathways in vertebrate cells. To analyze the epistatic relationship of Pol to the NER pathway, we generated cells deficient in the XPA protein (xpa cells) as well as cells deficient in both XPA and Pol (xpa pol cells). The XPA protein is involved in an early step of NER, and its defect causes hypersensitivity to UV and skin cancer. We found that both xpa and xpa pol cells proliferated with normal kinetics (data not shown). Their UV sensitivity showed an additive effect of the mutations of XPA and Pol on UV sensitivity (Fig.  2G), suggesting that Pol does not function in the NER pathway. We also examined the involvement of Pol in the HR pathway by assessing HR capability in pol cells. Analyses of  gene-targeting efficiencies (Table I) and IR sensitivity (Fig. 2B) consistently showed that HR capability is not reduced in pol cells. Thus, defective HR does not account for increased UV sensitivity in pol cells. In summary, the increased UV sensitivity of pol cells is most likely to be explained by a defect in TLS, rather than a defect in NER or HR. Increased Levels of SCE, an HR-mediated Repair Process in pol Cells-We have shown that SCE, cytogenetically detectable crossover, reflects postreplication repair mediated by HR (36,41). Interestingly, pol cells exhibited a significant increase in the level of spontaneous SCEs (Fig. 5). This observation suggests that Pol plays a role in maintaining chromosomal DNA during the cell cycle. To analyze SCE induced by genotoxic treatments, we exposed cells to 4-NQO, which is known to form DNA adducts similar to UV damage and to stimulate SCE (1). The level of induced SCE events was also higher in pol cells when compared with wild-type cells (Fig. 5). These observations support the notion that defective TLS in pol cells is compensated by HR, the other major postreplicational repair pathway.
Pol-dependent TLS Can Function in the Absence of Rad18 -To examine the epistatic relationship between Pol and Rad18, we generated cells deficient in both Rad18 and Pol (rad18 pol cells). We have previously shown that rad18 cells are sensitive to various DNA damaging agents including IR, UV, and CDDP (27); consistent with the phenotype of the yeast rad18 mutant. Interestingly, defects in both Rad18 and Pol have an additive effect on UV sensitivity (Fig. 2H), indicating that Pol may work even in the absence of the Rad18 protein in vertebrate cells. DISCUSSION Here we report the phenotypic analysis of pol cells. Remarkably, the mutant cells showed elevated UV sensitivity. We postulate that the elevated UV sensitivity in pol cells indeed reflects an important role for Pol in processing UV damage from the following data. Colony survival assays showed that depletion of Pol in wild-type and rad18 cells both increased their UV sensitivity ( Fig. 2A). Likewise, pol cells exhibited increased levels of UV-induced chromosomal aberrations when compared with wild-type cells (Fig. 4). The elevated UV sensitivity of pol cells is not likely to be attributed to chickenspecific characteristics, because both the human and chicken POLK gene normalized UV sensitivity of Pol-deficient DT40 cells ( Fig. 2A).
The discrepancy between this in vivo data and no detectable activity of purified Pol of copying UV damage could be explained as followings. First, experimental conditions used for the biochemical studies of Pol are not optimal for evaluating the function of Pol in vivo. For example, unknown cofactors required for Pol in the bypass of UV damage may have been absent from in vitro reactions. Second, Pol may be involved in an extension step following nucleotide incorporation of a damaged template DNA, based on the two step-two polymerase hypothesis proposed by several groups (Refs. 11, 42, and 43; reviewed in Refs. 44 and 45). A recent biochemical study shows that human Pol is very efficient at extending from a G opposite the 3Ј-T of a T-T dimer (46). This observation as well as our genetic study suggests that Pol can play a significant role in TLS of UV lesions.
pol DT40 cells showed elevated UV sensitivity, but did not show elevated sensitivity to AAF or BaP, although Pol can replicate a template strand containing these adducts in in vitro reactions (10,12,17,19,20). Presumably, this is explained as follows: Several TLS polymerases have redundancy in bypassing a given DNA damage (47), and the relative contribution of each TLS polymerase to processing a given type of DNA lesion may be distinctly different between each cell line and tissue. Thus, the involvement of Pol in bypassing DNA damage caused by AAF and BaP should be analyzed using DT40 mutants deficient in both Pol and another TLS polymerase.
It is interesting that defects in Rad18 and Pol showed an additive effect on UV sensitivity. The data suggest that Pol participates in Rad18-independent TLS of UV lesions and that Rad18 does not necessarily regulate all TLS polymerases in vertebrate cells. Recent genetic studies of S. cerevisiae also suggested that the existence of Rad18-independent TLS of UV lesion (48).
Pol Contributes to the Maintenance of Chromosomal DNA in Cooperation with HR Pathway-A significant increase in the level of SCE was observed in pol cells not only after 4-NQO treatment but also during the cell cycle. Likewise, the level of spontaneously arising SCE events is increased in cells from XP-V patients (49). We interpret these increased levels of SCE as more frequent usage of the HR pathway for postreplicational repair. Presumably, gaps that should be filled by Pol may accumulate in pol cells and stimulate HR with the other sister chromatid, leading to an increase in SCE events. Importantly, increased levels of spontaneous SCE in pol cells imply the functional cooperation between the Pol-dependent TLS and HR pathways in the maintenance of chromosomal DNA. This conclusion leads to the notion that the relative usage of POLK in postreplication repair may determine mutation frequency in cycling cells, since Pol-dependent TLS is highly mutagenic (16). Recent studies showed that the expression of mammalian POLK is inducible by environmental mutagens, BaP and dioxin (50), and that Pol is overexpressed in human non-small cell lung cancer. Thus, these studies and our study shed light on the understanding of mutagenesis and carcinogenesis.