Xeroderma Pigmentosum Group A Suppresses Mutagenesis Caused by Clustered Oxidative DNA Adducts in the Human Genome

Clustered DNA damage is defined as multiple sites of DNA damage within one or two helical turns of the duplex DNA. This complex damage is often formed by exposure of the genome to ionizing radiation and is difficult to repair. The mutagenic potential and repair mechanisms of clustered DNA damage in human cells remain to be elucidated. In this study, we investigated the involvement of nucleotide excision repair (NER) in clustered oxidative DNA adducts. To identify the in vivo protective roles of NER, we established a human cell line lacking the NER gene xeroderma pigmentosum group A (XPA). XPA knockout (KO) cells were generated from TSCER122 cells derived from the human lymphoblastoid TK6 cell line. To analyze the mutagenic events in DNA adducts in vivo, we previously employed a system of tracing DNA adducts in the targeted mutagenesis (TATAM), in which DNA adducts were site-specifically introduced into intron 4 of thymidine kinase genes. Using the TATAM system, one or two tandem 7,8-dihydro-8-oxoguanine (8-oxoG) adducts were introduced into the genomes of TSCER122 or XPA KO cells. In XPA KO cells, the proportion of mutants induced by a single 8-oxoG (7.6%) was comparable with that in TSCER122 cells (8.1%). In contrast, the lack of XPA significantly enhanced the mutant proportion of tandem 8-oxoG in the transcribed strand (12%) compared with that in TSCER122 cells (7.4%) but not in the non-transcribed strand (12% and 11% in XPA KO and TSCER122 cells, respectively). By sequencing the tandem 8-oxoG-integrated loci in the transcribed strand, we found that the proportion of tandem mutations was markedly increased in XPA KO cells. These results indicate that NER is involved in repairing clustered DNA adducts in the transcribed strand in vivo.


Introduction
Genomic DNA is constantly exposed to both exogenous and endogenous genotoxic agents. Among them, ionizing radiation (IR) induces various DNA adducts in the genome because of inactivated horse serum (JRH Biosciences), 200 μg/ml sodium pyruvate, 100 U/ml penicillin, and 100 μg/ml streptomycin at 37°C in an atmosphere of 5% CO 2 and 100% humidity.

Construction of XPA knockout cells
We purchased custom zinc finger nuclease (ZFN) for targeting XPA from Sigma-Aldrich. The design, assembly, and validation of ZFN were performed using the CompoZr1 Custom ZFN Service (Sigma-Aldrich). The target sequence for ZFN, that is located in exon 1 of the XPA gene, was as follows: 5 0 -CAGGCCCGGCTGGCTGCCCggcccTACTCGGCGACGGCGGCT-3 0 . ZFN mRNA (2 μg) was transfected into TSCER122 (5 × 10 6 ) cells supplemented with 0.1 ml Nucleofector solution V (Lonza) using Nucleofector I according to the manufacturer's recommendations. After 24 h in culturing medium, cells were seeded into 96-microwell plates at 1.6 cells/well (i.e., 8 cells/ml) and then incubated at 37°C for 7-10 days. Genomic DNA was isolated from colonies and subjected to PCR using KOD FX (Toyobo) with the primers 5 0 -AGCT AGGTCCTCGGAGTGG-3 0 and 5 0 -GGACAGGACGCTTTGACAAG-3 0 . The amplified DNA fragment was then sequenced to confirm deletion around the ZFN target site.

Irradiation conditions
Prior to irradiation with 254 nm ultraviolet (UVC), cells were washed and resuspended with RPMI 1640 without phenol red. Cells (2.5 × 10 6 ) in 5 ml medium were exposed to UVC in a 35 × 100 mm petri dish using a UV germicidal lamp at a fluence of 2.7 W/m 2 , which was monitored by a UV-light meter (Lutron). After irradiation, cells were seeded into 96-microwell plates at 1.6 cells/well (i.e., 8 cells/ml) to determine cell survival.
PCR-based preparation of a site-specific modified targeting vector containing 8-oxoG adducts Targeting vectors pvIT 1x8oG , pvIT 2x8oG , and pvINT 2x8oG , carrying a single 8-oxoG in TS, two 8-oxoG in TS, and two 8-oxoG in NTS, respectively; the control vector pvIT G without DNA adduct were prepared by a PCR-based method with the plasmid pTK15 as previously described [30,32]. The targeting vector comprised 6.1 kb of the four individual TK genes, encompassing exons 5-7, and part of the I-SceI sequence in intron 4, which carries a loss-of-function TTAT deletion. We inserted an 8-oxoG into pvIT 1x8oG in place of the underlined guanine at the BssSI site in TS (5 0 -CACGAG), two 8-oxoG into pvIT 2x8oG in TS (5 0 -CACGAG), and two 8-oxoG into pvINT 2x8oG in NTS (5 0 -CTCGTG) (S1A-S1C Fig). We labeled a 5 0 -TTCA-sequence (MseI R ) near the 8-oxoG-modified BssSI site that was resistant to MseI digestion and thereby distinguished between targeted and non-targeted revertants of TK according to interallelic recombination (Fig 1). The vectors were sequenced to confirm 8-oxoG presence at the expected site. Overview of the TATAM system. Principle of the TATAM system. The targeting vector pvIT G , pvIT 1x8oG , pvIT 2x8oG , or pvINT 2x8oG and the I-SceI expression plasmid pCBASce were co-transfected into TSCER122 cells. DSB at the I-SceI site enabled high gene-targeting efficiency for the TATAM system by inducing DSB repair-enhanced site-specific homologous recombination. The targeting vector contained a MseI R site that was resistant to MseI digestion and thereby distinguished between targeted and non-targeted revertants of TK. Genomic DNAs of the revertants were prepared, and part of the TK gene containing the DNA adduct-integrated site was amplified by PCR. The amplified fragment was sequenced as described in Materials and Methods. Transfection and cloning of TK revertant cells Prior to transfection, the concentrations of targeting vectors were determined with a Qubit1 2.0 Fluorometer with Qubit1 dsDNA HS Assay Kit (Life technologies). The targeting vector (1 μg) and I-SceI expression plasmid pCBASce (50 μg) were co-transfected into 5 × 10 6 cells that were suspended in 0.1 ml Nucleofector Solution V using Nucleofector I according to the manufacturer's recommendations [33]. After 72 h of incubation, cells were seeded into 96-microwell plates in the presence of 200 μM hypoxanthine, 0.1 μM aminopterin, and 17.5 μM thymidine (HAT) in order to isolate 8-oxoG-integrated revertant clones. After 2 weeks of incubation, drug-resistant colonies (TK revertants) were counted.

Mutation analysis
Genomic DNA templates for PCR were prepared from TK-revertant colonies using the alkaline lysis method according to a protocol provided by Toyobo Co., Ltd. Cells were incubated with 18 μl of 50 mM NaOH at 95°C for 10 min and then neutralized with the addition of 2 μl of 1 M Tris-HCl (pH 8.0). The resulting cell lysates were used as templates for PCR to amplify the TK gene fragments containing the 8-oxoG-integration site. PCR amplification was performed using KOD FX with forward and reverse primers: (Pri#291, intron 4), 5 0 -GCT CTT ACG GAA AAG GAA ACA GG-3 0 ; (Pri#292, intron 5), 5 0 -CTG ATT CAC AAG CAC TGA AG-3 0 , respectively, as previously described [30]. Genomic regions around the BssSI and MseI R sites were sequenced using an ABI 3730xl DNA analyzer (Applied Biosystems) and clones harboring the MseI R sequence were counted to determine the frequency of 8-oxoG integration and numbers of mutations at 12-bp (5 0 -TCC CAC GAG GCT) around the BssSI site. The integration frequency of 8-oxoG adducts in the targeting vector was calculated by dividing the number of MseI R clones by the total number of revertant clones analyzed. Single-point mutation was defined as a single-base substitution, one-base insertion, or one-base deletion detected at an 8-oxoG. Tandem mutations were multiple-base substitutions, deletions, and/or insertions that were detected at sites, including 8-oxoG. Base substitutions, deletions, and/or insertions found at sites other than 8-oxoG were defined as non-targeted. Mutant proportions were calculated by dividing the number of mutants by the number of MseI R -bearing clones.

Complementation assay
For the complementation assay, the coding region of XPA was amplified from cDNA of TSCER122 cells by PCR with primers 5 0 -AGCTAGGTCCTCGGAGTGG-3 0 and 5 0 -TTTTGA ATTTTGAAAAGGACCAA-3 0 . The resulting fragment was further amplified by PCR with primers 5 0 -TGGGAATTCGCCACCATGGCGGCGGCCGA-3 0 and 5 0 -TTTCAGTCGACTC ACATTTTTTCATATGTCAGTT-3 0 . The PCR product was digested with EcoRI and SalI and was cloned into the EcoRI/SalI sites of pCl-neo vector (Promega). The obtained plasmid was named pCl-XPA, which contained XPA cDNA. XPA KO (5 × 10 6 ) cells were transfected with 10 μg of BamHI-linearlized pCl-XPA using Nucleofector I. After 48 h in the culturing medium, cells were seeded into 96-microwell plates in the presence of 1 mg/ml G418 (Roche). G418-resistant clones were then isolated and the expression of XPA was confirmed by Western blot analysis. The XPA-expressing clone was subjected to UV-irradiation and TATAM analyses.

Statistical analysis
Statistical significance was evaluated by Student's t-test or two-tailed Fisher's exact test; P < 0.05 was considered to be significant.

Establishment of XPA knockout cells
The XPA gene is located on chromosome 9 and contains six exons. To knock out XPA in human lymphoblastoid TSCER122 cells, a customized ZFN was designed to target exon 1 of XPA. The target sequence of ZFN is shown in Fig 2A. After transfecting ZFN mRNA into TSCER122, we obtained one homozygous disrupted clone. The resulting XPA KO cells had a 308-bp deletion in exon 1 of both the alleles (Fig 2A). The protein expression was evaluated by Western blot analysis with anti-XPA monoclonal antibody (Fig 2B), showing that XPA protein was expressed in TSCER122 but not in XPA KO cells. OGG1, APE1, and Pol β were also detected in both TSCER122 and XPA KO cells, which indicates that the disruption of XPA did not significantly alter the expression level of BER components. XPA protects cells against UVinduced cytotoxicity [34]; accordingly, we compared the sensitivity of TSCER122 and XPA KO cells to UV irradiation. As shown in Fig 2C, XPA KO cells were hypersensitive to UV irradiation. Such hypersensitivity against UV was rescued by expressing the wild-type XPA in XPA KO (XPA KO + pCl-XPA) cells. The population doubling time of XPA KO cells (13 ± 0.083 h) was similar to that of TSCER122 cells (13 ± 0.047 h), indicating that XPA knockout did not markedly influence the proliferation rate of TSCER122 cells.

Mutagenic events induced by oxidative DNA adducts in TSCER122 cells
We prepared targeting vectors pvIT G , pvIT 1x8oG , pvIT 2x8oG , and pvINT 2x8oG , which carried no DNA adduct, a single 8-oxoG in the TS, tandem 8-oxoG in the TS, and tandem 8-oxoG in the NTS, respectively (refer to Materials and Methods and S1A-S1C Fig) as previously reported [30]. An overview of the TATAM system is shown in Fig 1. First, the mutant proportion induced by integration of each targeting vector was determined in TSCER122 cells. As shown in the eighth column of Table 1, the total proportion of mutants induced by pvIT 1x8oG was 8.1%, which was approximately 6-fold higher than that induced by the control vector pvIT G (1.3%). The total proportion of mutants induced by pvIT 2x8oG (7.4%) was comparable with that induced by pvIT 1x8oG , although the number of 8-oxoG in pvIT 2x8oG was twice that in pvIT 1x8oG . In contrast, the mutant proportion induced by pvINT 2x8oG (11%) was somewhat higher than that induced by pvIT 2x8oG . This was an unexpected result because the mutant proportion by a single 8-oxoG has been reported to be similar between TS and NTS [30].
Effect of knockout of XPA on mutagenic events induced by a single 8-oxoG Next, we analyzed the effect of XPA knockout on the mutant proportions induced by integration of unmodified vector or a single 8-oxoG. First, the mutant proportion induced by pvIT G integration was not altered by XPA disruption (1.3 ± 0.061 and 1.3 ± 0.043% in TSCER122 and XPA KO cells, respectively) ( Fig 3A). Second, as shown in Fig 3B, the proportion of mutants associated with pvIT 1x8oG integration was also similar between TSCER122 and XPA KO cells (8.3 ± 0.93 and 7.6 ± 0.016%, respectively). In addition, we analyzed mutation spectra that was induced by an 8-oxoG. As shown in the fifth column of Table 1, the mutation most predominantly observed was GÁC to TÁA transversion at the 8-oxoG site in both TSCER122 and XPA KO cells. Smaller numbers of GÁC to CÁG transversions, GÁC to AÁT transitions, and one-base deletions at the 8-oxoG position and "non-targeted" mutations, which were detected at sites other than 8-oxoG, were also observed in both the cells, which were similar to results of earlier studies [30]. No significant increase in tandem mutations was observed in XPA KO cells (sixth  Table 1 and Table 2). Overall, XPA disruption did not markedly affect mutagenesis that was induced by an isolated 8-oxoG in the genome.
Effect of the knockout of XPA on mutagenic events induced by tandem 8-oxoG in the transcribed or non-transcribed strand Furthermore, we investigated the effect of XPA knockout on mutations caused by tandem 8-oxoG integrated in either TS or NTS. The mutant proportion induced by pvIT 2x8oG integration was significantly higher in XPA KO cells (12 ± 1.6%) than in TSCER122 cells (7.4 ± 0.59%) (Fig 3C). In contrast, the proportion of mutants that were induced by pvINT 2x8oG integration was not increased in XPA KO cells (11 ± 0.83%) compared with TSCER122 cells (11 ± 1.7%) (Fig 3D). Thus, XPA deficiency enhanced clustered 8-oxoG-induced mutagenesis in TS but not in NTS.
We then analyzed the mutation spectra that was induced by tandem 8-oxoG integrated in either TS or NTS. As shown in the fifth column of Table 1, the most frequent mutation was a single GÁC to TÁA transversion at one of the 8-oxoG positions in either TS or NTS, in both TSCER122 and XPA KO cells. Lesser proportions of single GÁC to CÁG transversions, GÁC to AÁT transitions, or one-base insertions/deletions were also detected irrespective of XPA expression. Interestingly, the proportion of tandem mutations was higher in XPA KO cells (3.0%)   Table 2. Spectra of single and tandem mutations induced by integration of pvIT 1x8oG , pvIT 2x8oG , and pvINT 2x8oG .
pvIT 1x8oG pvIT 2x8oG pvINT 2x8oG than in TSCER122 cells (1.2%) in TS (Fisher's exact test, P < 0.05), although that proportion was comparable between TSCER122 and XPA KO cells in NTS (2.0% and 2.6%, respectively). As shown in Table 2, the tandem mutations most predominantly observed were GNG to TNT in both TSCER122 (0.46 and 1.4% in TS and NTS, respectively) and XPA KO cells (1.1 and 1.1% in TS and NTS, respectively). Smaller numbers of different types of tandem mutations were also detected. Unexpectedly, the proportion of non-targeted mutations was significantly higher in XPA KO cells (2.8%) than in TSCER122 cells (1.1%) in TS (Fisher's exact test, P < 0.05) (seventh column of Table 1 and S1 Table). Notably, the proportion of non-targeted mutations was not significantly altered between TSCER122 and XPA KO cells when pvIT G , pvIT 1x8oG , or pvINT 2x8oG was integrated into the genome.
To verify that XPA was actually involved in the suppression of tandem 8-oxoG-induced mutations in TS, the mutant proportion induced by pvIT 2x8oG integration was examined in XPA KO + pCl-XPA cells. As shown in the eighth column of Table 1, the total mutant proportion was comparable between TSCER122 and XPA KO + pCl-XPA cells (7.4% and 7.4%, respectively). Similarly, the proportions of tandem mutations and non-targeted mutations in XPA KO + pCl-XPA cells (1.5% and 1.5%, respectively) were as low as those detected in TSCER122 cells (1.2% and 1.1%, respectively) ( Table 2 and S1 Table).
The targeting efficiencies of the vectors to XPA KO cells were similar to those to TSCER122 cells (data not shown), suggesting that XPA disruption did not influence homologous recombination efficiency.

Discussion
Because clustered DNA adducts are poorly repairable by BER [8][9][10][11]35], we postulated that an alternative repair machinery is involved in the removal of DNA adducts in vivo. To investigate the role of NER on the oxidative DNA damage-induced mutagenesis, XPA can be a good candidate gene to be examined without altering the expression of BER proteins (Fig 2B), whereas some other NER components change the status of OGG1 [36,37]. Loss of XPA does not sensitize cells to IR [38]. XPA disruption in TSCER122 cells does not alter cell survival after IR irradiation (data not shown). However, decreased XPA expression results in increased IR-induced chromosomal aberrations levels [39], implying the importance of XPA-mediated NER in  repairing IR-induced DNA damage. However, the biological relevance of NER and IR-induced mutagenesis has remained unclear. In this study, we present evidence that NER plays a substantial role in the suppression of mutations that were induced by clustered oxidative DNA adducts in human cells.
Our results indicate that NER is not crucial for suppressing mutations caused by a single 8-oxoG (Fig 3B). This observation may be because of the abundant repair activities of BER proteins, such as OGG1, MYH [30], and Nei endonuclease VIII-like glycosylases [40], to remove 8-oxoG in cells. Alternatively, the small helix distortion due to a single 8-oxoG may not be enough for efficient recognition by NER [41].
The observation that NER disruption significantly increased the proportion of mutants that were induced by tandem 8-oxoG in TS but not in NTS (Fig 3C and 3D) is consistent with the results of a previous study that revealed transcription coupled repair of 8-oxoG [28]. In fact, in TSCER122 cells, the mutant proportion of tandem 8-oxoG in NTS was higher than that in TS (sixth column of Table 1). This enhanced mutagenesis of tandem 8-oxoG is probably due to delayed BER at the tandem DNA adducts [11]. In addition, at the site of clustered DNA adducts, the binding of DNA glycosylase on one DNA adduct inhibits repair of the neighboring damage [35].
NER is most likely the most efficient mechanism to repair clustered DNA adducts because of its capability to remove all adducts at once [42]. Furthermore, it is plausible that NER removes clustered 8-oxoG by processing the BER intermediate [43]; once DNA glycosylase cleaves one of the 8-oxoG at the tandem site, a resulting AP site can initiate NER-mediated repair of clustered DNA damage. Obviously, NER proteins recognize 8-oxoG and also facilitate tandem DNA adducts containing an AP-site analog in vitro [27,42], indicating that both 8-oxoG and its BER intermediate in the clustered DNA damage attract NER. Moreover, the unique structural changes within the DNA helix induced by complex damage may contribute to the damage recognition by NER [44].
As shown in the sixth and seventh column of Table 1, the enhanced mutagenesis of tandem 8-oxoG in TS in XPA KO cells is because of the increased proportions not only of tandem mutations but also of non-targeted mutations, which are induced near the adduct position. Although the exact reason for the increased proportion of non-targeted mutations in XPA KO cells remains unclear, it may be because of stalled replication or retarded repair of sites of clustered DNA adducts in XPA absence.
It is noteworthy that IR produces a variety of DNA adducts, which are formed in close proximity in one or both DNA strands. NER may be involved in repairing different types of clustered DNA damage, which can lead to slow or stalled BER, thus protecting the genome from the lethal strand break formation. Further studies that examine how NER contributes to repair of different kinds of clustered DNA damage will help in understanding its functions in IRinduced DNA damage.
We conclude that NER suppresses mutations induced by clustered oxidative DNA adducts. To the best of our knowledge, this is the first report providing the in vivo evidence that NER is involved in the repair of clustered oxidative DNA damages. This role for NER may contribute to the oxidative stress leading to neurodegeneration in patients with XPA [45].  Table. The spectra of non-targeted mutations induced by integration of pvIT G , pvIT 1x8oG , pvIT 2x8oG , and pvINT 2x8oG . a Underlined sequences are observed mutations. "Δ" indicates onebase deletion. del., deletion; ins., insertion. b "8" in the sequence indicates 8-oxoG. c XPA KO + pCl-XPA cells. (XLSX)