Processing of a single ribonucleotide embedded into DNA by human nucleotide excision repair and DNA polymerase η

DNA polymerases often incorporate non-canonical nucleotide, i.e., ribonucleoside triphosphates into the genomic DNA. Aberrant accumulation of ribonucleotides in the genome causes various cellular abnormalities. Here, we show the possible role of human nucleotide excision repair (NER) and DNA polymerase η (Pol η) in processing of a single ribonucleotide embedded into DNA. We found that the reconstituted NER system can excise the oxidized ribonucleotide on the plasmid DNA. Taken together with the evidence that Pol η accurately bypasses a ribonucleotide, i.e., riboguanosine (rG) or its oxidized derivative (8-oxo-rG) in vitro, we further assessed the mutagenic potential of the embedded ribonucleotide in human cells lacking NER or Pol η. A single rG on the supF reporter gene predominantly induced large deletion mutations. An embedded 8-oxo-rG caused base substitution mutations at the 3′-neighboring base rather than large deletions in wild-type cells. The disruption of XPA, an essential factor for NER, or Pol η leads to the increased mutant frequency of 8-oxo-rG. Furthermore, the frequency of 8-oxo-rG-mediated large deletions was increased by the loss of Pol η, but not XPA. Collectively, our results suggest that base oxidation of the embedded ribonucleotide enables processing of the ribonucleotide via alternative DNA repair and damage tolerance pathways.

no such mutations in the genome of tumor cells derived from RNase H2 deficient mice 5 . The mutagenic potential of ribonucleotides embedded into DNA in mammalian cells remains to be elucidated. Cellular DNA/RNA and its precursors are subject to oxidation from endogenous and exogenous sources of reactive oxygen species. The exposure of DNA to oxidative stress has been shown to promote the formation of a substantial amount of ribonucleotides and oxidized base damages in vitro and in vivo 16 . 7,8-Dihydro-8-oxo-riboguanosine-5′-triphosphate , the oxidized form of rNTPs, can be generated in the nucleotide pool by the action of oxygen radicals 17 . 8-Oxo-rGTP is also utilized as the substrate during DNA synthesis by Pols [18][19][20] . Thus, the modified ribonucleotides could exist in DNA via the oxidation of the embedded ribonucleotide and/or the insertion of damaged rNTPs during DNA replication. Base oxidation in embedded ribonucleotide inhibits RNase H2-mediated excision repair 21,22 . Furthermore, the base excision repair (BER) pathway, involved in the repair of 7,8-dihydro-8-oxo-deoxyguanosine (8-oxo-dG), cannot excise the oxidized ribonucleotide 7,8-dihydro-8-oxo-riboguanosine (8-oxo-rG) in the DNA 21,22 . Therefore, a modified ribonucleotide would exhibit greater mutagenic potential compared with unmodified ribonucleotide in cells. Otherwise, an alternative repair pathway may participate in the repair of a damaged ribonucleotide. A possible mechanism is nucleotide excision repair (NER) involved in the removal of various helix-distorting DNA lesions from DNA 23 . A previous study has revealed that NER proteins derived from thermophilic eubacteria can excise ribonucleotides embedded into DNA 24 . However, another study has suggested that a ribonucleotide embedded into DNA is a poor substrate for NER function in vitro in both human and Escherichia coli (E. coli) 25 . Based on these findings, the involvement of NER in ribonucleotide removal in cells remains controversial.
Ribonucleotides that escape from repair can interfere with DNA replication. In vitro studies have revealed that single and multiple ribonucleotides in the template DNA influence DNA synthesis mediated by Pols in yeast and humans 2,26 . Furthermore, the presence of 8-oxo-rG in DNA strongly hinders the primer extension reaction catalyzed by human replicative Pol α 21 . To counteract these replication-blocking lesions, organisms possess specialized Pols that carry out translesion DNA synthesis (TLS) past DNA lesions. Pol η is one of TLS Pols that can bypass various DNA damages, including ultraviolet (UV) light-induced cyclobutane pyrimidine dimers 27 . Our in vitro experiments revealed that Pol η efficiently and accurately bypassed undamaged and damaged ribonucleotides (rG and 8-oxo-rG, respectively) in a more error-free manner compared with 8-oxo-dG 21 . Thus, it is of interest to investigate whether TLS Pols exert protective effects against the mutagenicity of incorporated ribonucleotides in cells.
In the current study, we investigated the possibility whether the reconstituted NER system can excise the ribonucleotide from the DNA. On the basis of in vitro analyses, we also examined the mutagenic potential of an embedded ribonucleotide in human lymphoblastoid cells lacking xeroderma pigmentosum group A (XPA), an essential factor for NER, and Pol η using a supF shuttle vector containing a single dG, rG, 8-oxo-rG, or 8-oxo-dG. Based on these biochemical and cell culture investigations, our findings provide new mechanistic insights underlying ribonucleotide-induced mutagenesis and the protective roles of DNA repair and tolerance processes in mammals.

Results
Reconstituted neR can excise an embedded oxidized ribonucleotide. To test the possibility that NER is involved in the ribonucleotide removal from DNA, we analyzed whether the human cell-free NER system reconstituted with purified protein factors possesses the ribonucleotide excision activity against a single rG or 8-oxo-rG in the plasmid DNA (Fig. 1a). As shown in Fig. 1b, single rG was a poor substrate for human NER, a finding consistent with the previous report 25 . Conversely, 8-oxo-rG was recognized and excised by NER machinery. In addition, the incision activity of NER against 8-oxo-rG was greater than that against 8-oxo-dG (Fig. 1c). The ladder pattern of the excised products was also slightly different between 8-oxo-dG and 8-oxo-rG ( Fig. 1b), which can be due to the difference in the sugar pucker conformation of deoxyribose and ribose that may affect the recognition and excision processes of NER 28 . Because the efficiency of NER in excising the ribonucleotide was lower compared with the canonical substrate for NER, i.e., UV-induced 6-4 photoproducts ( Fig. 1b) ( Supplementary Fig. S1), we further assessed the cellular impact of the NER deficiency on the mutagenic potential of an embedded ribonucleotide in the following section. experimental design to determine the mutagenic potential of an embedded ribonucleotide in human cells. To examine the mutagenic potential of an embedded ribonucleotide in human cells, double-stranded plasmids containing rG or 8-oxo-rG were constructed as the model substrates. Ribonucleotides were inserted at position 112 (G 112 ) on supF (Fig. 2a). The control plasmids harboring dG or 8-oxo-dG at the same position were also synthesized to clarify the effect of the ribose sugar backbone on mutations. The phenotype of cells used in this study was confirmed prior to the experiments; XPA −/− and POLH −/− cells were hypersensitive to UV exposure (Fig. 2b), as previously described 29,30 . The stable expression of XPA and POLH in each knock-out cells, i.e., "XPA −/− + XPA" and "POLH −/− + POLH", rescued the phenotype, which was then utilized for the complementation experiment as described in the following section.
A single ribonucleotide predominantly induces large deletions in human cells. First, the plasmid DNA containing a single rG was transfected and replicated in wild-type (WT), XPA −/− , and POLH −/− cells. Plasmids replicated in cells were then recovered and introduced into KS40/pOF105 indicator strain, which was plated on titer and selection plates to calculate the frequencies of supF mutants. The mutant frequency induced by rG was comparable between the cells (3.9 ± 0.19, 3.6 ± 0.76, and 3.7 ± 0.056 × 10 −3 in WT, XPA −/− , and POLH −/− cells, respectively), which was approximately four-fold higher than the control dG (0.88 ± 0.35, 0.79 ± 0.30, and 0.63 ± 0.12 × 10 −3 in WT, XPA −/− , and POLH −/− cells, respectively) ( Fig. 3a,b). We further analyzed the mutation spectrum induced by rG. Interestingly, large deletions were the predominant mutation (67%, 75%, and 82% in In vitro dual incision assays were performed using purified NER proteins and internally 32 P-labeled DNA substrates containing a 6-4PP, dG, 8-oxo-dG, rG, or 8-oxo-rG as described in the Methods. DNA samples were subjected to 10% denaturing polyacrylamide gel electrophoresis followed by autoradiography. The full-length gel is shown in Supplementary  Fig. S1. (c) Quantification of the products formed by NER-mediated dual incision reaction. Values are presented as mean ± S.E. of three independent experiments. Significant differences are indicated by asterisks (**P < 0.01, ****P < 0.0001 by two-way ANOVA).  Table S1). The length of deletions varied from 10 to 244 bp, with the majority of them being >100 bp in all cell lines (Supplementary Table S2). Notably, no sequence homology was found at each deletion junction. Because the frequency of large deletions observed was only approximately 10% in the control experiments involving dG ( Fig. 4a) (Supplementary Table S1), large deletion mutations were specifically induced by rG embedded into the DNA.
Disruption of XPA or POLH enhances the mutagenic potential of an oxidized ribonucleotide. Next, we assessed whether base oxidation in a ribonucleotide affected its mutagenic potential in cells. The mutant frequency induced by 8-oxo-rG (3.0 ± 0.19 × 10 −3 ) was not higher than that induced by rG (3.9 ± 0.19 × 10 −3 ) and 8-oxo-dG (5.7 ± 0.15 × 10 −3 ) in WT cells (Fig. 3b,c). In contrast with 8-oxo-dG that exclusively induced G:C to T:A transversion mutations at the position G 112 (87%) (Fig. 4d) (Supplementary Table S1), 8-oxo-rG preferentially induced base-substitution mutations at a 3′-base neighboring the 8-oxo-rG site, i.e., GG 112 T to GGG (46%), followed by G:C to T:A (19%) and one-base deletions (14%) at the site of 8-oxo-rG (Fig. 4c). The frequency of large deletions induced by 8-oxo-rG was 7.9%, which was lower than that induced by rG (67%) (Fig. 4b). As some extent of G:C to T:A mutation (18%) was also induced by the control dG (Fig. 4a), mutations specifically induced by 8-oxo-rG were GG 112 T to GGG and one-base deletions, rather than large deletions and G:C to T:A transversion mutations. GG 112 T to GGG mutation was also observed with 8-oxo-dG in WT cells (6.7%) (Fig. 4d), suggesting that this mutation depends on the sequence context around G 112 .

Discussion
The absence of RER leads to the accumulation of ribonucleotides in the genome, which is associated with increased genomic instability 31 . Although several studies have reported the importance of RER for maintaining genome integrity in higher eukaryotes 3,32 , the mutational specificity of embedded ribonucleotides remains largely unclear in mammals. To visualize mutations induced by a ribonucleotide that is readily repaired by RER in the genome, we employed the shuttle vector system that can rapidly propagate in cells. Our results suggest that an embedded rG can cause the deleterious mutations in human cells. On the other hand, its oxidative derivative 8-oxo-rG did not display the enhanced mutagenic potential compared with rG and 8-oxo-dG, which can be explained by the alternative repair and damage tolerance pathways based on in vitro investigations using the reconstituted NER and Pol η.
It has been controversial whether NER can recognize a small lesion such as 8-oxo-dG-and ribonucleotide-containing DNA as the substrate 24,25,30,[33][34][35] . The in vitro NER reconstitution experiments in this study implicated that 8-oxo-rG is the preferable substrate for human NER compared with rG and 8-oxo-dG in the plasmid DNA (Fig. 1c). Further, our results revealed that the loss of the NER factor XPA does not alter the mutant frequency and mutation spectra induced by rG and 8-oxo-dG, but significantly enhanced the mutagenic potential of 8-oxo-rG (Fig. 3). Hence, these observations suggest that the oxidation of guanine base in ribonucleotide enables ribonucleotide removal by human NER.
In a previous in vitro study, both replicative and TLS Pols bypassed rG in an error-free manner 21 . Conversely, the oxidation of a ribonucleotide, i.e., 8-oxo-rG, on the template DNA strongly inhibited DNA synthesis by Pols and contributed to miscoding events. Because 8-oxo-rG also could not be repaired by either RER or BER 21,22 , we hypothesized that such modification of the ribonucleotide enhances the mutagenic potential in cells. Contrary to the initial hypothesis, it seems that 8-oxo-rG did not display greater mutagenic potential than rG in WT cells (Fig. 3b,c). The frequency of large deletions induced by rG (67%) was much higher than that induced by 8-oxo-rG (7.9%) (Fig. 4b,c). These observations may be attributed to the recognition of the damaged ribonucleotide by the alternate repair and/or damage tolerance pathways. While 8-oxo-rG is the replication blocking lesion, Pol η rapidly and accurately bypasses the lesion 21 . Along with the in vitro study, our results suggest that the oxidized ribonucleotide on the template DNA is recognized by the DNA damage tolerance pathway and Pol η performs TLS in cells. Interestingly, the proportion of GG 112 T to GGG mutation caused by 8-oxo-rG was decreased in POLH −/− cells (Fig. 4c), indicating that such mutation is induced during Pol η-mediated TLS across 8-oxo-rG. www.nature.com/scientificreports www.nature.com/scientificreports/ Although the structural impact of the presence of 8-oxo-rG in DNA is still unclear, the mutation could be because of the sequence specific misalignment induced by 8-oxo-rG. For example, the nucleotide incorporated opposite T 113 is encoded by 5′-neighboring 8-oxo-rG through the misalignment-mediated nucleotide insertion followed by realignment of the template DNA during TLS. It has been suggested that such mutations due to misalignment can be induced by DNA damage at the specific sequence context 36 . While the loss of POLH resulted in a three-fold increase in the frequency of large deletion mutations induced by 8-oxo-rG (Fig. 4c), the loss of XPA did not enhance the frequency of such deletions despite the increase in the frequency of mutations (Figs 3c and 4c). Based on these observations, ribonucleotide-induced large deletions may occur in conjunction with DNA synthesis. Our results also imply that an embedded ribonucleotide causes mutations not only in tandem repeat sequences but also in other genomic loci.
In contrast with findings of the current study, previous studies have reported that such large deletion mutations at the non-repetitive sequences were rare in the genome of the yeast strain lacking RER [13][14][15] . This difference may be because of a distinct repair mechanism activated during the mutagenic processing of the ribonucleotide. During Top1-mediated ribonucleotide processing, Top1 incises the 3′-side of the ribonucleotide and forms a Top1-DNA cleavage complex (Top1cc) between the catalytic tyrosine residue and the 3′-phosphate of the ribonucleotide 14,37 . Top1cc is further reversed by the nucleophilic attack of 2′-OH of the ribonucleotide to generate a 2′,3′-cyclic phosphate and to release Top1 from the complex. The subsequent processing of the released DNA by Top1 can lead to DNA double strand breaks (DSBs) 38 . Possible repair pathways following DSBs formation include homologous recombination (HR) and non-homologous end-joining (NHEJ). The predominant DSB repair pathway is NHEJ throughout the cell cycle in mammalian cells 39 , whereas HR is prevalent especially during the G2/S phase in yeast 40 . This pathway choice may explain the differential mutagenic property of embedded ribonucleotides between yeast and mammalian cells. Interestingly, large deletions induced by rG and 8-oxo-rG were often extended from the site of the ribonucleotide (G 112 ) to a position 5ʹ-upstream in supF (Supplementary  Table S2). These mutations could result from the Top1-mediated processing of a ribonucleotide in the following manner. The incision of a ribonucleotide by Top1 generates Top1cc. Upon the generation of a 2′,3′-cyclic phosphate-terminated nick and the release of Top1, one possible scenario is the formation of a second Top1cc 5′-upstream of the nick, leaving a gap between 3′-Top1cc and 5′-OH 41,42 . Because Top1 is coupled with DNA replication 43 , the generation of such a gap or nick during replication can lead to DSBs, in which the 3′-end of the strand contains Top1cc. In the present study, the deletion mutations observed toward the 5′ side of supF may have occurred via the processing of such complex DSBs, e.g., the resection of the dirty terminus by NHEJ. Reportedly, even a single DNA damage, e.g., cyclobutane pyrimidine dimer, abasic site, and DSB is sufficient to induce large deletion mutations in the episomal and genomic DNA in mammalian cells [44][45][46] . Our results provide a further insight into the mechanism of ribonucleotide-induced genomic instability and its suppression by alternative factors involving DNA repair and damage tolerance. Figure 5 shows the probable model of ribonucleotide-induced mutagenesis and the mechanism for resolving a damaged ribonucleotide in DNA. In summary, unrepaired ribonucleotides in DNA can cause severe genomic instability via the formation of deleterious mutations in mammals. Our findings also suggest that NER and Pol η are employed the back-up pathways for suppressing ribonucleotide-induced mutagenesis under circumstances where embedded ribonucleotides cannot be properly resolved. These findings can deepen our understanding of the mechanism underlying genome instability caused by the genomic ribonucleotide accumulation. www.nature.com/scientificreports www.nature.com/scientificreports/ Methods cell culture. Human lymphoblastoid TSCER2 cells, derived from TK6 cell line 47 , were cultured in RPMI-1640 medium (Nacalai Tesque) supplemented with 200 μg/mL sodium pyruvate, 100 U/mL penicillin, and 100 μg/ mL streptomycin, and 10% (v/v) heat-inactivated fetal bovine serum at 37 °C in an atmosphere of 5% CO 2 and 100% humidity. XPA −/− and POLH −/− cells were established as described previously 29 . protein preparation. Expression  In vitro dual incision assay. The oligonucleotide containing 6-4PP was prepared as described previously 52 .
The oligonucleotides containing dG, 8-oxo-dG, rG, or 8-oxo-rG were synthesized and purified by Tsukuba Oligo Service Co., Ltd. (Ibaraki, Japan). The oligonucleotides were purified by polyacrylamide gel electrophoresis. After the purification step, the homogeneity and no detectable degradation of the oligonucleotides were confirmed by the same method prior to use. The internally 32 P-labeled DNA substrates, containing a site-specific lesion, were prepared as previously described 53 . In vitro dual incision assays were performed basically according to the previously described procedures 54 30 . To construct the POLH expression plasmid, the coding region of POLH was amplified from the cDNA of TK6 cells by PCR with primers 5′-GAGCTCGGATCCTGAAAAATGGCTACTGGACAGGATC -3′ and 5′-CAGCGGGTTTAAACCTAATGTGTTAATGGCTTAAAAAATGATTCCAATG-3′. The resulting product was digested with BamHI and PmeI and cloned into the BamHI/PmeI sites of pEF6/myc-His (ThermoFisher Scientific). The plasmid obtained was named pEF-POLH. POLH −/− cells (1 × 10 6 ) were transfected with 10 µg of ScaI-linearized pEF-POLH using a NEPA21 Super Electroporator (Nepa Gene Co., Ltd.) according to the manufacturer's instructions. After incubating for 48 h in RPMI-1640 medium, cells were plated into 96-microwell plates in the presence of 10 µg/ ml blasticidin S. Blasticidin S-resistant clones expressing POLH were then separated. The XPA-and POLH-expressing clones were then subjected to UV irradiation and supF forward mutation assay. UV irradiation. Cells in the exponential phase were washed and resuspended with RPMI-1640 medium without phenol red prior to the irradiation. Moreover, cells (2.5 × 10 6 ) suspended in 5 ml medium were placed in a 100 mm petri dish and were exposed to UVC light. After irradiation, cells were cultured for 3 days. The relative growth rate was determined by counting cells over the 3 day period using Coulter Counter Z2 (Beckman).
supF forward mutation assay. The position 112 of the supF gene was selected as the site for embedding rG, 8-oxo-rG, or 8-oxo-dG according to the previous study 55 . The 5′-phosphorylated 21-mer primer (5′-CGACTTCGAAGNTTCGAATCC-3′, where N represents dG, rG, 8-oxo-rG, or 8-oxo-dG) was synthesized and provided by Tsukuba Oligo Service Co., Ltd. (Ibaraki, Japan). Each primer was annealed with pMY189 single-stranded DNA that was prepared using the VCSM13 helper phage. The closed circular double-stranded DNA, containing a single dG, rG, 8-oxo-rG, or 8-oxo-dG, was then synthesized and purified with ultracentrifugation as previously described 56 .