Deletion of the GAPDH gene contributes to genome stability in Saccharomyces cerevisiae

Cellular metabolism is directly or indirectly associated with various cellular processes by producing a variety of metabolites. Metabolic alterations may cause adverse effects on cell viability. However, some alterations potentiate the rescue of the malfunction of the cell system. Here, we found that the alteration of glucose metabolism suppressed genome instability caused by the impairment of chromatin structure. Deletion of the TDH2 gene, which encodes glyceraldehyde 3-phospho dehydrogenase and is essential for glycolysis/gluconeogenesis, partially suppressed DNA damage sensitivity due to chromatin structure, which was persistently acetylated histone H3 on lysine 56 in cells with deletions of both HST3 and HST4, encoding NAD+-dependent deacetylases. tdh2 deletion also restored the short replicative lifespan of cells with deletion of sir2, another NAD+-dependent deacetylase, by suppressing intrachromosomal recombination in rDNA repeats increased by the unacetylated histone H4 on lysine 16. tdh2 deletion also suppressed recombination between direct repeats in hst3∆ hst4∆ cells by suppressing the replication fork instability that leads to both DNA deletions among repeats. We focused on quinolinic acid (QUIN), a metabolic intermediate in the de novo nicotinamide adenine dinucleotide (NAD+) synthesis pathway, which accumulated in the tdh2 deletion cells and was a candidate metabolite to suppress DNA replication fork instability. Deletion of QPT1, quinolinate phosphoribosyl transferase, elevated intracellular QUIN levels and partially suppressed the DNA damage sensitivity of hst3∆ hst4∆ cells as well as tdh2∆ cells. qpt1 deletion restored the short replicative lifespan of sir2∆ cells by suppressing intrachromosomal recombination among rDNA repeats. In addition, qpt1 deletion could suppress replication fork slippage between direct repeats. These findings suggest a connection between glucose metabolism and genomic stability.


Scientific Reports
| (2020) 10:21146 | https://doi.org/10.1038/s41598-020-78302-5 www.nature.com/scientificreports/ Chromatin is composed of DNA fiber and chromatin-binding proteins such as histones that package chromosomal DNA into nuclei. The condensed structure usually becomes an obstacle for the execution of nuclear activities on chromatin. Therefore, chromatin regulator proteins (e.g., histone modifiers and chromatin remodeling factors) create an environment that allows the replisome, transcriptional machinery and DNA damage repair machinery to perform on chromatin 10,11 . The canonical histones (H2A, H2B, H3 and H4) harbor various posttranslational modifications (such as acetylation, methylation, phosphorylation and ubiquitination) 10 . Histone acetyltransferase (HAT) and histone deacetylase (HDAC) acetylate and deacetylate lysine residues in histone proteins, respectively. HAT and HDAC cooperatively manage gene transcription 12,13 , chromatin remodeling 14 , and DNA damage repair 15,16 . Hst3 and Hst4, NAD + -dependent deacetylases in budding yeast 17 , mutually deacetylate histone H3 on lysine 56 (H3-K56) [18][19][20] . Histone H3-K56 is persistently acetylated in hst3∆ hst4∆ cells, which causes sensitivity to DNA damaging agents because of the loosened chromatin structure 18,19 . Because the deletion of the RTT109 gene, encoding the HAT responsible for histone H3-K56 acetylation, allows histone H3-K56 to be deacetylated and confers severe DNA damage sensitivity, proper control of both acetylation and deacetylation on histone H3-K56 in chromatin is needed to perform DNA repair 21 . Sir2 is another NAD + -dependent deacetylase, and one substrate of Sir2 is histone H4 on K16 22 . Sir2 is involved in gene silencing of genes inserted within ribosomal DNA (rDNA) repeats and at telomere loci and the silent mating type loci HML and HMR 23 . sir2 deletion causes persistence of histone H4 acetylation on K16 in chromatin, which elevates the frequencies of intrachromosomal recombination in rDNA repeats by replication fork slippage and generates extrachromosomal ribosomal DNA circles (ERCs) 24,25 . The accumulation of ERCs reduces the replicative lifespan in sir2∆ cells 24,25 . Thus, persistence of either H3-K56 acetylation on chromatin or H4 K16 acetylation on rDNA repeats leads to genome instability.
In this study, we found that the alteration of glucose metabolism suppressed the genome instability caused by aberrant chromatin structure. Deletion of the TDH2 gene, which encodes glyceraldehyde 3-phosphate dehydrogenase (GAPDH), a metabolic enzyme in glycolysis/gluconeogenesis, partially suppressed the DNA damage sensitivity of hst3∆ hst4∆, although the tdh2 gene deletion did not affect the H3-K56 acetylation in hst3∆ hst4∆ cells. tdh2 deletion could also restore the short replicative lifespan of sir2∆ cells. In addition, tdh2 deletion could suppress replication fork slippage between direct repeats in both wild-type and hst3∆ hst4∆ cells. Furthermore, we examined the role of quinolinic acid (QUIN), which is a metabolic intermediate in the de novo nicotinamide adenine dinucleotide (NAD + ) synthesis pathway (or kynurenine pathway) and was accumulated in tdh2∆ cells, and tested whether QUIN could suppress replication fork instability as observed in tdh2∆ cells. The cells with deletion of QPT1, quinolinate phosphoribosyl transferase, which synthesizes nicotinate mononucleotide (NaMN) from QUIN, experienced elevated intracellular QUIN concentrations and partially suppressed DNA damage sensitivity in hst3∆ hst4∆ cells similarly as observed in tdh2∆ cells. The qpt1 deletion restored the short replicative lifespan of sir2∆ cells. Furthermore, the qpt1 deletion suppressed replication fork slippage between direct repeats more compared to wild-type cells. These findings suggest that metabolic alteration contributes to preventing the genomic instability caused by the impaired function of chromatin regulation, and QUIN may be a candidate to stabilize the DNA replication fork to prevent DNA damage.

Results
Deletion of the TDH2 gene partially suppresses the DNA damage sensitivity of hst3∆ hst4∆ cells. The TDH2 gene encodes one of the yeast GAPDH genes (TDH1/2/3) involved in both glycolysis and gluconeogenesis. A previous study reported that hst3∆ hst4∆ cells activate gluconeogenesis in the presence of glucose, and the tdh2 gene deletion represses gluconeogenesis in enhanced hst3∆ hst4∆ cells and restores slow growth 26 . Because the slow growth of hst3∆ hst4∆ cells is due to the frequent occurrence of DNA damage during cell cycle progression 18,19 , we tested whether the tdh2 gene deletion was able to suppress the DNA damage sensitivity of hst3∆ hst4∆ cells. This was examined by monitoring cell growth in YPD (Yeast extract-Polypeptone-Dextrose) solid medium supplemented with the following DNA damaging agents: methyl methanesulfonate (MMS), hydroxyurea (HU) and camptothecin (CPT). The alkylating agent MMS attaches alkyl groups to DNA bases, and HU inhibits ribonucleotide reductase to reduce intracellular deoxyribonucleotide levels. CPT inhibits Type I topoisomerases, thus DNA SSBs are not repaired, and then creates a DNA DSB after passing the replication fork. These agents promote replication collapse and eventually cause DNA DSBs, which are usually fixed by homologous recombination (HR) 8,16,27 . hst3∆ hst4∆ tdh2∆ cells grew in YPD solid medium amended with each DNA damaging agent better than hst3∆ hst4∆ cells; however, the growth was not restored to the levels of the wild-type and tdh2∆ cells ( Fig. 1a and Fig. S1). Rad53 is a member of both the DNA damage and the intra-S phase checkpoint systems and harbors multiple phosphorylation sites phosphorylated by checkpoint activation in response to DNA lesions or replication fork arrests [28][29][30][31][32][33][34] . The extent of Rad53 phosphorylation depends on the intensity of DNA damage. We monitored Rad53 phosphorylation to examine whether deleting tdh2 represses the occurrence of DNA damage in hst3∆ hst4∆ cells. Multiple phosphorylated Rad53 was indicated in wild-type and tdh2∆ cells treated with MMS by the presence of slow-migrated bands in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), although a single band of unphosphorylated Rad53 was detected in cells without MMS ( Fig. 1b; lanes 1, 2, 5 and 6). Phosphorylated Rad53 bands were detected in hst3∆ hst4∆ and hst3∆ hst4∆ tdh2∆ cells treated with MMS, similar to the patterns of wild-type and tdh2∆ cells ( Fig. 1b; lanes 2, 4, 6 and 8). Even in the absence of MMS, a smeared Rad53 band was detected in hst3∆ hst4∆ cells (Fig. 1b; lane 3), indicating that the chromatin acetylated at histone H3-K56 becomes fragile and induces DNA damage during nuclear activities 18,19 . Interestingly, the smeared Rad53 band was not detected in hst3∆ hst4∆ tdh2∆ cells without MMS treatment or in wild-type cells ( Fig. 1b; lanes 1 and 7). Thus, tdh2 deletion suppresses the occurrence of DNA damage originating from the chromatin associated with persistently acetylated histone H3-K56 in hst3∆ hst4∆ cells. Next, we asked whether DNA damage suppression by tdh2 deletion was restricted to the Scientific Reports | (2020) 10:21146 | https://doi.org/10.1038/s41598-020-78302-5 www.nature.com/scientificreports/ chromatin acetylated histone H3-K56. Rtt109 functions as an acetyltransferase for histone H3-K56 21 . rtt109∆ cells, which did not acetylate histone H3-K56, showed greater sensitivity to MMS, CPT and HU than hst3∆ hst4∆ cells 21 ( Fig. 1c and Fig. S2). The DNA damage sensitivity of hst3∆ hst4∆ rtt109∆ cells was the same as that of rtt109∆ cells, in that H3-K56 in chromatin remained deacetylated in hst3∆ hst4∆ rtt109∆ ( Fig. 1c and Fig. S2). Both rtt109∆ tdh2∆ and hst3∆ hst4∆ rtt109∆ tdh2∆ cells exhibited the same DNA damage sensitivities as rtt109∆ cells ( Fig. 1c and Fig. S2), suggesting that tdh2 deletion does not suppress the occurrence of DNA damage caused by chromatin deacetylated histone H3-K56. Next, we asked whether tdh2 deletion might reduce the acetylation level of histone H3-K56 in hst3∆ hst4∆ cells and gain resistance to DNA damage sensitivity. Our previous analysis showed that the level of acetyl-CoA, which is used as a substrate to acetylate lysine residue on target proteins, including histones, was increased in hst3∆ hst4∆ tdh2∆ cells more than in wild-type, hst3∆ hst4∆ and tdh2∆ cells 26 . To exclude the possibility that another histone modifier reduced the acetylation level of histone H3-K56 to a level sufficient to promote DNA damage repair in hst3∆ hst4∆ tdh2∆ cells, we examined whether tdh2 deletion influenced the H3-K56 acetylation level in hst3∆ hst4∆ cells. The acetylation of histone H3-K56 is regulated in a cell cycle-dependent manner, and this acetylation appears from late G1 to early G2 phase 16,35 . When the G2/M phase was arrested by treatment with nocodazole, an inhibitor of microtubule polymerization, H3-K56 was deacetylated in both wild-type and tdh2∆ cells ( Fig. 1d; lanes 1 and 3). In contrast, the acetylation level of histone H3-K56 remained the same between hst3∆ hst4∆ and hst3∆ hst4∆ tdh2∆ cells treated with nocodazole ( Fig. 1d; lanes 2 and 4). Thus, tdh2 deletion does not affect the acetylation level in hst3∆ hst4∆ cells. Altogether, tdh2 deletion is able to suppress the genome instability caused by the aberrant chromatin structure constitutively acetylated histone H3 on K56.
tdh2 deletion suppresses deletion between direct repeats (DRs) on the chromosome. Next, we tested whether tdh2 deletion could suppress the genome instability caused by chromatin acetylation other than H3-K56. The sir2 deletion increases the acetylation level of histone H4 on K16 at ribosomal DNA (rDNA) repeats and elevates the ratio of intrachromosomal recombination among rDNA repeats by replication fork slippage to bear ERCs 25 . Accumulation of ERCs reduces the replicative lifespan of sir2∆ cells 36 . Pedigree analysis monitors the replicative age of mother cells to count the number of divided daughter cells. The replicative lifespan of wild-type cells was approximately 22 divisions (50% viability), with a maximum of approximately 40 ( Fig. 2a; wild-type). The lifespan of sir2∆ cells was significantly reduced to half the level (approximately 13 divisions (50% viability)) of wild-type cells ( Fig. 2a; wild-type vs. sir2∆ (P = 4.457E-115)). tdh2∆ cells had an www.nature.com/scientificreports/ increased replicative lifespan compared to wild-type cells ( Fig. 2a; wild-type vs. tdh2∆ (P = 0.017)) 26 . sir2∆ tdh2∆ cells exhibited a significantly extended lifespan compared to the sir2∆ cells but did not reach the lifespan of the wild-type cells ( Fig. 2a; sir2∆ tdh2∆ vs. sir2∆ (P = 3.676E-65), and sir2∆ tdh2∆ vs. wild-type (P = 0.018)). Fob1 functions as a replication fork barrier in rDNA repeats, which elevates the opportunities for intrachromosomal recombination among rDNA repeats 37 . The fob1 deletion restored the lifespan of sir2∆ cells such that the lifespan of sir2∆ fob1∆ cells was almost the same as that of wild-type ( Fig. 2b; wild-type vs. sir2∆ fob1∆). The replicative lifespans were almost the same among wild-type, sir2∆ fob1∆ and sir2∆ fob1∆ tdh2∆ cells ( Fig. 2b; wild-type vs. sir2∆ fob1∆ vs. sir2∆ fob1∆ tdh2∆), suggesting that tdh2 deletion suppresses the intrachromosomal recombination among rDNA repeats in the same manner as fob1 deletion. These data suggest that the tdh2 gene deletion can suppress replication fork slippage, causing intrachromosomal recombination among rDNA repeats in sir2∆ cells. To examine whether the genomic instability suppressed by tdh2 deletion is due to aberrant intrachromosomal recombination between direct repeats (DRs) in hst3∆ hst4∆ cells, we used a strain to monitor the ratio of the CaURA3 gene deletion by recombination between DRs due to replication fork slippage (Fig. 2c top). The CaURA3 deletion cells can grow in SC solid medium containing 5-floroorotic acid (5-FOA), a counterselecting agent for URA3 gene deletion. The frequency of CaURA3 gene deletion in hst3∆ hst4∆ tdh2∆ cells was significantly (P = 0.046) reduced to levels lower than those of hst3∆ hst4∆ cells ( Fig. 2c; hst3∆ hst4∆ vs. hst3∆ hst4∆ tdh2∆). Furthermore, the frequency of CaURA3 gene deletion in tdh2∆ cells was significantly (P = 0.034) reduced to levels lower than those of wild-type cells ( Fig. 2c; wild-type vs. tdh2∆). Thus, tdh2 deletion is involved in suppressing deletions among DRs due to replication fork slippage. Next, we examined whether tdh2 deletion supported the DNA repair machinery in preventing deletion by replication fork slippage. DNA repair machinery has a pivotal role in preventing genome rearrangement in DNA replication 38 . Sgs1, a RecQ family nucleolar DNA helicase, suppresses DNA replication-associated genome rearrangement 39,40 . sgs1∆ cells exhibited MMS and HU sensitivity, leading to DNA damage due to DNA replication fork stalling ( Fig. S3; wild-type and sgs1∆). Because the HU and MMS sensitivities of tdh2∆ cells were almost the same as those of sgs1∆ cells ( Fig. S3; sgs1∆ and sgs1∆ tdh2∆), the replication fork stability by tdh2 deletion depends on the DNA repair machinery. Thus, tdh2 deletion increases the replication fork stability together with the DNA repair machinery.
Quinolinic acid (QUIN) is a metabolic candidate to suppress replication fork instability. We examined whether a metabolite increased in tdh2∆ cells might contribute to replication fork stability in both wild-type and hst3∆ hst4∆ cells. Using capillary electrophoresis time-of-flight mass spectrometry (CE-TOF/ MS) analysis to compare the levels of cellular metabolites 26 , we focused on quinolinic acid (QUIN), which was significantly increased in tdh2∆ cells. QUIN is a metabolic intermediate in the de novo NAD + synthetic pathway, or the kynurenine pathway (Fig. 3a). QPT1 encodes quinolinate phosphoribosyltransferase to converts QUIN to nicotinic acid mononucleotide (NaMN). We confirmed that the amount of intracellular QUIN significantly (P < 0.05) accumulated in both tdh2∆ and qpt1∆ cells more than wild-type cells ( Fig. 3b; wild-type, tdh2∆ and qpt1∆). We used the qpt1 deletion strain to examine whether QUIN was able to suppress the DNA damage sensitivity of hst3∆ hst4∆ cells. hst3∆ hst4∆ qpt1∆ cells exhibited better tolerance to DNA damaging agents (MMS, HU and CPT) than hst3∆ hst4∆ cells. However, the tolerance was not recovered to the levels of the wild-type and tdh2∆ cells (Fig. 3c and Fig. S4; wild-type, qpt1∆, hst3∆ hst4∆ qpt1∆ and hst3∆ hst4∆). Furthermore, the sensitivity of hst3∆ hst4∆ tdh2∆ qpt1∆ cells was almost the same as that of hst3∆ hst4∆ tdh2∆ cells ( Fig. 3c and Fig. S4; hst3∆ hst4∆ tdh2∆ and hst3∆ hst4∆ tdh2∆ qpt1∆). Thus, qpt1 deletion can partially restore the DNA damage www.nature.com/scientificreports/ sensitivity of hst3∆ hst4∆ cells in a Tdh2-dependent manner. Next, we examined whether qpt1 deletion was able to extend the lifespan of sir2∆ cells by pedigree analysis. As shown in Fig. 3d, sir2∆ qpt1∆ cells experienced significantly extended replicative lifespans compared to sir2∆ cells (sir2∆ qpt1∆ vs. sir2∆ (P = 8.324E-65)). Furthermore, the lifespan of sir2∆ tdh2∆ qpt1∆ cells was almost the same as that of sir2∆ tdh2∆ cells ( Fig. 3e; sir2∆  tdh2∆ qpt1∆ vs. sir2∆ tdh2∆), suggesting that qpt1 deletion extends the lifespan of sir2∆ in a Tdh2-dependent manner. Next, we investigated whether the qpt1 deletion suppressed spontaneous URA3 gene deletion between DRs as well as the tdh2 deletion. The ratio of CaURA3 gene deletion in DRs was significantly (P = 0.042) reduced in qpt1∆ cells compared to wild-type cells (Fig. 3f). Altogether, these findings suggest that the elevated intracellular QUIN levels in the tdh2 deletion contribute to genome stability.
Intracellular NAD + , which is elevated in tdh2∆ cells, does not suppress the DNA damage sensitivity of hst3∆ hst4∆ cells. QUIN is utilized in the NAD + salvage pathway to synthesize NAD + (Fig. S5A).
We hypothesized that QUIN accumulated to synthesize NAD + de novo and that NAD + functioned as an end product to suppress replication fork instability in tdh2∆ cells. To test this hypothesis, we examined whether tdh2 deletion elevated intracellular NAD + levels. The cells were cultured, and the intracellular NAD + concentration was measured by performing acid extractions of nucleotides over the course of cell growth in liquid cultures. The intracellular NAD + concentration was almost the same between wild-type and tdh2∆ cells at first, although the concentration was severely decreased in cells lacking NPT1, the nicotinate phosphoribosyl transferase in the NAD + salvage pathway ( Fig. S5A and B) 41 . Intracellular NAD + concentrations in both wild-type and npt1∆ cells steadily declined over the course of cell culture (Fig. S5B) because yeast cultures reached the end of the log phase and approached the diauxic shift, which was due to depletion of the limiting NAD + precursor nicotinic acid from the growth medium 41,42 . In contrast, the NAD + concentration remained constant in tdh2∆ cells over the course of cell culture (Fig. S5B), indicating that tdh2 deletion promotes NAD + synthesis. In addition, tdh2 deletion www.nature.com/scientificreports/ slightly increased the intracellular NAD + concentration in hst3∆ hst4∆ tdh2∆ cells, although the concentration declined in hst3∆ hst4∆ cells (Fig. S5C). Next, we examined whether de novo NAD + synthesis induced by tdh2 deletion is necessary for the growth of hst3∆ hst4∆ cells. To monitor the contribution of tdh2 deletion to declining intracellular NAD + levels in hst3∆ hst4∆ cells, we constructed npt1∆ hst3∆ hst4∆ and npt1∆ hst3∆ hst4∆ tdh2∆ strains containing the PHM286 plasmid harboring the HST3 and URA3 genes. Strains were streaked on SC solid medium supplemented with 5-FOA, a counterselecting agent, to induce loss of the URA3 plasmid. Although npt1∆ hst3∆ hst4∆ cells did not grow in SC medium with 5-FOA, npt1∆ hst3∆ hst4∆ tdh2∆ cells did grow (Fig. S5D). Thus, an increase in intracellular NAD + levels by tdh2 deletion is necessary for the cell growth of hst3∆ hst4∆ cells. To reveal the mechanism that elevates intracellular NAD + levels by tdh2 deletion, we compared the transcription levels of genes involved in NAD + synthesis among strains. The BNA2 and BNA4 genes belong to the kynurenine pathway and provide QUIN for tryptophan (Fig. 3a), and TNA1 encodes the high affinity nicotinic acid plasma membrane permease responsible for the uptake of nicotinic acid (Fig. S5A). Although the expression levels of the BNA2 and BNA4 genes were almost the same between wild-type and tdh2∆ cells and the level of TNA1 was slightly increased in tdh2∆ cells compared with wild-type cells (Fig. S5E), the expression levels of these genes were significantly (P < 0.05) increased in hst3∆ hst4∆ tdh2∆ cells compared with hst3∆ hst4∆ cells ( Fig. S5E and F). Thus, tdh2 deletion induces the gene expression involved in intracellular NAD + synthesis in hst3∆ hst4∆ cells, which can improve the cell growth of hst3∆ hst4∆ cells. Next, we examined whether elevated intracellular NAD + levels suppressed the DNA damage sensitivity of hst3∆ hst4∆ cells. Isonicotinamide (INAM), an isostere of nicotinamide, raises intracellular NAD + levels in budding yeast 43 . As shown in Fig. S6A, the intracellular NAD + concentration was significantly (P < 0.05) elevated in the presence of INAM in both wild-type and hst3∆ hst4∆ cells. Treatment with INAM allowed both wild-type and hst3∆ hst4∆ cells to tolerate MMS (Fig. S6B). However, the addition of INAM did not suppress the sensitivities of hst3∆ hst4∆ cells to HU and CPT (Fig. S6B). These results indicate that elevated intracellular NAD + levels do not contribute to the DNA damage sensitivity of hst3∆ hst4∆ cells. Altogether, these data suggest that NAD + does not contribute to the suppression of replication fork instability in tdh2∆ cells.

Discussion
In this study, we elucidated that the deletion of TDH2 suppresses replication fork instability in the chromatin environment in which composed histones were acetylated. The acetylation of histone molecules loosens the chromatin structure by weakening the DNA-histone interaction. In particular, histone H3 on K56 is located within the histone core region and interacts with the DNA strand, and acetylation directly weakens the DNAhistone interaction 16 . The loosened chromatin structure becomes fragile, causing SSBs and DSBs during nuclear activities and exposes the naked DNA region to possible replication fork slippage at repeats. Both sir2∆ and hst3∆ hst4∆ cells exhibit high frequencies of loss of heterozygosity (LOH) 20,44 . LOH occurs in diploid cells that lose the chromosome arm or regenerate another chromosome arm instead of losing the arm by break-induced replication (BIR) 4,45 . In this study, tdh2 deletion suppresses recombination by replication fork slippage in both hst3∆ hst4∆ and sir2∆ cells. The frequency of recombination between DRs was reduced in tdh2∆ cells compared to wild-type cells (Fig. 2c), which contributes to the extension of the replicative lifespan of tdh2∆ cells (Fig. 2a) 26 .
ROS, hazardous byproducts of mitochondrial respiration, are well recognized as mediators of DNA damage 46 . DNA damage caused by ROS appears as an oxidized base, a sugar modification, a DNA or protein crosslink or a DNA strand break [47][48][49] . 8-Oxoguanine, a representative oxidized DNA adduct, works as a replication fork blocker 50,51 . The accumulation of DNA damage causes replication fork stalling, which provides an opportunity for HR between direct repeats on the DNA strand. DNA damage agents used in this study (MMS, HU and CPT) are involved in ROS production. MMS indirectly inhibits respiratory chain in mitochondria by caused mitochondrial DNA (mtDNA) damage, which induces ROS production in budding yeast 52 . CPT also inhibits DNA topoisomerase I in mitochondria to induce mtDNA damages, which inhibits the respiratory chain in mitochondrion to induce ROS production in mammals 53 . HU also induces ROS production in budding yeast 54 . ROS is usually scavenged by several cellular metabolites, NADPH, a reduced form of nicotinamide adenine dinucleotide phosphate (NADP), functions as a major scavenger of ROS and is provided from the pentose phosphate pathway (PPP) branched from glycolysis mainly. In addition of generating phosphopentoses and ribonucleotides, PPP plays a pivotal role to combat oxidative stress 55 . Our previous study showed that tdh2∆ and hst3∆ hst4∆ tdh2∆ cells accumulated NADP + , metabolic intermediates of PPP and various ribonucleotides 26 . The phenotype of tdh2∆ is similar to that of fbp1∆, essential gene in gluconeogenesis 26 . The combination with interruption of gluconeogenesis caused by fbp1∆ and triple sirtuin gene deletions (hst3∆ hst4∆ sir2∆) alters the metabolic flux from glycolysis to PPP and increase the ribonucleotide levels 56 . These suggests that tdh2∆ alters the metabolic flux from glycolysis to PPP, and increases NADPH level to scavenge ROS. Therefore, tdh2∆ brings the resistance for DNA damage agents (HU, MMS and CPT) in hst3∆ hst4∆ genetical background, and stabilizes replication fork to prevent homologous recombination between DRs.
In mammalian, QUIN has been shown to increase the production of free radicals, leading to oxidative stress, DNA damage, and increased poly(ADP-ribose) polymerase 1 (PARP-1) activity 57,58 . However, moderate QUIN level can induce resistance to oxidative stress through increased NAD + production 59 . NAD + influences DNA repair and gene expression through its role as a substrate for PARP-1 in mammal cells. Although there is no PARP in budding yeast, another mechanism supports to allow QUIN to induce resistance for oxidative stress. It has been shown that QUIN chelates ferrous iron to generate ROS 60 . Iron-Sulfur (Fe-S) centers are metallic cofactors with electronic properties that are associated with proteins 61 . It is well known that numerous DNA-transacting proteins such as DNA replication machinery contains Fe-S centers 54 . Because Fe-S centers are sensitive to oxidative agents, a moderate QUIN level can activate the oxidative stress response, and leads to resistance for www.nature.com/scientificreports/ oxidative stress. In the future, approaches to repair or prevent DNA damage will focus more heavily on altering the metabolic state of cells.

Methods
Strains and media. The genotypes of the strains, plasmids and primers used in this study are listed in Table S1. The parental budding yeast strain used in the present study was BY4742 (MATα his3∆ leu2∆1 met15∆0 ura3∆0) 62 . Yeast cells were routinely grown at 30 °C in YPD (1% yeast extract, 2% peptone, 2% glucose) or appropriate synthetic complete (SC) medium 63,64 . If necessary, the media were solidified with 2% agar. A yeast strain harboring a single gene deletion was commercially available from the haploid yeast open reading frame deletion collection 65 (GE Dharmacon, Lafayette, CO, USA). To construct a double or triple gene deletion strain, the different mating type single gene deletion haploid strains were crossed, and sporulation was subsequently induced. After dissection, the spores were germinated on YPD medium. The deletion of each gene was confirmed using either antibiotics or auxotrophic markers and checking for growth on agar plates containing antibiotics or SC agar plates without selective amino acids. We employed YPD media supplemented with the following antibiotics: G418 (Sigma-Aldrich, St. Louis, MO, USA) at a final concentration of 100 μg/ml for the kan gene, hygromycin B at a final concentration of 200 μg/ml for the hph gene and ClonNAT (Werner Bioagents, Germany) at a final concentration of 100 μg/ml for the nat gene. SC-histidine medium was employed to select the his5 + strain 66,67 .
The strains with deletions in both HST3 and HST4 harbor the PHM286 URA3 plasmid, which contains the wild-type HST3 gene and prevents spontaneous DNA damage and genomic instability. These strains were counterselected for loss of the PHM286 plasmid by selecting colonies that grew in SC medium supplemented with 5-fluoroorotic acid (5-FOA) at a final concentration of 100 µg/ml prior to use in subsequent experiments. To construct the strains monitoring CaURA3 gene deletion frequencies, the plasmid PHM764 was digested with Bsu 36I and integrated into the TRP1 locus. The URA + strains were selected and confirmed for correct integration of the PHM764 plasmid by PCR. A standard method was used for isolation of the yeast genomic DNA 64 . E. coli strain DH5α 68 and standard media and methods were used for plasmid manipulations 69  DNA damage sensitivity test. Yeast strains were cultured in 5 ml of YPD liquid medium at 25 °C overnight and then adjusted to 5 × 10 6 cells/ml in 5 ml of YPD medium. Next, cell culture was continued for 3 h at 25 °C, and then cells were harvested and suspended in DIW at 5 × 10 7 cells/ml. Two hundred microliters of cell suspension (1 × 10 7 cells) was transferred to a 96-well plate and sequentially diluted tenfold to a concentration of 5 × 10 3 cells/ml. A small portion of the diluted cell suspension in each well was set on YPD medium containing each concentration of DNA damaging agent using a replica plater (Sigma-Aldrich, St. Louis, MO, USA). The plates were incubated at 25 °C for 3 to 5 days. YPD media was supplemented with the following DNA damaging agents: methyl methanesulfonate (MMS), hydroxyurea (HU) and camptothecin (CPT). The concentration of each agent is listed in figures.
Western blotting. Proteins were separated by 15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes (Amersham PROTORAN) (GE healthcare, Little Chalfont, Buckinghamshire, England). The protein level in each lane on a nitrocellulose membrane was adjusted equally and confirmed by staining with 0.1% Ponceau S solution (Sigma-Aldrich, St. Louis, MO, USA). Rabbit polyclonal anti-acetylated lysine 56 on the yeast histone H3 antibody (1:1000 dilution) was used 16 . For detection of phosphorylated species of the Rad53 protein, anti-Rad53 rabbit polyclonal antibody (the equivalent antibody is commercially available (Santa Cruz Biotechnology, Dallas, TX, USA)) was used.
Detection of Rad53 phosphorylation in response to DNA damage. Yeast strains were cultured in 5 ml of YPD liquid medium at 25 °C overnight and then adjusted to a concentration of 5 × 10 6 cells/ml in 5 ml of YPD medium containing MMS at a final concentration of 0.03%. Afterwards, cell culture was continued for 3 h at 25 °C, and then cells were harvested. The whole cell extraction method was described previously 72 . Multiple species of phosphorylated Rad53 and unphosphorylated Rad53 were separated by 7.5% SDS-PAGE and then detected by western blotting using an anti-Rad53 antibody.
Cell cycle arrest during G2/M phase. Yeast strains were cultured in 5 ml of YPD liquid medium at 25 °C overnight and then adjusted to a concentration of 5 × 10 6 cells/ml in 5 ml of YPD medium containing nocodazole www.nature.com/scientificreports/ at a final concentration of 10 µg/ml. Afterwards, cell culture was continued for 3 h at 25 °C, and then cells were harvested. The whole cell extraction method was previously described 72 .
Replicative lifespan assay (pedigree analysis). The pedigree analysis procedure was described previously 26 . Typically, a minimum of 50 mother cells was counted for each strain tested. To compare the difference in replicative lifespans among strains statistically, we performed the unpaired t-test (two-tailed).
CaURA3 deletion assay. The yeast strains were streaked on SC-Ura plates to select URA3-positive strains prior to the assay. The URA3-positive colonies were inoculated in 5 ml of YPD liquid medium and cultured at 25 °C overnight. A small aliquot of culture (5 × 10 4 or 5 × 10 5 cells/strain) was plated in a YPD solid plate supplemented with 5-FOA at a final concentration of 100 µg/ml to select strains containing the ura3 gene deletion. The number of colonies was counted, and the frequencies of the CaURA3 gene deletion were calculated for plated cells. Three replicates were analyzed for each strain.
Synthetic lethality test. Yeast strains with or without the PHM286 plasmid (URA3) were grown on YPD medium at 25 °C overnight; then, the strains were streaked on SC solid medium supplemented with 5-FOA at a final concentration of 100 µg/ml at 25 °C for 3 days to counterselect for loss of the PHM286 plasmid.  Table S1.
The measurement of NAD + concentration and INAM treatment. The yeast strains were cultured in 250 ml of YPD liquid medium at 25 °C. During cell culture, the OD 600 nm of a 1:10 cell dilution and the cell number were recorded, and 20 ml of culture was pelleted and washed with water. After harvesting, the cell pellet was stocked at − 80 °C until use. The preparation and quantification of intracellular NAD + was described previously 41 . Over 10 independent cultures were routinely used to determine the NAD + concentrations in duplicate. For INAM treatment, the yeast strains (5 × 10 6 cells/ml at start) were cultured in 20 ml of YPD liquid medium supplemented with or without INAM at a final concentration of 25 mM at 25 °C until an OD 600 of ~ 1.5 was reached and then harvested as described previously 43 . Cell pellets were employed to measure the intracellular NAD + concentration.

Measurement of intracellular quinolinic acid (QUIN) concentration. For metabolite extraction,
yeast cells (1 × 10 8 cells) were suspended in 50% methanol and immediately frozen in liquid nitrogen. Then, frozen samples were ground by a Multi Beads Shocker (Yasui Kikai, Osaka, Japan) then centrifuged at 13,000 × g for 10 min at 4 °C. The supernatant was mixed with an equal volume of chloroform, and the mixture was centrifuged again. The upper aqueous phase was transferred to a tube and evaporated using SpeedVac SPD 1010 (Thermo Fischer Scientific, Waltham, MA, USA). Levels of quinolinic acid (QUIN) in yeast cells were determined using an Agilent 6460 Triple Quad mass spectrometer coupled to an Agilent 1290 HPLC system with multiple reaction monitoring (MRM) mode. The MRM transition for QA was optimized as m/z 166 to 78. MS settings and chromatographic conditions were described previously 73 . The amount of QA was calculated by integrating the sum of the area using Mass Hunter Quantitative software (Agilent Technologies, Santa Clara, CA, USA).