DNA damage response activates respiration and thereby enlarges dNTP pools to promote cell survival in budding yeast

The DNA damage response (DDR) is an evolutionarily conserved process essential for cell survival. Previously, we found that decreased histone expression induces mitochondrial respiration, raising the question whether the DDR also stimulates respiration. Here, using oxygen consumption and ATP assays, RT-qPCR and ChIP-qPCR methods, and dNTP analyses, we show that DDR activation in the budding yeast Saccharomyces cerevisiae, either by genetic manipulation or by growth in the presence of genotoxic chemicals, induces respiration. We observed that this induction is conferred by reduced transcription of histone genes and globally decreased DNA nucleosome occupancy. This globally altered chromatin structure increased the expression of genes encoding enzymes of tricarboxylic acid cycle, electron transport chain, oxidative phosphorylation, elevated oxygen consumption, and ATP synthesis. The elevated ATP levels resulting from DDR-stimulated respiration drove enlargement of dNTP pools; cells with a defect in respiration failed to increase dNTP synthesis and exhibited reduced fitness in the presence of DNA damage. Together, our results reveal an unexpected connection between respiration and the DDR and indicate that the benefit of increased dNTP synthesis in the face of DNA damage outweighs possible cellular damage due to increased oxygen metabolism.

the "checkpoint kinases" is the key component of the DDR (Fig.  1 and Table 1) (6 -9). In the presence of damaged DNA, the sensor kinases (ATM/ATR in mammals and Tel1/Mec1 in budding yeast) become active and phosphorylate the effector kinases (CHK1 and CHK2 in mammals and Chk1p and Rad53p in budding yeast) (10). Rad53p, the yeast ortholog of CHK2, is an essential intermediate kinase in the checkpoint pathway, as it connects the upstream kinases and downstream effectors to mediate an array of cellular outcomes in response to DNA damage (6,11,12). One of the Rad53p targets is another checkpoint kinase Dun1p (13).
Activation of the checkpoint kinases results in cell cycle arrest, activation of DNA repair, and reprogramming of transcription. One of the key outcomes of the DDR in yeast is the enlargement of the deoxyribonucleoside triphosphate (dNTP) pools, which is a prerequisite for effective DNA repair ( Fig. 1) (14,15). The rate-limiting step of dNTP synthesis is the reduction of ribonucleoside diphosphates into corresponding deoxyribonucleoside diphosphates, catalyzed by ribonucleotide reductase (RNR) (16). In most eukaryotes, RNR enzymes are ␣2␤2 heterotetramers, in which the ␣2 homodimer and the ␤2 homodimer represent the large and small subunits, respectively. In yeast, however, the small subunit is a heterodimer of Rnr2p and Rnr4p; the large subunit is a homodimer of Rnr1p. The catalytic site is contained within the large subunit of both mammalian and yeast RNR enzymes. Both mammalian and yeast RNR genes are regulated transcriptionally, and the enzymes are regulated allosterically (17)(18)(19). In yeast, transcription of RNR2, RNR3, and RNR4 genes is induced following checkpoint activation and Dun1p-mediated phosphorylation and inactivation of the transcriptional repressor Crt1p (20). Transcription of RNR1 is regulated in a cell cycle-dependent manner by the transcriptional complex MBF and by high mobility group-domain protein Ixr1p, but not by Crt1p (21)(22)(23)(24). Dun1p regulates RNR activity and dNTP synthesis by at least two additional mechanisms. Dun1p phosphorylates Dif1p, a protein required for nuclear localization of Rnr2p and Rnr4p. Phosphorylation of Dif1p by Dun1p releases Rnr2p and Rnr4p into the cytoplasm, where they assemble with Rnr1p to form an active RNR enzyme (25)(26)(27)(28)(29)(30). During S phase or after DNA damage, Dun1p also phosphorylates and induces degradation of Sml1p, a protein that binds and inhibits the Rnr1p subunit ( Fig.  1) (31)(32)(33)(34).
Proliferating cells need to maintain a delicate balance between histone and DNA synthesis to ensure correct stoichiometric amounts for chromatin assembly and to avoid genome instability (35,36). Treatment with genotoxic agents that damage DNA or interfere with DNA replication triggers repression of histone genes (37)(38)(39). We have previously shown that a decrease in histone expression induces respiration (40). This poses an intriguing question: does DDR induce mitochondrial respiration? One of the sources of reactive oxygen species (ROS) is the oxidative electron transport chain (ETC) in the mitochondria. It is widely believed that DDR results in downregulation of respiration to protect DNA from endogenous ROS (41)(42)(43). Surprisingly, our data show that DDR and growth in the presence of sublethal concentrations of genotoxic chemicals activate respiration to increase ATP production and to elevate dNTP levels, which are required for efficient DNA repair and cell survival upon DNA damage.

DDR stimulates aerobic respiration
To determine whether DDR stimulates respiration, we used two approaches to introduce DDR. The first approach utilized the genotoxic chemicals bleocin and 4-nitroquinoline 1-oxide . Bleocin belongs to the antibiotic bleomycin family and causes DNA double-strand breaks (44). 4-NQO mimics the effect of UV light and forms DNA adducts (45). Both bleocin and 4-NQO trigger DDR. When compared with control cells, cells grown in the presence of sublethal concentrations of either chemical consumed more oxygen and produced more ATP, two parameters reflecting the activity of aerobic respiration in the mitochondria (Fig. 2, A and B) (40,46). Oxygen consumption of cells treated with bleocin or 4-NQO increased 1.8-and 1.5-fold, respectively, whereas the cellular ATP level increased 2.6-and 2.0-fold, respectively (Fig. 2, A and B).
The second approach to introduce DDR employed rad52⌬ mutation. RAD52 is required for DNA double-strand break repair and homologous recombination. Inactivation of RAD52 renders cells unable to repair DNA strand breaks and thereby triggers DDR (47). Compared with WT cells, rad52⌬ cells consumed 1.6 times more oxygen and displayed 3.2 times increased ATP levels (Fig. 2C).

Checkpoint kinases Mec1p and Rad53p are required for DDRinduced respiration
DDR is mediated through activation of checkpoint kinases and their cellular targets, which coordinate cell cycle arrest and repair of damaged DNA (Fig. 1). To investigate the requirement

DNA damage response activates respiration
of the checkpoint kinases Mec1p, Tel1p, Chk1p, Rad53p, and Dun1p for induction of respiration, we introduced the corresponding mutations into rad52⌬ cells and determined oxygen consumption (Fig. 3A). The oxygen consumption of tel1⌬ cells was elevated compared with WT cells, and introducing the tel1⌬ mutation into the rad52⌬ cells further increased oxygen consumption above the rad52⌬ level. chk1⌬ cells consumed less oxygen than WT cells, and the oxygen consumption of the rad52⌬chk1⌬ cells was attenuated compared with rad52⌬ cells. Because mec1⌬ and rad53⌬ cells are viable only if harboring crt1⌬ or sml1⌬ mutations (20, 31), we measured oxygen consumption in mec1⌬sml1⌬, mec1⌬crt1⌬, rad53⌬sml1⌬, and rad53⌬crt1⌬ strains. Surprisingly, mec1⌬sml1⌬ and rad53⌬sml1⌬ strains displayed increased oxygen consumption compared with WT cells. This increase can be attributed to the sml1⌬ mutation, because sml1⌬ cells also displayed slightly increased oxygen consumption. This finding is not entirely surprising, because sml1⌬ cells display increased copy number of mitochondrial DNA, likely due to the elevated dNTP level (48). The oxygen consumption in rad53⌬crt11⌬ was comparable with WT cells, while the oxygen consumption of mec1⌬crt1⌬ cells was decreased. Importantly, introducing the rad53⌬sml1⌬ and rad53⌬crt1⌬ mutations into rad52⌬ cells completely abrogated the elevated oxygen consumption of rad52⌬ cells. Extremely slow growth of rad52⌬mec1⌬sml1⌬ and rad52⌬mec1⌬crt1⌬ did not allow culturing these cells in sufficient quantity for further analysis. The oxygen consumption of dun1⌬ cells was elevated compared with WT cells, and introducing dun1⌬ mutation into rad52⌬ cells did not diminish the oxygen consumption of rad52⌬ cells. This result suggests that induction of respiration in rad52⌬ cells is not due to degradation of Sml1p and increased dNTP synthesis, because degradation of Sml1p requires Dun1p. The requirement of Rad53p for DDR-induced respiration is consistent with suppression of the elevated ATP levels in rad52⌬ cells by introducing rad53⌬sml1⌬ and rad53⌬crt1⌬ mutations (Fig. 3B).
To corroborate these results, we induced DDR by growing cells bearing deletions of the individual checkpoint kinases in the presence of sublethal concentrations of bleocin. Growth in the presence of bleocin induced respiration in WT, tel1⌬, chk1⌬, and dun1⌬ cells. The induction of respiration was completely absent in mec1⌬sml1⌬, mec1⌬crt1⌬, rad53⌬sml1⌬, and rad53⌬crt1⌬ cells (Fig. 3C). We interpret these results to mean that Mec1p and its downstream effector kinase Rad53p are required, whereas Tel1p, Chk1p, and Dun1p are not required for DDR-induced respiration. As a control, we also included the cyt1⌬ strain. CYT1 encodes cytochrome c 1 , and cyt1⌬ cells are not able to respire (40). The results show that the ETC is responsible for about 95% of the oxygen consumed by the cells growing in the absence or presence of bleocin (Fig. 3C).
To test the possibility that the increased oxygen consumption in the WT cells grown in the presence of genotoxic chemicals or in rad52⌬ cells is caused by a delayed progression through the S phase of the cell cycle, we arrested WT and rad52⌬ cells in G 1 phase with ␣-factor and compared oxygen consumption of arrested and asynchronous cells (Fig. 3D). Because the oxygen consumption in the two cell populations does not significantly differ for both WT and rad52⌬ cells, we conclude that the DDR induces respiration in a cell cycle-independent manner.

DDR down-regulates histone levels through activation of Rad53p
Because the connection between DDR and respiration is not immediately obvious and to some extent is counterintuitive, we asked the following question. What is the molecular mechanism underlying this phenomenon? We have recently reported that decreased histone expression results in reduced nucleosome occupancy across the genome and altered chromatin structure, which triggers respiration (40). To determine whether a similar mechanism is responsible for DDR-induced respiration, we assessed histone expression in cells grown in the presence of sublethal concentrations of bleocin or 4-NQO. Under these conditions, the expression of all four histone genes as well as the protein levels of histone H3 were markedly decreased (Fig. 4, A-C). A very similar trend of decreased histone gene

DNA damage response activates respiration
expression and protein level of histone H3 was observed in rad52⌬ cells, where the DDR is induced genetically (Fig. 4, D-H). Furthermore, down-regulation of histone transcripts and protein levels depended on Rad53p, as introducing rad53⌬ mutation into rad52⌬ cells restored histone transcript and protein levels in rad52⌬rad53⌬crt1⌬ and rad52⌬rad53⌬sml1⌬ cells to the WT levels ( Fig. 4, D-G). Down-regulation of histone transcripts in rad52⌬ cells was not suppressed by introducing tel1⌬ or dun1⌬ mutations into rad52⌬ cells and was only partially suppressed by chk1⌬ mutation.

DNA damage response activates respiration DDR-mediated repression of histone levels alters chromatin structure and induces expression of TCA cycle and ETC genes
Yeast genes can be categorized as growth genes and stress genes, with each group featuring distinct nucleosomal architecture of their promoters (49). The promoter of a growth gene contains a "nucleosome-free region," whereas the promoter of a stress gene is usually occupied by delocalized nucleosomes. As a result, growth genes are constitutively expressed in contrast to stress genes, which are regulated by factors that affect chromatin structure, including abundance of histone proteins. Considering that respiratory genes in Saccharomyces cerevisiae belong to the stress gene category (50) and that reduced histone

DNA damage response activates respiration
expression induces respiration (40), we reasoned that by downregulating histone expression DDR might affect chromatin structure and induce respiratory genes. To test this possibility, we determined histone H3 and RNA pol II occupancy in the promoters of CIT1, IDH1, and QCR7, genes encoding enzymes of the TCA cycle and ETC. The histone H3 occupancy of CIT1, IDH1, and QCR7 promoters was significantly reduced in cells treated with bleocin or 4-NQO compared with control cells, whereas the occupancy of RNA pol II at the same set of promoters was increased (Fig. 5, A and B). Similarly, the histone H3 occupancy of CIT1, IDH1, and QCR7 promoters was decreased, whereas RNA pol II occupancy at the same promoters was increased in rad52⌬ cells (Fig. 5, C and D). The changes in chromatin structure and RNA pol II occupancy were accompanied by increased transcription of the corresponding genes upon treatment with bleocin or 4-NQO and in rad52⌬ cells (Fig. 5E). In this analysis, we also included the COX1 gene that is encoded by the mitochondrial genome. As we have shown previously, decreased histone expression and chromatin changes of the nuclear genome affect transcription of genes encoded by the mitochondrial genome. The corresponding mechanism involves elevated expression of nuclear genes RPO41 and MTF1, encoding mitochondrial RNA polymerase and its associated factor, respectively (40).
Inactivation of RAD52 increased respiration (Fig. 2) and down-regulated histone levels (Fig. 4) in a Rad53p-dependent manner. To determine whether the increased expression of genes required for the TCA cycle, ETC, and OXPHOS in rad52⌬ cells also requires Rad53p or other checkpoint kinases, we determined transcript levels for CIT1, IDH1, QCR7, and COX1 genes in rad52⌬ cells containing deletions of the individual checkpoint kinases (Fig. 6). We found that only inactivation of RAD53 (in rad53⌬sml1⌬ and rad53⌬crt1⌬) but not inactivation of TEL1, CHK1, or DUN1 suppressed the elevated expression of CIT1, IDH1, and QCR7 genes in rad52⌬ cells.

Elevated histone levels suppress respiration in rad52⌬ cells
To test whether the decreased histone levels and altered chromatin structure are indeed responsible for induction of respiration when cells grow in the presence of sublethal concentrations of genotoxic chemicals, we elevated histone levels in WT cells by ectopic expression of extra histones. A high copy number plasmid encoding all four core histones significantly reduced oxygen consumption when cells were grown in the presence of bleocin (Fig. 7, A and B). In addition, overexpression of histones suppressed oxygen consumption in rad52⌬ cells (Fig. 7C).
In another approach, we tested whether the histone levels can be increased in rad52⌬ cells by introducing tom1⌬ or hir1⌬ mutations. Tom1p functions in a pathway responsible for degradation of free histones. Free histones that are not assembled into chromatin are degraded in a pathway that depends on phosphorylation by Rad53p and ubiquitylation by Ubc4p, Ubc5p, and Tom1p (51,52). When TOM1 was deleted in rad52⌬ cells, the protein level of histone H3 was restored almost to the WT level, confirming the usefulness of this approach (Fig. 7D). Hir1p is a subunit of the HIR complex that acts as a histone chaperone and a repressor of the majority of histone genes (53,54). Inactivation of HIR1 induces expression of histone genes (35). Introducing tom1⌬ or hir1⌬ mutations into rad52⌬ cells significantly suppressed oxygen consumption, suggesting that it is indeed the decreased level of histones that is responsible for induction of respiration in rad52⌬ cells (Fig. 7E).

DDR-induced respiration activates dNTP synthesis
Does the DDR-induced respiration provide any advantage to yeast cells? One major outcome of DDR is the increase in activity of RNR, the key enzyme that catalyzes the rate-limiting step in dNTP synthesis (16). The complex regulation of RNR activity suggests that the control of the dNTP pools is very important for maintaining genome integrity and cell survival under genotoxic stress. Indeed, the enlargement of dNTP pools is essential for effective DNA repair (14,15).
In our previous work, we have shown that reduced histone expression and altered chromatin structure induce respiration and significantly elevate cellular ATP levels (40). To determine whether increased respiration and elevated ATP levels can drive up dNTP synthesis, we evaluated the sizes of dNTP pools in swi6⌬ and asf1⌬ cells (Fig. 8A). Swi6p is the transcriptional activation subunit of the SBF and MBF complexes that regulate histone gene transcription (35,36). Asf1p is a histone chaperone involved in chromatin assembly. Both swi6⌬ and asf1⌬ cells display markedly up-regulated oxygen consumption and ATP levels (40). We found that the dNTP pools are significantly increased in swi6⌬ and asf1⌬ cells, and the increase is abolished in swi6⌬cyt1⌬ and asf1⌬cyt1⌬ cells (Fig. 8A). CYT1 encodes the cytochrome c 1 subunit, and deletion of CYT1 inactivates the ETC. When we induced DDR in WT cells either chemically or genetically, the dNTP pools were also significantly increased. However, blocking respiration by introducing the cyt1⌬ mutation appreciably decreased dNTP levels in rad52⌬ cells or WT cells treated with bleocin or 4-NQO (Fig. 8B). These results indicate that decreased histone expression and the defect in chromatin structure or growth in the presence of genotoxic chemicals activate respiration and increase ATP and dNTP levels.
Survival of yeast cells in the presence of DNA damage depends on the availability of dNTPs for effective DNA repair (14,19). Consistently with this role of increased dNTP pools for effective DNA repair, cyt1⌬ cells that cannot up-regulate respiration and are thus unable to effectively enlarge their dNTP pools are more sensitive to bleocin or hydroxyurea (Fig. 8C). Inactivation of SML1 does not significantly change growth on bleocin or hydroxyurea, but sml1⌬cyt1⌬ cells are more sensitive to bleocin or hydroxyurea than sml1⌬ cells. Interestingly, sml1⌬cyt1⌬ cells appear to be more sensitive to hydroxyurea than cyt1⌬ cells. The difference in sensitivity to hydroxyurea between cyt1⌬ and sml1⌬cyt1⌬ cells is quite subtle but reproducible. Because both ATP and Sml1p are allosteric regulators of RNR (55), this observation may indicate that under conditions of direct inhibition of RNR by hydroxyurea, the absence of Sml1p sensitizes RNR to low ATP levels.
To test whether the lethality of rad53⌬ mutation can be suppressed by increasing dNTP synthesis by up-regulating respiration and ATP synthesis, we introduced mbp1⌬ mutation into rad53⌬ cells. MBP1 encodes the DNA-binding subunit of the DNA damage response activates respiration DNA damage response activates respiration MBF transcriptional factor. We have shown previously that mbp1⌬ cells display increased respiration and ATP levels (40). The rad53⌬mbp1⌬ cells are viable and grow significantly better than the rad53⌬mbp1⌬cyc1⌬ cells (Fig. 8D). CYC1 encodes cytochrome c, and cyc1⌬ cells are not able to respire (56). This result indicates that increasing respiration and presumably ATP and dNTP synthesis represent the major mechanism for suppression of rad53⌬ lethality by mbp1⌬ mutation.

Only long-term but not acute genotoxic stress induces respiration
Several studies found that DDR represses transcription of genes encoding enzymes of the TCA cycle, ETC, and OXPHOS. These studies evaluated acute exogenous genotoxic stress, typically created by 1-2-h cell exposure to genotoxic chemicals (37,60). In contrast, our results show that chronic activation of DDR rendered by rad52⌬ mutation or by growing WT cells in the presence of sublethal concentrations of genotoxic chemicals activates transcription of the TCA cycle, ETC, and OXPHOS genes and elevates oxygen consumption. To deter-mine whether the duration of the genotoxic stress is responsible for the difference between our results and studies that found the inhibitory effect of DDR on transcription of respiratory genes (37, 60), we performed a time-course experiment and measured oxygen consumption and transcription of CIT1, IDH1, QCR7, and COX1 genes for several hours after addition of bleocin (Fig. 9). Indeed, the most significant increase in oxygen consumption (Fig. 9A) and transcription of the respiratory genes ( Fig. 9B) is observed after more than 2 h of growth in the presence of bleocin. These results show that increased respiration is a long-term response to a chronic sublethal genotoxic stress and suggest that long-term survival and growth under chronic genotoxic stress requires increased respiration to support ATP and dNTP synthesis.

Discussion
The key finding of this study is that the DDR activates respiration to increase ATP production and elevate dNTP levels, which are required for efficient DNA repair. Based on our results, we propose a model in which DDR regulates dNTP

DNA damage response activates respiration
synthesis by a bifurcating mechanism (Fig. 10). In one wellestablished branch of the pathway, Dun1p inactivates Crt1p, Sml1p, and Dif1p, leading to increased RNR activity and dNTP synthesis (31)(32)(33)(34). In the second branch of the pathway, Mec1p and Rad53p down-regulate transcription of histone genes (Fig.  10). Decreased histone levels result in altered chromatin structure and induction of TCA cycle and ETC genes required for respiration (40). A direct outcome of elevated respiration is increased production of ATP, a potent allosteric activator for the RNR enzyme (16,18), increased synthesis of dNTPs, and improved cell survival (Fig. 10). This "histone" branch of the pathway does not require Dun1p and inactivation of Crt1p, Sml1p, and Dif1p, because inducing DDR in dun1⌬ cells activates respiration almost to the same level as in the WT cells (Fig. 3, A and C).
Based on chromatin architecture, yeast genes belong to one of two broad groups: growth genes and stress genes (49). Growth genes are expressed rather constitutively, and their promoters feature a nucleosome-free region where transcription factors bind upstream of the ORF. Stress genes are expressed at a lower level, and their promoters are dominated by delocalized nucleosomes rather than by the nucleosome-free region. Consequently, stress genes are regulated by factors that affect the structure of chromatin, including histone level. The respiratory genes in S. cerevisiae belong to the stress category, unlike respiratory genes in higher eukaryotes (50). Consequently, reduced histone expression or a defect in chromatin assembly induces respiration by allowing increased activation of the TCA cycle, ETC, and OXPHOS genes by the Hap2/3/ 4/5p complex (40).

DNA damage response activates respiration
Although it is well-established that DNA replication is coordinated with transcription of histone genes and that DDR represses histone transcription, the role of Rad53p in this process is not fully understood. The expression of histone genes is regulated by two G 1 /S-specific transcription complexes SBF and MBF, in addition to other transcription regulators (35,36). Swi4p/Swi6p and Mbp1p/Swi6p form SBF and MBF, respec-tively. DDR induces Rad53p-dependent phosphorylation of Swi6p, which results in down-regulation of CLN1 and CLN2 transcription and delayed G 1 to S progression (58,59). On the other hand, Rad53p phosphorylates and inactivates Nrm1p, the co-repressor of MBF, which results in activation of MBF targets (23,24). Systematic phosphoproteomics screen identified Swi6p, Swi4p, and Mbp1p as direct targets of Rad53p (60).

DNA damage response activates respiration
Because the transcription of the histone genes is reduced in swi4⌬ and mbp1⌬ cells (61), the simplest explanation for the role of Rad53p in regulation of histone gene transcription is that Rad53p phosphorylates and down-regulates SBF and MBF complexes.
How does elevated respiration and ATP levels regulate dNTP synthesis? The large subunit of both mammalian and yeast RNR contains the catalytic site as well as two allosteric sites (16,17,19). One of the allosteric sites, the "specificity site," binds dTTP, dGTP, and dATP and regulates the appropriate ratios among the four dNTP pools. The second allosteric site, the "activity site," binds ATP or dATP and regulates the total dNTP pool size by monitoring the dATP/ATP ratio. When the cellular ATP-to-dATP ratio increases, the binding of ATP to this allosteric site activates RNR, promoting synthesis of all dNTPs. When the dNTP concentration reaches a certain level, the RNR activity is allosterically inhibited by binding of dATP to the "activity site." DNA replication fidelity requires correct absolute and relative concentrations of the dNTPs (62,63), and mutations in both the "specificity" and "activity" sites of yeast RNR result in significantly reduced replication fidelity (14,63).
In yeast RNR, the allosteric dATP feedback inhibition is more relaxed, allowing the increase of dNTP pools upon DNA damage (14). The increase in dNTP pools significantly improves survival following DNA damage; however, it also results in higher mutation rates (14,57,64). Our results suggest that the DDR-induced expansion of the dNTP pools is partly facilitated by elevated respiration and ATP production.
The relationship between respiratory metabolism and DNA replication and repair is contentious. Leakage of electrons from the ETC is one of the endogenous sources of ROS, which damage cellular structures, including DNA, contributing to the pathogenesis of cardiovascular diseases, inflammatory diseases, and cancer, and a shorter life span (65,66). However, mitochondria can also function as a cellular antioxidant defense, and increased mitochondrial activity enables more efficient operation of the ETC, limiting ROS production and increasing antioxidant capacity (67). In addition, DNA replication and repair are energetically costly (45), and ETC and OXPHOS generate significantly more ATP than glycolysis. This energetic aspect of DNA repair is evolutionarily conserved, as illustrated by increased fatty acid oxidation, oxidative phosphorylation, and oxygen consumption in response to both chronic endogenous and acute exogenous genotoxic stress in mice (68).
DNA damage induced by methylmethane sulfonate (MMS) was reported to suppress respiration (69), down-regulate yeast AMP-activated protein kinase ortholog Snf1p (43), and suppress transcription of genes regulated by the Hap1p and Hap2/ 3/4/5p complex (37, 60). Hap1p and Hap2/3/4/5p activate genes encoding enzymes of the TCA cycle, ETC, and OXPHOS.  In one well-established branch of the pathway, Dun1p inactivates Crt1p, Sml1p, and Dif1p, leading to increased RNR activity and dNTP synthesis. In the second branch of the pathway, Mec1p and Rad53p downregulate transcription of histone genes. Decreased histone levels result in altered chromatin structure and induction of TCA cycle and ETC genes required for respiration. A direct outcome of elevated respiration is increased production of ATP, a potent allosteric activator for the RNR enzyme, increased synthesis of dNTPs, and improved cell survival.

DNA damage response activates respiration
The repression of Hap1p and Hap2/3/4/5p targets was independent of checkpoint kinases and was not observed in cells exposed to ionizing radiation (37). The authors concluded that the MMS-induced repression of Hap1p and Hap2/3/4/5p targets was not specific for DNA damage and was a consequence of oxidative stress or other effect of MMS on Hap1p and Hap2/3/ 4/5p signaling (37,60). These results are in agreement with our observations. When we tested different genotoxic chemicals for respiration inducers, we also included MMS. Even with MMS concentrations spanning from 0.0001 to 0.01%, we could not detect any stimulatory effect on oxygen consumption, although we observed consistent and marked stimulation of oxygen consumption in cells treated with bleocin or 4-NQO (Fig. 2). We conclude that the effect of MMS on respiration is not representative of genotoxic chemicals and DDR but rather reflects the particular properties of MMS and/or the specific pathways MMS affects. This conclusion is supported by direct inhibition of respiration in isolated mitochondria by MMS (69).
This study connects respiratory metabolism and DDR, two processes deemed not to be very compatible. We speculate that the benefit of increased ATP and dNTP synthesis for cell survival offsets the deleterious effect of respiratory metabolism on DNA repair.

Yeast strains, media, and plasmid construction
All yeast strains used in this study are listed in Table 2. Standard genetic techniques were used to manipulate yeast strains and to introduce mutations from non-W303 strains into the W303 background (71). Cells were grown at 30°C in yeast extract/peptone/dextrose (YPD) medium containing 2% glucose or under selection in synthetic complete medium containing 2% glucose and, when appropriate, lacking specific nutrients to select for a particular genotype. Cell cycle arrest in G 1 phase by ␣-factor was carried out by adding ␣-factor to 10 g/ml to cells exponentially growing in YPD medium. Following ␣-factor addition, the cultures were incubated for 3 h, and the arrest was monitored by examining cell morphology (72). For construction of plasmid pRS426-HTA1/HTB1-HHF1/ HHT1, the HTA1/HTB1 locus was amplified by PCR using forward primer 5Ј-TTCACACGAGCGAATTCTCTGAAG-3Ј and reverse primer 5Ј-AGCAACAGTGCTCGAGGAACC-TAA-3Ј. The HHF1-HHT1 locus was amplified using forward primer 5Ј-AAATACGAGCTCCGTGTAAGTTACAGAC-3Ј and reverse primer 5Ј-TTTCGAGGGGATCCCCAGGAAAA-3Ј. The HHF1-HHT1 fragment was digested with SacI and BamHI and ligated into pRS426. The HTB1-HTA1 fragment was digested with EcoRI and XhoI and ligated into the pRS426-HTA1/HTB1 plasmid.

Oxygen consumption measurement
Oxygen consumption measurements were performed essentially as described (40,46). Cells were grown to an A 600 nm of 0.6 in YPD medium containing 2% glucose, and 3 A 600 nm units (9 ϫ 10 7 ) of yeast cells were harvested by centrifugation. Cells were resuspended in a buffer containing 10 mM HEPES, 25 mM K 2 HPO 4 , pH 7.0, and incubated at 30°C in an oxygen consumption chamber (Instech Laboratories, Inc.) connected to a Neo-FOX fluorescence-sensing detector using NeoFOX software (Ocean Optics, Inc.). Results were calculated as picomoles of O 2 /10 6 cells/s and expressed as percentages of the WT value. The oxygen consumption rate in WT cells grown in YPD medium was 5.08 pmol/10 6 cells/s and was set as 100%.

Cellular ATP assays
Cellular ATP levels were determined as described (40,46). Cells were grown to an A 600 nm of 0.6 in YPD medium containing 2% glucose, and 3 A 600 nm units (9 ϫ 10 7 ) of yeast cells were harvested, and cells were harvested by centrifugation and lysed in 5% TCA with pre-chilled glass beads. Cell lysate was neutralized to pH 7.5 with 10 M KOH and 2 M Tris-HCl, pH 7.5. ATP levels were measured using the ENLITEN ATP assay (FF2000, Promega) according to the manufacturer's instructions and normalized by the number of cells. The ATP level in WT cells grown on YPD medium was 0.58 mol/10 10 cells and was set as 100%.

Western blotting
Whole-cell lysates were prepared, and Western blotting was performed as described previously (40). Briefly, cells were grown in YPD medium containing 2% glucose to an A 600 nm of 0.6. Four A 600 nm units (1 A 600 nm unit is equal to 3 ϫ 10 7 cells) of yeast cells were harvested and immediately boiled in SDS sample buffer. Anti-histone H3 polyclonal antibody (Abcam, ab1791) was used at a dilution of 1:1000, and anti-Pgk1p (Invitrogen, 459250) was used at a dilution of 1:3000.

dNTP quantitative analysis
Four A 600 nm units (12 ϫ 10 7 ) of yeast cells were harvested and lysed in 5% TCA with pre-chilled glass beads. Cell lysate was neutralized to pH 7.5 with 10 M KOH and 2 M Tris-HCl, pH 7.5. The lysate was used immediately for probe-based quantitative PCR analysis of individual dNTP levels or aliquoted and stored at Ϫ70°C until the time of analysis. The procedure for the fluorescence-based dNTP quantitative analysis was essentially as described previously with minor modifications (74). Specifically, both detection templates (DT) 1 and 2 for each dNTP were used, and the results were compared. Fluorescence signal was recorded every 5 min (1 cycle) for a total of 50 min (10 cycles) to monitor the kinetics of individual dNTP incorporation, and the fluorescence signal after a 40-min incubation (8 cycles) was used for calculation of the normalized fluorescence units. Standard curves using appropriate individual dNTPs were established for assay validation.

Spotting assay
Cells were grown to log phase at 30°C, and 10-fold serial dilutions were spotted on the YPD plates with or without genotoxic chemicals and incubated at 30°C for 48 -72 h.

Statistical analysis
The results represent at least three independent experiments. Numerical results are presented as means Ϯ S.D. Data were analyzed by using an InStat software package (Graphpad, San Diego). Statistical significance was evaluated by one-way analysis of variance, and p Ͻ 0.05 was considered significant.