SIRT2 promotes base excision repair by transcriptionally activating OGG1 in an ATM/ATR-dependent manner

Abstract Sirtuin 2 (SIRT2) regulates the maintenance of genome integrity by targeting pathways of DNA damage response and homologous recombination repair. However, whether and how SIRT2 promotes base excision repair (BER) remain to be determined. Here, we found that independent of its catalytic activity SIRT2 interacted with the critical glycosylase OGG1 to promote OGG1 recruitment to its own promoter upon oxidative stress, thereby enhancing OGG1 promoter activity and increasing BER efficiency. Further studies revealed that SIRT2 was phosphorylated on S46 and S53 by ATM/ATR upon oxidative stress, and SIRT2 phosphorylation enhanced the SIRT2-OGG1 interaction and mediated the stimulatory effect of SIRT2 on OGG1 promoter activity. We also characterized 37 cancer-derived SIRT2 mutants and found that 5 exhibited the loss of the stimulatory effects on OGG1 transcription. Together, our data reveal that SIRT2 acts as a tumor suppressor by promoting OGG1 transcription and increasing BER efficiency in an ATM/ATR-dependent manner.


Introduction
A rise in genomic instability may aberrantly inactivate tumor suppressor genes while activating oncogenes, thereby promoting tumorigenesis ( 1 ,2 ).In response to different kinds of DNA damage, various types of DNA repair pathways have evolved to efficiently eliminate the damage to preserve genomic stability, therefore suppressing tumorigenesis ( 3 ,4 ).Among all types of DNA damage, damage to DNA bases arising from endoge-nously generated reactive oxygen species (ROS), exogenous chemicals or irradiation occurs very frequently.It is estimated that thousands of DNA base lesions are generated per human cell per day ( 5 ).Therefore, efficient base excision repair (BER) is critical to the maintenance of genome integrity.The BER pathway is activated by the removal of the damaged bases by a glycosylase such as OGG1.The resulting abasic sites are further cleaved by the AP endonuclease APE1, generating DNA nicks that need to be processed by a number of DNA end processors, including TDP1 / 2, PNKP, Aprataxin, APE1 and Pol β.The processed DNA ends with a one-nucleotide gap are filled by Pol β, and this event is followed by XRCC1-Lig3or Lig1-mediated ligation of the single-strand break (SSB) to complete the BER process (6)(7)(8).Although knocking out key BER factors such as Pol β and XRCC1 in mice often leads to embryonic lethality ( 9 ,10 ), cells of numerous different types of cancer tissues harbor mutated Pol β, and several XRCC1 polymorphisms have been associated with breast cancer or lung cancer in Asian patients ( 11 ,12 ).In addition, Pol β expression negatively regulates the progression of breast cancer and lung cancer by promoting the demethylation of the CDH13 promoter ( 13 ).Moreover, simultaneously knocking out the two redundant glycosylases, Ogg1 and Myh increases the incidence of tumorigenesis in mice ( 14 ).
The ATM (ataxia-telangiectasia mutated) and ATR (ATMand Rad3-Related) kinases are two of the farthest upstream kinases that respond to a variety of types of DNA damage (15)(16)(17).Although most studies on ATM and ATR are related to DNA double-strand breaks (DSBs) or defects in DNA replication progression, accumulating evidence indicates that ATM, ATR or their downstream substrates are involved in regulating oxidative stress-induced DNA damage ( 18 ,19 ).ATM directly participates in BER by phosphorylating TDP1 at Ser81 to promote XRCC1-mediated recruitment of TDP1 to base damage sites ( 20 ).CHK2, the downstream kinase of ATM, phosphorylates XRCC1 at Thr284 to facilitate its recruitment to DNA damage sites ( 21 ).Although the direct roles of ATR and its downstream kinase CHK1 in the BER pathway remain to be established, the 9-1-1 complex, which is regulated by the ATR-CHK1 axis, promotes BER by targeting several BER factors (22)(23)(24)(25)(26).
Here, we demonstrated that SIRT2 promotes BER independent of its deacetylase activity.In response to oxidative stress, A TM and A TR phosphorylate SIRT2 at S46 and S53.Phosphorylation of SIRT2 enhances its interaction with the glycosylase OGG1, resulting in the recruitment of OGG1 to its own promoter to increase OGG1 transcription and BER effi-ciency.We also identified several cancer-associated SIRT2 mutants with loss of the stimulatory effects on OGG1 transcription and BER efficiency, indicating that SIRT2 functions as a tumor suppressor by promoting BER for genome stabilization.

Plasmid transfection and reagents
The ORF of SIRT2 fused to Flag or GFP at the C-terminus was inserted into the backbone of pEGFP-N1.The ORF of OGG1 fused to His or GFP at the C-terminus was inserted into the backbone of pEGFP-N1.Mutants of SIRT2 and OGG1 were generated based on wild-type plasmids.
For HC A2-hTER T fibroblasts, transfection was performed on a Lonza 4D electroporator with the program DT-130.Eukaryotic expression plasmids were introduced into HEK293 cells using the polyethylenimine (PEI) transfection method.

FACS analysis
Forty-eight to seventy-two hours post-transfection, cells were harvested and resuspended in 0.2 ml PBS for FACS analysis on a FACSVerse flow cytometer (BD Biosciences, USA).At least 20 000 events were analyzed.All data were subsequently analyzed using FlowJo software.

Survival assay and FACS analysis of apoptosis
HC A2-hTER T fibroblasts were seeded at a density of 8 × 10 4 cells per plate.On Day 2 post-seeding, the cells were treated with 20 mM KBrO 3 for 1 hour and then cultured in drugfree medium for the indicated times.Cells were counted and harvested according to the manufacturer's instructions for the apoptosis assay with an Annexin V-FITC Apoptosis Detection Kit (Beyotime, Cat.# C1062L).Samples were analyzed on a BD Accuri C6 flow cytometer (BD Biosciences, USA).

Analysis of BER efficiency
The pEGFP-N1 vector was incubated with methylene blue at a concentration of 500 μM, and the mixture was irradiated with visible light at a distance of 18 cm for 3 h.The damaged pEGFP-N1 vector was then purified and cotransfected with a vector encoding DsRed2 into cells.On Day 3 posttransfection, cells were harvested for FACS analysis.The ratio of GFP + cells to DsRed2 + cells was used to measure BER efficiency.

Luciferase assay
HC A2-hTER T fibroblasts were electroporated with the indicated firefly luciferase reporter vectors and pRL-SV40 Renilla vectors prior to a 2-day incubation.On Day 3 posttransfection, cells were collected and subsequently lysed according to the manufacturer's instructions.Relative luciferase activities were measured by a dual-luciferase reporter system (Promega, Cat.# E1910) using a GloMax Luminometer (Promega, Cat.# E5311).Each assay was performed at least three times independently.

Immunoprecipitation
Cells were harvested, washed once with 1 × PBS, and lysed with lysis buffer (20 mM HEPES, pH 8.0; 0.2 mM EDTA; 5% glycerol; 150 mM NaCl; 1% NP-40) for 15 min on ice.The lysate was then sonicated on ice for 3 min before being centrifuged at 13 500 rpm for 15 min at 4 • C. The supernatant was then prepared for immunoprecipitation by two methods.In the first method, the supernatant was pretreated with protein A / G agarose (Abmart, Cat.# A10001M) and IgG (Santa Cruz, #sc-2025) for 1 h at 4 • C.After centrifugation, the supernatant was collected and incubated with the indicated antibodies overnight.Afterward, the samples were incubated for 2 h at 4 • C after 30 μl of protein A / G agarose was added to the mixture.In the other method, cell lysates were incubated with anti-DYKDDDK magnetic beads (Selleck, Cat.# B26102) or GFP-Trap (ChromoTek, Cat.# gta-100) overnight at 4 • C. All precipitated samples were washed three times with lysis buffer and boiled for 10 min with 2 × sample buffer (1% SDS; 10 mM EDTA; 50 mM Tris-HCl, pH 8.1; and protease inhibitors).Whole-cell lysates were used as input as a control.

ChIP assay
HEK293 cells were transfected with 5 μg of the indicated plasmids and cultured for 24 h before being treated with 40 mM KBrO 3 .Thirty minutes after KBrO 3 treatment, cells were harvested for the ChIP assay as previously described ( 76 ).After precipitation, chromatin and input fragments were used as templates for real-time PCR with the following primers: 5 -CGGCTCTCGGA GAA CGGGCTG-3 and 5 -GTTCTCC AGGGGC AGACC AC A-3 .

Alkaline comet assay
Cells were seeded at a density of 2 × 10 4 cells / well in 6-well plates.On Day 3, cells were treated with KBrO 3 at a concentration of 40 mM for 0.5 h or with 100 μM H 2 O 2 for 4 h.Then, the cells were collected, resuspended in PBS and diluted to 3 × 10 5 cells per ml before the comet assay was performed.The detailed procedure is described in the manufacturer's instructions (Trevigen, Cat.# 4250-050-K).For formamidopyrimmidine DNA glycosylase (FPG) treatment, on Day 2 post seeding, cells were treated with KBrO 3 at a concentration of 15 mM for 0.5 h, and then recovered in free-drug medium for indicated time.After lysis, the slides were washed 3 times with enzyme reaction buffer (40 mM HEPES, 0.1 M KCl, 0.5 mM EDTA, 0.2 mg / ml BSA, pH 8.0 with KOH), and then incubated with FPG in a 1:5000 dilution with the enzyme buffer at 37 • C for 1 h.Tail moments were used to quantify the amount of DNA damage using CometScore software (cas-plab_1.2.3b2).

SIRT2 increases BER efficiency
To test whether SIRT2 participates in the maintenance of genome integrity in the presence of oxidative stress, we first performed a comet assay to analyze the changes in genomic stability in control or SIRT2-depleted HCA2-hTERT cells upon oxidative stress induced by KBrO 3 or H 2 O 2 .We found that the tail moment, which reflects the degree of genomic instability, was significantly increased in SIRT2-depleted HCA2-hTERT cells in the presence of KBrO 3 or H 2 O 2 (Figure 1 A-C, Supplementary Figure S1 a, b).In contrast, under normal conditions, depleting SIRT2 had either a very weak or nonsignificant effect on genomic instability (Figure 1 A, B, Supplementary Figure S1 a, b).These data suggest that SIRT2 might be involved in regulating the repair of oxidative stressinduced DNA damage.
Therefore, we manipulated the expression of SIRT2 and analyzed the BER efficiency in HC A2-hTER T cells using our well-established plasmid reactivation assay for quantifying BER efficiency ( 52 ).The pEGFP-N1 plasmid was treated with methylene blue together with visible light to induce 8-oxodG damage ( 53 ), and the damaged plasmids were transfected into cells together with a plasmid expressing DsRed2 as an internal control for monitoring the difference in the transfection efficiency between experiments.Afterward, the ratio of GFP + cells to DsRed2 + cells was employed as a measure of BER efficiency (Figure 1 D).We observed an increase of approximately 2-fold in BER efficiency in HC A2-hTER T cells overexpressing SIRT2 ( Supplementary Figure S1 c-d).Consistent with the results of SIRT2 overexpression, depleting SIRT2 significantly reduced BER efficiency by approximately 2-fold (Figure 1 E).Restoration of SIRT2 expression in SIRT2 knockdown cells rescued the decline in BER efficiency (Figure 1 E).Consistently, using a formamidopyrimmidine DNA glycosylase (FPG)-modified alkaline comet assay, we observed that the tail moment of control cells was reduced by 36.14% from 1.5 to 3 h post KBrO 3 treatment, while that of SIRT2 depleted cells was less pro-nounced with a reduction by 18.56% (shSIRT2-1) or 25.01%(shSIRT2-2) ( Supplementary Figure S1 e, f), further confirming that SIRT2 participates in the repair of KBrO 3 -induced DNA base damage.
Then, we analyzed how SIRT2 affects cell proliferation upon oxidative stress.We found that depleting SIRT2 significantly reduced the proliferation of HC A2-hTER T cells in the presence of KBrO 3 (Figure 1 F).We also examined how SIRT2 affected the apoptosis rate in HC A2-hTER T cells upon oxidative stress.We found that depleting SIRT2 significantly increased oxidative stress-induced early apoptosis by ∼2-fold and late apoptosis by ∼2to 3-fold (Figure 1 G, H).
Together, these results suggest that SIRT2 increases BER efficiency for genome stabilization, thereby promoting cell survival and inhibiting apoptosis.

SIRT2 promotes the expression of OGG1 at the transcriptional level
We first examined whether SIRT2 is present at the oxidative stress-induced DNA damage sites using a well-established killer-red assay ( 54 ,55 ).The assay revealed that SIRT2 is not recruited to DNA damage sites to regulate BER ( Supplementary Figure S2 ).
We therefore tested whether SIRT2 affects the expression of important BER factors in SIRT2-depleted cells.We found that knocking down SIRT2 led to a decrease in the protein level of the glycosylase OGG1 but not other key BER factors (Figure 2 A).Consistent with this finding, overexpressing SIRT2 increased the protein level of OGG1 in a dose-dependent manner (Figure 2 B).Analysis of the OGG1 mRNA level through reverse transcription-quantitative PCR revealed that depleting SIRT2 reduced the mRNA level of OGG1, while overexpressing SIRT2 increased the mRNA level of OGG1 (Figure 2 C-E), indicating that the regulation of OGG1 expression by SIRT2 occurs through transcriptional mechanisms.We then cloned a 2.3-kb OGG1 promoter that drives the expression of the firefly luciferase gene (Figure 2 F).Using the resulting P OGG1 -firefly luciferase vector, we found that depleting SIRT2 suppressed OGG1 promoter activity, while overexpressing SIRT2 stimulated OGG1 promoter activity (Figure 2 G, H), confirming that SIRT2 regulates OGG1 transcription.
We then separated the OGG1 promoter into three fragments: -2147 to -1198, -1197 to -298 and -297 to +3.By overexpressing SIRT2, we analyzed the changes in the activity of the promoter fragments.We found that SIRT2 retained its stimulatory effect on OGG1 promoter activity only when the -297 to +3 region was intact, whereas the stimulatory effect of SIRT2 was abolished with the other two fragments (Figure 2 I, J).Further chromatin immunoprecipitation (ChIP) assays revealed that SIRT2 was recruited to the OGG1 promoter at the -297 to +3 region in response to oxidative stress (Figure 2 K).
Importantly, data mining revealed that the mRNA level of SIRT2 was positively correlated with the OGG1 mRNA level in 18 of 33 different types of cancers, confirming that SIRT2 promotes OGG1 expression at the transcriptional level (Figure 2 L).
Taken together, these data indicate that SIRT2 is recruited to the OGG1 promoter to increase its transcriptional activity, thereby promoting OGG1 expression and the repair of DNA base damage.

SIRT2 physically interacts with the OGG1 protein
Since SIRT2 and other members of the Sirtuin family participate in DNA repair by directly interacting with and modifying repair factors, we proposed that SIRT2 might regulate BER by targeting important BER factors.A vector encoding Flagtagged SIRT2 was transfected into HEK293 cells, and co-IP experiments were then performed with beads coated with an antibody against Flag.The data indicated that OGG1 but no other core BER factors interacted with SIRT2 ( Supplementary Figure S3 a).Then, we performed co-IP experiments with an antibody against the His tag using lysates of cells transfected with vectors expressing His-tagged OGG1 and Flag-tagged SIRT2.The results demonstrated that OGG1 interacted with SIRT2 in cells (Figure 3 A).The reciprocal co-IP experiment further confirmed this conclusion (Figure 3 B).We also performed in vitro co-IP experiments with purified recombinant SIRT2 and OGG1 proteins.The data revealed that SIRT2 and OGG1 interacts directly (Figure 3 C).
Then, we determined whether the SIRT2-OGG1 interaction is enhanced upon oxidative stress.To this end, we treated cells with KBrO 3 or H 2 O 2 to induce oxidative stress and then performed co-IP experiments.We found that the interaction between SIRT2 and OGG1 was enhanced in response to KBrO  Furthermore, we performed co-IP experiments to determine which domains of the two factors interact with the other protein.We found that removing either the N-terminal or Cterminal domain or removing both the N-and C-terminal domains of SIRT2 did not abolish its interaction with OGG1 (Figure 3 E, F), suggesting that the central catalytic domain of SIRT2 is responsible for its interaction with OGG1.We also observed that the N-terminal and central domains but not the C-terminal domain of OGG1 interacted with SIRT2 (Figure 3 G, H).

SIRT2 facilitates the recruitment of OGG1 to its own promoter to increase OGG1 transcription
We hypothesized that SIRT2 might deacetylate OGG1 to regulate BER.To test this hypothesis, we performed an in vitro deacetylation assay using purified recombinant SIRT2 and OGG1 proteins.We found that SIRT2 did not deacetylate OGG1 in vitro (Figure 4 A).Further co-IP experiments in SIRT2-depleted cells revealed that SIRT2 depletion did not change the acetylation level of OGG1 ( Supplementary Figure S4 a).These data indicate that the regulation of BER by SIRT2 probably does not occur through deacetylation of OGG1.Further analysis of BER efficiency using vectors encoding the enzymatically dead SIRT2 H187Y mutant indicated that SIRT2 promoted BER efficiency independent of its enzymatic activity (Figure 4 B, C).
In addition to its function in the BER pathway, OGG1 has been reported to function as an important transcription modulator and bind to specialized promoter structures ( 56 ).Since SIRT2 promotes BER efficiency independent of its enzymatic activity and its interaction with OGG1 is enhanced by oxidative stress, we proposed that SIRT2 might facilitate the recruitment of OGG1 to its own promoter to stimulate its expression in response to oxidative stress.
We therefore first examined whether OGG1 is recruited to its own promoter to enhance its own transcriptional activity.The ChIP assay revealed that OGG1 was recruited to its promoter at the -297 to +3 region, and oxidative stress further stimulated its recruitment to its own promoter (Figure 4 D, E).In addition, we found that overexpression of OGG1 stimulated the activity of its own promoter in a dose-dependent manner (Figure 4 F) and that depletion of OGG1 impaired its promoter activity (Figure 4 G).Further luciferase assay revealed that the -197 to +3 region in OGG1 promoter is required for OGG1-mediated activation of its own promoter ( Supplementary Figure S4 b, c).By analyzing the OGG1 promoter activity in cells overexpressing OGG1 catalytically dead mutants OGG1 K249A or D268N ( 57 ,58 ), we also demon- strated that OGG1 stimulated its promoter activity independent of its canonical enzymatic activity ( Supplementary Figure S4 d).A previous report indicates that OGG1 might cooperate with transcription factors including TFIID, SP1 and NF-κB to regulate transcription of different genes ( 56 ), so we further examined whether OGG1 stimulates its own promoter activity through these transcription factors.The luciferase assay revealed that depleting OGG1 diminished the stimulatory effect of SP1 and TFIID on OGG1 promoter activity ( Supplementary Figure S4 e, f), indicating the two transcription factors are involved in the regulation of OGG1 promoter by OGG1 itself.Next, we aimed to understand whether SIRT2 interacts with OGG1 to promote the recruitment of OGG1 to its own promoter, thereby stimulating its expression and BER efficiency.The ChIP assay revealed that overexpressing SIRT2 promoted the recruitment of OGG1 to its own promoter (Figure 4 H), while knocking down SIRT2 suppressed this recruitment in response to KBrO 3 -induced oxidative stress (Figure 4 I).Further in vitro Electrophoretic Mobility Shift Assay (EMSA) revealed that OGG1 but not SIRT2 directly bound to its own promoter (-197 to +3), while supplementing SIRT2 greatly enhanced the binding of OGG1 to its own promoter ( Supplementary Figure S4 g-i).
Consistent with this finding, depleting endogenous OGG1 reduced its promoter activity and completely abolished the stimulatory effect of SIRT2 on OGG1 promoter activity (Figure 4 J).ChIP assay also revealed that knocking down endogenous OGG1 reduced the recruitment of SIRT2 to OGG1 promoter ( Supplementary Figure S4 j).Simultaneously overexpressing SIRT2 and OGG1 synergistically increased OGG1 promoter activity by ∼7.5-fold, while overexpressing SIRT2 or OGG1 alone increased OGG1 promoter activity by ∼2.5fold or ∼2.0-fold, respectively (Figure 4   These data demonstrate that SIRT2 probably forms a complex with OGG1 to stimulate the recruitment of OGG1 to its own promoter, thereby enhancing its promoter activity to increase OGG1 expression.After OGG1 binds to the promoter, the SIRT2-OGG1 complex might fall apart and SIRT2 might quickly fall off the OGG1-DNA complex.

The promotion of BER by SIRT2 is dependent on the ATM and ATR kinases
Since the SIRT2-OGG1 interaction was enhanced under oxidative stress (Figure 3 D, Supplementary Figure S3 b), we proposed that SIRT2 is regulated by DNA damage response factors such as the ATM and ATR kinases.However, ATM and ATR are primarily involved in regulating DSB repair; thus, we first examined whether they are activated by oxidative stress.We found that treating cells with KBrO 3 caused upregulation of ATM or ATR phosphorylation (Figure 5 A), indicating that the two kinases also participate in the regulation of the oxidative stress-induced DNA damage response.The conserved motif that ATM and ATR recognize and phosphorylate is the SQ / TQ cluster ( 59 ).To test whether SIRT2 is phosphorylated by A TM / A TR upon oxidative stress, we performed co-IP experiments followed by Western blot analysis using a widely used antibody against phosphorylated SQ / TQ.The results revealed that in response to DNA damage induced by KBrO 3, SIRT2 was phosphorylated (Figure 5 B).Further co-IP experiments revealed that SIRT2 interacted with ATM or ATR (Figure 5 C, D).Pretreating cells with ATM or ATR inhibitors abolished the oxidative stress-induced phosphorylation of SIRT2 (Figure 5 E).Moreover, inhibiting ATM or ATR kinase activity with their specific inhibitors attenuated the stress-induced enhancement of the SIRT2-OGG1 interaction (Figure 5 F) and abolished the stimulatory effect of SIRT2 on OGG1 promoter activity (Figure 5 G, H).By analyzing the amino acid sequence of SIRT2, we predicted that the amino acid residues Ser46 and Ser53 were two potential sites that could be phosphorylated by ATM and ATR, and we constructed vectors expressing SIRT2 S46A, S53A and S46AS53A.We further examined whether the mutations abolished the enhancing effect of SIRT2 on BER efficiency.We found that mutating both of these serine residues into alanine residues abolished the stimulatory effect (Figure 5 I, Supplementary Figure S5 a), indicating that both residues are critical to the promotion of BER by SIRT2.In vitro phosphorylation experiments demonstrated that ATM or ATR may phosphorylate recombinant SIRT2 WT protein but not SIRT2 S46AS53A mutant ( Supplementary Figure S5 b, c).Further co-IP experiments revealed that the two mutations attenuated the KBrO 3 -induced phosphorylation of SIRT2 (Figure 5 J) and abolished the KBrO 3 -induced enhancement of the SIRT2-OGG1 interaction (Figure 5 K).Overexpressing the SIRT2 S46AS53A mutant also failed to stimulate OGG1 pro-moter activity (Figure 5 L), promote the recruitment of SIRT2 to OGG1 promoter region (Figure 5 M) and increase the OGG1 mRNA level (Figure 5 N).
Taken together, our data demonstrate that in response to oxidative stress, A TM and A TR phosphorylate SIRT2 at amino acid residues S46 and S53 to promote its interaction with OGG1, thereby increasing OGG1 transcription and improving BER efficiency.

Several cancer-associated SIRT2 mutations impair its capacity to promote BER
To determine the functional consequences of SIRT2-mediated regulation of OGG1 transcription, we identified 37 cancerassociated mutations in SIRT2 through data mining using the TCGA database (Figure 6 A, Supplementary Figure S6 a).We then constructed vectors expressing these SIRT2 mutants.Since cancer-associated mutations often cause rapid protein  degradation ( 60 ,61 ) and these mutations apparently affect protein functions, we first examined whether these mutations affect the protein level of SIRT2.We found that 5 of the 37 mutations led to downregulation of the protein expression of SIRT2 (T63I, R97C, R163C, E173D and G177W) (Figure 6 B); thus, we excluded these mutants from further analysis.Using the remaining 32 mutants, we analyzed the efficiency of BER in cells overexpressing these SIRT2 mutants.We found that the stimulatory effect of 18 of the 32 mutants was partially impaired or completely abolished (Figure 6 C).The relative BER efficiency in cells overexpressing these mutants was at most 80% of that in cells overexpressing SIRT2 WT.We then further examined whether these SIRT2 mutants affected OGG1 promoter activity.We found that 10 of them failed to enhance or partially lost their stimulatory effects on OGG1 promoter activity (Figure 6 D).The relative OGG1 promoter activity in cells overexpressing the 10 mutants was at most 70% of that in cells overexpressing SIRT2 WT.Real-time quantitative PCR analysis further confirmed that the 5 SIRT2 mutants (R69L, F124S, P128L, R153H and G291D) were unable to promote OGG1 mRNA expression as high as SIRT2 WT (Figure 6 E).The mRNA level of OGG1 in cells overexpressing the five mutants was less than 80% of that in cells overexpressing SIRT2 WT.Data mining revealed that patients with the five mutations in SIRT2 had high mutation rates the genomes ( Supplementary Figure S6 a).Co-IP experiments revealed that three of the five mutants R69L, R153H and G291D could not be phosphorylated in response to oxidative stress ( Supplementary Figure S6 b).Furthermore, we performed softagar assay to examine whether these mutations affect tumorigenesis ( 62 ).Indeed, we found that overexpressing SIRT2 WT reduced the rates of tumorigenesis while all the five mutations abolished the suppressive effect of SIRT2 on tumorigenesis ( Supplementary Figure S6 c).
Taken together, our data demonstrate that at least five cancer-derived SIRT2 mutations reduce the capability of SIRT2 to promote OGG1 transcription to enhance BER efficiency, thereby contributing to tumorigenesis.

Discussion
In this report, our data reveal that in response to oxidative stress SIRT2 is phosphorylated at the two serine residues S46 and S53 by A TM / A TR, and the phosphorylated SIRT2 promotes the recruitment of OGG1 to its own promoter to activate its transcription, thereby increasing BER efficiency and preventing tumorigenesis ( Supplementary Figure S7 ).
In the past two decades, numerous targets of SIRT2 have been identified ( 38 ,63-65 ).As an NAD+-dependent deacetylase, SIRT2 deacetylates histone or nonhistone substrates to regulate physiological processes ( 36-38 , 63 , 65 , 66 ).SIRT2 is generally considered a tumor suppressor because its expression is downregulated in most cancers, and it restricts cell cycle progression by deacetylating cell cycle-related factors and deacetylating H4K16 to allow chromatin condensation ( 30 , 37 , 41 , 67-71 ).Our present study adds another piece of evidence that SIRT2 functions as a tumor suppressor by promoting DNA repair by BER for genome stabilization.However, previous reports also indicated that SIRT2 increases the protein stability of the oncoprotein c-Myc by transcriptionally repressing the expression of the E3 ligase NEDD4 ( 39 ), and inhibiting SIRT2 might be a good therapeutic approach in cancer treatment ( 27 , 72 , 73 ).We propose that the opposite oncogenic and tumor-suppressive roles of SIRT2 might rely on biological contexts such as the type and stage of cancer.
Other Sirtuins, such as SIRT1 and SIRT6, have been reported to function to regulate transcription and DNA repair independent of their enzymatic activities ( 74 ,75 ).To our knowledge, we are the first to demonstrate that SIRT2 plays an important role in DNA repair independent of its enzymatic activity.In addition, this finding also indicates a positive role of SIRT2 in regulating transcription, distinct from its suppressive roles in transcription through targeting histones ( 36 ).Our results indicate that beyond targeting the deacetylase activity of SIRT2, developing novel methods to enhance the association of SIRT2 with OGG1 might be a promising approach for cancer prevention and the treatment of neurodegenerative diseases that are associated with the rise in genomic instability resulting from oxidative stress ( 76 ).
In addition to functioning in tumorigenesis, SIRT2 is also reported to play important roles in neurodegenerative diseases ( 77 ), which are associated with changes in oxidative stress homeostasis ( 76 ).Moreover, although it is well established that ATM or ATR mutations may lead to genomic instability, thereby promoting tumorigenesis, patients with ATM or ATR mutations are also reported to have neurodegenerative diseases ( 78 ,79 ).Since ATM and ATR are primarily recognized as the transducers functioning in response to DSBs, while failure to repair DSBs is usually not associated with phenotypes of neurodegeneration, our finding that ATM and ATR regulate BER by modifying SIRT2 may provide a mechanistic explanation for the etiology of neurodegenerative diseases.
Notably, A TM / A TR-SIRT2 axis-mediated recruitment of OGG1 to its own promoter enhances the transcriptional activity of OGG1, which creates a positive feedback loop.Since the phosphorylation of SIRT2 by A TM / A TR and the autophosphorylation of A TM / A TR are the triggers of the positive feedback loop, we speculate that the oxidative stress-induced triggers need to be released after the oxidative stress-induced DNA damage is repaired.There might be different mechanisms to achieve this.For instance, the phosphorylation of SIRT2 could be removed via the action of phosphatases.The phosphatases that are responsible for the dephosphorylation of A TM / A TR have been identified ( 80 ,81 ), but the potential phosphatases that dephosphorylate SIRT2 remain to be determined.In addition, phosphorylated SIRT2 might have a short half-life due to possibly stronger interaction with E3 ligases.
Although our data indicate that SIRT2 regulates BER by increasing the expression of the glycosylase OGG1, which specifically removes 8-oxoG, a number of other glycosylases also participate in the BER pathway by removing different types of damaged bases, such as alkylated bases.Whether SIRT2 regulates the expression of other glycosylases remains to be determined, and the potential underlying mechanisms remain to be identified.
It is of note that both N-and C-terminals of SIRT2 prevent its interaction with OGG1.We speculate that this is a protective mechanism that helps cells to avoid overactivation of DNA repair.Only in the condition of oxidative stress, the occurrence of phosphorylation at the two serine residues in the N-terminal results in increased negative charges and hydrogen-bond networks that might change the structure of SIRT2 to allow stronger interaction with OGG1, thereby stimulating OGG1 transcription to improve DNA repair.
Notably, in this work, we analyzed 37 cancer-associated SIRT2 mutants, and only five of them exhibited loss of the stimulatory effect on OGG1 promoter activity and mRNA expression.Three mutations of them affected the phosphorylation of SIRT2, thereby inhibiting its own promoter activity.For the remaining two mutations, we propose that these mutations might affect other aspects of SIRT2.For instances, the subcellular localization of SIRT2, or the interaction between SIRT2 and OGG1, might be changed by the mutations.However, this warrants further investigation.The other 32 SIRT2 mutants might have gained oncogenic functions, such as promoting the expression of oncogenes, or lost other cancer-preventive functions, such as mediating HR repair and the stabilization of tumor suppressors.Moreover, in the 32 SIRT2 mutants, we observed some inconsistency between BER efficiency and transcriptional activity of OGG1 mediated by SIRT2.This can be explained by that there exists other SIRT2mediated regulatory mechanisms in BER.Indeed, a previous report indicates that SIRT2 might deacetylate the critical BER factor PARP1, leading to the WWP2-mediated upregulation in ubiqutination and degradation of PARP1 (49).These mutations in SIRT2 might change its regulation in PARP1 protein stability and complicate the results in BER efficiency.However, great efforts are still needed to clarify whether and how exactly these mutations affect BER pathway.

Figure 1 .
Figure 1.SIRT2 promotes BER for genome stabilization.( A ) The analysis of genomic inst abilit y in control and SIRT2-depleted HCA2-hTERT cells in the presence of the o xidativ e stress inducer KBrO 3 using the comet assay.The tail moment was employed as the measure of genomic inst abilit y, and at least 100 cells were analyzed using the software CometScore.The error bars indicate the s.e.m. values.The Mann-Whitney U test was used for statistical analysis.**** P < 0.0001; n.s., not significant.( B ) R epresentativ e images of the comet assay.( C ) Western blot analysis of SIRT2 expression in control and SIRT2-depleted HCA2-hTERT fibroblasts.( D ) Schematic depiction of the experimental principle of BER efficiency analysis.( E ) Analysis of BER efficiency in the indicated HCA2-hTERT cells.The error bars indicate the S.D. values.Student's t test was used for statistical analysis.*** P < 0.001; **** P < 0.0 0 01; n.s., not significant.( F ) The effect of depleting SIRT2 on cell proliferation.HCA2-hTER T cells w ere treated with KBrO 3 at a concentration of 20 mM for 1 h and maintained in culture for the indicated days.At different time points post-KBrO 3 treatment, cells were harvested and counted.The error bars indicate the S.D. values.Student's t test was used for statistical analysis.* P < 0.05; ** P < 0.01; *** P < 0.001; ( G, H ) Analysis of apoptosis rates of control and SIRT2-depleted HCA2-hTERT cells in the presence of KBrO 3 .On Day 3 post-KBrO 3 treatment, cells were harv ested f or analy sis of apoptosis.R epresentativ e FACS traces are sho wn in ( G ). T he cells in the upper right rectangles w ere considered early apoptotic cells, while those in the lo w er right rectangles were considered late apoptotic cells.The quantitative data are shown in ( H ).

3 or H 2 O 2 (
Figure 3 D, Supplementary Figure S3 b), suggesting that the two factors function cooperatively to promote the repair of oxidative stress-induced DNA damage.

Figure 2 .
Figure 2. SIRT2 regulates OGG1 expression through transcriptional mechanisms.( A ) Analysis of the protein expression of different BER factors in control and SIRT2-depleted HEK293 cells.( B ) Analysis of the protein expression of OGG1 in HEK293 cells transfected with the control vector or the SIR T2 o v ere xpression v ector.Quantification of O GG1 protein le v els w as perf ormed with ImageJ softw are (A, B). ( C ) T he tw o amplification regions in the OGG1 gene (R1: 1 58-31 9; R2: 350-51 4).( D ) Analysis of OGG1 mRNA levels in control and SIRT2-depleted HEK293 cells.RNA was extracted for reverse transcription and quantitative PCR with the indicated primers for amplifying the indicated regions in (C).* P < 0.05; ** P < 0.01.( E ) The mRNA level of OGG1 in control and SIRT2-overexpressing HEK293 cells.** P < 0.01; *** P < 0.001.( F ) Schematic diagram of the firefly luciferase gene driven by the OGG1 promoter (P OGG1 -Firefly Luciferase).( G ) Analysis of OGG1 promoter activity in control and SIRT2-depleted HCA2-hTERT cells.The P OGG1 -firefly luciferase vector was transfected into the cells, and 72 h post-transfection, the cells were lysed for luciferase activity measurement.* P < 0.05; *** P < 0.001.( H ) Analysis of OGG1 promoter activity in control and SIR T2-o v ere xpressing HCA2-hTER T cells.Cells w ere transfected with the luciferase reporter vector and either the control vector or a vector encoding SIRT2 at different concentrations.Sevent y-t wo hours post-transfection, cells were lysed for luciferase activity measurement.** P < 0.01.( I ) Schematic diagrams of the firefly luciferase gene driven by OGG1 promoter fragments.( J ) Analysis of fold changes in the activity of OGG1 promoter fragments in the presence versus absence of SIRT2 overexpression.** P < 0.01.( K ) ChIP analysis of the recruitment of SIRT2 to the OGG1 promoter in HEK293 cells.Cells with exogenously introduced control plasmids or pSIRT2-GFP expression vectors were treated with KBrO 3 and were then lysed and subjected to a ChIP assay with an anti-GFP antibody.**** P < 0.0 0 01.( L ) The Cleveland dot plot shows the correlation between SIRT2 and OGG1 mRNA expression in 33 different cancer types.The red dot indicates a significant Pearson correlation. K).

Figure 3 .
Figure 3. SIRT2 physically interacts with OGG1.( A, B ) Co-IP analysis of the SIRT2-OGG1 interaction.HEK293 cells were transfected with vectors encoding SIRT2-Flag and OGG1-His.Cells were lysed for co-IP with an anti-His antibody ( A ) or an anti-Flag antibody ( B ) followed by Western blot analysis.( C ) In vitro analysis of the SIR T2-O GG1 interaction.T he purified recombinant SIRT2-Flag and O GG1-GFP w ere coincubated in vitro prior to co-IP with an anti-GFP antibody.( D ) Co-IP analysis of the SIRT2-OGG1 interaction in HEK293 cells treated with KBrO 3 .HEK293 cells transfected with vectors encoding SIRT2-Flag-GFP and OGG1-His were treated with 40 mM KBrO 3 for 30 min before being lysed for co-IP with Flag beads followed by Western blot analysis.(E, F) Co-IP analysis of the interaction between OGG1 and SIRT2 mutant fragments.The SIRT2 mutant fragments are shown in ( E ).Cells were transfected with vectors encoding these mutants tagged with Flag and OGG1 tagged with GFP and were lysed for co-IP with GFP-Trap followed by Western blot analysis ( F ). ( G, H ) Co-IP analysis of the interaction between SIRT2 and OGG1 mutant fragments.The OGG1 mutant fragments are shown in ( G ). Cells were transfected with vectors encoding these mutants tagged with HA and GFP, and SIRT2 tagged with Flag and GFP, f ollo w ed b y the ly sis of cells for co-IP with Flag beads and Western blot analysis ( H ).

Figure 4 .
Figure 4. SIRT2 facilitates the recruitment of OGG1 to its own promoter to enhance OGG1 transcription.( A ) In vitro analysis of the acetylation level of OGG1 in the presence of SIRT2.Purified recombinant OGG1 and SIRT2 from HEK293 cells were incubated with each other in the absence or presence of NAD + prior to Western blot analysis with the indicated antibodies.( B ) Analysis of BER efficiency in HCA2-hTERT cells overexpressing SIRT2 WT or the enzymatically dead H187Y mutant.**** P < 0.0001; n.s., not significant.( C ) Western blot analysis of the expression of SIRT2 WT or the enzymatically dead H187Y mutant in HCA2-hTERT cells.( D, E ) ChIP analysis of the recruitment of OGG1 to its promoter in HEK293 cells.Cells with e x ogenous introduction of the control plasmid or pO GG1-GFP e xpression v ector w ere left untreated ( D ) or treated with KBrO 3 ( E ), f ollo w ed b y ly sis and a ChIP assay with GFP-Trap.** P < 0.01.( F ) Analysis of OGG1 promoter activity in control and OGG1-overexpressing HCA2-hTERT cells.Cells were transfected with the luciferase reporter vector and a control vector or a vector encoding OGG1 at different concentrations.Sevent y-t wo hours post-transfection, cells w ere ly sed f or luciferase activity measurement.The Western blot image shows the analy sis of O GG1 e xpression in HEK293 cells transfected with increasing amounts of the OGG1 vector.* P < 0.05.( G ) Analysis of OGG1 promoter activity in control and OGG1-depleted HCA2-hTERT cells.* P < 0.05; ** P < 0.01.( H ) ChIP analysis of the recruitment of OGG1 to its promoter in HEK293 cells transfected with the control or SIRT2 plasmid and the OGG1-GFP plasmid.Cells were transfected with the indicated plasmids and were then lysed and subjected to a ChIP assay with GFP-Trap.** P < 0.01; *** P < 0.001; **** P < 0.0 0 01.( I ) ChIP analysis of the recruitment of OGG1 to its promoter in control and SIRT2-depleted cells transfected with OGG1-GFP plasmids.**** P < 0.0 0 01.( J ) Analysis of OGG1 promoter activity in control and OGG1-depleted HCA2-hTERT cells in the presence or absence of SIRT2 overexpression.* P < 0.05; n.s., not significant.( K ) Analysis of OGG1 promoter activity in HCA2-hTERT cells transfected with vectors encoding OGG1 and / or SIRT2.*** P < 0.001; **** P < 0.0001.

Figure 5 .
Figure 5. SIRT2 regulates OGG1 transcription in an A TM / A TR-dependent manner .( A ) Analysis of A TM or A TR phosphorylation in cells treated with KBrO 3 .HEK293 cells treated with 40 mM KBrO 3 were harvested for protein extraction followed by Western blot analysis with indicated antibodies.( B ) Analysis of SIRT2 phosphorylation in KBrO3-treated cells.HEK293 cells transfected with SIRT2-Flag plasmids were treated with or without KBrO 3 and ly sed f or co-IP with an anti-Flag antibody f ollo w ed b y Western blot analy sis with indicated antibodies.(C, D) Co-IP analy sis of the interaction betw een SIRT2 and ATM ( C ) or ATR ( D ). ( E ) The effect of ATM or ATR inhibitors on SIRT2 phosphorylation in cells treated with KBrO 3 .KBrO 3 -treated HEK293 cells were incubated with an ATM inhibitor (15 μM) or an ATR inhibitor (10 μM) for 4 hours and were then lysed for co-IP with an anti-Flag antibody followed by Western blot analysis with the indicated antibodies.( F ) Co-IP analysis of the interaction between SIRT2 and OGG1 in cells treated with ATM or ATR inhibitors.( G, H ) Analysis of OGG1 promoter activity in HCA2-hTERT cells with or without SIRT2 overexpression in the presence or absence of the ATM inhibitor ( G ) or ATR inhibitor ( H ). * P < 0.05; **** P < 0.0 0 01; n.s., not significant.( I ) T he analy sis of BER efficiency in HCA2-hTER T cells transfected with vectors encoding SIRT2 WT, S46A mutant, S53A mutant or S46AS53A mutant.* P < 0.05; ** P < 0.01; *** P < 0.001; n.s., not significant.( J ) Analysis of SIRT2 WT and S46AS53A mutant phosphorylation.HEK293 cells transfected with Flag-tagged WT or mutant SIRT2 were treated with or without KBrO 3 and then lysed for co-IP with an anti-p-(S / T)Q antibody f ollo w ed b y Western blot analy sis.( K ) Co-IP analy sis of the interaction betw een O GG1 and SIRT2 WT or SIRT2 S46AS53A.( L ) Analysis of OGG1 promoter activity in HCA2-hTERT cells transfected with vectors encoding SIRT2 WT or SIRT2 S46AS53A.** P < 0.01; n.s., not significant.( M ) ChIP analysis of the recruitment of SIRT2 to OGG1 promoter in cells transfected with the indicated plasmids or subjected to the indicated treatment.** P < 0.01; *** P < 0.001;**** P < 0.0 0 01.( N ) Analysis of the OGG1 mRNA level in cells transfected with vectors expressing SIRT2 WT or the S46AS53A mutant.**** P < 0.0001; n.s., not significant.

Figure 6 .
Figure 6.Se v eral cancer-associated SIR T2 mutants do not e x ert a stimulatory effect on O GG1 transcription and f ail to impro v e BER. ( A ) L ollipop plot of the SIRT2 protein with the alterations present in cancers.( B ) Western blot analysis of the expression of all SIRT2 mutants in HCA2-hTERT cells.HCA2-hTERT cells were transfected with vectors encoding SIRT2 WT or the indicated 37 mutants at 5 μg.On Day 2 post-transfection, cells were harv ested f or Western blot analy sis.( C ) Analy sis of BER efficiency in HCA2-hTER T cells o v ere xpressing SIR T2 WT or SIR T2 mutants.( D ) Analy sis of OGG1 promoter activity in HCA2-hTERT cells overexpressing SIRT2 WT or SIRT2 mutants.( E ) Analysis of OGG1 mRNA levels in HEK293 cells o v ere xpressing SIR T2 WT or SIR T mutants.