RNF8 promotes efficient DSB repair by inhibiting the pro‐apoptotic activity of p53 through regulating the function of Tip60

Abstract Objectives RING finger protein 8 (RNF8) is an E3 ligase that plays an essential role in DSB repair. p53 is a well‐established tumour suppressor and cellular gatekeeper of genome stability. This study aimed at investigating the functional correlations between RNF8 and p53 in DSB damage repair. Materials and methods In this article, wild‐type, knockout and shRNA‐depleted HCT116 and U2OS cells were stressed, and the roles of RNF8 and p53 were examined. RT‐PCR and Western blot were utilized to investigate the expression of related genes in damaged cells. Cell proliferation, apoptosis and neutral cell comet assays were applied to determine the effects of DSB damage on differently treated cells. DR‐GFP, EJ5‐GFP and LacI‐LacO targeting systems, flow cytometry, mass spectrometry, IP, IF, GST pull‐down assay were used to explore the molecular mechanism of RNF8 and p53 in DSB damage repair. Results We found that RNF8 knockdown increased cellular sensitivity to DSB damage and decreased cell proliferation, which was correlated with high expression of the p53 gene. RNF8 improved the efficiency of DSB repair by inhibiting the pro‐apoptotic function of p53. We also found that RNF8 restrains cell apoptosis by inhibiting over‐activation of ATM and subsequently reducing p53 acetylation at K120 through regulating Tip60. Conclusions Taken together, these findings suggested that RNF8 promotes efficient DSB repair by inhibiting the pro‐apoptotic activity of p53 through regulating the function of Tip60.

Upon DSB formation, the following processes occur: the histone H2AX is phosphorylated by ataxia telangiectasia mutated (ATM), and mediator of DNA damage checkpoint 1 (MDC1) binds to phosphorylated H2AX (γ-H2AX). E3 ubiquitin ligase RING finger protein 8 (RNF8) binds to phosphorylated MDC1 at DSB sites and promotes the recruitment of RNF168, another E3 ubiquitin ligase. RNF8 and RNF168 conjugate Lys63-linked ubiquitin chains onto histone H2A with their cognate E2 conjugating enzyme UBC13 to induce chromatin remodelling. 7-10 RNF8/RNF168-UBC13-dependent ubiquitination of histones at DSBs promotes the assembly of a series of repair proteins, including BRCA1 and 53BP1. Thus far, RNF8 has been shown to promote DSB damage repair mainly by promoting the binding of key proteins involved in DSB damage repair to DSB sites through RNF8 ubiquitination activity. 9,[11][12][13][14] In addition to the classic role of RNF8 described above, other effects of RNF8 have been reported. Studies have confirmed that RNF8 depletion partially suppresses HR defects caused by BRCA1 depletion, and the degree of HR suppression due to RNF8 depletion is significantly less than that due to RNF168 depletion. 15 Moreover, scientists revealed a newly discovered non-catalytic role of RNF8 in unfolding higher-order chromatin structures. 16 Furthermore, RNF8 has been shown to promote cellular resistance to replication stress and is important for meiosis. [17][18][19][20][21] Thus, RNF8 may maintain genomic stability through multiple pathways.
Different levels of damage may trigger different responses; low levels of DNA damage trigger repair and cell survival, and high levels of DNA damage trigger cell death. 22 p53 is a well-established tumour suppressor and cellular gatekeeper of genome stability. 23 p53 stabilization may contribute to the different expression levels of pro-survival and pro-apoptosis genes. 24,25 p53 has been shown to determine cell fate through promoting cell survival or apoptosis in many works, but the specific mechanism of this role remains to be revealed. RNF8 and p53 are both key repair factors in the DSB damage response.
In this report, we found that the expression of p53 was increased in RNF8-deficient cells compared to that in control cells and that increased p53 was highly correlated with the increased sensitivity of RNF8-deficient cells to DSB damage. Although RNF8 and p53 are present in the same complex, RNF8, an E3 ubiquitin ligase, neither degrades nor ubiquitinates p53. However, RNF8 can effectively improve the efficiency of DSB damage repair through weakening p53 acetylation at the K120 site by regulating Tip60 activity. The present results indicate that RNF8 promotes efficient DSB repair by inhibiting the pro-apoptotic function of the p53 gene and reveals a novel way by which RNF8 regulates DSB damage repair efficiency.

| Cell culture and treatment
The A03_1 cell line carrying an array of LacO operators were kindly given by Dr Guohong Li (Institute of Biophysics, Chinese Academy of Sciences, China). 26 Human HCT116 WT and HCT116 p53−/− cells were kindly given by Dr Yong Cai (Jilin University, China). 27  supplemented with 10% foetal bovine serum (FBS) (Gibco) and 1% penicillin/streptomycin (Invitrogen). All cells above were cultivated at 37°C and 5% CO 2 in a humidified incubator. For DSB induction, cells were incubated with the indicated concentrations of Eto or CPT for 20 min and then released for repair. AsiSI-ER-U2OS cells were treated with 300 nM 4-OHT for 3 h to induce DSBs.  Table 1 and Table 2.

| Neutral cell comet assay
The neutral comet assay was performed using the Comet Assay Kit from Trevigen (Gaithersburg, MD) following the manufacturer's guidance. Images were captured using the fluorescent microscope (ECLIPSE, 80i, Nikon, Japan). Tail moment was tested using CometScore software (TriTek, Sumerduck, USA).

| Immunoprecipitation (IP) and Western blotting
The cells were lysed with RIPA lysis buffer, and the whole cell lysates were incubated with appropriate antibodies at 4°C for 3 hours.
Samples were then incubated for another 3 hours with protein A/G Sepharose beads. After washing the samples with RIPA lysis buffer, the immunoprecipitates were resolved on SDS-PAGE and probed with antibodies as indicated.

| Mass spectrometry
After staining protein in SDS-PAGE gels with coomassie blue, gel lanes were sliced into different bands. The specific analysis process was completed by Shanghai Applied protein Technology (APT), China.

| Quantitative real-time polymerase chain reaction (qRT-PCR)
Total RNA was isolated from cultured cells using the TRIzol reagent, and the resulting messenger RNA (mRNA) was reversely transcribed into complementary DNA (cDNA) using PrimeScript ™ RT reagent Kit according to the instructions. QRT-PCR analysis was performed using SYBR ® Premix Ex Taq ™ following the introductions. GAPDH, 18SrRNA and β-actin were all served as the internal reference. The method of 2 −ΔΔCt was used to calculate the relative mRNA expression of the target genes. According to the mRNA sequences of target genes and reference genes in GenBank, the primers were all designed by Primer Premier 5 (Table 3).

| Cell proliferation assay
Different types of cells were first treated with the indicated con-

| Statistical analysis
The data were analysed using the Student t test and presented as the mean ± SEM. The quantifications were based on the results of at least three independent experiments. The levels of significance are designated *P < .05, **P < .01 and ***P < .001.

| RNF8 knockdown increases the sensitivity of cells to DSBs
In response to DSB damage, RNF8 provides a structural chromatin platform for the subsequent recruitment of key DNA repair factors, such as 53BP1 and BRCA1. 9

TA B L E 3 Primers used for qRT-PCR analyses
from RPA, a single-stranded binding protein ( Figure 1A). 12,32,33 RNF8 deficiency impeded the recruitment of BRCA1 and 53BP1 at DSB sites ( Figure 1B, C and Figure S1A). Restrictive endonuclease can also induce DNA splicing and cause endogenous DSB damage. U2OS cells stably expressing the HA-tagged AsiSI-ER restriction enzyme were employed to generate fragments of the human genome longer than 1 Mb. 34 RNF8 was also recruited at restriction enzyme-induced DSB sites and co-localized with γ-H2AX ( Figure S1B-D). Then, the drug etoposide (Eto), a topoisomerase II inhibitor, was used to induce DSB damage. Eto interferes with topoisomerase II activity and causes DSBs through the formation of a DNA-drug-enzyme cleavage complex. 35 RNF8 was recruited to Eto-induced DSBs. Trends in the average number and size of RNF8 foci in each damaged cell were similar to those of γ-H2AX, the number of which peaked after 1 hour of repair ( Figure 1D).
The above results indicate that RNF8 is a chromatin-associated protein widely involved in the repair of damage from DSBs from different sources, including laser irradiation, restriction enzymes and chemical drugs, and which is consistent with the classical way by which RNF8 regulates DSB damage repair efficiency.
To explore the function of RNF8 in DNA damage, the effect of RNF8 expression on DSB repair efficiency following Eto treatment was examined. RNF8-proficient and RNF8-deficient U2OS cells were established, and changes in the formation of γ-H2AX foci following Eto treatment were examined. Compared with Eto-treated control cells, U2OS-shRNF8 cells contained more γ-H2AX foci, and more time was needed for the signal from these foci to decrease to the background level (Figure 2A), suggesting reduced repair efficiency. In addition, RNF8 knockdown significantly increased cell apoptosis compared with that in U2OS-shCTR cells ( Figure 2B). The

| The increased sensitivity of RNF8-deficient cells to DSB damage is correlated with high expression levels of the p53 gene
Since cell apoptosis is correlated with the expression and activation of p53, 36-38 the protein and mRNA levels of p53 were first examined. p53 at both the protein and mRNA levels was increased in shRNA-transfected U2OS cells ( Figure 3A, B) and also in MCF7 cells treated the same way ( Figure S2A cells was slower than that of HCT116 WT -shCTR cells. However, the proliferation of HCT116 p53−/− -shRNF8 cells was faster than that of HCT116 WT -shRNF8 cells ( Figure 3D, Figure S3A, B). Then, a neutral comet assay was conducted, and the degree of DNA breaks in HCT116 p53−/− -shRNF8 cells was significantly lower than that in HCT116 WT -shRNF8 cells ( Figure 3E). Furthermore, the proportion of apoptotic HCT116 p53−/− -shRNF8 cells was much lower than that of apoptotic HCT116 WT -shRNF8 cells ( Figure 3F). The above results indicate that the increased sensitivity of RNF8-deficient cells to DSB damage is closely related to the high-level expression of the p53 gene.

| RNF8 can improve DSB repair efficiency by inhibiting the pro-apoptotic function of p53
Based on the above results, we speculated that RNF8 improves the efficiency of DSB damage repair, promotes cell survival and maintains genomic stability by inhibiting the functional activity of p53.
To verify the effect of RNF8 expression on the two DSB damage repair pathways, HR and NHEJ, two well-characterized GFP-based reporter systems, DR-GFP and EJ5-GFP, were utilized under conditions of p53 activation. The system was described previously. 29 A schematic diagram is presented in Figure 4A, B, and the cells were treated as shown in Figure S4. and RNF8 reduced the level of BAX ( Figure 4G, H and Figure S6A, B). Therefore, our collective data suggest that RNF8 promotes DSB damage repair by inhibiting the pro-apoptotic function of p53.

| RNF8 indirectly regulates the pro-apoptotic function of p53
The results above suggest that RNF8 can inhibit the pro-apoptotic activity of p53 during DSB damage repair. To explore the exact mechanism of this inhibitory effect, we first tested whether p53 and RNF8 are present in the same complex during DSB damage repair. GFP-tagged RNF8 was expressed in HEK-293T cells, and RNF8 complexes were then isolated and subjected to mass spectrometry.
A number of known RNF8-associated proteins (UBC, 40 H3 41 and NONO 42 ) were co-purified with RNF8, and p53 was also present in the RNF8 complex ( Figure 5A). In addition, GST pull-down and immunoprecipitation assays confirmed the presence of RNF8 and p53 in the same complex ( Figure 5B, C). Furthermore, by using a lac operator-repressor (LacO-LacI) targeting system, we found that p53 indeed co-localized with RNF8 ( Figure 5D), suggesting a functional correlation between p53 and RNF8. RNF8 is an E3 ubiquitin ligase that ubiquitinates many substrates and promotes the degradation of substrates such as H3 and NONO. 42,43 However, in our experimental system, RNF8 could neither degrade p53 nor promote its ubiquitination ( Figure 5E, F and Figure S7). Therefore, although RNF8 and p53

| RNF8 inhibits the pro-apoptotic function of p53 by regulating the activity of Tip60
Based on the results above, we preliminarily speculated that the inhibition of the pro-apoptotic function of p53 by RNF8 is indirectly mediated by other proteins. The pro-apoptotic activity of p53 depends on its acetylation at K120 by the acetyltransferase Tip60. 44 So we first checked whether RNF8 can regulate the function of Tip60.

Protein immunoprecipitation experiments showed that RNF8 and
Tip60 are present in the same complex under normal or DNA damage conditions ( Figure 6B). In order to detect whether there is direct interaction between RNF8 and Tip60, we conducted the proximity ligation assay (PLA). The results showed that both endogenously expressed ( Figure S8A) and exogenously expressed ( Figure 6C i) RNF8 and Tip60 were directly interacted. In addition, a direct interaction F I G U R E 5 RNF8 indirectly regulates the pro-apoptotic function of p53. A, Co-immunoprecipitation (Co-IP) purification experiment was performed using HEK-293T cells expressing GFP-tagged RNF8 as indicated in the method. The part of hits from mass spectrometry analysis was shown. B, Co-IP assays were performed to check the interaction between RNF8 and p53 in Eto or CPT treated HEK-293T cells. IP samples were resolved by SDS-PAGE and immunoblotted with the indicated antibodies. C, GST pull-down assay of p53 was conducted using the indicated proteins expressed in bacteria. GFP-tagged RNF8 was expressed in HEK-293T cells, and the lysates were incubated with purified GST or GST-fused p53. Bound proteins were separated by SDS-PAGE and immunoblotted with anti-GFP antibody. D, (i) A schematic diagram showing the lac operator-repressor (LacO-LacI) targeting system for assaying co-localization and interaction of proteins. (ii) Colocalization of p53 with GFP and GFP-RNF8. E, U2OS cells were transfected with indicated plasmids for 48 h, then treated with or without Eto (10 μmol/L). The immunoblot showed the expression of the indicated proteins. GFP was used as the loading control. F, The cells were treated as E and exposed to 10 mmol/L MG132 (a potent cell-permeable 20S proteasome inhibitor) for 6 h before protein lysate extraction. GFP was used as the loading control between Tip60 and p53 was also detected ( Figure 6C ii and Figure   S8B). We also found that the interaction between Tip60 and p53 was attenuated both in immunoprecipitation assay ( Figure 6D) and PLA ( Figure 6E) when RNF8 was overexpressed. Furthermore, RNF8 significantly inhibited the Tip60-mediated acetylation of p53 at K120, and when RNF8 and Tip60 were simultaneously overexpressed, p53 acetylation at K120 ( Figure 7A and Figure S9A) and the expression of p53-induced downstream apoptosis-related genes, such as BAX, p21 and PUMA ( Figure 7B and Figure S9B), were significantly decreased, compared to that in cells overexpressing Tip60 alone. Consistent with the above results, the apoptosis rate in cells overexpressing RNF8 and Tip60 together was significantly lower than that in cells overexpressing Tip60 alone ( Figure 7C).
Therefore, RNF8 can directly attenuate the interaction between Tip60 and p53.
To further investigate whether the regulation of RNF8 on the pro-apoptotic function of p53 is dependent on the presence of Tip60, we stably interfered Tip60 ( Figure S10A) and found that RNF8 does not affect the expression of apoptosis-related target genes downstream of p53 ( Figure 7D) and the death of cells ( Figure S10B) with low expression of Tip60. In order to further explore the mechanism of RNF8 in regulating Tip60, we identified whether RNF8 affects the acetyl- Additionally, there is often interplay among these responses that ultimately determines the fate of the stressed cells. 47,48 RNF8 is an E3 ubiquitin ligase localized primarily in the nucleus during interphase.
In contrast, under genotoxic stress, RNF8 is localized to sites of DNA damage, where it ubiquitinates many chromatin substrates, including histones H2A, H2AX and H1, the ubiquitination of which is of great importance to chromatin structure remodelling. 12,49 In addition to its classic features above, after DSB damage, RNF8 can also promote efficient DSB damage repair by decreasing the pro-apoptosis activity of p53 through regulating Tip60 protein activity.
We found that RNF8 is widely involved in the repair of DSB damage induced by multiple sources, such as drugs, restrictive endonucleases and laser irradiation, and foci containing labelled RNF8 overlapped with those containing γ-H2AX (Figure 1). The loss of teraction between RNF8 and p53 in the protection against genomic instability and tumorigenesis. 50 In the present study, we further showed that the proliferation of HCT116 p53−/− -shRNF8 cells was faster than that of HCT116 WT -shRNF8 cells. In addition, the level of apoptotic HCT116 p53−/− -shRNF8 cells was significantly lower than that of HCT116 WT -shRNF8 cells (Figure 3). Through the DR-GFP and EJ5-GFP repair systems, we also found that RNF8 may promote DSB damage repair by inhibiting the pro-apoptotic function of p53 ( Figure 4).
To further explore the mechanism by which RNF8 inhibits the pro-apoptotic function of p53, RNF8-associated complexes were isolated and subjected to mass spectrometry. RNF8 and p53 were present in the same complex; however, RNF8 neither degraded nor ubiquitinated p53 ( Figure 5), suggesting that RNF8 indirectly regulates p53 function through mediating other proteins. p53 protein activity in regulating apoptosis is highly correlated with its own post-translational modification, and the Tip60-mediated acetylation of p53 at the K120 site is crucial for the activation of PUMA and BAX, which are essential for apoptosis. 51,52 Indeed, RNF8, p53 and Tip60 were present in the same complex, and the high level of Tip60 expression significantly increased the level of p53 acetylated at the K120 site. However, when RNF8 was overexpressed together with Tip60, the acetylation of p53 at the K120 site was significantly reduced and the expression of apoptosis-related target genes downstream of p53 (eg, PUMA and BAX) exhibited the same trend ( Figure 6 and 7). The results also showed that, compared with the cells expressing Tip60 alone, the protein level of γ-H2AX in the cells co-expressing RNF8 and Tip60 was significantly lower. Because the formation of γ-H2AX is mediated by activated ATM kinase, we suspect that RNF8 may inhibit the ATM activation mediated by Tip60. 45 To test our hypothesis, we examined the phosphorylation of ATM-ser1981 at different expression levels of RNF8 and Tip60. We showed that overexpression of Tip60 can significantly enhance the auto-phosphorylation of ATM-ser1981; however, the phosphorylation of ATM-ser1981 is significantly down-regulated in the cells co-expressing RNF8 and Tip60 and the cells with low expression of Tip60 (Figure 7).
We speculate that RNF8 can inhibit Tip60-dependent ATM activation.
The mechanism of functional regulation of RNF8 for Tip60 is very complex. We suggest that RNF8 may inhibit the p53 pro-apoptotic function mediated by Tip60 through two pathways. One is the continued expression of γ-H2AX will activate p53 and promote the apoptosis by inducing the function of p53. RNF8 can inhibit the generation of excessive γ-H2AX mediated by Tip60. Another is the MDC1 scaffold protein can bind directly to γ-H2AX, which provide a platform to recruit RNF8 at DSB site. RNF8 can further stabilize the binding of Tip60 to chromatin, which are crucial for the apoptotic response induced by Tip60. 53 On the contrary, the reduction of γ-H2AX inhibits the recruitment of RNF8, thus inhibiting the apoptosis-inducing function of Tip60.
In addition to p53 acetylation, p53 phosphorylation also plays a crucial role in the activation and promotion of apoptosis. The phosphorylation of p53 at Ser46 following severe DNA damage increases the affinity of p53 to the promoters of pro-apoptotic genes such as p53AIP1. 54,55 Naoe Taira et al demonstrated that dual-specificity tyrosine-phosphorylation-regulated kinase 2 (DYRK2) can directly phosphorylate p53 at Ser46 upon exposure to genotoxic stress. 56 In addition, Takenori Yamamoto and colleagues found a direct interaction between DYRK2 and RNF8 in regulating the DNA damage response. 57 Here, we showed that p53, RNF8 and DYRK2 are co-localized in the nucleus ( Figure   S11A); however, we did not detect the direct interaction between RNF8 and DYRK2 reported by Takenori Yamamoto ( Figure S11B). This is perhaps due to the different cell types and injury types examined in these studies. In addition, there was no significant difference in the expression of apoptosis-related genes induced by phosphorylation of p53 at Ser46 under the regulation of RNF8 and DYRK2, such as p53AIP1 and BAX ( Figure S11C). Based on the above results, we believe that DYRK2 has a small effect on the regulation of p53 pro-apoptotic activity by RNF8.
In summary, our data indicate that RNF8 expression promotes efficient DSB damage repair by not only promoting the efficient recruitment of repair proteins at DSB sites but also inhibiting the pro-apoptotic function of p53 through attenuating Tip60 activity.
RNF8 expression hinders the recruitment of key repair factors at the F I G U R E 8 Description of a pattern diagram. Schematic diagram of the possibility of RNF8 in DSBs damage repair. Upon DSBs damage, the expression of RNF8 promotes efficient DSB damage repair by promoting efficient recruitment of repair proteins at DSB sites. In addition to its classic features, RNF8 can also promote DSB damage repair by inhibiting the pro-apoptotic function of p53 through regulating the acetyltransferase activity of Tip60. RNF8 can reduce the Tip60-dependent phosphorylation of ATM-ser1981 and inhibit the over-activation of ATM kinase. When the expression of RNF8 is low, the acetylation activity of Tip60 cannot be regulated, while ATM kinase will be over-activated. The synergistic interaction of Tip60 and activated ATM can effectively promote acetylation modification at p53-K120 site, improve the expression of p53 targeted apoptotic genes and lead to the activation of apoptosis pathway DSB site. Furthermore, the loss of Tip60 activity regulation effectively promotes the over-activation of ATM and the acetylation of p53 at K120, which improves the expression of p53-targeted apoptotic genes and increases apoptosis (Figure 8). Our work not only reveals a new way by which RNF8 regulates the efficiency of DSB damage repair but also provides new details on the effects of p53 on cell fate.

ACK N OWLED G EM ENTS
The authors would like to give thanks to Dr Guohong Li

CO N FLI C T S O F I NTE R E S T
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The authors declare that the data supporting the findings of this study are available from the corresponding author upon reasonable request.