TRIM29 is required for efficient recruitment of 53BP1 in response to DNA double‐strand breaks in vertebrate cells

TRIM29 is involved in DNA double‐strand break repair but the specific roles of TRIM29 in DNA repair are not clearly understood. We proposed that TRIM29 regulates the choice of DNA DSB repair pathway by facilitating 53BP1 accumulation to promote NHEJ, possibly through RNF8 or TIP60.

Tripartite motif-containing protein 29 (TRIM29) is involved in DNA double-strand break (DSB) repair. However, the specific roles of TRIM29 in DNA repair are not clearly understood. To investigate the involvement of TRIM29 in DNA DSB repair, we disrupted TRIM29 in DT40 cells by gene targeting with homologous recombination (HR). The roles of TRIM29 were investigated by clonogenic survival assays and immunofluorescence analyses. TRIM29 triallelic knockout (TRIM29 À/À/À/+ ) cells were sensitive to etoposide, but resistant to camptothecin. Foci formation assays to assess DNA repair activities showed that the dissociation of etoposideinduced phosphorylated H2A histone family member X (ɣ-H2AX) foci was retained in TRIM29 À/À/À/+ cells, and the formation of etoposide-induced tumor suppressor p53-binding protein 1 (53BP1) foci in TRIM29 À/À/À/+ cells was slower compared with wild-type (WT) cells. Interestingly, the kinetics of camptothecin-induced RAD51 foci formation of TRIM29 À/À/À/+ cells was higher than that of WT cells. These results indicate that TRIM29 is required for efficient recruitment of 53BP1 to facilitate the nonhomologous end-joining (NHEJ) pathway and thereby suppress the HR pathway in response to DNA DSBs. TRIM29 regulates the choice of DNA DSB repair pathway by facilitating 53BP1 accumulation to promote NHEJ and may have potential for development into a therapeutic target to sensitize refractory cancers or as biomarker of personalized therapies.
Radiotherapy and the majority of chemotherapeutic drugs, including topoisomerase (TOP) 1 and 2 inhibitors, induce DNA double-strand breaks (DSBs) that are considered as one of the most biologically lethal damage to a cell. Unrepaired or mis-repaired DNA DSBs can cause cell death and genomic instability [1]. In higher eukaryotic cells, homologous recombination (HR) and nonhomologous end joining (NHEJ) are the two predominant pathways, together with alternative end joining (alt-EJ) and single-strand annealing (SSA), to repair DNA DSBs [2,3]. HR precisely repairs DNA DSBs by copying information from a DNA template, which is usually a sister chromatid. At the initial step of HR, the DSB end is resected to generate a 3 0 single-stranded DNA (ssDNA) overhang. The MRE11/RAD50/NBS1 (MRN) complex and CtBR-interacting protein (CtIP) are also required for this end resection process [4]. Replication protein A (RPA) then coats the 3 0 ssDNA overhang to destabilize the DNA secondary structure and to protect 3 0 ssDNA overhang from nucleolytic degradation. Subsequently, RAD51 replaces RPA facilitated by breast cancer type 1 susceptibility protein (BRCA1), partner, and localizer of BRCA2 (PALB2) and breast cancer type 2 susceptibility protein (BRCA2) [5,6]. In contrast to HR, the intact template and intensive resection step are not required for NHEJ. NHEJ is initially triggered by the binding of Ku70/80 heterodimers at DNA DSB sites. Subsequently, other repair proteins, such as DNA-dependent protein kinase catalytic subunit (DNA-PKcs), Artemis, X-ray repair cross-complementing 4 (XRCC4), XRCC4-like factor (XLF), newly identified paralog of XRCC4 and XLF (PAXX), and DNA ligase 4 (LIG4), are recruited to DNA DSB sites [7]. HR is relatively active during S and early G2 cell cycle phases, while NHEJ is active throughout all phases [8]. In addition to cell cycle phases, the choice between HR and NHEJ pathways is influenced by types of damaged DNA ends and DNA end resection [9]. For example, intensive DNA end resection required for HR can be suppressed by tumor suppressor p53-binding protein 1 (53BP1), which is thought to promote NHEJ. 53BP1 not only suppresses HR, but also foster fidelity of HR by limiting mutagenic singlestranded annealing (SSA), which is triggered by hyper-resection of DNA ends [10]. However, 53BP1 requires other protein partners to efficiently function in DNA repair because 53BP1 alone is not enough to block DNA end resection or protect DNA ends from nucleolytic activities [11] that draws attention to the identification of novel factors associated with 53BP1-dependent DNA repair. For instance, researchers have recently discovered a four-subunit protein complex called shieldin, which is composed of REV7, SHLD1, SHLD2, and SHLD3 [11]. Shieldin complex also promotes many cellular processes that are associated with 53BP1, such as protection of DNA ends, immunoglobulin class switching, and NHEJ [12][13][14]. It has been reported that shieldin complex also interacts with CTC1/STN1/TEN1 (CST) complex, which is able to antagonize end resection [15].
It is known that DNA repair partially plays a role in the cancer therapeutic response. Either elevated or reduced activities of DNA repair can determine outcomes of cancer treatments [16]. Upregulation of DNA repair can result in cancer resistance to therapies [9]. Although numerous proteins associated with DNA DSB repair are characterized well, the functions of some identified proteins are not clearly described in DNA DSB repair. Therefore, a better understanding of DNA DSB repair is crucial to identify potential targets to sensitize treatment-resistant cancers and improve outcomes.
Tripartite motif-containing protein 29 (TRIM29) is also known as ataxia-telangiectasia group D complementing protein (ATDC). Structurally, TRIM family proteins contain three conserved domains: a RING-finger domain, one or two B-box domains, and a coiledcoil domain [17]. Although TRIM29 is a member of TRIM family proteins, TRIM29 lacks RING-finger domain. It has been recently shown that TRIM29 has a remnant E3 ligase activity mediated by its B-box domains [18]. TRIM29 is highly expressed in many tumor types such as pancreatic, esophageal, bladder, lung, breast, head and neck, and colorectal cancers [19,20]. It has been reported that TRIM29 knockdown in SiHa, BxPC3, and Panc1 cell lines results in radiosensitivity [21,22]. In addition, TRIM29 can be phosphorylated by MAPKAP kinase 2 in an ATM-dependent fashion contributing to radioresistant phenotypes [22]. TRIM29 also interacts with DNA-PKcs, BRCA1-associated surveillance complex (BASC), Tat-interactive protein 60 (TIP60), cohesion, and histone proteins. TRIM29 might function as a scaffold protein for DNA repair protein accumulation at DNA DSB sites resulting in the efficiently active DNA damage response (DDR) [23]. Although previous studies of TRIM29 indicate that it is associated with DNA repair [23,24], little is known about the roles of TRIM29 in DDR and DNA repair induced by exogenous DNA-damaging compounds.
To analyze the functions of TRIM29, we disrupted TRIM29 in DT40 cells. Our data showed that TRIM29 triallelic knockout (TRIM29 À/À/À/+ ) cells were sensitive to etoposide, but resistant to camptothecin. The recruitment of 53BP1 to DNA DSBs was also decreased in TRIM29 À/À/À/+ cells; however, RAD51 localization was increased. We propose that TRIM29 is a choice regulator of DNA DSB repair pathways by promoting NHEJ and suppressing HR in response to DNA DSBs in vertebrate cells.
The cell cycle distribution in asynchronous populations of WT and TRIM29 À/À/À/+ cells was examined by analyzing DNA contents. Under normal conditions, the differences of cell cycle phase distribution between WT and TRIM29 À/À/À/+ cells were not statistically significant (Table 2 and Fig. S2). Moreover, mitotic indices observed in asynchronous populations of WT and TRIM29 À/À/À/+ cells were not significantly different under normal conditions (Table 3).

TRIM29 À/À/À/+ cells are defective for DNA DSB repair
To investigate the DNA DSB repair efficiency in TRIM29 À/À/À/+ cells, phosphorylated H2A histone family member X at Ser 139 (ɣ-H2AX) foci formation assays was conducted to measure the DNA DSB repair efficiency induced by etoposide in WT, TRIM29 À/À/À/+ , and Ku70 À/À cells. Cells containing more than four ɣ-H2AX foci were classified as positive. H2AXs near DNA DSBs are phosphorylated on Ser 139 after formation of DNA DSBs. Therefore, phosphorylation of H2AX on Ser 139 was used as a DNA DSB marker [27]. After pulse treatment with etoposide, ɣ-H2AX foci formed in a similar fashion in all cell lines (Fig. 3). However, as shown in Fig. 4A-C, the percentages of ɣ-H2AX-positive cells and median numbers of ɣ-H2AX foci per cell of WT, TRIM29 À/À/À/+ , and Ku70 À/À cells were decreased in a different time-dependent manner after pulse-treated with 1 µM etoposide for 2 h. The delayed in the dissolution of ɣ-H2AX foci in TRIM29 À/ À/À/+ and Ku70 À/À cells indicating that DNA DSB repair efficiency in response to etoposide treatment of TRIM29 À/À/À/+ cells was lower than in WT cells. In contrast to etoposide, the DNA DSB repair kinetics induced by camptothecin of WT, TRIM29 À/À/À/+ , and Ku70 À/À cells were not statistically different (Fig. 5). Taken together, the results suggest that NHEJ-mediated DSB repair induced by etoposide is defective in TRIM29 À/À/À/+ cells.
53BP1 localization at DSBs is defective in TRIM29 À/À/À/+ cells In response to DNA DSBs, DDR is activated to detect, signal, and recruit DNA repair proteins. Ataxia-telangiectasia mutated (ATM), ATM's substrates, mediator of DNA damage checkpoint protein 1 (MDC1), and 53BP1 are key factors in DDR [28]. 53BP1 is one of the important pathway choice regulators of DNA DSB repair pathways by promoting NHEJ and suppressing HR [29,30]. To investigate the DDR to DNA DSBs, 53BP1 and ɣ-H2AX foci induced by 1 µM etoposide were monitored at the indicated time points. Cells containing more than four 53BP1 foci were classified as positive. The results showed that the percentages of 53BP1-positive TRIM29 À/À/À/+ cells were lower than those of 53BP1-positive WT and Ku70 À/À cells after exposure to etoposide, and the median numbers of 53BP1 foci formation after DNA damage were lower than WT and Ku70 À/À cells at 1 and 2 h (Fig. 6), although the percentages of ɣ-H2AX-positive TRIM29 À/À/À/+ cells were not statistically different from those of ɣ-H2AX-positive WT and Ku70 À/À cells (Fig. 3). These results suggest that the Table 3. Mitotic index of WT and TRIM29 À/À/À/+ cells under normal conditions (mean AE SD).
As shown in Fig. 7A-C, the percentages of RAD51positive TRIM29 À/À/À/+ were greater than those of RAD51-positive WT cells. The same effect was also observed in Ku70 À/À cells at 1 and 2 h after exposure to camptothecin. We found that the difference in RAD51 recruitment between WT and TRIM29 À/À/À/+ cells was not correlated with DNA damage as the number of ɣ-H2AX-positive TRIM29 À/À/À/+ cells were not significantly different from WT and Ku70 À/À after camptothecin treatment (Fig. 8). These results showed that the kinetics of RAD51 foci formation in TRIM29 À/ À/À/+ and Ku70 À/À cells induced by camptothecin were faster than WT cells. Taken together, our findings indicate that the activity of the DNA DSB-induced HR pathway is increased in TRIM29 À/À/À/+ cells.

Discussion
We investigated which DNA repair pathway involved TRIM29 in the repair of DNA damage induced by exogenous, genotoxic agents using clonogenic survival assays. In this study, we found that TRIM29 was       responsible for the repair of etoposide-induced DNA DSBs. TRIM29 À/À/À/+ cells displayed the sensitivity to etoposide, but not cisplatin, olaparib, or UV-C. In general, TOP2 is associated with relaxation, catenation/decatenation, and winding/unwinding of the DNA double helix to resolve topological entanglement by forming a reversible, cleavable complex with a DNA molecule to generate transient DNA DSBs and then religating DNA ends at the end of this reaction. Etoposide stabilizes the cleavable complex leading to DNA DSBs [26]. Etoposide-mediated DSBs are predominantly repaired by the NHEJ pathway [31]. Interestingly, TRIM29 À/À/À/+ cells were resistant to camptothecin that induces an irreversible, cleavable complex of TOP1-DNA. Generally, the function of TOP1 is to unwind supercoiled DNA molecules associated with DNA replication by inducing DNA singlestrand breaks (SSBs) [32]. When the irreversible TOP1-DNA complex encounters a replication fork, the unrepaired SSB is converted into a DSB. Thus, camptothecin cytotoxicity is specific to S phase [33,34]. TOP1-mediated DNA DSBs generated by camptothecin are preferentially repaired by the HR pathway [35]. Likewise, NHEJ-defective DT40 cells, such as Ku70 À/À , DNA-PKcs knockout (DNA-PKcs À/À/À ), and LIG4 knockout (LIG4 À/À ) cells, are hypersensitive to etoposide, and NHEJ deficiency confers camptothecin resistance in DT40 cells [31,36]. Our results indicated that TRIM29 might be involved in NHEJ pathway. It has been reported that the retention of ɣ-H2AX foci indicates defects in DNA DSB repair and is correlated with loss of the clonogenic potential [37,38]. The retention of ɣ-H2AX foci in TRIM29 À/À/À/+ cells was observed when TRIM29 À/À/À/+ cells were treated with etoposide, not camptothecin. It suggested that TRIM29-deficient cells were defective for etoposide-induced DNA DSB repair. Using ɣ-H2AX, 53BP1, and RAD51 foci formation assays, we demonstrated that TRIM29 was involved in 53BP1 localization at DNA DSB sites induced by etoposide. 53BP1 protein functions upstream of NHEJ. It has the ability to promote the NHEJ pathway by inhibiting end resection, a key step in the HR pathway [29,39]. Furthermore, increased recruitment of RAD51, a key factor of the HR pathway, after exposure of TRIM29 À/À/À/+ cells to camptothecin was similar to the previous observation of RAD51 foci formation induced in XRCC4 mutant fibroblasts that were defective in the NHEJ pathway. Bee et al. reported that recruitment of RAD51 at DNA DSB sites induced by IR increased in XRCC4 mutant fibroblasts compared with WT fibroblasts. This compensation of HR in NHEJ-defective cells is normally observed in higher eukaryotes [8]. Moreover, the analysis of HR-mediated DNA DSB repair using I-SceI-based assays reveals that the frequency of HR repair is increased in Ku70 À/À , XRCC4 knockout (XRCC4 À/À ), and DNA-PKcs À/À embryonic stem cell lines [40]. This phenomenon probably proceeds via error-free RAD51-mediated gene conversion (GC) rather than mutagenic RAD52-mediated SSA as recent literature suggested SSA might function as an alternative DSB repair pathway in HR-deficient cells in higher eukaryotic models [3,41].
Mechanistically, it has been speculated that 53BP1 and other NHEJ factors that suppress HR by inhibiting DNA end resection at DSB sites need to be removed by BRCA1 [42,43]. Therefore, HR factors bind to DNA DSBs with less competition in NHEJ-defective cells. Collectively, our data suggest that TRIM29 is essential for the recruitment of 53BP1 to promote the NHEJ pathway, thereby suppressing the HR pathway. However, the mechanism underlying how TRIM29 facilitates 53BP1 recruitment at DNA DSB sites is unclear. It has been reported that a physical interaction between TRIM29 and RNF8 promotes DNA DSB repair [24], and overexpression of TRIM29 inhibits TIP60 functions by stimulating degradation and changing the localization of TIP60 [44]. In addition, TIP60 suppresses the binding of 53BP1 to histone H4 dimethylated at Lys 20 (H4K20 Me2 ) by acetylating of histone H4 on Lys 16 (H4K16 Ac ) [45,46]. Therefore, TRIM29 may facilitate the recruitment of 53BP1 to DNA DSBs by two possible mechanisms. One mechanism involves TRIM29 assisting RNF8 in the RNF8-RNF168 pathway to remove JMJD2A and L3MBTL1 that compete with 53BP1 for H4K20 Me2 [47,48], allowing 53BP1 to bind to H4K20 Me2 . The other mechanism involves TRIM29 diminishing TIP60-mediated H4K16 acetylation, leading to an increase in 53BP1 binding to H4K20 Me2 .

Conclusion
In summary, whether TRIM29 functions through RNF8 or TIP60 in order to facilitate choice of DNA DSB repair pathways remain unknown. Further functional studies on TRIM29 using additional models will help elucidate molecular functions of TRIM29 in DNA DSB repair. According to this study, TRIM29 acts as one of DNA DSB repair pathway choice regulators, promoting 53BP1 recruitment to DNA DSB sites. The better understanding of functions of TRIM29 may facilitate establishment of new cancer treatments. For instance, targeting TRIM29 may sensitize refractory cancers to therapies. Although we did not clearly clarify the association of TRIM29 with DNA DSB repair, this study defines a novel role of TRIM29 in facilitating NHEJ and sheds light on TRIM29 in the field of DNA repair.

Cell proliferation assay
For the proliferation assay, cells were seeded in 24-well plates and cultured for 72 h. The cells were maintained in the exponential phase of growth. The number of cells was counted using an improved Neubauer hemocytometer with trypan blue (Gibco) every 24 h. Three independent experiments were conducted.

Determination of the cell cycle distribution by flow cytometry and mitotic index
To analyze the cell cycle phase distribution, cells were washed with PBS and fixed with 70% ice-cold ethanol at À20°C for at least 2 h. The cells were then stained with propidium iodide from a Muse Ò Cell cycle Assay Kit (Merck Millipore, Burlington, MA, USA), according to the manufacturer's protocol. The cell cycle distribution was analyzed by flow cytometry using a Navios flow cytometer and KALUZA analysis 2.1 software (Beckman Coulter, Inc., Brea, CA, USA). The mitotic index was calculated by a number of cells undergoing mitosis divided by a total number of cells, which were manually counted from fluorescence images at least 1000 cells.

Statistical analysis
Comparisons between groups were made by Student's t-test using GRAPHPAD PRISM 7 (Software, Inc., La Jolla, CA, USA). Results are displayed as the mean AE standard deviation (SD) or median. Statistical significance was accepted at P < 0.05.