ATDC (Ataxia Telangiectasia Group D Complementing) Promotes Radioresistance through an Interaction with the RNF8 Ubiquitin Ligase*

Background: ATDC/TRIM29 promotes resistance to ionizing radiation, but the factor(s) that mediate this effect are incompletely understood. Results: ATDC/TRIM29 binds to RNF8, promoting DNA repair and resistance to IR. Conclusion: Following DNA damage, ATDC/TRIM29 is phosphorylated and interacts with RNF8, promoting DNA repair and cell survival. Significance: The interaction between ATDC/TRIM29 and RNF8 is novel and is important for the DNA damage response. Induction of DNA damage by ionizing radiation (IR) and/or cytotoxic chemotherapy is an essential component of cancer therapy. The ataxia telangiectasia group D complementing gene (ATDC, also called TRIM29) is highly expressed in many malignancies. It participates in the DNA damage response downstream of ataxia telangiectasia-mutated (ATM) and p38/MK2 and promotes cell survival after IR. To elucidate the downstream mechanisms of ATDC-induced IR protection, we performed a mass spectrometry screen to identify ATDC binding partners. We identified a direct physical interaction between ATDC and the E3 ubiquitin ligase and DNA damage response protein, RNF8, which is required for ATDC-induced radioresistance. This interaction was refined to the C-terminal portion (amino acids 348–588) of ATDC and the RING domain of RNF8 and was disrupted by mutation of ATDC Ser-550 to alanine. Mutations disrupting this interaction abrogated ATDC-induced radioresistance. The interaction between RNF8 and ATDC, which was increased by IR, also promoted downstream DNA damage responses such as IR-induced γ-H2AX ubiquitination, 53BP1 phosphorylation, and subsequent resolution of the DNA damage foci. These studies define a novel function for ATDC in the RNF8-mediated DNA damage response and implicate RNF8 binding as a key determinant of the radioprotective function of ATDC.

Anticancer treatments such as ionizing radiation (IR) 3 induce DNA damage (including both single-strand and doublestrand breaks (DSB)), which activates a complex DNA damage response (DDR) program. This DDR coordinates cell cycle checkpoint activation, transcriptional regulation, and DNA repair, resulting in either restoration of the genome and cell survival or induction of programmed cell death in a process known as apoptosis (1). A key component of the DDR is the phosphorylation of H2AX primarily by ATM, ATR (ataxia telangiectasia and Rad3-related), and DNA-PK at sites of DNA DSB, resulting in subsequent recruitment of DNA repair and chromatin-modifying enzymes. This recruitment process requires a series of chromatin and repair enzyme covalent modifications including ubiquitination and sumoylation (2).
RNF8 is a crucial component of the DDR. RNF8, an E3 ubiquitin ligase, catalyzes the addition of poly-ubiquitin chains to H2AX and the DNA DSB repair complex, Mre11-Rad50-NBS1 (1). RNF8 is recruited to DNA damage sites by the binding of its forkhead-associated (FHA) domain to ATM-phosphorylated MDC1 (3,4). Once recruited, RNF8 catalyzes poly-ubiquitination of downstream repair proteins such as RAP80, 53BP1, and BRCA1, which allows their recruitment to the DNA DSB to facilitate DNA repair (5).
Ataxia telangiectasia group D complementing (ATDC), also known as TRIM29, was identified for its capacity to induce resistance to IR in cells derived from patients with ataxia telangiectasia, a disorder characterized by ATM deficiency (6). ATDC (TRIM29) is a member of the tripartite motif (TRIM) protein family, which is defined by a conserved RING domain, one or two B-box domains, and a coiled-coil region (7). ATDC does not have a RING domain, but does have typical B-box and coiled-coil domains. ATDC is known to physically interact with the intermediate filament protein, vimentin, as well as the proteins hPKCI-1, p53, HDAC9, and Tip60 (8 -12).
Interestingly, ATDC is highly expressed in multiple tumor types and is typically a marker of invasive/aggressive tumors (13)(14)(15)(16). In pancreatic cancer cells, ATDC has been shown to interact with DVL-2 through its coiled-coil domain, leading to the promotion of cell proliferation and invasion (17). ATDC also binds to p53, influencing cell cycle progression (10,17). We have recently demonstrated that ATDC is a downstream phosphorylation target of ATM and MAPKAPK2 (MAPKactivated protein kinase 2) following exposure to IR. Phosphorylation of ATDC is required to mediate resistance to IR (17). ATDC has also been shown to bind to chromatin and activate the DDR following DSB formation (18). The mechanism(s) downstream of ATDC that mediate radioresistance, however, are not known.
In this study, we performed a mass spectrometry screen to identify ATDC binding partners and elucidated the downstream mechanisms by which ATDC mediated radioresistance. We describe a direct protein-protein interaction between ATDC and RNF8 that is enhanced by DNA damage. The interaction between ATDC and RNF8 promoted ATDC trafficking to the nucleus and enhanced DNA DSB repair, ␥H2AX monoubiquitination, 53BP1 phosphorylation, and foci recovery of RNF8 and BRCA1 after IR treatment. Our results demonstrate a novel interaction between ATDC and RNF8 and reveal novel mechanistic insight into how this interaction participates in the radioprotective role of ATDC.

Experimental Procedures
Cells and Antibodies-HEK 293 cells were maintained in DMEM supplemented with 10% FBS and penicillin and streptomycin. The HEK 293 and human pancreatic ductal adenocarcinoma cell lines Panc1, BxPc3, and CAPAN2 were purchased from ATCC (Manassas, VA). The bladder cancer cell line, UC14, was obtained from Monica Liebert and subjected to DNA fingerprinting to confirm identity (19). All cells were grown in DMEM or RPMI 1640. Monoclonal anti-FLAGtagged antibody was purchased from Sigma. H2AX-, ␥H2AX-, and Myc-tagged antibodies were purchased from Cell Signaling (Danvers, MA). An anti-BRCA1 antibody was purchased from Millipore (Billerica, Massachusetts). Polyclonal anti-53BP1 and anti-p53BP1 antibodies were purchased from Bethyl Laboratories, Inc. (Montgomery, TX). Monoclonal anti-ATDC (B2), anti-H3, and anti-BRCA1 antibodies were obtained from Santa Cruz Biotechnology (San Jose, CA). The polyclonal anti-RNF8 antibody used was described previously (20). HEK 293 cell lines with stable expression of ATDC and its C-terminal deletion mutant have been described previously (17).
Constructs-The cDNA of human ATDC was kindly provided by J. Murnane (University of California, San Francisco). The full-length cDNA and the N-terminal or C-terminal truncation mutants of ATDC were subcloned into N-terminal p3ϫFLAG-CMV expression vector as described previously (17).
Cell Fractionation-Nuclear, cytoplasmic, membrane, and cytoskeleton proteins were extracted using a compartmental protein extraction kit (Millipore, Billerica, Massachusetts) according to manufacturer's instructions.
Chromatin Extraction-HEK 293 cells expressing either ATDC or ATDC⌬C (lacking amino acids 348 -588) were irradiated with 10 Gy and incubated for different times (1, 3, 6, and 24 h). Proteins bound to chromatin were released by treatment with 0.2 N HCl and analyzed by Western blotting.
Western Blot Analysis-Cells expressing full-length or deletion mutants of ATDC or RNF8 were lysed in cell lysis buffer (50 mM Tris-HCl, pH 8.0, containing 1% Triton X-100, 25 g/ml aprotinin and leupeptin, 1 mM PMSF, and 10% glycerol). The cell lysates were then centrifuged at 10,000 ϫ g for 10 min to remove debris. Protein concentrations were measured using the Bradford assay (Bio-Rad). Proteins separated by SDS-PAGE were transferred to nitrocellulose membranes. After blocking with blocking buffer (Tris-buffered saline (TBS), pH 7.4, with 0.1% Tween 20 (TBST) containing 5% skim milk powder) for 1 h at room temperature, the membranes were incubated for 1 h at room temperature with primary antibody. After incubation with secondary antibodies conjugated with HRP, the proteins were visualized using an ECL detection kit (Thermo Scientific).
Co-immunoprecipitation-Cells were transiently transfected with plasmids for expression of FLAG-tagged ATDC, Myctagged RNF8, or their various deletion mutants. Cells were resuspended in ice-cold lysis buffer (50 mM Tris, 150 mM NaCl, 0.5% IGEPAL, pH 7.4) containing freshly added protease inhibitors (Roche Applied Science, Mannheim, Germany) and sonicated for 5 s. The lysates were centrifuged for 15 min at 10,000 ϫ g to remove debris. As described previously, FLAGtagged ATDC or Myc-tagged RNF8 was immunoprecipitated by incubating with anti-FLAG, anti-ATDC (B2, Santa Cruz Biotechnology), or anti-Myc antibodies and protein G-agarose beads (Gibco, Life Technologies) at 4°C overnight (17). Immunoprecipitates were washed four times with ice-cold lysis buffer and resolved by reducing SDS-PAGE. Co-immunoprecipitated ATDC and its mutants were detected with anti-FLAG antibody (Sigma) or anti ATDC antibody (Santa Cruz Biotechnology). Co-immunoprecipitated RNF8 or its mutants were detected with anti-Myc antibody (Cell Signaling).
Trypsin Digestion-HEK 293 cells were transfected with FLAG-ATDC expression vector or control vector. 24 h after transfection, cell lysates were made and subjected to immunoaffinity purification with anti-FLAG conjugated to agarose beads. The bound proteins were eluted with 50 mM ethanolamine at pH 11.5 or 50 mM glycine buffer at pH 2.5. The eluates were lyophilized to 20-l volume. Proteins were reduced with 10% (v/v) DTT at 60°C for 30 min and then alkylated with 15 mM iodoacetamide in the dark at room temperature for 30 min. A 1:20 ratio of TPCK-treated trypsin (Promega, Madison, WI) was added to each sample, and the samples were vortexed and then incubated at 37°C overnight. The tryptic digestion was terminated by the addition of 2.5% v/v TFA.
LC-MS Analysis-Peptides from a tryptic digest of each sample were analyzed using a nano-LC/MS system consisting of an HPLC NanoACQUITY system (Waters). A C12 trap column (Jupiter Proteo-Phenomenex; particle size 90 Å, 75 m ϫ 3 cm) was utilized before analytical column for desalting at a flow rate 10 l/min. 30-l aliquots of the peptide solutions were loaded for each run. The trapped peptides were separated on an analytical column (Jupiter Proteo-Phenomenex: particle size 90 Å, 75 m ϫ 25 cm, C12) with a 300 nl/min flow rate ϫ 1-h acetonitrile gradient. The mobile phases A and B were 0 and 100% acetonitrile, respectively, each containing 0.1% formic acid. The gradient began at 5% B and was ramped to 18% B by 32 min, to 35% by 47 min, to 50% by 50 min, and finally to 80% by 52 min. In each mass analysis, one high mass resolution (60,000 full width at half maximum) MS spectra was acquired and scanned from 300 -1600 m/z in MS mode, followed by analysis of the 15 most abundant data-dependent MS/MS analyses (with dynamic exclusion for 45 s) throughout the collision-induced dissociation phase.
Collected MS/MS spectral data were analyzed for identification using the following analysis protocols. MS/MS spectra were searched with Mascot (version 2.3.02) under the conditions of 10 ppm parent ions mass tolerances, 0.8 Da product ion mass tolerances, and two missed cleavages. Carbamidomethylated cysteine (fixed), and Oxidation (M), N-acetyl (N-term), Deamidation (N,Q), Pyro-Glu (Q) (variable) were chosen as modifications.
Immunofluorescent Staining-HEK 293 cells expressing FLAG-tagged ATDC or ADTC⌬C were treated with ionizing radiation with 4 Gy and analyzed using immunofluorescence assays at the indicated time points after IR using antibodies against 53BP1 and BRCA1 at a dilution of 1:200 or RNF8 at a 1:500 dilution.
Pulldown Assays-Escherichia coli transformed with an inducible expression vector for RNF8 were cultured at 37°C, until A 600 reached ϳ0.8, and then 1 mM isopropyl-␤-D-1-thiogalactopyranoside was added for an overnight incubation at room temperature to induce RNF8 expression. The bacteria were harvested and lysed by sonication on ice in 50 mM Tris-Cl (pH 8.0), 150 mM NaCl, 0.2% sodium deoxycholate, and 0.1% Triton X-100. The lysate was cleared by centrifugation, and RNF8 prebound glutathione-Sepharose beads were prepared according to protocol as described previously (21). FLAG-ATDC was in vitro translated directly from the PCR products that contain the T7 promoter by TNT quick coupled transcription/translation system according to manufacturer's protocol (Promega). Pulldown reactions were then initiated by the addition of GST-RNF8 or GST prebound to glutathione-Sepharose beads to in vitro translated products or lysates of HEK 293 cells with stable expression of ATDC. After rocking overnight at 4°C, the beads were washed three times with TBS containing 250 mM NaCl and 0.25% Triton X-100 and analyzed by Western blotting.
Comet Assays-ATDC-or ATDC⌬C-expressing HEK 293 cells growing in complete DMEM medium at 80% confluence were irradiated (15 Gy). DNA double-strand repair was analyzed in neutral comet assays using the Trevigen comet assay kit (4250-050-K) according to the manufacturer's instructions.
Statistical Analysis-Data are represented as mean Ϯ S.E. or S.D. as indicated from at least three independent experiments. Significance of differences between groups was evaluated by Student's t test or analysis of variance. p Ͻ 0.05 was considered statistically significant.

Results
ATDC Binds Directly to RNF8 -To better characterize the biological function of ATDC, we performed a screen to identify ATDC binding partners by affinity purification and mass spectrometric analysis. FLAG-tagged ATDC was expressed in HEK 293 cells (which have no endogenous ATDC expression), and trypsin-digested immunoprecipitates were eluted using acidic or basic elution buffers and analyzed by mass spectrometry (17). Polypeptide sequences from ϳ250 distinct proteins were identified using this methodology, including previously described binding partners such as vimentin, providing validation of the screen methodology (Table 1). We also found peptide sequences from histones (including histones H1.2, H2B, and H4), heat shock protein 70, ribosomal proteins (40S and 60S), and PARP1, as well as Prohibitin-2 and myristoylated alaninerich protein kinase C substrate (MARCK), which have putative roles in invasion and metastasis ( Fig. 1, Table 1) (8,(22)(23)(24). The most striking ATDC-interacting protein identified by this screen was RNF8 (37% peptide coverage by mass spectroscopy, Fig. 1), suggesting that ATDC and RNF8 are interaction partners. Of the putative interaction partners of ATDC, we decided to focus on RNF8 due to its known role in the DNA damage response and our prior data indicating a role for ATDC in IR resistance (25).
To confirm that the ATDC and RNF8 proteins physically associate, we next performed co-immunoprecipitation experiments using HEK 293 cells (which have low/absent endogenous ATDC expression), transfected with either empty vector or overexpressing FLAG-ATDC. RNF8 consistently co-immunoprecipitated with FLAG-ATDC ( Fig. 2A). Immunoprecipitation of endogenous ATDC in BxPC3 pancreatic cancer cells that express high endogenous levels of ATDC also confirmed co-IP of endogenous RNF8 (Fig. 2B). To determine whether ATDC binds directly to RNF8, we performed GST pulldown assays and found that GST-RNF8 pulled down both immunoprecipitated FLAG-ATDC and in vitro translated FLAG-ATDC, indicating that the interaction between the proteins is direct (Fig. 2C). These results establish a novel and direct physical interaction between ATDC and RNF8.
The ATDC-RNF8 Interaction Is Mediated by the C Terminus of ATDC and the RING Domain of RNF8 -To define the domains critical for the interaction between ATDC and RNF8, we first created a series of truncation mutants of FLAG-tagged ATDC (Fig. 3A). Co-IP of FLAG-tagged full-length and truncated ATDC proteins with RNF8 revealed that the deletion of the C terminus (⌬C) abolished the interaction of ATDC with RNF8 and that expression of the ATDC C terminus alone (ATDC⌬348) was sufficient to mediate interaction with RNF8 (Fig. 3, B and C). To identify whether the FHA or RING domains of RNF8 were required for interaction with ATDC, we created Myc-tagged constructs lacking these domains (Fig. 4A), which were then transfected and co-expressed with ATDC in HEK 293 cells (Fig. 4, B and C). Although deletion of the RNF8 FHA domain (⌬FHA) did not disrupt interaction with ATDC, loss of the RING domain (⌬RING) of RNF8 completely blocked the interaction with ATDC, demonstrating that the RING domain of RNF8 is required for ATDC binding (Fig. 4B). These results define a novel interaction between the C terminus of ATDC and the RING domain of RNF8.
We have previously characterized phosphorylation of ATDC on the Ser-550 residue as a critical determinant of the radioprotective function of ATDC, which is mediated by MAPKAPK2 (25). This residue maps to the C-terminal interaction domain of ATDC, and so we hypothesized that the interaction between ATDC and RNF8 might require this Ser-550 residue. To determine this, we performed co-IP experiments in HEK 293 lysates co-expressing RNF8 and either WT ATDC or ATDC S550A mutant proteins. We observed that mutation of Ser-550 to alanine abrogated ATDC co-IP with RNF8 (Fig. 4C). The RING domain of RNF8 is required for its ubiquitin ligase function as well as interaction with ATDC (Fig. 4B). The typical function of RING domains is ubiquitin ligation. Mutation of the RNF8 RING domain Cys-403 residue to Ser blocks ligation function. To determine whether the ubiquitin ligase function of RNF8 was required for the ATDC-RNF8 interaction, we also performed co-IP with RNF8 C403S mutant protein. RNF8 C403S co-immunoprecipitated with ATDC (Fig. 4C), indicating that ligase function was dispensable for the RNF8-ATDC interaction.
To further confirm that the RNF8-ATDC interaction occurs under physiologic conditions, we expressed RNF8 and its ⌬FHA and ⌬RING truncation mutant proteins in UC14 and CAPAN2 tumor cells, which have endogenous ATDC expression. As seen in HEK 293, wild type, but not ⌬RING RNF8, co-immunoprecipitated with endogenous ATDC (Fig. 4D). Taken together, these data demonstrate that the C terminus of ATDC interacts with the RING domain of RNF8 and that this interaction requires the Ser-550 residue of ATDC, but not the RNF8 ubiquitin ligase function.
ATDC Is Present in the Cytoplasm and Nucleus and Requires Its Coiled-coil Domain for Nuclear Shuttling-ATDC has previously been reported to localize predominantly to the cytoplasm and cytoskeleton of cells (10). Other groups have reported that in certain cell lines, it is predominately present in the nucleus (18). RNF8 has been shown to have nuclear localization and to localize to sites of DNA damage after irradiation (4). Because ATDC expression leads to resistance to multiple forms of DNA damage, including IR, and because we demonstrated a physical interaction between these molecules, we next wanted to determine where ATDC and RNF8 interact within cells (25). Under basal conditions, RNF8 localized exclusively to the nuclear fraction when overexpressed in HEK 293 cells (Fig.  5A), whereas ATDC was present in cytoplasmic, cytoskeletal, and nuclear fractions in both HEK 293 cells (Fig. 5B) and the pancreatic cell line BxPC3 (Fig. 5C), which has high endogenous levels of ATDC. To confirm that ATDC localized to the cytoplasm and the nucleus, cells expressing ATDC were immunostained and examined by confocal microscopy. Although ATDC was predominately cytoplasmic, 20% was nuclear, confirming the cell fractionation experiments (Fig. 5D). To explore whether certain domains of ATDC were responsible for nuclear localization of ATDC, HEK 293 cells stably transfected with full-length or truncated FLAG-ATDC mutants were examined by cell protein fractionation. Interestingly, although loss of the ATDC C terminus, which binds to RNF8 (ATDC⌬C), did not block trafficking to the nucleus, loss of the coiled-coil domain (ATDC⌬348) did block detection of ATDC in the nucleus (Fig. Ionizing Radiation Increases ATDC Nuclear Trafficking and Binding to RNF8 -We next hypothesized that IR may induce ATDC translocation to the nucleus and promote binding to RNF8, facilitating DNA repair. To determine whether IR induces ATDC nuclear translocation, HEK 293 cells expressing ATDC or ATDC⌬C were treated with IR (10 Gy) and the cytoplasmic and nuclear fractions were isolated at various time points (1, 3, 6, and 24 h) following IR. We observed an increase in ATDC in the nuclear fraction starting at 1 h and increasing until 24 h after IR, which also corresponded to a decrease in the cytoplasmic portion of ATDC (Fig. 6A, top three panels). Similar results were seen with the ATDC⌬C construct, indicating that IR induces ATDC trafficking to the nucleus and that this process does not require the C terminus of ATDC (Fig. 6A, bottom three panels).
To determine whether IR also promoted the ATDC-RNF8 interaction, we next performed co-immunoprecipitation of  Myc-RNF8 and FLAG-ATDC in cells before and 1 or 3 h after irradiation with 10 Gy. Treatment with IR increased ATDC pulldown with RNF8 at 1 h after irradiation, but this increase returned to baseline levels by 3 h (Fig. 6B). Conversely, immunoprecipitation of FLAG-ATDC or FLAG-ATDC⌬C from whole cell extracts following irradiation resulted in increased pulldown of RNF8 by wild type ATDC, but not ATDC⌬C, confirming the importance of the ATDC C terminus for RNF8 binding (Fig. 6C).
RNF8 is recruited to nuclear DNA repair foci (marked by H2AX phosphorylation) following IR (1). Given that ATDC is recruited to the nucleus and binds to RNF8, we hypothesized that ATDC might be recruited to IR-induced foci. To examine this, we immunostained BxPC3 cells before and after IR for expression of ␥-H2AX and ATDC. Although we observed formation of numerous ␥-H2AX foci, we did not observe ATDC recruitment to the IR-induced H2AX foci (Fig. 7A). To further confirm the lack of ATDC recruitment to DNA damage foci, we transfected Panc1 cells with GFP-ATDC and microlaser-irradiated cells, a mechanism to induce high levels of localized DNA damage. Although GFP-RNF8 is rapidly recruited to laser-irradiated DNA damage tracts, we found no evidence that ATDC accumulated at sites of DNA damage (Fig. 7, B and C) (1).
Although ATDC is not recruited to sites of DNA damage, we postulated that it may influence the kinetics of RNF8 recruit-ment to the chromatin, and thus influence the DNA damage response. To test this, we examined recruitment of RNF8 and ATDC to chromatin following IR. In control cells lacking ATDC, RNF8 was rapidly (by 1 h) and persistently (Ͼ24 h) recruited to the chromatin (Fig. 6D, top panels). In contrast, in ATDC-expressing HEK 293 cells, RNF8 chromatin loading was less rapid (peak at 6 h) and returned to basal levels by 24 h (Fig.  6D, middle panels). Furthermore, although ATDC was not recruited to DNA damage sites, IR did induce ATDC loading to chromatin with similar kinetics to RNF8 (Fig. 6D). Deletion of the ATDC RNF8 binding domain reduced association of ATDC with chromatin, suggesting that ATDC loading to chromatin is dependent on interaction with RNF8 (Fig. 6D, bottom panels). Taken together, these results establish ATDC nuclear trafficking and binding to RNF8 as a regulator of ATDC recruitment and binding to chromatin following exposure to IR and suggest that this interaction may influence the stability of RNF8 binding to chromatin following DNA damage.
ATDC-mediated Radioresistance Requires Both N Terminal and C Terminal Regions-We have previously shown that ATDC expression leads to enhanced growth and resistance to IR and that knockdown or loss of ATDC sensitizes cells to IR (17,25). Others have demonstrated that knockdown of ATDC sensitizes cells to UV irradiation (26). ATDC has also been reported to interact with p53 through its N terminus and to inhibit p53 nuclear activity (10). To confirm the importance of ATDC for radioresistance, we next investigated the survival rate of HEK 293 cells with stable expression of full-length ATDC, ATDC⌬220(⌬N), or ATDC⌬C (Fig. 8A). Interestingly, although full-length ATDC increased resistance to IR as expected, cells expressing either of the ATDC truncation mutants ATDC⌬N or ATDC⌬C showed a partial resistance to IR (versus full-length ATDC) but still had a significant increase in survival when compared with HEK 293 cells lacking ATDC (Fig. 8, B and C). These data indicate that both the N termini and the C termini of ATDC participate in protection against IR.
The primary cytotoxic lesion induced by IR is DNA DSBs. To examine whether the radioprotective effect of ATDC correlated with increased DNA DSB repair, we used neutral comet assays to measure whether ATDC expression altered accumulation and resolution of DSBs. When compared with control cells, cells stably expressing ATDC showed shorter comet tails 6 h after exposure to IR, and this effect was more pronounced 24 h after IR (Fig. 8D), suggesting that expression of ATDC increased the kinetics of DNA repair. In contrast, cells expressing ATDC⌬C (which lack the ability to bind RNF8) showed no significant change in length of comet tails when compared with control cells (Fig. 8, D and E). These results indicate that ATDC promotes radioresistance by improved DNA DSB repair kinetics and that this effect requires its C-terminal RNF8 interaction domain.
The ubiquitin ligase function of RNF8 promotes phosphorylation of 53BP1 and its recruitment to DNA DSB repair foci (27). To determine whether the RNF8-ATDC interaction promoted phosphorylation of 53BP1 following IR, we measured phospho-53BP1 levels following IR in cells with ATDC,  ATDC⌬C, or no ATDC (Fig. 9, C and D). Expression of fulllength ATDC, but not ATDC⌬C, promoted increased phosphorylation of 53BP1 following IR. These data suggest that the interaction of ATDC with RNF8 influences RNF8-mediated ␥-H2AX mono-ubiquitination and enhances phosphorylation of 53BP1 in response to IR.
Phosphorylation of H2AX occurs rapidly following induction of DNA DSB and is a marker of persistent DNA damage (28). To assess the effect of the ATDC-RNF8 interaction on IR-induced H2AX phosphorylation, HEK 293 cells with and without ATDC or ATDC⌬C were treated with IR (10 Gy) and H2AX phosphorylation and mono-ubiquitination was measured by Western blotting. As seen in Fig. 9A, ATDC expression promoted mono-ubiquitination following IR (Fig. 10A). In control vector-and ATDC⌬C-expressing cells, IR treatment also resulted in H2AX phosphorylation (␥H2AX band) after as early as 10 min, which persisted for 24 h (Fig. 10). Irradiation of ATDC-expressing cells resulted in a similar pattern of induc-  tion of H2AX phosphorylation, but a more rapid return to basal levels after 3 h (Fig. 10, B and C).
H2AX and BRCA1 DNA damage foci are markers of DNA DSBs. To examine whether ATDC and ATDC⌬C influenced kinetics of foci formation and resolution, we stained HEK 293 cells without and with either ATDC or ATDC⌬C following IR (Fig. 11). We noted that RNF8 and BRCA1 formed nuclear foci rapidly with similar numbers of foci and fluorescence intensity in both ATDC-expressing and control cells 1 h following IR (4 Gy) (Fig. 11, data not shown). However, ATDC-expressing cells had a more rapid resolution of the RNF8 and BRCA1 foci at the 6-h time point after IR and almost complete resolution of foci at 24 h. In contrast, ATDC⌬C stably expressing cells showed no difference in the time course of foci recovery when compared with control cells lacking ATDC (Fig. 11, B and C). These results show that the interaction between ATDC and RNF8 enhances recovery of DNA damage and correlates with enhanced resolution of H2AX phosphorylation and RNF8/BCRA1 foci following IR in a time-dependent manner.  . ATDC promotes IR induced ␥H2AX mono-ubiquitination and 53BP1 phosphorylation following IR. A, wild type ATDC, but not ATDC⌬C, promoted mono-ubiquitination of ␥H2AX (␥H2AX-ub (mono)) following IR (10 Gy) as measured by Western blot. B, quantitation of -fold increase in monoubiquitination of ␥H2AX normalized to H2AX relative to untreated control (mean Ϯ S.D., **, p Ͻ 0.05 versus vector control). C, wild type ATDC, but not ATDC⌬C, enhanced 53BP1 phosphorylation (p-53BP1) as measured by Western blot. D, quantitation of -fold increase in phosphorylation of 53BP1 normalized to unphosphorylated 53PB1 and relative to untreated control (mean Ϯ S.D., **, p Ͻ 0.05 versus vector control).

Discussion
DNA damage-induced cell death is an integral component of anticancer therapy. Understanding the DNA damage response that restores genomic integrity following cancer cell treatment and thus mitigates cell death is essential to understand therapy resistance and develop ways to reverse it. We have previously reported that ATDC becomes phosphorylated by MK2 downstream of p38 and ATM and induces resistance to IR, but the downstream mechanism(s) by which ATDC provides resistance to IR-induced DNA damage remained elusive (25). To identify potential downstream effectors, we performed a mass spectrometry-based screen for proteins that physically interact with ATDC. This screen identified the DNA repair factor, RNF8, as a putative binding partner for ATDC (3,4). In this study, we confirm that ATDC shares a direct physical contact with RNF8 that is enhanced following induction of DNA dam-  . Expression of ATDC promoted RNF8 and BRCA1 DNA repair foci resolution. A, immunofluorescent staining for RNF8 following treatment with 4 Gy IR. B and C, cells expressing ATDC or ATDC⌬C were treated with IR for various time points, and RNF8 (B) and BRCA1 (C) foci were quantitated. Error bars ϭ S.E. *, p Ͻ 0.01 when compared with vector or ATDC⌬C controls.
age. We have further refined the interaction domains of these proteins to the C terminus of ATDC and the RING domain of RNF8 and demonstrate that the interaction requires the ATDC Ser-550 residue (which is phosphorylated following IR) but is not dependent on the ubiquitin ligase function of RNF8. Furthermore, we demonstrate that this interaction results in improved DNA DSB repair kinetics as measured by neutral comet assays, increased clonogenic survival, and a more rapid resolution of DNA repair foci following exposure to IR. These results establish the ATDC-RNF8 interaction as a key mechanism of ATDC-induced resistance to IR. ATDC has previously been described as a driver of resistance to DNA damage by UV and ionizing radiation that binds to chromatin, but the exact means by which it promotes resistance to DNA damage was unclear (18,25,26). Like ATDC, RNF8 is activated in an ATM-dependent manner and participates in DDR by facilitating DNA DSB processing through ubiquitination of histones and DNA repair factors promoting the assembly of DNA repair complexes at sites of DNA DSB (3,4). Here we characterize a direct interaction between ATDC and RNF8 that is enhanced by exposure to IR and that contributes to ATDC-induced IR resistance. Interestingly, the RNF8-ATDC interaction depended on the ATDC Ser-550 residue, which we have previously shown to be phosphorylated by MK2 downstream of ATM and which is required for the role of ATDC in radioresistance (25). We also find that although ATDC is predominantly cytoplasmic in most cell lines expressing it, some ATDC is present in the nucleus. Unexpectedly, ATDC does not itself appear to localize to nuclear repair foci but does bind to the chromatin and influence RNF8 chromatin binding following induction of DNA damage (Fig. 6D). These findings suggest that the ATDC-RNF8 interaction may have a more global effect on RNF8 nuclear dynamics, which occurs outside of the DNA repair foci but which is nonetheless important for efficient DNA repair (Table 1). These results are also consistent with a recent study identifying ATDC/TRIM29 as a nuclear DDR participant that binds to the chromatin following DNA damage (18).
RNF8 and its partner RNF168 are RING domain-containing ubiquitin ligases that are recruited to DNA DSB following H2AX phosphorylation by MDC1 (29). Following DNA damage, RNF8 and RNF168 catalyze the addition of ubiquitin to H2A and H2AX and recruitment of 53BP1 and BRCA1 to DNA damage sites, facilitating repair. Although the ATDC-RNF8 interaction does not influence initial RNF8 or BRCA1 foci formation, it did result in faster disappearance of these foci, suggesting that it speeds DNA DSB repair processes. Although the exact means by which the ATDC-RNF8 interaction impacts the function of RNF8 is incompletely understood, ATDC is known to bind the acetyltransferase Tip60, which is involved in chromatin remodeling and promotes RNF8 ubiquitin ligase activity and DNA repair (12,30). Furthermore, our mass spectrometry screen also identified a putative interaction between ATDC and RuvBL2 (Tip48), another member of the p400/NuA4 complex that modulates RNF8 activity (Table 1) (18). We therefore hypothesize that ATDC may promote RNF8 activity, perhaps by physically binding to and coordinating the activity of the chromatin-remodeling complexes necessary for RNF8 to promote efficient DNA DSB repair.
Although ATDC is important for cell survival, proliferation, and DNA DSB repair, it is notable that it lacks intrinsic enzymatic function. Indeed, ATDC seems to function as a direct protein link between proteins involved in DNA damage sensing (ATM-p38-MK2) and cell cycle regulation (DVL2, p53) (10,17,25). These results implicate RNF8 as an additional binding partner that influences DNA repair and resistance to damage. Together these findings suggest ATDC as a multifunctional scaffold protein that undergoes covalent modifications (phosphorylation, acetylation) following cellular insults and subsequently binds to multiple proteins in the cytoplasm and nucleus to coordinate cellular survival and DNA repair (10,17,25).
Induction of DNA damage leading to apoptosis and cell death by agents such as chemotherapy and ionizing radiation remains a critical aspect of modern cancer treatments. ATDC is highly expressed in many tumor types, binds to the DNA repair factor RNF8 and thus may be a determinant of resistance to both cytotoxic chemotherapy and ionizing radiation. RNF8 is required for robust DNA DSB repair, is expressed in many tumors, and appears to participate in acquired resistance to PARP inhibition (COSMIC (Catalogue of Somatic Mutations in Cancer)) (31). The identification of this novel interaction between ATDC and RNF8 establishes a new mechanism whereby ATDC and RNF8 could function in a coordinate manner to produce resistance to DNA damage-based anticancer therapies. Expression of these proteins may represent prognostic biomarkers that, if targeted, could result in tumor sensitization to therapy.