PTIP Regulates 53BP1 and SMC1 at the DNA Damage Sites*

Although PTIP is implicated in the DNA damage response, through interactions with 53BP1, the function of PTIP in the DNA damage response remain elusive. Here, we show that RNF8 controls DNA damage-induced nuclear foci formation of PTIP, which in turn regulates 53BP1 localization to the DNA damage sites. In addition, SMC1, a substrate of ATM, could not be phosphorylated at the DNA damage sites in the absence of PTIP. The PTIP-dependent pathway is important for DNA double strand breaks repair and DNA damage-induced intra-S phase checkpoint activation. Taken together, these results suggest that the role of PTIP in the DNA damage response is downstream of RNF8 and upstream of 53BP1. Thus, PTIP regulates 53BP1-dependent signaling pathway following DNA damage.

The DNA damage response pathways are signal transduction pathways with DNA damage sensors, mediators, and effectors, which are essential for maintaining genomic stability (1)(2)(3). Following DNA double strand breaks, histone H2AX at the DNA damage sites is rapidly phosphorylated by ATM/ATR/ DNAPK (4 -10), a family homologous to phosphoinositide 3-kinases (11,12). Subsequently, phospho-H2AX (␥H2AX) provides the platform for accumulation of a larger group of DNA damage response factors, such as MDC1, BRCA1, 53BP1, and the MRE11⅐RAD50⅐NBS1 complex (13,14), at the DNA damage sites. Translocalization of these proteins to the DNA double strand breaks (DSBs) 3 facilitates DNA damage checkpoint activation and enhances the efficiency of DNA damage repair (14,15).
Recently, PTIP (Pax2 transactivation domain-interacting protein, or Paxip) has been identified as a DNA damage response protein and is required for cell survival when exposed to ionizing radiation (IR) (1, 16 -18). PTIP is a 1069-amino acid nuclear protein and has been originally identified in a yeast two-hybrid screening as a partner of Pax2 (19). Genetic deletion of the PTIP gene in mice leads to early embryonic lethality at embryonic day 8.5, suggesting that PTIP is essential for early embryonic development (20). Structurally, PTIP contains six tandem BRCT (BRCA1 carboxyl-terminal) domains (16 -18, 21). The BRCT domain is a phospho-group binding domain that mediates protein-protein interactions (17,22,23). Interestingly, the BRCT domain has been found in a large number of proteins involved in the cellular response to DNA damages, such as BRCA1, MDC1, and 53BP1 (7, 24 -29). Like other BRCT domain-containing proteins, upon exposure to IR, PTIP forms nuclear foci at the DSBs, which is dependent on its BRCT domains (16 -18). By protein affinity purification, PTIP has been found in two large complexes. One includes the histone H3K4 methyltransferase ALR and its associated cofactors, the other contains DNA damage response proteins, including 53BP1 and SMC1 (30,31). Further experiments have revealed that DNA damage enhances the interaction between PTIP and 53BP1 (18,31).
To elucidate the DNA damage response pathways, we have examined the upstream and downstream partners of PTIP. Here, we report that PTIP is downstream of RNF8 and upstream of 53BP1 in response to DNA damage. Moreover, PTIP and 53BP1 are required for the phospho-ATM association with the chromatin, which phosphorylates SMC1 at the DSBs. This PTIP-dependent pathway is involved in DSBs repair.
The siRNA duplexes were purchased from Dharmacon Research (Lafayette, CO). The siRNA sequence targeting PTIP and 53BP1 are 5Ј-AAG GAA GAA GAG GAA GAG GAA-3Ј and 5Ј-AAG AUA CUC CUU GCC UGA UAA-3Ј, respectively. siRNAs were transfected into the cells using oligofectamine (Invitrogen) according to the manufacturer's instructions.
Cell Culture and Treatment with Ionizing Radiation-U2OS cells were cultured in RPMI 1640 medium with 10% fetal bovine serum. H2AX-, MDC1-, RNF8-, and 53BP1-deficient MEFs have been described elsewhere (32)(33)(34)(35). The ptip flox/Ϫ MEFs were previously described (31). RAP80-deficient MEFs were provided by Dr. Junjie Chen. All of the MEFs were cultured in Dulbecco's modified Eagle's medium with 10% fetal bovine serum. The ptip flox/Ϫ MEFs were infected with Cre-expressing adenovirus (ad-Cre, University of Michigan Vector Core). For IR, the cells were irradiated by JL Spepherd 137 Cs radiation source with indicated doses. The cells were then maintained in the culture conditions for indicated time points specified in the figure legends.
Cell Lysis and Western Blot-The cells were lysed with buffer containing 0.5% Nonidet P-40, 50 mM Tris-HCl, pH 8.0, 2 mM EDTA, and 300 mM NaCl for further Western bolt analysis. For soluble and insoluble (chromatin) fractions, the cells were lysed with buffer containing 0.5% Nonidet P-40, 50 mM Tris-HCl, pH 8.0, 2 mM EDTA, and 100 mM NaCl, and insoluble pellets were resuspended in 0.2 M HCl and then were neutralized with 1 M Tris-HCl, pH 8.5.
Immunofluorescence Staining-To visualize IR-induced foci (IRIF), the cells were cultured on coverslips and treated with 5 Gy of IR followed by recovery for 4 h and then were fixed in 3% paraformaldehyde for 20 min and permeabilized in 0.5% Triton X-100 in phosphate-buffered saline for 5 min at room temperature. The samples were blocked with 5% goat serum and then incubated with primary antibody for 60 min. The samples were washed three times and incubated with secondary antibody for 30 min. The coverslips were mounted onto glass slides and visualized by a fluorescence microscope.
In Vitro Kinase Assay-Wild type, PTIP, and 53BP1-deficient cells were treated with 10 Gy of IR and allowed to recover for 1 h. The cells were lysed in NTN 300 buffer (0.5% Nonidet P-40, 50 mM Tris-HCl, pH 8.0, and 300 mM NaCl). The extract was incubated with ATM antibody and protein G beads at 4°C for 2 h. After washing three times with the kinase buffer (10 mM HEPES, pH 7.5, 50 mM glycerophosphate, 50 mM NaCl, 10 mM MgCl, 10 mM MnCl, and 1 mM dithiothreitol), the beads were resuspended in the kinase buffer with or without ATP and incubated with purified GST-CHK2 (amino acids 1-250) fusion protein at 37°C for 30 min. The proteins were eluted with SDS loading buffer, separated by SDS-PAGE, and analyzed by Western blot using anti-phospho-CHK2-T68 antibody.
Neutral Comet Assays-Single-cell gel electrophoretic comet assays were performed under neutral conditions (36). Briefly, ptip flox/Ϫ MEFs with or without treatment of Ad-Cre were arrested at G 1 by treated with 2 mM thymidine for 24 h. The cell cycle profile was confirmed by flow cytometry (data not shown). The cells were irradiated with 25 Gy and incubated in culture medium at 37°C for the indicated times. For cellular lysis, the slides were immersed in neutral lysis solution (2% sarkosyl, 0.5 M EDTA, 0.5 mg/ml proteinase K, pH 8.0) overnight at 37°C. On the second day, after electrophoresis at 15 V for 25 min (0.6 V/cm), the slides were stained for 20 min with 10 g/ml propidium iodide and viewed in a fluorescence microscope. The comet tail moment was analyzed by CometScore software.
Pulsed Field Gel Electrophoresis-MEFs were exposed to 80-Gy IR. At the indicated time after IR, the cells were har-vested by trypsinization, washed, and resuspended in phosphate-buffered saline. Low melting point agarose was added to a final concentration of 0.8%, and 0.1-ml plugs were cast containing ϳ300,000 cells of each. The plugs were incubated in a 0.5 M EDTA solution containing 2% sarkosyl and 10 g/ml proteinase K at 56°C for 24 h, followed by wash plugs twice with TE buffer (pH 8.0) for 2 h. Pulsed field gel electrophoresis was performed on a Bio-Rad Chef II Mapper for 36 h optimized to the 1-4-Mb range. The gels were stained with Sybr green I (Invitrogen) overnight in the dark at 4°C and then were scanned by using a phosphorimaging device in fluorescence mode (GE HealthCare). The DNA fragmentation was detected as DNA migrated into the gel and calculated as a ratio of DNA migrated in each lane of the gel relative to the total DNA loaded (separated and nonseparated DNA). The DNA double strand breaks level was calcu-FIGURE 1. IRIF of PTIP is controlled by H2AX, MDC1, and RNF8. Wild type (WT) and H2AX-, MDC1-, RNF8-, 53BP1-, and RAP80-deficient MEFs were exposed to 5 Gy of IR. Four hours after IR, the cells were fixed and immunostained with rabbit anti-mouse PTIP polyclonal antibody and anti-␥H2AX monoclonal antibody. DAPI, 4Ј,6-diamidino-2-phenylindole. lated as a ratio of the amount of DNA fragmentation at each recovery time point to that at time 0 following DNA damage. The mean value at each time point is from three independent experiments.
Radioresistant DNA Synthesis Assay-ptip flox/Ϫ MEFs were treated with or without adeno-cre and plated at 40% confluency in 60-mm dishes in fresh medium containing 0.01 Ci ml Ϫ1 of 14 C-labeled thymidine (Amersham Biosciences). After 24 h, the medium was replaced with thymidine-free medium, and the cells were treated with 5-Gy IR in a Gam-maCell 132 Cesium source. The cells were returned to the incubator for 30 min before the addition of 20 Ci/ml of 3 H-labeled thymidine (Amersham Biosciences). At the indicated time points, the cells were washed, trypsinized, pelleted, and frozen on dry ice until all of the samples were collected. The samples were lysed in 500 l of 0.1 N NaOH, 1% SDS, 1 mM EDTA. After incubation at 65°C for 30 min, an equal volume of 20% trichloroacetic acid was added to the samples, which were immediately and thoroughly mixed. Samples were applied to Whatman GF/A filters on a vacuum apparatus and were washed extensively with 10% trichloroacetic acid and then with 100% ethanol. After air-drying, the filters were placed in vials with 3 ml of scintillation fluid and counted on a Packard Tri-Carb 2000 liquid scintillation analyzer using the standard dual label procedure. The data are presented as the percentage of DNA synthesis at the various time points, where the percentage of DNA synthesis ϭ ( 3 H dpm/ 14 C dpm) treated/ ( 3 H dpm/ 14 C dpm) untreated ϫ 100.

RESULTS
PTIP Is Downstream of H2AX, MDC1, and RNF8 following DNA DSBs-Following DNA damage, PTIP is quickly accumulated at DSBs and colocalizes with ␥H2AX (16 -18). However, it remains unclear how this dynamic process is regulated. Following DNA damage, ␥H2AX recruits MDC1, then RNF8, to DSBs (24 -27, 29, 37-39). This pathway regulates several downstream proteins relocation to the DNA damage sites, such as 53BP1 and the RAP80⅐BRCA1 complex (37)(38)(39)(40). To study the role of PTIP in this DNA damage response pathway, we have examined IRIF of PTIP in several DNA damage responsedeficient cells. Wild type, H2AX, MDC1, RNF8, 53BP1, and RAP80deficient MEFs were treated with 5 Gy of IR. As shown in Fig. 1, PTIP failed to relocate to DSBs in the absence of H2AX, MDC1, or RNF8. But IRIF of PTIP did not require 53BP1 and RAP80, both of which are downstream of RNF8. These results suggest that the ␥H2AX-MDC1-RNF8 pathway controls PTIP IRIF.
PTIP Regulates 53BP1 Foci Formation following DNA Damage-Because PTIP foci formation is independent of 53BP1 and RAP80, we next examine whether PTIP is upstream of 53BP1 and the RAP80-BRCA1 complex following DNA damage. PTIP-deficient MEFs were generated from ptip flox/Ϫ MEFs by introducing Cre recombinase (31). We confirmed that PTIP was not detected by Western blot in the PTIP-deficient MEFs ( Fig. 2A). Moreover, the IRIF of RNF8, RAP80, and BRCA1 was intact in the PTIP-deficient MEFs (Fig. 2B). However, 53BP1 foci formation was abrogated in the absence of PTIP (Fig. 2B), indicating that PTIP regulates IRIF of 53BP1 but not that of the RAP80-BRCA1 complex.
PTIP and 53BP1 Regulate Phospho-ATM Association with Chromatin-It has been shown that MDC1, the upstream regulator of PTIP and 53BP1, modulated ATM association with chromatin (34), which prompted us to examine the role of PTIP and 53BP1 in the ATM association with chromatin. Following DNA damage, chromatin-associated ATM are dramatically increased (34). We also confirmed this phenomenon in the normal MEFs following DNA damage (data not shown). However, compared with that in the normal MEFs, chromatin-bound phospho-ATM and ATM were significantly reduced in the PTIP-and 53BP1-deficient MEFs (Fig. 3A). These results suggest that PTIP and 53BP1 regulate the phospho-ATM association with chromatin.
We noticed that although chromatin-associated phospho-ATM are dramatically decreased in PTIP-and 53BP1-deficient cells, the phospho-ATM in soluble fraction are significantly increased, suggesting that PTIP and 53BP1 may not affect the phosphorylation and kinase activity of ATM following DNA damage. This hypothesis was confirmed by a Western blot to detect the phospho-ATM. As shown in Fig.  3B, compared with that in the normal MEFs, ATM was still phosphorylated in the PTIP-and 53BP1-deficient MEFs following IR, although the phosphorylation level of ATM was modestly reduced. Moreover, the kinase activity of ATM from the PTIP-and 53BP1-deficient MEFs was not significantly altered in an in vitro kinase assay using recombinant CHK2 as the substrate (Fig. 3C). These results together suggest that PTIP and 53BP1 regulate the phospho-ATM and ATM association with chromatin, although they cannot affect the phosphorylation and kinase activity of ATM.
PTIP and 53BP1 Regulate ATM-dependent SMC1 Phosphorylation at the DNA Damage Sites-Because PTIP and 53BP1 regulate chromatin-bound ATM, we hypothesize that ATM substrates on the chromatin may also be regulated by PTIP and 53BP1. SMC1 is a subunit of sister chromatin cohesion complex that is incorporated into the chromatin during S phase and maintains the association between the sister chromatids (41,42). Following DNA damage, chromatin-associated SMC1 is phosphorylated by ATM at Ser 957 and Ser 966 (43)(44)(45). Phosphorylated SMC1 is important for DSBs repair (43)(44)(45)(46)(47). To examine the role of PTIP and 53BP1 in ATMdependent SMC1 phosphorylation, we used siRNA to downregulate PTIP and 53BP1 in U2OS cells and found that the phosphorylation of SMC1 was abolished in the PTIP and 53BP1 depleting cells following IR (Fig. 4A), suggesting that PTIP and 53BP1 regulate the ATM-dependent SMC1 phosphorylation. Again, in the PTIP and 53BP1 depleting cells, ATM phosphorylation was not significantly altered (Fig. 4A), which is consistent with our observation in the PTIP-and 53BP1-deficient MEFs (Fig. 3B). In addition, we also examined the phosphorylation level of p53 and Chk2, another two known substrates of ATM, and found that Ser 15 phosphorylation of p53 and Thr 68 phosphorylation of Chk2 were only modestly reduced in PTIP-and 53BP1-depleting cells (supplemental Fig. S1), indicating that other phosphoinositide 3-kinases including ATR and DNAPK may function redundantly with ATM to phosphorylate downstream substrates such as Chk2 and p53 following DNA damage.
Next, we examined the IRIF of phospho-SMC1 in the presence or absence of PTIP and 53BP1. As shown in Fig. 4B, phospho-SMC1 formed nuclear foci and colocalized with ␥H2AX and 53BP1 in control siRNA-treated cells. In contrast, the IRIF of phospho-SMC1 was abolished in the PTIP-and 53BP1-depleting cells. Furthermore, we could not detect phospho-SMC1 signals by immunostaining (Fig. 4B). In addition, the IRIF of 53BP1 was also abolished in the PTIP depleted cells, further confirming that 53BP1 foci formation is dependent on PTIP. Although the Ser 957 of SMC1 is conserved from human to mouse, the ϩ4 Asp in human has exchanges to Glu in mouse. Thus, we failed to use this anti-human phospho-SMC1 to detect mouse SMC1 by immunostaining. Nevertheless, by using siRNA to knockdown PTIP and 53BP1, we demonstrate that PTIP and 53BP1 regulate the phosphorylation of ATM substrate on the chromatin possibly because of control phospho-ATM accumulation at the DNA damage sites.
PTIP Participates in DNA Damage Repair and Intra-S Phase Checkpoint-53BP1 participates in nonhomologous end joining (NHEJ) following DNA DSBs and other types of NHEJ under physiological condition, such as immunoglobulin class switching recombination during B cell development (48 -50). Because PTIP participates in the DNA damage response and recruits 53BP1 to DSBs, we wonder whether PTIP is involved in NHEJ repair. We measured the efficiency of NHEJ in the PTIPdeficient MEFs by using the comet assay under the neutral condition. As shown in Fig. 5A, the average length of the comet tail FIGURE 3. PTIP and 53BP1 regulate phospho-ATM association with chromatin. A, PTIP and 53BP1 regulate IR-induced phospho-ATM association with chromatin. ptip flox/Ϫ MEFs treated with or without adeno-cre,53BP1 ϩ/ϩ and Ϫ/Ϫ MEFs were irradiated with or without IR (10 Gy). One hour after treatment, the Nonidet P-40 soluble fraction and chromatin fraction were prepared and subjected to Western blot analysis with indicated antibodies. B, PTIP and 53BP1 are dispensable for the IR-induced ATM phosphorylation. ptip flox/Ϫ MEFs treated with or without adeno-cre, 53BP1 ϩ/ϩ and Ϫ/Ϫ MEFs were exposed to 0 or 10 Gy of IR. One hour after treatment, the cells were harvested and analyzed by Western blot using anti-pS1981-ATM and anti-ATM antibodies. C, PTIP and 53BP1 do not affect the ATM kinase activity. ptip flox/Ϫ MEFs treated with or without adeno-cre and 53BP1 Ϫ/Ϫ MEFs were exposed to 10 Gy of IR. One hour after treatment, ATM was immunoprecipitated (IP) using anti-ATM antibody and then incubated with GST-CHK2 at 37°C for 30 min. The ATM kinase activity was determined by a Western blot using anti-pT68-CHK2 antibody. A Coomassie Brilliant Blue (CBB) staining was included to ensure the equal loading of recombinant GST fusion protein.
of PTIP-deficient MEFs was much longer than that in control MEFs in a time course following IR. We also performed a pulsed field gel electrophoresis-based DSB assay after IR to directly measure DNA repair efficiency (51). The fragmentation of chromosomes can be detected as DNA entering the gel and is quantized as a ratio of DNA within the lane relative to the total DNA loaded. We observed that PTIP Ϯ cells displayed progressively less DNA damage in a time course from 0 to 18 h after IR, whereas PTIP Ϫ/Ϫ cells displayed a moderate defect in DNA repair (Fig. 5B). These results suggest that PTIP is involved in NHEJ follow DNA double strand breaks.
Both ATM-deficient cells and cells transiently expressing SMC1 S957A mutant display defects in the intra-S phase checkpoint following DNA damage (44,45,52). We have tested whether PTIP participates in the S phase checkpoint in response to IR. Similar to that in the 53BP1-deficient cells (32) and SMC1 mutant cells (44), DNA damage-induced intra-S phase checkpoint activation was slightly impaired in the PTIP-deficient cells, suggesting that PTIP may play a subtle role in intra-S phase regulation.

DISCUSSION
In summary, our results show that the role of PTIP in the DNA damage response is downstream of ␥H2AX, MDC1, and RNF8 and upstream of 53BP1. Following DNA damage, ␥H2AX, MDC1, and RNF8 form a complex connected by phosphorylation-dependent protein-protein interactions (24 -26, 37-39). This pathway is required for the relocation of 53BP1 to the DNA damage sites (37)(38)(39)(40), which is mediated by PTIP. The relocation of 53BP1 to the DNA damage sites is also dependent on its tudor domain that recognizes methylated histones (53)(54)(55)(56). Interestingly, PTIP associates with a histone methyltransferase complex (30,31). Thus, it is possible that PTIP-dependent histone methylation may recruit 53BP1 to the DNA damage sites. In addition, PTIP foci formation is controlled by RNF8, an important E3 ubiquitin ligase during the DNA damage response (21,(37)(38)(39)(40). However, we neither detect any interactions between RNF8 and PTIP nor find that RNF8 directly ubiquitinates PTIP. Thus, it is likely that RNF8 indirectly regulates PTIP IRIF. Recently, we and others have shown that RNF8 regulates histone ubiquitination following DNA damage (37,39). Because ubiquitination is a bulky modification for the nucleosome, it is possible that histone ubiquitination modifies the chromatin structure and allows PTIP and its associated complex to access the DSBs. The details underlying these mechanisms need to be further examined.
In this study, we have also found that the PTIP-53BP1 pathway regulates ATM association with chromatin and subsequently controls ATM-dependent phosphorylation of SMC1 at the DNA damage sites. Interestingly, ATM is considered as one of the initial kinases to activate DNA damage response pathways and phosphorylates numerous substrates even including H2AX, MDC1, RNF8, and 53BP1 (57). However, we found that phospho-ATM association with chromatin was apparently regulated by PTIP and 53BP1. One possibility is a small amount of ATM is activated by MRE11⅐RAD50⅐NBS1 complex immedi- FIGURE 4. PTIP and 53BP1 regulate ATM-dependent SMC1 phosphorylation at DSBs. A, PTIP and 53BP1 are required for the phosphorylation of SMC1 after IR. U2OS cells were treated with control siRNA or PTIP siRNA or 53BP1 siRNA. siRNA treated cells were irradiated with or without IR (10 Gy). After recovery for one hour, the cell lysates were analyzed by Western blot with indicated antibodies. B, PTIP and 53BP1 regulate the IRIF of phospho-SMC1. U2OS cells treated with control siRNA, PTIP siRNA, or 53BP1 siRNA were exposed to 5 Gy of IR. Four hours after IR, the cells were fixed and immunostained with indicated antibodies. DAPI, 4Ј,6-diamidino-2-phenylindole.
ately follow DNA damage (58 -61). The activated ATM subsequently sparks the signal transduction by phosphorylating other DNA damage mediators including H2AX, MDC1, 53BP1, and RNF8. Once the pathway is activated, PTIP and 53BP1 induce chromatin remodeling at the DNA damage sites, which promotes the association of phospho-ATM with chromatin and induces the phosphorylation of chromatin-bound ATM substrates like SMC1. This process may lead to a positive feedback loop to amplify the DNA damage-induced signal transduction and facilitate DNA damage repair at the chromatin lesions.
We have found that the level of ATM phosphorylation is slightly reduced in the PTIP-and 53BP1-deficient MEFs (Fig.  3), although Ser 1981 phosphorylation of ATM does not represent the ATM activation (34,(62)(63)(64). Consistent with ATM autophosphorylation, we have found that the phosphorylation of other substrates of ATM, including Chk2 and p53, are also slightly impaired in the PTIP-and 53BP1-deficient cells (supplemental Fig. S1). Interestingly, the phosphorylation of SMC1, another ATM substrate that tightly associates with the chromatin, is dramatically reduced in the PTIP-and 53BP1-deficient cells (Fig. 4), indicating that chromatin-bound ATM phosphorylates SMC1. Because the phosphorylation of SMC is involved in the S phase checkpoint activation, we have found that PTIP, like 53BP1 (32), might be also involved in the S phase checkpoint activation. Moreover, the function of ATM retention to chromatin is also likely to participate in DSBs repair. Both 53BP1 and SMC1 have been shown to participate in NHEJ both in vitro and in vivo (47,48,65,66). Here, we examined and found that like 53BP1 and SMC1, PTIP was involved in NHEJ, which further confirmed that PTIP functionally mediates, facilitates, and cooperates with 53BP1-ATM-SMC1 pathway for DNA DSBs repair. Besides NHEJ, homologous recombination (HR) is another major type of DNA DSBs repair. However, like 53BP1, we could not detect the significance of PTIP in HR repair by using an IsceI-dependent reporter assay (data not shown). It is possible that the assay per se is not sensitive for analyzing a subtle function of PTIP in HR because SMC1 may participate in HR repair in yeast through an indirect genetic analysis (47). Nevertheless, our results position PTIP in the DNA damage response pathway and demonstrate the function PTIP following DNA damage response. In addition, PTIP also regulates histone methylation, which could potentially affect chromatin remodeling following DNA damage. Further examination on PTIP-dependent histone methylation may uncover other mechanisms underlying PTIP-dependent DNA damage response.