ATM Activation by Ionizing Radiation Requires BRCA1-associated BAAT1*

ATM (ataxia telangiectasia mutated) is required for the early response to DNA-damaging agents such as ionizing radiation (IR) that induce DNA double-strand breaks. Cells deficient in ATM are extremely sensitive to IR. It has been shown that IR induces immediate phosphorylation of ATM at Ser1981, leading to catalytic activation of the protein. We recently isolated a novel BRCA1-associated protein, BAAT1 (BRCA1-associated protein required for ATM activation-1), by yeast two-hybrid screening and found that BAAT1 also binds to ATM, localizes to double-strand breaks, and is required for Ser1981 phosphorylation of ATM. Small interfering RNA-mediated stable or transient reduction of BAAT1 resulted in decreased phosphorylation of both ATM at Ser1981 and CHK2 at Thr68. Treatment of BAAT1-depleted cells with okadaic acid greatly restored phosphorylation of ATM at Ser1981, suggesting that BAAT1 is involved in the regulation of ATM phosphatase. Protein phosphatase 2A-mediated dephosphorylation of ATM was partially blocked by purified BAAT1 in vitro. Significantly, acute loss of BAAT1 resulted in increased p53, leading to apoptosis. These results demonstrate that DNA damage-induced ATM activation requires a coordinated assembly of BRCA1, BAAT1, and ATM.

Breast cancer is the most common cancer and the second leading cause of cancer mortality in women, with approximately one of nine women being affected in her lifetime (1). Inheritance in breast cancer families follows the classic mendelian pattern of autosomal dominant transmission, with 50% of carriers' children inheriting BRCA1 2 mutations. This inheritance pattern, as well as loss-of-heterozygosity studies in tumors from affected members of BRCA1-linked families, supports the hypothesis that BRCA1 fits the model of a classic tumor suppressor gene, with loss of the normal allele in the tumors of all informative cases (2). Female mutation carriers are estimated to have an 85% lifetime risk of breast cancer (3) and a 40 -50% risk of ovarian cancer (4). BRCA1, first identified as a breast cancer susceptibility gene, encodes a 1863amino acid protein with an N-terminal RING finger domain and a C-terminal acidic domain termed the BRCT domain (5). Mutations in both alleles of BRCA1 greatly increase the risk of breast and ovarian cancers, identifying this gene as a tumor suppressor. Gene disruption experiments in mice result in early embryonic lethality and have therefore provided no information regarding BRCA1 function in adult animals (6 -8). Mice resulting from conditional knockout show immature mammary development and form tumors after long latency with p53 mutation (9). The BRCA1 protein may act at a number of points in nuclear function and growth control (10 -12). For example, the immunofluorescence pattern of BRCA1 dramatically changes from discrete nuclear dots to a dispersed pattern when cells are treated with chemical compounds or IR, implying that BRCA1 is involved in a replication checkpoint after DNA damage (10,(13)(14)(15). More recently, we found that BRCA1 plays a crucial role in activating caspase-3 upon UV damage (16).
An appropriate response to DNA damage is crucial for maintenance of genome stability. Several cellular proteins have been implicated in such processes, such as ATM/ATR protein Ser/Thr kinases, the MRN complex, Fanconi anemia proteins, and the BRCA1 breast cancer tumor-susceptible protein (17)(18)(19)(20). In human and murine cells, ATM is required for early response to agents such as IR that induce DSBs. Despite considerable overlap between the processes regulated by ATM and its relative ATR, cells that lack ATM are extremely sensitive to IR. Thus, ATM plays a unique and essential role in determining survival following IR. In its unstimulated state, ATM is proposed to exist as a homodimer in which the kinase domain of one subunit faces the autophosphorylation of another (21). Upon stimulation, the intermolecular phosphorylation site of the subunits promotes dissociation, and the monomers are free to phosphorylate other substrates (21)(22)(23). Use of the phosphorylation site-specific antibody shows that almost maximal ATM autophosphorylation occurs within minutes of IR, even following very low doses. Indeed, ATM autophosphorylation is readily detectable when as few as two or four DSBs are generated enzymatically by ectopic expression of the restriction enzyme I-SceI (21). Curiously, exposure of cells to mildly hypotonic buffers or to chromatin-modifying drugs, treatments that do not induce DSBs, also leads to rapid and near-maximal autophosphorylation of ATM. Together, these data suggest that a change in chromatin structure, rather than the presence of DSBs per se, is the stimulus to ATM autophosphorylation (21).
To further understand the role of BRCA1 in the DNA damage pathway, we studied a novel BRCA1-interacting protein isolated by yeast two-hybrid screening, which we named BAAT1. Biochemical and biological analyses revealed that BAAT1 is required for ATM activation under conditions of IR damage. Given recent studies of the pivotal roles of the MRN complex in ATM activation (22,23), it is suggested that the ATM kinase activity is tightly regulated by the coordination of cellular proteins.

EXPERIMENTAL PROCEDURES
Identification and Cloning of BAAT1-The CytoTrap vector kit (Stratagene) was used to identify proteins that interact with the BRCT domain. An NcoI/NotI-digested 0.6-kilobase pair DNA fragment of the BRCT domain containing amino acids 1650 -1863 in pcDNA3/BRCA1 was subcloned into the pSOS vector and designated as the bait. The human SOS-BRCT domain fusion construct was coexpressed in yeast strain cdc25H expressing the myristoylated target cDNA library derived from human thymus and allowed to incubate at 37°C. Plasmids from the yeast colonies that grew at 37°C were isolated, and the DNA was sequenced for identification and cloning.
Cell Culture, Transfections, and Western Blotting-NMECs (184B5 cells) and HeLa, U2OS, and 293T cells were grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and 100 units/ml penicillin/streptomycin. SNU251 cells were grown in RPMI 1640 medium containing 10% fetal bovine serum and 100 units/ml penicillin/streptomycin and supplemented with 1 mM sodium pyruvate. For transfections, the calcium phosphate method was used, except when stably expressing BAAT1, small interfering RNA (siRNA) was generated using FuGENE 6 reagent (Roche Applied Science) according to the manufacturer's protocol. Anti-PP2A antibody was provided by Dr. Kazuhiko Yamamoto (Mount Sinai School of Medicine).
Plasmids and Antibodies-IMAGE clone BC015632 in the pOTB7 vector was purchased (ResGen Inc.). GST-BRCA1 constructs were as described previously (11,12) or were generated by PCR. Anti-BRCA1 FIGURE 1. BAAT1 interacts with BRCA1. a, Northernblot analysis of BAAT1 expression in human tissues. PBL, peripheral blood lymphocytes. b, identification of the interacting region between BAAT1 and BRCA1. The indicated regions of GST-BRCA1 and FLAG-BAAT1 were transiently expressed in 293T cells. GST-BRCA1 fragments were precipitated with GSH beads, and samples were immunoblotted with anti-FLAG antibody. To determine the BRCA1-binding region, FLAG-tagged forms of BAAT1 were expressed in 293T cells. After immunoprecipitation with anti-BRCA1 antibody (21A1), samples were immunoblotted with anti-FLAG antibody. GST fusion proteins and FLAG-BAAT1 were immunoblotted with anti-GST and anti-FLAG antibodies, respectively, to confirm their expression. NS, nonspecific signals. c, immunoprecipitation (IP) and immunoblot (IB) analysis showing that endogenous BRCA1 and ATM can immunoprecipitate endogenous BAAT1 from HeLa cells. Two different anti-BRCA1 antibodies (C-20 and 21A1) and anti-ATM antibodies (N-17 and 2C1) were used for the assay. Aff. prep., affinity precipitation.
Co-immunoprecipitation and GST Pull-down Assays-HeLa cells or NMECs were lysed for 30 min in 500 ml of lysis buffer (50 mM Tris-HCl (pH 7.5), 100 mM NaCl, 0.5% Nonidet P-40, 0.5% deoxycholate, 5 mM EDTA, and protease inhibitors (100 mg/ml phenylmethylsulfonyl fluoride, 2 mg/ml aprotinin, and 2 mg/ml leupeptin)). Total cell lysates (1 mg) was incubated with 10 mg/ml of ethidium bromide and the primary antibody for 3 h, followed by overnight incubation with either protein A or G-agarose beads at 4°C. The immunocomplexes were separated on SDS-polyacrylamide gels and visualized by Western blotting. To map the interaction site of BAAT1 and BRCA1, 293T cells were cotransfected with both GST-BRCA1 and FLAG-BAAT1. After pull-down with GSH beads, samples were separated on SDS-polyacrylamide gels and visualized by Western blotting.
Northern Blotting-Full-length BAAT1 and ␤-actin cDNAs were radiolabeled with radioactive [␣-32 P]dCTP (Ambion, Inc.) according to the manufacturer's protocol. A human poly(A) ϩ RNA Northern blot membrane (12 tissues; OriGene Technologies, Inc.) was blotted with the radiolabeled probes according to the manufacturer's protocol.
Purification of Baculoviral BAAT1-His-tagged full-length BAAT1 was produced in Sf9 cells and purified on a nickel column using the protocol described previously (11).
bodies from Jackson ImmunoResearch Laboratories, Inc. (Texas Red-X for mouse IgG and fluorescein isothiocyanate for rabbit IgG) and were used at 1:100 dilution throughout the experiments. The cell nuclei were stained with 4Ј,6-diamidino-2-phenylindole (Sigma). The fluorescence patterns were detected by laser confocal microscopy (Zeiss LSM 510 META confocal microscope), and the data were analyzed and collected using the Zeiss LSM 510 program (Version 3.2). To minimize the possibility of signal cross-talk, a multitrack configuration was used during data collection.
Establishment of siRNA-expressing Cells-The BAAT1 siRNA target sequence 5Ј-AACGGUCACUGAAGGAGA-3Ј (Dharmacon, Inc.) was subcloned into the BglII site of the pSUPER vector (a kind gift from Dr. R. Agami). The control plasmid pSUPER/Luc-siRNA (5Ј-GTTACGCT-GAGTACTTCGA-3Ј) was a kind gift from Dr. A. Schoenfeld (Mount Sinai School of Medicine). About 1 ϫ 10 6 NMECs were transfected with the pBABEpuro vector (provided by Dr. T. Akagi, Osaka Bioscience Institute) at a 1:10 ratio and the pSUPER/Luc-siRNA control or pSU-PER/BAAT1 siRNA. Cell colonies were isolated (2 g/ml puromycin), and the down-regulation of BAAT1 expression was analyzed by Western blotting. For the okadaic acid assay, cells were preincubated with okadaic acid (0.5 mM) for 3 h and treated with IR (5 grays (Gy)). After 30 min, cell lysates were immunoblotted. For transient transfection of U2OS cells, double-strand RNA of the same sequence given above was introduced into the cells with Oligofectamine (Invitrogen). After 48 h, cells were subjected to annexin V staining or immunoblot analysis.
Annexin V/Propidium Iodide Staining and Trypan Blue Staining-Cells were washed with phosphate-buffered saline and harvested by trypsiniza-tion, and the concentration was adjusted to 1 ϫ 10 6 cells/ml. Annexin V was detected using an annexin V-fluorescein isothiocyanate apoptosis detection kit (Oncogene Research Products) as recommended by the manufacturer. For the trypan blue assay, cells were treated with IR, and the surviving cells were counted 24 h after staining them with trypan blue (0.4%, w/v).

Screening and Isolation of BRCT Domain-interacting Proteins-Be-
cause the BRCT domain contains transactivation activity (24), it is difficult to isolate interacting proteins using the conventional yeast twohybrid screen. We used a modified yeast two-hybrid screening (see "Experimental Procedures") with amino acids 1650 -1863 of BRCA1 as the bait. We obtained four cDNA fragments that encode partial sequences of metalloproteinase, ribosomal protein L13, elongation factor G-like protein, and a C-terminal part of the uncharacterized C7ORF27 protein. Previously known BRCT domain-associated proteins such as CtIP (C-terminal binding protein-interacting protein) and BACH1 (BTB and CNC homology-1) were not isolated from this screening.
The isolated C7ORF27 gene encodes an open reading frame of 546 amino acids, which is a C-terminal part of the full-length cDNA registered in the GenBank TM Data Bank (accession number BC015632). The full-length protein consists of 821 amino acids of 100 kDa and does not show significant homology to proteins registered in the data base or motifs such as a nuclear localizing signal. The amino acid sequences of the human and mouse proteins (NCBI accession number BAC39362) share 73% identity. Homologous proteins of other species are available from the Ensembl Database (available at www.ensembl.org). We named this protein BAAT1 on the basis of its possible roles in the BRCA1/ATM pathway (described below). Northern blot analysis using multiple human tissues showed that a 2.9-kb transcript of BAAT1 is ubiquitously expressed, although slightly increased levels of mRNA were detected in testis and pancreas (Fig. 1a).
BAAT1 Associates with BRCA1 and ATM-Interaction of C-terminal BRCA1 and BAAT1 was confirmed using FLAG-tagged BAAT1 and GST-tagged BRCA1 fragments coexpressed in 293T cells. After pulldown of the GST segments using GSH beads, it was found that FLAG-BAAT1 preferentially bound to amino acids 1600 -1749 and 1749 -1863 of the BRCA1 protein (25), consistent with the results of the yeast two-hybrid screen by which BAAT1 was isolated (Fig. 1b). Long exposure of the x-ray film showed that BRCA1 amino acids 1-324 and 502-802 weakly bound to BAAT1. Significantly, a breast cancer-associated mutant form of the BRCT domain, M1775R, did not bind to BAAT1, suggesting that interaction of BRCA1 with BAAT1 is crucial for the tumor-suppressing activity of BRCA1. The C-terminal BAAT1 domain (amino acids 176 -821) showed stronger interaction with BRCA1 compared with full-length BAAT1 in this assay (Fig. 1b, right panel). It has been shown that BRCA1 forms a large complex with cellular proteins termed BASC (BRCA1-associated genome surveillance complex), including ATM (26). Several lines of investigation have also demonstrated that BRCA1 is a substrate of ATM and is involved in the ATM/ ATR pathway (27)(28)(29)(30). Co-immunoprecipitation of BAAT1 with BRCA1 was detected with two different anti-BRCA1 antibodies (C-20 and 21A1), and physical interaction of BAAT1 with ATM was detected when ATM was immunoprecipitated with two different antibodies (N-17 and 2C1) from HeLa cells, suggesting that BAAT1 is involved in the DNA damage pathway regulated by BRCA1 and ATM (Fig. 1c).   APRIL 7, 2006 • VOLUME 281 • NUMBER 14 BAAT1 Is Involved in the IR Damage Pathway-Functional interaction of BAAT1 with ATM was studied in human NMECs after IR treatment. BAAT1 levels were found to increase after DNA damage, and the amount of BAAT1 associated with ATM also increased (Fig. 2a, left  panel). Similarly, increased interaction between BRCA1 and BAAT1 after IR treatment (1 h) was confirmed by co-immunoprecipitation assay (Fig. 2a, right panel). The involvement of ATM in DSB signaling and the localization of its substrate histone ␥-H2AX to DSBs have been thoroughly studied (21,31). Although no significant ␥-H2AX signals had been detected in NMECs before IR treatment, co-localization of BAAT1 and ␥-H2AX was increased 1 h after treatment, demonstrating that BAAT1 is present in DSBs (Fig. 2b). The specificity of anti-BAAT1 antibody for immunocytochemical analysis was confirmed with NMECs transfected with BAAT1 siRNA (Fig. 2b, lower panels; see Fig.  4a). These results demonstrate that BAAT1 is involved in the ATMmediated DNA damage pathway.

ATM Regulation by BAAT1
Co-localization of BAAT1, BRCA1, and ATM-NMECs were immunostained with anti-BRCA1 and anti-BAAT1 antibodies under conditions of IR damage. Before IR treatment, both nuclear and cytoplasmic BAAT1 proteins (green) were detected (Fig. 3a, panels a and c), and co-localization with BRCA1 (red) was not prominent (Fig. 3a, panels a  and g). After IR treatment, however, overlapped signals (yellow) increased in the nucleus (Fig. 3a, panels a and h). Of note, not all of BAAT1 was co-localized with BRCA1 in IR-treated cells. BAAT1 and ATM were further studied by immunocytochemical analysis using SNU251 cells, an ovarian cancer cell line carrying truncated BRCA1. SNU251 cells were infected with control or BRCA1 adenovirus for 48 h. BRCA1 expression was confirmed by immunoblot analysis and immunocytochemistry, which showed the nuclear dot pattern of the protein.
Cells were then treated with IR, and localization of BAAT1 and ATM was determined using specific antibodies after 5 h. Immunocytochemical analysis showed nuclear and cytoplasmic ATM (red) and BAAT1 (green) in the control virus-infected SNU251 cells (Fig. 3c). Although BRCA1 expression induced nuclear localization of ATM and BAAT1, no significant co-localization of both proteins was observed (Fig. 3c,  lower panels). However, co-localization of both proteins (yellow) was increased after IR treatment in BRCA1-expressing cells (Fig. 3c, lower panels). As determined by counting BAAT1 foci detected by immunostaining in Ͼ50 cells, 55, 42, and 44% of BAAT1 co-localized with BRCA1, ␥-H2AX, and ATM, respectively. These results demonstrate that co-localization of BAAT1 and ATM is BRCA1-dependent.
BAAT1 Is Required for ATM Ser 1981 Phosphorylation under Conditions of IR Damage-Recent studies have revealed that phosphorylation of the ATM protein at Ser 1981 is crucial for activation of its catalytic activity induced by DNA damage (21). We explored the role of BAAT1 in the regulation of ATM activation by decreasing the levels of BAAT1 by siRNA under conditions of DNA damage. NMECs were stably transfected with the BAAT1-specific pSUPER construct (32), and several clones were isolated (Fig. 4a, left panel). Cells were treated with IR, and phosphorylation of ATM was studied 0.5, 1, 3, 6, and 12 h after treatment. As shown in Fig. 4b, phosphorylation of Ser 1981 was significantly impaired in BAAT1 knockdown cells after IR treatment. Consistent with this, phosphorylation of CHK2 (a known ATM substrate) at Thr 68 was also decreased in BAAT1 siRNA-expressing cells. Interestingly, depletion of BAAT1 without IR treatment increased p53, which was slightly increased 3 h after treatment (Fig. 4b). Impaired phosphorylation of ATM by IR treatment was also detected in U2OS cells transiently transfected with BAAT1 siRNA (Fig. 4, a, right panel; and b, right panel).
The requirement for BAAT1 in ATM activation was studied in mutant BRCA1-expressing HCC1937 and SNU251 cells (Fig. 4c). After treatment of these cells with IR (3 or 5 Gy), cells lysates were immunoblotted for ATM Ser 1981 phosphorylation. Interestingly, ATM Ser 1981 was phosphorylated in response to IR damage in both cell types. Interestingly, BRCA1 expression in SNU251 cells enhanced ATM phosphorylation by IR (Fig. 4c, right panel). These results demonstrate that, although BRCA1 is not essential for ATM activation, it can enhance its phosphorylation.
BAAT1 Depletion Causes Apoptosis without DNA Damage-Because BAAT1 siRNA-expressing cells showed high levels of p53, we examined whether these cells show any biological phenotype. Annexin V analysis revealed that BAAT1 siRNA-expressing cells showed constitutive apoptosis (23.3%) compared with control cells (7.3%), and IR treatment did not enhance this phenotype (19.4%) (Fig. 5, upper panels). A similar phenotype was observed in U2OS cells transiently transfected with BAAT1 siRNA. These cells showed increased cell death upon depletion of BAAT1 from 4.7 to 12%. These results demonstrate that BAAT1 is involved in the regulation of ATM phosphorylation after DNA damage and that loss of BAAT1 results in activation of the p53-mediated checkpoint, leading to apoptosis.
Phosphorylation of ATM Substrates Is Decreased in BAAT1 Knockdown Cells-Impaired Ser 1981 phosphorylation under conditions of DNA damage was also detected by immunocytochemical analysis. Although control cells showed increased nuclear signals of phosphorylated Ser 1981 (green) after IR treatment (Fig. 6f ), such signals were greatly reduced in BAAT1 siRNA-expressing cells (Fig. 6h). Consistent with these results, phosphorylation of ␥-H2AX (Fig. 6, n and p) and NBS1 (Fig. 6, v and x) was similarly impaired in BAAT1 siRNA-expressing cells after IR treatment. As one possible mechanism of impaired ATM phosphorylation in BAAT1 siRNA-expressing cells, loss of BAAT1 may result in increased activity of ATM phosphatase. Recently, it has been demonstrated that PP2A physically binds to ATM and regulates Ser 1981 phosphorylation (33). In these studies, a phosphatase inhibitor, okadaic acid, was shown to induce autophosphorylation of ATM. On the basis of these results, possible roles of BAAT1 in ATM dephosphorylation were studied using okadaic acid in BAAT1 siRNA-expressing cells. Control and BAAT1 siRNA-expressing cells were treated with okadaic acid before IR treatment. Impaired phosphorylation of ATM Ser 1981 in BAAT1 siRNAexpressing cells was markedly restored after okadaic acid treatment (Fig. 7a, left panels). Interestingly, interaction between ATM and PP2A was not changed in BAAT1 siRNA-expressing cells (Fig. 7a, right panels). We further examined whether BAAT1 is involved in the protection of phosphorylated ATM. Hemagglutinin-tagged BAAT1 was expressed in 293T cells, and cell lysates were treated with PP2A for 4 h at 37°C. As shown in Fig. 7b, expression of hemagglutinin-tagged BAAT1 partially protected ATM Ser 1981 (but not p53 Ser 15 ) phosphorylation induced by IR. The protection of phosphorylated ATM from PP2A was further examined by in vitro phosphatase assay. ATM was immunoprecipitated from NMECs after IR, followed by incubation with PP2A in phospha-tase buffer. Significantly, dephosphorylation of ATM Ser 1981 was partially blocked in the presence of baculovirus-produced full-length BAAT1 (Fig. 7c, lower panel). These results suggest that BAAT1 plays a role in the protection of ATM Ser 1981 phosphorylation from ATM phosphatase under conditions of DNA damage.

DISCUSSION
The regulation of cellular responses to DNA damage is clearly important for the maintenance of genome stability. Many proteins are involved in the early stage of this pathway, including BRCA1, ATM/ ATR, and the MRN complex (26, 27, 34 -36). Recent studies have demonstrated a complicated mechanism of ATM activation: alteration of the chromatin/chromosome structures directly affects its activation (21,37), or the MRN complex stimulates ATM activity by facilitating stable binding of substrates when NBS1 phosphorylation is present (22,23). Thus, the MRN complex was found to be a sensor and effector of ATM activation and signaling in response to DSBs. Complementary results indicate that the NBS1 subunit serves as a bridge between ATM and the DNA-bound MRE11/RAD50 heterodimer (38). Although the original model of ATM activation involved transmission of the DNA damage signal to ATM through structural changes in chromatin, recent FIGURE 7. Protection of ATM Ser 1981 phosphorylation by BAAT1. a: left panels, okadaic acid treatment restored ATM Ser 1981 phosphorylation after IR. Control (cont) or BAAT1 siRNA-expressing NMECs were preincubated with okadaic acid (0.5 mM) for 3 h and treated with IR (5 Gy). Cells were collected after 30 min and subjected to immunoblot analysis with the indicated antibodies. Right panels, the interaction of ATM and PP2A was examined in control and BAAT1 siRNA-expressing cells. ATM Ser1981-P, anti-phospho-Ser 1981 ATM antibody. b: human embryonic kidney 293T cells were transfected with hemagglutinin (HA)-tagged BAAT1 for 48 h. One hour after IR treatment, cells were lysed and incubated with PP2A (protein phosphatase (PPase)). Samples were immunoblotted with the indicated antibodies. p53 Ser15P, anti-phospho-Ser 15 p53 antibody. c: upper panels, in vitro PP2A assay of phosphorylated ATM. NMECs were treated with IR and lysed after 1 h. After immunoprecipitation (IP) of ATM, the optimal concentration of PP2A was determined by in vitro phosphatase assay. Lower panels, similarly, ATM phosphorylation was studied by in vitro PP2A assay using 0.1 unit of PP2A. Purified BAAT1 (100 -800 ng) was simultaneously added to the samples. studies suggest another model in which the MRN complex forms a bridge between ATM and DSB sites and delivers the signal that triggers ATM Ser 1981 autophosphorylation and monomer formation (31).
Of interest, a pathway involving 53BP1 may sense DSBs. 53BP1 was recently identified as a sensor of DSBs through a mechanism involving unstacking of nucleosomes at sites of DNA DSBs that expose a 53BP1binding site on histone H3 (39). The evidence indicates that 53BP1 and the MRN complex activate ATM through distinct pathways (40). Furthermore, 53BP1 is recruited to chromatin in cells treated with hypotonic media (39), but ATM activation does not require NBS1 under such conditions (41). Although our present results support the notion that assembly of BRCA1/ATM/BAAT1 is crucial for ATM activation after IR treatment, it remains to be determined whether BAAT1 is involved in MRN pathways. However, we observed co-localization and co-immunoprecipitation of MRE11 and BAAT1 (data not shown), suggesting that at least a certain fraction of BAAT1 may participate in the MRN pathway.
We found that, in BAAT1 siRNA-expressing cells, ATM phosphorylation is significantly decreased after IR treatment. There are several possibilities to explain this finding. (a) ATM is not recruited to the sites of DSBs; (b) ATM is recruited to the sites of DSBs, but it remains in its dimeric form; and (c) the activity of ATM phosphatase is increased in IR-treated cells. It has been shown recently that PP2A interacts with ATM in undamaged cells and that IR induces phosphorylation-dependent dissociation of PP2A from ATM (33). It was also shown that the PP2A inhibitor okadaic acid induces autophosphorylation of ATM at Ser 1981 in undamaged cells. We found that okadaic acid treatment restored ATM autophosphorylation induced by IR in BAAT1 siRNAexpressing cells. Because okadaic acid treatment does not induce ␥-H2AX foci, it is likely that it induces ATM autophosphorylation by inactivation of a protein phosphatase rather than by inducing DSBs (33). We examined the physical interaction between ATM and PP2A in BAAT1 siRNA-expressing cells; however, no significant changes in their interaction were detected (data not shown). These results suggest that, although BAAT1 is involved in the regulation of ATM phosphatase, a precise role of the protein in this pathway remains to be elucidated.
To understand more complex nuclear events, further investigation should focus on the specificity of BAAT1 function under various stimuli that activate ATM activity. The generation of mice carrying a mutant baat1 locus should provide definitive evidence of the physiological roles of BAAT1 in the DNA damage pathway.