Ionizing Radiation Activates Nuclear Protein Phosphatase-1 by ATM-dependent Dephosphorylation*

Ionizing radiation (IR) is known to activate multiple signaling pathways, resulting in diverse stress responses including apoptosis, cell cycle arrest, and gene induction. IR-activated cell cycle checkpoints are regulated by Ser/Thr phosphorylation, so we tested to see if protein phosphatases were targets of an IR-activated damage-sensing pathway. Jurkat cells were subjected to IR or sham radiation followed by brief32P metabolic labeling. Nuclear extracts were subjected to microcystin affinity chromatography to recover phosphatases, and the proteins were analyzed by two-dimensional gel electrophoresis. Protein sequencing revealed that the microcystin-bound proteins with the greatest reduction in 32P intensity following IR were the α and δ isoforms of protein phosphatase 1 (PP1). Both of these PP1 isoforms contain an Arg-Pro-Ile/Val-Thr-Pro-Pro-Arg sequence near the C terminus, a known site of phosphorylation by Cdc/Cdk kinases, and phosphorylation attenuates phosphatase activity. In wild-type Jurkat cells or ataxia telangiectasia (AT) cells that are stably transfected with full-length ATM kinase, IR resulted in net dephosphorylation of this site in PP1 and produced activation of PP1. However, in AT cells that are deficient in ATM, IR failed to induce dephosphorylation or activation of PP1. IR-induced PP1 activation in the nucleus may be a critical component in an ATM-mediated pathway controlling checkpoint activation.

Ionizing radiation (IR) 1 is known to activate multiple signaling pathways resulting in diverse stress responses including apoptosis, cell cycle arrest, and gene induction (1,2). Many of these responses are controlled by phosphorylation and dephosphorylation of Ser and Thr residues in proteins. PP1 is a major protein Ser/Thr phosphatase conserved among eukaryotic species that regulates a variety of key steps in metabolism, replication, transcription, and the cell cycle (3)(4)(5)(6). PP1 is required for completion of mitosis in many eukaryotic organisms. For example, in Aspergillus nidulans and in Schizosaccharomyces pombe proteins with 80% identity to mammalian PP1 are required for separation of daughter nuclei, completion of anaphase, and chromosome segregation (7,8). The Aspergillus mutant bimG11, which encodes a protein similar to mamma-lian PP1, prevents normal mitotic progression and normal polar growth. Expression of mammalian PP1 fully complements the bimG11 phenotype. Additionally, microinjection of neutralizing PP1 antibody into mammalian cells in early mitosis causes metaphase arrest (9,10).
Although PP1 is required for cell cycle progression, few substrates of PP1 have been identified, and therefore, its role in these processes remains poorly defined. Several lines of evidence suggest that in the absence of DNA damage, Rb is regulated by PP1. Mitotic extracts dephosphorylate Rb in the absence but not the presence of inhibitor 2, a protein specific for PP1 (11). Microinjection of PP1 into G 1 cells results in accumulation of dephosphorylated Rb and inhibition of S phase progression (12). Other potential PP1 substrates that regulate mitotic progression include Cdk1 and lamin B (13), the dephosphorylation of which may be necessary for the formation of the nuclear envelope. Histones H1 and H3 may also be dephosphorylated by PP1 in mitosis (14 -16).
PP1 plays an important role in cell cycle progression and is a potentially important downstream effector of IR-stimulated damage-sensing pathways resulting in checkpoint activation. Here we show that IR causes dephosphorylation of a Thr site in PP1 catalytic subunit (PP1c) resulting in activation of PP1. The IR-induced dephosphorylation of this site is shown to be dependent on ATM (mutated in ataxia telangiectasia), the gene product that is deficient in the human autosomal recessive disease.

MATERIALS AND METHODS
Cell Culture-Jurkat cells (a human T cell lymphoma cell line) were grown in RPMI 1640 medium (Life Technologies, Inc.) with penicillin and streptomycin and 10% fetal bovine serum. FT/pEBS7 and FT/ pEBS7-YZ5 cells were both derived from the AT22IJE-T line (17), an immortalized fibroblast line containing a homozygous frameshift mutation at codon 762 of the ATM gene. AT22IJE-T cells were transfected with the mammalian expression vector pEBS7 (18) containing either the hygromycin resistance marker to yield FT/pEBS7 cells or with full-length ATM open reading frame to yield FT/pEBS7-YZ5 cells. FT/ pEBS7 and FT/pEBS7-YZ5 were generously provided by Y. Shiloh (Tel Aviv University) and grown in Dulbecco's modified Eagle's medium with 15% fetal bovine serum and 100 g/ml hygromycin B. All cells were in an exponential growth phase at the time of radiation.
Radiation Treatment-Cell cultures were irradiated with a Varian linear accelerator at a dose rate of 9 Gy per min. During irradiation, the cultures were maintained in a container designed to mimic the conditions of the cell culture incubator (5% CO 2 and 95% air at 37°C).
Preparation of Nuclear Extracts-Cells were collected by centrifugation in an ice-cold preparation buffer consisting of 20 mM HEPES pH 7.4, 110 mM potassium acetate, 2 mM magnesium acetate, and 5 mM EGTA. The cell pellet was resuspended in the same buffer containing 1 g/ml aprotinin, 1 mM Pefabloc TM , 0.2 mM PMSF, and 1 mM dithiothreitol at 5 ϫ 10 5 cells/ml. Digitonin was added to a final concentration of 50 g/ml to permeabilize the plasma membrane and release the cytosol. The cell suspension was placed on ice for 5 min and then diluted 10-fold in complete preparation buffer. Following centrifugation, the pellet containing intact nuclei was extracted with either nuclear lysis buffer * This study was supported by National Institutes of Health Grants CA 72622 (to J. M. L.) and GM 56362 and CA 40042 (to D. L. B.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Western Analyses-Samples (80 -100 g of protein) were run on 12% SDS-polyacrylamide gels and transferred (with Tris/glycine/methanol buffer, 100 V for 1 h) to nitrocellulose. Reactive proteins were detected with horseradish peroxidase-conjugated antibodies and detected by chemiluminescence. The following commercial antibodies were used: anti-phospho-PP1␣ (Thr-320) antibody (Cell Signaling Technology), anti-PP1c (Santa Cruz Biotechnology, Inc.), and anti-Cdk2 (Upstate Biotechnology). Anti-PP1␣ and anti-PP1␦ were raised and purified against synthetic peptides corresponding to the C-terminal regions of the ␣ and ␦ isoforms, respectively.
Microcystin Affinity Purification-80 l of a 50% slurry of microcystin-agarose (Upstate Biotechnology Inc.) was added to nuclear extract containing 4.0 mg of total protein and incubated for 3 h at 4°C, after which the beads were washed three times with ice-cold nuclear lysis buffer. The protein was then dissociated from the beads with SDS boiling sample buffer and assayed by two-dimensional gel analysis.
Histone H1 Kinase Assays-Cells were collected, washed with cold phosphate-buffered saline, and resuspended in a lysis buffer. The suspensions were kept on ice for 30 min, and the lysate was collected by centrifugation at l5,000 ϫ g for 20 min at 4°C. Protein concentrations were determined using a Bradford assay. Extracts were diluted to 1 mg/ml with lysis buffer. Immunoprecipitation with anti-Cdk2 was performed by incubating 0.5 mg of extract and 4 g of antibody for 1 h at 4°C. The immune complexes were then incubated at 4°C with 20 l of a 50% suspension of protein A-agarose and washed three times first with lysis buffer and then with kinase buffer (l0 mM Tris, pH 7.4, 150 mM NaCl, l0 mM MgCl 2 , and 0.5 mM dithiothreitol) at 4°C. The pellets were incubated for 15 min at 37°C with 40 l of the kinase assay buffer containing 25 M ATP, 2.5 Ci of [ 32 P]ATP, and histone H1 at 1.0 mg/ml. 8 l of 6ϫ electrophoresis sample buffer was added to 40 l of the supernatant and boiled for 5 min and run on a 12.5% SDS-polyacrylamide gel. The gel was fixed and stained with 0.25% Coomassie Blue (in 45% methanol, 10% acetic acid), destained (40% methanol, 10% acetic acid), dried, and exposed to x-ray film. For quantification, the histone H1 bands were excised and 32 P incorporation was determined by liquid scintillation counting. The amount of Cdk2 protein was determined by Western analysis.
Phosphorylation Assay-The purified catalytic subunit of rabbit skeletal muscle PP1c/␣ (19) (1.6 g) was incubated for 30 min at 30°C with varying concentrations of purified Cdk2/cyclin A kinase in a total volume of 40 l of kinase buffer (10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 10 mM MgCl 2 , 0.5 mM dithiothreitol, and 0.1 mM ATP). An aliquot of the reaction mixture was added to 6ϫ sample buffer and heated at 100°C for 4 min and resolved by 12% SDS-PAGE. Phosphorylated PP1c and PP1c were analyzed by Western analysis.
To assay the ability of endogenous Cdk2 to phosphorylate the Thr-320 site in PP1c, Jurkat cells were subjected to 10 Gy, and Cdk2 was immunoprecipitated from 0.5 mg of cellular extract at various times post-IR with 4 g of Cdk2 antibody. The immunoprecipitates were washed three times with ice-cold lysis buffer and three times with ice-cold kinase buffer without ATP and then incubated with purified catalytic subunit of rabbit skeletal muscle PP1c (1.6 g) at 30°C for 45 min in a total volume of 40 l of kinase buffer. Phosphorylated and total PP1c from the supernatant and Cdk2 in the pellet were analyzed by immunoblot analysis.
PP1 Activity Assay-PP1 was immunoprecipitated with 5 g of anti-PP1c monoclonal antibody from either an aliquot of the purified PP1c-Cdk2 reaction mixture (0.5 g of PP1c diluted in 1 ml of PP1 activity lysis buffer (see "Preparation of Nuclear Extracts")) or from nuclear lysis solution (1.0 mg of nuclear protein) prepared from Jurkat, FT/ pEBS7, and FT/pEBS7-YZ5. The immune complexes were then incubated with 30 l of a 50% suspension of protein A-agarose beads washed three times with ice-cold lysis buffer followed by ice-cold Ser/Thr assay buffer (50 mM Tris-HCl, pH 7.0, 0.1 mM CaCl 2 ). PP1 activity was assayed using a Ser/Thr phosphatase assay kit (Upstate Biotechnology Inc.). The PP1 immune complex beads in 50 l of Ser/Thr assay buffer were incubated with the phosphopeptide KRpTIRR at 30°C for 30 min. The beads were pelleted, and 25 l of supernatant was analyzed for free phosphate in the malachite green assay by dilution with 100 l of developing solution (malachite green). After incubation for 15 min, the release of phosphate was quantified by measuring the absorbance at 620 nm in a microtiter plate reader.

IR Induces in Vivo
Dephosphorylation of PP1-We tested if protein Ser/Thr phosphatases were targets of an IR-activated damage-sensing pathway. Jurkat cells were labeled with inorganic 32 P for 45 min after irradiation or sham irradiation. Nuclear extracts were prepared and subjected to microcystin affinity chromatography, followed by two-dimensional gel analysis (Fig. 1). Microcystin binds with nanomolar affinity to the catalytic cleft of protein phosphatases (20) and can be used to rapidly and quantitatively recover the various forms of PP1 and PP2A from extracts (21) together with their multiple regulatory subunits (22). On the two-dimensional gel, more than 50 distinct silver-stained proteins were visualized in the elution from the microcystin beads, and about a dozen of these were 32 P-labeled. The levels of 32 P incorporated into one group of proteins retained by the microcystin matrix dramatically decreased following IR, and we centered our attention on these proteins (Fig. 1). The major form, both by silver staining and 32 P labeling, was the protein in spot 3. However, all of these spots were dephosphorylated in response to IR, based on 32 P labeling. Tandem mass spectrometric sequencing revealed these proteins were the ␣ and ␦ isoforms of PP1 (Table I). The differences in migration in the first dimension (isoelectric focusing) for these PP1 isoforms are probably because of formation of intramolecular disulfides and/or differences in oxidation state of the multiple Cys residues in the catalytic subunits. Regardless, the results show IR caused loss of 32 P indicating that PP1 is phosphorylated in living cells and IR causes net dephosphorylation.
Effects of IR on Nuclear PP1 in Jurkat and AT Cells-PP1 ␣ and ␦ isoforms contain a C-terminal RP(I/V)TPPR motif known to be a preferred site for Cdk/Cdc phosphorylation (23). Phosphorylation of the threonine in this motif occurs in yeast as well as mammalian PP1 (24) and attenuates PP1 activity. We and others (25,26) have shown that IR inhibits Cdk2/cyclin A and Cdk2/cyclin E activity, and both of these kinases phosphorylate PP1␣ at Thr-320. We reasoned that DNA damage might activate PP1 through reduced phosphorylation of this Thr site. Jurkat cells were irradiated and nuclear extracts were assayed for phosphatase activity at various times post-IR (Fig. 2a). There was more than a doubling in PP1 activity over 90 min. In parallel nuclear extracts were subjected to immunoblot analy-FIG. 1. IR dephosphorylates PP1 ␣ and ␦. Asynchronously growing Jurkat cells were subjected to either 0 or 10 Gy. Post-IR, the cells were labeled with [ 32 P]orthophosphate (0.6 mCi/ml) for 45 min after which they were washed with cold phosphate-buffered saline. Extracts were prepared and equal amounts of protein subjected to microcystin affinity chromatography as previously described (38). The pellets were heated with sample buffer and were subjected to two-dimensional gel electrophoresis. The gels were then silver-stained and exposed to films to make autoradiographs. The four proteins circled were sequenced by tandem mass spectroscopy.
sis with a phospho-PP1␣ (Thr-320) antibody. As shown in Fig.  2, a and b, 10 Gy resulted in a time-dependent dephosphorylation of Thr-320 of PP1, which parallels the IR-induced in-crease in PP1 activity. PP1 levels themselves, which serve as a loading control, were unchanged, arguing that there was increased specific activity, not synthesis or import of PP1. FIG. 2. a, IR activates nuclear PP1. Jurkat cells (circles) were subjected to 10 Gy, and nuclear extracts were prepared at various times post-IR. PP1c was then immunoprecipitated, and its activity was measured using the phosphopeptide substrate KRpTIRR as described under "Materials and Methods." Quantitation (squares) of the optical density of the immunoblot shown in b was measured by Image Quant 5.0 (Molecular Dynamics, Sunnyvale, CA). b, IR induces dephosphorylation on Thr-320 of nuclear PP1 catalytic subunit. Asynchronously growing Jurkat cells received either 0 or 10 Gy irradiation, and nuclear extracts were prepared at the indicated times post-IR and subjected to immunoblot analysis with phospho-PP1c (Thr-320) and anti-PP1c.
FIG. 3. Thr-320 phosphorylation influences PP1 activity. Purified catalytic subunit of rabbit skeletal muscle PP1 was incubated for 30 min at 30°C with the indicated concentrations of purified Cdk2 kinase as described under "Materials and Methods." a, an aliquot of the reaction mixture was subjected to immunoblot analysis with phospho-PP1c/␣ Thr-320 antibody and anti-PP1c. Quantitation (squares) of the optical density of the immunoblot is shown in b. b, PP1c was immunoprecipitated from the remaining reaction mixture and its activity was measured as described for Fig. 2a. To verify that Thr-320 phosphorylation regulates PP1 activity under our assay conditions, purified skeletal muscle PP1 protein was incubated with purified Cdk2 kinase and subjected to immunoprecipitation with anti-PP1c. PP1 activity in the immunoprecipitates was assayed with a phosphopeptide as substrate. As expected, increasing doses of Cdk2 resulted in increased phosphorylation of the Thr-320 site (Fig. 3a). Phosphorylation of this site in PP1 corresponded to decreased PP1 activity (Fig. 3b). Thus, in vitro and in living cells the level of Thr-320 phosphorylation inversely correlated with the specific activity of PP1.
To establish that IR inhibited the activity of the endogenous Cdk2 kinase that was phosphorylating PP1, Cdk2 was immunoprecipitated at various intervals after 10 Gy. As shown in Fig. 4, the activity of endogenous Cdk2 toward PP1 at the Thr-320 site decreased in a time-dependent manner following IR.
IR Both Decreases Cdk2 Kinase and Increases PP1 Phosphatase Activity in an ATM-dependent Manner-We previously found that IR-induced dephosphorylation of histone H1 is ATMdependent (14). This is because of reduced Cdk2 kinase and increased nuclear phosphatase activity. IR (10 Gy) inhibits Cdk2 activity in an ATM-dependent manner as shown in Fig.  5a. Cdk2 from AT (PEBS) cells lacking active ATM kinase was inhibited only 40% whereas Cdk2 from reconstituted AT cells (YZ-5) that express ATM was inhibited by 84%. The radiationinduced dephosphorylation of PP1 Thr-320 was ATM-dependent. We compared AT cells (PEBS) and AT cells transfected with full-length ATM (YZ-5) that were irradiated with a dose of 10 Gy. Nuclear extracts were subjected to immunoblot analysis that showed IR failed to dephosphorylate PP1 Thr-320 in AT cells (Fig. 5b). However, in AT cells transfected with fulllength ATM, IR-induced dephosphorylation of Thr-320 occurred in a time-dependent manner starting ϳ15 min postirradiation. This genetic system was used to demonstrate that IR-induced dephosphorylation of the Thr-320 site of PP1 is dependent on ATM.
The ATM kinase is also necessary for the IR-induced increase in PP1 activity. Nuclear extracts were prepared from AT cells deficient in ATM and AT cells expressing ectopic fulllength ATM. Samples were subjected to immunoprecipitation with anti-PP1 and assayed for PP1 activity. As shown in Fig.  5c, the IR-induced time-dependent 2-fold increase in PP1 activity was only observed in cells expressing ATM. Interestingly, the time course of IR activation of PP1 parallels that of IR-induced H1 dephosphorylation, which we have previously reported (14).
Three genes code for four distinct isoforms of PP1 in mammals called ␣, ␥1, ␥2, and ␦ (3). Using isoform-specific antibodies to immunoprecipitate PP1c ␣ and ␦ from cell lysates, we tested if the Thr-320 site was dephosphorylated in response to IR in Jurkat, AT, and reconstituted AT cells. The numbering of the Thr-320 is not identical in ␣ and ␦ but lies in the corre-sponding sequence motif in both these PP1 isoforms. Consistent with the metabolic 32 P-labeling results (Fig. 1), both the ␣ and ␦ PP1 isoforms were dephosphorylated at the Thr-320 site in response to IR (Fig. 6a). Furthermore, the dephosphorylation of both the ␣ and ␦ isoforms was ATM-dependent, as FIG. 4. IR decreases the ability of endogenous Cdk2 to phosphorylate the Thr-320 site in PP1c. Jurkat cells were subjected to 10 Gy, and Cdk2 was immunoprecipitated from nuclear extracts at various times post-IR. The pellet was incubated with purified PP1c at 30°C for 45 min as described under "Materials and Methods." The supernatant was subjected to immunoblot analysis with phospho-PP1c/␣ (Thr-320) antibody and anti-PP1c.
FIG. 5. IR-induced nuclear PP1c dephosphorylation and activation are ATM-dependent. a, in vitro H1 kinase activity was determined on immunoprecipitates with antibodies to Cdk2 at various times following 10 Gy. All experiments were performed 3 times with similar results, and a representative experiment is shown. Equal loading of H1 and Cdk2 is shown as controls. Percentage inhibition of kinase activity was determined by excising H1 bands from the gel and measuring 32 P incorporation by liquid scintillation counting. b, AT fibroblasts transfected with either empty vector FT/pEBS7 (PEBS) or recombinant wild type ATM FT/pEBS7-YZ5 (YZ-5) were irradiated with 10 Gy, and nuclear extracts were subjected to immunoblot analysis with either a phospho-PP1c (Thr-320) antibody or anti-PP1c. c, AT fibroblasts PEBS and YZ-5 were irradiated with 10 Gy, and PP1c was immunoprecipitated from nuclear extracts and assayed for PP1 activity as described for Fig. 2a. shown in Fig. 6b. Decreased phosphospecific staining of PP1 ␣ and ␦ was only seen in irradiated cells that were expressing ATM (YZ-5 clone). Cells that did not express ATM (PEBS) did not show diminished phospho-Thr-320 staining in response to IR. DISCUSSION IR is known to activate the G 1 , S, and G 2 checkpoints (2,27) in an ATM-dependent manner. The ATM gene is mutated in the autosomal recessive disease ataxia telangiectasia, characterized by neuronal degeneration resulting in ataxia, oculocutaneous telangiectasia, immune dysfunction, and cancer predisposition (28). ATM is thought to be an upstream sensor of DNA damage as well as oxidative stress. ATM transmits the damage signal downstream through its C-terminal phosphoinositol 3-kinase domain. ATM has been shown to directly phosphorylate several proteins and to enhance the phosphorylation of other proteins by activating downstream protein kinases. These targets include the nuclear tyrosine kinase c-Abl, Chk2, nibrin, p53, and BRCA1, all of which have been implicated in DNA damage responses (29 -33).
Interestingly, ATM has also been shown to control the damage-induced dephosphorylation of Ser-376 in p53 (34) and to regulate H1 dephosphorylation following IR. Thus, ATM activates the phosphatase(s) responsible for dephosphorylating H1 and p53. We have recently demonstrated that IR causes dissociation of the B55␣ regulatory subunit, which has been implicated in mitotic progression, from heterotrimeric nuclear PP2A in an ATM-dependent manner (38). Although the functional significance of IR-induced dissociation of B55␣ from PP2A heterotrimer is unknown, it is likely that IR-induced subunit exchange allows PP2A to execute diverse functions in the damage response. PP1 and PP2A are both downstream effectors of an IR-activated and ATM-dependent signaling pathway. It seems likely that IR-induced PP1 activation as well as PP2A subunit exchange requires the function of the kinase domain of ATM, but this remains to be determined.
We imagine two potential but not mutually exclusive mechanisms by which ATM could regulate PP1 activity. First, ATM is known to inhibit Cdc2 and Cdk2 in response to IR. The PP1 Thr-320 site is phosphorylated by Cdc2/Cdk2. Thus, IR-activated ATM could increase PP1 activity indirectly through inactivation of Cdc2/Cdk2 kinase, thereby decreasing the phosphorylation of Thr-320. Because ATM has recently been shown to regulate Cdc25A (the activator of Cdk2) through Chk2, it is possible that ATM regulates PP1 through the Chk2-Cdc25A pathway. If this were the case, then a dual specificity phospha-tase (Cdc25) would be regulating another phosphatase (PP1) via a phosphatase-kinase-phosphatase cascade (Cdc25-Cdk2-PP1). Our data do not establish that Cdk2 is the in vivo kinase of PP1. There may be other kinases that phosphorylate the Thr-320 site in PP1, but they would also need to be regulated in response to IR by an ATM-dependent pathway. Alternatively, ATM could directly (or indirectly) phosphorylate a nuclear PP1 subunit that restrains PP1 in its phosphorylated, low activity form. Phosphorylation would release inhibition and allow autodephosphorylation of Thr-320 and activation of PP1 toward other substrates. NIPP is one example of a nuclear PP1 regulatory subunit whose activity is regulated by phosphorylation (35,36).
What are the downstream nuclear substrates of PP1 that may be critical in the damage response? Unlike protein serine/ threonine kinases, PP1 catalytic subunit does not manifest a high degree of sequence specificity and dephosphorylates multiple substrates. The substrate specificity of PP1 is thought to be modulated through the formation of heterodimeric complexes with regulatory subunits. Regulatory proteins that target PP1 to its substrates in response to DNA damage are not known, and these may bind both ␣ and ␦ isoforms or there may be individual regulatory subunits for these isoforms. Specific PP1 isoforms have been shown to have distinct functions and locations. For example, the PP1␣ isoform has been shown to regulate the G 1 /S transition by dephosphorylating Rb (12), and the PP1␦ isoform is chromatin-associated. Because histone H1 and histone H3 have been implicated as PP1 substrates (14,16,37), it is likely that the PP1␦ isoform may regulate IR-induced H1 and H3 dephosphorylation.
In summary, the PP1 ␣ and ␦ isoforms are dynamically phosphorylated in Jurkat cells and in response to IR become activated by dephosphorylation of Thr-320 to function as a downstream effector of an ATM-dependent damage-sensing pathway. Regardless of the mechanism by which ATM activates PP1, defining the regulatory subunits that target PP1 to its respective substrates may reveal novel targets for chemo and radiation sensitizers. Drugs that function at the level of these targets may have low toxicity and, therefore, be more efficacious than less specific inhibitors.
FIG. 6. IR-induced dephosphorylation of PP1 ␣ and PP1 ␦ is ATM-dependent. Jurkat (a) and AT fibroblasts PEBS and YZ-5 (b) were irradiated with 10 Gy. Nuclear extracts were prepared and subjected to immunoprecipitation with the indicated isoform-specific anti-PP1c antibodies. Immunoblot analysis was performed as described for Fig. 2. IP, immunoprecipitate.