Regulatory Interactions between the Checkpoint Kinase Chk1 and the Proteins of the DNA-dependent Protein Kinase Complex* □ S

Checkpoints are biochemical pathways that provide cells a mechanism to detect DNA damage and respond by arresting the cell cycle to allow DNA repair. The conserved checkpoint kinase, Chk1, regulates mitotic progression in response to DNA damage by blocking the activation of Cdk1/cyclin B. In this study, we investigate the regulatory interaction between Chk1 and members of the Atm family of kinases and the functional role of the C-terminal non-catalytic domains of Chk1. Chk1 stimulates the kinase activity of DNA-PK (protein kinase) complexes, which leads to increased phosphorylation of p53 on Ser-15 and Ser-37. In addition, Chk1 stimulates DNA-PK-dependent end-joining reactions in vitro . We also show that Chk1 protein complexes bind to single-stranded DNA and DNA ends. These results indicate a connection between components that regulate the checkpoint pathways and DNA-PK complex proteins, which have a role in the repair of double strand breaks. In cells signal transduction checkpoints cell cycle transcription These re-sponses DNA and chromosome are completed fidelity. in checkpoint on a 6% acrylamide gel along with Bst eII-digested (cid:4) DNA as a molecular weight marker. The products were transferred to nitrocellulose and probed with a 32 P-labeled fragment of pDEA-7Z. a , reconstitution of end joining in MO59J extracts. DNA-PK complexes (10 and 150 units) were added to extracts from MO59J, and the joining of Bsa I-cut plasmid was monitored by Southern analyses. b , Chk1 stimulates end-joining reactions. Bacterially produced GST (cid:2) Chk1 was added with or without 50 units of DNA-PK complexes to the reaction to determine the effect on formation of dimers in extracts from M059J cells (5 and 15 (cid:2) g of extract for reactions using cohesive or blunt ends, respectively). The labeled DNA added to the reaction was digested with either Bsa I or Sma I to determine the effect of Chk1 on ligation of cohesive or blunt ends, respectively. c , Chk1 (cid:3) protein lacking the C-terminal domains blocks end joining in vitro . GST (cid:2) Chk1 (cid:3) (20, 50, and 200 ng and 1 (cid:2) g) and GST alone (1 (cid:2) g) were added to the reaction to determine the effect on formation of dimers in extracts from M059K cells. For all experiments the products were transferred to nitrocellulose membrane and hybridized with a 1.1 Kb Rsa I fragment from pDEA-7Z and visualized by autoradiography ( t , trimer; d , dimer; m , monomer).

In response to DNA damage and replication interference, cells activate signal transduction pathways known as checkpoints that prevent cell cycle progression and induce the transcription of genes that facilitate DNA repair (1,2). These responses ensure that DNA replication and chromosome segregation are completed with high fidelity. Defects in checkpoint response can result in genomic instability, cell death, and a predisposition to cancer in higher organisms (3).
The tumor suppressors p53 and Atm and the product of the essential gene ATR are checkpoint genes of primary importance in mammals. The Atm protein is required for several aspects of the response to DNA damage, including cell cycle arrest and the kinetics of p53 activation (4,5). Atm, Atr, and DNA-dependent protein kinase (DNA-PK) 1 are members of a family of proteins that shows homology to phosphatidylinositol-3-kinases. This family includes the checkpoint proteins Mec1 and Tel1 from Saccharomyces cerevisiae, and rad3 from Schizosaccharomyces pombe. Protein kinase activity has been attributed to the members of the Atm family despite their homology to phosphatidylinositol 3-kinase family members (6 -9). The mammalian Atm-like kinases have overlapping substrate specificities (8 -14), yet they appear to control different aspects of the response to DNA damage. The mammalian Atm family member that has been best characterized by biochemical methods is DNA-PK cs (for review, see Ref. 15). DNA-PK cs is the catalytic subunit of a protein complex that plays a role in double strand break repair. DNA-PK is activated by singlestranded DNA (16) and double strand breaks (17). Although the catalytic subunit has been shown to bind DNA on its own, the heterodimeric protein complex Ku70⅐Ku80 acts as the DNA binding partner of DNA-PK cs . Thus, the DNA-PK complex is composed of DNA-PK cs and Ku70⅐Ku80 (15). DNA-PK cs -deficient mice display abnormalities in V(D)J recombination, a process that is essential for the rearrangement of genes to generate functional immunoglobulin and T-cell receptor genes (15). Thus, DNA-PK cs -deficient animals display severe immune deficiencies, and DNA-PK(Ϫ/Ϫ) cells display increased sensitivity to ionizing radiation and double strand break repair defects. This is despite the fact that these cells are proficient in cell cycle arrest following DNA damage and in the up-regulation of DNA damage-inducible genes (reviewed in Ref. 15). These findings suggested that either DNA-PK cs has a redundant role with Atm and Atr in the checkpoint response or the role of DNA-PK cs is specific for repair of the DNA lesions.
Chk1 (checkpoint kinase 1) was first identified in S. pombe because of its essential role for checkpoint arrest at G 2 /M (18,19). In S. pombe and S. cerevisiae, the Chk1 proteins function downstream of the Atm-like proteins to regulate the checkpoint response that prevents chromosome segregation after DNA damage. In mammalian cells and S. pombe, Chk1 blocks mitotic progression via inactivation of the mitotic inducer Cdc25 (20 -22).
A primary integrator of checkpoint and stress signals is the product of the tumor suppressor gene p53. p53 regulates many aspects of the DNA damage response, including cell cycle arrest, the transcriptional response, and apoptosis (reviewed in Ref. 23). Therefore, p53 function is highly regulated by different forms of post-translational modifications that include phosphorylation on multiple regulatory residues (23). Atm, Atr, and DNA-PK phosphorylate p53 on serine residues (Ser-15 by Atm, Atr, and DNA-PK and Ser-37 by DNA-PK and Atr) that have been shown to be phosphorylated in vivo following DNA damage (8 -10, 12, 24). Phosphorylation of Ser-15 and -37 may be important for modulating the stability and activity of p53 (10).
In this study, we investigated 1) the regulatory interaction between Chk1 and the members of the Atm family of kinases and 2) the functional role of the C-terminal non-catalytic domains of Chk1. We show that Chk1 was found in complexes that bind single-stranded DNA and double strand breaks and stimulated the kinase activity of DNA-PK complexes in vitro. Stimulation of DNA-PK resulted in enhanced phosphorylation of p53 on Ser-15 and -37 and required the C-terminal domains of Chk1. Chk1 stimulated DNA end joining in vitro, and Chk1 mutants lacking the C-terminal 115 amino acids acted in a dominant fashion to block DNA-PK kinase activity and DNA repair.

MATERIALS AND METHODS
Cell Culture, Treatment of Cells, and Transfection-HeLa and 293T cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. SF9 cells were grown in Insect Xpress (Biowhittaker) supplemented with 2.5% fetal bovine serum, and Hi5 cells were grown in Excell 405 (JRH Biosciences, Inc.). Cells were treated with 5 mM hydroxyurea (HU) for 24 h, unless otherwise stated. HeLa and 293T cells were transfected using FuGENE 6 transfection reagent (Roche Applied Science) and LipofectAMINE (Invitrogen), respectively, following the manufacturers' instructions.
Cloning, Mutagenesis, DNA, and Protein Purification-The GFP⅐Chk1 and GFP⅐Chk1 KD plasmids were generated by excising Chk1 and Chk1 KD cDNAs from PYS43 and PYS64, respectively, with KpnI and XmaI and ligating them to KpnI-and BspEI-cut pEYFB vector (Clontech). The GST⅐Chk1 C-terminal deletion mutants were generated by excising a BamHI fragment from pYS43 and pYS65 (20) and from the EYFP⅐Chk1 constructs. Site-directed mutagenesis was performed using the QuikChange kit (Stratagene) following the manufacturer's instructions. GST fusion proteins were expressed in Escherichia coli (BL21). Soluble proteins were concentrated using glutathione-agarose beads, or the fusion proteins were expressed in Hi5 cells and collected as previously described (20).
In Vitro Kinase Assays-Twenty units of DNA-PK (Promega) were preincubated with either Chk1 or Chk1 KD (40 -80 pmol) in kinase buffer (see above) containing 200 ng of sonicated calf thymus DNA, 15 Ci of [␥-32 P]ATP, 200 M ATP, 1.3 mM spermidine, and 1 mM DTT. The preincubation step was carried out at 30°C for 15 min. The reactions were placed on ice, and GST⅐p53 substrate was added. The reactions were then incubated at 30°C for 30 min. Kinase reactions were terminated by the addition of SDS-sample buffer with 50 mM DTT as a final concentration and boiled. Proteins were separated by SDS-PAGE on 4 -12% acrylamide gels. Where indicated, the GST⅐Chk1 proteins were pretreated with 10 units of DNase I prior to being used in the kinase reaction.
Western Blotting, Data Quantification, and Analysis-Western blot analysis was performed as previously described (20). Chk1 was detected with antibodies generated against a C-terminal peptide of Chk1 (anti-PEP antibodies; Ref. 20). p53 was detected with DO-1 (Santa Cruz Biotechnology). Phosphorylated Ser-15 and -20 on p53 were detected with phospho-specific antibodies (New England Biolabs). 32 P-labeledproteins were visualized by autoradiography, and their relative densities were quantified using phosphorimaging (Molecular Dynamics) and ImageQuant. The enhanced chemiluminescence signals were quantified from exposures in the linear range using NIH image 1.6.
DNA Binding and Immunoprecipitation Assays-Nuclear extracts and DNA cellulose binding assays were carried out as previously described (31) except that the nuclear extracts were dialyzed against buffer Z (31) containing 1 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, 1 mM DTT, and 1ϫ protein inhibitor mixture (20). 12 g of nuclear extract were incubated with DNA cellulose for 15 min at 30°C in complete kinase buffer in a 50-l volume. After preincubation, p53 was added and incubated an additional 30 min. For the DNA end binding assays, 50 pmol of each 78-mer oligonucleotide (one oligonucleotide was biotinylated) were annealed and incubated with 25 l of streptavidin magnetic beads according to the manufacturer's instructions (Dynal). The DNA end binding assay was carried out by incubating 30 -80 g of nuclear extract with either annealed oligos bound to magnetic beads or magnetic beads alone as control and incubated at room temperature for 45 min-1 h in kinase buffer. ATP was added to the reaction buffer (20 M) where indicated. The beads were washed eight times with kinase buffer and the proteins separated on 4 -12% gradient gels (Invitrogen) and processed for Western analyses as above. The immunoprecipitation assays were carried out with 1 mg/ml nuclear extract in buffer Z containing 20 micromolar ATP and protease inhibitors as above. The extracts were precleared with protein A-Sepharose beads and incubated with 2 g of antibody at 4°C overnight. The proteins were visualized by enhanced chemiluminescence (PerkinElmer Life Sciences) using the following antibodies: Chk1 anti-PEP antibody (20) or Santa Cruz Biotechnology (C19 or G4); DNA-PK cs (Ab1) and (Ab2), Oncogene Research; Ku70 (M19), Santa Cruz Biotechnology.
Immunoprecipitation of FLAG⅐Chk1 from 293T cells was carried out by first enriching for chromatin-associated proteins. 293T cells transfected with either CMV-FLAG⅐Chk1 or pCDNA3.1 were treated or not with 5 mM HU for 18 h, collected and washed with phosphate-buffered saline, and resuspended in solution A (10 mM HEPES, pH 7.9, 10 mM KCl, 1.5 mM MgCl 2 , 0.34 M sucrose, 10% glycerol, 1 mM DTT, 10 mM NaF, 1 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, protease inhibitors). Cytoplasmic proteins were separated from nuclei by lowspeed centrifugation (4,800 rpm for 4 min at 4°C using the St-Micro rotor, Sorvall Super T21). Isolated nuclei were washed with solution A, resuspended in solution A containing 1 mM CaCl 2 and 150 units of micrococcal nuclease (Sigma), and incubated at 4°for 1 h before stopping by adding EGTA 1 mM as final concentration. Sodium chloride was added to 300 mM final concentration and incubated for 10 min on ice. The nuclear preparation was then sonicated twice for 15 s and cleared by centrifugation at 13,200 rpm (same rotor as above) for 15 min at 4°C. The extracts were diluted in Buffer Z and precleared with Protein A-Sepharose followed by incubation with FLAG-affinity resin at 4°C overnight. The proteins were precipitated by gentle centrifugation, washed extensively, and separated on a 4 -12% gradient SDS-PAGE gels.
DNA Repair Assays-Extracts were prepared and used for end-joining assays as previously described (32) with the following modifications. 10 ng of pDEA-7Z digested with BsaI were incubated with 15 g of total cell extract in a 10-l volume for 90 min at 37°C. GST and GST⅐Chk1⌬ proteins were added at approximately equal concentrations at the beginning of the reaction. The reactions were treated with 10 g of Proteinase K for 30 min at 37°C followed by phenol chloroform extraction. The deproteinated DNA products were separated on 0.7% agarose gels containing 1 g/ml ethidium bromide, transferred to nitrocellulose, hybridized with a 32 P-labeled RsaI fragment from pDEA-7Z, and visualized by autoradiography.

DNA-PK and Chk1 Display a Synergistic Interaction in the
Phosphorylation of p53 on Ser-15 and Ser-37-Chk1 proteins from S. pombe and S. cerevisiae function downstream of the Atm homologues rad3 and Mec1 to regulate mitotic progression following DNA damage (20,33,34). Chk1 and DNA-PK phosphorylate p53 on Ser-15 and -37 (30), and the phosphorylation of these residues is increased following DNA damage. To determine whether the Atm-related kinases regulate Chk1, we performed experiments to test whether purified DNA-PK complexes (DNA-PK cs ϩ Ku70 and Ku80) would increase p53 phosphorylation by Chk1 in vitro. We found that preincubation of DNA-PK and Chk1 in the presence of sheared DNA led to a 13-55-fold increase in p53 phosphorylation (with an average induction of 35-fold) (Fig. 1a). The increased p53 phosphorylation was dependent both on Chk1 concentration and the presence of sheared (activated) DNA and was not observed when DNA-PK was preincubated with GST alone. 2 To rule out the possibility that the enhancement seen in our assay was because of contaminating DNA in the Chk1 preparations, kinase assays were carried out with increasing concentrations of Chk1 with and without the addition of sheared DNA. We observed that the dramatic enhanced phosphorylation of p53 increased with Chk1 concentration only in reactions where sheared DNA had been added to the reaction (Fig. 1b, top). To further rule out the possibility that the Chk1 preparation contained some form of DNA that would activate DNA-PK, we performed the kinase assays with Chk1 protein preparations that had been pretreated with DNase I. Treatment with DNase I did not reduce the ability of Chk1 to synergize with the DNA-PK complex ( Fig.  1b, bottom). These experiments indicated that the synergism observed in our assays was not because of contaminating DNA in the Chk1 protein preparations (Fig. 1b).
We used phosphorylation-defective p53 mutated proteins and anti-phospho-specific antibodies to determine the residues phosphorylated by DNA-PK and Chk1. Mutation at Ser-37, a site known to be phosphorylated by both DNA-PK and Chk1, resulted in considerable loss of the enhanced phosphorylation of p53 mediated by DNA-PK and Chk1 (see Fig. 1, c-f). However, phosphorylation of this mutated p53 was still enhanced when both DNA-PK and Chk1 were present (Fig. 1f). These data suggest that DNA-PK and Chk1 show synergistic activity toward more than one site on p53. Phosphorylation of Ser-15 was slightly reduced in the mutated Ser-37 (Fig. 1c, lane 4), and the mutation of Ser-15 caused a mild reduction of 32 P incorporation into GST⅐p53 (1-80). Mutations at both Ser-15 and -37 in GST⅐p53 (1-80) resulted in undetectable phosphorylation by the kinases whether they were present alone or together (Fig. 1e, last lane). These results indicate that DNA-PK and Chk1 act synergistically to phosphorylate p53 on serines 15 and 37, with Ser-37 being the major phospho-acceptor of the enhanced phosphorylation observed in these assays. Recently it has been shown that phosphorylation of Ser-20 and not Ser-15 regulates stability of p53 following DNA damage (27). In our assays, mutation of Ser-20 also decreased the 2 D. M. Goudelock and Y. Sanchez, unpublished data.  5). b, enhancement of DNA-PK was dependent both on Chk1 concentration and the presence of DNA. DNA-PK complexes were incubated with GST⅐p53 (1-80) and increasing amounts of Chk1 in the presence or absence of 200 ng of double-stranded DNA. Bottom, kinase assay with bacterially produced Chk1 proteins that had been pretreated with DNaseI. c, parallel kinase assays were performed with (Top) or without (Bottom) [␥-32 P]ATP to determine phosphorylation levels of wild-type p53 or mutated p53 at Ser-15, -20, or -37 by DNA-PK and/or bacterially produced GST⅐Chk1. The ␥-32 P incorporation was visualized by autoradiography, and the level of Ser-15 phosphorylation was determined by Western analysis with anti-pSer-15 antibodies. The amount of Chk1 added to the reaction was determined by Western analysis using anti-Chk1 antibodies as previously described (20). d, graph depicting the fold induction of phosphorylation of wild-type or mutated p53 (1-80) by DNA-PK and bacterially produced GST⅐Chk1 relative to phosphorylation by DNA-PK alone. e, phosphorylation of wild-type and mutated p53 by DNA-PK and GST⅐Chk1 expressed in insect cells. f, graph comparing the fold induction of phosphorylation levels of the GST⅐p53 (1-80)WT and S37A by DNA-PK and/or GST⅐Chk1 expressed in insect cells. The phosphorimaging values used to plot the data are written beneath each column. enhanced levels of 32 P incorporation and phospho-Ser-15 as measured by antibody staining. This reduction may indicate that Ser-20 is required for Ser-15 or -37 to be phosphorylated, or it may be that a change in the protein at residue 20 affects the adjacent residues and indirectly affects Ser-15 and -37 phosphorylation.
To rule out that the DNA-PK protein fraction (Promega) contained other p53 kinases that could be enhanced by the recombinant Chk1, such as Atr, the proteins in the DNA-PK fraction (20-fold of the amount used in each reaction or 400 units) were analyzed by immunoblotting for the presence of DNA-PK, Atr, Atm, replication protein A (Rpa) 2, and Chk1. Although we observed a strong signal for DNA-PK and Ku80, we did not detect Atm, Atr, or Chk1 signals in the fraction used as a source of DNA-PK/Ku in our assays ( Supplementary Fig.  2). 3 However, the possibility exists that other p53 kinases are present at such low levels that they are below the level of detection by the antibodies used. Atm does not phosphorylate p53 on Ser-37, which is the major phospho-acceptor residue in our enhancement assay, indicating that it is not Atm that is being activated in our assay.
The C-terminal Domain of Chk1 but Not Its Catalytic Activity Is Required for the Stimulation of DNA-PK-To address the direction of signaling between Chk1 and DNA-PK, we found that addition of wortmannin (a phosphatidylinositol kinase inhibitor) during or after the preincubation step resulted in a dramatic reduction in p53 phosphorylation (data not shown), whereas wortmannin had no effect on phosphorylation of p53 by Chk1. This suggested that Chk1 was enhancing the phosphorylation of p53 by DNA-PK. To further investigate the role of Chk1 in this synergistic interaction, we tested a catalytically defective Chk1 protein (Chk1 KD D130A) (20) for its ability to act synergistically with DNA-PK. We found that Chk1 KD had a comparable ability to that of wild-type Chk1 in the stimulation of p53 phosphorylation by DNA-PK (Fig. 2). This effect was observed whether we used the first 80 amino acids of p53 (Fig.  2b) or the full-length p53 protein (Fig. 2a) fused to GST as the substrate. These results indicated that the role of Chk1 is not dependent on catalytic activity and suggested that Chk1 could be exerting a conformational change on the DNA-PK complex. The Chk1 kinases from different species share conserved motifs in their C-terminal domain (Ref. 20). The high conservation between the Chk1 kinases suggests that they may be regulated by common mechanisms and that the C-terminal tail is likely to contain regulatory domain(s). To address whether regions outside the kinase domain were required for the effect on DNA-PK activity, we generated mutants of Chk1 that lacked the Cterminal 115 amino acids in both the catalytically proficient and catalytically defective mutant of Chk1 (Fig. 2, b and c).
Deletion of the C-terminal region of Chk1 abrogated the enhancement of p53 phosphorylation by DNA-PK (Fig. 2c, compare lanes 2 and 4, 3 and 5). In fact, the GST⅐Chk1⌬ mutant, which retained some catalytic activity as measured by phosphorylation of Cdc25C, 4 inhibited the DNA-dependent phosphorylation of p53 by DNA-PK at the concentrations tested (Fig. 2c, compare lanes 1 and 4).
Thus Chk1 stimulates the kinase activity of DNA-PK via its C-terminal domains. Although the role of DNA-PK in p53 phosphorylation in vivo is not clear, our data suggest that high levels of Chk1 or the catalytically inactive Chk1D130A (Chk1 KD ) mutant could stimulate the kinase activity of DNA-PK.
Chk1 Binds DNA and Double Strand Ends and This Inter-action Does Not Require the Catalytic Activity of DNA-PK-Although we did not detect binding of Chk1 purified from bacterial cells to double-stranded oligonucleotides by gel retardation assays under phosphorylation permissive conditions (see above), the regulatory interactions between DNA-PK complexes and Chk1 suggested that Chk1 may interact with these proteins in the presence of DNA. To investigate whether Chk1 may complex with proteins bound to DNA, we examined the ability of Chk1 from nuclear extracts prepared from the glioblastoma cell lines MO59J and MO69K to be precipitated using DNA cellulose beads under phosphorylation-permissive conditions. Chk1 from nuclear extracts bound to the DNA cellulose beads in a DNA-PK cs -and ATM-independent manner (Fig. 3a).
Because DNA cellulose beads contain both double-and singlestranded DNA, we next determined the form of DNA bound by Chk1-containing complexes by performing the binding reactions with biotinylated double-or single-stranded 78-mers bound to magnetic streptavidin beads (Dynal). Chk1 from MO59J (lacking functional DNA-PK cs ) and MO59K (functional DNA-PK cs ) glioblastoma cells was precipitated with both DNA cellulose (data not shown) and DNA ends (Fig. 3b) 2. Chk1 stimulates the kinase activity of the DNA-PK. a, the catalytic activity of Chk1 is not required for the stimulation of DNA-PK. Kinase assays were carried out as described above except that the full-length p53 protein fused to GST was used as a substrate (FL) and the phosphoproteins were visualized by autoradiography or with anti-phosphoserine 15 antibodies. b, the C-terminal domain of Chk1 is required for enhancing the activity of DNA-PK. Kinase assays using bacterially produced full-length Chk1 proteins and deletion mutated proteins were carried out as above. The phosphoproteins were visualized by autoradiography. c, expression of wild-type, catalytically inactive mutant, and the deletion mutants in bacteria was confirmed by Western analyses using anti-Chk1 antibodies raised against the entire Chk1 protein (20). Chk1 to interact with both forms of DNA. We also observed that the interaction of Chk1 with dsDNA cellulose did not require Atm, a checkpoint kinase that has been shown to bind DNA (35). 5 Ku70 also bound the double-stranded oligonucleotides in the absence of functional DNA-PK cs . Chk1 from the same extracts was also precipitated by single-stranded oligonucleotides bound to streptavidin beads (Fig. 3b). Furthermore, the interaction of Chk1 from nuclear extracts derived from MO59J cells with DNA was increased after treatment with the replication inhibitor HU (Fig. 3b, lanes 11 and 12, 13 and 14). These results indicate that Chk1 complexes with proteins that bind strand breaks and single-stranded DNA and suggest a potential mechanism of Chk1 activation and/or possible role in the processing of DNA damage lesions. Although binding of Chk1 to DNA did not require the catalytic activity of the DNA-PK, we cannot rule out the possibility that the DNA binding activity of Chk1 could be mediated by the catalytically inactive DNA-PK cs in MO59J cells and/or enhanced by other proteins involved in DNA strand break or ssDNA recognition or processing, such as the Ku70, Ku86, or Rpa. As previously documented, although DNA-PK and Ku bound ssDNA they preferentially interacted with double-stranded DNA (Fig. 3b) (16,36).

Chk1 and Ku70 Are in the Same Protein Complexes in Vivo-
The interaction between Chk1 and DNA-PK cs could also be detected by pull-down analyses with GST⅐Chk1. GST⅐Chk1 can complex with DNA-PK cs from HeLa cell nuclear extracts prepared from cells treated with HU (Fig. 4a). To determine whether the in vitro interactions we observed between Chk1 and proteins from the DNA-PK complex were occurring in vivo, we analyzed the proteins present in Chk1 and Ku70 complexes by co-immunoprecipitation assays. Nuclear extracts prepared from HeLa cells were incubated under phosphorylation-permissive conditions with antibodies against Chk1 or Ku70 (Fig. 4b).
The complexes were precipitated with protein A-Sepharose beads and subjected to Western analyses using antibodies against Chk1 and Ku70. Ku70 co-precipitated with Chk1 (Fig.  4b, lane 2), and Chk1 was present in the Ku70 immunocomplexes (Fig. 4b, lane 5). Ku80 was also observed in the Chk1 and Ku70 complexes (Fig. 4C and data not shown). To determine whether the interaction between Chk1 and Ku could be The products were separated on a 4 -12% gradient gel, blotted, and the proteins visualized by enhanced chemiluminescence. b, nuclear extracts from the glioblastoma cell lines MO59K (functional DNA-PK catalytic subunit) and MO59J (no functional DNA-PK catalytic subunit) treated or not with 5 mM (HU) were incubated with streptavidin beads bound to double-stranded 78-mer oligonucleotides (ds: lanes 3, 4, 11, 12), singlestranded oligonucleotide (ss: lanes 5,6,13,14) or not bound (lanes 7, 8,15,16). ATP (20 M) was added to the reactions. After washing, the proteins were separated on a 4 -12% gradient gel blotted and processed for Western analyses. The blots were incubated with antibodies to Chk1, Ku70, or DNA-PK cs . The proteins were visualized by enhanced chemiluminescence.

FIG. 4. Co-immunoprecipitation of Chk1 and Ku70 from human cells.
a, the catalytic subunit of the DNA-PK forms a complex with GST⅐Chk1. Whole cell extracts from HeLa cells treated with HU were incubated with GST and GST⅐Chk1 proteins under phosphorylation-permissive conditions. The complexes were precipitated with glutathione beads, washed with buffer containing Tween, separated by SDS-PAGE, and transferred to polyvinylidene difluoride membranes. Proteins were detected with antibodies against Chk1 (G4; Santa Cruz Biotechnology) and DNA-PK cs (Oncogene Research). b, Chk1 and Ku70 were precipitated with antibodies (G4 and C19; Santa Cruz Biotechnology, respectively) from nuclear extracts prepared from HeLa cells treated or not with 5 mM hydroxyurea (lanes 1-3). The complexes were formed under phosphorylation-permissive conditions, separated by SDS-PAGE, and blotted with antibodies against Chk1 (G4; Santa Cruz Biotechnology), Ku70 (M19; Santa Cruz Biotechnology). Ku was precipitated using an antibody raised in goat, and Chk1 in these complexes was visualized by Western analyses using an anti-Chk1 antibody generated in goat. A goat antibody raised against the yeast protein Sic1 was used as control (lane 6); the asterisk denotes the IgG heavy chain. c, Ku-86 co-precipitates with FLAG⅐Chk1. Chk1 complexes were precipitated with antibodies that recognize the FLAG epitope (M2; Sigma) from chromatin-enriched nuclear extracts prepared from 293T cells expressing FLAG⅐Chk1. Vec, immunoprecipitation from extracts prepared from cells transfected with vector. The complexes were separated by SDS-PAGE and subjected to Western analyses with antibodies against Chk1 (FLAG M5; Sigma) and Ku86 (Oncogene Research). enhanced in S phase and/or by a checkpoint signal, we also performed the immunoprecipitation reactions using extracts from HeLa cells that had been treated with 5 mM HU for 24 h. Although we detected phosphorylated Chk1 in the complexes precipitated by the Chk1 antibody, we did not observe an increase in the amount of Ku70 present in the Chk1 complexes (Fig. 4b, compare lanes 2 and 3). Chk1 and Ku have many roles in the cell, which predicts, in turn, that they would have many different partners. Thus, it is not surprising that only a small amount of Chk1 associates with Ku. This is not different from the observation that only a small fraction of those proteins are found in a complex at any given time, despite the abundance of Ku70/80 and DNA-PK cs in the cell. One possible explanation is that the DNA lesion itself stimulates assembly of proteins on DNA breaks, and as such we would not expect to identify a large number of these complexes in the cell. We have recently observed that Chk1 proteins associate with chromatin (37). We carried out immunoprecipitation experiments from chromatinenriched nuclear extracts prepared from 293T cells expressing FLAG⅐Chk1. Using this approach, we observed co-immunoprecipitation of Ku80 with FLAG⅐Chk1, and this association was not changed following treatment with HU (Fig. 4c). Thus, we conclude that Chk1 and Ku70/Ku86 are present in the same protein complex, supporting the regulatory interactions we had observed between these proteins in vitro and suggest a role for this complex in signaling and/or repair.
Chk1 Stimulates DNA-PK-dependent End Joining by Whole Cell Extracts-The regulation of the kinase and DNA binding activities of the DNA-PK complex, which is involved in DNA repair, by Chk1 suggested that Chk1 would also enhance the DNA-PK-mediated end joining of double strand breaks. We next examined whether Chk1 could have an effect on DNA-PKdependent DNA end processing and whether this effect required the C-terminal domains (contained in the last 115 amino acids) of Chk1. For this, we examined the ability of extracts to process 5Ј-overhangs generated by restriction digest of plasmid DNA (pDEA-7Z) with the enzyme BsaI (32). To visualize the DNA fragments, they were transferred to nitrocellulose and hybridized with ␥-32 P-labeled 1.1-kb RsaI probe derived from pDEA-7Z. We detected the appearance of a labeled product of the correct molecular weight for a dimer (32) and higher molecular weight products when the DNA was incubated with cell extracts. The processing of the DNA ends was dependent on the DNA-PK cs because we observed little to no processing when the DNA was incubated with extracts from MO59J cells (Fig. 5a, DNA-PK cs Ϫ). To show that the factor missing in the MO59J required for this reaction was the DNA-PK cs , we reconstituted the processing reactions in the extracts from MO59J cells by addition of purified DNA-PK fraction (Promega) (Fig. 5a). We then chose a concentration of DNA-PK that would restore low levels of end joining to extracts from MO59J cells (50 units) and tested whether Chk1 could stimulate the end-joining activity under those conditions. Addition of Chk1 had a synergistic effect with DNA-PK on end joining of both cohesive and blunt ends (Fig. 5b) when DNA-PK was added to the MO59J extracts, further supporting the interaction between the checkpoint kinase and the proteins regulating DNA end joining. To examine the role of C-terminal domains of Chk1 in stimulating DNA-PK-dependent end joining, we examined whether addition of Chk1⌬ C-terminal to the reaction had any effect on end processing in extracts from MO59K (DNA-PKϩ) cells. Chk1⌬ protein blocked processing of the DNA ends in a dose-dependent manner, whereas addition of molar excess amounts of GST had little effect on the reaction (Fig. 5c). These results support our model derived from previous experiments indicating a functional interaction between Chk1 and the pro-teins in the DNA-PK complex and the importance of the C terminus of Chk1 for this interaction.
Cells with defects in the DNA-PK cs display a sensitive phenotype to ionizing radiation (38,39). We examined the effect of expression of Chk1 or Chk1⌬ proteins on the appearance of cells with abnormal, pignotic, or fragmented nuclei following treatment with ionizing radiation (Fig. 6). We observed an ϳ2-fold increase in the number of cells with pignotic or fragmented DNA 6 h following treatment with 10 Gy of ionizing radiation in HeLa cells expressing Chk1⌬ compared with cells expressing Chk1 or an empty vector. This phenotype could be explained if the cells failed in the checkpoint response, as has been documented for conditional embryonic stem cells after removal of the remaining wild-type Chk1 allele (40). The phenotype could also be because of a defect in DNA repair. When the cells expressing Chk1⌬ were treated with nocodazole after ionizing radiation, we did not observe an increase in the mitotic index of these cells 8 h after radiation. 6 These results suggest that the effect of Chk1⌬ on the radiosensitivity of HeLa cells is not because of a defect in the G 2 /M checkpoint-induced cell cycle arrest.  5. Chk1 stimulates end joining in vitro. pDEA-7Z plasmid cut with BsaI or SmaI was incubated with extracts from M059J (DNA-PK cs Ϫ) and M059K (DNA-PK cs ϩ). Formation of dimers was monitored by separating the deproteinated DNA products on a 6% acrylamide gel along with BsteII-digested DNA as a molecular weight marker. The products were transferred to nitrocellulose and probed with a 32 Plabeled fragment of pDEA-7Z. a, reconstitution of end joining in MO59J extracts. DNA-PK complexes (10 and 150 units) were added to extracts from MO59J, and the joining of BsaI-cut plasmid was monitored by Southern analyses. b, Chk1 stimulates end-joining reactions. Bacterially produced GST⅐Chk1 was added with or without 50 units of DNA-PK complexes to the reaction to determine the effect on formation of dimers in extracts from M059J cells (5 and 15 g of extract for reactions using cohesive or blunt ends, respectively). The labeled DNA added to the reaction was digested with either BsaI or SmaI to determine the effect of Chk1 on ligation of cohesive or blunt ends, respectively. c, Chk1⌬ protein lacking the C-terminal domains blocks end joining in vitro. GST⅐Chk1⌬ (20, 50, and 200 ng and 1 g) and GST alone (1 g) were added to the reaction to determine the effect on formation of dimers in extracts from M059K cells. For all experiments the products were transferred to nitrocellulose membrane and hybridized with a 1.1 Kb RsaI fragment from pDEA-7Z and visualized by autoradiography (t, trimer; d, dimer; m, monomer). cycle progression and induce the transcription of genes that facilitate DNA repair (34,41). In the work described here, using three different approaches we established a link between components that regulate checkpoint pathways and the pathways that mediate DNA repair.
There are two major pathways that regulate the repair of double strand breaks in mammalian cells. The pathway that mediates recombination between homologous sequences is regulated by proteins, including Rad51, -52, -54, -55, and -57 (42). The pathway that mediates the joining of non-homologous sequences (NHEJ) is regulated by the DNA-PK/Ku complex, Xrcc4, DNA ligase IV, Rad50, and nucleases such as Mre11 and Fen1 (reviewed in Refs. 43 and 44). Rad50 and Mre11, which form a complex involved in processing of double strand breaks, have also been implicated in checkpoint signaling in yeast and mammals (42,45,46). Extensive biochemical and genetic studies have implicated DNA-PK in recombination and double strand break repair, yet little is known about the regulation of this complex or about the signaling pathways that dictate the mechanism that the cell uses to repair DNA lesions. There is evidence for a requirement of DNA-PK in lesion repair during G 1 /S and S phases of the cell cycle. Additionally, DNA-PK function is required to stop DNA replication in vitro following UV radiation (47).
We reconstituted DNA damage signaling in vitro with DNA-PK complexes in order to address the regulatory and substrate relationship between Chk1 and this member of the Atm family of kinases. During these studies, we determined that Chk1 activates the DNA binding and kinase activities of the DNA-PK/Ku complex. We show that Chk1 stimulates DNA-PK to phosphorylate p53 on Ser-15 and Ser-37, with Ser-37 being the major phospho-acceptor. Although the role of DNA-PK cs in the regulation of p53 remains controversial (10, 24, 48 -50), the data that are used to argue that DNA-PK does not phosphorylate p53 in vivo are derived from studies carried out with cells lacking DNA-PK cs . The only conclusion that could be reached from those studies was that DNA-PK cs is not required for such signaling to take place. However, those studies do not rule out the possibility that DNA-PK cs participates in signaling to p53. There are three kinases of the ATM family that phosphorylate p53 in vitro: Atm, Atr, and DNA-PK, which could argue for overlapping roles or redundancy. Chk1 also phosphorylates p53 on Ser-15 and -37 (30) and Chk1 and Chk2 phosphorylate Ser-20 (27,29,30). The studies described here used p53 as readout to determine that Chk1 stimulated the kinase activity of DNA-PK in vitro. In addition, we describe an interaction between Chk1 and DNA-PK that results in the phosphorylation of p53 on residues that are not phosphorylated by Atm or Chk2 (Ser-37) and that are postulated to regulate the transcriptional role of p53 or be involved in the protein stability following ultraviolet radiation (51)(52)(53).
Evidence to support the regulation of DNA-PK complexes came from studies that showed that interaction of DNA-PK with Rpa or Ku is regulated by DNA damage during S phase and involves the DNA-PK-dependent phosphorylation of the p34 subunit of Rpa (54). The damage-induced activation of DNA-PK in cells, which leads to the phosphorylation of p34, is abrogated by 7-hydroxystaurosporine (UCN-01), a drug that inactivates the G 2 /M checkpoint and has been shown to inactivate Chk1 but not DNA-PK in vitro (54 -56). Furthermore, the activation of DNA-PK by Chk1 also leads to enhanced phosphorylation of p34. 7 Unlike the catalytic activity, the C-terminal domains of Chk1 are required to enhance kinase activity of DNA-PK. A dominant deletion mutant of Chk1 that fails to activate DNA-PK also blocks DNA-PK-dependent processing of double strand breaks in vitro, corroborating an interaction between the proteins that process the lesions and Chk1. It is of interest to note that deletion of the C-terminal 115 amino acids of Chk1 does not result in increased kinase activity. 8 In this regard our mutant behaves differently from other published C-terminal deletion mutants that display high kinase activity. The reason for the different results could be that, unlike other deletion mutants (30,57), we only removed 115 amino acids from the C terminus. It is possible that our mutated protein contains regulatory domains that block promiscuous interaction of the catalytic domain with all substrates.
Chk1 Complexes at DNA-We speculate that the proteins that mediate Chk1 binding to dsDNA and ssDNA are different. However, the possibility exists that the double-stranded DNA is processed by helicases and nucleases in the extract to generate ssDNA regions that are then bound by proteins that target Chk1 to ssDNA. We detected small amounts of Rpa2, a component of the Rpa single-stranded DNA binding complex in the dsDNA pull-downs, suggesting that this could be the case. 9 There are several mechanisms that could lead to a double strand break during DNA replication that would require immediate checkpoint signaling and repair. Examples include a  6. Expression of CHK1⌬ increases the sensitivity of cells to ionizing radiation. HeLa cells transfected with either EYFP, EYFP⅐CHK1, or EYFP⅐CHK1⌬ were treated with 10 Gy ionizing radiation using a Cesium source. The cells were fixed with 4% paraformaldehyde, stained with 4Ј,6-diamidino 2-phenolindol dyhydrochloride (DAPI), and analyzed by fluorescence microscopy. The percent of abnormal, pignotic, or fragmented nuclei per 250 -300 cells is plotted on the graph. The hatched bar represents the fraction of abnormal nuclei in which expression of the EYFP proteins was detected. The picture on left illustrates the morphology of nuclei in cells transfected with CHK1⌬. single strand nick in the leading or lagging strand template or at a stalled replication fork. We propose that proteins such as the Ku and Rpa complexes are involved in the decision of the cell to repair a lesion that is encountered during DNA replication and that Chk1 could play a regulatory role in this process as previously suggested (16,54). Whether the activation of DNA-PK acts as an amplification signal to increase the signaling to p53 to signal repair or apoptosis (50) and whether this synergism results in increased repair of double strand breaks in vivo remains to be determined.