Specific Role of Chk1 Phosphorylations in Cell Survival and Checkpoint Activation^{▿ †}

Establishment of cells expressing mutants of Chk1.

Twenty micrograms of pCAGGS-Chk1 (wild type) and its mutants (S317A, S345A, S317AS345A, S280A, S280E, S301A, S357A/S366A, S317E, and S345E) and 5 μg of PGK-Hyg vector were linearized with SalI and KpnI, respectively. The linearized DNA was electroporated into Chk1^lox/− embryonic stem (ES) cells with a Gene Pulser II (27). Cells were selected with 0.15 mg/ml hygromycin B for 7 days. Single colonies were screened by reverse transcription-PCR with a set of primers: Chk1sc-05 (5′-CAGCCCCTCATACATTGATAAATT-3′) and Myc-04 (5′-TGCATATTCAGATCCTCTTCTGAG-3′).

Antibodies.

The antibodies used were as follows: anti-Chk1 (sc-8408 [Santa Cruz] for Western blotting and C9358 [Sigma] for immunohistochemistry), anti-phospho-Ser345-Chk1 (no. 2341; Cell Signaling), anti-phospho-Ser317-Chk1 (ab2834-50; abcam), anti-phospho-Ser10 histone H3 (06-570; Upstate), anti-Chk2 (sc-9064; Santa Cruz), anti-p53 (NCL-p53-505; Novocasta Laboratory), anti-IKKα (sc-7182; Santa Cruz), anti-Orc2 (sc-13238; Santa Cruz), anti-Nek2 (sc-33167; Santa Cruz), anti-lamin B (33-2000; Zymed laboratories), anti-pericentrin (PRB-432C [Covance] and 611815 [BD Transduction Laboratory]), anti-phospho-Tyr15-Cdk1 (219440; Calbiochem), and anti-β-actin (ab2676-100; abcam).

G₂ checkpoint assay.

The G₂ checkpoint assay was assessed using phospho-histone H3 staining as described previously (41). In brief, cells were harvested, washed with phosphate-buffered saline (PBS), and fixed in 70% cold ethanol overnight. The fixed cells were washed twice with PBS, permeabilized in 0.25% Triton-PBS on ice for 10 min, and then washed with PBS and suspended in 100 μl of α-phospho-histone H3 antibody in 1% bovine serum albumin-PBS for 3 h at room temperature. The resultant cells were washed with PBS and incubated in α-rabbit immunoglobulin G fluorescein isothiocyanate-conjugated antibody diluted 1:100 with 1% bovine serum albumin-PBS for 30 min at room temperature. The immunostained cells were then washed twice with PBS and stained in 50 μg/ml propidium iodide-100 μg/ml RNase A solution for 1 h and subjected to fluorescence-activated cell sorter analysis.

Radio-resistant DNA synthesis.

Chk1-disrupted ES cells (1 × 10⁶) were cultured with ES medium containing 10 nCi/ml [¹⁴C]thymidine (Amersham) for 24 h, washed with PBS, and cultured again for 30 min in ES medium. The cells were then X-irradiated (10 Gy), cultured for 1 h, and then pulse-labeled with ES medium containing 2.5 μCi/ml [³H]thymidine for 30 min. The labeled cells were harvested and fixed in ice-cold 70% ethanol overnight. The fixed cells were applied to membranes on the filtration plate, and the membranes were washed with 70% ethanol and then 95% ethanol (Nalgen Nunc; no. 255984). Radioactivity on the membranes was determined with a liquid scintillation counter. Radio-resistant DNA synthesis (RDS) was determined by the radioactivity of ³H divided by that of ¹⁴C.

For experiments using cells expressing Chk1 transiently, 3 × 10⁶ Chk1^Δ/− ES cells were transfected with 9 μg of pCAGGS-Chk1-PACT or pCAGGS-Chk1, cultured for an additional 12 h, and then analyzed for RDS assay.

Subcellular fractionation.

Subcellular fractionation of mammalian cells was performed according to a previously reported method (25, 35). In brief, 6 × 10⁶ ES cells were suspended in 400 μl of solution A (10 mM HEPES 7.9, 10 mM KCl, 1.5 mM MgCl₂, 0.34 M sucrose, 10% glycerol, 1 mM dithiothreitol, protease and phosphatase inhibitors), and Triton X-100 was added to a final concentration of 0.1%. The cells were incubated on ice for 5 min, and cytoplasmic (S1) and nuclear fractions were harvested by centrifugation at 1,300 × g for 4 min. The isolated nuclei were then washed in solution A, lysed in 400 μl solution B (3 mM EDTA, 0.2 mM EGTA, 1 mM DTT, protease and phosphatase inhibitors), and incubated on ice for 10 min. The soluble nuclear (S2) and chromatin fractions were harvested by centrifugation at 1,700 × g for 4 min. The isolated chromatin (P2) was then washed in solution B, centrifuged at 10,000 × g for 1 min, and resuspended in 400 μl Laemmle sample buffer. The P2 fraction was sonicated four times for 30 s each time using a Biorupter.

Immunostaining in situ.

Cells on glass slides were fixed in −20°C methanol-acetone (1:1) for 7 min. Immunofluorescence microscopy was performed with the combinations of antibodies specified in the figure legends as described previously (18).

Transfection and TUNEL staining.

Chk1^lox/− ES cells were infected with adenoviruses expressing Cre recombinase to disrupt the endogenous Chk1 allele. Cells (1 × 10⁶) deficient in endogenous Chk1 were cultured for 24 h. The cells were then cotransfected with 1 μg of pME18S-GFP and 3 μg of pCAGGS-Chk1-PACT or pCAGGS-Chk1 using Lipofectamine 2000 Reagent (Invitrogen). Twenty-four hours after transfection, cells were harvested and fixed in 2% paraformaldehyde for 30 min at room temperature. Terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling (TUNEL) staining was performed with an In Situ Cell Death Detection Kit, TMR red (Roche Diagnostics). TUNEL-positive, GFP-positive cells were counted under a microscope.

Establishment of Chk1^lox/− ES cells that exogenously express Chk1 phosphorylation site mutants.

In mammalian cells, Chk1 is known to be phosphorylated at multiple sites by PIKK and other kinases in response to DNA damage or a DNA replication block (Fig. 1A). However, it is not known whether these phosphorylations are essential for Chk1 functions or what effect they have on cell cycle arrest or cell survival. To address these questions, we performed knockout-knockin experiments using Chk1^lox/− ES cells. To avoid any effect arising from a difference in the amounts of mutant proteins expressed and clonal effects, we selectively used cell lines that express the mutant proteins at a level similar to that of endogenous Chk1 protein (Fig. 1B). To confirm whether wild-type or mutant Chk1 proteins exogenously expressed in ES cells really have kinase activity, we examined this activity in vitro using immunoprecipitates (lysates from clone no. 1) obtained with anti-Myc antibody and a Cdc25C fragment as a substrate and found that these mutants exhibited kinase activity similar to that of wild-type Chk1 (Fig. 1C).

Cells expressing S317A and S345A mutants were defective in cell cycle checkpoints.

In the previous study, we found that Chk1 protein was not detected and cell cycle checkpoints were completely defective at 30 h after infection with adenoviruses expressing Cre, as described previously (27). Therefore, we examined the abilities of wild-type or mutant Chk1 proteins using cell lines of clone no. 1 to reverse the loss of cell cycle checkpoint functions in response to several genotoxic stresses at this time point. With respect to the G₂/M checkpoint, prevention of mitotic entry in response to infrared radiation was almost completely reversed by wild-type Chk1, S280A, and S301A, but not by the mutants S317A, S345A, and S317A/S345A (Fig. 2, left panel). Similar results were observed when cells were treated with hydroxyurea (HU) (Fig. 2, middle panel). With respect to the intra-S phase checkpoint, as with the G₂/M checkpoint, Chk1-dependent RDS was not restored by S317A, S345A, or S317A/S345A mutants (Fig. 2, right panel). Similar results were also observed when clone no. 2 cell lines were used (data not shown). Taken together, these results suggested that phosphorylation of Chk1 at S317 and S345 may be indispensable for cell cycle checkpoint functions. Although S357 and S366 are also conserved phosphorylation sites targeted by PIKK, a previous report demonstrated that these sites were not likely to be phosphorylated in vivo in the event of a DNA replication block (43). Consistent with this, the S357A/S366A double mutant reversed the cell death observed in Chk1-depleted cells (data not shown). S317E and S345E did not restore checkpoint functions (data not shown), suggesting that replacement of the Ser residue with Glu did not mimic the phosphorylation status in this case, as suggested previously (4, 11).

Mutation at S317, but not at S345, did not cause mitotic catastrophe.

Chk1 is a multifunctional kinase that regulates cell cycle progression and cell cycle checkpoints. Accordingly, we next asked whether mutants of Chk1 could prevent the cell death observed in Chk1-depleted ES cells. We found a significant increase in the population of the sub-G₁ fraction in Chk1-depleted cells (Fig. 3A, none), and a similar increase was observed in cells expressing S345A and S317A/S345A mutants. In contrast, cells expressing S317A, as well as the wild type, S357A/S366A, and S301A, did not undergo cell death (Fig. 3A). Consistent with this, premature mitosis measured by the counting of phospho-histone H3-positive cells with less than a 4 N DNA content was significantly increased in cells expressing S345A mutants, whereas this increase was not detected at all in cells expressing the wild type or the S317A mutant (Fig. 3B). These results suggested that phosphorylation of S317 was dispensable for the prevention of premature mitosis.

Previously, we demonstrated that apoptosis induced by Chk1 depletion was a mitotic catastrophe that occurred via the ATM/ATR-p53-dependent DNA damage checkpoint (27). Therefore, we determined whether apoptosis in cells expressing S345A was really due to mitotic catastrophe through the activation of the DNA damage checkpoint. A significant increase in the amount of p53 protein and modification of Chk2 were readily detected in cells expressing S345A (Fig. 3C), suggesting activation of the Chk2-dependent DNA damage checkpoint pathway. These changes were not detected in cells expressing the wild type or the S317A mutant. The knockout of the endogenous Chk1 was equally efficient in wild-type, S317A, and S345A cell lines. Thus, phosphorylation of S317 appeared to be dispensable for prevention of mitotic catastrophe, but phosphorylation of S345 may be required. Because the status of p53 activity in ES cells is contentious, we examined whether Chk1 depletion caused the induction of Noxa and Puma expression, which was known to be p53 dependent. A significant induction of Noxa and Puma expression levels was detected in Chk1-deficient ES cells (see Fig. S1 in the supplemental material), suggesting that p53 was active in our ES cells.

Increased chromosomal abnormalities in cells expressing S317A.

Given that cells expressing S317A were defective in cell cycle checkpoint function but capable of reversing a nonviable phenotype, we then asked whether Chk1-dependent checkpoint signaling is of crucial importance for the maintenance of chromosomal integrity, which protects cells from tumorigenesis. To address this question, we performed cytogenetic studies with S345A cells or S317A cells. The metaphase spreads 36 h after Chk1 depletion from different genotypes were analyzed for chromosome alterations, such as dicentric chromosomes, fragments (acentric and centric), and breaks (chromosome and chromatid). Chromosome numbers were also calculated for each sample. As many as 150 metaphases per genotype were analyzed in this study. In cells expressing wild-type Chk1, 3.7% of the cells showed chromosomal aberrations. Surprisingly, although S345A cells showed defects in phenotypes at cell cycle checkpoints, only 7.6% of metaphase chromosomes displayed structural abnormalities, presumably due to elimination of abnormal cells by mitotic catastrophe. S317A cells (15.1%) were more prone to chromosomal abnormalities than were S345A cells. In addition, 25% of S317A cells with abnormal chromosomes exhibited tetraploidy, which was not detected at all in control cells or S345A cells (Fig. 3D). These chromosomal abnormalities were not detected in cells expressing endogenous Chk1 (data not shown).

Phosphorylation of Chk1 during normal S phase and subcellular localization of its phosphorylated forms.

Our results suggested that phosphorylation of Chk1 at S345 is a prerequisite for prevention of premature mitosis during normal S phase. We next examined whether this site is really phosphorylated in vivo under unperturbed conditions. ES cells were synchronized at the G₁/S boundary by a nocodazole/mimosine treatment and then released into S phase. Fluorescence-activated cell sorter analysis revealed that S phase started at 1 h and was completed 9 h after release (Fig. 4A, left). Immunoblotting using antibodies specific to phospho-Chk1 that is phosphorylated at S317 and S345 revealed that phosphorylation at S317 was detected at late G₁ to late S phase. In contrast, Chk1 phosphorylated at S345 was not detected at late G₁ phase but peaked at early to middle S phase and then decreased at late S phase. Similar patterns of Chk1 phosphorylation were also observed in cells synchronized at M phase by nocodazole treatment alone, although synchronization appeared less effective, eliminating the possibility that Chk1 phosphorylations were due to DNA damage induced by mimosine treatment (see Fig. S2 in the supplemental material). The bands of Chk1 phosphorylated at S317 and at S345 were enhanced by HU treatment and almost completely disappeared after treatment with phosphatase, confirming the specificities of these phosphospecific antibodies (Fig. 4A, right).

To investigate the subcellular localization of phosphorylated Chk1, we fractionated extracts of synchronized ES cells by a centrifugation-based method and obtained fractions of cytoplasmic proteins (S1) and soluble nuclear proteins (S2) and a chromatin-enriched fraction (P2). Immunoblotting using anti-Orc2, anti-lamin B, anti-Nek2, and anti-IKKα antibodies revealed predominant localization of Orc2 in P2, lamin B in S2, and Nek2 and IKKα proteins in the S1 fraction, indicating that fractionation of the extracts was successful (5, 15). Chk1 was detected in all fractions at almost the same level, although it was decreased in the S1 fraction at late S phase. The Chk1 phosphorylated at S317 predominantly appeared in the S1 fraction at early S phase, but it decreased during late S phase. Inversely, it was detected in the S2 fraction at late S phase. Interestingly, Chk1 phosphorylated at S345 was detected only in the S1 fraction. Taken together, these results suggested that phosphorylation of Chk1 at S345, but not at S317, might be important for cytoplasmic localization of Chk1 during S phase.

Very recently, Smits et al. reported that PIKK-dependent phosphorylation of Chk1 leads to its release from chromatin, which is important for transmitting the DNA damage signal (35). Therefore, we next determined which Chk1 phosphorylation site was necessary for chromatin dissociation upon DNA damage. Exogenously expressed wild-type Chk1 effectively dissociated from the chromatin fraction in response to DNA damage induced by UV (Fig. 4C). This dissociation was not observed in S317A cells but was obvious in S345A cells. S345A proteins were not present in the S1 fraction in the presence or absence of DNA damage, further supporting the idea that phosphorylation of Chk1 at S345 is important for its cytoplasmic localization. Consistent with this, a mobility shift in Chk1 upon DNA damage was detected only in the cytoplasmic fraction. Taken together with the fact that Chk1 that was phosphorylated at S317 was not detected in the chromatin fraction (Fig. 4B), phosphorylation at this site appeared to be important for chromatin dissociation in response to DNA damage.

We then asked whether one phosphorylation of Chk1 affects the other phosphorylation upon DNA damage. In the soluble fraction (S1 plus S2), the protein amounts of wild-type Chk1 and its mutants were at the same level. Antibodies specific to phospho-S317 and phospho-S345 did not detect S317A and S345A proteins at all in the absence or presence of DNA damage, respectively, confirming the specificity of these antibodies. Interestingly, phosphorylation of Chk1 at S317 was strongly enhanced in wild-type Chk1 and S345A upon DNA damage induced by UV, suggesting that the phosphorylation of Chk1 at S345 did not affect the phosphorylation at S317 (Fig. 4D). In contrast, phosphorylation of Chk1 at S345 was strongly enhanced in wild-type Chk1 but was very weak in S317A upon DNA damage, suggesting that phosphorylation of S317 is required for the proper phosphorylation of S345 in response to activation of DNA damage checkpoints.

Centrosomal targeting of Chk1 complements its phosphorylation site mutations.

Given that the mutation at S345 abolished all functions tested, the possibility was raised that cytoplasmic localization of Chk1 might be important for its function. In this regard, centrosome-associated Chk1 has been reported to shield centrosomal Cdk1 from premature activation (18). Therefore, we first confirmed centrosomal localization of the endogenous Chk1 and examined the effect of DNA damage on its localization. We used anti-pericentrin antibodies to detect centrosomes. Although centrosomal signals of Chk1 appeared in trace amounts in the absence of DNA damage, they were significantly enhanced by bleomycin treatment. Interestingly, anti-phospho-S345 antibodies revealed the predominant localization of the phosphorylated form of Chk1 at S345 in the cytoplasm and its colocalization with centrosomes (Fig. 5), confirming the results of cell fractionation (Fig. 4B). In addition, similar results were also observed in Chk1-depleted cells expressing wild-type Chk1 or the S317A or S345A mutant, showing that trace centrosomal signals of wild-type Chk1 were enhanced by DNA damage. S345A did not appear to localize at centrosomes (see Fig. S3 in the supplemental material). Thus, we speculated that phosphorylation of Chk1 at S345 was required only for its centrosomal targeting and that targeting centrosomal Chk1 with mutations in its phosphorylation sites could prevent apoptosis in the presence or absence of genotoxic stressors. Interestingly, a recent study revealed that immobilizing a kinase-dead mutant of Chk1 within the centrosome by fusing it to the PACT domain of AKAP450 suppressed endogenous Chk1 activity and thus induced premature activation of Cdk1 at centrosomes (18). Using a similar approach, we transfected wild-type or mutant pCAGGS-Chk1-PACT. With the use of a green fluorescent protein (GFP) expression vector, we confirmed that transfection efficiency was more than 85% under these experimental conditions (data not shown). The results showed that immobile S345A-PACT and S317A/S345A-PACT were capable of inhibiting apoptosis in Chk1-deficient cells to an extent similar to that of wild-type Chk1-PACT, but the kinase-dead mutant (K38M)-PACT was not (Fig. 6A). S345A and S317A/S345A mutants without a PACT domain failed to inhibit apoptosis. Immunoblotting revealed that the expression levels of these proteins were almost the same (Fig. 6B). Immunostaining in situ revealed that wild-type Chk1-PACT and its mutants strictly localized at centrosomes (Fig. 6C). Interestingly, immobile S317A/S345A-PACT was also capable of inhibiting apoptosis in the presence of HU (Fig. 6D). Consistent with this, wild-type Chk1-PACT, but not K38M-PACT, colocalized with centrosome-associated phosphorylation of Cdk1 at Y15 (see Fig. S4 in the supplemental material), suggesting the inhibition of Cdk1 at centrosomes. In contrast, wild-type PACT and S317A/S345A-PACT failed to rescue the defect in intra-S phase checkpoint function, presumably due to the fact that the target of the intra-S phase checkpoint may be nuclear Cdk2 but not centrosomal Cdk1 (Fig. 6D). Taken together, these results suggested that centrosomal targeting of Chk1 suppressed both checkpoint defects and nonviable phenotypes arising from mutations in its phosphorylation sites.