Chk1 is phosphorylated at Ser317 and Ser345 by ATR in response to stalled replication and genotoxic stresses. This Chk1 activation is thought to play critical roles in the prevention of premature mitosis. However, the behavior of Chk1 in mitosis remains largely unknown. Here we reported that Chk1 was phosphorylated in mitosis. The reduction of this phosphorylation was observed at the metaphase-anaphase transition. Two-dimensional phosphopeptide mapping revealed that Chk1 phosphorylation sites in vivo were completely overlapped with the in vitro sites by cyclin-dependent protein kinase (Cdk) 1 or by p38 MAP kinase. Ser286 and Ser301 were identified as novel phosphorylation sites on Chk1. Treatment with Cdk inhibitor butyrolactone I induced the reduction of Chk1-S301 phosphorylation, although treatment with p38-specific inhibitor SB203580 or siRNA did not. In addition, ionizing radiation (IR) or ultraviolet (UV) light did not induce Chk1 phosphorylation at Ser317 and Ser345 in nocodazole-arrested mitotic cells. These observations imply the regulation of mitotic Chk1 function through Chk1 phosphorylation at novel sites by Cdk1.
Cell cycle checkpoints are fundamental mechanisms not only to monitor genomic stability but also to coordinate repair and cell cycle progression (Hartwell & Weinert 1989). These checkpoints ensure that damaged or unreplicated DNA is not segregated into the daughter cells in mitosis (Zhou & Elledge 2000). Defects in these processes cause genomic instability and predispose to cancer (Bartek & Lukas 2003; Kastan & Bartek 2004). In the center of these pathways, there exists two parallel but partially overlapping protein kinase cascades (Rhind & Russell 2000; Abraham 2001; Bartek & Lukas 2003; Shiloh 2003). One is the ataxia-telangiectasia mutated (ATM)-Chk2 pathway activated in response to DNA double strand break. ATM induces Chk2 phosphorylation at Thr68, which triggers Chk2 activation through the induction of its autophosphorylation at Thr383 and Thr387. The other is ATM- and Rad3-related (ATR)-Chk1 pathway activated by a broader spectrum of genotoxic stimuli including ultraviolet (UV) and DNA replication stresses (e.g., DNA replication inhibitors). ATR induces Chk1 phosphorylation at Ser317 and Ser345 (Zhao & Piwnica-Worms 2001), which is considered to facilitate Chk1 function (Zhou & Elledge 2000; Kastan & Bartek 2004).
With regard to mitotic delay, two pathways are considered to share a common mechanistic endpoint (Zhou & Elledge 2000; Neely & Piwnica-Worms 2003). Chk1/Chk2 phosphorylates Cdc25 family phosphatase (Rhind & Russell 2000; Bartek & Lukas 2003). This phosphorylation leads to the inhibition of Cdc25 family phosphatase. Since Cdc25 family phosphatase activates cyclin dependent kinase 1 (Cdk1, also called Cdc2) by dephosphorylating two residues (Thr14 and Tyr15), the inactivation of Cdc25 family phosphatase increases the level of Cdk1 inhibitory phosphorylation (Zhou & Elledge 2000). This mechanism is critical for G2/M checkpoint, since Cdk1 activation is essential for mitotic entry (Doree & Hunt 2002; Nurse 2002).
During mitosis, Chk1/Chk2 was also reported to be phosphorylated even in the absence of DNA damage stress (Tsvetkov et al. 2003; Ng et al. 2004). Chk2-Thr68 phosphorylation occurred at the centrosomes and midbody where Polo-like kinase 1 (Plk1) was localized, and was induced by Plk1 in vitro (Tsvetkov et al. 2003). Chk1 phosphorylation occurred at cells arrested in mitosis by treating with microtubule depolymerization or stabilization reagent (Ng et al. 2004). This phosphorylation did not occur at Ser345 or in ATM-dependent manner (Ng et al. 2004). However, it is largely unknown about the regulation of mitotic phosphorylation of these checkpoint-related kinases. In this study, we reported that mitotic Chk1 phosphorylation occurred at novel phosphorylation sites and was governed by Cdk1.
Chk1 phosphorylation in mitosis
We first examined the phosphorylation state of Chk1 in cells arrested at early mitosis by the addition of the microtubule depolymerization reagent, nocodazole. The electrophoretic mobility of Chk1 in mitosis was slower than that in interphase (Fig. 1A). The mitotic mobility shift was completely abolished by treatment with λ protein phosphatase (λPPase) but not using dephosphorylation buffer only (Fig. 1A). This mobility shift was diminished at 40 min, and was almost abolished at 60 min after the release of nocodazole (Fig. 1B). The decrease of mobility shift coincided with the reduction of cyclin B1 protein level, one of the markers at metaphase-anaphase transition (Nurse 2002; Doree & Hunt 2002) (Fig. 1B). These results suggested that Chk1 was phosphorylated specifically in early mitosis.
Identification of mitosis-specific Chk1 phosphorylation sites
The detection of Chk1 phosphorylation in mitosis raised the question of its responsible kinase. Since Chk1 was dephosphorylated after metaphase-anaphase transition (Fig. 1B), the responsible kinase may be activated specifically in early mitosis. Cyclin-dependent kinase 1 (Cdk1, also called Cdc2) is known to be activated at entry into mitosis and be inactivated at metaphase-anaphase transition through cyclin B degradation (Doree & Hunt 2002; Nurse 2002). p38 MAP kinase was also reported be activated in the spindle assembly checkpoint during early mitotic process (Takenaka et al. 1998). So, we investigated whether Cdk1 or p38 can phosphorylate Chk1 at site(s) occurring in early mitosis. Two-dimensional phosphopeptide mapping analyses revealed that Cdk1 and p38 are likely candidates for mitotic Chk1 kinase (Fig. 1C).
Cdk1 (Doree & Hunt 2002; Nurse 2002) or p38 (Miyata & Nishida 1999; Chang & Karin 2001) preferentially phosphorylates (Ser/Thr)-Pro sites: human Chk1 contains four putative phosphorylation sites (three Ser-Pro and one Thr-Pro). Since Chk1 was phosphorylated only at Ser residues (Fig. 1D), we produced Chk1 in which Ser286, Ser301, or Ser331 is changed into Ala. Mutation at Ser286 or Ser301 to Ala induced the disappearance of the radioactive spot 1 or 2 on the thin layer plate (Fig. 2A), although mutation at Ser331 did not (data not shown). Ser286 or Ser301 mutation reduced both the mobility shift (Fig. 2B) and the phosphorylation of Chk1 by Cdk1 (Fig. 2C): both were almost impaired by double mutation (Fig. 2B,C). These results suggested that Cdk1 phosphorylated Chk1 at Ser286 and Ser301 in vitro.
Pro after a putative phosphorylation site is considered to be essential for the kinase recognition of Cdk1 (Doree & Hunt 2002; Nurse 2002) or p38 (Miyata & Nishida 1999; Chang & Karin 2001). However, Pro287 in human Chk1 is changed into Ser in mouse and rat Chk1 (Fig. 2D), suggesting the possibility that Ser286 phosphorylation may not occur in mouse or rat mitotic cells. On the other hand, Ser301-Pro302 sequence is conserved among human, mouse, rat and chicken Chk1 (Fig. 2D). So, we produced a site- and phosphorylation state-specific antibody (Nishizawa et al. 1991; Yano et al. 1991; Kawajiri & Inagaki 2004) for Ser301 on Chk1, using a phosphopeptide pS301 corresponding to Ser301-phosphorylated Chk1. Rat monoclonal (mono) or rabbit polyclonal (poly) antibody recognized Chk1 phosphorylated by Cdk1 or by p38, but not non-phosphorylated one (control; Fig. 2E). The immunoreactivity toward each phosphorylated Chk1 was abolished by the preincubation with the phosphopeptide pS301, but not with non-phosphorylated peptide S301 (Fig. 2E). These results suggested that our antibodies against pS301 specifically recognized Chk1 phosphorylation at Ser301.
Our rat monoclonal or rabbit polyclonal anti-phosphoSer301 on Chk1 crossreacted with about 54 kDa band which was not corresponding to Chk1 in the mitotic cell lysate (data not shown). Therefore, these antibodies cannot be used for immunoblotting of the cell lysate or immunostaining. So, for the detection of Chk1 phosphorylation at Ser301 in cells, we used anti-Chk1 immunoprecipitates in the following experiments.
Cdk1 is the most likely candidate for mitotic Chk1 kinase
To examine which kinase phosphorylates Chk1 in mitosis, we first checked the effect of p38-specific inhibitor SB203580 (Lee et al. 1999) on mitotic Chk1 phosphorylation. HeLa cells were synchronized at G1/S boundary by the method of double thymidine block. After the release of the second thymidine block, cells were treated with or without SB203580 in the presence of nocodazole. As shown in Fig. 3A, mitotic cells accumulated in a time-dependent manner: there was only marginal difference of mitotic cell accumulation between two cell groups. SB203580 did not appear to affect mitotic Chk1 mobility shift, although it inhibited HSP27-Ser82 phosphorylation likely by MAPKAP-2 (Landry et al. 1992; Rouse et al. 1994) downstream of p38 (Fig. 3B). Rather, this mobility shift appeared to coincide with the increase of immunoreactivity of 4A4 (Tsujimura et al. 1994), which specifically recognizes vimentin-Ser55 phosphorylation by Cdk1 (Fig. 3B).
To demonstrate p38 independence of mitotic Chk1 phosphorylation more clearly, HeLa cells were subjected to double thymidine cell synchronization combined with p38 α and β siRNA (Fig. 3C). This treatment effectively induced p38 depletion in HeLa cells (Fig. 3C). Like SB203580 treatment, p38 siRNA transfection did not appear to affect Chk1-Ser301 phosphorylation, Cdk1-induced vimentin-Ser55 phosphorylation (Tsujimura et al. 1994) (Fig. 3C) and mitotic cell accumulation (Fig. 3D) in the presence of nocodazole at 16 h after the release of the second thymidine block. These results suggested that p38 is unlikely to be responsible for mitotic Chk1 phosphorylation.
Next, we checked the effect of Cdk inhibitor butyrolactone I (Kitagawa et al. 1993) on mitotic Chk1 phosphorylation. As shown in Fig. 3C, butyrolactone I treatment induced the reduction of Chk1-S301 phosphorylation together with vimentin-Ser55 phosphorylation by Cdk1 (Tsujimura et al. 1994). This observation implied that Cdk1 is the most likely candidate for mitotic Chk1 phosphorylation.
Cdk1-induced phosphorylation may not affect the subcellular localization of Chk1
Recent analysis suggested the possibility that Chk1 phosphorylation at Ser317 and Ser345 may play important roles not only in its kinase activation but also in the nuclear accumulation of Chk1 (Jiang et al. 2003). So, we constructed a plasmid expressing either wild-type (WT) or S286D/S301D (DD) mutant Chk1, modified by addition of C-terminal 7Myc tag and by three silent mutations to render the mRNA resistant to RNA interference (RNAi). Immunoblotting analysis confirmed that small interfering RNA (siRNA) transfection reduced the expression level of endogenous Chk1 but not that of exogenous Chk1-7Myc (Fig. 4A). In this experimental condition, there were only marginal differences of Chk1 localization between WT (Fig. 4B) and phospho-mimic mutant (DD; Fig. 4C) Chk1. These results suggested that Chk1 phosphorylation at Ser286 and Ser301 might not affect its subcellular localization.
Ionizing radiation (IR) or ultraviolet (UV) light does not induce the phosphorylation of Ser317 and Ser345 on Chk1 in nocodazole-arrested mitotic cells
In order to assess the likely biological significance of mitotic Chk1 phosphorylation, we examined whether or not ionizing radiation (IR) or ultraviolet (UV) light induced Chk1 phosphorylation at Ser317 and Ser345 during early mitosis when Chk1 was phosphorylated at Ser286 and Ser301. In interphase cells treated with nocodazole, 10 Gy IR or 10 J/m2 UV induced Chk1-Ser317 and -Ser345 phosphorylation (Fig. 5A), which are thought to be essential for the facilitation of Chk1 function (Zhou & Elledge 2000; Zhao & Piwnica-Worms 2001). However, the same dose of IR or UV did not induce Chk1 phosphorylation at Ser317 and Ser345 in nocodazole-arrested mitotic cells (Fig. 5A). These observations suggested that Chk1 phosphorylation did not occur at Ser317 and Ser345 in response to genotoxic stresses during mitosis although Ser286 and Ser301 were phosphorylated (Fig. 5B).
A major new finding in this study is that mitotic Chk1 phosphorylation occurred at Ser286 and Ser301, phosphorylation sites different from ones by ATR (Zhao & Piwnica-Worms 2001). In addition, Cdk1 is the most likely candidate for this mitotic Chk1 kinase.
Our analyses revealed that Cdk1 and p38 can phosphorylate Chk1 at sites occurring in mitosis (Fig. 1). Treatment with p38-specific inhibitor or siRNA induced only marginal change in Chk1 phosphorylation at Ser301 (Fig. 3B,C). So, p38 was unlikely to phosphorylate Chk1 in mitosis. On the other hand, treatment with Cdk-specific inhibitor reduced the level of Chk1 phosphorylation at Ser301 (Fig. 3C). Since this treatment also blocked the mitotic accumulation to some extent (Fig. 3D), we could not completely rule out the possibility that the block of mitotic entry might lead to the reduction of Chk1 phosphorylation at Ser301 rather than the inhibition of Cdk1 activity. However, mitotic Chk1 phosphorylation occurred only at Ser prior to Pro (Fig. 2) and disappeared at metaphase-anaphase transition (Fig. 1A): Cdk1 has these features of mitotic Chk1 kinase (Doree & Hunt 2002; Nurse 2002). Therefore, Cdk1 governed mitotic Chk1 phosphorylation likely through the direct-enzyme reaction.
Our observations also raised the question whether or not p38 is activated in the spindle assembly checkpoint, for the following reasons. In our experimental conditions, p38 phosphorylation at its activation sites does not appear to occur in nocodazole-arrested mitotic cells (Fig. 3B). And, treatment with p38-specific inhibitor or siRNA induced the only marginal change of vimentin-Ser55 phosphorylation by Cdk1 (Tsujimura et al. 1994) but also of mitotic index (Fig. 3). These observations contrast with the previous report that treatment with the same p38-specific inhibitor reduced H1 kinase activity in mitotic NIH3T3 cells arrested by nocodazole (Takenaka et al. 1998). Recently, p38α was reported to be activated at the antephase checkpoint, which acts at prophase to decondense the chromosomes and to return to G2 interphase (Matsusaka & Pines 2004). This checkpoint, which HeLa cells lack, was activated in response to another microtubule depolymerizing reagent, colcemid (Matsusaka & Pines 2004). So, our observations are consistent with this notion that p38 participates in the antephase checkpoint rather than in the spindle assembly checkpoint.
Detection of mitotic Chk1 phosphorylation at novel sites raised the question of its biological significance. One possibility is the change of Chk1 kinase activity. Ng et al. (2004) reported that mitotic Chk1 phosphorylation reduced its basal kinase activity. However, we observed only marginal difference of the kinase activity toward Cdc25A, Cdc25C, Wee1A and Wee1B between interphase and mitotic Chk1 immunocomplexes (data not shown). So, mitotic Chk1 phosphorylation might not change its basal kinase activity. The second possibility is the change of Chk1 localization. We tried to examine the subcellular localization of a WT or S286A/S301A in mitotic cells. However, we could not observe such mitotic cells at least in the transient expression system: this may be due to the checkpoint activation induced by the exogenous expression of Chk1 protein. So, we alternatively examined the subcellular localization of either WT or phospho-mimic mutant (DD) Chk1 in interphase cells. Since we observed only marginal difference of the localization between two proteins (Fig. 4), Chk1 phosphorylation at Ser286 and Ser301 might not affect its subcellular localization. The third possibility is the prevention of Chk1 activation in response to genotoxic stresses. In nocodazole-arrested mitotic cells, IR did not induce Chk1 phosphorylation at Ser345, one of ATR phosphorylation sites (Huang et al. 2005). We also observed that IR or UV treatment did not induce Chk1 phosphorylation at Ser317 and Ser345 in nocodazole-arrested mitotic cells (Fig. 5A). These observations raised the possibility that mitotic Chk1 phosphorylation might play some roles in the inactivation of ATR-Chk1 pathway during early mitosis (Fig. 5B). Interestingly, Cdc25C phosphorylation at Ser214 by Cdk1 was reported to play an important role in G2/M transition through prevention of Cdc25C phosphorylation at Ser216, a Chk1 phosphorylation site, in mitosis (Bulavin et al. 2003). So, in early mitosis, Chk1-dependent checkpoint pathway might be inactivated at several steps.
In conclusion, we demonstrated that mitotic Chk1 phosphorylation occurred at novel sites and was regulated by Cdk1. The biological significance of this mitotic Chk1 phosphorylation will be addressed in the future work.
pENTR-3C-7myc plasmid is based on pENTR-3C (Invitrogen, Carlsbad, CA, USA), in which 7myc-tagged sequence with stop codon is inserted using NotI and XhoI sites. To create EcoRI site at 5′ end and NotI site at 3′ end and to delete stop codon, human Chk1 gene was amplified by polymerase chain reaction (PCR) using pcDNA-Chk1 (Kaneko et al. 1999) as a template. Then, the amplified Chk1 gene was inserted into pENTR-3C-7myc using EcoRI and NotI sites. The S286D/S301D (DD) mutation was created by site-directed mutagenesis using Quickchange protocol (Stratagene, La Jolla, CA, USA). To obtain RNAi-insensitive Chk1 constructs, three silent point mutations within the siRNA target sequence (see “Transfection”) were also introduced by site-directed mutagenesis. For the expression of Chk1 proteins in HeLa cells, pDEST-26 carrying WT or DD Chk1-7myc gene with silent mutations was created by the homologous recombination reaction between pDEST-26 vector (Invitrogen) and each pENTR vector described above.
Peptides and antibodies
Chk1 peptides S301 (Cys-Ser-Asn-Leu-Asp-Phe-Ser-Pro-Val-Asn-Ser-Ala) and pS301 (Cys-Ser-Asn-Leu-Asp-Phe-phosphoSer-Pro-Val-Asn-Ser-Ala) were chemically synthesized (Peptide Institute Inc., Osaka, Japan). Site- and phosphorylation state-specific (rabbit polyclonal and rat monoclonal) antibodies for Ser301 on Chk1 were produced by methods using pS301 as an antigen for immunization, which we had first established (Nishizawa et al. 1991; Yano et al. 1991). Other antibodies were purchased from following companies: rabbit polyclonal anti-Chk1 (FL-478) from Santa Cruz Biotechnology (Santa Cruz, CA, USA); mouse monoclonal anti-Chk1 (DCS-310) or anti-Myc (9E10) from Sigma (St Louis, MO, USA); rabbit polyclonal anti-p38, anti-p38-pT180/pY182, anti-HSP-pSer82, anti-Chk1-pSer317, or anti-Chk1-pSer345 from Cell Signaling Technology (Beverly, MA, USA); rabbit polyclonal anti-Cdk1 (Ab-1) or mouse monoclonal anti-Cyclin B1 (Ab-2) from Calbiochem (San Diego, CA, USA).
Production of Chk1 protein
His-tagged human Chk1 kinase dead mutant (K38M) protein was expressed in High Five cells using a baculovirus expression system (Kaneko et al. 1999) and purified by nickel-nitrilotriacetic acid column chromatography according to the manufacturer's protocol (Qiagen). GST-fusion human Chk1 WT and mutants in which Ser286 and/or Ser301 are changed into Ala were expressed in BL21-Codon Plus (Stratagene) and purified as reported previously (Zhao & Piwnica-Worms 2001): each GST-fusion protein exhibited little kinase activity in our experimental condition.
Active Cdk1 were purified from Sf9 cells infected with baculoviruses encoding Cdk1 and GST-human cyclin B1 (Watanabe et al. 1995). GST-p38α protein was purchased from Chemicon International. The phosphorylation assay was performed for 60 min at 30 °C in 20 µL of 25 mm Tris-Cl (pH 7.5), 0.1 mm ATP, 10 mm MgCl2, 0.1 µm Calyculin A, and 200 µg/mL His-tagged Chk1 or 150 µg/mL GST-Chk1 in the presence of 20 µg/mL Cdk1 or 10 µg/mL GST-p38α with or without 2 µCi [γ-32P] ATP.
Cell culture, synchronization and metabolic labeling
HeLa or CHO cells were grown in Dulbecco's modified Eagle's medium (DMEM) or DMEM/F12 (v/v = 1/1) with 10% calf serum or 10% fetal bovine serum, respectively. For preparation of interphase (I) or mitotic (M) cells, HeLa cells or CHO cells (Fig. 1B) were synchronized in early mitosis by the addition of 200 ng/mL or 50 ng/mL nocodazole for 16–18 h or 4 h, respectively. Early mitotic cells were collected by mechanical shaking off, and then adherent cells served as interphase cells. For metabolic labeling, after mitotic HeLa cells were washed with labeling medium (phosphate-free DMEM plus 10% dialyzed calf serum), cells were labeled with [32P] orthophosphate (Amersham Biosciences, Piscataway, NJ, USA) at a final concentration of 0.7 mCi/mL for an additional 4 h. Double thymidine block synchronization was performed as follows. HeLa cells were first arrested by addition of 2 mm thymidine for 16–18 h and then released. At 9–10 h after release, cells were arrested in G1/S boundary by addition of 2 mm thymidine for 16–18 h. After the release of the second thymidine block, cells were incubated with or without 20 µm SB203580 in the presence of 200 ng/mL nocodazole at the indicated time. For the experiment using butyrolactone I, cells had been treated with 100 µm butyrolactone I in the presence of 200 ng/mL nocodazole from 12 h to 16 h after the release of the second thymidine block.
Tryptic phosphopeptide mapping and phosphoamino acid analyses
Tryptic phosphopeptide mapping and phosphoamino acid analyses were performed as described (Boyle et al. 1991).
21-nucleotide double strand RNAs were commercially synthesized in Qiagen (Valencia, CA, USA): the target sequence of human p38α, p38β, or Chk1 was (AA)GAAGCTCTCCAGACCATTT (AA)CTGGATGCATTACAACCAA, or (AA)GCGTGCCGTAGACTGTCCA, respectively. HeLa cells were transfected with each siRNA using Oligofectamine reagent (Invitrogen) or with each pDEST-26 vector described above using Lipofectamine Plus reagent (Invitrogen). The combination of double thymidine block synchronization and siRNA transfection was done as described (Hirota et al. 2003).
Detection of Chk1 phosphorylation at Ser301 in the immunoblotting analysis
In mitotic HeLa cell lysate, rat monoclonal or rabbit polyclonal anti-phosphoSer301 on Chk1 crossreacted with about 54 kDa band which was not corresponding to Chk1 (data not shown). Thus, we used following anti-Chk1 immunoprecipitates for the detection of Chk1 phosphorylation at Ser301. HeLa cells were lyzed in modified RIPA buffer (50 mm Tris-Cl (pH 8.0), 150 mm NaCl, 1 mm EDTA, 20 mmβ-glycerophosphate, 20 mm NaF, 1 mm Na3VO4, 1% NP40, 0.5% deoxycholate, 0.1% SDS and protease inhibitor cocktail (Sigma)). The cell extract was incubated with rabbit polyclonal anti-Chk1 (FL-478) for 1 h, followed by an additional 1 h incubation with Protein-A-Sepharose beads. Each immunoprecipitate was subjected to immunoblotting with mouse monoclonal anti-Chk1 (DCS-310) or rat monoclonal anti-phosphoSer301 on Chk1.
We thank T. Takahashi for continuous encouragement and support. We also thank T. Yamaguchi, Y. Hayashi, and C. Yuhara for technical assistance, and N. Watanabe (RIKEN, Japan) for providing baculovirus encoding GST-Cyclin B1 or Cdk1. HG thanks Y. Yamakita and S. Yamashiro (Rutgers University) for technical support and helpful discussion. This work was supported in part by Grants-in-Aid for Scientific Research and Cancer Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan, by a grant-in-aid for the Third Term Comprehensive 10-Year Strategy for Cancer Control from the Ministry of Health, Labour and Welfare, Japan, by The Naito Foundation, and by the Uehara Memorial Foundation.