Replication Checkpoint Enforced by Kinases Cds1 and Chk1

See allHide authors and affiliations

Science  08 May 1998:
Vol. 280, Issue 5365, pp. 909-912
DOI: 10.1126/science.280.5365.909


Cdc2, the kinase that induces mitosis, is regulated by checkpoints that couple mitosis to the completion of DNA replication and repair. The repair checkpoint kinase Chk1 regulates Cdc25, a phosphatase that activates Cdc2. Effectors of the replication checkpoint evoked by hydroxyurea (HU) are unknown. Treatment of fission yeast with HU stimulated the kinase Cds1, which appears to phosphorylate the kinase Wee1, an inhibitor of Cdc2. The protein kinase Cds1 was also required for a large HU-induced increase in the amount of Mik1, a second inhibitor of Cdc2. HU-induced arrest of cell division was abolished incds1 chk1 cells. Thus, Cds1 and Chk1 appear to jointly enforce the replication checkpoint.

Cell cycle checkpoints ensure that chromosomal DNA is replicated and repaired before nuclear division (1, 2). In mammalian cells, loss of checkpoint control results in rearrangements, amplification, and loss of chromosomes, events that are causally associated with cancer. In yeasts, checkpoint controls are vital for survival when DNA is damaged or replication is inhibited.

Cdc2, the cyclin-dependent kinase that initiates mitosis, is the ultimate target of the DNA replication and repair checkpoints. In the fission yeast Schizosaccharomyces pombe, Cdc2 is inhibited by phosphorylation on Tyr15. This phosphorylation is catalyzed by the kinases Wee1 and Mik1 and reversed by the phosphatase Cdc25. Inhibitory phosphorylation of Cdc2 is crucial for replication and repair checkpoints in fission yeast and human cells (3-6). Chk1 may enforce the DNA repair checkpoint by phosphorylating and inhibiting Cdc25 (6-8). Chk1 is only required for the repair checkpoint and not for the replication checkpoint evoked by HU (9). However, an HU-induced arrest is rapidly abrogated after inactivation of temperature-sensitive Wee1 protein in amik1 background (10), showing that the replication checkpoint requires inhibitory phosphorylation of Cdc2. Wee1 and Mik1 are not individually required for an HU arrest, suggesting that Wee1 and Mik1 may be coordinately regulated by the replication checkpoint. Experiments were designed to test this hypothesis.

The abundance of Wee1 was unaffected by HU (11). We tested whether Wee1 was phosphorylated by an HU-activated kinase, with attention focused on the NH2-terminal regulatory domain (12). GST:Wee1152, a fusion of glutathione S-transferase to amino acids 11 to 152 of Wee1, was expressed in bacteria. Protein preparations yielded full-length GST:Wee1152 and GST:Wee170 (Fig.1A), the latter being a truncation product containing ∼70 amino acids of Wee1 (13). These proteins were bound to glutathione (GSH)-Sepharose and mixed with lysates from cells harvested at intervals during incubation with HU (14). The proteins bound to GSH-Sepharose were washed, incubated with [γ-32P]adenosine triphosphate, and analyzed by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) (Fig. 1B). GST:Wee1152 was phosphorylated in the 0-min sample, whereas GST:Wee170 phosphorylation was negligible. Phosphorylation of both proteins increased during incubation in HU. Unfused GST was not phosphorylated. Thus, GST:Wee170 associates with and is phosphorylated by an HU-regulated kinase, whereas GST:Wee1152 is phosphorylated by this kinase and another that is active before HU treatment.

Figure 1

Association of an HU-regulated kinase with the NH2-terminus of Wee1. (A) SDS-polyacrylamide gel of GST:Wee1152 and degradation products produced in bacteria and purified with GSH-Sepharose. The major degradation products are GST fused to ∼70 amino acids of Wee1 (GST:Wee170) and unfused GST. Proteins were stained with Coomassie blue. (B) Cells were treated with 12 mM HU for a 1-hour time course at 30°C. Cell lysates were incubated with GST:Wee1 proteins bound to GSH-Sepharose. The GSH-Sepharose was washed and assayed for associated kinase activity. The positions of GST:Wee1152 and GST:Wee170 after 12% SDS-PAGE are shown. (C) Mutational inactivation ofcdc22, encoding the large subunit of ribonucleotide reductase, enhances phosphorylation of GST:Wee170. Cells carrying the temperature-sensitivecdc22-M45 allele or wild-type cells were grown at 25°C. The cultures were split and maintained at 25°C or incubated at 35.5°C for 90 min. GST:Wee1 phosphorylation assays were performed as described above. (D) Activation of the kinase that phosphorylates GST:Wee170 in extracts fromcdc25-22 cells released into HU from a G2arrest. A culture of cdc25-22 cells grown at 25°C was shifted to 35.5°C for 4 hours. The culture was split and treated with 12 mM HU (+HU, □) or mock treated (−HU, ◊) for 45 min at 35.5°C. Both cultures were returned to 25°C, and samples were taken every 10 min. The septation index marks the completion of mitosis (top). Lysates from each time point were assayed for GST:Wee170 kinase (bottom). GST:Wee170 phosphorylation occurred in the HU-treated samples after exit from M, whereas phosphorylation of GST:Wee170 remained low in samples from the mock-treated culture. In the mock-treated culture, activity of the kinase that phosphorylates GST:Wee1152increased as cells underwent mitosis.

HU prevents DNA replication by inhibiting ribonucleotide reductase and thereby reducing deoxyribonucleotide pools, but it may also have other effects. However, mutational inactivation of cdc22, which encodes the large subunit of ribonucleotide reductase (15), caused phosphorylation of GST:Wee170 in cell lysates (Fig. 1C), indicating that the HU effect is due to inhibition of DNA replication.

GST:Wee170 phosphorylation was assayed in lysates from temperature-sensitive cdc25-22 cells that were arrested in G2 phase, incubated in HU for 45 min, and then released from G2 arrest in medium containing HU. HU failed to induce phosphorylation of GST:Wee170 in lysates prepared from cdc25-22 cells arrested in G2 (Fig. 1D). Upon release from G2, GST:Wee170 phosphorylation and cell septation increased coincidentally. Septation marks the end of mitosis (M) and the beginning of DNA replication (S); thus, GST:Wee170phosphorylation increased as cells encountered the HU arrest. No increase in GST:Wee170phosphorylation was detected in a mock HU-treated culture (Fig. 1D). These observations indicate that the kinase that phosphorylates GST:Wee170 is activated when cells attempt DNA replication with insufficient deoxyribonucleotides.

Phosphorylation of GST:Wee170 was abolished in lysates made from rad1 and rad3 checkpoint mutants (Fig. 2A), suggesting that the kinase that phosphorylates GST:Wee170 may be involved in the replication checkpoint. Experiments were undertaken to identify this kinase. Chk1 was excluded by experiments with lysates made fromchk1 cells (Fig. 2B). The involvement of Cds1, a protein kinase that is important for survival of HU treatment (16), was indicated by the observation that GST:Wee170phosphorylation was abolished in a lysate prepared from HU-treated cds1 cells (Fig. 2B). An experiment was performed to determine whether Cds1 coprecipitated with GST:Wee170. Lysates from HU-treated cells in which genomic cds1 encoded a protein with a COOH-terminal HAHIS tag [consisting of two hemagglutinin (HA) epitopes followed by hexahistidine] were incubated with GST:Wee1 or GST proteins, washed, and analyzed by immunoblotting. Cds1HAHIS precipitated with GST:Wee1 but not with unfused GST (Fig. 2C). Rad3 did not associate with GST:Wee1 (11). Cds1 purified from fission yeast phosphorylated GST:Wee170 in vitro, producing a two-dimensional tryptic phosphopeptide map that was identical to a map made from GST:Wee170 phosphorylated by its associated kinase (11). These data strongly suggest that Cds1 is the HU-regulated kinase that associates with and phosphorylates GST:Wee170.

Figure 2

Phophorylation of GST:Wee170 by the kinase Cds1. (A) Phosphorylation of GST:Wee170 is dependent on Rad1 and Rad3. Wild-type, rad1, and rad3 strain cultures were split and treated for 1 hour with HU or mock-treated and then assayed for GST:Wee170 phosphorylation as described in Fig. 1. (B) GST:Wee170phosphorylation is dependent on Cds1 but not Chk1. Wild-type, chk1, and cds1 cultures were split and treated for 1 hour with HU or mock treated. Phosphorylation of GST:Wee1 was assayed. (C) Cds1 interacts with GST:Wee1 but not with GST. Lysates from HU-treated cells expressing epitope-tagged Cds1 [Cds1HAHIS(Cds1:HA)] expressed from the cds1 genomic locus were incubated with GST or GST:Wee1 proteins bound to GSH-Sepharose. After extensive washing, samples were resolved by 10% SDS-PAGE and subjected to immunoblotting with antibodies to HA and GST (anti-HA and anti-GST). (D) Wee1 interacts with GST:Cds1 but not with GST:Chk1 in vivo. Strains carrying integratednmt1:GST:chk1 + ornmt1:GST:cds1 + constructs were grown in minimal medium (EMM2) without thiamine at 30°C for 16 hours to induce expression of GST fusion proteins. Cells were lysed, and GST fusion proteins were purified with GSH-Sepharose. After extensive washing, samples were resolved by 10% SDS-PAGE and subjected to immunoblotting with anti-Wee1, anti-Cdc25, and anti-GST. A ∼110-kD protein corresponding to Wee1 was detected in the GST:Cds1 sample, whereas a ∼80-kD protein corresponding to Cdc25 was detected in the GST:Chk1 sample. The ∼110-kD protein was not detected in samples prepared from Δwee1 cells (11). GST:Cds1 and GST:Chk1 migrate at positions corresponding to ∼82 and ∼80 kD, respectively.

Association of Cds1 and Wee1 was also examined in vivo. GST:Cds1 was expressed from the thiamine-repressible nmt1 promoter (17). As a control, we used a strain that expressed GST:Chk1. The GST fusion proteins were purified with GSH-Sepharose and analyzed by immunoblotting. Cdc25 (but not Wee1) associated with GST:Chk1, whereas Wee1 (but not Cdc25) associated with GST:Cds1 (Fig.2D).

HU may increase the abundance of Cds1, it may make Cds1 competent to associate with Wee1, or perhaps the replication checkpoint enhances the kinase activity of Cds1. Cells expressing Cds1HAHIS were treated with HU or mock-treated, and lysates were made and precipitated with GST:Wee1 proteins. Association of Cds1HAHIS with GST:Wee1 was detected in lysates from cells treated with HU, whereas Cds1HAHIS did not associate with GST:Wee1 in mock-treated samples (Fig. 3A). Ni2+–nitrilotriacetic acid (NTA) precipitation of Cds1HAHIS by the hexahistidine sequence showed that Cds1HAHIS was equally abundant in HU- and mock-treated cell lysates. Therefore, HU treatment enhanced association of Cds1 with GST:Wee1 (18). Cds1HAHIS was also purified from HU- or mock-treated cells and assayed in an autophosphorylation assay. Phosphorylation of Cds1 was increased in samples from cells treated with HU (Fig. 3B). Similar results were obtained with GST: Wee170 as a substrate (11). Immunoblot analysis confirmed that similar amounts of Cds1HAHIS were purified from HU- and mock-treated cells (Fig. 3B). Thus, HU apparently both activates Cds1 as a kinase and stimulates binding to Wee1.

Figure 3

Stimulation by HU of Cds1 kinase activity, binding of Cds1 to GST:Wee170 in lysates, and accumulation of Mik1. (A) HU stimulates Cds1 binding to GST:Wee170. A culture of cells expressing Cds1HAHIS from the cds1 genomic locus was split and treated for 3.5 hours with HU or mock-treated. The cultures were split again, and samples were processed for Ni2+-NTA purification of Cds1HAHIS in denaturing conditions or purification of proteins that bind GST:Wee1 proteins in nondenaturing conditions. After extensive washing, samples were resolved by 10% SDS-PAGE and subjected to immunoblot analysis. (B) HU activates Cds1 kinase activity. In a separate experiment, Cds1HAHIS was purified in native conditions and assayed for autophosphorylation activity (top). Immunoblot analysis confirmed that similar amounts of Cds1HAHIS were recovered in the HU- and mock-treated samples (bottom). (C) Cds1- and Rad3-dependent increase in Mik1 abundance in HU-treated cells. Wild-type, cds1, and rad3 cells in which genomic mik1 encoded Mik1HAHIS (Mik1:HA) protein were incubated for 4 hours in YES at 25°C with 12 mM HU or mock-treated. Mik1HAHIS protein was purified by Ni2+-NTA purification in denaturing conditions and detected by immunoblotting with anti-HA. (D) Mik1 is important for survival of HU treatment. Serial dilutions (1X, 500 cells; 0.2X, 100 cells) of wild-type and mik1 cells were plated in YES medium supplemented with 0 or 6 mM HU and incubated for 2 days at 30°C.

Phosphorylation of Wee1 by Cds1 may play a role in the DNA replication checkpoint. However, wee1 mutants arrest division in response to HU (3); thus, another protein that controls Cdc2 Tyr15 phosphorylation must also be regulated by the replication checkpoint. The kinase Mik1 phosphorylates Cdc2 on Tyr15 (4); therefore, the abundance of Mik1 was monitored in HU-treated cells with a strain expressing Mik1HAHIS from the genomic locus. Immunoblot analysis showed that the amount of Mik1HAHIS increased in cells treated with HU (Fig. 3C). The abundance of mik1 mRNA was unchanged in HU-treated cells (11); thus, HU must affect Mik1 synthesis or turnover. The HU-induced increase of Mik1HAHIS was largely abolished in cds1 cells and rad3 cells (Fig. 3C), indicating that this response may be regulated by the replication checkpoint. Mik1 is a dose-dependent inhibitor of mitosis (4); thus, the increase of Mik1 abundance in HU-treated cells should help to enforce the checkpoint. Indeed, mik1 mutants exhibited enhanced sensitivity to HU (Fig. 3D).

Expression of a large amount of GST:Cds1 from the nmt1promoter caused a cell cycle arrest, as indicated by highly elongated cells (Fig. 4). The arrest occurred inrad3 and chk1 mutants, indicating that Cds1 functions downstream of Rad3 and independently of Chk1. Expression of GST:Cds1 had no effect in cells that express Cdc2Y15F, a form of Cdc2 that cannot be phosphorylated by Wee1 or Mik1. These findings support the conclusion that Cds1 transmits a checkpoint signal.

Figure 4

Cell cycle arrest caused by expression of large amounts of GST:Cds1. Cells having an integratednmt1:GST:cds1 + construct were incubated innmt1-repressing medium (promoter off) ornmt1-inducing medium (promoter on) for 19 hours at 30°C. GST:Cds1 expression induced cell cycle arrest in wild-type,rad3, and chk1 strains. Flow cytometry analysis confirmed that these cells were arrested with a 2C DNA content (11). Cells expressing Cdc2Y15F, a form of Cdc2 in which tyrosine at position 15 is replaced with phenylalanine, were insensitive to GST:Cds1.

Cells lacking Cds1 arrest division in response to HU, indicating that the Cds1 signal is supplemented by other checkpoint proteins (16). We therefore examined the replication checkpoint in synchronous cultures of cds1 chk1 double-mutant cells;cds1, chk1, and rad3 single-mutant cells; and wild-type cells (19). HU arrested division in wild-type cells and in cds1 and chk1single-mutant cells (Fig. 5A). In contrast, cds1 chk1 double-mutant cells underwent division with kinetics that were similar to rad3 cells. Thus, the checkpoint is abolished in cds1 chk1 cells. Indeed, thecds1 chk1 double-mutant cells were acutely sensitive to killing by HU (Fig. 5B). These findings are consistent with a recent study (20) and support a model in which Cds1 has a direct role in enforcing the replication checkpoint.

Figure 5

Abolishment of replication checkpoint incds1 chk1 cells. (A) Synchronous cultures were obtained by centrifugal elutriation. Progression through mitosis in the presence of 12 mM HU was monitored by scoring binucleate or septated cells. The cds1 chk1 double-mutant (□) andrad3 (⧫) cells underwent mitosis in the presence of HU, whereas division was arrested in control strains [wild-type (○) andcds1 (▵) and chk1 (×) single-mutant cells]. (B) The cds1 chk1 double-mutant andrad3 cells were equally hypersensitive to killing by 12 mM HU. Wild-type and chk1 cells were not sensitive to HU, whereas cds1 cells were moderately sensitive.

These findings indicate that Cds1 and Chk1 are dual effectors of the replication checkpoint. Chk1 may regulate Cdc25, whereas Cds1 appears to phosphorylate Wee1 and is required to increase the abundance of Mik1. The effect of phosphorylation of Wee1 by Cds1 is unknown and remains to be confirmed with in vivo studies, but it is striking that the replication checkpoint both increases the kinase activity of Cds1 and stimulates binding to Wee1 in cell extracts. The large HU-induced increase in the abundance of Mik1 helps to explain whycdc25 wee1 double-mutant cells arrest division in response to HU and reaffirms the importance of Cdc2 Tyr15phosphorylation in the replication checkpoint (21). Cds1 shares substantial sequence identity with the kinase Rad53p, a protein involved in the DNA replication checkpoint inSaccharomyces cerevisiae (2). These facts suggest that studies of Cds1 and Rad53p may help to frame investigations of replication checkpoints in human cells.

  • * To whom correspondence should be addressed. E-mail: prussell{at}


View Abstract

Navigate This Article