Suppressors of Cdc25p Overexpression Identify Two Pathways That Influence the G2/M Checkpoint in Fission Yeast

Corresponding author: Kristi Chrispell Forbes, Department of Genetics, Harvard Medical School, 200 Longwood Ave., Boston, MA 02115..

Communicating editor: M. D. ROSE

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Checkpoints maintain the order of cell-cycle events. At G2/M, a checkpoint blocks mitosis in response to damaged or unreplicated DNA. There are significant differences in the checkpoint responses to damaged DNA and unreplicated DNA, although many of the same genes are involved in both responses. To identify new genes that function specifically in the DNA replication checkpoint pathway, we searched for high-copy suppressors of overproducer of Cdc25p (OPcdc25⁺), which lacks a DNA replication checkpoint. Two classes of suppressors were isolated. One class includes a new gene encoding a putative DEAD box helicase, suppressor of uncontrolled mitosis (sum3⁺). This gene negatively regulates the cell-cycle response to stress when overexpressed and restores the checkpoint response by a mechanism that is independent of Cdc2p tyrosine phosphorylation. The second class includes chk1⁺ and the two Schizosaccharomyces pombe 14-3-3 genes, rad24⁺ and rad25⁺, which appear to suppress the checkpoint defect by inhibiting Cdc25p. We show that rad24 mutants are defective in the checkpoint response to the DNA replication inhibitor hydroxyurea at 37° and that cds1 rad24 mutants, like cds1 chk1 mutants, are entirely checkpoint deficient at 29°. These results suggest that chk1⁺ and rad24⁺ may function redundantly with cds1⁺ in the checkpoint response to unreplicated DNA.

CONTROL mechanisms called checkpoints help to maintain the correct order of cell-cycle events (HARTWELL and WEINERT 1989 ). In wild-type cells, checkpoints ensure that later events are dependent upon the completion of earlier events. The G2/M checkpoint ensures that mitosis does not take place when DNA replication is incomplete or chromosomes are damaged. As a result of this checkpoint, wild-type fission yeast cells with damaged or incompletely replicated DNA undergo cell-cycle arrest at G2/M. The cells continue to grow while arrested, becoming highly elongated. In contrast, checkpoint mutants are unable to delay their cell cycle in response to incomplete DNA replication or DNA damage (ENOCH and NURSE 1990 ). They enter mitosis and form a spindle, but the unreplicated chromosomes remain in the center of the cell. Ultimately the nucleus is cleaved by the septum resulting in inviable cells. Such aberrant mitoses are referred to as "cuts" because they resemble the phenotype of cut^- mutants (HIRANO et al. 1986 ).

In fission yeast, inhibitory tyrosine phosphorylation of Cdc2p, the catalytic subunit of cyclin-dependent kinase, is required for the G2/M checkpoint (ENOCH and NURSE 1990 ; ENOCH et al. 1991 ; RHIND et al. 1997 ). The balance between the activities of the Cdc2p inhibitory kinases, Wee1p and Mik1p, and the Cdc2p activating phosphatases, Cdc25p and Pyp3p, determines whether a cell will pass the G2/M boundary (DUNPHY 1994 ). Cells in which this balance is altered are checkpoint defective. For example, a strain that constitutively overproduces Cdc25p (OPcdc25⁺) lacks the DNA replication checkpoint. A cdc2-3w strain, which renders Cdc2p activation independent of Cdc25p, is also defective in this checkpoint (ENOCH and NURSE 1990 ). A wee1-50 mik1 strain, which has greatly decreased tyrosine kinase activity because of a temperature-sensitive allele of wee1⁺ and the deletion of the mik1⁺ gene, is checkpoint defective even at the permissive temperature for wee1-50. At the nonpermissive temperature, these cells are inviable (SHELDRICK and CARR 1993 ).

A picture of the molecular events that may underlie the checkpoint response to damaged DNA is beginning to emerge. It is hypothesized that Rad3p kinase is activated in the presence of DNA damage. The products of five other genes termed the "checkpoint rad" genes, rad1⁺, rad9⁺, rad17⁺, rad26⁺, and hus1⁺, are also required for the early phase of the checkpoint response (HUMPHREY and ENOCH 1995 ). Many of these genes are evolutionarily conserved (STEWART and ENOCH 1996 ); for example, the rad3⁺ gene is functionally and structurally similar to the MEC1 and TEL1 genes from Saccharomyces cerevisiae and the human ATM gene, which is mutated in the severe cancer prone syndrome, ataxia-telangiectasia (LAVIN et al. 1995 ). Another protein kinase, Chk1p, is phosphorylated in a Rad3p-dependent manner (WALWORTH and BERNARDS 1996 ). In mammalian cells, and presumably also in fission yeast (FURNARI et al. 1997 ), Chk1p phosphorylates Cdc25p on serine residues, creating binding sites for 14-3-3 proteins (PENG et al. 1997 ; SANCHEZ et al. 1997 ). Binding of 14-3-3p inhibits Cdc25p activity by a mechanism that is not clear (PENG et al. 1997 ). In the absence of active Cdc25p, Cdc2p remains in a tyrosine phosphorylated, inactive conformation. Chk1p may also positively regulate Wee1p (O'CONNELL et al. 1997 ).

The checkpoint response to unreplicated DNA is less well understood. The response requires rad3⁺ and the other checkpoint rad genes, because mutations in these genes abolish cell-cycle arrest in the presence of the DNA synthesis inhibitor, hydroxyurea (HU). The response also requires tyrosine phosphorylation of Cdc2p (ENOCH et al. 1991 ). However, the mechanisms linking Rad3p to tyrosine phosphorylation of Cdc2p in response to the unreplicated DNA are not known. Although mutations in chk1⁺ or rad24⁺ abolish the response to damaged DNA, they do not ordinarily disrupt the response to unreplicated DNA (WALWORTH et al. 1993 ; AL-KHODAIRY et al. 1994 ), though a recent study shows that chk1 mutants have a partial replication checkpoint defect at high temperatures (FRANCESCONI et al. 1997 ). Other effectors may link Rad3p to the cell-cycle machinery when DNA replication is blocked. A possible effector is the kinase encoded by cds1⁺; however, cds1^- mutants arrest normally in HU although they lose viability rapidly (MURAKAMI and OKAYAMA 1995 ). Recently, BODDY et al. 1998 have shown that Cds1p phosphorylates Wee1p in vitro. However, the role of this interaction in checkpoint control has not been established.

To identify transducers of the incomplete DNA replication checkpoint signal in Schizosaccharomyces pombe, we overexpressed known checkpoint genes and evaluated their ability to suppress the checkpoint defect of OPcdc25⁺ in the presence of HU. We find that overexpression of chk1⁺ is able to suppress the HU sensitivity of OPcdc25⁺. To identify new genes specifically involved in the DNA replication checkpoint, we performed a screen for high-copy plasmid suppressors of the checkpoint defect of OPcdc25⁺. Two classes of suppressors were isolated, which appear to suppress OPcdc25⁺ by distinct mechanisms. One class includes a new gene encoding a putative DEAD box helicase, suppressor of uncontrolled mitosis (sum3⁺). This gene negatively regulates the cell-cycle response to stress when overexpressed and restores checkpoint response by a mechanism that is independent of Cdc2p tyrosine phosphorylation. The second class includes chk1⁺ and two genes encoding 14-3-3 proteins, rad24⁺ and rad25⁺, which appear to suppress the checkpoint defect by inhibiting Cdc25p. The finding of this class of genes was unexpected, as they were previously thought to be involved only in the response to damaged DNA, and indeed chk1^-, rad24^-, and rad25^- mutants arrest normally in HU at 29° (WALWORTH et al. 1993 ; AL-KHODAIRY et al. 1994 ). We show that rad24 mutants are defective in the checkpoint response to HU at 37° and that cds1 rad24 mutants, like cds1 chk1 mutants, are entirely checkpoint deficient at 29°. These results suggest that chk1⁺ and rad24⁺ may function redundantly with cds1⁺ in the checkpoint response to unreplicated DNA.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Growth of S. pombe:
Standard media and growth conditions were used as described (MORENO et al. 1991 ). Cells were transformed by electroporation as described (PRENTICE 1992 ). Phase-contrast micrographs were obtained using an Axiophot microscope (Carl Zeiss, Inc., Thornwood, NY) and a Photonic Microscope Image Processor C1966 (Hamamatsu Photonic Sys. Corp., Bridgewater, NJ). Cells were counted using a Coulter counter. All strains and plasmids are listed in Table 1. Cells were grown at 29° unless otherwise noted.

Screening for high-copy plasmid suppressors of OPcdc25⁺:
TE387 (OPcdc25⁺) was transformed with a LEU2 S. pombe cDNA library in which cDNA expression is regulated by the thiamine-repressible nmt1⁺ promoter (B. EDGAR and C. NORBURY, unpublished results; MAUNDRELL 1993 ). Leu⁺ transformants were selected on Edinburgh minimal media (EMM) plates with 2 µM thiamine, and colonies were replica plated to EMM plates without thiamine overnight to induce the nmt1⁺ promoter. Colonies were then replica plated to EMM plates with the vital dye, phloxine B (Fisher Scientific, Pittsburgh, PA) with and without 5 mM HU, and grown for 2 days. Colonies that grew in both the presence and absence of HU were examined microscopically. Out of the 89,000 transformants screened, 150 formed colonies in the presence and absence of HU. Plasmids were recovered from most of these transformants and were retested by a new transformation into the TE387 strain and replica plating as described. Twenty-nine transformants that were normal in size in the absence of HU and elongated in the presence of HU were analyzed further. The cDNA inserts of 19 plasmids (Table 2), which had consistently strong suppressor phenotypes, were sequenced.

Testing for suppression of OPcdc25⁺, cdc2-3w, or wee1-50 mik1:
Strains TE387, TE361, or TE386 (Table 1) were transformed with LEU2 REP plasmids containing the thiamine-repressible nmt1⁺ promoter (B. EDGAR and C. NORBURY, unpublished results; MAUNDRELL 1993 ). Leu⁺ transformants were selected on EMM plates with 2 µM thiamine, and colonies were replica plated to EMM plates without thiamine overnight to induce the nmt1⁺ promoter. Colonies were then replica plated to EMM plates with phloxine B, with and without 5 mM or 10 mM HU, and grown for 2 days at 29° (or 25° in the case of TE386). Cells were examined microscopically for viability and elongation phenotypes. To investigate cell number increase, TE 387 transformants were grown in liquid EMM with thiamine, washed once with EMM lacking thiamine, and inoculated in EMM lacking thiamine. Cells were fixed every 2 hr in formal saline and the cell number was determined using a Coulter counter. To demonstrate colony growth in the presence or absence of HU, Leu⁺ transformants were grown on EMM without thiamine for 24 hr to induce the nmt1⁺ promoter, then streaked on EMM plates with or without 5 mM HU, and grown for 5 days. Images of petri plates were captured using the Stratagene (La Jolla, CA) Eagle Eye II.

Description of sum2⁺ subclones:
To examine the "sum2N" and the 40S ribosomal p40 protein portions of the "sum2N+p40" fusion (pTE306) separately, each portion of the fusion was subcloned into a REP3X plasmid (MAUNDRELL 1993 ). Plasmid pTE452 (REP3X sum2N) contains nucleotides 1 to 661 of the fusion, including 224 amino acids of the sum2⁺ sequence, followed by an SphI-BamHI linker used to aid in cloning. Plasmid pTE458 (REP3X sum2p40) contains nucleotides 661 to 1660 (the 3' end) of the fusion, preceded by a XhoI-SphI linker used to aid in cloning. Plasmid pTE458 includes the entire open reading frame of the 40S ribosomal p40 protein gene. A complete clone of sum2⁺ on a REP3X vector was obtained (a gift from J. Bähler and J. Pringle, which we have called pTE462). The complete sum2⁺ was subcloned on a BamHI-SalI fragment into the vectors REP41X (pTE490), and REP81X (pTE491; MAUNDRELL 1993 ).

Nucleotide sequence accession number:
The cDNA sequence of sum3⁺ has been deposited with GenBank under accession number AF025536. The complete sequence of sum2⁺ can be found under accession number D89169.

RNA analysis:
Strain TE235 (wild type) transformed with plasmids pTE101, pTE301, and pTE304 (Table 1) and strain TE640 (sty1^-/spc1^-) transformed with pTE101 were grown to midlog phase in EMM media. KCl was added to a final concentration of 1 M to half of each culture 60 min before harvesting. The pelleted cells were lysed with glass beads in 1 ml of solution containing 0.32 M sucrose, 20 mM Tris-HCl (pH 7.5), and 10 mM EDTA and diluted with 4 ml of the above solution containing 1% SDS. Phenol extraction was performed at 60° for 3 min followed by further phenol/chloroform extraction at 22°, and total RNA was ethanol precipitated. For Northern hybridizations, 7 µg of total RNA was separated on a denaturing formaldehyde agarose gel (RAVE et al. 1979 ). Following transfer to nitrocellulose (GeneScreen; New Life Science Products, Boston), the bound RNA was hybridized to either gpd1⁺ (PIDOUX et al. 1990 ) or act1⁺ (MERTINS and GALLWITZ 1987 ) probes that were generated as described (HUMPHREY and ENOCH 1998 ). Probes were labeled as described (FEINBERG and VOGELSTEIN 1984 ).

Analysis of phosphotyrosine levels of Cdc2 protein:
Strain TE235 (wild type) and TE22 (hus1-14) were grown to midlog phase in EMM at 29°. HU was added to a final concentration of 10 mM at t = 0, and cells were harvested at t = 0, 2, 4, and 6 hr. Pelleted cells were lysed with glass beads (Sigma Chemical Co., St. Louis) into lysis buffer H containing 0.1% NP-40, 10% glycerol, 50 mM Tris-HCl (pH 7.5), 15 mM EDTA, 100 mM sodium chloride, 0.1 mM NaF, 2 mM sodium orthovanadate, 10 µg/ml aprotinin, 10 µg/ml leupeptin, 10 µg/ml pepstatin A, 20 µM TPCK, 1 mM PMSF, 60 mM ß-glycerophosphate, 15 mM paranitrophenol phosphate, and 1 µM okadaic acid. The Cdc2 protein was isolated by affinity purification using p13^suc1 beads (BRIZUELA et al. 1987 ). Protein was resolved by 12% SDS-PAGE and transferred electrophoretically to a nitrocellulose membrane. The membrane was immunoblotted with anti-Cdc2p (PN24) and anti-pTyr (4G10; Upstate Biotechnology, Lake Placid, NY) antibodies, which were detected by enhanced chemiluminescence (ECL; Amersham, Arlington Heights, IL) according to the manufacturer's instructions. Between primary antibodies, the membrane was stripped according to the manufacturer's instructions (Amersham). Bands were quantitated using densitometry [Molecular Dynamics (Sunnyvale, CA) ImageQuant program].

To investigate the effects of overexpressing suppressor genes on Cdc2p tyrosine phosphorylation, strain TE235 was transformed with plasmids pTE101, pTE301, pTE302, pTE303, and pTE304 (Table 1). These strains were grown to midlog phase in EMM media for 22 hr in media with 2 µM thiamine to repress the nmt1⁺ promoter, or without thiamine to derepress the nmt1⁺ promoter. Tyrosine phosphorylated Cdc2p was measured as described above.

Western blot analysis of Cdc25p:
Strain TE235 (wild type) was transformed with plasmids pTE102, pTE170, pTE301, pTE303, pTE304, and pTE413 (Table 1). These strains were grown to midlog phase in EMM media for 22 hr with 2 µM thiamine to repress the nmt1⁺ promoter, or without thiamine to derepress the nmt1⁺ promoter. Strain TE79 (cdc25 cdc2-3w) was grown to midlog phase in YE5S media. HU was added to all the cultures to a final concentration of 10 mM 3 hr before harvesting. Pelleted cells were lysed with glass beads (Sigma) in 2x Laemmli buffer and boiled immediately. The proteins were resolved by 10% SDS-PAGE and transferred electrophoretically to an Immobilon P membrane (Millipore Corp., Bedford, MA). The membrane was immunoblotted with anti-Cdc25p antibody (BP2 serum, gift of Sergio Moreno), and then blotted with anti-rabbit secondary antibody (Amersham), which was detected by enhanced chemiluminescence (ECL; Amersham) according to the manufacturer's instructions.

Analysis of HU and ultraviolet (UV) response:
Wild-type (TE235), rad24 (TE465), cds1 (TE700), chk1 (TE548), rad3 (TE890), chk1 rad24 (TE 922), cds1 rad24 (TE919), and cds1 chk1 (TE856) cells were grown to midlog phase in rich media and HU was added to a final concentration of 10 mM. The culture was divided in half, with half the cells remaining at 29° and half the cells being shifted to 37°. Samples from each half were collected at 2-hr intervals. Cells were fixed for microscopy and analyzed for "cut" formation as previously described (ENOCH et al. 1992 ). At least 100 cells were counted at each time point. For analysis of the response to UV radiation, the same strains were grown to early log phase, plated on YE5S, and irradiated with increasing doses of UV. The number of colonies growing on two plates after 5 days was counted for each viability measurement.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Overexpression of chk1⁺ restores OPcdc25⁺ checkpoint function during HU treatment:
To learn more about the transducers of the incomplete DNA replication checkpoint signal in S. pombe, high-copy suppressors of the checkpoint defect of strains overexpressing Cdc25p (OPcdc25⁺) were sought. OPcdc25⁺ mutants are unable to delay cell-cycle progression in the presence of unreplicated DNA and therefore attempt to undergo mitosis and segregate a single set of chromosomes when treated with HU, an inhibitor of ribonucleotide reductase (ENOCH and NURSE 1990 ). In these inappropriately dividing cells, the septum bisects the single nucleus. This results in aneuploid or anucleate cells with phenotypes resembling the morphology of cut^- mutants (HIRANO et al. 1986 ). Consequently, OPcdc25⁺ cells fail to form colonies in the presence of HU. In contrast, wild-type cells initially undergo cell-cycle arrest in the presence of HU, and then resume the cell cycle with a longer S-phase, forming slowly growing colonies consisting of highly elongated cells.

Overexpression of upstream components of the G2/M checkpoint pathway or positive regulators of the checkpoint response might be expected to amplify the checkpoint signal and thus suppress this defect. Negative regulators of Cdc25p or positive regulators of Wee1p or Mik1p might also suppress OPcdc25⁺. Overexpression of such suppressors may allow growth of OPcdc25⁺ on HU as elongated cells. These can be distinguished from genes that counteract the effects of HU, such as the catalytic subunit of ribonucleotide reductase, because cells overexpressing genes that counteract the effects of HU divide at a normal length on HU. Suppressors can also be distinguished from general negative regulators of the cell cycle, because those genes block cell division both in the presence and absence of HU (HUMPHREY and ENOCH 1998 ).

Before performing a screen for new genes, we overexpressed some known checkpoint genes and evaluated their ability to suppress the checkpoint defect of OPcdc25⁺ in the presence of HU. OPcdc25⁺ mutant cells were transformed with the rad1⁺, rad3⁺, rad9⁺, rad17⁺, rad26⁺, chk1⁺, cds1⁺, and hus1⁺ genes under the control of the thiamine-repressible nmt1⁺ promoter on a vector carrying the LEU2 gene (see MATERIALS AND METHODS; Table 1). Leu⁺ transformants were selected in the presence of thiamine. High-level expression of each gene was then activated by replica plating cells to media lacking thiamine for at least 18 hr, after which cells were replica plated to media without thiamine and with or without 10 mM HU. Cells were examined microscopically for viability and elongation phenotypes.

Overexpression of the rad1⁺, rad3⁺, rad9⁺, rad17⁺, rad26⁺, cds1⁺, or hus1⁺ genes did not allow OPcdc25⁺ transformants to survive in the presence of HU better than the vector control (Figure 1). Overexpression of rad26⁺ was somewhat difficult to evaluate as cells overexpressing this gene did not grow well. Only the overexpression of chk1⁺ permitted growth of OPcdc25⁺ in the presence of HU (Figure 1). This is surprising because Chk1p is thought to be specifically responsible for transmitting the checkpoint signal for DNA damage, and chk1^- cells arrest normally in HU (WALWORTH et al. 1993 ; AL-KHODAIRY et al. 1994 ; WALWORTH and BERNARDS 1996 ). Because examination of some known checkpoint genes detected only a gene involved in the DNA damage checkpoint pathway, but not genes specifically involved in the DNA replication checkpoint, a screen for new genes was undertaken.

View larger version (39K): In this window In a new window Download PPT slide   Figure 1. Overexpression of chk1+ rescues the checkpoint defect of OPcdc25+. OPcdc25+ (TE387) was transformed with the plasmids rad1+ (pTE567), rad3+ (pTE157), rad9+ (pTE479), rad17+ (pTE478), rad26+ (pTE169), hus1+ (pTE32), cds1+ (pTE531), chk1+ (pTE170), or vector control (pTE102, pREP1), where expression of each gene was controlled by the thiamine-repressible nmt1+ promoter (MAUNDRELL 1993 ). Each transformant was grown in the absence of thiamine for 24 hr to induce the nmt1+ promoter, and then patched on EMM in the absence (top) or presence (bottom) of 5 mM HU and grown for 5 days at 29°.

Isolation of suppressor genes that restore OPcdc25⁺ checkpoint function during hydroxyurea treatment:
To identify new genes specifically involved in the DNA replication checkpoint, OPcdc25⁺ was transformed with an S. pombe cDNA library under the control of the thiamine-repressible nmt1⁺ promoter on a vector encoding the S. cerevisiae LEU2 gene (B. EDGAR and C. NORBURY, unpublished results; MAUNDRELL 1993 ). Plasmids that allowed the cells to form colonies on minimal media plates in the presence of HU following derepression of the nmt1⁺ promoter (see MATERIALS AND METHODS) were identified. Colonies were examined microscopically, and transformants that showed an elongated phenotype in the presence of HU were selected. Genes that counteract the effects of HU, such as the catalytic subunit of ribonucleotide reductase, should not be identified by this screening method because cells overexpressing those genes divide at a normal length on HU rather than as elongated cells (K. CHRISPELL FORBES, T. HUMPHREY and T. ENOCH, unpublished observations). Plasmids that caused cell-cycle arrest in the absence of HU, or plasmids that suppressed by inhibiting septation, were not examined further. A total of 89,000 OPcdc25⁺ transformants were screened, and 29 plasmids were chosen for further study.

Further analysis of the cDNA inserts of the suppressors showed that the most frequently isolated suppressor had been identified 10 times, and the second most common suppressor was identified 4 times (Table 2). While this study was in progress, these two suppressor genes were independently identified as the two S. pombe 14-3-3 genes, rad25⁺ and rad24⁺, respectively (FORD et al. 1994 ).

Disrupting rad24⁺ activity reduces the DNA damage checkpoint response, making the cells radiation sensitive (FORD et al. 1994 ), but was not reported to affect the response to unreplicated DNA. It is intriguing that the 14-3-3 proteins, which were hypothesized to be involved in the damage checkpoint response but not in the DNA replication checkpoint, were isolated in a screen for checkpoint genes responding to incomplete replication. Because the results of this screen indicate that overexpression of rad24⁺ and rad25⁺ enhances the checkpoint signal, while other studies have shown that lack of rad24⁺ causes a loss of the DNA damage checkpoint (FORD et al. 1994 ), it seems possible that these proteins are directly involved in transducing the checkpoint signal.

In addition to the plasmids containing the 14-3-3 genes, two other sets of plasmids were found to allow OPcdc25⁺ to grow in the presence of HU (Table 2). Sequencing revealed that one set of plasmids contains an identical artifactual fusion, which we call sum2N+p40. The N-terminal portion of each clone contained the first 372 nucleotides of a 1113-nucleotide open reading frame (ORF) similar to the S. cerevisiae SCD6 gene, a multicopy suppressor of clathrin deficiency (D. GELPERIN and S. LEMMON, personal communication). We have named this S. pombe SCD6-related gene sum2⁺. The C terminus of the fusion contains the complete coding sequence of a gene closely related to 40S ribosomal p40 proteins, in a different reading frame from the truncated sum2⁺ gene. To determine which ORF was responsible for the suppressor phenotype, the N- and C-terminal ORFs were subcloned into REP vectors (MATERIALS AND METHODS; Table 1; MAUNDRELL 1993 ). In addition, a complete clone of sum2⁺ (pTE462, gift of Jurg Bähler and John Pringle) was studied. Transformation of these plasmids into checkpoint-deficient yeast showed that the truncated sum2⁺, sum2N, gave a poor rescue of OPcdc25⁺ in the presence of HU (data not shown). The 40S ribosomal p40 gene did not allow OPcdc25⁺ to survive treatment with HU (pTE452, data not shown). The complete clone of sum2⁺ was lethal when highly overexpressed (pTE462, REP3X vector), and was unable to allow OPcdc25⁺ to survive treatment with HU when overexpressed at lower levels (pTE490, REP41X vector, and pTE491, REP81X vector, data not shown). We conclude that the N-terminal fragment of sum2⁺, sum2N, has some capacity to rescue the OPcdc25⁺ checkpoint defect, which is enhanced by fusion to the 40S ribosomal p40 protein for unknown reasons. Because the full-length gene does not appear to have suppressing activity, the biological significance of suppression by sum2N is not clear.

Surprisingly, chk1⁺ was not isolated in this screen. This may be because it is not sufficiently represented in the library that was used. Alternatively, it may have been discarded because OPcdc25⁺ cells overexpressing chk1⁺ are not as elongated in HU as cells overexpressing rad24⁺, rad25⁺, or sum3⁺.

sum3⁺ encodes a member of the DEAD box helicase family:
Three other plasmids isolated in the screen contained a novel ORF that we have called sum3⁺ (Table 2). Sequencing of the cDNA inserts of these sum3⁺ plasmids revealed a 1908-nucleotide ORF encoding a 636-amino-acid protein, with a predicted molecular weight of ~70 kD. Comparisons of sum3⁺ with GenBank sequences show that sum3⁺ encodes a member of the DEAD box family of ATP-dependent RNA helicases. Members of this family share nine regions of amino acid conservation, including the Asp-Glu-Ala-Asp (DEAD) motif (FULLER-PACE 1994 ). The predicted protein product of sum3⁺ contains these motifs. The predicted protein product of sum3⁺ shows 74% amino acid similarity and 61% identity to DED1, its closest S. cerevisiae homologue (Figure 2). Like DED1, the sum3⁺ gene is essential (STRUHL 1985 ; B. GRALLERT and K. LABIB, personal communication).

View larger version (49K): In this window In a new window Download PPT slide   Figure 2. Alignment of the amino acid sequence of S. pombe sum3+ predicted protein product with that of DED1p from S. cerevisiae. Identical residues are indicated by white lettering on a black background. The sequence alignment was performed using the DNASTAR Megalign program. The cDNA sequence of sum3+ has been deposited with GenBank under accession number AF025536.

The suppressors enhance the DNA replication checkpoint response of OPcdc25⁺:
To determine whether the suppressors isolated in the screen acted as general negative regulators of cell division or if they were specifically restoring the checkpoint response, the phenotypes of OPcdc25⁺ cells overexpressing the suppressors were studied. If the suppressors were acting as general inhibitors of the cell cycle, they should have caused the cells to become elongated and should have blocked division even in the absence of HU. OPcdc25⁺ cells were transformed with a vector control, rad24⁺ or sum3⁺, where expression of each gene was controlled by the thiamine-repressible nmt1⁺ promoter. Cells were grown in the absence of thiamine to induce the nmt1⁺ promoter, and the cell number was counted every 2 hr. The number of cells in the samples overexpressing rad24⁺ or sum3⁺ continued to increase as rapidly as the number of cells containing the vector control, indicating that the cells overexpressing the suppressors were not arrested or significantly delayed in their cell cycles (Figure 3A; data for the first 12 hr after the removal of thiamine is not shown). In general, the length of the cells overexpressing rad24⁺ or sum3⁺ was comparable to the length of cells containing the vector control (Figure 3B, -HU), demonstrating that the cell cycle was not being delayed in these cells. Results for cells overexpressing rad25⁺ or chk1⁺ were similar (data not shown). OPcdc25⁺ cells overexpressing rad24⁺, rad25⁺, or sum3⁺ are able to form colonies (Figure 4A, -HU), which also suggests they are not cell-cycle arrested.

View larger version (62K): In this window In a new window Download PPT slide   Figure 3. Suppressors rescue the checkpoint defect of OPcdc25+ but do not affect proliferation in the absence of HU. (A) OPcdc25+ cells overexpressing rad24+ or sum3+ continue to divide in the absence of HU. OPcdc25+ (TE387) was transformed with the vector (pTE102), rad24+ (pTE303), or sum3+ (pTE304), where expression of each gene was controlled by the thiamine-repressible nmt1+ promoter (MAUNDRELL 1993 ). Cells were grown in the absence of thiamine to induce the nmt1+ promoter. Samples were fixed every 2 hr and the number of cells was counted using a Coulter counter. (B) Overexpression of rad24+ or sum3+ restores OPcdc25+ checkpoint function during hydroxyurea treatment. The transformants in A were grown in the absence of thiamine for 12 hr to induce the nmt1+ promoter. HU (10 mM) was added to half of the cells. Cells in the presence or absence of HU were fixed 6 hr later, stained with DAPI, and photographed. Arrows indicate cuts.

View larger version (59K): In this window In a new window Download PPT slide   Figure 4. rad24+ and rad25+ interact specifically with cdc25+. (A) Overexpression of rad24+, rad25+, or sum3+ restores OPcdc25+ checkpoint function during HU treatment. OPcdc25+ (TE387) transformed with vector (pTE102), rad24+ (pTE303), rad25+ (pTE302), or sum3+ (pTE304) was grown in the absence of thiamine for 24 hr to induce the nmt1+ promoter, then patched on EMM in the absence (top) or presence (bottom) of 5 mM HU, and grown for 5 days at 29°. (B) Overexpression of sum3+ restores cdc2-3w checkpoint function during HU treatment, but overexpression of rad24+ or rad25+ does not. cdc2-3w (TE361) was transformed with the same plasmids and grown as described in A.

To determine whether overexpression of the suppressors isolated in the screen was enhancing the checkpoint response of OPcdc25⁺, we studied the phenotypes of OPcdc25⁺ cells overexpressing rad24⁺ or sum3⁺ in the presence and absence of HU. Overexpression of vector, rad24⁺ or sum3⁺ was induced by the removal of thiamine, and HU was added to half the cells for 6 hr. As shown in Figure 3B, OPcdc25⁺ cells containing a vector control do not elongate, and many form cuts in the presence of HU because the cells lack the unreplicated DNA checkpoint (Figure 3B, +HU; arrows indicate cuts). In contrast, OPcdc25⁺ cells overexpressing rad24⁺ or sum3⁺ in the presence of HU did not cut and show an elongated phenotype indicative of cell-cycle arrest (Figure 3B, +HU). Cells overexpressing rad25⁺ are indistinguishable from those overexpressing rad24⁺ (data not shown). These results show that the suppressors enhance the checkpoint response of OPcdc25⁺ in the presence of HU, but do not affect cell-cycle progression under normal conditions. However, all of the suppressors caused significant cell-cycle delays when overexpressed in wild-type cells, indicating that the suppressors inhibit cell-cycle progression when not counteracted by mutations in cell-cycle regulators (data not shown).

rad24⁺ and rad25⁺ restore checkpoint control through a specific interaction with cdc25⁺:
To investigate the mechanism of action of the suppressors, it was of interest to determine whether they could restore checkpoint control when expressed in other checkpoint mutants. Like OPcdc25⁺, cdc2-3w or wee1-50 mik1 cells lack a checkpoint response to unreplicated DNA and are inviable when replica plated onto media containing HU, even at the permissive temperature for wee1-50 mik1. As we have stated, overexpression of rad24⁺, rad25⁺, or sum3⁺ restores checkpoint function in OPcdc25⁺ (Figure 4A, +HU). To determine whether the suppressors could also rescue cdc2-3w or wee1-50 mik1 in the presence of HU, plasmids containing rad24⁺, rad25⁺, chk1⁺, sum3⁺, and sum2N+p40 were transformed into these strains.

cdc2-3w or wee1-50 mik1 cells overexpressing rad24⁺, rad25⁺, or chk1⁺ did not form colonies in the presence of HU (Figure 4B and data not shown). In contrast, overexpression of sum2N+p40 or sum3⁺ allowed cdc2-3w or wee1-50 mik1 to survive in the presence of HU (Figure 4B and data not shown). Therefore, we conclude that the suppressors we have identified restore checkpoint control by two different mechanisms. The finding that rad24⁺, rad25⁺, and chk1⁺ worked specifically to rescue OPcdc25⁺ suggests that these genes may act to negatively regulate Cdc25p. In contrast, sum2N+p40 and sum3⁺ apparently suppress checkpoint defects by a more global mechanism as they are able to suppress several checkpoint mutants.

Overexpression of sum3⁺ negatively regulates the cell-cycle response to osmotic stress:
When wild-type fission yeast are exposed to conditions of high osmolarity, their entry into mitosis is stimulated (MILLAR et al. 1995 ). This has two beneficial effects: the cells enter G1, so they can mate or sporulate, and cell size is reduced, reducing the metabolic requirements for doubling. This response requires the stress-activated sty1⁺/spc1⁺ MAPK pathway (Figure 5A), but the molecular mechanism of the stress-activated MAPK's effects on the cell cycle are not understood. Loss-of-function mutants for the kinases in this pathway or strains overexpressing either of the two phosphatases (pyp1⁺ and pyp2⁺) that negatively regulate sty1⁺/spc1⁺ MAPK cannot stimulate mitosis, and become abnormally long under conditions of high osmolarity (MILLAR et al. 1992 ; SHIOZAKI and RUSSELL 1995 ). We have recently identified a gene, sum1⁺, that suppresses the checkpoint defect of cdc2-3w and also inhibits the osmotic cell-cycle response (HUMPHREY and ENOCH 1998 ). Like overexpression of sum3⁺, overexpression of sum1⁺ allows survival of cdc2-3w, OPcdc25⁺, or wee1-50 mik1 in the presence of HU. Moreover, we have shown that overexpression of pyp1⁺, or mutation of sty1⁺/spc1⁺, suppresses the checkpoint defect of cdc2-3w (HUMPHREY and ENOCH 1998 ). These results suggest that the stress response pathway may counteract checkpoint controls. To investigate whether any of the suppressors we had identified restored checkpoint control through a similar mechanism, we examined wild-type cells overexpressing rad24⁺, rad25⁺, sum2N+p40, and sum3⁺ under conditions of osmotic stress. Like mutants lacking sty1⁺/spc1⁺ activity, sum3⁺ overexpressors were found to be elongated on 1 M KCl (Figure 5B) or 1.5 M sorbitol (data not shown). The phenotype of sum2N+ p40 is similar (data not shown). This phenotype is similar to that of cells overexpressing pyp1⁺ (Figure 5B, middle; SHIOZAKI and RUSSELL 1995 ), although the cells in which sum3⁺ is overexpressed look somewhat wider. Thus, like sum1⁺ (HUMPHREY and ENOCH 1998 ), overexpression of sum3⁺ and sum2N+p40 apparently inhibits the stress response to high-osmolarity conditions. Overexpression of rad24⁺, rad25⁺, and chk1⁺ did not cause elongation under these conditions, which confirms that they are likely to restore checkpoint control by a different mechanism (data not shown).

View larger version (51K): In this window In a new window Download PPT slide   Figure 5. sum3+ negatively regulates the stress response. (A) The stress-activated MAPK pathway in fission yeast. Stress activates a kinase cascade (not shown) leading to activation of the MAPK, Sty1p/Spc1p. Active Sty1p/Spc1p phosphorylates Atf1p, which promotes the transcription of genes including gpd1+ and advances mitosis through an unidentified mechanism (G2 M). A plausible position for Sum3p, downstream of Sty1p/Spc1p on the branch of the pathway that leads to activation of mitosis, is shown (dashed ). (B) Overexpression of sum3+ disrupts the cell-cycle response to osmotic stress. Wild-type cells (strain TE235) with vector (pTE101) or overexpressing sum3+ (pTE304) or pyp1+ (pTE304) were grown on EMM plates containing thiamine, replica plated onto EMM plates for 24 hr to derepress the nmt1+ promoter, and then replica plated to EMM plates containing 1 M KCl. Cells were photographed after 24 hr. (C) Northern blot analysis of gpd1+ transcription levels in wild-type cells (strain TE235) overexpressing vector (pTE101), sum3+ (pTE304), or pyp1+ (pTE301) and sty1-/spc1- (strain TE640) cells with vector (pTE101). Overexpression of sum3+ does not inhibit the transcriptional response of the MAPK pathway. KCl was added to a final concentration of 1 M for 60 min where indicated (KCl, + lanes). Total RNA generated from the above strains was separated on a Northern gel, blotted to nitrocellulose, and transcripts visualized with gpd1+ and act1+ probes as indicated.

Overexpression of sum3⁺ does not inhibit the stress-induced transcriptional response:
The osmotic stress phenotype of overexpressed sum3⁺ suggests that Sum3p could be negatively regulating the stress response pathway. For example, Sum3p could negatively regulate one of the MAP kinases, or it could positively regulate pyp1⁺ or pyp2⁺. As shown in Figure 5A, the stress response pathway bifurcates after activation of the Sty1p/Spc1p MAPK. One branch, requiring the transcription factor Atf1p, leads to transcriptional activation of genes required for the stress response (DEGOLS et al. 1996 ; SHIOZAKI and RUSSELL 1996 ; WILKINSON et al. 1996 ). A second branch leads to stimulation of cell division (MILLAR et al. 1995 ), although the components of this branch have not yet been identified. In atf1^- mutants, the transcriptional response to stress is abolished, but the cell-cycle response is not affected (SHIOZAKI and RUSSELL 1996 ; WILKINSON et al. 1996 ). Loss-of-function mutations in any of the MAPK genes, or overexpression of pyp1⁺ or pyp2⁺, causes Atf1p-dependent transcription to be uninducible (DEGOLS et al. 1996 ; SHIOZAKI and RUSSELL 1996 ; WILKINSON et al. 1996 ). If Sum3p were negatively regulating one of the kinases or positively regulating pyp1⁺ or pyp2⁺, then overexpression of sum3⁺ should prevent expression of the genes downstream of Atf1p, such as gpd1⁺, even in the presence of stress.

A Northern blot of total RNA from wild-type cells grown in the presence or absence of osmotic stress was probed with gpd1⁺ (Figure 5C, top) and reprobed with actin as a loading control (Figure 5C, bottom). In wild-type cells carrying a vector control, transcription of gpd1⁺ was strongly induced by exposure to 1 M KCl, whereas wild-type cells overexpressing pyp1⁺, or sty1^-/spc1^- cells, showed no increase in gpd1⁺ transcription. Overexpression of sum3⁺ did not reduce the induction of gpd1⁺ transcription (Figure 5C). This result suggests that overexpression of sum3⁺ interferes with the stress response at a point downstream of MAP kinase activation, on the branch of the pathway that leads to activation of mitosis (Figure 5A). Similar results were found for sum1⁺ (HUMPHREY and ENOCH 1998 ). Overexpression of rad24⁺, rad25⁺, or chk1⁺ allowed normal induction of gpd1⁺ transcription (data not shown).

Activation of the G2/M checkpoint leads to tyrosine phosphorylation of Cdc2p in wild-type cells, but not in hus1^- mutants:
rad24⁺, rad25⁺, and chk1⁺ interact genetically with cdc25⁺. Cdc25p is required for removing an inhibitory tyrosine phosphate from Cdc2p. We conjecture that in the presence of unreplicated DNA, these gene products and others inhibit Cdc25p, thus resulting in an accumulation of tyrosine-phosphorylated Cdc2p. Genetic studies are consistent with this model (ENOCH and NURSE 1990 ; RHIND et al. 1997 ); however, biochemical studies have challenged this view with the finding that Cdc2p kinase activity is maintained in G2/M-arrested cells after treatment with HU (KNUDSEN et al. 1996 ). To test whether Cdc2p tyrosine phosphorylation increases in response to unreplicated DNA, we examined Cdc2p tyrosine phosphorylation in wild-type cells and in the checkpoint-deficient mutant hus1^- (Figure 6A). In wild-type cells, a fourfold increase in tyrosine-phosphorylated Cdc2p can be seen at 4 and 6 hr after the addition of HU, while in the checkpoint-defective hus1^- cells there is no increase in tyrosine-phosphorylated Cdc2p relative to the total amount of Cdc2p (Figure 6A). These data support the model that incompletely replicated DNA activates a checkpoint in wild-type cells that causes Cdc2p to become phosphorylated on tyrosine. Similar results were reported by others while this article was under review (RHIND and RUSSELL 1998 ).

View larger version (31K): In this window In a new window Download PPT slide   Figure 6. Stimulation of Cdc2 tyrosine phosphorylation by rad24+, rad25+, and chk1+. (A) Activation of the G2/M checkpoint by HU increases tyrosine phosphorylation of Cdc2p in wild-type cells, but not in checkpoint mutants. Wild-type cells (TE235) and hus1-14 (TE22) mutant cells were grown to midlog phase in EMM at 29°. HU was added to a final concentration of 10 mM, and cells were harvested every 2 hr. Cdc2p was isolated by affinity purification, resolved by SDS-PAGE, and transferred to a nitrocellulose membrane. The blot was probed sequentially with a phosphotyrosine antibody and a Cdc2p antibody. (B) Wild-type cells (TE235) were transformed with vector (pTE102), rad24+ (pTE303), sum3+ (pTE304), and pyp1+ (pTE301). These strains were grown to midlog phase with thiamine (promoter off, - lanes) or without thiamine (promoter on, + lanes). Cdc2p was analyzed as described above. Phosphotyrosine and total Cdc2p were detected by sequential immunoblotting, and the bands were quantitated using a densitometer (Molecular Dynamics ImageQuant program). The amount of the phosphotyrosine signal was normalized to the amount of total Cdc2p, and the ratio of phosphotyrosine to total Cdc2p with the promoter repressed was set equal to one.

The two classes of suppressors have different effects on tyrosine phosphorylation of Cdc2p in wild-type cells:
One mechanism by which suppressors could restore the G2/M checkpoint would be to increase tyrosine phosphorylation of Cdc2p. To see if this was the case, plasmids overexpressing rad24⁺, rad25⁺, chk1⁺, pyp1⁺, and sum3⁺ were transformed into wild-type cells, and the effects on Cdc2p tyrosine phosphorylation were evaluated. Unlike OPcdc25⁺ cells overexpressing these genes, wild-type cells overexpressing these genes become elongated in the absence of HU (data not shown). Cdc2p was affinity purified from cellular lysates using p13^suc1 beads (BRIZUELA et al. 1987 ), resolved by SDS-PAGE, and transferred to an Immobilon membrane that was sequentially probed with an antibody to phosphotyrosine and an antibody to Cdc2p. The amount of tyrosine-phosphorylated Cdc2p and total Cdc2p was quantitated using a densitometer, and the amount of tyrosine-phosphorylated Cdc2p was divided by the amount of total Cdc2p. To normalize the results, the ratio of tyrosine-phosphorylated Cdc2p to total Cdc2p with the promoter off was defined as equal to one. Typical results are shown in Figure 6B. Qualitatively similar results were obtained in several independent experiments.

Induction of a vector control has no effect on Cdc2p tyrosine phosphorylation (Figure 6B). When rad24⁺ or rad25⁺ is overexpressed in wild-type cells, the amount of tyrosine-phosphorylated Cdc2p increases; similar results are seen for chk1⁺ (data not shown). Thus, these suppressors can stimulate tyrosine phosphorylation of Cdc2p. This is consistent with the hypothesis that this class of suppressors negatively regulates Cdc25p. Stimulation of Cdc2p tyrosine phosphorylation by rad24⁺ and rad25⁺ in OPcdc25⁺ cells was much weaker (data not shown). This is not surprising because the excess of Cdc25p in these strains should make the activity of rad24⁺ and rad25⁺ more difficult to detect. We also did not see consistent differences in Cdc2p tyrosine phosphorylation after HU treatment in cells overexpressing any of the suppressors compared to control cells (data not shown). This is probably because HU treatment alone stimulates Cdc2p tyrosine phosphorylation substantially, making it difficult to detect any further enhancement due to overexpression of suppressors.

Overexpression of sum3⁺ does not cause an increase in Cdc2p tyrosine phosphorylation (Figure 6B). This suggests that this class of suppressors regulates cell-cycle progression by a mechanism independent of tyrosine phosphorylation of Cdc2p. Overexpression of pyp1⁺ also does not increase in Cdc2p phosphorylation (Figure 6B), suggesting that inactivation of the stress-activated MAPK pathway arrests cells at G2/M by a mechanism independent of tyrosine phosphorylation of Cdc2p.

Overexpression of rad24⁺, rad25⁺, or chk1⁺ increases levels of Cdc25p:
To investigate the effects of rad24⁺, rad25⁺, chk1⁺, and sum3⁺ on the biochemical status of Cdc25p, cell lysates of wild-type cells overexpressing each of these genes were prepared and analyzed by Western blotting for Cdc25p. Inducing overexpression of a vector control or sum3⁺ has no major effect on the level or mobility of Cdc25p (Figure 7). In contrast, inducing overexpression of rad24⁺, rad25⁺, or chk1⁺ leads to a marked increase in the levels of Cdc25p. No further increase is observed in response to HU (Figure 7). Overexpression of rad24⁺ or chk1⁺ also results in increased Cdc25p accumulation in OPcdc25⁺ cells (data not shown). Cdc25p appears to shift to a slower migrating form in the presence of HU in cells containing a vector control or overexpressing sum3⁺ (Figure 7, +HU). In cells overexpressing rad24⁺, rad25⁺, or chk1⁺, Cdc25p accumulates to such high levels that it is not possible to determine if its mobility is altered in response to HU.

View larger version (50K): In this window In a new window Download PPT slide   Figure 7. Overexpression of rad24+, rad25+, or chk1+ increases levels of Cdc25p. Wild-type cells (TE235) overexpressing vector (pTE102), chk1+ (pTE170), rad24+ (pTE303), rad25+ (pTE302), and sum3+ (pTE304) were grown in media with thiamine (promoter off, - lanes) or without thiamine (promoter on, + lanes). HU was added to the cultures to a final concentration of 10 mM 3 hr before harvesting (HU, + lanes). Total proteins were resolved by 10% SDS-PAGE and transferred to an Immobilon P membrane, stained with Ponceau S to ensure equal loading, and then immunoblotted with anti-Cdc25p antibody. cdc25 (TE79) and wild-type cells overexpressing Cdc25p (pTE413) are shown as controls.

The accumulation of Cdc25p that was observed in wild-type cells overexpressing rad24⁺, rad25⁺, or chk1⁺ could be the result of increased transcription or translation of Cdc25p, or perhaps the stabilization or sequestration of the Cdc25p, so that it is not degraded at normal rates. The accumulation of Cdc25p cannot explain the suppression of the checkpoint defect by these suppressors, because increased Cdc25p levels would be predicted to stimulate mitosis and reduce the efficiency of checkpoint control. Therefore, we believe that overexpression of rad24⁺, rad25⁺, or chk1⁺ may cause Cdc25p to be sequestered in an inactive state where it is less susceptible to proteolysis, which is a normal part of its regulation (MORENO et al. 1990 ; KOVELMAN and RUSSELL 1996 ; NEFSKY and BEACH 1996 ).

Loss of chk1⁺ or rad24⁺ function compromises the checkpoint response to unreplicated DNA:
Our results suggest that Chk1p and 14-3-3 proteins could function in the checkpoint response to unreplicated DNA by inhibiting Cdc25p activity, which prevents Cdc2p tyrosine dephosphorylation and thus blocks mitosis. However, previous studies have shown that chk1 and rad24 mutants arrest normally in response to unreplicated DNA, although they lack a checkpoint response to damaged DNA (WALWORTH et al. 1993 ; AL-KHODAIRY et al. 1994 ; FORD et al. 1994 ). While this study was in progress, FRANCESCONI et al. 1997 demonstrated that chk1 mutants have a partial replication checkpoint defect at high temperatures. Given that overexpression of rad24⁺, rad25⁺, or chk1⁺ seems to enhance the checkpoint signal in HU, we speculated that rad24 might show a similar replication checkpoint defect at high temperatures. We did not examine rad25 because it has a much weaker defect in the DNA damage checkpoint (FORD et al. 1994 ). The rad24 rad25 double mutant is inviable (FORD et al. 1994 ), so its checkpoint responses cannot be evaluated.

The response of rad24 and chk1 cells to HU at high temperature was evaluated. Wild-type, rad3, rad24, and chk1 cells were shifted to 37° and HU was added. Cell phenotypes were examined at 2-hr intervals. As reported by FRANCESCONI et al. 1997 , under these conditions chk1 cells show a partial replication checkpoint defect (Figure 8A). rad24 cells are as defective as chk1 cells, with 27% cuts after 8 hr in the presence of HU at 37° (Figure 8A), indicating that under these conditions, rad24⁺ is also required for the checkpoint response to unreplicated DNA. However, neither strain is fully defective in comparison to rad3, in which cuts accumulate sooner and reach higher levels (Figure 8A).

View larger version (23K): In this window In a new window Download PPT slide   Figure 8. rad24+ and chk1+ are required for the checkpoint response to unreplicated DNA. (A) Wild-type (TE235), rad24 (TE465), chk1 (TE548), and rad3 (TE890) cells were grown to midlog phase. HU was added to a final concentration of 10 mM and cells were shifted to 37° at t = 0. Samples were collected at 2-hr intervals, fixed, and stained with DAPI, and cuts were counted. At least 100 cells were counted for each data point. (B) The strains listed in A along with cds1 (TE700), chk1 rad24 (TE 922), cds1 rad24 (TE919), and cds1 chk1 (TE856) were grown and HU was added as described in A, except that cells were at 29° throughout the experiment. The percentage of cuts at 6 hr is shown. (C) Cells from the same experiment as B were examined at 2-hr intervals. (D) cds1 rad24 (TE919) cells and (E) cds1 chk1 (TE856) cells after 6 hr in 10 mM HU at 29°, stained with DAPI. Arrows indicate cut cells.

cds1 rad24, like the cds1 chk1 double mutant, is completely checkpoint defective:
Our results suggest that Chk1p and 14-3-3 proteins could be directly involved in the checkpoint response to unreplicated DNA. It has also been proposed that Cds1p, a protein kinase with significant similarity to Rad53p, which is required for the checkpoint response to unreplicated DNA in S. cerevisiae (ALLEN et al. 1994 ; WEINERT et al. 1994 ), is involved in the checkpoint response to unreplicated DNA. cds1⁺ cannot be the only gene necessary for checkpoint arrest, because cds1^- mutants arrest normally in response to HU, though they quickly lose viability (MURAKAMI and OKAYAMA 1995 ). To explain why single mutants in rad24⁺, cds1⁺, and chk1⁺ are only modestly checkpoint defective, we considered the possibility that redundant pathways are involved in the checkpoint response to unreplicated DNA. To investigate further the roles of rad24⁺, cds1⁺, and chk1⁺ in the checkpoint response to unreplicated DNA, we constructed double mutants among these genes and studied their checkpoint responses by examining their phenotypes in the presence of HU at 29°. Like rad24 single mutants (FORD et al. 1994 ), cds1 rad24 and chk1 rad24 were semi-wee and had a round morphology under normal conditions (data not shown).

In the presence of HU, chk1 rad24 cells were only slightly more checkpoint defective than the chk1 or rad24 single mutants, showing 17% cuts after 6 hr (Figure 8B and Figure C). As the double mutant is not more checkpoint defective than the single mutants, rad24⁺ and chk1⁺ may function together in the checkpoint response to unreplicated DNA. In contrast, cds1 rad24 double-mutant cells were much more checkpoint defective than rad24 or cds1 single mutants, showing 70% cuts after 6 hr in the presence of HU (Figure 8B and Figure D). As has been reported elsewhere (BODDY et al. 1998 ; LINDSAY et al. 1998 ; ZENG et al. 1998 ), the cds1 chk1 cells were also more severely checkpoint defective than chk1 or cds1 cells, with 62% cuts at 6 hr in the presence of HU (Figure 8B and Figure E). As shown in Figure 8C, the kinetics and extent of cut formation in cds1 rad24 and cds1 chk1 cells suggest that these mutants are completely defective in the checkpoint response to unreplicated DNA (compare to rad3). These results suggest that rad24⁺ and chk1⁺ function in parallel with cds1⁺ in the checkpoint response to unreplicated DNA, as the checkpoint response can be fully eliminated by combining mutations in either rad24⁺ or chk1⁺ with cds1⁺.

We also examined the DNA damage checkpoint response of the double mutants by evaluating their survival at increasing doses of UV irradiation. Again, the chk1 rad24 cells were somewhat checkpoint defective, showing UV-sensitivity comparable to chk1, while cds1 rad24 and cds1 chk1 cells were severely UV-sensitive like rad3 cells (data not shown). These results suggest that cds1⁺ may function in parallel with rad24⁺ and chk1⁺ in the checkpoint response to damaged DNA, as the checkpoint response can be fully eliminated by combining mutations in either rad24⁺ or chk1⁺ with cds1⁺.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We have conducted a screen for genes involved in the checkpoint response to unreplicated DNA, by searching for high-copy suppressors of the replication checkpoint defect of cells overexpressing cdc25⁺ (OPcdc25⁺). None of the suppressors that were identified affect the cell-cycle progression of OPcdc25⁺ under normal conditions, so they are not simply negative regulators of the G2/M transition. Rather, they suppress by somehow enhancing the ability of OPcdc25⁺ cells to respond to checkpoint signals. As summarized in Figure 9, the suppressors fall into two classes. One group of suppressors includes rad24⁺, rad25⁺, and chk1⁺, three genes known to be involved in the DNA damage checkpoint. Our analysis suggests that these genes negatively regulate Cdc25p, thus rescuing the checkpoint defect caused by Cdc25p overexpression. We have shown that the same interactions are important during the checkpoint response to unreplicated DNA in wild-type cells, as rad24^- and chk1^- mutants have replication checkpoint defects, particularly when combined with mutations in cds1⁺. Thus, we believe that the damage checkpoint proteins Rad24p, Rad25p, and Chk1p function in concert with the protein kinase Cds1p in the checkpoint response to unreplicated DNA. Another group of suppressors includes sum3⁺. Overexpression of sum3⁺ may block the stress response pathway that inhibits cell-cycle progression by a mechanism that is independent of Cdc2 tyrosine phosphorylation.

View larger version (19K): In this window In a new window Download PPT slide   Figure 9. Model of genetic interactions influencing checkpoint control. The checkpoint response to unreplicated DNA uses parallel pathways, one requiring Cds1p and the other requiring Chk1p and 14-3-3 proteins. The targets of Cds1p are not known. The stress response counteracts the checkpoint response, and inhibition of this pathway indirectly enhances the checkpoint response. Sum3p inhibits the cell-cycle response to stress through unknown effectors. For further discussion, see the text.

sum3⁺ regulates the osmotic stress and checkpoint responses:
sum3⁺ encodes a member of the DEAD box family of ATP-dependent RNA helicases. DEAD box proteins have been found to have roles in many cellular processes including RNA splicing, RNA degradation, ribosome biogenesis, ribosome assembly, translation, and regulation of maternally expressed RNAs or developmentally regulated mRNAs (SCHMID and LINDER 1992 ; PY et al. 1996 ). Members of this family include mouse eIF-4A and PL10, numerous S. cerevisiae genes, and Drosophila WM6 (WARBRICK and GLOVER 1994 ). The latter was identified in a screen for Drosophila cDNAs able to suppress the mitotic catastrophe phenotypes of wee1-50 OPcdc25⁺ or wee1-50 mik1 at the restrictive temperature (SCHMID and LINDER 1992 ; WARBRICK and GLOVER 1994 ). Compared to many other DEAD helicases, the protein encoded by WM6 is relatively distantly related to Sum3p with 29% identity and 52% similarity, and it is not known whether it can rescue checkpoint mutants in the presence of HU. The closest known homologue of sum3⁺ is the S. cerevisiae gene, DED1 (Figure 2). Ded1p was previously believed to function in splicing, but recent work implicates it in the initiation step of translation (CHUANG et al. 1997 ; DE LA CRUZ et al. 1997 ). Both DED1 and sum3⁺ are essential for viability (STRUHL 1985 ; B. GRALLERT and K. LABIB, personal communication). Intriguingly, we have recently isolated another essential gene potentially implicated in translation initiation, sum1⁺, in a screen for suppressors of the checkpoint mutant cdc2-3w (HUMPHREY and ENOCH 1998 ). sum1⁺ is homologous to the S. cerevisiae WD repeat protein TIF34, which is associated with an essential multiprotein complex, eIF3, required for translational initiation (NARANDA et al. 1997 ). We believe that these suppressors are not working by stimulating translation of ribonucleotide reductase protein (the target of HU), because cells overexpressing the sum genes in the presence of HU are elongated, while cells overexpressing ribonucleotide reductase are not.

Given that the cell-cycle effectors of the Sty1p/Spc1p stress-activated MAPK pathway are not known, at least two models can be drawn to account for Sum3p's negative regulation of the cell-cycle response to osmotic