Protein Phosphatase 2C Acts Independently of Stress-activated Kinase Cascade to Regulate the Stress Response in Fission Yeast*

Stress-activated signal transduction pathways, which are largely conserved among a broad spectrum of eukaryotic species, have a crucial role in the survival of many forms of stress. It is therefore important to discover how these pathways are both positively and negatively regulated. Recent genetic studies have implicated protein phosphatase 2C (PP2C) as a novel negative regulator of stress response pathways in both budding and fission yeasts. Moreover, it was hypothesized that PP2C dephosphorylates one or more components of protein kinase cascades that are at the core of stress-activated signal transduction pathways. Herein we present genetic and biochemical studies of the fission yeastSchizosaccharomyces pombe that disprove this hypothesis and indicate that PP2C instead negatively regulates a downstream element of the pathway. First, high expression of PP2C produces phenotypes that are inconsistent with negative regulation of the Wik1-Wis1-Spc1 stress-activated kinase cascade. Second, high expression of PP2C leads to sustained activating tyrosine phosphorylation of Spc1. Third, Spc1-dependent phosphorylation of Atf1, a transcription factor substrate of Spc1, is unaffected by high expression of PP2C. Fourth, high expression of PP2C suppresses Atf1-dependent transcription of a stress-response gene. These studies strongly suggest that PP2C acts downstream of Spc1 kinase in the stress-activated signal transduction pathway.

Eukaryotic organisms frequently encounter environmental conditions that cause cytotoxic damage; hence, they have developed sophisticated systems of sensing and responding to physiological stress. Protein kinase cascades are at the core of these stress sensor pathways (1). These cascades follow the paradigm established for mitogen-activated protein kinase (MAPK) 1 cascades: a MAPK kinase kinase (MAPKKK) phos-phorylates a MAPK kinase (MAPKK), which in turn phosphorylates a MAPK. The MAPKKK and MAPK components are serine-threonine kinases, whereas MAPKKs are dual specificity enzymes, activating MAPK substrates by phosphorylating threonine and tyrosine residues in a conserved motif (2). It is not understood why protein kinase cascades are used to transmit stress signals, although it is likely that the spatial distribution of the cascade elements facilitates rapid signaling from the cell surface to the nucleus where MAPK homologs phosphorylate transcription factor substrates.
Recent studies have revealed impressive functional and structural conservation of stress response pathways in yeast, plants, and various metazoan species, including humans (3). The fission yeast Schizosaccharomyces pombe has been focus of some of the most interesting investigations, in part because studies of fission yeast have uncovered a link between stress response pathways and cell cycle control (4,5). The S. pombe stress-activated kinase cascade consists of Wik1-Wis1-Spc1 kinases (6). The Spc1 MAPK homolog, which is also known as Sty1 and Phh1 (4,7), is highly similar to mammalian p38 kinases (8) and Hog1p kinase of the budding yeast Saccharomyces cerevisiae (9). Like p38 kinase, Spc1 is broadly responsive to many forms of stress, including high osmolarity, oxidative stress, UV irradiation, and heat stress, as well as carbon and nitrogen starvation (4,5,10,11). Recent investigations have further revealed that mammalian p38 and fission yeast Spc1 share similar substrates. In the case of Spc1, the substrate is Atf1, a transcription factor containing a bZIP (basic leucine zipper) domain. Remarkably, Atf1 is highly similar to ATF-2, a mammalian transcription factor that is widely believed to be an important substrate of p38 kinase in vivo (11,12).
It is evident that fission and budding yeasts can serve as useful model systems for uncovering novel modes of regulation of stress-activated protein kinase cascades. This accounts for the interest in recent studies suggesting that stress-activated kinase cascades in yeast are negatively regulated by two types of protein phosphatases: tyrosine-specific enzymes (4,5,10) and serine-threonine phosphatases of the type 2C class (PP2C) (5,13). Evidence for a role of tyrosine-specific phosphatases in the negative regulation of stress-activated kinase cascades first arose from genetic studies of S. cerevisiae. The gene PTP2 was cloned as a high copy suppressor of mutations that cause lethality due to hyperactivation of the Hog1p kinase cascade. High expression of PTP2 results in decreased tyrosine phosphorylation of Hog1p, a finding consistent with the conclusion that Ptp2p directly dephosphorylates Hog1p in vivo (13). Ptp2p is assumed to be a tyrosine-specific enzyme because it is more closely related to tyrosine-specific phosphatases than to dual specificity tyrosine/threonine phosphatases that have been identified as MAPK phosphatases in mammalian cells, although the enzymatic specificity of Ptp2p has not been examined. In the case of S. pombe, a critical advance was made with the discovery that the synthetic lethal phenotype observed in a strain lacking two tyrosine phosphatase genes, pyp1 and pyp2, was effectively suppressed by null mutations in genes encoding elements of the Wis1-Spc1 kinase cascade. Moreover, the stress-sensitive and G 2 cell cycle delay phenotypes exhibited by spc1 and wis1 mutants were replicated by high expression of Pyp1 and Pyp2 enzymes (4,5). These genetic studies were followed by definitive biochemical findings showing that Pyp1 and Pyp2 directly dephosphorylated tyrosine 173 of Spc1 both in vivo and in vitro (5,10). Interestingly, pyp2 ϩ expression is induced in response to stress by a process that requires the Wis1-Spc1 kinase cascade and Atf1 transcription factor, indicating a mechanism of negative feedback regulation of the Wik1-Wis1-Spc1 signal transduction pathway (11,12).
In contrast to the situation with tyrosine-specific phosphatases, the role of PP2C in the negative regulation of stressactivated kinase cascades is poorly understood. The initial insights came from studies of S. cerevisiae that revealed that two PP2C genes, PTC1 and PTC3, were multicopy suppressors of mutations that caused hyperactivation of the Hog1p kinase cascade (13). The suggestion that PP2C negatively regulates the Hog1p stress-activated kinase cascade was further supported by the discovery of a synthetic lethal interaction involving ptp2 and ptc1 mutations, which was suppressed by hog1 mutations (14). These findings led to the hypothesis that PP2C negatively regulates the stress signaling pathway by dephosphorylating Hog1p or Pbs2p, the MAPKK homolog that activates Hog1p. Independent evidence for a role of PP2C in the negative regulation of stress-activated kinase cascades came from studies of S. pombe. In S. pombe the genes ptc1 ϩ , ptc2 ϩ , and ptc3 ϩ account for ϳ90% of the total PP2C activity (15,16). Mutants carrying deletions of two or three of these genes are hypersensitive to increased levels of calcium. These defects are rescued by wis1 and spc1 mutations (16). These findings are consistent with the notion that PP2C negatively regulates the Wis1-Spc1 kinase cascade in S. pombe, possibly via direct dephosphorylation of Spc1 and/or Wis1.
Studies of PP2C in yeast are consistent with an increasing body of evidence suggesting a role for PP2C in the regulation of signal transduction systems. In mammals, PP2C is believed to play a role in Ca 2ϩ -dependent signal transduction in the brain (17). Findings implicating PP2C in Ca 2ϩ -related signal transduction have also arisen from studies of Arabidopsis thaliana, whose abi1 gene, which is essential for the response to the plant hormone abscisic acid, encodes a protein homologous to PP2C that also has a putative Ca 2ϩ binding site (18,19). PP2C also appears to be important for cell maturation and development, since its expression and activity is reported to be upregulated during the monocytic differentiation evoked by vitamin D 3 in the human leukemic HL-60 cells (20). Moreover, a recent study demonstrated that the FEM-2 gene of Caenorhabditis elegans encodes a PP2C enzyme that is required to promote male development (21).
Although genetic studies of fission and budding yeasts point toward a role for PP2C in regulating stress-activated protein kinase cascades, as of yet there have been no studies that directly test this hypothesis. Herein we describe such studies carried out with S. pombe. We report that expression of one of the PP2C genes, ptc1 ϩ , is induced after osmotic stress by a mechanism that requires Spc1 and Atf1. Contrary to the hypothesis, we have found that high expression of ptc1 ϩ does not reduce activating tyrosine phosphorylation of Spc1 or the phosphorylation of Atf1 that is catalyzed by Spc1. In fact, high expression of ptc1 ϩ causes stress-induced tyrosine phosphorylation of Spc1 to be sustained, suggesting that Ptc1 acts downstream of Spc1 to negatively regulate the stress response. This finding is consistent with the observation that the level of stress-induced transcription of several genes is reduced in cells that overproduce Ptc1.
RNA Isolation and Hybridization-RNA isolation and detection of leu1, pyp2, and gpd1 mRNA have been described (10). NdeI-NotI fragments from the pREP1-ptc1HA6H, pREP1-ptc2HA6H, or pREP1-ptc3HA6H plasmids were used to probe ptc1, ptc2, or ptc3 mRNA. Cells from various strains were grown to midlog phase at 30°C, stressed by adjusting the medium to 0.6 M KCl, and harvested by filtration. Quantification of signals was performed using a Molecular Dynamics PhosphorImager.
Purification and Detection of Spc1-HA6H and Atf1-HA6H Proteins-KS1376 and FG1761 strains carry a genomic copy of spc1 ϩ tagged with a sequence encoding two copies of the influenza virus hemagglutinin (HA) epitope followed by six consecutive histidines residues. For ptc1 ϩ overexpression experiments, KS1376 and KS1479 cells were transformed with the pART1 control vector and the pART1-ptc1 ϩ vector in which ptc1 ϩ expression is driven by the strong constitutive adh1 promoter. Cells were grown to midlog phase at 30°C and stressed by the addition of 0.6 M KCl. Samples were harvested by filtration. Spc1 or Atf1 proteins were purified by Ni 2ϩ -nitrilotriacetic acid (NTA)-agarose chromatography, resolved by SDS-polyacrylamide gel electrophoresis, electroblotted to a nitrocellulose membrane, and detected by immunoblotting with either the anti-HA (12CA5) or antiphosphotyrosine (4G10 antibody (Upstate Biotechnology, Inc.) (5,11). Immunoreactive bands were revealed with horseradish peroxidase-conjugated secondary antibodies and the ECL Western blotting detection system (Amersham Corp.).
Analysis of PP2C Activity-Fractionation of S. pombe crude lysates by Mono Q chromatography was performed as described (15,16). Preparation of 32 P-labeled casein and the procedures of the PP2C phosphatase assay were as described by McGowan and Cohen (24). Okadaic acid (100 nM; Calbiochem) was used to inhibit PP2A activity in the crude extracts.

Synthetic Lethal Interactions Involving Mutations of Tyrosine Phosphatase and PP2C
Genes-The first aim of our work was to expand the genetic studies implicating PP2C in the negative regulation of the Spc1 stress-activated signal transduction pathway. Pyp1 tyrosine phosphatase has a major role in the negative regulation of Spc1 kinase (5); therefore, we explored the genetic interactions involving mutations of pyp1 and the genes encoding PP2C. Strains harboring the ⌬pyp1 mutation in combination with either ⌬ptc1 or ⌬ptc3 mutations were found to grow as well as wild type, as judged by colony size on rich YES agar medium (Fig. 1). However, whereas a ⌬ptc1 ⌬ptc3 double mutant also grew as well as wild type, the ⌬pyp1 ⌬ptc1 ⌬ptc3 triple mutant was only weakly viable ( Fig. 1).
This result suggests that Pyp1 tyrosine phosphatase together with Ptc1 and Ptc3 PP2C enzymes contribute to the negative regulation of the stress-activated signal transduction pathway that includes the Wis1-Spc1 kinase cascade. To further explore this possibility, we asked whether the ⌬pyp1 ⌬ptc1 ⌬ptc3 triple mutant phenotype was suppressed by inactivation of Wis1. A genetic cross was performed to generate a ⌬pyp1 ⌬ptc1 ⌬ptc3 ⌬wis1 mutant. This quadruple mutant grew much better than the ⌬pyp1 ⌬ptc1 ⌬ptc3 triple mutant ( Fig. 1), indicating that inactivation of the Wis1-Spc1 kinase cascade suppresses the ⌬pyp1 ⌬ptc1 ⌬ptc3 triple mutant phenotype. These findings are most simply interpreted to suggest that the Ptc1 and Ptc3 enzymes contribute to the negative regulation of the stress-activated signal transduction pathway. ptc1 ϩ Expression Is Regulated by Spc1 and Atf1-Heat shock leads to phosphorylation and the activation of Spc1 and increased expression of ptc1 ϩ mRNA (10,15). These observations prompted us to explore whether ptc1 ϩ mRNA expression is responsive to other forms of stress that are known to activate Spc1 kinase, such as high osmolarity, and whether the activation of Spc1 and increased expression of ptc1 ϩ are causally related. Fig. 2A shows that ptc1 ϩ mRNA increases in response to exposure to high salt media (0.6 M KCl). Maximal ptc1 ϩ expression was reached within 30 -40 min of exposure to high osmolarity conditions. Induction of ptc1 ϩ expression was completely abolished in spc1-M13 cells ( Fig. 2A). Thus, ptc1 ϩ expression is elevated in response to osmotic stress by a Spc1-dependent mechanism.
Atf1 protein, a bZIP transcription factor that is a substrate of Spc1 kinase, is required for the transcription of several stress response genes such as pyp2 ϩ , which encodes a tyrosine phosphatase that inactivates Spc1, and the gpd1 ϩ gene encoding glycerol-3-phosphate dehydrogenase, which is believed to be crucial for the elevation of the internal glycerol concentration necessary to survive high osmolarity stress (11,12). Therefore, we analyzed the abundance of ptc1 ϩ mRNA in an atf1 Ϫ strain after stress. The data in Fig. 2A show that high osmolarity stress-induced ptc1 ϩ expression is dependent on Atf1.
The ptc1 ϩ gene is one of three PP2C genes to have been identified in fission yeast. We explored whether the other two genes, ptc2 ϩ and ptc3 ϩ , were also transcriptionally regulated in response to stress. The ptc2 ϩ mRNA was detected as two weakly expressed species whose level was unchanged in response to high osmolarity stress (Fig. 2B). The ptc3 ϩ mRNA was detected as an abundant species whose level was also unresponsive to stress (Fig. 2B). Therefore, of the three known PP2C genes in S. pombe, ptc1 ϩ is uniquely responsive to stress.
High Expression of Ptc1 and Pyp Phosphatases Cause Different Cell Morphology Phenotypes-The discovery that ptc1 ϩ undergoes stress-induced transcription by a mechanism dependent on Spc1 and Atf1 was highly reminiscent of the regulation of pyp2 ϩ expression (11,12). As mentioned above, pyp2 ϩ encodes a tyrosine phosphatase that dephosphorylates and thereby inactivates Spc1 kinase. Spc1-dependent induction of pyp2 ϩ mRNA expression constitutes a negative feedback control mechanism. Genetic studies in fission yeast suggest that the Wis1-Spc1 kinase cascade and Ptc1 phosphatase have counteractive roles in stress response (16). Moreover, similar findings involving the Hog1p stress-activated kinase and Ptc1p and Ptc3p PP2C enzymes in S. cerevisiae have led others to suggest that PP2C might directly dephosphorylate stress-activated kinases (13). Thus, one hypothesis that explains our genetic findings is that Ptc1 negatively regulates Spc1 by dephosphorylating Spc1, Wis1, or another upstream element of the signal transduction cascade that is activated by serine or threonine phosphorylation, as observed for the known Spc1 negative regulators, Pyp1 and Pyp2 tyrosine phosphatases (5,11,12).
Mutants defective for Spc1 or Wis1 activity display characteristic phenotypes, including a delay of the onset of mitosis. Thus, spc1 Ϫ and wis1 Ϫ cells divide at a cell length of approximately 20 m in EMM2 medium, whereas wild type cells divide at approximately 14 m (5). High expression of pyp1 ϩ or pyp2 ϩ genes produces an identical cell elongation phenotype, a finding consistent with the demonstration that Pyp1 and Pyp2 inactivate Spc1 kinase via direct dephosphorylation of tyrosine 173 (4, 5, 10). If PP2C enzymes dephosphorylate and thereby inactivate stress-activated protein kinases, then high expression of PP2C should induce phenotypes that are very similar to those observed in spc1 Ϫ and wis1 Ϫ mutants as well as Pyp1 and Pyp2 overproducer strains. We therefore examined the cell size phenotype of strains that express ptc1 ϩ at a high level. We employed a strain transformed with a plasmid (pART1-ptc1 ϩ plasmid) expressing ptc1 ϩ under the control of the strong adh1 promoter (25). Assays of Mono Q chromatographic fractionation samples from extracts of cells transformed with pART1-ptc1 ϩ confirmed that there was a large increase in PP2C activity (Fig. 3A). The increased activity was predominantly detected in fractions 6 -8, which corresponds to the chromato- Cells were collected at 10-min intervals, and total RNA was extracted and submitted to Northern hybridization analysis with the ptc1 ϩ probe. The leu1 ϩ probe was used as an RNAloading control. B, wild type (PR109) cells were stressed by growth in medium containing 0.6 M KCl, total RNA was extracted every 20 min and submitted to Northern blot hybridization using the ptc2 ϩ or ptc3 ϩ probes.
graphic properties of Ptc1 established in previous studies (15,16). Interestingly, overexpression of ptc1 ϩ leads to a decrease in cell length, 15.1 Ϯ 0.9 m for control cells and 12.8 Ϯ 1 m for pART1-ptc1 ϩ -transformed cells, respectively (Fig. 3B). This phenotype is the opposite of the cell elongation phenotype that is caused by overexpression of Pyp1 and Pyp2 or by mutational inactivation of Spc1 (5,26,27). This phenotype is inconsistent with the suggestion that PP2C inactivates Wis1, Spc1, or upstream components of the signal transduction pathway.
Stress-induced Tyrosine Phosphorylation of Spc1 Is Sustained in Ptc1 Overproducer Cells-To further explore the possibility that Ptc1 negatively regulates Wis1 or another upstream element, we examined Spc1 tyrosine phosphorylation in a Ptc1 overproducer strain. Plasmid pART1-ptc1 ϩ was transformed into a strain in which the chromosomal copy of spc1 ϩ was tagged with a sequence encoding the hemagglutinin antigen epitope and six histidines (Ha6H), allowing easy purification of Spc1 with Ni 2ϩ -NTA-agarose beads and detection with anti-HA antibodies. Exposure of cells transformed with a control plasmid (pART1) to osmotic stress, performed in this experiment by suspension of cells in growth medium containing 0.6 M KCl, led to a rapid and transient increase in Spc1 tyrosine phosphorylation (Fig. 4). Spc1 tyrosine phosphorylation reached a maximum level approximately 10 min after the exposure to osmotic stress, a finding that is consistent with previous studies. As shown in Fig. 4, exposure of pART1-ptc1 ϩ transformant cells to osmotic stress also led to a rapid increase in Spc1 tyrosine phosphorylation. Thus, high expression of ptc1 ϩ had no effect on the initial large increase in Spc1 tyrosine phosphorylation. However, whereas the increase in Spc1 tyrosine phosphorylation was quite transient in the control culture, decreasing to the basal amount within 50 min following exposure to osmotic stress, the level of Spc1 tyrosine phosphorylation remained high in the pART1-ptc1 ϩ transformant cells for the duration of the experiment (Fig. 4). These findings directly contradict the prediction of the model, which postulates that PP2C negatively regulates Wis1 or other upstream elements of the Spc1 stress response pathway.
PP2C Does Not Regulate Spc1-dependent Phosphorylation of Atf1-Having found that Ptc1 overproduction does not inhibit Spc1 tyrosine phosphorylation, we next turned our attention to the possibility that Spc1 or Atf1 is a substrate of Ptc1 phosphatase. In theory, Ptc1 could inhibit Spc1 by dephosphorylating threonine 171, which together with tyrosine 173 is a site of activating phosphorylation carried out by Wis1 kinase (5). Another possibility was that Ptc1 counteracted Spc1 by dephosphorylating the substrates of Spc1. Atf1 transcription factor was recently identified as a key substrate of Spc1 (11,12). Spc1-mediated phosphorylation of Atf1 causes Atf1 to migrate with reduced electrophoretic mobility; thus, the phosphorylation state of an Spc1 substrate can be quite easily monitored. This analysis was performed with a strain in which the genomic copy of atf1 ϩ encoded the Atf1 protein having the Ha6H tag that was described above (11). Plasmid pART1-ptc1 ϩ and a control plasmid were transformed into this strain. In cells transformed with the control plasmid pART1, exposure to medium containing 0.6 M KCl caused Atf1 to migrate with reduced mobility in SDS-polyacrylamide gel electrophoresis (Fig. 5). This mobility shift is completely dependent on Spc1 kinase (11). Fig. 5 shows that high expression of ptc1 ϩ had no effect on the Atf1 mobility shift. These findings strongly suggest that Ptc1 does not inactive Spc1; nor does Ptc1 appear to FIG. 3. Effects of the overexpression of Ptc1 and Pyp phosphatases. A, high expression of the ptc1 ϩ gene in pART1-ptc1 ϩ transformants results in a large increase in PP2C activity. PP2C assays of Mono Q chromatographic fractionation samples from extracts of cells transformed with pART1-ptc1 ϩ or a control plasmid (pART1) were performed. The assays were performed by measuring the release of radioactive phosphate from 32 P-labeled casein (see "Experimental Procedures"). B, overexpression of Ptc1 or Pyp phosphatases cause different cell phenotypes. Wild type (PR109), ⌬wis1 (JM544), spc1-M13 (KS1136), and cells overexpressing ptc1 ϩ were plated on EMM2 medium and grown 48 h at 30°C. For pyp1 ϩ and pyp2 ϩ overexpression, PR109 cells were transformed with plasmids expressing those genes under the control of the S. pombe nmt promotor (plasmids pREP1-pyp1 ϩ and pREP1-pyp2 ϩ ). Transformed cells were grown on EMM2 agar medium containing thiamine, which repressed the nmt promotor. Expression was induced by plating cells on EMM2 agar lacking thiamine for 48 h at 30°C. Bar, 20 m.

FIG. 4.
Overproduction of ptc1 ؉ causes tyrosine phosphorylation of Spc1 to be sustained at a high level following osmotic stress. Cells containing a chromosomal copy of Ha6H epitope-tagged spc1 ϩ and transformed with the control plasmid pART1 (control) or pART1-ptc1 ϩ (ov. ptc1 ϩ ) were incubated for the indicated times in the presence of 0.6 M KCl. Spc1 protein was precipitated with Ni 2ϩ -NTA beads and probed by Western blotting for the presence of phosphotyrosine (␣-pTyr). Immunoblot analysis with an anti-HA antibody confirmed that the Spc1 levels were approximately constant throughout the experiment. dephosphorylate the sites on Atf1 that are phosphorylated by Spc1 kinase.
Transcriptional Induction of Stress Response Genes Is Reduced in Cells That Overexpress ptc1 ϩ -Although our findings strongly suggested that Ptc1 does not directly regulate Spc1 or dephosphorylate residues of Atf1 that are phosphorylated by Spc1 kinase, these studies did not directly address the possibility that Ptc1 negatively regulates the Atf1-dependent transcriptional induction of stress response genes. Therefore, studies were carried out to measure the transcriptional induction of pyp2 ϩ and gpd1 ϩ in pART1-ptc1 ϩ transformant cells exposed to osmotic stress. As shown in Fig. 6, incubation of cells harboring the control plasmid pART1 in medium containing 0.6 M KCl led to the rapid and transient increase in the level of both pyp2 ϩ and gpd1 ϩ mRNA. These findings are consistent with previous studies (4,10). In cells transformed with pART1-ptc1 ϩ , the pattern of pyp2 ϩ and gpd1 ϩ mRNA transcriptional induction was broadly similar, although the maximum level of induction was reduced by approximately 50%. These findings support the notion that Ptc1 negatively regulates the Spc1-dependent transcriptional induction of stress response genes.

DISCUSSION
Genetic Studies Indicate That PP2C Is Involved in the Negative Regulation of Stress Sensor Pathways-The major aim of the studies described herein was to test the hypothesis that type 2C protein phosphatases have as one of their major in vivo activities the negative regulation of stress-activated protein kinase cascades. Before discussing the molecular evidence presented in this paper, it may be helpful to review the genetic data that led to the formation of the model. The hypothesis arose from independent genetic studies carried out with two organisms: the budding yeast S. cerevisiae and the distantly related fission yeast S. pombe. The key findings in regard to S. cerevisiae are as follows: 1) overproduction of two PP2C genes, PTC1 and PTC3, rescues lethality induced by hyperactivation of Hog1p, a stress-activated protein kinase most similar to S. pombe Spc1 kinase; and 2) ⌬ptc1 mutations exhibit synthetic lethal interactions with mutations of PTP2, which encodes a tyrosine phosphatase (PTP) implicated in the negative regulation of Hog1p (14). These findings led Saito and colleagues (13) to propose that Ptc1p and Ptc3p negatively regulate Hog1p or Pbs2p, the latter protein being the kinase that phosphorylates and thereby activates Hog1p kinase.
The initial indication that PP2C enzymes might also be involved in the negative regulation of stress-activated kinases in fission yeast arose from the discovery via a mutant screen that the Ca 2ϩ -sensitive phenotype observed in a ⌬ptc1 ⌬ptc3 strain is suppressed by spc1 and wis1 loss-of-function mutations (16). One explanation for this observation was that loss of PP2C activity led to hyperactivation of Spc1 kinase and consequent lethality. In an attempt to extend the genetic data that relate to the model, we began our studies by investigating interactions involving PP2C and PTP genes. Genetic and biochemical studies have proven that Pyp1 and Pyp2 negatively regulate Spc1 via direct dephosphorylation of Tyr-173 (4, 5, 10). If PP2C enzymes also negatively regulate the Spc1 kinase cascade, then one might expect to observe genetic interactions involving mutations of the PTP and PP2C genes. Indeed, we found that whereas double mutant combinations of the ⌬pyp1, ⌬ptc1, and ⌬ptc3 mutations produced cells that grew very well on standard rich growth medium, the ⌬pyp1 ⌬ptc1 ⌬ptc3 triple mutant grew very poorly. This finding is consistent with the studies of S. cerevisiae showing synthetic lethal interactions with ptc1 and ptp2 mutations (14). This correlation strongly suggests that PP2C enzymes have similar roles in regulating the stress responses in the two evolutionarily divergent yeasts. Our studies went a step further by showing that the ⌬pyp1 ⌬ptc1 ⌬ptc3 phenotype was quite effectively suppressed by ⌬wis1 mutations. This finding is reminiscent of the suppression of ⌬pyp1 ⌬pyp2 synthetic lethality by spc1 and wis1 mutations (4, 5). These observations are most easily interpreted to FIG. 5. Phosphorylation-induced electrophoretic mobility shift of Atf1 is unaffected by overproduction of Ptc1. Cells carrying a chromosomal copy of Ha6H epitope-tagged atf1 ϩ (KS1479) and transformed with the control pART1 plasmid (lanes 1 and 2; WT) or with pART1-ptc1 ϩ plasmid (lanes 3 and 4; ov. ptc1) were grown in YES medium at 30°C, and aliquots were harvested before (lanes 1 and 3) or after (lanes 2 and 4) 15 min of osmotic stress achieved by adjusting the culture medium to a concentration of 0.6 M KCl. Atf1-tagged protein was then purified on Ni 2ϩ -NTA beads and analyzed by SDS-polyacrylamide gel electrophoresis followed by immunoblotting with anti-HA antibody. The retarded migration of Atf1 observed in the control sample (lane 2) was also seen in the sample made from cells overexpressing ptc1 ϩ (lane 4). suggest that Ptc1 and Ptc3 phosphatases are involved in the negative regulation of the Spc1 signal transduction system, although other more complex explanations cannot be excluded.
Expression of ptc1 ϩ Is Regulated by the Stress-regulated Signal Transduction System-With a background of genetic data suggesting that PP2C enzymes negatively regulate the Spc1 kinase cascade, we proceeded to explore whether any of the known PP2C genes in S. pombe were transcriptionally induced in response to stress by an Spc1-dependent mechanism. Expression of pyp2 ϩ is strongly elevated in cells exposed to stress, and this induction of pyp2 ϩ mRNA transcription is entirely dependent on Spc1 and Atf1. In analogous studies, we found that the level of ptc1 ϩ mRNA underwent a large increase in response to osmotic stress. Moreover, this response was abolished in spc1 and atf1 mutants. The similarity of findings involving the transcriptional regulation of pyp2 ϩ and ptc1 ϩ further supports the idea that Ptc1 may be involved in the negative regulation of the Spc1 kinase signal transduction system. As is the case for pyp2 ϩ , we propose that transcriptional induction of ptc1 ϩ is part of a negative feedback or attenuation mechanism that serves to fine tune the cellular response to stress.
Wis1 and Spc1 Are Not Regulated by Ptc1-Tyrosine phosphorylation of Spc1 is absolutely dependent on Wis1 in vivo (4,5). Wis1 is a MAPKK homolog, and therefore like all MAPKK homologs it is very likely that Wis1 is activated by serine and/or threonine phosphorylation carried out by a MAPKKK homolog. Indeed, recent studies have identified a MAPKKK homolog, Wik1, which activates Wik1 in response to osmotic stress (6). Wis1 can be considered to be a reasonable candidate target of Ptc1 phosphatase. If Wis1 is negatively regulated by Ptc1, then it is predicted that overproduction of Ptc1 should depress Spc1 tyrosine phosphorylation and thereby cause cells to divide at a larger size. However, we carried out such a study and found that Ptc1 overproduction did not decrease Spc1 tyrosine phosphorylation. In fact, the major effect of Ptc1 overproduction was to cause Spc1 tyrosine phosphorylation to be sustained at a high level in cells exposed to osmotic stress. Likewise, the cells that overproduced Ptc1 underwent division at a reduced cell size relative to wild type. These findings contradict a model in which Ptc1 negatively regulates Wis1. In addition, these findings are inconsistent with Spc1 inhibiting the activity of any protein that acts upstream of Wis1, such as Wik1, to promote the activation of Spc1.
Spc1 is activated by phosphorylation of two residues: Thr-171 and Tyr-173 (5). PP2C enzymes could potentially dephosphorylate Thr-171; therefore, it was reasonable to propose that Spc1 might be a direct substrate of Ptc1. Atf1 undergoes an electrophoretic mobility shift when phosphorylated by Spc1, thereby providing a convenient assay of the activation state of Spc1 in vivo. If Ptc1 dephosphorylates threonine 171 of Spc1, then Ptc1 overproduction should reduce or eliminate the Atf1 electrophoretic mobility shift that occurs in response to osmotic stress. However, we found that Ptc1 overproduction had no effect on the Atf1 electrophoretic mobility shift, suggesting that Ptc1 does not negatively regulate Spc1. Furthermore, these observations strongly suggest that Ptc1 does not dephosphorylate the residues on Atf1 that are phosphorylated by Spc1 and that cause the electrophoretic mobility shift. However, the fact that Ptc1 overproduction does not inhibit the Atf1 electrophoretic mobility shift does not eliminate the possibility that Atf1 is a substrate of Ptc1 phosphatase. It is possible that Atf1 undergoes activating phosphorylation by another protein kinase. In an attempt to address this possibility, we have asked whether Atf1 and Ptc1 co-precipitate from fission yeast cell lysates. Preliminary studies indicate no significant amount of Atf1 and Ptc1 co-precipitation, but these findings do not exclude a more transient or unstable interaction involving Atf1 and Ptc1.
Regulation of Stress-induced Gene Expression by Ptc1-Although our findings argue against a model in which Ptc1 dephosphorylates Spc1 or Wis1 or counteracts the phosphorylation of Atf1 that is catalyzed by Spc1, some of our molecular studies provide support for the possibility that Ptc1 regulates some aspect of the Spc1 signal transduction pathway. One important finding is that high expression of Ptc1 causes a decrease in the transcriptional induction of the pyp2 ϩ and gpd1 ϩ genes. This observation suggests that Ptc1 may indirectly regulate Atf1, perhaps by dephosphorylating an Atf1 co-factor involved in transcriptional induction. It was recently shown that Atf1 forms heterodimers with the protein encoded by the pcr1 ϩ gene (28). ⌬pcr1 cells display a phenotype similar to ⌬atf1 mutant cells, suggesting that Atf1-Pcr1 heterodimers may have an important role in the stress-induced gene transcription. However, current evidence suggests that Pcr1 is not phosphorylated in vivo, suggesting that it cannot be a substrate of Ptc1 phosphatase. Nevertheless, it is likely that Atf1-Pcr1 heterodimers interact with other proteins to regulate gene expression; thus, it is plausible that these unknown factors may be substrates of PP2C enzymes as indicated in the model (Fig. 7).
The action of PP2C on this unknown cofactor could be indirect and involve another stress-activated kinase cascade. In S. cerevisiae, PTC1 was identified as a suppressor of mutation in the PKC1 gene, which encodes a protein kinase C that regulates the MAP kinase Mpk1p implicated in cell wall metabolism (29). The temperature sensitivity of these mutants can be FIG. 7. Model of the Spc1 cascade regulation. Upon various forms of stress, the Wik1-Wis1-Spc1 kinase cascade is stimulated, leading to the activation of the transcription factor Atf1. Active Atf1 promotes the transcription of several important genes involved in stress response, such as gpd1 ϩ . The inactivation of Spc1 is carried out by two tyrosine phosphatases: Pyp1, which is constitutively expressed; and Pyp2, whose stress-induced expression creates a negative feedback loop. The type 2C serine/threonine phosphatase encoded by ptc1 ϩ is also transcribed in an Spc1-Atf1-dependent manner, and Ptc1 negatively regulates the pathway. Ptc1 acts downstream of Spc1, perhaps by dephosphorylating a cofactor of the Atf1-Pcr1 transcription complex (X). rescued by osmostabilizers. Since Ptc1p was first identified as a component of the Hog1p/Pbs2p kinase cascade, it appears that Ptc1p is involved in the regulation of more than one kinase cascade. These observations suggest that some cross-talk may exist between stress-activated kinase cascades. Recently, the cloning of a novel protein kinase gene, SOK1, implicated in stress response has been reported in S. cerevisiae (30). Although Sok1p is only activated by oxidative stress, it appears that several kinase cascades can contribute to the stress response. The situation is significantly different in fission yeast in which the Spc1 cascade seems to be required for the cellular response to a variety of forms of stress, including high osmolarity stress, heat stress, oxidative stress, and nutrient limitation (4, 5, 10). However, it is possible that another stressactivated signal transduction pathway is activated in parallel with the Spc1 system and that PP2C could be important to coordinate the cellular adaptation to stress.
It is striking that high expression of Ptc1 causes the stressinduced increase of Spc1 tyrosine phosphorylation to be sustained for an extended period of time. It is evident that this finding is inconsistent with any model that proposes that Ptc1 negatively regulates Wis1 or other upstream elements of the signal transduction pathway that have a positive role in activating Spc1. But why should Ptc1 overproduction cause Spc1 tyrosine phosphorylation to be sustained following exposure to osmotic stress? We propose that Ptc1 overproduction causes osmostress-induced Spc1 tyrosine phosphorylation to be sustained because cells are defective in adapting to the osmotic stress. In wild type cells the reduction of Spc1 tyrosine phosphorylation that occurs after the initial large increase in Spc1 tyrosine phosphorylation may come about largely because cells have adapted to the osmotic stress itself. Future experiments will be aimed at addressing this possibility.