Evidence for Antagonistic Regulation of Cell Growth by the Calcineurin and High Osmolarity Glycerol Pathways in Saccharomyces cerevisiae

Because Ca 2 (cid:1) signaling of budding yeast, through the activation of calcineurin and the Mpk1/Slt2 mitogen-activated protein kinase cascade, performs redundant function(s) in the events essential for growth, the simultaneous deletion of both these pathways ( (cid:2) cnb1 (cid:2) mpk1 ) leads to lethality. A PTC4 cDNA that encodes a protein phosphatase belonging to the PP2C family was obtained as a high dosage suppressor of the lethality of (cid:2) cnb1 (cid:2) mpk1 strain. Overexpression of PTC4 led to a decrease in the high osmolarity-induced Hog1 phosphorylation, and HOG1 deletion remarkably suppressed the synthetic lethality, indicating an antagonistic role of the high osmolarity glycerol (HOG) pathway and the Ca 2 (cid:1) signaling pathway in growth regulation. The cal-cineurin-Crz1 pathway was required for the down-reg-ulation of the HOG pathway. Analysis of the time course of actin polarization, bud formation, and the onset of mitosis in synchronous cell cultures demonstrated that calcineurin negatively regulates actin polarization at the bud site, whereas the HOG pathway positively regulates bud formation at a later step after actin has polarized. Cells of the yeast Saccharomyces cerevisiae evaluate and cope with their external environment by using rapidly respond-ing, highly calcineurin negatively regulates actin polarization at the bud site, whereas the HOG pathway promotes bud formation at a step after We also show that the calcineurin-Crz1 pathway is involved in the of the activated by Ca or

Because Ca 2؉ signaling of budding yeast, through the activation of calcineurin and the Mpk1/Slt2 mitogenactivated protein kinase cascade, performs redundant function(s) in the events essential for growth, the simultaneous deletion of both these pathways (⌬cnb1 ⌬mpk1) leads to lethality. A PTC4 cDNA that encodes a protein phosphatase belonging to the PP2C family was obtained as a high dosage suppressor of the lethality of ⌬cnb1 ⌬mpk1 strain. Overexpression of PTC4 led to a decrease in the high osmolarity-induced Hog1 phosphorylation, and HOG1 deletion remarkably suppressed the synthetic lethality, indicating an antagonistic role of the high osmolarity glycerol (HOG) pathway and the Ca 2؉ signaling pathway in growth regulation. The calcineurin-Crz1 pathway was required for the down-regulation of the HOG pathway. Analysis of the time course of actin polarization, bud formation, and the onset of mitosis in synchronous cell cultures demonstrated that calcineurin negatively regulates actin polarization at the bud site, whereas the HOG pathway positively regulates bud formation at a later step after actin has polarized.
Cells of the yeast Saccharomyces cerevisiae evaluate and cope with their external environment by using rapidly responding, highly complex signaling pathways. Of particular importance are the signaling pathways known collectively as the mitogen-activated protein kinase (MAPK) 1 cascade, which couples an extracellular input signal to a variety of outputs. The budding yeast possesses five MAPKs acting in five functionally distinct cascades, each mediating physiologically distinct responses (1). Adaptation to high osmolarity is mediated by the HOG (high osmolarity glycerol) pathway, in which the Hog1 MAPK cascade couples the high osmolarity signal to glycerol accumulation. The Hog1 MAPK of this pathway is related to the members of the family of human stress-activated p38 MAPKs. The activation of the HOG pathway is regulated by bifurcated upstream branches, one mediated by the Sho1 sensor and the other by the two-component system consisting of Sln1, Ypd1, and Ssk1 (2)(3)(4)(5)(6). The HOG pathway is negatively regulated by Ptp2 and -3, which are protein-tyrosine phosphatases, and by Ptc1, -2, and -3 serine/threonine-protein phosphatases belonging to the PP2C family (3,(7)(8)(9). The HOG pathway regulates various stress genes for osmoprotection including GPD1, which encodes glycerol-3-phosphate dehydrogenase, an enzyme involved in glycerol synthesis (10). This pathway is also suggested to be involved in the regulation of the cell cycle, budding pattern, and cell division after osmotic stress. However, the details of the regulatory mechanisms of these events are unknown (11,12).
In contrast to the HOG pathway, the Pkc1-Mpk1/Slt2 MAPK cascade responds to low osmolarity conditions. An important function of the Pkc1 pathway is to maintain cell wall integrity in response to low osmolarity and high temperature conditions by controlling the assembly of the cell wall (1).
Another major mechanism for adaptation to environmental changes in yeast is Ca 2ϩ signaling. Calcineurin, essential only under specific environmental conditions, is an important mediator of the Ca 2ϩ signal. Calcineurin is required for the recovery from the G 1 arrest induced by mating pheromones and for homeostasis of ions, including Na ϩ /Li ϩ , Mn 2ϩ , and OH Ϫ (13)(14)(15)(16). Calcineurin regulates the expression of ion transporter genes by a mechanism mediated by Crz1, a calcineurin-dependent transcription factor (17,18).
Although the defect of either calcineurin or the Mpk1 MAPK cascade alone does not lead to lethality under normal conditions, simultaneous defects of these pathways are synthetically lethal, indicating that the two pathways perform redundant function(s) in same essential event(s) for cell growth (19,20). Upon activation by Ca 2ϩ or various stimuli, the Mpk1 MAPK cascade maintains cell wall integrity by controlling the cell wall (1). In addition to these mechanisms, the two signaling pathways coordinately regulate the G 2 -M transition through the activation of Swe1, a negative regulator of Cdc28/Clb in the G 2 phase (21). In this process, calcineurin up-regulates Swe1 function in two ways, one by the activation of the transcription of SWE1 gene and the other by the destabilization of Hsl1 kinase, a negative regulator of Swe1 kinase. A high concentration of CaCl 2 in the medium causes the cells to activate these pathways and leads the cells to a G 2 cell-cycle delay. It was proposed that the activation of these pathways occurs in response to a membrane stretch stress (21).
In this present report, we show that the HOG pathway and the Ca 2ϩ signaling pathways antagonize one another in the regulation of cell growth. We further demonstrate that these pathways are antagonistic in the regulation of bud emergence. In the presence of CaCl 2 in the medium, the activation of calcineurin negatively regulates actin polarization at the bud site, whereas the HOG pathway promotes bud formation at a step after actin polarization. We also show that the calcineurin-Crz1 pathway is involved in the down-regulation of the HOG pathway activated by Ca 2ϩ or hyperosmotic stress.
Isolation of Suppressor cDNA Clones-The ⌬cnb1 ⌬mpk1 mutant (TNP36) was transformed with cDNA libraries constructed in the yeast expression vector pNV7, in which yeast cDNA sequences are placed under the control of the inducible GAL1 promoter (gift from K. Matsumoto, Nagoya University, Nagoya, Japan) on selection plate (SD minus uracil) containing 1 M sorbitol. The transformants were screened for growth on YPG plate (without 1 M sorbitol). The 85 positive transformants were isolated from 35,000 transformants. Six cDNA, including MPK1 gene, were isolated by the screening.
Cell-cycle Synchronization-The procedure for cell-cycle synchronization was described previously (21). Cells were grown in YPD to a density of 2-3 ϫ 10 6 cells/ml at 28°C. Synthetic ␣-factor was added to a final concentration of 5 g/ml, and cells were incubated for 135 min. After washing the cells with ice-cold water to remove ␣-factor, the cells were suspended in fresh YPD or YPD containing CaCl 2 .
Immunoblotting-The cells were washed with TEG buffer (50 mM Tris-HCl (pH 7.5), 10 mM EDTA, 5 mM EGTA, 25 mM NaCl, 0.05% Tween 20, and 10% glycerol) containing 1 mM dithiothreitol, phosphatase inhibitors (10 mM NaF and 1 mM sodium orthovanadate), and protease inhibitor mixture (␣-complete, Roche Molecular Biochemicals) and harvested by centrifugation. Pellets were frozen in liquid N 2 and stored at Ϫ80°C until use. Frozen pellets were resuspended in 100 l of 2ϫ sample buffer and lysed by vigorous vortexing for 12 min with an equal volume of glass beads at 4°C. After removal of glass beads, lysates were boiled for 5 min. Proteins were resolved by SDS-PAGE on a 7.5 or 12% gel for 2 h at 30 mA and analyzed by immunoblotting using either anti-Myc antibody (9E10, BAbCO) or anti-PSTAIRE antibody (this recognizes Cdc28 and Pho85; Santa Cruz Biotechnology). Preparation of cell lysates for analysis of Hog1 phosphorylation was as described (3), and phosphorylated Hog1-3HA was detected by anti-phospho-p38 antibody (Cell Signaling Technology) and Hog1-3HA was detected by anti-HA antibody (HA.11, BAbCO) as a loading control. For all immunoblots, horseradish peroxidase-conjugated second antibodies were used, and blots were visualized by ECL detection systems (Amersham Biosciences).
Hog1 Kinase Assay-Hog kinase assay in PTC4 or PTP2 overexpressed cells was performed as follows. A Hog1-HA integrated strain (SHI119) was transformed with vector only (YEp51) or pGAL-PTC4 (YEp51-PTC4). The cells were grown to 4 ϫ 10 6 cells/ml in SR minus leucine medium at 28°C and shifted to SG minus leucine medium and induced PTC4 gene expression. After 1.5 h of induction, cells were exposed to 0.4 M KCl for various periods of time. Cells were harvested by centrifugation and lysed by vigorous vortexing for 12 min with an equal volume of glass beads at 4°C in buffer A (50 mM Tris-HCl (pH 7.5), 50 mM NaCl, 5 mM EGTA, 5 mM MgCl 2 , 0.1% Triton X-100, and 1 mM DTT, containing protein phosphatase inhibitors (50 mM ␤-glycerophosphate, 10 mM sodium vanadate, and 10 mM NaF) and protease inhibitors (1 mM PMSF and inhibitor mixture; Roche). The lysates were This study clarified by centrifugation, 14,000 rpm for 20 min at 4°C, and Hog1-HA was immunoprecipitated by incubation with 2 mg of total protein, 5 l of anti-HA antibody (HA.11, 1 mg/ml), and 30 l of protein A beads (Sigma) for 90 min at 4°C. The immunoprecipitates were washed once with buffer A, twice with buffer B (50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 5 mM EGTA, 5 mM MgCl 2 , 1 mM DTT, 10 mM sodium vanadate, 10 mM NaF, and 1 mM PMSF), and once with kinase assay buffer (50 mM Tris-HCl (pH 7.5), 50 mM NaCl, 5 mM EGTA, 5 mM MgCl 2 , 1 mM DTT, 1 mM sodium vanadate). The 25-l supernatants of immunoprecipitates were used for immunoblotting as a loading control. To assay Hog1 kinase activity, Hog1-HA bound to protein A beads was incubated with [␥-32 P]ATP (5 Ci) and 45 g of myelin basic protein (MBP, Sigma) for 30 min at 30°C. Incorporation of 32 P into MBP was assessed by SDS-PAGE and quantified by BAS-1800 Bioimaging analyzer. In the experiments shown in Fig. 2, ⌬cnb1 ⌬mpk1 strain (TNP36) was transformed with vector only, pGAL-PTC4 (pNV7-PTC4), or pGAL-PTP2 (pSSP25, Ref. 3). The cells were grown to 6 -8 ϫ 10 6 cells/ml in SR minus uracil medium containing 1 M sorbitol at 28°C and shifted to SG minus uracil medium without sorbitol. Hog1 protein was immunoprecipitated by anti-Hog1 antibody (y-215, Santa Cruz Biotechnology.) with protein A beads and assayed for kinase activity as described above.
Fluorescence-activated Cell Sorting Analysis, Fluorescence Staining, and Microscopy-Approximately 1 ϫ 10 7 cells from yeast culture were harvested by centrifugation and resuspended in 300 l of 0.2 M Tris-HCl (pH 7.5). Cells were fixed by addition of 700 l of cold ethanol gradually with vortexing, followed by incubation for overnight at Ϫ20°C. The fixed cells were washed with the same buffer containing RNase A (1 mg/ml) for 3 h at 30°C. The cells were stained with 100 l of propidium iodide (50 g/ml) solution in 4 mM sodium citrate, 10 mM NaCl, and 0.1% Nonidet P-40 for 15 min on ice. The cells were analyzed by FACSCalibur (Becton Dickinson). F-actin was visualized by staining with rhodamine-conjugated phalloidin, as described (24). Quantification of cumulative percentage of bud formation and nuclear division during cell cycle progression in various strains that were synchronized with ␣-factor was determined by microscopic observation and propidium iodide staining. At least 300 cells were counted for each time point.

Identification of PTC4 cDNA as a High Dosage Suppressor of the Synthetic Lethality of the ⌬cnb1 ⌬mpk1
Double Deletion-To clarify the cellular event(s) that are regulated by the Ca 2ϩ signaling pathways (calcineurin and Mpk1 mitogen-activated protein kinase cascade), we screened for the complementary DNA (cDNA), overexpression of which could rescue the synthetic lethality of the ⌬cnb1 ⌬mpk1 strain. Because the double deletion mutant can grow in the presence of an osmotic stabilizer (e.g. 1 M sorbitol), a yeast cDNA library constructed in an expression vector (pNV7) driven by the galactose-inducible GAL1 promoter was introduced into the ⌬cnb1 ⌬mpk1 cells grown in the presence of sorbitol. The cells were spread on selection plates (SD minus uracil) containing sorbitol. From ϳ35,000 Ura ϩ transformants, 85 transformants that could grow on YPG plate, but not on YPD plate (both plates without sorbitol), were selected. Six plasmids that reproducibly conferred viability to the ⌬cnb1 ⌬mpk1 strain on the YPG plate were recovered. The cDNA contained in the plasmids was designated DSL (for dose suppressor of lethality as a result of the ⌬cnb1 ⌬mpk1 double deletion). Nucleotide sequences of the inserts were determined. One of the isolates (DSL1) contained a PTC4 cDNA, encoding a PP2C family protein phosphatase of unknown function (22).
The Lethality of the ⌬cnb1 ⌬mpk1 Strain Is Suppressed by the Additional Deletion of the HOG1 Gene-The PP2C family of protein phosphatases are commonly known as negative regulators of the MAPK cascade in various organisms, including the StyI/Spc1 pathway in Schizosaccharomyces pombe, and the p38 and c-Jun N-terminal kinase pathways in mammalian cells (25,26). Likewise, the HOG pathway of S. cerevisiae is negatively regulated by the Ptc1, -2, and -3 PP2C family of phosphatases (3,9). However, the physiological function of Ptc4 is not understood. We presumed that the phenotypic suppression might have been the result of the down-regulation of the HOG pathway by PTC4 overexpression. If this were the case, the lethality of the ⌬cnb1 ⌬mpk1 strain should be suppressed by a defective HOG pathway. As expected, the ⌬cnb1 ⌬mpk1 ⌬hog1 triple deletion mutant was viable on YPD plates at 30°C and even at 37°C (Fig. 1A). In contrast, the ⌬cnb1 ⌬mpk1 double deletion had no significant effect on the osmosensitivity of the ⌬hog1 strain (Fig. 1A), suggesting that the Ca 2ϩ signaling pathway and the HOG pathway are antagonistic in the growth regulation, but not in the high osmolarity response.
Hog1 activation is the result of the phosphorylation of both Thr-174 and Tyr-176. To examine whether the PTC4 gene down-regulates Hog1, we compared the levels of Hog1 phosphorylation in the strains with or without overexpressing WT (W303), ⌬hog1 (SOS5), ⌬cnb1 (DHT14), ⌬mpk1 (TNP46), ⌬cnb1 ⌬mpk1 (TNP36), or ⌬hog1 ⌬cnb1 ⌬mpk1 (SOS48) cells were spotted on agar plates of YPD, or YPD ϩ 1 M sorbitol (ϩSor), and cultured for 2 days at the indicated temperature. B, effect of PTC4 overexpression on the high osmolarity-induced Hog1 phosphorylation. WT (W303) cells transformed with YEp51 (vector) or YEp51-PTC4 (pGAL-PTC4) were grown in SR minus leucine (same composition as SD minus leucine, except 2% raffinose was used instead of dextrose) to early log phase and shifted to SG minus leucine to induce the PTC4 gene expression and incubated for 1.5 h before exposure to 0.3 M KCl. Samples were taken at the indicated times. Whole-cell extracts were prepared, and immunoblotting was performed with anti-phospho-p38 or anti-HA as a loading control. C, Hog1 kinase activity was measured using the same cell extracts. Immunoprecipitated Hog1-HA was incubated with MBP and [␥-32 P]ATP. The level of Hog1-HA in lysate was assessed by immunoblotting with anti-HA antibody. The radiolabel incorporated into MBP was quantified by BAS1800 Bioimaging analyzer.
PTC4 gene from the GAL1 promoter by using polyclonal antibody against phospho-p38, which cross-reacts with phospho-Hog1. As expected, the level of the high osmolarity-induced Hog1 phosphorylation was significantly reduced by PTC4 overexpression (Fig. 1B). This result was further verified by measuring the Hog1 kinase activity using [␥ 32 -P]ATP and MBP as a substrate. Indeed, the kinase activity was reduced by PTC4 overexpression (Fig. 1C). To see whether the suppression of this lethality is the result of the inhibition of the HOG pathway, we examined whether the overexpression of other negative regulators (PTC1, -2, and -3 and PTP2) known for the HOG pathway can suppress the lethality of the ⌬cnb1 ⌬mpk1 strain. Overexpression of the PTC1, -2, or -3 gene could suppress the lethality to a degree similar to the overexpression of the PTC4 gene ( Fig. 2A). However, overexpression of the PTP2 gene failed to suppress the lethality ( Fig. 2A). We next examined whether overexpression of PTC4 or PTP2 gene truly reduced the Hog1 kinase activity in the ⌬cnb1 ⌬mpk1 strain. As the result, overexpression of these genes reduced the basal level activity of Hog1 kinase in the ⌬cnb1 ⌬mpk1 strain (Fig. 2B). The PTP2 overexpression inhibited the Hog1 activity, but it could not suppress the growth defect (Fig. 2B). For the difference in the effect of the overexpression of these genes on the growth of the ⌬cnb1 ⌬mpk1 strain, see "Discussion." ⌬hog1 Mutant Displays Ca 2ϩ -induced G 2 Delay and Polarized Bud Growth-Deletion of the HOG pathway has been shown to partially suppress the defect of the mutants in cell wall integrity signaling in a manner mediated by the SVG (STE vegetative growth) pathway that regulates the cell wall integrity in parallel with the Mpk1 MAPK pathway (27). In this process, the HOG pathway seemed to negatively regulate cell wall integrity through the SVG pathway, antagonizing the Mpk1 MAPK pathway. Supporting this possibility, it was shown that the defect of the HOG pathway causes an activation of the STE pathway (28), which in turn may lead to the activation of cell wall integrity.
The HOG pathway seems to be also involved in G 2 /M cellcycle regulation. Deletion of HSL1 gene, which encodes a negative regulator of Swe1 kinase, suppressed the lethality caused by the double deletions of the CNB1 and MPK1 genes, suggesting that the HOG pathway and the Ca 2ϩ signaling pathways also antagonize in G 2 /M cell-cycle progression (data not shown). In the Ca 2ϩ -triggered, Swe1-mediated G 2 /M cell-cycle regulatory mechanism, Zds1 inhibits SWE1 transcription in G 2 phase, antagonizing the Ca 2ϩ signaling pathways (21. 29). Therefore, the ⌬zds1 strain exhibits various Ca 2ϩ phenotypes, including hypersensitivity to CaCl 2 (Ͼ50 mM), G 2 cell-cycle delay, and polarized bud growth. If the HOG pathway and the Ca 2ϩ signaling pathways antagonize one another in cell-cycle regulation, it would be expected that the ⌬hog1 strain growing in the presence of CaCl 2 would display a phenotype analogous to that of the ⌬zds1 strain. This was indeed the case. The ⌬hog1 strain exhibited the Ca 2ϩ phenotypes similar to those of the ⌬zds1 strain (Fig. 3, A and B). The similar phenotypes were not observed with the ⌬hog1 strain by sorbitol or other cations (e.g. Mg 2ϩ and Na ϩ ), suggesting that the Ca 2ϩ phenotypes exhibited by the ⌬hog1 strain were distinct from those of the high osmolarity of the solutes (Fig. 3A and data not shown).
The Ca 2ϩ phenotypes of the ⌬zds1 strain were shown earlier to be suppressed by the deletion of the CNB1, MPK1, or SWE1 gene (21). So we asked whether calcineurin, Mpk1, or Swe1 is required for the phenotype displayed by the ⌬hog1 strain. The effects of CaCl 2 on the ⌬hog1 strain were all rescued by the deletion of CNB1 or SWE1, but not that of MPK1 (Fig. 3, A and  B). However, because the effect of the SWE1 deletion was only partial, it was suggested that the HOG pathway has additional target(s) in the Ca 2ϩ -mediated growth regulation.
Swe1 Degradation in the ⌬hog1 Strain Is Delayed by Ca 2ϩ -The SWE1 mRNA level of WT cells fluctuates during the cell cycle, peaking in G 1 /S and declining just before the onset of mitosis (21,29). We first examined whether the HOG pathway is involved in the transcriptional regulation of the SWE1 gene by measuring the SWE1 mRNA levels in the presence or absence of 50 mM CaCl 2 in a synchronized cell cultures by ␣-factor. The fluctuation pattern was not dependent on the presence of the HOG1 gene, indicating that Hog1 is not important for the transcriptional regulation of the SWE1 gene (data not shown).
We next examined whether the HOG pathway is involved in post-transcriptional Swe1 regulation. The accumulation of Swe1 is periodic during the cell cycle, peaking at the time of bud emergence and declining before nuclear division (30, 31). We constructed various strains (WT, ⌬cnb1, ⌬hog1, and ⌬hog1 ⌬cnb1, each on the ⌬swe1 background) expressing a Myc . The transformants were spotted on YPD, or YPD ϩ 1 M sorbitol (ϩSor) and YPG and cultured for 2 days at 30°C. B, the ⌬cnb1 ⌬mpk1 (TNP36) cells transformed with galactose-inducible plasmids bearing PTC4 or PTP2 gene (pNV7-PTC4 and pSSP25) or empty vector were grown in SR minus uracil containing osmotic stabilizer (1 M sorbitol) to mid-log phase and shifted to SG minus uracil and induced gene expression. Samples were taken at the indicated times, and Hog1 kinase activity was measured (see "Materials and Methods"). The level of immunoprecipitated Hog1 was assessed by immunoblotting with anti-Hog1 antibody. The radiolabel incorporated into MBP was quantified by BAS1800 Bioimaging analyzer. epitope-tagged Swe1 from the chromosomally integrated gene and compared the fluctuation patterns of Myc-tagged Swe1 by immunoblot analysis in synchronized cell cultures. The Myc-Swe1 construct was fully functional in vivo (data not shown). The G 1 cells synchronized with ␣-factor were released into YPD medium with or without 50 mM CaCl 2 . The cell-cycle progression in synchronous cell cultures was monitored by the determination of nuclear division and DNA contents (Fig. 4, B and C). As previously reported (30,31), the Swe1 level of the WT strain fluctuated during the cell cycle. The slower migrating species of Swe1 represent a phosphorylated isoform, and the phosphorylation occurs prior to its degradation coincidental with the onset of mitosis (31). Immediately after the degradation, the faster migrating species of Swe1 re-appeared. In YPD medium, there was no significant difference among the strains in the periodic patterns of the Swe1 level and the timing of nuclear division. However, in the ⌬hog1 cells cultivated with CaCl 2 , the Swe1 level was sustained for 30 min longer, and correspondingly, the onset of mitosis was delayed by 30 min, in comparison with those in WT cells treated similarly. This result suggests that Ca 2ϩ caused a delay in Swe1 degradation in the ⌬hog1 strain. Both Ca 2ϩ -induced Swe1 stabilization and the delay of nuclear division in the ⌬hog1 strain were shortened to 15 min by CNB1 deletion, indicating that the Swe1 stability is antagonistically regulated in a manner mediated by the HOG pathway and calcineurin (compare ⌬hog1 and ⌬hog1 ⌬cnb1 strains in Fig. 4A).
The HOG Pathway Promotes Bud Emergence, whereas Calcineurin Inhibits Actin Polarization-Using the same synchronous cell cultures described in the previous section, we also determined the time course of bud emergence (Fig. 4B). In YPD medium, no significant difference was observed among the strains in the timing of bud emergence. In the presence of 50 mM CaCl 2 , bud emergence in the ⌬hog1 strain was delayed by 30 min compared with that in the WT strain. In the presence of sorbitol, the ⌬hog1 strain showed a delay of bud emergence by 15 min compared with that of the WT strain. The Ca 2ϩ -induced delay in the ⌬cnb1 strain was shortened to 15 min (compare (SHI63), ⌬hog1 (SHI27), or ⌬hog1 ⌬cnb1 (SHI30) strains were grown in YPD to early-log phase, and the cells were then synchronized as described under "Materials and Methods." Whole-cell extracts were prepared and analyzed by immunoblotting with an antibody against Myc-Swe1 and one against Cdc28 as a loading control. B, quantification of cumulative percentage of nuclear division and bud formation during cell-cycle progression in various strains in the cell cultures prepared as in A was determined by microscopic observation following PI staining; at least 300 cells were counted for each time point. The cells were released into YPD medium with or without 50 mM CaCl 2 (ϩCa 2ϩ ) or 150 mM sorbitol (ϩSor). Samples were taken at an interval of 15 min after removing the ␣-factor. C, DNA content of various strains in a synchronized culture prepared as in A after release from arrest in G 1 . ⌬hog1 and ⌬hog1 ⌬cnb1 in Fig. 4B). In contrast, the sorbitolinduced delay was not altered in the ⌬cnb1 strain (Fig. 4B). Moreover, no significant delay was observed in DNA replication of various strains in the presence of 50 mM CaCl 2 , indicating that the Ca 2ϩ -induced delay of bud formation is not simply a reflection of the decelerated cell cycle by Ca 2ϩ (Fig. 4C). Taken together, these results indicated that bud emergence is antagonistically regulated, positively by the HOG pathway and negatively by calcineurin.
To address whether Ca 2ϩ , rather than the osmotic stress, is specifically responsible for the severe delay in mitosis caused by CaCl 2 , we examined whether osmotic stress of sorbitol might induce a similar effect. In YPD medium containing 150 mM sorbitol, which gives an osmolarity equivalent to that of 50 mM CaCl 2 , both Swe1 degradation and the nuclear division of the ⌬hog1 cells were delayed by 15 min compared with those of WT cells. This delay was shorter than that caused by CaCl 2 (30 min; Fig. 4, A and B). Moreover, the delays in Swe1 degradation and cell-cycle progression caused by sorbitol were not alleviated by the deletion of calcineurin, suggesting that the mechanisms underlying the delays were different for Ca 2ϩ and sorbitol (Fig. 4, A and B).
To further clarify the roles of these pathways in the control of bud emergence, we examined the effects of CaCl 2 on bud emergence and actin polarization at the bud site in ␣-factor-synchronized cells (Fig. 5). Actin polarization was determined by rhodamine-phalloidin staining. No significant difference was observed in the kinetics of actin polarization among the strains (WT, ⌬cnb1, ⌬hog1, and ⌬hog1 ⌬cnb1) after release of the cells to YPD medium. The timing of actin polarization of WT and the ⌬hog1 cells was only slightly delayed after release to the medium containing 50 mM CaCl 2 (compare ϮCa 2ϩ at 15 min for WT and ⌬hog1 cells). However, bud emergence of the ⌬hog1 cells was delayed from that of WT cells by 30 min (compare WT and ⌬hog1 cells, ϩCa 2ϩ at 45-60 min), although the actin patches were localized apparently normally at the presumptive bud site. The delay in actin polarization was nearly completely abolished in the ⌬cnb1 cells, although bud emergence was still delayed by 15 min (compare WT and ⌬hog1 cells, ϩCa 2ϩ at 15-30 min). The Ca 2ϩ -induced delay of bud emergence in the ⌬hog1 strain was only partially suppressed by the deletion of calcineurin (compare ⌬hog1 and ⌬hog1 ⌬cnb1 cells, ϩCa 2ϩ at 45 min).
Calcineurin is known to regulate cellular Ca 2ϩ homeostasis. Defect of calcineurin leads to increase Ca 2ϩ tolerance by increasing vacuolar sequestration of Ca 2ϩ from the cytosol (32,33). Therefore, the effect of ⌬cnb1 on actin polarization could be the result of an alteration of cellular Ca 2ϩ homeostasis. We investigated whether calcineurin truly regulates actin polarization by using a constitutively active form of calcineurin, containing the C-terminal truncated form of catalytic subunit (CMP2⌬C). The galactose-inducible GAL1-CMP2⌬C construct was introduced to the URA3 locus of WT strain and an experiment similar to that described above was performed without addition of CaCl 2 . Overexpression of CMP2⌬C inhibited actin polarization to the bud site (Fig. 6, compare ϩglucose (off) and ϩgalactose (on)). The overexpression of CMP2⌬C did not lead to a delay in the progression of DNA replication (Fig. 6B) consistently with the results obtained with the addition of Ca 2ϩ (Fig. 4C). These results indicated that the HOG pathway promotes bud emergence after actin polarization, whereas calcineurin inhibits actin polarization (see "Discussion").
The Calcineurin-Crz1/Tcn1 Pathway Is Required for the Down-regulation of the HOG Pathway-To further investigate the antagonizing mechanism between the 2 pathways, we examined whether calcineurin down-regulates the HOG pathway by determination of phospho-Hog1 with anti-phospho-p38 antibody. In WT cells, the Hog1 phosphorylation induced by CaCl 2 was transient, peaking at 15 min and disappearing by 30 min of the treatments (Fig. 7A). In the ⌬cnb1 strain, Hog1 phosphorylation was induced by CaCl 2 similarly as in WT strain, but the phosphorylation level increased to a much higher level and sustained longer than in the WT one (Fig. 7A).
We further examined whether Crz1/Tcn1, a transcription factor activated by calcineurin, was involved in this process (17,18). The level of Ca 2ϩ -induced Hog1 phosphorylation in the ⌬crz1 strain was similarly high as that in the ⌬cnb1 strain (Fig.  7A). Moreover, ⌬crz1 and ⌬cnb1 mutations did not exhibit an additive effect on the level of Hog1 phosphorylation, suggesting that calcineurin and Crz1 regulate Hog1 phosphorylation in a common pathway (data not shown). The Ca 2ϩ -induced Hog1 hyperphosphorylation in the ⌬crz1 strain led to an increased expression level of GPD1 mRNA compared with that in the WT strain, suggesting that the negative regulation of the HOG pathway was mediated by the Crz1 transcription factor (data not shown). We next examined the possibility that various protein phosphatase(s) known to inhibit the HOG pathway are involved in the regulation of Hog1 by Crz1. We examined by Northern blot analysis the effect of CaCl 2 in WT cells on the mRNA levels of various phosphatases, which included the PP2C family phosphatases (Ptc1, -2, -3, and -4) and the proteintyrosine phosphatases (Ptp2 and -3). Of the mRNA levels of these phosphatase genes, only that of PTP2 gene was moderately decreased by the deletion of CRZ1 gene (data not shown). However, the calcineurin-Crz1 pathway-dependent regulation of PTP2 transcription seemed inadequate to explain the Hog1 hyperphosphorylation induced by Ca 2ϩ . The activation of the HOG pathway in response to high osmolarity is mediated by two upstream branches, one through the Sho1 sensor and the other through the two-component system consisting of Sln1, Ypd1, and Ssk1 (2-6). To address which of the two branches is responsible for the inhibition of Hog1 by Ca 2ϩ , we compared the Ca 2ϩ -induced Hog1 phosphorylation levels among the strains deleted for a component of the HOG pathway, together with the regulatory subunit of calcineurin (i.e. ⌬cnb1 ⌬pbs2, ⌬cnb1 ⌬ste11, and ⌬cnb1 ⌬ssk1 strains) (Fig. 7A). Hog1 phosphorylation was not detectable in the ⌬cnb1 ⌬pbs2 strain, indicating that CaCl 2 activates Hog1 through Pbs2 (Fig. 7A). The Ca 2ϩ -induced Hog1 hyperphosphorylation was observed in the ⌬cnb1 ⌬ste11 strain, but not in the ⌬cnb1 ⌬ssk1 strain, suggesting that the calcineurin-Crz1 pathway inhibits the Sln1, but not Sho1, branch, of the HOG pathway. Moreover, the enhancement of Hog1 phosphorylation in ⌬cnb1 strain was also observed by sorbitol (300 mM) that causes an osmolarity equivalent to that of 100 mM CaCl 2 (Fig. 7B). It was recently shown that a transient increase in the cytosolic Ca 2ϩ is induced from the vacuolar stores by hyper-osmotic stress (34). The calcineurin-Crz1 pathway is important in the regulation of intracellular Ca 2ϩ homeostasis. Therefore, these results suggested that the increase in cytosolic Ca 2ϩ may down-regulate the HOG pathway through the regulation of the Sln1 branch (see "Discussion"). DISCUSSION HOG Pathway and Calcineurin Antagonize in Growth Regulation-We have shown here that the HOG1 deletion remarkably suppressed the lethality caused by the simultaneous loss of calcineurin and the Mpk1 MAPK pathway. Thus, the ⌬cnb1 ⌬mpk1 ⌬hog1 triple disruption mutant grew apparently normally on YPD plate at 30°C and even at 37°C, a restrictive temperature for the ⌬mpk1 strain. The lethality of the ⌬cnb1 ⌬mpk1 strain is attributable mainly to a defect in the cell wall construction for the following reasons. 1) The lethality is rem- 5. Calcineurin negatively regulates actin polarization and bud emergence, whereas Hog1 positively regulates bud formation after actin polarization. WT (W303), ⌬cnb1 (DHT14), ⌬hog1 (SOS5), or ⌬hog1 ⌬cnb1 (SOS31), ⌬ptp2 (SH7687), or ⌬ptp2 ⌬cnb1 (SHI334) was grown in YPD medium to early-log phase, and cells were synchronized as described under "Materials and Methods." The cells were released into edied by osmotic stabilizers added to the medium (20). 2) The expression of the FKS2 gene, which encodes a glucan synthetase, is activated by the calcineurin-Crz1 pathway (35). 3) The Mpk1 MAPK cascade has been implicated in the expression of various genes involved in cell wall synthesis (1). The observation that the triple disruption mutant was viable on YPD plates even at 37°C indicates that the HOG pathway negatively regulates the cell wall construction, antagonizing the Ca 2ϩ signaling pathways.
Besides the cell wall, calcineurin and the Mpk1 pathway coordinately regulate the G 2 /M cell-cycle progression through the regulation of Swe1 (21). Because the suppression of the synthetic lethality by the HOG1 deletion was very potent, the HOG pathway and the Ca 2ϩ signaling pathways were thought to counteract each other in diverse events of growth regulation, including cell wall synthesis and cell-cycle progression. In fact, the synthetic lethality was partially suppressed by the deletion of the Nim1-like kinase Hsl1, the negative regulatory kinase of Swe1 (data not shown). The involvement of the HOG pathway in cell-cycle regulation was further supported by the observations that the HOG pathway regulates stability of Swe1 (Fig.  3A) and the effect of HOG1 deletion was partially suppressed by the SWE1 deletion (data not shown).
The ⌬cnb1 ⌬mpk1 synthetic lethality was suppressed by overexpression of PTC1, -2, -3, or -4, encoding the serine/threonine phosphatase PP2C family known as the negative regu-lator of the HOG pathway, supporting the notion that the HOG pathway and the Ca 2ϩ signaling pathways antagonize each other in growth regulation. However, overexpression of the tyrosine phosphatase Ptp2 failed to suppress the lethality. This result suggests that the PP2C overexpression suppressed the growth of the ⌬cnb1 ⌬mpk1 strain by dephosphorylating a target protein phosphorylated by the serine/threonine kinase Hog1. Further investigation is necessary to elucidate the mechanism of the lethality.
Calcineurin and the HOG Pathway Regulate Distinct Processes of Bud Emergence-The HOG pathway and calcineurin/ Mpk1 MAPK pathways perform antagonistic roles in the regulation of bud emergence. The detailed analysis of the kinetics of the bud emergence and the actin localization in synchronized cell cultures revealed that calcineurin and the HOG pathway regulate distinct processes of bud emergence (Fig. 5). In the WT strain, the actin polarization at the presumptive bud site and bud emergence were shortly delayed (ϳ15 min) by Ca 2ϩ . The delays were abrogated in the ⌬cnb1 strain, indicating that the delays occurred in a calcineurin-dependent manner. Compared with the wild-type strain, the delay in bud emergence induced by Ca 2ϩ in the ⌬hog1 strain was longer. Although bud emergence in the ⌬hog1 strain was severely delayed by Ca 2ϩ , actin polarization was only slightly affected. These results indicated that the Hog1 activation stimulated bud emergence at a step after actin had correctly become localized at the bud site. The Ca 2ϩ -induced delay of bud emergence in the ⌬hog1 strain was only partially suppressed by the deletion of calcineurin, suggesting that the CNB1 deletion can suppress the defect in actin polarization, but not the bud emergence. Taken together, these results indicated that the processes of the bud emergence that are regulated by calcineurin and the HOG pathway are distinct. Calcineurin inhibits actin polarization, an early step of bud emergence, whereas Hog1 promotes bud emergence at a step after actin has become polarized. The identification of the targets of calcineurin in regulating actin polarization and that of the HOG pathway in bud emergence still remains to be made.
The G 2 delay caused by 50 mM CaCl 2 in ⌬hog1 strain may have been the result of the activation of the morphogenesis checkpoint. Supporting this notion, bud formation in the ⌬hog1 strain was severely delayed, although DNA replication was not significantly affected by CaCl 2 (Fig. 4, B and C). The morphogenesis checkpoint monitors abnormal bud formation and is functional in late G 1 to S phases, but is nonfunctional once the bud has grown to a critical size (31). It is possible that the defect of bud emergence caused by deregulated synthesis of the cell wall and the inhibition of the actin polarization in the ⌬hog1 may activate the morphogenesis checkpoint. Why did the ⌬hog1 strain exhibit a severe defect in bud emergence in the presence of Ca 2ϩ ? As a yeast cell buds, the cell wall must expand to fit the expanding plasma membrane. During expansion of the rigid cell wall, it must be weakened at the point of new growth to add newly made cell wall material. The severe defect in bud emergence in the ⌬hog1 cells would be explained by the deregulated synthesis of the cell wall at the bud site by the hyperactivation of the cell wall synthesis as a result of the absence of its negative regulator (Hog1) and by the activation of the Ca 2ϩ signaling pathways. Alternatively, it is also possible that the delay in bud emergence might have resulted from the YPD medium with or without 50 mM CaCl 2 (ϩCa 2ϩ ). Samples were taken at an interval of 15 min after removing ␣-factor and fixed with formaldehyde. A, actin staining with rhodamine-phalloidin. The arrowheads indicate polarized actin at the incipient bud site, and the asterisks indicate defective bud emergence in ⌬hog1 cells. B, the results were quantified and plotted as time after release versus percentage of actin patches localized at the incipient bud site (solid column) and bud formation (line); at least 300 cells were counted for each time point. White bars and circular symbols, in YPD medium. Light gray bars and square symbols, in YPD plus 50 mM CaCl 2 .
FIG. 6. Overexpression of constitutively active calcineurin inhibits the actin polarization to the bud site. A, WT (W303) strain and GAL-CMP2⌬C integrating strain (SHI290) was grown in YPR to early-log phase, and cells were synchronized as described under "Materials and Methods." The cells were released into YPD or YPG medium. Samples were taken at an interval of 15 min after removing the ␣-factor, and cells were fixed with formaldehyde. The actin polarization and the bud formation of the cells were monitored by microscopic observation with rhodamine-phalloidin staining. The results were quantified and plotted as time after release versus percentage of actin patches localized at the incipient bud site (solid column) and bud formation (line); at least 300 cells were counted for each time point. White bars and circular symbols, in YPD medium. Light gray bars and square symbols, in YPG medium. B, DNA content of various strains in a synchronized culture prepared as in A after release from arrest in G 1 . defect in recovery from the inhibition of protein synthesis. It was recently reported that the ⌬hog1 mutant was defective in the recovery from hypertonic stress-induced inhibition of protein synthesis (36).
It was suggested that Ca 2ϩ signal inhibited bud emergence through inhibition of Hog1 in a manner mediated by the calcineurin-Crz1 pathway and through inhibition of actin polarization by calcineurin. If so, the inhibitory effect of Ca 2ϩ on bud emergence would be abolished by the activation of Hog1. This was indeed the case. The delay in bud emergence that was still seen in the ⌬cnb1 strain was apparently totally abolished by the deletion of the PTP2 gene (Fig. 5B).
We showed here that calcineurin was involved in the inhibition of actin polarization at the bud site. In the pathogenic fungus Cryptococcus neoformans, it was earlier suggested that calcineurin was required for the polar cell growth in mating and haploid fruiting (37). In the nerve cells, calcineurin was implicated in cell polarization (38).
The Ca 2ϩ -induced inhibition of actin polarization in ⌬hog1 strain was not suppressed by the deletion of CRZ1 (data not shown), suggesting that other target molecule(s) are involved in the regulation of actin polarization by calcineurin.
Cross-talk between the HOG Pathway and Ca 2ϩ Signaling Pathway-Because the HOG pathway and the Ca 2ϩ signaling pathways are regulated opposingly by high and low external osmolarity, respectively, the question arises as to whether these pathways regulate each other. Our data suggested that the calcineruin-Crz1 pathway was involved in the down-regu-lation of the HOG pathway through the regulation of the Sln1, but not Sho1, branch (Fig. 7A; see a model in Fig. 8). Further, Hog1 phosphorylation was sustained with the ⌬ste11 ⌬cnb1 strain, but not with the ⌬ssk1 ⌬cnb1 strain, suggesting that the calcienrin-Crz1 pathway is involved in the inhibition of the Sln1 pathway. The phosphorylation level of Hog1 in the ⌬ssk1 ⌬cnb1 strain decreased after 45 min of its activation, suggesting that the down-regulation of the Sln1 branch by the cal-  7. The calcineurin-Crz1 pathway required for the down-regulation of HOG pathway. A, the effect of deletion of CNB1, CRZ1/ TCN1, PBS2, STE11, or SSK1 deletion on the Ca 2ϩ -induced hyperphosphorylation of Hog1. WT (SHI119), ⌬cnb1 (SHI121), ⌬crz1 (SHI288), ⌬cnb1 ⌬pbs2 (SHI318), ⌬cnb1 ⌬ste11 (SHI131), or ⌬cnb1 ⌬ssk1 (SHI313) was grown in YPD medium to early-log phase, and the cells were then exposed to 100 mM CaCl 2 (ϩCa 2ϩ ). Samples were taken at an interval of 15 min after addition of CaCl 2 . Whole-cell extracts were prepared and immunoblotted with an antibody against phospho-p38 and with one against HA as a loading control. B, the effect of sorbitol on Hog1 phosphorylation in ⌬cnb1 strain. WT (SHI119) or ⌬cnb1 (SHI121) strain was grown in YPD medium to early-log phase, and the cells were then exposed to 300 mM sorbitol. Immunoblotting was performed as described in A. C, outline of the upstream branches of HOG pathway. cineurin-Crz1 pathway is important in the cross-talk regulation of the HOG pathway (Fig. 7A). The Hog1 phosphorylation in the ⌬cnb1 strain induced by 100 mM Ca 2ϩ was significantly higher and sustained longer than that by 300 mM sorbitol, suggesting that Ca 2ϩ is more potent than sorbitol in Hog1 activation (compare the phosphorylation levels of "⌬cnb1" samples in Fig. 7, A and B). Because hyperosmotic stress induces a transient increase in cytosolic Ca 2ϩ from vacuolar stores (34), the difference in the effect of CaCl 2 and sorbitol on Hog1 phosphorylation may simply reflect the difference in the effect of exogenous CaCl 2 and hyperosmotic stress on the elevation of intracellular Ca 2ϩ . How does Ca 2ϩ activate the Sln1 branch of the HOG pathway? As previously suggested, the osmo-activated Sln1-Ypd1-Ssk1 phosphorelay system may be under a negative feedback regulation by accepting a phosphate group from glycerol 3-phosphate or related metabolic intermediate donors, which may be overproduced by glycerol-3-phosphate dehydrogenase, a target of the HOG pathway (5). The loss of calcineurin activity leads to elevation of the cytosolic Ca 2ϩ concentration (32,39,40). If such a feedback regulatory mechanism that is mediated by a phosphorylated metabolite does exist for the HOG pathway, the elevation of intracellular Ca 2ϩ may cause the formation of an insoluble complex with the phosphate compound, interfering with re-phosphorylation of the two-component system. Alternatively, it is possible that unknown Crz1 target genes are involved in this process. The mechanism by which the calcineurin-Crz1 pathway regulates the Sln1 branch still remains to be seen. Our observations provide evidence for a cross-talk between the Ca 2ϩ signaling pathway and the HOG pathway. The activation of the p38 MAPK cascade by Ca 2ϩ signal was recently reported with the p38 MAPK cascade of Caenorhabditis elegans (41). In mammalian cells, the p38 MAPK pathway is inhibited by calcineurin in a manner mediated by the transcriptional up-regulation of MAPK phosphatase-1 (42). It is possible that regulation of the p38 family MAPK pathway by the Ca 2ϩ signaling pathways is conserved from yeast to higher eukaryotes.
On the basis of these data, we propose a model for the regulation of cell growth by the HOG pathway and calcineurin (Fig. 8).