The Elm1 Kinase Functions in a Mitotic Signaling Network in Budding Yeast

Abstract

In budding yeast, the Clb2 mitotic cyclin initiates a signaling network that negatively regulates polar bud growth during mitosis. This signaling network appears to require the function of a Clb2-binding protein called Nap1, the Cdc42 GTPase, and two protein kinases called Gin4 and Cla4. In this study, we demonstrate that the Elm1 kinase also plays a role in the control of bud growth during mitosis. Cells carrying a deletion of the ELM1 gene undergo a prolonged mitotic delay, fail to negatively regulate polar bud growth during mitosis, and show defects in septin organization. In addition, Elm1 is required in vivo for the proper regulation of both the Cla4 and Gin4 kinases and interacts genetically with Cla4, Gin4, and the mitotic cyclins. Previous studies have suggested that Elm1 may function to negatively regulate the Swe1 kinase. To further understand the functional relationship between Elm1 and Swe1, we have characterized the phenotype of Δelm1 Δswe1 cells. We found that Δelm1 Δswe1 cells are inviable at 37°C and that a large proportion of Δelm1 Δswe1 cells grown at 30°C contain multiple nuclei, suggesting severe defects in cytokinesis. In addition, we found that Elm1 is required for the normal hyperphosphorylation of Swe1 during mitosis. We propose a model in which the Elm1 kinase functions in a mitotic signaling network that controls events required for normal bud growth and cytokinesis, while the Swe1 kinase functions in a checkpoint pathway that delays nuclear division in response to defects in these events.

Members of a family of proteins called cyclin-dependent kinases control the events of the eukaryotic cell cycle (33, 34). These proteins associate with members of the cyclin family of proteins to form active kinase complexes that induce specific cell cycle events. The events leading to activation of cyclin-dependent kinase complexes are well understood; however, little is known about the molecular pathways that are initiated by cyclin-dependent kinases to control specific cell cycle events. One possibility is that cyclin-dependent kinases act to directly phosphorylate proteins involved in specific cell cycle events, such as nuclear lamins, components of the DNA replication machinery, or microtubule-associated proteins. Alternatively, cyclin-dependent kinases may activate intricate signaling pathways involving multiple additional kinases that are ultimately responsible for phosphorylating the many proteins involved in cell cycle events.

The control of bud growth during mitosis in Saccharomyces cerevisiae provides an excellent model system in which to understand the pathways used by cyclin-dependent kinases to induce specific cell cycle events (1, 8, 21, 25, 44). A new bud emerges from the mother cell during interphase and grows in a polar manner with actin localized at the bud tip. Upon entry into mitosis, the mitotic cyclins induce a reorganization of the actin cytoskeleton that causes the bud to grow over its entire surface, and cells that lack the function of the mitotic cyclins therefore develop highly elongated buds (2, 37). This switch in the pattern of bud growth is not necessary for viability, and mutations that disrupt the switch can be easily identified because they cause cells to have highly elongated buds and to form colonies with an unusual morphology (1, 5). Although cellular division by budding does not occur in all organisms, many of the proteins that function in the pathway that controls bud growth during mitosis are highly conserved, suggesting that similar pathways are used by other organisms to control mitotic events.

The switch in the pattern of bud growth that occurs during mitosis is primarily under the control of the Clb2 mitotic cyclin, although it is clear that the switch can also be induced by other redundant mitotic cyclins, perhaps through independent pathways (20, 24, 25). We have used a combination of genetics and biochemistry to identify proteins that are required for the mitotic switch in the pattern of bud growth (1, 8, 21, 44). These experiments have provided evidence for the existence of an intricate signaling network that includes a Clb2-binding protein called Nap1, the Cdc42 GTPase, members of the septin family, and two protein kinases called Cla4 and Gin4. The Gin4 kinase binds tightly to Nap1 and is activated by hyperphosphorylation during mitosis. The mitosis-specific activation of Gin4 is dependent in vivo upon the function Nap1, Cla4, the GTP-bound form of Cdc42, and the septins (1, 44). The Cla4 kinase is also hyperphosphorylated during mitosis, and hyperphosphorylation requires the activity of Clb2, Cdc28, Nap1, and the GTP-bound form of Cdc42, but not Gin4 (44). The hyperphosphorylated form of Cla4 appears to be responsible for relaying the signal to activate the Gin4 kinase (44).

We do not yet understand how the Clb2-Cdc28 kinase complex may relay the signal to activate the Cla4/Gin4 pathway, nor do we understand the molecular mechanisms underlying the mitosis-specific regulation of the Gin4 and Cla4 kinases. To learn more about this signaling network, we have used a genetic approach to identify new mutations that disrupt the switch from polar to isotropic bud growth. This work has demonstrated that a protein kinase called Elm1 functions in the signaling network that controls bud growth during mitosis. In addition, we have found evidence suggesting that the Swe1 kinase functions in a checkpoint that delays nuclear division in response to defects in the pathway that includes Elm1. Our results are consistent with previous work that has suggested a role for Swe1 in a checkpoint that monitors the proper organization of actin and/or septin filaments during the cell cycle (4, 23, 24).

MATERIALS AND METHODS

Strains and culture conditions.

Except where noted, all cells were grown in yeast extract-peptone-dextrose (YPD) media. All strains are in the W303 strain background (leu2-3,112 ura3-52 can1-100 ade2-1 his3-11 trp1-1 ssd ho). The additional features of the strains used in this study are listed in Table 1.

TABLE 1.

Features of the strains used in this study

Strain	Genotype	Source or reference
DK133	MATa LEU+	See text
AS1	MATa Δelm1::TRP1 Δbar1	See text
AS22	MATa Δelm1::URA Δbar1 Δclb1 Δclb3::TRP1 Δclb4::HIS3	See text
DK186	MATa Δbar1	1
DK212	MATa Δclb1 Δclb3::TRP1 Δclb4::HIS3 Δbar1	1
HT1	MATa Δcla4::URA3 Δbar1	44
AS29	MATa Δcla4::URA3 Δelm1::TRP1 Δbar1	See text
RA5	MATa Δgin4::LEU2 Δbar1	1
AS14	MATa Δelm1::TRP1 Δgin4::LEU2 Δbar1	See text
AS15	MATa Δelm1::TRP1 Δnap1::LEU2 Δbar1	See text
AS40	MATa Δelm1::LEU2 URA3::gal1-CLB2Δ176/gal10-CDC42V12 Δbar1	See text
HT31	MATaURA3::gal1-CLB2Δ176/gal10-CDC42V12 Δbar1	44
AS19	MATa Δswe1::URA3 Δbar1	See text
AS43	MATa Δswe1::URA3 Δelm1::TRP1 Δbar1	See text
AS36	MATa Δcla4::LEU2 Δswe1::URA3 Δbar1	See text

Mutagenesis and screening.

Mutagenesis with ethylmethane sulfate was carried out as previously described with strain DK133 (22). After mutagenesis, half of the mutagenized culture was plated on YPD medium at a density of 1,000 colonies per 100-mm plate and screened for rough colonies. The other half of the mutagenized culture was used to inoculate 1 liter of YPD liquid medium overnight at 30°C. This culture was then passed through a 5-μm-pore-size mesh screen to separate severely clumpy cells from single cells and small clumps (15a). The cells that remained on the screen were washed into YPD and were then plated onto YPD medium at a density of 1,000 colonies per 100-mm plate and were grown at 30°C for 2 days before screening for rough colonies with a dissecting microscope. A total of 125,000 colonies were screened. Cells producing rough colonies were screened with phase-contrast optics to identify mutations that cause the formation of elongated buds.

Cloning of the ELM1 gene and construction of strains.

To clone the gene corresponding to the ecm41 mutation, we transformed mutant cells with a genomic library carried in a CEN-containing vector and screened for rescue of both the rough colony morphology and the elongated bud morphology. We obtained one rescuing plasmid, and sequencing revealed that it contained the ELM1 gene. The ELM1 gene was first identified in a similar screen for mutants that exhibited an elongated bud morphology (5). To delete the ELM1 gene, we used PCR to amplify the TRP1 gene from the pRS304 vector with short regions of homology to the ELM1 5′ and 3′ ends of the open reading frame at each end (oligonucleotides: ACTTACTCGCATAGATATTATTTTTTGAACGCCAGGTTAACAATAATTACTTAGCATGAAATGCGGCATCAGAGCAGA and CACATCGGCTATACGATTATCAGCTAACCCAATCCGACAGATATCATCCTGTAGTTTCATTCTGTGCGGTATTTCACA). The PCR product was transformed into a wild-type strain (DK186), and transformants that carry a deletion of the ELM1 gene were identified by their rough colony morphology and elongated buds and confirmed by PCR. Meiotic linkage experiments were used to demonstrate that the original ecm41 mutation cosegregates with the ELM1 gene deletion.

Double-mutant strains were generated by mating strain AS1 with RA5, HT1, AS19, or DK96, followed by sporulation and tetrad analysis. In each case 16 tetrads were dissected and analyzed for each double mutant.

Immunofluorescence methods and FACS analysis.

Fixation and staining of cells with antibodies were carried out as previously described (36). For the data shown in Fig. 3, more than 200 cells were counted for each time point. For fluorescence-activated cell sorter analysis (FACS), strains were grown overnight at 30°C to an optical density (OD) of 0.6, fixed in 70% ethanol for 1 h, and treated with 1 mg of RNase per ml overnight at 4°C and 5 mg of proteinase K per ml for 5 min at 30°C. The cells were then stained with propidium iodide, and DNA content was measured by flow cytometry.

FIG. 3.

Elm1 is required for normal progression through mitosis. (A) Wild-type and Δelm1 cells were released from an α-factor arrest, and the percentage of cells with a short mitotic spindle was determined as a function of time during the cell cycle. (B) Western blots show the amount of the Clb2 protein present in wild-type and Δelm1 cells as a function of time after release from an α-factor arrest. (C) Western blots showing the amount of the Clb2 protein present as a function of time after addition of α-factor to Δclb1,3,4 and Δelm1 Δclb1,3,4.

Cell cycle arrests, Western blotting, and kinase assays.

For all experiments, strains that are Δbar1 are arrested with 1 to 2 μg of α-factor per ml. Arrest with α-factor was carried out for 3 to 3.5 h at room temperature, with the exception of the experiment shown in Fig. 4, which was carried out for 5 h at 30°C. Arrest with benomyl is carried out for 3.5 h at room temperature at a final concentration of 30 μg/ml in YPD liquid medium. For all Western blotting experiments, 1.6-ml samples of culture were taken at each of the indicated time points. The cells were then rapidly pelleted in a 1.8-ml screw-top tube, the supernatant was removed, and the tube is frozen on liquid nitrogen. After all of the samples are collected, 300 ul of glass beads were added to each tube, followed by 130 μl of 1× protein gel sample buffer (65 mM Tris-HCl, pH 6.8; 3% sodium dodecyl sulfate [SDS], 5% β mercaptoethanol, 10% glycerol, bromphenol blue, 2 mM phenylmethylsulfonyl fluoride [PMSF], and 2 μg of leupeptin, 2 μg of pepstatin, and 2 μg of chymotrypsin per ml). The protease inhibitors were added immediately before the sample buffer was used. The tubes were immediately placed in a Biospec Multibeater-8 and beaten at top speed for 2 min, centrifuged briefly, and immediately incubated in a boiling water bath for 5 min. Then, 10 μl of each sample is loaded onto an SDS-containing polyacrylamide gel, and Western blotting was carried out as previously described. To see the phosphorylation-induced shift in the electrophoretic mobility of Gin4, samples were electrophoresed for 3 h at 180 V on 9% polyacrylamide gels (3). To observe the phosphorylation induced shift in the electrophoretic mobility of Swe1, samples were electrophoresed for 2 h at 180 V on 10% polyacrylamide gels. Cla4 and Clb2 Western blots and Gin4- and Clb2-associated kinase assays were done as previously described (21, 44).

FIG. 4.

Elm1 is not required for the formation of active Clb2-Cdc28 kinase complexes. A time course shows the appearance of Clb2-associated kinase activity during mitosis in wild-type and Δelm1 cells. Cells were released from an α-factor arrest and then assayed for Clb2-associated kinase activity during the cell cycle as previously described (21).

Overexpression of Clb2^Δ176 and Cdc42^V12.

Plasmid pHT19, which expresses Clb2^Δ176 and Cdc42^V12 from the Gal1 and Gal10 promoters, respectively, was integrated into Δelm1 strains as previously described (44). To induce expression of proteins in cells arrested in interphase, the cells were first grown overnight at 25°C to an OD₆₀₀ of 0.7 in YEP medium containing 2% glycerol and 2% ethanol. The cells were then arrested in interphase by the addition of α-factor to 2 μg/ml, followed by incubation at 30°C for 2 h. Protein expression was induced by the addition of galactose to 2%, and 1.6 ml samples of the culture were taken at the indicated times and prepared for electrophoresis and Western blotting as described above.

Generation of anti-Swe1 antibodies.

Antibodies that recognize Swe1 were raised by immunizing rabbits with a glutathione S-transferase (GST) fusion that included the C terminus of Swe1. The GST fusion was made by amplifying the C terminus of Swe1 by PCR (primers GCGGGATCCATGAGTT-CTTTGGACGAGG and AACCGAATTCCTTGCTCTTTT) and cloning into the BamHI and EcoRI sites of the pGEX-4T3 vector. The second primer incorporates an EcoRI site found 900 bases from the 5′ end of the SWE1 gene. The GST fusion is expressed in bacteria and purified by glutathione affinity chromatography as previously described (8).

A full-length MBP-Swe1 fusion was constructed by PCR amplification of the full-length SWE1 gene, followed by cloning into the vector pMal-C2 (primers GCGGGATCCATGAGTTCTTTGGACGAGG and CGCGCTGCAGTCATATAAAAAATTTTGGCTTAG). To purify the MBP-Swe1 fusion, cells carrying the MBP-Swe1 expression construct were grown in 6 liters of 2XYT media containing 100 μg of ampicillin per ml at room temperature until an OD of 0.7 was reached. IPTG (isopropyl-β-d-thiogalactopyranoside) was then added to a final concentration of 0.1 mM, and the cells were incubated for another 3 to 4 h at room temperature. The cells were harvested by centrifugation at room temperature, and the pellet was scraped out of the bottles and frozen directly in liquid nitrogen. We have found that proteolysis of fusion proteins is minimized if the cells are rapidly lysed into a high-salt buffer that inhibits proteases. Therefore, the frozen pellet was ground to a fine powder under liquid nitrogen with a mortar and pestle (total grinding time, ca. 10 min). The powder was transferred to a beaker that had been prechilled with liquid nitrogen and was then allowed to warm at room temperature until the powder was just beginning to thaw around the edges. An 80-ml portion of room temperature buffer containing 80 mM Tris-HCl (pH 7.4), 0.8 M NaCl, 0.2% Tween 20, 5 mM β-mercaptoethanol (BME), and 1 mM PMSF was added to the powder and rapidly stirred at 4°C until a uniform suspension was obtained. The extract was sonicated for 1 min, allowed to cool on ice, and then sonicated again for 1 min. All of the remaining steps were carried out at 4°C. The extract was centrifuged at 20,000 rpm for 10 min and then at 40,000 rpm for 1 h. The supernatant was diluted by adding 3 volumes of ice-cold H₂O and loaded onto a 5-ml amylose column over a period of 4 to 5 h. The column was washed at a flow rate of 10 column volumes/h with buffer containing 20 mM Tris-HCl (pH 7.4), 200 mM NaCl, and 5 mM BME. The column was eluted with 20 mM Tris-HCl (pH 7.4), 200 mM NaCl, 5 mM BME, and 10 mM maltose, and the presence of the fusion protein was determined by Bradford assay. The peak fractions were pooled and dialyzed extensively into buffer containing 50 mM HEPES (pH 7.6), 0.25 M KCl, and 30% glycerol. The purified protein was frozen on liquid nitrogen and stored at −80°C. More-detailed protocols for the purification of GST and MBP fusions are available on the Kellogg lab website (21a).

Antibodies generated against GST-Swe1 were affinity purified against MBP-Swe1 as previously described (19). The affinity-purified antibodies do not recognize any proteins in extracts from strains carrying a deletion of the SWE1 gene.

Treatment of Swe1 with phosphatase.

Swe1 was immunoprecipitated from wild-type (DK186) cells by using the anti-Swe1 polyclonal antibody and treated with lambda phosphatase as previously described for Gin4 (1).

RESULTS

Identification of mutations that disrupt the mitotic control of bud growth.

In previous work we demonstrated that a simple genetic screen can be used to identify mutations that disrupt the switch from polar to isotropic bud growth (1). This screen relies on the fact that such mutations cause cells to develop highly elongated buds, leading to the formation of colonies that have an unusual rough morphology that can easily be identified. In our original screen, we wanted to focus on the identification of genes encoding proteins that function in pathways used by the Clb2 mitotic cyclin to control bud growth. We therefore carried out the screen in a strain that is dependent solely upon the Clb2 cyclin for the control of mitotic events because it carries deletions of the genes for the other redundant mitotic cyclins (CLB1, CLB3, and CLB4). This screen led to the identification of the Gin4 and Cla4 protein kinases and to members of the septin family of proteins (1, 8, 44).

To identify additional genes required for the mitotic control of bud growth, we have performed a similar screen to identify mutations that cause an elongated bud phenotype in a wild-type background. Such mutations should identify proteins that play a role both in the Clb2-dependent pathway and in redundant pathways used by the other mitotic cyclins to control bud growth. Since mutations that disrupt the mitotic control of bud growth cause the formation of clumpy cells, we were able to enrich for the desired mutants by passing the mutagenized cells through a 5-μm-pore-size mesh screen (15a). Using this method we isolated 53 mutant strains, which we have called ecm mutants (elongated cell morphology). These mutants fell into five complementation groups that include 2 alleles of CLA4 and 22 alleles of a gene that did not correspond to any of the genes that we had previously identified as playing a role in the mitotic control of bud growth. An example of the elongated bud phenotype observed for a representative of the latter complementation group (ecm41) is shown in Fig. 1A. A low-copy-number plasmid library was used to clone the gene corresponding to the ecm41 mutation, and sequence analysis of a plasmid that rescues the ecm41 mutation revealed that it contained the ELM1 gene. Meiotic linkage analysis confirmed that the original mutation was in the ELM1 gene. The ELM1 gene was first identified in a similar screen for mutants that exhibit an elongated bud morphology (5).

FIG. 1.

The phenotype of cells lacking the function of the Elm1 kinase. The cells shown in panels A, B, and C were grown to log phase in liquid YPD medium at 30°C and photographed with Nomarski optics. (A) The phenotype of the ecm41 mutation identified in a screen for mutations that cause an elongated cell morphology. (B) Deletion of the ELM1 gene in a wild-type background results in an elongated bud phenotype. (C) Deletion of the ELM1 gene results in a severe elongated bud phenotype in a Clb2-dependent background. (D) Cells carrying a deletion of the ELM1 gene in a Clb2-dependent background are barely viable. The indicated strains were grown on a YPD plate at 37°C.

Elm1 is required for the proper control of bud growth during mitosis.

To determine whether Elm1 functions in mitotic control pathways, we first generated a strain that carries a deletion of the ELM1 gene. We found that deletion of the ELM1 gene in a wild-type background results in an elongated bud phenotype that is identical to the phenotype caused by the elm1 mutations isolated in our screen (Fig. 1B). We also deleted the ELM1 gene in a Clb2-dependent strain, since we found in previous work that deletion of the genes for proteins that function in Clb2-dependent mitotic control pathways causes a much more severe phenotype in Clb2-dependent cells than in wild-type cells (1, 8, 21, 44). We found that deletion of the ELM1 gene in Clb2-dependent cells results in an elongated bud phenotype that is significantly more severe than the phenotype observed in a wild-type background (Fig. 1C). In addition, Δelm1 Δclb1,3,4 cells grow at a much slower rate than Δelm1 cells at 30°C and are nearly inviable at 37°C (Fig. 1D). Previous work has demonstrated that Clb2-dependent cells have no obvious morphological defects (12, 37).

The severe phenotype observed in the Clb2-dependent background suggests that Elm1 plays an important role in pathways initiated by the mitotic cyclin Clb2. However, the fact that loss of Elm1 function exhibits a pronounced phenotype in a wild-type background suggests that Elm1 may also play an important role in pathways used by the other mitotic cyclins to control bud growth during mitosis.

ELM1 exhibits genetic interactions with the CLA4 and GIN4 genes.

To further establish that Elm1 functions to control bud growth during mitosis, we determined whether ELM1 interacts genetically with the GIN4 or CLA4 kinases. We found that Δelm1 Δcla4 cells grow extremely slowly, while cells carrying either single deletion form colonies at a normal or nearly normal rate (Fig. 2A). In addition, Δelm1 Δcla4 cells are nearly inviable at 37°C and have extremely elongated buds when grown at 30°C (Fig. 2B). Previous work has demonstrated that deletion of the CLA4 gene causes an elongated bud phenotype that is more mild than the phenotype caused by deletion of the ELM1 gene (10a, 44). We also found that Δelm1 Δgin4 cells grow more slowly than cells carrying either single deletion; however, this phenotype is not as severe as the phenotype of the Δelm1 Δcla4 cells (data not shown). These results suggest that Elm1, Cla4, and Gin4 share related functions within the cell.

FIG. 2.

ELM1 interacts genetically with CLA4. (A) Strains carrying deletions of the genes for ELM1 and CLA4 either alone or in combination were grown on a YPD plate at 37°C. (B) Cells carrying deletions of both the ELM1 and CLA4 genes have a severe elongated bud phenotype. Cells were grown to log phase in liquid YPD medium at 30°C and photographed with Nomarski optics.

Elm1 is required for normal progression through mitosis.

Loss of function of Cla4, Gin4, Nap1, or septins in cells that are dependent upon Clb2 causes a prolonged delay early in mitosis at the short spindle stage (1, 8, 10a, 21). To determine whether Δelm1 cells undergo a similar mitotic delay we used α-factor to synchronize Δelm1 cells in G₁, released the cells from the arrest, and then collected samples every 10 min during the cell cycle and determined the fraction of cells with a short mitotic spindle (Fig. 3A). We found that the control cells begin to form short spindles at 60 min, while the Δelm1 cells form short spindles at 70 min, exhibiting a slight delay. The Δelm1 cells then undergo a prolonged delay at the short spindle stage (Fig. 3A). We also measured Clb2 protein levels in both the control cells and in the Δelm1 cells (Fig. 3B). We found that the Clb2 protein in the control cells appears at 70 min, reaches peak levels at 90 min, and then begins to disappear. In Δelm1 cells, the Clb2 protein appears at the same time as in the control cells but then persists at high levels throughout the remainder of the time course, a result consistent with a prolonged mitotic delay.

We carried out similar experiments to determine whether Δelm1 Δclb1,3,4 cells also undergo a mitotic delay. We found, however, that even after treatment of Δelm1 Δclb1,3,4 cells with α-factor for more than 3 h many cells still carry high levels of the Clb2 protein, suggesting that they are arrested in mitosis. We were therefore unable to synchronize Δelm1 Δclb1,3,4 cells in G₁. As an alternative means of assaying the delay in mitosis, we added α-factor to log-phase cultures of Δelm1 Δclb1,3,4 cells and to control cells and then used Western blotting to follow the levels of the Clb2 protein (Fig. 3C). After 90 min of treatment with α-factor, the control cells become synchronously arrested in G₁ with no Clb2 protein, as expected for an α-factor-induced arrest. In comparison, the Clb2 protein levels in the Δelm1 Δclb1,3,4 strain remain essentially constant for 5 h, suggesting that the majority of the Δelm1 Δclb1,3,4 cells are arrested in mitosis. The fact that Δelm1 cells and Δelm1 Δclb1,3,4 cells undergo prolonged mitotic delays provides further evidence that Elm1 plays a role in the control of mitosis.

Elm1 is not required for the formation of active Clb2-Cdc28 kinase complexes.

The previous experiments demonstrate that Elm1 is required for normal progression through mitosis. A possible explanation for these results might be that Elm1 is required for the formation of active Clb2-Cdc28 kinase complexes. To determine whether this is the case, we assayed the activity of Clb2-Cdc28 kinase complexes during the cell cycle in Δelm1 cells and in control cells. We found that Clb2-associated kinase activity rises to nearly normal levels in Δelm1 cells, although the kinase activity reaches peak levels slightly later than in the control cells (Fig. 4). Loss of function of Nap1 or Gin4 also causes a slight delay in the appearance of Clb2-associated kinase activity (1, 21). These results demonstrate that Elm1 is not required for the activation of Clb2-Cdc28 kinase complexes but is required for the timely appearance of kinase activity.

Elm1 is required for the mitosis-specific activation of the Gin4 kinase.

The Gin4 kinase undergoes mitosis-specific activation and plays an important role in the pathway used by Clb2 to control bud growth during mitosis (1). We therefore wanted to determine whether the activation of Gin4 during mitosis is dependent upon Elm1 in vivo. Gin4 is activated during mitosis by hyperphosphorylation, and activation can be readily assayed by Western blotting to detect a shift in the electrophoretic mobility of Gin4 that is caused by hyperphosphorylation (1). We synchronized Δelm1 cells and wild-type cells in G₁ with the mating pheromone α-factor and then released the cells from the cell cycle block and assayed Gin4 hyperphosphorylation as the cells progressed through a single cell cycle. We found that the mitosis-specific hyperphosphorylation of the Gin4 kinase completely fails to occur in Δelm1 cells, indicating that Elm1 is required in vivo for activation of the Gin4 kinase (Fig. 5A). We have also assayed Gin4 kinase activity in Δelm1 cells, which has further demonstrated that Gin4 is not activated in Δelm1 cells (see below).

FIG. 5.

Elm1 is required for the mitosis-specific hyperphosphorylation of the Gin4 and Cla4 kinases. (A) A time course showing the appearance Gin4 hyperphosphorylation during mitosis. Wild-type and Δelm1 cells were released from α-factor arrest and then assayed for Gin4 hyperphosphorylation during the cell cycle as previously described (1). (B) Cells were arrested in interphase with α-factor and induced to express Cdc42^V12 and Clb2^Δ176 in a control strain and in a Δelm1 strain. Samples were taken every hour for 3 hours and were then immunoblotted with affinity-purified anti-Cla4 antibodies to observe Cla4 hyperphosphorylation as previously described (44).

Elm1 is required for the Clb2- and Cdc42-dependent hyperphosphorylation of the Cla4 kinase.

The Cla4 protein is required in vivo for the activation of Gin4 and undergoes hyperphosphorylation that is dependent upon Clb2, Cdc28, Nap1, and the GTP-bound form of Cdc42 (44). The hyperphosphorylation of Cla4 can be induced experimentally by expression of Clb2 and the GTP-bound form of Cdc42 (44). For these experiments, the Gal10 promoter was used to drive expression of a mutant form of Cdc42 that is locked in the GTP-bound state (called Cdc42^V12), and the Gal1 promoter was used to drive expression of a truncated form of the Clb2 cyclin that lacks the destruction box that targets Clb2 for degradation during interphase (called Clb2^Δ176). Before expression of Cdc42^V12 and Clb2^Δ176 was induced, cells were arrested in interphase so that the effects of expression of Clb2^Δ176 could be studied without the complication of other cyclins being present. As with Gin4, the hyperphosphorylation of Cla4 can be assayed by Western blotting to detect an electrophoretic mobility shift. In previous studies, we were able to use this system to demonstrate that Clb2 and the GTP-bound form of Cdc42 act synergistically to induce hyperphosphorylation of both Cla4 and Gin4 (44).

To test whether Elm1 is required for Cla4 hyperphosphorylation, we generated a Δelm1 strain that expresses Cdc42^V12 and Clb2^Δ176 under the control of the Gal1 and Gal10 promoters. We arrested the cells in interphase and then induced the expression of Clb2^Δ176 and Cdc42^V12 by adding galactose. We found that the hyperphosphorylation of Cla4 is significantly reduced in Δelm1 cells, providing further evidence that Elm1 plays a role in the control of mitotic events by Clb2 (Fig. 5B).

Elm1 does not activate the Gin4/Cla4 signaling pathway by inhibiting Swe1 activity.

These experiments demonstrated that the normal regulation of the Cla4 and Gin4 kinases is dependent upon the Elm1 kinase. One possible explanation for these results is that Elm1 functions to inhibit the activity of a negative regulator of the Clb2-Cdc28 kinase complex. For example, although Clb2-Cdc28-associated kinase activity rises to normal levels in Δelm1 cells, perhaps a negative regulator prevents the active kinase complex from signaling the activation of Gin4. If this model were correct, deletion of the gene for the negative regulator that Elm1 is inhibiting should result in a suppression of the Δelm1 phenotype. A known negative regulator of mitotic cyclin-dependent kinase complexes is the Wee1 kinase, which is referred to as Swe1 in budding yeast (6, 35). Furthermore, previous work has demonstrated that the elongated bud phenotype caused by deletion of the ELM1 gene can be suppressed by deletion of the SWE1 gene, suggesting that Elm1 is a negative regulator of Swe1 (11, 27).

To determine whether deletion of the SWE1 gene is able to fully suppress the Δelm1 phenotype, we tested whether the mitosis-specific hyperphosphorylation of the Gin4 kinase is restored in Δelm1 Δswe1 double mutants. For this experiment, we arrested cells in either interphase or mitosis and then used Western blotting to assay Gin4 hyperphosphorylation (Fig. 6). We also immunoprecipitated the Gin4 kinase using an anti-Gin4 polyclonal antibody and assayed its kinase activity in vitro. We found that both the hyperphosphorylation and the activation of the Gin4 kinase fail to occur in Δelm1 Δswe1 cells, just as in Δelm1 cells. As a control, we probed the same samples for the Clb2 protein to demonstrate that the cells were arrested in interphase or mitosis. These results demonstrate that deletion of the SWE1 gene does not fully suppress the Δelm1 phenotype and are inconsistent with the idea that Elm1 functions simply to allow entry into mitosis by negatively regulating Swe1. Note also that the hyperphosphorylation of Cla4 in response to the activity of Clb2 and the GTP-bound form of Cdc42 is defective in Δelm1 cells arrested in interphase (Fig. 5B). Since there is no Swe1 protein present during interphase (29, 42; see also data presented below), these results are inconsistent with the idea that Elm1 exerts its effects on Cla4 by negatively regulating Swe1.

FIG. 6.

Deletion of the SWE1 gene does not restore the hyperphosphorylation and activation of the Gin4 kinase in Δelm1 cells. A wild-type control strain and strains carrying deletions of the SWE1 or ELM1 genes either alone or in combination were grown to log phase in liquid YPD medium and arrested with either α-factor or benomyl. Gin4 hyperphosphorylation was assayed by Western blotting, and Gin4 kinase activity was assayed by an immunoprecipitation kinase assay as previously described (1). The same samples were also probed with an anti-Clb2 antibody to demonstrate that the cells were arrested in mitosis. We always observed that there is a small amount of the Clb2 protein present after treatment with α-factor in Δelm1 and in Δelm1 Δswe1 cells, a finding consistent with the idea that some of the cells in these strains have difficulty exiting mitosis. The nature of the band that migrates slightly more slowly than Clb2 is unknown. This band also appears in the Δelm1 Δswe1 strain upon longer exposure of the blot.

To study the phenotype of Δelm1 Δswe1 cells more closely, we stained the cells with antitubulin antibodies and a DNA stain (Fig. 7A and B). We found that a large proportion of Δelm1 Δswe1 cells grown at 30°C are multinucleate and contain multiple microtubule organizing centers (Fig. 7A and B). In log-phase cells, we observed that 12% (25 of 205) of the Δelm1 Δswe1 cells are unambiguously multinucleate. However, if we synchronize cells by α-factor arrest and release, we find that 52% (104 of 200) of the cells in mitosis are multinucleate. We suspect that it is easier to detect multinucleate cells during mitosis because microtubule-organizing centers and nuclei are pushed apart by microtubule-based mechanisms. Only 4% (8 of 210) of the cells in a Δelm1 log-phase population are clearly multinucleate, although the abnormal cell morphology makes it somewhat difficult to define single cells. We did not observe cells with more than two nuclei in the Δelm1 strain, whereas we frequently observed Δelm1 Δswe1 cells with at least five nuclei. To further confirm the presence of multiple nuclei, we performed flow cytometry to analyze the DNA content of log-phase cultures of wild-type, Δswe1, Δelm1, and Δelm1 Δswe1 cells. We found that a large percentage of Δelm1 Δswe1 cells have a DNA content greater than 2 N, supporting the idea that these cells contain more than one nucleus (Fig. 7C). Note that a large proportion of the Δelm1 cells have a 2 N DNA content, providing further evidence for the existence of a prolonged mitotic delay in these cells. Deletion of the SWE1 gene appears to partially eliminate the mitotic delay in the Δelm1 strain, as indicated by the reappearance of a 1 N peak in the Δelm1 Δswe1 cells. In addition to showing defects in cytokinesis, we found that Δelm1 Δswe1 cells are inviable at 37°C (Fig. 8A).

FIG. 7.

Δelm1 Δswe1 cells are multinucleate. Δelm1 Δswe1 cells were grown to log phase at 30°C and stained with antitubulin antibodies and the DNA stain DAPI (4′,6′-diamidino-2-phenylindole). A high-magnification image (A) and a lower-magnification view that includes more cells (B) are shown. In panel C, the indicated strains were grown to log phase at 30°C, stained with propidium iodide, and analyzed for DNA content by flow cytometry.

FIG. 8.

Δelm1 Δswe1 cells are inviable at 37°C and are morphologically abnormal. (A) Strains carrying deletions of SWE1 and ELM1 either alone or in combination were grown on YPD plates at 30 and 37°C. (B) Morphology of Δelm1 Δswe1 cells. Cells were grown to log phase at 30°C in YPD liquid medium and photographed with Nomarski optics.

These results suggest that Δelm1 cells are defective in executing mitotic events that are required for cytokinesis and that Swe1 functions in a checkpoint that detects these defects and delays nuclear division until cytokinesis can be carried out properly. Previous work has suggested the existence of a Swe1-dependent checkpoint that delays nuclear division in response to perturbations in the organization of actin and/or septin filaments (4, 23, 29). We did find that deletion of the SWE1 gene causes a partial suppression of the elongated bud phenotype of Δelm1 cells, as previously described (Fig. 8B) (11). This partial suppression also occurs when the SWE1 gene is deleted in Δcla4 cells (data not shown). Although the elongated bud phenotype of the double deletion is less severe than the single Δelm1 deletion, there are still numerous cells with elongated buds and aberrant morphologies in the Δelm1 Δswe1 strain, indicating that deletion of the SWE1 gene does not entirely suppress the Δelm1 morphological phenotype.

Septin organization is abnormal in Δelm1 cells and in Δelm1 Δswe1 cells.

Previous work has demonstrated that the Gin4 kinase is required for the normal localization of the septins and that the septins are required for the activation and localization of the Gin4 kinase (8, 26). In addition, Elm1 is localized to the bud neck and the Δelm1 phenotype strongly resembles the phenotype caused by loss of septin function (8, 13, 14, 31). We were therefore interested in determining whether septin organization is normal in Δelm1 and Δelm1 Δswe1 cells. We used a polyclonal antibody that recognizes Cdc11 to localize the septins in log-phase populations of wild-type cells, Δelm1 cells, Δswe1 cells, and Δelm1 Δswe1 cells. In wild-type cells, Cdc11 is found at the tip of the emerging bud and as a tight double ring at the bud neck (Fig. 9). In the Δelm1 strain, many cells no longer have tight double rings at the bud neck and there are often additional diffuse septin rings along the length of the elongated buds (Fig. 9). We observed the same defects in Δelm1 Δswe1 cells, although at a lower frequency. In both Δelm1 and Δelm1 Δswe1 strains, we observed that some cells have apparently normal septin localization at the bud neck. Cells carrying a deletion of the GIN4 gene also show weak and diffuse septin rings at the bud neck, as well as some cells with normal septin localization (26).

FIG. 9.

The septins are mislocalized in Δelm1 and Δelm1 Δswe1 cells. Wild-type, Δelm1, and Δelm1 Δswe1 cells were grown to log phase at 30°C in YPD liquid medium and stained with a DNA stain and a polyclonal antibody that recognizes Cdc11.

Swe1 hyperphosphorylation is dependent upon Elm1.

To further investigate the relationship between Swe1 and the pathway that includes Elm1, Cla4, and Gin4, we used an anti-Swe1 polyclonal antibody to follow the behavior of the Swe1 protein during the cell cycle by Western blotting (Fig. 10A). In addition, we monitored the behavior of the Clb2 protein in the same samples as a marker for when the cells enter mitosis (Fig. 10A). We found that the Swe1 protein is absent during interphase and then appears slightly before Clb2 and immediately undergoes a dramatic electrophoretic mobility shift that is suggestive of hyperphosphorylation. Mitosis-specific hyperphosphorylation of Wee1 has also been observed in Xenopus embryo extracts (32, 43). In order to ensure that the Swe1 mobility shift is due to hyperphosphorylation, we immunoprecipitated Swe1 from mitotic cells and then treated the protein with phosphatase. We found that the electrophoretic mobility shift collapses in samples treated with phosphatase, indicating that it is indeed due to hyperphosphorylation (Fig. 10B). Note that hyperphosphorylation of Swe1 is correlated with the appearance of the Clb2 protein and peaks when Clb2 protein levels peak, suggesting a causative relationship.

FIG. 10.

Elm1 is required for the mitosis-specific hyperphosphorylation of the Swe1 kinase. (A) A time course shows the appearance of Swe1 hyperphosphorylation during mitosis. Wild-type and Δelm1 cells were released from α-factor arrest and then assayed for Swe1 hyperphosphorylation during the cell cycle by using affinity-purified anti-Swe1 antibodies. In this experiment, less total protein was loaded for the Δelm1 samples because the starting culture had fewer cells (due to the aberrant morphology of Δelm1 cells, it is difficult to quantitate the total number of cells in culture). We have not found evidence for decreased levels of the Swe1 protein in Δelm1 cells. (B) Western blot of phosphatase-treated and untreated samples of Swe1 immunoprecipitates probed with Swe1 polyclonal antibody.

We next examined the behavior of the Swe1 protein in Δelm1 cells. We found that Swe1 hyperphosphorylation is significantly decreased in Δelm1 cells, and it appears as though a specific subset of Swe1 hyperphosphorylations is lost (Fig. 10A). We observed an identical loss of Swe1 hyperphosphorylation in Δcla4 cells, in Δgin4 Δclb1,3,4 cells, in Δnap1 Δclb1,3,4 cells, and in a temperature-sensitive cdc11 strain at the restrictive temperature (data not shown).

DISCUSSION

The Clb2-Cdc28 cyclin-dependent kinase complex functions in budding yeast to control the events of mitosis, including a switch in the pattern of bud growth that occurs as cells enter mitosis. Our previous work has suggested that the Clb2-Cdc28 kinase complex controls this switch through a signaling network that includes a Clb2-binding protein called Nap1, as well as the Cdc42 GTPase and two protein kinases called Cla4 and Gin4 (1, 21, 44). In the present study, we demonstrate that the Elm1 kinase also plays a role in this signaling network. Furthermore, our data suggests that the Elm1 kinase may also play a role in cytokinesis and that the Swe1 kinase may function to delay the cell cycle in response to defects caused by loss of Elm1 function. These results provide further evidence for the existence of intricate signaling networks that function to control mitotic events.

The Elm1 kinase functions in a mitotic signaling network that controls bud growth during mitosis.

The Elm1 kinase was originally identified in a screen for mutations that cause an elongated cell morphology, which led to the suggestion that Elm1 is a negative regulator of pseudohyphal growth (5). A number of the experiments that we have carried out suggest that Elm1 plays a critical role in signaling pathways that control mitotic events. First, we identified the ELM1 gene in a screen for mutations that cause defects in the control of polar bud growth during mitosis, and we found that the elongated bud phenotype caused by loss of Elm1 function is similar to the phenotype caused by mutations that inactivate other proteins known to function in the mitotic control of polar bud growth. Second, the phenotype caused by deletion of the ELM1 gene is considerably more severe in cells that are dependent upon Clb2 for survival than it is in wild-type cells, suggesting that Elm1 plays a critical role in the control of mitotic events by Clb2. Third, Elm1 interacts genetically with both Cla4 and Gin4. Fourth, Δelm1 cells undergo a prolonged mitotic delay with short spindles and active Clb2-Cdc28 kinase complexes, similar to the delay caused by deletion of the genes for other proteins that function in the pathway that controls polar bud growth during mitosis. Finally, Elm1 is required in vivo for the proper regulation of both Gin4 and Cla4. Taken together, these results demonstrate that Elm1 plays an important role in mitotic control pathways.

The finding that some Δelm1 cells and many Δelm1 Δswe1 cells are multinucleate suggests that Elm1 may also play an important role in cytokinesis, although it is possible that defects in cytokinesis may be a secondary defect caused by hyperpolarization of actin to the tip of the highly elongated buds observed in cells that fail to make the switch from polar to isotropic bud growth (21).

Elm1 is required for the mitosis-specific hyperphosphorylation of the Gin4 and Cla4 kinases.

In previous work, we found that the Gin4 and Cla4 kinases undergo mitosis-specific hyperphosphorylations that are dependent upon the function of Clb2, Cdc28, Nap1, and the GTP-bound form of Cdc42 (1, 44). The hyperphosphorylation of Gin4 requires Cla4 and appears to be signaled by the hyperphosphorylated form of Cla4. Importantly, it has been shown that Clb2 and the GTP-bound form of Cdc42 can act synergistically to induce the hyperphosphorylation of Cla4 and Gin4 in cells arrested in interphase. Furthermore, the hyperphosphorylation of Cla4 and Gin4 in this context is dependent upon Cdc28 (44). These results argue that hyperphosphorylation of Cla4 and Gin4 is induced by the activity of the Clb2-Cdc28 kinase complex and that hyperphosphorylation can be induced in a manner that is largely independent of the events of mitosis.

In this study, we have found that the Elm1 kinase is required in vivo for the complete hyperphosphorylation of both Cla4 and Gin4. One possible model for these results is that Elm1 functions upstream of the Cla4 and Gin4 kinases in a mitotic signaling pathway initiated by the Clb2-Cdc28 kinase complex. A second possible model is that Elm1 functions in a parallel signaling pathway that is required for the proper regulation of Gin4 and Cla4. A third model is that Elm1 is required to inactivate an inhibitor of the pathway that leads to activation of Cla4 and Gin4. A number of experiments demonstrate that Elm1 does not function simply to inhibit the activity of Swe1, a known negative regulator of mitosis (discussed below). However, our experiments do not rule out the possibility that Elm1 allows activation of Cla4 and Gin4 through the inactivation of an unknown inhibitor.

Elm1 is required for proper progression through mitosis.

Clb2-associated kinase activity rises to normal levels in Δelm1 cells, indicating that Elm1 is not required for formation of active Clb2-Cdc28 kinase complexes. However, there is a slight delay in the appearance of Clb2-associated kinase activity and the cells undergo a prolonged delay in mitosis at the short spindle stage even though Clb2-associated kinase activity is present. A very severe mitotic delay is observed when the ELM1 gene is deleted in cells that are dependent upon the Clb2 cyclin for control of mitotic events. In this case, the cells are barely viable, and it appears that the majority of the cells never exit mitosis. Since the Clb2 protein and Clb2-associated kinase activity rise to normal levels in the Δelm1 mutants, the failure to activate Gin4 does not appear to be due simply to a failure to enter mitosis, although it is possible that the Clb2-Cdc28 kinase complex in Δelm1 cells has a lower specific activity and therefore has difficulty promoting mitotic events.

The primary reason for the mitotic delay observed in Δelm1 cells is unclear. One possibility is that Elm1 plays a critical role in pathways used by Clb2 to control many different mitotic events and that without Elm1 the cells are simply unable to execute these events. Another possibility is that Elm1 functions in a pathway that is required only for one specific event and that the failure of this event activates a checkpoint that delays the other events of mitosis. We found that the septins are not properly localized in Δelm1 cells, which may lead to activation of a checkpoint that is the primary cause of the mitotic delay observed in Δelm1 cells. At this point, however, it is difficult to rule out the possibility that Elm1 functions to control other mitotic events, such as elongation of the mitotic spindle. For example, although we observe that spindle elongation eventually occurs in Δelm1 cells after a prolonged delay, this may be due to the existence of a partially redundant pathway that can compensate for a lack of Elm1 function in this event. It is also possible that Elm1 plays a role in pathways that regulate mitotic cyclin destruction, which could contribute to the mitotic delays observed in Δelm1 cells.

Δelm1 Δswe1 cells have severe defects in cell cycle control.

The Wee1 kinase was first identified genetically in fission yeast as a negative regulator of entry into mitosis, and evidence from a variety of systems has suggested that Wee1 phosphorylates mitotic cyclin-dependent kinases on a tyrosine residue near the N terminus of the protein, resulting in the inactivation of cyclin-dependent kinase activity (35, 39). Recent work in budding yeast cells has demonstrated that Swe1, the budding yeast homolog of Wee1, functions as part of a checkpoint that delays entry into mitosis in response to defects in the organization of actin and/or septin filaments (4, 23, 29). Work in budding yeast cells has also demonstrated that loss of function of the Hsl1 kinase causes a mitotic delay and an elongated bud phenotype that can be suppressed by deletion of the SWE1 gene (4, 27). Similarly, it has been reported that the elongated bud phenotype of Δelm1 cells can be suppressed by deletion of the SWE1 gene (11). These results suggest that Elm1 and Hsl1 are negative regulators of Swe1 and that at least some of the defects observed in Δelm1 cells and Δhsl1 cells could be due to failure to inactivate Swe1.

To gain a better understanding of the functional relationship between Elm1 and Swe1, we looked closely at the phenotype of Δelm1 Δswe1 cells. Consistent with previous reports, we found that deletion of the SWE1 gene partially suppresses the elongated bud phenotype of Δelm1 cells (11). However, we found that deletion of the SWE1 gene does not suppress the defects that we observed in the regulation of the Gin4 kinase in Δelm1 cells. Specifically, we observed that the mitosis-specific hyperphosphorylation and activation of Gin4 fail to occur in Δelm1 Δswe1 cells, just as in Δelm1 cells. We also found that hyperphosphorylation of Cla4 in response to Clb2 and the GTP-bound form of Cdc42 is defective in interphase cells, which lack Swe1. Finally, we observed that Δelm1 Δswe1 cells are temperature sensitive for growth and that a large proportion of Δelm1 Δswe1 cells grown at a permissive temperature contain multiple nuclei, suggesting severe defects in cytokinesis. Taken together, these results demonstrate that the Δelm1 phenotype is not rescued by loss of Swe1. In addition, these results are inconsistent with a model in which Elm1 exerts its effects on Gin4 and Cla4 simply by negatively regulating Swe1.

Elm1 is required for proper localization of the septins.

Previous work has shown that the Gin4 kinase is required for the normal localization of the septins, and that the septins are required for the localization and activation of Gin4 (8, 26). It has also been shown that the septins are mislocalized in cla4 mutant strains (10a). In this study we have demonstrated that the Elm1 kinase is also required for the proper localization of the septins. The septin rings at the bud neck in Δelm1 and Δelm1 Δswe1 strains are often more diffuse, and the septins frequently form additional ring-like structures along the length of the elongated buds observed in these cells. These results, when combined with previous results, demonstrate that at least three different kinases are required for normal localization of the septins.

The Swe1 kinase undergoes complex regulation during the cell cycle that is dependent upon the function of Elm1.

Using a polyclonal antibody that recognizes Swe1, we found that Swe1 undergoes complex regulation during the cell cycle. Specifically, the Swe1 protein is absent during interphase and then begins to appear slightly before the Clb2 mitotic cyclin as cells enter mitosis. When Swe1 first appears it migrates as a single band on polyacrylamide gels, and then it rapidly shifts to a series of higher-molecular-weight bands as Swe1 undergoes dramatic hyperphosphorylation during mitosis. These results are consistent with previous studies showing that the Xenopus and S. pombe Wee1 homologs undergo dramatic hyperphosphorylation during mitosis in Xenopus embryo extracts and that human Wee1 protein levels peak during mitosis and decline sometime during late mitosis or early G1 (28, 32, 43, 45). These results are also consistent with previous experiments in budding yeast cells demonstrating that Swe1 protein levels fluctuate during the cell cycle (23, 42). However, our results differ somewhat from previous work in that we find that the Swe1 protein undergoes dramatic hyperphosphorylation and that protein levels peak during mitosis (as judged by levels of the Clb2 mitotic cyclin) and then decline as cells exit mitosis. Previous studies on Swe1 did not detect a dramatic hyperphosphorylation of Swe1 and found that protein levels peak during S/G₂ and then decline prior to and during nuclear division (as judged by bud emergence and nuclear division) (23, 42). It is likely that the apparent differences in Swe1 behavior are due to differences in the methods used to detect Swe1 and to assess cell cycle progression.

Interestingly, we found that a subset of Swe1 hyperphosphorylations appears to be lost in Δelm1 cells, in Δcla4 cells, in Δgin4 Δclb1,3,4 cells, in Δnap1 Δclb1,3,4 cells, and in a temperature-sensitive cdc11 strain at the restrictive temperature (Fig. 8 and reference 42a). Studies carried out in Xenopus embryo extracts have demonstrated that Wee1 is phosphorylated on both the N terminus and the C terminus and that at least two distinct kinases are responsible for hyperphosphorylation of Wee1 (10, 43). It seems possible, therefore, that inactivation of Elm1 results in a loss of the hyperphosphorylations occurring at either the N terminus or the C terminus of Swe1. Studies carried out in Xenopus embryo extracts have also demonstrated that full hyperphosphorylation of Wee1 results in a 7- to 20-fold reduction in its ability to phosphorylate cyclin-dependent kinases in vitro (32, 43). The failure to fully hyperphosphorylate Swe1 in Δelm1 cells may therefore result in persistence of Swe1 kinase activity during mitosis.

These results suggest that Swe1 may be a direct target of mitotic signaling pathways that require the function of Elm1, Cla4, Gin4, and the septins. Alternatively, inactivation of Elm1, Cla4, Gin4, or the septins pathway may cause defects in specific mitotic events that lead to activation of a checkpoint that delays passage through mitosis by preventing the full hyperphosphorylation of Swe1. These possibilities are not mutually exclusive. For example, it is possible that the Elm1, Cla4, and Gin4 kinases function to control mitotic events and to maintain Swe1 in a hyperphosphorylated state as long as these events are occurring properly. It is interesting to note that the hyperphosphorylation of Swe1 is correlated with the appearance of the Clb2 protein, perhaps suggesting that Swe1 is hyperphosphorylated in response to activation of an Elm1/Cla4/Gin4 pathway by the Clb2-Cdc28 kinase complex. Recent coimmunoprecipitation experiments have demonstrated an interaction between the S. pombe homologs of Swe1 and Gin4, suggesting that the Gin4 kinase may play a direct role in the hyperphosphorylation of Swe1 (18).

Roles of the Elm1 and Swe1 kinases in the control of mitotic events.

To explain the results that we have obtained, we propose that the Elm1 kinase functions in a mitosis-specific signaling network that controls the reorganization of actin filaments that is necessary both for the switch from polar to isotropic bud growth and for cytokinesis. In the absence of Elm1, these events fail to occur properly and a Swe1-dependent checkpoint detects the defects and induces a delay in nuclear division. Thus, in Δelm1 Δswe1 cells nuclear division continues in the absence of proper cytokinesis, leading to the formation of multinucleate cells.

We have attempted to detect elimination of the mitotic delay in Δelm1 Δswe1 cells by monitoring Clb2 protein levels in cells synchronized with α-factor. However, we found that significant levels of the Clb2 protein are still present in Δelm1 Δswe1 cells after prolonged treatment with α-factor, indicating that many cells have difficulty passing through mitosis. Furthermore, we found that Clb2 levels remain high in Δelm1 Δswe1 cells for a prolonged period after cells enter mitosis, a finding inconsistent with elimination of the mitotic delay. We are concerned, however, that the presence of multiple nuclei in Δelm1 Δswe1 cells may lead to the activation of other checkpoints, such as the spindle assembly checkpoint, that delay the cell cycle in an Swe1-independent manner. The generation of temperature-sensitive elm1 mutants will be required to conclusively determine whether Swe1 functions to induce a mitotic delay in response to inactivation of Elm1.

Our results demonstrate that the normal regulation of Cla4 and Gin4 is dependent upon Elm1. To determine whether Elm1 regulation is dependent upon Cdc28, Cla4, or Gin4, we generated an antibody against Elm1 and used it to assay the kinase activity of Elm1 in immunoprecipitates during the cell cycle. However, we were unable to detect any significant changes in the in vitro kinase activity of Elm1 during the cell cycle (42a). We were also unable to detect any cell cycle-dependent changes in the electrophoretic mobility of Elm1 that might suggest posttranslational modifications. These results do not rule out the possibility that Elm1 kinase activity is highly regulated in vivo.

The results that we have obtained for Elm1 and Swe1 are similar to the results of previous studies on the fission yeast homologs of Swe1 and Gin4, called Wee1 and Cdr2 (7). In those experiments, it was found that deletion of the gene for Cdr2 causes a G2 delay, as well as defects in cell morphology and cytokinesis. Inactivation of the Wee1 gene in cdr2⁻ cells eliminates the mitotic delay but does not eliminate the defects in cytokinesis and cell morphology.

Our model does not explain why deletion of the SWE1 gene causes partial suppression of the elongated bud phenotype of Δelm1 cells. One possibility is that Swe1 has targets other than Cdc28, and the importance of Swe1 in regulating these targets is only seen in Δelm1 cells.

Control of mitotic events by intricate signaling pathways.

The finding that the Elm1 kinase works with Cla4 and Gin4 to control specific mitotic events provides further evidence for an intricate signaling network that functions during mitosis. A large number of highly conserved kinases have now been demonstrated to function during mitosis and cytokinesis in budding yeast, including Cdc28, Swe1, Elm1, Cla4, Ste20, Gin4, Hsl1, Kcc4, Cdc5, Ipl1, Cdc15, Mps1, Bub1, and Dbf2 (1, 4, 6, 8a, 9, 10a, 15–17, 27, 35a, 38, 40, 46). In addition, several highly conserved GTPases have been demonstrated to function in the control of mitotic events and cytokinesis in budding yeast, including Cdc42, Tem1, and Ras (30, 41, 44). It seems likely, therefore, that many mitotic events will be controlled by intricate and highly conserved signaling networks. Elucidation of these networks represents an important challenge for our future understanding of the cell cycle.