The GTPase-Activating Protein Rga1 Interacts with Rho3 GTPase and May Regulate Its Function in Polarized Growth in Budding Yeast

In budding yeast, Rga1 negatively regulates the Rho GTPase Cdc42 by acting as a GTPase-activating protein (GAP) for Cdc42. To gain insight into the function and regulation of Rga1, we overexpressed Rga1 and an N-terminally truncated Rga1-C538 (a.a. 538-1007) segment. Overexpression of Rga1-C538 but not full-length Rga1 severely impaired growth and cell morphology in wild-type cells. We show that Rga1 is phosphorylated during the cell cycle. The lack of phenotype for full-length Rga1 upon overexpression may result from a negative regulation by G1-specific Pho85, a cyclin-dependent kinase (CDK). From a high-copy suppressor screen, we isolated RHO3, SEC9, SEC1, SSO1, SSO2, and SRO7, genes involved in exocytosis, as suppressors of the growth defect caused by Rga1-C538 overexpression. Moreover, we detected that Rga1 interacts with Rho3 in two-hybrid and bimolecular fluorescence complementation (BiFC) assays. Rga1 preferentially interacts with the GTP-bound form of Rho3 and the interaction requires the GAP domain and additional sequence upstream of the GAP domain. Our data suggest that the interaction of Rga1 with Rho3 may regulate Rho3’s function in polarized bud growth.


Introduction
Rho-family GTPases in eukaryotic cells regulate cytoskeletal rearrangement, vesicular trafficking, and cell cycle progression [1,2]. They act as molecular switches, cycling between the GTP-and GDP-bound states. Only in the GTP-bound form can they bind to downstream effectors to transduce signals. The cycling of Rho GTPases requires the guanine nucleotide exchange factors (GEFs) and GTPase-activating proteins (GAPs). GEFs catalyze the exchange of GDP for GTP of Rho proteins, leading to their activation. GAPs promote the intrinsic GTP-hydrolyzing activity of Rho proteins, thereby leading to their rapid conversion to the inactive GDP-bound state [3].
There are six Rho GTPases in the budding yeast Saccharomyces cerevisiae: Cdc42 and Rho1 to Rho5 [4]. Cdc42, Rho1, Rho3, and Rho4 are important for polarized organization of the cytoskeleton. This function requires their interaction with the formins, Bni1 or Bnr1 [5]. Cdc42 is critical for bud emergence because temperature-sensitive cdc42 mutants became large and round without a bud at restrictive temperature [6]. Rho3 and Rho4 are important for the maintenance of bud growth after bud emergence since rho3Δ rho4Δ cells often lysed at the smallbudded stage with a large and round morphology [7]. Besides their roles in actin organization, Cdc42 and Rho3 also have a direct role in exocytosis (polarized secretion). cdc42-6 and rho3-V51 mutants accumulated post-Golgi secretory vesicles while the actin cytoskeleton was still polarized [8,9]. The function of Rho3 in exocytosis is thought to be mediated by two downstream effectors, Myo2 (myosin V, involved in the transport of vesicles) and Exo70 (exocyst component, implicated in the docking of vesicles on the plasma membrane) [10,11], while the role of Cdc42 in exocytosis is thought to be mediated by its effectors, Sec3 and Exo70, two exocyst components [11,12]. It is interesting to note that the functions of Cdc42 and Rho3 overlap to some extent. High-copy RHO3 rescued the growth defect of cdc42-6 mutant whereas highcopy CDC42 and BEM1 rescued the lethality of rho3Δ rho4Δ cells [7,9].
The physiological role of Rga1 appears to down-regulate Cdc42 activity since high-copy RGA1 impaired the growth of temperature-sensitive cdc42-1 mutant at 35°C while rga1Δ restored growth to the temperature-sensitive cdc24-H mutant (Cdc24 is a Cdc42GEF) [14]. Rga1 shares a redundant function with Rga2 and Bem3 in Cdc42-controlled conversion of a septin "cloud" into a septin ring at the presumptive bud site [20]. However, Rga1 appears to have certain cellular function not redundant with Rga2 and Bem3. We previously reported that Rga1 plays a critical role in the elimination of active Cdc42-GTP at the bud neck after cytokinesis. This function is crucial for proper bud-site selection at the beginning of the next cell cycle, as rga1Δ cells, unlike rga2Δ or bem3Δ cells, budded within the old budding site [21]. Rga1 also appears to act much more potently than Rga2 and Bem3 in the negative regulation of the pheromone response pathway [13]. Apart from these, whether or not Rga1 may have other unique cellular functions is not clear.
Here, we show that a high dose of N-terminally truncated Rga1 impairs polarized bud growth and this defect could be rescued by high-copy RHO3 and genes encoding components of the exocytic apparatus but not by high-copy CDC42. In addition, Rga1 interacts with Rho3 in vivo. These results suggest that Rga1 may normally engage in the regulation of Rho3 in polarized growth. We also show that Rga1 is phosphorylated in the cells and the phosphorylation event negatively regulates its function.
Cdc42-regulated functions. Here, we tried to examine Rga1's cellular function by overexpressing Rga1. In theory, overexpression of Rga1 would reduce the activity of Cdc42, generating a cdc42 mutant-like loss-of-polarity phenotype. Surprisingly, we found that overexpression of full-length Rga1 in wild-type cells under the control of a galactose-inducible promoter did not dramatically impair growth (Fig 1A), or cell morphology ( Fig 1B). We then overexpressed an N-terminally truncated Rga1-C538 (a.a. 538-1007) segment ( Fig 1A), a presumably dominantactive mutant based on its ability to suppress the growth and morphological defects of temperature-sensitive bem2 mutants [22]. Interestingly, overexpression of Rga1-C538 was toxic to the cells. No colonies were formed on SRG-Ura agar medium within 4 days at 30°C (Fig 1A). Cell morphology was also affected.
Rga1-C538 overexpression caused a loss-of-polarity phenotype. Microscopic examination of cells induced for Rga1-C538 overexpression for 16 hours at 30°C revealed that 63% of the cells examined (n = 238) became large and round, indicative of a loss of cell polarity. Among the large and round cells, 58% (n = 263) were unbudded. The remaining 42% were budded, but usually had a small or tiny bud (Fig 1B). Some cells even carried two or more buds (mostly two) on the same mother cell (Fig 1B, see arrows). All the budded cells examined (n = 72) displayed a depolarized actin cytoskeleton as actin patches were no longer enriched in the buds. Actin cables were barely visible (Fig 1C, bottom panel). These findings suggest that Rga1-C538 overexpression caused a defect in polarized bud growth. Staining of nuclear DNA revealed that 40% of large and round cells (n = 198) contained two or more nuclei in the mother cell (  bottom panel), which was not seen in wild-type control cells, nor in cells overexpressing fulllength Rga1. In the multinucleated population of large and round cells, 36% (n = 224) were unbudded. The remaining 64% carried at least one bud. This observation indicates that nuclear segregation was defective.
The growth defect caused by Rga1-C538 overproduction could be effectively suppressed by the addition of 1 M sorbitol, an osmolarity-stabilizing agent, to the culture medium ( Fig 1D). The large and round cell morphology was also partially suppressed, though most cells examined (59%; n = 276) were still markedly bigger than control cells (Fig 1E). This observation suggests that cell wall may be defective in cells overexpressing Rga1-C538. In support of this view, we observed that after 12 hours' induction for Rga1-C538 overexpression, 30% of cells were dead as indicated by staining with methylene blue, a vital staining dye that selectively stains dead cells blue, but not live cells [23] (Fig 1E). Interestingly, among the dead cells, nearly all (92%; n = 304) had a normal morphology and 73% (n = 304) were small-budded. For comparison, only 16% of live cells (n = 258) were small-budded. Moreover, after the addition of sorbitol to the medium, the death of normal looking cells was almost completely suppressed (3% dead cells; n = 245) ( Fig 1E). These results suggest that Rga1-C538 overexpression may affect cell wall integrity and cause cell lysis at the small-budded stage.
The defect in polarized bud growth caused by Rga1-C538 overexpression seems to depend on an interaction with Cdc42 because overexpression of the Rga1-C538 K872A mutant neither impaired growth (Fig 1A), nor enlarged the cells (Fig 1B). It has been shown that the K872A mutation in Rga1 not only eliminated its GAP activity but also its interaction with Cdc42 [15]. This result implies that Cdc42 may be the direct target of Rga1-C538 action. An increase of Rga1 activity may down-regulate Cdc42 activity, causing a loss of cell polarity. The lack of phenotype for full-length Rga1 upon overexpression suggests that the GAP activity of Rga1 may normally be regulated (see next section).
Taken together, our results suggest that an excess of Rga1 GAP activity affects polarized bud growth, causing depolarized growth and cell lysis at the small-budded stage.

Rga1 may be negatively regulated by phosphorylation
The observation that overexpression of the N-terminally truncated Rga1-C538 segment impaired growth whereas overexpression of full-length Rga1 did not implies that the GAP activity of Rga1 may normally be negatively regulated in the cells. The truncated Rga1 segment may lack this regulation and became hyperactive.
HA-tagged HA-Rga1 immunoprecipitated from yeast cell lysate showed a small mobility shift in SDS-PAGE ( (Fig 2A), indicating that Rga1 is phosphorylated in the cells, in agreement with previous reports [24,25]. To monitor the phosphorylation of Rga1 during cell cycle progression, we synchronized the cells at late G1 by α factor. The cells were then released into fresh medium and samples were collected every 12 min for Western blotting. We observed that Rga1's protein level varied dramatically during the cell cycle. The amount of protein was low at G1 but high at M phase ( Fig 2B). Rga1 phosphorylation was prominent at G2 and M phases ( Fig 2B).
Rga1 is thought to be phosphorylated by Cdc28-Cln2, a G1-specific CDK (cyclin-dependent kinase) [24] and Cdc28-Clb2, a mitotic CDK [25]. Rga2, the paralogue of Rga1, is known to be negatively regulated by Cdc28 and Pho85, another G1-specific CDK, by phosphorylation [26,27,28]. Because Rga1 shares extensive sequence homology over their entire lengths with Rga2, we asked if Rga1 might also be a target of Pho85 CDK. To test this, we overexpressed full-length Rga1 in pcl1Δ pcl2Δ cells. Pcl1 and Pcl2 are cyclins that specifically bind to and activate Pho85. We found that overexpression of full-length Rga1 severely inhibited the growth of pcl1Δ pcl2Δ cells, but not wild-type cells ( Fig 2C). pcl1Δ pcl2Δ cells became large and round, similar to the effect of Rga1-C538 overexpression ( Fig 2D).
There are 15 consensus CDK phosphorylation sites (S/T-P) in Rga1. Among them, only one site locates within the RhoGAP domain. The remaining 14 sites locate in the middle region between the two LIM domains and the RhoGAP domain. Because 10 out of 15 CDK consensus sites locate in the region a.a. 1-537, which is missing in the dominant-active Rga1-C538 segment, the N-terminal half of the protein may contain the key phosphorylation sites. Interestingly, Thr 278 locates in the only full consensus CDK phosphorylation site (S/T-P-X-K/R) in Rga1. It could be a major CDK phosphorylation site. However, we found that Rga1 T278A mutant that mimics the unphosphorylated form did not impair growth in wild-type cells upon overexpression (data not shown). It is likely that the phosphorylation of multiple sites may be required for the negative regulation, like the case of Rga2, which involves in the phosphosphorylation of at least 8 sites [28].
Together, our results indicate that Rga1 is a phosphoprotein. Rga1 may be an important substrate of the G1-specific Pho85 CDK and the phosphorylation of Rga1 by CDK may negatively regulate Rga1 activity to ensure proper Cdc42 activation during bud emergence.

An excess of Rga1 activity may impair exocytosis
To gain insight into how Rga1-C overexpression affected cell growth, we performed a highcopy suppressor screen looking for genes that, when mildly overexpressed, could rescue the lethality caused by Rga1-C538 overproduction (see details of the screen in Materials and Methods). RHO3, SEC9, SSO1, SSO2, SRO7, and SEC1 genes were isolated from the screen. Among them, RHO3, SEC9, and SEC1 showed the strongest suppression for the growth defect ( Fig 3A). It is interesting to note that all the protein products encoded by these genes are implicated in exocytosis. Rho3 controls the transport and fusion of secretory vesicles into the growing bud. Sec9, Sso1, and Sso2 are plasma membrane t-SNAREs that participate in the fusion of secretory vesicles with the plasma membrane [29][30][31]. Sro7 interacts with both Sec9 and Exo84, a component of the exocyst, and functions in the docking and fusion of post-Golgi vesicles with the plasma membrane [32,33]. Sec1 binds to the assembled SNARE complex and is involved in the activation of membrane fusion [34,35]. The observation that these genes could suppress the lethality caused by Rga1-C538 overexpression suggests that excess Rga1-C538 may impair exocytosis.
High-copy SEC10, which encodes a component of exocyst [36], did not suppress the lethality. The lethality was not rescued by high-copy SEC4 or SNC2, either (Fig 3A). Sec4 is a Rab GTPase that plays a crucial role in vesicular transport [37,38]. Snc2 is the v-SNARE located on secretory vesicles [39]. Interestingly, high-copy EXO70 weakly suppressed the lethality caused by Rga1-C538 overproduction (Fig 3A). Exo70 is another component of exocyst and is a downstream effector of Rho3 [10,11]. Because a subset of exocytic genes implicated in the vesicle fusion step suppressed the lethality caused by Rga1-C538 overproduction, Rga1-C538 overexpression may affect the vesicle fusion step of exocytosis.
To test whether vesicular transport may also be affected, we overexpressed full-length Rga1 in temperature-sensitive myo2-66 and several post-Golgi secretory mutants. We found that Rga1 overexpression reduced the growth of myo2-66 (at 24°C) and sec4-8 (at 30°C) mutants, but not the growth of sec3-2, sec6-4, or sec9-4 mutants (Fig 3B). Because Myo2 (myosin V) and Sec4 are essential for vesicular transport, this result suggests that an excess of Rga1 activity may also affect vesicular transport.
Together, our results suggest that excess Rga1 activity may impair polarized growth by affecting exocytosis.
Rho GTPases other than Rho3 did not rescue the lethality caused by Rga1-C538 overexpression Cdc42 is a known in vivo target of Rga1. Rga1-C538 overexpression would in theory inactivate Cdc42, causing a growth defect. Surprisingly, we isolated RHO3 but not CDC42 from the highcopy suppressor screen. An assay using high-copy CDC42 showed that a high dose of Cdc42 indeed did not suppress the growth defect caused by Rga1-C538 overexpression (Fig 4A). The morphology of cells was not suppressed, either (Fig 4B). In contrast, high-copy RHO3 efficiently suppressed the growth defect ( Fig 4A). The morphology defect was also partially suppressed by RHO3 as the percentage of large and round cells was 31% (n = 233) after 20 hours' induction for Rga1-C538 overexpression, lower than that of control cells (66%; n = 229). Dead cells (indicated by methylene blue staining) was just 8% (n = 278), significantly lower than that of control cells (27%; n = 215).
Among the six Rho GTPases, Rho4 is the closest to Rho3 with respect to cellular functions although they do not share a high degree of amino acid sequence similarity [40]. We found that high-copy RHO4 did not suppress the growth and cell morphology defect (Fig 4A and 4B). High-copy RHO1 or RHO2 also failed to suppress the defects (Fig 4A and 4B). The specific suppression by Rho3 but not by other Rho GTPases suggests that Rho3 might be another target of Rga1 in vivo. Rga1-C538 overexpression may down-regulate Rho3 together with Cdc42, leading to a growth defect. In support of this idea, we found that Rho3's suppression activity requires its cycling between the GTP-and GDP-bound states because constitutively active rho3 Q74L or inactive rho3 T30N mutants did not suppress either the growth or cell morphology defect caused by Rga1-C538 overexpression (Fig 4C and 4D).
Does Rga1 normally interact with Rho3 in vivo? To answer this question, we performed a bimolecular fluorescence complementation (BiFC) assay using full-length Rga1 and Rho3 Q74L . Rho3 is palmitoylated at the cysteine 5 residue at the N-terminus [41]. Moreover, palmitoylation is important for Rho3's localization to the plasma membrane [42]. We observed that GFP-Rho3 fusion protein rarely localized to the plasma membrane (Fig 5C), presumably due to the lack of palmitoylation. We then added the region a.a. 1-11 of Rho3 to the N-terminus of In the BiFC assay, we thus used PMT-Venus-N-Rho3 Q74L (Venus is a GFP variant). We found that PMT-Venus-N-Rho3Q74L interacted with Venus-C-Rga1 at the bud cortex and bud neck (Fig 5D, see arrows), locations where Rga1 localized [20].
Taken together, our result showed that Rga1 interacts with Rho3 and preferentially with the active GTP-bound form of Rho3.

Additional sequence upstream of the RhoGAP domain of Rga1 is required for efficient interaction with Rho3
To locate the region in Rga1 that interacts with Rho3, we generated several Rga1-C constructs that share intact C-terminus and examined their interaction with Rho3 and Cdc42 using twohybrid assay. We found that Rga1-C662 (a.a. 662-1007) and Rga1-C700 (a.a. 700-1007) all interacted with Rho3 Q74L . In contrast, the shorter segment, Rga1-C783 (a.a. 783-1007), did not interact with Rho3 Q74L (Fig 6A). All these Rga1-C segments interacted with Cdc42 Q61L efficiently like Rga1-C598 did. This result suggests that the interaction of Rga1 with Rho3 may require both the GAP domain and the region a.a. 700-783 upstream of the GAP domain. Interestingly, we observed that overexpression of Rga1-C783, which no longer interacts with Rho3 in two-hybrid assay, did not impair cell growth (Fig 6B). Nor did it cause a loss of cell polarity (Fig 6C). This observation supports the idea that Rga1 may regulate Rho3 in polarized growth.

Rga1 may be a novel GAP for Rho3
The regulation of Rho GTPases requires guanine nucleotide-exchange factors (GEFs) and GTPase-activating proteins (GAPs). Cdc42 have several GAPs [13,16]. In contrast, there is only one GAP, Rgd1, identified so far for Rho3 [43]. Rgd1 also functions as a GAP for Rho4, which has an overlapping function with Rho3. In this study, we identified an interaction between Rho3 and Rga1, a GAP specific for Cdc42. Rga1 preferentially interacts with the GTPbound form of Rho3 and the interaction requires the GAP domain of Rga1 and additional sequences upstream of the GAP domain. There are two possibilities about the outcome of this interaction. In the first possibility, Rga1 acts as a GAP for Rho3. This interaction would stimulate the intrinsic GTPase activity of Rho3. In the second possibility, Rga1 binds to Rho3-GTP but does not stimulate Rho3's GTPase activity. Because it is not clear at present whether Rga1 may possess GAP activity towards Rho3 in vitro, we are not sure which possibility may be true. Given that high-copy RHO3 but not rho3 Q74L suppressed the effect of Rga1-C overexpression, we favor the first possibility.
The Rho3GAP, Rgd1, localizes to the presumptive bud site and the bud tip in small-budded cells [44]. Rga1 shares a localization pattern similar to, though not identical with, that of Rgd1 [20]. In addition, our BiFC assay showed that Rga1 and Rho3 interact at the bud cortex of small-budded cells and at the bud neck in large-budded cells. These observations suggest that Rga1 may function as a GAP for Rho3. Rga1 may play a role in the regulation of both Cdc42 and Rho3 during the cell cycle.
In this study, we showed that Rga1 is phosphorylated during the cell cycle and the activity of Rga1, presumably the GAP activity, is negatively regulated by Pho85 CDK. Rga1 may also be regulated by mitotic Clb2-Cdc28 CDK complex because we detected Rga1 hyperphosphorylation around mitosis. A previous study identified that Rga1 is negatively regulated by the epsins, which function as adaptors in endocytosis. The epsin Ent2 interacts with Rga1 via a region upstream of the GAP domain in Rga1 [45]. These results suggest that the regulation of Rga1 could be rather complex in the cells.

Rga1 may regulate polarized growth through its interaction with Rho3
A polarized actin cytoskeleton is essential for bud emergence and bud growth. The large, round cell morphology and depolarized actin organization in cells overexpressing Rga1-C suggest that an excess of Rga1 activity impairs actin organization. Because Cdc42 and Rho3 both regulate actin organization and the phenotype of cells with excess Rga1 activity is reminiscent of that of cdc42 and rho3Δ rho4Δ mutants [6,7], the functions of Cdc42 and Rho3 in actin organization are likely affected by excess Rga1 activity.
Cdc42 and Rho3 also have a direct role in exocytosis independent of their functions in actin organization. cdc42-6 and rho3-V51 mutants exhibited a normal actin cytoskeleton but accumulated secretory vesicles at 25°C and 14°C, respectively [8,9]. Moreover, Cdc42 interacts with Sec3 and Exo70, two exocyst components [11,12], whereas Rho3 interacts with Exo70 and Myo2 (myosin V) [10,11]. All three proteins are involved in exocytosis. We speculate that the toxicity of excess Rga1 activity may be partly due to the down-regulation of both Cdc42 and Rho3 functions in exocytosis. This is supported by the observation that high-copy SEC9 and SRO7, which suppressed the growth defect of both cdc42-6 and rho3-V51 mutants [9], also effectively suppressed the toxicity of Rga1-C.
The functions of Cdc42 and Rho3 in exocytosis appear to be partially overlapping because cdc42-6 is synthetically lethal with rho3-V51 mutant. In addition, high-copy RHO3 suppressed the growth defect of cdc42-6 mutant [9]. Conversely, high-copy CDC42 suppressed the growth defect of rho3-V51 mutant [8]. CDC42 or BEM1 also suppressed the growth defect of rho3Δ rho4Δ cells [7]. Despite this, some major differences exist between the functions of Cdc42 and Rho3 in exocytosis. First, Rho3 seems to control both vesicular transport and vesicle fusion while Cdc42 mainly controls vesicle fusion because rho3-V51 mutant accumulated exocytic vesicles in both the bud and the mother cell at 14°C while cdc42-6 mutant accumulated vesicles mainly in the bud [8,9]. Moreover, Rho3 interacts with Myo2 whereas Cdc42 does not (Fig 5B). Second, Cdc42 function in exocytosis may be restricted to early stage of bud development whereas Rho3 may be implicated in both early and subsequent stages of bud growth because cdc42-6 mutant accumulated vesicles only in the small-budded stage [9]. We speculate that the ability of RHO3 to suppress the toxicity of Rga1-C on high-copy plasmids is likely because Rho3 regulates both the transport and fusion of exocytic vesicles while Cdc42 only regulates vesicle fusion. A higher dose of Rho3 could not only counteract the negative regulation by Rga1, allowing the cells to polarize actin cytoskeleton and deliver secretory vesicles into the bud, it could also replenish Cdc42 at the bud tip in small-budded cells because Cdc42 is known to be delivered to the bud by secretory vesicles [46]. In contrast to this scenario, a higher dose of Cdc42 may only strengthen its own force but not that of Rho3. As a result, down-regulation of Rho3-regulated transport of Cdc42 into the growing bud by excess Rga1 activity may, in turn, dampen the effect of an increase in Cdc42 level. This may explain why high-copy CDC42 did not suppress the toxicity of Rga1-C.
Rga1 may normally play a role in exocytosis via its interaction with Rho3. By modulating the functions of both Cdc42 and Rho3, Rga1 may coordinate Cdc42 and Rho3 to ensure proper bud growth.

Strains and genetic methods
Yeast strains used in this study are listed in Table 1. Standard culture media and genetic techniques were used expect where noted. Escherichia coli strains DH12S (Life Technologies, Gaithersburg, MD) and DH5α (TaKaRa, Japan) were used as hosts for plasmid manipulation. SRG-Ura medium containing 2% galactose and 1% raffinose was used to overexpress genes driven by the GAL1 promoter. SC-Ura medium contained 2% dextrose. Oligonucleotide PCR primers were purchased from Sangon Biotech (Shanghai, China).

Immunoprecipitation, dephosphorylation and cell synchronization
Yeast cells were lysed in lysis buffer (10 mM HEPES-NaOH, 50 mM NaCl, 10% glycerol, 0.1% NP-40, pH7.5, plus 1×protease inhibitor cocktail and 1×phosphatase inhibitor cocktail) by vortexing in the presence of glass beads at 4°C. Cell debris was removed by centrifugation at 10,000 × g for 15 min. For immunoprecipitation, the supernatant was diluted with lysis buffer and incubated with mouse monoclonal anti-HA antibody HA.11 affinity matrix (Covance, Richmond, CA) for 2 hr at 4°C. The beads were washed three times. Bound proteins were eluted by boiling for 5 min in 2 × SDS sample buffer. Samples were then analyzed by SDS-PAGE, followed by standard immunoblotting procedure using enhanced chemiluminescence (ECL) reagents. Mouse monoclonal anti-HA and rabbit polyclonal anti-porin antibodies were used for the blotting.
For the dephosphorylation of immunoprecipitates, bound proteins were treated with 1 unit of shrimp alkaline phosphatase (SAP, USB Corporation, Cleveland, OH) in the dephosphorylation buffer (50 mM Tris-HCl, 5 mM MgCl 2 , pH 8.5, plus 1× protease inhibitor cocktail) for 1 hr at 37°C before eluted from the beads. For the synchronization of cells of strain YZT52 in G1, cells were incubated with α factor at the final concentration of 0.5 μg/ml for 120 min at 30°C. Cells were then washed and resuspended in fresh medium.

Microscopy
An Olympus BX51 microscope (Tokyo, Japan) and a Retiga 2000R CCD camera (QImaging Corporation, Canada) were used to visualize cell morphology by differential interference contrast (DIC) and fluorescent microscopy. The images were acquired using QCapture Suite (QImaging Corporation, Canada). Image processing was performed with ImagePro Plus (Glen Mills, PA). For the examination of F-actin and DNA, yeast cells were stained with 1.0 μg/ml phalloidin-TRITC (tetramethyl rhodamine isocyanate) (Sigma-Aldrich) and 1.0 μg/ml of Hoechst 33258 (Polysciences, Inc.) at room temperature for 30 min and washed three times with 1×phosphate-buffered saline (PBS) (pH 8.0). Methylene blue staining of yeast cells was performed as described previously [23]. Methylene blue staining solution (0.001% methylene blue, 2% sodium citrate) was mixed with an equal volume of cell culture, and cells were observed immediately.