Gic1 and Gic2 are two Cdc42/Rac interactive binding (CRIB) domain-containing effectors of Cdc42-GTPase that promote polarized cell growth in S. cerevisiae. To identify novel genes that functionally interact with Gic1 and Gic2, we screened for high-copy suppressors of a gic1 gic2 temperature-sensitive strain. We identified two pairs of structurally related genes, SKG6-TOS2 and VHS2-MLF3. These genes have been implicated in polarized cell growth, but their functions have not previously been characterized. We found that overproduction of Skg6 and Tos2 in wild-type cells causes aberrant localization of Cdc3 septin and actin structures as well as defective recruitment of Hof1 and impaired formation of the septum at the mother-bud neck. These data suggest a negative regulatory function for Skg6 and Tos2 in cytokinesis. Consistent with this model, deletion of SKG6 suppresses the growth defects associated with loss of HOF1, a positive regulator of cytokinesis. Our analysis of the second pair of gic1 gic2 suppressors, VHS2 and MLF3, suggests that they regulate polarization of the actin cytoskeleton and cell growth and function in a pathway distinct from and parallel to GIC1 and GIC2.
THE establishment of cell polarity is required for differentiation in most cell types and is a critical step in cellular morphogenesis (Drubin and Nelson 1996; Johnson 1999; Nelson 2003). Budding in Saccharomyces cerevisiae involves polarized cell growth during which the actin cytoskeleton is asymmetrically organized. This process is both spatially and temporally regulated to synchronize with progression of the cell cycle. Bud emergence is accompanied by entry of cells into a new cell cycle (G1-to-S transition) and establishment of the mother–daughter axis of polarization. F-actin structures including actin cables and cortical actin patches are directed to the growing bud cortex to promote apical bud growth until G2–M phase when they are distributed in the daughter cell to promote isotropic bud growth. Upon sufficient bud growth, actin cables and patches are briefly depolarized in both the mother and the daughter cell until cell cycle cues signal their repolarization toward the mother-bud neck to promote cytokinesis (Pruyne and Bretscher 2000b).
Cytokinesis in budding yeast is accomplished by concerted actions of actomyosin ring closure and septum formation at the site of cell division (Bi et al. 1998). Evolutionarily conserved structural GTPases, called septins, form a system of filaments and are essential for cytokinesis (Moffat and Andrews 2003). Early in the cell cycle, septins mark the site for future cell division, which coincides with the site of bud emergence (Longtine and Bi 2003). Subsequently, septins promote cytokinesis by recruiting proteins to the bud neck required for formation of the actomyosin ring (e.g., Myo1, F-actin, and Cyk1) and the primary septum (e.g., chitin synthase II components) (Schmidt et al. 2002). Septins also are required for reorientation of actin cables to the bud neck at cytokinesis, which targets vesicle delivery for septum formation at the division plane (Adams and Pringle 1984).
Hof1 (also known as Cyk2) is a homolog of Schizosaccharomyces pombe Cdc15, which is a founding member of the PCH/FCH family of conserved proteins involved in actin-based processes and shown to organize sterol-rich membrane domains at the site of cell division (Lippincott and Li 2000; Takeda et al. 2004). Hof1 plays an important role in cytokinesis in S. cerevisiae, localizing as a ring structure to the mother-bud neck and stabilizing the actomyosin ring during contraction (Lippincott and Li 1998). However, on the basis of genetic evidence the primary cytokinetic function of Hof1 appears to be in septum formation (Vallen et al. 2000). Cells lacking Hof1 display cytokinesis defects appearing as chains of cells with connected cytoplasms. This phenotype of hof1 cells results from failure in septum formation between two cell bodies before subsequent rounds of budding (Lippincott and Li 1998; Vallen et al. 2000)
Polarized growth is accompanied not only by reorganization of the actin cytoskeleton, but also by insertion of a new membrane at the cell cortex, biosynthesis of cell-wall material at the sites of active growth, and remodeling of the cell wall required for cell expansion. Both actin cytoskeleton reorganization and cell-wall remodeling are controlled by Rho-type GTPases. The evolutionarily conserved GTPase Cdc42 is recruited to the incipient bud site and (through activation of its downstream effectors) directs polarized redistribution of the actin cytoskeleton (Ziman et al. 1993). Actin cables serve as tracks for the myosin V-dependent delivery of secretory vesicles containing new plasma membrane and cell-wall material to sites of polarized growth (Pruyne and Bretscher 2000a). Actin patches, in addition to being sites of dynamic actin assembly and endocytosis (Pruyne and Bretscher 2000a), provide sites where deposition of new cell-wall material occurs (Mulholland et al. 1994). In contrast to Cdc42, Rho1 and Rho2 affect cell-wall synthesis, both by directly stimulating the glucan synthase complex and by indirectly increasing expression of cell-wall biosynthesis machinery through upregulation of the Pkc1-MAP kinase pathway (Cabib et al. 1998; Schmidt and Hall 1998).
Gic1 and Gic2 are two related Cdc42/Rac-interactive binding domain-containing effectors of Cdc42 first identified as multicopy suppressors of a BEM2 (Rho-type GTPase) mutant allele and shown to bind an activated form of Cdc42 (Brown et al. 1997; Chen et al. 1997). Deletion of both GIC1 and GIC2 results in loss of cell polarity and cell growth at elevated temperatures. gic1 gic2 cells are enlarged and abnormally rounded with defects in bud-site selection and actin polarization (Chen et al. 1997). However, the mechanism by which Gic1 and Gic2 contribute to polarized cell growth remains unclear. To gain new insights, we performed a high-copy suppression screen on a gic1 gic2 strain and identified two pairs of structurally related genes that function with GIC1 and GIC2 in directing polarized growth and cell division.
MATERIALS AND METHODS
Strains, media, genetic techniques, and growth conditions:
Yeast strains used in this study are listed in Table 1. Preparation of rich medium (YEPD), synthetic minimal medium (SD), and SD with necessary supplements was performed as described (Rose et al. 1990). Standard methods of yeast genetics and recombinant DNA manipulations were carried out (Sambrook et al. 1989; Guthrie and Fink 1991). Construction of gene deletion mutants and replacement of the chromosomal native promoters by the inducible GAL1 promoter were carried out by a one-step PCR-mediated homologous recombination technique (Longtine et al. 1998). For overexpression of genes under the control of the GAL1 promoter, cultures were grown at 26° in rich medium containing raffinose (2%) to a cell density of 1 × 107 cells/ml, followed by dilution to 2.5 × 106 cells/ml with rich medium containing galactose (4%) and incubation at 26° or 37° for indicated periods of time. Strains expressing chromosomally tagged HOF1-GFP and MYO1-GFP fusions were generated by transforming yeast strains with the YCplac111-based integration plasmids pSK1051 and pSK1052 (gift from K. S. Lee, National Cancer Institute, Bethesda, MD) bearing HOF1-GFP∷LEU2 and MYO1-GFP∷LEU2 fragments, respectively.
Cloning of high-copy suppressors of gic1 gic2:
Briefly, a temperature-sensitive (Ts−) gic1-Δ1∷LEU2 gic2-1∷HIS3 ura3-52 strain (CCY1024-19C) that fails to grow at ≥33° was transformed with a yeast genomic DNA library constructed in the high-copy number URA3-plasmid YEp24 (Carlson and Botstein 1982). Ura+ transformants were selected on supplemented SD agar lacking uracil. After 20–24 hr of growth at 26°, the plates containing Ura+ transformants were shifted to 35°. Three days later, viable Ura+ transformants were identified. Approximately 100,000 library transformants were screened, which yielded 426 transformants able to grow at 35°. Plasmids from these transformants were recovered and amplified in Escherichia coli. Colony hybridization of E. coli strains bearing each of these plasmids with genes that were known to suppress the temperature-sensitive growth defect of gic1 gic2 cells revealed that 232 of the suppressor plasmids contained previously identified suppressor genes (including BEM1, CDC42, GIC1, GIC2, MSB3, and SSD1-v1) (Brown et al. 1997; Chen et al. 1997; Bi et al. 2000). Restriction enzyme digestion and partial sequencing of the remaining 194 candidates revealed that they represented 11 unique classes of plasmids (Figure 1A). Comparison of the insert sequences from different classes of plasmid against the Saccharomyces Genome Database enabled prediction of a single ORF on each plasmid that likely was responsible for suppression of the temperature-sensitive growth defect of gic1 gic2 cells. This was confirmed by subcloning individual ORFs where possible or by disrupting/deleting that ORF from the library plasmid and testing the ability of modified plasmids to suppress the growth defects of gic1 gic2 cells at restrictive temperatures.
Microscopy was performed using Zeiss Axioscope (Carl Zeiss, Oberkochen, Germany). Images were captured using a MicroMax CCD camera (Princeton Instruments, Trenton, NJ) and processed using IPLab spectrum (Scanalytics, Fairfax, VA) software.
For examining the morphological features, cells were fixed with formaldehyde (3.7%; EM Sciences, Gibbstown, NJ) for 0.5–1 hr at room temperature and washed twice with PBS buffer. Actin staining was carried out essentially as described previously (Pringle et al. 1989). For staining nuclei, fixed cells were resuspended in PBS buffer containing 1 μg/ml 4′,6-diamidino-2-phenylindole (DAPI; Sigma Chemical, St. Louis) and incubated for 10 min before visualization. Chitin at the bud neck was examined by staining fixed cells with 0.01 mg/ml calcofluor (Sigma Chemical). To evaluate cell separation, fixed cells were washed twice with PBS, once with sorbitol buffer (1 m sorbitol in 50 mm K-phosphate, pH 7.5), and incubated with 0.2 mg/ml zymolyase 20T (ICN Biomedicals, Aurora, OH) in sorbitol buffer containing 2 mm DTT at 37° for 10 min. More than 90% of such treated cells lost their refractile appearance as observed by light microscopy, indicating cell-wall removal was efficient.
For experiments to detect the subcellular localization of Cdc3-GFP, Myo1-GFP, and Hof1-GFP, cells were cultured at 37° for indicated periods of time, fixed with 1% formaldehyde on ice for 10 min, washed once with ice-cold PBS buffer, and scored immediately thereafter. To detect the subcellular localization pattern of Vhs2 and Mlf3, GFP-VHS2 and GFP-MLF3 fusions were expressed under the ACT1 promoter from a low-copy plasmid pTD125 and visualized in live cells from logarithmically growing cultures.
Identification of high-copy suppressors of gic1 gic2 cells:
gic1 gic2 cells are viable but exhibit a Ts− growth defect at ≥33° (Chen et al. 1997). We used this phenotype to isolate genes, which upon overexpression from high-copy number plasmids could alleviate the Ts− growth defect of gic1 gic2 cells. With this approach, we anticipated the identification of genes that functionally interact with GIC1 and GIC2 either in a common linear pathway or in redundant parallel pathways to facilitate polarized cell growth. We identified as suppressors many known polarity-related genes, including AXL2, BNI1, CLN2, MSB1, MSB2, RSR1, and STE20 (lacking autoinhibitory sequences upstream of nucleotide 355 of the STE20 ORF, and hence NΔ-118-STE20) (Figure 1A). Surprisingly, overexpression of SSN6, which encodes a conserved transcriptional repressor that controls the expression of ∼3% of yeast genes (Smith and Johnson 2000), also complemented the Ts− growth defect of gic1 gic2 cells. Since Ssn6 functions in a complex with the Tup1 transcriptional repressor (Keleher et al. 1992), we examined whether overexpression of TUP1 also could complement the Ts− growth defect of gic1 gic2 cells. Indeed, a 2μ plasmid bearing TUP1 suppressed the Ts− defect of gic1 gic2 cells (Figure 1A), albeit slightly less efficiently than the SSN6 plasmid.
Interestingly, three ORFs identified from our screen had no well-characterized function in polarized cell growth at the time that we identified them (see below for recent reports implicating these genes in polarity pathways). These ORFs were YHR149c (SKG6), YGR221c (TOS2), and YIL135c (VHS2) (Figure 1A). Skg6 and Tos2 share sequence homology with each other, and Vhs2 is homologous in sequence to another S. cerevisiae protein, Mlf3. Therefore we tested whether overexpression of MLF3 could suppress the Ts− phenotype of gic1 gic2 cells. Indeed, a 2μ plasmid bearing MLF3 partially suppressed the Ts− defect of gic1 gic2 cells (Figure 1A). Note that simultaneous overexpression of these homologous gene pairs, SKG6 and TOS2, or VHS2 and MLF3, showed no additive effects in suppressing the gic1 gic2 phenotype (not shown).
We next examined the ability of these suppressors to rescue actin organization defects of gic1 gic2 cells. Most suppressor plasmids decreased the fraction of gic1 gic2 cells with a depolarized actin cytoskeleton, reflected by an overall decrease in the fraction of unbudded cells (Figure 1B). The ability of these suppressors to rescue actin polarization defects of gic1 gic2 cells correlated with their ability to suppress the Ts− growth defect (Figure 1A), suggesting that these phenotypes may be interdependent.
Since SKG6, TOS2, VHS2, and MLF3 suppressed the Ts− growth and actin organization defects of gic1 gic2 cells, we further investigated their roles in polarized cell growth.
Skg6 and Tos2 are not essential for polarized cell growth:
Skg6 and Tos2 proteins are 734 and 622 amino acids in length, respectively. Their primary sequences share 34% identity and 47% similarity (supplemental Figure 1 at http://www.genetics.org/supplemental/). Each is predicted to contain a single transmembrane domain within the N-terminal region (residues 71–99 in Skg6 and residues 37–64 in Tos2), and Skg6 contains two putative peptide cleavage sites (after residues 23 and 95). Both Skg6 and Tos2 localize to sites of polarized growth (i.e., throughout the bud cortex in small-budded cells, as a cap along the bud cortex in medium-budded cells and as a double-ring structure at the bud neck in large-budded cells; Drees et al. 2001 and our unpublished results). Skg6 is known to interact in the two-hybrid assay with the Cdc42 effector Cla4 (Uetz et al. 2000) and two potential negative regulators of Cdc42, Zds1, and Zds2 (Bi and Pringle 1996), which themselves interact in the two-hybrid assay with Gic1 and Gic2 (Uetz et al. 2000; Drees et al. 2001). Tos2 is known to interact in the two-hybrid assay with Pkc1 and Cdc24, the guanine nucleotide exchange factor for Cdc42 (Drees et al. 2001). In fact, one study indicated that Tos2 may help anchor Cdc24 to sites of polarized growth (Toenjes et al. 2004). These observations, together with our identification of SKG6 and TOS2 as high-copy suppressors of gic1 gic2, suggest that these two genes likely regulate polarized cell growth.
To investigate further the potential function of SKG6 and TOS2 in polarized growth, we generated strains that lack one or both genes. skg6 and tos2 single-mutant cells and skg6 tos2 double-mutant cells each were viable and showed normal rates of growth at a range of temperatures from 13° to 37°. Mutant cells also displayed normal cell size and morphology and had no obvious defects in bud-site selection, actin cytoskeleton organization, mating projection formation, or sporulation (not shown).
Skg6 or Tos2 overproduction leads to defects in apical-to-isotropic growth switch and cytokinesis:
We next tested whether overproduction of Skg6 or Tos2 perturbs polarized growth in wild-type cells. To this end, we replaced the chromosomal native promoter of SKG6 and TOS2—either singly or in combination—with an inducible GAL1 promoter. Cells were grown in rich medium containing galactose to induce the overproduction of Skg6 or Tos2 at 26°. This resulted in the appearance of a small fraction of cells with abnormally elongated buds, an effect that was greatly exacerbated at 37° (Figure 2A and Table 2). Furthermore, simultaneous overproduction of Skg6 and Tos2 increased the fraction of cells with elongated buds. Interestingly, a significant fraction of Skg6-overproducing cells (and, to a lesser extent, Tos2-overproducing cells) had a second bud that in most instances emerged from the previously formed elongated bud (denoted by arrowheads in Figure 2A). The appearance of multibudded cells was accompanied by a reduction in the fraction of cells that were unbudded or had a single bud. Together, these phenotypic changes suggested that overproduction of Skg6 or Tos2 disrupts the switch from apical-to-isotropic growth and causes a delay or defect in the normal execution of cytokinesis. To ascertain whether the elongated/multibudded phenotype of Skg6- and Tos2-overproducing cells results from a defect in cell division or, alternatively, from a defect in DNA duplication and segregation as seen for the Ts− cdc4 mutant cells (Goh and Surana 1999), we examined their nuclei by DAPI staining. Like wild-type cells, Skg6- and Tos2-overproducing cells showed normal nuclear division and segregation regardless of the shape and number of their buds (Figure 2B). Additionally, the multibudded phenotype of Skg6- or Tos2-overproducing cells could not be resolved by zymolyase treatment (Figure 2C), further suggesting that the defect lies in cytokinesis rather than cell separation.
Skg6 or Tos2 overproduction causes septin mislocalization:
Current models for cytokinesis in S. cerevisiae suggest that there are two parallel pathways—one involving assembly and contraction of an actomyosin ring and the other involving formation of a septum—both of which are spatially and temporally coordinated. In the absence of actomyosin ring formation, the septum pathway is sufficient to drive cell division. Key components of each of these pathways, Myo1 and Hof1, respectively, localize independently of each other at the bud neck. However, both proteins depend on septins for their recruitment and maintenance at the bud neck, making septins essential for cytokinesis (Field and Kellogg 1999; Bi 2001).
The cytokinesis defect in Skg6- and Tos2-overproducing cells prompted us to examine the localization of septins in these cells. CDC3 is an essential gene encoding one of the septin proteins, and cdc3 conditional mutants arrest with hyperelongated buds (Kim et al. 1991; Longtine and Bi 2003). In budded wild-type cells, we found that Cdc3-GFP localizes as an hourglass structure at the bud neck (Figure 2D, a and b), often appearing as two closely apposed rings as previously described (Kim et al. 1991). However, in elongated/multibudded cells overproducing Skg6 or Tos2, Cdc3-GFP was aberrantly organized and/or mislocalized (72 and 20%, respectively; Figure 2D). In many cells, Cdc3-GFP was found at some distance from the bud neck, skewed toward either the daughter (Figure 2D, g) or the mother (not shown). In other cells, Cdc3-GFP appeared as small dots or short bars that were mislocalized against one side of the bud (Figure 2D, e, h, and j). Occasionally, the two rings of Cdc3-GFP were split asymmetrically (Figure 2D, f) or separated excessively (Figure 2D, c). Since localization of different septin proteins is interdependent (Haarer and Pringle 1987; Kim et al. 1991), it is likely that localization of the other septins is also impaired in Skg6- and Tos2-overproducing cells.
Localization of Myo1, F-actin, and Hof1 in Skg6-overproducing cells:
The results above suggest that overproduction of Skg6 (and Tos2) interferes with the localization of the cell division machinery and thereby with the normal execution of cytokinesis. To further dissect which of the aforementioned cytokinesis pathways is likely impaired, we examined in Skg6-overproducing cells the localization patterns of several cytokinetic components known to be downstream of, and dependent on, septins: Myo1, F-actin, and Hof1.
Myo1–GFP expressed from its chromosomal locus under the endogenous promoter was found at the mother-bud neck of medium- to large-budded cells in both wild-type cells and Skg6-overproducing cells (74 vs. 61%, respectively; n = 100 cells; Figure 3A). Further, the fraction of cells showing Myo1-GFP as a dot (indicating a contracted actomyosin ring) was similar for wild-type and Skg6-overproducing cells. However, 64% of elongated/multibudded Skg6-overproducing cells showed a Myo1–GFP band that was off center from the mother-bud junction (Figure 3A, c), skewed toward either the mother or the daughter cell (Figure 3A, e and f, respectively), a pattern never observed in wild-type cells. Thus, while overexpression of Skg6 does not impair recruitment of Myo1 to the bud neck, it affects Myo1 spatial organization. It is possible that these defects in Myo1 arise as a consequence of the defects in septin organization and localization caused by Skg6 overexpression (above), especially given that Myo1 localization depends on septins.
Next, we examined the effects of Skg6 overproduction on polarized actin organization and Hof1-GFP localization at the bud neck, both of which depend on septins and are necessary for septum formation and cell separation (Adams and Pringle 1984; Lippincott and Li 1998; Vallen et al. 2000). Approximately 72% of Skg6-overproducing cells with elongated buds showed actin repolarization to the bud neck that precedes cytokinesis. However, unlike the condensed double ring of actin patches seen in wild-type cells, >74% of these cells demonstrated a fanned-out actin patch organization that was often biased toward one side of the mother-bud neck (Figure 3B and Table 3). Analysis of Hof1-GFP expressed from its chromosomal locus under the endogenous promoter also revealed an interesting phenotype, specifically in Skg6-overproducing cells with abnormal cell morphology. For the majority of medium- to large-budded wild-type and Skg6-overproducing cells with normal morphologies, Hof1–GFP showed a similar localization pattern, found either as a single band or as a doublet at the bud neck. However, in 67% of the elongated/multibudded cells overproducing Skg6, Hof1-GFP was absent from the mother-bud neck (Figure 3C).
skg6 suppresses defects of the hof1Δ mutant:
Hof1 restricts the septal chitin to the bud neck and promotes septum formation, which is necessary for cell separation after actomyosin ring contraction (Vallen et al. 2000). hof1 cells fail to form septa and hence accumulate as chains of cells with a continuous cytoplasm (Lippincott and Li 1998; Vallen et al. 2000). Since we observed that Skg6 overproduction impairs Hof1 recruitment to the cell division site, we hypothesized that the cytokinesis defect of Skg6-overproducing cells may arise from negative regulation of Hof1. If such a mechanism exists, deletion of skg6 would be predicted to suppress the phenotype of hof1 mutant cells. To test this hypothesis, we examined genetic interactions among SKG6, TOS2, and HOF1 by tetrad analysis. In the ssd1-d genetic background of our laboratory strains, hof1 cells formed tiny spore colonies and exhibited growth defects at temperatures ranging from 26° to 37° (Figure 3, D and E). The tos2 mutation had no effect on this hof1 phenotype (Figure 3, D and E). However, hof1 skg6 cells and hof1 skg6 tos2 cells consistently formed spore colonies that were of regular size, indicative of normal cellular growth (Figure 3D). These cells also exhibited normal growth at 26° (Figure 3E) and improved growth (relative to hof1 cells) at 33° on YEPD agar (not shown). This genetic interaction is consistent with observed cytokinesis defects upon Skg6 overproduction and supports the hypothesis that Skg6 performs an inhibitory role in cytokinesis, which is antagonistic to the positive role of Hof1.
Skg6 and Tos2 overproduction causes defect in primary septum formation:
The results above prompted us to next examine primary septum formation in Skg6-and Tos2-overproducing cells. Formation of the septum requires cell-wall synthesis across the bud neck (Bi 2001). To assess deposition of cell wall at the neck, we treated wild-type and Skg6- or Tos2-overproducing cells with calcofluor white, which stains cell-wall chitin. All wild-type budded cells showed an intensely stained band at the bud neck, indicative of proper septum formation and closure of the neck (Figure 3F, open arrowheads). In contrast, elongated/multibudded cells resulting from Skg6 or Tos2 overproduction displayed calcofluor staining that was extremely faint or absent at the neck (Figure 3F, solid arrowheads). This phenotype suggests that the bud necks remained open in these cells as a result of defective septum formation.
Taken together, these data suggest that SKG6 and TOS2 when overproduced negatively regulate cytokinesis by disrupting septum formation. One possible mechanism for this disruption is their mislocalization of Hof1 and/or septins.
Vhs2 and Mlf3 are essential for polarized growth at elevated temperature:
We next investigated the genetic roles of the second pair of gic1 gic2 suppressors, VHS2 and MLF3. Vhs2 and Mlf3 proteins are 436 and 452 amino acids in length, respectively, and their primary sequences share 30% identity and 42% similarity (supplemental Figure 2 at http://www.genetics.org/supplemental/). VHS2 was previously identified as a high-copy suppressor of the lethality of hal3 sit4 cells, which are defective in G1-to-S-phase transition (Munoz et al. 2003). MLF3 was identified as a high-copy suppressor of the growth defects caused by the immunosuppressive drug leflunomide (Fujimura 1998). However, the functions of Vhs2 and Mlf3 have not been characterized and their sequences provide no obvious clues regarding their cellular functions. Our identification of VHS2 and MLF3 as high-copy suppressors of the gic1 gic2 phenotype suggests that they likely contribute to polarized cell growth.
To gain further insights into VHS2 and MLF3 functions, we generated deletion strains lacking one or both genes. Analysis of the growth phenotypes of vhs2 and mlf3 mutants revealed that neither VHS2 nor MLF3 is essential for the viability of haploid cells grown at a range of temperatures from 13° to 37°, although mlf3 cells formed slightly smaller colonies on YEPD agar at 37° (not shown). However, vhs2 mlf3 cells were severely impaired for growth at 37° (Figure 4A). Interestingly, these defects were more severe in diploids, as homozygous diploid vhs2 and mlf3 single-mutant and vhs2 mlf3 double-mutant cells exhibited stronger Ts− growth defects (and calcofluor white sensitivities; see below) than their haploid counterparts (Figure 4A). The enhanced phenotype in diploids compared to haploids has been reported for other cell polarity mutants, including bni1 and msb3 msb4 (Bi et al. 2000).
When incubated at 37° in liquid YEPD medium, diploid vhs2 mlf3 cells became increasingly large and round over time, a phenotype suggesting loss of polarity (Figure 4B). However, unlike cdc42 or gic1 gic2 cells (Chen et al. 1997; Johnson 1999), vhs2 mlf3 cells did not arrest exclusively as unbudded cells. Instead, an enrichment of unbudded vhs2 mlf3 cells occurred gradually after shift to the nonpermissive temperature (Figure 5B). Unbudded vhs2 mlf3 cells that appeared large and round also showed a characteristic “wrinkled” morphology (Figure 4B, arrowhead), suggesting that loss of cell polarity may lead to weakening of the cell wall and lysis of these cells. As expected, the morphological defects (loss of polarity and “wrinkled” appearance) were much weaker in diploid vhs2 and mlf3 single-mutant cells compared to diploid vhs2 mlf3 double-mutant cells (Figure 4B).
We next localized Vhs2-GFP and Mlf3-GFP fusion proteins expressed from low-copy plasmids under control of the ACT1 promoter. These constructs complemented the temperature-sensitive growth defects of vhs2 mlf3 double-mutant cells (not shown), indicating that they are functional. In both cases, cytoplasmic localization was observed (Figure 4C). Similar results were obtained for chromosomally tagged GFP fusions of VHS2 and MLF3 expressed from their endogenous promoters (not shown).
Actin organization defects in vhs2 mlf3 cells:
Since a polarized actin cytoskeleton (patches and cables) is required for polarized cell growth (Pruyne and Bretscher 2000b), the morphological defect of vhs2 mlf3 cells prompted us to examine actin organization in these cells. When grown at 37°, ∼50% of unbudded wild-type diploid cells had a polarized pattern of actin patches that were concentrated at one end of the cell (Figure 5, A and C), indicative of cells poised for bud emergence. Essentially all small-budded wild-type cells had a highly polarized pattern of actin organization, with actin patches being concentrated in the small buds (Figure 5, A, open arrowheads, and D). In contrast, diploid vhs2 mlf3 cells showed depolarized actin organization, which intensified with increasing time of incubation at 37°. After 7.5 hr growth at 37°, only ∼2% of unbudded cells showed a polarized distribution of actin structures (Figure 5C). Furthermore, ∼25% of small-budded vhs2 mlf3 cells had actin patches concentrated in the mother, with a complete loss of patches from the bud (Figure 5, A, solid arrowheads and inset, and D). This subset of cells had atypically round and large mother-cell bodies, suggesting that abnormal F-actin distribution may have led to inappropriate expansion of the mother-cell bodies at the expense of bud growth. In addition, the majority of budded vhs2 mlf3 cells that showed a fairly normal polarized distribution of actin had an unusually large number of actin patches in the mother-cell bodies, indicating that the actin cytoskeleton in these cells might not be fully polarized. Together, these data suggest that the combined loss of Vhs2 and Mlf3 impairs polarized actin organization, which in turn disrupts polarized growth.
Defect in mating projection formation in vhs2 mlf3 cells:
Next, we examined whether the actin polarization defect of vhs2 mlf3 cells was reflected in their ability to form mating projections upon pheromone treatment. At 26°, MATa haploid vhs2 mlf3 cells responded to α-factor in a manner similar to wild-type cells and initiated the formation of mating projections. At 37°, however, unlike wild-type cells that formed elongated mating projections, vhs2 mlf3 cells became somewhat larger and rounder with a pointed structure restricted to one side of the cell surface that apparently failed to elongate as a mating projection (Figure 6). This result further reinforces the roles of Vhs2 and Mlf3 in cell polarization.
Cell-wall defects in vhs2 mlf3 cells:
We noted that after 5 hr of growth at 37°, vhs2 mlf3 cell cultures became visibly flocculent and contained small clumps of cells that were not separable by mild sonication (Figure 4B). This observation, taken together with the wrinkled cell morphology mentioned above, suggests that vhs2 mlf3 cells may have cell-wall defects. To investigate this possibility further, we used two complementary assays. First, we compared the sensitivity of vhs2 and mlf3 single-mutant and vhs2 mlf3 double-mutant cells to calcofluor white, which interferes with cell-wall assembly and exacerbates the growth defects of strains defective in this process (Ram et al. 1994). Second, we tested whether addition of sorbitol (an osmotic stabilizer) to the growth medium could alleviate growth defects of the mutants. Sorbitol is known to suppress the cell lysis defects of some cell-wall integrity mutants, including those defective in the Rho1-Pkc1 pathway (Kamada et al. 1996; Heinisch et al. 1999). Our results showed that, in a haploid background, vhs2 mlf3 double-mutant cells were supersensitive to calcofluor white (Figure 4A). In a diploid background, this effect was exacerbated so that even vhs2 and mlf3 single-mutant cells were supersensitive to calcofluor (Figure 4A). In addition, the growth defects of mlf3 and vhs2 mlf3 mutants at 37° were almost completely rescued by addition of 1 m sorbitol to the growth medium (Figure 4A). Since the onset of the actin depolarization phenotype precedes the wrinkled morphology phenotype (Figure 4B and Figure 5, C and D), we believe that the cell-wall defect in vhs2 mlf3 cells likely occurs as a consequence of actin depolarization. Consistent with this, we found that providing osmotic support in the growth medium is not sufficient to prevent the actin polarization defects of vhs2 mlf3 cells. The number of unbudded cells with depolarized actin cytoskeleton and small-budded cells with buds devoid of actin patches remained roughly similar in vhs2 mlf3 cultures grown in the presence or absence of 1 m sorbitol (not shown).
All of these data are consistent with vhs2 mlf3 mutants being defective in actin polarization at 37°, leading to defects in cell-wall assembly and cell lysis, which in turn causes the wrinkled cell morphology.
GIC1 and G1 cyclin genes as high-copy suppressors of vhs2 mlf3 cells:
We next investigated the functional relationships of Vhs2 and Mlf3 with other proteins known to regulate polarized cell growth. Toward this end, we individually overexpressed 59 different “polarity” genes to test their effects on the Ts− growth of haploid vhs2 mlf3 cells (for a complete list of genes tested, see supplemental Figure 3 at http://www.genetics.org/supplemental/). We identified GIC1 (and to a lesser extent GIC2) as an efficient high-copy suppressor of the vhs2 mlf3 growth defect at 37° (Figure 7A). Since VHS2 and MLF3 also are high-copy suppressors of gic1 gic2 cells (Figure 1A), this pattern of reciprocal suppression suggests that Vhs2 and Mlf3 may function in a pathway that is genetically redundant with a Gic1/Gic2 pathway. Consistent with this interpretation, vhs2 mlf3 gic1 gic2 cells showed a more severe Ts− growth defect than vhs2 mlf3 or gic1 gic2 cells (Figure 7B).
We also identified the G1 cyclins CLN1, CLN2, and PCL1 as robust high-copy suppressors of vhs2 mlf3 cells (Figure 7A). Cln1 and Cln2 associate with Cdc28 to form an active Cdk–cyclin complex that promotes G1-to-S-phase transition. Likewise, Pcl1 and Pcl2 associate with Pho85 to form an active Cdk–cyclin complex that promotes G1-to-S phase transition, and this complex becomes essential in the absence of Cln1 and Cln2 (Measday et al. 1997). The identification of G1 cyclin genes as high-copy suppressors of vhs2 mlf3 cells prompted us to investigate whether vhs2 mlf3 mutants are defective in G1-to-S transition. FACS analysis of vhs2 mlf3 cells released from α-factor (G1) arrest into the cell cycle at 37° revealed that they entered S phase normally and progressed through the cell cycle with wild-type kinetics (not shown). This result suggests that the ability of G1 cyclins to suppress vhs2 mlf3 growth defects may not be related to their role in promoting cell cycle progression. Instead, suppression may stem from the additional roles of these cyclins in regulating Cdc42-related polarity functions. Both Cdc28-Cln1/2 and Pho85-Pcl1,2 are required for phosphorylation of Cdc24, the guanine nucleotide exchange factor for Cdc42, and for assembly of the septin ring (Moffat and Andrews 2004). In addition, several lines of genetic evidence suggest that Pho85–Pcl1,2 complexes promote Cdc42 activity or a process that substitutes for Cdc42 function (Lenburg and O'Shea 2001). Since Cdc42 is a master regulator of polarized actin assembly and cell growth, increased levels of these cyclins may act through Cdc42 to restore growth to vhs2 mlf3 cells at 37°.
Here, we identified two pairs of structurally related genes, SKG6 and TOS2 and VHS2 and MLF3, as multicopy suppressors of the growth and actin organization defects of gic1 gic2 cells. One of our initial goals in performing a dosage suppression screen was to gain insights into how Gic1 and Gic2 regulate the actin cytoskeleton and polarized growth. Although our analyses have not yet clarified this mechanism, they demonstrate the involvement of both new gene pairs in cell functions that involve Gic1 and Gic2: cytokinesis (SKG6 and TOS2) and actin organization and polarized cell growth (VHS2 and MLF3). Below, we discuss each gene pair separately.
Skg6 and Tos2 help regulate cytokinesis and the switch from apical-to-isotropic cell growth:
All of our data characterizing Skg6 and Tos2 lead us to propose that they have an inhibitory function in cytokinesis, primarily via regulation of septin organization and/or Hof1 localization and, in turn, septum formation. This is supported by the relocalization of Skg6-GFP and Tos2-GFP from the bud tip to the bud neck preceding cytokinesis (Drees et al. 2001 and our unpublished data) and by the hyperelongated- and multiple-bud phenotype of cells overexpressing Skg6 or Tos2 (Figure 2A), their failure to undergo cell division (Figure 2C), and their mislocalization and/or disrupted organization of septins, Myo1, actin, and Hof1 at the bud neck (Figure 2D; Figure 3, A–C). We favor the model that Skg6 and Tos2 act through Hof1 and septins to affect cytokinesis rather than Myo1, because we observed no obvious defects in recruitment or closure of the actomyosin ring. Further, deletion of SKG6 suppressed the temperature-sensitive growth of hof1 cells (Figure 3, D and E) and septum formation was impaired in multibudded Skg6- or Tos2-overproducing cells (Figure 3F).
Previous studies have suggested that Gic1 and Gic2 may regulate cytokinesis on the basis of localization of Gic proteins at the bud neck in large-budded cells and the synthetic lethal interaction between gic1 gic2 and cla4 (Chen et al. 1997). Further, a recent study showed that Gic1 and Gic2 are required for the ability of Cdc42 to recruit and organize the septin collar at the bud neck (Iwase et al. 2005). The similar localization pattern of Skg6 and Tos2 to that of Gic1 and Gic2 and their co-implication in regulating cytokinesis may explain their identification as suppressors of gic1 gic2 mutations.
Although we have primarily focused on the functions of Skg6 and Tos2 in cytokinesis, it may not be their exclusive function, given their localization to sites of bud emergence and as a polarized cap in medium- to large-budded cells. Skg6 and Tos2 may have additional roles in polarized growth—possibly in regulating the switch from apical to isotropic cell growth, as inferred from the elongated bud phenotype caused by their overproduction (Figure 2A and Table 2). In fact, Toenjes et al. (2004) reported an elongated cell phenotype upon TOS2 overexpression and suggested that this phenotype may result from prolonged retention of Cdc24 and/or other polarity factors at sites of polarized growth.
Vhs2 and Mlf3 help regulate actin cytoskeleton organization, cell-wall integrity, and polarized cell growth:
Normal polarized morphogenesis in S. cerevisiae requires orientation of the actin cytoskeleton, which in turn targets secretory vesicles to deliver membrane and cell-wall synthesis/modifying enzymes toward regions of active growth. The importance of regulating this process is underscored by the fact that ∼27% of SBF target genes (those expressed during late G1 when cell polarity is established) are involved in cell-wall biogenesis/maintenance and/or polarized growth (Iyer et al. 2001). Our results have implicated Vhs2 and Mlf3 in this process. vhs2 mlf3 cells have a depolarized actin cytoskeleton and morphological phenotypes, indicating loss of polarity (Figures 5 and 4B, respectively). For example, vhs2 mlf3 cells become enriched in depolarized unbudded cells. This phenotype is not likely to result from cell arrest in G1 since we did not find any obvious defects in cell cycle progression. Instead, we suspect that vhs2 mlf3 cells may be impaired in apical polarization of the actin cytoskeleton. This is suggested further by our observation that small-budded vhs2 mlf3 cells are completely devoid of visible F-actin structures in the bud (Figure 5A, insert, and D). In addition, vhs2 mlf3 cells are defective in forming normal elongated mating projections (Figure 6). Thus, Vhs2 and Mlf3 clearly appear to have roles in regulating actin cytoskeleton organization and cell polarity. However, both of these proteins are unlikely to be stably associated with components of the actin cytoskeleton on the basis of their diffused cytoplasmic localization (Figure 4C).
Our data also demonstrate Vhs2 and Mlf3 involvement in regulating cell-wall integrity. However, three observations suggest that this function is likely to be indirect, stemming from the roles of Vhs2 and Mlf3 in regulating actin polarization. First, in vhs2 mlf3 cells the onset of actin polarization defects precedes the onset of cell lysis defects. Second, addition of sorbitol to the growth medium rescues the cell-wall defects but not the actin polarization defects of vhs2 mlf3 cells. Third, vhs2 mlf3 cells accumulate as large and round cells, whereas all known mutants of the cell integrity pathway lyse and die at the small-budded stage (Heinisch et al. 1999).
Vhs2 and Mlf3 functions in regulating actin organization are redundant with Gic1 and Gic2:
Although VHS2 and MLF3 overexpression suppresses the Ts− growth and actin polarization defects of gic1 gic2 cells, multiple lines of evidence presented here suggest that Vhs2 and Mlf3 function in a pathway that is distinct and parallel to Gic1 and Gic2:
We have demonstrated reciprocal high-copy suppression between these gene pairs; gic1 gic2 is suppressed by VHS2 and MLF3, and vhs2 mlf3 is suppressed by GIC1 (Figures 1 and 7A). This is not expected for genes that function in the same linear pathway.
The growth phenotype of gic1 gic2 cells is exacerbated by deletion of VHS2 and MLF3 (Figure 7B).
The Ts− growth defects of vhs2 mlf3 but not gic1 gic2 cells are rescued by sorbitol.
Haploid gic1 gic2 cells but not vhs2 mlf3 cells exhibit defects in bud-site selection.
Thus, all lines of available evidence point to Vhs2 and Mlf3 having functions that are separate from Gic1 and Gic2, operating in a parallel (and perhaps partially overlapping) pathway directing polarized growth. One of the only common phenotypes shared by gic1 gic2 and vhs2 mlf3 mutant cells is defective actin cytoskeleton polarization, and perhaps this point of similarity provides the basis for their reciprocal high-copy suppression.
To summarize, we have characterized the genetic and cellular roles of two pairs of proteins whose functions until now were unknown. We show that Skg6 and Tos2 regulate organization of proteins (septins, actin, and Hof1) at the bud neck to influence septum formation and cytokinesis, while Vhs2 and Mlf3 regulate polarization of the actin cytoskeleton and cell growth. A more precise understanding of the cellular functions of each of these proteins awaits identification of their in vivo ligands and determination of their biochemical activities. However, we have demonstrated the involvement of these proteins in specific physiological processes and defined their genetic interactions, providing a strong foundation for future analyses.
We are grateful to Guang-Chao Chen and Liaoteng Wang for the initial isolation of multicopy suppressors. We thank James Moseley and Avital Rodal for critical reading and editing of the manuscript. This work was supported by a National Institutes of Health (NIH) grant (GM63691) to B.L.G. and an NIH grant (GM45185) and a Texas Higher Education Coordinating Board advanced research program grant (003658-0427b) to C.S.M.C.
Communicating editor: M. D. Rose
- Received March 14, 2006.
- Accepted June 23, 2006.
- Copyright © 2006 by the Genetics Society of America