Saccharomyces cerevisiae whi2Δ cells are unable to halt cell division in response to nutrient limitation and are sensitive to a wide variety of stresses. A synthetic lethal screen resulted in the isolation of siw mutants that had a phenotype similar to that of whi2Δ. Among these were mutations affecting SIW14, FEN2, SLT2, and THR4. Fluid-phase endocytosis is severely reduced or abolished in whi2Δ, siw14Δ, fen2Δ, and thr4Δ mutants. Furthermore, whi2Δ and siw14Δ mutants produce large actin clumps in stationary phase similar to those seen in prk1Δ ark1Δ mutants defective in protein kinases that regulate the actin cytoskeleton. Overexpression of SIW14 in a prk1Δ strain resulted in a loss of cortical actin patches and cables and was lethal. Overexpression of SIW14 also rescued the caffeine sensitivity of the slt2 mutant isolated in the screen, but this was not due to alteration of the phosphorylation state of Slt2. These observations suggest that endocytosis and the organization of the actin cytoskeleton are required for the proper response to nutrient limitation. This hypothesis is supported by the observation that rvs161Δ, sla1Δ, sla2Δ, vrp1Δ, ypt51Δ, ypt52Δ, and end3Δ mutations, which disrupt the organization of the actin cytoskeleton and/or reduce endocytosis, have a phenotype similar to that of whi2Δ mutants.
TO achieve balanced growth and proliferation it is necessary that cells cease division when there are insufficient nutrients. When cells of the budding yeast Saccharomyces cerevisiae are starved of nutrients, they arrest in the G1 phase of the cell cycle in an unbudded, phase-bright state (Pringle and Hartwell 1981). A series of physiological changes occurs, which allows cells to survive adverse environmental conditions (Snow 1966; Schenberg-Frascino and Moustacchi 1972; Deutch and Parry 1973; Parryet al. 1976; Lillie and Pringle 1980). Cells carrying a whi2 mutation fail to show this response (Sudberyet al. 1980; Saulet al. 1985). As whi2Δ cells approach stationary phase, cell division continues beyond the point where it ceases in wild-type cells and is accompanied by prolonged expression of the G1 cyclins Cln1 and Cln2 (Radcliffeet al. 1997). As a result of continued division without growth, whi2Δ cells are abnormally small in stationary phase. Furthermore, they fail to acquire the stress resistance shown by wild-type cells (Saulet al. 1985). Exponentially growing whi2Δ cells also show increased sensitivity to certain environmental stresses, such as caffeine and 1.0 m NaCl (Binleyet al. 1999).
Recently it has been shown that Whi2 acts in the stress-response pathway (Kaidaet al. 2002). When whi2Δ cells are exposed to 37°, 0.4 m NaCl, or 0.4 mm H2O2, the expression of stress-response genes mediated by stress-response elements (STREs) in their promoters was reduced to 50% of wild-type levels (Kaidaet al. 2002). STRE-mediated gene expression is also induced as wild-type cells enter stationary phase, but induction is delayed by several hours in whi2Δ cells and cell division continues (Kaidaet al. 2002). Whi2 physically interacts with Psr1, one of a pair of redundant phosphatases located in the plasma membrane (Kaidaet al. 2002). A psr1Δ psr2Δ mutant has phenotypes similar to those of whi2Δ with respect to sensitivity to Na+ ions and small size in stationary phase. Furthermore, expression of STRE-mediated stress-response genes is also reduced in a psrΔ1 psr2Δ mutant. In addition to Psr1, Whi2 also physically interacts with Msn2, a transcription factor that plays a key role in the regulation of stress-response genes (Martinez-Pastoret al. 1996). Msn2 is hyperphosphorylated in whi2Δ or psr1 psr2 mutants (Kaidaet al. 2002) and overexpression of MSN2 rescues the heat sensitivity of whi2Δ or psr1 psr2 mutants (Kaidaet al. 2002). While Psr1 and Psr2 are cell surface proteins (Siniossoglouet al. 2000), Msn2 shuttles between the nucleus and cytoplasm (Gorneret al. 1998). Whi2 interacts with both Msn2 and Psr1/2 and is located throughout the cell (Kaidaet al. 2002). These observations suggest that Whi2 and Psr1/2 are functional partners that regulate the activity of Msn2 through its state of phosphorylation (Kaidaet al. 2002).
In this article, we describe the use of a colony-sectoring assay to isolate mutants that have a more severe effect on fitness when combined with a whi2Δ mutation. Seven independent mutants that affect a variety of cell functions were recovered. None of the mutations were strictly colethal with whi2Δ, so we have called the mutants siw (synthetic interaction with whi2Δ). We show that whi2Δ and several of the siw mutants have defects in actin organization and endocytosis. Further, we demonstrate a genetic interaction between SIW14 and ARK1 and PRK1, which encodes a pair of redundant protein kinases that regulate the stability of actin cortical patches. Finally, we show that mutations known to cause defects in the actin cytoskeleton and in endocytosis fail to show cell cycle arrest upon nutrient deprivation. Thus, we conclude that the normal function of the actin cytoskeleton and endocytosis are required to coordinate cell proliferation with nutrient availability.
MATERIALS AND METHODS
Strains and culture conditions: Strains used in this article are described in Table 1. They were routinely cultured at 30° on YEPD [1% Difco yeast extract (Becton Dickinson, Sparks, MD), 2% Difco peptone, plus 2% glucose]. Plates were solidified by the addition of 2% Difco agar. Cells were grown to stationary phase as follows: 2.5 ml of YEPD in a 50-ml conical flask was inoculated with yeast from a fresh YEPD plate culture and incubated for 48 hr at 26° on a rotary shaker at a speed of 150 rpm. By this time, wild-type cells have arrested as unbudded cells that have a bright appearance when examined by phase-contrast microscopy. The high ratio of flask size to culture volume is important for the development of the whi2Δ phenotype, which is manifested only under conditions of vigorous aeration (Rahmanet al. 1988). Cell volume was measured with a ZB1 Coulter Counter and Channelyser as described previously (Sudberyet al. 1980). Values reported are the median of the volume distribution.
Cloning and characterization of SIW genes: Because of the similarity of the siw mutants to mutants affecting the PKC1/SLT2 mitogen-activated protein (MAP) kinase pathway, siw mutants were transformed with plasmids carrying wild-type copies of all the genes in this pathway. The mutant originally designated siw9 was complemented by a centromeric plasmid carrying the SLT2 gene, suggesting that siw9 is an allele of SLT2. This was confirmed by the observation that siw9 failed to complement an slt2Δ mutation. This mutation is referred to as slt2siw9. The remaining mutants were transformed with a genomic library constructed in YCp50, a centromeric URA3 vector (Roseet al. 1987; purchased from the ATCC, http://www.atcc.org/). Ura+ transformants, representing approximately five-genome equivalents, were replica plated onto YEPD plates containing 5 mm caffeine, and transformants that grew were retained for further study. Normally the clones that grew produced red sectors, indicating that there was no longer selection against the loss of plasmid pPR14 (ADE2 WHI2) used in the colony-sectoring screen. The dependency of caffeine-resistant growth on the presence of a plasmid from the library was examined by growing transformants on media containing 5′-fluroorotic acid (5′-FOA) that selects for cells that have lost the YCp50-based plasmid containing the URA3 gene (Boekeet al. 1984). Colonies that grew on 5′-FOA were no longer able to grow on a YEPD plate containing 5 mm caffeine. Plasmids were recovered from independent colonies by transformation of Escherichia coli with total yeast DNA preparations. Recovered plasmids were transformed into the original siw mutant. Plasmids that both induced sectoring of pPR14 and conferred wild-type levels of caffeine resistance were retained for further study. Two primers (Table 2) were used to sequence ∼400 bp at either end of the genomic insert in the plasmids recovered. These sequences were compared to the S. cerevisiae genome database (http://genome-www.stanford.edu/Saccharomyces/) and the intervening sequence was retrieved. Subcloning was carried out using the derived restriction map of the insert to identify the minimum complementing region.
Gene deletions were carried out in two different strains: the Y763 strain used as the parent for the colony-sectoring assay (Costiganet al. 1992) and W303a, which is known to be ssd1-1d (Cvrckovaet al. 1995). Deletions were carried out as follows. FEN2 (SIW 1) was deleted using a URA3-based disruption cassette kindly supplied by G. Lucchini. The deletion was verified by Southern hybridization. For ALG9 (SIW5), ZDS1 (SIW7), THR4 (SIW12), and SIW14 (YNLO32w), pairs of PCR primers were designed (Table 2) to amplify a URA3 template so that the resulting DNA molecule consisted of the URA3 gene flanked by 45 bp of DNA homologous to the sequence immediately upstream of the start codon and downstream of the stop codon of the gene to be deleted. In the case of SIW14 (YNLO32w), BamHI and HindIII restriction sites were incorporated into the primers between the YNLO32w and URA3 sequences to aid subsequent Southern analysis of putative disruptants. PCR amplification of the URA3 template was performed for 35 cycles with an annealing temperature of 57° and the resulting PCR fragment was used to transform yeast strains Y763 and W303a, both of which are ura3. Deletion of the target open reading frame was verified by Southern hybridization or PCR analysis.
In all cases, the deleted strain was more sensitive to caffeine than was the parent. A diploid was formed by crossing the disruptant and the original mutant. Allelism was demonstrated by noncomplementation of mutant phenotypes in the diploid and 4:0 segregation of caffeine sensitivity in tetrads dissected after sporulation. The deleted strain was also crossed to a strain with the wild-type SIW allele and tetrads were dissected. In all cases, caffeine sensitivity cosegregated with the auxotrophic marker used to engineer the deletion.
Characterization of the slt2siw9 mutation: PCR primers (Table 2) were designed to amplify the locus from the slt2siw9 mutant in two segments with an overlapping region spanning the SacI site at nucleotide 600, each fragment containing a BglII or HindIII site present in the respective 5′ and 3′ flanking regions. The resulting PCR products were digested with either BglII and SacI (5′ fragment) or HindIII and SacI (3′ fragment) and ligated to pUC19 digested with HindIII and BglII. The full-length sequence of the resulting slt2siw9 clone was determined and compared to the wild-type sequence.
Mutagenesis of SIW14: Mutagenesis of SIW14 cloned with its own promoter into the multicopy plasmid pYES2 (Invitrogen, San Diego) was carried out using a QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). Oligonucleotides used are listed in Table 2. Following mutagenesis, the entire SIW14 gene was resequenced for each mutant allele generated.
Western blotting: Cells were disrupted with glass beads and total soluble protein was prepared as described (Boyneet al. 2000). Proteins were fractionated by SDS/polyacrylamide gel electrophoresis and electroblotted to a Hybond C nitrocellulose membrane (Amersham Pharmacia Biotech, Little Chalfont, UK). Primary antibody was rabbit polyclonal anti-phospho-p44/42 MAP kinase (Thr202/Tyr204; Cell Signaling Technology, Beverly, MA) used at 1:1000 dilution. Two rounds of antibody binding were used to amplify the weak signal that resulted from this primary antibody. Secondary antibody was mouse anti-rabbit immunoglobulins (Jackson Immunoresearch, Cambridge, UK) used at 1:3500 dilution. Tertiary antibody was goat anti-mouse immunoglobulins conjugated to horseradish peroxidase (Dako A/S, Glostrup, Denmark) diluted 1:5000. Binding of tertiary antibody was visualized by enhanced chemiluminescence (Amersham Pharmacia Biotech) and recorded by a Gene Gnome chemiluminescence recording system (Syngene, Cambridge, UK).
Glycogen staining: Cells from 1 ml of stationary-phase cultures were collected by centrifugation and resuspended in 1 ml of 0.2% w/v iodine and 0.4% w/v potassium iodide. After 5 min, cells were collected by centrifugation and washed 3× in distilled water and finally resuspended in 0.2 ml of water. Each sample was placed in a well of a 96-well microtiter plate and photographed with transmitted light. Cells producing glycogen stained dark purple, while cells not producing glycogen remained light brown.
Actin staining: Cultures were stained with phalloidin conjugated to tetramethylrhodaminyl-isothiocyanate (TRITC) conjugated (Sigma, St. Louis) as previously described (Adams and Pringle 1983).
Heat shock: Stationary-phase cultures of 500 μl were heated using a Hybaid thermocyler for 10 min. Samples were diluted in triplicate and plated on YEPD and the mean number of colonies that grew after 3 days was recorded. For control cultures, unheated samples of the same cultures were diluted and plated in triplicate. Viability is expressed as percentage viability of the heated cultures relative to control.
Plate dilution tests: Cells were harvested from a freshly grown YEPD plate and resuspended in YEPD to an OD600 of 8.0. This suspension was then serially diluted and 5 μl of the undiluted and 10-1, 10-2, 10-3, and 10-4 dilutions were spotted onto the surface of the agar plate and cultures were incubated for 3 days at 30°. These dilutions resulted in ∼20 cells in the spot from the most diluted suspension. All growth tests were carried out using YEPD media supplemented as indicated except for the 37° growth test in which minimal medium was used.
Lucifer yellow uptake assays: Lucifer yellow assays were carried out as described (Dulicet al. 1991).
Immunocytofluorescence: Immunocytofluorescence was carried out as described (Ayscough and Drubin 1998). Mouse anti-actin antibody was purchased from Sigma and used at a dilution of 1:200. Rabbit polyclonal anti-Cdc11 was purchased from Santa Cruz and used at 1:250 dilution. Mouse polyclonal antisera to Cof1, Sac6, and Abp1 were a kind gift from Kathryn Ayscough.
Microscopy: For differential interference contrast (DIC) and fluorescence microscopy, cells were examined with a Leica DMLB fluorescence microscope. Digital images were recorded by a low-temperature CCD camera (model RTE Princeton Instruments, Princeton, NJ) controlled by an Apple Macintosh G4 computer running Open Lab software, version 2.2.5 (Improvison, Warwick, UK). Images were exported as tif files and edited for contrast and brightness in Adobe Photoshop, version 5.5. Composite figures were assembled using Microsoft Power Point 2000.
A colony-sectoring assay identifies seven genes that interact with whi2Δ: To identify genes that may interact with WHI2, we carried out a colony-sectoring assay for mutations that either are colethal or show an enhanced growth defect with the whi2Δ allele. For this screen, we applied the methodology used by Costigan and Snyder (Costiganet al. 1992). Briefly, the screen is based on a whi2Δ ade2-1 host that harbors an unstable plasmid carrying a wild-type copy of WHI2. Mutations that have a more severe phenotype in conjunction with whi2Δ result in selection against loss of the WHI2 plasmid and therefore produce homogeneous white colonies that lack red sectors. In this way we isolated seven mutants, which subsequent analysis showed affected the following genes: ALG9, FEN2, PRS3, SIW14, SLT2 (MPK1), THR4, and ZDS1 (see materials and methods).
SIW14 encodes a protein that contains the sequence IHCNRGKHRTGCL, which is an exact match to the canonical sequence for tyrosine phosphatases, [LIVMF]HC xxGxxx[STC][STAG]x[LIVMFY], listed in the Prosite database (http://ca.expasy.org/prosite/). SLT2 (MPK1) encodes the MAP kinase of the PKC1 MAP kinase pathway (Torreset al. 1991; Leeet al. 1993). We refer to the allele isolated in our screen as slt2siw9. We cloned slt2siw9 by PCR and showed that the mutation was E225K, an amino acid substitution in the conserved region IX of the kinase domain. FEN2 was first identified in a screen for mutants resistant to fenproprimorph, an inhibitor of ergosterol biosynthesis (Marcireauet al. 1996) and it was subsequently shown that Fen2 is the membrane pantothenate transporter (Stolz and Sauer 1999). PRS3 is one of a family of five genes that encodes phosphoribosyl pyrophosphate synthetase, an enzyme that catalyzes the first step in a variety of biosynthetic pathways, including amino acid and nucleotide biosynthesis (Carter et al. 1994, 1995). We have previously published a detailed characterization of prs3Δ (Binleyet al. 1999). ALG9 encodes a mannosyl transferase that transfers core oligosaccharides from the lipid carrier dolichol pyrophosphate to the asparagine residues of secreted proteins in the endoplasmic reticulum (Burdaet al. 1996). THR4 encodes threonine synthetase, required for threonine biosynthesis. ZDS1 has been isolated in many genetic screens, hence its name, which is derived from “zillion different screens” (Bi and Pringle 1996; Yuet al. 1996). Deletion of both ZDS1 and ZDS2 genes results in a mitotic delay and elongated buds with constrictions where cytokinesis has been attempted but not completed (Yuet al. 1996; Mizunumaet al. 1998).
Deletion phenotype of genes identified in the colony-sectoring assay: ALG9, FEN2, PRS3, SIW14, THR4, and ZDS1 were deleted in two strain backgrounds: Y763, the host strain for the synthetic lethal screen, and W303a, a strain carrying the ssd1-1d allele in which we found that the stress sensitivity of the whi2Δ phenotype is enhanced (see Table 3). Strains with a slt2Δ allele in two different backgrounds were obtained elsewhere (Table 1). We carried out appropriate crosses to congenic strains containing a whi2Δ allele to determine whether the deletion alleles were colethal with whi2Δ. In each case, viable double mutants could be constructed, showing that the mutations were not colethal with whi2Δ. However, upon reintroduction of the unstable WHI2-bearing plasmid into each of the double mutants, it was either absolutely stable or lost at a very low frequency compared to a wild-type host. Often, where a red sector appeared it clearly grew more slowly than the rest of the colony (data not shown). Moreover, in some cases enhanced phenotypes were observed in the double mutant. For example, a fen2Δ whi2Δ mutant was unable to grow on minimal medium and an slt2Δ whi2Δ strain had a grossly abnormal morphology (data not shown). Thus, although not colethal, these mutations have a more severe phenotype when combined with whi2Δ. For this reason we designated these mutants siw (synthetic interaction with whi2Δ).
Most of the siw mutants showed stress sensitivities that were similar to those of a whi2Δ strain. Most were hypersensitive to caffeine and 1 m NaCl (Table 3). In addition, alg9Δ, slt2siw9, and zds1Δ mutants were temperature sensitive on minimal medium; fen2Δ and slt2siw9 mutants were hypersensitive to 0.7 m CaCl2; the fen2Δ mutant was sensitive to Calcofluor white, a characteristic of cell wall mutants; siw14Δ and slt2siw9 were unable to utilize glycerol; and, finally, siw14Δ cells were resistant to 6.6 mm MnCl2, which inhibits the growth of wild-type cells.
When grown to stationary phase in YEPD, siw14Δ, fen2Δ, and slt2Δ arrested as phase-dark, budded cells, which failed to accumulate glycogen (Figure 1). Moreover, siw14Δ cells were hypersensitive to a 53° heat shock (Table 3). Thus, like whi2, these mutants failed to have the normal physiological responses to nutrient limitation. There was no reduction in cell size. However, many of the mutants showed very large vacuoles and it is possible that the cytoplasmic mass is less than that of a wild-type cell. This is particularly marked in the case of the slt2Δ mutant where, interestingly, some very small cells are evident (arrows in Figure 1). In contrast, alg9Δ, thr4Δ, and zds1Δ mutants showed the normal responses to nutrient limitation (data not shown). However, the alg9 and zds1 alleles isolated in the colony-sectoring screen did arrest as phase-dark budded cells in stationary phase (data not shown). Thus, it is possible that the mutant alleles isolated in the screen have a different phenotype from that of the deletion alleles. However, the genetic complexity of the original mutants makes such a conclusion uncertain and we did not pursue this matter further.
Cytoskeletal abnormalities in stationary-phase whi2Δ and siw14Δ cells: We previously reported that whi2Δ cells contain abnormal actin clumps in stationary phase (Binleyet al. 1999). We examined the siw mutants for any abnormalities in the actin cytoskeleton. While the actin distribution of growing cells was normal, we observed that stationary-phase siw14Δ cells had a single intensely staining clump of actin in >95% of the cells examined (Figure 2A). The clumps also contained other actin-associated proteins such as Abp1, Sac6, and Cof1 (Figure 2B). This abnormality was present in all strain backgrounds containing the whi2Δ and siw14Δ mutations. We also observed that siw14Δ stationary-phase cells contained abnormal clumps of the septin Cdc11, which is normally located in a ring structure at the bud neck (Figure 2C).
Endocytosis is defective in whi2Δ, fen2Δ, thr4Δ, and siw14Δ mutants: Defects in the cortical actin cytoskeleton are often associated with defects in endocytosis. To determine if fluid-phase endocytosis was functioning normally in whi2Δ and siw mutants, we carried out lucifer yellow uptake assays. The whi2Δ, fen2Δ, thr4Δ, and siw14Δ mutants were defective in fluid-phase endocytosis (Figure 3). The defect in fen2Δ mutants was particularly profound.
SIW14 interacts with ARK1 and PRK1: Ark1 and Prk1 are two homologous protein kinases that are thought to regulate the association of the actin cortical patch complex with the endocytic machinery by phosphorylating Pan1 and Sla1 (Copeet al. 1999; Zeng and Cai 1999; Zenget al. 2001). The actin clumps in whi2Δ and siw14Δ mutants are strikingly similar to the actin clumps seen in ark1Δ prk2Δ cells (Copeet al. 1999). Because of this phenotypic similarity, we searched for interactions between ARK1 and PRK1 with WHI2 and SIW14. We found that overexpression of SIW14 from a GAL1 promoter on a multicopy plasmid was lethal in a prk1Δ strain and reduced growth in an ark1Δ strain (Figure 4A). A high proportion of cells with no visible actin patches or cables accumulated within 4 hr upon induction of SIW14 in a prk1Δ strain (Figure 4B).
Mutants in endocytosis and the actin cytoskeleton are defective in cell cycle arrest in stationary phase: The inability of whi2Δ and the siw mutants to arrest the cell cycle upon nutrient limitation could be a consequence of the defects in the actin cytoskeleton and endocytosis. To test this hypothesis, we determined whether mutations in genes known to function in endocytosis and the actin cytoskeleton have defects in cell cycle arrest. We monitored the effect on cell size and appearance in the stationary phase of deletion alleles of the following genes that have defects in actin organization or are listed in the Saccharomyces Genome Database as having defects in endocytosis: AKR1, ARK1, ARL1, CLC1, DNM1, END3, ENT1, ENT2, PKH1, PRK1, SWA2, RVS161, RVS167, SLA1, SLA2, TLG2, VAN1, VPS 4, VPS34, VRP1, YPK1, YPK2, VPS21 (YPT51), YPT52, and YPT53. Of these, we found that end3Δ, rvs161Δ, sla1Δ, vrp1Δ, sla2Δ, vps21Δ (ypt51Δ), and ypt52Δ arrested in stationary phase as phase-dark budded cells, instead of the phase-bright unbudded appearance of wild-type cells (Figure 5). Furthermore, vrp1Δ and rvs161Δ cells were much smaller than wild-type cells: 52 and 65%, respectively, of the size of wild-type cells. The ark1Δ and prk1Δ mutants showed some indication that cell cycle arrest was not normal, but this was not so marked and an ark1Δ prk1Δ double mutant did not show a more severe defect in nutrient arrest (data not shown). All the other mutants listed had a similar appearance and size to wild-type cells except for the akrΔ mutant, which as noted before has an unusual “peanut” shape (Pryciak and Hartwell 1996).
We examined end3Δ, rvs161Δ, sla1Δ, sla2Δ, vrp1Δ, vps21Δ (ypt51Δ), and ypt52Δ mutants to see if they also showed clumps of actin similar to that seen in whi2Δ and siw14Δ cells. Figure 6 shows that end3Δ, sla1Δ, prk1Δ, and ypt52Δ did indeed show such clumps. Thus, this type of actin disorganization is a feature of cells that are unable to respond appropriately to nutrient deprivation. Clumps of actin in sla1Δ mutants have been reported previously (Holtzmanet al. 1993).
SIW14 shows a genetic interaction with SLT2: The phenotypes of whi2 and siw mutants and mutants affecting the PKC1/Slt2 MAP kinase share the following phenotypic abnormalities: (a) they are sensitive to caffeine and other stresses, and these sensitivities are rescued by 1 m sorbitol and exacerbated by a ssd1-1d allele; (b) they fail to show a normal response to nutrient limitation; and (c) they display defects in the actin cytoskeleton, which in slt2Δ mutants is marked by loss of polarity and the appearance of actin bars in the cytoplasm (Costigan et al. 1992, 1994; Mazzoniet al. 1993). Moreover, we isolated an allele of SLT2 in the synthetic lethal screen. We therefore searched for interactions between the siw mutants and Slt2. We found that SIW14 on a multicopy plasmid could rescue the caffeine (Figure 7), but not the temperature sensitivity of the slt2siw9 mutation (data not shown). The minimum inhibitory concentration of caffeine of slt2siw9 is 4 mm, whereas that of a congenic slt2Δ strain is 2 mm. It is therefore likely that the slt2siw9 kinase retains some activity. Interestingly, multicopy SIW14 exacerbated the caffeine sensitivity of the congenic slt2Δ strain, so that growth on 1 mm caffeine was prevented (Figure 7). Thus, multicopy SIW14 interacts with SLT2 in an allele-specific fashion: it rescues the caffeine sensitivity of the slt2siw9 allele but enhances the caffeine sensitivity of an slt2Δ allele. To determine whether the putative tyrosine phosphatase active site was required for the interaction, we generated the following point mutations in three conserved residues in the putative tyrosine phosphatase active site: C214S, G217A, and R220K. The C214S mutation changes a critical cysteine residue that participates in catalysis and is known to abolish tyrosine phosphatase activity (Johnsonet al. 1992). Multicopy plasmids carrying these catalytically dead versions of Siw14 were unable to rescue the caffeine sensitivity of slt2siw9 and failed to enhance the caffeine sensitivity of slt2Δ (Figure 7). Thus, the interaction between Siw14 and Slt2 depends on a functioning tyrosine phosphatase active site. Surprisingly, the mutants were still able to rescue the caffeine sensitivity of an siw14Δ mutation (Figure 7), indicating that the tyrosine phosphatase active site is not required for at least some of the normal functions of Siw14.
Slt2 is activated by dual phosphorylation on threonine 190 and tyrosine 192 (Leeet al. 1993). Since Siw14 is a putative tyrosine phosphatase, it is possible that multicopy SIW14 affects the phosphorylation state of Slt2. To determine if this were the case, we used an antibody specific to the activated form of Slt2 to monitor the phosphorylation state of Slt2 (Martinet al. 2000). Phosphorylation of Slt2 increased upon exposure to caffeine as reported previously (Figure 8; Martinet al. 2000). Although the level of activating phosphorylation was reproducibly reduced in an slt2siw9 mutant upon caffeine exposure, perhaps explaining its reduced activity, the level of phosphorylation was unaffected by multicopy SIW14. Furthermore, the level of phosphorylation of wild-type Slt2 was also unaffected by deletion or overexpression of SIW14. Thus, Siw14 does not act directly to dephosphorylate Slt2.
Whi2Δ cells are defective in the organization of the actin cytoskeleton and endocytosis: We have shown here that the effect of a whi2Δ allele is pleiotropic. In addition to the failure to respond to nutrient limitation by ceasing cell division, whi2Δ cells are sensitive to a wide range of stresses, have a disorganized actin cytoskeleton upon nutrient limitation, and are defective in fluid-phase endocytosis. Overexpression of WHI2 has been reported to rescue the stress sensitivity of an rsp5 mutation (Kaidaet al. 2002). Rsp5 is a ubiquitin protein ligase that is thought to regulate the endocytic machinery and is required for normal stress response (reviewed in Rotinet al. 2000). It is responsible for the mono-ubiquitination and consequent internalization of many membrane proteins such as permeases, transporters, and receptors, including the Ste2 α-factor receptor. Rsp5 interacts with Pan1, a protein that links the endocytic machinery to the plasma membrane (Zoladeket al. 1997), and its location to the plasma membrane and perivacuolar structures is Sla2 dependent (Wanget al. 2001). The suppression of the rsp5 phenotype by WHI2 overexpression suggests that Whi2 may be a positive regulator of endocytosis, which would be consistent with the defects in endocytosis reported here in a whi2Δ mutant.
A synthetic lethal screen isolated further mutants with the same pleiotropic defects as whi2Δ: Using a colony-sectoring assay, we isolated a number of other mutants that had some or all of the phenotypes of a whi2Δ mutant. Of these prs3Δ, fen2Δ, and siw14Δ were particularly interesting since they showed sensitivity to stresses such as caffeine and 1 m NaCl and failed to respond by halting cell proliferation in response to nutrient depletion.
The amino acid sequence of Siw14 suggests that it is a member of a small family of tyrosine phosphatases. There are two homologs of SIW14 in the yeast genome: OCA1 (33% identity, 51% similarity) and YNL056w (35% identity and 58% similarity). Oca1 is required for cell cycle arrest in response to acid linoleic hydroperoxide, a lipid peroxidation product that accumulates during oxygen stress (Alicet al. 2001), so other members of this family are also concerned with stress response. We mutated the active site of Siw14 and found that catalytically dead alleles still complement the caffeine sensitivity of the siw14Δ mutation. Clearly, Siw14 has a function that is not dependent on tyrosine phosphatase activity. However, it is likely that Siw14 does have tyrosine phosphatase activity, because the catalytically dead mutants are unable to suppress the caffeine sensitivity of the slt2siw9 allele or to enhance the sensitivity of the slt2Δ allele. Thus, Siw14 appears to be a bifunctional protein.
The inability of slt2Δ mutants to arrest upon nutrient deprivation and an abnormal distribution of actin in slt2Δ cells has been reported previously (Mazzoniet al. 1993; Costigan and Snyder 1994). Null alleles of SIW14 also show cytoskeletal defects. When depleted of nutrients, siw14Δ cells contain a single large clump of actin and associated proteins such as Abp1, Cof1, and Sac6. Furthermore, the septin cytoskeleton shows a similar phenotype, as Cdc11 accumulates in large bars. This phenotype is strikingly similar to that of an ark1Δ prk1Δ mutant. Ark1 and Prk1 are two protein kinases that regulate the actin cytoskeleton (Copeet al. 1999; Zeng and Cai 1999; Zenget al. 2001). In an ark1Δ prk1Δ double mutant, actin was found in large clumps, similar to the appearance of actin clumps that we observe in stationary-phase whi2Δ and siw14Δ cells (Copeet al. 1999). Overexpression of either Prk1 or Ark1 results in a different type of actin abnormality: the formation of large cytoplasmic bars. We showed that overexpression of SIW14 in a prk1Δ results in an apparent loss of all forms of filamentous actin and is lethal. The similar actin cytoskeletal defects in both siw14Δ and ark1Δ prk1Δ mutants as well as the genetic interaction between prk1Δ and SIW14 overexpression suggest that SIW14 may influence budding during nutrient deprivation through the actin cytoskeleton.
The actin cytoskeleton participates in endocytosis, and siw14Δ mutants are defective in fluid-phase endocytosis. Siw14 has been shown to physically interact with Ypt53, a Rab5-like GTPase required for vesicle-mediated transport (Uetzet al. 2000). YPT53 shares homology with YPT52 and VPS21 (YPT51). It is thought that Vps21, Ypt52, and Ypt53 are required for the early steps in endocytosis (Singer-Krugeret al. 1994; Prescianotto-Baschong and Riezman 2002). A null allele of VPS21 causes an impairment of endocytosis that is more severe when combined with ypt52Δ and ypt53Δ alleles. We found that, like siw14Δ mutants, vps21Δ and ypt52Δ mutants upon nutrient limitation show a failure to arrest cells and have a disorganized actin cytoskeleton. We also found that vps21Δ and ypt52Δ mutants are sensitive to caffeine and 1 m NaCl and display actin clumps in stationary phase. We also examined a ypt53Δ strain but failed to see any phenotypic abnormalities. This may be because the function of Ypt53 is redundant with Vps21 and Ypt52. However, taken together, these observations support the conclusion that a normal response to nutrient and other stresses involves actin organization and endocytosis.
FEN2 was first identified in a search for mutants resistant to fenproprimorph, an inhibitor of ergosterol biosynthesis (Marcireauet al. 1996). Cells harboring a fen2Δ mutation were shown to have a threefold decrease in ergosterol levels. Subsequently, it was shown that Fen2 is a membrane pantothenate transporter (Stolz and Sauer 1999). Pantothenate is essential for the biosynthesis of coenzyme A, which is a carrier of activated C2 units in sterol biosynthesis. Thus, the reduction in ergosterol levels in a fen2 mutant is due to reduced availability of pantothenate. Ergosterol is required for the proper fluidity and function of cell membranes and for endocytosis (Hongayet al. 2002). Thus, the reduction in endocytosis in a fen2 mutant that we observed can be satisfactorily explained by the known function of Fen2.
We isolated a thr4 mutation in the synthetic lethal screen and we found that a thr4Δ mutant was sensitive to 1 m NaCl and was defective in endocytosis. The involvement of threonine synthetase was initially puzzling. However, a thr4Δ mutation was recovered in a screen for mutations that are colethal with sec13Δ, which is defective in the transport of the general amino acid permease from the Golgi to the cell surface (Roberget al. 1997). Thus, it is possible that Thr4 does have an unexpected role in protein trafficking.
Mutations known to affect the organization of the actin cytoskeleton and reduce endocytosis decouple growth and cell division: Although the molecular functions of the Whi2, Siw14, Fen2, and other Siw genes are diverse, they share the common property that mutations result in a disorganized actin cytoskeleton and/or reduce endocytosis. This observation led us to examine the phenotype of other mutants known to be defective in the organization of the actin cytoskeleton and endocytosis. Many mutants defective in endocytosis did not show any obvious deficiencies in response to nutrient stress. However, we found that rvs161Δ, sla1Δ, sla2Δ, vrp1Δ, vps21Δ, ypt52Δ, ypt53Δ, and end3Δ had the same phenotype as the original whi2Δ mutation. The phenotypes of vrp1Δ and rvs161Δ were particularly strong, resulting in a marked reduction in cell size compared to wild-type cells. It is not clear why only some mutants studied here show a reduction in cell size. However, one important point to be considered is that we have measured cell volume, which includes both the cytoplasmic and vacuolar compartments. Many of the mutants clearly result in enlarged vacuoles, which could be masking a reduction in the size of the cytoplasmic compartment. It is also important to note that the sizes measured here refer to stationary-phase cells. We observed no size reduction in exponentially growing cells.
Recently, two systematic surveys have yeast deletion sets to identify genes that regulate cell size in S. cerevisiae. The work of Jorgensen et al. (2002) focused on exponentially growing cells and thus is not directly comparable to the experiments reported here. Zhang et al. (2002) initially screened the deletion collection for mutants that had an abnormally small cell size in stationary-phase cultures and then tested for size abnormalities in exponentially growing cells. This work identified whi2Δ as one of the 20 mutations that caused abnormally small size in stationary phase. None of the other mutations identified here that cause a reduction in cell size were present in the list of Zhang et al. (2002). However, FEN2 (YCR028c) was listed in a table of mitochondrial mutants that showed a reduction in cell size. The criterion for mitochondrial mutations was a failure to grow on glycerol. As discussed above, the molecular defect has been identified in fen2 and it seems likely that it is misclassified as a mitochondrial mutant. VPR1 was also present in a list of genes whose deletion resulted in small cell size, but that for various reasons failed a quality control test. The rvs161Δ allele also showed a size reduction, but was also omitted for quality control reasons (B. Schneider, personal communication). Thus, the results presented here are broadly consistent with the results of the systematic gene deletion survey.
The proteins that are required for the proper nutrient response function either in the first steps of endocytosis or in the actin complex that mediates endocytosis. Taken together, these observations suggest that the actin/endocytosis complex is required for cells to cease cell division in response to nutrient stress. We suggest two possible explanations. First, endocytosis could result in the internalization of membrane proteins required for growth and budding. Second, fluid-phase endocytosis could be required for the uptake of small molecules from the medium used to sense cell density. Such molecules may accumulate to form the signal to cease cell division in the high cell densities found in stationary-phase culture. A precedent for this may be the way that Candida albicans cells undergo only the yeast-to-hypha transition at low cell densities. It has recently been shown that the accumulation of farnesol provides the signal that is used by C. albicans cells to sense cell density (Hornbyet al. 2001; Ohet al. 2001).
We thank Mike Snyder, Kevin Costigan, Kathryn Ayscough, Mike Tyers, and Giovanna Lucchini for strains and plasmids and Kathryn Ayscough for critical reading of the manuscript. This work was supported by a project grant from the Biotechnology and Biological Sciences Research Council (BBSRC). A.C. received a BBSRC Cooperative Award in Science and Engineering studentship. K.B. and P.R. were supported by BBSRC research training studentships.
Communicating editor: B. Andrews
- Received July 28, 2003.
- Accepted October 27, 2003.
- Copyright © 2004 by the Genetics Society of America