In the absence of a successful mating, pheromone-arrested Saccharomyces cerevisiae cells reenter the mitotic cycle through a recovery process that involves downregulation of the mating mitogen-activated protein kinase (MAPK) cascade. We have isolated a novel gene, POG1, whose promotion of recovery parallels that of the MAPK phosphatase Msg5. POG1 confers α-factor resistance when overexpressed and enhances α-factor sensitivity when deleted in the background of an msg5 mutant. Overexpression of POG1 inhibits α-factor-induced G1 arrest and transcriptional repression of the CLN1 and CLN2 genes. The block in transcriptional repression occurs at SCB/MCB promoter elements by a mechanism that requires Bck1 but not Cln3. Genetic tests strongly argue that POG1 promotes recovery through upregulation of the CLN2 gene and that the resulting Cln2 protein promotes recovery primarily through an effect on Ste20, an activator of the mating MAPK cascade. A pog1 cln3 double mutant displays synthetic mutant phenotypes shared by cell-wall integrity and actin cytoskeleton mutants, with no synthetic defect in the expression of CLN1 or CLN2. These and other results suggest that POG1 may regulate additional genes during vegetative growth and recovery.
THE yeast Saccharomyces cerevisiae has a and α haploid cells that mate to produce a/α diploids (Sprague and Thorner 1992). The haploids secrete pheromones, a-factor and α-factor, that act on haploid cells of the opposite mating type. The pheromone binds to and activates the Ste2 receptor in a cells and the Ste3 receptor in α cells. Receptor activation turns on a signal transduction cascade that induces the transcription of a number of genes involved in mating and causes cells to undergo cell-cycle arrest in G1 phase and morphological changes (termed shmoo formation).
The receptors transduce the pheromone signal to a heterotrimeric G-protein consisting of the Gpa1 (Gα), Ste4 (Gβ), and Ste18 (Gγ) subunits (Dietzel and Kurjan 1987; Miyajimaet al. 1987; Whitewayet al. 1989). In the absence of pheromone, Gpa1 binds to and maintains Ste4/Ste18 (Gβγ) in an inactive state. Upon activation of the receptor, Gpa1 is released from Ste4/Ste18, allowing Ste4 to transduce the signal to a highly conserved mitogen-activated protein kinase (MAPK) cascade. Ste4 activates the MAPK cascade by binding to Ste20 (Leeuwet al. 1998), a Cdc42-activated kinase (Simonet al. 1995; Peteret al. 1996; Lebereret al. 1997), and to Ste5, a LIM/Ring-H2 domain protein (Inouyeet al. 1997; Fenget al. 1998). Ste5 acts as a scaffold (Elion 1995) for Ste11 (a MAPK kinase kinase), Ste7 (a MAPK kinase), and Fus3/Kss1 (MAPKs, of which Fus3 is most critical for mating; Elionet al. 1991; Madhaniet al. 1997). Ste20 activates Ste11 through an unknown mechanism that requires the presence of Ste5 (Lebereret al. 1992; Wuet al. 1995; Fenget al. 1998).
Once activated, the MAPKs act on a transcription factor, Ste12 (Elion et al. 1991, 1993), which induces the expression of numerous genes required for signal transduction, G1 arrest, and mating (Songet al. 1991). In addition, Fus3 activates a cyclin-dependent kinase inhibitor, Far1, which inhibits the activity of the G1 cyclin-dependent kinase, Cdc28 (Chang and Herskowitz 1990; Peter and Herskowitz 1994; Jeounget al. 1998). Fus3 and Kss1 also repress the transcription of the G1 cyclin genes, CLN1 and CLN2 (Cherkasovaet al. 1999). The combination of transcriptional repression of cyclin genes and inactivation of the Cln/Cdc28 complexes leads to arrest in the G1 phase of the cell cycle (Valdiviesoet al. 1993; Cherkasovaet al. 1999).
The response elicited by pheromone is transient. In the absence of mating, cells reenter the cell cycle through a process of recovery or desensitization (Sprague and Thorner 1992). Regulators of recovery include Gpa1 and Sst2, which inhibit the activity of Ste4/Ste18 (Dietzel and Kurjan 1987; Dohlman and Thorner 1997) and phosphatases such as Msg5, Ptp2, and Ptp3, which inactivate Fus3 (Doiet al. 1994; Zhanet al. 1997). In addition to driving the G1- to S-phase transition (Hadwigeret al. 1989), the G1 cyclin Cln2 may also have a function in recovery because overexpression of CLN2 blocks the ability of cells to arrest in the presence of α-factor (Oehlen and Cross 1994). Cln2 overproduction inhibits the mating MAPK pathway between the Ste4 and Ste11 steps, suggesting that Ste5, Ste20, or Ste11 may be direct targets of Cln2/Cdc28 (Wassmann and Ammerer 1997).
To identify other components involved in the recovery process, we isolated genes that block pheromone-induced G1 arrest when overexpressed. Among the genes isolated, we found a novel gene, POG1. Double mutant analysis suggests that POG1's promotion of recovery parallels that of MSG5. POG1 requires CLN2 but not the other G1 cyclins to promote recovery. Consistent with this, POG1 overexpression leads to elevated levels of CLN1 and CLN2 mRNAs in the presence of α-factor. This loss of transcriptional repression occurs through SCB/MCB promoter elements and requires Bck1, a MAPK kinase kinase known to upregulate Swi4-dependent cell-cycle box (SCB)/MluI cell-cycle box (MCB) promoter elements during vegetative growth (Maddenet al. 1997). Additional genetic evidence suggests that STE20 may be a key target of control for the promotion of recovery of POG1 and CLN2. Finally, POG1 has a vegetative function that may be redundant with CLN3 and distinct from its ability to regulate the CLN1 and CLN2 genes.
MATERIALS AND METHODS
Yeast strains, media, and genetic manipulation: Yeast strains and plasmids are listed in Table 1. Standard methods were used for microbial and molecular manipulations (Guthrie and Fink 1991). msg5::Leu2 from pSPdel was introduced as described (Doiet al. 1994). hmlα::LEU2 from pCW9-1 (provided by C. White, Frederick Cancer Institute, Frederick, Maryland) was introduced as a BamHI fragment. All strain constructions by gene replacement were confirmed by Southern analysis (Sambrooket al. 1989).
cDNA library screen: EY1118 cells were transformed with a yeast cDNA plasmid library that expresses cDNA inserts from the GAL1 promoter (Liuet al. 1992). Ura+ transformants were first selected on glucose-uracil plates. A total of 80,000 colonies were then screened for α-factor resistance by replica plating them onto galactose-uracil plates spread with 4.2 μgof α-factor. In a secondary screen, α-factor-resistant transformants were retested for dependence on galactose for growth in the presence of α-factor. In a tertiary screen, transformants were passaged over 5-fluoroorotic acid + uracil plates to select for loss of the plasmid DNA to confirm the dependence of α-factor resistance on the presence of the plasmid. Positive plasmids were then rescued from yeast (Hoffmann and Winston 1987) and retransformed into EY1118 to confirm galactose-dependent growth in the presence of α-factor.
Recombinant DNA techniques: Standard methods were used for all recombinant DNA techniques (Sambrooket al. 1989). Plasmids were transformed into Escherichia coli HB101. DNA sequencing was performed by the dideoxy chain termination method (Sambrooket al. 1989) with Sequenase (United States Biochemical Corp., Cleveland). Terminal sequences of the isolated cDNAs were determined using a 3′ T7 primer and a 5′ primer (5′-TCGAGGTCGACCCACGC) synthesized to match polylinker sequences in the vector.
Plasmids: ZM43 and ZM44 are GAL1 promoter derivatives of YCplac33 (URA3 CEN) and YCplac 111 (LEU2 CEN), respectively, (provided by Z. Moqtaderi, Harvard Medical School, Boston; Gietz and Sugino 1988). pGALSTE20, a BamHI-Pvu-III fragment containing the STE20 coding sequence (CDS) except for amino acid residues 1–69, was isolated from pSTE20-5 (Lebereret al. 1992) and cloned into the BamHI-HindIII (blunt-ended) sites of ZM44. A BstYI-BamHI fragment containing STE20 amino acid residues 1–69 then was cloned into the BamHI site to generate pML52. YEpMSG5, a XbaI-PSTI fragment from YCpMSG5 (Doiet al. 1994), was transferred to XbaI-PstI-cleaved YEplac195, generating pML31.
Cloning of POG1 and gene deletion: The POG1 cDNA in pML2 was isolated from a pRS316-based cDNA library (Liuet al. 1992). The genomic copy of POG1 was cloned as follows: pML32, containing a genomic copy of POG1, was isolated from a Yep24-based genomic yeast library (Carlson and Botstein 1982) probed with POG1 cDNA sequences. A 4-kb SalI-HindIII fragment from pML32 containing the POG1 CDS was then cloned into the SalI-HindIII sites of YEplac195 and YEplac181 to generate pML33 and pML34.
pog1::HIS3 deletion mutation was as follows: A fragment containing 1100 bp of POG1 5′ flanking sequences (–171/–1278 from the ATG) was isolated from pML58 (pBlueScript containing a SalI-EcoRV fragment from YEpPOG1) as a SphI-BamHI fragment and transferred to YIplac211. A fragment containing 1155 bp of POG1 3′ flanking sequences (+153/+1308 from the stop codon) was amplified by polymerase chain reaction (PCR) using primers: (A) 5′-CCGTCAGGATCCACTCCTTATCTCATTTCA-3′ (a BamHI site that is added is underlined) and (B) 5′-CCGTCGAATTCGTTCCTCTTTG TTTCTGG-3′ (an EcoRI site that is added is underlined). The BamHI-EcoRI PCR product was cloned into YIplac211 containing the –171/–1278 piece to generate pML59. A BamHI HIS3 gene fragment from pUC18-HIS3 (provided by D. Kodosh) was then introduced into pML59 to generate pML60. For gene replacement, pML60 was digested with SphI and EcoRI and the resulting pog1::HIS3 fragment was used for transformation. Replacement of the genomic POG1 locus was confirmed by Southern analysis (Sambrooket al. 1989).
Epitope tagging of POG1: To place the green fluorescent protein (GFP) tag on the N terminus of Pog1, an AccI fragment from pML33 containing amino acid residues 48–351 of Pog1 was blunt-ended and cloned into the BamHI (blunt-ended) site of pCGF-1A (Leeet al. 1996), creating pML61. To HA-tag Pog1 at the C terminus the POG1 CDS was amplified by PCR using primers: (C) 5′-CCGTCAGAATTCATGAAGCAG GAGCCACAT-3′ (an added EcoRI site is underlined) and (D) 5′-CGCTCAGTCGACGAATGAAGGTTAGGAAGG-3′ (an added SalI site is underlined). The EcoRI-SalI PCR fragment was cloned into pBluescript to generate pML62. A HA triple tag from pGTE1 (Tyerset al. 1992) was cloned into the BseRI site within the POG1 CDS in pML62, introducing the HA tag between amino acids 345–347 of the POG1 CDS, generating pML68. Then the EcoRI-SalI fragments from pML62 and pML68 were transferred to pDAD2 (URA3 2μ), placing the untagged and HA-tagged copies of POG1 under the control of the GAL1 promoter and resulting in plasmids pML67 and pML72. None of the plasmids containing tagged copies of Pog1 conferred α-factor resistance in halo assays.
Halo and spotting assays: α-Factor sensitivity was measured by halo assay as described (Elionet al. 1990) using 50 μl of an overnight culture or of cultures equalized for cell density. α-Factor peptide (synthesized by C. Dahl, Harvard Medical School, Boston) was dissolved in 90% methanol and stored at –20°. Unless indicated otherwise, 420 ng of α-factor was used for all sst1 strains. All halo assays were done at least twice using independent transformants. For spotting assays, cells were diluted to the same A600 (0.4–0.5) and then diluted serially 100× over a 10,000-fold range before spotting 5 μl of each dilution on solid medium. For α-factor resistance tests, yeast cells were spotted onto plates spread with 4.2 μg of α-factor.
Growth conditions: Yeast strains were grown in selective synthetic complete (SC) medium containing either 2% dextrose or 2% galactose. All strains were grown at 30° except for C699-59 and its control strain, which were grown at 25°. For α-factor inductions, logarithmically growing cells were adjusted to the same A600 (0.4–0.6) and divided into two with one-half receiving α-factor. Cultures were incubated with shaking for 2 hr and then harvested. For galactose induction of genes under the GAL1 promoter, the cells were grown in 2% galactose for 2–2.5 hr prior to the addition of α-factor. All sst1 cultures were treated with 50 nm α-factor unless otherwise stated. SST1 cultures were treated with 1 mm α-factor.
Preparation of yeast protein extracts: Cells were harvested at 4°, washed twice with cold sterile water, and frozen in dry ice. Whole cell extracts were prepared by lysis with glass beads as described (Elionet al. 1993) using modified H buffer adjusted to 125 mm NaCl, 2.5 mm benzamidine, and 1 μg/ml aprotinin. Protein concentration was determined with the Bio-Rad (Richmond, CA) protein assay.
β-Galactosidase assays: Yeast strains transformed with pYBS45 (Lyonset al. 1996) were either left untreated or treated with 50 nm (sst1) or 2 mm (SST1) α-factor for 2 hr. Protein extracts were prepared and assayed for β-galactosidase (lacZ) activity as described (Elionet al. 1995).
Western blots: Yeast protein extract (50–100 μg) was loaded in 8% SDS-PAGE gels. 12CA5 mouse monoclonal antibody (ascites fluid from Harvard University Antibody Facility) at a 1/10,000 dilution was used to detect Cln2HA (Tyerset al. 1993). Western blots were done as described (Harlow and Lane 1988). Blots were developed with the Amersham (Arlington Heights, IL) ECL kit according to the manufacturer's instructions using Fuji RX X-ray film.
Northern blots: Total RNA was prepared and 10–20 μg of RNA was loaded in duplicate 1% formaldehyde-agarose gels, transferred to nitrocellulose (Schleisser and Schuller) and probed by Southern analysis (Sambrooket al. 1989). DNA probes were labeled by the random-primed method. The probes used were a 1.8-kb SalI-NotI CLN2 fragment from pGALCLN2 (pML1), a 1.6-kb BamHI CLN1 fragment from EB608 (Elionet al. 1991), a 1.4-kb HindIII-SpeI CLN3 fragment from pBF30 (provided by B. Futcher), and XhoI-HindIII ACT1 from pYEE15 (Elionet al. 1991). Blots were probed simultaneously with ACT1 as a control for loading and CLN1, CLN2, or CLN3. Relative amounts of mRNA were quantified using a Fuji Imaging plate.
RNase protection: lacZ and ACTI probes were synthesized from pSPCTV and pSACTall (Stuart and Wittenberg 1994) using the Promega (Madison, WI) Riboprobe kit. RNase protection was performed as described in Stuart and Wittenberg (1994) using a protocol generously provided by D. Stuart.
Cell morphology and indirect immunofluorescence: All microscopy was performed using an Axioscope fluorescence microscope (Carl Zeiss, Thornwood, NY). To determine the percentage of budded cells, cells were fixed in 3.7% formaldehyde and briefly sonicated before quantitation as described (Elionet al. 1990). Fluorescence signal from GFP-tagged Pog1 was examined in the FITC channel after cells were grown in 2% galactose for 1 hr. Detection of C-terminal HA-tagged Pog1 through the use of 12CA5 monoclonal antibody and DTGF-labeled goat anti-mouse antibody was done as described (Leeet al. 1996). Images shown in Figure 6C were captured with a Toshiba 3CCD camera and Phase 3 Imaging System by Media Cybernetics.
Protein sequence analysis: The Pog1 amino acid sequence was analyzed using BLASTP/X (Altschulet al. 1990), BEAUTY (Worleyet al. 1995), TFASTA/FASTP (Pearson and Lipman 1988), MOTIFS, and BLOCKS (Henikoff and Henikoff 1994) database search programs.
Isolation of cDNAs that block pheromone-induced G1 arrest: To isolate genes that promote recovery, we transformed a MATa sst1Δ strain (EY1118) with a yeast cDNA library under the control of the GAL1 promoter (Liuet al. 1992) and screened for α-factor-resistant colonies on selective medium containing galactose. sst1Δ strains are supersensitive to α-factor (Chan and Otte 1982) due to the loss of the Sst1 protease that degrades α-factor (Ciejek and Thorner 1979; Sprague and Herskowitz 1981). We identified seven plasmids that conferred galactose-dependent resistance to α-factor. Partial sequencing of the cDNA inserts revealed that they encoded ∼278 amino acids of the C terminus of SIR4 and the entire coding sequences of CLN2, GPA1, and a novel gene we named POG1 (for Promoter of Growth; GenEMBL AC Z46833, ORF YI8277.07, and SwissProt. AC P40473). CLN2 and GPA1 encode positive regulators of recovery (Dietzel and Kurjan 1987; Miyajimaet al. 1987; Oehlen and Cross 1994), confirming the validity of the screen. In contrast, the Sir4 C-terminal fragment confers α-factor resistance by an indirect mechanism that is linked to derepression of silent MATα information at HMLα (data not shown; Marshallet al. 1987). The POG1 cDNA does not promote α-factor resistance through derepression of HMLα as it confers α-factor resistance in a strain that is deleted for HMLα (MLY4; data not shown).
To rule out the possibility that the α-factor resistance of the POG1 cDNA was the result of cDNA library construction and/or growth of cells on galactose, we isolated the native POG1 gene from a 2μ-based yeast library (YEpPOG1; see materials and methods). MATa sst1Δ cells overexpressing POG1 (YEpPOG1) displayed obvious resistance to α-factor compared to cells transformed with a vector control (Figure 1A; compare turbid halo to clear halo). The ability of the YEpPOG1 plasmid to confer α-factor resistance is similar to that of the ADH-CLN2 and YEpMSG5 plasmids (Figure 1A). Thus, in this assay, the POG1 gene promotes α-factor resistance as well as two other known promoters of recovery.
POG1 encodes a protein of 351 amino acids with a predicted mass of ∼40 kD and no significant homologies to known proteins or protein motifs on the basis of protein-sequence comparison searches (see materials and methods). Pog1 has an acidic N-terminal half and a basic C-terminal half rich in proline residues. Localization studies on hemagglutinin (HA)- and GFP-tagged versions of Pog1 suggest that the protein localizes in the nucleus (data not shown). However, all tagged derivatives of Pog1 examined to date are nonfunctional, so these data must be viewed with caution. On the basis of Northern blot analysis, POG1 mRNA levels are similar in a and α haploid and a/α diploid cells and are not affected by pheromone exposure (data not shown), suggesting that POG1 has a function that is not restricted to regulation of the pheromone response pathway.
POG1 overexpression blocks G1 arrest and inhibits expression of FUS1: We determined whether pGAL-POG1 promotes α-factor resistance by interfering with G1 arrest and shmoo formation. Cells overexpressing POG1 were treated with a slightly subsaturating concentration of α-factor (F. Farley and E. A. Elion, unpublished results) and the percentage of budded cells was determined before and after exposure to α-factor. Cells overexpressing POG1 had a slightly greater percentage of budded cells than the control cells prior to exposure to α-factor, with essentially identical cell morphology (Figure 1B). After a 2-hr exposure to α-factor, the POG1 overexpressing cells maintained a high percentage of budded cells while nearly all of the control cells underwent G1 arrest. Cells overexpressing POG1 in the presence of α-factor were also blocked for shmoo formation. Only a small fraction (∼10%) of the unbudded POG1-transformed cells displayed the projections typical of cells responding to pheromone (compared to ∼70% of the unbudded control cells). Thus, overexpression of POG1 blocks pheromone-induced G1 arrest and cell morphological changes.
We determined whether the POG1-induced block in G1 arrest and projection formation correlated with inhibition of the MAPK cascade, by assaying the effect of POG1 overexpression on α-factor-induced transcription of the FUS1 gene. The FUS1 gene is strongly induced by pheromone and is dependent upon an active MAPK pathway (McCaffreyet al. 1987; Trueheartet al. 1987; Elionet al. 1991). For comparison, we monitored the effect of pGALMSG5, a known inhibitor of Fus3 (Doiet al. 1994). MATa sst1Δ cells were cotransformed with vector pGALPOG1 or pGALMSG5, and a second plasmid containing a FUS1-lacZ reporter gene (pYBS45). Strains were first grown in medium containing galactose to induce the expression of POG1 and MSG5 and then incubated with pheromone and assayed for β-galactosidase levels. Overexpression of POG1 causes a reproducible twofold decrease in the levels of FUS1 expression (Figure 1C). This level of inhibition is similar to that caused by overexpression of MSG5 (Figure 1C; Doiet al. 1994).
Deletion of POG1 increases the α-factor sensitivity of an msg5 mutant: We next examined the effect of a POG1 deletion on pheromone response and growth. A pog1 null strain was created by replacing one chromosomal copy of the POG1 gene in a wild-type diploid strain (MLY16) with a pog1::HIS3 allele lacking the POG1 coding sequence (MLY17; see materials and methods). Upon sporulation and tetrad dissection of the heterozygous diploid, all four spores were equally viable. Compared with isogenic wild-type spore clones, the pog1::HIS3 spore clones exhibited no obvious growth defects, no heightened α-factor sensitivity in halo assays, and no differences in levels of FUS1-lacZ expression in either the absence or presence of α-factor (data not shown).
It was possible that the absence of a phenotype for the pog1 null was due to the fact that recovery is regulated at multiple levels, any of which might operate in parallel with POG1. For example, deletion of the MSG5 gene causes only a slight increase in α-factor sensitivity (Doiet al. 1994), most likely because it is only one of three phosphatases that regulates Fus3 (Zhanet al. 1997). We tested whether POG1 regulates recovery parallel to MSG5 by constructing a msg5 pog1 double mutant. An isogenic msg5::LEU2 disruption strain (MLY81) was mated to a pog1Δ mutant (MLY30). MATa wild-type, single, and double-mutant spore clones were tested for α-factor sensitivity using a spotting assay. As shown in Figure 1D, the pog1 msg5 double mutant is more sensitive to α-factor than either single mutant, suggesting that POG1 promotes recovery in parallel with MSG5.
POG1 requires CLN2 to promote recovery: We investigated whether POG1 requires any of the genes known to regulate the recovery response, as further evidence for a physiological role in regulating recovery. The ability of pGALPOG1 to promote α-factor resistance was tested in a variety of strains harboring deletions in regulators of recovery, using halo assays as a monitor. pGAL-POG1 conferred α-factor resistance to both sst2Δ and msg5Δ deletion strains (data not shown), suggesting that POG1 functions in parallel to both SST2 and MSG5. By contrast, pGALPOG1 was unable to confer α-factor resistance to a cln2 deletion mutant (EY1027; Figure 2). The requirement for the CLN2 G1 cyclin is remarkably specific, as POG1 overexpression confers significant α-factor resistance to isogenic cln1 (EY1028) and cln3 (ML201) mutants. Thus, POG1 requires CLN2, but not CLN1 or CLN3, to promote growth in the presence of pheromone.
POG1 blocks α-factor-induced repression of CLN1 and CLN2 mRNAs: The requirement for CLN2 for POG1-dependent α-factor resistance suggested that POG1 might upregulate the levels of Cln2 in the presence of α-factor. We first examined Cln2 levels in a MATa sst1Δ strain containing an integrated copy of a HA-tagged CLN2 gene (CY326). Normally, the level of Cln2-HA is significantly reduced by α-factor (Figure 3A, vector control) as a result of transcriptional repression of the CLN2 gene (Valdiviesoet al. 1993; Cherkasovaet al. 1999). However, the abundance of Cln2-HA is only slightly reduced when POG1 is overexpressed. POG1 overexpression overcomes the α-factor inhibition of Cln2-HA to almost the same degree as GPA1val50, a gene known to cause hyperadaptation (Miyajimaet al. 1989). This loss of negative control is at the level of CLN2 mRNA. Overexpression of POG1 and GPA1val50 prevents α-factor-induced inhibition of both CLN1 and CLN2 mRNAs (Figure 3B). POG1 overexpression has no obvious effect on the level of CLN3 mRNA, in contrast to GPA1val50, which clearly affects CLN3 mRNA levels.
POG1 stimulates transcription via SCB/MCB elements: We determined whether POG1 was regulating the expression of CLN1 and CLN2 through a common promoter element. Cell-cycle-dependent transcription of CLN1 and CLN2 is primarily regulated by the Swi4 and Swi6 transcription factors that act as a complex on SCB and MCB elements upstream of the transcription start of both genes (Nasmyth and Dirick 1991; Ogaset al. 1991). In the CLN2 promoter, SCB and MCB elements are present in UAS1, with a second UAS, UAS2, contributing only slightly to cell-cycle transcription (Figure 3D; Stuart and Wittenberg 1994). We tested whether POG1 acts through the SCB/MCB elements using a CLN2-lacZ promoter fusion gene containing either the full-length CLN2 promoter (–729 – 256) or two mutant derivatives, one lacking UAS2 and the other lacking both UAS2 and functional SCB/MCB elements within UAS1 (Stuart and Wittenberg 1994). RNase protection analysis showed that POG1 is able to increase expression of the lacZ gene when it is controlled by either the full-length CLN2 promoter or the derivative lacking UAS2. Mutation of the SCB and MCB elements blocked the effect (Figure 3C), suggesting that POG1 acts through SCB or MCB elements.
This result was confirmed independently using a chromosomally integrated lacZ reporter gene under the control of synthetic SCB elements (CY3557; Table 2). POG1 overexpression induced the levels of β-galactosidase activity ∼2- to 2.5-fold in the presence of α-factor compared to control. A parallel experiment was conducted in an isogenic strain harboring a multicopy plasmid containing the lacZ reporter gene under the control of synthetic MCB elements (Di Comoet al. 1995). Overexpression of POG1 had no obvious effect on the expression of the MCB-lacZ reporter gene (data not shown). However, the absence of an effect could be due to the high dosage of the reporter gene coupled with the stability of β-galactosidase or to the absence of proper control by the MCB elements when they are out of the context of the CLN2 promoter. From these results we conclude that overexpression of POG1 either indirectly or directly stimulates SCB-dependent transcription in the presence of α-factor. Further work is needed to demonstrate whether POG1 acts through MCB elements.
POG1 and CLN2 both require STE20 to promote efficient recovery: We performed genetic epistasis tests to determine the step(s) at which POG1 might regulate the mating signal transduction pathway. Overexpression of POG1 blocked the ability of pGALSTE4 (Gβ) to induce G1 arrest, as demonstrated by the restoration of growth on galactose plates for the pGALSTE4 strain harboring YEpPOG1 (Figure 4). POG1 also efficiently promotes α-factor resistance in the presence of STE11-4 (Figure 5A), a gain-of-function allele of STE11 that causes high constitutive and induced signaling (Stevensonet al. 1992). Note that in the STE11-4 strains shown in Figure 5, A and B, FAR1 is deleted in order to reduce STE11-4-induced growth inhibition in the absence of pheromone. In these strains, α-factor mediates G1 arrest through enhanced transcriptional repression of CLN1 and CLN2 (Cherkasovaet al. 1999). Thus, POG1 acts a step(s) below Gβ in the pathway that may be distinct from Ste11.
Two experiments suggest that POG1 may act at the STE20 step of the pathway. First, overexpression of STE20 in a MATa sst1Δ strain greatly reduces the ability of POG1 to promote α-factor resistance, although MSG5 is still able to efficiently promote recovery (Figure 5D). Second, POG1 loses most of its ability to promote α-factor resistance in the absence of STE20, as assessed in a pheromone responsive STE11-4 ste20Δ strain (compare Figure 5, A and B). STE11-4 partially bypasses the requirement for STE20 for α-factor-dependent activation of the MAPK cascade and G1 arrest, providing a means to assess α-factor sensitivity in the absence of STE20 (Lyonset al. 1996; Fenget al. 1998). These results suggest that POG1 requires STE20 to promote cell division in the presence of α-factor.
We tested the possibility that POG1 requires STE20 to upregulate CLN2 transcription in the presence of α-factor, by determining the ability of POG1 to block repression of CLN2 transcription in the STE11-4 ste20Δ strain. Northern blots show that STE20 is not required for the enhanced levels of CLN2 mRNA produced by POG1 (Figure 5C). Thus, the inability of POG1 to induce recovery is not due to a loss in its ability to induce CLN2 expression, raising the possibility that CLN2 requires STE20 to promote recovery. Consistent with this, ADH-CLN2 also has a reduced ability to confer α-factor resistance in a STE11-4 ste20Δ (EY2022) strain compared with a STE11-4 strain (EY1298; data not shown). Furthermore, we find that in the presence of excess STE20, CLN2-overexpressing cells display a reduced ability to recover (Figure 5D). However, CLN2 has a greater ability than POG1 to promote recovery in the presence of excess STE20, as shown by the slightly greater turbidity of the pGALCLN2 halo compared with the pGALPOG1 halo. This difference could be due to quantitative differences in the levels of CLN2 mRNA produced in the two strains. For example, the level of CLN2 mRNA produced from the ADH-CLN2 gene is approximately five-fold greater than the level of CLN2 mRNA generated by YEpPOG1 in the presence of α-factor (data not shown). Taken together, the results suggest that STE20 is an important target of control for recovery events that are mediated by CLN2.
POG1 has a vegetative function that is redundant with CLN3: Deletion of POG1 does not affect the rate of appearance or levels of CLN1 and CLN2 mRNAs during recovery (data not shown), raising the possibility that POG1 regulates G1 cyclin transcription through a redundant mechanism. CLN3 is a positive regulator of CLN1 and CLN2 transcription (Tyers and Futcher 1993; Nasmyth 1996) and functions in parallel with other genes to control expression of the G1 cyclins (Di Comoet al. 1995). We therefore compared the ability of MATa wild-type, single, and double-mutant spore clones from a pog1Δ (MLY30) × cln3Δ (ML201) cross to undergo α-factor-induced G1 arrest and regulate CLN1 and CLN2 transcription. pog1 cln3 double mutants were more sensitive than the wild-type strain, but as sensitive to α-factor as the cln3 single mutants. Northern blot analysis of cln3 and pog1 cln3 strains did not reveal an obvious difference in the rate of appearance or absolute levels of CLN1 and CLN2 mRNAs during recovery (data not shown).
In contrast, pog1 cln3 double mutants were found to exhibit two vegetative growth defects. First, a pog1 cln3 double mutant has a more pronounced temperature-sensitive growth defect than a cln3 single mutant, as shown by spotting equal numbers of cells on YPD plates and incubating them at 37° (Figure 6A). The pog1 single mutant has no obvious temperature sensitivity. The temperature sensitivity of the pog1 cln3 double mutant is remedied by the inclusion of either 12 mm Mg2+, 1 m sorbitol (Figure 6A), or 25 mm Ca2+ (data not shown) in the medium. Sorbitol functions as an osmotic stabilizer and remedies cell lysis (Cidet al. 1995). Ca2+ and Mg2+ may also stabilize the cell membrane and enhance cell wall biosynthesis (Lin and Macey 1978; Levinet al. 1990; Cyertet al. 1991; Mariniet al. 1996). Microscopic examination of cells grown on solid medium at 37° overnight and then resuspended in 1 m sorbitol for osmotic support show that the pog1 cln3 cells are, on average, larger sized than the single mutants or wild type (Figure 6C).
Second, a pog1 cln3 double mutant has greatly reduced viability when it is starved for a carbon source (Figure 6B). The effect of starvation for a carbon source was examined by patching wild-type, single-, and double-mutant strains onto solid SC medium lacking dextrose. After several days at 25°, the patches were transferred to fresh YPD plates for further incubation. A loss of viability after carbon source starvation suggests that the pog1 cln3 double mutant is unable to enter Go (Werner-Washburneet al. 1993). Collectively, these results suggest that, parallel to CLN3, POG1 promotes vegetative growth.
The pog1 cln3 mutant shares some characteristics of Pkc1 pathway mutants: the temperature sensitivity of the double mutant is rescued by sorbitol, Mg2+, and Ca2+, and the morphological defects are accentuated in solid medium (Lee and Levin 1992; Paraviciniet al. 1992; Mariniet al. 1996). In contrast to Pkc1 pathway mutants (Costiganet al. 1992), the pog1 cln3 double mutant is not sensitive to nitrogen starvation (Figure 5B) or caffeine and can grow in the presence of nonfermentable carbon sources such as ethanol and glycerol (data not shown). In addition, overexpression of POG1 does not rescue growth defects of Pkc1 pathway mutants (e.g., bck1Δ and mpk1Δ, data not shown). For these reasons, POG1 is unlikely to be directly involved in the Pkc1 pathway.
POG1 requires BCK1 to increase CLN2 expression in the presence of α-factor: The partial overlap of phenotypes between the pog1 cln3 double mutant and Pkc1 pathway mutants suggested that POG1 may regulate a subset of the same genes that are also regulated by Pkc1. The Pkc1 pathway plays a major role in cell-wall biosynthesis and contributes to the activation of G1 cyclin expression via the Swi4 and Swi6 transcription factors (Maddenet al. 1997) and G1 cyclins are thought to be activators of the Pkc1 pathway (Zarzovet al. 1996; Grayet al. 1997). We determined whether POG1 requires the Pkc1 pathway to regulate the expression of CLN2 in the presence of α-factor using a bck1Δ deletion strain. BCK1 encodes a MAPK kinase kinase that functions downstream of Pkc1 (Costiganet al. 1992) and upstream of the MAPK Mpk1 thought to regulate Swi4 (Maddenet al. 1997). Northern blots showed that CLN2 mRNA levels are not elevated to an obvious degree in the bck1 cells overexpressing POG1 compared to the wild-type strain (Figure 7A). In halo assays, the POG1-overexpressing bck1 cells exhibited partially filled-in halos indicating that POG1 can stimulate recovery in the abence of BCK1 (Figure 7B), although not nearly as well as in its presence (Figure 1A). These results have two implications. First, POG1 may require BCK1 to regulate the expression of the CLN2 gene. Second, POG1 may promote recovery through additional routes besides upregulation of CLN2.
POG1 is a novel regulator of recovery that operates through CLN2: We have isolated and characterized a novel gene, POG1, that blocks α-factor-induced G1 arrest when overexpressed. Several lines of evidence argue that POG1 positively regulates recovery in the presence of α-factor. First, as a multicopy suppressor, POG1 provides a rate-limiting function that promotes cell division and, in the presence of α-factor, interferes with the ability of cells to arrest in G1 phase and form shmoos (Figure 1). Second, a pog1 null mutation enhances the α-factor sensitivity of an msg5 mutant, suggesting that Pog1 promotes recovery in parallel with Msg5, a known regulator of recovery (Figure 1). Third, POG1 specifically requires the CLN2 G1 cyclin to promote cell division (Figure 2), and CLN2 is implicated in having a role in recovery (Oehlen and Cross 1994; Wassmann and Ammerer 1997). Fourth, POG1 elevates CLN2 and CLN1 mRNA levels in the presence of α-factor (Figure 3), providing a molecular explanation for the requirement for CLN2. Taken together, these data provide strong support for a model in which POG1 promotes recovery by overcoming transcriptional repression of CLN2 and CLN1. The resulting increased levels of Cln2 and Cln1 protein may be sufficient to promote recovery. However, the strict requirement for the CLN2 gene suggests that the Cln2 protein performs essential functions for recovery that are not shared by Cln1, in accordance with previous observations (Oehlen and Cross 1994).
POG1 may regulate transcription of SCB/MCB elements: POG1 is likely to operate at the level of transcription of the CLN1 and CLN2 genes. Excess POG1 stimulates the expression of β-galactosidase promoter fusions containing either the CLN2 SCB/MCB elements or synthetic SCB elements (Figure 3), ruling out a post-transcriptional effect. Two observations suggest that Pog1 may play a direct role in regulating CLN1 and CLN2 transcription. First, HA- and GFP-tagged versions of Pog1 localize in the nucleus (data not shown). While these proteins are nonfunctional, their nuclear localization may accurately reflect a nuclear function for Pog1. Second, the predicted Pog1 protein contains two hallmarks of transcription factors, an acidic domain in the N-terminal half and a proline-rich domain in the C-terminal half.
Two interpretations are possible to explain the mechanism by which Pog1 increases the expression of CLN1 and CLN2 in the presence of pheromone. Pog1 may be a transcriptional activator that activates by binding directly to the SCB/MCB elements or it may positively regulate the activity of the Swi4/Swi6 complex. Alternatively, Pog1 could be an inhibitor of α-factor-induced repression of the CLN1 and CLN2 promoters. Previous work has shown that Fus3 and Kss1 repress transcription of the CLN1 and CLN2 genes to promote G1 arrest (Cherkasovaet al. 1999). Pog1 could interfere with the ability of Fus3 and Kss1 to repress transcription by blocking an α-factor-induced repressor that abrogates transcription through the SCB/MCB elements, or by affecting directly the repressive function of either Fus3 or Kss1.
It seems likely that the ability of Pog1 to upregulate the CLN1 and CLN2 genes is dependent upon α-factor, because neither overexpression nor deletion of POG1 has an obvious effect on CLN1 CLN2 mRNA levels during vegetative growth, even in the absence of the CLN3 gene. Potential control by pheromone is not at the level of POG1 gene expression; the POG1 gene is not pheromone inducible and it is expressed in diploid cells as well as haploid cells. Perhaps the Pog1 protein or its target(s) are modified in response to α-factor. In this regard, two-hybrid analysis suggests that Pog1 interacts with Kss1 but not Fus3, and Pog1 confers significantly less α-factor resistance when the KSS1 gene is deleted (data not shown). These observations raise the possibility that Kss1 regulates Pog1 to promote recovery or vice versa. Pog1 also requires Bck1 to stimulate expression of the CLN2 gene (Figure 7), suggesting that Pog1 or another protein must first be modified (directly or indirectly) by Bck1 in order for Pog1 to function. Swi4 is a possible candidate as it has been implicated as a downstream target of Mpk1 (Maddenet al. 1997).
POG1 may regulate additional genes during vegetative growth and recovery: Our analysis suggests that Pog1 has additional functions for vegetative growth and recovery that are distinct from transcriptional control of CLN1 and CLN2. POG1 promotes vegetative growth in parallel with CLN3, as shown by the enlarged size, temperature-sensitive, and cellular lysis phenotypes of a pog1 cln3 double mutant (Figure 6 and data not shown). However, this growth defect does not correlate with reduced rates of appearance or levels of CLN1 or CLN2 mRNAs (data not shown). The phenotypes of the pog1 cln3 double mutant are reminiscent of mutants defective in cell-wall integrity or the actin cytoskeleton (Cidet al. 1995; Roemeret al. 1998). The Pkc1 pathway regulates cell-wall integrity and nutrient response (Levin and Errede 1995). Mutants with defects in the protein kinases under the control of Pkc1 undergo temperature-sensitive lysis that is remedied by osmotic support, similar to a pog1 cln3 double mutant (Levin and Errede 1995). Likewise, the actin cytoskeleton mutant rvs167 is also sensitive to osmotic stress and to starvation for carbon, nitrogen, or sulfur sources (Baueret al. 1993). Thus, POG1 may regulate genes that control cell-wall integrity or the actin cytoskeleton. Such genes may be controlled by SCB or SCB-like elements.
POG1 may also positively regulate recovery through functions distinct from transcriptional control of the CLN2 gene. First, although a bck1 mutation nearly completely blocks the ability of overexpressed POG1 to upregulate the CLN2 gene, POG1 still confers partial α-factor resistance (Figure 7). Second, in the presence of the STE11-4 mutation, a POG1 multicopy plasmid confers significantly more α-factor resistance than does an ADH-CLN2 plasmid, even though the POG1 plasmid induces less CLN2 mRNA (data not shown). The STE11-4 mutation is thought to bypass an inhibitory effect of Cln2/Cdc28 kinase at the Ste11 step of the pathway (Wassmann and Ammerer 1997). Thus, POG1 overexpression can still promote recovery under conditions that block either the expression or function of the CLN2 gene, arguing that POG1 regulates additional genes to promote recovery. Candidate genes include the PCL2 gene that is regulated by Swi4 and whose expression is highly induced by α-factor (Measday et al. 1994, 1997) and the PCL1,10 genes that are repressed by α-factor (Measdayet al. 1997). PCL1,2,10 genes encode 3 of 10 cyclin partners for the Pho85 kinase (Measdayet al. 1997), which promotes budding in parallel with Cln1/Cdc28 and Cln2/Cdc28 (Measdayet al. 1994) and regulates cell morphology (Measdayet al. 1997). pho85 mutants are more sensitive to α-factor than wild-type cells (Measdayet al. 1997) and delay recovery from G1 arrest (Cherkasovaet al. 1999), consistent with the possibility that transcriptional control of Pho85 cyclin partners may play a role in recovery.
Ste20 may be a critical target of control for recovery from G1 arrest: It is noteworthy that disruption or overexpression of STE20 greatly decreases the ability of both POG1 and CLN2 to confer α-factor resistance (Figure 5). Thus, POG1 and CLN2 may promote recovery by inhibiting the ability of Ste20 to activate the pheromone response pathway. Consistent with this, overexpression of POG1 and CLN2 modestly inhibits the expression of the FUS1 gene (Figure 1; Oehlen and Cross 1994). Because POG1 is able to induce the expression of CLN2 in the absence of a functional STE20 gene (Figure 5), it seems likely that Pog1 activates the expression of the CLN2 gene, which in turn promotes recovery through an effect on Ste20. Previous work suggests that Cln2/Cdc28 inhibits the mating signal transduction pathway at or near the Ste11 step (Wassman and Ammerer 1997). Ste20 is essential for the activation of Ste11 (Lebereret al. 1992; Fenget al. 1998). Furthermore, a STE11-4 mutation partially bypasses the requirement for Ste20 for signal transduction (Lyonset al. 1996), explaining why STE11-4 can bypass the inhibitory effects of excess CLN2 (Wassman and Ammerer 1997).
How might Cln2/Cdc28 regulate Ste20 to promote recovery? Ste20 can either promote G1 arrest and shmoo formation by interacting with Gβ (Leeuwet al. 1998) or budding and pseudohyphal growth by interacting with Cdc42 (Peteret al. 1996; Lebereret al. 1997). Cln2/Cdc28 is thought to act upstream of Ste20 and Cla4 to promote budding (Cvrckováet al. 1995), suggesting that Cln2/Cdc28 may directly or indirectly alter the specificity of Ste20 from mating to budding. We find that the Cdc42 binding site of Ste20 is not required for POG1 or CLN2 overexpression to promote recovery (data not shown), suggesting that the control is not at the level of association with Cdc42. Further studies are required to clarify how Ste20 is regulated to promote recovery.
We thank H. Liu, Z. Moqtaderi, D. Stuart, M. Whiteway, C. Wittenberg and B. Errede for their generous gift of plasmids and strains and Andrew Neuwald for kindly analyzing the Pog1 protein sequence with the Probe program. We thank S. Buratowski, H. Saito, K. Struhl, D. Takemoto, F. Winston, and members of the Elion laboratory for insightful comments on the manuscript. This research was supported by grants to E.A.E. from the March of Dimes (#1-FY96-0925), Council for Tobacco Research, and National Institutes of Health (GM 46962).
Communicating editor: M. Carlson
- Received September 9, 1998.
- Accepted November 3, 1998.
- Copyright © 1999 by the Genetics Society of America