Defects in Protein Glycosylation Cause SHO1-Dependent Activation of a STE12 Signaling Pathway in Yeast

Corresponding author: George F. Sprague, Jr., Institute of Molecular Biology, University of Oregon, Eugene, OR 97403-1229., gsprague{at}molbio.uoregon.edu (E-mail)

Communicating editor: A. P. MITCHELL

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In haploid Saccharomyces cerevisiae, mating occurs by activation of the pheromone response pathway. A genetic selection for mutants that activate this pathway uncovered a class of mutants defective in cell wall integrity. Partial loss-of-function alleles of PGI1, PMI40, PSA1, DPM1, ALG1, MNN10, SPT14, and OCH1, genes required for mannose utilization and protein glycosylation, activated a pheromone-response-pathway-dependent reporter (FUS1) in cells lacking a basal signal (ste4). Pathway activation was suppressed by the addition of mannose to hexose isomerase mutants pgi1-101 and pmi40-101, which bypassed the requirement for mannose biosynthesis in these mutants. Pathway activation was also suppressed in dpm1-101 mutants by plasmids that contained RER2 or PSA1, which produce the substrates for Dpm1. Activation of FUS1 transcription in the mannose utilization/protein glycosylation mutants required some but not all proteins from three different signaling pathways: the pheromone response, invasive growth, and HOG pathways. We specifically suggest that a Sho1 Ste20/Ste50 Ste11 Ste7 Kss1 Ste12 pathway is responsible for activation of FUS1 transcription in these mutants. Because loss of pheromone response pathway components leads to a synthetic growth defect in mannose utilization/protein glycosylation mutants, we suggest that the Sho1 Ste12 pathway contributes to maintenance of cell wall integrity in vegetative cells.

VEGETATIVE yeast cells respond to environmental cues by activating signal transduction pathways that enable them to mount the appropriate physiological response. Among the cues that yeast responds to are mating pheromone, nutrients, osmolarity of the growth medium, and changes in turgor pressure at the plasma membrane. Each of these cues is dealt with by distinct signaling mechanisms to cause the appropriate response to a given stimulus.

Activation of the mating factor signal transduction pathway by binding of pheromone to its cognate receptor leads to a new pattern of gene transcription, to arrest of the cell division cycle in the G1 phase, and to reorientation of cell polarity toward the perceived mate (see reviews from BARDWELL et al. 1994 ; ERREDE et al. 1995 ; PRINGLE et al. 1995 ; HERSKOWITZ 1997 ; LEBERER et al. 1997 ; MADDEN and SNYDER 1998 ; CHANT 1999 ). The signaling pathway is comprised of biochemical modules that are common to many transduction pathways in eukaryotic cells. At the head of the pathway is a serpentine seven-transmembrane receptor coupled to a heterotrimeric G protein. In ways that are incompletely understood, the activated G protein is coupled to a mitogen-activated protein (MAP) kinase cascade. This coupling involves Ste20, a protein kinase of the PAK family, and Ste5, a scaffolding protein for the MAP kinase cascade. The cascade itself is composed of Ste11 (an MAPKKK), Ste7 (an MAPKK), and Fus3 or Kss1 (two MAPKs). Either MAPK is capable of functioning in the pheromone pathway, but it is likely that Fus3 does so in vivo (MADHANI and FINK 1998 ). Targets of the terminal MAPK include Ste12, a transcription factor required for transcription of pheromone-responsive genes, and Far1, a CDK inhibitor. These targets, therefore, constitute at least part of the mechanism whereby pheromone regulates transcription and cell cycle progression.

The third physiological effect of pheromone, reoriented cellular polarity, requires a different biochemical module, Cdc42, a p21 GTPase of the Ras superfamily. The membrane-tethered, activated G protein is thought to lead to localized activation of the guanine nucleotide exchange factor (GEF) for Cdc42 and, thereby, to localized activation of Cdc42 (SIMON et al. 1995 ; ZHAO et al. 1995 ; NERN and ARKOWITZ 1999 ). Once activated, Cdc42 can organize the actin cytoskeleton as it does in vegetative cells. Cdc42 may also activate Ste20 and influence signaling through the MAP kinase cascade (STEVENSON et al. 1995 ; PRYCIAK and HUNTRESS 1998 ; JOHNSON 1999 ).

Although there is a fairly sophisticated understanding of signal transmission in the pheromone pathway, how signaling is regulated—for example, how the signal is attenuated so that the cell cycle can reinitiate once mating occurs—is poorly understood. Likewise, how the pheromone pathway may interface with other aspects of cell physiology is also poorly understood. These issues are brought into sharp focus by the realization that pheromone pathway components participate in other signal transduction pathways. For example, the haploid invasive growth phenotype, which is a response to nutrient limitation, requires Ste20, Ste11, Ste7, and Kss1 (GIMENO et al. 1992 ; LIU et al. 1993 ). Likewise, one branch of the HOG pathway, which is activated in response to osmotic stress, requires Ste20 and Ste11 (O'ROURKE and HERSKOWITZ 1998 ).

To learn more about regulation of the pheromone pathway and its possible interface with other pathways, we have sought mutants that showed increased expression of FUS1, a pheromone-responsive gene. A substantial number of these mutants had an ancillary slow-growth phenotype, and they proved to be defective at various steps in the pathway leading to the synthesis of mannosyl moieties or in the utilization of these moieties for protein glycosylation. Double-mutant studies argue that a distinct signaling pathway that begins with the membrane protein Sho1 for the HOG pathway and culminates at Ste12 from the pheromone response pathway is required for enhanced FUS1 transcription. In addition, we found that mutations that debilitate protein glycosylation show synthetic interactions with mutations in pheromone pathway component structural genes, namely STE20, STE11, STE7, KSS1, and STE12. These results, which complement and extend those reported by LEE and ELION 1999 , suggest that these components are part of a pathway that operates in vegetative cells to sense or respond to defects in the cell wall.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains, media, and microbiological techniques:
The yeast strains used in this work are listed in Table 1. The majority of the strains are isogenic with SY2002 (STEVENSON et al. 1995 ), a derivative of Sc252 (Jim Hopper lab strain). The selection for His⁺ mutants that activate a FUS1-HIS3 reporter was performed in strains SY2428 (a) and SY2431 (), derived from SY2002. These strains are ste4 GAL-STE4 FUS1-HIS3 mfa2::FUS1-lacZ and have complementary nutritional markers (arg4 and lys2, respectively) to facilitate complementation analysis (Table 1). A strain containing the alg1-1 allele was kindly provided by Phillip Robbins, and the strain containing the spt14-2 allele was provided by Jan Fassler. Deletion of STE genes for epistasis analysis was performed using the one-step replacement plasmids ste7::URA3, ste12::URA3, ste11::URA3, ste5::URA3, ste20::URA3, kss1::URA3, fus3-6::LEU2, ste12::LEU2, and ste4::LEU2. The mata1^- strain was created by two-step replacement of construct YIp5-MATaXho1 to create a frame-shift at the HMR locus followed by mating-type switching (using pGAL-HO) to obtain SY2428. The STE11-4 allele was introduced into strains by transformation of the two-step replacement construct pSL1655. Other strains were constructed using one-step integration plasmids bck1::URA3, slt2::URA3, rga1::URA3, pbs2::URA3, och1::LEU2, mnn10::LEU2, and gas1::URA3. All strains that carried defined deletions were confirmed by Southern analysis or PCR Southern analysis, and by phenotypic analysis when possible. For construction of the pmi40-101 and pgi1-101 homozygous diploids, the a and strains were transformed with complementary plasmids, and growth, mating, and selection for diploids were performed on 2% glucose with 10 mM mannose for pmi40-101, and on 2% mannose and 0.1% glucose for pgi1-101.

Yeast and bacterial strains were propagated using standard methods (SHERMAN et al. 1982 ). YPD and SD media have been described (ROSE et al. 1990 ). Supplementation of media with 10 mM mannose (Sigma, St. Louis) was performed for analysis of strains carrying the pmi40-101 allele. For strains that contained the pgi1-101 allele, 2% mannose (or 2% fructose) and 0.1% glucose were used instead of 2% glucose. Yeast transformations were performed as described (GIETZ et al. 1995 ). Calcofluor (white fluorescent brightener 28 Calcofluor white M2R; Sigma) was used at the concentrations described in the text. Bacterial transformations, bacterial DNA preparations, and plasmid constructions were performed by standard methods (SAMBROOK et al. 1989 ).

Plasmids:
For genetic analysis of the mutants, deletion derivatives were produced by one- or two-step gene replacement using the following constructs. The ste7::URA3 (pSL1077), ste12::URA3 (pSL1311), ste11::URA3 (pSL1094), and STE11-4 (pSL1655) integration constructs were previously reported (STEVENSON et al. 1992 ). The mfa2-1::FUS1-lacZ (pDH17), arg4 (YIp5arg4-EcoRVXhoI), rad16::pGAL::STE4 (pDH15), and his3::FUS1-HIS3 (pDH106) plasmids have been described (HORECKA and SPRAGUE 1996 ). Constructs ste5::URA3 (pSURE; J. Thorner), ste20::URA3 (pEL45; LEBERER et al. 1992 ), kss1::URA3 (pGA1850; G. Ammerer), fus3-6::LEU2 (pYEE98; ELION et al. 1990 ), ste12::LEU2 (pSUL-16; FIELDS and HERSKOWITZ 1987 ), and ste4::LEU2 (p121; WHITEWAY et al. 1989 ) have also been described (STEVENSON et al. 1992 ). Plasmids pRS313, 314, 315, and 316 have been described (SIKORSKI and HIETER 1989 ). The bck1::URA3 and slt2::URA3 plasmids were provided by D. Levin. The rga1::URA3 and pbs2::URA3 (BREWSTER et al. 1993 ) plasmids have been described previously (STEVENSON et al. 1995 ). The och1::LEU2 construct (NAKAYAMA et al. 1992 ) and the mnn10::LEU2 plasmids (DEAN and POSTER 1996 ) were described. Deletion of the lys2 gene was performed with pCP7, using two-step integration, and YIp5-MATaXhoI (TATCHELL et al. 1981 ) was used to make strains mata1^-. The gas1::URA3 plasmid pSL1359 was made by digestion of the URA3 gene from pSP65 with SalI and insertion into the GAS1 gene. Digestion with HindIII and EcoRI liberated the gas1::URA3 fragment. For linkage analysis of PGI1, an EcoRI-SalI fragment directly downstream of the PGI1 gene was cloned from a complementing YEp24 plasmid and introduced into the EcoRI and SalI sites of pRS306 to create pPGI306. Digestion with ClaI was used to target integration directly downstream of PGI1. For linkage analysis of PMI40, a BamHI-SalI genomic fragment from a complementing YEp24 library plasmid was introduced into pRS306 by a directed sticky-end ligation to create pPMI306. The resultant plasmid was digested with ClaI to direct integration directly downstream of PMI40, creating a URA3-marked allele of PMI40. Plasmid pDPM306 was used to target integration of URA3 directly adjacent to DPM1, creating a URA3-marked allele of DPM1.

Mutant isolation and analysis:
Pilot experiments indicated that >30% of His⁺ colonies that were isolated in a ste4 FUS1-HIS3 his3 background also had a slow-growth defect. To gather a collection of independently isolated and spontaneous slow-growth mutants, two single colonies, one MAT (SY2431) the other MATa (SY2428), were inoculated into rich medium and grown in YPD for 16 hr at 30°. Approximately 5 x 10⁸ cells were plated onto individual SD-HIS plates. His⁺ colonies that had a slow-growth phenotype were picked from separate plates and characterized. Mating of GAL-STE4 ste4 strains for dominance/recessive and complementation tests was performed by first incubating cells for 5 hr in 2% galactose to induce the expression of GAL-STE4. Cells were then spotted on YPD plates in appropriate mating mixes for 18 hr and then spread onto selective plates. To facilitate the dominant/recessive test, SY2428 contained a mutation of the MATa1 gene, such that mata1/MAT diploids derived from this strain still expressed haploid-specific genes. A CEN-based plasmid containing the wild-type MATa gene was introduced into diploids to allow for sporulation. For a subset of the mutants, tetrad analysis was performed on synthetic medium due to the growth defect on YPD. For all of the mutants, segregation analysis demonstrated a 2 His⁺:2 His^- ratio, indicating that the defect was due to a single gene. In addition, all the His⁺ spores had a slow-growth defect, which showed that activation of the FUS1 reporter cosegregated with slow growth. Complementation analysis was performed by mating the mutants in the a strain to the mutants in the strain (as above) and assaying the diploids for growth on SD-HIS. Other tests for mating-specific functions were performed essentially as described (SPRAGUE 1991 ), except that for pgi1-101 and pmi40-101 matings were performed in 2% mannose + 0.1% glucose and 2% glucose + 10 mM mannose, respectively. Linkage analysis was used to demonstrate that PGI1, PMI40, and DPM1 were linked to the mutations represented by complementation groups 1, 2, and 3, respectively, in the following manner. Plasmids that contained genomic DNA directly adjacent to PGI1, PMI40, and DPM1 were integrated into the genome of a wild-type strain. The resultant strains were mated to otherwise wild-type pgi1-101, pmi40-101, and dpm1-101 mutants. Diploids were sporulated and tetrad analysis demonstrated that the Ura⁺ phenotype segregated in repulsion to the temperature-sensitive phenotype for pgi1-101, pmi40-101, and dpm1-101. More than 20 tetrads containing four viable spores were assayed for each mutant.

Cloning the genes:
Genes that complemented the slow-growth defect of the mutants were obtained by plasmid complementation using the library of CARLSON and BOTSTEIN 1982 . The DNA sequence of the complementing plasmids was obtained by the dideoxy chain termination method (SANGER et al. 1977 ) using double-stranded DNA templates and the Sequenase Version 2.0 DNA sequencing kit (United States Biochemical, Cleveland). Subcloning of the library plasmids containing PGI1, PMI40, DPM1, and GAS1 was used to show that these were the genes responsible for complementation of the slow-growth and FUS1-activation phenotypes. Additional PMI40 clones were isolated using the yes-galactose-inducible library (ELLEDGE et al. 1991 ). Suppressors of the slow-growth defect of the ste4 dpm1-101 strain were isolated using the yes library and a CEN-based library (ROSE et al. 1990 ; ROSE 1991 ). The yes library yielded RER2, and the CEN library yielded PSA1. Although the strains used in this study are isogenic to Sc252, we initially isolated pgi1, pmi40, and gas1 alleles that activate FUS1-HIS3 from the strain background 246-1-1 (provided by Kelly Tatchell; STEVENSON et al. 1992 ), and in W303, suggesting that the relationship between protein glycosylation and FUS1 activation is not a strain-dependent occurrence.

ß-Galactosidase assays:
For ß-galactosidase assays, cells were prepared and assayed as described previously (JARVIS et al. 1988 ). Cells from an overnight culture were grown for 5 hr in YPD unless otherwise indicated. In the case of strains containing plasmids, cells were grown to saturation in selective medium before growth for 5 hr in YPD. The ß-galactosidase activities reported are the average of three assays; standard deviation was <20%. For experiments involving tunicamycin, which is solubilized in DMSO, we confirmed that addition of DMSO alone had no effect on FUS1-lacZ activity.

Sequence analysis of dpm1-101 and subcellular fractionation of the Dpm1 protein:
The DNA sequence of the dpm1-101 allele was determined by the sequencing of a single sample that contained a mix of five separate polymerase chain reactions (run for 20 cycles) of chromosomal DNA prepared from wild-type cells and from the dpm1-101 mutant using primers that flank the DPM1 gene. The dpm1-101 allele had a single alteration from the wild-type sequence. Subcellular fractionation was performed by low- and high-speed centrifugation of cell lysates (HORAZDOVSKY and EMR 1993 ). Protein fractions were precipitated with trichloracetic acid, run on a 10% SDS-PAGE gel, transferred to nitrocellulose, and probed with monoclonal antibodies specific for Dpm1. The blots were also probed for Pep12 and Vps10 as controls.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

A subset of mutants that activate FUS1 have an associated cell wall integrity defect:
In a previous study, negative regulators of the pheromone response pathway were isolated by a selection for mutants that exhibited autonomous activation of the FUS1 promoter (STEVENSON et al. 1992 , STEVENSON et al. 1995 ). The selection was performed in a ste4 his3::FUS1-HIS3 strain. As a consequence of the ste4 mutation, the strain lacks basal signaling in the pheromone response pathway and the FUS1 promoter is inactive. Hence, the strain is a histidine auxotroph. Selection of His⁺ derivatives identified dominant and recessive mutations that activated pheromone response pathway signaling (STEVENSON et al. 1992 ). The latter mutations are presumed to reveal genes specifying negative regulators of pheromone response pathway signaling. Indeed, one such mutation identified RGA1, which encodes a GAP for Cdc42, and suggested a link between Cdc42 and pheromone pathway components (STEVENSON et al. 1995 ). The ste4 mutation precluded organizing mutants into complementation groups. We therefore elected to carry out a large-scale selection in ste4 his3 FUS1-HIS3 strains that also contain GAL-STE4 (Table 1). Preliminary examination of 150 His⁺ mutants showed that >30% had a slow-growth defect. Ten independently isolated His⁺ colonies with a growth defect were characterized in greater detail. In addition to stimulation of the FUS1-HIS3 reporter, the isolates showed enhanced expression of an integrated FUS1-lacZ reporter, as determined by ß-galactosidase activity (Table 2). To assess whether the phenotype was dominant or recessive and due to a single mutation, the mutants were mated to complementary parent strains. All the diploids were His^-, indicating that the mutations were recessive. Tetrad analysis (of >20 tetrads per mutant) showed that for each mutant a defect in a single gene was responsible for FUS1-HIS3 activation. Furthermore, for all the mutants the growth defect and FUS1-HIS3 activation cosegregated, implying that the same mutation was responsible for both phenotypes. Complementation analysis demonstrated that the mutations were not cell type dependent and that the mutants comprised at least four different complementation groups (Table 2).

We examined the slow-growth phenotype of the mutants in greater detail. FAR1-mediated G₁ arrest likely did not account for the growth defect of the mutants since other mutants that activated the pathway to the same or higher levels did not have a growth defect (STEVENSON et al. 1992 ). The mutants grew more slowly at 37° than at 30°, and they also grew more slowly in YPD than in synthetic medium. Microscopic examination showed that the mutants had a common morphology: mutant cells were larger than wild-type cells, had multiple buds, and had a higher percentage of random or bipolar budding pattern (Table 2). A significant fraction of cells were observed to lyse when incubated at 37° or on YPD. We examined the possibility that cell lysis was due to a cell wall integrity defect by testing the mutants for sensitivity to the cell wall toxin Calcofluor. All the mutants of the slow-growth group were sensitive to Calcofluor (RAM et al. 1994 ; Table 2). However, not all His⁺ mutants were sensitive to Calcofluor or exhibited the altered morphology (e.g., rga1; STEVENSON et al. 1995 , and pce1; P. J. CULLEN and G. F. SPRAGUE, unpublished results; see Table 2), suggesting that the slow-growth mutants define a class of negative regulators of the pheromone response pathway. Strains containing the previously characterized STE11-4 hyperactive allele (STEVENSON et al. 1992 ) were also found to be sensitive to Calcofluor (Table 2). Because these mutants show increased expression of FUS1, a gene whose transcription is influenced by the pheromone response pathway but not other known pathways, we refer to the pathway activated in the mutants as the pheromone response pathway. However, this interpretation is revisited in the DISCUSSION.

Defects in mannose utilization and protein glycosylation cause activation of FUS1 transcription:
PMI40: To identify the genes responsible for the mutant phenotype, we selected for complementation of the growth defect after transformation with a plasmid library. A representative mutant from complementation group 1 was transformed with a YEP24-based yeast genomic library (CARLSON and BOTSTEIN 1982 ), and transformants were selected for wild-type growth at 37°. Three independent isolates rescued the growth defect and also conferred a His^- phenotype to group 1 mutants (Fig 2). These plasmids contained overlapping genomic DNA fragments, each of which included the PMI40 gene. Subsequent deletion analysis revealed that PMI40 was indeed the complementing gene. To determine whether PMI40 corresponds to group 1, PMI40 DNA was cloned into pRS306 and the resulting construct was integrated at the PMI40 locus by homologous recombination. In a genetic cross, the marked PMI40 and a group 1 mutation segregated as alleles (no recombinants in >20 tetrads; see MATERIALS AND METHODS), demonstrating that a defect in the PMI40 gene was indeed responsible for FUS1 reporter activation.

View larger version (12K): In this window In a new window Download PPT slide   Figure 1. Defects in mannose utilization and protein glycosylation lead to pheromone response pathway activation. Partial loss-of-function alleles of each gene shown in bold stimulated the FUS1-lacZ reporter in a ste4 background. Abbreviations are as follows: Glc, glucose; Fru, fructose; Man, mannose; P, phosphate; Dol, dolichol; GPI, glycophospholipid. Double arrows indicate that a number of enzymes are required for the given process. High-copy suppressors of the dpm1-101 allele that also suppress pheromone response pathway activation are underlined. GAS1 is the only target shown of the many targets of the N- and O-linked glycosylation and GPI anchor pathways.

View larger version (26K): In this window In a new window Download PPT slide   Figure 2. The pmi40-101 allele activates FUS1-HIS3 in a ste4 background. Identical numbers of cells containing the indicated plasmids were spotted onto synthetic medium lacking uracil and with the following additional conditions: box 1, no modification; box 2, 37°; box 3, plus 10 mM mannose at 37°; box 4, minus histidine; box 5, minus histidine plus 10 mM mannose.

PMI40 encodes mannose-6-phosphate isomerase (SMITH et al. 1992 ), which is required for synthesis of mannosyl moieties, important constituents of glycosylated proteins, and is therefore an essential gene (Fig 1). Addition of 10 mM mannose to growth media obviates the PMI40 essential function (SMITH et al. 1992 ). We therefore assessed the effect of mannose supplementation on the phenotypes of the pmi40 allele isolated here, designated pmi40-101. ste4 pmi40-101 cells were able to grow at 37° on synthetic medium supplemented with 10 mM mannose, but they were histidine auxotrophs (Fig 2). That is, mannose supplementation rescued both the growth and pheromone pathway phenotypes of the mutants. To quantify the effect of mannose on FUS1 transcription, we measured expression of FUS1-lacZ. The ste4 pmi40-101 strain grown in the presence of 10 mM mannose had a ß-galactosidase activity identical to that found in the ste4 parent strain (Table 3). Curiously, ste4 pmi40-101 cells grown in YPD had greater ß-galactosidase activity than the same cells grown in synthetic glucose medium (Table 3). Expression of FUS1-lacZ in wild-type or ste4 cells was not affected by mannose or by YPD. Together, these experiments imply that it is loss of Pmi40 enzymatic activity, not loss of an uncharacterized biochemical activity of the protein, that leads to activation of the FUS1 reporters.

PGI1: Transformation of a group 2 mutant with the genomic library yielded independent complementing plasmids that contained the PGI1 gene. Deletion mapping and subcloning verified that PGI1 was the complementing gene, and integrative mapping showed that PGI1 segregated as an allele of group 2 mutations. PGI1 encodes glucose-6-phosphate isomerase, which converts glucose-6-phosphate to fructose-6-phosphate (AGUILERA and ZIMMERMANN 1986 ). Synthetic medium containing fructose as a carbon source (SF) supplemented with 5 mM glucose was previously shown to restore viability to pgi1 mutants (AGUILERA 1986 ). We showed that incubation in SF + 5 mM glucose medium suppressed both the growth defect and the His⁺ phenotype of ste4 pgi1-101 (not shown), suggesting that, as for Pmi40, it is loss of (or reduced) Pgi1 enzyme activity that leads to increased transcription from the FUS1 promoter. Synthetic medium containing mannose as a carbon source and supplemented with 5 mM glucose also suppressed both phenotypes (see Fig 1). In addition, we observed an increase in FUS1-lacZ activity for ste4 pgi1-101 mutants grown in YPD as compared to synthetic medium, as observed for ste4 pmi40-101 cells (Table 3). The growth defects of mutants in the other complementation groups were rescued by neither SF + 5 mM glucose nor SD +10 mM mannose, suggesting that they did not owe their phenotype to diminished Pmi40 or Pgi1 activity. Thus, defects in 2-hexose isomerases that operate at the head of the mannose utilization pathway stimulated FUS1 expression. Basal FUS1-lacZ activity in pmi40 and pgi1 strains containing an intact pheromone pathway (STE4) was more than twofold higher than in wild-type cells, which suggests that there is a physiological role for pheromone response pathway activation in mannose utilization mutants (see also the tunicamycin experiment below).

DPM1: Complementation of the slow-growth phenotype of a group 3 mutant yielded sets of plasmids from three distinct regions of the genome. One set of plasmids contained the DPM1 gene, and tetrad analysis showed that complementation group 3 was linked to DPM1 (see MATERIALS AND METHODS). DPM1 is required for synthesis of dolichol-P-mannose from dolichol and GTP-mannose (ORLEAN et al. 1988 ; ORLEAN 1990 ; Fig 1). A plasmid containing DPM1 conferred wild-type growth and a His^- phenotype to all three mutants in this complementation group, and we therefore chose the strongest mutant in terms of FUS1-lacZ activation, dpm1-101, for further studies. As for the hexose isomerase mutants, FUS1-lacZ activity in ste4 dpm1-101 was higher in YPD than in synthetic medium (Table 3). To ascertain the nature of the defect of the Dpm1 protein, we examined dpm1-101 in more detail. Western analysis using monoclonal antibodies to Dpm1 demonstrated that the amount and size of Dpm1 was identical in wild-type and dpm1-101 strains (not shown). Subcellular fractionation profiles were also identical for strains containing wild-type and mutant Dpm1 proteins and were in accord with the reported endoplasmic reticulum localization of Dpm1 (not shown). The DNA sequence of dpm1-101 contained a single nucleotide substitution (A to G) predicted to result in a glycine-to-serine substitution at amino acid 205. This position is adjacent to the predicted dolichol-binding site of Dpm1 (ALBRIGHT et al. 1989 ; ORLEAN 1992 ). Because the dpm1-101 allele was suppressed by overexpression of genes that provide either of the substrates for Dpm1 (see below), and because the level and localization of Dpm1G205S was identical to wild type, we suggest that Dpm1G205S has a defect in substrate binding or in the ligation of dolichol to mannose.

The screen for library plasmids that complemented the dpm1-101 mutation yielded plasmids from two genomic regions distinct from the DPM1 locus. One overlapping set of these suppressing plasmids contained the PSA1 gene, which is required for the synthesis of GTP-mannose, one of the substrates for the Dpm1 enzyme (HASHIMOTO et al. 1997 ). The second set of plasmids contained RER2, whose product is important for dolichol synthesis (SATO et al. 1999 ), the other substrate for Dpm1. In addition to suppression of the growth defect of ste4 dpm1-101, the RER2- and PSA1-containing plasmids suppressed the His⁺ and FUS1-lacZ activation phenotypes (Fig 3). We have also shown that a temperature-sensitive allele of the PSA1 gene is capable of causing FUS1-lacZ activation in a ste4 strain at the nonpermissive temperature (not shown).

View larger version (29K): In this window In a new window Download PPT slide   Figure 3. Suppressors of the ste4 dpm1-101 slow-growth phenotype also prevent FUS1 transcription. Identical numbers of cells containing the indicated plasmids were spotted onto synthetic medium lacking uracil (left), YPD plates at 37° (center), and synthetic medium lacking both uracil and histidine (right). ß-Galactosidase activity was determined as described in MATERIALS AND METHODS. The reported values are the average of three independent determinations.

Other protein glycosylation steps: We hypothesized that mutants defective in subsequent steps of protein glycosylation might also show activation of the FUS1 reporters, and we found support for this hypothesis in the following four experiments: (1) The OCH1 gene, although not essential, encodes a protein that catalyzes mannose outer chain elongation of N-linked oligosaccharides. Deletion of OCH1 in a ste4 background stimulated FUS1-lacZ expression by >15-fold (Table 4) and supported growth on SD-His medium. The ste4 och1 strain was sensitive to Calcofluor and grew more slowly on YPD than on synthetic medium. Microscopic examination of the ste4 och1 strain revealed the same morphology as that of pgi1, pmi40, and dpm1 strains: cells were larger than wild type, and were observed to lyse in YPD or Calcofluor. (2) Loss of ALG1 and MNN10, which are also required for N-linked glycosylation, activated FUS1 transcription in ste4 cells (Table 4). (3) The consequence of diminished GPI anchor addition was investigated by use of a temperature-sensitive allele of SPT14, which is required for the first step of GPI anchor addition (for review see TAKEDA and KINOSHITA 1995 ). The FUS1-lacZ gene was integrated into a strain containing a temperature-sensitive allele of SPT14, spt14-2, and the STE4 gene was deleted. The spt14-2 allele stimulated FUS1-lacZ by more than fourfold (even at 30°), and the strain was sensitive to Calcofluor (Table 4). (4) We hypothesized that inhibitors of protein glycosylation, such as the drug tunicamycin (KUO and LAMPEN 1974 ), might also activate FUS1. Tunicamycin prevents N-linked glycosylation by acting as an inhibitor of the Alg7 protein (LEHLE and TANNER 1976 ; KUKURUZINSKA and LENNON 1995 ). We found that addition of tunicamycin caused a >20-fold increase in FUS1-lacZ activity in ste4 cells (Table 4), comparable to loss of OCH1. Enhanced FUS1-lacZ expression was observed by the addition of tunicamycin to cells in YPD, reminiscent of the increased expression of FUS1-lacZ observed for pgi1, pmi40, and dpm1 mutants. In wild-type cells grown on YPD medium, tunicamycin caused an approximately twofold increase in FUS1-lacZ expression. Therefore, FUS1 transcription was activated by mutants defective in protein glycosylation and by tunicamycin, an inhibitor of protein glycosylation. Taken together, these results show that obstruction of the protein glycosylation machinery itself causes activation of the pheromone response pathway.

Loss of Gas1, a target of the protein glycosylation pathway, causes activation of FUS1:
Isolation of plasmids that complement the slow-growth phenotype of a fourth mutant revealed that this group corresponds to GAS1. GAS1 encodes an abundant and heavily N- and O-glycosylated protein that is the predominant GPI-anchored cell surface protein and is thus a major target of protein glycosylation (POPOLO et al. 1997 ; Fig 1). Loss of GAS1 results in the release of ß-1,3-glucan into the medium and compromises the integrity of the cell wall, as shown in part by its synthetic lethality with loss of PKC1 and KRE6 (POPOLO et al. 1997 ; RAM et al. 1998 ; POPOLO and VAI 1999 ). As measured by degree of resistance to AT or by expression of FUS1-lacZ, the gas1-101 allele caused only a modest activation of the FUS1 reporters. GAS1 is not an essential gene, and so we constructed a deletion allele to examine its effects. The phenotype of the ste4 gas1 strain was indistinguishable from that of the ste4 gas1-101 strain (Table 5). The ste4 gas1 strain had a similar morphology to other protein glycosylation mutants that activate FUS1, and gas1 was sensitive to Calcofluor. To address the possibility that loss of Gas1 function in some way precludes greater activation of the FUS1 reporters, we measured ß-galactosidase expression in a ste4 gas1 dpm1-101 strain. As seen in Table 5, ß-galactosidase activity in this strain was identical to that of a ste4 dpm1-101 strain. One interpretation of these results is that the presumed global perturbation of cell wall structure caused by dpm1-101 has a greater effect on pathway signaling than does the loss of a single cell wall protein.

View this table: In this window In a new window   Table 5. Loss of GAS1 activates FUS1 reporters in a ste4 strain

Components required for activation of FUS1 transcription in protein glycosylation mutants:
To determine the STE gene requirements for activation of the FUS1 reporters in protein glycosylation mutants, ste mutations were introduced into the dpm1-101 and pmi40-101 strains. These strains were chosen because they show the highest FUS1-lacZ activity. Activation of FUS1-lacZ by dpm1-101 was completely dependent upon STE50, STE20, the MAP kinase components STE11, STE7, and KSS1, and upon the STE12 transcription factor (Table 6). Loss of FUS3, STE4, or STE5 did not have an effect on FUS1 transcription in dpm1-101 cells, suggesting that proteins whose known functions are limited to the pheromone response pathway are not required for signaling to FUS1 in the glycosylation mutants. Similar results were obtained for pmi40-101 (not shown). Since STE4 and STE5 are haploid-specific genes and are not required for FUS1 transcription in glycosylation mutants, we asked whether any haploid function was required by assessing FUS1 expression in diploids defective for protein glycosylation. In diploids, the basal level of FUS1 expression is dramatically reduced due to repression of haploid- and mating-specific genes (TRUEHEART and FINK 1989 ; MCCAFFREY et al. 1987 ). In the absence of mannose, the FUS1-HIS3 and FUS1-lacZ reporters were activated in a homozygous pmi40-101 diploid, as indicated by growth on media lacking histidine, as well as by a >15-fold increase in FUS1-lacZ activity. Neither reporter was activated in the homozygous pmi40-101 diploid in medium supplemented with 10 mM mannose. Enhanced transcription of FUS1 was also observed in homozygous pgi1-101 diploid cells grown in the absence of mannose, whereas no significant increase in FUS1 transcription was observed in wild-type diploids grown with or without mannose (data not shown).

The haploid invasive growth pathway requires the same subset of STE genes as are required for activated FUS1 transcription in the protein glycosylation mutants (ROBERTS and FINK 1994 ; TEDFORD et al. 1997 ). We therefore tested the role of genes thought to be involved in invasive growth, not mating. Loss of Tec1, a transcription factor required for invasive growth but not mating or FUS1 transcription (MADHANI et al. 1997 ), had no effect on FUS1-lacZ expression in either dpm1 or pmi40 cells (Table 6; data not shown). Likewise, loss of Hsl7, a negative regulator of Ste20 function in invasive growth (FUJITA et al. 1999 ), had no effect on FUS1-lacZ expression in dpm1-101 strains (Table 6), which further supports the fact that components specific for the invasive growth pathway are not required for signaling to FUS1.

In some HOG pathway mutants, namely pbs2 and hog1 mutants, Ste20-dependent cross talk to the pheromone response pathway is observed (O'ROURKE and HERSKOWITZ 1998 ; SPRAGUE 1998 ). We therefore asked whether loss of either the Ssk1 (MAEDA et al. 1994 ) or the Sho1 (MAEDA et al. 1995 ; POSAS and SAITO 1997 ) branch of the HOG pathway affected activation of FUS1 reporters. We found that SHO1 was absolutely required for signaling to FUS1 in dpm1-101 ste4 cells (Table 6) and in pmi40-101 ste4 cells (not shown). Loss of SHO1 did not affect signaling in wild-type or ste4 cells, consistent with previously published evidence that SHO1 is not required for FUS1-lacZ transcription or for mating (data not shown; O'ROURKE and HERSKOWITZ 1998 ). Deletion of SSK1 had no effect on signaling in ste4 dpm1-101 (Table 6) or ste4 pmi40-101 cells (not shown).

Loss of STE pathway components exacerbates the growth defect of protein glycosylation mutants:
In the course of the epistasis analysis, we noticed that deletion of STE20, STE11, or STE12 exacerbated the growth defect of pmi40-101 mutants (Fig 4), whereas deletion of FUS3, STE4, or STE5 (which are not required for FUS1 signaling by pmi40-101) had no effect on cell growth (not shown). The growth defect was completely rescued by supplementation with 10 mM mannose, which bypasses the requirement for Pmi40. Loss of STE11 in pmi40-101 cells caused a slightly more severe growth defect than loss of the other STE genes, suggesting that STE11 has a function in pmi40-101 cells in addition to its participation in the pheromone and invasive growth pathways, perhaps as a participant in the HOG pathway (see below). Loss of STE gene function also conferred a synthetic growth defect in the dpm1-101 background. The synthetic defect was less severe than in the pmi40-101 background, perhaps because pmi40-101 alone has a more severe growth defect than does dpm1-101 (Table 2). We also found that the previously characterized hyperactive STE11-4 allele caused Calcofluor sensitivity and lysis (Table 2). Epistasis analysis showed that STE7, FUS3/KSS1, and STE12 (but not STE4 and STE5) were required to mediate Calcofluor sensitivity in a STE11-4 strain (not shown). These sets of results imply that pheromone response pathway components participate in maintaining cell wall integrity in vegetative cells.

View larger version (82K): In this window In a new window Download PPT slide   Figure 4. Synthetic interactions of pmi40. Equal concentrations of cells were spotted onto synthetic media containing 2% glucose + 10 mM mannose at 30° (left), 2% glucose at 30° (middle), or 2% glucose at 37° (right).

Intact protein kinase C and HOG pathways are essential in protein glycosylation mutants:
Given that the protein glycosylation mutants exhibit a propensity to lyse and given that Ste20, Ste50, and Ste11 have a known role in the HOG pathway as well as in the pheromone and invasive growth pathways, we postulated that glycosylation pathway mutations would show synthetic interactions with the HOG pathway mutations. Indeed, loss of PBS2 caused a growth defect in pmi40-101 cells (Fig 4). The growth defect of pbs2 pmi40-101 cells was comparable to ste20 pmi40-101 or ste12 pmi40-101 cells, but less severe than that of ste11 pmi40-101 cells. Together, these results imply that components from both the pheromone/invasive growth pathway (Ste7, Kss1, and Ste12) and the HOG pathway (Pbs2) are required to maintain viability in protein glycosylation mutants.

We also hypothesized that the protein kinase C (PKC) pathway, which controls cell wall integrity and can be modulated by the pheromone response pathway (YASHAR et al. 1995 ; BUEHRER and ERREDE 1997 ), might have a positive function in mutants compromised for protein glycosylation. As in the pheromone and HOG signaling pathways, a MAP kinase cascade is required for PKC-dependent signaling (LEE and LEVIN 1992 ; IRIE et al. 1993 ; MAZZONI et al. 1993 ). Loss of the MAPKKK Bck1 or of the MAPK Slt2 was lethal in a pmi40 strain (Fig 4; data not shown). Thus, loss of PKC pathway components caused a more severe growth defect than loss of the HOG pathway components in pmi40-101 cells. Deletion of either Bck1 or Slt2 exacerbated the slow-growth phenotype of dpm1-101 cells. However, the double mutants were viable, which allowed FUS1-lacZ analysis in these strains. Loss of BCK1 or SLT2 decreased FUS1-lacZ expression in the dpm1-101 ste4 background by >10-fold, and in dpm1-101 by >2-fold (Table 7). This latter result suggests that the PKC pathway may influence signaling in the pheromone response pathway in wild-type cells. As another means to test this possibility, we used a hyperactive allele of BCK1, BCK1-20 (LEE and LEVIN 1992 ). Plasmid-borne BCK1-20 stimulated the FUS1-lacZ reporter by more than threefold in the wild-type strain (Table 7). In addition, BCK1-20 caused enhanced growth of wild-type cells on medium lacking histidine, presumably due to enhanced expression of the FUS1-HIS3 reporter. Epistasis analysis showed Ste11, Ste7, and Ste12 were required for the BCK1-20 effect (not shown). Thus, the PKC pathway is required for viability, and may contribute to FUS1 expression, in protein glycosylation mutants.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Defects in protein glycosylation activate FUS1 transcription:
We have presented several lines of evidence that defects in protein glycosylation activate the pheromone response pathway reporter FUS1. First, partial loss-of-function mutants that compromise early (pgi1, pmi40, psa1, and dpm1) or late (och1, alg1, mnn10, and spt14) steps in protein glycosylation stimulate expression of FUS1 in ste4 mutants. We suggest that mutations in other genes required for protein glycosylation, such as SEC53 and genes required for O-linked glycosylation, would also stimulate FUS1 transcription, although these possibilities have not been tested. Second, supplementation with mannose can suppress both the slow-growth and FUS1 transcription phenotypes of pgi1-101 and pmi40-101 mutants, presumably by providing an alternate route to synthesis of mannose-6-P. Third, multicopy plasmid suppressors of the growth defects of dpm1-101 mutants (PSA1 and RER2) also suppressed FUS1 expression. We have isolated multicopy suppressors of the growth defect of pmi40-101 cells, and these also significantly reduce FUS1 expression (P. J. CULLEN and G. F. SPRAGUE, unpublished results). Fourth, tunicamycin, which prevents N-linked glycosylation, stimulates FUS1 transcription. It has been shown that defects in protein glycosylation also induce the unfolded protein response (UPR; for review see SIDRAUSKI et al. 1998 ), which results in an accumulation of endoplasmic reticulum-specific chaperones. Although we favor the hypothesis that protein glycosylation defects lead to Sho1 activation, it is conceivable that the UPR influences FUS1 transcription in the protein glycosylation mutants.

The results of LEE and ELION 1999 also support a function for STE pathway components in protein glycosylation mutants. Our results are largely consistent with theirs but differ in three notable points. First, we have shown that defects in any part of the protein glycosylation pathway (not simply OCH1) lead to activation of the pheromone response pathway (see Fig 1). Second, we have shown that Sho1, an osmosensor in the HOG pathway, is an essential component of this Ste-dependent signaling cascade (Table 6). Third, we have shown that pathway activation is not constitutive, but rather dependent upon the osmotic environment of the cell, an observation that is consistent with the requirement for the Sho1 osmosensor in this pathway (Table 3 and see below).

Components from the HOG, pheromone response, and invasive growth pathways are required for signaling in protein glycosylation mutants:
Activation of FUS1 transcription in protein glycosylation mutants requires proteins known to operate in one or more signal transduction pathways: the pheromone response pathway, the HOG pathway, and the invasive growth pathway. On the basis of the known relationships of these proteins, we suggest the following amalgamated pathway: Sho1-Ste20/Ste50-Ste11-Ste7-Kss1-Ste12 (see Fig 5). This genetic pathway most closely resembles the invasive growth pathway, especially since Sho1 has been implicated in filamentous growth (O'ROURKE and HERSKOWITZ 1998 ). For example, KSS1 is known to be exclusively required for the invasive pathway (COOK et al. 1997 ; MADHANI et al. 1997 ), and we have shown that in protein glycosylation mutants activation of FUS1 transcription is also mediated by KSS1 and not FUS3, consistent with observations by LEE and ELION 1999 . However, the two pathways differ in several important respects. First, Hsl7 and Tec1, which are involved in the invasive growth pathway (FUJITA et al. 1999 and MADHANI et al. 1997 , respectively), have no discernible effect on FUS1 transcription in protein glycosylation mutants. Second, activation of the invasive growth response has not been observed to activate FUS1 transcription (MADHANI et al. 1997 ; our unpublished observations). Third, the glycosylation mutants do not exhibit physiological changes associated with activation of the invasive growth pathway, namely elongated cell morphology and unipolar budding pattern. Likewise, although the pathway we have deduced includes proteins that function in HOG and pheromone response, not all proteins from these pathways are required for the transcription effect we observe. Specifically, Ste4, Ste5, and Fus3 from the pheromone pathway and Pbs2 from the HOG pathway are not required.

View larger version (17K): In this window In a new window Download PPT slide   Figure 5. A Sho1 Ste12 signaling pathway. The inputs, signal transducers, and outputs of each Ste11-dependent process are shown. Orange represents components specific for the pheromone pathway, green represents components for the HOG pathway, and dark yellow depicts components for the invasive growth pathway. Proteins in purple are used in each of the pathways.

The requirement of Sho1 for FUS1 transcription in protein glycosylation mutants is notable for several reasons. First, a link from Sho1 through Ste7, Kss1, and Ste12 to FUS1 transcription has not been observed previously in cells containing wild-type pathways, although cross talk from the HOG pathway to FUS1 transcription can occur in pbs2 and hog1 mutants (O'ROURKE and HERSKOWITZ 1998 ). Second, the requirement for Sho1, a predicted glycosylated membrane protein, provides a possible mechanism by which defects in protein glycosylation activate FUS1 transcription. Defects in glycosylation lead to changes in the osmotic sensitivity of the cell, thereby activating Sho1. Whether FUS1 (and other pheromone responsive genes) is an intended target when Sho1 is activated in this way or whether its activation represents loss of pathway specificity is not known. Third, the requirement for SHO1 provides an explanation as to why Ste11 is in both the pheromone and HOG pathways: to simultaneously activate both pathways under special conditions (see below).

Osmosensitivity in protein glycosylation mutants causes Sho1-dependent activation of FUS1:
Defects in protein glycosylation components result in an osmosensitive cell (e.g., STATEVA et al. 1991 ). We have been able to partially rescue the slow-growth phenotype of protein glycosylation mutants (including pmi40-101 and och1) by the addition of salt (NaCl or KCl) to the medium. Addition of salt caused a concomitant reduction in FUS1-lacZ activity of more than fivefold (not shown). In addition, growth of mutants in YPD caused a more severe growth defect in protein glycosylation mutants, which correlates with an increase in expression of FUS1-lacZ. Thus, it appears that FUS1 expression responds to the general osmotic environment in these mutants. These observations may also rationalize why loss of a target of the protein glycosylation pathway, GAS1, causes activation of FUS1, since gas1 mutants are known to have severe cell wall and osmolarity defects (POPOLO et al. 1997 ). Together, these results lead us to hypothesize that Sho1 is activated in response to the osmotic stress these mutants suffer.

Defects in protein glycosylation lead to Sho1-dependent activation of the pheromone-responsive FUS1 gene, whereas other treatments that activate Sho1—addition of osmolyte to the medium or nutrient limitation—do not have this effect. We do not know the reason for this difference, but two possibilities merit consideration. One possibility is that the glycoslyation defect in some way causes a loss of pathway specificity and in this sense the effect on FUS1 transcription is unintended. For example, mutation of Hog1 allows cross talk between the HOG and pheromone pathways (O'ROURKE and HERSKOWITZ 1998 ). Perhaps Hog1 is inactivated in the protein glycosylation mutants. We suggest, however, that the effect is intended. Among the genes whose transcription is activated in response to pheromone are FKS2 (a component of ß-1,3-glucan synthase; MAZUR et al. 1995 ) and CHS1 (a chitin synthase; APPELTAUER and ACHSTETTER 1989 ; DIAZ et al. 1992 ; PHILIPS and HERSKOWITZ 1997 ; CAPPELLARO et al. 1998 ), which function in cell wall biosynthesis. Increased expression of these genes in protein glycosylation mutants would provide one means for the cell to compensate for the cell wall defect. Indeed, increased FKS2 expression has been observed in och1 mutants (LEE and ELION 1999 ). Ste12 may be targeted to promoters of cell wall genes in part by heterodimerization with Mcm1, a DNA-binding protein whose targets include cell wall genes and whose phosphorylation state is sensitive to osmolarity (KUO et al. 1997 ). Thus, the Sho1-Ste12 link provides a way to tie changes in cell wall structure/function detected by Sho1 to expression of genes needed to strengthen or remodel the cell wall. Furthermore, it seems likely that this link is important physiologically in wild-type cells given the synthetic interactions we observe between Sho1-Ste12 pathway mutations and protein glycosylation mutations. Thus, the Sho1-Ste12 pathway may be a signaling pathway that communicates changes in protein glycosylation and thereby adjusts accordingly the transcription of cell wall integrity components.

	FOOTNOTES

¹ Present address: Oregon State Police, Forensic Sciences Division, 1111 S.W. 2nd Ave., Portland, OR 97204-3258.
² Present address: Ludwig Institute for Cancer Research Office of Information Technology, CH-1066 Epalinges, Switzerland.

	ACKNOWLEDGMENTS

We thank Tom Stevens for monoclonal antibodies and helpful advice; and David Levin, Ira Herskowitz, Gerald Fink, Steve Oliver, Jan Fassler, and Sabine Strahl-Bolsinger for strains and plasmids. We thank Bee Na Lee and Elaine Elion for sharing results before publication. Thanks to Aaron Rogat and Christa Fink for preliminary mutant characterization. Thanks also to Sean O'Rourke, Greg Smith, David Mitchell, April Goehring, Hilary Kemp, Megan Keniry, David Rivers, Sonya Carlson, and Elizabeth Monika for helpful comments and suggestions. This work was supported by research (GM-30027 to G.F.S.) and training (GM19188 for P.J.C. and GM14591 for J.S.) grants from the U.S. Public Health Service. J.H. was supported by a U.S.-National Science Foundation/Japan-STA International Research Fellowship (INT-9600341), and Y.J. was supported by Japan STA Center of Excellence and National Institute of Bioscience and Human Technology grants.

Manuscript received February 23, 2000; Accepted for publication March 13, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AGUILERA, A., 1986  Deletion of the phosphoglucose isomerase structural gene makes growth and sporulation glucose dependent in Saccharomyces cerevisiae.. Mol. Gen. Genet. 204:310-316[Medline].

AGUILERA, A. and F. K. ZIMMERMANN, 1986  Isolation and molecular analysis of the phosphoglucose isomerase structural gene of Saccharomyces cerevisiae.. Mol. Gen. Genet. 202:83-89[Medline].

ALBRIGHT, C. F., P. ORLEAN, and P. W. ROBBINS, 1989  A 13-amino acid peptide in three yeast glycosyltransferases may be involved in dolichol recognition. Proc. Natl. Acad. Sci. USA 86:7366-7369[Abstract/Free Full Text].

APPELTAUER, U. and T. ACHSTETTER, 1989  Hormone-induced expression of the CHS1 gene from Saccharomyces cerevisiae.. Eur. J. Biochem. 181:243-247[Medline].

BARDWELL, L., J. G. COOK, C. J. INOUYE, and J. THORNER, 1994  Signal propagation and regulation in the mating pheromone response pathway of the yeast Saccharomyces cerevisiae.. Dev. Biol. 166:363-379[Medline].

BUEHRER, B. M. and B. ERREDE, 1997  Coordination of the mating and cell integrity mitogen-activated protein kinase pathways in Saccharomyces cerevisiae.. Mol. Cell. Biol. 17:6517-6525[Abstract].

BREWSTER, J. L., T. DE VALOIR, N. D. DWYER, E. WINTER, and M. C. GUSTIN, 1993  An osmosensing signal transduction pathway in yeast. Science 259:1760-1763[Abstract/Free Full Text].

CAPPELLARO, C., V. MRSA, and W. TANNER, 1998  New potential cell wall glucanases of Saccharomyces cerevisiae and their involvement in mating. J. Bacteriol. 180:5030-5037[Abstract/Free Full Text].

CARLSON, M. and D. BOTSTEIN, 1982  Two differentially regulated mRNAs with different 5' ends encode secreted with intracellular forms of yeast invertase. Cell 28:145-154[Medline].

CHANT, J., 1999  Cell polarity in yeast. Annu. Rev. Cell Dev. Biol. 15:365-391[Medline].

COOK, J. G., L. BARDWELL, and J. THORNER, 1997  Inhibitory and activating functions for MAPK Kss1 in the S. cerevisiae filamentous-growth signalling pathway. Nature 390:85-88[Medline].

DEAN, N. and J. B. POSTER, 1996  Molecular and phenotypic analysis of the S. cerevisiae MNN10 gene identifies a family of related glycosyltransferases. Glycobiology 6:73-81[Abstract/Free Full Text].

DIAZ, S., S. ZINKER, and J. RUIZ-HERRERA, 1992  Alterations in the cell wall of Saccharomyces cerevisiae induced by the alpha sex factor or a mutation in the cell cycle. Antonie Van Leeuwenhoek 61:269-276[Medline].

ELION, E. A., P. L. GRISAFI, and G. R. FINK, 1990  FUS3 encodes a cdc2+/CDC28-related kinase required for the transition from mitosis into conjugation. Cell 60:649-664[Medline].

ELLEDGE, S. J., J. T. MULLIGAN, S. W. RAMER, M. SPOTTSWOOD, and R. W. DAVIS, 1991  Lambda YES: a multifunctional cDNA expression vector for the isolation of genes by complementation of yeast and Escherichia coli mutations. Proc. Natl. Acad. Sci. USA 88:1731-1735[Abstract/Free Full Text].

ERREDE, B., R. M. CADE, B. M. YASHAR, Y. KAMADA, and D. E. LEVIN et al., 1995  Dynamics and organization of MAP kinase signal pathways. Mol. Reprod. Dev. 42:477-485[Medline].

FIELDS, S. and I. HERSKOWITZ, 1987  Regulation by the yeast mating-type locus of STE12, a gene required for cell-type-specific expression. Mol. Cell. Biol. 7:3818-3821[Abstract/Free Full Text].

FUJITA, A., A. TONOUCHI, T. HIROKO, F. INOSE, and T. NAGASHIMA et al., 1999  Hsl7p, a negative regulator of Ste20p protein kinase in the Saccharomyces cerevisiae filamentous growth-signaling pathway. Proc. Natl. Acad. Sci. USA 96:8522-8527[Abstract/Free Full Text].

GIETZ, R. D., R. H. SCHIESTL, A. R. WILLEMS, and R. A. WOODS, 1995  Studies on the transformation of intact yeast cells by the LiAc/SS-DNA/PEG procedure. Yeast 11:355-360[Medline].

GIMENO, C. J., P. O. LJUNGDAHL, C. A. STYLES, and G. R. FINK, 1992  Unipolar cell divisions in the yeast S. cerevisiae lead to filamentous growth: regulation by starvation and RAS. Cell 68:1077-1090[Medline].

HASHIMOTO, H., A. SAKAKIBARA, M. YAMASAKI, and K. YODA, 1997  Saccharomyces cerevisiae VIG9 encodes GDP-mannose pyrophosphorylase, wh