The Filamentous Growth MAPK Pathway Responds to Glucose Starvation Through the Mig1/2 Transcriptional Repressors in Saccharomyces cerevisiae
Sheelarani Karunanithi, Paul J. Cullen

Abstract

In the budding yeast S. cerevisiae, nutrient limitation induces a MAPK pathway that regulates filamentous growth and biofilm/mat formation. How nutrient levels feed into the regulation of the filamentous growth pathway is not entirely clear. We characterized a newly identified MAPK regulatory protein of the filamentous growth pathway, Opy2. A two-hybrid screen with the cytosolic domain of Opy2 uncovered new interacting partners including a transcriptional repressor that functions in the AMPK pathway, Mig1, and its close functional homolog, Mig2. Mig1 and Mig2 coregulated the filamentous growth pathway in response to glucose limitation, as did the AMP kinase Snf1. In addition to associating with Opy2, Mig1 and Mig2 interacted with other regulators of the filamentous growth pathway including the cytosolic domain of the signaling mucin Msb2, the MAP kinase kinase Ste7, and the MAP kinase Kss1. As for Opy2, Mig1 overproduction dampened the pheromone response pathway, which implicates Mig1 and Opy2 as potential regulators of pathway specificity. Taken together, our findings provide the first regulatory link in yeast between components of the AMPK pathway and a MAPK pathway that controls cellular differentiation.

IN eukaryotes, evolutionarily conserved signaling pathways regulate the cellular response to nutrients (Mantovani and Roy 2011). In the budding yeast Saccharomyces cerevisiae, nutrient sensing and regulation have been extensively studied, and receptors that sense nutrients and their cognate pathways have been defined (Dechant and Peter 2008; Xue et al. 2008; Ljungdahl 2009). Glucose is the preferred carbon source in yeast, and its detection and metabolism are highly regulated (Gancedo 2008; Zaman et al. 2008). Among the pathways that control the cellular response to glucose levels are the Snf3/Rgt2 pathway (Ozcan and Johnston 1999), the Ras2-cAMP-protein kinase A (PKA) pathway (Thevelein and Voordeckers 2009), and a glucose repression pathway that is regulated by the ubiquitous nutrient-sensing AMP-dependent kinase (AMPK) Snf1 (Carlson 1999). Snf3 and Rgt2 are glucose sensors that regulate the expression of hexose transporters (Hxt) of appropriate affinities through the transcriptional repressor Rgt1 (Ozcan and Johnston 1996; Sabina and Johnston 2009). Gpr1 is another glucose sensor (Nakafuku et al. 1988; Kraakman et al. 1999; Harashima and Heitman 2002; Peeters et al. 2007; Thevelein and Voordeckers 2009) that, together with the major nutrient regulatory GTPase Ras2 (Kataoka et al. 1984), converges on adenylate cyclase to regulate cellular cAMP levels and PKA activity (Toda et al. 1987; Robertson and Fink 1998; Pan and Heitman 1999; Robertson et al. 2000). Communication and signal integration among the different pathways results in a unified response to fluctuating nutrient levels (Kaniak et al. 2004; Kim and Johnston 2006).

Under nutrient-limiting conditions, Snf1 regulates the utilization of poor carbon sources (Celenza and Carlson 1989; Woods et al. 1994; Lesage et al. 1996; McCartney and Schmidt 2001). One target of Snf1 is the transcriptional repressor Mig1. Snf1 phosphorylates Mig1 (Ostling and Ronne 1998; Treitel et al. 1998; Smith et al. 1999), which relieves its transcriptional repression function and promotes its export from the nucleus. Mig2 is a functional homolog of Mig1 (Lutfiyya and Johnston 1996), but Mig2 is not regulated by Snf1 and has a localization (Lutfiyya et al. 1998; Fernandez-Cid et al. 2012) and turnover pattern (Lim et al. 2011) that is distinct from Mig1. In high-glucose conditions, Mig1 and Mig2 repress the expression of genes involved in the metabolism of poor carbon sources, in part through Mig1-dependent recruitment of the co-repressor Tup1/Ssn6 (Treitel and Carlson 1995).

In response to nutrient limitation, yeast not only prepare to utilize less preferred carbon sources but also can switch their growth pattern. Depending on cell type and the specific nutritional challenge, yeast can undergo quiescence [G0 arrest (Gray et al. 2004)], sporulation (Neiman 2011), filamentous/invasive/pseudohyphal growth (Cullen and Sprague 2012), and/or biofilm/mat formation (Reynolds and Fink 2001). Many fungal species undergo similar responses. In pathogens like Candida albicans, filamentous growth (Lo et al. 1997) and biofilm/mat formation (Blankenship and Mitchell 2006) are required for virulence. In S. cerevisiae, filamentous growth occurs in response to limiting nitrogen (Gimeno et al. 1992) or glucose (Cullen and Sprague 2000) and is regulated by multiple signaling pathways including TOR (Rohde and Cardenas 2004), Ras-cAMP-PKA (Gimeno et al. 1992), and Snf1 (Cullen and Sprague 2000) through the transcriptional repressors Nrg1 and Nrg2 (Kuchin et al. 2002, 2003), and a mitogen-activated protein kinase (MAPK) pathway commonly referred to as the filamentous growth pathway (Liu et al. 1993; Roberts and Fink 1994). Nutrient limitation stimulates the filamentous growth pathway (Pitoniak et al. 2009), although the plasma-membrane regulators Msb2 and Sho1 (O’Rourke and Herskowitz 1998; Cullen et al. 2000, 2004) are not thought to sense nutrients directly. Rather, the MAPK pathway is sensitized to nutrient levels by regulatory inputs from Ras2-cAMP-PKA (Mosch et al. 1996; Chavel et al. 2010) and regulated processing of Msb2 (Cullen et al. 2004) by starvation-dependent induction of genes that encode its cognate proteases (Vadaie et al. 2008). Whether other nutrient-sensing pathways also regulate the filamentous growth pathway in response to nutrient levels is an open question (Figure 1A, question mark).

Figure 1 

Opy2 regulates the filamentous growth pathway. (A) The MAPK pathway that regulates filamentous growth and biofilm/mat formation. It is not clear how glucose limitation leads to pathway activation. (B) The plate-washing assay. Wild-type, msb2Δ, opy2Δ, ste50Δ, and ste12Δ strains were spotted onto YEPD medium and incubated for 2 days at 30°. The plate was photographed (left panel, YEPD), washed under a stream of water, and photographed again to reveal invaded cells (right panel, Washed). (C) Microscope-based assays and colony morphology assays to evaluate the role of Opy2 in the filamentous growth pathway. (Far left column) The single-cell invasive growth assay. Wild-type, opy2Δ, and ste12Δ cells were spread at low concentrations onto synthetic complete medium lacking glucose (SC-GLU). Plates were incubated for 24 hr and examined by DIC microscopy at ×100 magnification. More than 100 microcolonies were examined, and representative images are shown. Arrows point to axial (nondistal) budding patterns commonly seen in the opy2Δ and ste12Δ mutants. Bar, 5 µm. (Second column from left) Biofilm/mat assay. Strains were grown for 16 hr in YEPD and spotted onto YEPD media with 0.3% agar (YEPD + 0.3% agar). Plates were incubated at 30° for 4 days and photographed. The opy2 and ste12 mats were smaller and less ruffled. Bar, 1 cm. (Third column from left) Strains were spotted onto YEPD + 4% agar atop a nitrocellulose filter and incubated at 30° for 7 days and photographed. Bar, 1 cm. (Far right column) Cells were spotted onto YEP + 0.3% agar for 21 days. Microscopic images (×20 magnification) of mat perimeters are shown. Dense mats containing pseudohyphae at borders can be seen for wild type but not for the opy2Δ and ste12Δ mutant. Bar, 200 µm. (D) Phosphorylation of the MAP kinase Kss1 in wild-type cells and the indicated mutants. Immunoblots are shown to detect phosphorylated Kss1 (P∼Kss1, 42 kDa) and Fus3 (P∼Fus3, 40.7 kDa) with p44/42 antibodies. Blots using antibodies to Pgk1 (44.7 kDa) are shown as a loading control for protein levels. Blots using antibodies to the Kss1 protein (Kss1, 42 kDa) are shown. The asterisk refers to a nonspecific band. (E) β-Galactosidase assays to measure expression of the pFRE-LacZ and pYLR042c-LacZ reporters in wild-type and opy2Δ strains. Enzymatic activity is expressed in Miller units. Two independent experiments were performed, with error bars showing differences between trials. (F) Immunoblots showing the level of the Flo11-HA protein (>220 kDa) in culture supernatant (S) and pellet (P) fractions. The level of Flo11 in supernatants probably reflects the levels seen in pellets.

In this study, we identified a new connection between the major glucose-sensing (AMPK) pathway and the filamentous growth (MAPK) pathway. This connection was uncovered by characterizing the newly identified plasma-membrane regulator Opy2 (Wu et al. 2006; Ekiel et al. 2009; Yang et al. 2009; Yamamoto et al. 2010; Cappell and Dohlman 2011), which regulates the Ste11 branch of the high-osmolarity glycerol response (HOG) MAPK pathway (Hohmann et al. 2007; Saito 2010). Opy2 is also thought to regulate the filamentous growth pathway (Yang et al. 2009; Yamamoto et al. 2010). We confirmed that Opy2 is a major regulator of the filamentous growth pathway, and we identified Mig1 and Mig2 as Opy2-interacting proteins that coregulate the filamentous growth pathway. Mig1 and Mig2, and the AMPK Snf1, were required for induction of the filamentous growth pathway in response to glucose limitation. We also show that Mig1 and Opy2 attenuated the pheromone response pathway, which suggests a role for these proteins in regulating pathway specificity. The regulatory connection between Mig1 and Mig2 and the filamentous growth pathway provides a direct link between glucose sensing (via AMPK) and cell differentiation (via MAPK).

Materials and Methods

Yeast strains and plasmids

Yeast strains are listed in Table 1. Plasmids are listed in Table 2. Yeast and bacterial strains were manipulated by standard methods (Sambrook et al. 1989; Rose et al. 1990). All yeast strains were grown in standard YEP media supplemented with 2% glucose (GLU or D) or 2% galactose (GAL) for 16 hr at 30° unless otherwise indicated. Epitope tagging with hemagglutinin (HA) and c-MYC (MYC) epitopes were performed as described (Schneider et al. 1995). Gene disruptions and GAL1 promoter fusions were made by PCR-based methods (Baudin et al. 1993; Longtine et al. 1998), including cassettes containing antibiotic-resistance markers (Goldstein and McCusker 1999). PCR analysis and phenotype were used to confirm integrations. Epitope tagging and GAL1 promoter fusions were modified to generate N-terminal myristoylation tags by introducing codons for the amino acid sequence MGCTVSTQTI (Gillen et al. 1998; Jansen et al. 2005). The plate-washing assay (Roberts and Fink 1994) and single-cell invasive growth (Cullen and Sprague 2000) assay were performed as described. Biofilm and mat assays were carried out on YEP or YEPD plates containing 0.3 or 4% agar (Reynolds and Fink 2001). Mat perimeters were screened by microscopy as described (Karunanithi et al. 2012). β-Galactosidase assays were performed as described (Cullen et al. 2000). Values reported are the average of at least two independent experiments. Halo assays were performed as described (Jenness et al. 1987).

View this table:
Table 1  Yeast strains used in this study
View this table:
Table 2  Plasmids used in this study

Bacterial strains were grown in Luria–Bertani broth at 37° (Sambrook et al. 1989). Plasmid pBS34 was obtained from the University of Washington Yeast Resource Center (Shaner et al. 2004). Filamentous growth pathway reporter plasmids pYLR042c-lacZ (Roberts et al. 2000) and pFRE-lacZ (Madhani and Fink 1997) have been described. Plasmids expressing a library of PGAL-driven ORFs were obtained from Open Biosystems (Gelperin et al. 2005). pSTE11-4 (Stevenson et al. 1992), pRS316-SHO1-GFP (Marles et al. 2004), pHA-MSB2 (Vadaie et al. 2008), and pMSB2-GFP (Cullen et al. 2004) have been described. Genomic full-length OPY2-GFP was cloned into SalI and EcoRI sites in pRS316 (Sikorski and Hieter 1989) and pRS426 (Christianson et al. 1992) by engineering restriction sites in the primers. The cytoplasmic domain (116–360 amino acids) of Opy2 was cloned into EcoRI and SalI sites in pGBDU-C1 (James et al. 1996) for two-hybrid analysis. Truncations of the cytosolic domain of Opy2 were generated by PCR amplification of codons corresponding to amino acids 246–360 (pGBDU-C1-OPY2Δ1–245), 116–268 (pGBDU-C1-OPY2Δ1–115, 269–360), amd 116–323 (pGBDU-C1-OPY2Δ1–115, 324–360) from the parent plasmid pGBDU-C1-OPY2CT (James et al. 1996). The cytosolic domain of Opy2 was subcloned from two-hybrid constructs into pMAL-C2 vector (New England Biolabs, Ipswich, MA) for expression in Escherichia coli. Full-length MIG1 from pBM3315 (De Vit et al. 1997) was used as a template to PCR-amplify and clone into BamHI and SalI sites in pGAD-C1 (James et al. 1996) and pGEX4T1 (Amersham Biosciences, Piscataway, NJ). Genomic full-length FUS3 was amplified by PCR and cloned into EcoRI and SalI sites in pGBDU-C1. Other two-hybrid constructs are described by A. Pitoniak, C. Chavel, N. Vadaie, J. Smith, S. Karunanithi, D. Camara, and P. Cullen (unpublished results).

Microscopy

Differential interference contrast (DIC) and fluorescence microscopy using GFP and rhodamine filter sets were performed using an Axioplan 2 fluorescent microscope (Zeiss) with PLAN-APOCHROMAT 100X/1.4 (oil) objective (numerical aperture 0.17). Digital images were obtained with the Axiocam MRm camera (Zeiss). Axiovision 4.4 software (Zeiss) was used for image acquisition.

OPY2 plasmid mutagenesis

Plasmid mutagenesis was performed as previously described (Vadaie et al. 2008). Approximately 25 µg of pRS316-OPY2-GFP DNA was mutagenized using hydroxyl-amine mutagenesis (Rose and Fink 1987). Mutation efficiency was ∼12% as assessed by the frequency of opy2 null mutations. Null alleles were selected on synthetic media lacking histidine, using the FUS1-HIS3 reporter integrated into the opy2Δ ste4Δ Σ1278b strain (PC3786). A total of 150 URA+ colonies were assessed, of which 14 were initially identified as HIS. Of the 14, 2 isolates were found to be plasmid-dependent and were confirmed by sequencing at the Roswell Park sequencing facility (Roswell Park Cancer Institute, Buffalo, NY) to contain mutation in the extracellular domain (OPY2C30Y) and in the cytosolic tail (OPY2G207D), respectively.

Two-hybrid analysis

Two-hybrid analysis was used to assess protein interactions (Fields and Song 1989). Two-hybrid analysis using the cytosolic domain of Opy2 against an ordered library of yeast ORFs fused downstream of the Gal4-activation domain constructs (pOAD) (Uetz et al. 2000) was performed. Opy2’s cytosolic tail was cloned downstream of a Gal4-binding domain vector (pGBDU-C1) and transformed into the library by standard transformation protocols. A total of 200,000 transformants were screened. Transformants were selected on synthetic media lacking uracil and leucine to select for the plasmids as well as lacking histidine, in the presence of 7 mM 3-amino-1,2,4-triazole (ATA), which allowed the detection of 40 ATA-resistant colonies. Isolates were selected on 5-fluoroorotic acid (Boeke et al. 1987) containing media to lose the bait plasmid (pGBDUC1-Opy2CT). Twenty-eight isolates activated the GAL1-HIS3 reporter in the absence of pGBDUC1-Opy2CT. Plasmid rescue and retransformation with the bait plasmid identified 11 isolates that were plasmid-dependent. Insert ORFs were sequenced by Roswell Park Sequencing facility (Roswell Park Cancer Institute, Buffalo, NY), which were identified to be Ste50 (5 isolates), Mig1 (5 isolates), and Mtc1 (1 isolate). Mig2 was not identified in the screen, which suggests that some clones might be underrepresented.

Immunoblots and protein analysis

Immunoblots were performed as described (Cullen et al. 2004). Proteins were separated by SDS-PAGE on 12% gels and transferred to nitrocellulose membranes (protran BA85, VWR International Inc., Bridgeport NJ). Membranes were incubated in 25 ml of blocking buffer (5% nonfat dry milk, 10 mM Tris–HCl, pH 8, 150 mM NaCl, and 0.05% Tween 20) for 1 hr at 25° or for 16 hr at 4°. Preblocked membranes were incubated in blocking buffer containing primary antibodies for 1 hr at 25°. Blots were washed three times for 5 min each in TBST (10 mM Tris–HCl, pH 8, 150 mM NaCl and 0.05% Tween 20). Blots were incubated in horseradish peroxidase (HRP)-conjugated secondary antibodies for 1 hr at 25° and washed as above. A WesternBright MCF fluorescent Western blotting kit from Advansta Inc. (Menlo Park, CA) was used to visualize the proteins. Shedding of Flo11 protein was evaluated as described (Karunanithi et al. 2010). Anti-GFP (#11814460001) and anti-HA (12CA5, #11666606001) antibodies were from Roche Applied Science (Indianapolis). Anti-GST antibodies (#600-101-200) were from Rockland Immunochemicals (Gilbertsville, PA). Anti-MBP antiserum (#E8030S) was obtained from New England Biolabs (Ipswich, MA). HRP-conjugated goat anti-mouse secondary antibodies were purchased from Bio-Rad (Hercules, CA).

Phospho-Kss1/Fus3 assay

The assay for monitoring the phosphorylation of Kss1 and Fus3 has been described (Lee and Dohlman 2008; Takahashi and Pryciak 2008). Cultures were grown to logarithmic phase for 6 hr in YEPD or YEPGAL for basal and induced conditions, respectively. Kss1 phosphorylation levels under induced (galactose) conditions are shown unless indicated. An equal number of cells were harvested and cell extracts were prepared by trichloracetic acid precipitation (Cox et al. 1997; Lee and Dohlman 2008). Protein concentrations were normalized using a protein BCA kit (Thermo-Fisher, Waltham, MA). Equal amounts of protein were loaded onto SDS-PAGE gels. Immunoblots were performed using p44/42 MAPK antibodies (#4370s) from Cell Signaling Technology (Danvers, MA) in 5% bovine serum albumin (BSA) at 4° for 16 hr. Blots were washed at 25°. Secondary antibody incubation was performed for 1 hr at 25° with HRP-conjugated goat anti-rabbit antibodies (Jackson Immunoresearch, Westgrove, PA). Incubations with anti-Kss1 (rabbit, Santa Cruz Biotechnology Inc., Santa Cruz, CA; #6775) and anti-Fus3 (goat, Santa Cruz Biotechnology; #6773) were performed in 5% BSA at 4° for 16 hr. Antibody detection for mouse and goat primary antibodies was achieved using HRP-conjugated secondary antibodies raised in goat (Bio-Rad) and donkey (Santa Cruz Biotechnology; #2020), respectively. To detect phosphorylation of Hog1 (Lee and Dohlman 2008), log-phase cultures were induced with 1 M KCl for 5 min. p38 MAPK antibodies (Cell Signaling Technology) were used. Proteins were detected using the WesternBright MCF fluorescent Western blotting kit from Advansta Inc. (Menlo Park, CA).

Protein purification and in vitro pull-down assays

pMAL-C2-Opy2CT and the truncated versions were expressed in E. coli strain [(K12) F-lacZ(am) pho(am) lon tyrT[supC(ts)] trp(am) rpsL(StrR) rpoH(am)165 zhg::Tn10)]. Cells (100 ml) were grown to an OD of 0.6 when isopropyl β-d-1-thiogalactopyranoside (IPTG) was added to a concentration of 0.3 mM and grown for 3 hr. Cell pellets were resuspended in 4 ml lysis buffer (50 mM Tris, pH 7.5, 140 mM NaCl, 1 mM EDTA, 10% glycerol, 0.1% TritonX-100) and frozen at −20°, and 1 mM dithiothreitol (DTT) and 1 mM phenylmethylsulfonylfluoride (PMSF) were added to the cell suspension after thawing and lyzed by sonication. Batch purification was performed using 50 µl of amylose resin (per 1 ml of lysate) (New England Biolabs). Purification was carried out as per manufacturer’s protocol and eluted using lysis buffer containing 10 mM maltose.

In vitro protein binding has been described (Truckses et al. 2006). BL21(DE3) cells containing pGEX4T1 or pGEX4T1-Mig1 were grown to logarithmic phase in 2XYT medium (25 ml) with carbenicillin (100 µg/ml) to select for plasmids. Cells were induced at an OD of 0.6 with 0.3 mM IPTG for 3 hr at 37°. Cell pellets were resuspended in lysis buffer and frozen. After thawing the cell suspension, 1 mM DTT and 1 mM PMSF were added and sonicated. Fifty microliters of glutathione-sepharose beads (Pierce, Rockford, IL) were added to the cell lysates and incubated at room temperature for 30 min. Beads were washed three times, and purified MBP-Opy2CT/truncations or His-Msb2CT (25 µl of 1 µg/µl) was added along with BSA (0.26 mg/ml) to a final volume of 300 µl. Incubation was carried out at 4° for 1 hr followed by three washes with lysis buffer. Bound proteins were extracted in SDS-PAGE buffer by boiling the beads at 100° for 5 min.

Databases

Mig1-binding sites were determined using PatchMatch (http://gfx.cs.princeton.edu/pubs/_2011_PAF/index.php). The Saccharomyces Genome Database was used for gene annotations (http://www.yeastgenome.org/). ImageJ was used for densitometric analysis (http://rsbweb.nih.gov/ij/).

Results

Opy2 regulates the filamentous growth pathway by plasma-membrane recruitment of Ste50

Opy2 functions in the HOG pathway (Wu et al. 2006; Ekiel et al. 2009; Yang et al. 2009; Yamamoto et al. 2010; Cappell and Dohlman 2011) and may also regulate the filamentous growth pathway (Yang et al. 2009; Yamamoto et al. 2010), although this possibility has not been extensively explored (Wu et al. 2006; Chen and Thorner 2007). To determine whether Opy2 regulates the filamentous growth pathway, the OPY2 gene was deleted in strains of the filamentous (∑1278b) background (Liu et al. 1996). The opy2Δ mutant had an invasive growth defect by the plate-washing assay (Roberts and Fink 1994) that was similar to the defect seen in mutants lacking other MAPK pathway regulators (Figure 1B, msb2Δ, ste50Δ, and ste12Δ). The filamentous growth pathway induces the reorganization of cell polarity [e.g., budding pattern, (Gimeno et al. 1992; Roberts and Fink 1994; Chant and Pringle 1995; Cullen and Sprague 2002)] and a delay in the G2 phase of the cell cycle (Ahn et al. 1999; Madhani et al. 1999). The single-cell invasive growth assay (Cullen and Sprague 2000) showed that opy2Δ cells were defective for filamentous growth and distal-unipolar budding (Figure 1C, SC-GLU, arrows) and were rounder than wild-type cells (Figure 1C, SC-GLU). Phosphorylation of the MAP kinase that regulates the filamentous growth pathway, Kss1, by the MAP kinase kinase Ste7 is required for pathway activation (Courchesne et al. 1989; Ma et al. 1995; Cook et al. 1997; Madhani et al. 1997) and provides a readout of pathway activity. The opy2 mutant had reduced levels of phosphorylated Kss1 (Figure 1D). Activated Kss1 induces transcriptional activators and inactivates transcriptional repressors (Liu et al. 1993; Roberts and Fink 1994; Cook et al. 1997; Bardwell et al. 1998), which results in the expression of genes that regulate filamentous growth. Transcriptional reporters of the filamentous growth pathway, FRE-LacZ and YLR042c-LacZ (Roberts et al. 2000), were dependent on Opy2 for expression (Figure 1E). Like other pathway regulators, Opy2 was required for mat formation on low-agar (Figure 1C, YEPD 0.3% agar) and high-agar media (Figure 1C, YEPD 4% agar; Supporting Information, Figure S1). Opy2 was required for cell differentiation into pseudohyphae at mat perimeters, which contain densely aggregated pseudohyphal cells (Figure 1C, YEP 0.3% agar) (Karunanithi et al. 2012). Flo11 is the major adhesion molecule that regulates biofilm/mat formation (Reynolds and Fink 2001; Karunanithi et al. 2010) and filamentous/invasive growth (Lambrechts et al. 1996; Lo and Dranginis 1998; Guo et al. 2000). FLO11 expression is regulated by the filamentous growth pathway (Lo and Dranginis 1998) and other pathways (Rupp et al. 1999; Barrales et al. 2008; Vinod et al. 2008) and provides another measure of pathway activity. The opy2 mutant, like other pathway mutants (e.g., ste12Δ), showed reduced Flo11 levels (Figure 1F, “P”). Shedding of Flo11 from cells, which regulates mat expansion (Karunanithi et al. 2010), was reduced in the opy2Δ mutant (Figure 1F, “S”). Together, these results demonstrate that Opy2 regulates the filamentous growth pathway.

Opy2 is a single-pass transmembrane protein with an N-terminal extracellular domain and a cytosolic C-terminal domain. Opy2-GFP localized to the cell periphery and vacuoles (Wu et al. 2006; data not shown). Colocalization of Opy2-mCherry with Msb2-GFP and Sho1-GFP, two plasma membrane regulators of the filamentous growth pathway (O’Rourke and Herskowitz 1998; Cullen et al. 2000, 2004), showed that these proteins all localized to the plasma membrane (data not shown). To determine how Opy2 regulates the filamentous growth pathway in the context of these plasma-membrane regulators, genetic suppression analysis was performed. Hyperactive versions of Msb2*, which lacks the inhibitory N-terminal mucin-homology domain (Cullen et al. 2004), and Sho1P120L (Tatebayashi et al. 2007; Vadaie et al. 2008), did not bypass the signaling defect of the opy2Δ mutant (Figure 2A). Thus, Msb2 and Sho1 may function through Opy2 in the pathway. Msb2 and Sho1 connect to the polarity regulatory GTPase Cdc42 (Peter et al. 1996; Leberer et al. 1997; Park and Bi 2007). Deletion of RGA1, which encodes the GTPase-activating protein for Cdc42 in the filamentous growth pathway (Stevenson et al. 1995; Smith et al. 2002), failed to restore signaling to the opy2Δ mutant (Figure 2A, rga1), which indicates that Opy2 functions below Cdc42 in the pathway. A hyperactive allele of the gene encoding the MAP kinase kinase kinase Ste11, STE11-4 (Stevenson et al. 1992), restored signaling in the opy2Δ mutant (Figure 2A, STE11-4), indicating that Opy2 functions above Ste11 in the pathway.

Figure 2 

Opy2 is required for plasma-membrane recruitment of Ste50. (A) Genetic suppression analysis showing phosphorylated Kss1 in an opy2Δ mutant with a control plasmid (CTL) or plasmids containing hyperactive alleles of MSB2, SHO1, and STE11. For rga1Δ, an rga1Δ opy2Δ strain was examined. Blots were performed as in Figure 1D. (B) Levels of phosphorylated Kss1 in wild-type and opy2Δ mutant strains that expressed Myr-Ste50 under the control of its own promoter or the GAL1 promoter (PGAL-Myr-Ste50). (C) Strains in B were spotted onto YEPGAL media for 48 hr and examined by the plate-washing assay to measure agar invasion. (D) Phosphorylation of Kss1 in cells overexpressing a wild-type (GAL-OPY2) or membrane-anchored version of the cytosolic domain of Opy2 (PGAL-Myr-Opy2CT). (E) Phosphorylation of Kss1 in an opy2Δ mutant containing a control plasmid (Ctl) and wild-type (WT) Opy2, Opy2C30Y, or Opy2G207D. (F) Two-hybrid analysis of the interaction between pGADC1-STE50 (Ste50), pGBDUC1-OPY2 (Opy2), or pGBDUC1-OPY2G207D (G207D). Interactions are represented by growth on synthetic medium lacking adenine (SD-Ura-Leu-Ade, bottom panel). Growth of strains on control media is also shown (SD-Ura-Leu, top panel).

Several proteins function between Cdc42 and Ste11 in the filamentous growth pathway (including Ste20 and Ste50; Figure 1A). In the HOG pathway, Opy2 functions to recruit Ste50 to the plasma membrane (Wu et al. 2006; Yamamoto et al. 2010). Ste50 is a general adaptor that associates with Ste11 (Posas et al. 1998; Ramezani Rad et al. 1998). Like Opy2, Ste50 functions in the filamentous growth pathway (Figure 1, B and D, and Figure S1) (Ramezani Rad et al. 1998; Ramezani-Rad 2003; Truckses et al. 2006). To determine whether Opy2 regulates the filamentous growth pathway by plasma-membrane recruitment of Ste50, the Ste50 protein was tethered to the plasma membrane by an N-terminal myristoylation signal (Myr-Ste50) (Gillen et al. 1998; Jansen et al. 2005). Myr-Ste50 partially restored MAPK signaling to an opy2Δ mutant (Figure 2B, Myr-Ste50 opy2Δ). Overproduction of Myr-Ste50 caused a more effective bypass than Myr-Ste50 expressed under its own promoter (Figure 2B, PGAL-Myr-Ste50 opy2Δ). Myr-Ste50 and PGAL-Myr-Ste50 also bypassed the invasive growth defect of the opy2Δ mutant (Figure 2C and Figure S2, A and B).

Overexpression of a membrane-anchored version of the cytosolic domain of Opy2 was sufficient to activate the filamentous growth pathway (Figure 2D, PGAL-Myr-Opy2CT), which demonstrates that the cytosolic domain of Opy2 plays the major role in pathway regulation. Chemical mutagenesis was used to identify loss-of-function mutations that reduce Opy2 function in the filamentous growth pathway. Two alleles were identified: OPY2G207D was partially defective for pathway activity (Figure 2E, P∼Kss1). OPY2G207D contains a mutation in a Ste50-binding site (Figure 3A) (Ekiel et al. 2009; Yamamoto et al. 2010) and was partially defective for binding to Ste50 by two-hybrid analysis (Figure 2F).

Figure 3 

Opy2 interacts with the transcriptional repressor Mig1. (A) The Opy2 ORF and deletions used in the study. Yellow vertical lines, conserved cysteines; red bars, Ste50-binding sites; green bars, Mig1-binding sites. (B) Opy2 associates with Ste50, Mtc1, Mig1, and Mig2 by two-hybrid analysis. (Left) Growth of strains on control media (SD-URA-LEU). (Right) Growth of strains represents protein interactions (SD-URA-LEU-HIS+3-ATA). (C) Two-hybrid deletion mapping. Two-hybrid analysis of versions of Opy2 lacking specified deletions in its cytoplasmic tail with Mig1. (D) Pull-down assay. GST (∼26 kDa) or GST-Mig1 (∼81.5 kDa) were purified (Input, anti-GST) and bound to glutathione-sepharose beads (Pull Down, anti-GST). Beads were incubated with purified MBP-Opy2CT (Input, anti-MBP, ∼69.4 kDa). Coprecipitated MBP-Opy2CT is shown (Pull Down, anti-MBP).

A second allele, OPY2C30Y, also showed reduced filamentous growth pathway activity (Figure 2E), which contained a substitution of a conserved cysteine residue in the extracellular domain (Figure 3A). Opy2G207D-GFP and Opy2C30Y-GFP were expressed at wild-type levels (Figure 2E) and were localized in a similar manner as wild-type Opy2-GFP (data not shown). A lower band was also seen in Opy2C30Y-GFP (Figure 2E, asterisk). Thus, the extracellular domain of Opy2 may participate in regulating the filamentous growth pathway. The Opy2G207D and Opy2C30Y proteins were also defective for HOG pathway signaling, based on the salt-dependent phosphorylation of the MAPK that functions in the HOG pathway, Hog1 (Figure S2C). Together, our results support and extend the conclusion put forward in Yang et al. (2009) that Opy2 is a general adaptor for the HOG and filamentous growth pathways that functions by plasma-membrane recruitment of Ste50.

Mig1 and Mig2 proteins interact with Opy2

Three observations suggest that Opy2 has a function in regulating the filamentous growth pathway separate from Ste50: (1) the opy2Δ mutant had a more severe invasive growth defect than the ste50Δ mutant (Figure 1B); (2) Myr-Ste50, even when overexpressed, did not fully bypass the signaling defect of the opy2Δ mutant (Figure 2B); and (3) the opy2Δ ste50Δ double mutant had a more severe invasive growth defect than the ste50Δ single mutant (Figure S2B). An additional role for Opy2 in regulating the filamentous growth pathway might result from Opy2 regulating a component of the pathway in addition to Ste50. The cytosolic domain of Opy2 (Opy2CT) was tested for interactions with pathway regulators. Opy2CT did not associate with the cytosolic domain of Msb2 (Msb2CT), Cdc24, Cdc42, Ste11, Ste7, or Kss1 by two-hybrid analysis (data not shown). Cdc24, Cdc42, and Ste11 (A. Pitoniak, C. Chavel, N. Vadaie, J. Smith, S. Karunanithi, D. Camara, and P. J. Cullen, unpulished results) and Msb2CT, Ste7, and Kss1 (see below) mediated two-hybrid interactions with control proteins. Opy2 did not interact with Ras2, which also regulates the filamentous growth pathway (Mosch et al. 1999; Chavel et al. 2010) or with Bem4, a newly identified regulator of the filamentous growth pathway (A. Pitoniak, C. Chavel, N. Vadaie, J. Smith, S. Karunanithi, D. Camara, and P. J. Cullen, unpublished results).

Opy2 might regulate the pathway by associating with a pathway regulator that has yet to be identified. A two-hybrid screen with the cytosolic domain of Opy2 using a pooled ORF library (Uetz et al. 2000) identified Ste50, which was expected, Mig1, and Mtc1 (Figure 3B). Specifically, five isolates of Ste50, five isolates of Mig1, and one isolate of Mtc1 were identified. Mig2 is a homolog of Mig1 (Lutfiyya and Johnston 1996; Lutfiyya et al. 1998; Westholm et al. 2008). Mig2 was not identified in the screen but associated with the cytosolic domain of Opy2 by two-hybrid analysis (Figure 3B). These interactions were not identified in comprehensive screens (Uetz et al. 2000; Ho et al. 2002), possibly because full-length versions of Opy2 were used, whereas we used only the cytosolic domain of Opy2. Deletion of MTC1, which encodes a protein of unknown function (Fleischer et al. 2006), did not affect the filamentous growth pathway and was not explored further.

Mig1 is a transcriptional repressor of genes required for growth on nonfermentable carbon sources (Nehlin and Ronne 1990; Carlson 1999). Two-hybrid deletion mapping showed that Mig1 associated weakly with the cytosolic domain of Opy2 that lacked amino acids 116–245 (Opy2CTΔ116–245), 269–360 (Opy2 CTΔ1–115,269–360), and 324–360 (Opy2CTΔ1–115,324–360) (Figure 3C). The sites that showed interaction with Mig1 also contained the Ste50-binding sites (Figure 3A) (Yamamoto et al. 2010). Opy2G207D, which was partially defective for binding to Ste50 (Figure 2F), was also partially defective for binding to Mig1 (Figure 3C; see Figure 3A for deletions). To verify the interaction between Mig1 and Opy2, epitope fusions of Mig1 (GST-Mig1) and the cytosolic domain of Opy2 (MBP-Opy2CT) were overexpressed in E. coli and purified by affinity chromatography. Beads coated with GST-Mig1 pulled down MBP-Opy2CT in vitro (Figure 3D). Thus, Mig1 interacts directly with the cytosolic domain of Opy2 in vivo and in vitro.

Mig1 and Mig2 regulate the filamentous growth pathway in response to glucose depletion

The identification of Mig1 and Mig2 as interacting partners with Opy2 suggested a connection between the AMPK pathway and the filamentous growth pathway. To test whether Mig1 and/or Mig2 regulate the filamentous growth pathway, the MIG1 and MIG2 genes were disrupted in strains of the filamentous background. The mig1Δ and mig2Δ single mutants were not defective for pathway activity, but the mig1Δ mig2Δ double mutant was defective (Figure 4A). Specifically, the mig1Δ mig2Δ double mutant failed to show induction of the pathway in response to glucose limitation (Figure 4A, P∼Kss1; compare GLU to GAL). The mig1Δ mig2Δ double mutant was also defective for invasive growth (Figure 4B). Mig3 is a poorly defined transcriptional repressor that is thought to function separately from Mig1 and Mig2 (Lutfiyya et al. 1998). Disruption of the MIG3 gene in wild-type cells or cells lacking Mig1 and/or Mig2 did not cause a further reduction in pathway activity (Figure S5), which indicates that Mig1 and Mig2 primarily coregulate the filamentous growth pathway.

Figure 4 

The Mig1 and Mig2 proteins regulate starvation-dependent induction of the filamentous growth pathway. (A) Phospho-Kss1 blots of wild type, opy2Δ, mig1Δ, mig2Δ, and mig1Δ mig2Δ strains grown to mid-log phase in YEPD or YEPGAL media. (B) Strains in A were grown on YEPD media for 48 hr and assessed by the plate-washing assay. (C) Two-hybrid analysis of Mig1 and Mig2 with Msb2CT.

To define how Mig1 and Mig2 regulate the filamentous growth pathway, genetic suppression analysis was performed. Unlike opy2Δ (Figure 2A), the signaling defect of the mig1Δ mig2Δ double mutant was suppressed by an activated version of Sho1 (data not shown), which indicates that Mig1 and Mig2 regulate the filamentous growth pathway at/above the level of Sho1 in the pathway (Figure 1A). To determine how Mig1 and Mig2 regulate the pathway at this level, two-hybrid analysis was performed against known regulators of the pathway. In addition to the cytosolic domain of Opy2, Mig1 and Mig2 also associated with the cytosolic domain of Msb2 (Figure 4C). The cytosolic domain of Msb2 has been shown to associate with only a single other protein, Cdc42, by two-hybrid analysis (Cullen et al. 2004), which suggests that the interaction with Mig1 and Mig2 is highly specific. The fact that Mig1 and Mig2 associate with the cytosolic domain of Msb2 might explain how it regulates the pathway at/above the level of Sho1 in the pathway. The Mig proteins may also modulate the Opy2/Ste50 interaction, given that Opy2 associates with Mig1 and Ste50.

Mig1 shuttles between the cytoplasm and the nucleus in a glucose-dependent manner (De Vit et al. 1997; DeVit and Johnston 1999). In response to glucose limitation, Mig1 is exported from the nucleus, which attenuates its transcriptional repression function (De Vit et al. 1997). The regulation of Mig2 function is less clear (Lutfiyya and Johnston 1996). Mig1 may regulate the filamentous growth pathway in the cytosol, under conditions in which it is exported from the nucleus. In cells undergoing filamentous growth, a functional N-terminally tagged GFP-Mig1 protein was predominately in the cytosol (data not shown), which supports the possibility that Mig1 associates with and may regulate Msb2 and Opy2 at the plasma membrane.

Snf1 regulates starvation-dependent induction of the filamentous growth pathway

Nucleo-cytoplasmic shuttling of Mig1 is regulated by Snf1 (De Vit et al. 1997; DeVit and Johnston 1999), the yeast AMPK (Mantovani and Roy 2011). Snf1 is required for filamentous growth (Cullen and Sprague 2000) by regulating the activity of another set of transcriptional repressors, Nrg1 and Nrg2 (Kuchin et al. 2002). Whether Snf1 also regulates the filamentous growth pathway is not clear. Like the mig1Δ mig2Δ double mutant, the snf1Δ mutant was defective for induction of the filamentous growth pathway in response to glucose depletion (Figure 5A). One possibility is that, in the snf1 mutant, Mig1 cannot be phosphorylated and exit the nucleus to regulate Msb2 and Opy2 at the cytosol–plasma membrane interface.

Figure 5 

Snf1 regulates the filamentous growth pathway independently of the HOG pathway. (A) Phospho-Kss1 levels for wild type, opy2Δ, and snf1Δ strains grown in YEPD (GLU) and YEPGAL (GAL) media. (B) Wild-type cells were grown in YEPD (GLU) or YEPGAL (GAL) for 6 hr. After 6 hr in galactose, cells harvested by centrifugation and transferred to YEPD media (GLU). Equal numbers of cells were withdrawn at 5, 15, and 30 min, and cell extracts were prepared and probed by immunoblot analysis. (C) Hog1 phosphorylation analysis. Phospho-blot of the MAP kinase Hog1 in response to 0.5 M KCl for 5 min was performed in the indicated strains in YPD media.

In response to glucose, Mig1 translocates into the nucleus, where it represses the expression of genes involved in gluconeogenesis and poor carbon source utilization (De Vit et al. 1997; data not shown). If Mig1 regulates the filamentous growth pathway in the cytosol, then its nuclear entry might result in attenuation of the filamentous growth pathway. The addition of glucose to filamentous cells causes a rapid switch in budding pattern (to the vegetative or axial pattern) and a round-cell morphology within one cell division cycle (Cullen and Sprague 2000). The effect of glucose addition on the activity of the MAPK pathway has not been tested. Within 15 min of glucose addition, the activity of the filamentous growth pathway dropped to basal levels (Figure 5B). The attenuation of the filamentous growth pathway in response to glucose was slower than the kinetics of the nuclear entry of Mig1, which occurs in 1–2 min (De Vit et al. 1997; DeVit and Johnston 1999). This difference might be explained by the fact that Mig2 is regulated differently from Mig1 in terms of its localization and turnover in the cytosol (Lutfiyya et al. 1998; Fernandez-Cid et al. 2012). We also cannot exclude the possibility that other proteins might also regulate the pathway under this condition.

The above results support the idea that Mig1 functions in the cytosol to regulate the filamentous growth pathway, acting as a cytosolic indicator of low glucose levels. In support of this possibility, overexpression of MIG1 (PGAL-MIG1) under conditions that favor its cytosolic localization (2% galactose) caused hyper-invasive growth (Figure S6A), in line with a previous report (Jin et al. 2008). Co-overexpression of MIG1 and MIG2 caused an additive hyper-invasive growth phenotype (Figure S6A). Although the cytosolic domains of Msb2 and/or Opy2 might enter the nucleus to modulate the transcriptional repression function of Mig1 and Mig2, we failed to detect the cytosolic domains of these proteins in the nucleus under any condition tested (data not shown).

The AMPK pathway also regulates the response to environmental stresses such as salt, elevated pH, and hydrogen peroxide (Hong and Carlson 2007). Snf1 is phosphorylated in response to these stresses and is required for growth under these conditions (Hong and Carlson 2007). The mig1Δ mig2Δ double mutant shows elevated expression of some HOG pathway targets, including the gene encoding the Na+ transporter ENA1 (Proft and Serrano 1999). Thus, we considered whether the HOG pathway might be hyperactive in the mig1Δ mig2Δ double mutant. Given that the HOG pathway can inhibit the filamentous growth pathway (Saito 2010), a hyperactive HOG pathway might attenuate the filamentous growth pathway. To test this possibility, the activity of the HOG pathway was examined in wild-type cells and in the mig1Δ mig2Δ double mutant. Under normal growth conditions (no salt), neither wild-type cells nor the mig1Δ mig2Δ mutant showed HOG pathway activity, based on the levels of phosphorylated Hog1 (Figure 5C, left panels). In response to salt (0.5 M KCl), wild-type cells and the mig1Δ mig2Δ mutant showed similar HOG pathway activity (Figure 5C, right panels). In the pbs2Δ mutant, which lacks the MAP kinase kinase for the HOG pathway, no phosphorylated Hog1 was detected (Figure 5C). Therefore, the HOG pathway is not hyperactive in the mig1Δ mig2Δ mutant.

Mig1 has been implicated in regulating cell-wall integrity (Krause et al. 2008), which is controlled by the protein kinase C (PKC) pathway (Levin 2005). In some settings, the PKC pathway can stimulate the filamentous growth pathway (Yashar et al. 1995; Birkaya et al. 2009). The activity of the PKC pathway, as measured by phosphorylation of the MAPK Slt2, Slt2∼P, did not correlate with filamentous growth pathway activity in the snf1Δ or mig1Δ mig2Δ double mutants (data not shown). Thus, Mig1 and Mig2 do not regulate the filamentous growth pathway through the PKC pathway. We also tested whether Mig1 and Mig2 were required to stabilize cell-surface components of the filamentous growth pathway. Loss of Mig1 and Mig2 did not affect the levels of Msb2, Sho1, or Opy2 in the cell (Figure S3). Deletion or overexpression of MIG genes might alter cellular growth rate, which might lead to altered MAPK activity. No differences were observed in growth of the mig1Δ mig2Δ mutant or the PGAL-MIG1 strain compared to wild type (Figure S6B). To summarize, Snf1 and the Mig1 and Mig2 proteins are required for full activation of the filamentous growth pathway, where they sensitize the pathway to cellular nutrient levels.

Mig1 and Opy2 contribute to a pathway-specific response

The filamentous growth pathway shares components with the pheromone response (Chen and Thorner 2007) and HOG (Saito 2010) pathways. Mig1 and Mig2 might also regulate these pathways. The mig1Δ mig2Δ double mutant responded to pheromone (Figure S4A). The mig1Δ mig2Δ ssk1Δ triple mutant, which lacks the Sln1 branch of the HOG pathway (Ota and Varshavsky 1993; Chang and Meyerowitz 1994), was not sensitive to osmotic stress (Figure S4B). Thus, Mig1 and Mig2 specifically regulate the filamentous growth pathway and are not required to activate the pheromone response or HOG pathways.

Opy2 (overproduction induced pheromone-resistant yeast, or Opy) was previously identified in a screen for genes that, when overexpressed, overcome pheromone-induced growth arrest (Edwards et al. 1997). We also saw that overexpression of OPY2 reduced sensitivity to mating pheromone (Figure 6A). Likewise, overexpression of OPY2 dampened the activity of a mating pathway growth reporter (Figure 6B, FUS1-HIS3) while causing hyper-invasive growth (Figure 6C). Overexpression of MIG1 similarly dampened the sensitivity to mating pheromone (Figure 6A) and caused a reduction in FUS1-HIS3 reporter expression (Figure 6B), while stimulating filamentous growth (Figure 6C). The hyper-invasive growth induced by overexpression of MIG1 was partially independent of the MAPK pathway (data not shown), suggesting that Mig1 may have additional roles in regulating filamentous growth in addition to regulating the MAPK pathway. Inhibition of the mating pathway by overexpression of OPY2 and MIG1 was not seen in cells overexpressing other pathway regulators, including Msb2 and Sho1 (Cullen et al. 2004). Thus, Opy2 and Mig1 have a specific function in regulating signal discrimination between the mating and filamentous growth pathways.

Figure 6 

Different roles for Opy2 and Mig1/Mig2 in regulating the mating and filamentous growth pathways. (A) Halo assay. Wild type, PGAL-OPY2, and PGAL-MIG1 were spread on YEPGAL media, onto which was spotted 2, 10, and 20 µl of 1 mg/ml α-factor. Plates were incubated for 48 hr and photographed. Bar, 1 cm. Inset portraying the loss in pheromone sensitivity is also shown. (B) Overexpression of OPY2, MIG1, and STE50 suppresses expression of the FUS1-HIS3 reporter. Strains were spotted onto synthetic medium with galactose (SC-GAL) or SC-GAL lacking histidine (SC-GAL-HIS), grown for 24 hr, and photographed. In the bottom panels, Myr-Ste50 was expressed in wild-type cells or in an opy2Δ mutant, as indicated. (C) Overexpression of OPY2 and MIG1 induces hyper-filamentous growth. Cells were grown in YEPGAL and examined by DIC microscopy at ×100; representative images are shown. Bar, 5 µm. (D) Mig1 and Mig2 associate with Ste7 and Kss1 but not with Fus3 by two-hybrid analysis.

Different MAP kinases regulate the filamentous growth (Kss1) and mating (Fus3) pathways. We found that Mig1 and Mig2 associated with the MAP kinase that regulates the filamentous growth pathway (Figure 6D, Kss1) but not with the MAP kinase that regulates the mating pathway (Figure 6D, Fus3). Mig1 and Mig2 also associated with the filamentation/mating MAP kinase kinase, Ste7, but not with Cdc24, Cdc42, or Ste11 (Figure 6D, shown for Ste7). Thus, Mig1 and Mig2 may specifically regulate the filamentation MAPK. The inhibitory effect of Opy2/Mig1 might occur through Ste50, which is a target of Kss1 (Jansen et al. 2001; Ramezani-Rad 2003; Ekiel et al. 2009; Yamamoto et al. 2010). Indeed, overexpression of Ste50 also dampened FUS1-HIS3 expression (Figure 6B). When anchored to the plasma membrane (PGAL-Myr-Ste50), reporter expression was restored to wild-type levels (Figure 6B) in an Opy2-independent manner (Figure 6B). Thus, Opy2 and Mig1 may differentially activate the filamentous growth pathway at the levels of Ste50 and/or Kss1.

MAPK pathway does not regulate the transcriptional repression function of Mig1 and Mig2

The filamentous growth pathway might regulate the glucose repression function of Mig1 and Mig2. In medium containing glucose and galactose, wild-type cells do not express genes under the control of the GAL promoter (Figure 7, wild-type GAL + GLU) (Nehlin et al. 1991; Lutfiyya et al. 1998). In cells lacking Mig1 and Mig2, GAL-driven genes are expressed (Figure 7, GAL + GLU, mig1Δ mig2Δ). Deletion of pathway regulators Opy2 or Ste12 did not alter gene expression under this condition (Figure 7, GAL + GLU, opy2Δ and ste12Δ). Thus, the filamentous growth pathway does not appear to regulate glucose repression mediated by Mig1 and Mig2. In support of this possibility, whole-genome DNA microarray analysis of genes induced by MSB2 overexpression or hyperactivation did not show induction or repression of Mig targets (Chavel et al. 2010), and the expression of GAL-dependent genes was not affected by the MAPK pathway under inducing conditions (Figure 7, GAL). Furthermore, cells lacking or expressing hyperactive MAPK regulators did not influence the ability of cells to grow on medium in which galactose was the preferred carbon source.

Figure 7 

The filamentous growth pathway does not regulate the glucose repression function of the Mig proteins. Immunoblots with antibodies to HA-tagged Act1, expressed from a GAL promoter (PGAL), PGAL-ACT1 was obtained from Gelperin et al. (2005). Wild-type, mig1Δ mig2Δ, opy2Δ, and ste12Δ strains were grown in YEP media supplemented with glucose and galactose (GLU + GAL) or galactose (GAL). Pgk1, loading control.

Discussion

Opy2 regulates the filamentous growth pathway

The identification of proteins that regulate signal transduction pathways is an important step in understanding what pathways sense and how they function. Opy2 had previously been implicated in regulating the filamentous growth and HOG pathways (Yang et al. 2009; Yamamoto et al. 2010). Here, we confirm and extend these findings by showing that Opy2 is a major regulator of the filamentous growth pathway. Opy2 is required for invasive growth, filament formation, and biofilm/mat expansion. Previously, our lab and other labs have shown that the mucin-like glycoprotein Msb2 (Cullen et al. 2004) and tetraspan protein Sho1 (O’Rourke and Herskowitz 1998; Cullen et al. 2000, 2004) regulate the filamentous growth pathway. Thus, three transmembrane proteins regulate the filamentous growth pathway. Our findings show that Msb2 and Sho1 function through Opy2. Our data also corroborate the idea that Opy2 is a cortical tether for the adaptor protein Ste50 (Wu et al. 2006; Yamamoto et al. 2010). In principle, cortical recruitment of Ste50 might promote pathway activation through the interacting partners Cdc42 (Truckses et al. 2006) or Ste11 (Wu et al. 2006). Genetic suppression analysis indicates that Opy2 regulates the filamentous growth pathway at the level of Ste11. Although the cytosolic domain of Opy2 plays the major role in Ste50 recruitment and pathway activation, the extracellular domain of Opy2, which contains eight cysteine residues that may form a cysteine knot (Craik et al. 2001; Cemazar et al. 2008), may also play a minor role in filamentous growth pathway regulation.

Mig1 and Mig2 proteins connect glucose signaling to MAPK regulation

We also show that the transcriptional repressors Mig1 and Mig2 regulate the filamentous growth pathway. Hints that Mig1 and Mig2 regulate the pathway have come from genome-wide screens conducted in our lab and other labs. The mig1 mutant was identified among an ordered collection of nonessential deletion mutants (Winzeler et al. 1999) as defective for Msb2-HA secretion (Chavel et al. 2010), an indicator of MAPK activity (Vadaie et al. 2008). MIG1 and MIG2 were identified in a whole-genome screen for genes that, when overexpressed, induce Msb2-HA secretion (Chavel et al. 2010). Overexpression of MIG1 (and OPY2) came out of a genomic screen for genes that, when overexpressed, dampen the activity of the mating pathway FUS1-HIS3 reporter (H. S. Dionne and P. J. Cullen, unpublished data). MIG1 was also identified in a large-scale survey for genes that, when overexpressed, induce filamentous growth (Jin et al. 2008).

The Mig1 and Mig2 proteins function in the nucleus as transcriptional repressors that bind to specific DNA elements (Nehlin and Ronne 1990; Lutfiyya and Johnston 1996; Ozcan and Johnston 1996). Our findings support the idea that Mig1 and Mig2 regulate the filamentous growth pathway in the cytosol. This possibility is supported by the following data: (1) Mig1 is found mainly in the cytosol under conditions when it regulates the filamentous growth pathway; (2) Mig1 and Mig2 associate with the cytosolic domains of transmembrane proteins Msb2 and Opy2, which are not found in the nucleus; and (3) Mig1 and Mig2 associate with other regulators of the filamentation pathway such as Ste7 and Kss1, which are primarily cytosolic. Given that Mig1 and Mig2 associate with the cytosolic domains of both Msb2 and Opy2, they may play a role in clustering the cell-surface proteins into a complex at the plasma membrane. Mig1 may also modulate the Opy2Ste50 interaction because Mig1 and Ste50 bind to overlapping sites in the cytosolic domain of Opy2. Finally, Mig1 and Mig2 associate with the MAP kinase that regulates the filamentous growth pathway, Kss1, but not with the mating pathway MAP kinase Fus3. Thus, Opy2 and Mig1 may promote pathway discrimination at the level of Kss1.

A growing list of transcription factors moonlight in other cellular compartments to regulate cellular functions in diverse ways. The major tumor suppressor and transcription factor p53 regulates apoptosis at the mitochondria (Murray-Zmijewski et al. 2008). β-Catenin is both a transcription factor and a structural component of adherens junctions (Logan and Nusse 2004). Transcription factors TFII-I (Park and Dolmetsch 2006) and the tumor suppressor PML (Giorgi et al. 2010) regulate calcium signaling outside the nucleus. Indeed, Mig2 may regulate cellular respiration in the mitochondria under glucose-limiting conditions (Fernandez-Cid et al. 2012). Likewise, noncanonical transcription factors can function as coincidence detectors in the nucleus: the glycolytic enzyme hexokinase 2 (Hxk2) coregulates Mig1-dependent gene repression in the nucleus (Moreno et al. 2005; Pelaez et al. 2012). Together, these findings reveal a high diversity in the molecular connections that integrate nutrition, gene expression, and cellular signaling.

Mig1 and Mig2 may alternatively regulate the filamentous growth pathway through transcriptional regulation. One possibility is that the proteins function as activators of filamentation-specific genes. Mig1 can activate gene expression in some contexts, particularly in cells lacking the chromatin-remodeling proteins Tup1/Ssn6 (Treitel and Carlson 1995). Recently, Tup1 has been shown to exit the nucleus and associate with internal membranes, resulting in modification of its gene expression function (Han and Emr 2011). It is possible that Mig1 and Mig2 are modified in some way by associating with Msb2, Opy2, Ste7, or Kss1, which results in alteration of their normal function. Mig1-binding sites (AATGCGGGGK or SYGGGG) are not found upstream of the filamentation target genes MSB2, KSS1, PGU1, SVS1, YLRO42C, STE12, and TEC1, which might suggest that simply switching these proteins from repressors to activators cannot explain the change in the function of these proteins. DNA microarray data on genes upregulated by Mig1 and Mig2 does not include filamentous growth pathway regulators (Westholm et al. 2008). Nevertheless, Mig1 and Mig2 may modulate nutritional scavenging by repression of GAL/SUC and other genes in the nucleus. Mig1 and Mig2 may also repress a subset of filamentation genes in high glucose, which is relieved under glucose-limiting conditions. At least one gene, YLR042c (Roberts et al. 2000), falls into this category (Lutfiyya et al. 1998).

Cellular glucose control involves the coordination of multiple proteins and pathways. A role for Mig1 and Mig2 in regulating the filamentous growth pathway provides the first regulatory connection between the AMPK pathway and a MAPK pathway in yeast. Previously, the AMPK pathway and MAPK pathway have been shown to coordinate filamentation responses by converging at the Flo11 promoter, at the level of the Nrg1 and Nrg2 (Kuchin et al. 2002, 2003). The regulatory connection that we describe brings to light an additional layer of signal integration, which may benefit the overall precision and timing of the differentiation response. The critical hole that this discovery fills is to help explain how the filamentous growth pathway becomes sensitized to nutrient levels. We show that the activity of the MAPK pathway is induced by glucose starvation and reduced by glucose addition through the action of Mig1, Mig2, and Snf1. These proteins might also sensitize MAPK activity to other nutrient levels. In particular, filamentous growth can be regulated by limiting nitrogen (Gimeno et al. 1992), which is sensed by the Snf1-activating kinase Sak1 (Orlova et al. 2010). A link between the AMPK and MAPK pathways has been actively sought in higher eukaryotes, and several connections have been reported. In skeletal muscle, contraction-induced AMPK activation leads to MAPK-dependent transcriptional induction of glucose transporters (Tremblay et al. 2003). AMPK can also induce MAPK-dependent induction of diverse targets, including MUC5A (Bae et al. 2011).

Acknowledgments

We thank Peter Pryciak (University of Massachussetts, Worcester, MA) for help with phospho-MAPK blots; Anuj Kumar (University of Michigan, Ann Arbor, MI) for suggestions and reagents; Mark Johnston (University of Colorado, Denver) for reagents and for providing comments on the manuscript; Stanley Fields (University of Washington, Seattle, WA) for two-hybrid plasmids; and Michael Yu (SUNY-Buffalo) for reagents. We thank L. Lutfiyya and M. Devit for reading the manuscript and for their helpful comments. Unnati Dev and Colin Chavel made several constructs and provided assistance. Preliminary data from screens were shared by Heather Dionne.

Footnotes

  • Communicating editor: J. Heitman

  • Received June 7, 2012.
  • Accepted August 10, 2012.

Literature Cited

View Abstract