In Saccharomyces cerevisiae, adhesive growth on solid surfaces is mediated by the flocculin Flo11 to confer biofilm and filament formation. Expression of FLO11 is governed by a complex regulatory network that includes, e.g., the protein kinase A (PKA) signaling pathway. In addition, numerous regulatory genes, which have not been integrated into regulatory networks, affect adhesive growth, including WHI3 encoding an RNA-binding protein and YAK1 coding for a dual-specificity tyrosine-regulated protein kinase. In this study, we present evidence that Whi3 and Yak1 form part of a signaling pathway that regulates FLO11-mediated surface adhesion and is involved in stress resistance. Our study further suggests that Whi3 controls YAK1 expression at the post-transcriptional level and that Yak1 targets the transcriptional regulators Sok2 and Phd1 to control FLO11. We also discovered that Yak1 regulates acidic stress resistance and adhesion via the transcription factor Haa1. Finally, we provide evidence that the catalytic PKA subunit Tpk1 inhibits Yak1 by targeting specific serine residues to suppress FLO11. In summary, our data suggest that Yak1 is at the center of a regulatory cascade for adhesive growth and stress resistance, which is under dual control of Whi3 and the PKA subunit Tpk1.
THE budding yeast Saccharomyces cerevisiae has evolved complex signaling networks to adapt cell growth and development in response to changing nutritional states and stress situations (De Virgilio and Loewith 2006; Zaman et al. 2008). These networks control expression of numerous genes that are involved in cell division, metabolic pathways, stress resistance, and cell adhesion. A central target of a growing number of signaling pathways is FLO11, a gene that belongs to the fungal adhesin family and encodes a GPI-anchored cell wall-associated glycoprotein (Lo and Dranginis 1996). Flo11 confers adhesion of yeast cells to agar substrates and plastic surfaces and it plays a critical role during biofilm formation and filamentous growth (Lo and Dranginis 1998; Guo et al. 2000; Reynolds and Fink 2001). The regulation of FLO11 is highly complex, and this is reflected by the unusually large FLO11 promoter region with a size of >3 kb (Rupp et al. 1999). The FLO11 promoter is under control of several conserved signaling cascades, including the cAMP-PKA pathway (Robertson and Fink 1998; Pan and Heitman 1999; Rupp et al. 1999), the Fus3/Kss1-MAPK cascade (Madhani et al. 1999; Rupp et al. 1999), the Snf1 pathway (Kuchin et al. 2002), the general amino acid control system (Braus et al. 2003), and the TOR network (Vinod et al. 2008). To regulate FLO11, these pathways have been shown to target numerous transcription factors, including Flo8 and Sfl1 (Robertson and Fink 1998; Pan and Heitman 2002), Tec1 and Ste12 (Madhani et al. 1999; Rupp et al. 1999), Nrg1 and Nrg2 (Kuchin et al. 2002), Gcn4 (Braus et al. 2003), Sok2 and Phd1 (Gimeno and Fink 1994; Ward et al. 1995; Pan and Heitman 2000), Mss1 (Gagiano et al. 1999), and Mga1 (Lorenz and Heitman 1998; Borneman et al. 2006). In addition, a number of regulatory genes have been identified that control adhesion in S. cerevisiae, including WHI3 encoding an RNA binding protein (Mösch and Fink 1997; Gari et al. 2001) and YAK1 encoding a dual-specificity tyrosine-regulated protein kinase (DYRK) (Garrett et al. 1991; Zhang et al. 2001). The exact mechanisms by which Whi3 and Yak1 regulate yeast adhesion, however, are largely unknown. Finally, FLO11 expression is under epigenetic control mechanisms that involve the histone deacetylases Hda1 and Rpd3L and a pair of cis-interfering noncoding (nc) RNAs (Halme et al. 2004; Bumgarner et al. 2009).
The WHI3 gene is involved in the regulation of cell division and cellular development. Deletion mutants bud and enter S phase at a smaller volume than wild-type cells and overexpression of WHI3 produces a lethal G1 phase arrest, indicating that Whi3 plays a role in setting the cell size at the G1/S transition (Nash et al. 2001). WHI3 negatively regulates CLN3, which encodes a G1 cyclin that promotes G1/S phase transition (Gari et al. 2001). The Whi3 protein contains a C-terminal RNA recognition motif (RRM) that binds CLN3 mRNA and an N-terminal domain that interacts with Cdc28 (Wang et al. 2004). It has thus been suggested that Whi3 might act as a cytoplasmic retention factor that sequesters Cdc28p and associated cyclins, thereby restricting the nuclear accumulation of these complexes to late G1 phase. The importance of WHI3 in coordinating cell cycle progression is further emphasized by the fact that whi3 mutations affect transcription by the cell cycle-regulated SBF and MBF protein complexes and cell cycle regulation of histone modifications (Schulze et al. 2009). WHI3 is further involved in several developmental events, including mating and adhesive growth of haploid strains as well as filament formation and meiosis of diploids (Mösch and Fink 1997; Gari et al. 2001). Global studies have identified a large number of mRNAs that are bound by the Whi3 RRM domain (Colomina et al. 2008, 2009). A significant fraction of these Whi3 targets encode membrane and exocytic proteins involved in transport and cell wall biogenesis, indicating that Whi3 might be a general modulator of protein fate that affects cell integrity and organization of the actin cytoskeleton. How exactly Whi3 controls these diverse developmental options of yeast cells is not known.
Yak1 is a member of the family of dual-specificity tyrosine-regulated protein kinases, which autophosphorylate their activation loop on an essential tyrosine, but phosphorylate their substrates on serine and threonine (Becker and Joost 1999). In S. cerevisiae, YAK1 was originally identified as a gene whose deletion suppresses loss of function of the cAMP-PKA pathway and whose expression is highly induced by arrest in early cell cycle, suggesting that Yak1 acts downstream and/or in parallel to PKA (Garrett and Broach 1989; Garrett et al. 1991). In vitro, Yak1 is phosphorylated by the catalytic PKA subunit Tpk1 (Zhu et al. 2000b; Budovskaya et al. 2005), but the physiological consequence of this modification is not known. In vivo, Yak1 is phosphorylated at several serine and threonine residues (for overview see www.uniprot.org), whose functions, however, have not been investigated. The nucleocytoplasmic distribution of Yak1 is also regulated, and it has been shown that glucose starvation or rapamycin-induced inhibition of TOR stimulates nuclear accumulation of this protein kinase (Martin et al. 2004). Several targets of Yak1 have been identified. In response to glucose starvation, the regulatory PKA subunit Bcy1 is phosphorylated and restricted to the cytoplasm in an Yak1-dependent manner (Griffioen et al. 2001), but the physiological role of this regulation is not clear. Glucose starvation further stimulates Yak1-mediated phosphorylation of Pop2/Caf1, which forms part of the Ccr4-Caf1-Not deadenylation complex that controls expression of numerous genes involved in stress response and carbohydrate metabolism (Moriya et al. 2001). Yak1-induced phosphorylation of Pop2 has thus been thought to be essential for arrest in G1 at the end of postdiauxic growth prior to entry of cells into stationary phase. A further target of Yak1 is the corepressor Crf1, which translocates to the nucleus and inhibits transcription of ribosomal genes upon phosphorylation by Yak1 (Martin et al. 2004). Activation of Crf1 by Yak1 is under negative control of TOR via PKA, thereby connecting nutrient sensing to ribosome biogenesis. Finally, the stress responsive transcription factors Hsf1 and Msn2/Msn4 are direct targets for Yak1 phosphorylation (Lee et al. 2008). It has been shown that these transcription factors are under negative control of PKA, but that they are activated by Yak1 (Smith et al. 1998). In summary, Yak1 appears to be part of a nutrient responsive signaling pathway that acts in parallel to PKA, but with opposite effects. Interestingly, YAK1 has also been found to positively regulate cell adhesion and filamentation (Zhang et al. 2001), but the exact mechanisms for this regulation are not known.
In this study, we focused on the functions of Whi3 and Yak1 in regulating FLO11-mediated cell adhesion. We find that Whi3 and Yak1 form part of a signaling pathway that controls FLO11 expression and adhesion by targeting the transcription factors Sok2, Phd1, and Haa1. We also find that this pathway controls adaptation of yeast cells to acidic stress. Furthermore, our study shows that the Yak1 pathway is connected to the cAMP-PKA pathway via the catalytic subunit Tpk1 that negatively regulates Yak1 function by targeting three serine residues within conserved PKA phosphorylation sites.
MATERIALS AND METHODS
Yeast strains and growth conditions:
All yeast strains used in this study are of the Σ1278B genetic background and are listed in Table 1. Yeast strain YHUM468 is identical to MY1398 and is a kind gift of Microbia (Lexington, MA). Full deletions of WHI3, YAK1, TPK1, TPK2, and TPK3 were introduced using appropriate whi3Δ∷HIS3, yak1Δ∷kanMX4, tpk1Δ∷loxP-kanR-loxP, tpk2Δ∷loxP-kanR-loxP, or tpk3Δ∷loxP-kanR-loxP deletion cassettes. Deletions of MSN2, MSN4, SOK2, and HAA1 were introduced as described (Gueldener et al. 2002), using plasmid pUG72 as template and gene-specific primers (supporting information, Table S1). Yeast culture medium was prepared as described (Guthrie and Fink 1991). Adhesive growth tests were described earlier (Roberts and Fink 1994).
Screen for high-copy suppressors of whi3Δ:
A URA3-based yeast genomic high-copy plasmid library (Connelly and Hieter 1996) was transformed into the nonadhesive whi3Δ strain YHUM1105, and ∼10,000 transformants were selected by growth on solid SC medium lacking uracil. After replica plating and incubation for 4 days at 30°, plates were subjected to a wash test to identify adhesive colonies. Plasmids conferring adhesive growth were rescued and verified by retransformation into YHUM1105. The genomic segment present in each of the selected clones was determined by sequence analysis, revealing two different clones containing YAK1 and a single clone carrying PHD1.
Plasmids used in this study are listed in Table 2. Plasmids BHUM1058, BHUM1187, and BHUM1188 were obtained by subcloning of a 3.6-kb HindIII fragment carrying YAK1 from the original isolate of the pRS202-based genomic library (Connelly and Hieter 1996). Plasmids BHUM1189, BHUM1190, and BHUM1740–BHUM1752 were obtained by site-directed mutagenesis using BHUM1187 as the template and appropriate primers (Table S1). Mutations were verified by sequence analysis and the resulting YAK1 alleles were inserted into pRS316 or pRS202, respectively. Plasmid BHUM1737 was obtained by PCR amplification of a SalI/BamHI fragment carrying HAA1 from yeast genomic DNA using appropriate primers (Table S1) and subsequent insertion into plasmid YEplac195. Plasmids BHUM1738 and BHUM1739 were created by PCR amplification of SmaI/PstI fragments carrying either MSN2 or MSN4 and insertion into YEplac195. Construction of plasmid BHUM1574 carrying the FLO11 open reading frame under control of the PGK1 promoter will be described elsewhere.
Yeast cultures were grown to exponential growth phase and total RNA was prepared following the instruction manual of TRIZOL reagent (Invitrogen, Karlsruhe, Germany). For each sample, 17 μg of total RNA was separated on a 1.2% agarose gel. After transfer to a positively charged nylon membrane FLO11, YAK1, and ACT1 transcripts were detected by using gene-specific DIG-labeled DNA probes following the instruction manual for DIG filter hybridization (Roche Diagnostics, Mannheim, Germany). Signals were detected by a Chemo Star Imager (INTAS Science Imaging Instruments, Göttingen, Germany) and quantified using the Lab Image 1D software (Kapelan Bio-Imaging, Leipzig, Germany).
Transcriptional profiling of yeast strains YHUM468 (control) and YHUM1105 (whi3Δ), carrying plasmid B2338 and of strain YHUM1313 (yak1Δ), carrying the plasmids B2338 (control), BHUM1058 (high-copy YAK1), or BHUM1189 (high-copy YAK1K398R), respectively, was performed using Yeast Genome 2.0 expression arrays (Affymetrix, High Wycombe, United Kingdom) following standard protocols. Briefly, two independent colonies of each yeast strain were inoculated in 10 ml YNB medium supplemented with tryptophan and grown overnight. Cultures were diluted into 30 ml of fresh medium to an OD600 of 0.25 and grown to an OD600 of 1.0 before cells were harvested by centrifugation and rapidly frozen in liquid nitrogen. Total RNA was prepared using an RNeasy Mini kit (Qiagen, Hilden, Germany) following yeast protocol 1c for mechanical disruption. RNA yield and purity were determined using an Agilent 2100 bioanalyzer (Agilent Technologies, Böblingen, Germany). All subsequent steps were conducted according to the Affymetrix Gene Chip Expression Manual (Affymetrix, High Wycombe, United Kingdom). Briefly, one cycle of cDNA synthesis was performed with 5–8 μg of total RNA. In vitro transcription labeling was carried out for 16 hr. The fragmented samples were hybridized for 16 hr on Affymetrix Yeast Genome 2.0 expression arrays, washed, and stained using the GeneChip Hybridization Wash and Stain kit (P/N 900720) and an Affymetrix Fluidics 450 station. Arrays were scanned on an Affymetrix GeneArrayScanner 3000 7G, and nonscaled RNA signal intensity (CEL) files were generated using the Affymetrix Command Console software. The resulting CEL files were loaded into the Affymetrix Expression Console software and CHP files were created using quantile normalization and probe logarithmic intensity error estimation (PLIER). Differentially expressed genes were defined as having an average signal intensity of >50 and a fold change of at least 1.5 (calculated as the fold change of the average expression in the duplicate measurements). Array data are available from ArrayExpress database (http://www.ebi.ac.uk) under accession nos. E-MEXP-2810 (WHI3) and E-MEXP-2885 (YAK1).
Preparation of yeast cell extracts and immunoblot analysis:
Preparation of total yeast cell extracts was performed as described (Kushnirov 2000). Two milliliter yeast cultures of OD600 = 1.0 were concentrated to 1 ml, mixed with 150 μl lysis buffer (1.85 m NaOH, 7.5% (v/v) β-mercapto-ethanol) and incubated on ice for 10 min. After mixing with 150 μl trichloroacetic acid, samples were centrifuged for 10 min at 13,000 rpm. After removal of the supernatant, 100 μl urea buffer (5% SDS, 8 m urea, 92.6 mm Na2HPO4, 107.4 mm NaH2PO4, 1.5% dithiothreitol, 0.1 mm EDTA, pH 6.8, bromophenol blue) and 2 μl of 2 m Tris-Cl were added to the pellet and shaken for 20 min at 37°. Samples were subjected to 5-min centrifugation at 13,000 rpm and 5–10 μl of the supernatant were used for further analysis. For immunoblot analysis, equal amounts of protein were separated by SDS–PAGE using 12% gels and transferred to nitrocellulose membranes by electrophoresis at 35 V overnight using a Mini-PROTEAN 3 electrophoresis system (Bio-Rad). GFP-Yak1, Yak1-HA, Tub1, or Cdc28 proteins were detected using enhanced chemiluminescence (ECL) technology after incubation of membranes with monoclonal mouse anti-GFP (Santa Cruz Biotechnology, Santa Cruz, CA), monoclonal mouse anti-HA (Sigma Aldrich, Munich, Germany), monoclonal mouse anti-tubulin (Calbiochem/Merck, Darmstadt, Germany), or polyclonal goat anti-Cdc28 antibodies (Santa Cruz Biotechnology). As secondary antibodies, horseradish peroxidase-coupled goat anti-mouse (Dianova, Hamburg, Germany) or mouse anti-goat (Santa Cruz Biotechnology) antibodies were used, respectively. Signals were detected by the Chemo Star Imager (INTAS Science Imaging Instruments) and quantified using the Lab Image 1D software (Kapelan Bio-Imaging).
Yeast strains carrying the PFLO11-lacZ reporters were grown to OD600 of 1.0 in appropriate YNB media, and extracts were prepared and assayed for β-galactosidase activity as previously described (Heise et al. 2010). β-Galactosidase activity was normalized to the total protein in each extract with the following formula: (optical density at 420 nm × 0.304)/(0.0045 × protein concentration × extract volume × time). Assays were performed in triplicate on at least three transformants, and the mean values were calculated. Standard deviations did not exceed 20%.
Yeast strains expressing GFP-YAK1 were cultivated in appropriate medium to exponential growth phase, before cells were visualized on a Zeiss Axiovert by differential interference contrast (DIC) and fluorescence microscopy microscope using a GFP filter set (AHF Analysentechnik AG, Tübingen, Germany). Cells were photographed using an Orca ER digital camera (Hamamatsu, Bridgewater, NJ) and the Openlab software (Improvision). Image processing was carried out using the ImageJ software (Open Source). Nucleocytoplasmic distribution of GFP-Yak1 was quantified as described (Leclerc et al. 1998). Briefly, nuclear and cytoplasmic staining of at least 80 cells was quantified using the ImageJ software and the mean values were calculated and corrected for background fluorescence measured in control strains carrying no GFP. The nucleocytoplasmic (n/c) ratio R was then calculated using the formula R = N−C/N+C, where N and C correspond to the corrected mean values obtained for the nuclear (N) and cytoplasmic (C) staining, respectively.
WHI3 controls adhesive growth by regulating the FLO11 promoter:
Previous studies have shown that whi3Δ mutants are suppressed for agar adhesion and adhesive filament formation (Mösch and Fink 1997; Gari et al. 2001). To test how Whi3 affects adhesion, we measured WHI3-dependent expression of the FLO11 gene. We found that FLO11 transcript levels drop more than sevenfold in the absence of WHI3, which is paralleled by a complete loss of agar adhesion (Figure 1, A and B). In addition, absence of WHI3 caused a more than sixfold decrease in the expression of a PFLO11-lacZ reporter gene (Figure 1C), indicating that Whi3 regulates adhesion by controlling activity of the FLO11 promoter. In support of this conclusion, we found that expression of the FLO11 open reading frame from the heterologous PGK1 promoter is sufficient to confer efficient adhesion and FLO11 transcription in a whi3Δ mutant strain (Figure 1, A and B).
YAK1 is a high-copy suppressor of a whi3 mutation and a dosage-dependent activator of FLO11:
We next isolated high-copy suppressors of the adhesion defect caused by a whi3Δ mutation, to identify putative effectors of Whi3. We found two genes that were able to induce adhesion in the absence of WHI3, YAK1, and PHD1. We did not uncover FLO11 as a multicopy suppressor of a whi3 mutation, which might reflect that the screen was not saturated or that the library does not contain functional FLO11-expressing plasmids. However, we first focused on YAK1 and found that overexpression of this gene was sufficient to confer efficient FLO11 expression and adhesion in a whi3Δ mutant strain (Figure 2, A and B). Moreover, high-copy expression of YAK1 led to hyperadhesive growth and enhanced FLO11 transcript levels in a WHI3 control strain. These data suggest that YAK1 is a dosage-dependent activator of adhesion. To corroborate this conclusion, we measured FLO11 transcript levels and agar adhesion in mutant strains that either completely lack YAK1 or that express YAK1KD, a variant in which the kinase domain is inactivated by a K398R kinase-dead (KD) mutation (Martin et al. 2004). These experiments revealed that absence of Yak1 or expression of the Yak1KD variant led to a strong decrease of FLO11 expression and a loss of adhesion (Figure 2, C, D, and E). Deletion of YAK1 also caused a decrease in PFLO11-lacZ reporter gene expression comparable to strains lacking WHI3 (Figure 1C). Taken together, these data demonstrate that the Yak1 kinase is an efficient regulator of FLO11-mediated adhesion and they suggest that Yak1 acts downstream of Whi3.
Whi3 controls YAK1 expression at the post-transcriptional level:
We reasoned that if Yak1 is an effector of Whi3, expression of YAK1 might be affected in whi3Δ mutant strains. Therefore, we expressed a functionally tagged version of YAK1, YAK1-HA (Martin et al. 2004), in WHI3 and whi3Δ strains and determined both YAK1-HA transcript levels and corresponding amounts of Yak1-HA protein. We found that absence of Whi3 did not affect YAK1-HA transcripts, but led to a threefold decrease of Yak1-HA protein levels (Figure 3A). This suggests that Whi3 affects YAK1 expression at the post-transcriptional level. To validate these data and to measure the influence of Whi3 on the localization of Yak1, we also expressed a functional GFP-YAK1 variant (Moriya et al. 2001) in WHI3 and whi3Δ strains. Again, absence of Whi3 did not affect GFP-YAK1 transcript levels, but led to a strong reduction in GFP-Yak1 protein amounts as determined by quantitative immunoblot analysis (Figure 3B). A strong decrease of GFP-Yak1 in whi3Δ mutant strains could also be observed by live cell imaging (Figure 3C). However, absence of Whi3 did not affect the nucleocytoplasmic distribution of GFP-Yak1. These data demonstrate that Whi3 controls YAK1 expression at the post-transcriptional level and they further support the view that Yak1 acts downstream of Whi3.
Whi3 and Yak1 are required for stress-induced stimulation of FLO11:
We next asked whether Whi3 and Yak1 might be required for stimulation of FLO11 in response to certain stress signals. Specifically, we tested glucose and amino acid starvation, which are nutritional stress conditions that stimulate FLO11 expression (Kuchin et al. 2002; Braus et al. 2003). In agreement with previous studies, we measured between roughly five- to sixfold higher FLO11 transcript levels in yeast strains that were grown in low glucose medium or in the presence of 3AT, a histidine analog that induces histidine starvation (Figure 4A). In strains lacking Whi3, we found a complete block in glucose starvation-induced stimulation of FLO11. Also, whi3Δ mutant strains were significantly suppressed for stimulation of FLO11 by 3AT. Thus, Whi3 appears to be involved in regulation of FLO11 by glucose and amino acids. In yak1Δ mutant strains, we also found that stimulation of FLO11 by glucose starvation was completely blocked, whereas induction by amino acid starvation was not significantly affected. This indicates that Yak1 is involved in glucose regulation of FLO11, which is supported by the fact that glucose starvation, but not amino acid starvation, stimulated expression and nuclear accumulation of GFP-Yak1 (Figure 4, B and C).
Yak1 is known to target the heat shock transcription factor Hsf1 (Lee et al. 2008). We asked, whether FLO11 might be regulated by heat stress. Indeed, growth at elevated temperature was paralleled by an increase in FLO11 transcript levels, and this effect was completely blocked in whi3Δ and yak1Δ mutant strains (Figure 4D). Thus, Whi3 and Yak1 also seem to be involved in heat stress-induced stimulation of FLO11 expression.
Yak1 targets FLO11 via Sok2 and Phd1, but not via Msn2/Msn4:
In a next step we wanted to identify transcription factors that act downstream of Whi3 and Yak1 to control FLO11. For this purpose, we measured transcriptional profiles of whi3Δ and yak1Δ strains using high-density oligonucleotide arrays (Affymetrix GeneChips). We also included strains overexpressing YAK1 in our analysis, because we found that YAK1 is a dosage-dependent activator of FLO11 expression and because YAK1 expression itself is highly upregulated by nutritional starvation (Garrett et al. 1991). Strains overexpressing WHI3 were not analyzed, because they are not viable (Nash et al. 2001). Analysis of these transcript profiles revealed that 282 genes are regulated at least 1.5-fold by Whi3 (Figure 5A; comparison of whi3Δ and WHI3 strains) and that 290 genes are regulated by Yak1 (Figure 5A; comparison of yak1Δ and YAK1 overexpressing strains). Moreover, 46 genes were found to be regulated by both Whi3 and Yak1. A representation factor of 3.3 (RF; a measure of the observed number of overlapping genes compared with the expected number) was calculated for this overlap (P < 1.5 × 10−13) supporting significant cooperative interaction between Whi3 and Yak1. To identify transcription factors that might act downstream of Yak1, we analyzed 209 genes that are positively regulated by Yak1 by concurrent enrichment analysis for gene ontology (GO) terms and transcription factor binding sites (Abascal et al. 2008; Morris et al. 2010). This class of genes, which includes FLO11, is enriched for functions associated with heat response, stress response, and carbohydrate metabolism (Figure 5B). Moreover, these genes are enriched for binding by the known stress regulators Msn2, Msn4, Hsf1, and Skn7 (Wiederrecht et al. 1988; Brown et al. 1993; Martinez-Pastor et al. 1996) and for the transcription factor Sok2 (Ward et al. 1995). This analysis corroborates the previous finding that Yak1 targets Msn2 and Hsf1 to control gene expression (Lee et al. 2008) and it identifies Sok2 and Skn7 as potential Yak1 targets.
We next tested the possibility that Yak1 might regulate FLO11 via the transcription factors Msn2, Msn4, or Hsf1. We found that Msn2 and Msn4 are required for maximal expression of FLO11 (Figure 6A). However, overexpression of neither MSN2 nor MSN4 was sufficient to suppress the invasive growth defects of a yak1Δ mutant strain (Figure 6B). Moreover, yeast strains lacking either MSN2 or MSN4 or both genes were not suppressed for agar adhesion and did not block stimulation of adhesion induced by overexpression of YAK1 (Figure 6C). These data indicate that Msn2/Msn4 do not mediate Yak1 regulation of FLO11-dependent adhesion. Genetic interaction between YAK1 and HSF1 with respect to adhesion could not be measured, because deletion or overexpression of the HSF1 gene causes inviability (data not shown).
We then tested the genetic interaction between YAK1 and SOK2 with respect to FLO11 expression and adhesion. Sok2 is a known repressor of FLO11 expression (Pan and Heitman 2000) and Sok2 was previously found to bind to the FLO11 promoter in vivo (Borneman et al. 2006). Here we found that deletion of SOK2 led to an increase in FLO11 transcript levels both in control and yak1Δ mutant strains (Figure 6D). Also, a sok2Δ mutation caused a stimulation of yeast cell adhesion in yak1Δ mutant strains (Figure 6E). Together, these data suggest that Yak1 regulates adhesion via Sok2. Finally, we measured the genetic interaction between YAK1 and PHD1, because the transcription factor Phd1 was previously proposed to act downstream of Sok2 in regulating FLO11 (Pan and Heitman 2000) and because we have isolated PHD1 as a high-copy suppressor of a whi3Δ mutation. Indeed, the defects in FLO11 expression and adhesion caused by a yak1Δ mutation is suppressed by overexpression of PHD1 (Figure 6, D and E), supporting the view that Yak1 controls FLO11 via Sok2 and Phd1.
Yak1 and Haa1 coregulate resistance to acidic stress and FLO11-mediated adhesion:
Not all yeast transcription factors have been analyzed for genome-wide in vivo promoter binding, but for some of these regulators transcriptional profiles are available. Therefore, we analyzed the class of Yak1-regulated genes for above-average coregulation by transcription factors that are not present in chip-on-chip databases. Here, we found that Yak1-regulated genes include 5 of the 10 genes that were previously identified to be regulated by the transcription factor Haa1 (Keller et al. 2001), namely GRE1, PHM1, TPO2, YGP1, and YRO2 (Figure 7A). Previous work has shown that Haa1 and Haa1-regulated genes are required for adaptation of yeast to weak acid stress, e.g., to millimolar concentrations of acetic acid (Fernandes et al. 2005). Therefore, we tested whether Yak1 is required for resistance to acetic acid. Indeed, yak1Δ mutant strains are more sensitive to acetic acid than a control strain, albeit to a lesser degree than strains lacking Haa1 (Figure 7B). In addition, acetic acid sensitivity of a yak1Δ strain can be suppressed by overexpression of HAA1. Thus, Yak1 seems to be involved in the response to acidic stress and might function upstream of Haa1.
We next tested, whether FLO11 might be regulated by Haa1 and acidic stress. Previous studies did not address this question, because they were done in nonadhesive yeast strains that are suppressed for FLO11 expression. For this purpose, we measured FLO11 transcript levels in haa1Δ mutant strains and in cultures exposed to acidic stress. Here, we found that exposure of control strains to acidic stress did not affect FLO11 expression (Figure 7C). Also, FLO11 transcript levels were not altered in haa1Δ mutant strains grown under nonstress conditions. Upon exposure to acidic stress, however, FLO11 transcript levels dropped significantly in strains lacking Haa1 and the reduction was comparable to strains lacking Yak1 (Figure 7C). Thus, Haa1 appears to be an important activator of FLO11 under acidic stress conditions. The effect of acidic stress on agar adhesion in haa1Δ mutant strains could not be assayed due to their growth defect under these conditions.
We next reasoned that the activating potential of Haa1 toward FLO11 might become apparent under nonstress conditions, if the HAA1 gene was overexpressed. Indeed, control strains carrying HAA1 on a high-copy plasmid were found to be hyperadhesive (Figure 7D). This finding enabled us to perform genetic interaction studies between HAA1 and YAK1. Here, we found that overexpression of HAA1 is sufficient to suppress the adhesion defects of a yak1Δ mutant strain (Figure 7D). Also, overexpression of HAA1 was sufficient to induce adhesive growth in a whi3Δ mutant, but not as efficiently as in yak1Δ strains. Together, these data suggest that Haa1 acts downstream of Yak1 to regulate not only acidic stress resistance, but also FLO11-mediated adhesion.
Tpk1 negatively regulates FLO11 and adhesion by phosphorylation of Yak1 at S206, S240, and S295:
A number of previous studies indicate that Yak1 is under control of PKA (Garrett and Broach 1989; Martin et al. 2004) and others have shown that the catalytic subunit Tpk1 efficiently phosphorylates Yak1 in vitro (Zhu et al. 2000b; Budovskaya et al. 2005). Therefore, we further investigated the role of Yak1 phosphorylation, specifically with respect to controlling FLO11 and adhesion. For this purpose, we individually mutated all serine, threonine, and tyrosine residues outside of the conserved kinase domain that were previously identified to be phosphorylated in vivo by phosphoproteome analysis studies (Figure 8A), and tested the mutated YAK1 variants for functionality. None of the mutations led to a significant decrease in FLO11 transcript levels (Figure 8B) or suppressed adhesion (Figure 8C), indicating that phosphorylation of the mutated residues is not required for Yak1 activity. In contrast, mutation of either S206, S240, or S295 led to a stimulation of FLO11 expression and adhesion, indicating that phosphorylation of these residues inhibits Yak1. This conclusion was supported by the finding that a combination of S206A with S295A or mutation of all three residues had additive effects and caused an up to fivefold stimulation of FLO11 expression and hyperadhesive growth (Figure 8, B and C).
We noticed that S206, S240, and S295 are all part of conserved PKA phosphorylation sites (Figure 8A), indicating that PKA might inhibit Yak1 through these residues. To test this possibility, we measured genetic interactions between YAK1 and the three TPK genes encoding the different catalytic PKA subunits. Specifically, we expressed the hyperactive YAK1S206A,S240A,S295A variant in yeast strains that carry only one of the three TPK genes, which allows distinguishing between subunit-specific activities. In agreement with previous studies, we found that a strain that carries only the TPK1 gene is defective for agar adhesion. Moreover, additional expression of the regular YAK1 gene from a low-copy plasmid was not sufficient to suppress the adhesive growth defect of this strain (Figure 8D). However, low-copy expression of the YAK1S206A,S240A,S295A variant efficiently suppressed the adhesion defect, indicating that Tpk1 negatively regulates Yak1 via S206, S240, and S295. The conclusion that Tpk1 is a negative regulator of Yak1 and FLO11 is further supported by our finding that the adhesion defect of the TPK1-expressing strain was suppressed by high-copy expression of YAK1 (Figure 8D) and that a tpk1Δ strain has increased FLO11 transcript levels (Figure S1).
In the case of a strain expressing only TPK2, we observed hyperadhesive growth (Figure 8D). A positive role of Tpk2 in controlling FLO11 has previously been shown (Robertson and Fink 1998; Pan and Heitman 1999) and is supported by our finding that FLO11 expression is drastically reduced in a tpk2Δ mutant strain (Figure S1). However, hyperadhesive growth of a TPK2-expressing strain was not significantly affected by additional expression of YAK1 or YAK1S206A,S240A,S295A or by high-copy expression of YAK1 (Figure 8D). This suggests that Tpk2 controls adhesion independently of Yak1. Finally, analysis of a yeast strain carrying only the TPK3 gene revealed that in this genetic background expression of YAK1S206A,S240A,S295A or YAK1 from a high-copy plasmid did not lead to a significant increase in agar adhesion when compared to a strain expressing YAK1 from a low-copy plasmid (Figure 8D). This suggests that Tpk3 does not regulate FLO11 via Yak1 phosphorylation at S206, S240, and S295. However, Tpk3 seems to regulate FLO11 by Yak1-independent mechanisms because we found that FLO11 expression is slightly induced in a tpk3Δ strain, which is in agreement with previous studies (Robertson and Fink 1998; Pan and Heitman 1999).
In S. cerevisiae, Yak1 is the sole member of the eukaryotic DYRK family protein kinases that regulate numerous cellular processes (Becker and Joost 1999). Previous studies have shown that Yak1 acts downstream or in parallel to the PKA pathway (Garrett and Broach 1989; Martin et al. 2004) and that Yak1 controls transcription factors involved in stress response (Lee et al. 2008) and ribosomal biogenesis (Martin et al. 2004). Here, we have identified Sok2, Phd1, and Haa1 as transcriptional regulators that mediate Yak1 control of FLO11-mediated cell surface adhesion and, in the case of Haa1, Yak1-dependent resistance to acidic stress. Moreover, we provide novel insights into the mechanisms for the regulation of Yak1 by the RNA-binding protein Whi3 and the PKA catalytic subunit Tpk1.
Our transcriptional profiling and genetic analysis suggests that Yak1 activates FLO11-mediated adhesion by targeting the transcriptional repressor Sok2, but not via the transcriptional activators Msn2/4. Previous studies have shown that Sok2 negatively regulates FLO11 by suppressing PHD1 (Pan and Heitman 2000) and that Sok2 binds directly to both the promoters of PHD1 and FLO11 (Borneman et al. 2006). Our data indicate that Yak1 negatively controls Sok2, thus relieving FLO11 expression by inhibition of a feed-forward repression mechanism (Figure 9A). Whether Yak1 regulates Sok2 by direct phosphorylation, as has been shown for Msn2 and Hsf1 (Lee et al. 2008), needs to be investigated in the future.
We have identified the transcription factor Haa1 as a further putative target of Yak1 that regulates FLO11 by transcriptional profiling and data mining. It is not known whether Haa1 directly binds to FLO11, but our data indicate that Haa1 is specifically required for FLO11 expression under acetic acid stress (Figure 9A). Under these conditions, Yak1 might therefore specifically depend on Haa1 to activate FLO11 and not be able to control adhesion via Sok2 and Phd1. Moreover, our data demonstrate an important role of Yak1 in controlling not only adhesion, but also acidic stress resistance. In this context, it is interesting to note that Yak1 positively regulates expression of SPI1, a gene that is highly upregulated under acetic acid stress in an Haa1-dependent manner and encodes a GPI-anchored cell wall protein that confers acidic stress resistance (Simoes et al. 2006). Thus, Yak1 might contribute to acidic stress resistance by ensuring the Haa1-mediated expression of Spi1 to protect the cell by, e.g., decreasing the porosity of the cell wall. In addition, expression of the GPI-anchored cell wall-associated adhesin Flo11 might further contribute to resistance by conferring the formation of protective biofilms.
Our study suggests that Yak1 acts downstream of the RNA binding protein Whi3 and mediates a subset of Whi3-controlled cellular functions, including stress response and cellular adhesion (Figure 9A). This conclusion is supported by several lines of evidence, including (i) genetic suppression of whi3 mutations by YAK1 overexpression, (ii) control of Yak1 protein levels, but not YAK1 mRNA, by WHI3, and (iii) significant overlap of WHI3 and YAK1 regulated genes. This suggests that Whi3 controls YAK1 at the post-transcriptional level. It has been shown that Whi3 might be a general modulator of protein fate, which either controls translation or affects the localization of mRNAs and encoded proteins by, e.g., cytoplasmic retention (Colomina et al. 2008). However, this previous study did not uncover YAK1 mRNA to be associated with Whi3 under standard conditions. This indicates that Whi3 may regulate YAK1 expression indirectly by controlling, e.g., translational regulators of YAK1. Alternatively, Whi3 might be able to bind YAK1 mRNA under nonstandard conditions and thereby control its translation.
We also found evidence that Whi3 is able to control FLO11 expression and adhesion by Yak1-independent mechanisms. This conclusion is based on the observations that (i) whi3Δ mutant strains have a strong adhesion defect but only a threefold reduction of Yak1 protein levels; (ii) yak1Δ mutants are able to induce FLO11 in response to amino acid starvation, while whi3Δ mutants are defective; and (iii) high-copy expression of HAA1 fully induces adhesion in strains lacking YAK1, but only partially in whi3Δ strains. We do not know how Whi3 controls FLO11-dependent adhesion independently of Yak1. However, our transcriptional profiling reveals that genes, which are regulated by Whi3 but not by Yak1, are enriched for binding by Ste12, Fkh1, and Fkh2 (Figure S2). These transcription factors are known to be involved in regulation of adhesion and pseudohyphal growth (Roberts and Fink 1994; Rupp et al. 1999; Hollenhorst et al. 2000; Zhu et al. 2000a), which indicates that they might confer regulation of FLO11 by Whi3.
Our data provide novel insights into the regulation of Yak1 by the PKA pathway. Previous studies have shown that PKA negatively acts on Yak1, but neither the in vivo roles of the different PKA catalytic subunits nor the identity and function of putative PKA target sites in Yak1 have been investigated. Here, we have identified Tpk1 as a central catalytic PKA subunit that negatively controls Yak1 activity in vivo with respect to controlling FLO11-mediated adhesion. Our data are in agreement with previous studies that have shown that Tpk1 is able to phosphorylate Yak1 in vitro (Budovskaya et al. 2005; Ptacek et al. 2005). More importantly, we have identified three major PKA phosphorylation sites in Yak1 that are likely to be targeted by Tpk1 in vivo and that mediate inhibition of Yak1 upon phosphorylation. The position of these sites are all located within the Yak1 fragment that has been identified to be phosphorylated by Tpk1 in vitro (Deminoff et al. 2009). We do not know how exactly Tpk1 phosphorylation affects Yak1, but Tpk1 might, e.g., affect the Yak1 kinase activity and/or its nuclear localization. It has been shown that Yak1 shuttles between the cytoplasm and the nucleus and that nucleocytoplasmic distribution is responsive to glucose starvation and rapamycin (Moriya et al. 2001; Martin et al. 2004). Here, we have corroborated nuclear accumulation of Yak1 in response to glucose starvation. As these conditions are known to negatively act on PKA (Santangelo 2006), phosphorylation by Tpk1 might be a mechanism that inhibits nuclear transport of Yak1. A detailed analysis of the different Yak1 variants addressing Tpk1 phosphorylation, kinase activity, nuclear transport, and substrate specificity will be required to further resolve these issue. With respect to the other catalytic subunits of PKA, our study excludes regulation of Yak1 by Tpk3, which is in agreement with previous large-scale in vitro phosphorylation studies (Ptacek et al. 2005). For Tpk2, in vitro studies have indicated only weak phosphorylation of Yak1 (Ptacek et al. 2005). Our study does not exclude direct regulation of Yak1 by Tpk2, but we have shown that Yak1 and Tpk2 positively regulate FLO11 in parallel-acting pathways that confer adhesion control in an additive manner.
In summary, our genetic analysis shows that Yak1 is at the center of a signaling pathway for FLO11-mediated biofilm formation and stress resistance that is under positive control of Whi3 and negatively regulated by PKA via Tpk1. Interestingly, our study reveals that PKA controls FLO11 not only positively by Tpk2 (Robertson and Fink 1998; Pan and Heitman 1999), but also negatively via Tpk1. What might be the reason for such a dual control mode? One advantage might be differential regulation of multiple target processes under varying PKA activity levels. Dual control of FLO11 by PKA ensures a high degree of adhesion independent of PKA activity and concomitant differential regulation of other processes (Figure 9B). Indeed, FLO11 expression and adhesion are high both during exponential growth and in stationary phase (Rupp et al. 1999), whereas other processes such as glycogen accumulation or growth rate are differentially regulated (Tamaki 2007). Thus, such a regulatory circuit might ensure cell surface adhesion in rapidly growing cells during initiation of biofilm formation as well as in resting cells within biofilms that have entered stationary phase. As such, our study contributes to a better understanding of the physiological consequences of the immensely complex protein kinase interaction network in S. cerevisiae (Breitkreutz et al. 2010), and it also paves the way for studies in other yeasts such as human pathogenic Candida species, where PKA and Yak1 have been found to control adhesion to host cells (Sonneborn et al. 2000; Iraqui et al. 2005).
We are grateful to D. Kruhl for technical assistance. We thank D. Martin, M. Hall, H. Moriya, R. Birke, and Microbia for generous gifts of plasmids and yeast strains. We are thankful to G. Döhlemann, R. Kahmann, J. Kämper, and M. Vranjes for support with transcriptional profiling. This work was supported by grants from the Deutsche Forschungsgemeinschaft, DFG MO 825/1-4 and GRK 1216, by the International Max Planck Research School (IMPRS) for Environmental, Cellular and Molecular Microbiology and by the Center for Synthetic Microbiology (SYNMIKRO).
Supporting information is available online at http://www.genetics.org/cgi/content/full/genetics.110.125708/DC1.
ArrayExpress accession nos. E-MEXP-2810 and E-MEXP-2885.
Communicating editor: M. Hampsey
- Received October 6, 2010.
- Accepted December 7, 2010.
- Copyright © 2011 by the Genetics Society of America