In Saccharomyces cerevisiae, the TEA transcription factor Tec1 controls several developmental programs in response to nutrients and pheromones. Tec1 is targeted by the pheromone-responsive Fus3/Kss1 mitogen-activated protein kinase (MAPK) cascade, which destabilizes the transcription factor to ensure efficient mating of sexual partner cells. The regulation of Tec1 by signaling pathways that control cell division and development in response to nutrients, however, is not known. Here, we show that Tec1 protein stability is under control of the nutrient-sensitive target of rapamycin complex 1 (TORC1) signaling pathway via the Tip41-Tap42-Sit4 branch. We further show that degradation of Tec1 upon inhibition of TORC1 by rapamycin does not involve polyubiquitylation and appears to be proteasome independent. However, rapamycin-induced Tec1 degradation depends on the HECT ubiquitin ligase Rsp5, which physically interacts with Tec1 via conserved PxY motives. We further demonstrate that rapamycin and mating pheromone control Tec1 protein stability through distinct mechanisms by targeting different domains of the transcription factor. Finally, we show that Tec1 is a positive regulator of yeast chronological lifespan (CLS), a known TORC1-regulated process. Our findings indicate that in yeast, Tec1 links TORC1 and MAPK signaling pathways to coordinate control of cellular development in response to different stimuli.
FOR cells to appropriately respond to diverse external signals, intracellular signal-specific pathways are often interconnected to allow integration of multiple signals and to ensure activation of the correct cellular programs. The budding yeast Saccharomyces cerevisiae is well suited to study the mechanisms of signal integration in eukaryotes, because numerous evolutionarily conserved signaling pathways are present in this organism, including a pheromone-responsive mitogen-activated protein kinase (MAPK) pathway (Bardwell 2005) as well as nutrient-sensitive target of rapamycin complex 1 (TORC1) (De Virgilio and Loewith 2006) and Ras/protein kinase A (PKA) pathways (Tamaki 2007). All of these pathways have been implicated in the control of cell division and development (Zaman et al. 2008), but their interconnection is not understood in detail.
The pheromone-responsive MAPK pathway of S. cerevisiae has been well studied (Chen and Thorner 2007). It controls not only sexual mating, but also vegetative adhesion required for the formation of biofilms and filaments (Liu et al. 1993; Roberts and Fink 1994), and therefore is also referred to as the mating and adhesive/filamentous growth MAPK cascade. Execution of these distinct cellular programs is under control of both shared and program-specific components. The shared components include the two MAPKs Fus3 and Kss1, the MAPK kinase Ste7, and the transcription factor Ste12, which positively control expression of both mating and vegetative adhesin genes (Roberts and Fink 1994). The program-specific components include the transcription factor Tec1, which belongs to the TEA domain (TEAD) family of transcriptional regulators, which control cellular development in many eukaryotes (Andrianopoulos and Timberlake 1991; Anbanandam et al. 2006). In S. cerevisiae, Tec1 is required for adhesion and the expression of the vegetative adhesin gene FLO11 (Gavrias et al. 1996; Mösch and Fink 1997; Lo and Dranginis 1998; Rupp et al. 1999). Tec1 is not required for expression of most mating-specific genes, e.g., the sexual agglutinin gene FUS1 (Zeitlinger et al. 2003). The mating and vegetative adhesion programs not only differentially control gene expression, but also cell division. Whereas mating depends on sexual partner cells arresting in G1, robust biofilms and filaments are best formed by dividing cells. In haploid cells, this difference is reflected by the fact that G1 cyclins are down-regulated when yeast cells switch from vegetative adhesion to mating (Wittenberg et al. 1990; Brückner et al. 2004). The cyclin gene CLN1 is activated by the transcription factor Tec1 during biofilm formation (Madhani et al. 1999). In contrast, Tec1-dependent activation of CLN1 is lost during mating due to Fus3-mediated phosphorylation of the transcription factor at residue Thr273 within a conserved CPD motif (Cdc4-phosphodegron), which confers ubiquitylation of Tec1 by the ubiquitin-ligase SCFCdc4 and subsequent degradation (Bao et al. 2004; Brückner et al. 2004; Chou et al. 2004). A block of Tec1-degradation during mating not only interferes with down-regulation of CLN1 expression, but also with an efficient G1 arrest (Brückner et al. 2004). In addition, Tec1 is stabilized by complex formation with Ste12 by a mechanism that is independent of the CPD, but involves the C-terminal part of Tec1 (Chou et al. 2006; Heise et al. 2010).
In S. cerevisiae, cell division and cellular development are also controlled by the nutrient-sensitive TORC1 signaling network, which is negatively regulated by the macrocyclic lactone rapamycin (Crespo and Hall 2002; De Virgilio and Loewith 2006). Inhibition of TORC1 by higher doses of rapamycin causes downregulation of G1 cyclins and leads to a G1 arrest (Heitman et al. 1991; Barbet et al. 1996). Lower doses of rapamycin lead to the inhibition of filament formation (Cutler et al. 2001) and to an extension of yeast chronological lifespan (CLS) (Powers et al. 2006). Numerous TORC1-regulated processes have been elucidated and a number of signaling pathways are known that either act downstream of TORC1 or that function in parallel, but share common targets (De Virgilio and Loewith 2006). Important signaling components acting downstream of TORC1 include the PP2A-like protein phosphatases Sit4 and Pph21/22, their regulatory subunit Tap42, and the Tap42-interacting protein Tip41 (Düvel and Broach 2004). These components regulate a major branch of the TORC1 network and control genes involved in stress regulation, nitrogen catabolite repression, and retrograde signaling, and they connect TORC1 with the general amino acid control system (Rohde et al. 2008). A further important branch of TORC1 signaling is regulated by the protein kinase Sch9, which is a direct target of TORC1 and controls ribosome synthesis and cell size (Urban et al. 2007). In addition, Sch9 regulates replicative lifespan (RLS) (Kaeberlein et al. 2005) and CLS (Fabrizio et al. 2001; Powers et al. 2006; Wanke et al. 2008). Finally, amino acid uptake is also regulated by TORC1 and involves the nitrogen-regulated protein kinase Npr1 and the HECT ubiquitin ligase Rsp5 (Schmidt et al. 1998; De Craene et al. 2001; Crespo et al. 2004).
The TORC1 signaling network is interconnected to a number of other signaling cascades that control cell division and development (Rohde et al. 2008). A prominent example is the Ras/PKA pathway, which includes the small GTP-binding protein Ras2 and the cAMP-dependent protein kinase A, composed of the regulatory subunit Bcy1 and the catalytic subunits Tpk1, Tpk2, and Tpk3 (Zaman et al. 2008). With respect to controlling cell division, the Ras/PKA pathway has been suggested to control G1 progression and entry into stationary phase by targeting the transcription factors Msn2/Msn4 (Smith et al. 1998) and the protein kinase Rim15 (Pedruzzi et al. 2003), regulatory components that are also targets of TORC1. These findings lead to the current view that the TORC1 and Ras/PKA pathways act in parallel in the control of cell division (Rohde et al. 2008). With respect to controlling cellular development, the Ras/PKA pathway is known to regulate adhesive and filamentous growth via the catalytic subunit Tpk2, which targets the transcription factors Flo8 and Sfl1 to regulate expression of FLO11 (Robertson and Fink 1998; Pan and Heitman 2002). Like TORC1, the Ras/PKA pathway also regulates chronological yeast lifespan (Reinders et al. 1998; Longo 2003).
In this study, we explored the possibility that the pheromone-responsive Fus3/Kss1 MAPK cascade and the nutrient-sensitive TORC1 pathways might be interconnected, as both pathways control G1 arrest, vegetative adhesion, and filamentous growth. We focused on Tec1, because this transcription factor is an activator of G1 cyclin and vegetative adhesin genes, which are down-regulated during mating through degradation of Tec1. In contrast, nutrient starvation causes down-regulation of G1 cyclin genes, but induces expression of the adhesin gene FLO11, raising the question of how Tec1 is regulated by known nutrient-sensitive signaling pathways. Here we show that the TORC1 signaling pathway regulates Tec1 protein stability and Tec1-target gene expression by a mechanism that is distinct from the stability control exerted by the Fus3/Kss1 MAPK cascade and that does not involve the Ras/PKA pathway. We also find that Tec1 controls chronological lifespan. We suggest that in S. cerevisiae, the Fus3/Kss1 MAPK and the TORC1 pathways are interconnected by Tec1 to ensure proper control of cell division and cellular development in response to pheromone and nutrients.
Materials and Methods
Yeast strains and growth conditions
All yeast strains used in this study are described in Table 1. Strains carrying tec1Δ::HIS3, fus3Δ::TRP1, tpk1Δ::loxP-kanMX-loxP, tpk2Δ::loxP-kanMX-loxP, tpk3Δ::loxP-kanMX-loxP, npr1Δ::URA3, bul2Δ::URA3, bul1Δ::kanMX4, ste7Δ::loxP-kanMX-loxP, sit4Δ::loxP-kanMX-loxP, rsp5Δ::kanMX4, and pdr5∆::kanMX deletion alleles were obtained by transformation using respective deletion cassettes and verified by Southern blot analysis. The URA3(P)-OLE1 allele was introduced by targeting the integrative plasmid BHUM1053 to the leu2::hisG locus of RH2754.
Standard yeast culture medium was prepared essentially as described (Guthrie and Fink 1991). For rapamycin-based experiments, cultures grown to exponential growth phase were treated with 200 ng/ml of rapamycin (LC Laboratories) for 70 min or with solvent alone. For pheromone induction, 1 µm synthetic α-factor (Novabiochem) was applied for 60 min. Adhesive growth tests were performed as described by Roberts and Fink (1994). For nitrogen and amino acid starvation, cultures were grown in SC medium to exponential growth phase, washed with water, incubated in fresh SC medium (no starvation) or in SC medium lacking ammonium sulfate and/or amino acids (starvation medium), and grown for 70 min before further investigations.
Yeast chronological lifespan was analyzed as previously described (Wanke et al. 2008). Briefly, cultures of yeast strain YHUM928 (tec1∆) carrying one of the plasmids YCplac33, BHUM169, or BHUM1306 were grown overnight, diluted to an OD600 of 0.2, and grown into exponential growth phase before 2 ng/ml of rapamycin or drug vehicle was added. Colony-forming units (CFU ml−1) were determined by collecting culture aliquots regularly every day and plating serial dilutions on solid medium.
Plasmids used in this study are listed in Table 2 and primers in Supporting Information, Table S1. Plasmid BHUM570 was constructed by subcloning of a 3.1-kb PstI/BamHI fragment containing TEC1(P)-BglII-TEC1 from pME2068 into YEplac195, followed by insertion of a GFPuv fragment from pME1771 into the BglII site. BHUM576 was the result of a PCR-based mutagenesis (Köhler et al. 2002) of TEC1 in pME2068, where a frameshift led to a stop codon. BHUM750 was isolated from a yeast two-hybrid library (Fashena et al. 2000) and identified as RSP5G372-E809, representing the C-terminal half of RSP5. BHUM752 was obtained by cloning a PCR-based EcoRI/BamHI fragment using primers SB/TEC1-2 and SB/TEC1-3 in EcoRI/BamHI-digested pEG202. For BHUM765, the ORF of RSP5 was amplified using primers SB/RSP5-1 and SB/RSP5-2 and cloned as a SalI fragment into pYGEX-2T. Blasting the URA3 marker of BHUM765 with a LEU2 fragment from pUC4-ura3::LEU2 (kind gift from Y. Kassir, Haifa, Israel) resulted in BHUM1305. BHUM1521 was obtained from BHUM1305 by removing the SalI fragment containing RSP5. Plasmid BHUM1158 was obtained by PCR amplification of the URA3 promoter using primers URA3-1 and URA3-2 and insertion of the resulting fragment into pRS305 after SalI/BamHI digestion. For plasmid BHUM1053, OLE1 was amplified with primer OLE1-1 and OLE1-2 and inserted into BamHI/XbaI-digested BHUM1158. BHUM1122 was obtained by site-directed mutagenesis using primers SK-TEC1-3 and SK-TEC1-4, pME2068 as template, and the QuickChange Site-Directed Mutagenesis kit (Stratagene, Amsterdam, The Netherlands). BHUM1306 was the result of three subsequently performed site-directed mutagenesis steps using primers SB-TEC1-m1f, SB-TEC1-m1r, SB-TEC1-m2f, SB-TEC1-m2r, SB-TEC1-m3f, SB-TEC1-m3r, and pME2068 as the first template.
Two-hybrid screen and analysis
For yeast two-hybrid analysis, plasmids pEG202 and pJG4-5, a yeast genomic library cloned in vector pJG4-5 (Fashena et al. 2000), and the yeast strain EGY48-p1840 were used (kindly provided by Erica Golemis and Roger Brent). For the isolation of Tec1-interaction partners, yeast strain EGY48-p1840 carrying TEC1281-486 as a bait (BHUM752) was transformed with the yeast genomic library on pJG4-5 to obtain ∼5 × 106 transformants. Transformants were collected as a pool, and Leu-prototrophic strains were selected by growth on SC medium lacking Leu, Trp, and His, and containing 2% galactose and 1% raffinose, and assayed for β-galactosidase activity as described (Fashena et al. 2000). Library plasmids were isolated as described (Hoffman and Winston 1987) and analyzed by DNA sequencing using the ABI Prism Big Dye terminator sequencing kit and an ABI 310 Genetic Analyzer (Applied Biosystems, Weiterstadt, Germany). Interactions were verified by reintroducing library plasmids into the parental strain EGY48-p1840 carrying either BHUM752 or pEG202, followed by a growth test for Leu prototrophy.
Two-hybrid interactions between Tec1 and Rsp5 were quantified by measuring specific β-galactosidase activities of strain EGY48-p1840 that was cotransformed with the plasmids pEG202 or BHUM752 together with either pJG4-5 or BHUM750. Transformants were grown on SC −His −Trp containing 2% galactose and 1% raffinose to an optical density at 600 nm of between 1 and 2 before β-galactosidase assays were performed as described below. All assays were performed in triplicate on at least three independent transformants for each combination of plasmids.
Yeast strains carrying plasmids for two hybrid analysis or the TEC1-lacZ reporter were grown to the exponential growth phase in appropriate media, and extracts were prepared and assayed for β-galactosidase activity as described previously (Brückner et al. 2004). β-Galactosidase activity was normalized to the total protein in each extract with the following formula: (optical density at 420 nm × 1.7)/(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%.
Purification of glutathione S-transferase fusion proteins from yeast:
Extracts of strains expressing glutathione S-transferase (GST) or GST-RSP5 together with Tec1 variants were prepared from cultures grown for 4 hr to exponential growth phase in appropriate SC medium. When required, 200 ng/ml rapamycin was added 20 min prior to protein extraction. Cells were harvested by centrifugation (5 min, 3000 rpm), washed in 2% galactose solution, and transferred to SC medium containing 2% galactose. After incubation for 5 hr at 30°, cultures were chilled on ice, harvested by centrifugation at 4°, washed once in B buffer [50 mm HEPES (pH 7.5), 50 mm KCl, 5 mm EDTA (pH 7.5)], resuspended in 300 µl ice-cold B buffer containing protease inhibitors (50 mm DTT, 1 mm PMSF, 0.5 mm TPCK, 0.5 mm TLCK, 0.5 mm Pepstatin A) and transferred to 2 ml-reaction tubes. Cells were broken by vortexing with glass beads at 4° for 10 min, followed by addition of 300 µl B buffer plus protease inhibitors and Triton X-100 to a final concentration of 0.1%. Samples were mixed again by vortexing at 4° for 1 min, followed by centrifugation for 5 min at 13,000 rpm to remove glass beads and large cell debris. A total of 10 µl of extracts was removed to determine total protein concentration. For input control, 50 µl of the supernatant were transferred to a 1.5 ml-reaction tube and denatured by addition of SDS sample buffer and heating for 5 min at 95°. A total of 200 µl of the remaining extract was mixed with 800 µl B buffer plus protease inhibitors plus 0.1% Triton X-100 and 100 µl 50% glutathione sepharose and incubated for 2 hr at 4°. Beads were repeatedly washed in B buffer plus 0.1% Triton X-100 and collected to purify GST fusion proteins and any associated proteins. Samples were denatured by heating at 95° for 5 min in SDS sample buffer.
Preparation of total cell extracts:
Preparation of total cell extracts was performed as described (Kushnirov 2000).
Equal amounts of proteins were subjected to SDS–PAGE using 12% gels (Laemmli 1970). Proteins were separated and then transferred to nitrocellulose membranes (Schleicher und Schuell, Dassel, Germany) by electrophoresis for 1 hr at 100 V using a Mini-PROTEAN 3 electrophoresis system (Bio-Rad, Munich, Germany). Tec1 variants as well as GST fusion proteins or Cdc28 were detected using enhanced chemiluminescence (ECL) technology (Amersham, Buckinghamshire, UK) after incubation of membranes with polyclonal rabbit anti-Tec11-280 antibodies (Heise et al. 2010), goat polyclonal anti-Cdc28 antibodies, or polyclonal rabbit anti-GST antibodies (Santa Cruz Biotechnology). Phosphorylation of Fus3 and Kss1 was detected using a Phospho-p44/42 MAPK (Thr202/Tyr204) antibody from Cell Signaling Technology. As secondary antibodies peroxidase-coupled goat antirabbit, goat antimouse, or donkey antigoat immunoglobulin G (Santa Cruz Biotechnology) were used. For quantification of signals, a scanner and the Quantity One software (Bio-Rad, Munich, Germany) or a Chemo Star Imager (INTAS Science Imaging Instruments, Göttingen, Germany) and the Lab Image 1D software (Kapelan Bio-Imaging, Leipzig, Germany) were used.
Cultures of yeast strain YHUM1803 carrying plasmids pME2280 (3myc-TEC1) or BHUM1177 (TEC1) as well as YEplac181 (empty vector) or pHU821 (His-UBI), respectively, were grown to exponential phase in the presence of 0.1 mm CuSO4 and treated with 10 μm MG132. An equivalent volume of DMSO was added to untreated control cultures. After 2 hr, an additional 5 μm MG132 (or DMSO) was added along with 200 ng/ml rapamycin where indicated, and incubation was continued for 70 min. Extracts were prepared under denaturing conditions, and total ubiquitin conjugates were isolated by Ni-NTA affinity chromatography as described (Davies et al. 2008, 2010). Ubiquitylated forms of Tec1 were detected in the isolated material by anti-myc Western blotting. In parallel, Tec1, total ubiquitin conjugates, and Pgk1 (as a loading control) were detected in the total extracts with monoclonal antibodies 9E10 (anti-myc), P4D1 (antiubiquitin), and 22C5 (anti-Pgk1).
Yeast cultures were grown to exponential growth phase prior to addition of rapamycin (200 ng/ml) or drug vehicle and incubation for 70 min. Total RNA was then prepared following the instruction manual of Trizol reagent (Invitrogen, Karlsruhe, Germany) and 8 µg was separated on a 1.2% agarose gel. After transfer to a positively charged nylon membrane TEC1, FLO11, CLN1, 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).
GFP fluorescence microscopy
Yeast strains expressing GFP-TEC1 were grown to mid-log phase in YNB medium before addition of rapamycin to 200 ng/ml. Cells were harvested at appropriate time points and immediately viewed in vivo on a Zeiss Axiovert microscope using (i) differential interference microscopy (DIC) and (ii) fluorescence microscopy using a GFP filter set (AHF Analysentechnik, Tübingen, Germany). Cells were photographed using a Hamamatsu Orca ER digital camera and the Improvision Openlab software (Improvision, Coventry, UK).
Reduction of Tec1 protein stability in response to rapamycin treatment
We directly tested whether Tec1 is under control of the TORC1 pathway by treatment of yeast cells with rapamycin. For this purpose, we used polyclonal anti-Tec1 antibodies that specifically detect endogenous Tec1 protein levels in yeast extracts (Heise et al. 2010). We found that inhibition of TORC1 by addition of rapamycin to yeast cultures led to a roughly fivefold decrease in endogenous Tec1 protein levels within 70 min after addition of the drug (Figure 1A). Addition of rapamycin did not significantly alter either TEC1 transcript levels (Figure 1B) or expression of a TEC1-lacZ reporter gene (Figure 1C), indicating a post-translational mechanism. To measure the effect of rapamycin on Tec1 protein stability, we performed a promoter shutoff experiment using the TEC1 gene under the control of the glucose-repressible GAL1 promoter. We found that rapamycin reduces the half-life of Tec1 approximately threefold from 67 to 24 min (Figure 1, D and E). These data suggest that Tec1 protein stability is under positive control of TORC1.
Rapamycin-induced degradation of Tec1 depends on the TORC1 network
To further explore how TORC1 regulates Tec1 stability, we screened an array of yeast mutant strains carrying deletions in genes encoding known TORC1-controlled signaling components for defects in rapamycin-induced degradation of Tec1. We first tested components of the TORC1-regulated type 2A and 2A-related protein phosphatase modules including the catalytic PP2A subunits Pph21/Pph22 and Pph3 as well as the PP2A-related catalytic subunit Sit4. We found that rapamycin-induced degradation of Tec1 was blocked in sit4∆, but not in pph21∆ pph22∆ or pph3∆ mutant strains (Figure 2A), indicating that TORC1 controls Tec1 via Sit4. This conclusion is further supported by our finding that Tec1 protein degradation is blocked in strains lacking the Sit4 regulatory subunit Sap190 and the nonessential protein Tip41, which controls Sit4 activity in response to rapamycin (Jacinto et al. 2001). We then investigated the involvement of a number of protein kinases that have been reported to act either downstream or in parallel to TORC1 and Sit4 including Npr1, Rim15, Yak1, and Sch9. We found that rapamycin-induced Tec1 degradation was affected in npr1∆ strains, but not in mutants strains lacking any of the other protein kinases tested (Figure 2B). This finding indicates that Npr1 might be involved in the TORC1-mediated control of Tec1. Finally we tested whether the Ras/PKA pathway might be involved in conferring control of Tec1 stability by rapamycin. However, rapamycin-induced Tec1 degradation was not affected by deletions in any of the genes encoding the catalytic PKA subunits Tpk1, Tpk2, or Tpk3 or in the absence of Ras2 (Figure 2C). Tec1 degradation was also not altered in strains expressing the hyperactive RAS2Val19 allele (data not shown). Together these data suggest that control of Tec1 stability by rapamycin is mediated by the TORC1-controlled protein phosphatase Sit4 and the protein kinase Npr1, but does not involve action of the Ras/PKA pathway.
Given the role of TORC1 in nutrient sensing, we also tested how nitrogen and amino acid deprivation affects Tec1. In comparison to rapamycin treatment, starvation for nitrogen and amino acids caused a significant increase in Tec1 protein levels when measured at a comparable time point (Figure S1). Moreover, starvation-induced up-regulation of Tec1 was not efficiently blocked in strains lacking SIT4. This suggests that Tec1 protein levels are nitrogen regulated by TORC1-independent mechanisms.
Control of Tec1 stability in the absence of the Fus3/Kss1 MAPK cascade
We next tested whether rapamycin-induced degradation of Tec1 involves elements of the Fus3/Kss1 MAPK cascade, because Tec1 is known to be degraded in response to mating pheromone. However, strains carrying deletions in the MAPK genes KSS1 and FUS3 or in the MAPK kinase gene STE7 are not suppressed with respect to rapamycin-dependent degradation of Tec1 (Figure 3A). This finding indicates that rapamycin-induced degradation of Tec1 does not involve action of the Fus/Kss1 MAPK cascade. We also measured how rapamycin affects the intracellular amounts of the phosphorylated forms of Fus3 and Kss1 by using phospho-specific antibodies against the two MAPKs. As expected, vegetatively growing cells contain low levels of phosphorylated Fus3 and Kss1, whereas high levels of the activated MAPK forms are present in pheromone-stimulated cells (Figure 3B). Interestingly, addition of rapamycin to vegetative cells led to a significant decrease in phosphorylated Kss1, but not Fus3. Moreover, this effect was dependent on the presence of Fus3, because the amount of phosphorylated Kss1 did not decrease in response to rapamycin in cells lacking FUS3. In pheromone-stimulated cells, rapamycin led to a reduction of the phosphorylated forms of both Kss1 and Fus3 (Figure 3B). These data indicate that the TORC1 pathway might feed into Fus3/Kss1 MAPK cascade at or above the level of the two MAPKs. However, this cross-pathway control does not appear to be crucial for pheromone-induced degradation of Tec1. Finally, we measured how concomitant addition of both rapamycin and mating pheromone to yeast affects Tec1 protein levels. This experiment revealed that the two effectors seem to act in a synergistic manner (Figure 3C), and it indicates that rapamycin and mating pheromone control Tec1 stability by distinct mechanisms.
Rapamycin-induced Tec1 degradation does not involve polyubiquitylation and appears to be proteasome independent
Previous studies have shown that pheromone treatment triggers polyubiquitylation of Tec1 and proteasome-dependent degradation (Bao et al. 2004; Chou et al. 2004). We therefore tested whether rapamycin-induced Tec1 degradation also involves ubiquitylation of the transcription factor. For this purpose, we expressed an epitope-tagged version of TEC1 (3myc-TEC1) together with an epitope-tagged variant of the ubiquitin gene (His-UBI) and performed pull-down experiments under denaturing conditions to detect ubiquitylated forms of Tec1. These experiments revealed that rapamycin treatment did not induce detectable polyubiquitylation of the transcription factor, not even under conditions of proteasome inhibition, where total ubiquitin conjugates accumulate to a high level (Figure 4A). However, we detected a monoubiquitylated form of Tec1, whose levels decreased in response to rapamycin. Analysis of the total protein extracts further revealed that treatment of cells with the proteasome inhibitor MG132 did not prevent degradation of Tec1 (Figure 4A). Because this finding indicates that rapamycin might induce a proteasome-independent degradation pathway, we also measured rapamycin-induced Tec1 degradation in temperature-sensitive proteasome mutants. Indeed, genetic inactivation of different proteasomal subunits was not sufficient to efficiently block Tec1 degradation (Figure 4B). Taken together, these data indicate that rapamycin triggers Tec1 degradation independent of polyubiquitylation and the proteasome.
Identification of the HECT ubiquitin ligase Rsp5 as a Tec1-interacting protein
We next aimed at identifying components that might participate in degradation of Tec1 in response to rapamycin. For this purpose, we analyzed an array of 22 yeast mutant strains carrying single deletions in nonessential genes encoding components of different protein modifying systems (Figure S2). However, none of the mutants tested was significantly suppressed by rapamycin-induced degradation of Tec1. This indicates that none of the tested E3 ubiquitin ligases, F-box proteins, or SUMO/Smt3 ligases, which are lacking in these mutants, are involved in the process. To more directly identify proteins involved in Tec1 protein degradation, we performed a yeast two-hybrid screen. For this purpose, a lexA-TEC1 construct lacking the Tec1 DNA-binding domain (TEAD) was used that alone led to only a weak induction of a lexA-driven lacZ reporter gene (Figure 5, A and B). Roughly 100,000 B42-fused yeast cDNAs (Fashena et al. 2000) were screened for a two-hybrid interaction with lexA-TEC1 and led to the identification of the HECT ubiquitin ligase Rsp5 as a potential interaction partner for Tec1 (Figure 5B). Sequence analysis of the isolated RSP5 cDNA revealed that it encoded residues 372–809 of Rsp5, encompassing the WW domain that confers interaction of this ubiquitin ligase with proteins that contain PxY motifs (Hoppe et al. 2000; Shcherbik et al. 2004). Analysis of the Tec1 protein sequence revealed that it harbors three PxY sequence motifs in the C-terminal part (Figure 5A). To further investigate the Tec1–Rsp5 physical interaction, we performed GST coaffinity purification experiments using Tec1 and Rsp5 proteins. We found that Tec1 could be copurified with GST–Rsp5 and that the interaction remained stable upon rapamycin treatment (Figure 5, C and D and Figure S3). Moreover, mutation of the three PxY motifs of Tec1 to AxY sequences or deletion of the C-terminal part of Tec1 efficiently blocked interaction of the transcription factor with Rsp5 (Figure 5C). Together, these data demonstrate that the ubiquitin ligase Rsp5 is a specific Tec1-interacting protein.
Dependence of rapamycin-induced degradation of Tec1 on Rsp5 and C-terminal PxY motifs
To address the physiological significance of the Tec1–Rsp5 interaction, we measured rapamycin-induced degradation of Tec1 in strains lacking RSP5. Normally, yeast strains carrying a full deletion of RSP5 are not viable, due to insufficient expression of the OLE1 gene that is required for production of essential amounts of oleic acid (Hoppe et al. 2000). To circumvent this problem, we constructed a strain that expresses the OLE1 gene from the Rsp5-independent URA3 promoter. In this genetic background, a full deletion of the RSP5 gene could be introduced, resulting in a viable rsp5∆ mutant strain. Using this strain, we found that rapamycin-induced degradation of Tec1 was efficiently blocked in the absence of RSP5 (Figure 6A). In contrast, deletion of RSP5 did not block the increase of Tec1 protein levels observed in response to nitrogen and amino acid starvation (Figure S1). Thus, Rsp5 might be involved in destabilization of the transcription factor in response to rapamycin. This conclusion is supported by our finding that rapamycin-induced degradation of Tec1 is blocked by deletion of the C-terminal domain of Tec1, which is required for association of Tec1 with Rsp5 (Figure 6B). Moreover, mutation of the three PxY motifs in Tec1 to AxY sequences also generated a rapamycin-resistant variant of Tec1 (Figure 6B). Importantly, the Tec1AxY variant, which does not interact with Rsp5, is fully functional in vivo when assayed in an adhesion test (Figures 6C and Figure S4). We also measured whether destabilization of Tec1 depends on Bul1, Bul2, Rog3, or Rup1, which represent Rsp5-interacting proteins that have been suggested to confer substrate specificity (Yashiroda et al. 1998; Andoh et al. 2002; Kee et al. 2005). However, rapamycin-induced degradation of Tec1 was not affected in strains lacking these components (Figure 6A).
Finally, we tested how rapamycin and Rsp5 affect the cellular localization of Tec1 in living cells by using a functional GFP-TEC1 fusion gene. We found that in the control strain, a functional GFP-Tec1 protein is rapidly degraded upon addition of rapamycin when analyzed by immunoblot analysis (Figure 6D) or by fluorescence microscopy (Figure 6E). In contrast, rapamycin-induced degradation of GFP-Tec1 was blocked in the rsp5∆ deletion strain (Figure 6E), corroborating the data obtained with strains expressing endogenous Tec1. However, addition of rapamycin did not significantly alter the nucleo-cytoplasmic distribution of GFP-Tec1, neither in the presence nor in the absence of Rsp5 (Figure 6E), indicating that subcellular localization of Tec1 is not under control of rapamycin.
Rapamycin and mating pheromone control Tec1 stability by distinct mechanisms
We then focused on the differences in the mechanisms that confer destabilization of Tec1 in response to either rapamycin or mating pheromone. For this purpose, we measured how rapamycin affects the stability of known pheromone-resistant variants of Tec1 that carry mutations within the Cdc4-phosphodegron motif (Chou et al. 2004). We found that the pheromone-resistant variants Tec1T273M and Tec1P274S were rapidly degraded in response to rapamycin (Figure 6A). Similarly Tec1K54A, a variant that blocks sumoylation of Tec1 at residue Lys54 and confers higher stability of the protein (Wang and Dohlman 2006), was not resistant to rapamycin treatment (Figure 7A). We next asked, how mating pheromone affects stability of the Tec1AxY variant that we here uncovered to be resistant to rapamycin. As shown in Figure 7B, Tec1AxY is rapidly degraded in response to pheromone treatment. In addition, pheromone-induced Tec1 degradation was not efficiently blocked in a rsp5∆ mutant strain (Figure 7C). These experiments demonstrate that pheromone- and rapamycin-induced destabilization of Tec1 depend on distinct residues of the transcription factors and indicate that rapamycin and pheromone are two cues that control Tec1 stability by distinct mechanisms.
Differential regulation of Tec1 targets CLN1 and FLO11 by rapamycin
We have previously shown that degradation of Tec1 in response to mating pheromone causes a significant loss in the transcription of the two Tec1-target genes CLN1 and FLO11 (Brückner et al. 2004). Therefore, we measured how rapamycin affects expression of CLN1 and FLO11 in the absence and presence of TEC1. In the case of CLN1, we found that rapamycin caused a significant decrease in CLN1 transcript levels, but only in the TEC1 strain and not in the tec1∆ mutant, in which expression of this cyclin gene was very low (Figure 8A). This finding indicates that rapamycin-induced repression of CLN1 depends on TEC1 and suggests that loss of CLN1 transcription might be caused by rapamycin-induced degradation of Tec1. This conclusion is supported by the fact that rapamycin does not cause a decrease in CLN1 transcript levels in a strain expressing the TEC1AxY variant (Figure 8B). A different expression pattern was observed in the case of the Tec1-target gene FLO11, which was slightly up-regulated by rapamycin in the presence of Tec1 (Figure 8A). In the absence of Tec1, FLO11 transcript levels were very low, even when rapamycin was added, indicating that Tec1 is essential for efficient FLO11 expression, independent of the TORC1 activity. Finally, rapamycin also caused a slight increase in FLO11 transcript levels in strains expressing TEC1AxY or TEC1T273M (Figure 8B). In summary, these data show that the Tec1-target genes CLN1 and FLO11 are differentially regulated by rapamycin and that regulation of these genes by TORC1 involves Tec1.
Tec1 is a regulator of TORC1 controlled chronological lifespan
Our finding that repression of CLN1 gene expression by rapamycin involves Tec1 prompted us to test whether Tec1 might be involved in rapamycin-induced cell cycle arrest. However, we could not detect a significant difference in rapamycin sensitivity between yeast strains lacking or overexpressing TEC1 (data not shown). We next tested whether Tec1 might affect yeast CLS, because previous studies have shown that this process is under control of TORC1 (Pedruzzi et al. 2003; Powers et al. 2006). In accordance with previous studies, low doses of rapamycin led to an extension of CLS, as shown for the TEC1 expressing control strain (Figure 9). However, rapamycin-induced CLS extension was not observed before prolonged stay of yeast cells in stationary phase. Interestingly, CLS was also significantly reduced in yeast strains that lack Tec1. As found for rapamycin treatment, the influence of Tec1 became obvious only after a prolonged stay of cells in stationary phase. However, absence of Tec1 did not fully suppress rapamycin-induced CLS extension. Finally, CLS was also reduced in strains expressing the TEC1AxY allele, but not as pronounced as in strains completely lacking the transcription factor. Interestingly, rapamycin-induced CLS extension was almost completely blocked in the presence of TEC1AxY. In summary, these data indicate that Tec1 might be a positive regulator of CLS.
In this study, we have presented evidence that in S. cerevisiae (i) the TORC1 and Fus3/Kss1 MAPK pathways are interconnected by the TEA domain transcription factor Tec1, (ii) both pathways control this transcription factor by affecting its stability through independent mechanisms, (iii) the connection between TORC1 and MAPK signaling is important to coordinate G1 cyclin and vegetative adhesin gene expression, and (iv) Tec1 affects chronological lifespan. These findings provide several novel insights into the mechanisms that enable integration of TOR and MAPK kinase signaling in eukaryotic cells and ensure coordination of cell division and development. Importantly, our study pinpoints the connection between these two pathways to a transcriptional regulator of the TEA domain family, whose members are known to control development in many eukaryotes (Andrianopoulos and Timberlake 1991; Anbanandam et al. 2006). Previous studies have described other connections between TOR and MAPK pathways. In S. cerevisiae, the TORC1 pathway is connected to the PKC-Mpk1 MAPK pathway (Angeles De La Torre-Ruiz et al. 2002; Kuranda et al. 2006), and TOR and stress MAPK pathways are linked in Schizosaccharomyces pombe (Petersen and Nurse 2007). In both cases, however, the connecting mechanisms have not been uncovered. With Tec1, we provide a precise molecular link between the TORC1 and Fus3/Kss1 MAPK pathways in S. cerevisiae. Our study further opens the possibility that in higher eukaryotes, TEAD transcription factors are not only regulated by PKA (Gupta et al. 2000), PKC (Jiang et al. 2001), and MAPK signaling pathways (Ambrosino et al. 2006), but also by TORC1. This would hook up TEAD-controlled eukaryotic development to metabolic stimuli such as amino acids, which regulate mTORC1 in mammals (Polak and Hall 2009).
Our data further provide insights into the physiological relevance of the TORC1-Tec1-MAPK connection in S. cerevisiae. Previous studies have shown that Tec1 is a positive regulator of the G1 cyclin CLN1 and the vegetative adhesin FLO11 (Madhani et al. 1999), that mating pheromone inhibits CLN1 and FLO11 expression (Brückner et al. 2004), and that rapamycin inhibits CLN1 expression and FLO11-dependent yeast filamentous growth (Barbet et al. 1996; Cutler et al. 2001; Zinzalla et al. 2007). Moreover, Tec1 protein stability is reduced by the MAPK Fus3 in response to mating pheromone (Bao et al. 2004; Brückner et al. 2004; Chou et al. 2004). Thus, our finding that rapamycin negatively regulates Tec1 protein stability implies that active TORC1, in contrast to the pheromone-activated MAPK, positively controls cell division and vegetative adhesion by targeting Tec1 (Figure 10). Our data also show that TORC1 seems to regulate the Tec1-target genes CLN1 and FLO11 by different mechanisms. In the case of CLN1, expression is activated by TORC1 and correlates with Tec1 protein levels. This indicates that TORC1 regulates the promoter by simply varying Tec1 amounts. In contrast to CLN1, the much more complex FLO11 promoter is repressed by TORC1. In addition, FLO11 expression does not correlate to varying Tec1 levels, but is strongly reduced in the complete absence of the transcription factor. More than one model might explain these observations. (i) The Tec1 protein amounts required for efficient FLO11 activation might be significantly lower than in the case of CLN1, explaining why rapamycin-induced reduction of Tec1 is not sufficient to block FLO11 expression. (ii) TORC1 might control not only Tec1 stability, but also its transcriptional activity. In this case, rapamycin treatment would lead to lower Tec1 protein amounts, but also induce its activity at the FLO11, but not at the CLN1 promoter. (iii) TORC1 might also regulate the FLO11 promoter by activating a repressor, whose function requires only minimal amounts of Tec1. Via this repressor, TORC1 would be able to control FLO11 at low and high Tec1 amounts. In the complete absence of Tec1, however, efficient FLO11 expression and its regulation by TORC1 would be lost, due to low repressor concentrations.
We have also uncovered that Tec1 affects yeast chronological lifespan, a process that is regulated by rapamycin and TORC1 (Powers et al. 2006). Here we found that absence of Tec1 causes a significantly shortenend CLS, but does not block CLS extension by rapamycin. This indicates that TORC1 is able to control CLS by Tec1-independent mechanisms, which might involve other factors that positively control yeast lifespan (Wei et al. 2008). It will be interesting to see how Tec1 affects CLS and whether Tec1 is required before and/or after entry into stationary phase. Remarkably, both the complete absence of Tec1 and the presence of a rapamycin-resistant variant of the transcription factor result in a similar CLS phenotype. This indicates that Tec1 might act at several stages and that its activity must be accurately regulated to achieve maximal CLS.
Our study shows that rapamycin-induced degradation of Tec1 is independent of the known mechanisms that confer pheromone-induced destabilization, but depends on (i) specific elements of the TORC1 signaling pathway, namely Sit4, Tip41, Sap190, and Npr1; (ii) the HECT-ubiquitin ligase Rsp5; and (iii) three conserved PxY motives in the C-terminal part of Tec1 that are required for physical interaction with Rsp5. An important question is how these components could work together to confer degradation of Tec1 in response to rapamycin. Previous studies have shown that Tip41 controls activity of Sit4 via interaction with Tap42 (Jacinto et al. 2001), that Sit4 can associate with Sap190 (Luke et al. 1996), and that regulation of Npr1 by TORC1 depends on Sit4 (Schmidt et al. 1998). Thus, our data are compatible with a model, in which TORC1 regulates Tec1 protein stability via control of Sit4/Sap190 and Npr1 (Figure 10). However, Sit4 and Npr1 might not act in a strict linear way, because we found that Tec1 degradation is only partially blocked in npr1∆ mutants, but is fully inhibited in the absence of SIT4. In addition, regulation of Tec1 by TORC1 and nutrients might be more complex, because we found that nitrogen deprivation leads to increased levels of the transcription factor, even in the absence of Sit4 and Rsp5. Here, future detailed studies are required to determine how different nutrient signals and responsive pathways control Tec1 stability.
Although we cannot exclude an indirect effect of Rsp5 on Tec1 stability, our finding that the transcription factor physically interacts with the ligase via three PxY motives suggests that Rsp5 acts by directly ubiquitylating Tec1. Similarly, single PxY motifs located in the C terminus of the homologous S. cerevisiae transcription factors Spt23 and Mga2 have been shown to be sufficient to mediate physical and functional interaction with Rsp5 (Shcherbik et al. 2004). Surprisingly, however, we found that the interaction between Rsp5 and Tec1 appears to be constitutive rather than regulated by rapamycin. In addition, the mechanism of Tec1 degradation remains to be established. Rsp5 is known for monoubiquitylation as well as polyubiquitylation via K63-linked chains and contributes to the modification of membrane-associated as well as soluble nuclear proteins (Rotin and Kumar 2009). Downstream processing of its targets may involve the 26S proteasome, but proteasome-independent degradation pathways are also well known. Hence, given the variety of mechanisms by which Rsp5 controls protein stability, regulation of Tec1 by Rsp5 might be much more complicated than anticipated and will have to be addressed in detail in the future.
In summary, our study has identified a member of the highly conserved TEA domain family of eukaryotic developmental regulators as a common target for MAPK and TOR signaling pathways, which control this type of transcription factor at the level of protein stability by independent mechanisms. Complex regulation of transcription factor stabilization by post-translational modifications is an important theme in eukaryotic cell growth and development with implications for disease control (Desterro et al. 2000; Hay 2006). A prominent example is the tumor suppressor p53, a transcription factor whose stability is regulated by multiple proteins including several ubiquitin ligases (Lavin and Gueven 2006). Modulation of p53 degradation is now becoming a promising therapeutic approach (Dey et al. 2008). Thus, future studies directed at the stability control of the TEA domain regulator Tec1 in S. cerevisiae may contribute to a more precise understanding of the molecular mechanisms underlying the complex regulation of transcription factors during cellular development.
We thank Bruno Andre, Kim Arndt, Gerhard Braus, Roger Brent, Claudio DeVirgilio, Erica Golemis, Satoshi Harashima, Stefan Irniger, Yona Kassir, Hiten Madhani, and Raffael Schaffrath for generous gifts of plasmids and strains. We are grateful to Diana Kruhl for technical assistance. 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 for Environmental, Cellular and Molecular Microbiology, and by the Marburg Center for Synthetic Microbiology.
Supporting information is available online at http://www.genetics.org/content/suppl/2011/08/12/genetics.111.133629.DC1.
- Received March 8, 2011.
- Accepted August 4, 2011.
- Copyright © 2011 by the Genetics Society of America