S. pombe TORC1 activates the ubiquitin-proteasomal degradation of the meiotic regulator Mei2 in cooperation with Pat1 kinase Free

Target of rapamycin (TOR) is a highly conserved serine/threonine kinase, which is thought to regulate various growth-related functions in response to environmental changes (Wullschleger et al., 2006). TOR forms two types of multi-protein complexes, TORC1 and TORC2. In mammalian cells, a single catalytic subunit, mTOR, is present in both TORC1 and TORC2 (mTORC1 and mTORC2). In yeast species, TORC1 and TORC2 have a similar subunit composition to mTORCs, but they carry different TOR proteins as the catalytic subunit (Wullschleger et al., 2006).

Despite the physiological importance of TORCs, only a limited number of TORC substrates have been identified to date. S6K1, an AGC family protein kinase, and 4E-BP1, an eIF4E-binding protein, are the best-characterized substrates of mTORC1. mTORC1 regulates translation through these molecules. Although not as well characterized as S6K1 or 4E-BP1, additional proteins, including Maf1, a repressor of RNA polymerase III, and Atg13, an autophagy-related protein that acts as a conserved regulator of Atg1 kinase, are emerging as direct targets of TORC1 in mammals and yeast. TORC1 appears to promote cell growth and cell cycle progression through the phosphorylation of these substrates (Wullschleger et al., 2006; Laplante and Sabatini, 2012). An even smaller number of targets are known for TORC2, and these include mammalian Akt and yeast Ypk2.

In fission yeast, Schizosaccharomyces pombe, TORC1, containing Tor2 as its catalytic subunit, and TORC2, containing a Tor1 catalytic subunit, fulfill opposing functions with respect to the promotion of G1 arrest and sexual development (Kawai et al., 2001; Weisman and Choder, 2001; Alvarez and Moreno, 2006; Uritani et al., 2006; Matsuo et al., 2007; Weisman et al., 2007) (reviewed in Otsubo and Yamamato, 2008). TORC1 is active during vegetative growth and regulates sexual development negatively, apparently by transmitting signals from a sufficient nitrogen supply (Matsuo et al., 2007). TORC1 is known to phosphorylate Psk1, a fission yeast homolog of S6K1 (Nakashima et al., 2012), but the effectors of TORC1 that function in the suppression of sexual development remain largely unidentified.

An RNA-binding protein, Mei2, governs the entry into meiosis in fission yeast (Watanabe and Yamamoto, 1994; Watanabe et al., 1997; Yamashita et al., 1998). Mei2 has the ability to switch off the mitotic cell cycle and promote meiosis, enabling cells to stably express meiosis-specific mRNAs (Harigaya et al., 2006). In addition to eliciting meiosis, Mei2 participates in earlier steps of sexual development, namely G1 arrest and mating, with its loss causing a partial deficiency in these stages (Watanabe and Yamamoto, 1994) (supplementary material Fig. S1A). We have previously reported a physical interaction between Mei2 and Mip1 (Shinozaki-Yabana et al., 2000), which later turned out to be the fission yeast ortholog of mammalian Raptor, an essential component of mTORC1 (Hara et al., 2002; Kim et al., 2002; Alvarez and Moreno, 2006; Hayashi et al., 2007; Matsuo et al., 2007). A physical interaction has been also found between Mei2 and Tor2 (Alvarez and Moreno, 2006). These observations suggest an intimate relationship between Mei2 and TORC1. In this study, we demonstrate that Mei2 is a novel target of TORC1.

We observed that cells of the temperature-sensitive tor2 mutant (tor2-ts6), which are de-repressed for sexual development at the restrictive temperature (Matsuo et al., 2007), tended to accumulate Mei2 protein at a high temperature (Fig. 1A). To examine its stability, we tagged Mei2 with three copies of the hemagglutinin epitope (Mei2–3HA) and expressed it under the control of the thiamine-repressible promoter in wild-type and tor2-ts6 cells. In wild-type cells, Mei2 disappeared rapidly following the addition of thiamine and cycloheximide to repress transcription and translation (Fig. 1B), as was reported previously (Kitamura et al., 2001). By contrast, Mei2 was retained more stably in tor2-ts6 cells (Fig. 1B). Mei2 protein was also accumulated in the temperature-sensitive mip1 mutant (Fig. 1C). These results suggested that active Tor2 (TORC1) could destabilize Mei2.

We next analyzed genetic interactions between tor2 and mei2. Although the tor2-ts6 mutant underwent extensive mating under nutrient-rich conditions at the restrictive temperature (Matsuo et al., 2007) (Fig. 2A), this mating was much reduced when combined with deletion of mei2 (Fig. 2A). By contrast, the sterility of an activated tor2 mutant, tor2-s65, which was isolated in our screen for sterile tor2 mutants (Urano et al., 2007), was suppressed by the overexpression of wild-type mei2 (Fig. 2B). This suppression was also observed with another activated tor2 mutant, tor2-s69 (supplementary material Fig. S1B). Furthermore, overexpression of mei2 induced mating more efficiently in the tor2-ts6 mutant than in the wild-type strain (Fig. 2C). These results suggest that Tor2 and Mei2 might interact and modulate the function of Mei2 to promote mating.

To investigate whether Tor2 could directly phosphorylate Mei2, we performed an in vitro Tor2 phosphorylation assay. FLAG-tagged Tor2 was immunoprecipitated from fission yeast cell extracts using an anti-FLAG antibody. We used the same immunoprecipitation conditions as described in our previous study, in which other components of TORC1 were co-precipitated (Matsuo et al., 2007). The kinase preparation was then mixed with a GST–Mei2 fusion protein, which harbors Mei2 residues 30–750 and was shown previously to be functional (Watanabe et al., 1997), and incubated with radioactive ATP. Tor2 phosphorylated GST–Mei2 efficiently in this system, whereas a kinase-dead form of Tor2 (Tor2KD), used as a negative control, displayed only marginal phosphorylation toward GST–Mei2 (Fig. 3A).

To identify the amino acid residues of Mei2 that are phosphorylated by Tor2, we carried out mass spectrometric analysis of recombinant GST–Mei2 that had been subjected to in vitro phosphorylation. We identified nine phosphorylated residues in this analysis – S34, S35, S39, S43, T44, T231, T339, S343 and S694 (Fig. 3B; supplementary material Fig. S2A). Given these results, we constructed a plasmid that expressed GST–Mei2-9A, in which all of the nine phosphorylation sites were replaced with alanine. GST–Mei2-9A, which carries Mei2 residues 1–750, was barely phosphorylated by Tor2 in vitro, indicating that Tor2 principally phosphorylates the nine identified Ser/Thr residues (Fig. 3C). It has been shown previously that Pat1 kinase phosphorylates Mei2 on S438 and T527 during the mitotic cell cycle (Watanabe et al., 1997). The above results indicated that these residues are not preferentially phosphorylated by Tor2. Consistently, Tor2 phosphorylated GST–Mei2-SATA, in which the two sites phosphorylated by Pat1 were replaced with alanine, as efficiently as it phosphorylated wild-type Mei2 (Fig. 3D; supplementary material Fig. S2B).

As expected, Mei2 protein recovered from growing cells was phosphorylated, showing a mobility shift after phosphatase treatment (Fig. 4A). We next examined whether TORC1 participated in this in vivo phosphorylation of Mei2. It was difficult, however, to detect an obvious mobility shift between the Mei2 protein derived from wild-type cells and that derived from tor2-ts6 cells, when they were extracted under the native conditions as in Fig. 1A, partly owing to their instability. Therefore, we extracted Mei2 protein under denaturing conditions (see Materials and Methods). We also tried to load the protein samples in equal amounts for ease of comparison. Consequently, the Mei2 protein recovered from tor2-ts6 cells revealed faster migration than that recovered from wild-type cells, even at low temperature (Fig. 4B). Mei2-9A recovered from wild-type cells also migrated faster, indicating that the nine sites are likely to be phosphorylated in vivo. Unexpectedly, however, the mobility of Mei2 protein from tor2-ts6 cells slowed after a longer incubation at 30°C (3 h). This might be due to phosphorylation of Mei2 by some kinase(s) activated under starved conditions, as a reduction in Tor2 activity mimics nitrogen starvation (Alvarez and Moreno, 2006; Uritani et al., 2006; Matsuo et al., 2007; Weisman et al., 2007). In support of this idea, phosphatase treatment of Mei2-9A-SATA recovered from growing cells indicated that it was still phosphorylated, suggesting the presence of a kinase(s) other than Tor2 and Pat1 that could phosphorylate Mei2 (supplementary material Fig. S2C). We also examined Mei2N, carrying only residues 1–428, which contains eight of the nine Tor2 phosphorylation sites but lacks the Pat1 sites. A similar mobility shift was observed with Mei2N prepared from tor2-ts6 cells, although the shift was smaller compared with that of full-length Mei2 (supplementary material Fig. S2D). These results indicate that Mei2 is phosphorylated in vivo at least in a TORC1-dependent manner, if not directly by TORC1.

We constructed a mei2 mutant strain in which the nine Tor2 phosphorylation sites were replaced with alanine. However, substitution of alanine for S694 was found to result in a sporulation deficiency (supplementary material Fig. S3A). Replacement of S694 with aspartic acid or glutamic acid also compromised the function of Mei2 (data not shown), suggesting that this residue should be serine to allow Mei2 to function properly. Hence, we decided to characterize the phenotype of a mei2-8A mutant, in which all the phosphorylation sites except S694 were replaced with alanine. The mei2-8A strain initiated sexual development more frequently than its wild-type counterpart under nutrient-rich conditions, in a similar fashion to the tor2-ts mutants at high temperature (Fig. 4C) (Matsuo et al., 2007). A series of alanine substitution mutants altered at one or more of the eight Tor2 phosphorylation sites (S694 was excluded from this analysis) revealed an increase in the mating and sporulation frequency to nearly the same degree (supplementary material Fig. S3B), suggesting that all of the Tor2 target sites must be phosphorylated to repress sexual development effectively. Furthermore, the overexpression of the mei2-8A allele could suppress the mating deficiency in the tor2 active mutant more efficiently than the overexpression of wild-type mei2 (supplementary material Fig. S1B). These results indicate that Tor2, and hence TORC1, contributes to the suppression of sexual development during vegetative growth through the phosphorylation of Mei2.

To examine the effect of the Tor2-dependent phosphorylation of Mei2, we analyzed the stability of Mei2-9A in wild-type cells, as was performed in Fig. 1B. The mutant Mei2 protein was more stable than the wild-type (Fig. 5A). Mei2-8A was also more stable (supplementary material Fig. S3C,D). It has been shown that Mei2 protein phosphorylated by Pat1 undergoes degradation through the ubiquitin-proteasome pathway, involving Ubc2 and Ubr1 (Kitamura et al., 2001). We confirmed that wild-type Mei2 protein in growing cells, presumably phosphorylated by both Pat1 and Tor2 kinases, became more stable in the mts2 mutant, which is defective in the 19S proteasome (supplementary material Fig. S4A). This observation implies that the phosphorylation of Mei2 by Tor2 accelerates its degradation through the previously identified ubiquitin-proteasome pathway (Kitamura et al., 2001). Indeed, polyubiquitylation of Mei2 was decreased greatly by the mei2-9A mutation (Fig. 5B). Mei2-8A was also less polyubiquitylated than wild-type Mei2 (supplementary material Fig. S4B). Furthermore, mts2 mutant cells initiated sexual development readily under nutrient-rich conditions (supplementary material Fig. S4C). These observations suggest that the proteasomal degradation of Mei2 is important for the repression of sexual development. Although Mei2-SATA, in which the Pat1 phosphorylation sites are replaced with alanine, was shown previously to be more stable than wild-type Mei2 in tor2⁺ cells (Kitamura et al., 2001) (Fig. 6A), its stability was comparable to that of wild-type Mei2 in the tor2-ts6 mutant cells at high temperature (Fig. 6B). This might indicate that the phosphorylation of Mei2 by both TORC1 and Pat1 kinase is necessary to induce proteasomal degradation of the protein.

This study has shown that TORC1 phosphorylates Mei2 in vitro and is also likely to do so in vivo. Alanine substitution at the identified phosphorylation sites stabilized Mei2 protein and enhanced sexual differentiation under nutrient-rich conditions. The TORC1-dependent phosphorylation of Mei2, in conjunction with the phosphorylation mediated by Pat1 kinase, accelerated its degradation through the ubiquitin-proteasome pathway. These observations led us to propose that TORC1 participates in the repression of sexual differentiation during vegetative growth through phosphorylation of Mei2. However, because the enhanced mating of tor2-ts cells on nutrient-rich medium was not completely blocked by the mei2 deletion (Fig. 2A), it is likely that TORC1 has some other target(s), phosphorylation of which would also contribute to the repression of untimely mating during vegetative growth.

We found that nine residues of Mei2 were phosphorylated by TORC1. Alanine substitution at one target site or at a combination of sites resulted in de-repression of sexual development, although alanine substitution at the eight sites excluding S694, which was indispensable for Mei2 function, was most effective. These observations suggest that TORC1 must maintain a high activity to phosphorylate all of the target sites on Mei2 during vegetative growth, and presumably each Mei2 target site is a good substrate for TORC1. It will be intriguing to investigate whether the latter is indeed the case, because it has been suggested that the quality of a phosphorylation site as a TORC1 substrate can be a key determinant of how downstream effectors respond in the signaling pathway (Kang et al., 2013).

Mei2 has been characterized in considerable detail as the master regulator of meiosis in fission yeast (Harigaya et al., 2006; Harigaya and Yamamoto, 2007; Yamamoto, 2010). It remains puzzling, however, how Mei2 participates in the earlier steps of sexual differentiation, i.e. G1 arrest and mating, even though it is not absolutely necessary for these steps. One possible explanation has emerged from our recent study; activated Mei2 was found to enhance the expression of ste11 through a positive-feedback loop involving the stress-responsive MAP kinase pathway (Sukegawa et al., 2011). Ste11 is a major transcription factor controlling the switch from cellular proliferation to sexual differentiation (Sugimoto et al., 1991; Otsubo and Yamamoto, 2012). Thus, lack of the full expression of ste11 in the absence of Mei2 function might result in lower mating efficiency.

Nutrient starvation induces sexual development in fission yeast cells. Both TORC1 and Pat1 kinase phosphorylate Mei2 and cooperate to destabilize it under conditions of rich nutrition. However, the phenotypes of the mutants defective in the respective kinases are largely different. Cells defective in tor2, encoding the catalytic subunit of TORC1, initiate the whole program of sexual development, including mating, whereas cells defective in pat1 skip mating and start meiosis directly (Iino and Yamamoto, 1985; Nurse, 1985; Matsuo et al., 2007). Considering all of the above observations, we propose the following stepwise activation of Mei2 as a plausible model (Fig. 6C). During vegetative growth, both TORC1 and Pat1 phosphorylate Mei2 to destabilize and inactivate it. Upon nitrogen starvation, TORC1 is downregulated and Mei2 becomes stable. Partially dephosphorylated Mei2 that accumulates at this stage might promote mating. Enhanced production of Mei2 under nitrogen starvation might also aid the accumulation of dephosphorylated Mei2 protein (Watanabe et al., 1988; Sugimoto et al., 1991). Following mating, Pat1 kinase is inactivated in zygotes (McLeod and Beach, 1988; Li and McLeod, 1996), and fully dephosphorylated Mei2 acquires the capacity to elicit meiosis. Consistent with this model, ectopic expression of Mei2-8A-SATA resulted in the induction of meiosis without mating (data not shown).

Although Mei2 appears to be an unusual substrate of TORC1, as it regulates specific switching rather than general translation or cell metabolism, there is a report that Arabidopsis AtRaptor1B, a Raptor homolog, binds to a Mei2-like protein, AML1 (Anderson and Hanson, 2005). Hence, it will be intriguing to determine whether AML1 is a substrate of Arabidopsis TORC1. Furthermore, it will also be of interest to determine whether this phosphorylation event has any role in linking environmental conditions to cellular development or in promoting ubiquitin-proteasomal degradation.

The Schizosaccharomyces pombe strains used in this study are listed in supplementary material Table S1. Yeast media YE, SD, MM and SSA were used for routine culture of S. pombe strains (Gutz et al., 1974; Moreno et al., 1991). Vectors carrying the thiamine-repressible nmt1 promoter or its derivatives have been described previously (Maundrell, 1990; Basi et al., 1993). General genetic procedures for S. pombe have also been described previously (Gutz et al., 1974). Transformants of S. pombe were generated using a lithium acetate method (Okazaki et al., 1990). Alanine substitution alleles of mei2 were created using the PrimeSTAR Mutagenesis Basal Kit (TaKaRa). Cells of JZ127 (h⁹⁰ mei2::ura4⁺) were transformed with linear DNA fragments carrying a mutated mei2 allele, and transformants were selected on media containing fluoroorotic acid.

Strains carrying a plasmid expressing Mei2–3HA or its derivative from a weakened thiamine-repressible nmt1 promoter (pREP41 or 81) were grown in minimal medium MM for 14 h in the absence of thiamine. Thiamine (5 µM) and cycloheximide (100 µg/ml) were then added to stop transcription and translation, and cell extracts were prepared at specific intervals and subjected to immunoblotting analysis. When analyzing Mei2-SATA in vivo, the protein was given an additional FA mutation (Fig. 6B) to suppress its ability to induce ectopic meiosis, as described previously (Watanabe et al., 1997).

Harvested cells were disrupted with glass beads in buffer B [50 mM Tris-HCl pH 7.6, 150 mM KCl, 5 mM EDTA, 1 mM dithiothreitol (DTT), 10% glycerol, 0.2% NP-40, 20 mM β-glycerophosphate, 0.1 mM Na₃VO₄, 15 mM p-nitrophenyl phosphate, 1 mM phenylmethylsulfonyl fluoride and protease inhibitor cocktail (Complete Mini EDTA-free; Roche)]. Mouse monoclonal antibodies against tubulin (TAT-1; a gift from Keith Gull, University of Oxford, Oxford, UK), HA (12CA5, Sigma), GFP (Roche) and FLAG (M2 Sigma) were used to detect target proteins. Horseradish-peroxidase-conjugated sheep anti-mouse-IgG (GE Healthcare) and a chemiluminescence system (ECL, GE Healthcare) were used to visualize bound antibodies. For phosphatase treatment, immunoprecipitated Mei2–GFP was incubated with 400 U of λ-phosphatase (BioLabs) at 30°C for 20 min, either with or without an inhibitor mixture (10 mM EGTA, 10 mM Na₃VO₄, 20 mM β-glycerophosphate, 15 mM p-nitrophenyl phosphate). For the detection of mobility shift of Mei2 in vivo, harvested cells were disrupted with glass beads in 20% trichloroacetic acid. The amounts of Mei2 protein were determined by measuring the HA signal intensity.

The proteasome-defective mts2 strain (JT922) was transformed with either pREP81-mei2-3HA, pREP81-mei2-9A-3HA or pREP81-mei2-8A-3HA, together with a plasmid that expresses 6His–ubiquitin from the nmt1 promoter (pREP2-6His-Ubi). The cells were then cultured at 25°C in the absence of thiamine and were shifted to 36°C for 1 h. Cell extracts were prepared, and the ubiquitylated fraction was purified by using Ni²⁺–NTA beads as described previously (Shiozaki and Russell, 1997). Precipitated proteins were separated by electrophoresis and immunoblotted with an anti-HA antibody to detect Mei2–3HA ubiquitylation.

The JY450 strain was transformed with either pREP41-His6Flag2-tor2 or pREP41-His6Flag2-tor2KD (kinase dead allele; D2140A). Tor2 or Tor2KD was immunoprecipitated with an anti-FLAG antibody as described previously for the analysis of TORC1 (Matsuo et al., 2007). Precipitates were washed three times with buffer B with no protease inhibitors, and then twice with buffer Ktor (20 mM HEPES-KOH pH 7.5, 1 mM DTT, 20 mM β-glycerophosphate, 0.1 mM Na₃VO₄, 15 mM p-nitrophenyl phosphate, 10 mM MnCl₂). Bacterially purified GST-fusion proteins carrying either wild-type Mei2 residues 30–750, which retain normal Mei2 function (Watanabe et al., 1997), full-length Mei2 (1–750) or Mei2 (1–750) with alanine substitutions at TORC1 or Pat1 kinase phosphorylation sites (Mei2-SATA) were prepared as substrates. An in vitro kinase assay was carried out as follows: 2 µg (Fig. 3A,D) or 1 µg (Fig. 3C; supplementary material Fig. S2B) of GST–Mei2 was mixed with immunoprecipitated Tor2 or Tor2KD in buffer Ktor supplemented with 2 µg of bovine serum albumin (BSA). The reaction was initiated by adding 5 µCi [γ-³²P]-ATP together with 25 µM cold ATP, and terminated by adding SDS-PAGE buffer after a 35-min incubation at 32°C.

To visualize Mei2–GFP, strains to be examined were observed under a fluorescence microscope (Axioplan 2; Carl Zeiss), and images were recorded using a chilled CCD camera (Quantix; Photometrics) and the MetaMorph software (Universal Imaging).

GST–Mei2 phosphorylated by Tor2 in vitro was prepared as described above and analyzed by mass spectrometry undertaken by Filgen. The data were also analyzed by Filgen using ProteinPilot (AB SCIEX).

We thank Tomomi Kato and Kumiko Aikawa (Kazusa DNA Research Institute, Chiba, Japan) for technical support.