TORC1 of fission yeast is rapamycin-sensitive

Authors

  • Terunao Takahara,

    1. Institute of Molecular and Cellular Biosciences, The University of Tokyo, Bunkyo-ku, Tokyo, Japan
    Search for more papers by this author
    • T.T. was a recipient of the Research Fellowship of JSPS for Young Scientists.
  • Tatsuya Maeda

    Corresponding author
    • Institute of Molecular and Cellular Biosciences, The University of Tokyo, Bunkyo-ku, Tokyo, Japan
    Search for more papers by this author

  • Communicated by: Masayuki Yamamoto

Correspondence: maeda@iam.u-tokyo.ac.jp

Abstract

The target of rapamycin (TOR) protein kinase plays central roles in the regulation of cell growth in response to nutritional availability. TOR forms two distinct multiprotein complexes termed TOR complex 1 (TORC1) and TORC2. Typically, only the activity of TORC1 is inhibited by the immunosuppressant rapamycin. Although rapamycin strongly inhibits cell growth of the budding yeast Saccharomyces cerevisiae through inhibition of TORC1, growth of the fission yeast Schizosaccharomyces pombe appears to be resistant to rapamycin. Here, we demonstrate that rapamycin inhibits the kinase activity of S. pombe TORC1 in vitro in a similar manner to TORC1 of other organisms. We furthermore show that incomplete inhibition of TORC1 by rapamycin underlies the apparent rapamycin resistance of S. pombe. In the presence of caffeine, which potentially lowers TORC1 activity, the growth of wild-type S. pombe cells is sensitive to rapamycin in a TORC1-dependent manner. Moreover, treatment of S. pombe cells with rapamycin plus caffeine induces starvation-specific gene expression and autophagy, similarly to cells with reduced TORC1 activity. These results indicate that rapamycin does inhibit TORC1 in S. pombe, but the inhibition is not sufficient to cause a growth defect. These findings establish a universal action of rapamycin on TORC1 inhibition.

Introduction

The target of rapamycin (TOR) is an evolutionarily conserved Ser/Thr kinase that controls cell growth and proliferation in response to nutrient availability (Wullschleger et al. 2006; Avruch et al. 2009). TOR was originally identified in the budding yeast Saccharomyces cerevisiae as the target of the immunosuppressive and anticancer drug, rapamycin (Heitman et al. 1991; Wullschleger et al. 2006; Avruch et al. 2009). Together with its cellular receptor FKBP12, rapamycin binds to the FKBP12-rapamycin-binding (FRB) domain of TOR and inhibits the activity of TOR (Harris & Lawrence 2003), although the detailed mechanism of this inhibition is still unknown. Budding and fission yeast have two TOR proteins (Tor1 and Tor2), whereas higher eukaryotes generally have only one TOR. TOR forms two structurally and functionally distinct multiprotein complexes, termed TOR complex 1 (TORC1) and TORC2 (Wullschleger et al. 2006; Avruch et al. 2009). In S. cerevisiae, TORC1 contains the proteins Kog1 and Lst8 in addition to either Tor1 or Tor2, and TORC2 contains Tor2, Avo1, Tsc11/Avo3, and Lst8 proteins. The subunits of each complex are also highly conserved from yeast to mammals. In general, TORC1 is rapamycin-sensitive and regulates various processes related to cell growth such as protein synthesis, autophagy, and ribosome biogenesis, whereas TORC2 regulates the organization of the actin cytoskeleton (Wullschleger et al. 2006). The best-characterized direct substrates of mammalian TORC1 (mTORC1) are eukaryotic translation initiation factor 4E-binding protein 1 (4E-BP1) and p70 ribosomal protein S6 kinase 1 (S6K1), both of which are involved in protein synthesis (Wullschleger et al. 2006; Ma & Blenis 2009). More recently, Sch9 (the ortholog of S6K1) and Atg13 (a critical regulator of autophagy) were identified as direct substrates of TORC1 in S. cerevisiae (Urban et al. 2007; Kamada et al. 2010).

In the fission yeast Schizosaccharomyces pombe, TORC1 is composed of Tor2 together with Mip1 (the Kog1 ortholog) and Wat1 (the Lst8 ortholog), TORC2 is composed of Tor1 together with Sin1 (the Avo1 ortholog), Ste20 (the Tsc11/Avo3 ortholog), and Wat1 (Hayashi et al. 2007; Matsuo et al. 2007; Otsubo & Yamamato 2008). Tor2 is essential for S. pombe survival, and recent reports have demonstrated that inactivation of Tor2 results in phenotypes similar to those of nitrogen-starved cells, including cell cycle arrest in the G1 phase, initiation of sexual development, and induction of nitrogen starvation responsive gene expression (Alvarez & Moreno 2006; Uritani et al. 2006; Matsuo et al. 2007; Weisman et al. 2007; Otsubo & Yamamato 2008). On the other hand, Tor1, as well as the other components of TORC2, are required for G1 arrest and sexual development in response to starvation, and for resistance to stresses such as osmotic and oxidative stresses (Hilti et al. 1999; Kawai et al. 2001; Weisman et al. 2001; Ikeda et al. 2008). The molecular mechanisms by which TOR regulates these events remain poorly understood. One remarkable feature of S. pombe is that rapamycin does not inhibit cell growth, which is in marked contrast to the effect of rapamycin in S. cerevisiae. However, rapamycin can inhibit some physiological processes in S. pombe including sexual development and amino acid uptake (Weisman et al. 1997, 2005; Weisman & Choder 2001) and also advances the onset of mitosis (Petersen & Nurse 2007). Moreover, a recent report has shown that treatment of S. pombe cells with rapamycin leads to dephosphorylation of the ribosomal protein S6 (Rps6) (Nakashima et al. 2010). Thus, rapamycin can have adverse effects on wild-type cells without affecting cell growth. Interestingly, several S. pombe mutants that are sensitive to rapamycin have been described (Weisman 2010). Rapamycin inhibits the growth of leucine auxotrophs by inhibiting the uptake of leucine, although this inhibition appears to occur in a Tor1-dependent manner (Weisman et al. 2005). Furthermore, S. pombe mutants that are expected to have altered TORC1 activity are also sensitive to rapamycin (Hayashi et al. 2007; Slegtenhorst et al. 2007). For example, rapamycin inhibits the growth of cells harboring a specific tor2 temperature-sensitive allele, tor2-287 and leads to arrest in the G1 phase (Hayashi et al. 2007). However, it remains to be clarified whether rapamycin can directly inhibit fission yeast TORC1 activity in wild-type cells and the relevance of this inhibition to the growth of wild-type prototrophic cells.

Here, we demonstrate that fission yeast TORC1 kinase activity is indeed inhibited by the rapamycin-FKBP12 complex in vitro. This observation prompted us to determine whether incomplete inhibition of TORC1 by rapamycin may underlie the rapamycin resistance of S. pombe. We show that the growth of wild-type fission yeast cells becomes sensitive to rapamycin in a TORC1-dependent manner when these cells are treated with caffeine, which potentially decreases TOR activity as reported in S. cerevisiae (Reinke et al. 2006) and mammals (Sarkaria et al. 1999). Moreover, consistent with TORC1 inactivation, the addition of rapamycin together with caffeine results in the induction of nitrogen starvation responsive gene expression and of autophagy, which are similar to the responses of cells with inactivated TORC1.

Results

Fission yeast TORC1 kinase activity is inhibited by the rapamycin-FKBP12 complex

Previous studies using S. pombe indicated that rapamycin treatment inhibits Tor2-dependent Rsp6 phosphorylation (Nakashima et al. 2010) and that the tor2-287 mutant exhibits hypersensitivity to rapamycin (Hayashi et al. 2007), suggesting that rapamycin can also inhibit TORC1 activity in S. pombe. To directly test this possibility, we examined the effect of rapamycin on S. pombe TORC1 kinase activity in vitro. As Mip1 is the defining component of TORC1 (Otsubo & Yamamato 2008), TORC1 was immunopurified from the lysates of cells co-expressing HA-tagged Mip1 (Mip1-HA) and Flag-tagged Tor2 (Flag-Tor2) using an anti-HA antibody, and the immunoprecipitates were subjected to a kinase assay in the presence of Mn2+, under which mTORC1 kinase activity is enhanced (Sato et al. 2009). Because the endogenous direct substrate of TORC1 in S. pombe is not yet known, we used 4E-BP1, a direct substrate of mTORC1 (Ma & Blenis 2009), as the substrate in this assay. There are several precedents for the use of 4E-BP1 as a substrate for the evaluation of TORC1 and TORC2 kinase activities in other organisms (Audhya et al. 2004; Jacinto et al. 2004; Wullschleger et al. 2005). As shown in Fig. 1A, like mTORC1, S. pombe TORC1 also phosphorylated 4E-BP1 at Thr37/46 in a time-dependent fashion (Fig. 1A). We therefore used the phosphorylation levels of these sites as a read-out of the kinase activity of S. pombe TORC1 in vitro.

Figure 1.

Fission yeast TORC1 kinase activity is inhibited by the rapamycin-FKBP12 complex. (A) Schizosaccharomyces pombe TORC1 phosphorylates 4E-BP1 in vitro. Cells expressing HA-tagged Mip1 (Mip1-HA, JT176) or untagged Mip1 (JY450) were transformed with a plasmid encoding Flag-Tor2 (pREP41-His6Flag2-tor2) and were subjected to immunoprecipitation with an anti-HA antibody. Kinase reactions of the precipitates were carried out as described in Experimental Procedures for the indicated times using 4E-BP1 as the substrate. The phosphorylation level of 4E-BP1 was examined by Western blotting with an anti-phospho-4E-BP1 (Thr37/46) antibody. (B) Rapamycin inhibits the kinase activity of S. pombe TORC1. The TORC1 immunoprecipitates described in (A) were preincubated with rapamycin and/or GST-FKBP12 for 10 min before initiation of the kinase reaction, which was carried out as in (A). (C) TORC1 kinase assay was carried out in the presence of Mg2+, as in (B). (D) Rapamycin does not induce dissociation of Mip1 from Tor2 in vitro. TORC1, immunoprecipitated as in (A), was treated with rapamycin and/or GST-FKBP12 for 10 min. After separation of the supernatant (Sup) from the beads, the amount of Mip1-HA and Flag-Tor2 in each fraction was examined by Western blotting. (E) Rapamycin does not induce dissociation of Mip1 from Tor2 in vivo. Logarithmically growing cells were treated with the indicated concentrations of rapamycin for 45 min. After harvesting the cells, Mip1-HA was immunoprecipitated from cell lysates with an anti-HA antibody and the amount of Flag-Tor2 associated with Mip1-HA was determined by Western blotting. TOR, target of rapamycin.

The addition of either rapamycin or GST-FKBP12 alone did not affect the kinase activity of TORC1 (Fig. 1B). However, addition of rapamycin and GST-FKBP12 together specifically inhibited the kinase activity of TORC1 in a rapamycin dose-dependent manner (Fig. 1B). The TORC1 kinase activity was also inhibited by FKBP12-rapamycin in the presence of Mg2+ (Fig. 1C). The concentration of rapamycin that inhibited the activity of S. pombe TORC1 was comparable with that which inhibits mTORC1 (Jacinto et al. 2004; Oshiro et al. 2004). These results suggest that rapamycin can inhibit the TORC1 kinase activity of S. pombe as well as of S. cerevisiae and mammals.

In mammalian cases, rapamycin together with FKBP12 decreases mTOR-raptor interaction in vivo and in vitro, although the decrease is not likely to be the cause of acute mTORC1 inhibition (Kim et al. 2002; Oshiro et al. 2004). To determine whether rapamycin decreases Tor2 interaction with the raptor ortholog Mip1 in S. pombe, we examined the effect of the rapamycin-FKBP12 complex on the interaction between Tor2 and Mip1 in vitro. As shown in Fig. 1D, the addition of both rapamycin and GST-FKBP12 to TORC1 in vitro had little, if any, effect on the interaction of Mip-1-HA with Flag-Tor2, despite the fact that this treatment strongly inhibited TORC1 kinase activity (Fig. 1B,C). Furthermore, the addition of rapamycin in vivo also did not appear to decrease the association of Flag-Tor2 and Mip1-HA (Fig. 1E). Thus, in S. pombe, rapamycin is unlikely to promote dissociation of Mip1 from Tor2. This result is consistent with the fact that TORC1 integrity in S. cerevisiae is not disrupted in the presence of rapamycin (Loewith et al. 2002). Therefore, rapamycin generally cannot disrupt TORC1 integrity and likely inhibits TORC1 activity by a mechanism other than destabilization of TORC1 integrity.

Rapamycin inhibits fission yeast growth through inhibition of TORC1 in the presence of caffeine

The above results suggest that fission yeast TORC1 is inhibited by rapamycin treatment in vivo. Nevertheless, the growth of wild-type prototrophic cells is not inhibited by rapamycin (Fig. 2A). This discrepancy could be due to that rapamycin does inhibit TORC1 in vivo in S. pombe, but that this inhibition alone is unable to sufficiently lower TORC1 activity to a level at which it would cause a growth defect. Interestingly, the mutants that possibly result in the decrease in the TORC1 activity show rapamycin-sensitive growth (Weisman et al. 2010). In addition, recent studies on mTORC1 using novel mTOR-specific inhibitors have also revealed that considerable mTORC1 functionality is resistant to rapamycin (Feldman et al. 2009; García-Martínez et al. 2009; Thoreen et al. 2009). To test the idea, we determined the effect of rapamycin on cell growth under conditions of reduced TORC1 activity. We took advantage of the fact that caffeine targets TOR (Reinke et al. 2006) and therefore potentially decreases the basal TORC1 activity in S. pombe. As shown in Fig. 2A, cell growth was indeed dramatically inhibited by rapamycin in the presence of 6 mm caffeine, at a caffeine concentration which alone only marginally affected cell growth. To examine whether these growth defects were indeed dependent on rapamycin inhibition of TORC1, we constructed Tor2 mutants in which Ser1837 in the FRB domain was substituted with one of several different amino acids. Homologous mutations are known to confer resistance to rapamycin in both S. cerevisiae and mammalian TORs (Cafferkey et al. 1993; Helliwell et al. 1994; Chen et al. 1995). We then examined cell growth of these mutants in the presence of both rapamycin and caffeine. Substitution of Arg, Thr, or Glu, but not of Ala, for Ser1837 conferred resistance to rapamycin in the presence of caffeine, suggesting that the observed growth defect is indeed because of TORC1 inhibition by rapamycin. As the S1837T mutant exhibited comparable growth with wild-type cells even in the presence of a higher concentration of caffeine (9 mm) and was also strongly resistant to rapamycin, we used this Tor2 mutant in the following analyses to test whether the observed responses were induced through the rapamycin-dependent inhibition of Tor2.

Figure 2.

Rapamycin inhibits the cell growth of Schizosaccharomyces pombe in the presence of caffeine. (A) Tenfold serial dilutions of WT (TS205), tor2S1837R (TS208), tor2S1837T (TS217), tor2S1837A (TS214), and tor2S1837E (TS211) cells were spotted onto yeast extract (YE) containing the indicated concentrations of rapamycin and/or caffeine and were grown for 3 days at 30 °C. (B) Tenfold serial dilutions of wild-type; tor2 (TS205), AV + WV; tor2A1822V, W2041V (TS765), ST; tor2S1837T (TS217), and ST + AV + WV; torS1837T, A1822V, W2041V (TS766) cells were spotted onto YE containing the indicated concentrations of rapamycin and/or caffeine and were grown for 3 days at 30 °C. TOR, target of rapamycin.

To further confirm whether the effect of caffeine on the growth in the presence of rapamycin mainly attributed to the inhibition of TORC1, we sought to make Tor2 mutants conferring caffeine resistance. A previous study in S. cerevisiae identified several mutations within Tor1 that caused resistance to caffeine: the mutation at Trp2176 as a result of decreased binding of caffeine, and the mutations at Ile1954 or Ala1957 probably due to increased activity of Tor1 (Reinke et al. 2006; Sturgill & Hall 2009) As shown in Fig. 2B, tor2 mutant (AV + WV), which has the substitutions at Ala1822 to Val and at Trp2041 to Val (the homologous sites to Ala1957 and Trp2176 in S. cerevisiae Tor1, respectively), was resistant to the condition of caffeine plus rapamycin at a comparable level with the rapamycin-resistant tor2S1837T mutant. Moreover, tor2 mutant having these three mutations (ST + AV + WV) exhibited further resistance to the condition of caffeine plus rapamycin. These results suggested that caffeine also mainly targets Tor2 in S. pombe.

Previous reports have demonstrated that inactivation of tor2 temperature-sensitive mutants by shifting to the nonpermissive temperature, or repression of tor2 expression, leads to arrest at the G1 phase, which is similar to the response of nitrogen-starved cells (Alvarez & Moreno 2006; Uritani et al. 2006; Hayashi et al. 2007; Matsuo et al. 2007; Weisman et al. 2007). To examine the effect of rapamycin and/or caffeine treatment on cell cycle progression of wild-type cells, we carried out flow cytometric analysis. As reported previously, nitrogen starvation led to an increase in the population of cells arrested at G1 (Fig. 3, designated 1C) (Kohda et al. 2007). Neither rapamycin nor caffeine alone caused G1 arrest as expected from their lack of effect on cell growth. Unlike the Tor2 inactivation, however, the treatment with rapamycin plus caffeine, although inhibits the cell growth, did not cause G1 arrest of wild-type cells. The failure of this treatment to cause G1 arrest was probably due to the effect of caffeine, because simultaneous treatment with caffeine during nitrogen starvation significantly reduced the population of G1 arrested cells (Fig. 3). Thus, the growth defect observed under conditions of rapamycin plus caffeine treatment was not necessarily because of G1 arrest.

Figure 3.

Rapamycin plus caffeine does not cause G1 arrest. WT (TS205) cells were grown to log phase in YE and were then treated with rapamycin (+Rapa, 200 ng/mL) and/or caffeine (+Caff, 6 mm) or were starved for nitrogen by transferring into EMM-N without (−N) or with caffeine (−N + Caff, 6 mm). Aliquots of the cells were taken after incubation for the indicated times and were subjected to flow cytometry to measure their DNA content. The positions for 1C and 2C contents were indicated. EMM, edinburgh minimal medium; YE, yeast extract.

Cells treated with rapamycin plus caffeine exhibit nitrogen starvation-like responses

Inactivation of tor2 has been shown to induce the expression of a specific set of genes, most of which are also induced upon nitrogen starvation (Sato et al. 1994; Mata et al. 2002; Alvarez & Moreno 2006; Uritani et al. 2006; Matsuo et al. 2007). We therefore examined whether cells treated with rapamycin plus caffeine phenocopy tor2 deficiency in inducing the expression of two of these genes, the mating inducer ste11 and the vacuolar protease isp6 genes, in S. pombe. As shown using northern analysis, neither rapamycin nor caffeine alone induced the expression of these genes, but combined treatment of wild-type cells with rapamycin and caffeine did induce the expression of both ste11 and isp6 genes after 2 h (Fig. 4). The induction of these genes was markedly inhibited in the tor2S1837T mutant, although nitrogen starvation was able to induce the expression of these genes in both wild-type and tor2S1837T cells. These results suggest that rapamycin indeed inactivates Tor2 activity in the presence of caffeine, which results in induction of specific gene expression. The induction of these genes appears to be primarily regulated by inactivation of Tor2 but not by arrest at the G1 phase.

Figure 4.

Genes that are specifically up-regulated by nitrogen starvation are also induced by rapamycin in the presence of caffeine in a Tor2-dependent manner. Total RNA was extracted from WT (TS205) and tor2S1837T (TS217) cells treated with rapamycin (+Rapa, 200 ng/mL), caffeine (+Caff, 6 mm), or both (+Rapa + Caff), or deprived of nitrogen (−N) for the indicated times. The mRNA expression levels of ste11 and isp6 were examined by Northern blotting and rRNA is included as a loading control. TOR, target of rapamycin.

In other organisms, nutrient starvation promotes autophagy induction through inactivation of TORC1 (Noda & Ohsumi 1998; Ohne et al. 2008). In fission yeast, although autophagy is also induced following nitrogen starvation (Nakashima et al. 2006; Kohda et al. 2007; Mukaiyama et al. 2010), the involvement of TORC1 has not yet been clarified. Because we observed that TORC1 activity was inactivated by rapamycin in the presence of caffeine, we determined whether autophagy is induced under this condition. To this end, we used GFP-Atg8 as a marker of autophagy induction. In the vegetative growth phase, GFP-Atg8 was dispersed throughout the cytoplasm (Fig. 5). However, following nitrogen starvation for 3 h, GFP-Atg8 formed a few foci, whose formation is an indicator of autophagy induction (Mukaiyama et al. 2010) as in mammalian cells. No foci were formed following addition of rapamycin alone, indicating that rapamycin does not induce autophagy. Addition of rapamycin plus caffeine induced foci formation of GFP-Atg8 in wild-type cells, but not in the tor2S1837T mutant. The number of cells with GFP-Atg8 foci following treatment with rapamycin plus caffeine was less than that following nitrogen starvation, however, probably due to incomplete inactivation of TORC1. These results suggest that autophagy is also induced under conditions of TORC1-repression that is triggered by rapamycin in the presence of caffeine.

Figure 5.

Autophagy is induced by rapamycin in the presence of caffeine in a Tor2-dependent manner. WT (TS205) and tor2S1837T (TS217) cells were grown to log phase in edinburgh minimal medium (EMM) and were then treated with rapamycin (200 ng/mL) and/or caffeine (9 mm) or were shifted to EMM-N for 3 h. GFP-Atg8 foci that are indicative of autophagy were observed under a fluorescence microscope and are indicated by arrowheads. Scale bar, 10 μm. TOR, target of rapamycin.

Discussion

A long-standing mystery in TOR signaling of the fission yeast S. pombe is whether rapamycin can directly inhibit TOR activity, because, in contrast to the budding yeast S. cerevisiae, the growth of S. pombe is not inhibited by rapamycin. In the present study, we have shown that the TORC1 kinase activity of S. pombe is inhibited by the rapamycin-FKBP12 complex in vitro, in a similar manner to TORC1 of other organisms. To our knowledge, this is the first report to demonstrate that TORC1 kinase activity is directly inhibited by rapamycin in S. pombe.

Nakashima et al. (2010) have recently demonstrated that rapamycin inhibits Rps6 phosphorylation in S. pombe as it does in other organisms. Especially, the observation that the effect of rapamycin was reversed by the rapamycin-binding defective Tor2S1837E mutant (Nakashima et al. 2010) strongly suggested that TORC1 is the direct target of rapamycin. However, the mechanism of rapamycin action through the binding to TORC1 was not conclusive as it had been attributed to either inactivation of TORC1 kinase activity or activation of counteracting protein phosphatase activity. According to a study in S. cerevisiae, the latter of which occurs by liberating the TORC1-associated phosphatases upon rapamycin-binding to TOR (Yan et al. 2006), and therefore, the result obtained from the Tor2S1837E mutant was also consistent with the failure of this mechanism. In contrast, our result clearly demonstrated that the TORC1 kinase activity of S. pombe is indeed inhibited by rapamycin.

In mammalian cells, rapamycin is known to dissociate raptor from mTOR (Kim et al. 2002; Oshiro et al. 2004). However, we demonstrated that in vitro addition of rapamycin together with FKBP12 to TORC1 did not cause dissociation of Mip1 from Tor2 (Fig. 1D), despite the apparent reduction in the kinase activity of TORC1 toward 4E-BP1 (Fig. 1B,C). In addition, rapamycin did not appear to affect the association of these proteins in vivo (Fig. 1E). Similarly, the association of their orthologs in S. cerevisiae appears insensitive to rapamycin (Loewith et al. 2002). These data suggest that the mechanism by which TORC1 kinase activity is inhibited does not necessarily involve dissociation of the scaffold protein (e.g., raptor, Kog1, and Mip1) that mediates the action of the kinase activity of TOR. Indeed, using autophosphorylation as a read-out of mTORC1 kinase activity, Soliman et al. (2010) have recently suggested that rapamycin can directly inhibit the intrinsic catalytic activity of mTORC1 in addition to its effect of dissociating raptor from mTOR. Our present results, together with the lack of the effect of rapamycin on TORC1 integrity in S. cerevisiae, are consistent with an ability of rapamycin to directly inhibit TORC1 catalytic activity, although we cannot exclude the possibility that changes in other protein(s) associated with TORC1 may account for the reduction in TORC1 kinase activity.

We have also shown here that rapamycin treatment together with caffeine prevents cell growth and induces starvation-specific gene expression and autophagy in S. pombe. The concentration of caffeine used in our experiments did not have significant effects on these parameters when treated alone (Figs 2, 4, 5). However, caffeine can potentially reduce TORC1 activity to some extent in S. pombe as it does in S. cerevisiae (Reinke et al. 2006). We speculate that the slight inhibition of TORC1 activity induced by caffeine enables rapamycin to inhibit critical TORC1 effector(s) for these parameters that are not sufficiently affected by rapamycin alone, thus leading to apparent effects. Rapamycin probably inhibits such critical effector(s) by more completely inhibiting TORC1 activity in the presence of caffeine. A similar mechanism may underlie the rapamycin-sensitive growth of the tor2-287 temperature-sensitive mutant at the permissive temperature (Hayashi et al. 2007), where TORC1 activity is potentially lowered to a level sufficient to cause rapamycin-sensitivity but not sufficient to cause growth arrest.

The different effects between following direct Tor2 inactivation, growth arrest at G1 (Alvarez & Moreno 2006; Uritani et al. 2006; Hayashi et al. 2007; Matsuo et al. 2007; Weisman et al. 2007), and following the treatment with rapamycin plus caffeine, growth inhibition without G1 arrest (Fig. 3), suggest that caffeine has additional target(s) other than Tor2. Caffeine probably inhibits Tor1 and Tor2 similarly, because G1 arrest caused by Tor2 inactivation is no longer observed in a tor1Δ background (Uritani et al. 2006; Weisman et al. 2007), which is similar situation to what is caused by the treatment with rapamycin plus caffeine, and the same concentration of caffeine led to the reduction in G1-arrested cells under the nitrogen-starved condition. In addition, because caffeine targets phosphatidylinositol 3-kinase-related kinase (PIKK) family proteins, ataxia-telangiectasia mutated (ATM) and ATM- and Rad3-related (ATR) in mammalian cells (Sarkaria et al. 1999), it is possible that members of the PIKK family are down-regulated to some extent in S. pombe as well, although combination of caffeine with rapamycin specifically inhibits TORC1 activity (Fig. 2B). In contrast, we and others have observed no hint of TORC2 inhibition by rapamycin in S. pombe. Indeed, a recent study demonstrated that most downstream events of TORC2 are rapamycin-insensitive (Schonbrun et al. 2009).

To date, Rps6 is the only known downstream effector of TORC1 in S. pombe whose phosphorylation is strongly inhibited by rapamycin alone (Nakashima et al. 2010). Apparently, inactivation of the TORC1-Rps6 axis alone is not sufficient to strongly inhibit the regulation of growth, of gene expression dictated by TORC1, or of autophagy, because these parameters are only affected by treatment with rapamycin plus caffeine. Interestingly, a similar partial effect of rapamycin has recently been revealed in mammalian cells by using novel mTOR inhibitors (Feldman et al. 2009; García-Martínez et al. 2009; Thoreen et al. 2009). Moreover, the homologous mTORC1-S6K1-S6 axis in mammalian cells is potently inhibited by rapamycin (Choo et al. 2008), whereas the effect of rapamycin on induction of autophagy or inhibition of cell proliferation is modest and highly cell type-dependent (Thoreen et al. 2009). The result that the in vitro TORC1 kinase activity of both S. pombe and mammals can be comparably inhibited by rapamycin (Fig. 1) also supports the fundamental similarity of the responses to rapamycin between them. Although it remains uncertain how the specificity of downstream effectors that are inhibited by rapamycin is determined, the effectors that can escape from rapamycin inhibition might require only a weak TORC1 kinase activity to be phosphorylated. The fact that caffeine treatment converts rapamycin-insensitive parameters to rapamycin-sensitive (Figs 2, 4, 5) also suggests that the difference in rapamycin-sensitivity among effectors is quantitative rather than qualitative, as caffeine inhibits TORC1 in an ATP-competitive manner and therefore likely affects any effectors indiscriminately. Future identification of TORC1 effectors by differential phosphoproteome analyses following the treatment of cells with rapamycin in the presence or absence of caffeine will highlight the factors that are involved in rapamycin resistance in S. pombe, which may also be conserved in mammals.

Experimental procedures

Yeast strains and growth media

The fission yeast strains used in this study are listed in Table 1. Yeast extract media (YE) and Edinburgh minimal medium (EMM) supplemented with 50 mg/L adenine (when needed), which contain 2% glucose and 93.5 mm NH4Cl as carbon and nitrogen sources, respectively, were used as growth media. For nitrogen starvation experiments, EMM-N, which is the nitrogen-free version of EMM, or SD-N, which contains 0.13% Difco yeast nitrogen base without amino acids and ammonium sulfate (BD Biosciences, Sparks, MD, USA) and 2% glucose was used.

Table 1. Schizosaccharomyces pombe strains used in this study
StrainGenotypeaSource
  1. a

    <<Insertion of resistance cassette.

JY1hKohda et al. (2007)
TS205htor2<<kanrThis study
TS208htor2S1837R<<kanrThis study
TS211htor2S1837E<<kanrThis study
TS214htor2S1837A<<kanrThis study
TS217htor2S1837T<<kanrThis study
JT268hatg8.NGFPKohda et al. (2007)
JT176h90ade6-M216 leu1 mip1-3HA<<kanrMatsuo et al. (2007)
JY450h90ade6-M216 leu1Matsuo et al. (2007)
TS244hatg8.NGFP tor2<<kanrThis study
TS357hatg8.NGFP tor2S1837T<<kanrThis study
TS765htor2A1822V, W2041V<<kanrThis study
TS766htor2S1837T, A1822V, W2041V<<kanrThis study

Construction of yeast strains

For construction of a series of tor2 mutants, we generated plasmids in which the Ser residue at 1837 of Tor2 was replaced with Arg, Glu, Ala, or Thr by site-directed mutagenesis using the pUG6-tor2-CT plasmid (Urano et al. 2007) as a template. The plasmids harboring the substitutions of both Ala1822 to Val (A1822V) and Trp2041 to Val (W2041V) of Tor2 were similarly constructed. The resultant plasmids were linearized with BamHI and integrated into the genome of the wild-type strain (JY1) by homologous recombination using a lithium acetate method (Okazaki et al. 1990).

Immunoprecipitation and in vitro kinase assay

JY450 and JT176 (mip1-HA) cells harboring the pREP41-His6Flag2-tor2 plasmid (Matsuo et al. 2007) were grown on EMM at 30 °C to an OD600 of 0.5. Cells were collected by centrifugation, washed twice with water, and then lysed in buffer A [40 mm HEPES-NaOH (pH 7.5), 120 mm NaCl, 1 mm EDTA, 0.3% CHAPS, 50 mm NaF, 10 mm β-glycerophosphate, 40 μg/mL aprotinin, 40 μg/mL leupeptin, 20 μg/mL pepstatin A, 1 mm PMSF] using ceramics beads and a Multi-Beads Shocker (Yasui Kikai, Osaka, Japan) with 10 × 30 s pulses. After centrifugation at 13 000 g for 10 min, the supernatants were incubated overnight with an anti-HA antibody (12CA5, ascites fluid) and then with protein A Sepharose (GE Healthcare, Little Chalfont, UK) for 1 h. Immune complexes were washed four times with buffer A, and twice with kinase buffer [25 mm Hepes-NaOH (pH 7.5), 50 mm NaCl, 50 mm NaF]. The kinase assay was carried out in 30 μL volume of kinase reaction buffer [25 mm Hepes-NaOH (pH 7.5), 50 mm NaCl, 50 mm NaF, 10 mm β-glycerophosphate, 10 mm MnCl2, 500 ng 4E-BP1 (P7992; Sigma, St. Louis, MO, USA)] for 10 or 20 min at 30 °C. In Fig. 1C, 10 mm MnCl2 was replaced with 10 mm MgCl2 and the kinase reaction was carried out for 30 min at 30 °C. The indicated concentration of rapamycin (LC Laboratories, Woburn, MA, USA) and GST-FKBP12 (600 ng) were preincubated with immune complexes for 10 min at 30 °C before starting the kinase assay by adding ATP at a final concentration of 500 μm. Kinase reaction was stopped by adding 10 μL of 4× Laemmli SDS sample buffer and then boiled for 5 min. Reaction mixture was resolved with SDS-PAGE and transferred to Immobilon-P polyvinylidene difluoride membrane (Millipore, Billerica, MA, USA). After blocked with 1% skim milk, membranes were immunoblotted with the following primary antibodies: anti-Flag M2 (F3165; Sigma), anti-HA (12CA5, ascites fluid) and anti-Phospho-4E-BP1 (Thr37/46) (#9459; Cell Signaling Technology, Danvers, MA, USA). Then, membranes were incubated with secondary antibody conjugated with horseradish peroxidase, anti-mouse IgG (NA931; GE Healthcare) or anti-rabbit IgG (NA934; GE healthcare), and developed with ECL Plus (GE Healthcare). GST-FKBP12 (human) was purified from the E. coli BL21(DE3) strain harboring pGEX-5X-3-FKBP12 using glutathione Sepharose (GE Healthcare), according to the manufacturer's recommendations. For examination of the effect of rapamycin in vivo (Fig. 1E), the indicated concentration of rapamycin was added into logarithmically growing cells for 45 min, and the cell lysates were subjected to immunoprecipitation with the anti-HA antibody as described previously.

Flow cytometry

Cells were collected by centrifugation, washed once with water, and resuspended in 0.3 mL of water. After mixing with 0.7 mL of ethanol, the cells were left overnight at −20 °C. The fixed cells were washed once with 0.5 mL of FACS buffer [0.2 m Tris–HCl (pH 7.5), 50 mm EDTA] and were resuspended in 0.5 mL of FACS buffer. The cells were briefly sonicated, were treated with RNase A for 4 h at 37 °C, and were then resuspended in FACS buffer containing 10 μg/mL propidium iodide. After incubation for 1 h at 4 °C, the cells (2 x 104) were analyzed using FACSCalibur with CellQuest software (BD Biosciences).

Northern blotting

Cells were grown to log phase in YE and were then treated with rapamycin (200 ng/mL) and caffeine (6 mm) or were transferred into SD-N for the indicated times. Total RNA was extracted from the cells by vortexing the cells in extraction buffer [0.2 m Tris–HCl (pH 7.5), 0.5 m NaCl, 10 mm EDTA, 1% SDS] with glass beads using 10 × 30 s pulses, and 10 μg of total RNA were then subjected to Northern analysis using the entire ORF of isp6 or ste11 as a probe. The DIG High Prime DNA Labeling and Detection Starter Kit II (Roche Applied Science, Penzberg, Germany) coupled with CDP-Star (Applied Biosystems, Carlsbad, CA, USA) was used for detection according with the manufacturers' instructions.

Fluorescence microscopy

Cell were grown to log phase in EMM and treated as indicated earlier. All images were acquired using a fluorescence microscope (BX51; Olympus, Tokyo, Japan) with an UplanApo 100x/1.35 oil objective lens (Olympus).

Acknowledgements

We thank Yutaka Hoshikawa, Yoko Otsubo, Fuyuhiko Tamanoi, and Masayuki Yamamoto for various materials and Ayano Kagami and Yoshinori Watanabe for access to FACSCalibur. We also thank Kenji Kitamura for helpful suggestion, and all members of the Maeda laboratory for help and discussion. This work was supported in part by a Grant-in-Aid for Scientific Research on Priority Areas (KAKENHI 21025008) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), a Grant-in-Aid for Exploratory Research (KAKENHI 20658027) and a Grant-in-Aid for Challenging Exploratory Research (KAKENHI 23651233) from the Japan Society for the Promotion of Science (JSPS), grants from the Noda Institute for Scientific Research and the Salt Science Research Foundation (No. 11D2) (all to T.M.), and a Grant-in-Aid for JSPS Fellows (to T.T.).

Ancillary