The target of rapamycin (TOR) signaling pathway plays crucial roles in the regulation of eukaryotic cell growth. In Saccharomyces cerevisiae, nitrogen sources in the extracellular environment activate the TOR signaling pathway. However, the precise mechanisms underlying the regulation of TOR activity in response to extracellular nitrogen sources are poorly understood. Here, we report that degradation of Stp1, a transcription factor for amino acid uptake and a key effector of the SPS amino-acid-sensing pathway, is controlled by TOR activity in S. cerevisiae. Using a genome-wide protein localization study, we found that Stp1 disappeared from the nucleus upon inactivation of TOR complex 1 (TORC1) by rapamycin, suggesting the involvement of Stp1 in the TOR signaling pathway. Supporting this notion, a knockout mutant for the STP1 gene was found to be hypersensitive to rapamycin, and overexpression of STP1 conferred resistance to rapamycin. Interestingly, we found that the rapamycin-induced disappearance of Stp1 from the nucleus resulted from Stp1 degradation, which was dependent on the activity of a protein phosphatase 2A (PP2A)-like phosphatase, Sit4, which is a well-known downstream effector of TORC1. Taken together, our findings highlight an intimate connection between the amino-acid-sensing pathway and the rapamycin-sensitive TOR signaling pathway.
The target of rapamycin (TOR) kinases belong to a family of phosphatidylinositol kinase-related kinases and function as a key regulator of eukaryotic cell growth in response to nutrient availability (Schmelzle and Hall, 2000). In yeast, Saccharomyces cerevisiae, TOR kinases are encoded by TOR1 and TOR2, and they exist in two functionally distinct multi-protein complexes, TOR complex 1 (TORC1) and TOR complex 2 (TORC2) (Loewith et al., 2002). TORC1 is responsible for the rapamycin-sensitive branch of the TOR signaling pathway. In S. cerevisiae, rapamycin forms a complex with yeast FKBP12 homolog, Fpr1, and this rapamycin-Fpr1complex inhibits TORC1 activity (Heitman et al., 1991; Koltin et al., 1991). Inactivation of TORC1 by rapamycin results in the induction of transcription of nitrogen catabolite-repressed genes (Beck and Hall, 1999; Cardenas et al., 1999; Hardwick et al., 1999). These changes in the transcriptional profile by rapamycin are carried out mainly by TORC1-controlled transcription factors, Gln3, Msn2 and Rtg1, which are relocated from the cytoplasm to the nucleus by inactivation of TORC1 (Beck and Hall, 1999; Komeili et al., 2000). It is known that Tap42-phosphatase complexes are major downstream targets of TORC1 and important in many TORC1-mediated responses (Di Como and Arndt, 1996; Duvel et al., 2003). Tap42-phosphatase complexes exist mainly on the membrane structure through interaction with TORC1. Rapamycin abrogates interaction of the Tap42-phosphatase complexes with TORC1, leading to activation of Tap42-associated phosphatases (Yan et al., 2006). In addition, it has been shown that nutrients stimulate association of Tap42 with a protein phosphatase 2A (PP2A)-like phosphatase, Sit4, whereas rapamycin induces the dissociation of the two proteins. Dissociation from Tap42 stimulates the phosphatase activity of Sit4, which then performs regulatory roles in the inactive TOR pathway (Di Como and Arndt, 1996; Rohde et al., 2004).
It is essential for yeast, as a unicellular organism, to adjust its growth according to the availability of extracellular nutrients. Amino acids are important nutrients for the activation of the TOR signaling pathway (Beck and Hall, 1999; Cardenas et al., 1999; Crespo et al., 2002; Hardwick et al., 1999; Magasanik and Kaiser, 2002), and are known to be sensed by a so-called Ssy1-Ptr3-Ssy5 (SPS) sensor, a multi-protein complex on the plasma membrane (Forsberg and Ljungdahl, 2001; Klasson et al., 1999). The amino acid signal derived from the SPS sensor is transmitted to the inside of the cell and regulates transcription of many genes involved in nitrogen uptake (Forsberg et al., 2001; Kodama et al., 2002). A transcription factor, Stp1, plays important roles in this amino acid signaling from the plasma membrane to the nucleus. Stp1 induces transcription of amino acid permease (AAP) genes by extracellular amino acid availability (Andreasson and Ljungdahl, 2002). When amino acids are abundant in the extracellular environment, Stp1 is activated by the endoproteolytic removal of N-terminal cytoplasmic retention motif, which is mediated by proteolytic activity of Ssy5, a component of the SPS sensor complex (Andreasson et al., 2006; Andreasson and Ljungdahl, 2004). It has also been shown that the SCFGrr1 ubiquitin E3 ligase complex and casein kinase-dependent phosphorylation are critical for processing of Stp1 N-terminal motif (Abdel-Sater et al., 2004). This processed Stp1 moves from the cytoplasm to the nucleus and activates the expression of its target AAP genes (Andreasson and Ljungdahl, 2002).
The facts that extracellular amino acids are important nutritional cues for activating the TOR pathway, and that Stp1 mediates the amino acid signaling from the plasma membrane to the nucleus for uptake of extracellular amino acids raise the possibility that Stp1 may be involved in the TOR pathway. In the present study, we show that Stp1 disappears from the nucleus upon inactivation of TORC1 by rapamycin. In addition, we show that the expression level of Stp1 influences rapamycin sensitivity of cells, and inactivation of TORC1 results in the degradation of Stp1 in a PP2A-like phosphatase Sit4-dependent manner. This suggests that Stp1 provides a novel link between the amino-acid-sensing pathway and the TOR pathway. Based on the finding that inactivated TORC1 regulates the activity of Stp1 by promoting protein degradation, we propose that Stp1-dependent amino acid uptake is intimately connected with the rapamycin-sensitive TOR signaling pathway.
Genome-wide protein localization study of rapamycin-treated cells
Protein localization information is a valuable resource in elucidating eukaryotic protein function and interaction. Several studies have shown that alteration in subcellular localization as a result of post-translational modification is an important mechanism for regulating protein function. Thus, alteration of the subcellular localization of a protein under a specific condition may suggest the possibility of functional change of the protein under the condition. Based on this correlation, we hypothesized that translocalization of a protein by rapamycin, a specific inhibitor of TORC1, might indicate its involvement in the TOR signaling pathway. We assumed that functions of the proteins involved in the TOR signaling pathway might be changed by inactivation of TORC1 and some of these functional changes might lead to alteration in subcellular localization of those proteins. To test this hypothesis, we carried out a genome-wide protein localization study of rapamycin-treated cells using 4159 green fluorescent protein (GFP)-tagged yeast strains (Huh et al., 2003). Cells of each GFP-tagged strain were treated with rapamycin for 2 hours and then subcellular localization of the GFP fusion protein was monitored by fluorescence microscopy. As a result, we observed alteration in subcellular localization of 98 proteins (Table 1).
However, there was a possibility that the proteins identified from the above screening included those whose translocalization might simply reflect an indirect consequence of TORC1 inactivation or cell growth inhibition. To further narrow down the candidate effectors of the TOR pathway, the positive proteins from the above screening were subjected to a second screening. The cells with these GFP fusion proteins were exposed to rapamycin for only 30 minutes, before the protein localization was monitored by fluorescence microscopy. This was done under the assumption that, if a protein were directly involved in the TOR pathway, functional change or translocalization of the protein would occur quickly in response to rapamycin. From this second screening, we found alteration in subcellular localization of 54 proteins (Table 1). Among them, Stp1 showed a dramatic alteration in localization; the GFP signal for Stp1 disappeared rapidly from the nucleus as a result of rapamycin treatment (Fig. 1A).
The expression level of Stp1 affects rapamycin sensitivity of cells
Stp1 is a well-known transcription factor that regulates the expression of several AAP genes (Eckert-Boulet et al., 2004). Stp1 is proteolytically activated by extracellular amino acids, and the processed Stp1 translocalizes to the nucleus where it activates transcription of its target AAP genes (Andreasson and Ljungdahl, 2002). The observation that rapamycin treatment resulted in loss of Stp1 from the nucleus in our screening strongly suggested that the activity of Stp1 would be downregulated by inactivation of the TOR pathway. To further investigate the relationship between Stp1 and the TOR pathway, we analyzed the effect of stp1Δ mutation on cell growth in a medium containing rapamycin. We found that the stp1Δ HY0331 strain derived from BY4741 was much more sensitive to rapamycin than the parent BY4741 strain (Fig. 1B). Considering that Stp1 plays a crucial role in amino acid uptake by inducing AAP genes, there was a possibility that deletion of the STP1 gene might cause severe amino acid limitation in the auxotrophic BY4741-derived strain and the rapamycin hypersensitivity of the stp1Δ HY0331 strain might result from this amino acid limitation. To check this possibility, we also analyzed the effect of stp1Δ mutation on rapamycin sensitivity of cells using a prototrophic strain FY1679-01B. Deletion of the STP1 gene in the FY1679-01B strain also made cells more sensitive to rapamycin (Fig. 1B), indicating that Stp1 is closely involved in the TOR pathway. To confirm the involvement of Stp1 in the TOR pathway, we also examined whether the cells overexpressing Stp1 would show resistance to rapamycin. As expected, overexpression of Stp1 conferred resistance to rapamycin (Fig. 1C). Taken together with the fact that Stp1 is activated by the SPS sensor in response to an extracellular amino acid signal, the close relationship between Stp1 expression level and rapamycin sensitivity suggests that Stp1 plays an important role in conveying the extracellular amino acid signal to the rapamycin-sensitive TOR pathway.
Inhibition of the TOR pathway promotes degradation of nuclear-localized Stp1
To determine whether the rapamycin-induced disappearance of Stp1 from the nucleus was the result of translocalization to the cytoplasm or a decrease in Stp1 protein level, we measured the Stp1 protein level by western blot analysis after rapamycin treatment or deprivation of nitrogen sources. Surprisingly, the rapamycin-induced disappearance of Stp1 from the nucleus turned out to be due to a decrease in the Stp1 level (Fig. 2A, upper panel). In addition to rapamycin treatment, nitrogen starvation also caused a rapid decrease in the level of processed low molecular mass form of Stp1, which is known to localize to the nucleus (Andreasson and Ljungdahl, 2002). We also confirmed a rapamycin-induced decrease in the Stp1 levels in prototrophic yeast cells (Fig. 2A, lower panel). Inactivation of the TOR pathway by rapamycin is known to result in mRNA turnover and inhibition of translation (De Virgilio and Loewith, 2006) and Stp1 has been shown to have a relatively short half-life (Belle et al., 2006). Therefore, we wanted to determine whether the decrease in Stp1 resulted from the downregulation of transcription or translation, or protein instability. To check this, STP1 mRNA was measured after rapamycin treatment or nitrogen starvation. We found that neither of the two conditions caused any downregulation of the STP1 mRNA level (Fig. 2B). We also tested whether blocking of translation by cycloheximide led to a decrease in the level of Stp1 similar to that caused by rapamycin treatment or nitrogen starvation. Treatment with cycloheximide alone caused a significant decrease in the Stp1 level (Fig. 2C), indicating a short half-life of Stp1 as previously reported (Belle et al., 2006). However, treatment with rapamycin or nitrogen starvation together with cycloheximide caused a further decrease in the level of Stp1 than treatment with cycloheximide alone. Taken together, we conclude that the disappearance of Stp1 from the nucleus as a result of rapamycin treatment or nitrogen starvation results from degradation of Stp1, not from transcriptional or translational downregulation. Moreover, rapamycin treatment did not cause any degradation of Stp1 in fpr1Δ cells (Fig. 2D), indicating that Stp1 degradation specifically results from inhibition of TORC1 by rapamycin.
Interestingly, there was a remarkable difference in the western blot pattern between rapamycin treatment and nitrogen starvation; in case of nitrogen starvation, unprocessed high molecular mass form of Stp1 was detected and appeared to accumulate with time, whereas it was not detected after rapamycin treatment (Fig. 2A). It is already known that a region of about 10 kDa at the N-terminus of Stp1 is endoproteolytically removed by Ssy5 in the SPS sensor when extracellular amino acids are abundant, but Ssy5 loses its activity when amino acids are depleted and then the unprocessed high molecular mass form of Stp1 accumulates in the cytoplasm (Andreasson and Ljungdahl, 2004). We observed that rapamycin treatment under nitrogen starvation had no effect on accumulation of the unprocessed high molecular mass form of Stp1 (Fig. 2E), indicating that Stp1 is synthesized normally during rapamycin treatment. Thus, absence of the unprocessed high molecular mass form of Stp1 after rapamycin treatment alone indicates that, even when TORC1 is inactivated by rapamycin, the SPS sensor is still active and proteolytic processing of Stp1 occurs normally. In other words, processing of Stp1 is not related to the activity of TORC1 but is solely dependent on the presence of extracellular amino acids. Our result also suggests that only the processed form of Stp1, which is localized to the nucleus, is susceptible to degradation when the TOR pathway is inactivated.
We next asked whether degradation of the processed form of Stp1 was regulated by the ubiquitin-proteasome system. To examine this possibility, we used the proteasome inhibitor MG-132 to block proteasome-mediated protein degradation. As shown in Fig. 2F, the processed form of Stp1 increased considerably in cells treated with MG-132 alone, suggesting that the ubiquitin-proteasome pathway is involved in maintaining the steady-state level of Stp1 under normal conditions. However, MG-132 did not prevent degradation of the processed form of Stp1 in rapamycin-treated or nitrogen-starved cells. This result suggests that, as a result of rapamycin treatment or when nitrogen is limiting, degradation of the processed form of Stp1 is mediated by a different mechanism other than the ubiquitin-proteasome pathway.
rrd1Δ cells show rapamycin resistance, which depends on the presence of Stp1
To gain further insight into the mechanism underlying degradation of Stp1 resulting from rapamycin treatment, we sought to identify genes affecting the steady-state level of Stp1. Based on the observation that the level of Stp1 affected rapamycin sensitivity of cells, we assumed that if degradation of Stp1 was mediated by a certain protein during rapamycin treatment, a deletion mutant lacking this protein would show rapamycin resistance as a result of increased steady-state level of Stp1.
Genome-wide functional profiling of rapamycin sensitivity using the yeast deletion library has identified 101 mutant strains showing considerable resistance to rapamycin (Xie et al., 2005). We selected three of these mutants, tip41Δ, rrd1Δ and pbi2Δ, taking into account the degree of rapamycin resistance and known facts about involvement in the TOR pathway. We first introduced deletion mutation of the STP1 gene into each of the three mutants to determine whether rapamycin resistance of the mutants was dependent on the existence of Stp1. Among the three genes, RRD1 showed the strongest genetic interaction with STP1; rapamycin resistance of rrd1Δ cells almost completely disappeared after introduction of stp1Δ mutation. Moreover, rrd1Δ stp1Δ cells showed similar hypersensitivity to rapamycin to that of stp1Δ cells (Fig. 3A). To further investigate the role of Rrd1 in the degradation of Stp1, we asked whether the Stp1 level was maintained in rapamycin-treated rrd1Δ cells. As expected, in rrd1Δ cells treated with rapamycin, Stp1 was very resistant to degradation compared with that of wild-type cells (Fig. 4A). These results suggest that Rrd1 plays an important role in degradation of Stp1 during rapamycin treatment. Presumably rapamycin resistance of rrd1Δ cells is, at least in part, due to an increased steady-state level of Stp1, which may keep transmitting amino acid signals to the inside of the cells and finally make cells partly overcome the growth inhibitory effect of rapamycin. Interestingly, pbi2Δ stp1Δ cells also showed a dramatic increase in rapamycin sensitivity (Fig. 3A). However, pbi2Δ cells treated with rapamycin did not show any difference in the degree of Stp1 degradation compared with wild-type cells (data not shown), indicating that rapamycin resistance of pbi2Δ cells is not related to the activity of Stp1 but is affected by other factors.
The PP2A-like phosphatase Sit4 is important for degradation of Stp1
Rrd1 is a yeast homolog of mammalian phosphotyrosyl phosphatase activator (PTPA) (Rempola et al., 2000). Rrd1 is known as an activator of Sit4, a PP2A-like phosphatase found in yeast. Rrd1 physically interacts with the Tap42-Sit4 complex and Rrd1-Sit4 dimer as a functional phosphatase unit is released from Tap42 by rapamycin treatment (Zheng and Jiang, 2005). Our results showed that Rrd1 is intimately involved in the degradation of Stp1; hence, we then asked whether the sit4Δ mutation would have a similar effect to that of the rrd1Δ mutation regarding degradation of Stp1. As expected, Stp1 in rapamycin-treated sit4Δ cells was also very resistant to degradation (Fig. 4A).
To further confirm the role of Rrd1 and Sit4 in the degradation of Stp1, we analyzed the temperature-sensitive allele of TAP42, tap42-11. It has been reported that the tap42-11 product can bind to Sit4 at 25°C even in the presence of rapamycin, but dissociates from Sit4 at 30°C even in the absence of rapamycin (Cherkasova and Hinnebusch, 2003; Di Como and Arndt, 1996). This indicates that Sit4 in tap42-11 cells is not activated by rapamycin treatment at 25°C, but it is abnormally activated at 30°C even in the absence of rapamycin. We checked whether altered Sit4 activity in tap42-11 cells would lead to a corresponding change in Stp1 degradation. As in sit4Δ cells, rapamycin could not fully induce degradation of Stp1 in tap42-11 cells at 25°C (Fig. 4B). Moreover, upon a temperature shift to 30°C, Stp1 in tap42-11 cells was degraded even without rapamycin treatment (Fig. 4C). Taken together, these results confirmed that upon rapamycin-induced inactivation of TORC1 degradation of Stp1 is mediated by the activity of Rrd1-Sit4 phosphatase unit.
We next analyzed whether the undegraded Stp1 in rrd1Δ and sit4Δ cells after rapamycin treatment was functionally intact for its target gene expression. For analysis, we measured the expression level of a well-known Stp1 target gene AGP1 (Abdel-Sater et al., 2004). As shown in Fig. 4D, rapamycin treatment of rrd1Δ and sit4Δ cells resulted in higher expression levels of AGP1 mRNA than in wild-type cells in both the BY4741 and FY1679-01B background, indicating that the undegraded Stp1 in rapamycin-treated rrd1Δ and sit4Δ cells maintains its activity as a transcription factor for target gene expression. Consistent with this, we found that the undegraded Stp1 in rapamycin-treated rrd1Δ and sit4Δ cells was localized to the nucleus (data not shown).
Intriguingly, unlike rrd1Δ cells, sit4Δ cells did not show any resistance to rapamycin despite an increased steady-state level of Stp1 (Fig. 3B). Sit4 is known to play crucial roles as a negative downstream effector of the TOR pathway (Beck and Hall, 1999; Jacinto et al., 2001), but its function does not seem to be limited to negative regulation of TOR pathway. sit4Δ cells showed growth retardation in rich medium without rapamycin (doubling time of 135 minutes, compared with 96 minutes for wild-type cells; data not shown; see also Fig. 3B), indicating that Sit4 has some important function in normal cell growth other than negative regulation of the TOR pathway. Sit4 has been shown to function in the G1/S transition of the mitotic cycle and modulate functions mediated by Pkc1 including cell wall and actin cytoskeleton organization (Angeles de la Torre-Ruiz et al., 2002; Sutton et al., 1991). Presumably these functions of Sit4 are so critical in cell growth that the growth inhibitory effect of their loss overrides the growth stimulatory effect of increased Stp1 level in rapamycin-treated sit4Δ cells.
There is another yeast PTPA, Rrd2 (Rempola et al., 2000), and it has been shown that Rrd2 has a preferential binding specificity for the catalytic subunits of Tap42-regulated PP2A (PP2Ac), and that Rrd2-PP2Ac is released from Tap42 by rapamycin treatment as Rrd1-Sit4 is (Zheng and Jiang, 2005). We checked whether Rrd2 and PP2Ac, encoded by PPH21 and PPH22, could also play a role in rapamycin-induced Stp1 degradation. However, rrd2Δ, pph21Δ and pph22Δ cells treated with rapamycin did not show any difference in Stp1 degradation compared with wild-type cells (data not shown), indicating that Rrd2-PP2Ac is not involved in the rapamycin-induced degradation of Stp1.
The half-life of Stp1 is known to be short (Belle et al., 2006). Because the steady-state level of Stp1 remained relatively unchanged in rapamycin-treated rrd1Δ and sit4Δ cells, we wondered whether the half-life of Stp1 was increased by the rrd1Δ and sit4Δ mutations. To check this possibility, we measured the level of Stp1 in rrd1Δ and sit4Δ cells treated with cycloheximide and with or without rapamycin. Stp1 was rapidly reduced in wild-type, rrd1Δ and sit4Δ cells treated with cycloheximide alone (Fig. 4E). However, the decrease in rrd1Δ and sit4Δ cells was delayed slightly as compared with that in wild-type cells, suggesting that the basal activity of Rrd1-Sit4 is maintained and partly responsible for the degradation of Stp1 under normal conditions. Retardation of the decrease in the level Stp1 in rrd1Δ and sit4Δ cells was more obvious when treated with cycloheximide together with rapamycin (Fig. 4F). Surprisingly, however, the Stp1 levels were still significantly reduced in rrd1Δ and sit4Δ cells under this condition. Taken together with Fig. 2F and Fig. 4A, this result indicates that the degradation of Stp1 is regulated by at least two distinct mechanisms: the TORC1-independent ubiquitin-proteasome pathway and the TORC1-dependent Rrd1-Sit4 pathway.
It has been shown that rapamycin stimulates eIF2α phosphorylation by Gcn2, with attendant induction of Gcn4 translation. Phosphorylation of eIF2α by Gcn2 is responsible for inhibition of translation initiation by rapamycin (Cherkasova and Hinnebusch, 2003). To determine whether the high steady-state level of Stp1 in rrd1Δ and sit4Δ cells treated with rapamycin might be due to the defect in inhibition of translation initiation, we analyzed the phosphorylation of eIF2α in rrd1Δ and sit4Δ cells. The level of phosphorylated eIF2α in rapamycin-treated rrd1Δ and sit4Δ cells was considerably increased, as it was in rapamycin-treated wild-type cells (data not shown). This result indicates that phosphorylation of eIF2α occurs normally in rrd1Δ and sit4Δ cells, and high steady-state level of Stp1 in rapamycin-treated rrd1Δ and sit4Δ cells is not due to the defect in inhibition of translation initiation.
As Stp1 was degraded as a result of nitrogen starvation as well as rapamycin treatment (Fig. 2A), we looked at whether degradation of Stp1 in nitrogen-starved cells was also mediated by Rrd1 and Sit4. Under nitrogen starvation, the processed form of Stp1 in wild-type cells disappeared in less than 30 minutes (Fig. 4G; see also Fig. 2A). Although the processed form of Stp1 still rapidly disappeared in rrd1Δ and sit4Δ cells during nitrogen starvation, the rate of degradation was slightly slower than in wild-type cells (Fig. 4G), indicating that Rrd1 and Sit4 also participate in degradation of Stp1 during nitrogen starvation. However, we cannot exclude the possibility that another mechanism may also be involved in the degradation of Stp1.
Degradation of Stp1 is related to its phosphorylation state
We observed that in rapamycin-treated rrd1Δ and sit4Δ cells, the processed low molecular mass form of Stp1 was not only resistant to degradation but also migrated slowly on a polyacrylamide gel (Fig. 4A). Based on the fact that Rrd1-Sit4 is a functional phosphatase unit (Zheng and Jiang, 2005), we conjectured that slow migration of Stp1 might be due to its phosphorylation. To determine whether Stp1 was a phosphoprotein, we treated cell extracts from wild-type, rrd1Δ and sit4Δ cells under normal condition with alkaline phosphatase and then analyzed the migration pattern of Stp1 on a polyacrylamide gel. When treated with alkaline phosphatase, the relatively broad Stp1 band from all cell types became sharp and migrated fast (Fig. 5A). Inhibition of alkaline phosphatase activity by the addition of phosphatase inhibitors restored the broad band pattern of Stp1 (Fig. 5A), indicating that Stp1 is a phosphoprotein and is phosphorylated under normal conditions in all wild-type, rrd1Δ and sit4Δ cells. To further confirm the phosphorylation of Stp1, we performed two-dimensional gel electrophoresis. Consistent with the result in Fig. 5A, the Stp1 spot was shifted to the high pH end by treatment with alkaline phosphatase (Fig. 5B). We next asked whether phosphorylation of Stp1 was related to its degradation. To check this, the migration pattern of Stp1 was further analyzed in rapamycin-treated wild-type, rrd1Δ and sit4Δ cells. We found that degradation-resistant Stp1 in rrd1Δ and sit4Δ cells was predominantly present as a phosphorylated slow-migrating form (Fig. 5C; see also Fig. 4A). Taken together, these results suggest that Rrd1-Sit4 is involved in the dephosphorylation of Stp1 during rapamycin treatment and this dephosphorylation of Stp1 may be a prerequisite for its degradation promoted by TORC1 inactivation.
A cyclin-dependent protein kinase, Ssn3, regulates the expression of Stp1 at a post-transcriptional level
We then investigated which kinases are responsible for Stp1 phosphorylation. It has been reported that Stp1 is phosphorylated by casein kinase I (CKI), encoded by YCK1 and YCK2, and this CKI-mediated phosphorylation is essential for endoproteolytic processing of Stp1 (Abdel-Sater et al., 2004). However, there was no significant change in band intensity or migration pattern of the processed Stp1 in yck1Δ and yck2Δ cells compared with wild-type cells (data not shown), suggesting that CKI-mediated phosphorylation is not important for regulation of degradation of the processed Stp1. Recent studies have shown that a cyclin-dependent protein kinase, Ssn3, is involved in phosphorylation and regulation of degradation of some transcription factors with short half-lives, such as Gcn4 and Hac1 (Chi et al., 2001; Pal et al., 2007). To examine the possibility that Ssn3 might be involved in phosphorylation of Stp1, we tested whether the phosphorylation state of Stp1 was altered in ssn3Δ cells. We found that the level of processed Stp1, especially in the phosphorylated form, was considerably reduced in ssn3Δ cells (Fig. 6A). Correlating with this observation, the Stp1-GFP signal was hardly detected in the nucleus of ssn3Δ cells (data not shown).
We next examined whether the low steady-state level of Stp1 in ssn3Δ cells resulted from rapid degradation caused by lack of phosphorylation or downregulation of expression. To determine whether Ssn3 played an opposite role to Rrd1-Sit4 in Stp1 phosphorylation, we performed western blot analysis of Stp1 in rrd1Δ ssn3Δ and sit4Δ ssn3Δ double mutant cells. We assumed that if Ssn3 kinase and the Rrd1-Sit4 phosphatase unit acted on the same site on Stp1, the western blot pattern of Stp1 in rrd1Δ ssn3Δ or sit4Δ ssn3Δ double mutant cells would be similar to that of ssn3Δ cells. However, both of the double mutant cells showed the median levels of Stp1 between those of the single mutants (Fig. 6B). Moreover, in rapamycin-treated rrd1Δ ssn3Δ and sit4Δ ssn3Δ cells, as in rrd1Δ and sit4Δ cells, Stp1 was still mainly present as a phosphorylated slow-migrating form. This result suggests that Ssn3 is not responsible for Stp1 phosphorylation. We then checked whether Stp1 was normally expressed in ssn3Δ cells. STP1 mRNA was not reduced in ssn3Δ cells (data not shown), suggesting that reduced expression of Stp1 in ssn3Δ cells is not due to downregulation of transcription. Remarkably, we found that, when starved of nitrogen, accumulation of the unprocessed high molecular mass form of Stp1 was significantly delayed in ssn3Δ cells compared with wild-type cells (Fig. 6C). This result indicates that low steady-state level of Stp1 in ssn3Δ cells is largely due to reduced expression rather than rapid degradation, and the reduced expression of Stp1 in ssn3Δ cells is regulated at a post-transcriptional level. The mechanism by which Ssn3 mediates post-transcriptional regulation of Stp1 is not clear at present.
In S. cerevisiae, inhibition of the TOR pathway by rapamycin is mediated by a complex of rapamycin and the yeast FKBP12 homolog, Fpr1, and this rapamycin-Fpr1 complex in turn binds to TORC1 to inhibit its activity (Heitman et al., 1991; Koltin et al., 1991). Therefore, rapamycin shows no inhibitory effect on the growth of fpr1Δ cells (Cardenas et al., 1999). It is widely accepted that in eukaryotic cells rapamycin treatment induces similar responses as under condition of nitrogen depletion (De Virgilio and Loewith, 2006; Rohde et al., 2001). However, it is reasonable to assume that the response of cells to nitrogen depletion would be different from that to rapamycin treatment. This is because rapamycin directly exerts its inhibitory effect on TORC1 via Fpr1 but not on the upstream effectors of TORC1, whereas nitrogen depletion shuts down amino acid signaling, which may affect the upstream effectors of TORC1. Stp1, identified in the present study, is a clear example of the differences between response of cells to nitrogen depletion and that to rapamycin treatment. Although the processed low molecular mass form of Stp1 in the nucleus is degraded to a similar degree under both conditions, the unprocessed high molecular mass form of Stp1 is detected only in nitrogen-starved cells (Fig. 2A). This demonstrates that processing of Stp1 normally occurs in rapamycin-treated cells whereas Stp1 is not processed at all in nitrogen-starved cells. This result indicates that processing of Stp1 is governed by the availability of extracellular amino acids irrespective of the TORC1 activity. Taken together with the fact that Stp1 is a direct target of the SPS sensor, and the observation that the expression level of Stp1 affects rapamycin sensitivity of cells, it is likely that Stp1 plays an important role in conveying the extracellular amino acid signal to the rapamycin-sensitive TOR pathway.
Here, we have shown that inactivation of TORC1 leads to degradation of Stp1 in the nucleus, which is dependent on the action of PP2A-like phosphatase Sit4. It is intriguing that Stp1, a transcription factor essential to amino acid signaling, is degraded upon inactivation of TORC1. We hypothesize that there may be a TORC1-dependent regulatory mechanism that controls the activity of the TOR pathway by modulating Stp1-dependent amino acid uptake. If yeast cells growing in a nutrient-rich medium are treated with rapamycin, it will bring about two opposing effects on TORC1. In the presence of rapamycin, TORC1 is inhibited by the rapamycin-Fpr1 complex. However, it can still receive activating signals from extracellular nutrients through its upstream effectors because rapamycin itself does not remove any nutrients from the culture medium nor does it inhibit the activity of a nutrient-sensing system such as Stp1-processing activity of the SPS sensor. To decrease upstream amino acid signals useless to TORC1 under this situation, cells preferentially eliminate Stp1 rather than switch off the SPS sensor. It is plausible that cells keep amino acid sensors turned on for the purpose of sensing and responding quickly to a changeable extracellular environment. We observed that the inactive, high molecular mass form of Stp1 accumulates in the cytoplasm under nitrogen starvation (Fig. 2A). By maintaining a certain amount of unprocessed Stp1 in the cytoplasm and keeping the SPS sensor turned on, yeast cells would be able to respond immediately to extracellular amino acids, when available, and start growing by establishing an amino acid uptake system and activating the TOR pathway. Taken together, our findings suggest a model for the close relationship between the amino acid signaling and the rapamycin-sensitive TOR pathway. In this model, inactivation of TORC1 causes activation of the Rrd1-Sit4 phosphatase unit that in turn promotes degradation of Stp1 and, consequently, leads to a decrease in the intake of amino acids.
It is evident that Stp1 activates the expression of several AAPs, by which amino acids become abundant inside cells. How do the intracellular amino acids activate TORC1? Do intracellular amino acids directly activate TORC1 through allosteric regulation? If not, what effector can recognize the abundance of intracellular amino acids to activate TORC1? We previously found that N-terminally GFP-tagged Tor1 is mainly localized to the vacuolar membrane (our unpublished results). It is known that Tor1 associates with Kog1, Lst8 and Tco89 in TORC1 (Loewith et al., 2002; Reinke et al., 2004). Among the components of TORC1, Kog1 and Tco89 are localized to the vacuolar membrane (Huh et al., 2003). Since the vacuole is a major nutrient reservoir in yeast, it is conceivable that the activity of TORC1 might be closely related to the nutrient level in the vacuole. In accordance with this notion, it has been suggested that the EGO complex in the vacuolar membrane is involved in the TOR pathway in controlling cell growth (Dubouloz et al., 2005). Furthermore, a recent study showed that Sch9 is phosphorylated by TORC1 on the vacuolar membrane (Urban et al., 2007). Therefore, it is plausible that some specific proteins involved in the vacuolar function might be important for regulating the TOR pathway. Future studies focusing on the connection between the vacuolar function and the TOR pathway are expected to provide further insights into the precise regulatory mechanisms of the TOR pathway.
Materials and Methods
Yeast strains, media and reagents
Yeast strains used in this study are listed in Table 2. Rich medium (yeast extract, peptone, glucose; YPD) and synthetic complete (SC) medium lacking appropriate amino acids, for selection, were prepared as previously described (Sherman, 2002). For nitrogen starvation, cells grown to early log phase in SC medium were washed with distilled water and incubated in SD-N medium (0.17% yeast nitrogen base without amino acids and ammonium sulfate, 2% glucose). All cultures were incubated at 30°C. Rapamycin (Tecoland) was dissolved in DMSO and used at appropriate concentrations. Cycloheximide (A.G. Scientific) and MG-132 (A.G. Scientific) were dissolved in DMSO and used at a final concentration of 100 μg/ml and 50 μM, respectively. For MG-132 treatment, cells were cultured in a special synthetic medium (0.17% yeast nitrogen base without amino acids and ammonium sulfate, 0.1% proline, appropriate amino acids, 2% glucose) and treated with MG-132, as previously described (Liu et al., 2007).
Microscopic analysis of protein localization after rapamycin treatment
Exponentially growing cells of GFP-tagged yeast strains (Huh et al., 2003) were treated with 200 ng/ml rapamycin for 2 hours at 30°C in a shaking incubator and then analyzed on 96-well glass-bottom microplates (Whatman) pretreated with concanavalin A (Sigma) to ensure cell adhesion. The cells were then observed using a Zeiss Axiovert 200M inverted microscope, as previously described (Sung and Huh, 2007).
Plasmid and strain construction
For construction of overexpression vectors, appropriate DNA fragments were subcloned into p416GPD vector (Mumberg et al., 1995). Gene disruption was carried out by the PCR-based gene deletion method (Wach, 1996). Construction of chromosomally TAP-tagged strains was performed as previously described (Ghaemmaghami et al., 2003), except that HB0078 (pFA6a-TAP-KlURA) was used as a PCR template. Transformation to introduce a plasmid or DNA segment was performed by the lithium acetate method (Gietz et al., 1992).
Quantification of STP1 and AGP1 mRNA
Total RNA was isolated from yeast cells using the RNeasy MiniKit (Qiagen). cDNA for reverse transcription-PCR was generated using the ProtoScript First Strand cDNA Synthesis Kit (New England Biolabs). The amount of STP1, AGP1 and ACT1 mRNA was analyzed by quantitative real-time reverse transcription-PCR using the Applied Biosystems 7300 Real-Time PCR system. Amplification efficiencies were validated and normalized against ACT1, and fold increases were calculated using the comparative CT method (Livak and Schmittgen, 2001). Values are the mean of three independent experiments and error bars indicate standard deviations.
Gel electrophoresis and western blot analysis
Yeast cell extracts were prepared by suspending mid-logarithmic phase cells in 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% NP-40, protease inhibitors (10 mM phenylmethyl sulfonyl fluoride, 1 mM pepstatin, 1 mM leupeptin, 1 mM benzamide) and phosphatase inhibitors (1 μM sodium orthovanadate, 10 mM β-glycerol phosphate, 10 mM sodium pyrophosphate, 10 mM sodium fluoride), followed by bead beating. Supernatant was recovered by centrifugation at 16,000 × g) for 10 minutes at 4°C and subjected to one- or two-dimensional gel electrophoresis. For two-dimensional gel electrophoresis, cell extracts were separated in the first dimension by isoelectric focusing on ReadyStrip IPG strips (pH 3-6, 7 cm; Bio-Rad Laboratories) using a commercial isoelectric focusing system (PROTEAN IEF Cell; Bio-Rad Laboratories). The focused strips were then run in the second dimension on 10% SDS-polyacrylamide gels. SDS-polyacrylamide gel electrophoresis and western blot analysis were performed by standard methods using HRP-conjugated anti-mouse IgG antibody (Sigma).
Alkaline phosphatase treatment of cell extracts
Crude cell extracts were prepared as described above and treated with alkaline phosphatase (Roche Diagnostics) as previously described (Abdel-Sater et al., 2004). To inhibit alkaline phosphatase activity, phosphatase inhibitors (1 μM sodium orthovanadate, 10 mM β-glycerol phosphate, 10 mM sodium pyrophosphate, 10 mM NaF) were added to the cell extracts.
We thank Alan G. Hinnebusch (NIH, Bethesda, MD) for generously providing yeast strains. This work was supported by grants from the Basic Research Promotion Fund of the Korea Research Foundation (KRF-2006-311-C00537 and KRF-2007-314-E00036) and the SRC/ERC Program of MOST/KOSEF (R11-2005-009-05001-0), Republic of Korea. C.-S.S. and S.Y.K. were supported by the BK21 Research Fellowship from the Ministry of Education, Science and Technology, Republic of Korea.
- Accepted March 16, 2009.
- © The Company of Biologists Limited 2009