Roles of phosphatidylserine and phospholipase C in the activation of TOR complex 2 signaling in Saccharomyces cerevisiae

Target of rapamycin (TOR) is a Ser/Thr protein kinase that controls cell growth and metabolism and is highly conserved among eukaryotes (Heitman et al., 1991; Saxton and Sabatini, 2017; Tafur et al., 2020; Tatebe and Shiozaki, 2017). TOR forms two distinct complexes, TORC1 and TORC2 (Tafur et al., 2020). TORC1 is a rapamycin-sensitive form that regulates translation initiation, ribosome biogenesis, transcription, autophagy, nutrient uptake and cellular senescence (Saxton and Sabatini, 2017; Morozumi and Shiozaki, 2021). TORC2 is a rapamycin-insensitive form that regulates the actin cytoskeleton, biosynthesis of sphingolipids, endocytosis and signal transduction (Riggi et al., 2020). Unlike most eukaryotes, which possess only a single TOR homolog, budding and fission yeasts have two TOR proteins, Tor1 and Tor2 (Heitman et al., 1991; Jacinto and Lorberg, 2008). TORC1 and TORC2 are well known to target the AGC (named initially for cAMP- and cGMP-dependent protein kinases and protein kinase C) kinases, including Akt, PKC and SGK1, and phosphorylate Thr/Ser residues within the turn motif and hydrophobic motif in these substrates (Jacinto and Lorberg, 2008).

In the budding yeast Saccharomyces cerevisiae, TORC1 phosphorylates Sch9 and Ypk3, and TORC2 phosphorylates Ypk1 and Ypk2 (hereafter referred to as Ypk1/2) and Pkc1 (Kamada et al., 2005; Urban et al., 2007; Nomura and Inoue, 2015; González et al., 2015), all of which belong to the AGC kinase family. TORC1 mainly localizes to the vacuolar membrane (Betz and Hall, 2013), and several regulators of TORC1–Sch9 signaling have been identified, including the Rag GTPases Gtr1, Gtr2 and Pib2 (Morozumi and Shiozaki, 2021). A recent study reported spatial regulation of TORC1 signaling (Hatakeyama et al., 2019). Conversely, TORC2 localizes at the domain just beneath the plasma membrane, which is known as the membrane compartment containing TORC2 (MCT) (Sturgill et al., 2008; Berchtold and Walther, 2009; Riggi et al., 2020). TORC2 regulates sphingolipid biosynthesis via Ypk1 and Ypk2 (Muir et al., 2014; Riggi et al., 2020; Tabuchi et al., 2006; Aronova et al., 2008; Roelants et al., 2011), the orthologs of mammalian SGK1 (Casamayor et al., 1999). It has been reported that membrane stress, caused by the inhibition of sphingolipid biosynthesis, or membrane stretch causes the activation of TORC2–Ypk1/2 signaling via plasma membrane-associated proteins, Slm1 and Slm2, which ensure the interaction between TORC2 and Ypk1/2 (Berchtold et al., 2012; Niles et al., 2012; Riggi et al., 2020). A recent study has reported that an invaginated plasma membrane structure in which phosphatidylinositol 4,5-bisphosphate [PtdIns(4,5)P₂] accumulates [called the PtdIns(4,5)P₂-enriched structure; PES] inactivates TORC2–Ypk1/2 signaling when the plasma membrane tension is decreased (Riggi et al., 2018).

Methylglyoxal (MG) is a natural metabolite derived from glycolysis. Since MG is a highly reactive carbonyl compound, it forms advanced glycation end products (Phillips and Thornalley, 1993; Inoue et al., 2011). MG is also associated with the regulation of intracellular signaling pathways from yeast to mammals, suggesting that MG functions as a signaling molecule that activates signal transduction (Inoue et al., 2011; Zemva et al., 2019). We previously showed that TORC2 phosphorylates Pkc1, an ortholog of mammalian protein kinase Cs (PKCs), and that MG activates TORC2–Pkc1 signaling, thereby modulating the Mpk1 mitogen-activated protein kinase (MAPK) cascade (Nomura and Inoue, 2015). Activation of the Mpk1 MAPK cascade via Pkc1 is induced by heat shock and cell wall stress (Kamada et al., 1995; de Nobel et al., 2000). Heat shock stress activates Pkc1 signaling via the membrane proteins Wsc1 and Mid2, the heat shock stress sensors (Kamada et al., 1995; Philip and Levin, 2001), although these sensor proteins are dispensable for MG-induced activation of TORC2–Pkc1 signaling (Nomura and Inoue, 2015). Correct activation of TORC2–Pkc1 signaling requires the conserved region 1 (C1) domain of Pkc1, which physically interacts with the membrane-associated small GTPase Rho1 (Nomura and Inoue, 2015). The binding of the C1 domain to lipids such as diacylglycerol (DAG), is known to activate some of mammalian PKC isozymes (Rosse et al., 2010). However, DAG did not activate yeast Pkc1 and was unable to bind to its C1 domain in an in vitro lipid overlay assay (Antonsson et al., 1994; Watanabe et al., 1994; Nomura et al., 2017b; Heinisch and Rodicio, 2018). Instead, the C1 domain of Pkc1 is bound by phosphatidylserine (PS), which ensures the interaction between Pkc1 and Rho1, along with the correct localization of Pkc1 at the bud (Nomura et al., 2017b; Heinisch and Rodicio, 2018). This suggests that Rho1 is necessary for Pkc1 to be a target of TORC2. However, the regulatory mechanisms of TORC2 in both TORC2–Pkc1 and TORC2–Ypk1/2 signaling are poorly understood when compared to those of TORC1 signaling.

Phosphatidylinositol-specific phospholipase C (PLC) hydrolyzes PtdIns(4,5)P₂ to DAG and inositol 1,4,5-triphosphate, and is involved in ubiquitous phosphoinositide-based regulatory systems, further contributing to a myriad of biological processes in eukaryotic cells (Suh et al., 2008). There are at least 13 different PLC isozymes in mammals, whereas S. cerevisiae only has phosphoinositide-specific PLC, encoded by PLC1 (Flick and Thorner, 1993; Payne and Fitzgerald-Hayes, 1993; Yoko-o et al., 1993). Plc1 is involved in various cellular processes, including cell cycle control (Flick and Thorner, 1998), activation of plasma membrane H⁺-ATPase (Coccetti et al., 1998), gene expression (Guha et al., 2007), histone acetylation (Galdieri et al., 2013) and kinetochore function (Lin et al., 2000), and hence, plc1Δ mutant exhibits slower growth (Flick and Thorner, 1993; Yoko-o et al., 1993). The catalytic activity of Plc1 is activated under hypoosmotic stress conditions, where the levels of PtdIns(4,5)P₂ at the plasma membrane are decreased within 2 min (Perera et al., 2004). In a genetic study of TORC2, multicopy expression of PLC1 suppressed the growth defect of tor2-21^ts cells, the temperature-sensitive TORC2 mutant, at non-permissive temperatures (Helliwell et al., 1998a). It has been also reported that Plc1 physically interacts with Tor2 (Lin et al., 1998), but molecular mechanisms underlying the relationship between Plc1 and TORC2 remains to be elucidated.

In the present study, we showed that MG activates TORC2–Ypk1/2 signaling as well as TORC2–Pkc1 signaling, and that PS is involved in the activation of TORC2–Ypk1/2 signaling following treatment with MG or inhibition of sphingolipid biosynthesis. Particularly, we found that the contribution of PS in the activation of TORC2–Ypk1/2 signaling under inhibitory conditions of sphingolipid biosynthesis is mediated via Cdc42. We also found that Plc1 modulates the activation of both TORC2–Ypk1/2 and TORC2–Pkc1 signaling and is crucial for the correct localization of Pkc1, thereby allowing TORC2-Pkc1 signaling activation. Our results provide insights into the mechanism by which PS and Plc1 contribute to the activation of TORC2 signaling.

Several genetic studies have indicated that TORC2 is involved in the cell wall integrity pathway, which consists of Pkc1 and the Mpk1 MAPK cascade, one of the downstream effectors of Pkc1 (Schmidt et al., 1997; Helliwell et al., 1998a,b; Ho et al., 2005). We previously found that TORC2 phosphorylated Pkc1 directly (Nomura and Inoue, 2015). The glycolytic metabolite MG activated TORC2–Pkc1 signaling, leading to increase in the phosphorylation of the Mpk1 MAPK (Nomura and Inoue, 2015). In addition to Pkc1, TORC2 targets Ypk1 and Ypk2, which are orthologs of mammalian SGK1 (Casamayor et al., 1999), and phosphorylates the hydrophobic motif (Thr662 and Thr659 in Ypk1 and Ypk2, respectively) of Ypk1/2 (Kamada et al., 2005; Jacinto and Lorberg, 2008). TORC2–Ypk1/2 signaling regulates sphingolipid biosynthesis and is activated by mechanical membrane stress and inhibition of sphingolipid metabolism (e.g. mediated by inhibitors including myriocin and aureobasidin A, AbA) (Berchtold et al., 2012; Muir et al., 2014; Riggi et al., 2020). Previously, we found that mutants lacking CSG2 and ISC1, both involved in sphingolipid biosynthesis, showed MG sensitivity (Nomura et al., 2017a), implying that MG is associated with the regulation of TORC2–Pkc1 and TORC2–Ypk1/2 signaling. Therefore, we determined whether MG activates TORC2–Ypk1/2 signaling by monitoring the phosphorylation levels of Ypk1/2. A previous report demonstrated that in cells treated with AbA, a phospho-specific antibody for the hydrophobic motif of Ypk1/2 could detect the increase in its phosphorylation (Ishino et al., 2022). Similar to what was seen with AbA, MG treatment led to enhanced phosphorylation of Ypk1/2 (Fig. 1A), suggesting that MG induced the activation of TORC2–Ypk1/2 signaling. Immunoblots of Ypk1 showed multiple bands when whole-cell extracts were resolved using SDS-PAGE, which was attributed to the flippase kinases Fpk1 and/or Fpk2 (Roelants et al., 2010). The inhibition of sphingolipid biosynthesis by AbA or myriocin treatment is known to decrease the activities of Fpk1 and Fpk2, thereby reducing the Ypk1 mobility shift (Roelants et al., 2010). In contrast to what was seen AbA treatment, the mobility shift of the Ypk1 bands detected by the Ypk1 antibody was enhanced in the case of MG treatment (Fig. 1A). We confirmed that the band shift of Ypk1 was dependent on Fpk1 (Fig. S1A).

TORC2 consists of Tor2, Avo1, Avo2, Avo3, Bit61 and Lst8. Tor2, Avo1 and Avo3 are essential components for TORC2 function (Tafur et al., 2020). MG-induced phosphorylation of Ypk1/2 was diminished in avo3-30 mutant cells, a temperature-sensitive mutant of AVO3 (Aronova et al., 2008), at non-permissive temperatures (Fig. 1B). TORC2–Ypk1/2 signaling-deficient mutants show increased susceptibility to AbA and myriocin (Roelants et al., 2004; Aronova et al., 2008; Muir et al., 2014). As shown in Fig. 1C, ypk1Δ and ypk1-1^tsypk2Δ cells were highly susceptible to AbA and MG. These results indicate that MG activates TORC2–Ypk1/2 signaling in addition to TORC2–Pkc1 signaling.

TORC2 localizes in a punctate pattern beneath the cell membrane (Sturgill et al., 2008; Berchtold and Walther, 2009; Berchtold et al., 2012). We examined the effect of MG on the TORC2 localization by assessing the position of Avo3–3GFP. The punctate pattern of Avo3–3GFP was not significantly altered following treatment with MG (Fig. S1B).

We previously reported that PS is necessary for the correct localization of Pkc1 at the bud tip to allow MG-induced activation of TORC2–Pkc1 signaling, and that the deletion of PS synthase encoded by CHO1 enhanced the susceptibility to MG (Nomura et al., 2017b). The CHO1 mutant showed increased susceptibility to myriocin (Hatakeyama et al., 2017). We found that cho1Δ cells were also sensitive to AbA (Fig. 2A), implying that PS modulates TORC2–Ypk1/2 signaling. To examine whether PS is involved in the activation of TORC2–Ypk1/2 signaling, we verified the effect of CHO1 deletion on the phosphorylation levels of Ypk1/2 following treatment with MG or AbA. The increase in phosphorylation levels of Ypk1/2 following MG or AbA treatment was significantly attenuated in cho1Δ cells (Fig. 2B,C). In addition, overexpression of the C2 domain of lactadherin (Lact-C2), which specifically binds to PS (Yeung et al., 2008) to compete with PS interactors, reduced the phosphorylation of Ypk1/2 in cells treated with MG or AbA (Fig. 2D,E). These results suggest that PS contributes to the activation of TORC2–Ypk1/2 signaling.

In addition to functioning as an enzyme cofactor, PS plays an important role in cellular processes, such as the establishment of cell polarity, vesicle transport in the secretory pathway and modulation of peripheral proteins interacting with the membrane (Lenoir et al., 2021). Previous reports have indicated a role for PS in the localization of Rho family GTPase Cdc42, which is essential for the establishment of cell polarity, in budding and fission yeasts (Fairn et al., 2011; Haupt and Minc, 2017). To examine the involvement of Cdc42 in the activation of TORC2–Ypk1/2 signaling, we investigated the phosphorylation levels of Ypk1/2 in a cdc42 mutant following treatment with MG or AbA. Since CDC42 is an essential gene, we used a temperature-sensitive cdc42 mutant, cdc42-13 (Iwase et al., 2006). As shown in Fig. 3A,B, AbA-induced phosphorylation of Ypk1/2 was significantly decreased in cdc42-13 cells at non-permissive temperatures, although MG-induced phosphorylation was comparable to that in wild-type cells. The cdc42-13 cells also showed a decrease in phosphorylation of Ypk1/2 post myriocin treatment (Fig. S2A). Additionally, a defect in AbA-induced phosphorylation of Ypk1/2 was observed in another cdc42 mutant, cdc42-1 (Fig. S2B). Rdi1 is a Rho guanine nucleotide dissociation inhibitor (GDI) that negatively regulates Cdc42 by extracting it from the membrane (Tiedje et al., 2008). PS has been reported to be involved in preventing the extraction of Cdc42 by Rdi1 (Das et al., 2012). Overexpression of RDI1 under the GAL1 promoter abolished the AbA-induced phosphorylation of Ypk1/2 but not in case of MG treatment, similar to the findings with cdc42 mutants (Fig. 3C). Rdi1 also interacts with the Rho-family GTPases Rho1 and Rho4 (Tiedje et al., 2008), although the mutation of RHO1 (Qadota et al., 1996) or RHO4 did not significantly decrease the AbA-induced phosphorylation of Ypk1/2 (Fig. S2C,D). These results suggest that Cdc42 is necessary for the activation of TORC2–Ypk1/2 signaling by AbA and myriocin, but not by MG, implying that the mechanism underlying the activation of TORC2–Ypk1/2 signaling induced by MG is different from that of the inhibition of sphingolipid metabolism (Fig. 3D).

To explore the activation mechanism of TORC2–Ypk1/2 signaling, we focused on phosphoinositide-specific PLC. Genetic and physical interactions between PLC1 and TORC2 have been reported, where PLC1 has been identified as a multicopy suppressor of the growth defect of tor2-21^ts, the temperature-sensitive TORC2 mutant (Helliwell et al., 1998a), and Plc1 physically interacts with Tor2 (Lin et al., 1998). Despite the possibility that Plc1 contributes to TORC2 signaling, its role in TORC2 signaling remains unknown. As a first approach to determine the role of Plc1 in TORC2 signaling, we determined its susceptibility to stressors that activates TORC2 signaling, such as MG and AbA. Because Plc1 is involved in a variety of biological processes in yeast, plc1Δ cells show a slow-growth phenotype (Flick and Thorner, 1993; Yoko-o et al., 1993). As shown Fig. 4A,B, plc1Δ cells grew more slowly than wild-type cells under non-stressed conditions, and the plc1Δ mutant exhibited the enhanced sensitivity to MG and AbA. Enhanced susceptibility caused by PLC1 deficiency was also observed with myriocin (Fig. S3A). The increased phosphorylation levels of Ypk1/2 following treatment with MG or AbA in plc1Δ cells were significantly repressed compared to those in wild-type cells (Fig. 4C,D; Fig. S3B). The repression was also observed in a lipase-inactive mutants of PLC1, namely PLC1^H393A/N394A (Jun et al., 2004) (Fig. 4E,F) and PLC1^G781D (Yoko-o et al., 1995) (Fig. S3C). The overexpression of PLC1 has been shown to suppress the temperature sensitivity of tor2-21^ts mutant, which is specifically defective in TORC2 function (Helliwell et al., 1998a). We next examined whether the lipase activity of Plc1 is required for this suppression. As shown in Fig. 4G, multicopy expression of neither PLC1^H393A/N394A nor the PLC1^G781D mutant could suppress the growth defect of tor2-21^ts cells at non-permissive temperatures. The expression levels of Plc1^H393A/N394A were comparable with those of Plc1 (Fig. S3D) and similar to those reported for Plc1^G781D (Yoko-o et al., 1995). These results suggest that Plc1 contributes to the activation of TORC2–Ypk1/2 signaling, for which the lipase activity is required.

Plc1 hydrolyzes PtdIns(4,5)P₂, which is abundant in the plasma membrane (Flick and Thorner, 1993; Payne and Fitzgerald-Hayes, 1993; Yoko-o et al., 1993). We examined the cellular distribution of PtdIns(4,5)P₂ in the plasma membrane of plc1Δ cells using the previously described PtdIns(4,5)P₂ sensor GFP–2xPH^PLCδ (Stefan et al., 2002). The fluorescence of GFP–2xPH^PLCδ in wild-type cells was observed in the plasma membrane and weakly in the cytosol, as previously reported (Fig. 4H). The distribution of GFP–2xPH^PLCδ in plc1Δ cells was essentially the same as that in the wild-type cells, indicating that PLC1 deletion did not affect the cellular distribution of PtdIns(4,5)P₂ (Fig. 4H; Fig. S4A,B). It has been reported that the emergence of PESs, the PtdIns(4,5)P₂-enriched structures, at the plasma membrane inhibits TORC2–Ypk1/2 signaling when the plasma membrane tension is decreased (Riggi et al., 2018). PESs were observed after the addition of 1 M sorbitol, which led to a hyperosmotic shock (Riggi et al., 2018) (Fig. 4H). The emergence of PESs under hyperosmotic shock conditions in plc1Δ cells was comparable to that in the wild-type cells (Fig. 4H). These findings suggest that the mechanism by which PLC1 deletion represses the increased phosphorylation levels of Ypk1/2 is not mediated by decreased plasma membrane tension.

We next examined whether Plc1 is also involved in the MG-induced activation of TORC2–Pkc1 signaling. MG activates TORC2–Pkc1 signaling through increased phosphorylation of Ser1143 within the hydrophobic motif of Pkc1 (Nomura and Inoue, 2015). Mutants defective in TORC2-targeted sites of Pkc1, as well as a pkc1Δ mutant, exhibit susceptibility to MG (Nomura and Inoue, 2015). To verify whether Plc1 is involved in MG-induced activation of TORC2–Pkc1 signaling, the phosphorylation levels of Ser1143 in Pkc1 were determined following treatment with MG in plc1Δ cells. As shown in Fig. 5A,B, MG-induced phosphorylation of Pkc1 at Ser1143 was significantly repressed in plc1Δ cells. Because Plc1 is involved in a variety of biological processes in yeast, plc1Δ cells showed a slow-growth phenotype even under non-stressed conditions, although the impaired growth of plc1Δ cells in the presence or absence of MG was rescued by overexpression of PKC1 (Fig. 5C). The rescue effect of PKC1 overexpression on the phenotype of plc1Δ cells is known to be abolished by alanine residue substitutions in the TORC2 phosphorylation sites in Pkc1 (Nomura and Inoue, 2015). We next examined whether Plc1 is involved in TORC2 activity by performing an in vitro TORC2 kinase assay using the Pkc1 peptide (Nomura and Inoue, 2015) (Fig. S5A). The activity of TORC2 purified from plc1Δ cells decreased as compared with that from wild-type cells, suggesting that Plc1 contributes to TORC2 activity. Additionally, the Plc1^H393A/N394A mutant also showed a decrease in TORC2 activity (Fig. S5B).

MG-induced activation of TORC2–Pkc1 signaling causes an increase in phosphorylation of the Mpk1 MAPK, and the increased phosphorylation of Mpk1 was not observed in tor2-21^ts mutants grown at non-permissive temperatures (Nomura and Inoue, 2015). To verify the role of Plc1 in the MG-induced activation of TORC2–Pkc1 signaling, we determined the phosphorylation levels of Mpk1 and found that the deletion of PLC1 inhibited the phosphorylation of Mpk1 following MG treatment (Fig. 5D,E). Meanwhile, the increase in levels of phosphorylated Mpk1 by heat shock, which activates Pkc1–Mpk1 signaling via the membrane-integrated sensor proteins Wsc1 and Mid2 (Kamada et al., 1995; Philip and Levin, 2001), was not significantly inhibited in plc1Δ cells (Fig. 5F,G). We have previously suggested that Mpk1 phosphorylation induced by heat shock does not require the activation of TORC2–Pkc1 signaling (Nomura and Inoue, 2015). The phosphorylation levels of Ser1143 in Pkc1 were not increased under heat shock conditions (Fig. S5C), indicating that heat shock does not activate TORC2–Pkc1 signaling and that Plc1 is dispensable for activation of Pkc1–Mpk1 signaling induced by heat shock. Additionally, the increase in the levels of phosphorylated Mpk1 in response to treatment with zymolyase, which causes cell wall damage, was also not affected by the deletion of PLC1 (Fig. S4D). These results suggest that Plc1 is necessary for MG-induced activation of TORC2–Pkc1 signaling.

Microscopy studies using GFP-tagged Pkc1 have shown that Pkc1 is localized at the plasma membrane, bud tip and bud neck (Denis and Cyert, 2005; Nomura et al., 2017b). Proper membrane localization of Pkc1 requires its membrane-targeting module, that is, the C1 domain (Denis and Cyert, 2005; Nomura et al., 2017b; Heinisch and Rodicio, 2018). The C1 domain is also necessary for the interaction of Pkc1 with the small GTPase Rho1, which activates Pkc1 (Nonaka et al., 1995; Nomura et al., 2017b). A mutation in the C1 domain abolishes the phosphorylation of Ser1143 in Pkc1 by TORC2 following treatment with MG (Nomura and Inoue, 2015; Nomura et al., 2017b). Therefore, proper membrane localization of Pkc1 is needed to activate TORC2–Pkc1 signaling. In wild-type cells, the proportion of cells with bud tip localization of Pkc1–3GFP, a genome-integrated triple tandem GFP-tagged PKC1 construct, was found to be more than 50% of the cells counted, and that with bud neck-located Pkc1–3GFP was ∼20% (Nomura et al., 2017b) (Fig. 6A,B); however, the proportion of cells with bud tip-located Pkc1–3GFP in plc1Δ cells was reduced to less than half of that in wild-type cells (Fig. 6A,B). No significant differences were observed in the proportion of cells with bud neck-located Pkc1–3GFP between the wild-type and plc1Δ cells (Fig. 6A,B). Additionally, a reduction in the proportion of bud tip-located Pkc1–3GFP was observed in the PLC1^H393A/N394A mutant (Fig. 6C). The protein levels of Pkc1–3GFP in plc1Δ cells were comparable to those in the wild-type cells (Fig. S6). These findings suggest that Plc1 participates in MG-induced activation of TORC2–Pkc1 signaling by ensuring the proper localization of Pkc1, which is required for the activation of this signaling pathway (Fig. 6D).

TORC2 signaling is associated with various cellular functions involving cell growth and homeostasis, such as sphingolipid biosynthesis, the polarity of actin cytoskeleton and the cellular stress response (Riggi et al., 2020). However, the lack of specific inhibitors for TORC2 and insufficient information on the initiators or modulators of TORC2 signaling have hindered the understanding of its regulatory mechanisms. In the present study, we demonstrated that MG functions as an initiator of both TORC2–Ypk1/2 and TORC2–Pkc1 signaling, and that PS and Plc1 are involved in the activation of these signaling pathways (Fig. 7). The levels of phosphorylation of Ypk1/2 induced by MG were comparable to those induced by AbA (Fig. 1A), and mutants defective in Ypk1 and/or Ypk2 and genes encoding enzymes involved in sphingolipid biosynthesis showed higher sensitivity to MG (Fig. 1C) (Nomura et al., 2017a). These findings suggest that MG negatively affects the machinery involved in sphingolipid biosynthesis, which activates TORC2–Ypk1/2 signaling. Meanwhile, Fpk1 and/or Fpk2-dependent phosphorylation of Ypk1, which can be detected by the shift in electrophoretic mobility, is regulated by sphingolipid levels (Roelants et al., 2010). Therefore, treatment of yeast cells with AbA, which leads to a decrease in sphingolipid levels, represses the band shift in SDS-PAGE gels (Roelants et al., 2010). However, MG treatment had the opposite effect in terms of mobility shift (Fig. 1A). Unlike AbA, MG does not appear to have a negative effect on sphingolipid biosynthesis. Additionally, it has been reported that treatment of yeast cells with myriocin, an inhibitor of sphingolipid biosynthesis, which is known to activate TORC2–Ypk1/2 signaling (Berchtold et al., 2012; Muir et al., 2014; Riggi et al., 2020), decreases the phosphorylation levels of Mpk1 (Olson et al., 2015). We confirmed this using AbA, another inhibitor of sphingolipid biosynthesis (data not shown). If MG inhibits sphingolipid biosynthesis, it is likely that Mpk1 is negatively regulated; however, MG enhances the phosphorylation levels of Mpk1 through TORC2–Pkc1 signaling (Nomura and Inoue, 2015). Collectively, the activation of TORC2–Ypk1/2 signaling following MG treatment is most likely not due to the inhibition of sphingolipid biosynthesis. Conversely, we revealed that Cdc42 was a factor modulating the AbA- and myriocin-induced activation of TORC2–Ypk1/2 signaling; however, it was dispensable for MG-induced activation of TORC2–Ypk1/2 signaling (Fig. 3). This result supports the conclusion that the activation of TORC2–Ypk1/2 signaling by MG and inhibition of sphingolipid biosynthesis are distinct mechanisms. Given that MG activates both TORC2–Ypk1/2 and TORC2–Pkc1 signaling, MG might act on the regulation of TORC2 itself involving PS; however, we have previously shown that MG does not affect the kinase activity of TORC2 in an in vitro kinase assay (Nomura and Inoue, 2015). Involvement of Ras and Rho family GTPases in Dictyostelium (Khanna et al., 2016; Charest et al., 2010; Cai et al., 2010; Senoo et al., 2019) and the Rab family GTPase in fission yeast (Tatebe et al., 2010) and budding yeast (Locke and Thorner, 2019) has been reported in the regulation of TORC2 signaling. Although we have not clarified how Cdc42 is involved in the regulation of TORC2–Ypk1/2 signaling at present, we showed experimental evidence demonstrating that this Rho family GTPase also seems to contribute to the regulation of TORC2 signaling in budding yeast.

Ypk1 has been reported to negatively regulate Fpk1 and/or Fpk2, thereby inhibiting the aminophospholipid flippase that controls the subcellular distribution of PS, which influences the localization of Rho1 and actin organization (Roelants et al., 2010; Hatakeyama et al., 2017). Meanwhile, our findings also showed that PS modulates the activation of TORC2–Ypk1/2 signaling, indicating that PS, whose distribution is regulated downstream of Ypk1/2, also contributes to Ypk1/2 activation. A previous study has shown that a combination of deletion mutations in metabolic enzymes involved in sphingolipid biosynthesis and repression of CHO1 causes a synthetic growth defect and abnormal endosomal trafficking, although the decreased levels of complex sphingolipids in such deletion mutants under steady-state conditions were not significantly affected by the repression of CHO1 (Tani and Kuge, 2012). These findings imply that a feedback control system exists in the regulatory mechanism of TORC2–Ypk1/2 signaling activation under stress conditions, in which PS might act as a modulator. A mutation in CHO1 affects transmembrane- and plasma membrane-localizing proteins that are regulated by PS, including Cdc42 (Lenoir et al., 2021). A recent study has been shown that a stable large protein-depleted membrane area, which is referred to as the ‘void zone’, is detected in the plasma membrane when cho1 mutant cells are exposed to high temperatures in S. cerevisiae (Mioka et al., 2022). Although we revealed the involvement of Cdc42 on the activation of TORC2–Ypk1/2 signaling, it is also possible that changes in membrane environment associated with PS depletion might affect TORC2–Ypk1/2 signaling.

The turnover of PtdIns(4,5)P₂ has been shown to be associated with TORC2 signaling. A multicopy expression of MSS4, which encodes PtdIns(4)P 5-kinase, or a deletion of INP51, which encodes PtdIns(4,5)P₂ 5-phosphatase, leads to increased levels of PtdIns(4,5)P₂, thereby suppressing the growth defect of tor2-21^ts cells at non-permissive temperatures (Helliwell et al., 1998a; Morales-Johansson et al., 2004). These observations indicate that PtdIns(4,5)P₂ has a positive effect on TORC2. Nevertheless, a multicopy expression of PLC1, the gene product of which hydrolyzes PtdIns(4,5)P₂, has also been shown to suppress the growth defect of tor2-21^ts cells (Helliwell et al., 1998a). Additionally, we found that the deletion of PLC1 attenuated the activation of TORC2 signaling (Fig. 4D). These results prompted us to evaluate the relationship between PtdIns (4,5)P₂ levels and TORC2 signaling. However, the steady-state levels of PtdIns(4,5)P₂ in plc1Δ cells are not higher than those in wild-type cells (Perera et al., 2004), suggesting that the contribution of Plc1 to the turnover of PtdIns(4,5)P₂ under steady-state conditions is small. Previously, a physical interaction between Tor2 (not TORC2) and Plc1 have been reported; namely, that GST protein containing a part of Tor2 (1693–1946 aa) physically interacts with Plc1 in vitro (Lin et al., 1998). This observation raises the possibility that Plc1 directly regulates TORC2; however, we were unable to confirm the physical interaction between Plc1 and TORC2 by immunoprecipitation in the experimental conditions as used with our yeast strain. The physical interaction between Plc1 and Tor2 may be strain specific, or such an interaction might be involved in an aspect other than TORC2 function. Meanwhile, the lipase inactive mutation as well as the deletion of PLC1 reduced TORC2 activity (Fig. S5A,B), suggesting that the lipase activity of Plc1 is engaged in the TORC2 activity; however, the precise mechanism underlying how Plc1 contributes to the activation of TORC2 signaling is not clear at this stage. PtdIns(4,5)P₂ is enriched in the plasma membrane. The C-terminal domain of Avo1 has been reported to contain a pleckstrin homology (PH) domain-like structure (Berchtold and Walther, 2009; Pan and Matsuura, 2012), and can bind to PtdIns(4,5)P₂in vitro (Berchtold and Walther, 2009). In addition, Slm1 and Slm2 also possesses the PH domain, and Seg1, one of the components of eisosomes, which are involved in the regulation of TORC2–Ypk1/2 signaling, also has an affinity for PtdIns(4,5)P₂ (Moreira et al., 2012). Thus, Plc1 might contribute to the activation of TORC2 signaling by participating in local PtdIns(4,5)P₂ turnover in the plasma membrane, which might affect these proteins.

It is well known that mammalian TORC2 (mTORC2) is activated in response to insulin and growth factors, and regulates downstream effectors such as protein kinase B (Akt), which belong to the same AGC kinase family as Ypk1/2 and Pkc1. Recent studies have shown that plasma membrane tension modulates the activation of mTORC2 signaling (Diz-Muñoz et al., 2016; Roffay et al., 2021). It is also known that PLC is involved in response to both hyper- and hypo-osmotic stresses in mammalian cells (Shah et al., 2014; Qifti et al., 2021). Thus, understanding the role of Plc1 in yeast TORC2 signaling might aid in understanding the regulatory mechanisms underlying TORC2 signaling that are conserved between yeast and mammals.

YPD (2% glucose, 1% yeast extract, and 2% peptone), synthetic dextrose (SD; 2% glucose, 0.67% yeast nitrogen base without amino acids) or raffinose medium (2% raffinose, 0.67% yeast nitrogen base without amino acids) was used for cell culture. Appropriate amino acids and bases were added, as necessary. Methylglyoxal, myriocin, and ethanolamine were purchased from Sigma-Aldrich (St Louis, MO, USA). Aureobasidin A was obtained from TaKaRa Bio, Inc. (Otsu, Japan). Materials used in this study are listed in Table S3.

The yeast strains that were used are listed in Table S1. To construct the plc1Δ mutant, the plc1Δ::HIS3 allele of YJF32 (Flick and Thorner, 1993) was amplified via PCR using the primers: 5′-CCCTTATCACAGTTACTTTCACCAAGAGAA-3′ and 5′-TTTGATGGTGATATCATCTTTCAGTGTCTC-3′. The product was then introduced into YPH250. The PKC1–3GFP allele was introduced into the PKC1 locus using the integration plasmid Pkc1-3GFP (Nomura et al., 2017b).

The plasmids used in this study are summarized in Table S2. The plasmid YEplac181-pGAL1-RDI1 was constructed as follows: the open-reading frame (ORF) of RDI1 and the 3′ untranslated region of RDI1 were amplified using the following primers: 5′-AACAGATCTATGGCCGAAGAAAGTACCGACTTTAG-3′ and 5′-CTGGTCGACAATTATACAAGGGTAAGAAGTACCA-3′. The PCR product was digested with BglII and SalI, and the resulting fragment was introduced into the BamHI and SalI sites of YCpLG, a plasmid for protein expression under the GAL1 promoter (YCpLG-RDI1). The pGAL1-RDI1 fragment obtained by digestion of YCpLG-RDI1 with EcoRI and SalI was cloned into the EcoRI and SalI sites of YEplac181.

The plasmid pRS316-PLC1^G781D was constructed using a KOD -Plus- Mutagenesis Kit (TOYOBO) with pRS316-PLC1 (Jun et al., 2004) as a template. pRS316-PLC1, pRS316-PLC1^H393A/N394A, or pRS316-PLC1^G781D were digested with BamHI, and each resulting fragment was introduced into the BamHI site of pRS315 or pRS426 to generate pRS315-PLC1, pRS315-PLC1^H393A/N394A, pRS426-PLC1, pRS426-PLC1^H393A/N394A or pRS426-PLC1^G781D, respectively.

In order to construct the integration plasmid Avo3–3GFP, the C-terminal of AVO3 was amplified with the following primers: 5′-GTTCTCGAGGCAACTATACGACCTGAGCCC-3′ and 5′-TTTCTGCAGCCACGTGTAAAATTAGCCGGC-3′. The PCR product was digested with XhoI and PstI, and the resulting fragment was introduced into the XhoI and PstI sites of PB1994 (Buttery et al., 2007). The plasmid Avo3–3GFP was digested with StuI, and linearized DNA was introduced at the locus of AVO3 to replace the gene containing a 3GFP tag.

The plasmids pRS424-3HA-PLC1 and pRS424-3HA-PLC1^H393A/N394A were constructed as follows: the fragment containing the 5′ untranslated region of PLC1 and the 3HA in pRS403-3xHA-PLC1 (Jun et al., 2004) was amplified using the following primers: 5′-AAGAGCTCCAGACTCCTTCATTACATCTCG-3′ and 5′-GTCATCTATAGCACTTTCAGTCATTCTAGA-3′. The PCR product was digested with SacI and XbaI, and the resulting fragment was introduced into the SacI and SpeI sites of pRS424, which was named pRS424-pPLC1-3HA. The ORF of PLC1 and the 3′ untranslated region of PLC1 in pRS316-PLC1 or pRS316-PLC1^H393A/N394A (Jun et al., 2004) was amplified using the following primers: 5′-TGTCTGCAGATGACTGAAAGTGCTATAGAT-3′ and 5′-TCTCTCGAGGGATCCTAATTCAGTAATGCT-3′. Each PCR product was digested with PstI and XhoI, and the resulting fragment was introduced into the PstI and XhoI sites of pRS424-pPLC1-3HA to generate pRS424-3HA-PLC1 or pRS424-3HA-PLC1^H393A/N394A.

The cells were cultured until an absorbance at 610 nm (A₆₁₀) of 0.3–0.5 was obtained. Cells harvested by centrifugation (1750 g for 1 min) were resuspended in 200 µl of 20% trichloroacetic acid with glass beads, and disrupted with a Micro Smash MS-100 (TOMY). Supernatants were collected and glass beads were rinsed with 400 µl of 5% trichloroacetic acid (the final concentration of trichloroacetic acid in the sample was 10%), and then the samples were incubated for 15 min more on ice. Samples were centrifuged at 106,200 g for 10 min at 4°C, and pellets were rinsed with ice-cold acetone. Following centrifugation, pellets were dried and resuspended with 2× Laemmli sample buffer. The total protein concentration was determined using the RC DC^TM Protein Assay (Bio-Rad Laboratories). The samples were subjected to SDS-PAGE, and the separated proteins were transferred onto a polyvinylidene difluoride membrane (Millipore). Anti-phosphorylated Ypk1^T662 and Ypk2^T659 (1:25,000, Ishino et al., 2022), anti-Ypk1 (1:3000, sc12054; Santa Cruz Biotechnology), anti-phospho-p44/42 MAPK (1:5000, #9101; Cell Signaling Technology), anti-Mpk1 (1:3000, sc6803; Santa Cruz Biotechnology), anti-Myc (1:8000, M192-3; MBL), anti-Pgk1 (1:10,000, #A6457; Molecular Probes), anti-HA (1:8000, M180-3; MBL), and anti-phosphorylated Pkc1^S1143 (1:3000, Nomura and Inoue, 2015) were used as primary antibodies. Immunoreactive bands were visualized with Immobilon Western chemiluminescent horseradish peroxidase substrate (Millipore) using a Fuji Las 4000 Mini System (Fujifilm). The intensity of immunoreactive bands was quantified using ImageJ. The source data for the quantitative analysis are in Table S4. Uncropped blots of western blotting images are provided in Fig. S7.

The cells were cultured in SD medium until an A₆₁₀ of 0.5–1.0 was obtained. Cells were then collected (1750 g for 2 min), washed with a 0.85% NaCl solution, and suspended in lysis buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.1 mM EDTA, 0.5% Tween 20) containing protease inhibitor cocktail (Nacalai Tesque) and phosphatase inhibitor cocktail (Nacalai Tesque). Cells were disrupted with glass beads using Micro Smash MS-100, and cell homogenates were centrifuged at 700 g for 10 min at 4°C to remove cell debris. Protein concentrations in the cell extracts were determined using the DC protein assay (Bio-Rad Laboratories). The Pkc1–3HA or Avo3–13myc protein was immunoprecipitated from the cell extracts (1.5 or 2 mg protein) by incubation with anti-HA antibodies coupled with agarose resin (MBL) or anti-Myc-tag mAb-magnetic beads (MBL) for 2 h at 4°C in lysis buffer. After incubation, agarose resins or magnetic beads were precipitated by centrifugation (700 g for 1 min) or magnet, washed four times with lysis buffer, and then suspended in SDS-PAGE sample buffer or used for the in vitro kinase assay. SDS-PAGE was performed, followed by western blotting.

Cells were cultured in YPD or SD medium until the early log phase of growth and then diluted to an A₆₁₀ value of 0.1 using sterilized 0.85% NaCl solution. The cells were serially diluted (1:10) with sterilized 0.85% NaCl solution and aliquots (5 µl) were spotted onto YPD or SD agar plates.

Microscopy analysis using a micro slide glass was performed with the fluorescence microscope BX51 equipped with an Olympus DP70 digital camera (OLYMPUS, Tokyo, Japan) and BX63 (OLYMPUS, Tokyo, Japan) equipped with a digital camera ORCA-Fusion (Hamamatsu Photonics, Shizuoka, Japan).

In an in vitro kinase assay performed as described previously (Nomura and Inoue, 2015). Briefly, TORC2 was immunopurified using Myc-tagged Avo3 (Avo3–myc) from cell extracts of wild-type (YPH250) and plc1Δ cells with an AVO3-13myc integration in the AVO3 locus. Immunopurified TORC2 was incubated with synthetic Pkc1 peptides (APPTLTPLPSVLTTSQQEEFRGFSFMPDDL) in buffer (25 mM Hepes-KOH pH 7.4, 10 mM MnCl₂ and 10 mM p-nitrophenylphosphate). The reaction was initiated by adding ATP (100 µM). After being incubated for 30 min at 30°C, the reaction was terminated by the addition of SDS-PAGE sample buffer, and the samples were then incubated for 5 min at 65°C. Samples were subjected to Tricine-SDS-PAGE (Schägger, 2006) or SDS-PAGE, and phosphorylated peptides were detected by western blotting using anti-phosphorylated Pkc1^S1143 antibodies (Nomura and Inoue, 2015).

Data are presented as mean±s.d. Statistical significance of differences was evaluated using one-way ANOVA with Tukey's multiple comparison test. Differences were considered significant at P<0.05.

We are grateful to Drs E. Bi, S. D. Emr, M. N. Hall. Y. Ohya, D. Pellman, T. Powers, J. Thorner, W. Wickner, and the National Bio-Resource Project (NBRP) of MEXT, Japan, for providing plasmids and yeast strains. We thank K. Ikeda for providing the technical support.