Modulation of TOR complex 2 signaling by the stress-activated MAPK pathway in fission yeast

Stress-activated protein kinases (SAPKs) comprise a mitogen-activated protein kinase (MAPK) subfamily that is responsive to environmental stress conditions. Following the discovery of budding yeast MAPK Hog1 as a SAPK prototype (Brewster et al., 1993), identification of SAPKs c-Jun N-terminal kinases (JNKs) and the p38 proteins demonstrated the evolutionary conservation of this stress-responsive subtype of MAPKs (Kyriakis and Avruch, 2001). Spc1 (also known as Sty1, Phh1) in the fission yeast Schizosaccharomyces pombe (Kato et al., 1996; Millar et al., 1995; Shiozaki and Russell, 1995a) is an ortholog of budding yeast Hog1 and mammalian p38, and plays a crucial role in cellular survival of diverse forms of stress, such as high osmolarity, oxidative stress, heat shock and starvation (Nguyen and Shiozaki, 2002; Toone and Jones, 2004). In response to stress stimuli, MAPK Spc1 is phosphorylated by the MAPK kinase (MAPKK) Wis1 (Shiozaki and Russell, 1995b; Warbrick and Fantes, 1991). Activated Spc1 then phosphorylates the transcription factor Atf1 (Gaits et al., 1998; Shiozaki and Russell, 1996; Wilkinson et al., 1996) to induce a set of stress resistance genes (Chen et al., 2003). Among the stress resistance genes regulated by the Spc1–Atf1 pathway is gpd1⁺, which encodes glycerol-3-phosphate dehydrogenase, a key enzyme in glycerol synthesis (Aiba et al., 1995; Degols et al., 1996; Shiozaki and Russell, 1996; Wilkinson et al., 1996). Induced expression of gpd1⁺ results in cellular accumulation of glycerol, which serves as a major cytoplasmic solute to counterbalance extracellular hyperosmotic stress (Ohmiya et al., 1995).

SAPK-interacting protein 1 (Sin1) was isolated by a yeast two-hybrid screen as a protein that interacts with MAPK Spc1, and was proposed to regulate the Spc1-mediated expression of stress resistance genes (Wilkinson et al., 1999). A later examination, however, found that the Spc1-dependent phosphorylation of Atf1 and its function in gene induction upon stress are not affected by the sin1 null (Δsin1) mutation, failing to confirm the functional link between Sin1 and the MAPK Spc1 cascade (Ikeda et al., 2008). A mammalian Sin1 ortholog, SIN1 (also known as MAPKAP1, MIP1), was also reported to interact with SAPK JNKs and their upstream MAPKK kinase, MEKK2 (also known as MAP3K2) (Cheng et al., 2005; Schroder et al., 2005). It was suggested that SIN1 interacts with MEKK2 and prevents its dimerization and activation, leading to suppression of the SAPK cascade (Cheng et al., 2005).

Conversely, Sin1 orthologs in budding yeast and higher eukaryotes have been identified as a component of TOR complex 2 (TORC2), a high molecular mass protein kinase complex that contains target of rapamycin (TOR) kinases as its catalytic subunit (Frias et al., 2006; Jacinto et al., 2006; Lee et al., 2005; Loewith et al., 2002; Wedaman et al., 2003; Yang et al., 2006). The mammalian TORC2 (mTORC2) complex functions as a key activator of a set of the AGC family of protein kinases, such as the AKT proteins, PKCα (also known as PRKCA) and SGK1, through phosphorylation of the ‘hydrophobic motif’ conserved among this kinase family (García-Martínez and Alessi, 2008; Hresko and Mueckler, 2005; Sarbassov et al., 2004; Sarbassov et al., 2005). Also in fission yeast, mass spectrometry analysis of TORC2 identified Sin1 as a TORC2 component (Hayashi et al., 2007). Furthermore, the Δsin1 mutant shares phenotypes with strains lacking the other TORC2 subunits, such as kinase Tor1 and Ste20, a S. pombe ortholog of mammalian RICTOR; the Δsin1, Δtor1 and Δste20 mutants are sterile and exhibit hyper-sensitivity to environmental stresses, including high osmolarity (Ikeda et al., 2008; Kawai et al., 2001; Matsuo et al., 2003; Weisman and Choder, 2001; Wilkinson et al., 1999). In addition, these mutants are defective in phosphorylation and activation of the AGC family kinase Gad8, which is structurally related to mammalian AKT proteins and SGK1 (Ikeda et al., 2008; Matsuo et al., 2003; Tatebe et al., 2010). These biochemical and genetic studies indicated that Sin1 is an essential subunit of TORC2 also in fission yeast and, consistently, a more recent study has demonstrated that Sin1 serves as a substrate-binding subunit of TORC2 (Tatebe et al., 2017). Sin1 specifically binds Gad8 through a domain highly conserved among Sin1 orthologs, thus named the conserved region in the middle (CRIM) domain (Schroder et al., 2007). Nuclear magnetic resonance (NMR) analysis found that CRIM is a ubiquitin-fold domain of ∼120 amino acid residues, and the CRIM domain of human SIN1 also binds specifically to the mTORC2 substrates, such as AKT proteins, PKCα and SGK1 (Furuita et al., 2015; Kataoka et al., 2015; Tatebe et al., 2017). Mutations to CRIM impair TORC2 signaling both in fission yeast and human cells, demonstrating the critical role of Sin1 as a functional subunit of TORC2.

Contrary to the comprehensive characterization of Sin1 as a TORC2 subunit, the physiological significance of its interaction with the SAPK cascade remains unclear in both fission yeast and mammals. In this study, we further characterized the interaction between Sin1 and MAPK Spc1, as well as the role of Spc1 in the regulation of TORC2 signaling. Detailed analysis of the Sin1–Spc1 interaction suggested that the common docking (CD) domain of MAPK Spc1 interacts with a cluster of basic amino acid residues in Sin1. Interestingly, inactivation of MAPK Spc1 results in reduced TORC2 activity, independently of the Spc1–Sin1 interaction. TORC2 is transiently inhibited upon high osmolarity stress and the swift recovery of TORC2 activity following this stress is dependent on the Spc1–Atf1 pathway that induces the glycerol synthesis enzyme Gpd1 for cellular adaptation to osmolarity stress. These results have uncovered coordinated actions of the SAPK and TORC2 pathways, both of which mediate cellular responses to changing environmental conditions.

Full-length MAPK Spc1 was used as bait in the yeast two-hybrid screen that isolated Sin1 (Wilkinson et al., 1999). To further narrow down the Sin1-binding region within Spc1, truncated Spc1 fragments were tested for their interaction with Sin1 in yeast two-hybrid assays. Spc1 N-terminal fragments of 313 residues and 109 residues failed to interact with Sin1 (Fig. 1A), implying that the C-terminus of Spc1 is required. Interestingly, the region C-terminal to the kinase catalytic domain of Spc1 contains the common docking (CD) domain (residues 299–313), a sequence motif conserved among MAPK family members (Tanoue et al., 2000). Because the CD domain is known to mediate interactions of MAPKs with their regulators and substrates, we examined whether mutations to the CD domain affect the interaction of Spc1 with Sin1. Deletion of the CD domain (ΔCD in Fig. 1A), as well as Asn substitutions of the conserved, critical Asp residues within the CD domain (Asp-304 and Asp-307) (Tanoue et al., 2000), abrogated the Spc1–Sin1 interaction (2DN), whereas mutations to the other acidic residues (Glu-308, Asp-312 and Glu-313) did not (DENQ). These results suggest that the CD domain of Spc1 is required for its interaction with Sin1.

Similar yeast two-hybrid assays using a series of Sin1 truncations (Fig. 1B) showed that the N-terminal 2–523 amino acid fragment as well as the C-terminal 509–665 amino acid fragment can interact with Spc1, indicating that residues 509–523 of Sin1 are required to bind Spc1. This region contains a cluster of positively charged amino acids, a known characteristic of the docking sites for the CD domains of MAPKs (Tanoue et al., 2000). Indeed, deletion of the basic stretch (Δ511–523 in Fig. 1B) prevented Sin1 from interacting with Spc1. Moreover, the full-length Sin1 with mutations to the three consecutive Lys residues within this region (residues 513–515; asterisks in Fig. 1B) failed to interact with Spc1 (3KQ in Fig. 1B), though Arg-517, Lys-519 and Lys-520 appeared to be dispensable (RKHQ).

These results suggest that the interaction between Spc1 and Sin1 is mediated by the Spc1 CD domain, with the domain's acidic residues possibly interacting with a cluster of basic residues in Sin1, which resembles MAPK docking sites found in substrates and regulators of the MAPK family members (Tanoue et al., 2000).

Sin1 has been reported as a phosphorylated protein, and its phosphorylation status is reflected by the electrophoretic mobility of the protein (Wilkinson et al., 1999). In SDS-PAGE analysis, the Sin1 protein expressed from its chromosomal locus with a FLAG epitope tag ran as somewhat diffuse bands (Fig. 2A, lane 1), which converged to a fast-migrating band following phosphatase treatment (lane 2). Disruption of the spc1⁺ gene (Δspc1) also resulted in the appearance of a fast-migrating band, together with a slow-migrating band similar to that observed in wild-type cells; thus, some fraction of Sin1 seems to be hypo-phosphorylated in Δspc1 cells (lane 4). We observed no significant change to the electrophoretic mobility of the other TORC2 subunits, such as Tor1, Ste20, Wat1 (also known as Pop3), and Bit61 (Fig. 2B). By contrast, it was noticeable that the amounts of the Sin1, Wat1 and Bit61 proteins somewhat increased in Δspc1 cells when compared to those in wild-type cells.

Disassembly of mammalian TORC2 has been reported as a regulatory mechanism for TORC2 signaling under starvation stress (Chen et al., 2013). In order to examine whether the TORC2 integrity is affected by the hypo-phosphorylation of Sin1 and/or altered levels of the TORC2 subunits in the absence of the stress-responsive MAPK (Fig. 2B), physical interactions among TORC2 subunit proteins were evaluated in the Δspc1 mutant. When tandem affinity purification (TAP)-tagged Tor1 was collected onto IgG beads from the wild-type and Δspc1 strains, no significant difference between the two strains was observed for the co-purification of Sin1 (Fig. 2C), Ste20 (Fig. 2D) or Wat1 (Fig. 2E). Bit61 associates with the Ste20 subunit of TORC2 (Tatebe and Shiozaki, 2010), and their interaction was also not affected by the Δspc1 mutation (Fig. 2F). These observations suggest that Spc1 does not notably affect integrity of the TORC2 complex.

Sin1 functions as a substrate-binding subunit of TORC2 by specifically recruiting Gad8, so that Tor1, the catalytic subunit of TORC2, phosphorylates the C-terminal hydrophobic motif of Gad8 (Tatebe et al., 2017). We found that the TORC2-dependent phosphorylation of Gad8 was significantly reduced in Δspc1 cells, suggesting that Spc1 positively regulates TORC2 activity toward Gad8 (Fig. 3A).

In order to test whether the Spc1–Sin1 interaction is involved in the Spc1-dependent regulation of TORC2 activity, we constructed a fission yeast strain whose chromosomal sin1 gene carries the 3KQ mutation that disrupts Sin1 interaction with Spc1 in the yeast two-hybrid assay (Fig. 1B). No significant difference in Gad8 phosphorylation was detected between the wild-type and sin1-3KQ mutant strains (Fig. 3B). In addition, the electrophoretic mobility of the Sin1-3KQ mutant protein is very similar to that of the wild-type protein both in spc1⁺ and Δspc1 cells (Sin1–FLAG in Fig. 3B). Therefore, the Spc1–Sin1 interaction detectable in yeast two-hybrid assays does not appear to be essential for Spc1-dependent regulation of TORC2 nor for Sin1 phosphorylation.

A previous mass spectrometry analysis of fission yeast TORC2 identified multiple phosphorylation sites in the Sin1 protein (Fig. S1A) (Hayashi et al., 2007). Among those are Ser-62, Ser-301 and Ser-530, all of which are followed by a proline residue, and which can be phosphorylated by MAPKs. We mutated the chromosomal sin1 gene to substitute these serine residues individually with alanine, and then examined Gad8 phosphorylation in these strains, but no significant difference was observed in comparison with that in wild-type strains (Fig. S1B). Other phosphorylation sites that do not match the MAPK phosphorylation site consensus were also mutated to alanine, with no apparent effect on TORC2-dependent phosphorylation (Fig. S1C).

Cellular localization of TORC2 can be visualized by fusing three copies of GFP to Ste20, the fission yeast ortholog of the RICTOR subunit; Ste20–3GFP shows punctate signals throughout the cell surface as well as the cell division septum (Tatebe et al., 2010). Similar cortical localization of TORC2 was observed in Δspc1 cells (Fig. 3C), which are elongated owing to a cell cycle delay (Shiozaki and Russell, 1995a). Δspc1 cells also showed no significant change in the distribution of Gad8 tagged with a single copy of GFP at the C-terminus, and fluorescent signals were detectable throughout the cell except within vacuoles, as in wild-type cells (Fig. 3C). Although it has previously been reported that the majority of Gad8 was detected in the nuclear fraction of the cell lysate (Cohen et al., 2016), we did not observe such nuclear enrichment of untagged, endogenous Gad8 either in wild-type or Δspc1 cells (Fig. S2), consistent with our microscopy results (Fig. 3C).

Taken together, these results indicate that Spc1 positively regulates TORC2 activity, but this regulation is independent of the Spc1–Sin1 interaction and may be rather indirect.

Like the Spc1 cascade (Shiozaki and Russell, 1995a), the TORC2–Gad8 pathway is required for fission yeast cells to grow under high osmolarity stress (osmostress) (Ikeda et al., 2008; Tatebe et al., 2010). In addition, it has been reported that the TORC2-dependent activation of Gad8 is inhibited in response to high osmostress (Cohen et al., 2014). We therefore characterized the kinetics of the TORC2 osmo-response and found that TORC2-dependent phosphorylation of Gad8 disappeared within 5 min of osmostress induced by the addition of 0.6 M KCl, followed by a gradual, somewhat oscillating recovery of phosphorylation after 20 min (pGad8 in Fig. 4A). The prompt inactivation of TORC2 upon osmostress seemed to be correlated to Spc1 activation, which was monitored through its activation loop phosphorylation (pSpc1 in Fig. 4A). Therefore, we examined whether high osmostress inhibits the TORC2–Gad8 pathway through activation of Spc1. Like wild-type cells, Δspc1 mutant cells showed transient attenuation of the Gad8 phosphorylation upon osmostress, though phosphorylation recovery was only slight at later time points (Fig. 4B). A very similar osmo-response of Gad8 phosphorylation was observed in the strain expressing Wis1AA, an inactive mutant of the MAPKK Wis1 that cannot phosphorylate Spc1 (Shiozaki et al., 1998) (Fig. 4C). By contrast, the strain expressing a constitutively active Wis1DD mutant exhibited transient inactivation and recovery of TORC2 activity, as in the wild-type strain (Fig. 4D). These results indicate that the Spc1 cascade is not required for the osmostress-induced inactivation of TORC2, but Spc1 activity promotes the re-activation of TORC2 after osmostress.

Although active Spc1 plays a role in the reactivation of TORC2 after osmostress (Fig. 4), we found that the osmo-response kinetics of Gad8 phosphorylation in the sin1-3KQ mutant is very similar to that in the wild type (Fig. 5A), negating the involvement of the Spc1–Sin1 interaction. As shown in Fig. 5B, the osmoregulation of TORC2 was also not altered in a sin1ΔC strain expressing Sin1 that lacked the pleckstrin homology (PH) domain, which is implicated in the phosphoinositide-dependent regulation of mTORC2 activity (Liu et al., 2015).

In order to explore how Spc1 contributes to the recovery of TORC2 activity after osmostress, Gad8 phosphorylation was monitored in the null mutants of reported Spc1 targets, such as Atf1 (Shiozaki and Russell, 1996), Hal4 (Wang et al., 2005), Cmk2 (Sánchez-Piris et al., 2002), Srk1 (Smith et al., 2002), Lsk1 (Sukegawa et al., 2011), Sds23 (Jang et al., 2013; Yamada et al., 1997) and Wsh3 (also known as Tea4) (Tatebe et al., 2005). In wild-type cells, the TORC2-dependent phosphorylation of Gad8 starts recovering within 30 min after osmostress (Fig. 4A); however, a very reduced re-phosphorylation of Gad8 was observed after 30 min in the Δatf1 strain compared with wild type and the other null mutants tested (Fig. 5C; Fig. S3A).

Being phosphorylated and activated by Spc1, transcription factor Atf1 induces expression of a set of the genes important for cellular adaptation to stressful conditions (Chen et al., 2003; Shiozaki and Russell, 1996; Wilkinson et al., 1996). We tested some of the genes under the regulation of Spc1–Atf1 for their involvement in the recovery of TORC2 activity after osmostress. It was found that a strain lacking gpd1⁺ failed to induce significant re-phosphorylation of Gad8 even after 100 min under osmostress (Fig. 5D; Fig. S3B). gpd1⁺ encodes glycerol-3-phosphate dehydrogenase, a protein involved in biosynthesis of glycerol that is important for cellular adaptation to high osmolarity (Ohmiya et al., 1995). Conversely, such a defect was not observed in a strain lacking gpd2⁺, a gpd1⁺ paralog with no apparent role in cellular osmo-resistance (Yamada et al., 1996) (Fig. S3B). In addition, the gene disruption of pmk1⁺ and slm1⁺, which encode a stress-responsive MAPK (Cohen et al., 2014; Madrid et al., 2016) and an ortholog of the Slm proteins in S. cerevisiae (Riggi et al., 2018), respectively, did not affect the Gad8 phosphorylation during high osmostress (Fig. S4).

These results suggest that Spc1 promotes reactivation of TORC2 after osmostress, through Atf1, which induces expression of gpd1⁺. Indeed, the Δspc1 Δatf1 and Δspc1 Δgpd1 double mutants showed defects similar to the respective single mutants, consistent with the idea that Spc1, Atf1 and Gpd1 function together during the recovery of TORC2 inactivated by osmostress (Fig. 6A,B). However, in normal osmolarity media, the Δgpd1 mutation did not affect Gad8 phosphorylation (Fig. 6C), indicating that the positive regulation of TORC2 by Spc1 in the absence of osmostress (Fig. 3A) is not dependent on gpd1⁺. By contrast, as in Δspc1 cells, Gad8 phosphorylation was reduced in Δatf1 cells (Fig. 6C). Unexpectedly, the Δspc1 and Δatf1 mutations appeared to be additive, with Gad8 phosphorylation in the Δspc1 Δatf1 double mutant significantly lower than in the respective single mutants. It is likely that, under normal growth conditions without osmostress, Spc1 and Atf1 independently affect TORC2 activity.

Genetic analysis in fission yeast demonstrated that the Spc1 cascade and the TORC2–Gad8 pathway are both required for cellular adaptation to high osmostress, though the stress elicits opposite responses to these two signaling pathways; activation of the Spc1 cascade and inhibition of the TORC2 pathway (Cohen et al., 2014; Ikeda et al., 2008; Millar et al., 1995; Shiozaki and Russell, 1995b). Because Sin1 was identified as a SAPK-interacting protein (Wilkinson et al., 1999) and also as a TORC2 subunit (Hayashi et al., 2007; Matsuo et al., 2007), Sin1 seems to be a candidate molecule that links Spc1 to TORC2 in cellular stress response. Having found that TORC2 does not affect Spc1 signaling (Ikeda et al., 2008), in this study we pursued the opposing question of whether Spc1 modulates TORC2 signaling.

We successfully reproduced the previously reported interaction between Spc1 and Sin1 in the yeast two-hybrid assay (Wilkinson et al., 1999), and further showed that the CD domain of Spc1 and a cluster of basic residues in Sin1 are involved in the interaction. The specificity of this Spc1–Sin1 interaction was further corroborated by a reciprocal yeast two-hybrid screen of a S. pombe cDNA library using a C-terminal Sin1 fragment of residues 401–665 as bait. A short, C-terminal Spc1 fragment that includes the CD domain (residues 304–349) was identified in this screen (data not shown). Thus, the interaction of Spc1 with Sin1 may be similar to the interactions of other MAPKs with their substrates and regulators (Tanoue et al., 2000). By contrast, we failed in our attempt to detect the Spc1–Sin1 interaction by co-purification assays (data not shown), and the mass spectrometry analysis of fission yeast TORC2 detected Sin1, but not Spc1 (Hayashi et al., 2007). The interaction between Spc1 and Sin1 may not be stable enough for these biochemical approaches.

We found that mutational inactivation of Spc1 results in compromised TORC2-dependent phosphorylation of Gad8, indicating that Spc1 positively regulates the TORC2–Gad8 pathway. However, TORC2 activity is not altered by the sin1-3KQ mutation that disrupts the interaction of Sin1 with Spc1 and thus, the Sin1–Spc1 interaction is not required for the observed Spc1-dependent regulation of the TORC2 pathway. In addition, the loss of Spc1 has no apparent impact on TORC2 integrity nor on the cellular localization of TORC2 and its substrate Gad8. These observations imply a rather circuitous regulatory mechanism by which Spc1 positively regulates TORC2–Gad8 signaling.

We found that TORC2 is inhibited upon high osmostress in a manner independent of the stress-induced activation of Spc1. Although Pmk1, another stress-responsive MAPK in fission yeast, is implicated in the negative regulation of the TORC2–Gad8 pathway (Cohen et al., 2014; Madrid et al., 2016), we found that Pmk1 is not required for the osmo-inhibition of TORC2 signaling (Fig. S4A). A recent study in budding yeast proposed that decreased plasma membrane tension under high osmolarity induces clustering of phosphatidylinositol-4,5-bisphosphate [PI(4,5)P₂], to which TORC2 is tethered as clumps segregated from its activators Slm1 and Slm2 (Riggi et al., 2018). Whereas high osmostress induces prominent clustering of PI(4,5)P₂ also in fission yeast (Kabeche et al., 2015), the Slm1/2 ortholog in fission yeast has no apparent role in TORC2 activation both in the presence and absence of osmostress (Fig. S4B). In addition, Slm orthologs are not found in mammals, where inactivation of TORC2 signaling upon high osmostress is also observable (Meier et al., 1998). Thus, the underlying mechanisms of the osmostress sensitivity of TORC2 may be different from species to species.

There may also be a difference between budding yeast and fission yeast in the process of TORC2 reactivation after osmostress. Whereas SAPK Hog1 is not important for the TORC2 reactivation in S. cerevisiae (Riggi et al., 2018), we found that Spc1 contributes to the recovery of TORC2 activity after osmostress through Atf1 and its target gene, gpd1⁺. This observation is probably not surprising, considering the essential role of the Spc1–Atf1 pathway in cellular adaptation to hyperosmolarity through the induction of the glycerol synthesis enzyme Gpd1 (Degols et al., 1996; Gaits et al., 1998; Ohmiya et al., 1999, 1995; Shiozaki and Russell, 1996; Wilkinson et al., 1996). Unexpectedly, however, our study also uncovered a role of Atf1, but not Gpd1, in TORC2 activation under normal growth conditions (Fig. 6C). TORC2 activity is severely compromised in the Δspc1 Δatf1 double mutant, suggesting that Atf1 contributes to TORC2 activity independently of Spc1, most likely through expression of unknown target genes.

In summary, the data presented in this paper shed light on the intertwining relationship between Spc1 and TORC2, both of which play critical roles in osmostress resistance of fission yeast cells. The Spc1–Atf1 pathway positively regulates TORC2 signaling both in the presence and absence of osmostress, independently of the Spc1–Sin1 interaction detectable in the yeast two-hybrid assay (Fig. S4C). Although both Hog1 and TORC2 in budding yeast are involved in the regulation of cellular glycerol accumulation during osmostress (Lee et al., 2012; Muir et al., 2015), crosstalk between the two signaling modules has not been reported. Further genetic studies in S. pombe and S. cerevisiae may unravel a novel mode of interaction between the SAPK and TORC2 pathways that are highly conserved among diverse eukaryotes.

Schizosaccharomyces pombe strains used in this study are listed in Table S1. Growth media and basic techniques for fission yeast were previously described (Alfa et al., 1993). Epitope tagging of chromosomal genes was carried out by the PCR-based method (Bähler et al., 1998). Site-directed mutagenesis was performed using the PrimeSTAR Max DNA polymerase (Takara Bio) according to the supplier's manual. Oligo DNAs for PCR are listed in Table S2. Stress treatment of S. pombe cells was carried out as previously described (Shiozaki and Russell, 1997). For high osmolarity treatment, one-third volume of pre-warmed medium containing 2.4 M KCl was added to the culture. Protein concentrations were determined using Protein Assay Reagent (Bio-Rad Laboratories).

Yeast two-hybrid assay was performed as previously described (Tatebe et al., 2005). The ORF encoding the inactive form of Spc1 (Spc1-T171A) was subcloned in the bait plasmid, pGBT9 (Clontech Laboratories) using NdeI and PstI sites, as ectopic expression of active Spc1 causes a growth defect. The complementary DNA of Sin1 (nucleotides 3 to 1998) was subcloned in the prey plasmid, pGAD GH (Clontech Laboratories) using BamHI and ApaI sites. Plasmids used in the assay are listed in Table S3. HF7c budding yeast strain (Clontech Laboratories) was used as host. Interaction was judged by histidine auxotrophy. Cells harboring either or both empty vector(s) were used as negative controls.

Co-purification of epitope-tagged proteins was performed (Morigasaki and Shiozaki, 2010) using buffers as described below. Lysis buffer containing 1×PBS, 10% (w/v) glycerol, 0.25% (w/v) Tween20, 10 mM NaF, 10 mM sodium pyrophosphate, 10 mM NaN₃, 10 mM beta-glycerophosphate 2Na, 10 mM p-nitrophenylphosphate 2Na, 1 mM PMSF, and 1:200-volume protease inhibitor cocktail (P8849, Sigma-Aldrich). The lysis buffer without protease inhibitors was used as washing buffer. Protein bound to beads was eluted with the Laemmli sample buffer without 2-mercaptoethanol for 15 min at room temperature. After removing the beads, the eluate was mixed with 1:19-volume of 2-mercaptoethanol and heated at 65°C for 15 min. IgG-Sepharose 6 Fast Flow (GE Healthcare) and EZview Anti-c-Myc Affinity Gel (Sigma-Aldrich) were used for precipitation of NTAP–Tor1 and Bit61–myc, respectively.

Whole-cell protein extract was prepared by trichloroacetic acid (TCA) extraction. Yeast cells in early log phase (OD₆₀₀=0.4, 25 ml) were harvested on a 0.4 µm-porosity filter membrane and resuspended in 200 µl of 10% (w/v) TCA solution. Cells were disrupted by beating with glass beads (ø=0.5 mm) at 2500 rpm for 4.5 min (30 s on and 30 s off, 9 cycles) using the Multibeads shocker (Yasui Kikai Co.). After removing glass beads, the cell homogenate was centrifuged for 10 min at 9000 g at room temperature, and the precipitate was resuspended in 200 µl of the Laemmli sample buffer containing 0.5M Tris-HCl, pH 8.0. The sample was then heated at 65°C for 15 min and centrifuged for 10 min at 9000 g at room temperature to remove cell debris. The supernatant was used as ‘TCA extract’ in mobility shift assays (see below). The protein concentration of the TCA extract was adjusted to 1 mg protein/ml with the standard Laemmli sample buffer.

For the Sin1 mobility shift assay, the TCA extract of Sin1–FLAG-expressing cells was subjected to SDS-PAGE using 6.5% T/2.67% C polyacrylamide gel [% T, total monomer concentration (g/100 ml); % C, weight percentage of crosslinker]. Sin1 was detected by immunoblotting using anti-FLAG antibodies. Phosphatase treatment was performed according to Tatebe et al. (2008) with some modification. Briefly, 10 µg protein of the TCA extract was diluted 180 times with the lambda-protein phosphatase (PPase) buffer (New England BioLabs). The diluted extract was dispensed into three tubes (A, B, and C). One-tenth volume of the buffer, 60 units of PPase in the buffer, or 60 units of PPase+10× phosphatase inhibitor mix in the buffer were added to dilution A, B or C, respectively. After mixing gently, the reaction mixtures were incubated at 30°C for 30 min. To stop the reaction, 1:7 volume of 100% (w/v) TCA was added. Protein was precipitated by centrifugation at 18,700 g for 10 min at 4°C, after standing on ice for 30 min. The precipitate was then resuspended in 20 µl of the Laemmli sample buffer containing 0.5 M Tris-HCl, pH 8.0 and heated at 65°C for 15 min. The 10× phosphatase inhibitor mix is composed of 20 mM Na₃VO₄, 100 mM NaF, 100 mM EDTA, 100 mM beta-glycerophosphate, 40 mM p-nitrophenylphosphate.

The activating phosphorylation of Thr171 and Tyr173 in Spc1 (pSpc1), Spc1, phosphorylation of Ser546 in Gad8 (pGad8), and the Gad8 protein were detected by immunoblotting using rabbit polyclonal antisera at 1:100,000, 1:15,000, 1:2000 and 1:5000 dilution, respectively (Tatebe et al., 2010; Tatebe and Shiozaki, 2003). Anti-histone H2B antiserum (1:5000) was a gift from Dr M. Yanagida (Maruyama et al., 2006). Rps6 was detected with anti-RPS6 antibody (1:1500; ab40820, Abcam plc.). For detection of FLAG-, HA- and Myc-tagged proteins, anti-FLAG (1:8000; M2, Sigma-Aldrich), anti-HA (1:2000; 12CA5, Roche Diagnostics), and anti-c-myc (1:2000; 9E10, Covance) mouse monoclonal antibodies were used, respectively. NTAP–Tor1 was detected with anti-calmodulin binding protein epitope tag (1:2000; Cat. no. 07-482, Merck Millipore). Anti-rabbit IgG (H+L) HRP-conjugated (1:2000; Cat. no. W4011, Promega) or anti-mouse IgG (H+L) HRP-conjugated (1:2000; Cat. no. W4021, Promega) were used as secondary antibodies.

In immunoblotting, Pierce ECL Plus Western Blotting Substrate (Thermo Fisher Scientific) was used for detection. The image of chemiluminescence was obtained using the imaging analyzer LAS4000 (GE Healthcare) and the signal intensity was measured with the software Multi Gauge 3.0 (Fujifilm). For quantification, the signal intensity of phospho-Gad8 (pGad8) was compensated by that of the Gad8 protein.

Cells were cultured in EMM medium (Alfa et al., 1993) until reaching early log phase in the dark and mounted on a thin layer of EMM+agar. Fluorescence images were taken with DeltaVision Elite Microscopy System (GE Healthcare) as described previously (Chia et al., 2017; Tatebe et al., 2010).

As reported by Cohen et al. (2016) and Keogh et al. (2006), the nucleus-rich fraction was prepared from S. pombe cells: 972 h- (PR37) and Δspc1 (KS1616). An aliquot of spheroplast was used as whole cell extract (WCE). After fractionation using 1.2 M sucrose cushion, the upper layer and pellet were collected as cytoplasmic (Cyt) and nucleus-rich (Nuc) fractions, respectively. Gad8 in each fraction was analyzed by immunoblotting using antiserum against Gad8. In addition, distribution of Rps6 and histone H2B (H2B) were analyzed as markers of cytosol and nucleus, respectively.

We thank Dr M. Yanagida (Okinawa Institute of Science and Technology Graduate University, Onna-son, Japan), Dr T. Toda (Hiroshima University, Higashi-Hiroshima, Japan) and Dr J. Quinn (Newcastle University, Newcastle, UK) for their generous gifts of antibodies and fission yeast strains, and K. H. Chia (Nara Institute of Science and Technology, Ikoma, Japan) for his help in microscopy.