Yeast Hog1 proteins are sequestered in stress granules during high-temperature stress

Mitogen-activated protein kinases (MAPKs), signaling proteins that are conserved from yeasts to human, are serine-threonine protein kinases activated by extracellular stimuli, such as cytokines, growth factors, hormones and various environmental stresses (Brewster et al., 1993; Herskowitz, 1995; Horvitz and Sternberg, 1991). MAPKs are catalytically inactive in their baseline forms, and phosphorylation of the activation loop is required for activation. When cells sense certain stimuli, small GTPases are activated that trigger downstream signaling pathways, including the MAPK cascade. Dephosphorylation of MAPKs by MAPK phosphatase leads to their inactivation (Cobb and Goldsmith, 1995).

In the baker's yeast Saccharomyces cerevisiae (indicated in constructs by the prefix Sc) Hog1 (ScHog1), the ortholog of mammalian MAPK p38 (MAPK14), plays a central role as a signaling mediator during osmoregulation (Brewster and Gustin, 2014). Activated Hog1 is localized to the nucleus and upregulates nearly 600 genes by phosphorylating osmoresponsive transcription factors, such as Msn2 and Msn4, Sko1, Hot1, and Smp1 (Hohmann, 2002), resulting in expression of chaperones and other general stress-response genes (Morano et al., 2012). One major role of the Hog1 pathway is elevation of the intracellular levels of glycerol – which acts as an osmolyte – by inducing glycerol synthesis and accumulation. Loss of Hog1 activity results in abnormal cell morphology (Brewster et al., 1993) and reduced cell growth on high-osmolarity medium. In addition to high osmolarity, Hog1 is partially activated by high-temperature stress through the Sho1 cascade. Inactivation of the Hog1 phosphatases Ptp2 or Ptp3 results in sensitivity to high temperatures (Winkler et al., 2002).

Conditions of high-temperature induce several heat stress-responsive pathways, including the cell wall integrity and Hog1 pathways, together with the formation of two intracellular granule structures, i.e. processing bodies (P-bodies) and stress granules (SGs) (Brengues et al., 2005; Kedersha et al., 2005), which contain non-translating mRNPs (messenger ribonucleoproteins). These granules harbor mRNA decay machinery, and various factors involved in translational impairment and transient storage of mRNAs and, thus, play important roles in quality control of cytosolic mRNAs (Borbolis and Syntichaki, 2015). In addition to these roles, roles of SGs have been recently expanded to include regulation of intracellular signaling pathway by sequestering signaling molecules under stress conditions. For example, signaling through the target of rapamycin complex 1 (TORC1) pathway, which coordinates nutrient availability, was shown to be modulated by sequestration of TORC1 in SGs under conditions of high temperature in S. cerevisiae (Takahara and Maeda, 2012). In the fission yeast Schizosaccharomyces pombe, calcinurin, the Ca²⁺/calmodulin-dependent protein phosphatase, localized to SGs in response to heat shock and negatively regulated the assembly of SGs (Higa et al., 2015). As for regulation of the MAPK pathway, RACK1, the scaffold protein binding to MTK1 MAPKKK in mammalian cells, was sequestered in SGs under certain types of stress, such as hypoxia, heat shock and arsenite, to suppress MTK1-mediated signaling pathways (Arimoto et al., 2008).

The methylotrophic yeast Candida boidinii survives and proliferates on growing plant leaves, assimilating methanol and nitrate for its survival (Kawaguchi et al., 2011; Shiraishi et al., 2015). Therefore, C. boidinii needs to respond and adapt to a wide variety of stresses and environmental changes, including nutrient limitation, high flux of UV radiation, and changes in temperature and osmolarity. In this study, we have investigated the intracellular dynamics of C. boidinii (indicated in constructs by the prefix Cb) Hog1 (CbHog1) by using a fusion protein of CbHog1 and the YFP Venus (CbHog1-Venus). Under conditions of high temperature, CbHog1-Venus formed intracellular dot structures that colocalized with the SG marker polyadenylate-binding protein 1 (Pab1) of C. boidinii (CbPab1) fused to Venus (CbPab1-Venus). The N-terminal region of CbHog1 was necessary and sufficient for recruitment of CbHog1 to SGs. Further analyses revealed that intracellular sequestration of Hog1 under high-temperature conditions is involved in downregulating the activity of phosphorylated Hog1 in vivo.

We searched the draft genome sequences of C. boidinii, and found an 1194-bp open reading frame (ORF) encoding a protein of 398 amino acids (aa), i.e. CbHog1, sharing a high degree of identity with ScHog1 (71%) as well as other yeast Hog1 proteins, Pichia pastoris (indicated in constructs by the prefix Pp) Hog1 (PpHog1; 78%) and Schizosaccharomyces pombe (indicated in constructs by the prefix Sp) Sty1 (SpSty1; 72%) (Fig. S1A). Based on aa alignment of CbHog1 with ScHog1 (Fig. S1A), we found that CbHog1 contained key conserved aa, including the K50 residue necessary for kinase activity and two phosphorylation sites, T172 and Y174 (Murakami et al., 2008). When compared with Hog1 proteins from other yeast species, ScHog1 had a non-conserved extended region of 37–87 aa at its C-terminus.

The CbHOG1 gene was disrupted by replacing the ORF with the CbURA3 gene as a selective marker. In S. cerevisiae, Hog1 is localized to the nucleus under hyperosmotic stress, and the ScHOG1 deletion (Schog1Δ) strain exhibits hyper-osmosensitivity (Brewster et al., 1993). We expressed CbHog1-Venus under its own promoter in the CbHOG1 deletion (Cbhog1Δ) strain. CbHog1-Venus was localized to the nucleus 5 min after a shift to medium containing 0.5 M NaCl (Fig. S1B). Stress resistance analysis was also performed under conditions of high osmolarity in wild type, Cbhog1Δ, and CbHog1-Venus expressing Cbhog1Δ strains. The Cbhog1Δ strain exhibited severe growth defects on YPD medium containing 0.5 M NaCl, and CbHog1-Venus completely complemented the growth defect of the Cbhog1Δ strain under 0.5 M NaCl (Fig. S1C). These results indicate that, similar to ScHog1, CbHog1 functions in osmotic stress tolerance in C. boidinii.

To examine the intracellular dynamics of CbHog1-Venus under high-temperature stress, C. boidinii cells expressing CbHog1-Venus were incubated at different temperatures (34–45°C) for 30 min, and then observed under a fluorescent microscope. C. boidinii was unable to proliferate at temperatures above 35.5°C (Fig. S2A), and C. boidinii cells incubated at high temperatures contained intracellular dot structures of CbHog1-Venus in the cytosol (Fig. 1A). Morphometric analysis revealed a positive correlation between the incubation temperature and the number of CbHog1-Venus dots from 35.5–42°C (Fig. 1B).

Nuclear localization of CbHog1-Venus under high-temperature condition was examined with cell samples incubated at 42°C for 30 min. Cells stained with DAPI were microscopically observed and showed that significant localization of CbHog1-Venus to the nucleus was not observed in C. boidinii at 42°C, whereas a few dots of CbHog1-Venus were associated with the blue fluorescence of DAPI (Fig. S3A). This indicated that nuclear localization of CbHog1 was not a major event in C. boidinii under high-temperature conditions.

Interestingly, we observed that PpHog1 fused to YFP (PpHog1-YFP) in P. pastoris and SpSty1-GFP in S. pombe formed similar dots under high-temperature stress, whereas ScHog1 fused to GFP (ScHog1-GFP) in S. cerevisiae did not (Fig. 1C).

To examine whether this dot formation was due to irreversible aggregation of the protein, we monitored CbHog1-Venus fluorescence after reducing the high temperature (42°C) again. As shown in Fig. 1D and E, CbHog1-Venus dots gradually disappeared after the shift from 42°C to 28°C, and very few cells harbored CbHog1-Venus dots 60 min after release from heat. These findings indicated that dot formation of CbHog1-Venus is a reversible process, rather than a consequence of irreversible aggregation of the protein.

SGs are cytosolic complexes composed of proteins and RNAs that appear in response to external stresses, such as heat shock and severe ethanol stresses (Buchan et al., 2011; Kato et al., 2011).

First, we cloned the DNAs coding for SG marker proteins in C. boidinii. A search of the C. boidinii draft genome sequence revealed putative SG proteins, including the poly-A-binding protein Pab1 and the Pab1-binding protein Pbp1 (Table S1). As in S. cerevisiae, high-temperature condition induced formation of cytosolic dots of both CbPab1-Venus and CbPbp1 fused to mCherry (CbPbp1-mCherry), and dots corresponding to these two SG proteins were colocalized in the cytosol (Fig. 2A). When C. boidinii cells expressing CbHog1-mCherry and CbPab1-Venus were shifted to 42°C, both CbHog1-mCherry and CbPab1-Venus formed dot structures, and most CbPab1-Venus dots colocalized with CbHog1-mCherry (Fig. 2B,C). These results indicate that, after shift to high temperature, SG formation is induced and CbHog1 is recruited to SGs.

SpSty1 was reported to affect formation of SGs under stress conditions of high osmolarity in S. pombe (Nilsson and Sunnerhagen, 2011). In the SpSTY1 deletion (Spsty1Δ) strain, the appearance of S. pombe poly-A-binding protein (SpPabp) fused to red fluorescent protein (SpPabp-RFP) as dots was markedly delayed during high-osmolarity stress but not under glucose deprivation condition. We investigated whether CbHog1 affected formation of SGs under high-temperature stress in C. boidinii. When the Cbhog1Δ strain expressing CbPab1-Venus was shifted to 42°C, CbPab1-Venus formed dots (Fig. 2D) as observed in the wild-type background (Fig. 2A). Therefore, the effect of CbHog1 was limited, if any, on the formation of SGs after shift to high temperature.

Next, to determine whether the activated form of CbHog1 is necessary for recruitment to SGs, we tested mCherry- or Venus-tagged CbHog1(K50R), a kinase-dead mutant lacking the phosphotransfer domain necessary for activity, for its recruitment to SGs. This mutation, indeed, decreased survival of C. boidinii in medium containing 0.5 M NaCl (Fig. S2B), suggesting loss of CbHog1 activity. Cytosolic dot structures of CbHog1(K50R)-Venus were observed under high-temperature stress (Fig. 2E,F), and CbHog1(K50R)-mCherry colocalized with CbPab1-Venus, a SG marker protein (Fig. 2G,H). Therefore, recruitment of CbHog1 to SGs was independent of CbHog1 activity.

Next, we investigated whether CbHog1-mCherry dots colocalized with P-bodies, another type of RNP granule. However, only 10% of dots indicating the presence of the P-body marker protein CbEdc3-Venus were associated with CbHog1-mCherry dots (Fig. S3B and Table S2).

Interestingly, CbHog1-mCherry formed dot structures not only in C. boidinii, but also in S. cerevisiae under high-temperature stress (Fig. S3C), whereas ScHog1-Venus did not form dot structures in S. cerevisiae (Fig. 1C). Therefore, we speculate that recruitment of CbHog1-mCherry to SGs is due to its aa sequence.

CbHog1 contains three conserved sequences, the common docking (CD) site (aa 300–314), the PBD-2 binding domain (PBD-2; aa 318–348), and the kinase domain (KD; aa 21–299) including the putative phosphotransfer domain (PPD; aa 47–52) and two phosphorylation sites (T174 and Y176) (Murakami et al., 2008) (Fig. 3A). We wanted to determine which regions on CbHog1 were necessary and sufficient for its recruitment to SGs.

Wild-type CbHog1-Venus formed around six fluorescent dots per cell (Fig. 3B,C). We constructed various truncated variants, and observed dot formation after shift to high temperature. The N-terminal deletion analysis (Fig. 3B) revealed that removal of the N-terminal region (aa 1–41) of CbHog1 drastically decreased dot formation. However, deletion of neither the KD domain nor the PBD-2 domain in the C-terminal region decreased CbHog1-Venus dot numbers (Fig. 3B,C). Further deletion analyses revealed that the N-terminal 30-aa sequence of CbHog1 was sufficient for partial localization, and the N-terminal 50-aa sequence of CbHog1 was sufficient for full recruitment of CbHog1-GFP to SGs (Fig. 3C).

Prediction of protein structure based on homology modeling revealed that the N-terminal region of CbHog1 is likely to assume a β-sheet structure with high hydrophobicity (Fig. 4A). CbHog1Δ(1-20)-Venus was partially localized to SGs, whereas CbHog1Δ(1-41)-Venus was not recruited to SGs (Fig. 3B). Therefore, we speculated that the second β-sheet structure composed of aa ²⁵NPVG²⁸ and ³⁵VCAAKD⁴⁰ has a greater contribution to SG localization than the first β-sheet structure composed of aa ⁵FTRT⁸ and ¹⁵FETT¹⁸. More-detailed deletion analyses revealed that CbHog1Δ(38-45)-Venus did not exhibit SG localization but, instead, dispersed in the cytosol (Fig. 4B,C). CbHog1Δ(35-40)-Venus and CbHog1Δ(39-45)-Venus were also defective in recruitment to SGs (Fig. 4B,C). Homology remodeling predicted that the 38-45 aa deletion of CbHog1 disrupted the second β-sheet structure (Fig. 4A). Taken together, these observations show that the β-sheet structures in the N-terminus region of CbHog1 (20-41) plays a critical role in recruitment of CbHog1 to SGs under high-temperature stress.

As shown above, and in contrast to other yeast Hog1-family proteins, ScHog1-Venus did not form dot structures under high-temperature stress in S. cerevisiae (Fig. 1C). Likewise, in C. boidinii, ScHog1-Venus was not recruited to SGs under high-temperature stress (Fig. 5A). Alignment of the aa sequences of Hog1 homologs (Fig. S1A) revealed that ScHog1 harbors a long C-terminal region that is not conserved in the other three yeast species. Based on the predicted structure (Fig. 4A), we speculated that this C-terminal region interferes with the β-sheet structure in the N-terminal region that is involved in the recruitment of ScHog1-Venus to SGs. To test this idea, we introduced ScHog1(1-350)-Venus lacking the C-terminal 85 aa residues into the Cbhog1Δ strain. Under high-temperature stress, ScHog1(1-350)-Venus was recruited to SGs not only in S. cerevisiae, but also in C. boidinii (Fig. 5B). However, CbHog1-ScHog1(351-435)-Venus formed dots under high-temperature stress in C. boidinii (Fig. S3D), indicating that the non-conserved C-terminal region of ScHog1 is necessary but not sufficient to prevent dot formation of ScHog1 in C. boidinii.

What is the physiological role of recruitment of CbHog1 to SGs under high-temperature stress? To answer this question, we first focused on the phosphorylation of CbHog1-Venus. As shown in Fig. S4, a band shift to a higher molecular mass was observed on Phos-tag-SDS-PAGE in cells exposed to high-temperature stress for 15 min, and also in cells exposed to 0.5 M NaCl. These results indicated that CbHog1 is phosphorylated and exists in the activated form under high-temperature stress. Because the N-terminal region of CbHog1 critical for recruitment to SGs is close to K50 – a critical aa residue for Hog1 activity in the KD domain – we sought to identify a mutant that lost the ability to be recruited to SGs but retained Hog1 kinase activity. Among the mutants impaired in recruitment to SGs (Fig. 3B,C, and Fig. 4C) – with Cb-Hog1Δ(38–45)-Venus and CbHog1Δ(35–40)-Venus not supporting growth in the presence of 0.5 M NaCl – the CbHog1Δ(1–17)-Venus mutant fully complemented and CbHog1Δ(25–28)-Venus partially complemented the growth of Cbhog1Δ in medium containing 0.5 M NaCl (Fig. 6). Therefore, CbHog1Δ(1–17)-Venus and CbHog1Δ(25–28)-Venus were identified as mutants that retained Hog1 activity while their efficiency of recruitment to SGs was reduced to ∼50% (Figs 3B and 4C).

Next, we spotted serially diluted cell suspensions onto YPD plates and incubated at 37°C for 24 h, cellular survival under high-temperature stress (after growth at 28°C) was then assessed. Although CbHog1-Venus (wild type), CbHog1Δ(38-45)-Venus (inactive Hog1), CbHog1Δ(35-40)-Venus (inactive Hog1) and Cbhog1Δ exhibited comparable levels of cellular survival under high-temperature stress (Fig. 6), survival was diminished in CbHog1Δ(1-17)-Venus and CbHog1Δ(25-28)-Venus (active but impaired in recruitment to SGs) (Fig. 6A-C). We also examined the cellular survival in cells expressing CbHog1Δ(1-10)-Venus, which fully complemented the growth of the Cbhog1Δ strain in medium containing 0.5 M NaCl (Fig. 6C). No growth defect was detected with CbHog1Δ(1-10)-Venus. These results showed that cytosolic and active CbHog1 inhibited the growth of C. boidinii and decreased cellular survival under high-temperature stress, suggesting that recruitment of active CbHog1 to SGs impedes Hog1-induced cell death under high-temperature stress.

We also examined ScHog1-Venus protein and its variants in C. boidinii. Similar to CbHog1-Venus, ScHog1-Venus was transported to the nucleus and complemented the growth defect of Cbhog1Δ in the presence of 0.5 M NaCl (Fig. 7A,B). Thus, ScHog1-Venus was confirmed to be active and functional in C. boidinii. Under conditions of high temperature, ScHog1-Venus dispersed into the cytosol in C. boidinii (Fig. 7A), and C. boidinii cells expressing ScHog1-Venus lost cell viability under high-temperature stress (Fig. 7B). Therefore, we used well-characterized ScHog1 and mutant proteins to study the relationship between Hog1 activity and the decrease in cellular survival in response to high-temperature stress in C. boidinii.

ScHog1 is phosphorylated on aa residues T174 and Y176 by the MAPKK Pbs2 (Brewster et al., 1993). Activation of ScHog1 leads to its Nmd5p-dependent nuclear import (Ferrigno et al., 1998), and it subsequently functions as a kinase through its PPD containing K52 (Hall et al., 1996). Neither the Cbhog1Δ strain expressing ScHog1(K52R)-Venus (KD mutant) nor the Cbhog1Δ strain expressing ScHog1(T174A/Y176F)-Venus (the phosphorylation defective mutant) was able to grow under 0.5 M NaCl (Fig. 7B), although ScHog1(K52R)-Venus and ScHog1(T174A/Y176F)-Venus translocated into the nucleus in C. boidinii (Fig. 7A). Under conditions of high temperature, they dispersed in the cytosol in C. boidinii (Fig. 7A), and cells expressing these inactive Hog1 variants exhibited the same cellular survival level as the wild-type strain under high-temperature stress (Fig. 7B). These results confirmed that active ScHog1-Venus dispersed in the cytosol induced cell death under high-temperature stress.

Finally, we studied whether intracellular sequestration of active ScHog1 affects its cytotoxicity. As shown above, ScHog1(1-350)-Venus was recruited to SGs after a shift to high temperature in C. boidinii (Fig. 5B). ScHog1(1-350)-Venus was able to complement the growth defect of the Cbhog1Δ strain under 0.5 M NaCl, confirming that the protein was functional. As shown in Fig. 7C, deletion of 85 aa from the C-terminal region [ScHog1(1-350)-Venus] was able to reduce the cytotoxicity of ScHog1-Venus under high-temperature stress. Accordingly, we speculated that sequestration of ScHog1(1-350)-Venus to SGs impeded cytotoxicity.

Plasma membrane targeting of Ras proteins requires posttranslational modification of the C-terminal CAAX box – in which C=cysteine, A any aliphatic aa and X=any aa – and either palmitoylation or the presence of a polybasic domain (Casey et al., 1989; Deschenes et al., 1989; Hancock et al., 1990). To investigate whether anchoring of active ScHog1-Venus to the plasma membrane can suppress cytotoxicity under high-temperature stress, ScHog1-Venus-CAAX^Ras2 was expressed in Cbhog1Δ cells. ScHog1-Venus-CAAX^Ras2 was anchored to the plasma membrane (Fig. 7A). In addition, as shown in Fig. 7B, ScHog1-Venus-CAAX^Ras2 complements the growth defect of Cbhog1Δ strain under 0.5 M NaCl and also suppress the cytotoxicity of active ScHog1-Venus under high-temperature stress (Fig. 7B). Therefore, the intracellular sequestration of ScHog1-Venus not only to SGs but also the plasma membrane may impede the function of ScHog1, suggesting that sequestration of yeast Hog1 plays an important role in downregulating yeast Hog1 activity.

Recent work has shown that TORC1 is recruited to lysosomal membranes to activate starvation signaling, whereas recruitment to SGs downregulates TORC1 activity (Demetriades et al., 2014; Takahara and Maeda, 2012). Intracellular sequestration of several signaling molecules in SGs has been investigated in relation to their functions (Arimoto et al., 2008; Kim et al., 2005). However, the molecular basis of the recruitment of these signaling factors to SGs and the physiological role of sequestration of these proteins to SGs remain to be elucidated in detail. Here, we revealed that important MAP kinases, the Hog1 proteins, exhibit drastic changes in intracellular localization to SGs under high-temperature stress in different yeast species.

We found that recruitment of CbHog1 to SGs depended on a hydrophobic β-sheet structure on its N-terminal 50 aa, and this is region necessary and sufficient for recruitment to SGs. In other words, recruitment of Hog1 to SGs depends on its own aa sequence. Aa sequences responsible for localization to SGs were analyzed in several proteins that did not affect formation of SGs as with CbHog1 (Bentmann et al., 2012; Weissbach and Scadden, 2012); however, no consensus sequences were determined. CbHog1 was phosphorylated and activated under high-temperature stress. Because the inactivated form of Hog1 protein was recruited to SGs, activation of Hog1 itself is not required for its recruitment.

Another interesting finding was that removal of the non-conserved C-terminal region of ScHog1 enabled recruitment of ScHog1 to SGs (Fig. 5B) in S. cerevisiae. This shows an inherent potential of ScHog1 to be recruited to SGs, which might be negatively regulated by its C-terminal region. Although this C-terminal region of ScHog1(351-435) is necessary to prevent dot formation of ScHog1 proteins in C. boidinii and S. cerevisiae under high-temperature, contrary to our presupposition, CbHog1-ScHog1(351-435)-Venus formed dots in C. boidinii (Fig. S3D). We speculate that there is some regulatory mechanism that involves interaction between ScHog1(1-350) and ScHog1(351-435) regions to prevent and regulate dot formation of ScHog1. A database search revealed that Hog1 proteins from Zygosaccharomyces rouxii, Torulaspora delbrueckii, and Candida glabrata also possess C-terminal non-conserved regions similar to that in ScHog1. Although we did not observe recruitment of Hog1 to SGs in these yeasts, regulation of Hog1 recruitment to SGs in S. cerevisiae, as well as in the three species mentioned above, remains an open question that awaits further study.

More than 20 proteins were identified as components of yeast SG, and they seem to have some interactions with P-bodies (Buchan and Parker, 2009; Buchan et al., 2011). It should be noted that, although most CbPab1-Venus dots colocalized with CbHog1-mCherry dots, two-thirds of CbHog1-mCherry dots did not show colocalization with CbPab1-Venus dots (Fig. 2B,C,G,H). Two explanations are possible for this observation. First, CbPab1-Venus does not equally represent the total SG population. So colocalizaton of CbPab1-Venus and CbHog1-mCherry inevitably shows only a part of the total colocalization of CbHog1-mCherry and SG. Second, there is a high possibility that CbHog1-mCherry associates with the P-body. Indeed, we observed that about 10% of CbHog1-mCherry dots associated with the P-body marker CbEdc3-Venus (Fig. S3B). As is the case, the dynamic feature of SGs is still to be further explored.

We then investigated the physiological significance of Hog1 sequestration in SGs by using different mutants of CbHog1-Venus and ScHog1-Venus proteins. Initially, we screened for a CbHog1 mutant protein that lacked the ability to be recruited to SGs (as determined by fluorescence microscopy) but retained kinase activity (as determined by the ability to grow in 0.5 M NaCl). Among various strains tested, Cb-Hog1Δ1-17 and Cb-Hog1Δ25-28 exhibited such a phenotype (Fig. 6). Our results with these mutants indicated that the active form of CbHog1 exerted a cytotoxic effect, and that impairment of recruitment to SGs promoted cell death. A previous study reported a negative effect of hyperactivated Hog1 kinase in the absence of two protein phosphatases, Ptp2 and Ptp3, during cell growth at high temperature (Winkler et al., 2002).

To further investigate the effect of Hog1 activity and SG recruitment on cell death, we expressed non-functional ScHog1-Venus mutants (kinase-dead and phosphorylation-defective) in C. boidinii. On the one hand, the wild-type ScHog1 protein completely complemented the growth of the Cbhog1Δ strain in medium containing 0.5 M NaCl, and previously characterized inactive ScHog1 mutations lost the ability to grow in the presence of high salt concentration (Fig. 7). On the other hand, under high-temperature stress, the wild-type ScHog1 protein dispersed in the cytosol and induced cell death after high-temperature shift, and this activity depended on Hog1 activity. Therefore, ScHog1 protein that was activated under high-temperature stress induces cell death when present in the cytosol. Although ScHog1(1-350) was active and able to complement the growth of Cbhog1Δ strain in 0.5 M NaCl, this mutant exhibited less cytotoxicity under high-temperature stress than wild-type ScHog1. These results all suggest that active Hog1 is sequestered to SGs for its downregulation.

The site of intracellular sequestration of Hog1 for downregulation is not necessarily limited to SGs. When ScHog1 was anchored to the plasma membrane (ScHog1-Venus-CAAX^Ras2), cytotoxicity under high-temperature stress was reduced, as observed for ScHog1(1-350) (Fig. 7). Therefore, the positive effect of activated Hog1 at SGs is likely to be small. Instead, we speculate that sequestration of cytosolic Hog1 to SGs is a strategy for modulation of Hog1 activity under high-temperature stress.

However, some part within active Hog1 proteins may also be translocated to the nucleus for gene activation. SpSty1-GFP formed cytosolic dots but, simultaneously, localized to the nucleus under high-temperature stress (Fig. 1C). These features indicate that SpSty1 upregulates gene expression in the nucleus to help the cell adapt to high-temperature stress but, at the same time, activity of a fraction of SpSty1 is suppressed by sequestration into dot structures. Compared with other yeast species, SpSty1 might play a more significant and active role in adaptation to high temperature. Consistent with this idea, previous studies have demonstrated that ScHog1 is only important for rapid recovery from high-temperature stress (Winkler et al., 2002), and that the Spsty1Δ strain loses the ability to grow at high temperatures (Millar et al., 1995).

In summary, yeast Hog1 is sequestered in SGs under high-temperature condition through its N-terminal hydrophobic β-sheet structure; and we propose that activated Hog1 in the cytosol can be downregulated by this sequestration. As for S. cerevisiae, the experimental strain is a conventional baker's yeast that has been preserved in artificial environments for more than 1000 years. By contrast, C. boidinii, which proliferates on plant leaves, must adapt to severe environmental changes under natural conditions. Therefore, recruitment of Hog1 to SGs for downregulation might confer a survival advantage in this species. Our results reveal the importance of intracellular sequestration of active Hog1 kinase into SGs, and we believe that elucidation of the underlying molecular mechanisms will – at molecular level – advance our understanding of the physiological role of SGs and the intricate regulation of intracellular signaling pathways.

Escherichia coli DH10B was used for plasmid propagation. E. coli was grown at 37°C in lysogeny broth (LB) medium (0.5% yeast extract, 1% tryptone, 0.5% NaCl) supplemented as required with ampicillin (100 µg/ml). Solid medium was prepared by adding 2% agar to LB medium.

The yeast strains used in this study are listed in Table S3. S. cerevisiae, C. boidinii, and P. pastoris cells were grown on yeast extract peptone dextrose (YPD) medium (1% Bacto-yeast extract, 2% Bacto-peptone, 2% glucose), synthetic defined (SD) medium (0.67% Yeast Nitrogen Base without aa, 2% glucose), or SD-Trp medium (SD medium supplemented with 1.92 g/l yeast synthetic drop-out medium supplements without tryptophan; Sigma-Aldrich). The pH of SD medium was adjusted to 6.0 with NaOH. For selection of strain Cbhog1Δ (ura3), 2.4 µg/ml of uracil or 0.08% w/v of 5-fluoroorotidic acid (5-FOA) was added to SD medium. S. pombe cells were grown on YES medium (0.5% Bacto-yeast extract, 3% glucose, 75 mg/l adenine, leucine, uracil and lysine). Cultivation was performed at 28°C under aerobic conditions, and cell growth was monitored by measuring the optical density of the medium at OD₆₁₀. Solid media were prepared by adding 2% agar to the appropriate liquid media. The Spsty1Δ strain expressing Sty1-GFP (FY14931) was provided by the NBRP (YGRC), Japan.

Plasmid DNAs from E. coli were isolated using Wizard Plus SV Minipreps DNA Purification System (Promega, Madison, WI). Transformation of E. coli was performed by the method of Dower et al. (1988). Yeast DNA was isolated by the method of Cryer et al. (1975). Transformation of S. cerevisiae was performed using the Fast yeast transformation kit (G-Biosciences, Maryland Heights, MO). Transformation of C. boidinii was performed using a modified version of the lithium acetate method (Sakai and Tani, 1992). Transformation of P. pastoris was performed as described previously (Higgins and Cregg, 1998).

The oligonucleotide primers used for PCR reactions are listed in Table S4. The plasmids used in this study are listed in Table S5. A deletion cassette for the CbHOG1 gene was constructed as follows: Primers CbHOG1_UP_Fw and CbHOG1_UP_Rv were used to amplify a 1.0-kb fragment, using genomic DNA of C. boidinii as template. The PCR product was fused with the 7.6-kb KpnI-digested fragment of pSK+SPR using the In-fusion HD cloning kit (TaKaRa, Kyoto, Japan), yielding pHC100. Then, primers CbHOG1_DOWN_Fw and CbHOG1_DOWN_Rv were used to amplify a second 1.0-kb fragment, using genomic DNA of C. boidinii as template. The PCR product was fused with the 8.6-kb SacI-digested fragment of pHC100 using the In-fusion HD cloning kit, yielding the CbHOG1 disruption vector pHC101.

The CbHOG1 promoter and ORF region without the STOP codon were amplified with primers CbHOG1_Vn_Fw and CbHOG1_Vn_Rv, using genomic DNA of C. boidinii as template. The PCR product was fused with the 6.4-kb AflIII–SalI fragment of plasmids pKK001 and pSPM001 using the In-fusion HD cloning kit, yielding pHC200 and pHC400, respectively. The ScHOG1 ORF region without the STOP codon was amplified with primers P_CbACT1ScHOG1_Vn_Fw and P_CbACT1ScHOG1_Vn_Rv. The PCR product was fused with the 8.1-kb SalI-digested fragment of pKK001 using the In-fusion HD cloning kit, yielding pHC300. The CbPAB1 promoter and ORF region without the STOP codon were amplified with primers CbPab1_Fw and CbPab1_Rv, using genomic DNA of C. boidinii as template. The PCR product was fused with the 6.4-kb AflIII–SalI fragment of plasmid pKK001 using the In-fusion HD cloning kit, yielding pHC401. pHC404 was constructed in a similar manner, with primers CbEdc3_Fw and CbEdc3_Rv. The CbPBP1 promoter and ORF region without the STOP codon were amplified with primers CbPbp1_Fw and CbPbp1_Rv, using genomic DNA of C. boidinii as template. The PCR product was fused with the 6.4-kb AflIII–SalI fragment of plasmid pSPM001 using the In-fusion HD cloning kit, yielding pHC403. Vector pHC402, encoding CbPab1-Venus under the control of the PAB1 promoter with the LEU2 marker (Sakai and Tani, 1992) was constructed by ligation of an AatII–NarI fragment from pHC401 with an AatII–NarI fragment of the LEU2 gene. Vector pHC405 was constructed in a similar manner from pHC404.

C. boidinii strains expressing truncated variants of CbHog1 were constructed using the plasmids pHC200-pHC213, which were transformed into Cbhog1Dura3. pHC201, a deletion mutant of CbHOG1, was constructed as follows: Primers CbHOG1_Vn_Δ(1-10)_Fw and CbHOG1_Vn_ΔN_Rv were phosphorylated at their 5′-ends with T4 nucleotide kinase, and used to amplify an 8.6-kb fragment by inverse PCR using pHC200 as template. The PCR product was self-ligated, yielding vector pHC201. Other deletion mutants of CbHOG1 were constructed in a similar way with primer sets CbHOG1_Vn_Δ(1-17)_Fw/CbHOG1_Vn_ΔN_Rv, CbHOG1_Vn_Δ(1-20)_Fw/CbHOG1_Vn_ΔN_Rv, CbHOG1_Vn_Δ(1-30)_Fw/CbHOG1_Vn_ΔN_Rv, CbHOG1_Vn_Δ(1-41)_Fw/CbHOG1_Vn_ΔN_Rv, HOG1_Vn_ΔC_Fw/CbHOG1_Vn_Δ(348-398)_Rv, HOG1_Vn_ΔC_Fw/CbHOG1_Vn_Δ(318-398)_Rv, HOG1_Vn_ΔC_Fw/CbHOG1_Vn_Δ(300-398)_Rv, HOG1_Vn_ΔC_Fw/CbHOG1_Vn_(1-20)_Rv, HOG1_Vn_ΔC_Fw/CbHOG1_Vn_(1-30)_Rv, HOG1_Vn_ΔC_Fw/CbHOG1_Vn_(1-40)_Rv, HOG1_Vn_ΔC_Fw/CbHOG1_Vn_(1-50)_Rv, and HOG1_Vn_ΔC_Fw/CbHOG1_Vn_(346-398)_Rv, yielding vectors pHC202, pHC203, pHC204, pHC205, pHC206, pHC207, pHC208, pHC209, pHC210, pHC211, pHC212, and pHC213, respectively. Deletion mutants of ScHOG1 were also constructed in a similar manner. The 1.6-kb SalI–EcoRI fragment of plasmid pSPM001 and the 4.1-kb SalI–EcoRI fragment of plasmid pHC200 were ligated, yielding vector pHC400.

C. boidinii strains expressing other series of truncated variants of CbHog1 were constructed using the plasmids pHC214-pHC221, which were transformed into Cbhog1Dura3. pHC214, a deletion mutant of CbHOG1, was constructed as follows. The CbHOG1 promoter and ORF region were amplified with primers CbHOG1_Vn_Δ(25-28)_Fw and CbHOG1_Vn_Δ(25-28)_Rv, using genomic DNA of C. boidinii as template. The PCR product was purified using the NucleoSpin Gel and PCR Clean-up kit and directly transformed into E. coli DH10B. The fragment was circularized in the host cells, yielding vector pHC214. Additional deletion mutants of CbHOG1 were constructed in a similar manner using primer sets CbHOG1_Vn_Δ(35-40)_Fw/CbHOG1_Vn_Δ(35-40)_Rv, CbHOG1_Vn_Δ(38-45)_Fw/CbHOG1_Vn_Δ(38-45)_Rv, CbHOG1_Vn_Δ(39-45)_Fw/CbHOG1_Vn_Δ(39-45)_Rv, CbHOG1_Vn_Δ(40-45)_Fw/CbHOG1_Vn_Δ(40-45)_Rv, CbHOG1_Vn_Δ(43-45)_Fw/CbHOG1_Vn_Δ(43-45)_Rv, and CbHOG1_Vn_Δ(K50R)_Fw/CbHOG1_Vn_Δ(K50R)_Rv, yielding vectors pHC215, pHC216, pHC217, pHC218, pHC219, and pHC220, respectively. Vector pHC221, encoding CbHog1(K50R)-mCherry under the control of the HOG1 promoter with the LEU2 marker, was constructed by ligation of an AatII–NarI fragment of pHC220 with an AatII–NarI fragment of the LEU2 gene.

C. boidinii strains expressing variant ScHog1 were constructed using the plasmids pHC500-pHC504, which were transformed into Cbhog1Dura3. pHC501, encoding a kinase-inactive form of ScHOG1, was constructed as follows. Primers ScHOG1_KD_Fw and ScHOG1_KD_Rv were phosphorylated at their 5′-ends with T4 nucleotide kinase, and used to amplify an 8.6-kb fragment by inverse PCR using plasmid pHC500 as template. The PCR product was self-ligated, yielding vector pHC501. Other plasmids harboring mutated ScHOG1 were constructed in a similar way using primer sets ScHOG1_nP_Fw/ScHOG1_nP_Fw and ScHOG1_CAAX_Fw/ScHOG1_CAAX_Rv, yielding vectors pHC502, and pHC503, respectively. pHC504 was constructed as follows. Primers HOG1_Vn_ΔC_Fw and ScHOG1_350_Rv were phosphorylated at their 5′-ends with T4 nucleotide kinase, and used to amplify an 8.3-kb fragment by inverse PCR using plasmid pHC300 as template. The PCR product was self-ligated, yielding vector pHC504.

The C. boidinii strain expressing CbHog1-ScHog1(351-435) was constructed using the plasmid pHC505, which was transformed into Cbhog1Dura3. pHC505, encoding CbHOG1 and the C-terminal region of ScHOG1, was constructed as follows. The CbHOG1 promoter and ORF region were amplified with primers CbHOG1_Vn_(1-398)_Fw and CbHOG1_Vn_(1-398)_Rv, using the plasmid pHC200 as template. The C-terminal region of ScHOG1 was amplified with primers ScHog1_(351-435)_Fw and ScHog1_(351-435)_Rv, using the plasmid pHC300 as template. Those two fragments were fused using the In-fusion HD cloning kit, yielding pHC505.

A deletion cassette for the CbHOG1 gene was amplified with CbHOG1_SPR_Fw/CbHOG1_SPR_Rv, using plasmid pHC101 as template. The PCR product was transformed into C. boidinii TK62 using a modified version of the lithium acetate method (Sakai and Tani, 1992). Proper gene disruption was confirmed by colony PCR. The constructed Cbhog1Δ strain was converted to uracil auxotrophy by 5-FOA selection, yielding the Cbhog1Δura3 strain. Restoration of the marker gene URA3 was confirmed by PCR analysis. The nucleotide sequence of CbHOG1 was deposited in DNA Data Bank of Japan (DDBJ) under the accession number LC312444. The nucleotide sequences of CbPAB1, CbPBP1 and CbEDC3 were also deposited in DDBJ under accession numbers, LC312445, LC312446, and LC312447, respectively.

The oligonucleotide primers used for PCR reactions are listed in Table S4. The plasmids used in this study are listed in Table S5. The S. cerevisiae strain expressing ScHog1-GFP was constructed as follows: the GFP-LEU2 region was PCR-amplified with primers ScHOG1CTag_Fw and ScHog1CTag_Rv, using vector pMO152 as template. The fragment containing the GFP-LEU2 gene was directly integrated into the C-terminal region of the ScHOG1 gene, yielding a S. cerevisiae strain expressing ScHog1-GFP.

The oligonucleotide primers used for PCR reactions are listed in Table S4. The plasmids used in this study are listed in Table S5. The deletion strain of the PpHOG1 gene was constructed as follows. Deletion cassettes were amplified with primers Zhog1_UPfw/Zhog1_UPrv and Zhog1_DWfw/Zhog1_DWrv, using genomic DNA of P. pastoris as template. The PCR products were linked by PCR with primers Zhog1_UPfw and Zhog1_DWrv, yielding a 2.0-kb fragment containing the upstream and downstream regions of the PpHOG1 gene. The zeocin resistance marker, which was used for the selection of disruptants of the PpHOG1 gene, was PCR-amplified using vector pPICZ A as template, yielding a 1.9-kb fragment. These two fragments were fused using the In-fusion HD cloning kit, yielding PpHOG1 disruption vector pHP001. pHP001 vector was linearized with XbaI and transformed into the host strain PPY12. Proper gene disruption was confirmed by colony PCR.

The vector harboring PpHog1-YFP was constructed as follows. The PpHOG1 promoter and the ORF region without the STOP codon were amplified using primers PpHog1IF_fw and PpHog1IF_rv, using genomic DNA of P. pastoris as template. The PCR product was fused with the 3.4-kb BamHI–SpeI fragment of plasmid pNT204 using the In-fusion HD cloning kit, yielding pHP100. Vector pHP100 was linearized with XbaI and transformed into the PpHOG1 deletion (Pphog1Δ) host strain.

Yeast cells were pre-cultured to stationary phase at 28°C in 5 ml YPD (for S. cerevisiae, C. boidinii or P. pastoris) medium or YES (for S. pombe) medium. Subsequently, 20-μl aliquots of the cultures were transferred to 5 ml SD-Trp (for S. cerevisiae), SD (for C. boidinii or P. pastoris), or YES (for S. pombe) medium, and the cells were grown to early-log phase at 28°C. The cultures were centrifuged at 6000 rpm for 2 min, and cells were transferred to 5 ml fresh SD-Trp medium containing 0.5 M NaCl (for S. cerevisiae), SD medium containing 0.4 M NaCl (for C. boidinii), SD medium containing 0.5 M NaCl (for P. pastoris), or YES medium containing 1 M KCl (for S. pombe) at 28°C for 5 min (for S. cerevisiae, C. boidinii or P. pastoris) or 10 min (for S. pombe). The cells were then harvested by centrifugation for microscopic observation.

Yeast cells were pre-cultured to the stationary phase at 28°C in 5 ml YPD (for S. cerevisiae, C. boidinii, or P. pastoris) or YES (for S. pombe) medium. Subsequently, 20-μl aliquots of these cultures were transferred to 5 ml SD-Trp (for S. cerevisiae), SD (for C. boidinii or P. pastoris), or YES (for S. pombe) medium, and the cells were grown to early-log phase at 28°C. The cells were then subjected to the indicated high-temperature stresses, and harvested by centrifugation for microscopic observation.

After cell cultivation for the indicated time period, sample cells were harvested, washed once and fixed with 1 ml of 70% ethanol for 30 min at room temperature. Fixed cells were then washed twice, resuspended in 150 µl of sterilized water, and stained with 150 µl of 0.125 µg/ml DAPI (4′, 6′-diamidino-2-phenylindole) solution. After 10 min of incubation, samples were used for microscopic observation.

Fluorescence microscopy was performed using an IX81 inverted microscope (Olympus) equipped with an UPLAN-Apochromat 100×/1.35 NA oil iris objective lens. Venus and mCherry signals were acquired using a Plan Fluor 100×lens (Carl Zeiss) with pinhole set to 1.02 airy units for YFP acquisition. Images were analyzed using the METAMORPH imaging software (Molecular Devices) and Adobe Photoshop CS6 (Adobe). All scale bars shown in microscope photographs are 2 μm. Fluorescence observations of cells cultured in vitro were repeated at least twice, and three images of different fields were acquired at each time point.

Cells (>50 per field) were counted in at least three fluorescence microscope fields, and independent observations were repeated at least three times.

To prepare samples for Phos-tag SDS-PAGE, harvested cells were suspended in lysis buffer (0.2 N NaOH, 1% [v/v] 2-mercaptoethanol) and incubated at 4°C for 10 min. Next, trichloroacetic acid was added (final concentration 10%). Then, the samples were vortexed, incubated at 4°C for 10 min and centrifuged. Subsequently, the pellet was washed twice with acetone and resuspended in sample buffer (16.7 mM Tris-HCl pH 6.8, 1% SDS, 10% Glycerol, 1% 2-mercaptoethanol, 1 mM MnCl₂) and boiled for 3 min.

The prepared samples were electrophoresed on a 6% Phos-tag SDS polyacrylamide gel (containing 30 μM Phos-tag [Wako], 100 μM MnCl₂) or 10% normal SDS polyacrylamide gel. Western analysis was performed by the method of Towbin et al. (Towbin et al., 1979). The proteins on a gel were transferred to a PROTRAN nitrocellulose transfer membrane (Schleicher & Schuell BioScience, Inc., Dassel, Germany) in a tank blotting system (BIO-RAD) at 100 V for 1 h. After the transfer, the membrane was blocked for 1 h in TBS-T buffer (0.3% Tris, 0.8% NaCl, 0.02% KCl, 0.1% Tween 20 pH 8.0) supplemented with 5% skimmed milk. The blots were then incubated for 1 h with monoclonal anti-GFP antibody (Clontech, cat. no. 632381) at 1:1000 dilution in TBS-T buffer, and washed three times with TBS-T. Subsequently, the blots were incubated for 1 h with an anti-mouse IgG (MBL, cat. no. 330) at 1:10,000 dilution in TBS-T, and washed three times in TBS-T. Immunoreactive bands were visualized using the Western Lightning Chemiluminescence Reagent Plus (Perkin-Elmer Life Science), and the signals were detected on a Light-Capture II (ATTO).