INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The cellular effects of changes in tonicity have been investigated extensively (Robbins et al., 1970; Banuett, 1998; Gustin et al., 1998; Lang et al., 1998). In particular, analysis of the yeast Saccharomyces cerevisiae has elucidated major roles for two MAP kinase signaling pathways: the Pkc1p "cell integrity" pathway, which is stimulated by hypotonic shock (Watanabe et al., 1994; Davenport et al., 1995), and the Hog1p pathway, which is stimulated by hypertonic shock (Brewster et al., 1993; Posas and Saito, 1997). Each pathway is required for long-term survival in the corresponding medium. Hypertonic shock profoundly affects the actin cytoskeleton (Chowdhury et al., 1992; Mulholland et al., 1994; Delley and Hall, 1999).

The Pkc1p pathway can receive input from plasma membrane glycoproteins of the Wsc family and from the Rho GTPases and their regulatory factors, for example in the context of control of cell polarization and cell wall biosynthesis (Gustin et al., 1998). Interestingly, some signaling events that require Pkc1p appear not to require the corresponding downstream MAP kinase cascade, which is linked to activation of the transcription factors Rlm1p and Swi4p/Swi6p (Lee and Levin, 1992; Gustin et al., 1998; Delley and Hall, 1999; Li et al., 2000). In contrast, the Hog1p pathway is initiated by a two-component phosphorelay involving the cell surface transmembrane protein Sln1p and the cytosolic protein Ypd1p (Posas et al., 1996). Stimulation of this pathway causes phosphorylation and transient nuclear entry of the terminal kinase, Hog1p, as well as of the transcription factor Msn2p (Ferrigno et al., 1998; Görner et al., 1998; Reiser et al., 1999). The downstream substrate(s) of Hog1p kinase is not known.

We have observed recently that arrest of the secretory pathway in yeast inhibits nuclear import and causes many nucleolar and nucleoplasmic proteins to relocate to the cytoplasm. These events (the "arrest of secretion response") are reversible and are prevented if Hog1p is overexpressed or if Pkc1p is deleted (Nanduri et al., 1999; our unpublished observations). These observations caused us to investigate the impact of changes in tonicity on the yeast nucleus.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Yeast Strains and Growth Conditions

Yeast strains (Table 1) were grown to midlog phase at 23°C in YEPD or in synthetic media lacking individual components to select for the presence of plasmids of interest. The osmotically sensitive strain DL376 was grown in YEPD supplemented with 0.5 M sorbitol. Chromosomal integration of GFP-NSP1 (pSW956) and GFP-NIC96 (pSW950) into YPH500 was as described by Bucci and Wente (1998). Cells were transformed with pPS1739 (HOG1-GFP, URA3, CEN [Ferrigno et al., 1998]), pBM3495 (MIG _217-400-GFP-lacZ, URA3, 2 µ [De Vit et al., 1997]), pSW636 (NUP49-GFP, LEU2, CEN [Bucci and Wente, 1998]), or pGFP-TBP (GFP-TBP, HIS3, CEN [Patterson et al., 1998]) by standard procedures and maintained on appropriate selective plates. DL376 transformants carrying pDL468 (GAL-PKC1^wt-HA, URA3, CEN [Gray et al., 1997]), pDL469 (GAL-PKC1p^K853R-HA, URA3, CEN [Gray et al., 1997]), pBM743 (GAL-PKC1p^R398A, URA3, CEN [Delley and Hall, 1999]), pHPS29 (PKC1^{C434S, C437S, C514S, C517S}, TRP1, CEN [Jacoby et al., 1997]), or pHPS30 (pPKC1^wt, TRP1, CEN [Jacoby et al., 1997]) were maintained on selective plates containing 0.5 M sorbitol.

For hypertonic shock, the osmolarity was increased by adding 3 M sorbitol, 5 M NaCl, or 5 M KCl to reach a final concentration of 1 M sorbitol or 0.5 M NaCl or KCl. For heat shock, cultures were transferred to a shaking water bath at 37°C. For hypotonic shock, cultures in YEPD were diluted with 4 volumes of water.

Immunostaining

Cells were fixed and processed as described (Nanduri et al., 1999). A 1/10 volume of 37% formaldehyde was added directly to growing cultures for 10 min, followed by sedimentation and further fixation in 3.7% formaldehyde, 10% methanol, in 0.1 M potassium phosphate buffer (pH 6.5) for 10 min. Cells were spheroplasted with Zymolyase (ICN, Costa Mesa, CA; catalog number 320921), allowed to adhere to polylysine-coated slides, dehydrated with the use of 20°C methanol for 5 min, 20°C acetone for 30 s (Wente et al., 1992), and immunostained with the use of the antibodies described by Liu et al. (1996). GFP-tagged proteins were detected without fixation. DNA was stained with DAPI. Cells were examined with a Nikon Microphot-FX microscope with the use of a 100× objective. Images were collected with the use of a Diagnostic Instruments (Sterling Heights, MI) 1.1.0 SPOT camera. Final figures were produced by with the use of Adobe Photoshop.

Immunoblotting

Glass-bead extracts of cells prepared with 10 mM Tris-HCl, pH 7.5, containing 1 mM PMSF, 1 µg/ml leupeptin and aprotinin, and 1% SDS were electrophoresed on a 8% SDS-polyacrylamide gel and blotted onto nitrocellulose membranes. The blots were blocked with 2% nonfat dry milk, probed with polyclonal anti-Fpr3p (J. Thorner, University of California, Berkeley), polyclonal anti-phospho-specific p38 (Hog1p) (New England Biolabs, Beverly, MA, catalog numbers 9211 and 9212), and developed with the use of ECL chemiluminescence (Amersham/Pharmacia, Arlington, IL).

RNA Synthesis

Cells were exposed to hypertonic shock for 0-4 h. Aliquots were pulse-labeled with [³H-methyl]-methionine (NEN Life Sciences [Boston, MA]; NET-061x, 70-85 Ci/mmol) for 5 min over this period. Total RNA was then extracted with acid phenol at 65°C for 1 h. Aliquots were precipitated with trichloroacetic acid and counted.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Reversible Relocation of Nuclear Proteins

When wild-type cells are exposed to 0.5 M NaCl, many nuclear proteins rapidly relocate to the cytoplasm. Extensive relocation was seen within 2 min and relocation was complete by 10 min. Figure 1A illustrates the relocation of the nucleolar proteins Cbf5p, Nop1p, and Ssb1p (Jordan and Shaw, 1995; Liu et al., 1996). The nucleolar proteins Fpr3p (Figure 1B) and Nsr1p (our unpublished results) also relocated. In addition, several nucleoplasmic proteins relocated, e.g., the shuttling nucleoplasmic hnRNP-like protein Npl3p/Mtr13p (Flach et al. 1994; Singleton et al., 1995) (Figure 1A) and the transcription factor fusion Mig1p-GFP--galactosidase (De Vit et al., 1997) (our unpublished results). In contrast, Figure 1A shows that the hnRNP-like protein Nab2p (Anderson et al., 1993) did not relocate. This was also the case for a chromatin-associated protein, a GFP fusion of TBP (Patterson et al., 1998) (our unpublished results). To learn whether hypertonic shock per se is responsible for the relocation of Fpr3p and Nop1p, we have also used 0.5 M KCl and 1 M sorbitol. Both had effects comparable to those of 0.5 M NaCl (our unpublished results), but strong hypotonic shock did not cause relocation (Figure 1B). To simplify the further description of these phenomena, although we have monitored several proteins that relocate, we focus on the nucleolar prolyl isomerase Fpr3p.

View larger version (50K): [in this window] [in a new window]   Figure 1.   Relocation of nuclear proteins during osmotic shock. (A) Several nuclear proteins were localized in wild-type cells (strain YPH500) before and after 10 min of hypertonic shock (0.5 M NaCl). (B) Time course of transient relocation of the nucleolar protein Fpr3p in YPH500 cells during hypertonic shock, lack of effect of 100 µg/ml cycloheximide added 15 min before the osmotic shock, and lack of effect of hypotonic shock (see MATERIALS AND METHODS) on Fpr3p localization. (C) YPH500 cells were subjected to hypertonic shock as in B, and samples containing equal amounts of protein were analyzed to detect activation of Hog1p by Western blotting.

On continued incubation in hypertonic medium, Fpr3p reentered the nucleus within 1 h (Figure 1B). This cycle was not affected by adding 100 µg/ml cycloheximide 15 min before the shock (Figure 1B). Judging from these observations and from Western blotting (our unpublished results), the copies of Fpr3p that leave the nucleus are those that return. Interestingly, the kinetics of transient Hog1p activation and nuclear entry were comparable to the kinetics of Fpr3p exit and return to the nucleus (Figure 1C) (Ferrigno et al., 1998).

The Pathway of Signaling

To define the pathway by which hypertonicity signals to the nucleus, we have examined the effects of 1 M sorbitol and 0.5 M NaCl on the localization of nucleolar proteins in strains that are deleted for individual candidate genes of interest.

Hog1p Pathway. Because the Hog1p pathway is stimulated by hypertonic shock, we hypothesized that it would be required for the relocation or return of proteins that relocate. Nevertheless, sho1, ssk1, ste11, pbs2, and hog1 strains showed reversible relocation of Fpr3p (Figure 2), and of Nop1p and Npl3p/Mtr13p (our unpublished results), that was equivalent to that in wild-type cells. The impact of hypertonicity on the nucleus thus involves intermediates other than the Hog1p MAP kinase cascade.

View larger version (75K): [in this window] [in a new window]   Figure 2.   Lack of involvement of the Hog1p map kinase cascade in the relocation of Fpr3p. Fpr3p was localized 0, 10, and 60 min after hypertonic shock. The deletion strains are sho1 (V1273), ssk1 (V1275), ste11 (V1276), pbs2 (V1274), and hog1 (JBY13).

Pkc1p Pathway. The Pkc1p pathway is not known to be stimulated by hypertonic shock in yeast, but it is stimulated by hypotonic shock. Nevertheless, we observed that a pkc1 strain resisted relocation of nuclear proteins during hypertonic shock for 10-60 min (Figure 3A, four panels at the left). To inquire whether the kinase activity of Pkc1p is needed to relocate Fpr3p, we have transformed pkc1 with plasmids that drive expression of wild-type Pkc1p or an active site mutant (K853R) from a galactose-inducible promoter. As shown in Figure 3A (middle four panels), hypertonic shock relocated Fpr3p only in the former strain. Figure 3A (four panels at the right) also shows that relocation did not occur in cells that express a mutant form of Pkc1p in which the putative diacylglycerol-binding site has been destroyed (Pkc1p^4C/S) (Jacoby et al., 1997). In all of the above experiments, wild-type and pkc1 strains from the same background (EG123, DL376) were grown with osmotic support, because it is required for survival in the absence of functional Pkc1p.

View larger version (97K): [in this window] [in a new window]   Figure 3.   Involvement of Pkc1p, but not of functionally related proteins, in relocation. (A) Localization of Fpr3p in wild-type cells (EG123); pkc1 (DL376); DL376 carrying pDL468 for expression of wild-type Pkc1p-HA; DL376 carrying pDL469 for expression of the active site mutant Pkc1K853R-HA; DL376 carrying either pHPS30, encoding wild-type Pkc1p, or pHPS29, encoding Pkc1p4C/S, which is purported not to bind DAG. All cells were shocked for 0 or 10 min. The strains in which Pkc1p expression is driven by a galactose-regulated promoter (middle panels) were grown in raffinose/sorbitol medium that was supplemented with 2% galactose for 4 h before the hypertonic shock. (B) Localization of Fpr3p in pkc1 (DL376) cells carrying pDL468 for expression of Pkc1p-HA or a corresponding plasmid, pBM743, for expression of the autophosphorylation site (hyperactive) mutant Pkc1pR398A-HA. Because both plasmids drive Pkc1p expression from a galactose-regulated promoter, both strains were grown in raffinose medium, received 2% galactose at time zero, and were processed for localization of Fpr3p after 0, 30, or 60 min. (C) Localization of Fpr3p in wild-type cells (EG123), bck1 (DL253), mpk1 (TS45-1a), and wsc1 (PA39-1b) after hypertonic shock for 0, 10, or 60 min.

pkc1

pkc1

bck1

mpk1

wsc1

Protection of the Nucleus Against the Effects of Hypertonic Shock

Because Pkc1p and Hog1p often have opposing effects, we inquired whether overexpression of Hog1p might protect the nucleus from the effects of osmotic shock. Indeed, as shown in Figure 4 (left four panels), cells that carry a high-copy plasmid for expression of Hog1p did not relocate nuclear proteins during hypertonic shock.

View larger version (67K): [in this window] [in a new window]   Figure 4.   Conditions that protect against relocation of nuclear proteins. Fpr3p was localized 0 and 10 min after hypertonic shock in YPH500 cells carrying either the control plasmid pRS314 or a high-copy plasmid (pPS1739) for expression of Hog1p. In parallel experiments, YPH500 or hsf1-1 (MYY260) cells were incubated for 1 h at 37°C before being returned to 23°C and subjected to hypertonic shock for 0 or 10 min.

When cells are transferred to 37°C, the induction of heat-shock proteins can protect them against subsequent stress (Liu et al., 1996). We therefore inquired whether 1 h of exposure to 37°C would allow cells to withstand subsequent hypertonic shock at room temperature. As shown in Figure 4 (third pair of images), this pretreatment greatly reduced the relocation of Fpr3p in wild-type cells, suggesting that some heat-shock proteins do contribute to protection. To test this hypothesis further, we used an equivalent protocol to study a heat-shock transcription factor mutant, hsf1-1. When this strain was incubated at 37°C for 1 h before being subjected to osmotic shock at 23°C, Fpr3p relocated as in wild-type cells that had not been preincubated at 37°C (Figure 4, right panels). Thus, some consequences of heat shock that depend on heat-shock transcription factor are responsible for the protection.

Functional Consequences of Shock

In protocols equivalent to those that documented relocalization of other nuclear proteins, the localization of at least three nucleoporins, Nsp1p, Nic96p, and Nup49p, remained unchanged (Figure 5). Moreover, nuclear import continued for both Hog1p-GFP and a Mig1p-GFP fusion (Figure 6). Both are known to be imported by importin  family members (De Vit et al., 1997; Ferrigno et al., 1998; Reiser et al., 1999).

View larger version (83K): [in this window] [in a new window]   Figure 5.   Integrity of the nuclear pore complex during hypertonic shock. Cells expressing Nsp1p-GFP (strain YAT502), Nic96p-GFP (strain YAT501), or Nup49p-GFP (strain YPH500 carrying plasmid pSW636) were exposed to hypertonic shock (0.5 M NaCl) for 0 or 10 min at 23°C. Note: the magnification of the cells in these images is twice that in the other figures.

View larger version (93K): [in this window] [in a new window]   Figure 6.   Nuclear import of GFP-tagged proteins. Top, YPH500 cells carrying pPS1739 (HOG1-GFP) were grown overnight in glucose medium to localize Hog1p-GFP to the cytoplasm (control). The cells were then exposed to hypertonic shock for 10 min (NaCl). Bottom, YPH500 cells carrying pBM3495 (encoding Mig1p-GFP--galactosidase) were grown in glucose medium and then diluted into medium containing 5% glycerol for 60 min to cause Mig1p-GFP--galactosidase to concentrate in the cytoplasm (glycerol). Cells were then exposed to hypertonic shock while being supplemented (or not) with 2% glucose to initiate nuclear import. After 10 min, aliquots were washed with 2% glucose in water and examined.

To evaluate nucleolar function during hypertonic shock, we monitored RNA synthesis after exposing cells to 0.5 M NaCl. Incorporation of label into RNA was radically inhibited for >1 h and then gradually returned to normal after several hours (Figure 7). These kinetics of normalization are notably slower than those of protein reimport. This suggests that some yet-unidentified protein returns more slowly, that some component has been degraded and must be resynthesized, or that, on return, the nuclear and nucleolar proteins must "reorganize" before rRNA synthesis resumes.

View larger version (16K): [in this window] [in a new window]   Figure 7.   RNA synthesis during hypertonic shock. YPH500 cells were exposed to hypertonic shock for 0-4 h. Aliquots were pulse-labeled with [3H-methyl]-methionine, and incorporation into RNA was quantitated.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Dramatic transient alterations of nuclear organization occur during hypertonic shock. These changes must accompany standard manipulations of yeast that are widely used, e.g., spheroplasting in the presence of sorbitol. They also raise a set of fundamental questions regarding signaling pathways and the factors that govern the stable organization of the nucleus, which clearly entails the dynamic flux of many macromolecules.

Hypertonic shock quickly causes water efflux across the plasma membrane and corresponding fluxes across intracellular membranes. The resulting reduction of turgor pressure must alter the surface tension of the plasma membrane and therefore affect the relationship of plasma membrane proteins to the cell wall. Hypertonic shock is followed within minutes by a decrease in the permeability of the yeast plasma membrane and by an increase of cytoplasmic solutes (e.g., glycerol), which reduces the initial osmotic insult (Banuett, 1998; Gustin et al., 1998). Hypertonic shock characteristically activates the Hog1p MAP kinase cascade via the plasma membrane proteins Sln1p and Sho1p. This cascade is implicated in many events, including the driving of transcription (e.g., of genes responsible for synthesis of glycerol) and the reorganization of Golgi functions (Reynolds et al., 1998). It is opposed by RAS-cAMP-PKA signaling (Gustin et al., 1998). Strikingly, both Hog1p and one of the factors implicated in STRE-mediated stress responses, Msn2p, quickly concentrate in the nucleus during osmotic shock (Ferrigno et al., 1998; Görner et al., 1998; Reiser et al., 1999).

In contrast, hypotonic shock and heat shock activate the Pkc1p cell integrity MAP kinase cascade, which is implicated in cell wall synthesis, bud morphogenesis, and polarized cell growth. This pathway is also stimulated by mating pheromone and G1 cyclin-Cdc28p. Upstream factors that activate Pkc1p, at least during heat shock, are plasma membrane glycoproteins of the Wsc family and the Rho GTPases (Levin and Bartlett-Heubusch, 1992; Igual et al., 1996; Kamada et al., 1996; Zarzov et al., 1996; Gustin et al., 1998).

Given this background, it is surprising that Pkc1p, not Hog1p, is required for the hypertonic stress-induced relocation of nuclear proteins. Since the kinase activity of Pkc1p is required and since activation of Pkc1p itself can cause relocation, it becomes important to identify the upstream and downstream factors that function in conjunction with Pkc1p. Judging from our observations, the Wsc glycoproteins are not required, but stimulation by diacylglycerol may be involved. In addition, kinases that are part of the Pkc1p MAP kinase cascade (Bck1p, Mpk1p) are not required. We therefore conclude that Pkc1p can stimulate an "alternative signaling pathway." Consistent with this conclusion, other investigators have shown that deletion of Pkc1p has more severe phenotypic consequences than does deletion of individual downstream kinases (Lee and Levin, 1992). Moreover, Pkc1p, but not the downstream kinases, is required both for heat shock-mediated depolarization of the actin cytoskeleton and for attenuation of ribosome biogenesis during treatment of cells with tunicamycin (Delley and Hall, 1999; Li et al., 2000). We observed that HA-tagged Pkc1p is normally distributed throughout the cytoplasm and the nucleus, but that, like many nuclear proteins, it is excluded from the nucleus during hypertonic shock (our unpublished results). Its immediate targets along the alternative pathway may therefore be outside the nucleus. The normally broad distribution of wild-type Pkc1p may nevertheless be a dynamic average, judging from the evidence on the dynamics of Pkc localization in animal cells (Mochly-Rosen and Kauvar, 2000).

Our observations have also indicated a further complexity in the nuclear response to hypertonic shock: excess Hog1p antagonizes those effects of hypertonicity that are mediated by Pkc1p. It is not clear at what level this cross-talk occurs; however, Hog1p activity may normally play a stabilizing role with regard to the integrity of the nucleus.

Relocation of nuclear proteins has also been reported to occur in other situations. For example, slower Pkc1p-dependent relocation occurs when transport along the secretory pathway is inhibited (Nanduri et al., 1999) (our unpublished observations). Pkc1p itself therefore may target key factors that are needed for stable retention of nuclear proteins. Among other conditions that inhibit ribosome biosynthesis are amino acid starvation and treatment with rapamycin (Warner, 1989; Moehle and Hinnebusch, 1991; Neuman-Silberberg et al., 1995; Powers and Walter, 1999); however, these treatments did not cause relocation of Fpr3p over a 2-h period (our unpublished results).