The Yeast hnRNP-like Protein Hrp1/Nab4 Accumulates in the Cytoplasm after Hyperosmotic Stress: A Novel Fps1-dependent Response

Plasmids and Strains

Plasmids used in this study are listed in Table 1. The yeast strains used in this study are listed in Table 2. Growth and maintenance of yeast strains, as well as genetic manipulations were performed as described previously (Rose et al., 1990). Liquid sporulation media was prepared as described previously (Kassir and Simchen, 1991). Yeast strains were transformed using the lithium acetate method (Ito et al., 1983) modified according to Gietz et al. (1992). Loss of URA3 plasmids from yeast cells was accomplished by plating on solid media containing 5-fluoroorotic acid (Boeke et al., 1984). Strain MHY705 (kap104-16) was generated by curing strain KAP104/kap104-16 of pRS316-KAP104. Strains MHY834 and MHY836 were constructed by transforming plasmid pHPS13 (LEU2 PKC1) into diploid strain 23710, sporulating the resulting transformants and screening for Gent^r spore clones, which were also Leu^– (MHY835) or Leu⁺ (MHY837). Strain MHY862 was generated by exchanging plasmid pHPS13 (LEU2 PKC1) in strain MHY836 with plasmid p636 (URA3 BCK1-20).

Stress Conditions

Cells were routinely grown in defined synthetic complete medium (SC) supplemented with 2% glucose as a carbon source (Rose et al., 1990). When required, strains harboring plasmids were grown in SC media lacking the appropriate amino acids and/or nutrients. When required, cycloheximide was added at a final concentration of 10 μg/ml. To apply a hyperosmotic shock in liquid medium, cells were pregrown in medium supplemented with 2% glucose to 1 × 10⁷ cells/ml. The cells were then sedimented and resuspended in medium containing either 0.7 M NaCl, 0.7 M KCl, or 1.4 M sorbitol. Samples were taken at the time points indicated in the figures. For hypotonic shock, cultures were diluted with 4 volumes of water. For heat stress, cultures were transferred to a shaking water bath at 42°C. For carbon starvation, cells were grown exponentially in SC supplemented with 2% glucose, galactose, or glycerol where appropriate and were then washed and resuspended in SC media lacking a carbon source. For nitrogen starvation, cells were grown exponentially in SC supplemented with 2% glucose, washed, and then resuspended in media lacking yeast nitrogen base. For oxidative stress, cells were resuspended in media containing 0.4 mM H₂O₂. For the growth of Δpkc1 mutants, all media contained 0.5 M sorbitol to stabilize the cells.

Radiolabeling and Immunoprecipitation

Δhrp1 cells containing the HRP1-Myc plasmid (MHY371) were grown in SC glucose medium lacking leucine and methionine. For labeling, 1000 μCi (90 μl) of EXPRE³⁵S³⁵S protein labeling mix (>1000 Ci/mmol; PerkinElmer Life Sciences) was added to 100 ml of exponential cells (0.5 × 10⁷ cells/ml), and cells were allowed to proceed for 1 h at 30°C. For the t = 0 time point, 20 ml of cells was pelleted, washed twice with ice-cold water, and frozen at –20°C. For samples that were chased after labeling, cells were washed twice in SC glucose medium lacking leucine and were resuspended in the original volume of the same medium containing 0.7 M NaCl. Incubation was continued and 20-ml samples were collected as described above at 5, 30, and 60 min. For immunoprecipitation, frozen cell pellets were resuspended in 200 μl of lysis buffer (50 mM Tris-HCl at pH 7.5, 100 mM NaCl, 5 mM EDTA, 1% Triton X-100) supplemented with protease inhibitors (0.5 mM phenylmethylsulfonyl fluoride and 3 μg/ml each of pepstatin, leupeptin, aprotinin, and chymostatin). Glass beads (50 μl, 0.5 mm) were added and the cells were lysed by repeated vortexing at 4°C (4× for 30 s each). Lysates were clarified by centrifugation for 10 min at 12,000 × g at 4°C and precleared with a 1-h incubation with Sepharose CL-4B beaded agarose (Sigma-Aldrich) at 4°C. Myc–Hrp1p–antibody complexes were precipitated by addition of 20 μl of a 50% slurry of c-Myc (9E10) antibody agarose conjugate (Santa Cruz Biotechnology) to the precleared lysate followed by incubation for 1 h with rocking at 4°C. Beads were washed three times with lysis buffer, and immunoprecipitated proteins were visualized by separation on a 10% SDS-polyacrylamide gel followed by autoradiography.

Preparation of Rabbit Polyclonal anti-Hrp1 Antiserum

A rabbit polyclonal antiserum against Hrp1 was raised using an affinity-purified glutathione S-transferase (GST)-Hrp1 fusion protein. The plasmid expressing GST-Hrp1 (pGEX-HRP1) has been described previously (Kessler et al., 1997). The protein was expressed in BL21 (DE3) cells (Studier et al., 1990) and purified by glutathione affinity chromatography as specified by the supplier (Amersham Biosciences). Immunization of an elite New Zealand White rabbit with this material was performed by a commercial antibody service (Covance Research Products, Denver, PA).

Immunoblot Analysis

Total cell extracts were prepared (Bossie et al., 1992) from cells collected at each time point. Proteins were resolved by SDS-PAGE, transferred, and immunoblotted as described previously (Bossie et al., 1992). The blots were first exposed to anti-Hrp1 antisera at a concentration of 1:5000, followed by horseradish peroxidase (HRP)-conjugated anti-mouse antibodies at 1:5000. Secondary antibodies were detected with enhanced chemiluminescence Western blotting reagents (Amersham Biosciences).

Indirect Immunofluorescence

Indirect immunofluorescence microscopy was performed as described previously (Sadler et al., 1989). Anti-Hrp1 antisera were used at 1:1000.

Hrp1 Rapidly Relocalizes to the Cytoplasm in Response to Osmotic Stress

When yeast cells are grown continuously in either rich (YEPD) or SC media, Hrp1p/Nab4p is abundant throughout the nucleoplasm and is faintly visible in the cytoplasm at 25–36°C, as judged by indirect immunofluorescence (Figure 1A; Henry et al., 1996; Lee et al., 1996). We know however, that under these conditions Hrp1p is capable of exiting and reentering the nucleus (Lee et al., 1996; Shen et al., 1998). Thus, under steady-state conditions, Hrp1p seems to localize to the nucleus because its rate of reimport into the nucleus after export is quite rapid (Figure 1B). During standard growth conditions the export of Hrp1p is thought to be coupled in some manner to the export of poly(A)⁺ RNA (Kessler et al., 1997; Shen et al., 1998; Valentini et al., 1999).

The ability of Saccharomyces cerevisiae to respond to stress is highly regulated (reviewed in Mager and De Kruijff, 1995; Estruch, 2000). Given this tight regulation, it was of interest to analyze the shuttling dynamics and localization of Hrp1 under different types of stress conditions. Accordingly, yeast cells were exposed to heat, salt, peroxide, and carbon or nitrogen starvation (see MATERIALS AND METHODS), and the intracellular localization of Hrp1p was determined by indirect immunofluorescence with anti-Hrp1p antibodies. Of the stress conditions tested, only exposure to high salt (0.7 M) was found to significantly impact the localization of Hrp1 in exponentially growing cells (Figure 2A). After exposure to 0.7 M NaCl, we found that Hrp1p rapidly accumulated in the cytoplasm. Relocalization was complete within 5 min (Figure 2A, top row). It should be noted that a weak cytoplasmic accumulation of Hrp1p was sometimes also observed after glucose depletion (our unpublished data). However, this level was significantly less than that observed after salt exposure. Together, these results led us to postulate that the appearance of Hrp1p in the cytoplasm was specific to osmotic shock rather than to a general stress response.

To confirm that osmotic shock, per se, is responsible for the relocation of Hrp1p, we repeated these experiments with 0.7 M KCl and 1 M sorbitol. Both had effects comparable with those of 0.7 M NaCl (our unpublished data), but strong hypotonic shock did not result in Hrp1p relocalization (Figure 2B). Furthermore, the nuclear export of Hrp1p is an active process that can be inhibited if cells are incubated at 4°C after salt exposure (Figure 2A, bottom row). Hrp1p was found to remain intact throughout the experiment as determined by immunoblot analysis (Figure 2C).

Because new Hrp1p protein is being synthesized during the course of the experiment, the origin of the Hrp1p that relocalizes to the cytoplasm is uncertain. It may either be preexisting cytoplasmic Hrp1p, newly synthesized Hrp1p, or a combination of both. Two experiments were used to answer this question. In the first, a pulse-chase experiment was designed to determine the stability of protein synthesized before high-osmolarity exposure throughout the experiment. The cytoplasmic accumulation of Hrp1 after exposure to osmotic stress conditions is not only due to new protein synthesis. Hrp1p remains stable throughout the incubation period of the experiment; therefore, the cytoplasmic signal is at least partially due to export of preexisting Hrp1p from the nucleus (Figure 2D). In the second experiment, the cytoplasmic accumulation of Hrp1 was deduced to be independent of new protein synthesis because it was unaffected by addition of the protein translational inhibitor cycloheximide (our unpublished data).

Steady-State Location of Hrp1p Remains Cytoplasmic during High Osmolarity

Interestingly, the steady-state location of Hrp1p remains cytoplasmic with continued exposure to high osmolarity. Once relocated, Hrp1p remained in the cytoplasm at 15, 30, and 60 min after salt exposure (Figure 2A). In fact, Hrp1p remained cytoplasmic in cells grown continuously (>10 generations) in 0.7 M NaCl (Figure 2A). This result is in contrast to the transient relocation observed for several nuclear and nucleolar proteins after hypertonic shock (Krebber et al., 1999; Nanduri and Tartakoff, 2001). As a control, we tested a protein reported to exhibit transitory relocation after hyperosmotic shock, the hnRNP-like protein Npl3. Consistent with previous reports, Npl3p reentered the nucleus within 1 h after hyperosmotic shock (our unpublished data; Krebber et al., 1999; Nanduri and Tartakoff, 2001). In sum, these results indicate that Hrp1p remains in the cytoplasm during hyperosmotic conditions.

Cytoplasmic Hrp1p Relocalizes to the Nucleus after the Removal of Hyperosmotic Conditions

Because hyperosmotic shock triggers Hrp1p relocalization to the cytoplasm, we predicted that a return to standard osmotic conditions would result in the nuclear accumulation of Hrp1p. To test this hypothesis, cells grown in 0.7 M NaCl were washed, resuspended in salt-free media, and examined over an extended period. The intracellular localization of Hrp1p was determined by immunofluorescence with anti-Hrp1 antibodies. As expected, Hrp1 relocalized to the nucleus within 5 min of stress removal and remained nuclear at all subsequent time points tested (Figure 3). Similar results were seen in the presence of the protein translational inhibitor cycloheximide, indicating new protein synthesis is not required (our unpublished data).

Nuclear Protein Export and RNA Synthesis

Because the nuclear export of Hrp1p requires ongoing RNA synthesis during standard osmolarity conditions (Shen et al., 1998), we tested whether the same was true during hyperosmotic shock. To accomplish this, we used a temperature-sensitive allele of RNA Pol II termed rpb1-1 (Nonet et al., 1987). In the rpb1-1 strain, RNA Pol II transcription is inhibited at the nonpermissive temperature. After a 30-min shift to the restrictive temperature, rpb1-1 cells were exposed to high osmolarity (0.7 M NaCl), and the localization of Hrp1p was determined by immunofluorescence with anti-Hrp1 antibodies. At both the permissive and restrictive temperatures for Rpb1 protein function, Hrp1p had relocalized to the cytoplasm within the first examined time point (15 min) (Figure 4, left). These results indicate that RNA Pol II transcription is not required for the export of Hrp1p from the nucleus during hyperosmotic shock.

We also examined the effect of a complete RNA synthesis block on RNA export. To accomplish this, we used the antifungal agent thiolutin, an inhibitor of all three RNA polymerases in yeast (Tipper, 1973). When cells are exposed to 0.7 M NaCl in the presence of thiolutin, Hrp1 rapidly relocalizes to the cytoplasm at a rate equivalent to that of cells lacking thiolutin (Figure 4, middle). This result indicates that after hyperosmotic shock the export of Hrp1 is no longer coupled to RNA synthesis.

Effect of Methylation on the Nuclear Export of Hrp1p after Exposure to Hyperosmotic Conditions

Because arginine methylation facilitates the nuclear export of Hrp1p during standard osmolarity conditions (Shen et al., 1998), we tested whether the same was true during hyperosmotic shock. As shown in Figure 4 (right), when ΔHMT1 cells are exposed to high osmolarity (0.7 M NaCl), Hrp1 rapidly relocalizes to the cytoplasm at a rate equivalent to that of wild-type cells (Figure 3, middle). Furthermore, the nuclear accumulation of Hrp1 after a return to standard osmolarity was also found to be independent of arginine methylation. When a ΔHMT1 strain initially grown at high osmolarity was washed and resuspended in salt-free media, cytoplasmic Hrp1 relocalizes to the nucleus within 5 min (our unpublished data). Together, these data demonstrate that cellular Hmt1 arginine methyltransferase activity does not effect Hrp1p localization after hyperosmotic exposure.

Hrp1 Relocalization by Hyperosmotic Stress Does Not Occur through the HOG Pathway

To properly adapt to hyperosmotic conditions, the cell has to sense osmotic changes and transmit the signal to the nucleus. Many, if not all, eukaryotic cells use mitogen-activated protein (MAP) kinase pathways for this purpose. The primary osmosensing MAP kinase pathway in S. cerevisiae is the HOG pathway outlined in Figure 5A (reviewed in O'Rourke et al., 2002). Because this pathway is stimulated by hypertonic shock, we predicted that it would be required for the cytoplasmic accumulation of Hrp1p. To test this hypothesis, we examined the effects of 0.7 M NaCl on the localization of Hrp1p in strains deleted for key genes of the HOG1 pathway. Unexpectedly, Hrp1p still relocalized to the cytoplasm in a strain lacking either 1) the HOG1 gene itself, 2) the SHO1 and SLN1 transmembrane osmosensors, or 3) four putative terminal transcription factors (MSN1, MSN2, MSN3, and HOT1) (Figure 6A). Strains harboring deletions of other HOG pathway genes (STE11 and PBS2), yielded similar results (our unpublished data). In sum, these data indicate that the cytoplasmic accumulation of Hrp1p after hyperosmotic stress is independent of the HOG MAP kinase cascade.

To complete our analysis of the HOG pathway we also tested whether these factors were required to reimport Hrp1p into the nucleus after a return to standard osmolarity. As noted in Figure 6A (bottom), when HOG-pathway mutants initially grown at high osmolarity are washed and resuspended in salt-free media, cytoplasmic Hrp1 relocalizes to the nucleus within 5 min. These data demonstrate that the nuclear reimport of Hrp1p after a shift from high to standard osmolarity is also independent of the HOG pathway.

Hyperosmotic Localization of Hrp1 Is Pkc1p Independent

The PKC pathway (Figure 5B) is a second osmosensing pathway in yeast that has been shown to affect expression of genes after an osmotic shift (Davenport et al., 1995). Interestingly, Nanduri and Tartakoff, 2001 have recently reported that Pkc1p is required for the transient relocation of several nuclear proteins after hyperosmotic shock. Because Hrp1p was not examined in this study, we monitored its localization in a Δpkc1 strain after an osmotic shift. Hrp1p rapidly relocalized to the cytoplasm in both a Δpkc1 strain (Figure 6B, left) and a Δpkc1 strain harboring wild-type PKC1 on a plasmid (Figure 6B, right). In these experiments, cells were initially grown in 0.5 M sorbitol and then shifted to 1.4 M sorbitol. The starting concentration of sorbitol (0.5 M) provided the osmotic support required for Δpkc1 cell viability yet did not induce relocalization (Figure 6B, top). The final sorbitol concentration (1.4 M) chosen closely mimics the osmolarity provided by 0.7 M NaCl used in previous experiments. This data suggests that Hrp1p relocalization does not require Pkc1p function.

Pathway Components Upstream and Downstream of Pkc1p Are Not Required for the Hyperosmotic Relocalization of Hrp1

Although Pkc1p seemed not to be required for the cytoplasmic accumulation of Hrp1p, it was possible that downstream kinases (refer to Figure 5B) could play a role independent of Pkc1p. However, isogenic strains lacking either Bck1p, Mkk1p, Mkk2p, or Slt2p showed Hrp1p relocalization after hypertonic shock (our unpublished data). Furthermore, in these mutants cytoplasmic Hrp1p reentered the nucleus at a rate equivalent to wild-type cells after salt removal (our unpublished data). Thus, the kinase cascade downstream of Pkc1p is also not necessary for the hyperosmotic localization of Hrp1p.

The absence of required mediators downstream of Pkc1p allowed us to reexamine the role of Pkc1p in the absence of any complications attributable to the presence of osmotic support. To accomplish this, we used a constitutively active allele of BCK1 termed BCK1–20 (Lee and Levin, 1992). In a BCK1-20 Δpck1 double mutant strain, partial suppression of the Δpkc1 allele permits growth in the absence of osmotic stabilizers. Both wild-type and BCK1-20 Δpck1 cells were exposed to high osmolarity (0.7 M NaCl), and the localization of Hrp1p was determined by indirect immunofluorescence with anti-Hrp1 antibodies. In both strains, Hrp1p had relocalized to the cytoplasm by the first time point examined (30 min) (Figure 6C, left). These results confirm our previous results that Hrp1p relocalization does not require Pkc1p function.

We also tested upstream components of the Pkc1p pathway (Figure 5B). Several putative cells membrane sensors have been linked to the Pkc1p pathway. These include Slg1p/Wsc1 (Gray et al., 1997; Jacoby et al., 1997; Verna et al., 1997), Wsc2p-Wsc4p (Verna et al., 1997), and Mid2p (Ono et al., 1994). During shifts from standard to high osmolarity and back to standard osmolarity, the localization of Hrp1p in these cells was equivalent to that of wild-type cells (our unpublished data). Deletion of additional proteins linked to the Pkc1p pathway (Bni1p, Fks1p, Rho2p, Rom2p, and Skn7p) also did not affect Hrp1p relocalization (our unpublished data).

Hyperosmotic Localization of Hrp1 Requires the Glycerol Export Protein Fps1

A universal strategy for cellular osmoregulation is the accumulation of a compatible solute to control the osmolarity of the cytosol (Yancey et al., 1982). In yeast, the compatible solute primarily used is glycerol (Brown, 1978). The accumulation of glycerol is controlled by both the rapid closure of the glycerol channel Fps1p and elevated de novo synthesis of glycerol (Figure 7A). In wild-type cells, the rate of glycerol export is decreased by hyperosmotic shock and is increased by hypoosmotic shock (Luyten et al., 1995; Sutherland et al., 1997). Interestingly, although the glycerol transport rate is rapidly regulated by external osmolarity, it seems to be independent of the HOG and PKC signaling pathways (Tamas et al., 1999). Given the unique role of Fps1p in osmoregulation, we inquired whether this protein might also be required for the hyperosmotic relocalization of Hrp1p. Indeed, as shown in Figure 7B (first column), a strain deleted for FPS1 was unable to relocate Hrp1p from the nucleus to the cytoplasm after hypertonic shock. The hyperosmotic-dependent localization of Hrp1p was restored in this strain when transformed with a plasmid encoding wild-type FPS1 (Figure 7B, second column).

Given that Δfps1 and wild-type strains are reported to accumulate equivalent amounts of glycerol after osmotic shock (Luyten et al., 1995; Tamas et al., 1999), we reasoned that Hrp1p relocalization was not linked to intracellular glycerol levels. To test this hypothesis, we determined whether increasing the intracellular glycerol levels of cells grown at standard osmolarity could trigger Hrp1p relocalization. Glycerol levels were increased by overexpression of Gdp1p. High-copy expression of Gdp1p in a wild-type strain leads to a twofold increase in glycerol content compared with a wild-type strain carrying a control plasmid. (Philips and Herskowitz, 1997; Tao et al., 1999). As shown in Figure 7B (fourth column), Hrp1p is not relocalized to the cytoplasm in cells grown at standard osmolarity, although they contain twofold more intracellular glycerol than wild-type cells.

As a second test, we examined a strain expressing an N-terminal truncation of Fps1p (Δ13–230) in which glycerol export is constitutive. This truncation results in an inability to accumulate significant glycerol levels within 3 h after a shift to high osmolarity (Tamas et al., 1999). Thus, although intracellular levels are significantly decreased in this strain, Hrp1p relocalization was unaffected (Figure 7B, third column). In sum, these results indicate that though Hrp1 relocalization requires Fps1p function, it is independent of intracellular glycerol levels. Furthermore, the data suggests that the amino terminal portion of Fps1p (aa 13–230) is not required for osmolarity-dependent Hrp1 relocalization.

Hrp1 Export after Hypertonic Stress Is Crm1/Xpo1 Dependent

We next investigated the nuclear export pathway of Hrp1p after hyperosmotic shock. Within the amino-terminal 60 aa of Hrp1, aa 42–52 (LAALQALSSSL) are predicted to form an amphipathic helix with leucine residues predominantly on one face. Such a structure is characteristic of a leucine-rich nuclear export signal (Rittinger et al., 1999), which acts as the ligand for the export receptor Crm1p/Xpo1p (Fornerod et al., 1997; Stade et al., 1997). This prompted us to focus on the Crm1p export receptor. To test for the involvement of this receptor, we monitored Hrp1 localization in the crm1 (xpo1-1) mutant (Stade et al., 1997). After a 30-min shift to the restrictive temperature, cells were exposed to high osmolarity (0.7 M NaCl) and the localization of Hrp1p was determined by indirect immunofluorescence with anti-Hrp1 antibodies. Under these conditions, Hrp1p remained in the nucleus in the majority of the cells at all time points tested (Figure 8). This result is not attributable to a lack of protein synthesis because, when the assay was performed in the presence or absence of the protein synthesis inhibitor cycloheximide, Hrp1p accumulated in the cytoplasm at equal levels (our unpublished data). In contrast, Hrp1p rapidly relocalized to the cytoplasm in strains lacking either the Los1p or Msn5p exportins. (Figure 8; Arts et al., 1998; Hellmuth et al., 1998; Kutay et al., 1998; Blondel et al., 1999; DeVit and Johnston, 1999). In sum, these results suggest that after exposure to high osmolarity Hrp1p, export from the nucleus requires functional Crm1p/Xpo1p.

Cytoplasmic Accumulation of Hrp1 Is Consistent with Reduced Nuclear Reimport

One explanation for the cytoplasmic accumulation of Hrp1p after hyperosmotic stress is that its rate of nuclear import is simply decreased relative to its rate of export. Consistent with this idea, Hrp1p rapidly relocalizes to the cytoplasm in a strain carrying a temperature-sensitive allele of its import receptor (KAP104) when it is shifted to the restrictive temperature (Aitchison et al., 1996; Figure 9A). We reasoned that if the cytoplasmic accumulation of Hrp1p is due to reduced Kap104p-mediated import, then other nuclear proteins that share this receptor would also shift to the cytoplasm after hyperosmotic shock. The Kap104 receptor is known to mediate the nuclear import of only one other protein; the Nab2 RNA-binding protein (Aitchison et al., 1996). Thus, we compared the cellular localization of both Nab2p and Hrp1p during a shift from standard osmolarity, to high osmolarity (0.7 M NaCl), and back to normal osmolarity. At the time points examined, the localization of Nab2p was equivalent to that previously observed for Hrp1p (Figure 9B). Furthermore, Nab2 reentered the nucleus within 5 min after the salt removal (Figure 9B). Three additional nuclear/nucleolar proteins tested (Npl3p, Nop1p, and Nsr1p) whose import is Kap104p-independent did not share this relocalization pattern (our unpublished data). Thus, these results are consistent with the idea that the cytoplasmic accumulation of the Hrp1 and Nab2 proteins could both be controlled by reduced Kap104p-mediated import.

To examine this further, we treated crm1 (xpo1-1) cells with 0.7 M NaCl for 15 min to induce an increase in cytoplasmic Hrp1p, raised the temperature, and then monitored Hrp1 localization at 15, 60, and 120 min (Figure 9C). As previously reported in this work, Hrp1p nuclear export is blocked in crm1 (xpo1-1) cells grown at the restrictive temperature. Thus, if the cytoplasmic accumulation of Hrp1p after hyperosmotic stress results from an increase its nuclear export, we expected Hrp1p to become trapped in the nucleus of crm1 (xpo1-1) cells. However, if Hrp1p cytoplasmic accumulation is due to a reduction or inhibition of Hrp1p nuclear import the protein would remain cytoplasmic. As shown in Figure 9C, the Hrp1 protein remains cytoplasmic after a shift of the crm1 (xpo1-1) cells to the restrictive temperature. This result indicates that Hrp1p nuclear import is either completely or partially blocked during conditions of high osmolarity.