Proline- and alanine-rich Ste20-related kinase associates with F-actin and translocates from the cytosol to cytoskeleton upon cellular stresses.

Proline- and alanine-rich Ste20-related kinase (PASK) is a Ste20-related protein kinase isolated from rat brain. Cell fractionation studies showed that PASK was present both in the cytosol and in Triton X-100-insoluble cytoskeletal fraction in rat tissues. In brain, PASK associated with protein complexes that contained actin and tubulin, confirming the association of PASK with the cytoskeleton in vivo. Glutathione S-transferase-PASK fusion protein cosedimented with F-actin, indicating that PASK binds to F-actin. In contrast to rat tissues, PASK was detected only in the Triton X-100-soluble cytosolic fraction in cultured PC12 and NIH 3T3 cells. Cytosolic PASK translocated to the cytoskeleton when these cells were stimulated with severe cellular stresses such as hypertonic sodium chloride, hydrogen peroxide, and heat shock at 45 degrees C. Our results suggest that PASK may be involved in the regulation of the cytoskeleton in response to cellular stresses such as hyperosmotic shock.

Ste20 is a yeast protein kinase that acts upstream of the pheromone-responsive mitogen-activated protein kinase cascade (1,2). Ste20-related protein kinases, which have a catalytic domain highly homologous to that of Ste20, have been identified in various eukaryotes, and their family is expanding. Members of this family can be divided into two groups based on their structure and regulation. The kinases of the first group, including PAKs 1 and Ste20, have a catalytic domain at the carboxyl terminus and a regulatory domain at the amino terminus, which contains a binding site for Rac1 and Cdc42 (3). The kinases of the second group have a catalytic domain at the amino terminus and a putative regulatory domain at the carboxyl terminus and are structurally related to yeast Sps1. Some of the Sps1 subfamily members, including germinal center kinase, hematopoietic progenitor kinase, Nck-interacting kinase, and kinase homologous to SPS1/STE20, have been shown to activate the stress-activated protein kinase/JNK pathway in transfection experiments (4 -8; for review, see Ref. 9).
In addition to their roles as upstream activators of mitogenactivated protein kinase pathways, members of the Ste20 family have been implicated in the regulation of cytoskeletal reorganization. The small GTPase Rac1 and Cdc42, upstream activators of PAK/Ste20 subfamily members, are considered key regulatory molecules that link surface receptors to the organization of the actin cytoskeleton (10). A Drosophila homolog of PAK, DPAK, has been reported to colocalize with focal adhesion and focal complexes (11). PAK1 in fibroblast cell lines has been shown to translocate from the cytosol to Rac-and Cdc42-dependent actin structures when cells were stimulated with platelet-derived growth factor (12). It has been reported that expression of activated mutants of PAK1 in mammalian cells induces actin reorganization (13,14). For Sps1/germinal center kinase subfamily members, there are no reports concerning their roles in the regulation of cytoskeletal proteins except that Dictyostelium Severin kinase has been shown to phosphorylate actin binding protein severin (15).
We have recently cloned a rat Ste20-related protein kinase of the Sps1/germinal center kinase subfamily and named it PASK (proline-and alanine-rich Ste20-related kinase; Ref. 16). PASK was present in both the cytosol and particulate fraction in the rat brain. The particulate PASK was not solubilized by extraction with Triton X-100, suggesting that it was associated with complexes of high density such as cytoskeleton. In this report, we identified actin and tubulin as the major constituents of PASK-associated proteins in the brain, confirming the association of PASK with the cytoskeleton in vivo. In addition, we examined changes in the subcellular distribution of PASK in cultured PC12 and NIH 3T3 cells in response to extracellular stimuli. We demonstrated that PASK translocates from the cytosol to the cytoskeleton upon stimulation of the cells with severe cellular stresses. PASK is the first example of Sps1/ germinal center kinase subfamily kinase that is shown to associate with the cytoskeleton.

EXPERIMENTAL PROCEDURES
Fractionation of Rat Tissues-All procedures were carried out at 0 -4°C. The experimental protocol was approved by the Committee for Animal Research of Mie University. Tissues from adult Harlan Sprague Dawley rats (13 weeks old) were homogenized with a Polytron in 10 volumes of extraction buffer containing 20 mM Tris-HCl, pH 7.5, 10 mM EDTA, 10 mM 2-mercaptoethanol, and 1 mM p-phenylmethylsulfonyl fluoride (buffer A). To the homogenates, we added 1 volume of buffer A containing 4% Triton X-100. After incubation on ice for 10 min, the homogenates were centrifuged at 10,000 ϫ g for 10 min. The supernatants (Triton X-100-soluble fraction) were saved, and the pellets were washed with buffer A containing 1% Triton X-100 to obtain the Triton X-100-insoluble cytoskeletal fractions.
Preparation of Anti-PASK Antibody-Antibody against the carboxylterminal region of PASK (amino acid residues 424 -553) was prepared as described previously (16). This antibody recognized a protein of 66 kDa on Western blot in COS-7 cells transfected with a full-length PASK cDNA but not in cells transfected with an empty vector. The antibody also detected a single protein of 66 kDa in many rat tissues, indicating that the antibody specifically recognizes PASK. In addition, the antibody immunoprecipitated PASK from the lysate of the transfected cells.
Immunoaffinity Purification of PASK from Rat Brain-Anti-PASK antibody (0.3 mg) was covalently bound to 1 ml of N-hydroxysuccimideactivated Sepharose (HiTrap affinity column; Amersham Pharmacia Biotech) to prepare an anti-PASK immunoaffinity column. Frozen rat brains (4 g) were homogenized with Polytron in 20 ml of extraction buffer containing 20 mM Tris-HCl, pH 7.5, 1 mM EDTA, 10 mM 2-mercaptoethanol, and 1 mM p-phenylmethylsulfonyl fluoride. The homogenate was centrifuged at 30,000 ϫ g for 15 min, and the resulting supernatant (210 mg of protein) was applied to the anti-PASK antibody column. After washing the column with phosphate-buffered saline and 10 mM Tris-HCl, pH 7.5, the bound proteins were eluted with 0.1 M glycine-HCl, pH 2.5.
Immunoprecipitation of PASK from Rat Brain Extract-Rat brain extract (14 mg of protein) prepared as described above was incubated with 10 g of anti-PASK antibody for 1 h at 4°C, and the immune complex was precipitated with 10 l of protein A-Sepharose CL-4B (Amersham Pharmacia Biotech). After washing the gel with phosphatebuffered saline containing 0.1% Nonidet P-40, 10 mM Tris-HCl, pH 7.5, containing 0.5 M NaCl, and 10 mM Tris-HCl, pH 7.5, bound proteins were eluted with 0.1 M glycine-HCl, pH 2.5.
Actin Sedimentation Assay-Actin sedimentation assay was performed according to den Hartigh et al. (18). Briefly, 2 g of GST fusion proteins were incubated with 10 g of rabbit muscle G-actin (Worthington Biochemical, Lakewood, NJ) for 10 min in 200 l of buffer containing 2 mM Tris-HCl, pH 7.4, 0.2 mM CaCl 2 , 0.2 mM dithiothreitol, and 0.5 mM ATP. Actin polymerization was induced by adding 75 mM KCl and 2 mM MgCl 2 , and after 1 h at room temperature, the samples were centrifuged at 100,000 ϫ g for 1 h at 25°C. Proteins in the supernatant and the pellet were separated on a 10% SDS-polyacrylamide gel and analyzed by silver staining. To determine whether PASK could bind to unpolymerized G-actin, 2 g of glutathione-Sepharosebound GST-PASK fusion protein were incubated with 10 g of G-actin under the same conditions as for the cosedimentation assay, except that KCl and MgCl 2 were omitted. After a 1-h incubation at room temperature with gentle agitation, the Sepharose beads were washed three times with the incubation buffer containing 1% Triton X-100 and boiled in SDS-PAGE loading buffer. Proteins were separated on a 10% SDSpolyacrylamide gel and stained with Coomassie brilliant blue.
Isolation of Triton X-100-soluble and -insoluble Fractions of PC12 and NIH 3T3 Cells-PC12 cells grown on poly-D-lysine-coated 60-mm dishes (Becton Dickinson, San Jose, CA) and NIH 3T3 cells grown on normal tissue culture dishes were treated with various stimuli as de-scribed. After stimulation, dishes were immediately placed on ice. The cells were washed twice with ice-cold phosphate-buffered saline and harvested in 250 l of extraction buffer containing 20 mM Tris-HCl, pH 7.4, 1% Triton X-100, 2 mM EDTA, 1 mM p-phenylmethylsulfonyl fluoride, 10 mM 2-mercaptoethanol, 50 mM NaF, 50 mM ␤-glycerophosphate, 10 mM sodium pyrophosphate, 1 mM Na 3 VO 4 , and 120 mM NaCl using a rubber policeman. After incubation on ice for 10 min, the lysates were centrifuged at 10,000 ϫ g for 10 min. The supernatants (Triton X-100soluble fraction) were saved, and the pellets were washed with 0.25 ml of extraction buffer to obtain Triton X-100-insoluble fractions. These fractions were subjected to immunoblot analysis with anti-PASK antibody as described above. The amount of PASK protein in these fractions was determined by densitometric scanning (Sharp JX-330M scanner) of the blot, using NIH Image software (version 1.61). 32 P Labeling of PC12 Cells-PC12 cells grown on poly-D-lysinecoated 60-mm dishes were washed three times with serum-and phosphate-free RPMI 1640 medium (Life Technologies) and incubated in this medium (1.35 ml) containing 0.75 mCi of carrier-free [ 32 P] orthophosphate (ICN, Costa Mesa, CA) for 2 h. Then the cells were stimulated for 1 h with hypertonic NaCl by the addition of 0.15 ml of 5 M NaCl. After stimulation, the cells were washed twice with ice-cold phosphate-free medium and harvested in 250 l of extraction buffer. Triton X-100-soluble supernatants and -insoluble pellets were prepared as described above. PASK immunoprecipitation was performed according to the method of Moore and Sefton (19) with slight modifications. Triton X-100-insoluble pellets were suspended in 50 l of SDS lysis buffer (0.5% SDS, 50 mM Tris-HCl, pH 8.0, and 1 mM dithiothreitol) and boiled for 3 min. To the suspensions, 200 l of radioimmunoprecipitation assay correction buffer (1.25% Nonidet P-40, 1.25% sodium deoxycholate, 12.5 mM sodium phosphate, pH 7.2, 2 mM EDTA, 62.5 mM NaF, 62.5 mM ␤-glycerophosphate, 12.5 mM sodium pyrophosphate, and 1.25 mM Na 3 VO 4 ) were added, and the supernatants were obtained by centrifugation. These supernatants were incubated for 3 h with non-immune rabbit IgG preadsorbed on protein A-Sepharose beads to remove proteins that bind to the resin and IgG nonspecifically and clarified by centrifugation. PASK was immunoprecipitated by incubation with 2.5 g of anti-PASK antibody for 1 h at 4°C followed by protein A-Sepharose precipitation. Sepharose beads were washed five times with immunoprecipitate wash buffer (1% Nonidet P-40, 1% sodium deoxycholate, 0.1% SDS, 0.15 M NaCl, 10 mM sodium phosphate, pH 7.2, 2 mM EDTA, 50 mM NaF, 50 mM ␤-glycerophosphate, 10 mM sodium pyrophosphate, 1 mM Na 3 VO 4 , 1 mM p-phenylmethylsulfonyl fluoride, and 10 mM 2-mercaptoethanol) and boiled in SDS-PAGE loading buffer. Immunoprecipitates were separated on a 10% SDS-polyacrylamide gel and transferred onto a PVDF membrane. Phosphorylation of blotted proteins was quantified by scanning on a Fuji BAS1000 bioimaging analyzer. The same membrane was subjected to immunoblot analysis with anti-PASK antibody.

RESULTS
Subcellular Distribution of PASK-We performed fractionation studies to determine whether PASK associates with the cytoskeleton in rat tissues and cultured PC12 and NIH 3T3 cells. Homogenates of rat brain, adrenal gland, heart, stomach, testis, epididymis, PC12 cells, and NIH 3T3 cells prepared in 2% Triton X-100 were centrifuged at 10,000 ϫ g for 10 min to obtain Triton X-100-soluble and -insoluble fractions. The amount of PASK protein in each fraction was analyzed by immunoblotting with anti-PASK (424 -553) antibody. Consistent with the results of a previous study (16), ϳ50% of PASK in rat brain was recovered in the Triton X-100-insoluble cytoskeletal fraction. As shown in Fig. 1, PASK was detected in the Triton X-100-insoluble fractions of other tissues, although the ratios of PASK protein in the Triton X-100-insoluble fraction to that in the soluble fraction varied considerably among these tissues and were generally smaller than that of brain. In contrast to these tissues, PASK was not detected in the Triton X-100-insoluble fractions of PC12 or NIH 3T3 cells. These results suggested that at least a component of PASK associates with the cytoskeleton in rat tissues but not in PC12 or NIH 3T3 cells.
To confirm that PASK associates with the cytoskeleton, proteins associated with PASK in rat brain were analyzed by immunoaffinity chromatography and immunoprecipitation Ste20-related Kinase PASK and Cytoskeleton with anti-PASK antibody. Brain extracts were applied to the anti-PASK immunoaffinity column. After washing and elution of bound proteins, the adsorbed proteins were analyzed by SDS-PAGE. Fig. 2A, lane 2, shows silver staining of the proteins eluted from the anti-PASK antibody affinity column. A band at 45 kDa and a doublet at 50 kDa were the major proteins copurified with PASK. As shown in Fig. 2B, left panel, these proteins were identified as actin and tubulin by immunoblotting with anti-actin and anti-␣-tubulin antibodies, respectively. Actin and tubulin were also coprecipitated with PASK by anti-PASK antibody and protein A-Sepharose, whereas none of these proteins was precipitated by nonimmune IgG (Fig. 2B, right panel). These results indicate that PASK forms molecular complexes with actin and tubulin.
PASK Associates with F-actin in Vitro-To assess the actinbinding properties of PASK and to determine which region of PASK binds actin, the behavior of various GST-PASK and truncated PASK fusion proteins (Fig. 3A) in an actin sedimentation assay was tested. Rabbit muscle actin was incubated with GST-PASK fusion proteins for 1 h in the presence or absence of KCl and MgCl 2 . Subsequently, the actin filaments were sedimented by centrifugation. Both pellet and supernatant were analyzed by SDS-PAGE. As shown in Fig. 3B, when actin was polymerized by the addition of KCl and MgCl 2 , GST-PASK (full-length), GST-PASK (66 -553), and GST-PASK (245-424) cosedimented with F-actin and were found almost entirely in the pellet. GST-PASK (1-360) also cosedimented with F-actin, although a portion of it remained in the supernatant. GST-PASK (1-72) and GST-PASK (424 -553) did not cosediment with F-actin and were found entirely in the supernatant. When actin was left unpolymerized (Fig. 3B, Ϫ(KCl,  MgCl 2 )), all the fusion proteins remained in the supernatant.
To determine the specificity of the association between PASK and F-actin, control experiments were performed with GST-JNK2 and GST-xylanase (33-547), which are similar in size to GST-PASK. These proteins did not cosediment with F-actin under the assay conditions (Fig. 3B). These results indicated that PASK associates with F-actin in vitro and that PASK interaction with F-actin is mediated by the region within amino acid residues 245-424.
To determine whether PASK could bind to unpolymerized G-actin, GST-PASK and GST bound to glutathione-Sepharose were incubated with G-actin. Sepharose beads were then washed with incubation buffer containing 1% Triton X-100. Sepharose-bound proteins were analyzed by SDS-PAGE. As shown in Fig. 3C, no actin was detected in association with GST-PASK or GST, indicating that PASK did not bind to Gactin under the assay conditions.

Translocation of PASK from the Cytosol to Cytoskeleton Is Induced by Cellular Stresses in PC12 and NIH 3T3
Cells-It is known that certain protein kinases, including PKCs and p160ROCK, interact with cytoskeletal proteins and translocate from the cytosol to cytoskeleton in response to extracellular stimuli (20 -22). We therefore tested whether translocation of PASK from the cytosol to the cytoskeleton was induced upon stimulation of PC12 and NIH 3T3 cells. For this purpose, PASK distribution was analyzed by cell fractionation after treatment of these cells with various stimuli. As shown in Fig. 4, A and B,   FIG. 1. Subcellular distribution of PASK in the brain, adrenal gland (gl.), heart, stomach, testis, epididymis, PC12 cells, and NIH 3T3 cells. Homogenates of rat tissues, PC12 cells, and NIH 3T3 cells were prepared in 2% Triton X-100 and centrifuged at 10,000 ϫ g for 10 min to obtain Triton X-100-soluble (S) and -insoluble (I) fractions. Equal portions of both fractions were subjected to SDS-PAGE, transferred to a PVDF membrane, and probed with anti-PASK antibody.

FIG. 2. Copurification of actin and tubulin with PASK by immunoaffinity chromatography and immunoprecipitation with anti-PASK antibody.
Purification of PASK from rat brain extract by immunoaffinity chromatography was performed as described under "Experimental Procedures." A, portions of the crude brain extract (lane 1) and the affinity-purified proteins (lane 2) were subjected to SDS-PAGE and stained with silver. B, left panel, crude brain extract (lane 1) and purified proteins (lane 2) were resolved by SDS-PAGE as in A and transferred to PVDF membranes, which were probed with anti-PASK, anti-␣-tubulin, or anti-actin antibody. Right panel, rat brain extract was incubated with anti-PASK antibody or nonimmune IgG, and the immune complex was precipitated with protein A-Sepharose as described under "Experimental Procedures." Portions of proteins immunoprecipitated with anti-PASK antibody (lane 3) or nonimmune IgG (lane 4) were subjected to immunoblotting as described above.

Ste20-related Kinase PASK and Cytoskeleton
upper two panels, PASK was detected in the Triton X-100insoluble cytoskeletal fraction when PC12 and NIH 3T3 cells were osmotically stressed with NaCl or sorbitol, oxidatively stressed with H 2 O 2 , or heat-shocked at 45°C. In these cells, the amount of PASK in the Triton X-100-insoluble fraction increased in a time-dependent manner (Fig. 4, C and D). To quantitate the changes in the distribution of PASK, dilution series of the Triton X-100-soluble and -insoluble fractions were resolved in the same gel and subjected to densitometric analysis. In cells stimulated with ϩ0.5 M NaCl for 1 h and in cells treated at 45°C for 1 h, ϳ30 and 80% of total PASK for PC12 cells and 20 and 80% of it for NIH 3T3 cells, respectively, were found in the Triton X-100-insoluble fraction. There was a concomitant stoichiometric decrease in the amount of soluble PASK. In Fig. 4, the reduction of soluble PASK is less obvious than the increase of insoluble PASK, because 5-fold excess amounts of the insoluble fractions were subjected to the immunoblot analysis. These results indicated that PASK translocated from the cytosol to the cytoskeleton in response to cellular stresses. The translocation was unique to PASK and not attributable to an artifact of cell fractionation, because the distribution of JNK1 and JNK2 in these stimulated cells was considerably different from that of PASK (Fig. 4, A and B, lower two panels): there was no significant increase in JNK2 in the insoluble fraction by any of these stresses; JNK1 did not translocate by hypertonic stress in NIH 3T3 cells, although its amount in the insoluble fraction increased by the other stresses.
In contrast to the translocation by these stresses, PASK in PC12 cells did not translocate by hypotonic stress, stimulation The cells were lysed in 1% Triton X-100 to obtain Triton X-100-soluble and -insoluble fractions. Portions of these fractions were subjected to SDS-PAGE followed by immunoblotting with anti-PASK antibody. Anti-JNK1 (fulllength) antibody was used to detect JNK1 and JNK2 (A and B, lower two panels). Relative amounts of the soluble and insoluble fractions subjected to immunoblot analysis were 1:5. Asterisks indicate an unidentified cross-reactive protein distinct from JNK2. G, PC12 cells were incubated in an isotonic medium for the indicated time intervals after 50-min hypertonic (ϩ0.5 M NaCl) treatment. PASK distribution was analyzed as above.
of acetylcholine receptor with 1 mM carbachol, inhibition of protein synthesis with 50 g/ml anisomycin, or exposure of the cells to 50 ng/ml nerve growth factor for 1 h (Fig. 4E and data not shown). Relatively mild stresses such as ϩ0.2 M NaCl, 0.5 mM H 2 O 2 , and heat shock at 42°C also failed to induce PASK translocation ( Fig. 4F and data not shown). These results suggested that PASK translocation was specifically induced by severe cellular stresses. As shown in Fig. 4G, hypertonic NaClinduced PASK translocation was reversible: the amount of PASK in the Triton X-100-insoluble fraction returned to a control level when cells were incubated in an isotonic medium for 4 h after hypertonic stress.
To investigate the mechanism of hypertonic NaCl-induced PASK translocation, we examined whether such a process involved phosphorylation of PASK. PC12 cells were labeled with [ 32 P]P i and stimulated with a hypertonic medium (ϩ0.5 M NaCl) for 1 h. The cells were extracted to obtain Triton X-100soluble and -insoluble fractions. PASK was immunoprecipitated with anti-PASK antibody directly from the Triton X-100soluble fraction and after extraction in 0.5% SDS from the Triton X-100-insoluble fraction. Immunoprecipitates were resolved by SDS-PAGE, transferred to a PVDF membrane, and subjected to autoradiography. The same membrane was probed with anti-PASK antibody to determine the amount of PASK protein in each lane. As shown in Fig. 5, phosphorylated PASK was immunoprecipitated from the soluble fraction of control cells. This indicated that PASK is probably constitutively phosphorylated in PC12 cells. The level of PASK phosphorylation in the Triton X-100-insoluble fraction of the stimulated cells was reduced to ϳ20 and 40% of that in the soluble fraction of the control and the stimulated cells, respectively. In cells stimulated with the hypertonic medium, a labeled band that migrated slightly faster than PASK was detected in the Triton X-100-soluble fraction. This band was not counted, because it was undetectable by immunoblot with anti-PASK antibody. These results indicate that PASK translocation to the cytoskeleton correlates with PASK dephosphorylation.
We examined the effect of two cytoskeletal poisons on hypertonic NaCl-induced PASK translocation. PC12 cells were first treated with 20 M cytochalasin B or 100 M colchicine for 1 h to block the polymerization of actin or tubulin, respectively. This was followed by stimulation of cells with the hypertonic medium (ϩ0.5 M NaCl) for 1 h. As shown in Fig. 6, these inhibitors had no effect on PASK distribution in both the control cells and NaCl-stimulated cells. These results suggest that rearrangement of the cytoskeleton is not involved in PASK translocation.

DISCUSSION
Hydropathy plots of the primary structure of PASK suggest that it is a soluble protein present in the cytosol. However, cell fractionation studies showed that PASK existed not only in the cytosol but also as protein complexes of high density such as the cytoskeleton. Immunoaffinity chromatography and immunoprecipitation using anti-PASK antibody revealed that actin and tubulin were the major constituents of PASK-associated protein complexes. These results indicate that PASK actually associates with the cytoskeleton.
In cosedimentation experiments, PASK was shown to bind F-actin in vitro. However, it is not clear whether PASK binds directly to actin, because it is possible that the actin preparation used in our experiments was contaminated by an adapter protein that binds PASK to actin. In addition, PASK has no known actin binding motifs. Ste20 has been reported to associate through Bem1 protein with actin cytoskeleton (23). It is possible that PASK associates with the cytoskeleton through such an adapter molecule, although further investigation is necessary to identify the cytoskeletal component to which PASK directly binds.
Cell fractionation of rat tissues showed that the distribution of PASK between soluble and cytoskeletal fractions varied considerably depending on the tissue type: ϳ50% of total PASK was present in the Triton X-100-insoluble cytoskeletal fraction of brain and testis, whereas only 10% of total PASK was found in this fraction of adrenal gland and epididymis. PASK was undetectable in the cytoskeletal fraction of PC12 and NIH 3T3 cells. Although the molecular mechanisms that regulate the association of PASK with the cytoskeleton remain to be elucidated, these differences in the distribution of PASK suggest that the interaction between PASK and the cytoskeleton may be regulated by some factor and may reflect different amounts and/or activities of such factors in different tissues.
Our results showed that translocation of PASK from the cytosol to cytoskeleton occurred in PC12 and NIH 3T3 cells when these cells were stimulated with severe cellular stresses. The observation that hypertonic stress-induced PASK translocation was unaffected by cytochalasin B or colchicine treatment suggests that PASK translocation is not induced by rearrangement of the cytoskeleton. Cytosolic PASK probably does not associate with actin or tubulin monomers, because none of these proteins was coimmunoprecipitated with PASK from a 100,000 ϫ g supernatant of brain extract by anti-PASK antibody (data not shown). This is also supported by the observation that GST-PASK did not bind to unpolymerized G-actin in vitro. Therefore, we do not believe that cytosolic PASK associated with actin or tubulin monomers translocates to the cy- toskeleton by polymerization of these cytoskeletal proteins. PASK translocation was also not affected by inhibition of protein synthesis with cycloheximide or anisomycin (data not shown), suggesting that PASK translocation is not induced by de novo synthesis of adapter proteins, which bind PASK to the cytoskeleton. In vivo 32 P labeling of PC12 cells showed that the phosphorylation level of PASK translocated to the cytoskeleton decreased to ϳ40% of that in the cytosol. Thus, it is conceivable that the level of PASK phosphorylation may modulate the affinity of PASK for the target molecule of the cytoskeleton. At this point, it is unclear whether PASK translocation is induced by PASK dephosphorylation. Further experiments using mutants lacking the phosphorylation site(s) will be needed to establish a causal relationship.
It has been reported that Ste20-like oxidant stress response kinase-1 and mammalian sterile twenty-like/kinase regulated by stress, Ste20 family members to which PASK is most closely related, are activated by cellular stresses (24,25). To investigate whether PASK is similarly activated by cellular stresses, PASK was immunoprecipitated from Triton X-100-soluble extracts of hypertonic or oxidant stress-stimulated PC12 cells, and its activity was determined in vitro with myelin basic protein as an exogenous substrate. However, we could not detect any significant changes in PASK activity, because it was very weak under the assay conditions used (data not shown). Taken together with the observation that cytosolic PASK was constitutively phosphorylated, these results suggest that the phosphorylated PASK in the cytosol may be inactive. Mammalian sterile twenty-like-1 has been reported to be activated by its dephosphorylation (26). Although PASK activity in the Triton X-100-insoluble fraction was not determined because of difficulties in immunoprecipitation under nondenaturing conditions, it is possible that only translocated PASK is activated at the cytoskeleton.
PASK is localized to a distinct set of cells, including neurons and transporting epithelia such as epithelial cells of brain choroid plexus, distal tubules, and collecting ducts of the kidney, ducts of salivary gland, and parietal cells of the stomach (16). These PASK-containing cells actively transport electrolytes and water to regulate the osmolarity and ionic composition of body fluids such as cerebrospinal fluid, urine, and saliva. As shown here, cytosolic PASK translocates to the cytoskeleton upon stimulation with hypertonic sodium chloride. It is thus conceivable that PASK may be involved in the regulation of cytoskeleton when these cells respond to changes in osmolarity of body fluids.