Osmotic Stress Inhibits p70/85 S6 Kinase through Activation of a Protein Phosphatase*

While studying the stress regulation of p70/85 S6 kinase (S6K), we observed that anisomycin and UV light stimulated S6K activity, but that sorbitol inactivated S6K. Pretreatment with hyperosmotic stress also prevented the activation of S6K by both 12-O-tetradecanoylphorbol-13-acetate and anisomycin. Comparison of sorbitol and rapamycin revealed that both agents inactivated S6K and caused dephosphorylation of Ser/Thr-Pro sites in the COOH terminus of S6K, including Thr412, a residue essential to S6K regulation, as determined by phospho-specific antibodies. Rapamycin-resistant S6K truncation mutants were similarly resistant to deactivation by sorbitol. Additionally, the PHAS-1 mobility shift, which is sensitive to rapamycin, was also found to be sensitive to osmotic stress. Experiments using the p38 inhibitor SB203580 and dominant negative mutants involving both stress-activated protein kinase/c-Jun NH2-terminal kinase and p38 stress pathways indicated that these pathways are probably not involved in osmotic stress inhibition of S6K. Examining the potential involvement of a phosphatase, we found that sodium pyrophosphate, sodium vanadate, cyclosporin A, tautomycin, and okadaic acid had no effect on osmotic stress inhibition of S6K. However, calyculin A prevented both rapamycin- and sorbitol-mediated deactivation of S6K. Our results suggest that osmotic stress and rapamycin act through a calyculin A-sensitive phosphatase to cause dephosphorylation and deactivation of S6K.

S6K function is also controlled via inhibitory domains at both the NH 2 terminus and COOH terminus of the kinase (12,13). Deletion of these domains confers relative resistance of these truncated kinases to the inhibitory effects of rapamycin. A model that explains these findings is that phosphorylation of S6K negates the inhibitory effect of these domains. Rapamycin might thus promote dephosphorylation events within these domains, a principle first suggested by Ferrari et al. (5).
Upon activation, S6K phosphorylates the S6 protein of the 40S ribosomal subunit (14). Phosphorylated S6 directs the translational machinery toward increased translation of transcripts containing an oligopyrimidine tract (5Ј TOP), such as those encoding ribosomal proteins and elongation factors (15)(16)(17). Thus, the most well understood function of S6K involves a contribution to cell growth by increasing the production of translational machinery. It is possible, however, that S6K also regulates other cell processes such as transcription and cell cycle progression (18,19).
Whereas the activation and function of S6K have become more clear of late, the rapid inactivation of S6K that is induced by rapamycin or amino acid withdrawal is somewhat less well understood (20). Rapamycin functions by binding to its intracellular receptor, FKBP12, and this complex can inhibit the activity of RAFT1/FRAP/mTOR without affecting other kinases such as Akt/PKB or MAP kinases (21)(22)(23)(24)(25). Whereas the rapamycin target RAFT1/FRAP/mTOR is associated with an activity that phosphorylates S6K, rapamycin-mediated deactivation of RAFT1/FRAP/mTOR does not explain the rapid dephosphorylation of S6K (8). It is possible that rapamycin activates a S6K-specific phosphatase, or that continuously active phosphatases reduce S6K phosphate content once the activating kinases are blocked by rapamycin.
While studying the effects of cell stress on the phosphorylation and activation of S6K, we were surprised to observe that whereas the cell stress agents anisomycin and UV light each activated S6K, osmotic stress inactivated the kinase. Consequently, we sought to characterize this method of S6K inhibition. We have found that osmotic stress inactivates S6K through a mechanism other than stimulation of MAP kinase stress cascades. This mechanism appears to be similar to that regulated by rapamycin and involves the activation of a calyculin A-sensitive phosphatase.

MATERIALS AND METHODS
Cell Culture, Cell Treatments, and Transient Transfections-Human embryonic kidney 293 cells and monkey kidney CV1 cells were maintained in Dulbecco's modified Eagle's medium with 10% calf serum (CS) and 1% penicillin-streptomycin. 293 cells (10 6 ) were transiently transfected with pEBG-based vectors (1 g/expression vector) using the calcium phosphate method (26). CV1 cells (10 6 ) were transfected with recombinant vTF7-3 vaccinia virus and co-transformed by lipofection with pTM1-based vectors (1 g/expression vector) as described previously (27)(28)(29). After transfection, the cells were serum-starved or fed as * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
described in the figure legends. The following day, cells were treated with 500 nM anisomycin, 400 mM sorbitol, 1 g/ml TPA, 10 ng/ml rapamycin (Suren Sehgal; Wyeth Ayerst), 20 M SB203580 (Calbiochem), 5 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 100 nM calyculin A, 5 M cyclosporin A, 1 M okadaic acid, 100 M tautomycin (Alexis Biochemicals), or various combinations of each for the incubation times described in the figure legends. All reagents were obtained from Sigma unless otherwise noted. UV light stimulation was performed by removing the media, exposing cells to 160 J/m 2 of energy in a Stratagene UV Stratalinker 1800, and then replacing the media and allowing the cells to recover for 30 min before harvesting.
Expression Vector Constructions and Mutagenesis-Truncation vector mutants of S6K were constructed as described previously (13). Site-directed mutants were constructed using a variation of the megaprimer mutagenesis method as described previously (30). Diagnostic restriction sites were engineered into mutant sites, and mutated sequences were subcloned into unmutagenized plasmids. The 5A mutant and NH 2 -terminal and COOH-terminal deletions of S6K have been described previously (13). The 5A mutant contains five alanine residues substituted for threonine or serine residues located at codons 434, 441, 444, 447, and 452. Amino acid numbering for S6K includes the NH 2terminal 23 amino acids present in the p85 isoform. Akt-encoding expression plasmids were obtained from Phil Tsichlis (Thomas Jefferson University, Philadelphia, PA). SAPK-␤1 and p85-S6K cDNAs were obtained from James Woodgett (Ontario Cancer Institute, Toronto, Canada). MKK6 and p38 clones were obtained from Brent Zanke (Ontario Cancer Institute), and SEK1 cDNA was obtained from Leonard I. Zon (Harvard Children's Hospital, Boston, MA).
Immunoprecipitations and Immunoblotting-Cells (10 6 unless otherwise noted) were lysed on ice for 25 min in MLB. Extracts were clarified in a microcentrifuge at 14 krpm at 4°C for 10 min before immunoprecipitation or Western blotting.
Endogenous S6K, SAPK, and PHAS-1 were immunoprecipitated with 2 g of the polyclonal antibody CWR#23 (recognizing a COOHterminal 12-amino acid peptide), CWR#17 (recognizing full-length p54 SAPK␤1), or P21NR (J. Lawrence; University of Virginia, Charlottesville, VA), respectively, and collected using 20 l of a 50% slurry of protein A-Sepharose (RepliGen). Hemagglutinin-tagged S6K was immunoprecipitated with 2 g of monoclonal antibody 12CA5 and protein A-Sepharose as described above. S6K tagged with the EE epitope, which was derived from polyomavirus middle T antigen, was immunoprecipitated with 50 l of a 50% slurry of EE antibodies (Gernot Walter (University of California, San Diego, CA); now available commercially from Babco (Richmond, CA)) covalently coupled to Affi-Gel 10 (Bio-Rad). GST-S6K was purified by affinity chromatography using 20 l of a 50% slurry of glutathione-agarose beads. All immunoprecipitations took place at 4°C for 1-4 h. Purified complexes were washed four times with 1 ml of lysis buffer MLB (containing 2 M LiCl for S6K precipitates) and washed once (twice for S6K precipitates) with 1 ml of 50 mM Tris-HCl, pH 7.4, and 1 mM dithiothreitol before use in kinase assays.
Endogenous S6K, endogenous PHAS-1, and EE-tagged and hemagglutinin-tagged proteins were detected using the same antibodies used for the immunoprecipitation described above. GST-tagged proteins were detected using polyclonal antibody 29.3. Detection of S6K phosphorylation on Thr 412 , Ser 434 , and Thr 444 /Ser 447 was done using polyclonal phospho-specific antibodies (New England Biolabs). Detection of p38 phosphorylation of Thr 180 /Tyr 182 from MLB lysate supernatants was performed using phospho-specific p38 antibodies (New England Biolabs). Bands were visualized using alkaline phosphatase-conjugated secondary antibodies after SDS-PAGE (10% SDS-PAGE for phospho-S6K blots, 12.5% SDS-PAGE for phospho-p38 blots, and 15% SDS-PAGE for PHAS-1 blots) and transferring to Immobilon-P (Millipore). SAPK Kinase Assays-Washed precipitates were incubated as described above in kinase buffer containing 1 g of GST-Jun (5-73) substrate (James Woodgett) and 10 Ci of [␥-32 P]ATP per reaction. Reactions were stopped as described above, proteins were separated by 12.5% SDS-PAGE, and radiolabeled GST-Jun was quantified with a Packard Imager.

RESULTS
Osmotic Stress Blocks S6K Activity-To investigate stress signaling to S6K, we treated CV1 cells with a variety of cell stressors (Fig. 1A). Whereas all stress stimuli activated endogenous SAPK (bottom panel), only anisomycin and UV light stimulated endogenous S6K activity (middle panel). However, sorbitol did not stimulate S6K; in contrast, it appeared to decrease S6K activity below basal levels. This was apparent in both the loss of S6 phosphorylation activity in S6K immunoprecipitates and the loss of slowly migrating S6K species, which correlate with S6K activity, on immunoblot analysis (top panel). To determine whether osmotic stress could block stimulation of S6K in addition to blocking basal activity, we pretreated 293 cells with sorbitol and then activated S6K with anisomycin or TPA (Fig. 1B). Osmotic stress pretreatment effectively blocked S6K stimulation by these agents to levels below basal activity. Osmotic stress blocked S6K activation even when added 15 min after the stimulus, and other osmotic stressors such as sodium chloride also inhibited S6K activation (data not shown). Thus, whereas chemical and DNA-damaging stress agents activate S6K, osmotic stress prevents S6K activation.
Osmotic Shock Inhibits S6K Similarly to Rapamycin-To further understand how osmotic stress might inhibit S6K, we compared osmotic stress and rapamycin inhibition of S6K. First, CV1 cells were treated with sorbitol or rapamycin, and endogenous S6K activity and mobility shift were analyzed ( Fig.  2A). Both treatments inactivated S6K and caused a similar increase in the S6K mobility shift. Next we studied the kinetics Immunoprecipitated endogenous S6K and SAPK were assayed for in vitro kinase activity and immunoblotted as described under "Materials and Methods." Anisomycin-and UV-treated cells showed increased S6K activity (middle panel) and retarded S6K electrophoretic mobility (top panel), whereas sorbitol-treated cells showed decreased S6K activity and faster S6K electrophoretic mobility. SAPK was activated by all three stress stimuli (bottom panel). These results have also been observed in 293 cells (data not shown). B, 293 cells were starved as described above, pretreated for 30 min with or without 400 mM sorbitol, and then treated for an additional 30 min with 500 nM anisomycin or 1 g/ml TPA. S6K was assayed as described above. Anisomycin and TPA (lanes 2 and 3) stimulated S6K activity, but sorbitol pretreatment prevented this stimulation (lanes 5 and 6) and resulted in levels below the basal activity. of both S6K inactivation and recovery after treatments followed by withdrawal of these drugs. As seen in Fig. 2B, both rapamycin and osmotic stress caused inactivation of endogenous S6K within 5 min after treatment (lanes 2 and 7). However, S6K activity recovered within 1 h after the removal of sorbitol (Fig. 2C, lane 10), whereas S6K remained inactive for up to 2 h after the removal of rapamycin (lane 5). This result might mean that although rapamycin and osmotic shock may share a common mechanism, rapamycin forms a more stable inhibitory association with its target. This is further supported by a recent report that rapamycin dissociates from FKBP12 with a t1 ⁄2 of about 17.5 h (31).
The ability of S6K to recover after the removal of osmotic stress implies that upon return to normal osmolarity, S6K is activated through a basal activation mechanism. We asked whether this kind of S6K activation could be inhibited by rapamycin (Fig. 2D). S6K activity recovered 1 h after the removal of sorbitol from the media (lane 3), but when rapamycin was added upon removal of the sorbitol, S6K activity did not recover (lane 4). This indicates that stimulation of S6K activity upon return to normal osmolarity is a rapamycin-sensitive event.
Site-specific Dephosphorylation of S6K after Osmotic Shock-Inhibition by rapamycin is associated with the removal of phosphates from a number of serine and threonine residues on S6K (4, 5, 32). To determine whether osmotic stress and rapamycin cause dephosphorylation of similar residues, we used phospho-specific S6K antibodies. In the active state, S6K migrated on SDS-PAGE as a series of slow-migrating bands, the slowest of which were detected by S6K phospho-specific antibodies to Ser 434 and Thr 444 /Ser 447 , as indicated by the arrows in lanes 4 and 7, respectively (Fig. 3A). Neither of these antibodies detected the faster-migrating inactive S6K that was treated with rapamycin or sorbitol, indicating that both treatments cause dephosphorylation of these residues. Importantly, both sorbitol and rapamycin treatment caused dephosphorylation of Thr 412 , a residue essential to S6K regulation (4), as detected by a phospho-specific antibody (Fig. 3B, lanes 5 and 6). Consequently, S6K inhibition by sorbitol and rapamycin involves dephosphorylation of at least four of the same residues, implying that they may both act through a similar mechanism.
Osmotic Stress and Rapamycin Inactivate Mutant S6Ks and Target PHAS-1 Similarly-Previous reports have shown that mutation or truncation of S6K can affect its activity and response to rapamycin (13,33). Mutation of five Ser/Thr-Pro sites in the COOH-terminal tail of S6K to Ala-Pro (S6K-5A) has been shown to partially reduce basal S6K activity while maintaining sensitivity to rapamycin. Similarly, sorbitol treatment of this mutant also inhibited its activity (Fig. 4A, lane 4). Truncation of the NH 2 -terminal 77 amino acids (⌬N77) or the NH 2 -terminal 77 amino acids and the COOH-terminal 78 amino acids (⌬N77⌬C434) caused a constitutive increase in S6K activity and relative rapamycin resistance. Likewise, these mutants were also relatively resistant to sorbitol treatment (Fig. 4B,  lanes 6 and 9). Thus, mutations in S6K that alter its response to rapamycin identically alter its response to osmotic stress.

FIG. 2. Osmotic stress and rapamycin inactivate S6K similarly.
A, CV1 monkey kidney cells were grown in 10% CS for 16 h and then treated for 30 min with 400 mM sorbitol or 10 ng/ml rapamycin. S6K was assayed as described in Fig. 1A. Both rapamycin and sorbitol decreased S6K activity and increased its electrophoretic mobility. These results have also been observed in 293 cells (data not shown). B, 293 cells were grown as described above and then treated for various times with sorbitol or rapamycin before assaying S6K activity as described in Fig. 1A. Both rapamycin and sorbitol caused S6K inhibition within 5 min (lanes 2 and 7). C, 293 cells were grown and treated as described for B. Recovery of S6K activity after 30 min of 400 mM sorbitol or 10 ng/ml rapamycin treatment was then measured by replacing the media with fresh Dulbecco's modified Eagle's medium containing 10% CS without sorbitol or rapamycin. S6K activity recovered 1 h after the removal of sorbitol (lane 10) but did not recover within 2 h after the removal of rapamycin (lane 5). D, 293 cells were grown for 18 h in 10% CS and treated for 30 min with or without 400 mM sorbitol. After this treatment, cells were harvested (lanes 1 and 2), or the media were replaced for 60 min with conditioned media (lane 3) or conditioned media containing 10 ng/ml rapamycin (lane 4). Endogenous S6K was assayed as described above. S6K activity that was inhibited by sorbitol (lane 2) recovered 1 h after the removal of sorbitol (lane 3) but did not recover if rapamycin was present (lane 4).
Another rapamycin-regulated protein involved in translational regulation is PHAS-1 (34). Rapamycin causes PHAS-1 dephosphorylation and mobility shift change through a pathway parallel to S6K inhibition. To determine whether osmotic shock affects PHAS-1, we analyzed PHAS-1 mobility shift in 293 cells after rapamycin or sorbitol treatment (Fig. 5). In both cell lysates and PHAS-1 immunoprecipitations, sorbitol caused an even greater PHAS-1 mobility shift (lanes 3 and 6) than rapamycin (lanes 2 and 5), indicating that osmotic stress can also cause dephosphorylation of PHAS-1.
Osmotic Stress Inhibition of S6K Is Not Mediated by the p38 or SAPK Pathways-Two kinase cascades that are activated by osmotic stress are the p38 pathway and the SAPK pathway (35,36). To determine whether activation of these pathways by osmotic stress is a mechanism by which sorbitol inhibits S6K, we co-expressed dominant negative mutants of stress pathway components with S6K and measured S6K activity after osmotic shock (Fig. 6A). Neither p38, MKK6, SAPK, nor SEK dominant negatives prevented the sorbitol inhibition of S6K (lanes 3-6). Additionally, blockade of both stress pathways by co-expression of dominant negative p38 and SAPK with S6K did not curtail sorbitol inhibition of S6K (lane 7). Another means of blocking the p38 pathway is through use of SB203580, a drug that causes deactivation and dephosphorylation of p38 (37). Whereas treatment of CV1 cells with SB203580 did block the phosphorylation of p38 in response to sorbitol (Fig. 6B, bottom  panel, lane 4), it failed to abrogate inhibition of S6K (middle panel, lane 4). Unexpectedly, SB203580 alone partially inhibited S6K activity (lane 2). This may indicate that p38 positively regulates S6K activity or that SB203580 has effects other than inhibiting p38. However, our data suggest that the SAPK and p38 stress pathways are not involved in mediating sorbitol inhibition of S6K.
Calyculin A, but Not Other Phosphatase Inhibitors, Blocks Osmotic Stress Inhibition of S6K-To test the hypothesis that sorbitol inhibits S6K by activating a phosphatase, we pretreated 293 cells with phosphatase inhibitors before osmotic shock. Neither sodium pyrophosphate, sodium orthovanadate (Fig. 7A), nor cyclosporin A (Fig. 7B) restored S6K activity. However, calyculin A, which is a broad specificity phosphatase inhibitor that inhibits PP1 and PP2A equipotently (38,39), prevented both rapamycin and osmotic shock inhibition of S6K (Fig. 7C, lanes 5 and 6). Neither okadaic acid, which is more specific for PP2A, nor tautomycin, which is more specific for  4). B, CV1 cells were transiently transfected with EE-pTM1-S6K-wt, ⌬N77, or ⌬N77⌬C434. After 17 h in 10% CS, cells were treated for 30 min with 400 mM sorbitol (S) or 10 ng/ml rapamycin (R). Transfected S6K was assayed as described under "Materials and Methods." Both ⌬N77 and ⌬N77⌬C434 have increased basal activity (lanes 4 and 7), and this activity is only partially decreased by rapamycin (lane 5) or sorbitol (lanes 6 and 9) treatment, whereas wt activity is completely blocked by these treatments (lanes 2 and 3). The slight decrease in S6K activity in lane 9 was not observed in repetitions of this experiment and is thought to be due to slight differences in the amount of immunoprecipitated S6K for the kinase assay.
FIG. 6. Sorbitol inhibition of S6K is not mediated by MAP kinase stress pathways. A, CV1 cells were transiently transfected with hemagglutinin-pTM1-S6Kwt and with or without EE-pTM1-p38-KR, MKK6-KA, SAPK-KR, SEK-KR, or p38-KR plus SAPK-KR. After 24 h of serum starvation, cells were treated with or without 400 mM sorbitol for 30 min. Transfected S6K was assayed as described under "Materials and Methods." Sorbitol inhibited S6K activity (lane 2), even in the presence of dominant negative mutants of the p38 and/or SAPK pathways (lanes 3-7). B, CV1 cells were grown in 10% CS for 18 h, treated with or without 20 M SB203580 (p38 inhibitor) for 30 min, and then treated with or without 400 mM sorbitol for an additional 30 min. Endogenous S6K was assayed as described in Fig. 1A, and phospho-p38 was visualized as described under "Materials and Methods." Sorbitol inhibited S6K (lane 3), even in the presence of SB203580 (lane 4), which prevented p38 phosphorylation (Phospho-p38 Blot).
PP1, was successful in preventing S6K inhibition by osmotic shock (Fig. 7D, lanes 4 and 5) (40). However, they did cause a slight retardation in S6K mobility on Western blot. To date, we have been unable to effect specific dephosphorylation of S6K with extracts of osmotically shocked cells made in buffer containing 0.1% Nonidet P-40 or 5 mM EDTA (data not shown). This suggests the existence of a calyculin A-sensitive but okadaic acid-and tautomycin-insensitive S6K phosphatase that is either insoluble or unstable in Nonidet P-40 or EDTA.
Akt Stimulation of S6K Is Blocked by Osmotic Shock-Although it is not necessary for all means of S6K activation, Akt/PKB can activate S6K in co-expression experiments and is probably a mediator of S6K activation by stimuli such as insulin that pass through phosphatidylinositol 3-kinase (41,42). Akt activation involves the phosphorylation of Thr 473 , and the mutation of this site to aspartate can increase Akt activity. We asked whether this mutationally activated Akt could stimulate S6K, and whether this stimulation could still be inhibited by sorbitol (Fig. 8). Co-expression of ⌬N60-S473D Akt with S6K stimulated S6K activity 1.6-fold (lane 5) over co-expression with ⌬N60, and stimulation of S6K by both Akt mutants was inhibited by sorbitol treatment (lanes 4 and 6). This suggests that sorbitol blocks S6K through a mechanism other than deactivation of Akt.

DISCUSSION
Our results demonstrate the ability of hyperosmotic shock to inhibit both endogenous and transfected S6K in a manner similar to that of rapamycin. In two different cell types, basal activity as well as TPA-or anisomycin-stimulated S6K activity is inhibited by osmotic stress. Additionally, rapamycin and osmotic shock cause dephosphorylation of similar and crucial regulatory sites on S6K. This behavior is another similarity between osmotic shock and rapamycin. Furthermore, S6K mutants that are relatively resistant to rapamycin are also resistant to sorbitol. Taken together, these data imply that rapamy-cin and osmotic stress operate through common effector mechanisms, possibly through a common phosphatase.
SAPK and p38 pathways do not appear to mediate inhibition of S6K in response to osmotic stress. Dominant negative co-transfections and chemical inhibitors of the p38 pathway have no effect on the osmotic stress inhibition of S6K. Because yeast studies have shown that primary osmosensors can regulate transcription without the use of MAP kinase pathways, it is quite possible that mammalian cells can also regulate osmotically stimulated cellular processes independent of SAPK or p38 pathway activation (35,43). Such osmosensors and their mechanism of regulation of both MAP kinase-dependent and -independent processes remain to be elucidated in mammalian cells.
Our data suggest that both rapamycin and osmotic stress inhibit S6K through stimulation of a calyculin A-sensitive FIG. 8. Akt stimulation of S6K is blocked by osmotic shock. CV1 cells were transiently transfected with hemagglutinin-pTM1-S6Kwt and with or without EE-pTM1-⌬N60-Akt or ⌬N60-T473D-Akt. After 24 h of serum starvation, cells were treated with or without 400 mM sorbitol for 25 min. Transfected S6K was assayed as described under "Materials and Methods." Mutationally activated ⌬N60-T473D-Akt activated S6K (lane 5) 1.6-fold compared with ⌬N60 (lane 3), and S6K activation by both constructs was prevented by sorbitol treatment (lanes 4 and 6). 7. Calyculin A, but not other phosphatase inhibitors, prevents osmotic stress inhibition of S6K. A, 293 cells were grown in 10% CS for 18 h, pretreated for 30 min with or without 5 mM sodium pyrophosphate (PP) or 1 mM sodium orthovanadate (OV), and then treated for an additional 30 min with or without 400 mM sorbitol. Endogenous S6K was assayed as described in Fig. 1A. Sodium pyrophosphate and sodium orthovanadate had little effect on S6K (lanes 2 and 3) and failed to prevent sorbitol inhibition of S6K (lanes 5 and 6). B, 293 cells were grown as described above, pretreated for 30 min with 5 M cyclosporin A (CsA), and then treated for an additional 30 min with or without 400 mM sorbitol and assayed as described above. Cyclosporin A had no effect on basal S6K activity (lane 3) or on S6K inhibition by sorbitol (lane 4). C, 293 cells were grown as described above, pretreated for 25 min with 100 nM calyculin A, and then treated for an additional 30 min with or without 400 mM sorbitol (S) or 10 ng/ml rapamycin (R) and assayed as described above. Calyculin A (lanes 5 and 6) prevented S6K inhibition by rapamycin (lane 2) or sorbitol (lane 3). D, CV1 cells were grown as described above; pretreated with calyculin A (CA) for 25 min, okadaic acid (OA) for 90 min, or tautomycin (TAU) for 4 h and 45 min; and then treated for an additional 20 min with or without 400 mM sorbitol. Endogenous S6K was assayed as described above. Whereas all three phosphatase inhibitors caused morphology changes resulting in rounded cells that fell off the plate (data not shown), only calyculin A (lane 3) prevented S6K inhibition by sorbitol. Although okadaic acid and tautomycin did not prevent S6K inhibition by sorbitol, they did cause a slight retardation in S6K mobility on the Western blot (top panel, lanes 4 and 5).
phosphatase. This is in agreement with others who have shown that calyculin A can stimulate S6K activity and can separately increase PHAS-1 phosphorylation (44,45). Because calyculin A is a broad specificity phosphatase that inhibits both PP2A and PP1 equipotently, we also tested the effects of okadaic acid, which is more specific for PP2A, and tautomycin, which is more specific for PP1, on osmotic stress inhibition of S6K (40). Neither of these more specific phosphatase inhibitors appeared to block sorbitol inhibition of S6K, suggesting the possibility of a separate class of phosphatases or the necessity to block both PP2A and PP1 to effect the reversal of osmotic inhibition of S6K. Two recent reports describe a physical interaction between PP2A and S6K (46,47). Curiously, the report by Westphal et al. (46) failed to observe a definitive okadaic acid effect on S6K activity, and the report by Peterson et al. (47), although documenting a reversal of rapamycin inhibition of S6K by calyculin A, did not show any data regarding okadaic acid effects on S6K. Consequently, speculation exists regarding the true nature of the S6K phosphatase based upon phosphatase inhibitor data.
Recently, calyculin A has been shown to activate Akt and prevent osmotic stress inhibition of Akt (48). The coincident inhibition of both Akt and S6K by sorbitol and the prevention of this inhibition by calyculin A suggest the possibility that sorbitol could inhibit S6K via Akt deactivation. Our data and those in the literature suggest that this is not the case. First, anisomycin and TPA stimulation of S6K can be prevented by sorbitol, but anisomycin and TPA have been reported to have no effect on Akt activity (42,49,50). Furthermore, the use of dominant negative Akt-S308A/T473A suggests that Akt does not participate in stimulation of S6K by agents such as TPA (50), although it may contribute to insulin-or heat shockactivated S6K. In addition, rapamycin and sorbitol have similar inhibitory effects on S6K, but rapamycin does not block Akt activation (22). Finally, we observe that osmotic shock can block S6K activation by the constitutively active Akt mutant ⌬N60-T473D. Consequently, our data and that of others suggest that S6K and Akt are regulated independently by osmotic shock.
Osmotic stress has been known for decades to inhibit the translational machinery (51). Data presented here suggest that translational inhibition by sorbitol may occur in part through the activation of a calyculin A-sensitive phosphatase that inactivates S6K and PHAS-1. The identification of this phosphatase will be important to further elucidate the mechanism of translational inhibition by osmotic stress and the negative regulatory mechanisms of the S6K pathway.