Cytosolic Alkalinization Increases Stress-activated Protein Kinase/c-Jun NH 2 -terminal Kinase (SAPK/JNK) Activity and p38 Mitogen-activated Protein Kinase Activity by a Calcium-independent Mechanism*

Stress-activated protein kinases (SAPK) are stimulated by a variety of agents and conditions that also activate the Na (cid:49) /H (cid:49) exchanger (NHE). Activation of the exchanger results in a rapid increase in intracellular pH (pH i ), rais- ing the possibility that cytosolic alkalinization may con-tribute to SAPK activation. This hypothesis was tested by manipulating the pH i of U937 cells using permeant weak bases. Three different bases increased pH i and caused a 4–12-fold increase in SAPK activity with a time course that paralleled intracellular alkalinization. p38, a related stress kinase, was also stimulated by the weak bases. Stimulation of the stress kinases was not accompanied by changes in cytosolic free calcium nor was the activation of SAPK achieved when calcium was elevated by thapsigargin or calcium ionophores. Weak bases not only alter the pH of the cytosol but also alkalinize endomembrane compartments such as endosomes and lysosomes. How- ever, the latter do not appear to mediate the stimulation of SAPK, since neither bafilomycin A 1 nor desipramine, agents that neutralize acidic endomembrane compartments, activated the kinase. Because hyperosmolarity acutely activates the NHE, we considered whether the resulting cytosolic alkalinization mediates the activation of SAPK upon cell shrinkage. The addition of amiloride or the omission of Na (cid:49) temper-ature.

Stress-activated protein kinase/c-Jun NH 2 -terminal kinase (SAPK/JNK) 1 and p38 mitogen-activated protein kinase (p38 MAPK) are members of a family of enzymes that are activated by a variety of agents and conditions that generate cellular stress. These kinases are members of two parallel, yet independent cascades, with distinct upstream activators and downstream targets. SAPK is activated by SEK (MKK4), which is in turn stimulated by MEKK or MLK3 (1)(2)(3). Similarly, p38 MAPK is activated by MKK3 and MKK6 (4,5). The effector pathways triggered by the stress kinases also differ: the preferred substrate of SAPK is c-Jun, a component of the transcription regulator AP-1 (6,7), whereas MAPKAPK-2, a serine/ threonine kinase, and the transcription factor ATF-2 are main targets of p38 MAPK (8,9).
The agonists and conditions that activate SAPK and/or p38 are remarkably varied. They include hyperosmolarity, inflammatory cytokines, heat shock, ultraviolet light, and protein synthesis inhibitors such as cycloheximide and anisomycin (for review, see Refs. 10 and 11). Several of these stimuli also activate the Na ϩ /H ϩ exchanger (NHE), a ubiquitous family of transmembrane proteins involved in the regulation of cytosolic pH, cellular volume, and transepithelial ion transport (for review, see Refs. 12 and 13). For example, hyperosmotic exposure, in most cells, leads to the rapid activation of the NHE (14). Heat shock, another activator of SAPK, also increases NHE activity in Vero cells (15). Moreover, cytokines, including interleukin-1 and TNF␣, activate NHE in myocytes (16) and fibroblasts (17), respectively. Ischemia/reperfusion, which activates SAPK (18) and enhances binding of ATF-2 and c-Jun to DNA (19), has likewise been associated with increased transport by the NHE (20). Finally, cycloheximide, a protein synthesis inhibitor that stimulates SAPK, has been shown to increase NHE activity as well (21).
It is unclear whether the activation of the stress kinases and the stimulation of the ion exchanger are related. The kinases may mediate the stimulation of the antiporter, although direct phosphorylation has been ruled out as the mechanism of NHE activation (22). Conversely, the ionic changes initiated by the exchanger may trigger kinase activation. Enhanced NHE activity is expected to increase the intracellular Na ϩ concentration and elevate the cytosolic pH (pH i ). These parameters could in turn activate SAPK and/or p38 MAPK. In this work, we examined the relationship between the cytosolic pH and the activation of the stress kinases. In particular, we analyzed the effects of alkalinization on SAPK and p38 MAPK and considered the possibility that shrinkage-induced activation of NHE is required for activation of the kinases.

EXPERIMENTAL PROCEDURES
Materials-Radiolabeled ATP was purchased from Mandel/Dupont (Guelph, Ontario, Canada). Fetal bovine serum was purchased from Life Technologies, Inc., and the cell culture medium was prepared by the Media Department at Princess Margaret Hospital (Toronto, Ontario, Canada). BCECF, indo-1, nigericin, and ionomycin were purchased from Molecular Probes (Eugene, OR). All other materials were purchased from Sigma.
Antibodies-Antibodies to SAPK and p38 MAPK were raised in rabbit against a pGEX vector containing full-length p54␣ SAPK or fulllength p38, respectively.
Cell Culture-U937 cells (American Type Culture Collection, Bethesda, MD) were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum in a humidified environment under 5% CO 2 . For kinase assays, U937 cells were preincubated for 24 h in medium supplemented with only 0.5% fetal calf serum.
Cytosolic pH Determinations-To measure pH i , U937 cells were sedimented and resuspended in a Na-HEPES-buffered solution (NHB) containing (in mM): 117 NaCl, 25 Na-HEPES, 5.36 KCl, 1.66 MgSO 4 , 1.36 CaCl 2 , and 25 glucose, pH 7.4, at 37°C, at a density of 2 ϫ 10 5 cells/ml. This suspension was then incubated with 1 M acetoxymethyl ester form of BCECF for 15 min at room temperature. Cells were sedimented, resuspended in fresh NHB, placed into a polystyrene cuvette, and inserted into the thermally regulated (37°C) holder of a Perkin-Elmer 650-40 fluorescence spectrophotometer. BCECF was continually excited at 495 nm, and emission was recorded at 525 nm. For each experiment, fluorescence emission was calibrated internally versus pH i by using the high KCl/nigericin technique (23).
Cytosolic Calcium Determinations-U937 cells at a density of 2 ϫ 10 5 cells/ml were sedimented, resuspended in NHB, and incubated with 2 M acetoxymethyl ester precursor of indo-1 for 25 min at room temperature. Cells were once again sedimented, resuspended in NHB, and incubated for an additional 15 min in NHB at 37°C. Indo-1 fluorescence was monitored as described above, with the excitation at 331 nm and the emission recorded at 410 nm. Free cytosolic calcium ([Ca 2ϩ ] i ) was calculated as described previously (24). Briefly, F max and F auto were obtained by adding 5 M ionomycin and 1 mM MnCl 2 , respectively, and a dissociation constant of 250 nM for the indo-1-Ca 2ϩ complex (25) was used to calculate [Ca 2ϩ ] i .
Immunoprecipitation of SAPK and p38 -Following incubation under the conditions specified in the text, aliquots of 5 ϫ 10 6 U937 cells were lysed in hypotonic lysis buffer containing 10 mM NaCl, 20 mM PIPES, pH 7.0, 5 mM EDTA, 0.5% Nonidet P-40, 0.05% ␤-mercaptoethanol, 0.1 mM phenylmethylsulfonyl fluoride, 100 M Na 3 VO 4 , 20 g/ml leupeptin, 50 mM NaF 2 , and 1 mM benzamidine. After 20 min, the lysate was sheared through a 23-gauge needle, and insoluble material was removed by centrifugation at 21,000 ϫ g for 10 min. Samples were normalized for the amount of protein, and SAPK and p38 were immunoprecipitated from the lysates by incubating for 1 h at 4°C with anti-SAPK (1:500 dilution) or anti-p38 (1:250) antibodies, respectively. Immune complexes were collected by adding 20 l of protein A-Sepharose beads to the lysate and incubating for an additional 30 min at 4°C. The beads were then washed four times with an ice-cold solution which contained 150 mM NaCl, 16 mM Na 2 HPO 4 , 4 mM NaH 2 PO 4 , and 0.1% Triton X-100.
Kinase Assays-After immunoprecipitation, the beads were sedimented and resuspended in 20 l of kinase buffer containing 50 mM Tris-Cl, 1 mM EGTA, 10 mM MgCl 2 , and 100 M (800 nCi) [␥-32 P]ATP, pH 7.5. For SAPK kinase assays we added 5 g of a fusion protein of glutathione S-transferase with residues 5-89 of c-Jun and incubated the samples at 30°C for 30 min. The kinase reaction was terminated by adding 40 l of 2 ϫ Laemmli's sample buffer. Samples were resolved by SDS-polyacrylamide gel electrophoresis on 10% acrylamide gels, stained with Coomassie Blue, destained, and dried. Autoradiography was performed using Mandel/Dupont Reflection film, and radioactivity was quantified by PhosphorImager analysis using ImageQuant (Molecular Dynamics). Kinase assays for p38 were performed similarly, except that a glutathione S-transferase fusion protein comprising the 178 carboxyl-terminal amino acids of the NHE-1 isoform of the Na ϩ /H ϩ exchanger was used as a substrate. The construct encoding this fusion protein was the kind gift of Dr. L. Fliegel (University of Alberta, Edmonton, Alberta, Canada). Equal loading of the immunoprecipitated kinases was confirmed in some of the experiments by immunoblotting with either anti-SAPK or anti-p38 antibodies. To resolve SAPK, which co-migrates with the heavy immunoglobulin chain, the samples were not reduced to avoid dissociation of the IgG complex.
Statistical Analyses-All values are reported as the mean Ϯ S.E. of the number of experiments specified. Statistical differences between the control and individual experimental conditions were evaluated using Student's t test. Levels of significance are indicated in the figures (*, p Ͻ 0.05; **, p Ͻ 0.01; and ***, p Ͻ 0.001).

RESULTS
Increasing Intracellular pH Activates SAPK-Many of the substances and conditions that increase SAPK activity, can also lead to an increase in pH i by activating NHE (see the introduction). It was therefore of interest to establish whether these events are causally related. To this end, we studied the effects of cytosolic alkalinization on SAPK activity. pH i was manipulated by means of weak electrolytes, and the imposed changes were monitored fluorimetrically using BCECF. As illustrated in Fig. 1A, the pH i of suspended U937 cells increased rapidly upon addition of 30 mM NH 4 Cl. In six similar experiments, pH i rose from a resting value of 7.36 Ϯ 0.04 to 7.76 Ϯ 0.03 within 10 s. The alkalinization was transient with pH i returning to near-basal levels within 10 min 7.42 Ϯ 0.03 ( Fig.  1, A and B). This recovery likely reflects gradual entry of NH 4 ϩ via K ϩ channels and/or the Na ϩ /K ϩ pump. We next determined the effect of this alkalinization on SAPK activity. A comparable exposure of the cells to 30 mM NH 4 Cl resulted in a reproducible 4-fold increase in SAPK activity (Fig. 1C), which was noticeable as early as 5 min and was sustained for up to 30 min (Fig. 1, C and D). Treatment with the weak base did not alter the efficiency of SAPK immunoprecipitation (Fig. 1C, lower panel).
The data described above cannot discern whether NH 4 Cl exerts its stimulatory effect on SAPK by alkalinizing the cytosol or by other means. To ascertain the mechanism of activation, we compared the effects of two other weak bases, namely trimethylamine (TMA) and triethylamine (TrEA). As shown in Fig. 1, A and B, these bases also induced a brisk increase in pH i from 7.43 Ϯ 0.01 to 7.85 Ϯ 0.06 and from 7.41 Ϯ 0.01 to 7.86 Ϯ 0.04 (n ϭ 3), respectively. Unlike the effects of NH 4 Cl, however, the alkalinization induced by TMA and TrEA persisted after 10 min (Fig. 1, A and B). The much slower decay of the alkalinization reflects the slower permeation of the protonated forms of the organic amines. Both TMA and TrEA also activated SAPK (Fig. 1C). In fact, the stimulation of the kinase was greater and progressed over the course of the experiment. After a 5-min incubation in either TMA or TrEA, SAPK activity increased 6-fold (Fig. 1D). By 10 min, the stimulation was 8-fold, and it was nearly 10-fold by 30 min. The more pronounced activation of the kinase parallels the sustained alkalinization of the cytosol. Jointly, these results suggest that SAPK is responsive to changes in the cytosolic pH, regardless of the agent used to impose them.
The effect of the weak bases was not osmotic in nature, since care was taken to maintain the total osmolarity constant by reducing the content of NaCl when the bases were added. In fact, the cell volume (measured electronically using the Coulter Channelyzer) was not detectably reduced even when the bases were added on top of the normal osmotic complement of the medium, likely because the permeation of the weak base increased the osmotic content of the cells, which compensated at least in part for the increased extracellular osmolarity.
Role of Calcium in the Alkalinization-induced Increase In SAPK Activity-Changes in intracellular pH are often accompanied by an increase in cytosolic calcium concentration ([Ca 2ϩ ] i ). Because it has been previously shown in T lymphocytes that elevated [Ca 2ϩ ] i in conjunction with TPA stimulation activates SAPK (26), we evaluated the role of this divalent cation in SAPK activation by alkalosis. For this purpose, we compared the effects of NH 4 Cl to those of calcium ionophores and of thapsigargin, an inhibitor of endomembrane Ca 2ϩ -ATPases. By inhibiting pumping into the endoplasmic reticulum, thapsigargin unmasks an endogenous calcium "leak" which results in a transient elevation of [Ca 2ϩ ] i . The capacitative coupling between depleted stores and the plasmalemmal calcium channels facilitates calcium influx, inducing a sustained elevation of [Ca 2ϩ ] i when cells are suspended in calcium-containing media (27). These phenomena were readily reproduced in U937, as shown in Fig. 2A. When exposed to 30 nM thapsigargin, U937 cells responded with a large and sustained rise in [Ca 2ϩ ] i from a steady-state level of 375 Ϯ 30 (n ϭ 10) to 1040 Ϯ 104 nM (n ϭ 5) (Fig. 2, A and B). Even higher levels of [Ca 2ϩ ] i were attained by exposure to 1 M ionomycin, a nonfluorescent calcium ionophore. The precise [Ca 2ϩ ] i levels attained with ionomycin could not be defined, because they exceeded the dynamic range of the probe used (indo-1, K d 250 nM). By contrast, [Ca 2ϩ ] i did not significantly increase at any time after addition of 30 mM NH 4 Cl (after 5 min [Ca 2ϩ ] i was 238 Ϯ 8 nM (n ϭ 5) (Fig. 2, A and B)). In fact, exposure to NH 4 Cl after [Ca 2ϩ ] i had been elevated by thapsigargin resulted in a sizable decrease in the cytosolic concentration of the cation. After 5 min of incubation with TMA or TrEA [Ca 2ϩ ] i averaged 260 Ϯ 1 and 343 Ϯ 4 nM, respectively. It therefore appears unlikely that alkalinization-induced activation of SAPK is mediated by an increase in [Ca 2ϩ ] i . This notion was confirmed by comparing the effects of the weak base on SAPK with those elicited by thapsigargin or the calcium ionophores. Thapsigargin, ionomycin, and A23187 produced only a modest stimulation of SAPK up to 10 min after addition, much smaller than that induced by anisomycin (Fig. 1, C and D). The effects of these calcium mobilizing agents are considerably smaller than those of the weak bases (cf. Fig. 1), despite the much greater effects of the former on [Ca 2ϩ ] i . Thus, an increase in [Ca 2ϩ ] i cannot explain the stimulatory effects of weak bases on SAPK.
Role of Acidic Endomembrane Compartments in SAPK Activation by NH 4 Cl-Exposure of cells to permeating weak bases will alkalinize not only the cytoplasm but also intracellular compartments, particularly those that are maintained at an acidic pH by vacuolar H ϩ -ATPases. It is therefore possible that the stimulatory effect of the weak bases on SAPK is mediated by a pH change within an endomembrane compartment. This is particularly relevant since Verheij et al. (28) showed that TNF␣ activates SAPK via ceramide, which is generated by hydrolysis of sphingomyelin within both neutral and acidic compartments. Therefore, it was conceivable that NH 4 Cl induced activation of SAPK through modulation of sphingomyelinase activity in an acidic compartment. Two approaches were used to test this possibility. First, we preincubated cells with desipramine, which has previously been shown to inhibit acidic sphingomyelinase by neutralizing the acidic compartment (29). As illustrated in Fig. 3, A and B, preincubation with 10 M desipramine for 1 h had no effect on SAPK activity, but obliterated the ability of TNF␣ to activate SAPK. In contrast, desipramine had no effect on the ability of NH 4 Cl to activate SAPK.
The role of endomembrane acidic compartments was also evaluated using bafilomycin A 1 , a potent and very selective inhibitor of vacuolar-type H ϩ -ATPases. This inhibitor permeates into the cells, reaches the ATPases of intracellular compartments, and thereby dissipates their pH gradients, while affecting the cytosolic pH minimally. Unlike the weak bases, treatment with 100 nM bafilomycin for up to 1 h had only a marginal statistically insignificant (p Ͼ 0.1) effect on SAPK

FIG. 1. Weak bases increase pH i and activate SAPK in U937 cells.
A and B, the intracellular pH of U937 cells was monitored using BCECF as described under "Experimental Procedures." A representative experiment is shown in A. Cells were suspended in NHB and, where indicated by the arrow, 30 mM NH 4 Cl or trimethylammonium chloride was added. A summary of multiple experiments using NH 4 (n ϭ 6), TMA (n ϭ 3), and triethylammonium (TrEA; n ϭ 3) is shown in B. C and D, U937 cells were pre-equilibrated for 1 h in NHB, then exposed to 10 g/ml anisomycin for 30 min or to 30 mM NH 4 , TMA, or TrEA for the indicated times. SAPK was subsequently immunoprecipitated, and its activity was determined using c-Jun as a substrate, as detailed under "Experimental Procedures." A representative radiogram is shown in the top panel of C, while the average of multiple experiments, quantified by phosphorimaging, is presented in D. The bottom panel in C is a SAPK immunoblot, confirming that comparable amounts of the kinase were immunoprecipitated in each instance. Non-reducing conditions were used to avoid interference with the heavy immunoglobulin chain. Data in B and D are means Ϯ S.E. of the number of experiments indicated. In this and all subsequent figures, the significance of the difference versus the control is indicated (*, p Ͻ 0.05; **, p Ͻ 0.01; and ***, p Ͻ 0.001). activity (Fig. 3, C and D). In addition, pretreatment with bafilomycin did not preclude the ability of NH 4 Cl to activate SAPK. Therefore, neutralization of acidic endomembrane compartments is not likely the mechanism whereby weak bases activate SAPK.
Changes in the pH of intracellular compartments are similarly unlikely to play a role in the activation of SAPK effected osmotically or by anisomycin. This conclusion was derived from the experiments in Fig. 4. Neither desipramine nor bafilomycin, at concentrations known to inhibit acidic sphingomyelinase and the H ϩ pump, respectively, had a significant inhibitory effect on the activation of SAPK by anisomycin or by hypertonic sorbitol (Fig. 4, A-D).
Cytosolic Alkalinization Also Activates p38 MAPK-p38 MAPK, a homolog of the yeast Hog1 protein, is also a member of the stress-activated protein kinase family (30). Like SAPK, p38 MAPK is activated by anisomycin, hyperosmolarity, and the cytokines interleukin-1 and TNF␣ (10). We therefore considered the possibility that, like SAPK, p38 MAPK could also be activated by changes in cytosolic pH. Thus, U937 cells were exposed to either NH 4 Cl or TMA, and the activity of immunoprecipitated p38 MAPK was assessed in vitro using as a substrate the carboxyl-terminal domain of NHE-1, which we had found earlier to be effectively phosphorylated by this kinase. 2 As illustrated in Fig. 5, exposure to NH 4 Cl resulted in a 5-fold increase in p38 MAPK activity detectable within 5 min and maintained through 30 min. Similar results were obtained with the organic base TMA, suggesting that, like SAPK, p38 MAPK is responsive to changes in pH i . As in the case of SAPK, the

FIG. 2. Role of calcium in the alkalinization-induced activation of SAPK.
A and B, cytosolic calcium determinations. Cells were suspended in NHB, and intracellular calcium was determined fluorimetrically using indo-1. Where indicated, the cells were treated with 30 nM thapsigargin (TG) followed by 30 mM NH 4 Cl (upper trace) or with NH 4 Cl alone (lower trace). A representative experiment is shown in A and summarized data from multiple experiments are presented in B. C and D, U937 cells were pre-equilibrated for 1 h in NHB and then exposed to anisomycin (10 g/ml, 30 min), A23187 (1 g/ml, 30 min), ionomycin (Iono; 2 M, 30 min) or thapsigargin (TG; 30 nM, 30 min). SAPK was subsequently immunoprecipitated, and its activity was determined using c-Jun as a substrate, as in  3. Role of endomembrane compartments in the alkalinization-induced activation of SAPK. U937 cells were pre-equilibrated for 1 h in NHB, then exposed to the agents indicated. SAPK was subsequently immunoprecipitated, and its activity was determined using c-Jun as a substrate, as in Fig. 1. Representative radiograms are shown in A and C, while the average of multiple experiments, quantified by phosphorimaging, is presented in B and D. Data in B and D are means Ϯ S.E. of the number of experiments indicated. A and B, cells were preincubated for 1 h with or without 10 M desipramine (Des). SAPK was next stimulated with either TNF␣ (100 ng/ml for 30 min) or with NH 4 Cl (30 mM) for the specified times. C and D, cells were preincubated for 1 h with or without 100 nM bafilomycin A 1 (Baf). SAPK was next stimulated with NH 4 Cl (30 mM) for the specified times. weak bases did not affect the efficiency of p38 immunoprecipitation (lower panel in Fig. 5A).
Is the Alkalinization Resulting from Shrinkage-induced Activation of the Na ϩ /H ϩ Exchanger Responsible for the Shrinkage-induced Activation of SAPK?-In addition to activating SAPK and p38 MAPK, hyperosmotic treatment also increases pH i . This cytosolic alkalinization in most cells is mediated by the shrinkage-induced activation of the Na ϩ /H ϩ exchanger (NHE). Since our present data show that alkalinization suffices to activate SAPK as well as p38 MAPK, we entertained the possibility that shrinkage-induced alkalinization, mediated by the NHE, is responsible for the observed activation of the kinases. When exposed to a hyperosmotic solution, U937 cells underwent a rapid intracellular alkalinization (Fig. 6A). pH increased at a rate of 0.05 Ϯ 0.007 pH unit/min to a new steady-state pH i of 7.58 Ϯ 0.05 (n ϭ 6). As reported for other cell types, this shrinkage-induced alkalinization in U937 cells was Na ϩ -dependent and inhibited by amiloride (Fig. 6A). Indeed, in the absence of external Na ϩ or in the presence of the diuretic, pH i became more acidic, at a rate of Ϫ0.02 Ϯ 0.005 (n ϭ 3) and Ϫ0.06 Ϯ 0.05 pH units/min (n ϭ 3), respectively. These findings confirm that cell shrinkage activates the NHE in U937 cells.
We then tested whether the alkalinization generated by the antiporter is responsible for stimulation of SAPK. Cells were stimulated with hypertonic sorbitol, and the extent of SAPK activation was tested in otherwise untreated cells or under conditions shown above to preclude antiporter-mediated alkalinization of the cytosol. As shown in Fig. 6B, SAPK was comparably activated in Na ϩ -containing and Na ϩ -free media, in the presence and absence of amiloride. Summarized data from multiple experiments are presented in Fig. 6C. We conclude that, while alkalinization alone can activate SAPK and p38 MAPK, osmotic stimulation of the NHE is not responsible for the activation of SAPK. DISCUSSION Because a variety of stimuli concomitantly activate SAPK and the NHE, we considered the possibility that these events are related. Our data indicate that a cytosolic alkalinization of a magnitude comparable to that attained by stimulating the antiport suffices to activate SAPK and p38 MAPK. While these data are suggestive of a causal relationship, subsequent exper- FIG. 5. NH 4 Cl and TMA activate p38 in U937 cells. U937 cells were pre-equilibrated for 1 h in NHB and then exposed to sorbitol (NHB ϩ 400 mM for 30 min) or to 30 mM either NH 4 Cl and TMA for the indicated times. p38 was subsequently immunoprecipitated, and its activity was determined using a fusion of glutathione S-transferase with the carboxyl-terminal domain of NHE-1 (fpNHE) as a substrate. A representative radiogram is shown in the top panel of A, while the average of three experiments, quantified by phosphorimaging, is presented in B. The bottom panel in A is a p38 immunoblot, confirming that comparable amounts of the kinase were immunoprecipitated in each instance. The data in B are means Ϯ S.E.

FIG. 4. Role of endomembrane compartments in the activation of SAPK induced by anisomycin and hyperosmolarity.
U937 cells were pre-equilibrated for 1 h in NHB, then exposed to the agents indicated. SAPK was subsequently immunoprecipitated, and its activity was determined using c-Jun as a substrate, as in Fig. 1. Representative radiograms are shown in A and C, while the average of multiple experiments, quantified by phosphorimaging, is presented in B and D. Data in B and D are means Ϯ S.E. of the number of experiments indicated. A and B, cells were preincubated for 1 h with or without 10 M desipramine (Des). SAPK was next stimulated with either anisomycin (10 g/ml) or sorbitol (400 mM) for 30 min. C and D, cells were preincubated for 1 h with or without 100 nM bafilomycin A 1 (Baf). SAPK was next stimulated with anisomycin or sorbitol as in A.
iments demonstrated that activation of the kinases occurs even when NHE-induced alkalinization is precluded pharmacologically or by ionic substitution. These findings rule out that the activation of SAPK and p38 MAPK is secondary to the activation of Na ϩ /H ϩ exchange. The converse relationship, namely that the NHE is stimulated by a pathway involving the stress kinases, remains a viable possibility. Alternatively, the two events may lie on parallel pathways, which could conceivably share common upstream elements. In this regard, independent studies have shown that members of the Rho family of small GTP-binding proteins can stimulate SAPK (2, 31) as well as NHE activity (32). Cdc42 has been found to activate Rac which in turn can activate Rho (33). These GTP-binding proteins regulate the formation of filopodia, lamellipodia, and stress fibers, and it is noteworthy that NHE-1, the "housekeeping" isoform of the antiporter, has been reported to accumulate at or near these structures (34). Hence, it is possible that activation of Cdc42, Rac, and/or Rho promotes the interaction between the NHE and the cytoskeleton, thereby increasing antiport activity, as well as activating SAPK pathways.
Alternatively, heterotrimeric G proteins could be the common step leading to the parallel activation of NHE and the stress kinases. Prasad et al. (35) demonstrated that constitutively active GTPase-deficient mutants of G␣ 12 and G␣ 13 promote the activation of SAPK. Interestingly, G␣ 13 has also been shown to activate the NHE (36), seemingly via pathways involving small GTP-binding proteins of the Rho family and MEKK1 (32). Thus, G␣ 13 may give rise to the coordinate, yet independent activation of SAPK and NHE by cell shrinkage or other stimulants.
Parallel yet independent activation of the stress kinases and of the NHE is also suggested by the diverging time courses of these events in cells challenged with hypertonic solutions: cation exchange is noticeable and attains maximal rate within seconds, while full osmotic activation of SAPK or p38 MAPK is delayed, reaching maximal level tens of minutes after cell shrinkage (see Ref. 37). This temporal disparity suggests that the two responses may have different functional roles in cell volume homeostasis. We speculate that the early response of the antiporter is intended to accomplish the acute phase of regulatory volume increase, an immediate defense against osmotic perturbation. In addition, however, chronic exposure of cells to hyperosmolarity is known to be counterbalanced by a slower accumulation of organic osmolytes (38). The latter proc-ess depends on an increase in biosynthetic enzymes (39) and in the abundance of organic osmolyte transporters (40), which are in turn associated with elevated mRNA levels (41). Therefore, it is conceivable that activation of the stress kinases signals an increase in transcription via activation of c-Jun or ATF-2, to prepare the cell for a prolonged period of hyperosmotic exposure.
In summary, stressful situations seemingly activate both NHE as well as SAPK and p38 MAPK. While alkalinization such as that generated by Na ϩ /H ϩ exchange is capable of stimulating the stress kinases, neither chemical (anisomycin) nor physical stresses (hypertonicity) require a pH change to exert their stimulatory effect on the kinases. Nevertheless, it is conceivable that other situations leading to stimulation of Na ϩ /H ϩ exchange may secondarily lead to activation of the stress kinases. Stimulation of the exchanger can be induced by integrin engagement, activation of mitogenic receptors, and by some hormones, and some of these treatments also result in activation of stress kinases.
Because upon osmotic cell shrinkage stimulation of NHE precedes activation of the kinases, we find it unlikely that SAPK and/or p38 MAPK mediates the stimulation of the antiporter. Instead, we favor the hypothesis that the two events are parallel yet independent responses, perhaps triggered by a common early event such as activation of small or heterotrimeric G proteins. The divergent activation of these pathways may provide the cell with separate complementary responses to the early and sustained phases of stressful perturbations. FIG. 6. Role of the Na ؉ /H ؉ exchanger in the osmotic activation of SAPK. A, cytosolic pH measurements. U937 cells were suspended in NHB with or without 1 mM amiloride (Amil) or in Na ϩ -free medium. Where indicated, the osmolarity of the medium was increased by addition of 250 mM NaCl. B and C, SAPK activity determinations. Cells were pre-equilibrated in NHB with or without amiloride or in Na ϩ -free medium, then challenged osmotically by addition of 400 mM sorbitol. SAPK was subsequently immunoprecipitated, and its activity was determined using c-Jun as a substrate, as in Fig. 1. A representative radiogram is shown in B, and the average of multiple experiments, quantified by phosphorimaging, is presented in C. The data are means Ϯ S.E. of the number of experiments indicated.