(Received for publication, August 20, 1996, and in revised form, January 16, 1997)
From the Division of Cell Biology, Hospital for Sick
Children, Toronto, Ontario, Canada M5G 1X8 and the ¶ Ontario
Cancer Institute, Princess Margaret Hospital, Toronto,
Ontario, Canada M5G 2M9
Stress-activated protein kinases (SAPK) are stimulated by a variety of agents and conditions that also activate the Na+/H+ exchanger (NHE). Activation of the exchanger results in a rapid increase in intracellular pH (pHi), raising the possibility that cytosolic alkalinization may contribute to SAPK activation. This hypothesis was tested by manipulating the pHi of U937 cells using permeant weak bases. Three different bases increased pHi 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. However, the latter do not appear to mediate the stimulation of SAPK, since neither bafilomycin A1 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+, which were verified to inhibit NHE, did not prevent the osmotically induced activation of SAPK. We conclude that cytosolic alkalinization increases the activity of SAPK and p38 by a calcium-independent mechanism that does not involve acidic intracellular organelles. In addition, even though cell shrinkage is accompanied by alkalinization due to the activation of NHE, the increased pHi is not the main cause of the observed stimulation of SAPK upon hyperosmotic challenge.
Stress-activated protein kinase/c-Jun NH2-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 (1-3). 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+/H+ 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+ concentration and elevate the cytosolic pH (pHi). 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.
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.
AntibodiesAntibodies to SAPK and p38 MAPK were raised in
rabbit against a pGEX vector containing full-length p54 SAPK or
full-length p38, respectively.
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% CO2. For kinase assays, U937 cells were preincubated for 24 h in medium supplemented with only 0.5% fetal calf serum.
Cytosolic pH DeterminationsTo measure pHi, 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 MgSO4, 1.36 CaCl2, and 25 glucose, pH 7.4, at 37 °C, at a density of 2 × 105 cells/ml. This suspension was then incubated with 1 µM 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 pHi by using the high KCl/nigericin technique (23).
Cytosolic Calcium DeterminationsU937 cells at a density of 2 × 105 cells/ml were sedimented, resuspended in NHB, and incubated with 2 µM 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 ([Ca2+]i) was calculated as described previously (24). Briefly, Fmax and Fauto were obtained by adding 5 µM ionomycin and 1 mM MnCl2, respectively, and a dissociation constant of 250 nM for the indo-1-Ca2+ complex (25) was used to calculate [Ca2+]i.
Immunoprecipitation of SAPK and p38Following incubation
under the conditions specified in the text, aliquots of 5 × 106 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 µM Na3VO4, 20 µg/ml leupeptin, 50 mM
NaF2, 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 × 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 µl 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 Na2HPO4, 4 mM
NaH2PO4, and 0.1% Triton X-100.
After immunoprecipitation, the beads were
sedimented and resuspended in 20 µl of kinase buffer containing 50 mM Tris-Cl, 1 mM EGTA, 10 mM
MgCl2, and 100 µM (800 nCi)
[-32P]ATP, pH 7.5. For SAPK kinase assays we added 5 µg 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 µl of 2 × 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+/H+ 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.
All values are reported as the mean ± 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 < 0.05; **, p < 0.01; and ***, p < 0.001).
Many of the
substances and conditions that increase SAPK activity, can also lead to
an increase in pHi 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. pHi was manipulated by means
of weak electrolytes, and the imposed changes were monitored
fluorimetrically using BCECF. As illustrated in Fig.
1A, the pHi of suspended U937 cells
increased rapidly upon addition of 30 mM NH4Cl.
In six similar experiments, pHi rose from a resting value of
7.36 ± 0.04 to 7.76 ± 0.03 within 10 s. The
alkalinization was transient with pHi returning to near-basal
levels within 10 min 7.42 ± 0.03 (Fig. 1, A and
B). This recovery likely reflects gradual entry of
NH4+ via K+ channels and/or
the Na+/K+ pump. We next determined the effect
of this alkalinization on SAPK activity. A comparable exposure of the
cells to 30 mM NH4Cl 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 NH4Cl 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 pHi from 7.43 ± 0.01 to 7.85 ± 0.06 and from 7.41 ± 0.01 to 7.86 ± 0.04 (n = 3), respectively. Unlike the effects of NH4Cl, 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 ActivityChanges in intracellular pH are often accompanied by an
increase in cytosolic calcium concentration
([Ca2+]i). Because it has been previously shown
in T lymphocytes that elevated [Ca2+]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 NH4Cl to those of
calcium ionophores and of thapsigargin, an inhibitor of endomembrane
Ca2+ATPases. By inhibiting pumping into the
endoplasmic reticulum, thapsigargin unmasks an endogenous calcium
"leak" which results in a transient elevation of
[Ca2+]i. The capacitative coupling between
depleted stores and the plasmalemmal calcium channels facilitates
calcium influx, inducing a sustained elevation of
[Ca2+]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 [Ca2+]i from a steady-state
level of 375 ± 30 (n = 10) to 1040 ± 104 nM (n = 5) (Fig. 2, A and
B). Even higher levels of [Ca2+]i were
attained by exposure to 1 µM ionomycin, a non-fluorescent calcium ionophore. The precise [Ca2+]i levels
attained with ionomycin could not be defined, because they exceeded the
dynamic range of the probe used (indo-1, Kd 250 nM). By contrast, [Ca2+]i did not
significantly increase at any time after addition of 30 mM
NH4Cl (after 5 min [Ca2+]i was
238 ± 8 nM (n = 5) (Fig. 2,
A and B)). In fact, exposure to NH4Cl
after [Ca2+]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
[Ca2+]i averaged 260 ± 1 and 343 ± 4 nM, respectively. It therefore appears unlikely that
alkalinization-induced activation of SAPK is mediated by an increase in
[Ca2+]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 [Ca2+]i. Thus, an increase in
[Ca2+]i cannot explain the stimulatory
effects of weak bases on SAPK.
Role of Acidic Endomembrane Compartments in SAPK Activation by NH4Cl
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+-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
NH4Cl 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 µM 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
NH4Cl to activate SAPK.
The role of endomembrane acidic compartments was also evaluated using bafilomycin A1, a potent and very selective inhibitor of vacuolar-type H+-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 > 0.1) effect on SAPK activity (Fig. 3, C and D). In addition, pretreatment with bafilomycin did not preclude the ability of NH4Cl 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+ 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 NH4Cl
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 NH4Cl 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
pHi. As in the case of SAPK, the 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+/H+ Exchanger Responsible for the Shrinkage-induced Activation of SAPK?
In addition to activating
SAPK and p38 MAPK, hyperosmotic treatment also increases pHi.
This cytosolic alkalinization in most cells is mediated by the
shrinkage-induced activation of the Na+/H+
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 ± 0.007 pH unit/min to a new steady-state
pHi of 7.58 ± 0.05 (n = 6). As reported
for other cell types, this shrinkage-induced alkalinization in U937
cells was Na+-dependent and inhibited by
amiloride (Fig. 6A). Indeed, in the absence of external
Na+ or in the presence of the diuretic, pHi became
more acidic, at a rate of 0.02 ± 0.005 (n = 3)
and
0.06 ± 0.05 pH units/min (n = 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+-containing and Na+-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.
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 experiments 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+/H+ 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 G12 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 process 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+/H+ 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+/H+ 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.