(Received for publication, August 23, 1996, and in revised form, October 5, 1996)
From the Centre de Biochimie, CNRS, Université de Nice, Parc Valrose, 06108 Nice Cedex 2, France
The ubiquitously expressed Na+/H+ exchanger NHE1 is the target of multiple signaling pathways, including those activated by tyrosine kinase receptors, G protein-coupled receptors, and integrins. The intracellular pathways leading to activation of NHE1 are poorly understood. To gain more insight into these activation pathways, we examined the role of mitogen-activated protein kinases (MAPKs) as potential mediators of NHE1 activation by extracellular stimuli such as growth factors and hyperosmotic stress.
Whereas p44 MAPK does not appear to phosphorylate NHE1 in
vitro, we found that inhibition of the p42/p44 MAPK signaling by expression of a dominant negative form of p44 MAPK, by expression of
the MAP kinase phosphatase MKP-1, or by inhibition of MAPK kinase 1 (MKK1) with the PD 98059 compound reduced by 50-60% NHE1 activation
in response to growth factors. This inhibitory effect also was observed
in C-terminal NHE1 deletion mutants in which the major phosphorylation
sites have been deleted. Furthermore, the use of a CCL39-derived cell
line expressing an estradiol-regulated form of oncogenic Raf-1
(CCL39-Raf-1:ER) revealed that the exclusive activation of the
Raf
MKK1
p42/p44 MAPK cascade was capable of inducing NHE1
activation to the same extent as potent growth factors like
thrombin.
Together, our findings demonstrate that the p42/p44 MAPK cascade plays a predominant role in the regulation of NHE1 by growth factors, an action that is mediated via accessory proteins that remain to be identified. In contrast, we found no evidence in favor of the contribution of any MAPK, p42/p44, p38 MAPKs, and Jun kinase, in NHE1 activation by osmotic stress.
Na+/H+ exchangers are vital membrane transporters involved in multiple cellular functions. NHE1,1 the first Na+/H+ exchanger isoform to be cloned (1), is ubiquitously expressed and appears to be the predominant species in nonepithelial cells, where it has been shown to play a major role in intracellular pH homeostasis and cell volume regulation. NHE1 activity can be modulated by a remarkably wide variety of stimuli including growth factors, tumor promoters, and hormones as well as physical factors such as changes in cell volume or cell spreading (reviewed in Ref. 2). On the basis of its hydropathy profile, NHE1 exhibits two distinct domains: a largely hydrophobic NH2-terminal region of 500 residues proposed to span the membrane 10-12 times, followed by a hydrophilic C-terminal domain of more than 300 residues. Structure-function relationship studies have demonstrated that the N-terminal transmembrane part of NHE1 is necessary and sufficient to catalyze the amiloride-sensitive ion exchange, whereas the cytoplasmic C-terminal domain is crucial for mediating the activation of NHE1 by growth factors, hormones, and osmotic stress (3).
Extracellular stimuli have been shown to activate NHE1 by increasing its sensitivity to intracellular H+ (4, 5, 6, 7), resulting in cytoplasmic alkalinization; the mechanism underlying this shift in pHi sensitivity is not yet fully elucidated. NHE1 is constitutively phosphorylated in unstimulated cells, and mitogenic stimulation is accompanied by an increase in phosphorylation of NHE1 on serine residues, with a time course similar to the rise in intracellular pH (8). Moreover, okadaic acid, a serine/threonine protein phosphatase inhibitor can by itself trigger activation of NHE1 in correlation with stimulation of its phosphorylation (9, 10). Based on these results, phosphorylation of NHE1 had been hypothesized to directly trigger the activation of the antiporter. Recent evidence, however, shows that phosphorylation of NHE1 cannot fully account for its activation (11). The picture of NHE1 regulation emerging from recent studies is as follows: at least three different mechanisms appear to control NHE1 regulation: 1) the effect of (a) regulatory protein(s), potentially phosphorylated, interacting with the cytoplasmic domain of NHE1; 2) elevation of [Ca2+]i, leading to the binding of Ca2+/calmodulin complex to the high affinity calmodulin-binding domain (amino acids 636-656); and 3) direct phosphorylation of the cytoplasmic domain of NHE1 (11, 12, 13).
The mechanisms by which extracellular signals are propagated from the cell surface to modify membrane targets such as NHE1 have not been completely resolved. The mitogen-activated protein kinases (MAPKs) have been identified as important mediators in a wide array of physiological processes: they constitute a large family of Ser/Thr protein kinases activated by separate cascades conserved through evolution (14), which regulate multiple processes stimulated by extracellular agents. Transmission of the signal occurs through sequential activation of cytosolic protein kinases; eventually, these phosphorylation cascades activate nuclear and cytosolic regulatory molecules to initiate cellular responses. In mammalian cells, the first and best characterized MAPK cascade is the p42/p44 MAPK cascade (15, 16); it involves activation of p42/p44 MAPK by direct phosphorylation by the dual specificity kinases MKK1 and MKK2 (17) which are themselves phosphorylated and activated by the serine/threonine kinase Raf (18). Raf is recruited to the membrane by Ras upon activation of either G protein-coupled receptors (19) or tyrosine kinase receptors (20). Recently, two other MAPK subtypes, p38 MAPK (21, 22, 23, 24) and p46/p54 JNKs (25, 26), have been discovered. The p42/p44 MAPK cascade has been shown to be essential for the propagation of growth factors and differentiating signals (27, 28, 29), whereas p38 MAPK and JNK, also named stress-activated protein kinase or SAPK, mediate signals in response to cytokines and environmental stress such as hyperosmolarity (21, 22, 23, 24, 25, 26). Although these cascades are initiated by stimuli that induce NHE1 activation and might therefore be involved in the signaling pathways leading to activation of the exchanger, reports establishing a direct link between MAPK cascades and regulation of ion transporters such as NHE1 are scarce (30).
The purpose of the present study was therefore to gain a better
understanding of the activation pathways of NHE1 by investigating the
putative role of the p42/p44 MAPK, p38 MAPK, and JNK cascades in the
regulation of NHE1 by growth factors and osmotic stress. We report that
in the Chinese hamster lung fibroblast cell line CCL39 the Raf MKK1
p42/p44 MAPK cascade plays an important role in activation of
NHE1 by growth factors, whereas neither p42/p44 and p38 MAPK nor JNK
appeared to be involved in NHE1 activation by osmotic stress.
Estradiol, horseradish peroxidase-conjugated
anti-rabbit IgG, myelin basic protein, insulin, and anisomycin were
obtained from Sigma; IL1 from Boehringer Mannheim,
protein A-Sepharose CL4B and protein G-Sepharose from Pharmacia
Biotech; and Nikkol (octaethylene glycol mono-n-dodecyl
ether) from Fluka.
-Thrombin was kindly provided by Dr. J. W. Fenton
II (New York State Department of Health, Albany, NY). HOE 694 was
kindly provided by Hoechst Laboratories. [
-32P]ATP and
[32P]orthophosphate were from ICN, and
[7-14C]benzoic acid was from DuPont NEN. The specific p38
MAPK inhibitor SB 203580 was provided by SmithKline Beecham
Pharmaceuticals (King of Prussia, PA). The specific MKK1 inhibitor PD
98059 was from Biolabs. 12CA5 monoclonal antibody that recognizes the
HA epitope was from BABCO (Emeryville, CA); P5D4 monoclonal antibody
against the VSVG epitope (31) was provided by Dr. B. Gould (Institut Pasteur, Paris); 9E10 monoclonal antibody against the Myc epitope was
provided by Dr. G. Evan (Imperial Cancer Research Fund, London). The
affinity-purified polyclonal rabbit anti-NHE1 antibody (RP-c28) raised
against a
-galactosidase fusion protein containing the last 157 residues (amino acids 658-815) of the human NHE1, has already been
described (8).
The parental Chinese hamster lung
fibroblast cell line CCL39, the Na+/H+
antiporter-deficient cell line PS120, and the corresponding
transfectants were maintained in Dulbecco's modified Eagle's medium
(H21 catalog number 52100; Life Technologies, Inc.) containing 25 mM NaHCO3. The CCL39-Raf-1:ER clone derived
from the CCL39 cell line was maintained in H21 medium without phenol
red and supplemented with glutamine and glucose to reach the
concentrations of normal H21 (H21 without phenol red; catalog number
11880). Both culture media were supplemented with 7.5% fetal calf
serum (Life Technologies, Inc.), penicillin (50 units/ml), and
streptomycin (50 µg/ml). Cells were maintained at 37 °C in the
presence of 5% CO2. The CCL39-
Raf-1:ER clonal cell line
was obtained by transfection of CCL39 cells with the plasmid
pLNC
Raf-1:ER (32) and selection of clones resistant to Geneticin
(G418). The clone that displayed the highest stimulation of MAPK
activity upon estradiol addition was selected and recloned (33).
PS120 cells
are Na+/H+ antiporter-deficient mutants derived
from the CCL39 cell line by the H+ suicide method (34). A
series of stable transfectants was obtained by expression of different
constructs of human NHE1 cDNA in PS120 cells. These included the
full-length cDNA deleted from the 5-untranslated region
(pEAP
construct), and C-terminal truncations at positions
698 and 635 (constructs
698 and
635, respectively) (3).
pEAP
plasmid was NH2-terminally-tagged with a
Myc epitope (NHE1-Myc) by Dr. J. Noël; the Myc epitope was
subcloned in the N terminus of
698 (
698-Myc) and
635
(
635-Myc). pEAP
was also
NH2-terminally-tagged with a double VSVG epitope
(NHE1-VSVG) by Dr. R. C. Poole.
The plasmids p44 MAPK and p44 MAPK-T192A are derived from the hamster cDNA of p44 MAPK as described previously (27). The human cDNA of the phosphatase MKP-1 kindly provided by Dr. S. Keyse (35) was subcloned in the expression vector pcDNAneo (Invitrogen, San Diego, CA) by J.-M. Brondello. The plasmid p38 MAPK is derived from murine cDNA of p38 MAPK (36) and was kindly provided by Dr. R. J. Ulevitch; the plasmid JNK was kindly provided by Dr. R. J. Davis.
For stable transfections, PS120 cells were transfected with each plasmid construct by using the calcium phosphate co-precipitation technique. The transfected cells were selected after three consecutive tests of pHi recovery after cytoplasmic acidification as described previously (3).
For transient expression, a H+-killing selection was also
used. PS120 cells (3 × 106/10-cm plate) were
cotransfected by the calcium phosphate technique with 7 µg of
pEAP expression vector encoding the
Na+/H+ antiporter (1, 3) and 43 µg of the
relevant construct. Forty-eight hours after transfection, cells were
subjected to an acid load selection that killed nontransfected cells,
usually >90% of the cell population (1, 3). pHi determination experiments were done the next day in surviving cells.
Changes in the intracellular pH were estimated from the distribution of [7-14C]benzoic acid (37) as described previously (38).
[32P]Orthophosphate Cell Labeling and ImmunoprecipitationThe cells expressing wild type or mutant exchangers were grown to confluence in 10-cm dishes and labeled for 5 h at 37 °C in phosphate-free, serum-free medium containing [32P]orthophosphate (100-300 µCi/ml). The cells were then stimulated in the same medium with 10% FCS for 15 min.
Immunoprecipitation of NHE1 was carried out as described previously (11). For immunoprecipitation with the monoclonal 9E10 antibody, a suspension of protein G-Sepharose beads was used instead of protein A-Sepharose beads. Immunoprecipitated proteins were solubilized by boiling in Laemmli sample buffer. Samples were analyzed by SDS-PAGE on 7.5% polyacrylamide gels. Phosphoproteins were visualized by autoradiography.
Immune Complex Kinase AssaysCells were serum-deprived
overnight or for 5 h and stimulated with various agonists. Kinase
assays were performed as described previously (40). Briefly, the cells
were lysed in Triton X-100 lysis buffer. Equal amounts of proteins from
cell lysates were immunoprecipitated on protein A-Sepharose beads
coupled with the 12CA5 anti-HA antibody. Activity of the kinases was
assayed in 40 µl of kinase buffer with various substrates:
glutathione S-transferase-ATF2-(1-109) for HA-p38MAPK and
HA-JNK; MBP for HA-p44MAPK; and 50 µM, 1-3 µCi
[-32P]ATP for 20 min at 30 °C. For the immune
complex kinase assays using immunoprecipitated NHE1 as substrate, after
washing with the Triton X-100 lysis buffer, the beads coming from
HA-p44MAPK and pEAP
or NHE1-VSVG immune complex were
mixed and then washed with kinase buffer. HA-p44 MAPK activity was
assayed by incubating the mixture of beads in 40 µl of kinase buffer
containing 5 µM, 1-3 µCi [
-32P]ATP
for 30 min at 30 °C. The reactions were stopped by Laemmli sample
buffer. The samples were heated at 95 °C for 5 min, protein was
separated by SDS-PAGE, and phosphoproteins were detected by autoradiography.
Western blotting analysis of NHE1 was performed as described previously (3).
Other MethodsProtein concentration was measured using the bicinchoninic acid protein assay kit (Pierce) with bovine serum albumin as standard.
Autoradiograms are representative of three separate experiments. Data are presented as the mean ± S.E. of at least three independent experiments performed in quadruplicate.
In Vivo Phosphorylation of NHE1 and NHE1 Deletion Mutants: Effect of Serum
The cytoplasmic domain of NHE1, which has been determined by
mutagenesis studies to be crucial for activation by growth factors (3),
has also been found to contain both the basal and stimulus-induced phosphorylation sites (11). To clarify the contribution of direct phosphorylation of NHE1 in the response to growth factors, we expressed
in PS120 cells (a Na+/H+ exchanger-deficient
derivative of the CCL39 cell line (34)) wild type and C-terminally
truncated forms of the human NHE1. The behavior of these truncated
mutants has been previously characterized as follows (38): deletion of
117 amino acids (698-815) of the cytosolic tail was found to largely
preserve the cytoplasmic alkalinization induced by the potent mitogenic
combination of -thrombin and insulin (
pHi = 0.15 ± 0.01, n = 10, for the
698 cells compared with
pHi = 0.22 ± 0.03, n = 9, for the
pEAP
cells). A larger deletion of the cytosolic tail of
180 amino acids (635-815) still preserved the alkalinization induced
by
-thrombin and insulin, although the extent of the pHi increase was reduced by 50% by the deletion (
pHi = 0.10 ± 0.02; n = 6).
We investigated the phosphorylation status of these NHE1 truncation
mutants both in quiescent cells and upon stimulation with growth
factors. The specific anti-NHE1 polyclonal antibody RP-c28 raised
against a -galactosidase fusion protein containing the last
C-terminal 157 amino acids of the human NHE1 (658-815) (8), was found
to recognize only the wild type exchanger pEAP
and not
the
698 and
635 deletions (38). To allow immunoprecipitation of
the NHE1 deletions, a Myc epitope tag recognized by the anti-Myc monoclonal antibody 9E10 was introduced at the N terminus of these cytoplasmic truncated forms of NHE1 (see "Experimental
Procedures"). PS120 cells were stably transfected with cDNAs
encoding the Myc-tagged full-length (NHE1-Myc) or truncated versions of
NHE1 (
698-Myc and
635-Myc). Immunoprecipitation of NHE1-Myc,
698-Myc and
635-Myc from 32P-labeled cells is shown
in Fig. 1. In NHE1-Myc quiescent cells (lane
1), the RP-c28 antibody recognized a protein running as a diffuse
32P-labeled band at a molecular mass of about 110 kDa as
reported for NHE1 (8); the anti-Myc 9E10 antibody was found to
recognize the same protein (data not shown). As described previously
for NHE1 (8), phosphorylation of NHE1-Myc was found to increase upon
stimulation with FCS (lane 2). The
698-Myc deletion
mutant was shown to run as a diffuse 32P-labeled protein at
a lower molecular mass than NHE1-Myc, consistent with the molecular
mass expected from the deletion of 117 amino acids (lane 3).
Interestingly, the
698-Myc deletion mutant did not demonstrate any
increase in phosphorylation upon serum stimulation but rather a
decrease in total phosphorylation. No 32P-labeled protein
at the expected molecular mass for the
635-Myc deletion could be
detected when
635-Myc was immunoprecipitated from resting cells
(lane 5) or cells stimulated with 10% FCS (lane 6).
These results indicate that the 635-Myc mutant is no longer
phosphorylated, therefore allowing the mapping of NHE1 phosphorylation sites to the cytoplasmic tail (amino acids 636-815).
It is noteworthy that these results confirm the mapping that had
already been obtained by indirect means, by comparison of phosphopeptide maps of four NHE1 variants: the wild type NHE1, two
internal deletion mutants (515-566 and
567-635), and the expressed NHE1 cytoplasmic domain (11).
The 698-Myc mutant was phosphorylated in unstimulated cells,
allowing the localization of basal phosphorylation site(s) for NHE1 in
the 635-698 domain. More importantly, no increase in the total
phosphorylation of the
698-Myc mutant could be detected upon
addition of serum, suggesting that the growth factor-sensitive phosphorylation site(s) are distal to the 698 residue. These results should however be taken with caution, since considering the total phosphorylation of the NHE1 deletion mutant could be misleading; an
increase in phosphorylation in one region could be masked by a decrease
in phosphorylation on a distinct site in another region. Phosphopeptide
mapping would be necessary to resolve this uncertainty, but
unfortunately the signals obtained from
32P-immunoprecipitates with the 9E10 antibody were too weak
to allow phosphopeptide mapping for the Myc-tagged NHE1 deletions.
It has been established that two classes of mitogens initiating their
signals either through receptor tyrosine kinases (i.e. epidermal growth factor) or through G protein-coupled receptors and
protein kinase C (i.e. thrombin) both stimulated NHE1
phosphorylation exclusively on serine residues. Moreover, the patterns
of NHE1 phosphorylation in response to epidermal growth factor and
thrombin were found to be identical. These findings led us to postulate that MAPKs that can integrate signals from multiple transmembrane receptors might play an important role in the cascade leading to NHE1
activation and presumably phosphorylation (10). As stated above,
truncation of the NHE1 cytoplasmic tail from residue 635 eliminated all
major sites of phosphorylation, yet the resulting truncated mutant was
still activable (although to a lesser extent) by all growth factors
tested (11). In contrast, deletion of the 566-635 domain abolished
NHE1 activation by growth factors, raising the possibility of the
existence of one or multiple regulatory protein(s) potentially
interacting with the cytoplasmic domain of NHE1. This regulatory
protein could itself be the target of a MAPK cascade. In this respect,
it is worth noting that the 635 transfectant was found to still
respond to the serine/threonine phosphatase inhibitor okadaic acid
(data not shown; Ref. 11), suggesting that phosphorylation reactions
upstream of NHE1 play a role in the exchanger activation pathway. We
therefore investigated the potential involvement of the various MAPK
cascades in this mechanism of activation. To this end, we first
compared the effect of various stimuli on activation of these kinase
cascades with their effects on NHE1 regulation in CCL39
fibroblasts.
Differential Effects of Growth Factors and Stress Signals on p44 and p38 MAPK Activities
HA-p44 MAPK was immunoprecipitated from CCL39 stable transfectants
submitted to various stimuli, and activation of HA-p44 MAPK was
measured by the capacity to phosphorylate its preferential substrate
MBP. As shown in Fig. 2 and as reported previously (39), mitogenic stimuli such as FCS and -thrombin plus insulin were demonstrated to induce a very large stimulation of p44 MAPK activity. In contrast, two stress agents, anisomycin and sorbitol (300 mM) were found to only weakly stimulate p44 MAPK.
In addition to p42/p44 MAPKs, novel members of the family of the MAPKs
(p38 MAPK and JNK) have recently been identified. These kinases are
selectively activated by proinflammatory cytokines such as IL1 but
also by environmental stresses such as UV radiation and
hyperosmolarity. HA-p38 MAPK was immunoprecipitated from PS120 stable
transfectants submitted to various stimuli, and activation of p38 MAPK
was measured by the capacity to phosphorylate its preferential
substrate ATF2. As reported for other cell lines (36, 41),
growth-promoting agents such as FCS or
-thrombin plus insulin that
maximally activated p44 MAPK only had a small effect on p38 MAPK
activation in hamster fibroblasts. The best stimulus for p38 MAPK was
found to be anisomycin (10-fold stimulation); p38 MAPK was also
stimulated by an osmotic shock of 300 mM sorbitol but to a
lesser extent than by anisomycin, whereas an osmotic shock of 100 mM sucrose was totally ineffective on both p38 and p44
MAPKs regardless of the time of stimulation.
We can conclude from these experiments that, in CCL39 fibroblasts, p44
MAPK is preferentially activated by growth factors (FCS, -thrombin
plus insulin), whereas p38 MAPK is selectively stimulated by stress
(sorbitol, anisomycin).
Effect of p44 and p38 MAPK Activators on NHE1 Regulation
We next investigated the effect of p44 and p38 MAPK activators on
NHE1 regulation. As shown in Fig. 3, anisomycin and
IL1, which were shown to be specific activators of p38 MAPK in
hamster fibroblasts (36), failed to activate NHE1 in these cells. The pHi experiments were done in PS120 fibroblasts stably transfected with wild type NHE1 (pEAP
), therefore
expressing only their endogenous level of p38 MAPK, but similar results
were obtained on HA-p38 MAPK transfectants (data not shown). In
contrast, as reported previously (12, 38), an osmotic shock of 100 mM sucrose that had no effect on p38 MAPK activation (Fig.
2) had the capacity to induce a marked stimulation of NHE1, arguing
against the involvement of p38 MAPK in the osmotic induced activation
of NHE1.
Effect of Osmotic Shocks of Varying Intensities on Activation of p38 MAPK and JNK
The stress-activated protein kinase JNK has also been reported to
be activated by osmotic stress in a number of cell lines (41, 42). We
therefore compared the effect of osmotic stress ranging from 100 to 500 mM on the activity of p38 MAPK and JNK. A 15-min
stimulation time was chosen in accordance with the time course of NHE1
activation by osmotic stress, which peaked at around 15 min (data not
shown). As shown in Fig. 4, activation of both kinases
was found to be osmolarity-dependent. No stimulation of p38
MAPK could be detected when the intensity of the osmotic stress was
lower than 300 mM sorbitol. An osmotic stress of 300 mM was found to produce a 2-3-fold activation of p38 MAPK,
whereas a 4-fold stimulation was obtained with 400 and 500 mM sorbitol. The response of JNK to hyperosmolarity was
more pronounced, but an intensity of 400 mM sorbitol was
required to produce activation of the kinase.
Inhibition of p38 MAPK (SB203580) Does Not Affect Stimulation of NHE1 by Sorbitol (or Thrombin plus Insulin)
The role of the p38 MAPK cascade in NHE1 activation by growth factors or osmotic stress was finally assessed using a chemical compound, pyridinyl imidazole (SB203580). This compound has been recently described as a specific inhibitor of the p38 MAPK, since it had no effect on a wide range of other kinases including p42 and p54 MAPKs and phosphatases (23, 43).
As shown in Fig. 5, a pretreatment of several hours in
the presence of 5 µM SB203580 that reduces p38 MAPK
activity by more than 90%2 had absolutely
no effect on the stimulation of NHE1 either by osmotic stress (300 mM sorbitol) or by growth factors (-thrombin plus
insulin).
Taken together, these findings argue strongly against a role of p38
MAPK in NHE1 activation pathway by osmotic stress or -thrombin plus
insulin.
NHE1 Is Not an in Vitro Substrate for p44 and p38 MAPKs
p42 and p44 MAPKs have been shown to phosphorylate the sequence
P-X-S/T-P in their target proteins (44). Interestingly, such
a consensus site (two adjacent serine residues, 722 and 725) can be
identified in the region of the cytoplasmic domain of NHE1 where the
potential phosphorylation sites have been localized (amino acids
700-746). To examine whether p44 MAPK can regulate NHE1 directly, we
tested the ability of the kinase to phosphorylate NHE1 in
vitro. For this purpose, we performed kinase assay experiments where, as a substrate for the kinase reaction, we used NHE1,
immunoprecipitated from pEAP-transfected fibroblasts. As
shown in Fig. 6, upper panel, stimulation of
the cells with 20% FCS induced the phosphorylation of several proteins
immunoprecipitated with the monoclonal anti-HA antibody, therefore
coimmunoprecipitating with HA-p44 MAPK. In particular, a protein
migrating just under the 112-kDa molecular mass marker could be
detected: this protein has been identified as p90rsk, which was
reported to coimmunoprecipitate with p44 MAPK and to be phosphorylated
in response to growth factors when associated with p44 MAPK (36, 24,
45). Note that p44 MAPK autophosphorylation could also be detected. The
same pattern of phosphorylation was obtained when fibroblasts deleted
for the Na+/H+ exchanger (PS120 cells) were
used as negative controls instead of NHE1 expressors (lanes
1 and 2). NHE1 protein could be detected with the
anti-NHE1 RP-c28 antibody (Fig. 6, lower panel; lanes 3 and 4) but did not appear as a
32P-labeled protein, suggesting that p44 MAPK does not
phosphorylate NHE1 in vitro. The antibody used to
immunoprecipitate NHE1 is directed against the last C-terminal 157 amino acids, a domain where the serine residues consensus sites are
located. One could therefore imagine that the binding of the antibody
to the cytoplasmic domain could mask the consensus sites and prevent
their access for the p44 MAPK. In order to circumvent this difficulty,
we performed similar experiments using fibroblasts expressing an N
terminus VSVG epitope-tagged NHE1. Insertion of the VSVG epitope
removed the N-glycosylation of NHE1, and NHE1-VSVG was shown
to migrate at the level of the 84-kDa molecular mass marker (Fig. 6,
lower panel; lanes 5 and 6), therefore
precluding potential interference of the migration of NHE1-VSVG with
the migration of p90rsk. As shown in Fig. 6, upper
panel, lanes 5 and 6, immunoprecipitation of
NHE1-VSVG with the monoclonal anti-VSVG antibody did not modify the
pattern of phosphorylation; NHE1-VSVG could be detected by Western
blotting with the anti-NHE1 antibody but did not appear as a
32P-labeled protein.
We can conclude from these experiments that p44 MAPK does not appear to phosphorylate NHE1 directly.
It is noteworthy that similar experiments using HA-p38 MAPK instead of HA-p44 MAPK allowed us to show that, as found in the case of p44 MAPK, p38 MAPK failed to phosphorylate NHE1 in vitro (data not shown).
Inhibitors of the p42/p44 MAPK Cascade Inhibit Growth Factor Activation of NHE1
p44 MAPK Dominant Negative Mutant Inhibits Growth Factor-mediated Activation of NHE1To investigate the role of the
p42/p44 MAPK signaling pathway in the regulation of NHE1 by growth
factors, we first examined the effects of blocking the MAPK cascade at
the level of MAPK itself. As reported previously, inhibition of
endogenous p42/p44 MAPK activity can be achieved by several approaches
including transient overexpression of the inactive p44 MAPK-T192A form
(27). In Fig. 7A, we have assessed the effect
of the expression of the vector p44 MAPK-T192A compared with wild type
p44 MAPK or empty vector on the pHi response of NHE1 to growth
factors (-thrombin plus insulin or serum). Expression of the
dominant negative mutant was found to reduce by 50% the response of
NHE1 to growth factors (
pHi = 0.14 ± 0.04, n = 7 for the cells expressing the p44 MAPK-T192A
compared with
pHi = 0.27 ± 0.02, n = 7 for the cells transfected with the control vector), suggesting the
involvement of p42/p44 MAPK in the NHE1 activation pathway initiated by
-thrombin plus insulin or serum.
Expression of MKP-1, a MAPK Phosphatase, Inhibits Growth Factor-mediated Activation of NHE1
As an alternative approach to
reduce endogenous MAPK, we used the same transient transfection assay
to overexpress MKP-1, a dual specificity phosphatase shown, in
vivo and in vitro, to dephosphorylate and inactivate
MAPKs (46). As shown in Fig. 7B, in the cells transfected
with the control vector, the combination -thrombin plus insulin
induced a cytoplasmic alkalinization of 0.16 ± 0.02, n = 4, and expression of MKP-1 was found to
significantly inhibit the pHi response of NHE1 to
-thrombin + insulin (
pHi = 0.09 ± 0.02, n = 4).
This inhibition is of the same order of magnitude as the inhibition
obtained with the dominant negative form of p44 MAPK.
To confirm the involvement of the p42/p44 MAPK cascade in the activation pathway of NHE1, we took advantage of the recent discovery of a specific MKK1 inhibitor, PD 98059, and investigated the effect of this compound on the activation of NHE1 by various stimuli.
The MKK1 inhibitor has been shown to specifically inhibit the
activation of p42/p44 MAPK without affecting the activity of other
MAPKs such as p38 and JNK (47, 48). In CCL39 cells, PD 98059 at 10 µM was demonstrated to inhibit by 70% the activation of
p44 MAPK by growth factors (data not
shown).3 As shown in Fig. 7C, PD
98059 (10 µM) was found to significantly inhibit the
response of NHE1 to -thrombin + insulin (65% inhibition); this
inhibitory effect could also be detected and was even more pronounced
when the cells were stimulated with thrombin alone. In contrast,
stimulation of NHE1 with sorbitol (300 mM) remained unaffected by the same pretreatment with PD 98059.
These results demonstrate that blocking the MKK1 inhibits the
activation of NHE1 by -thrombin plus insulin but not by sorbitol, reinforcing the notion that the p42/p44 MAPK cascade plays an important
role in the activation of NHE1 by growth factors and suggesting that,
in contrast, this cascade does not seem to be involved in the osmotic
activation of NHE1.
Specific Activation of the p42/p44 MAPK Cascade Stimulates NHE1
MKK1 and MKK2 constitute the preferential targets of Raf-1 (18).
We therefore finally assessed the contribution of the p42/p44 MAPK
cascade in the mitogen-induced activation of NHE1 by using a
CCL39-derived cell line, CCL39-Raf-1:ER, that stably expresses an
estradiol-regulated form of human oncogenic Raf-1 (32, 33). The
addition of estradiol to these cells was demonstrated to be sufficient
to elicit rapid and persistent activation of MKK1 and MAPK (33).
Fig. 8A shows
the effect of estradiol on the pHi of quiescent
CCL39-Raf-1:ER cells. Treatment of these cells with 1 µM estradiol resulted in a marked cytoplasmic
alkalinization. This change in pHi is due to activation of
NHE1, since it was completely abolished in presence of the
NHE1-specific inhibitor HOE 694 (49, 50) (data not shown). The addition
of the same concentration of estradiol to the parental cell line CCL39
had no effect on pHi, demonstrating that the estradiol-induced activation of NHE1 resulted from Raf activation. The effect of estradiol on pHi is comparable with that observed with growth
factors like
-thrombin in CCL39-
Raf-1:ER or CCL39 cells.
The Time Course of Activation of NHE1 by Estradiol Parallels That of p42/p44 MAPK
Fig. 8B shows a time course of
activation of NHE1 by estradiol and thrombin. After a lag of 2 min,
thrombin induced a rise in pHi that peaked at around 5 min and
declined after 15 min but nevertheless persisted for 1 h. The
effect of estradiol on pHi was barely detectable 5 min after
stimulation but reached a maximal value within 15 min to remain at a
maximum level for 1 h. The time course of activation of NHE1 by
estradiol is perfectly compatible with the time course of its
stimulation of p42/p44 MAPK. As described previously (33), activation
of Raf-1:ER induced the mobility shift-up of both p42 and p44 MAPKs in a time-dependent manner. This retardation of MAPK
mobility in SDS-PAGE results from phosphorylation of the two activation sites of MAPK and strictly correlates with MAPK activation (51). The
activation of MAPK was detected 5 min after the addition of estradiol
and reached its maximum 15-30 min later.
It is important to note that in the case of -thrombin, the
time course of activation of p44 MAPK in hamster fibroblasts, readily
detectable after 2 min and reaching a maximum within 5 min (data not
shown; Ref. 52), is similarly compatible with an involvement of p44
MAPK in the pathway of NHE1 activation (Fig. 8B).
We
next investigated the effect of MKK1 inhibition on estradiol-induced
activation of NHE1 (Fig. 8C). The inhibitory effect of PD
98059 on the -thrombin-induced stimulation of NHE1 in
CCL39
Raf-1:ER cells was comparable with that observed in the
parental cell line CCL39. The effect of PD 98059 on the stimulation of
NHE1 induced by estradiol was more pronounced, suggesting that the
effects of estradiol are mediated by the only substrates of Raf
identified so far, MKK1 and MKK2. We also investigated whether the
estradiol- and
-thrombin-induced alkalinizations are cumulative.
Combination of estradiol and
-thrombin produced a cytosolic
alkalinization similar to that obtained with each agent alone,
suggesting that the NHE1 activation pathway initiated by
-thrombin
is mainly mediated by the Raf
MAPK cascade.
In conclusion, these results demonstrate that exclusive activation of
the Raf MKK1
p42/p44 MAPK cascade is able to induce NHE1 activation.
Together, these results strongly suggest an important contribution of the p42/p44 MAPK signal transduction pathway in the regulation of NHE1 in response to growth signals.
Inhibition of MKK1 Also Reduces the Growth Factor Response of
Truncated Forms of NHE1 (698 and
635)
We next investigated the effect of the MKK1 inhibitor on the
activation of the NHE1 deletions 698 and
635 by
-thrombin. As
shown in Fig. 9, the inhibitory effect of PD 98059 on
activation by
-thrombin plus insulin is preserved in the
698 and
635 cells (although the effect was found to be less pronounced for
the
635 mutant), thus suggesting that the alkalinization observed in
the deletion mutants upon stimulation with
-thrombin plus insulin is
mediated, in part, by a MKK1-dependent pathway. Since the
698 deletion removes the consensus sites for phosphorylation by p44 MAPK and the
635 deletion removes all major NHE1 phosphorylation sites, these results support the notion that the effect of p44 MAPK on
NHE1 is not mediated by direct phosphorylation of NHE1.
The ubiquitously expressed form of Na+/H+ exchanger, NHE1, is involved in a variety of physiological functions by virtue of its ability to govern intracellular pH. One of the most remarkable features of NHE1 is its capacity to respond to multiple extracellular stimuli such as hormones, growth factors, vasoactive peptides, and integrins as well as to hyperosmotic stress. Because many of these extracellular stimuli share the ability to activate a common protein kinase signaling cascade, we previously proposed (8, 53) that this pathway known as the p42/p44 MAPK cascade could be involved in the activation of NHE1. However, although attractive, this notion remained purely hypothetical until the very recent report by Barber and colleagues (30) and the present study. By using selective molecular tools that positively or negatively interfere at each level of the signaling cascade (Raf-1, MKK1 (MEK1), and p44 MAPK (extracellular regulated kinase 1)), we have provided strong evidence for the contribution of the p42/p44 MAPK cascade in the activation of NHE1.
Perhaps the most persuasive results pointing to the role of this
cascade in NHE1 activation were derived from the cell line expressing
an estradiol-inducible form of oncogenic Raf-1. There, by the ability
to bypass multiple upstream signals known to be generated by many
growth factors (phosphatidylinositol 3-kinase, phospholipase C ,
Ca2+ rise, Rho/Rac-activated pathways), we showed that the
rapid and exclusive activation of the p42/p44 MAPK cascade, at the
level of Raf-1, was able to activate NHE1 to the same extent and with a
comparable time course as that induced by
-thrombin in these cells.
Raf-1 activation was found to be sufficient to produce maximal
activation of NHE1, and this effect was not cumulative with the effect
of
-thrombin. In addition, NHE1 activation mediated either by
estradiol or by
-thrombin was inhibited by the specific MKK1
inhibitor PD 98059. These results, therefore, support the notion that
-thrombin action is mediated via the MAPK cascade. The
-thrombin-mediated NHE1 activation, however, was incompletely inhibited by the MKK1 inhibitor. Two explanations can account for this
result. First, the concentration of PD 98059 that we used in our
experiments (10 µM) was shown to inhibit by 70% the activation of p44 MAPK by
-thrombin; the residual p44 MAPK activity could therefore be sufficient to produce NHE1 activation. Note that
even the activation of NHE1 mediated by direct stimulation of Raf,
which is acting directly upstream of MKK, was not fully abolished by PD
98059 (Fig. 8C). Second, thrombin-mediated activation of
NHE1 might involve, in addition to the p42/p44 MAPK cascade, another
pathway, in particular the Ca2+/calmodulin activating step
triggered by the rapid and sustained elevation of Ca2+
(12). On the other hand, thrombin has been reported to stimulate p38
MAPK in human platelets (54). In CCL39 fibroblasts, our results show
that the effect of the combination
-thrombin plus insulin on p38
MAPK is much less pronounced than its effect on p44 MAPK (Fig. 2); in
addition, inhibition of p38 MAPK was found to have no effect on the
activation of NHE1 by
-thrombin plus insulin (Fig. 5). Taken
together, these results argue against an involvement of p38 MAPK in the
thrombin-mediated activation of NHE1.
Finally, our findings demonstrating that reducing p44 MAPK activity by expression of a dominant negative form of p44 MAPK or by expression of the dual specificity phosphatase MKP-1 inhibited NHE1 response to growth factors provided additional evidence for a p42/p44 MAPK-dependent regulation of NHE1. It is noteworthy that these results are in agreement with a recently published report by Barber and colleagues (30). Using a different approach, transient expression of constitutively active kinases and GTPases and dominant negative kinases in CCL39 fibroblasts, they reported that constitutively active Ras V12 stimulates NHE1 through a Raf-1- and MKK-dependent mechanism.
In the context of these two independent studies stressing the
contribution of the p42/p44 MAPK cascade in the activation of NHE1, it
is important to note that NHE1 does not appear to be a direct
target for p42/p44 MAPK. We have demonstrated that p44 MAPK does not
phosphorylate NHE1 in vitro (Fig. 6); in addition, deletion
of the consensus sites for phosphorylation by p44 MAPK (698 mutant)
did not prevent the inhibitory effect of PD 98059 on growth
factor-mediated activation of NHE1 (Fig. 9). These results allow the
mapping of the site sensitive to activation by the p42/p44 MAPK pathway
proximal to the 698 residue, presumably even proximal to the 635 residue, since an effect of the MKK1 inhibitor could still be detected
for the
635 mutant.
Our results therefore reinforce the hypothesis formulated by Wakabayashi et al. (11), postulating the existence of one or multiple proteins that would regulate NHE1 activity presumably through interaction with its cytoplasmic domain (566-635 region). Identification of these regulatory proteins should undoubtedly result in valuable information concerning the molecular mechanism of activation of NHE1. Exploitation of the two-hybrid system led P. Fafournoux in our group to isolate several cDNA clones encoding distinct proteins that specifically interact with NHE1 cytoplasmic domain. These putative regulatory proteins as targets of the p42/p44 MAPK cascade are currently under investigation.
A second aim of the present study was to gain better understanding of
the molecular mechanism underlying regulation of NHE1 by osmotic
stress. Like growth factors, osmotic stress has been shown to activate
NHE1 by increasing its sensitivity for H+ (7). Despite
these kinetic similarities suggesting a common mode of stimulation,
activation of NHE1 during volume regulation was not found to be
associated with an increase in NHE1 phosphorylation (55). The Raf
MKK
MAPK cascade was shown to be activated by
hyperosmolarity in the renal cell line, Madin-Darby canine kidney cells
(56), and in 3Y1 fibroblasts (42). Our results show that, in CCL39
fibroblasts, although the osmotic stress was capable of activating p44
MAPK (Fig. 2), inhibition of MKK1 did not prevent NHE1 stimulation by
hyperosmolarity, arguing against the involvement of the p42/p44 MAPK
cascade in the osmotically induced activation of NHE1. In this context,
it is important to note that the effects observed by Terada and
colleagues (56) were strictly dependent upon the intensity of the
osmotic stress applied to the cells. In the experiments reported there
(56), a minimum of 500 mM raffinose was required to get a
stimulatory effect on the MAPK cascade. Here, we show that, in hamster
fibroblasts, whereas we could detect a strong stimulation of NHE1 with
only 100 mM sucrose (Fig. 3), this osmotic stress was not
found to stimulate p44 MAPK (Fig. 2). In view of the results summarized above, it appears that the activation pathways initiated by growth factors and osmotic stress to stimulate NHE1 are distinct. In this
regard, it is noteworthy that in a recent study, McSwine and colleagues
(57) reported that serum, but not hypertonicity, activates NHE1 by a
Ca2+-dependent process: these results support
the notion of two distinct pathways for both stimuli.
New members of the MAPK family, namely p38 MAPK and JNK, have recently
been cloned. Among the stimuli that selectively activate these kinases,
environmental stresses such as osmotic challenge have been identified.
For instance, hyperosmolarity was reported to induce a strong
stimulation of both p38 MAPK and JNK in HeLa cells (41),
phosphorylation of p38 MAPK in a murine pre-B cell line (21), and
activation of JNK in Chinese hamster ovary cells (58). In addition,
Barber and colleagues (30) have recently reported that a GTPase
described to stimulate the Jun kinase cascade, G13, activates NHE1
(through a Cdc42- and MEK kinase-dependent mechanism). It
was therefore of particular interest to analyze the potential
involvement of p38 MAPK and JNK in the osmotically induced activation
of NHE1. Several lines of evidence argue against this hypothesis.
First, as already observed in the case of p44 MAPK, an osmotic shock of
100 mM sucrose susceptible to activate NHE1 was unable to
stimulate neither p38 MAPK nor JNK (Fig. 4). Second, activation of NHE1
by hyperosmolarity remained unaffected by inhibition of p38 MAPK (Fig.
5). Taken together, our results do not argue in favor of the
contribution of any of the MAPK p44, p38, or JNK cascades in the
regulation of NHE1 by osmotic stress.
In conclusion, we have demonstrated that in addition to the well
characterized Ca2+/calmodulin activating pathway (12), NHE1
is activable by the Raf MKK
p42/p44 MAPK cascade and that
this pathway contributes primarily to NHE1 activation by potent growth
factors such as
-thrombin. In addition, we demonstrated that this
action does not directly involve phosphorylation of NHE1, indicating
the role of NHE1 accessory proteins in this activation process.
Finally, we found no evidence in favor of the involvement of p38 MAPK
and JNK cascades in osmotically induced activation of NHE1.
P. Lenormand is gratefully acknowledged for
providing the CCL39-Raf-1:ER cell line. We thank J. Noël and
R. C. Poole, respectively, for the Myc- and VSVG-tagged NHE1
constructions. We thank J. Lavoie, A. Brunet, and F. R. McKenzie for
providing unpublished results. J.-M. Brondello, P. Fafournoux, and J. Noël are acknowledged for fruitful discussions.