(Received for publication, October 30, 1996, and in revised form, February 21, 1997)
From the Institute of Biochemistry and Molecular
Biology, Albert-Ludwigs-University, D-79104 Freiburg, Germany and
the ¶ Department of Internal Medicine, Eberhard-Karls-University,
D-72076 Tübingen, Germany
Mitogen-activated protein (MAP) kinases are
important mediators of the cellular stress response. Here, we
investigated the relationship between activation of the MAP kinase p38
and transcription factor NF-B. Different forms of cellular stress
were found to preferentially trigger either p38 or NF-
B. Arsenite or
osmotic stress potently activated p38 but were ineffective in inducing NF-
B activation. Tumor necrosis factor-
and hydrogen peroxide, in
contrast, led to NF-
B activation but only modestly stimulated p38.
The activation of NF-
B was strongly abolished by antioxidants, while
the activity of p38 and transcription factor AP-1 were increased. Inhibition of small GTPases including Rac and Cdc42 prevented p38 and
AP-1 activation without interfering with NF-
B. In addition, inhibition of p38 by a pharmacological inhibitor or a dominant-negative mutant of MAP kinase kinase-6, an activator of the p38 pathway, interfered with NF-
B-dependent gene expression but not
its DNA binding activity. Our results indicate that activation of p38 and NF-
B are mediated by separate pathways, which may converge further downstream in the cell nucleus. Different forms of cellular stress, however, initially trigger distinct signaling cascades involving either oxidative stress or GTPase-coupled pathways.
Gene induction by cellular stresses is mediated through biochemical processes that mostly involve the interplay of multiple signaling pathways. Depending on the nature of the stimulus, intricate protein kinases are activated that ultimately phosphorylate transcription factors and result in gene expression. Central mediators that propagate signals from the cell membrane to the nucleus are protein kinases related to the mitogen-activated protein (MAP)1 kinase superfamily. To date, at least three different subtypes of MAP kinases are known (reviewed in Refs. 1-3). These are in turn activated by distinct upstream dual specificity kinases thus revealing the existence of protein kinase modules that can be independently and simultaneously activated. Whereas mitogens and growth factors lead to activation of protein kinase cascades resulting in activation of ERK family MAP kinases, many forms of cellular stress preferentially trigger two related signaling pathways (4-8). These center on two MAP kinase homologues called stress-activated protein kinases or Jun NH2-terminal kinases (JNKs), and p38, also termed reactivating kinase. Targets of stress-activated protein kinase and JNK include several transcription factors such as c-Jun, JunD, ATF-2, and Elk-1, which become activated after exposure to cellular stresses (9-12).
Mammalian p38 was originally identified in murine pre-B cells transfected with the lipopolysaccharide complex receptor CD14 and in macrophages where it is activated in response to lipopolysaccharide (13). The amino acid sequence of p38 is most similar to HOG-1, a MAP kinase homologue that lies in a signaling pathway that restores the osmotic gradient across the plasma membrane of Saccharomyces cerevisiae in response to increased external osmolarity (14). Like ERKs and JNKs, p38 requires phosphorylation of a closely spaced tyrosine and threonine for activation. However, the enzyme is distinguished by the sequence TGY in the activation domain, which differs from the TEY sequence found in ERKs, and the TPY sequence in the JNK and MAP kinase homologues.
In addition to hyperosmotic shock, p38 is activated by a wide spectrum
of stimuli such as physicochemical stresses, lipopolysaccharide, and cytokines (13, 15-17). Despite rapid progress in the elucidation of structural elements of the p38 pathway, the physiological
consequences of its stimulation by stress agents largely remain to be
defined. One of the downstream targets of p38 is the MAPKAP kinase-2,
which can activate the transcription factors CREB and ATF-1 (18) and phosphorylate heat-shock protein hsp27 (19, 20). It is assumed that
hsp27 phosphorylation promotes the polymerization of actin and so
counteracts the disruptive effects of stress on actin microfilaments (21). A number of evidences further suggest that p38 plays a key role
in regulating apoptotic and inflammatory responses. Highly specific
inhibitors of p38 such as pyridinyl imidazoles potently block the
synthesis of TNF and interleukin-1 (22). Prevention of interleukin-1
and TNF synthesis was shown to be due to inhibition of translation
initiation, which is regulated through the AUUUA motif in the
3-untranslated regions of their transcripts (23). In addition,
inhibition of p38 has been found to prevent interleukin-6 and
granulocyte-macrophage colony-stimulating factor mRNA synthesis indicating that p38 may regulate transcriptional events (24). In this
respect, p38 has been implicated in the activation of multiple
transcription factors, such as ATF-2 (17, 25), Elk-1 (25), c-Fos
and c-Jun (26), CHOP (27), Max (28), and NF-
B (24).
The transcriptional activator NF-B is a key component controlling
the synthesis of cytokines and many other immunoregulatory gene
products (reviewed in Refs. 29 and 30). The factor is triggered by a
great variety of proinflammatory or pathogenic stimuli including
inflammatory cytokines, phorbol ester, T cell mitogens, and
lipopolysaccharide, as well as ionizing and UV irradiation. In
unstimulated cells, NF-
B resides in the cytoplasm as an inactive heterodimeric complex bound to a third inhibitory subunit called I
B.
Upon stimulation of cells, NF-
B is rapidly activated by the
dissociation of I
B. This exposes nuclear localization sequences in
the remaining NF-
B heterodimer, leading to nuclear translocation and
subsequent binding of NF-
B to DNA-regulatory elements within target
genes. It has been found that the dissociation of I
B is preceded by
its phosphorylation, an event which subsequently leads to the
proteolytic degradation of I
B (31). A number of evidences further
suggest that reactive oxygen intermediates (ROIs) are the essential
second messengers involved in the course of NF-
B activation (30).
This is supported by the findings that (i) NF-
B is readily activated
in many cell types exposed to hydrogen peroxide (32, 33), and (ii)
activation in response to all stimuli tested so far is blocked by a
variety of antioxidants (32, 34). Therefore, it is currently assumed
that the intracellular production of ROIs, which is stimulated by
stress inducers, activates a redox-sensitive kinase that in turn
phosphorylates I
B.
As many of the stimuli leading to NF-B and p38 activation are
similar, we wished to establish the relationship between p38 and
NF-
B-mediated transcription. A potential role of p38 in NF-
B activation has recently been implicated by the finding that
p38-specific inhibitors potently attenuate NF-
B activity (24). We
report here that although some stress inducers such as TNF activate
both p38 and NF-
B, other stimuli including sodium arsenite and
hyperosmotic shock selectively trigger p38. In addition, the
antioxidants that prevented NF-
B activation did not inhibit but even
increased p38 activity. However, blockade of p38 activation by a
specific p38 inhibitor or a dominant-negative MAP kinase kinase-6
(MKK6) mutant inhibited NF-
B-dependent gene expression
without interfering with NF-
B DNA binding activity. These data
indicate that activation of NF-
B and p38 lie on primarily separate
stress-responsive pathways but may converge downstream at the level of
NF-
B-mediated transactivation.
The human embryonic kidney cell line 293 and HeLa cells were maintained in Dulbecco's modified Eagle's medium
(Life Technologies, Inc.) supplemented with 10% heat-inactivated fetal
calf serum, 100 units of penicillin/ml, and 0.1 mg of streptomycin/ml.
All reagents were obtained from Sigma unless otherwise specified. Recombinant human TNF- was a gift (Knoll AG, Ludwigshafen, Federal Republic of Germany). Toxin A from Clostridium difficile was
kindly provided by Drs. F. Hofmann and K. Aktories (Freiburg, FRG) and the p38 specific inhibitor SB203580 by Dr. J. Lee (King of Prussia, PA).
p38 kinase activity was measured
following expression of a glutathione S-transferase (GST)
p38 fusion plasmid that was driven by the human EF-1 promoter and
kindly provided by Dr. L. E. Zon. GST-ATF-2 (amino acids 1-109), which
served as a p38 substrate, was expressed in Escherichia coli
and purified by glutathione affinity chromatography (10). To analyze
the substrate specificity of p38 the following recombinant fusion
proteins were expressed in E. coli and purified by standard
techniques: GST-I
B-
(35), His-p50,2
and His-Gal4-p65 (36). To measure the transactivating activity of
NF-
B, cells were transfected with a luciferase reporter gene construct controlled by six
B-binding sites (37). The plasmid pCMV-p65 (38) served as a positive control. A CMV-driven eukaryotic ATF-2 expression plasmid has been described (9). A dominant-negative expression plasmid of MKK6 (K82A) (25) was originally obtained from Dr.
R. J. Davis. Transfections of cells were performed by the calcium
phosphate coprecipitation method (37). Briefly, 1.5 × 106 cells for kinase assays and 2 × 105
cells for luciferase assays were seeded on 10-cm and 6-cm plates, respectively, and transfected after overnight incubation. Eighteen h
later, cells were either left untreated or stimulated with the indicated reagents. After 1 or 6 h cells were harvested and
assayed for p38 kinase or luciferase activity, respectively.
Cells were harvested 24 h post-transfection, lysed in 25 mM glycylglycine, pH 7.8, 1% Triton X-100, 15 mM MgSO4, 4 mM EGTA, 1 mM DTT, and centrifuged at 13,000 × g at 4 °C for 5 min. Fifty microliters of the supernatant were assayed in 100 µl of assay buffer (25 mM glycylglycine, 15 mM MgSO4, 4 mM EGTA, 15 mM potassium Pi, pH 7.8, 1 mM DTT, 1 mM ATP) using a LB 96 P luminometer (Berthold, Bad Wildbach, FRG). Following injection of 100 µl of luciferin (0.3 mg/ml) light emission was measured in 30-s intervals. Results are given in relative light units.
Solid-phase Kinase Assayp38 activity was determined by a
solid-phase kinase assay using GST-ATF-2 as substrate. Following
transfection with the GST-p38 construct, 293 cells (1.5 × 106) were extracted in 100 µl of lysis buffer containing
20 mM HEPES, pH 7.5, 300 mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 1 mM DTT, 10 mM NaF, 1 mM
Na3VO4, 2 mM phenylmethylsulfonyl
fluoride, and 1% Triton X-100. After centrifugation to remove cellular
debris, 300 µl of equilibration buffer (20 mM HEPES, pH
7.5, 2.5 mM MgCl2, 1 mM EDTA, and
0.05% Triton X-100) were added. p38 was adsorbed by the addition of 30 µl of a 50% slurry of glutathione-coupled Sepharose beads (Pharmacia
Biotech Inc.). After incubation at 4 °C for 30 min, the beads were
washed twice with washing buffer (20 mM HEPES, pH 7.5, 50 mM NaCl, 2.5 mM MgCl2, 1 mM EDTA, 0.1 mM Na3VO4,
0.05% Triton X-100), twice in LiCl buffer (20 mM Tris-HCl, pH 7.6, 500 mM LiCl, 0.1 mM
Na3VO4), and twice in washing buffer. The beads
were then equilibrated by two washing steps in kinase buffer (20 mM HEPES, pH 7.5, 2.5 mM MgCl2, 40 µM ATP, 2 mM DTT, 0.1 mM
Na3VO4). The kinase reaction was performed by
adding 25 µl of kinase buffer containing 5 µCi of
[-32P]ATP and 5 µg of GST-ATF-2 as substrate. After
incubation for 30 min at 30 °C, the reaction was terminated by the
addition of 4 × Laemmli sample buffer. Phosphorylated proteins
were resolved on a 15% SDS-polyacrylamide gel electrophoresis gel,
visualized by autoradiography, and quantitated by PhosphorImager
analysis.
For detection
of NF-B and AP-1 DNA binding activities, cells were essentially
processed as described (39). Binding reactions were performed in a
20-µl volume containing 3-4 µg of extract, 4 µl of 5 × binding buffer (20 mM HEPES, pH 7.5, 50 mM KCl,
1 mM DTT, 2.5 mM MgCl2, 20%
Ficoll), 2 µg of poly(dI-dC) as nonspecific competitor DNA, 2 µg of
bovine serum albumin, and 10,000-15,000 cpm Cerenkov radiation of the
labeled oligonucleotide. After a 30-min binding reaction at room
temperature, samples were loaded on a 4% non-denaturing polyacrylamide
gel and run in 0.5 × TBE buffer (89 mM Tris, 89 mM boric acid, 1 mM EDTA), pH 8.3. The sequences of the oligonucleotides used to detect the DNA binding activity of NF-
B and AP-1 were as follows (binding sites are underlined): NF-
B, 5
-AGCTTCAGAGGGGATTTCCGAGAGG-3
;
AP-1, 5
-CGCTTGATGAGTCAGCCGGAA-3
. The oligonucleotides
were annealed with their complementary strands and labeled using T4
polynucleotide kinase (Boehringer Mannheim) and
[
-32P]ATP. The labeled probes were purified from free
nucleotides on push columns (Stratagene, Heidelberg, FRG).
p38 MAP kinase is known to be
activated by a variety of cellular stresses. To measure kinase activity
we expressed GST-tagged p38 in 293 cells and determined p38 activity in
a solid-phase assay with ATF-2 as substrate. The classical stress
inducers such as the sulfhydryl agent sodium arsenite and hyperosmotic
concentrations of sorbitol stimulated p38 activity to as much as 11- and more than 20-fold of the control, respectively (Fig.
1, A and C). p38 activity was also
strongly induced by the phorbol ester PMA and to a lesser extent by
TNF. The combined incubation with both TNF and PMA resulted in additive
stimulatory effects. Treatment of cells with hydrogen peroxide caused
only a moderate increase in p38 activity, and in some experiments
stimulation by hydrogen peroxide was only slightly above basal
activity.
In parallel experiments we measured the activation of transcription
factor NF-B, which is activated by various stimuli many of those
resembling p38 activators. In contrast to p38, NF-
B was virtually
not activated by sorbitol and sodium arsenite (Fig. 1, B and
C). Maximum activation of NF-
B was observed after
treatment with TNF. Phorbol ester did not activate NF-
B in 293 cells, although in other cell types activation by PMA was obtained. As
reported earlier (32, 33), NF-
B was also activated in response to hydrogen peroxide. We further analyzed the effects of okadaic acid, a
phosphatase inhibitor, and thapsigargin, an inhibitor of
endoplasmic Ca2+ ATPase, which have been shown to activate
NF-
B (40). Okadaic acid strongly induced p38 and NF-
B activation
(Fig. 1D). Thapsigargin, although weaker, activated NF-
B
but had virtually no effects on p38 activity. Collectively, these data
demonstrate that activation of p38 and NF-
B are only partially
overlapping. Several stress inducers activate selectively either p38
MAP kinase or NF-
B, thus indicating the existence of alternate
stress-inducing pathways.
Blockade of p38
activation has previously been shown to inhibit NF-B-transactivating
activity (24). We therefore investigated the ability of p38 MAP kinase
to phosphorylate NF-
B subunit proteins in vitro. In the
experiments, the potential substrates were expressed in E. coli as GST or Gal4 fusion proteins and incubated with purified p38 expressed in 293 cells. Fig. 2 shows that none of
the NF-
B proteins including p50, the transactivating C terminus of
p65, and I
B-
were phosphorylated by p38. In contrast, ATF-2 was
strongly phosphorylated by p38 MAP kinase. The results indicate that
NF-
B is unlikely to serve as a direct target of p38.
Activation of AP-1 Largely Parallels p38 MAP Kinase Activation
Besides NF-B another important stress-responsive
transcription factor is AP-1. The factor comprises a homo- or
heterodimeric complex composed of proteins of the Fos, Jun, and ATF
families. ERKs, JNKs, and presumably p38 are involved in activation of
c-Jun and ATF-2 (9, 41). We analyzed activation of AP-1 DNA binding using an oligonucleotide with the classical AP-1 site from the collagenase promoter. This motif preferentially binds to Jun/Jun and
Jun/Fos but not to ATF-2/Jun dimers (41). In identical cellular extracts used for the analysis of NF-
B, it was found that AP-1 DNA
binding activity was significantly increased in response to all stress
stimuli (Fig. 3). PMA and TNF, either alone or in
combination, potently induced AP-1 activation. A strong activation was
also observed in response to treatment with sodium arsenite, sorbitol, and hydrogen peroxide. We did not investigate further the subunit composition or requirement of de novo protein synthesis
in response to each individual stress stimulus. Based on previous
reports it can be presumed that activation of AP-1 by PMA, TNF, and
other cellular stressors involves different types of MAP kinases
targeting distinct AP-1 subunits. Thus, physicochemical stresses
preferentially lead to ATF-2 activation, whereas phorbol ester and TNF
mainly activate c-Jun (41, 42). The experiments demonstrate that, in
contrast to NF-
B, a larger variety of extracellular stimuli converge at the level of AP-1.
p38 MAP Kinase and NF-
The activation of both NF-B and AP-1 is controlled by
redox-sensitive signaling pathways (30). NF-
B is a
prooxidant-inducible factor whose activation is inhibited by a variety
of antioxidants. In contrast, antioxidants have been shown to increase
the activation of AP-1 by inducing c-fos and
c-jun transcription (43, 44). To investigate the involvement
of ROIs in p38 activation we used the antioxidant pyrrolidine
dithiocarbamate (PDTC), an iron chelator and glutathione precursor.
Fig. 4A demonstrates that preincubation of
293 cells with PDTC considerably increased p38 kinase activity in
response to all stimuli analyzed. Consistent with this and previous
reports, AP-1 activation was also enhanced by PDTC. In contrast,
NF-
B activation in response to TNF, when measured in identical
extracts, was completely abolished (Fig. 4A). The data demonstrate that p38 and NF-
B are oppositely regulated by
redox-dependent processes. Similar to AP-1, p38 activity is
enhanced by antioxidants.
To investigate whether antioxidants have a direct effect on p38 kinase activity, we incubated p38 extracted from PMA-treated cells with PDTC. Virtually no change in kinase activity was observed following treatment of p38 with different concentrations of PDTC (Fig. 4B) or glutathione (data not shown) in vitro. This shows that antioxidants could potentiate p38 activity only in intact cells.
We further sought to dissect the signaling pathways by studying the
involvement of GTPase proteins. The small GTPases Rac and Cdc42
have recently been identified as important intermediates to the JNK and
p38 activation cascades (4, 5). The activity of Rac and Cdc42 but not
of Ras is specifically inhibited by enterotoxin toxin A from C. difficile (45). Pretreatment of 293 cells with toxin A caused a
strong inhibition of p38 activation indicating the involvement of Rac
and Cdc42 in p38 kinase activation (Fig. 4A). In line with
the previous results, AP-1 activation was also inhibited by toxin A
(Fig. 4A) in response to all stimuli. Virtually no effect
was observed for NF-B suggesting that signals leading to NF-
B
activation are not transduced through Rho-related GTPases.
In the previous experiments it could not
be excluded that additional components lying in the p38 kinase pathway
cooperate with NF-B-mediated transactivation. We therefore
overexpressed p38 and ATF-2 in 293 and HeLa cells and investigated the
effects on expression of a NF-
B-controlled reporter gene.
Transfections with different concentrations of the p38 encoding plasmid
did not markedly change NF-
B activity in unstimulated or stimulated cells (Fig. 5). Likewise, overexpression of increasing
amounts of ATF-2, either alone or in combination with p38, was
ineffective in enhancing expression of the NF-
B-controlled
luciferase construct (Fig. 6). In contrast, transfection
of a NF-
B p65-encoding construct, which served as a positive
control, strongly increased NF-
B-mediated transactivation (Fig.
6A).
Effects of p38 Inhibition and MKK6 on NF-
The activity of p38 is specifically inhibited by the
pyridinyl imidazole SB203580. In 293 cells we found that SB203580 also strongly prevented expression of a NF-B-controlled reporter gene in
response to a combination of PMA and TNF (Fig.
7A). The drug concentrations required for
inhibition of NF-
B activity roughly corresponded to the effects on
p38 kinase activity (Fig. 7B). As measured in EMSA
electrophoretic mobility shift assay, SB203580, however, was unable to
prevent activation of NF-
B DNA binding activity (Fig.
7B).
We further studied the effects of the dual specificity kinase MKK6,
which has been identified as a selective activator of p38 (25, 46).
Expression of a kinase inactive MKK6 mutant abolished p38 activation
triggered by PMA and TNF (Fig. 8A). In parallel experiments, a potent inhibition of NF-B transactivating activity was also observed (Fig. 8B). Collectively, these
results therefore indicate a complex mode of regulation of p38 and
NF-
B. As suggested by the differential activation pattern of both
molecules, triggering of NF-
B by the release of I
B in the
cytoplasm involves effectors other than p38 activation. Convergence may
occur further downstream in the cell nucleus, either by effects of p38
on NF-
B-mediated transactivation or post-transcriptional
mechanisms.
NF-B represents a prototypic stress-responsive transcription
factor that is activated by a plethora of noxious conditions that also
cause activation of MAP kinases. In the course of activation, its
inhibitor I
B is phosphorylated and proteolytically degraded thereby
allowing the DNA binding NF-
B dimer to enter the cell nucleus.
Although the critical protein kinase mediating this key event is
unknown, various kinases including Src, Raf, protein kinase C isoforms
and
, and a proteasome-associated kinase have been proposed to
either directly or indirectly participate in this process (reviewed in
Refs. 29 and 47). Molecules of the MAP kinase modules have further been
implicated in NF-
B activation. Activation of JNK by upstream MEK
kinase (MEKK-1) resulted in NF-
B activation, and kinase inactive
MEKK-1 blocked TNF-induced NF-
B activity in some cellular models
(48). However, since this manuscript was submitted, it has been shown
that at least in some cell types JNK and NF-
B activation are two
separate TNF responses (49, 50). In addition, p38 may be linked to
NF-
B activation since a specific pharmacological inhibitor of p38
blocked NF-
B-dependent transactivation (Ref. 24 and
present data).
In this study we demonstrate that the primary signaling pathways
leading to p38 and NF-B activation can be dissociated, since many
stimuli activating either of these molecules are not overlapping. Although several stress-inducing agents such as cytokines, Taxol, mitogens, and UV irradiation activate both p38 and NF-
B (6, 13, 17,
19, 30), in particular, efficient agonists of p38 including arsenite
and hyperosmotic shock virtually do not activate NF-
B. Also other
classical stimuli of stress-responsive MAP kinases such as DNA-damaging
agents or heat shock were not found to trigger NF-
B
activation.3 This observation indicates
that no common cellular stress sensor or denominator of a stress
response exists where different forms of cellular stress converge. A
distinct responsiveness to cellular stresses is underlined by the fact
that PMA, which is a potent NF-
B activator in several cell types,
selectively induced p38 but not NF-
B activation in 293 cells.
In contrast to NF-B many stress stimuli were found to lead to the
activation of AP-1. This convergence of multiple stress stimuli at the
level of AP-1 is reasonable, because different forms of stimuli
preferentially affect distinct MAP kinase cascades (9, 41, 42). These
in turn all modulate AP-1 activation at different control points
including regulation at the transcriptional and post-translational
level and AP-1 subunit composition. The involvement of p38 in induction
of AP-1 has been implicated recently by the finding that inhibition of
p38 kinase prevented the synthesis of c-Fos and c-Jun in response to
stress stimuli (26).
ROI formation has been considered as a common denominator involved in
signal transduction of a variety of different forms of stress. Indeed,
activation of NF-B by all stimuli analyzed so far is inhibited by
antioxidants indicating that ROIs are the key messengers for NF-
B
activation (29, 30). Activation of ERK1 and ERK2, as well as JNKs in
response to growth factors, and cytokines has also been shown to be
inhibited by antioxidants (51, 52). Thus, we were surprised to find
that p38 activity was not inhibited but even augmented by antioxidants.
An increase in p38 activity was paralleled by an antioxidant-induced
activation of AP-1 confirming earlier findings that several unrelated
antioxidants cause AP-1 activation (43, 44). Similar to AP-1 (43), p38 was also modestly activated by hydrogen peroxide. This is in line with
reports showing that, in contrast to JNKs and ERKs, the p38 pathway is
only weakly activated by several prooxidants such as hydrogen peroxide,
nitric oxide, or organic peroxides (53, 54). Although many stimulatory
conditions of JNKs overlap that of p38, there are other examples of
distinct modes of regulation (55, 56). A physiological correlate of
antioxidant treatment may be hypoxia. Interestingly, it has been
observed that in hypoxia-reperfusion treatment of cells JNKs are
specifically activated during reperfusion where ROI formation occurs,
whereas p38 is only activated in initial hypoxia (57, 58). Thus, a
shift in the cellular redox balance may allow cells to rather
selectively trigger p38 or the JNK and ERK pathways.
An independent regulation of the pathways leading to activation of p38
and NF-B was found by studying the involvement of small GTPase
proteins. Members of the Rho subfamily such as Rac and Cdc42 that
mediate p38 and JNK activation (4, 5) are specifically inhibited by the
Clostridium enterotoxin toxin A (45). Whereas toxin A did
not interfere with NF-
B DNA binding activity, p38 activation was
strongly reduced.
Although we show that p38 and NF-B activation are not correlated
indicating divergent primary pathways, our results also suggest that
convergence may occur further downstream. Thus, a dominant-negative
mutant of MKK6 as well as a p38-specific pharmacological inhibitor
strongly abrogated NF-
B activity. However, activation of NF-
B DNA
binding activity was not modulated suggesting that p38 selectively
interfered with the transactivation potential of NF-
B. Using
in vitro kinase assays we show that in contrast to ATF-2
neither I
B-
nor the NF-
B DNA binding subunits p50 or the
transactivating C-terminal half of p65 are phosphorylated by p38. The
failure to phosphorylate p65 excludes the possibility that p38 directly
modulates the transactivation potential of NF-
B, which has been
shown to become activated by phosphorylation at one of the two p65
transactivation domains (59, 60). However, it is still conceivable that
a kinase downstream in the p38 pathway, such as a member of the
MAPKAP-K family, phosphorylates NF-
B subunits.
NF-B has been shown to physically and functionally interact with a
number of unrelated transcription factors including AP-1, C/EBP, SRF,
members of the ATF/CREB family, steroid receptors, TFII-D, and
coactivator proteins (reviewed in Refs. 59 and 61). To investigate
whether transcription factors that may become activated by p38
cooperate with the transactivating potential of NF-
B, we analyzed
the effect of p38 and ATF-2 overexpression in NF-
B reporter gene
assays. In these experiments we did not find an increase of
NF-
B-dependent gene expression, although we cannot exclude that endogenous p38 may exert an intrinsic stimulatory effect.
A possible explanation for a convergence of both pathways may be a
broader stimulatory effect of p38 on transcriptional and
post-transcriptional mechanisms. Inhibition of p38 activation by
dominant-negative MKK6 or the related MKK3 has been reported to inhibit
gene expression driven by several promoters and controlled by several
transcription factors, among them ATF-2 and Elk-1 (25). However, p38
activators do not generally increase gene expression as certain
promoters, such as the SV40 promoter, are not affected (25). This
certainly indicates some mode of selectivity of the p38 pathway on gene
expression.
In summary, our data indicate that the stress-responsive pathways
involved in p38 and NF-B activation are primarily separate. Obviously, one important stress response is selectively transduced through ROIs culminating in the activation of NF-
B in the cytoplasm; other types of cellular stresses initiate an independent pathway that
is mediated by GTPase proteins. Convergence of both pathways may occur
further downstream in the cell nucleus at the level of NF-
B-mediated
transactivation.
We thank Drs. R. Beyaert, R. J. Davis, N. Jones, C. V. Paya, R. J. Ulevitch, and L. E. Zon for plasmid constructs, Drs. F. Hofmann and K. Aktories for providing purified toxin A, and Dr. J. Lee for SB203580. We also thank Dr. R. Kapeller for helpful advice.