(Received for publication, July 8, 1996, and in revised form, February 3, 1997)
From the Department of Biochemistry, Institute of Medical Biology, University of Tromsø, 9037 Tromsø, Norway
The transformed phenotype of v-Ras- or
Bacillus cereus phosphatidylcholine-hydrolyzing
phospholipase C (PC-PLC)-expressing NIH 3T3 cells is reverted by
expressing a kinase-defective mutant of protein kinase C (
PKC).
We report here that extracellular signal-regulated kinase (ERK)-1 and
-2 are constitutively activated in v-Ras- and PC-PLC-transformed cells
in the absence of added growth factors. Interestingly, the activated
ERKs were exclusively localized to the cell nucleus. Consistently, the
transactivating potential of the C-terminal domain of Elk-1, which is
activated upon ERK-mediated phosphorylation, was strongly induced in
serum-starved cells expressing v-Ras or PC-PLC. Reversion of v-Ras- or
PC-PLC-induced transformation by expression of dominant negative
PKC
abolished the nuclear ERK activation suggesting
PKC as a novel,
direct or indirect, activator of mitogen-activated protein kinase/ERK kinase in response to activated Ras or elevated levels of
phosphatidylcholine-derived diacylglycerol. Transient transfection
experiments confirmed that
PKC acts downstream of Ras but upstream
of mitogen-activated protein kinase/ERK kinase. We found both the
v-Ras- and PC-PLC-transformed cells to be insensitive to stimulation
with platelet-derived growth factor (PDGF). No detectable receptor
level, autophosphorylation, or superinduction of DNA synthesis could be
observed in response to treatment with PDGF. Reversion of the
transformed cell lines by expression of dominant negative
PKC
restored the receptor level and the ability to respond to PDGF in terms
of receptor autophosphorylation, ERK activation, and induction of DNA
synthesis.
A growing body of evidence suggests that the induction of the
mitogen-activated protein kinase (MAPK)1
pathway leading to activation of extracellular signal-regulated kinase
(ERK)-1 and -2 is essential for mitogenic signal transduction (1-4).
ERK1 and -2 are rapidly activated after growth factor stimulation or
overexpression of constitutively active Ras, Raf, or MAPK/ERK kinase
(MEK) (2, 4). Following activation the ERKs translocate to the nucleus
to phosphorylate their nuclear substrates (5-8). Results obtained by
the use of dominant interfering mutants and antisense RNA suggest that
activation of the ERKs may be required for proliferation of fibroblasts
(9). The only direct activators of the ERKs identified so far are MEK1
and -2 (2, 4, 10). These dual specificity kinases activate ERK1 and -2 by phosphorylating both the threonine and the tyrosine residue in the
sequence motif TEY (3). MEK1, but not MEK2, forms a ternary complex
with Ras and Raf-1 (11). The Raf-1 kinase has been regarded as the
major MEK1 activator in most cell systems (2, 12). However, it is now
evident that several proteins may contribute to MEK1 activation (10,
13-16). Of particular relevance to this study is the finding that the
atypical protein kinase C subtype (
PKC) is activated by tyrosine
kinase receptors (17) and is shown to phosphorylate and activate MEK1
in vitro as well as in vivo (6, 18, 19). We (20,
21) and others (22, 23) have previously used the notion
PKC for this
PKC subspecies cloned from Xenopus laevis. However, the more
recent description of
/
PKC (24, 25) identified this atypical PKC subspecies originating from Xenopus as
PKC (19).
Polypeptide-derived growth factors and activation of Ras or Src cause an increase in the hydrolysis of phosphatidylcholine (PC) resulting in a sustained elevation in intracellular levels of PC-derived diacylglycerol (DAG) (26-32). Chronic stimulation of PC hydrolysis by stable expression of the gene (plc) encoding Bacillus cereus phosphatidylcholine-hydrolyzing phospholipase C (PC-PLC) causes severe growth deregulation and morphological transformation of NIH 3T3 fibroblasts without activating Ras (20, 21). Moreover, expression of B. cereus PC-PLC was able to release NIH 3T3 cells from a block to proliferation imposed by expression of dominant negative N-17 Ras but not so when the block was due to expression of dominant negative Raf-1 (33, 34). Consistently, expression of dominant negative Raf-1 reverted the transformed phenotype induced by plc expression (20). Furthermore, addition of purified bacterial PC-PLC to quiescent NIH 3T3 cell cultures induced Raf-1 kinase activity, whereas an inhibitor of endogenous PC-PLC activity blocked Raf-1 activation in response to serum (33-35). These findings suggest that the generation of DAG by PC hydrolysis is located downstream of Ras but upstream of Raf in the mitogenic signal transduction pathway. Thus, PC-derived DAG may directly or indirectly be involved in the poorly defined activation of the Raf kinase family members (12).
Several proteins, including PKC, bind to Ras in a
GTP-dependent manner suggesting that several pathways are
involved in relaying Ras-mediated signals (20, 36-42).
PKC (the
human homolog is named
PKC) constitutes together with
PKC, the
atypical PKCs. These PKC subtypes contain only one cysteine-rich zinc
finger and is not activated by phorbol esters or Ca2+ ions
(43). A requirement for functional
PKC in the Ras-mediated insulin-induced maturation pathway in Xenopus oocytes and
for serum-activated DNA synthesis in murine fibroblasts has been
demonstrated (22, 23). Consistent with the notion of
PKC as a
downstream target of activated Ras, a dominant negative mutant of
PKC, as well as dominant negative Raf-1, was found to revert both
v-ras- and plc-induced transformation (20). The
atypical PKCs have been reported to be involved in several signaling
pathways activating different downstream components.
PKC is
important for NF
B activation in different cell lines, including NIH
3T3, probably through activation of an I
B kinase (18, 19).
Furthermore,
PKC plays a critical role during stromelysin promoter
activation by PDGF in fibroblasts (44). Recent experiments, using
antisense oligonucleotides against different PKC subtypes, have shown
that
PKC acts as a mediator of PDGF-induced
2-integrin gene expression in human dermal fibroblasts (45). Both
PKC and
PKC have been suggested as downstream
components for PI 3-kinase-mediated signaling (17, 46). Thus, atypical PKCs clearly play critical roles in several signaling pathways.
In this paper we report that ERK1 and -2 are constitutively activated
and localized to the nucleus in both v-ras- and
plc-transformed cells. Expression of a dominant negative
mutant of PKC abolished ERK activation in response to
v-ras or plc expression. Consistently, v-ras- and plc-induced GAL-ElkC-transactivation
was found to be dependent on functional
PKC but did not involve
activation of JNK. Moreover, the dominant negative mutant of
PKC
blocked v-ras but not activated MEK induction of ERK kinase
activity. We also show that v-ras or plc
transformation lead to PDGF receptor down-regulation leading to
abolished superinduction of DNA synthesis in response to PDGF. This was
also the case for a cell line stably overexpressing an activated mutant
of MEK1. Reversion of the v-ras- and
plc-transformed phenotype by expression of a dominant
negative mutant of
PKC re-established the receptor level and
PDGF-induced receptor autophosphorylation, ERK kinase activation, and
induction of DNA synthesis.
NIH 3T3
fibroblasts (passage 123) were purchased from the American Type Culture
Collection (ATCC CRL 1658) and cultured in Dulbecco's modified
Eagle's medium (DMEM) supplemented with 10% calf serum (HyClone,
Logan, UT), penicillin (100 units/ml), and 100 µg/ml streptomycin
(Life Technologies, Inc.) in a CO2 incubator (5%
CO2) at 37 °C. NIH 3T3 cells transformed with the
v-Ha-ras or v-src oncogenes (26, 28) were grown
in the same medium. A well characterized PC-PLC expressing clone (clone
P18)(20) was grown in the presence of hygromycin B (Calbiochem) at 300 µg/ml, whereas the doubly transfected cell lines,
v-ras-dnPKC and plc-dn
PKC (20), were grown
in the presence of both hygromycin B and 400 µg/ml G418 (Life
Technologies, Inc.).
Measurements of the induction of DNA synthesis were performed as described previously (21). The MEK inhibitor PD 098059 (New England Biolabs) was dissolved in dimethyl sulfoxide and added to the the cell cultures to a final concentration of 0.5% (v/v) of vehicle.
Preparation of Whole Cell, Cytosolic, and Nuclear ExtractsFor preparation of cytosolic extracts, serum-starved or
PDGF-stimulated cell cultures were rinsed twice with ice-cold
phosphate-buffered saline and lysed in the dishes by adding ice-cold
lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM
NaCl, 2 mM EDTA, 1 mM EGTA, 1% Triton X-100,
25 µg/ml leupeptin, 25 µg/ml aprotinin, 1 mM
phenylmethylsulfonyl fluoride, and 1 mM
Na3VO4) for 30 min on ice. The lysates were collected with a rubber policeman, and insoluble material was removed
by centrifugation in a microcentrifuge for 5 min at 13,000 rpm. The
supernatant was aliquoted and stored at 70 °C. The protein concentrations in the cytosolic extracts were determined using a
detergent-compatible protein assay kit (Bio-Rad DC Protein Assay) with
bovine serum albumin as the standard.
Whole cell and nuclear extracts were prepared as described by Westwick
and Brenner (47) and by Sjøttem et al. (48), respectively, in the presence of the following protease/phosphatase inhibitors: 2 µg/ml aprotinin, 40 µg/ml bestatin, 0.5 µg/ml leupeptin, 0.5 mM phenylmethylsulfonyl fluoride, 0.7 µg/ml pepstatin A,
20 mM -glycerophosphate, 50 µM
Na3VO4, and 10 mM
p-nitrophenyl phosphate (all from Sigma).
Activation of ERK1/2 in situ was determined using a rabbit polyclonal antibody raised against a synthetic phosphotyrosine peptide corresponding to residues 196-209 of human ERK1/p44 MAP kinase (9101, New England Biolabs, Inc., Beverly, MA). This antibody specifically recognizes activated ERK1 and -2. The different cell lines were seeded into 8-well chamber slides (Nunc, Inc., Naperville, IL) at subconfluent cell densities (less than 30% confluence) in DMEM supplemented with 10% calf serum and left for 24 h. The cultures were subsequently serum-starved for 20 h in DMEM supplemented with 0.1% calf serum and either stimulated with 10 ng/ml PDGF (BB homodimer, Sigma) for 15 min or left untreated. The immunostaining was performed essentially as described in the protocol obtained from the supplier (New England Biolabs). Briefly, the cells were fixed in 3% paraformaldehyde, permeabilized using a buffer containing 0.1% Triton X-100, and preincubated for 1 h in 5.5% normal horse serum. The primary antibody was added for 48 h at 4 °C in a humidified chamber, and the staining was developed using biotinylated secondary antibodies and preformed streptavidin-peroxidase complexes in a Ni(II)-enhanced DAB reaction (Vectastain, Vector Laboratories).
ImmunoblottingPhosphorylated/activated ERK1 and -2 were
specifically detected using the phospho-ERK-specific antibody described
above. The total level of ERK2 was determined using a monoclonal
anti-ERK2 antibody (clone B9, Upstate Biotechnology Inc.) recognizing
both phosphorylated and non-phosphorylated forms. The cell cultures were treated as indicated in the respective figure legends. Thrombin (-thrombin) was purchased from Hoffman-La Roche. The cell cultures were harvested directly into 1 × SDS-PAGE gel load buffer and immediately heated to 95 °C for 5 min and sonicated briefly on ice.
The extracts (20 µl, corresponding to 2 × 105 cells
per lane) were separated by SDS-PAGE, electrotransferred onto
polyvinylidene difluoride membranes (Millipore), and developed following protocols obtained from the respective suppliers of antibodies. Tyrosine phosphorylation of the PDGF receptor was determined using a phosphotyrosine-specific monoclonal antibody following the protocol obtained from the supplier (clone 4G10, Upstate
Biotechnology Inc.). To verify protein loading, some of the immunoblots
were stripped for 2 h in 0.2 M glycine, pH 2.4, 1%
SDS at 65 °C, blocked, and reprobed with the anti-ERK2 antibody. The
level of PDGF receptor was determined as described by Vaziri and Faller
(49) using a polyclonal PDGF receptor antibody (PDGFR-
-specific, Santa Cruz Biotechnology). All immunoblots were developed using alkaline phosphatase-conjugated secondary antibodies and the
chemiluminescent substrate CDP-Star (New England Biolabs or Boehringer
Mannheim).
ERK activities in cytosolic or nuclear
extracts were determined as described by Sale et al. (50).
The activity of hemagglutinin (HA)-tagged ERK1 in extracts from
transient transfected cell lines was measured by an immune complex
kinase assay using MBP as the substrate. Subconfluent cell cultures
were transfected using lipofectamine (Life Technologies, Inc.)
according to the instructions of the manufacturer. Following 4 h
of incubation with DNA, the cells were incubated for 24 h in the
presence of 10% serum and then starved in 0.1% serum for another
24 h. Preparation of cell extracts and immunoprecipitation of
HA-ERK1 using a monoclonal antibody recognizing the HA epitope (12CA5,
Boehringer Mannheim) was performed as described (6). The immune
complexes were washed three times with cell lysis buffer containing 0.5 M NaCl and twice with MBP kinase buffer (20 mM
HEPES, pH 7.4, 10 mM MgCl2, 0.1 M
Na3VO4). The complexes were resuspended in
kinase buffer containing 1 mM dithiothreitol and 0.3 mg/ml
MBP, and kinase reactions were initiated by adding
[-32P]ATP (0.1 mM, 5 cpm/fmol final).
Following incubation for 15 min at 30 °C, the kinase reactions were
terminated by adding 5 × SDS-PAGE gel load buffer and boiled
immediately for 5 min. The phosphorylated proteins were separated on a
12.5% polyacrylamide gel and electrotransferred to a polyvinylidene
difluoride membrane (Millipore). The phosphorylated proteins were
detected by autoradiography and quantitated using a PhosphorImager
(Molecular Dynamics). The activation status of the c-Jun N-terminal
kinase (JNK) in the different cell lines was measured by a solid-phase
kinase assay. To make GST-Jun5-115 a sequence encoding
amino acids 5-115 of c-Jun was amplified by polymerase chain reaction
(5
-GAATTGGATCCATGGAAACGACCTTCTATGAC-3
and
5
-GAATTCTCGAGTGCTCATCTGTCACGTTCTTG-3
) and inserted into the
BamHI and XhoI sites of pGEX-4T-3 (Pharmacia
Biotech Inc.) producing an in-frame fusion with glutathione
S-transferase (GST). The fusion protein
(GST-Jun5-115) was expressed in Escherichia coli LE392 (51). GST-Jun5-115 was coupled to
glutathione-agarose beads and stored as a 20% suspension in NETN
buffer (20 mM Tris-HCl, pH 7.4, 100 mM NaCl, 1 mM EDTA, 5 mM phenylmethylsulfonyl fluoride, 5 mM benzamidine, 0.5% (v/v) Nonidet P-40, 2 mM
dithiothreitol, 2 µg/ml aprotinin, 0.5 µg/ml leupeptin, and 0.7 µg/ml pepstatin A). The kinase assays were performed as described by
Westwick and Brenner (47). Phosphorylated proteins were resolved on
10% SDS-polyacrylamide gels that were dried and subjected to
autoradiography at
70 °C with intensifying screens. Equal protein
loading was verified by immunoblotting with an antibody against JNK1
(C-17; Santa Cruz Biotechnology).
To measure the activity of the C-terminal transactivation domain of Elk-1 in the different cell lines, pGAL4-ElkC was made by inserting a BglII-XbaI fragment from pBS-Elk-1 into the BamHI and XbaI sites of pSG424 (52). Subconfluent cultures of the different cell lines in 100-mm diameter Petri dishes were transfected with 2 µg of pGAL4-ElkC or pSG424 (vector control), 2 µg of the reporter vector pG5E1bTATA-CAT (53), and 2 µg of salmon sperm carrier DNA using the calcium phosphate coprecipitation method. The precipitates were left on the cells for 4 h in medium containing 10% serum, after which the cells were glycerol-shocked for 90 s in 15% (v/v) glycerol in HBS buffer (137 mM NaCl, 5 mM D-glucose, 0.9 mM NaH2PO4, 21 mM HEPES, pH 7.08). The cells were washed three times with DMEM and serum-starved for 48 h in 0.1% serum. The cells were harvested and chloramphenicol acetyltransferase assays were performed as described previously (20).
We have previously shown that
chronic stimulation of PC hydrolysis by expression of B. cereus PC-PLC caused induction of DNA synthesis, enhanced
proliferation in the absence of added growth factors, and led to
transformation of NIH 3T3 cells (21). Expression of a dominant negative
mutant of PKC (dn
PKC) led to reversion to a non-transformed,
normal phenotype of both plc- and
v-ras-expressing cells (20). Here, we wanted to further
evaluate the signal mediators downstream of PC-derived DAG. Recently,
PKC was found to activate MEK1 both in vitro and in
vivo (6). To begin elucidating if the MEK/ERK pathway is involved
in transduction of the mitogenic signal generated by PC-derived DAG, we
used a novel inhibitor, PD098059, that specifically interferes with the
activation of MEK1 and MEK2 (54). The ability of plc- and
v-ras-expressing cells to induce DNA synthesis in the
absence of added growth factors was completely lost in the presence of
PD098059 (Fig. 1A). Also, prolonged treatment
of the plc- and v-ras transformed cell lines with
this inhibitor caused complete reversion to a normal, non-transformed phenotype (data not shown), indicating that MEK is a necessary downstream component for mitogenic signaling and transformation mediated by PC-derived DAG and v-ras. However, we found no
significant increase in cytosolic ERK activity in serum-starved cell
cultures of v-ras- or plc-transformed cells (Fig.
1B). Furthermore, treatment of these cell lines with PDGF
did not induce increased ERK activity. Consistently, immunoblotting
analyses of whole cell extracts using an antibody specifically
recognizing phosphorylated and activated ERK1 and -2 did not reveal any
potent ERK activation in response to stable expression of either
v-ras or plc (Fig. 1C). Interestingly, unlike their parental transformed cells, the cell lines reverted by
expression of dn
PKC-induced phosphorylation of myelin basic protein
in response to PDGF (Fig. 1B). The reverted cell lines also
consistently showed a pronounced PDGF-induced phosphorylation of ERK1
and -2 (Fig. 1C). The activation of ERK1 and -2 in response to serum was indistinguishable between the different cell lines, suggesting that the differential activation of ERK in transformed and
reverted cell lines was specific for PDGF-induced signaling.
Activation of the MAPK pathway is accompanied by translocation of ERK1
and -2 to the nucleus (4-8). Thus, we next analyzed the ERK activity
in nuclear extracts from cells expressing v-ras or
plc alone or together with dnPKC (Fig. 2).
The basal nuclear ERK activity was increased in both v-ras-
and plc-expressing cells compared with quiescent NIH 3T3
cells. However, the increased basal ERK activity displayed by the
transformed cells was inhibited by expression of dn
PKC. To further
analyze this, serum-starved or PDGF-stimulated cell cultures were
immunostained with an antibody specifically recognizing phosphorylated
ERK1 and -2 (Fig. 3). These experiments confirmed that
ERK1 and -2 were constitutively activated both in v-ras- and
in plc-transformed cell lines. Moreover, phosphorylated ERKs
were exclusively localized to the nuclei. Treating the transformed cell
lines with PDGF for 15 min did not further increase the intensity of
the nuclear staining (data not shown). Reversion of the
v-ras- or plc-induced transformation by
expression of dn
PKC abolished the chronic activation of the nuclear
ERKs. However, the phosphorylation of ERK1 and -2 in the reverted cell
lines could be induced by PDGF. These data suggest that expression of
v-ras as well as increased intracellular levels of
PC-derived DAG caused by expression of B. cereus PC-PLC both result in activation and nuclear translocation of ERK1 and -2. Furthermore, the activation of ERKs in response to Ras and PC-PLC is
dependent on functional
PKC.
Expression of dn
We have previously shown that dnPKC
abolished the Ras- or PC-PLC-induced activation of both NF-
B and
AP-1 (20). However, since we found constitutive activation and nuclear
translocation of ERKs in Ras- and PC-PLC-transformed cell lines, the
activation status of the transcription factor Elk-1 was of interest.
For this purpose, a chimeric transcription factor composed of the DNA
binding domain of yeast GAL4 and the transactivating domain of Elk-1
(GAL4-ElkC) was constructed. The transactivation potential of this
nuclear fusion protein is strongly enhanced by specific MAPK
phosphorylation within the transactivation domain of Elk-1 (55).
Consistent with the nuclear localization of activated ERKs in the
transformed cell lines, we found that the transactivating potential of
GAL4-ElkC was strongly induced in response to transformation by
v-ras or plc (Fig. 4A).
Reversion of the transformed cell lines by expression of dn
PKC
coincided with a large decrease in the transactivation potential of
GAL4-ElkC down to the background level seen in quiescent NIH 3T3 cells.
Thus, both v-ras and plc induced the
transactivating potential of Elk-1 in a
PKC-dependent manner. Recently, it was reported that another subclass of the MAPK
family, termed the c-Jun N-terminal kinases (JNKs) or stress-activated protein kinases, phosphorylates and activates the transactivation domain of Elk-1 (56-58). To determine the activity of the JNKs in the
transformed and reverted cell lines, whole cell extracts were used in a
solid phase kinase assay using GST-Jun5-115 as a substrate
(Fig. 4B). GST-Jun5-115 was extensively
phosphorylated when treated with extract from UV-stimulated NIH 3T3
cells. However, we found no increase in JNK activity in
ras-transformed cells and only a slight increase in
plc-transformed cells. The same results were obtained by
carrying out immune complex kinase assays using an antibody
specifically recognizing JNK1 (data not shown). Prolonged treatment of
ras- and plc-transformed cells with the specific
MEK inhibitor PD 098059 totally abolished the GAL4-ElkC activity (data
not shown). Altogether, these results clearly suggest that the
increased transactivation potential observed for GAL4-ElkC is due to
activation by nuclear ERK1 and -2 with very little, if any,
contribution by JNKs.
Expression of dn
The expression of dnPKC blocked ras- and
plc-induced activation of ERK as well as Elk-1
transactivation suggesting that
PKC acts downstream of Ras and
PC-PLC but upstream of ERK. To more firmly localize
PKC with respect
to MEK, we performed transient transfections measuring the activation
of HA-tagged ERK1 (59) in response to Ha-Ras V-12 (22) or an activated
mutant of MEK1 (aMEK) (60), in the presence or absence of dn
PKC
(19). The activation of ERK1 induced by Ras was blocked by
cotransfection with the dn
PKC expression vector. However, dn
PKC
did not affect MEK1-induced activation of ERK1 in NIH 3T3 cells (Fig.
5). Immunoblot analysis of cell extracts from the
different transfected cell cultures showed that the differences in ERK
activity were not due to differences in the expression of HA-tagged
ERK1 (data not shown). Collectively, these results suggest that
PKC
is located downstream of Ras but upstream of MEK in the MEK/ERK
pathway. To further evaluate the effect of dn
PKC on MEK-mediated
signaling in NIH 3T3 cells, we first made a cell line that stably
overexpressed aMEK. The aMEK expressing cells displayed a transformed
phenotype, induced DNA synthesis in the absence of added growth
factors, showed a constitutively increased nuclear ERK activity and an increased GAL-ElkC transactivation potential (data not shown). However,
several attempts to establish clones that stably coexpressed dn
PKC
and aMEK failed. Of more than 50 clones analyzed from two experiments
performed with two different expression vectors for dn
PKC, none
expressed dn
PKC. All of them had retained the transformed phenotype
(data not shown). This indicates that dn
PKC interferes with the
survival of aMEK-transformed cells without affecting the activation of
ERK. Thus, in addition to the MEK/ERK pathway,
PKC also acts in
other signaling pathways (17-20, 22, 44-46, 61), some of which may be
critical for cell survival.
Expression of dn
While determining ERK activities in whole cell extracts, we
noted that the reverted cell lines responded more potently to PDGF than
their parental transformed cell lines did (see Fig. 1). Expression of
v-ras has previously been shown to cause suppression of PDGF
receptor autophosphorylation (49, 62-64). Several mechanisms may be
involved in this desensitization to PDGF stimulation including Ras-induced activation of protein-tyrosine phosphatase activity (64),
induction of a membrane-bound inhibitor of receptor phosphorylation (63), and reduced transcription of the gene encoding the receptor (49).
Thus, it was of interest to determine whether plc
transformation inhibited PDGF receptor autophosphorylation as found for
v-ras and if reversion of the transformed phenotype by
dnPKC also restored receptor autophosphorylation. We therefore
performed immunoblot analyses of tyrosine-phosphorylated proteins in
extracts prepared from cells that were either untreated or stimulated
with PDGF using thrombin stimulation as a negative control (Fig.
6A). Interestingly, a
phosphotyrosine-containing protein with a migration corresponding to
the PDGF receptor was only detected after PDGF stimulation of NIH 3T3
cells and the reverted cell lines. This tyrosine-phosphorylated protein
could not be detected after PDGF stimulation of the transformed cell
lines. Immunoblotting with an antibody against the PDGF receptor revealed complete down-regulation of the receptor in both
v-ras and plc-transformed cells, whereas
reversion of the transformed phenotype by expression of dn
PKC
coincided with restored receptor levels (Fig. 6B). PDGF
treatment of the v-ras- and plc-transformed cell
lines was not able to further increase the induction of DNA synthesis,
whereas addition of 10% serum caused a superinduction of DNA synthesis
in v-ras- and plc-transformed cells (Fig.
6C). This suggests that the insensitivity to PDGF was not
due to a maximum stimulation of DNA synthesis by Ras or PC-PLC. On the other hand, PDGF did induce DNA synthesis in the reverted cell lines.
Thus, reversion of the transformed cells by dn
PKC expression prevented the constitutive down-regulation of the PDGF receptor allowing receptor autophosphorylation and ERK activation in response to
PDGF. Since both v-ras and plc activate other
signaling pathways in addition to the MEK/ERK pathway, we looked at
PDGF receptor expression in cells stably transfected with aMEK. Western
blot analyses revealed down-regulation of the PDGF-
receptor showing that constitutive ERK activation is alone able to lead to receptor down-regulation (data not shown).
In this study we present evidence for chronic activation and
nuclear translocation of ERK1 and -2 in NIH 3T3 cells transformed with
v-ras or plc. Expression of a dominant negative
mutant of PKC blocked the activation of ERK1 and -2. A sustained
activation of ERK1 and -2 has been suggested to be a prerequisite for
proliferation of fibroblasts and differentiation of PC12 cells (4).
Interestingly, the ERKs located in the cytosol were not
phosphorylated/activated in the transformed cells. Previously, Gardner
et al. (65) reported no significant activation of MEK in
v-ras-transformed NIH 3T3 cells. These workers found only
modest constitutive activation of ERKs in Ras-transformed NIH 3T3 cells
and no persistent ERK activity in similarly transformed Rat 1a
fibroblasts (66). Their conclusion was based on the use of cytosolic
extracts. These previous results are therefore entirely consistent with
our present findings showing no constitutive activation of ERKs in the
cytosol of the transformed cells. However, we show here that both
v-ras- and plc-transformed cells contain
constitutively elevated nuclear ERK activity. This may suggest that the
mechanisms for terminating ERK activation by dephosphorylation mediated
by protein phosphatases are probably operating normally in the cytosol,
whereas the regulation of nuclear ERK activity is subverted in the
transformed cells. Our results together with the recent demonstration
that Ras transformation is inhibited by coexpression of
kinase-defective mutants of ERK1 and -2 (38) indicate that constitutive
nuclear ERK activity may be a prerequisite for transformation by
v-ras or plc. Thus, it seems logical to assume
that the transformed phenotype is dependent on persistent ERK-mediated
phosphorylation of transcription factors. Consistent with this notion
we found a potent activation of the chimeric transcription factor
GAL4-ElkC, which contains the DNA binding domain and nuclear
localization signal from yeast GAL4 and the transactivation domain of
Elk-1, in the transformed cells. Intriguingly, the ERKs do not contain
any known nuclear localization signals, and mutants that lack the TEY
motif (T192A, Y194F) can be translocated to the nucleus (5, 7, 67).
Therefore, although sustained ERK activation always seems to be
associated with nuclear translocation, transient activation does not
lead to nuclear translocation (4), and activation is not required for
nuclear localization.
The activation of ERKs in response to Ras and PC-PLC was blocked by
expression of a dominant interfering mutant of PKC. Thus,
PKC may
be critical for the Ras- and PC-PLC-mediated activation of MEK1.
Although Raf-1 has been regarded as the major MEK1 activator in most
cell systems, it is now evident that several protein kinases may serve
as MEK1 activators (10, 13-15, 18, 68). In fact, Raf-1 may not be an
important activator of MEK1 upon stimulation with serum or insulin
(15). Also, a MEK1 mutant unable to bind to Raf-1 and B-Raf was still
potently activated in response to serum, thrombin, and v-Ras (10). Of
direct relevance to our present results is the finding that
PKC was
able to phosphorylate and activate a MEK1 preparation from COS-1 cells
in vitro (18). Furthermore, by the use of transient
transfections, it was recently demonstrated that both MEK and ERK were
activated in vivo by expressing activated
PKC and that
the dominant negative mutant of
PKC severely impaired the activation
of both kinases following stimulation with serum or tumor necrosis
factor-
(6). The N-terminal regulatory domain of
PKC has been
found to bind to Ras in a GTP-dependent manner in
vitro and
PKC co-immunoprecipitated with Ras-GTP in vivo (37). Therefore, the full-length construct of a dominant interfering mutant of
PKC used in this and other studies (6, 20, 22,
37) could simply function to sequester Ras from interacting
productively with other downstream targets. However, cotransfection
studies using a mutant construct expressing only a catalytically
inactive C-terminal kinase domain of
PKC, unable to interact with
Ras-GTP, inhibited MEK1 activation as potently as the full-length
dominant negative mutant (6). Consistent with these observations we
found that the dominant negative
PKC mutant blocked Ras- but not
MEK1-induced ERK activation when assayed by transient overexpression.
Together, these results place
PKC upstream of MEK in Ras- and
PC-PLC-mediated signaling. However, several attempts to establish
clones that coexpressed dn
PKC with aMEK failed. This is not due a
blockade of the MEK/ERK pathway since transient transfections of aMEK
cells with a dn
PKC expression plasmid did not abolish the GAL-ElkC
activity nor did it affect the activation of HA-tagged ERK. This
suggests that dn
PKC interferes with the survival of aMEK-transformed
cells without affecting the activation of ERK. Interestingly, Moscat
and co-workers (61) have recently reported that the atypical PKCs are
clearly involved in cell survival. The zinc finger domain of both
PKC and
PKC was shown to interact with the product of the
par4 gene that is involved in growth inhibition and
induction of apoptosis. The interaction of Par4 with
PKC/
PKC
reduced the activity of the kinases dramatically. Furthermore,
overexpression of Par4 as well as dn
PKC/dn
PKC in NIH 3T3 cells
induced apoptosis, whereas the cells survived when Par4 was coexpressed
with the wild-type kinases. Altogether, these findings suggest that
atypical PKCs have a role in mediating cell survival signals. It has
been reported that PI 3-kinase is required for generating survival
signals in PC12 cells (69). Interestingly, the kinase activity of
PKC, which is closely related to
PKC, is stimulated in
vitro by phosphatidylinositol 3,4,5-triphosphate (46).
Furthermore, PI 3-kinase was recently suggested as an in
vivo activator of
PKC based on measurements of AP-1
transactivation (17). We found that reversion of Ras- or
PC-PLC-transformed cells by stable expression of dn
PKC was accompanied by loss of growth factor-independent AP-1- and
NF
B-mediated transactivation (20). Constitutively active mutants of
Ras relay signals through several parallel pathways (37, 38, 40, 70). Hydrolysis of phosphatidylcholine has also been reported to be involved
in several signaling pathways (20, 27, 71, 72). However, the ERKs are
the only known substrates for MEK. Since we were unable to establish
cell lines where activated MEK1 and dn
PKC were stably coexpressed,
activation of the ERKs is by itself evidently not sufficient to
overcome the apoptotic effect of overexpressing dn
PKC. We found that
the ERK activation in response to PDGF or serum was not blocked in
cells coexpressing Ras or PC-PLC and dn
PKC. Expression levels of
dominant interfering mutants that completely block
Ras-dependent signaling would clearly be lethal to the
cells. Therefore, clones that stably express cytotoxic levels of
dn
PKC cannot be established as stably transfected cell lines. Our
previous findings that expression of dn
PKC reverted v-ras
and plc-induced transformation and our present results, demonstrating a block in the activation (and nuclear translocation?) of
ERK1 and -2, implicate
PKC as an important mediator of ERK activation in response to Ras and PC-PLC. However, although
PKC acts
upstream of MEK1 in the MEK/ERK pathway, it is also required in other
pathways where one or more of these may be critical for cell
survival.
It was recently reported that murine fibroblasts transformed by either
the v-ras or v-src oncogenes or following
prolonged growth factor stimulation of normal cells expressed low
levels of PDGF receptors compared with quiescent nontransformed cells (49). The reduced receptor levels coincided with a lower transcription of the PDGF- receptor gene and reduced receptor mRNA levels. We
found that the down-regulation of the PDGF-
receptor in response to
v-ras was mimicked by transformation induced by elevated
levels of PC-derived DAG. Furthermore, overexpression of aMEK also led to a reduced PDGF-
receptor level. Reversion of both
v-ras and plc-transformed cells by expression of
dn
PKC restored the receptor levels. Taken together, our results
suggest that constitutive nuclear ERK activity leads to down-regulation
of the PDGF-
receptor perhaps via effects on transcription factors
regulating the transcription of the PDGF-
receptor gene.
Together with previous findings (6, 20), the results reported here
strongly suggest that PC-derived DAG acts via the ERK1 and -2 MAPK
pathway with PKC as a downstream mediator of MEK activation.
PC-derived DAG acts late in the G1 phase of the cell cycle
as suggested from the kinetics of induction of DNA synthesis in
quiescent fibroblasts treated with B. cereus PC-PLC (73).
Furthermore, in keratinocytes the production of PC-derived DAG is
inhibited by transforming growth factor-
(74) which induces
G1 arrest by inhibiting various cyclin-Cdk kinases through the cooperative action of at least two Cdk inhibitors (75). Interestingly, a recent report demonstrated that in the yeast cell
cycle the only PKC found in this organism, PKC1, functions downstream
of the activation of the CDC28 kinase at the late G1 restriction point START (76). Strikingly, the activation of PKC1, which
acts via a MAPK module involving a MEK kinase (BCK1), two MEKs (MKK1
and -2), and a MAPK (MPK1) (77), was linked to a CDC28-dependent
production of PC-derived DAG most probably due to activation of a
PC-PLC. These findings strongly support our notion outlined above
concerning the location of Ras and PC-derived DAG in the mitogenic MAPK
pathway and provide evidence that a signaling pathway generally thought
of as solely stimulated by membrane-bound receptors is also employed by
a cell cycle regulatory kinase in the absence of extracellular stimuli.
In light of the conserved nature of MAPK modules and the cell cycle
regulatory machinery, future work aimed at elucidating the functions of
PC-derived DAG will clearly profit from parallel experimental
approaches involving both mammalian and yeast cell systems.
We thank P. E. Shaw for the gift of pBS-Elk-1. We are grateful to J. Moscat for discussions and generous gifts of reagents. The skillful technical assistance of Randi Ystborg and Turid Holm is gratefully acknowledged.