(Received for publication, March 31, 1997, and in revised form, May 14, 1997)
From the Howard Hughes Medical Institute and the Departments of Medicine and Biochemistry, Duke University Medical Center, Durham, North Carolina 27710 and the ¶ Department of Molecular Biochemistry, Glaxo Wellcome Research and Development, Research Triangle Park, North Carolina 27709
Many receptors that couple to
heterotrimeric guanine-nucleotide binding proteins (G proteins) have
been shown to mediate rapid activation of the mitogen-activated protein
kinases Erk1 and Erk2. In different cell types, the signaling pathways
employed appear to be a function of the available repertoire of
receptors, G proteins, and effectors. In HEK-293 cells, stimulation of
either 1B- or
2A-adrenergic receptors (ARs) leads to rapid
5-10-fold increases in Erk1/2 phosphorylation. Phosphorylation of
Erk1/2 in response to stimulation of the
2A-AR is effectively
attenuated by pretreatment with pertussis toxin or by coexpression of a
G
subunit complex sequestrant peptide (
ARK1ct) and
dominant-negative mutants of Ras (N17-Ras), mSOS1 (SOS-Pro), and Raf
(
N-Raf). Erk1/2 phosphorylation in response to
1B-AR stimulation
is also attenuated by coexpression of N17-Ras, SOS-Pro, or
N-Raf,
but not by coexpression of
ARK1ct or by pretreatment with pertussis
toxin. The
1B- and
2A-AR signals are both blocked by
phospholipase C inhibition, intracellular Ca2+
chelation, and inhibitors of protein-tyrosine kinases. Overexpression of a dominant-negative mutant of c-Src or of the negative regulator of
c-Src function, Csk, results in attenuation of the
1B-AR- and
2A-AR-mediated Erk1/2 signals. Chemical inhibitors of calmodulin, but not of PKC, and overexpression of a dominant-negative mutant of the
protein-tyrosine kinase Pyk2 also attenuate mitogen-activated protein
kinase phosphorylation after both
1B- and
2A-AR stimulation. Erk1/2 activation, then, proceeds via a common Ras-, calcium-, and
tyrosine kinase-dependent pathway for both Gi-
and Gq/11-coupled receptors. These results indicate that in
HEK-293 cells, the G
subunit-mediated
2A-AR- and the
G
q/11-mediated
1B-AR-coupled Erk1/2 activation
pathways converge at the level of phospholipase C. These data suggest
that calcium-calmodulin plays a central role in the
calcium-dependent regulation of tyrosine phosphorylation by
G protein-coupled receptors in some systems.
GTP-binding protein (G
protein)1-coupled receptors (GPCRs)
comprise a family of heptahelical membrane-bound receptors that mediate
responses to a vast array of ligands (1). While the effects of these
receptors on intermediary metabolism have been extensively studied,
recent data have suggested that they play important roles in the
regulation of cell growth and differentiation. Constitutively
activating mutations of the thyrotropin and luteinizing hormone
receptors are associated with hyperfunctioning thyroid adenomas and
idiopathic male precocious puberty (1, 2). Expression of a
constitutively active mutant of the 1B-adrenergic receptor (AR) in
myocardial cells induces myocardial hypertrophy in transgenic animals
(3), and
1-adrenergic agonists stimulate hypertrophy in cultured
neonatal rat ventricular myocytes (4).
Mitogen-activated protein (MAP) kinases represent a point of convergence for cell surface signals regulating cell growth and division. The MAP kinases comprise a family of serine/threonine kinases, which include the extracellular signal-regulated kinases Erk1 and Erk2, the Jun N-terminal kinase/stress-activated protein kinase, and p38mapk (5). MAP kinases are regulated via protein phosphorylation cascades whose basic pattern has been highly conserved throughout evolution. In the mammalian Erk1/2 pathway, the proximal kinases Raf-1 and B-Raf phosphorylate and activate the dual function threonine/tyrosine kinases MAP/Erk kinases 1 and 2, which in turn phosphorylate Erk1/2. Once phosphorylated, activated Erk1/2 translocate to the cell nucleus, where they phosphorylate and activate nuclear transcription factors (6). Many signals received at the cell surface, including those mediated by growth factor receptor tyrosine kinases (7) and integrins, which mediate cell adhesion (8), initiate the MAP kinase cascade via activation of the low molecular weight GTP-binding protein, p21ras (9). Association with GTP-bound p21ras localizes Raf to the plasma membrane, which is sufficient to induce its activation (10).
Recently, receptors that couple to heterotrimeric G proteins, including
the lysophosphatidic acid (LPA) (11, 12), bombesin (13), thromboxane
A2/prostaglandin H2 (14), prostaglandin F2 (15),
-thrombin (16), angiotensin II (17, 18),
1B-adrenergic (13),
2A-adrenergic (13, 19), M1 muscarinic acetylcholine (13), D2
dopamine (13), and A1 adenosine (13) receptors, have been shown to
activate MAP kinases (13, 20, 21). The signal transduction pathways
employed by these receptors are heterogeneous. In Rat-1 and COS-7
cells, receptors coupled to pertussis toxin-sensitive G proteins
mediate Erk1/2 activation via a G
subunit complex-mediated
pathway that is dependent upon tyrosine protein phosphorylation and
p21ras activation (13, 19, 22). These signals are independent of receptor-mediated effects on phosphatidylinositol hydrolysis, calcium influx, or inhibition of adenylyl cyclase (22, 23). In
contrast, receptors coupled to pertussis toxin-insensitive G proteins
mediate Erk1/2 activation via a G
subunit pathway that is
p21ras-independent and may involve PKC (13). Direct activators
of PKC, such as phorbol esters, have been reported to stimulate
activation of MAP kinases through both
p21ras-dependent and -independent pathways (4,
9).
Significant heterogeneity may also exist between cell types. Activation
of Erk1/2 by 1-adrenergic receptors in neonatal rat ventricular
myocytes (4) and by prostaglandin F2
receptors in
NIH-3T3 cells (15) is G
q/11-mediated and
p21ras-dependent, suggesting that
G
q/11 subunits also activate p21ras in some cell
types. This pathway differs markedly from the Gq/11-coupled receptor-mediated p21ras-independent MAP kinase
activation that has been described in COS-7 cells (13). In this paper,
we characterize the mechanisms of Erk1/2 activation employed by the
Gi-coupled
2A- and by the Gq/11-coupled
1B-adrenergic receptors, heterologously expressed in HEK-293 cells.
We find that both receptors mediate
p21ras-dependent Erk1/2 activation via
phospholipase C and calcium-dependent activation of Src
family kinases. These data suggest that, in some cell types, the
G
subunit complex-dependent
2A-AR and G
q/11 subunit-dependent
1B-AR signals
converge at the level of PLC and proceed via a common,
p21ras-dependent, signaling pathway.
Phorbol 12-myristate 13-acetate (PMA) and
1,2-bis(2-aminophenoxy)ethane-N,N,N,N
-tetraacetic
acid (BAPTA) were from Sigma. UK-14304 was from Pfizer. Fluphenazine,
calmidazolium, U73122, A23187, bisindolylmaleimide I (GFX), genistein,
and herbimycin A were from Calbiochem. Ophiobolin A was from Biomol.
Pertussis toxin was from List Biologicals. Rabbit polyclonal anti-Pyk2
IgG was a kind gift of J. Schlessinger.
The cDNAs for the 1B- and the
2A-adrenergic receptors were cloned in our laboratory (24, 25). The
-adrenergic receptor kinase 1 carboxyl-terminal (
ARK1ct)
peptide-encoding minigene, containing cDNA encoding the
carboxyl-terminal 195 amino acids of
ARK1, and the dominant-negative
SOS-Pro construct, encompassing the proline-rich carboxyl-terminal
fragment of mSOS1, were prepared in our laboratory as described
previously (19, 26). The cDNA encoding a constitutively active
mutant of G
q (G
q-Q209L), as described
previously (27), was prepared in our laboratory by R. Premont. The
cDNA encoding a constitutively active mutant of G
i2
(G
i2-Q204L) was from H. Bourne. The cDNAs encoding
G
1 and G
2 were from M. Simon. The
p21N17ras dominant negative mutant was from D. Altschuler, the
p74raf-1 dominant negative mutant (
NRaf) was from L. T. Williams, the p112pyk2 dominant negative mutant (PKM) was from
J. Schlessinger, p60c-src was from D. Fujita, and
p50csk was from H. Hanafusa. The constitutively activated Y530F
p60c-src (TAC(Y)
TTC(F)), in which the
regulatory carboxyl-terminal tyrosine residue has been mutated, and
catalytically inactive K298M p60c-src (AAA(K)
ATG(M)) were prepared as described (28-31). All cDNAs were
subcloned into pRK5, pcDNA, pCMV, or pRS
eukaryotic expression vectors for transient transfection.
HEK-293, Rat-1, and PC12 cells were from the American Type Culture Collection. HEK-293 cells were maintained in minimum essential medium with Earle's salts (Life Technologies, Inc.) supplemented with 10% fetal bovine serum (Life Technologies) and 100 µg/ml gentamicin (Life Technologies), at 37 °C in a humidified 5% CO2 atmosphere. Rat-1 cells were maintained in Dulbecco's modified Eagle's medium (Life Technologies) supplemented with 10% fetal bovine serum and 100 µg/ml gentamicin under similar conditions. PC12 cells were maintained in RPMI medium 1640 (Life Technologies) supplemented with 10% heat-inactivated horse serum (Life Technologies), 5% fetal bovine serum, 100 µg/ml gentamicin, and 20 µg/ml L-glutamic acid (Life Technologies) under similar conditions. Transfections of HEK-293 cells were performed on 80-90% confluent monolayers in six-well dishes. Cells were transfected using the calcium phosphate coprecipitation method as described previously (32). Empty pRK5 vector was added to transfections as needed to keep the total mass of DNA added per well constant within an experiment.
Prior to stimulation, transfected monolayers were serum-starved in minimum essential medium with Earle's salts (HEK-293 cells) or Dulbecco's modified Eagle's medium (Rat-1 cells) supplemented with 0.1% bovine serum albumin (fraction V, protease-free) (Boehringer Mannheim), 100 µg/ml gentamicin, and 10 mM HEPES, pH 7.4, for approximately 24 h.
Pyk2 ImmunoblottingUnstimulated PC12 and HEK-293 cell monolayers were lysed directly with 100 µl/well Laemmli sample buffer. Cell lysates were sonicated briefly, and approximately 30 µg of protein/lane were loaded for resolution via SDS-polyacrylamide gel electrophoresis. Pyk2 was detected by protein immunoblotting using a 1:1000 dilution of rabbit polyclonal anti-Pyk2 IgG with horseradish peroxidase-conjugated goat anti-rabbit IgG (Santa Cruz Biotechnology) as secondary antibody. Chemiluminescent detection of Pyk2 was performed after development of membranes with ECL reagent (Amersham Corp.), according to the manufacturer's instructions and exposure to Biomax XR scientific imaging film (Eastman Kodak Co.).
MAP Kinase Assay and ImmunoblottingStimulations were
carried out at 37 °C in serum-starving medium as described in the
figure legends. After stimulation, monolayers were lysed directly with
100 µl/well Laemmli sample buffer. Cell lysates were sonicated
briefly to disrupt DNA, and proteins (30 µg/lane) were resolved by
SDS-polyacrylamide gel electrophoresis. Phosphorylation of Erk1/2 was
detected by protein immunoblotting using a 1:1000 dilution of rabbit
polyclonal phospho-specific MAP kinase IgG (New England Biolabs) with
alkaline phosphatase-conjugated goat anti-rabbit IgG (Amersham) as
secondary antibody. Quantitation of Erk1/2 phosphorylation was
performed after development of membranes with Vistra ECF reagent
(Amersham) by scanning on a Storm PhosphorImager (Molecular Dynamics).
After scanning, membranes were treated for 30 min with 40% methanol to
remove the Vistra ECF reagent, stripped by treatment with stripping
buffer (62.5 mM Tris-Cl, pH 6.8, 2% SDS, 100 mM -mercaptoethanol) for 30 min at 50 °C, and
reprobed with rabbit polyclonal anti-Erk2 IgG (Santa Cruz
Biotechnology) to quantitate total p42mapk.
LPA receptor-mediated Erk1/2
activation in Rat-1 fibroblasts is mediated by G subunits derived
from PTX-sensitive G proteins (33) and is independent of changes in
intracellular cAMP, calcium, or PKC (23). As shown in Fig.
1A, stimulation of endogenous LPA or thrombin
receptors in these cells resulted in a 3-6-fold increase in Erk1/2
phosphorylation, which was completely inhibited by treatment with PTX.
Acute activation of PKC by treatment with phorbol ester resulted in a
less than 2-fold increase in Erk1/2 phosphorylation. Exposure to the
calcium ionophore A23187 resulted in a less than 2-fold increase in
Erk1/2 phosphorylation.
HEK-293 cells exhibit a distinct pattern of Erk1/2 activation. As shown in Fig. 1B, Erk1/2 phosphorylation via endogenous LPA and thrombin receptors is mediated via both PTX-sensitive and -insensitive G proteins in these cells. LPA and thrombin induce a similar 8-10-fold increase in Erk1/2 phosphorylation. Like Rat-1 cells, the LPA signal is PTX-sensitive. In contrast, the thrombin receptor mediates PTX-insensitive Erk1/2 phosphorylation, indicating that a distinct Erk1/2 activation pathway, mediated by PTX-insensitive G proteins, exists in these cells. Also, acute stimulation of HEK-293 cells with phorbol ester or with A23187 resulted in 24- and 15-fold stimulations of Erk1/2 phosphorylation, respectively, in stark contrast to the above results obtained in Rat-1 cells.
Erk1/2 Phosphorylation in HEK-293 Cells Is Mediated by GTo
characterize the mechanisms of Erk1/2 activation via PTX-sensitive and
insensitive G proteins in HEK-293 cells, we employed a transiently
transfected model system in which Gi-coupled 2A-AR and
Gq/11-coupled
1B-AR were heterologously expressed. As
shown in Fig. 2A, stimulation of the
2A-AR
resulted in PTX-sensitive Erk1/2 phosphorylation, while
1B-AR-mediated Erk1/2 phosphorylation in response to stimulation was
insensitive to pretreatment with PTX. Since G
subunits mediate
Erk1/2 activation by several GPCRs, we determined whether cellular
expression of a G
-sequestrant polypeptide derived from
ARK1ct
would inhibit
1B-AR- and
2A-AR-mediated MAP kinase activation. As
shown in Fig. 2B, expression of the
ARK1ct peptide
attenuated Erk1/2 phosphorylation in response to
2A-AR, but not
1B-AR, stimulation. EGF-stimulated Erk1/2 phosphorylation was not
sensitive to pretreatment of cells with PTX or to overexpression of
cDNA coding for the
ARK1ct peptide. This suggests that the
2A-AR signals primarily via the G
subunit complex from
PTX-sensitive G proteins, whereas the
1B-AR signal is mediated by
the G
subunit from PTX-insensitive G proteins. As shown in Fig.
2C, overexpression of a constitutively active mutant of
G
q (G
q-Q209L), but not of
G
i2 (G
i2-Q204L), was sufficient to induce
Erk1/2 phosphorylation. Overexpression of
G
1
2 resulted in a consistent 2-fold
stimulation of Erk1/2 phosphorylation, unlike the 6-8-fold
stimulations of MAP kinase activity observed previously in COS-7 cells
(13, 34).
PTX and ARKct sensitivity of
1B-AR- and
2A-AR-mediated Erk1/2 phosphorylation in HEK-293 cells and effect of
overexpression of constitutively active mutant forms of
G
q and G
i2 and of wild type G
and G
on Erk1/2 phosphorylation. A, HEK-293 cells were transiently
transfected with plasmid DNA encoding
2A-AR (0.2 µg/well),
1B-AR (0.2 µg/well), or empty vector. Where indicated, cells were
preincubated overnight with PTX. Serum-starved cells were stimulated
with 10 µM UK-14304 (
2A-AR), 20 µM
phenylephrine (
1B-AR), or EGF for 5 min prior to determination of
Erk1/2 phosphorylation. B, HEK-293 cells were transiently
transfected as in A with the addition of either 1 µg/well
plasmid DNA coding for
ARK1ct or 1 µg/well empty vector, and
stimulated as described. C, HEK-293 cells were transiently
transfected with plasmid DNA encoding G
q-Q209L (1 µg/well), G
i2-Q204L (1 µg/well), empty vector (1 µg/well), or both G
1 (0.5 µg/well) and
G
2 (0.5 µg/well) as described. Erk1/2 phosphorylation
in these cells was determined after 24 h of serum starvation. Data
are expressed as -fold Erk1/2 phosphorylation, in which the Erk1/2
phosphorylation produced in unstimulated, empty vector-transfected
cells was defined as 1.0. In the absence of transfected
1B-AR and
2A-AR, phenylephrine and UK-14304 resulted in 2.3- and
0.9-fold stimulation of Erk1/2 phosphorylation, respectively. Values
shown represent means ± S.E. from three separate experiments each
performed in duplicate. *, greater than not stimulated (NS) (p < 0.05, two-tailed T test).
Both
In
COS-7 cells, 2A-AR stimulation results in Erk1/2 activation by a
p21ras-dependent mechanism, whereas
1B-AR-mediated Erk1/2 activation is insensitive to overexpression of
a p21ras dominant-negative mutant and is inhibited by
down-regulation of PKC (34). To determine the role of p21ras in
adrenergic receptor-mediated Erk1/2 phosphorylation in HEK-293 cells,
cDNA coding for either the
1B-AR or
2A-AR was coexpressed with cDNA coding for dominant-negative mutant forms of
p21ras (N17-Ras), mSOS1 (SOS-Pro), or p74raf-1
(
N-Raf). As shown in Fig. 3, phosphorylation of
Erk1/2 in response to stimulation of both the
1B-AR and the
2A-AR
was attenuated in cells coexpressing N17-Ras, SOS-Pro, or
N-Raf.
Acute stimulation with phorbol esters was attenuated by
overexpression of
N-Raf, but not by overexpression of N17-Ras or
SOS-Pro, indicating that PKC-mediated Erk1/2 activation is
Ras-independent. As expected, EGF-stimulated Erk1/2
phosphorylation was sensitive to the effects of overexpressed N17-Ras,
SOS-Pro, and
N-Raf. Phosphorylation of Erk1/2 as a result of
G
q-Q209L expression was similarly attenuated in cells
coexpressing N17-Ras (data not shown). These data suggest that in
HEK-293 cells, stimulation of both Gq/11- and
Gi-coupled receptors leads to Erk1/2 phosphorylation in a
manner that is dependent upon mSOS, p21ras, and
p74raf-1 activation.
Erk1/2 Phosphorylation Mediated by
Pertussis toxin-sensitive G
subunit-mediated activation of p21ras in COS-7 cells is
sensitive to inhibitors of tyrosine kinases and requires recruitment of
the Ras guanine-nucleotide exchange factor, mSOS (19). The G
subunit effectors responsible for these signals are unknown. Since
PLC
isoforms are regulated by both G
q/11 and G
subunits and PLC
overexpression results in Erk1/2 activation in
COS-7 cells (34), we tested whether PLC activation was required for
1B-AR- and
2A-AR-mediated Erk1/2 phosphorylation in HEK-293
cells. As shown in Fig. 4, pretreatment of HEK-293 cells
with the PLC inhibitor, U73122, markedly attenuated both
1B-AR- and
2A-AR-mediated Erk1/2 phosphorylation. Phorbol ester-mediated Erk1/2
phosphorylation was insensitive to the effects of U73122. The results
suggest that one or more isoforms of PLC are required for both
Gq/11- and Gi-coupled receptor-mediated MAP
kinase activation in HEK-293 cells.
Recent reports have suggested that Ras-dependent Erk1/2
activation in vascular smooth muscle cells (35) and in neuronal cells
(36) may be calcium-dependent. As shown in Fig.
5, treatment of HEK-293 cells with the calcium ionophore
A23187 resulted in 5-10-fold increases in Erk1/2 phosphorylation,
similar to that observed after 5-min stimulation of cells expressing
transfected 1B- and
2A-AR. Pretreatment of HEK-293 cells with the
cell membrane-permeable Ca2+ chelating agent BAPTA
abrogated
1B-AR- and
2A-AR-mediated, as well as A23187-induced,
Erk1/2 phosphorylation. Erk1/2 phosphorylation after stimulation of
BAPTA-pretreated cells with EGF was unaffected. These data suggest that
increased intracellular Ca2+ concentration, resulting from
G
- or G
-mediated PLC activation, is required for Erk1/2
activation in HEK-293 cells.
Tyrosine
phosphorylation of the Shc adaptor protein, which supports the SH2
domain-mediated recruitment of Grb2-Sos to the plasma membrane, has
been implicated in both receptor-tyrosine kinase- and GPCR-mediated
Erk1/2 activation in some cell types (37). To test whether the
calcium-dependent 1B- and
2A-AR signals in HEK-293
cells are also dependent upon tyrosine protein phosphorylation, we
determined the effects of two tyrosine kinase inhibitors, genistein and
herbimycin A, on
1B- and
2A-mediated Erk1/2 phosphorylation in
HEK-293 cells. As shown in Fig. 6A, pretreatment of HEK-293 cells with the tyrosine kinase inhibitors markedly attenuated
1B-AR-,
2A-AR-, and EGF-R-mediated Erk1/2 phosphorylation. Erk1/2 phosphorylation induced by the calcium ionophore A23187 was also tyrosine kinase inhibitor-sensitive, suggesting that elevation of intracellular Ca2+ levels is
sufficient to induce tyrosine phosphorylation in these cells.
Involvement of the c-Src tyrosine kinase in
1B-AR- and
2A-AR-mediated Erk1/2 phosphorylation in HEK-293
cells. A, HEK-293 cells were transiently transfected with
plasmid DNA encoding
1B-AR,
2A-AR, or empty vector. Prior to
stimulation, serum-starved cells were treated with Me2SO
(0.1%; control) or herbimycin A (1 µM) for
24 h or with Me2SO (0.1%; control) or
genistein (50 µM) for 15 min. Cells were stimulated with
the appropriate agonist for 5 min, and Erk1/2 phosphorylation was
determined. B, HEK-293 cells were transiently transfected
with plasmid DNA encoding either wild-type c-Src (0.5 µg/well) or
Src-Y530F (0.5 µg/well) as described. Erk1/2 phosphorylation in these
cells was determined after 24 h of serum starvation. C,
HEK-293 cells were transiently cotransfected with plasmid DNA encoding
1B-AR,
2A-AR, or wild-type c-Src (0.5 µg/well) plus plasmid DNA
encoding either Src-K298M (1.0 µg/well) or Csk (1.0 µg/well). Cells
were serum-starved for 24 h and stimulated as indicated for 5 min,
and Erk1/2 phosphorylation was determined. Data are expressed as -fold
Erk1/2 phosphorylation, in which the Erk1/2 phosphorylation produced in unstimulated, empty vector-transfected cells was
defined as 1.0. Values shown represent means ± S.E. from three
separate experiments each performed in duplicate. NS, not
stimulated.
Both pertussis toxin-sensitive (38) and -insensitive (39) activation of
Src family nonreceptor tyrosine kinases have been described by several
laboratories, and Src activity appears to be required for G
subunit-mediated Erk1/2 activation in COS-7 cells (40). To determine
whether Src family tyrosine kinases are involved in GPCR-mediated
Erk1/2 activation in HEK-293 cells, we measured
1B-AR- and
2A-AR-mediated Erk1/2 phosphorylation in cells coexpressing either a
catalytically inactive mutant of p60c-src,
Src-K298M, or a negative regulatory protein of
p60c-src, p50csk, which phosphorylates and
inactivates p60c-src (41). As shown in Fig.
6B, overexpression of either a constitutively activated
mutant form of p60c-src (Src-Y530F) or of wild-type
p60c-src in HEK-293 cells was sufficient to induce
Erk1/2 phosphorylation. As shown in Fig. 6C, overexpression
of either Src-K298M or p50csk significantly attenuated Erk1/2
phosphorylation induced by either
1B-AR and
2A-AR stimulation or
treatment with calcium ionophore. Erk1/2 phosphorylation induced by
acute stimulation with phorbol ester or by overexpression of wild-type
c-Src was unaffected. The inability of overexpressed p50csk to
significantly inhibit Erk1/2 phosphorylation mediated by overexpressed wild-type c-Src probably reflects ineffective competition between the
two overexpressed proteins. These data suggest that both
1B- and
2A-AR signals in HEK-293 cells are calcium-dependent and mediated by Src family tyrosine kinase activity.
Acute stimulation of PKC with phorbol ester is sufficient to
induce Erk1/2 phosphorylation in HEK-293 cells. Unlike the 1B-AR- and
2A-AR-mediated signals, the acute PMA signal is insensitive to
the effects of N17-Ras, SOS-Pro, overexpressed p50csk,
Src-K298M, and tyrosine kinase inhibitors. As shown in Fig. 7, the PKC inhibitor GFX, which abolished acute
PMA-stimulated Erk1/2 phosphorylation, had no effect on Erk1/2
phosphorylation after stimulation of cells with adrenergic agonists,
EGF, or calcium ionophore. Similar results were obtained through
down-regulation of endogenous PKC expression after chronic treatment of
cells with phorbol ester. These data suggest that PKC activation
mediates Erk1/2 phosphorylation via a pathway that is distinct from the calcium- and tyrosine kinase-dependent pathway employed by
1B- and
2A-ARs.
Calcium-dependent activation of a novel focal adhesion
kinase family protein-tyrosine kinase, Pyk2, has been shown to mediate calcium ionophore-, phorbol ester-, and Gq/11-coupled
receptor-stimulated Erk1/2 activation in neuronal cells (36) via a
direct interaction with c-Src (42). Although Pyk2 is expressed at high
levels only in cells of neuronal origin (36), it is possible that this
or a related kinase might link calcium flux to tyrosine kinase
signaling pathways in other cell types. However, as shown in Fig.
8A, protein immunoblots of HEK-293 cell
lysates using anti-Pyk2 antisera detect only low levels of Pyk2
expression. To determine whether Pyk2 is involved in the
calcium-dependent activation of Erk1/2 in these cells,
1B-AR- and
2A-AR-mediated Erk1/2 phosphorylation was assayed in
cells expressing a dominant negative mutant of Pyk2 (PKM; Ref. 42). As
shown in Fig. 8B, Erk1/2 phosphorylation in response to
adrenergic receptor stimulation or treatment with calcium ionophore was
significantly attenuated, with no effect on EGF- or phorbol
ester-induced signals.
The mechanism whereby calcium influx regulates the activity of the
focal adhesion kinase family member Pyk2 is unknown. Neither Ca2+ nor PKC directly activate Pyk2 in vitro
(36). Recently, calmodulin inhibitors have been shown to inhibit
p21ras-dependent Erk1/2 activation in cultured rat
vascular smooth muscle cells (35). To determine whether calmodulin
might play a role in AR-mediated Erk1/2 activation in HEK-293 cells, we
determined the effect of three different calmodulin inhibitors on
1B- and
2A-AR-stimulated Erk1/2 phosphorylation. As shown in Fig.
9, pretreatment of HEK-293 cells with fluphenazine,
calmidazolium, or ophiobolin resulted in marked attenuation of the
phospho-MAP kinase signal, compared with Me2SO-pretreated
controls. Erk1/2 phosphorylation resulting from stimulation of
endogenous EGF receptors was unaffected. These data suggest that
calcium-calmodulin may directly or indirectly contribute to the
regulation of Pyk2 kinases and the Src-dependent activation
of Erk1/2 by GPCRs.
These data suggest a model for 1B-AR- and
2A-AR-mediated
Erk1/2 activation that is mediated by calcium-dependent
regulation of protein-tyrosine kinases. In HEK-293 cells, as in COS-7
cells (34), the
1B-AR-mediated signal is dependent upon the
subunit of a pertussis toxin-insensitive G protein, while the
2A-AR-mediated signal is sensitive to PTX treatment and is dependent
upon the release of free G
subunit complexes. Fig.
10 depicts a model of GPCR-mediated Erk1/2 activation
in HEK-293 cells that is consistent with our data.
G
q/11- and G
-dependent activation of
PLC increases cytoplasmic levels of inositol 1,4,5-trisphosphate,
resulting in an increase in cytoplasmic calcium concentration. High
intracellular concentrations of calcium, perhaps through calmodulin,
lead to activation of Pyk2 or a closely related tyrosine kinase, which regulates the activity of p60c-src.
Src-dependent tyrosine phosphorylation of adaptor proteins, such as Shc, results in recruitment of the Grb2-SOS complex to the
plasma membrane, where it catalyzes p21ras guanine nucleotide
exchange. Ras-dependent recruitment of p74raf-1
kinase to the membrane initiates the phosphorylation cascade leading to
activation of Erk1/2. In this system the G
subunit- and
G
q/11 subunit-mediated pathways each require the
PLC
-dependent stimulation of calcium influx. This early
convergence is distinct from findings in COS-7 and CHO cells (34) and
more closely resembles the calcium- and Ras-dependent
activation of Erk1/2, which has been reported in primary cultures of
vascular smooth muscle cells and ventricular myocytes (35, 43, 44). The
observation that dominant interfering mutants of p21ras,
p74raf-1, and SOS do not fully attenuate
1B-AR- and
2A-AR-mediated Erk1/2 phosphorylation may reflect incomplete
inhibition of receptor-mediated p74raf-1 activation.
Alternatively, these data may be indicative of another, p21ras-independent, mechanism of Erk1/2 phosphorylation, as has
been described for Gq- and Go-coupled MAP
kinase activation in Chinese hamster ovary cells (34, 45).
Activation of p60c-src is required for
Gi-coupled receptor-mediated, G
subunit-dependent activation of Erk1/2 in COS-7 cells (37),
and Gi- and Gq/11-coupled receptor-stimulated
Erk1/2 activation in PC12 cells (42). Src family kinase recruitment
into Shc-containing protein complexes has been demonstrated following
stimulation of formyl-methionyl peptide receptors in human neutrophils
(46) and following stimulation of LPA and
2-adrenergic receptors in COS-7 cells (37). Our data indicate that Src kinases function as key
intermediates in calcium-dependent regulation of Erk1/2, mediated by both G
and G
q/11 subunits in some cell
types. Collectively, these findings indicate that regulation of Src
family protein-tyrosine kinase activity, potentially via multiple
mechanisms, is a common requirement for GPCR-mediated Erk1/2
activation.
In neuronal cells, association of p60c-src with the
calcium-regulated focal adhesion kinase family member Pyk2 mediates
both Shc phosphorylation and Erk1/2 activation (42). Pyk2 was
previously thought to be active only in neuronal cells. The detection
of Pyk2 in HEK-293 cell lysates as well as the sensitivity of 1B-AR- and
2A-AR-mediated Erk1/2 activation in HEK-293 cells to both the
dominant negative mutant of Pyk2 and specific inhibitors of p60c-src suggest that a Pyk2-mediated
Src-dependent mechanism of p21ras activation may
represent a paradigm for mitogenic signaling in a variety of
non-neuronal cell types. The mechanism of Ca2+-mediated
Pyk2 activation remains unclear, however, since calcium does not
directly modulate Pyk2 activity (36). Perhaps significantly, both
adrenergic receptor- and calcium ionophore-mediated Erk1/2 phosphorylation in HEK-293 cells is sensitive to chemical inhibitors of
calmodulin. Eguchi et al. (35) have suggested that
calmodulin regulates Erk1/2 activation in cultured rat vascular smooth
muscle cells. In NG108 cells, depolarization induces
calcium-dependent Erk1/2 activation, which is mediated by
calmodulin-dependent kinase IV (47). Our data suggest that
the calcium-mediated regulation of Src family tyrosine kinases proceeds
through a calmodulin-dependent mechanism. These data also
suggest that, if Pyk2 directly activates p60c-src
in HEK-293 cells, then perhaps calcium/calmodulin is involved in
activation of Pyk2, either directly or through a calcium/calmodulin effector protein.
The elucidation of GPCR-mediated mitogenic signaling pathways has
revealed significant degrees of heterogeneity between cell types. In
Rat-1 fibroblasts, Gi-coupled, but not
Gq/11-coupled, receptors mediate tyrosine
kinase-dependent Erk1/2 activation via a calcium- and
PLC-independent mechanism (23). In Chinese hamster ovary cells,
Gq/11-coupled receptor stimulation leads to
Gq/11-mediated activation of PKC, p74raf-1, and
Erk1/2 in a tyrosine kinase- and p21ras-independent manner
(34). In PC12 neuroblastoma cells, both Gi- and
Gq/11-coupled receptors have been shown to activate Erk1/2 via calcium-dependent regulation of p112pyk2,
p60c-src, and p21ras (42). Our data suggest
that calcium-dependent regulation of Ras by both
Gi- and Gq/11-coupled receptors may represent a
common mechanism of GPCR-mediated Erk1/2 activation in many
non-neuronal cell types. Indeed, Gq/11-coupled receptors
mediate calmodulin inhibitor-sensitive, Ras-dependent
Erk1/2 activation in cultured vascular smooth muscle cells (35), a
calcium-sensitive tyrosine kinase has been cloned from calf uterus
(48), and Gq/11-coupled receptor-mediated hypertrophy of
cultured rat ventricular myocytes is reportedly
Ras-dependent (43). Characterization of receptor- and
kinase-specific differences in the mechanisms of GPCR-mediated mitogenic signal transduction may permit the development of strategies for selective antagonism of distinct G protein-coupled
receptor-mediated mitogenic signaling pathways, which ultimately may
permit selective modulation of cell proliferation in a variety of
pathophysiologic states.
We thank Donna Addison and Mary Holben for expert secretarial assistance, Sameena Rahman for technical assistance, and Dr. John Raymond for insightful discussion.