In vascular smooth muscle (VSM) and many other cells, G protein receptor-coupled activation of mitogen-activated protein kinases has been linked, in part, to increases in free intracellular Ca2+. Previously, we demonstrated that ionomycin-, angiotensin II-, and thrombin-induced activation of extracellular signal-regulated kinase (ERK)1/2 in VSM cells was attenuated by pretreatment with KN-93, a selective inhibitor of the multifunctional Ca2+/calmodulin-dependent protein kinase (CaM kinase II). In the present study, we show that the Ca2+-dependent pathway leading to activation of ERK1/2 is preceded by nonreceptor proline-rich tyrosine kinase (PYK2) activation and epidermal growth factor (EGF) receptor tyrosine phosphorylation and is attenuated by inhibitors of src family kinases or the EGF receptor tyrosine kinase. Furthermore, we demonstrate that pretreatment with KN-93 or a CaM kinase II inhibitor peptide inhibits Ca2+-dependent PYK2 activation and EGF receptor tyrosine phosphorylation in response to ionomycin, ATP, and platelet-derived growth factor but has no effect on phorbol 12,13-dibutyrate- or EGF-induced responses. The results implicate CaM kinase II as an intermediate in the Ca2+/calmodulin-dependent activation of PYK2.
- calcium/calmodulin-dependent protein kinase
- proline-rich tyrosine kinase
- extracellular signal-regulated kinase 1/2
- epidermal growth factor receptor transactivation
- vascular smooth muscle cells
- autoinhibitory peptide
mitogen-activated protein (MAP) kinases such as extracellular signal-regulated kinase (ERK)1/2 are involved in the regulation of essential cellular processes, including gene expression and cell proliferation (5,23). Signaling pathways involving ERK1/2 are triggered by diverse cellular stimuli, including receptor tyrosine kinase-coupled growth factors (25), G protein-coupled receptor (GPCR) agonists (8), mechanical stimuli (19), and integrin-dependent cell/matrix interactions (7). Although the signaling cascade initiated by receptor tyrosine kinase-coupled growth factors such as epidermal growth factor (EGF) or platelet-derived growth factor (PDGF) is understood in considerable molecular detail, the signaling pathways by which GPCR ligands activate ERK1/2 are not fully understood.
A number of proximal signaling pathways, including G protein βγ-subunits (26), protein kinase C, and/or Ca2+ (4), may converge on the ERK signaling cascade. The nonreceptor tyrosine kinases, proline-rich tyrosine kinase (PYK2) and src, have been proposed to be points of convergence for Ca2+- and protein kinase C-dependent pathways leading to ERK1/2 activation (6). PYK2 andsrc may act via a pathway involving tyrosine phosphorylation and activation of the EGF receptor, scaffolding of adaptor proteins and guanine nucleotide exchange factors (SHC/GRB2/sos), and consequent ras-dependent activation of the well-described protein kinase cascade, culminating in ERK1/2 activation (10,12). Alternatively, PYK2 and src may lead to activation of a protease that produces heparin/EGF from extracellular matrix that then acts as a ligand for the EGF receptor (20). Also, src has been proposed to directly activate the SHC/GRB2/sos complex (12). In all of these models, the proximal events that couple increases in free intracellular Ca2+ or protein kinase C to PYK2/src activation are not known.
In vascular smooth muscle (VSM), ERK1/2 activation in response to angiotensin II (ANG II), thrombin, or ATP appears to involve both Ca2+ and protein kinase C-dependent mechanisms (2, 4,6). We previously reported that activation of ERK1/2 in cultured rat aortic VSM in response to receptor and nonreceptor Ca2+-mobilizing stimuli (ANG II, thrombin, and ionomycin) is temporally preceded by activation of the multifunctional serine/threonine kinase Ca2+/calmodulin-dependent protein kinase II (CaM kinase II) (2). Inhibition of CaM kinase II by pretreatment with a calmodulin antagonist (calmidazolium) or a CaM kinase II inhibitor (KN-93) attenuated ERK1/2 activation in response to these stimuli. This pharmacological evidence, along with limited molecular studies, suggests that CaM kinase II, like protein kinase C, is an intermediate in GPCR-mediated activation of ERK1/2 in VSM cells. However, these studies provided no insight into the mechanisms by which CaM kinase II could activate the ERK signaling pathway.
In the present study, we used the Ca2+ ionophore ionomycin to selectively stimulate Ca2+-dependent signaling pathways in VSM cells independent of other GPCR pathways involving either activated G protein βγ-subunits or protein kinase C and determined the effects of inhibitors of tyrosine kinases and CaM kinase II on the activation of PYK2, the EGF receptor, and ERK1/2. The results indicate that Ca2+-dependent activation of ERK1/2 in VSM cells is completely dependent on PYK2/src activation and transactivation of EGF receptor. Pretreatment with the CaM kinase II inhibitors, KN-93 or autoinhibitory peptide (AIP), attenuates PYK2 and EGF receptor activation in response to Ca2+-mobilizing stimuli in VSM, supporting a role for CaM kinase II in the activation of the nonreceptor tyrosine kinases PYK2 and src.
VSM cells were dispersed from the medial layer of thoracic aortas of Sprague-Dawley rats weighing 200–300 g as described earlier (14), cultured in DMEM/F-12 medium containing 10% fetal bovine serum (Hyclone) at 37°C and 5% CO2, and split twice a week. Cells from passages 3–10 were used in the experiments. Confluent cultures were growth arrested for 16–24 h in DMEM/F-12 containing 0.4% serum. One hour before the experiments, the 0.4% serum-containing medium was removed from the cells and replaced with Hanks' balanced salt solution (HBSS) supplemented with Ca2+/Mg2+ and 10 mM HEPES, pH 7.4.
Cell lysate preparation.
VSM cells were pretreated with inhibitors for 30 min before being exposed to stimulants for the indicated times. Reactions were stopped by transferring the dishes to ice, removing the HBSS, and adding ice-cold lysis buffer [in mM: 25 MOPS, pH 8.6, 1.5 EGTA, 125 NaCl, 50 NaF, 50 sodium pyrophosphate, 1 dithiothreitol, and 0.1 phenylmethylsulfonyl fluoride, with 0.5% NP-40 and 0.2 U/ml aprotinin (1 ml/100 mm dish, 0.3 ml/60 mm dish)]. The cells were then scraped and collected into 1.5-ml microcentrifuge tubes and stored at −20°C.
Permeabilization of smooth muscle cells.
AIP, a CaM kinase II-specific inhibitor peptide (16), was introduced into the VSM cells by reversible permeabilization, as described previously (17). Briefly, cells were growth arrested for 24 h in serum-free media followed by a series of 2-min incubations in PBS to reduce their temperature to 4°C. The cells were then incubated in a permeabilization buffer (20 mmol/l HEPES, pH 7.4, 10 mmol/l EGTA, 140 mmol/l KCl, 50 μg/ml saponin, 5 mmol/l NaN3, and 5 mmol/l oxalic acid dipotassium salt) containing 10 μM CaM kinase II inhibitor peptide (myristoylated-AIP, autocamtide-2-related inhibitory peptide; Biomol Research Labs, Plymouth Meeting, PA) for 10 min on ice. The cells were then washed with PBS (4°C) three times and allowed to recover in PBS for an additional 20 min at 4°C. After incubation at room temperature in PBS for 2 min, the cells were placed in HBSS supplemented with Ca2+/Mg2+ and 10 mmol/l HEPES, pH 7.4, and placed in the incubator (37°C) for 15 min before treatment with agonists.
Immunoprecipitations and Western blotting.
Cell lysates were thawed and cleared by centrifugation at 10,000g. The supernatants were transferred to fresh microcentrifuge tubes and incubated with primary antibody for 90 min with rocking at 4°C. Washed protein A beads (25 μg; Pierce) were then added and incubated at 4°C with continued rocking overnight. The beads were recovered by centrifugation and washed three times with lysis buffer. Immunoprecipitated proteins were solubilized in 3× gel loading buffer and heated for 5 min at 95°C. The sample was cleared of beads by brief centrifugation, and the supernatant was loaded onto SDS-PAGE gels, transferred to either Nitro plus or PVDF-plus membranes (MSI). The membranes were blocked in Tris-buffered saline with 0.1% Tween 20 (TBST) containing either 5% nonfat dry milk or 3% BSA. After being blocked, the membranes were incubated with the primary antibody for 1 h at room temperature or overnight at 4°C, washed three times with TBST, and incubated with horseradish peroxidase-conjugated secondary antibody (1:1,500 dilution; Amersham) for 1 h at room temperature, followed by another three TBST washes. The membranes were developed using enhanced chemiluminescence substrate (Amersham) followed by exposure to ECL Hyper film (Amersham).
Antibody to the phosphorylated sequence of CaM kinase II was obtained through the services of HTI Bio-Products (Ramona, CA). Antigen peptide was 283HRQET(P)VDCLKKF294, corresponding to the Thr287 site in the δ2-CaM kinase II subunit. The antibody was specific and reacted minimally with either unphosphorylated CaM kinase II or other phosphorylated sequences from the kinase. Production and specificity of the antibody for detection of δ2-subunits of CaM kinase II has been documented in previous publications (18, 27). The polyclonal antibody specific for PYK2 and the monoclonal antibody specific for tyrosine-phosphorylated protein (4G10) were purchased from Upstate Biotechnology (Lake Placid, NY). Polyclonal antibodies specific for the phosphorylated PYK2 were purchased from Biosource International (Camarillo, CA). The polyclonal antibody specific for the EGF receptor was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The polyclonal antibodies for the phosphorylated and unphosphorylated ERK1/2 were purchased from New England Biolabs (Beverly, MA).
All tissue culture media were purchased from GIBCO BRL (Rockville, MD) unless otherwise specified. Tissue culture dishes and disposable plastic pipettes were purchased from Fisher Scientific (Pittsburgh, PA). KN-93 and KN-92 were purchased from Seikagaku America (Falmouth, MA). Ionomycin, phorbol 12,13-dibutyrate, 4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d]pyrimidine (PP2), AG-1478, and AG-1296 were obtained from Calbiochem (La Jolla, CA). All other chemicals were obtained from Sigma (St. Louis, MO).
Ca2+-dependent activation of ERK1/2 is attenuated by inhibitors of CaM kinase II.
Once activated by Ca2+/calmodulin, CaM kinase II subunits rapidly autophosphorylate on a conserved threonine residue (Thr287 in the δ-subunit) (1). Phosphorylation of Thr287 results in partial activation of the kinase, even in the absence of Ca2+/calmodulin. Thus either Ca2+/calmodulin-independent (autonomous) activity or the degree of Thr287 phosphorylation can be used as an index of CaM kinase II activation in situ. Previous work in this lab established the in situ Ca2+ dependency for CaM kinase II activation (by assaying generation of autonomous activity) in VSM cells by using ionomycin, a Ca2+ ionophore, as a stimulus (1). A selective inhibitor of CaM kinase II, KN-93, attenuated this activation in situ with an IC50 of 14 μM and maximal inhibition at 30 μM (2). Figure1 A shows Western blots of VSM cell lysates using an antibody that specifically recognizes the peptide sequence in CaM kinase II containing phosphorylated Thr287(27). Both KN-93 and a peptide (AIP) modeled on the CaM kinase II autoinhibitory domain, the latter introduced into the cells by reversible permeabilization, were found to inhibit ionomycin-stimulated Thr287 autophosphorylation, confirming their effectiveness as inhibitors of CaM kinase II activity in the VSM cells. Ionomycin-induced activation of ERK1/2 was blocked by both KN-93 and AIP (Fig. 1 B), consistent with previous findings (2) that implicated CaM kinase II as an intermediate in the Ca2+-dependent ERK1/2 signaling pathway. In contrast, EGF-induced ERK1/2 activation was unaffected by pretreatment with KN-93 or AIP.
Tyrosine kinases are intermediates in the Ca2+-dependent activation of ERK1/2.
Activation of ERK1/2 in response to GPCR agonists depends on activation of the nonreceptor tyrosine kinases, PYK2 and src, which results in the tyrosine phosphorylation and activation of EGF receptors (11, 12). To determine if the Ca2+-dependent activation of ERK1/2 specifically required activation of srcand the EGF receptor tyrosine kinase, VSM cells were pretreated with PP2, a src family-selective tyrosine kinase inhibitor, or AG-1478, a selective EGF receptor tyrosine kinase inhibitor. Both PP2 (Fig. 2, B and C) and AG-1478 (Fig. 2) inhibited ERK1/2 activation in response to addition of ionomycin. Transactivation of PDGF receptor following stimulation with GPCR agonists has also been linked to ERK1/2 activation (15). However, pretreatment of VSM cells with the PDGF receptor tyrosine kinase inhibitor AG-1296, which inhibited control PDGF-induced activation of ERK1/2, had no effect on ionomycin-induced activation of ERK1/2 (Fig. 2 A). To rule out nonspecific effects of these tyrosine kinase inhibitors on CaM kinase II activation, the effects of AG-1478 and PP2 on CaM kinase II autophosphorylation were determined. Neither drug inhibited the peak activation-dependent autophosphorylation of the kinase after 30-s stimulation with ionomycin (Fig. 3).
Ca2+-dependent activation of the EGF receptor is inhibited by KN-93, Ca2+ chelators, and inhibitors of src family kinases.
Inhibition of ionomycin-stimulated ERK1/2 activation by AG-1478 implicates the EGF receptor tyrosine kinase as an intermediate in the Ca2+-dependent signaling pathway. As shown in Fig.4, this was confirmed by directly documenting tyrosine phosphorylation of the EGF receptor in response to addition of ionomycin. Tyrosine phosphorylation of the EGF receptor was maximal 3 min after ionomycin treatment (Fig. 4 A), temporally lagging peak CaM kinase II activation at 30 s (Fig. 3) but preceding peak ERK1/2 activation at 5 min (Fig. 2 B). KN-93 blocked ionomycin-stimulated EGF receptor tyrosine phosphorylation but had no effect on phosphorylation stimulated by EGF binding. Activation of CaM kinase II is dependent on increases in intracellular Ca2+. Chelation of free intracellular Ca2+ with 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA), introduced into the cells by addition of the cell-permeable BAPTA-AM, prevented the ionomycin-dependent activation of CaM kinase II (data not shown) and EGF receptor (Fig. 4 B), confirming the Ca2+ dependence of the response.
As reported previously (3), src andsrc family kinases mediate GPCR-dependent activation of PYK2 and transactivation of the EGF receptor. To determine src's role in the Ca2+-dependent activation of these tyrosine kinases, VSM cells were treated with the selective inhibitor ofsrc family kinases, PP2. At 10 μM, PP2 blocked ionomycin-dependent activation of both PYK2 and the EGF receptor, strongly suggesting that src or an src-like kinase mediates the CaM kinase II-dependent activation of PYK2 and the subsequent transactivation of the EGF receptor (Fig.5). It should be noted, however, that PP2 also partially inhibits the EGF-stimulated EGF receptor activation (data not shown). This may indicate that PP2 has a nonspecific effect on the EGF receptor tyrosine kinase or that src or ansrc-like kinase may have some role in the ligand-dependent activation of the receptor tyrosine kinase.
Activation of PYK2 is dependent on CaM kinase II.
Previous studies have implicated the nonreceptor tyrosine kinase PYK2 in the Ca2+-dependent pathway leading to EGF receptor transactivation and ERK1/2 activation. Addition of ionomycin to VSM cells resulted in the rapid (within 1–3 min) tyrosine phosphorylation of PYK2 (Fig.6 A) that has been previously shown to correlate with PYK2 activation (22). Pretreatment of the cells with the CaM kinase II inhibitors KN-93 or AIP (Fig. 6) significantly attenuated ionomycin-induced PYK2 tyrosine phosphorylation. Chelation of intracellular free Ca2+ with BAPTA (Fig. 7 A) also inhibited the ionomycin-dependent activation of PYK2. The physiological stimuli ATP and PDGF have also been shown to cause an increase in intracellular Ca2+ as well as to activate CaM kinase II in VSM cells (27, 28). Treatment with these agonists resulted in the rapid activation of PYK2 in the VSM cells, responses that were partly inhibited by pretreatment with KN-93 (Fig. 7 B).
Phorbol ester activators of protein kinase C have also been shown to activate PYK2 (22). Pretreatment with KN-93 (Fig.8 A) or AIP (data not shown) had no effect on phorbol 12,13-dibutyrate-induced activation of PYK2 (Fig. 8 B) or ERK1/2 (2). These experiments document the relative specificities of the CaM kinase II inhibitors that are restricted to a Ca2+-dependent pathway proximal to the activation of PYK2. Overall, the results implicate CaM kinase II as an intermediate in the Ca2+/calmodulin-dependent activation of PYK2 and provide a pathway for Ca2+-dependent activation of ERK1/2, independent of protein kinase C or by direct coupling through GPCRs (Fig. 9).
Agonists for growth factor receptors and GPCR (8), membrane depolarization (28), and mechanosensitive mechanisms (19) are among the diverse stimuli that have been reported to result in an activation of MAP kinases, including ERK1/2. In the case of stimuli acting through GPCR, there appear to be at least three general pathways leading to ERK1/2 activation; a G protein βγ-subunit-dependent pathway, a protein kinase C-dependent pathway, and a Ca2+-dependent pathway. Elucidation of the precise mechanisms involved for any one of these or the relative importance of each for a given agonist is complicated by the fact that the pathways may be redundant and activated in both an agonist- and cell type-specific manner. For example, in VSM cells and cardiac fibroblasts, ANG II-induced ERK1/2 activation occurs through redundant protein kinase C- and Ca2+-dependent pathways. Inhibition of both pathways is required to block ANG II-stimulated ERK1/2 activation (4). To specifically study the Ca2+-dependent pathways leading to ERK1/2 activation in VSM cells, we have relied on the Ca2+ ionophore, ionomycin, to selectively activate these pathways independent of heterotrimeric G protein activation or protein kinase C. Results from previous studies have implicated the multifunctional serine/threonine kinase CaM kinase II as an intermediate in the Ca2+-dependent pathway leading to ERK1/2 (2). In the present study, we have narrowed down the likely site of CaM kinase II involvement to a point proximal to activation of the nonreceptor tyrosine kinases PYK2 and src.
There is a substantial body of evidence implicating growth factor receptors, such as the EGF receptor, as intermediates in the activation of ERK1/2 following GPCR activation in a number of cell types, including VSM (9, 10, 12, 13, 15, 22). This circuitous pathway requires activation of PYK2 and src or srcfamily kinases to initiate EGF receptor “transactivation” (10) or to stimulate a protease that releases latent EGF from the extracellular matrix (9, 13, 20, 21) with consequent EGF receptor activation leading to ras-dependent activation of the MAP kinase signaling cascade. The results shown here, indicating essentially complete inhibition of ionomycin-stimulated ERK activation with an EGF receptor tyrosine kinase inhibitor (AG-1478) or a src family tyrosine kinase inhibitor (PP2), suggest a requirement for these intermediate tyrosine kinases in the Ca2+-dependent pathways leading to ERK1/2 activation in VSM cells. However, because of potential specificity problems with the tyrosine kinase inhibitor, it is still not possible to conclude with certainty which src family tyrosine kinases are involved or whether they act proximal to, coincident with, and/or distally to PYK2 activation.
In VSM cells, transactivation of both the PDGF and EGF receptors has been implicated in ANG II-dependent activation of ERK1/2 (10,15). However, in the present study we found that treatment with AG-1296, a selective PDGF receptor tyrosine kinase inhibitor, had no effect on ionomycin-stimulated ERK1/2. Together, these results suggest that even within the same cell type there can be differences in signaling pathways, possibly due to the source of the cells, differences in primary cell culture, or relative abundance or differential cellular compartmentalization of receptor types. It is also possible that PDGF receptor transactivation, although not Ca2+ dependent, may be activated by alternative pathways stimulated by GPCR agonists, for example involving protein kinase C-dependent activation of PYK2/src.
Because the primary approach used in this study was pharmacological, we were careful to document the intended effect of the CaM kinase II inhibitors on CaM kinase II activation and to establish their relative specificity by demonstrating a lack of effect of the inhibitors on steps downstream of CaM kinase II or on pathways independent of Ca2+/calmodulin. Two chemically and mechanistically distinct inhibitors of CaM kinase II were used in the studies with similar results. KN-93 prevents activation of CaM kinase II by interfering with calmodulin-dependent activation of the kinase (24). On the other hand, AIP is a peptide modeled on the autoinhibitory domain of CaM kinase II and appears to act as a competitive inhibitor of CaM kinase II substrates (16). Treatment of VSM cells with either drug inhibited ionomycin-stimulated CaM kinase II and ERK1/2 activation with similar potencies. Although maximal concentrations of KN-93 and AIP strongly inhibited Ca2+-dependent activation of PYK2, EGF receptor tyrosine kinase, and ERK1/2, no effects were observed on the same responses stimulated by phorbol 12,13-dibutyrate or EGF. KN-92, the inactive analog of KN-93, also had no effect on ionomycin-stimulated PYK2 activation. Conversely, the two tyrosine kinase inhibitors (AG-1478 and PP2), which effectively blocked ionomycin-stimulated ERK1/2 activation, had no effect on activation of CaM kinase II. Collectively, these controls support the conclusions that are based on the specificity of these drugs.
Recently, it was reported that membrane depolarization stimulated phosphorylation and activation of PYK2, the EGF receptor, and ERK1/2 in PC12 cells, and, on the basis of experiments that used a calmodulin antagonist (W-7) and a CaM kinase II inhibitor (KN-62), it was concluded that these responses were CaM kinase II dependent (17). Interestingly, activation of these responses by either bradykinin or ionomycin was unaffected by the CaM kinase II inhibitors. It is possible to rationalize the lack of effect of the inhibitors on bradykinin-induced responses on the basis of activation of redundant signaling pathways, for example, involving protein kinase C. However, without a direct demonstration of the efficacy of the inhibitors on activation of CaM kinase II in the PC12 cells, it is difficult to reconcile their lack of effect on ionomycin-induced EGF receptor activation. Previously, we reported that the activation of CaM kinase II in VSM cells in response to ionomycin was due to a release of Ca2+ from intracellular pools (14). Activation of CaM kinase II in PC12 cells by KCl is due primarily to influx of Ca2+ due to membrane depolarization. Thus another possible explanation for the results in PC12 cells may be a relative selectivity of the inhibitors used for CaM kinase II activated via Ca2+influx as opposed to release of intracellular Ca2+.
Activated src has been shown to physically associate with both PYK2 (10) and EGF receptor (12). However, a number of molecular studies using active and dominant-negative PYK2 constructs have implicated PYK2 activation as a proximal step in the transactivation of the EGF receptor (3,28). The data presented in this study support a model of Ca2+/calmodulin-dependent ERK1/2 activation that is dependent on a CaM kinase II at a step proximal to activation of PYK2 and src and, consequently, the EGF receptor tyrosine kinase (Fig. 9). Although the exact mechanisms by which CaM kinase II leads to activation of nonreceptor tyrosine kinases such as PYK2 orsrc family kinases remains to be determined, the model potentially explains the Ca2+/calmodulin dependency for PYK2 activation and suggests that CaM kinase II and protein kinase C may act via a common substrate or set of substrates proximal to PYK2 and src family kinases. The requirement for PYK2 andsrc family kinases in the GPCR-induced transactivation of the EGF receptors has recently been established using genetic (knockout) approaches (3). Importantly, these studies also established the requirement for src family kinases in the GPCR-induced activation of PYK2. This suggests that future studies aimed at elucidating the intermediate steps that couple activation of CaM kinase II, and possibly protein kinase C, to PYK2 might be directed toward proteins known to be involved in regulating srcfamily kinases.
We thank Wendy Hobb for expert assistance in the preparation of the manuscript and Dr. Kevin Pumiglia and Paul Pfleiderer for helpful discussions of the results.
This work was supported by a Trustee Scholarship from Albany Medical Center to R. Ginnan and grants to H. Singer from the National Heart, Lung, and Blood Institute (HL-49426 and HL-40992).
Address for reprint requests and other correspondence: H. A. Singer, Center for Cardiovascular Sciences, Albany Medical College (MC-8), 47 New Scotland Ave., Albany, NY 12208 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2002 the American Physiological Society