The purpose of this study was to investigate the potential role of mitogen-activated protein (MAP) kinase in smooth muscle contraction by monitoring MAP kinase activation, caldesmon phosphorylation, and contractile force during agonist stimulation. Isometric tension in response to KCl and phenylephrine (PE) was measured from strips of ferret aorta. MAP kinase activation was monitored by Western blot using a phosphospecific p44/p42 MAP kinase antibody. Caldesmon phosphorylation was assessed using specific phosphocaldesmon antibodies. We report here that treatment of smooth muscle strips with PD-098059, a specific inhibitor of MAP kinase kinase, did not detectably modify the KCl-evoked contraction but significantly inhibited the contraction to PE in the absence of extracellular Ca2+. In this experimental condition, where the contraction occurs in the absence of increases in 20-kDa myosin light chain phosphorylation, PD-098059 also inhibited significantly MAP kinase and caldesmon phosphorylation. Collectively, these results demonstrate a direct cause-and-effect relationship between MAP kinase activation and Ca2+-independent smooth muscle contraction and support the concept of caldesmon phosphorylation as the missing link between both events.
- mitogen-activated protein kinase
- calcium-independent contraction
- ferret aorta
- thin-filament regulation
although calcium-dependent phosphorylation of the 20-kDa myosin light chain (LC20) has been shown to be a major determinant of excitation-contraction coupling in smooth muscle, there is accumulating evidence for additional regulatory mechanism(s) (for review, see Ref. 13). Indeed, during prolonged contraction of some smooth muscle types, a dissociation between intracellular Ca2+ concentration, LC20 phosphorylation levels, cross-bridge cycling rates, and force levels has been reported (4, 14,19, 22). In the smooth muscle cells of ferret aorta, more specifically, we have shown that a significant part of the contraction to phenylephrine (PE) and the entirety of the contraction to phorbol esters persist at resting intracellular Ca2+ concentration (3, 12, 16). This Ca2+-independent contraction appears to be completely abolished by protein kinase C (PKC) inhibitors (16) and significantly diminished by tyrosine kinase inhibitors (5,17). Moreover, in isolated smooth muscle cells from ferret aorta bathing in a Ca2+-free medium, we have shown a PKC-dependent translocation of cytosolic mitogen-activated protein (MAP) kinase to the membrane before contraction, followed by a second redistribution to the contractile filaments during cell contraction (18) . These results led us to propose that MAP kinases could be specifically involved in the Ca2+-independent smooth muscle activation of these cells.
Extracellular-signal-regulated kinase (ERK)-MAP kinases are serine/threonine kinases that are activated by phosphorylation on both threonine and tyrosine residues through an upstream kinase, MEK (MAPk and ERKk kinase). Besides their recognized roles in cell growth, ERK-MAP kinases can be activated in nonproliferative differentiated cells (2, 11). Stimulation of MAP kinase in response to both mechanical load and pharmacological stimulation has been reported in vascular smooth muscle by Katoch and Moreland (15) and Adam et al. (1) and, more recently, in nonvascular smooth muscles by Gerthoffer et al. (8, 9) and Nohara et al. (21). Although activated MAP kinase can phosphorylate many proteins in vitro, caldesmon, an actin-binding protein, was proposed by these different authors as the most likely substrate to play a role in smooth muscle contraction. Indeed, Adam and Hathaway (2) had previously reported that the sites of caldesmon phosphorylated in situ are identical to those phosphorylated in vitro by MAP kinase but not by PKC or other kinases, suggesting that caldesmon is an endogenous substrate for ERK-MAP kinase in differentiated smooth muscle. Moreover, caldesmon is abundantly expressed in smooth muscle, and its phosphorylation by MAP kinase has been shown to reduce its capacity to inhibit actin sliding velocity in a motility assay (9). Caldesmon therefore appeared as a possible target of MAP kinase-dependent regulation of smooth muscle contractile tone. However, although caldesmon phosphorylation is known to be increased simultaneously with ERK-MAP kinase activation in smooth muscle (8, 9), there was so far no direct demonstration in intact muscle of a link between agonist-mediated activation of MAP kinase and caldesmon phosphorylation or between MAP kinase phosphorylation and muscle contraction.
In this study, we have investigated the role of ERK-MAP kinase activation in ferret aorta smooth muscle contraction by using a specific MEK inhibitor (PD-098059). Combined with the use of a recently developed antibody directed against caldesmon phosphorylated at the MAP kinase site, this approach allowed us to demonstrate a direct cause-and-effect relationship between MAP kinase activation and Ca2+-independent smooth muscle contraction and to strengthen the hypothesis of caldesmon phosphorylation being the missing link between both events.
Tissue preparation and force measurements. Ferrets were anesthetized with chloroform in a ventilation hood in accord with procedures approved by the Institutional Animal Care and Use Committee. The aorta was quickly removed and immersed in an ice-cold and oxygenated (95% O2-5% CO2) physiological saline solution (PSS) composed of (in mM) 120 NaCl, 5.9 KCl, 25 NaHCO3, 17.5 dextrose, 2.5 CaCl2, 1.2 MgCl2, and 1.2 NaH2PO4(pH 7.4). The aorta was cleaned of all adherent connective tissue, and the endothelium was removed by gentle abrasion with a rubber policeman. Circular strips were prepared as previously described (14) and attached to a force transducer. They were allowed to equilibrate for at least 1 h and then were challenged with a depolarizing saline solution containing 51 mM KCl (PSS in which 51 mM of NaCl has been stoichiometrically replaced by KCl; KPSS). Muscle strips were frozen at the desired time points following agonist stimulation (by using a dry ice-acetone, TCA slurry) and stored at −80°C until used.
Phospho-MAP kinase and MAP kinase Western blot. Previously frozen samples were homogenized in a buffer containing 50 mM Tris, 10% glycerol, 140 mM NaCl, 1% Triton X-100, and the following protease inhibitors: 5.5 μM leupeptin, 5.5 μM pepstatin, 20 KIU/ml aprotinin, 2 mM Na3VO4, 1 mM NaF, 100 μM ZnCl2, 20 mM β-glycerophosphate, and 20 μM phenylmethylsulfonyl fluoride (PMSF). Sample homogenates were then centrifuged for 2 min, and the supernatants were collected. Aliquots were subjected to electrophoresis on 10% SDS-polyacrylamide gel and transferred onto nitrocellulose membrane. Membranes were blocked in a PBS solution con taining 5% dry milk for 2 h before an overnight incubation in a Trisbuffered saline (TBS) solution containing 1% milk and either a phosphospecific p44/p42 MAP kinase (Tyr-202/Tyr-204) antibody (at 1:1,000 dilution; New England Biolabs) or a panERK antibody (at 1:500 dilution; Signal Transduction). Membranes were washed using TBS containing 1% milk and then incubated with horseradish peroxidase-conjugated secondary antibody (dilution 1:2,000; Calbiochem) for 1 h. Immunoreactive bands were visualized by enhanced chemiluminescence (Pierce). Developed films from enhanced chemiluminescence were scanned and analyzed using NIH Image; care was taken to avoid saturation of exposures for densitometry.
Caldesmon and phosphocaldesmon Western blot. Previously frozen tissue samples were homogenized in a buffer containing (in mM) 50 imidazole (pH 7.0), 300 KCl; 10 NaF, 1 EGTA, 0.5 MgCl2, 10 β-glycerophosphate, 1 Na3VO4, 0.1 PMSF, 1 dithiothreitol, and 1 benzamidine and were boiled for 10 min. Proteins denatured by heating were eliminated by centrifugation, and TCA precipitation (10% wt/vol) was used to concentrate caldesmon, which is heat resistant. Samples were subjected to electrophoresis on 7.5% SDS-polyacrylamide gel and transferred onto nitrocellulose. Membranes were blocked in a TBS solution containing 5% BSA for 1 h before an overnight incubation in blocking solution containing antisera (see below). Immunoreactive bands were revealed with an incubation in blocking solution containing125I-labeled protein A (DuPont NEN) and visualized by autoradiography.
Rabbit polyclonal antisera were generated against full-length caldesmon or a phosphopeptide based on the sequence surrounding one of the MAP kinase phosphorylation sites on caldesmon [Ser-702 of chicken caldesmon or Ser-759 of human caldesmon; see Franklin et al. (7)]. Affinity-purified polyclonal antibodies against phosphocaldesmon (used at a 1:100 dilution) were shown to be specific for detecting ERK-phosphorylated vs. PKC-phosphorylated caldesmon, and no cross-reactivity was observed with ERK-phosphorylated myelin basic protein (7). The antiserum raised against full-length caldesmon (used at 1:50,000 dilution) was used to normalize the amount of phosphocaldesmon present in each sample.
Relative amounts of phosphocaldesmon were quantitated by densitometry of the exposed and scanned X-ray films using NIH Image.
Statistics. The statistical significance of the results expressed as means ± SE was assessed using an unpaired bilateral Student’st-test or a one-way repeated-measures ANOVA when appropriate (followed by a Student-Newman-Keuls multiple-comparison test when significant).
Inhibition of the ERK-MAP kinase cascade does not detectably modify the PE- and KCl-evoked contractions in presence of extracellular Ca2+.
To explore the regulatory role of ERK-MAP kinase on ferret aorta smooth muscle contraction, we used PD-098059. This compound prevents the activation of MAP kinase through inhibition of activation of the upstream kinase MEK and has been used widely to investigate the role of ERK-MAP kinase (6, 23). Aortic strips were preincubated for 30 min in the presence of PD-098059 (30 μM) or its vehicle (DMSO) and were then submitted to cumulative addition of PE. These experiments were performed in PSS containing 2.5 mM Ca2+, and the results were expressed as percentage of a control KCl-evoked contraction elicited at the beginning of the experiment. As depicted in Fig.1, the PE-induced contraction measured in the presence of the inhibitor was slightly but not significantly attenuated (P > 0.05 for all points) compared with the control concentration-response curve. Furthermore, PD-098059 did not significantly (P > 0.1) modify the contractile response of aortic strips to a depolarizing solution (Fig. 1, right).
Inhibition of the ERK-MAP kinase cascade reduces the Ca2+-independent contraction to PE.
In this set of experiments, ferret aortic strips were stimulated with a single dose of PE (0.3 or 10 μM). When the evoked contraction reached a steady state, 3 mM EGTA was added. The steady-state portion of the contraction that persists after the addition of EGTA corresponds to the previously described Ca2+-independent contractile tone. As shown on Fig.2 A, in control strips, PE elicited significant and concentration-dependent levels of maintained tone in the absence of extracellular Ca2+. In contrast, when strips were pretreated with 30 μM PD-098059, the Ca2+-independent contraction was either completely abolished (0.3 μM PE) or significantly inhibited (10 μM PE).
To demonstrate that the effect of PD-098059 observed on the contraction is actually related to MAP kinase activation, we measured the PE-evoked phosphorylation of MAP kinase using an ERK1/ERK2 phospho-MAP kinase antibody (Tyr-202/Tyr-204) on protein-matched samples. As shown in Fig.2 B, PE elicited a significant and concentration-dependent increase in phosphorylation of MAP kinase, which was completely inhibited by PD-098059 (10 μM). These experiments were also performed in the presence of Ca2+ to verify that the MAP kinase phosphorylation observed in our model was not an artifact of the Ca2+-free bathing medium, due possibly to the inactivation of a Ca2+-dependent process. The results confirmed that, when muscles were treated with PE (10 μM in PSS containing 2.5 mM Ca2+), the relative amount of phospho-MAP kinase was also increased (2.6 ± 0.3-fold vs. 2.2 ± 0.3-fold in absence of Ca2+;P > 0.05). PD-098059 was also able to prevent MAP kinase phosphorylation in both conditions. When muscles were pretreated with the MEK inhibitor before the treatment with PE, the relative amount of steady-state phospho-MAP kinase was not significantly different from that in the control preparations (n = 5,P > 0.05). The time course of changes in phospho-MAP kinase was also determined in the presence of extracellular Ca2+ (Fig.2 C). Peak increases in phospho-MAP kinase occurred at 5 min. We have previously reported that, in the absence of extracellular Ca2+, a similar time course is obtained with a peak value at 4 min. (17).
Caldesmon phosphorylation evoked by PE in absence of extracellular Ca2+ is ERK-MAP kinase dependent.
We next used a newly developed antibody, shown to be selective for caldesmon phosphorylated at a MAP kinase site, to assess the ERK-MAP kinase-dependent phosphorylation of caldesmon in ferret aorta, specifically in the Ca2+-independent component of agonist-evoked contraction.
As shown in Fig. 3,top, when aortic strips were incubated in the absence of extracellular Ca2+, the level of phosphorylated caldesmon, detected by immunoblotting, was significantly increased following treatment with 10 μM PE. This stimulatory effect of 10 μM PE on caldesmon phosphorylation was not observed when the strips were preincubated with 30 μM PD-098059. The level of phosphocaldesmon, normalized for the total amount of caldesmon present, was almost threefold (2.7 ± 0.5, P< 0.05) higher after treatment with 10 μM PE than in control or PD-098059-pretreated conditions. Figure 3,bottom, illustrates that, when these experiments were performed in the presence of Ca2+, a very similar pattern of caldesmon phosphorylation was observed. No differences from the Ca2+-free conditions were observed for the basal and PE-stimulated levels of phosphocaldesmon.
In this study, we have shown that PE-induced activation of ERK-MAP kinase in ferret aorta is responsible for the Ca2+-independent component of PE-evoked contraction. Inhibition of the MAP kinase cascade by PD-098059, a synthetic inhibitor that selectively prevents the activation of the upstream kinase MEK (6, 23), was shown to dramatically attenuate the contraction evoked in the absence of extracellular Ca2+ by submaximally and maximally effective concentrations of PE. Furthermore, we also present direct evidence, by using a recently developed antibody directed against phosphorylated caldesmon, that in our model caldesmon is a physiological substrate for agonist-evoked ERK-MAP kinase and that, when phosphorylation of caldesmon is prevented, so is the PE-induced contraction. It is important to note that, although several studies have recently focused on the role of ERK-MAP kinase and of its potential substrate caldesmon in smooth muscle contraction (see below), our use of PD-098059 to knock out MAP kinase activation allows the first demonstration of a cause-and-effect relationship between ERK-MAP kinase activation and smooth muscle contraction.
Adam and colleagues (1) showed, by partially purifying MAP kinase from porcine carotid arteries, that the level of MAP kinase activity is increased by both mechanical load and pharmacological stimulation. Simultaneously, Katoch and Moreland (15) reported that, in swine carotid artery, the temporal profile of MAP kinase activation/inactivation in response to histamine stimulation and membrane depolarization was similar to that of contraction/relaxation. Both studies proposed that, among the thin-filament-based regulatory proteins, caldesmon was the substrate for activated MAP kinase in differentiated smooth muscle. This assumption was based on a previous observation by Adam and Hathaway (2) that the caldesmon residues phosphorylated in intact vascular tissue are consistent with MAP kinase-catalyzed phosphorylation sites in purified caldesmon. However, these studies only reported in vitro assay of caldesmon phosphorylation using purified caldesmon, and the data were not correlated with contraction measurements. Nixon et al. (20), using recombinant activated p42mapk, showed that caldesmon was effectively phosphorylated by p42mapk but without any change in contraction, arguing against a role for phosphorylation by MAP kinase in regulating contraction when added to a Triton X-100-permeabilized fiber. Although the species (rabbit vs. ferret in our study) constitutes an obvious difference between their studies and ours, the Triton X-100-skinned preparations used by Nixon and colleagues (20) could per se account for subtle but critical changes in smooth muscle biology such as the loss or mistargeting of proteins involved in contraction (particularly for its Ca2+-independent component, see below). In the present study, the use of a specific phosphocaldesmon antiserum allowed us to detect in situ an increase in the relative amount of phosphorylated caldesmon after exposure to PE. Combined with the inhibitory effect of the MEK inhibitor PD-098059 (30 μM) on this agonist-induced caldesmon phosphorylation, our results are the first to show a direct relation between agonist-induced activation of ERK-MAP kinase and caldesmon phosphorylation in intact muscle.
Another major finding in our study arises from the isolation of the Ca2+-independent component of smooth muscle contraction. Different authors have indeed reported the possible involvement of the ERK-MAP kinase in the contraction process (8, 9, 21, 24) but did not address the relationship with the LC20 phosphorylation-mediated pathway. In ferret aorta smooth muscle, we have previously shown that, whereas the Ca2+-dependent phosphorylation of the 20-kDa regulatory light chain of myosin is, like in other smooth muscle types, a major determinant of smooth muscle contraction, a significant part of agonist-evoked contraction persists in Ca2+-free bathing solution in the absence of any detectable increase in LC20 phosphorylation above basal level (19). This allows a Ca2+-independent pathway to be studied in isolation, separate from agonist-induced changes in LC20 phosphorylation. In the present study, we found that this Ca2+-independent contraction was dramatically reduced by the MEK inhibitor PD-098059 (30 μM), contrary to the Ca2+-dependent component of the PE-evoked contraction that is not detectably affected by the inhibitor. Importantly, the phosphorylation of both ERK-MAP kinase and caldesmon and their inhibition by PD-098059 were also observed when Ca2+ was present in our experimental protocol, indicating that this process is not due to a nonphysiological alteration of Ca2+ homeostasis. Our findings suggest, therefore, that, in addition to the LC20 phosphorylation pathway, the MAP kinase signaling pathway is also involved in the ferret aorta contraction. The time course of MAP kinase activation is similar to that of contractile activation, both in the presence of extracellular Ca2+ (Fig.2 C) and in the absence of extracellular Ca2+ (17). Most importantly, during steady state of the contraction (as opposed to the early development of the contraction) and throughout the Ca2+-free contraction, LC20 phosphorylation levels are diminished to basal or near-basal values, pointing to the need for additional signaling mechanisms to explain the contraction (3, 12,16-19).
The absence of a significant effect of the MEK inhibitor on the depolarization-evoked contraction (this study and Ref. 24) and on the agonist-induced maximal contraction (10) suggests that the ERK-MAP kinase pathway may not play a major role in situations in which the LC20 phosphorylation levels are high. Thus, although we cannot exclude that the activation of MAP kinase is involved in the development of smooth muscle contraction, our results indicate that MAP kinase phosphorylation is certainly involved in the maintenance of vascular tone, independent of any Ca2+-dependent processes such as phosphorylation of LC20 by myosin light chain kinase. Finally, it is noteworthy that PD-098059 is unable to completely inhibit the Ca2+-independent contraction in response to the highest concentration of PE, whereas, under the same conditions, this compound can completely inhibit the agonist-induced MAP kinase phosphorylation. This raises the possibility that additional pathways may be involved in the regulation of smooth muscle contraction.
In summary, we have demonstrated a direct relationship between ERK-MAP kinase activation and smooth muscle contraction, and we have also provided new evidence that phosphorylation of caldesmon could be part of the process linking agonist-induced ERK-MAP kinase activation and contraction, independently of Ca2+-evoked LC20 phosphorylation.
We are very grateful to Dr. Leonard P. Adam (Boston Biomedical Research Institute, Boston, MA) for generously providing the phosphocaldesmon antibodies and for helpful discussions.
Address for reprint requests: K. G. Morgan, Boston Biomedical Research Institute, 20 Staniford St., Boston, MA 02114-2500.
C. Dessy is a fellow of the Belgian American Educational Foundation and of the D. Collen Foundation.
This work was supported by National Heart, Lung, and Blood Institute Grants HL-31704 and HL-42293 to K. G. Morgan and by a grant from the Korea Science and Engineering Foundation to I. Kim.
- Copyright © 1998 the American Physiological Society