Ezrin, radixin, and moesin (ERM) proteins are known to be substrates of Rho kinase (ROCK), a key player in vascular smooth muscle regulation. Their function in arteries remains to be elucidated. The objective of the present study was to investigate ERM phosphorylation and function in rat aorta and mesenteric artery and the influence of ERM-binding phosphoprotein 50 (EBP50), a scaffold partner of ERM proteins in several cell types. In isolated arteries, ERM proteins are phosphorylated by PKC and ROCK with different kinetics after either agonist stimulation or KCl-induced depolarization. Immunoprecipitation of EBP50 in noradrenaline-stimulated arteries allowed identification of its interaction with moesin and several other proteins involved in cytoskeleton regulation. This interaction was inhibited by Y27632, a ROCK inhibitor. Moesin or EBP50 depletion after small interfering RNA transfection by reverse permeabilization in intact mesenteric arteries both potentiated the contractility in response to agonist stimulation without any effect on contractile response induced by high KCl. This effect was preserved in ionomycin-permeabilized arteries. These results indicate that, in agonist-stimulated arteries, the activation of ROCK leads to the binding of moesin to EBP50, which interacts with several components of the cytoskeleton, resulting in a decrease in the contractile response.
- ezrin-radixin-moesin-binding phosphoprotein 50
- vascular smooth muscle
ezrin, radixin, and moesin (ERM) proteins are established Rho kinase (ROCK) substrates. These proteins possess a very similar structure with a COOH-terminal actin-binding domain and a NH2-terminal FERM-interacting domain. Stimulation of ERM proteins requires the phosphorylation of a carboxyl terminal threonine residue (7, 25, 49), which is mainly mediated by ROCK (44, 51), but PKC (29, 71), SLIK (9, 38), and lymphocyte-oriented kinase (LOK) (4) have also been identified as mediators of ERM phosphorylation. After stimulation, the FERM domain of ERM proteins binds to their interacting partner either directly or through the interaction with ERM-binding phosphoprotein 50 kDa (EBP50; also known as NHERF1) (60). Studies on the function of ERM proteins have revealed their importance in the regulation of the cytoskeleton in several cell types. Ezrin is known to be critical for the formation of microvilli in intestinal epithelial cells (66, 76). Moesin is involved in the formation of the mitotic spindle by a stabilization of microtubules (9, 38). Radixin is linked to the formation of cochlear stereocilia, and its impairment causes nonsyndromic deafness (32, 35). ERM proteins have also been associated with Rho and Rac-dependent formation of focal adhesions in fibroblasts (43), uropod formation in migrating lymphocytes (39), changes in cell shape after osmotic stress (59, 71) or microtubule rearrangements induced by human immunodeficiency virus infection (3, 23, 37). However, little is known about the physiology of ERM proteins in vascular smooth muscle. Upregulation of moesin expression and ROCK-dependent ERM phosphorylation are described in neointima coronary smooth muscle cells after arterial injury (5, 45), and increase in phospho-ERM is found in atherosclerotic plaque in carotid artery (61), suggesting that these proteins could be involved in the development of atherosclerosis. However, it is not known whether they contribute to the regulation of vascular tension.
The contraction of vascular smooth muscle is controlled via the phosphorylation of myosin light chain MLC20 by myosin light chain kinase (MLCK) and its dephosphorylation by myosin light chain phosphatase (MLCP). MLCK is activated by the calcium-calmodulin complex following an increase in cytosolic Ca2+. By catalyzing inhibitory phosphorylation of MYPT, the regulatory myosin-binding subunit of MLCP, ROCK increases the level of phosphorylation of MLC20 and exerts an important calcium-independent regulation of contraction (14, 69, 70). There is now wide evidence that the actin cytoskeleton is involved in the regulation of smooth muscle contraction and contributes to the adaptation of contractility to external mechanical stimulation or agonist stimulation (22, 33). Actin filaments are fixed to the plasma membrane at the level of dense plaques, where they interact with the cytoplasmic domain of transmembrane integrins through linker proteins such as vinculin, actinin, and talin. The extracellular domain of integrins binds to extracellular matrix elements so that dense plaques form a link between the contractile machinery and the extracellular matrix (8, 73). In cultured cells, ROCK is known to mediate actin dynamics and to be responsible for cytoskeleton changes (19, 48, 54), and Y27632, a specific ROCK inhibitor, has been shown to dismantle stress fibers (42) and focal adhesions (31). However, the contribution of ROCK-dependent cytoskeleton changes to the regulation of arterial tone and the involvement of ERM in this process are unknown.
The aim of the present work was to investigate the functional role of ERM proteins and their binding partner EBP50 in vascular smooth muscle cells. The stimulation of ERM proteins was tested in intact aorta and mesenteric arteries in response to various stimuli. The use of intact arteries preserved the integrity of the cytoskeleton and the interactions between cells, which are important in the signal transduction of the ROCK pathway. Indeed, cultured vascular smooth muscle cells exhibit a noncontractile, proliferative, and migratory phenotype, which is not relevant to the study of processes involved in the development of contraction (52, 62). The interaction between ERM and EBP50 was tested by immunoprecipitation in unstimulated and stimulated arteries, and the nature of other partners recruited by EBP50 was investigated by mass spectrometry. To determine the function of EBP50 and ERM in an intact artery, their contribution to the regulation of the vascular tone was investigated in small interfering RNA (siRNA) transfected mesenteric artery. Results indicate that a ROCK-dependent moesin/EBP50 pathway involving several elements of the cytoskeleton is activated by artery stimulation and contributes to the regulation of vascular tone.
MATERIALS AND METHODS
Male Wistar rats (220–250 g body wt) were anesthetized with diethyl ether and killed by decapitation in accordance with international guidelines and with the approval of the local ethics committee for animal experiments. The thoracic aorta and the superior mesenteric artery were removed and cleaned of connective tissue.
ERM phosphorylation assay in aorta and mesenteric artery.
The superior mesenteric artery or rings cut from the aorta (4 mm) were used, and the endothelium was denuded and immersed in physiological solution (in mM: 122 NaCl, 5.9 KCl, 15 NaHCO3, 10 glucose, 1.25 MgCl2, and 1.25 CaCl2) maintained at 37°C and continuously gassed with 95% O2-5% CO2 for at least 2 h. When required, artery rings were incubated with antagonist for 20 min before stimulation. Artery rings were then transferred to physiological solution containing either noradrenaline or angiotensin II, or to a 100 mM KCl solution (in mM: 27 NaCl, 100 KCl, 15 NaHCO3, 10 glucose, 1.25 MgCl2, and 1.25 CaCl2), with or without antagonist. One ring from the same artery was not stimulated and was used as a control. After stimulation, artery rings were immediately frozen in liquid nitrogen. The rings were fixed in cold acetone-10% TCA for 1 h at −80°C and washed twice with cold acetone (−20°C). Proteins were extracted in a denaturating buffer containing 9.5 M urea, 2/100 Igepal CA-630, 5/100 β-mercaptoethanol, 1/100 pharmalytes (pI: 3–10), 1 mM vanadate, 5 mM NaF, and a protease inhibitor cocktail (Roche). Proteins were resolved by SDS-PAGE (4–12% NuPAGE, Invitrogen) and blotted onto polyvinylidene difluoride (PVDF) membrane (iBlot, Invitrogen). Membranes were probed with phospho-ERM [phospho-ezrin (Thr567)/radixin (Thr564)/moesin (Thr558)] antibody (1/1,000 Cell Signaling) and actin antibody (1/5,000, Santa Cruz) as loading control. IRDye-conjugated fluorescent secondary antibodies (680 nm and 800 nm) were used to detect primary antibodies. Bands were detected and quantified by an Odyssey infrared imaging (Li-Cor). Phospho-ERM expression was normalized to the expression of actin in each sample.
siRNA transfection of mesenteric arteries.
Superior mesenteric artery was used for siRNA transfection experiments: artery was removed under sterile condition, cut in 2-mm rings, and endothelium-denuded. The rings from one artery were separated into two groups for transfection with scramble siRNA or the siRNA of interest (Stealth, Invitrogen). The rings were successively immersed in solution containing (in mM) 1) 10 EGTA, 120 KCl, 5 ATP, 2 MgCl2, and 20 TES (pH 6.8) for 20 min at 4°C; 2) 120 KCl, 5 ATP, 2 MgCl2, 20 TES, and 40 nM siRNA (pH 6.8) for 3 h at 4°C; 3) 120 KCl, 5 ATP, 10 MgCl2, 20 TES, and 40 nM siRNA (pH 6.8) for 30 min at 4°C; and 4) 140 NaCl, 5 KCl, 10 MgCl2, 5 glucose, 2 MOPS (pH 7.2), and Ca2+ gradually increased from 0.01 to 0.1 to 1.8 mM every 15 min, from 4°C to room temperature (12). Mesenteric artery rings were then immersed in DMEM culture medium supplemented with 1 mM l-glutamine, 50 U/ml penicillin, and 50 μg/ml streptomycin and maintained in an incubator (37°C, 5% CO2) for 3 or 4 days. Duration of culture was determined on the basis of preliminary experiments to give a reproducible effect on artery contractile responses. siRNA duplexes were designed with Block-iT RNAi designer software (Invitrogen) with accession number AF154336 for EBP50 and NM030863 for moesin. The scramble siRNA was the Stealth RNAi control low GC siRNA (Invitrogen). The sequences of siRNA duplexes were AUAAAUGUCAUCAUUGAGAAUGCC and GGGCAUUCUCAAUGAUGACAUUUAU for moesin and GAGAAAGGCAAGGUGGGCCAGUUUA and UAAACUGGCCCACCUUGCCUUUCUC for EBP50.
siRNA delivery into mesenteric arteries was assessed by transfection of a BLOCK-iT Alexa Fluor Red Fluorescent Control (40 nM, Invitrogen). Transfected arteries were immersed into Hanks' solution (composition in mM: 140 NaCl, 4.2 KCl, 1.2 KH2PO4, and 10 HEPES, pH 7.4 with Tris) containing 4.5 mg/ml DTT, 6 mg/ml papain, and 6.5 mg/ml bovine serum albumin for 14 min at 37°C. Myocytes were isolated by gentle trituration and mounted on coverglass with Prolong antifade reagent (Invitrogen). Nuclei were stained with DAPI and cells were observed with a Zeiss Axiovert S-100 microscope (Zeiss, Oberkochen, Germany) (Supplemental Fig. S3; Supplemental Material for this article is available online at the Journal website).
Knockdown efficiency was assessed by real-time quantitative PCR (RTQ-PCR) and Western blotting. RNA was extracted from mesenteric arteries with TriPure isolation reagent (Roche). RNA extraction was quantified by measuring absorbance with a Nanodrop spectrophotometer (ThermoScientific), and RT was performed with 1 μg total RNA for 1 h at 37°C with Ready-TO-GO You-Prime First-Strand beads (Amersham Biosciences) and oligo(dT)12–18 (Amersham) as primers. RTQ-PCR analyses were performed in duplicates using iQ SYBR Green Supermix (Bio-Rad). The sequences of primers were (5′ to 3′) as follows: moesin, AGTTCCGGGCCAAGTTCTAT and CCCACTGGTCCTTGTTGAGT; ezrin, ACCAAGAAGGCTCAGCAAGA and AGCTCTGCCATTTCTGAAGC; radixin, CTCCATGCTGAGAACGTCAA and CCTCGGGTTCTGCTAGTGAG; ebp50, AGAAGGAGAACAGCCGTGAA and TGCTCAGAGGTTGCTGAAGA; and gapdh, AGACAGCCGCATCTTCTTGT and CTTGCCGTGGGTAGAGTCAT (Eurogentech). PCR conditions were as follows: 95°C for 3 min followed by 40 10-s cycles at 95°C, 1 min at 60°C and 71 15-s cycles at 55°C (My iQ, iQ5, Bio-Rad). The relative changes in target gene/GADPH mRNA ratio were determined by the following formula: 2ΔΔCT, where CT is the threshold cycle number.
Immunoblotting was performed as described above. Membranes were probed with moesin antibody (1/1,000, Cell Signaling) or EBP50 antibody (1/1,000, Sigma), and actin antibody (1/5,000, Santa Cruz) was used as loading control. Bands were detected and quantified by an Odyssey infrared imaging (Li-Cor).
Measurement of artery contraction.
Rings from siRNA-transfected mesenteric artery were mounted on two 40-μm- diameter wires for isometric tension recording in a wire myograph (model 500A, Danish Myo Technology) as described previously (20). Incubation baths were filled with physiological solution gassed with 95% O2-5% CO2 at 37°C supplemented with Nω-nitro-l-arginine (100 μM) to prevent any effect of endogenous nitric oxide. Basal tension was set at a tension equivalent to that generated at 0.9 times the diameter of the vessel at 100 mmHg. A first contraction was evoked by changing the solution in the bath to the 100 mM KCl solution. After washout and recovery in physiological solution, contraction was evoked by adding increasing concentrations of noradrenaline or angiotensin II. Contractions evoked by increasing KCl concentrations were performed in the presence of phentolamine (1 μM). To determine the calcium sensitivity of contraction, mesenteric artery rings were permeabilized with ionomycin (1 μM, 3 min). The arteries were then washed with a Ca2+-free solution (in mM: 122 NaCl, 5.9 KCl, 15 NaHCO3, 10 glucose, 10 HEPES, and 1 EGTA) maintained at 37°C and gassed with 95% O2-5% CO2. Contraction was induced by increasing free-calcium concentration to 1 μM, determined with Maxchelator software (Stanford University, Stanford, CA). When tension was stabilized, increasing concentrations of noradrenaline were added. Contractile responses were expressed as the percentage of the maximal response or as the percentage of the contraction induced by 100 mM KCl stimulation in the same artery ring.
Immunoprecipitation of protein extracts from arteries.
Superior mesenteric arteries were allowed to recover for 2 h in physiological solution at 37°C. Where required, arteries were incubated with Y27632 (10 μM) for 20 min. Arteries were then stimulated with noradrenaline (1 μM) for 2 min with or without Y27632, immediately flash-frozen with liquid nitrogen, and stored at −80°C. Five arteries were used for each condition. Proteins were extracted with a nondenaturating extraction buffer containing 0.5/100 Igepal CA-630, 1 mM dithiothreitol, 50 mM HEPES (pH 7.6), and 1/100 Halt protease/phosphatase (ThermoFischer). Protein concentration was determined by the Bradford method (6). Protein A Sepharose was incubated with 2 μg anti-EBP50 antibody (Sigma) for 1 h at 4°C and then washed three times with cold PBS buffer (in mM: 137 NaCl, 6 KCl, 1.2 MgCl2, 2 CaCl2, 10 glucose, and 10 HEPES, pH 7.4 with Tris). Subsequently, protein extracts were incubated with protein A Sepharose and EBP50 antibody complex for 2 h at 4°C and were then washed three times with cold PBS (once with 2× concentrated PBS) followed by boiling for 5 min in NuPAGE sample buffer diluted with nondenaturating extraction buffer.
Identification of proteins by mass spectrometry.
Immunoprecipitates were loaded on a 10% (wt/vol) polyacrylamide-SDS gel. After electrophoresis the gel was stained with Colloidal Coomassie Blue (Fermentas). Bands of interest were excised and digested with trypsin. Peptides were analyzed by capillary LC-tandem mass spectrometry (LC/MS-MS) in a LTQ XL ion trap mass spectrometer (ThermoScientific) fitted with a microelectrospray probe. The data were analyzed with Proteome Discoverer software (ThermoScientific), and the proteins were identified with SEQUEST against a target-decoy nonredundant rat protein database obtained from EMBL-EBI IPI RAT v 3.53. The false discovery rate was estimated by target-decoy database search and was set below 5% (2).
Immunoprecipitates were loaded on a 4–12% NuPAGE SDS-PAGE gel (Invitrogen). After electrophoresis, proteins were transferred onto PVDF membrane (iBlot). Membranes were probed with anti-actin (Santa Cruz), anti-vinculin (Abcam), or anti-moesin (Cell Signaling) antibodies (1/500 for all). Bands were detected as described above.
Values are expressed as means ± SE from at least three determinations. Statistical analysis was made using either a two-way ANOVA with a Bonferroni post test to evaluate treatment differences or a one-way ANOVA to compare two groups of data. P < 0.05 was considered to be significant. EC50 was calculated by nonlinear curve fitting, using a classical sigmoïde concentration-response function (Prism, GraphPad software).
Noradrenaline and vasopressin were purchased from Sigma-Aldrich, GÖ6983 and Y27632 were purchased from Tocris, and H1152 was purchased from Calbiochem under the name “ROCK inhibitor IV.”
Phosphorylation of ERM proteins in rat artery.
Stimulation of rat aorta with noradrenaline (1 μM) induced a rapid phosphorylation of ERM, which was maximum at 2 min and slightly decreased after 10 min. (Fig. 1A). Similar increase in ERM phosphorylation was observed with vasopressin (Supplemental Fig. S1). Depolarization of aorta with high-KCl (100 mM) solution led to a similar rapid (30 s) but transient increase in ERM phosphorylation, followed by a secondary response observed after 10 min (Fig. 1A). These results indicated that aorta stimulation by agonist or by depolarization increased the phosphorylation of ERM proteins. Similar results were obtained in mesenteric artery (Supplemental Fig. S2). ERM phosphorylation was further characterized in aorta, which provided larger amount of material.
To determine whether ROCK was involved in ERM phosphorylation, we compared the level of phosphorylated ERM in response to noradrenaline or KCl in segments of the same aorta preincubated with or without the ROCK inhibitor Y27632. Y27632 (10 μM) did not affect the rapid increase in ERM phosphorylation induced either by noradrenaline or by high KCl (Fig. 1, B and C) but markedly decreased ERM phosphorylation measured after 2 min or 10 min stimulation with noradrenaline (Fig. 1B) and after 10 min stimulation with KCl (Fig. 1C). As Y27632 has also been reported to inhibit other kinases (13), we tested the effect of a newer and more specific ROCK inhibitor, H1152 (1 μM) (72), on ERM phosphorylation evoked by noradrenaline stimulation. The same results were obtained in the presence of H1152 as observed with Y27632 (Fig. 1B).
The involvement of PKC in ERM phosphorylation was tested under the same condition by using the broad spectrum PKC inhibitor Gö6983 (21). Gö6983 (1 μM) suppressed the rapid (30 s) phosphorylation evoked by both noradrenaline and KCl but did not affect the phosphorylation induced by 2 min stimulation with noradrenaline (Fig. 2A), which on the contrary was sensitive to the ROCK inhibitor (Fig. 1B).
Considering that KCl-induced depolarization activates voltage-operated calcium channel (VOCC) and produces a massive calcium entry into the cell, we tested the effect of nimodipine, a selective inhibitor of L-type VOCC, to assess the calcium dependence of ERM phosphorylation. Nimodipine (1 μM) completely inhibited ERM phosphorylation measured after 30 s of depolarization with KCl (Fig. 2B) but did not significantly affect ERM phosphorylation seen after longer stimulation. The calcium ionophore ionomycin (1 μM, applied for 30 s) also significantly increased ERM phosphorylation (Fig. 2C). Ionomycin-induced phosphorylation was abolished by Gö6983 but was unaffected by Y27632 (Fig. 2C), suggesting that calcium could activate a rapid phosphorylation of ERM through the activation of conventional Ca2+-dependent PKC.
Immunoprecipitation of EBP50: identification of interacting proteins by mass spectrometry.
ERM proteins are known to interact with EBP50, a scaffold protein that possesses a FERM domain that binds ERM proteins and two distinct PDZ domains capable of interacting with other proteins (60). To identify potential interaction of ERM proteins with EBP50 in isolated artery, EBP50 was immunoprecipitated in tissue extracts from unstimulated mesenteric arteries and from arteries stimulated with noradrenaline for 2 min. Mesenteric artery was used because protein recovery after nondenaturating extraction was better than in aorta and because it allowed efficient transfection experiments (see below). One series of mesenteric arteries was preincubated with 10 μM Y27632 before stimulation with noradrenaline. As shown on Fig. 3A, immunoglobulin bands were observed in all the samples incubated with EBP50 antibodies. Compared with controls performed without antibody or without protein extract, specific additional bands were observed in noradrenaline-stimulated arteries only, but not in unstimulated arteries or in arteries stimulated with noradrenaline in the presence of Y27632. More than 12 different specific bands, with apparent molecular weight >35,000, were excised for digestion with trypsin and analyzed by mass spectrometry (Fig. 3A). The EBP50 interacting proteins identified by LC/MS-MS are summarized in Table 1. They included moesin but not ezrin or radixin. Several other proteins related to the cytoskeleton were also identified. These results were validated by performing coimmunoprecipitation of moesin, actin, or vinculin, three proteins for which antibodies were available, with EBP50 in protein extracts from unstimulated and noradrenaline-stimulated mesenteric artery. As shown on Fig. 3B, 2 min noradrenaline stimulation of mesenteric arteries induced the interaction of EBP50 with moesin, actin, and vinculin, while no interaction was detected in unstimulated arteries, confirming the data obtained by mass spectrometry.
Knockdown of moesin and EBP50: functional consequence for the vascular tone.
To determine whether the EBP50-moesin complex identified in noradrenaline-stimulated artery contributes to the regulation of vascular contractility, we tested the effect of the inhibition of these proteins on the contractile response of mesenteric artery. Because pharmacological inhibitors of ERM proteins are unavailable, siRNA transfection by reverse permeabilization was used to deplete mesenteric arteries in moesin or EBP50. The delivery of siRNA by reverse permeabilization was validated by transfecting a red fluorescent oligonucleotide into intact mesenteric artery. Vascular smooth muscle cells were then isolated by enzymatic digestion and observed with an epifluorescence microscope. A large proportion of elongated red cells was visible with few isolated nuclei, confirming the entry of red fluorescent siRNA (Supplemental Fig. S3).
Anti-moesin siRNA or scramble siRNA was transfected in different segments from the same mesenteric artery. After 72 h of culture, moesin mRNA expression was inhibited by 63 ± 5% (n = 5) compared with artery segments transfected with scramble siRNA, while no significant change was observed for ezrin, radixin, and EBP50 mRNA levels. Moesin protein expression was inhibited by 41 ± 9% (n = 5) (Fig. 4, A and B).
To assess the role of moesin in artery contraction, we measured contractility in transfected arteries in response to different stimulations. No significant difference was observed in contraction induced by various KCl concentrations comparing artery rings transfected with scramble siRNA to those transfected with moesin-specific siRNA (Fig. 5A; KCl EC50 values were 14 mM and 13 mM for moesin siRNA and control scramble siRNA transfected arteries, respectively; P = 0.38, n = 6), and maximal contraction induced by 100 mM KCl stimulation was of similar amplitude: arteries transfected with scramble siRNA and moesin siRNA produced maximal contraction of 3.7 ± 0.4 mN/mm and 3.8 ± 0.4 mN/mm, respectively (n = 13, P = 0.9). However, the concentration-response curve to noradrenaline was shifted to the left in moesin siRNA-transfected arteries compared with scramble siRNA-transfected arteries (P < 0.0001, n = 8). Noradrenaline EC50 values were 84 nM and 17 nM, in scramble siRNA- and moesin siRNA-transfected arteries, respectively (Fig. 5B). The amplitude of maximal contraction induced by noradrenaline (3 μM) was significantly increased in moesin-depleted arteries: values were 65 ± 9% and 83 ± 4% of the contraction evoked by 100 mM KCl in scramble siRNA-transfected arteries (n = 7) and moesin siRNA-transfected arteries (n = 7), respectively (P < 0.05). A similar leftward shift of the concentration-response curve for angiotensin II was also observed in moesin siRNA-transfected artery rings compared with scramble siRNA-transfected artery rings (P = 0.001, n = 5); angiotensin II EC50 values were 0.97 nM and 2.5 nM, respectively (Fig. 5C).
To determine whether moesin was directly involved in the regulation of contractility in response to agonists or whether it was acting through EBP50, we knocked down EBP50 by transfection of an anti-EBP50 siRNA. After 96h of culture, EBP50 mRNA expression was reduced by 62 ± 11% (n = 10) and no significant changes were observed for ezrin, radixin, and moesin mRNA. EBP50 protein expression decreased by 61 ± 5% (n = 4) (Fig. 4, C and D). In EBP50 siRNA-transfected artery rings, the noradrenaline concentration-response curve was shifted to the left compared with scramble siRNA-transfected artery rings (P < 0.0001, n = 9). Noradrenaline EC50 values were 98 nM and 16 nM for scramble siRNA and EBP50 siRNA-transfected arteries, respectively (Fig. 5D). The amplitude of maximal contraction induced by noradrenaline (3 μM) was significantly increased in EBP50-depleted arteries: values were 85 ± 8% and 145 ± 27% of the contraction evoked by 100 mM KCl in scramble siRNA-transfected arteries (n = 9) versus EBP50 siRNA-transfected arteries (n = 9), respectively (P < 0.05). Thus, decreased expression of moesin or EBP50 resulted in a similar potentiation of agonist-evoked contraction, suggesting that this effect could be mediated by an interaction between moesin and EBP50.
To further characterize the mechanism of action of moesin and EBP50 on artery contraction, we measured the contraction in artery rings permeabilized with ionomycin, which allowed to equilibrate the intracellular and extracellular concentrations of calcium. The contraction induced by noradrenaline in the presence of a calcium-EGTA buffer in the physiological solution is then not dependent on change in cytosolic calcium. There was no difference between segments of artery transfected with anti-moesin or anti-EBP50 siRNA and segments of the same artery transfected with scramble siRNA in the contractile response induced by 1 μM free-calcium (pCa 6) after permeabilization with 1 μM ionomycin for 3 min. The amplitude of contraction was 7.1 ± 0.4 mN/mm and 7.2 ± 0.6 mN/mm for arteries transfected with a scramble or anti-EBP50 siRNA and 7.6 ± 0.8 and 7.8 ± 0.7 mN/mm for arteries transfected with a scramble or anti-moesin siRNA (n = 6). Meanwhile, the noradrenaline concentration-response curve was significantly shifted to the left in arteries transfected with siRNA anti-moesin (P < 0.0001, n = 6) or with siRNA anti-EBP50 (P < 0.0001, n = 6), compared with segments of the same artery transfected with scramble siRNA (Fig. 6, A and B). Noradrenaline EC50 values were 14 nM and 61 nM for moesin siRNA- and scramble siRNA-transfected artery segments, respectively. For EBP50 siRNA-transfected artery segments, EC50 value for noradrenaline after ionomycin permeabilization was 14 nM and EC50 value for noradrenaline in scramble siRNA-transfected artery segments was 150 nM.
The present study shows that, in agonist-stimulated artery, the activation of ROCK leads to the phosphorylation of ERM proteins and to the formation of a complex between EBP50-moesin and several components of the cytoskeleton, resulting in a decrease in the contractile response to vasocontractile agonists.
Phosphorylation of ERM proteins at a carboxyl terminal threonine residue is required for their unfolding and their interaction with their partners (68). The present results indicate that stimulation of isolated arteries increased ERM phosphorylation in response to noradrenaline or 100 mM KCl solution. However, an important difference between agonists and KCl-induced ERM activation was the time course of the phosphorylation. Noradrenaline-stimulation resulted in a fast and relatively sustained ERM phosphorylation, while depolarization of the artery induced a transient increase in ERM phosphorylation. ERM proteins are considered as targets of ROCK (1, 27, 65), which is known to be activated by contractile agonists as well as by KCl depolarization (46). Involvement of ROCK in ERM activation was supported by the effects of the ROCK inhibitors Y27632 and H1152, which almost completely abolished ERM phosphorylation evoked by noradrenaline. Surprisingly, Y27632 did not affect the initial phase of phosphorylation observed in arteries stimulated with KCl or noradrenaline.
The lack of selectivity of kinase inhibitors is a major concern in modern pharmacology. Y27632 is a widely used selective ROCK inhibitor but has been identified to inhibit PKC, PKA, myristoylated alanine-rich PKC substrate (MARCKS), and PKN in vitro (13, 26). H1152, a more specific inhibitor of ROCK (72), exerted a similar effect as Y27632, suggesting that ROCK is indeed the major kinase involved in ERM phosphorylation in response to noradrenaline. Nevertheless, we cannot totally exclude the participation of other kinases in ERM phosphorylation, as for example PKN, which has been reported to be involved in stress fibers and focal adhesion formation but which sensitivity to H1152 is unknown (19). Moreover, neither Y27632 nor H1152 discriminates between the different isoforms of ROCK (ROCKI and ROCKII) (50). Their respective role cannot be identified in the present study.
The level of phospho-ERM at 30 s was significantly depressed in the presence of the PKC inhibitor Gö6983 both in noradrenaline and in KCl-stimulated arteries. Phosphorylation of ERM by PKC can be predicted because of the similarities of the phosphorylation consensus sequence motifs of ROCK and PKC (29). PKC-dependent phosphorylation of moesin has been well documented in collecting duct CD8 cells submitted to hypotonic stress (71), in leukocytes and in endothelial cells (56). The initial phosphorylation of ERM appeared to be calcium dependent, as indicated by its sensitivity to nimodipine, which suggests that PKC was activated via an increase in cytosolic calcium. This hypothesis was confirmed by the effect of the calcium ionophore ionomycin. The increased level of phospho-ERM induced by Ca2+ in ionomycin-permeabilized arteries was blocked by Gö6983, but unaffected by Y27632, indicating that ERM phosphorylation was mediated by a calcium-activated PKC and not by ROCK. We did not test the effect of nimodipine on ERM phosphorylation evoked by noradrenaline because agonist stimulation also activates non-voltage-dependent calcium entry through transient receptor potential (TRP) channels and intracellular calcium release, which are not inhibited by nimodipine (20).
These observations suggest that the activation of ERM results from two distinct pathways, which have different kinetics: a rapid phosphorylation involved Ca2+-activated PKC, and was followed by a slower ROCK-mediated phosphorylation. It is interesting to note that, likewise, two distinct signal transduction pathways have been described in phenylephrine-induced contraction of rabbit femoral artery: an initial, rapid, Ca2+-dependent and a slow Ca2+-independent mechanism, driven by PKC and ROCK, respectively (14).
Coimmunoprecipitation experiments were used to investigate whether, following activation, moesin bound to its known partner EBP50. Indeed, moesin-EBP50 interaction was first described in human placenta (60) and thereafter reported with purified moesin (68) and in CD4 T cells (18). The interaction between the moesin FERM domain and EBP50 has been characterized by crystallography (17). This interaction was hardly detectable in unstimulated artery but was markedly stronger after 2 min artery stimulation with noradrenaline. This is in agreement with the observation that 2 min noradrenaline stimulation induced the phosphorylation of ERM, which is required for its binding to EBP50 (68). The main particularity of EBP50 is that it possesses 2 PDZ domains, one of the most common protein-protein interacting motifs (24). The identification by mass spectrometry of the proteins interacting with EBP50 in noradrenaline-stimulated arteries revealed that almost all the EBP50 interacting partners were related to the cytoskeleton. They can be subdivided into four different groups: cytoskeletal filaments (actin, vimentin, tubulin, lamin, nuclear mitotic apparatus), motor proteins (nonmuscular myosin IIa, smooth muscle myosin, dynein), protein anchors (annexin, moesin, actinin, vinculin, talin, spectrin, filamin and plectin), and others (calreticulin, nucleolin). Some of the mass spectrometry data were confirmed by coimmunoprecipitation experiments. One of the most striking results was the identification of moesin but not ezrin or radixin as an interacting partner of EBP50. Interestingly, specific interactions between EBP50 and its partners were not detected in extracts of arteries stimulated with noradrenaline in the presence of the ROCK inhibitor Y27632, suggesting that the interaction between EBP50 and its partners required an intact ROCK activity. Because we observed that artery stimulation induced ERM phosphorylation by ROCK, we propose that ROCK-dependent phosphorylation of ERM is required for the interaction of EBP50 with moesin but could also be implicated in its interaction with its other partners. In smooth muscle cells, almost all the proteins identified as binding partners of EBP50 are found in specialized cytoskeletal domains: focal adhesion and dense body. The formation of such structures, and especially focal adhesion, is a complex dynamic event with a battery of assembly/disassembly and maturation processes, which is known to involve Rho family members (11, 19, 48, 54, 75). The present results suggest that moesin-EBP50 may represent a link between ROCK and focal adhesion related proteins. The following pathway may be proposed: in noradrenaline-stimulated artery, ROCK-dependent phosphorylation of ERM proteins results in the subsequent activation of moesin, allowing its interaction with the FERM domain of EBP50, which in turn interacts with other cytoskeleton-associated proteins as detected by the immunoprecipitation experiment, through its two PDZ domains. Among these proteins, α-actinin has already been detected as an interacting partner of PDZ domain containing proteins (10, 28, 36, 74). Similarly, vimentin (55), filamin (74), and myosin (74, 77) also bind to PDZ domains. The interaction of EBP50 with moesin and cytoskeleton-related proteins in response to agonist stimulation could then be involved in the modulation of vascular tone.
In the absence of pharmacological inhibitors of moesin and EBP50 proteins, knockdown was performed to investigate their role in the regulation of vascular tone in isolated arteries. Since classical transfection reagents are ineffective to transfect whole arteries, a reverse permeabilization protocol was applied (12, 15, 41, 47, 64). This protocol did not affect the ability of artery to contract, and there was no difference between arteries cultured for 72 h with or without prior reverse permeabilization with or without scramble siRNA, indicating that this protocol does not trigger any process that could affect contraction (data not shown). We focused on the role of moesin because ezrin and radixin were found not to interact with EBP50 in stimulated arteries and thus could have different targets. Surprisingly, depletion of moesin produced a significant increase in the contractile sensitivity to agonist and in the contractile force developed. This effect was specific for the responses to agonist-induced stimulations because contraction produced by KCl-depolarization was unaffected. Interestingly, the same effect was observed after depletion of EBP50, suggesting that the inhibitory action of moesin on the force development could be mediated through its interaction with EBP50. Experiments performed with ionomycin-permeabilized arteries showed that moesin or EBP50 depletion also potentiated noradrenaline-induced contraction in the presence of a fixed intracellular calcium concentration, suggesting that this effect does not involve a calcium-dependent component of the contractile machinery of the cell. Several studies point out the contribution of cytoskeleton elements to artery contraction. For example, microtubule depolymerization increases isometric force development in a way that requires ROCK (40, 53, 57, 58, 67). Besides, actin elongation has been shown to be necessary for force development in arteries (33) and inhibition of nonmuscular myosin II with blebbistatin decreases arterial tone (16, 30, 63). These results strongly suggest the involvement of cytoskeleton components in the regulation of contractile processes in arteries.
Taken together, the present results indicate that in stimulated artery, the activation of ROCK leads to the binding of moesin to EBP50, resulting in a decrease in contractile response to agonists. The implication of the cytoskeleton is suggested by the interaction of EBP50 with several elements of the cytoskeleton such as myosin, α-actinin, the major component of the smooth muscle dense bodies, or tubulin, the major component of microtubule network. The future challenge will be to identify how cell architecture could modulate the development of smooth muscle contraction. Moreover, since vascular pathologies have been associated with increased moesin phosphorylation (61), further studies should investigate whether its binding partner EBP50 might be involved in the development of these diseases.
This work was supported by grants from the Ministère de l'Education et de la Recherche Scientifique of the Belgian French Community (Action Concertée no. 06/11-339), the Fonds pour la Recherche Scientifique Médicale, the Interuniversity Attraction Poles Program - Belgian Science Policy (P6/28), the Directorate General Higher Education and Scientific Research, French Community of Belgium, and the EXGENESIS Integrated Project (LSHM-CT-2004-005272) from the European Commission. N. Baeyens was supported by a fellowship from the Fund for training in Research in Industry and Agriculture. D. Vertommen is Collaborateur Logistique of the Fonds National de la Recherche Scientifique.
No conflicts of interest, financial or otherwise, are declared by the authors.
The authors thank Greet Vandenberg and Steve Calberson for excellent technical assistance.
- Copyright © 2010 the American Physiological Society