We have been searching for a mechanism to induce smooth muscle contraction that is not associated with phosphorylation of the regulatory light chain (RLC) of smooth muscle myosin (Nakamura A, Xie C, Zhang Y, Gao Y, Wang HH, Ye LH, Kishi H, Okagaki T, Yoshiyama S, Hayakawa K, Ishikawa R, Kohama K. Biochem Biophys Res Commun 369: 135–143, 2008). In this article, we report that arachidonic acid (AA) stimulates ATPase activity of unphosphorylated smooth muscle myosin with maximal stimulation (Rmax) of 6.84 ± 0.51 relative to stimulation by the vehicle and with a half-maximal effective concentration (EC50) of 50.3 ± 4.2 μM. In the presence of actin, Rmax was 1.72 ± 0.08 and EC50 was 26.3 ± 2.3 μM. Our experiments with eicosanoids consisting of the AA cascade suggested that they neither stimulated nor inhibited the activity. Under conditions that did not allow RLC to be phosphorylated, AA stimulated contraction of smooth muscle tissue with an Rmax of 1.45 ± 0.07 and an EC50 of 27.0 ± 4.4 μM. In addition to the ATPase activities of the myosin, AA stimulated those of heavy meromyosin, subfragment 1 (S1), S1 from which the RLC was removed, and a recombinant heavy chain consisting of the myosin head. The stimulatory effects of AA on these preparations were about twofold. The site of AA action was indicated to be the step-releasing inorganic phosphate (Pi) from the reaction intermediate of the myosin-ADP-Pi complex. The enhancement of Pi release by AA was supported by computer simulation indicating that AA docked in the actin-binding cleft of the myosin motor domain. The stimulatory effect of AA was detectable with both unphosphorylated myosin and the myosin in which RLC was fully phosphorylated. The AA effect on both myosin forms was suggested to cause excess contraction such as vasospasm.
- skinned fiber
smooth muscle myosin can be purified with an unphosphorylated regulatory light chain (RLC) of 20 kDa. The myosin is activated when the RLC is phosphorylated by myosin light chain kinase (MLCK) in the presence of Ca2+ and calmodulin (Ca/CaM) (53). Dephosphorylation of the RLC by myosin light chain phosphatase (MLCP) makes myosin inactive (14). Myosin with unphosphorylated RLC exhibits a folded monomeric conformation (41, 46, 54). The fold of the tail part of the heavy chain is commonly observed in the RLC region, and the tail of the heavy chain is extended when the RLC is phosphorylated (6). The phosphorylated myosin assembles into the myosin filaments that are commonly observed in smooth muscle, including vascular smooth muscle (VSM) cells. Therefore, the RLC is thought to play a key role in regulation of structure and function of the myosin (54).
Arachidonic acid (AA) is one of the bioactive lipids and is produced in eukaryotic cells mainly through the catalytic action of phospholipase A2, although a few other mechanisms are known, and it is then further processed by cyclooxygenase (COX), lipoxygenase, and P-450 enzymes to yield metabolites. They are secreted outside of the cells and bind to their respective receptors of the target cells (37, 49). In the cytoplasm of the cells, AA itself can stimulate a few enzymes, including protein kinase C (31) and Rho-kinase (2, 9, 10), and can inhibit MLCP (12). This indicates that AA can act as an intracellular messenger that works to increase the level of RLC phosphorylation. Therefore, the mechanism underlying the role of AA as an agonist (44) or intercellular messenger (50) has been ascribed to RLC phosphorylation.
Despite a flood of examples of the RLC phosphorylation mechanisms, a few pharmacological agents contract intact smooth muscle without any signs of RLC phosphorylation (22). Therefore, we have been attempting for many years to find an intracellular mechanism to activate ATPase of smooth muscle myosin directly (11, 22, 35). In the case of the demembranated smooth muscle tissue, contraction can be induced by incubating it in a high concentration of Mg2+ solution including ATP (45). According to Moreland and Moreland (32), when CTP was used instead of ATP, it contracted with little sign of RLC phosphorylation.
Blebbistatin was found as an agent to inhibit cytokinesis (52) and was characterized as an inhibitor of the ATPase activities of a few myosin isoforms by Sellers and colleagues (25, 27). The inhibitory mechanism by the blebbistatin does not follow the mechanism of the RLC phosphorylation mentioned above. They proposed that it docks in the actin-binding cleft of the myosin motor domain to inhibit the release of inorganic phosphate (Pi) that is produced by the hydrolysis of ATP (25), inhibiting the myosin ATPase activity (57). Blebbistatin not only inhibits the ATPase activity but also relaxes VSM cells in tissue (8) and in culture (20). The inhibition and relaxation did not associate with the changes in RLC phosphorylation.
The effect of AA was reported to contract VSM tissues when added to the intact tissues (4). The contractile activity is interpreted that the metabolites of AA, such as prostaglandins, induce contraction through receptors specific to the metabolites (4, 44). The mechanism underlying the inhibitory effect of blebbistatin creates a new possibility, which is that AA induces contraction by binding myosin and stimulating myosin ATPase activity without requiring RLC phosphorylation. Because the AA metabolites exert little effect on the ATPase activity, we attempted to prove the possibility of a new pathway for AA to induce contraction.
In this study, we analyzed the stimulatory effect of AA on the ATPase activity of purified smooth muscle myosin and the contraction of demembranated smooth muscle tissue. The analysis indicated that the effect is not associated with RLC phosphorylation and suggested that AA stimulates the step of Pi release of the ATPase activity. Docking molecular simulation (33) showed that the binding site of AA is located in the motor domain and, on closer view, in the actin-binding cleft that neighbors the docking site of blebbistatin. In addition, we have presented evidence that AA can stimulate the contraction of smooth muscle while the RLC remains unphosphorylated.
MATERIALS AND METHODS
AA, ML-9, ATP, adenosine 5′-(γ-thio)triphosphate (ATPγS), cytidine 5′-triphosphate (CTP), β-escin, calcium ionophore A-23187, 7-methylguanosine (m7Guo), and diisopropylfluorophosphate (DFP) were purchased from Sigma-Aldrich Chemical (St. Louis, MO). Dithiothreitol (DTT), CaM, α-chymotrypsin, V8 protease, calyculin A, and 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride (ABSF) were purchased from Wako Pure Chemical Industries (Osaka, Japan). 2′-(or -3′)-o-(N-methylanthraniloyl) adenosine 5′-triphosphate (mant-ATP) and 2′-(or -3′)-o-(N-methylanthraniloyl) adenosine 5′-diphosphate (mant-ADP) were purchased from Invitrogen (Carlsbad, CA). Y-27632 was a generous donation from Yoshitomi Pharmaceutical Industries (Osaka, Japan). AA derivatives whose double bonds were cleaved, i.e., (±)5,6-dihydroxy-8Z,11Z,14Z-eicosatrienoic acid [(±)5,6-DiHETrE], (±)8,9-dihydroxy-5Z,11Z,14Z-eicosatrienoic acid [(±)8,9-DiHETrE], (±)11,12-dihydroxy-5Z,8Z,14Z-eicosatrienoic acid [(±)11,12-DiHETrE], and (±)14,15-dihydroxy-5Z,8Z,11Z-eicosatrienoic acid [(±)14,15-DiHETrE], and eicosanoids that were located in downstream of AA cascades described in the Supplemental Data, were purchased from Cayman Chemical (Ann Arbor, MI). (Supplemental data for this article is available online at the American Journal of Physiology-Heart and Circulatory Physiology website.) Other chemicals were of the highest analytical grade available.
Myosin was purified with unphosphorylated RLC from the smooth muscle of chicken gizzard based on the methods of Ebashi (7). When phosphorylated myosin was used, the RLC of the myosin was thiophosphorylated by 0.1 μM MLCK in 60 mM KCl, 5 mM MgCl2, 1 mM DTT, 0.5 mM ATPγS, 20 mM Tris·HCl (pH 7.5), 0.3 μM CaM, and 0.1 mM CaCl2 for 25 min at 25°C (39). MLCK and CaM were removed by precipitation of the phosphorylated myosin after dialysis against 15 mM MgCl2, 1 mM EGTA, 1 mM DTT, and 20 mM Tris·HCl (pH 7.5). We used the phosphorylated myosin after resolving it in 0.6 M KCl, 5 mM MgCl2, 1 mM DTT, and 20 mM Tris·HCl (pH 7.5). The phosphorylation of RLC was confirmed by urea-glycerol polyacrylamide gel electrophoresis (PAGE) (43). MLCK was purified from chicken gizzard smooth muscle by using the method of Adelstein and Klee (1) with slight modifications (15). A cDNA for CaM, which was cloned in our laboratory from chicken gizzard (GenBank accession no. M36167), was expressed in Escherichia coli (BL21) and purified using the method of Matsuura et al. (30), except that we used a butyl-Tyopal column. In some experiments we used the recombinant CaM instead of the bovine brain CaM (Wako). Actin was purified from the acetone powder of chicken skeletal muscle by using the method of Spudich and Watt (51) with slight modifications (23) and was used as actin after polymerization. Heavy meromyosin (HMM) and subfragment 1 (S1) were produced by digesting myosin with V8 protease, followed by termination with DFP (17). They were purified from the digests via Superose 6 100/300 GL column chromatography (GE Healthcare, UK). For the RLC-deficient S1, we digested the RLC of myosin with α-chymotrypsin under conditions to produce HMM, from which the RLC was digested (42). After termination by ABSF, the digest was further subjected to digestion by V8 protease, followed by termination with DFP. The myosin digested by both proteases was subjected to Superose 6 chromatography and used as the RLC-deficient S1. HMM and S1 in the phosphorylated form were prepared from the phosphorylated myosin by proteolysis, followed by Superose 6 chromatography as mentioned above. The purities of the proteins were routinely monitored by sodium dodecyl sulfate (SDS)-PAGE (26) with slight modifications (18).
Expression and purification of S1 heavy chain consisting of smooth muscle myosin heads.
The cDNA sequence coding Met-1 to Glu-729 was transferred from smooth muscle myosin heavy chain clone (Ref. 60; GenBank accession no. NM205274) into a baculovirus transfer plasmid vector containing the His tag sequence at its NH2 terminus and the myc tag sequence at its COOH terminus (pFastBacHTa; Invitrogen) (38). The recombinant virus was produced from the above plasmid according to the manufacturer's instructions for the Bac-To-Bac baculovirus expression system (Invitrogen). To express the recombinant myosin S1 heavy chain, we infected Sf-9 cells with the virus at a multiplicity of infection of 1. The infected cells were grown for 24 h at 27°C in Grace's insect cell culture medium (Invitrogen) supplemented with 10% fetal calf serum and antibiotic and were then lysed by pipetting in the I-PER insect cell protein extraction reagent (Pierce Chemical, Rockford, IL) containing the complete Mini EDTA-free protease inhibitor cocktail (Roche Applied Science, Indianapolis, IN). The unlysed cells and its debris were removed by centrifugation for 30 min at 100,000 g. The supernatant was loaded on the Sepharose 6 Fast Flow (GE Healthcare) at 4°C. The column was washed with a 10-fold volume of buffer consisting of 500 mM NaCl, 5 mM imidazole, 10% glycerol, 1% Nonidet P-40, and 50 mM Tris·HCl (pH 8.0). The recombinant S1 heavy chain was eluted with buffer consisting of 150 mM NaCl, 500 mM imidazole, 10% glycerol, and 50 mM Tris·HCl (pH 8.0). The buffer used for the elution was exchanged with buffer consisting of 600 mM KCl, 1 mM DTT, and 20 mM Tris·HCl (pH 7.5) using the Amicon Ultra centrifugal filter unit (Millipore, Billerica, MA).
ATPase and CTPase activity assays.
The ATPase and CTPase activities of myosin were assayed in the conventional manner at 0.25 μM for the unphosphorylated form or at 0.05 μM for the phosphorylated form of myosin, at 0.1 μM for both forms of HMM, at 0.1 μM for S1, at 0.1 μM for RLC-deficient S1, and at 0.1 μM for S1 heavy chain. In short, 0.5 mM ATP or CTP was hydrolyzed by myosins and their fragments for 10 min at 25°C in a solution containing 60 mM KCl, 5 mM MgCl2, 0.2 mM EGTA, 1 mM DTT, and 20 mM Tris·HCl (pH 7.5) in the presence of dimethyl sulfoxide (DMSO) as the vehicle or in various concentrations of AA dissolved in DMSO. The hydrolysis was terminated by adding 300 mM perchloric acid in the final concentration, followed by the colorimetric method using malachite green (21). The specific activities per head per second are expressed (s−1). Rmax was determined from the maximal activity of the relationship between AA concentration and the activities. EC50 was the concentration of AA that caused half-maximal stimulation. In the case of the tension measurement, Rmax and EC50 were determined by the relationship between AA concentration and tension in a way similar to the ATPase activity measurement. We confirmed the specific activities (s−1) by using the Enzchck phosphate assay kit (Invitrogen) to monitor time-dependent Pi release. This method was also used to measure the detection of RLC phosphorylation after detection of ATPase activities.
Single turnover experiments.
Mant-ATP at 0.25 μM was mixed with 0.5 μM myosin in the unphosphorylated form for mant-ATP fluorescence measurements in buffer consisting of 60 mM KCl, 5 mM MgCl2, 0.2 mM EGTA, 1 mM DTT, and 20 mM Tris·HCl (pH 7.5) in the vehicle or various concentrations of AA (61). Changes in the fluorescence due to the binding to and release from the myosin were detected as the energy transfer from tryptophan excited at 290 nm, and the emitted light at 440 nm was recorded by a fluorometer (F-4500; Hitachi, Tokyo, Japan). Data sets were fitted to the double-exponential equations with a floating end point using WinCurveFit (Kevin Raner Software, Victoria, Australia). The reason why we fitted with double exponentials is explained in the Supplemental Data.
Release of Pi from S1 and release of ADP from S1.
To detect Pi release from the S1-ADP-Pi complex (3), we mixed 1 μM ATP with 2 μM S1 in the unphosphorylated form, 20 μM m7Guo, and purine nucleoside phosphorylase (PNP) in the above solution. The mixture was excited at 250 nm, and the emission at 450 nm was recorded by a fluorometer (F-4500). Fitting of the single experimental traces was performed using WinCurveFit software. To detect ADP release from the S1-ADP complex, S1 solution containing 1 μM S1, 25 μM ADP, 60 mM KCl, 5 mM MgCl2, 0.2 mM EGTA, 1 mM DTT, and 20 mM Tris·HCl (pH 7.5) in DMSO or 20 μM AA was mixed rapidly with solution containing 500 μM ATP, 60 mM KCl, 5 mM MgCl2, 0.2 mM EGTA, 1 mM DTT, and 20 mM Tris·HCl (pH 7.5). The increase in tryptophan fluorescence was recorded using the model 05-109 stopped-flow instrument (Applied Photophysics, Leatherhead, UK).
Measurement of isometric tension.
All experimental procedures conformed to the “Guidelines for Proper Conduct of Animal Experiments” approved by the Science Council of Japan and were carried out under the “Rules and Regulation of the Animal Studies Committee of Tokyo Medical University” and “The Principles of Animal Care and Experimental Committee of Gunma University.” In addition, the institutional Animal Care and Use Committee of Tokyo Medical University and Gunma University approved all procedures involving animals.
Male Hartley guinea pigs (250–350 g) were euthanized under deep anesthesia with diethyl ether, and then their taenia caeci were removed. A small smooth muscle strip (3.0–3.5 mm long, 0.15–0.25 mm wide) was connected to a force transducer (BG-10; Kulite Semiconductor Products, Leonia, NJ) in a bubble chamber (16, 58), followed by skinning with β-escin and A-23187 as described previously (58). The tension measurement was described in detail previously (58). In short, a skinned preparation was stretched in a relaxing solution consisting of 61.6 mM K(methanesulfonate, Ms), 3.1 mM Mg(Ms)2 to give 0.85 mM Mg2+, 1.3 mM Na2ATP to give 1.0 mM Mg·ATP, 20 mM creatine phosphate, and 10 mM EGTA with pH 7.0. After the tension reached a steady level at about 10 μN, the preparation was immersed in Ca2+ contraction buffer consisting of 41.6 mM K(Ms), 2.5 mM Mg(Ms)2 to give 0.85 mM Mg2+, 1.3 mM Na2ATP to give 1.0 mM Mg·ATP, 20 mM creatine phosphate, 1 μM CaM, and 10 mM Ca-EGTA to give pCa 4.3 with pH 7.0 to induce Ca2+ contraction. After the strip was relaxed in the relaxing solution, the preparation was incubated with a solution in which ATP was replaced by CTP and various concentrations of AA existed. Mg2+ contraction was induced by immersing the strip in the following Mg2+ contraction buffers. The 10 mM Mg2+ contraction buffer consisted of 39.4 mM K(Ms), 21.0 mM Mg(Ms)2, 1.0 mM Na2CTP, 20 mM creatine phosphate, and 10 mM EGTA with pH 7.0, containing the vehicle or various concentrations of AA. The 30 mM Mg2+ contraction buffer was prepared by changing K(Ms) and Mg(Ms)2 to 34.4 and 36.8 mM, respectively, whereas other components were unchanged. We used CTP instead of ATP because CTP is a poor substrate for phosphorylating RLC (5). In all the above solutions, the ionic strength was kept at 200 mM by addition of K(Ms) and pH was adjusted with 20 mM PIPES and KOH to 7.0 according to the methods of Horiuti (16). The tension in the Mg2+ contraction buffer was expressed relative to that in the Ca2+ contraction buffer, and the contractions and relaxations were elicited at 25 ± 1.0°C throughout the study.
Detection of RLC phosphorylation in smooth muscle tissue.
After contraction with the Mg2+ contraction buffer containing various concentrations of AA, the skinned smooth muscle strip was quickly frozen by immersion in an ice-cold acetone containing 10% trichloroacetic acid (TCA) and 10 mM DTT for 10 min, followed by washout of TCA with acetone containing 10 mM DTT. The muscle strip was then incubated in a urea sample buffer consisting of 20 mM Trizma base, 22 mM glysine, 10 mM DTT, 8 M urea, and 0.1% bromophenol blue for 8 h at 4°C. The extracts by incubation were subjected to glycerol-PAGE coupled with Western blotting using nitrocellulose membrane (Hybond-ECL; GE Healthcare) (20, 43). RLC was detected with anti-RLC antibody made in our laboratory and was visualized with horseradish peroxidase-labeled anti-rabbit IgG (NIF 824; GE Healthcare) and enhanced chemiluminescence using the ECL Western blotting detection system (GE Healthcare).
Electron microscopy of myosin.
Chicken gizzard myosin in the unphosphorylated form was dissolved in a buffer of 150 mM KCl, 2 mM ATP, 3.5 mM MgCl2, and 20 mM Tris·HCl (pH 7.5) in the presence of the vehicle or 20 μM AA. Samples were absorbed to freshly cleaved mica and then subjected to rotary shadowing (28). The shadowed images were observed with an electron microscope (JEM-1010; JEOL Electron, Tokyo, Japan).
Results are means ± SE. Statistical hypotheses on the differences between means were tested with Student's t-test for paired samples unless noted otherwise. The null hypotheses were rejected when P was <0.05. The 95% confidence limit was calculated and the statistical analysis performed using Excel (Microsoft Office 2003).
Effects of AA on myosin ATPase activities.
Myosin ATPase activity was measured in the presence of various concentrations of AA. When myosin was untreated by MLCK, and hence in an unphosphorylated form, its ATPase activity in the presence of a vehicle was low, i.e., 0.00892 ± 0.00136 s−1 (n = 10). The activity was elevated with an increase in the AA concentration (Fig. 1A). The maximum stimulation relative to that of the vehicle (Rmax) was 6.84 ± 0.51 (n = 10) when the extent of stimulation was expressed relative to the activity in the vehicle, and the EC50 was 50.3 ± 4.2 μM (n = 10) (Table 1). The stimulating effect was also observed with the myosin that had been phosphorylated by MLCK in the presence of Ca/CaM. The activity in the vehicle was 0.0678 ± 0.0065 s−1 (n = 6), as indicated in Fig. 1A, confirming elevation of the activity by the RLC phosphorylation (54). AA increased the activity by an Rmax of 2.15 ± 0.17 (n = 6), with an EC50 of 8.20 ± 0.84 μM (n = 6) (Table 1).
When unphosphorylated myosin is used, a question may be raised whether the stimulation will be caused by the increase in the RLC phosphorylation by the enzymes that may be contained as ingredients in the myosin preparation (36). We subjected the myosin after ATPase activity measurement by Enzchck methods to urea-glycerol PAGE to detect any change in the phosphorylation of RLC to answer this question. However, we never detected any changes in RLC phosphorylation (data not shown). We also examined the effect of AA in the presence of the MLCP inhibitor, calyculin A; the Rho-kinase inhibitor, Y-27632; and the MLCK inhibitor, ML-9. The ATPase activities were within the 95% confidence range as indicated by the dotted lines in Fig. 1A. These data indicate that stimulation by AA in the present study was not caused by an increase in RLC phosphorylation. It also should be noted that AA was able to stimulate ATPase activity of the unphosphorylated myosin when it was activated by actin. The stimulation was definite, but the extent as expressed by Rmax was 1.72 ± 0.08 (n = 4), a value that is lower than the Rmax of ATPase activity of unphosphorylated myosin without actin. However, AA stimulated at a comparable EC50 of 26.3 ± 2.3 μM (n = 4).
The stimulating effect of AA was further examined with the proteolytic products of myosin: HMM and S1. The ATPase activities of HMM with unphosphorylated RLC and HMM with phosphorylated RLC in the presence of a vehicle were 0.0248 ± 0.0029 s−1 (n = 7) and 0.0636 ± 0.0144 s−1 (n = 7), respectively. Rmax was ∼2, and EC50 was ∼9 μM, regardless of whether RLC was phosphorylated or not, as presented in Table 1. The ATPase activities of S1 are unchanged irrespective of RLC phosphorylation (47). We confirmed that they were 0.0484 ± 0.0098 s−1 (n = 8) for S1 with unphosphorylated RLC and 0.0516 ± 0.0127 s−1 (n = 8) for S1 with phosphorylated RLC. Moreover, the extent of stimulation of ATPase activity by AA of S1 was similar, i.e., an Rmax of ∼2 with an EC50 of ∼10 μM for S1's with both unphosphorylated and phosphorylated RLC (Table 1). The effect of AA on the RLC-deficient S1 was also studied. The ATPase activity in the absence of AA was 0.0475 ± 0.0045 s−1 (n = 6), a value similar to the ATPase activity of intact S1 (Table 1). The extent of stimulation of ATPase activity by AA was an Rmax of 2.06 ± 0.19 (n = 6) with an EC50 of 9.70 ± 1.49 μM (n = 6), indicating a similar stimulating effect.
The S1 heavy chain consisting of myosin heads without any myosin light chains was obtained as a recombinant protein as described in materials and methods. As shown in Table 1 as S1 heavy chain, it was active in hydrolyzing ATP at the rate of 0.0101 ± 0.0028 s−1. AA stimulated the S1 heavy chain with an Rmax of 2.50 ± 0.33 (n = 5) and an EC50 of 83.3 ± 12.3 μM (n = 6). Because the S1 heavy chain was associated with neither the essential light chain nor RLC, the site of AA action was demonstrated to be in the S1 heavy chain. As we discuss later, this is in good conformity with the proposal that AA docked in the cleft actin binding of S1 heavy chain.
Does AA stimulate the ATPase activity of the myosin in which RLC was phosphorylated? Our experiments shown in Table 1 answer this question. The activity of myosin with phosphorylated RLC was 0.0678 ± 0.0065 s−1 (n = 6). AA stimulated the activity of the myosin with phosphorylated RLC with an Rmax of 2.15 ± 0.169 (n = 6) and an EC50 of 8.20 ± 0.84 μM (n = 6). The extent of stimulation was less than that observed with myosin with unphosphorylated RLC but comparable to that observed with HMM with unphosphorylated RLC and S1 with unphosphorylated RLC.
Control experiments with AA derivatives: their effects on myosin ATPase activity.
AA has four double bonds at the 5–6, 8–9, 11–12, and 14–15 positions. We obtained (±)5,6-DiHETrE, (±)8,9-DiHETrE, (±)11,12-DiHETrE, and (±)14,15-DiHETrE, for which the double bonds were cleaved at the respective sites, to examine the role of respective double bonds in stimulating myosin activity. Under the conditions where 80 μM AA stimulated the ATPase activity of unphosphorylated myosin with an Rmax of 5.06 ± 0.69 (n = 5), the extents of the stimulation by the derivatives were in the range of 1.22 ± 0.44–1.58 ± 0.10 as expressed relative to DMSO vehicle and irrespective of the double bond positions (Fig. 2). This failure of stimulation indicates that all four double bonds are required for AA to stimulate the ATPase activity. The experiments of the above derivatives were also carried out in the presence of various concentrations of DMSO vehicle up to 160 μM to confirm the failure in stimulating the ATPase activity as shown in the Supplemental Table.
To test the effects on the metabolites of AA that were located in the downstream of the signal transductory pathways involving COX (49), we measured the effect of eicosanoids, including prostaglandins (prostaglandin I2, prostaglandin D2, and prostaglandin B2), thromboxane (thromboxane B2), leukotrienes (leukotriene B2 and leukotriene C4), and 20-hydroxyeicosatetraenoic acid, on the ATPase activity of unphosphorylated myosin. We observed neither definite stimulatory effects nor definite inhibitory effects on the myosin ATPase activity (Supplemental Table). In detail, the slight inhibition (0.62–0.99 compared with DMSO vehicle) and slight stimulation (1.03–1.93 compared with DMSO vehicle) when examined at 0.16 μM to 10 mM eicosanoids (Supplemental Table).
We also examined the effect of the fatty acids with C-18 on the ATPase activity of unphosphorylated myosin. We chose stearic acid because it has no double bond. Furthermore, we chose oleic acid because it has one double bond. We failed (stearic acid, 0.97–1.03; oleic acid, 0.99–1.02) to detect their stimulatory effect on the ATPase activity (Supplemental Table).
Effects of AA on contraction elicited under conditions where myosin remains unphosphorylated.
The skinned preparation of smooth muscle tissue developed tension when the Ca2+ and ATP in the contraction buffer were replaced by Mg2+ and CTP (32, 58), respectively. Figure 3A shows the typical records of the time courses of the contraction of skinned smooth muscle fiber in the presence of 10 or 30 mM Mg2+. When the contraction was induced in 30 mM Mg2+, the effect of AA was detectable soon after the addition of AA, but the tension development was quickly diminished. On the other hand, the tension developed in 10 mM Mg2+ showed a plateau for a much longer period. Therefore, the conditions under 10 mM Mg2+ enabled us to detect the effect of AA on the tension development. We chose to measure the contraction in the presence of 10 mM Mg2+, although the effect of AA was not detectable during the initial 10 min as shown in Fig. 3B (see discussion, for the delay of the effect). Figure 3C shows the contraction 15 min after the addition of the various concentrations of AA. AA stimulated the contraction with an Rmax of 1.45 ± 0.07 (n = 10) and an EC50 of 27.0 ± 4.4 μM (n = 10). The increase in the contractile force is in conformity with Rmax and EC50 values for the actin-activated ATPase activity of unphosphorylated myosin (Table 1) in terms of AA concentration. This view was also supported by the stimulatory effect of AA on CTPase activity of unphosphorylated myosin as shown in Fig. 1B, which shows the relationship between CTPase activity and the concentration of AA. The CTPase activity in the absence of AA was 0.0143 ± 0.0010 s−1 (n = 5), which was comparable to that of ATPase activity (Table 1), and was increased with increasing AA concentration [Rmax = 6.74 ± 0.82 (n = 5) and EC50 = 47.6 ± 9.7 μM (n = 5)]. Thus CTPase activity of the myosin was able to be activated by AA in terms of contraction and supports the idea that contraction in Fig. 3B was brought about by the CTPase activity of unphosphorylated myosin.
Single turnover of myosin ATPase.
To rule out the possibility that a small population of highly active molecules might influence the measurement (Fig. 4A), we conducted single turnover experiments with mant-ATP in the presence of various concentrations of AA. The addition of mant-ATP to the myosin solution caused an abrupt increase in the fluorescence due to mant-ATP binding to myosin. We added a 100-fold excess of ATP at the peak of this transient binding to prevent additional mant-ATP binding and measured the subsequent decay of mant-nucleotide fluorescence that corresponded to the release of the hydrolysis products of mant-ADP and Pi. Time courses of the fluorescence decay were fitted to a double-exponential curve (for the rationale, see Supplemental Data). As shown in Fig. 4B, the fast and slow components had the rate constants of ∼2.8 × 10−2 and 1.1 × 10−3 s−1, respectively, which were independent of the AA concentrations as shown in the inset. The slow component could be ascribed to the products released by the myosin. Increasing the AA concentration to 80 μM resulted in the loss of the slow component, and the fast component was almost dominated, supporting the idea of a major population of stimulated myosin in the presence of more than 80 μM of AA. The EC50 for the component was 41.5 ± 7.3 μM (n = 4), which conformed well with that of the EC50 for the ATP-dependent stimulation of the ATPase activity of myosin with unphosphorylated RLC shown in Fig. 1A and Table 1.
Detection of Pi release from myosin heads.
According to the knowledge of the hydrolysis of ATP using skeletal muscle myosin, the release of phosphate (Pi) that is hydrolyzed by myosin ATPase from the reaction intermediate complex with ADP and Pi is the rate-limiting step. ATPase hydrolysis of smooth muscle myosin is generally thought of as subject to the scheme of Ref. 29. We measured the rate of Pi release under single turnover conditions from the intermediate complex of myosin with ADP and Pi using m7Guo, which detects the release of Pi by PNP. A high concentration of myosin, i.e., at least 1 μM, is required due to the detection limit of m7Guo. Because of the high viscosity of the myosin solution causing difficulties in the assay, we used S1 with unphosphorylated RLC at 2 μM instead of the myosin. The time course of the change in m7Guo fluorescence in the presence of 20 μM AA was much faster than that of the change in vehicle as illustrated in Fig. 5A. A curve fitting indicated that the rate constants of Pi release in the vehicle and in AA were 0.0174 ± 0.0 052 s−1 (n = 6) and 0.0339 ± 0.0056 s−1 (n = 6), respectively. Thus AA at 20 μM enhanced the rate of Pi release 1.95-fold, confirming the extent of stimulation measured by the S1 ATPase activity (Table 1). We also measured the release of ADP from the S1-ADP complex. Compared with the step of Pi release, the step is fast (29) enough to use the stopped-flow technique, which was based on the increase in the tryptophan fluorescence. We estimated the effect of AA on ADP release by the chasing ADP from S1-ADP complex by using excess ATP (59). When the release of ADP in vehicle was compared with the release of ADP in 20 μM AA, we could not detect any differences, indicating the step of the releasing ADP was not the site of AA action (Supplemental Data).
Structural changes of myosin.
It is well known that smooth muscle myosin with unphosphorylated RLC has a folded tail with heads in the down direction when Mg-ATP exists in low ionic strength (9, 41, 54) and that the myosin with phosphorylated RLC has an extended tail with heads in the up direction. We confirmed these conformations of myosin in the vehicle (Fig. 6A for unphosphorylated myosin and Fig. 6B for phosphorylated myosin). However, the extended tails were dominated in the presence of 20 μM AA, whereas the other conditions remained the same (Fig. 6C). As a result, we observed myosin filaments that were assembled with the extended tails (Fig. 6D) similar to that for myosin with phosphorylated RLC. A quantitative evaluation with over 200 molecules including those incorporated into filaments indicated that the myosin with extended tails was increased from 5.5% (Fig. 6E) to 58.5% (Fig. 6F) by 20 μM AA. To confirm the assembly of the myosin into filament (Fig. 6F) on a biochemical basis, we subjected the AA-treated myosin to centrifugation. The increase of the myosin in the precipitation after the centrifugation was confirmed as shown in the Supplemental Data.
The present study shows that AA acts on smooth muscle myosin to stimulate its ATPase activities whether or not its RLC is phosphorylated (Table 1). What we would like to stress is that a new pathway of the signal transduction presented in this report leads to induced contraction without any signs of RLC phosphorylation (Fig. 3). The site of AA action was proved to be in the myosin motor domain, i.e., heavy chain consisting of myosin heads. As discussed later, the docking of AA in the actin-binding cleft of myosin motor domain may enhance the activity of hydrolyzing ATP to form the reaction product of ADP + Pi as an intermediate. The step of releasing Pi was suggested to be enhanced (Fig. 5). We also report that the stimulatory effect of AA could be detected in myosin whose ATPase activity was fully activated by the RLC phosphorylation mechanism. Such an additional effect of AA supports the proposal that smooth muscle myosin has a latent capacity to interact with actin, inducing an excess contraction (55).
Blebbistatin is a small molecule to inhibit the ATPase activity of myosin, including smooth muscle myosin (27). The site of action is suggested by the docking pattern that blebbistatin binds to the myosin motor domain by closing the back door so that Pi cannot be released from the myosin (25). Although the effect of AA is quite the opposite from that of blebbistatin, we thought that the search for the putative binding site of AA in a way similar to that of blebbistatin should enable us to discuss the site. The search of AA on the myosin head was performed using the program AutoDock 3.0 (33) with the AutoDockTools 1.5.0 (http://mgtools.scripps.edu) graphical interface. Atomic structures for head domains of ADP·BeFx-bound smooth muscle myosin, which is shown in the Protein Data Bank (PDB) entry 1br4, were applied as target molecules during the calculations. For head domains and AA, 100 docking runs were evaluated. Representative groups, which are the binding patterns, were collected and ordered based on the calculated binding free energy and the occurrence of the complexes. To obtain the final binding model of AA, 100 conformations of the ligand were generated using the grid parameter calculated for the actin-binding cleft and then crystallized using the AutoDockTools. The computation suggested that AA docks to the actin-binding cleft as expected (Fig. 7). The docking in the cleft is quite reasonable, because it is located near but distinct from the blebbistatin docking site (single arrow for AA, double arrows for blebbistatin). The cleft located in the myosin head consisting of heavy chain explains the idea that AA stimulation is detectable irrespective of RLC phosphorylation.
The result of the docking simulation is in conformity with the biochemical data shown in Table 1. The ATPase activity of S1 heavy chain, without light chains, i.e., the myosin motor domain, could be activated by AA with comparable Rmax and EC50 values. The S1 heavy chain was expressed in Sf-9 cells and purified by the His-tag affinity column. The expression and following purification procedures yielded questions whether the S1 heavy chain was associated with myosin light chain(s) of Sf-9 myosin or CaM of Sf-9 and whether AA activated through the associated protein(s). However, when the concentrate after purification was subjected to SDS-PAGE, we observed only a sole band of ∼100 kDa, ruling out the above concerns.
The activity of smooth muscle myosin has been known for many years to be closely related to its structure (54). Examining myosin with unphosphorylated RLC shows that its tail exhibits a folded form, with the heads directing downward to show the head-down structure. When RLC is phosphorylated, the heads extend upward against the tail of myosin and the tail becomes extended, forming myosin filaments (54). Concerning the tail structure, we found that AA extends the tail of myosin with unphosphorylated RLC (Fig. 6C). As a result, AA is able to assemble myosin into filaments without any signs of RLC phosphorylation (Fig. 6D). A few proteins are known to induce myosin assembly without RLC phosphorylation (40, 48). To our knowledge, AA is the first example of this induction among bioactive lipids. Concerning the head structure, the myosin heads with phosphorylated RLC stand straight against the tail to confirm the head-up structure (Fig. 6B). The myosin heads with unphosphorylated RLC in the presence of AA showed neither the head-down nor the head-up structure. What we found with isolated myosin molecule was that both heads were arranged in a line without showing any angle between the head and tail. Such an intermediate nature of the head conformation may explain why the ATPase activity stimulated by AA was definitely lower than the activity stimulated by RLC phosphorylation (Fig. 1).
The present report strongly suggests that the myosin tail appears to have an active role in mediating the stimulatory effect of AA. The extent of the activation of myosin was about sevenfold, but the stimulation of HMM and S1, each of which lost its tail, was only about twofold (Table 1), indicating the importance of the tail for unphosphorylated myosin being stimulated by AA. In other words, a low level of stimulation may be related to the absence of the tails. It has been known for many years that smooth muscle myosin assembles to thick filaments through its tail in the presence of 10–30 mM Mg2+, if not phosphorylated, and that incubation for about 60 min is required for the assembly (7). We speculate that myosin assembly should occur in the skinned smooth muscle in 10 mM Mg2+ during the lag time for AA to stimulate contraction and that the assembly through tail is required for AA to stimulate the contraction. The observation that the effect of AA was detectable in 30 mM Mg2+ soon after addition of AA (Fig. 3A) would suggest that myosin might already be assembled into filament in 30 mM Mg2+ before the addition of AA.
It has been explained that AA acts as an agonist by exerting its action after the metabolism to prostaglandins, for example (44). The metabolites exert their effect on the contractile activity through the respective receptors (49). Although AA for COX is stimulated by an EC50 of 10–20 μM AA (56), its metabolite, such as prostaglandins, exerted no effect on the myosin ATPase activity (results and Supplemental Data). The data, in turn, suggest the possibility that AA is metabolized by COX and that AA cannot stimulate myosin. However, the fatty acid binding protein that stabilizes the fatty acids may prevent AA metabolism (34). The concentration of AA after the agonist stimulation was reported to be 10–100 μM (13), and the serum AA concentration was reported to be at similar levels (24). Because of the hydrophilic nature of AA, AA may freely enter into the cytoplasm. Together, the above data at the cellular AA level should be maintained, stimulating myosin ATPase.
We have reported the stimulatory effect of AA on the ATPase activity of myosin with unphosphorylated RLC. What we also want to discuss is how AA stimulates the ATPase activity of myosin with phosphorylated RLC (Fig. 1A and Table 1). The fluorescently labeled actin filaments slid over a myosin-coated surface in an ATP-dependent way. When an actin filament was attached to an ultracompliant glass microneedle, the force generated by a single myosin head could be monitored. The force increased with the length of the actin filament. By extrapolating the length, we found the smooth muscle myosin developed more force than skeletal muscle myosin (55). We observed that AA enhanced the movement of actin filaments on the surface of phosphorylated myosin (unpublished observation). This means that AA stimulates the additional myosin ATPase activity that has been activated by the RLC phosphorylation. Although we did not conduct the experiments under pathophysiological conditions, we suppose that AA develops the above latent ability for smooth muscle to induce an excess contraction such as vasospasms.
EC50 values for AA to stimulate the myosin ATPase activity and the myosin CTPase activity (Table 1 and Fig. 1) were 50 and 48 μM, respectively. Similar EC50 values were reported for AA to stimulate the enzymic activities of Rho-kinase (2, 9, 10), COX (56), and protein kinase C (50) and to inhibit those of MLCP (12). Over 100 μM AA was required to stimulate contraction of skinned smooth muscle through the MLCP inhibition pathway (12) and through the Rho-kinase stimulation pathway (2), indicating a discrepancy in the concentration between enzymic activity and tension development. However, in the new pathway of the direct binding of AA, the EC50 value for stimulating tension development was 27 μM AA (see results, Effects of AA on contraction elicited under conditions where myosin remains unphosphorylated), a concentration that is explained by stimulating enzymic activity of myosin. We must note, however, that the new pathway was not tested under pathophysiological conditions such as a subarachnoid hemorrhage model, and we expect that the pathway may cause vasospasm after subarachnoid hemorrhage as discussed above.
This work was supported by grants from the Smoking Research Foundation and by Grants-in-Aid for Scientific Research of the Ministry of Education, Culture, Sports, Science and Technology of Japan.
- Copyright © 2010 the American Physiological Society