Myosin Loop 2 Is Involved in the Formation of a Trimeric Complex of Twitchin, Actin, and Myosin*

Molluscan smooth muscles exhibit a low energy cost contraction called catch. Catch is regulated by twitchin phosphorylation and dephosphorylation. Recently, we found that the D2 fragment of twitchin containing the D2 site (Ser-4316) and flanking immunoglobulin motifs (TWD2-S) formed a heterotrimeric complex with myosin and with actin in the region that interacts with myosin loop 2 (Funabara, D., Hamamoto, C., Yamamoto, K., Inoue, A., Ueda, M., Osawa, R., Kanoh, S., Hartshorne, D. J., Suzuki, S., and Watabe, S. (2007) J. Exp. Biol. 210, 4399–4410). Here, we show that TWD2-S interacts directly with myosin loop 2 in a phosphorylation-sensitive manner. A synthesized peptide, CAQNKEAETTGTHKKRKSSA, based on the myosin loop 2 sequence (loop 2 peptide), competitively inhibited the formation of the trimeric complex. Isothermal titration calorimetry showed that TWD2-S binds to the loop 2 peptide with a Ka of (2.44 ± 0.09) × 105 m−1 with two binding sites. The twitchin-binding peptide of actin, AGFAGDDAP, which also inhibited formation of the trimeric complex, bound to TWD2-S with a Ka of (5.83 ± 0.05) × 104 m−1 with two binding sites. The affinity of TWD2-S to actin and myosin was slightly decreased with an increase of pH, but this effect could not account for the marked pH dependence of catch in permeabilized fibers. The complex formation also showed a moderate Ca2+ sensitivity in that in the presence of Ca2+ complex formation was reduced.

Molluscan smooth muscles, such as mussel anterior byssus retractor muscle (ABRM) 2 and adductor muscle, exhibit a low energy cost phase of tension maintenance termed catch. Catch muscle develops active tension following an increase of the intracellular [Ca 2ϩ ] induced by secretion of acetylcholine. Myosin is activated by direct binding of Ca 2ϩ to the regulatory myosin light chain and initiates a relative sliding between thick and thin filaments (1). After a decrease of intracellular [Ca 2ϩ ] to resting levels, the catch state is formed where tension is maintained over long periods of time with little energy consumption (2,3). Catch tension is abolished by secretion of serotonin and an increase of intracellular [cAMP] with the resulting activation of cAMP-dependent protein kinase and phosphorylation of twitchin (4,5). Twitchin phosphorylation is required for relaxation of the muscle from catch. For this cycle to repeat, dephosphorylation of twitchin is necessary (6). Thus, in this scheme, twitchin is a major regulator of the catch state.
Molluscan twitchin is known as a myosin-binding protein belonging to the titin/connectin superfamily. It is a single polypeptide of 530 kDa containing multiple repeats of immunoglobulin (Ig) and fibronectin type 3-like motifs in addition to a single kinase domain homologous to the catalytic domain of myosin light chain kinase of vertebrate smooth muscle (7). There are several possible phosphorylation sites in molluscan twitchin recognized by cAMP-dependent protein kinase, and two, D1 and D2, have been identified. The D1 phosphorylation site (Ser-1075) is in the linker region between the 7th and 8th Ig motifs (numbering from the N terminus). The D2 site (Ser-4316) is in the linker region between the 21st and 22nd Ig motifs. Additional sites are found close to D1, but are thought not to be vital for catch regulation.
The molecular mechanisms underlying development and maintenance of the catch state have been controversial for several years. One theory proposes that catch reflected attached frozen or slowly cycling cross-bridges (8,9). What distinguished the attached cross-bridge from the detached relaxed state is not clear. Also it was suggested that interactions between thick filaments, other than cross-bridges, or between thin and thick filaments are responsible for the catch contraction (10). In either of the latter cases, the cross-bridge (myosin head) was not involved.
Recently we found that a twitchin fragment including the D2 phosphorylation site and its flanking Ig motifs (TWD2-S) interacted with myosin and actin in a phosphorylation-sensitive manner, and it was suggested that this trimeric complex contributed to tension maintenance in catch (11). TWD2-S bound to a region of the actin molecule known also to interact with loop 2 of myosin that is involved in the ATP-driven movement of myosin with actin (12). In the present study, we show that the myosin loop 2 binds to TWD2-S using competitive cosedimentation assays and isothermal titration calorimetry (ITC). These techniques were applied to also study in more detail the interactions of the twitchin-binding peptide of actin (identified in the previous study (11)). In addition, the effects of pH and Ca 2ϩ on the binding of TWD2-S to myosin and actin were investigated.

EXPERIMENTAL PROCEDURES
Protein Preparations-Live specimens of the mussel Mytilusgalloprovincialis were used. Mytilus ABRM was dissected from shells, frozen immediately in liquid nitrogen, and stored at Ϫ80°C until use. ABRM myosin was purified as previously reported (11). F-actin was prepared from acetone powder of chicken fast skeletal muscle (13). Protein concentrations were measured by the Bradford method (Bio-Rad).
Competitive Cosedimentation Assays-All procedures were carried out at 4°C or on ice. A reaction mixture of 10 l contained 20 g/ml myosin (ϳ42 nM), 1 mg/ml actin (ϳ24 M), and 20 g/ml TWD2-S (ϳ9.2 M) in 20 mM MOPS-NaOH, pH 7.4, 30 mM KCl, 4 mM MgCl 2 , 1 mM ATP, 0.1 mM EGTA. The 10-l reaction mixture also contained the actin or myosin loop 2 peptide to provide molar ratios (to TWD2-S) of 1:10, 1:5, 1:2, 1, and 5:1. The reaction mixture thus prepared was kept on ice for 15 min and subjected to low speed centrifugation (5,000 ϫ g) to sediment filamentous myosin. The supernatant and pellet were applied to SDS-PAGE. The in vivo molar ratio of twitchin/ myosin is 1:15 (6), and the higher levels of TWD2-S were used to facilitate its detection on gels.
The effects of Ca 2ϩ on formation of the trimeric complex were monitored by carrying out the cosedimentation assays as above but with 1 mM CaCl 2 and 1 g/ml TWD2-S. Proteins were visualized by silver staining. The concentration of protein in gels was determined by the computer software Image J.
Isothermal Titration Calorimetry-ITC was performed using a VP-ITC calorimeter (MicroCal LLC). Experimental parameters were: total injections, 25 times; cell temperature, 25°C; reference power, 10 mCal/s; initial delay, 600 s; syringe concentration, 400 M; cell concentration, 10 M; stirring speed, 300 rpm. Injection parameters were: volume, 10 l; duration, 20 s; spacing, 300 s; filter period, 2 s. Titration was performed at 25°C by injecting 10 l of 400 M synthetic peptide (actin subdomain I or myosin loop 2 peptide) into the ITC cell containing 1.5 ml of 10 M TWD2-S or S-TWD2-S. All samples were dissolved in 30 mM potassium phosphate buffer, pH 7.0, 4 mM MgCl 2 , and 1 mM ATP. No EGTA was added to the solution as it disturbed ITC measurements. The data thus obtained were corrected for the heat of dilution of the peptide and analyzed by MicroCal Origin 6.0 software.
Solid Phase Binding Assay-The pH dependence of binding of TWD2-S to native actin, myosin loop 2 peptide, and native myosin was determined with a solid phase binding assay (11) using the Protein Detector ELISA kit (KPL). Native actin and myosin used in the cosedimentation assay were utilized for the solid phase binding assay. Actin was dissolved in 20 mM Trismaleate, pH 7.0, containing 50 mM KCl, and myosin in 20 mM MOPS-NaOH, pH 6.8 containing 0.6 M KCl and 0.1 mM dithiothreitol. Wells in a plastic immunoplate were coated with 50 g/ml actin, 50 g/ml myosin, or 50 g/ml myosin loop 2 peptide and then bound with 30 g/ml TWD2-S in 50 mM MOPS, 1 mM EGTA, and 200 mM KCl adjusted to pH 6.7, 7.2, 7.4, and 7.7 after blocking with 1ϫ bovine serum albumin diluent/ blocking solution in the kit. The salt concentration of the TWD2-S solution was referred to by Höpflinger et al. (15). Binding was carried out for 1 h at room temperature. Mouse anti-His tag antibody (GE Healthcare) and anti-mouse IgG antibody conjugated with alkaline phosphatase were used as the primary and secondary antibodies, respectively. Absorbance at 630 nm was measured after maximum color development. Bovine serum albumin was used as a control in these experiments and also was used for standardization of data obtained with actin, myosin, and myosin loop 2.
The effects of Ca 2ϩ on binding of TWD2-S to the myosin loop 2 and actin peptides were examined by the solid phase binding assays. Wells in a plastic immunoplate were coated with 50 g/ml myosin loop 2 peptide or 50 g/ml actin peptide and then treated with 30 g/ml TWD2-S in 1ϫ coating solution in the ELISA kit. Binding/incubation was carried out for 1 h at room temperature, and color detection was performed as described above.

RESULTS
Competitive Cosedimentation Assays-Unphosphorylated TWD2-S was used because the phosphorylated TWD2-S does not form the trimeric complex (11). As shown in Fig. 1A, the pellet of the control sample contained myosin, actin, and TWD2-S, i.e. the trimeric complex. In the presence of low levels of the loop 2 peptide, the complex is dissociated, and TWD2-S is found with both myosin (pellet) and actin (supernatant). At the 1:10 ratio (loop 2 peptide/TWD-2S), the concentration of the loop 2 peptide is about 10ϫ greater than the concentration of myosin heads (ϳ84 nM). As the level of loop 2 peptide is increased, more TWD-2S is found in the supernatant, and at the 1:1 ratio is found only in the supernatant. Thus, at relatively low concentrations of loop 2 peptide (in the presence of ATP and absence of Ca 2ϩ ), the trimeric complex is disrupted.
The effects of the actin peptide on complex formation also were tested (Fig. 1B) using the same molar ratios of peptide to TWD2-S. The myosin pellet in each sample contained TWD2-S but decreasing amounts of actin as the ratio of the actin peptide was increased, in contrast to the loop 2 peptide. At a molar ratio of 5 (actin peptide/TWD2-S), actin was not present in the pellet. As expected, at the higher levels of the actin peptide, TWD2-S maintained its binding to myosin, but binding of TWD2-S to F-actin was decreased.
It is considered that in the catch state the intracellular [Ca 2ϩ ] is at resting level (2), and under these conditions the complex of twitchin, myosin, and actin is thought to be formed. It was reported that unphosphorylated twitchin has decreasing effects on contractions induced at higher [Ca 2ϩ ] (6). Thus, it is possible that the formation of the trimeric complex is Ca 2ϩ -dependent. To address this point, the cosedimentation assays also were carried out in the presence of both ATP and Ca 2ϩ . In the absence of ATP and the presence of Ca 2ϩ (rigor conditions), actin was cosedimented with myosin without TWD2-S (data not shown). In the absence of Ca 2ϩ and the presence of TWD2-S (molar ratio to myosin of 1) part of the actin was cosedimented with myosin ( Fig. 2A). However, in the presence of Ca 2ϩ the amount of actin in the pellet was decreased and was calculated to be 46 Ϯ 12% of that cosedimented in the absence of Ca 2ϩ (n ϭ 3) (p Ͻ 0.05). The binding of TWD2-S to the actin and loop 2 peptides in the presence of Ca 2ϩ slightly decreased as revealed by the solid phase binding assay, supporting the data of the cosedimentation assay (Fig. 2B). Thus, the presence of Ca 2ϩ partly inhibited the formation of the complex. Part of this effect could be due to a reduced affinity of TWD2-S to myosin and actin in the presence of Ca 2ϩ and a resultant decrease in formation of the trimeric complex (TWD2-S was detected in the supernatant fraction in the presence of Ca 2ϩ ; see arrowhead in Fig. 2A).
Isothermal Titration Calorimetry-TWD2-S was titrated by ITC with the loop 2 peptide as shown in Fig. 3A, providing the stoichiometry (N), binding constant (K a ), binding enthalpy (⌬H), and binding entropy (⌬S) in the endothermic reaction. The N value of 1.90 Ϯ 0.02 obtained in a solution containing TWD2-S and the loop 2 peptide indicated that the TWD2-S fragment contained two binding sites for the loop 2 peptide with K a ϭ (2.44 Ϯ 0.09) ϫ 10 5 M Ϫ1 . On the other hand, no signal was detected in the reaction mixture containing S-TWD2-S and the loop 2 peptide, indicating that interaction between TWD2-S and the loop 2 peptide is lost on phosphorylation of the D2 site in the twitchin molecule. The same experiment was performed using the actin peptide, which was shown previously to bind to TWD2-S (11). The thermogram and binding isotherm in the titration of the actin peptide with TWD2-S are shown in  tively, again indicating that TWD2-S contained two binding sites for the actin peptide. Thiophosphorylation of TWD2-S diminished the binding of S-TWD2-S to the actin peptide as in the case with S-TWD2-S and the myosin loop 2 peptide (Fig. 3, A and B, lower panels). Data from ITC together with the above competitive cosedimentation assay clearly demonstrate that both myosin loop 2 and the actin subdomain I regions are involved in the formation of the trimeric complex of myosin, actin, and twitchin.
Effects of pH on Binding of TWD2-S to Myosin, the Loop 2 Peptide, and Actin-It was reported that a change of pH affects catch contraction, and higher pH (in the range of 7.2-7.7 in various studies) (see Ref. 15) in skinned muscle fibers caused relaxation of muscle from catch. Using the solid phase binding assay, it was investigated whether a pH dependence in various interactions could be demonstrated. Myosin, the loop 2 peptide, and actin immobilized in the wells of a microplate were subjected to binding/incubation with TWD2-S at pH 6.7, 7.2, 7.4, and 7.7. TWD2-S interacted with all of the three preparations at each pH value (Fig. 4). The binding of TWD2-S to myosin was maximum at pH 6.7 and decreased to 62, 59, and 59% (p Ͻ 0.01) at pH 7.2, 7.4, and 7.7, respectively (relative to the pH 6.7 value). Similarly, the affinity of TWD2-S to actin was reduced significantly to 76% at pH 7.2, compared with pH 6.7 (p Ͻ 0.05). No significant changes were detected in the binding of TWD2-S to the loop 2 peptide. These results indicate that although alkalization influences some of the binding parameters, it is not a dominant factor, as expected from the fiber studies.

DISCUSSION
In the present study, it is shown that the myosin loop 2 region is involved in the formation of the trimeric complex of myosin, actin, and twitchin. The twitchinbinding region of actin, located in subdomain I, also is known as the myosin loop 2-binding region. Thus, it is reasonable to assume that the formation of the trimeric complex involves interactions of twitchin with actin subdomain I and with the myosin head loop 2 region. The colocalization of twitchin in the region of the D2 site with myosin and actin is supported by previous electron microscope observations of isolated thick filaments (11). The electrostatic interaction of myosin loop 2 with actin is implicated in the first step of the ATP-driven movement (16). An increase in [Ca 2ϩ ] i initiates the active state by Ca 2ϩ binding to myosin and also leads to dephosphorylation of twitchin by calcineurin (17). As suggested previously (11), the formation of the high force state is thought to displace the twitchin interactions with myosin and actin and thus prevent formation of catch. This is consistent with observations made with reconstituted fibers (18). Thus, during an active contraction, the competition between the actin-myosin interaction sites and the sites on actin subdomain I and on myosin loop 2 for twitchin would be lessened by formation of the high force state. From the present study, it can be added that Ca 2ϩ also decreased the binding of actin to myosin in the presence of TWD2-S, probably reflecting a reduction in binding of the twitchin peptide to myosin and actin, as shown. This would reduce the possibility for formation of catch in the presence of Ca 2ϩ , facilitating the interaction between actin and myosin filaments.
It is interesting that ITC experiments indicated two binding sites on the twitchin fragment for both the actin peptide and the loop 2 peptide. How these four sites are regulated by phosphorylation of the D2 site is not known, and the relationship, if any, between the loop 2 sites for actin and twitchin and the actin sites for loop 2 and for twitchin requires clarification. The data from the cosedimentation studies suggest that the link between twitchin (TWD2-S) and myosin is more sensitive to competition by the loop 2 peptide. At 1:10 loop 2 peptide/TWD2-S, the interaction between actin and myosin is lost.
Under these conditions, the concentration of loop 2 peptide is about 10-fold higher than the myosin head concentration. These findings indicate that there is no direct interaction between myosin and actin when the two proteins form the trimeric complex with twitchin. Under catch condition, myosin is inactivated and does not bind to actin ( Fig. 2A), indicating that the loop 2 region does not bind to actin. In the cosedimentation assay, loop 2 peptide would have competed with the loop 2 region in myosin for binding to TWD2-S. It is noted that the cosedimentation assay was performed at cold temperature to avoid any possible denaturation of myosin and actin, whereas the ITC experiments were carried out at 25°C to facilitate the quick reaction required for the experiments between the peptides. However, proteins or peptides are well known to change their structures dramatically depending on environmental temperatures. Therefore, such possible structural differences must be carefully considered to interpret the data obtained in the two series of the experiments in the present study.
Galler and co-workers (15) reported that structures within the myofilament that are sensitive to pH changes are responsible for catch. Our study on the pH dependence of the interactions of TWD2-S showed that its binding to myosin and actin was reduced on increasing pH, but its interaction with the loop 2 peptide was not affected. With the fiber studies, the pH was critical for formation of catch, and catch was not observed at pH values around 7.4 (15). The binding of TWD2-S although slightly reduced at these pH values still occurred. Thus, the marked pH dependence seen with fibers cannot be attributed entirely to a pH dependence of the interactions involved in formation of the trimeric complex. These results are consistent with previous observations (11) where the trimeric complex was formed at pH 7.4. However, the salt concentration adopted in the present study was 200 mM KCl, which is somewhat higher than those used elsewhere for fibers. Therefore, experiments in different salt concentrations are required for a more correct interpretation of the results obtained in the present study.
A controversial and long standing issue is the identity of the physical interaction(s) responsible for maintaining tension in catch with low energy consumption. Recent studies have proposed that the myosin cross-bridge-actin interaction is not involved (10, 19 -21). Our results are consistent with this idea, and we propose that interaction between thick and thin filaments is due to the formation of a complex between actin and myosin that is cross-linked by twitchin, specifically in the region around the D2 site. The myosin site for interaction with twitchin is in the myosin head (the loop 2 region) but distinct µ ∆ ∆ ∆ ∆

Twitchin Binds to Myosin Loop 2
from the cross-bridge. The trimeric complex, however, may be sensitive to cycling cross-bridge-actin interactions because it is proposed that the complex is not formed or is displaced in the high force state.
Further studies are necessary to define and localize the various sites involved in formation of the trimeric complex. It is also important to determine whether or not the trimeric complex is formed by the reaction of twitchin with myosin and actin, cooperatively.