Role of the N-terminal Region of the Regulatory Light Chain in the Dephosphorylation of Myosin by Myosin Light Chain Phosphatase*

Myosin regulatory light chain (RLC) is phosphorylated at various sites at its N-terminal region, and heterotrimeric myosin light chain phosphatase (MLCP) has been assigned as a physiological phosphatase that dephosphorylates myosin in vivo. Specificity of MLCP toward the various phosphorylation sites of RLC was studied, as well as the role of the N-terminal region of RLC in the dephosphorylation of myosin by MLCP. MLCP dephosphorylated phosphoserine 19, phosphothreonine 18, and phosphothreonine 9 efficiently with almost identical rates, whereas it failed to dephosphorylate phosphorylated serine 1/serine 2. Deletion of the N-terminal seven amino acid residues of RLC markedly decreased the dephosphorylation rate of phosphoserine 19 of RLC incorporated in the myosin molecule, whereas this deletion did not significantly affect the dephosphorylation rate of isolated RLC. On the other hand, deletion of only four N-terminal amino acid residues showed no effect on dephosphorylation of phosphoserine 19 of incorporated RLC. The inhibition of dephosphorylation by deletion of the seven N-terminal residues was also found with the catalytic subunit of MLCP. Phosphorylation at serine 1/serine 2 and threonine 9 did not influence the dephosphorylation rate of serine 19 and threonine 18 by MLCP. These results suggest that the N-terminal region of RLC plays an important role in substrate recognition of MLCP.

nine 9 of RLC. Phosphorylation of these sites does not activate myosin motor activity, but rather inhibits it because of a decrease in affinity for actin (7)(8)(9).
One of the enzymes which catalyzes dephosphorylation of the RLC of myosin is called myosin light chain phosphatase (MLCP). Serine/threonine protein phosphatases can be classified into several types based upon their substrate specificity, sensitivity to inhibitors, and the requirement of divalent cations for their activation (10 -12), i.e. types 1, 2A, 2B, and 2C. A number of myosin phosphatases have been prepared from various smooth muscle tissues. Pato and co-workers (13,14) purified four distinct protein phosphatases from avian gizzards, i.e. SMP-I, SMP-II, SMP-III, and SMP-IV. SMP-I and SMP-II dephosphorylated isolated myosin regulatory light chain, but not intact myosin (13), and are classified as PP2A and PP2C, respectively. SMP-III and SMP-IV dephosphorylated intact myosin, and the characterization of these enzymes indicates that they are similar to type I phosphatases (14,15). However, they are not inhibited by protein phosphatase inhibitor 2 (14,15), which is one of the characteristics of type I phosphatase (10).
It has been shown that myosin preparations contain significant myosin phosphatase activity, suggesting that MLCP associates with myosin. MLCP purified from a smooth muscle myosin preparation (16) is a type I phosphatase based upon biochemical characterization. MLCP prepared from the actomyosin fraction of gizzard smooth muscle is composed of three subunits, 130, 37, and 20 kDa, of which the 37 kDa is the catalytic subunit (17). Similar phosphatases were subsequently purified by other groups (18,19). Because these smooth muscle MLCPs show an affinity to myosin, it is thought that these phosphatases are responsible for the dephosphorylation of myosin in situ. Consistent with this notion, it was shown that the major myosin phosphatase in skinned smooth muscle is the type I phosphatase.
Whereas the isolated RLC can be dephosphorylated by several types of protein phosphatases, only the myosin-associated phosphatase efficiently dephosphorylates intact myosin (16 -19) but types 2A and 2C (13) phosphatases do not. This suggests that there may be a specific substrate recognition mechanism for dephosphorylation of myosin by MLCP. In the present paper, we studied the role of the N-terminal residues of RLC in dephosphorylation of smooth muscle myosin. The site specificity of the myosin-associated MLCP was also studied using RLC phosphorylated at various sites.

Preparation of Proteins and Construction of RLC Expression Vector-
Smooth muscle myosin (20) and myosin light chain kinase (21) were prepared from turkey gizzards as described. Myosin was washed several times with high MgCl 2 (30 mM) containing buffer to remove residual MLCP activity. HMM and S1 were prepared from gizzard myosin as described (22). Calmodulin was prepared from frozen bull testes according to Walsh et al. (23). MLCP was prepared from turkey gizzard as described (17). The catalytic subunit of smooth muscle MLCP was prepared as described (16). cDNA of smooth muscle myosin RLC was obtained from a chicken gizzard gt11 cDNA library and subcloned into a PT7-7 Escherichia coli expression vector (24) as described (25). Truncation of RLC cDNA was done using a PCR-based method as described previously (25). The expression of recombinant RLC and its mutants in the E. coli strain BL21 (DE3) was performed according to Kamisoyama et al. (26). The expressed RLC in E. coli cells was extracted in ureacontaining buffer and purified with a series of liquid chromatography steps according to the method of Ikebe et al. (25).
Protein Biochemical Procedures-Phosphorylation of RLC at serine 19 was achieved by incubating RLC (4 mg/ml) with 2 g/ml MLCK and 5 g/ml calmodulin in buffer containing 0.5 mM [␥-32 P]ATP, 150 mM KCl, 1 mM MgCl 2 , 0.1 mM CaCl 2 , 1 mM DTT, and 30 mM Tris-HCl, pH 7.5 (buffer A), at 25°C for 20 min. The extent of phosphorylation was 0.9 mol of phosphate/mol of RLC, and no phosphorylation at threonine residues was detected based upon phosphoamino acid analysis, indicating that only serine 19 was phosphorylated. Phosphorylation at both serine 19 and threonine 18 was achieved by incubating RLC (4 mg/ml) with 50 g/ml MLCK and 30 g/ml calmodulin in modified buffer A containing 1 mM [␥-32 P]ATP and 30 mM KCl at 25°C for 40 min. Incorporation of 1.8 mol of phosphate/mol of RLC was obtained, and an equal amount of phosphoserine and phosphothreonine was detected using phosphoamino acid analysis, indicating that both serine 19 and threonine 18 were phosphorylated (27). Phosphorylation of RLC (2 mg/ml) at serine1/serine2 and threonine 9 by PKC was done by incubating with a solution containing 3 g/ml PKC, 100 ng/ml phorbol 12-myristate 13-acetate, 0.1 mg/ml phosphatidylserine, 1 mM [␥-32 P]ATP, 30 mM KCl, 1 mM MgCl 2 , and 30 mM Tris-HCl, pH 7.5 at 24°C for 50 min. Two mol of phosphate/RLC was incorporated, and both serine and threonine were phosphorylated as judged by phosphoamino acid analysis (Ref. 7, also see Fig. 6). The phosphorylated RLCs were precipitated with 5% trichloroacetic acid and dissolved in and dialyzed against buffer B (30 mM KCl, 1 mM DTT, and 30 mM Tris-HCl, pH 7.5).
RLC-deficient gizzard myosin was prepared according to the method of Trybus et al. (28) with modification (29). Phosphorylated RLC (2 molar excess) was added to RLC-deficient myosin in buffer C (30 mM KCl, 1 mM DTT, 2 mM MgCl 2 , and 30 mM Tris-HCl, pH 7.5) at 0°C. After 10 min, the myosin was centrifuged for 2 min at 10,000 ϫ g. The pellets were suspended with buffer C and centrifuged again. This step was repeated three times. The pellets were then dissolved with 0.4 M KCl, 5 mM DTT, and 30 mM Tris-HCl, pH 7.5, and used for experiments. The obtained myosin contained phosphorylated RLC as judged by SDSpolyacrylamide gel electrophoresis analysis followed by autoradiography (not shown). The extent of phosphorylation of the obtained myosin was stable at least for 6 h at 25°C.
Phosphoamino acid analysis was done as described previously (27). The extent of phosphorylation of RLC was determined as described (27).

Dephosphorylation of Myosin by Smooth Muscle MLCP-
Smooth muscle myosin containing various truncated RLCs ( Fig. 1) was prepared as described under "Materials and Methods." Dephosphorylation of native smooth muscle myosin phosphorylated at serine 19 of RLC by smooth muscle MLCP was described by single exponential kinetics (Fig. 2). The dephosphorylation time course of myosin incorporating exogenous serine 19 phosphorylated RLC was identical to that of native myosin, suggesting that the exogenous RLC was properly incorporated into myosin. The rate of dephosphorylation was somewhat reduced with NDL-4 RLC-incorporated myosin, but the decrease was not significant. On the other hand, when the N-terminal seven residues of RLC were deleted (NDL-7), the dephosphorylation by smooth muscle MLCP was markedly and significantly inhibited. The rate of dephosphorylation was 1/10 of that of myosin incorporating wild type RLC. Further deletion of three amino acid residues did not affect the rate of dephosphorylation. These results suggest that Ala 5 -Lys 6 -Ala 7 is critical for dephosphorylation of RLC at serine 19 by MLCP.
Dephosphorylation of Isolated RLC-Because many protein phosphatases can dephosphorylate isolated RLC but not intact myosin, we wished to assay whether or not the effect of deletion of the N-terminal amino acid residues observed above is also found with isolated RLC as a substrate. To address this notion, the dephosphorylation reaction by smooth muscle MLCP was carried out with isolated truncated mutants of RLC. As shown in Fig. 3, the rate of dephosphorylation was not markedly affected by the deletion of any N-terminal residues of RLC. The rates of dephosphorylation of NDL-4 and NDL-7 were slightly lower than that of wild type RLC, but the decrease was not significant (Fig. 3). These results indicate that the deletion of Ala 5 -Lys 6 -Ala 7 diminished dephosphorylation of serine 19 of RLC only in the presence of myosin heavy chain but not in its absence. These results also suggest that myosin heavy chain plays a role in the smooth muscle MLCP-RLC interaction.

Dephosphorylation of Myosin by the Catalytic Subunit of Smooth Muscle MLCP-Smooth
muscle MLCP is composed of three subunits, i.e. the myosin binding subunit, catalytic subunit, and small subunit (17)(18)(19). It was shown that phosphatase activity against myosin is enhanced in the presence of the myosin binding subunit (30), suggesting a difference in the nature of the dephosphorylation reaction between apoenzyme and holoenzyme. To examine whether the decrease in phosphatase activity of smooth muscle MLCP by the deletion of the N-terminal region of myosin-incorporated RLC is because of the function of the myosin binding subunit or because of the characteristics of the catalytic subunit of smooth muscle MLCP, we used catalytic subunit of smooth muscle MLCP to dephosphorylate truncated phosphorylated RLC incorporated into myosin. As shown in Fig. 4, dephosphorylation of myosin by the catalytic subunit of smooth muscle MLCP was also significantly attenuated by deletion of the N-terminal residues. The rate of dephosphorylation of NDL-7-containing myosin was 9% of the rate of wild type RLC-containing myosin. This was also confirmed using a different myosin fragment which con- 25°C. An aliquot was removed at various times, and the reaction was stopped by the addition of 5% TCA. The phosphorylation level of myosin was measured as described (21). Solid lines represent single exponential decay curve of the dephosphorylation of wild type RLC and NDL-7, respectively. E, wild type RLC; ƒ, NDL-4; , NDL-7; q, NDL-10; Ⅺ, native myosin. tained truncated RLC generated by a brief tryptic proteolysis of myosin which cleaves off the N-terminal 16 residues of RLC (31). Myosin was first phosphorylated with MLCK, then subjected to tryptic digestion and used as a substrate for smooth muscle MLCP holoenzyme and apoenzyme as described under "Materials and Methods." The rate of dephosphorylation of myosin incorporating the truncated RLC which lacked the Nterminal 16 residues was significantly decreased (not shown).
Dephosphorylation of Myosin Phosphorylated at Various Sites on RLC-It is known that RLC of smooth muscle and nonmuscle myosins are phosphorylated at various sites in the N-terminal region (7,8,32) (see Fig. 1). Serine 19 and threonine 18 are phosphorylated by MLCK (32). On the other hand, serine 1/serine 2 and threonine 9 are phosphorylated by PKC (7,8) and cdc2 kinases (9). Myosin phosphorylated at these different sites was prepared (see "Materials and Methods"), and the rate of dephosphorylation by smooth muscle MLCP was determined. The dephosphorylation of myosin phosphorylated at both serine 19 and threonine 18 showed a single exponential decay curve, and the rate of dephosphorylation was identical to that for dephosphorylation of myosin phosphorylated at serine 19 alone (Fig. 5). This suggests that the rate of dephosphorylation of serine 19 and threonine 18 of RLC by smooth muscle MLCP is indistinguishable. Consistent with this notion, phosphoamino acid analysis revealed that both phosphoserine and phosphothreonine were decreased to the same extent after 10 min of the dephosphorylation reaction where approximately 70% of the total incorporated phosphate was removed by the phosphatase (Fig. 6). On the other hand, the dephosphorylation of myosin phosphorylated at PKC sites (i.e. serine 1/serine 2 and threonine 9) by smooth muscle MLCP showed dual phases. Myosin was initially dephosphorylated by smooth muscle MLCP with a rate similar to that for the dephosphorylation of the serine 19/threonine 18 sites but then became resistant to dephosphorylation by MLCP. To determine the sites resistant to smooth muscle MLCP, a sample was taken at 30 min after the addition of smooth muscle MLCP and subjected to phosphoamino acid analysis. As shown in Fig. 6, both phosphoserine and phosphothreonine were detected before MLCP addition, whereas only phosphoserine was observed at 30 min after the addition of MLCP, indicating that phosphorylation at serine 1/serine 2 is resistant to dephosphorylation by smooth muscle MLCP. This difference in the MLCP susceptibility of serine and threonine sites was also found using isolated RLC as a substrate (not shown).
Effect of PKC Phosphorylation on the Dephosphorylation of MLCK Sites-The results described above indicate that the N-terminal region of RLC is important for the dephosphorylation of RLC at serine 19/threonine 18 when it is incorporated into myosin heavy chain. Because PKC and cdc2 kinases phosphorylate RLC in the N-terminal region, it is of interest to examine whether or not the phosphorylation of RLC by these kinases can affect the dephosphorylation of serine 19/threonine 18. RLCs were first phosphorylated by PKC under conditions incorporating 2.0 mol of phosphate/mol of RLC with nonradioactive ATP as a substrate. RLCs were then precipitated with 5% TCA, dissolved, and dialyzed against neutral pH buffer. RLC was then phosphorylated with MLCK using radioactive ATP to incorporate 32 P into both serine 19 and threonine 18. RLCs were then hybridized with RLC-deficient myosin as described under "Materials and Methods." As shown in Fig. 7, the dephosphorylation of serine 19/threonine 18 by smooth muscle MLCP was not significantly affected by phosphorylation at serine 1/serine 2 and threonine 9. The rate was decreased by 18% because of serine 1/serine 2 and threonine 9 phosphorylation. To determine the effect of phosphorylation at MLCK sites on the dephosphorylation of PKC sites, RLC was first phosphorylated by MLCK at serine 19 and threonine 18 with nonradioactive ATP and then phosphorylated with radioactive ATP by PKC. The phosphorylated RLC was incorporated into myosin heavy chain and subjected to the MLCP-catalyzed dephosphorylation reaction. The time course of dephosphoryl-ation of RLC at PKC sites was unaffected by MLCK phosphorylation (not shown).

DISCUSSION
While isolated RLC can be dephosphorylated by various serine/threonine protein phosphatases (1)(2)(3)(4)(5), not many protein phosphatases dephosphorylate RLC incorporated into the myosin molecule. A type 1 serine/threonine protein phosphatase has been identified as the physiological myosin light chain phosphatase in smooth muscle (16 -19); therefore, we studied the nature of myosin dephosphorylation catalyzed by this myosin light chain phosphatase as presented in this paper. It was shown that the dephosphorylation of myosin at serine 19/threonine 18 of RLC is significantly influenced by the presence of the N-terminal amino acid residues of RLC. The critical residues are Ala 5 -Lys 6 -Ala 7 because the deletion of seven N-terminal amino acid residues markedly attenuates smooth muscle MLCP-induced dephosphorylation while the deletion of four N-terminal amino acid residues does not. An interesting result is that the decrease in smooth muscle MLCP-dependent dephosphorylation is found only when RLC associates with myosin heavy chain. A possible scenario to account for this finding would be that the N-terminal region of RLC is involved in the interaction of RLC with smooth muscle MLCP and that the serine 19/threonine 18 of RLC are somewhat occluded from the smooth muscle MLCP catalytic site in the presence of the heavy chain. It is plausible that the interaction of smooth muscle MLCP with the N-terminal region of RLC increases the accessibility of the MLCP catalytic site to the phosphate moieties of phosphorylated RLC.
Previously, it was found that the dephosphorylation of myosin at serine 19 by smooth muscle MLCP is significantly affected by the conformation of myosin, i.e. a folded conformation of myosin is highly resistant to dephosphorylation by MLCP (33). It was suggested that the myosin tail interacts with the N-terminal region of RLC, thus stabilizing the folded conformation, because the mutation of the basic residues at the N-terminal region of RLC abolishes the folded conformation (25). That result is consistent with the present finding that the N-terminal region of RLC plays a role in the substrate recognition of smooth muscle MLCP. It is plausible that in the folded conformation smooth muscle MLCP cannot access the substrate site because of obstruction by the myosin tail binding to the N-terminal region of RLC.
It is known that the rates of phosphorylation of serine 19 and threonine 18 by MLCK are significantly different from each other. In contrast, the rate of dephosphorylation of these sites by smooth muscle MLCP is practically the same. The results suggest that the difference in the phosphorylation level at serine 19 and threonine 18 in cells reflects the difference in the phosphorylation rate of these sites but not by the dephosphorylation process. On the other hand, the dephosphorylation of PKC/cdc 2 kinase sites showed a marked difference in their susceptibility to smooth muscle MLCP. It has been known that PKC phosphorylates threonine 9 several times faster than the serine site in vitro (7), whereas phorbol ester, an activator of PKC, induces only serine phosphorylation but not threonine phosphorylation in cells (34). The present result provides a clear answer to this apparent discrepancy. The phosphorylation at threonine 9 would be rapidly dephosphorylated by MLCP in cells but the phosphorylated serine 1/serine 2 would be resistant to dephosphorylation by MLCP. Other protein phosphatases may slowly dephosphorylate the serine 1/serine2 sites in cells because a spontaneously active aorta phosphatase was shown to dephosphorylate both threonine and serine sites with the same rate constant (7). However, the fact that serine 1/serine 2 sites remained phosphorylated in cells after phorbol  7. Effect of phosphorylation at PKC sites on dephosphorylation of serine 19/threonine 18. RLC was first phosphorylated by PKC as described under "Materials and Methods" with nonradioactive ATP. Phosphorylated RLC was precipitated with 5% TCA and then dissolved and dialyzed against 50 mM Tris-HCl, pH 7.5, 50 mM KCl, 1 mM DTT. RLC phosphorylated by PKC was then phosphorylated at serine 19 by MLCK with [␥-32 P]ATP. The phosphorylated RLC was incorporated into myosin as described under "Materials and Methods." The dephosphorylation reaction was done as described in the legend to Fig. 2. Semilogarithmic plots of time course data are shown. E, control myosin; q, myosin phosphorylated at serine 1/serine 2 ϩ threonine 9 with nonradioactive ATP. ester stimulation provides further evidence that MLCP is the physiological myosin light chain phosphatase. The slow dephosphorylation of the serine 1/serine 2 sites was previously shown with crude gizzard phosphatase (7) so the present results are consistent with the earlier results.
The time course of dephosphorylation of myosin at serine 19 by smooth muscle MLCP is explained by a single rate constant. Furthermore, the rate of dephosphorylation of serine 19 is virtually the same for HMM and S1 (not shown). These results indicate that the dephosphorylation process of RLC at serine 19 is random and independent relative to the other head of myosin. In phosphorylation reactions, it has been reported that myosin is phosphorylated sequentially by MLCK, i.e. phosphorylation of the first head is faster than that of the second head (35)(36)(37)(38), although this is still controversial. If one accepts the sequential phosphorylation of myosin by MLCK, it would be expected that significantly higher MLCK activity would be required for phosphorylation of the second head and that the majority of phosphorylated myosin population at lower overall levels of RLC phosphorylation in cells would be singly phosphorylated myosin. To date, it is controversial whether or not the motor activity of phosphorylated myosin head is dependent on the phosphorylation of the other head of myosin, but the actomyosin contractile activity as a function of overall myosin phosphorylation in cells could be complex.
While the N-terminal region of RLC is important for determining the dephosphorylation rate of myosin at serine 19 by smooth muscle MLCP, phosphorylation at this region, i.e. serine 1/serine2 and threonine 9, failed to influence the dephosphorylation rate of myosin at serine 19. This is in contrast to the MLCK reaction in which the rate of phosphorylation at serine 19 is decreased by phosphorylation at serine1/serine 2 and threonine 9 (7,8). This difference might be because of a difference in the manner of substrate recognition between the two enzymes because MLCK requires basic residues at the N-terminal side of the phosphorylation sites. This charge interaction may not be critical for the MLCP reaction.