Regulatory Mechanism of Dictyostelium Myosin Light Chain Kinase A*

In this study, we examined the activation mechanism of Dictyostelium myosin light chain kinase A (MLCK-A) using constitutively active Ca2+/calmodulin-dependent protein kinase kinase as a surrogate MLCK-A kinase. MLCK-A was phosphorylated at Thr166 by constitutively active Ca2+/calmodulin-dependent protein kinase kinase, resulting in an ∼140-fold increase in catalytic activity, using intact Dictyostelium myosin II. Recombinant Dictyostelium myosin II regulatory light chain and Kemptamide were also readily phosphorylated by activated MLCK-A. Mass spectrometry analysis revealed that MLCK-A expressed by Escherichia coli was autophosphorylated at Thr289 and that, subsequent to Thr166 phosphorylation, MLCK-A also underwent a slow rate of autophosphorylation at multiple Ser residues. Using site-directed mutagenesis, we show that autophosphorylation at Thr289 is required for efficient phosphorylation and activation by an upstream kinase. By performing enzyme kinetics analysis on a series of MLCK-A truncation mutants, we found that residues 283–288 function as an autoinhibitory domain and that autoinhibition is fully relieved by Thr166 phosphorylation. Simple removal of this region resulted in a significant increase in the kcat of MLCK-A; however, it did not generate maximum enzymatic activity. Together with the results of our kinetic analysis of the enzymes, these findings demonstrate that Thr166 phosphorylation of MLCK-A by an upstream kinase subsequent to autophosphorylation at Thr289 results in generation of maximum MLCK-A activity through both release of an autoinhibitory domain from its catalytic core and a further increase (15–19-fold) in the kcat of the enzyme.

tein kinases, induces a high degree of myosin motor activity by specifically phosphorylating the myosin II regulatory light chain (MRLC) (1)(2)(3)(4). Myosin II phosphorylation has been shown to be involved in smooth muscle contraction and cell motility in non-muscle cells (5)(6)(7). In Dictyostelium discoideum myosin II, phosphorylation of MRLC at Ser 13 induces its actindependent ATPase activity and its motor activity (8 -11). A recent study indicated that filamentous assembly is required for efficient regulation of Dictyostelium myosin II by MRLC phosphorylation (12). In contrast to vertebrate MLCK, an unconventional MLCK (MLCK-A) has been identified and characterized in Dictyostelium and shown not to require Ca 2ϩ /CaM for its activity (13,14). The results of a genetic study using cells lacking MLCK-A indicate that MLCK-A plays a role in efficient cytokinesis (15). It has been shown that MLCK-A undergoes autophosphorylation at Thr 289 , resulting in increased activity of the enzyme (13); but autophosphorylation is not required for MLCK-A activity in vivo (16). Thus, the role of autophosphorylation at Thr 289 remains unclear. With regard to the activation mechanism of this enzyme, MLCK-A has been shown to be activated and phosphorylated by concanavalin A treatment of cells (15). This activation is likely due to phosphorylation at Thr 166 in the activation loop because the activity of the T166E mutant is 12-fold higher than that of wild-type MLCK-A (17). Interestingly, cGMP activates endogenous MLCK-A in cell lysates, suggesting that MLCK-A kinase can be either directly or indirectly up-regulated by cGMP signals and thereby induce phosphorylation and activation of MLCK-A (16). Such activation would be analogous to the activation mechanism of protein kinase cascades such as the mitogen-activated protein kinase cascade (18) and the CaM kinase cascade. For example, the multifunctional Ca 2ϩ /CaM-dependent protein kinases (CaM-Ks), CaM-KI and CaM-KIV, are activated through phosphorylation at the Thr residue in the activation loop by an upstream kinase, Ca 2ϩ /CaM-dependent protein kinase kinase (CaM-KK) (19,20).
Because neither the MLCK-A kinase nor MLCK-A phosphorylated at Thr 166 has been investigated, the regulatory mechanism of MLCK-A remains unclear. Here, we found that mammalian CaM-KK, an activating kinase for CaM-KI and CaM-KIV, was capable of phosphorylating Thr 166 of MLCK-A, resulting in a large increase in the catalytic efficiency of MLCK-A. This allowed us to examine several aspects of MLCK-A activation, including the role of autophosphorylation at Thr 289 , the specific location of the autoinhibitory region, and the relationship between Thr 166 phosphorylation and the autoinhibitory mechanism of MLCK-A.
Activation of Dictyostelium MLCK-A by CaM-KKc-Purified recombinant MLCK-As (0.16 mg/ml) were incubated at 30°C for the indicated times in a solution containing 50 mM HEPES (pH 7.5), 10 mM Mg(Ac) 2 , 1 mM DTT, 1 mM EGTA, and 200 M ATP in either the presence or absence of 6.6 g/ml CaM-KKc. The reaction was initiated by the addition of ATP and terminated by 21-fold dilution with 50 mM HEPES (pH 7.5), 2 mg/ml bovine serum albumin, 10% ethylene glycol, and 1 mM EDTA. Five l of the diluted sample (37.5 ng of MLCK-A) was then subjected to the protein kinase assay.
MLCK-A Activity Assay-MLCK-A activity was measured at 30°C for 5-10 min (or for 2 min for His-tagged MRLC) in a solution (25 l) containing 50 mM HEPES (pH 7.5), 10 mM Mg(Ac) 2 , 1 mM DTT, 40 M smooth muscle MRLC peptide (Kemptamide, Lys-Lys-Arg-Pro-Gln-Arg-Ala-Thr-Ser-Asn-Val-Phe-Ser-NH 2 ) (24) or various concentrations of recombinant Dictyostelium MRLC, and 400 M [␥-32 P]ATP (ϳ1000 cpm/ pmol) in the presence of 1 mM EGTA. The reaction was initiated by the addition of [␥-32 P]ATP and terminated either by spotting aliquots (15 l) onto phosphocellulose paper (Whatman P-81), followed by several washes with 75 mM phosphoric acid (25) for peptide phosphorylation, or by the addition of SDS-PAGE sample buffer, and the samples were then subjected to SDS-15% PAGE, followed by quantifying 32  Phosphorylation of Dictyostelium MLCK-A by CaM-KKc-Purified recombinant MLCK-As (0.16 mg/ml) were assayed at 30°C for the indicated times in a solution containing 50 mM HEPES (pH 7.5), 10 mM Mg(Ac) 2 , 1 mM DTT, 1 mM EGTA, and 200 M [␥-32 P]ATP (500 -1000 cpm/pmol) in either the presence or absence of 6.6 g/ml CaM-KKc. The reaction was initiated by the addition of [␥-32 P]ATP and terminated either by spotting 15 l of the sample onto P-81 paper, followed by measurement of 32 P incorporation into MLCK-A as described above, or by the addition of the SDS-PAGE sample buffer. The samples were then subjected to SDS-10% PAGE, followed by autoradiography.
Dictyostelium Myosin II Phosphorylation-Purified Dictyostelium myosin II (0.2-1 mg/ml) was incubated with either activated or unactivated MLCK-A (1.5 g/ml) at 30°C for the indicated times in a solution containing 150 mM NaCl, 50 mM HEPES, 10 mM Mg(Ac) 2 , 1 mM DTT, 1 mM EGTA, 0.4 mg/ml bovine serum albumin, and 1 mM [␥-32 P]ATP. The reaction was terminated by the addition of SDS-PAGE sample buffer, and the samples were then subjected to SDS-15% PAGE, followed by either autoradiography or quantifying 32 P incorporation into MRLC by Cerenkov counting of the excised gels.
Identification of Phosphorylation Sites in Dictyostelium MLCK-A by Mass Spectrometry-Either purified recombinant MLCK-A (ϳ10 g) or MLCK-A (ϳ10 g) phosphorylated by incubation with CaM-KKc for 120 min as described above was separated by SDS-10% PAGE, followed by in-gel digestion with 17 g/ml chymotrypsin (Roche Applied Science) overnight at 37°C. The digested peptides were eluted by 0.1% formic acid and subjected to liquid chromatography-tandem mass spectrometry (MS/MS) analysis. Liquid chromatography-MS/MS analysis was performed using a Micromass Q-Tof2 quadruple/time-of-flight hybrid mass spectrometer interfaced with a Micromass CapLC TM capillary reverse-phase liquid chromatography system. A 90-min linear gradient from 5 to 45% acetonitrile in 0.1% formic acid was produced and split at a 1:20 ratio, and the gradient solution was injected into a PepMap C 18 Nano LC column (75 m ϫ 150 mm; LC Packings, Amsterdam, The Netherlands) at ϳ25 nl/min. The eluted peptides were sprayed directly into the mass spectrometer. MS/MS data were acquired using Micromass MassLynx software and converted to a single text file (containing the observed m/z of the precursor peptide and fragment ion m/z and intensity values) using Micromass ProteinLynx software. The file was analyzed with the Mascot MS/MS Ions search (Matrix Science) 2 to assign non-phosphorylated and phosphorylated peptides to Dictyostelium MLCK-A amino acid sequence (GenBank TM /EBI accession number 2 Available at www.matrixscience.com.

FIG. 2. Phosphorylation and activation of Dictyostelium MLCK-A by CaM-KKc. A, phosphorylation of MLCK-A by CaM-KKc.
Purified wild-type MLCK-A (0.16 mg/ml) was incubated at 30°C for the indicated times (10 -120 min) in a solution containing 50 mM HEPES (pH 7.5), 10 mM Mg(Ac) 2 , 1 mM DTT, 1 mM EGTA, and 200 M [␥-32 P]ATP in the presence (q) or absence (E) of 6.6 g/ml CaM-KKc. 32 P incorporation into MLCK-A was measured as described under "Experimental Procedures." Results represent duplicate experiments. B, activation of MLCK-A by CaM-KKc. Purified wild-type MLCK-A (0.16 mg/ml) was incubated with 200 M ATP at 30°C for the indicated times (0 -120 min) as described for A in the presence (q) or absence (E) of 6.6 g/ml CaM-KKc. The reaction was terminated, and MLCK-A (1.5 g/ml) was subjected to the protein kinase assay at 30°C for 10 min in the presence of 400 M [␥-32 P]ATP using 40 M Kemptamide as a substrate as described under "Experimental Procedures." Results represent duplicate experiments. C, phosphorylation of Dictyostelium MRLC by MLCK-A. Purified Dictyostelium myosin II (1 mg/ml) was incubated in the presence of 1 mM [␥-32 P]ATP for 10 min at 30°C with (ϩ) or without (Ϫ) 1.5 g/ml MLCK-A, which was preincubated with 200 M ATP as described for B in the absence (Ϫ) or presence (ϩ) of 6.6 g/ml CaM-KKc for 60 min. After terminating the reaction, samples were subjected to SDS-15% PAGE, stained with Coomassie Brilliant Blue (left panel), and subjected to autoradiography (right panel). MHC, Dictyostelium myosin II heavy chain; BSA, bovine serum albumin. D, time course of MRLC phosphorylation by MLCK-A. Purified Dictyostelium myosin II was incubated in the presence of 1 mM [␥-32 P]ATP for the indicated times (5-60 min) at 30°C with 1.5 g/ml unactivated MLCK-A (E) or CaM-KKc-activated MLCK-A (q) as described for C. After terminating the reaction, samples were subjected to SDS-15% PAGE, followed by measurement of 32 P incorporation into MRLC as described under "Experimental Procedures." Results represent duplicate experiments.   Table I. The singly charged ion of a peptide (A-1, residues 283-295) derived from untreated MLCK-A (A) and that of a peptide (B-4, residues 163-179) derived from activated MLCK-A (B) were subjected to MS/MS analysis as described under "Experimental Procedures." The observed fragment ions are indicated above and below each peptide sequence. The phosphorylated Thr residues are indicated. Cys*, carboxyamidomethylcysteine. M64176). We set search parameters as follows: data base, NCBInr; taxonomy, Dictyostelium discoideum; enzyme, none; fixed modifications, carbamidomethyl; variable modifications, phospho (Ser/Thr); peptide tolerance, Ϯ0.2 Da; MS/MS tolerance, Ϯ0.2 Da; and peptide charge, 1ϩ, 2ϩ, and 3ϩ.
Protein Phosphatase 2A Treatment of Dictyostelium MLCK-A-Either wild-type MLCK-A (3.5 g) or the T289A mutant (1.2 g) was incubated with or without 0.025 unit of protein phosphatase 2A (Upstate Biotechnology, Inc., Lake Placid, NY) at 30°C for 60 min in a solution containing 50 mM Tris-HCl (pH 7.5), 5 mM MgCl 2 , and 1 mM DTT. The reaction was terminated by the addition of SDS-PAGE sample buffer, and the samples were then subjected to SDS-7.5% PAGE, followed by protein staining.
Miscellaneous-Protein concentration was estimated by staining with Coomassie Brilliant Blue (Bio-Rad) using bovine serum albumin or rabbit skeletal myosin (for Dictyostelium myosin II) as a standard (26).

RESULTS
Unlike vertebrate MLCKs, Dictyostelium MLCK-A is not activated by Ca 2ϩ /CaM (13). Mutation of Thr 166 to Glu in MLCK-A results in a 12-fold increase in its catalytic activity compared with the wild-type enzyme, suggesting that MLCK-A is likely to be activated through phosphorylation at Thr 166 by an upstream protein kinase in Dictyostelium (17). This is also consistent with the finding of a previous study that T166A and T289A mutations abolish the increase in 32 P incorporation into MLCK-A in metabolically labeled cells after stimulation with concanavalin A, which activates MLCK-A (17). However, there is no direct evidence of MLCK-A activation by Thr 166 phosphorylation because a MLCK-A kinase capable of phosphorylating Thr 166 has not been identified in Dictyostelium. Therefore, the activation mechanism of MLCK-A remains unclear. A search for metazoan homologs of Dictyostelium MLCK-A in the cDNA data base revealed that the catalytic domain of MLCK-A is most similar (ϳ50% identical) to mammalian and Caenorhabditis elegans CaM-KI ( Fig. 1) (27, 28). It has been demonstrated that both CaM-KI enzymes are phosphorylated by an upstream protein kinase (CaM-KK) at a Thr residue in their activation loop (Thr 177 in human CaM-KI and Thr 179 in C. elegans CaM-KI) that is equivalent to Thr 166 in MLCK-A, resulting in a large increase in their catalytic efficiencies (21,27,28). We therefore investigated whether MLCK-A can be phosphorylated at Thr 166 by CaM-KK, resulting in activation. If this is the case, the precise activation mechanism of MLCK-A could be determined using the surrogate MLCK-A kinase.
Phosphorylation and Activation of Dictyostelium MLCK-A by CaM-KKc-To assay for phosphorylation and activation of MLCK-A by CaM-KK, we used a minimum catalytic domain mutant of rat CaM-KK␣ (GST-CaM-KK-(84 -434); CaM-KKc) lacking the N-terminal 83 amino acid residues and lacking a C-terminal regulatory region (residues 435-505) that includes an autoinhibitory domain and Ca 2ϩ /CaM-binding segments because the mutant has been shown to be constitutively active (29,30). As shown in Fig. 2A, when we incubated recombinant GST-fused MLCK-A with CaM-KKc in the presence of MgATP and EGTA, robust phosphorylation was observed, whereas autophosphorylation of MLCK-A itself was very weak. Approximately 3.2 mol of phosphate was incorporated into MLCK-A within 120 min upon incubation with CaM-KKc. We measured MLCK-A activity under the same conditions using a peptide substrate corresponding to chicken gizzard MRLC (Kemptamide) (Fig. 2B). The activity of recombinant MLCK-A was very low or undetectable under these conditions (1.5 g/ml MLCK-A) and was not altered by incubation with MgATP in the absence of CaM-KKc. However, when we used a high concentration of the enzyme (20 g/ml MLCK-A), we could detect wild-type MLCK-A activity (4.0 nmol/min/mg), as shown in MLCK-A (Fig. 2A). These results indicate that CaM-KKc is capable of phosphorylating and activating MLCK-A, resulting in a large increase in its catalytic efficiency for the peptide substrate as well as autophosphorylation activity. We also confirmed CaM-KK-mediated induction of MLCK-A activity using purified Dictyostelium myosin II as a substrate (Fig. 2, C and  D). The protein kinase activity of MLCK-A for 18-kDa Dictyostelium MRLC was very low (1.5 nmol/min/mg) and was dra-matically enhanced by phosphorylation with CaM-KKc (0.23 mol/min/mg), resulting in efficient and stoichiometric phosphorylation of MRLC (ϳ0.8 mol of phosphate/1 mol of MRLC). We also determined the specificity constant (k cat /K m ) by measuring the phosphorylation rate at several subsaturating myosin concentrations (0.2-1 mg/ml) as described previously (17). The protein kinase activity of unactivated MLCK-A for 18-kDa Dictyostelium MRLC was very low (k cat /K m ϭ 350 M Ϫ1 s Ϫ1 ). This is slightly lower that the value measured previously (670 M Ϫ1 s Ϫ1 ) (17), most likely because of the GST tag present at the N terminus of the enzymes used in this study. For CaM-KKctreated MLCK-A, the activity was ϳ140-fold higher (k cat /K m ϭ 49,500 M Ϫ1 s Ϫ1 ).
Identification of Phosphorylation Sites in Dictyostelium MLCK-A-Although it has been demonstrated that recombinant MLCK-A undergoes autophosphorylation at Thr 289 (14, 17), we did not detect stoichiometric autophosphorylation of MLCK-A in this present study ( Fig. 2A). We therefore attempted to identify the phosphorylation sites in both unactivated and activated MLCK-As by mass spectrometry. When we analyzed our recombinant unactivated MLCK-A by chymotrypsin digestion, followed by mass spectrometry analysis, we obtained peptide sequences that covered 84% of the total amino acid sequence of MLCK-A (data not shown). Among them, we detected a single phosphopeptide corresponding to residues 283-295 in MLCK-A (peptide A-1 in Table I). MS/MS analysis revealed a single phosphorylation site at Thr 289 in the peptide (Fig. 3A), indicating that MLCK-A was already autophosphorylated at Thr 289 in E. coli, which is consistent with a previous report (17). Therefore, we could not detect stoichiometric autophosphorylation at Thr 289 in MLCK-A. In addition, we found that wild-type MLCK-A migrated on SDS-7.5% polyacrylamide gel slower than the protein phosphatase 2A-treated enzyme, whereas the mobility of the T289A mutant on SDS-polyacrylamide gel was not altered by the phosphatase treatment (Fig.  4C), indicating that Thr 289 in wild-type MLCK-A was apparently fully autophosphorylated. We also thought that activation of MLCK-A by CaM-KKc was due to phosphorylation at Thr 166 by CaM-KKc. Indeed, mass spectrometry analysis of MLCK-A incubated with CaM-KKc in the presence of MgATP for 120 min revealed that Thr 166 was phosphorylated in MLCK-A (peptide B-4 in Fig. 3B and Table I). This was also confirmed by the finding that CaM-KKc-induced P i incorporation into MLCK-A and activation of MLCK-A were completely abolished by the T166A mutation (Fig. 4, A and B). Although we observed residual autophosphorylation of the T166A mutant in either the presence or absence of CaM-KKc (Fig. 4A), this was likely autophosphorylation at Thr 289 because P i incorporation was not observed with the T166A/T289A double mutant. Based on the mass spectrometry analysis, we also observed that MLCK-A-(1-277), lacking Thr 289 , possessed a single phosphorylation site at Thr 166 after 30 min of incubation with CaM-KKc (data not shown). In addition to autophosphorylation sites at Thr 289 and phospho-Thr 166 , we detected four other phosphorylation sites (Ser 42 , Ser 155 , Ser 251 , and Ser 270 ) in activated MLCK-A (Table I). Because CaM-KKc could phosphorylate only Thr 166 in MLCK-A (Fig. 4A), these Ser residues are likely autophosphorylation sites. Autophosphorylation at multiple Ser residues of MLCK-A subsequent to activation is consistent with the results shown in Fig. 2A. When we analyzed the phosphorylation sites in the T166E mutant, which has been shown to be a phosphorylation-mimicking mutant (17), we detected two phosphorylation sites at Ser 251 and Ser 270 in addition to phospho-Thr 289 by mass spectrometry, but we could not observe phosphorylation at Ser 42 and Ser 155 in the mutant (data not shown). This may indicate that autophospho-  (Table I). However, this autophosphorylation subsequent to phosphorylation at Thr 166 was slow and did not appear to be involved in the enzymatic regulation in MLCK-A (Fig. 2B), suggesting that it may not be physiological. In addition, only Thr phosphorylation of MLCK-A has been observed in vivo (17).
Taken together, these results indicate that CaM-KKc is capable of activating MLCK-A through phosphorylation at Thr 166 . This is consistent with the finding that maximum activation of MLCK-A by CaM-KKc was achieved within 30 min and that, at this time, ϳ1 mol of phosphate was incorporated into the enzyme (Fig. 2, A and B). Notably, the maximum activity of activated MLCK-A (1.3 mol/min/mg) was ϳ300-fold higher than that of the unactivated enzyme under our conditions, indicating that mutation of Thr 166 to Glu only partially (12-fold higher compared with the wild-type enzyme) mimicked the effect of phosphorylation on the activity (17). Also the activity measured for CaM-KKc-treated MLCK-A is comparable with that measured in crude Dictyostelium lysate containing activated MLCK-A (16). Therefore, we could address the activation mechanism of MLCK-A using this surrogate MLCK-A kinase, CaM-KKc.
Requirement of Thr 289 Autophosphorylation for Efficient Activation of Dictyostelium MLCK-A-First, we examined the role of autophosphorylation at Thr 289 in activation of MLCK-A (Fig.  4). P i incorporation into the T289A mutant was significantly reduced (ϳ20%) compared with the wild-type enzyme when the mutant was incubated with CaM-KKc in the presence of MgATP for 30 min (Fig. 4A). In addition, the phosphorylationmimicking mutant T289E was more efficiently phosphorylated by CaM-KKc compared with the T289A mutant within 30 min, but the mutant was less efficiently phosphorylated by CaM-KKc compared with the wild-type enzyme under these conditions, indicating that the Glu mutation partially mimicked the effect of phosphorylation at Thr 289 . We therefore performed a kinetic experiment to examine activation of the Thr 289 mutants (Fig. 4D). As shown in Figs. 2B and 4D, wild-type MLCK-A, which was autophosphorylated at Thr 289 , was rapidly activated by CaM-KKc (t1 ⁄2 ϳ 8 min). In contrast, the T289A mutant was activated very slowly (t1 ⁄2 Ͼ 90 min), and activation of T289E was more rapid (t1 ⁄2 ϳ 20 min) than that of the T166A mutant, which is consistent with the P i incorporation into these MLCK-A mutants induced by CaM-KKc (Fig. 4A). These results indicate that autophosphorylation at Thr 289 is required for efficient phosphorylation at Thr 166 by an upstream kinase, which results in maximum activity.
Identification of an Autoinhibitory Domain in Dictyostelium MLCK-A-In the absence of phosphorylation at Thr 166 in MLCK-A, the activities of both wild-type MLCK-A and the T289A mutant were very low or below the detectable level under the conditions employed (1.5 g/ml MLCK-A) (Figs. 2  and 4). A previous study demonstrated that a truncated form lacking the C-terminal 37 amino acid residues exhibits ϳ10fold higher activity compared with full-length MLCK-A, indicating the presence of an autoinhibitory domain in its C-terminal region (14). We wished to map this autoinhibitory domain and to determine the relationship between autoinhibition and activation of MLCK-A. A series of C-terminal truncation mutants were expressed (Fig. 5A), and we used a high concentration of enzyme (20 g/ml) to accurately measure the activities of these enzymes prior to activation by CaM-KKc (Fig. 5B). Whereas truncation at residue 288 did not alter MLCK-A activity, truncation at residue 282 induced a significant increase in activity (ϳ15-fold), and the activities of additional truncation mutants (MLCK-A-(1-277) and MLCK-A-(1-270)) were indistinguishable from that of MLCK-A-(1-282) (Fig. 5B). This indicates that residues 283-288 function as an autoinhibitory region in MLCK-A to suppress MLCK-A activity. In contrast, phosphorylation and activation of all mutants by CaM-KKc were similar to those of the wild-type enzyme (Fig. 5C), suggesting that activation of MLCK-A by an upstream kinase is not affected by its autoinhibitory mechanism. Although the activities of the mutants lacking an autoinhibitory region (MLCK-A-(1-282), MLCK-A-(1-277), and MLCK-A-(1-270)) were ϳ15-fold higher than that of autoinhibited wild-type MLCK-A, they were ϳ20-fold lower than that of the activated form under these conditions.
Effect of Activation on the Kinetic Parameters of Dictyostelium MLCK-A-It was important to determine how activation by an upstream kinase alters the kinetic parameters of MLCK-A. The high molecular mass of myosin (274 kDa) made it impossible to obtain saturating concentrations in our assays; so for this substrate, k cat /K m could not be resolved into its components, k cat and K m . We therefore determined the kinetic constants of wild-type MLCK-A (autophosphorylated at Thr 289 ), activated MLCK-A, and unactivated and activated MLCK-A-(1-277) lacking an autoinhibitory region (Fig. 5) for a peptide substrate (Kemptamide), recombinant Dictyostelium MRLC, and ATP. Whereas the K m values of basal MLCK-A for both ATP and the peptide substrate were very similar to those of the activated enzyme, the k cat value was dramatically increased (ϳ800-fold) by Thr 166 phosphorylation (Table II). Deletion of the autoinhibitory region (MLCK-A-(1-277)) increased the k cat value (ϳ57-fold) compared with basal MLCK-A without significant effects on the affinities for both ATP and Kemptamide. The k cat value of MLCK-A-(1-277) was further increased (ϳ19-fold) by Thr 166 phosphorylation, which is similar to the k cat value (6.4 s Ϫ1 ) of activated wild-type MLCK-A. When we used recombinant Dictyostelium MRLC as a substrate, the kinetic changes were similar to what we observed with Kemptamide, except that the catalytic efficiency (k cat ) of basal MLCK-A for MRLC was ϳ8-fold higher than that for the peptide substrate. As a result, the k cat value of wild-type MLCK-A   was increased ϳ100-fold by activation without a significant decrease in the K m for MRLC. Similar to what we observed with the peptide substrate, the k cat value of MLCK-A-(1-277) was further increased (ϳ15-fold) by Thr 166 phosphorylation without any effects on the affinity for MRLC. These results suggest that phosphorylation at Thr 166 is accompanied by a large increase in the catalytic efficiency of MLCK-A, which is associated with suppression of the autoinhibitory mechanism and an increased k cat value. The autoinhibitory region (residues 283-288) plays a role in the complete suppression of MLCK-A activity in the absence of activation; however, releasing the autoinhibitory region from the catalytic core is not enough to generate maximum MLCK-A activity.

DISCUSSION
The activation mechanism of MLCK-A shown in this study is somewhat analogous to that of the CaM kinase cascade, which is composed of an upstream CaM-KK and downstream CaM-KI and CaM-KIV. CaM-KK phosphorylates a specific Thr residue (Thr 177 in CaM-KI and Thr 196 in CaM-KIV) in each of their activation loops, resulting in the remarkable activation of both downstream CaM-Ks in the presence of Ca 2ϩ /CaM. Previous studies demonstrated that CaM-KKc is unable to phosphorylate and activate CaM-KI and CaM-KIV in the absence of Ca 2ϩ / CaM because phosphorylation in the activation loop is blocked by the autoinhibitory regions of CaM-KI and CaM-KIV (29,30). Thus, both downstream CaM-Ks require release of autoinhibitory regions from their catalytic domains through Ca 2ϩ /CaM binding to their regulatory domains to expose the activation Thr residue to CaM-KK (22,27,29). In contrast, MLCK-A does not seem to require either cofactors or release of autoinhibition for subsequent phosphorylation at Thr 166 by an upstream kinase, but it does requires Thr 289 autophosphorylation for efficient phosphorylation at Thr 166 by an upstream kinase (Fig. 4). Unlike CaM-KI phosphorylated at Thr 177 , whose activity is completely suppressed by the autoinhibitory region in the absence of Ca 2ϩ /CaM, the activity of MLCK-A phosphorylated at Thr 166 is no longer suppressed by its autoinhibitory region (residues 283-288). This is similar to the fact that CaM-KIV generates a significant degree of the autonomous activity through phosphorylation at Thr 196 by CaM-KK (31). However, the mechanism of generation of the Ca 2ϩ /CaM-independent activity of activated CaM-KIV has remained unclear. It is intriguing that MLCK-A undergoes autophosphorylation only at Thr 289 in the absence of activation, whereas the enzyme is maintained in an autoinhibited state by the interaction of the autoinhibitory region (residues 283-288) with the catalytic core. Autophosphorylation at Thr 289 has been shown to be an intramolecular event (13), which could be explained by the likely localization of Thr 289 in or close to the catalytic cleft by interaction of the adjacent autoinhibitory segment (residues 283-288) with the catalytic domain. It then follows that autophosphorylation at Thr 289 might induce a conformational change in the MLCK-A regulatory region that fully exposes Thr 166 in the catalytic domain to an upstream kinase. This is supported by the fact that mutation of Thr 289 to non-phosphorylatable Ala significantly reduced the efficiency of phosphorylation at Thr 166 by CaM-KKc. Notably, autophosphorylation at Thr 289 did not in itself significantly suppress the autoinhibitory function to generate MLCK-A activity compared with the enhanced activity of MLCK-A mutants lacking an autoinhibitory region (Fig. 5B). Taken together, these results show that, once MLCK-A was phosphorylated at Thr 166 by an upstream kinase, MLCK-A exhibited a maximum level of kinase activity by releasing the autoinhibitory region from the catalytic core as well as by increasing the efficiency of catalysis. Because we examined the activation mechanism of MLCK-A, including the role of Thr 289 autophosphorylation and the effect of Thr 166 phosphorylation in vitro, activation of the enzyme in intact Dictyostelium remains to be examined.
It has been shown that the addition of cGMP to crude lysates from vegetative and developing cells (but not from starved cells lacking MLCK-A) induces MRLC phosphorylation in vitro, indicating that cGMP stimulates MLCK-A activity (16). However, cGMP itself is not capable of directly activating MLCK-A (16), indicating two possibilities for the activation mechanism of MLCK-A: 1) MLCK-A kinase is directly regulated by cGMP signaling; and/or 2) MLCK-A is regulated by a cGMP-regulated protein to expose Thr 166 to MLCK-A kinase, analogous to the activation mechanism of the CaM kinase cascade, in which both the upstream and downstream kinases are tightly regulated by Ca 2ϩ /CaM to constitute a signaling pathway. Although MLCK-A kinase in Dictyostelium remains to be identified, our present results suggest that the second explanation is quite unlikely because MLCK-A does not require any cofactors or regulators for phosphorylation at Thr 166 and activation by a constitutively active upstream kinase. Therefore, the upstream MLCK-A kinase activity is likely to be either directly or indirectly controlled by cGMP signaling. However, we cannot rule out the possibility that the regulatory protein(s) may bind to MLCK-A to modulate the efficiency of Thr 166 phosphorylation. Interestingly, recent studies have identified at least four candidate cGMP-binding targets in Dictyostelium (32), and cells with a deletion of the two cGMP-binding targets GbpC and GbpD show a reduced increase in MRLC phosphorylation in response to chemoattractants, suggesting a possible involvement of cGMP-binding proteins in chemotaxis-induced MRLC phosphorylation in Dictyostelium (33). Our study raises the possibility that the MLCK-A kinase could be structurally related to CaM-KK, especially with respect to its catalytic domain, which might be regulated by cGMP signaling. Another possibility is that a CaM-KK homolog that is also a Ca 2ϩ /CaMdependent enzyme may exist in Dictyostelium and render MLCK-A Ca 2ϩ -dependent, as in the case of Ca 2ϩ /CaM-regulated vertebrate MLCK, because a CaM-KK homolog has been identified in lower eukaryotes such as C. elegans (34) and Aspergillus nidulans (35). Therefore, further studies on the identification and characterization of the MLCK-A kinase will be required to clarify the detailed mechanism of the MLCK-A activation cascade and its physiological significance during cytokinesis and chemotaxis in Dictyostelium.