Myosin Regulatory Light Chain Diphosphorylation Slows Relaxation of Arterial Smooth Muscle*

Background: The regulatory light chains of smooth muscle myosin are phosphorylated at Ser19 and Thr18. Results: Phosphorylation at Thr18 does not increase force elicited by Ser19 phosphorylation, but reduces the rate of relaxation. Conclusion: Diphosphorylation slows relaxation compared with monophosphorylation at Ser19. Significance: Knowledge of the functional effects of myosin diphosphorylation is important for understanding the underlying causes of hypercontractility. The principal signal to activate smooth muscle contraction is phosphorylation of the regulatory light chains of myosin (LC20) at Ser19 by Ca2+/calmodulin-dependent myosin light chain kinase. Inhibition of myosin light chain phosphatase leads to Ca2+-independent phosphorylation at both Ser19 and Thr18 by integrin-linked kinase and/or zipper-interacting protein kinase. The functional effects of phosphorylation at Thr18 on steady-state isometric force and relaxation rate were investigated in Triton-skinned rat caudal arterial smooth muscle strips. Sequential phosphorylation at Ser19 and Thr18 was achieved by treatment with adenosine 5′-O-(3-thiotriphosphate) in the presence of Ca2+, which induced stoichiometric thiophosphorylation at Ser19, followed by microcystin (phosphatase inhibitor) in the absence of Ca2+, which induced phosphorylation at Thr18. Phosphorylation at Thr18 had no effect on steady-state force induced by Ser19 thiophosphorylation. However, phosphorylation of Ser19 or both Ser19 and Thr18 to comparable stoichiometries (0.5 mol of Pi/mol of LC20) and similar levels of isometric force revealed differences in the rates of dephosphorylation and relaxation following removal of the stimulus: t½ values for dephosphorylation were 83.3 and 560 s, and for relaxation were 560 and 1293 s, for monophosphorylated (Ser19) and diphosphorylated LC20, respectively. We conclude that phosphorylation at Thr18 decreases the rates of LC20 dephosphorylation and smooth muscle relaxation compared with LC20 phosphorylated exclusively at Ser19. These effects of LC20 diphosphorylation, combined with increased Ser19 phosphorylation (Ca2+-independent), may underlie the hypercontractility that is observed in response to certain physiological contractile stimuli, and under pathological conditions such as cerebral and coronary arterial vasospasm, intimal hyperplasia, and hypertension.

saturates the four Ca 2ϩ -binding sites of calmodulin (1). (Ca 2ϩ ) 4 -calmodulin activates myosin light chain kinase (MLCK), 2 which catalyzes phosphorylation of the motor protein myosin II at Ser 19 of its two 20-kDa regulatory light chain subunits (LC 20 ) (2). This simple phosphorylation reaction markedly increases the actin-activated MgATPase activity of myosin, which provides the energy for cross-bridge cycling and the development of force or shortening of the muscle (3). MLCK is also capable of phosphorylating LC 20 at Thr 18 in vitro, but this requires very high (unphysiological) concentrations of the kinase (4,5). Relaxation follows the removal of Ca 2ϩ from the cytosol, which inactivates MLCK, and myosin is dephosphorylated by myosin light chain phosphatase (MLCP), a type 1 Ser/Thr phosphatase (6).
We and others have demonstrated that smooth muscle contraction can be elicited in the absence of Ca 2ϩ by treatment with inhibitors of type 1 protein phosphatases (7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19). For example, treatment of Triton-skinned rat caudal arterial smooth muscle strips with the membrane-impermeant phosphatase inhibitor microcystin in the absence of Ca 2ϩ (presence of EGTA) elicited a slow, sustained contractile response that correlated with LC 20 phosphorylation (16). Further investigation revealed that this Ca 2ϩ -independent phosphorylation occurred at both Ser 19 and Thr 18 , referred to as diphosphorylation (16). The kinase responsible was shown not to be MLCK on the basis of the following observations: (i) purified MLCK is inactive in the absence of Ca 2ϩ (20 -22); (ii) LC 20 diphosphorylation requires unphysiologically high MLCK concentrations (5); (iii) MLCK inhibitors have no effect on Ca 2ϩ -independent, microcystin-induced LC 20 diphosphorylation and contraction of Triton-skinned tissue (16,19); (iv) removal of endogenous calmodulin by treatment of Triton-skinned smooth muscle strips with the calmodulin antagonist trifluoperazine in the presence of Ca 2ϩ does not affect Ca 2ϩ -independent, microcystin-induced LC 20 diphosphorylation and contraction (23); (v) endogenous LC 20 in smooth muscle myofilaments is phosphorylated in the absence of Ca 2ϩ at Ser 19 or Thr 18 alone, as well as at both sites (16), whereas purified MLCK (at high concentration) only phosphorylates Thr 18 after Ser 19 has been phosphorylated (4); (vi) stimuli that induce maximal activation of MLCK in smooth muscle tissues (e.g. membrane depolarization of intact vascular smooth muscle strips with an optimal KCl concentration, or addition of a maximal concentration of Ca 2ϩ to permeabilized strips) induce LC 20 phosphorylation exclusively at Ser 19 (23,24); (vii) Ca 2ϩ -independent LC 20 kinase activity can be separated from MLCK chromatographically (16); and (viii) the Ca 2ϩ -independent LC 20 kinase, unlike MLCK, does not use ATP␥S as a substrate (this study). We purified this Ca 2ϩ -independent LC 20 kinase activity from chicken gizzard myofilaments and identified it as integrin-linked kinase (ILK) (17). Bacterially expressed ILK phosphorylated LC 20 in intact myosin in a Ca 2ϩ -independent manner (17). Approximately 50% of cellular ILK was retained in Triton-skinned smooth muscle and may be associated with MLCP because purified phosphatase preparations contain co-purifying ILK (19). It should be noted that ILK has often been described as a pseudokinase (25), but the evidence for its bona fide kinase activity is substantial (26,27). Zipper-interacting protein kinase (ZIPK) has also been implicated in the diphosphorylation of LC 20 (18,28), although inhibition of ZIPK activity in Triton-skinned rat caudal arterial smooth muscle did not affect microcystin-induced LC 20 diphosphorylation or contraction (19), suggesting that ILK is likely the responsible kinase in these conditions.
The diphosphorylation site in LC 20 is highly evolutionarily conserved: the sequence around Thr 18 -Ser 19 (Arg-Ala-Thr-Ser-Asn-Val-Phe-Ala-Met-Phe; residues 16 -25), is identical throughout the animal kingdom and is also found in a homolog of LC 20 (29) in the genome of the unicellular choanoflagellate Monosiga brevicollis (30); choanoflagellates appear to be the closest living relatives of metazoans (30,31). LC 20 isoforms are also found in non-muscle myosin II, and contain phosphorylation sites corresponding to Thr 18 and Ser 19 of smooth muscle LC 20 that play an important role in regulation of motility (32).
The functional effects of phosphorylation of LC 20 at Ser 19 and Thr 18 have been investigated in vitro using purified LC 20 or intact myosin as substrates at high concentrations of MLCK. Ikebe and Hartshorne (4) showed that the actin-activated MgATPase activity of diphosphorylated myosin was 2-3-fold greater than that of myosin phosphorylated exclusively at Ser 19 . This increase in actomyosin MgATPase activity can be attributed to a doubling of the V max when both sites are phosphorylated (33)(34)(35). In the in vitro motility assay, however, myosin phosphorylated at both Ser 19 and Thr 18 moved actin filaments at a rate similar to myosin phosphorylated at Ser 19 alone (35,36).
Tissue Preparation and Force Measurements-Caudal arteries were removed from male Sprague-Dawley rats (300 -350 g) that had been anesthetized with halothane and euthanized according to protocols consistent with the standards of the Canadian Council on Animal Care and approved by the University of Calgary Animal Care and Use Committee. The arteries were cleaned of excess adventitia and adipose tissue in Ca 2ϩ -free H-T buffer. Segments were placed over a 0.31-mm needle and moved back and forth 40 times to remove the endothelium, cut into helical strips (1.5 ϫ 6 mm), mounted on a Grass isometric force transducer (model FT03C) connected to a PowerLab (ADInstruments) 8-channel recording device with a resting tension of 0.45 g and incubated for 20 min in H-T buffer (bath volume ϭ 0.8 ml). Tissues were stimulated at least twice with H-T buffer containing 87 mM KCl (the increase in [KCl] was balanced by a decrease in [NaCl]) with a 20-min interval of relaxation in Ca 2ϩ -free H-T buffer. Muscle strips were then incubated in Ca 2ϩ -free H-T buffer and either used for experiments with intact tissue or were skinned (demembranated) as follows. Tissues for skinning were incubated for 5 min in Buffer A and subsequently demembranated by incubation for 2 h in Buffer A containing 1% (v/v) Triton X-100. Skinned tissues were then washed 3 times (5 min each) in pCa 9 solution prior to treatments described in the figure legends.
Quantification of LC 20 Phosphorylation Levels-At selected times during experimental protocols, tissues were immersed in cold 10% trichloroacetic acid, acetone, 10 mM DTT, washed three times (1 min each) with acetone/DTT, and lyophilized for 36 h. Dried tissues were immersed in 1 ml of SDS gel sample buffer (2% (w/v) SDS, 100 mM DTT, 10% (v/v) glycerol, 0.01% bromphenol blue, 60 mM Tris-HCl, pH 6.8), heated to 95°C for 2 min, cooled to room temperature, and rotated overnight at 4°C. Samples (40 l) were subjected to phosphate affinity SDS-PAGE using an acrylamide-pendant phosphate-binding tag (Phos-tag SDS-PAGE with 12.5% acrylamide) at 30 mA/gel for 70 min in mini-gels in which 0.05 mM Phos-tag acrylamide (NARD Institute, Japan) and 0.1 mM MnCl 2 were incorporated into the running gel (49). Separated proteins were transferred to PVDF membranes (Roche Applied Science) overnight at 27 volts and 4°C in 25 mM Tris-HCl, pH 7.5, 192 mM glycine, 10% (v/v) methanol. Proteins were fixed on the membrane by treatment with 0.5% glutaraldehyde in phosphate-buffered saline (137 mM NaCl, 2.68 mM KCl, 10 mM Na 2 HPO 4 , 1.76 mM KH 2 PO 4 ) for 45 min. Membranes were then incubated with 5% nonfat dried milk in Tris-buffered saline containing Tween (TBST: 20 mM Tris-HCl, pH 7.5, 137 mM NaCl, 3 mM KCl, 0.05% Tween 20) for 1-2 h, followed by primary antibody in TBST overnight at 4°C. Following washout of the primary antibody, membranes were incubated with secondary antibody (anti-rabbit or anti-mouse IgG-horseradish peroxidase conjugate in TBST at 1:10,000 dilution) for 2 h at room temperature, washed with TBST (4 ϫ 5 min), and then with TBS (1 ϫ 5 min) before chemiluminescence signal detection using the Super-Signal West Femto reagent (Thermo Scientific, Rockford, IL). The emitted light was detected and quantified with a chemiluminescence imaging analyzer (LAS3000mini; Fujifilm) and images were analyzed with MultiGauge version 3.0 software.
Data Analysis-Values are presented as the mean Ϯ S.E., with n indicating the number of animals used; several muscle strips were used from each animal. Statistical analyses were performed with SigmaPlot and data were analyzed by Student's t test, with p Ͻ 0.05 considered to indicate statistically significant differences. 20 Diphosphorylation and Contraction- Fig. 1A show the time course of Ca 2ϩindependent contraction of Triton-skinned rat caudal arterial smooth muscle strips in response to the phosphatase inhibitor microcystin (t1 ⁄ 2 ϭ 451.1 Ϯ 13.4 s (n ϭ 8)). Tissues were immersed in TCA/acetone/DTT at the indicated times during the contractile response, washed with acetone, lyophilized, and tissue proteins were extracted in SDS gel sample buffer. Phosphorylated and unphosphorylated forms of LC 20 were separated by Phos-tag SDS-PAGE (49) and detected by Western blotting with anti-pan LC 20 , which recognizes all forms of the protein (Fig. 1B, panel a). The three separated bands were identified by Western blotting with phosphospecific antibodies to LC 20 (Fig. 1B, panels b-d). In resting tissue in the absence of Ca 2ϩ (lane 1), only unphosphorylated LC 20 was detected. Treatment with microcystin in the absence of Ca 2ϩ induced a time-dependent increase in mono-and diphosphorylated LC 20 . The monophosphorylated band contained a mixture of LC 20 phosphorylated exclusively at Ser 19 (Fig. 1B, panel b) and LC 20 phosphorylated exclusively at Thr 18 (Fig. 1B, panel c). The antibody to Thr(p) 18 -LC 20 also recognized diphosphorylated LC 20 (Fig. 1B, panel c), identified as containing both Thr(P) 18 and Ser(P) 19 in Fig. 1B, panel d. The cumulative quantitative data in Fig. 1C show the time-dependent increase in mono-and diphosphorylation, and the corresponding decrease in unphosphorylated LC 20 in response to microcystin in the absence of Ca 2ϩ .

Ca 2ϩ -independent, Microcystin-induced LC
Ca 2ϩ -independent, Calyculin-A-induced LC 20 Diphosphorylation and Contraction-Treatment of intact rat caudal arterial smooth muscle with the membrane-permeant phosphatase inhibitor calyculin-A in Ca 2ϩ -free solution also induced LC 20 mono-and diphosphorylation, which correlated with force development with a t1 ⁄ 2 of 1326 Ϯ 96 s (n ϭ 6) (Fig. 2). In this case, the amount of monophosphorylated LC 20 detected was significantly less (Fig. 2C) than was observed in the Tritonskinned tissue in response to microcystin (Fig. 1C). It is also noteworthy that the steady-state force achieved in response to calyculin-A in the absence of Ca 2ϩ appeared to be significantly higher than the force induced by a strong depolarizing stimulus (87 mM KCl) ( Fig. 2A). This prompted us to address the question: does LC 20 diphosphorylation elicit more steady-state isometric force than monophosphorylation?
KCl-induced LC 20 Monophosphorylation and Contraction-We first demonstrated that an increase in cytosolic free Ca 2ϩ concentration induced exclusively monophosphorylation of LC 20 at Ser 19 . Ca 2ϩ entry via voltage-gated Ca 2ϩ channels was activated by KCl-induced membrane depolarization of intact rat caudal arterial smooth muscle strips, which induced a rapid contractile response (t1 ⁄ 2 ϭ 10.2 Ϯ 0.2 s (n ϭ 29)) (  (Fig. 3B, panels a and d). LC 20 phosphorylation stoichiometry peaked at ϳ0.6 mol of P i /mol of LC 20 (Fig. 3C).
Effects on Force and LC 20 Phosphorylation of Sequential Treatment with Ca 2ϩ and Microcystin-Similarly, addition of Ca 2ϩ to Triton-skinned rat caudal arterial smooth muscle induced phosphorylation of LC 20 exclusively at Ser 19 (Fig. 4G, lanes A in panels a-d) with a t1 ⁄ 2 of 151.7 Ϯ 4.8 s (n ϭ 23) and an LC 20 phosphorylation level of ϳ0.5 mol of P i /mol of LC 20 ( Table 1). Addition of microcystin at the plateau of a Ca 2ϩinduced contraction resulted in a further increase in force of ϳ25% ( Fig. 4B and Table 2), which correlated with LC 20 diphosphorylation (Fig. 4G, lanes B in panels a-d, and Table 1). If microcystin and Ca 2ϩ were added together, a rapid contraction occurred (t1 ⁄ 2 of 65.3 Ϯ 2.3 s (n ϭ 15) compared with 151.7 Ϯ 4.8 s (n ϭ 23) for Ca 2ϩ alone and 451.1 Ϯ 13.4 s (n ϭ 8) for microcystin at pCa 9), which was again accompanied by LC 20 diphosphorylation (Fig. 4G, lanes C in panels a-d, and Table 1).
No force development or LC 20 phosphorylation was observed in the absence of Ca 2ϩ and phosphatase inhibitor (Fig. 4, D and G, lanes D in panels a-d, and Table 1). If contraction was evoked by addition of microcystin in the absence of Ca 2ϩ , subsequent addition of Ca 2ϩ elicited further force development (ϳ20%; Fig. 4F and Table 2) and LC 20 diphosphorylation (Fig.  4G, lanes F in panels a-d, and Table 1) compared with control ( Fig. 4, E and G, lanes E, and Tables 1 and 2). A more detailed analysis of the (Ca 2ϩ ϩ microcystin)-induced contraction revealed rapid phosphorylation of LC 20 at Ser 19 that can be attributed to MLCK activation by Ca 2ϩ , and a slower rate of phosphorylation at Thr 18 , due to ILK activity that is unmasked by the phosphatase inhibitor (Fig. 5).
Effects on Force and LC 20 Phosphorylation of Combined Treatment with KCl and Calyculin-A-Calyculin-A treatment of intact rat caudal arterial smooth muscle in the presence of extracellular Ca 2ϩ elicited a slow, sustained contraction (Fig. 6, green trace) with a t1 ⁄ 2 of 1206 Ϯ 102 s (n ϭ 6), which was indistinguishable from the calyculin-A-induced contraction in Ca 2ϩ -free solution (t1 ⁄ 2 ϭ 1326 Ϯ 96 s (n ϭ 6)) ( Fig. 2A). Membrane depolarization in the presence of extracellular Ca 2ϩ elicited a rapid increase in force (t1 ⁄ 2 ϭ 10.2 Ϯ 0.2 s (n ϭ 29)), which subsequently declined to a steady-state level (Figs. 3A and 6, red trace). The simultaneous application of KCl and calyculin-A in the presence of extracellular Ca 2ϩ elicited a contractile response (Fig. 6, black trace) that matched the superimposed contractions due to membrane depolarization (Fig. 6, red trace) and phosphatase inhibition (Fig. 6, green trace): the initial rapid contractile response in the presence of KCl and calyculin-A occurred with a t1 ⁄ 2 of 11.2 Ϯ 0.6 s (n ϭ 6), i.e. similar to the contraction induced by KCl treatment alone (t1 ⁄ 2 ϭ 10.2 Ϯ 0.2 s (n ϭ 29)), whereas the slow, sustained contractile response occurred with a t1 ⁄ 2 of 1110 Ϯ 84 s (n ϭ 3), i.e. similar to the contraction induced by calyculin-A in Ca 2ϩ -free solution (t1 ⁄ 2 ϭ 1326 Ϯ 96 s (n ϭ 6)). We hypothesize that the biphasic contractile response to KCl and calyculin-A involves two distinct mechanisms: the rapid response is attributable to membrane depolarization-mediated Ca 2ϩ entry and MLCK activation, and

Smooth Muscle Myosin Light Chain Diphosphorylation
the slow response to calyculin-A-mediated inhibition of MLCP with unmasking of Ca 2ϩ -independent LC 20 kinase activity. These mechanisms are supported by measurements of site-specific LC 20 phosphorylation during the time course of contraction in the presence of extracellular Ca 2ϩ and following addition of both KCl and calyculin-A (Fig. 7). Thus, there was a rapid initial increase in LC 20 monophosphorylation (Fig. 7B,  panel a), which occurred exclusively at Ser 19 (Fig. 7B, panels b  and c), followed by a slight dephosphorylation (Fig. 7C) leading to partial relaxation (Fig. 7A). It was only at prolonged incubation times that diphosphorylation of LC 20 was observed (Fig.  7B, panels a and d), which correlated with the slow, sustained phase of contraction (Fig. 7A).
Stoichiometric Phosphorylation of LC 20 at Ser 19 in Tritonskinned Tissue-The results described above suggest that phosphorylation of LC 20 at Thr 18 may increase the level of force that is achieved in intact or Triton-skinned rat caudal arterial smooth muscle as a result of Ser 19 phosphorylation. Alterna-tively, the observed increases in force could be due to an increase in the total level of Ser 19 phosphorylation, rather than phosphorylation at Thr 18 . To distinguish between these possibilities, it would be necessary to achieve stoichiometric phosphorylation exclusively at Ser 19 and then observe whether or not phosphorylation at Thr 18 has an additional effect on steadystate force. The next step, therefore, was to achieve stoichiometric phosphorylation exclusively at Ser 19 . Unfortunately, treatment of intact tissue with an optimal KCl concentration to elicit a maximal increase in [Ca 2ϩ ] i , leading to maximal activation of MLCK, does not lead to stoichiometric phosphorylation of LC 20 at Ser 19 (Fig. 3). This is due to competing dephosphorylation of LC 20 by MLCP, which is constitutively active. Likewise, in Triton-skinned tissue, addition of a maximal [Ca 2ϩ ] fails to elicit stoichiometric LC 20 phosphorylation at Ser 19 for the same reason (Fig. 4G, lane A in panel a, and Table 1). We tested the possibility that the stoichiometry of LC 20 phosphorylation could be increased by addition of exogenous calmodu- lin and MLCK to Triton-skinned tissue in the presence of Ca 2ϩ , recognizing the caveat that, if the MLCK concentration was too high, it would phosphorylate Thr 18 as well. Whereas the addition of calmodulin in the absence or presence of MLCK did increase LC 20 phosphorylation slightly, there remained a significant amount of unphosphorylated LC 20 , and a low level of LC 20 diphosphorylation was observed (supplemental Fig. S1 and Table S1). This approach was, therefore, unsuitable for achieving stoichiometric phosphorylation at Ser 19 in the absence of Thr 18 phosphorylation.
An alternative approach to achieve stoichiometric LC 20 phosphorylation was to use ATP␥S to thiophosphorylate LC 20 : MLCK uses ATP␥S as a substrate (50), but the thiophosphorylated protein is not a substrate for MLCP (51). This approach was used successfully with Triton-skinned rat caudal arterial smooth muscle (Fig. 8). Triton-skinned tissues were shown to be viable by contraction at pCa 4.5 in the presence of ATP and an ATP regenerating system, and relaxation following removal of Ca 2ϩ (Fig. 8A). Following removal of ATP, incubation with ATP␥S in the presence of Ca 2ϩ , but absence of ATP or an ATP regenerating system, resulted in stoichiometric thiophosphorylation of LC 20 at Ser 19 (Fig. 8B, lanes 2 and 3). It is noteworthy that thiophosphorylated LC 20 migrates more rapidly upon Phos-tag SDS-PAGE than does phosphorylated LC 20 , which enables clear discrimination between phosphorylated and thiophosphorylated forms of the protein.
ATP␥S is not hydrolyzed by activated myosin and therefore does not support cross-bridge cycling and contraction (20,52,53). Stoichiometric thiophosphorylation at Ser 19 (Fig. 8B, lanes  2 and 3) was, therefore, not accompanied by contraction (Fig.  8A). Transfer to pCa 9 solution containing ATP and an ATP regenerating system following washout of ATP␥S resulted in a rapid contractile response (t1 ⁄ 2 ϭ 21.2 Ϯ 0.2 s (n ϭ 8)) and steady-state force corresponding to 85.4 Ϯ 1.9% (n ϭ 8) of the pCa 4.5-induced contraction (Fig. 8A). Once the steady-state force was achieved, microcystin was added at pCa 9 in the presence of ATP and an ATP regenerating system. No additional force development was observed (77.3 Ϯ 4.2% (n ϭ 5) of pCa 4.5-induced contraction), although significant di(thio)phosphorylation of LC 20 did occur (Fig. 8B, lanes 6 and 7, and Table 3).
The identities of the thiophosphorylated LC 20 species as depicted in Fig. 8B were verified by the use of phosphospecific antibodies (supplemental Fig. S2). Incubation of Tritonskinned rat caudal arterial smooth muscle strips with ATP␥S and microcystin at pCa 9, in the absence of ATP and an ATP regenerating system, failed to elicit thiophosphorylation of LC 20 (supplemental Fig. S3, lanes 3 and 4). This is in contrast to incubation with ATP␥S at pCa 4.5, in the absence of ATP and an ATP regenerating system, which led to LC 20 monothiophosphorylation (supplemental Fig. S3, lane 2) at Ser 19 (supplemental Fig. S2, lanes 3-5). 20 on the Rates of Dephosphorylation and Relaxation-Finally, we investigated the possibility that LC 20 diphosphorylation may affect relaxation, rather than contraction, by comparing the rates of dephospho- rylation and relaxation of Triton-skinned rat caudal arterial smooth muscle following monophosphorylation of LC 20 at pCa 4.5 or diphosphorylation of LC 20 at pCa 9 in the presence of okadaic acid. Okadaic acid was chosen as the phosphatase inhibitor for these experiments, rather than microcystin, because its effects are readily reversible (54), whereas microcystin can covalently modify the catalytic subunit of type 1 protein phosphatase, resulting in irreversible inhibition of the phosphatase (55). Indeed, we have observed that microcystin-induced contractions cannot be reversed by washout of the inhibitor (data not shown).

Effects of Diphosphorylation of LC
Comparable levels of phosphorylation of LC 20 were achieved with pCa 4.5 (0.48 Ϯ 0.02 mol of P i /mol of LC 20 (n ϭ 4)) or okadaic acid treatment at pCa 9 (0.49 Ϯ 0.09 mol of P i /mol of LC 20 (n ϭ 5)), with monophosphorylation occurring exclusively in response to Ca 2ϩ and both mono-and diphosphorylation being detected in the presence of okadaic acid, as expected (Fig. 9C). The steady-state force generated by okadaic acid at pCa 9 was 83.3 Ϯ 1.4% (n ϭ 9) of that at pCa 4.5 (supplemental Fig. S4). Relaxation was initiated by transfer to pCa 9 solution and the time courses of LC 20 dephosphorylation and relaxation were quantified (Fig. 9, A and B, respectively). The rate of dephosphorylation of LC 20 was markedly reduced in the tissues in which LC 20 had been diphosphorylated compared with tissues containing exclusively monophosphorylated LC 20 (Fig.  9A): t1 ⁄ 2 values were 83.3 s for Ca 2ϩ -treated tissue and 560 s for okadaic acid-treated tissue. This correlated with a reduction in the rate of relaxation (Fig. 9B): t1 ⁄ 2 values were 560 s for Ca 2ϩtreated tissue and 1293 s for okadaic acid-treated tissue. The slower rate of dephosphorylation following okadaic acid treatment cannot be explained by slow washout of the inhibitor because MYPT1-Thr 697 and -Thr 855 (the inhibitory phosphorylation sites in the myosin targeting subunit of MLCP) (56) were maximally dephosphorylated at the first time point analyzed during the relaxation, i.e. when force was at 90% (supplemental Fig. S5). LC 20 diphosphorylation has been observed in several smooth muscle tissues treated with various contractile stimuli, including carbachol- (37) and neurally stimulated bovine tracheal smooth muscle (38), prostaglandin-F 2␣ -stimulated rabbit thoracic aorta (39,40), and angiotensin II-stimulated rat renal efferent arterioles (41). LC 20 diphosphorylation has also been observed in pathological cases of smooth muscle hypercontractility, for example, coronary artery spasm (44,45), cerebral vasospasm after subarachnoid hemorrhage (43,46), and intimal hyperplasia (42). More recently, Cho et al. (57) provided evidence for enhanced Ca 2ϩ -independent LC 20 diphosphorylation and force generation in ␤-escin-permeabilized mesenteric arterial smooth muscle rings of spontaneously hypertensive rats compared with normotensive Wistar Kyoto controls. Furthermore, phenylephrine induced significant LC 20 diphosphorylation in the spontaneously hypertensive rat arteries. Evidence was also presented that ZIPK contributes to the Ca 2ϩindependent LC 20 diphosphorylation through phosphorylation of MYPT1 at Thr 697 and possibly direct phosphorylation of LC 20 , and the expression level of ZIPK, but not ILK, was greater in spontaneously hypertensive rats than Wistar Kyoto tissues (57). Collectively, these data suggest that LC 20 diphosphorylation may account for the hypercontractility observed in smooth muscle tissues in response to certain contractile stimuli and in pathological situations. It was, therefore, important to determine the functional effect of LC 20 phosphorylation on smooth muscle contractility. The results of these studies led to the following conclusions.

DISCUSSION
(i) Treatment of Triton-skinned rat caudal arterial smooth muscle with the phosphatase inhibitor microcystin in the absence of Ca 2ϩ induced a slow, sustained contraction, as previously observed (16), which correlated with LC 20 phosphorylation at Ser 19 and Thr 18 (Fig. 1).
(ii) Similar results were obtained when intact tissues were treated with the membrane-permeant phosphatase inhibitor calyculin-A in the absence of extracellular and stored Ca 2ϩ (Fig.  2). However, an interesting difference between the Tritonskinned and intact tissues was observed: microcystin treatment of skinned tissue induced monophosphorylation at Ser 19 and Thr 18 at similar rates (Fig. 1B, panels b and c), in addition to diphosphorylation (Fig. 1B, panel d), whereas no monophosphorylation was observed at Thr 18 following calyculin-A treatment of intact tissue in the absence of extracellular Ca 2ϩ (Fig.  2B, panel c), but instead Ser 19 monophosphorylation was followed by Thr 18 phosphorylation to form the diphosphorylated species (Fig. 2B). This suggests that LC 20 phosphorylation at the two sites was random in the Triton-skinned tissue experiments but sequential in the intact tissue experiments. A possible explanation would be that distinct kinases are involved in the two situations, the most likely candidates being ILK and ZIPK, and we have provided evidence that ILK is responsible for  LC 20 phosphorylation levels were quantified by Phos-tag SDS-PAGE (see Fig. 4G, panel a) in tissues treated as described in the legend to Fig. 4, A-F microcystin-induced Ca 2ϩ -independent contraction of Tritonskinned rat caudal arterial smooth muscle (19).
(iii) The level of steady-state force induced by calyculin-A in the absence of Ca 2ϩ is significantly greater than that induced by a maximally effective concentration of KCl, i.e. an optimal Ca 2ϩ signal ( Fig. 2A). This would be consistent with diphosphorylation of LC 20 increasing steady-state force compared with Ser 19 monophosphorylation. Indeed, addition of microcystin to Triton-skinned tissue pre-contracted at pCa 4.5 (Fig. 4B), or of Ca 2ϩ to tissue pre-contracted with microcystin in the absence of Ca 2ϩ (Fig. 4F), evoked a significant increase in steady-state force (Table 2), which correlated with increases in LC 20 diphosphorylation ( Fig. 4G and Table 1). However, Ser 19 phosphorylation stoichiometry also increased under these conditions (from ϳ0.5 mol of P i /mol of LC 20 to ϳ1 mol of P i /mol of LC 20 ) ( Table 1), suggesting that the enhanced force responses could be due to increased phosphorylation at Ser 19 (whether in the form of monophosphorylated or diphosphorylated LC 20 ).
(v) The fact that the rate of contraction of Triton-skinned rat caudal arterial smooth muscle in response to Ca 2ϩ was signifi-  cantly faster (t1 ⁄ 2 ϳ 150 s) than that in response to microcystin at pCa 9 (t1 ⁄ 2 ϳ 450 s) suggested that it may be possible to induce maximal phosphorylation at Ser 19 before achieving diphosphorylation, and thereby determine more convincingly if diphosphorylation causes additional force development. Furthermore, treatment with microcystin at pCa 4.5 caused a significant increase in the rate of contraction (t1 ⁄ 2 ϳ 65 s) compared with Ca 2ϩ alone (t1 ⁄ 2 ϳ 150 s) or microcystin alone (t1 ⁄ 2 ϳ 450 s). Detailed analysis of the (Ca 2ϩ ϩ microcystin)-induced contraction of Triton-skinned rat caudal arterial smooth muscle revealed rapid phosphorylation of LC 20 at Ser 19 (which can be attributed to MLCK activation by Ca 2ϩ ) and a slower rate of phosphorylation at Thr 18 (due to ILK activity that is unmasked by the phosphatase inhibitor) (Fig. 5). The observation that no additional force was evoked as diphosphorylated LC 20 appeared argues that Thr 18 phosphorylation likely does not increase steady-state force beyond that achieved by phosphorylation at Ser 19 .
(vi) The combination of KCl and calyculin-A in the presence of Ca 2ϩ induced a biphasic contractile response of intact tissue (Fig. 6), which corresponds to the combined contractile responses to KCl in the presence of Ca 2ϩ and calyculin-A in the absence or presence of Ca 2ϩ . In this case, the initial rapid phasic contraction correlated with Ser 19 phosphorylation, and the slow sustained contractile response with the diphosphorylation of LC 20 (Fig. 7). The contractile effects of KCl and calyculin-A, however, could be explained entirely by Ser 19 phosphorylation.
It was necessary, therefore, to devise a way to achieve stoichiometric phosphorylation at Ser 19 without Thr 18 phosphorylation, and then observe whether subsequent phosphorylation at Thr 18 has an effect on steady-state force development. This was achieved by using ATP␥S to evoke close-to-stoichiometric thiophosphorylation at Ser 19 with very little dithiophosphorylation ( Fig. 8B and Table 3). Subsequent phosphorylation of LC 20 at Thr 18 (Fig. 8B) failed to elicit an increase in force (Fig.  8A). We conclude, therefore, that phosphorylation at Ser 19 of LC 20 accounts for maximal force development, and no further force results from additional phosphorylation at Thr 18 .
We then turned our attention to the possibility that diphosphorylation may affect relaxation rather than contraction by comparing the time courses of dephosphorylation of LC 20 and  relaxation of Triton-skinned muscle strips that had been precontracted under conditions that evoked phosphorylation exclusively at Ser 19 or at both Ser 19 and Thr 18 to the same overall phosphorylation stoichiometry. The rates of dephosphorylation and relaxation were significantly slower in the case of diphosphorylated LC 20 (Fig. 9). We conclude, therefore, that diphosphorylation of LC 20 at Thr 18 and Ser 19 has a marked effect on relaxation compared with monophosphorylation at Ser 19 .
The mechanism underlying the reduction in the rate of dephosphorylation of diphosphorylated LC 20 compared with Ser 19 -monophosphorylated LC 20 remains to be determined. A possibility is that the K m of MLCP for diphosphorylated LC 20 may be significantly higher than that for LC 20 phosphorylated exclusively at Ser 19 . Although such kinetic comparisons have not been performed to date, in vitro assays indicated that dephosphorylation of diphosphorylated LC 20 (whether free or in intact myosin) occurred by a random mechanism, with dephosphorylation at Ser 19 and Thr 18 occurring at similar rates (5).
The principal conclusions from this study are: (i) the level of steady-state force is dictated by the level of Ser 19 phosphorylation and is unaffected by Thr 18 phosphorylation; and (ii) Thr 18 phosphorylation reduces the rate of LC 20 dephosphorylation and relaxation, supporting a sustained contractile response. There is abundant literature indicating that most contractile stimuli elicit phosphorylation exclusively at Ser 19 and this can be explained by Ca 2ϩ -induced activation of MLCK, with or without a modest degree of Ca 2ϩ sensitization due to MLCP inhibition (58). Specific stimuli and pathophysiological situations associated with hypercontractility induce LC 20 diphosphorylation at Thr 18 and Ser 19 . This can be explained by  increased MLCP inhibition, unmasking constitutive Ca 2ϩ -independent LC 20 kinase activity (ILK and/or ZIPK), and potentially an increase in activity of Ca 2ϩ -independent LC 20 kinases, leading to an increase in Ser 19 phosphorylation (force) and Thr 18 phosphorylation (sustained contraction). ILK and ZIPK are therefore potential therapeutic targets for the treatment of cerebral and coronary vasospasm, intimal hyperplasia, hypertension, and other conditions associated with hypercontractility.