Arachidonic Acid Inhibits Myosin Light Chain Phosphatase and Sensitizes Smooth Muscle to Calcium*

Arachidonic acid (AA) increased, at constant Ca2+, the levels of force and 20-kDa myosin light chain (MLCzo) phosphorylation in permeabilized smooth muscle, and slowed relaxation and MLCzo dephosphorylation. The Ca2+-sensitizing effect of AA was not inhibited by inhibitors of AA metabolism (indometha- cin, nordihydroguaiaretic acid, or propyl gallate), of protein kinase C (pseudopeptide) or by guanosine-5‘- 0-(B-thiodiphosphate) and was abolished by oxidation of AA in air. A non-metabolizable AA analog, 5,8,11,14-eicosatetraynoic acid) also had Ca2+-sensi- tizing effects. Extensive treatment with saponin abolished the Ca2+-sensitizing effects of phorbol 12,13-dibutyrate and guanosine-5’-O-(y-thiotriphosphate), but not that of AA. A purified, oligomeric MLCzo phosphatase isolated from gizzard smooth muscle was dissociated into subunits by AA, and its activity

Contraction of smooth muscle is triggered by the activation, by Ca2+-calmodulin, of myosin light chain kinase that phosphorylates Ser-19 on the regulatory (20 kDa) myosin light chain (MLC20),' permitting the activation of myosin ATPase by actin (reviewed by Hartshorne (1987)). The major mechanism of relaxation, in turn, is dephosphorylation of MLCzo by myosin light chain phosphatase(s). Therefore, we had suggested that inhibition of MLCPO phosphatase is the mechanism of G-protein-coupled Ca2+ sensitization , and subsequently demonstrated that Ca*+-sensitizing agonists and G T P r S inhibit MLCzo phosphatase in permeabilized smooth muscle (Kitazawa et d . , 1991b). However, MLCzo phosphatase is thought to be strongly associated with myosin filaments Dent et al., 1992;Sellers and Pato, 1984)) whereas the surface receptors of Ca2+-sensitizing agonists and, presumably, the G-proteins that couple the effector system that inhibits MLCzo phosphatase, are associated with the plasma membrane. Therefore, the messenger or cascade that relays the inhibitory message from a surface membrane-bound G-protein to the filament-bound protein phosphatase remains to be identified. In the present study, we explored the possibility that AA, or related lipid messenger(s), may be involved in Ca2+ sensitization.
Arachidonic acid (AA) can be produced in eukaryotic cells through, a t least, three mechanisms: through the direct action of phospholipase A2 on phospholipids, through sequential action of phospholipase C and diacylglycerol lipaseb), or through the action of diacylglycerol lipase on phosphatidic acid generated by phospholipase D. AA can be further processed by cyclooxygenase, lipoxygenase, and P450 enzymes to yield metabolites that can affect their respective surface receptors and act as potential autocrine regulators (Needleman et d . , 1986). In addition, AA itself can activate protein kinase C (Kikkiawa et al., 1988;Nishizuka, 1989;McPhail et al., 1984), and phorbol esters, well-known activators of kinase C, have Ca2+-sensitizing effects on smooth muscle (Chatterjee and Tejada, 1986;Park and Rasmussen, 1985;Itoh et al., 1988; The abbreviations used are: MLCZo, 20-kDa myosin light chain; AA, arachidonic acid; ML-9, (l-(5-chloronaphthalenesulfonyl)-lHhexahydro-1,4-diazepine); ETYA, 5,8,11,14-eicosatetraynoic acid; Me2S0, dimethyl sulfoxide; HMM, heavy meromyosin; PDBu, phorbol-12,13-dibutyrate; PP1, protein phosphatase-1; SM-PPlM, smooth muscle protein phosphatase-lM; GTPrS, guanosine-5'-0-(y-thiotriphosphate; GDTBS, guanosine-5'-O-(B-thiodiphosphate). Nishimura et al., 1990;Ruzycky and Morgan, 1989;Gupta et al., 1990). In several cell systems, agonists can activate phospholipase A2 and increase the production of AA through a Gprotein-coupled pathway that can also be activated by GTPyS or aluminum fluoride (Buckley et al., 1991;Burch et al., 1986;Jelsema, 1987;Narasimhan et al., 1990). Consequently, it has been proposed that AA can act as a cellular messenger (Axelrod et al., 1988, Burch, 1989. In the present study, we examined the effects of AA on force development and relaxation, and on myosin light chain phosphorylation and dephosphorylation, in permeabilized smooth muscle. We also determined the effects of AA on the activity and oligomeric state of an isolated smooth muscle myosin light chain phosphatase. Our results show that (unmetabolized) AA can inhibit both purified myosin light chain phosphatase and myosin light chain dephosphorylation in situ, while sensitizing contraction to Ca2+ through a mechanism that requires neither activation of protein kinase C nor the "downstream" intervention of a G-protein.

MATERIALS AND METHODS
Small strips (2-3 mm long and 100-200 pm wide for tension measurements or 300-400 pm wide for MLC,, phosphorylation measurements) of rabbit femoral artery were dissected and stretched to 1.5 times resting length. Isometric tension was measured with a force transducer (AE 801; AME, Horten, Norway) in a well on a "bubble" plate . In preliminary experiments, we found that reducing the temperature from 37 "C to room temperature (22-24 "C) lengthened the lag time (between addition of AA to the bath and the onset of contraction) and the contraction half-time from, respectively, 1 2 0.3 min (n = 5) to 12 & 3.5 min ( n = 5) ( p < 0.05) and from 14 f -0.4 min (n = 4) to 46 f 11.6 min (n = 4) ( p < 0.01). Therefore, unless noted otherwise, experiments were conducted a t 37 "C. After steady responses to high K+ were observed, the strips were incubated i n normal relaxing solution (Ca2+-free, 1 mM EGTA) for about 5 min, and permeabilized by 60-75 min of incubation, a t room temperature, with 5,000-10,000 units/ml (based on rabbit red blood cell hemolysis) of Staphylococcus aurem a-toxin (GIBCO). The higher concentration of a-toxin (10,000 units/ml) and longer incubation time (75 min) were used when multiple strips were permeabilized simultaneously. For permeabilization with saponin, the strips were incubated with 200 pg/ml saponin for 22 min. Saponin was freshly dissolved in Me2S0 just before use: the final Me2S0 concentration in the bath was 1%.
For the introduction of protein kinase C pseudosubstrate peptide into receptor-coupled smooth muscle, the cells were permeabilized by 20 min of incubation with 50 pM p-escin (Kobayashi et al., 1989).
To deplete the sarcoplasmic reticulum of calcium, all permeabilized strips were treated with A23187 (10 p~) for 10 min in relaxing solution Kobayashi et al., 1991a).
Phosphorylation of MLC20 was measured with two-dimensional isoelectric focusing and sodium dodecyl sulfate-gel electrophoresis, as previously published (Kitazawa et al., 1991a). The dephosphorylation and relaxation rate measurements were carried out a t 20 "C (Kitazawa et al., 1991b) on strips rapidly frozen in freon-22 cooled by liquid nitrogen at the indicated intervals.
Details of the solutions used for studies on permeabilized strips were described previously Kobayashi et al., 1989Kobayashi et al., , 1991. Calmodulin (1 pM) was added to the Ca2+ containing activating solution for experiments employing saponin and 0-escin permeabilization.
S. aureus cu-toxin was purchased from GIBCO/BRL, GDPBS, and GTPyS from Boehringer Mannheim, and saponin from ICN Nutritional Biochemicals, Cleveland, OH. AA, myristic acid, oleic acid, phosphatidic acid (dimyristoyl-and dioleitoyl-phosphatidic acid), lysophosphatidic acid (myristoyal-and oleitoyal-lysopbosphatidic acid), p-escin, PDBu (phorbol-12,13-dibutyrate), and ML-9 (1-(5chloronaphthalenesulfonyl)-lH-hexahydro-l,4-diazepine) were purchased from Sigma and ETYA (5,8,11,14-eicosatetraynoic acid) from Cayman Chemical Co., Ann Arbor, MI. Protein kinase C pseudosubstrate peptide (19-36 sequence: Arg-Phe-Ala-Arg-Lys-Gly-Ala-Leu-Arg-Gln-Lys-Asn-Val-His-Glu-Val-Lys-Am)) was purchased from Peninsula Laboratories, Belmont, CA. AA was dissolved in ethanol or in M e 8 0 and stored a t -20 "C in the dark under nitrogen. The stock solutions were diluted immediately before use, and the final solvent concentrations in bath were 1%. In control experiments in which the strips were exposed to Me,SO or ethanol (1% each) for the same length of time as the exposures to AA, 1% of either solvent was added to the strips before addition of the fatty acids with the given solvent, to compensate for any solvent effect. 1% ethanol had a slight Ca2+-sensitizing effect, whereas 1% Me2S0 had a slight depressant effect on contraction. The solutions containing MezSO 1% or ethanol 1% were exchanged for a solution containing MezSO 1% or ethanol 1% in addition to AA or the other fatty acids used. The effect of AA was not significantly different regardless of the solvent used ( p > 0.05 a t 1, 30, 100, and 300 p~ AA), and the results are combined. Similarly, there was no significant difference between the CaZ'sensitizing effect of AA purchased from either Sigma or from Cayman Chemical Co.
Protein phosphatase assays were carried out as described (Cohen et al., 1988Cohen, 1991 in 50 mM Tris-HC1, 0.1 mM EGTA (pH 7.0 a t 25 "C), 0.1% (v/v) 2-mercaptoethanol, 0.03% Brij 35 (solution A). The substrate concentrations were 10 p~ (phosphorylase) and 0.3 p~ (heavy meromyosin (HMM)). Assays (30 pl) were carried out a t 30 "C in the presence of varying concentrations of AA or oleic acid, and caffeine (5 mM) was added when phosphorylase was the substrate. The AA was obtained from a 100 mM stock solution dissolved in MezSO and stored a t -70 "C. The small amount of MezSO carried over into the assays (<0.5% by volume) had no effect on either phosphorylase phosphatase or myosin phosphatase activity. Oleic acid (sodium salt) was stored a t -20 "C, as a 1 mM aqueous solution in solution A. The dephosphorylation of both heavy meromyosin and phosphorylase was limited to <20% to ensure that rates of dephosphorylation were linear with respect to time. 1 milliunit of activity was that amount of phosphatase which catalyzed the release of 1.0 nmol of phosphate from each substrate in 1 min.
Statistical significance was evaluated by Student's t test, and the numbers are expressed in mean f S.E.
The dose-response curve to AA and to another unsaturated free fatty acid, oleic acid (C18:1), and to a saturated fatty acid, myristic acid (C14:0), were determined noncumulatively by adding single doses of the fatty acid to individual strips, and the forces were normalized to the second maximal Ca2+induced contraction. Oleic acid and myristic acid also caused some Ca'+ sensitization, but their effects were much smaller than that of AA (Fig. 2).
AA (300 p~) caused a smaller (16 f 8.0%, n = 8) increase in force in pCa > 8,lO mM EGTA-containing solution. Lower concentrations (30 and 100 p~) of AA did not cause force development under these Ca2+-free conditions. The contraction in the absence of Ca2+ was presumably due to the presence of a constitutively active MLC kinase in smooth muscle (Gong et al., 1992) Intact, unlike permeabilized, rabbit femoral artery strips ( n = 4) did not develop force when incubated with 300 p~ AA in normal Krebs' solution for 30-60 min. In fact, contractions induced by 154 mM K+ and by 100 p~ phenylephrine were inhibited by preincubation with AA (300 ~L M ) to, respectively, 36 f 7.4% ( n = 3) and 11 f 5.9% (n = 4) of their original responses. When cytosolic Ca'+ was increased by partially depolarizing the membrane with 30 mM K', 300 p~ AA caused a small increase in force, evidenced as a definite upward deflection in each force record, from 26 f 6.0% (n = 3) to 37 f 5.0%, ( n = 3 , p < 0.01) of the maximal, 154 mM K+-induced contraction; these contractions were transient (reached peak at about 5 min) and were followed by relaxation to near basal tension. ETYA (100 p~) , a non-metabolizable analog of AA, also had a biphasic (small contraction followed by relaxation) effect, similar to that of AA, on submaximally contracted, non-permeabilized smooth muscle ( n = 3 ).
Ca2+ Sensitization Is Not Mediated by the Metabolic Products of Fatty Acids-AA can be metabolized to biologically active products in cells (Needleman et al., 1986). T o determine whether Ca2+ sensitization is the direct effect of AA or of its metabolites, we determined the effects of inhibition of AA metabolism on the Ca2+-sensitizing effect of AA.
Propyl gallate (100 p~) , an antioxidant that inhibits both cyclooxygenase and lipoxygenase (Smith et al., 1985;Rainesford, 1988), did not inhibit, but rather increased the effect of AA by shifting the dose-response curve to the left (Fig. 3). The combination of both a cyclooxygenase inhibitor, indomethacin (100 p~) , and a lipoxygenase inhibitor, nordihydroguaiaretic acid (100 p~) , also did not inhibit AA-induced Ca2+ sensitization (data not shown).
ETYA is an isomorphic competitive inhibitor of AA in which four alkyne bonds replace four alkene bonds. ETYA competitively inhibits the uptake of AA into cellular membranes and inhibits lipoxygenase and, at higher concentrations, cyclooxygenase enzymes (Anderson et al., 1989). ETYA also had a Ca'+-sensitizing effect on a-toxin-permeabilized femoral artery. 100 p~ ETYA caused 28 f 7.6% (n = 4) of the maximal Ca2+-induced contraction in the presence of a submaximally activating concentration of Ca2+ (pCa 6.7).
The Ca2+-sensitizing effect of AA was significantly decreased by exposing it for 48 h to air and light at room temperature. 100 p~ of this oxidized AA did not cause any Ca'+ sensitization of a-toxin-permeabilized femoral artery ( n = 4), but freshly prepared AA subsequently added to the same strips caused 47 f 2.8% ( n = 4) of the maximal Ca2+-induced contraction. This response is lower than in the absence of oxidized AA and suggests an inhibitory effect of the latter. in the same strips by 100 p~ phenylephrine with 10 p~ GTP (Fig. 4).
MLCz0 Dephosphorylation and Relaxation Are Inhibited by Arachidonic Acid-Agonist-and GTPyS-induced Ca2+ sensitization is mediated through inhibition of MLCPO phosphatase (Kitazawa et al., 1991b) Therefore, to determine whether AA operates through the same mechanism, we examined its effect on MLCzo dephosphorylation.
The effects of AA on the rates of contraction and relaxation and on the rate of MLCzo dephosphorylation were determined at 20 "C. In the absence of Ca2+, at this temperature, the preincubation with AA itself did not cause contraction. Preincubation with 300 p~ AA for 30 min did not change the rate of force development in response to pCa 5, but it significantly slowed relaxation. The half-time of pCa 5-induced contraction was 1.9 f 0.06 min ( n = 11) in the absence and 2.0 f 0.05 min ( n = 13, p > 0.05) in the presence of AA (300 p M , 30 min of preincubation). To measure the rate of relaxation, strips were activated with pCa 5 and, after force reached a plateau, incubated in relaxing solution containing 10 mM EGTA (pCa > 8) and 100 p~ ML-9 to inactivate MLCIO kinase (Kitazawa et al., 1991b). AA (300 p~) prolonged the half-time of relaxation from 3.5 -t 0.45 min ( n = 6) to 6.7 & 0.82 min ( n = 6, p e 0.01).
To assess more directly the effect of AA on MLCZo phosphatase activity in situ, we determined the rate of MLCIO dephosphorylation of muscles in Ca2+-and ATP-free, 10 mM EGTA and 100 p~ ML-9-containing relaxing solution. After activation with maximal Ca2+ (pCa 5) for 15 min, strips were incubated in the above Ca2+-free solution and rapidly frozen after 1, 2, 3, and 6 min. As shown in Fig. 6, AA slowed down the rate of MLClo dephosphorylation ( p < 0.001 at 2 and 3 min) .

Arachidonic Acid and Oleic Acid Inhibit Purified Smooth
Muscle Protein Phosphatase 1~ by Dissociating It into Subunits-The major protein phosphatase that dephosphorylates MLCSO of smooth muscle myosin was recently purified from avian gizzard and shown to be composed of three subunits with apparent molecular masses of 130, 37, and 20 kDa. The 37-kDa component was identified as the p-isoform of protein phosphatase-1 (PPl), while the 130-and 20-kDa components formed a regulatory complex that enhanced, by about %fold, the myosin phosphatase activity of PP1 and suppressed phosphorylase phosphatase activity by 80% Cohen et al., 1992).
In the present study, heavy meromyosin was used in the assays because of its much greater solubility in the low ionic strength buffers used to assay the phosphatase. As shown in Fig. 7A, AA (20 phi) produced a small (30%) activation of the myosin phosphatase activity of SM-PP1M but was strongly inhibitory at higher concentrations, with 50% inhibition occurring at about 60 p~ AA. In contrast, phosphorylase phosphatase activity was inhibited by about 50% up to 20 W M AA and activated at the higher concentrations which suppressed myosin phosphatase activity. Oleic acid had effects similar to those of AA, except that higher concentrations were required

t o observe equivalent inhibition and activation
of myosin phosphatase and phosphorylase phosphatase activity (Fig.  7 B ) . For example, 250 p~ oleic acid was required for 50% inhibition of the myosin phosphatase activity as compared to 60 p~ for AA.
The inhibition of myosin phosphatase which occurs above 30 ~L M AA, or above 150 ~L M oleic acid, is similar to that observed previously when the regulatory 130-20 kDa complex is dissociated from the 37-kDa catalytic subunit in the presence of high concentrations of the chaotrope LiBr (Alessi et al., 1992). In order to investigate whether the effects of AA and oleic acid also resulted from dissociation of the subunits of SM-PPlM, the purified phosphatase was incubated for 15 min at 30 "C with 300 ~L L M AA and then subjected to gel filtration on Superose 12 equilibrated in buffer containing 300 p~ AA. As shown in Fig. 8, the catalytic activity eluted from the column with an apparent molecular mass of about 35 kDa, whereas the native enzyme elutes from Superose 12 with an apparent molecular mass of about 500 kDa (Fig. 8, open  circles). Similar results were obtained when SM-PP1M was incubated with 400 ~L M oleic acid and subjected to gel filtration in buffer containing 400 ptM oleic acid (data not shown). The dissociated catalytic subunit of SM-PP1M was not inhibited by AA up to 200 p M .

DISCUSSION
The major findings of this study are that AA can sensitize smooth muscle to Ca'+ by inhibiting dephosphorylation of MLCzo in situ, and that it also dissociates and inhibits an oligomeric myosin light chain (MLCZo) phosphatase isolated from smooth muscle.
The increase in force elicited by AA at constant Ca'+ concentration was accompanied by an increase in MLCZo phosphorylation. This response is identical to that seen during Ca'+ sensitization induced by excitatory agonists (Kitazawa et al., 1991a) and GTPyS (Kitazawa et al., 1991a;Fujiwara et al., 1989;Nishimura et al., 1990;Kubota et al., 1992). The slowing of relaxation and MLCzo dephosphorylation by AA (this study) also parallels the inhibition of relaxation and dephosphorylation by Cat+-sensitizing agonists and GTPyS ; GTPyS also inhibits the HMMphosphatase activity of smooth muscle homogenates (Kubota et al., 1992).
The Cat+-sensitizing effects of AA do not appear to be mediated by any of the products of the known metabolic pathways of AA, as they were not inhibited by either cyclooxygenase (indomethacin) or lipooxygenase (nordihydroguaiaretic acid) inhibitors, but to the contrary, were enhanced by propyl gallate (Fig. 3), a general inhibitor of AA metabolism. Furthermore, both oleic acid, that is not a direct source of AA metabolites, and a non-metabolizable AA analog, ETYA, had similar, although less pronounced, Ca'+-sensitizing effects on force development, while oxidation of AA in air abolished its Cat+-sensitizing effect. Finally, the inhibitory effect of AA on a purified smooth muscle myosin light chain phosphatase provides clear evidence of the effectiveness of the non-metabolized fatty acid.
The effects of AA do not appear to be mediated by a Gprotein because, unlike the effects of Ca'+-sensitizing agonists and GTPyS (Fig. 4) (Himpens et al., 1990;Kitazawa et al., 1989;Somlyo et al., 1990), they were not inhibited by GDPPS. Similarly, the inhibition of purified HMM phosphatase was observed in the absence of added guanine nucleotide or Gprotein. Therefore, although some effects of AA are mediated through its action on G-proteins (Nakajima et al. 1990), the effects of AA described in our study do not require the involvement of a downstream G-protein intermediate. Indeed, the Ca*+-sensitizing effects of AA were observed even after the effects of GTPyS were abolished by extensive permeabilization with saponin (Fig. 5).
Protein kinase C is also unlikely to be a necessary mediator of the Ca2+-sensitizing effect of AA, although AA and other unsaturated fatty acids (Nishizuka, 1989;Kikkiawa et al., 1988;Khan et al., 1992;McPhail et al., 1984;Sutton and Haeberle, 1990) can activate kinase C in several systems. Phorbol esters, activators of kinase C, are also reported to stimulate phospholipase Az (Carter et al., 1989) and have Ca2+-sensitizing effects on smooth muscle (Chattejee and Tejada, 1986;Ruzycky and Morgan, 1989). However, a pseudosubstrate peptide inhibitor of kinase C (House and Kemp, 1988) did not inhibit AA-induced Ca2+-sensitization that, furthermore, was also present after extensive permeabilization with saponin that abolished the effect of PDBu (Fig. 5).
The effects of AA on permeabilized smooth muscle could also not be ascribed to stimulation of MLCzO kinase: AA did not affect the rate of force development and, furthermore, it is reported to inhibit purified MLCzo kinase (Kigoshi et al., 1990). The relaxant effect of AA (and of its non-metabolizable analog, ETYA) on depolarized, submaximally contracted, intact (non-permeabilized) smooth muscle may have been due t o inhibition of Ca2+ influx through voltage-gated Ca*+ channels (Shimada and Somlyo, 1992) and/or to stimulation of guanylate cyclase (Laychock, 1989). However, since we neither monitored nor buffered cytoplasmic Ca2+ in these experiments (on intact smooth muscle), the mechanism(s) of AA action on intact smooth muscle remains to be identified.
The catalytic subunit of MLCSO phosphatase in avian smooth muscle is identical to PP1, and it is now well established that PP1 is complexed to a variety of regulatory subunits in cells that target it to particular locations, modify its specificity, and permit its regulation by extracellular signals Dent et al., 1990Dent et al., , 1992. The major phosphatase which dephosphorylates MLCIO in avian smooth muscle is a trimeric enzyme in which PP1 (37 kDa) is complexed t o a 130-kDa subunit which, in turn, interacts with a 20-kDa protein (Cohen et al., 1992;Alessi et al., 1992). The inhibition of this trimeric MLCPO phosphatase by AA is of considerable interest, not only because it provides evidence of the direct effect of AA on the enzyme, but also because the inhibitory effect appears to result from dissociation of the catalytic subunit. It has been established previously that dissociation of the catalytic subunit induced by high concentrations of LiBr decreases the rate of dephosphorylation of myosin, while at the same time increasing the rate of dephosphorylation of glycogen phosphorylase, and that these effects are reversed when the catalytic subunit and the regulatory subunits are recombined . Similarly, AA (above 20 ,AM) increases phosphorylase phosphatase and decreases myosin phosphatase activity (Fig. 7) and promotes dissociation of the MLCZo phosphatase holoenzyme (Fig. 8). It will be of great interest to determine whether in muscle, as in solution, such dissociation of the oligomeric enzyme can account for the inhibitory effect of AA and possibly other endogenous (e.g. fatty acid) phosphatase inhibitors. The reason for the increase in myosin phosphatase activity and decrease in phosphorylase phosphatase activity in the presence of up to 20 ,AM AA and up to 100-150 PM oleic acid is unknown, although preliminary experiments suggest that it may be explained by dissociation of the 20-kDa subunit from the 130-kDa component.* However, the low concentrations of AA that activated HMM phosphatase in solution did not relax pre-contracted smooth muscle (unpublished observation), and 30 PM AA had a contractile effect, indicative of inhibition of M L C~O phosphatase (Fig. 2). The reason for this discrepancy, whether due to different experimental conditions or dissociation between force and MLCPO phosphorylation due to cooperativity , remains to be determined.
The question arises whether, as suggested in the scheme below, AA is also a Ca*+-sensitizing messenger released by agonists under physiological conditions. Agonist "t receptor + G-protein -T PLA, +t AA -J MLC?, phosphatase The concentrations of AA required for Ca2+ sensitization and for inhibition of MLCzo phosphorylation (present study) are similar to the increase (38-75 ,AM) in AA content of pancreatic islets stimulated with glucose (Wolf et al., 1991). Prolonged elevation (6-fold over 8 min) of AA following stimulation with vasopressin has been observed in a cultured smooth muscle cell line (Grillone et al., 1988). Ca2+ sensitization by agonists and GTPyS  is also similar to that by AA (present study), in being associated with a relatively long lag period preceding force development.
However, at room temperature (present study), the time course of the response to AA is much slower than the Ca'+-sensitizing effect of agonists. It will be necessary, therefore, to pursue the kinetics of AA transients (in progress) to determine whether the requisite increase in AA concentration occurs during agonistinduced Ca2+ sensitization in smooth muscle, and whether the slower time course of the effect of added AA than that of agonists reflects the slower diffusion and access of exogenous, rather than endogenous AA to its myosin light chain phosphatase target. It is also possible that transport of endogenous AA (or other lipid) is accelerated by a fatty acid-binding protein. Finally, as in the case of G-protein-mediated MLCzo phosphatase inhibition (Kitazawa et al., 1991), we suggest that inhibition of protein phosphatase(s) by AA and/or other fatty acid messengers may also have regulatory functions in other cell systems.