The calcium-dependent association and functional coupling of calmodulin with myocardial phospholipase A2. Implications for cardiac cycle-dependent alterations in phospholipolysis.

Herein we demonstrate the calcium-dependent regulation of myocardial phospholipase A2 activity, which is mediated by a cytosolic protein constituent that can be chromatographically resolved from, and subsequently reconstituted with, purified myocardial phospholipase A2. Purification of this protein by sequential column chromatographies revealed an 18-kDa doublet, which was identified as calmodulin by Western blotting, calcium-dependent precipitation with W-7 agarose beads, and reconstitution of calcium-mediated phospholipase A2 inhibition with authentic homogeneous calmodulin. Calcium-induced calmodulin-mediated inhibition of myocardial phospholipase A2 was titrated by physiologic increments of calcium ion (Kd approximately 200 nM). Moreover, ternary complex affinity chromatography with calmodulin-Sepharose demonstrated that inhibition of myocardial phospholipase A2 activity by calmodulin resulted from the direct interaction of calmodulin with the myocardial phospholipase A2 catalytic complex. Exposure of cultured A-10 muscle cells to three structurally disparate calmodulin antagonists (W-7, trifluoperazine, and calmidazolium) resulted in the robust release of arachidonic acid, which was entirely ablated by pretreatment of cells with (E)-6-(bromomethylene)-3-(1-naphthalenyl)-2-H-tetrahydropyran-2-one. Collectively, this study identifies a novel mechanism whereby latent phospholipase A2 activity can be released from tonic inhibition by alterations in the interactions between the phospholipase A2 catalytic complex, calcium ion, and the intracellular calcium transducer, calmodulin.

In myocardium the majority of phospholipase A 2 activity is catalyzed by a calcium-independent phospholipase A 2 , which is selective for plasmalogen substrate containing arachidonic acid (1)(2)(3). Since this enzyme neither requires calcium as an oblig-atory cofactor in catalysis nor employs calcium for membrane association, it has traditionally been assumed that calcium ion does not directly regulate the activity of this enzyme (1)(2)(3)(4)(5)(6). However, alterations in calcium homeostasis play prominent roles in cardiac physiology, the predominant phospholipid constituents in the electrically active membrane of myocytes are plasmalogens containing arachidonic acid (7,8), and both reaction products of phospholipase A 2 catalysis are potent modulators of ion channel function (9 -12). Since prior results demonstrated that calcium ion could inhibit the activity of crude (i.e. cytosolic) myocardial phospholipase A 2 activity (1), we sought to identify a pathway that could integrate alterations in myocardial phospholipase A 2 activity with changes in myocytic calcium homeostasis and electrophysiologic function. Herein we describe a novel mechanism through which calcium ion regulates nominally "calcium-independent" myocardial phospholipase A 2 through its physical association with, and functional coupling to, the intracellular calcium transducer, calmodulin. We now report that myocardial phospholipase A 2 specifically and tightly binds to calmodulin in a calcium-dependent fashion, that this interaction is titrated over physiologic increments of calcium ion, that the interaction between calmodulin and the phospholipase A 2 catalytic complex regulates phospholipase A 2 activity in vitro, and that pharmacologic ablation of this interaction by three structurally disparate calmodulin antagonists results in the release of arachidonic acid by nominally "calcium-independent" phospholipase A 2 in intact muscle cells.
Purification of the Calcium-dependent Inhibitor of Myocardial Phospholipase A 2 -Ventricular myocardium from New Zealand White rabbits was placed in ice-cold buffer (250 mM sucrose, 10 mM imidazole, 10 mM KCl, 5 mM EDTA, pH 7.5) and homogenized utilizing a Potter-Elvehjem homogenizer. Homogenates were centrifuged at 20,000 ϫ g max for 20 min, and the resultant supernatant was heat-treated (90°C for 3 min) and rapidly cooled, and precipitated protein was pelleted by centrifugation (20,000 ϫ g max for 20 min). The resultant cytosol was dialyzed against 20 mM imidazole, 1 mM magnesium acetate, pH 7.5 (Buffer A) at 4°C and loaded onto a DEAE-Sephacel column (2.5 ϫ 8.0 cm), and adsorbed protein was eluted by a linear gradient of NaCl (100 mM-1 M NaCl). The calcium-dependent inhibitor was concentrated (Amicon Centriplus-10) and diluted 3-fold with buffer A, and 500 g of protein was loaded onto a PC 1.6/5 Mono-Q column (Pharmacia Biotech Inc.) prior to elution with a linear 1 M NaCl gradient in buffer A.
Enzyme Assays-Calcium-dependent inhibition of myocardial phospholipase A 2 was assessed by incubating partially purified phospholipase A 2 (150 g) with column fractions (25 l) or bovine brain calmodulin in a final volume of 200 l in 100 mM Tris-HCl, pH 7.0, containing either 10 mM CaCl 2 or 4 mM EGTA for 2 min at 25°C. Phospholipase A 2 activity was quantified by the release of [ 3 H]oleic acid from 1-O-(Z)-hexadec-1Ј-enyl-2-[9,10-3 H]octadec-9Ј-enoyl-sn-glycero-3-phosphocholine (plasmenylcholine) (2 M) as described previously (2). The * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  indicated free calcium ion concentrations employed in the assays were prepared using buffered CaCl 2 /EGTA solutions (17).
Calmodulin-Sepharose and W-7 Agarose Chromatography-Protein samples (500 M CaCl 2 final concentration) were loaded onto calmodulin-Sepharose columns pre-equilibrated with 50 mM Tris-HCl (pH 7.0) (buffer B) containing 500 M CaCl 2 . After application of 10 column volumes of equilibration buffer, columns were washed with buffer B containing 4 mM EGTA. For experiments involving W-7 agarose, 75 l of Mono-Q-purified calcium-dependent inhibitor was adjusted to either 1 mM CaCl 2 or 4 mM EGTA prior the addition of W-7 agarose equilibrated in 20 mM imidazole, pH 7.5. After a 60-min incubation at 4°C, the W-7 agarose was pelleted by centrifugation, and the pellet was washed (3 times) with 20 mM imidazole, pH 7.5, containing either 1 mM CaCl 2 or 4 mM EGTA prior to resuspension.

RESULTS AND DISCUSSION
Previously, we have demonstrated that calcium ion inhibits over 80% of crude myocardial cytosolic phospholipase A 2 activity (1) and have repeatedly observed this effect in multiple independent preparations (x ϭ ϳ96 pmol/mg⅐min in the presence of EGTA and ϳ 18 pmol/mg⅐min in the presence of 10 mM calcium ion). During the course of investigating this phenomenon, we observed that highly purified preparations of myocardial phospholipase A 2 did not manifest the calcium-mediated inhibition of enzymic activity, which was present in the crude cytosolic fraction (Fig. 1). The calcium-mediated inhibition of purified phospholipase A 2 activity could be reconstituted by the addition of heat-treated cytosol (which did not contain meas-urable phospholipase A 2 activity) to highly purified preparations of myocardial phospholipase A 2 (Fig. 1A). Furthermore, incubation of heat-treated cytosol with trypsin completely ablated its ability to reconstitute calcium-mediated inhibition of purified myocardial phospholipase A 2 activity. Collectively, these experiments demonstrate that a heat-stable protein constituent present in myocardial cytosol was responsible for the calcium-mediated inhibition of myocardial phospholipase A 2 activity.
To identify the protein constituent mediating these effects, heat-treated cytosol was subjected to anion exchange chromatography. A single, well resolved peak of calcium-dependent inhibition was identified (Fig. 1B), pooled, and subjected to Mono-Q chromatography (Fig. 1C). The calcium-dependent inhibitor of myocardial phospholipase A 2 eluted as a single, well resolved UV peak at ϳ300 mM NaCl, which contained an 18-kDa doublet that precisely cochromatographed with the observed calcium-dependent inhibition of myocardial phospholipase A 2 (Fig. 1D). SDS-PAGE in the presence of EGTA demonstrated the collapse of the doublet to a single band.
Since calmodulin is a heat-stable, acidic 18-kDa protein that migrates as a doublet on SDS-PAGE in the presence of calcium ion and a single band in the presence of EGTA (18,19), the active fractions from Mono-Q chromatography were further analyzed by Western blotting utilizing antibodies directed FIG. 1. Characterization and purification of the calcium-dependent inhibitor of myocardial phospholipase A 2 . A, the presence and trypsin sensitivity of the cytosolic calcium-dependent inhibitor of myocardial phospholipase A 2 were assessed in the absence or presence of calcium after the addition of heat-treated cytosol to highly purified myocardial phospholipase A 2 as described under "Experimental Procedures." B, the supernatant from heat-treated myocardial cytosol was chromatographed on a DEAE-Sephacel column, and individual column fractions were assayed for the calcium-dependent inhibition of myocardial phospholipase A 2 as described under "Experimental Procedures." E, 10 mM CaCl 2 ; q, 4 mM EGTA. C, the calcium-dependent inhibitor from DEAE-Sephacel chromatography was loaded onto a Mono-Q column prior to elution utilizing a linear gradient of 1 M NaCl as described under "Experimental Procedures." E, 10 mM CaCl 2 ; q, 4 mM EGTA. D, proteins from Mono-Q chromatography were resolved by SDS-PAGE and subjected to either silver staining (top) or Western blotting utilizing mouse monoclonal anti-calmodulin-IgG 1 (bottom) as described under "Experimental Procedures." Calmodulin Modulates Myocardial Phospholipase A 2 20990 against calmodulin. The 18-kDa protein doublet, which cochromatographed with calcium-dependent inhibition, was recognized by anti-calmodulin monoclonal antibody (Fig. 1). Moreover, authentic bovine brain calmodulin both cochromatographed with the homogeneous calcium-dependent inhibitor of myocardial phospholipase A 2 (utilizing identical conditions for Mono-Q chromatography) and inhibited myocardial phospholipase A 2 activity in a calcium-dependent fashion.
To further substantiate the identity of the calcium-dependent inhibitor of myocardial phospholipase A 2 as calmodulin, W-7 agarose affinity resin was employed (20). W-7 agarose beads bound the Mono-Q-purified calcium-dependent inhibitor in the presence of calcium ion, and the inhibitor was completely released by subsequent incubation with 4 mM EGTA ( Fig. 2A). No binding was manifest in the presence of buffer containing 4 mM EGTA. The partitioning of the protein mediating calciumdependent inhibition onto W-7 agarose in the presence of calcium ion and its subsequent release with EGTA correlated with the amount of calcium-dependent inhibition of phospholipase A 2 (Fig. 2, B and C). Since the 18-kDa polypeptide, which cochromatographed with calcium-dependent inhibition of myocardial phospholipase A 2 , was: 1) recognized by antibodies directed against calmodulin; 2) cochromatographed with authentic calmodulin; 3) bound to W-7 agarose beads in a calciumdependent fashion; 4) possessed identical electrophoretic characteristics as authentic calmodulin; and 5) calmodulin entirely reproduced the calcium-dependent inhibition of myocardial phospholipase A 2 , we conclude that the cytosolic protein constituent mediating calcium-dependent inhibition of myocardial phospholipase A 2 was calmodulin.
To determine the calcium dependence of calmodulin-mediated inhibition of myocardial phospholipase A 2 activity, measurements of phospholipase A 2 activity were conducted with highly purified myocardial phospholipase A 2 (50,000-fold purified; 10 ng of protein) and authentic calmodulin. Half-maximal inhibition was manifest at 200 nM calcium ion (Fig. 2D), which closely parallels the K d of calcium ion for calmodulin (21-23). The calcium-induced calmodulin-mediated inhibition was reversible by subsequent chelation of calcium ion.
To distinguish between the potential mechanisms that were responsible for the calmodulin-induced alterations in myocardial phospholipase A 2 activity (i.e. phospholipid substrate sequestration versus a direct interaction between calcium-activated calmodulin and myocardial phospholipase A 2 ) calmodulin-Sepharose chromatography was employed (24,25). The application of myocardial cytosol (in buffer containing 500 M CaCl 2 ) to a calmodulin-Sepharose affinity column resulted in the adsorption of less than 5% of the applied protein while myocardial phospholipase A 2 activity was completely adsorbed. Phospholipase A 2 activity was quantitatively desorbed by application of buffer containing 4 mM EGTA (Fig. 3A). The load and the void contained an identical protein banding pattern and protein masses (39 versus 38.5 mg for load and void, respectively), while the EGTA eluent contained only 0.5 mg of protein, which possessed a completely different banding pattern (Fig. 3B). Myocardial cytosolic phospholipase A 2 activity applied to a calmodulin-Sepharose column in the presence of EGTA was not adsorbed, demonstrating that activated calmodulin was required for association with the myocardial phospholipase A 2 catalytic complex. Moreover, 100-fold increases in substrate concentration (from 2 to 200 M) did not attenuate calmodulin-mediated phospholipase A 2 inhibition.
To determine whether the interaction between activated calmodulin and myocardial phospholipase A 2 present in myocardial cytosol was an inherent component of the myocardial phospholipase A 2 complex, highly purified myocardial phospholipase A 2 (Ͼ50,000-fold purified from the ATP affinity chromatography eluent in Ref. 2) was utilized. Similar to results with crude fractions, activity was quantitatively absorbed in the presence of calcium ion and desorbed by washing with 4 mM EGTA. For comparison, highly purified recombinant 85-kDa calcium-dependent phospholipase A 2 was not adsorbed onto calmodulin-Sepharose.
To determine the biologic significance of calmodulin-mediated regulation of phospholipase A 2 in intact cells, we examined the effects of three structurally disparate calmodulin antagonists on arachidonic acid release in cultures of A-10 muscle cells, a cell line that predominantly contains "calcium-independent" phospholipase A 2 (14). Exposure of prelabeled A-10 muscle cells to W-7 resulted in over an 8-fold increase in [ 3 H]arachidonic acid release (Fig. 4). Furthermore, half-maximal effects were manifest at a concentration of W-7, which parallels the K d for binding of W-7 by calmodulin (26 -28). Moreover, preincubation of A-10 muscle cells with BEL prior to the addition of W-7 ablated the release of [ 3 H]arachidonic acid (Fig. 4). Finally, utilization of two structurally disparate calmodulin antagonists, calmidazolium and trifluoperazine, resulted in similar increases in [ 3 H]arachidonic acid release from A-10 cells with effective concentrations that approximated the K d for the binding of calmidazolium and trifluoperazine to calmodulin (Ͻ1 M and ϳ1 M, respectively) (29 -33). No significant differences in [ 3 H]arachidonic acid incorporation into phospholipids were manifest in the presence of W-7.
The results of this study identify the physical association of calmodulin with the myocardial phospholipase A 2 catalytic complex, the role of calcium ion in facilitating this association, and the importance of this interaction in modulating phospholipase A 2 activity in intact muscle cells. Since calcium ion is neither an obligatory cofactor in catalysis nor necessary for the membrane association of this enzyme (in contrast to secretory phospholipase A 2 (34) and calcium-dependent phospholipase A 2 (35)(36)(37)), it has traditionally been assumed that calcium ion does not play an important role in modulating the activity of this enzyme in intact cells (1)(2)(3)(4)(5)(6). These results clearly demonstrate the importance of calcium ion and the intracellular calcium transducer, calmodulin, in regulating the activity of nominally "calcium-independent" myocardial phospholipase A 2 . The calmodulin antagonists resulted in release of [ 3 H]arachidonic acid from cellular phospholipids even at ambient cytoso-lic calcium concentrations (Ϸ100 nM) implying that at least a portion of the calmodulin⅐phospholipase A 2 catalytic complex is likely present in subcellular loci containing calcium ion in concentrations greater than that present in the cytosolic compartment (e.g. membranous intracellular calcium pools, subsarcolemmal locations at or near calcium channels). During the cardiac cycle, concentrations of calcium vary from 150 nM during diastole to 600 nM during systole (38 -40). The electrophysiologically active membrane of myocytes is highly enriched in plasmalogen molecular species (7,8) containing arachidonic acid (the preferred substrate of this enzyme), and nonesterified arachidonic acid is a potent modulator of ion channel function (9 -12). Thus, these results constitute the first description of a biochemical mechanism that can mediate the cardiac cycle-dependent alterations in phospholipase A 2 -catalyzed liberation of arachidonic acid, thereby modulating the function of critical sarcolemmal proteins in a cardiac cycle-dependent fashion.
FIG. 3. Calmodulin-Sepharose chromatography of phospholipase A 2 from myocardial cytosol. A, dialyzed rabbit heart cytosol (39 mg of protein) was loaded onto a 1-ml column of calmodulin-Sepharose, myocardial phospholipase A 2 activity was eluted by the application of buffer containing 4 mM EGTA, and phospholipase A 2 activity was quantified as described under "Experimental Procedures." The void volume contained 38.5 mg of total protein, and the EGTA 1 fraction contained 0.5 mg of total protein. B, calmodulin-Sepharose column fractions were individually diluted to a final concentration of 100 g/ml protein with 10% SDS and 0.05% ␤-mercaptoethanol, and the proteins were resolved on a 10% polyacrylamide gel and subsequently stained with silver. Calmodulin-Sepharose column fractions were assayed for calcium-dependent phospholipase A 2 activity as described under "Experimental Procedures."