Bidirectional control of phospholipase A2 activity by Ca2 /calmodulin-dependent protein kinase II, cAMP-dependent protein kinase, and casein kinase II

In preparations of synaptic terminals (synap-tosomes) isolated from rat brain, the activity of phospholipase A2 (PLA2), a phospholipid hydrolase that serves a central function in signal transduction, was inhibited in a Ca2+-dependent manner by incubation with 60 mM K+ or with the Ca2+-selective ionophore ionomycin. Reversal by alkaline phosphatase treatment suggested that this inhibitory effect resulted from phosphorylation of a synaptosomal protein substrate. When lysed synaptosomes were incubated with Ca2+/calmodulin (CaM), purified Ca2+/CAM-dependent protein kinase II (Ca2+/CaM-dependent PK H) and ATP, PLA2 activity in lysates was nearly abolished within 10 min. This effect was accompanied by a marked decrease in the V. of the enzyme and little or no change in the Km. Furthermore, Ca2+/CaM with ATP but without exogenous Ca2+/CaM-dependent PK II partially inhibited PLA2 activity, and this effect was prevented by treating the lysates with a selective peptide inhibitor of Ca2+/CaM-dependent PK I. In contrast, incubation of intact synaptosomes with 4B-phorbol 12- myristate 13-acetate or of lysed synaptosomes with purified protein kinase C had little or no effect on PLA2 activity. The results srgly s t that the Ca2+-dependent inhibition of PLA2

effect was accompanied by a marked decrease in the V. ofthe enzyme and little or no change in the Km. Furthermore, Ca2+/CaM with ATP but without exogenous Ca2+/CaMdependent PK II partially inhibited PLA2 activity, and this effect was prevented by treating the lysates with a selective peptide inhibitor of Ca2+/CaM-dependent PK I. In contrast, incubation of intact synaptosomes with 4B-phorbol 12myristate 13-acetate or of lysed synaptosomes with purified protein kinase C had little or no effect on PLA2 activity. The results srgly s t that the Ca2+-dependent inhibition of PLA2 activity observed In intact nerve endings was produced by activation of the m Ca2+/CaM-dependent PK II. A membrane-permeable adenylyl cyclase activator, forskolin, enhanced PLA2 activity In intact synaptosomes, and cAMPdependent protein kinase potentiated PLA2 activity in lysed synaptosomes. Furthermore, another broad-spectrum protein kinase present in synaptic terminals, casein kinase H, also potentiated PLA2 activity in lysed synaptosomes. The effects of both protein kinases were associated with a decrease in K. and no change in V,.. The results suggest that PLA2 activity in synaptic terminals is subject to bidirectional control by distinct signal transduction pathways. Moreover, mutually antagonistic effects of the Ca2+/CaM-dependent PK H and PLA2 pathways provide a possible molecular mechanism for bidirectional modulation of neurotransmitter release. Phospholipase A2 (PLA2) catalyzes the receptor-stimulated hydrolysis of membrane phospholipids, providing substrates for the biosynthesis of potent cellular regulators, which include arachidonic acid metabolites (eicosanoids) and platelet-activating factor (1)(2)(3). In neurons, the eicosanoids may participate in various forms of signal transduction and synaptic plasticity, including K+ channel modulation, presynaptic inhibition, and long-term potentiation (for review, see ref. 4). Despite these important functions in intracellular signaling, the mechanism of regulation of PLA2 activity is only poorly understood. In many tissues, a membranebound, Ca2+-dependent form of PLA2, which is thought to be responsible for the receptor-mediated liberation of arachidonic acid, is stimulated directly by free Ca2l (1,2). The high (millimolar) concentration of Ca2l required for stimulation in vitro suggests, however, that a different mechanism may be responsible for PLA2 activation in intact cells. In many tissues, guanine nucleotide-binding proteins (G proteins) coupled to neurotransmitter and hormone receptors have been shown to activate PLA2 and promote arachidonic acid liberation (5), possibly by reducing the enzyme requirement for free Ca2' (6,7). In agreement with this interpretation, G protein-mediated activation of PLA2 by certain hormones was shown to occur independently of hormone-induced increases in Ca2' (8).
In this study, we have investigated the possible role of protein phosphorylation in the regulation of PLA2 activity in neural tissue. Using a preparation of intact synaptic nerve endings, we found that membrane depolarization produces a Ca2+and phosphorylation-dependent inhibition of PLA2 activity. Studies using purified protein kinases indicate that this inhibitory effect results from activation of the multifunctional Ca2+/calmodulin-dependent protein kinase II (Ca2+/ CaM-dependent PK II). Conversely, two other broadspectrum protein kinases, casein kinase II and cAMPdependent protein kinase, were found to potentiate PLA2 activity in synaptic endings.
EXPERIMENTAL PROCEDURES Synaptosome Preparation. Synaptosomes, prepared from rat cerebral cortex by discontinuous density gradient centrifugation as described (9), were resuspended in oxygenated Hepes-buffered saline (142 mM NaCI/2.4 mM KCl/1.2 mM K2HPO4/1 mM MgCl2/5 mM D-glucose/10 mM Hepes buffer, pH 7.4) to 1 mg of protein per ml. Standard incubations were carried out at 37°C in the presence of 1 mM CaCl2 and stopped by the addition of 5 volumes of an ice-cold hypotonic buffer (lysis buffer) containing 1 mM EDTA, 1 mM EGTA, 10 mM sodium pyrophosphate, 10 mM Hepes (pH 7.4), and protease inhibitors (10). PLA2 Assay. The fluorogenic PLA2 substrate 1,2-bis-(1pyrenedecanoyl) phosphatidylcholine (2 ,uM for standard assays) was added to 1.5 ml of 0.1 M Tris-HCI (pH 9.0) in a stirred quartz cuvette maintained at 37°C within a Perkin-Elmer LS-5B luminescence spectrometer. After 2 min, samples of the lysed synaptosomes (0.2 mg of protein) were transferred to the cuvette, and the rate of increase in fluorescence due to the liberation of pyrenedecanoic acid was monitored (excitation wavelength, 340 nm; emission wave-length, 380 nm). The changes in fluorescence due to released pyrenedecanoate were calibrated against known concentrations of the unesterified fatty acid. Both Ca2l-dependent and Ca2l-independent PLA2 activities were present in the lysed tissue. Ca +-dependent PLA2 activity was maximal at alkaline pH (pH optimum, 9.0-9.5) in the presence of 2 mM Ca2 . In addition, this activity was inhibited by the histidine reagent p-bromophenacyl bromide (10 AtM) and entirely associated with membranes. In contrast, Ca2+-independent PLA2 activity was optimal at acidic pH (4-4.5), 1/10th as sensitive to inhibition by p-bromophenacyl bromide, and distributed approximately equally in both soluble and particulate fractions. Ca2+-independent and Ca2+-dependent PLA2 activities were measured before and after the addition of Ca2+ (2 mM final concentration), respectively. Ionomycin, forskolin, and 4/phorbol 12-myristate 13-acetate were added to the synaptosomes from dimethyl sulfoxide stock solutions kept at -200C. 1,2-Bis-(1-pyrenedecanoyl) phosphatidylcholine and pyrenedecanoic acid were stored in dimethyl sulfoxide at -20°C for up to 1 month. Final dimethyl sulfoxide concentration in the assay buffer was 0.2%. No change in activity occurred in lysed synaptosomes that had been incubated in control conditions in the absence or presence of this concentration of dimethyl sulfoxide for up to 10 min.

RESULTS
The effect ofdepolarization ofintact nerve terminals on PLA2 activity, measured subsequently in synaptosomal lysates, is shown in Fig. 1

(line a). K+-induced depolarization reduced
Ca2+-dependent PLA2 activity to 50o of its initial value within 1 min of incubation. Inhibition was complete within 10 min. When the synaptosomes were depolarized with K+ in the absence of external Ca2 , PLA2 activity remained unchanged ( Fig. 1, line d). When intact synaptosomes were incubated with the Ca2+ ionophore ionomycin (10 AM), PLA2 activity was inhibited to an extent comparable to that observed with the elevated K+ level and with a similar time course (Fig. 1, line b). In contrast to these actions on Ca2+-dependent PLA2 activity, neither K+ nor ionomycin had any effect on Ca2'-independent PLA2 activity; thus, after incubation of synaptosomes with Ca2+ and a high level of K+ or ionomycin for 10 min, Ca2+-independent PLA2 activity was 110 ± 9% of control (n = 8) or 114 ± 30%o of control (n = 3), respectively.
The depolarization-induced inhibition of PLA2 activity might have been caused by Ca2+-dependent protein kinase, Ca2+-dependent protein phosphatase, or Ca2+-dependent protease activity. To determine whether protein phosphorylation was responsible for this effect, lysates of K+depolarized synaptosomes (10 min) were incubated in the presence of alkaline phosphatase before measuring PLA2 activity. Ca2+-dependent PLA2 activity, which had been undetectable (the 10-min time point of line a in Fig. 1 returned upon treatment with alkaline phosphatase to the value of control, unstimulated synaptosomes (Fig. 1, point f). Alkaline phosphatase treatment did not affect PLA2 activity of control synaptosomes (98 ± 8%, n = 3). These results suggest that inhibition of PLA2 activity results from Ca2+dependent phosphorylation of a protein substrate rather than from the activation of a protease or a protein phosphatase. In synaptic terminals, voltage-dependent or receptorregulated Ca2+ influx activates two broad-spectrum protein kinases, Ca2+/CaM-dependent PK II and Ca2+/phospholipiddependent PKC (14,15). No specific, membrane permeable activator of Ca2+/CaM-dependent PK II has been described. Therefore, to test the possibility that this protein kinase participates in the K+-induced regulation of PLA2 activity, we incubated lysed synaptosomes in the presence of purified Ca2+/CaM-dependent PK II (in a reaction medium containing Ca2+, CaM, and ATP). Within 2 min of incubation, PLA2 activity was reduced to 50%o of its control value (Fig. 2, line  a). Almost complete inhibition of activity occurred after 10 min of incubation. A kinetic analysis indicated that Ca2"/ CaM-dependent PK II inhibited PLA2 activity by decreasing the Vm. of the enzyme, with no significant change in the apparent Km (Table 1).
PLA2 activity was not affected when Ca2+/CaM, ATP (Fig.  3), Ca2+, or CaM (not shown) was added separately to lysed synaptosomes. In contrast, addition of a combination of Ca2+, CaM, and ATP resulted in a 70%o inhibition of Ca2+dependent PLA2 activity, most likely through activation of endogenous Ca2W/CaM-dependent PK II (which is abundant in synaptic terminals; ref. 15). The inhibition by Ca2+, CaM, and ATP in combination could be enhanced further by the addition of exogenous Ca2+/CaM-dependent PK II (Fig. 3) but not by the addition of Ca2+/CaM-dependent PK I or Ca2+/CaM-dependent PK III (data not shown). Furthermore, inhibition of PLA2 by combined Ca2+ CaM, and ATP was prevented by treatment with a selective peptide inhibitor of Effects of purified protein kinases on Ca2+-dependent PLA2 activity in lysed synaptosomes. Synaptosomes, purified and lysed as described, were incubated at 370C in 0.1 ml of lysis buffer containing 5 mM magnesium acetate, 0.5 mM ATP, and one of the following protein kinases: "10 nM Ca2+/CaM-dependent protein kinase II with 1 mM CaCI2 and 1 ,uM CaM (curve a), 100 nM PKC with 1 mM CaC12 and 1 ,uM 4jB-phorbol 12-myristate 13-acetate (line b), 150 nM catalytic subunit of cAMP-dependent protein kinase (line c), or 10 nM casein kinase II (curve d). At the indicated times PLA2 activity was measured. Results were calculated as the percentage of activity in the absence of added protein kinase and are expressed as means ± SEM (vertical bars) of three to nine experiments. Ca2+dependent PLA2 activity was 6.5 ± 1 nmol/min per mg of protein (n = 27) at 0 min and was not affected by incubation for 10 min under control conditions (not shown). Ca2+/CaM-dependent PK II (13) (Fig. 3). Addition of exogenous Ca2+/CaM-dependent PK II in the absence of Ca2+/ CaM/ATP did not inhibit PLA2 activity (Fig. 3).
Ca"+-dependent PLA2 activity of rat cortical synaptosomes is entirely membrane bound (see Experimental Procedures). We prepared membranes from lysed synaptosomes and incubated them in the presence of purified Ca2+/CaMdependent PK II, Ca2W/CaM, and ATP. After 10 min of incubation (37°C) membrane-associated PLA2 activity was reduced from 44 ± 3 to 14 ± 0.4 nmol/min per mg of protein (n = 4). In contrast to these results with synaptosomal PLA2, Ca2+/CaM-dependent PK II had no effect on an extracellular form of Ca2+-dependent PLA2, purified from porcine pancreas (data not shown). Lysed synaptosomes were incubated for 10 min at 37°C in lysis buffer (1 mg of protein per ml) supplemented with 5 mM magnesium acetate, 0.5 mM ATP, and the purified protein kinase (PK) indicated.
Incubations with Ca2+/CaM-dependent PK II were carried out in the presence of 1 mM Ca2+ and 1 AM CaM. PLA2 activity was measured as described with concentrations of substrate ranging from 2 to 10 ,uM. Values are means ± SEM of three to eight experiments. *To determine the kinetics of inhibition by Ca2+/CaM-dependent PK II, an amount of kinase was used that resulted in only partial (-65%) inhibition under Vma,, conditions. Activators of PKC were previously shown to enhance PLA2 activity in various tissues (16,17). When we incubated intact synaptosomes with the tumor-promoting phorbol ester 4.3-phorbol 12-myristate 13-acetate (1 jM), which is very potent in stimulating PKC-dependent protein phosphorylation in intact synaptosomes (14), only a small and transient inhibition of Ca2+-dependent PLA2 activity was seen (Fig. 1,  line c). Moreover, purified PKC had no effect on PLA2 activity in synaptosomal lysates (Fig. 2, line b). These findings indicate that PKC-dependent protein phosphorylation does not contribute to the Ca2l-dependent modulation of PLA2 activity in synaptic nerve endings.
To investigate further the role of protein phosphorylation in the regulation of PLA2 activity, we examined cAMPdependent protein kinase and casein kinase II, two other broad-spectrum protein kinases present in synaptic endings (12,18), for their effects on this phospholipase. To study the possible action of cAMP-dependent protein kinase on PLA2 activity in intact synaptosomes, we used the membranepermeable adenylate cyclase activator forskolin. When synaptosomes were treated with 50 jM forskolin, Ca2+dependent PLA2 activity was potentiated in 10 min to 160% of control (Fig. 1, line e). Consistent with this finding, the addition of the purified catalytic subunit of cAMP-dependent protein kinase to lysed synaptosomes enhanced PLA2 activity (Fig. 2, line c), an effect that was dependent on the presence of magnesium and ATP (not shown). Kinetic analysis of this response showed that cAMP-dependent protein kinase potentiated PLA2 activity by decreasing the Km, with little effect on the Vn. of the enzyme (Table 1). In contrast to Ca2+/CaM-dependent PK II, cAMP-dependent protein kinase did not affect PLA2 activity when added to isolated synaptosomal membranes (not shown).
Purified casein kinase II also enhanced PLA2 activity of synaptosomal lysates (Fig. 2, line d). This increase in enzyme activity was associated with a decrease in the Km of PLA2 for its substrate by a factor of 3 (Table 1). Like cAMP-dependent protein kinase, casein kinase II had no effect on isolated synaptosomal membranes (data not shown). repetitive nerve activity [which also increases intracellular Ca2+ (not shown)] may stimulate Ca2+/CaM-dependent PK II activity, enhance phosphorylation of synapsin 1 (a synaptic vesicle-associated phosphoprotein), and potentiate release of neurotransmitter evoked by a nerve impulse (for a review, see ref. 21). Conversely, neurotransmitters that activate PLA2 cause the formation of free arachidonic acid (AA) and its metabolites, which have been shown to enhance the opening of K+ channels and reduce neurotransmitter release (4,23,24). Thus, these two receptor-operated cascades, may exert opposing functions on transmitter release (solid arrows). In addition, they may also exhibit mutually antagonistic actions (broken arrows), with Ca2+/CaM-dependent PK II inhibiting PLA2 activity (present study) and arachidonic acid metabolites inhibiting Ca2+/ CaM-dependent PK II activity (11). These antagonistic effects would be expected to amplify the effects of the relevant receptors on neurotransmitter release.

DISCUSSION
Our experiments suggest that Ca2+ entry into presynaptic terminals inhibits Ca2+-dependent PLA2 activity through activation of Ca2+/CaM-dependent PK II and subsequent phosphorylation of a membrane protein substrate. This substrate may be PLA2 itself or an associated PLA2 regulatory protein (19,20). The ability of Ca2+/CaM-dependent PK II to inhibit PLA2 activity was unexpected in view of the stimulatory action of Ca2+ on this phospholipase (for review, see refs. 1 and 2). However, even though free Ca2+ is known to enhance PLA2 activity in vitro, increasing evidence now indicates that, in intact cells, receptor-dependent activation of PLA2 may occur via a G protein-mediated mechanism and may take place at physiological Ca2+ concentrations (5-8).
Our results further underscore the dissociation between increases in intracellular Ca2+ and PLA2 activation, suggesting that, in tissues containing high levels of Ca2+/CaMdependent PK II, such as the brain, the predominant effect of increasing Ca2+ levels may be to inhibit, rather than to stimulate, PLA2 activity.
A variety of evidence indicates that activation of Ca2+/CaMdependent PK II enhances neurotransmitter release (see ref. 21 for a review). For instance, purified Ca2+/CaM-dependent PK II injected into the presynaptic digit of the squid giant synapse (22) or introduced by freeze-thaw cycles into rat cortical synaptosomes (13) potentiated the depolarization-induced release of neurotransmitter. In contrast, arachidonic acid and its lipoxygenase metabolites, products ofPLA2 activity, have been shown to stimulate K+ channels and reduce neurotransmitter release in both mollusk and mammalian neurons (4,23). Thus, at least in certain instances, these two receptor-operated cascades appear to exert opposing actions on neurosecretion. The inhibition by Ca2+/CaM-dependent PK II of PLA2 activity, observed in the present study, and the reciprocal inhibition of Ca2+/CaMdependent PK II activity by arachidonic acid and its metabolites (11), products ofPLA2 activity, suggest that these two receptoroperated cascades exert mutually antagonistic actions. These actions, in turn, may contribute to the opposing effects ofthe two signal-transduction pathways on neurotransmitter release (Fig.  4).
Using lysed synaptosomes, we found that cAMPdependent protein kinase and casein kinase II potentiate the activity of Ca2+-dependent PLA2. In agreement, incubation with the catalytic subunit of cAMP-dependent protein kinase has been shown to increase PLA2 activity in lysed peritoneal macrophages (25). The lack of effect of cAMP-dependent protein kinase and casein kinase II on isolated synaptosomal membranes suggests that these protein kinases may act by stimulating the phosphorylation of a cytosolic PLA2modulating protein. The identity of the hypothetical activating factor(s) remains to be determined, however.
The opposing effects of Ca2+/CaM-dependent PK II, cAMP-dependent protein kinase and casein kinase II on the activity of Ca2+-dependent PLA2, shown in the present study, suggest that various signal-transduction pathways may have opposing effects on the activity of this phospholipase (and therefore on the release of arachidonic acid and platelet-activating factor) by modulating the activities of distinct protein kinases. Thus, neurotransmitters and neurohormones that activate cAMP-dependent protein kinase (26) or casein kinase 11 (27) may potentiate PLA2 activity and enhance arachidonic acid release, whereas repetitive nerve activity and neurotransmitters that activate Ca2+/CaMdependent PK 11 (15) may inhibit it. This bidirectional control of PLA2 activity may represent a final common pathway for interactions amongst a variety of first messengers.