Phosphorylation-dependent Regulation of Phospholipase A2 by G-proteins and Ca2+ in HL60 Granulocytes*

We studied the regulation of arachidonic acid (AA) release by guanosine 5‘-0-(3-thiotriphosphate (GTPyS) and Ca2+ in electropermeabilized HL60 granulocytes. Stimulation of AA release by GTPyS and Ca2+ was mediated by phospholipase Az (PLA2) and required the presence of MgATP (EC50: 100-250 p ~ ) . The nucleotide effects were Ca2+-dependent (maximal effects detected at 1 p~ free cation). UTP and ATPyS, which stimulate AA release in intact HL60 granulocytes with potencies and efficacies similar to those of ATP, were ineffective in supporting the effects of GTPrS in elec- tropermeabilized cells. Pretreatment with pertussis toxin affected stimulation of AA release by ATP in intact cell, without altering the nucleotide effects in permeabilized cells. We observed the protein kinase C-dependent phosphorylation of PLAz in permeabilized HL60 granulocytes, together with a correlation between the effects of phorbol esters and staurosporine on this reaction and on AA release. ATP-independent activation of PLAz by GTPyS and/or Caz+ was meas- ured in subcellular fractions prepared from HL60 granulocytes. These data appear consistent with a model in which PLAz activity in resting HL60 granulocytes is subjected to an inhibitory constraint that prevents its activation by Ca2+ and G-proteins. Re-moval of this

In HL60 cells (7) as well as in a variety of other cell types platelets, C62B glioma, and Chinese hamster ovary cells (8,9,11). Furthermore, a recently purified cytosolic PLAP from U937 cells undergoes substantial activation by Ca2+ at concentrations similar to those observed in stimulated cells (12)(13)(14). Elevated intracellular Ca2+, therefore, seems to constitute an indispensable factor in the activation of PLAz by agonists in a variety of cells. Despite these studies, however, it is still unclear whether the physiological elevation in [Ca"]i is sufficient to activate PLAz, as opposed to being necessary for the manifestation of receptor/G-protein-mediated effects on the enzyme activity. The relationship between PKC and PLAz activities has also received considerable attention (e.g. Refs. 5,6,[10][11][15][16][17][18][19][20]. In general, short term treatment of cells with PKC activators potentiates the release of AA stimulated by Ca2+ ionophores or agonists, while down-regulation of the kinase results in inhibition, or even complete abolishment, of AA release. There is evidence linking the differential activation of the a isoenzyme of PKC with AA release in Madin-Darby canine kidney cells (20). Although these studies undoubtedly involve PKC in the regulation of PLAz activity, little is known about the identity of the components which are modified as a consequence of the kinase activation.
In the present study we have studied the regulation of AA release in electropermeabilized HL60 granulocytes and observed that PKC-dependent phosphorylation of PLAz is associated with the activation of this enzyme by Ca2+ and Gproteins.  Culture and Differentiation of HMO Cells-HL6O cells were cultured and differentiated into a neutrophil-like phenotype, as described previously (7). Briefly, cells were routinely grown in IMDM supplemented with 25 mM HEPES, pH 7.4,35 mM NaHC03, and 10% calf bovine serum, under a 92.5% air, 7.5% CO, atmosphere at a density of 0.2-1.0 X IO6 cells/ml. Before differentiation, cells were adjusted to approximately 0.5 X IO6 cells/ml and grown for 12-24 h in serumfree IMDM supplemented with 5 pg/ml insulin, 5 pg/ml transferrin, 5 ng/ml sodium selenite, 100 units/ml penicillin, and 100 pg/ml streptomycin (IMDM with defined supplements) to a density of about 0.8 X IO6 cells/ml. Differentiation into a neutrophil-like phenotype was then induced by culturing the cells for 72 h in the presence of 0.5 mM Bt2cAMP.
Labeling of Cellular Phospholipids with PHIAA-At the end of differentiation, HL60 cells were harvested and resuspended in the IMDM with defined supplements (see previous paragraph) at a concentration of 3.5 X lo6 cells/ml and labeled at 37 "C for 1.5 h with 1 pCi/ml of [3H]AA (specific activity = 76 Ci/mmol, 13 nM final concentration). In some cases, cells were treated with PMA and/or staurosporine by adding them to the medium during the last 5-90 min (as indicated in the corresponding figure legends) of the labeling period. Such treatment did not affect the labeling of cellular phospholipids with [3H]AA.
Electropermeabilization of HL60 Cells-Electropermeabilized cells were prepared as described by Xie and Dubyak (21). Briefly, [3H]AAlabeled HL60 granulocytes were washed twice with Ca2+ and M efree ice-cold electropermeabilization buffer (EPB) consisting of 20 mM HEPES, pH 7.4, 120 mM potassium glutamate, 20 mM potassium acetate, 3 mM NaC1, 1 mM EGTA, 1 mg/ml glucose, and 3 mg/ml BSA. Cells were then resuspended in ice-cold EPB buffer at a density of 3 X 10' cells/ml. Permeabilization was induced by nine consecutive discharges (separated by 2-5 intervals, mixing the suspension after every three pulses) of an electric field (2.5 kV/cm), using a 4-pfarad capacitor to provide a time constant of about 100 ps. This resulted in the permeabilization of more than 95% of the cells for at least 1 h after the electric discharges.
Release of PHIAA in Intact or Electropermeabilized HL60 Cells- The protocol for the assay of [3H]AA release in intact cells was described in detail elsewhere (7). Briefly, the cells were washed twice with ice-cold Hank's balanced salt solution (HBSS) containing 1 mM M F , 1.6 mM Ca2+, 1 mg/ml glucose, and 3 mg/ml BSA, followed by resuspension in the same buffer. The reaction was started by mixing the cells with agonists and placing the assay tubes (1 X lo6 intact or permeabilized cells in 200 pl of final volume) into a 37 "C water bath, with gentle and constant shaking. EPB was used as assay buffer when measuring [3H]AA release in permeabilized cells. Immediately after electropermeabilization, cells were distributed into ice-cold test tubes containing the indicated concentrations of Ca2+, M$+, and agonists, and the reaction was started as described for intact cells. The assays for [3H]AA release in both intact and permeabilized cells were stopped following 8-and 20-min incubations, respectively, by 15-fold dilution with an ice-cold stop solution consisting of 50 mM Tris-HC1, pH 7.5, 100 mM KC], 5 mM EGTA, and 5 mM EDTA, followed by centrifugation for 20 min at 1400 X g and 4 "C. Two ml of the supernatants were added to 10 ml of scintillation liquid and counted. Unlabeled arachidonate (0.2 mM) was added to the incubation medium for measurement of AA release in intact and electropermeabilized cells (as well as subcellular fractions), in order to increase the assay sensitivity (7).
Release of PHlAA in Homogenates and Plasma Membrane Fractions Derived from HL60 Granulocytes-Differentiated and [3H]AAlabeled HL60 cells were washed twice with ice-cold Ca" and M Pfree EPB supplemented with protease inhibitors (2 p~ pepstatin, 2 pg/ml leupeptin, and 200 p~ PMSF, freshly added), followed by incubation on ice for 30 min in this solution with additional 4 mM DIFP. Cells were subsequently pelleted and resuspended in the above mentioned buffer without DIFP. The cell suspensions (3-4 X 10' cells/ml) were subjected to N, cavitation at 500 p.s.i. for 20 min on ice. Aliquots of the cavitate representing 1 X 10' cells were immedi-ately distributed to assay tubes containing test substances in EPB (final volume of reaction: 200 pl). Reactions were started as in assays using intact or permeabilized cells and terminated, after 20 minincubations at 37 "C, by addition of chloroform/methanol (L2, by volume). Lipids were extracted according to Bligh and Dyer (22). Free [3H]AA was separated by thin layer chromatography as described (7).
Plasma membranes were isolated according to Uhing et al. (23). The cell cavitates, obtained as described above, were centrifuged at 600 X g for 10 min to pellet nuclei and unbroken cells. Aliquots of 5 ml of the post-nuclear supernatants were layered on top of 8 ml of 43% (w/w) sucrose cushions and ultracentrifuged in a Beckmann SW 28.1 rotor at 100,000 X g and 4 "C for 60 min. The cytosolic fractions floating on top of the sucrose cushions were saved; the plasma membranes banding at the interface were washed by dilution in Ca2+and Mg2"free EPB supplemented with protease inhibitors, pelleted by ultracentrifugation (100,000 X g, 60 min at 4 'C), and resuspended in the same buffer for subsequent assays. The assay for AA release in subcellular fractions were performed as indicated for cell cavitates, using aliquots representing 2 X lo6 cells in a final volume of 200 pl.
Phosphoinositide-specific Phospholipase C Activity in HL60 Granulocyte Membranes-HL6O cells were simultaneously labeled with L-my0[2-~H]inositol and differentiated (7). Nitrogen cavitation, membrane isolation, and assays were carried out as described by Cowen et al. (24). Incubations were carried out for 6 min in the presence of 350 nM free Ca2+, 3 mM MgCl,, 10 mM LiC1, and the other specified additions. Sample extraction, column separation, and coelution of inositol bisphosphates and trisphosphates were as described (7).

Labeling of Electropermeabilized HL60 Granulocytes with [-y-3' P]
ATP-Differentiated HL60 cells were incubated at 37 "C for 1 h in the presence or absence of 1 p~ staurosporine added to the regular differentiation medium (see above), immediately before start of the experiments. Cells were subsequently washed twice with glucose-free HBSS and incubated with 5 p~ antimycin A and 6 mM 2-deoxy-~glucose for 5 min at 37 "C, to reduce the levels of endogenous ATP (25). Cells were then washed in EPB supplemented with 2 p~ pepstatin, 2 pg/ml leupeptin, 200 p M PMSF, 10 mM NaF, 4 mM Na3V04, 0.92 mM CaC1, (resulting in 1 p M free ca2+), and 3 mM MgC1, (EPB plus protease and phosphatase inhibitors: "EPB + PPI"). The washed pellets were resuspended in EPB + PPI further supplemented with 4 mM DIFP and incubated on ice for 30 min. Cells were centrifuged again, resuspended in EPB + PPI, and subjected to electropermeabilization as described above. Following electropermeabilization, aliquots representing 4 X lo7 cells were diluted in a final volume of 1 ml of EPB + PPI containing 60 pCi of [Y-~'P]ATP (specific activity = 4,500 Ci/mmol, final concentration = 13.2 nM), 50 p~ GTP+, with or without 100 nM PMA, and incubated on ice for 20 min. After further incubation at 37 "C for 10 min, cells were diluted 30-fold with ice-cold stop solution (as described for AA release assay) supplemented with protease and phosphatase inhibitors, and pelleted by centrifugation. Pellets were extracted by resuspension in 300 pl of ice-cold extraction buffer consisting of 50 mM Tris-HC1, pH 7.5, 1% Nonidet P-40, 150 mM NaCI, 1 mM EGTA, 1 mM EDTA, 10 mM NaF, 4 mM Na3V04, 2 p M pepstatin A, 2 pg/ml leupeptin, 200 p~ PMSF, and 2 mM DIFP (protease inhibitors added freshly), followed by incubation at 4 "C for 40 min. The samples were centrifuged for 10 min at 16,000 X g and 4 "C, and the supernatants (cell extracts) saved for the immunoprecipitation procedures.
Immunoprecipitation, Electrophoresis, and Autoradiography of Phosphorylated cPLAZ-Experiments were carried out using a modification of the procedure described by Carlin and Knowles (26). The first step consisted of "preclearing" the extracts from proteins interacting with irrelevant IgGs present in the rabbit anti-cPLA2 antiserum and/or with the protein A-coated Pansorbin cells. To this end, 20O-pl aliquots of 10% Pansorbin suspension (pretreated with 3% BSA) were combined with 20 pl of normal rabbit serum and incubated on ice for 2 h, followed by centrifugation. The Pansorbin cell pellets bearing antibodies unrelated to PLAz were washed with extraction buffer and resuspended in 200 p1 of cell extracts (or in 200 p1 of extraction buffer containing 50 ng of cPLA2 standard) and incubated for 2 h on ice, followed by centrifugation (16,000 X g, 10 min at 4 " C ) . The supernatants (precleared extracts) obtained after two cycles of preclearing were subsequently treated with 2 p1 of the rabbit anti-cPLAz antiserum and incubated on ice for 2 h. The antigen-antibody complexes thus formed were used to resuspend BSA-pretreated Pansorbin pellets (equivalent to 200 pl of 10% Pansorbin) and incubated on ice for 2 h, followed by centrifugation (12,000 X g, 2 min at 4 'C). The pellets, containing the immune complexes bound to protein A, were washed three times with 50 mM Tris-HC1 pH 7.5,350 mM NaCl,

25968
Phosphorylation-dependent Regulation of Phospholipase AP 6 mM EDTA, 0.5% Nonidet P-40, and supplemented with protease inhibitors, and once using this solution without Nonidet P-40. The washed pellets were resuspended in Laemmli buffer, boiled for 10 min, centrifuged at 16,000 X g for 10 min at room temperature, and the supernatants subsequently subjected to SDS-PAGE (10% acrylamide). At the end of electrophoresis, the gels were dried and exposed a t -70 "C using Kodak XAR films and intensifying screens. Western Blotting of cPLAz-The position of cPLA2 on the autoradiograms was determined by subjecting either cPLA2 standards (50 ng) or unlabeled neutrophil-like HL60 cells (3 X lo7 cells) to the extraction (in 200 p1 of extraction buffer), preclearing, immunoprecipitation, washes, and PAGE procedures described above (in parallel with processing of the electroporated cells incubated with [y3'P] ATP). Samples were transferred to nitrocellulose (pore size = 0.45 pm) by semi-dry electrophoresis in Towbin transfer buffer consisting of 192 mM glycine, 25 mM Tris, and 20% methanol (27), supplemented with 0.02% sodium azide. Non-specific binding sites in the blots were blocked first with 1% Tween-20 in PBS (25 mM NazHPOa, pH 7.4, 150 mM NaC1, and 0.02% sodium azide) for 30 min and, subsequently, using a 3% solution of BSA in PBS for 1 h at room temperature. The blocked filters were incubated first with a 1:lOOO dilution of rabbit anti-cPLA2 in 3% BSA in PBS for 1 h at room temperature, followed by washing and reaction for 2 h with 0.5 pCi/ml lZ5I-goat anti-rabbit IgG antiserum in a 3% solution of BSA in PBS. The membranes were subjected to repeated washings both, between incubations with primary and secondary antiserum, and after reaction with the secondary antiserum. Washing steps were carried out at room temperature using first 1% Tween-20 in PBS (three times, 5 min each) and then PBS (two times, 5 min each). Blots were subsequently air-dried and exposed, as described above. Bands in addition to PLA, were visible in all of the immunoprecipitated samples, corresponding to the recognition of rabbit IgG heavy and light chains (approximately 55 and 25 kDa, respectively) by the iodinated goat anti-rabbit IgG antiserum during Western blotting.
Preparation of EGTA-Bivalent Cation Buffers-Calculation of free Caz+ and M e in EGTA-containing buffers was performed using the FREECA computer program (28).
Presentation of Data-The release of [3H]arachidonate was expressed as the percentage of the total radioactivity incorporated into intact or electroporated cells (or their subcellular fractions). Data are representative of results obtained in two to four independent experiments.

RESULTS
Requirement of ATP f o r G T P y S and Ca2+-promoted AA Release in Electropermeabilized HL.60 Granulocytes-As shown in Fig. 1A GTPyS did not exert significant effects on AA release in electropermeabilized HL60 granulocytes incubated in the absence of ATP. However, when 1 mM ATP was included, addition of GTPyS resulted in a significant, Ca2+dependent, stimulation of AA release. Maximal activation by GTPyS was detected at about 1 pM Ca2+, consistent with [Ca2+]i levels in agonist-stimulated HL60 cells. The ATPdependent effect of G T P y S was concentration-dependent, displaying an EC50 of approximately 2 p~ (Fig. 1B). Addition of 80-300 p~ GTPyS resulted in the progressive inhibition of t h e maximal stimulatory effects detected at 10-50 p~ nucleotide. It is also worth noting that addition of Ca2+ (up t o 1 0 p~) did not elicit AA release in the absence of ATP. T h e effect of ATP was concentration dependent, displaying a n EC50 of approximately 0.25 mM, regardless of the presence or absence of G T P y S (Fig. 2 A ) . M 2 + was also required to support the effect of ATP as shown in and efficacies identical to those of ATP in stimulating AA release in intact HL60 cells (7), could not replace ATP in supporting the action of GTPyS and Ca2+ (Fig. 4) Furthermore, other nucleotides that are poor substrates for protein kinases, such as AMP(NH)P (29), did not support the effects of GTPyS and Ca2+ on AA release (Fig. 4). These results, together with the high ECso values observed for the ATPpermissive effect (Fig. 2 4 ) and the M$+ requirement (Fig.  2B), suggest that phosphorylation of certain factors may be was not attenuated by pretreatment with 100 nM PMA (Fig.  5B). In fact, both A1F;-and A23187-stimulated AA release were significantly potentiated by such pretreatment. These differential effects of PMA on agonist and Cap+ ionophoreinduced release of AA resemble those reported in human platelets (15) and rabbit neutrophils (16).
Pretreatment of HL60 granulocytes for 10 min with 100 nM PMA resulted in a significant increase in ATP and Cap+dependent release of AA measured after electroporation (Fig.  6). Addition of GTPyS resulted in further stimulation of AA release in permeabilized cells treated with PMA. The effects of PMA in permeabilized cells could be partially reversed by the pretreatment of cells with 1 p~ staurosporine, a PKCselective inhibitor which interacts with its catalytic domain. Staurosporine, per se, completely abolished the ATP-dependent AA release stimulated by GTPyS and Ca2+ in permeabilized cells (Fig. 6). Addition of 1 p~ staurosporine also resulted in inhibition of agonist-stimulated AA release in intact cells (data not shown). These results suggest a model of dual regulation of AA release by PKC: on one hand, activation of this enzyme blocks agonist-promoted (but not A1F;-or Ca2+ ionophore-dependent) release of AA, consistent with an impairment of the interaction between agonist-receptor complexes and G-proteins. On the other, PKC activation appears t o be required for G-protein-and Ca2+-driven release, as measured in electropermeabilized cells.
Ca2+ and GTPyS Induce AA Release in HL60 Cell Homogenates and Isolated Plasma Membranes in an ATP-independent Manner-Ca2+, per se, could induce a substantial release of AA in HL60 granulocyte cavitates (Fig. 7A). Maximal effect of Cap+ was obtained a t about 1 p~. Addition of 50 p~ GTPyS resulted in a 2-fold increase in the efficacy of AA release. The effect of GTPyS in cell homogenates was also Cap+ dependent, as measured in the presence of ATP in electropermeabilized cells. However, in contrast with the results obtained with that system, ATP was not required to support the effects of Ca2+ and GTPyS on the release of AA in the cell homogenates. Addition of ATP actually resulted in a small, but reproducible, inhibition of GTPyS-stimulated AA release in cell homogenates. The ATP-independent effects were also detected when measuring AA release by a combination of plasma membrane and cytosolic fractions derived from HL60 granulocytes (Fig.  7B). The requirement of cytosol was explained by the presence of cPLAz in this fraction, as measured in immunoblotting experiments using an anti-cPLAp antiserum (results not shown). Taking into account that the cavitation and homogenization procedures were carried in Ca2+-free buffers containing 1 mM EGTA, this finding was consistent with the Ca2+-dependent translocation of PLAz activity from membranes to cytosol observed in RAW 264.7 (30) and U937 cells (14, 31). A similar requirement of a combination of plasma membranes and cytosol was also observed when measuring Effects of PMA and staurosporine on AA release in electropermeabilized HL60 granulocytes. Cells were pretreated with or without 100 nM PMA or 1 WM staurosporine, or their combination, during the last 10 min of labeling with [3H]AA. These compounds were also present, in the corresponding conditions, throughout the washing procedures and the assay for AA release. Assays were carried out as described in the legend to Fig. 1; 1  phospholipase D activity toward endogenously labeled substrates in subcellular fractions derived from HL60 granulocytes (32).
The Release of AA in Electropermeabilized HL60 Cells and Subcellular Fractions Is Mediated Through PLA2--We previously showed that mepacrine, a widely used PLAz inhibitor, abolishes agonist-stimulated AA release, but not phospholipase C activity, in intact HL60 cells. This effect suggested the involvement of PLA2 in AA release in this cell type (7).
Mepacrine also inhibited AA release driven by ATP, Ca2+, and GTPyS in electropermeabilized HL60 cells (Fig. 8A). More importantly, pretreatment of cytosol and plasma membranes with specific antibodies directed against a cPLA2 cloned from U937 cells (12, 14) significantly attenuated the basal as well as GTPyS-and Ca2+-stimulated AA release (Fig.   8B). A similar treatment, using membranes from HL60 cells t h a t were prelabeled with [3H]inositol, did not result in inhibition of inositol phosphate release driven by receptors (chemotactic or purinergic), GTPyS or high Ca2+ (Fig. 8C). These results confirmed that release of AA in this system is mediated Granulocytes-The data presented so far suggested the hypothesis that PLA, is normally under a certain inhibitory constraint which prevents its activation by Ca2+ and Gproteins and that phosphorylation by a PKC-dependent mechanism, or physical disruption (e.g. by N P cavitation), can release this constraint. T o investigate the locus of this regulation, we examined the phosphorylation of cPLA2 by carrying out immunoprecipitations of extracts derived from HL60 granulocytes which were previously electropermeabilized and incubated with [y3'P]ATP. The autoradiogram (Fig. 9A) clearly evidenced the phosphorylation of cPLAz in HL60 granulocytes. Furthermore, in good agreement with their effects on AA release in permeabilized cells (Fig. 6), PMA and staurosporine significantly enhanced and inhibited, respectively, the phosphorylation of PLA2. The effect of PMA was substantially reversed by staurosporine, suggesting the spe-by PLAz.  A, cells were preincubated with (triangles) or without (circles) 50 pM mepacrine a t 37 "c for 10 min before electropermeabilization and AA release assay. AA release was assayed in the presence of 1 mM ATP, 3 mM MgCI2, 1 mM EGTA, and various concentrations of CaC12 (to generate the indicated calculated concentrations of free Ca") with (closed symbols) or without (open symbols) 50 p~ GTP$% B, plasma membrane and cytosolic fractions were isolated as described under "Experimental Procedures." The combined fractions were incubated in the absence of added serum (empty bars) or in the presence of 1:50 dilution of normal rabbit serum (filled bars) or rabbit anti-cPLA2 antiserum (hatched bars) for 80 min at 4 "C before assay. AA release was measured in the presence of 3 mM total MgC12 and the other indicated conditions. C, Phosphoinositidespecific phospholipase C activity in membranes (20,000 X g pellet) prepared from nitrogen cavitates of HL60 granulocytes prelabeled with ~-myo [2-~H]inositol. Membranes were preincubated for 80 min under the same conditions explained in B (same symbols are used), followed by assay of inositol phosphate release driven by receptor occupancy (UTP or fMLP plus GTP), receptor-independent G pro- dependency of the enzyme phosphorylation was also investigated (Fig. 10). The results showed that increasing free Ca2+ from 2 nM (as calculated when using buffers containing 1 mM EGTA and an estimated 10-20 PM total cation arising from contamination of different reagents) to 1 PM resulted in increased cPLA2 phosphorylation, either in the presence or absence of PMA. Addition of PMA resulted in stimulation of cPLA2 phosphorylation in the absence of significant free Ca", although to a level lower than that detected at 1 p M free cation.

DISCUSSION
In the present report we have studied the mechanism underlying the ATP-dependent stimulation of AA release by GTPyS and Ca2+ in electropermeabilized HL60 granulocytes.
The electropermeabilization technique used in the present study has many advantages over other procedures widely utilized to introduce substances into living cells (e.g. detergent permeabilization) (33,34). The relatively small pores induced by electric fields permits the introduction of small (Mr < 1000) intracellular substances but prevents the loss or dilution of enzymes and relevant regulatory proteins, problems which are frequently encountered in detergent-induced permeabilization. More importantly, the interaction of the detergents with the cholesterol groups present in the plasma membrane results in disruption of the lipid microenvironment. On the contrary, the effects of electric fields are only localized on cell sites oriented a t 0 and 180" with respect to the direction of the field, leaving other regions of the plasma membrane intact (33). This is relevant in view of the importance of normal lipid microenvironments in the molecular assembly of phospholipases, accessory regulatory proteins, and plasma memtein activation (GTP-yS), or 1 mM Ca2+. Release of inositol phosphates was expressed as percentage of total incorporated counts. For details see "Experimental Procedures." CTRL, control. We have applied this technique to the study of the regulation of PLA2 activity, under conditions that closely resemble the physiological situation, and observed the ATP dependence of Ca2+-and GTPyS-stimulated AA release. Our results agree with the report by Cockcroft (39) who also observed an ATPdependent stimulation of AA release by GTPyS in human neutrophils permeabilized with streptolysin 0. A previous study did not find evidence of ATP requirement when measuring PLA2-dependent stimulation of AA release by GTPyS in digitonin-permeabilized platelets (6). This may represent a n intrinsic characteristic of the platelet enzyme or an alteration in enzyme regulation caused by the use of detergents. In this regard it resembles the situation that we observed when measuring AA release by cell homogenates or isolated fractions derived from HL60 granulocytes, where ATP is not required for AA release (Fig. 7).

Phosphorylation-dependent Regulation
The dependence of AA release on ATP in electropermeabilized HL60 granulocytes represents a substrate requirement for the PKC-mediated phosphorylation of the enzyme. The evidence for this conclusion is summarized as follows: 1) the Echo for the effect of MgATP2-in electropermeabilized HL60 cells is relatively high (100-250 p~, Fig. 2, A and R ) , as opposed to the Echo observed when measuring stimulation of AA release by ATP in intact cells (3 p~, Ref. 7 ) ; 2) the effect of ATP in electropermeabilized cells is less sensitive to pertussis toxin than that in intact cells (Fig. 3); 3) UTP and ATPyS, which are equipotent to ATP in the stimulation of AA release in intact cells (7), are unable to support AA release in permeabilized cells; 4) unlike the allosteric regulation of a Ca"-independent PLA, in myocytes by ATP and its non-hydrolyzable analogues (41,42). AMP-P(NH)P, which is a poor substrat.e for protein kinases (Ref. 29), cannot support AA release (Fig. 4); 5) PMA enhanced, and staurosporine inhibited, ATP-dependent AA release in permeabilized cells (Fig. 6), suggesting the involvement of I'KC in the phosphorylation-dependent. regulation of AA release; 6) PLA, is phosphorylated by a PKC-dependent mechanism (Fig. 9), under the same conditions where AA release is induced by Ca2+ and GTPyS in the presence of ATP.
The results shown in Fig. 10 are worthy of discussion. On one hand PMA, stimulated cPLA, phosphorylation in elect ropermeabilized HL60 granuloc.ytes at nanomolar concentrations of Ca". We believe this could be attributed to the small but measurable fraction of phorbol ester receptors bound to membranes at resting Ca2+ (43). to the action of Ca'+-insensitive PKC isoforms (such as 6 and t ) , or to the reported ability of phorbol esters to drive PKC binding to membranes in buffers containing 1 mM EGTA and no added cation (although with a potency lower than that observed in its presence, Ref. 44). On the other hand, the observed phosphorylation of cPLA2 driven by PMA in the absence of added Ca2+ does not correlate with the low release of AA detected under this condition in the electroporated cells (Fig. 6). We take this observation as an indication that the phosphorvlation is necessaty but not sufficicnt for the expression of P I A 2 activity (which also requires the presence of Ca" and. prohably, direct G-protein coupling, as discussed below). This is in agreement with the observed requirement for Ca" when measuring in oitro the PLA, activity present in extracts derived from ATP-treated Chinese hamster ovarv cells overexpressing cPLA, (45). In these cells, the pretreatment with agonist produced a 100% shift in the SDS-PAGE mobility of cPLA2 (which is due to phosphorylation).
The release of AA in the systems used in the present studv is mediated by PLA2, as evidenced bv the inhibitorv effects of mepacrine and the anti-PLA, antiserum (Fig. 8. A and H ) . The effects of GTPyS on AA release can be at least partially attributed to the direct interaction of a G-protein with PIA,. as opposed to a mechanism wherebv G-protein-dependent activation of phospholipase C results in the generation of diacylglycerol and Cap+-mobilizing inositol phosphates, activation of PKC, and subsequent events leading to the phosphorylation and activation of PLA2. This conclusion is supported by the experiments using cell homogenates and isolated membranes, where ATP was not required to support AA release by GTPyS (Fig. 7, A and R). The fact that GTPyS could further enhance AA release in the presence of a saturating concentration of PMA (100 nM) also supports a direct interaction of a G-protein with PLA, (Fig. 6). In agreement with this notion is the demonstration of the differential regulation of phospholipase C and PLAp activity bv (;-proteins in HL60 granulocytes (2, 46).
These results are consistent with a model where PLA? is normally under an inhibitory constraint that prevents its activation by Ca2+ and G-proteins and that phosphorylation of PLA2 by a PKC-dependent mechanism, or phvsical disruption (e.g N2 cavitation) of the molecular assembly of the enzyme system, can release this constraint. We have also excluded the possibility that the N, cavitation procedure may cause the phosphorylation of the PLA, (as an explanation for the ATP-independent effects measured in homogenates and subcellular fract,ions). This was investigated hv comparing the levels of phosphorylation attained in permeahilized The above mentioned model can interpret the stimulatory effects brought about by PKC activation, and the inhibitory effects of its down-regulation, on AA release in various cell types. In view of the close correlation between the effects of activators and inhibitors of PKC on AA release in electropermeabilized HL60 cells (Fig. 6) and their effects on the phosphorylation of PLA, (Fig. 9), it is unlikely that phosphorylation of additional factors (such as lipocortins, Ref. 47) represents the only mechanism responsible for the release of the inhibitory constraint. In this context, Lin et al. (45) have recently observed that agonists and phorbol esters increase serine phosphorylation of cytosolic PLA, in Chinese hamster ovary cells and RAT-2 fibroblasts. The phosphorylation of PLA2 in HL60 granulocytes may be directly catalyzed by PKC (as observed in uitro for a cytosolic PLA? purified from rat mesangial cells)? or, alternatively, by a protein kinase or protein kinase cascade dependent on PKC activity. A likely candidate for a PKC-dependent pathway that may lead to PLAz phosphorylation is the MAPK kinase/MAPK cascade. This possibility is based on the reported involvement of PKC in the phosphorylation and activation of MAPK by mitogenic agonists (48,49), the serine/threonine phosphorylation-dependent activation of MAPK kinase (50), and on in uitro experiments showing the direct phosphorylation (and activation) of purified forms of PLA, by MAPK.'l3 Cockcroft et al. (2) observed that depletion of endogenous ATP completely abolished the stimulation of PLAp-mediated AA release by agonists in HL60 granulocytes and neutrophils. This result could be explained by the inability of the PKCdependent pathway to phosphorylate PLA, in the absence of ATP. Interestingly, Lu et al. (25) reported that under conditions similar to those in the present study, GTPyS-stimulated superoxide production was dependent on ATP and could be inhibited by staurosporine. In view of the stimulation of superoxide anion production by AA (51, 52), it seems reasonable to propose that the ATP dependence of GTPyS-stimulated superoxide anion generation reported by Lu et al. is linked to the PKC-dependent phosphorylation of PLA, reported here.
In summary, we have investigated the mechanism for the regulation of PLA2 by G-proteins and Ca2+. We propose that PLA, is maintained in situ under an inhibitory constraint that prevents its activation by G-proteins and Ca2+. This constraint can be released by PKC-dependent phosphorylation of PLA2. The role of Ca2+ in the regulation of PLA2catalyzed AA release is apparently 3-fold. First, it is involved in the activation of PKC and, hence, subsequent phosphorylation of PLA,. Second, once the inhibitory constraint is released, Ca2+ per se becomes directly stimulatory to PLA,, as evidenced by the ability of Ca2+ to stimulate PLA, in cell homogenates and isolated membranes (Fig. 7) and its ability to increase the activity of purified PLA, from rat mesangial (53) and U937 (12-14) cells in vitro. Third, Ca2+ is required for the activation of PLA, by G-proteins after the inhibitory constraint of PLA, is released (Figs. 1,6, and 7). The present study therefore suggests that the activation of PLA, by Gprotein-coupled receptors in intact cells requires the parallel activation of phospholipase C, with generation of diacylglycerol and Ca2+-mobilizing inositol phosphates. As a result, diacylglycerol, in the presence of Ca", activates PKC, which either directly, or through other protein kinases (e.g. MAPK), leads to the phosphorylation of PLAz and its activation by Gproteins and Ca2+. Fig. 11 illustrates this model.