A Calcium-dependent Mechanism for Associating a Soluble Arachidonoyl-hydrolyzing Phospholipase A2 with Membrane in the Macrophage Cell Line RAW 264.7*

Arachidonoyl-hydrolyzing phospholipase A2 plays a central role in providing substrate for the synthesis of the potent lipid mediators of inflammation, the eicosanoids, and platelet-activating factor. Although Ca2+ is required for arachidonic acid release in vivo and most phospholipase A2 enzymes require Ca2+ for activity in vitro, the role of Ca2+ in phospholipase A2 activation is not understood. We have found that an arachidonoyl-hydrolyzing phospholipase A2 from the macrophage-like cell line, RAW 264.7, exhibits Ca2(+)-dependent association with membrane. The intracellular distribution of the enzyme was studied as a function of the Ca2+ concentration present in homogenization buffer. The enzyme was found almost completely in the 100,000 x g soluble fraction when cells were homogenized in the presence of Ca2+ chelators and there was a slight decrease in soluble fraction activity when cells were homogenized at the level of Ca2+ in an unstimulated cell (80 nM). When cells were homogenized at Ca2+ concentrations expected in stimulated cells (230-450 nM), 60-70% of the phospholipase A2 activity was lost from the soluble fraction and became associated with the particulate fraction in a manner that was partly reversible with EGTA. Membrane-associated phospholipase A2 activity was demonstrated by [3H]arachidonic acid release both from exogenous liposomes and from radiolabeled membranes. With radiolabeled particulate fraction as substrate, this enzyme hydrolyzed arachidonic acid but not oleic acid from membrane phospholipid, and [3H]arachidonic acid was derived from phosphatidylcholine, phosphatidylethanolamine, and phosphatidylinositol/phosphatidylserine. We suggest a mechanism in which the activity of phospholipase A2 is regulated by Ca2+: in an unstimulated cell phospholipase A2 is found in the cytosol; upon receptor ligation the cytosolic Ca2+ concentration increases, and the enzyme becomes membrane-associated which facilitates arachidonic acid hydrolysis.

Arachidonoyl-hydrolyzing phospholipase AZ plays a central role in providing substrate for the synthesis of the potent lipid mediators of inflammation, the eicosanoids, and platelet-activating factor. Although Ca2+ is required for arachidonic acid release in vivo and most phospholipase A, enzymes require Ca2' for activity in vitro, the role of Ca" in phospholipase A2 activation is not understood.
We have found that an arachidonoyl-hydrolyzing phospholipase AZ from the macrophage-like cell line, RAW 264.7, exhibits Ca2+dependent association with membrane. The intracellular distribution of the enzyme was studied as a function of the Ca2+ concentration present in homogenization buffer.
The enzyme was found almost completely in the 100,000 x g soluble fraction when cells were homogenized in the presence of Ca2+ chelators and there was a slight decrease in soluble fraction activity when cells were homogenized at the level of Ca2+ in an unstimulated cell (80 nM We suggest a mechanism in which the activity of phospholipase AZ is regulated by Ca2+: in an unstimulated cell phospholipase A2 is found in the cytosol; upon receptor ligation the cytosolic Ca2+ concentration increases, and the enzyme becomes membraneassociated which facilitates arachidonic acid hydrolysis.
*This work was conducted in the F. L. Bryant, Jr. Research Laboratory for the Mechanisms of Lung Disease and supported by National Institutes of Health Grant HL 34303. The costs of publication of this article were defrayed in part by the payment oE page charges. This article must therefore be hereby marked "aduertisemrnt" in accordance with 18  Hydrolysis of the ether-linked phospholipid, l-0-alkyl-2-arachidonoyl-glycerophosphocholine, by phospholipase A2 also releases the precursor of another potent lipid mediator, platelet-activating factor. Hence, an understanding of the regulation of arachidonoyl-hydrolyzing phospholipase A2 may ultimately allow us to control for some of the deleterious aspects of inflammation.
The mechanisms involved in the regulation of intracellular phospholipase AZ enzymes in general are incompletely understood. A role for Ca2+ in arachidonoyl-hydrolyzing phospholipase AZ activation is strongly suggested by the following findings: incubation of permeabilized rat mesangial cells with increasing concentrations of Ca'+ resulted in increased arachidonic acid release (1); in inflammatory cells Ca '+ ionophore could induce arachidonic acid release (2, 3) and immune complex-and calcium ionophoreinduced arachidonic acid release was absolutely dependent on extracellular Ca2+ (3). Furthermore, many arachidonoyl-hydrolyzing enzymes showed an absolute requirement for Ca*+ in vitro (4)(5)(6)(7)(8). However, it is not known whether the role of Ca*' is to regulate the activity of the enzyme itself or to regulate some other transduction mechanism(s) leading to phospholipase A2 activation. In platelets (6) and mesangial cells (8), the intracellular distribution of arachidonoyl-hydrolyzing phospholipase AP activity has been shown to be dependent on the Ca*+ content of the buffer used for cell disruption.
When cells were homogenized in the presence of Ca"' chelators, phospholipase AS activity was almost entirely cytosolic (5)(6)(7)(8); in the presence of a high concentration of Ca2+ (>l mM), there was a loss of cytosolic activity and a concomitant increase in membrane-associated activity (6,8). These observations may reflect a regulatory role for Ca*+ in the activation of phospholipase A2 by promoting its association with membrane, thus facilitating arachidonic acid hydrolysis. However, the level of Ca2+ that is required for this phenomenon has not been investigated previously. We have described the purification of an arachidonoyl-specific phospholipase A* from the macrophage-like cell line RAW 264.7 (7). We now report that the intracellular distribution of this phospholipase A2 changes from cytosolic, when cells are homogenized in Ca'+ concentrations expected in unstimulated cells, to membrane-associated, when cells are homogenized at Ca*+ concentrations expected in stimulated cells. When membrane-associated, the phospholipase Al? was found to hydrolyze arachidonic acid, but not oleic acid, from membrane phospholipid. cultured as described previously (9). fractions were isolated, and the phospholipase A2 activity of each fraction was determined in the presence of 10 mM Ca2+ using a liposomal substrate. As shown in Fig. 1, soluble fraction phospholipase A2 activity was highest when cells were homogenized in the presence of 100 pM EGTA (2.6 f 0.3 nmol/min/mg; n = 6). Activity decreased only slightly when cells were homogenized in the presence of 78 nM Ca'+, which is the concentration of Ca2+ found in an unstimulated cell (13). A decrease in soluble fraction phospholipase A2 activity of 60 and 70% was seen when cells were homogenized in the presence of 235 and 450 nM Ca'+, respectively, the range of Ca2+ in a stimulated cell (13). Only a small additional decrease in soluble fraction phospholipase AP activity was seen when cells were homogenized in the presence of higher concentrations of Ca'+, demonstrating that the process was almost complete at physiological Ca*+ levels. For these experiments an homogenization buffer containing intracellular concentrations of Na' and K' was used to simulate the cytosolic environment of the enzyme. The possibility that these results could in part be due to depolarization caused by the high [K'] of this buffer was excluded as similar results were seen when cells were homogenized in 5 mM HEPES or in buffer containing 140 mM NaCl and 10 mM KC! (data not shown). Neither could this effect be explained by major differences in protein distribution following exposure to Ca*+ as approximately 50% of the total protein was found in the soluble fraction for each of the homogenates.
Similar results were obtained whether cells were homogenized by sonication or by nitrogen cavita- When cells were homogenized in the presence of 100 FM EGTA and then the homogenate was adjusted to 100 PM Ca2+ before ultracentrifugation, the same loss of soluble fraction phospholipase AZ was observed, suggesting an intact cell was not required for this phenomenon to be observed. We then tested the hypothesis that the decrease in soluble fraction phospholipase A, activity, seen when increasing amounts of Ca*+ were present during homogenization, was concomitant with an increase in membrane-associated phospholipase A, activity. A direct relationship between the decrease in soluble fraction phospholipase AS activity and the increase in membrane-associated activity was found (Fig. 1). Very little activity was detected in the particulate fraction when cells were homogenized in the presence of 100 PM EGTA, and highest activity was detected when cells were homogenized in the presence of 450 nM Ca'+. When cells were homogenized at Ca2+ concentrations greater than 230 nM, the activity recovered from the particulate fraction was less than the decrease in soluble fraction activity. Fig. 1 shows a maximum of 33% of the decrease in soluble fraction phospholipase Al activity recoverable from the particulate fraction when cells were homogenized in the presence of 450 nM Ca'+. Recovery was found to be variable and at best up to 75% of the decrease in soluble fraction phospholipase AP activity was recoverable on the particulate fraction under the optimal conditions described below. The inability to quantitatively recover the enzyme from the particulate fraction may be due to losses of enzyme activity sustained by experimental procedures, to the presence of Ca*+-dependent proteases inactivating phospholipase A, even in the presence of PMSF and leupeptin, or to a fraction of the phospholipase AZ becoming integral membrane protein as has been suggested for protein kinase C (14).
To detect phospholipase AS activity from the particulate fraction that was isolated from cells homogenized in the presence of 100 ELM Ca", it was necessary to dissociate the enzyme from the membrane with the Ca*+ chelator, EGTA (Fig. 2) presence of Ca'+, suggesting a Ca'+-dependent association of the enzyme with the membrane. Increasing the ionic strength of the incubation buffer to 1 M NaCl only slightly increased the amount of phospholipase A2 activity recovered. If the particulate fraction was incubated with 10 mM EGTA and 1 M NaCl but was not ultracentrifuged, the total phospholipase Al activity was always less than that measurable following an ultracentrifugation step. When the pellet was resuspended in incubation buffer for an additional 30 min on ice and ultracentrifuged, approximately 50% of its phospholipase AZ activity was now recoverable in the resulting supernatant.
A third incubation and ultracentrifugation did not release any further phospholipase AZ activity into the supernatant. In experiments not shown, particulate fraction was also tested with a variety of detergents in an attempt to improve recovery of the enzyme. However, the phospholipase was completely inhibited by detergents at their critical micellar concentration and it was necessary to dilute the detergent lo-to loo-fold in the phospholipase AS assay to minimize inhibition. Using this approach, no further enzyme was recoverable. The properties of the soluble enzyme recovered from the particulate fraction were similar to those described previously for soluble fraction arachidonoyl-specific phospholipase A2 (7): activity was dependent on the protein concentration and increased linearly up to 26 pg; when assayed as a function of incubation time, a nonlinear relationship was observed after 30 s to 1 min, corresponding to 2% hydrolysis of substrate. The Activity of Arachidonoyl-hydrolyzing Phospholipase A, with Membrane Phospholipid as Substrate-An alternate approach of demonstrating membrane-associated arachidonoylhydrolyzing phospholipase AZ activity was to measure arachidonic acid release from the phospholipids in the membrane that the enzyme became bound to. Initially, to ensure that this enzyme could hydrolyze phospholipids presented as membrane, we measured the activity of the partially-purified phospholipase AZ from the soluble fraction using ["Hlarachidonic acid-labeled membranes as substrate. As shown in Fig. 3, the phospholipase did hydrolyze [3H]arachidonic acid from this substrate, exhibiting nonlinear kinetics after 1-min incubation when more than 5% of substrate was hydrolyzed, which is similar to the kinetics described with liposomal substrate At all time points there was much greater arachidonic acid release from the particulate fraction from cells homogenized with Ca2+ than with EGTA (Fig. 3), demonstrating that an increase in membrane-associated phospholipase AP activity induced by Ca2+ can be detected by this approach. As we have reported previously (7), both the crude and partially purified soluble fraction phospholipase A2 exhibit a biphasic calcium dose response in which activity is measurable at physiological levels of calcium followed by a sharp rise in activity using millimolar concentrations of calcium. The calcium dose response of membraneassociated phospholipase AZ also showed measurable activity at physiological calcium levels, but 3-to 4-fold higher activity in the presence of 10 mM calcium (data not shown), the level used in these assays.
It was interesting to determine the membrane phospholipid substrates that were hydrolyzed by membrane-associated phospholipase AZ and to compare them to those hydrolyzed by partially purified soluble fraction phospholipase AZ. Table  I  fractions, consistent with earlier observations in which arachidonic acid hydrolysis from liposomes containing PC, PE, or PI was observed using soluble fraction phospholipase AZ (7). We then examined the specificity of membrane-associated phospholipase AZ for the fatty acid of membrane phospholipid using cells radiolabeled with either ["Hlarachidonic acid or [3H]oleic acid. Cells were homogenized in the presence of either 100 PM EGTA or 100 pM Ca"+, and the kinetics of 3H- accord with the properties of the arachidonoyl-specific phos-occurs at physiological levels of Ca", supporting a biological role for this process. The partial translocation of phospholipase Az from soluble to particulate fraction has been reported in mouse bone marrow-derived macrophages stimulated with l-oleoyl-Z-acetyl-glycerol (16) and in rat mesangial cells stimulated with phorbol myristate acetate (8). However, either the Ca2+ concentration of the homogenization buffer was not specified (16)  Radiolabeled cells were homogenized in the presence of 10 mM EGTA or varying Ca" concentrations and soluble and particulate (0) fractions were isolated as described under "Experimental Procedures." Particulate fraction phospholipase A, activity was determined in the presence of 10 mM Call> (or in the presence of 10 mM CaCl? in excess of the EGTA concentration) in a 30 min assay using radiolabeled membranes (100,000 dpm) as substrate as described under "Experimental Procedures." Pbospholipase AZ activity is expressed as the difference in ["Hlarachidonic acid release at 30 min minus ["HI arachidonic acid release at 0 min (which ranged between 875 and 1800 dpm). The protein content of particulate fractions resulting in radioactivity of 100,000 dpm ranged from 5 to 7 gg. Data are representative of n = 2 experiments. pholipase AX we have described previously (7). The low level of membrane-associated phospholipase activity that is evident on membranes from EGTA-homogenized cells may represent the membrane-bound, Ca'+-dependent phospholipase A2 with an alkaline pH optimum, which has been described in the macrophage-like cell line, P388D1 (15). Its substrate specificity has not been studied in detail, but it is known to hydrolyze both sn-2 palmitic and arachidonic acids. The Effect of the Ca'+ Concentration of Homogenization Buffer on Membrane Association of Arachidonoyl-hydrolyzing Phosphotipase AP, Detected Using ~HlArachidonic Acid-labeled Membrane as Substrate-The effect of Cap' concentration on the membrane association of phospholipase Az was then measured using \"H]arachidonic acid-labeled membranes as substrate. Fig. 5 shows that particulate fraction phospholipase A2 activity was lowest when cells were homogenized in the presence of Ca" chelators, increased only slightly when cells were homogenized at levels of Ca"+ expected in an unstimulated cell (80 nM), and showed the greatest increase when cells were homogenized at levels of Ca2+ expected in stimulated cells (200-650 nM). Soluble fraction phospholipase A:! activity showed an identical trend to that depicted in Fig.  1 (data not shown). These results provide further evidence that the decrease in soluble fraction phospholipase AS activity seen when increasing amounts of Ca*+ are present in homogenization buffer is exactly paralleled by an increase in particulate fraction phospholipase AZ activity. Moreover, a functional association of the phospholipase A, with membrane phospholipid is induced by Ca"', leading to the hydrolysis of arachidonic acid, but not oleic acid. Hence we suggest a mechanism whereby the activity of phospholipase AZ is regulated: in an unstimulated cell phospholipase A, is found in the cytosol where, in the absence of substrate, there is no arachidonic acid release; upon receptor ligation the cytosolic Ca" concentration increases, and the enzyme becomes membrane-associated predominantly in a Ca"-dependent manner which facilitates arachidonic acid release.