A Guanine Nucleotide-binding Protein Participates in IgE Receptor-mediated Activation of Endogenous and Reconstituted Phospholipase A2 in a Permeabilized Cell System*

M free Ca’+. These data indicate that receptor-mediated activation of a guanine nucleotide-binding protein can shift the Ca2+ dependence of PLA2 activity resulting in greatly enhanced activity at phys- iological concentrations of intracellular free Ca’+. The partial reconstitution of various PLA2 forms into such a broken-cell system offers a new approach for study- ing the mechanisms of G-protein-mediated activation of PLA2.

Emerging paradigms to explain stimulus-secretion coupling in a variety of cells have included receptor-mediated activation of cellular phospholipases.
In addition to a possible role for the well-documented activation of cellular phospholipase C (PLC)' in the stimulation of exocytosis (1,2), there is  (4), and rat basophilic leukemia cells (5). The activation of PLAz within cells primarily results in the increased hydrolysis of the sn-2 acyl linkage of phosphatidylcholine (PC) and, to a lesser extent, phosphatidylethanolamine (6). The fatty acid at the 2-position of PC is largely arachidonic acid, and the activation of cellular PLA, represents a major pathway for mobilization of cellular arachidonic acid. The arachidonic acid thus mobilized serves as the ratelimiting component in the synthesis of cyclooxygenase and lipooxygenase products in cells (7), and it has been shown to be capable of releasing calcium from intracellular stores (8) as well as activating Ca*+-dependent protein kinases (9). The activation of PLAz occurs within minutes of receptor activation (lo), but the mechanism by which receptors stimulate cellular PLA, activities is not well understood.
A number of recent reports have suggested that the receptor-mediated activation of PLA, in several cellular systems involves the activation of guanine nucleotide-binding proteins (G-proteins).
These conclusions are based primarily on the findings that non-hydrolyzable GTP analogs, such as GTPrS, can stimulate PLA, activities in permeabilized cell preparations (11,12) and that receptor-stimulated arachidonic acid production in a variety of cell systems is susceptible to certain bacterial toxins known to modify subsets of cellular G-proteins (11, 42). Studies by Jelsema and Axelrod (13,14) have indicated that the light-dependent activation of PLAz in isolated rod outer segments can be mediated by the free fly subunits of the visual G-protein transducin. A convenient experimental system to study the receptormediated activation of PLA, in secretory cells is the RBL-2H3 subline (15). This cell line appears to be of mucosal mast cell lineage (16) and expresses FccRI, the high-affinity cellsurface receptor for the Fc portion of immunoglobulin E (15, 17,18). Aggregation of these receptors by any one of a variety of means results in cellular degranulation and the accompanying secretion of histamine and other mediators of the immediate hypersensitivity response (15,19). Early biochemical events following the aggregation of IgE receptors on RBL cells include (i) increased hydrolysis of polyphosphoinositides (20,21); (ii) a rise in the cytoplasmic free ionized calcium levels (22); (iii) an increased influx of calcium across the plasma membrane (23,24); and (iv) increased cellular PLA, activity (10,23).
In this report we describe the IgE receptor-mediated, guanine nucleotide-dependent activation of both cellular and exogenously introduced PLAz in permeabilized RBL-2H3 Activation of Phospholipase Az in RBL Cells cells. The results reported here indicate that at least some forms of both intracellular and secreted PLAs enzymes can be activated by a process that involves cellular G-proteins. This experimental system should allow further investigation of the mechanism for G-protein-mediated activation of PLA2.
For measurements of PLA, activity, exogenous substrate was pre-Dared bv drvine a mixture of 1-stearovl-2-13Hlarachidonvl-L-3-phosphatidyichohne ([3H]AA-PC; Amersham Corp.) and carrier soybean PC (Sigma) under a stream of nitrogen. The lipids were then sonicated for lo-15 min into GAME buffer (&a 8) with a probe sonicator (Vibra-Cell, Sonics, at 20 watts). 50 il of sonicated lipids (1500-2000 dnm/ul. 20 Ire/ml with 0.1 me/ml carrier PC) was added to 1 ml of pkrmkabilizedcells along with-appropriate nucleotide analogs and/or antigen (DNPz,Ref. 28). Following incubation at 37 "C for 10 min the reactions were terminated by addition of 3 ml of isopropyl alcohol, n-heptane, 1 M acetic acid (4O:lO:l) and vortexed for 30 s. 1 ml of n-heptane was then added to each reaction and the resultant phases clarified by centrifugation (2000 X g, 5 min) (13). 1.5 ml of the upper phase was then added to 150 mg of silica gel (Fisher, 240 mesh)-and centrifuged (2000 x g, 5 min). 1 ml of the supernatant from this centrifugation was counted in 10 ml of Liuuiscint (National Diagnostics) to determine the release of 3H label from phospholipid substrate.
PLAs activity with endogenous substrate was measured after labeling of cellular lipids with ('HH]arachidonic acid ([3H]AA).
[3H]AA (0.1 &i/ml: Du Pont-New Eneland Nuclear) was added to adherent RBL cells, and the cells were ciltured at 37 "C for an additional 16-20 h. The cells were then harvested, permeabilized, and assayed as described above. Inactivation of Cellular PLA? by p-Bromoacetophenone (PBAJ-Cells were harvested as described above and resuspended in a modified Tyrode's buffer (135 mM NaCl, 20 mM HEPES, 5 mM KCI, 2 mM MgC&, 1.8 mM CaCl*, 5.6 mM glucose) containing a saturated solution ofPBA prepared by vigorous mixing and sonication of finely ground crvstals (2 umol/ml). Following incubation on a rotator at 37 "C for 3d min, the cells were decanted from the settled PBA crystals that remained undissolved, and then the cells were washed and resuspended in GAME buffer containing appropriate [Ca"] and permeabilized as described above. Preparation of Exogenous PLAz-Porcine pancreas and Baja naja PLA;? stock solutions (Sigma) were extensively dialyzed against GAME buffer (pCa 8) at 4 "C, diluted l:lO, and stored at 4 "C. For some experiments the PLAz was chemically inactivated by exposure to PBA as described elsewhere (29,30). Briefly, 10 pM stock PLAz in GAME buffer was incubated with 50 pM PBA for 2 h at 37 "C. The reaction was extensively dialyzed against GAME buffer (&a 8) at 4 "C, diluted l:lO, and stored at 4 'C

Measurements
of PLC and PLD Activities in Permeabilized Cells-Measurements of PLC and phospholipase D (PLD) activities toward exogenous substrate were performed using substrate prepared as described above except that l-stearoyl-2-arachidonyl-3-[3H]choline phosphatidylcholine (1.5-2.0 x 155 dpm/rl) was added as the radiolabel. The release of 3H label as phosphocholine or choline was determined as described elsewhere (31). Inositol-specific PLC activity with endogenous substrate was measured as described previously (28).

RESULTS
Guanine Nucleotide-stimulated PLAz Activity in Permeabilized RBL Cells-Permeabilization of RBL cells is effected by exposure to the cholesterol-binding bacterial toxin, streptolysin 0 (26)(27)(28), which yields relatively large pores >1'20 A in diameter (32). This permeabilized cell system has been pre-viously reported to retain the activation of cellular PLC in response to the aggregation of IgE receptors (27). In the presence of EGTA-containing buffers these pores permit strong buffering of the free cytoplasmic Ca2+ concentration ([Ca"]; Ref. 28). We found that these permeabilized cells also retain PLA2 activity. Exogenous sn-2 [3H]arachidonyl-phosphatidylcholine ([3H]-AA-PC) was initially used as a substrate for the cellular enzyme(s) because of its chemical homogeneity and because potential interference from endogenous phosphatidylinositol-specific PLC activities would be minimized. The exogenous substrate consists of small unilamellar vesicles consisting of radiolabeled PC with carrier soybean PC. PLA:! activities appear to be intracellular because no detectable basal PLA2 activity is observed with biosynthetically labeled intact cells in the presence of millimolar [Ca"], whereas permeabilized RBL cells display a significant basal PLA, activity under similar conditions (Table I). Furthermore, we do not observe any detectable antigen-stimulated PLA2 activity in intact cells as measured by the hydrolysis of exogenous substrate (data not shown).  (Fig. 1B). At the highest levels of [Ca'+] employed (i.e. 1 mM), no effect of antigen addition on PLA2 activity is observed.
In order to assess whether the effect of antigen on PLAz activity involves the activation of a G-protein, we tested the PLAz activity in the presence of low levels of GTP-yS. Low concentrations of GTPyS (0.1 PM) alone have little stimulatory effect on PLAB activity at [Ca"] of 0.1 to 1 PM, and no effect of GTPrS at this concentration is seen when [Ca"+] = 1 mM (Fig. 1B). Addition of antigen together with 0.1 PM GTPyS results in the marked 5-6-fold stimulation of PLA2  (Fig. lB), while no effect of antigen and GTPyS is observed at the highest [Ca"']. It is notable that the effect of antigen and GTPyS on PLAz activity is significantly more than the sum of the effects of either agent added alone (Fig. 1B). All the stimulatory effects of antigen added alone or with GTPyS (0.1 PM) described above could be inhibited (>80%) by the addition of 100 PM GDPpS (data not shown). These data suggest that a cellular G-protein mediates at least a part of the IgE receptor-mediated stimulation of cellular PLAz activity that is induced by the aggregation of IgE receptor complexes. Furthermore, the receptor-mediated stimulation of cellular PLA2 in permeabilized cells appears to be optimal in the range 0.1-l FM [Ca*+], which is similar to the activation of the PLAz observed at higher concentrations of GTPrS in the absence of antigen (Fig. L4).
Similar experiments were performed with permeabilized cells that had been biosynthetically labeled with [3H]AA. The resulting incorporation of this label into a variety of endogenous phospholipids gives rise to a more heterogeneous substrate than the exogenous [3H]AA-PC employed in the experiments of Fig. 1 (33). However, a larger percentage of labeled phospholipids could be hydrolyzed under the stimulating conditions, probably because they were generally more accessible to the activated PLA2. As shown in Table I Exogenous PLA, from Heterologous Sources Can Reconstitute Cellular G-protein-PLA2 Coupling-In order to explore the mechanism for the effect of cellular G-proteins on PLA, activity, we introduced purified exogenous PLAz into permeabilized RBL cell preparations together with the exogenous [3H]AA-PC substrate. Measurement of exogenous PLAz activity against a negligible background of endogenous cellular PLA2 activities was possible if the cells were pretreated with the histidine-modifying agent PBA prior to permeabilization. The pretreatment of intact RBL cells with PBA has been shown previously to inhibit cellular PLAz activity by >95% (5). We observe that pretreatment of RBL cells with PBA (2 mM, 20 min, 37 "C) completely inhibits the stimulation of cellular PLAz activity in permeabilized cells with or without GTPyS (40 PM) at [Ca"'] between 0.1 and 1000 FM (Fig. 2A).
In the presence of these PBA-treated, permeabilized RBL cells and [3H]AA-PC, the addition of GTPyS (40 PM) stimulates the activity of exogenously supplied purified PLA, from porcine pancreas or the N. naja snake venom. The stimulation is maximally 5-fold at [Ca"'] = 1 PM when PLA2 from either N. naja ( Fig. 2A) or porcine pancreas (not shown) is used. Several other experiments (data not shown) provided further characterization of the measured PLA, activity. GTP$S (40 PM) was found to have no effect on the activities of purified PLA2 from porcine pancreatic or N. naja venom sources when measured separately in the same buffers as Fig. 2A of exogenous PLAz from either source into a fixed amount of permeabilized cells ( lo7 cells/ml) shows that maximal stimulation of PLAp activity by GTPyS is achieved when 60 nM PLAz is added. The stimulated PLAz activity is not seen when the exogenous PLAz is chemically inactivated with PBA prior to addition to the permeabilized cells.
Experiments were also performed in which the activity of exogenous PLAz was measured in the presence of permeabilized cells with biosynthetically labeled endogenous phospholipids as substrate. As for the experiments of Table I, the cellular phospholipid pool was labeled with ["HJAA, and the subsequent release of this label from the permeabilized cells was quantified.
Control experiments showed that biosynthetically labeled cells that are treated with PBA and then permeabilized are completely inhibited in GTP+-or antigendependent release of radiolabel over a range of [Ca"]. In contrast, antigen-stimulated PLC hydrolysis of endogenous phosphatidylinositol (28) was found to be unaffected by the PBA treatment (data not shown). Together these results indicate that the PBA-sensitive endogenous PLAz activity is responsible for a major portion of the [3H]AA released with the permeabilized cells. Table II shows that introduction of exogenous porcine pancreatic PLA2 (60 nM/107 cells) results in GTPyS-dependent enhancement in the release of [3H]AA when [Ca'+] is buffered between 0.1 and 10 pM. At these [Ca"'], aggregation of the IgE receptor in the presence of 0.1 pM GTPyS results in release of [3H]AA that is significantly greater than the amount of label released when either 0.1 pM GTPyS or antigen is added alone. These data are consistent with Fig. 2B.

DISCUSSION
The aggregation of the IgE receptors by antigen (0.1 pg/ml Our results demonstrate that receptor-mediated stimula-DNP,,-BSA) inpermeabilized RBL cells prepared from PBA-tion of intracellular PLA2 can be accomplished via a Gpretreated intact cells does not stimulate any endogenous protein-dependent pathway. In these experiments, streptoly-PLA2 activity as expected from the previous results of Mc-sin 0 permeabilization of RBL cells allows manipulation of Givney et al. (5). However, when PLAz from porcine pancreas intracellular composition without disrupting potentially labile (Fig. 2B) and N. naja sources (data not shown) is supplied, receptor activation pathways. Activation of PLA2 was monithis activity is stimulated by the addition of antigen. This tored using both biosynthetically labeled endogenous phosstimulation of exogenous PLA2 activity by antigen is observed pholipid (Tables I and II)  = 1 mM, no effect of antigen on PLAz activity is conditions, a greater number of 3H counts and a greater observed. The addition of 0.1 pM GTPrS alone to permeabil-fraction of the total counts are released with the endogenous ized cells has little effect on the activity of exogenously added substrate than with the exogenous substrate. There may be a PLAz derived from either porcine pancreas (Fig. 2B) or N.
number of factors contributing to this difference, but since naja PLAz (data not shown). However, the simultaneous we have no way of determining the specific activity of the addition of 0.1 pM GTP-yS and antigen stimulates the activity endogenous substrate, the rates of substrate hydrolysis using of exogenous PLAz from both sources >lO-fold ( Fig. 2B and these two methods cannot be directly compared. The different data not shown). This stimulation occurs only when [Ca"] levels are set between 0.1 and 100 pM, and it is significantly greater than the sum of the effects of either agent added alone.
percentages of label hydrolyzed may be explained by differences in the amount of substrate that is accessible to the PLA, and/or G-protein components. It is possible, for example, that only exogenous 13H]AA-PC that becomes incorporated into the cellular membranes can serve as a substrate for the activated PLA,. Despite these differences the results obtained with both methods lead to the same conclusion, that activation of PLA:! enzymes by antigen-mediated cross-linking of IgE receptors is enhanced synergistically by GTPyS, implicating the involvement of a G-protein in the antigenstimulated pathway. It is notable that this activation is seen to occur maximally at [Ca"] in the expected range for a physiologically relevant process (0.1-l @M) can be inhibited by GDP@,. We found that added GTP does not cause a similar synergistic response with antigen, possibly because significant amounts of GTP and/or GDP remain associated with the unwashed permeabilized cells used in our experiments.
Although our results strongly implicate a role for G-proteins in the receptor-mediated activation of cellular PLAZ, the nature of the endogenous G-protein in RBL cells responsible for the activation of endogenous PLAz remains to be elucidated. We have found that pretreatment of cells with either pertussis toxin (100 rig/ml, 3 h) or cholera toxin (1 pg/ml, 3 h) prior to permeabilization does not affect the ability of antigen (or GTP+yS) to activate either the endogenous or exogenous PLA2 activities in these permeabilized cells, suggesting that neither Gi, G,, nor G, is involved.' Cholera toxin pretreatment of RBL cells also does not alter receptor-activated, inositol-specific PLC activity that is assayed after cell permeabilization (28). In experiments with intact RBL cells, IgE receptor-mediated activation of PLAz and PLC has been found to be insensitive to pretreatment with pertussis toxin (Ref. 18 and Footnote 2). We have recently shown that the pretreatment of RBL cells with cholera toxin potentiates the antigen-stimulated Ca'+ influx into intact RBL cells (28), indicating that IgE receptors may stimulate Ca2+ influx and PLA:! via distinct G-proteins.
Several different pathways have been proposed to account for the receptor-mediated activation of PLA, in various cell types, including those involving elevated cytosolic Ca2+ (34), protein kinase C (35), calcium-binding proteins (36, 37), and protein synthesis (38). The synergistic effect of antigen and GTP$S on endogenous PLA2 activity in the permeabilized cell system occurs under conditions where the Ca2+ levels are strongly buffered at 0.1 pM and thus does not appear to require the elevation of [Ca"]. Soluble mediators are unlikely to play a critical role in PLA2 activation in the permeabilized cell system since supernatants from GTPYS-stimulated cells do not possess the ability to activate exogenous PLA2. We have found that pretreatment of cells with cycloheximide and emetine to inhibit protein synthesis does not alter the antigenstimulated PLA,! activity.* Exposure of permeabilized RBL cells to the protein kinase C-activating phorbol ester, tetradecanoyl phorbol acetate (100 nM), does not significantly elevate cellular PLA2 activity under conditions where GTP+S and antigen strongly stimulate cellular PLAz (data not shown). This indicates that the activation of protein kinase C is not sufficient for the activation of PLAz in this system. Thus, our data are most consistent with either a direct interaction of relevant G-proteins with endogenous PLA2 or an indirect interaction that is mediated by other as yet unidentified cellular component(s).
Our ability to reconstitute GTP+yS-and antigen-stimulated PLA, activities in the permeabilized PBA-treated cells using * V. Narasimhan, D. Holowka, and B. Baird, unpublished results. secreted forms of PLA2 is further evidence that the stimulated release of [3H]AA observed in untreated permeabilized cells is catalyzed by endogenous PLAZ. As with the endogenous PLA, activity, the exogenous PLAz activity that is dependent on stimulation by GTP-yS and antigen at low [Ca"] can be detected with either exogenous (Fig. 2) or endogenous (Table  II) substrates. Our results are consistent with previous observations that anti-porcine pancreatic PLAz antibodies can inhibit cellular forms of PLAz that appear to co-localize with the ras protein in ras-transformed cells (39).
Reconstitution experiments with well-characterized secreted forms of PLAz should allow further investigation of the mechanism by which G-proteins mediate the activation of this important class of cellular lipases. For example, a consensus tyrosine-X-glycine-X-glycine sequence involved in Ca*+ binding to both mammalian pancreatic and IV. nuja venom PLAz enzymes has been characterized (40, for review see Ref. 34). As a mechanism for G-protein action on PLA*, our data are consistent with the possibility that the either the G-protein itself or an intermediary protein acts allosterically to enhance Ca2+ binding directly to the consensus Ca*+binding site. An alternative possibility is that the G-proteinactivated species leads to the enhanced binding of PLAz to phospholipid substrate, thereby indirectly increasing the effective affinity of the PLAZ enzyme for Ca*+. Such cooperativity between binding of Ca*+ and phospholipid substrate to distinct sites on the PLA2 from mammalian pancreatic and snake venom sources has been reported (34, 41). The use of this novel experimental system to analyze the G-proteinmediated activation of site-directed mutant PLA, and purified intracellular PLA2 forms should serve to provide greater insight into the role of specific structural features of PLA, enzymes involved in their cellular activation.