The role of phospholipase A activity in rat liver microsomal lipid peroxidation.

The involvement of phospholipase(s) A in lipid peroxidation of rat liver microsomes was investigated by: (a) determining the effects of phospholipase A inhibitors (p-bromophenylacyl bromide, chlorpromazine, mepacrine) on the accumulation of thiobarbituric acid reactivity or on levels of oxidized phospholipids in response to selected oxidative stimuli and (b) measurement of phospholipase A activities in response to these agents. Lipid peroxidation in response to various peroxidation systems was inhibited completely by exposure of microsomes to p-bromophenylacyl bromide (250 microM). The effectiveness of p-bromophenylacyl bromide was dependent on the presence of glutathione (200 microM) in preincubation mixtures. Chlorpromazine (100 microM) and mepacrine (100 microM) also effectively inhibited peroxidation, and their potency was independent of glutathione. The accumulation of oxidized phospholipids in response to the potent peroxidation stimulus alloxan/ferrous ion was similarly inhibited by p-bromophenylacyl bromide, although the level of oxidized phospholipid in response to the initiator ADP/ferrous ion was not affected. Microsomal phospholipase A1 activity, assessed using a liposomal substrate, was substantially enhanced by promoters of lipid peroxidation. Phospholipase A2 activity was not detected using a liposomal substrate but was evident using radiolabeled microsomes as endogenous substrate and was enhanced by oxidative stimuli. We conclude that phospholipase A activity may play an integral role in the microsomal lipid peroxidation mechanism. Based on this study, we hypothesize a role for phospholipases in facilitating propagation reactions.

The Role of Phospholipase A Activity in Rat Liver Microsomal Lipid Peroxidation" (Received for publication, September 29, 1986) Jeffrey H. Beckman, Stephen M. Borowitz, and Ian M. Burr From the Department of Pediatrics, Vanderbilt University Medical Center, Nashville, Tennessee 37232 The involvement of phospholipase(s) A in lipid peroxidation of rat liver microsomes was investigated by: (a) determining the effects of phospholipase A inhibitors (p-bromophenylacyl bromide, chlorpromazine, mepacrine) on the accumulation of thiobarbituric acid reactivity or on levels of oxidized phospholipids in response to selected oxidative stimuli and (b) measurement of phospholipase A activities in response to these agents.
Lipid peroxidation in response to various peroxidation systems was inhibited completely by exposure of microsomes to p-bromophenylacyl bromide (250 PM). The effectiveness of p-bromophenylacyl bromide was dependent on the presence of glutathione (200 PM) in preincubation mixtures. Chlorpromazine (1 00 PM) and mepacrine (100 PM) also effectively inhibited peroxidation, and their potency was independent of glutathione. The accumulation of oxidized phospholipids in response to the potent peroxidation stimulus alloxan/ ferrous ion was similarly inhibited by p-bromophenylacyl bromide, although the level of oxidized phospholipid in response to the initiator ADPIferrous ion was not affected.
Microsomal phospholipase AI activity, assessed using a liposomal substrate, was substantially enhanced by promoters of lipid peroxidation. Phospholipase AP activity was not detected using a liposomal substrate but waa evident using radiolabeled microsomes as endogenous substrate and was enhanced by oxidative stimuli. We conclude that phospholipase A activity may play an integral role in the microsomal lipid peroxidation mechanism. Based on this study, we hypothesize a role for phospholipases in facilitating propagation reactions.
An association between lipid peroxidation and enhanced phospholipase A activity has been demonstrated in various membranes (1-4) including rat liver mitochondria (l), hepatic lysosomes (3), and rat hepatic microsomes (2). Furthermore, both site-specific (4) and nonspecific (5)(6)(7) phospholipase inhibitors have been demonstrated to inhibit lipid peroxidation. It has been tentatively concluded that fatty acyl hydroperoxides may be preferred substrates for phospholipases, explaining the "increase" in measured phospholipase activity (8,9) in response to peroxidizing agents.
To date, this association between lipid peroxidation and phospholipase A activity has not been fully explored. In particular, the concept that phospholipase A activation may * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
be necessary for lipid peroxidation has not been directly addressed nor has possible differential effects of phospholipase A activation on initiation and propagation of peroxidation. Lipid peroxidation consists of distinct phases of initiation (formation of catalytic levels of lipid hydroperoxides) and propagation (extension of lipid hydroperoxide formation upon degradation of initiating species via formation of reactive radicals (10)). Promoters of lipid peroxidation include various forms of chelated iron (11) or of copper (12) which act as either initiators or propagators. Reducing agents such as ascorbic acid (13) or superoxide anion (14) may promote peroxidation by reducing ferric ion or ferric chelates to the ferrous form. A recent report (15) has also suggested that specific ferric to ferrous ion ratios are required for initiation reactions and that oxidants may promote ferrous ion-induced peroxidation. We have recently observed that the diabetogenic agent alloxan, in the presence of ferrous ion and glutathione, acts as a potent promoter of lipid peroxidation (data described in the Miniprint Section'), although its precise role in the peroxidation mechanism remains to be clarified. Alloxan is a known source of free radicals and hydrogen peroxide (16) and thus may promote peroxidation reactions indirectly. Ferric ion may substitute with equal effectiveness for ferrous ion, although the peroxidative response to alloxan/ferric ion is sensitive to inhibition by superoxide dismutase, whereas the alloxan/ferrous ion-induced peroxidation is insensitive to the enzyme, suggesting that superoxide-mediated reduction of ferric ion is required for the alloxan/ferric ion-induced peroxidative response. The alloxan/ferrous ion-induced peroxidative response was not affected by catalase or hydroxyl radical scavengers, whereas it was prevented by the iron chelator, DETAPAC.2 In the present study, we have explored the role of phospholipase A in microsomal lipid peroxidation by determining the effects of phospholipase A inhibition on peroxidative responses promoted by selected stimuli and by investigating the effects of these agents on microsomal phospholipase A activity. Our assessments of lipid peroxidation were made by both a general assay (thiobarbituric acid reactivity) and the more Portions of this paper (including part of "Experimental Procedures," part of "Results," Figs. 8, 10, and Table 11) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814. Request Document No. 86111-3368, cite the authors, and include a check or money order for $2.00 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press.

Phospholipase A and
Microsomal Lipid Peroxidation specific measurement of oxidation products of the phosphatidylcholine fraction.
Preparation of Microsomes-Male Sprague-Dawley rats (200-300 g) were obtained from Harlan Sprague-Dawley, Inc. (Indianapolis, IN). Liver microsomes were prepared from 20 g of tissue by preparing a 10% homogenate in 0.15 M KCl, 10 mM Tris buffer, pH 7.4, and performing differential centrifugation. The pellet obtained by centrifuging the 20,000 X g supernatant at 100,000 X g was resuspended in homogenization buffer (5-10 mg of protein/ml) and assayed protein content (17).
Lipid Peroxidation Assay-Microsomes (500 pg of protein/ml) were incubated at 37 "C in 100 mM Tris buffer, pH 7.4 (1-ml total volume) in the presence of promoters of lipid peroxidation. GSH (200 pM) was included in reaction mixtures unless otherwise noted. The effects of phospholipase A inhibition were investigated by exposing microsomes to PBB (250 JLM) for 30 min prior to addition of the initiator or propagator to reaction mixtures. Following incubation, the method of Mihara et al. (18) was used to derivatize TBA-reactive substances (malondialdehyde plus other peroxidative products) which assayed spectrofluorometrically by the method of Yagi (19).
Preparation of Radiolabeled S~bstrate-~~C-Labeled and unlabeled phosphatidylcholine were dried under nitrogen and redissolved in water by sonication with a probe sonicator (Model W225, Heat Systems-Ultrasonics, Inc., Farmingdale, NY) at setting 3 for two 30s intervals. Phosphatidylcholine was dissolved such that each 200 pl of the 6-12 ml of substrate preparation contained 100 nmol of lipid containing 10,000 cpm.
Phospholipase Assay-Microsomes (500 pg of protein/ml) were incubated for 1 h at 37 "C in the presence of 100 nmol of liposomal substrate/ml, 100 mM Tris buffer, pH 7.5, and the described agents. Alloxan was dissolved in 10 mM HCl prior to dispensing. Incubations were halted by the addition of MeOH/CHC13 (2:1), and the lipids were extracted by the Bligh and Dyer method (20). The chloroform layer was evaporated to dryness under nitrogen and redissolved in 50 pl of chloroform prior to application onto a TLC plate. Thin layer plates were developed 50-60% of their maximum in chloroform/ methanol/acetic acid/water (65:25:84), dried, and then developed the entire length of the plate in hexane/diethyl ether/glacial acetic acid (80201). Lipids were visualized by exposure to iodine, the lanes were marked, and the iodine was allowed to sublime overnight. Lysophosphatidylcboline-, phosphatidylcholine-, and free fatty acid-containing spots were scraped into scintillation vials and counted in a toluenebased scintillant. Phospholipase A, activity was defined as the percentage of recovered counts obtained in lysophosphatidylcholine, and phospholipase A2 activity was defined by the percentage of counts in free fatty acid.
Endogenous Substrate Phospholipase Assay-As an alternative to the use of a liposomal substrate, microsomes were radiolabeled by exposure to [3H]arachidonic acid (0.1 pCi, 83.6 Ci/nmol) for 30 min at 37 "C in 100 mM Tris buffer, pH 7.4, in the presence of 80 p M coenzyme A, 1 mM MgCl,, and 1.6 mM ATP. Incubations labeled the microsomal phospholipids with the phosphatidylcholine fraction containing the majority of the radioactivity. Incubations were continued in the same buffer system in the presence of oxidants using the concentrations described in the text. Reaction mixtures were processed as described for the liposomal assay except that radioactivity was measured for all phospholipid and neutral lipid fractions.
Assay of Phospholipid Oxidation Products-The procedure of Poli et al. (21) was used to prepare 2,4-dinitrophenylhydrazine derivatives of carbonyl compounds of microsomal reaction mixtures. Microsomal reaction mixtures (1 ml) were mixed with 1 ml of dinitrophenylhydrazine reagent (0.34 mg/ml in 1 M HCl) and allowed to react at room temperature in the dark for 2 h. The total lipid fraction was then extracted (20), and extracts were evaporated to dryness under nitrogen redissolved in methanol and applied to silicic acid TLC plates. The plates were developed initially in methylene chloride which separated unreacted reagent and neutral lipids from the phospholipid fraction (which remained at the origin). The plates were developed until at least one-half of their length was free of reagent or neutral lipid-associated dinitrophenylhydrazines. Phospholipids were then separated on this section of the plates by subsequent development using the first solvent system described above for the phospholipase assay. The phosphatidylcholine band (identified using a standard) was scraped off of the plates, eluted with methanol, and measured spectrophotometrically (22). Carbonyl-containing phospholipids were originally described by Tam and McCay (23) who determined that following derivitization with dinitrophenylhydrazine, the resulting phenylhydrazone functional groups were localized on the &position polyunsaturated fatty acids of the phospholipids.

RESULTS
Exposure of rat liver microsomes in the initiator (11) ADP/ ferrous ion (10 p M ) resulted in an early (first 4-5 mins) phase of rapid induction of peroxidative activity (as measured by TBA activity), followed by a slower accumulation of TBAreactive products. Pre-exposure of microsomes to PBB (250 p~) completely blocked lipid peroxidation (Fig. 1). These reaction mixtures contained glutathione (200 p~) (a necessary factor for alloxan-induced propagation reactions (see Fig. 9)). In the absence of glutathione, the peroxidative response to ADP/ferrous ion is not affected by PBB. The effectiveness of PBB (250 p~) as an inhibitor of microsomal phospholipase A, activity in the presence/absence of glutathione paralleled its effectiveness as an inhibitor of lipid peroxidation (Fig. 2). PBB (with or without GSH), incubated for periods up to 30 min with microsomal reaction mixtures in which peroxidation was complete (in response to alloxan/iron), did not affect subsequent measurement of TBA reactivity (data not shown). Chlorpromazine (100 g~) inhibited ADP/ferrous ion-induced lipid peroxidation in the absence of GSH (99% inhibition of relative intensity, p 5 0.001), as did mepacrine (100 g~) (83% inhibition of relative intensity, p 5 0.001).
Initiation of lipid peroxidation by ascorbic acid and ADP/ ferric ion in the presence of glutathione exhibited a "lag period during which lipid peroxidation proceeds at a slow rate, followed by a period of rapid lipid peroxidation, as described by others (24). Pre-exposure of microsomes to PBB (250 p~) completely blocked lipid peroxidation (Fig. 3).
We have observed that exposure of rat liver microsomes to Microsomes (control and PBB-treated) were then exposed to ADP (1 mM final concentration) and ferrous sulfate (10 FM) for the given duration. Lipid peroxidation was assessed by measurement of TBA reactivity as described under "Experimental Procedures." Reaction mixtures were then exposed to ADP (1 mM)/ferrous sulfate (10 p~) or control incubation for the given duration. Lipid peroxidation ( A ) was assessed (TBA reactivity), and phospholipase A, activity ( B ) was measured as described under "Experimental Procedures." Open bars, control incubations; solid bars, ADP/ferrous ionexposed microsomes; cross-hatched bars, PBB-treated microsomes exposed to ADP/ferrous ion. Reaction mixtures were then exposed to ascorbic acid (500 pM final concentration), ADP (1 mM), and ferrous sulfate (10 pM) for the given duration. Lipid peroxidation was assessed by measurement of TBA reactivity as described under "Experimental Procedures." alloxan in the presence of 200 pM glutathione results in enhanced levels of lipid peroxidation (data detailed in the Miniprint Section). Incubation of microsomes with PBB following exposure to ADP/ferrous ion prevented peroxidative responses attributed to alloxan (Fig. 4). Similar peroxidative responses to copper (a reported propagator (12)) were also inhibited by PBB (Fig. 4). Alloxan/ferrous ion-induced lipid Microsomes (500 pg/ml) were exposed to the peroxidation initiator ADP (1 mM)/ferrous ion (10 p~) for 5 min. Incubation was then continued in the presence or absence of PBB for 30 min, followed by an exposure for 5 min to either alloxan (100 pM) (0) or copper (10 p~) (0) or no additional agent (A) without PBB (-) or with PBB (---). Lipid peroxidation was then assessed as described under "Experimental Procedures." Peroxidation was assessed at various intermediate stages as well as final levels stopping reaction at appropriate times.

TABLE I Effects of initiators and propagators of lipid peroxidation on phospholipase A, activity of rat liver microsomes at pH 7.5
Reaction mixtures (1.0-ml total volume) contained rat liver microsomes (500 pg of protein) and radiolabeled phosphatidylcholine liposomes (100 nmol) in 50 mM Tris buffer, pH 7.5. Unless otherwise noted, reaction mixtures contained glutathione (200 p~) . Exposure of microsomes to phosphatidylcholine liposomes and the described agents was performed, and phospholipase A1 activity was measured as described under "Experimental Procedures." peroxidation was also inhibited by chlorpromazine (85% inhibition by 100 p M , p 5 0.001) and mepacrine (78% by 100 p M , p 5 0.001). Microsomal phospholipase AI activity (assessed using a liposomal phosphatidylcholine substrate) was enhanced by both initiators and propagators of lipid peroxidation (Table  I).
ADP/ferrous ion elicited a higher level of phospholipase A, than did ADP/ferric ion (272 uersus 162% of control), whereas the effect of ascorbic acid/ADP/ferric ion on phospholipase AI activity was nearly equal to that of ADPlferrous ion.
Alloxan enhanced microsomal phospholipase AI activity in a manner which strictly paralleled its effect on lipid peroxidation. That is, alloxan-induced phospholipase AI activity required glutathione (data not shown), was potentiated by iron (Table I), and was inhibited completely by DETAPAC (data not shown). Furthermore, the alloxan dose response for phospholipase AI exhibited a biphasic dose-response pattern (Fig. 5) which closely resembled the peroxidation response (Fig. 8).
Phospholipase A2 activity was not detected in response to any of the peroxidation stimuli using the liposomal assay system. Using an assay in which microsomes were preradiolabeled with [3H]arachidonic acid (80% incorporation into phosphatidylcholine) and subsequently exposed to alloxan/ GSH/ferrous ion, we measured a rapid accumulation of radiolabeled lysophosphatidylcholine (indicative of phospholipase A, activity) and of radiolabeled free fatty acids (representative of phospholipase A2) with a parallel loss of radioactivity in phosphatidylcholine (Fig. 6, A and B ) . The enhanced accumulation of radioactivity of free fatty acids was transient, however, suggesting that significant reacylation may have occurred. Indeed, accumulation of radioactivity in the phospholipid precursor phosphatidic acid was prominent in response to alloxan (Fig. 6C). The accumulation of radioactivity into lysophospholipids, free fatty acids, and phosphatidic acid as well as the loss of radioactivity in phosphatidylcholine was completely blocked by 250 p~ PBB (data not shown).
Alloxan/ferrous ion and ADP/ferrous ion each promoted the formation of oxidized phospholipid molecules containing carbonyl groups reactive with 2,4-dinitrophenylhydrazine Microsomes (500 pg of protein/ml) were exposed to alloxan for 1 h at pH 7.5 and 37 "C in the presence of 100 nmol of radiolabeled (2-position) phosphatidylcholine liposomes. Incubations were halted by addition of methanol/chloroform (2:1), lipids were extracted, and thin layer chromatography was performed as described under "Experimental Procedures." Bands corresponding to phosphatidylcholine, lysophosphatidylcholine, and free fatty acids were scraped into counting vials, scintillant was added, and radioactivity was counted. Phospholipase A, activity was defined by the ratio of counts in lysophosphatidylcholine to that in intact phosphatidylcholine. In addition to alloxan, incubations included no additional agents fa) or ferrous sulfate (  lipid (phosphatidylcholine-bound carbonyl groups) by rat liver microsomes exposed to alloxan/ferrous ion or to ADP/ ferrous ion. Microsomes (500 pg of protein/l ml of incubation mixture) were exposed to PBB (250 p M ) or to control incubation for 30 min at 37 "C. Reaction mixtures were then exposed to: ( a ) alloxan (100 p~) and ferrous sulfate (10 pM), ( b ) ADP (1 mM) and ferrous sulfate (10 p~) , or (c) control incubation in the absence of peroxidative stimuli. Reactions were halted by the addition of 2,4-dinitrophenylhydrazine reagent, and oxidized phospholipids (phosphatidylcholine-bound carbonyl groups) were measured as described under "Experimental Procedures." Open bars indicate reactions in the absence of PBB, whereas cross-htched bars represent parallel reactions in the presence of PBB. ADP/ferrous ion; 2.6 f 0.25 p M for control; 15-min incubations). PBB inhibited alloxan/ferrous ion-induced formation of oxidized phospholipid but did not affect the formation of this product in response to ADP/ferrous ion (Fig. 7).

DISCUSSION
We have demonstrated a positive relationship between the process of microsomal lipid peroxidation and phospholipase A activity. Inhibition of phospholipase A activity by PBB (at a level verified effective) prevented peroxidative responses to each of the peroxidative stimuli utilized. PBB acts as a sitespecific phospholipase A inhibitor which binds a histidine residue at the active site of phospholipase A enzyme (25,26). PBB is not specific for phospholipase A (27), however; and its use must be viewed with caution. Therefore, we have verified that other inhibitors of phospholipase(s) A, chlorpromazine and mepacrine, also inhibited lipid peroxidation in response to both ADP/ferrous ion and to alloxan/ferrous ion.
The effectiveness of PBB as an inhibitor of iron-promoted initiation reactions was dependent on the presence of glutathione during the PBB preincubation with microsomes. Similarly, inhibition of microsomal phospholipase A1 activity by PBB required glutathione, suggesting that PBB-induced inhibition of lipid peroxidation is indeed related to phospholipase A, inhibition. In view of the reported reactivity of PBB with thiols (27), we investigated the possibility that PBB might react with GSH to in some manner inhibit the TBA assay. PBB (with or without glutathione) does not inhibit the TBA reaction, and preincubation of PBB with GSH does not render the PBB into an inhibitor of the TBA reaction. Phospholipase assays are similarly not affected by addition of PBB and GSH to reaction mixtures immediately prior to extraction.
The phospholipase activity of rat liver microsomes was measured in the presence of various promoters of lipid peroxidation. The liposomal assay detected only phospholipase AI activity (the predominant phospholipase of microsomes (28) as phospholipase A2 activity was minimal and not influenced by our interventions (including the addition of calcium (5 mM)). The membrane-embedded nature of phospholipase A, (29) or reacylation of hydrolyzed fatty acids (see below) may explain its apparent lack of activity against liposomal substrate. Phospholipase Az activity in response to oxidative stimuli was evident using the assay based on preradiolabeled microsomes. The enhanced accumulation of radioactivity in free fatty acids was transient, however, and would be missed using only long-term incubations. Reacylation of fatty acids into phospholipids was indicated by the accumulation of radioactivity in phosphatidic acid.
Agents which promote lipid peroxidation substantially enhance microsomal phospholipase A activities. Others have reported that phospholipid epoxides are readily hydrolyzed by phospholipases and suggested that phospholipid oxidation products are preferred substrates for phospholipase enzymes (8,9). Indeed, we observed a strong correlation between the biphasic dose-response relationships of alloxan with respect to lipid peroxidation and phospholipase AI activity. Besides the possibility that the oxidized phospholipids are preferred substrates, it is possible that free radicals or other oxidants directly activate phospholipases.
Regardless of the mechanism for apparent phospholipase activation, an important consideration is the role that this activity plays in the lipid peroxidation mechanism. One possibility is that the inhibition of peroxidation by phospholipase A inhibitors was simply a result of confining peroxidation products to a form that was not TBA-reactive, i.e. as intact phospholipids. However, the formation of phospholipid oxidation products in response to alloxan/ferrous ion was also inhibited by PBB, indicating that the peroxidative response to phospholipase inhibitors cannot be explained in this manner. In contrast to its substantial inhibition of alloxan-induced phospholipid oxidation products (reactive with 2,4dinitrophenylhydrazine), PBB did not affect the level of these products in response to the initiator, ADP/ferrous ion. We hypothesize that the phospholipid oxidation products measured in response to ADP/ferrous ion are remnants of phospholipid hydroperoxides derived from primary initiation reactions. We further hypothesize that the oxidized phospholipids induced by alloxan/ferrous ion represent an extension of primary initiation reactions (secondary initiation or propagation) which is phospholipase A-dependent. Unesterified fatty acids released by phospholipases may be more effective in interacting with initiators and may be involved in a "second wave" of initiation. Alternatively, unesterified fatty acid radicals may be better able to reinitiate peroxidative reactions ( i e . propagate peroxidation) than are intact phospholipid radicals or radicals of scission products released nonenzymatically from them. The latter suggestion is analogous to one developed by Parthasarathy et al. (4) in a related study. The authors demonstrated that PBB inhibits the lipid peroxidation associated with endothelial cell modification of low density lipoprotein. They interpreted their data to indicate that free fatty acids might more readily propagate lipid peroxidation due to facilitated interaction with low density lipoprotein molecules and/or endothelial cells. Our results obtained using the rat liver microsomal lipid peroxidation model system indicate that enhanced lipolytic activity may prove to be an integral component of the mechanism of lipid peroxidation.
Although phospholipase activity appears to be directly involved in the lipid peroxidation mechanism, it also may play a role in protecting membranes against peroxidative injury. Glutathione peroxidase reduces fatty acyl hydroperoxides as substrates (30, 31) and thus requires phospholipase activity to provide the substrate with which it can react. Conditions favoring the accumulation of initiation products may thus be alleviated by phospholipase activity.
In summary, we have demonstrated a dependence of rat liver microsomal lipid peroxidation on phospholipase A activities and have shown that a variety of promoters of lipid peroxidation act directly or indirectly as stimuli to increase phospholipase A activity. The data presented suggest that phospholipase activity is involved at a stage of the lipid peroxidation mechanism beyond primary initiation.