Studies of the Role of Group VI Phospholipase A2 in Fatty Acid Incorporation, Phospholipid Remodeling, Lysophosphatidylcholine Generation, and Secretagogue-induced Arachidonic Acid Release in Pancreatic Islets and Insulinoma Cells*

An 84-kDa group VI phospholipase A2 (iPLA2) that does not require Ca2+ for catalysis has been cloned from Chinese hamster ovary cells, murine P388D1 cells, and pancreatic islet β-cells. A housekeeping role for iPLA2 in generating lysophosphatidylcholine (LPC) acceptors for arachidonic acid incorporation into phosphatidylcholine (PC) has been proposed because iPLA2 inhibition reduces LPC levels and suppresses arachidonate incorporation and phospholipid remodeling in P388D1 cells. Because islet β-cell phospholipids are enriched in arachidonate, we have examined the role of iPLA2 in arachidonate incorporation into islets and INS-1 insulinoma cells. Inhibition of iPLA2 with a bromoenol lactone (BEL) suicide substrate did not suppress and generally enhanced [3H]arachidonate incorporation into these cells in the presence or absence of extracellular calcium at varied time points and BEL concentrations. Arachidonate incorporation into islet phospholipids involved deacylation-reacylation and not de novo synthesis, as indicated by experiments with varied extracellular glucose concentrations and by examining [14C]glucose incorporation into phospholipids. BEL also inhibited islet cytosolic phosphatidate phosphohydrolase (PAPH), but the PAPH inhibitor propranolol did not affect arachidonate incorporation into islet or INS-1 cell phospholipids. Inhibition of islet iPLA2 did not alter the phospholipid head-group classes into which [3H]arachidonate was initially incorporated or its subsequent transfer from PC to other lipids. Electrospray ionization mass spectrometric measurements indicated that inhibition of INS-1 cell iPLA2 accelerated arachidonate incorporation into PC and that inhibition of islet iPLA2 reduced LPC levels by 25%, suggesting that LPC mass does not limit arachidonate incorporation into islet PC. Gas chromatography/mass spectrometry measurements indicated that BEL but not propranolol suppressed insulin secretagogue-induced hydrolysis of arachidonate from islet phospholipids. In islets and INS-1 cells, iPLA2 is thus not required for arachidonate incorporation or phospholipid remodeling and may play other roles in these cells.

An 84-kDa group VI phospholipase A 2 (iPLA 2 ) that does not require Ca 2؉ for catalysis has been cloned from Chinese hamster ovary cells, murine P388D1 cells, and pancreatic islet ␤-cells. A housekeeping role for iPLA 2

in generating lysophosphatidylcholine (LPC) acceptors for arachidonic acid incorporation into phosphatidylcholine (PC) has been proposed because iPLA 2 inhibition reduces LPC levels and suppresses arachidonate incorporation and phospholipid remodeling in P388D1 cells. Because islet ␤-cell phospholipids are enriched in arachidonate, we have examined the role of iPLA 2 in arachidonate incorporation into islets and INS-1 insulinoma cells. Inhibition of iPLA 2 with a bromoenol lactone (BEL) suicide substrate did not suppress and generally enhanced [ 3 H]arachidonate incorporation into these cells in the presence or absence of extracellular calcium at varied time points and BEL concentrations. Arachidonate incorporation into islet phospholipids involved deacylation-reacylation and not de novo synthesis, as indicated by experiments with varied extracellular glucose concentrations and by examining [ 14 C]glucose incorporation into phospholipids. BEL also inhibited islet cytosolic phosphatidate phosphohydrolase (PAPH), but the PAPH inhibitor propranolol did not affect arachidonate incorporation into islet or INS-1 cell phospholipids. Inhibition of islet iPLA 2 did not alter the phospholipid head-group classes into which [ 3 H]arachidonate was initially incorporated or its subsequent transfer from PC to other lipids. Electrospray ionization mass spectrometric measurements indicated that inhibition of INS-1 cell iPLA 2 accelerated arachidonate incorporation into PC and that inhibition of islet iPLA 2 reduced LPC levels by 25%, suggesting that LPC mass does not limit arachidonate incorporation into islet PC. Gas chromatography/mass spectrometry measurements indicated that BEL but not propranolol suppressed insulin secretagogue-induced hydrolysis of arachidonate from islet phospholipids. In islets and INS-1 cells, iPLA 2 is thus not required for arachidonate incorporation or phospholipid remodeling and may play other roles in these cells.
Phospholipases A 2 (PLA 2 ) 1 catalyze hydrolysis of the sn-2 fatty acid substituent from glycerophospholipid substrates to yield a free fatty acid and a 2-lysophospholipid (1)(2)(3)(4)(5)(6)(7). PLA 2 are a diverse group of enzymes, and the first members to be well characterized have low molecular masses (ϳ14 kDa), require millimolar [Ca 2ϩ ] for catalytic activity, and function as extracellular secreted enzymes designated sPLA 2 (3,6). The first PLA 2 to be cloned that is active at [Ca 2ϩ ] that can be achieved in the cytosol of living cells is an 85-kDa protein classified as a group IV PLA 2 and designated cPLA 2 (3,5). This enzyme is induced to associate with its substrates in membranes by rises in cytosolic [Ca 2ϩ ] within the range achieved in cells stimulated by extracellular signals that induce Ca 2ϩ release from intracellular sequestration sites or Ca 2ϩ entry from the extracellular space, is also regulated by phosphorylation, and prefers substrates with sn-2 arachidonoyl residues (5).
Recently, a second PLA 2 that is active at [Ca 2ϩ ] that can be achieved in cytosol has been cloned (8 -10). This enzyme does not require Ca 2ϩ for catalysis, is classified as a group VI PLA 2 , and is designated iPLA 2 (3,4). The iPLA 2 enzymes cloned from hamster (8), mouse (9), and rat (10) cells represent species homologs, and all are 84-kDa proteins containing 751-752 amino acid residues with highly homologous (ϳ95% identity) sequences. Each contains a GXSXG lipase consensus motif and eight stretches of a repeating sequence motif homologous to a repetitive motif in the integral membrane protein binding domain of ankyrin (8)(9)(10). The substrate preference of these iPLA 2 enzymes varies widely with the mode of presentation (8), but each is susceptible to inhibition (8-10) by a bromoenol lactone (BEL) suicide substrate (11,12) that is not an effective inhibitor of sPLA 2 or cPLA 2 enzymes at comparable concentrations (4,(11)(12)(13)(14).
It is not yet certain that the observations with P388D1 cells reflect a general mechanism for incorporation of arachidonic acid into phospholipids of all cells, and P388D1 cells exhibit some atypical features of arachidonate incorporation. Arachidonate represents about 25% of the total esterified fatty acyl mass in native murine macrophage phospholipids (25,26) but only 3% of that in murine P388D1 macrophage-like cells (27)(28)(29). A significant portion of exogenous arachidonate provided to P388D1 cells is also found in the intracellular free fatty acid fraction (29), and this is not the case in other cell types (30 -32), in which nonesterified arachidonate is maintained at a very low level by an efficient esterification system (26,(33)(34). In many cells, nonesterified arachidonate imported from the extracellular space or released intracellularly by PLA 2 enzymes is quickly converted to arachidonoyl-CoA in an ATP-dependent step and then rapidly incorporated into phospholipids by acyltransferases (35,36). The low levels of esterified arachidonate and the relatively high levels of nonesterified arachidonate observed in P388D1 cells (27)(28)(29) suggest that they may be deficient in arachidonate incorporation mechanisms expressed by other cells.
One biomedically important cell that may require especially effective arachidonate incorporation mechanisms is the insulin-secreting pancreatic islet beta cell, the function of which is impaired in diabetes mellitus. Arachidonate represents 30 -36% of the total esterified fatty acyl mass in phospholipids of normal rat and human islets and isolated beta cells (37,38), and arachidonate-containing species are the most abundant components of all major islet phospholipid head-group classes (39). The abundance of arachidonate-containing phospholipid species is higher in islets than in many other tissues (39), and islet plasma membranes and secretory granule membranes are especially enriched in such species (38,39). Fusion of secretory granule and plasma membranes is the final event in insulin exocytosis, and the high content of certain arachidonate-containing phospholipids in those membranes may facilitate their fusion (39,40). Because both rat and human beta cells also express iPLA 2 (10, 41), we have examined the participation of iPLA 2 in arachidonic acid incorporation into islet and insulinoma cell phospholipids, and our findings differ from those in P388D1 cells.
Isolation and Culture of Pancreatic Islets and Insulinoma Cells-Islets were isolated aseptically from male Sprague-Dawley rats by collagenase digestion of minced pancreas, density gradient isolation, and manual selection under microscopic visualization (43). Isolated islets were placed in cCMRL-1066 under a humidified atmosphere of 95% air, 5% CO 2 and cultured at 37°C. Islet total insulin and protein contents were determined as described previously (44). Islets were isolated from human pancreata in the Core Laboratory of the Washington University Diabetes Research and Training Center (45) and were cultured as described (39). INS-1 insulinoma cells provided by Dr. Christopher Newgard (University of Texas-Southwestern, Dallas, TX) were cultured under previously described conditions (46 -48).
Incubation of Islets and Insulinoma Cells with Radiolabeled Fatty Acids, BEL, BAPTA, IL-1, and Other Additives-Islets or INS-1 cells were washed three times in KRB medium containing 5.5 mM glucose and 0.1% BSA, resuspended in that medium, and preincubated for 30 min at 37°C. The islets (100 per condition) or INS-1 cells (2 ϫ 10 5 per condition) were then placed in fresh KRB medium that contained glucose (5.5, 11, or 22 mM), 0.1% BSA, and either 2.5 mM CaCl 2 or 1 mM EGTA and no added Ca 2ϩ and were incubated (30 min, 37°C) after addition of vehicle only (control) or BEL (1-25 M). In experiments examining iPLA 2 activity, control and BEL-treated islets or INS-1 cells were then either homogenized immediately or placed in fresh cCMRL medium and incubated at 37°C for 8 -68 h before homogenization, and iPLA 2 activity was determined in homogenates. In some experiments examining radiolabeled fatty acid incorporation, effects of loading islets or INS-1 cells with BAPTA-AM (100 M, 30 min) or propranolol (250 M, 30 min) under described conditions (49) before adding radiolabeled fatty acid were determined. After the preincubation and loading steps described above, fatty acid incorporation experiments were initiated by adding either [ 3 H]arachidonic acid (final concentration 0.5 Ci/ml, 5 nM) or [ 3 H]palmitic acid (final concentration 0.5 Ci/ml, 10 nM) to the medium, and incubation was performed for 10 -60 min at 37°C. Islets or INS-1 cells were then washed three times in KRB medium containing 5.5 mM glucose and 0.1% BSA to remove unincorporated 3 H-fatty acid.
In some experiments, cellular lipids were then immediately extracted by the method of Bligh and Dyer (50) and analyzed by TLC or HPLC. In other experiments, after initial incubation with [ 3 H]arachidonic acid, islets were washed to remove unincorporated radiolabel, placed in cCMRL medium, and incubated (24 h, 37°C) in the absence of 3 H-fatty acid. The islets were then washed and their lipids extracted and analyzed. In experiments examining effects of IL-1 on islet fatty acid incorporation, islets were incubated (24 h, 37°C) with no additions, with IL-1 (5 units/ml) alone, or with IL-1 plus NMMA (0.5 mM) under described conditions (51). The islets were then washed, preincubated, and incubated with 3 H-fatty acids as above. Under all loading and incubation conditions, islet and INS-1 cell viability exceeded 98% by trypan blue exclusion performed as described (49).
Assay of iPLA 2 Activity in Islets and INS-1 Cells-Control and BEL-treated islets and INS-1 cells were homogenized by sonication as described (52). Protein content was measured with Coomassie reagent (Pierce) against bovine serum albumin as standard. Ca 2ϩ -independent PLA 2 activity was assayed by incubating (30 min, 37°C) aliquots of homogenate in buffer (100 mM Hepes, pH 7.5, 5 mM EDTA, 400 M Triton X-100) with 100 M 1-palmitoyl-2-[ 14 C]linoleoyl-sn-glycero-3phosphocholine in a final volume of 500 l with or without 1 mM ATP. The substrate was presented in mixed micelles of Triton X-100/phospholipid at a molar ratio of 4/1, obtained by a combination of heating, vortex mixing, and bath sonication. Products were analyzed by TLC as described below, and PLA 2 specific activity was calculated from disintegrations/min of released [ 14 C]linoleate and protein content as described (52). Chromatographic Analyses of Radiolabeled Phospholipids-Incorporation of 3 H-fatty acids into islet and INS-1 cell phospholipids was determined by TLC (15,16) or by HPLC (37,49,53). TLC was performed on silica gel G plates with hexane/diethyl ether/acetic acid Immunoblotting Analyses of INS-1 Cell iPLA 2 Protein-INS-1 cell cytosolic proteins were analyzed by SDS-PAGE and transferred to a nylon membrane that was subsequently blocked (3 h, room temperature) with Tris-buffered saline/Tween (TBS-T, which contains 20 mM Tris-HCl, 137 mM NaCl, pH 7.6, and 0.05% Tween 20) containing 5% milk protein. The blot was then washed (TBS-T, 5 min, five times) and incubated (1 h, room temperature) with a polyclonal antibody (1:4000 dilution in TBS-T) to iPLA 2 provided by Dr. Simon Jones (Genetics Institute, Boston). The nylon membrane was then washed in TBS-T (5 min, five times) and incubated (1 h, room temperature) with a secondary antibody coupled to horseradish peroxidase (Roche Molecular Biochemicals) at 1:20,000 dilution in TBS-T containing 3% BSA. The antibody complex was visualized by enhanced chemiluminescence (ECL, Amersham Pharmacia Biotech).
Isolation of RNA from INS-1 Cells, Reverse Transcription, and Polymerase Chain Reaction Amplification of an iPLA 2 cDNA Fragment-Total RNA was isolated from INS-1 cells and first strand cDNA prepared by reverse transcription (RT) using standard methods (8,55). Polymerase chain reactions (PCR) were performed and products analyzed as described elsewhere (8). Primers used to amplify a rat iPLA 2 cDNA fragment were sense (5Ј-TGCAGACCAGTTAGTATGGC-3Ј) and antisense (5Ј-TGGACAAGCTTCTGGAAC-3Ј). The expected fragment length is 528 bp, and digestion with restriction endonuclease SmaI is expected to yield 164-and 364-bp fragments.
Attempts at Suppressing INS-1 Cell iPLA 2 Expression with Antisense Oligonucleotides-Many attempts to suppress INS-1 cell iPLA 2 expression with antisense oligonucleotides were performed using different oligonucleotide sequences at various concentrations, durations of exposure, and presentation conditions. The first mimicked that applied to P388D1 cells (16) and employed a 20 base-long antisense oligonucleotide corresponding to nucleotides 59 -78 in the rat group VI iPLA 2 sequence (10). The oligonucleotide sequence is 5Ј-CTCCTTTGCCCG-GAATG-GGT-3Ј, and a sense oligonucleotide was used as control. Both contained phosphorothioate linkages to limit degradation. For transfection, 2.5 ϫ 10 5 cells were cultured with oligonucleotide in Iscove's modified Dulbecco's medium for 4 h in a humidified atmosphere at 90% air and 10% CO 2 . A final concentration of 10% fetal bovine serum was then added, and the cells were cultured for an additional 48 h in medium supplemented with 2 mM glutamine, 100 units/ml penicillin, 100 g/ml streptomycin, and nonessential amino acids. At oligonucleotide concentrations from 0.10 to 20 M, there was no suppression of INS-1 cell iPLA 2 protein or activity, and cell detachment occurred at higher concentrations, as for P388D1 cells (16). Oligonucleotide presentation in various concentrations of Lipofectin (56) to cells attached to surfaces and in suspension was also examined, as were two other antisense oligonucleotide sequences. One (5Ј-AGGCGTCCAAAG-AACTGC-3Ј) was designed hybridize to the mRNA region containing the initiator codon and the other (5Ј-TTGTTGTCATGTACATCC-3Ј) to a region containing a methionine codon upstream of the catalytic center coding region, but neither was effective.
Islet de Novo Synthesis of Glycerolipids from [ 14 C]Glucose-Islet synthesis of glycerolipids from [ 14 C]glucose was examined by a modification of described methods (60). Islets suspended in KRB medium containing 5.5 mM glucose and 0.1% BSA were placed (100/condition) in silanized test tubes and preincubated (30 min, 37°C) with vehicle only (control) or with BEL (25 M). [ 14 C]Glucose (final concentration 20 Ci/ml) was then added, and the islets were incubated for 60 min at 37°C. Medium was then removed and the islets washed with KRB medium containing 5.5 mM glucose. Islet lipids were then extracted (50), and the extract was concentrated and analyzed by TLC on silica gel G plates with toluene/diethyl ether/ethyl acetate/acetic acid (80/10/ 10/0.2). Phospholipids, diacylglycerol, and triacylglycerol were recovered from the plate and their [ 14 C] contents determined by liquid scintillation spectrometry.

Incubation of INS-1 Cells with Arachidonic Acid to Induce Phospholipid Remodeling-INS-1 insulinoma cells cultured in RPMI medium
supplemented with penicillin, streptomycin, fungizone, and gentamicin at a concentration of 0.1% (w/v) each. Cells (1.2 ϫ 10 6 per condition) were treated (30 min, 37°C) with vehicle only or with BEL (25 M). Medium was then removed and replaced with fresh medium containing no supplements other than those described above or containing arachidonic acid (final concentration 70 M), and cells were cultured at 37°C for various periods. After 0, 8, or 24 h, cells were washed twice with phosphate-buffered saline, suspended in homogenization buffer, and disrupted by sonication. One aliquot of homogenate was used to measure protein content and iPLA 2 activity, and the remainder was extracted by the method of Bligh and Dyer (50). The extract was concentrated and analyzed by NP-HPLC to separate phospholipid head-group classes. GPC lipids were analyzed by ESI/MS.

Effects of BEL on Islet Lysophosphatidylcholine (LPC) Mass and on the [ 3 H]LPC Content of Islets Labeled with [ 3 H]Choline-
To determine islet LPC mass, islets (1500 per condition) were washed twice with and then resuspended and incubated (30 min, 37°C) in KRB medium containing 5.5 mM glucose and 0.1% BSA. Islets were then placed in fresh medium and incubated (30 min, 37°C) with vehicle only (control) or with BEL (25 M). To one set of islets, bee venom sPLA 2 (50 units/ml, Sigma) and CaCl 2 (2.5 mM) were added, and incubation was continued (30 min, 37°C). Islets were then washed twice with phosphate-buffered saline and extracted by the method of Bligh and Dyer (50) with the modifications that 150 mM LiCl in water was aqueous phase and that the chloroform phase contained 17:0-LPC internal standard (300 pmol per condition). Concentrated extracts were analyzed by silica gel G TLC on heat-activated ( (20 Ci/ml) and then washed with and resuspended (300 islets/condition) in KRB medium containing 5.5 mM glucose and 0.1% BSA and incubated (30 min, 37°C). Medium was removed and replaced with fresh medium supplemented with vehicle only (control) or with BEL (25 M), and incubation was performed (30 min, 37°C). Medium was removed and replaced with fresh medium, and incubation was performed (10 min, 37°C). Islets were washed with KRB containing 5.5 mM glucose and 0.1% BSA and extracted (50). Concentrated extracts were analyzed by TLC to isolate LPC, and its 3 H content was determined.
Electrospray Ionization Mass Spectrometric Analyses of Choline-containing Lipids-PC and LPC were analyzed as Li ϩ adducts by ESI/MS on a Finnigan (San Jose, CA) TSQ-7000 triple stage quadrupole mass spectrometer with an ESI source controlled by Finnigan ICIS software. Phospholipids were dissolved in methanol/chloroform (9/1, v/v) containing LiOH (2 nmol/l), infused (1 l/min) with a Harvard syringe pump, and analyzed under described instrumental conditions (39,49,62). For tandem MS, precursor ions selected in the first quadrupole were accelerated (32-36 eV collision energy) into a chamber containing argon (2.3-2.5 millitorr) to induce collisionally activated dissociation, and product ions were analyzed in the final quadrupole. Identities of GPC species in the total ion current profile were determined from their tandem spectra (62). To quantitate LPC species, constant neutral loss scanning was performed to monitor LPC [M ϩ Li] ϩ ions that undergo loss of 59 (trimethylamine) upon collisionally activated dissociation. Quantitation was achieved by comparing the intensity of the ion for the 17:0-LPC internal standard (m/z 516) to intensities of ions for the endogenous islet LPC species (63) 16:0-LPC (m/z 502) and 18:0-LPC (m/z 530). Standard curve experiments in which a constant amount of standard 17:0-LPC and varied amounts of standard 16:0-LPC and 18: 0-LPC were added to a series of tubes and analyzed as Li ϩ adducts by ESI/MS/MS indicated that this method was linear over a wide range that included levels observed in islets.
Examination of Insulin Secretagogue-induced Hydrolysis of Arachidonic Acid from Islet Phospholipids by Isotope Dilution Gas Chromatography-Negative Ion Electron Capture-Mass Spectrometry-Accumulation of nonesterified arachidonic acid in islet incubations was determined by modifications of described methods (42,49). Islets (300 per condition) were incubated (30 min, 37°C) in KRB medium containing 3 mM glucose, 2.5 mM CaCl 2 , 0.1% BSA, and vehicle only (control), BEL (25 M), or propranolol (250 M). Medium was then removed and replaced with fresh medium. To some tubes no supplements were added, and to others either supplemental glucose (final concentration 17 mM) or glucose plus carbachol (final concentration 0.5 mM) was added. Incubation was then performed (30 min, 37°C) and terminated by extraction with 2 ml of chloroform/methanol (1/1) containing 825 pmol of [ 2 H 8 ]arachidonic acid internal standard. Extracts were analyzed by RP-HPLC to isolate arachidonic acid, which was converted to a pentafluorobenzyl ester derivative and analyzed by GC/MS in negative ion electron capture mode under described conditions (42,49). Selected monitoring of carboxylate anions of arachidonate (m/z 303) and [ 2 H 8 ]arachidonate (m/z 311) was performed to quantitate arachidonate by reference to a standard curve (42,49).
Statistical Analyses-Student's t test was used to compare two groups, and multiple groups were compared by one-way analysis of variance with post hoc Newman-Keul's analyses.

RESULTS
In the report that motivated our study, 25 M BEL maximally inhibited iPLA 2 activity and suppressed [ 3 H]arachidonic acid incorporation into P388D1 cell phospholipids during a 10-min incubation but only slightly suppressed [ 3 H]palmitic acid incorporation, and [ 3 H]arachidonic acid incorporation was unaffected by chelation of extracellular or intracellular [Ca 2ϩ ] (15). To examine the generality of these phenomena, we performed similar experiments with islets. Control islets exhibited PLA 2 activity in the absence of Ca 2ϩ that was stimulated by ATP, and activity was 98% inhibited in 25 M BEL-treated islets (Fig. 1, left panel), consistent with properties of recombinant iPLA 2 expressed from cDNA cloned from an islet library (10). Despite effective iPLA 2 inhibition, no suppression of [ 3 H]arachidonic acid incorporation into phospholipids was observed in BEL-treated islets during 10-min incubations in Ca 2ϩ -replete medium, in Ca 2ϩ -free EGTA-containing medium, or in islets that had been loaded with the intracellular Ca 2ϩchelator BAPTA-AM and incubated in Ca 2ϩ -free EGTA-containing medium (Fig. 1, center panel). dium than in Ca 2ϩ -replete medium, consistent with reports that exogenous arachidonate achieves greater intracellular access in islets (64,65) and other tissues (66) in Ca 2ϩ -free compared with Ca 2ϩ -replete medium and that this also occurs with palmitate (67).
In Ca 2ϩ -free EGTA-containing medium, non-BAPTA-loaded islets incorporated more [ 3 H]arachidonic acid into phospholipids than did BAPTA-loaded islets (Fig. 1, center panel), under conditions where BAPTA reduces islet beta cell cytosolic [Ca 2ϩ ] by 50% and prevents ionophore A23187-induced rises in cytosolic [Ca 2ϩ ] (49). This indicates that [ 3 H]arachidonic acid incorporation into islet phospholipids is not completely Ca 2ϩindependent. Incorporation of [ 3 H]palmitic acid into islet phospholipids (Fig. 1, right panel) exhibited features similar to those for [ 3 H]arachidonic acid, and incorporation of both fatty acids was generally higher in BEL-treated than control islets. This effect was observed at all incubation times between 10 and 60 min (Fig. 2), and no suppression of [ 3 H]arachidonic acid incorporation into islet phospholipids was observed at any BEL concentration that reduced islet iPLA 2 activity (Fig. 3). Incorporation of both fatty acids into human islet phospholipids was affected by BEL and by manipulating Ca 2ϩ in ways similar to those with rat islets (not shown).
To determine whether it is possible to suppress fatty acid incorporation into islet phospholipids under our conditions, we examined effects of incubating islets with interleukin-1 (IL-1) (Fig. 4), which stimulates islet production of nitric oxide and causes accumulation of nonesterified arachidonic acid by an NO-dependent mechanism that is thought to reflect impaired arachidonoyl-CoA generation (51). Incubating islets with IL-1 impairs incorporation of [ 3 H]arachidonic acid into islet phospholipids in Ca 2ϩ -replete medium (51), and we found that this is also true in Ca 2ϩ -free medium and that incorporation of [ 3 H]palmitic acid is similarly suppressed (Fig. 4). Coincubation of islets with IL-1 and the nitric oxide synthase inhibitor NMMA prevented IL-1-induced suppression of fatty acid incor-poration into islet phospholipids (Fig. 4), reflecting the NO dependence of this effect. Fatty acid incorporation into islet phospholipids thus can be suppressed, even though inhibiting islet iPLA 2 activity does not do so.
To attempt to inhibit beta cell iPLA 2 activity with antisense oligonucleotides, we examined INS-1 insulinoma cells because islets are a multicellular aggregate in which it is difficult to achieve effective intracellular levels of oligonucleotide. INS-1 cells are derived from rat islet beta cells (46 -48) and were found to express a PLA 2 activity in the absence of Ca 2ϩ that is stimulated by ATP and inhibited by BEL (Fig. 5A), an 84-kDa protein recognized by an iPLA 2 antibody (Fig. 5B), and an mRNA species that is amplified in RT-PCR reactions with rat iPLA 2 -specific primers and that contains the expected SmaI restriction endonuclease cleavage site (Fig. 5C). We were unable to suppress INS-1 cell iPLA 2 activity or protein with an antisense oligonucleotide corresponding in sequence, except for minor differences between species, to one that suppresses P388D1 cell iPLA 2 expression (16). Two other antisense oligonucleotide sequences presented under a variety of conditions were also ineffective.
Nonetheless, because INS-1 cell iPLA 2 activity is inhibited by BEL, we compared effects of BEL on fatty acid incorporation into phospholipids of islets and INS-1 cells. We also examined effects of propranolol because both BEL and propranolol inhibit phosphatidate phosphohydrolase (PAPH) in some cells (57,70). At 250 M, propranolol maximally inhibits PAPH in islets (71) and amniotic WISH cells (70) and suppressed islet de novo synthesis of triacylglycerol, as do PAPH inhibitors in P388D1 cells (29,57). BEL did not suppress incorporation of [ 3 H]arachidonic acid or [ 3 H]palmitic acid into islet (Fig. 6, A and B) or INS-1 cell phospholipids (Fig. 6, C and D) in Ca 2ϩ -replete or Ca 2ϩ -free EGTA-containing medium, and incorporation of neither fatty acid was significantly affected by propranolol in islets or INS-1 cells (Fig. 6).
One possible explanation for the failure of BEL to suppress Fig. 1 and is expressed as disintegrations/ min per 20 g of islet protein (n ϭ 6). fatty acid incorporation into islet phospholipids is that, by inhibiting PAPH, BEL causes accumulation of phosphatidic acid (PA) and enhances de novo phosphatidylinositol (PI) synthesis (24,72). This might mask suppression of [ 3 H]arachidonic acid incorporation into GPC lipids by deacylation/reacylation. We first determined whether BEL inhibits islet PAPH activity. PAPH isoforms include PAPH-1, which is Mg 2ϩ -dependent and inhibited by NEM (58,73) and BEL, and PAPH-2, which is Mg 2ϩ -independent, NEM-insensitive, and resistant to BEL (57). Islets were found to express Mg 2ϩ -dependent, NEM-sensitive PAPH activity both in cytosol and membranes (Table I), consistent with a previous report (74). BEL (25 M) reduced islet cytosolic Mg 2ϩ -dependent, NEM-sensitive PAPH activity by 48% but had no effect on membranous Mg 2ϩ -dependent, NEM-sensitive PAPH activity (Table I). Neither BEL nor NEM affected islet PAPH activity in the absence of Mg 2ϩ , and the combination of BEL and NEM achieved no greater inhibition of cytosolic PAPH activity than NEM alone (not shown).

FIG. 2. Time course of incorporation of [ 3 ]arachidonic acid or [ 3 ]palmitic acid into islet phospholipids. Islets were incubated with vehicle (open symbols) or 25 M BEL (closed symbols) for 30 min at 37°C and were then incubated with [ 3 H]arachidonic acid (left panel) or [ 3 H]palmitic acid (right panel) for 10 -60 min. The 3 H content of islet phospholipids was determined as in
The inhibition of islet cytosolic PAPH-1 activity by BEL caused us to examine the possibilities that BEL stimulates incorporation of fatty acids into PA and PI via de novo synthesis while it reduces incorporation into phosphatidylcholine (PC) by deacylation/reacylation. Islet phospholipids into which [ 3 H]arachidonic acid is incorporated were analyzed by NP-HPLC (Fig. 7). After 10 min of incubation with [ 3 H]arachidonic acid, the majority of [ 3 H] was incorporated into PC in both control, as previously observed (63), and BEL-treated islets. BEL-treated islets exhibited greater [ 3 H]arachidonic acid incorporation into PC than control islets, and the profile of [ 3 H]arachidonic acid-labeled phospholipid head-group classes was similar in the two groups (Fig. 7A). Upon removal of exogenous [ 3 H]arachidonic acid and continued culture, a timedependent decline in the [ 3 H]arachidonate content of PC and a rise in that of phosphatidylethanolamine (PE) occurred, as previously observed in islets (63) and other cells (34), and these effects occurred in both BEL-treated and control islets (Fig.  7B). RP-HPLC analyses of PC molecular species labeled with [ 3 H]arachidonic acid were also similar in control and BEL-treated islets (Fig. 7C). PI contained only a minor fraction of [ 3 H]arachidonic acid incorporated into phospholipids in BELtreated or control islets. These findings indicate that accumu-

FIG. 4. Suppression of [ 3 ]arachidonic acid and of [ 3 ]palmitic incorporation into islet phospholipids after incubation with interleukin-1 and prevention of this effect by a nitric oxide synthase inhibitor.
Islets were incubated with vehicle (CONTROL), with 5 units/ml interleukin-1 alone (IL-1), or with IL-1 plus 0.5 mM NMMA (IL-1ϩNMMA) for 24 h at 37°C. Islets were then incubated with [ 3 H]arachidonic acid (dark bars) or with [ 3 H]palmitic acid (crosshatched bars) for 10 min at 37°C in medium containing 1 mM EGTA and no added Ca 2ϩ . The 3 H content of islet phospholipids was then determined as in Fig. 1 and is expressed as disintegrations/min per 20 g of islet protein (n ϭ 6).

FIG. 3. Effects of varied concentrations of BEL on islet iPLA 2 activity and on incorporation of [ 3 ]arachidonic acid into islet phospholipids.
Islets were incubated with vehicle or with 1, 3, or 10 M BEL for 30 min at 37°C. Islets were then homogenized and assayed for iPLA 2 Fig. 1 and is expressed as disintegrations/min per 20 g of islet protein (n ϭ 9).

lation of PA and PI does not mask a suppression by BEL of [ 3 H]arachidonate incorporation into islet PC and PE and that BEL has little effect on the phospholipid head-group classes into which [ 3 H]arachidonate is initially incorporated or on its subsequent transfer from PC to other lipids.
To examine further any contribution of de novo phospholipid synthesis (24) to [ 3 H]arachidonic acid incorporation into islet phospholipids, effects of medium glucose concentration were determined (Table II, top). De novo phospholipid synthesis from glucose requires glycolytic metabolism to triose phosphates and acylation of glycerol 3-phosphate and then lysophosphatidic acid to yield PA (24). Kinetic properties of the predominant islet glucose transporter and glucokinase cause islet glycolytic flux to rise severalfold as the medium glucose concentration increases from 5 to 20 mM (75), and this promotes islet de novo glycerolipid synthesis (60). Incorporation of [ 3 H]arachidonic acid into phospholipids was found not to be significantly affected by medium glucose concentration in control or BELtreated islets, although the latter exhibited greater incorporation at each glucose concentration (Table II, top). This indicates that de novo synthesis is not a major contributor to incorporation of [ 3 H]arachidonic acid into islet phospholipids under our conditions and cannot account for effects of BEL on incorporation. This is supported by the finding that conversion of [ 14 C]glucose to [ 14 C]glycerolipids is reduced rather than increased in BEL-treated islets (Table II, bottom).
Recovery of iPLA 2 activity in BEL-treated islets occurs slowly and substantial inhibition persists for many hours (51), raising the possibility that longer term arachidonate incorporation studies can be performed without repeated manipulation of the cells. When islets were exposed to 25 M BEL and then cultured for various periods, 99, 92, and 80% inhibition of iPLA 2 activity was observed at 0, 8, and 24 h after treatment, respectively (Fig. 8). Mean iPLA 2 activity for the entire 24-h period following BEL treatment was about 10% that in control islets. Similar experiments with INS-1 cells indicated that iPLA 2 activity was inhibited by 99, 92, and 60% at 0, 8, and 24 h, respectively, after treatment and that the mean iPLA 2 activity in the 24-h period following BEL treatment was about 20% of control values (Fig. 8).
Similar experiments were performed to examine effects of iPLA 2 inhibition on incorporation of arachidonate mass into INS-1 cell phospholipids to complement the radiochemical data. Electrospray ionization (ESI) mass spectrometric (MS) analyses of control INS-1 cell glycerophosphocholine (GPC) lipids as Li ϩ adducts indicated that arachidonate-containing species are not abundant in unsupplemented cells but that their abundance increases substantially within hours after adding arachidonic acid to the medium (Fig. 9).  Fig. 9C illustrates that at 24 h the abundance of this species increases further and that 18:0/20:4-GPC (m/z 816) also becomes abundant, a sequence consistent with reports that 16:0/20:4-GPC is a primary remodeling product and that 18:0/20:4-GPC is a secondary product from remodeling at both sn-1 and sn-2 positions (76 -80). Remodeling of INS-1 cell phospholipids with arachidonate was not suppressed by BEL treatment and was accelerated at 8 h after treatment (Fig. 10), despite the fact INS-1 cell iPLA 2 activity was inhibited by an average of about 96% during the first 8 h after treatment (Fig. 8).
Suppression of arachidonate incorporation by iPLA 2 inhibition in P388D1 cells is proposed to reflect reduction in lysophosphatidylcholine (LPC) acceptor molecules, the abundance of which is thought to limit the rate of arachidonate incorporation and to be governed by iPLA 2 activity (15,16). To determine effects of iPLA 2 inhibition on islet LPC mass, ESI tandem MS experiments were performed (Fig. 11). After incubation of islets with BEL or vehicle, internal standard 17:0-LPC, which does not occur naturally in islets, was added to islet lipid extracts, and LPC species were isolated by TLC and analyzed as Li ϩ adducts by ESI/MS/MS scanning for neutral loss of trimethylamine, an approach that yields a better signal to noise ratio than does monitoring the total ion current. Comparing relative intensities of ions for 17:0-LPC-Li ϩ (m/z 516) and 16:0-LPC-Li ϩ (m/z 502) or 18:0-LPC-Li ϩ (m/z 530) yields a linear standard curve over a wide concentration range that includes LPC levels observed in islets. as template in RT-PCR reactions with rat iPLA 2 -specific primers that were expected to yield a product of 528 bp in length, based on the rat iPLA 2 cDNA sequence (10). RT-PCR products were analyzed by agarose gel electrophoresis and visualized with ethidium bromide. Lane 1 represents a control reaction performed without an RT step to exclude contamination of the RNA with genomic DNA. Lanes 2 and 3 represent reaction products obtained from RT-PCR reactions performed with two separate preparations of INS-1 cell RNA. Lanes 4 and 5 represent products obtained from digestion of the DNA species visualized in lanes 2 and 3 with the restriction endonuclease SmaI, which was expected to yield products of 164 and 364 bp lengths, based on the rat iPLA 2 cDNA sequence (10). Fig. 11 illustrates that the two most abundant LPC species in islets are 16:0-LPC and 18:0-LPC, consistent with a previous report (63), that treatment of islets with BEL induces a modest decline in the abundance of these species relative to the internal standard, and that treatment with bee venom sPLA 2 induces the expected increase in abundance of these species. A series of similar experiments indicated that the resting islet LPC mass measured by this method (5.11 pmol/islet) corresponds closely to a measurement of 5.55 pmol/islet based on GC/MS quantitation islet LPC fatty acid content (63) and that the mean fall in LPC mass in BEL-treated islets is about 25% (Table III, (15,16). Although BEL induces a modest decline in islet LPC content (Table III), this is not associated with reduced [ 3 H]arachidonic acid incorporation into islet PC (Fig.  7), suggesting LPC level does not limit arachidonate incorporation into islet PC.
Although iPLA 2 activity is not required for arachidonate incorporation or phospholipid remodeling in islets, we have suggested that iPLA 2 might participate in insulin secretagogue-induced hydrolysis of arachidonic acid from islet phos-

TABLE I Inhibition of phosphatidate phosphohydrolase activity in pancreatic islet cytosol and membranes by NEM and by BEL
Cytosol and membranes were prepared from homogenized pancreatic islets and assayed for phosphatidate phosphohydrolase activity in medium containing 3 mM MgCl 2 by following conversion of [glycerol-U-14 C]phosphatidic acid to [ 14 C]diacylglycerol, and effects of treating islet cytosol with 8 mM NEM or 25 M BEL on the conversion were determined (n ϭ 12). The p values for the differences between the control condition and the NEM and BEL conditions were each less than 0.05, as was that between the NEM and BEL conditions.  pholipids, based on the finding that BEL suppresses insulin secretagogue-induced release of arachidonate metabolites from islets (81). To determine whether this might reflect inhibition of PAPH-1 (4, 57) rather than iPLA 2 , we compared effects of BEL and the PAPH inhibitor propranolol (29, 70) on accumulation of nonesterified arachidonate in islets stimulated with the insulin secretagogue 17 mM D-glucose alone or in combination with the muscarinic agonist carbachol (Fig. 12). As measured by isotope dilution GC/MS with [ 2 H 8 ]arachidonate as in-ternal standard, 17 mM D-glucose induced a doubling in accumulation of nonesterified arachidonic acid in islet incubations (Fig. 12), consistent with previous reports (82,83), and this response was greatly amplified by carbachol (Fig. 12). The effect of 17 mM glucose alone was completely prevented by BEL, and the response to the combination of 17 mM glucose plus  carbachol was reduced by about 70%. Propranolol did not reduce accumulation of nonesterified arachidonate induced either by 17 mM glucose alone or by the combination of 17 mM glucose plus carbachol (Fig. 12). DISCUSSION A housekeeping role for iPLA 2 in phospholipid remodeling has been suggested from iPLA 2 inhibition studies in P388D1 cells (4,15,16), but we find no evidence that iPLA 2 activity is required for incorporation of arachidonate or phospholipid re-modeling in pancreatic islets or insulinoma cells. A difficulty in studying iPLA 2 functions is that the most effective pharmacologic inhibitor of the enzyme, BEL (8 -12), also inhibits PAPH-1 (57), although BEL does not inhibit sPLA 2 or cPLA 2 enzymes at concentrations that effectively inhibit iPLA 2 (11)(12)(13)(14)57). One approach to this problem is to compare effects of BEL and of the PAPH inhibitor propranolol to distinguish functional consequences of inhibition of iPLA 2 or of PAPH (29,70), and a second employs antisense oligonucleotide suppression of iPLA 2 expression (16). We used the first approach here but were unable to suppress iPLA 2 activity in INS-1 insulinoma cells with antisense oligonucleotides using conditions applied to P388D1 cells (16) and several variations on them. It has been suggested that the efficacy of antisense suppression may be affected by iPLA 2 expression level (29). Amniotic WISH cells, for example, apparently express much higher iPLA 2 levels than P388D1 cells (29), and our inability to suppress INS-1 cell iPLA 2 with antisense oligonucleotides might reflect high iPLA 2 expression. Others have also encountered difficulties with lack of efficacy and lack of specificity of antisense approaches (84), which may offer no greater selectivity than conventional pharmacologic agents even when suppression of the target is achieved (85).
Despite difficulties in achieving completely selective iPLA 2 inhibition, the fact that incorporation of arachidonic acid into islet or INS-1 cell phospholipids is not suppressed when iPLA 2 is inhibited by 98% indicates that iPLA 2 activity is not required for this process in these cells. Pathways for arachidonate incorporation into islet cells appear to differ in several respects from those in P388D1 cells. Such differences might be expected from the marked difference in the esterified arachidonate content in islet phospholipids, which is among the highest of any known tissue (37)(38)(39), compared with that of P388D1 macrophage-like cells, which contain far lower levels of esterified arachidonate (27)(28)(29) than do native macrophages (25,26) and higher levels of nonesterified arachidonate (29) than other cells (30 -32). P388D1 cells may be deficient in arachidonate incorporation mechanisms employed by other cells and forced to use mechanisms not employed by cells that can maintain a high content of arachidonate-containing phospholipids.
Although arachidonate incorporation is unaffected by chelation of extracellular or intracellular Ca 2ϩ in P388D1 cells (15), chelation of intracellular Ca 2ϩ reduces but does not eliminate incorporation of arachidonate and palmitate into islets, when incorporation is compared in non-BAPTA-loaded islets and in BAPTA-loaded islets incubated with 3 H-fatty acid in Ca 2ϩ -free EGTA-containing medium. This suggests that both Ca 2ϩ -dependent and Ca 2ϩ -independent processes participate in fatty acid incorporation into islet phospholipids. Although Ca 2ϩ -independent hydrolysis of phospholipids by iPLA 2 to generate lysophospholipid acceptors is proposed to be the Ca 2ϩ -independent event that limits arachidonate incorporation into P388D1 cells (15,16), only about 60% suppression of arachido-nate incorporation into P388D1 cells is achieved under conditions where iPLA 2 activity is completely inhibited (15). This suggests that P388D1 cells also employ Ca 2ϩ -independent mechanisms other than iPLA 2 to generate a substantial fraction of LPC acceptors that limit arachidonate incorporation into these cells (15,16). Ca 2ϩ -independent mechanisms for generating lysophospholipids that do not involve iPLA 2 exist in other cells (86). The Ca 2ϩ -dependent processes that participate in fatty acid incorporation into islet cells have not yet been identified, but islet beta cells express at least two classes of Ca 2ϩ -dependent PLA 2 enzymes that might contribute to islet lysophospholipid content (14,87).
The role of LPC content in regulating fatty acid incorporation into GPC lipids and the contribution of iPLA 2 to LPC levels also differ between islets and P388D1 cells. BEL induces a more modest reduction in LPC levels in islets than in P388D1 cells, and this reduction does not impair arachidonate incorporation into islet PC but does so in P388D1 cells (15,16). This suggests that iPLA 2 makes only a modest contribution to islet LPC content and that islet LPC is maintained at sufficiently high levels by non-iPLA 2 -dependent mechanisms that arachidonate incorporation into islet PC is not limited by LPC level. ESI/ MS/MS measurements here and previous measurements by other methods (63) yield similar estimates of islet LPC content and indicate that LPC levels in rat islets are substantially higher than that in rat liver (53), and this may be one means employed by islets to maintain a higher content of arachidonate-containing phospholipids than liver (39) or P388D1 cells (27)(28)(29). Although LPC levels do not limit arachidonate incorporation into islet phospholipids, our experiments with IL-1 here and elsewhere (51) indicate that generation of arachidonoyl-CoA may limit such incorporation under circumstances where production of ATP is impaired or CoASH degradation is accelerated. IL-1 stimulates islet production of nitric oxide, which impairs ATP generation by islet mitochondria (68) and reacts with free thiol groups (69), such as those of CoASH. Both ATP and CoASH are required to form arachidonoyl-CoA from arachidonic acid (26,33,34), an obligatory step for its incorporation into phospholipids (35,36).
Although we never observe suppression of fatty acid incorporation into islet or insulinoma cell phospholipids when iPLA 2 activity is inhibited, stimulation of incorporation is often observed. The simplest explanation is that net incorporation of fatty acid reflects the difference between acylation and deacylation and that iPLA 2 inhibition reduces deacylation but does not affect acylation because iPLA 2 -catalyzed generation of ly-  Fig. 11, and the quantity of 16:0-LPC (m/z 502) and 18:0-LPC (m/z 530) were determined relative to the internal standard 17:0-LPC (m/z 516) by interpolation from a standard curve (n ϭ 6). Bottom, islets were incubated with [ 3 H]choline for 40 h at 37°C and were treated with vehicle (Control) or with 25 M BEL for 30 min at 37°C. The islets were then incubated for 10 min at 37°C. At the end of the incubation, islet lipids were extracted and analyzed by TLC to isolate LPC, which was extracted from the plate. The 3 H content of the LPC fraction was determined by liquid scintillation spectrometry (n ϭ 6). sophospholipid acceptors is not required for fatty acid incorporation into islet phospholipids. The similar effects of BEL on incorporation of both arachidonate and palmitate are consistent with the fact iPLA 2 hydrolyzes phospholipids with palmitate substituents at least as readily as those with arachidonate substituents (8,88). Although palmitate incorporation into P388D1 cells has been taken to reflect primarily de novo phospholipid synthesis (15), palmitate incorporation into islet phospholipids under our conditions proceeds predominantly by deacylation/reacylation because BEL impairs de novo phospholipid synthesis but does not suppress palmitate incorporation. Deacylation/reacylation reactions with palmitate occur as rapidly as those with arachidonate in some cells (76 -80). We have not determined whether palmitate incorporated into islet phospholipids under our conditions resides primarily in the sn-1 or sn-2 position, but islets contain substantial amounts of GPC species with palmitate as the sn-2 substituent, including 16:0/ 16:0-GPC (39). Deacylation/reacylation of phospholipid sn-1 substituents also occurs (76 -80), and iPLA 2 can catalyze hydrolysis of sn-1 substituents of phospholipids (8).
The fact that iPLA 2 is not required for arachidonate incorporation or phospholipid remodeling in beta cells suggests that it may play other roles in these cells. Various signaling functions for iPLA 2 have been suggested (10,(17)(18)(19)(20)(21), and as demonstrated here and elsewhere (81)(82)(83), insulin secretagogues induce hydrolysis of arachidonic acid from islet phospholipids. Our findings that insulin secretagogue-induced accumulation of nonesterified arachidonate in islet incubations is reduced by BEL corroborates the observation that insulin secretagogueinduced release of arachidonate metabolites from islets is reduced by BEL (81). The failure of the PAPH inhibitor propranolol (29,70) to block insulin secretagogue-induced accumulation of nonesterified arachidonate in islet incubations suggests that PAPH is not the BEL-sensitive target that participates in insulin secretagogue-induced hydrolysis of arachidonate from islet phospholipids, although others (4,57) have suggested that this process involves sequential actions of phospholipase D, PAPH, and diglyceride lipase. Our findings are consistent with observations by others that 250 M propranolol maximally inhibits islet PAPH but does not block islet accumulation of nonesterified arachidonate and stimulates insulin secretion (71). PA produced by phospholipase D and hydrolyzed by PAPH may also be an unlikely source of arachidonic acid in signaling events because PA from this route contains primarily saturated or monounsaturated fatty acids and little arachidonate (89).
Signaling roles have been proposed for iPLA 2 in apoptosis induced by the Fas/Fas ligand system in U937 cells (20), in apoptosis induced by Ca 2ϩ store depletion in beta cells (90), and in leukotriene production by granulocytes (17). The possibility that iPLA 2 plays different roles and is subject to different mechanisms of regulation in different cells is suggested by the fact that stimuli which induce iPLA 2 -catalyzed arachidonate release and leukotriene production in human granulocytes fail to induce these events in human lymphocyte lines, even though both classes of cells express iPLA 2 and leukotriene biosynthetic enzymes (17,21). The facts that the active form of iPLA 2 is an oligomer (8,88), that multiple splice variants exist in some cells (21,41), and that heterooligomeric complexes among different splice variants exhibit altered catalytic activity (21) suggest that iPLA 2 may be subject to complex regulatory mechanisms that differ among cell types. Our findings that iPLA 2 activity is not required for arachidonate incorporation into beta cell phospholipids, even though iPLA 2 appears to participate in such incorporation in P388D1 cells (15,16), is consistent with the notion that the enzyme plays diverse roles in different cell types.