Electrospray ionization/mass spectrometric analyses of human promonocytic U937 cell glycerolipids and evidence that differentiation is associated with membrane lipid composition changes that facilitate phospholipase A2 activation.

Upon differentiation, U937 promonocytic cells gain the ability to release a large fraction of arachidonate esterified in phospholipids when stimulated, but the mechanism is unclear. U937 cells express group IV phospholipase A(2) (cPLA(2)), but neither its level nor its phosphorylation state increases upon differentiation. A group VI PLA(2) (iPLA(2)) that is sensitive to a bromoenol lactone inhibitor catalyzes arachidonate hydrolysis from phospholipids in some cells and facilitates arachidonate incorporation into glycerophosphocholine (GPC) lipids in others, but it is not known whether U937 cells express iPLA(2). We confirm that ionophore A23187 induces substantial [(3)H]arachidonate release from differentiated but not control U937 cells, and electrospray ionization mass spectrometric (ESI/MS) analyses indicate that differentiated cells contain a higher proportion of arachidonate-containing GPC species than control cells. U937 cells express iPLA(2) mRNA and activity, but iPLA(2) inhibition impairs neither [(3)H]arachidonate incorporation into nor release from U937 cells. Experiments with phosphatidate phosphohydrolase (PAPH) and phospholipase D (PLD) inhibitors coupled with ESI/MS analyses of PLD-PAPH products indicate that differentiated cells gain the ability to produce diacylglycerol (DAG) via PLD-PAPH. DAG promotes arachidonate release by a mechanism that does not require DAG hydrolysis, is largely independent of protein kinase C, and requires cPLA(2) activity. This may reflect DAG effects on cPLA(2) substrate state.

Upon differentiation, U937 promonocytic cells gain the ability to release a large fraction of arachidonate esterified in phospholipids when stimulated, but the mechanism is unclear. U937 cells express group IV phospholipase A 2 (cPLA 2 ), but neither its level nor its phosphorylation state increases upon differentiation. A group VI PLA 2 (iPLA 2

) that is sensitive to a bromoenol lactone inhibitor catalyzes arachidonate hydrolysis from phospholipids in some cells and facilitates arachidonate incorporation into glycerophosphocholine (GPC) lipids in others, but it is not known whether U937 cells express iPLA 2 . We confirm that ionophore A23187 induces substantial [ 3 H]arachidonate release from differentiated but not control U937 cells, and electrospray ionization mass spectrometric (ESI/MS) analyses indicate that differentiated cells contain a higher proportion of arachidonate-containing GPC species than control cells. U937 cells express iPLA 2 mRNA and activity, but iPLA 2 inhibition impairs neither [ 3 H]arachidonate incorporation into nor release from U937 cells. Experiments with phosphatidate phosphohydrolase (PAPH) and phospholipase D (PLD) inhibitors coupled with ESI/MS analyses of PLD-PAPH products indicate that differentiated cells gain the ability to produce diacylglycerol (DAG) via PLD-PAPH. DAG promotes arachido-
nate release by a mechanism that does not require DAG hydrolysis, is largely independent of protein kinase C, and requires cPLA 2 activity. This may reflect DAG effects on cPLA 2 substrate state.
U937 cells are derived from a human histiocytic lymphoma and differentiate into monocyte-like cells when treated with various agents, including dimethyl sulfoxide (Me 2 SO) 1 (1). Dif-ferentiated but not undifferentiated U937 cells release a large fraction of arachidonic acid esterified in their phospholipids when stimulated with various agonists, including Ca 2ϩ ionophores (2)(3)(4). This has been attributed to enhanced activation of group IV phospholipase A 2 (cPLA 2 ) in differentiated U937 cells because the response is suppressed by cPLA 2 inhibitors (3,4). Like other PLA 2 enzymes (5-9), cPLA 2 catalyzes hydrolysis of the sn-2 fatty acid substituent from glycerophospholipid substrates to yield a free fatty acid and a 2-lysophospholipid.
How cPLA 2 might be more readily activated in differentiated U937 cells than in control cells is incompletely understood. Undifferentiated U937 cells express cPLA 2 (2)(3)(4), and neither its level nor phosphorylation state increase upon differentiation (4). Both undifferentiated and differentiated U937 cells undergo a similar rise in cytosolic [Ca 2ϩ ] when treated with Ca 2ϩ ionophores (3,4). It has been suggested that cPLA 2 activation in differentiated U937 cells is coupled to depletion of internal Ca 2ϩ stores and that products of its action mediate store depletion-induced Ca 2ϩ entry (3).
A recently cloned (11)(12)(13)(14)(15) group VI PLA 2 (iPLA 2 ) does not require Ca 2ϩ for catalysis and is inhibited by a bromoenol lactone (BEL) suicide substrate (16,17) that does not inhibit sPLA 2 or cPLA 2 enzymes at comparable concentrations (7, 16 -18). The iPLA 2 enzyme appears to be activated by agents that induce Ca 2ϩ -store depletion in some cells (19 -21) and to catalyze arachidonate hydrolysis from phospholipids of such cells by a mechanism that does not require a rise in cytosolic [Ca 2ϩ ] (19,20). Because Ca 2ϩ -store depletion also induces arachidonate release from differentiated U937 cells (3), these observations (19 -21) raise the questions of whether U937 cells express iPLA 2 and whether differentiation increases its expression. Differentiation induced expression of Ca 2ϩ -independent PLA 2 activity does occur in HL-60 cells (22), a human myeloblast cell line that, like U937 cells, differentiates when treated with Me 2 SO (1).
U937 cells have been reported to express a Ca 2ϩ -independent PLA 2 activity that participates in Fas-induced apoptosis and arachidonate release under conditions in which cPLA 2 is inactivated proteolytically by caspases (23), but whether this activity resides in iPLA 2 gene product(s) has not been established. The possible participation of iPLA 2 in U937 cell apoptosis (23) is of interest because Ca 2ϩ -store depletion induces apoptosis of pancreatic islet ␤-cells by a mechanism that does not require a rise in cytosolic [Ca 2ϩ ] but that does require arachidonate release and metabolism (24). Ca 2ϩ -store depletion also induces arachidonic acid hydrolysis from ␤-cell phospholipids by a BEL-sensitive mechanism thought to reflect iPLA 2 action (20).
The iPLA 2 enzyme has also been proposed to provide lysophosphatidylcholine (LPC) acceptors for arachidonate incorporation into glycerophosphocholine (GPC) lipids in murine P388D1 cells (25,26), which, like U937 cells, are of monocytemacrophage lineage. If increased iPLA 2 expression occurred upon U937 differentiation and accelerated arachidonate incorporation into GPC, any resultant accumulation of arachidonate-containing GPC lipids might increase hydrolysis of arachidonate from U937 cell phospholipids by cPLA 2 . Increasing the mole fraction of arachidonate-containing GPC species in lipid vesicles has been reported to facilitate hydrolysis of substrates by cPLA 2 in a cooperative manner (27), although this effect is influenced by other lipids in such vesicles (28).
Here we examine effects of differentiation on U937 cell GPC lipid composition, expression of Ca 2ϩ -independent PLA 2 activity, sensitivity of this activity to BEL, and effects of BEL on arachidonate incorporation into and release from U937 cell phospholipids.
Extraction and Chromatographic Analyses of Phospholipids-Phos-pholipids from U937 cells were extracted with chloroform/methanol under neutral conditions (30). Incorporation of [ 3 H]arachidonate into phospholipids was determined by TLC, as described previously (25,26,31). Phospholipid head group classes were isolated by NP-HPLC, as described previously (31). Mass Spectrometric Analyses of Phospholipids and Diglycerides-ESI/MS and ESI/MS/MS analyses were performed on a Finnigan (San Jose, CA) TSQ-7000 triple stage quadrupole mass spectrometer, as described previously (32,33). To quantitate arachidonate mass in phospholipids, alkaline hydrolysis was performed in the presence of [ 2 H 8 ]arachidonate internal standard, and liberated fatty acids were extracted, converted to pentafluorobenzyl ester derivatives, and measured by isotope dilution GC/NIEC/MS, as described previously (31).
Isolation of RNA from U937 Cells, Reverse Transcription, and Polymerase Chain Reactions-RNA was isolated from U937 cells and RT-PCR performed using standard procedures, as described previously (13). To amplify iPLA 2 cDNA, primers used were sense (5Ј-TTCGGAG-CAGAAGTGGACAC-3Ј) and antisense (5Ј-TGAAAGTACATGCCG-CG-CATG-3Ј). PCR with these primers amplifies cDNA species for long (LH-iPLA 2 ) and short (SH-iPLA 2 ) isoforms of human iPLA 2 (15) to yield products of 515 and 353 bp, respectively, and both products were observed using U937 cell RNA as template. These products were then digested with endonuclease AvaI, which converted the RT-PCR product for LH-iPLA 2 into two products (396 and 119 bp) and for SH-iPLA 2 (353 bp) into two products (234 and 119 bp). These products correspond to the lengths expected from the corresponding cDNA sequences (15).
Isolation of Pancreatic Islets and Culture of INS-1 Insulinoma Cells-Islets were isolated from Harlan Sprague-Dawley rats, as described previously (34). INS-1 insulinoma cells provided by Dr. Christopher Newgard (University of Texas-Southwestern, Dallas) were cultured as described (35).
Immunoblotting Analyses of U937 Cell and INS-1 Cell iPLA 2 Protein-U937 cell or INS-1 cell cytosolic proteins were analyzed by SDSpolyacrylamide gel electrophoresis, and immunoreactive iPLA 2 was visualized with an antibody from Dr. Simon Jones (Genetics Institute, Boston), as described previously (31).

Time Course of Incorporation of [ 3 H]Arachidonic
Acid into U937 Cell Phospholipids-U937 cells (10 6 /ml) were preincubated (15 min, 37°C) with vehicle only (control) or BEL (25 M) in RPMI 1640 medium containing supplements described above. After preincubation, incorporation experiments (10 -60 min, 37°C) were initiated by adding [ 3 H]arachidonic acid (0.5 Ci/ml, 5 nM). Cells were washed three times to remove unincorporated label. Lipids were extracted and analyzed by TLC to isolate phospholipids, and their 3 H content was determined as described (31).
Determination of Phosphatidylethanol Accumulation-For radiochemical experiments, U937 cells were labeled (24 h) with [ 3 H]palmitic acid (3 Ci/10 6 cells), washed free of unincorporated label, and incubated without or with ionophore A23187 in the presence or absence of ethanol. Lipids were extracted, mixed with 18:1/18:1-phosphatidylethanol standard, and analyzed by TLC on silica gel G plates (Whatman) using the upper phase of a solvent system prepared with ethyl acetate/ isooctane/acetic acid/water (13:2:3:10). The phosphatidylethanol band was identified with iodine vapor and its 3 H content determined. For ESI/MS analyses of U937 cell phosphatidylethanol species, experiments were similar, except that cells were not labeled with [ 3 H]palmitate, and no standard phosphatidylethanol was added.
Determination of Diglyceride Accumulation-For radiochemical experiments, U937 cells were labeled with [ 3 H]palmitic acid as above and incubated without or with ionophore A23187 (1.5 M) for 1-10 min. Lipids were extracted, mixed with 10 g of dinonadecadienoin standard, and analyzed by TLC to isolate diacylglycerol (DAG) (31). The 3 H content of the DAG was then determined. For ESI/MS analyses of U937 cell DAG species, experiments were similar, except that cells were not labeled with [ 3 H]palmitate, and no dinonadecadienoin standard was added.
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.

Effects of Inducing Differentiation of U937 Cells with Dimethyl Sulfoxide on [ 3 H]Arachidonic Acid Incorporation into
and Release from Phospholipids-When control, undifferentiated U937 cells were labeled with [ 3 H]arachidonate and incubated with or without Ca 2ϩ ionophore A23187, less than 2% of incorporated 3 H was released into the medium. In similar experiments with Me 2 SO-differentiated U937 cells, A23187 induced release of 18% of incorporated 3 H, and this was suppressed by the PLA 2 inhibitor (39) PGBx (not shown). This confirms reports (2-4) that differentiated but not undifferentiated U937 cells release a large fraction of [ 3 H]arachidonate esterified in their phospholipids upon stimulation with Ca 2ϩ ionophores.
When phospholipids into which [ 3 H]arachidonic acid was incorporated were examined by NP-HPLC, control U937 cells incorporated the majority of 3 H into glycerophosphoethanolamine (GPE) lipids and a smaller amount into glycerophosphocholine (GPC) lipids (Table I). Differentiated U937 cells incorporated a greater proportion of 3 H into GPC lipids than did control cells (Table I). Quantitation of arachidonate content of phospholipid extracts by isotope dilution GC/NIEC/MS after saponification indicated that the amount of esterified arachidonate per nmol of lipid phosphorus in differentiated U937 cells exceeded that in control cells (Table I). This suggests that differentiation of U937 cells is associated with an increased content of arachidonate-containing phospholipid species, and this possibility was examined by electrospray ionization mass spectrometric (ESI/MS) analyses of U937 cell GPC and GPE lipids isolated by NP-HPLC.  (Table  II).

Electrospray Ionization Mass Spectrometric Analyses of U937 Cell GPC and GPE Lipids-
Comparison of ESI/MS spectra in Fig. 1, A and B, indicates that Me 2 SO-induced differentiation of U937 cells is associated with a decrease in the relative abundance of some species with lower m/z values (e.g. m/z 724 and 738) and an increase in species with higher m/z values (e.g. m/z 816, 814, and 788). The tandem spectrum obtained from CAD of the ion at m/z 816 from differentiated U937 cells is illustrated in Fig. 1D and contains an ion reflecting loss of trimethylamine plus stearic acid (m/z 473) and a less abundant ion (m/z 453) reflecting loss of trimethylamine plus arachidonic acid. These features indicate that the parent ion at m/z 816 represents 18:0/20:4-GPC-Li ϩ , as confirmed by ions at m/z 341 (loss of lithiated phosphocholine plus arachidonate as a ketene) and m/z 532 (loss of stearic acid). Tandem spectra obtained from ions at m/z 788 and 814 in the GPC-Li ϩ mixture from Me 2 SO-treated U937 cells indicated that these ions represent 16:0/20:4-GPC-Li ϩ and 18:1/20:4-GPC-Li ϩ , respectively. Table II summarizes other species identified in this mixture and indicates that the fraction of the ion current for the GPC-Li ϩ mixture contributed by arachidonate-containing species increases upon cell differentiation.
Negative ion ESI/MS analyses of U937 cell GPE lipids indicated that arachidonate-containing species contributed a 5-fold  Fig. 1. Cells were then collected by centrifugation and their lipids extracted. Lipid phosphorus was measured in 1 aliquot of the extract. Arachidonic acid mass was measured by isotope dilution GC/NIEC/MS in a second aliquot before and after saponification to determine esterified arachidonate content. The remainder of lipid extracts was analyzed by NP-HPLC, and the 3 H and lipid phosphorus contents of isolated phospholipid head group classes were determined. Values for cells incubated without A23187 are designated "Resting (R)." The difference between that value and the value observed for A23187-stimulated cells is designated "R-A23187." For the latter parameter, positive or negative values reflect A23187-induced decreases or increases, respectively. Mean values (ϮS.E.) are displayed (n ϭ 4). Statistically significant differences were observed between undifferentiated and Me 2 SO-treated cells only for resting and R-A23187 values of esterified arachidonate mass and phosphatidylcholine 3  higher fraction of the total ion current in undifferentiated cells than was the case for GPC lipids and that differentiation induced a modest rise in the relative abundance of such GPE species (not shown).
Characterization of U937 Cell Ca 2ϩ -independent Phospholipase A 2 Activity-The finding that differentiation of U937 cells is associated with an increased proportion of arachidonate-containing species in GPC lipids raises the question of whether this reflects increased expression of iPLA 2 , which is proposed to generate LPC acceptors for arachidonate incorporation into GPC lipids in P388D1 macrophage-like cells (25,26). RT-PCR analyses indicated that both control and Me 2 SOtreated U937 cells express iPLA 2 mRNA (not shown). Immunochemical analyses indicated that any expression of iPLA 2 protein in U937 cells is substantially lower than that in INS-1 insulinoma cells (not shown). To examine the possibility that expression of iPLA 2 mRNA by U937 cells is associated with production of functional enzyme, activity assays were performed. Features of iPLA 2 enzymatic activity that distinguish it from other characterized PLA 2 are that iPLA 2 does not require Ca 2ϩ for catalysis, is stimulated by ATP, and is susceptible to inhibition by BEL (11)(12)(13)(14)(15). Fig. 2A illustrates that both control and Me 2 SO-treated U937 cells express cytosolic Ca 2ϩ -independent PLA 2 activity, and addition of ATP to cytosol from Me 2 SO-treated cells induced a concentration-dependent increase in activity. Fig. 2B illustrates that adding BEL to cytosol from Me 2 SO-treated U937 cells induced a concentration-dependent decrease in Ca 2ϩ -independent PLA 2 activity up to a concentration of 20 M, but about 50% of total activity was resistant even to 100 M BEL. Control cell cytosol also exhibited both BEL-sensitive and BEL-resistant components of Ca 2ϩ -independent PLA 2 activity. The resistance to BEL of a large fraction of Ca 2ϩ -independent PLA 2 activity in U937 cell cytosol contrasts to the case in pancreatic islets (Fig. 3), which express iPLA 2 (13,15,31). At 10 M BEL, virtually all Ca 2ϩ -independent PLA 2 activity in islet cytosol is inhibited (Fig. 3), but higher BEL concentrations (IC 50 ϳ35 M) are required to inhibit islet cytosolic phosphatidate phosphohydrolase-1 (PAPH), which is also sensitive to BEL (31,37). Treating intact U937 cells with varied BEL concentrations yielded inhibition curves similar to those obtained from adding BEL to cytosol (not shown).
Effects of BEL on [ 3 H]Arachidonic Acid Incorporation into and Release from U937 Cell Phospholipids-Because U937 cells express iPLA 2 mRNA and some BEL-sensitive Ca 2ϩ -independent PLA 2 activity, BEL was used to examine iPLA 2 participation in incorporation of [ 3 H]arachidonate into and release from U937 cell phospholipids. Fig. 4 illustrates that treating Me 2 SO-differentiated U937 cells with 25 M BEL did not suppress incorporation of [ 3 H]arachidonate into phospholipids, and similar results were obtained with control cells (not shown). Inhibition of iPLA 2 with BEL also fails to suppress [ 3 H]arachidonate incorporation into islets or insulinoma cells (31) but does so in P388D1 cells (25). Any iPLA 2 activity expressed by U937 cells is thus not required for [ 3 H]arachidonate incorporation into their phospholipids. Fig. 5 illustrates that treating U937 cells with 12.5 M BEL does not significantly affect A23187-induced release of esterified [ 3 H]arachidonate from their phospholipids. At this concentration, about 90% of U937 cell cytosolic Ca 2ϩ -independent PLA 2 activity that is sensitive to any concentration of BEL up to 100 M is inhibited, and virtually all iPLA 2 activity is inhibited by 12.5 M BEL in other cells that express the enzyme (11)(12)(13)(14)(15). This suggests that iPLA 2 does not participate in A23187-induced release of [ 3 H]arachidonate from phospholipids of differentiated U937 cells. Further evidence that iPLA 2 does not participate in this phenomenon is that removing Ca 2ϩ from the medium and adding the Ca 2ϩ chelator EGTA prevents A23187-induced [ 3 H]arachidonic acid release from differentiated U937 cells (Fig. 5). These findings contrast with those in pancreatic islets or vascular myocytes, in which A23187 or other Ca 2ϩ store-depleting agents induce a BEL-sensitive re-lease of arachidonic acid even in the presence Ca 2ϩ chelators that is thought to reflect iPLA 2 -catalyzed hydrolysis of arachidonate from phospholipids (19,20).
Phosphatidate Phosphohydrolase and [ 3 H]Arachidonic Acid Release from U937 Cells-In contrast to findings with 12.5 M BEL, Fig. 6 illustrates that 50 M BEL substantially diminishes A23187-induced [ 3 H]arachidonic acid release from differentiated U937 cells. Because such high concentrations of BEL inhibit PAPH, we examined the potential involvement of PAPH in U937 cell [ 3 H]arachidonate release. Propranolol is a widely used PAPH inhibitor (37,42,43) and was also found to suppress A23187-induced release of [ 3 H]arachidonic acid from differentiated U937 cells (Fig. 6). In signaling events, substrate for PAPH is often provided by phospholipase D (PLD), which hydrolyzes phospholipids to yield phosphatidic acid (42)(43)(44)(45)(46)(47). In ions were subjected to CAD and tandem MS to identify molecular species. The combined chain length and number of double bonds in fatty side chains are tabulated, as is chain length and double bond number for each substituent of the major species at each m/z value. At some m/z values, more than one isomer was observed. The designation "e" denotes an sn-1 alkyl ether linkage and "a" an sn-1 acyl linkage.  3. BEL concentration dependence for inhibition of phosphatidate phosphohydrolase-1 and Ca 2؉ -independent PLA 2 activity in islet cytosol. One set of aliquots of islet cytosol was treated with varied [BEL] and a second with BEL-free vehicle. Ca 2ϩ -independent PLA 2 activity was measured as in Fig. 1. Phosphatidate phosphohydrolase activity was measured in the absence and presence of NEM, and PAPH-1 activity was calculated from the NEM-inhibited component (n ϭ 12). the presence of ethanol or n-butyl alcohol, PLD produces the transphosphatidylation products phosphatidylethanol (PtEt) or phosphatidylbutanol, which are resistant to hydrolysis by PAPH, and their accumulation provides a marker for PLD activation (42)(43)(44)(45)(46)(47). Because ethanol and butanol suppress diglyceride formation by the PLD-PAPH pathway, these alcohols inhibit functional responses that require such diglyceride generation in some systems (42)(43)(44)(45)(46)(47).
When [ 3 H]arachidonate-labeled differentiated U937 cells were incubated with A23187 and varied concentrations of ethanol or n-butyl alcohol, a concentration-dependent inhibition of [ 3 H]arachidonic acid release occurred with both alcohols (Fig.  7). To examine whether this might reflect effects on PLD, [ 3 H]palmitate-labeled U937 cells were incubated with A23187 and ethanol, and cell lipids were extracted and analyzed by TLC to isolate [ 3 H]PtEt. Fig. 8 illustrates that time-dependent accumulation of [ 3 H]PtEt occurred in A23187-stimulated differentiated U937 cells, reflecting PLD activation. No such effect was observed in undifferentiated cells (Fig. 8), indicating that ability to activate PLD upon A23187 stimulation is acquired upon differentiation. All of those ions represent species containing only saturated or monounsat-urated but not polyunsaturated fatty acid substituents, consistent with a report that phosphatidyl alcohols produced by PLD contain mainly saturated or monounsaturated fatty acid substituents and that polyunsaturated substituents are not abundant (46). The fact that arachidonate-containing species are not abundant in the U937 cell PtEt mixture (Fig. 9C) indicates that diglyceride lipase-catalyzed hydrolysis of diglycerides derived from sequential actions of PLD and PAPH is not a major source of arachidonate released by A23187-stimulated U937 cells, and the diglyceride lipase inhibitor RHC80267 (42) did not influence A23187-induced arachidonate release (not shown). (Fig.  10A). This was first observed after 5 min and increased further at 10 min. Because peak [ 3 H]PtEt accumulation occurred at 5 min in similar experiments performed in the presence of ethanol (Fig. 8), the time course of [ 3 H]diglyceride accumulation (Fig. 10A) is consistent with its generation by hydrolysis of the PLD product PA. At concentrations that impair arachidonate release (Fig. 7), ethanol or n-butyl alcohol suppressed A23187induced [ 3 H]diglyceride accumulation in differentiated U937 cells (not shown), suggesting that [ 3 H]diglycerides arise from sequential actions of PLD and PAPH and that endogenous diglycerides might participate in arachidonic acid release. Addition of 5-40 M (1-oleoyl, 2-acetyl)-sn-glycerol (OAG), a cellpermeant diglyceride (48), amplified A23187-induced [ 3 H]arachidonic acid release from differentiated U937 cells in a concentration-dependent manner (Fig. 10B) and reversed inhibition of release by ethanol or n-butyl alcohol (not shown). OAG did not affect [ 3 H]arachidonic acid release in the absence of A23187 or with undifferentiated U937 cells (Fig. 10B).

Accumulation of Endogenous Diglycerides and Effects of Exogenous Diglyceride in U937 Cells-When the [ 3 H]diglyceride content of [ 3 H]palmitate-labeled U397 cells was determined by TLC, A23187 induced time-dependent [ 3 H]diglyceride accumulation in differentiated but not in undifferentiated cells
To verify that diglyceride accumulation occurred in differentiated U937 cells incubated with A23187 and to identify molecular species, ESI/MS analyses of Li ϩ adducts of cellular diglycerides isolated by TLC were performed. Complexation of neutral acylglycerols with Li ϩ permits their identification by ESI/MS, and tandem spectra from CAD of [MϩLi] ϩ ions permits assignment of fatty acid substituents (49). Fig. 11 illustrates tandem spectra from the three most abundant [M ϩ Li] ϩ ions in the diglyceride mixture from differentiated U937 cells incubated with A23187, and the abundance of species represented by these ions increased markedly upon A23187 stimulation.  (Fig. 11A). Ions reflecting combined losses involving ␣,␤-unsaturated derivatives of fatty acid substituents are also observed in tandem spectra of triacylglycerol Li ϩ adducts (49). Ions representing Li ϩ complexes of palmitic (m/z 263) and oleic acid (m/z 289) are also observed (Fig. 11A). These features identify the parent diacylglycerol (DAG) species as 16:0/18:1-DAG, and it corresponds to the most abundant of the PtEt species (m/z 701, 16:0/18:1-PtEt) obtained from differentiated U937 cells incubated with A23187 and ethanol (Fig. 9).
Similar features of tandem spectra obtained from CAD of U937 cell DAG-Li ϩ ions at m/z 627 (Fig. 11B) and 573 (Fig.  11C) identify the parent species as 18:1/18:1-DAG and as 16:0/ 16:1-DAG, respectively. The former spectrum contains an ion representing the Li ϩ complex of oleic acid (m/z 289), as well as ions representing losses of oleate in various forms (Fig. 11B). The latter spectrum contains ions representing Li ϩ complexes of palmitic (m/z 263) and palmitoleic acid (m/z 261), as well as The facts that each diglyceride species in Fig. 11 contains only saturated or monounsaturated and not polyunsaturated fatty acid substituents and that these diglycerides correspond to PtEt species obtained from differentiated U937 cells incubated with A23187 and ethanol ( Fig. 9) are consistent with generation of the diglycerides by a PLD-PAPH pathway.
Because both diglycerides (Fig. 10) and Ca 2ϩ entry (Fig. 5) participate in A23187-induced [ 3 H]arachidonic acid release from differentiated U937 cells and because both diglycerides and Ca 2ϩ activate some protein kinase C (PKC) isoforms (50), effects of the PKC inhibitor staurosporine (51) were examined. Staurosporine inhibits cellular PKC activity at concentrations above 10 nM, and complete inhibition is achieved at 100 nM (52,53). Within this range, staurosporine caused a modest concentrationdependent attenuation of A23187-induced [ 3 H]arachidonic acid release from differentiated U937 cells (Fig. 12A). Incomplete suppression was achieved even at 100 nM staurosporine, however, and about two-thirds of release was resistant to PKC inhibition (Fig. 12A), indicating that PKC-independent mechanisms account for most A23187-induced [ 3 H]arachidonic acid release. That cPLA 2 activity is required for such [ 3 H]arachidonic acid release is suggested by the fact that the cPLA 2 inhibitor arachidonoyl trifluoromethyl ketone (54) suppressed the vast majority of release from A23187-treated, differentiated U937 cells (Fig. 12B) with a concentration dependence virtually identical to that for inhibition of cPLA 2 -catalyzed arachidonate release from human platelets (55). Arachidonoyl trifluoromethyl ketone also suppressed amplification of [ 3 H]arachidonate release from A23187-treated, differentiated U937 cells by exogenous OAG (not shown).
The observations in Figs. 6 -12 indicate that enhanced arachidonate release in response to A23187 that occurs upon differentiation of U937 cells requires diglyceride generation by a PLD-PAPH pathway, that most of the effect of diglycerides is independent of PKC, and that the majority of arachidonate release requires cPLA 2 activity.

DISCUSSION
One PKC-independent effect of diglycerides is to increase susceptibility of phospholipids in membrane mixtures to hydrolysis by PLA 2 (56 -60), and diglyceride generation might represent a mechanism whereby cells can regulate cPLA 2 activity by changing the physical state of its substrates. Maximal stimulation of cPLA 2 activity toward arachidonate-containing GPC substrates occurs at a concentration of 25 M with the diglyceride OAG (58), which is within the range in which OAG amplifies A23187-induced arachidonate hydrolysis from membranes of differentiated U937 cells. The effect of diglycerides to enhance susceptibility of phospholipids in lipid mixtures to hydrolysis by cPLA 2 could reflect the ability of diglycerides to separate neighboring phospholipid head groups and to increase access of cPLA 2 to the glycerol backbone (42, 56 -60). Diglycerides cause phase changes in membrane bilayers (56) and increase membrane penetration, hydrophobic membrane binding, and interfacial enzymatic activity of cPLA 2 (60). Exogenous diglycerides also enhance A23817-induced, cPLA 2 -catalyzed [ 3 H]arachidonate release from intact neutrophils by a PKC-independent mechanism that is thought to reflect an effect on substrate state (39).
These observations and our findings suggest that changes in membrane lipid composition facilitate activation of cPLA 2 in differentiated U937 cells in a manner that does not require changes in the level of the enzyme or its phosphorylation state. This model can rationalize some puzzling observations about the role of cPLA 2 in arachidonate release from U937 cells.
Upon differentiation induced by Me 2 SO or other agents, U937 cells acquire ability to release a large fraction of [ 3 H]arachidonate esterified in their phospholipids when stimulated with various agonists (2)(3)(4), including Ca 2ϩ ionophores, and there is a corresponding release of arachidonate mass into the medium and a decline in arachidonate mass esterified in phospholipids, as measured by mass spectrometry (62). The mechanism underlying this effect has remained obscure, although studies with the inhibitors arachidonoyl trifluoromethyl ketone and PGBx and with anti-cPLA 2 antibodies introduced into permeabilized cells indicate that cPLA 2 is responsible for arachidonate hydrolysis from phospholipids of differentiated U937 cells (2)(3)(4). Our finding that A23187-induced [ 3 H]arachidonate release from differentiated U937 cells requires Ca 2ϩ entry from the extracellular space and does not occur in EGTA-containing medium is consistent with participation of cPLA 2 or other Ca 2ϩ -dependent phospholipases in the release phenomenon.
Arachidonate release from differentiated U937 cells can be induced when Ca 2ϩ in the extracellular medium is replaced by Sr 2ϩ , which supports activation of cPLA 2 but not of sPLA 2 (3). This finding and the fact that neither control nor differentiated U937 cells express sPLA 2 activity or protein (4) indicate that sPLA 2 is not the Ca 2ϩ -dependent phospholipase involved in arachidonate release. The enzyme iPLA 2 appears to catalyze FIG. 11. ESI/MS identification of diglyceride species accumulating in Me 2 SO-differentiated U937 cells incubated with ionophore A23187. Control or Me 2 SO-differentiated U937 cells were incubated (10 min) with 1.5 M A23187 and their lipids analyzed by TLC to isolate diglycerides, which were analyzed by ESI/MS as Li ϩ adducts. A is the tandem spectrum from the [M ϩ Li] ϩ ion at m/z 601; B is that from the ion at m/z 627; and C is that from the ion at m/z 573. arachidonate release induced by Ca 2ϩ ionophores or other Ca 2ϩ store-depleting agents in vascular myocytes (19), islet ␤-cells (20), and in embryonic kidney cells transfected with iPLA 2 (63), but arachidonate release in these cases is inhibited by low concentrations of BEL and does not require Ca 2ϩ entry (19,20). The resistance of A23187-induced arachidonate release from differentiated U937 cells to inhibition by low concentrations of BEL and its requirement for Ca 2ϩ entry indicate that iPLA 2 does not catalyze the release in these cells.
Although these considerations suggest that cPLA 2 is responsible for A23187-induced arachidonate release from differentiated U937 cells, it has not been clear why cPLA 2 is more readily activated in differentiated than in control cells. U937 cells express cPLA 2 (10), but cPLA 2 activity or protein does not increase upon differentiation nor does cPLA 2 phosphorylation state (2)(3)(4). Both control and differentiated U937 cells also undergo a similar rise in cytosolic [Ca 2ϩ ] when stimulated with ionophores (4). These findings suggest that differentiation confers some new cPLA 2 activation mechanism in U937 cells that does not operate in undifferentiated cells.
Our findings suggest that cPLA 2 activation occurs more readily in differentiated U937 cells because of changes in the physical state of its membrane substrates resulting in part from diglyceride generation by a PLD-PAPH pathway that operates only in differentiated and not in undifferentiated cells. Evidence for involvement of PLD in A2187-induced arachidonate release from differentiated U937 cells includes the facts that adding ethanol to the incubation results in formation of the PLD product phosphatidylethanol and in suppression of arachidonate release. PLD activity is regulated by Ca 2ϩ entry and by G-proteins in some systems (64), consistent with the demonstrated requirement for Ca 2ϩ entry in arachidonate release from differentiated U937 cells and the reported involvement of G-proteins in this phenomenon (4). PtEt accumulation is induced only in differentiated and not in undifferentiated U937 cells upon incubation with A23187 and ethanol, suggesting that expression of or an activation mechanism for PLD is acquired upon differentiation. Both high concentrations of BEL and propranolol inhibit PAPH (31, 42, 43) and suppress A23187-induced arachidonate release from differentiated U937 cells, and accumulation of the PLD-PAPH pathway product diglyceride occurs only in differentiated cells upon stimulation with A23187.
We believe that our ESI/MS analyses of PtEt from U937 cells incubated with A23187 and ethanol represent the first direct examination of PtEt molecular species produced in cells. Previous analyses of diglyceride structures of PA and phosphatidyl alcohols produced by PLD have been complicated by difficulties in removal or derivatization of head groups (46), but intact PtEt molecules can be analyzed by ESI/MS. Our analyses of U937 cell PtEt support findings based on identification of fatty acids from saponified phosphatidyl alcohols that PLD products contain saturated or monounsaturated substituents but very low levels of polyunsaturated substituents (46). HPLC analyses of derivatized diglycerides generated from PAPH-catalyzed hydrolysis of PA produced by PLD also indicate that such diglycerides contain predominantly saturated or monounsaturated fatty acids (46), and our ESI/MS/MS analyses of intact, underivatized diglycerides support this conclusion. Such diglycerides appear not to activate some PKC isoforms in intact cells that are activated by polyunsaturated diglycerides (46). Although PKC inhibition modestly suppresses A23187-induced arachidonate release from U937 cells, which might reflect PKC-dependent stimulation of either cPLA 2 (5,8) or of PLD (64), the majority of arachidonate release is resistant to PKC inhibition. The low arachidonate content of phospholipid solvolysis products generated by PLD suggests that hydrolysis of diglycerides generated by the PLD-PAPH pathway contributes little to agonist-induced arachidonate release. These observations suggest that diglycerides exert most of their effect to stimulate U937 cell arachidonate release by mechanisms other than PKC or diglyceride lipase, such as altering the state of cPLA 2 substrates in membranes.
Another change in U937 cell lipid composition that might increase the rate of arachidonate hydrolysis from phospholipids by cPLA 2 is the rise in the proportion of GPC lipids comprised by arachidonate-containing species that occurs upon differentiation, as rates of arachidonate release from GPC lipids increase with the mole fraction of such lipids in vesicle mixtures (27,28). We examined the possibility that iPLA 2 participates in the differentiation-associated rise in arachidonate-containing GPC species in U937 cells because iPLA 2 is proposed to participate in phospholipid remodeling by generating LPC acceptors for incorporation of arachidonate into GPC lipids (25,26). This proposal is based on findings that suppression of iPLA 2 activity with antisense oligonucleotides (26) (31). These findings suggest that iPLA 2 does not play a general role in synthesis of arachidonate-containing GPC lipids.
Although induction of iPLA 2 expression does not account for the higher proportion of arachidonate-containing species in GPC lipids of differentiated compared with undifferentiated U937 cells, other processes could be involved. Undifferentiated U937 cells proliferate more rapidly than differentiated cells (1), and rapidly proliferating tumor cells contain substantially less arachidonate in their phospholipids than their differentiated counterparts (65). This has been attributed to preferential ␤-oxidation of arachidonic acid by tumor cells (65), but differentiation-associated changes in a number of other processes (66) could also be involved.
The fact that a large fraction of Ca 2ϩ -independent PLA 2 activity in U937 cells is resistant to inhibition by BEL suggests that these cells might express another PLA 2 in addition to iPLA 2 that does not require Ca 2ϩ for catalysis because iPLA 2 cloned from all reported sources is exquisitely sensitive to BEL (11)(12)(13)(14)(15). Differentiated HL-60 cells also express a Ca 2ϩ -independent PLA 2 activity that is relatively resistant to BEL (24), and a BEL-resistant, Ca 2ϩ -independent PLA 2 activity has been isolated from brain (61). A Ca 2ϩ -independent PLA 2 other than iPLA 2 might account for arachidonate release that occurs in U937 cells undergoing Fas-induced apoptosis, as this process is only modestly suppressed by 50 M BEL (23). Expression of BEL-resistant, Ca 2ϩ -independent PLA 2 activities is not observed in all cells because Ca 2ϩ -independent PLA 2 activity in pancreatic islets (36) and insulinoma cells (31) is completely inhibited by BEL at concentrations below 10 M. BEL-resistant, Ca 2ϩ -independent PLA 2 enzymes might play some specialized but as yet unidentified role in cells that do express them, as induction of expression of such activity occurs upon differentiation of HL-60 cells (22).