Incorporation and Distribution of Epoxyeicosatrienoic Acids into Cellular Phospholipids*

The different regioisomers of epoxyeicosatrienoic acids derived from cytochrome P-450 monooxygenase are readily esterified into phospholipids of mastocytoma cells. Incorporation of 14,15-epoxyeicosatrienoic acid was concentration-dependent, with K , = 1.1 p~ and V,,, = 36 pmol/min/l0‘cells. Half-maximal incorporation occurred in 30 min, reaching a steady-state concentration of 470 pmol/lOs cells. This was slightly lower than the values for arachidonic acid (665 pmol/ 10’ cells) or 5-hydroxyeicosatetraenoic acid (554 pmol/lO‘ cells). The distribution of 14,15-epoxyeicos-atrienoic acid was preferential in the order phosphatidylethanolamine > phosphatidylcholine > phosphatidylinositol > phosphatidyl serine >> neutral lipids plus fatty acids. This contrasted with 5(S)-hydroxyei-cosatetraenoic acid, which was distributed primarily into phosphatidylcholine. Fast atom bombardment/tan-dem mass spectrometry facilitated identification of molecular species containing epoxyeicosatrienoic acids without relying on radioisotopes. Phosphatidylethanolamine plasmalogens with 16:l of mem-brane l-['4C]palmitoyl-2-lyso-PC dissociation of TSQ 70 quadrupole in diethanolamine in triethanolamine matrix bombardment collision gas argon at millitorr in the quadrupole region of tandem spectrometer. collision offset energy was 30 eV; multiplier dynode

The different regioisomers of epoxyeicosatrienoic acids derived from cytochrome P-450 monooxygenase are readily esterified into phospholipids of mastocytoma cells. Incorporation of 14,15-epoxyeicosatrienoic acid was concentration-dependent, with K , = 1.1 p~ and V,,, = 36 pmol/min/l0'cells. Half-maximal incorporation occurred in 30 min, reaching a steady-state concentration of 470 pmol/lOs cells. This was slightly lower than the values for arachidonic acid (665 pmol/ 10' cells) or 5-hydroxyeicosatetraenoic acid (554 pmol/lO' cells). The distribution of 14,15-epoxyeicosatrienoic acid was preferential in the order phosphatidylethanolamine > phosphatidylcholine > phosphatidylinositol > phosphatidyl serine >> neutral lipids plus fatty acids. This contrasted with 5(S)-hydroxyeicosatetraenoic acid, which was distributed primarily into phosphatidylcholine. Fast atom bombardment/tandem mass spectrometry facilitated identification of molecular species containing epoxyeicosatrienoic acids without relying on radioisotopes. Phosphatidylethanolamine plasmalogens with 16:l or 182 at the sn-1 position, or an 18:0 acyl group, and phosphatidylcholine with 16:O alkyl ether or an acyl group at the sn-1 position incorporated all possible epoxyeicosatrienoic acid regioisomers. Under basal conditions, cells eliminated 14,15-cis-epoxyeicosatrienoic acid slowly with a half-life of 34.9 f 7 h. Cells stimulated with calcium ionophore A23187 eliminated 14,15-epoxyeicosatrienoic acid rapidly. It was notable that its rate of release from phosphatidylcholine and phosphatidylinositol exceeded that for arachidonic acid. A coenzyme A-independent transacylase also catalyzed the transfer of epoxyeicosatrienoic acids from mastocytoma cell membranes into 1-palmitoyl-2-lysophosphatidylcholine. The cellular incorporation, release, and distribution of epoxyeicosatrienoic acids is distinctive and contrasts with most other eicosanoids, suggesting that these compounds may have both autocoid and nonautocoid functions.
* This work was supported by National Institutes of Health Grant GM 41026. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisernent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Incorporation of EETs, DiHETEs, 5-HETE, and Arachidonic Acid into Mast Cells-Virus-transformed murine bone marrow-derived mast cells were maintained in tissue culture using RPMI media supplemented with 10% (v/v) fetal calf serum. These cells were selected as a convenient model system. Cells (1.6-2.6 X 10B/ml) were isolated by centrifugation at 50 X g for 15 min at 25 "C, washed with Hanks' balanced salt solution (0.5 volume) and finally resuspended in Hanks' balanced salt solution at a concentration of 0.7 X lo6 cells/ ml. 14,15-EET (5-50 nmol), dissolved in 500 pl of buffer containing 5 mg of BSA/ml, was added to mast cells (50 mls). The final concentration of EET ranged from 0.1 to 1 p~; the concentration of BSA was 0.05% (w/v) unless specified differently. Samples were incubated at 37 "C, and portions were removed at intervals from 5 min to 2 h, centrifuged at 100 X g for 15 min at 25 "C, and washed once with Hanks' balanced salt solution containing 0.25 mg of BSA/ml to stop the equilibration between EET and mast cells. Phospholipids were extracted from cells according to Bligh and Dyer (29). Aliquots of the supernatants were evaporated to dryness, dissolved in 0.5 ml of MeOH/HzO/acetic acid (50/50/0.05, v/v), pH 5.7, and purified on RP-HPLC. For molecular species analysis, 5,6-, 8,9-, 11,2-, and 14,15-EET (310 nmol) were incubated with 310 ml of cells for 2 h. Incorporation of arachidonic acid, 5-HETE, and 14,15-diHETE was also examined by a similar procedure. Steady-state levels were estimated by fitting data to the equation for a rectangular hyperbola using Graph-Pad.
High Performance Liquid Chromatography-PC, PE, PS, and PI were isolated by normal phase HPLC on Lichrosorb" eluted for 6 min at 1 ml/min with a mobile phase consisting of 53% solvent A, 47% solvent B and then programmed to 100% solvent B over 20 min (30). Solvent A was hexane/isopropanol (3/4, v/v); solvent B was hexane/isopropanol/water (3/4/0.7, v/v/v). Phospholipids eluted in the order PE < PI < PS < PC, with retention times of 19-22,29-32, 35-36, and 40-44 min, respectively (30). Individual molecular species of phospholipids were isolated by reversed phase HPLC on Ultrasphere' ODS eluted for 100 min at 2 ml/min with methanol/acetonitrile/water (90.5/2.5/7, v/v/v) containing 1 mM trifluoroacetic acid ammonium salt, pH 7.4 (31). A third mobile phase was used for chromatographic analysis of EETs or other eicosanoids released into the supernatant after stimulation of mast cells. This was a gradient from 45% solvent A (HzO, 0.05% acetic acid, pH 5.7) and 55% solvent B (methanol) to 70% solvent B during 20 min, to 80% solvent B during 15 min, and to 100% solvent B during 5 min. The complete analysis time was 60 min. Products were detected by scintillation counting of radioactive products.
Release of EETs Following Cellular Incorporation-Cells (5 ml, 0.7 X lo6 cells/ml) containing 3H-labeled EETs and/or [''Cc] arachidonic acid were stimulated with 2 p~ A23187. The reaction was stopped at intervals from 0-60 min by centrifuging at 1000 X g for 10 min, followed by washing with 5 ml of 0.25 mg/ml BSA. The radioactivity remaining in the pellet was quantified by scintillation counting. Phospholipids were extracted (29) and separated by HPLC to determine the amount of radioactivity remaining in PE, PC, and PI as a function of time after stimulation.
The reincorporation of EETs into lysophospholipids via transacylase was examined according to Uemura et al. (32). Mast cells were incubated for 18 h at 37 "C with 12.5 p~ [3H]-14,15-EET (4 pCi) in media containing 0.4% fetal calf serum. Cells (2.4 X 1 0 ' ) were washed and homogenized in 2.4 ml of 0.25 M sucrose, 10 mM HEPES, pH 7.0, 0.5 mM EGTA, 1 mM dithiothreitol, 1 pg/ml leupeptin. Homogenates were centrifuged at 500 X g for 10 min, and the supernatant fluid was then centrifuged at 100,000 X g for 60 min. The microsomal pellet was isolated and suspended in 1.7 ml of 0. Mass Spectrometry-Collision-induced dissociation mass spectra were obtained by fast atom bombardment of phospholipids (20-100 ng), using a TSQ 70 triple quadrupole mass spectrometer (Finnigan MAT, San Jose, CA). Negative ions from PE and PI were produced in a diethanolamine matrix and those from PC in a triethanolamine matrix (30). The fast atom bombardment gun (Ion Tech, Teddington, UK) was operated at 1 mA with xenon accelerated to 6 kV. The collision gas was argon at 0.5 millitorr in the second quadrupole region of the tandem mass spectrometer. The collision offset energy was 30 eV; the electron multiplier conversion dynode was operated at 12 kV.  (Fig. 1, lower panel). Hydrolytic instability of 14,15-EET might partly account for this difference, since 14,15-diHETE, the hydration product of 14,15-EET, was not readily incorporated, reaching a steady-state content of 151 pmol/106 cells. Increasing the albumin concentration in the buffer protected the EET from hydrolysis; however, there was no corresponding increase in cellular EET, because sequestration to albumin also limited the availability of free 14,15-EET for incorporation (Table I). After a 2-h incubation [3H]14,15-EET was distributed among phospholipids in the order PE > PC > PI (Fig. 1, top panel) Table 11). The peak with a retention time of 55 min is an unidentified material co-extracted with phosphatidylethanolamine. p~ (Fig. 2). When mast cells were washed and resuspended in Hanks buffer without BSA, 14,15-EET disappeared slowly from the phospholipids by a first-order process with a halflife of 34.9 f 7 h, determined from the nonlinear regression analysis of the loss of [3H]14,15-EET versus time.

Identification of Separate EET Isomers within Phospholipid
Classes and Molecular Species-Mast cells were incubated with 5,6-, 8,9-, ll,lZ-, or 14,15-EET for 2 h, and their phospholipids were analyzed by FAB/triple quadrupole mass spectrometry to generate (M-H)-ions for phosphatidylethanolamine and phosphatidylinositol or (M-CH3)-ions for phosphatidylcholine species (30, 33, 34). Fig. 3 typifies the chromatographic separation of phosphatidylethanolamine containing 8,9-EET in at least eight peaks corresponding to six molecular species. Collision-induced dissociation of the phospholipids yielded characteristic daughter fragment ions for each of the four EET regioisomers; for a specified EET regioisomer, these ions were indicative of its occurrence independent of the polar head group on the phospholipid. EET regioisomers were widely distributed among different molecular species of phospholipids. The identities of these species and semiquantitative estimates of the concentrations, determined by MS/MS, are depicted in Table 11. EET regioisomers were distributed among similar molecular species, with only minor quantitative differences. Molecular species with sn-1 substituents of 16:l plasmalogen, 182 plasmalogen, and 18:O acyl were richest in EETs at the sn-2 position. 5-HETE was distributed among the same molecular species containing EETs with some quantitative differences.
Comparative Release of 14,15-EET and Arachidonic Acid-Steady-state incorporation of a mixture of 14,15-EET (0.1 p~) and arachidonic acid (0.1 p~) was similar to the steady state achieved during incubations with each agent, individually. The processes involved in phospholipid remodeling are not saturated at these concentrations, which is consistent with data from Fig. 2. When mast cells containing both ["CC] arachidonic acid and [3H] 14,15-EET were subsequently stimulated with 2 p~ Ca2+ ionophore to activate lipases, liberation of 14,15-EET occurred more rapidly than arachidonic acid (Fig. 4). The pseudo first-order decline in cellular 14,15-EET had a rate constant k = 0.32 min-'; the decline in arachidonic acid had a rate constant k = 0.14 min". After 60 min, 14,15-EET declined from 100 to 38.  (Table 111).
To determine if a coenzyme A-independent transacylase could catalyze transfer of EETs among phospholipid classes, mast cell membranes containing [3H]14,15-EET were incubated with 1-['4C]-palmitoyl-2-lyso-PC. Chromatographic analysis indicated that the PC fraction contained both 3H and 14C species; PE and PI contained only 3H. When the molecular species of PC were separated by RP-HPLC, only a single peak

DISCUSSION
Cellular phospholipids containing oxidized forms of arachidonic acid, such as HETEs or EETs, could originate from two mechanisms: (i) direct oxidation of the fatty acid substituents at the sn-1 or sn-2 positions of glycerophospholipids and diacylglycerol and (ii) "remodeling" (re-esterification) of lysoglycerophospholipids with HETEs or EETs. Our data indicate that the latter process occurs readily in mastocytoma cells, which is consistent with recent results by Karara et al. (15) using rat liver. The distribution of EETs varies among the different systems examined. For instance, EETs were distributed almost uniformly between PC, PE, and PI in rat livers (15), but were enriched in the PE fraction of mastocytoma cells. Human platelets may release 14,15-EET from their PI fraction under some circumstances, but the identity of the 14,15-EET needs to be established more rigorously (17). These examples resemble the situation with HETEs whose distribution among phospholipids with different polar head groups also varies substantially. For instance, 15(S)-HETE accumulates in phosphatidylinositol in neutrophils (21) and bovine coronary artery cells (22) and into PC and PI in macrophages (35); 12(S)-HETE accumulates in phosphatidylcholine (20,23). At present, we cannot exclude the possibility that the stereo-or regioisomeric form of the EETs or HETEs could influence their uptake and distribution into different phospholipid classes; however, the cell type appears to be the most prominent variable governing these processes. Mastocytoma cells are a convenient model system in this regard. They are useful for comparison of their incorporation of oxidized fatty acids with that reported for other cells that undergo stimulus-secretion coupling (17,21,35).
It should be stressed that we identified glycerophospholipids containing oxidized fatty acids by FAB MS/MS without relying on radioactive tracers. Characteristic ions from EETs can be obtained directly from individual molecular species isolated by RP-HPLC (Fig. 3, Table 11) analogous to procedures employed to determine molecular species distributions of arachidonic acid (36). The direct analysis of glycerophospholipids also avoids problems associated with chemical or enzymatic deacylation procedures to determine the identity of unstable substituents at the sn-1 or sn-2 positions, and it allows unambiguous distinction between high affinity cell association and authentic covalent incorporation into phospholipids. An extension of this procedure may facilitate efforts to determine if glycerophospholipids containing oxidized fatty acids occur naturally and if particular conditions promote their formation in uiuo. Direct oxidation of polyunsaturated lipids within glycerophosphatide substrates has been demonstrated under certain conditions, particularly cell-free systems using exogenously supplied 15-lipoxygenase (37-39) or chemical oxidants (40)(41)(42). Oxidations of phospholipid substrates by other regiospecific lipoxygenases or by mixed function oxidase have been difficult to demonstrate, especially using physiologically relevant conditions. Our in vitro data, and data by others (14,16,21,22), argue for remodeling as the origin of oxidized substituents within glycerophospholipids. However, we speculate that molecular species distribution patterns from remodeling of oxidized lipids will differ from those produced by direct oxidation of polyunsaturated fatty acids via enzymatic or nonenzymatic processes. Distinction between these two mechanisms of formation could be important in view of the emerging roles for oxidatively modified phospholipids. For example, lipoxygenase-dependent oxidation products within membranes of rabbit reticulocytes may contribute to maturational breakdown of mitochondria (27); oxidatively modified phosphatidylcholines can activate neu-trophils via the receptor for platelet-activating factor (26) or compete as substrates for degradation of platelet-activating factor (43); and oxidized lipoproteins aggravate atherogenesis (25).
Others have observed a rapid, preferential release of oxidized polyunsaturated fatty acids from phospholipids when liposomes or cell membranes are incubated with phospholipase Az in vitro (40)(41)(42). Our results substantiate that intact cells release 14,15-EET at rates exceeding those observed for arachidonic acid. It has been speculated that such rate differences could be beneficial by preventing accumulation of oxidized lipids in cell membranes (40)(41)(42). However, there are two reasons why this seems debatable for EETs. First, the rapid depletion of EETs was only evident after activation of phospholipase Az; under basal conditions, depletion of 14,15-EET (approximately 5 half-lives) would require 175 h. Second, the differences in release kinetics were restricted to PC and P I (Table 111); however, PE is the predominant pool containing EET. These data seem incompatible with differential release rates serving a protective function. It is notable that we detected phospholipids containing only EETs, but not diHETEs, by FAB/MS/MS, suggesting that incorporation into cellular phospholipids protects the EETs from hydration. The lack of diHETEs in cellular phospholipids is consistent with either a low rate of uptake or a transient uptake followed by a rapid release; we are unable to distinguish between these two possibilities at present, since structural alterations in membranes due to lipid peroxidation have been associated with enhanced sensitivity to phospholipase Az (41).
The rapid release of 14,15-EET from phospholipids is also interesting in the context of platelet-activating factor biosynthesis via the remodeling pathway. Contemporary mechanisms stress the importance of a transacylase/phospholipasecoupled reaction requiring an initial generation of a lysophospholipid precursor to "accept" arachidonic acid freed from phosphatidylcholine; the lysophosphatidylcholine can then be acetylated to form platelet-activating factor (32, 44,45). The kinetics of EET release and its distribution among certain molecular species are favorably disposed for this process, and we have demonstrated that lysophosphatidylcholine can accept EETs from mast cell membranes, which is consistent with this proposal. However, it remains to be proved that EETs or HETEs actually participate in the remodeling pathway for platelet-activating factor formation (46).