Activated Platelets and Monocytes Generate Four Hydroxyphosphatidylethanolamines via Lipoxygenase*

12/15-Lipoxygenase (LOX) mediates immune-regulatory activities not accounted for by its known free acid eicosanoids, suggesting that additional lipids may be generated by activated cells. To characterize novel LOX-derived lipids, a lipidomic approach was utilized. Ionophore-activated interleukin-4-treated human peripheral monocytes generated up to 10-fold more esterified 15-hydroxyeicosatetraenoic acid (15-HETE) than free in a phosphatidylinositol 3-kinase- and protein kinase C-sensitive manner. Precursor scanning electrospray ionization/tandem spectroscopy for m/z 319 (HETE, [M-H]–) showed 4 ions at m/z 738, 764, 766, and 782 that were identified using tandem spectroscopy and MS3 as specific diacyl and plasmalogen 15-HETE phosphatidylethanolamines. Using H 182O water, the compounds were shown to form by direct oxidation of endogenous phosphatidylethanolamine (PE) by 15-LOX, with PE being the preferred phospholipid pool containing 15-HETE. Similarly, human platelets generated 4 analogous PE lipids that contained 12-HETE and increased significantly in response to ionophore, collagen, or convulxin. These products were retained in the cells, in contrast to free acids, which are primarily secreted. Precursor scanning of platelet extracts for the major platelet-derived prostanoid, thromboxane B2 (m/z 369.2), did not reveal PE esters, indicating that this modification is restricted to the LOX pathway. In summary, we show formation of PE-esterified HETEs in immune cells that may contribute to LOX signaling in inflammation.

12/15-LOX is expressed by monocytes/macrophages in response to interleukin-4 and -13 and oxidizes arachidonate at either C12 (mouse, rat, pig) or C15 (human, rabbit) (5,6). In contrast, platelet 12-LOX forms 12-HpETE. After their generation, 12/15-LOX-derived hydroperoxides are either reduced by glutathione peroxidases forming HETEs or further metabolized by LOX to epoxy or tri-hydroxy products, termed hepoxilins or lipoxins (7)(8)(9)(10)(11). Up to now these compounds have been assumed to be the sole arachidonate-derived products, although the detailed metabolic pathways for H(p)ETEs in immune cells are largely uncharacterized. One modification of endogenously generated H(p)ETEs is their esterification into membrane phospholipids or other complex lipid pools. Several years ago, before the advent of high sensitivity mass spectrometry, this question was addressed by the addition of radiolabeled HETEs to cells followed by examination of their fate by radio-TLC (12)(13)(14)(15)(16)(17). However, this is indirect, since exogenous H(p)ETE differs considerably in fate from endogenously generated product (18). Also, this approach is insensitive, and furthermore, does not allow identification of specific molecular species.
Expression of 12/15-LOX in macrophages without release of detectable free H(p)ETEs or lipoxin modulates immune cell function (19 -23). This suggests that H(p)ETE metabolites formed within or close to the membrane during LOX turnover may contribute; however, the identities of such lipids are currently unknown. Given the recent studies suggesting an important anti-inflammatory and pro-resolving function for 12/15-LOX, identification of novel products of this pathway is a clinically relevant goal (24,25).

Materials-15(S)-Hydroxy-[S-(E,Z,Z,Z)]-
Isolation and Activation of Human Monocytes-50 ml of buffy-coat blood was diluted 1:1 (v/v) with PBS/citrate/dextran (0.8% w/v citrate, 2% w/v dextran 400, pH 7.4). Red cells were allowed to sediment for 1-2 h. The straw-colored supernatant was collected and underlayed with Lymphoprep 2:1 (v/v, supernatant:Lymphoprep) and centrifuged 800 ϫ g for 20 min at 4°C. The interface was collected and diluted 1:1 (v/v) with ice-cold PBS containing 0.4% (w/v) citrate, pH 7.4, and spun at 400 ϫ g, 10 min at 4°C. The supernatant was discarded, and the cell pellet was washed 5ϫ with ice-cold PBS/citrate buffer, 400 ϫ g for 5 min at 4°C. The cell pellet was finally resuspended in a small volume of RPMI 1640 (10% fetal calf serum, 100 units/ml penicillin, 100 g/ml streptomycin, 2 mM glutamine). Approximately 10 8 cells were seeded per T75 flask and incubated at 37°C for 2 h to allow monocytes to adhere. The medium was replaced, and cells were cultured for 72 h with 700 pM IL-4 to induce 15-LOX1. Monocytes were harvested by centrifugation and resuspended in Krebs buffer (50 mM HEPES, 100 mM NaCl, 5 mM KCl, 1 mM NaH 2 PO 4 ⅐2H 2 O, 1 mM CaCl 2 , 2 mM glucose). 4 ϫ 10 6 cells in 1 ml were stimulated with A23187 (10 M) with/without 0.8 M phorbol 12-myristate 13-acetate at 37°C for 10 min. In some experiments H 2 18 O was used in place of water in Krebs buffer. Experiments using signaling inhibitors included a 10-min preincubation step at 37°C before the addition of stimulus. In some experiments 15-HETE-d8 (330 ng) was included during activation of cells. Murine-resident peritoneal cells were isolated from wild-type or 12/15-LOX Ϫ/Ϫ mice by peritoneal lavage into 2 ml of PBS and used without further purification (26).
Isolation and Activation of Washed Human Platelets-Whole blood was collected from healthy volunteers free from nonsteroidal anti-inflammatory drugs for at least 14 days into acid-citrate-dextrose (ACD; 85 mM trisodium citrate, 65 mM citric acid, 100 mM glucose) (blood:ACD, 8.1:1.9 v/v) and centrifuged at 250 ϫ g for 10 min at room temperature. The platelet-rich plasma was recentrifuged at 900 ϫ g for 10 min, and the pellet was resuspended in Tyrode buffer (134 mM NaCl, 12 mM NaHCO 3 , 2.9 mM KCl, 0.34 mM Na 2 HPO 4 , 1 mM MgCl 2 , 10 mM Hepes, 5 mM glucose, pH 7.4) containing acid-citrate-dextrose (1:9 v/v). The platelets were washed by centrifuging at 800 ϫ g for 10 min then resuspended in Tyrode buffer. Platelets were activated at 37°C in the presence of 1 mM CaCl 2 for 15 min, FIGURE 1. 15-HETE generated by IL-4-treated monocytes is predominantly esterified. Panel A, 15-H(p)ETE generation in response to ionophore activation. IL-4-treated human monocytes were activated with A23187 before analysis for 15-HETE using HPLC-UV as described under "Experimental Procedures." Panel B, generation of esterified 15-HETE parallels free acid but continues for longer. Shown is the time course of free and esterified 15-H(p)ETE generation by IL-4-treated monocytes activated with 10 M A23187 for 2-15 min. Panel C, peritoneal macrophages from 12/15-LOX Ϫ/Ϫ mice do not generate free or esterified 15-HETE. Free and esterified 12-HETE generated by murine peritoneal macrophages from wild-type and 12/15 LOX Ϫ/Ϫ was determined using LC-UV before and after activation with 10 M A23187 (n ϭ 3; mean Ϯ S.E.; *, p Ͻ 0.05 versus unactivated control, Student's t test). Panel D, involvement of phosphatidylinositol 3-kinase in generation of free and esterified 15-HETE. Generation of 15-HETE was determined after activation of cells with 10 M A23187 with/without 100 M wortmannin. Panel E, protein kinase C stimulates generation of free and esterified 15-HETE in monocytes. IL-4-induced human monocytes were activated with 10 M A23187 with/without 0.8 M phorbol 12-myristate 13-acetate (PMA) and/or 10 M bisindolylmaleimide (BIS) (n ϭ 3; mean Ϯ S.E.; *, p Ͻ 0.05 versus A23187 alone, Student's t test). with 10 M A23187, 10 g/ml collagen, or 60 ng/ml convulxin before lipid extraction as below.
Lipid Extraction-15(S)-Hydroxyeicosadienoic acid (10 g for HPLC-UV or 10 ng for LC/MS/MS) and/or 10 ng of di-14: 0-phosphatidylethanolamine was added to each sample before extraction as internal standards. Hydroperoxides were then reduced to their corresponding stable alcohols by adding 1 mM SnCl 2 (27). Lipids were extracted by adding a solvent mixture (1 M acetic acid, 2-propanol, hexane (2:20:30, v/v/v) to the sample at a ratio of 2.5 ml of solvent mixture/1 ml of sample, vortexing, and then adding 2.5 ml of hexane (27). After vortex and centrifugation, lipids were recovered in the upper hexane layer. The samples were then re-extracted by the addition of an equal volume of hexane followed by vortex and centrifugation. The combined hexane layers were dried under N 2 flow and analyzed for free 15-HETE using LC-UV or LC/MS/MS or esterified compounds using LC/MS/MS (as described below). To calculate esterified HETEs, free HETE was subtracted from total measured after saponification. For this, N 2 -dried lipids were resuspended in 1.5 ml of 2-propanol, and then fatty acids were released by base hydrolysis with 1.5 ml of 0. Precursor Scanning Mass Spectrometry-Electrospray mass spectra were obtained on a Q-Trap instrument (Applied Biosystems 4000 Q-Trap) operating in the negative mode. Lipid extracts were diluted (1:50 -1:100) and introduced at 10 l/min in methanol using a Hamilton syringe. Instrument settings were determined by tuning on an oxidized phosphatidylethanolamine (PE) standard and run with declustering potential Ϫ140 V and collision energy Ϫ45 V. Spectra were obtained from 550 -1000 atomic mass units over 12 s, with typically 10 scans acquired and averaged. For determination of -fold changes in specific 15-HETE-containing lipids, 10 l of diluted extracts were injected under flow (1 ml/min) in a methanol:water (50:50) mixture with specific multiple reaction monitoring (MRM) transitions monitored using m/z 319.2 as daughter ion and comparing ion intensity to the internal standard, di-14:0-PE (MRM m/z 634.5 3 227.2). For 15-HETE-d8, m/z 327.2 was used instead.
Normal Phase HPLC-UV of Phospholipid Classes-To fractionate monocyte phospholipid classes before MS analysis, extracts were separated on a Spherisorb S5W 4.6 ϫ 150-mm column (Waters Ltd) using a gradient of 50 -100% B over 25 min (A, hexane:2-propanol, 3:2; B, solvent A:H 2 O, 94.5:5.5) at a flow rate of 1.5 ml min Ϫ1 (28). Absorbance was monitored at 205 nm, and products were identified using a mixture of standard phospholipids (Sigma). 1-min fractions were collected for subsequent analysis by ESI/MS/MS. After drying down and  JULY 13, 2007 • VOLUME 282 • NUMBER 28

JOURNAL OF BIOLOGICAL CHEMISTRY 20153
resuspending in 100 l of methanol, 10-l samples of each fraction were injected under flow (1 ml/min) in a methanol:water (50:50) mixture into the electrospray source, with specific MRM transitions monitored using m/z 319.2 as the daughter ion, and areas for each transition were determined in each fraction by integration of the peaks. These were then replotted versus time to obtain normal phase elution profiles for each HETEcontaining ion.
Reverse Phase LC/MS/MS of Phospholipids-Online reverse phase separation of phospholipids to separate based on acyl chain was carried out using a Luna 3-m C18 (2)  15-HETE Quantitation in Phospholipid Classes-Lipid extracts from A23187-activated monocytes were separated by normal phase LC-UV, and 1-min fractions were collected from 0 to 20 min. Fractions were dried under N 2 and resuspended in methanol. Fractions containing each phospholipid class were identified by head-group precursor scanning (PE ϭ 7-9 min, phosphatidylcholine (PC) ϭ 18 -20 min, phosphatidylinositol ϭ 11-13 min, phosphatidylglycerol 6 -7 min, phosphatidylserine 14 -16 min) and hydrolyzed as already described, then analyzed by reverse phase LC/MS/MS monitoring for 15-HETE. Amounts of PE and PC were determined using normal phase LC-UV (205 nm).

Generation of 15-HETE-containing Phospholipids by Soybean LOX Oxidation of Commercial Phospholipid
Preparations-5 mg/ml L-␣-phosphatidylethanolamine (egg (Sigma) or brain, porcine plasmalogen (Avanti Polar Lipids, Alabaster, AL)) was incubated for 30 min at 37°C in PBS, pH 7.4, 4% sodium cholate, with 50 kilounits of soybean lipoxygenase type IV (Sigma). Samples were then reduced using 1 mM SnCl 2 for 10 min, 20°C, spiked with 10 ng of internal standard (di-14:0-PE), and extracted as before. Statistical Analysis-Data are representative of at least three separate donors, with samples run in triplicate for each experiment (mean Ϯ S.E.). Significance was examined using an unpaired t test, where p Ͻ 0.05 was considered significant (denoted by an asterisk on the figures).

Monocytes Generate Predominantly Esterified 15-H(p)ETE
after Ionophore Activation-To define pathways that regulate calcium-dependent activation of 15-LOX1, generation of 15-H(p)ETEs after stimulation of IL-4-treated human monocytes with A23187 was examined. Samples were reduced using SnCl 2 before lipid extraction to convert all HpETE into the more stable HETE (27). Most 15-H(p)ETE generated on activation was esterified to complex lipids of unknown structure (Fig.  1A). Also, basal levels of esterified 15-H(p)ETE were detected without ionophore activation. Esterified H(p)ETE generation occurred early and continued after free H(p)ETE had plateaued (Fig. 1B). Murine peritoneal macrophages that express 12/15-LOX, the functional equivalent of human 15-LOX1, generated Phosphatidylethanolamine Products of Lipoxygenase esterified 12-H(p)ETE, and this was absent in macrophages from 12/15-LOX Ϫ/Ϫ mice (Fig. 1C). These data show that the predominant H(p)ETEs generated by monocytes and macrophages on calcium activation are not free acid products, but complex lipids.

Regulation of 15-H(p)ETE Release by Intracellular
Signaling Pathways-Calcium facilitates membrane association of 12/15-LOX (29). However, it is unknown whether additional signaling pathways participate in monocyte 12/15-LOX regulation by calcium and whether differences exist between regulation of free versus esterified product formation. Inhibition of both H(p)ETE forms was observed on the addition of wortmannin, implicating phosphatidylinositol 3-kinase (Fig. 1D). The protein kinase C inhibitor bisindolylmaleimide did not significantly affect ionophore stimulation of 12/15-LOX; however, activation of protein kinase C using phorbol 12-myristate 13-acetate significantly promoted generation of both forms in a bisindolylmaleimide-sensitive manner (Fig. 1E). This indicates that stimulation of protein kinase C potentiates calcium-dependent activation of 12/15-LOX.  Fig. 2A). Also, precursor ESI/MS/MS of murine peritoneal macrophage extracts demonstrated identical ions (Fig. 2B). Next, MRM transitions of the parent-HETE daughter ion were monitored in activated and unactivated human monocytes and compared with an internal standard (di14:0-PE). All four ions increased significantly after calcium mobilization (Fig. 2C).

H(p)ETE Is Esterified to PE in Activated Human Monocytes and Mouse Peritoneal Macrophages-
These m/z values are consistent with nitrogen-containing HETE derivatives of two phospholipid species, PE and PC (e.g. m/z 766.5 could be 16:0p/15-HETE-PC or 18:0p/15-HETE-PE). However, the m/z values are not consistent with other phospholipids, including phosphatidylinositol, phosphatidylserine, phosphatidylglycerol, or phosphatidic acids (30). To determine phospholipid class, normal phase LC was undertaken that separates phospholipids according to head group. Fractions were collected, and aliquots of each 1-min fraction were monitored for the parent 3 m/z 319.2 transition using MS/MS. Greater than 95% of ion intensity of the transition for all 4 peaks co-eluted with di-14:0-PE standard (5-6-min fraction), with the remainder (Ͻ5%) eluting with PC in the 17-19-min fraction (Fig. 2D). These data suggest   . The ionization efficiency of equimolar amounts of PC in negative mode is approximately a third that of PE in our system (not shown), and so it is possible that HETE-PC was not readily detected in our experiments. Therefore, to conclusively determine the phospholipid site specificity of HETE formation, the phospholipid fractions from activated monocytes were purified using normal phase HPLC, then hydrolyzed to release free 15-HETE, which could then be directly quantified by LC/MS/MS. Using this approach, Ͼ92% of the 15-HETE was found in PE (Fig. 3). The preference of PE over PC as the site of HETE synthesis is observed whether it is expressed as a function of total phospholipid class or per g of phospholipids, with the level equating to ϳ1.5% of the PE pool in activated monocytes containing 15-HETE (based on detecting 0.7125 nmol of 15-HETE in 48 nmol of purified PE).
Structural Identification of Individual 15-HETE Ions-Because of isobaric peaks in the PE fraction, further LC was required to separate different 15-HETE-containing PEs. Reverse phase-LC/MS/MS, which separates based on acyl chain composition, was used on normal phase-purified PE fractions (see "Experimental Procedures"). The parent 3 m/z 219 transition (15-HETE ion) was monitored with an MS/MS spectrum triggered during elution of the MRM transition. The spectrum shown is obtained at the apex of the peak of elution for each compound, which was at the same retention time (4.6 min) for all 4 ions, likely due to the presence of identical sn-2 lipids. Using this approach, spectra were obtained that could be compared with PE oxidized in vitro using soybean 15-LOX.
The product ion spectrum of m/z 782.5 (Fig. 4A)    lower, and good quality spectra could not be obtained (Ref. 32; data not shown). Plasmalogen phospholipids are sensitive to acidic conditions and can be hydrolyzed by exposure to HCl fumes (32). However, not only did ions at m/z 738.5, 764.5, and 768.5 degrade relative to the diacyl internal standard (di14:0-PE) after acid exposure but also the m/z 782.5, which is not a plasmalogen (data not shown), degraded. This suggests that HETE-containing lipids, by virtue of their -OH group on the 20:4, are sensitive to acid hydrolysis. Brain PE, which contains ϳ50% plasmalogen and little ether-linked lipids was, therefore, oxidized using soybean 15-LOX. Spectra obtained during on-line reverse phase LC/MS/MS of normal phase-separated 15-LOXoxidized brain PE are shown (Figs.  4, F-H). These spectra are identical to those obtained from monocyte extracts, suggesting that the ions at m/z 738. 5 (30)). However, a variation in the double bond position cannot be conclusively ruled out. Finally, for all four monocyte 15-HETE-PEs, MS3 analysis confirmed that daughter ions m/z 219 and 175 conclusively originate from the 15-HETE fragment at m/z 319.2 (Fig. 5). No daughter ions that would be expected from other HETE isomers (e.g. 5-HETE, m/z 115; 8-HETE, m/z 155; 9-HETE, m/z 151; 11-HETE, m/z 167; 12-HETE, m/z 179) are found. To further confirm the isomer distribution, the PE fraction was hydrolyzed, and each HETE positional isomer was quantified using LC/MS/MS. As shown, only the 15-HETE isomer was detected (Fig.  6A). Furthermore, chiral phase LC-UV analysis of the purified 15-HETE from monocyte PE (collected from reverse phase LC of the hydrolyzed PE fraction) showed that Ͼ96% was the S enantiomer (Fig. 6B). These data collectively indicate that the compounds contain predominantly 15 (Fig. 7A and inset). In contrast, monocytes did not incorporate 18 O into PE-HETEs, with elevation in HETE-16 O  JULY 13, 2007 • VOLUME 282 • NUMBER 28 only after ionophore activation (Fig.  7, B-G). This indicates that formation of PE-HETEs likely does not involve esterification of HETE but direct oxygenation of arachidonatecontaining PE by 15-LOX.

Phosphatidylethanolamine Products of Lipoxygenase
To further validate the use of H 2 18 O as a tool to identify direct oxidation of PE by LOX, experiments were conducted using soybean 15-LOX and purified SAPE or SAPC in borate/deoxycholate buffer (32). Both were oxidized at similar rates (1.2 Ϯ 0.15 or 1.4 Ϯ 0.12 nmol/min/8 kilounits of 15-LOX for PE or PC, respectively, mean Ϯ S.E.). PC-HETE was more readily detected as a positive ion but was also observed in negative mode with CID spectra showing diagnostic 15-HETE daughter ions at m/z 319, 219 and 175 (Fig. 8, A and B). In these experiments where LOX directly oxidizes PE, inclusion of H 2 18 O had no effect on the isotope distribution of PE-HETEs formed (Fig. 8, C-F 8, C-F). We also confirm in these experiments that SnCl 2 , included to reduce the primary 15-LOX product, 15-HpETE, to the more stable 15-HETE, does not promote incorporation of 18 O (Fig. 8, C-F).
PE-HETEs Are Retained within the Cells-To determine whether PE-HETEs are primarily secreted similar to free acid eicosanoids or retained intracellularly, monocytes were pelleted after ionophore activation, and levels of free and PE-HETEs were determined separately in supernatant and cell pellets. As expected, free H(p)ETE was primarily secreted (70%, not shown); however, partitioning of PE-HETE was different, with the majority retained within the cells (Fig. 7H). This indicates different cell partitioning of esterified versus free H(p)ETE products and has implications for the potential signaling actions of this class of lipid.
Exogenous or Endogenous 15-HETE Is Not Incorporated into Phospholipids during Activation of 15-LOX-To determine whether 15-HETE becomes esterified during the time course of our studies, monocytes were co-incubated with 15-HETE-d8 at similar concentrations to what is synthesized by the cells with/ without ionophore activation. If this was incorporated, it would be observed by precursor scanning for m/z 327. Precursor scanning at m/z 319 or 219 showed that the presence of the deuterated analog did not prevent 15-LOX oxidation of PE, with the characteristic PE-HETE ions being detected (Fig. 9, A, B, and E). Scanning for m/z 327 showed prominent ions 8 atomic mass units higher than the PE-HETEs-d0, suggesting incorporation of the 15-HETE-d8 into PE; however, these were also present when the cells were not supplemented with 15-HETE-d8 ( Fig.  9, C with inset, and F) and, furthermore, did not show up in a precursor scan for the 15-HETE-d8 daughter ion at 226, which instead detected ions that were not seen in the precursor m/z 327 scan (Fig.  9D). It is, therefore, likely that the precursor 327 scan is detecting esterified docosahexanoic acid (m/z 327). Furthermore, levels of the ions detected in the precursor 327 scan do not change with cell activation (Fig. 9F). Therefore, it appears that exogenous 15-HETE-d8 is not being incorporated into the phospholipid pool during the timescale of our experiments. Furthermore, our earlier experiments using H 2 18 O also showed that endogenously generated 15-HETE is also not incorporated, since m/z 321 (15-HETE-18 O) was not found attached to monocyte phospholipids after ionophore activation (Fig. 7C).

Activated Human Platelets Generate the Same H(p)ETE-PE Products on Activation, however, with Exclusively 12-H(p)ETE-Human
platelets contain a different LOX isoform, 12-LOX, that forms 12-H(p)ETE as its sole product and can be activated by physiological agonists, such as collagen (34,35). Levels of free and esterified 12-H(p)ETE were determined after activation with either collagen, the glycoprotein VI-selective agonist convulxin or A23187. For all stimuli, esterified H(p)ETE was detected; however, the levels relative to free H(p)ETE were lower than those formed in human monocytes (Fig.  10A). Precursor ion ESI/MS/MS scanning of ionophore-activated human platelet extracts for m/z 319.2 showed identical ions to human monocytes at m/z 738, 764, 766, and 782 (Fig. 10B). In addition to 12-HETE, platelets also generate large quantities of a second arachidonate product, thromboxane B2, after activation of prostaglandin H synthase-1. However, precursor scanning for m/z 369.2 revealed no significant thromboxane B2-containing phospholipids (Fig.  10C). This indicates that PE-HETE adduct formation is restricted to LOX metabolites.
As for monocytes, the HETE-containing ions co-eluted with di-14:0-PE on normal-phase LC, indicating that they are PE phospholipids (not shown). By comparison with the internal standard, all 4 peaks increased in intensity after activation with collagen (approx 2-5-fold), convulxin (ϳ5-10 fold), and A23187 (10 -20-fold) (Fig. 10D). MS3 of the 4 ions using the 12-HETE daughter ion of m/z 319.2 revealed a characteristic 12-HETE fragmentation pattern with a diagnostic daughter ion at 179 (Fig. 11). No ions for 15-HETE or any other isomer are present, indicating that they originate specifically from 12-LOX turnover. Because the four ions have identical m/z as the monocyte products, they are proposed to be the same compounds as in Scheme 1 except for the substitution of 15with 12-H(p)ETE.

DISCUSSION
The cellular formation of complex LOX products is not characterized. Here we show that the majority of 15-HETE generated in IL-4 treated monocytes after cell activation is esterified and that the preferred phospholipid pool for this eicosanoid is PE (Figs. 1-6). Thus, novel complex LOX products are found in monocytes that may signal in immunity and inflammation resolution.
Comparison of levels of free 15-HETE after hydrolysis of either total lipid extract ( Fig. 1) or purified monocyte PE (Fig. 3) show that the PE-esterified HETE does not seem to account for all the esterified HETE formed. Specifically, total esterified HETE varied from 100 to 2000 ng/4 ϫ 10 6 cells (due to donor variation), but only 17.5 ng/4 ϫ 10 6 cells was detected after purification and hydrolysis of PE from one donor. This suggests that 15-HETE is esterified into additional nonphospholipid pools that have not been detected here, though perhaps not being ionized in negative mode. Examples could include diacylglycerides or triglycerides. Further experiments will address this issue and identify additional nonphospholipid sites of HETE esterification in the cells.
Previous studies examined sites of exogenous 15-HETE esterification into phospholipids and showed this to be mainly phosphatidylinositol or PC (12, 36 -40). Herein, we found that PE was the predominant phospholipid pool of 15-HETE after activation of endogenous 15-LOX. The key difference is that endogenously formed products are generated by direct oxidation of PE without prior release of arachidonate to form free 15-HETE ( Fig. 7 and Scheme 2). Therefore, this occurs via an entirely different process. Furthermore, during our experiments neither exogenous or endogenous 15-HETE was esterified into any phospholipid pool (Figs. 7C and 9). In human monocytes the predominant phospholipid is PC (40%), with PE only 30%; however, PC is enriched on the outer leaflet, with PE on the inside (41,42). This suggests that after LOX translocation to the membrane, direct oxidation of complex substrate occurs. Although the ability of purified 15-LOX to oxidize PC in vitro at low rates (compared with arachidonate) is known, PE oxidation by LOX has never been shown to be a source of endogenous products in monocytes or other cells. Phosphatidylinositol 3-kinase and protein kinase C regulated generation of monocyte HETE-PEs, identifying two new signaling pathways that control calciumdependent activation of 15-LOX (Fig. 1, D and E). Because PE-HETEs are formed independently of phospholipase A 2 , a direct action of these pathways on activity of the 15-LOX itself is inferred. The mechanisms involved remain to be determined. . Exogenous 15-HETE is not incorporated into PE. Monocytes were incubated with 15-HETE-d8 (330 ng/4 ϫ 10 6 cells) during ionophore activation, then lipids were reduced and extracted as described under "Experimental Procedures." Extracts were infused at 10 l/min into the ESI source, and precursor scanning was performed. amu, atomic mass units. Panels A and B, precursor scanning for endogenous 15-HETE at m/z 319 or 219 shows characteristic PE-HETEs in a representative ionophore-activated sample containing 15-HETE-d8. Panels C and D, precursor scanning for exogenous 15-HETE-d8 at m/z 327 or 226 in the same sample. Different patterns that are not consistent with formation of esterified HETEs are observed. Inset, precursor scan from an ionophore-activated sample not supplemented with 15-HETE-d8 shows the same ions as panel C. Panels E and F, levels of 15-HETE-PE but not 15-HETE-d8-PE are elevated upon activation. Diluted samples were injected under flow (1 ml/min methanol) into the ESI source in negative mode with specific MRM transitions monitored using either m/z 319 (15-HETE) or 327 (15-HETE-d8) as daughter ions and comparing intensity to the MRM transition of the internal standard (di-14:0-PE) (n ϭ 3, mean Ϯ S.E.). exposed to collision-induced dissociation, fatty acid esters preferentially fragment at the carboxyl, eliminating the free fatty acid as a neutral loss and also generating the acid anion which can be seen in negative mode. Additionally, eicosanoids fragment internally under these conditions, generating further ions diagnostic for the specific molecular species attached without the need for MS3. It works especially well with complex lipids that ionize in negative mode but may miss uncharged or positive species. Previous attempts to characterize the esterification of HETEs in cells involved the addition of exogenous radiolabeled eicosanoids followed by radio-TLC or TLC followed by saponification of individual lipids and radio-HPLC (12,16,43,44). However, these approaches suffer from several drawbacks. The behavior of exogenously added HETE is not the same as endogenously produced eicosanoid (18). Also, there are significant sensitivity issues precluding the detection and identification of individual lipid species within a class. It was possible to determine that HETE is contained within a particular phospholipid pool; however, identifying which molecular species contain HETE was never achieved. Precursor scanning MS/MS overcomes all these issues and provides a high sensitivity method with which to (i) identify series of ions that contain eicosanoids in crude lipid extracts without purification (Figs. 2, 10) and (ii) structurally identify each individual compound through subsequent MS/MS analysis of partially purified parent ions (Figs. 4). Additionally, using MS3, the exact HETE positional isomer composition of each lipid can be verified through fragmentation of the HETE (m/z 319.2 [M-H] Ϫ ) daughter ion generated by collision-induced dissociation of each individual parent (Figs. 5 and 11). In this study MS3 confirmed the enzymatic origin of each PE lipid generated by monocytes or platelets along with the more traditional chromatographic analysis of positional isomers and enantiomers (Fig. 6).
In summary, four specific PE-HETEs that form after activation of human monocytes and platelets have been identified and structurally characterized. The same pattern of products is conserved among two LOX isoforms. Furthermore, the utility of precursor ESI/MS/MS scanning for lipid mediators as a method for identifying families of esterified products that could signal in immune regulation and inflammation is demonstrated.