ANALYTICAL
BIOCHEMISTRY
198,
203-211
(1991)
Collisionally Induced Dissociation of Epoxyeicosatrienoic Acids and Epoxyeicosatrienoic Acid-Phospholipid Molecular Species Kerstin National
Received
Bernstrom, Jewish
April
Kathleen
Center for Immunology
Kayganich,
and Robert
and Respiratory
Medicine,
1400 Jackson
Street, Denver,
Colorado
80206
8, 1991
Four isomers of epoxyeicosatrienoic acid (EET) can be formed by cytochrome P-450 oxidation of arachidonic acid: 5,6-, 8,9-, 11,12-, and 14,15-epoxyeicosatrienoic acid. The collision-induced dissociation of the [M-H]anion at m/z 319 from each of these isomers, using negative-ion fast atom bombardment ionization and a triple quadrupole mass spectrometer, resulted in a series of common ions as well as ions characteristic of each isomer. The common ions were m/z 301 [M-H@and 257 [M-(H,O + CO,)]-. Unique ions resulted from cleavages (Y to the epoxide moiety to form either conjugated carbanions or aldehydes. Mechanisms involving charge site transfer are suggested for the origin of these ions. A distonic ion series that may involve a charge-remote fragmentation mechanism was also observed. The epoxyeicosatrienoic acids were also incorporated into cellular phospholipids following incubation of the free acid with murine mast cells in culture. Negative fast atom bombardment mass spectrometry of purified glycerophosphoethanolamine-EET species and glycerophosphocholine-EET species yielded abundant [M-H]and [M-CH,]ions, respectively. The collision-induced dissociation of these specific high-mass ions revealed fragment ions characteristic of the epoxyeicosatrienoic acids incorporated (m/z 319, 301, and 257) and the same unique ions as those seen with each isomeric epoxyeicosatrienoic acid. With this direct method of analysis, phospholipids containing the four positional isomers of EET, including the highly labile (5,6-EET), could be identified as unique molecular species in mast cells incubated with EET. o 1991 Academic press, IIN.
Arachidonic acid is known to be metabolized by various oxidative pathways including the cyclooxygenase, the lipoxygenase, and, most recently discovered, the cytochrome P-450-mediated pathway (1). Interest 0003-2697/91 Copyright All rights
C. Murphy
$3.00 0 1991 by Academic Press, of reproduction in any form
in oxygenation of this polyunsaturated fatty acid is largely derived from the biologically active metabolites formed by these pathways, including prostaglandins, thromboxane, and leukotrienes. The cytochrome P-450 pathway also yields biologically active molecules such as the epoxyeicosatrienoic acids (EETs)’ and the o/w-l hydroxy metabolites, which have attracted attention because of their effects on endocrine, renal, ocular, and secretory cell function (1). While interest in these P-450 metabolites is growing, it is still difficult to assesstheir physiological significance. Recent reports have demonstrated the physiological occurrence of several EETs, including 8,9-, 11,12-, and 14,15-EET in rat liver (2,3), 8,9- and 14,15-EET in renal tissue and urine (4,5), and the release of 14,15-EET from activated platelets (6). Most studies have reported the presence of the free carboxylic acid of the EETs; however, in rat liver the EETs were suggested to be esterified in phospholipids (7). In this study (7) the purified phospholipid classes were hydrolyzed with phospholipase A, and the EETs isolated and detected as pentafluorobenzyl esters using negative-ion chemical ionization mass spectrometry. Additional studies have followed the incorporation of specific EETs (14,15-EET) into the phospholipids of rat glomerular mesangial cells (8). Thus, it would appear that an important aspect of this route of arachidonic acid metabolism leads to the formation of oxidized products that become reincorporated into phospholipids following their formation. Direct analysis of the EET-containing phospholipid molecular species has not been previously reported,
i Abbreviations used: CID, collision-induced dissociation; EET, epoxyeicosatrienoic acid, FAB, fast atom bombardment; FCS, fetal calf serum; GPC, glycerophosphocholine; GPE, glycerophosphoethanolamine; HBSS, Hank’s balance salt solution; PLA,, phospholipase A,; rf, radiofrequency. 203
Inc. reserved.
204
BERNSTROM,
KAYGANICH,
perhaps in part due to difficulties inherent in the analysis of nonvolatile phospholipids. Here we report a direct method for the qualitative analysis of intact phospholipids containing epoxyeicosatrienoic acid, using negative-ion fast atom bombardment tandem mass spectrometry. With this technique, enzymatic degradation and subsequent derivatization are unnecessary to establish the exact isomer of EET as well as the other radyl group substituent on the giycerophospholipid backbone. This may help to minimize potential artifacts due to autooxidation or degradation of these polyunsaturated fatty acids.
MATERIAL
AND
METHODS
Arachidonic acid (>99% pure) was obtained from Nucheck Prep (Elysian, MN). [5,6,8,9,11,12,14,15-3H]Arachidonic acid (100 Ci/mmol) was obtained from DuPont-New England Nuclear (Boston, MA). 5,6-EET, 8,9-EET, 11,12-EET, and 14,15-EET were purchased from Cayman Chemical (Ann Arbor, MI) and their purity was checked by HPLC. All solvents were HPLC grade obtained from Fisher Scientific International. Reagents of the highest quality available were purchased from Aldrich Chemical Co. Epoxidution of [3Hjarachidonic acid. Radiolabeled EETs were prepared according to the method of Falck et al. (9) with minor modifications. [3H]Arachidonic acid (12 &i) diluted with 0.25 pmol cold arachidonic acid was evaporated to dryness using N,. 3-Chloroperoxybenzoic acid (Aldrich Chemical Co., 0.54 pmol) was dissolved in CH,Cl, (0.2 ml) and added to the arachidonic acid. After 12 h (stirring at O”C), 1 vol of CH,Cl, and 1.5 vol of NaHSO, were added and the mixture was kept at O’C for 30 min. The organic phase was washed with 1.5 vol of saturated NaCl and then taken to dryness. The radiolabeled EETs were purified by reverse-phase HPLC essentially as previously described (9). Phospholipids containing esteriHPLC separation. fied EETs were initially separated by normal-phase HPLC and then by reverse-phase HPLC as described below. Normal-phase HPLC (10) was performed using a 5 pm silica (Lichrosorb; Phenomenex, Torrence, CA) analytical column (4.6 X 250 mm). The mobile phase consisted of a gradient of 53% solvent A (hexaneiisopropanol; 3/4, v/v) held for 6 min then programmed to 100% solvent B (hexane/isopropanol/water; 31410.7, v/ v/v) over a 20-min period (system I). The solvent flow rate was 1 ml/min. Reverse-phase HPLC purification of phospholipids was performed essentially as described by Patton et al. (11) on Ultrasphere ODS columns (4.6 X 250 mm, 5 pm; Beckman, San Ramon, CA). The mobile phase consisted of methanol/water/acetonitrile (90.5/7/2.5, v/v/v) containing 1 mM NH,OOCCF, that had been adjusted to pH 7.4 (system II). This mobile
AND
MURPHY
phase contained a volatile buffer rather than that previously described. A flow rate of 2 ml/min was employed. Cell incubations. Virus-transformed murine bone marrow-derived mast cells (12) were maintained in tissue culture using RPM1 media (Whittaker Bioproducts Inc., Walkersville, MD) supplemented with 10% fetal calf serum (FCS). Fresh media (20 ml) were added every 2 days to the cell cultures and cell cultures were divided at weekly intervals when cell concentrations reached 1 X 106/ml. Cells were transferred to media containing 0.4% FCS, l-2 days prior to individual experiments. Cells were harvested for subsequent studies (1.6-2.6 X 10’ cells) from the tissue culture media by centrifugation at 50g for 15 min at room temperature, washed with Hank’s balance salt solution (HBSS), and then suspended to a final concentration of 0.7 x l@/ml. Cells were then incubated for 2 h unless otherwise indicated with 1 pM 14,15-EET, 11,12-EET, 8,9-EET, or 5,6-EET that had been dissolved in 1 ml HBSS containing 5 mg/ ml bovine serum albumin. After incubation at 37”C, the suspension was centrifuged at 1OOg for 15 min, and washed once with fresh HBSS containing 0.25 mg/ml bovine serum albumin. Phospholipids were extracted from the cell pellet following the method of Bligh and Dyer (13) and then purified with normal-phase (system I) and reverse-phase (system II) HPLC. Mass spectrometry. Collision-induced dissociation mass spectra of individual EETs were obtained from approximately 20-100 ng of material placed on a fast atom bombardment (FAB) probe of a Finnigan TSQ 70 (San Jose, CA) triple quadrupole mass spectrometer. Negative ions were produced using diethanolamine (J. T. Baker, Phillipsburg, NJ) as liquid matrix. The EETs were normally stored at -70°C and kept on ice just prior to FAB analysis. The fast atom bombardment gun (IonTech, Teddington, UK) was operated at 1 mA with xenon accelerated to 6 kV. Argon was used as the collision gas at a pressure approximately 0.5 mTorr in the second quadrupole region of the tandem mass spectrometer. The collision offset energy (Ebb) was 30 eV. The electron multiplier conversion dynode was maintained at 12 kV. Precursor and product ion spectra were obtained using the TSQ 70 data system and typically between four and six spectra were averaged. Identical fast atom bombardment mass spectrometry conditions were used to obtain the negative-ion spectra of isolated phospholipids from the virus-transformed mast cells that had been incubated with the EETs. Phospholipid collision-induced dissociation (CID) spectra shown were normalized to the most abundant product ion. Phospholipase A, assay. Purified EET-containing phospholipids were subjected to phopholipase A, (PLA,) hydrolysis as previously described (14) using 50-100 pg of the EET-phosphatidylcholine species. In addition, the lysoglycerophosphocholine products were
EET-PHOSPHOLIPID
ANALYSIS
BY
TANDEM
MASS
205
SPECTROMETRY
x 05
0 8 0
319
[M-HI
[M-H]
-
167 m/z319 s 9 f 2 t
so
100
179
208
-(Hz0
-Hz0
+ CO,,
59
301 257
I 150
200
250
300
M/Z
MI
x 05
x 05
D
319
100
Z
319
MO-
8 0
W-HI 8
W-RI P 8s
m/7.319
s
9 t -or*
0 + CO,)
2 Y z 2
-H,O Jo1
191
m/z 319
- 59
137
163
-(H,o
257 ,!,I-!,,,, so
100
150
219
,,,,
loo
250
I
,,,,(
Ml
39
so
MIZ
100
150
Is .I
11,llll,rrll,l 200
+ CO,) \
275 I
25) I.1
250
I
‘z; I
_
II
300
350
MIZ
FIG. 1. Collision-induced dissociation of the fast atom bombardment-generated carboxylate anions from four isomeric epoxyeicosatrienoic acids (EETs) in a tandem quadrupole mass spectrometer. (A) CID of 14,15-EET. (B) CID of 11,12-EET. (C) CID of 8,9-EET. (D) CID 5,6-EET. All spectra were obtained following reduction of the molecular anion abundance at m/z 319 to approximately 30% by the pressure argon in the rf-only quadrupole collision cell.
recovered from the PLA, incubations by Bligh-Dyer traction (13) and analyzed by FABIMS.
-
ex-
RESULTS
The negative ions obtained by fast atom bombardment mass spectrometry of the isomeric EETs, (5,6EET, 8,9-EET, 11,12-EET, and 14,15-EET) consisted of an abundant [M-H]ion at m/z 319. The ions present in the ion source that constitute the negative-ion FAB mass spectrum of these molecules were dominated by this [M-H]ion, most likely the closed shell carboxylate anion (data not shown). The CID of the [M-H]from the four isomeric EETs (m/z 319) produced structurally significant decomposition ions characteristic for each isomer (Fig. 1). In general, the fragmentation of these isomeric molecules was dominated by the position of the epoxide ring along the arachidonic chain. The ion structures suggested in Fig. 1 could originate either from charge remote rearrangement mechanisms or following charge transfer from the carboxylate anion to various sites by abstraction of a bis-allylic proton on each EET.
of of
Additionally, each EET isomer loses the elements of water to form an ion at m/z 301 and loses 62 D (presumably loss of water as well as CO,) to form an ion at m/z 257. Specifically, the CID product ions from 14,15-EET (Fig. 1A) yielded decomposition ions at m/z 219 (loss of the neutral aldehyde with a proton rearrangement), m/z 175 (loss of the neutral aldehyde as well as CO, with a proton rearrangement), and m/z 113 (an enolate anion fragment). The two possible mechanisms mentioned above leading to m/z 219 are illustrated in Scheme I. These same mechanisms can account for the ion at m/z 179 in 11,12-EET (Fig. 1B) and that at m/z 139 in 8,9EET (Fig. 1C). For the 11,12-EET (Fig. lB), the most abundant CID product ion (m/z 167), produced following decomposition of m/z 319 could arise from a rearrangement similar to that shown in Scheme I, resulting in cleavage of the lo,11 bond with delocalization of the anionic site by the allylic olefin. A similar ion was observed for 8,9EET at m/z 127. The ion at m/z 208 from 11,12-EET is rather curious since it is the only abundant odd electron
206
BERNSTROM,
KAYGANICH,
AND
SCHEME
ion species observed in any EET studied. It was therefore rather diagnostic of 11,12-EET. A homolytic cleavage mechanism of the epoxide ring followed by homolytic cleavage of the 12,13 carbon-carbon bond would account for the d&tonic ion structure shown in Fig. 1B. Less abundant ions in this series were observed at m/z 248 and 168 for 14,15-EET and 8,9-EET, respectively. The CID of 8,9-EET yielded the most numerous ion fragments of all the EETs, most likely due to the central location of the epoxide ring (Fig. 1C). In addition to the ion structures already presented above, abundant ions were observed at m/z 151, 155, 163, and 167; m/z 151 appears as a member of the ion series which is most abundant in 5,6-EET at m/z 191 presented below. The ions at m/z 155 and 167 may result from a homolytic epoxide cleavage reaction shown in Scheme II. The mechanisms suggested in Scheme II are charge remote processes similar to that suggested in Scheme I. The collisional activation of m/z 319 from 5,6-EET yielded quite abundant fragment ions. It is possible that the negative charge when localized at the carboxylate oxygen can attack the 5,6-epoxide in a favorable 6membered ring transition state, resulting in formation of the lactone alkoxide anion seen in Scheme III. This anion can rearrange to form the stabilized allylic anion at m/z 191. The collisional events in the rf-only quadrupole provide energy for this process, which is likely endothermic, as shown in Scheme III. The alkoxide anion could also change the charge site by an intramolecular proton transfer to yield a bis-allylic anion with the
MURPHY
I
charge site localized at Cl3 as illustrated in Structure A. Through further charge migrations and proton rearrangement this ion could decompose to the diene anion (m/z 137) or the triene anion (m/z 163). These structures are illustrated in Fig. 1D. Analysis of phospholipids containing esterified EETs. Incubation of cultured bone marrow-derived mast cells with radiolabeled and cold EETs resulted in the incorporation of EETs into various phospholipid classes as revealed by normal-phase HPLC separation of the lipid classes and scintillation counting of fractions (Fig. 2A). These results are summarized in Table
8.9.EET
m/z 167
ml?. 319
SCHEME
II
EET-PHOSPHOLIPID
ANALYSIS
BY
0 o
00
/ T/n+ \
\
m/z
SCHEME
191
III
1. Normal-phase HPLC revealed modest separation of the radiolabeled molecular species from the major components of glycerophosphoethanolamine (GPE) and glycerophosphocholine (GPC), typically eluting a few minutes after the majority of the phospholipid in each class (Fig. 2A). Reverse-phase HPLC separation of the radiolabeled GPE and GPC fractions resulted in the elution of several unique radiolabeled EET-phospholipid molecular species. These were well separated from other major fatty acyl molecular species (Fig. 2B). The methyl esters of the free fatty acids from isolated EET-phosphatidylcholines obtained by treatment with PLA, were analyzed using GC/MS. A significant amount of EET was seen, but other fatty acids were also seen. The results indicated that EET was present in the sn-2 position, but did not exclude the possibility of sn-1 esterification of EET. Designations of the carbon atoms of the glycerol portion of glycerophospholipids as sn-1 or sn-2 are shown in Fig. 3 (see legend). The additional PLA, products, 2-lysophosphatidylcholines, were analyzed by FAB mass spectrometry to determine the ratio of EET in the sn-1 position to that in the sn-2 position. The results indicated that approximately 80% of the EET was incorporated into the sn-2 position and 20% into the sn-1 position (on the basis of four different molecular species). The collision-induced dissociation of purified EETphospholipid molecular species yielded abundant fragment ions characteristic of both the phospholipid class and the unique EET acyl portion of the molecule. For example, analysis of one EET-glycerophosphocholine molecular species obtained following incubation of 11,12-EET with mast cells resulted in ions corresponding to M-15, M-60, and M-86 ions observed at m/z 782, 737, and 711 (data not shown), in agreement with previously reported behavior of glycerophosphocholine molecular species by negative-ion FAB mass spectrometry
TANDEM
MASS
207
SPECTROMETRY
(15). Abundant fatty acyl ions were observed at m/z 255 (hexadecanoate) and m/z 319 (EET). Collision-induced dissociation of the M-15 (m/z 782) from this GPC molecular species (16:Oa/EET-GPC) yielded abundant carboxylate anions, indicating the esterification of both fatty acyl groups (Fig. 3A) as previously described (14-16). Unlike previously reported CID spectra of M-15 from GPC, numerous fragment ions were also observed at masses lower than the expected acyl carboxylate anions, including m/z 167,179, and 208. These are the identical ions observed in the CID of m/z 319 from 11,12-EET (Fig. 1B). In this experiment, m/z 319 was produced in the second quadrupole collision cell from the glycerophospholipid ion at m/z 782 (M-15) and most likely underwent further collisional activation and decomposition to yield the characteristic ions discussed above for 11,12-EET. Such events of secondary collision activation have been previously described for triple quadrupole instruments (17). Identical behavior was observed for other phospholipid classes and subclasses containing 11,12-EET, as illustrated in Fig. 3B for 16:Op/ll,l2-EET-GPE. In this example, decomposition of m/z 738 ([M-H]) yielded the fatty acyl ion from the sn-2 substituent (EET, m/z 319), loss of the sn-2 (R,) substituent as ketene (-R,=C=O, m/z 436), and a neutral loss of the sn-2 carboxylic acid (-R,COOH, m/z 418). Again, abundant ions resulting from further collision-induced decomposition of the EET carboxylate anion were observed at m/z 167,179, and 208. The decomposition of the EET carboxylate anion (m/z 319) from isomeric GPE molecular species resulted in the same ion products characteristic of the EET isomer, as illustrated in Figs. 4A-4C. These phospholipid molecular species (16:Op/EET-GPE) were isolated from mast cells incubated with 14,15-EET, 8,9EET, and 5,6-EET, respectively. Decomposition of the intact phospholipid anion (m/z 738) thus provides specific information on the position of the epoxide ring on the arachidonate backbone. As seen in Fig. 4C, the 5,6EET-specific ions observed in the CID of m/z 319 from 5,6-EET (Fig. 1D) are produced at m/z 191, 163, and 137. Even the suspected lactone anion at m/z 99 is observed. Since this product ion scan came from decompo-
m/z319
STRUCTURE
A
208
BERNSTROM,
KAYGANICH,
AND
A
MURPHY
B
1500 1000
D s
E f B
z: 9
500
0
10
20 ELUTION
30 TIME,
40
3 /
MIN
ELUTION
TIME,
MIN
FIG. 2. (A) Normal-phase HPLC separation of phospholipids extracted from bone marrow-derived mast cells incubated with [11,12‘H]EET and cold 11,12-EET for 2 h. The HPLC chromatograph was monitored with uv absorption at 206 nm. Fractions were taken in 2-min intervals. Aliquots of each fraction were analyzed for radioactivity content expressed in dpm. (B) Reverse-phase HPLC of the GPE fraction of phospholipids purified by the normal-phase HPLC above. One-minute fractions were collected from the HPLC effluent and an aliquot was taken for determination of the radioactivity content. The HPLC-purified phospholipid molecular species indicated (1,2,3,4,5) were then subjected to fast atom bombardment mass spectrometric analysis.
sition of m/z 738 in all cases and m/z 319 is a prominent anion produced, it is not known with this experiment whether the fragment ions are formed directly from the [M-H]of the phospholipid or proceed from a carboxylate anion intermediate. The pressure of the collision gas in the rf-only quadrupole was sufficiently high for multiple collisions to take place. Under these conditions, fragment ions formed by CID of the parent ion may undergo collisions and fragment further (17). Over 20 molecular species of phospholipids that differ in polar head group and substitution at sn-1 have been studied and all have been found to produce CID product ions characteristic of the specific regioisomer of EET.
TABLE Incorporation of Two pholipids of Mast Cells Distribution of [3HJEETs
1
[3H]EET Isomers into Glycerophosfollowing 60 min of Incubation and into Glycerophospholipid Classes
14,15-EET Phospholipid class GPE GPI GPS GPC
% Added” 8.9 7.5 0.8 13
+ 2 + 3 f 0.2 &4
11,12-EET
% Incorporatedb 22 14 1.7 25
+ 7’ *4 * 0.4 +4
% Added
% Incorporated
15.@ 6.1 1.2 14.4
7.0 14.3 2.6 34.2
D Percentage of initially added [3H]EET recovered in phospholipid class. * Percentage of total phospholipid-associated radioactivity in phospholipid class. ’ Mean of four separate incubations, &EM. d Average of two separate incubations.
DISCUSSION
The fragmentation of even electron organic negative ions such as carboxylate anions is somewhat less well understood than the ion chemistry of positive ions typically encountered in electron impact ionization of molecules such as EET. The general decomposition behavior of negative ion species has been studied in some detail (18); however, the behavior of unsaturated epoxy acids has not been reported prior to this study. We have postulated that the decomposition of carboxylate anions from EETs (m/z 319) is dominated by reactions that either occur through homolytic cleavage of the epoxide moiety or are initiated following charge site transfer mechanisms involving the attack of the carboxylate anion upon bis-allylic protons forming stabilized anion sites. It is of interest to note that these reactions could be induced by the low-energy collisions that occur in the triple quadrupole mass spectrometer. Even though charge-remote fragmentation was initially observed under high-energy-collision conditions (19), it has been demonstrated that the energy required to induce charge-remote fragmentations varies widely depending on compound type, method of ionization, mass of collision target, and target gas pressure (17). Additionally, Wysocki and Ross (17) suggest that the compound dependence of charge-remote fragmentation processes is due to competition between charge-remote fragmentations and charge-driven reactions. Competitive fragmentation processes may be operative in compounds such as the EETs, which have acidic bis-allylic protons, capable of intramolecular transfer to the carboxyl moiety. Alternatively, some ions may be formed by reac-
EET-PHOSPHOLIPID
ANALYSIS
BY
TANDEM
MASS
209
SPECTROMETRY
M-H-
:
M-15
2 % d : ‘Z s ct
100
200
300
400
500
600
700
800
M/Z
436 _
167 179
46
-RjCOOH
-I
I 208
loo
200
41s
300
400
500
600
700
M/Z
FIG. 3. Tandem mass spectrometry of phospholipid molecular species purified by two steps of HPLC ment. R2 corresponds to a C,,H,,,O moiety from the EET esterified at sn-2 lost as ketene (R2=C=0) carbon atoms designated by convention as an-1 and m-2 are correspondingly numbered in each structure. glycerophosphocholine molecular species following incubation of 11,12-EET with murine mast cells. (M-15, m/z 782) was subjected to collision-induced dissociation, resulting in the formation of abundant and the decomposition ions of 11,12-EET. This molecular species was identified as l&Oa/ll,l2-EET-GPC. a major glycerolphosphoethanolamine molecular species obtained in the same 11,12-EET incubation. molecular anion at m/z 738 resulted in the formation of one acyl ion at m/z 319 as well as EET-specific
tion mechanisms in which an ion complex is formed with either a hydride ion or a hydroxyl ion followed by intramolecular reaction and decomposition of the molecule (18). Such a mechanism may be operating in the formation of the conjugated species at m/z 137 and 103 in 8,9-EET and 5,6-EET. While the suggested reaction mechanisms are consistent with known ion chemistry of negative ions, further experiments with stable isotopelabeled EETs would be required to confirm these suggested processes. Regardless of the exact mechanisms of these fragmentations, the mass spectra observed for each of the four isomeric EETs uniquely identify the epoxide position along the eicosanoid backbone. Collision-induced dissociation of the anions derived from phosphatidylcholine and phosphatidylethanolamine revealed esterification of EETs at the glycerol backbone through the appearance of an ion at m/z 319. In the CID spectra of the [M-H]and [M-CHJions from all of the diacylphospholipids studied, m/z 319 was less abundant than the other carboxylate anion. This would indicate that EET was esterified at sn-I, as the abundance of the sn-2 carboxylate anion is normally greater than that of the sn-1 carboxylate anion in the MS/MS experiment of these precursor ions (14-16). However, exceptions to the ion abundance of the sn-2 > sn-1 anion relationship have been reported for molecular species having acyl groups with extensive unsaturation and/or a substantial difference between the chain lengths of the two acyl groups (20). The presence of sn-l/sn-2 positional isomers of these EET-phospholipids, as revealed by the results of the PLA, hydrolysis
I
16:Opi 11,12-EET-GPE
-R2=C=0
and ionized by fast atom bombardor a carboxylic acid (R&OOH). The (A) Analysis of a major radioactive The major negative ion at high mass carboxylate anions (m/z 255 and 319) (B) Tandem mass spectrometry of Collision-induced dissociation of the ions from 16:Op/ll,lZ-EET-GPE.
experiments, further complicates the simple interpretation of the sn-1 and sn-2 substituents from the relative abundances of the carboxylate anions in the CID spectra. Until pure standards of these EET-phospholipids are available so that the CID behavior of the pure compounds can be characterized, the interpretation of these CID spectra regarding sn-1 and sn-2 substitution remains uncertain. CID analysis of the M-15 anion from GPC as well as the [M-H]anion from GPE molecular species reveals ions characteristic of the position of the epoxy group of the esterified EET in addition to the sn-1 fatty acyl group also esterified on the glycerolipid backbone. Thus, it is possible to uniquely determine not only the presence of EET in the glycerophospholipid molecular species by tandem mass spectrometry, but also the exact isomer of EET (5,6-, 8,9-, 11,12-, or 14,15-EET) that is present in the oxidized phospholipid molecular species. Such information has been impossible to obtain except by following extensive decomposition of the phospholipid followed by separate analysis of the various products. The use of the methodology described here for the direct analysis of EET-phospholipids minimizes the probability of sample autooxidation and decomposition as is likely to occur when conventional means are used to analyze these labile compounds. Indeed, the efficacy of this methodology is demonstrated by the characterization of several intact 5,6-EET-phospholipid species. The detection of 5,6-EET from in uivo sources has not been demonstrated before because of its facile hydrolysis to the 5,6-diol during conventional analytical
210
BERNSTROM,
KAYGANICH,
AND
MURPHY
‘001
I
436
I
M/Z
M/Z
83
191
1
eb A
436
I
191
[M-HI’ 16:0p/5,6-EET
319
GPE 73ll
X
x4I ‘Z zCL
137
M/Z
FIG. 4. Tandem mass spectrometry of the 16:0 plasmalogen EET-containing molecular species of glycerophosphoethanolamine purified by HPLC and analyzed by fast atom bombardment mass spectrometry following incubation with isomeric EETs. I!& corresponds to a CleHsoO moiety from the EET esterified at sn-2 lost as ketene (b=C=O) or a carboxylic acid (R&OOH). (A) 160 Plasmalogen obtained following incubation of 14,15-EET with mast cells. (B) Plasmalogen obtained following incubation of 8,9-EET with mast cells. (C) Plasmalogen obtained following incubation of 5,6-EET with mast cells.
workup. The analytical capability of negative FAB tandem mass spectrometry should enable a more detailed investigation of the endogenous production of EETs and their metabolic fate in terms of reincorporation into phospholipids. From such information it should be possible to better understand the P-450-mediated metabolism of arachidonic acid and the formation of the epoxygenase-derived eicosanoids as well as the role that they may play in various physiological and pathophysiologic processes. ACKNOWLEDGMENTS This work was supported, in part, by a grant from the National Institutes of Health (GM41026). The authors acknowledge helpful discussions with Dr. Veronica Bierbaum (University of Colorado) concerning the negative ion chemistry of these molecules.
REFERENCES 1. Fitzpatrick, 229-241.
F. A., and Murphy,
R. C. (1989)
Pharmacol. Rev. 40,
2. Capdevila, J., Pramanik, B., Napoli, J., Manna, S., and Falck, J. (1984) Arch. Biochem. Biophys. 231.511-517. A., Dishman, E., Blair, I., Falck, J., and Capdevila, J. 3. Karara, (1989) J. Biol. Chem. 264, 19,822-19,827. V., Jacobson, H., Siddhanta, A., Pramanik, 4. Falck, J., Schueler, B., and Capdevila, J. (1987) J. Lipid Res. 28, 640-646. A., Manna, S., Pramanik, B., Falck, J., and 5. Toto, R., Siddhanta, Capdevilla, J. (1987) Biochim. Biophys. Actu 919, 132-139. 6. Ballou, L., Lam, B., Wong, P.-K., and Cheung, W. Y. (1987) Proc. Natl. Acad. Sci. USA 84,6690-6694. 7. Capdevila,
B&hem. 8. Harris, (1990)
J. H., Kishore,
V., Dishman,
E., and Blair,
I. A. (1987)
Biophys. Res. Commun. 146,638-644. R. C., Homma,
T., Jacobson,
J. Cell Physiol. 144,429-437.
H. R., and
Capdevila,
J.
EET-PHOSPHOLIPID
ANALYSIS
9. Falck, J. R., Yadagiri, P., and Capdevila, J. (1990) Enzymology, p. 357, Academic Press, New York. 10. Rivnay,
B. (1984)
11. Patton,
G. M., Fasulo,
J. Chromatogr.
in Methods
BY in
S. J. (1982)
Res.
12. Pierce, J. H., DiFiore, P. D., Aaronson, S. A., Potter, Pumphrey, J., Scott, A., and Ihle, I. N. (1985) Cell 41,685. 13. Bligh, E. G., and Dyer, W. J. (1959) Con. J. Biochem. Physiol. 911-917.
M.,
23,190-196.
R. C. (1991)
15. Jensen,
N. J., Tomer,
16. Munster, J. Lipid
14. Kayganich, K., and Murphy trom. 2.45-54.
MASS
211
SPECTROMETRY K. B., and Gross,
M. L. (1986)
Lipids
21,
580-588.
294,303-315.
J. M., and Robins,
TANDEM
J. Am. Sot. Mass
37, Spec-
H., and
Budzikiewicz,
H. (1988)
Biol.
Chem.
Hoppe-
Seykr369,303-308. 17. Wysocki, zon Proc.
V. H., and Ross, 104,179-211.
M.
M.
(1991)
Int.
J. Muss
Spectrom.
18. Bowie, J. H. (1990) Mass Spectrom. Reu. 9, 349-379. 19. Jensen, N. J., Tomer, K. B., and Gross, M. L. (1985) J. Am. Chem. Sot. 10'7.1863-1868. 20. Huang, Z.-H., Gage, D. A., and Sweeley, Mass Spectrom., in press.
C. C. (1991)
J. Am.
Sot.
BIOCHEMISTRY
198,
203-211
(1991)
Collisionally Induced Dissociation of Epoxyeicosatrienoic Acids and Epoxyeicosatrienoic Acid-Phospholipid Molecular Species Kerstin National
Received
Bernstrom, Jewish
April
Kathleen
Center for Immunology
Kayganich,
and Robert
and Respiratory
Medicine,
1400 Jackson
Street, Denver,
Colorado
80206
8, 1991
Four isomers of epoxyeicosatrienoic acid (EET) can be formed by cytochrome P-450 oxidation of arachidonic acid: 5,6-, 8,9-, 11,12-, and 14,15-epoxyeicosatrienoic acid. The collision-induced dissociation of the [M-H]anion at m/z 319 from each of these isomers, using negative-ion fast atom bombardment ionization and a triple quadrupole mass spectrometer, resulted in a series of common ions as well as ions characteristic of each isomer. The common ions were m/z 301 [M-H@and 257 [M-(H,O + CO,)]-. Unique ions resulted from cleavages (Y to the epoxide moiety to form either conjugated carbanions or aldehydes. Mechanisms involving charge site transfer are suggested for the origin of these ions. A distonic ion series that may involve a charge-remote fragmentation mechanism was also observed. The epoxyeicosatrienoic acids were also incorporated into cellular phospholipids following incubation of the free acid with murine mast cells in culture. Negative fast atom bombardment mass spectrometry of purified glycerophosphoethanolamine-EET species and glycerophosphocholine-EET species yielded abundant [M-H]and [M-CH,]ions, respectively. The collision-induced dissociation of these specific high-mass ions revealed fragment ions characteristic of the epoxyeicosatrienoic acids incorporated (m/z 319, 301, and 257) and the same unique ions as those seen with each isomeric epoxyeicosatrienoic acid. With this direct method of analysis, phospholipids containing the four positional isomers of EET, including the highly labile (5,6-EET), could be identified as unique molecular species in mast cells incubated with EET. o 1991 Academic press, IIN.
Arachidonic acid is known to be metabolized by various oxidative pathways including the cyclooxygenase, the lipoxygenase, and, most recently discovered, the cytochrome P-450-mediated pathway (1). Interest 0003-2697/91 Copyright All rights
C. Murphy
$3.00 0 1991 by Academic Press, of reproduction in any form
in oxygenation of this polyunsaturated fatty acid is largely derived from the biologically active metabolites formed by these pathways, including prostaglandins, thromboxane, and leukotrienes. The cytochrome P-450 pathway also yields biologically active molecules such as the epoxyeicosatrienoic acids (EETs)’ and the o/w-l hydroxy metabolites, which have attracted attention because of their effects on endocrine, renal, ocular, and secretory cell function (1). While interest in these P-450 metabolites is growing, it is still difficult to assesstheir physiological significance. Recent reports have demonstrated the physiological occurrence of several EETs, including 8,9-, 11,12-, and 14,15-EET in rat liver (2,3), 8,9- and 14,15-EET in renal tissue and urine (4,5), and the release of 14,15-EET from activated platelets (6). Most studies have reported the presence of the free carboxylic acid of the EETs; however, in rat liver the EETs were suggested to be esterified in phospholipids (7). In this study (7) the purified phospholipid classes were hydrolyzed with phospholipase A, and the EETs isolated and detected as pentafluorobenzyl esters using negative-ion chemical ionization mass spectrometry. Additional studies have followed the incorporation of specific EETs (14,15-EET) into the phospholipids of rat glomerular mesangial cells (8). Thus, it would appear that an important aspect of this route of arachidonic acid metabolism leads to the formation of oxidized products that become reincorporated into phospholipids following their formation. Direct analysis of the EET-containing phospholipid molecular species has not been previously reported,
i Abbreviations used: CID, collision-induced dissociation; EET, epoxyeicosatrienoic acid, FAB, fast atom bombardment; FCS, fetal calf serum; GPC, glycerophosphocholine; GPE, glycerophosphoethanolamine; HBSS, Hank’s balance salt solution; PLA,, phospholipase A,; rf, radiofrequency. 203
Inc. reserved.
204
BERNSTROM,
KAYGANICH,
perhaps in part due to difficulties inherent in the analysis of nonvolatile phospholipids. Here we report a direct method for the qualitative analysis of intact phospholipids containing epoxyeicosatrienoic acid, using negative-ion fast atom bombardment tandem mass spectrometry. With this technique, enzymatic degradation and subsequent derivatization are unnecessary to establish the exact isomer of EET as well as the other radyl group substituent on the giycerophospholipid backbone. This may help to minimize potential artifacts due to autooxidation or degradation of these polyunsaturated fatty acids.
MATERIAL
AND
METHODS
Arachidonic acid (>99% pure) was obtained from Nucheck Prep (Elysian, MN). [5,6,8,9,11,12,14,15-3H]Arachidonic acid (100 Ci/mmol) was obtained from DuPont-New England Nuclear (Boston, MA). 5,6-EET, 8,9-EET, 11,12-EET, and 14,15-EET were purchased from Cayman Chemical (Ann Arbor, MI) and their purity was checked by HPLC. All solvents were HPLC grade obtained from Fisher Scientific International. Reagents of the highest quality available were purchased from Aldrich Chemical Co. Epoxidution of [3Hjarachidonic acid. Radiolabeled EETs were prepared according to the method of Falck et al. (9) with minor modifications. [3H]Arachidonic acid (12 &i) diluted with 0.25 pmol cold arachidonic acid was evaporated to dryness using N,. 3-Chloroperoxybenzoic acid (Aldrich Chemical Co., 0.54 pmol) was dissolved in CH,Cl, (0.2 ml) and added to the arachidonic acid. After 12 h (stirring at O”C), 1 vol of CH,Cl, and 1.5 vol of NaHSO, were added and the mixture was kept at O’C for 30 min. The organic phase was washed with 1.5 vol of saturated NaCl and then taken to dryness. The radiolabeled EETs were purified by reverse-phase HPLC essentially as previously described (9). Phospholipids containing esteriHPLC separation. fied EETs were initially separated by normal-phase HPLC and then by reverse-phase HPLC as described below. Normal-phase HPLC (10) was performed using a 5 pm silica (Lichrosorb; Phenomenex, Torrence, CA) analytical column (4.6 X 250 mm). The mobile phase consisted of a gradient of 53% solvent A (hexaneiisopropanol; 3/4, v/v) held for 6 min then programmed to 100% solvent B (hexane/isopropanol/water; 31410.7, v/ v/v) over a 20-min period (system I). The solvent flow rate was 1 ml/min. Reverse-phase HPLC purification of phospholipids was performed essentially as described by Patton et al. (11) on Ultrasphere ODS columns (4.6 X 250 mm, 5 pm; Beckman, San Ramon, CA). The mobile phase consisted of methanol/water/acetonitrile (90.5/7/2.5, v/v/v) containing 1 mM NH,OOCCF, that had been adjusted to pH 7.4 (system II). This mobile
AND
MURPHY
phase contained a volatile buffer rather than that previously described. A flow rate of 2 ml/min was employed. Cell incubations. Virus-transformed murine bone marrow-derived mast cells (12) were maintained in tissue culture using RPM1 media (Whittaker Bioproducts Inc., Walkersville, MD) supplemented with 10% fetal calf serum (FCS). Fresh media (20 ml) were added every 2 days to the cell cultures and cell cultures were divided at weekly intervals when cell concentrations reached 1 X 106/ml. Cells were transferred to media containing 0.4% FCS, l-2 days prior to individual experiments. Cells were harvested for subsequent studies (1.6-2.6 X 10’ cells) from the tissue culture media by centrifugation at 50g for 15 min at room temperature, washed with Hank’s balance salt solution (HBSS), and then suspended to a final concentration of 0.7 x l@/ml. Cells were then incubated for 2 h unless otherwise indicated with 1 pM 14,15-EET, 11,12-EET, 8,9-EET, or 5,6-EET that had been dissolved in 1 ml HBSS containing 5 mg/ ml bovine serum albumin. After incubation at 37”C, the suspension was centrifuged at 1OOg for 15 min, and washed once with fresh HBSS containing 0.25 mg/ml bovine serum albumin. Phospholipids were extracted from the cell pellet following the method of Bligh and Dyer (13) and then purified with normal-phase (system I) and reverse-phase (system II) HPLC. Mass spectrometry. Collision-induced dissociation mass spectra of individual EETs were obtained from approximately 20-100 ng of material placed on a fast atom bombardment (FAB) probe of a Finnigan TSQ 70 (San Jose, CA) triple quadrupole mass spectrometer. Negative ions were produced using diethanolamine (J. T. Baker, Phillipsburg, NJ) as liquid matrix. The EETs were normally stored at -70°C and kept on ice just prior to FAB analysis. The fast atom bombardment gun (IonTech, Teddington, UK) was operated at 1 mA with xenon accelerated to 6 kV. Argon was used as the collision gas at a pressure approximately 0.5 mTorr in the second quadrupole region of the tandem mass spectrometer. The collision offset energy (Ebb) was 30 eV. The electron multiplier conversion dynode was maintained at 12 kV. Precursor and product ion spectra were obtained using the TSQ 70 data system and typically between four and six spectra were averaged. Identical fast atom bombardment mass spectrometry conditions were used to obtain the negative-ion spectra of isolated phospholipids from the virus-transformed mast cells that had been incubated with the EETs. Phospholipid collision-induced dissociation (CID) spectra shown were normalized to the most abundant product ion. Phospholipase A, assay. Purified EET-containing phospholipids were subjected to phopholipase A, (PLA,) hydrolysis as previously described (14) using 50-100 pg of the EET-phosphatidylcholine species. In addition, the lysoglycerophosphocholine products were
EET-PHOSPHOLIPID
ANALYSIS
BY
TANDEM
MASS
205
SPECTROMETRY
x 05
0 8 0
319
[M-HI
[M-H]
-
167 m/z319 s 9 f 2 t
so
100
179
208
-(Hz0
-Hz0
+ CO,,
59
301 257
I 150
200
250
300
M/Z
MI
x 05
x 05
D
319
100
Z
319
MO-
8 0
W-HI 8
W-RI P 8s
m/7.319
s
9 t -or*
0 + CO,)
2 Y z 2
-H,O Jo1
191
m/z 319
- 59
137
163
-(H,o
257 ,!,I-!,,,, so
100
150
219
,,,,
loo
250
I
,,,,(
Ml
39
so
MIZ
100
150
Is .I
11,llll,rrll,l 200
+ CO,) \
275 I
25) I.1
250
I
‘z; I
_
II
300
350
MIZ
FIG. 1. Collision-induced dissociation of the fast atom bombardment-generated carboxylate anions from four isomeric epoxyeicosatrienoic acids (EETs) in a tandem quadrupole mass spectrometer. (A) CID of 14,15-EET. (B) CID of 11,12-EET. (C) CID of 8,9-EET. (D) CID 5,6-EET. All spectra were obtained following reduction of the molecular anion abundance at m/z 319 to approximately 30% by the pressure argon in the rf-only quadrupole collision cell.
recovered from the PLA, incubations by Bligh-Dyer traction (13) and analyzed by FABIMS.
-
ex-
RESULTS
The negative ions obtained by fast atom bombardment mass spectrometry of the isomeric EETs, (5,6EET, 8,9-EET, 11,12-EET, and 14,15-EET) consisted of an abundant [M-H]ion at m/z 319. The ions present in the ion source that constitute the negative-ion FAB mass spectrum of these molecules were dominated by this [M-H]ion, most likely the closed shell carboxylate anion (data not shown). The CID of the [M-H]from the four isomeric EETs (m/z 319) produced structurally significant decomposition ions characteristic for each isomer (Fig. 1). In general, the fragmentation of these isomeric molecules was dominated by the position of the epoxide ring along the arachidonic chain. The ion structures suggested in Fig. 1 could originate either from charge remote rearrangement mechanisms or following charge transfer from the carboxylate anion to various sites by abstraction of a bis-allylic proton on each EET.
of of
Additionally, each EET isomer loses the elements of water to form an ion at m/z 301 and loses 62 D (presumably loss of water as well as CO,) to form an ion at m/z 257. Specifically, the CID product ions from 14,15-EET (Fig. 1A) yielded decomposition ions at m/z 219 (loss of the neutral aldehyde with a proton rearrangement), m/z 175 (loss of the neutral aldehyde as well as CO, with a proton rearrangement), and m/z 113 (an enolate anion fragment). The two possible mechanisms mentioned above leading to m/z 219 are illustrated in Scheme I. These same mechanisms can account for the ion at m/z 179 in 11,12-EET (Fig. 1B) and that at m/z 139 in 8,9EET (Fig. 1C). For the 11,12-EET (Fig. lB), the most abundant CID product ion (m/z 167), produced following decomposition of m/z 319 could arise from a rearrangement similar to that shown in Scheme I, resulting in cleavage of the lo,11 bond with delocalization of the anionic site by the allylic olefin. A similar ion was observed for 8,9EET at m/z 127. The ion at m/z 208 from 11,12-EET is rather curious since it is the only abundant odd electron
206
BERNSTROM,
KAYGANICH,
AND
SCHEME
ion species observed in any EET studied. It was therefore rather diagnostic of 11,12-EET. A homolytic cleavage mechanism of the epoxide ring followed by homolytic cleavage of the 12,13 carbon-carbon bond would account for the d&tonic ion structure shown in Fig. 1B. Less abundant ions in this series were observed at m/z 248 and 168 for 14,15-EET and 8,9-EET, respectively. The CID of 8,9-EET yielded the most numerous ion fragments of all the EETs, most likely due to the central location of the epoxide ring (Fig. 1C). In addition to the ion structures already presented above, abundant ions were observed at m/z 151, 155, 163, and 167; m/z 151 appears as a member of the ion series which is most abundant in 5,6-EET at m/z 191 presented below. The ions at m/z 155 and 167 may result from a homolytic epoxide cleavage reaction shown in Scheme II. The mechanisms suggested in Scheme II are charge remote processes similar to that suggested in Scheme I. The collisional activation of m/z 319 from 5,6-EET yielded quite abundant fragment ions. It is possible that the negative charge when localized at the carboxylate oxygen can attack the 5,6-epoxide in a favorable 6membered ring transition state, resulting in formation of the lactone alkoxide anion seen in Scheme III. This anion can rearrange to form the stabilized allylic anion at m/z 191. The collisional events in the rf-only quadrupole provide energy for this process, which is likely endothermic, as shown in Scheme III. The alkoxide anion could also change the charge site by an intramolecular proton transfer to yield a bis-allylic anion with the
MURPHY
I
charge site localized at Cl3 as illustrated in Structure A. Through further charge migrations and proton rearrangement this ion could decompose to the diene anion (m/z 137) or the triene anion (m/z 163). These structures are illustrated in Fig. 1D. Analysis of phospholipids containing esterified EETs. Incubation of cultured bone marrow-derived mast cells with radiolabeled and cold EETs resulted in the incorporation of EETs into various phospholipid classes as revealed by normal-phase HPLC separation of the lipid classes and scintillation counting of fractions (Fig. 2A). These results are summarized in Table
8.9.EET
m/z 167
ml?. 319
SCHEME
II
EET-PHOSPHOLIPID
ANALYSIS
BY
0 o
00
/ T/n+ \
\
m/z
SCHEME
191
III
1. Normal-phase HPLC revealed modest separation of the radiolabeled molecular species from the major components of glycerophosphoethanolamine (GPE) and glycerophosphocholine (GPC), typically eluting a few minutes after the majority of the phospholipid in each class (Fig. 2A). Reverse-phase HPLC separation of the radiolabeled GPE and GPC fractions resulted in the elution of several unique radiolabeled EET-phospholipid molecular species. These were well separated from other major fatty acyl molecular species (Fig. 2B). The methyl esters of the free fatty acids from isolated EET-phosphatidylcholines obtained by treatment with PLA, were analyzed using GC/MS. A significant amount of EET was seen, but other fatty acids were also seen. The results indicated that EET was present in the sn-2 position, but did not exclude the possibility of sn-1 esterification of EET. Designations of the carbon atoms of the glycerol portion of glycerophospholipids as sn-1 or sn-2 are shown in Fig. 3 (see legend). The additional PLA, products, 2-lysophosphatidylcholines, were analyzed by FAB mass spectrometry to determine the ratio of EET in the sn-1 position to that in the sn-2 position. The results indicated that approximately 80% of the EET was incorporated into the sn-2 position and 20% into the sn-1 position (on the basis of four different molecular species). The collision-induced dissociation of purified EETphospholipid molecular species yielded abundant fragment ions characteristic of both the phospholipid class and the unique EET acyl portion of the molecule. For example, analysis of one EET-glycerophosphocholine molecular species obtained following incubation of 11,12-EET with mast cells resulted in ions corresponding to M-15, M-60, and M-86 ions observed at m/z 782, 737, and 711 (data not shown), in agreement with previously reported behavior of glycerophosphocholine molecular species by negative-ion FAB mass spectrometry
TANDEM
MASS
207
SPECTROMETRY
(15). Abundant fatty acyl ions were observed at m/z 255 (hexadecanoate) and m/z 319 (EET). Collision-induced dissociation of the M-15 (m/z 782) from this GPC molecular species (16:Oa/EET-GPC) yielded abundant carboxylate anions, indicating the esterification of both fatty acyl groups (Fig. 3A) as previously described (14-16). Unlike previously reported CID spectra of M-15 from GPC, numerous fragment ions were also observed at masses lower than the expected acyl carboxylate anions, including m/z 167,179, and 208. These are the identical ions observed in the CID of m/z 319 from 11,12-EET (Fig. 1B). In this experiment, m/z 319 was produced in the second quadrupole collision cell from the glycerophospholipid ion at m/z 782 (M-15) and most likely underwent further collisional activation and decomposition to yield the characteristic ions discussed above for 11,12-EET. Such events of secondary collision activation have been previously described for triple quadrupole instruments (17). Identical behavior was observed for other phospholipid classes and subclasses containing 11,12-EET, as illustrated in Fig. 3B for 16:Op/ll,l2-EET-GPE. In this example, decomposition of m/z 738 ([M-H]) yielded the fatty acyl ion from the sn-2 substituent (EET, m/z 319), loss of the sn-2 (R,) substituent as ketene (-R,=C=O, m/z 436), and a neutral loss of the sn-2 carboxylic acid (-R,COOH, m/z 418). Again, abundant ions resulting from further collision-induced decomposition of the EET carboxylate anion were observed at m/z 167,179, and 208. The decomposition of the EET carboxylate anion (m/z 319) from isomeric GPE molecular species resulted in the same ion products characteristic of the EET isomer, as illustrated in Figs. 4A-4C. These phospholipid molecular species (16:Op/EET-GPE) were isolated from mast cells incubated with 14,15-EET, 8,9EET, and 5,6-EET, respectively. Decomposition of the intact phospholipid anion (m/z 738) thus provides specific information on the position of the epoxide ring on the arachidonate backbone. As seen in Fig. 4C, the 5,6EET-specific ions observed in the CID of m/z 319 from 5,6-EET (Fig. 1D) are produced at m/z 191, 163, and 137. Even the suspected lactone anion at m/z 99 is observed. Since this product ion scan came from decompo-
m/z319
STRUCTURE
A
208
BERNSTROM,
KAYGANICH,
AND
A
MURPHY
B
1500 1000
D s
E f B
z: 9
500
0
10
20 ELUTION
30 TIME,
40
3 /
MIN
ELUTION
TIME,
MIN
FIG. 2. (A) Normal-phase HPLC separation of phospholipids extracted from bone marrow-derived mast cells incubated with [11,12‘H]EET and cold 11,12-EET for 2 h. The HPLC chromatograph was monitored with uv absorption at 206 nm. Fractions were taken in 2-min intervals. Aliquots of each fraction were analyzed for radioactivity content expressed in dpm. (B) Reverse-phase HPLC of the GPE fraction of phospholipids purified by the normal-phase HPLC above. One-minute fractions were collected from the HPLC effluent and an aliquot was taken for determination of the radioactivity content. The HPLC-purified phospholipid molecular species indicated (1,2,3,4,5) were then subjected to fast atom bombardment mass spectrometric analysis.
sition of m/z 738 in all cases and m/z 319 is a prominent anion produced, it is not known with this experiment whether the fragment ions are formed directly from the [M-H]of the phospholipid or proceed from a carboxylate anion intermediate. The pressure of the collision gas in the rf-only quadrupole was sufficiently high for multiple collisions to take place. Under these conditions, fragment ions formed by CID of the parent ion may undergo collisions and fragment further (17). Over 20 molecular species of phospholipids that differ in polar head group and substitution at sn-1 have been studied and all have been found to produce CID product ions characteristic of the specific regioisomer of EET.
TABLE Incorporation of Two pholipids of Mast Cells Distribution of [3HJEETs
1
[3H]EET Isomers into Glycerophosfollowing 60 min of Incubation and into Glycerophospholipid Classes
14,15-EET Phospholipid class GPE GPI GPS GPC
% Added” 8.9 7.5 0.8 13
+ 2 + 3 f 0.2 &4
11,12-EET
% Incorporatedb 22 14 1.7 25
+ 7’ *4 * 0.4 +4
% Added
% Incorporated
15.@ 6.1 1.2 14.4
7.0 14.3 2.6 34.2
D Percentage of initially added [3H]EET recovered in phospholipid class. * Percentage of total phospholipid-associated radioactivity in phospholipid class. ’ Mean of four separate incubations, &EM. d Average of two separate incubations.
DISCUSSION
The fragmentation of even electron organic negative ions such as carboxylate anions is somewhat less well understood than the ion chemistry of positive ions typically encountered in electron impact ionization of molecules such as EET. The general decomposition behavior of negative ion species has been studied in some detail (18); however, the behavior of unsaturated epoxy acids has not been reported prior to this study. We have postulated that the decomposition of carboxylate anions from EETs (m/z 319) is dominated by reactions that either occur through homolytic cleavage of the epoxide moiety or are initiated following charge site transfer mechanisms involving the attack of the carboxylate anion upon bis-allylic protons forming stabilized anion sites. It is of interest to note that these reactions could be induced by the low-energy collisions that occur in the triple quadrupole mass spectrometer. Even though charge-remote fragmentation was initially observed under high-energy-collision conditions (19), it has been demonstrated that the energy required to induce charge-remote fragmentations varies widely depending on compound type, method of ionization, mass of collision target, and target gas pressure (17). Additionally, Wysocki and Ross (17) suggest that the compound dependence of charge-remote fragmentation processes is due to competition between charge-remote fragmentations and charge-driven reactions. Competitive fragmentation processes may be operative in compounds such as the EETs, which have acidic bis-allylic protons, capable of intramolecular transfer to the carboxyl moiety. Alternatively, some ions may be formed by reac-
EET-PHOSPHOLIPID
ANALYSIS
BY
TANDEM
MASS
209
SPECTROMETRY
M-H-
:
M-15
2 % d : ‘Z s ct
100
200
300
400
500
600
700
800
M/Z
436 _
167 179
46
-RjCOOH
-I
I 208
loo
200
41s
300
400
500
600
700
M/Z
FIG. 3. Tandem mass spectrometry of phospholipid molecular species purified by two steps of HPLC ment. R2 corresponds to a C,,H,,,O moiety from the EET esterified at sn-2 lost as ketene (R2=C=0) carbon atoms designated by convention as an-1 and m-2 are correspondingly numbered in each structure. glycerophosphocholine molecular species following incubation of 11,12-EET with murine mast cells. (M-15, m/z 782) was subjected to collision-induced dissociation, resulting in the formation of abundant and the decomposition ions of 11,12-EET. This molecular species was identified as l&Oa/ll,l2-EET-GPC. a major glycerolphosphoethanolamine molecular species obtained in the same 11,12-EET incubation. molecular anion at m/z 738 resulted in the formation of one acyl ion at m/z 319 as well as EET-specific
tion mechanisms in which an ion complex is formed with either a hydride ion or a hydroxyl ion followed by intramolecular reaction and decomposition of the molecule (18). Such a mechanism may be operating in the formation of the conjugated species at m/z 137 and 103 in 8,9-EET and 5,6-EET. While the suggested reaction mechanisms are consistent with known ion chemistry of negative ions, further experiments with stable isotopelabeled EETs would be required to confirm these suggested processes. Regardless of the exact mechanisms of these fragmentations, the mass spectra observed for each of the four isomeric EETs uniquely identify the epoxide position along the eicosanoid backbone. Collision-induced dissociation of the anions derived from phosphatidylcholine and phosphatidylethanolamine revealed esterification of EETs at the glycerol backbone through the appearance of an ion at m/z 319. In the CID spectra of the [M-H]and [M-CHJions from all of the diacylphospholipids studied, m/z 319 was less abundant than the other carboxylate anion. This would indicate that EET was esterified at sn-I, as the abundance of the sn-2 carboxylate anion is normally greater than that of the sn-1 carboxylate anion in the MS/MS experiment of these precursor ions (14-16). However, exceptions to the ion abundance of the sn-2 > sn-1 anion relationship have been reported for molecular species having acyl groups with extensive unsaturation and/or a substantial difference between the chain lengths of the two acyl groups (20). The presence of sn-l/sn-2 positional isomers of these EET-phospholipids, as revealed by the results of the PLA, hydrolysis
I
16:Opi 11,12-EET-GPE
-R2=C=0
and ionized by fast atom bombardor a carboxylic acid (R&OOH). The (A) Analysis of a major radioactive The major negative ion at high mass carboxylate anions (m/z 255 and 319) (B) Tandem mass spectrometry of Collision-induced dissociation of the ions from 16:Op/ll,lZ-EET-GPE.
experiments, further complicates the simple interpretation of the sn-1 and sn-2 substituents from the relative abundances of the carboxylate anions in the CID spectra. Until pure standards of these EET-phospholipids are available so that the CID behavior of the pure compounds can be characterized, the interpretation of these CID spectra regarding sn-1 and sn-2 substitution remains uncertain. CID analysis of the M-15 anion from GPC as well as the [M-H]anion from GPE molecular species reveals ions characteristic of the position of the epoxy group of the esterified EET in addition to the sn-1 fatty acyl group also esterified on the glycerolipid backbone. Thus, it is possible to uniquely determine not only the presence of EET in the glycerophospholipid molecular species by tandem mass spectrometry, but also the exact isomer of EET (5,6-, 8,9-, 11,12-, or 14,15-EET) that is present in the oxidized phospholipid molecular species. Such information has been impossible to obtain except by following extensive decomposition of the phospholipid followed by separate analysis of the various products. The use of the methodology described here for the direct analysis of EET-phospholipids minimizes the probability of sample autooxidation and decomposition as is likely to occur when conventional means are used to analyze these labile compounds. Indeed, the efficacy of this methodology is demonstrated by the characterization of several intact 5,6-EET-phospholipid species. The detection of 5,6-EET from in uivo sources has not been demonstrated before because of its facile hydrolysis to the 5,6-diol during conventional analytical
210
BERNSTROM,
KAYGANICH,
AND
MURPHY
‘001
I
436
I
M/Z
M/Z
83
191
1
eb A
436
I
191
[M-HI’ 16:0p/5,6-EET
319
GPE 73ll
X
x4I ‘Z zCL
137
M/Z
FIG. 4. Tandem mass spectrometry of the 16:0 plasmalogen EET-containing molecular species of glycerophosphoethanolamine purified by HPLC and analyzed by fast atom bombardment mass spectrometry following incubation with isomeric EETs. I!& corresponds to a CleHsoO moiety from the EET esterified at sn-2 lost as ketene (b=C=O) or a carboxylic acid (R&OOH). (A) 160 Plasmalogen obtained following incubation of 14,15-EET with mast cells. (B) Plasmalogen obtained following incubation of 8,9-EET with mast cells. (C) Plasmalogen obtained following incubation of 5,6-EET with mast cells.
workup. The analytical capability of negative FAB tandem mass spectrometry should enable a more detailed investigation of the endogenous production of EETs and their metabolic fate in terms of reincorporation into phospholipids. From such information it should be possible to better understand the P-450-mediated metabolism of arachidonic acid and the formation of the epoxygenase-derived eicosanoids as well as the role that they may play in various physiological and pathophysiologic processes. ACKNOWLEDGMENTS This work was supported, in part, by a grant from the National Institutes of Health (GM41026). The authors acknowledge helpful discussions with Dr. Veronica Bierbaum (University of Colorado) concerning the negative ion chemistry of these molecules.
REFERENCES 1. Fitzpatrick, 229-241.
F. A., and Murphy,
R. C. (1989)
Pharmacol. Rev. 40,
2. Capdevila, J., Pramanik, B., Napoli, J., Manna, S., and Falck, J. (1984) Arch. Biochem. Biophys. 231.511-517. A., Dishman, E., Blair, I., Falck, J., and Capdevila, J. 3. Karara, (1989) J. Biol. Chem. 264, 19,822-19,827. V., Jacobson, H., Siddhanta, A., Pramanik, 4. Falck, J., Schueler, B., and Capdevila, J. (1987) J. Lipid Res. 28, 640-646. A., Manna, S., Pramanik, B., Falck, J., and 5. Toto, R., Siddhanta, Capdevilla, J. (1987) Biochim. Biophys. Actu 919, 132-139. 6. Ballou, L., Lam, B., Wong, P.-K., and Cheung, W. Y. (1987) Proc. Natl. Acad. Sci. USA 84,6690-6694. 7. Capdevila,
B&hem. 8. Harris, (1990)
J. H., Kishore,
V., Dishman,
E., and Blair,
I. A. (1987)
Biophys. Res. Commun. 146,638-644. R. C., Homma,
T., Jacobson,
J. Cell Physiol. 144,429-437.
H. R., and
Capdevila,
J.
EET-PHOSPHOLIPID
ANALYSIS
9. Falck, J. R., Yadagiri, P., and Capdevila, J. (1990) Enzymology, p. 357, Academic Press, New York. 10. Rivnay,
B. (1984)
11. Patton,
G. M., Fasulo,
J. Chromatogr.
in Methods
BY in
S. J. (1982)
Res.
12. Pierce, J. H., DiFiore, P. D., Aaronson, S. A., Potter, Pumphrey, J., Scott, A., and Ihle, I. N. (1985) Cell 41,685. 13. Bligh, E. G., and Dyer, W. J. (1959) Con. J. Biochem. Physiol. 911-917.
M.,
23,190-196.
R. C. (1991)
15. Jensen,
N. J., Tomer,
16. Munster, J. Lipid
14. Kayganich, K., and Murphy trom. 2.45-54.
MASS
211
SPECTROMETRY K. B., and Gross,
M. L. (1986)
Lipids
21,
580-588.
294,303-315.
J. M., and Robins,
TANDEM
J. Am. Sot. Mass
37, Spec-
H., and
Budzikiewicz,
H. (1988)
Biol.
Chem.
Hoppe-
Seykr369,303-308. 17. Wysocki, zon Proc.
V. H., and Ross, 104,179-211.
M.
M.
(1991)
Int.
J. Muss
Spectrom.
18. Bowie, J. H. (1990) Mass Spectrom. Reu. 9, 349-379. 19. Jensen, N. J., Tomer, K. B., and Gross, M. L. (1985) J. Am. Chem. Sot. 10'7.1863-1868. 20. Huang, Z.-H., Gage, D. A., and Sweeley, Mass Spectrom., in press.
C. C. (1991)
J. Am.
Sot.
