- Email: [email protected]
Hepoxilin A3 is oxidized by human neutrophils into its ω-hydroxy metabolite by an activity independent of LTB4 ω-hydroxylase1
Biochimica et Biophysica Acta 1348 Ž1997. 287–298
Hepoxilin A 3 is oxidized by human neutrophils into its v-hydroxy metabolite by an activity independent of LTB 4 v-hydroxylase 1 Denis Reynaud a , Olga Rounova a , Peter M. Demin a,b, Kazimir K. Pivnitsky b, Cecil R. Pace-Asciak a,c,) a b
Research Institute, Hospital For Sick Children, DiÕision of Neurosciences, 555 UniÕersity AÕenue, Toronto, Ont., Canada M5G 1X8 Institute of Experimental Endocrinology of National Endocrinology Scientific Center, National Academy of Medical Sciences, Moscow 115478, Russia c Department of Pharmacology, Faculty of Medicine, UniÕersity of Toronto, Toronto, Ont., Canada M5S 1A8 Received 13 February 1997; revised 14 April 1997; accepted 21 April 1997
Abstract Hepoxilin A 3-methyl ester is taken up by intact human neutrophils where it is first hydrolyzed into the free acid which is subsequently converted into a single major metabolite. The structure of this metabolite was determined through mass spectral analysis of several derivatives, and through identity with an authentic compound prepared by chemical synthesis. The metabolite was identified as v-hydroxy-hepoxilin A 3 showing that the epoxide functionality of the parent hepoxilin is not opened during incubation with human neutrophils. All attempts to investigate hepoxilin metabolism in broken cells, despite the presence of protease inhibitors ŽAproteinin, PMSF, DFP. and supplementation with NADPH were unsuccessful. Metabolism of hepoxilin A 3 required the intact cell, while parallel experiments with LTB 4 as substrate demonstrated that this eicosanoid was metabolized into its v-hydroxy metabolite regardless of whether intact or broken cell preparations were used provided that NADPH was present in the latter. Hepoxilin metabolism in intact cells was inhibited dose-dependently by CCCP Ž0.01–100 m M., a mitochondrial uncoupler, whereas LTB 4 metabolism was unaffected by CCCP. This data suggests that metabolism of hepoxilin A 3 occurs in intact human neutrophils through v-oxidation, is likely located in the mitochondrial compartment of the cell Žinhibition by CCCP. and is carried out by an activity that is independent of the well characterized, relatively stable microsomal LTB 4 v-hydroxylase. q 1997 Elsevier Science B.V. Keywords: Hepoxilin A 3; Metabolism; v-Oxidation; v-Hydroxy-hepoxilin A 3
1. Introduction Abbreviations: Me, methyl; FA, free acid; HxA 3 Ž8S., hepoxilin A 3 , 8ŽS.-hydroxy-11ŽS., 12ŽS.-trans-epoxyeicosa-5Z, 9E, 14Z-trienoic acid; HxB 3 Ž8S., hepoxilin B 3 , 10ŽS.-hydroxy-11ŽS., 12ŽS.-trans-epoxyeicosa-5Z, 8Z, 14Z-trienoic acid; RPMI 1640, Roswell Park Memorial Institute medium 1640; CCCP, carbonyl cyanide m-chlorophenylhydrazone; PMSF, phenylmethylsulfonylfluoride; DFP, diisopropylfluorophosphate; GCrMS, gas chromatographyrmass spectrometry ) Corresponding author. Fax: q1 Ž416. 813 5086; E-mail: [email protected]
Hepoxilins are derived through an intramolecular rearrangement of 12ŽS.-HPETE w1,2x. Although this process can take place nonenzymatically yielding two
1
The financial support of the MRC ŽMT-4181 to CRP-A. is gratefully acknowledged.
0005-2760r97r$17.00 q 1997 Elsevier Science B.V. All rights reserved. PII S 0 0 0 5 - 2 7 6 0 Ž 9 7 . 0 0 0 6 4 - 7
288
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
isomers, hepoxilin A 3 and B 3 , we have recently provided evidence of an enzymatic rearrangement in which mostly hepoxilin A 3 is formed w3x. Additionally, we have shown that the enzymatic process utilizes only 12Ž S. -HPETE as substrate for hepoxilin A 3 formation, the 12ŽR.-enantiomer is not selected in this transformation w3x. Two pathways of metabolism of hepoxilin A 3 have previously been reported. One involving an epoxide hydrolase resulting in epoxide ring opening to yield a biologically inactive trihydroxy metabolite w4x; the second pathway involves glutathione addition to the epoxide ring resulting in a biologically active peptido-hepoxilin w5,6x. Hepoxilins have been shown to possess a diversity of biological activities related to their actions on the mobilization of calcium and potassium ions in the cell Žfor recent reviews, see w7,8x.. We have recently shown the existence of a putative hepoxilin receptor in human neutrophils which binds specifically the hydroxy epoxide functional group as similar metabolites, e.g. 12-HETE, leukotriene B 4 , and various epoxy-containing metabolites of arachidonic acid, do not compete for the specific binding of hepoxilin A 3 w9–11x. Since this putative receptor is located inside the cell, it was necessary to employ the methyl ester of hepoxilin to demonstrate binding and calcium actions in the intact cell as the free acid of hepoxilin is inactive since it does not permeate the cell w10x. The free acid binds only when the cell is disrupted w9x. In this paper we provide evidence for the uptake in intact human neutrophils of the methyl ester of hepoxilin A 3 and its intracellular hydrolysis into the free acid which is then metabolized via a new pathway involving the oxidation of the terminal end of the compound to yield a v-hydroxy-hepoxilin A 3 in which the epoxide group is retained. Evidence is presented dissociating this hepoxilin metabolic activity from LTB 4 v-hydroxylase w12–14x. 2. Materials and methods
Žsilica gel G, ethyl acetateracetic acid 99r1, vrv. . w3 H 8 x-LTB 4 Žspec. act. 218 Cirmmol. was purchased from Amersham Ž Toronto. , and unlabeled LTB 4 was from Cayman Chemical Ž Ann Arbor. . RPMI 1640 medium and all chemicals for buffers were reagent grade and were purchased from Sigma, USA. Dextran T-500 and Ficoll–Paque were purchased from Pharmacia, Sweden. CCCP was purchased from Calbiochem, USA. 2.2. Preparation of human neutrophils Human neutrophils were prepared according to w16x with slight modifications according to Boyum ¨ Dho et al. w17x. Forty ml of venous blood was collected from normal drug-free human volunteers and anticoagulated with heparin sodium Ž Organon Tecknica Co., Durham, NC. . Erythrocytes were removed by 4.5% dextran sedimentation for 45 min. The neutrophil-rich plasma was centrifuged at 1500 rpm for 5 min. The resulting pellet was gently suspended in phosphate buffered saline whose composition was in mM: NaCl 140, NaH 2 PO4 10, Na 2 HPO4 10, pH 7.2. This suspension was subjected to Ficoll– Paque gradient centrifugation at 2600 rpm for 20 min. The entire supernatant was aspirated and discarded. The pellet containing the neutrophils was resuspended in 0.5 ml of RPMI 1640 Žfree of sodium bicarbonate but supplemented with L-glutamine and 25 mM HEPES buffered to pH 7.2. and the contaminating erythrocytes were eliminated by lysis in 10 ml of 0.85% ammonium chloride containing 17 mM Tris, pH 7.2, during 15 min at 378C. The suspension was centrifuged at 900 rpm for 5 min. The neutrophils were then resuspended in 5 ml of RPMI 1640 and allowed to re-equilibrate at 378C for 15 min. After an additional centrifugation at 900 rpm for 5 min, the cells were resuspended in RPMI 1640 medium and were counted in a Coulter counter ŽModel 901.. The cells were finally adjusted to a concentration of 10 7 cellsrml in RPMI 1640 medium. 2.3. Metabolism of HxA 3 by human neutrophils
2.1. Materials 3
8ŽS. HxA 3-methyl ester and w H 6 x-HxA 3 Ž 8S. -Me Žspec. act. 169 Cirmmol. were prepared in our laboratory by total chemical synthesis w15x. The purity of the compounds was established periodically by TLC
Freshly prepared human neutrophils Ž 100 = 10 6 cells. were incubated in 2 ml of Buffer A in a siliconized glass tube to which had been added w3 HxHxA 3 methyl ester Ž5 = 10 6 cpm. diluted with 30 m g of the unlabeled compound. The composition of
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
Buffer A consisted of the following in mM: NaCl 140, KCl 5, MgCl 2 1, CaCl 2 1, HEPES sodium-free 10, and glucose 10, pH 7.3. The mixture was incubated for various times, but for preparative runs, it was incubated for 60 min. The incubation mixture was then split in half; one-half was extracted with ethyl acetate without prior acidification to extract the intact hepoxilin structure, if so present in the incubation mixture, and its products. The other half was acidified to pH 3 to convert any hepoxilin structures present in the incubation mixture to the trioxilin structures through hydrolysis of the epoxide moiety. The trioxilins are considerably more polar on TLC than the hepoxilins. In order to determine whether the original methyl ester group of the substrate was hydrolyzed to the free acid during the incubation with neutrophils, one-half of each of the above extracts was subjected to methylation with diazomethane, and its TLC pattern was compared with the unmethylated half of the sample. Each sample was dissolved in ethyl acetate and spotted on TLC Ž silica gel G, ethyl acetateracetic acid 99.5r0.5, vrv. . The plate was developed for 60 min, then the solvent was dried with an air gun, and scanned for radioactivity on a TLC radiochromatogram scanner ŽBerthold.. Radioactive zones of silica gel were scraped into siliconized glass tubes and the compounds were extracted with ethyl acetatermethanol Ž 90r10, vrv. . In preparative runs, the zones corresponding to the metabolites 1 and 2 were isolated, the compounds were eluted with with methanol and extracted with ethyl acetate after dilution with water. The compounds of interest were recovered in the ethyl acetate phase. In order to confirm that the purified metabolites 1 and 2 contained an intact epoxide ring as in the substrate, a portion of the sample as well as authentic standard HxA 3-methyl ester were hydrolysed in 0.1% phosphoric acid in tetrahyrofuran for 16 h at 238C. The samples were then extracted into ether, washed neutral with water and evaporated to dryness. The reaction with acid converts the epoxide ring to an enantiomeric mixture of vicinal diols. Thus HxA 3 is converted into the triol, TrXA 3. 2.4. DeriÕatization Methyl esters were prepared by reaction with a freshly prepared ether solution of diazomethane for
289
10 min at 238C in the dark. Methyl esters were hydrolysed back to the free acid by reaction with ethanolr1 N KOH Ž1r1, vrv. for 30 min at 238C. The free acid of the compound was extracted into diethyl ether after gentle acidification to pH 6 with 0.01 N HCl. These mild conditions do not open the epoxide moiety of HxA 3. Pentafluorobenzyl esters were prepared from the free acid dissolved in 10 m l methanol and 50 m l acetonitrile by the rapid addition of 2 m l pentafluorobenzyl bromide Ž Pierce. and 1.5 m l of diisopropylethylamine Ž Pierce. . The mixture was heated at 658C for 5 min, followed by evaporation of the solvents and reagents with nitrogen gas in a fume hood. Trimethylsilyl ether derivatives were prepared by reacting the methyl esters or pentafluorobenzyl esters with 20 m l TRI-SIL Z Ž Pierce. for 5 min at 608C. The solvent was subsequently removed with nitrogen gas in a fume hood, the residue was dissolved in 10 m l dodecane, and the sample was placed on dry ice until analysed Ž usually within 60 min.. Hydrogenation of the methyl ester derivative was carried out using platinum oxide Ž Aldrich. as catalyst preactivated in methanol with hydrogen gas. The hydrogenation of the sample was carried out by bubbling hydrogen gas for 1 min. The catalyst was then filtered, and the solvent was taken to dryness with nitrogen gas. The residue was then silylated as described above. The following TLC R f values ŽSolvent system Žsilica gel G.: ethyl acetateracetic acid Ž99.5r0.5, vrv.. for the various derivatives were observed: HxA 3 Žmetabolite 1. , free acid 0.44, Me ester 0.67, PFB ester 0.72; v-hydroxy HxA 3 Žmetabolite 2., free acid 0.33, Me ester 0.42, PFB ester 0.54; TrXA 3 Žcompound 3., free acid 0.19–0.25, Me ester 0.30–0.37, PFB ester 0.40–0.44; v-hydroxy TrXA 3 Žcompound 4., free acid 0.06, Me ester 0.09, PFB ester 0.15–0.20. 2.5. Gas chromatography–mass spectrometry (GC r MS) Samples were analysed by GCrMS using both the electron impact ŽEI. mode and the negative ion chemical ionization Ž NICI. mode on a Hewlett Packard GCrMS ŽModel 5988. with direct on-column injection of sample on a capillary column Ž HP-1, 30 m.. The temperature was programmed from 1708C at injection to 3008C at 108Crmin. Helium was the
290
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
carrier gas and methane was the reactant gas in NICI mode. The following GC retention values Ž R t . were observed: Metabolite 2, MeTMSi 18.1 min, MeTMSi after hydrogenation 18.6 min, MeTMSi after acid opening of the epoxide 18.6 min; chemically synthesized 20-hydroxy HxA 3 MeTMSi 18.1 min, MeTMSi after hydrogenation 18.6 min, MeTMSi after acid opening of the epoxide 18.6 min.
knowing that the allylic epoxide of HxA 3 is quite unstable to acidic workup, and that the methyl ester of HxA 3 is relatively stable to normal handling Ž e.g.
2.6. Comparison of hepoxilin A 3 metabolism with LTB4 metabolism A series of experiments were carried out in parallel incubations of human neutrophil broken preparations with w3 H 6 xhepoxilin A 3 methyl ester or free acid and with w3 H 8 x-LTB 4 free acid or methyl ester. Freshly prepared neutrophils were either sonicated Ž 3 = 5 s; sonic dismembrator, Systems Corp., Farmingdale, NY., or freeze-thawed Žsequence of freezing on dry iceracetone and thawing at 238C repeated three times. , or homogenized then separated into mitochondrial Ž10 000 = g pellet. or microsomal Ž100 000 = g pellet. preparations and incubated in 1 ml of Buffer A in the presence or absence of 2 mM NADPH for various times Ž 0–60 min. . The incubation mixture was extracted with ethyl acetate, and analyzed by TLC as above. In other experiments, the protease inhibitors aproteinin ŽSigma. 0.17 Urml, PMSF ŽSigma. 0.1 mM, and DFP Ž Sigma. 1 mM, were added in separate experiments before cell disruption Žsonication or freeze-thawing. to investigate recovery of hepoxilin or LTB 4 metabolizing activity. Inhibition of hepoxilin A 3 but not LTB 4 metabolism by CCCP. Parallel incubations of intact human neutrophils Žprepared as above. with the two substrates were carried out at different concentrations of CCCP Ž 0–100 m M., during 60 min incubation at 378C. Sample extraction with ethyl acetate and TLC analysis was carried out as above.
3. Results When w3 Hx-HxA 3 methyl ester is incubated with intact human neutrophils, a single major metabolite with increased polarity is observed. We attempted to gain some quick information into its structure by a combination of simple techniques involving TLC,
Fig. 1. Thin layer radiochromatograms of ethyl acetate extracts of: ŽA. incubation of w 3 Hx-HxA 3-Me Ž5=10 6 cpm containing 30 m g of unlabeled hepoxilin. with intact human neutrophils Ž100= 10 6 cells. for 60 min at 378C in 2 ml of Buffer A; ŽB. a portion of the extract in ŽA. was acidified with 1 N HCl; ŽC. a portion of the extract in ŽA. was converted into the methyl ester form with diazomethane in diethyl ether; ŽD. a portion of the extract in ŽB. was subjected to methylation with diazomethane. TLC was performed on silica gel G plates which were developed with ethyl acetateracetic acid Ž99.5r0.5, vrv.. Note that in ŽA. two metabolites are formed; metabolite 1 was identified as the free acid form of the substrate HxA 3; metabolite 2 was identified as the free acid of v-hydroxy HxA 3. Note that very little, if any, of the epoxide hydrolase metabolite of HxA 3 , i.e. trioxilin A 3 Žcompound 3., is formed during the 60-min incubation demonstrating the virtual absence of epoxide hydrolase activity in human neutrophils. The epoxide ring opening of both metabolites is observed only when the metabolites are treated with acid to generate compounds 3 and 4 identified as trioxilin A 3 and v-hydroxy trioxilin A 3 , respectively Žfor mass spectra see Figs. 6–8..
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
extraction and TLC. . We therefore divided the incubation mixture into four parts. One part was extracted with ethyl acetate without acidification to ensure that the epoxide group, if present, would be extracted intact Ž Fig. 1A. ; a second part was acidified to pH 3 with HCl, a procedure which results in the hydrolysis of the epoxide to a more polar vicinal diol Že.g. trioxilin. ŽFig. 1B. . A third part was treated with an ether solution of diazomethane to form the methyl ester derivative Ž Fig. 1C.. Comparison of this chromatogram with that of the first part Ž Fig. 1A. would give information as to whether the carboxyl group of the metabolites was retained in the methyl ester form after incubation, or whether it had been hydrolysed into the free acid form. A fourth part of the original extract was acidified then subsequently methylated ŽFig. 1D., again to indicate which metabolites Ž in Fig. 1B. had a free carboxylic acid group. TLC analysis of these four extracts was quite informative as shown in Fig. 1. It is important to note Žsee Fig. 1A. that the methyl ester substrate is completely converted by the cells into a minor Ž metabolite 1. and a major product Žmetabolite 2.. That the epoxide group in the substrate was not opened during incubation was determined from the methylation of this extract ŽFig. 1C.. This panel shows that metabolite 1 reverts to the migration of the substrate, while metabolite 2 is clearly distinct from this. When subjected to methylation with diazomethane, metabolites 1 and 2 migrated upwards in the plate indicating that both metabolites 1 and 2 became methylated; hence both metabolites were present at the end of the incubation as free acids and were extracted as free acids. The hydrolysis of the methyl ester of HxA 3 by the cells is rationalised by the presence of esterases in the cell cytosol. It is also important to note that the epoxide is not opened into a diol during incubation with the cells, i.e. absence of significant epoxide hydrolases in the neutrophils, since very polar material corresponding to trioxilins is not present in appreciable amounts in the unacidified incubation mixture ŽFig. 1A.. That both metabolites Ž1 and 2. contained an intact epoxide was also shown by the fact that acid hydrolysis caused both metabolites to migrate with greater polarity Žcompare Fig. 1A with 1B, or Fig. 1C with 1D.. Metabolite 1 formed compound 3 upon acidification which migrated like trioxilin A 3 Ž TrXA 3 . , while metabolite 2 was converted
291
into compound 4 which was more polar than trioxilin A 3. Again both compounds 3 and 4 had free carboxylic acid groups since both reacted with diazomethane to form methyl esters Žcompare Fig. 1B and 1D. . We investigated the time course of the conversion of HxA 3-Me into metabolite 2. Fig. 2 shows that the methyl ester substrate is first hydrolysed into the free acid Ž metabolite 1. which is subsequently converted into metabolite 2. We have observed differences between subjects in this conversion although we have not yet carefully screened a large number of subjects to better study this potentially interesting observation. Attempts to modify this conversion through the addition of cofactors ŽATP, NAD, NADH, NADP or NADPH. were mostly unsuccessful as formation of metabolite 2 required the intact cell. All attempts to carry out the same metabolism with membranes from disrupted cells, as well as cell cytosol were unsuccessful. Also unsuccessful in recovering hepoxilin metabolizing activity during cell disruption was the use of the protease inhibitors, aproteinin, PMSF and DFP Ždata not shown.. In contrast, LTB 4 was metab-
Fig. 2. Time course of the metabolism of w 3 Hx-HxA 3-Me by intact human neutrophils. w 3 Hx-HxA 3-Me Ž0.55=10 6 cpm containing 5 m g of unlabeled hepoxilin. was incubated for times indicated with intact human neutrophils Ž16=10 6 cells. in 1 ml of Buffer A as in Fig. 1A at 378C. At the end of the incubation, the mixture was extracted with ethyl acetate, and this extract was evaporated with N2 gas and the extract was dissolved in a small amount of ethyl acetate and spotted on TLC as in Fig. 1. FA s free acid.
292
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
olized irrespective of whether the neutrophils were intact, or disrupted through sonication or freeze-thawing as long as NADPH was present during incubation of the disrupted cells. Fig. 3 compares the metabolism of hepoxilin A 3 and LTB 4 by cells subjected to freeze-thawing ŽFig. 3, panel A. and intact and sonicated cells ŽFig. 3, panel B.. As shown, hepoxilin metabolism disappears when the cell is disrupted, while LTB 4 metabolism is still relatively intact. That hepoxilin as well as LTB 4 metabolism was enzymatic was shown by the absence of any conversion of either substrate with cells that were placed in boiling water for 5 min Ždata not shown.. Use of the free acid of hepoxilin A 3 with disrupted cells did not result in any conversion into products. Conversely, LTB 4 was
v-oxidized by the intact or disrupted cell, irrespective of whether it was used in the free acid or methyl ester form. Additional experiments were carried out to dissociate hepoxilin metabolism from that which oxidizes LTB 4 . We discovered that CCCP, a mitochondrial uncoupler, dose-dependently inhibits the metabolism of HxA 3 , while LTB 4 is unaffected by the drug. Fig. 4 shows TLC patterns of extracts of incubations analyzing HxA 3 and LTB 4 metabolism in the presence of CCCP, 5 m M for HxA 3 and 60 m M for LTB 4 substrates. Fig. 5 shows the dose-dependency of the selective inhibition by CCCP of HxA 3 metabolism into the v-hydroxy metabolite, with essentially unaffecting LTB 4 v-oxidation. These data
Fig. 3. Radio TLC profiles comparing the effect of cell disruption — ŽA. freeze-thaw method, ŽB. sonication — of human neutrophils in their ability to metabolize hepoxilin A 3 and LTB 4 . For methods see legend to Fig. 2. Incubation time was 60 min.
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
suggest that the activity that oxidizes hepoxilin A 3 is different from that which oxidizes LTB 4 . Additionally, the data with CCCP suggest that hepoxilin metabolism occurs in the mitochondria, whereas LTB 4 metabolism occurs in the microsomal fraction of a cell homogenate. Since hepoxilin metabolizing activity was lost during cell disruption, direct evidence of the localization of hepoxilin metabolism to mitochondria could not be obtained but is suggested by its inhibition by CCCP. Metabolite 1 was identified as the free acid of HxA 3 by GCrMS, and compound 3 was the corresponding trioxilin. Metabolite 2 was subjected to structural analysis by GCrMS in two ionization modes. First, the free acid was isolated by TLC, converted into the PFB ester and analyzed as the TMSi ether derivative using the NICI mode. In this ionization mode, a prominent ion results from the loss of the PFB group essentially giving important information on the molecular weight of the compound being analysed. While the parent HxA 3 showed a major fragment ion at mrz 407 ŽM-PFB., in metabolite 2 this fragment ion shifted to mrz 495, corresponding to the presence in metabolite 2 of an additional hydroxyl group. Similar information was
293
Fig. 5. Concentration-dependent inhibition by CCCP of hepoxilin A 3 v-oxidation but not LTB 4 v-oxidation in intact human neutrophils. For methods see legend to Fig. 2. Incubation time was 60 min.
obtained when the acidified HxA 3 Ž which forms trioxilin A 3 , a trihydroxy derivative. was subjected to NICI-GCrMS. Its spectrum showed a fragment ion at mrz 569 Ž M-PFB. , while that of the corresponding derivative of metabolite 2, i.e. compound 4, showed a
Fig. 4. Inhibition of hepoxilin but not LTB 4 v-oxidation in intact human neutrophils by CCCP, a mitochondrial uncoupler. Note that CCCP does not interfere with the hydrolysis of the methyl ester of hepoxilin to the corresponding free acid, only with the v-oxidation step. For methods see legend to Fig. 2. Incubation time was 60 min.
294
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
fragment ion at mrz 657, confirming that metabolite 2 had an extra hydroxyl group. As expected metabolite 2 was more polar than the parent HxA 3 due to the presence in the former of an additional hydroxyl group Žsee Fig. 1. . In order to obtain more information on the location of the additional hydroxyl group in metabolite 2, it was subjected to EI-GCrMS ŽFig. 6B. as the MeTMSi derivative. Data is shown in comparison to that obtained from the parent HxA 3 ŽFig. 6A. . Fig. 6A shows the mass spectrum of HxA 3 with a prominent
fragmentation at the C7–C8 bond Ž M-141, C1–C7. giving rise to a fragment ion at mrz 281 ŽC8–C20.. In the corresponding spectrum of metabolite 2 Ž Fig. 6B., this fragment ion is shifted to mrz 369 representing the presence of an additional hydroxyl group Ž281 q 88. in the C8–C20 portion of the structure of HxA 3. An important fragment of structural significance locating the hydroxyl group at the v-position was the ion at mrz 103 in metabolite 2 Ž Fig. 6B., which was absent in the parent compound which lacks this moiety Ž Fig. 6A. . Localization of the addi-
Fig. 6. GCrMS ŽEI. of the MeTMSi derivatives of ŽA. authentic HxA 3 and ŽB. metabolite 2 identified as v-hydroxy HxA 3. Note the shift of fragment ions indicated in panel B in the relevant portions of the molecule containing the additional hydroxyl group, i.e. mrz 369 Ž281 q 88..
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
tional hydroxyl group of metabolite 2 to the terminal end of the molecule was derived from the mass spectra of the following derivatives. The acidified products derived from HxA 3 Ž i.e. TrXA 3 , compound 3. and metabolite 2 Ži.e. compound 4. showed structurally informative ions at mrz 213 ŽC12–C20. and 301 ŽC12–C20., respectively Ž Fig. 7A and B. . The fragment ion at mrz 243 due to the C1–C8 portion of the molecule is common to both compounds. Minor fragment ions occur in the spectrum of compound 4 ŽFig. 7B. at mrz 657 wM-15x and mrz 492
295
wM-Ž2 = 90.x although the corresponding fragments for compound 3 are almost undetectable due to easier fragmentation along the carbon–carbon bonds mentioned above. Further support of these fragmentations was obtained with the hydrogenated derivatives. Catalytic hydrogenation of HxA 3 led to the opening of the epoxide with insertion of a hydrogen atom at C11 w1x. Mass spectral analysis of this product gave an informative fragment at mrz 215 representing the fragment ion containing the C12–C20 portion of the molecule ŽFig. 8A. . The comparable fragment in the
Fig. 7. GCrMS ŽEI. of the acid-catalysed products obtained from ŽA. HxA 3 , i.e. trioxilin A 3 , and ŽB. compound 4, i.e. tetroxilin A 3 , derived from metabolite 2. Note the shift of fragment ions in ŽB. in the relevant portions of the molecule containing the additional hydroxyl group, i.e. mrz 301 Ž213 q 88..
296
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
mass spectrum of the hydrogenated metabolite 2 was shifted to mrz 303 Ž215 q 88. indicating that the double bond at D14,15 had not been affected in the formation of metabolite 2, and that the additional hydroxyl group must be located at the terminal end of the metabolite ŽFig. 8B.. The common fragment ion at mrz 243 in the original compounds is shifted to mrz 245 in the hydrogenated compounds confirming that the double bond at C5–C6 in the original substrate remained intact in the metabolite. Additional shifts in minor fragments due to wM-15x and wM-90x
are also seen between hydrogenated HxA 3 Ž mrz 487 and 412, respectively. and hydrogenated metabolite 2 Žmrz 575 and 500, respectively. Ž Fig. 8A, B.. The placement of the hydroxyl group at C20 in metabolite 2 is due to the low intensity fragments in the mrz 110–140 region of the spectra as compounds with a hydroxyl group at C19 typically have higher intensity fragment ions in this region Ž see mass spectra of 19and 20-hydroxy prostaglandins in w18x.. This data suggests that metabolite 2 is v-hydroxy-HxA 3 Ž or 20-hydroxy HxA 3 .. Finally, additional proof for the
Fig. 8. GCrMS ŽEI. of the catalytic hydrogenation products ŽPtO 2rmethanol. of ŽA. HxA 3 and ŽB. metabolite 2 identified as v-hydroxy HxA 3. Samples were analysed as the MeTMSi derivatives. As with Fig. 6 and Fig. 7, note the shift of fragment ions in ŽB. in the relevant portions of the molecule containing the additional hydroxyl group, i.e. mrz 303 Ž215 q 88. and 447 Ž359 q 88..
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
v-location of the hydroxyl group was obtained through identity of the mass spectrum and chromatographic properties of the metabolite with an authentic 20-hydroxy HxA 3 prepared by total chemical synthesis w19x.
4. Discussion Previous studies have shown two metabolic pathways for the metabolism of hepoxilins, i.e. by hepoxilin epoxide hydrolase w4x and a glutathione S-transferase w6,20,21x. In this study, we provide evidence that human neutrophils metabolize hepoxilins via a third pathway involving v-oxidation with retention of the hydroxy epoxide functional group of the parent hepoxilin. v-Oxidation of eicosanoids has been shown to take place in human neutrophils. 12-HETE and LTB 4 have been shown to be transformed into their voxidized products w12–14,22,23x. We have also found in this study that human neutrophils metabolize LTB 4 but in contrast to LTB 4 v-oxidation, we found that hepoxilin metabolism Žformation of metabolite 2. is abolished when cells are disrupted, both when the cellular components are retained Ž through freezethawing. or when they are separated Ž through sonica-
297
tion.. Additionally, we discovered that hepoxilin metabolism in intact neutrophils is inhibited by CCCP, a mitochondrial uncoupler, while LTB 4 metabolism is unaffected. This data strongly indicates that hepoxilin v-oxidizing activity is independent of LTB 4 v-oxidation and is coupled to the mitochondrial compartment whereas LTB 4 v-oxidation is carried out by a microsomal enzyme. It should be noted that mitochondria represent a small number of organelles in the neutrophil w24x. Hence, the loss of 20-hydroxylase activity described in this study may be compounded by the dilution of mitochondria during cell disruption. Some hitherto unidentified cofactor Ž other than adenine nucleotides. may also be required for the recovery of enzymatic activity. A scheme describing the structures of the products formed and elucidated in this paper is shown below ŽScheme 1.. HxA 3-Me is first taken up by the intact neutrophil where it is hydrolysed into the free acid Žmetabolite 1. by cytosolic esterases. This metabolite is subsequently converted into metabolite 2, the vhydroxy derivative of the free acid of HxA 3 Ž Fig. 2. . Reactions used to elucidate these structures involved acid catalysed epoxide ring opening into the corresponding trioxilin Žfrom HxA 3 . and v-hydroxy trioxilin Ž from v-hydroxy HxA 3 .. We additionally confirmed the terminal location of the hydroxyl group in
Scheme 1. Pathway of metabolism of hepoxilin A 3 methyl ester described in this study.
298
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
metabolite 2 through hydrogenation-induced ring opening of the epoxide. Finally we demonstrated equivalence of metabolite 2 with an authentic compound prepared by total chemical synthesis w19x. The chemically prepared compound showed identical mass spectra and retention times on GC. It is interesting to note that metabolite 2 retained an intact epoxide group even after 60 min incubation at 378C ŽFig. 2., indicating that neutrophils are quite deficient in epoxide hydrolase activity. References w1x C.R. Pace-Asciak, E. Granstrom, ¨ B. Samuelsson, J. Biol. Chem. 258 Ž1983. 6835–6840. w2x C.R. Pace-Asciak, J. Biol. Chem. 259 Ž1984. 8332–8337. w3x D. Reynaud, P. Demin, C.R. Pace-Asciak, J. Biol. Chem. 269 Ž1994. 23976–23980. w4x C.R. Pace-Asciak, W.-S. Lee, J. Biol. Chem. 264 Ž1989. 9310–9313. w5x C.R. Pace-Asciak, O. Laneuville, M. Chang, C.C. Reddy, W.-G. Su, E.J. Corey, Biochem. Biophys. Res. Commun. 163 Ž1989. 1230–1234. w6x C.R. Pace-Asciak, O. Laneuville, W.-G. Su, E.J. Corey, N. Gurevich, P. Wu, P.L. Carlen, Proc. Natl. Acad. Sci. USA 87 Ž1990. 3037–3041. w7x C.R. Pace-Asciak, D. Reynaud, P.M. Demin, Lipids 30 Ž1995. 1–8. w8x C.R. Pace-Asciak, Biochim. Biophys. Acta 1215 Ž1994. 1–8. w9x D. Reynaud, P. Demin, C.R. Pace-Asciak, Biochem. J. 313 Ž1996. 537–541.
w10x D. Reynaud, P.M. Demin, C.R. Pace-Asciak, Biochem. Biophys. Res. Comm. 207 Ž1995. 191–194. w11x Reynaud, D., Demin, P. and Pace-Asciak, C.R. Ž1995. in: Mediators in the Cardiovascular System — Regional Ischemia ŽSchror, ¨ K. and Pace-Asciak, C.R., eds.., Agents and Actions Supplements, Vol. 45, pp. 291–296, Birkhauser Verlag, Basel. w12x W.S. Powell, J. Biol. Chem. 259 Ž1984. 3082–3089. w13x R.J. Soberman, R.T. Okita, B. Fitzsimmons, J. Rokach, B. Spur, K.F. Austen, J. Biol. Chem. 262 Ž1987. 12421–12427. w14x M.C. Romano, R.G. Eckardt, P.E. Benders, T.B. Leonard, K.M. Straub, J.F. Newton, J. Biol. Chem. 262 Ž1987. 1590– 1595. w15x P.M. Demin, K.K. Pivnitsky, L.L. Vasiljeva, C.R. PaceAsciak, J. Labelled Compounds and Radiopharmaceuticals 34 Ž1994. 221–230. w16x A. Boyum, J. Clin. Lab. Invest. 21 Ž1968. 77–98. ¨ w17x S. Dho, S. Grinstein, E.J. Corey, W.G. Su, C.R. Pace-Asciak, Biochem. J. 266 Ž1990. 63–68. w18x Pace-Asciak, C.R. Ž1989. Adv. Prostagl. Thromb. Leuk. Res., 18. w19x P.M. Demin, T.A. Manukina, C.R. Pace-Asciak, K.K. Pivnitsky, Mendeleev Commun. 4 Ž1996. 130–132. w20x O. Laneuville, M. Chang, C.C. Reddy, E.J. Corey, C.R. Pace-Asciak, J. Biol. Chem. 265 Ž1990. 21415–21418. w21x O. Laneuville, E.J. Corey, R. Couture, C.R. Pace-Asciak, Biochim. Biophys. Acta 1084 Ž1991. 60–68. w22x A.J. Marcus, L.B. Safier, H.L. Ullman, M.J. Broekman, N. Islam, T.D. Oglesby, R.R. Gorman, Proc. Natl. Acad. Sci. USA 81 Ž1984. 903–907. w23x M. Nadeau, B. Fruteau de Laclos, B. Picard, P. Braquet, E.J. Corey, P. Borgeat, Can. J. Biochem. Cell Biol. 62 Ž1984. 1321–1326. w24x R.H. Kirschner, G.S. Getz, A.E. Evans, Enzyme 13 Ž1972. 56–76.
Hepoxilin A 3 is oxidized by human neutrophils into its v-hydroxy metabolite by an activity independent of LTB 4 v-hydroxylase 1 Denis Reynaud a , Olga Rounova a , Peter M. Demin a,b, Kazimir K. Pivnitsky b, Cecil R. Pace-Asciak a,c,) a b
Research Institute, Hospital For Sick Children, DiÕision of Neurosciences, 555 UniÕersity AÕenue, Toronto, Ont., Canada M5G 1X8 Institute of Experimental Endocrinology of National Endocrinology Scientific Center, National Academy of Medical Sciences, Moscow 115478, Russia c Department of Pharmacology, Faculty of Medicine, UniÕersity of Toronto, Toronto, Ont., Canada M5S 1A8 Received 13 February 1997; revised 14 April 1997; accepted 21 April 1997
Abstract Hepoxilin A 3-methyl ester is taken up by intact human neutrophils where it is first hydrolyzed into the free acid which is subsequently converted into a single major metabolite. The structure of this metabolite was determined through mass spectral analysis of several derivatives, and through identity with an authentic compound prepared by chemical synthesis. The metabolite was identified as v-hydroxy-hepoxilin A 3 showing that the epoxide functionality of the parent hepoxilin is not opened during incubation with human neutrophils. All attempts to investigate hepoxilin metabolism in broken cells, despite the presence of protease inhibitors ŽAproteinin, PMSF, DFP. and supplementation with NADPH were unsuccessful. Metabolism of hepoxilin A 3 required the intact cell, while parallel experiments with LTB 4 as substrate demonstrated that this eicosanoid was metabolized into its v-hydroxy metabolite regardless of whether intact or broken cell preparations were used provided that NADPH was present in the latter. Hepoxilin metabolism in intact cells was inhibited dose-dependently by CCCP Ž0.01–100 m M., a mitochondrial uncoupler, whereas LTB 4 metabolism was unaffected by CCCP. This data suggests that metabolism of hepoxilin A 3 occurs in intact human neutrophils through v-oxidation, is likely located in the mitochondrial compartment of the cell Žinhibition by CCCP. and is carried out by an activity that is independent of the well characterized, relatively stable microsomal LTB 4 v-hydroxylase. q 1997 Elsevier Science B.V. Keywords: Hepoxilin A 3; Metabolism; v-Oxidation; v-Hydroxy-hepoxilin A 3
1. Introduction Abbreviations: Me, methyl; FA, free acid; HxA 3 Ž8S., hepoxilin A 3 , 8ŽS.-hydroxy-11ŽS., 12ŽS.-trans-epoxyeicosa-5Z, 9E, 14Z-trienoic acid; HxB 3 Ž8S., hepoxilin B 3 , 10ŽS.-hydroxy-11ŽS., 12ŽS.-trans-epoxyeicosa-5Z, 8Z, 14Z-trienoic acid; RPMI 1640, Roswell Park Memorial Institute medium 1640; CCCP, carbonyl cyanide m-chlorophenylhydrazone; PMSF, phenylmethylsulfonylfluoride; DFP, diisopropylfluorophosphate; GCrMS, gas chromatographyrmass spectrometry ) Corresponding author. Fax: q1 Ž416. 813 5086; E-mail: [email protected]
Hepoxilins are derived through an intramolecular rearrangement of 12ŽS.-HPETE w1,2x. Although this process can take place nonenzymatically yielding two
1
The financial support of the MRC ŽMT-4181 to CRP-A. is gratefully acknowledged.
0005-2760r97r$17.00 q 1997 Elsevier Science B.V. All rights reserved. PII S 0 0 0 5 - 2 7 6 0 Ž 9 7 . 0 0 0 6 4 - 7
288
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
isomers, hepoxilin A 3 and B 3 , we have recently provided evidence of an enzymatic rearrangement in which mostly hepoxilin A 3 is formed w3x. Additionally, we have shown that the enzymatic process utilizes only 12Ž S. -HPETE as substrate for hepoxilin A 3 formation, the 12ŽR.-enantiomer is not selected in this transformation w3x. Two pathways of metabolism of hepoxilin A 3 have previously been reported. One involving an epoxide hydrolase resulting in epoxide ring opening to yield a biologically inactive trihydroxy metabolite w4x; the second pathway involves glutathione addition to the epoxide ring resulting in a biologically active peptido-hepoxilin w5,6x. Hepoxilins have been shown to possess a diversity of biological activities related to their actions on the mobilization of calcium and potassium ions in the cell Žfor recent reviews, see w7,8x.. We have recently shown the existence of a putative hepoxilin receptor in human neutrophils which binds specifically the hydroxy epoxide functional group as similar metabolites, e.g. 12-HETE, leukotriene B 4 , and various epoxy-containing metabolites of arachidonic acid, do not compete for the specific binding of hepoxilin A 3 w9–11x. Since this putative receptor is located inside the cell, it was necessary to employ the methyl ester of hepoxilin to demonstrate binding and calcium actions in the intact cell as the free acid of hepoxilin is inactive since it does not permeate the cell w10x. The free acid binds only when the cell is disrupted w9x. In this paper we provide evidence for the uptake in intact human neutrophils of the methyl ester of hepoxilin A 3 and its intracellular hydrolysis into the free acid which is then metabolized via a new pathway involving the oxidation of the terminal end of the compound to yield a v-hydroxy-hepoxilin A 3 in which the epoxide group is retained. Evidence is presented dissociating this hepoxilin metabolic activity from LTB 4 v-hydroxylase w12–14x. 2. Materials and methods
Žsilica gel G, ethyl acetateracetic acid 99r1, vrv. . w3 H 8 x-LTB 4 Žspec. act. 218 Cirmmol. was purchased from Amersham Ž Toronto. , and unlabeled LTB 4 was from Cayman Chemical Ž Ann Arbor. . RPMI 1640 medium and all chemicals for buffers were reagent grade and were purchased from Sigma, USA. Dextran T-500 and Ficoll–Paque were purchased from Pharmacia, Sweden. CCCP was purchased from Calbiochem, USA. 2.2. Preparation of human neutrophils Human neutrophils were prepared according to w16x with slight modifications according to Boyum ¨ Dho et al. w17x. Forty ml of venous blood was collected from normal drug-free human volunteers and anticoagulated with heparin sodium Ž Organon Tecknica Co., Durham, NC. . Erythrocytes were removed by 4.5% dextran sedimentation for 45 min. The neutrophil-rich plasma was centrifuged at 1500 rpm for 5 min. The resulting pellet was gently suspended in phosphate buffered saline whose composition was in mM: NaCl 140, NaH 2 PO4 10, Na 2 HPO4 10, pH 7.2. This suspension was subjected to Ficoll– Paque gradient centrifugation at 2600 rpm for 20 min. The entire supernatant was aspirated and discarded. The pellet containing the neutrophils was resuspended in 0.5 ml of RPMI 1640 Žfree of sodium bicarbonate but supplemented with L-glutamine and 25 mM HEPES buffered to pH 7.2. and the contaminating erythrocytes were eliminated by lysis in 10 ml of 0.85% ammonium chloride containing 17 mM Tris, pH 7.2, during 15 min at 378C. The suspension was centrifuged at 900 rpm for 5 min. The neutrophils were then resuspended in 5 ml of RPMI 1640 and allowed to re-equilibrate at 378C for 15 min. After an additional centrifugation at 900 rpm for 5 min, the cells were resuspended in RPMI 1640 medium and were counted in a Coulter counter ŽModel 901.. The cells were finally adjusted to a concentration of 10 7 cellsrml in RPMI 1640 medium. 2.3. Metabolism of HxA 3 by human neutrophils
2.1. Materials 3
8ŽS. HxA 3-methyl ester and w H 6 x-HxA 3 Ž 8S. -Me Žspec. act. 169 Cirmmol. were prepared in our laboratory by total chemical synthesis w15x. The purity of the compounds was established periodically by TLC
Freshly prepared human neutrophils Ž 100 = 10 6 cells. were incubated in 2 ml of Buffer A in a siliconized glass tube to which had been added w3 HxHxA 3 methyl ester Ž5 = 10 6 cpm. diluted with 30 m g of the unlabeled compound. The composition of
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
Buffer A consisted of the following in mM: NaCl 140, KCl 5, MgCl 2 1, CaCl 2 1, HEPES sodium-free 10, and glucose 10, pH 7.3. The mixture was incubated for various times, but for preparative runs, it was incubated for 60 min. The incubation mixture was then split in half; one-half was extracted with ethyl acetate without prior acidification to extract the intact hepoxilin structure, if so present in the incubation mixture, and its products. The other half was acidified to pH 3 to convert any hepoxilin structures present in the incubation mixture to the trioxilin structures through hydrolysis of the epoxide moiety. The trioxilins are considerably more polar on TLC than the hepoxilins. In order to determine whether the original methyl ester group of the substrate was hydrolyzed to the free acid during the incubation with neutrophils, one-half of each of the above extracts was subjected to methylation with diazomethane, and its TLC pattern was compared with the unmethylated half of the sample. Each sample was dissolved in ethyl acetate and spotted on TLC Ž silica gel G, ethyl acetateracetic acid 99.5r0.5, vrv. . The plate was developed for 60 min, then the solvent was dried with an air gun, and scanned for radioactivity on a TLC radiochromatogram scanner ŽBerthold.. Radioactive zones of silica gel were scraped into siliconized glass tubes and the compounds were extracted with ethyl acetatermethanol Ž 90r10, vrv. . In preparative runs, the zones corresponding to the metabolites 1 and 2 were isolated, the compounds were eluted with with methanol and extracted with ethyl acetate after dilution with water. The compounds of interest were recovered in the ethyl acetate phase. In order to confirm that the purified metabolites 1 and 2 contained an intact epoxide ring as in the substrate, a portion of the sample as well as authentic standard HxA 3-methyl ester were hydrolysed in 0.1% phosphoric acid in tetrahyrofuran for 16 h at 238C. The samples were then extracted into ether, washed neutral with water and evaporated to dryness. The reaction with acid converts the epoxide ring to an enantiomeric mixture of vicinal diols. Thus HxA 3 is converted into the triol, TrXA 3. 2.4. DeriÕatization Methyl esters were prepared by reaction with a freshly prepared ether solution of diazomethane for
289
10 min at 238C in the dark. Methyl esters were hydrolysed back to the free acid by reaction with ethanolr1 N KOH Ž1r1, vrv. for 30 min at 238C. The free acid of the compound was extracted into diethyl ether after gentle acidification to pH 6 with 0.01 N HCl. These mild conditions do not open the epoxide moiety of HxA 3. Pentafluorobenzyl esters were prepared from the free acid dissolved in 10 m l methanol and 50 m l acetonitrile by the rapid addition of 2 m l pentafluorobenzyl bromide Ž Pierce. and 1.5 m l of diisopropylethylamine Ž Pierce. . The mixture was heated at 658C for 5 min, followed by evaporation of the solvents and reagents with nitrogen gas in a fume hood. Trimethylsilyl ether derivatives were prepared by reacting the methyl esters or pentafluorobenzyl esters with 20 m l TRI-SIL Z Ž Pierce. for 5 min at 608C. The solvent was subsequently removed with nitrogen gas in a fume hood, the residue was dissolved in 10 m l dodecane, and the sample was placed on dry ice until analysed Ž usually within 60 min.. Hydrogenation of the methyl ester derivative was carried out using platinum oxide Ž Aldrich. as catalyst preactivated in methanol with hydrogen gas. The hydrogenation of the sample was carried out by bubbling hydrogen gas for 1 min. The catalyst was then filtered, and the solvent was taken to dryness with nitrogen gas. The residue was then silylated as described above. The following TLC R f values ŽSolvent system Žsilica gel G.: ethyl acetateracetic acid Ž99.5r0.5, vrv.. for the various derivatives were observed: HxA 3 Žmetabolite 1. , free acid 0.44, Me ester 0.67, PFB ester 0.72; v-hydroxy HxA 3 Žmetabolite 2., free acid 0.33, Me ester 0.42, PFB ester 0.54; TrXA 3 Žcompound 3., free acid 0.19–0.25, Me ester 0.30–0.37, PFB ester 0.40–0.44; v-hydroxy TrXA 3 Žcompound 4., free acid 0.06, Me ester 0.09, PFB ester 0.15–0.20. 2.5. Gas chromatography–mass spectrometry (GC r MS) Samples were analysed by GCrMS using both the electron impact ŽEI. mode and the negative ion chemical ionization Ž NICI. mode on a Hewlett Packard GCrMS ŽModel 5988. with direct on-column injection of sample on a capillary column Ž HP-1, 30 m.. The temperature was programmed from 1708C at injection to 3008C at 108Crmin. Helium was the
290
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
carrier gas and methane was the reactant gas in NICI mode. The following GC retention values Ž R t . were observed: Metabolite 2, MeTMSi 18.1 min, MeTMSi after hydrogenation 18.6 min, MeTMSi after acid opening of the epoxide 18.6 min; chemically synthesized 20-hydroxy HxA 3 MeTMSi 18.1 min, MeTMSi after hydrogenation 18.6 min, MeTMSi after acid opening of the epoxide 18.6 min.
knowing that the allylic epoxide of HxA 3 is quite unstable to acidic workup, and that the methyl ester of HxA 3 is relatively stable to normal handling Ž e.g.
2.6. Comparison of hepoxilin A 3 metabolism with LTB4 metabolism A series of experiments were carried out in parallel incubations of human neutrophil broken preparations with w3 H 6 xhepoxilin A 3 methyl ester or free acid and with w3 H 8 x-LTB 4 free acid or methyl ester. Freshly prepared neutrophils were either sonicated Ž 3 = 5 s; sonic dismembrator, Systems Corp., Farmingdale, NY., or freeze-thawed Žsequence of freezing on dry iceracetone and thawing at 238C repeated three times. , or homogenized then separated into mitochondrial Ž10 000 = g pellet. or microsomal Ž100 000 = g pellet. preparations and incubated in 1 ml of Buffer A in the presence or absence of 2 mM NADPH for various times Ž 0–60 min. . The incubation mixture was extracted with ethyl acetate, and analyzed by TLC as above. In other experiments, the protease inhibitors aproteinin ŽSigma. 0.17 Urml, PMSF ŽSigma. 0.1 mM, and DFP Ž Sigma. 1 mM, were added in separate experiments before cell disruption Žsonication or freeze-thawing. to investigate recovery of hepoxilin or LTB 4 metabolizing activity. Inhibition of hepoxilin A 3 but not LTB 4 metabolism by CCCP. Parallel incubations of intact human neutrophils Žprepared as above. with the two substrates were carried out at different concentrations of CCCP Ž 0–100 m M., during 60 min incubation at 378C. Sample extraction with ethyl acetate and TLC analysis was carried out as above.
3. Results When w3 Hx-HxA 3 methyl ester is incubated with intact human neutrophils, a single major metabolite with increased polarity is observed. We attempted to gain some quick information into its structure by a combination of simple techniques involving TLC,
Fig. 1. Thin layer radiochromatograms of ethyl acetate extracts of: ŽA. incubation of w 3 Hx-HxA 3-Me Ž5=10 6 cpm containing 30 m g of unlabeled hepoxilin. with intact human neutrophils Ž100= 10 6 cells. for 60 min at 378C in 2 ml of Buffer A; ŽB. a portion of the extract in ŽA. was acidified with 1 N HCl; ŽC. a portion of the extract in ŽA. was converted into the methyl ester form with diazomethane in diethyl ether; ŽD. a portion of the extract in ŽB. was subjected to methylation with diazomethane. TLC was performed on silica gel G plates which were developed with ethyl acetateracetic acid Ž99.5r0.5, vrv.. Note that in ŽA. two metabolites are formed; metabolite 1 was identified as the free acid form of the substrate HxA 3; metabolite 2 was identified as the free acid of v-hydroxy HxA 3. Note that very little, if any, of the epoxide hydrolase metabolite of HxA 3 , i.e. trioxilin A 3 Žcompound 3., is formed during the 60-min incubation demonstrating the virtual absence of epoxide hydrolase activity in human neutrophils. The epoxide ring opening of both metabolites is observed only when the metabolites are treated with acid to generate compounds 3 and 4 identified as trioxilin A 3 and v-hydroxy trioxilin A 3 , respectively Žfor mass spectra see Figs. 6–8..
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
extraction and TLC. . We therefore divided the incubation mixture into four parts. One part was extracted with ethyl acetate without acidification to ensure that the epoxide group, if present, would be extracted intact Ž Fig. 1A. ; a second part was acidified to pH 3 with HCl, a procedure which results in the hydrolysis of the epoxide to a more polar vicinal diol Že.g. trioxilin. ŽFig. 1B. . A third part was treated with an ether solution of diazomethane to form the methyl ester derivative Ž Fig. 1C.. Comparison of this chromatogram with that of the first part Ž Fig. 1A. would give information as to whether the carboxyl group of the metabolites was retained in the methyl ester form after incubation, or whether it had been hydrolysed into the free acid form. A fourth part of the original extract was acidified then subsequently methylated ŽFig. 1D., again to indicate which metabolites Ž in Fig. 1B. had a free carboxylic acid group. TLC analysis of these four extracts was quite informative as shown in Fig. 1. It is important to note Žsee Fig. 1A. that the methyl ester substrate is completely converted by the cells into a minor Ž metabolite 1. and a major product Žmetabolite 2.. That the epoxide group in the substrate was not opened during incubation was determined from the methylation of this extract ŽFig. 1C.. This panel shows that metabolite 1 reverts to the migration of the substrate, while metabolite 2 is clearly distinct from this. When subjected to methylation with diazomethane, metabolites 1 and 2 migrated upwards in the plate indicating that both metabolites 1 and 2 became methylated; hence both metabolites were present at the end of the incubation as free acids and were extracted as free acids. The hydrolysis of the methyl ester of HxA 3 by the cells is rationalised by the presence of esterases in the cell cytosol. It is also important to note that the epoxide is not opened into a diol during incubation with the cells, i.e. absence of significant epoxide hydrolases in the neutrophils, since very polar material corresponding to trioxilins is not present in appreciable amounts in the unacidified incubation mixture ŽFig. 1A.. That both metabolites Ž1 and 2. contained an intact epoxide was also shown by the fact that acid hydrolysis caused both metabolites to migrate with greater polarity Žcompare Fig. 1A with 1B, or Fig. 1C with 1D.. Metabolite 1 formed compound 3 upon acidification which migrated like trioxilin A 3 Ž TrXA 3 . , while metabolite 2 was converted
291
into compound 4 which was more polar than trioxilin A 3. Again both compounds 3 and 4 had free carboxylic acid groups since both reacted with diazomethane to form methyl esters Žcompare Fig. 1B and 1D. . We investigated the time course of the conversion of HxA 3-Me into metabolite 2. Fig. 2 shows that the methyl ester substrate is first hydrolysed into the free acid Ž metabolite 1. which is subsequently converted into metabolite 2. We have observed differences between subjects in this conversion although we have not yet carefully screened a large number of subjects to better study this potentially interesting observation. Attempts to modify this conversion through the addition of cofactors ŽATP, NAD, NADH, NADP or NADPH. were mostly unsuccessful as formation of metabolite 2 required the intact cell. All attempts to carry out the same metabolism with membranes from disrupted cells, as well as cell cytosol were unsuccessful. Also unsuccessful in recovering hepoxilin metabolizing activity during cell disruption was the use of the protease inhibitors, aproteinin, PMSF and DFP Ždata not shown.. In contrast, LTB 4 was metab-
Fig. 2. Time course of the metabolism of w 3 Hx-HxA 3-Me by intact human neutrophils. w 3 Hx-HxA 3-Me Ž0.55=10 6 cpm containing 5 m g of unlabeled hepoxilin. was incubated for times indicated with intact human neutrophils Ž16=10 6 cells. in 1 ml of Buffer A as in Fig. 1A at 378C. At the end of the incubation, the mixture was extracted with ethyl acetate, and this extract was evaporated with N2 gas and the extract was dissolved in a small amount of ethyl acetate and spotted on TLC as in Fig. 1. FA s free acid.
292
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
olized irrespective of whether the neutrophils were intact, or disrupted through sonication or freeze-thawing as long as NADPH was present during incubation of the disrupted cells. Fig. 3 compares the metabolism of hepoxilin A 3 and LTB 4 by cells subjected to freeze-thawing ŽFig. 3, panel A. and intact and sonicated cells ŽFig. 3, panel B.. As shown, hepoxilin metabolism disappears when the cell is disrupted, while LTB 4 metabolism is still relatively intact. That hepoxilin as well as LTB 4 metabolism was enzymatic was shown by the absence of any conversion of either substrate with cells that were placed in boiling water for 5 min Ždata not shown.. Use of the free acid of hepoxilin A 3 with disrupted cells did not result in any conversion into products. Conversely, LTB 4 was
v-oxidized by the intact or disrupted cell, irrespective of whether it was used in the free acid or methyl ester form. Additional experiments were carried out to dissociate hepoxilin metabolism from that which oxidizes LTB 4 . We discovered that CCCP, a mitochondrial uncoupler, dose-dependently inhibits the metabolism of HxA 3 , while LTB 4 is unaffected by the drug. Fig. 4 shows TLC patterns of extracts of incubations analyzing HxA 3 and LTB 4 metabolism in the presence of CCCP, 5 m M for HxA 3 and 60 m M for LTB 4 substrates. Fig. 5 shows the dose-dependency of the selective inhibition by CCCP of HxA 3 metabolism into the v-hydroxy metabolite, with essentially unaffecting LTB 4 v-oxidation. These data
Fig. 3. Radio TLC profiles comparing the effect of cell disruption — ŽA. freeze-thaw method, ŽB. sonication — of human neutrophils in their ability to metabolize hepoxilin A 3 and LTB 4 . For methods see legend to Fig. 2. Incubation time was 60 min.
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
suggest that the activity that oxidizes hepoxilin A 3 is different from that which oxidizes LTB 4 . Additionally, the data with CCCP suggest that hepoxilin metabolism occurs in the mitochondria, whereas LTB 4 metabolism occurs in the microsomal fraction of a cell homogenate. Since hepoxilin metabolizing activity was lost during cell disruption, direct evidence of the localization of hepoxilin metabolism to mitochondria could not be obtained but is suggested by its inhibition by CCCP. Metabolite 1 was identified as the free acid of HxA 3 by GCrMS, and compound 3 was the corresponding trioxilin. Metabolite 2 was subjected to structural analysis by GCrMS in two ionization modes. First, the free acid was isolated by TLC, converted into the PFB ester and analyzed as the TMSi ether derivative using the NICI mode. In this ionization mode, a prominent ion results from the loss of the PFB group essentially giving important information on the molecular weight of the compound being analysed. While the parent HxA 3 showed a major fragment ion at mrz 407 ŽM-PFB., in metabolite 2 this fragment ion shifted to mrz 495, corresponding to the presence in metabolite 2 of an additional hydroxyl group. Similar information was
293
Fig. 5. Concentration-dependent inhibition by CCCP of hepoxilin A 3 v-oxidation but not LTB 4 v-oxidation in intact human neutrophils. For methods see legend to Fig. 2. Incubation time was 60 min.
obtained when the acidified HxA 3 Ž which forms trioxilin A 3 , a trihydroxy derivative. was subjected to NICI-GCrMS. Its spectrum showed a fragment ion at mrz 569 Ž M-PFB. , while that of the corresponding derivative of metabolite 2, i.e. compound 4, showed a
Fig. 4. Inhibition of hepoxilin but not LTB 4 v-oxidation in intact human neutrophils by CCCP, a mitochondrial uncoupler. Note that CCCP does not interfere with the hydrolysis of the methyl ester of hepoxilin to the corresponding free acid, only with the v-oxidation step. For methods see legend to Fig. 2. Incubation time was 60 min.
294
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
fragment ion at mrz 657, confirming that metabolite 2 had an extra hydroxyl group. As expected metabolite 2 was more polar than the parent HxA 3 due to the presence in the former of an additional hydroxyl group Žsee Fig. 1. . In order to obtain more information on the location of the additional hydroxyl group in metabolite 2, it was subjected to EI-GCrMS ŽFig. 6B. as the MeTMSi derivative. Data is shown in comparison to that obtained from the parent HxA 3 ŽFig. 6A. . Fig. 6A shows the mass spectrum of HxA 3 with a prominent
fragmentation at the C7–C8 bond Ž M-141, C1–C7. giving rise to a fragment ion at mrz 281 ŽC8–C20.. In the corresponding spectrum of metabolite 2 Ž Fig. 6B., this fragment ion is shifted to mrz 369 representing the presence of an additional hydroxyl group Ž281 q 88. in the C8–C20 portion of the structure of HxA 3. An important fragment of structural significance locating the hydroxyl group at the v-position was the ion at mrz 103 in metabolite 2 Ž Fig. 6B., which was absent in the parent compound which lacks this moiety Ž Fig. 6A. . Localization of the addi-
Fig. 6. GCrMS ŽEI. of the MeTMSi derivatives of ŽA. authentic HxA 3 and ŽB. metabolite 2 identified as v-hydroxy HxA 3. Note the shift of fragment ions indicated in panel B in the relevant portions of the molecule containing the additional hydroxyl group, i.e. mrz 369 Ž281 q 88..
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
tional hydroxyl group of metabolite 2 to the terminal end of the molecule was derived from the mass spectra of the following derivatives. The acidified products derived from HxA 3 Ž i.e. TrXA 3 , compound 3. and metabolite 2 Ži.e. compound 4. showed structurally informative ions at mrz 213 ŽC12–C20. and 301 ŽC12–C20., respectively Ž Fig. 7A and B. . The fragment ion at mrz 243 due to the C1–C8 portion of the molecule is common to both compounds. Minor fragment ions occur in the spectrum of compound 4 ŽFig. 7B. at mrz 657 wM-15x and mrz 492
295
wM-Ž2 = 90.x although the corresponding fragments for compound 3 are almost undetectable due to easier fragmentation along the carbon–carbon bonds mentioned above. Further support of these fragmentations was obtained with the hydrogenated derivatives. Catalytic hydrogenation of HxA 3 led to the opening of the epoxide with insertion of a hydrogen atom at C11 w1x. Mass spectral analysis of this product gave an informative fragment at mrz 215 representing the fragment ion containing the C12–C20 portion of the molecule ŽFig. 8A. . The comparable fragment in the
Fig. 7. GCrMS ŽEI. of the acid-catalysed products obtained from ŽA. HxA 3 , i.e. trioxilin A 3 , and ŽB. compound 4, i.e. tetroxilin A 3 , derived from metabolite 2. Note the shift of fragment ions in ŽB. in the relevant portions of the molecule containing the additional hydroxyl group, i.e. mrz 301 Ž213 q 88..
296
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
mass spectrum of the hydrogenated metabolite 2 was shifted to mrz 303 Ž215 q 88. indicating that the double bond at D14,15 had not been affected in the formation of metabolite 2, and that the additional hydroxyl group must be located at the terminal end of the metabolite ŽFig. 8B.. The common fragment ion at mrz 243 in the original compounds is shifted to mrz 245 in the hydrogenated compounds confirming that the double bond at C5–C6 in the original substrate remained intact in the metabolite. Additional shifts in minor fragments due to wM-15x and wM-90x
are also seen between hydrogenated HxA 3 Ž mrz 487 and 412, respectively. and hydrogenated metabolite 2 Žmrz 575 and 500, respectively. Ž Fig. 8A, B.. The placement of the hydroxyl group at C20 in metabolite 2 is due to the low intensity fragments in the mrz 110–140 region of the spectra as compounds with a hydroxyl group at C19 typically have higher intensity fragment ions in this region Ž see mass spectra of 19and 20-hydroxy prostaglandins in w18x.. This data suggests that metabolite 2 is v-hydroxy-HxA 3 Ž or 20-hydroxy HxA 3 .. Finally, additional proof for the
Fig. 8. GCrMS ŽEI. of the catalytic hydrogenation products ŽPtO 2rmethanol. of ŽA. HxA 3 and ŽB. metabolite 2 identified as v-hydroxy HxA 3. Samples were analysed as the MeTMSi derivatives. As with Fig. 6 and Fig. 7, note the shift of fragment ions in ŽB. in the relevant portions of the molecule containing the additional hydroxyl group, i.e. mrz 303 Ž215 q 88. and 447 Ž359 q 88..
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
v-location of the hydroxyl group was obtained through identity of the mass spectrum and chromatographic properties of the metabolite with an authentic 20-hydroxy HxA 3 prepared by total chemical synthesis w19x.
4. Discussion Previous studies have shown two metabolic pathways for the metabolism of hepoxilins, i.e. by hepoxilin epoxide hydrolase w4x and a glutathione S-transferase w6,20,21x. In this study, we provide evidence that human neutrophils metabolize hepoxilins via a third pathway involving v-oxidation with retention of the hydroxy epoxide functional group of the parent hepoxilin. v-Oxidation of eicosanoids has been shown to take place in human neutrophils. 12-HETE and LTB 4 have been shown to be transformed into their voxidized products w12–14,22,23x. We have also found in this study that human neutrophils metabolize LTB 4 but in contrast to LTB 4 v-oxidation, we found that hepoxilin metabolism Žformation of metabolite 2. is abolished when cells are disrupted, both when the cellular components are retained Ž through freezethawing. or when they are separated Ž through sonica-
297
tion.. Additionally, we discovered that hepoxilin metabolism in intact neutrophils is inhibited by CCCP, a mitochondrial uncoupler, while LTB 4 metabolism is unaffected. This data strongly indicates that hepoxilin v-oxidizing activity is independent of LTB 4 v-oxidation and is coupled to the mitochondrial compartment whereas LTB 4 v-oxidation is carried out by a microsomal enzyme. It should be noted that mitochondria represent a small number of organelles in the neutrophil w24x. Hence, the loss of 20-hydroxylase activity described in this study may be compounded by the dilution of mitochondria during cell disruption. Some hitherto unidentified cofactor Ž other than adenine nucleotides. may also be required for the recovery of enzymatic activity. A scheme describing the structures of the products formed and elucidated in this paper is shown below ŽScheme 1.. HxA 3-Me is first taken up by the intact neutrophil where it is hydrolysed into the free acid Žmetabolite 1. by cytosolic esterases. This metabolite is subsequently converted into metabolite 2, the vhydroxy derivative of the free acid of HxA 3 Ž Fig. 2. . Reactions used to elucidate these structures involved acid catalysed epoxide ring opening into the corresponding trioxilin Žfrom HxA 3 . and v-hydroxy trioxilin Ž from v-hydroxy HxA 3 .. We additionally confirmed the terminal location of the hydroxyl group in
Scheme 1. Pathway of metabolism of hepoxilin A 3 methyl ester described in this study.
298
D. Reynaud et al.r Biochimica et Biophysica Acta 1348 (1997) 287–298
metabolite 2 through hydrogenation-induced ring opening of the epoxide. Finally we demonstrated equivalence of metabolite 2 with an authentic compound prepared by total chemical synthesis w19x. The chemically prepared compound showed identical mass spectra and retention times on GC. It is interesting to note that metabolite 2 retained an intact epoxide group even after 60 min incubation at 378C ŽFig. 2., indicating that neutrophils are quite deficient in epoxide hydrolase activity. References w1x C.R. Pace-Asciak, E. Granstrom, ¨ B. Samuelsson, J. Biol. Chem. 258 Ž1983. 6835–6840. w2x C.R. Pace-Asciak, J. Biol. Chem. 259 Ž1984. 8332–8337. w3x D. Reynaud, P. Demin, C.R. Pace-Asciak, J. Biol. Chem. 269 Ž1994. 23976–23980. w4x C.R. Pace-Asciak, W.-S. Lee, J. Biol. Chem. 264 Ž1989. 9310–9313. w5x C.R. Pace-Asciak, O. Laneuville, M. Chang, C.C. Reddy, W.-G. Su, E.J. Corey, Biochem. Biophys. Res. Commun. 163 Ž1989. 1230–1234. w6x C.R. Pace-Asciak, O. Laneuville, W.-G. Su, E.J. Corey, N. Gurevich, P. Wu, P.L. Carlen, Proc. Natl. Acad. Sci. USA 87 Ž1990. 3037–3041. w7x C.R. Pace-Asciak, D. Reynaud, P.M. Demin, Lipids 30 Ž1995. 1–8. w8x C.R. Pace-Asciak, Biochim. Biophys. Acta 1215 Ž1994. 1–8. w9x D. Reynaud, P. Demin, C.R. Pace-Asciak, Biochem. J. 313 Ž1996. 537–541.
w10x D. Reynaud, P.M. Demin, C.R. Pace-Asciak, Biochem. Biophys. Res. Comm. 207 Ž1995. 191–194. w11x Reynaud, D., Demin, P. and Pace-Asciak, C.R. Ž1995. in: Mediators in the Cardiovascular System — Regional Ischemia ŽSchror, ¨ K. and Pace-Asciak, C.R., eds.., Agents and Actions Supplements, Vol. 45, pp. 291–296, Birkhauser Verlag, Basel. w12x W.S. Powell, J. Biol. Chem. 259 Ž1984. 3082–3089. w13x R.J. Soberman, R.T. Okita, B. Fitzsimmons, J. Rokach, B. Spur, K.F. Austen, J. Biol. Chem. 262 Ž1987. 12421–12427. w14x M.C. Romano, R.G. Eckardt, P.E. Benders, T.B. Leonard, K.M. Straub, J.F. Newton, J. Biol. Chem. 262 Ž1987. 1590– 1595. w15x P.M. Demin, K.K. Pivnitsky, L.L. Vasiljeva, C.R. PaceAsciak, J. Labelled Compounds and Radiopharmaceuticals 34 Ž1994. 221–230. w16x A. Boyum, J. Clin. Lab. Invest. 21 Ž1968. 77–98. ¨ w17x S. Dho, S. Grinstein, E.J. Corey, W.G. Su, C.R. Pace-Asciak, Biochem. J. 266 Ž1990. 63–68. w18x Pace-Asciak, C.R. Ž1989. Adv. Prostagl. Thromb. Leuk. Res., 18. w19x P.M. Demin, T.A. Manukina, C.R. Pace-Asciak, K.K. Pivnitsky, Mendeleev Commun. 4 Ž1996. 130–132. w20x O. Laneuville, M. Chang, C.C. Reddy, E.J. Corey, C.R. Pace-Asciak, J. Biol. Chem. 265 Ž1990. 21415–21418. w21x O. Laneuville, E.J. Corey, R. Couture, C.R. Pace-Asciak, Biochim. Biophys. Acta 1084 Ž1991. 60–68. w22x A.J. Marcus, L.B. Safier, H.L. Ullman, M.J. Broekman, N. Islam, T.D. Oglesby, R.R. Gorman, Proc. Natl. Acad. Sci. USA 81 Ž1984. 903–907. w23x M. Nadeau, B. Fruteau de Laclos, B. Picard, P. Braquet, E.J. Corey, P. Borgeat, Can. J. Biochem. Cell Biol. 62 Ž1984. 1321–1326. w24x R.H. Kirschner, G.S. Getz, A.E. Evans, Enzyme 13 Ž1972. 56–76.