Biosynthesis of a linoleic acid allylic epoxide: mechanistic comparison with its chemical synthesis and leukotriene A biosynthesis

Abstract

Biosynthesis of the leukotriene A (LTA) class of epoxide is a lipoxygenase-catalyzed transformation requiring a fatty acid hydroperoxide substrate containing at least three double bonds. Here, we report on biosynthesis of a dienoic analog of LTA epoxides via a different enzymatic mechanism. Beginning with homolytic cleavage of the hydroperoxide moiety, a catalase/peroxidase-related hemoprotein from Anabaena PCC 7120, which occurs in a fusion protein with a linoleic acid 9R-lipoxygenase, dehydrates 9R-hydroperoxylinoleate to a highly unstable epoxide. Using methods we developed for isolating extremely labile compounds, we prepared and purified the epoxide and characterized its structure as 9R,10R-epoxy-octadeca-11E,13E-dienoate. This epoxide hydrolyzes to stable 9,14-diols that were reported before in linoleate autoxidation (Hamberg, M. 1983. Autoxidation of linoleic acid: Isolation and structure of four dihydroxy octadecadienoic acids. Biochim. Biophys. Acta 752: 353–356) and in incubations with the Anabaena enzyme (Lang, I., C. Göbel, A. Porzel, I. Heilmann, and I. Feussner. 2008. A lipoxygenase with linoleate diol synthase activity from Nostoc sp. PCC 7120. Biochem. J. 410: 347–357). We also prepared an equivalent epoxide from 13S-hydroperoxylinoleate using a “biomimetic” chemical method originally described for LTA₄ synthesis and showed that like LTA₄, the C18.2 epoxide conjugates readily with glutathione, a potential metabolic fate in vivo. We compare and contrast the mechanisms of LTA-type allylic epoxide synthesis by lipoxygenase, catalase/peroxidase, and chemical transformations. These findings provide new insights into the reactions of linoleic acid hydroperoxides and extend the known range of catalytic activities of catalase-related hemoproteins.

The plethora of products formed through oxygenation of polyunsaturated fatty acids was extended recently through characterization of the activities of an enzyme from Anabaena PCC 7120 (1), a photosynthetic cyanobacterium which is also termed Nostoc sp. PCC 7120 (2). The enzyme in question is a plasmid-encoded fusion protein with a catalase-related hemoprotein on the N-terminal side and a lipoxygenase at the C-terminal end. The catalase-related domain is 39 kDa in size, directly connected to the lipoxygenase (49 kDa), giving a molecular mass of 88 kDa for the full-length fusion protein. The full-length fusion protein was successfully expressed in Escherichia coli, and although the cDNA encoding only the catalase-related domain did not express as active protein, the cDNA of the lipoxygenase (LOX) domain expressed very well by itself, allowing the dual catalytic activities of the fusion protein to be clearly distinguished. The LOX domain converts C18 fatty acids to the corresponding 9R-hydroperoxide (3, 4). This is an unusual LOX activity, forming products enantiomeric to the 9S-LOX of plants (5), and characterized earlier in Hydra vulgaris (6), in the marine alga Ulva conglobata (7), and as an activity of Gersemia fruticosa arachidonate 11R-LOX (8). We showed that the catalase-related domain converts the 9R-hydroperoxide of α-linolenic acid to two unstable epoxides that we isolated and characterized (1). One is an allylic 9,10-epoxide with C13-C16 of the carbon chain cyclized into a [1.1.0]bicyclobutane, a unique structural motif in biology. A second product, an allylic epoxide of the leukotriene A (LTA) type, 9R,10R-octadeca-11E,13E,15Z-trienoic acid, is also formed from the same 9R-hydroperoxylinolenic acid substrate (1).

The metabolism of linoleic acid by this Anabaena fusion protein was studied by Feussner and colleagues and reported to form 9R-HPODE via the LOX domain and three more polar dihydroxy products (2). The major diol was identified as 9,14-dihydroxyoctadeca-10E,12E-dienoic acid. Formation of a 9,10-dihydroxyoctadeca-11E,13E-dienoic acid and an 8,11-dihydroxyoctadeca-9E,12E-dienoic acid was also described. Additional experiments using incubations conducted using H₂¹⁸O or ¹⁸O₂ indicated that the hydroxyl at C-9 in the main diol arises from molecular oxygen and the hydroxyl at C14 from water. The fusion protein was termed a linoleate diol synthase (2).

Based on the structures of the linoleic acid-derived 9,14- and 9,10-diols formed by the Anabaena fusion protein and the incorporation of water at C14 in the main diol, they are likely to be the stable end products of unstable epoxide intermediate(s) that hydrolyze under the typical conditions of incubation at room temperature. If this were the case, based on analogy to LTA₄ and other allylic epoxides (e.g., Refs. 9–11), hydrolysis should produce two diastereomeric 9,14-diols as the main products together with two minor 9,10-diols. In the first part of this article, we clarify this issue. We then report the isolation and characterization of the unstable epoxide precursor. Third, we compare the stereochemical structure of this product to the isomeric epoxides formed through a chemical transformation. The chemical method was developed originally as a “biomimetic” synthesis in starting from a fatty acid hydroperoxide and giving the natural epoxide as one of the products, although the steps in between are completely different from the enzymology (12–14). Originally developed for preparation of the conjugated triene-containing allylic epoxide LTA₄, here, we report its application to preparation of conjugated diene-containing epoxides.

EXPERIMENTAL PROCEDURES

Materials

Fatty acids were purchased from NuChek Prep (Elysian, MN). [1-¹⁴C]Linoleic acid was purchased from Perkin-Elmer Life Sciences. 9R-Hydroperoxy-C18.2ω6 (9R-HPODE) was prepared using the LOX domain of the Anabaena enzyme in pH 7.5 Tris buffer using methods described previously for the linolenic acid analog (1); the reaction of linoleic acid with the LOX domain in buffer alone exhibits strong substrate inhibition at fatty acid concentrations over 30 μM, necessitating large volumes for preparation of milligram quantities (e.g., 4 mg in 200 ml), but this can be largely overcome by including 0.1% CHAPS detergent in the solution (e.g., allowing reaction of 5 mg fatty acid in 50 ml buffer). After reaction was complete as judged by UV analysis, the 9R-HPODE product was extracted with dichloromethane at pH 4 and purified by straight-phase high-pressure liquid chromatography (SP-HPLC) (1).

Expression and purification of Anabaena enzyme constructs

The cDNA for the full-length peroxidase-LOX fusion protein from Anabaena sp. strain PCC 7120 was cloned into pET17b for expression in E. coli as described previously (1). The full-length construct included the native cDNA sequence with a (His)₆ tag on the N terminus (MHHHHHHDLNTY—–LMMSINI.). The protein was expressed in E. coli BL21 (DE3) cells (Novagen) using methodology described previously (15) and was purified on nickel-nitrilotriacetic acid agarose (Qiagen, Valencia, CA) according to the manufacturer's instructions and quantified from the main Soret band at 406 nm assuming ɛ = 100,000. The LOX-only domain (starting at amino acid 344 with the amino acid sequence KDDLPGK…., and comprising the last 430 amino acids of the full-length construct) was expressed with an N-terminal His₆ tag and purified by nickel affinity chromatography (3).

Enzyme incubations and extraction of stable end products

Incubations with purified enzyme were conducted in 50 mM Tris and 150 mM NaCl, pH 7.5, at room temperature. Small-scale reactions (0.5 ml and 1 ml) were monitored by UV spectroscopy by repetitive scanning in the range 200–300 nm or by monitoring the decrease in absorbance over time at 235 nm. For preparation of products for NMR analysis, reactions were conducted in 50 ml buffer with up to 200 μM 9R-HPODE as substrate, added in a small volume of ethanol (≤1% by volume). After 20 min, samples were acidified by addition of a previously titrated amount of 1 M KH₂PO₄ plus sufficient 1 N HCl to give pH 6 and applied to a preequilibrated Bond-Elut C18 cartridge (1 g cartridge per 25 ml original buffer), and after washing with water the products were eluted with methanol.

HPLC analysis of stable end products

Samples were run on an Agilent (Palo Alto, CA) 1100 HPLC equipped with a diode array detector (and for radiolabeled samples, connected online to a Flo-One A-100 radioactive detector; Radiomatic Instruments and Chemical, Meridian, CT). Samples were analyzed initially using a Waters C18 Symmetry column (25 × 0.46 cm) with a solvent system of methanol/water/glacial acetic acid (75/25/0.01, by volume), a flow rate of 1 ml/min, and with UV detection at 205, 220, 235, and 270 nm; for improved resolution, in some runs, the solvent proportions were changed to 65/35/0.01 (methanol/water/glacial acetic acid, by volume). For methyl esters, a solvent system of methanol/water 75/25 (v/v) was employed. Subsequent SP-HPLC of individual products used a Beckman Ultrasphere 5 μ silica column (25 × 0.46 cm), a solvent system of hexane/isopropanol/acetic acid (100:5:0.1, by volume), a flow rate of 1 ml/min for free acids, and a solvent of hexane/isopropanol (100:3, v/v) for methyl esters. For chiral phase HPLC, we employed a Daicel Chiralpak AD column (0.46 × 25 cm), hexane/methanol (100:2 or 100:5, by volume), and 1 ml/min (16).

Preparation and purification of the unstable epoxide

Enzyme reactions were conducted at 0°C essentially as previously described (1). Briefly, reaction was initiated by vigorous vortex mixing of a hexane solution of 9R-HPODE (100–400 μM in 20 ml) with the Anabaena enzyme (0.8 nmole) in 200 μl of phosphate buffer, pH 8. After 1 min of mixing, the hexane phase was collected, evaporated to ∼2 ml under nitrogen, and the fatty acids methylated using ethereal diazomethane in the presence of 1% ethanol. The sample was evaporated to dryness, redissolved in hexane, and stored at −80°C until further analyzed. Initially, we experimented with HPLC under cold conditions (silica column, hexane/diethyl ether solvent, 100:5, v/v, run at −10–15°C) as used previously for isolation of allene epoxides (17). Later, the use of HPLC conditions previously developed for chromatography of LTA₄ proved more satisfactory, either SP-HPLC of methyl esters using triethylamine in the solvent (14, 18) or our modification of a reversed-phase HPLC (RP-HPLC) procedure, suitable for methyl esters or acids (1, 19). For SP-HPLC, an Altex silica guard column (5 μm, 45 × 4.6 mm) was run with a solvent of hexane/triethylamine (100:0.5, by volume) eluted at 3 ml/min. For RP-HPLC, the hexane extracts were analyzed and purified using a Waters Symmetry C18 5 μm column (0.46 × 25 cm) eluted at a flow rate of 1 ml/min with methanol/20 mM potassium phosphate pH 8 (replaced with triethylamine for LC-MS analysis, see below) in the proportions 80:20 (by volume), with UV detection at 205, 220, 235, and 270 nm using an Agilent 1100 series diode array detector. The main products from RP-HPLC were recovered by extraction with cold hexane followed by evaporation to dryness under a strong stream of nitrogen; products were promptly redissolved in hexane/triethylamine (100:0.1, by volume) and stored at −80°C prior to further analysis.

Incubation in H₂¹⁸O and LC-MS analysis of the products

LC-MS of the stable products from reaction of 9R-HPODE with the Anabaena enzyme in H₂¹⁸O buffer (¹⁸O:¹⁶O, 3:1) was performed using a Thermo Finnigan LC Quantum instrument. A Waters Symmetry C18 column (15 × 0.2 cm) was eluted with methanol/10 mM ammonium acetate (80:20, by volume) at 0.2 ml/min. The heated capillary ion lens was operated at 220°C. Nitrogen was used as a nebulization and desolvation gas. The electrospray potential was held at 4 kV. Source-induced dissociation was set at −10 eV. Negative ion mass spectra were acquired over the mass range m/z 100 to 500 at 2 s/scan. Collision-induced dissociation was performed at −15 eV.

Biomimetic synthesis of C18.2 epoxides

The synthesis followed the method described for the conversion of 5- hydroperoxyeicosatetraenoic acid (5-HPETE) to LTA₄ as the methyl ester (14) and the corresponding reaction of 15-HPETE methyl ester to the 14,15-epoxide analog of LTA₄ (13), as used previously in our laboratory (18). For synthesis of the C18.2 allylic epoxides, 3 mg (10 μmol) 13S-HPODE methyl ester was evaporated to an oil under N₂ in a dry 300 μl Reactivial (Pierce) to which a small triangular Teflon stirrer was added. An argon atmosphere, instilled via a syringe port cap, and 80 μl freshly distilled THF:CH₂Cl₂ (1:1) was added. To this was added 6.0 μl 1,2,2,6,6-pentamethylpiperidene (35 μmol), freshly distilled under vacuum from CaH₂; the resulting solution was cooled to −70°C by means of a dry ice-acetone bath. Five minutes later, 3.0 μl trifluoromethane sulfonic anhydride (Alfa or Aldrich) was added, which gave a pale-yellow color. The reaction was maintained at −70°C and was complete by 45 min. Triethylamine (10 μl) was added, and the reaction warmed to room temperature and transferred to 500 μl hexane:Et₂O:triethylamine (50:50:1). This was washed twice with equal volumes of brine and the organic phase evaporated under N₂ to give a yellow oil that was promptly reconstituted with 800 μl hexane/triethylamine (100:1, v/v), layered with argon, and stored at −70°C for analysis. The crude mixture was purified in aliquots by SP-HPLC at room temperature using a Waters μPorasil column (10 μ, 3.9 × 250 mm) and a solvent system of hexane:triethylamine (100:1 v/v/) at 1.0 ml/min with UV detection at 250 nm. This synthesis was completed in our laboratory in 1984, and 1 mg samples of the two HPLC-purified epoxide methyl esters were stored at −70°C in HPLC column solvent since that time. Aliquots of the two epoxides were repurified by RP-HPLC in conjunction with this study using the HPLC conditions described above for the Anabaena C18.2 epoxide methyl ester. Using a Waters Symmetry C18 column (25 × 0.46 cm), a solvent of methanol/water/triethylamine (85/15/0.1) at a flow rate of 1 ml/min with UV detection at 250 nm, the epoxide methyl esters eluted in the same order as on SP-HPLC at retention times of 14 and 15.5 min, respectively. Their UV spectra were acquired with the online diode array detector, and ¹H-NMR spectra were recorded in d₆-benzene.

Derivatization and GC-MS analysis

Methyl esters were prepared using ethereal diazomethane/methanol (5:1). Catalytic hydrogenations were performed in 100 μL of ethanol using about 1 mg of palladium on alumina and bubbling with hydrogen for 2 min at room temperature. The hydrogenated products were recovered by the addition of water and extraction with ethyl acetate. Trimethylsilyl ester and trimethylsilyl ether derivatives were prepared by treatment with bis(trimethylsilyl)-trifluoracetamide (10 μL) and pyridine (5 μL) at room temperature for 2 h. Subsequently, the reagents were evaporated under a stream of nitrogen, and the samples were dissolved in hexane for GC-MS analysis. GC-MS was carried out in the positive ion electron impact mode (70 eV) using a Thermo Finnigan Trace DSQ ion trap GC-MS with the Xcalibur data system, version 1.3. Samples were injected at 150°C, and after 1 min, the temperature was programmed to 300°C at 12 or 20°C/min.

NMR analyses

¹H-NMR spectra were recorded on a Bruker 600 MHz instrument at 283K. Samples were dissolved in d₆-benzene or CDCl₃, and the chemical shifts are reported relative to the benzene (δ 7.16 ppm for hydrogen) or CHCl₃ (δ 7.24 ppm) signals. The chemically synthesized epoxide methyl esters were analyzed originally on a Bruker 400 MHz instrument using carbon tetrachloride as solvent. (Although the epoxides were stable in this solvent and satisfactory spectra were obtained, the use of CCl₄ for proton NMR necessitates the addition of tetramethylsilane or other reference compound to provide a lock signal for our current instruments.)

RESULTS

Metabolism of linoleic acid by the Anabaena fusion protein

When linoleic acid is metabolized by the recombinant Anabaena fusion protein, it is rapidly oxygenated to 9R-HPODE by the LOX domain (2–4) and then further transformed to polar products. Analysis of the polar products was carried out using purified 9R-HPODE as substrate, and a typical RP-HPLC chromatogram is shown in Fig. 1. The two most prominent peaks, which exhibit identical conjugated diene chromophores with λmax at 230.5 nm (Fig. 1, peaks 1 and 2, inset) were identified as 9,14-dihydroxyoctadeca-10E,12E-dienoic acids by GC-MS and ¹H-NMR (see supplementary data). Two smaller and later eluting peaks exhibited conjugated dienes with a very slight shift to higher wavelength (peaks 3 and 4, λmax 231.5 and 231.8 nm, respectively; Fig. 1, inset) and were identified as 9,10-dihydroxy-octadeca-11E,13E-dienoic acids (see supplementary data). Based on the stereochemistry of 9,14-diols deduced by Hamberg (20), and assuming that the original R chirality of the hydroperoxide is retained in the C9 hydroxyl, the two 9,14-diols are the 9R,14S (peak 1) and 9R,14R (peak 2) diastereomers. These are predicted to be the major products of hydrolysis of an unstable allylic 9,10-epoxide, with the 9,10-diols being the minor hydrolysis products. This was supported by LC-MS analysis of an incubation conducted in H₂¹⁸O, in which the incorporation of one atom of ¹⁸O from water was confirmed in both the 9,14- and 9,10-diols (data not shown).

Fig. 1.

RP-HPLC analysis of the stable end products of linoleic acid metabolism by the Anabaena peroxidase-LOX fusion protein. Column: Waters Symmetry C18, 25 × 0.46 cm; solvent, methanol/water/glacial acetic acid (65:35:0.01, by volume); flow rate, 1 ml/min; UV detection at 235 nm. Inset: the UV spectra of peaks 1–4; the spectra of peaks 1 and 2 are identical to each other (λmax 230.5 nm), while 3 is shifted to 231.5 nm and 4 (dashed line) to 231.8 nm.

Some very minor products (≤5% abundance of the main 9,14-diols) were structurally characterized. These included an unusual 9,12-dihydroxy-octadeca-10,13-dienoic acid hydrolysis product that chromatographs immediately after the second 9,14-diol peak on RP-HPLC (see supplementary data). We also found a minor epoxyalcohol, 9,10-epoxy-11-hydroxy-octadeca-12-enoic acid that chromatographed between the two 9,10-diols and gave a mass spectrum as the methyl ester TMS ether derivative similar to that attributed by Lang et al. to an 8,11-diol (2). Its relatively featureless and not very informative mass spectrum, dominated by the cleavage ion between the epoxide and hydroxyl that retains the OTMS moiety, is typical of fatty acid epoxides with an α-hydroxyl (hepoxilin B-type epoxyalcohols) (e.g., Refs. 21–23) (see supplementary data).

Isolation and characterization of an unstable epoxide intermediate

Using a protocol we developed for the synthesis and instantaneous extraction of unstable intermediates, we isolated a nonpolar conjugated diene-containing product from reaction of 9R-HPODE with the Anabaena fusion protein. After conversion to the methyl ester derivative, the product was purified by SP-HPLC using a solvent containing triethylamine to help stabilize acid-sensitive molecules or the use of cold conditions as in Fig. 2A. The UV spectrum of the product showed a conjugated diene chromophore with an unusually broad top and λmax at 241 nm (Fig. 2B). The complete covalent structure was elucidated by ¹H-NMR and correlation spectroscopy analysis (Fig. 3). The trans epoxide configuration follows from the coupling constant between the two epoxide protons (J = 2 Hz) and the coupling constants across the two double bonds (J_11,12 = 15.6 Hz; J_13,14 = 15.0 Hz) establish the trans-trans configuration of the conjugated diene. Assuming that the chirality at C9 is retained as 9R, then the complete structure of the epoxide product is 9R,10R-epoxyoctadeca-11E,13E-dienoic acid.

Fig. 2.

Chromatography of the linoleate allylic epoxide. A: The sample was run on an Altex/Beckman silica column (5 μ, 4.5 × 0.46 cm) immersed in an ethanol/ice bath at approximately −10 to −15°C with a solvent of hexane/diethyl ether (100:5, v/v) at a flow rate of 3 ml/min with UV detection at 235 nm. B: UV spectrum of the linoleate allylic 9,10-epoxide prepared using the Anabaena enzyme. For comparison, the spectrum of 9-HPODE is shown as the dotted line. Solvent: hexane.

Fig. 3.

¹H-NMR spectrum of the methyl ester of the unstable epoxide formed from linoleic acid by the Anabaena peroxidase-LOX fusion protein. The inset displays the olefinic protons in more detail. The product is identified as 9R,10R-epoxyoctadeca-11E,13E-dienoate methyl ester. Solvent: d6-benzene. Chemical shift δ, multiplicity, number of protons, proton number, J: 6.38, dd, 1, H12, J_11,12 = 15.6, J_12,13 = 10.6; 6.04, dd, 1, H13, J_12,13 = 10.7, J_13,14 = 15.0; 5.59, td, 1, H14, J_13,14 = 15.1, J_14,15 = 7.1; 5.35, dd, 1, H11, J_10,11 = 8.0, J_11,12 = 15.4; 3.36, s, 3, CH₃O; 3.02, dd, 1, H10, J_9,10 = 1.3, J_10,11 = 8.0; 2.69, dt, 1, H9, J_8,9 = 6.3, J_9,10 = 1.7; 2.10, t, 2, H2; 1.95, q, 2, H15; 1.53, p, 2, H3; 1.45–1.33, m, 2, H8; 1.28–1.26, m, 4, H16,17; 1.15–1.06, m, 8, H4,5,6,7; 0.86, t, 3, H18.

Using the fusion protein, an equivalent product was also prepared from 11R-HPETE (the main product of the Anabaena LOX domain from arachidonic acid) and identified as 11R,12R-epoxyeicosa-5Z,8Z,13E,15E-tetraenoate by ¹H-NMR (see supplementary data).

Biomimetic chemical synthesis of linoleate allylic epoxide

For the chemical reaction, we used 13S-HPODE as starting material as it is readily available in multimilligram quantities using soybean LOX-1, and one of the resulting 12,13-epoxyoctadeca-8,10-dienoate products has a symmetrical epoxydiene core structure to the corresponding epoxide formed by the Anabaena enzyme from 9-HPODE. In common with the chemical reaction using HPETE to produce LTA-type epoxides (14), the chemical transformation of methyl 13-HPODE gave three products. These eluted on SP-HPLC at retention volumes of 12.6, 13.8, and 19.2 ml, respectively, in relative proportions of approximately 8:7:5. The first two peaks were identified as 12S,13S-epoxyoctadeca-8,10-dienoates with 8Z,10E and 8E,10E configurations of the conjugated diene, respectively, and the third eluting product was 13-keto-octadeca-9,11-dienoate. The UV spectra of the two epoxides differed from each other and were quite distinctive (Fig. 4A). The NMR spectra showed that both are trans epoxides (J_12,13 = 2 Hz) and allowed designation of the 8Z,10E conjugated diene of the first eluting epoxide and the 8E,10E conjugation in the second (see supplementary data). The UV spectrum and NMR spectrum of the second 12,13-epoxide were superimposable with those of the 9,10-epoxide prepared using the Anabaena enzyme. Although arising from 13-HPODE and 9-HPODE, respectively, the two epoxides are symmetrical in the epoxy-diene moiety.

Fig. 4.

UV spectra of the 12,13-epoxides and a glutathione conjugate. A: The two epoxides formed by chemical synthesis from 13-HPODE methyl ester, recorded in hexane. B: The GSH conjugate of epoxide 2, the 12,13-epoxy-octadeca-8E,10E-dienoate recorded in methanol. The spectrum of 13-HPODE is included as the dotted line in the figures for comparison.

Hydrolysis of the Anabaena 9,10-epoxide

The purified linoleate 9,10-epoxide of the Anabaena enzyme was hydrolyzed as the methyl ester derivative, yielding two major 9,14-diols and two minor 9,10-diols that matched the corresponding methyl ester derivatives of the products of the enzyme incubations. The minor 9,12-diol that chromatographs immediately after the second 9,14-diol on RP-HPLC was also found at a similar (minor) level (data not shown).

Reaction of the synthetic 12,13-epoxyoctadecadienoate with glutathione

The original description of the biomimetic synthesis of LTA₄-related epoxides included the preparation of the LTC₄ analogs, the glutathione adducts (13). We used the reported methods to react the 12,13-epoxyoctadeca-8E,10E-dienoate with GSH and obtained the predicted 13-hydroxy-12-glutathionyl adduct, purified the adduct by RP-HPLC, and recorded its distinctive UV spectrum (Fig. 4B). The UV spectrum of authentic LTC₄ shows a bathochromic shift compared with plain fatty acid conjugated trienes, whereas only the longer wavelength shoulder in the spectrum of the conjugated diene-containing glutathione adduct of the linoleate epoxide is extended to higher wavelength.

DISCUSSION

Here, we show that the product from reaction of linoleic acid or 9R-HPODE with the Anabaena peroxidase-LOX fusion protein is an unstable allylic epoxide, a novel oxylipin of enzymatic origin. The existence of this type of epoxide was invoked previously to account for the formation of 9,14-diols during autoxidation of linoleic acid and linoleate hydroperoxides (10, 20), and an epoxide was speculated to be an intermediate of the linoleate diol synthase activity of this Anabaena enzyme (2). Actually, the latter name for the enzyme is not very appropriate, although we are not offering a more correct and succinct alternative. With all the polyunsaturated fatty acids tested, the heme domain acts as an allylic epoxide synthase, although with distinctive product characteristics from the individual substrates. There are true linoleate diol synthases in which a specific diol stereoisomer is formed under enzymatic control (24, 25) but that is not the case here. Hydrolysis of the linoleate epoxide in vitro is nonenzymatic, giving a 1:1 mixture of the two major 9,14-diol diastereomers. The enzymatic end product from linoleic acid is the allylic 9,10-epoxide.

Chemical synthesis of an analog of the Anabaena allylic epoxide was achieved using 13-HPODE methyl ester as starting material (Fig. 5C). Using a method originally developed for synthesis of LTA₄ and related conjugated trienoic epoxides (13, 14), we obtained the equivalent results with C18.2 starting material. In parallel to the more detailed original report (14), the chemical transformation of linoleate hydroperoxide gave two allylic epoxides plus a keto-octadecadienoate as products. The 12,13-trans-epoxide with trans-trans conjugated diene is analogous in structure to the Anabaena 9,10-epoxide. Although both the chemical reaction and the Anabaena epoxide synthesis are initiated by activation of the hydroperoxide moiety, the enzyme maintains control of the transformation and makes a single epoxide product.

Fig. 5.

Mechanisms of transformation to allylic epoxides. A: LOX-catalyzed LTA-type epoxide biosynthesis. B: Transformation of 9R-HPODE by the Anabaena enzyme. C: Chemical biomimetic synthesis of allylic epoxide.

Allylic epoxides with conjugated trienes are well known in mammalian systems, particularly the arachidonic acid-derived 5-LOX product LTA₄, the immediate precursor of the pro-inflammatory dihydroxy LTB₄, and the bronchoconstrictor glutathionyl adduct LTC₄ (26). Transformation of the fatty acid hydroperoxide to the allylic epoxide is mechanistically quite distinct in the 5-LOX and Anabaena reactions (Fig. 5A, B). Conventional LTA₄ synthesis by 5-LOX is initiated by abstraction of a bis-allylic hydrogen from the fatty acid hydroperoxide (Fig. 5A). The hydrogen abstraction is stereospecific and associated with a primary isotope effect, both typical features of a LOX-catalyzed reaction (27–29). The presence of the third double bond in the substrate is required to facilitate reaction; this reduces the bond dissociation energy of the bis-allylic hydrogen targeted for abstraction, thus permitting its LOX-catalyzed removal. Reaction then continues with cleavage of the hydroperoxide and formation of the LTA₄ epoxide with conjugated triene. The corresponding reaction from a dienoic hydroperoxide substrate is beyond the capabilities of a LOX enzyme because the corresponding hydrogens on the carbon chain (at C14 of 9R-HPODE) have only one allylic double bond and thus are more tightly bound. We checked using the isolated LOX domain of the Anabaena fusion protein and confirmed that the polar epoxide-derived products were completely absent. We propose that transformation of 9R-HPODE by the hemoprotein domain of the Anabaena enzyme is catalyzed via homolytic cleavage of the hydroperoxide moiety, formation of a carbocation, and subsequent elimination of a proton to give the allylic epoxide as product (Fig. 5, top of middle column). This is compatible with the mechanism of conversion of 9R-hydroperoxy-linolenic acid shown earlier, which produces an allylic epoxide with bicyclobutane and an LTA-type epoxide (1).

In a cellular environment, allylic epoxides are easily hydrolyzed, but they may also react with nucleophiles, such as glutathione, a reaction that would be promoted by glutathione transferases. Such a possibility might help explain the apparent absence of detectable linoleate diols in Anabaena strains expressing the hemoprotein-LOX enzyme. The presence of 9R-HODE and 9R-HOTrE is established (2, 30), but no further derivatives of the hydroperoxides have been reported. Analysis of the metabolic fate of the epoxides described in this and our earlier report (1) might help clarify the pathways of linole(n)ic acid metabolism in Anabaena and point toward their biological role.

Abbreviations

• H(P)ETE, hydro(pero)xyeicosatetraenoic acid

• H(P)ODE, hydro(pero)xyoctadecadienoic acid

• LT, leukotriene

• LOX, lipoxygenase

• RP-HPLC, reversed-phase high-pressure liquid chromatography

• SP-HPLC, straight-phase high-pressure liquid chromatography