Hidden stereospecificity in the biosynthesis of divinyl ether fatty acids Mats Hamberg Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Sweden Fatty acid hydroperoxides generated in plant tissues by lipoxygenases or a-dioxygenases are subject to sec- ondary modification by several enzymes [1]. One branch of hydroperoxide metabolism involves forma- tion of divinyl ether derivatives by carbon–carbon bond cleavage catalyzed by specific divinyl ether synth- ases. This type of transformation was first described in 1972 by Galliard and Phillips, who found that extracts of potato tuber catalyzed the conversion of linoleic acid 9(S)-hydroperoxide (9(S)-HPOD) into a divinyl ether they named colneleic acid [2]. More recent work has demonstrated that the divinyl ether synthase pro- ducing colneleic acid and the related colnelenic acid is induced in plant leaves during attack by fungal patho- gens [3,4], and that divinyl ether fatty acids inhibit mycelial growth and spore germination in certain fungi Keywords divinyl ether synthase; double bond configuration; mechanism; stereospecifically deuterated hydroperoxides; stereospecificity of hydrogen abstraction Correspondence M. Hamberg, Department of Medical Biochemistry and Biophysics, Division of Physiological Chemistry II, Karolinska Institutet, S-171 77 Stockholm, Sweden Fax: +46 8736 0439 Tel: +46 852487640 E-mail: Mats.Hamberg@mbb.ki.se (Received 13 October 2004, revised 29 November 2004, accepted 1 December 2004) doi:10.1111/j.1742-4658.2004.04510.x Incubations of [8(R)- 2 H]9(S)-hydroperoxy-10(E),12(Z)-octadecadienoic acid, [14(R)- 2 H]13(S)-hydroperoxy-9(Z),11(E)-octadecadienoic acid and [14(S)- 2 H]13(S)-hydroperoxy-9(Z),11(E)-octadecadienoic acid were per- formed with preparations of plant tissues containing divinyl ether synth- ases. In agreement with previous studies, generation of colneleic acid from the 8(R)-deuterated 9(S)-hydroperoxide was accompanied by loss of most of the deuterium label (retention, 8%), however, the opposite result (98% retention) was observed in the generation of 8(Z)-colneleic acid from the same hydroperoxide. Formation of etheroleic acid and 11(Z)-etheroleic acid from the 14(R)-deuterated 13(S)-hydroperoxide was accompanied by loss of most of the deuterium (retention, 7–8%), and, as expected, biosynthesis of these divinyl ethers from the corresponding 14(S)-deuterated hydroper- oxide was accompanied by retention of deuterium (retention, 94–98%). Biosynthesis of x5(Z)-etheroleic acid from the 14(R)- and 14(S)-deuterated 13(S)-hydroperoxides showed the opposite results, i.e. 98% retention and 4% retention, respectively. The experiments demonstrated that biosynthesis of divinyl ether fatty acids from linoleic acid 9- and 13-hydroperoxides takes place by a mechanism that involves stereospecific abstraction of one of the two hydrogen atoms a to the hydroperoxide carbon. Furthermore, a consistent relationship between the absolute configuration of the hydrogen atom eliminated (R or S) and the configuration of the introduced vinyl ether double bond (E or Z) emerged from these results. Thus, irrespective of which hydroperoxide regioisomer served as the substrate, divinyl ether synthases abstracting the pro-R hydrogen generated divinyl ethers having an E vinyl ether double bond, whereas enzymes abstracting the pro-S hydrogen produced divinyl ethers having a Z vinyl ether double bond. Abbreviations colneleic acid, 9-[1¢(E),3¢(Z)-nonadienyloxy]-8(E)-nonenoic acid; colnelenic acid, 9-[1¢(E),3¢(Z),6¢(Z)-nonatrienyloxy]-8(E)-nonenoic acid; etheroleic acid, 12-[1¢(E)-hexenyloxy]-9(Z),11(E)-dodecadienoic acid; etherolenic acid, 12-[1¢(E),3¢(Z)-hexadienyloxy]-9(Z),11(E)-dodecadienoic acid; 9(S)- HPOD, 9(S)-hydroperoxy-10(E),12(Z)-octadecadienoic acid; 13(S)-HPOD, 13(S)-hydroperoxy-9(Z),11(E)-octadecadienoic acid. 736 FEBS Journal 272 (2005) 736–743 ª 2005 FEBS [3,5]. Further research stimulated by these findings has resulted in the cloning of genes encoding divinyl ether synthases in tomato and potato and in the identifica- tion of these enzymes as cytochrome P-450 proteins [4,6]. The divinyl ether synthase-catalyzed formation of colneleic acid from 9(S)-hydroperoxy-10(E),12(Z)-octa- decadienoic acid in homogenate of potato tuber is believed to take place via an enzyme-bound epoxide carbocation [7–10]. Studies using stereospecifically deu- terated substrate have shown that the double bond- forming step in the biosynthesis of colneleic acid involves stereoselective removal of the pro-R hydrogen from C8 [8]. Several new divinyl ether synthases in addition to the colneleic acid-forming enzyme have been discovered during the last decade [11–14], and the aim of this study was to determine the stereospecifici- ties of these additional enzymes using hydroperoxides stereospecifically deuterated at the appropriate carbon atoms. Results Stereospecifically deuterated fatty acids and hydroperoxides Stereospecifically deuterated stearates were synthesized starting with 8- or 14-hydroxystearates of high optical purity. Deuterium was introduced by reduction of the corresponding p-toluenesulfonates with lithium alumin- ium deuteride, a reaction that takes place with clean inversion of configuration [15] (Fig. 1). The deuterated stearates were converted to the corresponding linole- ates by exposure to growing cultures of the flagellate Tetrahymena pyriformis. This biological desaturation proceeds without significant degradation of the carbon chain and without migration of the isotope [16], how- ever, considerable dilution with unlabeled material took place. Incubation of the deuterated linoleates with soybean lipoxygenase or tomato lipoxygenase resulted in deuterated 13(S)-HPOD and 9(S)-HPOD, respectively, which had a deuterium content of 34– 49% (Table 1). Stereospecificity of hydrogen eliminations from C8 in the biosynthesis of colneleic acid isomers It has been shown previously that biosynthesis of colne- leic acid from 9(S)-HPOD catalyzed by divinyl ether synthase in potato tuber is accompanied by stereoselec- tive loss of the pro-R hydrogen from C8 [8]. This result was confirmed here (Table 1). An isomer of colneleic acid, i.e. 8(Z)-colneleic acid, was recently isolated from leaves of the plant Clematis vitalba [14]. Interestingly, biosynthesis of this compound from [8(R)- 2 H]9(S)- HPOD took place with retention of the deuterium label, i.e. a result opposite to that observed in the biosynthesis of colneleic acid (Table 1, Fig. 2). Stereospecificity of hydrogen eliminations from C14 in the biosynthesis of etheroleic acid isomers In the biosynthesis of etheroleic acid and its two iso- mers, i.e. x5(Z)-etheroleic acid and 11(Z)-etheroleic acid, one hydrogen is lost from C14. As seen in Table 1, these conversions also took place with stereo- specific hydrogen removals. In the case of etheroleic acid (garlic) and 11(Z)-etheroleic acid (Ranunculus Fig. 1. Reactions used to prepare stereospecifically deuterated fatty acid hydroperoxides. The following reagents ⁄ treatments were used (typical percentage yields are given within parenthesis): i, cinchoni- dine (resolution; 10%); ii, acetyl chloride followed by treatment with water ⁄ acetone (95%); iii, anodic coupling with methyl hydrogen tridecanedioate (34%); iv, NaOH in methanol ⁄ water (95%); v, CH 2 N 2 (99%); vi, p-toluenesulfonyl chloride ⁄ pyridine (86%); vii, lith- ium aluminium deuteride (82%); viii, chromic acid (90%); ix, sodium acetate (90%); x, Tetrahymena pyriformis (10%); xi, soybean lip- oxygenase (95%). R 1 ¼ (CH 2 ) 7 -COOCH 3 ,R 2 ¼ (CH 2 ) 7 -COOH. M. Hamberg Stereospecificity of divinyl ether synthases FEBS Journal 272 (2005) 736–743 ª 2005 FEBS 737 lingua) an essentially complete loss of isotope was noted with the 14(R)-deuterated precursor, whereas incubation of the 14(S)-deuterated precursor afforded products that retained most of the deuterium label. The opposite labeling pattern was observed in the bio- synthesis of x5(Z)-etheroleic acid (Ranunculus acris), i.e. stereospecific removal of the pro-S hydrogen and retention of the pro-R hydrogen (Table 1, Fig. 2). Discussion Since the pioneering work by Schroepfer and Bloch on the stereochemistry of the desaturation of stearate into oleate [17], studies using stereospecifically deuterated or tritiated precursors have generated important infor- mation on enzyme-catalyzed reactions. In addition to providing insights into reaction mechanisms, results obtained with specifically labeled substrates have been useful in molecular modeling studies of enzyme–sub- strate complexes. An example from the oxylipin ⁄ eicos- anoid field is the use of precursor acids labeled with tritium in the 13(R) or 13(S) positions to elucidate the stereochemistry and mechanism of the cyclooxygenase reaction leading to prostaglandins [16], and the use of this knowledge for establishing the productive confor- mation of the arachidonic acid molecule bound to the active site of the cyclooxygenase enzyme [18]. Divinyl ether synthases are cytochrome P-450 enzymes [4,6] and related to other (hydro)peroxide- metabolizing P-450s such as allene oxide synthase [19], hydroperoxide lyase [20,21], thromboxane synthase [22] and prostacyclin synthase [22]. In the case of divi- nyl ether biosynthesis, the initial step is believed to consist of cleavage of the O-O bond of the hydroper- oxide to provide an alkoxy radical, which undergoes cyclization and one-electron oxidation into an epoxide carbocation [23]. Enzyme-assisted removal of a pro- ton a to the epoxide group and cleavage of the car- bon-carbon single bond of the epoxide group provides the final divinyl ether structure (Fig. 3). The nature of the hydrogen-abstracting group in divinyl ether synth- ases is unkown but may be either a basic amino acid residue [7] or the strongly basic Fe III -OH group of the P-450 heme [23]. This study confirms and extends previous work on the divinyl ether synthase class of cytochrome P-450s by showing that removal of one of the two hydro- gens a to the hydroperoxide group invariably takes place in a stereospecific way (Fig. 2). Furthermore, inspection of the data given in Table 1 and Fig. 2 Table 1. Isotope composition of deuterated fatty acid hydroperox- ides and divinyl ether fatty acids. Isotopic analysis was carried out with selected ion monitoring mass spectrometry using the ions m ⁄ z 225 and 226 (reduced[14– 2 H]13(S)-HPOD), m ⁄ z 311 and 312 (reduced[8- 2 H]9(S)-HPOD), and m ⁄ z 308 and 309 (divinyl ether fatty acids). Compound Monodeuterated molecules (%) Retention of deuterium (%) [14(R)- 2 H]13(S)-HPOD 38.3 100 Etheroleic acid 2.8 7 x5(Z)-Etheroleic acid 37.4 98 11(Z)-Etheroleic acid 3.2 8 [14(S)- 2 H]13(S)-HPOD 33.9 100 Etheroleic acid 31.8 94 x5(Z)-Etheroleic acid 1.5 4 11(Z)-Etheroleic acid 33.3 98 [8(R)- 2 H]9(S)-HPOD 48.8 100 Colneleic acid 4.0 8 8(Z)-Colneleic acid 48.0 98 Fig. 2. Stereospecificities of five divinyl ether synthases. R 1 ¼ (CH 2 ) 6 -COOH, R 2 ¼ (CH 2 ) 7 -COOH. DES, divinyl ether synthase. Fig. 3. Proposed sequence of reactions in the biosynthesis of divi- nyl ether fatty acids. R 1 ¼ C 5 H 11 and R 2 ¼ (CH 2 ) 6 -COOH (colneleic acid series), or R 1 ¼ (CH 2 ) 7 -COOH and R 2 ¼ C 4 H 9 (etheroleic acid series). Adapted from Grechkin [23]. Stereospecificity of divinyl ether synthases M. Hamberg 738 FEBS Journal 272 (2005) 736–743 ª 2005 FEBS reveals an interesting correlation, i.e. biosynthesis of divinyl ethers having the E configuration of the newly created vinyl ether double bond (colneleic acid, ethero- leic acid and 11(Z )-etheroleic acid) takes place with stereoselective loss of the pro-R hydrogen, whereas the Z vinyl ethers (8(Z)-colneleic acid and x5(Z)-etheroleic acid) are formed in a process which involves loss of the pro-S hydrogen. These results can be intepreted in terms of the conformations of the carbon–carbon sin- gle bond a to the epoxide of the epoxide carbocation intermediate, i.e. transoid and cisoid conformations are needed to produce E and Z vinyl ether double bonds, respectively (Fig. 4). As seen from this model, irres- pective of the detailed structure of the surrounding act- ive site, rotation of the carbon–carbon single bond to produce the two conformations moves either the pro-R or pro-S hydrogen in contact with the same region of the active site. This may be taken to suggest that the positioning of the hydrogen-abstracting group relative to the bound substrate is highly conserved in all divi- nyl ether synthases of higher plants. The stereochemi- cal data also show that the hydrogen eliminated consistently has a syn relationship to the vicinal oxy- gen atom (Fig. 4). This stereochemistry is in agreement with the notion [23] that the heme iron not only parti- cipates in the hydroperoxidase reaction but also serves as the hydrogen-abstracting group. Further interpret- ation of the stereochemical data presented must await access of the three-dimensional structures of the divi- nyl ether synthase P-450s. Experimental procedures Plant materials Specimens of Ranunculus acris L., Ranunculus lingua L and Clematis vitalba L. were obtained as described previously [12–14]. Leaves were either used directly or shock-frozen in liquid nitrogen and stored at )80 °C until use. Tubers of potato (var. Bintje) and bulbs of garlic were obtained from a local market. Stereospecifically deuterated hydroperoxides [8(R)- 2 H]9(S)-Hydroperoxy-10(E),12(Z)-octadecadienoic acid 9(S)-HPOD stereospecifically labeled with deuterium in the 8(R) position was prepared via [8(S)- 2 H]stearic acid (25 mg) and [8(R)- 2 H]linoleic acid as described previously [8]. An aliquot of the hydroperoxide was reduced by treatment with sodium borohydride in methanol at 0 °C, and the methyl ester ⁄ Me 3 Si derivative of the resulting hydroxide was analyzed by GC ⁄ MS. The isotopic composition of the sample was 48.8% monodeuterated and 51.2% undeute- rated molecules as determined by mass spectrometric analysis of the fragment [CH 3 OOC-(CH 2 ) 7 -(CH ¼ CH) 2 -CH ¼ O + SiMe 3 ](m ⁄ z 311 and 312 in the undeuterated and deuterated derivatives, respectively). As expected from the localization of the deuterium atom at C8, the fragment [(CH ¼ CH) 2 -CH(OSiMe 3 )-(CH 2 ) 4 -CH 3 ] + (m ⁄ z 225) was devoid of deuterium. 3(R,S)-Hydroxyheptanoic acid Methyl 3-oxoheptanoate (39.5 g; 0.25 mmol; Fluka Chemie GmbH, Buchs, Switzerland) was dissolved in methanol (250 mL) and sodium borohydride (4 g) was added at 0 °C over a period of 3 h under magnetic stirring. Subsequently, a solution of sodium hydroxide (12 g) in water (100 mL) was added and the mixture was stirred for 15 h at 23 °C. Extraction with diethyl ether provided 3(R,S)-hydroxyhept- anoic acid (36.1 g; 99%) as a colorless viscous oil which slowly solidified at room temperature. The purity as checked by GC ⁄ MS analysis of the methyl ester ⁄ Me 3 Si derivative was > 95%. 3(R)-Hydroxyheptanoic acid 3(R,S)-Hydroxyheptanoic acid (14.6 g, 0.1 mmol) and cinchonidine (29.4 g, 0.1 mmol) in carbon tetrachloride (1 L) was heated on a boiling water bath for 3 min and then allowed to cool to room temperature. The crystalline Fig. 4. Conformations and hydrogen abstractions in the final step of divinyl ether fatty acid biosynthesis. (A) Divinyl ether synthases from potato, garlic or Ranunculus lingua introducing an ‘E’ vinyl ether double bond, (B) divinyl ether synthases from Ranunculus acris or Clematis vitalba introducing a ‘Z ’ vinyl ether double bond. R 1 ¼ (CH 2 ) 6 -COOH and R 2 ¼ C 5 H 11 (colneleic acid series) or R 1 ¼ C 4 H 9 and R 2 ¼ (CH 2 ) 7 -COOH (etheroleic acid series). ‘B’ attached to the enzyme surfaces, base. M. Hamberg Stereospecificity of divinyl ether synthases FEBS Journal 272 (2005) 736–743 ª 2005 FEBS 739 mass that separated was collected on a Bu ¨ chner funnel and redissolved in CCl 4 (700 mL). The solution was left over- night at 23 °C and the crystals formed ( 20 g) were again subjected to crystallization from CCl 4 . After six such crys- tallizations, rosette-formed crystals (2.5 g) of the cinchoni- dine salt of 3(R)-hydroxyheptanoic acid were obtained. Regeneration of the acid by acidification and extraction with diethyl ether provided 3(R)-hydroxyheptanoic acid (0.73 g; yield, 10% of the theoretical) having [a] D 23 ¼ )23.3° (c 2.5, chloroform); earlier published for 3(R)- hydroxyhexanoic acid, [a] D ¼ )28° [24] and for 3(R)- hydroxyoctanoic acid, [a] D ¼ )21° [25]. An aliquot was esterified with diazomethane and converted to its 2(S)- phenylpropionyl derivative [26]. Analysis by GLC showed an enantiomeric composition of > 98% 3(R)- and < 2% 3(S)-hydroxyheptanoic acid. Methyl 14(R)-hydroxystearate 3(R)-Hydroxyheptanoic acid (438 mg, 3 mmol) was refluxed for 15 min with acetyl chloride (20 mL). The residue obtained following evaporation of the reagent was dissolved in acetone (20 mL), and water (13.3 mL) was added under magnetic stirring. After 5 h at 23 °C, the solution was extracted providing essentially pure 3(R)-acetoxyheptanoic acid. This material was dissolved in methanol (100 mL) containing methyl hydrogen tridecane-1,13-dioate (5.16 g, 20 mmol) and sodium methoxide (1.2 mmol). The solution was transferred to an electrolysis cell and a current of 1.8 A was passed through for 2 h. The resulting product was saponified, esterified by treatment with diazomethane, and subjected to silicic acid open column chromatography. Elution with diethyl ether ⁄ hexane (20 : 80, v ⁄ v) afforded methyl 14(R)-hydroxyoctadecanoate (299 mg; yield, 32%). Analysis by GC ⁄ MS showed a single peak (purity, > 96%) on which a mass spectrum showing the following prominent ions was recorded: m ⁄ z 296 (M + – 18; loss of H 2 O), 283 (M + – 31; loss of OCH 3 ), 257 (M + – 57; loss of (CH 2 ) 3 - CH 3 ), 225 (257–32; loss of CH 3 OH), 185, 143, 87, and 69. Methyl 14(R)-p-toluenesulfonyloxystearate Methyl 14(R)-hydroxystearate (157 mg, 0.5 mmol) was dis- solved in dry pyridine (4 mL), cooled to )25 °C and treated with p-toluenesulfonyl chloride (400 mg). After 12 h at )25 °C and 48 h at +4 °C, water was added and the solu- tion was extracted with diethyl ether. Purification by open column silicic acid chromatography afforded methyl 14(R)- p-toluenesulfonyloxystearate (200 mg, yield, 86%). [14(S)- 2 H]Stearic acid Methyl 14(R)-p-toluenesulfonyloxystearate (100 mg, 0.21 mmol) was refluxed with lithium aluminium deuteride (300 mg; 98 atom percentage deuterium; purchased from Sigma-Aldrich) in dry tetrahydrofuran (25 mL) for 18 h. The resulting octadecanol was dissolved in 6 mL of acetone and treated with 0.63 mL of Jones’ reagent at 23 °C for 30 min. The product was purified by open column silicic acid chromatography to afford [14(S)- 2 H]stearic acid (44 mg, yield, 74%). Analysis of an aliquot (methyl ester) by GC ⁄ MS showed a single peak. The mass spectrum showed the following prominent ions: m ⁄ z 299 (M + ), 256 (M + – 43; loss of C 3 H 7 ), 200 (M + – 99; loss of C 7 H 15 ), 143, 87, and 74. [14(S)- 2 H]Linoleic acid A culture of Tetrahymena pyriformis strain phenoset A (American Type Culture Collection #30327; Manassas, VA, USA) was added to culture medium (400 mL) consisting of glucose (0.5%, w ⁄ v), yeast extract (0.5%, w ⁄ v) and peptone (0.5%, w ⁄ v) in 0.004 m potassium phosphate buffer pH 7.0 and containing the sodium salt of [14(S)- 2 H]stearic acid (10 mg). The mixture was incubated under continuous sha- king at 32 °C for 92 h [8]. The cell pellet collected by cen- trifugation was suspended in 50% aqueous methanol (100 mL) containing sodium hydroxide (7 g) and the mix- ture was refluxed under an atmosphere of argon for 90 min. The isolated mixed fatty acids ( 9 mg) were subjected to semipreparative RP-HPLC using a column of Nucleosil C 18 100-7 (250 · 10 mm) purchased from Macherey-Nagel (Du ¨ ren, Germany) and a solvent system of acetonitrile ⁄ water ⁄ acetic acid (800 : 200 : 0.1, v ⁄ v ⁄ v) at 3mLÆmin )1 . [14(S)- 2 H]Linoleic acid (2 mg) was collected at 68–72 mL effluent. The mass spectrum of the methyl ester of this material showed molecular ions at m ⁄ z 294 and 295 corresponding to undeuterated and monodeuterated mole- cules, respectively (ratio, 0.51), the undeuterated linoleate being derived from the organism used to carry out the desaturation. [14(S)- 2 H]13(S)-Hydroperoxy-9(Z),11(E)-octadecadi- enoic acid [14(S)- 2 H]Linoleic acid (1.9 mg) was stirred for 10 min at 0 °C in 0.1 m sodium borate buffer pH 10.4 containing soy- bean lipoxygenase (soybean lipoxygenase type IV purchased from Sigma-Aldrich, 15 lL). Material isolated by extraction with diethyl ether was purified by silicic acid open column chromatography to provide 1.7 mg (yield, 80%) of the title compound showing k max (EtOH) ¼ 233 nm. An aliquot was reduced with NaBH 4 in methanol at 0 °C and the resulting deuterated 13-hydroxy-9(Z),11(E )-octadecadieno- ate was analyzed as the methyl ester ⁄ Me 3 Si derivative by GC ⁄ MS using selected ion monitoring of the mass spectral ions due to the fragment [(CH ¼ CH) 2 -CH(OSiMe 3 )- (CH 2 ) 4 -CH 3 ] + (m ⁄ z 225 and 226 in undeuterated and Stereospecificity of divinyl ether synthases M. Hamberg 740 FEBS Journal 272 (2005) 736–743 ª 2005 FEBS deuterated hydroxides, respectively). The isotopic composi- tion found after correction for the natural abundance of the m ⁄ z 226 ion was 33.9% monodeuterated and 66.1% un- deuterated molecules. As expected from the localization of the deuterium label at C14, the fragment CH 3 OOC-(CH 2 ) 7 - (CH ¼ CH) 2 -CH ¼ O + SiMe 3 was undeuterated and gave rise to an ion at m ⁄ z 311 with insignificant enrichment of m ⁄ z 312. [14(R)- 2 H]Stearic acid Methyl 14(R)-p-toluenesulfonyloxystearate (100 mg, 0.21 mmol) was dissolved in acetic acid (7 mL) containing sodium acetate (50 mg) and the mixture was kept at 60 °C for 17 h. The product was saponified by refluxing with 10% NaOH in 80% aqueous methanol for 90 min and then esterified by treatment with diazomethane and treated with p-toluenesulfonyl chloride in pyridine as described above. Reduction with lithium aluminium deuteride followed by Jones oxidation and silicic acid open column chromato- graphy afforded [14(R)- 2 H]stearic acid (22 mg; yield, 37%) (Fig. 1). [14(R)- 2 H]Linoleic acid [14(R)- 2 H]Stearic acid (22 mg) was incubated with Tetra- hymena pyriformis as described above and the free fatty acids subjected to RP-HPLC to provide [14(R)- 2 H]linoleic acid (4 mg; ratio of labeled ⁄ unlabeled molecules, 0.62). [14(R)- 2 H]13(S)-Hydroperoxy-9(Z),11(E)-octadecadi- enoic acid [14(R)- 2 H]Linoleic acid (4 mg) was incubated with soybean lipoxygenase as described above. Following purification by silicic acid open column chromatography, [14(R)- 2 H]13(S)- hydroperoxy-9(Z),11(E)-octadecadienoic acid (3 mg) was obtained. The isotopic composition as determined by GC ⁄ MS analysis of the methyl ester⁄ Me 3 Si derivative of the reduced compound was 38.3% monodeuterated and 61.7% undeuterated molecules. Enzyme preparations The following divinyl ether synthase preparations were used for study of the biosynthesis of the divinyl ethers indicated. Colneleic acid Tubers of potato were sliced and homogenized at 0 °Cin 0.1 m borate buffer pH 9.0 (2 : 1, v ⁄ w) using an Ultra-Tur- rax. The homogenate was filtered through gauze and centri- fuged at 9300 g for 15 min. Further centrifugation at 105 000 g provided a particulate fraction which was resus- pended in borate buffer (half the volume compared with the corresponding 105 000 g supernatant) [8]. Etheroleic acid Bulbs of garlic were sliced and homogenized at 0 °Cin 0.1 m potassium phosphate buffer pH 8.0 (2 : 1, v ⁄ w). The homogenate was filtered through gauze and successively centrifuged at 1100 g for 15 min and 105 000 g to provide a particulate fraction which was resuspended in phosphate buffer (half the volume compared with the corresponding 105 000 g supernatant) [11]. x5(Z)-Etheroleic acid Leaves of Ranunculus acris were minced and homogenized at 0 °C in 0.1 m potassium phosphate buffer pH 6.7 (5 : 1, v ⁄ w; buffer supplemented with 235 lm of salicylhydroxamic acid as an lipoxygenase inhibitor). The filtered homogenate was centrifuged at 1100 g for 15 min, and further centrifu- gation of the supernatant at 105 000 g provided a particu- late fraction that was resuspended in phosphate buffer (half the volume compared with the corresponding supernatant) [12]. 11(Z)-Etheroleic acid Leaves of Ranunculus lingua were treated in the same way as described for Ranunculus acris [13]. 8(Z)-Colneleic acid Leaves of Clematis vitalba were homogenized at 0 °Cin potassium phosphate buffer pH 6.7 (5 : 1, v ⁄ w) and the homogenate was filtered through gauze [14]. Incubations and isolation of products Suspensions of the particulate fractions of homogenates of garlic, R. acris or R. lingua (2–20 mL) were stirred at 23 °C for 20 min with 300 lm [14(R)- 2 H]- or [14(S)- 2 H]13(S)- HPOD. In the same way, filtered homogenates of C. vitalba (45 mL) or suspensions of the 105 000 g fraction of potato tuber homogenate were stirred with 300 lm [8(R)- 2 H] 9(S)-HPOD. Material obtained after extraction with diethyl ether was subjected to solid-phase extraction using an amino- propyl column (0.5 g; Supelco, Bellefonte, PA, USA) [12]. Material eluted with diethyl ether ⁄ acetic acid (98 : 2, v ⁄ v) was esterified by treatment with diazomethane and subjected to RP-HPLC using a column of Nucleosil C 18 100–7 (250 · 10 mm) and a solvent system of acetonitrile ⁄ water (80 : 20, v ⁄ v) at a flow rate of 4 mLÆmin )1 . The effluent was led to a Bischoff model DAD-100 diode-array detector (Bis- choff Chromatography, Leonberg, Germany), and divinyl M. Hamberg Stereospecificity of divinyl ether synthases FEBS Journal 272 (2005) 736–743 ª 2005 FEBS 741 ethers localized by their strong absorption at 250–253 nm were collected, esterified, and analyzed for deuterium content by GC ⁄ MS. Blank incubations were performed in which the divinyl ether synthase preparations were incubated in the absence of hydroperoxide and subsequently carried through the whole sequence. In the case of the enzyme preparation obtained from garlic bulbs, significant amounts of endo- genous etheroleic acid were detected, i.e. 3–4% of the levels achieved in incubations carried out in the presence of added 13(S)-HPOD. The isotopic compositions of etheroleic acid biosynthesized from deuterated precursors using this enzyme preparation were corrected for the dilution caused by this endogenous material. By contrast, no significant occurrence of endogenous divinyl ethers was detected in the other four preparations. GC / MS GC ⁄ MS was carried out with a Hewlett-Packard model 5970B mass-selective detector connected to a Hewlett-Pack- ard model 5890 gas chromatograph equipped with a capil- lary column of 5% phenylmethylsiloxane (12 m, 0.33 lm film thickness). Helium was used as the carrier gas, and the column temperature was raised at 10 °CÆmin )1 from 120 to 260 °C. Isotopic composition of analytes were determined using the selected ion monitoring mode and the following mass spectral ions: m ⁄ z 225 and 226 (undeuterated and monodeuterated [14- 2 H]13(S)-HPOD; reduced, methyl este- rified and trimethylsilylated), m ⁄ z 311 and 312 ([8- 2 H]9(S)- HPOD; reduced, methyl esterified and trimethylsilylated), and m ⁄ z 308 and 309 (methyl esters of divinyl ether fatty acids). Acknowledgements Mrs Gunvor Hamberg is thanked for expert technical assistance and for collection and identification of the plant materials used. This work was supported by a generous grant given by the late Professor Sune Bergs- tro ¨ m, Stockholm, and by grants from the Swedish Research. Council for Environment, Agricultural Sciences and Spatial Planning (project no. 2001-2553) and the European Union (project No. QLK5-CT-2001- 02445). References 1 Feussner I & Wasternack C (2002) The lipoxygenase pathway. Annu Rev Plant Biol 53, 275–297. 2 Galliard T & Phillips DR (1972) The enzymic conver- sion of linoleic acid into 9-(nona-1¢,3¢-dienoxy)non- 8-enoic acid, a novel unsaturated ether derivative isolated from homogenates of Solanum tuberosum tubers. Biochem J 129, 743–753. 3 Weber H, Che ´ telat A, Caldelari D & Farmer EE (1999) Divinyl ether fatty acid synthesis in late blight-diseased potato leaves. Plant Cell 11, 485–493. 4 Stumpe M, Kandzia R, Go ¨ bel C, Rosahl S & Feussner I (2001) A pathogen-inducible divinyl ether synthase (CYP74D) from elicitor-treated potato suspension cells. FEBS Lett 507, 371–376. 5 Grane ´ r G, Hamberg M & Meijer J (2003) Screening of oxylipins for control of oilseed rape (Brassica napus) fungal pathogens. Phytochemistry 63 , 89–95. 6 Itoh A & Howe GA (2001) Molecular cloning of a divi- nyl ether synthase. Identification as a CYP74 cyto- chrome P-450. J Biol Chem 276, 3620–3627. 7 Crombie L, Morgan DO & Smith EH (1987) The enzy- mic formation of colneleic acid, a divinyl ether fatty acid: experiments with [(9 S )- 18 O 2 ]hydroperoxyoctadeca- (10E),(12Z)-dienoic acid. J Chem Soc Chem Commun 502–503. 8 Fahlstadius P & Hamberg M (1990) Stereospecific removal of the pro-R hydrogen at C-8 of 9(S)-hydroper- oxyoctadecadienoic acid in the biosynthesis of colneleic acid. J Chem Soc Perkin Trans 1, 2027–2030. 9 Corey EJ, Nagata R & Wright SW (1987) Biomimetic total synthesis of colneleic acid and its function as a lipoxygenase inhibitor. Tetrahedron Lett 28, 4917– 4920. 10 Gerwick WH (1996) Epoxy allylic carbocations as con- ceptual intermediates in the biogenesis of diverse marine oxylipins. Lipids 31, 1215–1231. 11 Grechkin AN, Fazliev FN & Mukhtarova LS (1995) The lipoxygenase pathway in garlic (Allium sativum L.) bulbs: detection of the novel divinyl ether oxylipins. FEBS Lett 371, 159–162. 12 Hamberg M (1998) A pathway for biosynthesis of divi- nyl ether fatty acids in green leaves. Lipids 33, 1061– 1071. 13 Hamberg M (2002) Biosynthesis of new divinyl ether oxylipins in Ranunculus plants. Lipids 37, 427–433. 14 Hamberg M (2004) Isolation and structures of two divi- nyl ether fatty acids from Clematis vitalba. Lipids 39, 565–569. 15 Helmkamp GK & Rickborn BF (1957) Stereochemistry of some lithium aluminium deuteride reductions. J Org Chem 22, 479–482. 16 Hamberg M & Samuelsson B (1967) On the mechanism of the biosynthesis of prostaglandins E 1 and F 1a . J Biol Chem 242, 5336–5343. 17 Schroepfer GJ & Bloch K (1965) The stereospecific con- version of stearic acid to oleic acid. J Biol Chem 240, 54–63. 18 Malkowski MG, Ginell SL, Smith WL & Garavito RM (2000) The productive conformation of arachidonic acid bound to prostaglandin synthase. Science 289, 1933– 1937. Stereospecificity of divinyl ether synthases M. Hamberg 742 FEBS Journal 272 (2005) 736–743 ª 2005 FEBS 19 Song W-C & Brash AR (1991) Purification of an allene oxide synthase and identification of the enzyme as a cytochrome P-450. Science 253, 781–784. 20 Matsui K, Shibutani M, Hase T & Kajiwara T (1996) Bell pepper fruit fatty acid hydroperoxide lyase is a cytochrome P-450 (CYP74B). FEBS Lett 394, 21–24. 21 Grechkin AN & Hamberg M (2004) The ‘heterolytic hydroperoxide lyase’ is an isomerase producing a short- lived hemiacetal. Biochim Biophys Acta 1636, 47–58. 22 Hecker M & Ullrich V (1989) On the mechanism of prostacyclin and thromboxane A 2 biosynthesis. J Biol Chem 264, 141–150. 23 Grechkin AN (2002) Hydroperoxide lyase and divinyl ether synthase. Prostaglandins Other Lipid Med 68–69, 457–470. 24 Serck-Hanssen K (1956) Optically active higher aliphatic hydroxy-compounds. Arkiv Kemi 10, 135–149. 25 Lemieux RU & Giguere J (1951) Biochemistry of ustal- gines. Can J Chem 29, 678–690. 26 Hammarstro ¨ m S & Hamberg M (1973) Steric analysis of 3-, x4-, x3- and x2-hydroxy acids and various alkan- ols by gas–liquid chromatography. Anal Biochem 52, 169–179. M. Hamberg Stereospecificity of divinyl ether synthases FEBS Journal 272 (2005) 736–743 ª 2005 FEBS 743 . interesting correlation, i.e. biosynthesis of divinyl ethers having the E configuration of the newly created vinyl ether double bond (colneleic acid, ethero- leic. preparations The following divinyl ether synthase preparations were used for study of the biosynthesis of the divinyl ethers indicated. Colneleic acid Tubers of